ADVANCES IN
Immunology EDITED BY FRANK J. DIXON Scripps Clinic and Research Foundation La Jolla, California
ASSOCIATE ...
13 downloads
971 Views
22MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
ADVANCES IN
Immunology EDITED BY FRANK J. DIXON Scripps Clinic and Research Foundation La Jolla, California
ASSOCIATE EDITORS
K. FRANKAUSTEN LEROYE. HOOD JONATHAN W. UHR
VOLUME 48
ACADEMIC PRESS, INC. Harcourt Brace Jovanovich, Publishes San Diego New York Boston London Sydney Tokyo Toronto
This book is printed on acid-free paper.
@
COPYRIGHT 0 1990 BY ACADEMIC PRESS, INC. All Rights Reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher.
ACADEMIC PRESS,INC. San Diego, California 92101
United Kingdom Edition published by ACADEMIC PRESS LIMITED 24-28 Oval Road, London NWI 7DX
LIBRARY OF CONGRESS CATALOG CARD NUMBER: 61-17057
ISBN 0-12-022448-8 (alk. paper)
PRINTED IN THE UNITED STATES OF A M W C A 90 91 92 93
9
8
7 6 5 4 3 2
I
ADVANCES I N IMMUNOLOGY, VOL. 48
Internal Movements in Immunoglobulin Molecules ROALD NEZLIN Department of Chemical Immunology, The Weizmann Institute of Science, Rehovot 76100, Israel
I. Introduction
Protein molecules are dynamic structures, exhibiting a variety of internal motions, ranging from atom fluctuations and oscillations of amino acid side groups to the movement of large portions of the molecules, such as their domains and subunits (Karplus and McCammon, 1983). This mobility has important functional implications; for example, large scale motion is significant for the regulation of catalytic activity or for the independent movement of subunits, allowing them to react with different ligands (Huber and Bennett, 1983). The latter motion is especially important in the case of multifunctional molecules, in which each type of activity links with a particular subunit. Immunoglobulins (Ig’s)are among the most thoroughly studied multifunctional protein molecules. They unite the recognition functionnamely, the ability to form complexes with a vast array of antigen molecules-with effector functions, such as complement activation and the capacity to react with different cell receptors. The functional sites responsible for all of these reactions are located in various domains, linked together with greater or lesser, flexible segments of peptide chains. T h e hypothesis that Ig’s are flexible molecules was proposed about 25 years ago, based on experiments performed using classical hydrodynamic methods (Noelken et al., 1965) and by electron-microscopic observations (Feinstein and Rowe, 1965; Valentine and Green, 1967). Since these are indirect techniques for the study of molecular dynamics, and considering the novelty of the ideas at that time, tribute should be paid to these scientists. In the years following, a complete arsenal of physicochemical methods was implemented to study the different types of internal motion in Ig molecules. Each new approach o r modification of classical methodology was used by different groups at once, and over the last few decades interest in the problem has remained at a constantly high level. As a result, many new concepts and discoveries in the field of immunology 1 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
2
ROALD NEZLIN
have been applied by investigators interested in Ig flexibility. When myeloma proteins were recognized as homogeneous Ig’s, they were widely implemented in experimentation in this field. Later, monoclonal antibodies were successfully used, in particular, those against different reporter groups, such as fluorescent dyes or spin labels. Most recently, due to the enormous progress in genetic engineering, artificially constructed chimeric antibodies have become the object of investigations of the structural basis for flexibility and the functional importance of the different modes of subunit movements. T h e purpose of this chapter is to summarize the various studies of the movement of Ig subunits (i.e., segmental flexibility) as well as internal motion in the subunits themselves, and to discuss the functional implications of these motions. II. Methods for the Study of Ig Flexibility
A. HYDRODYNAMIC METHODS Classic hydrodynamic methods-such as sedimentation analysis, determination of the frictional coefficient ratio, and viscosity measurements-provide valuable information regarding the compactness of protein molecules. If the sedimentation coefficients and the molecular weights are known, it then becomes possible to calculate the frictional coefficient ratio (f :fmin),which characterizes both hydration and shape. For typical globular proteins the values of the ratio are 1.10- 1.25. Larger values may signal molecular asymmetry, as well as the possibility that certain regions of the polypeptide chains are flexible. Similarly, the viscosity data provide information which helps to reveal whether the investigated molecule is close to typical globular proteins. If the isolated subunits of a protein molecule resemble globular proteins, but the whole molecule does not, we can assume that the structure of the latter is not compact (Noelken et al., 1965). B. ELECTRON MICROSCOPY Contemporary electron microscopy enables recognition of the finest structural details of protein molecules (Green, 1969). Despite the fact that this technique is limited to the study of dried and fixed protein preparations, valuable information on the dynamic structure of protein molecules has been gained. For example, if we are able to alter the arrangement of some part of a molecule, we may assume that this molecule does not have a rigid, but has a mobile, conformation. Indeed, it was found that, after reaction with antigens, antibody configuration can
INTERNAL MOVEMENTS I N
Ig MOLECULES
3
change significantly and electron micrographs clearly differentiate the changes in the relative positions of antibody subunits (Feinstein and Rowe, 1965; Valentine and Green, 1967). C. X-RAYCRYSTALLOGRAPHY The most powerful method of obtaining high-resolution structural pictures of protein molecules is by X-ray crystallography, which enables us to determine the average position of atoms in a protein crystal. In addition, important information regarding the dynamic properties of protein molecules can also be derived from X-ray diffraction experiments (Bennett and Huber, 1984). First, if one or another part of the protein molecule is disordered-that is, it has a poorly defined electron density-we can assume that the corresponding subunit has some freedom of rotation and that the molecule under investigation is flexible. However, this assumption is correct only if the observed variation in electron density is unrelated to methodological problems, for example, peculiarities in crystal packing. Second, in some cases a protein can form different, but closely related, types of crystals. In each instance the molecule has a different general conformation-for example, a variation in the angle between subunits. Such an observation may be, primarily, the result of relatively free mobility of the subunits, which permits them to accommodate different positions during crystal formation. Third, the lack of close contact between subunits provides further indirect evidence that they could have greater or lesser freedom of rotation. Finally, difficulties in generating the crystallization of a protein molecule can also provide evidence of its flexible structure. D. FLUORESCENCE POLARIZATION Fluorescence polarization, introduced by Weber (1953), is the most direct technique for investigating the rotational properties of protein molecules. Initially, a protein molecule is labeled using a fluorescent dye, either covalently or specifically, by combination with the combining site of the corresponding antibody. A solution of labeled protein is then excited by vertically polarized light, which results in a set of excited dye molecules becoming arranged primarily in the vertical plane. Due to Brownian thermal motion, the initial orientation of the excited dye molecules becomes randomized. Depolarization of the emitted light gives information regarding the size and shape of the macromolecule, as well as its flexibility. The rotational correlation time (4 or T ) , ~which can be calculated from such experiments, provides information about the rota-
' + is used in the fluorescencepolarizationexperiments;
T,
in the spin-label experiments.
4
KOALD NEZLIN
tional properties of the macromolecule under investigation. If the experimental values of the rotational relaxation time are significantly lower than those calculated, assuming the rigidity of the macromolecule, the macromolecule is flexible and its separate parts are capable of more or less independent motion. Two types of fluorescent depolarization measurements have been used: steady state (Weber, 1953) and nanosecond, o r time resolved (Stryer, 1968; Yguerabide, 1972). Reliable data on the rotational correlation times can be obtained using the steady-state approach if the following considerations are taken into account: First, to calculate the rotational correlation time, it is necessary to estimate the values of the fluorescence lifetime of the bound dye molecule. This parameter has to be measured for each labeled protein, since there are significant differences for the various conjugates. If the macromolecule has several rotational correlation times related to the different types of motion, then in the experiments using short-lived dyes, the higher values of the rotational correlation time could be not determined. Second, fluorescence depolarization could not be assumed to be due solely to Brownian motion of the macromolecule, but could also be due to independent thermally activated rotation of the covalently linked dye molecule. This problem, however, can be partially overcome, and satisfactory results have been obtained. For example, values of the rotational correlation time found for the Fab fragment in experiments with a covalently bound label, o r with a label firmly bound in the antibody combining site, coincide well (Tumerman et al., 1972a,b).The steady-state method, however, has some limitations: The method is limited to determination of the averaged rotational correlation time only. Further, in order to determine the rotational correlation time by this technique, it is necessary to perform a series of measurements, in which increasing amounts of sucrose are added to the protein solution, to extrapolate the average anisotropy to infinite viscosity. The nanosecond pulse technique permits direct measurement of the rotational correlation time from a single experiment. The timedependent anisotropy curves for whole Ig molecules obtained in these experiments are curved, showing the presence of two main depolarizing motions with two correlation times. In initial experiments performed using the nanosecond fluorescence technique (Yguerabide et al., 1970), the dansyl dye was firmly bound in the combining sites of antidansyl antibodies; thus, the independent movement of the dye molecules was prevented. This method was later refined, and previously obtained data were reinterpreted (Hanson et al., 1981). Further progress was made by using homogeneous monoclonal anti-dansyl antibodies, which eliminates the problem of random orientation of the dye molecules within the
INTERNAL MOVEMENTS I N
Ig MOLECULES
5
combining sites when heterogeneous antibodies are used; however, antidye antibodies cannot be used in all experiments. Another problem is linked with the difficulty of interpreting the results obtained in real physical terms. E. SPIN-LABEL APPROACH The spin-label method introduced by McConnell was used relatively recently to determine the rotational relaxation time. This method is based on the fact that the distance between the outer wide extrema of electron spin resonance (ESR) spectra is related quantitatively to the rotational correlation motion. Shimshick and McConnell(l972) used this dependence when the spin label has no motional freedom relative to the macromolecule. Initially, the theoretical dependence of the shift of outer wide extrema (2A’) as a function of the rotational correlation time is determined; then the shift of outer wide extrema as a function of viscosity is determined experimentally and extrapolated to infinite viscosity. The difference between the position of extrema for a solution of infinite viscosity and that for an aqueous solution is calculared and used to determine the value of the rotational correlation time from the theoretically calculated dependence. If the spin label is able to rotate relatively independently of the macromolecule, it would be possible to find conditions (e.g., by increasing the viscosity by adding sucrose) in which the macromolecule is immobolized completely, while the label is still mobile. With increasing viscosity, the rotation of the spin label would eventually cease. Experimentally, the dependence of 2A’ versus the temperature : viscosity ratio has to be plotted (an isotherm). The linear part of such an isotherm, which characterizes the hindrance of macromolecular rotation, is extrapolated to infinite viscosity and the value found is used to calculate the rotational correlation time, as in the previous case. The values of the rotational correlation time obtained using this method are in good accordance with the data obtained using other techniques, particularly those by fluorescence polarization. 111. Segmental Flexibility of Ig Molecules
A. IgG 1. Early Studies
One of the important findings of classical studies by R. R. Porter on the papain fragmentation of IgC was that Fab and Fc fragments retain their principal biological activities. This observation could only be possible if
6
ROALD NEZLIN
the general structural features of the isolated fragments were preserved during papain cleavage. The hydrodynamic studies by Noelken et al. (1965) supported this notion. It was found that the papain Fab and Fc fragments of IgG behave ’ as typical globular proteins: Their frictional coefficient ratios at f : f M l N are 1.10-1.25 and their intrinsic viscosity is 4.0 ml/g. In the same experiments, however, whole IgG appears to be a more extended molecule (f : fMIN of 1.47, intrinsic viscosity of 6.0 ml/g). Based on these observations, Noelken et al. formulated the hypothesis that the IgG molecule is built from three compact subunits resembling papain fragments, linked by flexible parts of the heavy peptide chains. At about the same time the results of two electron-microscopic studies were published which also demonstrated the free movements of the Fab subunits. In both investigations the complexes of IgG antibodies with antigens were studied using the negative contrast method. Feinstein and Rowe (1965) found two main types of antibody complexes by using ferritin, a molecule which has many identical epitopes. In soluble complexes antibody molecules bind by combining sites to nearby epitopes and have a compact structure. But after cross-linking of two antigen molecules, these researchers noted that ‘the antibody molecule “clicks open” to varying degrees about a “hinge point”.’ Valentine and Green (1967), in their elegant studies on antibodies complexed with a small bivalent hapten, also found that the angle between antibody subunits bearing the combining site varies from nearly 0” in antibody dimers to 180” in ring complexes built from several antibody molecules. These observations form the experimental basis for a flexible model of the IgG molecule, according to which antibody arms could rotate more o r less freely and independently, due to the existence of a flexible hinge region, a stretch of heavy chain between Fab and Fc. T h e methods used in these studies, however, do not provide unequivocal evidence that protein subunits rotate freely in solution. The most direct ways to prove unambiguously that a protein molecule is constructed from independently rotating subunits is by measuring the rotational correlation time, using relaxation methods such as fluorescence polarization or the spinlabel method, and comparing the data obtained with the computed values for rigid models.
2 . Fluorescence Polarization Studies Determination of the rotational correlation time of Ig molecules by using measurements of fluorescence polarization of dye-protein conjugates has a long history. For nearly 30 years this approach has been used in many laboratories, and the quality of the results was correlated with the
INTERNAL MOVEMENTS I N
Ig MOLECULES
7
progressive development of the methodology. In early studies (reviewed by Dorrington and Tanford, 1970; Cathou et al., 1974) the steady-state method was used, which enables the determination of an average rotational correlation time, but the contradictory results obtained from different experiments were due to some methodological problems (see Section 11,D). In 1969 Zagyansky st al. performed experiments with dansyl-human IgC conjugates, trying to avoid possible methodological errors. First, the values of the lifetime of the excited state of conjugated dansyl groups were measured directly by phase fluorometry. This parameter was necessary to calculate the rotational correlation time. The values obtained (about 7.3 nsec) were significantly lower than the previously used values of the lifetime of the excited state of dansyl-albumin conjugates (i.e., 12 nsec). Also, immediately prior to commencing polarization measurements, the conjugates were freed, using gel filtration, from aggregates which are always present in Ig solutions and which could lead to overestimation of the values obtained. The results of this study were clear-cut: The experimentally found rotational relaxation time for IgC was several times lower (i.e., 20 nsec) than that calculated on the assumption that the IgC molecule is rigid (about 70 nsec) (Table I). This indicates that the observed polarization is determined by Brownian rotation not of the whole IgC molecule, but of its parts, which are substantially smaller, being interconnected by flexible bonds. In further experiments it was found that the correlation time for the Fab fragment is close to that in experiments with intact IgC (Nezlin et al., 1970). Due to the short lifetime of the excited state of the dansyl dye in these experiments, only Fab movements were registered, not the tumbling of the whole molecule. The contribution of thermally activated free rotation of the covalently bound dansyl groups was not very significant, since the values of the rotational correlation time for Fab were nearly the same, whether the dye molecule was fixed in the combining site of antidansyl Fab or was attached nonspecifically elsewhere (Tumerman et al., 1972a,b). It was later found that the values of the rotational correlation time for the whole IgC molecule could be slightly higher. Dudich et al. (1978) showed that this parameter for IgC from various species is concentration dependent: Determinations performed at protein concentrations below 2 phi give values about 20% lower than those from higher concentrations. This phenomenon is clearly dependent on Fc. It was found only with IgC and Fc, not with Fab or with complexes of dansyl with the specific antidansyl antibody. As was later established (Dudich and Dudich, 1983),the concentration dependence was due to the dissociation of
ROALD NE%I.IN
8
TABLE I ROTATIONAL CORRELATION TIMES OF IgCs A N D THEIR FRAGMENTS DETERMINED BY FLUORESCENCE POLARIZATION" Rotational correlation time (nsec), fluorescent dye localized Protein
Inside the Outside the combining site combining site
Steady-state fluorescent polarization (dansyl dye) 20-30 IgCl, human
IgG, rabbit Fab from human, rabbit, pig IgC Fab', F (ab'):! from human IgG Fc from human IgG IgC, rat antidansyl
27 21 20-22 12-16 33
Fab, Fab', rat 22 an tidansyl IgC, Pig. precipitating 41 anti-DNP IgG, pig 63 nonprecipitating anti-DNP Time-resolved fluorescence polarization' 4s 13.7-18.0 IgG, rabbit antidansyl 4' 82.7-105.1 IgC, rabbit 4s 24 antipyrene 4 L 131 Predicted global rotational correlation time IgG 75 nsec IpG 167 nsec
Reference Zagyansky et al. (1969), Dudich et al. (1978) Dudich et al. (1978) Nezlin el al. (1970), Dudich et al. (1978) Nezlin et al. (1970) Nezlin et al. (1970) Dudich et al. ( 1 978) Tumerman ct al. ( 1 972a,b) Tumerman et al. (1972a,b) Dudich et al. ( 1 978) Dudich et al. (1978)
Hanson et al. (1981) Hanson et al. ( 1 98 1) Lovejoy et al. (1977) Weltman and Edelman (1967) Hanson el al. (1981)
DNP, Dinitrophenyl. The short and long rotational correlation times characterizing two main depolarizing motions. a
small-molecular-weightpeptides also labeled by the dansyl dye. They usually adsorb on the Fc portion of the IgG molecule (not IgM or IgA) and dissociate after dilution. In 1970 Yguerabide et al. reported the results of direct measurements of fluorescence polarization of dansyl-antibody complexes using nanose-
INTERNAL MOVEMENTS IN
Ig MOLECULES
9
cond fluorescence spectroscopy. According to the time-dependent anisotropy curves of the whole IgC antibody, there are two main depolarizing motions: one with a correlation time of 33 nsec; the other, 168 nsec. The shorter value was interpreted in terms of Fab segmental flexibility, and the longer time was attributed to rotation of the whole molecule. The dansyl-Fab complex had only a single correlation time equal to the shorter time value for the dansyl-IgG complex, indicating that this fragment rotates as a rigid unit. Later Chan and Cathou (1977) repeated these experiments and confirmed that the rotation of IgG does not correspond to that of a compact sphere. The values of the correlation times were smaller (i.e., 26 and 110 nsec), due to the removal of protein aggregates by gel filtration. The results were not significantly different when, instead of the short-lived dansyl groups, the long-lived flourophore (pyrene) was used (Lovejoy et al., 1977). This permits the study of depolarization up to 300 nsec, whereas, in experiments with dansyl groups, measurements of the correlation times were limited to only 150 nsec. Two correlation times were found for rabbit antipyrene antibodies: 24-33 nsec and 13 1- 140 nsec. Hanson et al. (198 l), using new methodological approaches, carefully reinvestigated the fluorescence polarization of antidansyl rabbit antibodies freed from aggregates. It was found that the decay of fluorescence anisotropy can also be described by two rotational correlation times. The values, however, were significantly lower: 14- 17 nsec and 83- 104 nsec. The authors reinterpreted the data by Yguerabide et al. (1970) and concluded that the long correlation time reflects wagging or wobbling of the Fab subunits, rather than the global tumbling of the entire IgG molecule. The short correlation time was ascribed to the movement of variable parts of Fab, or Fab's twisting around its long axis. Steady-state and time-resolved fluorescence polarizations were also used to study the soluble complexes of antidansyl antibodies with staphylococcal protein A (Hanson et al., 1985). Protein A is capable of forming complexes of different sizes, with IgG molecules binding to their Fc portion, If, in the free IgG molecule, the Fab portion has considerable freedom of rotation, the formation of protein A-IgG complexes should not influence the rotational freedom of Fab, and the correlation time has been found to change only insignificantly. It was found that, despite fixation in complexes, IgC molecules retain their segmental flexibility. This also implies that IgC antibodies anchored to cells by their Fc portions also retain freedom of rotation of their Fab fragments. The finding by Slattery et al. ( 1985) that the segmental flexibility of IgE molecules does not undergo significant change after binding to cell Fc, receptors is in agreement with the data by Hanson et al. (1985). All of the data discussed above provide unequivocal proof of the
10
ROALD NEZLIN
existence of the segmental flexibility of IgG antibody molecules in solution. Indeed, the experimentally found rotational correlation times are lower than the correlation times calculated for a rigid sphere of hydrodynamically equivalent dimensions: 75 nsec, according to one method of calculation, and 155- 167 nsec to another (Hanson et al., 1981; Slattery et al., 1985). There is no obvious energy restriction for rotation of the Fab arms from the fully opened position at an angle of 180" and a closed position of less than 60" (Schumaker et al., 1980). The lack of energy constraints allows both Fab arms to move relatively freely to wide angles, and this of course has crucial functional significance. However, it is difficult to discuss the real physical meaning of the short and long rotational correlation times obtained using the nanosecond polarization technique. Due to methodological problems, the emission anisotropy kinetics of these experiments were usually fitted to the sum of two exponentials. However, the rotational motions of nonspherical particles are complex, and to describe them properly, more correlation times are needed (Oi et al., 1984). Therefore, the short and long correlation times obtained in time-resolved fluorescence experiments, in fact, are weighted averages of several values (Chan and Cathou, 1977). In the latest series of nanosecond polarization experiments, the mean rotational correlation time was used for comparison among the flexibilities of different Ig molecules (Dangl et al., 1988; Schneider et al., 1988). This parameter is calculated as a weighted average of the two experimentally obtained correlation times and does not enable the direct comparison of the correlation times derived from steady-statepolarization experiments. 3 . Spin-Label Approach
The possibility of using the spin-label method to determine the rotational correlation time of protein molecules is based on initial findings by Shimshick and McConnell(l972). They found that the differences among positions of the extrema of ESR spectra of spin-labeled protein, which had been immobilized and freely rotated in aqueous solution, could be attributed to the rotational motion of the macromolecule. Several modifications in the method were used in studies of rotational motion in Ig's (Kaivarainen, 1975; Dudich et al., 1977). As in the steady-state polarization technique, the rotational correlation times obtained using the spin-label technique represent mean values. The results were not significantly different if the label was bound either noncovalently in the combining site of the antilabel antibody or covalently elsewhere on the macromolecule. In our experiments for covalent labeling, the TEMPO-trichlorotriazine spin label was used (Kaivarainen and Nezlin, 1976), which distributed relatively randomly within the IgC molecule (with some preference to Fab) (Nezlin, 1986).
INTERNAL MOVEMENTS IN
Ig MOLECULES
11
Table 11 lists the values of the rotational correlation times obtained using the spin-label method. In general, they correspond well with the correlation times found in the steady-state polarization experiments (Table I), despite the obvious methodological differences between both approaches. Thus, the results obtained by two relaxation techniquesthe spin-label method and fluorescence polarization-coincide well and point to significant flexibility of the IgC molecules. 4 . X-Ray Crystallography
During the past two decades the three-dimensional structures of Fab and Fc fragments have been described in detail (Alzariet al., 1988; Davies et al., 1975; Huber et al., 1976; Davies and Metzger, 1983). X-Ray crystallography has also provided valuable information relating to the dynamic aspects of the IgG structure, as well as giving additional evidence for the existence of segmental flexibility in these molecules (Huber and Bennett, 1983; Bennett and Huber, 1984). Attempts by researchers to crystallize intact Ig’s however, have encountered difficulties. To date, only four IgC molecules have been crystallized, but two of them (i.e., Dob and Mcg) have deletions in the hinge region (Steiner and Lopes, 1979; Fett et al., 1973), resulting in considerably reduced freedom of Fab movement. The significant freedom of IgC subunit rotation is an obstacle for the formation of a crystal lattice. The most striking finding, however, was that, in crystals of intact IgG, the Fab fragments were well ordered, while the Fc subunit was disordered (Marquart et al., 1980; Ely et al., 1978). The Fc disorder is probably due to distribution of this part of the molecule among several different sites of the crystal lattice (Bennett and Huber, 1984). Fc is able to adopt at least four different conformations, which is why no density related to Fc was found. In contrast, no crystallographic evidence for the mobility of the Fc portion could be found in either hinge-deleted protein. In these proteins Fc is located tightly between the Fab portions, thereby limiting its mobility (Rajan et al., 1983). There are no such close contacts between subunits of the intact IgG (Huber et al., 1976). The flexibility of porcine IgC was studied using the neutron spin-echo method (Alpert et al., 1985). The data obtained are compatible with a flexible model of the molecule. 5. IgG Subclasses
Segmental flexibility depends on the structure of the part of heavy chains between the Fab and Fc subunits, the so-called hinge region. Structure of the hinge region of IgG isotypes varies greatly in the length and number of interheavy disulfide bridges; it is not surprising therefore that IgC subclasses express different degrees of flexibility.
12
ROALD N E Z L I N
TABLE I1 ROTATIONAL CORRELATION TIMES OF Ig's AND THEIR FRAGMENTS ON PROTEIN MOIETY' SPIN-LABELED Rotational correlation time (nsec) Protein
Label inside the combining site
Label outside the combining site
Kaivarainen el al. (1973, 1974)
Rabbit anti-spin label antibody IgG F (ab'h Fab' Rabbit IgC antiDNP antibody IgA MOPC 3 15 an ti-DNP mouse myeloma Fab Fv Mouse monoclonal IgG anti-DNP antibody, Fab Pig IgG antiDNP antibody Precipitating Nonprecipitating Rabbit IgC F (ab')z Fab' Rabbit, human pFc'
32 30 18 39
Hsia and Piette ( 1969) Dwek et al. (1975)
44.4
23 6.5 20
19 29
Anglister et al. ( 1984)
20 28 26 25 21 6
Human IgAl myeloma
32-4 1
Human IgA2 myeloma Human IgM myeloma Human IgE Yu myeloma
45
' DNP, dinitrophenyl.
Reference
50 60-62
Sykulev et al. (1979) Timofeev et al. (1978) Nezlin et al. (1985), Nezlin and Sykulev (1984) Dudich et al. (1980) Sykulev et al. (1984) Sykulev et al. (1984) Dudich and Dudich (1980) Nezlin et al. (1973)
INTERNAL MOVEMENTS I N
Ig MOLECULES
13
In early studies it was found that human IgG2, which has an unusual hinge region structure involving four closely separated disulfide bridges between heavy chains, has a higher value in correlation time than IgG1, which has only two disulfide bridges (Nezlin et al., 1973). Differences in flexibility exist between two subclasses of porcine IgG; for example, precipitating antibodies are more flexibile than nonprecipitating ones having the same specificity (Dudich et al., 1978; Sykulev et al., 1979). Cebra et al. (1977) studied two isotopes of guinea pig antibodies against the dansyl group. These subclasses have different hinge regions (i.e., the IgGl hinge region is shorter than that of IgG2) and differing abilities to cross-link hapten-protein conjugates (i.e., IgGl is less effective). IgGl antibodies mainly formed large complexes with the bivalent dansyl hapten. Some antibody molecules in these complexes are in the “open” form with the free antibody combining sites, whereas IgG2 antibodies easily formed “closed” dimers. Using nanosecond fluorescence polarization, it was found that IgGl molecules have restricted movement of their Fab fragments and hence are more rigid than IgG2 molecules, which have a more extended hinge region. A systematic study of flexibility differences between IgG isotypes was undertaken recently by a team from Stanford University (Reidler et al., 1982; Oi et al., 1984; Dangl et al., 1988; Schneider et al., 1988). Monoclonal antidansyl mouse antibodies of different IgG subclasses were produced using the heavy-chain switch variants derived from a hybridoma cell line, which originally synthesized IgGl antidansyl antibodies. All of these molecules have the same light chain and variable heavy-chain (VH) region, but different constant heavy-chain (CH)regions. According to nanosecond polarization experiments, the IgG2b variant is the most flexible molecule, whereas IgGz, is less flexible and IgGl is relatively rigid (Table 111) (Oi et al., 1984). This study was confirmed and extended by elegant experiments by Dangl et al. (1988). A family of six chimeric IgG molecules was produced by genetic manipulations of somatic cells (Schneider et al., 1987). T h e molecules had identical light chains and VH regions, but different CH regions: among four human CH isotypes, mouse C H and ~ rabbit CH regions. A strong direct correlation among segmental flexibility, length of the hinge region, and the ability to activate complement was found (Table IV). Distribution of murine IgG isotypes according to their flexibility is as follows (mean correlation time in parenthesis): (55 nsec) > IgGn, (63 nsec) > IgG3 (84 nsec) > IgGl (120 nsec). For human isotypes the distribution is: IgG3 (50 nsec) > IgGl (69 nsec) > IgG4 (84 nsec) > IgGn (120 nsec). Rabbit IgG has an intermediate flexibility (i.e.y72 nsec).
14
KOALD NEZLIN
TABLE I11
FLEXIBILITY OF MOUSEMONOCLONAL
ANTIBODIESWITH ANTIDANSYL ACTIVIT~ Mean rotational correlation time (nsec)c Proteinb
Without With dithiothreitol dithiothreitol
124 120
88
81 60 47
58 51 42
28 28 29
’
28 28 29
~~~
From Oi et al. (1984). Spontaneous mutants of a cell line producing IgGl antibody. All molecules have the same light chains and VH regions, but different CH regions. Determined by time-resolved fluorescence polarization. Myeloma human IgE ND was covalently coupled with a dansyl group (Cathou, 1978).
In another series of experiments a number of genetically engineered IgGl-IgGn hybrids were studied (Schneider et al., 1987, 1988). All of these molecules also possess identical variable parts, having antidansyl ability but differing constant regions, in which a section of one isotype was changed with the corresponding portion of a sequence from another isotype. The main finding of these experiments was that the proper combination of the C H1 domain and the hinge region is important for the flexibility of the IgG molecule (Table V). B. I ~ M A N D I ~ A The first evidence of IgM segmental flexibility was obtained by Feinstein and co-workers (Feinstein et al., 1971; Beale and Feinstein, 1976). Electron micrographs of complexes of bacterial flagella and IgM antibodies show that the IgM molecules are often in a “staple” conformation,
TABLE IV CORRELATION OF HINGE REGIONSTRUCTURE, FLEXIBILITY, AND COMPLEMENT BINDING ACTIVITY Complement fixationb Protein
Upper hinge
Core
Mouse 1 6 , Human 1 6 , Mouse IgGzb Human I g C l Mouse IgGz, Rabbit IgG Human l g G r Mouse IgGl Human I g G Guinea pig l g C 2 Guinea pig IgGl
EPRIPKPSTPPGSS ELKTPLGDTTHT EPSGPISTINP EPKSCDKTHT EPRGPTIKP APSTCSKPT ESKYGPP VPRDCG ERK EPIRTPZBPBP QSWGHT
C P CPRCP(EPKSCDTPPPCPRCP)s CPPCKECHK CP CP CPP CPPCK CP P C CP CPS CKPCI CT CCVECPP CP CTCPK CP CPPCIP C
Upper hinge length PGNILGGP APELLGGP APNLEGGP APELLGGP APNLLGGP PPELLGGP APEFLGGP VPSEVS APPVAGP PPPENLGGP GAPZLLGGP
14(9)’ 12
11 10 9 9 7(5)’ 6 3 I1 6
Correlation times (nsec) 78 f 3 50 f 2 55 f 2 69 2 3 63 f 3 72 f 3 84 f 3
81 f 3
120 f 5 Flexible Restricted flexibility
Human
Rabbit
Guinea pig
70
15
-
100 300 250 300
30
I00 300 250 450
-
+I-
80
25 400
-
+/250
-
+I-
-
+ -
a Shown are the mean correlation times and complement fixation of mouse chimeric (Dangl et al., 1988) and guinea pig IgC, and IgG2 (Cebra et al., 1977; data on complement activity from Sandberg et al., 1971). The amino acid sequences are aligned from residues 216 to 238 (human IgCl Eu numbering). Amount of DNS26-BSA (ng) required to activate one CHso with 10 p g of antibody. Hinge length when polyproline helical structures are considered restricting elements defining upper hinge length.
16
KOALD NEZLIN
TABLE V DOMAIN STRUCTURE AND SEGMENTAL FLEXIBILITY OF GENETICALLY ENGINEERED ANTIBODIES~ with or without dithiothreitol
CH lc Proteinb IgG w 2 a
Interdomain hybrids Hingeless Ig& Hybrid 2 Hybrid 3 Hybrid 9 Hybrid 10 Intra-CH1 hybrids Hybrid 11 Hybrid 12 Hybrid 13 Hybrid 14
Aminoterminal
Carboxyterminal
Hinge
c H 2 cH3
Without
With
Yl Y2a
Yl
y2a
Yl y2a
Yl
Yl y2a
83 61
71 49
Y2a Yl Yl Y2a Y2Y
y2a Yl Yl Y2a y2a
Deleted Y2 Yl Yl y2a
y2a Y2 y2ad Yl Yle
y2a Y2 Yl Yl
y2a
84 84 82 108 62
87 55 61 76 52
Y2a Yl
Yl
Yl y2a Yl Y2a
Yl y2a Yl y2a
Yl y2a Yl y2a
114 89 81 64
60 49 76 47
Yl
Y2a
y2a Y2a Yl
y2a
From Schneider et al. (1988). Antidansyl IgCs with identical light- and heavy-chain variable regions. Each CH domain (CH1, hinge, c&!,and cH3) was derived from mouse IgCl or IgC2,. The crossover points in the corresponding DNA sequences were either within introns or within the coding regions at the positions coding for the amino acids, as indicated in the footnotes in Eu protein numbering. The segmental flexibility was determined from nanosecond fluorescence anisotropy kinetics. The mean rotational correlation time (4) calculated from the two exponential fits is given. The amino-terminal half of CH1 is amino acids 118-161: the carboxy-terminal half is 162-2 15. y l to residue 238; y2a from residue 239 on. y2a to residue 238; y l from residue 239 on.
when the Fab arms, linked to a thread of a single flagellum, are folded down from the Fc region, which is seen as a central disc. Later a number of studies were performed on human myeloma IgM or IgM antibodies, using fluorescence polarization and spin-label techniques (Table VI) (Cathou, 1978). All of the data provided supporting evidence for segmental flexibility of the human myeloma IgM molecules. The experimental correlation times are significantly lower than those calculated for a rigid sphere with the IgM volume. Although the IgM ~ probably serves as molecule has no typical hinge region, the C Hdomain
INTERNAL MOVEMENTS IN
Ig MOLECULES
17
TABLE VI ROTATIONAL CORRELATION TIMES OF IgA, IgE, AND IgM DETERMINED BY FLUORESCENCE POLARIZATION Correlation time (nsec)
Protein Steady-state polarization" IgA, human myeloma
147
IgA, monomer human myeloma
26-32
IgA, dimer human myeloma
30
Fab from IgA human myeloma
25
IgAl monomer human myeloma IgE, Yu human myeloma IgE, mouse antidansyl IgM, human myeloma IgM, subunit human myeloma IgM, human myeloma
33(32)b 55(60)b 54' 73(50)* 57 27-47
IgM, subunit human myeloma
22-27
Time-resolved fluorescence polarizationd Fab from equine antidansyl IgM 32 IgM, equine antidansyl IgM IgM, porcine antidansyl IgM, nurse shark
4s 4s
61 >loo0 69 568 93
+L
>loo0
f#JL
4s
4L
Reference Weltman and Davies (1970) Zagyansky and Gavrilova (1974) Zagyansky and Gavrilova (1974) Zagyansky and Gavrilova (1974) Dudich et al. (1980) Nezlin et al. (1973) Slattery et al. (1985) Dudich el al. (1980) Dudich et al. (1980) Zagyansky et al. (1972) Zagyansky et al. (1972) Holowka and Cathou (1976) Holowka and Cathou (1 976) Holowka and Cathou (1976) Holowka and Cathou (1976)
' Covalent labeling outside the combining site.
In parentheses, the values of the rotational correlation time, as determined by the spin-label method. Dansyl dye in the combining site. Short and long correlation times which characterize the two main depolarizing motions. Dansyl dye is in the combining site.
its analog (Holowka and Cathou, 1976). IgMs of other species were also studied and it was found that IgMs of lower vertebrates were less flexible than those of human proteins (see Section VII). Most of the experimental data on IgA were obtained by steady-state polarization and spin-label techniques (Table VI). The data are compati-
18
KOALD NEZLIN
ble with those of the flexible model of these molecules. Results of the nanosecond polarization experiments on IgA also provide evidence in favor of segmental flexibility of this class of Ig's (Liu etal., 1981). The data were similar for monomeric, dimeric, and secretory forms of IgA, despite the fact that monomeric IgA antibodies are poor agglutinins. There are probably other factors which may explain the poor correlation between flexibility and the ability to agglutinate antigen. The mobile part of IgA molecules are most likely their Fab fragments since the correlation time values for whole IgA and their Fab are very similar. C. IgE IgE's have specific functions which are different from those of other classes: They serve as antigen receptors on mast cells and basophils and are an important participant in the chain of events which lead to specific allergic reactions after encounter with allergens. The concentration of IgE in serum is usually low, and all experiments were initially done on two proteins isolated from myeloma patients Yu and ND. In the experiments performed by steady-state polarization and the spin-label method, the IgE (Yu) molecule was significantly less flexible than were IgC and IgM molecules (Table VI) (Nezlin et al., 1973). Time-resolved polarization measurements of IgE (ND) coupled covalently with dansyl groups also pointed to the rigidity of these molecules (Cathou, 1978). Slattery et al. (1985) used steady-state polarization measurements to study specific complexes of dansyl groups with corresponding mouse monoclonal antibodies. The average value of the correlation time for IgE in solution was found to be 54 nsec, which corresponded well with the results of the above-mentioned experiments on IgE (Yu)(i.e., 55 nsec). Similar results were found for covalent complexes of another fluorophore-pyrenylmaleimide-with the same IgE antibodies (i.e., 64 nsec). The calculated value of the correlation time for a rigid sphere with IgE parameters is 78 or 155 nsec, depending on the method of calculation. Experiments with complexes of the dansyl hapten with IgE antibodies bound to receptors on membrane vesicles from basophilic cells supported the low flexible model of IgE. Despite the large volume of such a particle with bound IgE (calculated correlation time, 1000 nsec), the experimentally found correlation time was only 64 nsec. Hence, the IgE molecule is not completely rigid, and its Fab subunits have some freedom of rotation, although it is less pronounced than that in IgG or IgM. The ability of IgE antibodies to agglutinate erythrocytes is also evidence in favor of some freedom of rotation of the Fab, subunit (Ishizaka, 1973). Among the family of monoclonal mouse antibodies with identical com-
INTERNAL MOVEMENTS IN
Ig MOLECULES
19
bining sites, but differing CH regions, IgE molecules are the most rigid (Table 111) (Oi et al., 1984). Even if one takes into account the higher molecular weight of IgE molecules (190,000 as compared to 150,000 for IgC molecules), the value of the correlation time for IgE is still higher (i.e., 98 nsec) than that for IgCl, the most rigid IgG subclass (i.e., 81 nsec). IV. Internal Motions in Fab and Fc Subunits
All Ig Fab subunits are built from variable domains of both light and heavy chains and constant domains, CL and CH1 respectively. Both variable domains, as well as two constant domains, form separate entities linked by short parts (Lea,switch peptides) of each of the chains. According to fluorescence polarization and spin-label measurements, isolated Fab fragments behave in solution as rigid particles, with rotation correlation times between 20 and 30 nsec. These values correspond well with the calculated value for a rigid ellipsoid with a volume equal to that of Fab. Crystallographic data for Fab have shown that extensive contact is nonexistent between variable and constant parts of the fragment. Thus, it is not surprising that X-ray and electron-microscopic studies have demonstrated that the angle between the pseudodyad axis of V and C parts ("elbow bend") varies significantly in Fab's from various proteins under study (Huber et al., 1976).T o date, there are crystallographic data for 15 Fab's and the elbow angle has been found to vary from 132"to 172" (Davies et al., 1988). The elbow flexibility was also demonstrated by crystallographic studies of the light-chain dimer, which imitates the structure of Fab (Edmundson et al., 1978). One of the two identical chains of the dimer have the conformation of an amino-terminal half of the heavy chain, and the other, the light chain. Despite the identity of both light chains in a dimer, their elbow angles are quite different (i.e., 70" and 110"). Further evidence for flexibility in the switch region of Fab was obtained from studies of different crystals of the light-chain dimers. The dimers of light chains of a monoclonal IgG crystallize in two forms: trigonal and orthorhombic. In two different crystal lattices the bend angle varies significantly--115" and 132" (Abola et al., 1980; Ely et al., 1983). More recently, it was found that the same Fab' from a monoclonal antibody can exist in two or more forms and they are bent differently at the switch region. Sheriff et al. (1 987) described two variants of complexes of Fab from the antilysozyme monoclonal antibody with lysozyme, which differ in switch bending, and Prasad et al. (1988) reported similar data with the
20
R0AL.D NEZLIN
Fab from an antibody against an Escherichia coli protein. In both cases the angular variation was nearly identical (i.e., 7" and 8"). Two groups presented electron-microscopic evidence of Fab bending after the reaction of antibodies with protein antigens (Roux and Metzger, 1982; Wrigley et al., 1983; ROUX,1984). This kind of flexibility is particularly evident on micrographs of closed complexes of two molecules of influenza hemagglutinin with two antibody molecules, in which the angle between the V and C portions of Fab was 90". Thus, in solution in which the Fab is either free or is part of an intact Ig molecule, it rotates as a compact structure, but the bend angle between the V and C parts can be changed significantly by some external force when it complexes with an antigen or during formation of a crystal lattice. The Fc fragment has quite different rotational properties from Fab. Measurements of the correlation time revealed that this parameter for isolated Fc is significantly smaller than that for Fab (Table I) (Nezlin et al., 1970; Dudich et al., 1978; Timofeev et al., 1978). Both fragments are built from four domains of nearly equal dimensions, and only some internal lability of Fc could be responsible for the different rotational properties of both fragments. After elucidation of the three-dimensional structure of the Fc fragment (Deisenhofer et al., 1976; Deisenhofer, 1981) it became evident that the main reason for the internal motion in the nanosecond range is the lack of lateral contact between c H 2 domains. Other structural features promoting flexibility are common to both fragments-namely, few longitudinal contacts and extended conformation of switch peptides. In intact molecules the Fc conformation is stabilized by S-S bridges on the hinge region (Seegan et al., 1979). Despite this, the Fc, even in the intact IgG molecule, has a high degree of internal rotational freedom: The correlation times of Fc and IgG spin-labeled on carbohydrates are nearly the same (Timofeev et al., 1978) and are about twice as short as the correlation time for Fab (see also Table VIII). Using the proton relaxation enhancement technique, Burton et al. (1977) not only found that the Fc of rabbit IgG is not the rigid particle, but also that significant rotational freedom is characteristic for this part of the molecule. The peculiar nonstable structure of Fc is responsible for the high sensitivity of the fragment of external influence, for example, the action of low pH (Abaturov et al., 1969). In the intact IgG molecule and in the isolated Fc fragment, the CH3 domains interact closely. The isolated pepsin pFc' fragment, which is a noncovalent dimer of c H 3 , conserves the main features characteristic for Ig C domains (Phizackerlay et al., 1979). However, in solution pFc' is flexible; the correlation times for pFc' are half of those for Fc (Table 11)
INTERNAL M O V E M E N T S IN
Ig MOLECULES
21
(Nezlin and Sykulev, 1984; Nezlin et al., 1985). The mobility of the CH3 domains in the Fc evidently is restricted by some form of contact with C H domains. ~ T h e CH2-cH.3 switch region is another flexible point of the IgG molecule. Bending in this region is clearly seen after the reduction of interheavy S-S bridges. Romans et al. (1977) found that so-called “incomplete” IgG antibodies, which are unable to agglutinate red cells, are converted into direct agglutinins after mild reduction. Such treatment cleaves S-S bonds in the hinge region, opens the upper part of the Fc region, and allows Fab arms to bridge between two red blood cells. Seegan et al. (1 979) measured the length of the Fab arms in complexes of reduced antibodies with antigens and found that it increases by 23 A. Such results could be explained by the formation of a new hinge between the cH2 and CH3domains. Nanosecond polarization experiments (Chan and Cathou, 1977) directly point to increased flexibility after the mild reduction of rabbit IgG (see also last columns of Tables 111 and V). V. Role of the Hinge Region Structure for Segmental Flexibility
The heavy-chain portion between Fab and Fc, which has a unique structure, is coded by a separate DNA segment (Sakano et al., 1979). Its susceptibility to proteolytic attack points to a noncompact extended conformation, and early electron-microscopic data have already suggested that the segmental flexibility of the IgG molecule depended mainly on this region’s action as a hinge. The principle role which the hinge region plays in the freedom of Fab and Fc movement was confirmed after crystallographic studies of the Dob and Mcg hinge-deleted proteins (Silverton et al., 1977; Ely et al., 1978). In contrast to the intact Kol and Zie proteins (in which the Fc fragment adopts more than one position in the crystal lattice), Fc of Dob and Mcg, which lack segmental flexibility, contributes fully to the diffraction pattern and has a definite position in the crystal. As a result of intensive crystallographic and proton nuclear magnetic resonance investigations, the structure of the hinge region of IgGl heavy chains, as well as its role in Ig flexibility, is now well understood (Marquart et al., 1980; Ito and Arata, 1985; Endo and Arata, 1985). This region is built from a central portion (i.e., core) and two flanking segments (Fig. 1). In crystal form as well as in solution, the central portion with the sequence -Cys-Pro-Pro-Cysforms two parallel poly-Lproline double helices, linked by S-S bridges. The segment on the amino-terminal side of the core (Cy~~~’-Thr’~~) in crystal forms a oneturn helix with little inherent stability. It is exposed to solvent and is
22
ROALD NEZLIN
L
Cys
I
I
pepsin
trypsin papain
la ProGluLeiLeuGly Pro220 I
L
I
I
I
235 IaProGluLeuLeuGly ProJ
cys
FIG. 1. The hinge region of human IgG. The residues in the box (216-230) are deleted in IgGl Dob myeloma protein (Steiner and Lopes, 1979).The central part of the hinge (core)is shadowed.
mainly responsible for the segmental flexibility of the IgC molecule. T h e segment on the carboxy-terminal side of the core ( P r ~ ~ ~ ~ - L ehas u ~an ”) extended conformation, but is not as flexible as the segment on the left side. However, after reduction and alkylation of the core S-S bonds, both segments have comparable flexibility (Endo and Arata, 1985). Another point of flexibility is the Cys229 atom, which, according to nuclear magnetic resonance studies (Ito and Arata, 1985), has considerable freedom of internal motion around the NH-C, and Cp-S bonds. The roles of different portions in flexibility of the hinge region are demonstrated in Tables IV and V, which summarize the results of polarization experiments on mutant and genetically engineered Ig molecules (Schneider et al., 1988; Dangl et al., 1988). In the first series of experiments, discussed in part above, the relationship of flexibility to the structure of the amino-terminal side segment of the core (left segment or upper hinge) is clearly established: T h e shorter the segment, the lower the correlation time. This correlation is in agreement with the abovementioned nuclear magnetic resonance findings on the role of the amino-terminal segment in flexibility. In other experiments (Table V) the rotational motion of Fab was studied on molecular hybrids of two mouse isotypes-IgGl and IgG2which possess different degrees of flexibility. T h e hybrid molecules are distinguished from each other by the structure of one or more CH domains, the hinge region, or part of the CHIdomain. It was found that, for
INTERNAL MOVEMENTS I N
Ig MOLECULES
23
flexibility, not only that the hinge region structure is important, but that both the CH1 and hinge regions must be in harmony, that is, properly matched. For example, the hybrid 13, which differed from flexible IgG2, only by the amino-terminal part of CH1, was found to be less flexible than intact IgG2,. VI. Functional Implications of Segmental Flexibility
A. PRECIPITATION AND AGGLUTINATION A condition for the formation of a precipitate or agglutinate network is the ability of the antibody molecules to bind simultaneously to epitopes of two antigen molecules or two cells. Epitope arrangement varies greatly, and freedom of Fab movement is crucial for optimal binding to antigens and the formation of a bridge between antigen molecules and cells. T h e ability of Ig molecules to change their general conformation from Y to T type (Cser et al., 1981a,b) greatly facilitates the capacity of antibodymediated linkage to antigens. Direct measurements of the flexibility of two porcine IgG antibodies having differing abilities to precipitate antigen support this view. During the early stages of the immune response to dinitrophenyl hapten, the animals synthesized antibodies which formed mostly insoluble complexes (precipitates) with heavily dinitrophenyl-substituted proteins, while antibodies generated later mostly formed soluble complexes with the same antigen molecules. Fluorescence polarization (Dudich et al., 1978) and the spin-label method (Sykulev et al., 1979) revealed that precipitating antibodies are significantly more flexible than are nonprecipitating antibodies (Tables I and 11). Spin-label experiments gave nearly identical results when the label was either in the antibody-combining site (the correlation times were 19 and 29 nsec for precipitating and nonprecipitating antibodies, respectively) or covalently bound to some residues outside the combining site (20 and 28 nsec, respectively). Neutron diffraction studies (Cser et al., 1977) and electron micrographs (Ryazantsev et al., 1989) also provided evidence for a more compact structure of the nonprecipitating antibodies. Carp IgM antibodies are a further example of the correlation between restricted flexibility and weak precipitating activity (Richter e l al., 1972). These antibodies precipitated only the presence of a large excess of antigen and, in the case of antibodies against haptens, only if a large number of hapten molecules were present on the carrier protein. According to steady-state fluorescence polarization measurements, carp
24
ROA1.D NEZLIN
antibodies are much less flexible than more recently evolved IgC and even human IgM (Table VII). The importance of flexibility in allowing antibody molecules to bridge between two cells was discussed above, as far as the ability of mild reduction to convert the incomplete agglutinins to direct agglutinins (Section IV). Another aspect of segmental flexibility is linked to the classic experiments by Karush (1978). He found that the functional affinity of antibodies is enhanced 100 times by antibody attachment to antigen with repeating determinants such as bacterial polysaccharides or synthetic antigens, not with one, but with two or more, combining sites. Although the arrangements of these repeating antigenic determinants cannot be identical for different kinds of bacteria or viruses, the gross conformations of antibody molecules of different specificities are the same. Obviously, the ability of the Fab fragments to rotate relatively independently greatly facilitates the interaction of the antibody with two or more combining sites with antigens, resulting in a significant gain in binding strength provided by bivalency . B. COMPLEMENT ACTIVATION AND BINDING TO CELLRECEPTORS Abnormal IgG proteins with a deleted hinge region (Steiner and Lopes, 1979; Deutsch and Suzuki, 1971) lack a number of important
TABLE VII FLEXIBILITY OF Ig’s FROM Low VERTEBRATES Rotational correlation time (nsec)” Species
IgM
Shark Squalis acantias
147
Carp Cyprinus carpi0 Frog R a m temporaria Tortoise Testudo horstieldi
128 135 102
Hen Galus domesticus Human, rat, rabbit
IgC
67 68
43 23-73
20-30
Reference Zagyansky and Ivannikova ( 1 974) Richter et al. (1972) Zagyansky (1975) Zagyansky and Ivannikova (1 974) Zagyansky and Ivannikova (1974) Zagyansky et al. (1969, 1972), Dudich et al. (1978, 1980)
” As determined by steady-state fluorescence polarization.
INTERNAL M O V E M E N T S IN
lg MOI.ECULES
25
biological functions, such as binding of the first component of complement (Clq) and the ability to react with different Fc cell receptors (Klein et al., 1981). Reactive sites responsible for interaction with the C l q component and a variety of Fc cell receptors are located in the CH2 domains (Burton, 1985). According to crystallographic studies of a hinge-deleted IgGI Dob protein (Silverton et al., 1977), Fab portions of the molecule are close to the C Hdomain ~ and sterically obstruct the reactive sites, making it impossible for a molecule as large as C l q to reach them (Dorrington and Klein, 1982, 1983). IgC4 molecules, which have restricted flexibility, have no complement binding capacity (Isenman et al., 1975), but the Fc fragment isolated from IgC4 is as active as Fc from I&. Clearly, the Fab fragments somehow protect the C lq binding site, and this structural property of IgC4 is obviously correlated to restricted flexibility. Analyzing the primary structure of the hinge regions of different Ig’s, Beale and Feinstein (1976) paid attention to the direct correlation between the length of the so-called upper part of the region (i.e., the amino-terminal side of the core) and complement activation (see also Feinstein et al., 1986). It is also known that the upper part of the hinge is important for the segmental flexibility of Ig molecules (Section V). Both properties-flexibility and complement activation-were recently compared on different genetically engineered chimeric mouse antidansyl antibodies, which have the same variable parts, but differ in constant parts of the heavy chain (Oi et al., 1984; Dangl et al., 1988). A clear-cut relationship was found among the length of the upper part of the hinge, the mean correlation time which reflects Fab motion, and the ability to fix complement (Table IV). The shorter the hinge, the more rigid the molecule, and the complement binding activity is corresponding lower. An extreme example is the hingeless rigid IgC2, antibody, which is unable to activate complement in all, in this respect resembling the hingedeleted human Dob protein. This study is in agreement with the results by Leatherbarrow and Dwek (1984), who studied C lq binding to immobilized mouse monoclonal Ig isotypes. The order of activity was found to be Igcg, > > IgGI. Correlation between the length of the upper part of the hinge and properties of the isotypes was also found for guinea pig Ig’s (Cebra et al., 1977). The isotype IgCl, which has restricted flexibility and a short sequence in this part of the hinge, is unable to fix complement (Sandberg et al., 1971) (Table IV). Ig’s lacking the hinge region have reduced, or even totally lack, any affinity for receptors on the surface of different kinds of cells (Klein et al., 1981). The sites involved in the binding of these receptors are probably located in the CH2 region (Burton, 1985), and it is possible that the Fab
26
KOALD NEZLIN
can also modulate the activity of these sites in the same way as a complement binding site. However, to date, no direct comparison of segmental flexibility and Fc receptor binding activity of different IgG subclasses has been made. According to the attractive hypothesis proposed by Burton ( 1986), flexibility of the IgG molecule allows the rotation of the Fc portion perpendicular to the plane of Fab subunits. This dislocation eliminates and obstruction of the Fc functional sites after immunoprecipitation and promotes interaction with C l q and cell receptors. Such movements are impossible in nonflexible molecules. VII. Evolution of Ig Flexibility
Several studies were performed in the 1970s to determine the flexibility of Ig’s isolated from sera of different species. The mean correlation times obtained by steady-state polarization measurements for Ig’s of lower vertebrates are definitely higher than those for Ig’s of the higher vertebrates (Table VII). For example, the correlation times for shark and carp IgMs are five to six times higher than those for human IgM; those for 7 S Ig’s (IgG type) from the tortoise and the hen are more rigid than those for human IgC. The nanosecond fluorescence polarization measurements also point out that nurse shark IgM is less flexible than equine or porcine IgM (Holowka and Cathou, 1976). Study of the general structure of Ig’s of low vertebrates has attracted less interest, and it is thus difficult to explain the observed differences in flexibility in real structural terms. According to Holowka and Cathou (1976), the Fab units and the whole F (ab’):! fragment of equine and porcine IgMs have some freedom for independent rotation, but for shark IgM molecules only bending of F (ab‘):!is possible. In summary, there is reason to believe that the segmental flexibility of antibody molecules appeared relatively late in evolution. VIII. Mobility of Carbohydrate Components
Ig molecules are glycoproteins, having several oligosaccharides, usually linked to heavy chains (Torano et al., 1977). In IgG about 2-3% of carbohydrates were found; in other classes their content is considerably higher (i.e., 10-12%) (Nezlin, 1977). When considering the spatial relationship between oligosaccharide chains and the protein moiety in different glycoproteins, two situations can be expected. In the first case an.oligosaccharide rotates relatively freely and links to the protein moiety by covalently binding to an amino
INTERNAL M O V E M E N T S IN
Ig MOLI:CLII.ES
27
acid residue. In the second an oligosaccharide chain interacts with several amino acid residues in a particular region of the protein surface and rotates together with the adjacent protein subunit of the molecule. In the latter case carbohydrate residues could close a part of the protein surface, preventing further interactions with some ligands. The closed area would be protected from proteolysis. Sugar residues could construct completely new antigenic o r recognition sites on the surface of a glycoprotein molecule. From the above, the importance of understanding the mode of the behaviour of carbohydrate chains becomes clear. The main carbohydrate unit of IgG linked to the Am2’’ residue of the Fc is the most studied. In crystals this branched oligosaccharide is ordered and occupies a fixed position covering part of the C2H domain surface. In human Fc it forms a few hydrogen bonds with the CH2 amino acid residues, but its main interactions are hydrophobic (Deisenhofer, 1981; Deisenhofer and Huber, 1983). The structure of the Fc carbohydrate in the rabbit is stabilizing not only by contact with amino acid residues of cH2, but also by contact with the oligosaccharide linked to the cH2 domain of the other heavy chain. Nearly all nine sugar units of the oligosaccharide core and both branches have well-defined density and are therefore fixed and immobile (Sutton and Philipps, 1983). Spatial positioning of carbohydrates in other Ig’s has not been investigated. T o study carbohydrate mobility in Ig molecules in solution, we used the spin-label approach (Nezlin and Sykulev, 1982). For spin-labeling of Ig oligosaccharides, limited periodate oxidation was performed, followed by reductive amination with TEMPO-amine spin label (Fig. 2) (Willan et al., 1977; Nezlin et al., 1978; Gnewuch and Sosnowsky, 1986). It was found that less than two spin-labeled molecules are bound to an IgG molecule. Significantly, more spin labels are incorporated in IgE (8 mol
Nal04
TEMPO amine
OH
I-
0
FIG.2. Introduction of TEMPO-amine spin label into carbohydrate residues.
28
KOALD NEZLIN
per mole of protein) and IgM (about 30 mol per mole of protein). This may be the result of the higher carbohydrate content in both IgE and IgM molecules. Using the same method, high-molecular-weight polysaccharide isolated from Pneumococczls was spin-labeled. ESR spectra of this polysaccharide are represented by three sharp lines and are similar to the spectrum of a spin label freely mobile in solution (Fig. 3). This may be the result of flexibility of the bonds between sugar units in the carbohydrate chain. However, ESR spectra of Ig’s spin-labeled at carbohydrates were quite different and reflected partially immobilized rotation of the spin label due to the attachment of oligosaccharide chains with the bound spin label to the protein moiety. In Table VIII the rotational correlation times of human Ig’s and their fragments spin labeled on carbohydrates are presented. Correlation times for IgG and Fc are similar and are about half of those for Fab, the value predicted for a rigid sphere having the same volume. Similarity of the data for IgG and Fc means that the relative mobility of oligosaccharides is the same in both isolated Fc and the intact IgG molecule. T h e correlation times obtained in these experiments closely resemble those for Fc found in the fluorescence polarization experiments (see Section IV). Apparently, the Fc oligosaccharide rotates together with an adjoining subunit (i.e., the CH2domain). The main finding of these experiments is that the close attachment of the Fc oligosaccharide to the protein surface points to the oligosaccharide B
A FREE SPIN LABEL -14OOC
20oc
PNEUMOCOCCUS POLYSACCHARIDE
IMMUNOGLOBULIN G
-
20%, OSbruuoM
O.C.0 x 8ucrose OeC,34%aucross
FIG.3. Electron spin resonance spectra. (A) Spin label free in solution (right) and immobilized (left). (B, left) Spin-labeled high-molecular-weight Pneumococcus polysaccharide without (top) and with (bottom) rabbit antipolysaccharide antibodies (Nezlin and Sykulev, 1982). (Right)Human IgC spin-labeled at carbohydrates. 2A‘, Distance between outer wide extrema (Timofeev et al., 1978).
INTERNAL MOVEMENTS IN
Ig MOI.ECULES
29
TABLE VIII ROTATIONALCORRELATION TIMES OF HUMAN Ig's A N D HLA ANTIGENS SPIN-LABELED AT CARBOHYDRATE AND PROTEIN MOIETY Rotational correlation times (nsec) Protein
IgC (mainly IgCI) Fc from IgG IgM, myeloma IgM, myeloma IgM, subunit Fc5 from IgM IgE Yu, myeloma IgAl myeloma IgAl, myeloma IgA2, myeloma HLA antigen class I, extracellular part HLA antigen class 11, extracellular part
Spin-labeling at carbohydrates
Spin-labeling at protein moiety
11 12 9.5
6 12-16" 50
7
7 6 66b 45h 2.5
55-60 41 32 45 8
2.7
14
Reference Timofeev et al. (1978) Timofeev et al. (1978) Timofeev and Lapuk (1982) Sykulev and Nezlin (1982) Sykulev and Nezlin ( 1982) Sykulev and Nezlin (1982) Sykulev and Nezlin (1982) Sykulev et al. (1984) Dudich et al. (1980) Sykulev et al. (1984) Nezlin et al. (1987, 1988) Nezlin et al. (1987, 1988)
As determined by steady-state Huorescence polarization (Nezlin et al., 1970; Dudich et al., 1978). The correlation time of the spin label bound to oligosaccharide(s) with restricted mobility.
chains being fixed not only in protein crystals, but also in solution. T h e correlation between crystallographic and spin-label data indicates that there is reason to believe that the correlation times obtained in spin-label experiments indeed reflect the mobility of the carbohydrate components of glycoproteins. Mobility of the IgCl oligosaccharide has also been investigated by ''C nuclear magnetic resonance (Rosen et al., 1979). In the nuclear magnetic resonance spectra two broad bands are seen, due to resonances from carbohydrates. T h e broad form of the bands suggests restricted mobility of the oligosaccharides relative to the protein moiety. Other Ig's have not one, but four and more, oligosaccharides per heavy chain, and the spin label can probably be incorporated in all of them. We can expect that the ESR spectra of labeled Ig's with a higher content of
30
K 0 A I . D NEZLIN
carbohydrates will reflect movement of the differing oligosaccharides with varied mobility, and the experimentally found correlation times often represent the mean values. However, the ESR spectra of IgE and IgAl spin-labeled on carbohydrates register two or more outer wide extrema. The presence of more than one wide extremum (better seen in the high field of the spectra; see Fig. 4) can be explained by the existence of differing relative mobilities of various oligosaccharide chains. T h e wide extremum, located distant from the central line of the spectrum, reflects the rotation of the spin labels bound to oligosaccharides with more restricted mobility. The correlation times calculated using these outer wide extrema are 66 nsec for IgE and 45 nsec for IgAl (Sykulev and Nezlin, 1982; Sykulev et al., 1984).The IgE and IgA oligosaccharides which possess these characteristically high correlation times are immobilized and rotated together with the whole molecule. We do not know to which part of the IgE and IgAl molecules they are bound. If they are located in the Fc portions, we could speculate that perhaps the Fc subunits of these Ig's are rigid, with no loose structure similar to the Fc of IgG. The wide extrema of ESR spectra for a spin-labeled myeloma IgM are broad and poorly resolved into subcomponents (Sykulev and Nezlin, 1982). The mean correlation times in this case were very low and similar for the whole molecule (7 nsec), the 7 S subunit (7 nsec), and the Fc5 fragment (6 nsec). A nearly identical value was obtained for another myeloma IgM: 9.5 nsec (Timofeev and Lapuk, 1982). It is difficult to
-
20°C. 0% sucrose
OOC, 38% sucrose
A part of the spectrum ol higher gain FIG. 4. Electron spin resonance spectra of human myeloma IgE Yu spinlabeled at carbohydrates.The presence of two outer wide extrema at higher gain can be explained by the existence of various relative mobilities of different oligosaccharides. Oligosaccharide(s)with more restricted rotation is characterized by longer 2A'l distance (Sykulev and Nezlin, 1982).
2%-
I N T E R N A L MOVEMENTS I N
Ig MOLECULES
31
interpret the physical significance of these values; most likely they reflect the loose structure of the Fc fragment of IgM. In the IgA2 molecule the spin labels incorporated in carbohydrates rotate surprising freely. They cannot be immobilized, even when in solution with a high concentration of sucrose (i.e., about 30%),when the whole glycoprotein molecule stops its rotation (Sykulev et al., 1984). Such freedom of rotation was also observed for ovalbumin oligosaccharide (Nezlin and Sykulev, 1988). More recently, we have studied the mobility of oligosaccharides of human leukocyte antigen (HLA), glycoproteins belonging to the Ig superfamily (Nezlin et al., 1987, 1988; Pankratova, 1988). In these experiments the extracellular portions of HLA classes I and I1 proteins were isolated and then spin-labeled on oligosaccharides. The correlation times obtained (i.e., 2.5-2.7 nsec) suggest a significant restriction of the mobility of spin labels bound to carbohydrates, due to interactions of the latter with the protein moiety (Table VIII). Information is still lacking regarding the exact localization of the spin label in the above-mentioned experiments; namely, to which of the several carbohydrate chains does the spin label incorporate and what sugar residues are tagged? The correlation times are the mean values and reflect the interrelationship of the oligosaccharide chains and the protein moiety in general. However, such experiments can also provide definitive information. For example, the spin-label method could be useful in investigating the mobility of Ig oligosaccharides with abnormal structure, such as those found in some pathological conditions (Roitt et al., 1988).As discussed above, the Fc, oligosaccharide, whose main function is probably to interact with cell receptors, is located on the protein surface. T h e Fc oligosaccharides of patients with such diseases as rheumatoid arthritis lack terminal galactose residues, which are important for fixation on protein surface. Such abnormal Ig's suffer a complete loss in their ability to bind to macrophage and monocyte Fc receptors and also demonstrate a reduced ability to induce cellular cytotoxicity (Rademacher et al., 1988). According to the attractive hypothesis by Parekh et al. (1989), one of the reasons for this defect could be the increased mobility of oligosaccharides lacking terminal glactose (Fig. 5 ) . This suggestion, of course, could be verified experimentally by the spin-label technique. If an oligosaccharide is firmly fixed on the protein surface and rotates together with a protein subunit, the rotational correlation time provides information regarding the flexibility or rigidity of the structure of this subunit. Even such general information can be useful in many cases, especially if X-ray crystallography data are not available. The method would, of course, be much more informative if the exact position of the spin label is known.
32
KOA1.D NEZ1.1N
Fab
Fab
Fab
Fab
0 Mannose @ N-acetylglucosamine
Galactose
FIG. 5. (Left) A schematic representation of an IgC molecule indicating the positions of conserved N-glycosylation sites ( A d g 7in CH2 domains). T h e arrow indicates the site of interaction between the a1-3 arms of two oligosaccharides. T he residues of the al-6 arm of each oligosaccharide are in contact with the surface of the protein. X-Ray crystallographic data show a well-defined galactose binding site. (Right) An IgC molecule from patient with rheumatoic arthritis. T he oligosaccharide chains terminate with N-acetylglycosarnine (Modified from Parekh el al., 1989).
IX. Conclusions
Ig’s are complex multifunctional proteins. They can react with several types of ligands, many of which are large molecules. In some cases the formation of complexes with various ligands takes place simultaneously and in different parts of the molecule. It is therefore not surprising that the studies discussed in this chapter provide strong evidence for the presence of diverse kinds of internal motion in the Ig molecule, which have evolved in favor of optimal interactions with various ligands. There is little doubt that the Fab arms of Ig molecules have enough freedom to rotate relatively independently. This so-called hinge flexibility, which is primarily dependent, in IgG, on the structure of the part of the hinge region between the hinge core and the Fab portion, greatly facilitates bivalent recognition of variably spaced antigenic determinants and permits the binding of two antigenic molecules or cells. It is not such an easy task, however, to describe the precise character of Fab movements. The rotational relaxation times obtained from fluorescence polarization measurements usually represent an average of several correlation times. We can, however, imagine that the Fab arms rotate more or less independently about the joints in the hinge region in one
INTERNAL MOVEMENTS IN
Ig MOLECULES
33
of the following ways: (1) angular waggling, (2) conelike wobbling, (3) twisting about the long axis, and (4)some movements about the short axis (Fig. 6) (Cathou, 1978; Hanson et al., 1981; Bennett and Huber, 1984). Another type of flexibility is presented by F (ab’n-Fc rotation, d u e to properties of the part of the hinge situated below the core. In the nanosecond range Fab rotates as a compact particle. However, after reaction with antigens, the angle between V and C region modules changes (so-called elbow bending). Contrary to Fab, the Fc fragment is flexible, which not surprising, since the CH2 domains are spatially separated. T h e Ig molecules are glycoproteins. T h e mobility of oligosaccharide chains, usually bound to the heavy chains, varies significantly. In Fc, the oligosaccharide chain is fixed spatially, but other Ig subclasses possess oligosaccharides capable of rotation more o r less independently from the protein moiety. Investigation and understanding of the mobility of oligosaccharides are important because of their role as ligands to cell receptors; arid changes in their freedom of movement would result, obviously, in pathological conditions. In summary, one can only begin to imagine from this review the wide variety of movements common to the various Ig molecules. This exceptional property is in direct conjunction with the functional diversity of these proteins: They are able not only to recognize antigens, but also to participate in a number of other important reactions. T h e functional sites are located in different parts of the Ig molecule, and flexibility seems to be a crucial factor for optimal binding to several different ligands, often simultaneously. T h e functional importance of flexibility is obvious after comparison with findings that hinge-deleted nonflexible Ig’s lack major effector functions. Such proteins are probably not just a genetic curiosity. According to H. F. Deutsch (cited by Rajan et al., 1983), about 1-2% of all IgC
A
.
.
B
.
C
FIG.6. Possible modes of rotation of the Fab subunit in an intact IgG molecule. (A) Angular waggling. (B) Conelike wobbling. (C) Motions along short and long axes.
34
K 0 A I . D NEZLIN
antibodies have no hinges. The functions of these proteins, which probably are abnormal, are still unknown. Study of t h e properties of this nonflexible Ig population could lead t o unexpected-and interesting-
findings.
REFERENCES
Abaturov, L. V., Nezlin, R., Vepgerova, T. I., and Varshavsky, J. M. (1969). Conformational studies of IgG and its subunits by the methods of hydrogendeuterium exchange and infrared spectroscopy. Biochim. Biophys. Acla 194, 386-396. Abola, E. E., Ely, K. R., and Edmundson, A. B. (1980). Marked structural differences in the Mcg Bence-Jones dimer in two crystal systems. Biochemistry 19, 432-439. Alpert, Y., Cser, L., Farago, B., FranPk, F., Mezei, F., and Ostanevich, Y. M. (1985). Segmental flexibility in pig IgC studied by neutron spin-echo iechnique. Biopolymers 24, 1769-1 784. Alzari, P. M., Lascombe, M.-B., and Poljak, R. J. (1988). Three-dimensional structure of antibodies. Annu. Rev. Immunol. 6 , 555-580. Anglister, J., Frey, T., and McConnell, H. M. (1984). Magnetic resonance of a monoclonal anti-spin-label antibody. Biochemistry 23, 1 138- 1 142. Beale, D., and Feinstein, A. (1976).Structure and function of the constant regions of immunoglobulins. Q. Rev. Biophys. 9, 135-180. Bennett, W. S., and Huber, R. (1984). Structural and functional aspects of domain motion in proteins. CRC Grit. Rev. Biochem. 15, 291-384. Burton, D. R. (1985). Immunoglobulin G: Functional sites. Mol. Immunol. 22, 161-206. Burton, D. R. (1986). Is IgM-like dislocation a common feature of antibody function? Immunol. Today 7 , 165-167. Burton, D. R., Dwek, R. A., Forsen, S., and Karlstrom, G. (1977). A novel approach to water proton relaxation in paramagnetic ion-macromolecule complex. Biochemistry 216, 250-254. Cathou, R. E. (1978). Solution conformation and segmental flexibility of immunoglobulins. Contemp. Immunol. 5 , 37-83. Cathou, R. E., Holowka, D. A., and Chan, L. M. (1974). Conformation and flexibility of immunoglobulins. Prog. Immunol. 2, 63-73. Cebra, J., Brunhouse, R., Cordle, C., Daiss, J., Fechheimer, M., Ricardo, M., Thunberg, A., and Wolfe, P. B. (1977). Isotypes of guinea pig antibodies: Restricted expression and bases for interactions with other molecules. Prog. Immunol. 3,269-277. Chan, L. M., and Cathou, R. E. (1977).The role of the inter-heavy chain disulfide bond in modulating the flexibility of immunoglobulin G antibody. J . Mol. Biol. 112,653-656. Cser, L., FranPk, F., Gladkikh, I. A., Nezlin, R. S., Novotny, J., and Ostanevich, Y. M. (1977). Neutron small-angle scattering study on two different precipitin types of pig anti-DNP antibodies. FEES Lett. 80, 329-33 1. Cser, L., Gladkikh, I. A., Franek, F., and Ostanevich, Y. M. (l98la).Investigation of antibody structure by scattering techniques. Colloid Polym. Sci:' 259, 625-640. Cser, L., Frangk, F., Gladkikh, 1. A., Kunchenko, A. B., and Ostanevich, Y. M.
INTERNAL MOVEMENTS IN
Ig MOLECULES
35
(198 1b). General shape and hapten-induced conformational changes of pig antibody against dinitrophenyl. A small angle scattering study. Eur. J . Biochem. 116, 109-116. Dangl, J. L., Wensel, T. G., Morrison, S. L., Stryer, L., Herzenberg, L. A,, and Oi, V. T. (1988). Segmental flexibility and complement fixation of genetically engineered chimeric human, rabbit and mouse antibodies. EMBO J . 7, 1989- 1994. Davies, D. R., and Metzger, H . (1983). Structural basis of antibody function. Annu. Rev. Immunol. 1,87-117. Davies, D. R., Padlane, E. A,, and Segal, D. M. (1975). Three dimensional structure of immunoglobulins. Annu. Rev. Biochem. 44,639-667. Davies, D. R., Sheriff, S., and Padlan, E. A. (1988).Antibody-antigen complexes. J . B i d . Chem. 22, 10541-10544. Deisenhofer, J. (1981). Crystallographic refinement and atomic models of a human Fc fragment and its complex with fragment B of protein A from Staphylococcus aureus at 2.9- and 2.8A resolution. Biochemistry 20,2361-2370. Deisenhofer, J., and Huber, R. (1983).T he contribution of X-ray crystallography to our understanding of immunoglobulin function. Prog. Immunol. 5,47-59. Deisenhofer, J., Colman, P. M., Epp, O., and Huber, R. (1976). Crystallographic structural studies of a human Fc fragment. 11. A complete model based on a Fourier map at 3.5A resolution. Hoppe-Seyler’s Z. Physiol. Chem. 357, 14211434. Deutsch, H. F., and Suzuli, T. (1971). A crystalline yG1 human monoclonal protein with an excessive H chain deletion. Ann. N.Y. Acad. Sci. 190,472-486. Dorrington, K. J., and Klein, M. H. (1982).Binding sites for Fcy receptors on IgG and factors influencing their expression. Mol. Immunol. 19, 12 15-122 1. Dorrington, K. J., and Klein, M. H. (1983).T he Cy2 region plays a major role in mediating the biological activities of IgG. Prog. Immunol. 5, 37-46. Dorrington, K. J., and Tanford, C. (1970). Molecular size and conformation of immunoglobulins. Adu. Imwiunol. 12, 333-38 1. Dudich, E. I., and Dudich, I . V. (1983). Polarization fluorescence, spin label and ultracentrifugal studies of specific interaction of low molecular weight proteins with the Fc fragment of human IgG. Mol. Immunol. 20,1267-1272. Dudich, I. V., Timofeev, V. P., Volkenstein, M. V., and Misharin, A. Y. (1977). Measurement of the time of rotational correlation of macromolecules by the ESR method in the case of a covalently bound spin label. Mol. B i d . (Engl. Trmsl.) 11, 685-693. Dudich, E. I . , Nezlin, R. S., and Franek, F. (1978). Fluorescence polarization analysis of various immunoglobulins. Dependence of rotational relaxation time on protein concentration and on ability to precipitate with antigen. FEBS Lett. 89, 89-92. Dudich, E. I., Dudich, I. V., and Timofeev, V. P. (1980). Fluorescence polarization and spin-label study of human myeloma IgA and IgM. Presence of segmental flexibility. Mol. Immunol. 17, 1335-1339. Dwek, R. A., Jones, R., Marsh, D., McLaughlin, A. C., Press, E. M., Price, N. C., and White, A. I . (1975). Antibody-hapten interactions in solution. Philos. T r a m R . Soc. London Ser. B 272,53-74. Edmundson, A. B., Ely, K. R., and Abola, E. E. (1978).Conformational flexibility in immunoglobulins. Contemp Top. Mol. Immunol. 7,95-118. Ely, K. P., Colman, P. M., Abola, E. E., Hess, A. C., Peabody, D. S., Parr, D. M., Connell, G. E., Laschinger, C. A., and Edmundson, A. B. (1978). Mobile Fc
36
KOALD NEZLIN
region in the Zie IgG2 cryoglobulin: Comparison of crystals of the F (ab’)y fragment and the intact immunoglobulin. Biochemistry 17,820-823. Ely, K. R., Herron, J. N., and Edmundson, A. B. (1983).Three-dimensional structure of the orthorhombic form of the Mcg Bence-Jones dimer. Prog. Immunol. 5,6 1-66. Endo, S., and Arata, Y. (1985).Proton nuclear magnetic resonance study of human IgGl and their proteolytic fragments: Structure of the hinge region and effects of a hinge region deletion on internal flexibility. Biochemistry 24, 156 1-1568. Feinstein, A., and Rowe, A. J. (1965).Molecular mechanism of formation of an antigen-antibody complex. Nature (London) 205, 147- 149. Feinstein, A,, Munn, E. A., and Richardson, N. E. (1971).The three-dimensional conformation of yM and yA globulin molecules. Annu. Rev. N . Y . Acad. Sci. 190, 104-121. Feinstein, A., Richardson, N., and Taussig, M. J. (1986).Immunoglobulin flexibility in complement activation. Immunol. Today 7, 169- 174. Fett, J. W., Deutsch, H. F., and Smithies, 0. (1973). Hinge-region deletion localized in the IgGl-globulin Mcg. Immunochemistry 10, 1 15-1 18. Gnewuch, T., and Sosnowsky, G. (1986).Spin-labeled carbohydrates. Chem. Rev. 86,203-238. Green, N. M. (1969).Electron microscopy of the immunoglobulins. Adv. Immunol. 11, 1-30. Hanson, D. C., Yguerabide, J., and Schumaker, V. N. (1981).Segmental flexibility of IgG antibody molecules in solution: A new interpretation. Biochemistry 20,6842-6852. Hanson, D. C., Yguerabide, J., and Schumaker, V. N. (1985).Rotational dynamics of IgG antibodies anchored in protein A soluble complexes. Mol. Immunol. 22,237-244. Holowka, D. A., and Cathou, R. E. (1976).Conformation of IgM. 2.Nanosecond fluorescence depolarization analysis of segmental flexibility in anti-s- 1 dimethylamino-5-naphthalenesulphonyl-~-lysine anti-immunoglobulin from horse, pig, and shark. Biochemistry 15,3379-3390. Hsia, J. C., and Piette, L. H. (1969).Spin-labelingas a general method in studying antibody active site. Arch. Biochem. Biophys. 129,296-307. Huber, R., and Bennett, W. S. (1983).Functional significance of flexibility in proteins. Biopolymers 22,26 1-279. Huber, R., Deisenhofer, J., Colman, P. M., Matsushima, M., and Palm, W. (1976). Crystallographic structural studies of an IgC molecule and an Fc fragment. Nature (London) 264,415-420. Isenman, D. E., Dorrington, K. J., and Painter, R. H. (1975).The structure and function of immunoglobulin domains. 11. T h e importance of interchain disulphide bonds and the possible role of molecular flexibility in the interaction between IgG and comp1ement.J. B i d . Chem. 114, 1726- 1729. Ishizaka, K. (1973).Chemistry and biology of immunoglobulin E. Antigens 1, 479-528. Ito, W., and Arata, Y. (1985).Proton nuclear magnetic resonance study of the dynamics of the conformation of the hinge segment of human G1 immunoglobulin. Biochemislry 24,6467-6474. Kaivarainen, A. I. (1975).Differential determination of rotational correlation times of spin-label proteins and bound labels. Mol. B i d . (Engl. Transl.) 9, 805-8 1 1.
INTERNAL MOVEMEN’I‘S I N
Ig MOLECLJLES
37
Kaivirainen, A. I . , and Nezlin, R. S. (1976). Spin-label approach to conformational properties of immunoglobulins. Immunochemistry 13, 1001- 1008. Kaivarainen, A. I., Nezlin, R.S., and Volkenstein, V. M. (1973).Spin-spin interaction between iminoxyl radicals localised in antibody combining sites. FEBS Lett. 35, 306-308. Kaivariiinen, A. I., Nezlin, R. S., and Volkenstein, V. M. (1974). Distance between iminoxyl radicals localized in antibody combining sites and relative freedom of rotation of antibody subunits. Mol. Biol. (Engl. Trawl.) 8, 816-823. Karplus, M., and McCammon, J. A. (1983). Dynamics of proteins: Elements and function. Annu. Reu. Biochem. 53, 263-300. Karush, F. (1978). T h e affinity of antibody range, variability, and the role of multivalence. Compr. Immunol. 5, 85-1 16. Klein, M., Haeffner-Cavaillon, N., Isenman, D., Rivat, C., Navia, M. A., Davies, D. R., and Dorrington, K. J. (198 1). Expression of biological effector function by IgG molecules lacking the hinge region. Pruc. Natl. Acad. Sci. U.S.A. 78,524-528. Leatherbarrow, R. J., and Dwek, R. A. (1984). Binding of complement subcomponent C l q to mouse IgC1, IgG2a and IgGPb, a novel Clq binding assay. Mol. Immunol. 21,32 1-327. Liu, B. M., Cheung, H. C., and Mestecky, J. (1981). Nanosecond fluorescence spectroscopy of human immunoglobulin A. Biochemistry 20, 1997-2003. Lovejoy, C., Holowka, D. A., and Cathou, R. E. (1977).Nanosecond fluorescence spectroscopy of pyrenbutyrate-anti-pyrene an tibody complexes. Biochemistry 16,3668-3672. Marquart, M., Deisenhofer, J., Huber, R., and Palm, W. (1980). Crystallographic refinement and atomic models of the intact immunoglobulin molecule Kol and its antigen-binding fragment at 3.0 A and 1.9 A resolution.]. Mol. Biol. 141,369-391. Nezlin, R. ( 1977). “Structure and Biosynthesis of Antibodies.” Plenum, New York. Nezlin, R. (1986). Distribution of TEMPO-DT spin label on immunoglobulin molecule. Fed. Proc., Fed. Am. Soc. Exp. B i d . 45, 740. Nezlin, R., and Sykulev, Y. K. (1982). Structural studies of immunoglobulins spin-labeled at the carbohydrate moiety. Mol. Immunol. 19, 357-366. Nezlin, R., and Sykulev, Y. (1984).ESR study of pFc’ fragment of IgG. Fed. Proc., Fed. Am. SOC.Exp. Biol. 43, 1432. Nezlin, R., and Sykulev, Y. K. (1988). Incorporation of spin label in carbohydrate components of glycoproteins-Application for structural studies of immunoglobulins and other proteins. In “Biomacromolecules in Spin Labeling” (G. I. Lichtenstein and R. I . Zhdanov, eds.), pp. 181-220. Nauka, Moscow. Nezlin, R., Zagyansky, Y. A., and Tumerman, L. A. (1970). Strong evidence for the freedom of rotation of IgC subunits.j. Mol. Biol. 50, 569-572. Nezlin, R., Zagyansky, Y. A., Kaivarainen, A. I., and Stefani, D. V. (1973). Properties of myeloma immunoglobulin E(Yu). Chemical, fluorescence polarization and spin-labeled studies. Immunuchemishy 10,681 -688. Nezlin, R., Timofeev, V. P., Sykulev, Y. K., and Zurabyan, S. E. (1978). Spinlabeling of immunoglobulin carbohydrates. Zmmunochemistry 15, 143-144. Nezlin, R., Arutyunyan, A. E., and Timofeev, V. P. (1985).Segmental mobility of COOH-terminal fragment of IgG heavy peptide chain in solution. Biofizika 30, 161-162. Nezlin, R., Pankratova, E. V., Arutyunyan, A., and Timofeev, V. P. (1987).
38
ROALD NEZLIN
Extracellular portions of HLA antigens are not compact globulae. Mol. Immunol. 24,803-806. Nezlin, R., Pankratova, E. V., and Timofeev, V. P. (1988).A spin-label study on the mobility of carbohydrate components in major histocompatibility complex. Biol. Membr. (U.S.S.R.) 3, 258-262. Noelken, M. E., Nelson, C. A,, Buckley, C. E., and Tanford, C. (1965). Gross conformation of rabbit 7 Sy immunoglobulin and its papain-cleaved fragments. J. Biol. Chem. 240,218-223. Oi, V. T., Vuong, T. M., Hardy, R., Reidler, J., Dangl, J., Herzenberg, L. A., and Stryer, L. (1984). Correlation between segmental flexibility and effector function of antibodies. Nature (London)307, 136- 140. Pankratova, E. V. (1988). Structure and functions of antigen-recognizing proteins coded by major histocompatibility complex genes. Mol. Biol. (Engl. Trawl.) 22,921-926. Parekh, R., Isenberg, D., Rook, G., Roitt, I., Dwek, R. A., and Rademacher, T. ( 1989). A comparative analysis of disease-associated changes in the galactosylation of serum IgG. J . Autoimmun. 2, 101-1 14. Phizackerley, R. P., Wishner, B. C., Bryant, S. H., Amzel, L. M., Lopez de Castro, J. A., and Poljak, R. J. (1979). Three dimensional structure of the pFc' fragment of guinea pig IgGl. Mol. Immunol. 16,84 1-850. Prasad, L., Vandoselaar, M., Lee, J. S., and Delbaere, L. T. J. (1988). Structural determination of a monoclonal Fab fragment specific for histidinecontaining protein of the phosphoenolpyruvate : sugar phosphotransferase system of E. coli. J. Biol. Chem. 263,2571-2574. Rademacher, T. W., Parekh, R. B., Dwek, R. A., Isenberg, D., Rook, G., Axford, J. S., and Roitt, I. (1988).T h e role of IgG glycoforms in the pathogenesis of rheumatoid arthritis. Springer Semin. Immunopathol. 10, 23 1-249. Rajan, S. S., Ely, K. R., Abola, E. E., Wood, M. K., Colman, P. M., Athay, R. J., and Edrnundson, A. B. (1983). Three dimensional structure of the Mcg IgGl immunoglobulin. Mol. Immunol. 20, 787-799. Reidler, J., Oi, V. T., Carlsen, W., Vuong, T. M., Pecht, I., Herzenberg, L. A., and Stryer, L. ( 1 982). Rotational dynamics of monoclonal anti-dansyl irnmunoglobulins. J . Mol. Biol. 158, 739-746. Richter, R., Nuhn, P., Ambrosius, H., Zagyansky, Y. A., Tumerman, L. A., and Nezlin, R. (1972).Restricted rotation of carp 15s immunoglobulin molecules as revealed by fluorescence polarization. FEBS Lett. 27, 184-1 86. Roitt, I., Dwek, R. A., Parekh, R., Rademacher, T. W., Alavi, A., Axford, J. S., Bodman, K. B., Bond, A,, Cooke, A., Hay, F. C., Isenberg, D. A., Lydyard, P. M., Mackenzi, L., Rook, G., Smith, M., and Sumar, N. (1980). T h e role of antigen in autoimmune response with special reference to changes in carbohydrate structure of IgG in rheumatoid arthritis. J . Autoimmun. 1,499-506. Romans, D. G., Tilley, C. A., Crookston, M. C., Falk, R. E., and Dorrington, K. J. (1977). Conversion of incomplete antibodies to direct agglutinins by mild reduction: Evidence for segmental flexibility within the Fc fragment of IgC. Proc. Natl. Acad. Sci. U.S.A. 74, 2531-2535. Rosen, P., Pecht, I., and Cohen, J. S. (1979). Monitoring the carbohydrate component of the Fc fragment of human IgG by 'C nuclear magnetic resonance spectroscopy. Mol. Immunol. 16,435-436. Roux, K. H. (1984). Direct demonstration of multiple VH allotypes on rabbit Ig molecules: Allotype characteristics and Fab arms rotational flexibility revealed by immunoelectron microscope. Eur. J. Immunol. 14, 459-464.
INI’EKNAI. MOVEMENTS IN
Ig MOI.ECULES
39
Roux, K. H., and Metzger, D. W. (1982). Imniunoelectron microscopic localization of idiotypes and allotypes on immunoglobulin molecules. J. Immunol. 129,2548-2553. Ryazantsev, S. N., Vasiliev, V. D., Abramov, V. M., Franek, F., and Zavyalov, V. P. ( 1989). Electron microscopy study of non-precipitating anti-dinitrophenyl antibodies. FEBS Lett. 244,291-295. Sakano, H., Rogers, J. H., Huppi, K., Brack, C., Traunecker, A., Maki, R., Wall, R., and Tonegawa, S. (1979). Domains and the hinge region of an immunoglobulin heavy chain are encoded in separate DNA segments. Nature (London) 277,627-633. Sandberg, A. L., Oliveira, B., and Osler, A. G. (1971). Two complement interaction sites in guinea pig immunoglobulins. J. Immunol. 106, 282-285. Schneider, W. P., Oi, V. T., and Yanovsky, C. (1987). Hybrid immunoglobulin isotypes of identical specificity produced by genetic recombination in E. coli and expression in lymphoid cells. Proteins 2, 8 1-89. Schneider, W. P., Wensel, T. G., Stryer, L., and Oi, V. T. (1988). Genetically engineered immunoglobulins reveal structural features controlling segmental flexibility. Proc. Natl. Acad. Sci. U.S.A. 85, 2509-2513. Schumaker, V. N., Seegan, G. W., Smith, C. A., Ma, S. K., Rodwell, J. D., and Schumaker, M. F. (1980). The free energy of angular position of the Fab arms of IgG antibody. Mol. Immunol. 17,4 13-423. Seegan, G. W., Smith, C. A., and Schumaker, V. N. (1979). Changes in quaternary structure of IgG upon reduction of the inter heavy chain disulphide bond. Proc. Natl. Acad. Sci. U.S.A. 76,907-9 1 1. Sheriff, S., Silverton, E. W., Padlan, E. A., Cohen, G. H., Smith-Gill, S. J., Finzel, B. C., and Davies, D. R. (1987). Three-dimensional structure of an antibodyantigen complex. Proc. Natl. Acad. Sci. U.S.A. 84, 8075-8079. Sheriff, S., Silverton, E., Padlan, E., Cohen, G., Smith-Gill, S., Finzel, B., and Davies, D. R. (1988). I n “Structure and Expression” (R. H. Sarma and M. H. Sarma, eds.), Vol. 1, pp. 49-53. Adenine Press, Albany, New York. Shimsliick, E. J., and McConnell, 14. M . (1972). KoIiitiod correlation time o f spi~~-l;ibeleda-chymotrypsin. Hiochrm. HI‘ofihy.s. Rrs. Co~nmun.46, 32 I 327. Silverton, E. W., Navia, M. A., and Davies, D. R. (1977). Three-dimensional structure of an intact human immunoglobulin. Proc. Natl. Acad. Sci. U.S.A. 74,5 140-5 144. Slattery, J., Holowka, D., and Baird, B. (1985). Segmental flexibility of receptor bound immunoglobulin E. Biochemistry 24, 78 10-7820. Steiner, L., and Lopes, A. D. (1979). The crystallizable human myeloma protein Dob has a hinge-region deletion, Biochemistry 18,4054-4067. Stryer, L. (1968). Fluorescence spectroscopy of proteins. Science 162, 526-533. Sutton, B. J., and Phillips, D. C. (1983). The three-dimensional structure of the carbohydrate within the Fc fragment of IgG. Biochem. Soc. Trans. 11, 130-132. Sykulev, Y. K., and Nezlin, R. (1982). Spin labeling of IgM and IgE carbohydrates. Immunol. Lett. 5, 121-126. Sykulev, Y. K., Timofeev, V. P., Misharin, A., and Nezlin, R. (1979). Spin label study of segmental flexibility of antihapten antibodies. Precipitating pig anti-DNP antibody is more flexible than non-precipitating. FEBS Lett. 101, 27-30. Sykulev, Y. K., Nezlin, R., German, G. P., Chernokhvostova, E. V., and Lavren-
40
K 0 A I . D NEZLIN
tiev, V. V. (1984).Structural studies of human IgAl and IgA2 immunoglobulins tagged with two different spin labels. Biofizika 29, 744-748. Timofeev, V. P., and Lapuk, V. (1982).Study of the irreversible conformational transition in IgM using spin labels introduced into the carbohydrate and peptide moieties of its molecule. Mol. Biol. (Engl. Truml.) 16, 325-333. Timofeev, V. P., Dudich, I. P., Sykulev, Y. K., and Nezlin, R. (1978). Rotational correlation times of IgG and its fragments spin-labeled at carbohydrate o r protein moieties. Spatially fixed position of the Fc carbohydrates. FEBS Lett. 89,191-195. Torano, A., Tsuzukida, J., Liu, Y. V., and Putnam, F. W. (1977). Location and structural significance of the oligosaccharides in human IgAl and IgA2 immunoglobulins. Proc. Nutl. Acad. Sci. U.S.A. 74,230 1-2305. Tumerman, L. A., Nezlin, R., and Zagyansky, Y. A. (1972a). Increase of the rotational relaxation time of antibody molecule after complex formation with dansyl-hapten. FEBS Lett. 19,290-292. Tumerman, L. A., Zagyansky, Y. A., and Nezlin, R. (1972b). Experimental evidence of the flexibility of IgG molecules. Mol. Biol. (Engl. Trunsl.) 6, 135- 147. Valentine, R. C., and Green, N. M. (1967). Electron microscopy of an antibodyhapten comp1ex.J. Mol. Bid. 27, 615-617. Weber, G. (1953). Rotational Brownian motion and polarization of the fluorescence of solutions. Adv. Protein Chem. 8,415-459. Weltman, 1. K., and Davis, R. P. (1970). Fluorescence Dolarization studv of a huma: IgA myeloma protein: absence of segmentai flexibility.J. Moi. Biol. 54. 177-185. Weltman, J. K., and Edelman, G. M. (1967). Fluorescence polarization of human yG-immunoglobulins. Biochemistry 6, 1437-1447. Willan, K. J., Golding, B., Givol, D., and Dwek, R. A. (1977).Specific spin labeling of the Fc region of immunoglobulins. FEBS Lett. 80, 133-136. Wrigley, N. G., Brown, E. B., and Skehel, J. J. (1983). Electron microscopy evidence for the axial rotation and interdomain flexibility of the Fab regions of IgG. J. Mol. Biol. 169, 771-774. Yguerabide, J. (1972). Nanosecond fluorescence spectroscopy of macromolecules. In “Methods in Enzymology” (C. H. W. Hirs and S. N. Timasheff, eds.), Vol. 26, pp. 498-577. Academic Press, New York.. Yguerabide, J., Epstein, H. F., and Stryer, L. (1970). Segmental flexibility in an antibody mo1ecule.J. Mol. Biol. 51, 573-590.. Zagyansky, Y. A. (1975). Phylogenesis of the general structure of immunoglobulins. Arch. Biochem. Biophys. 166, 371-381. Zagyansky, Y. A., and Gavrilova, E. M. (1974). Segmental flexibility of human myeloma immunoglobulins A. Zmmunochemistry 11,68 1-682. Zagyansky, Y. A., and Ivannikova, E. I. (197 . T h e general structure of shark (Squalis acantis) and hen (Galus domesti us) immunoglobulins. Mol. Bio. Rep. 1,301-304. Zagyansky, Y. A., Nezlin, R., and Tumerman, L. A. (1969). Flexibility of immunoglobulin G molecules as established by fluorescent polarization measurements. Immunochemistry 6, 787-800. Zagyansky, Y. A., Tumerman, L. A., and Egorov, A. M. (1972). Segmental flexibility of IgM molecules. Zmmunochemistry 9 , 9 1-94.
P
This article was accepted for publication on 27 October 1989.
ADVANCES IN 1MML:NOLOC;Y. VOI.. 48
Somatic Diversification of the Chicken Immunoglobulin Light-C hain Gene WAYNE 1. McCORMACK AND CRAIG B. THOMPSON Departments o f Micmbiology and Immunology and Internal Medicine, Howard Hughes Medical Institute, Univenity of Michigan Medical School, Ann Arbor, Michigan 48 109
I. Introduction
A central requirement for the humoral immune system is the ability to specifically recognize a wide variety of molecular antigens. It has been estimated that both humans and mice generate between lo6 and lo8 different antibody molecules during the process of creating an immunological repertoire. In these species the heterogeneity of antibody molecules results from a series of somatic recombinations which leads to the production of a functional immunoglobulin (Ig) molecule during B cell differentiation (reviewed by Tonegawa, 1983; Alt et al., 1986; Hunkapiller and Hood, 1989). A functional heavy-chain gene is assembled in each B cell from an assortment of variable (V), diversity (D), and joining (J) elements. A functional light-chain gene is assembled from an assortment of V and J sequences. These joining events themselves lead to the generation of additional diversity through variations in the precise joining point and d e nova nucleotide addition at the joint (Alt and Baltimore, 1982). Whereas mammalian light chains might be expected to have less diversity than heavy-chain genes, because they lack D segments, additional light-chain diversity has been generated in mammals by having two independent genes ( K and A), either of which can recombine to encode a functional Ig molecule in conjunction with heavy-chain gene. Not all species, however, use somatic recombination as a means of generating the primary Ig gene repertoire. The chicken Ig light-chain (IgL)locus contains a single VL gene segment capable of rearrangement (Reynaud et al., 1985). Despite this, chickens have been shown to display considerable heterogeneity in their circulating Ig light chains, as demonstrated by isoelectric focusing (Jalkanen et al., 1984). Recent experiments have suggested that'chickens create diversity within their IgL genes by a process of gene conversion, using sequence templates derived from V region pseudogene segments (QV) located 5' from the single rearranged VL gene segment (Reynaud et al., 1987; Thompson and Neiman, 1987). 41 Copyright 0 1990 by Acddenlii Press. Inc. All rights of reproduction in any form reserved.
42
WAYNE T. McCORMACK A N D CRAIG B. THOMPSON
This diversification process is induced during the clonal expansion of Ig' B cells in the bursa of Fabricius and leads to the generation of a functional immunological repertoire from as few as 3 x lo4 Ig+ cell precursors. Understanding the molecular regulation of this extensive gene conversion process may have important implications for our understanding of somatic diversification of mammalian Ig genes, as well as allow us to characterize the regulation of somatic gene conversion in higher eukaryotes. In this chapter we review the current state of knowledge concerning how the chicken generates an immunological repertoire for its unique IgL gene during B cell development in the bursa of Fabricius. 11. Bursa of Fabricius Is Essential for Normal B Cell Development
The bursa of Fabricius first aroused the interest of immunologists when Glick et al. (1956) described the central role of the bursa of Fabricius in the production of antibodies in chickens. This work led to the separation of lymphocytes into B (bursal-derived) cells essential for antibody production and T (thymus-derived) cells required for delayed-type hypersensitivity and cell-mediated immunity (reviewed by Cooper et al., 1984; Cantor, 1984). This division of lymphocytes into B and T cells became a cornerstone of modern immunology, and the mammalian counterparts of both B and T cells were soon defined. Despite the apparently central role that the bursa of Fabricius plays in avian B cell development, a single mammalian counterpart for the bursa of Fabricius has never been defined. However, in the 30 years since Glick et al.'s original discovery, the role of the bursa of Fabricius in B cell ontogeny has continued to interest avian immunologists. As a result of the ease with which avian embryos can be experimentally manipulated, a cellular characterization of bursal development has been carried out in remarkable detail (for detailed reviews, see Grossi et al., 1976; Glick, 1977; Ratcliffe, 1985; Pink, 1986). In brief, the bursa of Fabricius is a lymphoepithelial organ that arises as an invagination at the base of the posterior cloaca. Beginning at day 4 of embryogenesis, the bursal epithelial anlage begins to proliferate, forming epithelial buds within the underlying lamina propria (Fig. 1). Between days 8 and 14 of embryogenesis, lymphoid cells colonize the epithelial buds to form bursal follicles (Moore and Owen, 1966; Houssaint et al., 1976, 1983). Each follicle is colonized with two to seven lymphoid cells, and subsequent growth of the follicles is due primarily to cellular proliferation (Le Douarin et al., 1975; Lydyard et al., 1976; Pink et al., 1985a; Pink, 1986). There are approximately 10,000 follicles in the bursa of a 4-week-old chick, each containing about lo5 lymphocytes (Olah and Glick, 1978).
SOMATIC DIVEKSIFICATION OF
Igl,
I Hatching
Embryogenesis
43 Posrharching
Granulocytes
9
Prebursal
Stem Cells 0 0
\
Bursa1 Stem Cells
7
-0
0 0
(Days) 5
10
15
20 I(Weeks) 1
2
3
4
FIG.1 . Schematic representation of the basic stages of B cell development in the bursa of Fabricius.
Both the lymphoid and epithelial components of the bursal follicle have been shown to be required for normal B cell development. If the bursal epithelial anlage is removed surgically at 60 hours of embryogenesis (Jalkanen et al., 1983a,b; Corbel et al., 1987) or the differentiation of the epithelial component is inhibited by testosterone treatment between days 6 and 8 of embryogenesis (Cooper et al., 1969; Huang and Dreyer, 1978),the resulting birds are profoundly immunodeficient (Eerola et al., 1983, 1984; Granfors et al., 1982; Jalkanen et al., 1984). Although these birds develop nearly normal numbers of Ig+ cells and levels of circulating Ig at 6 months of age, they are unable to mount primary or secondary immune responses. Nearly all of the surface and circulating Ig’s in these birds is IgM. T h e lymphoid component of the bursal follicle can be depleted specifically by treatment of the developing embryo with cyclophosphamide on days 15-17 of embryogenesis (Lerman and Weidanz, 1970). The grown birds not only demonstrate profound immunodeficiency, but also fail to develop significant numbers of Ig’ cells or circulating Ig’s. In these birds the epithelial component of the bursa remains intact, and, if reseeded with lymphocytes derived from an embryonic bursa, B cell immunity will develop normally (Toivanen and Toivanen, 1973).The embryonic bursal lymphocytes that can reseed the bursal epithelium and regenerate the secretory immune system have been termed “bursal stem cells.” T h e bursal stem cells are Ig’, are present at reasonable levels between day 15 and the time of hatching, and are rapidly lost posthatching (Pink et al., 1985b). They can no longer be detected by a cell transfer assay by 2-4 weeks of age. Interestingly, if the entire bursa is removed surgically
44
WAYNE
'r. M ~ C O K M A C K A N D
C R A I G B . THOMPSON
between days 17 and 18 of embryogenesis, close to 50% of the bursectomized birds will be completely deficient of Igf cells and circulating Ig (Cooper et al., 1969). It appears that all of the cells destined to generate the humoral immune system reside in the bursa during this developmental period. Posthatching, the bursa of Fabricius continues rapid growth for 2-4 weeks. Thereafter, it reaches a plateau and begins to involute between 4 and 6 months of age, as the bird reaches sexual maturity (Glick, 1977). Bursectomy after several days posthatching does not result in loss of the chicken's humoral immune system (reviewed by Pink et al., 1987; Toivanen et al., 1987). Based on the above observations, it has been concluded that the bursa of Fabricius is the normal developmental site where the precursors of mature B cells are expanded within the developing embryo and the primary immunological repertoire of the chicken is generated. Parallel studies in mammals have shown that mammalian B cell precursors are produced continuously from bone marrowderived precursors and that these cells create the primary immunological repertoire through recombination of their Ig genes from a large pool of precursor gene segments (Alt et al., 1986). 111. Structure of the Unique Chicken lgL Gene
Protein structural and sequence data (Hood et al., 1970; Grant et al., 1971) demonstrated that the circulating Ig's of the chicken all share a single A-like light-chain isotype, suggesting the presence of a single IgL gene. A chicken IgL cDNA was first cloned by Reynaud et al. (1983). T h e initial characterization of the chicken IgLgene (Fig. 2) showed that it was composed of the same functional domains as the IgL genes of mammals. A 2 1-amino-acid hydrophobic leader segment was followed by a 92amino-acid V gene segment linked to a J gene segment of 13 amino acids and a constant (C) domain of 103 amino acids. Interestingly, when the V and C region gene segments were used to probe Southern blots of DNA derived from developing polyclonal bursa1 lymphocytes, the cells were found to share a single functional rearrangement of the IgI>gene (Fig. 2) (Reynaud et al., 1985). In order to characterize this rearrangement, Reynaud et al. (1987) cloned and characterized the genomic organization of the chicken IgL gene. As shown in Fig. 3, these studies revealed that the gene contains a single functional VL gene segment separated by 1.8 kb of DNA from a single functional JI. gene segment, which is located 2 kb 5' from a single constant region segment. The functional V element, designated VL,is separated by a small 125-bp intron from the leader segment. The unique
SOMATIC DIVERSIFICAI'ION OF
S T
II
M A I
i i
E X
P N I
I
I I V REGION PROBE
M
G B
V REGION
PROBE
II
45
C
J
V
Igl.
c;
A I
' I
C REGION PROBE
M
G B
C REGION PROBE
FIG. 2. The chicken IgL cDNA structure and genomic organization. (Top) The chicken IgLcDNA structure, including germ-line-encoded restriction endonuclease sites, is depicted. (Bottom) Southern blcts containing germ-line DNA (G) derived from chicken erythrocytes and bursal cell-derived (B) DNA isolated from a polyclonal population of Ig' bursal lymphocytes at 6 weeks of age digested with the restriction endonuclease BclI and hybridized with probes specific for the Igl. V and C gene regions. (Left) The V gene region probe hybridizes to multiple homologous segments in the genome. In contrast (right), the C gene region probe hybridizes in germ-line DNA to a single band. In bursal DNA both probes identify a single new band migrating at 8 kb, which comigrates with both probes. The intensity of this rearranged band is of intensity approximately equal to that of the remaining germ-line band. Nearly 100% of the cells isolated from the bursa express surface Ig, suggesting that nearly all of the rearranged alleles present in this polyclonal population are functional. Marker bands (M) are endlabeled Hind111 fragments of A DNA (23.1, 9.4, 6.7, 4.4, 2.3, and 2.0 kb). UT, Un translated.
leader segment is 5' from a typical Ig promoter containing a conseived octamer box 32 bp upstream from a T A T A box. This organization is also characteristic of mammalian Ig V genes (Parslow et al., 1984). Within t h e 22 kb upstream from the single rearranging VL gene segment, Reynaud et al. (1987) identified 25 V gene-homologous gene segments in both
46
W A Y N E T. McCORMACK A N D C R A I G B . THOMPSON
FIG. 3. Schematic representation of the IgL locus for both rearranged and unrearranged alleles. The IgL locus is organized in the germ-line with a single functional V gene element, designated VLI,separated by 1.8 kb of intervening DNA from a single J gene segment. The J L segment is 2 kb upstream from a single C gene segment. During rearrangement, juxtaposition of V L and J L sequences occurs by deletion of the intervening DNA and creation of a V-J joint. Upstream from the single functional V1. gene element are 25 V gene-homologous segments numbered in order of their occurrence from 3' to 5' from the VL, gene and depicted as determined by Reynaud el al. (1987). Arrows designate the orientation relative to the V L gene. ~
transcriptional orientations. All 25 of these V gene segments lack leader segments, as well as recombination signal sequences, and have therefore been designated V segment pseudogenes (rjrV1-25; see Fig. 3). These data were consistent with the observation that there is a single functional rearrangement of the IgL gene which leads to the juxtaposition of the VL gene segment with the single J L gene segment in all developing bursa1 lymphocytes. The presence of single functional IgI. V and J gene segments and the resultant limited potential for functional rearrangement contradicted the hypothesis that the bursa of Fabricius induced the generation of the immunological repertoire in the chicken by generating combinatorial diversity through stimulation of Ig gene rearrangement. Therefore, in the last several years the regulation of Ig12 gene rearrangement and diversification during development has been the subject of intense investigation by several laboratories. IV. Rearrangement of the IgLGene
The chicken IgL locus, because of its small size and unique germ-line structure, has readily lent itself to studies characterizing the timing and molecular mechanism of V-J joining. Joining of the single functional V segment with the single J segment results in the deletion of the DNA between VI, and J L from the genome. This deletion was found to be accomplished by a molecular mechanism which results in the precise
SOMATIC DIVERSII;I<:AI'ION OF
Igl.
47
joining of the Vl, and JL recombination signal sequences, leading to the formation of a circular episomal element (McCormack et al., 1989c), similar to the circular DNA intermediate formed by T cell receptor gene rearrangement (Okazaki et al., 1987; Fujimoto and Yamagishi, 1987). Despite the precision observed in the generation of the signal joint, the joining of the chicken VL and J L coding sequences appears to occur randomly. Just as in mammals, the coding ends are susceptible to exonucleolytic activity prior to their ligation to form a coding joint. Whereas the random nature of V-J joining leads to a considerable proportion (67%) of rearrangements being out of frame, it also serves to generate diversity within the V-J joining region of in-frame clones. At least 16 independent functional V-J joints have been identified. Although chicken B cell progenitors apparently lack terminal deoxynucleotidy1 transferase activity, it appears that single nonrandom nucleotides are added to the end of the V-J coding segments during the rearrangeThese nucleotides do not resemment process (McCormack et al., 1989~). ble the N segments associated with mammalian V-D-J joining (Alt and Baltimore, 1982; Milner et al., 1986),because they are not G-rich polynucleotide additions, but instead are specific single-base pair additions to the 3' end of VL and/or the 5' end of JL. In addition to creating a small amount of additional non-germ-line diversity to V-J joints, the significance of these base pair additions at the V-J joints remains to be determined. T h e circular episome generated by joining of the conserved heptamers 3' from VL and 5' from JL.is apparently not propagated during cell division, because it was found to be absent in B cell lines and posthatching bursa1 lymphocytes (McCormack et al., 1989b). This is presumably because the 1.8-kb episome lacks an origin of replication and a centromere, both of which would be required for its propagation during cellular proliferation. Therefore, the presence of the episome can be used as a marker for the rearrangement process itself, because it is only present transiently following IgL gene rearrangement. By using the polymerase chain reaction to amplify the signal joint present on the episome (Saiki et al., 1988),it is possible to determine when and where IgL gene rearrangement occurs during embryonic development. The signal joint is readily detectable in all hematopoietic tissues between days 10 and 15 of embryogenesis (McCormack et al., 1989b), reaching peak abundance at approximately day 15 of embryogenesis and disappearing from all hematopoietic tissues by the day of hatching (Fig. 4). Despite the fact that there is a large expansion of the number of Ig' cells in the bursa of Fabricius between days 15 and 18 of embryogenesis (Lydyard et al., 1976), there is a dramatic decrease in the relative abun-
-
A
u3
+
I
v-J joining
I
B
FIG.4. Detection of the signaljoint episome during avian B cell development. (A) The experimental design to detect the signal joint formed during IgL gene rearrangement. Polymerase chain reaction (PCR) primers 994 and 993, which face toward the recombination signal sequences on VL and Jl,, respectively, were prepared and added to samples of DNA isolated from various tissues. In the presence of a circular episome, these primers would face toward each other across the signal joint, and therefore amplify DNA across the joint. This DNA can be specifically detected following a PCR reaction by hybridization of the products with the U 3 probe. (B) Bursa1 and splenic DNAs isolated from various developmental time points, from 12 days of embryogenesis to 6 weeks posthatching, were amplified by PCR, and the resultant products were separated on a 1% agarose gel and hybridized with the U 3 probe. In both tissues the transient appearance of signal joints was detected at 15 days of embryonic development. In separate experiments using additional rounds of PCR amplification, signal joints could also be detected at 12 and 18 days of development, at lower abundance. H, Hatching.
dance of the signal joint episome. These data argue that the increase in bursal lymphoid cells with rearranged IgL genes between days 15 and 18 of embryogenesis results from migration and/or proliferation of cells with a rearranged IgL gene rather than continued IgL gene rearrangement within the bursa. V. Early Role of the Bursa: Selection of lg' Precursors
One surprising feature of mature chicken B cells is that virtually every Ig' cell contains both a functionally rearranged IgL allele and a germ-line
allele. These data were originally interpreted to suggest that virtually every Igl, gene rearrangement might be functional (i.e., in frame) in nature. However, as mentioned, considerable junctional heterogeneity was observed in V-J joints obtained between days 10 and 15 of embryogenesis (Fig. 5). Two thirds of the time, the random nature of V-J joining leads to V-J joints that are out of frame (McCormack et al., 1989b). Despite this, few mature B cells in the chicken contain two rearranged IgL genes. Together, these data suggest that during B cell development (1) B cells rearrange only one allele and (2) there is positive selection for cells with a functionally rearranged IgL gene. As mentioned, IgL gene rearrangement occurs for only a limited time during embryonic development, most IgL rearrangements occurring between days 10 and 15 of development. T h e fact that both rearranged V-J joints and signal joints could be isolated from extrabursal hematopoietic tissue during this period suggests that the bursa is not required for IgL gene rearrangement. These data are consistent with previous studies that showed that removal of the bursal anlage after 60 hours of development did not prevent the eventual expression of normal levels of serum Ig or the appearance of circulating surface Ig' B cells (Granfors et al., 1982; Eerola et al., 1983; Jalkanen et al., 1983a). However, as demonstrated on isoelectric focusing gels, the serum Ig isolated from chickens bursectomized at 60 hours of development displays extremely limited IgI, diversity (Jalkanen et al., 1983b, 1984).T h e small amount of diversity observed may be due simply to the heterogeneity that exists in the functional V-J joints created by the rearrangement process. Although the bursa is not required for Ig gene rearrangement, it does influence the set of rearranged V-J joints expressed. More specifically, it appears that the bursa serves to selectively expand cells with productive IgL genes during embryonic development. When V-J joints were analyzed at three developmental time points from the bursa of Fabricius, the percentage of in-frame joints increased from 33% to 94% between days 10-12 and day 18 of embryonic development (Table I). Although the
50
W A Y N E T. Mc CORMACK A N D CRAIG B. THOMPSON
< germline
" L < > >--J
GAC AGC AGC AGT ACT GCT G
T GGT ATA TTT GGG GCC GGG
FIG.5. V-J joints detected between days 10 and 15 of embryogenesis. V-J joints were amplified by polymerase chain reactions from both splenic and bursa1 tissues between days 10 and 15 of embryogenesis, and the resulting joints are depicted. A wide variety ofjoining positions in both VLand J L are utilized during this developmental period in the creation of joints, and many of the joints identified are out of frame, as would be expected for a random process. The addition of non-germ-line base pairs to the coding ends prior to ligation, with the addition of a C to the 3' end of the germ-line VL gene segment and an A added to the 5' end of the full J L segment, is suggested from the analysis of nucleotides present at the joint.
mechanism for this selective expansion has not been clearly defined, it is possible that the target structure for positive selection is the surface Ig molecule. Thus, selective amplification of cells with productive gene rearrangements could result from antigen-induced proliferation of surface Ig' cells within the bursa. Such an interaction might be mediated through the interaction of germ-line-encoded surface Ig molecules with a bursal-specific antigen(s).
51 TABLE 1 PERCENTAGE OF PRODUCTIVE VERSUS NONPRODUCTIVE V-J Jorwrs CLONED FROM THE BURSAAT DIFFERENT DAYS OF DEVELOPMENT” V-J joints Day
Productive
Nonproductive
10-12 15 18
53% 79% 94%
67% 21% 6%
a
>30 clones analyzed at each time point.
Bursal antigen-induced proliferation could account for the observed exponential growth of embryonic surface Ig+ bursal cells (Lydyard et al., 1976; Reynolds, 1987). Although such a mechanism would account for the bursa’s ability to induce the large-scale amplification of B cell progenitors, it provides no explanation for how an effective immunological repertoire could be created from a single functional IgL gene (see below). VI. Bursal Stem Cell
As mentioned in Section 11, the Ig+ cells that can home to the bursal follicle and regenerate the lymphoid component of a follicle have been termed “bursal stem cells.” T h e frequency of bursal stem cells in the developing bursa can be estimated based on bursal reconstitution experiments. For example, lo6 lymphoid cells from an 18-day embryo can be used to repopulate the bursal follicles of a second 18-day embryo which has had the lymphoid component eliminated by cyclophosphamide treatment between days 15 and 17 of embryogenesis. Based on reconstitution with cells bearing congenic markers or distinct IgLalleles, it appears that individual follicles are reconstituted by one, or at the most two, stem cells (Pink et a/., 1985b; Weill et al., 1986). Because there are approximately lo4 bursal follicles, the frequency of bursal stem stems at 18 days of embryogenesis is, therefore, approximately one in 50- 100 cells. As in the previous section, we would suggest that the cell that accounts for the expansion of the lymphoid component of individual bursal follicles is a cell that expresses the unmodified Ig molecule encoded by germ-line gene segments. In order to determine the frequency of such cells, we sequenced 42 randomly chosen rearranged V-J segments from
52
W A Y N E T. Mi.<:OKMA(:K A N D CKAIC; B. T H O M P S O N
18-day bursal lymphocytes. One of 42 of these V-J segments retained the germ-line sequence of the VLgene segment. Therefore, the frequency of rearranged but undiversified V segments is of the same order of magnitude as the frequency of bursal stem cells in the population at this time point. The rapid decline in the presence of bursal stem cells between 18 days of embryogenesis and 2 weeks of age could be accounted for by the diversification of the rearranged V gene segment in these remaining cells. We would therefore like to propose a molecular definition of the bursal stem cell as a cell which retains functionally rearranged Ig heavychain (IgH)and IgLgenes that have not yet undergone extensive somatic diversification. VII. Developing Bunal lymphocytes Acquire Diversity within the Rearranged lg, V Gene Segment
Although a small amount of heterogeneity is created during the formation of the V-Jjoint within the IgL gene locus, the observed heterogeneity is insufficient to account for the large immunological repertoire of the chicken immune system (Pink et al., 1981; DuPasquier, 1982; Weill et al., 1987). Sequences of IgL cDNAs obtained from B cells in peripheral lymphoid organs revealed significant sequence alterations from the sequence of the germ-line V gene segment (Reynaud et al., 1985; Parvari et al., 1987). These suggested that some form of somatic diversification mechanism occurred during avian B cell development to modify the sequences of the IgL gene. T o characterize this somatic diversification process, the rates of loss of restriction endonuclease sites present within the rearranged and germ-line IgL genes during chicken B cell development were analyzed (Fig. 6). Because all IgL genes rearrange the same functional elements to become activated, the study of these genes provides a unique opportunity to gain insights into how Ig diversity developed in a heterogeneous population of B cells. These results indicated that extensive sequence modification occurred on the rearranged VLgene segment during normal B cell development (Thompson and Neiman, 1987). This diversification was found to occur at multiple sites throughout the V gene segment and was not confined to sequences involved in V-J joining. The diversification process occurred subsequent to V-Jjoining during B cell development in the bursa of Fabricius. By restriction analysis, the diversification was found to be highly selective. Diversification occurred in the rearranged V gene segment, but not in the unrearranged V gene segment. Coding sequences in the leader and J L gene segments of both the rearranged and
SOMATIC DIVEKSIE'ICA'I'ION C)E'
Igl
53
Normal Bursa i5d18d H 1 2 4 w k s -M -G
Sca / Sol
-R uo
XC
Germlme
(GI
Sol Sca Kpn
LVJ
C
LVJ
C
Rearranged ( R 1 Somatic Diversification ( M )
FIG. 6. Somatic diversification of the IgL gene locus during early bursal development. (Bottom) The germ-line and primary rearrangements of the chicken Igl. gene are shown. (Top) DNA isolated from bursal cells at various ages of development was digested simultaneously with ScaI and SalI. A Southern blot was prepared and hybridized with the AC probe. T he positions of the bands representing the germ-line (G), rearranged (R), and modified (M) rearranged bands are indicated (top), and the derivation of each band is diagrammed (bottom). Diversification, as detected by a loss of the ScaI site in the V L gene region of the rearranged allele, begins in the embryonic bursa soon after IgL rearrangement. By 4 weeks of age, over 90% of bursal cells have lost the VLScaI site within the rearranged allele as a result of gene diversification. H , Hatching.
unrearranged IgL genes do not undergo detectable rates of sequence modification. T h e diversification of the rearranged VL gene segment within the bursa of Fabricius was shown to lead to Ig sequence heterogeneity in mature B cells, because cells that have migrated from the bursa of Fabricius appear to be enriched for cells that have undergone the loss of germ-line VL restriction endonuclease sites. The above data demonstrate a second role for the bursa of Fabricius in
54
WAYNE T. Md:ORMACK A N D CRAIG B. THOMPSON
the development of avian B cells. In addition to generating the clonal expansion of B cells during early embryonic development, B cell development in the bursa of Fabricius is associated with extensive somatic diversification of the rearranged V gene segments of these cells. This process would be expected to alter the antigenic specificity of the V domain during B cell development in the bursa of Fabricius. This molecular mechanism may explain how the bursa might effectively generate an immunological repertoire, as well as select against self-reactive clones. If both exponential growth and gene conversion were activated in Ig' B cell progenitors migrating to the bursa of Fabricius, a given cell would continue to undergo exponential growth and clonal expansion until the surface Ig receptors of all of its clonal offspring no longer reacted with antigen@)in the embryonic bursal microenvironment. If the diversification process occurred independently in each individual cell, a wide degree of sequence heterogeneity within the rearranged variable gene segment could be generated by such a process. Such a model also allows for the selection of B cell tolerance to self-antigens expressed within the bursa during the somatic generation of the immunological repertoire. Consistent with this hypothesis, an increasing proportion of surface Ig' bursal cells cease exponential growth and enter a quiescent phase between the day of hatching and 4 weeks of age (Thompson et al., 1987), the time at which the mature repertoire is being completed by the somatic diversification process and cells migrate from the bursa to peripheral lymphoid tissues (Thompson and Neiman, 1987). VIII. Sequences That Lead to Diversification of the Rearranged VLGene Segment Have Identical Counterparts in the $V Gene Segments Located 5' from VL
The above data are consistent with the hypothesis that a significant immunological repertoire can be generated through somatic diversification of rearranged variable gene segments during B cell development in the bursa of Fabricius. They do not, however, provide insights into the molecular mechanisms which lead to the diversification of the VL gene immunological repertoire. If the diversification process occurred entirely at random, it would lead to a significant amount of cell waste through the generation of somatic mutations that altered structural features required for the generation and cell surface expression of an intact Ig molecule. Based on the initial observation that sequence blocks found in a splenic IgL cDNA clone were shared in a $VL gene segment, Reynaud et al. (1985) sequenced all 25 of the $V gene segments 5' from the rearranging VL gene segment in an inbred chicken strain. By comparing the sequences of the pseudogenes with the sequences present in 12 cDNAs
SOMATIC DIVERSIFICATION OF
55
Igi,
isolated from two distinct developmental time points, these investigators were able to demonstrate that the sequence heterogeneity contained within the rearranged V gene segments was reflected in the sequence heterogeneity that exists within the $V gene segments (Reynaud et al., 1987). Moreover, diversification of individual VL gene segments appeared to occur through the replacement of blocks of VL sequence with sequences derived from the $V segments (Fig. 7). These replacements ranged from 10 to 120 bp in length.
