REVIEWS Genomic views of STAT function in CD4+ T helper cell differentiation John J. O’Shea*, Riitta Lahesmaa‡, Golnaz Vahedi*, Arian Laurence* and Yuka Kanno*
Abstract | Signal transducer and activator of transcription (STAT) proteins are well known for their essential roles in transmitting cytokine-mediated signals and specifying T helper (TH) cell differentiation. Recent technological advances have revealed that STAT proteins have broad and complex roles in gene regulation and epigenetic control, including important roles as functional repressors. However, the challenge of how to link signal transduction, nucleosome biology and gene regulation remains. The relevance of tackling this problem is highlighted by genome-wide association studies that link cytokine signalling and STATs to various autoimmune or immune deficiency disorders. Defining exactly how extrinsic signals control the specification and plasticity of TH cells will provide important insights and perhaps therapeutic opportunities in these diseases. Next-generation sequencing High-throughput sequencing methods that rapidly and inexpensively produce accurate sequencing data that can cover entire genomes. Several different platforms, based on different chemistries, are available, including: the Illumina Genome Analyzer, the Roche 454 Sequencing System, the Applied Biosystems SOLiD System and the Helicos BioSciences HeliScope.
*Molecular Immunology and Inflammation Branch, National Institute of Arthritis and Musculoskeletal and Skin Diseases, National Institutes of Health, Bethesda, Maryland 20892, USA. ‡ Turku Centre for Biotechnology, University of Turku and Åbo Akademi University, P.O. Box 123, FI‑20521 Turku, Finland. Correspondence to J.J.O’S. e‑mail:
[email protected] doi:10.1038/nri2958
CD4+ T cells are essential for host defence, as exempli‑ fied by the effects of the CD4+ T cell depletion that is associated with HIV infection and AIDS. This loss of T cells leads to a profound impairment of the immune response and a range of opportunistic infections. Conversely, CD4+ T cells are also fundamental driv‑ ers of autoimmunity when a loss of tolerance occurs. CD4+ T cells mainly direct immune responses through the cytokines they produce, and our understanding of the range of cytokines produced by CD4+ T cells has increased considerably 1. In addition to T helper 1 (TH1) and TH2 cells, which produce interferon‑γ (IFNγ) and interleukin‑4 (IL‑4), respectively, new subsets of T cells continue to be recognized. These include regulatory T (TReg) cells2, of which there are both natural and induced subsets, and IL‑17‑producing2–4, IL‑9‑producing5–7 and IL‑22‑producing TH cells7,8. In addition, IL‑21‑producing follicular helper T (TFH) cells provide help to B cells, but their identity as a distinct ‘lineage’ and their relationship to other CD4+ T cell subsets remain a source of some controversy 9,10. What is clear is that the cytokine milieu is crucial for CD4+ T cell differentiation. Signal transducer and activa‑ tor of transcription (STAT) family proteins have essential roles in transmitting many cytokine-mediated signals and therefore have similarly crucial roles in TH cell dif‑ ferentiation1,11. The first STAT proteins to be discov‑ ered (STAT1 and STAT2) were identified as inducers of de novo gene transcription in response to interferons (IFNs)12,13. Since then, the essential, non-redundant functions of the seven members of the STAT family have
been extensively defined by generating individual gene knockout mice and by careful analysis of the effects on gene expression14–18. One of the challenges to interpret‑ ing such gene expression data is to distinguish the direct actions of STATs on individual genes from secondary, indirect effects of STAT deficiency. Recent technologies have enabled investigators to construct a genome-wide view of transcription factor binding to distinguish direct from indirect effects. In this Review, we discuss the impact of next-generation sequencing19,20, and illustrate how this technology has allowed us to begin to construct a quantitative map, not only of genome-wide transcription factor binding but also of the effects on genome-wide epigenetic changes. Specifically, we review the genome-wide STAT binding studies that have been reported so far. We discuss the relationship between STAT binding and local epigenetic patterns, and consider how STAT proteins can integrate extrinsic signals to influence epigenetic changes associ‑ ated with T cell lineage commitment. Finally, we review emerging new information regarding mutations and polymorphisms of STAT genes that are associated with human immune disorders.
New technology and genome-wide views STATs are transcription factors that induce the tran‑ scription of their target genes by recognizing and bind‑ ing specific DNA consensus sequences. The direct binding of STATs to DNA was initially analysed by electrophoretic mobility shift assays. Later, analysis of STAT binding to specific genes was carried out using
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 239 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Box 1 | Genome-wide chromatin immunoprecipitation to study protein–DNA interactions Chromatin immunoprecipitation (ChIP) has been used to profile protein–DNA interactions. By choosing appropriate antibodies specific for the protein or epigenetic modification of interest, both transcription factor binding and histone epigenetic marks can now be profiled on a genome-wide scale. After protein-associated DNA fragments have been enriched and purified through immunoprecipitation, the DNA fragments can be measured and mapped to reference genomes by either hybridization (in ChIP-on-chip techniques) or high-throughput next-generation sequencing (in ChIP–seq procedures). ChIP-on-chip, which is based on microarray hybridization technology133,134, has the intrinsic limitation that only pre-selected regions of the genome are included in the arrays, such as proximal gene promoter regions. Also, array-based methods are restricted by the variation and limitations implicit in nucleotide hybridization. By contrast, ChIP–seq covers the entire genome without any preconceived bias19,20,135. Because the DNA fragments of interest are sequenced directly instead of being hybridized to microarray chips, ChIP–seq provides higher resolution, greater genomic coverage, fewer artefacts and a larger dynamic range of signal strength than ChIP-on-chip. Although the relatively short reads (35–75 base pairs) generated by various next-generation sequencing platforms could pose technical difficulties for certain other applications, such as RNA-seq, the technology is well suited for a ChIP–seq approach136. In addition to mapping transcription factor binding and histone epigenetic marks, ChIP–seq has been applied to map the binding of CCCTC-binding factor (CTCF), which regulates chromatin architecture, and histone acetyltransferase p300, which marks enhancer elements27. This next-generation sequencing platform is also used to define nucleosome positioning and accessibility by coupling with micrococcal nuclease digestion (MNase–seq)137 or with the detection of DNase hypersensitivity sites (DNase–seq)138,139. Another application is comprehensive methylome (DNA methylation) mapping140, which will provide further insights into the stable and heritable aspects of the epigenome. Finally, next-generation sequencing has been used to profile various types of RNA, including microRNAs141,142, long non-coding intervening RNAs143 and enhancer RNAs144.
Chromatin immunoprecipitation (ChIP). A technique used to detect the DNA binding sites of specific proteins within chromatin. These assays involve chemical crosslinking of the bound proteins to the DNA, followed by immuno precipitation with an antibody that is specific for the protein of interest.
Epigenetic regulation The heritable, but potentially reversible, states of gene activity that are imposed by the structure of chromatin, such as covalent modifications of DNA or of nucleosomal histones. The epigenome pertains to the aspects of heritable cellular phenotype that are not explained by DNA sequence.
Nucleosome A nucleosome consists of a core of histone proteins with a segment of DNA wrapped around it. It is the minimum unit required to make up a chromosome.
ChIP–seq A technique in which chromatin immunoprecipitation is followed by high-throughput sequencing to generate a genome-wide distribution map of protein–DNA interactions. This technique can be used to measure transcription factor binding or histone modifications.
chromatin immunoprecipitation (ChIP) followed by PCR-
based detection of precipitated DNA using primers specific for pre-selected regions. The application of this type of targeted analysis was inevitably limited to a small subset of genes and regions. However, with the arrival of next-generation sequencing methods, an unbiased genome-wide view of protein–DNA associa‑ tion has become a reality (BOX 1; FIG. 1), allowing us to catalogue the entire range of STAT target genes on a genome-wide scale. Equally important has been the capability to map histone epigenetic marks throughout the entire genome to gain an insight into how chromatin accessibility relates to STAT binding and ultimately to transcriptional regulation. Evolving views of the epigenome. It has become clear since the original discovery of the STATs that, in addi‑ tion to transcription factor binding, a crucial part of gene regulation is epigenetic regulation and that the modi‑ fications comprised by this term are highly dynamic21. It is beyond the scope of the present Review to compre‑ hensively discuss this incredibly active field but, briefly, factors that influence the accessibility of chromatin for active transcription include DNA methylation, ATPdependent nucleosome remodelling and a large number of post-translational histone modifications. Acting together with transcription factors, these chromatin modifications have major effects on gene expression22–24. Early genome-wide maps of histone modifications gen‑ erated by ChIP–seq (ChIP followed by high-throughput sequencing) suggested novel functions for histone modi‑ fications and showed the importance of combinatorial patterns of modifications25,26. Although acetylation is always associated with active chromatin regions, the functional significance of histone methylation with respect to gene expression is more complex. For example, trimethylation of histone H3 lysine 4 (H3K4), H3K36
and H3K79 is associated with active genes (so these modifications are known as ‘permissive’ marks), whereas dimethylation or trimethylation of H3K27, H3K9 and H4K20 is linked to gene silencing (and these are known as ‘repressive’ marks). Importantly, in contrast to clas‑ sic views of the epigenome, it is now clear that some of these modifications can occur rapidly in response to exogenous signals27–29. Therefore, the nucleosome is increasingly viewed as a nuclear sensor that responds to various signals from the cellular environment. Cytokine signalling and the T cell epigenome. The abil‑ ity to measure genome-wide changes in histone modi‑ fications by ChIP–seq provided an opportunity to ask a simple but crucial question about T cell biology, namely whether the observed epigenetic modifications in T cells are more consistent with a model of stable terminal dif‑ ferentiation of CD4+ T cells or with intrinsic flexibility in T cell responses6,7. The stability of various T cell subsets continues to be intensively debated, and striking exam‑ ples of T cell plasticity have appeared in the literature30,31. In this regard, measuring the epigenetic landscape of TH cells has proved to be illuminating. Genome-wide maps of H3K4 (permissive) and H3K27 (repressive) trimethylation in naive CD4+ T cells and fully polarized TH1, TH2, TH17, induced TReg and natural TReg cells have now been obtained32. The data show that the histone modifications of genes encoding signature cytokines of particular TH cell subsets are consistent with a model of terminal commitment, such that permissive marks on a particular cytokine gene are selectively present in the relevant lineage that expresses that cytokine and repressive marks are present in other CD4+ T cell line‑ ages that do not express the cytokine. However, genes encoding ‘master regulator’ transcription factors, such as Tbx21 (T-box 21) for TH1 cells and Gata3 (GATA binding protein 3) for TH2 cells, were found to have ‘bivalent
240 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS poised domains’, meaning that both permissive and repressive histone marks are present in the lineages of alternative fates32. Bivalent domains were originally iden‑ tified in stem cells and they seem to allow for flexibility in gene expression once a cell receives signals for differentia‑ tion33,34. Bivalent domains were also found to be present on genes encoding other key transcription factors, including Runx3 (runt-related transcription factor 3), Bcl6 (B cell lymphoma 6) and Blimp1 (B lymphocyteinduced maturation protein 1; also known as Prdm1)32. Thus, the answer to the question initially posed is that epigenetic analysis has provided evidence for both ter‑ minal commitment (in the case of cytokine genes) and flexible plasticity (in the case of master regulator genes) of TH cells, depending on which genes are examined. As such, the extent to which T cell subsets really behave as ‘lineages’ or as flexible populations will continue to be the focus of ongoing research and controversy. Although many interesting observations have arisen from this genome-wide epigenetic profiling 32, many other questions remain. First, the cell preparation used for these studies was generated in vitro and character‑ ized at a single time point. The dynamic nature of chro‑ matin remodelling and modification over the course of T cell differentiation is yet to be fully elucidated by genome-wide assays. It will be interesting to determine whether the bivalent marks noted on the master tran‑ scriptional regulator genes are already in place in the very early phases of TH cell differentiation to guide the transcriptional programme or whether these marks gradually develop over time. Equally, it will be crucial to determine how the recruitment of STAT proteins affects the deposition or removal of epigenetic marks, and how all the aspects of nucleosome remodelling are acquired over time. We also do not yet know the degree of simi‑ larity between in vitro-generated cells and bona fide TH cells that arise in vivo during the course of infection or autoimmunity in mice or in humans. Profiling of genome-wide STAT binding. Initial work using ChIP–seq to map STAT1 binding sites in the genome revealed more than 11,000 sites in unstimulated HeLa cells and 40,000 sites after IFNγ stimulation35. However, it was not clear from these data whether STAT1 is an important initiator of gene regulation in all cases of binding or whether STAT1 has a major role in creating the local epigenetic patterns around these binding sites. Subsequent work showed that for most genes, deposi‑ tion of the local histone modifications preceded ligandinduced STAT1 binding 36. Although this study was an important breakthrough, it will be important to analyse STAT1 action in various primary cells and to link tran‑ scriptional and epigenetic changes using STAT1‑deficient cells (discussed further below). Since these initial reports, all STATs (with the exception of STAT2) have been pro‑ filed by ChIP–seq, and the original datasets are publicly available through the Gene Expression Omnibus (GEO) repository, as shown in TABLE 1. These datasets will pro‑ vide an enormous resource to promote further genomic research in the scientific community, and the deposition of original datasets to publicly accessible domains such
%JGOKECNN[ETQUUNKPM RTQVGKPs&0#EQORNGZ KPNKXKPIEGNNU
(TCIOGPVEJTQOCVKP D[UQPKECVKQPQT GP\[OCVKEFKIGUVKQP
+OOWPQRTGEKRKVCVG YKVJCPVKDQF[ 4GXGTUGETQUUNKPMU CPFRWTKH[&0#
2TGRCTGNKDTCT[CPFCPCN[UG D[FGGRUGSWGPEKPI 5GSWGPEGTGCFDFWFDWJFDWJDDDFFWJDFJFDJJ 5GSWGPEGTGCFFFJWDWFJDWJDJJDJWFWFWFDJJD 5GSWGPEGTGCFJFWDJWFJDWJDFFDDJWJFDJWFDJ /CRUGSWGPEGF VCIUQPTGHGTGPEG IGPQOG 8KUWCNK\GRGCMU QPIGPQOG DTQYUGT
Figure 1 | Experimental flow of ChIP–seq analysis. A technique to combine chromatin immunoprecipitation 0CVWTG4GXKGYU^+OOWPQNQI[ with next-generation sequencing to map protein–DNA interactions across the whole genome is shown. Chemically crosslinked protein–DNA complexes are immunoprecipitated and the protein-bound DNA fragments are isolated. The crosslinks are reversed, and the purified DNA is used to generate a library for sequencing. Automated reactions yield more than 20 million sequence reads of 36 nucleotides from each sample (using the Illumina Genome Analyzer platform). The sequence reads are aligned onto the reference genome and the distribution of protein–DNA interaction sites is visualized as ‘peaks’ on the genome browser.