G
L 1 CTC GCC CAC ACC TCA GGT TCC CTG GTG CAG GCA GCG CTG ACT CAG CCC TCC TCG GTG TCA GCA AAC CCG GW GAA ACC
WR 1
GTC AAG ATC ACC TGC TCC GGG W l AGG AGC TAC TAT GGC TGG TIC CAG CAG AAG GCA CCT GGC ACT GCC CCT GTC ACT (GGT AGC) c,54 ..................... .G.*..C ................................................... C
Jv5
AGC) .G.*..C ...........................
(GGT
.....................
*v7 . T .
7.7
............
-1.
......... 1.. .....................
CDR 2 ATC TAT GCT AAC ACC AAC AW CCC TCG GAC ATC CCT TCA CCA TTC TCC GCT TCC MA TCC GGC TCC ACA GCC ACA
G
CTC
(-154
......... TAC
... GA.
..G
................................................
..G .G. ...
pJ7
......... TIC ... GA.
..G
................................................
..G .G.
G
CDR 3 & J T T A ACC ATC ACT GGG CTC CIA GCC CAC W C GAG GCT GTC TAT TAC TGT GGG AGT CCA GAC AGC AGC ACT ACT GAT GGT
c,54
........................
..G
........................
..c
)V7
........................
..G
........................
..c
AG.
...
..................... .........
TA. GT. .G. A T G
FIG. 7. Sequence comparison of germ-line VL, an 18-day embryo cDNA (C154) and VL pseudogenes 5 (+V5) and 7 (+V7). In the cDNA there have been 24 bp substitutions in comparison to the VL gene segment, including a 6-bp insertion within CDRI. All 12 of the 5' sequence substitutions can be found within the +V5 gene segment. All 12 of the 3' sequence substitutions can be found within the $V7 sequences. Analysis of this cDNA and multiple additional ones has led to the hypothesis that V gene region diversification occurs through the copying of blocks of sequence present in the pseudogenes into the rearranged VL gene segment.
56
W A Y N E T. M(.CORMACK A N D C R A I G B. T H O M P S O N
Sequences derived from 6 weeks of bursal development showed significantly greater sequence divergence from VL than sequences derived from 18-day cDNAs. Sequence analysis revealed that this increased heterogeneity was consistent with repetitive sequence substitution events occurring within individual rearranged VLgene segments (Fig. 8). Rearranged V gene segments from 18-day embryos revealed one or two sequence substitution events, while V gene segments from 3-week-old chicks were composed of three to seven individual events (Reynaud et al., 1987). Although it appeared that the more proximal (or 3') $V gene segments were used more frequently as sequence donors than the more distant (or 5 ' ) $V gene segments, no other apparent rules for sequence substitutions can be derived from their data. Neither the 5' nor 3' position of sequence diversification nor the size of gene conversion events appeared to be conserved. The genomic organization surrounding the rearranged VL gene segment was unaltered by somatic diversification. IX. Diversification Results in a Cell in Which the VI Gene Segment Has Undergone Gene Conversion
Two potential molecular mechanisms (Fig. 9) could account for the sequence diversification observed by Reynaud et al. (1987). T h e first would be a recombination event that involves a sequence exchange between the rearranged VL gene segment and a donor $V gepe. This double-homologous recombination would transfer homologous gene segments between the functional gene segment and the individual pseudogenes. Alternatively, the diversification might occur through a gene conversion process, perhaps initiated by double-stranded breaks within the recipient rearranged IgL gene. The possibility that double-stranded breaks might be created within the rearranged variable gene segment was raised by the observation by Weill and colleagues that a consensus heptamer sequence is found in the germ-line V gene segment at multiple positions (Reynaud et al., 1987), and by the observation by Hope et al. (1986) that a heptamer-specific endonuclease is found in bursal cell extracts. Such a gene conversion process would result in the copying of pseudogene segments into doublestranded gaps initiated by endonuclease cleavage of the rearranged VL gene segments. T h e molecular details of such a gene conversion process have been reviewed by Szostak et al. (1983). An important distinction between these two molecular processes is that in double-homologous recombination, sequences originally derived from the rearranged VLgene segment would appear in the $V gene segments when these segments underwent diversification. In contrast, gene con-
SOMATIC DIVERSIFICATION O F
Igi
57
FIG. 8. Schematic representations of the positions and sizes of sequence substitutions in rearranged VL gene segments isolated from 18-day embryos. The positions of blocks of sequence substitutions in seven representative rearranged VL gene segments are depicted. Neither the size nor the position of the gene conversion events appears to be conserved. As few as 10 bp and as many as 240 bp can be involved in sequence diversification. Sequence diversification also appears to be able to occur sequentially during B cell development (e.g., second and sixth VL gene segments).
version will leave the pseudogene sequences unaltered by the process of diversification, and the germ-line VL sequences altered during diversification would be lost from the genome. In order to distinguish between these two molecular mechanisms, it was necessary to generate clonal B cell populations that were no longer undergoing gene conversion, so that
58
WAYNE T. M K O R M A C K A N D CRAIG B. T H O M P S O N
GENE CONVERSION
FIG. 9. Schematic representation of the potential mechanisms of IgL gene diversification.Two possible mechanisms for the substitution of sequences in a gene segment are depicted. (Top)Diversificationhas occurred as a result of gene conversion and leads to an additionalcopy of sequences derived from a 5' donor which are copied into an homologous position in the 3' recipient. (Bottom) The results of a double-homologous recombination event are depicted, in which homologous sequences are exchanged between the 5' and 3' homologous elements.
the organization of +V gene segments following gene conversion could be examined. Although mature B cell lines are relatively easy to generate in mammals, the generation of clonal mature B cell lines in the chicken has been difficult. Recently, Barth and Humphries (1988) have been able to generate a number of B cell lines from mature postbursal B cells by transformation with the v-rel oncogene. These cell lines do not display somatic diversification of their IgL V gene segments during either in uiuo or in vitro passage. Therefore, these cell lines are clonal expansions of a unique diversification event. Restriction enzyme sites within the +V and VLgene segments of these clonal cell lines can be used to detail the molecular consequences of the diversification of the rearranged IgL V gene segment. As demonstrated in Fig. 10, despite evidence of significant diversification on the rearranged VLgene segment by both sequence and restriction site analyses, we have been unable to detect any diversification of the
SOMATIC DIVERSIFICATION OF
V REGION PROBE
M
G
27ti130m~4i~i
Igl~
59
C REGION PROBE
M
27ti130s314i~i
G
FIG. 10. Three clonal mature B cell lines (27Li, 30S3, and 41Li) derived by transformation with a v-rebcontaining retrovirus are analyzed for the presence of KpnI sites within the JlV gene region. For comparison, germ-line tissue from chicken erythrocytes (G) was included. Southern blots of cell line DNA digested with KpnI were hybridized with V and C gene region probes. No alteration in the KpnI sites within the pseudogenes of this clonal population of B cells can be determined. Of particular interest, cell line 4 1Li contains two rearranged alleles, one of which has lost the KpnI site present at the end of the CDRl domain in the VL gene segment. Despite this, there is no appearance of a novel KpnI site in the JlV gene region cluster. This and additional data using six other restriction endonuclease digests known to cut within the pseudogene segments suggest that the molecular mechanism involved in IgL gene diversification is consistent with gene conversion rather than double-homologous recombination. Marker bands (M)are end-labeled Hind111 fragments of A DNA (23.1,9.4,6.7,4.4,2.3,2.0,and 0.56 kb).
60
WAYNE T. McCORMACK A N D CRAIG B. T H O M P S O N
JiV gene segments on either the rearranged or unrearranged alleles in the same cell. When a restriction site such as KpnI is lost from the rearranged VL gene segment, it appears to be lost from the genome rather than appearing in the pseudogene V gene segments. When a restriction site such as BarnHI is copied from a JiV segment into the rearranged VL gene segment, the JiV retains its original BamHI site. Therefore, in the cell that has undergone IgL gene diversification, the process most closely resembles gene conversion, a unidirectional transfer of sequence information from donor (JiV) to recipient (VL). Although it remains possible that diversification occurred through a tram-mediated double-homologous recombination following chromosome duplication in the S phase of the cell cycle, this appears unlikely to be the molecular mechanism. Based on this type of exchange, the chromosome receiving the VL gene segment would retain a functionally rearranged undiversified VL gene segment and would, therefore, be selectively amplified in the bursal environment. Over time, this cell would be subjected to further rounds of gene conversion. Therefore, one would expect to see the acquisition of modifications within the $V segments at positions distinct from those used to diversify the VL gene segment during development. N o evidence for such diversification by restriction enzyme analysis was observed in over 40 clonal v-re1 cell lines. Therefore, it appears that diversification of the IgL gene occurs as a result of replacement of the rearranged VL gene segment with copies of 10 to 120-bpsegments of pseudogenes. Additional molecular characteristics of this diversification process will undoubtedly be elucidated through further characterization of gene conversion events. Interestingly, avian leukosis virus-transformed B cell lines and tumors were found to undergo continued diversification of their rearranged VL gene segment during in uiuo or in uitro passage, suggesting that they represent transformed bursal stem cells (Thompson and Neiman, 1987). At least one avian leukosis virus-induced B cell line derived by Baba et al. (1985)constitutively undergoes gene conversion of its rearranged IgL gene segment. This cell line should provide an ideal model for in uitro study of the Ig gene conversion process. X. Somatic Diversification of the Ig VI Gene Repertoire by Gene Conversion Is a General Feature of Avian Ig Genes
McCormack et al. (1989a) have recently shown that rearrangement of a single functional VL gene segment and subsequent diversification by gene conversion appear to be general mechanisms for the generation of IgL diversity in a wide variety of avian species (Fig. 11). The chicken IgH
SOMATIC DIVERSIFICATION OF
V REGION PROBE
B G
M
IgL
61
C REGION PROBE
BG
M
FIG.11. Duck IgL gene rearrangement. DNAs isolated from the bursa (B) and erythrocytes (C) of a mallard duck are analyzed for IgL gene rearrangement. (Left) Southern blots cut with PstI are hybridized with the chicken V gene region probe. (Right) The same blots are hybridized with the chicken C gene region probe [marker (M) bands as in Fig. 101. Just as in chicken, the duck contains multiple V gene-homologous segments in the genome. However, based on the C gene region probe analysis, it appears that bursa1 lymphocytes undergo a single major gene rearrangement during B cell development in the bursa of Fabricius. This appears to be true of many avian species. Interestingly, however, some avian species have also developed some level of combinatorial diversity (for details, see McCormack et al., 1989a).
locus, like the IgL locus, appears to generate only very limited diversity by recombination (Weill et al., 1986; Parvari et al., 1988). In addition, Reynaud et al. (1989) have reported that gene conversion also appears to account for the immunoIogica1repertoire of the IgH gene. Based on their preliminary analysis of the chicken heavy-chain locus, Reynaud et al. have determined that there is a single functional V gene capable of rearranging with one of a cluster of D region segments and a single J region
62
WAYNE T. McCORMACK A N D CRAIG B . T H O M P S O N
segment. Further diversification apparently occurs as a result of gene conversion events utilizing pseudogenes located in the 60 kb upstream from the VH,gene segment. It appears that diversification by gene conversion occurred as an early evolutionary mechanism for the diversification of the Ig locus. It has previously been shown that DNA molecules microinjected into mammalian cell nuclei can undergo gene conversion (Rubnitz and Subramani, 1986). Therefore, the enzymes required to mediate gene conversion between homologous segments exists in a wide variety of cells. The possibility that at least some of the somatic diversification of mammalian Ig genes can be accounted for by gene conversion has been previously proposed (Maizels, 1989). In light of the new evidence generated by analysis of the chicken Ig diversification mechanism, it is likely that this mechanism of diversification will undergo further intensive study in mammalian Ig genes (Wysocki and Gefter, 1989). XI. Additional Mechanisms of lg Diversification
The sequences present in the JlV gene segments are not sufficient to account for all of the genetic variability observed in cDNAs derived from mature B cells. Eight of 214 modified nucleotides found in the 12 sequenced V-J gene segments analyzed by Reynaud et al. (1987) had no counterparts in the pseudogene pool. Such nucleotides appear to occur frequently at the apparent border of a gene conversion event. These data suggest the possibility that additional sequence diversity is generated by the gene conversion process through an error-prone mechanism that repairs the ends of the gene conversion events (Reynaud et al., 1987). In addition, sequence alterations in the J region segment, for which there were no apparent donor in the JlV segments, have been observed in cDNAs isolated from late developmental time points. This suggests that an additional type of somatic mutation may be occurring during the maturation of avian immune responses. Alternatively, these novel base pairs may be generated by a mismatch repair system used to resolve regions of heteroduplex DNA present at the end of gene conversion events, similar to the mutation model of Ripley (1982) involving quasipalindromic DNA sequences. Determining the etiology of these apparently untemplated nucleotides will undoubtedly contribute to our understanding of how somatic diversification of the genome occurs.
SOMATIC DIVERSIFICATION OF
I ~ L
63
XII. Summary
The bursa of Fabricius provides a unique organ for the study of lineage-specific development in a multicellular organism. Unlike mammalian B cells, B cells in the chicken develop in a single wave of differentiation, beginning with the commitment of progenitor cells to the B cell lineage between days 10 and 15 of embryogenesis. By day 18 of embryogenesis, all lymphoid progenitor cells capable of differentiation along the B cell lineage have migrated to the bursa of Fabricius. Following migration to the bursa, these lymphoid progenitors enter exponential growth and begin to populate each of the lo4 bursal follicles. Between day 18 of embryogenesis and 2-4 weeks of age, B cells undergo a stage of bursaldependent differentiation. By the end of this period, chickens are able to mount primary immune responses against virtually all antigens. In addition, by this time sufficient numbers of B cells have migrated from the bursa to peripheral lymphoid organs so that the B cell immune system can be maintained even if the bird is bursectomized. Bursectomy of chicks after 4 weeks of age has no long-term effects on the development and maintenance of the B cell immune system in adult birds. Because of the central nature of the surface Ig molecule to B cell development in mammals, the chicken IgL gene locus has been intensively studied during avian B cell development. The chicken IgLlocus is of particular interest because it has only one V region capable of rearrangement. Rearrangement of the IgL gene is not dependent on the bursal environment. B cell progenitors rearrange their IgL gene between days 10-15 of embryogenesis, prior to migration to the bursa. IgL gene rearrangement occurs by a deletional mechanism in which a precise joining of the IgL recombination signal sequences leads to a circular episomal element. During this deletion it appears that single nonrandom bases are added to both the V and J coding segments. Subsequent V-J joining occurs at random. Most progenitor B cells appear to rearrange only a single IgL allele. The high frequency of in-frame alleles observed in avian B cell lines appears to result from the selective amplification of cells with productive IgL rearrangements during bursal development between days 12 and 18 of embryogenesis. To create an immunological repertoire, chickens must diversify the coding sequence of this single functional V gene segment during development. This diversification occurs subsequent to IgL rearrangement during a strictly bursal-dependent phase of B cell development. IgL gene diversification is limited to the rearranged V gene segment and occurs by a gene conversion mechanism, using $V genes as sequence donors. Rear-
64
W A Y N E T. McCORMACK AND CRAIG B. THOMPSON
rangement of a single functional V gene and somatic gene conversion appear to account for the diversity observed in the IgH V regions as well. The molecular mechanisms of the gene conversion events that operate during chicken B cell development and the possible contributions of this process to mammalian Ig gene diversity are the subjects of ongoing investigations.
REFERENCES Alt, F. W., and Baltimore, D. (1982).Joining of immunoglobulin heavy chain gene segments: Implications from a chromosome with evidence of three D-JH fusions. Proc. Natl. Acad. Sci. U.S.A. 79,4118-4122. Alt, F. W.,Blackwell, T. K., DePinho, R. A., Reth, M. G., and Yancopoulos, G. D. (1986).Regulation of genome rearrangement events during lymphocyte differentiation. Immunol. Rev. 89,5-30. Baba, T . W., Giroir, B. P., and Humphries, E. H. (1985).Cell lines derived from avian lymphomas exhibit two distinct phenotypes. Virology 144, 139- 15 1. Barth, C. F., and Humphries, E. H. (1988).Expression of v-re1 induces mature B-cell lines that reflect the diversity of avian immunoglobulin heavy- and light-chain rearrangements. Mol. Cell. Biol. 8, 5358-5368. Cantor, H. (1984).T lymphocytes. I n “Fundamental Immunology” (W. E. Paul, ed.), pp. 57-69. Raven, New York. Cooper, M. D., Cain, W. A., Van Alten, P. J., and Good, R. A. (1969).Development and function of the immunoglobulin producing system. I. Effect of bursectomy at different stages of development on germinal centers, plasma cells, immunoglobulins, and antibody production. Znt. Arch. Allergy Appl. Immunol. 35,242-252. Cooper, M. D., Kearney, J., and Scher, I. (1984).B lymphocytes. In “Fundamental Immunology” (W. E. Paul, ed.), pp. 43-55. Raven, New York. Corbel, C., Belo, M.,Martin, C., and Le Douarin, N. M. (1 987).A novel method to bursectomize avian embryos and obtain quail -P chick .bursa1 chimeras. 11. Immune response of bursectomized chicks and chimeras and post-natal rejection of the grafted quail bursas. J . Immunol. 138,2813-282 1. DuPasquier, L. (1982).Antibody diversity in lower vertebrates. Why is it so restricted? Nature (London) 296, 31 1-313. Eerola, E.,Jalkanen, S., Granfors, K., and Toivanen, A. (1983).Immune capacity of the chicken bursectomized at 60 hours of incubation: Mitogen induced cell proliferation and immunoglobulin secretion. J. Immunol. 131, 120- 124. Eerola, E., Jalkanen, S., Granfors, K., and Toivanen, A. (1984).Immune capacity of the chicken bursectomized at 60H of incubation. S c a d . J . Immunol. 19, 493-500. Fujimoto, S., and Yamagishi, H. (1987).Isolation of an excision product of T-cell receptor a-chain rearrangements. Nature (London) 327,242-243. Glick, B. (1977).The bursa of Fabricius and immunoglobulin synthesis. Int. Rev. Cytol. 48, 345-402. Glick, B., Chang, T. S., and Jaap, R. G. (1956).The bursa of Fabricius and antibody production on the domestic fowl. Poult. Sci. 35, 224. Granfors, K.,Martin, C., Lassila, O., Suvitaival, R., Toivanen, A., and Toivanen, P. (1982).Immune capacity of the chicken bursectomized at 60 hours of
SOMATIC DIVEKSIFICATION OF
I@.
65
incubation: Production of the immunoglobulins and specific antibodies. Clin. Immunol. Immunopathol. 23,459-469. Grant, J. A., Sanders, B., and Hood, L. (197 1). Partial amino acid sequences of chicken and turkey immunoglobulin light chain. Homology with mammalian A chains. Biochemistry 10,3123-3132. Grossi, C. E., Lydyard, P. M., and Cooper, M. D. (1976). B-cell ontogeny in the chicken. Ann. Immunol. (Paris) 127C, 93 1-94 1 . Hood, L., Grant, J. A., and Sox,H. C., Jr. (1970). On thestructure ofnormal light chains from mammals and birds: Evolutionary and genetic implications. In “Developmental Aspects of Antibody Formation and Structure” (J. Sterz and I. Riha, eds.), pp. 283-309. Vol. 1, Academic Publ. House, Prague. Hope, T.J., Aguilera, R.J., Minie, M.E., and Sakano, H. (1986). Endonucleolytic activity that cleaves immunoglobulin recombination sequences. Science 231, 1141-1 145. Houssaint, E., Belo, M., and Le Douarin, N. M. (1976). Investigations on cell lineage and tissue interactions in the developing bursa of Fabricius through interspecific chimeras. Dev. Biol. 53,250-264. Houssaint, E., Torano, A., and Ivanyi, J. (1983). Ontogenic restriction of colonization of the bursa of Fabricius. Eur. J . Immunol. 13,590-595. Huang, H. V., and Dreyer, W. J. (1978).Bursectomy in ovo blocks the generation of immune diversity.j. Immunol. 121, 1738-1747. Hunkapiller, T., and Hood, L. (1989). Diversity of the immunoglobulin gene superfamily. A d a Immunol. 44, 1-63. Jalkanen, S., Granfors, K., Jalkanen, M., and Toivanen, P. (1983a). Immune capacity of the chicken bursectomized at 60 hr of incubation: Surface immunoglobulin and B-L (Ia-like) antigen-bearing cells. J . Immunol. 130, 2038204 1 . Jalkanen, S., Granfors, K., Jalkanen, M.,and Toivanen, P. (1983b). Immune capacity of the chicken bursectomized at 60 hr of incubation: Failure to produce immune, natural, and autoantibodies in spite of immunoglobulin production. Cell. Immunol. 80,363-373. Jalkanen, S., Jalkanen, M., Granfors, K., and Toivanen, P. (1984). Defect in the generation of the light-chain diversity in bursectomized birds. Nature (Loadon) 311,69-71. Le Douarin, N. M., Houssaint, E., Jotereau, F. V., and Belo, M. (1975). Origin of hemopoietic stem cells in embryonic bursa of Fabricius and bone marrow studied through interspecific chimeras. Proc. Natl. Acad. Sci. U.S.A. 72,27012705. Lerman, S. P., and Weidanz, W. P. (1970).The effect of cyclophosphamide on the ontogeny of the humoral immune response in chickens. J . Immunol. 105, 6 14-6 19. Lydyard, P. M., Grossi, C. E., and Cooper, M. D. (1976). Ontogeny of B cells in the chicken. I. Sequential development of clonal diversity in the bursa. J.Ex$. Med. 144,79-97. Maizels, N. (1989). Might gene conversion be the mechanism of somatic hypermutation of mammalian immunoglobulin genes? Trends Genet. 5,4-8. McCormack, W. T., Carlson, L. M., Tjoelker, L. W., and Thompson, C. B. (1989a). Evolutionary comparison of the avian IgL locus: Combinatorial diversity plays a role in the generation of the antibody repertoire in some avian species. Int. Immunol. 1, 332-341. McCormack, W. T., Tjoelker, L. W., Barth, C. F., Carlson, L. M., Petryniak, B.,
66
WAYNE T. McCORMACK A N D CRAIG B. T H O M P S O N
Humphries, E. H., and Thompson, C. B. (1989b). Selection for B cells with productive IgL gene rearrangements occurs in the bursa of Fabricius during chicken embryonic development. Genes Dev.3,838-847. McCormack, W. T., Tjoelker, L. W., Carlson, L. M., Petryniak, B., Barth, C. F., Humphries, E. H., and Thompson, C. B. (1989~).Chicken IgL gene rearrangement involves deletion of a circular episome and addition of single nonrandom nucleotides to both coding segments. Cell (Cambridge, Mass.) 56, 785-791. Milner, E. C. B., Meek, K. D., Rathbun, G., Tucker, P., and Capra, J. D. (1986). Are anti-arsonate antibody N-segments selected at both the protein and the DNA level? Immunol. Today 7,36-40. Moore, M. A. S., and Owen, J. J. T. (1966). Experimental studies on the development of the bursa of Fabricius. Dev.Biol. 14,40-5 1. Okazaki, K., Davis, D. D., and Sakano, H. (1987).T cell receptor /3 gene sequences in the circular DNA of thymocyte nuclei: Direct evidence for intramolecular DNA deletion in V-D-J joining. Cell (Cambridge, Mass.) 49,477-485. Olah, I., and Click, B. (1978). T h e number and size of the follicular epithelium and follicles in the bursa of Fabricius. Poult. Sci. 57, 1445-1450. Parslow, T., Blair, D., Murphy, W., and Granner, D. (1984). Structure of the 5’ ends of immunoglobulin genes: A novel conserved sequence. Proc. Natl. Acad. Sci. U.S.A. 81,2650-2654. Parvari, R., Ziv, E., Lentner, F., Tel-Or, S., Burstein, Y.,and Schechter, I. (1987). Analyses of chicken immunoglobulin light chain cDNA clones indicate a few germline VA genes and allotypes of the CA locus. EMBO J. 6,97-102. Parvari, R., Avivi, A., Lentner, F., Ziv, E., Tel-Or, S., Burstein, Y., and Schechter, I. (1988).Chicken immunoglobulin y-heavy chains: Limited VH gene repertoire, cornbinatorial diversification by D gene segments and evolution of the heavy chain locus. EMBO J. 7,739-744. Pink, J. R. L. (1986). Counting components of the chicken’s B cell system. Immunol. Rev. 91, 115-128. Pink, J. R. L., Koch, C., and Ziegler, A. (1981). Immuno-ornithological conversation. In “The Immune System” (C. M. Steinberg and I. Lefkovits, eds.), Vol. 1 , pp. 69-75. Karger, Basel. Pink, J. R. L., Vainio, O., and Rijnbeek, A.-M. (1985a). Clones of B lymphocytes in individual follicles of the bursa of Fabricius. Eur. J. Immunol. 15, 83-87. Pink, J. R. L., Ratcliffe, M. J. H., and Vainio, 0. (1985b). Immunoglobulinbearing stem cells for clones of B (bursa-derived) lymphocytes. Eur. J. Immunol. 15,617-620. Pink, J. R. L., Lassila, O., and Vainio, 0. (1987). B-lymphocytes and their selfrenewal. In “Avian Immunology: Basis and Practice” (A. Toivanen and Toivanen, P., eds.), Vol. 1, pp. 65-78. CRC Press, Boca Raton, Florida. Ratcliffe, M. J. H. (1985). T h e ontogeny and cloning of B cells in the bursa of Fabricius. Immunol. Today 6,223-227. Reynaud, C.-A., Dahan, A., and Weill, J.-C. (1983). Complete sequence of a chicken A light chain immunoglobulin derived from the nucleotide sequence of its mRNA. Proc. Natl. Acad. Sci. U.S.A. 80,4099-4103. Reynaud, C.-A., Anquez, V., Dahan, A., and Weill, J.-C. (1985). A single rearrangement event generates most of the chicken immunoglobulin light chain diversity. Cell (Cambridge, Mass.) 40, 283-29 1. Reynaud, C.-A., Anquez, V., Grimal, H., and Weill, J.-C. (1987). A hypercon-
SOMATIC DIVERSIFICATION O F
Igl.
67
version mechanism generates the chicken light chain preimmune repertoire. Cell (Cambridge, Mass.) 48,379-388. Reynaud, C.-A., Dahan, A., Anquez, V., and Weill, J.-C. (1989). Development of the chicken antibody repertoire. I n “Immunoglobulin Genes” (T. Honjo, F. W. Alt, and T. H. Rabbitts, eds.), pp. 151-162. Academic Press, San Diego, California. Reynolds, J. D. (1987). Mitotic rate maturation in the Peyer’s patches of fetal sheep and in the bursa of Fabricius of the chick embryo. Eur. J. Immunol. 17, 503-507. Ripley, L. S. (1982). Model for the participation of quasi-palindromic DNA sequences in frameshift mutation. Proc. Null. Acad. Sci. U.S.A. 79, 41284132. Rubnitz, J., and Subramani, S. (1986).Extrachromosomal and intrachromosomal gene conversion in mammalian cells. Mol. Cell. Biol. 6, 1608-1614. Saiki, R. K . , Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., Mullis, K. B., and Erlich, H. A. (1988). Primer-directed enzymatic amplication of DNA with a thermostable DNA polymerase. Science 239,487-49 1. Szostak, J. W., Orr-Weaver, J. L., and Rothstein, R. J. (1983).T h e double-strandbreak repair model for recombination. Cell (Cambridge, Mass.) 33,26-35. Thompson, C. B., and Neiman, P. E. (1987). Somatic diversification of the chicken immunoglobulin light chain gene is limited to the rearranged variable gene segment. Cell (Cambridge, Mass.) 48,369-378. Thompson, C. B., Humphries, E. H., Carlson, L. M., Chen, C.-L. H., and Neiman, P. E. (1987). T h e effect of alterations in myc gene expression on B cell development in the bursa of Fabricius. Cell (Cambridge, Mass.) 51,371-381. Toivanen, P., and Toivanen, A. (1973). Bursa1 and postbursal stem cells in chicken. Functional characteristics. Eur. J . Immunol. 3,585-595. Toivanen, P., Naukkarinen, A., and Vainio, 0. (1987). What is the function of bursa of Fabricius? I n “Avian Immunology: Basis and Practice” (A. Toivanen and P. Toivanen, eds.), Vol. 1, pp. 79-99. CRC Press, Boca Raton, Florida. Tonegawa, S. ( 1983). Somatic generation of antibody diversity. Nature (London) 302,575-58 1 . Weill, J.-C., Reynaud, C.-A., Lassila, 0..and Pink, J. R. L. (1986).Rearrangement of chicken immunoglobulin genes is not an ongoing process in the embryonic bursa of Fabricius. Proc. Nad. Acad. Sci. U.S.A. 83,3336-3340. Weill, J.-C., Leibowitch, M., and Reynaud, C.-A. (1987). Questioning the role of the embryonic bursa in the molecular differentiation of B lymphocytes. Curr. Top. Microbiol. Immunol. 135, 1 11-124. Wysocki, L. J., and Gefter, M. L. (1989). Gene conversion and the generation of antibody diversity. Annu. Rev. Biochem. 58,509-53 1 . This article was accepted for publication on 20 October 1989.
ADVANCES I N IMMUNOLOGY, VOL. 48
T lymphocyte-Derived Colony-Stimulating Factors ANNE KELSO AND DONALD METCALF The Walter and Eliza Hall Institute of Medical Reseamh, Victoria 3050, Austmlia
1. Introduction
The study of the control of hemopoiesis by colony-stimulating factors (CSFs) has been firmly based in physiology from its inception in the 1960s to its recent application to the therapy of hemopoietic disorders. This is due in large part to the development of clonal culture methods that allow the lineage commitment, proliferation, and terminal maturation of normal hemopoietic progenitor cells to be monitored in vitro. By contrast, parallel work on the production of CSFs by T lymphocytes has relied on immortalized or transformed producer cell lines, an approach which has yielded many insights into the cellular and molecular control of CSF synthesis, but isolates the cell from its normal context. Recent advances now suggest that a reconstruction of the physiology of CSF production by T cells may soon be possible. It therefore seems timely to attempt to draw together the large body of work on the synthesis of CSFs and other regulators by T cells in uitro, and our more fragmentary understanding of the in vivo correlates of this response. II. Colony-Stimulating Factors
It is generally accepted that functional interactions between lymphoid and granulocyte-macrophage populations constitute a crucial component in host resistance to microbial infections. Major defects in either cell population, such as T cell depletion in severe combined immunodeficiency or granulocyte depletion in congenital neutropenia, result in a profound susceptibility to recurrent major infections. In the past the interacting roles envisaged for these populations in responses to infections included ( 1) the processing and presentation by macrophages of antigens to responding lymphoid cells, (2) direct T lymphocyte cytotoxic effects, particularly on virus-infected cells, (3) production of specific antibodies by B lymphocytes, and (4)granulocyte- and macrophage-mediated phagocytosis and destruction of microorganisms, processes enhanced by specific antibodies. These various events have 69 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
70
A N N E KELSO A N D DONALD METCALF
been the subject of extensive investigation and are firmly established as major components in host resistance. Possibly less well emphasized is the likely sequential nature of these host responses. The granulocyte-macrophage responses, together with those of other components of the reticuloendothelial system, are the most rapidly activated and are therefore the most likely to mediate early host responses to infection. When these responses fail to eliminate the infection, the more slowly developing T and B lymphocyte responses become of progressively greater importance with direct cell-mediated cytolysis, cytotoxic antibody production, and antibody-mediated enhancement of granulocyte-macrophage responses. Lymphocyte responses are ultimately responsible for the immunized state that will ensure accelerated and more efficient responses to future infections by the same organism. More recently, it has been recognized that activated lymphocytes, particularly T lymphocytes, can play a quite different role in resistance through their ability to produce mediator molecules, or “cytokines,” such as interferon-y (IFN-y), interleukins (ILs), tumor necrosis factors (TNFs), and the CSFs (Table I). These agents can enhance the functional responses of other cells involved in resistance and, because their production following lymphocyte activation is rapid, lymphocyte responses of this type may in fact be involved earlier in host resistance than was previously thought. Recent studies have not only characterized these mediator molecules, but have led to the cloning of cDNAs for each and the mass production of recombinant molecules and initial clinical trials on some. Much of the discussion to follow is concerned with the specific glycoproteins regulating granulocyte-macrophage production and function, the CSFs, and in particular assesses the importance and implications of CSF production by T lymphocytes in the context of the known biology of these factors. A. DISCOVERY AND CHARACTERIZATION OF THE CSFs The CSFs were discovered following the development of semisolid cultures able to support the proliferation of committed progenitor cells to form colonies of maturing granulocytes and/or macrophages (Bradley and Metcalf, 1966; Ichikawa et al., 1966). It was evident that these cells, like all other hernopoietic cells, could not proliferate spontaneously and that proliferation required stimulation by the addition to the cultures of other cells, cell extracts, or medium conditioned by such cells. It was assumed initially that, as was the case with erythropoiesis, a single active factor might stimulate this proliferation, and the operational name “colony-stimulating factor” was applied to the unknown factor. A quanti-
T LYMPHOCYTE-DERIVED CSFS
71
TABLE I HEMOPOIETIC REGULATORS AND ALLIED CYTOKINES KNOWNTO BE PRODUCED BY ACTIVATED LYMPHOCYTES
Regulator Erythropoietiri
Responding cells'
E
M, G, Eo, Meg, E, endothelial G-CSF G, M?, endothelial M-CSF M, G, placenta Multi-CSF (IL-3) G, M, Eo, Meg, mast, E, stem IL- 1 Stem, endothelial, strornal, fibroblast,T, B IL-2 T, B, M IL-4 T, B, G, M, mast IL-5 Eo, T, B IL-6 B, T, G, stem, various IL-7 T, pro-B, pre-B LIF M, various IFN-.)I Strornal, fibroblast, endothelial, various GM-CSF
Lymphocyte production demonstrated NO
Yes No Nob
Yes
Yes
Yes Yes Yes
Yes No
Yes
Yes
a E, Erythroid; G, neutrophilic granulocytic; M, macrophage; Meg, megakaryocyte; Eo, eosinophil; mast, mast cell; stem, multipotential hemopoietic stem cell; T, T lymphocyte; B, B lymphocyte. M-CSF mRNA is constitutively expressed in some T cell clones (N. M. Cough, personal communication).
tative relationship in the cultures between the amount of CSF added and the number of colonies developing allowed the cultures to be used as a bioassay system, not only for detecting CSF, but also for monitoring its purification (Metcalf, 1984). Original purification efforts were made using CSF detectable in human urine (Stanley et al., 1975) and in L cell-conditioned medium (Stanley and Heard, 1977), the first CSF to be purified to homogeneity. This form proved to be a relatively selective stimulus for macrophage colony formation and is now known as M-CSF (or CSF-l). A second form of CSF of smaller size (molecular weight 23,000) was demonstrated in medium conditioned by mouse lung tissue (Sheridan and Metcalf, 1973), and this form stimulated granulocyte-macrophage colony formation. When this CSF was finally purified from mouse lung conditioned medium, it was given the name GM-CSF (Burgess et al., 1977). Parallel studies showed that an unusual form of CSF was present in the serum of endotoxin-injected mice and was characterized by an ability only to stimulate the forma-
72
ANNE KELSO AND DONALD METCALF
tion of small numbers of highly mature granulocytic colonies. When purified to homogeneity from mouse lung conditioned medium, this CSF (molecularweight 25,000) was given the name G-CSF (Nicola etal., 1983). Another type of CSF-indeed, what was initially believed possibly to represent a mixture of different CSFs-was detected in medium conditioned by the myelomonocytic leukemia WEHI-3B and medium from cultures of spleen cells stimulated by mitogens, particularly pokeweed mitogen. This material was able to stimulate not only granulocytemacrophage colony formation, but also the formation of eosinophil, megakaryocyte, erythroid, mast cell, and multipotential colonies. When purified, these multiple activities were found to be due to a single molecule (molecular weight 21,000-28,000), and in this laboratory it was given the name multi-CSF (Cutler et al., 1985). In a parallel series of studies, a factor able to elicit the expression of 20-a-hydroxysteroid dehydrogenase in T lymphocytes and to stimulate the proliferation of several continuous hemopoietic cell lines was purified under the name IL-3 (Ihle et al., 1982). Multi-CSF and IL-3 are identical. At this stage it is necessary to introduce three other hemopoietic regulators with proliferative actions on either granulocyte-macrophages or eosinophils, since the first two of these are discussed later as products of T cells. IL-4 was cloned as a factor with proliferative effects on mast cells and T cells and the ability to enhance immunoglobulin GI (IgGl), IgE, and la expression by B cells (BSF-1) (Lee et al., 1986; Noma et al., 1986). It was shown subsequently to enhance granulocyte-macrophage colony formation stimulated by multi-CSF or G-CSF (Rennick et al., 1987). IL-5 was originally discovered as a CSF for eosinophils (Metcalf et al., 1983),then purified as a factor stimulating the production and functional activity of eosinophils (eosinophil differentiation factor) (Sanderson et al., 1985b) and cloned as a T cell-replacing factor or B cell growth factor type I1 (Kinashi et al., 1986). The regulator IL-6 was discovered successively as a plasmacytoma growth factor (Metcalf, 1973; Nordan et al., 1987), an IFN-like molecule (Zilberstein et al., 1986), a lymphocyte differentiation-inducing factor (Hirano et al., 1986), a hepatocytestimulating factor (Gauldie et al., 1987),and an inducer of differentiation in myeloid leukemic cells (Shabo et al., 1988).IL-6 has proved to function as a CSF, at least for murine granulocytic cells (Suda et al., 1988; Metcalf, 1989a).
B. BIOCHEMICAL NATUREOF THE CSFs AND THEIR RECEPTORS All four CSFs are glycoproteins; three are composed of a single polypeptide chain, while the fourth (M-CSF) is a dimer of two apparently identical polypeptide chains linked by disulfide bridges (Table 11). The
T LYMPHOCYTE-DERIVED CSFS
73
polypeptide chain lengths are relatively uniform in size (molecular weight 18,000-22,000) and are held in an active three-dimensional configuration by mandatory disulfide bridges (Nicola, 1989). The carbohydrate moiety of the CSFs is relatively large and can vary according to the cellular source of the CSF. It is not involved in receptor binding, is not required for the biological activity of the molecules, and may serve to prolong the half-life of the molecules in viva Each CSF is encoded by a single unique gene. Surprisingly, in view of the similar actions of the molecules, no sequence homology exists between the four CSFs, and they differ in predicted secondary structure. Despite the unrelatedness of the CSFs at the mRNA and protein levels, some intriguing features of the location and structure of certain CSF genes suggest a possible common ancestral origin of these and allied hemopoietic growth factor genes. Thus, in humans the genes for GMCSF, multi-CSF, M-CSF, the M-CSF receptor (c-fms), IL-4, and IL-5 (Yang et al., 1988; Le Beau et al., 1986; van Leeuwen et al., 1989) are clustered on the long arm of chromosome 5. In the mouse also, the genes for GM-CSF, multi-CSF, IL-4, and IL-5, but not M-CSF, are clustered on chromosome 11 (Barlow et al., 1987; Lee and Young, 1989; Lee et al., 1989). Unique high-affinity membrane receptors exist for each CSF (Nicola, 1989). Only two receptors have so far been cloned and characterized. The M-CSF receptor is the c-fms protooncogene product and is a transmembrane glycoprotein of molecular weight 160,000 with an intracytoplasmic tyrosine kinase domain that becomes autophosphorylated following binding of M-CSF (Sherr et d., 1985). The other receptor now identified by cDNA cloning is the human GM-CSF receptor, a heavily glycosylated transmembrane protein of molecular weight 85,000 (Gear-
TABLE I1 MURINECSFs Factor GM-CSF G-CSF M-CSF (CSF-1) Multi-CSF (IL-3)
Molecular weight of peptide" 14,400 19,100 18,000-2 1,000* 16,200
Chromosomal location of gene 11A5-B1 11D-El 3F3 1 1A5-Bl
a In their native form all are glycosylated molecules, their carbohydratecontent varying according to the cellular source. M-CSF is a dimer; the molecular weight of the monomer is shown.
'
74
ANNE KELSO AND DONALD METCALF
ing et al., 1989). The cytoplasmic region of this receptor lacks a tyrosine kinase domain and must therefore transduce a signal by a different mechanism. There is some evidence implicating GTP-binding proteins and protein kinase C in such signaling (Farrar et al., 1985; GomezCambronero et al., 1989). Biological evidence indicates that signaling from CSF-receptor complexes must involve multiple cascades, some of which are unique for a given complex, while others may be shared (Metcalf, 1989b; Nicola et al., 1988). In hemopoietic populations the distribution of CSF receptors is restricted to cells of lineages shown to exhibit biological responses following stimulation, receptors being present both on immature proliferating cells and on mature postmitotic cells. Typically, each responding cells displays only a few hundred receptors, and biological responses can be elicited with very low levels of receptor occupancy (e.g., around 10%) (Nicola, 1989). Most granulocyte-macrophage cells simultaneously express receptors for more than one CSF, and occupancy by a CSF of its receptors can lead to “tram-down-modulation” of other types of CSF receptors (Walker et al., 1985). This arrangement of CSF receptors allows more than one CSF to act simultaneously on a responding cell, with the possibility not only of additive or superadditive responses, but also of competitive actions between the CSFs. ACTIONS OF THE CSFs C. BIOLOGICAL The outstanding characteristic of the CSFs is their ability to elicit multiple types of effects in responding cells, the exact nature of the response being dictated by the cells themselves. These effects can be grouped into five categories. 1. Proliferative Stimulation Each cell division in immature granulocytes and macrophages depends on stimulation by an adequate concentration of CSF (Metcalf, 1984). Stimulation seems to be required throughout most of the cell cycle in order for the cycle to be completed; even though CSF-receptor complexes have a relatively long intracellular half-life, CSF withdrawal usually prevents further progression through the cycle (Metcalf and Merchav, 1982; Metcalf, 1985).The quantitative responsiveness of individual cells varies, some cells requiring ten- to 100-fold higher concentrations of CSF to proliferate, but in general the CSFs elicit proliferative responses in the lo-’’ to lo-’* M concentration range. The concentration of CSF determines the mean cell cycle time and the total number of progeny generated during a given interval (Metcalf, 1980). The proliferative action of the CSFs exhibits some target cell selectivity,
T LYMPHOCYTE-DERIVED CSFS
75
but there is considerable overlap in the actions of the different CSFs (Metcalf, 1987). G-CSF is a relatively selective stimulus for granulocyte formation and M-CSF is a correspondingly selective stimulus for macrophage formation, while GM-CSF and multi-CSF are able to stimulate the proliferation of both cell types (Table I). T h e range of responding cells tends to broaden with increasing CSF concentration. This is best seen with GM-CSF, which at low concentrations is a selective stimulus for macrophage proliferation and at higher concentrations also stimulates granulocyte proliferation. As concentrations are progressively increased, GM-CSF stimulates eosinophils, then megakaryocytes, and finally the proliferation of some erythroid and multipotential precursors (Metcalf et al., 198613). As shown in Table I, at least three CSFs can stimulate the formation of granulocytes or macrophages. IL-6 is also a proliferative stimulus for granulocytes (Suda et al., 1988; Metcalf, 1989a), while IL-4 has weaker actions on both granulocyte and macrophage populations. Similarly, eosinophil proliferation can be stimulated by at least three factors: IL-5, multi-CSF, and GM-CSF. 2 . Cell Survival
At least in vitro, the survival of granulocyte-macrophage precursors is absolutely dependent on CSF (Metcalf and Merchav, 1982), and even with mature polymorphs and eosinophils, survival in vitro is extended in the presence of CSF (Begley et al., 1986). These actions may be based on an ability of CSFs to maintain intracellular ATP levels and membrane transport integrity (Whetton and Dexter, 1983; Hamilton et al., 1988; Vairo and Hamilton, 1988). 3. Differentiation Commitment
When acting on bipotential granulocyte-macrophage progenitors, the CSFs can dictate irreversibly the maturation lineage of the progeny cells. For example, M-CSF stimulation can result in irreversible commitment to the formation of macrophage progeny, while high concentrations of GM-CSF can result in a similar commitment to granulocyte formation (Metcalf and Burgess, 1982). 4 . Maturation Induction
Th e CSFs can at least initiate the maturation processes by which, for example, myeloblasts change progressively to the mature neutrophils that appear in the circulation (Valtieri et al., 1987). It is improbable that the CSFs directly regulate all of the transcriptional events involved in such maturation sequences, which may well be self-sustaining once a sequence is initiated.
76
ANNE KELSO AND DONALD METCALF
5. Functional Activation An important aspect of the biology of the CSFs is their ability to stimulate the functional activity of the mature cells in appropriate lineages. A number of the activities stimulated by the CSFs are of direct relevance to resistance to infections. These include chemotaxis (Wang et al., 1987, 1989), increased phagocytosis and cytotoxicity (Lopez et al., 1983, 1986; Fleischmann et al., 1986; Villalta and Kierszenbaum, 1986), increased superoxide production (Weisbart et al., 1987; Sullivan et al., 1987),increased Ia expression and antigen-presenting function (Fischer et al., 1988), and the production by monocytes of molecules involved in inflammatory responses, such as prostaglandin E (Kurland et al., 1979), plasminogen activator (Lin and Gordon, 1979),other CSFs (Metcalf and Nicola, 1985; Lindemann et al., 1989), IL-1, IFN, and TNF (Moore et al., 1980; Warren and Ralph, 1986; Cannistra et al., 1987).Enhanced tumoricidal activity and killing of intracellular parasites by CSF-treated macrophages have also been reported (Grabstein et al., 1986; Handman and Burgess, 1979; Weiser et al., 1987; Reed et al., 1987). Increased functional activity is a rapid response following CSF stimulation, is often achieved by concentrations lower than those required for proliferative stimulation, and usually declines promptly following removal of the CSF. The CSFs, therefore, have two important actions in host resistance. They elicit prompt rises in the functional activity of existing mature granulocytesand macrophages, and they amplify such responses by stimulating increased production of additional granulocytes and macrophages, allowing progressive escalation of responses during sustained infections. D. CELLULAR SOURCES OF THE CSFs The CSFs are produced by many cell types and at least some can simultaneously produce more than one type of CSF. Apart from lymphocytes, known CSF-producing cells include endothelial cells, fibroblasts, and macrophages-cells widely distributed throughout the body and likely to make early contact with invading microorganisms. Possibly because these cells are ubiquitous, CSF is extractable from all tissues in concentrations higher than in the serum, and such tissues actively produce CSF in organ culture. It is likely that CSFs can be produced by most cell types when appropriately stimulated. The undoubted ability of T lymphocytes to produce CSFs and other regulators, at least in vitro, and the enthusiasm aroused by advances in this area, have led to some overinterpretation of the role played by
?’ LYMPHOCYTE-DERIVED CSFS
77
lymphocytes in their biology. This applies particularly to the notion that these agents are exclusively of lymphocyte origin, as could be inferred from the term “Iymphokines.” While T cells may be the major source of multi-CSF and IL-5, lymphocytes are certainly not the only source of other members of this class of regulators or necessarily even an important source in many situations (Metcalf, 1984, 1988). For convenience in the following discussion, the term “lymphokine” is used when referring to products of T lymphocytes, but the limited context in which the term is being used should be kept in mind. 111. T lymphocyte Production of CSFs in Witnr
T h e history of the identification of T lymphocytes as CSF-producing cells has paralleled developments in lymphocyte culture and cloning methods and in the purification and molecular cloning of the CSFs themselves. Soon after the first reports that lymphocytes could release soluble regulatory factors in vitro, McNeill(l973) and Parker and Metcalf (1974a,b) found that antigen- o r mitogen-stimulated spleen cell supernatants contained bone marrow colony-stimulating activity. Parker and Metcalf further showed that production of this activity depended on the presence of T cells. Direct evidence that at least some T cells could synthesize CSFs came from the finding that certain murine T cell lymphomas and hybridomas released colony-stimulating activity, constitutively in some cases, but more commonly following stimulation with lectins [concanavalin A (Con A) o r phytohemagglutinin] or phorbol myristate acetate (PMA) (Ralph et al., 1978; Howard et al., 1979; Schrader et al., 1980). The production of CSFs by lymphomas and hybridomas could have been due to aberrant gene expression, as was later shown for some non-T cell tumors, such as WEHI-3B (Ymer et al., 1985). However, with the discovery of the T cell growth factor IL-2, it became possible to generate long-term lines and clones which retained many of the functional properties of normal activated T lymphocytes and which could be stimulated by antigen or mitogens to secrete CSFs (Schreier and Iscove, 1980; Nabel et al., 1981a,b; Elyetal., 1981; Staberetal., 1982; Kelsoetal., 1982; Griffinet al., 1984). Experiments in which filler cell-free and single-cell cultures of such clones produced CSFs demonstrated unequivocally that T cells could synthesize these factors (Kelso and Metcalf, 1985a; Kelso, 1986). Because of their very high production levels and, in some cases, restricted cytokine content, various polyclonal and monoclonal T cell populations were used as sources of protein for purification and of mRNA for cDNA cloning of GM-CSF, multi-CSF, IL-5, and other lymphokines
78
ANNE KELSO AND DONALD METCALF
(Burgesset al., 1980; Schrader and Clark-Lewis, 1982; Cutler et al., 1985; Metcalf et al., 1983; Sanderson et al., 1986; Gough et al., 1984; Yokota et al., 1984; Kinashi et al., 1986). Several groups screened large panels of T cell clones and hybrids looking for associations or dissociations in the synthesis of different factors. Some of the first evidence for the common identity of factors active in different assays was obtained in this way [e.g., macrophageactivating factor and IFN-y, eosinophil differentiation factor and B cell growth factor type I1 (IL-5)] (Zlotnik et al., 1983; Kelso and Glasebrook, 1984; Sanderson et al., 1986).Like transformed T cell lines, clones varied markedly in both the amounts and types of CSFs and other lymphokines they produced. Retesting of clones and analysis of subclones showed that the levels and proportions of different CSFs produced were reproducible and heritable (Kelso and Gough, 1987), suggesting that T lymphocyte maturation results in the acquisition of stable clonotypic patterns of lymphokine production. Clonal associationsand kinetic similarities in the production of distinct lymphokines following stimulation led to the idea that T cells synthesize different factors in a coordinate manner, and hence that the corresponding genes may be activated by a common mechanism (Watson, 1983; Kelso and Metcalf, 1985b; Sanderson et al., 1985a; Wiskocil et al., 1985). These early studies established three key points: CSF release depends on triggering the T cell with antigen or an equivalent stimulus, T cells can coordinately secrete several distinct CSFs, and individual clones vary in the types and quantities of CSFs they synthesize. Recent developments in these areas are discussed below. IV. Inducibility and Regulation of CSF Synthesis
A. INDUCTIONOF CSF SYNTHESIS Small resting T cells do not synthesize significant quantities of CSFs or other lymphokines. Primary T cell activation in uitro depends on the cross-linking of T cell antigen receptors (TCRs) by antigen complexed to major histocompatibility complex (MHC) molecules, by lectins that bind the TCR (e.g., Con A or phytohemagglutinin), or by solid phaseimmobilized anti-TCR antibodies, together with the delivery of “second signals” such as IL-1 or IL-2. The resultant activation of T cell populations to express IL-2 receptors and to proliferate in response to IL-2 or other growth factors is accompanied by the secretion of various lymphokines, including CSFs. As yet, little is known about the biochemical or genetic events that mediate transition of the naive T cell to an activated lymphokine-secreting state.
T LYMPHOCYTE-DERIVED CSFS
79
By comparison, the control of lymphokine production in T cell clones and tumors is less complex and better understood. Constitutive lymphokine production by such cell lines is usually low, whether or not their proliferation depends on growth factors such as IL-2 and IL-4. However, an often dramatic induction of lymphokine synthesis is seen when the cells are exposed to TCR ligands; unlike naive T cells, clones and lymphomas generally d o not require ancillary signals to retrigger lymphokine synthesis. T h e response is rapid and short-lived. For example, in some clones an increase in the rate of CM-CSF or multi-CSF secretion is detectable within 30 minutes of Con A or anti-TCR antibody addition; for high-producer clones CSFs continue to accumulate in the culture supernatant for 12-24 hours, while in low-producer clones both the rate and the duration of secretion are lower (Ely et al., 1981; Kelso et al., 1984; Gough and Kelso, 1989). Increases in the secretion of CSFs and other lymphokines following TCR activation are associated with rises in steady-state levels of the corresponding mRNAs, which then decline to low or undetectable levels after about 24 hours. The strong correlation that we and others have noted between mRNA and secreted protein levels indicates that the regulation of lymphokine synthesis is mainly pretranslational (Herold et al., 1986; Gough and Kelso, 1989). Evidence that the up-regulation of cytoplasmic mRNA at least partially reflects increased transcription rates has been reported (Kronke et al., 1985; Lindsten et al., 1989). Lymphokine synthesis can also be reactivated in T cell clones through receptors other than the TCR, in at least some cases via distinct intracellular signaling pathways (Table 111).For example, IL-2 can!trigger various clones to produce GM-CSF, IFN-.)I,and IL-4 (Howard et al., 1983; Inaba et al., 1983; Kelso et al., 1986; Farrar et al., 1986; Dunn et al., 1987; Kelso and Gough, 1989). We have noted, however, that IL-2 is a poor stimulus of multi-CSF production. In several T cell clones GM-CSF and IFN-.)I production can be stimulated by TCR ligands or by IL-2 (albeit more weakly in the latter case), but significant multi-CSF synthesis only occurs in response to TCR triggering (Kelso and Gough, 1989). Similar results have been obtained with polyclonal T cell blasts o r T cells activated in viuo in a graft-versus-host reaction. This suggests that activation of multi-CSF synthesis depends on signaling events not obligatory for GM-CSF o r 1FN--ysynthesis. It has now been shown by several groups working with T cell tumors and clones that TCR binding triggers the hydrolysis of phosphatidylinositol bisphosphate to yield diacylglycerol and inositol triphosphate, which in turn activate protein kinase C (PKC) and effect a rise in the intracellular free Ca2+ concentration (Taylor et al., 1984; Weiss et al., 1984; Imboden and Stobo, 1985). Activation of PKC and other Ca2+-
80
ANNE KELSO AND DONALD METCALF
TABLE 111 T w o PATHWAYS FOR ACTIVATION OF CSF SYNTHESIS IN T CELLS
Receptor TCR
IL-2R
Properties of response Stimulated by antigen, anti-TCR antibodies, Con A High titers of GM-CSF, multi-CSF, and IFN-y Inhibited by cyclosporine A or anti-CD4 antibody Mimicked by ionomycin k PMA
Stimulated by IL-2 Moderate titers of GM-CSF and IFNy, very low multi-CSF Not inhibited by cyclosporine A or anti-CD4 antibody Mimicked by PMA
dependent kinases then leads, by less well-defined steps, to the up-regulation of lymphokine synthesis. The same outcome can be achieved by combined stimulation with PMA, which is an analog of diacylglycerol and directly activates PKC, and a Ca2+ionophore such as ionomycin (Truneh et al., 1985; Imboden and Stobo, 1985). By contrast, interaction of IL-2 with its receptor does not cause phosphatidylinositol breakdown or a Ca2+ flux (Mills et al., 1985, 1986). Earlier indications that IL-2 can activate PKC have now been challenged by experiments with PKCdeficient cells which fail to respond to TCR ligands but still respond to IL-2 (Mills et al., 1988; Valge et al., 1988). Whatever they may be, the intracellular signaling events stimulated by IL-2 differ from those activated by antigen and other TCR ligands, and it is intriguing that these two pathways differ in their ability to stimulate multi-CSF synthesis. Two other observations also suggest that the GM-CSF and multi-CSF genes differ in minimal signaling requirements for their expression. Whereas ionomycin, on its own or in combination with PMA, stimulated both GM-CSF and multi-CSF synthesis, PMA on its own was like IL-2 in that it was a selective stimulus for GM-CSF production (Kelso and Gough, 1989). Furthermore, the immunosuppressive drug cyclosporine A, which blocks Ca2+-dependent activation at a point distal to the rise in intracellular Ca2+concentration triggered by ionomycin or TCR binding (Metcalfe, 1984), inhibited the TCR-dependent induction of GM-CSF, IFN-y, and multi-CSF synthesis, but did not prevent IL-2-induced lymphokine production or proliferation (Dunn et al., 1987; Kelso and Gough, 1989). Bickel et al. (1987) have reported that cyclosporine A inhibited the production of IL-2 and multi-CSF, but not GM-CSF, in Con A-stimulated spleen cells and PMA-stimulated EL4 thymoma cells; this result differs quantitatively from our finding that cyclosporine A reduced
T LYMPHOCYTE-DERIVED CSFS
81
Con A-stimulated GM-CSF production to a level attributable to PKC activation alone. Together, these data suggest that the induction of multiCSF synthesis depends absolutely on Ca2+-dependentsignaling events, whereas GM-CSF and IFN-y synthesis can be activated by Ca2+-independent pathways. The physiological significance of JL-2-induced lymphokine synthesis is not known, but its theoretical significance lies in the demonstration that some lymphokine genes can be activated by more than one receptorlinked signaling pathway. Since the production of IL-2 itself is TCR dependent, activation of both the TCR and IL-2 receptor pathways is ultimately determined by antigen recognition. It is interesting to note that the rules governing CSF production are probably different in other cell types. The inductive stimuli are clearly different in monocyte/macrophages, in which lipopolysaccharide or other cytokines stimulate GM-CSF secretion, and in mast cells, in which the production of GM-CSF, multi-CSF, IL-4, IL-5, and IL-6 are triggered by Fc receptor cross-linking (Wodnar-Filipowicz et al., 1989; Plaut et al., 1989). There is also evidence that GM-CSF expression in macrophages, unlike T cells, is regulated at a posttranscriptional level (Thorens et al., 1987).
B. TRANSCRIPTIONAL REGULATION OF CSF GENEEXPRESSION Although there have been few studies showing directly that the induction of lymphokine synthesis is due to increased transcription (Kronke et al., 1985; Lindsten et al., 1989), analysis of the control of lymphokine gene expression has concentrated on the identification of transcriptional regulatory sequences and nuclear factors. This has proceeded in three stages: (1) the search for noncoding sequences common to several lymphokine genes which might be responsible for their coordinate induction, (2) delineation of functional regulatory sequences, and (3) identification of nuclear proteins that bind to promoter regions. With the cloning and sequencing of CSF and other lymphokine genes, it was found that nucleotide similarities were generally weak in coding regions, but were significant in adjacent 5' and 3' regions. In particular, a conserved stretch of ten nucleotides was noted 100-300 bp upstream from the human and murine GM-CSF, G-CSF, IL-2, multi-CSF, IL-4, IL-5, and IFN-y genes (Stanley et al., 1985; Miyatake et al., 1985; Kelso and Cough, 1987; Shannon et al., 1988; Yokota et al., 1988) (CK-1 in the terminology of Shannon et al.). Shannon et al. also identified an adjacent seven-nucleotide sequence common to the GM-CSF and multi-CSF genes (which they called CK-2). By transfection of constructs containing 5'-flanking sequences of the
82
ANNE KELSO A N D D O N A L D METCALF
human GM-CSF gene linked to the chloramphenicol acetyltransferase (CAT) reporter gene, Chan et al. (1986) showed that a 660-bp fragment (from about -630 to +37) contained sequences required for inducible CAT expression in a lymphoblastoid cell line. This group then identified a sequence at -57 to +37 which was sufficient for GM-CSF induction in three T leukemic cell lines and which bound nuclear extracts from two of those lines (Nimer et al., 1988). This suggested that the CK-1 sequence further upstream is not necessary for GM-CSF promoter activity in these lines, although it does fall within a second positive regulatory sequence at - 106 to -68. A negative regulatory sequence was also localized between - 193 and - 179. Interestingly, in T leukemic cells cotransfected with the GM-CSF promoter-CAT construct and the tax (p40") gene of human T cell leukemia virus type I or 11, CAT expression was increased in the absence of exogenous stimulation, indicating that tax proteins can transactivate GM-CSF expression (Nimer et al., 1989). Transactivation did not occur in several non-T cell lines and a GM-CSF nonexpressing T cell line. Work by Miyatake et al. (1988) defined tax-responsive elements in a region including CK-1 and in the GC-rich region adjacent to CK-2. Shannon et al. (1988,1989) identified two proteins (NF-GMa and -b) in nuclear extracts of PMA-stimulated bladder carcinoma U5637 cells, which could bind specifically to an oligonucleotide encompassing the CK-1 and CK-2 sequences (-96 to -75). NF-GMa was present in unstimulated U5637 cells, which transcribed low levels of GM-CSF, and in several other hemopoietic cell lines. NF-GMb showed a more restricted distribution and, unlike NF-GMa but like GM-CSF itself, was up-regulated following stimulation. Extracts from a GM-CSF-nonexpressingcell line did not bind CK-1 and 2. Only NF-GMa bound to oligonucleotides spanning the CK-1 containing region of the G-CSF gene. Together, these data indicate that the 5' sequences of the human GM-CSF gene contain both positive and negative transcriptional control elements and include regions able to bind nuclear proteins whose expression can vary both with the cell line and with the inductive stimulus. Similar studies of the human IL-2 gene have also identified positive and negative 5'-regulatory elements and some corresponding nuclear factors (Fujita et al., 1986; Nabel et al., 1988; Crabtree, 1989). With further characterization of these and the NF-GM factors, it will be interesting to see whether the same DNA-binding proteins interact with the promoters of different coordinately expressed lymphokine genes, and whether different factors account for the differential expression of these genes by T cells and non-T cells or in response to different stimuli. It is important to note that the above studies used T cell lines that probably correspond to activated T cells in which the differentiative
T LYMPHOCYTE-DERIVED <;SFs
83
events have already occurred that allow lymphokine genes to be expressed. The accessibility of genes to transcriptional factors and polymerases is likely to be a primary determinant of the differences in lymphokine production between T cells and non-T cells, between naive and activated T cells, and perhaps also among T cells with different lymphokine production patterns. Indeed, for the human IL-2 gene numbers of DNase-hypersensitive sites (which correspond to regions of open chromatin structure) range from none in nonhemopoietic cell lines to three in nonexpressing hemopoietic lines and unstimulated T cells and four in stimulated T cells (Siebenlist et al., 1986). Posttranscriptional Regulation of CSF Gene Expression
Shaw and Kamen (1986) noted that AU-rich sequences were common in the 3'-untranslated regions of a number of transiently expressed cytokine and protooncogene mRNAs. They showed, moreover, that linkage of the AU-rich sequence of the human GM-CSF gene to the 3' end of the P-globin gene substantially reduced the half-life of the globin transcript. More recently, Wreschner and Rechavi (1988) reported that the stability of a large number of mRNAs was inversely proportional to the incidence of AU sequences in their 3'-untranslated regions. Experiments in which the 3'-untranslated region of the bovine IL-2 gene was linked to the 3' end of the CAT gene and transfected into various cell types demonstrated an IL-2 sequence-associated repression of CAT activity that depended on synthesis of a labile lymphocyte-specific protein (Reeves et al., 1987). AU-rich sequences in lymphokine transcripts may therefore be recognition signals for specific nucleases which thus contribute to the transience of lymphokine synthesis. The finding that antLCD28 enhancement of anti-CD3-induced production of GM-CSF, IL-2, IFN-y, and TNF-a in human blood T cells is associated with increased half-life of the lymphokine transcripts, but not of several other mRNAs with AU-rich 3'-untranslated regions, indicates that lymphokine production can be specifically and coordinately modulated at the level of mRNA stability (Lindsten el al., 1989). V. T lymphocyte Heterogeneity
From many studies of the effects of purified recombinant lymphokines and antilymphokine antibodies, it is now thought that the combination of lymphokines present in a culture or at the site of an immune response has a profound influence on the form of that response. T w o important examples occur in hemopoiesis and in antibody isotype switching by B lymphocytes.
84
ANNE KELSO AND DONALD METCALF
In the first case the mature cell types that develop from bone marrow stem cells are affected both in vitro and in vivo by CSFs of various lineage specificities, whose activities are further modulated by other cytokines such as IL-1 (hemopoietin-1, which augments the effects of multi-CSF on multipotential cells), IL-4 (which potentiates multi-CSF-dependent mast cell production, G-CSF-dependent granulocyte production, and erythropoietin-dependent erythrocyte production), and IFN-y and TNF (both of which inhibit colony formation) (Rennick et al., 1987; Peschel et al., 1987; Gajewski et al., 1988). Similarly, different lymphokines induce, enhance, or inhibit the switching of immunoglobulin gene expression from one heavy-chain class to another. In vitro, IL-4 induces B cells to switch to IgGl and, at higher concentrations, IgE synthesis; the IgGl response can be enhanced synergistically when IL-2 and IL-5 are also present, and both IgCl and IgE synthesis are strongly inhibited by IFN-y (Snapper and Paul, 1987; McHeyzer-Williams, 1989).Conversely, IFN-y stimulates switching to IgGP, and this response is inhibited by IL-4. The in vivo effects of anti-IL-4 and anti-IFN-y antibodies on serum antibody isotypes in several immune responses suggest that IL-4 and IFN-y play similar roles in vivo as in vitro (Finkelman et al., 1986,1988; Coffman et al., 1989).Reciprocal effects of IL-4 and IFN-y on the production of TNF-a, IL- 1, and prostaglandin E2 by human monocytes have also been described (Hart et al., 1989). In both systems the outcome depends on the balance between different regulators, which can act synergistically or antagonistically on the same target cells. The question then arises as to how the production of such factors is controlled to achieve an appropriate response to infection or other immune or hemopoietic stress. There are a number of possible mechanisms for the differential expression of lymphokine genes, but in recent years attention has focused on heterogeneity in the T cell compartment. This has led to a search for patterns in the lymphokine profiles of T cell clones that might identify distinct subsets, and to comparisons of clones from animals immunized with various antigens to look for preferential production of appropriate lymphokines. A. CD4' AND CD8' SUBSETS The major phenotypic subsets of peripheral T cells are the CD4' and CD8' cells, which recognize antigenic peptides in association with MHC class I1 and I molecules, respectively. Since MHC class I1 molecules are thought mainly to present peptide fragments of endocytosed extracellular antigens, whereas MHC class I molecules present intracellular peptides (Braciale et al., 1987), the CD4' and CD8' subsets would be expected to be activated by quite different immunogens and infectious
T LYMPHOCYTE-DERIVED CSFS
85
agents. Classically, "helper" T cell function for B cell responses (usually to exogenous proteins) has been attributed to CD4' cells, and cytolytic T cell activity (for virus-infected or tumor target cells expressing endogenous antigens) has been associated with CD8' cells. This has led to a common perception that lymphokines (some of which directly mediate B cell help) are the products of CD4' cells. It is interesting then that qualitative differences in lymphokine production patterns between CD4' and CD8' T cell clones have not been observed. Morris et al. (1982) first showed that a virus-specific CD8' cytolytic T cell clone released IFN-y when it interacted with virus-infected target cells, and many later studies have substantiated that CD8' cell populations and clones can secrete various lymphokines, including CSFs (Guerne et al., 1983, 1984; Kelso et al., 1982; Prystowsky et al., 1982). Our own analyses of primary and long-term alloreactive T cell clones have found striking variations in the combinations and levels of various lymphokines and their mRNAs, including GM-CSF, multi-CSF, and IL-5, and all of these variations could be found among both CD4' and CD8' clones (Kelso and MacDonald, 1982; Kelso and Glasebrook, 1984; Kelso and Metcalf, 198513; Kelso and Cough, 1988). However, although no consistent qualitative differences were detected, we did observe that average production levels from CD8' clones were ten- to 100-fold lower than for CD4' clones (Table IV). When assay sensitivities were limiting, this quantitative difference was reflected in lower frequencies of lymphokine-producing CD8' clones. A consequence of such a difference is that at least 90% of the secreted lymphokine would come from CD4' cells in an equal mixture of the two cell types. The conclusions of these clonal studies have recently been extended to T cells activated in vivo in an acute graft-versus-host reaction. Both CD4' and CD8' cells from mice with GVH disease produced GM-CSF and/or multi-CSF when restimulated in vitro for 24 hours, but the frequencies and average activities of CSF-secreting cells in the CD8' population were lower than for the CD4' cells (A. Kelso, unpublished observations). A low frequency of multi-CSF-secreting cells was also found among cytolytic CD8' T cells activated in vivo against allogeneic tumor cells. It is not known whether the quantitative differences between CD4' and CD8' clones are genetically determined-for example, by linkage to CD4 and CD8 expression-or whether they reflect differences in signaling strength under the conditions used. It is interesting to consider that differences between the cytolytic activities of CD4' and CD8' cells may also be quantitative rather than absolute, suggesting reciprocal regulation of lymphokine and cytolytic protein synthesis in the two subsets.
86
ANNE KELSO AND DONALD METCALF
TABLE IV COMPARISON OF LYMPHOKINE PRODUCTION LEVELS BY CD4' CD8' T CELLCLONES" mRNA
AND
Protein
Clones
GM-CSF
Multi-CSF
IFN--y
GM-CSF
Multi-CSF
IL-2
CD4' CD8'
35 8.6
18 4.5
84 48
11,184 1,869
1,469 172
197 25
" Mean values are shown for 39 CD4' and 40 CD8' alloreactive T cell clones assayed for lymphokine mRNA and secreted protein after 8 hours' incubation with antLCD3 antibody (Kelso and Cough, 1988). Other studies indicated that differences between high- and low-producer clones increased with longer incubation times. RNA values are relative hybridization levels in Northern analyses, and protein levels are units per lo6 cells, determined by biological assays. B. T h l AND Th2 SUBSETS Until 1986 several analyses of large panels of murine T cell clones had failed to find any evidence that clones could be classified into subsets based on their lymphokine profiles. In an early study we found that most alloreactive clones produced IFN-ylmacrophage-activating factor (Kelso and Glasebrook, 1984). Many of these also produced colony-stimulating activity and a minority secreted detectable IL-2, but there was no clear association or dissociation among the three activities. A tendency for IL-2 synthesis to be associated with a high production of CSF of IFN-y could be attributed to the direct enhancing effect of IL-2 on lymphokine synthesis (Kelso et al., 1984). In a later study of another group of clones, most produced GM-CSF, some also produced multi-CSF, and a fraction of multi-CSF' clones produced the eosinophil CSF, IL-5 (Kelso and Metcalf, 1985b). Interestingly, in view of the T h 1/Th2 division discussed below, five of six IL-5' clones secreted detectable IL-2. As before, there was a tendency for a high production of different factors to be associated. No evidence was obtained for the production of M-CSF or G-CSF by any of these clones. At the same time Sanderson et al. (1985a,b) screened a large group of primary alloreactive and Mesocestoides corti-reactive clones using various biological assays for GM-CSF, IL-2, multi-CSF, IL-4, IL-5, and IFN. While the specificity of these and the other lymphokine assays then available is not always clear, these studies were like our own in suggesting that clones made up a continuous distribution for the production of each activity. There was evidence, however, that antigen-specific clones from mice infected with the cestode M. corti contained a higher frequency of
T LYMPHOCYTE-DERIVED CSFS
87
IL-5 producers than the alloreactive group, consistent with the presence of serum IL-5 and peritoneal eosinophilia in the infected host (Sanderson et al., 198513; Strath and Sanderson, 1986). Much interest was then aroused when Mosmann and colleagues reported that murine CD4’ T cell clones could be classified into two major groups based on their lymphokine production patterns: T h l , which secrete IL-2, IFN-y, TNF-/3, GM-CSF, and multi-CSF; and Th2, which secrete IL-4, IL-5, IL-6, GM-CSF, and multi-CSF (Mosmann et al., 1986; Cherwinski et al., 1987; Mosmann and Coffman, 1989). Parallel studies by Mosmann, Bottomly, Janeway, and others have built u p a compelling picture of two discrete CD4’ T cell subsets whose mutually exclusive expression of “inflammatory” and “helper” lymphokines might account for many instances in which immunogens tend to activate one or the other type of response (reviewed by Bottomly, 1988; Janeway et al., 1988). Since our own studies had indicated that clones could coexpress IL-2 and IL-5 (Kelso and Metcalf, 1985b), screening of a new panel of CD4’ and CD8’ alloreactive clones was undertaken to test the dissociation of IL-4 from IFN-y and IL-2(Kelso and Gough, 1988). Lymphokine gene expression was assayed at the mRNA level using Northern hybridization and, to avoid the potentially selective or differentiative effects of longterm culture, clones were analyzed as early as possible, at an average of 22 days after cloning from 5-day mixed-leukocyte cultures. Data were obtained for about 30% of all primary clones. Most clones transcribed IFN-y, GM-CSF, and multi-CSF, and only a minority transcribed IL-4 (Table V). By biological assays for the protein product, 100% of the clones secreted GM-CSF, multi-CSF, and IL-2, indicating that the protein assays were more sensitive than the mRNA analyses and, more importantly, that no clones conformed to the IL-2Th2 pattern. Most IL-4’ clones also transcribed IFN-y and so were not of the T h 2 type. In fact, the frequencies of coexpression of IL-4 and IFN-y, as well as of GM-CSF and multi-CSF, were close to those predicted from the individual frequencies if overlap were random. Of the 16 possible patterns of expression of these four lymphokine genes, eight were observed at the mRNA level, and most of these were found in both CD4’ and CD8’ clones. In three other studies human clones of CD4’ and CD8’ phenotypes also failed to conform to the T h 1/Th2 classification and included many IFN-yf IL-4’ clones (Umetsu et al., 1988; Paliard et al., 1988; Maggi et al., 1988). Similarly, Firestein et al. (1989) have now reported that primary keyhole limpet hemocyanin-reactive clones from immunized mice included both single and double producers of IL-2 and IL-4; these authors
88
ANNE KELSO A N D D O N A L D METCALF
TABLE V PATTERNS OF LYMPHOKINE mRNA EXPRESSION I N SHORT-TERM ALLOREACTIVE T CELLCLONES“
GM-CSF Multi-CSF IFN-y
-
+ + + ++ +
IL-4
-
+ + +
Clones (%)
1 4 8 8
68
2
1 9
Data are shown for 102 T cell clones assayed for lyrnphokine transcripts by Northern blot hybridization after 8 hours’ stimulation with antiCD3 antibody (Kelso and Cough, 1988).
suggest that double producers represent a distinct subset, “Tho,” which might be precursors or a transition state between T h l and Th2 cells. Recent data from Mosmann’s group also now support the conclusion that, in the absence of intentional priming and/or early in culture, most clones do not display mutually exclusive Th 1 or Th2 patterns of lymphokine gene expression (Mosmann and Coffman, 1989). The tentative consensus, therefore, appears to be that peripheral T cells are not precommitted to the T h l and Th2 phenotypes, but that such restricted patterns of expression are more frequently found among clones from immunized mice and after prolonged in vitro culture. It has not been established whether T h l and Th2 clones are derived from uncommitted precursors or from minor preexisting populations. However, their preferential outgrowth in uitro can be promoted by some cytokines and accessory cells, such as IFN-y, which inhibits proliferation of Th2, but not T h l , clones; IL-1, for which it has been reported that only Th2 clones have receptors: and different antigen-presenting cells which selectively stimulate T h l or Th2 clones (Gajewski et al., 1988; Lichtman et al., 1988; Fernandez-Botran et al., 1988; Weaver et al., 1988). The Thl/Th2 model and its recent derivatives are based on studies of T cell clones in vitro, and its relevance to T cell function in vivo is not yet known. There are some interesting examples of in vivo responses which are now interpreted in the framework of the Thl/Th2 subdivision. One is the genetically determined response to Leishmania major, in which in
T LYMPHOCYTE-DERIVED CSFS
89
vitro studies have suggested a central role for IFN-y in parasite clearance from infected macrophages. Heinzel et al. (1989) found that C57BL/6 mice expressed mRNA for IFN-y, IL-1, and IL-2, but not IL-4, in their draining lymph nodes after infection and went on to eliminate the parasite; BALB/c mice, on the other hand, expressed IL-4, IL-1, and IL-2, but not IFN-y, mRNA and were unable to clear the infection. This and several other murine immune responses point to a dichotomy between IFN-y- and IL-4-mediated activation of host defense mechanisms, which, in the case of L . major, appear to be determined at the level of lymphokine production. While consistent with the selective activation of T h l or Th2 cells, the actual mechanisms leading to preferential lymphokine production have not been elucidated in any of these systems. It should also be noted that others have suggested a correlation between multi-CSF synthesis and susceptibility to L. major infection (Lelchuk et al., 1988; Feng et al., 1988). If T h l and T h 2 clones do indeed correspond to functionally significant subsets in vivo, it will be important to determine their lineage relationships and to define the signals that either stimulate their expansion from committed precursors or direct their differentiation from uncommitted cells. T h e CD45/CD45R activation markers and recently described murine T cell antigens that appear to mark populations expressing different lymphokine profiles (Powrie and Mason, 1988; Bottomly et al., 1989; Hayakawa and Hardy, 1988) may facilitate such studies. Attention should then shift to the various classes of antigenpresenting cells, whose role in delivering the secondary signals required for primary T cell activation may be critical for the selection or instruction of T cells expressing appropriate lymphokine genes. OF CSF GENES IN T CELLCLONES C. DIFFERENTIAL EXPRESSION As the above discussion indicates, most studies of the regulation of lymphokine synthesis in T cells have focused on the coordinate production of several lymphokines by a given clone and the synthesis of different lymphokine combinations by different clones. There is evidence, however, for the noncoordinate production of lymphokines by T cell clones at several levels, suggesting mechanisms for differential lymphokine synthesis other than those offered by interclonal heterogeneity. In studying the synthesis of GM-CSF and multi-CSF by several stable long-term CD4' T cell clones, we have observed dissociations in the production of these two factors under the following conditions. First, as noted in Section IV and Table 111, two functionally distinct activation pathways can be defined that differ in their relative induction of GM-CSF and multi-CSF mRNA accumulation or protein secretion. One is acti-
90
ANNE KELSO AND DONALD METCALF
vated by antigen, Con A, anti-TCR antibodies, or ionomycin f PMA and induces synthesis of high levels of GM-CSF, IFN-y, and multi-CSF. This response is inhibited by cyclosporine A and by anti-CD4 antibody in the case of TCR-dependent stimulation. The other is activated by IL-2 or PMA and induces the synthesis of lower levels of GM-CSF and IFN-y; clones vary in their IL-2induced multi-CSF titers, but these are always very low. This pathway is resistant to inhibition by cyclosporine A and anti-CD4 antibody (Kelso et al., 1986; Kelso and Owens, 1988a; Kelso and Cough, 1989). Thus, although optimal activation of the GM-CSF and multi-CSF genes seems to require similar signals, GM-CSF expression can also be induced under conditions that trigger little, if any, multi-CSF synthesis. Second, in the response of some clones to Con A or anti-TCR antibodies, GM-CSF and multi-CSF production is temporally dissociated (Cough and Kelso, 1989). GM-CSF is detectable earlier than multi-CSF at both the cytoplasmic mRNA and secreted protein levels in all such clones, including those in which multi-CSF is ultimately the major product. The approximate length of the lag period between the onset of GM-CSF and multi-CSF secretion is a stable property of each clone and can be as great as several hours. Third, in single-cell studies of anti-TCR-induced CSF synthesis, the distribution of GM-CSF and multi-CSF titers among cells of some clones suggests a difference in the threshold for activation of the two genes (Fig. 1) (Kelso and Owens, 1988b). Even within a clone, cells varied over a
FIG. 1. CSF secretion by single cells of three T cell clones. Two hundred cells of each clone were micromanipulated into wells coated with anti-TCR antibody and incubated for 24 hours. The CSF content of culture supernatants was then determined in assays with FDC-P1 cells (which respond to GM-CSF and/or multiCSF) or 32D c13 cells (which respond only to multi-CSF) (Cough and Kelso, 1989). Since the two assays have the same sensitivity to multi-CSF, cells synthesizing FDC-PI activity without 32D c13 activity are interpreted as producing GM-CSF without multi-CSF.
T LYMPHOCYTE-DERIVED CSFS
91
100-fold range in their 24-hour CSF titers. Cells at the lower end of the spectrum secreted GM-CSF without detectable multi-CSF, whereas higher-producing cells synthesized multi-CSF, apparently accompanied by GM-CSF. Multi-CSF- cells were not genetic variants, since such cells could give rise to multi-CSF-secreting progeny. Assuming that CSF output is related to the size of the signal received by the cell, these results are interpreted to suggest that a stronger signal is required to activate multiCSF synthesis than GM-CSF synthesis. Interestingly, for each of these responses in three clones analyzed in detail, there was a preference for production of GM-CSF over multi-CSF, even though GM-CSF was not always the major product induced under optimal conditions. Moreover, the three types of GM-CSF-multi-CSF dissociation outlined above were correlated (Table VI), suggesting that they might reflect a single underlying mechanism (Gough and Kelso, 1989). Too few clones have been analyzed for these traits to assess whether the three clones represent common phenotypes, but single-cell CSF assays of polyclonal T cell populations revealed a heterogeneous distribution of E9.D4-like and D1 .MG-like cells (A. Kelso, unpublished observations). Although a number of possibilities exist, only some of which have been excluded by experiment (Kelso and Gough, 1987; Gough and Kelso, 1989), we favor the idea that the bias toward GM-CSF synthesis arises from the contiguity of the GM-CSF and multi-CSF genes and their chromatin structures. In this model activation of GM-CSF transcription would TABLE VI PREFERENTIAL SYNTHESIS OF GM-CSF OVER MULTI-CSFIN THREE T CELLCLONES
~
~
Clone
GM-CSF-multi-CSF lag (minutes)"
GM-CSF: multi-CSF ratiob
$6 Multi-CSF- cells of all CSF+ cellsC
LB3 E9.D4 Dl.M6
94 51 11
106 21 3
54 61 5
a Lag times represent the delay between the onsets of GM-CSF and multi-CSF secretion after Con A stimulation (mean, seven to ten experiments). Ratios are the FDC-P1 titer:32D c13 titer ratios in culture supernatants after IL-2 stimulation for 24 hours (mean, four to nine experiments). The percentage of multi-CSF- cells was determined in assays of 200 cells and represents the percentage of FDC-P1' 32D- cells of all FDC-PI+ cells (mean of two or three experiments, one of which is shown in Fig. 1) (Gough and Kelso, 1989).
_
_
92
A N N E KELSO A N D D O N A L D METCALF
enable activation of the multi-CSF gene by reducing conformational constraints on the accessibility of the multi-CSF promoter. These constraints would be greater in LB3 and E9.D4 cells than in the Dl.M6 clone and would account for clonal variations in the GM-CSF/multi-CSF lag period and responsiveness of the multi-CSF gene to activation by suboptimal signaling events o r signal strength. The functional significance of the preference for GM-CSF expression is unclear. It may be a fortuitous consequence of the tight linkage of the GM-CSF and multi-CSF genes o r it may, for example, reflect a need for more stringent regulation of multi-CSF expression to maintain a balance between beneficial and pathological effects of this factor. It is not known whether the structurally linked genes for IL-4 and IL-5 are also functionally linked in this manner, either to each other or to GM-CSF and multi-CSF. However, it may be important that these genes are 100200 kb apart and are separated from GM-CSF/multi-CSF by at least 200 kb in humans and at least 600 kb in mice (van Leeuwen et al., 1989; Lee et al., 1989), whereas GM-CSF and multi-CSF are only 9 and 14 kb apart in humans and mice, respectively (Yang et al., 1988; Lee and Young, 1989). VI. CSF Production by 1lymphocytes in Vivo
A. In Vivo EFFECTSOF CSFs With the availability of purified recombinant CSFs and their cDNAs, it has become possible to investigate in vivo responses to these regulators, either by injection of the protein or by expression of the introduced DNA. On the whole such studies have indicated that the target populations and actions of the CSFs are similar in vitro and in vivo, but that the long-term consequences probably depend on the site and chronicity of exposure. In our own studies repeated intraperitoneal injection of GM-CSF (three times daily for 6 days) caused selective dose-dependent increases in monocyte/macrophage and neutrophil numbers in the blood, liver, lung, and peritoneal cavity, as well as in peritoneal eosinophil numbers (Metcalf et al., 1987b). A similar intraperitoneal administration regimen with multi-CSF affected a broader target cell range, inducing rises in monocyte/macrophage, neutrophil, and eosinophil numbers in the blood and peritoneal cavity and in promyelocyte, myelocyte, eosinophil, nucleated red cell, megakaryocyte, mast cell, and progenitor cell numbers in the spleen. The most dramatic of these effects was a rise of up to 100-fold in splenic mast cells, which was not observed with GM-CSF (Metcalf el al.,
1986a). Both GM-CSF and multi-CSF injections stimulated macrophage phagocytic activity for opsonized red blood cells. T h e serum half-lives of intraperitoneally injected CSFs were short, at about 35 minutes for GM-CSF and about 20 minutes for multi-CSF. In the protocols used local and circulating CSF levels would have persisted at levels stimulatory in vitro for only 2-4 hours after injection. However, the induced marrow and tissue changes indicated that, even at these limiting concentrations, both CSFs could exert significant central and peripheral effects. Intravenous injection of radiolabeled multi-CSF and the subsequent detection of labeled hemopoietic cells in the marrow and spleen further indicated that multi-CSF could gain access to these tissues and hence that the observed effects were likely to be direct (Metcalf and Nicola, 1988). T h e protective and therapeutic effects of injected CSFs have been investigated in several murine models of hemopoietic suppression by radiation and cytotoxic drugs, and of infectious disease. While CSFs exert a positive effect in the former case by accelerating hemopoietic recovery (Talmadge et ul., 1989),their effects on host defense appear to depend on the immune mechanisms required to eliminate the pathogen. For example, multi-CSF administration overcame the inability of athyniic mice to raise a ‘I‘ cell-dependent mucosal mast cell response to Strongyloides rutti and to expel the parasite (Abe and Nawa, 1988). On the other hand, multi-CSF injection of genetically susceptible mice infected with L. mujor increased lesion size and parasite multiplication, perhaps by increasing the pool of macrophages available for infection (Feng et al., 1988). Following various studies in primate models, GM-CSF, G-CSF, and multi-CSF are now being tested in phase I and I1 clinical trials for their ability to enhance hemopoietic recovery following radiotherapy or cytotoxic drug treatment in cancer patients and transplant recipients, and in infections such as the acquired immunodeficiency syndrome (reviewed by Clark and Kanien, 1987; Morstyn et ul., 1989). Of all of the cytokines that have been tested as therapeutic agents, the CSFs look particularly promising for the treatment of myelosuppression, because of their ability to achieve sustained increases in blood neutrophil numbers with few adverse side effects. ‘Their low toxicity (especially in the case of G-CSF) probably reflects their short serum half-lives, restricted target cell range, and apparently preferential action in vivo on the production of hemopoietic cells rather than on the activation of potentially damaging end-cell functions. It is therefore interesting that the chronic deregulated production of GM-CSF in vivo can be lethal. In transgenic mice expressing GM-CSF constitutively under the control of a retroviral promoter, high circulating
94
ANNE KELSO A N 0 DONA1.1) METCALF
GM-CSF levels were accompanied by macrophage accumulation in the eyes, pleural and peritoneal cavities, and striated muscle, and most mice died with muscle wasting (Lang et al., 1987). Transgene expression appeared to be restricted to the macrophage lineage, in which it may have exerted an autocrine stimulatory effect. A related syndrome occurred in irradiated mice transplanted with bone marrow cells infected with a GM-CSF-expressing retrovirus. In these mice, which died within 4 weeks of transplantation, high serum levels of GM-CSF were associated with infiltration of the spleen, liver, lung, peritoneal cavity, and muscles with macrophages and neutrophils, as well as eosinophils in muscle (Johnson et al., 1989).
B. In Vzz!o DETECTION OF CSFs Colony-stimulating activity is readily detectable in vivo in the serum, urine, and other fluids and in extracts of virtually any tissue, and levels rise dramatically following endotoxin injection (Metcalf, 1984). However, the major constitutive serum CSF is M-CSF, and the major endotoxin-induced serum type is G-CSF. Neither of these is likely to be T cell derived. Following endotoxin injection, GM-CSF is a minor activity in the serum (Metcalf, 1988), but is produced at elevated levels by mouse lung tissue and many other organs not likely to contain significant numbers of lymphocytes, such as the salivary gland and kidney (Sheridan and Metcalf, 1972, 1973). Multi-CSF, on the other hand, has not been demonstrated in the serum or tissue extracts from normal, endotoxin-injected, or parasitized mice, or in graft-versus-host or autoimmune disease (Garland et al., 1983; Strath and Sanderson, 1986; Metcalf et al., 1987a; Firestein et al., 1988), and it remains to be determined whether endogenous multi-CSF is ever a systemic mediator. It shares this property with some other T cell lymphokines, including IL-2 and IL-4, whose detection in the circulation o r other fluids has either not been reported or is controversial (e.g., Firestein et al., 1988). There has been one report of IL-5 detection in the serum, namely in mice infected with the cestode M. corti, in which a rise in the circulating IL-5 level precedes increases in eosinophil numbers in the bone marrow, spleen, and peritoneal cavity (Strath and Sanderson, 1986).Since eosinophilia does not occur in M. corti-infected athymic mice (Johnson et al., 1979) and IL-5-secreting M. corti-reactive T cell clones can be isolated from the spleens of infected euthymic animals (Sanderson et al., 1985b),it is probable that T cells are the in vivo source of IL-5 in this model. The infrequent detection of T cell-derived lymphokines in vivo contrasts with the many accounts of serum presence of those cytokines
whose major sources are thought to be macrophages, fibroblasts, and endothelial cells. M-CSF and IL-6 are present in normal human and mouse sera, and G-CSF, IL-1, IL-6, and TNF-a have been detected in serum or cerebrospinal fluid during bacterial, viral, or parasitic infections, autoimmune diseases, o r graft-versus-host, or host-versus-graft reactions (Maury and Teppo, 1987; Helfgott et al., 1989; Cheers et al., 1988). In addition, the presence of M-CSF, IL-1, IL-6, and TNFa mRNAs in many tissues of normal animals and individuals, and tissuespecific expression of GM-CSF, G-CSF, IL-1, and IL-6 mRNAs in endotoxin-injected mice, have been reported (Tovey et al., 1988; Troutt and Lee, 1989). It is not clear whether the failure to detect T cell products in the serum is due to their short half-lives o r the presence of specific serum inhibitors, or whether it indicates that they do not normally reach the circulation. However, the in vivo production of lymphokines can be demonstrated in other ways: by detection of mRNA in tissues using bulk o r in situ hybridization, by detection of the protein product in tissue sections using immunohistochemistry, and by the specific effects of injected antilymphokine antibodies. There are as yet only a few reports of lymphokine mRNA or protein detection in fresh tissues. One of these, the demonstration of IL-2 and IFN--y or IL-4 mRNA expression in the lymph nodes of L. major-infected mice, has already been mentioned in Section V (Heinzel et al., 1989). Several other studies have also detected the production of IL-2, IL-4 o r IFN--y in vivo (Buchan et al., 1988; Fox et al., 1985; Kasaian and Biron, 1989; Gessner et al., 1989; Carding et al., 1989). Similar studies of GMCSF, multi-CSF, and IL-5 expression in tissues have not been noted, although there are numerous instances in which one or more of these products have been detected within hours of in vitro stimulation of fresh T lymphocytes. Transcripts of IL-2, -3, -4, and -5 could not be detected in tissues of endotoxin-injected mice (Troutt and Lee, 1989); GM-CSF mRNA in the lungs and hearts of these animals was unlikely to be of T cell origin. T h e best evidence for both the production and the functional importance of these CSFs in immune responses comes from studies with neutralizing antibodies. T h e symptoms of cerebral malaria in mice infected with Plasmodium berghei have been attributed to TNF-a, since neurological signs correlated with an increase in serum TNF levels and were prevented by injecting anti-TNF-a antibody (Grau et a/., 1987). It has now been reported that treatment of P . berghei-infected mice with a combination of anti-GM-CSF and anti-multi-CSF antibodies inhibited the accumulation of macrophages in the spleen, the rise in serum TNF-a,
96
A N N E KELSO A N D DONALD METCALF
and cerebral pathology, suggesting that GM-CSF- and multi-CSF-stimulated macrophages may be the source of the TNF (Grau et al., 1988). Since cerebral malaria is CD4' T cell dependent, it is likely that T cells contribute to this cytokine cascade. In another murine model of parasitic disease, the dramatic blood and lung eosinophilia induced by the helminth Nippostrongylus brmiliensis was completely prevented by the injection of anti-IL-5 antibodies, but was unaffected by anti-IL-4; conversely, the serum IgE response was inhibited by anti-IL-4 but not by anti-IL-5 antibodies (Coffman et al., 1989; Finkelman et al., 1986). In other studies anti-IL-4 antibody blocked the serum IgE response toL. major in susceptible BALB/c mice, anti-IFN-y antibody blocked the serum IgG2a response to killed Brucellu abortus, and anti-TNF-a antibody prevented the development of granulomas in livers of bacille Calmette-Guerin-infected animals (Heinzel et al., 1989; Finkelman et al., 1988; Kindler et al., 1989). In general, these studies have found effects consistent with the in vitro target cells and activities of the corresponding cytokines. In summary, it appears that the cytokines whose major source in vitro is the T lymphocyte can only rarely be detected in the circulation. Their presence in vivo can nevertheless be revealed either at the mRNA level in tissues o r by the effects of specific antibodies. Although not formally demonstrated, the weight of circumstantial evidence favors the conclusion that T cells are indeed the sources of many of these lymphokines in vivo. C. WHYDo T CELLSSYNTHESIZE CSFs? In vitro, T cells can produce much higher levels of GM-CSF than are released by non-T cell sources such as macrophages and fibroblasts. In the case of multi-CSF, as well as IL-4 and IL-5, T cells and mast cells are the only known sources. Moreover, GM-CSF and multi-CSF can be synthesized by most T cell clones, including those of CD4' or CD8' and T h l or Th2 phenotypes. A significant proportion of in vivo-activated T cells also produce these factors. Why should this be so? Several points may be important. First, it seems likely that lymphokine production by T cells is usually a local response, in which biologically active concentrations of CSF are achieved only in the vicinity of the producing cells. This may in part reflect the unique activation requirement of the T cell, namely, that it must recognize antigen as peptide fragments presented by MHC molecules on the surface of another cell. This event usually occurs at the site of antigen deposition, particularly in draining lymph nodes, where the tissue architecture juxtaposes T cells and antigen-presenting dendritic cells or B cells.
There is evidence, moreover, that lymphokine secretion can be directed toward the antigen-presenting cell. Kupfer et al. ( 1987) found that antigen-specific binding of cloned T cells to presenting cells induced reorientation of the microtubule-organizing center to face the region of contact. By placing T cells and anti-TCR antibody on opposite sides of a nucleopore membrane, Po0 et al. (1988) showed that IL-4 was secreted preferentially toward the stimulus when the antibody concentration was limiting, whereas secretion was nondirectional at high stimulus concentrations. Together, these experiments suggest a mechanism for focusing the secreted lymphokine at the surface of the antigen-presenting cell. Effects on more distant target cells would only occur when high antigen densities or enhancing accessory signals induced the production of higher lymphokine titers and/or nonpolarized secretion. Other mechanisms may also exist to retain small quantities of secreted factors in the microenvironment of the producing cell, such as the recently described trapping of CSFs by the extracellular matrix of bone marrow stromal cells (Gordon et al., 1987; Roberts el al., 1988). T h e local polarized secretion of lymphokines by T cells contrasts with the likely situation for some other CSF-producing cells, in which secretion is induced by soluble stimuli such as endotoxin and certain cytokines. This may be particularly important in the bloodstream, where circulating cells (especially monocytes) and endothelial cells of the vessel walls may be responsible for the presence in the serum of G-CSF, IL- 1, IL-6, and T N F in many infections and disease states. It may then be more appropriate to consider the vascular system as a CSF-producing tissue than as a sink for CSFs from other sites. A consequence of localized production oflymphokines by T cells is that the CSFs produced would not have the opportunity to affect hemopoiesis in the bone marrow. Instead, their target cells would mainly comprise tissue macrophages and mast cells and circulating monocytes, neutrophils, eosinophils, and other blood cells passing through the lymphoid tissue or the site of immune activation. In that case the major effects of T cell-derived CSFs would be to attract and stimulate the functional activities of their targets, rather than to boost their production from progenitors. Although the evidence is still limited, the indications are that GM-CSF and multi-CSF production must generally be beneficial for the immune response, increasing antigen-presenting cell function and stimulating mechanisms for the clearance of extracellular and intracellular pathogens. As discussed above, toxicity associated with these factors has usually been noted when they are synthesized or administered chronically, at high concentrations and at inappropriate sites. Under normal conditions
98
ANNE KELSO AND
DONALD
METCALF
their transience is ensured by the multiple mechanisms limiting lymphokine production by T cells. The finding that most, if not all, T cell clones synthesize both GM-CSF and multi-CSF further suggests that expression of these regulators does not preclude the activation of host defense mechanisms by other lymphokines. Unlike JL-4 and IFN-y, the T cellderived CSFs are not known to antagonize the actions of other cytokines or to divert their target cells to other activities. In conclusion, CSF production by T lymphocytes is not necessary for normal hemopoiesis and, even during immune stress, might not directly affect the bone marrow. It is suggested instead that the role of T cellderived CSFs is to modulate the activities of hemopoietic cells at the site of antigenic activation. T h e strict dependence of T cells on recognition of antigen o r antigen/TCR-dependent signals (such as IL-2 and accessory molecules) for induction of CSF synthesis and the focusing of the secreted product on the cognate target cell place geographical constraints on the access of these potent regulators to their ubiquitous target cells. Such compartmentalization may be an important level of regulation which distinguishes the synthesis of GM-CSF and multi-CSF by T cells from the elevation of G-CSF, M-CSF, and other cytokines in systemic responses. REFERENCES Abe, T., and Nawa, Y. (1988).Immunology 63, 181-185. Barlow, D. P., Bucan, M., Lehrach, H., Hogan, B. L. M., and Cough, N. M. (1987).EMBO J . 6,617-623. Begley, C. G., Lopez, A. F., Nicola, N. A., Warren, D. J., Vadas, M. A., Sanderson, C. J., and Metcalf, D. (1986).Blood 68, 162-166. Bickel, M., Tsuda, H., Amstad, P., Evequoz, V., Mergenhagen, S. E., Wahl, S. M., and Pluznik, D. H. (1987). Proc. Null. Acud. Sci. U.S.A. 84,3274-3277. Bottornly, K. (1988).Immunol. Toahy 9,268-274. Bottomly, K., Luqman, M., Greenbaurn, L., Carding, S., West, J., Pasqualini, T., and Murphy, D. B. (1989). Eur. J. Immunol. 19,617-623. Braciale, T. J., Morrison, L. A., Sweetser, M. T., Sarnbrook, J., Gething, M.-J., and Braciale, V. L. (1987).Immunol. Rev. 98,95-114. Bradley, T. R., and Metcalf, D. (1966). A w l . J . Exp. B i d . Med. Sci. 44,287-300. Buchan, G., Barrett, K., Fujita, T., Taniguchi, T., Maini, R., and Feldrnann, M. (1988). Clin. Exp. Immunol. 71, 295-301. Burgess, A. W., Carnakaris, J., and Metcalf, D. (1977).J. Biol. Chem. 252, 19982003. Burgess, A. W., Metcalf, D., Russell, S. H. M., and Nicola, N. A. (1980).Bzochem.J. 185,301-314. Cannistra, S. A., Rarnbaldi, A., Spriggs, D. R., Herrrnann, F., Kufe, D., and Griffin, J. D. ( 1 9 8 7 ) Clin. ~ Invest. 79, 1720- 1728. Carding, S. R., Jenkinson, E. J., Kingston, R., Hayday, A. C., Bottornly, K., and Owen, J. J. T. (1989).Proc. Natl. Acud. Sci. U.S.A. 86, 3342-3345.
Chan, J. Y., Slarnon, D. J., Nimer, S. D., Golde, D. W., and Gasson, J. C. (1986). Proc. Nad. Acad. Sci. U.S.A. 83, 8669-8673. Cheers, C., Haigh, A. M., Kelso, A., Metcalf, D., Stanley, E. R., and Young, A. M. (1988).Infect. Immun. 56, 247-251. Cherwinski, H. M., Schumacher,J. H., Brown, K. D., and Mosrnann, T. R. (1987). J . Exp. Med. 166, 1229-1244. Clark, S. C., and Kamen, R. (1987). Science 236, 1229-1237. Coffrnan, R. L., Seymour, B. W. P., Hudak, S., Jackson, J., and Rennick, D. (1989).Science 245,308-310. Crabtree, G. R. (1989). Science 243, 355-36 1. Cutler, R. L., Metcalf, D., Nicola, N. A., and Johnson, G. R. (1985).J. Biol. Chem. 260,6579-6587. Dunn, D. E., Herold, K. C., Otten, G. R., Lancki, D. W.,Gajewski,T.,Vogel, S. N., and Fitch, F. W. (1987).j. Immunol. 139, 3942-3948. Ely, J. M., Prystowsky, M. B., Eisenberg, L., Quintans, J., Goldwasser, E., Glasebrook, A. L., and Fitch, F. W. (1981).J. Immunol. 127,2345-2349. Farrar, W. L., Thomas, T . P., and Anderson, W. B. (1985).Nature (London)315, 235-237. Farrar, W. L., Birchenall-Sparks, M. C., and Young, H. B. (1986).J . Immunol. 137, 3836-3840. Feng, Z . Y., Louis, J., Kindler, V., Pedrazzini, T., Eliason, J. F., Behin, R., and Vassalli, P. (1988).Eur. J. Immunol. 18, 1245- 125 1. Fernandez-Botran, R., Sanders, V. M., Mosrnann, T. R., and Vitetta, E. S. (1988). J . Exp. Med. 168,543-558. Finkelman, F. D., Katona, I. M., Urban, J. F., Snapper, C. M., Ohara, J., and Paul, W. E. (1986).Proc. NULL.Acad. Sci. U.S.A. 83, 9675-9678. Finkelman, F. D., Katona, I. M., Mosmann, T. R., and Coffrnan, R. L. (1988). J . Immunol. 140, 1022-1027. Firestein, G. S., Xu, W.-D., Townsend, K., Broide, D., Alvaro-Gracia, J., Glasebrook, A. L., and Zvaifler, N. J. (1988).J. Exp. Med. 168, 1573-1586. Firestein, G. S., Roeder, W. D., Laxer, J. A., Townsend, K. S., Weaver, C. T., Hom, J . T., Linton, J., Torbett, B. E., and Glasebrook, A. L. (1989). J. Immunol. 143,518-525. Fischer, H.-G., Frosch, S., Reske, K., and Reske-Kunz, A. B. (1988).J . Immunol. 141,3882-3888. Fleischmann, J., Golde, D. W., Weisbart, R. H., and Gasson, J. C. (1986).Blood 68, 708-7 1 I . Fox, R. I., Theofilopoulos, A. N., and Altrnan, A. (1985).J.Immunol. 135,3 10931 15. Fujita, T., Shibuya, H., Ohashi, T., Yarnanishi, K., and Taniguchi, T. (1986).Cell (Cambridge, Mass.) 46,401-407. Gajewski, T. F., and Fitch, F. W. (1988).J. Immunol. 140,4245-4252. Gajewski, T. F., Goldwasser, E., and Fitch, F. W. (1988).J. Immunol. 141, 26352642. Garland, J . M., Aldridge, A., Wagstaffe, J., and Dexter, T. M. (1983).Br.J. Cancer 48,247-259. Gauldie, J . , Richards, C., Harnish, D., Lansdorp, P., and Baurnann, H. (1987). Proc. Natl. Acad. Sci. U.S.A. 84, 7251-7255. Gearing, D. P., King, J. A., Cough, N. M., and Nicola, N. A. (1989).EMBO J. 8, 3667-3676.
100
ANNE KELSO A N D DONALD MEI'CALF
Gessner, A., Moskophidis, D., and Lehmann-Grube, F. (1989).J.Immunol. 142, 1293-1298. Gomez-Cambronero, J., Yamazaki, A., Metwally, F., Molski, T. F. P., Bonat, V. A., Huang, C.-K., Becker, E. L., and Sha'afi, R. I. (1989).Proc. Natl. Acad. Sci. U.S.A. 86, 3569-3573. Gordon, M. Y., Riley, G. P., Watt, S. M., and Greaves, M. F. (1987). Nature (London) 326,403-405. Cough, N. M., and Kelso, A. (1989). Growth Factors 1, in press. Cough, N. M., Cough, J., Metcalf, D., Kelso, A., Grail, D., Nicola, N. A., Burgess, A. W., and DUN, A. R. (1984). Nature (London) 309,763-767. Grabstein, K. H., Urdal, D. L., Tushinski, R. J., Mochizuki, D. Y., Price, V. L., Cantrell, M. A., Gillis, S., and Conlon, P. J. (1986). Science 232, 506508. Grau, G. E., Fajardo, L. F., Piguet, P.-F., Allet, B., Lambert, P.-H., and Vassalli, P. (1987). Science 237, 1210-1212. Grau, G. E., Kindler, V., Piguet, P.-F., Lambert, P.-H., and Vassalli, P. (1988). J. Exp. Med. 168, 1499-1504. Griffin, J. D., Meuer, S. C., Schlossman, S. F., and Reinherz, E. L. (1984). J.Immunol. 133,1863-1868. Guerne, P.-A., Piguet, P.-F., and Vassalli, P. (1983).J . Immunol. 130, 2225-2230. Guerne, P.-A., Piguet, P.-F., and Vassalli, P. (1984).J . Immunol. 132, 1869-1871. Hamilton, J. A., Vairo, G., and Lingelbach, S. R. (1988).J. Cell. Physiol. 134, 405-4 12. Handman, E., and Burgess, A. W. (1979).J. Immunol. 122, 1134-1 137. Hart, P. H., Vitti, G. F., Burgess, D. R., Whitty, G. A., Piccoli, D. S., and Hamilton, J. A. (1989). Proc. Natl. Acad. Sci. U.S.A. 86, 3803-3807. Hayakawa, K., and Hardy, R. R. (1988).J . Exp. Med. 168, 1825-1838. Heinzel, F. P., Sadick, M. D., Holaday, B. J., Coffman, R. L., and Locksley, R. M. (1989).J. Exp. Med. 169,59-72. Helfgott, D. C., Tatter, S. B., Santhanam, U., Clarick, R. H., Bhardwaj, N., May, L. T., and Sehgal, P. B. (1989).J. Immunol. 142,948-953. Herold, K. C., Lancki, D. W., Dunn, D. E., Arai, K., and Fitch, F. W. (1986).Eur.J. Immunol. 16, 1533-1538. Hirano, T., Yasukawa, K., Harada, H., Taga, T., Watanabe, Y., Matsuda, T., Kashiwamura, S., Nakajima, K., Koyama, K., Iwamatsu, A., Tsunasawa, S., Sakiyama, F., Matsui, H., Takahara, Y., Taniguchi, T., and Kishimoto, T. (1986).Nature (London) 324, 73-76. Howard, M., Burgess, A., McPhee, D., and Metcalf, D. (1979). Cell (Cambridge, MUSS.)18,993-999. Howard, M., Matis, L., Malek, T. R., Shevach, E., Kell, W., Cohen, D., Nakanishi, K., and Paul, W. E. (1983).J. Exp. Med. 158, 2024-2039. Ichikawa, Y., Pluznik, D., and Sachs, L. (1966). Proc. Natl. Acad. Sci. U.S.A. 56, 488-495. Ihle, J. N., Keller, J., Henderson, L., Klein, F., and Palaszynski, E. (1982). J . Immunol. 129,243 1-2436. Imboden, J. B., and Stobo, J. D. (1985).J. Exp. Med. 161,446-456. Inaba, K., Granelli-Piperno, A., and Steinman, R. M. (1983).J. Exp. Med. 158, 2040-2057. Janeway, C. A,, Carding, S., Jones, B., Murray, J., Portoles, P., Rasmussen, R., Rojo, J., Saizawa, K., West, J., and Bottomly, K. (1988). Imrnunol. Rev. 101, 39-80.
LYMPHOCYTE-DERIVED
CSFs
101
Johnson, G. R., Nicholas, W. L., Metcalf, D., McKenzie, I. F. C., and Mitchell, G. F. (1979).Int. Arch. Allergy Appl. Immunol. 59,315-322. Johnson, G. R., Gonda, T. J., Metcalf, D., Hariharan, I. K., and Cory, S. (1989). EMBO J . 8 , 4 4 1-448. Kasaian, M. T., and Biron, C. A. (1989).J. Immunol. 142, 1287-1292. Kelso, A. (1986).J. Immunol. 136, 2930-2937. Kelso, A., and Glasebrook, A. L. (1984).J. Immunol. 132, 2924-2931. Kelso, A., and Cough, N. (1987).Lymphokines 13,209-238. Kelso, A., and Cough, N. M. (1988). Proc. Natl. Acad. Sci. U.S.A. 85,9189-9193. Kelso, A., and Cough, N. M. (1989). Growth Factors 1, 165-177. Kelso, A., and MacDonald, H. R. (1982).J. Exp. Med. 156, 1366-1379. Kelso, A., and Metcalf, D. (1985a). Exp. Hematol. (Copenhagen) 13, 7-15. Kelso, A., and Metcalf, D. (1985b).J. Cell. Physiol. 123, 101-110. Kelso, A., and Owens, T. (1988a). Cell. Immunol. 116,99-I 1 1 . Kelso, A., and Owens, T. (1988b).J. Immunol. 140, 1159-1 167. Kelso, A., Glasebrook, A. L., Kanagawa, O., and Brunner, K. T. (1982). J . Immunol. 129,550-556. Kelso, A., MacDonald, H. R., Smith, K. A., Cerottini, J.-C., and Brunner, K. T. (1984).J . Immunol. 132,2932-2938. Kelso, A., Metcalf, D., and Cough, N. M. (1986).J. Immunol. 136, 1718-1725. Kinashi, T., Harada, N., Severinson, E., Tanabe, T., Sideras, P., Konishi, M., Azuma, C., Tominaga, A., Bergstedt-Linqvist, S., Takahashi, M., Matsuda, F., Yaoita, Y., Takatsu, K., and Honjo, T. (1986).Natu,re (London)324,70-73. Kindler, V., Sappino, A.-P., Grau, G. E., Piguet, P.-F., and Vassalli, P. (1989). Cell (Cambridge, Mass.) 56, 731-740. Kronke, M., Leonard, W. J., Depper, J. M., and Greene, W. C. (1985).J. Exp. Med. 161, 1593-1598. Kupfer, A., Swain, S. L., and Singer, S. J. (1987).J.Exp. Med. 165, 1565-1580. Kurland, J. I., Pelus, L. M., Ralph, P., Bockman, R. S., and Moore, M. A. S. (1979).Proc. Natl. Acad. Sci. U.S.A. 76, 2326-2330. Lang, R. A., Metcalf, D., Cuthbertson, R. A., Lyons, I., Stanley, E., Kelso, A., Kannourakis, G., Williamson, D. J., Klintworth, G. K., Gonda, T. J., and Dunn, A. R. (1987). Cell (Cambridge, Mass.) 51,675-686. Le Beau, M. M., Westbrook, C. A., Diaz, M. O., Larson, R. A., Rowley, J. D., Gasson, J. C., Golde, D. W., and Sherr, C. J. (1986).Science 231,984-987. Lee, F., Yokota, T., Otsuka, T., Meyerson, P., Villaret, D., Coffman, R., Mosmann, T., Rennick, D., Roehm, N., Smith, C., Zlotnik, A., and Arai, K. (1 986). Proc. Natl. Acad. Sci. U.S.A. 83,2061 -2065. Lee, J. S., and Young, I. G . (1989). Genomics 5, 359-362. Lee, J . S., Campbell, H. D., Kozak, C. A., and Young, I. G. (1989). Somatic Cell Mol. Genet. 15, 143-152. Lelchuk, R., Graveley, R., and Liew, F. Y. (1988). Cell. Immunol. 111,66-76. Lichtman, A. H., Chin, J., Schmidt, J. A., and Abbas, A. K. (1988).Proc. Null. Acad. Sci. U.S.A. 85,9699-9703. Lin, H.-S., and Gordon, S. (1979).J. Exp. Med. 150,231-245. Lindemann, A., Riedel, D., Oster, W., Ziegler-Heitbrock, H. W. L., Mertelsmann, R., and Herrmann, F. (1989).J . Clin. Invest. 83, 1308-1312. Lindsten, T., June, C. H., Ledbetter, J. A., Stella, G., and Thompson, C. R. (1989). Science 244,339-343. Lopez, A. F., Nicola, N. A., Burgess, A. W., Metcalf, D., Battye, F. L., Sewell, W. A., and Vadas, M. A. (1983).J. Immunol. 131,2983-2988.
102
A N N E KELSO A N D D O N A L D METCALF
Lopez, A. F., Williamson, D. J., Gamble, J. R., Begley, C. G., Harlan, J. M., Klebanoff, S. J., Waltersdorph, A., Wong, G., Clark, S. C., and Vadas, M. A. (1986).J. Clin. Invest. 78, 1220-1228. Maggi, E., Del Prete, G., Macchia, D., Parronchi, P., Tiri, A., Chrktien, I., Ricci, M., and Romagnani, S. (1988).Eur.1. Immunol. 18, 1045-1050. Maury, C. P. J., and Teppo, A.-M. (1987).J. Exp. Med. 166, 1132-1137. McHeyzer-Williams, M. G., (1989).Eur. J. Immunol., in press. McNeill, T. A. (1973).Nature (London, New Biol.) 244, 175-176. Metcalf, D. (1973).J. Cell. Physiol. 81, 397-410. Metcalf, D. (1980).Proc. Natl. Acad. Sci. U.S.A. 77,5327-5330. Metcalf, D. (1984). “The Hemopoietic Colony Stimulating Factors.” Elsevier, Amsterdam. Metcalf, D. (1985).Blood 65,357-362. Metcalf, D. (1987). Proc. R . SOC.London, Ser. B 230, 389-423. Metcalf, D. (1988).“The Molecular Control of Blood Cells.” Harvard Univ. Press, Cambridge, Massachusetts. Metcalf, D. (1989a). Leukemia 3, 349-355. Metcalf, D. (1989b). Nature (London) 339,27-30. Metcalf, D., and Burgess, A. W. (1982).J. Cell. Physiol. 111,275-283. Metcalf, D., and Merchav, S. (1982).]. Cell. Physiol. 112,411-418. Metcalf, D., and Nicola, N. A. (1985). Leuk. Res. 9,35-50. Metcalf, D., and Nicola, N. A. (1988).Proc. Natl. Acad. Sci. U.S.A. 85,3160-3164. Metcalf, D., Cutler, R. L., and Nicola, N. A. (1983). Blood 61,999-1005. Metcalf, D., Begley, C. G., Johnson, G. R., Nicola, N. A., Lopez, A. F., and Williamson, D. J. (1986a). Blood 68,46-57. Metcalf, D., Burgess, A. W., Johnson, G. R., Nicola, N. A., Nice, E. C., DeLamarter, J., Thatcher, D. R., and Mermod, J.-J. (1986b).j. Cell. Physiol. 128, 42 1-43 1. Metcalf, D., Begley, C. G., Nicola, N. A., and Johnson, G. R. (1987a).Exp. Hematol. (Copenhagen) 15,288-295. Metcalf, D., Begley, C. G., Williamson, D. J., Nice, E. C., De Lamarter, J., Mermod, J.-J., Thatcher, D., and Schmidt, A. (1987b).Exp. Hematol. (Copenhagen) 15, 1-9. Metcalfe, S. (1984). Transplantation 38, 161-164. Mills, G. B., Cheung, R. K., Grinstein, S., and Gelfand, E. (1985).J.Immunol. 134, 243 1-2435. Mills, G. B., Stewart, D. J., Mellors, A., and Gelfand, E. W. (1986).J.Immunol. 136, 3019-3024. Mills, G. B., Girard, P., Grinstein, S., and Gelfand, E. W. (1988).Cell (Cambridge, Mass.) 55,91-100. Miyatake, S . , Otsuka, T., Yokota, T., Lee, F., and Arai, K. (1985). EMBOJ. 4, 2561 -2568. Miyatake, S . , Seiki, M., Yoshida, M., and Arai, K. (1988). Mol. Cell. Biol. 8, 558 1-5587. Moore, R. N., Oppenheim, J. J . , Farrar, J. J., Carter, C. S., Waheed, A., and Shadduck, R. K. (1980).J. Immunol. 125,1302-1305. Morris, A. G., Lin, Y.-L., and Askonas, B. A. (1982).Nature (London) 295, 150152. Morstyn, G., Lieschke, G., Sheridan, W., Layton, J., Cebon,J., and Fox, R. (1989). Semin. Hematol. 26,9- 13. ~
T LYMPHOCYTE-DERIVED CSFS
103
Mosmann, T. R., and Coffman, R. L. (1989). Adv. Immunol. 46, 1 1 1-147. Mosmann, T. R., Cherwinski, H., Bond, M. W., Giedlin, M. A., and Coffman, R. L. (1986).J . Immunol. 136, 2348-2357. Nabel, G., Galli, S. J., Dvorak, A. M., Dvorak, H. F., and Cantor, H. (1981a). Nature (London) 291, 332-334. Nabel, G., Greenberger, J. S., Sakakeeny, M. A., and Cantor, H. (1981b). Proc. Null. Acad. Sci. U.S.A. 78, 1157-1161. Nabel, G. J., Gorka, C., and Baltimore, D. (1988).Proc. Natl. Acad. Sci. U.S.A. 85, 2934-2938. Nicola, N. A. (1989). Annu. Rev. Biochem. 58,45-77. Nicola, N. A., Metcalf, D., Matsumoto, M., and Johnson, G. R. (1983).J . Biol. Chem. 258,9017-9023. Nicola, N. A., Peterson, L., Hilton, D. J., and Metcalf, D. (1988). Growth Factors 1, 41-49. Nimer, S. D., Morita, E. A., Martis, M. J., Wachsman, W., and Gasson, J. C. (1988). Mol. Cell. Biol. 8, 1979-1984. Nimer, S. D., Gasson, J. C., Hu, K., Smalberg, I., Williams, J. L., Chen, I. S. Y., and Rosenblatt, J. D. (1989). Oncogene 4,671-676. Noma, Y., Sideras, P., Naito, T., Bergstedt-Lindquist, S., Azuma, C., Severinson, E., Tanabe, T., Kinashi, T., Matsuda, F., Yaoita, Y., and Honjo, T. (1986). Nature (London) 319,640-646. Nordan, R. P., Pumphrey, J. G., and Rudikoff, S. (1987).J. Immunol. 139,813817. Paliard, X., d e Waal Malefijt, R., Yssel, H., Blanchard, D., Chretien, I., Abrams, J., d e Vries, J., and Spits, H. (l988).J. Immunol. 141,849-855. Parker, J. W., and Metcalf, D. (1974a).Immunology 26, 1039-1049. Parker, J. W., and Metcalf, D. (1974b).J. Immunol. 112, 502-510. Peschel, C., Paul, W. E., Ohara, J., and Green, I. (1987).Blood 70,254-263. Plaut, M., Pierce, J. H., Watson, C. J., Hanley-Hyde, J., Nordan, R. P., and Paul, W. E. (1989).Nature (London) 339,64-67. Poo, W.-J., Conrad, L., and Janeway, C. A. (1988).Nature (London)332,378-380. Powrle, F., and Mason, D. (1988). Immunol. Today 9,274-277. Prystowsky, M. B., Ely, J. M., Beller, D. I., Eisenberg, L., Goldman, J., Goldman, M., Goldwasser, E., Ihle, J,,Quintans, J., Remold, H., Vogel, S. N., and Fitch, F. W. (1982).J. Immunol. 129, 2337-2344. Ralph, P., Broxmeyer, H. E., Moore, M. A. S., and Nakoinz, I. (1978).CancerRes. 38, 1414-1419. Reed, S. G . , Nathan, C. F., Pihl, D. L., Rodricks, P., Shanebeck, K., Conlon, P. J., and Grabstein, K. H. (1987).J. Exp. Med. 166, 1734-1746. Reeves, R., Elton, T. S., Nissen, M. S., Lehn, D., and Johnson, K. R. (1987).Proc. Natl. Acad. Sci. U.S.A. 84, 6531-6535. Rennick, D., Yang, G., Muller-Sieburg, C., Smith, C., Arai, N., Takabe, Y., and Cemmell, L. (1987). Proc. NatZ. Acad. Scz. U.S.A. 84,6889-6893. Roberts, R., Gallagher, J., Spooncer, E., Allen, T. D., Bloomfield, F., and Dexter, T. M. (1988). Nature (London) 332,376-378. Sanderson, C. J., Strath, M., Warren, D. J., O’Garra, A., and Kirkwood, T. B. L. ( 1 985a). Immunology 56,575-584. Sanderson, C. J.. Warren, D. J,, and Strath, M. (1985b).J.Exp. Med. 162,60-74. Sanderson, C. J., O’Garra, A., Warren, D. J., and Klaus, G. G. B. (1986). Proc. Null. Acad. Sci. U.S.A. 83,437-440.
104
ANNE KELSO AND DONALD METCALF
Schrader. J. W., and Clark-Lewis, I. (1982).J . Immunol. 129,30-35. Schrader, J. W., Arnold, B., and Clark-Lewis, I. (1980). Nature (London) 283, 197-199. Schreier, M. H., and Iscove, N. N. (1980).Nature (London) 287,228-230. Shabo, Y., Lotem, J., Rubenstein, M., Revel, M., Clark, S. C., Wolf, S. G., Kamen, R., and Sachs, L. (1988). Blood 72,2070-2073. Shannon, M. F., Gamble, J. R., and Vadas, M. A. (1988). Proc. Natl. Acad. Sci. U.S.A. 85,674-678. Shannon, M. F., Vadas, M. A., and Gamble, J. R. (1989).J. Cell. Biochem. in press. Shaw, G., and Kamen, R. (1986). Cell (Cambridge, Mass.) 46,659-667. Sheridan, J. W., and Metcalf, D. (1972).J. Cell. Physiol. 80, 129-140. Sheridan, J. W., and Metcalf, D. (1973).J . Cell. Physiol. 81, 11-24. Sherr, C. J., Rettenmier, C. W., Sacca, R., Roussel, M. F., Look, A. T., and Stanley, E. R. (1985). Cell (Cambridge, Mass.)41,665-676. Siebenlist, U., Durand, D. B., Bressler, P., Holbrook, N. J., Norris, C. A,, Kamoun, M., Kant, J. A., and Crabtree, G. R. (1986).Mol. Cell. Biol. 6,3042-3049. Snapper, C. M., and Paul, W. E. (1987). Science 236,944-947. Staber, F. G., Hultner, L., Marcucci, F., and Krammer, P. H. (1982). Nature (London) 298,79-82. Stanley, E. R., and Heard, P. M. (1 977).J. Biol. Chem. 252,4305-43 12. Stanley, E. R., Hansen, G., Woodcock, J., and Metcalf, D. (1975). Fed. Proc., Fed. Am. SOC.Exp. Biol. 34, 2272-2278. Stanley, E., Metcalf, D., Sobieszczuk, P., Gough, N. M., and Dunn, A. R. (1985). EMBO J . 4,2569-2573. Strath, M., and Sanderson, C. J. (1986).Exp. Hematol. (Copenhagen) 14, 16-20. Suda, T., Yamaguchi, Y., Suda, J., Miura, Y., Okano, A., and Akiyama, Y. (1988). Exp. Hematol. (Copenhagen) 16,891-895. Sullivan, R., Griffin, J. D., Simons, E. R., Schafer, A. I., Meshulam, T., Fredette, J. P., Maas, A. K., Gadenne, A.-S., Leavitt, J. L., and Melnick, D. A. (1987). J . Immunol. 139,3422-3430. Talmadge, J. E., Tribble, H., Pennington, R., Bowersox, O., Schneider, M. A., Castelli, P., Black, P. L., and Abe, F. (1989). Blood 73, 2093-2103. Taylor, M. V., Metcalfe, J. C., Hesketh, T. R., Smith, G. A,, and Moore, J. P. (1984).Nature (London) 312,462-465. Thorens, B., Mermod, J.-J., and Vassalli, P. (1987). Cell (Cambridge, Mass.) 48, 67 1-679. Tovey, M. G., Content, J., Gresser, I., Gugenheim, J., Balanchard, B., Guymarho, J., Poupart, P., Gigou, M., Shaw, A., and Fiers, W. (1988).J. Immunol. 141, 3106-3110. Troutt, A. B., and Lee, F. (1989).J. Cell. Physiol. 138, 38-44. Truneh, A., Albert, F., Golstein, P., and Schmitt-Verhulst, A,-M. (1985).Nature (London) 313,318-320. Umetsu, D. T., Jabara, H. H., DeKruyff, R. H., Abbas, A. K., Abrams, J. S., and Geha, R. S. (1988).J. Immunol. 140,421 1-4216. Vairo, G., and Hamilton, J. A. (1988).J. Cell. Physiol. 134, 13-24. Valge, V. E., Wong, J. G. P., Datlof, B. M., Sinskey, A. J., and Rao, A. (1988).Cell (Cambridge, Mass.) 55, 101-112. Valtieri, M., Tweardy, D. J., Caracciolo, D., Johnson, K., Mavilio, F., Altmann, S., Santoli, D., and Rovera, G. (1987).J. Immunol. 138,3829-3835. van Leeuwen, B. H., Martinson, M. E., Webb, G. C., and Young, I. G. (1989). Blood 73, 1142-1 148.
T LYMPHCKYTE-DERIVED C:SFS
105
Villalta, F., and Kierszenbaurn, F. (1986).J. Immunol. 137, 1703-1707. Walker, F., Nicola, N. A., Metcalf, D., and Burgess, A. W. (1985).Cell (Cambridge, Mass.) 43, 269-276. Wang, J. M., Colella, S., Allavena, P., and Mantovani, A. (1987).Immunology 60, 439-444. Wang, J. M., Rambaldi, A., Biondi, A., Chen, Z. G., Sanderson, C. J.. and Mantovani, A. (1989). Eur. J. Immunol. 19,701-705. Warren, M. K., and Ralph, P. (1986).J . Immunol. 137, 2281-2285. Watson, J. D. (1983).J. Immunol. 131,293-297. Weaver, C. T., Hawrylowicz, C. M., and Unanue, E. R. (1988). Proc. Null. Acad. Sci. U.S.A. 85,8181-8185. Weisbart, R. H., Kwan, L., Golde, D. W., and Gasson, J. C. (1987). Blood 69, 18-21. Weiser, W. Y., Van Niel, A., Clark, S. C., David, J. R., and Remold, H. G. (1987). J . Exp. Med. 166, 1436-1446. Weiss, A., Imboden, J., Shoback, D., and Stobo, J. (1984). Proc. Null. Acad. Sci. U.S.A. 81, 4169-4173. Whetton, A. D., and Dexter, T. M. (1983).Nature (London) 303,629-631. Wiskocil, R., Weiss, A., Irnboden, J., Kamin-Lewis, R., and Stobo, J. (1985). J . Immunol. 134, 1599-1603. Wodnar-Filipowicz, A., Heusser, C. H., and Moroni, C. (1989). Nalure (London) 339, 150-152. Wr-eschner, D. H., and Rechavi, G. ( I 988). Eur. J . Biochem. 172, 333-340. Yang, Y.-C., Kovacic, S., Kriz, R., Wolf, S., Clark, S. C., Wellems, T. E., Nienhuis, A., and Epstein, N. (1988). Blood 71,958-961. Ymer, S., Tucker, W. Q. J., Sanderson, C. J., Hapel, A. J., Campbell, H. D., and Young, I. G. (1985). Nalure (London) 317,255-258. Yokota, T., Lee, F., Rennick, D., Hall, C., Arai, N., Mosrnann, T., Nabel, G., Cantor, H., and Arai, K. (1984). Proc. Natl. Acad. Sci. U.S.A. 81, 1070-1074. Yokota, T., Arai, N., d e Vries, J., Spits, H., Banchereau, J., Zlotnik, A., Rennick, D., Howard, M., Takebe, Y . , Miyatake, S., Lee, F., and Arai, K. (1988). Immunol. Rev. 102, 137-187. Zilberstein, A., Ruggieri, R., Korn, J. H., and Revel, M. (1986). EMBO J. 5, 2529-2537. Zlotnik, A., Roberts, W. K., Vasil, A., Blumenthal, E., Larosa, F., Leibson, H. J., Endres, R. O., Graham, S. D., White, J., Hill, J,, Henson, P., Klein, J. R., Bevan, M. J., Marrack, P., and Kappler, J. W. (1983). J. Immunol. 131, 794-800. This article was accepted for publication on 4 October 1989.
ADVANCES IN IMMUNOLOGY, VOL. 48
Molecular Basis of Human leukocyte Antigen Class II Disease Associations DOMlNlQUE CHARRON labomtoire d’lmmunog6n&ique Mol6culaim, Institut Biom6dical des Codelien, and CHU Piti6 SalpBtrih, Univenit6 Paris Vl, Paris, Fmnce
1. Introduction
Current knowledge of the structures and biological functions of the human leukocyte antigen (HLA) system provides a framework for unraveling the mechanisms of one of the most puzzling enigmas of modern biology and medicine, which consists of the definite associations among components of the system and genetic susceptibility o r resistance to disease (Dausset, 1981). Although known for over 25 years, this phenomenon is only now beginning to be understood in molecular terms. Indeed, recent immunological, biochemical, and molecular biological developments have led to precise understanding of the structures and functions of the human major histocompatibility complex (MHC). This knowledge is prerequisite in evaluating the genetic and somatic aspects of the role of the HLA complex in the physiology of the immune response and their contribution to autoimmunity. T h e most significant initial observation has been an increase in the frequency of certain serologically defined alleles of the HLA complex in several autoimmune diseases. During the past 15 years fastidious attempts have been undertaken to improve the identification of the most significant HLA components involved in genetic associations. However, since the HLA system displays the unusual feature of a strong linkage disequilibrium between loci and alleles, the genetic traits found to be associated with disease did not emerge at random. The pattern of genetic associations followed an almost constant trend. The associations gained strength each time an additional locus centromeric to the precedent was individualized. Th e advances made in this respect almost paralleled the introduction of progressively more refined typing procedures, which allowed the division of former genetic entities (loci and alleles) into additional subtypes. This is best illustrated in the case of insulin-dependent diabetes mellitus (IDDM), which was initially found to be associated with alleles 107 Copyright 8 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
108
DOMINIQUE CHARRON
within the A, then the B, locus and finally the HLA-D region. Among the HLA-associated diseases, or at least for those diseases in which an autoimmune process is suspected to be directly relevant to the pathogenesis, the associations are with genes and molecules of the HLA-D region (HLA class I1 genes and products). The most recent data assign the disease susceptibility to common amino acid sequences present on a HLA class I1 molecule within its “active” site. It can be postulated that a specific antigen (or set of antigens), presumably in the form of short peptides, bind with exquisite affinity to the precise structure of the class I1 molecules associated with a particular disease. This would lead to inappropriate antigen presentation and T cell activation, resulting in an impaired immune response. However, the exact nature of the antigens (foreign or self) remains enigmatic as the potential influence of the T cell receptors (TcR) repertoires. Present knowledge is still insufficient to construct a definitive representation of the trimolecular structures presumably involved in controlling the immune response (MHC-antigen-TcR). Moreover, with regard to disease susceptibility, one cannot discriminate between two possibilities. Is the disease susceptibility associated with one or a few contiguous specific linearly displayed amino acids, or with several amino acids (or sets of amino acids) distantly spaced in the molecule, creating a unique three-dimensional conformation? These two hypotheses have practical implications for selecting the most pertinent typing approach in studies of disease susceptibility. Indeed, in the first case identification of the crucial amino acids (or stretch of amino acids) should suffice, while in the second case a three-dimensional structural model ultimately would be required to identify the exact composition of the disease susceptibility elements. This chapter focuses strictly on the HLA MHC class I1 genes and molecules with regard to how they contribute to better delineation of the genetic associations and how the current knowledge of their structure, expression, and functions can be used to speculate on their role in the pathogenesis of disease. Because of the strong linkage disequilibrium between loci and alleles, I feeljustified in restricting the description of the genetic associations to only the most recent data (mainly generated by molecular means), since they supercede in precision and accuracy the previous data obtained by serological methods. Furthermore, emerging studies of gene regulation and expression provide information concerning the function of these genes and molecules. Such approaches are intended to identify which HLA molecules confer the highest susceptibility o r resistance to a disease and how they contribute to disease development. I discuss the central contribution of three-dimensional conformational structures of the HLA class I1 mole-
H LA CLASS
I I DISEASE ASSOCIATIONS
109
cules and emphasize in particular the possibilities generated by the formation of hybrid HLA class I1 molecules, the prominent role of which has been highlighted by the most recent epidemiological studies of disease susceptibility. In addition, such genetic and molecular approaches have practical implications in predictive medicine (i.e., identifying individuals at risk for a disease) and may provide new rationales for therapeutic intervention. It. HLA Class It Structures: Genes and Molecules
Class I1 molecules have a central immunoregulatory role. They are cell surface glycoproteins expressed principally on macrophages, monocytes, and B lymphocytes, which present processed antigens (in the form of peptides) to antigen-specific T lymphocytes, leading to their specific activation and proliferation. Such T cell features are implicated in the development of autoimmunity. A brief description of the class I1 genes and molecules is required in order to discuss their role in disease susceptibility (Fig. 1). T h e human MHC class I1 genes are located on a segment of chromosome 6 and consist of several sets of a-/3 species, for a total of 14-15 individual genes per haplotype. Three major subregions-namely, DR, DQ, and DP-are identified by their cell surface-expressed products, consisting of one a-fl dimer each for DQ and DP and one o r two concomitantly expressed a-P dimers for DR, depending on the haplotype (e.g., DRA-DRB1 and DRA-DRB3 or DRA-DRB4) (Korman et al., 1985; Kappes and Strominger, 1988). In addition to the expressed a-/3 genes, MHC class I1 pseudogenes (e.g., DRB2, DPA2, and DPB'L),nonexpressible genes (e.g., DQA2 and DQB2), and weakly expressed genes (eg., DNA and DOB) are found (Trowsdale and Kelly, 1985; Tonnelle et al., 1985). Overall, MHC class I1 genes are characterized by two apparently opposing structural features, both of which are critical to our understanding of the system. First, the HLA class I1 genes and proteins are homologous to each other (as well as among species). Second, most of the class I1 genes are highly polymorphic at the population level. Therefore, gene duplication, gene conversion, and intralocus double crossings-over are the most likely mechanisms to have generated the present isotypic and allelic diversities (Auffray et al., 1984; Gustafsson et al., 1984; Gorski and Mach, 1986; Anderson et al., 1987).Together, these facts explain several features of the observed polymorphism. Most of the class I1 polymorphism is located at the first domain of the molecule and appears to be clustered into t w o to four hypervariable regions. These structural variations between the class I1 gene products are the
HLA
CLASS I1 DISEASE ASSOCIATIONS
111
fundamental differences permitting immune recognition and are therefore likely to be critical in disease susceptibility. The crystal structure of one HLA class I molecule has recently been obtained, and, because of the overall similarity in amino acid composition and general secondary conformation of classes I and I1 molecules, it was subsequently used to create a three-dimensional model for class I1 molecules (Bjorkman et al., 1987; Brown et al., 1988). Although hypothetical, such a model provides a structural framework for understanding the function of an MHC class I1 molecule. An antigen recognition site can be proposed which is composed of an internal cavity lined by two a-helical structures and closed at the bottom by a platform of eight @-pleatedsheet structures. When the model is compared with a bulk of published and unpublished serological, functional, and genetic data, residues which are monoclonal antibody epitopes (i.e., protruding into the solvent) or class II-restricted T cell recognition epitopes (i.e., facing into the site o r upward) can be predicted. Thus, a model is provided for the binding of foreign or self-peptides and for the identification of the putative contact sites between the peptides and the class I1 residues. T h e amino acid stretches of class I1 allelic variability previously identified can then be positioned in this three-dimensional model to delineate the components of the active sites of the molecules. For example, along the DR @ lchain the polymorphic residues are part of the @-pleatedsheet FIG. 1. The class I1 genes of the HLA system are clustered on the short arm of chromosome 6 into three subregions: DR, DQ, and DP. Overall, the genes are organized into exons coding for domains of the protein and noncoding introns. A class I1 molecule is a glycoprotein dimer composed of a transmembrane a chain coded for by an A gene, noncovalently associated with a transmembrane p chain coded for by a B gene. The DR subregion contains a nonpolymorphic DR A gene and three or four DR B genes, depending on the haplotype. The DRB 1 genes are highly polymorphic and correspond to the DR1-DRwl8 specificities. DRB2 is a pseudogene. The former are associated with either a DRB3 gene (in haplotypes 3, 5 , w6, and wS), which is moderately polymorphic and corresponds to the specificity DRw52, or a DRB4 gene (in haplotypes 4, 7, and w9), which is monomorphic and corresponds to the specificity DRw53. The DQ molecules are a-p dimers composed of a DQ a chain coded for by a DQA 1 gene and a DQP chain coded for by a DQB 1 gene. Both are highly polymorphic. The DQA2 and DQBP (formerly DX a and DX p) genes are homologous to the precedent, but are not expressed. The DQwl-DQw9 specificities reflect the DQ P chain. The DP subregion contains two expressed genes: DPAl and DPB1. DPAl has only two known alleles. DPBl polymorphism is more extensive and corresponds to the DPwl-DPw6 cellularly defined subtypes. DP a and p chains associate in a DPa-p dimer. Additional genes are present which are not regularly expressed or are expressed at low levels.
112
DOMINIQUE CHAKKON
platform for the two first hypervariable regions (amino acids 9- 13 and 25-33; HVRl and HVR2, respectively), while the major HVR region (HVR3) from amino acids 57 to 76 (extending to amino acid 85 and centered on amino acids 67-74) is located on turns of the a-helical structure. In the class I1 molecules both the a 1 and /3l domains participate in the putative antigen recognition site. Indeed, an intermolecular dimerization must be postulated in the case of the HLA class I1 model, in order to be consistent with the three-dimensional structural pattern of the class I molecule. This dimerization is likely to be formed by a disulfide bridge between the two spatially close amino acids at positions 78 and 15, respectively, on the cr and /3 chains. 111. HLA Class I1 Polymorphism and Typing
This section evaluates the contribution and potential of the different typing procedures used to identify disease susceptibility factors. A. SEROLOGY
Serologically defined HLA class I1 specificities consist of conformational motifs (or epitopes) present on at least one set of HLA class I1 molecules expressed at the surface of the B lymphocyte population and recognized by allospecific antibodies (monoclonal or polyclonal) (Fig. 2). Such discrete epitopes can be unique [monoclonal antibody (mAb) binding] or grouped (alloantibodies)and expressed on the same molecule; an epitope can eventually be found on different types of molecules (belonging to the same isotype or to several different isotypes). In the latter case the epitope represents a shared determinant between different alleles or different isotypes (i.e., interallelic or interisotypic determinant, respectively) and is termed “supertypic.” Serological typing can thus identify the presence of an epitope, but cannot directly assess the molecular species on which it is present. This can be deduced from immunoprecipitation or gene transfection studies which would, in addition, allow gene assignment of the epitope. Examples of supertypic epitopes are numerous; DR2 and DR4 specificities are found on several different HLA molecules (Lea,interallelic epitopes) (Wu et al., 1986; G. T. Nepom et al., 1983). DRw52 can be found on different alleles of the same gene (DRBS) but also on a different gene product (DRB1) (Haziot et al., 1986). Similarly, DRw53 is found on either DRB 1 or DRB2 (Matsuyama et al., 1988). If one excludes the rather unlikely case of an alloantibody recognizing by chance a genuine disease susceptibility epitope, serological typing reflects a complex situation of linkage disequilibrium. Indeed, the epi-
a
D Dwl Dw2 Dw3 Dw4 Dw5 Dw6 Dw7 Dw8 Dw9 DwlO Dwl l(w7) Dw12 Dw13 Dw14 Dw1S Dw16 Dw17(w7) DwlS(w6) Dw19(w6) Dw20 Dw21 Dw22 Dw23 Dw24 Dw2S Dw26
DR DRl DR2 DR3 DR4 DRS DRw6 DR7 DRw8 DR9 DRwlO DRwll (S) DRw12(S) DRw13(w6) DRw14(w6) DRwlS(2) DRw16(2) DRwl70) DRwlS(3) DRw52 DRwS3
DQ DQWl
DQw2 DQw3 DQw4 DQW~W
DQw6W) DQW7CW3) DQWW3) DQw9(w3)
DP DPwl DPw2 DPw3 DPw4 DPwS DPw6
b
DR2 DR3 DRS DRw6 DQWl
DQw3 Dw6 Dw7
u
HLA-D spccificitics Dwl. DwU) DWZ, Dwl2 DwZl ,DW22 Dw3 Dw4, Dw10, Dw13. Dw14. Dw15
DWS
Dw6, Dw18. Dw19 Dw9, Dw16 Dw7. Dwll. Dw17 Dw8 DwW Dw24. Dw25, Dw26
DRwlS, DRwl6 DRwl7. DRwl8 DRwll. DRwl2 DRwl3, DRwl4 W w 5 , DQw6 DQw7. DQW8, DQw9 Dw18, Dw19 Dwll. Dw17
AssaciatedDR Sp~citics DRl DRwlS(2) DRw 16(2) DR3 DR4 DRw 1l(5) DRw13(w6) DRw 14(w6) DR7 DRw8 DR9 drw52
FIG. 2. (a) Listing of the HLA-D region specificities, (b) listing of the new subdivisions (splits), and (c) correlation between HLA DR specificities and HLA D typing. (Data obtained from the HLA Nomenclature Committee.)
114
DOMINIQUE CHAKKON
tope recognized by an antibody can be present on the class I1 molecule involved in the susceptibility, while also present on a distinct molecule irrelevant to the disease. Assignment of the antibody specificity to disease susceptibility could only be allowed in the case of recombination between two genes. Such recombination is rarely found within a homogeneous population and is more frequently found between different ethnic groups. This explains the interest in studying HLA and disease associations in these groups. Since there are only 20 officially recognized HLA-DR typing serological specificities (HLA nomenclature) and the structural variation accounts for over 50 alleles, most of the serologically defined specificities are likely to be supertypic. Moreover, since serological reagents were selected through allorecognition, they are not necessarily suitable for detecting disease susceptibility epitopes. HLA serological typing thus may presently represent a rather incomplete approach to typing for HLA disease associations. New serological reagents are being developed to circumvent these difficulties. HLA class I1 transfectants or transgenics can be used to produce specific antiisotype and/or antiallele antibodies. A series of even more refined approaches will be undertaken when a precise sequence of the class I1 molecule is identified as the disease susceptibility element. Site-directed mutagenesis of the class I1 molecule or immunization with synthetic peptides should generate antibodies uniquely directed toward discrete sequences of the class I1 product (i.e., disease-specific epitope) (Atar et al., 1989). This approach would not only be invaluable for epidemiological studies, but should be ultimately considered for epitope-specific therapy. This approach is particularly pertinent in the case of hybrid determinants (i.e., determinants created by the formation of particular a-P heterodimers, as discussed below) for which no serological reagent is available. Apart from serology for the detection of cell surface-expressed (thus, conformational) epitopes, several molecular methods have been developed and have been widely used in HLA typing over the past 10 years. They include analysis of the polymorphism at either the protein or DNA level and therefore largely reflect linear amino acid or nucleic acid sequences. My purpose is not to review these approaches in detail, but to highlight some of their characteristics and the implications for HLA and disease association.
B. BIOCHEMISTRY Biochemical typing, which includes one-dimensional isoelectrofocusing (IEF) and two-dimensional polyacrylamide gel electrophoresis (PAGE), usually requires radioactive cell labeling and immunoprecipita-
HLA
115
CLASS I1 DISEASE ASS0C:IAI‘IONS
tion with specific mAb’s), steps which are both time consuming and expensive. Two-dimensional PAGE at least has proven to be highly efficient in detecting new HLA class I1 variants (Fig. 3). Furthermore, it represents the best technique for definition of the reactivity of mAb’s in association, eventually, with immunodetection by Western blotting of their individual chains or complex reactivity. The first dimension (of a two-dimensional PAGE) can be nonequilibrium pH gradient electrophoresis o r IEF, which provides better resolution of the class I1 fl and a chains, respectively. Since our initial description of the DR fl chain polymorphism, two-dimensional PAGE has been widely used to identify molecular variants belonging to DR2, DR4, DR5, DRw6, DR7, and DRw52.
b-
0 0
0
.a
Dw15
am
0
+ .W
3.1
.W
3.2
(normal)
OQU 3 . 3
IEF
FIG. 3. Two-dimensional gel electrophoresis analysis of HLA class I1 molecules. Schematic representation of (a) DR pl chain variants in DR4 individuals, (b) DQ p chain variants in DQw3 individuals, and (c) DR, DQ, and DP (Y chains of the different haplotypes. IDDM, Insulin-dependent diabetes mellitus; IEF, isoelectrofocusing.
116
D o M I N l Q U E CIIAKRON
So far, this is one of the best typing procedures for the identification of DQ a and DP a variants, with eight and two alleles, respectively (Charron and McDevitt, 1979, 1980; Knowles, 1989; Charron, 1990; Charron and Fernandez, 1990). The theoretical limitation is that in some instances conservative amino acid substitution will not be detected electrophoretically and thus will not be resolved by two-dimensional PAGE. In practice, this appears to be very rare. Further, in the case in which the initial (i.e., most basic) location of two different alleles is electrophoretically indistinguishable, the more processed forms (more acidic locations) are distinct. This reflects the fact that, although the net charges of two molecules may be similar, the type and the variety of amino acids which compose the molecule do, in fact, slightly affect their electrophoretic migration. This is clearly the case when the DR /3l chains from DR1 and DRw9 cells are analyzed (R. Fauchet and D. Charron, unpublished observations). In contrast, limited substitution (e.g., point mutation) at the genomic level may induce drastic changes in the net charge and variants, which are silent by genomic analysis [restriction fragment-length polymorphism (RFLP)] but are readily distinct by two-dimensional PAGE as in the case of DR-BON versus DRBl gene products (Coppin et al., 1987). Biochemical typing has a unique potential of importance in disease association, since it is at present the only typing technique which can detect hybrid molecules, such as intraisotypic hybrid molecules, obtained by trans-association and interisotypic association from &-and/or transcomplementation. (Charron et al., 1984; Lotteau et al., 198713). Although the detection of such hybrid molecules requires rare reagents (such as chain-haplotype-, or allele-specific mAbs), it is a critical step in studying HLA and disease association, since several epidemiological studies have documented the high incidence of such hybrid molecules in certain diseases (Nepom et al., 1984; Ronningen et al., 1989). The use of silver staining instead of radioactivity in two-dimensional PAGE and the development of Western blotting and IEF may provide more simple ways to perform biochemical typing for HLA class I1 in the near future (Hermans et al., 1989a; Rodriquez de Cordoba et al., 1989). Apart from the genetic polymorphism, biochemical studies of HLA class I1 molecules are also of interest in the study of the diversity of expression of class I1 molecules in cells of the diseased organs in which they may appear in an aberrant manner. Protein analyses are thus complementary to studies at the transcription level performed by Northern blot, run-on/ run-off, and S 1 nuclease protection assays. In brief, the biochemical approach encompasses the genetic and somatic aspects of HLA class 11, both of which are pertinent to the question of HLA and disease association.
HLA CLASS 11 DISEASE
ASSOCIATIONS
117
C. DNA (RFLP/ALLELE-SPECIFIC OLIGONUCLEOTIDES)
1. RFLP T h e combination of restriction endonucleases and specific probes enables one to study allelic polymorphism arising from genomic neucleotide sequence variations in both coding and noncoding regions of a gene. RFLP analysis is based on length variation of fragments, resulting from the cleavage of specific genomic DNA at polymorphic restriction sites, as revealed by Southern blotting. The use of RFLP became popular a few years ago, in the HLA field in particular, for the study of HLA class I1 and disease association (Cohen et al., 1985; Bell et al., 1985).The potential value of RFLP analysis in genotyping class I1 alleles was tested as soon as specific probes became available, and the method has since been developed and extensively refined, in particular during the Tenth International Histocompatibility Workshop (Dupont, 1989). A great contribution was made when it became possible to perform the typing of cells not expressing HLA (e.g., leukemic cells and cells from Bare lymphocyte syndrome) (Marcadet et al., 1985a). However, it was hoped that this genomic approach would identify new genetic markers uniquely associated with and specific for a disease. Indeed, had these markers existed, this approach may have localized in the HLA region non-HLA disease susceptibility genes. In such a case the previous association with certain HLA alleles detected serologically would have had to be interpreted as part of the strong linkage disequilibrium which is a constant feature of this gene segment. A few pitfalls emerged in using RFLP analysis, because of the presence of several HLA class I1 pseudogenes and inter- and intraloci nucleotide sequence homology. Cross-hybridization rendered problematic interpretation of the early RFLP studies using full-length probes. RFLP patterns reflected multiple hybridization signals, some being cross-hybridization. In general, it is the strong linkage disequilibrium among polymorphic restriction sites (a large number are present in noncoding region) and coding sequence variations that are highly informative in RFLP studies. Improvement came from the use of shorter and fewer cross-hybridizing probes “exon specific,” which can be considered operationally as locus o r gene specific (DR p, DP a, and DP p). For DQ a+, although high sequence homology with DX a+ results in cross-hybridization, the low polymorphism of the DX a-P genes enables identification of the relevant fragments and interpretation of the polymorphism as specific for the DQ genes. Integrated DNA RFLP typing strategies and procedures have been developed (Bidwell et al., 1988; Cohen, 1989) which are convenient tools for the investigation of HLA polymorphism in large population samples.
118
DOMINIQUE CHAKRON
However, these procedures do not provide any advantage with regard to definition and have some drawbacks. Since both haplotypes may share one or more allelic patterns, the identification of heterozygosity can be complicated. Moreover, some alleles may not be adequately resolved. This is illustrated in the cases of DR-BON versus DR1 (Coppin et al., 1987) and the splits of DR4 p chains (i.e., Dw4, -10, -13, -14, and -15), which are of great importance in rheumatoid arthritis (RA) studies (Gregersen et al., 1986). The development of nonradioactive detection may provide a means of introducing this technique into additional typing laboratories (Erlich et al., 1986; Saiki et al., 1985). Moreover, interest is increasing in studying the regulation of individual HLA class I1 gene expression which may affect disease states. It is at present unclear whether regulatory sequences (5’ upstream promoters and enhancers) have any genetic variability. If this is the case, RFLP analysis may provide a means to identify and characterize these regulatory polymorphisms and to study their epidemiology in diseased and control populations.
2. Allele-Spec$c Oligonucleotides (ASO) AS0 is an additional molecular technique which has been introduced to circumvent the inability of cDNA probes to identify some HLA class I1 splits, particularly the different DR4 Dw haplotypes. This technique was rapidly combined with the polymerase chain reaction procedure, which greatly facilitates the obtaining of specific DNA sequences suitable for further analysis (by RFLP or ASO) (Saiki et al., 1985, 1986). From the known nucleotide sequences of the individual DR, DQ, and DP a and p genes of the different haplotypes, it is feasible to select short sequences specific for one (or for a group of) alleles. These allele-specific oligonucleotides are synthesized and used in combination with a radioactive (or nonradioactive) detection system. Hybridization of AS0 following specific amplification by thermal cycles (polymerasechain reaction) of the hypervariable exons represents an elegant technique for the identification of a given short linear class I1 sequence within the genome. Ultimately, the comparison of sequences of every class I1 gene (a+) from normal versus diseased cells will be attained. This will undoubtedly focus interest on relevant amino acids and stretches of amino acids which are important for conferring high susceptibility to a particular disease. However, considering the threedimensional structure and recent functional data, it is likely that additional information will be revealed by the study of conformational parameters relevant to the identification of disease susceptibility factors. Since HLA and disease association may finally reflect specific immune
HLA CLASS I1
DISEASE ASSOCIATIONS
119
recognition of an antigen presented to a T cell by a class I1 molecule, the use of T cell clones would permit identification of the disease-susceptible versus non-susceptible individuals. The suggested conformational nature of the disease susceptibility factor predicts that it is likely to be composed of one or several amino acids or a series of amino acids localized on one class I1 chain (aor p) or on the two chains (aand p) of a particular class I1 dimer. This arouses interest in the role of the determinants created when hybrid HLA class I1 molecules are formed. Hybrid epitopes are consistent with the overwhelming epidemiological data showing a heterozygous (or haplotypic) effect in several diseases (Nepom et al., 1984; Ronningen et al., 1989), as initially suggested by Svejgaard et al. (1983) in IDDM. IV, HIA Class II Hybrid (Intra- or Interisotypic) Molecules
Cell surface expression of the class I1 molecule as a stable heterodimer appears to be a logical requirement for immunological function and is likely to play a major role in disease association. The expressed repertoire of the class I1 molecules is the combined result of allelic and isotypic variations. Thus, particularly for disease associations, an appreciation of the full repertoire of class I1 dimers has to be evaluated. In addition to the number of class I1 a-P loci combined two by two within a haplotype to form isotypes, a powerful way to increase class I1 antigen diversity would be the association of the a and p chains of the two haplotypes by tramcomplementation within an isotype. These molecules are found only in heterozygous individuals. An even more efficient way of generating additional diversity would be to pair a and p chains from different isotypes either in cis and/or tram. Here, I summarize the data on hybrid molecules and explore their potential for the understanding of associations between HLA and disease.
MOLECULES A. HLA-DQ HYBRID T h e HLA-DQ products have the potential of forming hybrid molecules, which are dimers created by a-/3 chain pairing resulting from gene trans-complementation (Fig. 4). Such molecules include chains belonging to both the paternal and maternal haplotypes. Indeed, the DQ a and p chains are both highly polymorphic (Goyert and Silver, 1983; Charron et al., 1984; Giles et al., 1985). T h e use of allele-specific mAb’s directed uniquely at the DQ a or P chain resulted in the isolation of hybrid (DQ) molecules. The use of allospecific sera and recombinant strains led to the initial description of such molecules in F1 mice (Jones et al., 1978; Silver et al., 1980).
120
DoMINlQUE CHAHKON
Gene
DQg
Gene
DQa
paternal haplotype
maternal haplotype
HU-DQ hybrid molecules
FIG. 4. Tram-Complementationleading to the formation of HLA-DQ hybrid molecules.
Our own approach included the use of two-dimensional PAGE to detect and identify electrophoretic variants of DQ molecules. A DQw2 (DR7)p-DQwl (DR1) a hybrid molecule was the first to be demonstrated in humans in a DR1-DR7 heterozygous cell line (Charron et al., 1984). Experiments with '251 surface-labeled antigens demonstrated that these hybrid molecules were also expressed at the cell surface. T h e lack of adequate reagents explains the limited work that has been published since our first demonstration of hybrid DQ a-DQ p molecules in heterozygous individuals. Giles et al. (1985) found the presence of a DQw3 P-DQw2 a dimer in a DR5-DR7 cell. Using the same anti-DQw3 mAb, Nepom et al. (1987) were able to identify, by two-dimensional PAGE and peptide mapping, a DQw3 P-DQw2 a chain complex in DR3-DR4 normal individuals. Moreover, the same DQ a (DQw2)-DQ /3 (DQw3) dimer was present at a comparable level in DR3-DR4 individuals, whether or not they had IDDM. Recent results using transfection suggest that all four DQ heterodimers are present in a DQw2-DQw3 cell (G. T. Nepom, personal communication). The ability to form hybrid DQ molecules by gene trans-complementation should not be surprising (Fig. 4). Indeed, when D Q a and p
HLA CLASS 11 DISEASE ASSOC:IAI‘IONS
121
alleles from distinct haplotypes are compared, a series of “naturally occurring” hybrid DQ molecules are found which are identical in their a and p chain composition to the hybrid molecules found in heterozygotes, but differ only in that their a and p genes are coded on the same haplotype (cis-configuration). Recent data at the nucleotide level described DQ dimers in DR7-DQw2 and DR7-DQw3 haplotypes sharing a DQ a chain, but differing in their DQ p chains (Song et al., 1987). Although the sequences provided are restricted to the first domain of the a and p chains, the data suggest that the DR7-DQw3 haplotypes encode in cis a natural hybrid molecule formed by a DQw3 a chain originating from a DR4 haplotype and the typical DQw2 p chain present in every DR7-DQw2 haplotype. Additional data were recently obtained on a larger panel of homozygous typing cells in which DQ a and p polymorphism was studied at the protein level by IEF analysis (Bontrop et al., 1987). Many, but not all, of the possible combinations of a and p chains were found. Thirty hypothetical dimers which could arise as a result of the various combinations between five DQ a and six DQ p chains were considered, but only 20 dimers were observed experimentally. Whether the absence of a given a-/3 combination is due to the inability to associate (i.e., forbidden a-p pairing) or whether the population studies are, as yet, insufficiently extensive to include all possible combinations is unresolved. Indeed, because of strong linkage disequilibrium between alleles of the DR and DQ loci, certain combinations of alleles predominate in a given population. Different combinations of genes which would generate czsderived hybrid molecules are most likely to be found in different populations. Natural hybrid molecules thus represent minor variant haplotypes. It is likely that these “natural” hybrid DQ molecules evolved by recombination between two distinct (heterozygous) haplotypes and were then fixed. This type of data argues in favor of a recombination “hot spot” between DQ a and p chains (Song et al., 1987) and justifies a search for sequences facilitating recombination events, as have been described in the mouse I region (Steinmetz et al., 1986; Smith et al., 1981). The demonstration in heterozygous individuals of the cell surface expression of hybrid HLA-DQ molecules raises several questions. Although there is no convincing report of alloantisera behaving as if they could recognize hybrid molecules, it may well be that such sera have been overlooked by serologists. Indeed, such sera would only recognize heterozygous individuals and would not segregate with a given HLA haplotype in family studies. Alternatively, it may be that these determinants are less immunogenic in terms of alloantibody response. Transfectants containing hybrid molecules will provide an exquisite tool for the production and screening of hybrid-specific antibodies which
122
DOMINIQUE CHARRON
can be subsequently used in epidemiological studies. While in the mouse the Ir gene tram-complementation phenomenon is unequivocally explained on the basis of hybrid Ia molecules, among the mixedlymphocyte reaction-stimulating determinants there appear to be some that represent conformational epitopes created by the trans-association of a and p chains (Fathman, 1980). Few functional data are available from humans. T cell clones have been found to be restricted by such hybrid molecules (Hansen et al., 1982). However, modification of the cloning procedure has recently demonstrated several clones which may reveal hybrid epitopes involved in celiac disease (CD) (Lundin et al., 1989). The case for a role for hybrid HLA-DQ determinants in disease susceptibility requires particular attention. Indeed, in numerous epidemiological studies, first in IDDM (Svejgaard et al., 1980) and later injuvenile RA (Nepom et al., 1984) and in CD (Betuel et al., 1980), an unexpectedly high incidence of the disease has been observed in particular combinations of DR haplotypes (i.e., heterozygous effect) (Charron, 1986). As a consequence, the relative risk is dramatically higher in some heterozygous situations than for any unique allele, even when present in a homozygous state. Such data cannot be explained by the model of monoallelic association. If HLA class I1 molecules are involved, it becomes logical to propose that the particular determinant derives from a combination of products, one from the first haplotype, the other from the second haplotype. Because of the lack of polymorphism in the DR (Y chain, it is unlikely that HLA-DR molecules themselves would be the structural basis of this heterozygous effect. The DQ loci fulfill the two requirements of being in strong linkage disequilibrium with DR alleles and having structurally polymorphic subunits. trans-association of DQ a-/I chains creates hybrid molecules which may be consistent with the observed heterozygous effect, since only these HLA-DQ hybrid molecules will bear conformational epitopes unique to the combination of paternal and maternal haplotypes. These possibilities are emphasized in the discussion concerning IDDM and CD. The expression of hybrid HLA-DQ molecules thus becomes an important parameter to consider. Recent work in the mouse using appropriate combinations of transfected genes has shown that, while haplotype-matched A a-p genes (haplotypes k,b, d) resulted in the optimum cell surface expression of the dimers (as one would expect from the biochemical studies previously conducted in normal cells), the level of expression of haplotype-mismatched A a-/I dimers was extremely variable and depended on the combination of haplotypes used. As an example, A ak-pbtransfectants had poor surface Ia expression, while A ak-pd transfectants had no detectable Ia at their
HLA CLASS 11 DISEASE
ASSOCIATIONS
123
surface. Furthermore, it was demonstrated that appropriate combinations of polymorphic sequences in the NHY-terminal half of the A a and /3 chains could control the pairing. This allele-specific control of Ia molecule surface expression was rather unexpected. It is, however, consistent with most functional and evolutionary features of class I1 genes (Braunstein and Germain, 1987; Sant et al., 1987). T h e inability of certain DQ a and /3 chains to form functional dimers has yet to be demonstrated in humans. Although some DQ a and /3 genes have never been found to occur in cis in the population, this may not preclude the possibility that they occur in rare cases o r even in trans (Bontrop et al., 1987). It is still not known whether the pairing efficiency is the same for DQ a-p gene products coded in cis versus in trans. Rules underlying the formation of DQ a-/3 pairs are still enigmatic, and data are not yet available on which part@) of the a-p chain is important in order for the dimer to be correctly assembled and transported to the cell surface. Mechanisms similar to those described in the mouse system are expected to be found in humans. However, several other factors have not yet been investigated which could interfere with hybrid molecule formation. These include numerous posttranslational processing events and quantitative regulatory mechanisms, some of which are noticeably different in mice and humans (Nee1 et al., 1987). These phenomena could affect gene repertoire and, subsequently, disease association. Finally, the possibility for other homologous hybrid molecules should be envisaged. While the total lack of structural polymorphism in the DR a gene and protein means that there is no difference in whether the constitutive a chain of a DR a-/3 dimer is coded for by either parental chromosome, other class I1 subsets are potential candidates for generating intraisotypic hybrid molecules. This is clearly a possibility for the DP subset, both subunits of which are structurally polymorphic (Ando et al., 1986; Lotteau et al., 1987a). Since several structurally distinct DP /3 chains can pair with the same DP a chain, this provides another example of naturally occurring hybrid molecules. Whether the two DP a chains can pair indiscriminately in heterozygous cells with any DP /3 chain is unknown. B. INTERISOTYPIC HYBRID HLA CLASSI1 MOLECULES It has been suggested that, due to a presumed higher affinity between the a and /3 chains when they are coded for by loci of the same isotype, the expression of class I1 dimers will occur only within one isotype. Isotypemismatched molecules were therefore considered forbidden pairs and were not expected to be found at the surface of normal cells (Travers et al., 1984). A series of laboratory experiments using normal Epstein-Barr
124
DOMINIQUE CHARRON
virus-transformed human B cell lines provides direct evidence for the assembly and the cell surface expression of mixed isotypes consisting of DR a and DQ p chains. Quantitation of the isotype-mismatched dimers versus the conventional class I1 dimers using immunoprecipitation of labeled cell lysates is not precise. It appears, nonetheless, that DR a is associated with DQ /?than with DR p and, similarly, less DQ p is associated with DR a than with DQ a. In the mouse transfection of Z-E and Z-A genes into L cells has recently demonstrated the unexpected formation of similar isotype pairs consisting of I-E a and I-A p chains (Germain and Quill, 1986; Malissen et al., 1986; Germain and Malissen, 1986). This unorthodox pairing appears to be predominantly influenced by the allelic polymorphism of the I-A /3 chains, since the I-A /I-I-E a complex could be detected as an I-E a-I-A pddimer, but not as an I-E a-1-A pkor I-A pbdimer. Furthermore, this work emphasized the unexpected influence of the polymorphic NH2terminal domain of the I-A p molecule in permitting pairing, since the I-E a chain is virtually nonpolymorphic. It should be remembered that, in the case of transfected genes, the Q and /3 chains have no alternative partner with which to pair. In normal cells the situation is more complex, since each a or /3 chain has the opportunity to pair with its homologous chain (in cis or trans) and eventually with the chains of other isotypes (in cis or trans). When we used the three-dimensional structural model of an HLA class I1 molecule proposed by Brown to test the different DR a-DQ p dimer formations, we identified three distinct areas in the a 1 and 0 1 domains which are the presumed to be interchains contact sites. Interestingly, while most of the bands between the DR a and DQ /3 residues are found in every DR a-DQ p combination tested, some are nevertheless restricted to particular dimers. Altogether, the data predict that DR a-DQw 1 j3 pairing is favored over DR a-DQw3 p and over DR a-DQw2 /3 (Hermans et al., 1989a; Charron et al., 1990a). This was verified at the product level, since we detected DR a-DQw 1 p molecules, while DR a-DQw2 ?/ molecules were not detectable in the appropriate haplotypes. Besides this qualitative regulation, the quantitative aspects of DR a-DQ p dimer expression were analyzed. We have observed that a permissive haplotype does not automatically result in the appearance of an interisotypic heterodimer. We therefore investigated the amount of individual a-p chain mRNA present in Epstein-Barr virus B cell lines in which biochemical analysis was unable to detect the presence or absence of isotype-mismatched a-/3 pairs. We have noted a correlation between the ratio of a-p transcripts of the different isotypes (i.e., DR and DQ) and the presence o r absence of the
HLA CLASS I1
DISEASE ASSOCIATIONS
125
DR a-DQ /3 product. DR a-DQw 1 /3 molecules could only be biochemically detected in cells in which we had observed a large excess of DR a transcripts over DR /3 and DQ /3 over DQ a (Lotteau et al., 1989b). In order to address some of the mechanisms underlying the formation of mixed-isotype dimers in a more direct manner, we undertook experiments designed to specifically modulate the expression of the DR a gene in Epstein-Barr virus-transformed human B cell lines. From the latter experiments we can conclude that transfection of a DR a gene into a cell line not expressing DR a-DQ /3 molecules induces the expression of DR a-DQ /3 dimers. In a second series of experiments, cells naturally expressing DR a-DQ /3 dimers were transfected with a DR a antisense cDNA. This, in turn, results in the extinction of the DR a-DQ /3 isotypemismatched pair. In conclusion, the absolute amount of each chain, the relative amount of each chain within an isotype, and the reciprocal affinity of chains for each other are potential regulatory parameters of DR a-DQ /3 hybrid formation (Lotteau et al., 1989a; Charron et al., 1990b). Whether isotype-mismatched molecules are important in the physiology of the immune response is an open question. The fact that few data are available may be due to the presumption that they were nonexistent. The reactivity of several published T cell clones has suggested that they could be restricted by such isotype-mismatched class I1 molecules (Gomard et al., 1986; Eckels et al., 1986). In serological terms allele“specific” human alloantisera are likely to have broad reactivity within a haplotype, rather than be strictly allele specific. These alloantisera may contain populations of antibodies recognizing interisotypic specificities (e.g., combination of DR a-DQ p). In this case the interisotypic hybrids would not be distinguished from the dominant allelic specificity, since DR and DQ are in strong linkage disequilibrium and DR a is monomorphic. New tools can be designed to search for such sera, and mutant cells (deleted of genes for one o r several chains) should be very useful. Similarly, transfectants containing various combinations of mixed a-/3 isotype pairs should be suitable targets. This approach will be of great importance in the future, especially if one o r several of these combinatorial determinants appear to be directly implicated in the pathophysiology of an HLA-associated disease. Interisotypic class I1 dimers are indeed suitable to provide a structural basis for any haplotypic effect within the HLA-D region which might be found in epidemiological studies. They could account for the contribution of both DR and DQ or DP subregions to susceptibility to a particular disease, as was recently suggested in CD and multiple sclerosis (MS). T h e suggestions that specific conformation within HLA class I1 dimers may represent a critical element for susceptibility and that hybrid class I1
126
DOMINIQUE CHARRON
molecules occur either by cis or trans-complementation within an isotype(s) are discussed below in the context of well-characterized diseases. For this purpose I review the data on RA, IDDM, CD, and MS which are most relevant to the structural and functional framework of the HLA class I1 molecules provided. V. Diseases
A. RHEUMATOID ARTHRITIS Although likely to be multifactorial, the search for an etiology of RA has focused mostly on a few immunological aspects: rheumatoid factors and HLA studies. The latter were initiated by the first reports over 10 years ago of the association of RA with the cellular subtype Dw4 (Stasny, 1976). Thus, the DR4 haplotypes became the most-studied HLA class I1 factors. Approximately 70% of RA patients are DR4, versus 28% in controls (Stasny, 1978). Shortly after its serological definition, the DR4 haplotypes were subdivided into five or six cellular specificitiesdefined by homozygous typing cells: w4, w10, w13, w14, w15, and KT2 (Reinsmoen and Bach, 1982). A general characteristic of the HLA system is that the different subtypes are unevenly distributed among normal populations in different parts of the world, DwlO being frequent in Jews and Asian Indians, Dw15 or KT2 occurring in Orientals. Among these subtypes only two-Dw4 and Dw14-are prevalent in RA and account for the found association with RA in Caucasian studies, while DwlO appears neutral. Great effort was then devoted to biochemical dissection of the molecular basis of the alloreactive splits of the DR4 haplotypes, which in turn helped to define the RA-associated haplotypes. Two-dimensional PAGE studies revealed large heterogeneity of DR /3 and DQ a-p electrophoretic patterns in DR4 (B. S. Nepom et al., 1983; G. T. Nepom et al., 1983) (Fig. 2).The different DRBl genes associated with a DRB4 gene (homogeneous in the DR4 haplotypes and corresponding to the DRw53 specificity) were subsequently sequenced. Only limited nucleotide differences were observed in circumscribed areas of the most external exon of the DR /3l gene. This HVR (i-e., HVR3) encompasses a stretch of ten amino acids between positions 65 and 75 and, by extension, a residue at position 86 (Gregersen et al., 1986).Dw4-, DwlO-, and Dwl4-DR /3l molecules have identical nucleotide sequences, except for seven nucleotides between Dw4 and Dw 10 and three between Dw4 and Dw14. These nucleotide differences result in only four amino acid differences for DwlO and two for Dw14 (Gregersen et al., 1986; Seyfried et al., 1987).
HLA CLASS I1
DISEASE ASSOCIATIONS
127
It is of interest to note that previous Southern blot RFLP studies, although usually efficient in revealing allelic diversity, were not very informative in the case of the DR4 haplotypes, notwithstanding their extensive diversity. This can be explained by the fact that the nucleotide variation affected a limited area of the gene, therefore probably corresponding to recent evolutionary processes with conservation of the restriction sites. Although several DR4 haplotypes are found to be positively associated with RA in different ethnic groups (e.g., Dw4 and Dw14 in Caucasians and Dw 15 in the Japanese), DR4 is not associated with RA in other ethnic groups, such as Jews or Asian Indians. While DwlO is the most common DR4 haplotype in Jews, it is almost completely absent among RA patients, suggesting that DWlO in this population is probably “protective,” or at least neutral. The position of the molecules which differ in amino acid composition between Dw 10 (not associated or protective) and Dw4Dw14 (both associated with disease susceptibility in RA) can be considered the most likely to be critical in determining the impaired immune response which underlies susceptibility to RA. These include primarily amino acids 67,70, 7 1, and 86. Moreover, DR 1 is overrepresented in the group of DR4- RA patients, irrespective of the ethnic group. This parallels the fact that DR1 is one of the least variable haplotypes among ethnic groups. Furthermore, the relationships between DR1 or DR4 haplotypes and RA include the supertypic specificity, MCl (Duquesnoy et al., 1984; Lepage et al., 1985). MC1 is a class I1 determinant defined by both serological (i.e., with alloantibodies or mAb’s and cellular (i.e., with cloned and primed lymphocytes cells) typing. DR1, DR4, and MC1 are strongly associated, since DRl and DR4 are entirely included within MC 1 (at least in Caucasians), although MC 1 is found in rare DR2, DRw9, and DRw 10 haplotypes. Interestingly, MC 1 represents highest single determinant associated with RA, since in a recent study of 80 patients MCl is present in 83% of the patients and 43% of the controls (Carpenter et al., 1988). This illustrates that, although supertypic (i.e., present on several distinct molecules), some “epitope” once identified may be of better value in predicting susceptibility haplotypes to a disease than a precise class I1 molecular subtype. Similar data were reported concerning the reactivity of 109 D6 mAb’s, which recognizes an epitope on DRw53 molecules (Lee et al., 1984). The risk for RA appears to be higher for 109 D6 than for DR4, since 109 D6 was found in a greater number of individuals with RA than those who are either DR4 or DRw53. However, DR7 individuals are also 109 D6’ (at least the one expressing the DRw53), but do not have a higher incidence
128
DOMINIQUE CHARKON
of RA. Thus, 109 D6 positivity does not itself confer full susceptibility, but could contribute to an increase of susceptibility in the context of other RA-associated alleles. Indeed, non-DR4 non-DRw53 109 D6’ RA individuals were found to possess DRw 10 molecules by assessed serology. Moreover, the individuals analyzed so far were reported to be also DR1 on the other haplotype. T h e 109 D6 serological epitope may therefore reflect a common class I1 structure present on both DRw53 p and DRwlO /3 chains (Winchester and Gregersen, 1988). MC1’ and 109 D6’ individuals may be more closely related than was previously thought. Although MCl is not present on DRw53 (since DR7 individuals are MC1-), the DRw53 sequence (amino acids 65-73) is identical to the one found in the DRw 10 p 1 chains, while some DRw10 individuals belong to the MCl’ group (Merryman et al., 1988; Carpenter et al., 1988). This suggests that among DRwlO individuals there is some heterogeneity interconverting 109 D6 and MC1. T h e 65-75 region of the pl area appears to be critical when the positively associated haplotypes (i.e., DRl, DR4w4, DR4w 14, and DR4w15) are analyzed. Positions 67,70, and 7 1 have conserved residues, except for a relatively conservative lysine-to-arginine substitution at position 7 1 in Dw4. This contrasts with DR4-DwlO (a neutral or protective haplotype), in which positions 67, 70, and 71 are consistently substituted similarly to those which are found in most of the DR2 p chain and which have been claimed to be protective for RA, at least in Caucasians (Dw2) and in the Japanese (Dw12) (Maeda et al., 1981; Ohta et al., 1982). Interestingly, the serological and functional epitope MC1, which is also an RA “allele,” could be contributed to by a conformational structure shared by DR1, DR4w4, and DR4w14 and thus could be potentially encoded by the same 65-75 region of the DR p l gene. However, not every amino acid may be important, even in this 65-75 short linear sequence. Only a few, perhaps the one exposed at the surface of the molecule (toward the TcR) or part of the peptide recognition structure will be involved in eliciting an immune response. T cell clone reactivities are informative in this respect. Indeed, alloreactive T cell clones have been generated against HLA-Dw 14 which have revealed a “cellular epitope” present in all seropositive RA patients tested, irrespective of their being DR4 (Goronzy el al., 1986). A similar reactivity pattern is found using an A S 0 whose sequence was derived from the DR4-Dw14 p chain 68-74 sequence. Furthermore, one can notice that a rare DRw6 (Dw16) subtype which possesses the DRwl4 epitope is also associated with RA (Seyfried et al., 1988). In other studies T cell clones have been found to recognize shared determinants between DR1 and DR4 RA haplotypes (Weyand et al., 1986). Overall, the data
HLA CLASS
I1 DISEASE ASS<)CXArIONS
129
localize the functional RA-associated (i.e., shared) epitope to an area of the DR p l molecule common to DR1, DR4w4, and DR4w14 (also Dw15 and Dw16). However, it was impossible to strictly correlate the recognized epitope with this sequence. This suggests that the disease susceptibility factor is not a common continuous linear sequence between all of the recognized epitopes, but is likely to be a three-dimensional structure of a conformational epitope, which may be of importance independently of the overall structure of the molecule on which it is found. Interestingly, in the mouse a mutant, Bm12, shows that a structural epitope can be fully exchanged between an I-E and an I-A p chain and remain fully functional (Mengle-Gaw et al., 1984). Considering the above concepts, it could be that T cell clones would be the most specific tools to be used in disease susceptibility analysis. If the molecular substratum of the disease-associated immune process resides in an antigen presentation phenomenon to trigger (auto)reactive T cells, it is conceivable that some variability is allowed for the linear structure of the class I1 molecule, so long as the three-dimensional conformation of several important residues is conserved: residues directly involved in association with the antigen and/or with the TcR, for which no variation would be allowed. The concept of molecular mimicry was logically proposed based on the assumption that an environmentally provided antigen could exist which would induce T cell recognition of the previously defined disease susceptibility epitope. This would result in the initiation of a self-T cell activation and further autoimmunity. A search for molecular sequences shared between a foreign antigen (e.g., a pathogen) and the “self” molecule was initiated. In this quest the gpl10 glycoprotein from Epstein-Barr virus revealed a QKRAA-QRAA sequence from residues 808-8 16. Since an arginine-to-lysine substitution is considered conservative, this sequence is reminiscent of the QRRAA or QKRAA sequences present at residues 70-74 in the hypervariable region of several RA susceptibility DR p alleles. These include DRl, DR4w4, DR4w14, and DR4w15 (Roudier et al., 1988). The nonassociated haplotypes DR4-Dw13 and DR4-Dw 10 differ noticeably (QRRAE and DERAA, respectively). Moreover, a hydrophobicity plot of the Epstein-Barr virus gpl10 reveals that the QKRAA-QRAA is located on the a-helical structure similarly to the class I1 HVR3 and is thus exposed. T cells recognizing and proliferating to the QKRAA determinant could be initiated and expanded by the initial pathogen sharing the QKRAA sequence. Perpetuation of the antiQKRAA response would then occur even after the initiating antigen (i.e., the Epstein-Barr virus) had disappeared. It was recently shown that Epstein-Barr virus induces T cells that
130
DOMINIQUE CHARRON
recognize QKRAA containing a gpl10 peptide as well as a Dw4 peptide. However, no evidence was provided that the T cells involved were recognizing identical epitopes, since they could also be overlapping epitopes (Roudier et al., 1989). Thus, the self-antigen-mimicry hypothesis requires further dissection of the T cell responses, the TcR repertoire, and the peptides involved. It also remains questionable why these “self ”specific T cell clones have not been deleted during thymic selection, since their specificity is self (class I1 peptide) in nature. Although T cells reactive with self-histocompatibility molecules are found in the synovial fluid and are increased, whether these autoreactive T cells recognizing DRl-Dw4, -Dw14, and -Dw15 would trigger RA and whether these T cells are expanded by recognition of Epstein-Barr virus gpl10 are still unknown. Besides self-mimicking antigens, more conventional antigens have also been considered as triggering agents in RA. These include Mycobacterium tuberculosis, for which T cell hyperresponsiveness has been described in DR4 individuals (Palacios-Boix et al., 1988). Moreover, Mycobacterium tuberculosis-specific T cells have been characterized from synovial T cells in patients with RA as well as anticollagen type I1 T cells. Although the anticollagen type I1 response is under MHC control in several animal models, the data from humans cannot conclusively validate the hypothesis of a central role for this latter molecule in RA. T h e isolation of heat-shock protein-specific T cell clone from an RA patient (Holoshitz et al., 1989) is puzzling, and the isolated clone’s being y-6 is even more enigmatic. Apart from the overwhelming evidence for a role for the DR 0 1 gene in RA susceptibility, the involvement of other locus products remains unestablished. Indeed, DR4 haplotypes include polymorphic DQ /3 genes. However, individuals with DR4-associated RA carry either the DQw3.1 or the DQw3.2 species, and the DQw3.1 and DQw3.2 specificities are equally distributed in RA patients and in normal controls. Thus, polymorphism of DQ does not appear to greatly contribute to the disease susceptibility, and any DQ /3 association in RA is likely to be secondary to DR4. However, DQ may influence the phenotypic expression of RA and superimpose some feature on the main genetic susceptibility, due to the DR /3l gene association. In this respect an increase of the DQw3.1 (DQw7) specificity has been found in severe (i.e.,seropositive) RA and in Felty’s syndrome or RA associated with nodules and/or erosions (Sansom et al., 198’7, 1988). Patients with Juvenile Rheumatoid Arthritis URA) constitute a group of clinically and serologically heterogeneous individuals, accounting for the diversity of the HLA data generated in this disease. Weak associations
HLA CLASS I 1
DISEASE ASSOCIATIONS
131
have been reported with DR4, DR5, and DRw8. Some particular aspects were revealed in the group of seropositive JRA patients associated with DR4. Although DR4 homozygosity was increased in serological studies, this in fact reflects an even higher relative risk conferred by a heterozygous state Dw4-Dw14 (relative risk, 116) (Nepom et al., 1984), contrasting with a relative risk of 7.2 for DR4. Since the epidemiological data suggest a contribution of both haplotypes to the disease susceptibility element in JRA, the conformational epitope responsible is likely to be located on a hybrid HLA class I1 molecule, the a and P chains of which have not yet been identified. It may be of interest to note that a recent association was reported between JRA and the DPw2 specificity (Odum et al., 1988). Determination of whether the DR and DP associations are independent, additive, o r synergistic should help to identify the JRA susceptibility conformation. In the latter case an interisotypic DR-DP o r DP-DQ molecule is a likely candidate. Th e latest data suggest that the DPBw2.1 allele found in JRA patients is independent of the DR5 and DRw8 specificities (H. Erlich, personal communication). However, one should wait until we know the DPA allele present in these patients, since the best disease linkage may well be subsequent to an isotype-mismatched DR P-DP a molecule. In addition to the genetic aspects, there are several somatic aspects of the disease association which are important. Cell surface expression of the HLA class I1 molecules is likely to elicit, at least in part, the pathogenic effect of the susceptibility gene(s). Indeed, numerous abnormalities of HLA expression have been observed in RA patients. T h e most relevant finding is an aberrantly high expression on the synovial tissue, in particular, on the adherent synovial lining cells of the HLA class I1 species. All three subsets-DR, DQ, and DP-are concomitantly expressed at levels comparable to those found on B lymphocytes of the same individual (Teyton et al., 1987; Charron and Teyton, 1987). This absence of differential isotype expression does not favor the prominent role of any class I1 subset in mediating the susceptibility. The capacity of synovial tissue to express HLA class I1 antigens may be interpreted as an argument in favor of an ability to initiate and perpetuate a local immune response processing and presenting an antigen (auto- or foreign) to T cells. This has yet to be rigorously demonstrated in the model. It is of hypothetical interest that the Ii chain [a proteoglycan associated with a-P class I1 dimers (Charron et al., 1983) and thought to be involved in antigen processing and/or presentation] is strongly expressed and hypersialated in adherent synovial lining cells (Teyton et al., 1987). Alternatively, HLA class I1 expression may provide a beneficial effect, allowing HLA class II-restricted cytotoxic T cells to eliminate altered or in-
132
DOMINIQUE CHARRON
fected synovial cells. Class I1 expression on synoviocytes correlates with their activation and proliferative capacity. Furthermore, we have recently shown that synovial cell proliferation could be modulated by anti-HLA class I1 antibodies (Teyton et al., 1990). The class I1 molecules could therefore provide to the synovial cells an intracellular signal which affects proliferation and/or activation. In addition to an immune function, the presence of HLA class I1 molecules of synovial cells may be important in down-regulation of their proliferation. Although y-interferon is a likely candidate (alone or in conjunction with other cytokines) for induction of the in vivo expression of HLA class I1 in synovial tissue, the exact mechanisms of such expression are unknown. In vitro inducibility of HLA class I1 expression by y-interferon in adherent synovial lining cells does not differ between cells from normal and RA individuals (Teyton et al., 1987). Thus, considering the hypothesis that the disease could correlate with hyperinducibilityof class I I molecules on synoviocytes, several key pieces are missing from the puzzle. Among them, the most important are probably identification of the antigen(s) involved in triggering of the autoimmune process and the composition of the T cell repertoire capable of responding to the above aggression.
B. INSULIN-DEPENDENT DIABETES MELLITUS IDDM has been the most investigated pathology with regard to the HLA class I1 system association and autoimmunity (Table I). This is remarkable for a disease which was not considered to involve the immune system less than two decades ago. Study of the HLA has been highly rewarding over the last years. The existence of an IDDM susceptibility gene within the MHC derives from two basic epidemiological observations: (1)the presence of particular HLA specificities, with a higher frequency in affected individuals than in the normal population, and (2) HLA haplotype sharing within a family, which confers a higher risk (Thomson, 1984;Svejgaard et al., 1983). However, the early data obtained in twins suggested that the disease is polygenic and that the HLA system contributes approximately 50% of the inheritability. Most of the efforts have since been concentrated toward the identification of the best HLA linkage from IDDM. Progress almost parelleled the improvement in HLA typing and reflected successive significant technological advances. The first association was described in 1973,when the specificity of B15 was reported, followed in 1974 by the report of a B8 association (Nerup et al., 1974;Singal and Blajchmann, 1973).The most exquisite recent association is with the absence of aspartic acid at position 57 in the DQP chain (Todd et al., 1987).This illustrates the development
HLA CLASS I1 DISEASE ASSOCIATIONS
133
of HLA typing procedures from serology to molecular biology and from sera to sequences and the subsequent subdivisions of loci and alleles. It is feasible that, during this period, improvements in I-ILA typing (mainly identification of variants within supertypic specificities) led to an increase in the relative risk for the disease. Thus, IDDM is an exemplary case for further dissection of the polymorphism of the HLA system in order to identify HLA factors related to the highest susceptibilities and the molecular basis for the fine specificities of the different alleles implicated. An extensive review of the individual steps which have improved the definition of IDDM and HLA association is not required, since they reflect intermediate approximations of the most accurate genetic linkage which is presently known. Briefly, the early B15 and B8 associations were due to high linkage disequilibrium of these haplotypes with DR4 and DR3 associations, respectively, found years later. Indeed, over 90% of Caucasian IDDM TABLE I SUMMARY OF HLA FACTORS CONTRIBUTING TO IDDM SUSCEPTIBILITY OR RESISTANCE^ Susceptibility Dr4 DR3 DRl DRw16 DRwl3 DQw8 DQw2 DQw5 DQw5 Dw19
Resistance
DRwl5 DRwl5 (DQwl.l) DR4 DR5 DRwl3 (DQw3.2) DQw6 DQw6 (DQw1.1) DQw7 (DQw 1.AZH) DQw7 (DQw1.19) Dw19
DQ357: Ala-Val-Ser
(DR2) (DR2)
(DRw6) (041.2) (DQ1.12) (DQw3.1) (DQw3.1) (DQw1.18) DQw6 DQ357:Asp
Heterozygous effect: hybrid HLA class I1 molecules. DR4 DQw8/DR3 DQw2 and DR4 DQw8/DRw8 DQw4: synergistic effect. DR7 DQw2 in blacks (DQw2P + DR4 DQa); DR9 DQw2 in blacks (DQw2P + DR4 DQa). T h e new HLA nomenclature is used whenever possible. T h e previous corresponding designation is provided in parenthesis. Data are taken from several sources (Todd et al., 1987; Todd el al., 1988a,b; Ronningen el al., 1989; Horn el a/., 1988).
134
DOMINIQUE CHARRON
patients are positive for DR3 and/or DR4, as compared to 45% of normal controls (Svejgaard et al., 1980). Subdivision (mainly by cellular and molecular means) of the DR4 haplotypes contributed to better definition of the disease susceptibility markers. However, the cellular splits of DR4 into five or six subtypes gave a rather confusing picture, with an increase in DR4-Dw4, but also in DR4-Dw14 (and DwlO in Jews), while the situation for DR4-Dw13 and DR4-Dw15 was uncertain (Bach et al., 1985). These uncertainties are now explained by the molecular composition of the DR and DQ a and /3 chains present in these cellularly defined haplotypes. Indeed, when the three segregant series-DR, DQ, and DP-are considered, the association with IDDM is the strongest with HLA-DQ (Owerbach et al., 1983; Cohen-Haguenauer et al., 1985; Bohme et al., 1986). This includes DQw3 and DQw2 subtypes. Among DQw3, at least three molecular subtypes have been found using serological, cellular, and molecular approaches (DQw3.1, DQw3.2, and DQw3.3, or DQw7, DQw8, and DQw9, respectively). A higher relative risk is associated with one of the DQ /3 subtypes (this subtype being shared with the different DQw subtypes of the DR4 haplotypes, which were found earlier to be susceptible). This linkage was first suspected by RFLP studies. In 1983 Owerbach et al. reported a 3.7-kDa BamHI DQ /3 gene fragment which was decreased in frequency in DR4' IDDM patients compared to DR4' controls. This fragment turned out to be allelic to the specificity DQw3.2 or DQw8, which was later shown (using several RFLP DQ /3 fragments') to be highly associated with IDDM (Owerbach et al., 1984; Cohen-Haguenauer et al., 1985). The importance of the RFLP data was strengthened by a perfect correlation of these fragments with a DQ p chain profile, as assessed by two-dimensional PAGE (Kim et al., 1985). Moreover, it linked the putative susceptibility epitope on the DQ p molecule of a given type (DQw8 or DQw3.2). The molecular subtype DQw3.2 (i.e.,DQw8) is thus the most prevalent accurate disease susceptibility linkage specificity found in IDDM in Caucasians, while in Orientals DR4, DRw8, and DRw9 are associated with IDDM. In contrast, DR2 appears to be highly protective in all studies (Bertrams and Baur, 1984). Further refinement of the disease susceptibility factor has been attempted. Indeed, since IDDM is positively associated not only with DR4, but also with DR3 and DR1, in different distinct populations, a search for a structure common to these positively predisposing haplotypes was conducted based on the available nucleotide and amino acid sequences of I
For example, 1.9-kb TaqI and 12-kb BamHI.
HLA CLASS I1
DISEASE ASSOCIATIONS
135
HLA class I1 genes. When inspecting the individual class I1 amino acid substitutions, Todd et al. (1987) noticed that all class I1 haplotypes which were not positively associated but neutral o r negatively associated with IDDM possessed in common an aspartic acid at position 57 in the DQ /3 chain. In contrast, there is no common polymorphic determinant (or structural stretch of amino acids) in the IDDM positively associated haplotypes (noticeably DR4, DR3, and DR1) and amino acid 57 of DQ /3 is an alanine, a valine, or a serine. Interestingly, Asp57is also found in every mouse I-A /3 gene sequenced so far, with the exception of the NOD mouse, which represents the best mouse model for spontaneously developing IDDM, in which it is serine (Acha-Orbea and McDevitt, 1987). This finding of specific DQ /3 amino acids associated with DR4’ IDDM has allowed the construction of oligonucleotide probes, which are helpful in population studies, since they can be used in dot blot analysis, a method adapted for large-scale studies. For example, when an oligonucleotide probe detecting DQw3.2 versus DQw3.1 was used in conjunction with an oligonucleotide probe detecting DQw2 and DQw 1.1 in a population of 39 Caucasian IDDM patients, only 10% of these patients were heterozygous for DQ /3 Asp57 (Todd et al., 1987). DQ /3 Asp“-negative homozygozity was thus correlated with disease in 90% of the patients. T h e results are therefore in full agreement with previous RFLP and two dimensional PAGE studies, showing a large increase of the DQw3.2 /3 chain over DQw3.1 and DQw3.3 (Monos et al., 1987). DQw3.1 and DQw3.2 can also be distinguished by their reactivity with the mAb TA 10 (Schreuder et al., 1986). Similarly, two-dimensional PAGE would have detected specifically the DQw2, DQw 1.1, and DQw 1 AZH /3 chains in the non-DR3-non-DR4 IDDM. Since the initial reports of the importance of position 57 of DQ /3 in IDDM susceptibility, the data have been both largely confirmed and extended. The analysis of haplotypes rarely associated with IDDM has been rewarding: DRw6, the DQ /3 sequence derived from DRw6’ IDDM, contained a Val57of the DQ fl chain, which corresponds to the DRw6Dw 19 subgroup of DRw6 individuals. This haplotype was present in all seven DRw6 IDDM patients observed, while it was present in only three of 13 DRw6 controls (most of the controls were Dw18 and possess an A~p”~-positive DQ /3 chain, DQJ31.6) (Horn et al., 1988; Todd et al., 1988a). Also, rare DR2 IDDM has been reported, since DR2 confers a dominant resistance to IDDM, as discussed later. However, in the DR2 IDDM examples (which are of a specific DR2 cellular subtype named AZH) the DQ /3 chain possesses a Ser57 DQ /3 chain, contrasting with the Asp57
136
DOMINIQUE CHAKRON
present in the two resistance-conferring haplotypes (DR2-Dw2 and DR2-Dw 12) in their respective DQ p chains. Similarly, the positive association reported with DR1 and the negative association with DR5 are consistent with the codon 57 pattern, since in DR1 the DQ /3 chain has a positive Val57,while the DQw3.1 /3 chain of DR5 contains Asp57.In the frequent case of DR3 association, the DQw p2 chain contains an Ala57. This is commonly found in Caucasians, while in blacks the DR3 is mostly associated with a specific DQw4 p chain. This chain contrasts to the regular DQw2 p chain, as it possesses an Asp57.However IDDM DR3 blacks appear to be DQw2+,and, given that DQw2’ serology reflects the DQ p chain, they are likely to be Asp57DQ p. Since associations of certain DR and DQ antigens are very different among ethnic groups, one should only consider the individual DR /3 and DQ a and p chains present (and their sequences) and not rely on the “classical,”but overestimated, linkage disequilibrium phenomenon. The emerging picture suggests that it is the charge of the polymorphic residue at position 57 of DQ p that is associated with IDDM susceptibility. Thus, the presence at position 57 of the DQ p chain of amino acids with a nonpolar hydrophobic R group (e.g., alanine or valine) or amino acids with a polar but uncharged R group (e.g., serine) is preferentially associated with an autoimmune response to an as yet unknown diabetes-related antigen, and this contrasts with the positively charged R of aspartic acid found in the IDDM negatively associated with haplotypes. Several important exceptions exist to the absence of Asp57DQ /3 as a requirement for IDDM susceptibility. The DRw9 haplotypes associated with IDDM in the Japanese have Asp57DQ p. This is also the case for two other susceptible haplotypes found in Japanese DR4 and DQw9. Moreover, the presence of a non-aspartic acid residue at this position may not be sufficient to confer the highest susceptibility to IDDM. In fact, a heterozygous effect was reported early in IDDM studies, which forces the consideration of the contribution of both HLA haplotypes in order to delineate the best disease susceptibility element (Svejgaard et al., 1983). It was already apparent in the early studies in which B8 and B 15 were found to be associated with IDDM that both haplotypes could contribute to the susceptibility of IDDM. A dramatically increased relative risk was further demonstrated for individuals possessing a particular heterozygous combination of HLA class I1 antigens, namely, DR3-DR4. This prompted a search for the class I1 molecule inducing the most susceptibility. Since the DR ct chain is monomorphic, it cannot therefore contribute to the heterozygous effect. The discovery that both Q and p chains of the DQ molecules were polymorphic led to the ideas that the formation of hybrid class I1 molecules obtained by trans-complementations could oc-
HLA CLASS I1
DISEASE ASSOCIATIONS
137
cur and that those molecules present only in heterozygous individuals will fulfill the requirement imposed to explain the heterozygous effect (similar to an FI effect in animal genetics) (Svejgaard et al., 1983; Charron et al., 1984). Ultimately, these ideas support the concepts that it is more the structural three-dimensional conformation than an individual sequence which is the likely genuine disease susceptibility element and that several distinct amino acids can contribute to its formation. Indeed, this would agree with our present understanding of the structural model for antigen presentation, which implies a series of dynamic interactions among the antigenic peptide, the MHC, and the TcR. Hybrid class I1 molecules provide the best model to explain the heterozygous effect observed for IDDM. The demonstration of the existence of homologous hybrid HLA-DQ molecules lends weight to this hypothesis (cf. Section 111). It was subsequently shown that these hybrid molecules are in fact expressed in DR3-DR4 individuals. This occurs in both normal and IDDM individuals, and no obvious quantitative difference was observed on the cells studied (EBV B cell lines) (Nepom et al., 1987). Nevertheless, these results support the possibility that hybrid conformational epitopes are created de now, which may affect the immune response and lead to the autoimmune process. Indeed, in the mouse such hybrid determinants would be only and specifically recognized by alloreactive T cells and T cell clones (Fathman, 1980). These hybrid molecules may bear a fixed three-dimensional structure, which contains a unique conformational entity itself bordered by a set of interacting residues (on the same, as well as opposing, chains). Some degree of degeneracy in the amino acid composition could be allowed by the model, leading to a gradient of “susceptibility” molecules. The data are consistent, the highest susceptibility being mediated by a conformational structure present in an hybrid molecule formed by the DQw3.2 (i.e., DQw8) fl chain of the DR4 haplotype and the DQw2 (Y chain of the DR3 haplotype. T h e “auto”-antigen, a foreign or autologous peptide, will be better accommodated by the molecule of highest susceptibility and less along the gradient of molecules. In this hybrid, or conformational, model similar conformations could be obtained in class I1 dimers resulting from trans-, but also from cis-, complementation. T h e minimal requirement at the DQ fl chain level would be the absence of Asp5’. Indeed, the heterozygous effect was observed not only in DR3-DR4 individuals, but also in DR4-DRl individuals. This could be interpreted in that the conformation of the DQ hybrid molecule formed in DR4-DR1 is very close to that formed in DR4-DR3 (functionally and/or for peptide binding), and thus confers a similar level of susceptibility.
138
DOMINIQUE CICAKKON
It is of interest to note that contrasting with DR4 is the case of DR3 association with IDDM. No distortion of allele frequency variation was found between patients and normal controls at either the DR /3 or DQ /3 locus. This was recently extended to the DQ a locus, for which the same heterogeneity was found in both patients and controls, thus not allowing identification of the residue(s) of the DQw2 a chain, which contribute(s) to disease susceptibility. When other DQ a-P combinations are present, the level of disease susceptibility appears to be lower, which may reflect less functional efficiency of the hybrid molecule created in these cases. There are even the possible combinations of DQ haplotypes in which the a-/3 associations are forbidden, as has been reported in the mouse for some I-A a-P. In this case only the homologous cis-encoded a-P DQ dimer would be expressed and could be the least active in terms of peptide recognition. Overall, this may mean that the structural requirement may be less stringent for DQ a than for DQ P. This would explain how the heterogenous effect can still be accounted for by a capacity of association of the DQw3.2 P chain identical to that of any type of DQ a chain found in DR3 haplotypes. Alternatively, the heterozygous effect may not concern a DQ a-P dimer, but some other type of class I1 molecules. In this respect, it is noteworthy that in Japanese IDDM-susceptible haplotypes DR4-DQw4 and DRw9-DQw9, the DQ P chain has Asp5’, but that it is the DR p chains which may contribute to disease susceptibility (i.e., the absence of aspartic acid or the presence of serine in IDDM DR /3 chains from Japanese patients). In this population a characteristic DR P-DQ a dimer could be equivalent to the DQ P-DQ a dimer found in Caucasians. In the IDDM-prone NOD mice which are I-E-, the expression of I-E (i.e., DR-like) molecules prevents the development of diabetes, while the I-A /3 (Lea,DQ-like) chain possesses a predisposing Ser5’ residue (Nishimot0 et al., 1987). While the authors suggest that protection may be due to an I-E-controlled suppression or to cross-tolerance to the self-antigen, I propose an alternative explanation. Introduction of an I-E a gene may impair the balance of the I-A and I-E a and P chains to the point that an excess of I-E a could associate with I-A P (interisotypic molecule) and thus prevent A a-/3 diabetogenic dimer expression at sufficient levels. The possible contribution of molecules other than DQ was raised earlier, when a group reported the existence of a DX a polymorphism (detected as a TuqI fragment in RFLP) associated with IDDM (Festenstein et al., 1986). However, the DX a gene has no known product, and several other groups did not find the DX a association. Again, the DX a polymorphism reflected a selection bias due to linkage disequilibrium. There is no reason to believe, therefore, that DX a alleles themselves add to IDDM
HLA CLASS 11 DISEASE
ASSOCIATIONS
139
susceptibility. Additional sequences should thus be present in the DQw2 a chain with which the sequence identified on the DQw3.2 p chain coded in trans would combine in heterozygotes to form a DQa-P dimer with full susceptibility capacity. T h e Brown model of a class I1 molecule predicts that the DQ p chain amino acid 57 is located at an extremity of an a helix and is pointing into the antigenic groove. Thus, the negatively charged n 0 n - A in ~ DQ ~ ~ p~ is likely to form a salt bridge with the positively charged Arg7gof the DQ a chain. However, the fact that the IDDM-associated DR3 haplotypes and the IDDM-neutral DR7 haplotypes, although having different DQ a chains, display the same two opposing amino acids does not favor any particular role for the Asp”7DQ /3-Arg79 DQ a interaction. Additional polymorphic adjacent sequences are evidently needed to explain the latter. Interestingly, the study of populations (and cases) in which IDDM susceptibility does not strictly follow the DQP57 pattern have provided support for the hypothesis that certain combinations of DQ a and DQ /3 obtained by trans- o r &-complementation are the crucial structures involved in IDDM susceptibility. In a series of 92 IDDM patients from Norway, an increased risk was found among DR4-DRw8 heterozygotic individuals, similar to that seen for DR3-DR4 heterozygotes. By RFLP and AS0 typing, the DR4-DRw8 patients appear to bear the DR4DQw8/DRw8-DQw4 haplotypes in eight of nine. This suggests that the two distinct combinations of haplotypes D R ~ - D Q w ~ / D R ~ - D Qand w~ DR4-DQwNDRw8-DQw4 share an equivalent DQ a-P conformation encoded in trans. Interestingly, DQw4 has an Asp57of its D Q P chain, and thus the DQw4-DQw8 heterozygotic individuals do not have two Asp57negative DQ p l alleles. In these cases Asp57DQ p is not protective as in DR2-Dw2. DR7 (and DRw9) haplotypes appear to confer disease susceptibility in blacks that contrasts with a neutral effect of DR7 found in Caucasians. Todd et al. (1988b) analyzed the DQ a and p chains found in these black haplotypes. While the sequences of the amino-terminal domain of the DR p chains (pl and p4) and DQ were identical in black and Caucasian DR7-DQw2 haplotypes, a difference was found in the DQ a chain. T h e black DQA 1 allele was identical to the DQA 1 allele found on DR4 Caucasian haplotypes. Thus, replacement of the DR7-DQA l allele (present in DR7 Caucasians) by a DR4-DQAl allele (present in DR7 blacks) appears to be sufficient to switch from a neutral to a susceptible haplotype for IDDM. Similar results were obtained in DRw9 black haplotypes in which DQ p was identical to the DQ p of black DR7 while the DRw9-DQ p from
140
DOMINIQUE CHARRON
Caucasians is different (DQw9 Asp57positive). This is consistent with DRw9’s also being a susceptible haplotype in blacks, although neutral in Caucasians. It should be pointed out that the DQAl of the black DRw9 haplotypes was found to be identical to Caucasian DR9 and black DR7 and to correspond to the regular DQAl of DR4. The DQa-/3 dimer composition found in DRw9-DRw9 blacks with IDDM is thus similar to that obtained in trans in the DR3-DR4 heterozygous Caucasians. Overall, the genetic data argue that position 57 of DQ p, although prominent, is not the only contributary residue to IDDM susceptibility. An unknown sequence(s) presumably on DQ Q is usually required to obtain full disease susceptibility. In some instances this DQ a factor may even overrule the causative effect of the specific DQ fl sequence. Moreover, one cannot exclude a role for additional sequences of the DQ 3./ chain. In some populations contribution to disease susceptibility by amino acids usually found on the DQ /3 chain could be absent and replaced by equivalent residues on DR p. The exact mechanisms by which islet p cells which provide insulin are progressively destroyed in IDDM are presently unknown. The way in which the genetic susceptibility is phenotypically translated remains largely speculative,although of critical importance. This central question has prompted many laboratories to investigate the somatic expression of HLA class I1 molecules in the pancreatic endocrine tissue and its regulation and its role in the physiopathology of the disease. Following the thyroid model, in which the aberrant expression of class I1 molecules on the epithelial cell population was proposed as the central event leading to the autoimmune phenomenon observed in thyroiditis, a similar hypothesis was proposed for IDDM (Hanafusa et al., 1983; Bottazo et al., 1986). Indeed, HLA class I1 expression was documented on p islet cells from a few IDDM patients, while other islet endocrine cells and the exocrine cells of the pancreas remained class I1 negative (Bottazo et al., 1985). Moreover, among human /3 cells only a small proportion of cultured cells are induced to express class I1 by y-interferon. Only when tumor necrosis factor a (or p), although not efficient alone, is combined with y-interferon do human /3 islet cells become strongly positive in culture (PujolBorrell et al., 1987). Recent data from mice have shown that class I1 expression alone is not sufficient to endow complete antigen-presenting cell function of /3 islet cells (Markmann et al., 1988). This casts doubt on the possibility that class I1 expression could be solely responsible for the initiation of autoimmunity. Moreover, the p cell specificity of HLA class I1 expression observed in vivo remains unexplained (Timsit et al., 1989). It could be that islet p cells are uniquely sensitive to an as yet unknown specific combination of lymphokines or, alternatively, that a p islet cell
HLA CLASS 11 DISEASE
ASSOCIATIONS
141
specific virus determines the class I1 induction via a mechanism similar to that which has been reported in rats, in which a noninfective neurotropic coronavinus induces class I1 expression in astrocytes (Massa et al., 1987). This may also be related to the class I1 expression of thyrocytes in culture which is observed following the introduction of simian virus 40 DNA (Belfiore et al., 1986). A series of elegant experiments was designed to precisely address the question of the role of aberrant HLA class I1 expression in the pathogenesis of IDDM. These experiments provided some unexpected results. Mice were made transgenic using constructs containing the insulin promoter gene and class I1 a and p chains. It was anticipated that these animals would only differ from nontransgenic animals by the expression of class I1 in pancreatic islet /3 cells. This was indeed the case (Lo et al., 1988; Sarvetnick et al., 1988). Moreover, diabetes results in almost all of the animals. However, one of the most unexpected findings was the absence of an inflammatory infiltrate surrounding the islet cells (noticeably T cells). This may be consistent with the incapacity of these cells to actually function as antigen-presenting cells. Thus, class I1 expression does not appear to be capable, by itself, of inducing autoimmune destruction of the islet /3 cells. However, the islet cells were altered and progressively disappeared. T h e decrease in insulin secretion and subsequent cell death observed in the p cells of these transgenic animals in the absence of T cell infiltration are puzzling, as is the observation that similar transgenic animals with class I genes, as well as with class I1 genes, became diabetic (Allison et al., 1986). Although several highly speculative explanations were proposed, none has yet been validated experimentally. This is the case for competition for transcriptional factors between the transgenic MHC and the endogenous insulin gene, interaction between the intracellular pathway of insulin secretion and MHC class I1 expression leading to degradation of the insulin or direct binding between insulin and the MHC proteins (Parham, 1988). Besides their immunological role in restricting the immune response, class I1 molecules are capable of transducing signals into the cell bearing them, which can result in activation of the cell (or proliferation, or differentiation). Using Sepharose-conjugated anti-HLA class I1 antibodies, we were able to directly activate human B cells via a secondary messenger pathway involving calcium flux, phospholipase C metabolism, and protein kinase C activation (Mooney and Charron, 1988; Mooney et al., 1989). Stimulation of interleukin 1 synthesis and release by the B cell were also observed. I suggest that such a role for class 11 molecule may well be relevant to the observation made in the transgenic mice and
142
DOMINIQUE C H A K R O N
subsequently to the pathophysiology of IDDM. Indeed, hyperexpression of HLA class I1 on pancreatic islet p cells, once established, may alter the cellular biology of the islet cells and transduce activation signals contributing directly or indirectly to cellular destruction (via interleukin 1 o r via the release of other cytotoxic agents). Such a hypothesis accommodates the different possibilities of HLA class 11 induction on the p islet cells (e.g., direct infection of the p cells, release of the cytokines, and gene transfer) and does not require any primary immunological event and effector. It is important to consider that immune mechanisms (e.g., antigen presentation and T cell cytotoxicity) and nonimmune mechanisms are not mutually exclusive possibilities toward the destruction of the p islet cells. Both reflect the somatic expression of the class I1 genotype. Furthermore, recent data suggest that in mice the signaling capacity of a HLA class I1 molecule may well be allele specific (Bishop and Frelinger, 1989). Fascinating questions remain to be answered in order to understand the role of the HLA class I1 expression in IDDM susceptibility. As the genetics have become more precise, the expression leading to the pathology has become more enigmatic. Identification of the self-peptide (or foreign peptide), which binds efficiently to the class I1 molecules involved in the genetic susceptibility, will obviously be an important breakthrough in our understanding and may also provide new therapeutical approaches. Alternatively, the demonstration that class I1 expression may alter in some way the cell biology of the pancreatic /3 cell would direct our thinking of pathophysiology and therapy onto different tracks. The strong negative association of IDDM with DR2 has been a constant observation which warrants discussion (Bertrams and Baur, 1984). T h e supertypic DR2 specificity is protective, but only in some subtypes, as the subdivision of DR2 into DRw 15 and DRw 16 correlates exactly with disease susceptibility, DRw 15 being dramatically decreased in IDDM, while DRw 16 (a rare haplotype in Caucasians) is associated with the disease (Tiwari and Terasaki, 1985). Taking into account the molecular composition of the DR2-DRw 15 and DR2-DRw 16 haplotypes and comparing them with the sequences of the other positively and negatively associated haplotypes, Todd et al. (1987) concluded that only residue 57 of the D Q p chain correlates strongly with resistance and susceptibility to IDDM. The presence of an aspartic acid residue at this position would therefore explain the protective effect of most DR2 haplotypes. Indeed, it perfectly fit the earlier epidemiological data, DR2-Dw2 and DR2-Dw 12 being protective and possessing an Asp5’ DQ p, while DR2 AZH is susceptible and has a serine at the same position.
HLA CLASS
11 DISEASE ASSOCIATIONS
143
In contrast, the reason for the dominant effect of Asp57DQP in IDDM resistance is much debated. Several alternative explanations are proposed. Cross-tolerance between an antigen and the precise class I1 epitope could explain the dominant effect. T h e possibility that molecular mimicry is responsible for immunological tolerance has also been investigated (Todd et al., 1988a). The envelope protein I-E2 of cytomegalovirus (CMV), a virus which may be involved in IDDM, possesses a stretch of six amino acids (81-88), five of these being present in the DR P 1 and DQw2 /3 chains at residues 52-57. Anti-CMV T cell responses specific for this I-E2 epitope could only occur in individuals in which the I-E2-specific repertoire would not have been deleted as self during the thymic education process (non-DR2 individuals thus lacking the homology I-E2 are self ). Such speculation justifies the study of the antigen- and MHC-specific T cell repertoire expressed in the patients. Alternatively, the nonresponder status could be an active phenomenon (Boitard et al., 1988). Suppression via T cells has been documented and found to be under MHC class I1 control. Moreover, specific suppression toward two distinct antigens (streptococcal cell wall antigen and schistosomal antigen extract) were found to be restricted to the DQ isotype (Sasazuki et al., 1983). In these two models anti-HLA-DQ mAb’s would abrogate the in vitro observed suppression (Matsushita et al., 1987). However, it is not the D Q P position 57 residue which is involved, since DW2’ individuals develop streptococcal cell wall-specific suppression, while Dw 12 individuals do not, although they possess the same Asp residue. T h e dominant resistance to IDDM cannot be unequivocally and solely explained by this suppression phenomenon. Several haplotypes, although possessing an Asp57, are only weakly negatively associated o r neutral, and some Asp5’-positive DQ P individuals are positively associated with IDDM. For example, in the Japanese DRw9, one could argue that a DQ a chain could abrogate the role of the Asp57DQ P. Additional class I1 sequences of the DQ P chain or of another chain may be required to confer full dominant suppression. It may be relevant to note that NOD transgenic mice possessing the DQw 1 molecule from a DR2-Dw 12 haplotype (an A~p~~-positive molecule) are not protected from becoming diabetic (Fukui et al., 1989). These data do not support a unique role for Asp5’ in suppression. C. CELIAC DISEASE Among autoimmune diseases CD, or gluten enteropathy, is unusual in two ways. It has one of the strongest associations with HLA class I1 of the known HLA-linked diseases, and its triggering factor (wheat gluten) is
144
DOMINIQUE CHARRON
known. Although the genetic susceptibility appears to be predominantly associated with the DR3 specificity, other alleles are involved, including DR7 and DR5 (Betuel et al., 1980). In serological studies of a large population of patients, the DR3-DQw2 haplotype was found in over 80%of the CD patients, in contrast with less than 30% in healthy controls (Tosi et al., 1983). CD is also characterized by the heterozygous effect observed concerning DR3 and DR7, which in considerable in some populations (Betuel et al., 1980). This highest relative risk found in the heterozygous DR3-DR7 individual suggests that possibility of a transcomplementation within the DQ system (Charron, 1986). Rare are the non-DR3 non-DR7 CD patients. In a study of such a selected population all 16 were DR4. This suggests either the presence of a common and unknown epitope pattern shared between DR3-DR7 and the DR4 in these CD patients or the possibility that two distinct types of CD exist, one associated with DQw2 haplotypes and the other with DR4 (Tosi et al., 1986). Studies were recently conducted at both the DNA and protein levels. A D Q p chain cDNA probe was reported to identify a polymorphic fragment which was a better marker of the CD than the DQw2 serological specificity (Howell et al., 1986). However, this was not confirmed in a group of homozygous DR3 CD patients in which the DR p and DQ p RFLP patterns were identical to those found in the control group (Sacks et al., 1987). In another study a 4-kDaRSaI class I1 fragment was found to discriminate the CD haplotype from matched controls (90%versus 18%).It was subsequently shown to be encoded by a DP p gene (Howell et al., 1988). This was unexpected, given the degree of recombination between the DQ and DP loci. A possible explanation could be that DQ-DP recombination is lower in the DR3-DQw2 haplotypes. Alternatively, although independent in normal DQ haplotypes, DP may be more strictly linked in the CD haplotypes. Moreover, since a DP a polymorphism (84%versus 36%)was also increased in the same patients, along with the DP p polymorphism, this study may indicate a role for interisotypic hybrid molecules and determinants formed by DP a and DQ /3 and/or DPP-DQa. An increased frequency of two specific DP alleles was found in a study of 23 CD patients from Italy (Bugawan et al., 1989). This study was conducted using a panel of DP-specific oligonucleotide probes on polymerase chain reaction-amplified material, which allowed the detection of 2 1 DP p and two DP a alleles. Two specific DP p alleles were increased in the patient population (DPB4.2 and DPBS), conferring relative risks of 93 and 2.8, respectively. The present data could either reflect a linkage disequilibrium or an
HLA CLASS I1
DISEASE ASSOCIATIONS
145
independent contribution of DP /3 alleles, in addition to the wellestablished role of DQw2. Interestingly, in the two studies the DP /3 polymorphism (Bugawan et al., 1989; Howell et al., 1988) associated with CD susceptibility demonstrated differences aJnong ethnic groups (Italy versus the United States). T h e U.S. patients had an increase in DPB 1 and DPB3 variants. Altogether, the data focus on position 69 of the DP /3 chain, at which both DPB4.2 and DPB3 share a lysine residue (as in DPB1). When the DQ polymorphism was further investigated at the DNA and protein levels in 30 white patients a 4-kDa BglII RFLP fragment was found in 97% of the CD patients, compared to 56% in the controls. Moreover, all of the CD patients tested possessed a biochemically detected DQ a chain variant, which was associated with the 4-kDa BglII fragment (Roep et al., 1988). The identification at the DNA and protein levels of a CD-specific DQ a (DQw2.3) chain can be further considered. In DR3 and DR7 individuals structural analyses have revealed that the DQw2 molecules have almost identical /3 chains (they differ by only one amino acid in the second domain), while they have distinct a chains, as shown by two-dimensional PAGE and confirmed by nucleotide sequence data (Song et al., 1987). T h e constant presence in CD patients (including a non-DR3 patient) of a specific DQw2 a chain (usually associated with the regular DR3 haplotype) would not have been detected without the use of molecular probes, since the serological DQw2 specificity reflects the DQw2 /3 chain, and since this chain is structurally almost totally identical in DR3, DR7, and some DR5 and DRw8 haplotypes. T h e molecule of most importance in CD appears to be a DQw2.3 a-DQw2 /3 dimer. Interestingly, this combination can be obtained by a-/3 chain complementation either in cis (same chromosome) or in trans (from both haplotypes). In the latter case the DQ molecule would belong to the hybrid type created de nouo in heterozygous individuals (Fig. 5). Thus, in DR3 nonDR7 individuals this molecule would be encoded in cis, while in DR5DR7 it would be encoded in trans. This would also be the case in DRw8DR7 patients. The hypothesis of a hybrid HLA-DQ molecule has received further support from a recent Norwegian study of 94 CD children, who were typed using two specific oligonucleotide probes (Sollid et al., 1989). T h e first AS0 detects the sequence 72-78 of DQ a present in the DR3-DQw2 and DR5-DQw7 haplotypes, and the second AS0 detects the sequence 26-33 of DQ /3 present in DR3-DQw2 and DR7-DQw2. All but one patient are positive with ASO, which encodes a specific DQ a-/3 dimer. Th e same DQ a-/3 heterodimer (CD specific) could be coded either in cis
146
DOMINIQUE CHARRON
DR3 DQW2
DQBl
DQAl
DR7 DQw2
DQBl
DQAl
DR5 DQw7
DQBl
DQA1
1
1
HsplOlyp
cenes
FIG.5. HLA class I1 DQ haplotypes, genes, and molecules in celiac disease. In DR3-DQw2 individuals the genes which provide the DQ a-p dimer are encoded on the same haplotype (&complementation), while in DR~-DQw~/DR~-DQw~ individuals the same DQ cr-p dimer is encoded by genes on both haplotypes (tramcomplementation).
or in trans (the haplotype carrying the cis combination is likely to have evolved by crossing over during evolution between the two haplotypes carrying the trans combination). These results would explain why DR3 appears to be a susceptibility haplotype, together with any other, while DR7 is associated in CD almost only with DR5 or DRw8 (and DR3). Since the DQw4 a chain (present in DRw8) differs in sequence from other chains in the 69-75 region, it may be relevant that there is a Ser75present in DQ a of both DR3-DQw2 and DR5-DQw7 which is not found in any other DQ a allele. The rare non-DR3 non-DR7 patients are usually DR4. In the study cited above the only patient who was negative with the two AS0 probes was DR4. Unless a common epitope(s) is found between the DR4 and non-DR4 groups, this result would again argue for two distinct forms of the disease (Tosi et al., 1986). It is conceivable that a gluten-derived peptide (acting as an antigen) binds similarly to the previously defined DQa-DQ3 dimer in classical CD and to a different class I1 molecule (with similar or different conformation of the binding site) in the DR4 patients. Alternatively, the antigen may be different. This possibility is raised by the observation that in experimental allergic encephalomyelitis, which is under MHC control in the mouse, the triggering antigen, myelin basic protein, possesses two distinct epitopes, both of which are encephalitogenic. CD is also concerned with the phenotypic expression of HLA class I1 antigens. It is of interest to note that while crypts and epithelium are DRf in CD, only the epithelial cells appear to bear the DQ' molecule, which are
HLA CLASS
11 DISEASE ASSOCIATIONS
147
also increased in the lamina propria. The direct role of these cells in CD is supported by the recent observation that Ia’ gut epithelial cells can function as accessory cells in allogeneic reactions and process and present soluble antigen to primed T cells in an MHC-restricted manner (Mayer and Salien, 1987). Thus, the genotypic association with HLA class I1 in CD may result in an impairing of the local gut immune response through the expression of the susceptibility HLA molecules. Although it has been not defined chemically, the triggering antigen is likely to be a gluten-derived peptide which could associate with some specific affinity to the MHC class I1 molecule involved in CD. It should be pointed out that, when the antigen is polysaccharidic, the HLA class I1 immune response appears to be preferentially restricted by DQ molecules (Durandy et al., 1986).
D. MULTIPLESCLEROSIS T h e first suggestion, dating from the early 1970s, that the HLA system contributes to the genetic susceptibility to MS is now well established. T h e association was initially found with HLA-A3 and -B7 (Jersild et al., 1975; Compston, 1982; Tiwari and Terasaki, 1985), then more closely with DR2 and, in particular, the Dw2 subtype. This cluster of associations reflects the strong linkage disequilibrium affecting the A3-B7 and DR2Dw2 haplotypes in Caucasians. In fact, a DR2-Dw2 association is found only in Northern Europeans and in North Americans. T h e association is with DRw6 in the Japanese and DR4 in Jordanian Arabs and Italians. This diversity along, with the fact that experimental allergic encephalomyelitis (a murine experimental model for MS) is linked to I-A (the mouse equivalent of DQ) (Fritz et al., 1984), provides a rationale to search for a DQ polymorphism which would be shared by the different DRsusceptible haplotypes and would represent the genuine susceptibility factor. Several groups have reported from both population and family studies an increased frequency of a particular allele of DQwl. Fauchet used a DQ p probe and four enzymes (i.e., BamHI, BglII, EcoRI, and EcoRV) to identify two RFLP variants-DQw l a and DQwlb-the latter corresponding to the cluster DQR2-6 previously defined (Cohen et al., 1984). T h e DQwla subtype is associated with DRI-DR2 “short” (DRw16) and DR-BON, while the DQwlb subtype is associated with DR2 “long” (DRw15) and some DRw6 (DRwl8). In 100% of the cases (14 DR2 and one DRwG), the patients were found to be DQwlbe2A segregation analysis demonstrated that the association was stronger with the DQwlb allele than with the DR2 allele (Semama et al., 1988). DQw la is the same as DQw5; DQw 1 b is the same as DQw6.
148
DOMINIQUE CHARRON
An earlier population study had provided evidence for an association of the DQR2-6 cluster with MS (Marcadet et al., 1985a). This was confirmed and extended in a study of 61 patients from a Norwegian population (Vartdal et al., 1989).Two oligonucleotideswere used: one recognizing a common DQ pl sequence of DQw6 (DR2), DQw8, and DQw9, the second recognizing a sequence common to DQw6 (DRw18) and DQw6 (DRwl9). Among the 59 patients which were DR2, DR4, or DRw6 by serology, 59 carried shared DQ p l polymorphic sequences, which can be identified as amino acids 10-29,31-52,58-69, and 71-83 of the respective DQ pl chains. The two patients who were negative with two oligonucleotide probes were DR3-. These may be related to previous work identifying that the Al-B8-Dw3 haplotype was found in severe progressive MS, while the A3-B7-DR2 haplotype was associated with the more common milder disease (Madigand et al., 1982). Biochemical identification of the class I1 molecules involved in MS susceptibility has been attempted (Sriram et al., 1985),but did not provide any definitive conclusion at the molecular level with regard to the disease susceptibility element. An ongoing study in this laboratory has provided some new interesting possibilities. Two-dimensionalPAGE of the class I1 products from DR2-DQw l b haplotypes reveals that a common set of class I1 dimers is found. This includes a DQa-p dimer corresponding to the DQw 1 a-DQw 1 p subset and two additional dimers: DR a-DQw 1 j3 and DP a-DQwl p (Fig. 6). Thus, in the DR2-DQwl haplotypes of controls and of MS-affected patients, we describe interisotypic hybrid molecules (Hermans et al., 1990). In view of the previous data on DQwlb polymorphism, this implies that the MS susceptibility could be borne by DQ a-DQ p, but also by a DR a-DQ p, a DP a-DQ /3 molecule, or combinations of the three. Thus, even if the epidemiological linkage data stress only polymorphic residues on one chain (e.g., DQwl p), determinants which may be monomorphic on an a chain (of the same or a different isotype) could be involved in constructing the three-dimensional structure which is involved and functional in disease susceptibility. An RFLP study of patients with MS from Northern Ireland and Scotland mentions a polymorphism of the DQ a gene (DQa-MSp1 3.25-kb fragment) which was found in increased frequency in DR2 MS patients (Heard et al., 1989). Since this polymorphism appears to be allelic to the DQw1.2 a usually found in DR2-Dw2 haplotypes, it is likely that in the cells of MS patients who are DR2-Dw2 DQw1.2, the new DQ a fragment is contributed by the other haplotype. This transcomplementation raises the possibility that in these patients the determi-
HLA CLASS
I1 DISEASE ASSOCIATIONS
149
FIG.6. Interisotypic hybrid HLA class I1 molecules DRa-DQwIp and DPaDQw I@. [35S]rnethionine-labeledextract from a DR2-DW2-DQwG EpsteinBarr virus-transformed B cell line immunoprecipitated with G25a, an anti-DQw 1 p chain monoclonal antibody (A) The a chain area of the gel (isoelectrofocusing, two-dimensional polyacrylamide gel electrophoresis) reveals DR a,DQ a,and DP a chains. (B) The /3 chain area of the gel (nonequilibrium pH gradient electrophoresis, two-dimensional polyacrylamide gel electrophoresis) reveals a DQw 1 p chain.
nant responsible for the highest susceptibility to MS is present on an intraisotypic HLA-DQ hybrid molecule. T h e participation of DP in the disease susceptibility element is still debated. Two Scandinavian studies reported an association with DPw4 (Moen et al., 1984; Odum et al., 1988), which was confirmed in a Japanese population (N. Odum, personal communication). The absence of a linkage between DPw4 and DR2 would argue that the two disease susceptibility elements have independent contributions. However, a determinant could be created by a dimer consisting of a DQ cy and a DP /3 chain, both genetically linked to MS, although independent, which would represent the key element. Alternatively, a disease susceptibility element could be formed by the
150
DOMINIQUE CHARRON
DQ p chain (DQw 1 p) present in MS and a specific DP a chain. We await studies of DP (Y polymorphism in MS in order to conclude this point. The way in which the HLA disease susceptibility factor participates in the mechanisms leading to the disease is still largely unknown. It seems that inducibility of Ia expression on astrocytes is under MHC control, at least in a rat model of infection with a coronavirus, giving a syndrome similar to experimental allergic encephalomyelitis (Massa et al., 1987). It is thus of the utmost interest to explore in human brain ceils their capacity for class I1 expression and the genetic (i.e., regulatory polymorphism) and somatic (i.e., virus o r lymphokine) factors involved. The T cell response is also a critical parameter in inducing and/or controlling the disease. Several reports mention modification of the TcR /3 repertoire in the MS population (Seboun et al., 1989; Beall et al., 1989). Whether the TcR p anomaly is in itself a disease susceptibility gene or a gene in linkage disequilibrium with another predisposing gene is still debated. It is also possible that other TcR genes contribute to the genetic background which interact with unknown environmental agents to create the disease. VI. Concluding Remarks and Perspectives
In the past 15 years an enigmatic and almost impressionistic view of HLA and disease association has been clarified by structural definition of the class I1 molecules involved in disease susceptibility. T h e First International Symposium on HLA and Disease was held in Paris in 1976 (Dausset and Svejgaard, 1976) and was an essential contribution, drawing the attention of the medical and scientific community to the importance of the HLA system in disease and stimulating further research. At the time the genetics and statistics provided the hard data, while the immunology was speculative. T h e situation was confused, but McDevitt (1976) stated that the “future prospects seemed bright” in his concluding remarks and requested more precise typing techniques, while foreseeing mapping of the Ir genes and an unraveling of their structure: “Availability of such typing techniques will probably result in the detection of very strong associations between particular alleles of genes in the HLA-D region and particular diseases.” Indeed, the prediction is now superbly fulfilled. Clearly, the HLA class I1 molecules involved in disease susceptibility have a normal structure and may be found in the normal population, although at a lower frequency. T h e data reviewed here support the concept that the disease susceptibility element is composed of amino acids (contiguous and/or distantly spaced on the same chain, or on different chains of the same or a distinct class I1 isotype) which delineate a specific three-dimensional
HLA CLASS
I 1 DISEASE ASSOCIATIONS
151
conformation. It is suggested that it is more the conformational epitope created than the linear stretch of amino acids which is important (Fig. 7). Clearly, even when a portion of one class I1 chain appears to be predominant in disease susceptibility, additive and often synergistic contributions are found for a second chain. In conjunction with the fact that HLA class I1 molecules are a-P dimers capable of forming hybrid molecules (intra- or interisotypic) by CZS- and/or tram-complementation, this provides an ideal structural framework for the localization of the conformational disease susceptibility epitope(s). T h e capacity of the conformational epitope to bind within a high-affinity range set of defined peptides varies, presumably according to the polymorphism of the epitope (Buus et al., 1987). Whether it is always the same conformation of the class I1 molecule or a slightly different, but equivalent, structure which is involved in one disease is unknown. Moreover, as a peptide receptor, an MHC class I1 molecule could be flexible, allowing for some variation in the amino acid composition of the antigenic peptide(s). Many peptides (but not all) are likely to bind to the disease susceptibility element. This questions the nature of the antigen involved in disease susceptiblity in autoimmunity. There is no obvious reason to exclude that the antigen itself has some degree of polymorphism, as it binds to the appropriate disease-specific conformational epitope with sufficient affinity and is able to induce T cell activation. Future key studies include the definition of antigenic peptides, whether they are environmental (foreign) or autologous (self ). T h e MHC molecule in itself may provide a molecular tool in this search, since the antigen can be defined through its complementary conformational interactions with the class I1 molecule. Candidate peptides are presently being tested for their ability to bind class I1 disease susceptibility elements. The TcR represents the third partner of the functional trimolecular complex involved in the control of the immune response. Already in animal models of autoimmunity (e.g., experimental allergic encephalomyelitis), experimental evidence for a selection of the T cell diversity gives support to the idea that the T cell repertoire is important (AchaOrbea et al., 1988). It is, however, unknown whether this reflects a secondary state or precedes and induces the disease. The T cell repertoire is the result of germ-line diversity, followed by thymic selection. The expression of class I1 molecules on the thymus is set to eliminate MHC self-reactivity. However, this process may require specific amounts of any class I 1 molecule and sufficient duration of expression of a particular class I1 molecule in order to efficiently result in a full depletion of autoreactive T cells (self-MHC). Hybrid molecules may be expressed at a lower level than regular class I1 species and their expression may not be as equally constant as the typical a-p isotypes.
n 0
z n n
a a n
a
n
I
cy
g
n
Recent data from mice and preliminary results from this laboratory in humans suggest that there are favored a-p associations, while some are almost totally forbidden. This is determined by the affinity of different p chains for a given a chain. T h e end result may be that autoreactive T cells specific for class I1 hybrid molecules may not be totally depleted in all individuals. This may also depend on the local presence of lymphokines in the thymus during the selection process. T h e TcR repertoire would thus vary in different individuals, more susceptible individuals possessing a small amount of autoreactive T cells specific for the HLA class I1 hybrid molecule. This offers a tentative explanation for the concordance rate in twins, which is, at most, 50% in the HLA-associated diseases. Ultimately, the HLA class I1 molecules are the potential target of therapeutic strategies. This possibility was suggested when anti-Ia mAb's were shown to prevent or reverse induced, as well as spontaneous, autoimmune diseases in animal models. This concept permits the perspective that either blocking or competing antibodies or peptides may be tailored toward the disease susceptibility element and will therefore display high specificity and predictably high efficacy against the disease-associated immune response. Alternatively, antilymphokine reagents (e.g., antibodies, antagonists, and drugs) which could down-regulate HLA class I1 expression may also be considered. However, the exact mechanisms of such treatments are not known. They may block immune responses at the effector level, but may also solicit the induction of tolerance and/or the stimulation of suppression. Adverse potential effects mediated by immunological o r nonimmunological pathways should not be ignored. ACKNOWLEDGMENTS This work was supported by grants from Institut National de la Sante et de la Recherche Medicale, Ligue Nationale Francaise Contrele Cancer, Association de Recherche contre la SclCrose en Plagues, and Ligue Francaise contre la SclCrose en Plagues. I thank M. Brandel for her excellent typing and editing. N. Mooney and G. Blanck provided helpful comments. I am particularly grateful to J. Dausset for his constant support.
FIG.7. Molecular localization of the HLA class I1 epitopes involves in disease susceptibility. ?, Amino acids not precisely mapped; + , heterozygous effect. (trans-complementation). RA, Rheumatoid arthritis; IDDM, insulin-dependent diabetes mellitus; CD, celiac disease; MS, multiple sclerosis. Adapted from Bjorkman et al. (1987) and Brown et al. (1988).
154
DOMINIQUE CHAKKON
REFERENCES
Acha-Orbea, H, and McDevitt, H. 0.(1987).Proc. Natl. Acad. Sci. U.S.A. 84,2435. Acha-Orbea, H., Mitchell, D. J., Timmerrnann, L., Wraith, D. C., Tausch, G. S., Waldor, M. K., Zamvil, S. S., McDevitt, H. O., and Steinrnan, L. (1988). Cell 54, 263. Allison, J., et al. (1986).Nature (London) 333, 529. Andersson, G., Larharnrnar, D., Widmark, E., Servenius, B., Peterson, P. A., and Rask, L. ( 1 9 8 7 ) ~Biol. . Chem. 262,8748. Ando, A., Hidetoshi, I., Kimura, M., Ogata, S., and Tsuji, K. (1986). Hum. Immunol. 17, 355. Atar, D., Dyrberg, T., Michelsen, B., Karlsen, A,, Kofod, H., Molvig, J., and Lernrnark, A. (1989).J. Immunol. 143,533. Auffray, C., Lillie, J. W., Arnot, D., Grossberger, D., Kappes, D., and Strorninger, J. L. (1984). Nature (London) 308,327. Bach, F. H., Rich, S. S., Barbosa, J., and Segall, M. (1985).Hum. Immunol. 12,59. Beall, S. S., Concannon, P., Charrnley, P., McFarland, H. F., Gatti, R. A., Hood, L. E., McFarlin, D. E., and Biddison, W. E. (1989).J. Neuroimmunol. 21,59. Belfiore, A., Mauerhoff, T., Pujol-Borrell, R., Mirakian, R., and Bottazzo, G. F. (1986).J. Endocrinol. Invest. 9,61. Bell, J. I., Denny, D. W., and McDevitt, H. 0. (1985).Immunol. Rev. 84, 51. Bertrams, J., and Baur, M. P. (1984). “Histocompatibility Testing,” pp. 348-358. Springer-Verlag, Berlin. Betuel, H., Gebuhrer, L., Descos, L., Percebois, H., Winaire, Y., and Bertrand, J . (1980). Tissue Antigens 15, 231. Bidwell, J. L., Bidwell, E. A., Savage, D. A., Middleton, D., Klouda, P. T., and Bradley, B. A. (1988). Transplantation 45, 640. Bishop, G. A., and Frelinger, J . A. (1989).Abstr. Int. Congr. Immunol. 7th 17, 86. Bjorkrnan, P. J., Saper, M. A,, Sarnraoui, B., Bennett, W. S., Strorninger, J. L., and Wiley, D. C. (1987).Nature (London) 329,506. Bohme, J., Carlsson, B., Wallin, J., Moller, E., Person, B., Peterson, P. A., and Rask, L. (1986).J. Immunol. 137,941. Boitard, C., Bendelac, M., Richard, C., Cornaud, C., and Bach, J. F. (1988).Proc. Natl. Acad. Sci. U.S.A. 85, 9719. Bontrop, R. E.. Baas, E. J., Otting, N., Schreuder, G. M., and Giphart, M. J. (1987).Immunogenetics 25, 305. Bottazo, G. F., Dean, B. M., McNally, J. M., Mackay, E. H., Swift, G. F., and Gamble, D. R. (1985). N . Engl. J . Med. 313,353. Bottazo, G. F., Pujol-Borrell, R., and Gale, E. (1986).In “The Diabetes Annual” (L. P. Krall and M. M. Albert, eds.), p. 3. Elsevier, Amsterdam. Braunstein, N., and Germain, R. N. (1987).Proc. Natl. Acad. Sci. U.S.A. 84,2921. Brown, J. H., Jardetzky, T., Saper, M. A., Samraoui, B., Bjorkman, P., and Wiley, D. C. (1988).Nature (London) 332, 845. Bugawan, T. L., Angelini, G., Larrick, J., Auricchio, S., Ferrara, G. B., and Erlich, H. A. (1989).Nature (London) 339,470. Buus, S., Sette, A,, Colon, S. M., Miles, C., and Grey, H. M. (1987). Science 235, 1353. Carpenter, A. B., Kahl, L. E., Bartkowiak, C. D., Duquesnoy, R. J., and Eisenbeis, C. H. (1988).J. Rheumatol. 15, 395. Charron, D. (1986). Pathol. Biol. 34, 795. Charron, D. (1990).Adv. Electrophoresis 4, in press.
HLA
CLASS I1 DISEASE ASSOCIATIONS
155
Charron, D., and Fernandez, N. (1990). In “Two Dimensional Gel Electrophoresis” (T. Endler and S. Hanash, eds.), in press. VCH Publ. Weinheim, FRG. Charron, D., and McDevitt, H. 0. (1979).Proc. Natl. Acad. Sci. U.S.A. 76,6567. Charron, D., and McDevitt, H. 0. (1980).J. Exp. Med. 152, 18s. Charron, D., and Teyton, L. (1987). Concepts Immunopathol. 4,200. Charron, D., Aellen Schulz, M.-F., St. Gerne, J., Erlich, H. A., and McDevitt, H. 0. (1983).Mol. Immunol. 20,2 1, Charron, D., Lotteau, V., and Tunnel, P. (1984). Nature (London) 312, 157. Charron, D., Herrnans, P., Lotteau, V., Arnesland, F., Turrnel, P., and Teyton, L. (1990a). Immunol. Res. in press. Charron, D., Lotteau, V., Herrnans, P., Burroughs, D., Faille, A,, Ju, L. Y., and Teyton, L. (1990b). In “Molecular Biology of HLA Class I1 Antigens” (J. Silver, ed.), in press. CRC Press, Boca Raton, Florida. Cohen, D. (1989). In “Irnrnunobiology of HLA” (B. Dupont, ed.), in press. Springer-Verlag, New York.. Cohen, D., Legall, I., Marcadet, A., Font, M. P., Lalouel, J. M., and Dausset, J. (1984).Proc. Natl. Acad. Sci. U.S.A. 81, 7870. Cohen, D., Paul, P., Le Gall, I., Marcadet, A., Font, M. P., Haguenauer, O., Sayagh, B., Cann, H., Lalouel, J. M., and Dausset, J. (1985).Immunol. Rev. 85, 87. Cohen-Haguenauer, O., Robbins, E., Busson, M., Descharnps, I., Hors, J., Lalouel, J. M., Dausset, J., and Cohen, D. (1985). Proc. Natl. Acud. Sci. U.S.A. 82, 3335. Cornpston, D. A. S. (1982). Lancet 2,98. Coppin, H., Villedieu, P., Thornsen, M., Giraud, P., Carnbon-Thornsen, A., and Charron, D. (1987).Inamunogenetics 25, 130. Ddusset,J. (1981). Science 213, 1469. Dausset, J., and Svejgaard, A. (1976). “HLA and Disease” Munksgaard, Copenhagen. Dupont, B. (ed.) (1989). “Irnrnunobiology of HLA,” Vol. 1, Springer-Verlag, New York. Duquesnoy, R. J., Marrari, M., Hackbarth, S., and Zeevi, A. (1984).Hum. Immunol. 10, 165. Durandy, A., Fischer, A,, Charron, D., and Griscelli, C. (1986).Hum.Immunol. 16, 114. Eckels, D. D., Hartzrnan, R. J., and Johnson, A. H. (1986).J.Immunol. 136,2515. Erlich, H. A., Sheldon, E. L., and Horn, G. (1986).Biotechnol. Rev. 4,975. Fathman, C. G. (1980). Transplantation 30, 1. Festenstein H., Awad, J., Hitrnan, G . A., Cutbush, S., Groves, A., Cassel, P., Ollier, W., and Sachs, J. (1986).Nature (London) 322,64. Fukui, Y., Iwanaga, T., Inarnitsu, T., Nishirnura, M. Yasurnani, M. Kirnura, A., and Sasasuki, T. (1989). Abstr. Int. Congr. Immunol., 7th 17,633. Gerrnain, R. N., and Malissen, B. (1986). Annu. Rev. Irnrnunol. 4, 281. Gerrnain, R. N., and Quill, H. (1986). Nature (London) 320,72. Giles, R. C., DeMars, R., Chang, C. C., and Capra, J. D. (1985).Proc. Natl. Acad. Sci. U.S.A. 82, 1776. Gornard, E., Henin, Y., Sterkers G., Masset, M., Fauchet, R., and Levy, J. P. (1986).J . Immunol. 136,3961. Goronzy, J., Weyand, C. M., and Fathrnan, C. G. ( 1986).J. Clin. Invest. 77, 1042. Gorski, J., and Mach, B. (1986).Nature (London) 322,67.