as GEO will be crucial for future discoveries. However, it will be important to bear in mind the degree of compat‑ ibility between different datasets (which have been gener‑ ated by different sequencing platforms and by different investigators under different experimental conditions), as this could impose certain limitations on comparable analysis. As the field matures, we await better ways to control the ‘quality’ of these sequencing datasets, includ‑ ing the use of appropriate reference controls to score ChIP–seq peaks37, to allow broader across-the-board analyses. Nevertheless, from the STAT binding data
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 241 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Table 1 | Summary of ChIP–seq analyses of STAT family proteins STAT protein
Stimulus
Cell type
Species
GEO accession number*
Refs
STAT1
IFNγ
HeLa cells
Human
GSE15353
35 37,145
IFNγ
HeLa cells
Human
GSE12783
STAT2
ND
ND
ND
–
STAT3
IL‑6
TH17 cells
Mouse
GSE21670
76
IL‑6
CD4+ T cells
Mouse
GSE26553
80
IL‑21
CD4 T cells
Mouse
GSE19198
89
LIF?
ES cells
Mouse
GSE11431
146
STAT4
IL‑12
TH1 cells
Mouse
GSE22105
42
STAT5A and STAT5B
IL‑2?
CD4 T cells
Mouse
GSE12346
68
STAT6
IL‑4
TH2 cells
Mouse
GSE22105
42
IL-4
TH2 cells
Human
GSE18017
59
+
+
–
ES, embryonic stem; IFN, interferon; IL, interleukin; LIF, leukaemia inhibitory factor; ND, not determined; STAT, signal transducer and activator of transcription; TH, T helper. *All of the original data from the experiments listed in the table are accessible through Gene Expression Omnibus (GEO) using the corresponding GEO accession number.
listed in TABLE 1, we can start to address the unique as well as shared functions of individual STAT proteins in directing epigenetic modifications and gene expres‑ sion in T cells. For the sake of brevity, we discuss only the studies listed in TABLE 1 that provide data derived from T cells.
ChIP-on-chip A technique that combines chromatin immuno precipitation (ChIP) with microarray technology (‘chip’) to investigate protein–DNA interactions in vivo on a genome-wide basis.
Enhancer element A control element in DNA that is bound by regulatory proteins that influence the rate of transcription of the associated gene(s). Enhancers function in an orientation- and positionindependent manner, so they can be located either upstream or downstream of the associated gene, or in an intron.
STAT4 in TH1 cells Landscape of STAT4 targets. Unlike other STATs, which are expressed by a wide range of cell types, STAT4 is pre‑ dominantly expressed by immune cells and the testis. Accordingly, its non-redundant functions are manifested mainly in immune cells, resulting in a very discrete phe‑ notype of STAT4 deficiency involving decreased IFNγ production38. STAT4 is activated mainly by IL‑12, IL‑23 and type I IFNs, and it functions predominantly in pro‑ moting TH1 cell differentiation. STAT4 is also the major regulator of Ifng expression in innate immune cells such as natural killer (NK) cells39,40. Before the advent of genomic approaches, only a limited number of direct STAT4 target genes had been identified (including Ifng, Il18r1 (IL‑18 receptor 1), Hlx (H2.0‑like homeobox), Map3k8 (mitogen-activated protein kinase kinase kinase 8) and Furin). The first effort to increase our knowledge of STAT4 target genes was through the use of ChIP-on-chip technology, which showed that STAT4 bound the promoters of many previ‑ ously unidentified target genes, such as Gadd45g (growth arrest and DNA-damage-inducible 45γ), Lcp2 (lympho cyte cytosolic protein 2) and Myd88 (myeloid differen‑ tiation primary response gene 88)41. The expanded list of STAT4 target genes also showed that not all genomic STAT4 binding events are equal. In some cases the bind‑ ing of STAT4 to a target gene following IL‑12‑mediated stimulation was not translated into a change in gene expression. Thus, binding per se is not the only deter‑ minant of STAT4‑dependent gene programming during TH1 cell differentiation.
Whereas the analysis provided by ChIP-on-chip tech‑ nology is limited to predefined regions of the genome, ChIP–seq data generate an unbiased genome-wide map of where STATs bind. Using ChIP–seq, STAT4 was found to have 10,000 binding sites in in vitrodifferentiated murine TH1 cells, 40% of which were local‑ ized to the promoters or gene bodies of approximately 4,000 annotated genes42. 60% of the STAT4 binding sites occurred in intergenic regions, where some distal enhancer elements are thought to reside, away from anno‑ tated genes. In sharp contrast to the implications of the STAT1 data described above, comparative epigenomic analysis of wild-type versus STAT4‑deficient TH1 cells provided evidence that of the ~4,000 genes bound by STAT4, nearly 1,000 had STAT4‑dependent alterations in epigenetic modifications. And of these 1,000 genes, 200 had highly STAT4‑dependent gene expression, as determined by microarray analysis of wild-type versus STAT4‑deficient cells. These genes therefore represent a core subset of direct STAT4 targets that are highly dependent on STAT4 for promoting gene expression and the local epigenetic signature. Importantly, their dependence on STAT4 cannot be compensated for by other STAT proteins or transcription factors. This gene subset included not only signature TH1 cell genes, such as Ifng and Tbx21, but also others, including Il18rap (IL‑18 receptor accessory protein), Icos (inducible T cell co-stimulator), Lilrb4 (leukocyte immunoglobulin-like receptor B4) and Nkg7 (NK group 7). This implies a potential role for these genes in maintaining the pheno type of fully polarized TH1 cells, and this might be of interest to examine in the future. The analysis of STAT4 target genes also showed that some cytokine genes that were previously considered to define other TH cell lineages were targets of STAT4 in TH1 cells. Although it was initially denoted as a TH2 cell cytokine in mice, subsequent work has shown that IL‑10 is expressed by multiple T cell subsets43. It was interest‑ ing to note that the Il10 gene was bound and positively regulated by STAT4 in TH1 cells, but also by STAT6 in TH2 cells. Similarly, IL‑21 was initially noted to be a product of T cells following T cell receptor stimulation, but was later reported to be produced by TH17 cells in a STAT3‑dependent manner 44–47. More recently, IL‑21 has also been reported as a lineage-defining cytokine for TFH cells10. However, ChIP–seq data have indicated that STAT4 can bind and regulate the Il21 gene, and this is consistent with the recent finding that IL‑12 (which sig‑ nals through STAT4) can induce the expression of IL‑21 in human T cells48. Thus, Il10 and Il21 are two examples of genes that can be regulated by multiple STATs. STAT4 as a transcriptional repressor. Generally, tran‑ scription factors that drive lineage commitment posi‑ tively regulate the expression of phenotype-defining genes, but they can also repress the expression of genes associated with alternative fates. Although STATs were originally discovered as activators of gene transcription and genome-wide analysis has confirmed that they func‑ tion in this way, there have been indications that STATs can also have roles as functional repressors. Microarray
242 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS ,#-s56#6 RCVJYC[
,#-
56#6
6*EGNN
,#-
2
2
#EVKXGEJTQOCVKP
NKPGCIGURGEKȮEIGPGU
2
*-OG 56#6
2
2
4GRTGUUGFEJTQOCVKP
IGPGUHQTQVJGTNKPGCIGU
2
56#6 56#6
*-OG 2
2
2
56#6
*-OG 2
2
*-OG
56#6
*-OG 2
2
2
0WENGWU
'ȭGEVQTE[VQMKPG UGETGVKQP
Figure 2 | Distinctive epigenetic patterns are formed by STAT proteins in differentiated T helper effector cells. A key role of signal transducer and activator of transcription (STAT) proteins includes shaping epigenetic patterns on target gene loci to maintain cell lineage specificity. Five distinct epigenetic patterns were found to be STAT4 dependent 0CVWTG4GXKGYU^+OOWPQNQI[ in T helper 1 (TH1) cells. These patterns included both permissive chromatin signatures (high levels of histone 3 lysine 4 trimethylation (H3K4me3), high levels of H3K36me3 or low levels of H3K27me3) and repressive chromatin signatures (high levels of H3K27me3 or low levels of H3K36me3). Permissive chromatin signatures are found on TH1 cell-expressed genes, whereas repressive chromatin signatures are found on TH2 cell-expressed genes in TH1 cells. JAK, Janus kinase. Figure is modified, with permission, from REF. 42 © (2010) Elsevier Science.
data have provided evidence that the expression of cer‑ tain genes is increased when a given STAT is deleted49,50, but there are few examples of genes for which STATs seem to function as direct transcriptional repressors51,52. In this regard, several possible ways in which a STAT protein could cause gene silencing have been reported, including the recruitment of DNA methyltransferase 1 (DNMT1) and histone deacetylase 1 (HDAC1) 53 or direct interaction with heterochromatin protein 1 (HP1) for heterochromatin formation54. In TH1 cells, STAT4‑dependent repressive histone marks have been identified on several TH2 cell-expressed genes, including
STAT6 target genes, which are actively repressed by STAT4 in TH1 cells. Although the total number of such genes is small (~40 genes), the data clearly point to a role for STAT4 as a transcriptional repressor, in addition to its more widely recognized role as a transcriptional activator (FIG. 2). It is not yet clear how a transcription factor can drive expression of one gene and repress the expression of another gene in the same cell, but it will be informative to analyse the associating factors and proteins that are locally recruited to genes that are bound and repressed by STAT proteins. The successful identification of specific chromatin modifications that
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 243 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS are associated with STAT binding 42 is a different way of using the genomic approach, aside from obtaining a list of target genes. A genomic approach that integrates different types of read-out enables us to ‘examine the forest’ rather than just ‘finding the trees’ in the genome. Not to say that the trees are not interesting, but the big picture is important as well.
STAT6 in TH2 cells STAT6 as a driver of TH2 cell differentiation. TH2 cell differentiation is induced by IL‑4, and the importance of STAT6 for this process has been well established in mice55–58. The actions of STAT6 and its downstream tar‑ gets in pathological TH2 cell responses such as asthma and allergy are also of great interest, given the public health impact of these diseases. Consequently, the func‑ tions of STAT6 have been studied in both mouse16,18 and human59 systems. In human cells, STAT6 mediates the expression of more than 80% of IL‑4‑regulated genes, a higher proportion than was reported in previous studies using mouse cells16. The functions and cellular distribu‑ tions of identified STAT6 targets are varied, reflecting the fundamental role of STAT6 in regulating multiple cellular activities. Genome-wide kinetic profiling of STAT6‑dependent gene expression and analysis of the STAT6‑dependent gene network in humans59 confirm that STAT6 is a major, 56#6 56#6
56#6# 56#6
'2#5
)(+
470: 56#6
$#6(
)#6#
56#6
&KTGEV56#6OGFKCVGFTGIWNCVKQP +PFKTGEV56#6OGFKCVGFTGIWNCVKQP 2TQVGKPsRTQVGKPKPVGTCEVKQP 1VJGTV[RGUQHKPVGTCEVKQPCPFQTTGIWNCVKQP
6$:
7RTGIWNCVGF GZRTGUUKQP &QYPTGIWNCVGF GZRTGUUKQP
Figure 3 | The STAT6 signalling network map identified during the initial TH2 cell differentiation stage. Transcription factors regulated by interleukin‑4 (IL‑4) and signal 0CVWTG4GXKGYU^+OOWPQNQI[ transducer and activator of transcription 6 (STAT6) form a core network of interacting nodes. The expression of proteins shown in red boxes is upregulated and of proteins in green boxes is downregulated by STAT6 in transcriptomics studies. STAT6‑mediated regulation of genes detected by ChIP–seq is marked with red arrows (using a solid line for direct regulation and a dashed line for indirect regulation). Furthermore, known direct interactions between the putative downstream transcriptional regulators of STAT6 in humans have been added to the figure; blue lines correspond to protein–protein interactions and black lines correspond to other types of interaction or regulation. The networks were generated through the use of Ingenuity Pathways Analysis with some modifications based on published reports. BATF, basic leucine zipper transcriptional factor ATF-like; EPAS1, endothelial PAS domain protein 1; GATA3, GATA binding protein 3; GFI1, growth factor independent 1; RUNX1, runt-related transcription factor 1; TBX21, T-box 21; TH, T helper. Figure is modified, with permission, from REF. 59 © (2010) Elsevier Science.