156
DOMINIQUE CHARRON
Goyert, S. M., and Silver, J . (1983).Proc. Natl. Acad. Sci. U.S.A. 80,5719. Gregersen, P., Shen, M., Song, Q., Merryman, P., Degar, S., Seki, T., Maccari, J., Goldberg, D., Murphy, H., Schenzer,J., Wang, C., Winchester, R., Nepom, G., and Silver, J. (1986).Proc. Natl. Acad. Sea. U.S.A. 83,2642. Gustafsson, K., Wiman, K., Emmoth, E., Larhammar, D., Bohme, J., HyldigNielsen, J. J., Ronne, H., Peterson, P. A., and Rask, L. (1984). EMBO J. 3, 1655. Hanafusa, T. R., Pujol-Borrell, R., Ciovato, L., Russel, R. C. G., Doniach, D., and Bottazo, G. F. (1983).Lancet 2, 1 1 11. Hansen, G. S., Svejgaard, A., and Claesson, M. H. (1982).J.Immunol. 128,2497. Haziot, A., Lepage, V., Freidel, C., Betuel, H., Degos, L., and Charron, D. J. (1986).J. Immunol. 136,37. Heard, R. N. S., Cullen, C., Middleton, D., Hawkins, S. A., Francis, D. A., Hern, J. E. C., and McDonald, W. I. (1989).Hum. Immunol. 25, 1 1 1. Hermans, P., Ju, L. Y., Ziai, S., Teyton, L., and Charron, D. (1989a). In “Two Dimensional Gel Electrophoresis” (T. Endler and S. Hanash, eds.), p. 130. VCH Publ. Weinheim, FRG. Hermans, P., Merlu, B., Lotteau, V., and Charron, D. (1989b).Abstr. Int. Congr. Immunol., 7th 17,90. Hermans, P., Amesland, F., Fauchet, R.,and Charron, D. (1990). Manuscript in preparation. Holoshitz, J., Konig, F., Coligan, J., DeBruyin, J., and Strober, S. (1989).Nature (London) 339, 226. Horn, G. T., Bugawan, T. L., Long, C. M., and Erlich, H. A. (1988).Proc. Natl. Acad. Sci. U.S.A. 85,6012. Howell, M. D., Austin, R. K., Nepom, G. T., and Kagnoff, M. F. (1986).J. Exp. Med. 164,333. Howell, M. D., Smith, J. T., Austin, R. K., Kelleher, D., Nepom, G. T., Volk, B., and Kagnoff, M. F. (1988). Proc. Natl. Acad. Sci. U.S.A. 85,222. Jersild, C., Dupont, B., Fog, T., Plate, P., and Svejgaard, A. (1975). Transplant. Rev. 22, 148. Jones, P. P., Murphy, D. B., and McDevitt, H. 0. (1978).J. Exp. Med. 148,925. Kappes, D., and Strominger, J. L. (1988). “Annual Review of Immunology” p. 991. Palo Alto, California. Kim, S. J., Holbeck, S. L., Nisperos, B., Hansen, J. A., Maeda, H., and Nepom, G. T. (1985).Proc. Natl. Acad. Sci. U.S.A. 82, 8139. Knowles, R. (1989).In “Immunobiology of HLA” (B. Dupont, ed.), Vol. 1 , p. 365. Springer-Verlag, New York. Korman, A. J., Boss, J. M., Spies, T., Sorrentino, R., Okada, K., and Strominger J. L. (1985).Immunol. Rev. 85,45. Lee, S. H., Gregersen, P. K., Shen, H. H., Nunez-Roldan, A., Silver, J., and Winchester, R. J. (1984).Rheumatol. Int., Suppl. 4, 17. Lepage, V., Alcalay, D., and Douay, C. (1985). Tissue Antigens 25,96. Lo, D., Burkly, L., Widera, G., Cowing, C., Flavell, R.,Palmiter, R., and Brinster, R. (1988). Cell (Cambridge, Muss.) 53, 159. Lotteau, V., Teyton, L., Tongio, M. M., Soulier, A., Thomsen, M., Sasportes, M., and Charron, D. (1987a). Immunogenetics 45,403. Lotteau, V., Teyton, L., and Charron, D. (1987b).Nature (London) 329,339. Lotteau, V., Sands, J., Teyton, L., Turmel, P., Charron, D., and Strominger, J . (1989a).J. Exp. Med. 169,351.
HLA CLASS I1 DISEASE ASSOCIATIONS
157
Lotteau, V., Teyton, L., Turmel, P., and Charron, D. (1989b). In “Immunobiology of HLA” (B. Dupont, ed.), Vol. 2, p. 273. Springer-Verlag, New York. Lundin, K., Sollid, L., Qvigstad, E., Markussen, G., and Thorsby, E. (1989).Abstr. Int. Congr. Immunol. 7th 17,93. Madigand, M., Oger, J. F., Fauchet, R., Sabouraud, O., and Genetet, B. (1982). J. Neurol. Sci. 53, 5 19. Maeda, H., Juji, T., Mitsui, H., Sonozaki, H., and Okitsu, K. (1981).Ann. Rheum. Dis. 40, 299. Malissen, B., Shastri, N., Pierres, M., and Hood, L. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 3958. Marcadet, A., Cohen, D., Dausset, J., Fischer, A., Durandy, A., and Griscelli, C. (1985a).N . Engl. J . Med. 312, 1287. Marcadet, A., Massart, C., Semana, G., Fauchet, R., Sabouraud, O., Merienne, M., Dausset, J., and Cohen, D. (1985b).Immunogenetics 22,93. Markmann, J., Lo, D., and Naji, A. (1988).Nature (London) 336, 476. Massa, P. T., et al. (1987).J . Exp. Med. 166, 259. Matsushita, S., Muto, M., Svemura, M., Saito, Y., and Sasazuki, T. (1987). J . Immunol. 138, 109. Matsuydma, T., Winchester, R., Lee, S., Shookster, L., and Nunez-Roldan, A. (1988).J . Immunol. 140,537. Mayer, L., and Shalien, R. (1987).J. Ex$. Med. 166, 1471. McDevitt, H. 0. (1976). In “HLA and Disease” (J. Dausset and A. Svjegaard, eds.), p. 3 1 1 . Munksgaard, Copenhagen. Mengle-Caw, L., Connor, S., McDevitt, H. O., and Fathman, C. G. (1984).J. Exp. Med. 160, 1184. Merryman, P., Gregersen, P. K., Lee, S., Silver, J., Nunez-Roldan, A., Crapper, R., and Winchester, R. (1988).J . Immunol. 140,2447. Moen, T., Stien, R., Bratlie, A., and Bondevik, E. (1984). Tissue Antigens 24, 126. Monos, D. S., Spielman, R. S., Gogolin, K. J., Radka, S. F., Baker, L., Zmijewski, C. M., and Karnoun, M. (1987). Immunogenetics 26,299. Mooney, N., and Charron, D. (1988). Int. J. Immunopathol. Pharmacol. 1,51. Mooney, N., Grillot-Courvalin, C., Hivroz, C., and Charron, D. (1989).J.Autoimmun. 2,215. Neel, D., Merlu, B., Turpin, E., Rabourdin-Comble, C., Mach, B., Goussault, Y., and Charron, D. J. (1987). Biochem. J . 244,433. Nepom, B. S., Nepom, G. T., Mickelson, E., Antonelli, P., and Hansen, J. A. (1983).Proc. Natl. Acad. Sci. U.S.A. 80,6962. Nepom, G. T., Nepom, B. S., Antonelli, P., Mickelson, E., Silver, J., Goyert, S. M., and Hansen, J. A. (1983).J. Exp. Med. 159, 394. Nepom, B. S., Nepom, G. T., Mickelson, E., Schaller, J. G., Antonelli, P., and Hansen J. A. (1984).J. Clin. Invest. 74,287. Nepom, B. S., Palmer, J., Kim, S. J,, Hansen, J. A., Holbeck, S. L., and Nepom, ’ G. T. (1986).J. Exp. Med. 164, 345. Nepom, B. S., Schwartz, D., Palmer, J. P., and Nepom, G. T. (1987).Diabetes 36, 114. Nerup, J., Platz, P., Anderson, 0. O., Christy, M., Lyngsoe, J., Poulsen, J. E., Ryder, L. P., Thomsen, M., Staub Nielsen, K., and Svejgaard, A. (1974). Lancet 2, 864. Nishimoto, H., Kikutani, H., Yamamura, K., and Kishimoto, T. (1987). Nature 328,432.
158
DOMINIQUE CHARRON
Odum, N., Hyldig-Nielsen, J. J., Morling, N., Sandberg-Wolheim, M., Platz, P., and Svejgaard, A. (1988). Tissue Antigens 31,235. Ohta, N., Nishirnura, Y. K., Tanimoto, K., Horiuchi, Y., Abe, C., Shiokowa, Y., Abe, T., Katagiri, M., Yoshiki, T., and Sasazuki, T. (1982). Hum. Immunol. 5, 123.
Owerbach, D., Lernmark, A. A., Platz, P., Ryder, L. P., Rask, L., Peterson, P. A., and Ludvigsson, J. (1983). Nature (London) 303,815. Owerbach, D., Lernrnark, A., Platz, P., Ryder, L., Rask, L., and Peterson, P. (1984). Diabetes 33,958. Palacios-Boix, A. A., Estrada, G., Colston, M. G., and Panayi, G. S. (1988). J . Immunol. 141, 1844. Parham, P. (1988). Nature (London) 333,500. Pujol-Borrell, R., Todd, I., Doshi, M., Bottazzo, G. F., Sutton, R., Gray, D., Adolf, G. R., and Feldmann, M. (1987). Nature (London) 326,304. Reinsmoen, N., and Bach, F. (1982). Hum. Immunol. 4,249. Rodriguez de Cordoba, S., Marshall, P., and Rubinstein, P. (1989).Immunogenetics 29,49.
Roep, B. O.,et al. (1988). Hum. Immunol. 23, 27 1. Ronningen, K. S., Markussen, G., Iwe, T., and Thorsby, E. (1989). Hum, Immunol. 24, 165.
Roudier, J., Rhodes, G., Petersen, J., Vaughan, J. H., and Carson, D. (1988). Scand. J . Immunol. 27,367. Roudier, J., Petersen, J., Rhodes, G. H., Luka, J., and Carson, D. A. (1989). Proc. Natl. Acad. Sci. U.S.A. 86,5104. Sacks, S. H., Bushell, A., Rust, N. A., Karagiannis, J. A., Tewell, D. P., Ledingham, I. G. G., Wood, K. J., McMichael, A. T. (1987). Hum. Immunol. 20, 175.
Saiki, R., Scharf, S., Faloona, F., Mullis, K., Horn, G., Erlich, H., and Arnheim, N. (1985). Science 23, 1350. Saiki, R., Bugawan, T., Horn, G., Mullis, K., and Erlich, H. (1986). Nature 324, 163.
Sansom, D. M., Bidwell, J. L., Maddison, P. J., Carnpion, G., Klouda, P. T., and Bradley, B. A. (1987). Hum. Immunol. 19,269. Sansom, D. M., Bidwell, J. L., Klouda, P. T., and Amin, S. N. (1988). Lancet 1,58. Sant, A. J., Braunstein, N. S., and Germain, R. N. (1987). Proc. Natl. Acad. Sci. U.S.A. 84, 8065. Sarvetnick, N., Liggit, D., Pitts, S., Hansen, S., and Stewart, T. (1988). Cell (Cambridge, Mass.) 52,773. Sasazuki, T., Nishimura, Y., Muto, M., and Ohta, N. (1983).Immunol. Rev. 70,5 1. Schreuder, G. M. T., Tilanus, M. J., Bontrop, R. E., Bruining, G. J., Giphart, M. J., van Rood, J. J., and d e Vries, R. R. P. (1986).J. Exp. Med. 164,938. Seboun, E., Robinson, M. A., Doolittle, T. H., Ciulla, T. A., Kindt, T. J., and Hauser, S. L. (1989). Cell (Cambridge, Mass.) 57, 1095. Semama, G., Quillivic, F., Madigand, M., Merienne, M., Geneted, B., Sabouraud, O., and Fauchet, R. (1988). In “Trends in European Multiple Sclerosis Research” (C. Confavreux et al., eds.). Elsevier, Amsterdam. Seyfried, C. E., Mickelson, E., Hansen, J. A., and Nepom, G. T. (1987). Mol. Immunol. 24,47 1. Seyfried, C. E., Mickelson, E., Hansen, J. A., and Neporn, G:T. (1988). Hum. Immunol. 21, 289.
HLA CLASS
I1 DISEASE ASSOCIATIONS
159
Silver, J., Swain, S. L., and Hubert, J. J. (1980). Nature (London) 286,272. Singal, D., and Blajchrnann, M. (1973).Diabetes 22,429. Smith, G. R., Kunes, S. M., Schultz, D. W., Taylor, A., and Trinan, K. L. (1981). Cell (Cambridge, Mass.) 24,429. Sollid, L. M., Markussen, G., Ek, J., Gjerde, H., Vartdal, F., and Thorsby, E. (1989). I . Ex#. Med. 169,345. Song, Q. L.:Gregersen, P. K., Karr, R. W., and Silver, J. (1987).J . Immunol. 139, 2993. Srirarn, S., Stewart, G. J., Buhler, M., Grurnet, C., and Engleman, E. G. (1985). Neurology 35,248. Stasny, P. (1976).J. Clin. Invest. 57, 1148. Stasny, P. (1978). N . Engl. J . Med. 298,869. Steinrnetz, M., Stephan, D., and Lindahl, K. F. (1986). Cell (Cambridge, Mass.) 44, 895. Stewart, G. J., Basten, A., and Kirk, R. L. (1979). Tissue Antigens 14, 86. Svejgaard, A., Platz, P., and Ryder, L. P. (1980). I n “Histocompatibility Testing 1980” (P. I. Terasaki, ed.), p. 656. Springer-Verlag, New York. Svejgaard, A., Platz, P., and Ryder, L. P. (1983).Immunol. Rev. 70, 193. Teyton, L., Lotteau, V., Turrnel, P., Arenzana-Seisdedos, F., Virelizier, J. L., Pujol, J. F., Loyau, G., Piatier-Tonneau, D., Auffray, C., and Charron, D. J. (1987).J. Immunol. 138, 1730. Teyton, L., Lotteau, V., Viguier, M., Pujol, J. P., and Charron, D. J. (1990). Submitted for publication. Thornson, G. (1984).Am. J. Hum. Genet. 36, 1309. Tirnsit, J., Savino, W., Boitard, C., and Bach, J. F. (1989).J . Autoimmun. 2, 115. Tiwari, J. L., and Terasaki, P.I. (1985). “HLA and Disease Associations.” Springer-Verlag, Berlin. Todd, J. A., Bell, J. I., and McDevitt, H. 0. (1987). Nature (London) 329, 599. Todd, J. A., Acha-Orbea, H., Bell, J. I., Chao, N., Fronek, Z., Jacob, C. O., McDerrnott, M., Sinha, A. A., Tirnrnerrnan, L., Steinrnan, L., and McDevitt, H. 0. (1988a). Science 240, 1003. Todd, J. A., Bell, J., and McDevitt, H. 0. (1988b). Nature (London) 333, 710. Tonnelle, C., DeMars, R., and Long, E. 0. (1985). EMBO J . 4,2839. Tosi, R., Visrnara, D., Tanigaki, N., Ferrara, G. B., Cicimarra, F., and Buffolano, W. (1983). Clin. Immunol. Immunopathol. 28, 395. Tosi, R., Tanigaki, N., Polonco, I., Dernarchi, M., Woochov, J. G., and Hetzel, P. A. (1986). Clin. Immunol. Immunopathol. 39, 168. Travers, P., Blundell, T. L., Sternberg, M. J. E., and Bodrner, W. F. (1984).Nature (London) 310, 235. Trowsdale, J., and Kelly, A. (1985). EMBO J. 4,2231. Vartdal, F., Sollid, L. M., Vandvik, B., Markussen, G., and Thorsby, E. (1989). Hum. Immunol. 25, 103. Weyand, C. M., Goronzy, J., and Fathrnan, G. ( 1986). Proc. Natl. Acad. Sci. U.S.A. 83. 762. Winchester, R., and Gregersen, P. (1988). Springer Semin. Immunopathol. 10, 119. Wu, S., Thomas, L., Saunders, F., and Bach, H. ( 1 986). Nature (London)324,676. ~~
This article was accepted for publication on 20 October 1989.
ADVANCES IN IMMUNOLOGY, VOI.. 48
Neuroimmunology E. J. GOETZL, D. C. ADELMAN, AND S. P. SREEDHARAN Division of Allergy and Immunology, Department o f Medicine and The Howad Hughes Medical Institute, Univerrity of California Medical Center, San Fmncisco, California 94 143
1. Introduction
T h e alterations in immunity resulting from specific neural stimulation or lesions and the neuroendocrine adaptations to immune reactions suggested initially that these systems cooperate in host defense ( 1-5). Characterization of the patterns of innervation of immune organs and the findings of some similar membrane proteins, antigens, enzymatic pathways, mediators, and receptors in these diverse tissues supported the evolving concept of a multisystem network of recognition and responses. Innervation of immune organs by the sympathetic nervous system, including noradrenergic elements, positions nerve terminals on o r near lymphocytes and macrophages (6). The sympathetic neurons of the spleen and other lymphoid organs also contain peptides, such as enkephalin and neuropeptide Y (7). Sympathetic nervous supplies to the thymus are established in development, prior to the organization of the cells for immunologic function (8).The interactions between rat intestinal mucosal mast cells and subepithelial peptidergic neurons are extensive, about two thirds of the mast cells having neural contact (9).Electrolytic or pharmacological alterations in neural supplies alter quite specifically the number and the subset distribution of murine lymphocytes. Defined hypothalamic ablations reduced the number of large granular lymphocytes, the activity of natural killer cells, the T helper : T suppressor cell ratio, and the level of circulating B cells (10, 11). T h e obliteration of a single mediator-specific population of neurons, such as those containing substance P (SP) by capsaicin treatment, reduced splenic lymphocyte responses in immunized rats (12). At the level of cellular constituents, some surface glycoproteins with immunoglobulin (1g)-like structure are expressed on both neural and immune cells (13, 14). Some of these plasma membrane glycoproteins, such as 1 B 236/myelin-associated glycoprotein, resemble the neural cell adhesion molecule (termed N-CAM), whereas others, such as contactin, 161 Copyright 8 1990 by Academic Press. Iiic. All rights of reproduction in any foroi reserved.
162
E. J. GOETLL ET AL.
are unique interneural adhesion factors (15, 16). T h e relationship between molecules suggested by identity of a small segment of protein structure has provided additional examples of neuroimmune similarities, such as the sharing of an epitope by myelin basic protein (MBP) and the T cell plasma membrane CD3 molecule (17). Cells of common origin, such as the macrophage and neural microglial cell, appear to maintain functional similarities in both antigen presentation and neural repair and regeneration (18, 19). The expression of major histocompatibility complex (MHC) antigens in the nervous system, in contrast to the immune system, is rare in the resting state, with the exception of endothelial cells, microglia, and some pericytes (20). A variety of stimuli induce the presentation of MHC class 11, and occasionally class I, proteins on astrocytes, microglia, and oligodendrocytes, which enables some of these cells to present antigen and become targets for cytotoxic T cells (20, 2 1). Immune cytokines, such as y-interferon (IFN-y) (22), and some viral infections (23) are potent inducers of neural cell MHC antigens. The level of the MHC class I1 antigens on neural cells is regulated similarly to that on immune cells by a range of biochemical signals (24). Hyperinducibility of MHC class I1 antigens on astrocytes is one index of the susceptibility of different strains of mice to the development of experimental allergic or, more properly, autoimmune encephalomyelitis ( 2 5 ) . The capacity of neuromediators to influence immune responses is an area of intense investigation, especially in view of the findings of functionally significant concentrations of some neuropeptides at sites of immune and inflammatory reactions (26). The neural delivery of sympathetic amines and some neuropeptides enhances systemic and regional immune responses in rodents (27, 28), but the specificity of such effects has not been defined in relation to other activities of the neuromediators. &-Adrenergic mediators contribute to the tissue injury of adjuvant arthritis in rats, and the most critical cellular targets appear to be elements of the immune system (29). Recent studies have shown that classical models of audiovisual conditioning permit physical challenges to activate mast cells and recruit mast cell mediators for neurally evoked responses (30). This chapter explores new findings in several areas of neuroirnmunology. The specificity of T cell receptors for neural antigens, which elicit inflammatory demyelinating reactions in the central nervous system, has been established recently and may represent a useful focus for therapy of autoimmune disorders. Immune cells produce factors considered earlier to be distinctively neural in origin, and they also respond to a wide range of neuromediators. Inversely, neural cells respond to stimulation by
NEUROIMMUNOLOGY
163
immunological mediators and generate immune factors. Subsets of receptors for structurally diverse mediators have been found in both systems. Th e molecular and cellular characteristics of neuroimmune responses suggest that some elements of each system are uniquely sensitive to influences of the other. II. T Cell Components of Autoimmune Reactions to the Nervous System
Distinctive characteristics of immune responses to neural tissue antigens have been elucidated by detailed studies of inflammatory demyelination in animal models and human diseases. Experimental autoimmune encephalomyelitis (EAE) is the predominant model of cellular immune diseases of the central nervous system, in which neural tissue invasion by lymphocytes leads to demyelination and paralysis (31). CD4' T helper (TH)cells specific for MBP are the primary mediators of this disease in rats and mice (32, 33). The histopathological features of demyelination and the resultant paralysis are induced in irradiated or nonirradiated rats by adoptive transfer of a small number of activated CD4' T cells, bearing interleukin-2 (IL-2) receptors, which are derived from affected rats immunized with MBP and then stimulated with MBP in vitro prior to transfer (34). T h e absense of a prominent infiltrate of unspecific lymphocytes in the irradiated rats, despite the similarity of demyelination in the two groups, confirmed the requirement for only a small number of MBP-specific CD4' lymphocytes. In contrast, classical cutaneous delayed-type hypersensitivity reactions are evoked by MBP only in nonirradiated rats receiving MBP-specific CD4' T cells, in which the total numbers of host and donor lymphocytes infiltrating cutaneous sites and neural lesions are far greater than in lethally irradiated rats (34). T cell clones developed from strains of mice susceptible to EAE, which respond to MBP in vitro, shared specificity for MHC protein as well as antigen, and induced EAE when injected intravenously into naive mice (33). T h e efficiency of adoptive transfer of EAE to rats was augmented by the presence of macrophages or IL-1 during in vitro stimulation with MBP (35)(Fig. l), presumably by facilitating activation of MBP-specific CD4' T cells. That a small number of MBP-specific and activated CD4+ T cells is critical to the primary pathogenesis of EAE was also confirmed by the preventive effect of monoclonal anti-CD4' in recipient rodents (36). More complex pathogenetic mechanisms may be involved in some forms of EAE and other neural autoimmune diseases. Although lacking the central features of an allergic reaction, EAE is blocked by agents that suppress the contributions of mast cell mediators. Some inhibitors of
MBP-Reactive Helperlhducer T CeNs
-
cr
I,
Ellector Mechanisms
I
Inducer T
factors
xlt":. KCD4 1i 4
1.-1
\ Anlr
MacrophagesIlL I
la
Galactocerebroside Cytotoxic T Cells
. Anti-Glycoprotein * - Mast Cell-Derived
- Inflammation
.Demyelination
Mediators
I I I
FIG. 1. Mediation and regulation of pathogenetic reactions in experimental autoimmune encephalomyelitis,q?, a chains, and ft,/3 chains of the T cell receptor with different specificities;THa, THb, subsets of helper T cells reactive with myelin basic protein (MBP); dashed arrow, stimulation or facilitation; wavy arrow, inhibition.
mast cell d e g r a d a t i o n and antagonists of the effects of histamine and serotonin prevented the development of passively transferred EAE in rats, without influencing lymphocyte functions (37). Neural cell-specific cytotoxic T cells also may contribute to autoimmune reactions by direct damage with or without MHC restriction (38) (Fig. 1). Analyses of the basic immune mechanisms of CD4' T cell-mediated demyelination in EAE have been the most informative, however, and have defined the dominant epitopes of the MBP antigen, the selective use of T cell receptor V genes, and the favorable effects of specific tolerance to MBP and of antibodies to the relevant T cell receptors. A. DOMINANT EPITOPES OF MBP AND OTHERNEURALAUTOANTIGENS In the mouse strains B1O.PL and PL/J ( H - 2 " ) and SJLlJ ( H - 2 ' ) , which are susceptible to the induction of EAE by MBP, the major and minor encephalitogenic epitopes of MBP have been identified using substituent synthetic peptides (39, 40). The vast majority of MBP-reactive T cell clones and hybridomas created by fusions of such T cell lines from the H-2" strains are specific for the amino-terminal nine amino acids of MBP (41,42). The minor epitopes, which become important after suppression of the response to the immunodominant epitope in several systems, also have been mapped with synthetic peptides of autologous and heterolo-
NEU ROIMM UNOLOC Y
165
gous MBP (43). In the SJL/J ( H - 2 ’ ) strain encephalitogenic determinants appear to be more complex and are composed of at least two nested epitopes of MBP, which may each be recognized by a genetically distinct T cell receptor (44). This pattern contrasts with the predominant use of only one T cell receptor for the response of the H-2p strain of mice to the major epitope of MBP (4 1, 42). Neurotropic viruses may induce or facilitate the development of EAE or other inflammatory demyelinating diseases in animal models by at least three immunological mechanisms. The first is the molecular mimicry that occurs when viral and host neural proteins share an immunogenic segment of amino acid sequence. One example is the six-aminoacid sequence found in both the encephalitogenic domain of rabbit MBP and hepatitis B virus polymerase, which elicited blood mononuclear leukocytes and antibodies reactive with both proteins as well as a histopathological state similar to EAE in the immunized rabbits (45) (Fig. 1). The second mechanism is the predisposition to EAE induced in otherwise resistant strains of mice by a previous viral infection (Fig. 1). Immunization with MBP or transfer of lymph node cells from MBP-primed syngeneic donors evoked typical EAE in adult C57BL/6 mice that had recovered from Semliki Forest virus infection, but were without effect on normal mice or those inoculated with noninfectious Semliki Forest virus (46). Whether the second challenge led to the expansion of preexisting clones of lymphocytes reactive to MBP or a heightened response by other pathways remains to be determined. The third type of virally induced autoimmune damage to neural tissues is mediated principally by “innocent bystander” mechanisms. In one such model susceptible mouse strains inoculated intracerebrally with Theiler’s murine encephalomyelitis virus (TMEV) develop a chronic infiltrate of mononuclear leukocytes in association with demyelination of the central nervous system (47). CD4’ (L3T4’) T cells predominated in the perivascular regions, but only CD8’ (Lyt-2’) T cells invaded the parenchyma (47). Virally infected EAE-susceptible SJL/J mice lacked CD4+ T cells capable of specifically mediating MHC class 11-restricted immune responses to MBP, proteolipid protein of myelin, o r spinal cord homogenates, but exhibited such T cell responses to TMEV and related viruses (48). Thus, the virally infected mice did not manifest a direct autoimmune reaction to myelin. That the immune response to TMEV and resultant neural tissue inflammation were responsible for the observed demyelination was suggested by the capacity of immunosuppressive drugs, antithymocyte sera, and monoclonal anti-la to decrease the extent of demyelination in mice of MHC-determined susceptibility (34, 36, 38,49).
166
E..J. GOETZL ET Al,.
B. RESTRICTED USEOF T CELLRECEPTOR V GENESI N EAE: EFFECTSOF ANTIBODIES TO THCELLRECEPTORS The current level of understanding of the molecular biology of T cell receptors and the consequent availability of immunochemical and genetic probes for specific domains of the receptors have permitted analyses of T cell receptor genes of encephalitogenic T H cells in mice (40-42). The T cell receptors of eight MBP-specific encephalitogenic T cell clones from five different PL/J or PL/J X SJL mice, which recognize MBP( 1-9) in association with I-Ap, used the genes predominantly of V p 8 and exclusively of one Va gene segment (41) (Fig. 1). Similarly, the T cell receptors of 33 clonally distinct B1O.PL THcells specific for MBP(1-9) and I-A4 used only two VCYgenes and two VP ( Vp8.2 and V p l 3 ) genes and showed the same level of restricted heterogeneity of Ja and J p gene use (42). Significant differences in the frequency of use of selected Va,Ca,and V p genes by T cell receptors have been observed for patients with multiple sclerosis as compared to normal subjects (50). In both studies of THcells from two different mouse strains, monoclonal antibodies to Vp8 completely blocked the recognition of MBP(1-9) by Vp8 THcells in vitro and prevented and reversed the EAE induced by MBP (4 1,42). C. OTHERIMMUNOMODULATING APPROACHES TO THE INFLAMMATORY OF EAE DEMYELINATION Although the development of EAE can be prevented by the administration of anti-Ia or anti-CD4 antibodies prior to MBP (35, 39) (Fig. l), the encephalitogenic subset of CD4' T cells was shown to be relatively refractory to mouse monoclonal anti-CD4 (35). The acute and relapsing EAE evoked in mice by the introduction of encephalitogenic CD4' T cells was delayed, but not prevented, by a single dose of monoclonal anti-CD4, which significantly depleted CD4' T cells without suppressing the total number of lymph node cells responding proliferatively to MBP in nitro (35). The modest effectiveness in suppressing EAE of monoclonal antibodies to lymphocyte surface proteins, other than those of the T cell receptor, led to investigations of other T cell-directed and antigendependent pathways of immunomodulation. Antiergotypic T cells, which recognize the state of activation rather than the idiotype of the encephalitogenic CD4' T cells, are induced by immunizing rats with activated syngeneic T cell clones that do not bind MBP (51). Rats treated with antiergotypic T cells were resistant to adoptively transferred EAE and, to a lesser extent, active EAE (51). Monoclonal antibody to T suppressor cell factors with inducer activity in a 10-day course reduced the incidence and severity of EAE elicited by
NEUROIMMUNOLOGY
167
MBP in mice (52). The MBP-specific Ts cells in mice treated with monoclonal antibody to the inducer T suppressor cell factors inhibited THcell responses to MBP (52). T h e induction of specific tolerance to MBP, as another method for reducing the severity of EAE, has been achieved with orally administered MBP (53), neural tissue proteins coupled to splenocytes (54), and the synthetic a-acetylated amino-terminal peptide of MBP coupled to splenocytes (55). The suppression of EAE in rats by orally administered MBP or synthetic substituent peptides of MBP revealed a greater effect of nonencephalitogenic substituent decapeptides than those capable of inducing EAE, suggesting additional subdivisions of molecular domains (53, 54). Neuroantigen-specific tolerance in adult mice, susceptible to MBP-induced EAE, often led to the suppression of delayed-type hypersensitivity in parallel with that of EAE (55).
111. Production and Functional Effects of Cytokines and Other Mediators of Neuroimmunological Reactions
Although considerable evidence had suggested previously that cells of the nervous and immune systems produce and respond to similar signals of several types, results of recent studies have revealed additional levels of molecular and cellular diversity. The definition of the structures for some mast cell- and leukocyte-derived variants of neuropeptides, the binding properties and protein components of lymphocyte receptors for some neuromediators, and the cellular localization and specificity of leukocytic neuropeptidases have begun to elucidate complex mechanisms for genetic adaptation and functional flexibility in such multisystem networks. For example, extracts of fluid and tissue samples from sites of immediate and delayed-type hypersensitivity reactions contain neuropeptide immunoreactivity, which is often chromatographically heterogeneous, as it is composed of intact authentic neuropeptides from nerves and immune cells, as well as genetically determined variants from immune cells and multiple peptidolytic products. A structural variant of a neuropeptide from the immune system not only has a distinctive cellular source, but may differ from the authentic neuropeptide in activity and potency, as a result of the unique specificity and affinity of the subset of receptors in each system. The multiple pathways of production of various mediators and their different effects in the immune and neuroendocrine systems are considered here, and the properties of the corresponding receptors are described in Section IV.
168
E. ,J. (;OKL'ZL ET AL.
A. NEURALGENERATION OF IMMUNOLOGICAL CYTOKINES Blood monocyte-derived microglial cells of the central nervous system expectedly exhibit many immune synthetic capabilities of other mononuclear phagocytes. Microglia also process antigens, express characteristic MHC class I1 and other surface antigens, and act as antigen-presenting cells (56, 57). It was initially more surprising to discover that astrocytes and a variety of glial cells also serve as sources of immunological mediators (Table I). Astrocytes from rats and humans do not normally carry surface MHC class I1 (Ia) antigens in culture, but are stimulated by IFN-y to express Ia and to present antigen in Ia-restricted interactions with T cells (58, 59). T h e elaboration of IL-1 by astrocytes also activates T cells to a state of more efficient immune function (60). Astrocytes stimulated by lipopolysaccharide also generate and release IL-3-like activity
TABLE I NEURALSOURCES OF IMMUNOLOGICAL MEDIATORS" Mediator
IL- 1
Astrocytes
+
Glioblastoma cells
+
IL-2 IL-3
-I-
+ +
+ +
IL-6 TNF-a IFNs
IFN-a/P
ND IFN-(u/P
GM-CSF
+
ND
+
TGF-j3 Membrane proteins MHC class I MHC class I1 IFN-.)I receptors
+b +b
+
V V
Method of identification Immunoassay, functional effects Immunoassay, functional effects Functional effects, size on gel filtration Functional effects Functional effects Functional effects, immunoassay Functional effects, immunoneutralization Functional effects, structure Flow cytometry, function Flow cytometry, function Function
IL, Interleukin; TNF-a, tumor necrosis factor-a; IFNs, interferons; GMCSF, granulocyte-macrophage colony-stimulating factor; TGF-j3, transforming growth factor-@;MHC, major histocompatibility complex; +, present; -, absent; ND, not done; V, varies with each cell line. Present only after stimulation.
'
NEU KO1 M M UNOLOC; Y
169
capable of enhancing the growth of microglial and other cells (61). Astrocytes incubated with tumor necrosis factor-a (TNF-a) release granulocyte-macrophage colony-stimulating factor (GM-CSF), which supports hemopoietic colony growth in culture, where anti-GM-CSF neutralizes approximately 70% of this activity (57). Astrocytes stimulated maximally by lipopolysaccharide or ionophores also serve as a source of TNF-a, which exerts cytotoxicity both for its characteristic cellular targets and for rat oligodendrocytes (62). The IFN activity produced by astrocytes, but not neurons, after superinduction with poly(rI.rC), cycloheximide, and actinomycin D consists of IFN-a and -p, but not -y, as assessed by function and immunoreactivity (63). T h e quantities of IFN-a and -p released achieve concentrations that elicit the expression of MHC class I antigens on one subset of astrocytes (63),which thus becomes susceptible to T cell cytotoxicity. Diverse other neural cells in culture also elaborate ILs (64). Cultured lines of murine, rat, and human glioblastoma cells also secrete IL-1, IL-3, IL-6, and IFN-a and -p (65-68) (Table I). In addition, another protein not found in astrocyte cultures exhibited T cell suppressive activity in vitro and was named “glioblastoma-derived T cell suppressor factor” (69). This factor was subsequently shown to be structurally identical to transforming growth factor-p (69). Human glioma cells also generate a monocyte chemotactic factor, which is structurally identical to the lymphocyte-derived chemotactic factor (70). Although only limited studies have been undertaken, several immune cytokines are found at significant concentrations in neural fluids from mice and humans with viral meningitis and encephalitis (71). Both IL-6 and IFN-y were detected, but at different times, with the early appearance of IL-6 and later peaks for IFN-y. That cytokine levels were lower in similarly infected athymic mice suggested that T cells are a major source of both IL-6 and IFN-y, which may be critical to local immunological reactions to the virus (7 1).
B. IMMUNOLOGICAL GENERATION OF NEUROENDOCRINE MEDIATORS Neuropeptides appear in lesional tissues and fluids of hypersensitivity and inflammatory reactions at concentrations and times suggesting primary contributions to pathogenesis (72-74). Nasal lavage fluids obtained serially for up to 24 hours from allergic patients after nasal challenge with antigen showed significant increases in calcitonin gene-related peptide (CGRP) and somatostatin (SOM), but not SP (72).In contrast to the rapid and transient rises in histamine concentration, those of SOM were highest at 6 hours, and those of CGRP reached maximal levels at u p to 24 hours after challenge. In the delayed-type hypersensitivity reaction
TABLE I1 IMMUNOLOGICAL SOURCES OF NEUROMEDIATORS" Neuromediator
=; 0
Authentic neuroendocrine peptide
Cellular source
ACTH
T lymphocytes
$&Endorphin
T lymphocytes
Growth hormone
Lymphocytes, leukemic cells
T h yrotropin
T lymphocytes
CRF
T lymphocytes
VIP
Eosinophils
SP
Eosinophils
Structure
Method of identification
Function, immunoreactivity, peptide structure, mRNA analysis Function, immunoreactivity, peptide structure Immunoreactivity, peptide structure, mRNA analysis Immunoreactivity, peptide structure Function, immunoreactivity, peptide structure Immunoreactivity, peptide structure Immunoreactivity, peptide structure, mRNA analysis
Variant neuropeptide VIPs
-2
SOMs
Mast cells, monocytes
VIPl0-28 free acid, amino-terminally extended VIPs Larger than SOMPS
Enkephalins
Lymphocytes, macrophages, mast cells
Amino-terminally extended en kephalins
Mast cells Lymphocytes Mast cells
vIP15-28, VIPzi-28, vIPz3-28 VIP4-28, VIPl5-28. VIP23-28 SPi-7, Sps-ii
T lymphocytes
18-kDa monomer, dimer
T lymphocytes
12-15 kDa
Cleavage products of neuropeptides VIP VIP SP Protein growth factor GGPF Oligodendrocyte growth factor
Mast cells
Immunoreactivity, peptide structure, mRNA analysis Immunoreactivity, amino acid composition Immunoreactivity, peptide structure
Functional effects, chromatography Functional effects, chromatography
a ACTH, Adrenocorticotrophic hormone; CRF, corticotropin-releasing hormone; VIP, vasoactive intestinal polypeptide; SP, substance P; SOMs, somatostatins; GGPF, glial growth-promoting factor.
172
E. J. COETZL E T
AL.
evoked by IL-2 and lymphokine-activated killer cell therapy of peritoneal carcinomatosis, ascitic fluid concentrations of SP and CGRP increased markedly to maximal levels at 8-72 hours (74). The SP, CGRP, and SOM from nasal and peritoneal fluids were largely the authentic neuropeptides, as determined by chromatographic analyses of immunoreactivities. Lymphocytes generate adrenocorticotrophic hormone (ACTH), pendorphin, thyrotropin, and corticotropin-releasing hormone (CRH), and eosinophils produce vasoactive intestinal polypeptide (VIP) and SP indistinguishable from the corresponding neuroendocrine peptides (75-78). Other studies in vitro have elucidated the nature of variant neuropeptides from nonneural sources. The VIP-related peptides from mast cells and basophils (79), SOM-like peptides from monocytes and mast cells (80), and enkephalinlike peptides from lymphocytes, macrophages, and mast cells (8 1) differ structurally from previously defined neuropeptides (Table 11). The immunological equivalents of SOM28and Met-enkephalin appear larger than the neuropeptides, but the amino acid sequences have not been completed. VIPs from rat basophilic leukemia cells and some cultured lines of murine mast cells consist of VIPl0-28 free acid and a mixture of amino-terminally extended “big” VIPs (79) (Table 11),which appear to be derived from a novel preproVIP encoded by an alternatively spliced mRNA. T h e isolation of distinctive fragments of VIPI-2s from suspensions of lymphocytes, mast cells, and other leukocytes after incubation in vitro confirmed the involvement of posttranslational peptidolysis (82) in the structural diversity of some VIPs in immunological responses (Table 11). The recent findings that genetic and proteolytic variants bind differently to lymphocyte than to neural receptors suggest that each member of the VIP family may have a unique immunoregulatory role (83). More recent genetic studies have partially delineated the messages for growth hormone and ACTH in human and rodent mononuclear leukocytes. The bioactivity, molecular weight, and antigenicity of mononuclear leukocyte-derived growth hormone indicate identity with pituitary growth hormone (Table 11). The maximal level of mRNA for growth hormone in the cytoplasm of blood mononuclear leukocytes was observed after 4 hours of incubation, and had an M , of 1.0 kbp (84). mRNA specific for proopiomelanocortin (POMC) was observed in normal human blood mononuclear leukocytes and leukocytes of lymphoid and myeloid malignancies (85).This message was translated into at least three forms of immunoreactive POMC-ACTH, exhibiting a chromatographic pattern similar to that of pituitary POMC-ACTH. The stimulation of
NEUKOIMMUNOLOGY
173
release of P-endorphin from B lymphocytes in blood mixed mononuclear leukocytes requires monocytes. The monocytes respond to stimuli as diverse as CRH and arginine vasopressin by generating IL-1, which is the actual mediator of P-endorphin release from B lymphocytes. IL- 1 thus acts as the final common signal for diverse stimuli of B lymphocyte production of POMC-ACTH (86). Immune cells that generate a neuropeptide in vitro may be the principal nonneural source of the neuropeptide in complex tissue reactions. The SP generated in granulomas of murine schistosomiasis is largely attributable to the eosinophils, as demonstrated by the localization of SP immunoreactivity and mRNA (87, 88). T cells produce proteins that influence the growth and differentiation of various neural cells (Table 11). In a few instances the protein factors have been identified and partially characterized. Some appear to be directed to astrocytes (89, 90) and others, to oligodendrocytes (90, 91). The glial growth-promoting factor is a 30-kDa dimer natively, which is composed of disulfide-linked 18-kDa subunits, and both forms are biologically active. T h e oligodendrocyte growth factor from T cells has an apparent size of 12-15 kDa and is immunochemically distinct from glial growth-promoting factor, as well as from other defined growth factors.
C. NEURALEFFECTS OF IMMUNOLOGICAL FACTORS Many types of neural cells express a range of antigens and respond functionally to mediators from the immune system (1-5). The patterns of neural responses observed may be considered to be grouped into four categories, based on the chemical nature of the immunological signal, the time course of the neural response, and whether the effect is neurally restricted or involves other systems (Fig. 2). The first two categories encompass the effects of immunologically derived cytokines on neural cell growth, survival, and differentiation that are either direct or the result of alterations in generation or reception of a neurally derived cytokine. Direct stimulation of growth and proliferation have been observed with a T and B cell-derived glial-stimulating factor and TNF for astrocytes (89, 90, 92), IL-1 for astrocytes and glial precursors (92, 93), IL-2 and glial growth-promoting factor for oligodendrocytes (93, 94), and IL-3 for microglia (93). Direct activation and differentiation of oligodendrocytes are elicited by IL-1 and IL-2, of pituitary cells by IL-1, and of astrocytes by IFN-y (93). IL-2 induces cellular maturation of oligodendrocytes, as assessed by biochemical criteria, and evokes the increased production of myelin (93). T h e critical trophic role of nerve growth factor (NGF) in determining sympathetic and some types of sensory innervation is reflected in the
174
E. J. COETZL ET AL.
FIG.2. Immunological mediation of neural cell activities.ACTH, Adrenocorticotrophic hormone; GGPF, glial growth-promoting factor; GSF, glialstimulating factor; IFN-y, y-interferon; IL, interleukin; MHC,major histocompatibility complex; NGF, nerve growth factor; Oligod., oligodendrocyte; PAF, platelet-activating factor; PEF, proliferation-enhancing factor; ST, somatotropin; TNF, tumor necrosis factor; TSH, thyrotropin; solid arrow, dashed arrow, wavy arrow, generation or secretion; stimulation or facilitation; inhibition.
correlation of the density of innervation with preceding tissue levels of NGF. In this second category increases in NGF synthesis during embryonic development and in the sciatic nerve after disruptive lesions depend on the arrival and activation of macrophages and their local production of IL-I (95).IL-1 augments the neural levels of NGF mRNA by increasing both transcription and stability of the message (95). This second category thus represents principally an indirect effect of the immune cytokine on tissue innervation. Another example of the second category is the cleavage of ingested myelin membranes by macrophages, which results in the generation of a polypeptide capable of evoking Schwann cell proliferation in vitro (96). In contrast to the slow persistent effects of immunological cytokines on neural cell growth and differentiation, their acute effects on the secretion of pituitary hormones influence multiple neuroendocrine and other systems substantially. Picomolar-nanomolar TNF and IL- 1 act directly on pituitary cells to stimulate the release of ACTH, growth hormone or somatotropin (ST),and thyrotropin, with or without suppression of pro-
NEUROIMMUNOLOGY
I75
lactin secretion (97, 98). One mechanism for this activity in the third category was suggested by the finding that IL-I enhances the expression of mRNA coding for POMC in pituitary cells (99). T cells appear not to be involved in IL-1 effects in the pituitary gland, as the magnitude of the endocrine response is as great in athymic mice (100). IL-1 also prepares some pituitary cells of cultured lines for much greater release of ACTH and @-endorphin by CRH and VIP, as a result of induction of protein kinase C (101). IL-1 also stimulates the synthesis of SOM in hypothalamic cells, in association with the reduced secretion of ST and thyrotropin, but the mechanisms for the delayed induction of SOM synthesis remain to be elucidated (102). Finally, in some rat models of inflammatory arthritis, the responses of plasma concentrations of ACTH and corticosteroids to IL-1 and CRH are impaired significantly, suggesting that this defect may play a role in determining susceptibility to the development of arthritis (103). In the fourth category of neural effects of immunological factors, low-molecular-weight mediators elicit immediate functional responses that occasionally persist or change qualitatively with time. Activation of mast cells in neural tissues alters the excitability of sensory C fibers and modifies synaptic transmission in sympathetic and parasympathetic ganglia (104). T h e results of application of exogenous histamine and pharmacological inhibitors suggest that histamine H- 1 action accounts for much of the observed effects. Platelet-activating factor rapidly activates the PC12 and NG108-15 lines of cultured neural cells, as reflected in immediate increases in the intracellular concentrations of [Ca2'], the acute release of ATP from PC 12 cells, and later neuronal differentiation of NG108-15 cells (105). Two distinct isomers of 8,15-dihydroxyeicosatetraenoic acid, which are generated by polymorphonuclear leukocytes and epithelial cells during inflammatory reactions, rapidly modify the threshold of cutaneous C fiber nociceptors (106, 107). The 8(R) isomer decreases and the 8(S) isomer increases the C fiber threshold to a variety of stimuli (107).
D. IMMUNOLOGICAL EFFECTS OF NEUROENDOCRINE MEDIATORS Neuroendocrine factors initiate and modulate some elements of all phases of immune responses and inflammatory reactions, but are especially potent mediators of macrophage and lymphocyte functions. I . Immediate Hypersensitivity and Acute Inflammation T h e introduction of endogenous neuromediators into tissues (by excitation of regional peptidergic neurons) or of exogenous synthetic neuromediators evokes responses resembling allergic and inflammatory reac-
176
E. J. GOETZL. E T A L .
tions in skin, lungs, gut, and joints (73, 108). The detection of nanomolar concentrations of several neuropeptides in allergic and inflammatory fluids confirmed the possibility of neural contributions. Further investigations have recently elucidated the nature and some mechanisms of the effects of neuropeptides on mast cells, basophils, and polymorphonuclear leukocytes in vitro (Fig. 3), but have also demonstrated low potency for almost all of the direct actions of neuropeptides on these cells. SP and a wide variety of other neuropeptides elicit the release of preformed and newly generated mediators from mast cells, at concentrations of 0.1-30 optimal potency being a function of the species and the distinct subset of mast cell (109). The specificity of mast cell activation is dependent on one structural domain of each peptide, suggesting the existence of a dedicated set of receptors, and independent of IgE and most of the early transductional steps coupled with IgE receptors. The low apparent potency of neuropeptide stimuli of mast cell activation may be attributable to the proteases released from the mast cells, which cleave most of the neuropeptide into less active fragments (1 10). Basophils are not susceptible to activation by most neuropeptides. The greatest potency of neuropeptides in immediate hypersensitivity is at the level of modulation of the effectiveness of IgE-dependent stimulation of basophils and mast cells (Fig. 3). At picomolar-nanomolar concentrations SOM inhibits by up to 80% the IgE-dependent release of histamine and LTD4 from basophils (111). Inhibition by SOM of mediator release from some types of mast cells, activated by anti-IgE or IgE antibody and antigen, is observed at 0.1-100 nM concentrations of SOM (1 12). Nanomolar levels of NGF enhance the release of histamine from IgE antibody-sensitized rat peritoneal mast cells stimulated by antigen or other secretagogues, as a result of apparent increases in the expression of mast cell receptors (113). As for mast cells, polymorphonuclear leukocytes are only chemotactically and biochemically responsive to 0.1-10 phf SP and some other neuropeptides ( 1 14) (Fig. 3). The principal effect of SP, as a representative neuropeptide, is to modulate the responses of polymorphonuclear leukocytes to more potent stimuli rather than to exert direct effects. At concentrations as low as 0.1 nM, SP enhances both the chemotactic and HnOe-generating responses of neutrophils to N-formyl-methionylleucyl-phenylalanine and C5a by 50-100% (1 15). The neuropeptide a-melanocyte-stimulating hormone (a-MSH), when administered intraperitoneally to mice in amounts as low as 0.1 nmol, suppresses neutrophil influx into subcutaneously implanted sponges containing IL-1, TNF, or C5a (116). Again, a-MSH lacked any direct inflammatory activity, but acted solely to modulate the responses to other mediators (Fig. 3).
w,
N EU K O I M M UNOLOGY
177
---_____---
FIG. 3. Neuropeptide mediation of functions of immunological cells. 'CT , Immunoglobulin E (IgE) antibody; -, antigen; solid arrow, chemotaxis; dashed arrow, stimulation or facilitation; wavy arrow, inhibition. Ab, antibody; CGRP, calcitonin gene-related peptide; CT M4, cytotoxic macrophage; fMLP, formylmethionyl-leucyl-phenylalanine; IFN-y, y-interferon; IL, interleukin; a-MSH, a-melanocyte-stimulating hormone; NGF, nerve growth factor; NK, neurokinin; N T , neurotensin; PL, prolactin; PMN, polymorphonuclear leukocytes; PP, phagocytosable particle; SK, substance K; SOM, somatostatin; SP, substance P; ST, somatotropin; TH,T helper cell; Thy, thymocyte; TNF-a, tumor necrosis factor-a; Ts, T suppressor cell; VIP, vasoactive intestinal polypeptide.
2 . Chronic Inflammation and Lymphocyte Functions Neuropeptides are potent mediators of many activities of macrophages a n d numerous types of lymphocytes (Fig. 3). SP is o n e of the most
selectively active chemotactic factors for monocytes and macrophages, as picomolar concentrations elicit maximal responses in vitro, in contrast to the micromolar concentrations required to stimulate neutrophil chemotaxis and activation (1 17). Other neuropeptides alter macrophage biochemical pathways, with or without the involvement of lymphocytes. Picomolar-eanomolar concentrations of @-endorphinsuppress the production of 0 2 a n d H202 by human monocytes incubated with phorbol myristate acetate or opsonized zymosan.
178
E. J. GOETZL E T AL.
In contrast, growth hormone enhances porcine macrophage oxidative metabolic responses in vitro and activates in vivo the responses of macrophages from hypophysectomized rats to a level similar to that achieved by IFN-y (118, 119). The generation of H202 by human macrophages and their function as antigen-presenting cells are inhibited by preincubation with picomolar concentrations of CGRP and calcitonin (120). In vivo, lymphocytes may be the target of some neuropeptide effects on macrophage functions. Inhibition of prolactin secretion in mice by bromocriptine prevents helper T cell induction of macrophage cytotoxicity and concomitantly suppresses lymphocyte proliferation and production of IFN-y, with reversal of all effects by exogenous prolactin (121). In some instances the reduction in generation of IFN-y results from cellular cooperation, as when P-endorphin evokes macrophage secretion of both prostaglandin E2 and oxygen metabolites that suppress T cell production of IFN-.)I(122). IL- 1 and neuropeptides interact functionally in several immunological reactions. The generation of IL-1, TNF-a, and IL-6 by human blood monocytes is induced by SP and substance K (SK), with maximal effects at concentrations of 2 10 nM of either neuropeptide (123). At similar concentrations SP and neurokinins A and B evoke the production of IL- 1 by cultured mouse macrophages (124). Neuropeptides that lack a primary effect on IL- 1 synthesis may substantially alter the generation and release processes, as for the enhancing activity of neurotensin (125). At another level of regulation, 10-100 nM a-MSH selectively blocks the stimulation of thymocytes and fibroblasts by IL-1 in vitro and the inflammatory effects of IL-1 and TNF in mice (116, 126). An array of neuropeptides have major effects on the differentiation, tissue homing and cycles, and immunological functions of lymphocytes (4,5). The regional accumulation of one or more subsets of lymphocytes, with distinct receptors and functional capabilities, in an organ system with special host defense requirements is accomplished by directed migration, local differentiation, and a variety of specific adherence mechanisms. Neurophysiological events influence each of these processes in embryonic development and during compartmental immune responses. The effects of VIP on T cell distribution is one example of recent elucidation of these processes. VIP-immunoreactive nerve fibers in mice are concentrated in the internodular regions of Peyer's patches and mesenteric lymph nodes, in close proximity to the endothelium of high endothelial venules within these patches, but not in spleen or subcutaneous lymph nodes (127). High-affinity receptors for VIP are found on murine T cells, with a predominance on the CD4+,rather than CD8+,subset, and can be down-regulated by exposure to VIP (127). Such a loss of VIP receptors from T cells of mesenteric lymph nodes altered the homing
NEUROIMMUNOLOGY
179
back to gut-associated lymphoid tissue of syngeneic recipients, due to a decrease in the rate of transfer from blood to the lymphoid tissue of origin, as contrasted with other immunological organs (127). Thus, the migration and retention of one subset of CD4' T cells are influenced critically by local delivery of the VIP neuropeptide and recognition by the T cell receptors for VIP. T h e most recently demonstrated effects of neuropeptides and other neuromediators include the development and functional priming of lymphocyte subsets, the substitution for one type of regulatory T cell, and the unmasking of autoimmune reactivity. Low concentrations of P-endorphin permit antigen-directed production of anti-Herpes antibodies by human blood mononuclear leukocytes from some subjects, where antigen alone has no effect (128). P-Endorphin and Metenkephalin facilitate the generation of cytotoxic T cells from murine spleen cells in one-way mixed-lymphocyte reactions. Arginine vasopressin replaces helper T cells in the generation of IFN-.)Iby Lyt-2' mouse splenic lymphocytes (129). Arginine vasopressin is recognized by mouse lymphocytes through one of its receptors, which exhibits a novel profile of affinities for defined antagonists (130). Estrogens and other potent steroid hormones mediate deletion of an intermediate developmental stage of T cell receptor-, CD3-, and IL-2 receptor-negative lymphocytes acutely and then both earlier and later forms of regulatory T cells after long-term treatment of mice (131). The production of antibodies to autologous erythrocytes by mouse CD5' B cells is augmented by estrogens, as a result of more efficient production by each responding B cell (132). T h e synthesis of Ig by mouse splenic and mesenteric lymph node lymphocytes increases in parallel with proliferative responses to mitogens after an intravenous infusion of SP, with the greatest effect on IgA, as expected from the results of in vitro studies (133). Thus, the stimulatory effects of SP and the inhibitory effects of VIP and SOM on diverse lymphocyte functions are observed in vivo, just as they were originally described in vitro. The finding that neuropeptides had their most potent influence on mononuclear leukocytic activities suggests the importance of elucidating the properties of lymphocyte receptors for neuropeptides and defining any variants of neuropeptides recognized more efficiently by such receptors than by the receptors of neuroendocrine cells. IV. lymphocyte Receptors for Neuromediators
T h e neuroendocrine modulation of immune functions necessitates the expression of specific receptors for neuroendocrine peptides on cells of the immune system. The following have been characterized in terms of
180
E. J. GOETZL ET AL.
the affinity and specificity of binding of the respective neuroendocrine peptide: opioid peptide receptors on human monocytes, granulocytes, and lymphocytes; SP and human ST receptors on cultured human IM-9 B lymphoblasts; prolactin receptors on natural killer cells; NGF receptors on rat mononuclear spleen cells; and ACTH receptors on human mononuclear leukocytes. A clone of immune cells may express more than one set of neuropeptide receptors, each of which has different structural and transductional properties. The IM-9 line of cultured human B lymphoblasts has receptors for SP, ST, and insulin (134-136). Ligand affinity cross-linking under similar conditions has shown proteins of 33, 58, 78, and 116 kDa with SP (134), 120 kDa with ST (135), and 95 and 135 kDa with insulin (136). The ST and insulin receptor proteins of IM-9 lymphoblasts resemble those of endocrine cells in size and carbohydrate constituents. Conversely, the receptors for some neuropeptides on different types of lymphocytes, appear to be similar in affinity and specificity. T and B cells and natural killer cells, represented by large granular lymphocytes, all bear receptors for prolactin with nanomolar dissociation constant values. In addition, low concentrations of cyclosporine A increase and high concentrations decrease the binding of prolactin to B and T cells, and natural killer cells similarly (137). The responsiveness of T cells to neural as well as nonneural mediators varies with the stage of lymphocyte development and activation. The proliferative activity of cloned T cells exposed to IL-2 is associated with greater CAMPincreases in response to histamine, prostaglandin E2, and P-adrenergic agents when challenged at later stages in the activation process (138). On an evolutionary ladder, as well as in any one species, the specificity of receptors for some neuromediators is similar in the neuroendocrine and immune systems. The receptors for cholecystokinin and gastrins in the brain are id&&cal in specificity to the respective receptors in pancreatic tissues at each evolutionary stage. At the level of divergence of endoderms from reptiles, both the cholecystokinin and gastrin receptors develop new specificity in neural and gastrointestinal tissues (139). The inhibitory activity of progesterone and some other steroids for m receptors is the same in brain and lymphoid tissues of mammals (140). ACTH high-affinity receptors on adrenal cells and mononuclear leukocytes are not only similar in number and binding affinity normally, but vary together in disease states (141). For example, a child with glucocorticoid deficiency due to ACTH insensitivity lacked high-affinity receptors for ACTH on mononuclear leukocytes as well as adrenal cortical cells (141). The three prototypic receptors for tachykinins were characterized initially by their localization in guinea pig ileal tissue, vas deferens and
NEUKOIMMUNOLOGY
181
bladder, and mammalian nervous system (142). Extensive studies with radiolabeled agonists and antagonists demonstrated their respective specificities for SP, SK, and neuromedin K (143). The SK receptor, but not that for SP, has been cloned recently from bovine stomach (144) and shown to be a member of the family with seven transmembrane domains, which includes adrenergic and muscarinic cholinergic receptors ( 144). Progress is evident in studies of the structure of the SP receptor on the IM-9 line of human cultured B lynphoblasts (134). Covalent affinity cross-linking of 1251-labeledSP to IM-9 B lymphoblasts labels a series of 33,58, 78, and 116 kDa proteins, in contrast to the 46-kDa protein of rat brain SP receptors derived by photoaffinity labeling with (3"251D-Tyr', 41-N3 Phe', N1e")-labeled SP (145). Perhaps of greater immunological significance is the finding of a much higher percentage of lymphocytes bearing SP receptors in murine gastrointestinal tissue than in the spleen and blood, as determined by flowcytometric quantification of binding of fluorescein-labeled SP ( 146). Means of 53% and 34% of the T cells-and 65% and 47% of the B cells-in Peyer's patches and the spleen, respectively, bound SP, as contrasted with fewer than 20% in blood. It has not been established whether SP and other neuropeptide-binding lymphocytes are attracted to and accumulate in these tissues o r whether they develop the receptors after tissue localization. T h e largely inhibitory effects of SOM and VIP on lymphocytes, as contrasted with the enhancing activities of SP, are also initiated by highaffinity receptors on both B and T cells (Table 111). T h e binding of ('*"I-Tyr'")-labeled VIP and ( 1251-Tyr")-labeledSOM to T and B cells of cultured lines is quantified at room temperature, in the presence of the peptidase inhibitors Dt-thiorphan and phenylmethylsulfonylfluoride, and bound peptide is separated from unbound peptide by centrifugation of the lymphocytes through a cushion of n-butyl phthalate and dinonyl phthalate (7 : 2 v/v) (83, 147). The PEER, Jurkat, and Molt-4 lines of cultured Tcells show high levels of specific binding of SOM, and the latter two lines also show evidence for VIP binding. In contrast, a greater variability is observed in studies of B cell lines. T h e MAK Epstein-Barr virus-infected human B cell line, the U266 IgE-producing human myeloma line, and the 5558 IgA-producing murine myeloma line exhibit high levels of specific binding of SOM. These lymphoid cells also demonstrate specific VIP binding, as d o the Dakiki IgA-producing human lymphoblasts and the U937 human monocyte line, both of which failed to bind SOM. T h e 2F.1 1.15 and S49 lines of B cells do not bind either SOM or VIP. Two representative human cell lines, the IL-2-secreting Jurkat leuke-
182
E. .J. GOETZL E 7 A L .
TABLE I11 HUMAN LYMPHOCYTE RECEPTORS FOR NEUROPEPTIDES' Lymphocyte Substance P Blood T cells (small subset) IM-9 B lymphoblasts Somatostatin U266 myeloma cellsb Isk B lymphoblasts Jurkat T cellsb MT-2 leukemic T cells Molt4F T lymphoblasts Vasoactive intestinal polypeptide Blood T cellsb (small subset) Molt-4b T lymphoblasts Jurkat T cells U266 myeloma cells Dakiki myeloma cells Nalm 6 pre-B cells a
Receptors/cell (mean x 35
Kd
(mean, nM) 180
2.3
0.6
1.31590 6.1 0.11290 6.5 0.3
0.005/ 100 1.1 0.003166 0.6 0.2
1.7/ND
0.5/80
15 12 41 156 156
7
5.2 7.6 9.1 13
ND, Not determined. Subsets of two different affinities.
mic T cells and the IgE-secreting U266 myeloma B cells, were studied by flow cytometry for binding of fluoresceinated VIP and SOM. The majority of the cells of both lines recognize SOM and VIP. The presence of an excess of unlabeled SOM or VIP prevents more than 50% and 40%, respectively, of the corresponding fluorescent labeling of the cells. The specific binding of ( 1251-Tyr11)-labeled SOM or ( 1251-Tyr10)-labeled VIP to Jurkat and U266 cells at 22°C is time dependent and can be reversed with the addition of excess unlabeled SOM and VIP, respectively. Computer-based Scatchard analyses of the competitive inhibition of binding of ( 1251-Tyr1')-labeled SOM to Jurkat and U266 cells, by different concentrations of unlabeled SOM, revealed 10' and lo3 high-affinity sites with K d values of 3 and 5 pM, respectively (147). A large number of low-affinity sites for SOM were also identified on both cell lines, with Kd values of 66 and 100 nM, respectively. The SOM receptors of the MT-2 adult T cell leukemia line, the Molt4F human T cell line, and the IsK Epstein-Barr virus-transformed B cell line have been studied in similar detail and have been shown to be of one class each, with respective Kd values of 0.64, 1.1, and 0.22 nM (148).
NEU KO1M M U N OLOG Y
183
T h e results of studies of the competitive inhibition of [ '251-Tyr1n]VIP binding to Jurkat and U266 cells by different concentrations of unlabeled VIP indicate the presence of lo4 sites, with mean Kd values of 5.2 and 7.6 nM, respectively (83). T h e VIP receptors of human blood mixed T cells, Molt-4b lymphoblasts, Nalm 6 human pre-B cells, and Dakiki plasma cells have respective mean Kd values of 0.47,7.3, 12.6, and 9.1 nM (33, 149). Neither U266 nor Jurkat cells bind SP. T h e VIP receptor protein of Molt4b lymphoblasts has a size of 47 kDa, resembling that of the liver, pancreas, and pituitary gland, but often is recovered in larger forms, due to oligomerization or association with guanine nucleotidebinding proteins (149). Analogs of SOM, including the naturally occurring 28-amino-acid SOM 28, mono-iodinated Tyr' '-labeled SOM, and (D-Trp', D-CYS)labeled SOM, also inhibit the binding of ( '251-Tyr11)-labeledSOM to both cell lines (147). Similarly, the principal mast cell-derived variant VIPlo-28, as well as the L-8-K and the (Ac-Tyr', D-Phe2)-labeledGRP1-29 amide peptide analogs of VIP displace ( '251-Tyr11)-labeledVIP from Jurkat and U266 cells (83).VIPIO-2' exhibits an affinity one tenth that of V1P1-28 for lymphocyte receptors, whereas it has only one tenthousandth the affinity of VIP,-28 for neural receptors of rodent tissues, which is consistent with tissue-specific receptors. SP, CGRP, and VIP failed to competitively inhibit the binding of [ 1251-Tyr1']SOMto Jurkat and U266 cells, respectively, whereas SP and SOM have no effect on binding of [ 1251-Tyr10]VIP to either cell line. As is found for SP receptors, more lymphocytes in tissues than in blood express receptors for SOM. Means of 49% and 35% of T cells-and 52% and 37% of B cells-in murine Peyer's patches and the spleen, respectively, bear receptors for fluorescent SOM, as assessed by flow cytometry (62). The presence of receptors for VIP, SOM, and other neuromediators on nonneural cells in the nervous system and other tissues raises the possibility that a neuropeptide may stimulate a nonneuronal cell to generate a neuroactive secondary mediator in a multicell response pattern. One example is the neurotrophic survival-promoting substances for neurons derived from astrocytes exposed to VIP (71) or monokines. T h e interaction of VIP with its receptor is known to elevate CAMP levels by stimulating adenylate cyclase in plasma membrane preparations of human blood mixed T cells and Molt-4b T lymphoblasts (149). In the same population of T cells, VIP also activates a CAMP-dependent protein kinase that phosphorylates a 41-kDa membrane protein. While it has been demonstrated that VIP significantly alters the synthesis of IgA and IgM, but not IgG, by mixed lymphocytes from murine spleen and Peyer's patches, there is no additional information about the functions of B cell receptors for VIP or the effects of mast cell-derived variants of VIP on T
184
E. J. COETZL ET AL.
and B lymphocytes. SOM and VIP receptors expressed on the same cell may exhibit functional interactions, as in the pituitary line of cultured GH& cells. SOM consistently and rapidly inhibits the stimulation of CAMPby VIP, and this inhibition is blocked in some types of cells by pertussis toxin, indicating a functional role for G; proteins. V. Conclusions
Molecular analyses of immune recognition and effector mechanisms, mediators, and receptors for mediators in neuroimmunological reactions have revealed additional levels of biological specificity and diversity not apparent from initial studies of the systemic and cellular events. T h e distinctive characteristics of immune responses to some neural antigens include the restricted use of T cell receptor V genes, modulation by neurotrophic viruses, and involvement of macrophage- and mast cellderived mediators. An increased understanding of these complex pathogenetic pathways has permitted the development of novel therapeutic approaches, such as the suppression of autoimmune encephalomyelitis by antibodies to subsets of T cell receptors. Immune cells produce variants of neuropeptides, not detected in the nervous system, both by alternative splicing of mRNA and by unique proteolytic cleavages of the preproneuropeptides. T and B cell receptors for neuropeptides differ in specificity, affinity, and transductional mechanisms from those in the nervous system. T h e immunologically specific variants of neuropeptide mediators and receptors for neuropeptides provide for the restriction of responses to the immune system, as well as for opportunities for immunological interactions with the nervous system. T h e evolving elucidation of the mechanisms of neuroimmune reactions promises both to reveal new molecular structures and cellular pathways of fundamental biological importance and to pave therapeutic avenues applicable to many immunological diseases. ACKNOWLEDGMENTS The studies cited were supported in part by grant PO1 A119784 from the National Institute of Allergy and Infectious Diseases. We are grateful to Dr. James L. Urban for his critical review of Section 11.
REFERENCES 1. Jankovic, B. D., and Isakovic, K. (eds.) (1973). “Microenvironmental
Aspects of Immunity.” Plenum, New York.
2. Ader, R. (ed.) (1981). “Psychoneuroimmunology.”Academic Press, New
York. 3. Blalock, J. E., and Smith, E. M. (1985).Fed. Proc., Fed. Am. SOC.Exp. B i d . 44, 108.
NEUROIMMUNOLOGY
185
4. Berczi, I. (ed.) (1986).“Pituitary Function and Immunity.” CRC Press, Boca Raton, Florida. 5. Goetzl, E. J., and Spector, N. H. (eds.) (1989). “Neuroimmune Networks: Physiology and Diseases.” Liss, New York. 6. Madden, K. S., Ackerman, K. D., Livnat, S., Felten, S. Y., and Felten, D. L. (1989).I n “Neuroimmune Networks: Physiology and Diseases” (E. J. Goetzl and N. H,. Spector, eds.), p. 1 . Liss, New York. 7. Fried, G., Terenius, L., Brodin, E., Efendic, S., Dockray, G., Fahrenkrug, J., Goldstein, M., and Hokfelt, T. (1986). Cell Tissue Res. 243,495. 8. Bulloch, K., Cullen, M. R., Schwartz, R. H., and Longon, D. L. (1989). J. Neurosci. Res. 18, 16. 9. Stead, R. H., Tomioka, M., Quinonez, G., Simon, G. T., Felten, S. Y., and Bienenstock, J. (1987). Proc. Natl. Acad. Sci. U.S.A. 84,2975. 10. Belluardo, N., Mudo, G., Cella, S., Santoni, A,, Forni, G., and Bindoni, M. (1987).Immunology 62,32 1. 1 1 . Katayama, M., Kobayashi, S., Kuramoto, N., and Yokoyama, M. M. (1987). Ann. N.Y. Acad. Sci. 496,366. 12. Nilsson, G., and Ahlstedt, S. (1988). Int. J . Immunopharmacol. 10, 747. 13. Salzer, J. L., Holmes, W. P., and Colman, D. R. (1987).J. CellBiol. 104,957. 14. Peault, B., Chen, C. H., Cooper, M. D., Barbu, M., Lipinski, M., and LeDouarin, N. M. (1987).Proc. Natl. Acad. Sci. U.S.A. 84,814. 15. Milner, R., I,ai, C., Sutcliffe, J. G., and Bloom, F. E. (1989).I n “Neuroimmune Networks: Physiology and Diseases” (E. J. Goetzl and N. H. Spector, eds.), p. 9. Liss, New York. 16. Ranscht, B. (1988).J. CellBiol. 107, 1561. 17. Jingwu, Z., Weber, W. E. J., Borst, J., Vandenbark, A. A., and Raus,J. C. M. (1989).J. Immunol. 142, 3917. 18. Perry, V. H., and Gordon, S. (1989).I n “Neuroimmune Networks: Physiology and Diseases” (E. J. Goetzl and N. H. Spector, eds.), p. 119. Liss, New York. 19. Perry, V. H., Brown, M. C., and Cordons, S. (1987).J. Exp. Med. 165,1218. 20. Hauser, S. L., and Hayashi, T. (1989).I n “Neuroimmune Networks: Physiology and Diseases” (E. J. Goetzl and N. H. Spector, eds.), p. 17. Liss, New York. 21. Manning, P. T., Johnson, Jr., E. M., Wilcox, C. L., Palmatier, M. A., and Russel, J. H. (1987). Am. J. Pathol. 128, 395. 22. Wong, G. H. W., Bartlett, P. F., Clark-Lewis, I., Battye, F., and Schrader, J. W. (1984). Nature (London) 310,688. 23. Massa, P. T., Brinkman, R., and Ter Meulen, V. ( 1 9 8 7 ) ~Exp. . Med. 166, 259. 24. Frohman, E. M., Vayavegula, B., Gupta, S., and Van Den Noort, S. (1988). Proc. Natl. Acad. Sci. U.S.A. 04, 1292. 25. Massa, P. T., T e r Meulen, V., and Fontana, A. (1987).Proc. Natl. Acad. Sci. U.S.A. 84,42 19. 26. Tissot, M., Pradelles, P., and Giroud, J. P. (1988). Inflammation 12, 25. 27. Fujiwara, R., and Orita, K. (1987).J . Immunol. 138,3699. 28. Helme, R. D., Eglezos, A., Dandie, G. W., Andrews, P. V., and Boyd, R. L. (1987).J . Immunol. 139, 3470. 29. Levine, J. D., Coderre, T. J., Helms, C., and Basbaum, A. I. (1988). Proc. Natl. Acad. Sci. U.S.A. 85,4553.
186
E. J. GOETZL ET AL.
30. MacQueen, G., Marshall, J., Perdue, M., Siegel, S., and Bienenstock, J. (1989).Science 243,83. 31. Paterson, P. Y.,and Halberg, M. K. (1989).I n “Neuroimmune Networks: Physiology and Diseases” (E. J. Goetzl and N. H. Spector, eds.), p. 25.Liss, New York.
32. Pettinelli, C. B., and McFarlin, D. E. (1981).J. Immunol. 127, 1420. 33. Zarnvil, S.,Nelson, P., Trotter, J., Mitchell, D., Knobler, R., Fritz, R., and Steinman, L. (1985).Nature (London) 317, 355. 34. Sedgwick, J., Brostoff, S., and Mason, D. (1987).J. Ex$. Med. 165, 1058. 35. Sriram, S., Carroll, L., Fortin, S., Cooper, S., and Ranges, G. (1988). J . Immunol. 141,464. 36. Dietsch, G. N., and Hinrichs, D. J. (1989).J. Immunol. 142, 1476. 37. Mannie, M. D., Dinarello, C. A., and Paterson, P. Y. (1989).J . Immunol. 138, 4229. 38. Jewtoukoff, V.,Lebar, R., and Bach, M. A. (1989).Proc. Natl. Acad. Sci. U.S.A. 86,2824. 39. Fritz, R. B.,Jen-Chou, C. H., and McFarlin, D. E. (1983).J. Immunol. 130, 191. 40. Urban, J. L., Kumar, V., Kono, D. H., Gomez, C., and Hood, L. (1989).In “Neuroimmune Networks: Physiology and Diseases” (E. J. Goetzl and N. H. Spector, eds.), p. 25.Liss, New York. 41. Acha-Orbea, H., Mitchell, D. J., Timmermann, L., Wraith, D. C., Tausch, G. S., Waldor, M. K., Zamvil, S. S., McDevitt, H. O., and Steinman, L. (1988).Cell (Cambridge, Mass.) 54,263. 42. Urban, J. L., Kumar, V., Kono, D. H., Gornez, C., Horvath, S. J., Clayton,J., Ando, D. G., Sercarz, E. E., and Hood, L. (1988).Cell (Cambridge, Mass.) 54, 577. 43. Hashim, G. A., and Day, E. D. (1987).J.Neurosci. Res. 18,209. 44. Sakai, K., Sinha, A. A., Mitchell, D. J., Zamvil, S. S., Rothbard, J . B., McDevitt, H. O., and Steinman, L. (1988).Proc. Natl. Acad. Sci. U.S.A. 85, 8608. 45. Fujinami, R. S.,and Oldstone, M.B. A. (1985).Science 230, 1043. 46. Mokhtarian, F., and Swoveland, P. (1987).J. Immunol. 138, 3264. 47. Lindsley, M. D.,and Rodriguez, M.(1989).J. Immunol. 142,2677. 48. Miller, S. D., Clatch, R. J., Pevear, D. C., Trotter, J. L., and Lipton, H. L. (1987).J . Immunol. 138,3776. 49. Paya, C. V., Patrick, A. K., Leibson, P. J., and Rodriguez, M. (1989). J. Immunol. 143,95. 50. Oskenberg, J. R., Sherritt, M., Begovich, A. B., Erlich, H. A., Bernard, C. C., Cavalli-Sforza, L. L., and Steinman, L. (1989).Proc. Natl. Acad. Sci. U.S.A. 86,988. 51. Lohse, A.W., Mor, F., Karin, N., and Cohen, I. R. (1989).Science 244,820. 52. Perry, L. L. (1988).J. Zmmunol. 141,3016. 53. Higgins, P. J., and Weiner, H. L. (1988).J. Immunol. 140,440. 54. Kennedy, M. K., Dal Canto, M. C., Trotter, J. L., and Miller, S. D. (1988). J . Immunol. 141, 2986. 55. Clayton, J. P., Gammon, G. M., Ando, D. G., Kono, D. H., Hood, L., and Sercarz, E.E.(1989).J . Exp. Med. 169, 1681. 56. Guilian, D., and Baker, T. J. (1986).J . Neurosci. Sci. 6,2163. 57. Frei, K., and Fontana, A. (1989).I n “Neuroirnmune Networks: Physiology and Diseases” (E.J. Goetzl and N. H. Spector, eds.), p. 127.Liss, New York.
NEUHOI MM U N O LOG Y
187
58. Hirsch, M. R., Wietzerbin, I., Pierres, M., and Goridis, C. (1989).Neurosci. Lett. 41, 199. 59. Fontana, A., Fierz, W., and Wekerle, H. (1984). Nature (London) 307, 273. 60. Fontana, A., Kristensen, F., Dubs, R., Gernsa, D., and Weber, E. (1982). I . Immunol. 129, 2413. 61. “Frei, K., Bodrner, S., Schwerdel, C., and Fontana, A. (1986).J. Immunol. 137.3521. 62. Robbins, D. S., Shirazi, Y., Drysdale, B. E., Liberman, A., Shin, H. S., and Shin, M. L. (1987).J. Immunol. 139,2593. 63. Borgeson, M., Tallent, M. W., and Keane, R. W. (1989).In “Neuroimmune Networks: Physiology and Diseases” (E.J. Goetzl and N. H. Spector, eds.), p. 51. Liss, New York. 64. Roberge, F. B., Caspi, R. R., and Nussenblatt, R. B. (1988).J. Immunol. 140, 2193. 65. Fontana, A., Hengartner, H., De Tribolet, N., and Weber, E. (1984). J . Immunol. 132, 1837. 66. Frei, K., Bodmer, S., Schwerdel, C., and Fontana, A. (1985).J. Immunol. 135,4044. 67. Yasukawa, K., Hirano, T., Watanabe, Y., Muratani, K., Matsuda, T., Nakai, S., and Kishirnoto, T. (1987).EMBO J . 6, 2939. 68. Larsson, I., Landstrom, L. E., Larner, E., Lundgren, E., Miorner, H., and Stannegard, 0. (1978). Immunity 22, 786. 69. Bodrner, S., Siepl, C., and Fontana, A. (1989).In “Neuroimrnune Networks: Physiology and Diseases” (E. J. Goetzl and N. H. Spector, eds.), p. 73. Liss, New York. 70. Robinson, E. A., Yoshirnura, T., Leonard, E. J., Tanaka, S., Griffin, P. R., Shabanowitz, J., Hunt, D. F., and Appella, E. (1989). Proc. Natl. Acad. Sci. U.S.A. 86, 1850. 71. Frei, K., Leist, T. P., Meager, A., Gallo, P., Leppert, D., Zinkernagel, R. M., and Fontana, A. (1988).J . Exp. Med. 168,449. 72. Walker, K. B., Serwonska, M. H., Valone, F. H., Harkonen, W. S., Frick, 0. L., Scriven, K. H., Ratnoff, W. D., Browning, J. G., Payan, D. G., and Goetzl, E. J. (1988).J. Clin. Immunol. 8, 108. 73. Levine, J. D., Goetzl, E. J., and Basbaurn, A. 1. (1987).Rheum. DZr. Clin. North Am. 13,369. 74. Schiogolev, S. A., Goetzl, E. J., Urba, W. J., and Longo, D. L. (1989).J. Clin. Immunol. 9, 169. 75. Smith, E. M., and Blalock, J. E. (1981).Proc. Natl. Acad. Sci. U.S.A. 78,7530. 76. Smith, E. M., Phan, M., Kruger, T. E., Coppenhaver, D. H., and Blalock, J. E. (1983).Proc. Natl. Acad. Sci. U.S.A. 80,6010. 77. Besedovsky, H., Del Rey, A,, Sorkin, E., Lotz, W., and Schwulera, U. (1985). Clin. Exp. Immunol. 59,622. 78. Aliakbari, J., Sreedharan, S. P., Turck, C. W., and Goetzl, E. J. (1987). Biochem. Biophys. Res. Commun. 148, 1440. 79. Goetzl, E. J., Sreedharan, S. P., and Turck, C. W. (1988).J . Biol. Chem. 263, 9083. 80. Goetzl, E. J., Chernov-Rogan, T., Cooke, M. P., Renold, F., and Payan, D. G. (1985).1. Immunol. 135, 2707. 81. MartinyJ., Prystowsky, M. B., and Angeletti, R. H. (1987).J. Neurosci. Res. 18,82.
188
E. J. COETZL E T AL.
82. Goetzl, E. J., Kodama, K. T., Turck, C. W., Schiogolev, S. A., and Sreedharan, S. P. (1989).Immunology 66,554. 83. Finch, R. J., Sreedharan, S. P., and Goetzl, E. J. (1989).J . Immunol. 142, 1977. 84. Weigent, D. A., Baxter, J. B., Wear, W. E., Smith, L. R., Bost, K. L., and Blalock, J. E. (1988).FASEB J . 2,2812. 85. Buzzetti, R., McLoughlin, L., Lavender, P. M., Clark, A. J. L., and Rees, L. H. (1989).J. Clin. Invest. 83,733. 86. Kavelaars, A., Ballieux, R. E., and Heijnen, C. J. (1989).J. Immunol. 142, 2338. 87. Weinstock, J. V., Blum, A., Walder, J., and Walder, R. (1988).J. Immunol. 141.961. 88. Weinstock, J. V., and Blum, A. M. (1989).J . Immunol. 142,3256. 89. Fontana, A., Otz, U., De Weck, A. L., and Grob, P. J. (1981).J . Neuroimmunol. 2,73. 90. Merrill, J. E., Kutsunai, S., Mohlstrom, C., Hofman, F., Groopman, J., and Golde, D. W. (1984). Science 224, 1428. 91. Benveniste, E. N., Merrill, J. E., Kaufman, S. E., Golde, D. W., and Gasson, J. C. (1985).Proc. Natl. Acad. Sci. U.S.A. 82,3930. 92. Lachman, L. B., Brown, D. C., and Dinarello, C. A. (1987).J. Immunol. 138, 2913. 93. Merrill, J. E. ( 1989).In “Neuroimmune Networks: Physiology and Diseases” (E. J. Goetzl and N. H. Spector, eds.), p. 89. Liss, New York. 94. Hofman, F. M., von Hanwehr, R. I., Dinarello, C. A., Mizel, S. B., Hinton, D., and Merrill, J. E. (1986).J. Immunol. 136,3239. 95. Henmann, P., Lindholm, D., Bandtlow, C., Meyer, M., Radeke, M. J., Misko, T. P., Shooter, E., and Thoenen, H. (1987). Proc. Natl. Acad. Sci. U.S.A. 84,8735. 96. Baichwal, R. R., Bigbee, J. W., and De Vries, G. H. (1988).Proc. Natl. Acad. Sci. U.S.A. 85, 1701. 97. Milenkovic, L., Rettori, V., Snyder, G. D., Beutler, B., and McCann, S. M. (1989).Proc. Natl. Acad. Sci. U.S.A. 86,2418. 98. Bernton, E. W., Beach, J. E., Holaday, J. W., Smallridge, R. C., and Fein, H. G. (1987). Science 238,519. 99. Brown, S. L., Smith, L. R., and Blalock, J. E. (1987).J.Immunol. 139,3181. 100. Besedovsky, H., Del Rey, A., Sorkin, E., and Dinarello, C. A. (1986).Science 233, 652. 101. Fagarasan, M. O., Eskay, R., and Axelrod, J. (1989). Proc. Natl. Acad. Sci. U.S.A. 86,2070. 102. Scarborough, D. E., Lee, S. L., and Reichlin, S. (1989).In “Neuroimmune Networks: Physiology and Diseases” (E.J. Goetzl and N. H. Spector, eds.), p. 83. Liss, New York. 103. Sternberg, E. M., Hili, J. M., Chrousos, G. P., Kamilaris, T., Listwak, S. J., Gold, P. W., and Wilder, R. L. (1989).Proc. Natl. Acad. Sci. U.S.A. 86,2374. 104. Undem, B. J., and Weinreich, D. (1989). In “Neuroimmune Networks: Physiology and Diseases” (E. J. Goetzl and N. H. Spector, eds.), p. 155. Liss, New York. 105. Kornecki, E., and Ehrlich, Y. H. (1988). Science 240, 1792. 106. Levine, J. D., Gooding, J., Donatoni, P., Borden, L., and Goetzl, E. J . (1985). J . Neurosci. 5, 3025.
NEUKOIMMUNO1,OGY
189
107. Levine, J. D., Lam, D., Taiwo, Y., Donatoni, P., and Goetzl, E. J. (1986). Proc. Natl. Acad. Sci. U.S.A. 83, 5331. 108. Goetzl, E. J., Rangi, S. P., Serwonska, M. H., and Sreedharan, S. P. (1989). Proc. Int. Assoc. Allergol. Clin. Immunol. 12th, in press. 109. Foreman, J. C., Jordan, C. C., Oehme, P. and Renner, H. (1983).J . Physiol. 355,449. 110. Caughey, G. H., Leidig, F., Viro, N. F., and Nadel, J. A. (1988).J . Pharmacol. Exp. Ther. 244, 133. 1 1 1 . Goetzl, E. J., and Payan, D. G. (1984).J . Immunol. 133, 3255. 112. Payan, D. G., and Goetzl, E. J. (1988). Int. J.Neurosci. 38,2 1 1 . 113. Tomioka, M., Stead, R. H., Nielsen, L., Coughlin, M. D., and Bienenstock,J. (1988).J . Allergy Clin. Immunol. 82, 599. 114. Serra, M. C., Bazzoni, F., Della Bianca, V., Greskowiak, M., and Rossi, F. (1988).J . Immunol. 141, 21 18. 115. Perianin, A,, Snyderman, R., and Malfroy, B. (1989).Biochem. Biophys. Res. Commun. 161,520. 116. Mason, M. J., and Van Epps, D. (1989).J . Immunol. 142, 1646. 117. Ruff, M. R., Wahl, S. M., Mergenhagen, S., and Pert, C. B. (1985).Neuropeplides 5, 363. 118. Peterson, P. K., Sharp, B., Gekker, G., Brummit, C., and Keane, W. F. (1987).J . Immunol. 138, 3907. 119. Edwards, C. K., 111, Ghiasuddin, S. M., Schepper, J. M., Yunger, L. M., and Kelley, K. W. (1988).Science 239, 769. 120. Nong, Y. H., Titus, R. G., Ribeiro, J. M. C., and Remold, H. G. (1989). J . Immunol. 143,45. 12 1 . Bernton, E. W., Meltzer, M. S., and Holaday, J. W. ( 1988). Science 239,40 1 . 122. Peterson, P. K., Sharp, B., Gekker, G., Brummitt, C., and Keane, W. F. (1987).J. Clin. Invest. 80,824. 123. Lotz, M., Vaughan, J. H., and Carson, D. A. (1988). Science 241, 1218. 124. Kimball, E. S., Persico, F. J., and Vaught, J. L. (1988).J . Immunol. 141,3564. 125. Lemaire, I . (1988).J. Immunol. 140,2983. 126. Cannon, J. G., Tatro, J. B., Reichlin, S., and Dinarello, C. A. (1986). J , Irnmunol. 137, 2232. 127. Ottaway, C. A. (1989). In “Neuroimmune Networks: Physiology and Diseases” (E. J. Goetzl and N. H. Spector, eds.), p. 235. Liss, New York. 128. Williamson, S. A., Knight, R. A., Lightman, S. L., and Hobbs, J. R. (1988). Immunology 65,47. 129. Torres, B. A., and Johnson, H. M. (1988).J. Immunol. 140,2179. 130. Johnson, H. M., and Torres, B. A. (1988).J. Immunol. 141,2420. 131. Screpanti, I., Morrone, S., Meco, D., Santoni, A., Gulino, A., Paolini, R., Crisanti, A., Mathieson, B. J., and Frati, L. (1989).J . Immunol. 142, 3378. 132. Ahmed, S. A., Su, T. P., Montoya, A. I., and Talal, N. (1989).J. Immunol. 142,2647. 133. Scicchitano, R., Bienenstock, J., and Stanisz, A. M. (1988). Immunology 63, 733. 134. Payan, D. G., McGillis,J. P., and Organist, M. L. (1986).J. Biol. Chem. 261, 14321. 135. Asakawa, K., Hedo, J. A., McElduff, A., Rouiller, D. G., Waters, M. J., and ’ Gorden, P. (1986). Biochem. J . 238, 379. 136. Rouiller, D. G., and Gorden, P. (1987).Proc. Natl. Acad. Sci. U.S.A. 84, 126.
190
E. .J. GOETZL E T AL.
137. Matera, L., Muccioli, G., Cesano, A., Bellussi, G., and Genazzani, E. (1988). Brain, Behav., Immun. 2, 1. 138. Dailey, M. O., Schreurs,J., and Schulman, H. (1988).j. Immunol. 140,2931. 139. Vigna, S. R., Thorndyke, M. C., and Williams,J. A. (1986).Proc. Natl. Acad. Sci. U.S.A. 83, 4355. 140. Su, T. P., London, E. D., and Jaffe, J. H. (1988). Science 240,219. 141. Smith, E. M., Brosnan, P., Meyer, W. J., 111, and Blalock, J. E. (1987). N. E n d . I . Med. 317, 1266. 142. Laufgr, R., Gilson, C., Chorev, M., and Selinger, Z. (198S).j. Biol. Chem. 261,10257. 143. Bergstrom, L., Beaujonan, J. C., Torrens, Y., Saffroy, M., Glowinski, J.,
Lavielle, S., Chassning, G., D’Orleans-Juste, P., Dion, S., and Regoli, D. (1987). Mol. Pharmacol. 32,764. 144. Masu, Y., Nakayama, K., Tamaki, H., Harada, Y., Kumo, M., and Nakanishi, S. (1987). Nature (London) 329, 836. 145. Dam, T. V., Escher, E., and Quirion, R. (1987). Biochem. Biophys. Rex Commun. 149,297. 146. Stanisz, A. M., Scicchitano, R., Dazin, P., Bienenstock, J., and Payan, D. G. (1987).J . Immunol. 139,749. 147. Sreedharan, S. P., Kodama, K. T., Peterson, K. E., and Goetzl, E. J. (1989). 1.Biol. Chem. 264,949. 148. -Nakamura, H., Koike, T., Hiruma, K., Sato, T., Tomioka, H., and Yoshida, S. (1987). Immunology 62,655. 149. Tseng, J., and ODorisio, M. S. (1989). I n “Neuroimmune Networks: Physiology and Diseases” (E.J. Goetzl and N. H. Spector, eds.), p. 105. Liss, New York. This article was accepted for publication on 20 October 1989.
ADVANCES IN IMMUNOLOGY, VOL. 48
Immune Privilege and Immune Regulation in the Eye JERRY Y. NIEDERKORN Department of Ophthalmology, University of Texas Southwestern Medical Center, Dallas, Texas 75235
1. Introduction
The eye has been likened to an imunological microcosm in which virtually all forms of immunological events can take place and, in doing so, often produce unique results (1). Indeed, when the topic of immunological privilege is raised, the anterior chamber of the eye (Fig. 1) is offered as the classic example of a privileged site in which histoincompatible allografts escape immunological recognition and enjoy prolonged, and sometimes permanent, residence (2). Likewise, discussions of other immunological phenomena, such as tolerance, often turn to the eye for examples to illustrate basic immunological principles. For example, one pathway for maintaining immunological tolerance of host tissue epitopes is the sequestration of self-antigens behind anatomical barriers. In this regard, the crystallin antigens of the lens are offered as examples of tissue-specific immunogens that are sequestered early in ontogeny but, although potentially capable of arousing an autoimmune response, do not dq so since they are incarcerated within an impervious capsule. Thus, tolerance of host antigens (e.g., lens crystallins) can be maintained, at least in part, by anatomical sequestration. T h e eye possesses other unique features that warrant the immunologist’s attention. Perhaps the oldest and most successful form of organ transplantation takes place in the eye. Corneal transplantation has been performed on animal subjects for over 150 years and on human patients for over 80 years (3). In the United States alone, 30,000 corneal transplants are performed each year, with a success rate well over 90%. T h e extraordinary success of corneal transplantation is often attributed to the mysterious and undefined “immunological privilege’’ of the cornea and the eye. What is the basis for this immunological privilege, and is it restricted to the eye? In this review I examine those features of the cornea and the eye which conspire to promote successful corneal transplantation. The putative immunological privilege of the cornea, corneal graft bed, and anterior chamber would seemingly create an environment free of 191 Copyright 8 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.
192 Area Of autoimmune activitv
I
JERRY Y . NIEDERKORN
,.
Choroid
,
Retina L
Anterior chamber
Area of immunologic privilege
Pupil Iris
FIG.1. Schematicrepresentation of ocular anatomy demonstratingregions of immunologicalprivilege and areas especially vulnerable to autoimmune activity. immune-mediated diseases. However, this is not the case, as the eye is vulnerable to an interesting array of autoimmune diseases uniquely suited to precise immunological analysis. Thus, the eye offers interesting opportunities for analyzing a panorama of immunological phenomena, including organ transplantation, immunological privilege, tolerance, and autoimmunity. Moreover, each of these immunological processes can have a profound bearing on the normal functioning of one of the most precious of the five senses: our vision. II. Corneal Allografts: lmmunogenically Privileged Grafts on Immunologically Privileged Graft Beds
The prospect of replacing an opaque diseased cornea with a healthy one was suggested as early as 1796 by Erasmus Darwin, who suggested that . . a slight and not painful operation might be facilitated by cutting the cornea with a kind of trephine, about the size of a thick bristle or a small crow-quill, an experiment I wish strongly to recommend to some ingenious surgeon or oculist” (4). In 1837 an Irish surgeon, Samuel Bigger, described how he had successfully placed a corneal allograft into a pet gazelle’s eye while he was a prisoner of the Egyptians (5). The first recorded attempt at therapeutic corneal transplantation on a human I‘.
IMMUNOLOGY OF T H E EYE
193
subject was performed by Richard Kissam, who sutured the cornea of a 6-month-old pig onto a blind Irishman (6). This xenograft ultimately failed, as did subsequent attempts at corneal grafting. However, succeeding decades brought the introduction of general anesthesia, antisepsis in surgery, improvements in ophthalmic surgical instruments, and, at the turn of this century, the first successful human corneal transplant (7). In the 75 years following the first successful human corneal allograft, literally thousands of corneal transplants have been performed each year to correct blndness produced by corneal edema, trauma, inflammation, or congenital abnormalities. The success of corneal transplantation is unrivaled by all other forms of organ transplantation. Corneas transplanted onto diseased, but otherwise avascular, graft beds (e.g., keratoconus), remain clear and healthy in greater than 90% of the recipients (8,9).This extraordinary success rate is even more impressive when one considers the conditions surrounding corneal transplantation. Human leukocyte antigen matching of donor and recipient is normally not performed, except in high-risk patients. Immunosuppressive drugs are restricted to topical corticosteroid eyedrops, which are gradually tapered to maintenance levels following suture removal. Such conditions would certainly lead to graft failure with any other form of organ transplantation, yet corneal allografts thrive in spite of such immunological handicaps. A. ESCAPING IMMUNOLOGICAL RECOGNITION The apparent ease with which corneal grafts avoid immunological recognition suggests that the cornea is endowed with unique immunological characteristics and thus possesses an immunological privilege not shared by other organ grafts (10). Historically, three basic hypotheses have been offered to account for the privileged existence of corneal allografts (10). The simplest explanation suggests that the cornea, like certain neuronal tissues, is devoid of conventional major histocompatibility complex (MHC) antigens. Accordingly, potentially alloreactive T cells would be “blind” to alien corneal cells. The putative absence of MHC antigens would not only prevent the arousal of an alloimmune response, but even if a response were initiated, the grafts would be immunologically invisible. This hypothesis, although appealing in its simplicity, has been unequivocally disproved. Several findings indicate not only that the cornea is vulnerable to immunological attack, but that it is also capable of eliciting an alloimmune response that results in graft rejection. Studies in rats, rabbits, and mice indicate that heterotopic transplantation of corneas to subdermal graft beds leads to rapid sensitization and swift rejection of the allografts (1 1-14).
194
JERRY Y. NIEDERKORN
Moreover, MHC antigens have been detected on all three cell layers of the cornea (15-18) (Fig. 2). In fact, the grafting procedure itself has been shown to elicit increased expression of MHC class I1 antigens on the corneal epithelium (19, 20). Although MHC antigens can be found on cells of the corneal epithelium, stroma, and endothelium, the density of antigen expression differs markedly among these three cell layers. In the rat, corneal endothelial cells express meager detectable amounts of MHC class I antigens and no detectable MHC class I1 antigens (2 1). Despite the paucity of MHC antigens, corneal endothelial grafts stimulate robust cytotoxic T lymphocyte (CTL) responses following heterotopic grafting in the rat (22). Although MHC class I antigen expression is greater in the epithelium than in the endothelium, heterotopic transplantation of isolated allogeneic corneal epithelium fails to induce detectable anti-MHC class I CTL responses in the rat (22). Thus, there are significant antigenic and immunogenic gradients within the corneal allograft. The mere expression of histocompatibility antigens on corneal cells does not ensure the induction of an alloimmune response. Nonetheless, the corneal allograft has the full potential to be both immunogenic and antigenic.
Basement membrane Bowman’s membrane
Descemet’s membrane Endothelium-
FIG. 2. The cornea is composed of three distinct layers: epithelium, stroma, and endothelium. MHC class I antigens are expressed on cells of each layer. [Reprinted from Niederkorn and Peeler (153) by permission of S. Karger Publishing, Inc.]
IMMUNOLOGY O F 'I'IIE EYE
195
The second hypothesis offered to explain the survival of corneal allografts suggests that the donor cells were rapidly replaced by host cellular components in the graft bed. According to this hypothesis, the cellular elements of the graft were replaced before the host's immune machinery could be aroused. Like the previous hypothesis, this explanation has been refuted. Animal studies using sex chromatin markers to distinguish donor cells from recipient cells have demonstrated the long-term survival of donor cells in corneal grafts (23,24). Other investigators came to similar conclusions by radiolabeling donor corneas with [3H]thymidine (25,26). Clinical findings also support this conclusion, since immunological rejection can occur over a decade after corneal transplantation (10). The third and most widely accepted hypothesis to account for the high acceptance of corneal allografts relates to the nature of the avascular corneal graft bed. It is a well-recognized clinical observation that vascularization of the corneal graft bed is a harbinger of graft failure (27,28). The absence of blood and lymph vessels at the interface of the graft and the graft bed is thought to prevent the escape of alloantigens to the regional lymphoid tissues, thereby resulting in an afferent blockade of the immunological reflex arc. Although the corneal allograft is potentially immunogenic and antigenic, the anatomical sequestration promotes graft survival due to the privileged location of the graft. Thus, the avascular graft and the graft bed conspire to produce a state of immunological ignorance that permits allograft survival. Heterotopic and orthotopic corneal allografts in rats and mice have been used to examine this hypothesis. Corneal allografts can be transplanted heterotopically (i.e., to an abnormal anatomical site) onto subdermal graft beds richly endowed with blood vessels and lymphatics, which would favor the induction and execution of alloimmunity. Such grafts can be compared to similar grafts placed orthotopically (i.e., to the normal anatomical site) onto avascular graft beds in the eye. If the avascular graft bed contributes to the survival of the allograft, one would predict that heterotopic corneal grafts would suffer significantly higher rejection rates than their orthotopic counterparts. This is indeed the case, as 100%of the fully allogeneic heterotopic corneal allografts are rejected in a mouse model of corneal transplantation (13, 14, 29), while only 55-57% of fully allogeneic grafts fail when grafted orthotopically (30, 31). The pioneering studies of Maumenee (32) provided strong support for the afferent blockade theory. In these studies rabbits bearing long-term orthotopic corneal allografts rejected skin grafts from the same donors which provided their corneal grafts. Skin graft rejection occurred at a tempo indicative of a first-set rejection, thereby supporting the notion
196
JERRY Y. NIEDERKORN
that the initial corneal allograft failed to stimulate alloimmunity and that the graft bed produced an afferent blockade of the immune response. However, skin graft rejection led to the rejection of 90%of the previously clear corneal grafts. Thus, the corneal grafts initially displayed immunogenic privilege, but were antigenically vulnerable to an ongoing systemic immune response: Afferent blockade was present, but efferent blockade was not. Callanan and co-workers (30) came to similar conclusions, using a rat orthotopic corneal allograft model. In these studies the appearance of antigen-specific CTL activity coincided with graft rejection, while the absence of CTL responses was a consistent feature of hosts bearing long-term corneal allografts. It is interesting that in both of these studies a small but significant number of corneal grafts were initially clear and avascular, yet subsequently underwent immunological rejection. Thus, the presence of an avascular graft bed does not necessarily ensure permanent graft survival or the maintenance of an afferent blockade. Recent studies from our laboratory lend further support to the afferent blockade theory of corneal graft survival. Ross et al. (33)have shown that orthotopic corneal grafts differing from their hosts only at MHC class I1 loci do not undergo rejection unless the host is systemically immunized with skin grafts from the same donor strain. The importance of an efferent blockade in promoting corneal graft survival is not clear. The previously mentioned studies by Maumenee (32)and by Ross et al. (33) argue against an effective efferent blockade. However, Khodadoust and Silverstein (34)confirmed earlier findings by Billingham and Boswell(1 l), which indicated that the corneal graft bed served as an effective barrier that shielded the graft from sensitized effector elements of the host. Results from the Khodadoust and Silverstein study indicated that 95% of the lamellar corneal grafts and 75% of the penetrating grafts remained healthy, even though the hosts had rejected large skin grafts from the same donors of the corneal grafts. LANCERHANS CELLSIN INITIATING CORNEAL B. ROLEOF DONOR GRAFTREJECTION Numerous studies have demonstrated the presence of MHC class I antigens on cells in all three layers of the cornea (15-18); however, MHC class I1 antigens are not normally expressed in detectable amounts on corneal cells. MHC Class I1 antigen-bearing Langerhans cells are abundant in the peripheral outermost regions of the cornea that interface with the conjunctiva (i.e., limbus), but are conspicuously absent from the central corneal epithelium in adult humans, mice, rabbits, and guinea pigs (35-38).
IMMUNOLOGY OF THE EYE
197
The absence of Ia’ Langerhans cells in the central cornea is of more than casual interest, since these cells represent an important immunogenic component of an allograft. Indeed, it has been suggested that Ia+ “passenger cells” are the major barrier to successful organ transplantation (39). Over three decades ago Snell (40) suggested that donor leukocytes present in transplanted tissues were a major source of tissue immunogenicity. Interest in passenger cells, however, remained dormant until Lafferty et al. (39) reconsidered the role of Ia’ cells in thyroid allografts and pancreatic islet grafts. Subsequently, numerous studies have confirmed that Ia’ passenger cells are indeed major obstacles to successful organ transplantation and that their prior removal has a favorable effect on graft survival. The conspicuous absence of Langerhans cells in the central corneal epithelium offers a unique opportunity to determine whether grafts typically devoid of passenger cells are capable of inducing the normal array of alloimmune responses and whether the presence or absence of such cells affects graft survival. Results from our laboratory indicate that the absence of resident Langerhans cells has a profound effect on the graft’s ability to induce allospecific delayed-type hypersensitivity (DTH) responses following heterotopic transplantation (4 1). Surprisingly, grafts differing from the host at MHC classes I and I1 loci as well as multiple minor H loci were consistently rejected, but failed to induce detectable DTH responses before, during, or after immunological rejection (20,29, 4 1). Although DTH responses were not induced, Langerhans cell-free grafts elicited potent CTL responses that coincided with the onset of graft rejection (20, 29,41). It is generally agreed that solid-organ transplants, such as skin or cornea, are rejected by either CTL- or DTH-mediated processes (42). Central corneal allografts fail to induce DTH, yet undergo immunological rejection at both orthotopic and heterotopic sites. Therefore, it is reasonable to conclude that DTH plays little, if any, role in the rejection of corneal allografts. Moreover, adoptive transfer studies in adult thymectomized bone marrow reconstituted mice (ATXBM) demonstrate that lymphoid cell suspensions depleted of CD4’ (i.e., T helper/DTH), but containing CD8’ (CTL/suppressor), cell populations promoted corneal graft rejection in T cell-depleted ATXBM recipients (29). Categorizing effector T cells into either CTL or DTH populations, based on the expression of cell surface markers, is an artificial and potentially misleading approach for understanding the mechanism of corneal graft rejection. CD4’ T cell clones can demonstrate cytolytic activity against MHC class I1 alloantigens (43),whereas some CD8’ T cell populations secrete lymphokines and demonstrate T helper cell activity
198
J E R R Y Y. N I E D E R K O R N
(44). Whether corneal allograft rejection is solely due to CTL or DTH is more of a semantic question than an immunological one. It is more productive to consider the relevance of the T cell phenotype in terms of the class of MHC antigen that it recognizes rather than with a presumed effector function. Recent studies of skin allograft rejection in mice have shown that the elimination of CD8' T cells (MHC class I-restricted T cells) enhances the survival of skin allografts differing from the host only at MHC class I loci (45). In similar studies Rosenberg et al. (46) found that the elimination of CD4' cells does not alter rejection of MHC class I-disparate skin grafts, whereas adoptive transfer of CD4+/CD8- T cell populations produced rapid rejection of MHC class I1 and minor H-disparate skin grafts. The rejection of a skin allograft, whether the MHC disparity is at class I only or class I1 only, most likely occurs by the same immunological process, even though different T cell populations are probably involved. Likewise, the basic mechanism of corneal allograft rejection, although fundamentally different from skin graft rejection, is most likely a CTLdominated process, regardless of the nature of the allodisparity. Although the absence of Langerhans cells does not affect the recognition and rejection of heterotopic corneal grafts involving MHC class I or I1 disparity (20, 29, 41), the situation with orthotopic grafts is markedly different. Using a rat model, Callanan et al. (30, 47) recently demonstrated that orthotopic central corneal allografts differing from the recipient at MHC classes I and I1 loci were rejected in 5 5 % of the naive recipients. However, if similar corneal grafts were pretreated with sterile latex beads to induce the infiltration of donor-derived Langerhans cells (before transplantation), 98% (and more recently, 100%) of the grafts were rejected (30; unpublished observations). Moreover, rejection was invariably accompanied by the development of antigen-specific CTL responses. Thus, the presence of graft-borne Langerhans cells almost doubled the risk of rejection of orthotopic corneal allografts involving combined MHC classes I and I1 disparities. Recent findings from our laboratory indicate that graft-borne Langerhans cells are the major immunogenic stimulus for the rejection of male corneal grafts by female recipients (19, 48). Langerhans cell-free maleheterotopic corneal grafts were incapable of inducing either a CTL or DTH response against the male-specific H-Y antigen (19,48). Moreover, male grafts were not rejected, even though they were residing on vascularized graft beds. The male grafts did, however, express H-Y antigen, since female mice preimmunized with male skin grafts rapidly rejected central corneal grafts from male donors. The immunogenic privilege of male grafts could, however, be breached by the presence of male donor
I M M U N O L O G Y O F T H E EYE
199
Langerhans cells. Corneas pretreated with latex beads, as a means of inducing the infiltration of donor Langerhans cells, induced strong CTL and DTH responses and underwent prompt rejection following heterotopic transplantation onto naive female recipients ( 19). Ray-Keil and Chandler (49) came to the same conclusions, using a similar heterotopic corneal graft model. However, in their study a significant number (i.e., 36%) of the male grafts were rejected by naive female hosts. The discordance between their results and ours might be explained by the presence of donor Langerhans cells in their grafts, since our grafts were approximately 2.5 mm in diameter, while theirs were slightly larger (3.0 mm in diameter) and therefore more likely to contain the peripheral regions that are known to contain large concentrations of Langerhans cells. Nonetheless, it is intriguing that Ray-Keil and Chandler found that elimination of donor Langerhans cells (i.e., hyperbaric oxygen treatment) virtually eliminated graft rejection: Only one of 23 grafts underwent rejection. Thus, in the case of at least one minor histocompatibility antigen system (i.e., H-Y), the absence of Langerhans cells has a profound effect on the immunogenicity and fate of the corneal graft. At this point the question arises as to whether it is the mere expression of alien MHC class I1 antigens on cells of the corneal graft or the expression of these antigens on the surface of a cell with antigen-presenting potential (i.e., a Langerhans cell) that is crucial in provoking an immune response and graft rejection. In vitro treatment with y-interferon readily induces the expression of MHC class I1 antigens on human (50) and rabbit (5 1) corneal endothelium. Male corneal grafts incubated in y-interferon (1) displayed extensive expression of Ia antigen, (2) induced potent H-Y-specific DTH and CTL responses, and (3) were promptly rejected following heterotopic transplantation to female C57BL/6 recipients (19). It is interesting, however, that the expression of donor Ia antigens on the corneal grafts occurs independently of the rejection process. Central corneal grafts from male donors become Ia' 2 days after grafting into syngeneic female recipients and remain Ia' through day 7 posttransplantation, yet the grafts fail to induce either a CTL or DTH response and survive indefinitely (19). By contrast, male grafts incubated in y-interferon, and therefore Ia' at the time of heterotopic grafting, are promptly rejected (19). Thus, the time at which Ia antigens are expressed determines whether the corneal graft is destined to survive or undergo rejection. A similar situation occurs with corneal grafts differing from the host only at MHC class I1 loci. Streilein et al. (52) found that MHC class
200
JERKY Y . NIEDERKORN
11-disparate heterotopic corneal grafts were not rejected and failed to induce alloimmunity. Such grafts failed to sensitize the host against the donor’s MHC class I1 antigens, because subsequent challenged orthotopic skin grafts from the corneal donor strain were rejected in a first-set tempo, indicating that the MHC class I1 antigens were being perceived by the host’s immune system for the first time. In subsequent studies we used the same donor-host combinations (i.e., A.TL and A.TH) and evaluated the ability of such grafts to induce allospecific DTH and CTL responses (20). Surprisingly, MHC class 11-disparategrafts containing the Langerhans cell-rich limbus region or grafts pretreated with latex beads, so that they contained donor Langerhans cells at the time of grafting, induced vigorous DTH and CTL responses, yet were not rejected (20). Using an orthotopic rat model, Katami et al. (53) reported a 95% rejection rate for grafts differing from the host at MHC classes I and I1 loci, but only a 25% rejection rate for MHC class 11-incompatible grafts. Ross et al. (33) observed similar results for orthotopic corneal allografts involving MHC class 11-disparategrafts. Although fully allogeneic grafts (i.e., MHC classes I and I1 disparate) were rejected in 55% of the hosts, MHC class 11-disparate grafts did not undergo rejection (i.e., experienced 100% survival), even though such grafts were found to express donor MHC class I1 antigens 4-7 days following transplantation. The transient expression of MHC class I1 antigens, however, is of sufficient duration to render the graft vulnerable to the efferent arm of the immune apparatus, since recipients preimmunized with MHC class IIdisparate skin grafts reject subsequent orthotopic corneal grafts (33). It is interesting, however, that grafts pretreated with latex beads prior to transplantation did not undergo immunological rejection unless the host was subsequently immunized with skin grafts from the same MHC class 11-disparatedonor strain (33). Under these circumstances both the orthotopic corneal graft and the skin graft underwent rejection. Thus, the MHC class 11-disparate corneal graft is not immunogenic, but can be induced to express alien MHC class I1 antigens that serve as targets for alloimmune effector elements. Collectively,the results indicate that in the case of corneal grafts representing only minor histoincompatibilities or MHC class I1 disparities, the timing of MHC class I1 antigen expression is critical and determines whether the graft is destined to survive or perish. C. DYNAMIC DISTRIBUTION OF CORNEAL LANGERHANS CELLS The well-defined geometric arrangement of Langerhans cells in the peripheral limbus and the abrupt exclusion of these cells from the central regions of the corneal epithelium suggest that the distribution of these
IMMUNOLOGY OF THE EYE
20 1
cells is under stringent regulation. Recent studies suggest that the juxtapositioning of corneal Langerhans cells is a dynamic physiological process. Various stimuli can induce corneal Langerhans cell migration from the periphery to the central regions. For example, electrocautery (54), bacterial and viral infections (54-58), and phagocytic stimuli (59) can induce Langerhans cell migration (Fig. 3). Since Langerhans cells’ rulson d’ftre is antigen presentation, one might suspect that the aforementioned stimuli somehow mimic signals that beckon Langerhans cells to sites for antigen processing. With this in mind, investigations were conducted to determine whether corneal cells could phagocytose antigens and, as a result, elaborate chemoattractants that induced the centripetal migration of peripheral Langerhans cells into the central regions of the cornea. It has been previously demonstrated that rabbit corneal cells elaborate an interleukin (1L)-1-like molecule, corneal epithelial cell-derived T cell-
Latex beads,
//
FIG.3. Distribution of epithelial Langerhans cells. IA’ Langerhans cells are concentrated at the periphery of the cornea and at the limbus (i.e., corneal/ conjunctival border). The central corneal epithelium is normally devoid of Langerhans cells. However, sterile latex beads deposited into shallow incisions in the corneal epithelium will induce the rapid migration of peripheral Langerhans cells into the central regions of the cornea. [Reprinted from Niederkorn and Peeler (153) by permission of S. Karger Publishing, Inc.]
202
JERRY Y. NIEDEKKOKN
activating factor (60). The quantity of this factor elaborated by rabbit corneal cells is increased 50-70% following phagocytosis of either Staphylococcus or latex beads (59). Moreover, Shams et al. (61) recently confirmed these findings, using human corneal epithelial cells, and demonstrated that either y-interferon or Staphylococcus aureus, would stimulate secretion of biologically active IL-1p. The hypothesis that IL- 1 and/or CETAF functioned as a chemoattractant to promote the migration of peripheral Langerhans cells into the central corneal epithelium was supported by findings in which intracorneal injection of purified IL-1 induced rapid centripetal migration of peripheral Langerhans cells to the site of cytokine inoculation (59). The specificity of the chemoattractant was supported by the observation that this effect could be blocked by anti-IL-1 antibodies (59). Chemotaxis could not be mimicked by intracorneal injection of other cytokines, such as IL-2, or by irrelevant proteins (e.g., hen egg lysozyme). Thus, corneal epithelial cells have the capacity to regulate the distribution and migration of potential antigen-presentingcells within this organ. The orderly distribution of Langerhans cells and the cornea’s ability to regulate their migration and distribution suggest that a dynamic regulatory process is at work. At this point one can only speculate on the mechanisms that maintain the normal circumferential arrangement of limbal Langerhans cells. Perhaps local chemotactic factors-either cell membrane bound or soluble-prevent Langerhans cell migration from the limbus, unless a stronger stimulus (e.g., IL-1) is released from the central corneal epithelium. An attractive, yet unproven, hypothesis is that the corneal epithelial cells might serve as accessory antigen-processing cells. Since the central cornea is devoid of conventional antigen-presenting cells, such as dendritic cells, some immunological provision must be made to accommodate pathogens and antigens that may insult the corneal epithelium. Potential antigens or pathogens (e.g., Staphylococcus or herpes simplex virus) are phagocytosed by corneal epithelial cells. This, in turn, would stimulate the elaboration of IL-I, which then induces the migration of peripheral Langerhans cells to the area of antigen accumulation. Partially processed antigens could then be displayed on the cell membranes of corneal epithelial cells and then be presented to the itinerant Langerhans cells for further processing and presentation to peripheral lymphoid elements. As the levels of IL-I dissipate, the Langerhans would be stimulated to return to the periphery and eventually to the regional lymphoid apparatus, where they would interact with antigen-specific T cells and thus initiate a conventional immune response.
IMMUNOLOGY OF n i E EYE
203
D. IMMUNOREGULATORY EFFECTS OF CORNEAL LANGERHANS CELLS:A DOUBLE-EDGED SWORD There is little doubt that corneal Langerhans profoundly influence the immunogenicity and fate of corneal allografts. Ridding corneal grafts of these passenger cells greatly reduces the risk of rejection in experimental animals and would presumably have the same effect on human subjects. Although the absence of Langerhans cells in the central cornea is an important parameter in explaining the immunological privilege of corneal allografts, other considerations are also involved. A thorough understanding of the corneal allograft’s exemption from immunological rejection may have a profound impact on other categories of organ transplantation. It is clear that Langerhans cells serve a vital function as antigenpresenting cells and that their presence in the eye is crucial for maintaining protective immune surveillance against potential pathogens. T h e peculiar distribution of ocular Langerhans cells, however, leads one to suspect that the behavior and functioning of ocular Langerhans cells is much different than cutaneous Langerhans cells. Understanding the biology of corneal Langerhans cells remains an intriguing and formidable challenge for the immunologist. 111. Immunological Privilege of the Anterior Chamber
T h e immunological privilege of the anterior chamber was recognized over 100 years ago, when researchers found that xenogeneic tumor grafts survived in this ocular compartment significantly longer than they did at other sites (62). In the 1940s Greene and associates (63) used the anterior chamber of the rabbit eye as a tool for propagating and passaging human tumors. Moreover, Greene raised an interesting, although unproven, hypothesis that the growth and metastasis of such tumors from the anterior chambers of experimental hosts could be used to predict the original tumor’s malignancy. Nonetheless, in the 40 years following Greene’s original observation, enormous insights have been gained regarding the nature of the immunological privilege in the anterior chamber of the eye. T h e known absence of lymphatic drainage to regional lymph nodes led early investigators to propose that the privilege of the anterior chamber could be traced to the sequestration of alloantigens within this compartment and, thus, afferent blockade of the immune apparatus (62). However, during the past decade it has become increasingly clear that the
204
JERRY
Y. NIEDERKORN
immunological privilege extended to allogeneic tissues transplanted into the anterior chamber is not an immunologically null event, but rather a dynamic immunoregulatory process of exceptional complexity. A. ANTERIOR CHAMBER-ASSOCIATED IMMUNE PRIVILEGE: THEBASIS FOR IMMUNOLOGICAL PRIVILEGE The notion that allografts placed into the anterior chamber are sequestered from the systemic apparatus was soon disproved. Raju and Grogan (64) and Franklin and Prendergast (65) demonstrated that not only were alloantigens capable of emigrating from the anterior chamber to the peripheral immune apparatus, but that alloimmune effectors were generated against the anterior chamber allografts which eventually underwent rejection. The first clue that antigens delivered directly into the anterior chamber could be processed in a manner that favored downregulation of alloimmunity came from studies by Kaplan and Streilein (66-68). In their studies semi-allogeneic (i.e., F1 donor) cells served as alloantigens and were injected directly into the anterior chambers of parental-strain rats. Evidence that the alloantigens had been perceived by the immune system came in the form of serum antibodies against the donor alloantigens. More importantly, these hosts demonstrated a modest, albeit significant, delay in their ability to reject challenged skin grafts from the donors of the intracameral inoculum, but were able to execute normal first-set rejection of third-party skin allografts (66-68). Thus, presentation of alloantigens via the anterior chamber produced an antigen-specific impairment of systemic cell-mediated immunity: Immune privilege was extended beyond the eye to extraocular sites. These results, along with those by Subba Rao and Grogan (69), indicated that the terms of immunological privilege were defined by the magnitude of the histocompatibility disparity between donor and hostimmunological privilege not being an all-or-none proposition. Subsequent studies in mice revealed that an impressive display of immune privilege could be demonstrated by transplanting tumor allografts into the anterior chamber. DBA/2 mastocytoma cells (P815) undergo swift rejection following subcutaneous transplantation into allogeneic BALB/c recipients, due to the host’s recognition of the minor histocompatibility antigens of the DBA/2 donor strain. However, equal numbers of DBAIP mastocytoma cells are not rejected following transplantation into the anterior chamber of BALB/c hosts. DBA/2 tumors not only grow progressively, but induce antigen-specific suppression of systemic cell-mediated immunity (70, 7 1). BALB/c hosts challenged orthotopically with DBA/2 skin grafts fail to reject donor-strain skin grafts, but are fully capable of rejecting third-party skin allografts.
IMMUNOLOGY OF T H E EYE
205
Th e progressive growth of intraocular tumor allografts and the failure
to reject relevant orthotopic skin grafts from the DBAI2 donor strain
suggested that systemic allograft immunity was paralyzed. Surprisingly, this was not the case, as these hosts developed serum antibodies as well as CTLs specific for the DBA/2 alloantigens (72). The latter finding represents a perplexing result, considering the long-term survival of the skin allografts and the ever-expanding intraocular tumor allografts. This unusual spectrum of immunological findings led to the generic term “anterior chamber-associated immune deviation” (ACAID) to convey the dynamic and diverse nature of this phenomenon (73). Investigations with other antigen systems revealed that the ACAID phenomenon was not restricted to minor histocompatibility antigens. Wetzig et al. (74) and Waldrep and Kaplan (75) demonstrated that anterior chamber inoculation of hapten-derivatized lymphoid cells resulted in ACAID, as demonstrated by the down-regulation of antigen-specific DTH responses. Moreover, these studies also established that the systemic down-regulation of cell-mediated immunity was attributable to the development of a suppressor T cell system. It is now apparent that a wide array of antigens elicit ACAID when delivered via the anterior chamber. Herpes simplex virus (76, 77), melanoma antigens (78), bovine serum albumin (79), hapten-derivatized cells (74,75), and retinal S antigen (80,81) induce ACAID following anterior chamber inoculation. The facility with which ACAID can be induced has permitted further exploration into the mechanisms behind the induction and maintenance of this immunoregulatory phenomenon. B. INDUCTION OF ACAID Th e absence of a patent lymphatic drainage route for grafts residing in the anterior chamber indicates that antigens must leave this site via the blood vascular route. Accordingly, it has been suggested that anterior chamber inoculation of antigen is tantamount to an intravenous injection and is merely another form of immune deviation, similar to the one described by Asherson and Stone (82), in which small amounts of deaggregated heterologous proteins injected intravenously led to the production of serum antibodies, but a conspicuous absence of DTH. However, several findings argue convincingly against the hypothesis that ACAID is simply a cumbersome method for producing intravenous immune deviation. Repeated attempts to mimic ACAID by intravenous injection of P815 cells have failed (83). More importantly, the nature of the suppressor mechanism of ACAID differs markedly from intravenously induced immune deviation. T h e suppressors of ACAID act at the efferent
206
JERRY Y . NIEDERKORN
level (84), while the suppressor system of immune deviation functions only at the afferent mode (85). As is discussed later, ACAID requires the presence of an intact spleen (73);by contrast, immune deviation is not ablated by splenectomy (86). In the case of herpes simplex virus (HSV),the intravenous injection of virus induces suppressed DTH, but vigorous CTL, responses (87, 88). However, anterior chamber inoculation of HSV leads to the suppression of both CTL (77) and DTH responses (76,88). The eye plays an active role in the inductive phase of ACAID. Removal of the eye 4-7 days after intracameral inoculation of antigen prevents the induction of ACAID (89). It is also within this time frame that allogeneic tumor cells (originating from the anterior chamber) can first be detected in the spleen of the host (90,91).Thus, significant events occur within the eye during the first 7 days; beyond this time the eye is superfluous. There are two considerations that may provide insights for understanding the induction of ACAID. One explanation is that the unique physicochemical environment of the anterior chamber alters antigens or antigen-perceiving cells. In this regard, Streilein et al. (92)and Granstein et al. (93) have reported the presence of transforming growth factor-p-a potent inhibitor of T cell proliferation-in the aqueous humor. Moreover, these investigators, as well as others, have demonstrated that aqueous humor inhibits antigen-specific and -nonspecific lymphocyteproliferative responses. A second, not mutually exclusive, explanation relates to the possible presence of unique antigen-processing cells within the anterior chamber or lining the outflow channels that communicate with the venous drainage system of the anterior chamber. It is conceivable that the eye possesses a specialized population of antigen-presenting cells committed to delivering a down-regulatory signal to the host’s T cells in a manner analogous to the “suppressor antigen-presenting cells” of the skin (94). Granstein et al. (94,95) have demonstrated the presence of ultraviolet radiation (UVR)-resistant antigen-presenting cells in the skin of mice. Under normal conditions, haptens painted onto the skin evoke positive contact sensitivity; however, hapten sensitization on UVR skin results in the suppression of a DTH response. Granstein (94) demonstrated that the suppression occurred via antigen presentation by UVR-resistant cells residing in the skin. It is feasible that similar suppressive antigenpresenting cells reside in the eye and are responsible for the obligatory ocular phase of ACAID. In searching for such cells, we have focused on the trabecular meshwork of the eye-the aqueous outflow system which serves as the conduit for removing the contents of the anterior chamber. Sparse, but consis-
IMMUNOLOGY OF T I I E EYE
207
tent, numbers of Ia' cells can be detected in the trabecular meshwork of both mouse and human eyes (96,97).The immunoregulatory capacity of these cells, however, remains to be established. Preliminary results suggest that cells of the iris and the ciliary body may have immunoregulatory functions. Mixed cultures of iris and ciliary body inhibited antigen-driven lymphocyte-proliferative responses in an antigen-nonspecific manner (92). It bears noting that at least a portion of this inhibitory effect may be attributable to the aqueous humor, since cells of the ciliary body are the major producers of the aqueous humor in situ. Recent findings lend support to the proposition that the induction of ACAID is due to aberrant antigen-processing cells, not unique humoral factors, in the eye. Williamson and Streilein (98) demonstrated that positive allospecific DTH responses could be induced via the anterior chamber if the alloantigen-bearing cells were admixed with cutaneous Langerhans cells of the recipient strain prior to anterior chamber inoculation. Presumably, the antigen-presenting capacity of the inoculated Langerhans cells competed with down-regulatory resident ocular cells and processed antigen in a manner that promoted the development of systemic DTH. Conversely, it might be suggested that the host Langerhans cells neutralized a putative inhibitor molecule present in normal aqueous humor. T h e latter suggestion is difficult to reconcile when one considers that ACAID can also be ablated by splenectomy, a procedure that would not be expected to alter the makeup of the aqueous humor. OF ACAID-INSIDE C. EXPRESSION
AND
OUTSIDE THE EYE
1. Effect on Intraocular Tumor Allografts In its simplest form ACAID has been offered as the underlying mechanism to explain the phenomenon of immunological privilege in the anterior chamber (99, 100). According to this hypothesis, allografts placed into the anterior chamber escape or delay immunological rejection, due to an active down-regulation of systemic cell-mediated immunity. Immune privilege, therefore, is a dynamic ongoing process manifested not only in the eye, but systemically as well. Using the DBA/2 mastocytoma (P815) model of ACAID, we have suggested that the antigen-specific down-regulation of DTH and the preservation of allospecific CTL and antibody responses suggested that immune privilege and ACAID translated into the active suppression of DTH (101). Moreover, in this model of ACAID, BALB/c hosts harboring anterior chamber P815 tumor allografts were incapable of rejecting orthotopic DBA/2 skin allografts. The suppression of skin allograft rejection was antigen specific, as the same hosts were able to reject third-party
208
JERRY Y. NIEDERKORN
skin allografts (102). However, as previously mentioned, ACAID can be ablated by splenectomy. Three interesting results occur in splenectomized BALB/c mice. First, and perhaps most striking, is the obvious loss of immune privilege: Intracameral P8 15 tumors undergo a violent rejection in splenectomized BALB/c mice (72). A second, and equally important, event is the restoration of systemic cellular immunity: Orthotopic DBA/2 skin allografts are not only rejected, but the tempo is indicative of a second-setrejection (i.e., the anterior chamber inoculum sensitized the host). The third noteworthy finding is the restoration of DTH that occurs in splenectomized mice. This mosaic of findings, summarized in Table I, has led to the following conclusions: (1) ACAID is principally an antigen-specific downregulation of systemic DTH; (2) rejection of minor H-incompatible orthotopic skin allografts is DTH dependent; (3) the rejection of intraocular tumor allografts is DTH dependent; and (4) progressive growth of tumor allografts occurs in the face of antigen-specific CTL and antibody. Coincidentally, these results also strongly suggest that skin allografts involving only minor H histoincompatibilities are rejected by a DTHdependent process if ACAID is ablated; however, if ACAID is established, the skin allografts survive in the face of allospecific CTLs and serum alloantibody. The weight of evidence supports the conclusion that the intraocular tumor allografts are also rejected by DTH-dependent mechanisms (Table I). In splenectomized BALB/c mice the acquisition of positive DTH coincides with the onset of intraocular tumor rejection. Moreover, the histopathological features of intraocular tumor rejection are reminiscent of a vigorous DTH reaction, with evidence of infarction of the microvasculature feeding the tumorous mass, ischemic necrosis e n masse, erosion of the vascular endothelium, and extensive “innocent bystander” damage. Although ablation of ACAID results in the prompt rejection of the intraocular tumor allograft, the consequences are severe, resulting in total destruction of the eye.
2. Effect on Ocular HSV Infections HSV- 1 infections of the cornea are one of the leading causes of corneal blindness in the United States. The damaging effects of corneal HSV infections are believed to be primarily due to the cytopathic effect of the virus, although there is a growing body of evidence suggesting that stromal disease is an immune-mediated process. For example, Metcalf et al. (103) demonstrated that athymic nude mice failed to develop stromal
TABLE I EFFECT OF ACAID ON IMMUNOLOGICAL PRIVILEGE AND INTRAOCULAR TUMOR ALLOGRAFT BEHAVIOR^ Immunological profile Host Eusplenic Splenectomized
Fate of allografts
ACAID
DTH
Ts
CTL
Antibody
Orthotopic skin
Intraocular tumor
+
-
+
+ +
+
Permanent survival Second-set rejection
Progressive growth Rejection (DTH-like pathology)
-
+
-
-
a Data are summarized from references cited in text. DTH, Delayed-type hypersensitivity; Ts, T suppressor cells that downregulate DTH, as shown in adoptive transfer assays; CTL, cytotoxic T lymphocyte activity.
210
JERKY Y. NIEDEKKORN
lesions following corneal infection with HSV- 1, while euthymic littermates displayed typical keratitis following ocular infection. However, nude mice could be rendered susceptible to stromal keratitis if provided T lymphocytes from sensitized euthymic donors (104). With this in mind, Ksander and Hendricks (105) attempted to mitigate HSV-1 keratitis in mice by inducing ACAID prior to corneal infection. As expected, topical corneal infections with HSV- 1 produced potent DTH and CTL responses. However, anterior chamber inoculation of HSV- 1 prior to topical infection resulted in a profound suppression of both DTH and CTL responses and virtually complete protection from corneal stromal lesions produced by HSV- 1. Further studies by Hendricks et al. (106) demonstrated that anterior chamber inoculation of a mutant strain of HSV- 1 having a deletion of the gene encoding the glycoprotein C resulted in the induction of normal DTH responses, but a conspicuous absence of CTL reactivity against HSV-1. Hosts primed in this manner had significantly reduced stromal disease following topical infection with the wild-type HSV- 1. Collectively, the results implicate HSV-specific CTLs, not DTH, in the pathogenesis of HSV stromal keratitis in this murine model. Thus, these studies not only offer a plausible hypothesis to account for the pathogenesis of HSV stromal keratitis, but they also offer a novel approach for using ACAID to reduce the severity of the putative immune-mediated ocular disease. Anterior chamber inoculation of infectious HSV- 1 not only induces the suppression of systemic DTH, but also results in a curious pattern of ocular inflammatory diseases in which the retina of the contralateral uninoculated eye undergoes inflammation, necrosis, and complete destruction (107, 108). Paradoxically, the retina of the eye initially inoculated with HSV-1 is preserved (107, 108). Atherton and Streilein (109) have shown that the development and severity of contralateral retinitis were intimately related to the spread of virus from the inoculated eye to the retina of the contralateral eye. Moreover, it has been suggested that in the absence of ACAID (e.g., hosts immunized via nonocular routes), virus-specific DTH responses contribute to the clearance of virus and prevent viral replication in the central nervous system, which in turn prevents the spread of virus to the contralateral eye (109, 110). Thus, the development of normal DTH reactivity to viral antigens would be expected to protect the host from contralateral retinitis induced by anterior chamber inoculation of HSV. It remains unclear, however, whether the suppression of DTH that occurs following anterior chamber inoculation promotes the development of contralateral retinitis.
IMMUNOLOGY OF THE EYE
21 1
D. EFFECTOR MECHANISM FOR SUSTAINING ACAID As previously described, a wide range of immunogens are capable of inducing ACAID. A common feature shared by all models of ACAID is the appearance of an antigen-specific suppression of DTH responses (99, 100).Wetzig et al. (74) were the first to demonstrate that the suppression induced via anterior chamber inoculation of hapten-derivatized splenic cells could be adoptively transferred with T cells that acted at the efferent mode, but did not bear detectable cross-reactive idiotype surface receptors. Subsequent studies by Waldrep and Kaplan (75) confirmed the presence of an efferent-level T suppressor system consisting of a primary pathway involving a cyclophosphamide-sensitive suppressor T cell and a secondary suppressor pathway that is antigen-nonspecific and mediated by a cyclophosphamide-resistant suppressor T cell population. The suppression of DTH induced by anterior chamber presentation of alloantigens, like the hapten models, is mediated by a suppressor T cell population that acts primarily at the efferent mode (1 11). However, unlike the previously described hapten models, the suppressor T cells in the alloantigen model of ACAID are I-J sensitive and contain a Thy-1.2' L3T4- Lyt-2.2' subpopulation and a second subpopulation that is Thy- 1 L3T4' Lyt-2.2- (11 1). In all studies to date, the suppressor cell population capable of suppressing systemic DTH following adoptive transfer to naive recipients resides in the spleen (99, 100). There is agreement among all current models of ACAID that, during the initial stages of ACAID, there is a strict requirement for both an intact eye and an intact spleen. Early removal of either the injected eye or the spleen prevents the induction of ACAID. Removal of the eye (i.e., within 7 days of anterior chamber inoculation) fails to induce either a positive or a negative signal for DTH. By contrast, splenectomy has a striking effect on the development of DTH in anterior chamber-primed animals: Vibrant DTH responses are detected, intraocular tumor allografts undergo a necrotizing rejection process, and orthotopic skin allografts are swiftly rejected in a manner indicative of second-set immunization (73). However, the strict requirement for the spleen is limited to the first 10 days following anterior chamber inoculation of alloantigen. Therefore, ACAID can be sustained and DTH thoroughly suppressed in the absence of an intact spleen, provided the spleen is in place during the 10 days following anterior chamber inoculation of alloantigen (73). Removal of the spleen after this does not alter the panorama of immunological findings that characterize ACAID. Therefore, the suppressor cells necessary to sustain the suppression of DTH have emigrated from the +
212
JERKY Y. NIEDEKKORN
spleen and are disseminated to other anatomical regions by day 10 postinoculation. Recent findings by Ferguson et al. (112) have provided further insights into the role of the spleen in the induction and maintenance of ACAID. Using trinitrophenyl (TNP)-derivatized T lymphocytes, these investigators proposed that hapten-derivatized T suppressor-inducer (Tsi) cells elaborated soluble suppressor-inducer factor that was antigen specific and immunoglobulin H restricted. The putative suppressor-inducer molecule was believed to leave the anterior chamber and enter the serum and was eventually filtered within the spleen, where it induced the development of a T suppressor-effector (Tse) population that maintained the persistent down-regulation of systemic DTH responses to TNP. The authors proposed that the serum-borne Tsi factor could not be detected in the eusplenic host, due to its rapid removal by high-affinity Tse cells in the spleen. However, sera from splenectomized hosts were capable of inducing the suppression of DTH when passively transferred to naive hosts. Collectively, these results suggest that in the hapten model of immunological privilege, derivatized T cells elaborate a molecule capable of inducing splenic effector cells that actively maintain down-regulation of DTH. However, it is not known whether a similar condition occurs with alloantigens or soluble antigens presented into the anterior chamber. Other antigens induce ACAID without being derivatized to histocompatible T cells prior to anterior chamber inoculation. Therefore, to integrate this paradigm to fit other antigen systems would require that antigen processing and presentation to T cells occur within the anterior chamber in order for the host T cells to elaborate a similar Tsi moleculea plausible, but unproven, proposition. IV. Effect of Immune Regulation on lntraocular Tumor Rejection
The high incidence of spontaneous neoplasms that occurs in patients suffering from immunological disorders and in transplant recipients subjected to prolonged immunosuppression has been cited as evidence in support of the immune surveillance theory (1 13). It is reasonable to suspect that the same situation occurs with immunologically privileged sites, such as the anterior chamber of the eye, where immunological censorship creates an environment that would be expected to thwart the induction and expression of immunological surveillance. Contrary to the immune surveillance theory, this is not the case (2). The incidence of spontaneous neoplasms in immunologically privileged sites is not higher than that occurring at other anatomic regions. In the case of the eye,
IMMUNOLOGY OF T H E EYE
213
tumors of the cornea are virtually unknown and the most common intraocular tumors occur in the posterior compartments of the eye, not in the anterior chamber (1 14). T h e topics of intraocular tumorigenesis and immune regulation raise interesting questions in the context of ACAID. For example, are highly immunogenic tumors that undergo spontaneous immunological rejection at extraocular sites exempt from immunological rejection within the immunologically privileged confines of the anterior chamber? If so, are DTH-dependent mechanisms disqualified from participation? Are CTLmediated tumor rejection mechanisms preferentially induced and executed in the anterior chamber? We have addressed these and other questions relating to tumor-specific immunity in the anterior chamber, using immunogenic syngeneic regressor tumors. P91 mastocytoma, a highly immunogenic mutant of P8 15 mastocytoma, and UV5C25, an immunogenic ultraviolet light-induced fibrosarcoma, undergo immunological rejection following subcutaneous transplantation in syngeneic hosts (1 15, 116).Rejection of both tumors is T cell mediated and tumor specific (115, 116). Following anterior chamber transplantation, P9 1 mastocytoma (of DBAI2 origin) grows progressively for approximately 3 weeks. Between the 3rd and 4th weeks posttransplantation, an intense inflammatory response is elicited and culminates in ischemic tumor necrosis en muse (115).Tumor resolution is completed within another 5-7 days, leaving the eye irreparably damaged (i.e., phthisis bulbi). The histopathological features of the resolving tumors are characteristic of DTH lesions and include (1) microvascular infarction, (2) erosion of vascular endothelium, (3) perivascular cuffing, (4) ischemic necrosis en muse,and (5) extensive innocent bystander damage. Moreover, immunological findings supported the proposition that rejection is predominantly a DTH-mediated process (117). Although the hosts develop tumor-specific CTLs and antibody, neither appears to be involved in the rejection of intraocular P91 tumors. Surprisingly, high levels of tumor-specific CTLs can be detected in regional lymph nodes as well as in the spleen, yet there is never any evidence of a lymphocytic infiltrate within the resolving intraocular tumors. Likewise, antibody is apparently not involved in intraocular P9 1 tumor rejection: Efforts to passively transfer tumor rejection have failed (117). Moreover, tumor rejection occurs in the absence of detectable serum antibody in splenectomized hosts and in the absence of complement in C3-deficient hosts (1 14). T h e results emphasize the importance of a tightly regulated immune response in the eye. Although the intense tumor-specific DTH responses eliminated the intraocular tumor and spared the host’s life, it did so at great expense: blindness.
214
JERRY Y. NIEDERKORN
The pathophysiology of intraocular P9 1 tumor rejection is of more than casual interest, not merely an interesting immunological anomaly. The most common intraocular malignancy of childhood, retinoblastoma, has one of the highest frequencies of spontaneous resolution of any human tumor, including nonocular tumors (118). Spontaneous resolution of retinoblastoma is known to produce extensive innocent bystander damage to ocular tissues and culminate in an atrophic eye-a pattern of tumor resolution not unlike the previously mentioned P9 1 rejection (1 18). Although these sequelae rid the eye of a life-threatening malignancy, the cost is high. Is intraocular tumor rejection always destined to result in immune-mediated blindness? Studies involving another regressor tumor, UV5C25, indicated that intraocular tumor resolution could occur via a CTL-mediated mechanism and simultaneously exclude the participation of DTH responses in oculi (1 16). Immunological rejection of intraocular UV5C25 tumors occurs over a prolonged period of 10-14 days, while maintaining the normal anatomical integrity of the affected eye. The immunological and histopathological features of UV5C25 rejection are consistent with a CTL-mediated process: (1) Thy-l+ Lyt-2' L3T4- lymphocytes are bound to individual tumor cells in situ, (2) resolution occurs by piecemeal necrosis of individual tumor cells, (3) ischemia and innocent bystander damage are absent, and (4) intraocular tumor resolution can be adoptively transferred with CD8' tumor-infiltrating lymphocytes isolated from resolving intraocular tumors (116, 119). Not only is UV5C25 tumor rejection a CTL-dominated process, but it occurs in hosts possessing tremendous DTH reactivity to UV5C25 tumor-specific antigens (116, 119). Therefore, it appears that the effector phase of DTH is actively excluded from the intraocular tumor nidus. The result of this exclusion is elimination of the malignancy without jeopardizing vision. Neither of these two categories of intraocular tumors is irrevocably committed to follow a specific pathway. For example, hosts bearing intraocular P9 1 tumors develop impressive tumor-specific CTL activity, yet there is no evidence of lymphocytes infiltrating resolving intraocular tumors. However, if the contralateral eye is challenged with P9 1 tumor cells, the intraocular tumor in the second eye undergoes a pattern of rejection characterized by a mononuclear infiltrate of predominantly Thy-1 Lyt-2' lymphocytes. Rejection occurs by piecemeal necrosis, with negligible innocent bystander damage. Thus, destructive and nondestructive patterns of intraocular tumor rejection can occur simultaneously in the same host. This further suggests that an active local process excludes cytolytic T cells from entering the first eye, but not the second eye. Recent studies indicate that this converse phenomenon occurs with +
IMMUNOLOGY OF T H E EYE
215
intraocular UV5C25 tumors. BALB/c mice reject intraocular UV5C25 tumors by a process dominated by cytolysis of individual tumor cells and mediated by infiltrating Thy- 1 Lyt-2’ lymphocytes, even though the hosts display intense systemic DTH responsiveness to the tumor antigens. However, the exclusion of DTH responses from the tumor-containing eye can be altered. BALB/c hosts treated systemically with anti-CD8 antibody develop rapidly growing intraocular tumors that eventually undergo a necrotizing pattern of tumor rejection, characterized by extensive innocent bystander damage and the complete destruction of the eye (unpublished observations.) Results from these studies, as well as investigations of ACAID, have led us to propose that two basic patterns of intraocular tumor rejection are available to the immunocompetent host (1 16, 119). T h e first pattern involves minimal damage to normal host tissues and occurs by piecemeal necrosis by cytolytic T lymphocytes. The second pattern is destructive not only to the intraocular tumor, but to the anatomical integrity of the entire eye. The latter form of immune rejection is characteristic of a DTH-like lesion in which tumor necrosis occurs en musse. Hemorrhagic necrosis of the tumor and the entire eye attest to the antigen-nonspecific characteristic of this form of intraocular tumor rejection. These findings demonstrate that intraocular immunological privilege is manifested to varying degrees and should not be viewed as an all-or-none proposition. Moreover, understanding intraocular immune regulation has important implications not only for intraocular tumors, but for more common sightthreatening ocular autoimmune diseases. +
V. Immune Regulation and Autoimmune Uveitir
Inflammatory diseases of the eye are important causes of blindness in the United States and throughout the world. The region of the eye composed of the choroid, ciliary body, and iris (i.e., the uveal tract) is particularly vulnerable to immune-mediated diseases, collectively termed “uveitis.” The notion that autoimmunity was the underlying cause of uveitic conditions was proposed by Elschnig over 80 years ago (120) and has been confirmed in numerous studies (12 1,122). In 1965 Wacker and Lipton ( 123) demonstrated that immunogens extracted from retinal tissue were effective in inducing an inflammatory disease in the uveal tracts of experimental guinea pigs. Scores of subsequent studies have established that systemic immunization with retinal antigens consistently induces an ocular inflammatory disease designated experimental autoimmune uveitis (EAU). EAU can be induced either with the well-characterized S antigen, a 48-kDa protein involved in light signal transduction, or with interphotoreceptor
216
JERKY Y. NIEDERKORN
retinoid-binding protein (IRBP), a 140-kDa protein that transports vitamin A derivatives between the photoreceptor cells and the retinal pigment epithelium (124, 125). Both proteins are major components of the photoreceptor cell layer, which is the primary target of immunemediated damage in EAU (1 2 1, 122). EAU has been a valuable tool for studying the immunopathogenesis of a variety of inflammatory diseases affecting primarily the posterior segment of the human eye (Fig. l ) , including sympathetic ophthalmia, birdshot retinochoroidopathy ,VogtKoyanagi-Harada syndrome, and Behcet’s syndrome (12 1, 122). Numerous studies have established that EAU is a T cell-mediated disease because (1) athymic nude rats do not develop the disease following immunization with S antigen (126); (2) EAU is inhibited completely by treatment with the immunosuppressive agent cyclosporine ( 1 27); (3) EAU can be adoptively transferred with helper T cell lines (128, 129) or T cells from donor rats immunized with S antigen (130);(4) the ability to transfer EAU with lymphoid cells can be eliminated by treatment with anti-CD4 antibody plus complement (130); (5) the development of EAU correlates with the appearance of systemic DTH to retinal S antigen (130); and (6) attempts to transfer EAU with hyperimmune serum have consistently failed (131). The spectrum of immunological findings and the histopathological features of EAU in rodents and subhuman primates indicate that this disease is predominantly a DTH-mediated disorder directed at a well-defined anatomical region of the eye: the posterior uvea and retina. Although IRBP and S antigen are retinal proteins residing close to each other, the pathogenesis of EAU induced by these two antigens differs considerably. EAU induced by S antigen is an acute explosive disease with an active phase lasting 7-10 days, during which the photoreceptor cell layer is totally destroyed (121, 122). The destruction of the photoreceptor layer is believed to prevent the recurrence of intraocular inflammation in S antigen-induced EAU, since the destruction of this cell layer eliminates the source of offending antigen in situ. By contrast, EAU induced by IRBP occurs later, the duration is longer, and the course is much less acute than EAU induced by S antigen (121,122).In addition to eliciting EAU, both retinal antigens also induce inflammation of the pineal gland, an organ that resides in another immunologically privileged location-the brain. VI. Uveitir: A Breakdown in Self-Tolerance and Immune Privilege
EAU has been an attractive model for studying the pathogenesis of a variety of human inflammatory ocular diseases, such as sympathetic ophthalmia. Key, and still unanswered, questions are what is the etiology for
IMMUNOLOGY OF T H E EYE
217
autoimmune diseases of the eye, and if retinal antigens are the inflicting immunogens, what triggers the autoimmune response to these proteins? In the simplest scenario one might suggest that physical trauma or infection might result in the release of large immunogenic quantities of retinal antigens. Moreover, a perforating injury to the eye might introduce bacterial contaminants that provide an adjuvant effect for the released ocular autoantigens. It has been proposed that this scenario occurs in the development of sympathetic ophthalmia, a relatively rare sequela of penetrating ocular trauma (132- 134). By clinical definition, sympathetic ophthalmia is a bilateral uveitis in which inflammation occurs first in the injured eye (inciting eye) and is followed by inflammation in the second, uninjured, eye (sympathizing eye). The appearance of inflammation in the second eye indicates that the pathological sequelae are the result of noninfectious organ-specific (i.e, autoimmune responses. Although immunohistological analyses have revealed a varied spectrum of inflammatory cells in sympathetic ophthalmia lesions (134, 135), the consensus is that the immunopathogenesis is largely due to an intraocular DTH reaction (134). The series of events that provoke immunopathogenic processes that culminate in EAU, sympathetic ophthalmia, and similar autoimmune diseases of the eye remains a mystery. The facility with which EAU can be induced in a wide range of laboratory animals and subhuman primates suggests that the eye is at considerable risk for the development of spontaneous autoimmune diseases. However, the low incidence of uveitis in the human population indicates that effective mechanisms guard against autoimmune responses within the eye. Maintenance of self-tolerance in the eye, like any other organ, can be achieved by four basic strategies: (1) sequestration of autoantigens from the immune system, (2) clonal deletion of self-reactive immune cells, (3) active suppression of anti-self-immune cells, and (4) programming immature cells with a down-regulatory signal that renders them hyporeactive without killing them (i.e.,“clonal anergy”). These strategies are not mutually exclusive, and there is evidence to suggest that at least two of the four strategies are involved in the self-tolerance of ocular autoantigens. Sequestration of self-antigens behind impervious anatomical barriers has been offered as the simplest strategy for preventing autoimmunity, especially in the case of ocular autoantigens. The sequestration of lens proteins within an impervious collagenous capsule is often cited as an example of this pathway for self-tolerance (136). The importance of this strategy in protecting the eye from autoimmune responses to uveal and retinal antigens is debatable. Under normal conditions, potential retinal antigens are unavailable for immunological recognition at the level of the inductive stage (i.e., the afferent mode) of the immune response, due
218
JERRY Y . NIEDERKORN
to their sequestration within the photoreceptor layer of the eye (12 1, 122, 137). The possibility that the deletion of retinal antigen-specific T cell clones is involved in ocular self-tolerance is remote. Recently, Caspi et al. (138) demonstrated that EAU could be induced in selected strains of inbred mice and that susceptibility seemed to be correlated with the H-2' MHC haplotype. Although the induction of EAU was limited to but a few mouse strains, all mouse strains tested demonstrated lymphocyteproliferative responses and antibody titers to both IRBP and S antigens. Thus, all hosts possessed potentially reactive T cell clones, but only a few host strains manifested autoimmunity. A similar situation may occur in humans. It has been reported that T cells capable of responding to retinal S antigen can be detected in the peripheral blood of normal human subjects with no history of uveitis (139). Depending on how we define clonal anergy and clonal suppression, either or both mechanisms could be involved in self-tolerance to autoantigens of the retina and the lens. A feeble response to autoantigens might be construed as evidence of clonal anergy or an incomplete manifestation of a suppressor mechanism. Indirect evidence suggests that a partial suppressor system is expressed in autoimmune responses to retinal antigens. Induction of EAU in mice is not only restricted to certain inbred strains, but the experimental protocol requires the use of potent adjuvants (i.e., Mycobacterium tuberculosis and Bordetella pertussis ) and treatment of the host with cyclophosphamide (138), a drug commonly used for eliminating suppressor cells in vivo (140). Recent studies by Caspi et al. (141) demonstrated that organ-resident nonlymphoid cells of the retina (retinal glial cells = Miiller cells) exercised suppressive effects on T helper lymphocytes specific for retinal antigens. The prospect that Muller cells are instrumental in preventing autoimmune responses in the uvea and the retina is especially appealing for several reasons. Muller cells are a major component of the neural retina, where they are closely associated with the photoreceptor cells and ensheath the retinal blood vessels (142). As a result of their anatomically strategic location, they are in a position to serve as a second barrier (after the retinal vascular blood-brain barrier) through which an infiltrating lymphocyte (e.g., autoreactive T cells?) must pass in order to enter the posterior segment of the eye. Moreover, it has been shown that Miiller cells exercise their suppressive effect on retinal antigen-specific cells by direct cell-cell contact (14 1)-a situation uniquely suited for the retina and Miiller cells. Recent studies have demonstrated that Muller cells become activated and undergo proliferation when exposed to lymphokines secreted by T
IMMUNOLOGY OF T l I E EYE
219
helper lymphocytes (14 1). This blend of characteristics leads to the following scenario: Due to their strategic location, retinal Muller cells could serve as an anatomical barrier for the migration of T cells into the posterior segment of the eye. If antigen-specific T cells find their way to the retina, the close physical association between the photoreceptor cells and the Miiller cells places the latter close to lymphokines released by autoreactive T cells. As a result, resident Muller cells would be activated and would function to dampen further activity by the autoreactive T cells. Finally, it has been reported that not only do Muller cells function as immunoregulatory cells, but that they are also active in healing and in scar formation in the terminal stages of EAU (143, 144). Although much needs to be learned and confirmed regarding the function of retinal Miiller cells, their potential in maintaining immunological homeostasis in the eye is a provocative topic for further analysis. In addition to the organ-resident Muller cells, there may be additional mechanisms for actively suppressing autoimmune responses directed at retinal antigens. Mizuno et al. (80, 81) have reported that intracameral injection of retinal S antigen induced suppression of systemic DTH responses to S antigen and significantly reduced the severity of uveitis induced by extraocular immunization with S antigen. Other studies in rats support the conclusion that the induction of ACAID to S antigen mitigates EAU. A suppressor T lymphocyte line was isolated from the spleens of rats primed via the anterior chamber with soluble retinal S antigen (145). Adoptive transfer of this suppressor T cell line was found to downgrade EAU in actively immunized syngeneic hosts. It is not known whether a similar suppressor T lymphocyte population is constitutively induced and maintained at a low level in normal individuals. It is interesting that in this experimental context an ocular antigen (i.e., S antigen) was introduced through the anterior chamber and the resulting suppressor T cell population served to mitigate an ocular autoimmune disease. One is tempted to entertain the hypothesis that ACAID may be an integral regulatory mechanism for maintaining the immunological homeostasis of the eye. VII. Is Cataract Formation an Immune-Mediated Disease?
No discussion on the immunology of the eye and autoantigens would be complete without commenting on the autoimmune potential of lens crystallins. As stated earlier, one of the simplest mechanisms for explaining self-tolerance is the anatomical sequestration of autoantigens.
220
JERRY Y . NIEDERKORN
By erecting an anatomical barrier between the host’s immune system and potential autoantigens, self-tolerance is assured. The lens serves as an excellent example of anatomical sequestration. Housed within a collagenous capsule, the lens is shielded from the systemic circulatory and lymphatic systems. Moreover, the lens capsule is suspended within an immunologically privileged compartment: the anterior chamber of the eye (Fig. 1). The early sequestration of the lens during ontogeny implies that its crystallins could act as autoantigens if released into the peripheral circulation. The suspicion that lens components are potential autoantigens was suggested over 80 years ago by Uhlenhuth (146), who demonstrated that antibodies to bovine lens reacted with lens proteins from a wide range of vertebrate species. Verhoeff and Lemoine (147) subsequently suggested that such immune factors might be related to the persistent ocular inflammatory diseases that occasionally follow penetrating injury to the eye. Recently, Angunawela ( 148) reconsidered the hypothesis that cataractogenesis is an autoimmune phenomenon, based on findings indicating a significant increase in the incidence of antibody directed against lens antigens in the sera of patients with cataractous lenses and in the sera of diabetic patients. Moreover, immunoglobulin deposits were detected on the cells of the cataractous lenses removed from nondiabetic and diabetic patients. The close relationship between diabetes and senile cataracts is well known; however, interpreting the presence of lens-specific antibodies in the serum of cataractous patients should be viewed with caution, since similar antibodies have been demonstrated in approximately 50% of the normal individuals tested (149, 150). It is possible that an increased incidence of antibody in these individuals was the result of cataractogenesis, rather than the cause. Nonetheless, these findings bring to our attention some interesting issues regarding immune regulation in the eye and the possible consequences of a malfunction of this regulation. It has long been assumed that the lens capsule was impervious to leakage of lens proteins and prevented their escape into the aqueous fluid of the anterior chamber. However, a and y lens crystallins can be readily detected in the aqueous humor of normal individuals (151). Leakage of lens antigens into the aqueous humor and transport to the outflow system may occur with sufficient frequency to explain the presence of serum antibody against lens crystallins in approximately 50% of the normal population (149,150). Moreover, homologous lens antigens are known to induce antibody synthesis in a variety of mammals, without inducing demonstrable T cell-mediated immune responsiveness ( 152). It is tempting to suggest that a constant leakage of lens antigens into the anterior
IMMUNOLOGY OF THE EYE
22 1
chamber results in the induction of ACAID. As in the case of the retina and the uvea, it appears that a vulnerable ocular tissue is protected from autoimmune attack by anatomical sequestration and possibly by a dynamic immunoregulatory process. This hypothesis, although appealing, remains to be verified. VIII. Conclusions
This chapter has addressed some of the unique immunological characteristics of the eye. The goal was not to provide an exhaustive review of ocular immunology, but to focus on the curious immunological “ground rules” peculiar to this organ. T h e weight of evidence suggests that the immune system is carefully regulated within the eye and, to a lesser degree, at the ocular surface (i.e., the cornea). Teleologically, it appears that this regulation is designed to restrict the expression of exuberant inflammatory responses-namely, DTH-that carry a heavy burden of innocent bystander damage to the ocular tissues possessing few regenerative capacities, if any. Although DTH is excluded, CTLs and antibody provide adequate coverage to protect the eye from pathogens without damaging juxtaposed normal ocular tissues. T h e rare incidence of immune-mediated inflammatory diseases of the eye is a testament to the effectiveness of the ocular irnmunoregulatory circuit. Gaining a better understanding of this circuitry will offer immunological implications that extend well beyond the boundaries of the eye. ACKNOWLEDGEMENTS These studies were supported in part by National Institutes of Health grants EY05631, CA30276, and EY07641 and by an unrestricted grant from Research to Prevent Blindness, Inc. The author is a Research to Prevent Blindness-Olga Keith Wiess Scholar.
REFERENCES 1. Silverstein, A. M. (1968). In “Clinical Aspects of Immunology” (P. G. Cell and R. R. A. Coombs, eds.), pp. 1160-1 175. Blackwell, Oxford. 2. Head, J . R., and Billingham, R. E. (1985). Perspect. Baol. Med. 29, 115-131. 3. Trevor-Roper, P. D. (1972). In “Corneal Grafting” (T. A. Casey, ed.), pp. 1-10. Appleton-Century-Crofts, New York. 4. Darwin, E. (1796). In “Zoonornia or the Laws of Organic Life” Vol. 2. J. Johnson, London. 5. Bigger, S. L. (1837). Dublin J . Med. Sci. 11,408-447. 6. Kissam, R. S. (1844). N.Y.J. Med. 2,281-289. 7. Zirrn, E. (1906). Albrecht von Graefes Arch. Ophthubol. 64,580-593. 8. Smith, R. B., McDonald, H. R.,Nesburn, A. B., and Minckler, D. S. (1980). Arch. Ophthalmol. (Chicago) 98, 1226- 1229. 9. Polack, F. M. (1973). I n “Corneal Graft Failure” (R. Porter and J. Knight, eds.), pp. 127-150. Elsevier/North-Holland Publ., Amsterdam.
222
JERRY Y. NIEDERKORN
10. Silverstein, A. M., and Khoudadoust, A. A. (1973). In “Corneal Graft Failure” (R. Porter and J. Knight, eds.), pp. 105-125. ElseviedNorth-Holland Publ., Amsterdam. 11. Billingham, R. E., and Boswell, T. (1953). Proc. R. SOC.London 141, 392406. 12. Khodadoust, A. A., and Silverstein, A. M. (1966). Sum. Ophthalmol. 11, 435-443. 13. Chandler, J. W., Ray-Keil, L., and Gillette, T. E. (1983). Cum. Eye Res. 2, 387-397. 14. Streilein, J. W., McCulley, J., and Niederkorn, J. Y. (1982). Invest. Ophthalmol. Visuul Sci. 23,489-500. 15. Mayer, D. J., Daar, A, S., Casey, T. A., and Fabre, J. W. (1983). Transplant. Proc. 15, 126-129. 16. Whitsett, C. F., and Stulting, R.D. (1984). Invest. Ophthalmol. Visual Sci. 25, 5 19-524. 17. Manski, W., and Whiteside, T. L. (1974). Invest. Ophlhalmol. Visual Sci. 13, 935-955. 18. Treseler, P. A., Foulks, G. N., and Sanfilippo, F. (1984). A m . ] . Ophthalmol. 98,763-772. 19. Peeler, J. S., Callanan, D. G., Luckenbach, M. W., and Niederkorn, J. Y. (1988).J. Exp. Med. 168, 1749-1766. 20. Peeler, J. S., and Niederkorn, J. Y. (1986).J. Immunol. 136,4362-4371. 21. Treseler, P. A., and Sanfilippo, F. (1986). Transplantation 41,248-252. 22. Treseler, P. A., Foulks, G. N., and Sanfilippo, F. (1986). Transplantation 41, 229-234. 23. Basu, P. K., Miller, I., and Ormsby, H. L. (1960). Am. J . Ophthalmol. 49, 5 13-5 15. 24. Chi, H. H., Teng, C. C., and Katzin, H. M. (1965). Ophthalmology (Rochester, Minn.) 59, 186- 191. 25. Hanna, C., and Irwin, E. S. (1962). Arch. Ophthalmol. (Chicago) 68,810-817. 26. Polack, F. M., Smelser, G. K., and Rose, J. (1964). Am. J . Ophthalmol. 57, 67-78. 27. Khodadoust, A. A. (1973). In “Corneal Graft Failure” (R. Porter and J. Knight, eds.), pp. 151-167. ElseviedNorth-Holland Publ., Amsterdam. 28. Batchelor, J. R., Casey, T. A., Gibbs, D. C., Lloyd, D. F., Werb, A., Prasad, S. S., and James, A. (1976). Lancet 1,551-554. 29. Matoba, A. Y., Peeler, J. S., and Niederkorn,J. Y . (1986).Invest. Ophthalmol. Visuul Sci. 27,1244- 1254. 30. Callanan, D., Peeler, J., and Niederkorn, J. Y. (1988). Transplantation 45, 437-443. 31. Williams, K. A., and Coster, D. J. (1985). Invest. Ophthalmol. Visuul Sci. 26, 23-30. 32. Maumenee, A. E.(1951). Am. J. Ophthalmol. 34, 142-152. 33. Ross, J., Callanan, D., Kunz, H., and Niederkorn, J. Y. (1989). Invest. Ophthalmol. Visual Sci., Suppl. 30,442. 34. Khodadoust, A. A., and Silverstein, A. M. (1972). Invest. Ophthalmol. 11, 137-148. 35. Klareskog, L., Forsum, U., Tjernlund, U. M., Rask, L., and Peterson, P. A. (1979). Invest. Ophthalmol. Visual Sci. 18,310-312. 36. Rodrigues, M. M., Rowden, G., Hackett, J., and Bakos, I. (1981). Invest. Ophthalmol. Visual Sci. 21,759-762.
IMMUNOLOGY OF T HE EYE
223
37. Gillette, T. E., Chandler, J. W., and Greiner, J. V. (1982). Ophthalmology (Rochester, Minn.) 89, 700-705. 38. Bergstresser, P. R., Fletcher, C. R., and Streilein, J. W. (1980). J. Invest. Dermatol. 74, 77-82. 39. Lafferty, K. J., Prowse, S. J., Silmenonvic, C. J., and Warren, H. S. (1983). Annu. Rev. Immunol. 1, 143-173. 40. Snell, G. D. (1957).Annu. Rev. Microbiol. 11,439-458. 4 1. Peeler, J. S., Niederkorn, J. Y., and Matoba, A. (1985). Invest. Ophthalmol. Visual Sci. 26, 1516-1523. 42. Steinmuller, D. (1985). Transplantation 40,229-233. 43. Blanchard, D., Els, C. V., Aubry, J. P., de Vries, J. E., and Spits, H. (1988). J. Immunol. 140, 1745- 1752. 44. Lafferty, K. J. (1985). Transplant. Proc. 17, 1549-1551. 45. Wheelahan, J., and McKenzie, I. F. C. (1987). Transplantation 44,273-280. 46. Rosenberg, A. S., Mizuochi, T., Sharrow, S. O., and Singer, A. (1987). J . Exp. Med. 165, 1296-1315. 47. Callanan, D. G., Luckenbach, M. W., Fischer, B. J., Peeler, J. S., and Niederkorn, J. Y. (1989).Invest. Ophthalmol. Visual Sci. 30,413-424. 48. Peeler, J. S., and Niederkorn, J. Y. (1987). Transplant. Proc. 19,316-319. 49. Ray-Keil, L., and Chandler, J. W. (1985). Transplantation 39,473-477. 50. Young, E., Stark, W. J., and Prendergast, R. A. (1985). Invest. Ophthalmol. Visual Sci. 26,57 1-574. 51. Donnelly, J. J., Weiye, L., Rockey, J. H., and Prendergast, R. A. (1985). Invest. Ophthalmol. Visual Sci. 26,575-580. 52. Streilein,J. W., Toews, G. B., and Bergstresser, P. R. (1979).Nature (London) 282,326-327. 53. Katarni, M., White, D. J. G., and Watson, P. G. (1989). Transplant. Proc. 21, 3147-3 149. 54. Williamson, J. S. P., DiMarco, S., and Streilein, J. W. (1987).Invest. Ophthalmol. Visual Sci. 28, 1527-1532. 55. Hazlett, L. D., Moon, M., Dawisha, A,, and Berk, R. S. (1986). Cuw. Eye Res. 5,343-355. 56. Lewkowicz-Moss,S. J., Shirneld, C., and Easty, D. L. (1985). Ophthalmic Res. 17,202-208. 57. Asbell, P. A., and Kamenar, T. (1987). Cum. Eye Res. 6, 179-186. 58. Pepose, J. (1989). Curr. Eye Res. 8,851-858. 59. Niederkorn, J. Y., Peeler, J. S., and Mellon, J. (1989). Reg. Immunol. 2, 83-90. 60. Grabner, G., Luger, T. A., Luger, B. M., Srnolin, G., and Oh, J. 0. (1983). Invest. Ophthalmol. Visual Sci. 24,589-595. 61. Shams, N. B. K., Sigel, M. M., and Davis, R. M. (1989). Reg. Immunol. 2, 136-148. 62. Barker, C. F., and Billingharn, R. E. (1977). Adv. Immunol. 25, 1-54. 63. Greene, H. S. N., and Lund, P. K. (1944). Cancer Res. 4,352-363. 64. Raju, S., and Grogan, J. B. (1969). Transplantation 7,475-483. 65. Franklin, R. M., and Prendergast, R. A. (1970).J. Immunol. 104,463-469. 66. Kaplan, H. J., and Streilein, J. W. (1977).J. Immunol. 118, 809-814. 67. Kaplan, H. J., Streilein, J. W., and Stevens, T. R. (1975).J. Immunol. 115, 805-810. 68. Kaplan, H. J., and Streilein, J. W. (1978).J.Immunol. 120,689-693. 69. Subba Rao, D. S. V., and Grogan, J. B. (1977). Transplantation 24,377-383.
224
jERRY Y . NIEDERKORN
70. Niederkorn, J., Streilein,J. W., and Shadduck,J. A. (1980).Invest. Ophthalmol. Visual Sci. 20,355-363. 71. Streilein,J. W., Niederkorn, J. Y., and Shadduck,J. A. (1980).J. Exp. Med. 152, 1121-1125. 72. Niederkorn, J. Y., and Streilein,J. W. (1982). Transplantation 33,573-577. 73. Streilein,J. W., and Niederkorn, J. Y. (1981).J. Exp. Med. 153, 1058-1067. 74. Wetzig, R. P., Foster, C. S., and Greene, M. I. (1982).J. Immunol. 128, 1753- 1757. 75. Waldrep, J. C., and Kaplan, H. J. (1983).Invest. Ophthalmol. Visual Sci. 24, 1086- 1092. 76. Whittum, J. A., Niederkorn, J. Y., McCulley, J. P., and Streilein, J. W. (1983).C u n . Eye. Res. 2,691-697. 77. Ksander, B. R., and Hendricks, R. L. (1987).Invest. Ophthalmol. Visual Sci. 28,1986-1993. 78. Niederkorn, J. Y. (1984).Invest. Ophthalmol. Visual Sci. 25,447-454. 79. Mizuno, K., Clark, A. F., and Streilein, J. W. (1989). Invest. Ophthalmol. Visual SCZ. 30, 11 12-1 119. 80. Mizuno, K . , Clark, A. F., and Streilein, J. W. (1988). Cuw. Eye Res. 7 , 627-632. 81. Mizuno, K., Altman, N. F., Clark, A. F., and Streilein,J. W. (1989).C u n . Eye Re.$.8, 112-121. 82. Asherson, G. L., and Stone, S. H. (1965).Immunology 9, 205-217. 83. Niederkorn, J. Y., and Streilein,J. W. (1982).In, “Immunologyand Immunopathology of the Eye” (G. R. OConnor and J. W. Chandler, eds.), pp. 222-225. Masson, New York. 84. Williamson,J. S. P., and Streilein,J. W. (1988).Reg. Immunol. 1, 15-23. 85. Sy, M. S., Miller, S. D., Kowach, H. B., and Claman, H. N. (1979). J . Immunol. 119,2095-2099. 86. Nash, A. A., and Ashford, N. P. N. (1982).Immunology 45,761-767. 87. Schrier, R. D., Pizer, L. I., and Moorhead, J. W. (1983).Infect. Immun. 40, 5 14-522. 88. Whittum, J. A., Niederkorn, J. Y., McCulley, J. P., Streilein,J. W. (1984). J . Virol. 51,556-558. 89. Niederkorn, J. Y., and Streilein,J. W. (1982).J. Immunol. 128,2470-2474. 90. Niederkorn, J. Y., and Streilein,J. W. (1983).Transplant. Proc. 15,401-403. 91. Niederkorn, J. Y., and Streilein,J. W. (1983).J. Immunol. 131,2587-2594. 92. Streilein,J. W., Bradley, D., and Williamson,J. S. P. (1989).Invest. Ophthalmol. Visual Sci., Suppl. 30,440. 93. Granstein, R., Staszewski, R., Zeira, E., Nazareno, R., Latina, M., and Albert, D. M. (1989).Invest. Ophthalmol. Visual Sci. Suppl. 30,441. 94. Granstein, R. D. (1985).J. Invest. Dermutol. 84,206-209. 95. Granstein, R. D., Lowy, A., and Greene, M. I. (1984).J. Immunol. 132, 563-565. 96. Lynch, M. G., Peeler, J. S., Brown, R. H., and Niederkorn, J. Y. (1987). Ophthalmology (Rochester, Minn.) 94,85 1-857. 97. Lynch, M. G., Luckenbach, M., Fischer, B. J., Davis, P., and Niederkorn, J. Y. (1989). Invest. Ophthalmol. Visual Sci., Suppl. 30,468. 98. Williamson,J. S. P., and Streilein,J. W. (1989).Transplantation 47,5 19-524. 99. Streilein,J. W. (1987).FASEB J. 1, 199-208. 100. Streilein, J. W., and Wegmann, T. G. (1987).Immunol. Today 8,362-366. 101. Niederkorn, J. Y., and Streilein,J. W. (1983).J. Immunol. 131,2670-2674.
IMMUNOLOGY OF THE EYE
225
102. Niederkorn, J. Y.,Shadduck, J. A., and Streilein,J. W. (1981).Immunogenetics 13, 227-236. 103. Metcalf, J. F.,Hamilton, D. S., and Reichert, R. W. (1979).Infect. Immun. 26, 1164-1171. 104. Russell, R. G., Nasisse, M. P., Larsen, H. S., and Rouse, B. T. (1984).Invest. Ophthalmol. Visual Sci. 25,938-944. 105. Ksander, B. R., and Hendricks, R. L. (1987).Invest. Ophthalmol. Visual Sci. 28,1986-1993. 106. Hendricks, R. L., Epstein, R. J., and Tumpey, T. (1989).Invest. Ophthalmol. Visual Sci. 30, 105-1 15. 107. Whittum, J. A., McCulley, J. P., Niederkorn, J. Y., and Streilein, J. W. (1984).Invest. Ophthalmol. Visual Sci. 25, 1065-1073. 108. Kielty, D., Cousins, S. W., and Atherton, S. S. (1987).Invest. Ophthalmol. Visual Sci. 28, 1994-1999. 109. Atherton, S. S.,and Streilein, J. W. (1987).Cum. Eye Res. 6, 133-139. 110. Streilein, J. W.,Atherton, S. S., and Vann, V. (1987).Curr. Eye Res. 6, 127-13 1. 111. Streilein,J. W., and Niederkorn, J. Y. (1985).J . Immunol. 134, 1381-1387. 112. Ferguson, T.A., Hayashi, J. D., and Kaplan, H. J. (1989).J.Immunol. 143, 821-826. 113. Burnet, F. M. (1970).Prog. Exp. Tumor Res. 13, 1-27. 114. Hogan, J.J.,and Zimmerman, L. E. (1962).“Ophthalmic Pathology.” Saunders, Philadelphia.
115. Niederkorn, J. Y., and Meunier, P. C. (1985).Invest. Ophthalmol. Visual Sci. 26,877-884. 116. Knisely, T.L., Luckenbach, M. W., Fischer, B. J., and Niederkorn, J. Y. (1987).J . Immunol. 138,4515-4523. 117. Niederkorn, J. Y., and Knisely, T. L. (1988).Curr. Eye Res. 7,515-526. 118. Khodadoust, A. A., Roozitalab, H. M., Smith, R. E., and Green, W. R. (1977).Sum. Ophthalmol. 21,467-478. 119. Knisely, T.L., and Niederkorn, J. Y. (1990).Cancer Immunol. Immother., in press.
120. Elschnig, A. (1910).Albrecht von Graefes Arch. Klin. Ex#. Ophthalmol. 76, 509-546. 121. Faure, J. P. (1980).Curr. Top. EyeRes. 2,215-302. 122. Gery, I., Mochizuki, M., and Nussenblatt, R. B. (1986).In “Progress in Retinal Research” (N. Osborne and J. Chader, eds.), pp. 75-109. Pergamon, New York.
123. Wacker, W.B.,and Lipton, M. M. (1965).Nature (London) 206,253-254. 124. Fong, S.-L., Liou, G. I., Landers, R.A., Alvarez, R. A., and Bridges, C. D. B. (1984).J. Biol. Chem. 259,6534-6542. 125. Liou, G. I., Bridges, C. D. B., Fong, S.-L., Alvarez, R. A., and GonzalezFernandez, G. (1982).Vision Res. 22, 1457- 1468. 126. Salinas-Carmona, M. C., Nussenblatt, R. B., and Gery, I. (1982).Eur. J . Immunol. 12,480-484. 127. Nussenblatt, R. B., Rodrigues, M. M., Wacker, W. B., Cevario, S. J., SalinasCarmona, M. C., and Gery, I. (1981).J.Clin. Invest. 67, 1228-1231. 128. Rozenszajn, L. A., Muellenberg-Coulombre, C., Gery, I., El-Saied, M., Kuwabara, T., Mochizuki, M., Lando, Z., and Nussenblatt, R. B. (1986). Immunology 57,559-565. 129. Caspi, R. K.,Roberge, F. G., McAllister, C. G., El-Saied, M., Kuwabara, T.,
226
JERRY Y. NIEDERKORN
Gery, I., Hanna, E., and Nussenblatt, R. B. (1986). J. Immunol. 136, 928-
933. 130. Mochizuki, M., Kuwabara, T., McAllister, C., Nussenblatt, R. B., and Gery, I. (1985). Invest. Ophthalmol. Visual Sci. 26, 1-14. 131. Aronson, S. B., and McMaster, R. B. (1971). Arch. Ophthalmol. (Chicago) 86, 557-563. 132. Rao, N.A.,and Wong,V. (1981). Trans. 0phthalmol.Soc. U.K. 101,357-360, 133. Rao, N. A., Wacker, W. B., and Marrak, G. E., Jr. (1979). Arch. Ophthalmol. (Chicago) 97, 1954-1958. 134. Chan, C. C., Nussenblatt, R. B., Fujikawa, L. S., Palestine, A. G., Stevens, G., Jr., Parver, L. M., Luckenbach, M. W., and Kuwabara, T. (1986). Ofihthalmology (Rochester, Minn.) 93,690-695. 135. Takobiec, F. A., Marboe, C. C., Knowles, D. M.. 11. Iwamoto. T.. Harrison. k.,Chang, S., and Coleman., J. (1983). Ophthalmology (Philadelphia) 90; 76-95. 136. Nossal, G. J. V. (1989). Science 245, 147-153. 137. McAllister, C. G., Wiggert, B., Chader, G. J., Kuwabara, T., and Gery, I. (1987).j. Immunol. 138, 1416-1420. 138. Caspi, R. R., Roberge, F. G., Chan, C. C., Wiggert, B., Chader, G. J., Rozenszajn, L. A., Lando, Z., and Nussenblatt, R. B. (1988).J.Immunol. 140, 1490- 1495, 139. Hirose, S., Tanaka, T., Nussenblatt, R. B., Palestine, A. G., Wiggert, B., Redmond, T. M., Chader, G. J., and Gery, I. (1988). C u n . Eye Res. 7, 393-402. 140. Gill, H. K., and Liew, F. Y. (1978). Eur. J. Immunol. 8, 172-176. 141. Caspi, R. R., Roberge, F. G., and Nussenblatt, R. B. (1987). Science 237, 1029- 1032. 142. Sigelman, J., and Ozanics, V. (1983). In “Biomedical Foundations of Ophthalmology” (T. D. Duane and E. A. Jaeger, eds.), pp. 45-49. Harper &
Row, New York.
143. Chan, C. C., Mochizuki, M., Nussenblatt, R. B., Palestine, A., McAllister, C., Gery, I., and Ben Ezra, D. (1985). Clin. Immunol. Immunopathol. 35,103-1 10. 144. Rao, N. A., Brown, C. J., and Marak, G. E. (1986). OphthalmicRes. 18,15-20. 145. Caspi, R. R., Kuwabara, T., and Nussenblatt, R. B. (1988). J. Immunol. 140, 2579-2584. 146. Uhlenhuth, P. (1903). In “Gestschrift zum Sechzigsten Gerburtstage von Robert Koch,” p. 49. Fischer, Jena. 147. Verhoeff, F., and Lemoine, A. N. (1922). Am. J. Ophthalmol. 5, 700-702. 148. Angunawela, I. I. (1987). Immunology 61,363-368. 149. Hackett, E., and Thompson, A. (1964). Lancet 2,663-666. 150. Sandberg, H . O., and Closs, 0. (1979). Scand. J . Immunol. 10,549-554. 151. Luntz, M. H. (1968). Exp. Eye Res. 7,561-569. 152. Goldschmidt, L., Goldbaum, M., Walker, S. M., and Weigle, W. 0. (1982). J . Immunol. 129,1652-1662. 153. Niederkorn, J. Y., and Peeler, J. S. (1988). Immunol. Res. 7,247-255.
This article was accepted for publication on 4 October 1989.
ADVANCES I N IMMUNOLOGY. VOL. 48
Molecular Events Mediating T Cell Activation AMNON ALTMN, K. MARK COGGESHALL, and T O M S MUSTELIN Depotintent of Immunology, Scripps Clinic and Reseatzh Foundation, l a Jolla, California 92037
1. Introduction
Triggering of resting thymus-derived (T)lymphocytes under physiological conditions by antigenic peptides associated with major histocompatibility complex (MHC) molecules gives rise to activated and differentiated T cells displaying genetically determined effector functions, such as secretion of immunoregulatory cytokines by T helper (Th) cells and specific lysis of target cells by cytotoxic T lymphocytes (CTLs). In addition to antigen-MHC complexes, other agonists, such as antibodies to various cell surface determinants or mitogenic lectins, can also stimulate T lymphocytes when they bind to their corresponding receptors. This binding event initiates an ordered cascade of biochemical changes and generates signals that are transmitted sequentially from the cell surface to the nucleus, leading to gene transcription. Replication of these activated T cells requires an additional signal mediated by the binding of growthpromoting lymphokines, such as interleukin-2 (IL-2), to their respective receptors. The ability to maintain functional antigen-specific T cell clones in long-term culture and the determination of the primary structure of the T cell antigen-receptor complex and other surface activation molecules have been some of the major advances in T cell immunobiology in recent years. This progress has allowed us to analyze the cascade of biochemical events that constitute the transmembrane signaling machinery. Furthermore, the existence of multiple ligands and the expression of their respective receptors on T cells, coupled with the ability to obtain such cells in a homogeneous form and in large quantities, make T cells an attractive model to cell biologists interested in studying signal transduction systems in general. Hence, numerous studies in the past few years have investigated various signaling events initiated by the binding of activating ligands to T cells. As a result of these studies, certain key features of the early biochemical events during T cell activation have emerged. According to a current dogma, triggering of the T cell antigen-receptor complex and some 227 Copyright 8 1990 by Academic Press, lnc. All rights of reproduction in any form reserved.
228
AMNON ALTMAN ET AL.
other surface molecules is coupled to the phosphodiesterase (phospholipase C)-mediated hydrolysis of membrane phosphoinositides (PIS),in particular, phosphatidylinositol4,5-bisphosphate(PIPp). PIP:! hydrolysis generates two products, inositol 1,4,5-trisphosphate (IPS) and diacylglycerol (DAG), which act in concert as second messengers to increase free intracellular Ca2+concentration ( [Ca2+]i)and activate protein kinase C (PKC), respectively, thereby stimulating subsequent events leading to cellular activation and proliferation. Although this aspect of T cell activation is generally accepted by most researchers in this area, nearly all other biochemical events are poorly understood. These include the mechanisms that couple receptor occupancy to PIP2 hydrolysis, the potential role of other signaling pathways, and the critical cytoplasmic and nuclear events that follow second messenger production. Here, we review the state of the art in the area of T lymphocyte signal transduction. This chapter is not intended to present an exhaustive review of the relevant voluminous literature, but rather a critical overview and synthesis of the important findings in this field. Other related reviews have recently been published (1 -3). II. T Cell Surface Activation Molecules
Activation of resting T cells can be initiated by a diverse group of cell surface receptors identified in recent years, originally, by appropriate monoclonal antibodies (mAbs) and, subsequently, by gene cloning and sequencing. The conclusion that such receptors play a role in T cell activation is based on the agonist or antagonist properties of the relevant mAb. The physiological ligands for some, but not all, of these receptors have been identified. Activation-related receptors on T cells can be generally classified into two categories. The first includes those that can, by themselves, generate primary activation signals leading to T cell activation in the absence of additional stimuli. The second group includes those receptors that do not initiate T cell activation independelitly, but rather provide accessory signals to the antigen-receptor complex (or other receptors included in the first category). In addition, other receptors that play an important role in the proliferation of activated T cells are not present on resting T cells, but are induced de nova as part of the activation process. The prototype of these is the a (p55, Tac) subunit of the IL-2 receptor (IL-2R). A. T CELLANTIGENRECEPTOR-cD3 COMPLEX The key recognition and activation element during physiological T cell responses to antigens is a complex receptor. It consists of, first, clonally
MOLECULAR EVENTS MEDIATING
‘I‘ CELL A C T I V A T I O N
229
distributed, highly polymorphic subunits, namely, the T cell antigen receptor (TCR), whose function is to recognize and specifically bind nominal antigenic peptides presented by MHC molecules, and, second, a complex termed “CD3” (formerly T3). CD3 is composed of at least five distinct invariant polypeptides, some or all of which are believed to function as the signal-transducing unit of the receptor complex. T h e TCR and CD3 associate noncovalently to form the complete and functional receptor. 1 . Structure of the TCR-CD3 Complex
Th e TCR was originally identified by antiidiotypic antibodies raised against T cell tumors (4) or antigen-specific T cell clones (5-7), and later by molecular cloning techniques (8, 9; reviewed in Refs. 10-13). It consists of four highly polymorphic (a,p, y, 6) proteins that form two heterodimers (a+ and 7-6) similar to immunoglobulins in their overall primary structure, gene organization, and rearrangement patterns. T h e a+ heterodimer clearly mediates the specific recognition of antigenic peptides (in the context of MHC molecules), as indicated by the ability of TCR a and p chain genes to transfer antigen specificity upon transfection into T cell lines (14, 15); the exact function of the 7-6 T cell receptor is largely unknown. Each of the four TCR polypeptides contains a single membrane-spanning region, but, notably, very short (up to 12 amino acids) cytoplasmic domains (8, 9, 16-20), suggesting that the polymorphic TCR does not play a direct role in signal transduction. The invariant CD3 complex was initially identified by an mAb, OKT3 (21), as a specific T cell marker on human T cells. Subsequent immunochemical and biochemical studies identified and characterized the polypeptides that comprise the CD3 complex in murine (22-24) and human (25-32) T cells. These analyses have been aided to a large extent by the ability to coimmunoprecipitate CD3 proteins with the TCR, using antiTCR antibodies and appropriate detergents, or to cross-link these two complexes. Individual components of the CD3 complex have been distinguished by their size, isoelectric point, and peptide maps, and by antibodies specific for the CD3 subunits and by cDNA cloning (33-40). A summary of the key features of murine and human CD3 polypeptides is presented in Table 1. To date, five distinct polypeptides, termed y , 6, E, 6, and 7,have been identified in noncovalent association with the clonotypic a-p TCR (Table I). These proteins range in size between 16 and 28 kDa. The y and 6 subunits contain one to three extracellular N-glycosylation sites, while E, 5, and 7 are nonglycosylated (23, 41). About 90% of 5 is found as a homodimer, and the rest forms a heterodimer with the r) subunit (23,32,
TABLE I COMPONENTS OF THE CD3 COMPLEX Primary structureb Chain Murine Human Murine Human Murine Human Murine Human Murine
Y 6 E
5 r)
Size (kDa)" 25-28
22-261 19-23
16 22
GIycosylation'
Phosphorylation
Ex.
TM
Cyt.
89
27
79
27
Yes
87 104
44 46 44
26
55
No
PKG'
9
21
113
No
TPK ( ~ p 5 6 ' ' ~ ? ) ~
?
Yes
No
(1) (2) i3j (2)
PKC (Ser- 126)d CaPK (Ser- 123, Ser- 126)'
-
Based on sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of immunoprecipitated proteins. Number of amino acid residues in extracellular (Ex.), transmembrane (TM), and cytoplasmic (Cyt.) domains deduced from cDNA cloning. In addition, cDNA sequences predict 2 1- to 22-amino-acid signal peptides. The number of N-glycosylation sites is shown in parentheses. Phosphorylation is induced by antigen, mitogen, anti-TCR-CD3 antibody, or PMA. Mediated by a putative Ca*+/calmodulin-dependentprotein kinase in response to calcium ionophore stimulation (420). f Documented in the mouse only in response to antigen, anti-TCR-CD3, anti-Thy-1, or PMA (358). gFound as a 5-5 homodimer (32 kDa) or a 5-7 heterodimer (38 kDa) in unreduced samples. The apparent size of the phosphorylated chain is 21 kDa. Murine and human cytoplasmic domains contain six or seven tyrosine residues, respectively. Tyrosine phosphorylation is induced by antigen, antireceptor antibodies, and mitogens, but not by PMA. a
MOL.E<:Ul.AK EVENTS MEDIATING
'I' CELL ACTIVATION
23 1
39-42). The q chain, which has been identified most recently and can be clearly distinguished from the other CD3 components by partial peptide mapping (41), is not as well characterized as the others. The homo- o r heterodimeric covalent linkage of 5 is most probably mediated by a cysteine residue in its transmembrane domain (39, 40). A CD3 subunit, termed p2 1, which becomes phosphorylated on tyrosine residues in response to antigen stimulation (42), was later found to represent the phosphorylated form of 5 (43). When T cells are activated by antigen, antireceptor antibodies, or mitogenic lectins, the y and E chains become phosphorylated on serine resides and 5, on tyrosine residues. Receptor phosphorylation is discussed in more detail in Section V. Molecular cloning of cDNAs encoding the y , 6, E, and 5 components of the receptor complex (33-40) revealed that these are typical membrane proteins with a 2 1- to 22-amino-acid signal peptide and a hydrophobic transmembrane domain consisting of 21-27 amino acids (Table I). Interestingly, the transmembrane domains contain, in all cases, a centrally located acidic residue (glutamic acid in y and aspartic acid in 6, E, and 5). This contrasts with the basic residues found in the corresponding domains of all clonotypic TCR subunits (8-13). These charged residues may form salt bridges between the TCR and CD3 and may stabilize the noncovalent association of the various components within the hydrophobic milieu of the T cell membrane. The nucleotide and amino acid sequences of the y, 6, and E chains are homologous with each other (35-38) and with several members of the immunoglobulin supergene family (38). T h e corresponding genes are all located within a 75-kb region of chromosome 11 in humans (44). Thus, these three CD3 subunits appear to constitute members of a receptor complex that arose by gene duplication. Th e 5 subunit, on the other hand, shares no homology with y, 6, or E and has several additional features that clearly distinguish it from the former and may have considerable implications for TCR-CD3 complexmediated signal transduction. First, unlike the y , 6, or E polypeptides, whose extracellular domains are relatively large (79- 104 amino acids) and considerably longer than the corresponding cytoplasmic domains (44-79 amino acids), 5 has a small extracellular domain of only nine amino acids and a much larger cytoplasmic domain of 1 12-1 13 amino acids. Second, the cytoplasmic region of 5 contains six (murine) or seven (human) tyrosine residues that are potential substrates for tyrosine protein kinase(s) (TPK), as well as a consensus ATP-binding sequence (39, 40). Based on these considerations, Weissman et al. (40) suggested that 5 (and perhaps q) should be considered as belonging to a distinct group
232
AMNON ALTMAN E T AL.
unrelated to the other three invariant components of CD3. These investigators (41) also proposed that two structural classes of receptor complexes are displayed on any T cell. One class, the majority, contains only (-6 homodimers, while the other class (10-20% of the total) contains the (-7 heterodimer. This model raises the intriguing possibility that the two receptor classes may be coupled to distinct signaling pathways. This concept is supported by the analysis of cellular variants deficient in the expression of ( or r) (45, 46; see Sections III,C,4 and X). Using antibodies specific for all of the subunits of the murine receptor complex in conjunction with pulse-chase experiments, Minami et al. (47) have shown that most chains are synthesized in large excess over the amount incorporated into the complete receptor and that the excess chains, either in their free form or as partially assembled complexes, are rapidly degraded in a lysosomal compartment. The ( chain appeared to be the limiting factor for assembly of the mature receptor complex, perhaps emphasizing again its critical role in the synthesis and function of the complete receptor. 2. TCR-CD3 Complex-Mediated T Cell Activation The physiologically relevant TCR-CD3 ligand is an antigenic peptide generated by antigen-presenting cells (APCs) in a proteolytic processing step and presented in the context of MHC molecules on such APCs (reviewed in Ref. 48). Ligand binding to the TCR-CD3 complex, together with some accessory molecules, initiates signal transduction and leads to full T cell activation, namely, expression of activation antigens such as IL-BR, synthesis and secretion of lymphokiries (by Th cells), cytolytic activity (by CTLs), and clonal expansion and proliferation. An obviously important question is which of the TCR-CD3 complex subunits plays the critical and primary role in signal transduction. Assessing the relative contribution of various subunits to receptor signaling and T cell activation is difficult, however, since the stable expression of a complete and functional receptor requires the assembly and simultaneous coexpression of all subunits, namely, a,p, y , 6, E, ( (and r)?). Thus, in T cell mutants selected for the loss of CD3 or TCR, a concomitant loss of both units was found (49); conversely, transfer of a TCR-j3 cDNA into a mutant human leukemic T cell line that did not express a functional 3./ chain mRNA and surface TCR-CD3 complex led to the expression of a structurally and functionally active receptor complex (50). More recent studies with a (-negative mutant of an antigen-specific murine T cell hybridoma (46) showed that some receptor complex assembly and cell surface expression can occur in the absence of ( and, moreover, that this [-lacking receptor complex can respond to some, but not
MOLECULAR EVENTS MEDIATING
'I'
CELL ACTIVATION
233
all, stimuli, as evidenced by PIS hydrolysis and IL-2 production. Whatever the role of individual CD3 polypeptides in signal transduction is, indirect evidence seems to indicate that the clonotypic TCR is not the primary signal transducer and that the CD3 complex plays the major role in the signaling apparatus (5 1). This conclusion is based on the findings that in activation-defective T cell mutants that express normal TCR-CD3 levels, some anti-CD3, but not anti-TCR, antibodies can stimulate PIS hydrolysis and second messenger production (52), and that heterokaryons formed between such mutant cells and T cell lines lacking TCR a or p chains respond to receptor triggering with second messenger production (53). Direct analysis of the role of antigen-MHC binding to the TCR-CD3 complex in T cell activation is hampered by the complexity of studying cellular interactions between T cells and APCs and the poorly understood structure of the immunogenic peptide-MHC complex. Thus, welldefined soluble T cell agonists, namely, anti-TCR-CD3 mAb, as well as lectins such as phytohemagglutinin (PHA) or concanavalin A (Con A) have been widely used in studying activation requirements and signaling events in T cells; however, such agonists may not fully mimic the events triggered by antigen binding. Indeed, instances have been recorded in which antigen and antLCD3 antibodies differ in their abilities to induce early activation events (see, e.g., Ref. 46). T h e lectins PHA and Con A act as polyclonal T cell activators and most likely bind to multiple surface glycoproteins. Although it is not known which target glycoprotein(s) is relevant in T cell activation, it is quite possible that the documented binding of PHA and Con A to isolated TCR or CD3 peptides (54), respectively, mediates their effects. Indeed, TCRCDPmutants of the human leukemic T cell line, Jurkat, failed to respond to PHA or Con A (as well as anti-TCR-CD3 antibodies) plus phorbol ester (49, 55, 56). Conversely, reconstitution of an active TCR-CD3 complex by gene transfection restored PHA responsiveness in these mutants (50). In addition, PHA-mediated activation has been reported to depend on the expression of the CD2 surface antigen (57), another T cell receptor that plays an important role in T cell activation and interactions with other cell types (see below). A wide range of anti-TCR o r -CD3 antibodies can activate T cells in a polyclonal manner (58-61) or, in the case of antigen-specific T cell clones, substitute for antigen (7, 62, 63). Similar antibodies, when presented in soluble form, can block the antigen-specific activation of T cell clones (5, 6, 63). Activation of anti-TCR-CD3 antibodies o r lectins requires cross-linking, perhaps reflecting the need for multivalent interactions between ligand and the TCR-CD3 complex to achieve optimal
234
AMNON ALTMAN E.r AL.
activation. In this respect these cross-linked T cell ligands may mimic the physiological T cell ligand (i.e., the peptide-MHC complex), most likely presented as a multimeric complex on the surface of APCs. The crosslinking requirement can be fulfilled by including in the culture accessory cells (e.g., monocytes) that bind the activating antibodies to their Fc receptors (FcRs) (58, 64-67) or by immobilizing the antibodies on Sepharose beads o r on the surface of the culture dishes (61, 63, 68). In addition to FcR-mediated cross-linking, monocytes may mediate another important function in T cell activation, that is, the secretion of soluble mediators, notably interleukin-1 (IL-1). However, the ability of immobilized anti-TCR-CD3 antibodies to activate monocyte-depleted highly purified T cells (60, 69) or of fibroblasts expressing the FcR for immunoglobulin G (IgG) to reconstitute the anti-CD3-mediated activation of such T cell preparations (70) argues that IL-1 (or other monocyte products) is not always critical, at least for some aspects of cellular activation. As is always the case with purified T cell populations, it is difficult to rule out the possibility that minor contamination by a few residual accessory cells provides the necessary mediators in these cases. Both functions normally provided by accessory cells can be replaced by phorbol myristate acetate (PMA) or other tumor promoters (1). Since the pleiotropic cellular effects of tumor promoters are generally attributed to their sustained PKC activation, this event appears to be crucial in T cell activation (see Section V,A,l). Conflicting reports exist regarding the requirements for T cell activation via the TCR-CD3 complex. However, in many instances such conflicting results are more apparent than real, since the exact activation requirements depend, among other things, on the T cell subset studied (e.g., CD4' versus CD8+), the activation status of the cell (e.g., resting versus activated), or the particular activation event measured. CD4' and CD8' T cells, restricted by MHC class I1 or class I molecules, respectively, require qualitatively different signals for the induction of functional IL-2R (61,71) or for proliferation (72-74). The activation requirements for previously stimulated T cells appear to be less stringent than those for naive resting T cells. While the latter require both functions provided by the accessory cells-that is, FcR-mediated cross-linking and secretion of soluble mediators (IL-l?)-the former require only the cross-linking function for activation. Thus, highly purified resting T cells can be activated by immobilized anti-CD3 antibodies only in the presence of an additional signal provided by IL-1(75,76) or by antibodies to other T cell surface antigens, such as CD2, CD5, or CD28 (1). In contrast, previously activated T cell tumors (76) or antigen-specific clones (63) require only immobilized anti-CD3 antibodies. In addition,
MOLECULAR EVENTS MEDIATING
T CX1.L
ACTIVATION
235
the early increase in [Ca2+]iby mobilization from intracellular stores (77) and the induction of IL-2R (60, 78) appear to require less exacting conditions than the induction of later activation events, such as IL-2 secretion and proliferation. While PKC activation or an increase in [Ca2+]ialone may be sufficient for IL-2R induction, both signals are required for IL-2 production and proliferation (73, 79, 80).
B. CD2 CD2 (previously known as T1 1) was originally identified as a T cellspecific sheep red blood cell receptor on human T cells (81). It was later defined by a specific mAb as a 50-kDa nonpolymorphic protein (82, 83) first expressed during early T cell development in the thymus (83-85), even prior to the expression of CD3. Molecular cloning of human (86, 87), rat (88), and murine (89) CD2 has revealed the structure of a typical transmembrane protein having a 185- to 186-amino-acid extracellular domain, with several N-glycosylation sites, a transmembrane segment of 25 residues, and a cytoplasmic domain of 116-126 amino acids. CD2 shows limited regions of homology with some members of the immunoglobulin supergene family (86-88). The physiological ligand for CD2 is lymphocyte function-associated antigen-3 (LFA-3) (90-93), a widely distributed cell surface glycoprotein of 55,000-70,000 molecular weight present on endothelial, epithelial, and connective tissue cells and on most blood cells, including erythrocytes (91, 94; reviewed in Ref. 95). CD2LFA-3 binding is thought to play an important role in the interactions between immature thymocytes and thymic epithelial cells (96)or between CTLs and their target cells (94). Interest in the structure and function of CD2 increased when it was found that certain combinations of mAbs to CD2 can induce IL-2dependent T cell proliferation (97,98),help for antibody responses (97), and antigen-specific or -nonspecific cytolytic activity by CTLs or natural killer cells, respectively (99).Studies of this “alternative pathway” of T cell activation (97) have defined, through the use of various mAbs, at least six distinct epitopes, some of which are induced upon activation of resting T cells. Unlike anti-TCR-CD3 antibodies, nearly all of which can singly activate T cells, only some antLCD2 mAbs can stimulate T cells, and even then, combination of two antibodies with distinct epitope specificities is required. When used singly, anti-CD2 mAbs usually block T cell activation and proliferation (94,95, 100). AntLCD2 mAb combinations can also activate thymocytes (85), including those of the immature TCR-CD3 subset, to express IL-2R (but not to proliferate). Depending on the particular combination of antibodies used, CD2-initiated T cell activation may (98) or may not (97) be accessory
236
AMNON ALTMAN E T A L .
cell dependent. T cell triggering via the CD2 pathway induces the same events as those stimulated by TCR-CD3 ligands, such as IL-2 secretion (97), increased [Ca2+]i(101), production of the phosphatidylinositol (PI) pathway-related secondary messengers (102), and opening of voltageinsensitive plasma membrane Ca2+channels (103). More recently, it has been shown that the purified natural CD3 ligand, LFA-3, can synergize with certain anti-CD2 mAbs to induce T cell activation (92, 104-106). The relationship between the “classical”TCR-CD3 complex and alternative CD2 pathways of T cell activation has been analyzed in a nutnber of studies. These two receptor complexes are not linked physically (107). Although some (94, loo), but not all (97, 108), studies documented the ability of antLCD2 mAbs to inhibit T cell activation via the TCR-CD3 pathway, these functional tests are difficult to interpret. The antLCD2 antibodies used in such blocking experiments could interfere with cellular contact between responding T cells and accessory cells, which is normally required for activation by anti-TCR-CD3 antibodies (58, 64-67). More evidence from the analysis of a CD2-CD3’ IL-2-dependent T cell clone (109) or a similar Jurkat mutant (110), which could be induced to proliferate, display cytolytic activity, and produce IL-2 following anti-TCR-CD3 antibody stimulation, clearly indicates that CD2 is not necessary for T cell activation via the TCR-CD3 complex, at least in cycling T cells. Anti-CD2 and anti-TCR-CD3 antibodies or the corresponding physiological ligands (LFA-3and antigen) can clearly synergize in T cell activation (74, 111, 1 12). This cooperation is not merely the result of a more avid adhesion between T cells and accessory cells due to CD2-LFA-3 interactions, but rather requires an active signaling role of CD2, as implied by the fact that deletion of the cytoplasmic CD2 domain, which did not affect LFA-3 binding, abolished this cooperation (112). In this respect CD2 triggering was found (113) to induce phosphorylation of the CD3 y chain (most likely a PKC-mediated serine phosphorylation), similar to that normally induced by TCR-CD3 ligands (see Section V,A, 1). However, tyrosine phosphorylation of the CD3 5 chain is not induced by the CD2 pathway (114). The ability of anti-CD2 mAb pairs to activate CD3- thymocytes (85), natural killer cells (99, 115, 116),or CD3-Jurkat mutants (117) indicates that activation via the CD2 pathway does not always require a functional TCR-CD3 complex. However, compelling evidence has been accumulated to indicate that TCR-CD3 expression is usually necessary for CD2mediated activation (97, 108, 113, 118-120). In this regard transfection of a TCR p chain-negative (and, hence, CDS) Jurkat mutant with a p chain cDNA restored TCR-CD3 expression as well as the activation potential via the CD2 pathway (1 19, 120).
MOLECULAR EVENTS MEDIATING
‘I’ CELL ACTIVATION
237
Collectively, the data indicate that, although stabilization of binding between T cells and other cell types (target cells for CTLs and natural killer cells, APCs for T h cells, and thymic epithelial cells for differentiating thymocytes) by CD2-LFA-3 interactions is an important element in CD2 function, this surface antigen plays an additional, more active role in T cell activation via a functional interaction with the TCR-CD3 complex. While CD2 expression is not essential for TCR-CD3-mediated triggering, the converse is not true. In most T cells the CD2 pathway is nonfunctional in the absence of TCR-CD3 expression. It is possible that during physiological T cell triggering (i.e., via the TCR-CD3 complex) a transient physical interaction between the two receptor systems occurs, thereby coupling CD2 to the TCR-CD3 signaling machinery. In some T cells, such as immature thymocytes or activated natural killer cells, the CD2 pathway does not require a functional TCR-CD3 complex. Perhaps in these situations the affinity of the CD2-LFA-3 interaction is higher than usual, resulting in an above-threshold activation signal that can lead to TCR-CD3-independent activation. Alternatively, in such cell types an undefined signal that is functionally analogous to TCR-CD3 triggering may synergize with the CD2 pathway. C. GLYCOSYL-PHOSPHATIDYLINOSITOL-ANCHORED ACTIVATION ANTIGENS Most membrane proteins are attached to the lipid bilayer via hydrophobic interaction between one or more protein segments functioning as transmembrane domains and the lipid bilayer. However, an increasing number of proteins have been found to utilize a more complex mode of membrane attachment, that is, by covalent linkage with a glycosylphosphatidylinositol (GPI) molecule located in the lipid bilayer (reviewed in Refs. 121 and 122). In this molecule the inositol ring of PI is glycosidically linked via glucosamine to a glycan structure. This glycan contains a terminal ethanolamine residue which, in turn, forms an amide bond with the carboxy terminus of the attached protein. GPI-linked proteins usually undergo a posttranslational modification, during which the conventional transmembrane (and cytoplasmic) domains are removed to facilitate the covalent attachment to GPI. Indeed, comparison of cDNApredicted and mature protein sequences of several GPI-anchored proteins indicated that a carboxy-terminal segment of 17-3 1 residues is removed, exposing the carboxyl group to which the GPI lipid will be attached. An increasing number of lymphocyte surface antigens have been found to be attached to the membrane by a GPI anchor (122),and the fact that most of these can function as activation antigens in T cells is of considerable interest. This function has been revealed by the ability of
238
AMNON ALTMAN E T AL.
mAbs to GPI-anchored T cell antigens to stimulate, either alone or (most frequently) in combination with other signals, T cell proliferation, and by the inhibitory effect of PI-specific phospholipase C (PI-PLC) on T cell proliferation (123, 124). The activating properties of GPI-anchwed proteins is of interest for two reasons. First, it is difficult to see how a protein lacking a cytoplasmic (and a transmembrane) peptide domain can transduce activation signals. Second, removal of GPI-anchored proteins by the hydrolytic action of PI-PLC or a GPI-specific phospholipase D would generate the second messengers DAG and phosphatidic acid, respectively. The following is a description of the structure and function of several GPI-anchored surface proteins that participate in T cell activation. 1. Thy-I
Thy- 1 was originally identified as a murine alloantigen expressed in two allelic forms on rodent thymocytes, lymphocytes, nervous, and some other tissues (125), but not on human T cells. It is a 25- to 30-kDa glycoprotein encoded by genes related to the immunoglobulin supergene family (126, 127).Isolation and amino acid sequencing of Thy- 1 revealed a protein composed of 111-1 12 amino acids and lacking a typical transmembrane domain (127). In contrast, gene cloning studies (126) revealed an additional segment that includes a stretch of 20 hydrophobic amino acids. It was subsequently found that Thy-1 undergoes a processing step in which newly synthesized Thy-1 molecules are cleaved by removing a carboxy-terminal segment to generate the mature protein, which then becomes covalently linked to a GPI structure (128, 129). This anchors Thy-1 to the membrane, and Thy-1 can be released by PI-PLC treatment (130). The ability of Thy-1 to act as an activation element was revealed in earlier studies that documented the mitogenic action of Thy- 1-specific rabbit anti-mouse brain antisera on T cells (131, 132).Subsequent studies with a collection of Thy- 1-specific mAbs directed at distant epitopes demonstrated that some, but not all, mAbs can induce the full spectrum of T cell activation, that is, IL-PR expression and lymphokine secretion and proliferation in mature T cells, or IL-2R expression only in thymocytes (133-136). While some mAbs were capable of activating T cells alone (133-135), a striking synergy was observed between nonmitogenic doses of two mAbs specific for different epitopes (135, 136) or by crosslinking a single mAb with a secondary antibody (136). Under these conditions T cell activation is accessory cell independent (135, 136). Induction of T cell proliferation required, in some studies (136), the addition of PMA. Differences in the activation requirements of anti-Thy-
MOLECULAR EVENTS MEDIAI'IN(;
T CELL ACTIVATION
239
1 mAbs most likely reflect the types of T cells used and the measured responses. As was noted for activation by anti-TCR-CD3 antibodies, the induction of early activation events (e.g., Ca2+ mobilization and IL-2R expression) o r the activation of T cell lines and hybridomas has less stringent requirements than the stimulation of late events (IL-2 production and proliferation) or the activation of resting T cells, respectively. Several studies addressed the question of whether Thy-1 can function as an independent signal-transducing molecule in the absence of TCRCD3 expression. From an earlier study that revealed the ability of antiThy-1 mAbs to increase [Ca2+]iin Thy-l-transfected murine B lymphoma cell lines (137), it was concluded that Thy-l-mediated activation is independent of TCR-CD3 o r other T cell-specific molecules. However, subsequent studies (138, 139) have shown that in the absence of a functional TCR-CD3 complex on T cells, at least some aspects of the Thy-1 activation pathway are nonfunctional. A TCR a chain deletion variant of a CTL clone, which expressed normal levels of Thy-1, could not be activated by anti-Thy- 1 mAbs to increase [Ca'+]i, secrete lymphokine, and display its cytolytic activity (138). Analysis of a TCR-CD3-negative variant of the human leukemic T cell line, Jurkat, that had been transfected with a murine Thy-1.2 gene, showed, however, that anti-Thy-1 mAbs can induce an increase in [Ca2+]i, but not IL-2 secretion (139). Cotransfection of the appropriate TCR a or p chain cDNA restored IL-2 production in response to Thy- 1-mediated activation. These apparently contradictory results with respect to Thy- 1-mediated intracellular Ca2+ mobilization in TCR-CD3- T cells may be due to differences in the epitopes recognized by the different Thy-1 mAbs, the T cells studied (e.g., normal versus leukemic), or the form of Thy-1 expression (e.g., endogenous versus transfected). It appears, therefore, that while early events of Thy- 1-mediated activation (i.e., Ca2+mobilization and perhaps IL-2R expression) may be independent of TCR-CD3 expression, the full spectrum of T cell activation via this pathway requires expression of a functional TCR-CD3 complex. In this respect it would be of interest to determine whether at least some aspects of the Thy-1 activation pathway are functional in Thy-1' TCRCDS early thymocytes. On the other hand, Thy-1 does not seem to be obligatory for T cell function, based on a recent study describing a Thy- 1(as well as CD4- CD8.) TCR-CD3' functional CTL clone (140). Furthermore, the lack of Thy-1 expression by human T cells also indicates that it is not essential for activation. Clearly, Thy-1 appears to constitute an important T cell activation molecule. Understanding its role in T lymphocyte activation would benefit greatly from the identification of its natural ligand.
240
A M N O N ALTMAN E T A L .
2. T Cell-Activating Protein (TAP, Ly-6, CD59) TAP was originally identified by a number of mAbs raised against allogeneic T cells and selected on the basis of their ability to stimulate IL-2 secretion by antigen-specific T cell hybridomas (141), an effect produced by the soluble mAbs in the absence of accessory cells. These antibodies were also found to be mitogenic for resting peripheral T cells in the presence of accessory cells or IL-1 (141), to be mitogenic for mature medullary thymocytes in the presence of PMA (142), and to modulate antigen-specific T cell responses (14 1). TAP is expressed on all CD4' peripheral T cells, -50% of CD8' cells (143), 20-30% of fetal or adult CD4- CD8- thymocytes, mature cortisone-resistant medullary thymocytes, and some T cell tumors. It is not expressed on bone marrow cells, resting B cells, or CTL precursors, but is induced on bacterial lipopolysaccharide-activatedB cells and activated CTLs (142- 144). The protein has a size of 10-12 kDa under nonreduced conditions and 16 kDa in reduced gels (14 1, 144, 145).Similarly to Thy- 1, it is anchored to the membrane by a covalently linked GPI; thus, it can be released from the cell surface by PI-PLC (145). Genetic analysis indicated that TAP maps to the Ly-6 locus (144), and amino-terinal sequencing confirmed homology with one of the Ly-6 alloantigens, Ly-6.1E (146). The Ly-6 locus controls the expression of, and/or encodes for, alloantigens found primarily on T and B cells. Two haplotypes, Ly-6.1 and Ly-6.2, map to this locus, and each may encode several alloantigens distinguished by their tissue distribution and by various mAbs. It is not clear whether Ly-6 antigens are encoded by a single polymorphic gene or by a multigene family (147, 148). Cloning of TAP cDNA identified it as a cysteine-rich 108-amino-acid protein with a hydrophobic carboxy-terminus typical of GPI-anchored proteins and lacking classical transmembrane or cytoplasmic domains and N-glycosylation sites (149). Restriction fragment-length polymorphism analysis by the Southern hybridization technique mapped TAP to the Ly-6 locus, and sequence comparison identified it as a product of the Ly-6.2 allele (149). TAP is not the only Ly-6 antigen that can participate in T cell activation, since anti-Ly-6 mAbs with distinct specificities have functional effects similar to those reported initially for anti-TAP antibodies (150).The exact role of TAP (or other Ly-6 alloantigens) in T cell physiology and their natural ligands are unknown, but a recent study (123) demonstrated that T A P mutants of antigen-specific T cell hybridomas are deficient in their ability to become activated by nominal antigen plus accessory cells, anti-CD3 antibodies, or mitogens. Furthermore, release
MOLECULAR EVENTS MEDIATING
?' CEI.1. A C T I V A T I O N
24 1
of TAP by PI-PLC inhibited the mitogen response of resting T cells (123). However, the same treatment could release other antigens involved in T cell activation (e.g., Thy- l), which are similarly anchored to the membrane (124).
3. Other GPI-Linked Activation Molecules In addition to the well-characterized Thy-1 and TAP, other GPI-linked proteins that can transduce activation signals in T cells have been described. LFA-3, the physiological CD2 ligand, also utilizes a GPI anchor (121,122). a. Decay-Accelerating Factor (DAF, CD55). DAF is a 70-kDa single-chain glycoprotein expressed on a variety of human cells, including erythrocytes and T cells, that plays a role in protecting cells from complementmediated lysis (151). Genetic deficiency of DAF expression results in erythrocyte sensitivity to such lysis, which is characteristic of paroxysmal nocturnal hemoglobinuria ( 151). The majority of membrane-bound DAF on lymphocytes is anchored to the membrane via a GPI linkage (152). DAF expression on T cells increased rapidly following human T cell activation by mitogens (153) and, moreover, a polyclonal antibody to human DAF stimulated resting T cell proliferation in the presence of phorbol ester (153). However, anti-DAF stimulated weaker responses than other mitogens and no increase in [Ca2+]iwas detected, even in the presence of a secondary, cross-linking, antibody (153). b. Ecto-5'-Nucleotidase (CD73). Ecto-5'-nucleotidase is one of several enzymes attached to the outer cell membrane surface via a GPI anchor (121, 122). It is expressed on subsets of human T and B cells, and its activity is markedly reduced in lymphocytes of patients with several immunodeficiency diseases ( 154). Ecto-5 '-nucleotidase was recently found to stimulate the proliferation of a human T cell subset when added to cultures together with PMA (155). c. Rat RT-6 Alloantigen. RT-6 is a rat T cell differentiation alloantigen existing in at least two allelic forms (i.e., RT-6.1 and RT-6.2) and linked to the membrane via a GPI anchor (156). RT-6 alloantisera were recently found to induce polyclonal activation of rat T lymphocytes (157). D. ACCESSORY ACTIVATION ANTIGENS Antibodies to TCR-CD3, CD2, and Thy-1 can, at least under certain circumstances (e.g., in the presence of accessory cells), stimulate T cell proliferation in the absence of additional exogenous signals, such as phorbol esters IL-1 or IL-2. These cell surface molecules are therefore thought to provide a primary activation signal to T cells. In contrast, a considerable number of other cell surface antigens were found to partici-
242
AMNON ALTMAN ET AL.
pate in T cell activation, albeit they were unable to provide a primary stimulus, Often, the function of these antigens was defined by testing the ability of the corresponding mAb to provide necessary activation signals in cultures of monocyte-depleted highly purified T cells activated with nonmitogenic forms of CD2- or CD3-specific antibodies. Furthermore, attempts at defining the mechanism of action of such accessory molecules were made by replacing the relevant mAb with defined soluble mediators, such as IL-1 or IL-2. Inclusion of various cell surface antigens in this group is somewhat arbitrary, since they function via distinct mechanisms and, in a few instances, appear to constitute a primary activation signal. While unable to induce the full spectrum of T cell activation, crosslinking of these antigens generally induces a common event, that is, mobilization of intracellular Ca2+ leading to a detectable increase of [Ca2+]i(158).
I. CD4lCD8 The T cell-specific CD4 and CD8 glycoproteins, which have been molecularly cloned and well characterized, separate most mature T lymphocytes into mutually exclusive subsets. CD8 or CD4 expression on the surface of T cells correlates with their ability to recognize nominal antigen in the context of MHC class I or I1 molecules, respectively (159-162). CD4 is a 55- to 67-kDa single-chain polypeptide. CD8 can be expressed as a heterodimer of a (34-38 kDa, originally termed Ly-2 in the mouse) plus /3 (28-30 kDa, Ly-3) subunits, or as an a-a homodimer (159, 159a). The function of CD4 or CD8 in T cell development and activation is still a matter of debate. Anti-CD4 or -CD8 antibodies modulated T cell responses in earlier studies (see, e.g., Ref. 163), and the ability of such antibodies to inhibit T cell function even in the absence of CD4 or CD8 ligand (i.e., MHC molecules) led to the suggestion that CD4 or CD8 transduces independent negative signals to T cells ( 164). However, this model is incompatible with several recent observations (see below). Based on recent evidence, the CD4 or CD8 glycoproteins are thought to play several distinct roles in T cell maturation and activation. First, they appear to stabilize and increase the avidity of the interaction between TCR-CD3 and the nominal antigen-MHC complex by binding to monomorphic determinants on MHC class I or I1 molecules. This association may actually link CD4 or CD8 to the TCR-CD3 complex, at least temporarily, during antigen presentation. Evidence for this role stems from studies in which the TCR-CD3 complex and CD4 or CD8 antigens were found to comodulate (165-167) or, more directly, by showing physical association between CD4 or CD8 and MHC molecules using coimmunoprecipitation techniques (168) or by direct binding measurements be-
MOLECULAR EVENTS MEDIATINC;
-r CELL ACTIVATION
243
tween CD4 or CD8 and MHC molecules expressed in transfected cells o r in liposomes (169- 171). Two amino acid residues in the invariant a3domain of class I MHC molecules involved in binding to CD8 have been tentatively identified (172, 173). In addition, several studies clearly demonstrated that: (1) transfer of CD8, together with transfected TCR a and /3 chain genes, into an acceptor T cell line was required for expression of the donor's antigen-specific cytolytic function (174); (2) CD4 or CD8 molecules augmented T cell activation by antigen (175-178); and (3) CD4 or CD8 cross-linking to the TCR-CD3 complex enhanced the effectiveness of T cell stimulation over that induced by anti-TCR-CD3 antibodies alone (162, 165, 179-182). Second, CD4 or CD8 may play an even more active role in T cell activation by transducing an independent signal. Consistent with this is the demonstration that antibody-mediated cross-linking of CD4 or CD8 induced an ethyleneglycoltetraacetic acid (EGTA)-resistant increase in [Ca"];, resulting from intracellulir Ca2+ mobilization (158). A soluble anti-CD4 antibody was found to inhibit Con A- or antigen-induced Ca2+ influx in an antigen-specific T cell hybridoma without affecting PIS hydrolysis or, presumably, intracellular Ca2+ mobilization (183). This suggested that CD4 may be functionally associated with a Ca2+channel. In another study (18 1) the accessory signal provided by CD8 cross-linking was highly sensitive to H-7, an inhibitor of PKC and cyclic nucleotidedependent protein kinases, implying activation of one (or more) of these kinases by CD8-mediated signaling. In a recent study CD4 triggering by the human immunodeficiency virus- 1 (HIV-1) envelope glycoprotein, gp120 (for which CD4 serves as the cellular receptor), stimulated PIS hydrolysis and increases in [Ca2+]i and IL-2R expression (184). One anti-CD4 antibody, B66, had the unique ability to provide a primary activation signal to human CD4+ T cells (185), that is, to stimulate, in its cross-linked form, proliferation and IL-2 production in the absence of additional stimuli. Another important clue to the function of CD4 and CD8 in T cell activation comes from the recent demonstration that the intracellular domains of CD4 and CD8 are physically associated with the T cell-specific TPK, pp!~6"~ (186, 187). This kinase appears to play a critical role in T cell development and activation (see Section V,B). On the basis of these associations, we have recently proposed (188) that CD4 or CD8 may ~ 1). This regulate T cell activation via its association with ~ p . 5 6 "(Fig. association could regulate the lateral mobility of, and hence the substrate accessibility to, ~ ~ 5 6Approximation " ~ . of CD4 or CD8 to the TCR-CD3 complex during T cell activation by antigen or mitogen might allow ~ ~ 5 6to' phosphorylate ' ~ physiological substrates, such as the 6 chain of
244
AMNON ALTMAN E T A L .
APC MHC
T cell
FIG. 1. Presentation of an antigenic peptide (Ag, in black) in the context of MHC molecules to a T cell. Association of CD4 or CD8 with MHC molecules during antigen presentation provides a means for bringing the TPK encoded by the lck gene, p56, to the immediate vicinity of the receptor complex, where it could phosphorylate substrates (e.g., the 5 chain) on tyrosine residues.
the TCR-CD3 complex (43; see Section V,B) or PI-PLC. Thus, CD4 or CD8 could play an important role in regulating tyrosine phosphorylation in T cells. In fact, antibody-mediated cross-linking of CD4 was recently found to stimulate the enzymatic activity of pp56Ickand to lead to rapid phosphorylation of the 5 subunit of the TCR-CD3 complex (189). Similar cross-linking could occur during physiological T cell activation as a result of the multivalent binding of MHC molecules on APCs to CD4 or CD8. Other data suggest that CD4 or CD8 may have more complex regulatory functions not easily explained on the basis of current models. Thus, anti-CD4 or -CD8 antibodies inhibit T cell activation via non-TCR-CD3 pathways, such as CD28 (Tp44) or Tp103 (see below), in the absence of MHC class I1 molecules, as well as activation by a combination of phorbol ester plus calcium ionophore (190). Therefore, our view of the functions of CD4 and CD8 may require considerable revision as more information about these important receptors is accumulated. Beyond their role in T cell activation per se, CD4 and CD8 have a profound influence on the selection of the T cell receptor repertoire in the thymus (162,19 1). This emerging area goes byond the scope of this chapter, except to state that signals delivered via CD4 o r CD8 in differentiating thymocytes may
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
245
act in concert with some yet undefined thymic ligands to influence the negative and/or positive selection of T cells. 2. CD28(Tp44)
CD28 was originally defined by an mAb, 9.3 (192), as a 80- to 90-kDa disulfide-linked homodimer consisting of 44-kDa subunits and expressed on the surface of all CD4' and -50% of CD8' human T cells (193,194). This surface glycoprotein is not linked physically to either the TCR-CD3 complex or CD2 (193-195). A murine surface antigen with structural and functional properties similar to those of human CD28 has been identified recently (196). Treatment of human T cells with anti-CD28 mAbs has two distinct effects. In some studies such antibodies provided a primary activation signal without requiring triggering by other ligands, such as anti-TCRCD3 or anti-CD2 antibodies (193, 194, 197). In one of these studies, mAb 9.3 stimulated proliferation and IL-2 production by resting T cells in the presence of monocytes (194), although the stimulation of IL-2 production by Jurkat cells required costimulation with PMA. More often, 9.3 was mitogenic and induced IL-2 secretion only in the presence of PMA (193, 197). Soluble (197) or cross-linked (158) antibody 9.3 can induce PIS hydrolysis, an EGTA-resistant increase in [Ca2']i (due to intracellular Ca2+ mobilization), and expression of IL-2R (158). However, another study failed to show PIS hydrolysis and intracellular Ca2+ mobilization in 9.3-stimulated Jurkat cells (198). Furthermore, immobilized 9.3, in the absence of additional stimuli, was recently found (199) to activate the HIV-1 enhancer included in the long terminal repeat of the virus. Thus, although occupation of CD28 alone is sufficient for triggering at least some early aspects of T cell activation, the physiological relevance of these effects is not clear. In the majority of studies, however, treatment with anti-CD28 antibodies provided an accessory activation signal that enhanced the response (e.g., IL-2 secretion and proliferation) of T cells stimulated by other ligands, such as PHA (195), Con A (195), or anti-TCR-CD3 (195, 197, 200) or -CD2 (105, 201) antibodies. Activation by anti-CD28 antibodies alone is accessory cell dependent (194); however, when this pathway provides an accessory signal to T cells activated by other pathways (e.g., with anti-CDS), CD28-mediated triggering actually substitutes for one of the accessory cell functions (195, 197, 200), which may involve a soluble mediator, but it does not overcome the requirement for antibody (e.g., anti-CD3) cross-linking ( 197). The difference between these two effects mediated by CD28 triggering may be related to the mAb 9.3 concentration. Relatively high saturating antibody concentrations provide a pri-
246
AMNON ALTMAN E T A L .
mary stimulus, whereas lower concentrations act as an accessory signal (197). The functional relationship between the TCR-CD3 and CD28 activation pathways is controversial. Some studies demonstrated that mAb 9.3 is nonfunctional in CD3-modulated T cells (194) or in a CD3- Jurkat mutant (1 17), leading to the conclusion that the CD28 activation pathway requires a functional TCR-CD3 complex. Other studies reported opposite results (193, 197). T h e reasons for this apparent conflict are unclear, but do not seem to be related to the antibody concentration used, since both studies with CD3- Jurkat mutants used saturating amounts of mAb 9.3 (117, 197). The mechanism of anti-CD28-mediated T cell activation was analyzed in more detail in a recent study (202).AntLCD28 augmented mitogen- o r anti-CD3-induced T cell activation by stimulating a five- to 50-fold increase in gene expression and secretion of several lymphokines (202), an effect resulting from increased stability of the corresponding mRNAs (203). This effect was observed even at an optimal mitogen or anti-CD3 concentration, although under these conditions CD28 activation did not enhance the proliferative response. Unlike other activation pathways that induce lymphokine production, the CD28-mediated response was not inhibited by cyclosporine. This suggests that CD28 is part of a biochemically distinct activation pathway. 3. CD5 (Lyt-1, Leu-1, Tp67)
CD5 is the human homolog of the murine alloantigen Lyt-1, defined originally as a specific marker of helpedinducer T cells. Earlier studies demonstrated that anti-Lyt- 1 mAbs augmented T cell proliferation and IL-2 production in an allogeneic mixed-leukocyte reaction, but were not themselves mitogenic ( 163). Similar effects have been observed with human T cells. Thus, while anti-CD5 alone generally does not activate T cells, it provides an accessory activation signal for T cells. Proliferation, IL-2R expression, and IL-2 production stimulated by antLCD3 antibodies were augmented by CD5-specific mAbs (179, 180, 200, 204). This effect required cross-linking (i.e., the presence of FcR' accessory cells or immobilization onto the plastic surface) of the anti-CD3 mAbs, indicating that CD5 triggering did not substitute for the cross-linking requirement. This conclusion is also supported by the finding that Fab fragments of anti-CD5 can generate the accessory signal (200). Although not fully mimicking the effect of IL-1, these antibodies appear to replace, at least to some extent, the effect of this monokine or some other accessory cell-derived soluble mediators (200,205). Cross-linking of CD5 also stimulated intracellular Ca2+mobilization ( 158).
MOLECULAR EVENTS MEDIATING
'1' CELL A C T I V A T I O N
247
4 . Other Accessory Signals for Resting T Cells
Additional T cell surface antigens identified by mAbs, and thought to play an accessory role in T cell activation, constitute a group of T cell proteins that are generally less well-characterized than the activation determinants described above. Some of these antigens are discussed below. a. LFA-I. LFA- 1 is a member of the integrin family of adhesion receptors, and one of its ligands is the intercellular adhesion molecule type 1 (CD54). It is expressed on various types of leukocytes and is composed of a 180-kDa a subunit (CD1 la) and a 95-kDa p subunit (CD18). mAbs to LFA- 1 inhibited a wide variety of adhesion-dependent leukocyte functions, such as interactions of cytolytic effector cells with their targets and those between T h cells and APCs (reviewed in Refs. 95 and 206). CD 11aand CD 18-reactive mAbs were found to synergize with submitogenic doses of cross-linked anti-CD3 or with PMA in inducing IL-2R expression, IL-2 secretion, and T cell proliferation (179, 206). A cross-linked antLCD18 antibody also induced an increase in [Ca2+]iin purified T cells (158). b. CD6, CD7. CD6 (120 kDa) is expressed on mature T cells and on a B lymphocyte subset. A cross-linked CD6-specific mAb was found to stimulate intracellular Ca" mobilization leadin4 to increased [Cazt]i (158) and, in combination with a F (ab')2 frament of anti-CD3 mAb, to synergize in inducing T cell proliferation (179, 180). CD7 is a 41-kDa glycoprotein expressed on all T cells. An mAb to CD7 was found to induce effects similar to those of the antLCD6 mAb, that is, Ca2+ mobilization ( 158) and anti-CD3-dependent T cell proliferation and IL-2 secretion (180, 207). As noted earlier, a common theme for all of the accessory molecules described in this section is their ability to stimulate an increase in [Ca2+]; upon cross-linking by their respective antibodies. However, this effect is not specific for the accessory molecules described above, since antibodies to other surface antigens, such as MHC class I antigens (180, 208) and P2-microglobulin (180), can produce similar effects and augment T cell proliferation mediated by anti-CD2 or -CD3 antibodies. T h e physiological relevance of increased [Ca2+]iin the absence of proliferation and lymphokine secretion is unclear. Furthermore, for most accessory activation molecules described here, it has not been determined whether this increase is caused by PIS hydrolysis and formation of the secondary messenger, IPS, or possibly by some other mechanism. This [Ca"]; increase may be a general response to the cross-linking of surface antigens and may be a signal for redistribution and polymerization of cytoskeletal
248
AMNON ALTMAN E T AL.
proteins that become associated with the membrane at the site of antigen capping ( 158). E. OTHERACTIVATION ANTIGENS As a corollary to the production of additional mAbs against various T cell types, several new activation antigens were identified recently. Some are present on resting T cells; others are up-regulated or induced de novo upon T cell activation. Most of these antigens have not been characterized beyond showing that they can activate T cells. Thus, the early signaling events they utilize and their functional relationship to the TCR-CD3 complex are, for the most part, undefined. A brief description of these antigens follows. I , CD43 (Sialophorin)
CD43 is a heavily sialylated 115-kDa glycoprotein with broad distribution in the hematopoietic system. Its expression is deficient in lymphocytes of Wiskott-Aldrich syndrome patients manifesting reduced T cell function (209). A CD43-specific mAb, L10, was found to stimulate proliferation of human peripheral blood T cells (2 10). Proliferation required monocytes and the Fc portion of the mAb, but was independent of other signals, such as PMA or mAbs to other T cell antigens. In contrast to this finding, two other mAbs to sialic acid-dependent CD43 epitopes were poorly mitogenic alone, but synergized effectively with PMA, soluble anti-CD3, or mitogenic lectins (211). This difference may be due to the different epitopes recognized by these antibodies. 2. Tp90 Tp90, a 90-kDa protein, was identified with an mAb prepared by immunizing mice with a mixture of two human T cell leukemia lines (212). It does not comodulate with CD3, but apparently depends on the TCR-CD3 complex, since its ability to stimulate IL-2 secretion by Jurkat cells, when combined with PMA, was abolished by modulation of the latter complex from the cell surface. The antibody was mitogenic for, and induced IL-2 secretion by, a subset (3- 14%of the total T cells) of human peripheral blood T cells included preferentially within the CD8' population. The action of this mAb on resting T cells was monocyte dependent and was less than that of the anti-CD3 mAb (212). 3. Early Activation Antigen-1 (EA-1, CD69) EA-1 was originally defined by a murine mAb generated against PMAactivated human T cells (213). The antibody defined a 60-kDa disulfidelinked homodimer consisting of phosphorylated and differentially glyco-
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
249
sylated subunits of 28-32 kDa, having a 24-kDa core protein (213,214). EA-1 is expressed on -60% of human thymocytes and at low levels on 6-7% of blood lymphocytes. Following PMA activation, EA-1 is rapidly induced on T cells and precedes the induction of IL-2R. T cells activated by other mitogens, as well as activated B cells, also express this antigen, albeit with delayed kinetics (213), and the induction of EA-1 requires mRNA and protein syntheses. PKC activation appears to be essential for EA-1 induction, and Ca2'-dependent pathways may also be involved in the process (214). A more recent study (215) indicated that EA-1 is most likely identical to the protein defined by the mAb, Leu-23 (216). EA- l-reactive mAbs can stimulate T cell proliferation, IL-2R expression, and IL-2 secretion in the presence of PMA, in a monocytedependent fashion (2 15). The cross-linked antibodies also stimulated an increase in [Ca2+]i. Similar structural and functional properties were ascribed to an antigen identified recently by another mAb, termed MLRS (2 17).Thus, EA- 1, Leu-23, and MLRS may define the same antigen. T h e exact function of this molecule in the activation of T cells and other cells remains to be determined. 4 . TplO3
An mAb raised against a human CD4' CTL clone has defined a 103-kDa T cell-specific activation antigen. Tp103 is not expressed, or is present only in low amounts, on resting T cells. It is found on all proliferating normal T cells, but not on several malignant T (or other) cell lines (218). Cross-linking of anti-TP103 mAbs by FcR' cells stimulated proliferation and cytolytic activity in human T cell clones via an IL-2dependent autocrine pathway (218). In the presence of accessory cells and exogenous IL-2, these antibodies were also mitogenic for the minor subset (-5%) of Tp103' resting T cells. Modulation with antLCD3 mAbs abrogated activation via the Tp103 antigen (but not IL-2-induced proliferation), suggesting that a functional TCR-CD3 complex may be required. Interestingly, T p 103-mediated activation was inhibited by antiCD4 or -CD8 antibodies (190). This effect could reflect a role for the TCR-CD3 complex in transducing the T p 103-generated signal, in agreement with the effect of CD3 modulation (218). 5. 2 H l
The 2H 1 mAb (2 19) immunoprecipitates, under reducing conditions, two major polypeptides of 140 and 105 kDa from human T cells. T h e antigen is expressed on 90% of T cells, 10% of thymocytes, and some leukemic T cell lines, but not on other lymphoid and hematopoietic cells. Its expression on thymocytes was dramatically increased by activation
250
AMNON ALTMAN El' AL.
with Con A plus PMA. This antibody synergized with PMA or with a non-mitogenic anti-CD2 mAb in inducing the proliferation of purified T cells. The antigen did not comodulate with CD3, indicating that it is not physically linked to the latter (219). 6 . Thymocyte-Activating Molecule (THAM)
T cell development and maturation are regulated by poorly defined signaling events involving cell-cell contact and lymphokines. Activation through alternative antigen-independent pathways may be physiologically relevant in immature thymocytes that do not yet express the TCRCD3 complex. Thus, identification of activation molecules expressed selectively on immature thymocytes would be an important step toward understanding of proliferation and differentiation events occurring in the thymus. THAM is a heterodimeric (110- and 128-kDa) protein expressed on the majority of fetal or adult thymocytes, and at particularly high levels on CD4- CD8- cells, the precursors of mature T cells. Its expression is correlated with IL-2R expression in early thymocytes (220). A crosslinked anti-THAM antibody induced the proliferation of both immature and mature (CD4+or CD8') thymocytes in the presence of either PMA or IL-1 and IL-2 (220). PMA or IL-1 acted to increase the level of IL-2R expression. Proliferation was abolished by anti-IL-2R antibodies, indicating that it is strictly IL-2 dependent. It would appear that THAMmediated activation is TCR-CD3-independent, since it occurs in immature thymocytes. Furthermore, THAM is also expressed by peripheral B cells and a fraction of bone marrow cells (220).
7 . Tp135-145 Tp135-145, a heterodimer consisting of 135- and 145-kDa peptides, was recently identified by a monoclonal antibody, MX24, and is expressed on ~ 4 0 - 7 5 % of resting peripheral blood T cells (CD4' and CD9+), on CD3-TCR- T cell lines, and on B cell lines (221). It did not comodulate with the TCR-CD3 complex, and stimulation of Jurkat or peripheral blood T cells with the monoclonal antibody led to 1L-2 production and proliferation. It would be of interest to determine whether this and the other heterodimers mentioned earlier (e.g., 2H1 and THAM), which are somewhat similar in size, are structurally related. F. IL- 1 RECEPTORIN T CELLACTIVATION IL-1 is a polypeptide produced in response to infection, injury, or antigenic stimulation and has a very broad spectrum of biological activities in nearly every organ system in the body. Although it is produced
MOLECULAR EVENTS MEDIATING
?' CELL AC:TIVAI'ION
25 1
primarily by macrophages, other cell types can also produce IL-1. Two distinct types of IL-1 that are encoded by two separate genes have been identified to date. The biology of IL-1 has been reviewed (222-225) and is beyond the scope of this chapter. However, a short discussion of its role in T cell activation and the receptor(s) and signaling pathways associated with its action is pertinent. Although it is clear that IL- 1 can enhance T cell activation, it may not be an obligatory factor, as indicated by the inability of neutralizing antiIL- 1 antibodies to affect mitogen- or antigen-induced T cell proliferation (223) and other criteria (226). Two classes of IL-1 receptors with Kd's of -5 X lo-'' and =5 X lo-'' M have been identified, and T cells express both of them. T h e 80-kDa IL-1 receptors originally identified by crosslinking studies with labeled IL-1, has been purified and molecularly cloned (227). T h e cDNA corresponds to a 65-kDa protein with a substantial cytoplasmic domain, indicating the potential to function as a signaltransducing molecule. However, since the concentration of IL- 1 required for half-maximal saturation of the cloned IL- 1-binding protein ("5 X 10'"' M) is considerably higher than the biologically active IL-1 concentrations, it has been suggested that additional protein(s) converting the receptor to a high-affinity state may exist (225). In fact, additional IL- 1-binding proteins of different sizes were identified in cross-linking experiments. The ability of IL-1 to produce clear biological effects at extremely low concentrations, even in cells that do not display detectable IL- 1 binding (225), has raised questions about the signaling machinery used by IL-1. It is possible, of course, that signal transduction can occur after occupation of few receptors. Several recent studies have dealt with IL- 1R-mediated signal transduction (228-232). In a simplisitic approach the ability of IL-1 to functionally mimic at least sorpe of the effects of phorbol esters in enhancing T cell activation (1, 220, 228) would indicate that PKC activation is a major component of the IL- 1-mediated signaling pathway. This is based on the potent PKC-activating properties of phorbol esters (see Section V,A,l). For example, the requirement of PHA and IL-1 for stimulation of high-level IL-2 secretion by the murine T cell lymphoma line LBRM33 can be replaced by ionomycin and PMA, respectively (228). However, the assumption that IL-1 and PMA generate the same signal is contradicted by the finding that IL-1 does not activate PKC, either directly or via PIS hydrolysis (228), indicating that IL- 1-mediated signaling is likely to be more complex. In an extension of this study, it was found (229) that, under conditions in which PHA-stimulated PIS hydrolysis and IL-2 production are inhibited by PMA, IL-1 overcame this inhibition and restored IL-2 secretion without affecting [Ca2+]i.However, another study (230)
252
AMNON ALTMAN E T AL.
suggested that IL-1 may modulate PKC activity, based on the finding that, in Jurkat cells, it enhanced the synthesis of phosphatidylserine, a membrane phospholipid known to be an important cofactor for PKC activation. Under the same conditions, IL-1 did not stimulate PIS hydrolysis or a rise in [Ca2+]i,nor did it modify the intracellular level of the cyclic nucleotide second messengers cGMP and cAMP (230). This agrees with our recent finding that IL-I stimulation of 11-2Ra (p55) chain expression on the human natural killer cell-like leukemic line Y T did not involve a detectable change in cAMP (233). However, another recent study (231) indicated that cAMP may act as a secondary messenger for IL-1. This conclusion was based on the findings that: (1) IL-1 stimulated cAMP synthesis in a variety of responsive cell types, including Y T cells that were negative in our laboratory (233); (2) cAMP analogs or CAMP-inducing agents replaced IL-1 in inducing IL-2R expression and thymocyte proliferation. However, it is important to note that the ability of two agents to produce the same biological effect by no means indicates that they use the same signaling pathways. While both IL-1 and cAMP can stimulate IL-2R expression on Y T cells (231, 233), our finding of a clear synergism between the two (233)suggests that they may, in fact, use distinct pathways. The most intriguing findings regarding IL- 1-induced early biochemical events emerged in a recent study by Rosoff et al. (232). Using peripheral blood T cells or T cell lines, they found that IL-1 stimulated production of the second messenger DAG via a pathway that does not involve PIS turnover, IPS production, or a rise in [Ca2+]i.The source of DAG was found to be phosphatidylcholine, since the accumulation of DAG was accompanied by a parallel hydrolysis of this lipid and accumulation of phosphorylcholine. The production of DAG from phosphatidylcholine has been documented in other systems (234, 235). The phosphatidylcholine-derived DAG differed in its fatty acid composition from the PI-derived DAG (232). If different DAG species can act at different locations in the membrane to activate distinct PKC isoenzymes (see Section V,A, l), this could explain the comitogenic effect of ligands stimulating PIShydrolysis (e.g., anti-CD3 antibodies) combined with those that do not utilize this pathway, such as IL-1. The intriguing aspect of this study (232) was the ability of IL-1 to stimulate DAG production rapidly (within 5 seconds) at exceedingly low concentrations (50% maxmial response at =3 X M) and, most remarkably, in the absence of detectable binding to Jurkat cells. On the basis of these findings, it was suggested that IL-I may utilize two signaling pathways, one involving the high-affinity IL-1 receptor coupled to an unknown transduction system, and the other operating via a novel, pre-
'r
MOLECULAR EVENTS MEDIATING C:EL.LACTIVATION
253
sumably receptor-independent, mechanism involving phosphatidylcholine breakdown. Direct activation of PLC by IL-1, or an intrinsic PLC activity of IL-1 itself, was postulated as a component of this secondary mechanism (232). Although all of the relevant studies agree that IL-1 does not stimulate PIS hydrolysis and increased [Ca2+]i,they differ in other aspects and clearly suggest that IL-1 may utilize more than one signaling pathway. This is hardly surprising in view of the fact that IL-1 exerts pleiotropic effects on a multitude of target cells via at least two receptor classes. Precedents exist for such a situation; for example, the action of glucagon was found to involve two distinct signal transduction systems, that is, the adenylate cyclase and the PIS hydrolytic pathway coupled to the high- o r low-affinity receptors for this hormone, respectively (236). Similar situations were reported for other hormones. 111. PI Turnover and Formation of Second Messengers in T Cells
Receptor-mediated hydrolysis of inositol phospholipids is a common mechanism for signal transduction of extracellular stimuli, such as hormones, neurotransmitters, some growth factors, antigens, and other biologically active substances. T h e rapid response of membrane PIS to the stimulation of cell surface receptors was first recognized in acetylcholinestimulated cells (237). It was later found that the rapid receptorstimulated labeling of PI and phosphatidic acid resulted from enhanced breakdown and resynthesis of PIS (238). T h e discovery that T cell stimulation by mitogens or anti-TCR-CD3 mAbs also leads to PIS hydrolysis (239-242) and an increase in [Ca2+]i(55) prompted many studies on the role of PI turnover in T cell activation. A. PICYCLE T h e PIS constitute less than 10% of the total membrane phospholipid in most mammalian cells, with PI being the most abundant form. PI can be sequentially phosphorylated at positions 4 and 5 of the inositol ring (Fig. 2), giving rise to the polyphosphoinositides phosphatidylinositol-4monophosphate (PIP) and PIP2, which account for 10% or less of PIS. These sequential phosphorylations are catalyzed by two highly active kinases, type I1 PI kinase (243) and PIP kinase. The former is a 60-kDa membrane protein (244) stimulated by polyamines (245), mitogenic lectins (241), and anti-CD3 antibodies (246) in T cells. It was earlier suggested (247,248) that this kinase is a product of the c-src or c-ros protooncogenes. Closer biochemical studies of the purified PI kinase, however, demonstrated significant differences in substrate specificity and ion re-
254
A M N O N ALTMAN E T AL.
quirements of the PI kinase and the protooncogene-encoded proteins (243). A recent development of considerable interest is the association of another type of PI kinase (type I) with TPKs, including growth factor receptors and transforming retroviral proteins. This 8 1- to 85-kDa PI kinase (reviewed in Ref. 249) phosphorylated PI on position 3 of the inositol ring, giving rise to PI-3-P, which could potentially become phosphorylated to PI-3,4-P:, and P1-3,4,5-P~.Furthermore, the relevant growth factors or the oncogenic TPKs activated this kinase, presumably by tyrosine phosphorylation and, as a result, increased intracellular levels of PIP (249). Thus, tyrosine phosphorylation is likely to affect the levels of PIS and the activity of the corresponding PI kinases in an important way. PIP kinase is probably mostly cytosolic and is still poorly characterized. It is inhibited by its product, PIP:, (250). Recently, PIP kinase was shown to be stimulated by nonhydrolyzable GTP analogs, suggesting that it might be regulated by a guanine nucleotide-binding regulatory protein (G protein) (251).A rapid dephosphorylation of PIP and PIP:, is catalyzed by two phosphomonoesterases, which have not been studied in any detail. It is generally thought that the different inositol phospholipids are in a rapid equilibrium with each other (252-254). PI-PLC generates cyclic and noncyclic inositol l-monophosphate (IP), inositol 1,4-bisphosphate (IP:,),and IPS from the cellular PIS pool. The other product of PI hydryolysis is DAG, a second messenger that activates PKC (see Section V,A, 1) This PI-PLC-mediated hydrolysis is a rapid response to receptor occupancy. Both DAG and inositol phosphates are used in the resynthesis of PIS. It has been suggested (254) that the production of IP, IP:,, and IPS is due to hydrolysis of all three corresponding forms of PIS. This conclusion was based on the findings that, first, a single enzyme purified to homogeneity from sheep seminal vesicles (255) is capable of hydrolyzing PI, PIP, and PIP:,. Second, hydrolysis of PIP:, in rat pituitary cells proceeded for less than 2 minutes, while PI hydrolysis continued for 30 minutes (256), kinetics similar to those in antigen-stimulated cytotoxic T cells (257). Finally, the stoichiometry of IPSversus DAG production in thrombin-stimulated platelets (258) or in vasopressin-stimulated hepatocytes (259) indicates that the amount of PIP:, hydrolyzed is not sufficient to account for the mass of DAG produced. In addition to the six inositol phosphates mentioned above, other isomers have been isolated from cells undergoing PIS hydrolysis. Many more could still be found, since the inositol moiety has six hydroxyl groups (Fig. 2) that can be phosphorylated, leading to a total of 66
MOLECULAR EVENTS MEDIATING
1'CELL ACTIVATION
255
FIG. 2. Structure of myo-inositol, the naturally occurring inositol stereoisomer. 1-6 indicate carbons 1-6; for example, inositol-1,4,5-trisphosphate(IPS) contains phosphates attached at positions 1,4, and 5 of the inositol ring.
possible arrangements of inositol phosphate compounds (260). However, only IP3 and IP4 have been shown to have biological activity. IP3 induces the release of Ca2+from intracellular storage sites (261), and IP4 appears to alter the plasma membrane permeability to extracellular Ca2+(262; see Section VI,A). IP4 is formed from IPS by an IPS-3-kinase (263), an enzyme also present in leukemic Jurkat T cells (264) and murine thymocytes (265). The activity of this kinase, and hence the levels of IPSand IP4, is regulated by Ca2+and calmodulin (CaM) (265, 266). It has been suggested that cyclic inositol phosphates may play a role in PI-associated signal transduction (254, 267). This notion is supported by the finding that both the cyclic and noncyclic forms of IPS can release Ca2+ from permeabilized cells (268). The inositol phosphates formed during agonist-induced hydrolysis of PIS are mainly removed by successive dephosphorylations ultimately yielding free inositol. The biologically active inositol phosphates, IPSand IP4, are dephosphorylated by a 5-phosphomonoesterase (5-PME) to inositol phosphates that are inert, at least with respect to Ca2+mobilization (254). 'Thus, 5-PME is analogous to the cyclic nucleotide phosphodiesterase, which hydrolyzes the biologically active cyclic nucleotides to inert 5'-monophosphates (254). 5-PME is specific for IPS and IP4, as no other bis- or trisphosphate compound is hydrolyzed (269). 5-PME is phosphorylated by PKC, resulting in a fourfold enhancement of its activity (270). The precise dephosphorylation pathways of inositol phosphate isomers are poorly understood. The identification of a variety of isomers having phosphates on positions 1-5 (Fig. 2) suggests a complex breakdown. The final dephosphorylating enzyme, inositol- 1monophosphatase, has a size of 58 kDa and is inhibited by Li+ ions (271, 272). As expected, Lit causes an accumulation of IP in cells (273). Thus, Li+ is used in measurements of PISturnover to prevent complete dephosphorylation of formed inositol phosphates.