direct contributor to the transcriptional profile asso‑ ciated with the TH2 cell phenotype. The findings also show that IL‑4‑induced regulation of gene transcription in human cells is highly dynamic; only a subset of the genes that were differentially regulated within the first few hours after IL‑4 treatment remained differentially expressed at later time points up to 72 hours. Thus, soon after IL-4 treatment, STAT6 is a major ‘switch signal’ to initiate the TH2 cell differentiation programme, but at later points other factors in addition to STAT6 are required to maintain the acquired TH2 cell phenotype. Genome-wide differential gene expression analysis using small interfering RNA (siRNA) identified 453 genes that are regulated by STAT6 in human cells 59. Only 6% of these genes had been previously identified as STAT6 targets, including GATA3, SOCS1 (suppres‑ sor of cytokine signalling 1) and IL24. The new target genes indicate new functions and processes that might be mediated by STAT6 signalling. In general, the find‑ ings underscore the importance of using genome-wide approaches to explore the species-specific roles of STAT proteins in humans and mice. The early signalling network: connection to different TH cell fates. By gene network analysis, the transcrip‑ tion factors regulated by IL‑4 and STAT6 were found to form a compact core interaction network of signalling 59 (FIG. 3). These data highlight the importance of com‑ binatorial signalling pathways that function together to determine T H cell commitment and fate. Of the newly identified direct STAT6 targets, three transcrip‑ tion factors that form hubs in the regulatory network are of particular interest: RUNX1, EPAS1 (endothe‑ lial PAS domain protein 1) and BATF (basic leucine zipper transcriptional factor ATF-like). RUNX proteins have a central role in regulating TH cell differentiation in general60, but RUNX1 preferentially inhibits TH2 cell differentiation by downregulating GATA3 expression61 and by binding to the IL4 silencer region62. In addi‑ tion, RUNX1 can form a complex with forkhead box P3 (FOXP3) or retinoic acid receptor-related orphan receptor-γ (RORγ; encoded by RORC), and these interactions are necessary for TReg and TH17 cell func‑ tion, respectively 63,64. Interestingly, EPAS1 binds to the RUNX1 promoter, potentially amplifying the effect of STAT6 on RUNX1 expression65. BATF, which is also a direct target of STAT6, regulates both TH17 and TH2 cell differentiation66,67. The connection between the TH2 cell differentiation programme and the programmes for other TH cell sub‑ sets can be further examined through the network of key transcription factors. It is notable that within the STAT6‑mediated TH2 cell differentiation programme there are close connections between STAT6 and other STAT family proteins involving only a few intermedi‑ ate molecules, as shown in FIG. 3. This underscores the importance of understanding the cooperative and antagonistic interactions between the STATs, as well as between downstream transcription factors, that direct TH cell fate. For example, a comparison of ChIP–seq data for STAT6 (REF. 59) and STAT5A68 showed that
244 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS they have overlapping targets and thus indicated that these two STATs might have cooperative roles59. This is consistent with the known contribution of STAT5 to STAT6‑independent TH2 cell commitment 68–70. A stabilizer of the T H2 cell phenotype. The crucial role of STAT6 in the initiation stage of TH2 cell differ‑ entiation is evident, but STAT6 also contributes to the maintenance of the TH2 cell phenotype in differentiated cells. This conclusion has been supported by a combi‑ natorial genome-wide analysis of STAT6 binding and STAT6‑induced epigenetic patterns and gene expres‑ sion42. Similarly to STAT4, STAT6 is responsible for the maintenance of distinct epigenetic patterns on selected target genes. STAT6 mainly functions as a transcrip‑ tional activator but, as is the case with STAT4, it is also a functional repressor for a subset of genes. In terms of activating target genes, STAT6 more frequently opposes the deposition of repressive epigenetic marks than it promotes permissive epigenetic marks; in this regard, STAT6 seems to be subtly different from STAT4. Of particular interest is a subset of genes that are bound by STAT4 in TH1 cells but by STAT6 in TH2 cells, for which the two STATs have opposing effects on local epigenetic patterns. A notable example is the Il18r1– Il18rap locus42. So, whereas one STAT (in this case STAT4) promotes permissive marks, the other STAT (STAT6) promotes repressive marks on the same locus. This divergent action of STAT4 and STAT6 on the same genes provides an insurance for inducing gene expres‑ sion in one lineage and repressing gene expression in the other lineage (FIG. 2).
STAT3 and TH17 cell differentiation TH cells that selectively produce IL‑17 (known as TH17 cells) are one of the newest T cell subsets to be recog‑ nized. They have crucial roles in host defence against extracellular bacteria and fungi, but also in the patho‑ genesis of various autoimmune diseases2,3. Cytokines that promote IL‑17 production include IL‑1, transforming growth factor‑β1 (TGFβ1), IL‑6, IL‑21 and IL‑23. The last three of these cytokines activate STAT3. Although STAT3 is activated by a large number of cytokines and has crucial functions in various tissues71,72, T cellspecific deletion of STAT3 mainly affects the expression of IL‑17 and IL‑21 (REFS 73–75), and consequently results in decreased severity of several autoimmune disease models73,76–79. Conversely, an increase in STAT3 activa‑ tion, through the deletion of Socs3, results in increased numbers of TH17 cells74. Landscape of STAT3 targets. ChIP–seq analysis of STAT3 binding in T cells, coupled with gene expression analy‑ sis, has confirmed that the Il17 and Il21 genes are direct targets of STAT3 (REFS 46,74,76). Of note, STAT3 binds to multiple sites in the Il17 locus80, the most prominent of which are intergenic regions that coincide with con‑ served non-coding sequences (CNS)81. These sites also bind histone acetyltransferase p300 and so are probably enhancer elements (G.V., Y.K. and J.J.O’S., unpublished observations). Furthermore, in an analogous role to that
of STAT6 in TH2 cells, STAT3 directly binds to multiple genes encoding transcription factors that are crucial for programming TH17 cells. These include Rorc 82, Rora83, Ahr (aryl hydrocarbon receptor)84, Batf 66, Irf4 (inter‑ feron regulatory factor 4)85 and Maf 86. Other important direct targets of STAT3 that define the TH17 cell phe‑ notype include Il23r and Il6ra45,76. Notably, the ability of STAT3 to positively regulate the expression of these genes is associated with the presence of permissive H3K4me3 marks. The prominent role of STAT3 in the specification of TH17 cells led to a re-evaluation of the factors involved in this process. Although TGFβ signal‑ ling is usually required for TH17 cell differentiation, an alternative mode of TH17 cell generation that can occur in the absence of TGFβ has also been recognized 87. It was found that activation of STAT3 together with IL‑1 was sufficient to promote expression of IL‑23R87. Acquisition of this receptor by T cells allowed respon‑ siveness to IL‑23, which has a major role in driving pathogenic IL‑17‑dependent responses88. Thus, patho‑ genic TH17 cells were generated in the absence of TGFβ signalling 87 through the activation of STAT3 and other cooperating factors. In this regard, it is interesting that genome-wide STAT3 binding sites overlap significantly with those for IRF4 after IL‑21 stimulation89. In addition to confirming the role of STAT3 in regu‑ lating the expression of TH17 cell-related cytokines and transcription factors, ChIP–seq analysis has pointed to a role for STAT3 in regulating T cell proliferation and sur‑ vival. Newly identified STAT3 target genes in T cells that might mediate these functions include the anti-apoptotic genes Bcl2 and Ier3 (immediate early response 3), and the proto-oncogenes Fos, Jun and Fosl2. Although no pheno‑ typical evidence for this function was reported in the ini‑ tial description of STAT3 deletion in T cells90, the delayed proliferation and poor clonal expansion of Stat3–/– T cells, particularly in the setting of inflammation76, are consistent with the ChIP–seq findings. Complex roles of STAT3 in TReg cells. IL‑6 inhibits FOXP3 expression, and this effect depends on STAT3 (REF. 91). Accordingly, deletion of Stat3 in T cells results in the clonal expansion of induced TReg cells in the set‑ ting of colitis, but not in the normal gut 76, which is consistent with the relief of IL‑6‑mediated inhibition of TReg cells. Curiously, when Stat3 was deleted only in the TReg cell population, the ability of TReg cells to constrain a pathogenic TH17 cell response was selectively impaired, whereas suppression of T H1 or T H2 cell responses remained intact 92. These data indicate that intrinsic acti‑ vation of STAT3 in TReg cells endows these cells with the ability to specifically suppress TH17 cell responses. Gene expression analysis of STAT3‑deficient TReg cells showed impaired expression of genes potentially contributing to the suppressor function of TReg cells. These genes included Prf1 (perforin 1), Gzmb (granzyme B), Klrg1 (killer cell lectin-like receptor G1), Ccr6 (CC-chemokine receptor 6), Il1r1 and Il6ra. As additional ChIP–seq datasets are gen‑ erated, it will be of considerable interest to dissect how STAT3 controls TReg cell-mediated suppression of specific TH cell subsets.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 245 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS
Genome-wide association study A study in which genome-wide genetic variation is linked to a particular phenotype, most often a clinical disorder, by applying high-throughput genotyping techniques to profile single nucleotide polymorphisms (SNPs) of control subjects compared with patients.
STAT5 and TReg cell differentiation Essential regulators of lymphoid development and peripheral tolerance. The STAT5A and STAT5B genes are adjacent on the same chromosome in both mice and humans, have 96% sequence similarity and have overlap‑ ping functions in diverse tissues93,94. Similarly to STAT3 deficiency, germline deletion of Stat5a and Stat5b (col‑ lectively referred to as Stat5) is embryonic lethal95,96. The few mice that survive are extremely runted and anaemic. Stat5 deletion has marked effects on all lymphoid line‑ ages, including T cells (thymic and peripheral), B cells and NK cells, pointing to a crucial role for STAT5 in lymphoid development. It should also be noted that the targeted disruption of Stat5a or Stat5b individually yields distinct phenotypes, which indicates that there are signalling mechanisms unique to each94. The target genes of STAT5 in T cells68 and other cells have been elucidated using ChIP–seq. These data have helped to explain the role of STAT5 in TH2 cell differen‑ tiation by showing that STAT5 upregulates expression of the IL‑4 receptor 68. However, there has been remark‑ ably little analysis so far of the non-redundant roles of STAT5 in regulating the development and survival of T cells. Given the profound effects of STAT5 defi‑ ciency in T cells, more extensive analysis of the STAT5 ChIP–seq datasets is warranted. The few peripheral T cells that develop in STAT5‑deficient mice have an activated phenotype, and this leads to the development of autoimmunity 97–99. One major factor underlying this autoimmune phenotype is impaired TReg cell development in both the thymus and the periphery owing to deletion of Stat5 in CD4+ T cells. Indeed, STAT5 binds to the promoter and first intron of the Foxp3 gene to activate transcription of this TReg cell master regulator 95,96,100. In addition, STAT5 influences the survival of TReg cells by regulating the expression of the IL‑2R α‑chain (also known as CD25) and the anti-apoptotic protein BCL‑2. Although it is probable that STAT5 regulates many aspects of lym‑ phoid survival, a direct comparison of STAT5 target genes in TReg cells and conventional CD4+ T cells has not been carried out. Equally, a direct comparison of STAT5 and STAT3 target genes in TReg cells would be particularly interesting. Whereas IL‑2‑mediated activation of STAT5 results in the upregulation of FOXP3 expression, which is indis‑ pensable for the maintenance of TReg cells, other STATs negatively influence FOXP3 expression. For example, IL‑4‑mediated STAT6 activation, IL‑12‑mediated STAT4 activation and IL‑6‑mediated STAT3 activation all decrease the expression of FOXP3 and affect chromatin modification at the Foxp3 locus73,95,101,102. However, the exact mechanisms by which STAT4 and STAT6 function to negatively regulate TReg cells have not been elucidated on a genome-wide scale. Similarly to how the activation of STAT4 and STAT6 determines TH1 versus TH2 cell differentiation, the acti‑ vation of STAT5 and STAT3 seems to dictate the dichot‑ omy of TReg cells and TH17 cells80. In addition to its role in positively regulating TReg cell function, STAT5 inhib‑ its TH17 cell differentiation103. To address the potential
mechanisms underlying this action, mapping of STAT5 targets in IL‑2‑activated TH17 cells was carried out by ChIP–seq. One important finding was the extensive over‑ lap between STAT3 and STAT5 binding sites in the Il17 gene80. It was found that STAT5 competes with STAT3 for binding to Il17 and inhibits the function of STAT3 in activating Il17 transcription. The opposing effects of STAT3 and STAT5 on Il17 transcription explain why IL‑2 inhibits IL‑17 production, although the effects of other signalling molecules that are activated by these distinct cytokines might also contribute to this phenotype. In many other cases, it is probable that STAT3 and STAT5 work together to enhance gene expression but, given the example of Il17, it is clear that these two highly related transcription factors can act in opposition. Exactly how these factors can accomplish global versus gene-specific effects warrants further investigation.