256
AMNON ALTMAN ET A L .
B. PI-PLC I . Structure and Function of PI-PLC PI-PLC is present in all cell types studied, and most of the activity is cytosolic, although its substrates are in the plasma membrane. Membrane-bound PI-PLC has also been described (reviewed in Ref. 254). Many of the earlier studies of PI-PLC described various forms of the enzyme, even within the same tissue or cell type. These forms could be distinguished by their size, Ca2+requirement, cellular localization, pH optimum, and preference for distinct substrates (274-280; reviewed in Ref. 254). The exact relationships among these forms are far from clear, although some of them have been distinguished by specific antisera or mAbs (275, 280, 281). PI-PLCs have also been isolated from lymphocytes (282-286). About 90%of the lymphocytes' enzymatic activity was present in the cytosol, and it displayed two pH optima (282-286). The cytosolic activity was resolved into several peaks that displayed selective activity toward PI or PIP2. In general, the lymphocyte-derived PI-PLCs preferentially hydrolyzed PIP2 at low Ca2+ concentrations (<1 -); however, PI was the preferred substrate at higher Ca2+ concentrations (282-286). These findings suggested that, at resting [Ca2+]ilevels, a receptor-stimulated PI-PLC activity hydrolyzes preferentially PIP2 prior to Ca2+mobilization, while hydrolysis of PI is initiated after [Ca2+]iincreases above a critical threshold level. Support for this hypothesis came from the observation that Jurkat T cells pretreated with a combination of the calcium ionophore and EGTA to deplete intracellular Ca2+stores still responded to PHA by PIP2 hydrolysis, although no increase in [Ca2+]iwas observed (287). In addition, Jurkat T cells stimulated with anti-TCR antibodies in the presence of EGTA (to block Ca2+influx) can mobilize intracellular Ca2+,a response to the formation of IPS and, hence, PIP2 hydrolysis (242). In contrast, however, PIP2 hydrolysis was not detectable in murine CTL clones stimulated with antigen (288) or Con A (288, 289) in the presence of EGTA. These discrepancies could reflect differences between human and murine systems or transformed and nontransformed cells, as well as between different sensitivity levels of the assays. The notion, derived from earlier biochemical studies, that multiple forms of PI-PLC exist was recently verified by the molecular cloning of cDNAs encoding four distinct forms of PI-PLC from bovine brain (290293) and guinea pig uterus (293a). The three brain enzymes having approximate molecular weights of 150-154, 145-148, and 85 kDa were originally termed PLC-I, -11, and -111, respectively (290), and, more recently, PLC-p, -7, and -6 (293b). Two homologous regions of 150 and
MOLECULAR EVENTS MEDIATING
'I' CELL A C T I V A T I O N
257
120 amino acids in length are shared among the p, y , and 6 forms. In addition, brain PLC-y displays >50% homology over a region of ~ 3 0 0 amino acids, with a well-conserved (regulatory?) domain of the src family of nonreceptor TPK (291, 292, 294). The recently described crk oncogene (294) also encodes this shared region, but not the catalytic domain of the TPKs; nevertheless, the crk protein coimmunoprecipitates with a TPK, and fibroblasts expressing the crk oncogene contain elevated levels of proteins phosphorylated on tyrosine residues (294). T h e arrangement of this homologous region in crk is more similar to PLC- 148 than to the src family of TPK, suggesting (294) that the crk gene arose by transduction of this (regulatory?) region of PLC-148 (PLC-8). It can be expected that the cloning and determination of primary structure of distinct PI-PLC enzymes will soon lead to understanding of their roles in different cellular functions. T h e differential expression of these enzymes in various cell lines and tissues supports the notion that they fulfill distinct, perhaps cell-specific, functions (reviewed in Ref. 293b). Their expression in T lymphocytes has not yet been reported. 2 . Regulation of PI-PLCs
PI-PLCs are regulated by multiple factors. In many systems the component that couples ligand-occupied receptors to PI-PLC is a G protein. Receptor-activated G proteins are thought to directly interact with PIPLC and increase its affinity for Ca2+,thereby activating it (295). Evidence for the control of PI-PLC by G proteins during T cell activation is discussed in more detail in Section IV. Th e fii.4ing that mammalian PI-PLC is far more efficient in hydrolyzing pure P1 in the form of lipid vesicles or detergent dispersions than in cell membranes (296,297) led to the suggestion that membrane phospholipids regulate PI-PLC activity. Studies of platelet-derived PI-PLC (297) indicated that phosphatidylcholine inhibits its activity by preventing interaction of the enzyme with PI. Phosphatidylserine and free fatty acids overcame this inhibition, and DAG further stimulated the enzyme 2.5fold. Both membrane-bound and cytosolic PI-PLC activities derived from mature murine T cells were inhibited by phosphatidylcholine and phosphatidylserine (298). In contrast, thymocyte PI-PLC activity was enhanced by these phospholipids (286). PIP2 was also shown to directly inhibit the activity of platelet-derived PI-PLC toward PI, by a mechanism distinct from substrate competition (299). It was estimated that, in unstimulated platelets, PI-PLC activity may be inhibited by >75%. Presumably, as PIP2 is consumed by receptor-stimulated hydrolysis, this level of inhibition decreases, thus allowing PI-PLC to efficiently hydrolyze PI. As mentioned earlier, the Ca2+concentration is an important regulator
258
A M N O N AL'I'MAN ET AL.
of PI-PLC, determining both the rate of PI hydrolysis and the preferred substrate in in vitro assays. The preference of PI-PLC for PIP2, as compared to PI or PIP, at low (0.1-1 p M ) Ca2+concentrations supports the view that PIP2 hydrolysis occurs at resting [Ca2+]ilevels and precedes Ca2+ mobilization, which subsequently results in an increase in [Ca2+]i and stimulation of PI hydrolysis by PI-PLC (254). PI-PLC may be regulated directly or indirectly by a TPK. Thrombininduced PI turnover in Chinese hamster fibroblasts was enhanced when the cells were costimulated with fibroblast growth factor (FGF) (300). FGF by itself had no effect on PIS turnover, but activated an FGF receptor-associated TPK. Several other TPK-activating growth factors similarly enhanced thrombin-induced PIS turnover in these cells. Other studies in Swiss 3T3 cells (301), BALB/c 3T3 cells (302), and human fibroblasts (303) revealed an enhancement of the basal rate of PIS turnover by several TPK-activating growth factors. Likewise, transformation of epithelial cell lines with the oncogenic TPKs v-src, v-fes, v-fms, and v-abl increased PIS hydrolysis by approximately fourfold over basal levels (304). Recent studies with epidermal growth factor-stimulated A43 1 epidermoid carcinoma cells demonstrated direct tyrosine phosphorylation and activation of PI-PLC (305). The ability of CAMP to inhibit PIS breakdown in various cells raises the possibility that it might exert a negative regulatory influence on PI-PLC. However, this possibility has not been formally proven. In T cells, cAMP does not block PI turnover stimulated by the G protein activator, aluminum fluoride (AlF;) (see Section IV,B), indicating that cAMP does not directly interfere with the function of PI-PLC. C. PIS TURNOVER IN T LYMPHOCYTES I. T Cell Activation b~ Phorbol Esters and Calcium Ionophores
One approach to studying the function of second messengers in cellular activation has been to use pharmacological agents that bypass receptor triggering and mimic downstream events. Although earlier studies documented an increase in [Ca2+]iin T cells stimulated by mitogens (306,307) or anti-TCR-CD3 antibodies (55),it has not been possible to completely replace the mitogenic effect of lectins with calcium ionophores (306), raising the possibility that an additional signal is required for T cell proliferation. T w o subsequent findings suggested that PKC activation provides the putative second signal: first, the realization that the [Ca2+]irise results from PIS breakdown (242), which generates two second messengers (i.e., IP3 and DAG) and, second, that DAG is the physiological PKC activator. Upon finding that phorbol esters bind to
(308, 309) and directly activate (310) PKC, the use of these agents to replace the signal provided by DAG has become widespread. Truneh et al. (3 11) first reported that the treatment of resting T cells with a combination of phorbol ester plus calcium ionophore leads to proliferation, and these two agents were subsequently found to replace the signal delivered by antigen to antigen-specific T cell lines and clones (72, 73, 312). T h e requirement for a combination of PMA plus calcium ionophore in inducing proliferation reflects the fact that neither agent alone stimulates IL-2 production, but the two together effectively accomplish this (72, 312-316). This combination is also effective in stimulating IL-2 secretion and proliferation by immature, CD4- CD8- thymocytes unresponsive to surface-acting ligands (317). Phorbol ester plus calcium ionophore, however, do not mimic the growth-promoting effect of IL-2 (72,73),indicating that IL-2 triggers a signaling pathway different from the TCR-CD3 complex (see Section IX,B). In contrast to IL-2 production, which requires a combination of phorbol ester plus calcium ionophore, either agent alone can induce IL-2Ra (p55) chain mRNA and cell surface expression (71, 73, 79, 80, 312, 314, 318-323), as well as phosphorylation of that receptor and some of the CD3 subunits (see Section V). The effects of phorbol esters are most likely due to PKC activation, while those mediated by calcium ionophores are probably mediated by other kinase(s); thus, Ca2+ alone does not activate PKC in intact T cells, as measured by translocation of the enzymatic activity to the membrane (324). Caution must be exerted in equating the effects of phorbol esters with those of the physiological PKC activator DAG. First, phorbol esters are metabolized much more slowly than is DAG, and this difference is probably responsible for the sustained and long-term PMA-induced PKC translocation, as opposed to the transient translocation stimulated by mitogens or anti-CD3 antibodies (205,324-329). In addition, it may also account for the difference in the functional effects of DAG and phorbol esters on T cells and the need to repeatedly stimulate T cells with DAG in order to activate them (312). Second, the existence of distinct PKC isoenzymes (see Section V,A, 1) with different activation requirements, and perhaps different substrate specificities, suggests that PKC isoenzymes are regulated independently (330). Thus, T cell activation with mitogens or antireceptor antibodies may activate some PKC isoenzymes selectively (331), whereas phorbol esters most likely activate them all. Finally, the possibility must be entertained that phorbol esters exert some of their pleiotropic effects via pathways independent of PKC activation. T h e signal requirements for IL-2 production and T cell proliferation raise a certain paradox: Ligands that trigger the TCR-CD3 complex
260
AMNON ALTMAN E T AL.
stimulate PIP2 hydrolysis and, hence, production of the two second messengers, IPS and DAG, whose action can be replaced by calcium ionophore and phorbol ester, respectively. It follows that TCR-CD3initiated PIP2 hydrolysis should provide the necessary and sufficient signals for optimal T cell activation. Nevertheless, in several experimental systems engagement of the TCR-CD3 complex was not sufficient for IL-2 production, and an accessory signal [e.g., PMA (55,313,332)or IL-1 (228)]was required. This requirement is difficult to explain, since PMA is considered to exert its effect via the activation of PKC, an event that is also stimulated by DAG. As suggested by others (see e.g., Refs. 228 and 232), this paradox could be resolved if it is assumed that PMA and DAG mediate qualitatively or quantitatively different effects on T cells. Similarly, the ability of IL-1 to stimulate production of a DAG species clearly distinct from PIP*-derivedDAG (232) could explain the requirement for the accessory signal of IL- 1, since different DAG species could selectively activate distinct PKC isoenzymes. Nevertheless, in IL- 1-independent T cell clones the complete set of transmembrane signals required for activation can be delivered through the TCR-CD3 complex in the absence of accessory signals (333). Although a combination of phorbol esters and calcium ionophore is clearly required for IL-2-dependent T cell proliferation, either one of these agents can stimulate T cell proliferation, particularly in humans, via a pathway that does not depend on IL-2 production (see Section IX).
2. Direct Measurements of PIS Turnover in T Cells Increased turnover of PISand hydrolysis of PIP2 are among the earliest biochemical changes in T h cells stimulated by mitogens (334, 335) or anti-TCR-CD3 antibodies (242), as well as in CTLs activated by their targets (336). The increase in IPScan be detected as early as 20 seconds after receptor triggering and is sustained for >30 minutes (242). A considerable proportion of IPS is converted to IP4 by a kinase (264), raising the possibility that IP4also plays a regulatory role in T cell activation. IPS and IP2 were found to be produced more rapidly than IP in TCR-CD3-stimulated Jurkat cells (242) or CTL clones (257). It is not clear whether these kinetics represent sequential phosphatase-mediated dephosphorylation steps of IPS or hydrolysis of each form of inositol phospholipids, that is, PIP2, PIP, and PI, respectively. 3 . Caz+Mobilization
Early studies measuring cell-associated 45Ca2+uptake suggested increased [Ca2+]iin mitogen-stimulated T cells (337, 338). With the development of improved methods of analysis based on the use of fluorescent
MOJ,ECUL.AK
EVENTS MEDIATING
T CELL A C T I V A T I O N
26 1
derivatives of the Ca2+chelator, EGTA (339, 340), it was found that, in resting T cells, [Ca"]; is -100 nM; a rise to -200-250 nM was seen upon stimulation with either PHA or Con A (307). A similar response has been documented in resting T cells, T cell lines, and tumors stimulated with mitogens, antigen, anti-TCR-CD3 antibodies, or antibodies to other T cell surface antigens (55, 57, 101, 116, 137-139, 242, 336, 341-345; reviewed in Refs. 346-348). T h e increase in [Ca2+]iduring T cell activation consists of two distinct phases. The addition of anti-TCR-CD3 antibodies to Jurkat cells induced a fivefold increase in [Ca2+]iwithin 60 seconds. [Ca2+]ithen fell to a plateau of 200-250 nM and remained elevated above baseline levels (80-100 nM) for >30 minutes (242). The early and transient increase is due to TCR-CD3-stimulated production of IPS, which acts as a secondary messenger to release Ca2+ from the endopalsmic reticulum. This response was mimicked by treating permeabilized cells with purified IPS and did not require extracellular Ca2+(242). Receptor cross-linking is not always necessary for intracellular Ca2+ mobilization. Monovalent Fab fragments of an antLCD3 antibody induced a similar response, albeit to a lower extent, in some (342), but not in other (346), instances. In contrast to intracellular Ca2+mobilization, the sustained increase in [Ca2+]iwas completely abolished when extracellular Ca2+was removed by EGTA o r when influx was abolished by Ca2+channel blockers (242,344, 349-35 1). Thus, in addition to mobilizing Ca2+from intracellular stores, TCR-CD3 triggering must also open a membrane Ca2+channel or regulate, in some other way, Ca2+transport across the plasma membrane (see Section VI,A). Ca2+influx seems to be an essential component of T cell activation and proliferation, since these events were inhibited in T cells cultured in a Ca2+-deficientmedium (352). Unlike the TCR-CD3 pathway, T cell triggering via CD2 does not appear to involve Ca2+mobilization from intracellular stores. Thus, the increase in [Ca2+]iinduced by a mitogenic combination of anti-CD2 antibodies (without cross-linking by a secondary antibody) was delayed relative to the anti-CD3 response, and was completely abolished by EGTA (101, 346). However, the CD2induced response was resistant to EGTA in another study (158) and was therefore considered to represent, at least in part, intracellular Ca2+ mobilization. This difference may reflect the effect of CD2 cross-linking (158). T h e sustained several-fold increase in [Ca2+]iobserved in receptortriggered T cells reflects a mean value for the T cell population. However, single-cell recordings of [Ca2+]iin activated B or T cells (353), as well as in other cell types (354), have revealed oscillations of [Ca2+]ifor a prolonged period. This repetitive rapid cycling between resting level
262
AMNON ALTMAN E T A L .
and transient peak levels in the micromolar range persisted for a few minutes, even when extracellular Ca2+ was removed by EGTA, suggesting cycles of Ca2+uptake and release by the endoplasmic reticulum. The significance of these oscillations is not clear at the moment. The increase in [Ca2+]iduring cellular activation could have multiple effects. In addition to activation of Ca2+-or CaM-dependent protein kinases or PKC (see Section V), it could also stimulate the activity of phospholipase AS. This effect could explain the release and increased metabolism of arachidonic acid in T cells (355, 356). The suppressive effects of lipoxygenase pathway inhibitors on T cell activation and proliferation (356) suggest that lipoxygenase metabolites of arachidonic acid are, in fact, important components of the signaling machinery during T cell activation. Although the increase in [Ca2+]iis an integral component in the T cell signaling machinery, this response in itself is not sufficient to trigger the whole spectrum of events associated with T cell activation, particularly IL-2-dependent proliferation. In fact, recent studies have suggested that an increase in [Ca2+]i,in the absence of additional signals and under nonmitogenic conditions, generates a tolerogenic signal that prevents T cells from responding to subsequent normal activation signals (357). Thus, treatment of antigen-specific T cell clones with antigen plus chemically fixed APCs induced unresponsiveness to further stimulation with normal unfixed APCs. This treatment was accompanied by very little inositol phosphate production, but a considerable increase in [Ca2+]i. Induction of the unresponsive state was abolished by the removal of extracellular Ca2+and, conversely, was mimicked by a calcium ionophore (357). It has been suggested that this biochemical mechanism could be the molecular basis for T cell tolerance induction in vivo, particularly during the intrathymic selection process of the T cell repertoire. 4 . TCR-CD3 Coupling to PIP2 Hydrolysis and a Ca" Response
The tight association between TCR-CD3 triggering and PIP2 hydrolysis implicates this biochemical event as a critical signaling mechanism during T cell activation. Several observations have been taken as evidence that PIP2 hydrolysis is essential in this respect. These include the ability of CAMP to uncouple PIP2 hydrolysis from TCR-CD3 triggering and to inhibit, in parallel, T cell proliferation (358), the minimal IL-2 production in a selected variant of an antigen-specific T cell hybridoma displaying deficient agonist-induced PIP2 hydrolysis (46), and the restoration of T cell proliferation by PMA plus calcium ionophore in CD4CD8- cells from mice homozygous for the lpr mutation, that display deficient PIP2 hydrolysis in response to mitogen or an antLCD3 antibody
MOLECULAR EVENTS MEDIATING
'1' CELL A C T I V A T I O N
263
(359). However, in the absence of highly specific inhibitors and wellcharacterized mutants, it has been difficult to directly address two important issues: Are PIS turnover and PIP:, hydrolysis, with the resulting rise in [Ca2]iand PKC activation, prerequisite events during T cell activation? And, if so, are these events sufficient? In fact, several recent studies indicated that, at least under certain circumstances, the answer to both questions is negative. Several examples have been documented for the uncoupling of PIS turnover and the associated increase in [Ca2+]ifrom subsequent activation events in T cells. First, mAbs to various T cell surface antigens that do not induce IL-2-dependent cellular proliferation can stimulate an increase in [Ca2+]iwhen cross-linked at the T cell surface. Such antigens include CD4, CD5, CD6, CD7, CD28, LFA-1, and others (158). Second, most immature CD4+ CD8+ thymocytes express CD3, which can function in signal transduction and mediate Ca2+mobilization; however, such cells cannot be induced to proliferate by mitogens or anti-TCRCD3 mAbs (360, 361). The use of EGTA to remove extracellular Ca2+ indicated that intracellular Ca2+ mobilization was intact in these immature thymocytes, whereas extracellular influx was markedly reduced (36 1). Third, Ca2+mobilization has been detected in TCR-CD3- antiThy-l-stimulated T cells (139) or in tolerized T cell clones (357) in the absence of cellular proliferation o r IL-2 production. Finally, monovalent Fab fragments of antLCD3 antibodies stimulated a marked increase in [Ca2+]i(342), but very little T cell proliferation (58). In all of the above examples, however, it has not been determined whether the increase in [Ca2+]iwas accompanied by PIS turnover and IPS production. The formal possibility exists, therefore, that the Ca2+mobilization documented in these studies was coupled to a signaling pathway other than PIP:, hydrolysis. In the case of immature thymocytes, a process that couples the TCR-CD3 complex to PIP2 hydrolysis could be an important step in their functional maturation. I n other instances, however, Ca2+ mobilization accompanied by IPS production was still not sufficient for IL-2 production. Using a variant of Jurkat with deficient receptor function, it was found (77) that certain anti-CD3 antibodies stimulated IPS production and a resulting rise in [Ca2+]iwithout inducing IL-2 production. However, by comparison with the parental Jurkat cells that produced IL-2, the mutant displayed diminished IP3 production and only a transient elevation of [Ca2+]i.It was suggested, therefore, that full activation of T cells (i.e., IL-2 production) may require sustained second messenger production, perhaps explaining the requirement for extended TCR-CD3 receptor occupancy in IL-2 production (56). Together, the studies described above indicate that
264
AMNON ALTMAN E T AL.
PIP2 hydrolysis and Ca2+mobilization are not sufficient per se for the complete activation of T cells. This could reflect, among other things, insufficient levels of PIP2-derived secondary messengers (77), the need for participation of additional signal transduction pathways in T cell activation, or a block at later steps. Other studies have suggested that PIP2 hydrolysis or Ca2+mobilization is not required for T cell activation. First, when T cells were triggered by PHA (229, 362) or by a single nonmitogenic anti-CD2 mAb (363) in the presence of PMA, marked proliferation and IL-2 production occurred in the absence of PISturnover and Ca2+mobilization. Thus, although PMA uncouples the TCR-CD3 (or CD2) complex from the PIS cycle (229, 364), it can clearly synergize with mitogen or antireceptor antibodies to induce complete activation of T cells. Second, target cell killing in Ca2+free medium by antigen-specific CTL clones was found to occur in the absence of detectable inositol phosphate production, including IPS(288). Another example (365) is provided by an antigen-specific T cell hybridoma variant that lacked TCR-CD3 expression. Reconstitution of receptor expression in this variant by gene transfection restored IL-2 production to normal levels, despite the fact that PIP2 hydrolysis and increased [Ca2+]iwere substantially reduced or undetectable (365). These findings raise the possibility that the production of PIP2-derived secondary messengers, usually observed soon after stimulation, are not essential events in the pathway leading to IL-2 secretion and T cell proliferation. This possibility reinforces once again the notion that alternative signaling pathways may play a causal role in T cell activation. A recent analysisof an activation-defective variant of a murine antigenspecific T cell hybridoma has, in fact, provided circumstantial evidence for the coupling of the TCR-CD3 complex to two distinct signal transduction pathways (45).This variant, EV.3, expressed normal levels of the homodimers, but was deficient in exreceptor complex containing pression of the complex having the 6-7 heterodimer (41). When stimulated with either APCs plus antigen or anti-CD3 mAbs, this variant did not produce significant levels of inositol phosphates or undergo PKCdependent CD3 phosphorylation. These responses were elicited, however, at normal levels by the G protein activator, A1F; (see Section IV), indicating that the variant possesses the intact machinery necessary for the production of PIP2-derived secondary messengers. Thus, the defect in the EV.3 variant probably lies at the level of PI-PLC coupling to the TCR-CD3 receptor complex. In contrast to the deficient PIP2 hydrolysis, however, phosphorylation of the CD3 chain in response to antigen or anti-CD3 stimulation was intact in this variant. This result suggested that the two receptor complexes on T cells that contain 6-q or 5-5 dimers
<-<
<
MOLECULAR EVENTS MEDIATING
'I' CELL ACTIVATION
265
may each be coupled to a different signaling pathway, that is, PIPz hydrolysis and activation of a putative TPK, respectively (see Sections V,B,2 and X). Moreover, experimental manipulations that uncouple PIP2 hydrolysis from the TCR-CD3 complex (i.e., CAMPor PMA treatment) did not affect the agonist-induced phosphorylation of the CD3 6 chain on tyrosine residues (358, 364). T h e existence of two structurally different antigen-receptor complexes in T cells (4 1) provides a logical and attractive model to account for at least some of the paradoxical situations described above, such as instances in which PIP2 hydrolysis is not sufficient for complete T cell activation and, conversely, experimental systems in which activation occurs in the absence of detectable PIS turnover and/or Ca2+ mobilization. However, the relative contribution of these two pathways to T cell activation and their possible interdependence remain to be established. IV. GTP-Binding Proteins in T Cell Signal Transduction
Transmembrane signal transduction pathways that utilize intracellular second messengers have several similar components. These include a ligand-specific transmembrane receptor, an intracellular effector enzyme that generates the second messenger(s), and a G protein that couples the occupied receptor to the effector enzyme. The bestcharacterized signaling systems in this regard are the receptors linked to adenylate cyclase (reviewed in Ref. 366). In this case, the hormonep, and y subunit) G protein receptor complex binds the heterotrimeric (a, through unique amino acid sequences on both the G protein and the receptor. The receptor induces the a subunit to exchange bound GDP for GTP, resulting in dissociation of the G protein subunits. The GTPbound a subunit then binds to, and stimulates or inhibits, the effector enzyme. T h e G protein possesses an intrinsic slow-rate GTPase activity; as a result it hydrolyzes bound GTP to GDP and, in the process, becomes deactivated. T h e cycle of G protein activation by nucleotide exchange continues as long as the receptor is occupied. The known G proteins having the trimeric subunit structure are the ubiquitous G, and G;, which display stimulatory or inhibitory activities toward adenylate cyclase, respectively; transducin (or GJ, a tissue-specific G protein that regulates retinal cyclic GMP phosphodiesterase; and GO, whose function is not well defined, but appears to regulate Caz+channels and perhaps other membrane conductance channels (366). Several experimental manipulations that were found to modulate the activity of the adenylate cyclase-coupled heterotrimeric G proteins have been widely used in attempts to identify a role for G proteins in coupling
266
AMNON ALTMAN ET AL.
receptors to other signaling pathways, particularly to PIP2 hydrolysis. These include the sustained activation of G proteins (and resulting biochemical events) by nonhydrolyzable GTP analogs, the ability of AlFi to mimic GTP and stimulate G proteins, and the modulation of the activity of these proteins by ADP ribosylation, a reaction catalyzed by certain bacterial toxins (e.g., cholera or pertussis and botulinum toxins) (366). A. PI-PLC COUPLING TO G PROTEINS A number of studies in recent years have suggested that receptorstimulated PIP2 hydrolysis also involves G proteins that couple the relevant receptors to PI-PLC. GTP was found to activate Ca2+-mobilizing receptors in mast cells (367) and to be essential for PI-PLC activation in various cell types (295, 368-372). Similar conclusions were drawn from the ability of AlF; to stimulate PIP2 hydrolysis (373) or the inhibition of chemotactic peptide-stimulated inositol phosphate formation by pertussis toxin (374). However, the putative G protein involved in PIP:, hydrolysis has not been defined. Attempts to identify G protein(s) essential for PI-PLC activation by susceptibility to bacterial toxins have been confusing, since these toxins have different effects on PI-PLC coupling to receptors in different cell types (375). The G proteins encoded by c-ras protooncogenes have also been implicated in the coupling of receptors to PI-PLC (375, 376). These proteins have a size of 2 1 kDa, contain GTP-binding consensus sequences, possess GTPase activity, and are localized at the inner surface of the plasma membrane by covalent modification with palmitic acid (376). Based on these properties, it has been predicted that the c-ras gene products would be components of signal transduction processes, particularly the inositol phospholipid signaling pathway. Indeed, increased PIS turnover and an enhanced production of inositol phosphates or DAG were documented in cells overexpressing normal c-ras or transformed by oncogenic ras genes (377-381). The source of increased DAG production in rastransformed cells could also be other than PIP2 (382). However, a more recent study (383) suggested that c-ras-encoded proteins do not control PI-PLC, but rather mediate their effect at a site downstream from that of PIS turnover. This conclusion was based on the finding that microinjection of a neutralizing anti-p2 1" mAb inhibited proliferation stimulated by a combination of phorbol ester plus a calcium ionophore (383). These agents duplicate the biochemical action of the PIP2-derived second messengers and therefore bypass the need for receptor-stimulated PIPLC activity and, hence, the putative G protein. Thus, although sufficient evidence has accumulated to indicate that G proteins control PI-PLC activation in various cell types, direct proof is still missing and the identity of the relevant G protein(s).is still elusive.
MOLECULAR EVENTS MEDIATING
‘I’ CELL A C T I V A T I O N
267
B. G PROTEINS IN T CELLS A number of biochemical, immunochemical, molecular, and functional studies indicate that T cells express various G proteins. Increased cAMP production in cholera toxin- or prostaglandin E-treated T cells (384-387) provides evidence for the presence of G,. However, since the predominant effect of adenylate cyclase stimulation in T cells is inhibition of cellular activation and proliferation (see Section V,A,3), it is highly unlikely that the TCR-CD3 complex or other T cell activation antigens are coupled to G,. Indeed, from most of the reports on the inhibition of T cell proliferation and signal transduction by adenylate cyclase stimulators, it was concluded that this effect is secondary to the increase in the intracellular cAMP concentration (384, 386-390). An exception to this conclusion is a study that examined the effect of cholera toxin on the activation of Jurkat cells (391). Treatment for 3 hours with the toxin inhibited IPS formation and the rise in [Ca2+];,down-regulated TCRCD3 expression, and induced ADP ribosylation of a 43-kDa substrate, a size similar to that of the a subunit of G,. However, these functional effects did not appear to result from the elevation of CAMP,since they were not mimicked by forskolin, an agent that elevates CAMP, or by 8-bromo CAMP. Thus, it was concluded that a cholera toxin substrate directly regulates signal transduction by the TCR-CD3 complex in a negative manner. Molecular cloning studies identified two Gi proteins in human T cells (392). However, hybridization studies with cDNA probes corresponding to these Gi proteins indicated that they are expressed in most cell types and are not, therefore, unique to T cells. Since Gi inhibits adenylate cyclase, thereby lowering the intracellular concentration of CAMP (which is an inhibitor of T cell activation), one could imagine a role for G; in T cell activation. Indeed, the G; activator, pertussis toxin, was mitogenic for T cells (393), stimulated PIP2 hydrolysis (390, 394, 395), and, together with PMA, induced IL-2 production by Jurkat cells (384). However, these effects of pertussis toxin were not mediated by direct ADP ribosylation of a G protein, since the B oligomer of the toxin, which does not catalyze ADP ribosylation, was fully active in the same functional assays (300,395, 396). Pertussis toxin was also found to inhibit the rise in [Ca2+]istimulated by cross-linking of various cell surface antigens on T cells (158). These findings underline the complexity of toxin effects on T (and other) cells and suggest that analysis of the functional effects of bacterial toxins, as an approach to implicate G proteins in PIP2-associated signal transduction, may be misleading. A photoaffinity labeling technique with CY-’~P-GTP, in conjunction with cholera- or pertussis toxin-mediated ADP ribosylation studies re-
268
AMNON ALTMAN E T AL.
vealed a number of G proteins in membranes of human peripheral blood T cells (397). Eight major bands of 64,52,45,41,35,31,26, and 22 kDa were labeled. Of these, the 45- and 41-kDa proteins most likely represent G, and Gi, since they served as substrates for cholera or pertussis toxins, respectively. The 41-kDa G protein is most probably not Go, since our studies, using a Go-specific anti-peptide antibody, did not detect any reactive protein in T cell membranes (K. M. Coggeshall and A. Altman, unpublished observations). Since the heterotrimeric G proteins (i.e., G,, Gi, Gt, and Go) are unlikely to play a direct role in T cell activation (i.e., by coupling to the TCR-CD3 complex), other candidate G proteins should be considered. In this respect a class of G proteins more homologous with the ras protooncogene products than with the “classical” trimeric G proteins have recently been described. Included in this group, which was identified by molecular cloning or conventional protein purification techniques, are the mammalian proteins encoded by the rho (398), rul(399), R-ras (400), rab (401), rup (402), and smg (403) genes. A 22-kDa botuh u m toxin substrate purified from human neutrophils (404)is probably the product of the rap-1 gene (402), based on its amino-terminal amino acid sequence. The common features of these proteins are their relatively small sizes (20-26 kDa) and the presence of conserved sequences thought to be involved in GTP binding. Their functions are unknown. At least the rup (402) and the rho (398) genes are expressed in T cells. Since botulinum toxin can ADP-ribosylate some of these small G proteins (404,405), we have recently begun a search for botulinum toxin substrates in human peripheral blood or leukemic (Jurkat) T cells. These studies have been aided by the use of mAb, 142-24E5,prepared against a synethetic p2 1” peptide (406), representing a relatively well-conserved sequence in some small G proteins. This mAb was found to recognize the neutrophil22-kDa botulinum toxin substrate (404). Botulinum toxin was found to ADP-ribosylate 22- and 26-kDa proteins, and mAb 142-24E5 reacted in immunoblots with proteins of 22,24,26, and 30 kDa in T cell membranes (T.Mustelin and A. Altman, unpublished observations).The 22-kDa protein recognized by the mAb and ADP-ribosylated by botulinum toxin is most likely related, if not identical, to the 22-kDa neutrophi1 G protein (rup-1) (404). Proteins of a size similar to those detected by mAb 142-24E5 were also identified by binding of radiolabeled GTP to electrophoresed T cell membrane proteins blotted on nitrocellulose (K. M. Coggeshall and A. Altman, unpublished observations).The function of these ras-related G proteins in T cells remains to be determined. The G protein activators AlFd and guanosine 5’-0-(3-thiotriphosphate) (GTPyS), a nonhydrolyzable GTP analog, have been used in
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
269
functional studies in an attempt to identify G proteins involved in T cell activation. Thus, GTPyS or AIF; was found to stimulate the activity of ornithine decarboxylase (ODC) (407, 408), an event closely associated with the activation of T and other cells (see Section VII), PIP:, hydrolysis (388, 409-412), PKC-mediated CD3 y and e chain phosphorylation (388),IL-2 production (384), and release of serine esterase-containing cytotoxic granules (4 13,414) in T cells or their membrane preparations. These functional studies support a role for G proteins in T cell activation by showing that GTPyS or AlFi stimulates the same spectrum of biological responses as those induced by TCR-CD3 triggering; however, they all fail to provide direct evidence that such G protein(s) is coupled to the TCR-CD3 complex. At present, one cannot rule out the possibility that AlFi (and perhaps also GTPyS) can stimulate PIS turnover by mechanisms totally different from the one activated by surface-active ligands. The need for such caution is evident in a recent study demonstrating that, in addition to GTP, other nucleotides also stimulated the hydrolysis of exogenous PIP:, by rabbit thymocyte membranes used as a source of PI-PLC (411). Nevertheless, the ability of GTPyS to induce cytotoxic granule exocytosis in CTL clones and, more important, abolition of the GTPyS effect in TCR-CD3-modulated CTLs, provide compelling evidence for the coupling of a G protein to the TCR-CD3 complex (413). Thus, the search for a putative G protein coupling the TCR-CD3 complex (or other T cell activation receptors) to PI-PLC is likely to be an active and intriguing area of research in future studies of signal transduction in T lymphocytes. V. Activation-Associated Phosphorylation in T lymphocytes
Protein phosphorylation is a basic mechanism for the modification of protein function in eukaryotic cells. It is mediated by a large family of protein kinases, approaching 100 members, many of which play an important role in cellular responses to external stimuli (415,416). All of the protein kinases characterized thus far fall within two broad classes, based on their substrate specificity, namely, serine/threonine specific o r tyrosine specific. Protein kinases are counteracted by protein phosphatases that dephosphorylate substrate proteins. In some cases protein kinases are integral parts of growth factor receptors and are activated directly by ligand binding. In many cases, however, protein kinases are regulated either by second messengers produced as a consequence of surface receptor occupancy or by phosphorylation as a part of protein kinase cascades. Activation of protein kinases is often accompanied by autophosphorylation. Despite the vast number of studies of protein phos-
270
A M N O N ALTMAN ET AL.
phorylation and dephosphorylation in eukaryotic cells, the exact role of these reactions in the regulation of cellular differentiation and growth, and the physiologically relevant substrates, are still poorly understood. Lymphocytes were found to possess all of the enzymatic machinery needed to phosphorylate and dephosphorylate proteins (417-419). At least four groups of protein kinases [i.e., PKC, CAMP-dependent protein kinase (PKA), TPK, and Ca2+/CaM-dependent protein kinase] participate in T cell activation and interact with each other in a complex, and as yet unknown, manner (358, 364, 420). Increased phosphorylation in lymphocytes following stimulation with T cell mitogens (421,422)or IL-2 (423) has been reported. However, the identity of the relevant protein kinases o r the phosphorylated residues (i.e., serinekhreonine versus tyrosine) has not been established in many earlier studies. Thus, in this section we primarily review only those reports that provide more specific information about phosphorylation events during T cell activation. A. SERINE/THREONINE PHOSPHORYLATION
This broad family of protein kinases is distinguished by its substrate specificity for serine and threonine residues in proteins, but members otherwise differ widely in their modes of regulation. 1. PKC
PKC was originally described as a proenzyme activated by a Ca2+dependent neutral protease or by trypsin (424); subsequently, the enzyme was found to be directly activated by Ca2+and phospholipid without proteolysis (425). PKC became the focus of attention among cellular biologists interested in signal transduction and tumorigenesis after it was discovered that it is activated by the inositol phospholipid-derived second messenger DAG (426), and that it is the cellular receptor for phorbol esters and other tumor promoters (308-3 10). Many structural and functional studies have since been conducted on this important subfamily of protein kinases (reviewed in Refs. 427-429). As a group, PKC enzymes are defined by their requirement for Ca2+, phospholipid, and a neutral lipid (e.g., DAG). DAG binding increases PKC affinity for Ca2+and phospholipid, and acidic phospholipids, particularly phosphatidylserine, are most efficient in activating the enzyme. The size of PKC is -77-80 kDa, and it consists of regulatory (=30 kDa) and catalytic ( ~ 5 kDa) 0 domains. The enzyme was found in all tissues examined and is particularly enriched in the brain. T o date, molecular cloning studies have revealed six distinct genes encoding seven forms of YKC denoted a,PI, &I, y , 6, E, and 5 (430-437; reviewed in Ref. 429). The isoenzymes differ in their affinities for Ca2+,different phospholi-
MOI.ECULAR EVENTS MEDIATINC
‘I’ C E L L ACTIVATION
27 1
pids, and neutral lipids. This, together with the distinct tissue distribution of the isoenzymes, suggests that they may be differently regulated and perhaps phosphorylate distinct substrates within the same cell or in different tissues (330,429). PKC enzymes are also regulated by autophosphorylation (438); by Ca2+-dependent proteases, termed calpains, that cleave the enzyme in the “hinge” region between the regulatory and catalytic domains (439) to generate a Ca2+- and lipid-independent constitutively active enzyme; and, finally, by an endogenous pseudosubstrate sequence that acts as a highly specific and potent PKC inhibitor (440). Following its isolation from peripheral blood lymphocytes (44 l), PKC was found to be abundant in lymphoid tissues and leukemic lines (428, 441,442). Nishizuka (443) first suggested that lymphocyte mitogens may stimulate PIS hydrolysis and DAG production, thereby activating PKC. Later, the finding that PKC is the cellular receptor for phorbol esters (308-310) shed new light on earlier reports of the mitogenic effects of these compounds on T cells (444,445)and suggested that PKC may play an important role in T cell activation. Recent studies indicated that peripheral blood T lymphocytes or leukemic T cells express at least two PKC isoenzymes, a and p (331,446-449a). Although PKCy is considered to be expressed specifically in the brain, w e found evidence for its lowlevel expression in T cells using isoenzyme-specific antipeptide antibodies and cDNA probes (449). Activation of PKC in T cells stimulated with phorbol esters, mitogens, or antLCD2 or anti-CD3 antibodies is accompanied by translocation of the enzymatic activity from the cytosol to the membrane (205, 324-329). However, a controversy exists with regard to whether IL-2 also stimulates PKC translocation in competent IL-2R-expressing T cells (see Section IX). Phorbol esters and calcium ionophores act in synergy to activate and translocate PKC in T cells (324). Although PKC stimulation is clearly associated with T cell activation (79, 80), it has become apparent that it can regulate the activation of T lymphocytes either positively o r negatively. This is consistent with the concept that PKC provides negative feedback control over various steps of the cell signaling processes (427, 429). In the case of T cells, clear examples for positive regulatory influences by PKC are the induction of IL-2R and IL-2 gene expression by phorbol esters. However, treatment of T cells with PKC-activating phorbol esters can also down-regulate the expression of various T cell surface antigens, including those participating in T cell activation-most importantly, the TCR-CD3 complex (80, 450-452), but also CD4 (80,452-455) and CD8 (80). Down-regulation of CD3 (and perhaps CD4 or CD8) may be associated with the anergy induced in T cell clones following the initial exposure to antigens, a
272
AMNON ALTMAN E T A L .
phenomenon that could represent a physiological mechanism to terminate T cell responses to antigen. In addition, treatment with phorbol esters was found to inhibit the mitogen-induced increase in PIS hydrolysis and [Ca2+]i(228,229), perhaps by uncoupling a putative G protein from the TCR-CD3 complex (413), and to limit IL-2 secretion by a murine T lymphoma line (456). In some instances, however (e.g., in PHAstimulated human peripheral blood T cells or thymocytes), treatment with PMA produced a 50% inhibition of PI and inositol phosphate labeling without altering the increase in [Ca2+]i(347). In vitro, PKC can phosphorylate multiple protein substrates from many cell types, including receptors, other membrane proteins, contractile and cytoskeletal proteins, enzymes, and others (428, 429). The list of substrates is expanding rapidly, but the physiological significance of most of these in vitro phosphorylation reactions remains to be explored. Interpretation of the results is compounded by the fact that, in most cases, the same proteins, and even the same serine/threonine residues in a given substrate, can be phosphorylated by other protein kinases (e.g., PKA). For the same reason, it has been difficult to determine the physiologically relevant substrates for PKC an vivo. Studies to determine sequences that define substrate specificity for different protein kinases (457) may be informative in this respect. Usually, PKC-mediated phosphorylation in intact cells has been studied by using PKC activators (i.e., phorbol esters and synthetic DAG) as well as PKC inhibitors. However, the most commonly used inhibitor, H-7, which was found to block T cell activation (458-460), can also inhibit other protein kinases, albeit at a lower affinity (461). Furthermore, if stimulation with mitogen or antLCD3 antibodies is found to phosphorylate the same proteins as phorbol ester induction, this does not necessarily mean that the ligand-stimulated phosphorylation is effected directly by PKC; other kinases could very well be stimulated in parallel with PKC, although PKC appears to be the most abundant serinelthreonine kinase in T cells (441). Another common approach to identifying PKC as a TCR-CD3-stimulated kinase has been to down-regulate its expression and function by chronic treatment with phorbol ester. Using this approach, it was shown that the CD3 y chain did not become phosphorylated in PKC-depleted T cells (364), although it was phosphorylated in response to antigen in untreated cells. In this respect a synthetic peptide corresponding to an endogenous PKC pseudosubstrate site, which is a highly specific PKC inhibitor (440), may be useful in pinpointing PKC substrates in vivo. Treatment of T cells with phorbol esters, synthetic DAG, or receptor (CD3 or CDZ)-binding ligands has been found to stimulate serine/
MOLECULAR EVENTS MEDIATING
T CELL A C T I V A T I O N
273
threonine phosphorylation of multiple substrates, presumably via PKC activation. Among those that are most relevant to the specific function of T cells are the CD3 y , 8, and E chains (23, 41, 113, 450, 451, 462-466); CD4 and CD8 (466-470); CD45 (47 1,472); the IL-2Ra (p55) chain (32 1, 423,472); MHC class I antigens (471,472); and pp56Ick(473). PKC may also participate in the phosphorylation of 63- and 67-kDa proteins mediated via the IL-2Rp (p75) chain in various T cell lines, since PMA or IL-2 induced a similar pattern of phosphorylation in these cells (474). Other PKC-phosphorylated membrane proteins in T cells may include CD5, CD7, CD43, and the /3 chain of LFA-1 (466), but not CD2 or CD28 (466). T h e physiological role and consequences of PKC-mediated cell surface antigen phosphorylation in T cells are largely unknown. In the case of CD3, phosphorylation may regulate the magnitude and duration of the T cell response to antigen. Initial exposure of T cell clones to antigen is known to induce an anergic state. Similarly, activation of PKC and CD3 phosphorylation induced by PMA are associated with a subsequent state of antigen unresponsiveness (320). Thus, phosphorylation may uncouple the CD3 complex from its signal transduction machinery, perhaps in analogy with the a]-(475) o r &adrenergic (476) receptor systems, in which receptor phosphorylation leads to its desensitization. Th e ribosomal S6 protein is phosphorylated in IL-2-dependent T cells following IL-2 or PMA treatment (477), probably reflecting a secondary response to the activation of S6 kinase by PKC. Recently, lamin B, a constituent protein of the nuclear lamina in most adult mammalian cells, was found to be phosphorylated in various cell lines, including T cells, in response to PMA (478). The phosphopeptide map of in vivo-phosphorylated lamin B was similar to that generated by in vitro phosphorylation with the catalytic domain of PKC, but not with PKA or Ca2+/CaMdependent protein kinase I1 (478). The ability of PKC, or a related kinase, to phosphorylate a nuclear protein may be relevant for pathways of signal transduction from the cytoplasm to the nucleus. T h e recent nuclear localization of PKC in liver cells (479) also suggests a role for this enzyme in nuclear signaling events. Another important target for PKC-mediated phosphorylation in T cells, which may participate in signal transduction from the cytoplasm to the nucleus (480), may be the serine/threonine protein kinase, Raf- 1, encoded by the Raf-1 protooncogene. Phorbol ester treatment was found to induce phosphorylation and activation of Raf-1 (481), and T cell stimulation was recently found to be associated with Raf- 1 phosphorylation (482). T h e ability of PKC to phosphorylate in vitro guanylate cyclase (483) or the catalytic subunit of adenylate cyclase (484), coupled with the finding that PMA or IL-2 can inhibit adenylate cyclase in IL-2-dependent
274
AMNON ALTMAN ET A L .
T cells (485), implies that T cell activation may be regulated in part by "cross-talk" among PKC and other kinases. The phosphorylation of the T cell-specific TPK, pp56Ick,in response to PKC-activating phorbol esters (473), is another example of interactions among different kinases in T cells (Fig. 5). The known heterogeneity of PKC isoenzymes raises the interesting possibility that distinct PKC isoenzymes may selectively participate in different phosphorylation events in T cells. There could be differences in their intracellular location and their activation requirements. Studies designed to overexpress certain PKC isozymes or, conversely, downregulate their expression in a selective manner could address this question.
2. Ca'+/CaM-Dependent Kinases Mammalian cells express several Ca2+/CaM-dependent protein kinases (486). CaM is a ubiquitous calcium-binding protein that regulates a variety of cellular processes. It has been implicated in lymphocyte responses to mitogenic stimuli, and its content was found to increase after mitogenic stimulation of human lymphocytes (487). Lymphocytes, including T cells and thymocytes, express several CaM-binding proteins, one of which is a 59-kDa subunit of a CaM-dependent phosphatase (488, 489). The function of Ca2+/CaM protein kinases in lymphocyte activation has been assessed by using Cd2+(490) or phenothiazine drugs (491) that act as CaM antagonists. Such drugs have been found to inhibit T cell mitogenesis induced by lectins (492) or by PMA (324). However, phenothiazines also inhibit PKC by competing with the phospholipid cofactor (493), thereby complicating interpretation of these results. Thus, phenothiazines probably inhibit the function of several kinases when T cells are stimulated with Ca2+-mobilizingligands that are likely to activate both PKC and Ca2+/CaMkinases. However, in the case of PMA-induced T cell proliferation, the inhibitory effect of CaM antagonists probably does not involve Ca2+/CaMkinases, since PMA stimulation is not accompanied by a rise in [Ca2+];(324). In leukemic Jurkat cells the calcium ionophore, ionomycin, was found to stimulate phosphorylation of two cytoplasmic proteins with sizes of 2 1,000 and 23,000 kDa that were also phosphorylated in response to an anti-TCR antibody (343). Ionomycin treatment also caused increased phosphorylation in human peripheral blood T cells (422), including phosphorylation of the CD3 y chain in quiescent human T cells (420). Unlike PMA, which induced phosphorylation of a single serine residue (Ser-126), ionomycin led to the phosphorylation of two residues (Ser- 123
MOLECULAR EVENTS MEDIATING
'I' CELL ACTIVATION
275
and Ser-126). Since ionomycin can induce a state of antigen unresponsiveness in T cell clones (357), it has been suggested (420) that this phosphorylation event may uncouple the CD3 complex, resulting in functional desensitization. Ionomycin treatment of murine T cells also caused serine phosphorylation of the Raf- 1 kinase, which was distinct from the pattern of phosphorylation stimulated by PMA, CAMP,or an anti-Thy- 1 antibody (482). PKC depletion by prolonged PMA treatment did not affect the ionomycin response, indicating that PKC was not involved. Based on the finding that, at the concentration range commonly used (200-300 nM), ionomycin alone did not activate PKC in intact T o r other cells (324,494),it would appear that ionomycin-induced protein phosphorylation is mediated by some Ca2+/CaMkinase(s), not by PKC. 3. Cyclic Nucleotide-Dependent Protein Kinases
Many studies examined the effects of the cyclic nucleotides, cAMP or cGMP, as well as manipulations that stimulate the corresponding cyclic nucleotide-dependent protein kinases, PKA and PKG, respectively, on T cell activation (reviewed in Refs. 495 and 496). T h e observed effects are presumed to reflect the activation of PKA o r PKG, leading to phosphorylation of their respective physiological substrates in T cells. However, relatively few studies directly examined PKA- or PKG-mediated phosphorylation events in T cells. CAMP-dependent or -independent protein kinases were identified in unseparated lymphocyte populations. Types I and I1 PKAs are localized in the cytoplasm and the inner surface of the plasma membrane in T cells, respectively (495). Earlier reports that the type I PKA plays a role in mitogen-induced T cell activation (497) were not confirmed. Pretreatment of human T cells with prostaglandin E l , a CAMP-inducing ligand, inhibited Con A-induced phosphorylation (422). This and other similar findings reaffirm the general concept that cAMP generates an inhibitory signal for T cell activation and proliferation (495), presumably resulting from the activation of PKA. With respect to PKA substrates in T cells, the treatment of T cells with cAMP agonists caused phosphorylation patterns different from those induced by mitogens, antireceptor antibodies, PMA, or ionomycin (482, 495, 498). CAMP-dependent phosphorylation of 17.5-, 23- to 25-, and 33.5-kDa proteins was documented in human T cells (499). This phosphorylation was sensitive to the cyclic nucleotidedependent protein kinase inhibitor, H-8. Treatment of T cells with adenylate cyclase activators o r cAMP analogs can stimulate some aspects of T cell activation [e.g., induction of IL-2Ra chain transcription and cell surface expression (233, 500-502) J and
276
AMNON ALTMAN ET AL.
protooncogene (e.g., c-fos and C-myb) transcription, which does not always correlate with translation (501). CAMP-dependent mechanisms may also play an important role in the aggregation and capping of CD3, CD4, and CD8 (499), since treatment of T cells with the PKA inhibitor, H-8 (461), inhibited these events as well as protein phosphorylation in intact human T cells (499). For the most part, however, manipulations leading to an increased intracellular cAMP concentration inhibited T cell activation at several distinct levels. This included inhibition of PIShydrolysis (358, 364); CD3 y subunit (serine) and CD3 6 subunit (tyrosine) phosphorylation (358,364, 388); lymphokine production and secretion (385,503,504);proliferation (73, 386, 505); and CTL activity (73, 387, 389) induced by TCR-CD3binding ligands. It is of interest that cAMP inhibited only the antigeninduced, but not the anti-CD3 or anti-Thy- l-stimulated, tyrosine phosphorylation of 6 (358). The CAMP-mediated inhibition of the increase in [Ca2+]iduring T cell activation (387) may be due not only to the blockade of PIShydrolysis, but, in addition, to PKA-mediated phosphorylation of Ca2+ channels (506) or a direct effect on the release of Ca2+ from intracellular stores resulting from PKA-dependent phosphorylation of the IPSreceptor (507). Based on the ability of the G protein activator, AlFi, to bypass CAMPmediated inhibition of T cell activation, it was suggested that PKA interferes primarily with PIShydrolysis and formation of the relevant second messengers (388). However, the ability of cAMP to suppress even PMA plus calcium ionophore-induced T cell activation (73,389) suggests additional downstream sites of inhibition. In addition, cAMP interfered with IL-2-induced T cell proliferation at the early GI phase of the cell cycle (73,485, 508). Treatment with a PKC-activating phorbol ester reconstituted serine and tyrosine phosphorylation of the CD3 y and 6 chains, respectively, in the presence of cAMP (364) and also overcame the CAMP-dependent inhibition of IL-2-induced proliferation (485). These findings suggest that PKC activation is dominant in the antagonism between PKA and PKC, perhaps as a result of adenylate cyclase phosphorylation (484). The complexity of the regulatory effects of PKA on T cell activation is also implied by the findings that CD3 6 subunit tyrosine phosphorylation by antigen, but not by antLCD3 or anti-Thy-1 antibodies, was sensitive to CAMP. In contrast, CD3 y subunit serine phosphorylation by all three ligands was sensitive to the same treatment (358). Several studies also examined the effects of the cGMP-dependent protein kinases, PKG, on T cell activation and proliferation (reviewed in Ref. 496). In general, cGMP has been considered to positively affect T
MOLECULAK EVENTS MEDIATING
T CEL.L
ACTIVATION
277
cell responses. T cell mitogens were found to activate guanylate cyclase and to increase intracellular cGMP levels, and products of the lipoxygenase pathway may play a role in these events (496). However, the substrates of PKG in T cells are unknown. B. TYROSINE PHOSPHORYLATION 1 . TPKs an T Lymphocytes
Protein phosporylation in tyrosine residues plays an important role in regulating normal growth and malignant transformation. However, little is known about its role in T lymphocytes. Several TPKs are expressed in lymphoid cells (Table 11). The TPK encoded by the lck protooncogene, pp56Ick(Fig. 3), is expressed at high levels only in T lymphocytes (509, 510). The lck gene is located on chromosome 1 at a site (lp32-35) where frequent chromosome abnormalities have been noted in T cell lymphomas (511).An active
TABLE I1 EXPRESSION OF TYROSINE PROTEIN KINASESIN T CELLSAND LYMPHOID TISSUES" Gene STC
Subfamily
Protein
59 56 lck tkl 52' abl Subfamily c-abl 120- 145 Insulin receptor subfamily INS-R p 90 ltk 52' PDGF receptor subfamily c-kit 145 ret 96 Others c-pim-1 p33P'" ? p3OC ? P40 ? p50d
fm
Expression in lymphoid cells
Reference
T cells T cells Spleen
187 509,510 52 1
Spleen, thymus
525
T cells Thymus, LAK cells
523 524
Spleen T cell lymphoma
526 527
T cells LSTRA T lymphoma Thymus Spleen
528 53 1 532 533 ~~~~
'
a
LAK, Lymphokine-activated killer; PDGF, platelet-derived growth factor. Based on predicted amino acid sequence. Possibly p33P'". Possibly tkl.
~
278
-AMNON ALTMAN E T AL.
Regulatory (7) Domain
1-H,N
Catalytic Domain
COOH-509
U
Unique to pp56ICk
0 ATP @ @ @ @
binding (GXGXFG)
Phosphotransfer reaction (K) Indicator of tyrosine specificity (DLRAAN) Tyr-394, autophosphorylation site Tyr-505, negative regulation
FIG.3. Structure of ~ ~ 5 6 'The ' ~ . lysine at site 2 is involved in transferring the y-phosphate of bound ATP to the substrate protein. The sequence at site 3 seems to distinguish tyrosine-specific protein kinases from those specific for serine/ threonine. A, Alanine; D, aspartic acid; F, phenylalanine; G, glycine; K, lysine; L, leucine; N, asparagine; R, arginine; Y, tyrosine; X, any amino acid.
autophosphorylating 58-kDa protein, perhaps identical to pp56Ick or ~ ~ 5 (encoded 7 ~ ' by ~ the closely related hcrk protooncogene), was identified in an acute myeloblastic leukemia (512). pp561Ckis equally abundant in CD4' and CD8' cells (187) and is located at the inner surface of the plasma membrane. The covalent attachment of myristic acid to the NH2 terminus of pp56Ickis probably important for its membrane localization (5 13,514). Recently, pp56Ickwas found to be associated with the cytoplasmic domain of CD4 or CD8 (186, 187). This was shown by antibody-induced comodulation and coimmunoprecipitation. Furthermore, cross-linking of CD4 by anti-CD4 antibodies plus secondary antibodies caused a severalfold increase in pp56Ick activity within minutes (189). This suggests that ligand binding to, or oligomerization of, CD4 could result in signal transduction to the interior of the cell by inducing tyrosine phosphorylation. Under physiological conditions of T lymphocyte activation by APCs, CD4 (or CD8) is thought to associate with the TCR-CD3 complex (166) by binding to invariable parts of MHC molecules (162, 165, 176). This brings pp56Ickto the immediate vicinity of the TCR-CD3 complex, through which the activation signal is transduced (Fig. 1). Thus, pp56ICkprobably plays an important role in T cell activation (188). In addition, activation of T lymphocytes with Con A plus PMA caused down-regulation of pp56ICkprotein
MOLECULAR EVENTS M E D I A T I N G
T CELL ACTIVATION
279
and lck mRNA (515). This suppression did not correlate with proliferation, but was maximal under conditions optimal for IL-2 and y-interferon production. In addition, ligands acting through the TCRCD3 complex or PKC-activating phorbol esters caused phosphorylation of pp56Ickat multiple serine residues in the amino-terminal region, resulting in altered electrophoretic mobility of pp56ICk,which now migrates, with an apparent size of 59 kDa (473, 5 16,517). In its active form pp56Ickautophosphorylates a tyrosine residue (Tyr394) corresponding to the major autophosphorylation site in all other TPKs (473,509,510,518).In cells this tyrosine residue is not phosphorylated (473), suggesting that pp56lckis, to a large extent, inactive in uivo. Instead, a carboxy-terminal tyrosine (Tyr-505) is highly phosphorylated in vivo by another TPK (473), and this is thought to result in suppression of the activity of pp56Ick(Fig. 3). The importance of Tyr-505 and the potential role of pp56Ickin the regulation of cell growth was perhaps best demonstrated by Marth et ul. (519) and by Amrein and Sefton (520), who transfected fibroblasts with a point-mutated form of lck, in which a phenylalanine residue substituted for Tyr-505. This mutation resulted in malignant transformation of the cells, and the pp56Ickin the transferants was catalytically much more active than in cells transfected with wild-type lck. These findings also indicate that crucial substrates for pp56Ickare present in fibroblasts. T h e role of pp56lck in T cell physiology is still unclear. T h e recent finding of its association with CD4 and CD8 (186, 187) is likely to advance our understanding of its role considerably in the near future. Identification of the TPK phosphorylating Tyr-505, the putative phosphatase counteracting it (see Section V,C), and relevant substrates for pp56lckis an important step in this direction (Fig. 4). T lymphocytes also express the TPKs encoded by the fm (187), and possibly tkl(521), protooncogenes. The latter was cloned from the spleen and encodes a putative TPK of about 52 kDa. The protein encoded by the fm gene, pp5gfY", is 59-60 kDa and differs from pp56Ickmainly in its amino-terminal region. It is also located at the plasma membrane and has a carboxy-terminal tyrosine (i.e., Tyr-53 1) analogous to Tyr-505 in ~ ~ 5 6 'As ' ~in. the case of the lck gene,fyn becomes acutely transforming when the codon for this carboxy-terminal tyrosine is altered (522). T h e fyn gene is expressed in a wide variety of cells, in addition to T cells, and its physiological function is unknown. It is not known whether pp5gfY" is associated with a surface glycoprotein in a manner analogous to CD4 or CD8 association with pp56ICk. Resting T lymphocytes express a low number of insulin receptors, which increases during T cell activation (523). Recently, a novel gene, ltk,
280
AMNON ALTMAN E T AL.
FIG.4. Regulation of pp56lCkcatalytic activity. The kinase is activated by the CD45 phosphotyrosine phosphatase (PTPase) and inactivated by a putative cytoplasmic TPK. Both events are thought to involve Tyr-505 of pp56"'. L, Putative ligand of CD45.
encoding a putative TPK, was cloned from a pre-B cell line (524). This TPK is related to the catalytic p chain of the insulin receptor and is expressed at high levels in the thymus, suggesting a possible role in T cell maturation. T cells also express the TPKs encoded by the c-abl(525), c-kit (526), ret (527), and pim-I (528) genes. The protein encoded by pim-I, p33P'", is a small cytosolic TPK (529) implicated in the development of T cell lymphomas (530). Its expression increases during T cell mitogenesis (528). Several less well-characterized TPKs have been purified from, or identified in, lymphoid cells, including a 30-kDa TPK from LSTRA T lymphoma cells (53 l ) , a 40-kDa enzyme from the thymus (532),and a 50-kDa kinase from the spleen (533). In conclusion, at least 12 different TPKs are present in T lymphocytes (Table 11). The reasons for the expression of many TPKs in a single cell type are not immediately obvious. Presumably, distinct TPKs are differentially regulated (e.g., pp56lCkby CD4 or CD8 and the insulin receptor by insulin), and they might have.different substrates in viva Different stimuli may also cause tyrosine phosphorylation of identical or overlapping sets of substrate proteins via the different kinases. In order to address these possibilities, it will be necessary to identify the physiologically relevant TPK substrates in T cells.
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
28 1
2. Substrates for TPKs an T Cells Several phosphotyrosine-containing proteins have been detected in T lymphocytes (534),and some reports indicate that changes in the pattern of tyrosine phosphorylation of these endogenous substrates occur during T cell activation. A 42-kDa protein in human T cells was rapidly phosphorylated at both tyrosine and serine residues following the addition of Con A, PHA, or anti-CD3 mAbs (535). This rapid response reached a maximum within 2 minutes. In another study a 66-kDa protein was phosphorylated on tyrosine residues in PHA- or Con A-stimulated human peripheral blood lymphocytes (536). Phosphorylation was observed as early as 2 minutes after mitogen stimulation and reached a maximum in 2 hours. On the other hand, Piga et al. (537) found no changes in tyrosine phosphorylation of T cell proteins during the first 24 hours after mitogen stimulation. Decreased phosphorylation of a 38-kDa protein and increased labeling of a 32-kDa protein were observed later. Many of the discrepancies in these (and other) studies stem from the difficulties in identifying phosphotyrosine-containing proteins. These have been detected by two alternative techniques. The first approach utilizes phosphoamino acid analysis to identify phosphotyrosine in 32Plabeled proteins, often enriched by treatment of the samples with strong alkali, based on the relative resistance of phosphotyrosine (compared to phosphothreonine and, in particular, phosphoserine) to dephosphorylation by 1M KOH (538).The major disadvantage of this method is the low phosphotyrosine content compared to the two other phosphoamino acids in normal cells. In addition, some phosphotyrosine-containing proteins might have slow phosphate turnover, resulting in poor labeling. The other technique is based on the immunological detection of phosphotyrosine-containing proteins, using antibodies specific for this phosphoamino acid (539,540). The use of such antibodies requires rigorous demonstration of specificity and is limited in its ability to detect minute quantities of phosphotyrosine. Treatment of T lymphocytes with anti-TCR-CD3, but not anti-CD2, antibodies resulted in rapid tyrosine phosphorylation of the 5 chain of TCR-CD3 (42,43,358,364). This 5 phosphorylation was not induced by phorbol esters, was intact in cells chronically treated with phorbol esters to deplete PKC, and was not effected by AlF; treatment, which directly stimulates PIP2 hydrolysis, presumably by the activation of G proteins(s) (388, 412). This indicates that triggering of the TCR-CD3 complex activates a cellular TPK by a mechanism independent of PIS turnover and PKC or, alternatively, that stimulation of TCR-CD3 causes a conformational change in the 5 subunit that facilitates its phosphorylation by a
282
A M N O N ALTMAN E T AL.
constitutively active TPK. The recent finding that pp56Ickis associated with CD4 and CD8 makes this kinase a good candidate for the TPK phosphorylating the ( chain, due to the association of CD4 or CD8 with the TCR-CD3 complex during antigen presentation, in association with MHC molecules (188; see Fig. 1). This proposition is supported by the recent finding (189) that cross-linking of CD4 activates pp56lckand results in rapid tyrosine phosphorylation of the ( chain. The role of ( chain tyrosine phosphorylation is unknown, but an interesting possibility arises from the observation that the T cells of mice homozygous for the lpr or gld mutations displayed constitutive tyrosine phosphorylation of ( (462) and increased levels of Ick mRNA (359). These mice suffer from a variety of autoimmune manifestations and an abnormal peripheral expansion of a normally minor population of CD4CD8-T cells that express normal or near-normal amounts of TCR-CD3. These T cells, nevertheless, are refractory to stimulation by mitogenic lectins or anti-TCR-CD3 antibodies in terms of their PIP2 hydrolysis and proliferative activity (359). They proliferate, however, in response to phorbol ester plus calcium ionophore (359) and undergo PIS turnover in response to A1F; (412). These observations indicate that, in lpr T cells, signal transduction from the TCR-CD3 complex is blocked proximally to the putative G protein regulating PI-PLC. Whether the tyrosine phosphorylation of the ( chain is the cause of this uncoupling or is related to the abnormal expansion of the cells in vim remains to be determined. It is noteworthy that the lpr or gld T cells lack both CD4 and CD8, which normally associate with, and probably regulate, pp56lck. The phospholipase inhibitor, lipomodulin, was phosphorylated on tyrosine in thymocytes stimulated with Con A, PMA, or calcium ionophore (541). The TPK mediating this event was not identified, but it was positively or negatively regulated by PKC or PKA, respectively. Since lipomodulin phosphorylation down-regulates its inhibitory activity toward phospholipases, this event may be important in T cell activation. A serinelthreonine protein kinase encoded by the Raf-1 protooncogene is an important target for several TPKs (481). This cytosolic kinase is believed to participate in signal transduction from the plasma membrane to the nucleus (480). It is not yet known whether TPKs in T lymphocytes phosphorylate the Raf- 1 kinase. It is clear that several protein kinases participate in regulating discrete steps in T cell activation and proliferation. These kinases appear to phosphorylate multiple, mostly unknown, substrates and also to interact with each other in complex ways (Fig. 5). The difficulties in assessing the role of each kinase in the absence of contribution by other kinases and the meager knowledge of the relevant physiological substrates will make
MOLECULAR EVENTS MEDIATING
‘1‘ CELL ACTIVATION
283
progress in this area slow and laborious. Isolation of protein kinase mutants and development of specific inhibitors will be extremely useful in this respect. C. PROTEIN PHOSPHATASES Protein phosphatases play an important role in determining the extent and duration of protein phosphorylation by reversing the action of protein kinases. In addition, several studies have indicated that protein phosphatases do not merely play a “passive” role to reverse the action of kinases, but are actually the target of hormonal control. For example, insulin or glucagon treatment in vivo was reported to induce acute changes in the activity of hepatic phosphorylase phosphatase (542). Compared to protein kinases, little is known about the function of phosphatases during T cell activation. However, the temporary nature of phosphorylation events in T cells suggests an active role for phosphatases (543). Human T cells were recently found to contain three serine/ threonine phosphatases i.e., types 1, 2A, and 2B (calcineurin) phosphatases (544)l. These enzymes are present in both the cytosolic and the particulate cellular compartments, but little is known about their substrates in viva Recently, the CD45 (T200) leukocyte common antigen, expressed at high levels on all leukocytes (545),was found to have extensive homology, in two homologous intracellular domains, to a major phosphotyrosine phosphatase purified and cloned from the human placenta (546). It was also directly shown that CD45 is indeed a phosphotyrosine phosphatase (547,547a).The finding that cross-linkingof CD45 to other T cell surface molecules can either inhibit or enhance T cell activation (548) lends support to the notion that the CD45 phosphatase may play an important part in regulating T cell activation. We have recently found that CD45 comprises >90% of membrane-bound phosphotyrosine phosphatase activity in T cells. Moreover, CD45 was capable of activating pp56Ick,apparently by dephosphorylating a regulatory tyrosine residue, Tyr-505 (549). Thus, pp56ICkactivation was absent in a CD45- mutant and, conversely, addition of a CD45 immunoprecipitate with demonstrated phosphotyrosine phosphatase activity to isolated pp56Ickactivated the latter, as indicated by its increased autophosphorylation and activity toward an exogenous substrate (549). The conclusion that pp56lckis a substrate for CD45 was reached in another recent study (549a), by showing that, in CD45‘ mutant T cell lymphoma lines, the phosphorylation of Tyr-505 on pp56lckis considerably higher than in the corresponding wild-type CD45’ cells. Furthermore, CD45 appears to be essential for T cell activation by antigen, as
A
Antigen + MHC
3
c-
PTPase
turnover
FIG. 5. Protein kinases and phosphatases immediately involved in signal transduction in T cells. (A) Regulation during signal input. Antigen-MHC binding to the TCR-CD3 complex and CD4 or CD8 stimulates PI turnover and brings pp56"' (p56) to the vicinity of TCR-CD3 complex (including the f chain). PIP2derived secondary messen ers activate protein kinase C (PKC) and, via an increase in [Ca*']i and Ca2 /calmodulin-dependent kinases (CaPK). Unknown ligands may activate or redistribute CD45, which then serves to activate pp56lckby virtue of its phosphotyrosine phosphatase (PTPase) activity. (B) The three protein kinases phosphorylate TCR-CD3 subunits; PKC and CaPK phosphory, 6 chain. These phosphorylation reactions could be late CD3 y , and p ~ 5 6 " ~the
Y
MOLECULAK EVENTS MEDIATING
1'CELL A C T I V A T I O N
285
demonstrated in a recent study in which a CD45- mutant of an antigenspecific T cell clone failed to respond to antigen with proliferation (549b). Response to antigen was regained in a revertant CD45' clone. Since CD45 is a putative receptor, the recent findings regarding its phosphatase activity suggest the possibility of a transmembrane signaling pathway in the form of a ligand-stimulated receptor phosphotyrosine phosphatase (54%). Understanding the exact role of the CD45 phosphatase in T cell activation will certainly benefit from the identification of its putative ligand. VI. Ion Channels in T Cells
Lymphocytes maintain a tightly regulated intracellular ionic milieu by virtue of ion channels and ATP-dependent ion pumps (Fig. 6). Together, these create large gradients between extra- and intracellular ion concentrations, which result in a membrane potential of 50-60 mV in T cells (550, 551). The concentration gradient across the plasma membrane is particularly large for Ca2+,of approximately four orders of magnitude, with an intracellular concentration of free calcium as low as 0.1 fl (3 0 7 ). A. Ca2+CHANNELS Mitogenic activation of T lymphocytes by antigen (34 l), lectins (307, 552), or anti-CD3 mAbs (55,342),is associated with a rapid and sustained increase in [Ca2+]i. Activation through CD2 also increases [Ca2+]iin CD3- T cells (101, 116). Calcium ionophores alone are mitogenic for some subpopulations of T lymphocytes (553,554). However, some T cell ligands increase [Ca" 1; without inducing mitogenesis. Conversely, T cell proliferation in the absence of increased [Ca2+]ihas also been reported (see Section III,C,4). As noted earlier (Section III,C,3), the rise in [Ca2+]iin stimulated T cells is initially caused by an IPS-mediated release of Ca2+ from the endoplasmic reticulum of a subcompartment thereof (242). This conclusion was drawn from the ability of IPS to release Ca2+in permeabilized cells (242,261) and the insensitivity of the initial rise in [Ca2+]ito removal
involved in positive or negative feedback regulation of CD3 function and cellular activation. (C) Cross-talk among the different kinases. pp56"' (p56) serves as a substrate for PKC (unknown consequences), CaPK (unknown consequences), a TPK hosphorylating Tyr-505 (505-TPK, which suppresses the activity of ~p56~''). and CD45 (activating pp56"'). PKC also phosphorylates the CD45 PTPase (unknown consequences) and CD4 (not shown).
286
AMNON ALTMAN E T A L .
/ ca2*
Na+
\ PI
turnover
FIG. 6. Regulation of ion channels during T cell activation. TCR-CD3induced PI turnover produces DAG and IPS.DAG activates PKC which, in turn, stimulates a Na' /H' antiporter (I), exchanging intracellular H + for extracellular Na'. The formed IPS liberates Ca2+from the endoplasmic reticulum (ER) and opens (perhaps in the form of IP4) voltage-sensitive Ca2+ channels (2) in the plasma membrane. The increase in [Ca2+]~, resulting from influx and intracellular mobilization, synergizes with PKC to stimulate the Na+/H+ antiporter (l), closes the Ca2+ channels (2), and opens potassium channels (3), causing K+ efflux.
of external Ca2+by EGTA (242). The initial rapid mobilization of internal CaZt is followed by a sustained influx of Ca2' from the external milieu (307, 552; reviewed in Ref. 347), which persists for over 24 hours (555). Ca2+channels regulated by membrane potential are found on excitable cell types, such as neurons and muscle cells. These channels are opened by depolarization (action potential) of the membrane, generally caused by increased sodium permeability. When triggered, these Ca2+ channels mediate an influx of Ca2+ down its concentration gradient, resulting in a rapid increase in [Ca2+]i.Attempts to detect this type of Ca2+channel in T lymphocytes have failed (347). Recently, a novel type of Ca2+channel was identified in T cell plasma membrane by the patchclamp technique (556). This sensitive technique allows recording of single-channel events and manipulation of the medium on both sides of the membrane. The Ca2+channels in T cells are voltage insensitive (556), that is, they do not open in response to changes in membrane potential. They have a conductance of 7 pS, and are mostly closed in resting T cells. When T cells are stimulated by PHA (556) or anti-CD3 (103) or antLCD2
MOLECULAR EVENTS MEDIATING
'I'
CELL A C T I V A T I O N
287
mAbs (103), a marked increase was seen in the opening pattern of these Ca2+ channels, resulting in an inward Ca2+ current. Signals through TCR-CD3 or CD2 activated an identical set of Ca2+channels physically separated from both receptor complexes ( 103). The secondary messenger mediating the signals for Ca2+channel opening from TCR-CD3 or CD2 appears to be IP3 (557), IP4 (262, 558), or a combination of both (559). These Ca2+ channels are autoregulated by Ca2+; when [Ca2+]; reaches a concentration of = lo6 M,almost all channels are closed (103). This negative feedback regulation is common to most Ca2+ channel types. The role of these channels and their mode of regulation in T cells await further characterization. B. POTASSIUM CHANNELS Treatment of T cells with calcium ionophores or other agents that increase [Ca2+]i(e.g., PHA and Con A), induces a shift of the membrane potential from about -50 to -70 mV (307). This hyperpolarization is caused by K+ efflux through voltage-sensitive K+ channels (560-562). A causal relationship between [Ca2+]iand K+ efflux is also suggested by their similar time courses. A number of studies indicate that K+ conductance is not required for Ca2+channel gating (374, 560, 563, 564) and might even be inhibitory (560). Instead, Ca2+ stimulates K+ channel opening (307, 564). An inhibition of K+ channels by a rise in [Ca2+]ihas also been reported (565). Three different types of K + channels, named n, n', and 1, have been identified in T cells, based on differences in their conductance, kinetics, and sensitivity to K + channel blockers (566, 567). Interestingly, the expression of the different types of K+ channel changed markedly during intrathymic T cell development (567). Immature CD4- CD8- thymocytes have only n-type channels, and their number increased during transition to the CD4+ CD8+ phenotype. Among single-positive thymocytes, the CD4+ CD8- cells expressed low numbers (=20 per cell) of n-type channels only, while the majority of the CD4- CD8+ cells had low numbers of both n' and 1 channels (but no n channels), with n' channels predominating in one subset and 1 channels predominating in another subset. In contrast, the abnormal CD4-CD8-T cells expanding in mice homozygous for the lpr mutation expressed a high number (= 220 per cell) of 1-type K+ channels (568, 569). The significance of this abundant expression is unknown. The importance of K+ channels and hyperpolarization in T cell activation is unclear. Inhibition of K+ conductance by compounds such as tetraethylammonium, 4-aminopyridine, or quinine blocked T cell mitogenesis (570), although the specificity of these compounds at the rela-
288
A M N O N ALTMAN ET AL.
tively high concentrations used in some of these studies has been questioned (563, 571). C. Na+/H+ANTIPORTER AND INTRACELLULAR pH Since pH is a critical factor in most, if not all, biochemical reactions, even small changes in intracellular pH (pHi), would be expected to profoundly affect many cellular functions. Thus, the finding that treatment of cells with mitogens or phorbol esters rapidly increases pHi by 0.1-0.2 units (572, 573) stimulated considerable interest in the role of pHi in cellular activation. This alkalinization is caused by an increased exchange of intracellular H+ for extracellular Na', mediated by a unique transmembrane protein, the Na+/H+antiporter. The increase in intracellular Na+ causes a decrease in membrane potential that is counteracted by K+ efflux. Na+ is extruded by a ouabain-sensitive Na+/K+ ATPase, which is activated secondarily to the increased intracellular Na'. The treatment of T lymphocytes with mitogenic lectins, anti-TCRCD3 mAbs, IL-2, phorbol esters or calcium ionophores caused a rapid and sustained increase of about 0.15 units in pHi (349, 574-576). This response was dependent on extracellular Na+ and was inhibited by amiloride or its analogs. In contrast, others have found no changes in pHi in mitogen-stimulated T cells (577,578). The effect of PMA was independent of Ca2+,while the effect of the calcium ionophore was obviously dependent on Ca2+,indicating two different pathways for activation of the Na+/H+antiporter. In addition, agents that stimulate TPK further potentiated the effects of PMA in fibroblasts (349), indicating a third pathway (579). The importance of mitogen-induced intracellular alkalinization remains unclear. In fibroblasts blockade of the Na+/H+ antiporter prevented mitogenesis (580), at least at neutral or acidic pH in the medium (581). In T lymphocytes, however, analogs of amiloride that completely block the rise in pHi had little effect on TCR-CD3- or IL-2-mediated proliferation (575, 582, 583), suggesting that intracellular alkalinization is not a prerequisite for T cell proliferation. The recent molecular cloning of a human cDNA encoding the amiloride-sensitive Na+/H+antiporter, an 894-amino-acid protein (584), should facilitate studies on the role of this protein in regulating cellular activation and growth. VII. Polyamine Metabolism
The polyamines putrescine, spermidine, and spermine are organic cations required for many growth-related cellular functions. They are essential fbr DNA synthesis (585),RNA and protein syntheses (586,587),
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
289
chromosomal integrity (588,589),and normal structure of microtubules and actin filaments (590). I n vitro, they also affect microtubule-associated protein kinases (591), stimulate phosphorylation of PIS(592) and nuclear proteins (593), and activate some phosphotyrosine phosphatases (593a). Polyamines can be used as substrates by transglutaminases (594), which are activated within 30 minutes following stimulation of T lymphocytes with mitogens (595). In all of the cases studied, cellular proliferation is associated with an increased turnover of polyamines (596). During growth onset, dramatic increases occur in the activities of ODC, the rate-limiting enzyme in polyamine biosynthesis (597), and S-adenosylmethionine decarboxylase, which is crucial for synthesis of the higher polyamines, spermidine and spermine. Stimulation of arginase, the enzyme that produces L-ornithine, was also seen during T lymphocyte activation (598). ODC activity correlates well with the growth rate of cells. Malignant transformation by oncogenes such as v-src (599) or Ha-rasVa'-'' (600) is accompanied by deregulation of ODC and a marked increase in its activity. The highest activity is usually seen in the late GI phase of the cell cycle, and a second, smaller, peak occurs in G2. ODC activity has an unusually short half-life in vivo (601), which is commonly thought to reflect a rapid degradation of the ODC protein (597). This short half-life correlates rather well with the amount of immunoreactive ODC in the cytosol(602), but direct evidence of a rapid ODC degradation (such as identification of proteolytic fragments) is still lacking. A specific protein inhibitor, the antizyme, has been isolated and characterized (603,604).It is a protein of apparently 19-22 kDa and has a high affinity for ODC in vitro. Indeed, ODC-antizyme complexes can be detected in vivo (605,606),but the fate of ODC in these complexes is unclear. Thus, the function of the antizyme, other than binding ODC, remains uncertain. The translation of ODC mRNA into protein is negatively regulated by polyamines (607,608). This negative feedback probably operates by the binding of some polyamine-regulated inhibitor of the mRNA (607). T h e activity of ODC in resting human T lymphocytes is low, and often undetectable (609). When cells were stimulated to proliferate by mitogenic lectins or anti-CD3 mAbs, a two-stage activation of ODC could be seen. First, the activity of ODC increased four- to fivefold within 5-10 minutes (609-61 1).This rapid activation was not seen when nonmitogenic ligands bound to surface structures of T cells and was independent of de novo protein synthesis (61 1). There is evidence suggesting the involvement of a G protein and inositol metabolism, but PKC is apparently not required (407, 609, 6 12). A rapid protein synthesis-independent activation of ODC has also been reported in T cells stimulated with IL-2 (613),
290
AMNON ALTMAN E T AL.
in fibroblasts treated with trypsin (614), and in epinephrine-treated kidney cells (615). The initial rapid increase in ODC activity was followed by a much larger increase starting about 3-4 hours after the addition of Con A or PHA (609, 616-618). This large increase was, at least primarily, mediated by transcription, and perhaps by stabilization of the mRNA (619). Thus, ODC regulation during T cell activation is complex and operates at several levels, namely, transcription, mRNA activity, translation, posttranslational activation/inactivation, and degradation. Regulation of transcription has been suggested to involve a PKC-mediated mechanism (620, 62 1). This pathway probably includes the AP- 1 (Jun) and Fos proteins (see Section VIII,A,l). The role of polyamines in T cell proliferation remains unclear. They modulate many cellular functions, and it will be difficult to determine which are important in growth promotion. VIII. bhi;lear Events
The cascade of early biocymical events that follows binding of different extracellular ligands to T cells ultimately triggers a genetic program for growth and expression of differentiated functions. As part of this program, small quiescent T cells are transformed into lymphoblasts with increases in cellular and nuclear size, a more prominent Golgi apparatus, and a polyribosome-rich cytoplasm (622). On the molecular level these morphological and functional changes are accompanied, and apparently caused by, a coordinated sequential activation of previously silent genes. These transcriptional events culminate in the initiation of DNA synthesis approximately 24 hours after the initial triggering. Among the de novo-induced genes during T cell activation, some may encode products associated with cell cycle progression and, thus, may be common to many cell types. Others are unique for T cells and are associated with specific immunologicalfunctions of these cells (623,624). A recent review listed more than 70 gene products regulated during the early and late phases of T cell activation (3),and many of these represent de novo-induced genes. However, even this list may considerably underestimate the number of genes induced in activated T cells, since a recent study (624) identified over 60 novel and distinct cDNA clones that constitute part of the early genetic response of resting human peripheral blood T cells to mitogens. Induction of the majority of these genes was resistant to, and in many instances augmented by, the protein synthesis inhibitor cycloheximide. This indicates that such genes represent an immediate protein synthesis-independent response to mitogens and, further, that
MOLECULAR EVENTS MEDIATING
‘I’ CELL A C T I V A T I O N
29 1
their expression is transcriptionally controlled by labile regulatory proteins. Indeed, regulated genes in T cells have been divided (3) into immediate (i.e., independent of protein synthesis), early (i.e., protein synthesis dependent, but induced prior to onset of proliferation), and late (i.e., following initiation of cell division). Many of the genes cloned from activated T cells have not been sufficiently characterized, and their function is unknown. The presence of hydrophobic leader peptides in some of these (e.g., Refs. 624 and 625) suggests that they may be novel lymphokines or cell surface activation antigens. Among the regulated genes studied more extensively during T cell activation are the cell cycle-regulated protooncogenes, c-fos, c-myc, and c-myb, and some of the genes related to the specific functions of T cells, namely, lymphokine and lymphokine receptor genes. A. PROTOONCOGENES The nuclear protooncogenes, c-fos, c-myc, and c-myb, are expressed transiently during the cell cycle, and their gene products either determine specific restriction points or are induced in response to cell cyclespecific regulatory signals. They are induced by multiple growth factors and other ligands in many cell types and appear to participate in intracellular events that govern proliferation at a point at which different signal transduction pathways converge in the nucleus (626). The expression of these, and other, protooncogenes has been examined in fresh, resting (Go), and previously activated Go- or GI-synchronized murine or human T cells in response to TCR-CD3-active ligands (e.g., anti-CD3 mAb, PHA, and Con A), agents that mimic the action of the second messengers generated in this pathway, namely, PMA and calcium ionophores or IL-2. 1. c-fos
c-fos is a protooncogene that encodes a 55-kDa nuclear DNA-binding phosphoprotein (627) found in a complex with a 39-kDa protein encoded by the cjun protooncogene (628, 629). The latter protein was recently found to be identical with the transcription factor AP-1, which binds to enhancer sequences responsive to PKC-activating agents (629). These findings suggest a role for the Fos protein in the transcriptional activation of other growth-related genes. As in other cell types, c-fos is the earliest identified gene induced during T cell activation. Its mRNA can be detected as early as 10-15 minutes after stimulation, peaks after approximately 30 minutes, and declines to undetectable levels by 1-2 hours (314, 323, 630-633). The mRNA was induced by mitogenic lectins, such as anti-CD3 mAb, PMA, or calcium
292
A M N O N ALTMAN E T A L .
ionophore alone, but PMA synergized with the other agents (314, 632). c-fos mRNA induction was resistant to cycloheximide, indicating that it is a primary event not requiring protein synthesis (314). Similarly, c-fos transcription was resistant to cyclosporine A, an inhibitor of T cell proliferation and lymphokine production (314). These results agree with studies in other cellular systems, in which c-fos was found to be expressed transiently during the early stage of transition from the GOto the GI phase of the cell cycle. It is not clear whether GI + S phase transition, which is induced by IL-2 in competent IL-2R' T cells, is also accompanied by c-fos mRNA induction. Go-synchronized human T cells previously activated with an antLCD3 mAb and IL-2 and rested for 2 days were found (634) to express c-fos mRNA, and expression was not increased upon reexposure to IL-2 (or phorbol ester). Similarly, IL-2 did not synergize with PHA in inducing c-fos mRNA in human peipheral blood lymphocytes (630). However, c-fos was induced by IL-2 in an IL-2-starved GI-arrested murine T cell clone (633).The response of this IL-2-dependent clone, CT6, would appear to be unrepresentative of the normal T cell response, and its physiological relevance is doubtful; thus, this same clone was found to hydrolyze PIP:, in response to IL-2 stimulation, a finding that could not be confirmed with other T cells studied (see Section IX,B). A CAMPanalog was also found to induce c-fos mRNA in IL-2-dependent CT6 cells, but to inhibit, in parallel, synthesis of the Fos protein (501).Since c-fos is regulated at several levels, this latter finding is of importance, and it will be necessary to ascertain whether c$os mRNA transcription documented in other studies with T cells is accompanied by protein synthesis.
2. c-my The c-my gene product is a 58- to 67-kDa protein believed to be important in cell cycle regulation. The addition of affinity-purified antibodies against the human c-myc protein to isolated nuclei from several cell types inhibited DNA synthesis and DNA polymerase activity (635), suggesting that c-myc encodes a protein that is involved in DNA synthesis. Regulation of this protooncogene during T cell activation has been studied extensively following the initial report that Con A stimulation of murine spleen cells rapidly induced c-my transcription (636). Like c-fos, c-my induction was independent of protein synthesis (314, 630, 636639). In fact, cycloheximide superinduced c-my mRNA in combination with T cell mitogens (636, 639) or slowed down its degradation (314), apparently by inhibiting the synthesis of a labile protein that downregulates c-my expression. The activation of c-myc transcription is slower than that of c-fos. The
MOLECULAR EVENTS MEDIATING
'r CELL ACTIVATION
293
mRNA could be detected 30 minutes after stimulation, peaked after 13 hours, and was still be seen at 24-48 hours. Expression declined to undetectable levels upon entry into the S phase (314,323,528,630-634, 636,638-642). Consequently, agents that inhibit the G --* S phase transition (e.g., hydroxyurea and antitransferrin receptor antibody) had no effect on induction of c-myc mRNA expression, but prevented its later down-regulation (640). Either activation of PKC by phorbol ester or an increase in [Ca2+]i by calcium ionophore induced c-myc expression, and a combination of the two was synergistic (314, 323, 631, 632, 639, 640). Unlike c-fos, the induction of c-myc mRNA by ligands that induce competence for proliferation in T cells (i.e., Go * G1 transition with IL-2R acquisition) was inhibited by cyclosporine A or dexamethasone (314,637,640),as well as by CAMPanalogs (632); however, similar to the former protooncogene, induction by a calcium ionophore, but not by PMA, required extracellular Ca2+ and a functional Na+/H+ antiporter (632). Moreover, c-myc differs from c-fos in that its mRNA was readily induced by IL-2 during the G I + S phase transition (314,633,637,640). This increase was resistant to cyclosporine A (314,367) and was inhibited by the anti-Tac mAb (640). T h e ability of both competence-inducing ligands and IL-2 to up-regulate c-myc mRNA during the G O + GI o r GI -+ S transition, respectively, c!early indicates that this protooncogene can be regulated by different pathways and at several points during the cell cycle. Although the exact function of c-myc in cellular activation and/or proliferation is unknown, treatment of T cells with c-myc-complementary antisense deoxyoligonucleotides almost completely inhibited T cell entry into the S phase, but not the Go+ G1 transition, and blastic transformation induced by PHA (643,644) or by IL-2 (644). This finding suggests that c-myc expression is not a prerequisite for the induction of competence in T cells. Alternatively, residual translation of the c-myc gene product in oligonucleotide-treated cells was sufficient to induce the Go + GI transition. Two recent studies indicated that the decreased c-myc mRNA expression by abnormally expanding CD4- CD8- T cells in autoimmune-prone lpr mice (639), or in T cells from aged mice (645), is not due to reduced transcription or mRNA stability, but is mediated by a posttranscriptional processing mechanism, possibly involving a labile protein inhibitor (645). 3. c-myb
T h e DNA-binding protein product of the c-myb protooncogene has been considered to play an important role in hematopoietic cell differen-
294
AMNON ALTMAN E 7 AL.
tiation because of its high-level expression in hematopoietic tissues and tumors and the decline in this expression during cell maturation (646, 647). In addition, a c-myb antisense deoxyoligonucleotide inhibited human hematopoiesis in vitro (648). However, this gene is expressed in other cell types during cell cycle progression (e.g., the GI phase) and is regulated in parallel with the histone H-3 gene (642, 647). The gene is activated transiently as a result of posttranscriptional mechanisms. In immature thymocytes, however, exceptionally high levels of c-myb mRNA are expressed in both quiescent and proliferating cells as a consequence of an increased transcription (647). With the exception of one study that reported constitutive c-myb expression in resting peripheral blood T cells without a detectable mitogeninduced increase (314), other studies failed to detect c-myb expression in such cells. In fact, these other studies reported the induction of c-myb gene expression upon resting T cell stimulation with PHA (630,642) or, in the case of GI-synchronized previously activated T cells, with IL-2 (633,634,649).Gene expression was first seen 14 (630)or 40 (642) hours after PHA stimulation, and the delay seen in one of these studies (642) was probably due to the use of accessory cell-depleted purified T cells, in which IL-2 production and proliferation are expected to be delayed. This c-myb mRNA induction was sensitive to a protein synthesis inhibitor (630). In contrast, IL-2-induced c-myb expression occurred earlier, with a peak response seen 1 (649) or 4-5 hours (633,634) after the addition of IL-2. This apparent discrepancy probably reflects the fact that PHAinduced c-myb expression is secondary to IL-2 production (630, 649). This conclusion is supported by the findings that IL-2 synergized with PHA in c-myb induction (630) and that only IL-2, but not phorbol ester, which induces T cell competence (e.g., Go+ GI transition), stimulated rapid c-myb mRNA expression in Go-synchronized human T cells (634). In addition to IL-2, a CAMPanalog also stimulated c-myb gene transcription in an IL-2-dependent T cell clone (501),supporting the notion that different signal transduction pathways converge in the nucleus to induce a similar, if not identical, result, that is, the induction of growth- and differentiation-related genes. Interestingly, increased c-myb expression was found in spleens and enlarged lymph nodes of autoimmune mouse stains homozygous for the lp or gld mutation (650,651). B. TRANSCRIPTIONAL REGULATION OF IL-2 AND IL-2Ra GENE EXPRESSION Activation of the IL-2 and IL-2R genes is a prerequisite for subsequent T cell proliferation. While a relatively short (i.e., 1-hour) exposure to mitogen is sufficient to induce maximal levels of the early protoonco-
MOLECULAR EVENTS M E DI AT I NG
?’ CELL A<;TIVATION
295
genes, c-fos and c-myc (323) a commitment to proliferation and IL-2 gene activation requires longer exposure to the activating ligand (i.e., 2-6 hours) (56, 323). This period correlates with the time which IL-2 and IL-2Ra (p55) gene transcription is first detected (56), suggesting that the induction of these two genes commits T cells to proliferation. This period may represent the need for a sustained PIS hydrolysis and production of the relevant second messengers (77) or some other, as yet undefined, events. Although the IL-2 and IL-2Ra genes are induced at about the same time after initial stimulation, there are clear differences in their induction requirements. First, either phorbol ester or calcium ionophore will stimulate IL-2Ra gene expression, but a combination of these two agents (or phorbol ester plus TCR-CD3-active ligands) is required to induce the IL-2 gene (73,79,80,314,323,350).Second, induction of the IL-2 gene (314, 652, 653), but not that of IL-2Ra (p55) (314), is sensitive to cyclosporine A. Third, IL-2 up-regulates the IL-2Ra gene (314,654-656), but not its own gene (630). Thus, transcription of these two genes is differentially regulated. Since commitment of T cells to the proliferative pathway requires the induction of IL-2 and IL-2Ra genes, understanding the transcriptional events regulating the expression of these genes is crucial to elucidating the molecular mechanisms of commitment in T cells. Considerable effort to analyze these regulatory events has been made in the last few years, and, as a result, transcriptional enhancer regions found upstreeam from the IL-2 and IL-2Ra genes have been identified. The approaches used to identify these sequences include mapping of DNase I-sensitive regions and a functional test of the ability of upstream sequences to direct transcription and, hence, translation of an indicator gene, most commonly chloramphenicol acetyltransferase, linked to a ubiquitously expressed promoter. 1 . Transcriptional Regulation of the ZL-2 Gene
The IL-2 enhancer is included in a region between -3 19 and -52 bp 5’ to the transcription initiation site (657-663; reviewed in Ref. 664). Four distinct sites regulated by nuclear DNA-binding proteins were mapped within this region (3, 661, 662). Two sites, A and D, bind a protein, NFIL-2A, found in uninduced T cells, and this binding contributes most to the activity of the enhancer (3).Another site, termed E, binds a different nuclear factor, NFAT-1 (originally termed NFIL-2E) that is not expressed constitutively and is induced following TCR-CD3 triggering. These sites were responsive to ligands that trigger the TCR-CD3 complex, but not to PKC-activating phorbol esters (3,661). The NFIL-2A
296
AMNON ALTMAN E T AL.
binding site A and NFAT-1 binding site E have consequently been termed antigen receptor response elements (ARRE) 1 and 2, respectively (3,661,662).Th e ability of cyclosporine A to block in parallel IL-2 gene transcription (324, 652, 653) and the ability of ARRE-2 to activate a linked promoter (3,661) imply that the cyclosporine A effect is mediated by direct or indirect interaction with the NFAT-1 binding element. Induction of NFAT-1 depended on RNA and protein syntheses. Both ARRE- 1 and ARRE-2 correspond to DNase I-hypersensitive sites mapped in an earlier study (659), of which one (ARRE- 1) is inducible by TCR-CD3 triggering and the other, ARRE-2, is constitutive. Interestingly, ARRE- 1 also corresponds to an element that negatively regulates IL-2 production (665). Presumably, the NFIL-2A protein that binds to this element has a dual activity, suppressing expression of the IL-2 promoter in resting T cells and activating it upon cellular stimulation. In addition, a short-lived de now-induced repressor protein appears to down-regulate IL-2 production by shutting off the transcription of the IL-2 gene shortly after mitogen stimulation (666). Since activation of both ARRE elements can be stimulated via the TCR-CD3 complex, but not by phorbol esters (3,661), a distinct PMAresponsive element is also likely to be found in the IL-2 gene promoter. Indeed, such a site binding the nuclear transcription factor, AP- 1, has also been identified between ARRE-1 and ARRE-2 (3).As noted earlier, AP-1 is the product of the c-jun protooncogene, which regulates gene transcription by forming a complex with the Fos protein. Deletion of the AP-1 binding site from the IL-2 enhancer reduces the PMA response in a murine IL-2-producing T lymphoma cell line (667). These findings indicate that the IL-2 enhancer is composed of elements that respond to TCR-CD3-derived signals, as well as signals associated with PKC activation. Although TCR-CD3 triggering activates PKC transiently via the production of DAG, it is possible that this activation is qualitatively o r quantitatively insufficent to induce active AP- 1 formation. Thus, it would be of interest to determine whether IL-1, which can be mimicked by PMA, is capable of inducing AP-1. These findings provide a molecular basis for the finding that the induction of IL-2 gene transcription requires costimulation with a TCR-CD3-active ligand plus PMA. The IL-2 promoter region is also responsive to activation signals delivered by Tax, the trans-activator protein product of the tax gene of human T cell leukemia virus type 1. Although transient o r stable expression of the tax gene in human leukemic T cells leads to weak and transient activation of the IL-2 promoter (658, 668-670), this effect can be augmented by costimulation with Ca2+-mobilizingT cell ligands o r phorbol ester (658, 669, 670). Moreover, tax expression largely overcomes the
MOLECULAR EVENTS MEDIAI‘ING
‘I’ CELL A C T I V A T I O N
297
cyclosporine A-mediated suppression of the IL-2 gene promoter activity in response to PHA plus PMA stimulation (669).These results emphasize again the requirement of two different signals for optimal IL-2 production. Since the production of several lymphokines by T cells appears to be coordinated, it is not unlikely that similar regulatory motifs exist in other lymphokine genes (664). Upstream region homologies of the IL-2 gene to the y-interferon gene (660, 67 1) and to the genes encoding IL-3 and granulocyte o r granulocyte-macrophage colony-stimulating factors (G-CSF and GM-CSF, respectively) (670,672)have been reported. This is supported by the finding that the expression of tax in T cell leukemia lines stimulates, in addition to IL-2 gene expression (see above), expression of the genes encoding GM-CSF, IL-3, and IL-4 (673). Moreover, the homology between IL-2 and IL-2Ra gene regulatory sequences (660, 674) may partially explain the coordinated induction of these two genes.
2. TranscriptionalRegulation of IL-2Ra (p55)Gene Considerable attention has also been given to transcriptional regulatory sequences in the IL-2Ra promoter region. Unlike IL-2 induction, which requires a combination of two different signals, expression of the IL-2Ra gene and its product (p55) requires one signal that can be provided by: mitogens; PKC-activating phorbol esters; calcium ionophores; antigen-MHC complexes; anti-TCR-CD3 antibodies (73, 79, 80, 314, 323); IL-1 (233, 675, 676); tumor necrosis factor-a (233, 677); CAMP (233, 500, 501); the human T cell leukemia virus type I trans-activator protein, Tax (502, 658, 668-670, 678-680); and, finally IL-2 itself (314,654-656). This multiplicity of inducing agents operating via different pathways suggests that distinct regulatory sequences probably exist in the IL-2Ra promoter. Indeed, analysis of this region with different deletion mutants and stimulating agents identified (678) multiple protein binding sites in the region encompassing nucleotides -476 to -225 (relative to the RNA transcription initiation site). Other studies indicated that the promoter region found 3’ to nucleotide -327 (679) o r -317 (502) contains the regulatory sequences responsive to the different agents that activate the IL-2Ra gene. However, different stimuli can be distinguished by the promoter sequences required for their activity. Thus, promoter inducibility by PHA plus PMA required sequences downstream from nucleotide -3 17 (502, 679); however, the region 3’ of nucleotide -265 was sufficient for promoter activation by Tax (502, 658, 679) in Jurkat cells or by PMA, IL-1, or forskolin (a stimulator of CAMP production) in Y T leukemic cells (502). A strong synergy in the activation of the IL-2Ra promoter was seen be-
298
AMNON ALTMAN E T AL.
tween Tax and Ca2+-mobilizingligands (670) or between Tax and PMA (502). In addition, negative response elements that decrease promoter activation in response to different stimuli have also been found (502, 674). Within the IL-2Ra promoter region, an 11-nucleotide sequence (nucleotides -265 to -255), which functions as an enhancer, was found to be critical for mitogen- or Tax-induced promoter activation (678,680-682). This sequence is very similar to enhancers found in the promoter region of the Ig K light-chain gene (683-685), HIV-1 (685),several viruses, and other cell surface molecules (680). These enhancer sequences bind and cross-compete for binding to at least two inducible proteins, namely, NF-KB(51 kDa) (678, 680, 683-685a) and HIVEN86A (86 kDa) (681, 682), which activate the HIV-1 long terminal repeat. NF-KBis constitutively expressed in mature Ig-producing B cells or plasma cells and can be induced by different stimuli in pre-B cells, T cells, and even in nonlymphoid cells. However, it acts as a lymphoid-specific enhancer (683, 684). This transcription factor is induced by a posttranslational mechanism that involves its dissociation from a specific cytoplasmic inhibitor, IKB, and subsequent translocation to the nucleus upon phorbol ester induction (686). The inducibility of NF-KBin T cells and its negative regulation by IKB suggest that the corresponding enhancer element plays a critical role in T cell activation. Moreover, the presence of this enhancer in both the IL-2Ra promoter and the HIV-1 long terminal repeat and their similar NF-KBmediated inducibility by T cell mitogens provide a molecular basis for the observation that T cell activation is accompanied by enhanced HIV-1 replication (674,687,688) and activation of the viral long terminal repeat (674,680,682,685,689,690). Further support for the concept that the HIV-1 long terminal repeat is regulated like a T cell activation gene (674) comes from a recent study that described a CD4’ T cell-specific negative regulatory gene, Rpt-1 (691). The 41-kDa protein product of this gene down-regulated the expression of a reporter gene directed by the IL-2Ra promoter and by the HIV-1 long terminal repeat. The KBenhancer of HIV-1 (674, 680, 689), but not the IL-2Ra KB element (680) that differs from it (and from the authentic Ig-KB enhancer) in two of 11 nucleotides, is responsive to phorbol ester. This suggests that another enhancer sequence responsive to PMA must exist within the IL-2Ra promoter, and that the KBenhancer is not required or is insufficent for PMA-inducible promoter activation. One possibility is that the PMA-responsive enhancer is represented by the CC.Ar.GG box (“Ar” stands for a six-nucleotide A,T-rich sequence) found close to the KB enhancer in the IL-2Ra upstream region (692). This sequence func-
tions as an enhancer and is also found in actin genes and in the c-fos gene, where it mediates the serum response of c-fos (692). However, the PMAresponsive element of the IL-2Ra gene has not been clearly mapped. In sum, it is clear that transcriptional regulation of the IL-2Ra gene is a complex event mediated by various positive and negative regulatory elements that respond to a variety of transcription factors. This regulation may depend on the differentiation and maturation status of the T cells, as well as on the particular stimulus that induces expression of the IL-2Ra chain and, hence, the high-affinity IL-2R. The highly complex nature of the nuclear events that control T cell activation is apparent in a recent study that analyzed the expression of nine novel mitogen-induced genes in response to various T cell-activating ligands (623). PHA o r anti-CD3 or anti-CD2 mAbs (all of which provide primary activation signals) induced all nine genes. However, stimulation with PMA, a calcium ionophore, or an anti-CD28 mAb, all of which provide incomplete mitogenic signals, revealed heterogenous gene expression encompassing five regulatory classes (623). Much of the effort in the area of T cell signal transduction is likely to focus in the future on defining the relevant regulatory genetic elements and their corresponding transcription factors. IX. Signaling by the 11-2 Receptor
IL-2 is a critical factor for lymphocytes. T cell activation initiated by the TCR-CD3 complex or by alternative pathways (e.g., CD2 and Thy-1) induces the expression of functional high-affinity IL-2R and the secretion of IL-2. Interaction of IL-2 with its receptor signals activated T cells to enter the S phase and initiate proliferation via an autocrine pathway. In addition, IL-2 is also a differentiation factor capable of inducing distinct lymphocyte functions (see below). Among all lymphokines identified to date, IL-2 and its respective receptor have been characterized to the greatest extent. The development of a sensitive IL-2 bioassay (693) and an mAb (Tac) reactive with the human IL-2R (694) are some of the key events that facilitated the relatively rapid progress in this area. However, despite the wealth of relevant knowledge, little is known about the signaling events initiated by the IL-2R, and the majority of the information consists of negative data. A. STRUCTURE OF THE IL-2 RECEPTOR Early studies revealed that activated T cells display high-affinity (Kd = M) IL-2R (695) and that Tac inhibits the binding of IL-2 to its receptor (694). The IL-2 binding protein recognized by the Tac mAb
300
A M N O N ALTMAN E T AL.
(p55 or a) was subsequently cloned and unambiguously identified as a 55-kDa IL-2R (696,697). However, several discrepancies led to the suggestion that another IL-2-binding protein exists. These included findings that Tac binding sites far outnumber IL-2 binding sites (694, 695), that IL-2 binding to its high-affinity receptor reduces the receptor number by about 50% while, at the same time, increasing Tac expression by tenfold (654), that high-affinity IL-2R turns over much faster than the Tac antigen (654), and that high IL-2 concentrations can stimulate Ig secretion by certain Tac- B cell lines (698). In addition, transfection of Tac antigen cDNA into fibroblasts or T cells resulted in the expression of low-affinity nonfunctional or high-affinity biologically active IL-2R, respectively (699,700), and fusion of Tac-containing membranes with TacT cell membranes led to the expression of high-affinity IL-2R (70 1,702). Finally, the availability of large quantities of pure IL-2 allowed the identification of an additional class of IL-2R with low affinity (Kd M) (703). These studies culminated in the discovery of an additional IL-2binding protein revealed by cross-linking of radiolabeled IL-2 to T cell membranes. This protein, termed p75 or p, has a size of 70-75 kDa, binds IL-2 with an affinity of = lo3 M, and is expressed on resting T and natural killer cells, among others (704-708). Association of IL-2Ra (p55; K d =lo-* M) with IL-2Rp (p75; K d =lo-' M) forms the high-affinity receptor (Kd =lo-" M) that combines the fast association kinetics of IL-2Ra with the slow dissociation rate of IL-2Rp (709, 710). Using mAbs specific for IL-2Rp (71 1) to screen an expression cDNA library, a cDNA encoding this subunit of the IL-2R has very recently been isolated (712). This cDNA encodes a protein that, in its mature form, consists of 525 amino acids, including a 286-amino-acid intracellular domain. As indicated by the lack of appropriate consensus sequences, the cytoplasmic domain is unlikely to encode a TPK (712). Cotransfection experiments with IL-2Ra and/or IL-2Rp cDNAs in lymphoid cells gave rise to the three known classes of IL-2R, namely, low (IL-SRa)-, medium (IL-BRP)-, and high (a+ P)-affinity receptors. A nonlymphoid cell line transfected with the IL-2Rp cDNA did not bind IL-2, suggesting the involvement of a cell type-specific processing mechanism and/or an additional component. In fact, cross-linking experiments revealed IL-2-binding proteins of other sizes, that is, 100, 135, and 180 kDa (713, 714), whose relationship to p55 (a)or p75 (p) is unknown.
B. FUNCTION OF THD IL-2R The IL-2Rp chain plays a pivotal role in IL-2-induced signal transduction. Already implied (699, 700) by the inability of IL-2Ra to function
MOLECULAR EVENTS MEDIAI'ING
T CELL A C T I V A T I O N
301
alone (with the exception of one report: 715), and the finding that only high-affinity IL-2Rs internalize IL-2 (716, 717), this thesis was corroborated by direct experiments showing that IL-2-mediated triggering of the p chain alone is sufficient to induce cellular differentiation (7 18),and that p, in the absence of a, mediates IL-2 internalization (500). Antibody blocking studies indicated that, although IL-2-dependent T cell growth at high IL-2 concentrations was not inhibited by anti-Tac mAbs (709, 71 l), it was almost completely blocked by an anti-IL-2RP mAb (71 1). This suggested that the /3 chain alone may be sufficient for proliferation, but does not rule out the possibility that IL-2Rp alone mediates cellular differentiation, while the high-affinity dimeric IL-2R is necessary for proliferation. In addition, the physiological significance of the I L 9 R P intermediate-affinity receptor expressed in the absence of the a subunit is questionable, since the concentrations of IL-2 needed to saturate this receptor are not likely to occur in viva The majority of the attempts to identify the transmembrane signaling pathway(s) associated with the IL-2R have yielded negative results. With the exception of studies by one group that documented IL-2-stimulated PKC activation (719), Ca2+ influx (720), and PIS hydrolysis (721), attempts by others to demonstrate the same events in IL-2-stimulated T cells have failed (257, 722-726; our unpublished observations). Moreover, recent studies using P K C mutants of a T cell line (446) or T cells depleted of their PKC by chronic phorbol ester treatment (727, 728) clearly indicated that PKC activity is not required for IL-2-mediated T cell growth (although it is necessary for TCR-CD3-mediated activation). While IL-2 was found to activate the Na+/H+ antiporter, a response usually associated with PKC activation o r Ca2+ mobilization (349), in leukemic cells that express IL-2R p, but not IL-2Ra (576), the biological significance of this effect is unclear; thus, the activation of Na+/H+ exchange and the resulting cellular alkalinization were not required for IL-2-induced T cell proliferation (575). Parenthetically, another lymphokine, IL-4, which is known to serve as a growth factor for both B and T cells, has also been reported not to stimulate PIS hydrolysis (729). Whether IL-2 shares some unique signaling pathway with other lymphokines remains to be determined. Several alternative signal transduction pathways can be considered for IL-2 action. One possibility is a cyclic nucleotide-dependent pathway, for example, the one involving adenylate cyclase. Although isolated reports of an IL-2-stimulated increase in cAMP (730) or cGMP (731) exist, the more common finding is inhibition of adenylate cyclase by IL-2 via a mechanism that involves PKC, but not direct activation of the inhibitory G protein (Gi (485). Conversely, cAMP analogs or adenylate cyclase stim-
302
AMNON ALTMAN ET AL.
ulators inhibited IL-2-induced T cell proliferation (73,508), phosphorylation (423), and gene transcription (501). In addition, a more recent study failed to find any evidence for the mediation of IL-2 effects by the cyclic nucleotide secondary messengers CAMPor cGMP (726). Another obvious possibility is that a TPK (or some other protein kinase) is physically, or at least functionally, associated with the IL-2R and plays a primary role in transducing the biological signal. The a chain of the IL-2R has a 13-amino-acid cytoplasmic domain, which is too short to encode kinase activity. Although the p chain has a much longer cytoplasmic domain, it does not appear to encode a TPK (712). The possibility of an additional receptor component with intrinsic kinase activity cannot be ruled out at present. A 78-kDa protein (IL-2R p?)included in anti-Tac immunoprecipitates was recently found to undergo IL-2-stimulated phosphorylation in vitro (732), implying a physical association of the IL-2R with a protein kinase. It was not established, however, whether tyrosine or serinelthreonine residues were phosphorylated. Other alternatives include the effects of IL-2 on arachidonic acid metabolism and the lipoxygenase pathway (355,356,733) or direct binding of IL-2 to the nucleus (717). The report of IL-2-stimulated pertussis toxin-sensitive binding and hybrolysis of GTP in membranes of an IL-2dependent T cell line (734) suggests that a G protein may be involved in signal transduction by the IL-2R. However, another study (725) showed that pertussis toxin does not block IL-2-mediated proliferation. In conclusion, it seems that binding of JL-2 to its receptor stimulates a signal transduction pathway different from that utilized by the TCRCD3 complex. This is hardly surprising, since antigen and IL-2 provide qualitatively different signals that can synergize in T cell activation. If these two ligands used identical signaling machinery, one would expect IL-2 to mimic the effects of TCR-CD3-active ligands, which is clearly not the case. This synergy between two different signals in T cell activation and growth is reminiscent of other cellular systems, such as 3T3 cells, in which agonists using different signal transduction pathways were found to synergize in stimulating cellular proliferation (579).
c. IL-2-INDEPENDENT T CELL PROLIFERATION Although conventional models of T cell proliferation attribute a central role to the interaction of IL-2 with its receptor on activated T cells, evidence has accumulated recently to indicate that IL-2-independent pathway@)of T cell proliferation exist. Earlier studies indicated that phorbol esters (444, 445) or calcium ionophores (553, 554) can each stimulate proliferation of human T cells. With the retrospective knowl-
MOLECULAR EVENTS MEDIATING
'1'CELL AC'IIVA'TION
303
edge that each of these agents by itself is incapable of stimulating IL-2 production, these results implied the existence of an IL-2-dependent pathway of T cell proliferation. This notion was verified in our studies examining the effects of PKCactivating tumor promoters on resting human T cells and an IL-2dependent murine T cell clone (735, 736). First, the mitogenic effect of the tumor promoters on T cells was IL-2 independent, since no IL-2 mRNA or IL-2 biological activity could be detected; second, tumor promoter (but not IL-2-)-mediated proliferation was resistant to neutralizing anti-IL-2 antibodies; third, PKC-activating agents stimulated the proliferation of a CTL clone incapable of producing IL-2 (735, 736). Finally, cyclosporine A did not inhibit the PMA-induced proliferation of human peripheral blood T cells, although it produced a near-complete inhibition of anti-CD3- or Con A-mediated proliferation (737). Similar to the phorbol ester response, no IL-2 mRNA or biological activity could be detected in Con A-stimulated peripheral blood lymphocytes, despite the strong proliferative response (736). These results were confirmed by others in PMA (738) -stimulated or anti-TCR mAb (739) -stimulated murine CTL clones, as well as in anti-CD3-treated human peripheral blood T cells o r T cell lines (740, 741). Attempts were made to identify other lymphokines o r monokines that could mediate the IL-2-independent T cell proliferation. Although IL-4 can function as a T cell growth factor (742), it was apparently not involved in the observed responses, since cyclosporine A blocked IL-4 synthesis, but had no effect on T cell proliferation (740). Similarly, IL-1 did not seem to mediate this response (743). However, another lymphokine, IL-6, was recently found to act as a T cell growth factor and stimulate, in a synergistic manner with PHA, strong proliferative responses of highly purified resting human T cells (743) o r mature murine thymocytes (744). IL-2 could not be detected in either case, and the proliferative response was not inhibited by anti-IL-2Ra antibody (743, 744), indicating that, for the most part, these responses were not mediated by the interaction of IL-2 with its high-affinity receptor. It remains to be determined whether IL-6, which can be produced by several cell types, including T cells and monocytes, is involved in the IL-2- independent responses observed by others (735-741). However, as with other lymphokines, IL-6 production was found to be inhibited by cyclosporine A (M. Lotz, personal communication). T h e cyclosporine A-resistant increase in lymphokine mRNA stability mediated by the CD28 activation pathway (202, 203) could also play a role in IL-2-independent T cell proliferation. The physiological relevance of IL-2-independent T cell proliferation is
304
A M N O N ALTMAN E T AL.
not known. However, since attempts to detect IL-2 production in vzvo have, in general, been unsuccessful, it is possible that this pathway of T cell growth is important in vzvo, for example, during intrathymic T cell maturation and differentiation, which are known to be accompanied by an intense proliferative activity. X. Perspectives
Studies on transmembrane signal transduction in T lymphocytes have gained considerable momentum in the last few years, owing in part to the characterization of various surface activation molecules, the ability to activate T cells via several pathways, and the realization that PIS turnover is tightly associated with T cell activation under normal conditions. It is clear, however, that, despite the wealth of information accumulated by many laboratories and summarized here, many gaps still exist in our knowledge. Rather than conclude this chapter with a brief summary of the major findings, we would like to present some of the questions that remain unanswered and offer perspectives for future studies in this field. A. ALTERNATIVE AND ACCESSORY ACTIVATION MOLECULES Although the central role of the TCR-CD3 complex in initiating signal transduction events in T cells is unquestionable, it has become clear that other T cell surface receptors also contribute to such events in important ways. These receptors can be divided into two groups. Included in the first group are those that initiate “alternative” activation pathways that bypass, at least under some conditions, the TCR-CD3 activation pathway. CD2 and Thy- 1 are prime examples of these receptors. T h e second group consists of “accessory” activation molecules, that is, molecules whose function generally depends on expression of a functionally intact TCR-CD3 complex. The best-characterized determinants in this category are CD4 o r CD8 and CD28. The distinction between these two classes of activation molecules is not always clear. Depending on the experimental conditions, some accessory molecules may appear to initiate alternative activation pathways and vice versa. Still, some of the more recently identified activation molecules have not been characterized to the extent that they can be unambiguously placed in either of the two categories. Understanding the exact role of distinct surface receptors in T cell activation is further complicated by the fact that the physiological ligands for many of them (e.g., Thy-1, CD5, CD28, and CD45) are unknown. Thus, defining the natural ligands for these activation molecules is likely to be a major undertaking in the near future. Identification of MHC class
MOLECULAR EVENTS MEDIATING
'I'
CELL A C T I V A T I O N
305
I1 or class I molecules and LFA-3 as the natural ligands for CD4 or CD8 and CD2, respectively, provide encouraging examples in this respect. Examples clearly exist for situations in which CD2 or Thy-1 can initiate signal transduction in cells that do not express a functional TCR-CD3 complex (85, 99, 115-1 17, 137). The alternative pathways of activation associated with these two T cell-specific surface molecules could play an important role during the intrathymic selection process of the T cell repertoire, at a stage prior to acquisition of a cell surface-expressed TCR-CD3 complex. The affinity or magnitude of ligand interactions with these receptors is perhaps higher in immature T cells relative to TCR-CD3' cells, or the TCR-CD3- thymocytes could express some unknown receptors, whose function is analogous to that of the TCRCD3 complex in more mature T cells. T h e high-level expression of the THAM activation antigen (220) on CD4- CD8- thymocytes is a possibly relevant example. First, the function of accessory activation antigens could be mediated by a number of mechanisms, and their combined action could mediate additive or synergistic effects on T cell activation. Some, such as CD4 or CD8 and LFA-1, as well as CD2, appear to function as adhesion receptors and, thus, could stabilize the interactions of T cells with APCs or with target cells, thereby facilitating transduction of activation signals via other receptors. The relative importance of such adhesion molecules in transmembrane signaling could be determined by the affinity and the level of expression of the primary activation antigens. Thus, CTL clones are known to differ in their sensitivities to anti-CD8-mediated inhibition. CD8 function is apparently less critical in those CTLs that express a high-affinity TCR (14). Similarly, memory T cells are more resistant than are naive T cells to the inhibitory effects of anti-CD4 or -CD8 mABs (745). Second, it has become clear that many leukocyte surface proteins are either enzymes in themselves [e.g., CD10, a peptidase (746); CD73, a 5'-ectonucleotidase (154); and CD45, a phosphotyrosine phosphatase (547)] or are associated with enzymes. Examples for the latter case are CD4 or CD8, which are associated with the T cell TPK, pp56lCk(186,187). Binding of ligands to receptors associated with enzymatic activity could modulate this activity and contribute in important ways to the complex biochemical changes occurring during lymphocyte activation. Some of the differentiation antigens possessing enzymatic activity [e.g., CD55 (DAF) or CD451 are not T cell specific and, thus, could affect the function and growth of other cell types. Defining the possible role of their associated enzymatic activities in T cell activation constitutes another important area for future studies.