STATs and human disease In addition to abundant data pointing to crucial func‑ tions of STAT proteins in animal models, evidence of the importance of STATs in humans is quickly emerging from studies of patients with primary immunodeficiency and autoimmune diseases. New insights into primary immunodeficiency disorders. Previous work has established that STAT5 mutations in humans are associated with impaired TReg cell func‑ tion104 and that STAT1 mutations are associated with susceptibility to viral and mycobacterial infections105,106. Furthermore, recent work has established that another classic primary immunodeficiency, hyper-IgE syndrome (HIES; also known as Job’s syndrome), is a result of dominant-negative mutations of STAT3 (REFS 107,108). This finding was interesting because germline deletion of Stat3 in mice is embryonic lethal. Thus, the restricted pathology seen in humans with STAT3 mutations was not anticipated. Presumably this is because the mutant allele interferes with but does not totally abrogate STAT3 function. A classic feature of HIES is infection without the typical signs of inflammation (such as redness and warmth), resulting in ‘cold abscesses’. It is tempting to speculate that this unique feature of HIES is related to the absence of TH17 cells in this disorder 109–112, as the lack of IL‑17 would lead to a failure to recruit neutrophils to sites of infection. It remains to be elucidated how the impaired function of STAT3 in other tissues contributes to the pathology seen in HIES in various tissues. Genetic polymorphisms and human autoimmunity. Although various animal models have implicated STATs and altered cytokine signalling in autoimmunity, the issue always arises as to whether these models really mirror immunopathogenic mechanisms in humans. However, large-scale genome-wide association studies have now pro‑ vided evidence that various genes involved in cytokine signalling, and STATs in particular, are linked to the devel‑ opment of autoimmunity in humans (TABLE 2). For exam‑ ple, polymorphisms in STAT3 are linked to susceptibility to Crohn’s disease (a form of inflammatory bowel dis‑ ease) and ankylosing spondylitis113,114. Equally compelling
246 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Table 2 | Human autoimmune diseases that are linked to JAK–STAT signalling pathways Disease
Immunological phenotype
Gene
Mutation and linkage
Ref
Hyper-IgE syndrome
Skin abscesses, cystic lung infection, increased serum IgE level, impaired TH17 cell generation
STAT3
Missense mutations result in a dominant-negative protein
Immune dysfunction and growth hormone insensitivity
Impaired TReg cell function
STAT5b
A missense mutation (A630P) in the SH2 domain leads to a failure to respond to activation signals
Immune dysfunction
Susceptibility to infection
STAT1
A missense mutation (L600P or L706S) or a truncating mutation (1757–1758delAG) leads to loss of function
105,106
Crohn’s disease
Overactive mucosal immune reaction in the gastrointestinal tract triggered by commensal intestinal bacteria
STAT3 and JAK2
11 previously reported loci and 21 additional loci, including STAT3 and JAK2, were linked to Crohn’s disease by meta-analysis
113
STAT4
A SNP (rs7574865) is linked to early onset and colonic Crohn’s disease
122
107,108
104
Rheumatoid arthritis
Immune reaction against the lining of small joints
STAT4
A SNP (rs7574865) is linked to both rheumatoid arthritis and SLE
121
Systemic lupus erythematosus (SLE)
Systemic immune reaction against own tissue and organs
STAT4
A SNP (rs7574865) is linked to both rheumatoid arthritis and SLE
121
Primary Sjögren’s syndrome (pSS)
Inflammation of salivary and lacrimal glands, leading to a dry mouth and dry eyes
STAT4
A SNP (rs7574865) is linked to pSS
123
Components of the JAK–STAT signalling pathway have been identified as causal genes for autoimmune diseases and have also been implicated in genetic linkage studies as having statistically significant differences between patients and controls. JAK, Janus kinase; SH2, SRC homology 2; SNP, single nucleotide polymorphism; STAT, signal transducer and activator of transcription; TH17, T helper 17; TReg, regulatory T.
is the evidence that polymorphisms of IL23R and JAK2 (Janus kinase 2) are linked to the same diseases115–120, sug‑ gesting a profound involvement of the IL‑23–STAT3 axis in the genesis of autoimmune diseases. In multiple studies, a variant allele of STAT4 has been found to be associated with an increased risk of developing systemic lupus erythematosus (SLE), rheumatoid arthritis, Sjögren’s syndrome and Crohn’s disease121–123. The connection between STAT4 and SLE is perhaps unexpected insofar as this disease is not a prototypical TH1 cell-mediated disease. However, it is worth noting that STAT4 can be activated by type I IFNs124, and an important aspect of the pathogenesis of SLE is the ‘interferon signature’125. Polymorphisms in TYK2 (tyrosine kinase 2), the gene product of which is activated by IFNs, IL‑12 and other cytokines, have also been reported to be associated with SLE126, pro‑ viding further evidence in support of the importance of STAT4 in the pathophysiology of this disorder. As the STAT4 polymorphisms do not fall within the cod‑ ing region of the gene, they presumably influence the level of gene expression, but clearly much more work is required to confirm this hypothesis.
Concluding remarks and future directions The powerful genome-wide approaches now available to researchers have enabled a comprehensive evaluation of the role of individual STAT proteins in specifying TH cell lineages and a quantitative determination of the target genes that are mobilized during the process of TH cell dif‑ ferentiation. These findings have established that STATs have multiple roles during the initiation stage, as well as the maintenance stage, of a TH cell fate decision. During
the maintenance stage, a key role of STATs involves the induction and/or preservation of epigenetic patterns on target gene loci. STAT proteins induce both permissive and repressive epigenetic modifications. Although a particular STAT can be assigned to each TH cell line‑ age as a dominant regulatory factor, it is clear that this is an overly simplistic way of defining TH cell lineages. Emerging evidence points to the existence of a functional network, in which STATs work both cooperatively and in opposition with each other and with other transcription factors to ensure the desired balance between different T cell fates. In certain cases, this network might also promote phenotypical plasticity. Fortunately, we now have genome-wide approaches to define the breadth of transcription factor action. In addition, we also have the ability to carry out many chromatin-related assays on a genome-wide scale to examine the activity of genomic regions127 through com‑ mon chromatin signatures128, and to determine the state of dynamic genomic organization (FIG. 4). In particular, the extensive coverage of the genome afforded by nextgeneration sequencing offers the possibility of exploring so called ‘gene desert’ or intergenic regions for distal enhancers and other types of regulatory element 21,24,129. This is an exciting opportunity to analyse previously unexplored regions of the genome and, in fact, recent reports have shown that the patterns of distal enhancers are quite unique and different in different cell types130. The challenge now, of course, is to understand how ‘master regulators’ of cell fate and other transcription factors, such as STATs, contribute to the activity of distal enhancers in a manner that creates cell identity 21,130. It is quite possible that some of the polymorphisms that
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 247 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS 2TQOQVGT
&*5UKVG
+HPI
56#6 *-OG 6*EGNNU
*-OG *-OG *-OG
6*EGNNU
*-OG %QPUGTXCVKQP %6%(DKPFKPI UKVG
)GPG DQF[
'PJCPEGT
%6%(DKPFKPI UKVG
Figure 4 | Markers of genomic organization to define activities of chromosome regions. Genomic organization encompassing the interferon‑γ (Ifng) locus in T helper (TH) cells. In TH1 cells, in which the Ifng locus is actively transcribed in a signal transducer and activator of transcription 4 (STAT4)-dependent manner, the promoter0CVWTG4GXKGYU^+OOWPQNQI[ is marked by permissive histone 3 lysine 4 trimethylation (H3K4me3) and STAT4 binding, and the gene body is marked by permissive H3K4me3 and H3K36me3 modifications. One of the distal enhancer elements (shaded in yellow) is marked by H3K4me1 and STAT4 binding in TH1 cells and by repressive H3K27me3 in TH2 cells. Further upstream of the Ifng locus, an insulator site marked by CCCTC-binding factor (CTCF) binding is located and all permissive histone marks and DNase hypersensitivity (DHS) sites are restricted beyond that point.
have been linked to autoimmune diseases reside in enhancer regions of the genome131 that are crucial for regulating tissue-specific patterns of gene expression from a distance. Equally, the entire notion of the epigenome is in the midst of a not so quiet revolution27,132. Increasingly, epigenetics is being viewed as an extension of signal transduction. Nonetheless, it is certainly not clear how all of the components of epigenetic information are linked to each other, to signalling and to transcrip‑ tion factor binding. Indeed, our understanding of how
1. 2. 3. 4.
5.
6. 7. 8. 9.
Zhu, J., Yamane, H. & Paul, W. E. Differentiation of effector CD4 T cell populations. Annu. Rev. Immunol. 28, 445–489 (2010). Littman, D. R. & Rudensky, A. Y. Th17 and regulatory T cells in mediating and restraining inflammation. Cell 140, 845–858 (2010). Korn, T., Bettelli, E., Oukka, M. & Kuchroo, V. K. IL‑17 and Th17 cells. Annu. Rev. Immunol. 27, 485–517 (2009). Veldhoen, M., Hocking, R. J., Atkins, C. J., Locksley, R. M. & Stockinger, B. TGFβ in the context of an inflammatory cytokine milieu supports de novo differentiation of IL‑17‑producing T cells. Immunity 24, 179–189 (2006). Veldhoen, M. et al. Transforming growth factor-β ‘reprograms’ the differentiation of T helper 2 cells and promotes an interleukin 9‑producing subset. Nature Immunol. 9, 1341–1346 (2008). Locksley, R. M. Nine lives: plasticity among T helper cell subsets. J. Exp. Med. 206, 1643–1646 (2009). O’Shea, J. J. & Paul, W. E. Mechanisms underlying lineage commitment and plasticity of helper CD4+ T cells. Science 327, 1098–1102 (2010). Wolk, K., Witte, E., Witte, K., Warszawska, K. & Sabat, R. Biology of interleukin‑22. Semin. Immunopathol. 32, 17–31 (2010). Yu, D., Batten, M., Mackay, C. R. & King, C. Lineage specification and heterogeneity of T follicular helper cells. Curr. Opin. Immunol. 21, 619–625 (2009).
signal transduction and nucleosome biology are related is really still in its infancy. Because the STAT pathway provides a rapid means of transmitting signals from the extracellular environment to the nucleus, investigating how cytokines drive TH cell specification will provide remarkable opportunities to link signal transduction, transcription factor binding, nucleosome biology and gene expression. Hopefully, valuable new information about gene regulation will come from systematic data‑ base analysis of STAT-mediated cytokine signalling in differentiating T cells.
10. Fazilleau, N., Mark, L., McHeyzer-Williams, L. J. & McHeyzer-Williams, M. G. Follicular helper T cells: lineage and location. Immunity 30, 324–335 (2009). 11. Levy, D. E. & Darnell, J. E. Jr. STATs: transcriptional control and biological impact. Nature Rev. Mol. Cell Biol. 3, 651–662 (2002). 12. Shuai, K., Stark, G. R., Kerr, I. M. & Darnell, J. E. Jr. A single phosphotyrosine residue of Stat91 required for gene activation by interferon-γ. Science 261, 1744–1746 (1993). One of the first reports to define the importance of STAT1 in interferon signalling. 13. Schindler, C. & Darnell, J. E. Jr. Transcriptional responses to polypeptide ligands: the JAK–STAT pathway. Annu. Rev. Biochem. 64, 621–651 (1995). 14. Leonard, W. J. & O’Shea, J. J. Jaks and STATs: biological implications. Annu. Rev. Immunol. 16, 293–322 (1998). 15. Hoey, T. et al. Distinct requirements for the naturally occurring splice forms Stat4α and Stat4β in IL‑12 responses. EMBO J. 22, 4237–4248 (2003). 16. Chen, Z. et al. Identification of novel IL‑4/Stat6‑ regulated genes in T lymphocytes. J. Immunol. 171, 3627–3635 (2003). 17. Lund, R. J., Chen, Z., Scheinin, J. & Lahesmaa, R. Early target genes of IL‑12 and STAT4 signaling in Th cells. J. Immunol. 172, 6775–6782 (2004). 18. Tuomela, S., Rautajoki, K. J., Moulder, R., Nyman, T. A. & Lahesmaa, R. Identification of novel Stat6 regulated
248 | APRIL 2011 | VOLUME 11
19.
20. 21. 22. 23. 24. 25.
26.
proteins in IL‑4‑treated mouse lymphocytes. Proteomics 9, 1087–1098 (2009). Mardis, E. R. Next-generation DNA sequencing methods. Annu. Rev. Genomics Hum. Genet. 9, 387–402 (2008). This review compares and discusses the basic chemistries used by currently available next-generation sequencing platforms. Shendure, J. & Ji, H. Next-generation DNA sequencing. Nature Biotech. 26, 1135–1145 (2008). Natoli, G. Maintaining cell identity through global control of genomic organization. Immunity 33, 12–24 (2010). Kouzarides, T. Chromatin modifications and their function. Cell 128, 693–705 (2007). Jenuwein, T. & Allis, C. D. Translating the histone code. Science 293, 1074–1080 (2001). Heintzman, N. D. & Ren, B. Finding distal regulatory elements in the human genome. Curr. Opin. Genet. Dev. 19, 541–549 (2009). Barski, A. et al. High-resolution profiling of histone methylations in the human genome. Cell 129, 823–837 (2007). One of the first epigenome mapping studies to report genome-wide distribution of 20 histone modification marks. Wang, Z. et al. Combinatorial patterns of histone acetylations and methylations in the human genome. Nature Genet. 40, 897–903 (2008).