306
AMNON ALTMAN ET AL.
Information is rapidly accumulating about the role of CD4 or CD8 and CD45 in T cell activation. The association of CD4 or CD8 with pp56Ick could provide an important mechanism for bringing this TPK to the vicinity of the TCR-CD3 complex, as well as other potential substrates that constitute part of the transmembrane signaling machinery (188). Although the physiological substrates of pp56"' have not been defined, the 6 chain associated with CD3 is one likely candidate. Phosphorylation of the 5 chain on tyrosine residues could affect its function in signal transduction. Our recent finding that pp56'"' can be rapidly activated by the CD45 PTPase (547a, 549) suggests that dephosphorylation events could be just as important in T cell activation. Although CD45 is not usually considered an accessory activation antigen, it is becoming increasingly clear that cross-linking of CD45 to other T cell differentiation antigens can inhibit or augment T cell activation, depending on the nature of the molecule cross-linked to CD45 (548). The recently discovered physical association between CD45 and CD2 (747) may underlie the cooperative interaction between the CD2 and CD3 pathways in T cell activation (74, 111, 112).Cross-linking between CD2 and CD3 would bring CD45 to the immediate vicinity of the TCR-CD3 complex, where the associated CD45 PTPase could act on several potential substrates. It is also interesting to note that the expression of CD2 and a low (180-kDa)-molecular-weightform of CD45 is markedly increased on the surface of memory T cells (745). This form of CD45 could be associated with an altered regulation of the associated PTPase activity, thereby facilitating the activation of memory T cells by comparison with naive T cells. The existence of different extracellular putative ligand-binding CD45 domains suggests that this PTPase could be differentially regulated in response to distinct leukocyte-specific ligands. It would be interesting to determine whether the PTPase activity of CD45 can be modulated by its cross-linking or by binding of its putative ligand once this ligand has been identified. CD45 may be a prototype for receptor-linked PTPases directly linked to a signal transduction mechanism involving protein dephosphorylation (547). Identifying the natural CD45 ligand will be critical to elucidating the regulatory role of this leukocyte common antigen in the activation and growth of T cells and other leukocytes. B. Is THE TCR-CD3 COMPLEX COUPLED TO A G PROTEIN? Studies in various cell types have reinforced the notion that PIP2 hydrolysis is coupled to receptors via G proteins (295, 368-372, 384, 388, 407-414). However, the identity of the G protein@) involved in this pathway is unknown. In the case of T cells, direct evidence for the
MOLECULAK EVENTS MEDIATING
T CELL ACTIVATION
307
coupling of a G protein to the TCR-CD3 complex is still missing. To date, no T cell-specific G proteins have been identified. Studies demonstrating functional effects of G protein activators (e.g., GTPyS or ALF4-) that mimic triggering of the TCR-CD3 complex provide only indirect evidence. Similarly, the effects of bacterial toxins known to modify G protein function by ADP ribosylation are difficult to interpret and may be mediated by mechanisms not involving a G protein (390, 395, 396). If a TCR-CD3-linked G protein exists, it may not contain sites that serve as substrates for cholera or pertussis toxins, and functional approaches based on the properties of the trimeric G proteins (Gs, Gi, Gt, and Go) may not be totally adequate. Other candidate G proteins need to be considered. One possibility is that one or more of the peptides associated with the TCR-CD3 complex (e.g., 5 and/or q) may couple this complex directly to PI-PLC. Although the primary structure of 5 does not indicate typical features of a G protein, it contains an ATP-binding motif (39) suggestive of some important functional property. The primary structure of q has not been reported to date; however, the finding that a variant of an antigen-specific T cell hybridoma with undetectable q expression is deficient in receptorstimulated PIP2 hydrolysis (45) suggests that q is essential for coupling the receptor complex to a PI-PLC. The size of r] (22 kDa) is within the range of the recently identified ras-related G proteins (398-404). Some of the ras-related proteins, as well as the products of ras genes themselves, could be components of the transmembrane signaling machinery in T cells. Although the ras gene products appear to function at a downstream site relative to the action of PI-PLC (383),it is possible that some related proteins are coupled to the TCR-CD3 complex. It would, therefore, be of interest to examine the expression of these genes in various T cell subsets. In addition, the use of antibodies (404, 406) or botulinum toxin (404, 405), which appear to react selectively with some ras-related proteins, combined with immunochemical and chemical cross-linking approaches, could also be useful for identifying the elusive TCR-CD3-coupled G protein. Thus, the search for this putative protein is likely to represent another active area of study in the future.
TRANSMEMBRANE SIGNALING PATHWAYS IN T CELLS C. ALTERNATIVE The dogma that PIP2 hydrolysis and the resulting formation of DAG and IPSconstitute the critical, and perhaps the only, signal transduction pathway coupled to the TCR-CD3 complex needs to be reexamined. Recent reports documented situations in which either T cell activation proceeds in the absence of detectable PIP2 hydrolysis and/or Ca2+mobilization (229, 288, 362, 363, 365) or, conversely, PIP2 hydrolysis is not
308
AMNON ALTMAN E T AL.
sufficient for T cell proliferation and IL-2 production (77, 139,357,360, 361). Thus, although PIP2 hydrolysis still appears to be a critical event for T cell activation, the current picture may require considerable revision to accommodate the contribution of additional (or alternative) signal transduction pathways. Increasing evidence points in the direction of a putative TPK as another signaling pathway involved in T cell activation in an important way. This TPK could well be the src-related T cell-specific kinase, pp56ICk, although the contribution of other TPKs is also possible. The evidence that pp56lckis a critical component of an alternative signal transduction pathway can be summarized as follows: First, pp56Ickis expressed at high levels only in T cells, and its expression is regulated during T cell activation (515). Second, this kinase is physically associated (186, 187) with the CD4 or CD8 glycoproteins that play an important role in antigen presentation to T cells, and it can be activated by cross-linking these surface glycoproteins (189). The association of pp56lckwith CD4 or CD8 may constitute an important mechanism to control the accessibility of this kinase to its physiological substrates, allowing it to phosphorylate relevant substrates (e.g., 6 ) during the critical time of TCR-CD3 triggering by a peptide-MHC complex (188). Third, pp56ICkseems to be the TPK mediating 6 chain phosphorylation, as indicated by the findings that 5 becomes phosphorylated on tyrosine residues when pp56lckis activated (1 89) and that 5 is constitutively phosphorylated in T cells overexpressing the pp56lCk-relatedTPK, pp60"" (748). Finally, overexpression of lck mRNA (359) and constitutive 6 tyrosine phopsphorylation (462) are associated with defective PIP2 hydrolysis and activation in lpr CD4- CD8- T cells (359). Molecular approaches designed to examine the effect of lck gene deletions or overexpression on T cell function and the use of specific TPK inhibitors may, therefore, be the methods of choice to address the role of pp56ICkin T cell signal transduction. Two findings reinforce the notion that two distinct signaling pathways play a role during T cell activation (Fig. 7). First, each T cell appears to express two distinct receptor complexes displaying either a 6-5 or a 5-7 heterodimer, and the ratio of these two complexes is tightly regulated (41). Second, tyrosine phosphorylation of 6 was intact in an 7-deficient T cell variant with minimal levels of receptor-stimulated PIP2 hydrolysis (45). Thus, PIP2 hydrolysis may be coupled to the 6-7-containing receptor complex, whereas TPK activation may be linked to the 6-6 homodimer (or both). Other receptors, for example, the platelet-derived growth factor (PDGF) receptor, can initiate, in parallel, PIS breakdown and activation of an associated TPK (749). Unlike the TCR-CD3 complex, however, the intracellular domain of the platelet-derived growth factor receptor is itself a TPK.
MOLECULAR EVENTS MEDIATING
?’ CELL A C T I V A T I O N
309
MHC
10%
90%
CD3
FIG. 7. Two classes of TCR-CD3 complexes are potentially coupled to two different signaling pathways. For explanations, see text. (Adapted from Ref. 45.)
Thus, if TPK activation is linked to TCR-CD3 triggering, a mechanism would be needed to enable the TCR-CD3 complex to “mobilize” such a TPK. T h e association of pp56Ickwith CD4-CD8 and the possible temporary linkage of CD4 or CD8 to TCR-CD3 during antigen presentation may provide precisely that putative mechanism. It remains to be seen whether these two signaling pathways are independent of each other, or perhaps interact in some manner. Possible modes of interaction include tyrosine phosphorylation of PI-PLC (305), PKC-mediated phosphorylation of pp56Ick(473, 516, 517), or the association of a PI kinase with a TPK (249). Although PKC activation seems to be an essential event for T cell activation, it is important to remember that DAG, the physiological PKC activator, could be generated via pathways unrelated to PIS breakdown, such as de nouo synthesis (235) or phosphatidylcholine hydrolysis (234). The PLC-mediated hydrolysis of GPI, a structure that anchors several proteins to the outer surface of the cell membrane (12 1,122),also generated DAG that could intercalate into the membrane and activate PKC on its inner leaflet. Thus, PKC activation can potentially occur in the absence of PIS breakdown. Furthermore, the hydrolytic action of a lipase on DAG can generate other products that play a potentially important role during T cell activation, namely, arachidonic acid and metabolites of the lipoxygenase pathway (355, 356, 733). In this regard arachidonic acid was recently found to function as a secondary messenger activating one of the PKC isoenzymes, PKCy (429, 750).
310
AMNON ALTMAN ET AL.
D. LYMPHOKINE-MEDIATED SIGNAL TRANSDUCTION The signaling pathway(s) mediating the action of various lymphokines on T and other cells is still an enigma. Although IL-2 was found to activate PKC, independent studies from several laboratories failed to detect IL-2-mediated PIS breakdown. However, IL-2 could potentially stimulate DAG production and, hence, PKC activation by a source other than PIS. The inability of I L 4 to stimulate PIS breakdown (729) suggests that lymphokines, as a group, utilize a different signaling machinery from the inositol phospholipid pathway used by lymphocyte antigen receptors. This would provide a logical explanation for the findings that antigen and lymphokines generate qualitatively different signal?, and, furthermore, that these two signals often synergize to induce the complete cycle of lymphocyte activation and proliferation. This model is analogous to the extensively studied 3T3 fibroblast system in which two distinct ligands utilizing different signal transduction pathways synergize by inducing, first, a Go + GI transition or competence, and, second, a GI --$ S phase transition or cell cycle progression (579). Lymphokine receptors could be associated with a putative TPK. However, the recent molecular cloning of the IL-2Rp (p75) subunit (712) indicated that this component of the IL-PR is unlikely to be a TPK. The possibility that another peptide with TPK activity is associated with the IL-2R is still viable. In this respect it has been suggested that the nonreceptor TPK encoded by the ltk gene may participate in signal transduction pathways initiated by hematopoietic growth factors (524). Other candidate pathways include arachidonic acid and its metabolites (355, 356, 733), some novel unknown second messenger(s), or a direction action of IL-2 (and other lymphokines) in the nucleus (717). E. SIGNAL TRANSDUCTION TO THE NUCLEUS Ligand binding to the TCR-CD3 complex activates PKC, Ca2+-CaMdependent protein kinase(s) and a TPK. However, these protein kinases most likely mediate their effects in the cytoplasm or at the inner surface of the cell membrane. Little is known about the biochemical events following secondary messenger formation and protein kinase activation that transduce the initial signals to the nucleus, thereby stimulating gene transcription. Proteins encoded by ras protooncogenes could serve as important intermediates that transduce activation signals to the cytoplasm. Several ”~ findings support such a role for ras gene products. First, intact ~ 2 1 ‘ was required for cellular proliferation, as indicated by the ability of microinjected rus-neutralizing antibodies to block cell division (751). Second, the
MOLECULAR EVENTS MEDIATING
‘1’ CELL ACTIVATION
31 1
same antibodies could block proliferation induced by phorbol esters plus the calcium ionophore (383), which, presumably, acts at a site distal to the PI-PLC-mediated production of second messengers. Thus, rasencoded proteins could be involved in signaling events occurring after the formation of DAG and IP3. It would therefore be interesting to apply similar approaches to T cell lines or tumors in an attempt to evaluate the role of rm-encoded proteins in T cell activation. Several protooncogenes whose products are cytoplasmic serine/ threonine protein kinases have been identified (480). At least one of these, Ruf-2, was recently found to be expressed in T cells and to be neutralizmodulated in response to ligand binding (482).An anti-p2 lraS ing antibody was found to interfere with cellular transformation induced by growth factor receptor-like viral oncogenes, but not with Ruf-2induced transformation (480), suggesting that the Ruf-1 gene product functions at a site distal to the rm-encoded proteins. This conclusion is supported by a similar study (48l), indicating that transformation by membrane-bound oncogene products, including p2 l“, but not by the nuclear oncogenes, v - m y and v-fos, induced phosphorylation of the Raf- 1 kinase on both serine/threonine and tyrosine residues and stimulated its kinase activity. The role of a cytoplasmic serinehhreonine protein kinase other than PKC is further supported by the finding that the nuclear products of thejun and fos protooncogenes are rapidly modified and/or phosphorylated in response to PMA treatment (752). One other important area for future studies is the nature of the regulatory events occurring in the nucleus itself as a result of T cell triggering by activating ligands. As is the case in many other cell types stimulated by agonists coupled to different signal transduction pathways, the nuclear protooncogene products of cfoos, c-myc, and c-myb seem to play an important role. Enhancer sequences responsive to several T cell-activating ligands have been identified in the promoter region of the IL-2Ra or some lymphokine genes, and characterization of the trans-acting transcription factors that bind to these enhancer regions has begun. Expansion of these analyses and application of similar experimental approaches to other genes involved in T cell-specific functions are likely to further clarify the details of the nuclear regulatory events occurring during T lymphocyte activation.
ACKNOWLEDGMENTS The authors are grateful to their past and present colleagues Dr. N. Isakov and Dr. W. Scholz, who participated in various aspects of the work, and to M. Kat Occhipinti for her patient and excellent editorial assistance. This is Publication 5916IMM from the Research Institute of Scripps Clinic, La Jolla, California
312
AMNON ALTMAN ET AL.
92037. The studies in our laboratory were supported by National Institutes of Health grants CA35299 and AR354 11, the Finska Lakaresdlskapet, Perklens Stiftelse, Duodecim, the Ella and Georg Ehrnrooth Foundation, a Leukemia Society of America Scholarship Award (to A.A.), and a Cancer Research Institute Fellowship (to K.M.C.).
REFERENCES 1. Weiss, A., and Imboden, J. B. (1987). T cell surface molecules and early events involved in human T lymphocyte activation. Adu. Immunol. 41, 1. 2. Clevers, H., Alarcon, B., Wileman, T., and Terhorst, C. (1988). The T cell receptorKD3 complex: A dynamic protein ensemble. Annu. Rev. Immunol. 6,629. 3. Crabtree, G. R. (1989). Contingent genetic regulatory events in T lymphocyte activation. Science 243,355. 4. Allison,J. P., McIntyre, B., and Bloch, D. (1982).Tumor-specific antigen or murine T-lymphoma defined with monoclonal antibody. J. Immunol. 129, 2293. 5. Meuer, S. C.., Fitzgerald, K. A., Hussey, R. E., Hodgdon, J. C., Schlossman, S. F., and Reinherz, E. L. (1983). Clonotypic structures involved in antigenspecific human T cell function. Relationship to the T 3 molecular complex. J. Exp. Med. 157, 705. 6. Haskins, K., Kubo, R., White, J., Pigeon, M., Kappler, J., and Marrack, P. (1983). The major histocompabitility complex-restricted antigen receptor on T cells. I. Isolation with a monoclonal antibody. J. Exp. Med. 157, 1149. 7. Kaye, J., Porcelli, S., Tite, J., Jones, B., and Janeway, C. A. (1983). Both a monoclonal antibody and antisera specific for determinants unique to individual cloned helper T-cell lines can substitute for antigen and antigenpresenting cells in the activation of T cells.J. Exp. Med. 158,836. 8. Yanagi, Y., Yoshikai, Y., Leggett, K., Clark, S. P., Aleksander, I., and Mak, T. W. (1984). A human T cell-specific cDNA clone encodes a protein having extensive homology to immunoglobulin chains. Nature (London)308, 145. 9. Hedrick, S. M., Nielsen, E. A,, Kavaler, J., Cohen, D. I., and Davis, M. M. (1984). Sequence relationships between putative T-cell receptor polypeptides and immunoglobulins. Nature (London)308, 153. 10. Schwartz, R. H. (1984). T lymphocyte recognition of antigen in association with gene products of the major histocompatibility complex. Annu. Rev. Immunol. 3,237. 11. Marrack, P., and Kappler, J. (1986). The antigen-specific, major histocompatibility complex-restricted receptor on T cells. Adu. Immunol. 38, 1. 12. Allison, J. P., and Lanier, L. L. (1987). The structure, function, and serology of the T-cell antigen receptor complex. Annu. Rev. Immunol. 5, 503. 13. Davis, M. M., and Bjorkman, P. J. (1988). T-cell antigen receptor genes and T-cell recognition. Nature (London)334,394. 14. Dembic, Z., Haas, W., Weiss, S., McCubrey,J., Kiefer, H., von Boehmer, H., and Steinmetz, M. (1986). Transfer of specificity by murine a and p T-cell receptor genes. Nature (London)320,232. 15. Saito, T., Weiss, A., Miller, J., Norcross, M. A., and Germain, R. N. (1987). Specific antigen-Ia activation of transfected human T cells expressing murine Ti @-human T 3 receptor complexes. Nature (London)325, 125.
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
313
16. Chien, Y., Becker, D. M., Lindsten, T., Okamura, M., Cohen, D. I., and Davis, M. M. (1984). A third type of murine T-cell receptor gene. Nature (London) 3 12, 3 1. 17. Saito, H., Kranz, D., Takagaki, Y., Hayday, A. C., Eisen, H. N., and Tonegawa, S. (1984). A third rearranged and expressed gene in a clone of cytotoxic T lymphocytes. Nature (London) 312,36. 18. Sim, G. K., Yague, J., Nelson, J., Marrack, P., Palmer, E., Augustin, A., and Kappler, J. (1984). Primary structure of human T-cell receptor a-chain. Nature (London) 312, 771. 19. Saito, H., Kranz, D. M., Takagaki, Y., Hayday, A. C., Eisen, H. N., and Tonegawa, S. (1984). Complete primary structure of a heterodimeric T-cell receptor deduced from cDNA sequences. Nature (London) 309, 757. 20. Chien, Y., Iwashima, M., Kaplan, K. B., Elliott, J. F., and Davis, M. M. (1987). A new T-cell receptor gene located within the alpha locus and expressed early in T-cell differentiation. Nature (London) 327,677. 21. Kung, P. C., Goldstein, G., Reinherz, E. L., and Schlossman, S. F. (1979). Monoclonal antibodies defining distinctive human T cell surface antigens. Science 206,347. 22. Allison, J. P., and Lanier, L. L. (1985). Identification of antigen receptorassociated structures on murine B cells. Nature (London) 314, 107. 23. Samelson, L. E., Harford,J. B., and Klausner, R. D. (1985). Identification of the components of the murine T cell antigen receptor complex. Cell (Cambridge, Mass.) 43,223. 24. Oettgen, H. C., Pettey, C., Maloy, W. L., and Terhorst, C. (1986). A T3-like protein complex associated with the antigen receptor on murine T cells. Nature (London) 320,272. 25. Meuer, S. C., Cooper, D. A., Hodgdon, J. C., Hussey, R. E., Fitzgerald, K. A., Schlossman, S. F., and Reinherz, E. L. (1983). Identification of the receptor for antigen and MHC complex on human inducer T lymphocytes. Science 222, 1239. 26. Borst, J., Alexander, S., Elder, J., and Terhorst, C. (1983). The T3 complex on human T lymphocytes involves four structurally distinct glycoproteins. J. Biol. Chem. 258,5 135. 27. Borst, J., Prendiville, M. A., and Terhorst, C. (1983). T h e T 3 complex on human thymus-derived lymphocytes contains two different subunits of 20 kDa. Eur. J. Immunol. 13,576. 28. Kanellopoulos, J. M., Wigglesworth, N. M., Owen, M. J., and Crumpton, M. J. (1983). Biosynthesis and molecular nature of the T3 antigen on human T lymphocytes. EMBO J. 2, 1807. 29. Borst, J., Coligan, J. E., Oettgen, H., Pessano, S., Malin, R., and Terhorst, C. F. (1984). The 6 and E chains of the human TS/T-cell receptor complex are distinct polypeptides. Nature (London) 312,455. 30. Oettgen, H. C., Kappler,J., Tax, W. J. M.,and Terhorst, C. (1984). Characterization of the two heavy chains of the T3 complex on the surface of human T lymphocytes.]. Biol. Chem. 259, 12039. 31. Brenner, M. B., Trowbridge, I. S., and Strominger, J. L. (1985). Crosslinking of human T cell receptor proteins. Association between the T-cell idiotypic /3 subunit and the T3 glycoprotein heavy subunit. Cell (Cambridge, Mass.) 40, 183.
314
AMNON ALTMAN E T AL.
32. Weissman, A. M., Samelson, L. E., and Klausner, R. D. (1986). A new subunit of the human T-cell antigen rtceptor complex. Nature (London)324, 480. 33. van den Elsen, P., Shepley, B.-A., Borst, J., Coligan, J. E., Markham, A. F., Orkin, S., and Terhorst, C. (1984). Isolation of cDNA clones encoding the 20K T 3 glycoprotein of human T-cell receptor complex. Nature (London) 312,413. 34. van den Elsen, P., Shepley, B.-A., Cho, M., and Terhorst, C. (1985). Isolation and characterization of a cDNA clone encoding the murine homologue of the human 20K TJ/T-cell receptor glycoprotein. Nature (London) 314, 542.
35. Krissansen, G. W., Owen, M. J., Verbi, W., and Crumpton, M. J. (1986). Primary structure of the T3 y subunit of the T3/T cell antigen receptor complex deduced from cDNA sequences: Evolution of the 1 3 y and 6 subunits. EMBO J. 5, 1799. 36. Gold, D. P., Puck,J. M., Pettey, C. L., Cho, M., Coligan,J., Woody,J. N., and Terhorst, C. F. (1986). Isolation of cDNA clones encoding the 20K nonglycosylated polypeptide chain of the human T-cell receptor/CD3 complex. Nature (London) 321,431. 37. Krissansen, G. W., Owen, M. J., Fink, P. J., and Crumpton, M. J. (1987). Molecular cloning of the cDNA encoding the T 3 y subunit of the mouse T3/T cell antigen receptor comp1ex.J. Zmmunol. 138,3513. 38. Gold, D. P., Clevers, H., Alarcon, B., Dunlap, S., Novotny, J., Williams, A. F., and Terhorst, C. (1987). Evolutionary relationship between the T 3 chains of the T-cell receptor complex and the immunoglobulin supergene family. Proc. Natl. Acad. Sci. U.S.A. 84,7649. 39. Weissman, A. M., Baniyash, M., Hou, D., Samelson, L. E., Burgess, W. H., and Klausner, R. D. (1988). Molecular cloning of the zeta chain of the T cell antigen receptor. Science 439, 1018. 40. Weissman, A. M., Hou, D., Orloff, D. G., Modi, W. S., Seuanez, H., OBrien, S. J., and Klausner, R. D. (1988). Molecular cloning and chromosomal localization of the human T-cell receptor 5 chain: Distinction from the molecular CD3 complex. Proc. Natl. Acad. Sci. U.S.A. 85,9709. 41. Baniyash, M., Garcia-Morales, P., Bonifacino, J. S., Samelson, L. E., and Klausner, R. D. (1988). Disulfide linkage of the 5 and 1) chains of the T cell receptor. Possible identification of two structural classes of receptors.J. Biol. Chem. 263,9874. 42. Samelson, L. E., Patel, M. D., Weissman, A. M., Harford, J. B., and Klausner, R. D. (1986). Antigen activation of murine T cells induces tyrosine phosphorylation of a polypeptide associated with the T cell antigen receptor. Cell (Cambridge, Moss.)46, 1083. 43. Baniyash, M., Garcia-Morales,P., Luong, E.,Samelson, L. E., and Klausner, R. D. (1988). The T cell antigen receptor 5 chain is tyrosine phosphorylated upon activation.J. Biol. Chem. 263, 18225. 44. Tunnacliffe, A., Olsson, C., Buluwela, L., and Rabbits, T. H. (1988). Organization of the human CD3 locus on chromosome 11. Eur. J . Zmmunol. 18, 1639. 45. Mercep, M., Bonifacino, J. S., Garcia-Morales, P., Samelson, L. E., Klausner, R. D., and Ashwell, J. D. (1988). T cell CD3-51) heterodimer expression and coupling to phosphoinositide hydrolysis. Science 242,57 1.
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
315
46. Sussman, J. J., Bonifacino, J. S., Lippincott-Schwartz, J., Weissman, A. M., Saito, T., Klausner, R. D., and Ashwell, J. D. (1988). Failure to synthesize the T cell CD3-5 chain: Structure and function of a partial T cell receptor complex. Cell (Cambridge, Mass.)52, 85. 47. Minami, Y., Weissman, A. M., Samelson, L. E., and Klausner, R. D. (1987). Building of a multichain receptor: Synthesis, degradation and assembly of the T-cell antigen receptor. Proc. Natl. Acad. Sci. U.S.A. 84, 2688. 48. Unanue, E. R., and Allen, P. M. (1987).T h e basis for the irnmunoregulatory role of macrophages and other accessory cells. Science 236,551. 49. Weiss, A., and Stobo, J. D. (1984).Requirement for the coexpression of T3 and the T cell antigen receptor on a malignant human T cell line. J. Exp. Med. 160, 1284. 50. Ohashi, P. S., Mak, T. W., van den Elsen, P., Yanagi, Y., Yoshikai, Y., Calman, A. F., Terhorst, C., Stobo, J. D., and Weiss, A. (1985). Reconstitution of an active surface TS/T-cell antigen receptor by DNA transfer. Nature (London) 316,606. 51. Goldsmith, M. A., and Weiss, A. (1988). New clues about T-cell antigen receptor complex function. Immunol. Today 9, 220. 52. Goldsmith, M. A., and Weiss, A. (1987). Isolation and characterization of a T-lymphocyte somatic mutant with altered signal transduction by the antigen receptor. Proc. Natl. Acad. Sci. U.S.A. 84,6879. 53. Goldsmith, M. A., Dazin, P. F., and Weiss, A. (1988). At least two nonantigen-binding molecules are required for signal transduction by the T-cell antigen receptor. Proc. Natl. Acad. Sci. U.S.A. 85,8613. 54. Kanellopoulos, J. M., DePetris, S., Leca, G., and Crumpton, M. J. (1985). The mitogenic lectin from Phaseolus vulgarzS does not recognize the T3 antigen of human T lymphocytes. Eur. J. Zmmunol. 15,479. 55. Weiss, A., Imboden, J., Shoback, D., and Stobo,J. (1984).Role of T3 surface molecules in human T-cell activation: T-cell-dependent activation results in an increase in cytoplasmic free calcium. Proc. Natl. Acad. Sci. U.S.A. 81,4 169. 56. Weiss, A., Shields, R.,Newton, M., Manger, B., and Imboden, J. (1987). Ligand-receptor interactions required for commitment to the activation of the interleukin 2 gene. J. Immunol. 138,2169. 57. O’Flynn, K., Krensky, A. M., Beverley, P. C. L., Burakoff, S. J., and Linch, D. C. (1985). Phytohaemagglutinin activation of T cells through the sheep red blood cell receptor. Nature (London) 313,686. 58. Van Wauwe, J. P., DeMey, J. R., and Goossens, J. G. (1980). OKT3: A monoclonal anti-human T lymphocyte antibody with potent mitogenic properties. J-Immunol. 124, 2708. 59. Von Wussow, P., Platsoucas, C. D., Wiranowska-Stewart, M., and Stewart, W. E. (1981). Human y interferon production by leukocytes induced with monoclonal antibodies recognizing T cells. J. Immunol. 127, 1197. 60. Tsoukas, C. D., Landgraf, B., Bentin, J., Valentine, M., Lotz, M., Vaughan, J. H., and Carson, D. A. (1985). Activation of resting T lymphocytes by antLCD3 (T3) antibodies in the absence of monocytes. J. Immunol. 135, 1719. 61. Crispe, I. N., Bevan, M. J., and Staerz, U. D. (1985). Selective activation of Lyt2+ precursor T cells by ligation of the antigen receptor. Nature (London) 317,627. 62. Kappler, J., Kubo, R., Haskins, K., White, J., and Marrack, P. (1983). T h e
316
AMNON ALTMAN E T A L .
mouse T cell receptor: Comparison of MHC-restricted receptors on two T cell hybridomas. Cell (Cambridge, Mass.) 34,727. 63. Meuer, S. C., Hodgdon, J. C., Hussey, R. E., Protentis, J. P., Schlossman, S. F., and Reinherz, E. L. (1983). Antigen-like effects of monoclonal antibodies directed at receptors on human T cell clones.J. Exp. Med. 158,988. 64. Chang, T., Kung, P., Gingras, S., and Goldstein, G. (1981). Does OKT3 monoclonal antibody react with an antigen-recognition structure on human T cells? Proc. Natl. Acad. Sci. U.S.A. 78, 1805. 65. Kaneoka, H., Perez-Rojas, G., Sasasuki, T., Benike, C., and Engleman, E. (1983). Human T lymphocyte proliferation induced by a pan-T monoclonal antibody (anti-Leu 4): Heterogeneity of response is a function of monocytes.J. Immunol. 131, 158. 66. Looney, R. J., and Abraham, G. (1984). The Fc portion of intact IgG blocks stimulation of human PBMC by anti-T3. J . Immunol. 133, 154. 67. Ceuppens, J. L., Bloemmen, F. J., and Van Wauwe, J. P. (1985). T cell unresponsiveness to the mitogenic activity of OKT3 antibody results from a deficiency of monocyte Fcy receptors for murine IgG2a and inability to cross-link the T3-Ti complex. J. Immunol. 135,3882. 68. Pichler, W. J., Walker, C., Bettens, F., Koponen, M., Tscharner, V. V., Kurrle, R., Snow, C., and de Weck, A. L. (1987). Differences of T cell activation by the anti-CD3 antibodies Leu 4 and BMAO3O. Cell. Immunol. 108, 175. 69. Tax, W. J. M., Hermes, F. F. M., Willems, R. W., Capel, P. J. A., and Koene, R. A. P. (1984). Fc receptors for mouse IgGl on human monocytes: Polymorphism and role in antibody-induced T cell proliferation. J. Immunol. 133,1185. 70. Peltz, G. A., Trounstine, M. L., and Moore, K. W. (1988). Cloned and expressed human Fc receptor for IgG mediates anti-CD3-dependent lymphoproliferation. J. Immunol. 141, 1891. 7 1. Erard, F., Nabholz, M.. Dupuy-d'Angeac, A., and MacDonald, H. R. (1985). Differential requirements for the induction of interleukin 2 responsiveness in L3T4+ and Lyt-2+ T cell subsets.J. Exp. Med. 162, 1738. 72. Albert, F., Hua, C., Truneh, A., Pierres, M., and Schmitt-Verhulst, A. M. (1985). Distinction between antigen receptor and IL2 receptor triggering events in the activation of alloreactive T cell clones with calcium ionophore and phorbol ester. J. Immunol. 139, 3649. 73. Isakov, N., and Altman, A. (1985). Tumor promoters in conjunction with calcium ionophores mimic antigenic stimulation by reactivation of alloantigen-primed murine T lymphocytes.J. Immunol. 135, 3674. 74. Yang, S. Y., Rhee, S., Welte, K., and Du ont, B. (1988). Differential in vitro activation of CD8-CD4' and CD4-CD8 T lymphocytes by combinations of antLCD2 and anti-CD3 antibodies. J. Immunol. 140,2115. 75. Williams, J. M., Deloria, D., Hansen, J. A., Dinarello, C. A., Loertscher, A,, Shapiro, H. M., and Strom, T. B. (1985). The events of primary T cell activation can be staged by use of Sepharose-bound anti-T3 (64.1) monoclonal antibody and purified interleukin 1.J. Immunol. 135, 2249. 76. Manger, B., Weiss, A., Weyand, C., Goronzy, J., and Stobo, J. D. (1985). T cell activation: Differences in the signals required for IL2 production by nonactivated and activated T cells.J. Immunol. 135,3669. 77. Goldsmith, M. A,, and Weiss, A. (1988). Early signal transduction by the antigen receptor without commitment to T cell activation. Science 240,1029.
P
MOLECULAR EVENTS M E D I A T I N G
?' <:ELI. ACTIVATION
3 17
78. Wakasugi, H., Bertoglio, J., Tursz, T., and Fradelizi, D. (1985). IL2 receptor induction on human T lymphocytes: Role for IL2 and monocytes. J. Immunol. 135, 32 1. 79. Isakov, N., Scholz, W., and Altman, A. (1986). Signal transduction and intracellular events in T-lymphocyte activation. Immunol. Today 7, 27 1. 80. Isakov, N., Mally, M. I., Scholz, W., and Altman, A. (1987). T-lymphocyte activation: The role of protein kinase C and the bifurcating inositol phospholipid signal transduction pathway. Immunol. Rev. 95, 89. 81. Lay, W. H., Mendes, N. F., Bianco, C., and Nussenzweig, V. (1971). Binding of sheep red blood cells to a large population of human lymphocytes. Nature (London) 230,53 1. 82. Howard, F. D., Ledbetter, J. A., Wang, J., Bieber, C. P., Stinson, E. B., and Herzenberg, L. A. (1984). A human T lymphocyte differentiation marker defined by monoclonal antibodies that block E rosette formation. J . Immunol. 126,2117. 83. Kamoun, M., Kadin, M. E., Martin, P. J., Nettleton, J., and Hansen, J. A. (198 1). A novel human T cell antigen preferentially expressed on mature T cells and shared by both well and poorly differentiated B cell leukemias and lymphomas. J . Immunol. 127,987. 84. Reinherz, E. L., Kung, P. C., Goldstein, G., Levey, R. H., and Schlossman, S. F. (1980). Discrete stages of human intrathymic differentiation: Analysis of normal thymocytes and leukemic lymphoblasts of T cell lineage. Proc. Natl. Acad. Sci. U.S.A. 77, 1588. 85. Fox, D. A., Hussey, R. E., Fitzgerald, K. E., Bensussan, A., Daley, J. F., Schlossman, S. F., and Reinherz, E. L. (1985). Activation of human thymocytes via the 50 kd T11 sheep erythrocyte binding protein induces the expression of IL2 receptors on both T3' and T3- populations. J . Immunol. 134,330. 86. Sewell, W. A., Brown, M. H., Dunne, J., and Owen, M. J. (1986). Molecular cloning of the human T-lymphocyte surface CD2 (T11) antigen. Proc. Natl. Acad. Sci. U.S.A. 83, 8718. 87. Sayre, P. H., Chang, H.-C., Hussey, R. E., Brown, N. R., Richardson, N. E., Spagnoli, G., Clayton, L. K., and Reinherz, E. L. (1987). Molecular cloning and expression of T11 cDNAs reveal a receptor-like structure on human T lymphocytes. Proc. Natl. Acad. Sci. U.S.A. 84, 2941. 88. Williams, A. F., Barclay, A, N., Clark, S. J., Paterson, D. J., and Willis, A. C. (1987). Similarities in sequences and cellular expression between rat CD2 and CD4 antigens. J. Exp. Med. 165, 368. 89. Yagita, H., Okumura, K., and Nakauchi, H. (1988). Molecular cloning of the murine homologue of CD2. Homology of the molecule to its human counterpart T 11.J. Immunol. 140,132 1. 90. Dustin, M. L., Sanders, M. E., Shaw, S., and Springer, T. A. (1987). Purified lymphocyte function-associated antigen 3 binds to CD2 and mediates T lymphocyte adhesion.J. Exp. Med. 165,677. 91. Plunkett, M. L., Sanders, M. E., Selvaraj, P., Dustin, M. L., and Springer, T. A. (1987). Rosetting of activated human T lymphocytes with autologous erythrocytes: Definition of'the receptor and ligand molecules as CD2 and lymphocyte function-associated antigen 3 (LFA-3).J. Exp. Med. 165,664. 92. Hunig, T., Tiefenthaler, G., Meyer zum Buschenfelde, K. H., and Meuer, S. C. (1987). Alternative pathway activation of T cells by binding of CD2 to its cell surface ligand. Nature (London) 326, 298.
318
AMNON ALTMAN ET AL.
93. Bierer, B. A,, Barbosa, J., Hermann, S., and Burakoff, S. J, (1988). Interaction of CD2 with its ligand, LFA-3, in human T cell proliferation. J. Immunol. 140,3358. 94. Krensky, A. M., Sanchez-Madrid, F., Robbins, E., Nagy, J., Springer, T. A., and Burakoff, S. J. (1983). The functional significance, distribution and structure of LFA- 1, LFA-2 and LFA-3: Cell surface antigens associated with CTL-target interactions. J. Immunol. 131,611. 95. Springer, T. A., Dustin, M. L., Kishimoto, T. K., and Marlin, S. D. (1987). The lymphocyte function-associated LFA- 1, CD2 and LFA-3 molecules: Cell adhesion receptors of the immune system. Annu. Rev. Immunol. 5,223. 96. Vollger, L. W., Tuck, D. T., Springer, T. A., Haynes, B. F., and Singer, K. H. (1987). Thymocyte binding to human thymic epithelial cells is inhibited by monoclonal antibodies to CD2 and LFA-3 antigens.J.Immunol. 138,358. 97. Meuer, S. C., Hussey, R. E., Fabbi, M., Fox,D., Acuto, O., Fitzgerald, K. A., Hodgdon, J. C., Protentis, J. P., Schlossman, S. F., and Reinherz, E. L. (1984). An alternative pathway of T cell activation: A functional role for the 50 kd T11 sheep erythrocyte receptor protein. Cell (Cambridge, Mass.) 36, 897. 98. Brottier, P., Boumsell, L., Gelin, C., and Bernard, A. (1985). T cell activation via CD2 (T,gp50) molecules: Accessory cells are required to trigger T cell activation via CD2-D66 plus CD2-9.6/Tll (sub 1) epitopes. J. Immunol. 135,1624. 99. Siliciano, R. F., Pratt, J. C., Schmidt, R. E.,Ritz, J., and Reinherz, E. L. (1985). Activation of cytolytic T lymphocyte and natural killer cell function through the T11 sheep erythrocyte binding protein. Nature (London) 317, 428. 100. Van Wauwe, J., Goosens, J., DeCock, W., Kung, P., and Goldstein, G. (1981). Suppression of human T-cell mitogenesis and E-rosette formation by the monoclonal antibody OKTl 1A. Immunology 44,865. 101. Alcover, A., Weiss, M. J., Daley, J. F., and Reinherz, E. L. (1986). The T11 glycoprotein is functionally linked to a calcium channel in precursor and mature T-lineage cells. Proc. Natl. A d . Sci. U.S.A. 83,2614. 102. Pantaleo, G., Olive, D., Poggi, A., Kozumbo, W. J., Moretta, L., and Moretta, A. (1987). Transmembrane signaling via the T11 dependent pathway of human T cell activation: Evidence for the involvement of 1,2 diacylglycerol and inositol phosphates. Eur. J. Immunol. 17,55. 103. Gardner, P., Alcover, A., Kuno, M., Moingeon, P., Weyand, C. M., Goronzy,J., and Reinherz, E. L. (1989). Triggering of T-lymphocytes via either T3-Ti or T11 surface structures opens a voltage-insensitive plasma membrane calcium-permeable channel: Requirement for interleukin-2 gene function. J. Biol. Chem. 264, 1068. 104. Tiefenthaler, G., Hunig, T., Dustin, M. L., Springer, T. A., and Meuer, S. C. (1987). Purified lymphocyte function-associated antigen-3 and T11 target structures are active in CD2-mediated T cell stimulation. Eur. J. Immunol. 17,1847. 105. Denning, S. M., Dustin, M.L., Springer, T. A., and Meuer, S. C. (1988). Purified lymphocyte function-associated antigen-3 (LFA-3)activates human thymocytes via the CD2 pathway. J. Immunol. 141,2980. 106. Dustin, M. L., Olive, D., and Springer, T. A. (1989). Correlation of CD2 binding and functional properties of multimeric and monomeric lymphocyte function-associated antigen 3. J . Exp. Med. 169,503.
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
319
107. Reinherz, E. L., Meuer, S. C., Fitzgerald, K. A,, Hussey, R. E., Levine, H., and Schlossman, S. F. (1982).Antigen recognition by human T lymphocytes is linked to surface expression of the T 3 molecular complex. Cell (Cambrulge, Mass.) 30, 735. 108. Pantaleo, G., Olive, D., Poggi, A., Pozzan, T., Moretta, L., and Moretta, A. (1987). Antibody-induced modulation of the CD3/T cell receptor complex causes T cell refractoriness by inhibiting the early metabolic step involved in T cell activation.J. Exp. Med. 166,619. 109. Faure, F.,Jitsukawa, S., Meuer, S., Bohuon, C., Triebel, F., and Hercend, T. (1988). Identification of a CD2- CD3' T cell receptor gamma' peripheral blood lymphocyte subpopulation. J . Immunol. 140,2 128. 110. Moingeon, P., Alcover, A., Clayton, L. K., Chang, H.-C., Transy, C., and Reinherz, E. L. (1988). Expression of a functional CD3-Ti antigen/MHC receptor in the absence of surface CD2. Analysis with clonal Jurkat cell mutants. J. Exp. Med. 168, 2077. 111. Yang, S. Y., Chouaib, S., and Dupont, B. (1986). A common pathway for T lymphocyte activation involving both the CD3-Ti complex and CD2 sheep erythrocyte receptor determinants. J. Immunol. 137, 1097. 112. Bierer, B. E., Peterson, A., Gorga, J. C., Herrmann, S. H., and Burakoff, S. J. (1988). Synergistic T cell activation via the physiological ligands for CD2 and the T cell receptor. J . Exp. Med. 168, 1145. 113. Breitmeyer, J. B., Daley, J. F., Levine, H. B., and Schlossman, S. F. (1987). The T11 (CD2) molecule is functionally linked to the T3/Ti T cell receptor in the majority of T cells.J. Immunol. 139,2899. 114. Weissman, A. M., Ross, P., Luong, E. T., Garcia-Morales, P., Jelachich, M. L., Biddison, W. E., Klausner, R. D., and Samelson, L. E. (1988). Tyrosine phosphorylation of the human T cell antigen receptor (-chain: Activation via CD3 but not CD2. J . Immunol. 141,3532. 115. June, C. H., Ledbetter, J. A., Rabinovich, P. S., Martin, P. J., Beatty, P. G., and Hansen, J. A. (1986). Distinct patterns of transmembrane calcium flux and intracellular calcium mobilization after CD2 or CD3 stimulation of human 1ymphocytes.J. Clin. Invest. 77, 1224. 116. O'Flynn, K., Knott, L., Russul-Saib, M., Abdul-Gaffar, R., Morgan, G., Beverley, P. C. L., and Linch, D. C. (1986). CD2 and CD3 antigens mobilize Ca2+independently. Eur. J. Immunol. 16,580. 117. Moretta, A., Poggi, A., Olive, D., Bottino, C., Fortis, C., Pantaleo, G., and Moretta, L. (1987). Selection and characterization of T-cell variants lacking molecules involved in T cell activation (T3 T-cell receptor, T44, and T 1 1): Analysis of the functional relationship among different pathways of activation. Proc. Natl. Acad. Sci. U.S.A. 84, 1654. 118. Holter, W., Majdic, O., Stockinger, H., Howard, B. H., and Knapp, W. (1988). Regulation of the CD2 alternate pathway of T cell activation by CD3: Evidence for heterologous desensitization.J. Immunol. 140, 1043. 119. Alcover, A., Alberini, C., Acuto, O., Clayton, L. K., Transy, C., Spagnoli, G., Moingeon, P., Lopez, P., and Reinherz, E. L. (1988). Interdependence of CD3-Ti and CD2 activation pathways in human T lymphocytes. EMBO J. 7, 1973. 120. Bockenstedt, L. A,, Goldsmith, M. A., Dustin, M., Olive, D., Springer, T. A,, and Weiss, A. (1988). The CD2 ligand LFA-3 activates T cells but depends on the expression and function of the antigen receptor. J. Immunol. 141, 1094.
320
AMNON ALTMAN E T A L .
121. Low, M. G., and Saltiel, A. R. (1988). Structural and functional roles of glycosyl-phosphatidylinositolin membranes. Science 239, 268. 122. Low, M. G. (1989).Glycosyl-phosphatidylinositol:A versatile anchor for cell surface proteins. FASEB J. 3, 1600. 123. Yeh, E. T. H., Reiser, H., Bamezai, A., and Rock, K. L. (1988). TAP transcription and phosphatidylinositol linkage mutants are defective in activation through the T cell receptor. Cell (Cambridge, Mass.) 52, 665. 124. Stiernberg,J., Low, M. G., Flaherty, L., and Kincade, P. W. (1987). Removal of lymphocyte surface molecules with phosphatidylinositol-specificphospholipase C: Effects on mitogen responses. J.Immunol. 138,3877. 125. Reif, A. E., and Allen, J. M. V. (1964). The AKR thymic antigen and its distribution in leukemias and nervous tissues.J. Exp. Med. 120, 413. 126. Seki, T., Chang, H.-C., Mariuchi, T., Denome, R., Ploegh, H., and Silver,J. (1985). A hydrophobic transmembrane segment at the carboxyl terminhus of Thy- 1. Science 227,649. 127. Williams, A. F., and Gagnon, J. (1982). Neuronal cell Thy-1 glycoprotein: Homology with immunology. Science 216,696. 128. Tse, A. G. D., Barclay, A. N., Watts, A., and Williams, A. F. (1985). A glycophospholipid tail at the carboxyl terminus of the Thy- 1 glycoprotein of neurons and thymocytes. Science 230, 1003. 129. Homans, S. W., Ferguson, M. A. J., Dwek, R. A., Rademacher, T. W., Anand, R., and Williams, A. F. (1988). Complete structure of the glycosyl phsophatidylinositol membrane anchor of rat brain Thy- 1 glycoprotein. Nature (London 333,269. 130. Low, M. G., and Kincade, P. W. (1985). Phosphatidylinositol is the membrane-anchoring domain of the Thy- 1 glycopEotein.’Nature (London) 318. 62. 131. Konaka, Y., Norcross, M. A., Maino, V. C., and Smith, R. T. (1981).AntiThy- 1-mediated T cell activation. Role of soluble factors and expression of interleukin 2 receptors. Eur. J,Immunol. 11,445. 132. Jones, B., and Janeway, C. A. (1981). Functional activities of antibodies against brain-associated T cell antigens. I. Induction of T cell proliferation. Eur. J. Immunol. 11,584. 133. Gunter, K. C., Malek, T. R., and Shevach, E. M. (1984). T cell-activating properties of an anti-Thy- 1 monoclonal antibody. Possible analogy to OKT3ILeu4.J. Exp. Med. 159, 716. 134. MacDonald, H. R., Bron, C., Rousseaux, M., Horvath, C., and Cerottini, J.-C. (1985). Production and characterization of monoclonal anti-Thy- 1 antibodies that stimulate lymphokine production by cytolytic T cell clones. Eur. J. Immunol. 15,495. 135. Pont, S., Regnier-Vigouroux, A., Naquet, P., Blanc, D., and Pierres, M. (1985). Analysis of the Thy-1 pathway of T cell hybridoma activation using 17 rat monoclonal antibodies reactive with distinct Thy-1 epitopes. Eur. J . Immunol. 15,1222. 136. Kroczek, R. A., Gunter, K. C., Seligmann, B., and Shevach, E. M. (1986). Induction of T cell activation by monoclonal anti-Thy- 1 antibodies.J.Immunol. 136,4379. 137. Kroczek, R. A., Gunter, K. C., Germain, R. N., and Shevach, E. M. (1986). Thy-1 functions as a signal transduction molecule in T lymphocytes and transfected B lymphocytes. Nature (London) 322, 181.
MOIXCULAK EVENTS MEDIATING
T CELL.
ACTIVATION
32 1
138. Schmitt-Verhulst, A.-M., Guimezanes, A., Boyer, C., Poenie, M., Tsien, R., Buferne, M., Hua, C., and Leserman, L. (1987). Pleiotropic loss of activation pathways in a T-cell receptor a-chain deletion variant of a cytolytic T-cell clone. Nature (London) 325, 628. 139. Gunter, K. C., Germain, R. N., Kroczek, R. A., Saito, T., Yokoyama, W. M., Chan, C., Weiss, A., and Shevach, E. M. (1987). Thy-I-mediated T-cell activation requires co-expression of CD3/Ti complex. Nature (London) 326, 505. 140. Wadsworth, S., Yui, K., and Greene, M. I. (1989). Major histocompatibility complex glass I-specific cytolytic T cells derived from gld mice, lacking Thy-1, CD4 and CD8. Proc. Nail.Acad. Sci. U.S.A. 86, 1018. 141. Rock, K. L., Yeh, E. T. H., Gramm, C. F., Haber, S. I., Reiser, H., and Benacerraf, B. (1986).TAP, a novel T cell-activating protein involved in the stimulation of MHC-restricted T lymphocytes. J. Exp. Med. 163, 315. 142. Yeh, E. T. H., Reiser, H., Benacerraf, B., and Rock, K. L. (1986). T h e expression, function and ontogeny of a novel T cell-activating protein, TAP, in the thymus. J . Immunol. 137, 1232. 143. Yeh, E. T. H., Reiser, H., Benacerraf, B., and Rock, K. L. (1986). Expression of T-cell-activating protein in peripheral lymphocyte subsets. Proc. N a d Acad. Sci. U.S.A. 83, 7424. 144. Reiser, H., Yeh, E. T. H., Gramm, C. F., Benacerraf, B., and Rock, K. L. (1986). Gene encoding T-cell-activating protein TAP maps to the Ly-6 locus. Proc. Natl. Acad. Sci. U.S.A. 83, 2954. 145. Reiser, H., Oettgen, H., Yeh, E. T. H., Terhorst, C., Low, M. G., Benacerraf, B., and Rock, K. L. (1986). Structural characterization of the TAP molecule: A phosphatidylinositol-linkedglycoprotein distinct from the T cell receptor/T3 complex and Thy-I. Cell (Cambridge, Mass.) 47,365. 146. Reiser, H., Coligan,J . , Benacerraf, B., and Rock, K. L. (1987). Biosynthesis, glycosylation, and partial N-terminal amino acid sequence of the T-cellactivating protein TAP. Proc. Nail. Acud. Sci. U.S.A. 84,3370. 147. Meruelo, D., Offer, M., and Rossmando, A. (1987). Evidence for a major cluster of lymphocyte differentiation antigens on murine chromosome 2. Proc. Natl. Acud. Sci. U.S.A. 79, 7460. 148. Kimura, S., Tada, N., Liu, Y., and Hammerling, U. (1984). Studies of the mouse Ly-6 region. Immunogenetics 20,47. 149. Reiser, H., Coligan, J., Palmer, E., Benacerraf, B., and Rock, K. L. (1988). Cloning and expression of cDNA for the T-cell-activating protein TAP. Proc. Natl. Acad. Sci. U.S.A. 85,2255. 150. Malek, T. R., Ortega, G., Chan, C., Kroczek, R. A., and Shevach, E. M. (1986). Role of Ly-6 in lymphocyte activation. 11. Induction of T cell activation by monoclonal anti-Ly-6 antibodies. J. Exp. Med. 164, 709. 151. Kinoshita, T., Medof, M. E., Silber, R., and Nussenzweig, V. (1985). Distribution of decay-accelerating factor in the peripheral blood of normal individuals and patients with paroxysmal nocturnal hemoglobinuria. J . Exp. Med. 162, 75. 152. Medof, M. E., Haas, R., Walter, E. I., and Rosenberry, T. L. (1986). Decayaccelerating factor (DAF) of complement is anchored to cells by a Cterminal glycolipid. Biochemistry 25,6740. 153. Davis, L. S., Patel, S. S., Atkinson, J. P., and Lipsky, P. E. (1988). Decayaccelerating factor functions as a signal transducing molecule for human T cells. J. Immunol. 141,2246.
322
AMNON ALTMAN ET AL.
154. Thompson, L. F., Ruedi, J. M., Low, M. G., and Clement, L. T. (1987). Distribution of ecto-5’-nucleotidase on subsets of human T and B lymphocytes as detected by indirect immunofluorescence using goat antibodies. J. Immunol. 139,4042. 155. Thompson, L. F., Ruedi, J. M., Glass, A., Low, M. G., and Lucas, A. H. (1989). Antibodies to 5’-nucleotidase (CD73), a glycosylphosphatidylinositol-anchored protein, cause human peripheral blood T cells to proliferate.J. Immunol. 143, 1815. 156. Koch, F., Thiele, H.-G., and Low, M. G. (1986). Release of the rat T cell antigen RT-6.2 from cell membranes by phosphatidylinositol-specificphospholipase C.J. Exp. Med. 164, 1338. 157. Wonigeit, K., and Schwinzer, R. (1987). Polyclonal activation of rat T lymphocytes by RT-6 alloantisera. Transplant. Proc. 19, 196. 158. Ledbetter, J. A., June, C. H., Crosmaire, L. S., and Rabinovitch, P. S. (1987). Cross-linking of surface antigens causes mobilization of intracellular ionized calcium in T cells. Proc. Natl. Acad. Sci. U.S.A. 84, 1384. 159. Littman, D. R. (1987). The structure of the CD4 and CD8 genes. Annu. Rev. Immunol. 5,56 1. 159a. Parnes, J. R. (1989). Molecular biology and function of CD4 and CD8. Adv. Immunol. 44,265. 160. Janeway, C. A. (1988). Accessories or coreceptors? Nature (London) 335, 208. 161. Fleischer, B., and Schrezenmeier, H. (1987). Do CD4 or CD8 molecules provide a regulatory signal in T-cell activation? Immunol. Today 9, 132. 162. Emmrich, F. (1988). Cross-linking of CD4 and CD8 with the T-cell receptor complex: Quaternary complex formation and T-cell repertoire selection. Immunol. Today 9,296. 163. Hollander, N., Pdlemer, E., and Weissman, I. L. (1981). Effects of Lyt antibodies on T-cell function: Augmentation by anti-Lyt-1 as opposed to inhibition by anti-Lyt-2. Proc. Natl. Acad. Sci. U.S.A. 78, 1148. 164. Bank, I., and Chess, L. (1985). Perturbation of the T4 molecule transmits a negative signal to T cells.J. Exp. Med. 162, 1294. 165. Saizawa, K., Rojo, J., and Janeway, C. A. (1987). Evidence for a physical association of CD4 and the CD3: a :p T-cell receptor. Nature (London) 328, 260. 166. Rivas, A., Takada, S., Koide, J., Sonderstrup-McDevitt, G., and Engleman, E. G. (1988). CD4 molecules are associated with the antigen receptor complex on activated but not resting T cells.]. Immunol. 140,2912. 167. Kupfer, A,, and Singer, S. J. (1988). Molecular dynamics in the membranes of helper T cells. Proc. Natl. Acad. Sci. U.S.A. 85,8216. 168. Bushkin, Y., Demaria, S., Le, J., andSchwab, R. (1988). Physical association between the CD8 and HLA class I molecules on the surface of activated human T lymphocytes. Proc. Natl. Acad. Sci. U.S.A. 85,3985. 169. Doyle, C., and Strominger, J. L. (1987). Interaction between CD4 and class I1 MHC molecules mediates T cell adhesion. Nature (London) 330, 256. 170. Norment, A. M., Salter, R. D., Parham, P., Engelhard, V. H., and Littman, D. R. (1988). Cell-cell adhesion mediated by CD8 and MHC class I molecule. Nature (London) 336,79. 171. Rosenstein, Y., Ratnofsky, S., Burakoff, S . J., and Herrmann, S. H. (1989).
MOLECULAR EVENTS MEDIATING
‘1’
CXLL ACTIVATION
323
Direct evidence for binding of CD8 to HLA class 1 antigens.J. Exp. Med. 169, 149. 172. Potter, T. A., Rajan, T. V., Dick, R. F., and Bluestone, J. A. (1989). Substitution at residue 227 of H-2 class I molecules abrogates recognition by CD8-dependent, but not CD8-independent, cytotoxic T lymphocytes, Nature (London) 337,73. 173. Salter, R. D., Norment, A. M., Chen, B. P., Clayberger, C., Krensky, A. M., Littman, D. R., and Parham, P. (1989). Polymorphisms in the a 3 domain of HLA-A molecules affects binding to CD8. Nature (London) 338, 345. 174. Gabert, J., Langlet, C., Zamoyska, R., Parnes, J. R., Schmitt-Verhulst, A.-M., and Malissen, B. (1988). Reconstitution of MHC class I specificity by transfer of the T cell receptor and Lyt-2 genes. Cell (Cambridge, Mass.) 50, 545. 175. Sleckman, B. P., Peterson, A., Jones, W. K., Foran, J. A., Greenstein, J. L., Seed, B., and Burakoff, S. J. (1987). Expression and function of CD4 in a murine T cell hybridoma. Nature (London) 328, 35 1. 176. Gay, D., Maddon, P., Sekaly, R., Talle, M. A., Godfrey, M., Long, E., Goldstein, G., Chess, L., Axel, R., Kappler, J., and Marrack, P. (1987). Functional interaction between human T-cell protein CD4 and the major histocompatibility complex HLA-DR antigens. Nature (London) 328,626. 177. Ballhausen, W. G., Reske-Kunz, A. B., Tourvieille, B., Ohashi, P. S., Parnes, J. R., and Mak, T. W. (1988). Acquisition of an additional antigen specificity after mouse CD4 gene transfer into a helper T cell hybrid0ma.J. Exp. Med. 167,1493. 178. Gay, D., Buus, S., Pasternak, J., Kappler, J., and Marrack, P. (1988). The T cell accessory molecule CD4 recognizes a monomorphic determinant on isolated la. Proc. Natl. Acad. Sci. U.S.A. 85,5629. 179. Walker, C., Bettens, F., and Pichler, W. J. (1987). Activation of T cells by cross-linking an antLCD3 antibody with a second anti-T cell antibody: Mechanism and subset-specific activation. Eur. J. Immunol. 17,873. 180. Walker, C., Bettens, F., and Pichler, W. J. (1987). T cell activation by cross-linking anti-CD3 antibodies with second anti-T cell antibodies: Dual antibody cross-linking mimics physical monocyte interactions. Eur.J. Zmmunol. 17, 1611. 181. Samstag, Y., Emmrich, F., and Staehelin, T. (1988). Activation of human T lymphocytes. Differential effectiveness of CD3- and CD8-mediated signaling. Proc. Natl. Acad. Sci. U.S.A. 85,9689. 182. Ledbetter, J. A., June, C. H., Rabinovitch, P. S., Grossman, A., Tsu, T. T., and Imboden, J. B. (1988). Signal transduction through CD4 receptors: Stimulatory vs. inhibitory activity is regulated by CD4 proximity to the CD3/T cell receptor. Eur. J. Immunol. 18,525. 183. Rosoff, P. M., Burakoff, S. J., and Greenstein, J. L. (1987). The role of the L3T4 molecule in mitogen and antigen-associated signal transduction. Cell (Cambridge, Mass.) 49,845. 184. Kornfeld, H., Cruikshank, W. W., Pyle, S. W., Berman, J. S., and Center, D. M. (1988). Lymphyocyte activation by HIV-1 envelope glycoprotein. Nature (London)335,445. 185. Carrel, S., Moretta, A., Pantaleo, G., Tambussi, G., Isler, P., Perussia, B., and Cerottini, J. C. (1988). Stimulation and proliferation of CD4’ periph-
324
AMNON ALTMAN E T AL.
era1 blood T lymphocytes induced by an anti-CD4 monoclonal antibody. Eur. J . Immunol. 18, 333. 186. Rudd, C. E., Trevillyan, J. M., Dasgupta, J. D., Wong, L. L., and Schlossman, S. F. (1988). The CD4 receptor is complexed in detergent lysates to a protein-tyrosine kinase (pp58) from human T lymphocytes.Proc. Nutl. Acad. Sci. U.S.A. 85,5190. 187. Veillette, A., Bookman, M. A., Horak, E. M., and Bolen, J. B. (1988). The CD4 and CD8 T cell surface antigens are associated with the internal membrane tyrosine-protein kinase ~56"'. Cell (Cambridge, Mass.) 55,301. 188. Mustelin, T., and Altman, A. (1989). Do CD4 and CD8 control T cell activation via a specific tyrosine protein kinase? Immunol. Today 10, 189. 189. Veilette, A., Bookman, M. A., Horak, E. M., Samelson, L. E., and Bolen, J. B. (1989). Signal transduction through the CD4 receptor involves the activation of the internal membrane tyrosine-protein kinase pp56ICk.Nature (London) 338,257. 190. Schrezenmeier, H., and Fleischer, B. (1988). A regulatory role for the CD4 and CD8 molecules in T cell activation.]. Immunol. 141, 398. 191. Von Boehmer, H., Teh, H. S., and Kisielow, P. (1989). The thymus selects the useful, neglects the useless and destroys the harmful. Immunol. Today 10, 57. 192. Hansen, J. A., Martin, P. J., and Nowinski, R. C. (1980). Monoclonal antibodies identifying a novel T-cell antigen and Ia antigens of human lymphocytes. Immunogenetics 10, 247. 193. Hara, T., Fu, S. M., and Hansen, J. A. (1985). Human T cell activation. 11. A new activation pathway used by a major T cell population via a disulfidebonded dimer of a 44 kilodalton polypeptide (9.3 antigen).J . Exp. Med. 161, 1513. 194. Moretta, A., Pantaleo, G., Lopez-Botet, M., and Moretta, L. (1985). Involvement of T44 molecules in an antigen-independent pathway of T cell activation. Analysis of the correlations to the T cell antigen-receptor complex. J . Exp. Med. 162,823. 195. Martin, P. J., Ledbetter, J. A., Morishita, Y., June, C. H., Beatty, P. G., and Hansen, J. A. (1986). A 44 kilodalton cell surface homodimer regulates lL2 production by activated human T lymphocytes. J . Immunol. 136, 3282. 196. Yokoyama, W. M., Koning, F., Kehn, P. J., Pereira, G. M. B., Stingl, G., Coligan,J. E., and Shevach, E. M. (1988). Characterization of a cell surfaceexpressed disulfide-linked dimer involved in murine T cell activation. J . Immunol. 141,369. 197. Weiss, A., Manger, B., and Imboden, J. (1986). Synergy between the T3/ antigen receptor complex and Tp44 in the activation of human T cells. J . Immunol. 137,819. 198. Pantaleo, G., Olive, D., Harris, D., Poggi, A., Moretta, L., and Moretta, A. (1986). Signal transducing mechanism involved in human T cell activation via the surface T44 molecules. Comparison with signals transduced via the T cell receptor complex. Eur. J . Immunol. 16, 1639. 199. Tong-Starksen, S. E., Luciw, P. A., and Peterlin, B. M. (1989). Signaling through T lymphocyte surface protens, TCR/CD3 and CD28, activates the HIV-1 long terminal repeat.]. Immunol. 142, 702. 200. Ledbetter, J. A., Martin, P. J., Spooner, C. E., Wofsy, D., Tsu, T. T., Beatty, P. G., and Gladstone, P. (1985). Antibodies to Tp67 and Tp44 augment and sustain proliferative responses of activated T cells.]. Immunol. 135,2331.
MOLECULAR EVENTS MEDIATING
?' CELL
AC?'IVA'rION
325
201. Yang, S. Y., Denning, S. M., Mizuno, S., Dupont, B., and Haynes, B. F. (1988). A novel activation pathway for mature thymocytes. Costimulation of CD2 (Tp50) and CD28 (Tp44) induces autocrine interleukin 2hnterleukin 2 receptor-mediated cell proliferation.,!. Exp. Med. 168, 1457. 202. Thompson, C. B., Lindsten, T., Ledbetter, J. A., Kunkel, S. L., Young, H. A., Emerson, S. G., Leiden, J. M., and June, C. H. (1989). CD28 activation pathway regulates the production of multiple T-cell-derived lymphokines/cytokines. Proc. Natl. Acad. Sci. U.S.A.86, 1333. 203. Lindsten, T., June, C. H., Ledbetter, J. A., Stella, C., and Thompson, C. B. (1989). Regulation of lymphokine messenger RNA stability by a surfacemediated T cell activation pathway. Science 244,339. 204. Ceuppens, J. L., and Baroja, M. L. (1986). Monoclonal antibodies to the CD5 antigen can provide the necessary second signal for activation of isolated resting T cells by solid-phase-bound OKT3. J . Immunol. 137, 1816. 205. Ledbetter,J. A., June, C. H., Martin, P. J. Spooner, C. E., Hansen,J. A., and Meier, K. E. (1986).Valency of CD3 binding and internalization of the CD3 cell-surface complex control T cell responses to second signals: Distinction between effects on protein kinase C, cytoplasmic free calcium, and proliferation.J. Immunol. 136, 3945. 206. Bierer, B. E., and Burakoff, S. J. (1988). T cell adhesion molecules. FASEB J. 2, 2584. 207. Carrera, A. C., Rincon, M., Sanchez-Madrid, F., Lopez-Botet, M., and de Landazuri, M. 0. (1988). Triggering of co-mitogenic signals in T cell proliferation by anti-LFA-1 (CD18, CDl ia), LFA-3, and CD7 monoclonal antibodies. J. Immunol. 141, 1919. 208. Geppert, T. D., Wacholtz, M. C., Davis, L. S., and Lipsky, P. E. (1988). Activation of human T 4 cells by cross-linking class I MHC molecules. J. Immunol. 140,2 155. 209. Remold-O'Donnell, E., Kenney, D. M., Parkman, R., Cairns, L., Savage, B., and Rosen, F. S. (1984). Characterization of a human lymphocyte surface sialoglycoprotein that is defective in Wiskott-Aldrich syndrome. J . Exp. Med. 159, 1705. 210. Mentzer, S. J., Remold-O'Donnell, E., Crimmins, M. A. V., Bierer, B. E., Rosen, F. S., and Burakoff, S. J. (1987). Sialophorin, a surface sialoglycoprotein defective in the Wiskott-Aldrich syndrome, is involved in human T lymphocyte proliferation.J. Exp. Med. 165, 1383. 21 1. Axelsson, B., Youseffi-Etemad, R., Hammarstrom, S., and Perlmann, P. (1988). Induction of aggregation and enhancement of proliferation and IL2 secretion in human T cells by antibodies to CD43.J. Immunol. 141,2912. 212. Carrel, S., Salvi, S., Giuffre, L., Isler, P., and Cerottini, J.-C. (1987). A novel SO-kDa polypeptide (Tp90) possibly involved in an antigen-independent pathway of T cell activation. Eur. J . Immunol. 17,835. 213. Hara, T.,Jung, L. K. L., Bjorndahl, J. M., and Fu, S. M. (1986). Human T cell activation. 111. Rapid induction of a phosphorylated 28 kD/32 kD disulfide-linked early activation antigen (EA-1) by TPA, mitogens and antigens. J. Exp. Med. 164, 1988. 214. Bjorndahl, J. M., Nakamura, S., Hara, T., Jung, L. K. L., and Fu, S. M. (1988). The 28-kDa/32-kDa activation antigen EA-1. Further characterization and signal requirements for its expression. J . Immunol. 141,4094. 215. Nakamura, S., Sung, S.-S. J., Bjorndahl,J. M., and Fu, S. M. (1989). Human
326
AMNON ALTMAN ET AL.
T cell activation. IV. T cell activation and proliferation via the early activation antigen EA-1. J. Exp. Med. 169,677. 216. Testi, R., Phillips, J. H., and Lanier, L. L. (1989). Leu 23 induction as an early marker of functional CD3/T cell antigen receptor triggering. J.Immunol. 142, 1854. 217. Cosulich, M. E., Rubartelli, A., Risso, A., Cozzolino, F., and Bargellesi, A. (1987). Functional characterization of an antigen involved in an early step of T-cell activation. Proc. Natl. Acad. Sci. U.S.A. 84,4205. 218. Fleischer, B. (1987). A novel pathway of human T cell activation via a 103 kD T cell activation antigen. J. Immunol. 138, 1346. 219. Morimoto, C., Rudd, C. E., Letvin, N. L., Hagan, M., and Schlossman, S. F. (1988). 2H1-A novel antigen involved in T lymphocyte triggering. J. Immunol. 140,2165. 220. Naquet, P., MacDonald, H. R., Brekelmans, P., Barbet, J., Marchetto, S., Van Ewuk, W., and Pierres, M. A. (1988). A novel T cell-activating molecule (THAM) highly expressed on CD4- CD8- murine thymocytes.J. Immunol. 141,4101. 221. Isler, P., Salvi, S., Rapin, C., Giuffre, L., Cerottini, J. C., and Carrel, S. (1988). Antigen-independent activation of T cells mediated by a novel cell surface heterodimer (Tp 135-145). Eur. J. Immunol. 18, 1491. 222. Durum, S. K., Schmidt, J. A., and Oppenheim, J. J. (1985). Interleukin 1: An immunological perspective. Annu. Rev. Immunol. 3,263. 223. Mizel, S. B. (1987). Interleukin 1 and T-cell activation. Immunol. Today 8, 330. 224. Dinarello, C. A. (1988). Biology of interleukin 1. FASEB J. 2, 108. 225. Dinarello, C. A., Clark, B. D., Puren, A. J., Savage, N., and Rosoff, P. M. (1989). The interleukin 1 receptor. Immunol. T o h y 10,49. 226. Lichtman, A. H., Chin, J., Schmidt, J. A., and Abbas, A. K. (1988). Role of interleukin 1 in the activation of T lymphocytes. Proc. Natl. Acad. Sci. U.S.A. 85,9699. 227. Sims, J. E., March, D. J., Cosman, D., Widmer, M. B., MacDonals, H. R., McMahan, C. J., Grubin, C. E., Wignall, J. M., Jackson, J. L., Call, S. M., Friend, D., Alpert, A. R., Gillis, S., Urdal, D. L., and Dower, S. K. (1988). cDNA expression cloning of the IL-1 receptor, a member of the immunoglobulin superfamily. Science 441,585. 228. Abraham, R. T., Ho, S. N., Barna, T. J., and McKean, D. J. (1987). Transmembrane signaling during interleukin 1-dependent T cell activation. Interaction of signal 1- and signal 2-type mediators with the phosphoinositide-dependent signal transduction mechanism. J. Biol. Chem. 262,2719. 229. Abraham, R. T., Ho, S. N., Barna, T.J., Rusovick, K. M., and McKean, D. J. (1988). Inhibition of T-cell antigen receptor-mediated transmembrane signaling by protein kinase C activation. Mol. Cell. Biol. 8, 5448. 230. Didier, M.,Aussel, C., Pelassy, C., and Fehlmann, M. (1988). IL1 signaling for IL2 production in T cells involves a rise in phosphatidylserine synthesis. J. Immunol. 141,3078. 231. Shirakawa, F., Yamashita, U., Chedid, M., and Mizel, S. B. (1988). Cyclic AMP-An intracellular second messenger for interleukin 1. Proc. Natl. A d . Sci. U.S.A. 85,8201. 232. Rosoff, P. M., Savage, N., and Dinarello, C. A. (1988). Interleukin-1 stimu-
MOLECULAR EVENTS MEDIATING
233. 234. 235. 236. 237. 238. 239. 240. 241. 242.
243. 244. 245. 246. 247.
248.
T CELL ACTIVATION
327
lates diacylglyerol production in T lymphocytes by a novel mechanism. Cell (Cambridge, Mass.)54,73. Scholz, W., and Altman, A. (1989). Synergistic induction of interleukin 2 receptor expression on Y T cells by interleukin 1 or tumor necrosis factor a in combination with CAMP-inducing agents. Cell Signaling, 1,367. Besterman,J. M., Duronia, V., and Cuatrecasas, P. (1986). Rapid formation of diacylglyceroi from phosphatidylcholine: A pathway for generation of a second messenger. Proc. Natl. Acad. Sci. U.S.A. 83,6785. Farese, R. V., Konda, T. S., Davis, J. S., Standaert, M. L., Pollet, R.J., and Cooper, D. R. (1987). Insulin rapidly increases diacylglycerol by activating de novo phosphatidic acid synthesis. Science 236,586. Wakelam, M. J. O., Murphy, G. J., Hruby, V. J., and Houslay, M. D. (1986). Activation of two signal transduction systems in hepatocytes by glucagon. Nature (London) 323,68. Hokin, L. E., and Hokin, M. R. (1958). Acetylcholine and the exchange of inositol and phosphate in brain phosphoinositide. J. Biol. Chem. 233, 818. Michell, R. H. (1975). Inositol phospholipids and cell surface receptor function. Biochim. Biophys. Acta 415,81. Fisher, D. B., and Mueller, G. D. (1971). Studies on the mechanism by which phytohemagglutinin rapidly stimulates phospholipid metabolism of human lymphocytes. Biochim. Biophys. Acta 248,434. Maino, V. C., Hayman, M. J., and Crumpton, M. J. (1975). Relationship between enhanced turnover of phosphatidylinositol and lymphocyte activation by mitogens. Biochem.J. 146,2 17. Taylor, M. V., Metcalfe, J. C., Hesketh, T. R., Smith, G. A., and Moore, J. P. ( 1984). Mitogens increase phosphorylation of phosphoinositides in thymocytes. Nature (London) 312,462. Imboden,J. B., and Stobo,J. D. (1985). Transmembrane signaling by the T cell antigen receptor. Perturbation of the T3-antigen receptor complex generates inositol phosphates and releases calcium ions from intracellular stores. J. Exp. Med. 161,446. OShea, J. J., Harford, J. B., and Klausner, R. D. (1986). Identification and characterization of the phosphatidylinositol kinase in membranes of murine T lymphocytes. J . Immunol. 137,971. Cockroft, S., Taylor, J. A., and Judah, J. D. (1985). Subcellular localisation of inositol lipid kinases in rat liver. Biochim. Biophys. Acta 845, 163. Vogel, S., and Hoppe, J. (1986). Polyamines stimulate the phosphorylation of phosphatidylinositol in membranes from A43 1 cells. Eur. J. Biochem. 154, 253. Cockcroft, S., Lamb, J. R., and Zanders, E. D. (1987). Inositol lipid metabolism in human T lymphocytes activated via the T 3 complex. Immunology 60, 209. Macara, I. G., Marinetti, G. V., and Balduzzi, P. C. (1984). Transforming protein of avian sarcoma virus UR2 is associated with phosphatidylinositol kinase activity: Possible role in tumorigenesis. Proc. Natl. Acad. Sci. U.S.A. 81, 2728. Sugimoto, Y., Whitman, M.,Cantley, L. C., and Erikson, R. L. (1984). Evidence that the Rous sarcoma virus transforming gene product phosphorylates phosphatidylinositol and diacylglycerol. Proc. Natl. Acad. Sci. U.S.A. 81, 21 17.
328
AMNON ALTMAN El' AL.
249. Blackshear, P. J., Nairn, A. C., and Kuo, J. F. (1988). Protein kinases 1988: A current perspective. FASEB J. 2,2957. 250. Van Rooiljen, L. A. A., Rossowska, M., and Bazan, N. G. (1985). Inhibition of phosphatidylinositol-4-phosphatekinase by its produce phosphatidylinositol-4,5-bisphosphate.Biochem. Biophys. Res. Commun. 126, 150. 251. Urumow, T., and Weiland, 0. H. (1986). Stimulation of phosphatidylinosito1 4-phosphate phosphorylation in human placenta membranes by GTPyS. FEBS Lett. 207,253. 252. Berridge, M. J. (1984). Inositol trisphosphate and diacylglycerol as second messengers. Biochem. J. 220, 345. 253. Majerus, P. W., Neufeld, E. J., and Wilson, D. B. (1984). Production of phosphoinositide-derived messengers. Cell (Cambridge, Mass.) 37, 70 1. 254. Majerus, P. W., Connolly, T. M., Deckmyn, H., Ross, T. S., Bross, T. E., Ishii, H., Bansal, V. S., and Wilson, D. B. (1986). The metabolism of phosphoinositide-derived messenger molecules. Science 234, 1519. 255. Wilson, D. B., Bross, T. E., Hofman, S. L., and Majerus, P. W. (1984). Hydrolysis of polyphosphoinositides by purified sheep seminal vesicle phospholipase C enzymes. J . Biol. Chem. 259, 117 18. 256. Imai, A., and Cershengorn, M. C. (1986). Phosphatidylinositol 4,5bisphosphate turnover is transient while phosphatidylinositol turnover is persistent in thyrotropin releasing hormone stimulated rat pituitary cells. Proc. Natl. Acad. Sca. U.S.A. 83, 8540. 257. Kozumbo, W. J., Harris, D. T., Gromkowski, S., Cerottini,J. C., and Cerutti, P. A. (1987). Molecular mechanisms involved in T cell activation. 11. T h e phosphatidylinositol signal transducing mechanism mediates antigeninduced lymphokine production, but not interleukin 2-induced proliferation in cloned cytotoxic T lymphocytes. J. Zmmunol. 138, 606. 258. Rittenhouse, S. E., and Sasson, J. P. (1985). Mass changes in myoinositol trisphosphate in human platelets stimulated by thrombin. Inhibitory effects of phorbol ester.J. Biol. Chem. 260,8657. 259. Bocckino, S. B., Blackmore, P. F., and Exton, J. H. (1985). Stimulation of 1,2-diacylglycerolaccumulation in hepatocytes by vasopressin, epinephrine and angiotensin 11.J. Biol. Chem. 260, 1420 1. 260. Majerus, P. W., Connolly, T. M., Bansal, V. S., Inhorn, R. C., Ross, T. S., and Lips, D. L. (1988). Inositol phosphates: Synthesis and degradation. J. Biol. Chem. 263,305 1. 261. Streb, H., Irvine, R. F., Berridge, M. J., and Schulz, I. (1983). Release of Ca2+from a nonmitochondrial intracellular store in pancreatic acinar cells by inositol-l,4,5-trisphosphate. Nature (London) 306,67. 262. Irvine, R. F., and Moor, R. M. (1986). Micro-injection of inositol 1,3,4,5tetrakisphosphate activates sea urchin eggs by a mechanism dependent on external Ca2+.Biochem. J. 240,917. 263. Irvine, R. F., Letcher, A. J., Heslop, T. P., and Berridge, M. J. (1986). The inositol tris/tetrakisphosphate pathway-Demonstration of Ins( 1,4,5)P3 3kinase activity in animal tissues. Nature (London) 320,63 1. 264. Stewart, S. J., Prpic, V., Powers, F. S., Bocckino, S. B., Isaacs, R. E., and Exton, J. H. (1986). Perturbation of the human T cell antigen receptor-T3 complex leads to the production of inositol tetrakisphosphate: Evidence for conversion from inositol trisphosphate. Proc. Natl. Acad. Sci. U.S.A. 83, 6098.
MOLECULAR EVENTS MEDIATING
1’CELL ACTIVATION
329
265. Zilberman, Y., Howe, L. R., Moore, T. P., Hesketh, T. R., and Metcalfe,J. C. (1987). Calcium regulates inositol 1,3,4,5-tetrakisphosphateproduction in lysed thymocytes and in intact cells stimulated with concanavalin. EMBO J. 6, 957. 266. Yamaguchi, K., Hirata, M., and Kuriyama, H. (1987). Calmodulin activates inositol 1,4,5-trisphosphate 3-kinase activity in pig aortic smooth muscle. Biochem. J. 244, 787. 267. Wilson, D. B., Bross, T. E., Sherman, W. R., Berger, R. A., and Majerus, P. W. (1985). Inositol cyclic phosphates are produced by cleavage of phosphatidylphosphoinositols (polyphosphoinositides) with purified sheep seminal vesicle phospholipase C enzymes. Proc. Natl. Acad. Sci. U.S.A. 82,40 13. 268. Wilson, D. B., Connolly, T. M., Bross, T. E., Majerus, P. W., Sherman, W. R., Tyler, A. N., Rubin, L. J., and Brown, J. E. (1985). Isolation and characterization of the inositol cyclic phosphate products of polyphosphoinositide cleavage by phospholipase C. Physiologicaleffects in permeabilized platelets and Limulus photoreceptor cells.J . Biol. Chem. 260, 13496. 269. Connolly, T. M., Bross, T. E., and Majerus, P. W. (1985). Isolation of a phosphomonoesterase from human platelets that specifically hydrolyzes the 5-phosphate of inositol 1,4,5-trisphosphate.J. Biol. Chem. 260, 7868. 270. Connolly, T. M., Lawing, W. J., Jr., and Majerus, P. W. (1986). Protein kinase C phosphorylates human platelet inositol trisphosphate 5’phosphornonoesterase, increasing the phosphatase activity. Cell (Cambridge, M a s . ) 46, 95 1. 271. Hallcher, L. M., and Sherman, W. R. (1980). Effects of lithium ion and other agents on the activity of myo-inositol- 1-phosphatase from bovine brain.J. Biol. Chem. 255, 10896. 272. Takimoto, K., Okada, M., Matsuda, Y., and Nakagawa, H. (1985). Purification and properties of myo-inositol- 1-phosphatase from rat brain. J. Biochem. (Tokyo) 98,363. 273. Allison, J. H., Blisner, M. E., Holland, W. H., Hipps, P. H., and Sherman, W. R. (1976). Increased brain myo-inositol 1-phosphate in lithium-treated rats. Biochem. Biophys. Res. Commun. 71, 664. 274. Keough, K. M. W., and Thompson, W. (1972). Soluble and particulate forms of phosphoinositide phosphodiesterase in ox brain. Biochim. Biophys. Acta 270, 324. 275. Hofman, S. L., and Majerus, P. W. (1982). Purification and properties of two distinct phosphatidylinositol-specific phospholipase C enzymes from sheep seminal vesicular glands. J. Biol. Chem. 257,6461. 276. Chau, L.-Y., and Tai, H.-H. (1982). Resolution into two different forms and study of the properties of phosphatidylinositol-specific phospholipase C from human platelet cytosol. Biochzm. Biophys. Acta 713, 344. 277. Low, M. G., and Weglicki, W. B. (1983). Resolution of myocardial phospholipase C into several forms with distinct properties. Biochem. J . 215, 325. 278. Nakanishi, H., Nomura, H., Kikkawa, U., Kishimoto, A., and Nishizuka, Y. (1985). Rat brain and liver soluble phospholipase C: Resolution of two forms with different requirements for calcium. Biochem. Biophys. Res. Commun. 132, 582. 279. Banno, Y., Nakashima, S., and Nozawa, Y. (1986). Partial purification of phosphoinositide phospholipase C from human platelet cytosol: Characterization of its three forms. Biochem. Biophys. Res. Commun. 136,713.
330
AMNON ALTMAN E T AL.
280. Ryu, S. H., Cho, K. S., Lee, K.-Y., Suh, P.-G., and Rhee, S. G. (1987). Purification and characterization of two immunologically distinct phosphoinisitide-specific phospholipase C from bovine brain. J. Biol. Chem. 262, 12511. 281. Ryu, S. H., Suh, P.-G., Cho, K. S., Lee, K.-Y., and Rhee, S. G. (1987). Bovine brain cytosol contains three immunologically distinct forms of inositolphospholipid-specific phospholipase C. Proc. Natl. Acad. Sci. U.S.A. 84,6649. 282. Allan, D., and Michell, R. H. (1974). Phosphatidylinositolcleavage catalyzed by the soluble fraction from lymphocytes. Activity at pH 5.5 and pH 7.0. Biochem.J. 142,591. 283. Allan, D., and Michell, R. H. (1974). Phosphatidylinositol cleavage in lymphocytes. Requirement for calcium ions at a low concentration and effects of other cations. Biochem. J. 142,599. 284. Carter, H. R., Bird, I. M., and Smith, A. D. (1986). Two species of phospholipase C isolated from lymphocytes produce specific ratios of inositol phosphate products. FEBS Lett. 204,23. 285. Carter, H. R., and Smith, A. D. (1987). Resolution of the phosphoinositidespecific phospholipase C isolated from porcine lymphocytes into multiple species. Partial purification of two isoenzymes. Biochem.J. 244,639. 286. Wang, P., Toyoshima, S., and Osawa, T. (1986). Partial purification and characterization of membrane-bound and cytosolic phosphatidylinositolspecific phospholipase C from murine thymocytes.J. Biochem. (Tokyo) 100, 1015. 287. Sasaki, T., and Hasegawa-Sasaki, H. (1985). Breakdown of phosphatidylinositol4,5-bisphosphate in a T cell leukemia line stimulated by phytohaemagglutinin is not dependent on Ca2+mobilization. Biochem.J. 227,97 1. 288. ORourke, A. M., and Mescher, M. F. (1988). T cell receptor-mediated signaling occurs in the absence of inositol phosphate production. J. Biol. Chem. 263, 18594. 289. Harris, D. T., Kozumbo, W. J., Testa, J. E., Cerutti, P. A., and Cerottini, J. C. (1988). Molecular mechanisms involved in T cell activation. 111. The role of extracellular calcium in antigen-induced lymphokine production and interleukin 2-induced proliferation of cloned cytoxic T lymphocytes. J. Immunol. 140,92 1. 290. Suh, P. G., Ryu, S. H., Moon, K. H., Suh, H. W., and Rhee, S. G. (1988). Cloning and sequence of multiple forms of phospholipase C. Cell (Cambridge, Mass.)54, 161. 291. Suh, P.-G., Ryu, S. H., Moon, K. H., Suh, H. W., and Rhee, S. G. (1988). Inositol phospholipid-specific phospholipase C: Complete cDNA and protein sequences and sequence homology to tyrosine kinase-related oncogene products. Proc. Natl. Acad. Sci. U.S.A. 85,5419. 292. Stahl, M. L., Ferenz, C. R., Kelleher, K. L., Kriz, R. W., and Knopf, J. L. (1988). Sequence similarity of phospholipase C with non-catalytic region of src. Nature (London)332,269. 293. Katan, M., Kriz, R. W., Totty, N., Philp, R., Meldrum, E., Aldape, R. A., Knopf, J. L., and Parker, P. J. (1988). Determination of the primary structure of PLC- 154 demonstrates diversity of phosphoinositide-specific phospholipase C activities. Cell (Cambridge, Mass.) 54, 171. 293a. Bennett, C. F., Balcarek, J. M., Varrichio, A., and Crooke, S. T. (1988). Molecular cloning and complete amino sequences of form-I phosphoinositide-specific phospholipase C. Nature (London)334,268.
MOLECULAR EVENTS MEDIATING
'I' CELL ACTIVATION
33 1
293b. Rhee, S. G., Suh, P.-G., Ryu, S.-H., and Lee, S. Y. (1989). Studiesofinositol phospholipid-specific phospholipase C. Science 244,546. 294. Mayer, B. J., Hamaguchi, M., and Hanafusa, H. (1988). A novel viral oncogene with structural similarity to phospholipase C. Nature (London)332, 272. 295. Smith, C. D., Cox, C. C., and Snyderman, R. (1986). Receptor-coupled activation of phosphoinositide-specific phospholipase C by an N protein. Science 232, 97. 296. Low, M. G., and Finean, J. B. (1976). The action of phosphatidylinositolspecific phospholipase C on membranes. Biochem. J. 154, 203. 297. Hofman, S. L., and Majerus, P. W. (1982). Modulation of phosphatidylinositol-specific phospholipase C activity by phospholipid interactions, diglycerides and calcium ions. J. Biol. Chem. 257, 14359. 298. Kamisaka, Y., Toyoshima, S., and Osawa, T. (1986). Phosphatidylinositolspecific phospholipase C of murine lymphocytes. Arch. Biochem. Biophys. 249, 569. 299. Graff, G., Nahas, N., Nikolopoulou, M., Natarajan, V., and Schmid, H. H. 0. (1984). Possible regulation of phospholipase C activity in human platelets by phosphatidylinositol 4,5-bisphosphate. Arch. Biochem. Biophys. 228, 299. 300. Paris, S., Chambard, J. C., and Pouyssegur, J. (1988). Tyrosine kinaseactivating growth factors potentiate thrombin- and ALFrinduced phosphoinositide breakdown in hamster fibroblasts. J. Biol. Chem. 263, 12893. 301. Helsop, J. P., Blakeley, D. M., Brown, K. D., Irvine, R. F., and Berridge, M. J. (1986). Effects of bombesin and insulin on inositol (1,4,5) trisphosphate and inositol (1,3,4) trisphosphate formation in Swiss 3T3 cells. Cell (Cambridge, Mass.)47,703. 302. Beaterman, J. M., Watson, S. P., and Cuatrecasas, P. (1986). Lack of association of epidermal growth factor-, insulin- and serum-induced mitogenesis with stimulation of phosphoinositide degradation in Balb/c 3T3 fibroblasts. J. Biol. Chem. 261,723. 303. Chu, S. H., Hoban, C. J., Owen, A. J., and Geyer, R. P. (1985). Plateletderived growth factor stimulates rapid polyphosphoinositide breakdown in fetal human fibroblasts. J. Cell. Physiol. 124, 391. 304. Jackowski, S., Rettenmier, C. W., Sherr, C. J., and Rock, C. 0. (1986). A guanine nucleotide-dependent phosphatidylinositol4,5-diphosphatephospholipase C in cells transformed by the v-fms and v-fes oncogenes. J. Biol. Chem. 261,4978. 305. Wahl, M. I., Daniel, T. O., and Carpenter, G. (1988). Antiphosphotyrosine recovery of phospholipase C activity after EGF treatment of A431 cells. Science 241, 968. 306. Hesketh, T. R., Smith, G. A., Houslay, M. D., Warren, G. B., and Metcalfe, J. C. (1977). Is an early calcium flux necessary to stimulate lymphocytes. Nature (London) 267,490. 307. Tsien, R. Y., Pozzan, T., and Rink, T. J. (1982). T cell mitogens came early changes in cytoplasmic free Ca" and membrane potential in lymphocytes. Nature (London) 295,68. 308. Niedel, J . E., Kuhn, L. J., and Vandenbark, G. R. (1983). Phorbol diester receptor copurifies with protein kinase C. Proc. Nad. Acad. Sci. U.S.A. 80, 36.
332
AMNON AL'I'MAN ET AL.
309. Kikkawa, U., Takai, Y., Tamaka, Y., Miyake, R., and Nishizuka, Y. (1983). Protein kinase C as a possible receptor protein of tumor-promoting phorbol esters. J. Biol. Chem. 258, 11442. 310. Castagna, M., Takai, Y.,Kaibuchi, K., Sano, K., Kikkawa, U., and Nishizuka, Y. (1982). Direct activation of calcium-activated, phospholipiddependent protein kinase by tumor-promoting phorbol ester.J.Biol. Chem. 247,7847. 31 1. Truneh, A., Albert, F., Goldstein, P., and Schmitt-Verhulst, A. M. (1985). Early steps of lymphocyte activation bypassed by synergy between calcium ionophores and phorbol ester. Nature (London) 313,318. 312. Koyasu, S., Suzuki, G., Asano, Y., Osawa, H., Diamantstein, T., and Yahara, 1. (1987). Signals for activation and proliferation of murine T cell clones. J. Biol. Chem. 262,4689. 313. Weiss, A., Imboden, J., Wiskocil, R., and Stobo,J. ((1984).The role ofT3 in ' the activation of human T cells.J. Clin. Immunol. 4, 165. 314. Granelli-Piperno, A., Andrus, L., and Steinman, R. M. (1986).Lymphokine and non-lymphokine mRNA levels in stimulated human T cells. Kinetics, mitogen requirements and effects of cyclosporin A. J . Exp. Med. 163, 922. 315. Goronzy, J., Weyand, C., Imboden, J., Manger, B., and Fathman, C. G. (1987). Heterogeneity of signal requirements in T cell activation with a panel of human proliferative T cell clones.J.Immunol. 138,3087. 316. Kaibuchi, K., Takai, Y.,and Nishizuka, Y. (1985). Protein kinase C and calcium ionophore in mitogenic response of macrophage depleted human peripheral T lymphocytes.J . Biol. Chem. 260, 1366. 317. Lugo, J. P., Krishnan, S. N., Sailor, R. D., and Rothenberg, E. V. (1986). Early precursor thymocytes can produce interleukin 2 upon stimulation with calcium ionophore and phorbol ester. Proc. Natl. Acad. Sci. U.S.A. 83, 1862. 318. Depper, J. M., Leonard, W. J., Kronke, M., Noguchi, P. D., Cunningham, R. E., Waldmann, T. A., and Greene, W. C. (1984). Regulation of interleukin 2 receptor expression: Effects of phorbol diester, phospholipase C and re-exposure to lectin or antigen. J.Immunol. 133,3054. 319. Leonard, W. J., Kronke, M., Peffer, N. J., Depper, J. M., and Greene, W. C. (1985). Interleukin 2 receptor gene expression in normal human T lymphocytes. Proc. Natl. Acad. Sci. U.S.A. 82,6281. 320. Ando, I., Hariri, G., Wallace, D., and Beverly, P. (1985). Tumor promoter phorbol esters induce unresponsiveness to antigen and expression of interleukin 2 receptor on T cells. Eur. J. Immunol. 15, 196. 321. Shackelford, D. A., and Trowbridge, I. S. (1984). Induction of expression and phosphorylation of the human interleukin 2 receptor by a phorbol diester. J. Biol. Chem. 259, 11706. 322. Shackelford, D. A., Smith, A. V., and Trowbridge, I. S. (1987). Changes in gene expression induced by a phorbol diester: Expression of lL2 receptor, T3 and T cell antigen receptor. J. Immunol. 138,613. 323. Kumagai, N., Benedict, S. H., Mills, G . B., and Gelfand, E. W. (1987). Requirements for the simultaneous presence of phorbol esters and calcium ionophores in the expression of human T lymphocyte proliferation-related genes.J. Immunol. 139, 1393. 324. Isakov, N., and Altman, A. (1987). Human T lymphocyte activation by tumor promoters: Role of protein kinase C. J. Immunol. 138, 3100.
MOI.EC1Il.AK EVENTS MEDIATING
'r CELL ACTIVATION
333
325. Kraft, A. S., and Anderson, W. B. (1983). Phorbol esters increase the amount of Ca'', phospholipid-dependent protein kinase associated with plasma membrane. Nature (London)301,62 1. 326. Farrar, W. L., and Ruscetti, F. W. (1986). Association of protein kinase C activation with IL2 receptor expression. J . Immunol. 136, 1266. 327. Ledbetter, J. A., Gentry, L. E., June, C. H., Rabinovitch, P. S., and Purchio, A. F. (1987). Stiniulation of T cells through the CD3/T cell receptor complex: Role of cytoplasmic calcium, protein kinase C translocation and phosphorylation of pp60-src in the activation pathway. Mol. Cell. Biol. 7, 650. 328. Nel, A. E., Bouic, P., Lattanze, G. R., Stevenson, H. C., Miller, P., Dirienzo, W., Stefanini, G. F., and Galbraith, R. M. (1987).Reaction of T lymphocytes with anti-T3 induces translocation of C-kinase activity to the membrane and specific substrate phosphorylation.J. Immunol. 138, 35 19. 329. Manger, B., Weiss, A., Imboden, J., Laing, T., and Stobo, J. D. (1987). The role of protein kinase C in transmembrane signaling by the T cell antigen receptor complex. Effects of stimulation with soluble or immobilized CD3 antibodies. J. Immunol. 139,2755. 330. Bell, R. M. (1986). Protein kinase C activation by diacylglycerol second messengers. Cell (Cambridge, M a s . ) 45,63 1. 33 1. Altman, A., Mally, M. I., and Isakov, N. (1990). Protein kinase C in human T lymphocytes: Synergy between ionomycin and TPA in selective isoenzyme activation and induction of related cellular functions. Submitted for publication. 332. Weiss, A,, Wiskocil, R. L., and Stobo, J. D. (1984). The role of T3 surface molecules in the activation of human T cells: A two-stimulus requirement for IL2 production reflects events occurring at a pre-translational level. J. Immunol. 131, 123. 333. Zlotnik, A., and Daine, B. (1986). Activation of IL1-dependent and IL1-independent T cell lines by calcium ionophore and phorbol ester. J. Immunol 136, 1033. 334. Resch, K., Gelfand, E. W., Hansen, K., and Ferber, E. (1972). Lymphocyte activation: Rapid changes in the phospholipid metabolism of plasma rnembrane during stimulation. Eur. J . Immunol. 2,598. 335. Hui, D. Y., and Harmony, J. A. K. (1980). Phosphotidylinositol turnover in mitogen-activated lymphocytes. Suppression by low density lipoproteins. @ Biochem. J . 192,91. 336. Treves, S., DiVirgilio, F., Cerundolo, V., Zanovello, P., Collavo, D., and Pozzan, T. (1987). Calcium and inositolphosphates in the activation of T cell-mediated cytotoxicity.J . Exp. Med. 166,33. 337. Allwood, G., Asherson, G. L., Davey, M. J., and Goodford, P. J. (197 1). T h e early uptake of radioactive calcium by human lymphocytes treated with phytohaemagglutinin.Immunology 21,509. 338. Whitney, R. B., and Sutherland, R. M. (1972). Enhanced uptake of calcium by transforming lymphocytes. Cell. Immunol. 5, 137. 339. Tsien, R. Y. (1981). A non-disruptive technique for loading calcium buffers and indicators into cells. Nature (London) 290,527. 340. Tsien, R. Y. (1980). New calcium indicators and buffers with high selectivity against magnesium and protons: Design, synthesis and properties of prototype structures. Biochemistly 19,2396.
334
AMNON ALTMAN ET AL.
341. Nisbet-Brown, E., Cheung, R. K., Lee, J. W. W., and Gelfand, E. W. (1985). Antigen-dependent increase in cytosolic free calcium in specific human T-lymphocyte clones. Nature (London) 316,545. 342. Oettgen, H. C., Terhorst, C., Cantley, L. C., and Rosoff, P. M. (1985). Stimulation of the T3-T cell receptor complex induces a membrane potential-sensitive calcium influx. Cell (Cambridge, Mass.)40, 583. 343. Imboden, J. B., Weiss, A., and Stobo,J. D. (1985). The antigen receptor on a human T cell line initiates activation by increasing cytoplasmic free calcium. J. Immunol. 134,663. 344. Utsunomiya, N., Tsuboi, M., and Nakanishi, M. (1986). Early transmembrane events in alloimmune cytotoxic T-lymphocyte activation as revealed by stopped-flow fluorimetry. Proc. Natl. Acad. Sci. U.S.A. 83, 1877. 345. Gray, L. S., Gnarra, J. R. Sullivan, J. A., Mandell, G. L., and Englehard, V. H. (1988).Spatial and temporal characteristics of the increase in intracelMar CaZt induced in cytotoxic T lymphocytes by cellular antigen.J. Immunol. 141,2424. 346. Alcover, A., Ramarli, D., Richardson, N. E., Chang, H.-C., and Reinherz, E. L. (1987). Functional and molecular aspects of human T lymphocyte activation via T3-Ti and T11 pathways. Immunol. REV.95,5. 347. Gelfand, E. W., Mills, G . B., Cheung, R. K., Lee, J. W. W., and Grinstein, S. (1987). Transmembrane ion fluxes during activation of human T lymphocytes: Role of Ca”, Na+/Ht exchange and phospholipid turnover. Immunol. Rev. 95,59. 348. Linch, D. C., Wallace, D. L., and O’Flynn, K. (1987).Signal transduction in human T lymphocytes. Immunol. Rev. 95, 137. 349. Hesketh, T. R., Moore, J. P., Morris, J. D., Taylor, M. V., Rogers,J., Smith, J. A., and Metcalfe, J. C. (1985). A common sequence of calcium and pH signals in the mitogenic stimulation of eukaryotic cells. Nature (London) 313, 481. 350. Mills, G. B., Cheung, R. K., Grinstein, S., and Gelfand, E. W. (1985). Increase in cytosolic free calcium concentration is an intracellular messenger for the production of interleukin 2 but not for the expression of the interleukin 2 receptor. J. Immunol. 134, 1640. 351. Gelfand, E. W., Cheung, R. K., Grinstein, S., and Mills, G. B. (1986). Characterization of the role for calcium influx in mitogen-induced triggera ing of human T cells. Identification of calcium-dependent and calciumindependent signals. Eur. J . Immunol. 16,907. 352. Whitney, R. B., and Sutherland, R. M. (1972). Requirement for calcium ions in lymphocyte transformation stimulated by phytohemagglutinin. J. Cell. Physiol. 80, 329. 353. Wilson, H. A., Greenblatt, D., Poenie, M., Finkelman, F. D., and Tsien, R. Y. (1987). Cross-linkage of B lymphocyte surface immunoglobulin by anti-Ig or antigen induces prolonged oscillation of intracellular ionized calcium. 354. Harootunian, A. T., Kao, J. P. Y., and Tsien, R. Y. (1988).Agonist-induced calcium oscillations in depolarized fibroblasts and their manipulation by photoreleased Ins( 1,4,5)Ps,Ca2+,and CaZt buffer. In “MolecularBiology of Signal Transduction,” p. 935. Cold Spring Harbor Lab., Cold Spring Harbor, New York.
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
335
355. Abraham, R.T., McKinney, M. M., Forray, C., Shipley, G. D., and Handwerger, B. S. (1986). Stimulation of arachidonic acid release and eicosanoid biosynthesis in an interleukin 2-dependent T cell line.J. Immunopharmacol. 8, 165. 356. Farrar, W. L., and Humes, J. L. (1985). The role of arachidonic acid metabolism in the activities of interleukins 1 and 2. J.Immunol. 135, 1153. 357. Jenkins, M. K., Pardoll, D. M., Mizuguchi, J., Chused, T. M., and Schwartz, R. H. (1987). Molecular events in the induction of a non-responsive state in interleukin 2-producing helper T lymphocyte clones. Proc. Natl. Acad. Sci. U.S.A.84,5409. 358. Klausner, R. D., O'Shea, J. J., Luong, H., Ross, P., Bluestone, J. A., and Samelson, L. E. (1987). T cell receptor tyrosine phosphorylation. Variable coupling for different activating ligands.J. B i d . Chem. 262, 12654. 359. Scholz, W., Isakov, N., Mally, M. E., Theofilopoulos, A. N., and Altman, A. (1988). lpr T cell hyporesponsiveness to mitogens linked to deficient receptor-stimulated phosphoinositide hydrolysis. J. Biol. Chem. 263,3626. 360. Havran, W. L., Poenie, M., Kimura, J., Tsien, R.,Weiss, A., and Allison,J. P. (1987). Expression and function of the CD3-antigen receptor on murine CD4'8+ thymocytes. Nature (London) 330, 170. 361. Finkel, T. H., McDuffie, M., Kappler, J. W., Marrack, P., and Cambier, J. C. (1987). Both immature and mature T cells mobilize Ca2' in response to antigen receptor crosslinking, Nature (London) 330, 179. 362. Gelfand, E. W., Cheung, R. K., Mills, G. B., and Grinstein, S. (1985). Mitogens trigger a calcium-independent signal for proliferation in phorbol ester-treated lymphocytes. Nature (London) 3 15,4 19. 363. Holter, W., Fischer, G. F., Majdic, O., Stockinger, H., and Knapp, W. ( 1986). T cell stimulation via the erythrocyte receptor. Synergism between monoclonal antibodies and phorbol myristate acetate without changes of free cytoplasmic Ca2' levels. J. Ex$. Med. 163,654. 364. Patel, M., Samelson, L. E., and Klausner, R.D. (1987). Multiple kinases and signal transduction. Phosphorylation of the T cell antigen receptor complex. J. Bad. Chem. 262,583 1. 365. Sussman, J. J., Mercep, M., Saito, T., Germain, R. N., Bonvini, E., and Ashwell, J. D. ( 1988). Dissociation of phosphoinositide hydrolysis and Ca2' fluxes from the biological responses of a T cell hybridoma. Nature (London) 334,625. 366. Gilman, A. G. (1987). G proteins: Transducers of receptor-generated signals. Annu. Rev. Biochem. 56, 615. 367. Gomperts, B. D. (1983). Involvement of guanine nucleotide-binding protein in the gating of Ca2+by receptors. Nature (London) 306,64. 368. Cockcroft, S., and Comperts, B. D. (1985). Role of guanine nucleotide binding protein in the activation of polyphosphoinositide phosphodiesterase. Nature (London) 314,534. 369. Smith, C. D., Lane, B. C., Kusaka, I., Verghese, M. W., and Snyderman, R. ( 1985).Chemoattractant receptor-induced hydrolysis of phosphatidylinosito1 4,5-bisphosphate in human polymorphonuclear leukocyte membranes. J. Biol. Chem. 260,5075. 370. Uhing, R. J., Prpic, V., Jiang, H., and Exton, J. H. (1986). Hormonestimulated polyphosphoinositide breakdown in rat liver plasma membranes. Roles of guanine nucleotides and calcium.J.Biol. Chem. 261,2140.
336
AMNON ALTMAN E T AL.
37 1. Straub, R. E., and Gershengorn, M. C. (1986). Thyrotropin-releasing hormone and GTP activate inositol trisphosphate formation in membranes isolated from rat pituitary cells.J. Biol. Chem. 261, 2712. 372. Baldassare, J. J., and Fisher, G. J. (1986). Regulation of membraneassociated and cytosolic phospholipase C activities in human platelets by guanosine triphosphate. J. Biol. Chem. 261, 11942. 373. Blackmore, P. F., Bocckino, S. B., Waynick, L. E., and Exton, J. H. (1985). Role of guanine nucleotide-binding regulatory protein in the hydrolysis of hepatocyte phosphophatidylinositol 4,5-bisphosphate by calcium mobilizing hormones and the control of cell ca1cium.J. Biol.Chem. 260, 14477. 374. Kikuchi, A., Kozawa, O., Kaibuchi, K., Katada, T., Ui, M., and Takai, Y. (1986). Direct evidence for involvement of a guanine nucleotide-binding protein in chemotactic peptide-stimulated formation of inositol bisphosphate and trisphosphate in differential human leukemia (HL-60) cells. J. Biol. Chem. 261, 11558. 375. Michell, B., and Kirk, C. (1986). G-protein control of inositol phosphate hydrolysis. Nature (London) 332, 112. 376. Hanley, M. R., and Jackson, T. (1987). The ras gene. Transformer and transducer. Nature (London) 328,668. 377. Fleischman, L. F., Chahwala, S. B., and Cantley, L. (1986). rm-transformed cells: Altered levels of phosphatidylinositol-4,5-bisphosphateand catabolites. Science 231,407. 378. Preiss, J., Loomis, C. R., Bishop, W. R., Stein, R., Neidel, J. E., and Bell, R. M. (1986). Quantitative measurement of sn-l,2-diacylglycerols present in platelets, hepatocytes, and ras-and sis-transformed normal rat kidney cells. J. Biol.Chem. 261,8597. 379. Wolfman, A., and Macara, I. G. (1987). Elevated levels of diacylglycerol and decreased phorbol ester sensitivity in ras-transformed fibroblasts. Nature (London) 325,359. 380. Wakelam, M. J. O., Davies, S. A., Houslay, M. D., McKay, I., Marshall, C. J., and Hall, A. (1986). Normal ~ 2 1 ~ -couples ’ ” ~ bombesin and other growth factor receptors to inositol phosphate production. Nature (London) 323, 173. 381. Lacal, J. C., de la PeAa, P., Moscat, J., Garcia-Barreno, P., Anderson, P. S., and Aaronson, S. A. (1987). Rapid stimulation of diacylglycerol production in Xen@w oocytes by micro-injection of H-ras p21. Science 238, 533. 382. Lacal, J. C., Moscat, J., and Aaronson, S.A. (1987). Novel source of 1,2diacylglycerol elevated in cells transformed by Ha-ras oncogene. Nature (London) 330,269. 383. Yu, C.-L., Tsai, M.-H., and Stacey, D. W. (1988). Cellular ras activity and phospholipid metabolism. Cell (Cambridge, Mass.) 52,63. 384. Aussel, C., Mary, D., Peyron, J.-F., Pelassy, C., Ferrua, B., and Fehlmann, M. (1988). Inhibition and activation of interleukin 2 synthesis by direct modification of guanosine triphosphate-binding proteins. J. Immunol. 140, 215. 385. Rappaport, R. S., and Dodge, G. R. (1982). Prostaglandin E inhibits the production of human interleukin 2. J . Exp. Med. 155,943. 386. Kim, D.-K., Nav, G. J., Lancki, D. W., Dawson, G., and Fitch, F. W. (1988). Cholera toxin discriminates between murine T lymphocyte proliferation stimulated by activators of protein kinase C and proliferation stimulated by IL-2. J . Immunol. 141,3429.
MOLECULAR EVENTS MEDIATING
’I’ C E I L AC’I’IVAIION
337
387. Gray, L. S., Gnarra, J., Hewlett, E. L., and Engelhard, V. H. (1988). Increased intracellular cyclic adenosine monophosphate inhibits T lymphocyte-mediated cytolysis by two distinct mechanism. J. Exp. Med. 167, 1963. 388. OShea, J. J., Urdahl, K. B., Luong, H. T., Chused, T. M., Samelson, L. E., and Klausner, R. D. (1987). Aluminum fluoride induces phosphatidylinosito1 turnover, elevation of cytoplasmic free calcium and phosphorylation of the T cell antigen receptor in murine T cells. J. Immunol. 139, 3463. 389. Takayama, H., Tren, G., and Sitkovsky, M. V. (1988). Locus of inhibitory action of cyclic AMP-dependent protein kinase in the antigen receptortriggered cytotoxic T lymphocyte activation pathway. J. Biol. Chem. 263, 2330. 390. Stewart, S. J., Prpic, U., Johns, J. A., Powers, F. S., Grager, S. E., Forbes, J. T., and Exton, J. H. (1989). Bacterial toxins affect early events of T cell activation. J. Clin. Invest. 83, 234. 391. Imboden, J . B., Shoback, D. M., Pattison, G., and Stobo, J. D. (1986). Cholera toxin inhibits the T cell antigen receptor-mediated increases in inositol trisphosphate and cytoplasmic free calcium. Proc. Natl. Acad. Sci. U.S.A. 83,5673. 392. Beals, C. R., Wilson, C. B., and Perlmutter, R. M. (1987).A small multigene family encodes Gi signal transducing proteins. Proc. Natl. Acad. Sci. U.S.A. 84,7886. 393. Morse, J. H., Kong, A. S., Lindenbaum, J., and Morse, S. I. (1977). T h e mitogenic effect of the lymphocytosis promoting factor from Bordetella pertussis on human T lymphocytes. J . Clin. Invest. 60,683. 394. Rosoff, P. M., Walker, R., and Winberry, L. (1987). Pertussis toxin triggers rapid second messenger production in human T lymphocytes. J. Immunol. 139,2419. 395. Gray, L. S., Huber, K. S., Gray, M. C., Hewlett, E. L., and Engelhard, V. H. (1989). Pertussis toxin effects on T lymphocytes are mediated through CD3 and not by pertussis toxin catalyzed modification of a G protein. J . Immunol. 142, 1631. 396. Tamura, M., Nogimori, K., Yajima, M., Ase, K., and Ui, M. (1983).A role of the B-oligomer moiety of islet-activating protein, pertussis toxin, in development of the biological effects on intact cells. J. Biol. Chem. 258,6756. 397. Pessa-Morikawa, T., Mustelin, T., and Anderson, L. C. (1989). Identification of GTP-binding proteins in T lymphocyte membranes by photoaffinity labeling with 32P-GTP. In “Cellular Basis of Immune Modulation,” G. Kaplan, R. Green, and R. C. Bleackley, eds.), p. 59. A. R. Liss, New York. 398, Madaule, P., and Axel, R. (1985). A novel ras-related gene family. Cell (Cambridge, Mass.) 41, 3 1 . 399. Chardin, P., and Tavitian, A. (1986). T h e ral gene: A new ras related gene isolated by the use of a synthetic probe. EMBO J. 5,2203. 400. Lowe, D. G., Capon, D. J., Delwart, E., Sakaguchi, A. Y., Naylor, S. L., and Goeddel, D. V. (1987). Structure of the human and murine R-ras genes, novel genes closely related to ras proto-oncogenes. Cell (Cambridge, &fuss.) 48, 13. 40 1 Touchot, N., Chardin, P., and Tavitian, A. (1987). Four additional members of the ras gene superfamily isolated by an oligonucleotide strategy: Molecular cloning of YPT-related cDNAs from a rat brain library. Proc. Natl. Acad. Sci. U.S.A. 84, 8210.
u.
338
AMNON ALTMAN ET AL.
402. Pizon, V., Chardin, P., Lerosey, I., Olofsson, B., and Tavitian, A. (1988). Human cDNAs rap1 and rap2 homologous to the Drosophilu gene Dras3 encode proteins closely related to ras in the “effector” region. Oncogene 3, 201. 403. Matsui, Y., Kikuchi, A., Kundo, J., Hishida, T., Teranishi, Y., and Takai, Y. (1988). Nucleotide and deduced amino acid sequences of a GTP-binding protein family with molecular weights of 25,000 from bovine brain. J. Biol. C h m . 263, 11071. 404. Bokoch, G. M., Parkos, C. A., and Mumby, S. M. (1988). Purification and characterization of the 22,000-dalton GTP-binding protein substrate for ADP-ribosylation by botulinum toxin, G .J. Biol. Chem. 263, 16744. 405. Narumiya, S., Sekine, A., and Fujiwara, @(1988). Substrate for botulinum ADP-ribosyltransferase, Gb, has an amino acid sequence homologous to a putative rho gene product. J.Bwl. Chem. 263, 17255. 406. Chesa, P. G., Rettig, W. A,, Melamed, M. R.,Old, L. J., and Niman, H. L. (1987). Expression of p21m in normal and malignant human tissues. Lack of association with proliferation and malignancy. Proc. Natl. Acad. Sci. U.S.A. 84,3234. 407. Mustelin, T., Poso, T., and Anderson, L. (1986). Role of G-proteins in T cell activation: Non-hydrolysable GTP analogues induce early ornithine decarboxylase activity in human T lymphocytes. EMBO J.5,3287. 408. Mustelin, T. (1987). GTP dependent of the transduction of mitogenic signals through the T 3 complex in T lymphocytes indicates the involvement of a G protein. FEBS Lett. 213, 199. 409. Sasaki, T., and Hasegawa-Sasaki, H. (1987). Activation of polyphosphoiand fluonositide phospholipase C by guanosine 5‘-0-(3-thio-triphosphate roaluminate in membranes prepared from a human T cell leukemia line, Jurkat. FEBS. Lett. 218,87. 410. Mire-Sluis, A. R.,Hoffbrand, A. V., and Wickremasinghe, R. G. (1987). Evidence that guanine nucleotide-binding regulatory proteins couple cell-surface receptors to the breakdown of inositol-containing lipids during T-lymphocyte mitogenesis. Biochem. Biophys. Res. Commun. 148, 1223. 41 1. Sommermeyer, H., Behl, B., Oberdisse, E., and Resch, K. (1989). Effects of nucleotides on the activity of phospholipase C in rabbit thymus lymphocytes. Differences in assays using endogenous [SH]-inositol-prelabeled membranes or exogenous [SH]-phosphatidylinositol 4,5-bisphosphate as substrate. J. Biol. Chem. 264,906. 412. Coggeshall, K. M., and Altman, A. (1989). Stimulation of PIP2 hydrolysis by aluminum fluoride in resting T cell subsets of normal and autoimmuneprone lpr mice. J. Immunol., 143,780. 413. Schrezenmeier, H., Anhert-Hilger, G., and Fleischer, B. (1988). Inactivation of a T cell receptor-associated GTP binding protein by antibodyinduced modulation of the T cell receptor/CD3 complex.J. Exp. Med. 168, 817. 414. Schrezenmeier, H., Anhert-Hilger, G., and Fleischer, B. (1988). A T cell receptor-associated GTP-binding protein triggers T cell receptor-mediated granule exocytosis in cytotoxic T 1ymphocytes.j. Immunol. 141,3785. 415. Krebs, E. G. (1986). The enzymology of control by phosphorylation. Enzymes 18,3.
MOLECULAR E V E N r S MEDIATING
T CELL ACTIVATION
339
416. Hanks, S. K., Quinn, A. M., and Hunter, T. (1988). The protein kinase 417. 418.
419.
420.
421.
family: Conserved features and deduced phylogeny of the catalytic domains. Science 241,42. Horenstein, A., Piras, M. M., Mordoh, J., and Piras, R. (1976). Protein phosphokinase activities of resting and proliferating human lymphocytes. Exp. Cell Res. 101, 260. Chaplin, D. D., Wedner, H. J., and Parker, C. W. (1979). Protein phosphorylation in human peripheral blood lymphocytes. I. Subcellular distribution and partial characterization of adenosine 3’,5’-monophosphatedependent protein kinase. Biochem. J. 182, 525. Johnson, E. M., Karn, J., and Allfrey, V. G. (1974). Early nuclear events in the induction of lymphocyte proliferation by mitogens. Effects of concanavalin A on the phosphorylation and distribution of non-histone chromatin proteins. J. Bwl. Chem. 249,4990. Cantrell, D. A., Friedrich, B., Davies, A. A., Gullberg, M., and Crumpton, M. J. (1989). Evidence that a kinase distinct from protein kinase C induces CD3 y-subunit phosphorylation without a concomitant down-regulation in CD3 antigen expression. J. Immunol. 142, 1626. Wedner, J. H., and Parker, C. W. (1975). Protein phosphorylation in human peripheral blood lymphocytes. Stimulation by phytohemagglutinin and N6-monobutyryl cyclic AMP. Biochem. Biophys. Res. Commun. 62,
808. 422. Chaplin, D. D., Wedner, H. J., and Parker, C. W. (1980). Protein phos-
phorylation in human peripheral blood lymphocytes. Mitogen-induced increases in protein phosphorylation in intact lymphocytes. J. Immunol. 124,
2390. 423. Gaulton, G. N., and Eardley, D. D. (1986). Interleukin 2-dependent phosphorylation of interleukin 2 receptors and other T cell membrane proteins. J. Immunol. 136,2470. 424. Inoue, M., Kishimoto, A., Takai, Y., and Nishizuka, Y. (1977). Studies on a
cyclic nucleotide-independent protein kinase and its proenzyme in mammalian tissues. 11. Proenzyme and its activation by calcium-dependent protease from rat brain. J. Biol. Chem. 252,7610. 425. Takai, Y., Kishimoto, A., Iwasa, Y., Kawahara, Y., Mori, T., and Nishizuka, Y. (1979). Calcium-dependent activation of a multifunctional protein kinase by membrane phospholipids. J. Biol. Chem. 254,3692. 426. Kaibuchi, K., Takai, Y., and Nishizuka, Y. (1981). Cooperative roles of various membrane phospholipids in the activation of calcium-activated, phospholipid-dependent protein kinase. J. Biol. Chem. 256, 7145. 427. Nishizuka, Y.(1986). Studies and perspectives of protein kinase C. Science
233,305. 428. Turner, R. S., and Kuo, J. F. (1985). Phospholipid-sensitive, Ca2+-
dependent protein kinase (protein kinase C): The enzyme, substrates, and regulation. I n “Phospholipids and Cellular Regulation” (J.F. Kuo, ed.), Vol. 2, p. 75. CRC Press, Boca Raton, Florida. 429. Nishizuka, Y. (1988). The molecular heterogeneity of protein kinase C and its implications for cellular regulation. Nature (London) 334,66 1. 430. Knopf, J. L., Lee, M.-H., Sultzman, M. A., Kriz, R. W., Loomis, C. R., Hewick, R. M., and Bell, R. M. (1986). Cloning and expression of multiple protein kinase C cDNAs. Cell (Cambridge, Mass.)46,491.
340
AMNON ALTMAN E T AI..
431. Parker, P. J., Coussens, L., Totty, N., Rhee, L., Young, S., Chen, E., Stabel, S., Waterfield, M. D., and Ullrich, A. (1986). The complete primary structure of protein kinase C-The major phorbol ester receptor. Science 233, 853. 432. Coussens, L., Parker, P. J., Rhee, L., Yang-Feng, T. L., Chen, E., Waterfield, D. M., Francke, U., and Ullrich, A. (1986). Multiple, distinct forms of bovine
and human protein kinase C suggest diversity in cellular signaling pathway. Science 233,859. 433. Ono, Y., Kikkawa, U., Ogita, K., Fujii, T., Kurokawa, T., Asaoka, Y., Sekiguchi, K., Ase, K., Igarashi, K., and Nishizuka, Y. (1987). Expression and properties of two types of protein kinase C: Alternative splicing from a single gene. Science 236, 1116. 434. Ohno, S., Yokikura, H., Sako, T., and Hidaka, H. (1987). Tissue-specific expression of three distinct types of rabbit protein kinase C. Nature (London) 325, 161. 435. Homey, G. M., O’Brian, C. A,, Johnson, M. D., Kirschmeyer, P., and Weinstein, I. B. (1987). Isolation of cDNA clones encoding protein kinase C: Evidence for a protein kinase C-related gene family. Proc. Natl. Acad. Sci. U.S.A. 84, 1065. 436. Ohno, S., Akita, Y., Konno, Y., Imajoh, S., and Suzuki, K. (1988). A novel 437. 438. 439.
440. 441.
442.
443.
phorbol ester receptor/protein kinase, nPKC, distantly related to the protein kinase C family. Cell (Cambridge, Mass.) 53,73 1. Ono, Y., Fuji, T., Ogita, K., Kikkawa, U., Igarashi, K., and Nishizuka, Y. (1988). The structure, expression, and properties of additional members of the protein kinase C. J. Biol. Chem. 263,6927. Mochly-Rosen, D., and Koshland, D. E. (1987). Domain structure and phosphorylation of protein kinase C. J. Biol. Chem. 262,2291. Kishimoto, A., Mikawa, K., Hashimoto, K., Yasuda, I., Tanaka, S., Tominaga, M., Kuroda, T., and Nishizuka, Y. (1989). Limited proteolysis of protein kinase C subspecies by calcium-dependent neutral protease (calpain). J. Biol. Chem. 264,4088. House, C., and Kemp, B. E. (1987). Protein kinase C contains a pseudosubstrate prototype in its regulatory domain. Science 238, 1726. Ogawa,Y.,Takai, Y., Kawahara,Y., Kimura,S.,and Nishizuka,Y. (1981).A new possible regulatory system for protein phosphorylation in human peripheral lymphocytes. I. Characterization of a calcium-activated, phospholipid-dependent protein kinase. J. Immunol. 127, 1369. Kuo, J. F., Anderson, R. G. G., Wise, B. C., Mackerlova, L., Salomonsson, I., Brackett, N. L., Katoh, N., Shoji, M., and Wrenn, W. R. (1980). Calciumdependent protein kinase: Widespread occurrence in various tissues and phyla of the animal kingdom and comparison of effects of phospholipid, calmodulin and trifluoperazine. Proc. Natl. Acad. Sci. U.S.A. 77, 7039. Nishizuka, Y. (1981). A new possible regulatory system for protein phosphorylation in human peripheral lymphocytes. 11. Possible relation to phosphatidylinositol turnover induced by mitogens. J. Immunol. 127,
1375. 444. Touraine, J.-L., Hadden, J. W., Tourame, F., Hadden, E. M., Estensen, R., and Good, R. A. (1977). Phorbol myristate acetate: A mitogen selective for a T lymphocyte subpopulation.J. Exp. Med. 145,460.
MOLECULAR EVENTS MEDIATING
‘1’
CE1.1. ACTIVATION
34 1
445. Abb, J., Bayliss, G. J., and Deinhardt, F. (1979). Lymphocyte activation by the tumor-promoting agent 12-0-tetradecanoyl phorbol- 13-acetate (TPA). J. Immunol. 122, 1639. 446. Mills, G. B., Girard, P., Grinstein, S., and Gelfand, E. W. (1988). Interleukin-2 induces proliferation of T lymphocyte mutants lacking protein kinase C. Cell (Cambridge, Mass.) 55,91. 447. Beyers, A. D., Hanekom, C., Rheeder, A,, Strachan, A. F., Wooten, M. W., and Nel, A. E. (1988). Characterization of protein kinase C and its isoforms in human T lymphocytes. J . Immunol. 141,3463. 448. Koretiky, G., Newton, M., Wahi, M., and Weiss, A. (1989). Heterogeneity of PKC isoenzymes in human T lymphocytes. Fed. Proc., Fed. Am. SOC.Exp. Biol. 3, A5 17. 449. Isakov, N., McMahon, P., and Altman, A. (1989). Selective post-transcriptional down-regulation of protein kinase C isoenzymes in leukemic T cells chronically treated with phorbol ester.]. Biol. Chem., in press. 449,. Shearman, M. S., Berry, N., Oda, T., Asa, K., Kikkawa, U., and Nishizuka, Y. (1988). Isolation of protein kinase C subspecies from a preparation of human T lymphocytes. FEBS Lett. 234,387. 450. Cantrell, D. A., Davis, A. A., and Crumpton, M. J. (1985). Activators of protein kinase C down-regulate and phosphorylate the TS/T-cell antigen receptor complex of human T lymphocytes. Proc. Null. Acad. Sci. U.S.A. 82, 8158. 451. Minami, Y., Samelson, L. E., and Klausner, R. D. (1987). Internalization and cycling of the T cell antigen receptor: Role of protein kinase C. J . Biol. Chem. 262, 13342. 452. Weyand, C. M., Goronzy, J., and Fathman, C. G. (1987).Modulation of CD4 by antigenic activati0n.J. Immunol. 138, 1351. 453. Solbach, W. (1982). Tumor promoting phorbol esters selectively abrogate the expression of the T 4 differentiation antigen expressed on normal and malignant (Sezary) T helper 1ymphocytes.J. Exp. Med. 156, 1250. 454. Hoxie, J. A., Matthews, D. M., Callahan. K. J., Cassel, D. L., and Cooper, R. A. (1986). Transient modulation and internalization of T 4 antigen induced by phorbol esters.J. Immunol. 137, 1194. 455. Wang, P. T. H., Bigby, M., and Sy, M.-S. (1987). Selective down modulation of L3T4 molecules on murine thymocytes by the tumor promotor, phorbol 12-myristate 13-acetate.J. Immunol. 139, 2 157. 456. Mills, G. B., May, C., Hill, M., Ebanks, R., Roifman, C., Mellors, A,, and Gelfand, E. W. (1989). Physiologic activation of protein kinase C limits IL2 secretion. J . Immunol. 142, 1995. 457. Kemp, B. E., Froscio, M., Rogers, A., and Murray, A. W. (1975). Multiple protein kinases from human lymphocytes. Identification of enzymes phosphorylating exogenous histone and casein. Biochem. J . 145,241. 458. Clark, R. B., Love, J. T., Sgroi, D., Lingenheld, E. G., and Sha’afi, R. I. (1987). The protein kinase C inhibitor H-7 inhibits antigen and IL2induced proliferation of murine T cell lines. Biochem. Biqbhys. Res. Commun. 145, 666. 459. Nel, A. E., Schabort, I., Rheeder, A., Bouie, P., and Wooten, M. (1987). Inhibition of antibodies to CD3 surface antigen and phytohemagglutininmediated T cellular responses by inhibiting Ca’+/phospholipid-dependent
342
AMNON ALTMAN ET AL.
protein kinase activity with the aid of 1-(5-isoquinolinylsulfonyl)-2-methylpiperazine dihydrochloride. J. Immunol. 139,2230. 460. Juszczak, R. J., and Russell, J. H. (1989). Inhibition of cytotoxic T lymphocyte-mediated lysis and cellular proliferation by isoquinoline sulfonamide protein kinase inhibitors. Evidence for the involvement of protein kinase C in lymphocyte function. J. Biol. Chem. 264,8 10. 461. Hidaka, H., Inagaki, M., Kawamoto, S., and Sasaki, Y. (1984). Isoquinolinesulfonamides, novel and potent inhibitors of cyclic nucleotidedependent protein kinase and protein kinase C. Biochemktry 23,5036. 462. Samelson, L. E., Davidson, W. F., Morse, H. C., 111, and Klausner, R. D. (1986). Abnormal tyrosine phosphorylation of T-cell receptor in lymphoproliferative disorders. Nature (London) 324,674. 463. Samelson, L. E., Harford, J., Schwartz, R. H., and Klausner, R. D. (1985). A 20-kDa protein associated with the murine T-cell antigen receptor is phosphorylated in response to activation by antigen or concanavalin A. Proc. Natl. Acad. Sci. U.S.A. 82, 1967. 464. Cantrell, D., Davies, A. A., Londei, M., Feldman, M., and Crumpton, M. J. (1987). Association of phosphorylation of the T 3 antigen with immune activation of T lymphocytes. Nature (London) 325,540. 465. Samelson, L. E., OShea, J. J., Luong, H., Ross, P., Urdahl, K. B., Klausner, R. D., and Bluestone, J. (1987). T cell antigen receptor phosphorylation induced by an anti-receptor antibody. J. Immunol. 139,2708. 466. Chatila, T. A., and Geha, R. S. (1988). Phosphorylation of T cell membrane proteins by activators of protein kinase C. J. Immunol. 140,4308. 467. Acres, R. B., Conlon, P. J., Mochizuki, D. Y., and Gallis, B. (1986). Rapid phosphorylation and modulation of the T 4 antigen on cloned helper T cells induced by phorbol myristate acetate or antigen. J. Biol. Chem. 261, 16210. 468. Acres, R. B., Conlon, P. J., Mochizuki, D. Y., and Gallis, B. (1987). Phosphorylation J. Immunol. 139,2268. 469. Blue, M. L., Hafler, D. A., Craig, K. A., Levine, H., and Schlossman, S. F. (1987). Phosphorylation of CD4 and CD8 molecules following T cell triggering. J. Immunol. 139,3949. 470. Hoxie, J., Rockowski,J. L., Haggarty, B. S., and Gaulton, G. N. (1988). T4 endocytosis and phosphorylation induced by phorbol esters but not by mitogen or HIV infection. J. Immunol. 140,786. 471. Autero, M., and Gahmberg, C. G. (1987). Phorbol diesters increase the phosphorylation of the leukocyte common antigen CD45 in human T cells. Eur. J. Immunol. 17, 1503. 472. Schackelford, D. A., and Trowbridge, I. S. (1986). Identification of lymphocyte integral membrane proteins as substrates for protein kinase C. Phosphorylation of the interleukin 2 receptor, class I HLA antigens, and T200 glycoprotein.J.Biol. Chem. 261, 8334. 473. Veillette, A., Horak, I. D., Horak, E. M., Bookman, M. A., and Bolen, J. B. (1988). Alterations of the lymphocyte-specific protein tyrosine kinase (pp56Ick)during T-cell activation. Mol. Cell. Biol. 8,4353. 474. Ishii, T., Takeshita, T., Numata, N., and Sugamura, K. (1988). Protein phosphorylation mediated by IL2/IL2 receptor /3 chain interaction. J. Immunol. 141, 174. 475. Leeb-Lundberg, L. M. F., Cotecchia, S., DeBlasi, A., Caron, M. G., and Lefkowitz, R. J. (1987). Regulation of adrenergic receptor function by
MOLECULAR EVENTS MEDIATING
'r CELL ACTIVATION
343
phosphorylation. I. Agonist-promoted desensitization and phosphorylation of a,-adrenergic receptors coupled to inositol phospholipid metabolism in DDTl MF-2 smooth muscle cells.J. Biol. Chem. 262,3098. 476. Bouvier, M., Hausdorff, W. P., DeBlasi, A., ODowd, B. F., Kobilka, B. K., Caron, M. G., and Lefkowitz, R. J. (1988). Removal of phosphorylation sites from the &-adrenergic receptor delays onset of agonist-promoted desensitization. Nature (London) 333,370. 477. Evans, S. W., and Farrar, W. L. (1987). Interleukin 2 and diacylglycerol stimulate phosphorylation of 40s ribosomal S6 protein. Correlation with increased protein synthesis and S6 kinase activation. J. Biol. Chem. 262, 4624. 478. Hornbeck, P., Huang, K.-P., and Paul, W.-E. (1988). Lamin B is rapidly phosphorylated in lymphocytes after activation of protein kinase C. Proc. Natl. Acad. Sci. U.S.A. 85,2279. 479. Masmoudi, A., Labourdette, G., Mersel, M., Huang, F. L., Huang, K.-P., Vincendon, G., and Malviya, M. N. (1989). Protein kinase C located in rat liver nuclei. Partial purification and biochemical and immunochemical characterization. J. Biol. Chem. 264, 1172. 480. Smith, M. R., DeGudicibus, S. J., and Stacey, D. W. (1986). Requirement for c-ras proteins during viral oncogene transformation. Nature (London) 320, 540. 481, Morrison, D. R.,Kaplan, D. R., Rapp, U., and Roberts, T. M. (1988). Signal transduction from membrane to cytoplasm: Growth factors and membranebound oncogene products increase Raf- 1 phosphorylation and associated protein kinase activity. Proc. Natl. Acad. Sci. U.S.A. 85,8855. 482. Siegel, J. N., Klausner, R. D., Rapp, U. R., and Samelson, L. E. (1989). Rapid changes in phosphorylation of c-raf following T cell activation. FASEB J. 3, A518. 483. Zwiller, J., Revel, M.-O., and Malviya, A. N. (1985). Protein kinase C catalyzes phosphorylation of guanylate cyclase in vitro.J . Biol. Chem. 260, 1350. 484. Yoshimasa, T., Sibley, D. R., Bouvier, M., Lefkowitz, R. J., and Caron, M. G. (1987). Cross-talk between cellular signalling pathways suggested by phorbol ester-induced adenylate cyclase phosphorylation. Nature (London) 327, 67. 485. Beckner, S. K., and Farrar, W. L. (1986). Interleukin 2 modulation of adenylate cyclase. Potential role of protein kinase C.J.Biol. Chem. 261,3043. 486. Stull, J. T., Nunnaly, M. H., and Michnoff, C. H. (1986). Calmodulindependent protein kinases. Enzymes 18, 113. 487. Bachvraroff, R. F., Miller, F., and Rapaport, F. T. (1984). The role of calmodulin in the regulation of human lymphocyte activation. Cell. Zmmunol. 85, 135. 488. Lee, T.-P., Venuti, J., Macara, I., Kawauchi, R., Davis, P. J., and Mookerjee, B. K. (1987). Characteristics of calmoddin binding to purified human lymphocyte plasma membranes.J. Immunol. 139,42. 489. Kincaid, R. L., Takayama, H., Billingsley, N. L., and Sitkovsky, M. V. ( 1 987). Differential expression of calmodulin-binding proteins in B,T lymphocytes and thyrnocytes. Nature (London) 330, 176. 490. Akerman, K. E. O., Honkaniemi, I., Scott, I. G., and Anderson, L. C. (1985). Interaction of Cd2+with the calmodulin activated Ca2+ Mg2+ATPase activity of human erythrocyte ghosts. Biochim. Biophys. Acta. 845,48.
344
AMNON ALTMAN E T AL.
491. Weiss, B., and Levin, R. M. (1978). Mechanism for selectively inhibiting the activation of cyclic nucleotide phosphodiesterase and adenylate cyclase by antipsychotic agents. Adv. Cyclic Nucleotide Res. 9, 285. 492. Cheung, R. K., Grinstein, S., and Gelfand, E. W. (1983). Permissive role of calcium in the inhibition of T cell mitogenesis by calmodulin antagonists. J. Immunol. 131,2291. 493. Uratsuji, Y., Nakanishi, H., Takeyama, Y., Kishimoto, A., and Nishizuka, Y. (1985). Activation of cellular protein kinase C and mode of inhibitory action of phospholipid-interacting-compounds. Biochem. Biophys. Res. Commun. 130.654. 494. Nishizuka, Y. (1984). The role of protein kinase C in cell surface signal transduction and tumour promotion. Nature (London) 308,693. 495. Kammer, G. M. (1988). The adenylate cyclase-CAMP-proteinkinase A pathway and regulation of the immune response. Immunol. Today 9,222. 496. Haddes, J. W. (1988). Transmembrane signals in the activation of T lymphocytes by mitogenic antigens. Immunol. Today 9,235. 497. Byus, C. V., Klimpel, G. R., Lucas, D. O., and Russell, D. H. (1977). Type I and type I1 cyclic AMP-dependent protein kinase as opposite effectors of lymphocyte mitogenesis. Nature (London) 268,63. 498. Dasch, J. R., and Stavitsky, A. B. (1985). Mitogen-induced phosphorylation of cytosolic proteins in rabbit T- and B-lymphocytes. Mol. Immunol. 22, 379. 499. Kammer, G. M., Boehm, C. A., Rudolph, S. A., and Schultz, L. A. (1988). Mobility of the human T lymphocyte surface molecules CD3, CD4 and CD8: Regulation by a CAMP-dependent pathway. Proc. Natl. Acad. Sci. U.S.A. 85, 792. 500. Robb, R. J., and Greene, W. C. (1987). Internalization of interleukin 2 is mediated by the p chain of the high-affinity interleukin 2 recept0r.j. Exp. Med. 165, 1201. 501. Farrar, W. L., Evans, S. W., Rapp, U. R., and Cleveland,J. L. (1987). Effects of anti-proliferative cyclic AMP on interleukin 2-stimulated gene expression.J , Immunol. 139,2075. 502. Lowenthal, J. W., Bohnlein, E., Ballard, D. W., and Greene, W. C. (1988). Regulation of interleukin 2 receptor a subunit (Tac or CD25 antigen) gene expression: Binding of inducible nuclear proteins to discrete promoter sequences correlates with transcriptional activation. Proc. Natl. Acad. Sci. U.S.A. 85,4468. 503. Baker, P. E., Fahey, J. V., and Munck, A. (1981). Prostaglandin inhibition of T cell proliferation is mediated at two levels. Cell. Immunol. 61, 52. 504. Chouaib, S., and Fradelizi, D. (1982). The mechanism of inhibition of human IL2 production. J. Immunol. 129,2463. 505. Goodwin,J. S., Bankhurst, A. D., and Messner, R. P. (1977). Suppression of T cell mitogenesis by prostaglandin. J. Exp. Med. 146, 1719. 506. Curtis, B. M., and Catterall, W. A. (1985). Phosphorylation of the calcium antagonist receptor of the voltage-sensitive calcium channel by CAMPdependent protein kinase. Proc. Natl. Acad. Sci. U.S.A. 82,2528. 507. Supattapone, S., Danoff, S. K., Theibert, A., Joseph, S. K., Steiner, J., and Snyder, S. H. (1988). Cyclic AMP-dependent phosphorylation of a brain inositol trisphosphate receptor decreases its release of calcium. Proc. Natl. Acad. Sci. U.S.A. 85,8748.
MOLECULAR EVENTS MEDIATING
'I'
CELL A<:TIVATION
345
508. Johnson, K. W., Davis, B. H., and Smith, K. A. (1988). CAMP antagonizes IL2-promoted T cell cycle progression at a discrete point in early G I . Proc. Natl. Acad. Sci. U.S.A. 85,6072. 509. Marth, J. D., Peet, R., Krebs, E. G., and Perlmutter, R. M. (1985). A lymphocyte-specific protein-tyrosine kinase is rearranged and overexpressed in the murine T cell lymphoma LSTRA. Cell (Cambridge, Mass.)43, 393. 510. Voronova, A. F., and Sefton, B. M. (1986). Expression of a new tyrosine protein kinase is stimulated by retrovirus promoter insertion. Nature (London) 319,682. 511. Marth, J. D., Disteche, C., Pravtcheva, D., Ruddle, F., Krebs, E. G., and Perlmutter, R. M. (1986). Localization of a lymphocyte-specific protein kinase gene (lck) at a site of frequent chromosomal abnormalities in human lymphomas. Proc. Natl. Acad. Sci. U.S.A. 83,7400. 512. Fischer, S., Fagard, R., Piau, J.-P,, Genetet, N., Blaineau, C., Reibel, L., LePrise, Y., and Gacon, G. (1985).Acute myeloblastic leukemia with active tyrosine protein kinase. L e d . Res. 9, 1345. 513. Marchildon, G. A., Casnellie, J. E., Walsh, K. A., and Krebs, E. G. (1984). Covalently bound myristate in a lymphoma tyrosine protein kinase. Proc. Natl. Acad. Sci. U.S.A. 81, 7679. 514. Voronova, A. F., Buss, J. E., Patschinsky, T., Hunter, T., and Sefton, B. M. (1984). Characterization of the protein apparently responsible for the elevated tyrosine protein kinase activity in LSTRA cells. Mol. Cell. Biol. 4, 2705. 515. Marth, J. D., Lewis, D. B., Wilson, C. B., Gearn, M. E., Krebs, E. G., and Perlmutter, R. M. (1987). Regulation of pp56Ick during T-cell activation: Functional implications for the src-like protein tyrosine kinases. EMBOJ. 6, 2727. 516. Casnellie, J. E. (1987). Sites of in viuo phosphorylation of the T cell tyrosine protein kinase in LSTRA cells and their alteration by tumor-promoting phorbol esters. J . Biol. Chem. 262,9859. 517. Casnellie, J. E., and Lamberts, R. J. (1986). Tumor promoters cause changes in the state of phosphorylation and apparent molecular weight of a tyrosine protein kinase in t lymphocytes. J. Bid. Chem. 261,4921. 518. Casnellie, J. E., Harrison, M. L., Pike, L. J., Hellstrom, K. E., and Krebs, E. G. (1982). Phosphorylation of synthetic peptides by a tyrosine protein kinase from the particulate fraction of a lymphoma cell line. Proc. Natl. Acad. Sci. U.S.A. 79,282. 519. Marth, J. D., Cooper, J. A., King, C. S., Ziegler, S. F., Tinker, D. A., Overell, R. W., Krebs, E. G., and Perlmutter, R. (1988). Neoplastic transformation induced by a n activated lymphocyte-specific protein tyrosine kinase (pp56'"). Mol. Cell. Biol. 8, 540. 520. Amrein, K. E., and Sefton, B. M. (1988). Mutation of a site of tyrosine phosphorylation in the lymphocyte-specific tyrosine protein kinase, pp56Ick, reveals the oncogene potential in fibroblasts. Proc. Natl. Acad. Sci. U.S.A. 85, 4247. 521. Strebhardt, K., Mullins, J. I., Bruck, C., and Rubsamen-Weigmann, H. ( 1 988). Additional member of the protein-tyrosine kinase family: T h e src-and lck-related protooncogene c-tkl. Proc. Natl. Acad. Sci. U.S.A. 84, 8778.
346
AMNON ALTMAN E T AL.
522. Kawakami, T., Kawakami, Y., Aaronson, S. A., and Robbins, K. C. (1988). 523. 524. 525. 526. 527. 528.
Acquisition of transforming properties by fyn, a normal src-related human gene. Proc. Natl. Acad. Sci. U.S.A. 85,3870. Helderman, J. H., and Strom, T. B. (1978). Specific insulin binding site on T and B lymphocytes as a marker of cell activation. Nature (London) 274,62. Ben-Neriah, Y., and Bauskin, A. R. (1988). Leukocytes express a novel gene encoding a putative transmembrane protein-kinase devoid of an extracellular domain. Nature (London) 333,672. Renshaw, M. W., Capozza, M. A., and Wang, J. Y. J. (1988). Differential expression of type-specific c-abl mRNAs in mouse tissues and cell lines. Mol. Cell. Biol. 8, 4547. Majumder, S., Brown, K., Qui, F.-H., and Besmer, P. (1988). C-kit protein, a transmembrane kinase: Identification in tissues and characterization. Mol. Cell. Biol. 8,4896. Takahashi, M., and Cooper, G. M.(1987). ret-Transforming gene encodes a fusion protein homologous to tyrosine kinases. Mol. Cell. Biol. 7, 1378. Mally, M. I., Vogt, M., Swift, S. E., and Haas, M. (1985). Oncogene expression in murine splenic T cells and in murine T-cell neoplasms. Virology 144,
115. 529. Telerman, A., Amson, R., Zant-Houri, R., and Givol, D. (1988). Identifica-
tion of the humanpim-1 gene product as a 33-kilodalton cytoplasmic protein with tyrosine kinase activity. Mol. Cell. Biol. 8, 1498. 530. Selten, G., Guypers, H. T., Boelens, W., Robannus-Maandag, E., Verbeek. J., Domen, J., van Beveren, C., and Berns, A. (1986). The primary structure of the putative oncogene pim-1 shows extensive homology with protein kinases. Cell (Cambridge, Mass.)46,603. 53 1. Geahlen, R. L., and Harrison, M.L. (1986). Detection of a novel lymphocyte protein-tyrosine kinase by renaturation in polyacrylamide gels. Biochem. Biophys. Res. Commun. 134,963. 532. Zioncheck, T. F., Harrison, M. L., and Geahlen, R. L. (1986). Purification and characterization of a protein-tyrosine from bovine thymus. J. Biol.
Chem. 261,15637. 533. Kong, S. K., and Wang, J. H. (1987). Purification and characterization of a protein tyrosine kinase from bovine sp1een.J. Biol. C h m . 262,2597. 534. Earp, H. S., Austin, K. S., Buessow, S. C., Dy, R., and Gillespie, C. Y. (1984).
535. 536. 537.
538.
Membranes from T and B lymphocytes have different patterns of tyrosine phosphorylation. Proc. Natl. Acad. Sci. U.S.A. 81,2347. Mire, A. R., Wickremasinghe, R. G., and Hoffbrand, A. V. (1986). Mitogen treatment of permeabilized human T lymphocytes stimulates rapid tyrosine and serine phosphorylation of a 42 kDa protein. FEBS Lett. 206,53. Wedner, H. J., and Bass, g. (1986). Tyrosine phosphorylation of a 66,000 Mr soluble protein in lectin-activated human peripheral blood T lymphocytes.J. Immunol. 136,4226. Piga, A., Wickremasinghe, R. G., Taheri, M. R., Yaxley, J. C., and Hoffbrand, A. V. ( 1985). Phytohemagglutinin-induced changes in tyrosine protein kinase and its endogenous substrates in human lymphocytes. Exp. Cell Res. 159, 103. Cooper, J. A., Sefton, B. M., and Hunter, T. (1983). Detection and quantitation of phosphotyrosine in proteins. In “Methods in Enzymology” (J. D. Corbin and J. G. Hardman, eds.), Vol. 99, p. 387. Academic Press, New York.
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
347
539. Frackelton, A. R., Ross, A. N., and Eisen, H. N. (1983). Characterization and use of monoclonal antibodies for isolation of phosphotyrosyl proteins from retrovirus-transformed cells and growth factor-stimulated cells. Mol. Cell. Biol. 3, 1343. 540. Wang, J. Y. J. (1985). Isolation of antibodies for phosphotyrosine by immunization with a v-abl oncogene-encoded protein. Mol. Cell. Biol. 5,3640. 54 1. Hirata, F., Matsuda, K., Notsu, Y., Hattori, T., and del Carmine, R. (1984). Phosphorylation at a tyrosine residue of lipomodulin in mitogen-stimulated murine thymocytes. Proc. Natl. Acad. Sci. U.S.A. 81, 4717. 542. Toth, B., Bollen, M., and Stalmans, W. (1988). Acute regulation of hepatic protein phosphatases by glucagon, insulin and glucose.J. Biol. Chem. 263, 1406 1. 543. Alexander, D. R., and Cantrell, D. A. (1989). Kinases and phosphatases in T cell activation. Immunol. Today 10, 200. 544. Alexander, D. R., Hexham, J. M., and Crurnpton, M. J . (1988). The association of type 1, type 2A and type 2B phosphatases with the human T lymphocyte plasma membrane. Biochem.J. 256,885. 545. Trowbridge, I. S., Ralph, P., and Bevan, M. J. (1975). Differences in the surface proteins of mouse B and T cells. Proc. Null. Acad. Sci. U.S.A.72, 157. 546. Charbonneau, H., Tonks, N. K., Walsh, K. A., and Fischer, E. H. (1988). The leukocyte common antigen (CD45).A putative receptor-linked protein tyrosine phosphatase. Proc. Natl. Acad. Sci. U.S.A. 85, 7182. 547. Tonks, N. K., Charbonneau, H., Diltz, C. D., Fischer, E. H., and Walsh, K. A. (1988). Demonstration that the leukocyte common antigen CD45 is a protein tyrosine phosphatase. Biochemistry 27, 8695. 547a. Mustelin, T., and Altman, A. (1990). Manuscript in preparation. 548. Ledbetter, J. A., Tonks, N. K., Fischer, E. H., and Clark, E. A. (1988). CD45 regulates signal transduction and lymphocyte activation by specific association with receptor molecules on T or B cells. Proc. Natl. Acad. Sci. U.S.A. 85, 8628. 549. Mustelin, T., Coggeshall, K. M., and Altman, A. (1989). Rapid activation of the T cell tyrosine protein kinase pp56"' by the CD45 phosphotyrosine phosphatase. Proc. Natl. Acad. Sci. U.S.A. 86,6302. 54%. Ostergaard, H. L., Shackelford, D. A., Hurley, T. R., Johnson, P., Hyman, R., Sefton, B. M. and Trowbridge, F. S. (1989). Expression of CD45 alters phosphorylation of the kk tyrosine protein kinase in murine lymphoma T cell lines. Proc. Natl. Acad. Sci. U.S.A. 86,8959. 549b. Pingel, J. T. and M. L. Thomas. (1989). Evidence that the leukocyte common antigen is required for antigen-induced T lymphocyte proliferation. Cell 58, 1055. 549c. Hunter, T. (1989). Protein-tyrosine phoshatases: The other side of the protein-tyrosine kinase coin. Cell 58, 1013. 550. Bashford, C. L., and Pasternak, C. A. (1984). Plasma membrane potential of LettrC cells does not depend on cation gradients but on pumps. J. Membr. Biol. 79, 275. 551. Kiefer, H., Blume, A. J., and Kaback, H. R. (1980). Membrane potential changes during mitogenic stimulation of mouse spleen lymphocytes. Proc. Natl. Acad. Sci. U.S.A. 77, 2200. 552. Freedman, M. H., Raff, M. C., and Gomperts, B. (1975). Induction of increased calcium uptake in mouse T lymphocytesby concanavalin A and its modulation by cyclic nucleotides. Nature (London) 255,378.
348
AMNON ALTMAN E T AL.
553. Luckasen, J. R., White, J. G., and Kersey, J. H. (1974).Mitogenic properties of a calcium ionophore, A23187. Proc. Natl. Acad. Sci. U.S.A. 71,5088. 554. Akerman, K. E. O., and Andersson, L. C. (1984). Direct mitogenic effect of ionophore A23 187 on isolated human helper lymphocytes. Eur.J. Immunol. 14,286. 555. Wolff, C. H. I., Akerman, K. E. O., and Andersson, L. C. (1985).Kinetics of long-term (73 hr) calcium content during mitogen activation of cultured human T lymphocytes. J. Cell. Physiol. 123,46. 556. Kuno, M., Goronzy, J., Weyand, C. M., and Gardner, P. (1986). Singlechannel and whole-cell recordings of mitogen-regulated inward currents in human cloned helper T lymphocytes. Nature (London) 323,269. 557. Kuno, M., and Gardner, P. (1987). Ion channels activated by inositol 1,4,5trisphosphate in plasma membrane of human T-lymphocytes. Nature (London) 362,301. 558. Michell, R. (1986). A second messenger function for inositol tetrakisphosphate. Nature (London) 324,613. 559. Morris, A. P., Gallacher, D. V., Irvine, R. F., and Petersen, 0. H. (1987). Synergism of inositol trisphosphate and tetrakisphosphate in activating Ca2'-dependent K t channels. Nature (London) 330,653. 560. Gray, L. S., Gnarra, J. R., Russel, J. H., and Engelhard, V. H. (1987). The role of Kt in the regulation of the increase in intracellular Ca" mediated by the T lymphocyte antigen receptor. Cell (Cambridge, Mass.) 50, 119. 561. DeCoursey, T. E., Chandy, K. G., Gupta, S., and Cahalan, M. D. (1985). Voltage-dependent ion channels in T lymphocytes. J. Neuroimnwnol. 10, 102. 562. Cahalan, M. D., Chandy, K. G., DeCoursey, T. E., and Gupta, S. (1985). A voltage-gated Kt channel in human T lymphocytes.J. Physiol. (London)358, 197. 563. Gelfand, E. W., Cheung, R. K., and Grinstein, S. (1986). Mitogen-induced changes in CaZt permeability are not mediated by voltage-gated Kt channe1s.J. B i d . Chem. 261, 11520. 564. Ishida, Y., and Chused, T. M. (1988). Heterogeneity of lymphocyte calcium metabolism is caused by T cell-specific calcium-sensitive potassium channel and sensitivity of the calcium ATPase pump to membrane potential. J. Exp. Med. 168,839. 565. Bregetovski, P., Redkozubov, A., and Alexeev, A. (1986). Elevation of intracellular calcium reduces voltage-dependent potassium conductance in human T cells. Nature (London) 319,776. 566. DeCoursey, T. E., Chandy, K. G., Gupta, S., and Cahalan, M. D. (1987). Two types of potassium channels in murine T lymphocytes. J. Gen. Physiol. 89,379. 567. Lewis, R. S., and Cahalan, M. D. (1988). Subset-specific expression of potassium channels in developing murine T lymphocytes. Science 239, 771. 568. Chandy, K. G., DeCoursey, T. E., Fischbach, M., Talal, N., Cahalan, M. D., and Gupta, S. (1986). Altered K' channel expression in abnormal T lymphocytes from mice with the lpr gene mutation. Science 233, 1197. 569. Grissmer, S., Cahalan, M. D., and Chandy, K. G. (1988).Abundant expression of type 1 K+ channels, a marker for lymphoproliferative diseases? J. Immunol. 141, 1137.
MOLECULAR EVENTS M E D I A T I N G
‘1’ CELL A C T I V A T I O N
349
570. DeCoursey, T. E., Chandy, K. G., Gupta, S., and Cahalan, M. D. (1984). Voltage-gated K t channels in human T lymphocytes: A role in mitogenesis? Nature (London) 307,465. 571. Schell, S. R., Nelson, D. J,, Fozzard, H. A., and Fitch, F. W. (1987). T h e inhibitory effects of K t channel-blocking agents on T lymphocyte proliferation and lymphokine production are “nonspecific.” J. Immunol. 139, 3224. 572. Moolenaar, W. H., Tertoolen, L. G. J., and de Laat, S. W. (1984). Phorbol ester and diacylglycerol mimic growth factors in raising cytoplasmic pH. Nature (London) 3 12,37 I. 573. Swann, K., and Whitaker, M. (1985).Stimulation of the N a f / H t exchanger of sea urchin eggs by phorbol ester. Nature (London) 314, 274. 574. Rosoff, P. M., and Cantley, L. C. (1985). Stimulation of the T3-T cell receptor-associated Ca2‘ influx enhances the activity of the Na+/Ht exchanger in a leukemic human T cell line. J . B i d . Chem. 260, 14053. 575. Mills, G. B., Cragoe, E. J., Gelfand, E. W., and Grinstein, S. (1985).Interleukin 2 induced a rapid increase in intracellular pH through activation of a Nat/H+ antiport. Cytoplasmic alkaiinization is not required for lymphocyte proliferation. J. B i d . Chem. 260, 12500. 576. Mills, G. B., and May, C. (1987). Binding of interleukin 2 to its 75 kDa intermediate affinity receptor is sufficient to activate Nat/Ht exchange. J . Immunol. 139,4083. 577. Rink, T. J., Tsien, R. Y., and Pozzan, T. (1982).Cytoplasmic pH and free Mg2+ in lymphocytes. J. Cell Biol. 95, 189. 578. Rogers,.J., Hesketh, T. R., Smith, G. A., and Metcalfe, J. C. (1983).IntracelM a r pH of stimulated thymocytes measured with a new fluorescent indicator. J. B i d . Chem. 258,5994. 579. Rozengurt, E. (1986). Early signals in the mitogenic response. Science 234, 161. 580. L’Allemain, G., Franchi, A., Cragoe, E., Jr., and Pouyssegur, J . (1984). Blockade of the Nat/Ht antiport abolishes growth factor-induced DNA synthesis in fibroblasts. J.Biol.Chem. 259,4313. 581. Pouyssegur, J., Sardet, C., Franchi, A., L’Allemain, G., and Paris, S. (1984). A specific mutation abolishing Na+/Ht antiport activity in hamster fibroblasts precludes growth at neutral and acidic pH. Proc. Natl. Acad. Sci. U.S.A. 81,4833. 582. Mills, G. B., Cheung, R. K., Cragoe, E. J.,Jr., Grinstein, S., and Gelfand, E. W. (1986).Activation of the Na+/Ht antiport is not required for lectininduced proliferation of human T lymphocytes. J.Zmmunol. 136, 1150. 583. Yamaguchi, D. T., Sakai, R., Bahn, L., Cragoe, E. J., Jr., and Jordan, S. C. (1986). Amiloride inhibition of DNA synthesis and immunoglobulin production by activated human peripheral blood mononuciear cells is independent of sodium/hydrogen antiport. J. Immunol. 137, 1300. 584. Sardet, C., Franchi, A., and Pouyssegur, J. (1989). Molecular cloning, primary structure, and expression of the human growth factor-activatable Nat/Ht antiporter. Cell (Cambridge, Mass.)56,271. 585. Fillingame, R. H., Jorstad, C. M., and Morris, D. R. (1975). Increased cellular levels of spermidine and spermine are required for optimal DNA synthesis in lymphocytes activated by concanavalin A. Proc. Natl. Acad. Sci. U.S.A. 72,4042.
350
AMNON ALTMAN ET A L .
586. Holtta, E., Janne, J., and Hovi, T. (1979). Suppression of the formation of
587. 588. 589. 590. 591. 592.
polysomes and macromolecules by m-a-difluoromethylornithine and methylglyoxal bis(guany1hydrazone) in phytohaemagglutinin-activatedhuman lymphocytes. Biochem. J. 178, 109. Holtta, E., and Hovi, T. (1985). Polyamine depletion results in impairment of polysome formation and protein synthesis before onset of DNA synthesis in mitogen-activated human lymphocytes. Eur. J. Zmmunol. 152,229. Pohjanpelto, P., and Knuutila, S. (1982). Polyamine deprivation causes major chromosome aberrations in a polyamine-dependent Chinese hamster ovary cell line. Exp. Cell Res. 141,333. Pohjanpelto, P., and Knuutila, S. (1984). Induction of major chromosome aberrations in Chinese hamster ovary cells by a-difluoromethylornithine. Cancer Res. 44, 4535. Pohjanpelto, P., Virtanen, I., and Holtta, E. (1981). Polyamine starvation causes disappearance of actin filaments and microtubules in polyamineauxotrophic CHO cells. Nature (London) 293,475. Nickerson, J. A., and Wells, W. W. (1986). Polyamine regulation of the microtubule-associated protein kinase. J. Neurochem. 46, 112. Vogel, S., and Hoppe, J. (1986). Polyamines stimulate the phosphorylation of phosphatidylinositol in membranes from A431 cells. Eur.J.Biochem. 154,
253. 593. Verma, R., and Chen, K. Y. (1986). Spermine inhibits the phosphorylation of the 11,000 and 10,000-dalton nuclear proteins catalyzed by nuclear protein kinase NI in NB-15 mouse neuroblastoma cells.J. Biol. Chem. 261, 2890. 599a. Tonks, N. K., Diltz, C. D,, and Fischer, E. H. (1988). Characterization of
the major protein-tyrosine-phosphatassof human placenta. J. Biol. Chem.
263,673 1. 594. Folk, J. E., Park, M.H., Chung, S. I., Schrode, J., Lester, E. P., and Cooper, H. L. (1980). Polyamines as physiological substrates for transglutaminases. J. Biol. C h m . 255,3695. 595. Novogrodsky, A., Quittner, S., Rubin, A. L., and Stenzel, K. H. (1978). 596. 597. 598. 599. 600. 601.
Transglutaminase activity in human lymphocytes: early activation by phytomitogens. Proc. Nad. Acad. Sci. U.S.A. 75, 1157. Janne, J., Pas6, H., and Raina, A. (1978). Polyamines in rapid growth and cancer. Biochim. Btophys. Acta 473,241. Pegg, A. E. (1986). Recent advances in the biochemistry of polyamines in eukaryocytes. Bi0chem.J. 234,249. Klein, D., and Morris, D. R. (1978). Increased arginase activity during lymphocyte mitogenesis. B i o c h . Biophys. Res. Commun. 81, 199. Haddox, M. K., Magun, B. E., and Russell, D. H. (1980). Ornithine decarboxylase induction during GIprogression of normal and Rous sarcoma virus-transformed cells. Cancer Res. 40,604. Sistonen, L., Keski-Oja, J., Ulmanen, I., Holtta, E., Wikgren, B.-J. P., and Alitalo, K. (1987). Dose effects of transfected c-Ha-ras"a"2 oncogene in transformed cell clones. Exfi. Cell Res. 168,5 18. Russell, D. H., and Snyder, S. H. (1969). Amine synthesis in regenerating rat liver: Extremely rapid turnover of ornithine decarboxylase. Mol. Pharmacol. 5,253.
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
35 1
602. Seely, J. E., and Pegg, A. E. (1983). Changes in mouse kidney ornithine
decarboxylase activity are brought about by changes in the amount of enzyme protein as measured by radioimmunoassay. J , Biol. Chem. 258,
2496. 603. Heller, J. S., Fong, W. F., and Canellakis, E. S. (1976). Induction of a protein inhibitor to ornithine decarboxylase by the end products of its reaction. Proc. Natl. Acad. Sci. U.S.A. 73, 1858. 604. Kitani, T., and Fujisawa, H. (1984). Purification and some properties of a
protein inhibitor (antizyme) of ornithine decarboxylase from rat liver. J. Biol. Chem. 259, 10036. 605. Murakami, Y., Fujita, K., Kameji, T., and Hayashi, S.-I. (1985). Accumulation of ornithine decarboxylase-antizyme complex in HMOAcells. Biochem.J. 225,689. 606. Murakami, Y., and Hayashi, S.-I. (1985). Role of antizyme in degradation of ornithine decarboxylase in HTC cells. Biochem. J. 226,893. 607. Kahana, C., and Nathans, D. (1985). Translational regulation of mammalian ornithine decarboxylase by polyamines. J . Biol. Chem. 260, 15390. 608. HBltta, E., and Pohjanpelto, P. (1986). Control of ornithine decarboxylase in Chinese hamster ovary cells by polyamines. J. Biol. Chem. 261,9502. 609. Mustelin, T., Pessa, T., Lapinjoki, S., Gynther,J., Jarvinen, T., Eloranta, T., and Andersson, L. C. (1989). In “Progress in Polyamine Research” (V. Zappia, ed.), p. 301. Plenum, New York. 610. Scott, I. G., Poso, H., Akerman, K. E. O., and Andersson, L. C. (1984). Mitogens cause a rapid induction of ornithine decarboxylase activity in human T-lymphocytes. Biochem. SOC.Trans. 13,934. 611. Scott, I. G., Poso, H., Akerman, K. E. O., and Andersson, L. C. (1985). Rapid activation of ornithine decarboxylase by mitogenic (but not by nonmitogenic) ligands in human T lymphocytes. Eur. J. Immunol. 15,783. 612. Mustelin, T., P6s6, H., Ilvanainen, A., and Andersson, L. C. (1986). myoInositol reverses Li’-induced inhibition of phosphoinositide turnover and ornithine decarboxylase induction during early lymphocyte activation. Eur. J. Immunol. 16,859. 613. Farrar, W. L., Vinocour, M., Cleveland, J. L., and Harel-Bellan, A. (1988). Regulation of ornithine decarboxylase activity by IL2 and cyclic AMP. J. Immunol. 141,967. 614. Gerner, E. W., Blass, J. R., and Fuller, D. J. M. (1985). Activation of ornithine decarboxylase in monolayer cells treated with trypsin. J. Cell Physiol. 125, 567. 615. Koenig, H., Goldstone, A. D., and Lu, C. (1983). Polyamines regulate calcium fluxes in a rapid plasma membrane response. Nature (London) 305, 530. 616. Kay, J. E., Lindsay, V. J., and Cooke, A. (1972). Ornithine decarboxylase in
phytohaemagglutinin stimulated lymphocytes: Control of degradation rate by amino acids. FEES. Lett. 21, 123. 617. Kay, J. E., and Lindsay, V. J. (1973). Polyamine synthesis during lymphocyte activation. Exp. Cell Res. 77,428. 618. Fidelus, R. K., Laughter, A. H., and Twomey, J. J. (1984). The role of mitogens and lymphokines in the induction of ornithine decarboxylase (ODC) in T lymphocytes.J. Immunol. 132, 1462.
352
AMNON ALTMAN E T AL.
619. Berger, F. G., Szymanski, P., Read, E., and Watson, G. (1984). Androgenregulated ornithine decarboxylase mRNAs of mouse kidney. J. Biol. Chem. 259,794 1 . 620. Sasakawa, N., Ishii, K., Yamamoto, S., and Kato, R. (1985). Induction of ornithine decarboxylase activity by I-oleyl-2-acetyl-glycerol in isolated mouse epidermal cells. Biochem. Biophys. Res. Commun. 128,913. 621. Jetten, A. M., Ganona, B. R., Vandenbark, G. R., Shirley, J. E., and Bell, R. M. (1985). Role of protein kinase C in diacylglycerol-mediated induction of ornithine decarboxylase and reduction of epidermal growth factor binding. Proc. Natl. Acad. Sci. U.S.A. 82, 1941. 622. Bloom, W., and Fawcett, D. W. (1975).“A Textbook of Histology,” 10th Ed. Saunders, Philadelphia. 623. Irving, S. G., June, C. H., Zipfel, P. F., Siebenlist, U., and Kelly, K. (1989). Mitogen-induced genes are subject to multiple pathways of regulation in the initial stages of T-cell activation. Mol. Cell. Biol. 9, 1034. 624. Zipfel, P. F., Irving, S. G., Kelly, K., and Siebenlist, U. (1989).Complexity of the primary genetic response to mitogenic activation of human T cells. Mol. Cell. Biol. 9, 1041 . 625. Lipes, M. A., Napolitano, M., Jeang, K.-T., Chang, N. T., and Leonard, W. J. (1988). Identification, cloning and characterization of an immune activation gene. Proc. Natl. Acad. Sci. U.S.A. 85,9704. 626. Heldin, C.-H., and Westermark, B. (1984). Growth factors: Mechanism of action and relation to oncogenes. Cell (Cambridge, Mass.) 37,9. 627. Curran, T., Miller, A. D., Sokas, L., and Verma, I. M. (1984). Viral and cellularfos proteins: A comparative analysis. Cell (Cambridge, Mass.)36, 159. 628. Franza, B. R., Rauscher, F. J., 111,Josephs, S. F., and Curran, T. (1988).T h e Fos complex related antigens recognize sequence elements that contain AP-1 binding sites. Science 239, 1150. 629. Angel, P., Allegretto, E. A., Okino, S. T., Hattori, K., Boyle, W. J., Hunter, T., and Karin, M. (1988). Oncogenejun encodes a sequence-specific tramactivator similar to AP- 1 . Nature (London)332, 166. 630. Reed, J. C., Alpers, J. D., Nowell, P. C., and Hoover, R. G. (1986). Sequential expression of protooncogenes during lectin-stimulated mitogenesis of normal human lymphocytes. Proc. Null. Acad. Sci. U.S.A. 83,3982. 631. Grausz, J. D., Fradelizi, D., Dautry, F., Monier, R., and Lehn, P. (1986). Modulation of c-fos and c-myc mRNA levels in normal human lymphocytes by calcium ionophore A23187 and phorbol ester. Eur. J. Immunol. 16, 1217. 632. Moore, J. P., Todd, J. A., Hesketh, T. R., and Metcalfe, J. C. (1986).c-fos and c-myc gene activation, inonic signals, and DNA synthesis in thymocytes. J.Biol. Chem. 261,8158. 633. Cleveland, J. L., Rapp, U. R., and Farrar, W. L. (1987). Role of c-myc and other genes in interleukin 2-regulated CT6 lymphocytes and their malignant variants. J.Immunol. 138,3495. 634. Stern, J. B., and Smith, K. A. (1986). Interleukin-2 induction of T-cell G I progression and c-myb expression. Science 233,203. 635. Studzinski, G. P., Brelvi, Z. S., Feldman, S. C., and Watt, R. A. (1986). Participation of c-myc protein in DNA synthesis of human T cells. Science 234,467.
MOLECULAK EVENTS MEDIATING
T CELL ACTIVATION
353
636. Kelly, K., Cochran, B. H., Stiles, C. D. and Leder, P. (1983). Cell-specific 637. 638. 639.
640. 641.
642.
643.
regulation of the c-myc gene by lymphocyte mitogens and platelet-derived growth factor. Cell (Cambridge, Mass.) 35,603. Reed, J. C., Sabath, D. E., Hoover, R. G., and Prystowsky, M. B. (1985). Recombinant interleukin 2 regulates levels of c-myc mRNA in a cloned murine T lymphocyte. Mol. Cell. Biol. 5, 3361. Kronke, M., Leonard, W. J., Depper, J. M., and Greene, W. C. (1985). Sequential expression of genes involved in human T lymphocyte growth and differentiation. J. Exp. Med. 161, 1593. Katagiri, K., Katagiti, T., Yokota, S., Eisenberg, R. A,, Cohen, P. L., and Ting, J. P. (1988). Two different molecular pathways account for low IL2 receptor and c-myc mRNA expression by lpr Lyt-2- L3T4- T cells.J. Immunol. 141, 2145. Reed, J. C., Nowell, P. C., and Hoover, R. G. (1985). Regulation of c-myc mRNA levels in normal human lymphocytes by modulators of cell proliferation. Proc. Natl. Acud. Sci. U.S.A. 82,4221. Schneider-Schaulies, J., Hunig, T., Schimpl, A., and Wecker, E. (1986). Kinetics of cellular oncogene expression in mouse lymphocytes. I. Expression of c-myc and c-rasHain T lymphocytes induced by various mitogens. Eur. J . Immunol. 16,3 12. Torelli, G., Selleri, L., Donelli, A., Ferrari, S., Emilia, G., Venturelli, D., Moretti, L., and Torelli, U. (1985). Activation of c-myb expression by phytohemagglutinin stimulation in normal human T lymphocytes. Mol. Cell. Biol. 5, 2874. Heikkila, R., Schwab, G., Wickstrom, E., Loke, S. L., Pluznik, D. H., Watt, R., and Neckers, L. M. (1987). A c-myc antisense oligodeoxynucleotide inhibits entry into S phase but not progress from Go to GI. Nature (London)
328,445. 644. Harel-Bellan, A., Ferris, D. K., Vinocour, M., Holt, J. T., and Farrar, W. L. ( 1988). Specific inhibition of c-myc biosynthesis using an antisense synthetic deoxy-oligonucleotide in human T lymphocyte.J. Immunol. 140,243 1. 645. Buckler, A. J., Vie, H., Sonenshein, G. E., and Miller, R. A. (1988). Defec-
646. 647. 648. 649. 650.
tive T lymphocytes in old mice. Dimimished production of mature c-myc RNA after mitogen exposure not attributable to alterations in transcription or RNA stabi1ity.J. Immunol. 140,2442. Gonda, T. J., and Metcalf, D. (1984). Expression of myb, myc and fos protooncogenes during the differentiation of a murine myeloid leukaemia. Nature (London) 310,249. Thompson, C. B., Challoner, P. B., Neiman, P. E., and Groudine, M. (1986). Expression of the c-myb proto-oncogene during cellular proliferation. Nuture (London 319, 374. Gewirtz, A. M., and Calabretta, B. (1988). A c-myb antisense oligodeoxynucleotide inhibits normal human hematopoiesis in vitro. Science 242, 1303. Pauza, C. D. (1987). Regulation of human T-lymphocyte gene expression by interleukin 2: Immediate response genes include the proto-oncogene c-myb. Mol. Cell. Biol. 7, 342. Mountz, J. D., Steinberg, A. D., Klinman, D. M., Smith, H. R., and Mushinski, J. F. (1984). Autoimmunity and increased c-myb transcription. Science 226,1087.
354
AMNON ALI'MAN E T AL.
651. Mountz, J. D., Mushinski, J. F., Mark, G. E., and Steinberg, A. D. (1985). Oncogene expression in autoimmune mice. J. Mol. Cell. Immunol. 2, 121. 652. Kronke, M., Leonard, W. J., Depper, J. M., Arya, S. K., Wong-Staal, F., Gallo, R. C., Waldmann, T. A., and Greene, W. C. (1984). Cyclosporin A inhibits T-cell growth factor gene expression at the level of mRNA transcription. Proc. Natl. Acad. Sci. U.S.A. 81,5214. 653. Elliott, J. F., Lin, Y., Mizel, S. B., Bleackley, R. C., Harnish, D. G., and Paetkau, V. (1984). Induction of interleukin 2 messenger RNA inhibited by cyclosporin A. Science 226, 1439. 654. Smith, K. A., and Cantrell, D. A. (1985). Interleukin 2 regulates its own receptors. Proc. Natl. Acad. Sci. U.S.A. 82,864. 655. Reem, G. H., and Yeh, N.-H. (1984). Interleukin 2 regulates expression of its receptor and synthesis of gamma interferon by T lymphocytes. Science 225,429. 656. Depper, J. M., Leonard, W. J., Drogula, C., Kronke, M., Waldmann, T. A., and Greene, W. C. (1985). Interleukin 2 (IL2) augments transcription of the IL-2 receptor gene. Proc. Natl. Acad. Sci. U.S.A. 82,4230. 657. Fujita, T., Shibuya, H., Ohashi, T., Yamanishi, K., and Taniguchi, T. (1986). Regulation of human interleukin-2 gene: Functional DNA sequences in the 5' flanking region for the gene expression in activated T lymphocytes. Cell (Cambridge, Mass.)46,401. 658. Maruyama, M., Shibuya, H., Harada, H., Hatakeyama, M., Seiki, M., Fujita, T., Inoue, J., Yoshida, M., and Taniguchi, T. (1987). Evidence for aberrant activation of the interleukin-2 autocrine loop by HTLV-I-encoded p4OX and T3/Ti complex triggering. Cell (Cambridge, Mass.)48,343. 659. Siebenlist, U., Durand, D. B., Pressler, P., Holbrook, N. J., Norris, C. A,, Kamoun, M., Kant, J. A., and Crabtree, G. R. (1986). Promoter region of interleukin-2 gene undergoes chromatin structure changes and confers inducibility on chloramphenicol acetyltransferase gene during activation of T cells. Mol. Cell. Biol. 6, 3042. 660. Durand, D. B., Bush, M. R., Morgan, J. G., Weiss, A., and Crabtree, G. R. (1987). A 275 basepair fragment at the 5' end of the interleukin 2 gene enhances expression from a heterologous promoter in response to signals from the T cell antigen receptor.]. Exp. Med. 165, 395. 661. Durand, D. B., Shaw, J.-P., Bush, M. R., Replogle, R. E., Belogaje, R., and Crabtree, G. R. (1988). Characterization of antigen receptor response elements within the interleukin-2 enhancer. Mol. Cell. Biol. 8, 1715. 662. Shaw, J.-P., Utz, P. J., Durand, D. B., Toole, J. J., Emmel, E. A., and Crabtree, G. R. (1988). Identification of a putative regulator of early T cell activation genes. Science 241,202. 663. Williams, T. M., Eisenberg, L., Burlein, J. E., Norris, C. A., Pancer, S., Yao, D., Burger, S., Kamoun, M., and Kant, J. A. (1988). Two regions within the human IL-2 gene promoter are important for inducible IL-2 expression. I. Immunol. 141.662. 664. Yaniguchi, T. (1988). Regulation of cytokine gene expression. Annu. Rev. Immunol. 6,439. 665. Nabel, G. J., Gorka, C., and Baltimore, D. (1988). T Cell-specificexpression of interleukin 2: Evidence for a negative regulatory site. Proc. Natl. Acad. Sci. U.S.A. 85,2934.
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
355
666. Efrat, S., and Kaempfer, R. (1984). Control of biologically active interleukin-2 messenger RNA formation in induced human lymphocytes. Proc. Natl. Acad. Sci. U.S.A. 81, 2601. 667. Serfling, E., Barthelmas, R., Pfeuffer, I., Schenk, B., Zarius, S., Swoboda, R., Mercurio, F., and Karin, M. (1989). Ubiquitous and lymphocyte-specific factors are involved in the induction of the mouse interleukin 2 gene in T lymphocytes. EMBO J. 8,465. 668. Inoue, J., Seiki, M., Taniguchi, T., Tsuru, S., and Yoshida, M. (1986). Induction of interleukin 2 receptor gene expression by p4OXencoded by human T cell leukemia virus type 1. EMBO J. 5,2883. 669. Siekevitz, M., Feinberg, M. B., Holbrook, N., Wong-Staal, F., and Greene, W. C. (1987). Activation of interleukin 2 and interleukin 2 receptor (Tac) promoter expression by the trans-activator (tat) gene produce of human T-call leukemia virus, type I. Proc. Natl. Acad. Sci. U.S.A. 84,5389. 670. Wano, Y., Feinberg, M., Hosking, J. B., Bogerd, H., and Greene, W. C. (1988). Stable expression of the tax gene of type I human T-cell leukemia virus in human T cells activates specific cellular genes involved in growth. Proc. Natl. Acad. Sci. U.S.A. 85,9733. 671. Fujita, T., Takaoka, C., Matsui, H., and Taniguchi, T. (1983). Structure of the human interleukin 2 gene. Proc. Natl. Acad. Sci. U.S.A. 80, 7437. 672. Miyatake, S., Seiki, M., Yoshida, M., and Ari, K. (1988). T-cell activation signals and human T-cell leukiemia virus type I-encoded p4OX protein activate the mouse granulocyte-macrophage colony-stimulating factor gene through a common DNA element. Mol. Cell. Biol. 8, 5581. 673. Miyatake, S., Seiki, M., Malefijt, R. D., Heike, T., Fujisawa, J., Takebe, Y., Nishida, J., Shlomai, J., Yokota, T., Yoshida, M., Arai, K., and Arai, N. (1988). Activation of T cell derived lymphokine genes in T cells and fibroblasts: Effects of human T cell leukemia virus type I p4OX protein and bovine papilloma virus encoded E2 protein. Nucleic Acids Res. 16,6547. 674. Siekevitz, M., Josephs, S. F., Dukovich, M., Peffer, N., Wong-Staal, F., and Greene, W. C. (1987). Activation of the HIV-1 LTR by T cell mitogens and the trans-activatory protein of HTLV-I. Science 238, 1575. 675. Shirakawa, F., Tanaka, Y., Eto, S., Suzuki, H., Yodoi, J., and Yamashita, U. (1986). Effect of interleukin 1 on the expression of interleukin 2 receptor (Tac antigen) on human natural killer cells and natural killer-like cell line Y T cells.J.Immunol. 137,55 1. 676. Lubinski, J., Fong, T. C., Babbitt, J. T., Ransone, L., Yodoi, J., and Bloom, E. T. (1988). Increased binding of IL2 and increased IL2 receptor mRNA synthesis are expressed by an NK-like cell line in response to IL-1. J. Immunol. 140, 1903. 677. Lee, J. C., Truneh, A., Smith, M. F., and Tsang, K. Y. (1987). Induction of interleukin 2 receptor (TAC) by tumor necrosis factor in Y T cells.J. Immunol. 139, 1935. 678. Ruben, S., Poteat, H., Tan, T.-H., Kawakami, K., Roeder, R.,Haseltine, W., and Rosen, C. A. (1988). Cellular transcription factors and regulation of receptor gene expression by HTLV-I tax gene product. Science 241,89. 679. Cross, S. L., Feinberg, M. B., Wolf, J. B., Holbrook, N. J., Wong-Staal, F., and Leonard, W. J. (1987). Regulation of the human interleukin-2 receptor a! chain promoter: Activation of a nonfunctional promoter by the transactivator gene of HTLV-I. Cell (Cambridge, Mass.) 49,47.
356
AMNON ALTMAN E T AL.
680. Leung, K., and Nabel, C . J. (1988). HTLV-I transactivator induces interleukin-2 receptor expression through an NF-kB-like factor. Nature (London) 333,776. 681. Ballard, D. W., Bohnlein, E., Lowenthal, J. W., Wano, Y., Franza, B. R., and Greene, W. C. (1988). HTLV-I Tax induces cellular proteins that activate the kB element in the IL2 receptor a gene. Science 241, 1652. 682. Bohnlein, E., Lowenthal, J. W., Siekevitz, M., Ballard, D. W., Franza, B. R., and Greene, W. C. (1988). The same inducible nuclear proteins regulate mitogen activation of both the interleukin-2 receptor-alpha gene and type 1 HIV. Cell (Cambridge, Mass.) 53,827. 683. Sen, R., and Baltimore, D. (1986). Induction of k immunoglobulin enhancer-binding protein NF-LB by a posttranslational mechanism. Cell (Cambndge, Mass.)47,921. 684. Pierce, J. W., Lenardo, M., and Baltimore, D. (1988). Oligonucleotide that binds nuclear factor NF-kB acts as a lymphoid-specific and inducible enhancer element. Proc. Natl. Acad. Sci. U.S.A. 85, 1482. 685. Nabel, G., and Baltimore, D. (1987). An inducible transcription factor activates expression of human immunodeficiency virus in T cells. Nature (London) 326, 7 11. 685a. Kawakami, K., Scheidereit, C., and Roeder, R. G. (1988). Identification and purification of a human immunoglobulin enhancer-binding protein (NF-LB) that activates transcription from a human immunodeficiency virus type 1 promoter in uitro. Proc. Natl. Acad. Sci.U.S.A. 85,4700. 686. Baeuerle, P. A., and Baltimore, D. (1988). ILB:,A specific inhibitor of the NF-kB transcription factor. Science 242, 540. 687. Zagury, D., Bernard, J., Leonard, R., Cheinyer, R., Feldman, M., Sarin, P. S., and Gallo, R. C. (1986). Long-term cultures of HTLV-111-infected T cells: A model of cytopathology of T-cell depletion of AIDS. Science 231, 850. 688. Folks, T., Powell, D. M., Lightfoote, M. M., Benn, S., Martin, M. A., and Fauci, A. S. (1986). Induction of HTLV-III/LAV from a non virusproducing T-cell line: Implications for latency. Science 231,600. 689. Kaufman, J. D., Valandra, G., Roderiquez, G., Bushar, G., Giri, C., and Norcross, M. A. (1987). Phorbol ester enhances human immunodeficiency virus-promoted gene expression and acts on a repeated 10-base-pair functional enhancer element. Mol. Cell. Biol. 7,3759. 690. Tong-Starksen, S. E., Luciw, P. A., and Peterlin, B. M. (1987). Human immunodeficiency virus long terminal repeat responds to T-cell activation signals. Proc. Natl. Acad. Sci. U.S.A. 84,6849. 691. Patarca, R., Schwartz,J., Singh, R.P., Kong, Q.-T., Murphy, E., Anderson, Y., Sheng, F. W., Singh, P., Johnson, K. A., Guarnagia, S. M., Durfee, T., Blattner, F., and Cantor, H. (1988). qt-1, an intracellular protein from helper/inducer T cells that regulates gene expression of interleukin 2 receptor and human immunodeficiency virus type 1. Proc. Natl. Acad. Sci. U.S.A. 85,2733. 692. Phan-Dinh-Tuy, F., Tuil, D., Schweighoffer, F., Pinset, C., Kahn, A., and Minty, A. (1988). The “CC.Ar.GG” box: A protein-binding site common to transcription-regulatory regions of the cardiac actin, c-fos and interleukin-2 receptor genes. Eur. J. Biochem. 173,507.
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
357
693. Gillis, S., Ferm, M. M., Ou, W., and Smith, K. A. (1978). T cell growth factor: Parameters of production and a quantitative microassay for activity. J. Immunol. 120,2027. 694. Leonard, W. J., Depper, J. M., Uchiyama, T., Smith, K. A., Waldmann, T. A., and Greene, W. C. (1982). A monoclonal antibody that appears to recognize the receptor for human T cell growth factor: Partial characterization of the receptor. Nature (London) 300, 267. 695. Robb, R. J., Munck, A., and Smith, K. A. (1981). T cell growth factor receptors. Quantitation, specificity and biological relevance. J . Exp. Med. 154,1455. 696. Leonard, W. J., Depper, J. M., Crabtree, G. R., Rudikoff, S., Pumphrey, J., Robb, R. J., Kronke, M., Svetlik, P. B., Peffer, N. J., Waldmann, T. A., and Greene, W. C. (1984). Molecular cloning and expression of cDNAs for the human interleukin-2 receptor. Nature (London) 311,626. 697. Nikaido, T., Shimizu, A., Ishida, N., Sabe, H., Teshigawara, K., Maeda, M., Uchiyama, T., Yodoi, J., and Honjo, T. (1984). Molecular cloning of cDNA encoding human interleukin-2 receptor. Nature (London) 311,631. 698. Ralph, P., Jeong, G., Welte, K., Mertelsmann, R., Rabin, H., Henderson, L. E., Souza, L. M., Boone, T. C., and Robb, R. J. (1984). Stimulation of immunoglobulin secretion in human B lymphocytes as a direct effect of high concentrations of IL2. J.Immunol. 133,2442. 699. Greene, W. C., Robb, R. J., Svetlik, P. B., Rusk, C. M., Depper, J. M., and Leonard, W. J. (1985). Stable expression of cDNA encoding the human interleukin 2 receptor in eukaryotic cells. J . Exp. Med. 162,363. 700. Hatakeyama, M., Minamoto, S., Uchiyama, T., Hardy, R. R., Yamada, G., and Taniguchi, T. (1985). Reconstitution of functional receptor for human interleukin-2 in mouse cells. Nature (London) 318,467. 701. Robb, R. J. (1986). Conversion of low-affinity interleukin 2 receptors to a high-affinity state following fusion of cell membranes. Proc. Nutl. Acad. Sci. U.S.A. 83, 3992. 702. Tsudo, M., Kozak, R. W., Goldman, C. K., and Waldmann, T. A. (1987). Contribution of a p75 interleukin 2 binding peptide to a high-affinity interleukin 2 receptor complex. Proc. Natl. Acad. Sci. U.S.A. 84,4215. 703. Robb, R. J., Greene, W. C., and Rusk, C. M.(1984). Low and high affinity cellular receptors for interleukin 2. Implications for the level of Tac antigen. J . Exp. Med. 160, 1126. 704. Sharon, M., Klausner, R. D., Cullen, B. R., Chizzonite, R., and Leonard, W. J. (1986). Novel interleukin-2 receptor subunit detected by cross-linking under high-affinity conditions. Science 234, 859. 705. Tsudo, M., Kozak, R. W., Goldman, C. K., and Waldmann, T. A. (1986). Demonstration of a non-Tac peptide that binds interleukin 2: A potential participant in a multichain interleukin 2 receptor complex. Proc. Nutl. A c d . Sci. U.S.A. 83,9694. 706. Teshigawara, K., Wang, H.-M., Kato, K., and Smith, K. A. (1987). Interleukin 2 high-affinity receptor expression requires two distinct binding proteins. J. Exp. Med. 165,223. 707. Robb, R. J., Rusk, C. M.,Yodoi, J., and Greene, W. C. (1987). Interleukin 2 binding molecule distinct from the Tac protein: Analysis of its role in formation of high-affinity receptor. Proc. Natl. Acad. Sci. U.S.A. 84, 2002.
358
AMNON ALTMAN E T AL.
708. Dukovich, M., Wano, Y., Thuy, L. B., Katz, P., Cullen, B. R., Kehrl, J. H., and Greene, W. C. (1987). A second human interleukin-2 binding protein that may be a component of high-affinity interleukin-2 receptors. Nature (London) 327, 518. 709. Wang, H.-M., and Smith, K. A. (1987). The interleukin 2 receptor. Functional consequences of its bimolecular structure.J. Exp. Med. 166, 1055. 710. Lowenthal, J. W., and Greene, W. C. (1987). Contrasting interleukin 2 binding properties of the a (p55) and P (p70) protein subunits of the human high-affinity interleukin 2 recept0r.J. Exp. Med. 166, 1156. 71 1. Tsudo, M., Kitamura, F., and Miyasaka, M. (1989). Characterization of the interleukin 2 receptor P chain using three distinct monoclonal antibodies. Proc. Natl. Acad. Sci. U.S.A. 86, 1982. 712. Hatakeyama, M., Tsudo, M., Minamoto, S., Kono, T., Doi, T., Miyata, T., Miyasaka, M., and Taniguchi, T. (1989). Interleukin-2 receptor P chain gene: Generation of three receptor forms by cloned human a and /3 chain cDNAs. Science 244, 55 1. 713. Herrmann, T., and Diamantstein, T. (1987). The mouse high-affinity IL-2 receptor complex. I. Evidence for a third molecule, the putative y chain, associated with the a-and/or P-chain of the receptor. Immunobiology 175, 145. 714. Saragovi, H., and Malek, T. R. (1988). Direct identification of the murine IL-2 receptor p55-p75 heterodimer in the absence of IL2.J. Immunol. 141, 476. 715. Rubin, L. A., Hoekzema, G. S., Nelson, D. L., Greene, W. C., and Jay, G. (1987). Reconstitution of a functional interleukin 2 receptor in a nonlymphoid cell.J. Immunol. 139,2355. 716. Weissman, A. M., Harford, J. B., Svetlik, P. B., Leonard, W. J., Depper, J. M., Waldmann, T. A., Greene, W. C., and Klausner, R. D. (1986). Only high-affinity receptors for interleukin 2 mediate internalization of ligand. Proc. Natl. Acad. Sci. U.S.A. 83, 1463. 717. Fujii, M., Sugamura, K., Sano, K., Nakai, M., Sugita, K., and Hinuma, Y. (1986). High-affinity receptor-mediated internalization and degradation of interleukin 2 in human T cel1s.J. Exp. Med. 163, 550. 718. Siegel, J. P., Sharon, M., Smith, P. L., and Leonard, W. J. (1987). The IL-2 receptor P chain (p70): Role in mediating signals for LAK, N K and proliferative activities. Science 238,75. 7 19. Farrar, W. L., and Anderson, W. B. (1985). Interleukin-:! stimulates association of protein kinase C in plasma membrane. Nature (London) 315,233. 720. Rossio, J. L., Ruscetti, F. W., and Farrar, W. L. (1986). Ligand-specific calcium mobilization in IL2 and IL3 dependent cell lines. Lymphokim Res. 5, 163. 721. Bonvini, E., Ruscetti, F. W., Ponzoni, M., Hoffman, T., and Farrar, W. L. (1987). Interleukin 2 rapidly stimulates synthesis and breakdown of polyphosphoinositides in interleukin 2-dependent, murine T cell 1ines.J. Eiol. Chem. 262,4 160. 722. Mills, G. B., Cheung, R. K., Grinstein, S., and Gelfand, E. W. (1985). Interleukin 2-induced lymphocyte proliferation is independent of increases in cytosolic-free calcium concentrations. J . Immunol. 134, 2431.
MOLECULAR EVENTS MEDIATING
T CELL ACTIVATION
359
723. Mills, G. B., Stewart, D. J., Mellors, A., and Gelfand, E. W. (1986). Interleukin 2 does not induce phosphatidylinositol hydrolysis in activated T cells. J. Immunol. 136, 3019. 724. LeGrue, S. J. (1987). Interleukin-2 stimulus-response coupling is calcium independent. Lymphokine Res. 6, 1. 725. LeGrue, S. J. (1988). Does interleukin 2 stimulus-response coupling result in generation of intracellular second messengers? Lymphokine Res. 7, 187. 726. Tigges, M. A., Casey, L. S., and Koshland, M. E. (1989). Mechanism of interleukin-2 signaling: Mediation of different outcomes by a single receptor and transduction pathway. Science 243, 78 1. 727. Valge, V. E., Wong, J. G. P., Datlof, B. M., Sinskey, A. J., and Rao, A. (1988). Protein kinase C is required for responses to T cell receptor ligand, but not to interleukin-2 in T cells. Cell (Cambridge, Mass.) 55, 101. 728. Redondo, J. M., Rivas-Lopez, A., Vila, V., Cragoe, E. J., and Fresno, M. (1988). T h e role of protein kinase C in T lymphocyte proliferation. Existence of protein kinase C-dependent and -independent pathways. J. Biol. Chem. 263, 17467. 729. Mizuguchi, J., Beaven, M. A., Ohara, J., and Paul, W. E. (1986). BSF-1 action on resting B cells does not require elevation of inositol phospholipid metabolism or increased [Ca2'],.j. Immunol. 137,2215. 730. Wickremasinghe, R. G., Mire-Sluis, A. R., and Hoffirand, A. V. (1987). Interleukin-2 binding to activated human T lymphocytes triggers generation of CAMPbut not of inositol phosphates. FEBS Lett. 220,52. 731. Hadden, J. W., Hadden, E. M., and Coffey, R. G. (1987). Interleukin 2 increases cyclic GMP levels in immature thymocytes and mitogen-primed T lymphocytes. Int. J . Immunophamacol. 9, 85 1. 732. Benedict, S. H., Mills, G. B., and Gelfand, E. W. (1987). Interleukin 2 activates a receptor-associated protein kinase. J. Immunol. 139, 1694. 733. Frasier-Scott, K., Hatzakis, H., Seong, D., Jones, C. M., and Wu, K. K. (1988). Influence of natural and recombinant IL2 on endothelial cell arachidonate metabolism. Induction of de novo synthesis of prostaglandin H synthase. J . Clin. Invest. 82, 1877. 734. Evans, S. W., Beckner, S. K., and Farrar, W. L. (1987). Stimulation of specific GTP binding and hydrolysis activities in lymphocyte membrane by interleukin 2. Nature (London) 325, 166. 735. Isakov, N., Bleackley, R. C., Shaw, J., and Altman, A. (1985). The tumor promoter teleocidin induces IL2 receptor expression and IL2-independent proliferation of human peripheral blood cells. J. Immunol. 135,2343. 736. Isakov, N., Bleackley, R. C., Shaw, J., and Altman, A. (1987). Teleocidin and phorbol ester tumor promoters exert similar mitogenic effects on human lymphocytes. Biochem. Biophys. Res. Commun. 130, 724. 737. Isakov, N., Scholz, W., and Altman, A. (1987). Effect of cyclosporin A on early stages of T cell activation. Transplant. Proc. 19, 1186. 738. Kim, D.-K., Otten, G., Moldwin, R. L., Dunn, D. E., Nau, G. J., and Fitch, F. W. (1986). PMA alone induces proliferation of some murine T cell clones but not others. J. Immunol. 137,2755. 739. Moldwin, R. L., Lancki, D. W., Herold, K. C., and Fitch, F. W. (1986). An antigen-receptor-driven, interleukin 2-independent pathway for proliferation of murine cytolytic T lymphocyte clones. J. Exp. Med. 163, 1566.
360
AMNON ALTMAN E T AL.
740. Laing, T.J., and Weiss, A. (1988). Evidence for IL2-independent proliferation in human T ce1ls.J. Immunol. 140, 1056. 741. Manger, B., Hardy, K. J., Weiss, A., and Stobo, J. D. (1986). Differential effect of cyclosporin A on activation signaling in human T cell 1ines.J. Clin. Invest. 77, 1501. 742. Mosmann, T. R., Bond, M. W., Coffman, R. L., Ohara, J., and Paul, W. E. (1986). T-cell and mast cell lines respond to B-cell stimulatory factor 1.Proc. Natl. Acad. Sci. U.S.A. 83,5654. 743. Ceuppens, J. L., Baroja, M. L., Lorre, K., Van Damme, J., and Biliau, A. (1988). Human T cell activation with phytohemagglutinin. The function of IL6 as an accessory signal.J. Immunol. 141,3868. 744. Le, J., Fredrickson, G., Reis, L. F. L., Diamantstein, T., Hirano, T., Kishimoto, T., and Vilcek, J. (1988). Interleukin 2-dependent and interleukin 2-independent pathways of regulation of thymocyte function by interleukin 6. Proc. Natl. Acad. Sci. U.S.A. 85,8643. 745. Sanders, M. E., Makgoba, M. W., and Shaw, S. (1988). Human naive and memory T cells: Reinterpretation of helper-inducer and suppressorinducer subsets. Immunol. Today 9, 195. 746. Letarte, M., Vera, S., Tran, R., Addis, J. B. L., Onizuka, R. J., Quackenbush, E. J., Jongeneel, C. V., and McInnes, R. R. (1988). Common acute lymphocytic leukemia antigen is identical to neutral endopeptidase. J. Ex$. Med. 168, 1247. 747. Shaw, S. (1989). Lymphocyte differentiation. Not all in a name. Nature ( ~ o n d m338,539. ) 748. OShea, J. J., Cross, S. L., Samelson, L. E., and Klausner, R. D. (1989). Expression of pp60"'"" in T cells results in constitutive tyrosine phosphorylation of the T cell receptor (TCR). FASEBJ. 3, A514. 749. Williams, L. T. (1989). Signal transduction by the platelet-derived growth factor receptor. Science 243, 1564. 750. Shearman, M. S., Naor, Z., Sekiguchi, K., Kishimoto, K., and Nishizuka, Y. (1989). Selective activation of the y-subspecies of protein kinase C from bovine cerebellum by arachidonic acid and its lipoxygenase metabolites. FEBS Lett. 243, 177. 75 1. Mulcahy, L. S., Smith, M.R., and Stacey, D. W. (1985). Requirement for ras proto-oncogene function during serum-stimulated growth of NIH 3T3 cells. Nature (London) 313,241. 752. Angel, P., Hattori, K., Smeal, T., and Karin, M. (1988). The jun protooncogene is positively autoregulated by its product, JunIAP-1. Cell (Cambridge, Mass.)55,875. This article was accepted for publication on 4 October 1989.
Index
A
Activation antigen, 248-250 accessory, 24 1-242 glycosyl-phosphatidylinositol,237-24 1 Adrenocorticotropic hormone, 170- 172 Afferent blockade anterior chamber, 203-204 corneal graft survival, 195- 196 Agglutination epitope, 23 segmental flexibility, 23-24 Allele-specific oligonucleotide, 118- 1 19 hybridization, 118 polymerase chain reaction, 118 Allograft, corneal. See Corneal allograft Alloimmune effector, anterior chamber, 204 Amino acid sequence, 108 Anatomical sequestration, lens, 220 Anterior chamber, 191, 192 afferent blockade, 203-204 alloimmune effector, 204 alloimmunity down-regulation, 204 antigen-processing cells, 206, 207 DBA/2 mastocytoma cell, 204 immunological privilege, 203-2 12 intraocular HSV-1 infection, 210 contralateral eye, 210 delayed-type hypersensitivity suppression, 210 mastocytoma, 2 13-2 15 physiochemical environment, 206, 207 systemic cell-mediated immunity, 204 tumor allograft, 204 tumorigenesis, 213-215 UV5C25.2 13-2 15 Anterior chamber-associated immune deviation, 205-21 1 cytotoxic T lymphocyte, 208,209 DBA/2 mastocytoma model, 207 delayed-type hypersensitivity, 2 19 antigen-specific down-regulation, 208, 209
intraocular tumor allograft, 208, 209 delayed-type hypersensitivity antigen-specific suppression, 21 1 anterior chamber, 21 1 efferent mode, 2 11 hapten model, 21 1 suppressor T cell, 2 11 effector mechanism, 2 11-2 12 expression, 207-210 eye, 206 induction, 205-207 ocular HSV infection, 208-210 S antigen, 219 spleen, 206, 2 11 tumor allograft, intraocular, 207-208 tumorigenesis, 2 13 vascular route, 205 AntLCD2 mAb, 235-236 CD3- Jurkat mutant, 236 CD3- thymocyte, 236 natural killer cell, 236 AntLCD28 mAbs, 245 Anti-CD28-mediated T cell activation, mechanism, 246 Anti-la antibody, experimental autoimmune encephalomyelitis, 164, 166 Anti-TCR antibody, 90 Antibody, 16-18 neutralizing, 95-96 Antibody isotype switching, 83-84 Antidansyl antibody, segmental flexibility, 13, 14 Antiergotypic T cell, 166 Antigen presentation, islet beta cell, 142 Antigen-presenting cell, 232 ASLC, 132 Asp5? insulin-dependent diabetes mellitus, 135-137 location, 139 Astrocyte, immunological mediator, 168-169 36 1
362
INDEX
Autoimmune response, retinal antigen, 217 Autoimmune uveitis, immune regulation, 215-216 Avian lg gene, gene conversion, 60-62 AZH. 136 B
B8 DQw2, 134 DQw3, 134 DQw3.1, 134 DQw3.2, 134 DQw3.3, 134 DQw7,134 DQw8, 134 DQw9, 134 HLA-DQ, 134 insulin-dependent diabetes mellitus, I37 B15 DQw2, 134 DQw3,134 DQw3.1, 134 DQw3.2, 134 DQw3.3, 134 DQw7, 134 DQw8, 134 DQw9, 134 HLA-DQ, 134 insulin-dependent diabetes mellitus, 137 1 B 236/myelin-associated glycoprotein, 161- 162 B cell, Ig diversity, 52 B cell development, bursa of Fabricius, 42-44 B cell growth factor type 11, 78 B lymphocyte, antibody isotype switching, 83-84 BALB/c mouse, splenectomized, 208 Bare lymphocyte syndrome, cell typing, 117 Basophil, 176, 177 Beta-adrenergic mediator, 162 Beta-endorphin, 170- 172 4-kDa BglII RFLP fragment, celiac disease, 145 Biochemical typing, HLA class 11, 114-1 16 interisotypic association, 116
intraisotypic hybrid molecule, 116 trans-association, 116 Black patient DQw9 Asp5’ positive, 140 DR9-DQ beta, 140 DR7 haplotype, 140 DR9 haplotype, 140 Bursa of Fabricius B cell clonal expansion, 54 B cell development, 42-44 development, 42-44 early role, 49-5 1 embryogenesis, 47 Ig gene rearrangement, 49 Ig+ precursor selection, 49-51 somatic diversification, 54 Bursal anlage, 49 Bursal follicle epithelial components, 42-43 lymphoid components, 42-43 Bursal lymphocyte diversity, 52-54 V gene segment, 52-54 Bursal stem cell, 5 1-52 decline, 52 embryogenesis, 52 frequency, 5 1 germ-line gene segment, 51-52 C
c-fos, T cell activation, 291-202 c-myb, T cell activation, 293-294 c-myc, T cell activation, 292-293 Ca2+,228 ion channel, 285-287 mobilization, 260-262 Calcitonin gene-related peptide, 169 Calcium ionophore, T cell activation, 258-260 Carbohydrate component, mobility, 26-32 Carp IgM antibody, 23-24 Cataract, immune-mediated, 2 19-22 1 Caucasian patient DQw9 Asp5’ positive, 140 DR9-DQ beta, 140 DR7 haplotype, 140 DR9 haplotype, 140 CD2.235-237 Abs, 235
INDEX
expression, 235 structure, 235 CD3, 162 CD4,242-245,304-306 independent signal transduction, 243 regulatory functions, 242-243 T cell-specific tyrosine protein kinase, 243-244 CD5,246 CD6,247 CD7,247 CD8,242-245,304-306 independent signal transduction, 243 regulatory functions, 242-243 CD28,244,245-246,304 CD43,248 CD55.241 CD59,240-241 CD69,248-249 CD73,241 CD4+ cell, 84-86, 198 clone, 197 lymphokine production patterns, 85-86 T helper, 163 CD8+ cell, 84-86, 197, 198 lyrnphokine production patterns, 8586 CD3 complex, components, 229,230 CD3- Jurkat mutant, anti-CD2 mAb, 236 CD45 leukocyte common antigen, 283 CD3- thymocyte, antLCD2 mAb, 236 Celiac disease 4-kDa BgIII RFLP fragment, 145 DP allele, 145 DP alpha allele, 145 DP alpha-DQw1 beta, 146-147 DP alpha polymorphism, 145 DP beta allele, 145 DP beta polymorphism, 145 DQw2 alpha chain, 146 DR3, 144, 147 DR5, 144, 147 DR7, 144, 147 DR alpha-DQw1 beta, 146-147 DR3-DQw2 haplotype, 144 DRw8, 147 gluten-derived peptide, 147- 148 HLA class 11, 144 HLA class I1 antigen, 147 hybrid HLA-DQ molecule, 147
363
4-kDa RSaI class I1 fragment, 145 triggering factor, 144 Cell receptor binding, segmental flexibility, 24-26 CH3 domain Fc, 20-21 IgG, 20-2 1 Chicken IgL gene, 41-64 genornic organization, 44-46 germ-line allele, 49-5 1 I g L allele, 49-51 J L gene segment, 44-46 rearrangement, 46-49 somatic diversification, 41-64 structure, 44-46 V segment, 52-54 V gene segment, 44-46 Chimeric antibody, 2 Cholecystokinin receptor, 181 CK- 1 sequence, 8 1-82 CK-2 sequence, 81-82 Class I1 alpha chain, 141-142 Class I1 beta chain, 141-142 Class I 1 molecule, transducing signals, 142 Clonal anergy, 21 7 , 2 18 Clonal suppression, 2 18 Colony-stimulating factor, 69-77 biochemical nature, 72-74 biological actions, 74-76 cell survival, 75 cellular sources, 76-77 characterization, 70-72 chronic deregulated production, 93-94 colony-stimulating activity, 94 detection in vivo, 94-96 differentiation commitment, 75 discovery, 70-72 eosinophil, 92 functional activation, 76 granulocyte, 75 hemopoietic recovery, 93 high-affinity membrane receptors, 73-74 macrophage, 75,92 maturation induction, 75 monocyte, 92 myelosuppression, 93 neutralizing antibody, 95-96 neutrophil, 92 proliferative stimulation, 74-75
364
INDEX
target cell selectivity, 74-75 secondary structure, 73 sequence homology, 73 serum half-lives, 93 single unique gene encoding, 73 T cell synthesis, 96-98 antigen-presenting cell, 97 local polarization secretion, 97 local response, 96 target cells, 97 T lymphocyte studies in vitro, 77-78 studies in vivo, 92-98 therapeutic effects, 93 Colony-stimulating factor gene expression posttranscriptional regulation, 83 transcriptional regulation, 8 1-83 Colony-stimulating factor receptor biochemical nature, 72-74 distribution, 74 Colony-stimulating factor synthesis induction, 78-81 pathways, 78-81 Complement activation, segmental flexibility, 24-26 Complement binding activity hinge region structure, 13, 15,22 segmental flexibility, 13, 15, 22 Concanavalin A, 90,233 Contactin, 161- 162 Cornea, MHC antigen, 193-194 Corneal allograft graft bed afferent blockade theory, 195- 196 avascular, 195 graft location, 195 host cellular component, 195 orthotopically placed, 195 immunological privilege, 192-203 Corneal endothelial graft, cytotoxic T lymphocyte, 194 Corneal epithelial cell, accessory antigen-processing cell, 202 Corneal epithelial cell-derived T cell-activating factor, 20 1-202 Corneal epithelium, Langerhans cell, 197 Corneal graft Ia antigen, 199 MHC class I1 antigen cell-surface expression, 199
expression, 199 expression timing, 199-200 MHC class I1 loci, 199-200 latex bead treatment, 200 rejection, donor Langerhans cells, 196-200 Corneal transplantation, 191 human leukocyte antigen matching, 193 immunological recognition, 193-196 immunosuppressive drug, 193 success, 193 Corticotropin-releasing hormone, 170- 172 CSF-1, 71 Cytokine, 70, 71 acute effects, 174-175 functional effects, 167-179 neural generation, 168-169 persistent effects, 173-174 production, 167-179 Cytotoxic T lymphocyte anterior chamber-associated immune deviation, 208,209 effector T cell, 197 Langerhans cell, 197- 198 ocular HSV infection, 210 transplant, 197-198 D
109 D6 mAb, rheumatoid arthritis, 128 DBA/2 mastocytoma cell, anterior chamber, 204 Decay-accelerating factor, 24 1 Delayed-type hypersensitivity anterior chamber-associated immune deviation, 2 19 antigen-specific down-regulation, 208, 209 intraocular tumor allograft, 208,209 effector T cell, 197 Langerhans cell, 197- 198 ocular HSV infection, 2 10 spleen, 2 11-2 12 transplant, 197 tumor allograft, 208, 209 UV5C25,214 Diabetes mellitus, insulin-dependent. See Insulin-dependent diabetes mellitus Diacylglycerol, 228, 309
INDEX
Disulfide bridge, 13 DNA RFLP typing strategy, 117-1 18 Dob, Fc, 21 Domain structure genetically engineered antibody, 14, 16,22 segmental flexibility, 14, 16, 22 Dominant resistance, insulin-dependent diabetes mellitus, 143-144 Double-homologous recombination, 56 DP, 109, 110-1 11 DP allele, celiac disease, 145 DP alpha allele, celiac disease, 145 DP alpha-DQ beta, multiple sclerosis, 149 DP alpha-DQw 1 beta, celiac disease, 146-147 DP alpha polymorphism, celiac disease, 145 DP beta allele, celiac disease, 145 DP beta polymorphism, celiac disease, 145 DPw4, multiple sclerosis, 149-150 DQ, 109,110-111 linkage disequilibrium phenomenon, 136-137 DQ allele, linkage disequilibrium, 12 1- 122 DQ alpha cis-complementation, 139-140 structural requirement, 138-139 tram-complementation, 139- 140 DQ alpha chain, functional dimer, 123 DQ alpha-DQ beta, multiple sclerosis, 149 DQ alpha gene polymorphism, multiple sclerosis, 149 DQ alpha polymorphism, 121 DQ beta cis-complementation, 139- 140 structural requirement, 138-139 tram-complementation, 139- 140 DQ beta chain amino acid 57, location, 139 functional dimer, 123 insulin-dependent diabetes mellitus, 134- 136 position 57 aspartic acid, 135-136, 137 DQ beta polymorphism, 121 DQ heterodimer, DQ2-DQ3 cell, 120-121 DQR2-6, multiple sclerosis, 148- 149 DQw 1, multiple sclerosis, 148
365
DQw2 B8, 134 B15, 134 DR3, 134 DR4, 134 DQw3 B8, 134 B15, 134 DR3, 134 DR4, 134 DQw3.I B8, 134 B15, 134 DR3, 134 DR4, 134 DQw3.2 B8, 134 B15, 134 beta chain, 136 DR3, 134 DR4, 134 DQw3.3 B8, 134 B15, 134 DR3, 134 DR4, 134 DQw7 B8, 134 B15, 134 DR3, 134 DR4, 134 DQw8 B8, 134 B15, 134 DR3, 134 DR4, 134 DQw9 B8, 134 B15, 134 DR3, 134 DR4, 134 DQw2 alpha chain, celiac disease, 146 DQw9 Asp5’ positive black patient, 140 Caucasian patient, 140 DQwla, multiple sclerosis, 148 DQw 1b, multiple sclerosis, 148 DR, 109, 110-111 linkage disequilibrium phenomenon, 136- 137
366
INDEX
DRl, 127 DR2, 112 insulin-dependent diabetes mellitus, 136, 143 AZH, 136 multiple sclerosis, 148- 150 DR3 celiac disease, 144, 147 DQw2, 134 DQw3, 134 DQw3.1, 134 DQw3.2, 134 DQw3.3, 134 DQw7,134 DQw8, 134 DQw9, 134 DR4, 112 DQw2, 134 DQw3, 134 DQw3.1, 134 DQw3.2, 134 DQw3.3, 134 DQw7, 134 DQw8, 134 DQw9, 134 HLA-DQ, 134 juvenile rheumatoid arthritis, 131 DR5 celiac disease, 144, 147 juvenile rheumatoid arthritis, 131 DR7, celiac disease, 144, 147 DR allele, linkage disequilibrium, 121122 DR alpha-DQ beta, multiple sclerosis, 149 DR alpha-DQw 1 beta, 125 celiac disease, 146- 147 DR alpha-DQw2 beta, 125 DR alpha-DQw3 beta, 125 DR alpha gene, 124 expression modulation, 125 DR alpha protein, 124 DR beta gene, rheumatoid arthritis, 131 DR9-DQ beta black patient, 140 Caucasian patient, 140 DR4 Dw haplotype, 118 DR4 haplotype, 127 DR7 haplotype black patient, 140 Caucasian patient, 140
DR9 haplotype black patient, 140 Caucasian patient, 140 DRw8 celiac disease, 147 juvenile rheumatoid arthritis, 13 1 DRw9, insulin-dependent diabetes mellitus, 137 DRwl5, insulin-dependent diabetes mellitus, 143 DRw 16, insulin-dependent diabetes mellitus, 143 DRw52, 112 DRw53, 112 DRw53 molecule, rheumatoid arthritis, 128 Dw4 position 67, 128-129 position 70, 128-129 position 71, 128-129 rheumatoid arthritis, 126- 127 DwlO, 127 rheumatoid arthritis, 126- 127 DW14, rheumatoid arthritis, 126-127 DX alpha polymorphism, 139
E EA-1,248-249 Early activation antigen-1,248-249 Ecto-5’-nucleotidase, 24 1 Effector T cell cytotoxic T lymphocyte, 197 delayed-type hypersensitivity, 197 Em bryogenesis bursa1 stem cell, 52 V-Jjoint, 49,50 Eosinophil differentiation factor, 78 Eosinophil proliferation, GM-CSF, 75 Epitope agglutination, 23 cell surface-expressed, 1 14 conformational, 114 HLA class 11, 112 precipitation, 23 Epstein-Barr virus, QKRAA Dw4 peptide, 130 gp 110 peptide, 130 Experimental autoimmune encephalomyelitis, 163-167
367
INDEX
anti-CD4 antibody, 164, 166 anti-la antibody, 164, 166 immunomodulation, 166- 167 inflammatory demyelination, 166- 167 T cell receptor V gene, 166 THcell receptor, 166 Experimental autoimmune uveitis immune privilege, 216-219 induction, 215-21 6 pathogenesis, 2 15-216 S antigen, 219 self-tolerance, 2 16-219 T cell, 216 Eye anterior chamber-associated immune deviation, 206 autoimmune diseases, 216-219 immune privilege, 191-221 immune regulation, 191-221 self-tolerance, strategies, 2 17 trabecular meshwork, 206-207 transplantation, 191 F
Fab
C portion, 20
constant domain, 19 elbow bend, 19 heavy chain, 19 hinge region structure, 21-23 internal motions, 19-21 light chain, 19 protein antigen antibodies, 20 switch region flexibility, 19 V portion, 20 Fab subunit, rotation modes, 33 Fc A d g 7 residue, carbohydrate unit, 27 C H domain, ~ 20-2 1 C H domain ~ lateral contact, 20 Dob, 21 hinge region structure, 21-23 internal motions, 19-21 Mcg, 21 rotational properties, 20 three-dimensional structure, 20 Fc carbohydrate, 27 Fc oligosaccharide, protein surface, 28-29
Fluorescence polarization IgA, rotational correlation times, 16- 17 IgE, rotational correlation times, 16- 17 ISM, rotational correlation times, 16- 17
G G-CSF, 72 G protein phosphoinositide-specific phospholipase, 266 T cell, 267-269 T cell antigen receptor-CD3 complex, 306-307 Gamma-interferon, 132, 162 Gastrin receptor, 181 Gene conversion avian Ig gene, 60-62 Ig VL gene, 60-62 psi-V gene segment, 58 VL gene segment, 56-60 Genetically engineered antibody domain structure, 14, 16,22 segmental flexibility, 14, 16, 22 Germ-line VL, sequence comparison, 55 Glial growth-promoting factor, 173 Glioblastoma cell, immunological mediator, 168-169 Gluten-derived peptide, celiac disease, 147-148 Gluten enteropathy. See Celiac disease Glycosyl-phosphatidylinositol,activation antigen, 237-241 Graft bed, corneal allograft afferent blockade theory, 195- 196 avascular, 195 graft location, 195 host cellular component, 195 orthotopically placed, 195 Granulocyte, 69 colony-stimulating factor, 75 IL-4, 75 IL-6, 75 Granulocyte-macrophage colony-stimulating factor, 71, 169 eosinophil proliferation, 75 IL-5, 75 Granulocyte-macrophage response, sequential nature, 70 Growth factor receptor, 253
368
INDEX
Growth hormone, 172 macrophage, 178 GTP-binding protein, T cell signal transduction, 265-269 H
2H 1,249-250 Haplotype 6, 123 Haplotype d, 123 Haplotype k, 123 Hapten-derived T suppressor-inducer cell,
212 Heavy-chain gene, 41 Hemopoiesis, 83-84 Herpes simplex virus, ocular infection, 208-210 Herpes simplex virus- 1, 2 10 Hinge region, upper part complement activation, 25 Fab motion, 25 length, 25 segmental flexibility, 25 Hinge region structure, 11 complement binding activity, 13, 15, 22 Fab, 21-23 Fc, 21-23 segmental flexibility, 13, 15.22 HIVENSBA, 298 HLA human leukocyte antigen, oligosaccharide mobility, 3 1 matching, corneal transplantation, 193 population distribution, 126 rheumatoid arthritis, 126 HLA-AS, multiple sclerosis, 148-150 HLA antigen, rotational correlation times, 28-29 HLA-B7, multiple sclerosis, 148-150 HLA class I, crystal structure, 1 11 HLA class I1 allelic variability amino acid stretches, 111-1 12 amino acid sequence, 108 antigen recognition site, 111 biochemical typing, 114-1 16 intensotypic association, 1 16 intraisotypic hybrid molecule, 116 trans-association, 1 16
celiac disease, 144, 147 disease associations, 107- 153 disease susceptibility, 108 DP, 109, 110-1 11 DQ, 109,110-1 11 DQ gene, 120- 121 DQ haplotype, 120-121 DQ molecule, 120-121 DR, 109,110-111 epitope, 112 molecular localization, 152- 153 gene location, 109 gene structure, 109- 1 12 homologous, 109 hybrid molecule, 119-126 cell surface expression, 119 stable heterodimer, 119 molecule structure, 109-1 12 one-dimensional isoelectrofocusing, 114-1 16 polymorphism, 109, 112-119 serology, 112-1 14 two-dimensional polyacrylamide gel electrophoresis, 114-1 16 silver staining, 116 theoretical limitations, 116 typing, 112-1 19 Western blotting, 116 HLA class I1 molecule interisotypic hybrid, 124- 126 transfected genes, 124 synovial cell, 132 HLA-D region, 108 HLA-D region specificity, 113-1 14 HLA D typing, HLA DR specificity, 113-1 14 HLA-DQ B8, 134 B15, 134 DR4,134 homologous hybrid, 137 HLA-DQ hybrid molecule, 119-124 celiac disease, 147 cell surface expression, 122 disease susceptibility, 122 trans-complementation, 119- 120 transectants. 122 HLA DR specificity, HLA D typing, 113-1 14 HLA-Dwl4, T cell clone reactivity, 129
INDEX
HLA factor, insulin-dependent diabetes mellitus, 133- 144 Host response, sequential nature, 70 Human B cell line, Epstein-Barr virus-transformed, 125 Human granulocyte, opioid receptor, 180 Human Ig, rotational correlation times, 28-29 Human IgG2, 13 Human IM-9 B lymphoblast, 180 Human leukocyte antigen. See HLA Human lymphocyte, opioid receptor, 180 Human lymphocyte receptor, neuropeptide, 180 Human monocyte, opioid receptor, 180 Human myeloma, 16-18 Human myeloma IgM, 16-18 Hypersensitivity, delayed-type. See Delayed-type hypersensitivity
I Ia antigen, corneal graft, 199 Ia+ cell, 197 Ig diversification evolutionary mechanism, 62 mechanisms, 62 Ig molecule carbohydrate component, mobility, 26-32 effector functions, I flexible, 1 internal movement, 1-34 ligand, 1 recognition function, 1 rotational correlation times, 28-29 segmental flexibility, 5-19 electron microscopy, 2-3 fluorescence polarization, 3-5 hydrodynamic study methods, 2 nanosecond, 4-5 spin-label approach, 5 steady state, 4 study methods, 1-34 X-ray crystallography, 3 Ig oligosaccharide, spin-labeling, 27-28 Ig+ precursor selection, bursa of Fabricius, 49-5 1 Ig VL gene gene conversion, 60-62
369
somatic diversification, 60-62 IgA fluorescence polarization, rotational correlation times, 16- 17 segmental flexibility, 14-18 IgAI, wide extrema, 30 IgAP, rotation freedom, 31 IgE fluorescence polarization, rotational correlation times, 16-1 7 functions, 18 segmental flexibility, 18- 19 wide extrema, 30 IgG c H 3 domain, 20-2 1 segmental flexibility, 5-19 early studies, 5-6 fluorescence polarization, 6- 10 hinge region structure, 21-23 IgG subclasses, 11-14 spin-label method, 10-11, 12 X-ray crystallography, 11 IgGi, 13 IgGl heavy chain, hinge region structure, 21-22 IgGl oligosacharide, nuclear magnetic resonance, 29 IgL gene, diversification, potential mechanisms, 58 IgL gene locus, somatic diversification, 52,53 IgM fluorescence polarization, rotational correlation times, 16- 17 segmental flexibility, 14-18 IgM antibody, 16-18 li chain, 132 IL. See also Interleukin IL-1, 168, 169 neuropeptide, 178 IL-2, 77 IL-3, 71, 72, 168, 169 IL-4, 71, 72 granulocyte, 75 macrophage, 75 IL-5, 71, 72, 78 GM-CSF, 75 IL-6, 71, 72, 168, 169 granulocyte, 75 Immortalized producer cell line, 69
370 Immune privilege experimental autoimmune uveitis, 2 16-2 19 eye, 191-221 Immune regulation autoimmune uveitis, 215-216 eye, 191-221 intraocular tumor rejection, 212-215 neoplasm, 2 12 spontaneous neoplasm, 212 Immune surveillance theory, 212 Immunoglobin. See Ig Immunological cell, neuropeptide mediation, 176, 177 Immunological factor low-molecular-weight mediators, 175 neural effects, 173-174 Immunological mediator astrocyte, 168-169 glioblastoma cell, 168- 169 microglial cell, immune synthetic capabilities, 168 neural sources, 168-169 Immunological privilege anterior chamber, 203-212 corneal allograft, 192-203 hapten model, 211-212 Imrnunomodulation, experimental autoimmune encephalomyelitis, 166-167 Immunosuppressive drug, corneal transplantation, 193 Inflammatory demyelination, experimental autoimmune encephalomyelitis, 166- 167 Inositol, inositol phosphate, 255 Inositol 1-monophosphate, phosphoinositidespecific phospholipase, 253 Inositol 1,4,5-trisphosphate, 228 Inositol phosphate, inositol, 255 Inositol phosphate isomer, dephosphorylation pathways, 255 Inositol stereoisomer, 255 Insulin-dependent diabetes mellitus aberrant HLA class I1 expression, 141- 142 Asp5’, 135-137 B8, 137 B15, 137
INDEX
dominant resistance, 143- 144 DQ beta chain, 134-136 position 57 aspartic acid, 135-136, 137 DR2, 136, 143 AZH, 136 DRw9,137 DRwl5, 143 DRwl6, 143 HLA factor, 133-144 suppression phenomenon, 143- 144 Insulin gene, transcriptional factor, 142 Insulin promoter gene, 141-142 Interferon-alpha, 168, 169 Interferon-beta, 168, 169 Interferon-gamma, 70, 7 1, 78 Interleukin, 70,71. See also IL Interleukin-2 gene, transcriptional regulation, 294-297 Interleukin-2 receptor function, 300-302 signaling, 299-304 structure, 299-300 Interleukin-7 receptor, T cell activation, 250-253 signal transduction, 251-252 Interleukin-2R alpha gene, transcriptional regulation, 294-295,297-299 Interphotoreceptor retinoid-binding protein, 215-216 Intraocular HSV- 1 infection, anterior chamber, 210 contralateral eye, 210 delayed-type hypersensitivity suppression, 210 Intraocular tumor rejection basic patterns, 2 15 cytolytic T lymphocyte, 215 hemorrhagic necrosis, 2 15 immune regulation, 2 12-2 15 immune rejection, 215 Ion channel Ca2+,285-287 potassium, 287-288 types, 287 T cell, 285-288 regulation, 285,286 Ionomycin, 80,90 Islet beta cell, 140-141 antigen presentation, 142
371
INDEX
lymphokine, 14 1 T cell cytotoxicity, 142 Isotype-mismatched molecule, 125 J
Jurkat leukemic T cell, 182 Juvenile rheumatoid arthritis DR4, 131 DR5, 131 DRw8, 131 1
Langerhans cell corneal epithelium, 197 corneal graft rejection, 196-200 cytotoxic T lymphocyte, 197-198 delayed-type hypersensitivity, 197-198 dynamic distribution, 200-202 epithelial distribution, 201 Ia+, 197 immunoregulatory effects, 203 male vs. female tissue, 198-199 migration, 20 1 orthotopic graft, 198 regulatory process, 202 Lens anatomical sequestration, 220 autoantigens, 220 crystallin, 219 alpha, 220 antigen, 191 gamma, 220 self-tolerance, 220 Leu- I , 246 Leukemic cell, cell typing, 117 LFA- 1,247 LFA-3,236 Light-chain gene, 41-64 diversity, 41 gene conversion, 41-42 independent genes, 41 joining events, 4 1 Linkage disequilibrium phenomenon DQ, 136-137 DR, 136-137 Ly-6, 240-24 1 Lymphocyte cytokines produced, 71
function, neuroendocrine mediator, 177-179 hemopoietic regulators produced, 7 1 neuropeptide, 177 differentiation, 178 immunological functions, 178 tissue cycles, 178 tissue homing, 178 retinal glial cell, 218 Lymphocyte receptor, neuromediator, 179- 184 Lymphoid cell, 69 Lymphoid tissue, tyrosine protein kinase, 277-280 Lymphokine fresh tissues, 95 islet beta cell, 141 Lymphokine synthesis, 78-8 1 IL-2-induced, 81 transcription, 8 1 Lyt- 1, 246 M
M-CSF, 71 Macrophage, 69 colony-stimulating factor, 75 growth hormone, 178 IL-4, 75 neuropeptide, 177 Macrophage-activating factor, 78 Major histocompatibility complex. See MHC Mast cell, 176, 177 Mastocytoma, anterior chamber, 213-215 MC 1, rheumatoid arthritis, 128 Mcg, Fc, 21 Met-enkephalin, 172 MHC, 78, 107 MHC antigen, cornea, 193- 194 MHC class 11 antigen, corneal graft cell-surface expression, 199 expression, 199 expression timing, 199-200 MHC class I1 loci, corneal graft, 199-200 latex bead treatment, 200 MHC control, anticollagen type I1 response, 130- 131 Mixed-isotype dimer, formation, 125
372
INDEX
Mouse monoclonal antibody, segmental flexibility, 13, 14 Mouse strain BIO.PL, 164-165 Mouse strain PL/J(H-2'""), 164-165 Mouse strain SLJ/J(H-2), 164-165 Muller cell, 2 18-2 19 Multi-CSF, 7 1,72 eosinophil proliferation, 75 Multiple sclerosis DP alpha-DQ beta, 149 DPw4,149-150 DQ alpha-DQ beta, 149 DQ alpha gene polymorphism, 149 DQR2-6,148-149 DQwl, 148 DQwla, 148 DQwlb, 148 DR2, 148-150 DR alpha-DQ beta, 149 Dw2 subtype, 148- 150 HLA-A3,148-150 HLA-B7, 148-150 T cell response, 150 Murine colony-stimulating factor, 73 Myelin basic protein alpha-acetylated amino-terminal peptide, 167 dominant epitope, 164- 165 specific tolerance induction, 167 Myeloma IgE, ESR spectra, 30-31 Myeloma IgM, ESR spectra, 30-31 Myeloma protein, 2 Myelosuppression, colony-stimulating factor, 93 Mylein basic protein, 162, 163-166 my-inositol, structure, 255 N
Na+/H+antiporter, intracellular pH, 288 Natural killer cell antLCD2 mAb, 236 prolactin receptor, 180 Neoplasm, immune regulation, 212 Nerve growth factor, 173- I74 Nervous system autoimmune reaction, T cell component, 163-167 Neural autoantigen, dominant epitope, 164-165
Neural cell, immunological mediation, 173- 174 Neuroendocrine mediator acute inflammation, 175-176 chronic inflammation, 177-179 immediate hypersensitivity, 175-176 immunological effects, 175-179 immunological generation, 169- 173 lymphocyte function, 177- 179 Neuroimmunology, 161-184 Neuromediator, 162 cellular source, 170- 172 identification method, 170-172 immunological sources, 170-172 lymphocyte receptors, 167, 179-184 structure, 170-172 Neuropeptidase cellular localization, 167 specificity, 167 Neuropeptide human lymphocyte receptor, 180 IL-I, 178 lymphocyte, 177 differentiation, 178 immunological functions, 178 tissue cycles, 178 tissue homing, 178 macrophage, 177 structure, 167 Neuropeptide mediation, immunological cell, 176, 177 Neurotrophic virus, 165 Neutralizing antibody, colony-stimulating factor, 95-96 Neutrotrophic coronavirus, 14 1 NFdB, 298 NOD mice, insulin-dependent diabetes mellitus-prone, 139 Nucleus, signal transduction, 3 11-3 11
0 Ocular anatomy, 192 Ocular HSV infection anterior chamber-associated immune deviation, 208-2 10 cytotoxic T lymphocyte, 210 delayed-type hypersensitivity, 210 Oligosaccharide, rotation, 26-27
INDEX
One-dimensional isoelectrofocusing, HLA class 11, 114-1 16 Opioid receptor human granulocyte, 180 human lymphocyte, 180 human monocyte, 180 Ornithine decarboxylase half-life, 289 polyamine metabolism, 289-290 P
p55, transcriptional regulation, 294-295, 297-299 pH, Na+/H+antiporter, 288 Phorbol ester, T cell activation, 258-260 Phorbol myristate acetate, 234 Phosphatase, T cell, signal transduction, 284 Phosphatidylinositol4,5-biphosphate,228 Phosphatidylinositol4,5-bipho~phate~, T cell antigenreceptor-CD3 complex Ca2+ response, 262-265 hydrolysis, 262-265 Phosphatidylinositol4.5-biphosphate kinase, 253 Phosphoinositide, 228 hydrolysis, 253-254 T cell, turnover direct measurements, 260 T lymphocyte, turnover, 258-265 Phosphoinositide cycle, 253-255 Phosphoinositide kinase type I, 253 Phosphoinositide kinase type 11,253 Phosphoinositide-specific phospholipase, 256-258 function, 256-257 G protein, 266 inositol 1-monophosphate, 253 lymphocyte, 256 multiple forms, 256 regulation, 257-258 structure, 256-257 Phosphoinositide turnover, T cell, secondary messenger formation, 253-265 Phosphorylation serine, 270-277 T lymphocyte, activation-associated, 269-285
373
threonine, 270-277 tyrosine, 277-283 Phytohemagglutinin, 233 PKC, 270-274 Plasma membrane glycoprotein, 161- 162 Platelet-derived growth factor, 308 Pneumococcus pol ysaccharide, 28 Polyamine metabolism, 288-290 ornithine decarboxylase, 289-290 Polymorphism analysis, 114 Posttranslational processing event, 123 Potassium, ion channel, 287-288 types, 287 Precipitation epitope, 23 segmental flexibility, 23-24 Prolactin receptor, 180-181 natural killer cell, 180 Proopiomelanocortin, 172- 173 Protein kinase, 228 Ca2+-dependent,274-275 CaM-dependent, 274-275 cyclic nucleotide-dependent, 275-277 T cell, signal transduction, 284 Protein kinase C, 79 Protein molecule, dynamic structure, 1 ligand, 1 Protein phosphatase, 283-285 Protooncogene, T cell activation, 291-204 psi-V gene segment diversification, 58-60 diversification sequences, 54-56 gene conversion, 58 VL, recombination event, 56 Putrescine, 288
Q QKRAA, Epstein-Barr virus Dw4 peptide, 130 gp 110 peptide, 130 Quantitative regulatory mechanism, 123 R
Rat RT-6 alloantigen, 241 Recombination, double-homologous, 56 Resting T cell, accessory signal, 247-248 Retina, organ-resident nonlymphoid cell, 218
374
INDEX
Retinal antigen, autoimmune response, 217 Retinal antigen-specific T cell clone, deletion, 217, 218 Retinal glial cell, T helper lymphocyte, 218 RFLP, 1 17- 119 Tuql fragment, 139 RFLP/allele-specificoligonucleotide, 117-1 19 RFLP analysis, pitfalls, 1 17 Rheumatoid arthritis 109 D6 mAb, 128 conformational epitope three-dimensional structure, 129 DR beta gene, 131 DRw53 molecule, 128 Dw4, 126-127 DwlO, 126-127 DW14, 126-127 etiology, 126-132 HLA, 126 MC1, 128 4-kDa RSaI class I1 fragment, celiac disease, 145 RT-6, 241
s S antigen, 215-216 anterior chamber-associated immune deviation, 219 experimental autoimmune uveitis, 219 Segmental flexibility agglutination, 23-24 antidansyl antibody, 13, 14 cell receptor binding, 24-26 complement activation, 24-26 complement binding activity, 13, 15,22 domain structure, 14, 16, 22 evolution, 26 functional implications, 23-26 genetically engineered antibody, 14, 16,22 hinge region structure, 13, 15, 22 Ig molecule, 5- 19 electron microscopy, 2-3 fluorescence polarization, 3-5 hydrodynamic study methods, 2 nanosecond, 4-5
spin-label approach, 5 steady state, 4 study methods, 1-34 X-ray crystallography, 3 IgA, 14-18 IgE,18-19 IgC, 5-19 early studies, 5-6 fluorescence polarization, 6- 10 IgG subclasses, 11- 14 spin-label method, 10-1 1, 12 X-ray crystallography, 11 IgG molecule, hinge region structure, 21-23 IgM, 14-18 mouse monoclonal antibody, 13, 14 precipitation, 23-24 Self-antigen sequestration, 2 17 Self-reactive immune cell, clonal deletion, 216,217 Self-tolerance experimental autoimmune uveitis, 216-2 19 eye, strategies, 2 17 lens, 220 Serine, phosphorylation, 270-277 Serology, HLA class 11, 112-1 14 Sialophorin, 248 Signal transduction lymphokine-mediated, 3 10 nucleus, 3 11-3 11 Skin, suppressor antigen-presenting cell, 206 SOM28, 172 Somatic diversification, Ig VL gene, 60-62 Somatostatin, 169 Spermidine, 288 Spermine, 288 Spleen anterior chamber-associated immune deviation, 206, 2 11 delayed-type hypersensitivity, 2 1 1-2 12 Spontaneous neoplasm, immune regulation, 2 12 Suppression phenomenon, insulin-dependent diabetes mellitus, 143-144 Suppressor antigen-presenting cell, 206 Sympathetic ophthalmia, 216, 217 Synovial cell, HLA class I1 molecule, 132
INDEX
T
TI I . See CD2 T cell. See ~ V Specific J type alternative transmenibrane signaling pathways, 307-309 experimental autoimmune uveitis, 2 16 <; protein, 267-269 ion channel, 285-288 regulation, 285, 286 phosphatase, signal transduction, 284 phosphoinosiride turnover direct measuremenis, 260 secondary messenger formation, 253-265 protein kiriase, signal transduction, 284 signal transduction, (;I‘I’-hinding protein, 265-269 tyrosine protein kinase, 277-280 substrates, 28 1-283 T cell-activating protein, 240-24 1 T cell activation accessory activation molecule, 304-306 alternative activation molecule, 304-3Oti alternative pathway, 235 c-fu\, 29 1-202 c-myb, 293-294 c-my, 292-293 calcium ionophore, 2.58-260 cell cycle progression, 290 de now-induced genes, 290 gene products, 290 IL-1 receptor, 250-253 signal transduction, 251-252 molecular events mediating, 227-3 1 1 nuclear events, 290-299 phorbol ester, 258-260 protooncogene, 291 -204 T cell antigen receptor-CL)S complex, 232-235 T cell antigen receptor, 78-81 T cell antigen receptor--(:D3 complex chains, 229-232 C; protein, 306-307 heterodimers, 229 phos~’hatidylinosito1 4,.5-liphosphate2 (;a’+ response, 262-265 hydrolysis, 262-26.5 polymorphic proteins, 229 structure, 22!)-232
375
synthesis, 232 ?’cell activation, 232-235 T cell clone, 78 colony-stimulating factor gene differential expression, 89-92 GM-CSF preferential synthesis, 91 lymphokine mKNA expression, 87-88 T cell clone reactivity, HLA-Dwl4, 129 T cell cytotoxicity, islet beta cell, 142 T cell-derived lymphokine, in vivo studies, 94-95 T cell hybrid, 78 T cell proliferation, interleukin-2-independerlt, 302-304 T cell receptor V gene experimental autoimmune encephalomyelitis, 166 THcell receptor, 166 T cell-specific tyrosine protein kinase, CD4,243-244 T cell surfice activation molecule, 228-253 T cell synthesis, colony-stimulating factor, 96-98 antigen-presenting cell, 117 local polarization secretion, 97 local response, 96 target cells, 97 T lymphocyte, 70, 7 1 colony-stimulating factor studies in vitro, 77-78 studies in vivo, 92-98 cytotoxic anterior chamber-associated immune deviation, 208, 209 effector T cell, 197 Langerhans cell, 197-198 ocular HSV infection, 2 10 transplant, 197- I98 heterogeneity, 83-92 phosphoinositide, turnover, 258-265 phosphorylation, activ;ition-associatetl, 269-28.5 tyrosine protein kinase, 277-280 T lynlphocyte-derived coI[tny-stimulating factor, 69-98 Tachykinin receptor, 181 TEMPO-amine spin label, 27-28 T h I , 86-89 T h 2 , 86-89
376
INDEX
T H cell receptor experimental autoimmune encephalomyelitis, 166 T cell receptor V gene, 166 Theiler’s murine encephalomyelitis virus, 165 Threonine, phosphorylation, 270-277 Thy-l,238-239 activation element, 238 amino acid sequencing, 238 independent signal-transducing molecule, 239 isolation, 238 Thymocyte-activating molecule, 250 Thyrotropin, 170-172 Tp44,244,245-246 Tp67,246 Tp90,248 Tp103,244,249 Tp135-45,250 Trabecular meshwork, 206-207 Trans-complementation, HLA-DQ hybrid molecule, 119- 120 Transcription, lymphokine synthesis, 8 1 Transcriptional factor insulin gene, 142 transgenic MHC, 142 Transcriptional regulation interleukin-2 gene, 294-297 interleukin-2R alpha gene, 294-295, 297-299 ~55,294-295,297-299 Transformed producer cell line, 69 Transforming growth factor-beta, 169 Transforming retroviral protein, 253 Transgenic MHC,transcriptional factor, 142 Transmembrane signaling, 227 Transplant cytotoxic T lymphocyte, 197-198 delayed-type hypersensitivity, 197 Trinitrophenyl-derived T lymphocyte, 2 12 Tumor allograft anterior chamber, 204 anterior chamber-associated immune deviation, intraocular, 207-208 delayed-type hypersensitivity, 208, 209 Tumor necrosis factor, 70, 71 Tumor necrosis factor alpha, 169 Tumorigenesis anterior chamber, 213-215
anterior chamber-associated immune deviation, 2 13 Two-dimensional polyacrylamide gel electrophoresis, HLA class 11, 114-116 silver staining, 116 theoretical limitations, 116 Tyrosine, phosphorylation, 277-283 Tyrosine protein kinase lymphoid tissue, 277-280 T cell, 277-280 substrates, 281-283 T lymphocyte, 277-280 U
U2666 myeloma B cell, 182 uv5c25 anterior chamber, 213-215 delayed-type hypersensitivity, 2 14 Uveitis, autoimmune. See Experimental autoimmune uveitis V
V gene segment, bursa1 lymphocyte, 52-54 V-J joining, 46-47 circular episome, 47 diversity, 47 random nature, 47 V-J jointing, embryogenesis, 49,50 V-re1 oncogene, 58 Vasoactive intestinal polypeptide, 170-72 VL gene, diversification sequences, 54-56 VL gene segment, 41 diversification, 56-60 gene conversion, 56-60 psi-V gene, recombination event, 56 sequence substitution position, 56, 57 sequence substitution size, 56, 57 VL pseudogene 5, sequence comparison, 55 VL pseudogene 7, sequence comparison, 55 W WEHI-SB, 77 Western blotting, HLA class 11, 116