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS 27. Mohammad, H. P. & Baylin, S. B. Linking cell signaling and the epigenetic machinery. Nature Biotech. 28, 1033–1038 (2010). 28. Liefke, R. et al. Histone demethylase KDM5A is an integral part of the core Notch–RBP‑J repressor complex. Genes Dev. 24, 590–601 (2010). 29. Zhang, X., Edwards, J. P. & Mosser, D. M. Dynamic and transient remodeling of the macrophage IL‑10 promoter during transcription. J. Immunol. 177, 1282–1288 (2006). 30. Zhou, X. et al. Instability of the transcription factor Foxp3 leads to the generation of pathogenic memory T cells in vivo. Nature Immunol. 10, 1000–1007 (2009). 31. Hegazy, A. N. et al. Interferons direct Th2 cell reprogramming to generate a stable GATA‑3+T-bet+ cell subset with combined Th2 and Th1 cell functions. Immunity 32, 116–128 (2010). 32. Wei, G. et al. Global mapping of H3K4me3 and H3K27me3 reveals specificity and plasticity in lineage fate determination of differentiating CD4+ T cells. Immunity 30, 155–167 (2009). The presence of bivalent epigenetic marks on transcription factor genes noted in this study helps to explain emerging views of TH cell plasticity. 33. Bernstein, B. E., Meissner, A. & Lander, E. S. The mammalian epigenome. Cell 128, 669–681 (2007). 34. Wang, Z., Schones, D. E. & Zhao, K. Characterization of human epigenomes. Curr. Opin. Genet. Dev. 19, 127–134 (2009). 35. Robertson, G. et al. Genome-wide profiles of STAT1 DNA association using chromatin immunoprecipitation and massively parallel sequencing. Nature Methods 4, 651–657 (2007). 36. Robertson, A. G. et al. Genome-wide relationship between histone H3 lysine 4 mono- and trimethylation and transcription factor binding. Genome Res. 18, 1906–1917 (2008). 37. Auerbach, R. K. et al. Mapping accessible chromatin regions using Sono-Seq. Proc. Natl Acad. Sci. USA 106, 14926–14931 (2009). 38. Kaplan, M. H. STAT4: a critical regulator of inflammation in vivo. Immunol. Res. 31, 231–242 (2005). An excellent review that summarizes the function of STAT4 in immune responses. 39. Uemura, A. et al. Natural killer cell is a major producer of interferon γ that is critical for the IL‑12‑induced anti-tumor effect in mice. Cancer Immunol. Immunother. 59, 453–463 (2010). 40. Miyagi, T. et al. High basal STAT4 balanced by STAT1 induction to control type 1 interferon effects in natural killer cells. J. Exp. Med. 204, 2383–2396 (2007). 41. Good, S. R. et al. Temporal induction pattern of STAT4 target genes defines potential for Th1 lineage-specific programming. J. Immunol. 183, 3839–3847 (2009). 42. Wei, L. et al. Discrete roles of STAT4 and STAT6 transcription factors in tuning epigenetic modifications and transcription during T helper cell differentiation. Immunity 32, 840–851 (2010). 43. Saraiva, M. & O’Garra, A. The regulation of IL‑10 production by immune cells. Nature Rev. Immunol. 10, 170–181 (2010). 44. Nurieva, R. et al. Essential autocrine regulation by IL‑21 in the generation of inflammatory T cells. Nature 448, 480–483 (2007). 45. Zhou, L. et al. IL‑6 programs TH‑17 cell differentiation by promoting sequential engagement of the IL‑21 and IL‑23 pathways. Nature Immunol. 8, 967–974 (2007). 46. Wei, L., Laurence, A., Elias, K. M. & O’Shea, J. J. IL‑21 is produced by Th17 cells and drives IL‑17 production in a STAT3‑dependent manner. J. Biol. Chem. 282, 34605–34610 (2007). 47. Spolski, R. & Leonard, W. J. Interleukin‑21: basic biology and implications for cancer and autoimmunity. Annu. Rev. Immunol. 26, 57–79 (2007). 48. Schmitt, N. et al. Human dendritic cells induce the differentiation of interleukin‑21‑producing T follicular helper-like cells through interleukin‑12. Immunity 31, 158–169 (2009). 49. Gil, M. P. et al. Biologic consequences of Stat1‑independent IFN signaling. Proc. Natl Acad. Sci. USA 98, 6680–6685 (2001). 50. Ramana, C. V. et al. Stat1‑independent regulation of gene expression in response to IFN-γ. Proc. Natl Acad. Sci. USA 98, 6674–6679 (2001). 51. Walker, S. R., Nelson, E. A. & Frank, D. A. STAT5 represses BCL6 expression by binding to a regulatory region frequently mutated in lymphomas. Oncogene 26, 224–233 (2007).
52. Tran, T. H. et al. Prolactin inhibits BCL6 expression in breast cancer through a Stat5a‑dependent mechanism. Cancer Res. 70, 1711–1721 (2010). 53. Zhang, Q. et al. STAT3- and DNA methyltransferase 1‑ mediated epigenetic silencing of SHP‑1 tyrosine phosphatase tumor suppressor gene in malignant T lymphocytes. Proc. Natl Acad. Sci. USA 102, 6948–6953 (2005). 54. Shi, S. et al. Drosophila STAT is required for directly maintaining HP1 localization and heterochromatin stability. Nature Cell Biol. 10, 489–496 (2008). 55. Kaplan, M. H., Schindler, U., Smiley, S. T. & Grusby, M. J. Stat6 is required for mediating responses to IL‑4 and for development of Th2 cells. Immunity 4, 313–319 (1996). 56. Shimoda, K. et al. Lack of IL‑4‑induced Th2 response and IgE class switching in mice with disrupted Stat6 gene. Nature 380, 630–633 (1996). 57. Takeda, K. et al. Essential role of Stat6 in IL‑4 signalling. Nature 380, 627–630 (1996). 58. Zhu, J., Guo, L., Watson, C. J., Hu-Li, J. & Paul, W. E. Stat6 is necessary and sufficient for IL‑4’s role in Th2 differentiation and cell expansion. J. Immunol. 166, 7276–7281 (2001). 59. Elo, L. L. et al. Genome-wide profiling of interleukin‑4 and STAT6 transcription factor regulation of human TH2 cell programming. Immunity 32, 852–862 (2010). The first report to show genome-wide binding of STAT6 in human TH2 cells. 60. Collins, A., Littman, D. R. & Taniuchi, I. RUNX proteins in transcription factor networks that regulate T‑cell lineage choice. Nature Rev. Immunol. 9, 106–115 (2009). 61. Komine, O. et al. The Runx1 transcription factor inhibits the differentiation of naive CD4+ T cells into the Th2 lineage by repressing GATA3 expression. J. Exp. Med. 198, 51–61 (2003). 62. Naoe, Y. et al. Repression of interleukin‑4 in T helper type 1 cells by Runx/Cbfβ binding to the Il4 silencer. J. Exp. Med. 204, 1749–1755 (2007). 63. Kitoh, A. et al. Indispensable role of the Runx1–Cbfβ transcription complex for in vivo-suppressive function of FoxP3+ regulatory T cells. Immunity 31, 609–620 (2009). 64. Zhang, F., Meng, G. & Strober, W. Interactions among the transcription factors Runx1, RORγt and Foxp3 regulate the differentiation of interleukin 17‑producing T cells. Nature Immunol. 9, 1297–1306 (2008). 65. Mole, D. R. et al. Genome-wide association of hypoxiainducible factor (HIF)-1α and HIF‑2α DNA binding with expression profiling of hypoxia-inducible transcripts. J. Biol. Chem. 284, 16767–16775 (2009). 66. Schraml, B. U. et al. The AP‑1 transcription factor Batf controls TH17 differentiation. Nature 460, 405–409 (2009). 67. Betz, B. C. et al. Batf coordinates multiple aspects of B and T cell function required for normal antibody responses. J. Exp. Med. 207, 933–942 (2010). 68. Liao, W. et al. Priming for T helper type 2 differentiation by interleukin 2‑mediated induction of interleukin 4 receptor α-chain expression. Nature Immunol. 9, 1288–1296 (2008). 69. Takatori, H. et al. Stat5a inhibits IL‑12‑induced Th1 cell differentiation through the induction of suppressor of cytokine signaling 3 expression. J. Immunol. 174, 4105–4112 (2005). 70. Zhu, J., Cote-Sierra, J., Guo, L. & Paul, W. E. Stat5 activation plays a critical role in Th2 differentiation. Immunity 19, 739–748 (2003). 71. Murray, P. J. The JAK–STAT signaling pathway: input and output integration. J. Immunol. 178, 2623–2629 (2007). 72. Akira, S. Roles of STAT3 defined by tissue-specific gene targeting. Oncogene 19, 2607–2611 (2000). 73. Yang, X. O. et al. STAT3 regulates cytokine-mediated generation of inflammatory helper T cells. J. Biol. Chem. 282, 9358–9363 (2007). 74. Chen, Z. et al. Selective regulatory function of Socs3 in the formation of IL‑17‑secreting T cells. Proc. Natl Acad. Sci. USA 103, 8137–8142 (2006). 75. Mathur, A. N. et al. Stat3 and Stat4 direct development of IL‑17‑secreting Th cells. J. Immunol. 178, 4901–4907 (2007). 76. Durant, L. et al. Diverse targets of the transcription factor STAT3 contribute to T cell pathogenicity and homeostasis. Immunity 32, 605–615 (2010). 77. Nishihara, M. et al. IL‑6–gp130–STAT3 in T cells directs the development of IL‑17+ Th with a minimum effect on that of Treg in the steady state. Int. Immunol. 19, 695–702 (2007).
NATURE REVIEWS | IMMUNOLOGY
78. Harris, T. J. et al. Cutting edge: an in vivo requirement for STAT3 signaling in TH17 development and TH17‑dependent autoimmunity. J. Immunol. 179, 4313–4317 (2007). 79. Liu, X., Lee, Y. S., Yu, C. R. & Egwuagu, C. E. Loss of STAT3 in CD4+ T cells prevents development of experimental autoimmune diseases. J. Immunol. 180, 6070–6076 (2008). 80. Yang, X., Ghoreschi, K., O’Shea, J. J. & Laurence, A. Opposing regulation of the locus encoding IL‑17 through direct, reciprocal actions of STAT3 and STAT5. Nature Immunol. 12, 247–254 (2011). 81. Akimzhanov, A. M., Yang, X. O. & Dong, C. Chromatin remodeling of interleukin‑17 (IL‑17)–IL‑17F cytokine gene locus during inflammatory helper T cell differentiation. J. Biol. Chem. 282, 5969–5972 (2007). 82. Ivanov, I. I. et al. The orphan nuclear receptor RORγt directs the differentiation program of proinflammatory IL‑17+ T helper cells. Cell 126, 1121–1133 (2006). 83. Yang, X. O. et al. T helper 17 lineage differentiation is programmed by orphan nuclear receptors RORα and RORγ. Immunity 28, 29–39 (2008). 84. Veldhoen, M. et al. The aryl hydrocarbon receptor links TH17‑cell‑mediated autoimmunity to environmental toxins. Nature 453, 106–109 (2008). 85. Brustle, A. et al. The development of inflammatory TH‑17 cells requires interferon-regulatory factor 4. Nature Immunol. 8, 958–966 (2007). 86. Bauquet, A. T. et al. The costimulatory molecule ICOS regulates the expression of c‑Maf and IL‑21 in the development of follicular T helper cells and TH‑17 cells. Nature Immunol. 10, 167–175 (2009). 87. Ghoreschi, K. et al. Generation of pathogenic TH17 cells in the absence of TGF-β signalling. Nature 467, 967–971 (2010). 88. McGeachy, M. J. et al. The interleukin 23 receptor is essential for the terminal differentiation of interleukin 17‑producing effector T helper cells in vivo. Nature Immunol. 10, 314–324 (2009). 89. Kwon, H. et al. Analysis of interleukin‑21‑induced Prdm1 gene regulation reveals functional cooperation of STAT3 and IRF4 transcription factors. Immunity 31, 941–952 (2009). 90. Takeda, K. et al. Stat3 activation is responsible for IL‑6‑dependent T cell proliferation through preventing apoptosis: generation and characterization of T cellspecific Stat3‑deficient mice. J. Immunol. 161, 4652–4660 (1998). 91. Bettelli, E. et al. Reciprocal developmental pathways for the generation of pathogenic effector TH17 and regulatory T cells. Nature 441, 235–238 (2006). 92. Chaudhry, A. et al. CD4+ regulatory T cells control TH17 responses in a Stat3‑dependent manner. Science 326, 986–991 (2009). 93. Liu, X., Robinson, G. W., Gouilleux, F., Groner, B. & Hennighausen, L. Cloning and expression of Stat5 and an additional homologue (Stat5b) involved in prolactin signal transduction in mouse mammary tissue. Proc. Natl Acad. Sci. USA 92, 8831–8835 (1995). 94. Grimley, P. M., Dong, F. & Rui, H. Stat5a and Stat5b: fraternal twins of signal transduction and transcriptional activation. Cytokine Growth Factor Rev. 10, 131–157 (1999). 95. Yao, Z. et al. Nonredundant roles for Stat5a/b in directly regulating Foxp3. Blood 109, 4368–4375 (2007). 96. Burchill, M. A., Yang, J., Vogtenhuber, C., Blazar, B. R. & Farrar, M. A. IL‑2 receptor β-dependent STAT5 activation is required for the development of Foxp3+ regulatory T cells. J. Immunol. 178, 280–290 (2007). 97. Nosaka, T. et al. Defective lymphoid development in mice lacking Jak3. Science 270, 800–802 (1995). 98. Thomis, D. C., Gurniak, C. B., Tivol, E., Sharpe, A. H. & Berg, L. J. Defects in B lymphocyte maturation and T lymphocyte activation in mice lacking Jak3. Science 270, 794–797 (1995). 99. Antov, A., Yang, L., Vig, M., Baltimore, D. & Van Parijs, L. Essential role for STAT5 signaling in CD25+CD4+ regulatory T cell homeostasis and the maintenance of self-tolerance. J. Immunol. 171, 3435–3441 (2003). 100. Zorn, E. et al. IL‑2 regulates FOXP3 expression in human CD4+CD25+ regulatory T cells through a STAT-dependent mechanism and induces the expansion of these cells in vivo. Blood 108, 1571–1579 (2006). 101. Takaki, H. et al. STAT6 inhibits TGF‑β1‑mediated Foxp3 induction through direct binding to the Foxp3 promoter, which is reverted by retinoic acid receptor. J. Biol. Chem. 283, 14955–14962 (2008).
VOLUME 11 | APRIL 2011 | 249 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS 102. O’Malley, J. T. et al. Signal transducer and activator of transcription 4 limits the development of adaptive regulatory T cells. Immunology 127, 587–595 (2009). 103. Laurence, A. et al. Interleukin‑2 signaling via STAT5 constrains T helper 17 cell generation. Immunity 26, 371–381 (2007). 104. Cohen, A. C. et al. Cutting edge: decreased accumulation and regulatory function of CD4+CD25high T cells in human STAT5b deficiency. J. Immunol. 177, 2770–2774 (2006). 105. Dupuis, S. et al. Impaired response to interferon-α/β and lethal viral disease in human STAT1 deficiency. Nature Genet. 33, 388–391 (2003). 106. Dupuis, S. et al. Impairment of mycobacterial but not viral immunity by a germline human STAT1 mutation. Science 293, 300–303 (2001). 107. Holland, S. M. et al. STAT3 mutations in the hyper-IgE syndrome. N. Engl. J. Med. 357, 1608–1619 (2007). 108. Minegishi, Y. et al. Dominant-negative mutations in the DNA-binding domain of STAT3 cause hyper-IgE syndrome. Nature 448, 1058–1062 (2007). References 107 and 108 were the first to describe the unexpected dominant negative phenotype of missense mutations of STAT3 in patients. 109. Milner, J. D. et al. Impaired TH17 cell differentiation in subjects with autosomal dominant hyper-IgE syndrome. Nature 452, 773–776 (2008). 110. van de Veerdonk, F. L. et al. Milder clinical hyperimmunoglobulin E syndrome phenotype is associated with partial interleukin‑17 deficiency. Clin. Exp. Immunol. 159, 57–64 (2010). 111. Ma, C. S. et al. Deficiency of Th17 cells in hyper IgE syndrome due to mutations in STAT3. J. Exp. Med. 205, 1551–1557 (2008). 112. Minegishi, Y. et al. Molecular explanation for the contradiction between systemic Th17 defect and localized bacterial infection in hyper-IgE syndrome. J. Exp. Med. 206, 1291–1301 (2009). 113. Barrett, J. C. et al. Genome-wide association defines more than 30 distinct susceptibility loci for Crohn’s disease. Nature Genet. 40, 955–962 (2008). 114. Danoy, P. et al. Association of variants at 1q32 and STAT3 with ankylosing spondylitis suggests genetic overlap with Crohn’s disease. PLoS Genet. 6, e1001195 (2010). 115. Duerr, R. H. et al. A genome-wide association study identifies IL23R as an inflammatory bowel disease gene. Science 314, 1461–1463 (2006). 116. Cargill, M. et al. A large-scale genetic association study confirms IL12B and leads to the identification of IL23R as psoriasis-risk genes. Am. J. Hum. Genet. 80, 273–290 (2007). 117. Burton, P. R. et al. Association scan of 14,500 nonsynonymous SNPs in four diseases identifies autoimmunity variants. Nature Genet. 39, 1329–1337 (2007).
118. Filer, C. et al. Investigation of association of the IL12B and IL23R genes with psoriatic arthritis. Arthritis Rheum. 58, 3705–3709 (2008). 119. Remmers, E. F. et al. Genome-wide association study identifies variants in the MHC class I, IL10, and IL23R–IL12RB2 regions associated with Behcet’s disease. Nature Genet. 42, 698–702 (2010). 120. Reveille, J. D. et al. Genome-wide association study of ankylosing spondylitis identifies non-MHC susceptibility loci. Nature Genet. 42, 123–127 (2010). 121. Remmers, E. F. et al. STAT4 and the risk of rheumatoid arthritis and systemic lupus erythematosus. N. Engl. J. Med. 357, 977–986 (2007). 122. Glas, J. et al. Evidence for STAT4 as a common autoimmune gene: rs7574865 is associated with colonic Crohn’s disease and early disease onset. PLoS ONE 5, e10373 (2010). 123. Korman, B. D. et al. Variant form of STAT4 is associated with primary Sjogren’s syndrome. Genes Immun. 9, 267–270 (2008). 124. Cho, S. S. et al. Activation of STAT4 by IL‑12 and IFN-α: evidence for the involvement of ligand-induced tyrosine and serine phosphorylation. J. Immunol. 157, 4781–4789 (1996). 125. Baechler, E. C. et al. Interferon-inducible gene expression signature in peripheral blood cells of patients with severe lupus. Proc. Natl Acad. Sci. USA 100, 2610–2615 (2003). 126. Hellquist, A. et al. Evidence for genetic association and interaction between the TYK2 and IRF5 genes in systemic lupus erythematosus. J. Rheumatol. 36, 1631–1638 (2009). 127. Hawkins, R. D., Hon, G. C. & Ren, B. Next-generation genomics: an integrative approach. Nature Rev. Genet. 11, 476–486 (2010). A review of the basic concepts and application of next-generation sequencing techniques by experts in the field. 128. Hon, G., Ren, B. & Wang, W. ChromaSig: a probabilistic approach to finding common chromatin signatures in the human genome. PLoS Comput. Biol. 4, e1000201 (2008). 129. Visel, A. et al. ChIP-seq accurately predicts tissuespecific activity of enhancers. Nature 457, 854–858 (2009). This study indicates that the location of tissue-specific distal gene enhancers can be accurately predicted by mapping genome-wide p300 binding. 130. Ghisletti, S. et al. Identification and characterization of enhancers controlling the inflammatory gene expression program in macrophages. Immunity 32, 317–328 (2010). 131. Kasowski, M. et al. Variation in transcription factor binding among humans. Science 328, 232–235 (2010). 132. Bonasio, R., Tu, S. & Reinberg, D. Molecular signals of epigenetic states. Science 330, 612–616 (2010).
250 | APRIL 2011 | VOLUME 11
133. Blat, Y. & Kleckner, N. Cohesins bind to preferential sites along yeast chromosome III, with differential regulation along arms versus the centric region. Cell 98, 249–259 (1999). 134. Ren, B. et al. Genome-wide location and function of DNA binding proteins. Science 290, 2306–2309 (2000). 135. Gupta, P. K. Single-molecule DNA sequencing technologies for future genomics research. Trends Biotechnol. 26, 602–611 (2008). 136. Park, P. J. ChIP-seq: advantages and challenges of a maturing technology. Nature Rev. Genet. 10, 669–680 (2009). 137. Schones, D. E. et al. Dynamic regulation of nucleosome positioning in the human genome. Cell 132, 887–898 (2008). 138. Crawford, G. E. et al. Identifying gene regulatory elements by genome-wide recovery of DNase hypersensitive sites. Proc. Natl Acad. Sci. USA 101, 992–997 (2004). 139. Sabo, P. J. et al. Genome-wide identification of DNaseI hypersensitive sites using active chromatin sequence libraries. Proc. Natl Acad. Sci. USA 101, 4537–4542 (2004). 140. Ji, H. et al. Comprehensive methylome map of lineage commitment from haematopoietic progenitors. Nature 467, 338–342 (2010). 141. Bartel, D. P. MicroRNAs: target recognition and regulatory functions. Cell 136, 215–233 (2009). 142. Kuchen, S. et al. Regulation of microRNA expression and abundance during lymphopoiesis. Immunity 32, 828–839 (2010). 143. Guttman, M. et al. Chromatin signature reveals over a thousand highly conserved large non-coding RNAs in mammals. Nature 458, 223–227 (2009). 144. Kim, T. K. et al. Widespread transcription at neuronal activity-regulated enhancers. Nature 465, 182–187 (2010). 145. Rozowsky, J. et al. PeakSeq enables systematic scoring of ChIP-seq experiments relative to controls. Nature Biotech. 27, 66–75 (2009). 146. Chen, X. et al. Integration of external signaling pathways with the core transcriptional network in embryonic stem cells. Cell 133, 1106–1117 (2008).
Competing interests statement
The authors declare no competing financial interests.
FURTHER INFORMATION John J. O’Shea’s homepage: http://www.niams.nih.gov/ Research/Ongoing_Research/Branch_Lab/Molecular_ Immunology_and_Inflammation/lcbs.asp Gene Expression Omnibus: http://www.ncbi.nlm.nih.gov/geo Ingenuity Pathways Analysis: http://www.ingenuity.com ALL LINKS ARE ACTIVE IN THE ONLINE PDF
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS
Recombination centres and the orchestration of V(D)J recombination David G. Schatz* and Yanhong Ji‡
Abstract | The initiation of V(D)J recombination by the recombination activating gene 1 (RAG1) and RAG2 proteins is carefully orchestrated to ensure that antigen receptor gene assembly occurs in the appropriate cell lineage and in the proper developmental order. Here we review recent advances in our understanding of how DNA binding and cleavage by the RAG proteins are regulated by the chromatin structure and architecture of antigen receptor genes. These advances suggest novel mechanisms for both the targeting and the mistargeting of V(D)J recombination, and have implications for how these events contribute to genome instability and lymphoid malignancy.
V(D)J recombination Somatic rearrangement of variable (V), diversity (D) and joining (J) regions of the genes that encode antigen receptors, leading to repertoire diversity of immunoglobulins and T cell receptors.
Chromosomal translocation An aberration of chromosome structure in which a portion of one chromosome is broken off and becomes attached to another.
*Department of Immunobiology and Howard Hughes Medical Institute, Yale University School of Medicine, 300 Cedar Street, Box 208011, New Haven, Connecticut 06520‑8011, USA. ‡ Department of Immunology & Microbiology, School of Medicine, Xi’an JiaoTong University, Xian Shaanxi, China 710049. Correspondence to D.G.S. e‑mail:
[email protected] doi:10.1038/nri2941 Published online 11 March 2011
Vertebrate adaptive immune responses are critically dependent on the expression of a diverse repertoire of antigen receptors — immunoglobulins and T cell receptors (TCRs) — by B and T cells. The genes that encode these antigen receptors are highly unusual in that they exist in a non-functional state in the germline, with the 5ʹ portion of each gene (which encodes the antigen binding domain) arranged as arrays of variable (V), diversity (D; only present in some loci) and joining (J) gene segments (FIG. 1). Assembly of these genes by V(D)J recombination generates antigen receptor diversity and is the central process around which early lymphocyte development is organized. Multiple V(D)J recombination events must occur during the genesis of each new lymphocyte, and each event requires the introduction of chromosomal DNA double strand breaks. Such lesions are among the most dangerous that can be inflicted on the genome and, hence, it is not surprising that elaborate mechanisms have evolved to regulate the generation of these DNA breaks and to ensure their efficient repair. Despite this, it has become increasingly clear that the occasional errors that occur during V(D)J recombination are an important source of genome instability (particularly chromosomal translocations ) and contribute to the development of lymphomas and leukaemias1–3. V(D)J recombination is initiated by the RAG recombinase (referred to hereafter as RAG) — a protein complex consisting primarily of the proteins encoded by recombination activating gene 1 (RAG1) and RAG2. RAG binds and cleaves the DNA at specific recombination signal sequences (RSSs) that flank each V, D and J gene segment. Thereafter, the DNA ends are processed and repaired by
a large group of DNA repair enzymes, many of which are components of the classical non-homologous end joining (NHEJ) repair pathway. The resulting recombination event deletes or inverts a segment of chromosomal DNA that ranges from as small as a few hundred base pairs to as large as several million base pairs. The initiation of V(D)J recombination is regulated at three distinct levels. First, the RAG proteins are expressed at high levels only during the early stages of lymphocyte development, thereby ensuring that the reaction does not occur in other tissues or cell types. Second, the ability of RAG to initiate V(D)J recombination is dictated by the ‘accessibility’ of RSSs within chromatin4. And third, V(D)J recombination is regulated by the position and three dimensional architecture of antigen receptor loci in the nucleus, with chromosome looping and condensation thought to have a vital role in allowing recombination between widely spaced gene segments. This Review attempts to integrate prior knowledge in these areas with recent advances in our understanding of how chromosome architecture and covalent histone modifications coordinate the early steps of V(D)J recombination. We discuss the implications of these advances for the physiological and pathophysiological targeting of V(D)J recombination and propose a model for the stepwise engagement of the components of the two RSSs during a V(D)J recombination event.
V(D)J recombination biochemistry DNA binding and cleavage by the RAG proteins. The RSSs that flank each gene segment consist of conserved heptamer and nonamer elements separated by a less well conserved spacer region of either 12 or 23 base
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 251 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS &* MD
8* /D
C+IJ
,* MD
'µ
%*
)GTONKPG 2&3
′44
4GEQODKPCVKQP EGPVTG
#UUGODNGF 8& ,
D+IM 8κ /D
&KUVCN
2TQZKOCN
,κ MD
K'κ
%κ
′'κ
E6ETD 8β MD
2&β &β
%β &β 2&β
,β MD
F6ETCs6ETF 8αQT8δ /D Non-homologous end joining (NHEJ). A DNA repair process that joins broken DNA ends (double strand breaks) without using homologous DNA as a template. Components of this pathway include the proteins Ku70 (also known as XRCC6), Ku80 (also known as XRCC5), Artemis, X‑ray repair crosscomplementing protein 4 (XRCC4), DNA ligase IV and the catalytic subunit of DNA-dependent protein kinase (DNA-PKcs).
Chromatin The combination of DNA, histones and other proteins that comprises eukaryotic chromosomes. The basic repeating unit of chromatin is the nucleosome, which consists of an octamer of histone proteins around which ~146 base pairs of DNA is wound.
Allelic exclusion In theory, every B cell has the potential to produce two immunoglobulin heavy chains and two immunoglobulin light chains. In practice, however, a B cell produces only one immunoglobulin heavy chain and the majority produce only one immunoglobulin light chain. Similarly, most T cells produce only a single T cell receptor β-chain protein. The process by which the production of two different chains is prevented is known as allelic exclusion. Allelic exclusion is accomplished primarily through regulated V(D)J recombination.
&δ
&δ
6'#
'δ
,δ ,δ
%δ 8δ
,β MD
%β 'β 8β
2,α
%α
'α
,α MD
Figure 1 | The structure of antigen receptor genes. Schematic diagrams of the four mouse antigen receptor loci for which the binding patterns of the recombination activating gene (RAG) proteins have been reported. Variable (V), diversity 0CVWTG4GXKGYU^+OOWPQNQI[ (D) and joining (J) gene segments are represented as yellow, red and blue rectangles, respectively, constant (C) regions are shown as grey rectangles, enhancer elements as green ovals and germline promoters associated with recombination centres as green diamonds with arrows. The promoters associated with each V gene segment and the recombination signal sequences (RSSs) are not shown. The approximate sizes of the regions within each locus are indicated, together with the approximate number of gene segments in each region (shown in parentheses). The recombination centres (blue shaded areas) are the regions that are bound by RAG1 and RAG2 when the loci are in their germline configurations85. a | The immunoglobulin heavy chain (Igh) locus in its germline and assembled configurations. b | The germline Igk locus. c | The germline T cell receptor β-chain (Tcrb) locus. d | The germline Tcra–Tcrd locus. The Tcra recombination centre is represented by two triangles to depict the 5ʹ to 3ʹ gradient of RAG binding observed downstream of the T early α (TEA) element and the increase observed downstream of the Jα49 germline promoter85.
pairs, which defines the 12RSS and 23RSS, respectively (BOX 1). There is a strong preference for recombination between a 12RSS and a 23RSS, a restriction known as the 12–23 rule5. RAG (BOX 2) recognizes RSSs and cleaves DNA during V(D)J recombination6. RAG1 is the principal DNA-binding component and contains most or all of the active site residues that catalyse DNA cleavage. RAG2 enhances RAG1–heptamer element interactions and is a vital cofactor for DNA cleavage6,7. RAG1 and RAG2 contain several important regulatory domains at their amino and carboxyl termini, respectively (BOX 2). DNA cleavage in vitro by RAG is strongly enhanced by high-mobility group protein B1 (HMGB1) or HMGB2 (which are ubiquitous, nonspecific DNA binding and bending proteins), but a role for HMGB proteins in V(D)J recombination in vivo has yet to be proven6. Hereafter, HMGB1 is used to refer to either protein. RSS recognition is thought to occur via a capture model (FIG. 2a) in which the full complement of RAG1, RAG2 and HMGB1 proteins binds to one RSS to form the signal complex and then captures a second RSS that lacks bound proteins, forming the synaptic or paired complex 8,9. DNA cleavage is a two-step process (FIG. 2b). RAG first introduces a single strand nick between the
heptamer and the gene segment. The 3ʹ hydroxyl group that is created then attacks the other strand to generate a DNA double strand break, a step referred to as hairpin formation because it generates a covalently sealed hairpin at the end of the gene segment. Hairpin formation only occurs in the paired complex and is thought to take place simultaneously at the two RSSs. Nicking is less tightly regulated and can occur in the signal or paired complex 6,10. Order during V(D)J recombination. Antigen receptor loci recombine in a well-defined order during lymphocyte development 5,11,12. The immunoglobulin heavy chain (Igh) and TCR β-chain (Tcrb) loci are the first to be assembled during the development of B cells and αβ T cells, respectively, with D‑to‑J recombination invariably preceding V‑to-DJ recombination. Following this, recombination of the immunoglobulin light chain locus (either Igk or Igl) or the TCR α-chain locus (Tcra) is initiated. Most mature B cells and αβ T cells express only one functional Igh or Tcrb allele, respectively, and this phenomenon of allelic exclusion is accomplished through tight control of the recombination process. Immunoglobulin loci recombine fully only in the B cell lineage and TCR loci recombine only in developing
252 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Box 1 | The recombination signal sequence The recombination signal sequence (RSS) lies immediately adjacent to each antigen receptor gene segment and contains two well-conserved DNA elements (the heptamer and the nonamer) separated by a spacer region (see the figure). The consensus sequences of the heptamer and nonamer are the most efficient in mediating V(D)J recombination, but most endogenous RSSs deviate from the consensus at the heptamer, the nonamer, or both. The length of the spacer is tightly conserved, varying at most by 1 base pair from either 12 base pairs or 23 base pairs. Although certain spacer residues show significant conservation, the spacer is much more variable in sequence than the heptamer or the nonamer. The nonamer is a sequence-specific binding site for the 5KVGQH&0# nonamer-binding domain of recombination activating gene 1 ENGCXCIG (RAG1) and functions to anchor the RAG proteins to the DNA. The #PVKIGPTGEGRVQT heptamer has at least two roles. It enhances binding of the RAG IGPGUGIOGPV 455 recombinase to the RSS and it specifies the site of DNA cleavage, probably owing to its ability to facilitate the formation of an altered DNA structure at the coding flank–heptamer border. The first three nucleotides of the heptamer (5ʹ-CAC) are particularly important in this regard and are essentially invariant in functional RSSs. The length of the spacer is crucial for efficient V(D)J recombination, suggesting that its primary role is to properly DR align the heptamer and nonamer. The sequence of the spacer can ′%#%#)6) #%#####%%′ also influence the efficiency of V(D)J recombination, although DR usually to only a small extent. Similarly to the heptamer and nonamer, the spacer is also involved in protein–DNA interactions in RAG–high-mobility group protein B (HMGB)–RSS complexes. *GRVCOGT 5RCEGT 0QPCOGT See REF. 6 for further details. 0CVWTG4GXKGYU^+OOWPQNQI[
Germline transcription Transcription of unrearranged antigen receptor gene loci that begins before or coincident with their activation. It is not thought to produce functional protein, and the promoter and initiation sites are often lost in the subsequent rearrangement events.
Nucleosome The fundamental structural unit of eukaryotic chromosomes. It consists of pairs of each of the core histones (H2A, H2B, H3 and H4), thereby creating the histone octamer, and a single molecule of the linker histone H1. The nucleosome spans ~146 base pairs of DNA.
Chromatin remodelling complex An enzymatic complex that remodels the DNA–nucleosome architecture and thus can determine transcriptional activity. The SWI–SNF ATPase is an example of a complex that remodels chromatin.
T cells. Such developmentally regulated and lineage-specific recombination occurs despite the use of fairly well conserved RSSs and the same recombination machinery for all of the loci. For the last 25 years, the dominant and highly successful model used to explain this has been that involving the accessibility of RSSs4,13. In the last 10 years, it has become clear that two additional levels of regulation — antigen receptor locus architecture and nuclear location — work together with locus accessibility to ensure order, fidelity and allelic exclusion during V(D)J recombination.
Accessibility and transcription Chromatin and the inaccessible state. The accessibility model emerged from the observation that Igh variable (VH) gene segments undergo germline transcription (also known as sterile transcription) coincident with their recombination4. The model was subsequently reinforced by a wealth of observations demonstrating that V(D)J recombination correlates with numerous markers of open chromatin, including germline transcription, activating histone modifications (such as histone acetylation), nuclease accessibility and DNA hypomethylation11,12,14–17. Together, these correlations indicated a tight connection between chromatin structure and the targeting of V(D)J recombination. This notion received strong support from experiments showing that when isolated lymphocyte nuclei were incubated with RAGcontaining nuclear extracts, cleavage could be detected at antigen receptor gene RSSs in a pattern that recap itulated the lineage and developmental specificity of V(D)J recombination18. These results demonstrated that chromatin can be permissive or repressive for V(D)J recombination and, importantly, that chrom atin structure affects the first (DNA cleavage) phase of the reaction.
One simple mechanism to explain the repressive effects of chromatin was revealed when it was discovered that wrapping an RSS around a nucleosome renders it refractory to binding or cleavage by the RAG proteins19–22. Proteolytic removal of N‑terminal histone tails20,22 or the action of ATP-dependent chromatin remodelling complexes such as the SWI–SNF ATPase complex 22–25 allows substantial cleavage of nucleosomal RSS substrates by RAG. By contrast, histone acetylation by itself has no20,21 or only a modest 22 ability to stimulate cleavage of nucleosomal RSS substrates, although it can enhance the effects of chromatin remodelling complexes in such assays22,24. Together, these biochemical studies strongly support the idea that the nucleosome is intrinsically inhibitory to DNA cleavage by RAG. To what extent are RSSs positioned on nucleosomes in the nucleus of developing lymphocytes? An initial study found that the RSS, particularly the nonamer element, serves as a nucleosome-positioning sequence in vitro and on plasmid substrates in vivo26. Therefore, RSSs should exhibit a high frequency of nucleosome occupancy in antigen receptor loci in the absence of mechanisms to reposition nucleosomes. However, a recent analysis of nucleosome occupancy in endogenous antigen receptor loci found no clear correlation between the location of RSSs and the position of nucleosomes27. Inaccessible chromatin was characterized by a high nucleosome density but even in this circumstance, some RSSs were at internucleosomal positions27. Thus, at least for the portions of the Tcra and Tcrb loci examined in this study, RSSs do not function as dominant nucleosome-positioning sequences. This is consistent with the emerging view that although the primary DNA sequence directly influences nucleosome positioning, many other factors (such as remodelling complexes, bound proteins, promoter-loaded RNA polymerase II
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 253 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Box 2 | The RAG proteins Recombination activating gene 1 (RAG1) and RAG2 are nuclear proteins that are well-conserved in jawed vertebrates. They interact with one another, and thousands of copies are estimated to be present in each developing lymphocyte104. ‘Core’ regions of each RAG protein have been defined as the minimal portion of the protein required for V(D)J recombination activity. The RAG1 core contains a well-defined nonamer-binding domain and a central region (amino acids 528–760) that is responsible for interactions with the heptamer and RAG2 (see the figure). Zinc finger region B (ZnB) is thought to be important for RAG2 interactions. The RAG1 core also contains the active site for DNA cleavage, which includes three acidic amino acids (D600, D708 and E962) that coordinate divalent metal ions and are essential for catalysis. The RAG1 amino‑terminal region (amino acids 1–383) enhances V(D)J recombination activity but the mechanism by which it does so is not well understood. This region contains several pairs of conserved cysteine residues (labelled C2) and ZnA, which homodimerizes, has E3 ubiquitin ligase activity and interacts with and ubiquitylates histone H3 (REF. 88). The RAG2 core is crucial for DNA cleavage activity, interacts with RAG1 and enhances the DNA-binding affinity and specificity of the RAG complex, although RAG2 has little or no DNA binding activity on its own. The RAG2 carboxy‑terminal region (amino acids 384–527) contains a plant homeodomain (PHD) finger that binds specifically to trimethylated histone H3 lysine 4 (H3K4me3), enhances the catalytic activity of the RAG complex and guides RAG2 to regions of active chromatin. Although RAG1 is expressed throughout the cell cycle, RAG2 is only stable in G0 or G1 phase cells owing to the phosphorylation of T490 in S, G2 and M phase cells105,106. RAG2 also contains an acidic region that interacts with histones79. Residue numbers refer to the mouse RAG proteins. See REFS 6,107,108 for further details. 0QPCOGTDKPFKPI &0#ENGCXCIG FQOCKP & &
4#) 0*
% %
r'WDKSWKVKPNKICUG r*KPVGTCEVKQP
4#) 0*
%QTGTGIKQP
2*& #EKFKETGIKQP
r%TKVKECNHQT&0#ENGCXCIG r'PJCPEGU&0#DKPFKPI
%11*
r*GRVCOGTDKPFKPI r4#)KPVGTCEVKQP
6
%QTGTGIKQP
'
%11* %GNNE[ENGTGIWNCVKQP
r$KPFU*-OG r'PJCPEGUECVCN[VKECEVKXKV[QH4#)
0CVWTG4GXKGYU^+OOWPQNQI[
Antisense transcription Transcription in the opposite direction and of the opposite strand from that used to generate the normal product of a gene. It is not thought to generate a protein product but instead might alter chromatin structure either directly (via the act of transcription) or indirectly (via the antisense RNA produced).
Heterochromatin High-density regions in the nucleus that are thought to contain compacted chromatin structures associated with silent genes.
and DNA methylation) also have important roles28–30. It is also interesting that the sequence AAAAA, which is found in many endogenous nonamers, is poorly incorporated into nucleosomes, which probably reflects intrinsic properties of homopolymeric deoxyadenosine (AAAAA) tracts28,31. Taken together, existing data suggest that inaccessibility to the V(D)J recombination machinery is mediated by both direct positioning of RSSs on nucleosomes and steric constraints created by nucleosome packing 21,27. Transcription and the generation of accessibility. The tight correlation that exists between germline transcription and V(D)J recombination suggests that transcriptional control elements (such as enhancers and promoters), transcription factors and perhaps transcription itself have important roles in rendering RSSs accessible for recombination11,12,14,32–35. The deletion of enhancer elements from an antigen receptor locus typically results in a near-total block in recombination of the affected
locus, accompanied by a loss of germline transcription and other markers of locus accessibility. By contrast, deletion of promoter elements has a more local effect, reducing recombination specifically in the region downstream of the promoter. For many years it was not known whether germline transcription was causally related to the generation of accessibility during V(D)J recombination or whether it was merely a side effect of other processes that were instead responsible for accessibility. This issue has now been conclusively resolved, at least for a portion of the Tcra locus, through elegant experiments involving the insertion of a transcription terminator into two different locations in the Jα cluster. The result in each case was a block in transcription and recombination of the Jα gene segments downstream of the terminator36,37. Hence, transcriptional elongation by RNA polymerase II has a direct and crucial role in creating the accessible chromatin that is required for Tcra recombination. Although this is likely to be the case for other antigen receptor loci as well, it is not yet clear if it is a universal rule. Several examples of a dissociation between V(D)J recombination and germline transcription have been reported12,16,38,39, and recent biochemical and in vivo experiments point to a central role for chromatin remodelling complexes in creating accessibility (BOX 3). The strong connection between transcription and V(D)J recombination extends also to antisense transcription. Developmentally regulated antisense transcription of VH gene segments40 and the DH–JH region is associated with Igh locus assembly 41–43. Antisense transcription has been suggested to have several different functions in V(D)J recombination. First, it might increase local RSS accessibility by mechanisms similar to that of germline (sense) transcription40,41. The strongest support for this comes from a recent study 44 in which the authors created a germline deletion of the interval between the VH and DH gene segments. This led to the production of long antisense transcripts running from DH into the 3ʹ part of the VH cluster, and also resulted in VH‑to-DHJH recombination in the thymus, where it is normally not detected. Second, based on the findings that the central DH gene segments are contained in a repeat unit, are poorly recombined, are transcribed bidirectionally and are associated with repressive histone marks, it has been proposed that antisense transcription contributes to the generation of heterochromatin and the suppression of V(D)J recombination by a double stranded RNA (dsRNA)dependent repeat-induced gene silencing mechanism42. Third, it has been hypothesized that the generation of non-coding RNAs (sense and/or antisense) or the noncoding RNAs themselves influence the three dimensional structure of local chromatin domains or the higher order configuration of the chromatin at antigen receptor loci34. Steps in local accessibility. These considerations and others led to the following model for how RSSs are freed from the repressive effects of chromatin and the nucleosome to allow RAG binding 11,14–17,32,34,45. The first step — initial locus opening — might be independent of well-characterized transcriptional control elements.
254 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS DNase I hypersensitive site A site of nuclease sensitivity when nuclei from cells are exposed to limiting concentrations of the enzyme DNase I. The digested regions of DNA correspond to sites of open DNA, which might be transcription factor binding sites or areas of altered nucleosome conformation.
Bromodomain A module of ~110 amino acids that is found in several transcriptional regulators. A bromodomain consists of a four-helix bundle with a single binding pocket for Nε-acetyl-lysine on histone tails.
In the Igh locus, loss of the repressive histone mark dimethylated histone H3 lysine 9 (H3K9) and an increase in the activating mark dimethylated H3K4 occur in the absence of the core Igh intronic enhancer (Eμ) and other known DNase I hypersensitive sites46. Although the mechanism for this is unknown, it is attractive to think that such changes allow transcription factors to gain access to Eμ46. Methylation of H3K9 is sufficient to suppress the recombination of an otherwise accessible minilocus47, and the presence of this marker correlates with inactive antigen receptor loci or gene segments48,49 and can trigger the formation of heterochromatin50. Hence, reducing H3K9 methylation is likely to be an important early step in creating accessibility. In the second step of the model, transcription factors bind to the enhancer and recruit histone acetyltransferases (HATs) that acetylate the N‑terminal tails of histones H3 and H4. This provides binding sites for the bromodomains in chromatin remodelling complexes, which in turn free up promoters for binding by RNA polymerase II, thereby allowing the initiation of germline transcription. Many factors (such as HATs, histone methylases, chromatin remodelling complexes and histone chaperones) associate and move with RNA polymerase II through the transcription unit, leading to further nucleosome remodelling and additional activating histone modifications51,52. One of the
most important of these modifications is H3K4 trimethylation (H3K4me3), which is strongly biased towards the 5ʹ end of the transcription unit and the region surrounding the promoter 53. In the fully activated antigen receptor locus, three principal mechanisms probably work together to free RSSs from repressive associations with nucleosomes: one, histone acetylation, which weakens the association of histone tails with DNA, loosens the chromatin fibre54 and enhances chromatin remodelling 22,24; two, chromatin remodelling, which could generate nucleosome-free RSSs by repositioning or evicting nucleosomes, or could act more subtly by transiently lifting a loop of DNA off of the nucleosome surface55; and three, transcriptional elongation by RNA polymerase II through RSSs, resulting in the disassembly of nucleosomes or their disassociation from DNA51. The relative importance of each of these mechanisms probably varies between different RSSs and different loci. In addition to these mechanisms, there are other, RSS-specific, pathways for the delivery of the RAG proteins. For example, the transcription factor FOS (a component of the AP1 dimer) has been implicated in the recruitment of RAG to the 3ʹ Dβ1 23RSS via an embedded AP1 binding site56, and the B cell-associated transcription factor PAX5 (paired box protein 5) has been suggested to facilitate RAG recruitment to VH gene segments57.
C2JCUG
D
)GPGUGIOGPV
455
455
4#) &0# UVTCPF
$KPFKPI CPFPKEMKPI 4#)4#) CPF*/)$
UKIPCNEQORNGZ
UKIPCNEQORNGZ
5[PCRUKU CPFPKEMKPI
2CKTGF EQORNGZ
%NGCXCIG
JCKTRKPHQTOCVKQP
2JCUG
%QFKPILQKPV
455
5KPINGUVTCPF PKEM 14
*CKTRKPQRGPKPI EQFKPIGPFRTQEGUUKPI
4#)
*1
0*',HCEVQTU CPF6F6
5KIPCNLQKPV
$NWPV UKIPCNGPF
1*
5GEQPFUVTCPF ENGCXCIG
*CKTRKP UGCNGF EQFKPIGPF
Figure 2 | Mechanism of V(D)J recombination. a | Steps in V(D)J recombination. Antigen receptor gene segments are flanked by a 12 recombination signal sequence (12RSS) or by a 23RSS. In the first phase, recombination activating gene 1 (RAG1) and RAG2 proteins, perhaps together with high-mobility group protein B1 (HMGB1), bind to the 12RSS or the 23RSS, forming the 12 signal complex or the 23 signal complex, respectively. Capture of the second RSS (a process termed synapsis) results in the formation of the paired complex, within which the RAG proteins introduce double strand breaks between the gene segments and the RSSs. In the second phase, the RAG proteins cooperate with non-homologous end joining (NHEJ) DNA repair factors to rejoin the DNA ends. Gene segment ends typically undergo non-templated nucleotide addition (light blue rectangle) by terminal deoxynucleotidyl transferase (TdT) and nucleotide loss before being joined to form the coding joint. RSS ends are typically joined without processing to form the signal joint. b | Steps in DNA cleavage by RAG. The RAG recombinase first introduces a nick on one DNA strand adjacent to the RSS by catalysing the hydrolysis of the phosphodiester bond. The 3ʹ hydroxyl (OH) group liberated by nicking then attacks the opposite DNA strand, resulting in a DNA double strand break and generating a hairpin-sealed coding end and a blunt signal end. 0CVWTG4GXKGYU^+OOWPQNQI[
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 255 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Box 3 | Creating accessibility: transcription versus chromatin remodelling In biochemical experiments using chromatinized substrates containing a 12 recombination signal sequence (12RSS) and a 23RSS, cleavage by the recombination activating gene (RAG) recombinase was observed only when a transcriptional activation domain was tethered in close proximity to both RSSs. The activation domain enhanced cleavage by recruiting the SWI–SNF chromatin remodelling complex, and transcription was neither required nor stimulatory for cleavage25. Therefore, in this well-defined system, chromatin remodelling was sufficient to confer accessibility in the absence of transcription. In vivo studies of the T cell receptor β-chain (Tcrb) locus also highlight the importance of chromatin remodelling. Recombination of the Tcrb locus depends on the action of multiple transcriptional control elements (FIG. 1c), including the Tcrb enhancer (Eβ), which is required for recombination and germline transcription of both Dβ–Jβ clusters38,109–113; PDβ1, the promoter associated with Dβ1, which is required for recombination and germline transcription of the Dβ1–Jβ1 cluster but not the Dβ2–Jβ2 cluster114,115; and PDβ2, a complex promoter that is a strong candidate for controlling recombination of the Dβ2–Jβ2 cluster116,117. Recombination of the Dβ1–Jβ1 cluster is associated with DNA looping and the formation of a holocomplex containing Eβ, PDβ1, transcription factors and components of the transcriptional apparatus112. The SWI–SNF chromatin remodelling complex is recruited to the Dβ1–Jβ1 region in a PDβ1- and Eβ-dependent manner and is required for efficient Dβ1‑to‑Jβ1 recombination118. Importantly, tethering of the chromatin remodelling factor BRG1 to a site immediately adjacent to Dβ1 compensated for deletion of PDβ1 in a recombination minilocus. However, not only were Dβ1‑to‑Jβ1 recombination and Dβ1 accessibility restored by BRG1 tethering, but germline transcription of the Dβ1–Jβ1 region was also restored118. Hence, although it is clear that recruitment of the SWI–SNF complex can substitute for the PDβ1 promoter, it remains unclear if chromatin remodelling was sufficient to create accessibility in the absence of transcription in this instance. A strong association between recruitment of the SWI–SNF complex, antisense transcription and recombination has also been observed at the immunoglobulin heavy chain locus (Igh)43.
Pro‑B cell A cell in the earliest stage of B cell development in the bone marrow. Pro-B cells are characterized by incomplete immunoglobulin heavy chain rearrangements and are defined as CD19+ and cytoplasmic IgM– or, sometimes, as B220+CD43+ (by the Hardy classification scheme).
Pericentric heterochromatin Regions of very densely packed chromatin fibres located near the centromere of each chromosome. These regions are typically inactive and often cluster to form discrete clumps in the nucleus.
Fluorescence in situ hybridization (FISH). The use of fluorescent probes to visually label specific DNA sequences in the nuclei of cells that are in the interphase or metaphase stages of mitosis.
Beyond accessibility Accessibility, although probably required for the recombination of any particular RSS, is not always sufficient. For example, in PAX5‑deficient pro‑B cells, 5ʹ VH gene segments fail to recombine despite exhibiting wild-type levels of germline transcription and histone acetylation58. Instances of accessible Tcrb variable (Vβ) gene segments that fail to undergo recombination have also been reported59. Current evidence suggests that at least two conditions, in addition to that of accessibility, must often be met before V(D)J recombination can occur. First, the recombining locus needs to move away from repressive chromatin compartments (such as the periphery of the nucleus and pericentric heterochromatin); and second, for long-distance recombination events, the locus must undergo a large-scale structural alteration (for example, compaction or looping) to allow widely separated gene segments to encounter one another with reasonable efficiency 34.
has been associated with interaction of the second allele with an Igk allele67 and relocation to pericentric heterochromatin68. The Tcra and Tcrd loci, which are not subject to allelic exclusion, do not undergo a developmentally regulated association with heterochromatin63. However, the connection between loss of accessibility, suppression of V(D)J recombination and association with a repressive chromatin compartment is neither direct nor simple. Germline transcription of the Vβ region has been demonstrated to occur on both alleles in pro‑T cells69, even though many or most of the alleles are associated with a repressive compartment63,66. And Igk undergoes biallelic germline transcription in pre‑B cells70,71 despite the fact that one allele is associated with heterochromatin. Hence, the mechanism by which locus association with pericentric heterochromatin might inhibit V(D)J recombination remains unknown and the inhibitory nature of the interaction remains to be proven.
Location in the nucleus. Both the periphery of the nucleus and regions of pericentric heterochromatin have been implicated in gene silencing 60 — perhaps most directly by the finding that forced recruitment of an expression cassette to the inner nuclear membrane results in transcriptional repression 61. Considerable evidence now indicates that antigen receptor loci move away from these nuclear compartments when they undergo recombination and associate with them following the termination of recombination 16,34,62. Regulated63–65 or stochastic66 associations of antigen receptor loci with repressive nuclear compartments have been implicated in the initiation and maintenance of allelic exclusion. After the successful recombination of one allele, recombination of the other must often be suppressed. In the case of the Igh locus, this
Antigen receptor locus architecture. Fluorescence in situ hybridization (FISH) studies have demonstrated a striking correlation between V(D)J recombination and changes in the large scale chromatin structure of antigen receptor loci. The Igh, Igk, Tcrb and Tcra loci all adopt a contracted configuration coincident with their recombination and revert to a decondensed state after their recombination is complete62,63,68,72. Contraction is at least in part the result of chromatin looping that brings distal V gene segments into close proximity with the (D)J segment 63,72,73, and this should facilitate long range recombination events involving distal V gene segments. The strongest evidence for this comes from the analysis of PAX5‑deficient pro‑B cells, in which Igh locus contraction and distal VH gene segment recombination both fail to occur; expression of PAX5 in these cells overcomes both defects74.
256 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS
Chromatin immunoprecipitation An experimental technique that analyses direct binding of an endogenous transcription factor to chromatin by fixation with formaldehyde followed by immunoprecipitation with a transcription factor-specific antibody. Gene-specific enrichment is then assessed by polymerase chain reaction analysis of the immunoprecipitated DNA.
Recombination centre A region of an antigen receptor locus that is characterized by strong binding of recombination activating gene 1 (RAG1) and RAG2 and high levels of germline transcription, RNA polymerase II, histone acetylation and trimethylated histone H3 lysine 4 (H3K4me3).
A major advance in our understanding of the structure and dynamics of antigen receptor loci comes from a recent high resolution (<50 nm) FISH analysis of the Igh locus75. Using data derived from 12 markers spanning the 3 Mb Igh locus, the authors could calculate an average trajectory for the Igh locus in haematopoietic cells before commitment to the B cell lineage (prepro‑B cells) and in pro‑B cells (in which Igh recombination occurs). In pre-pro‑B cells, the Igh locus is relatively decondensed, with distal VH gene segments in a distinct domain that is well separated from the DH–JH region. The VH portion of the locus could best be modelled as a series of 1 Mb DNA rosettes made up of multiple loops of ~120 kb, with the rosettes separated by linkers of ~60 kb. It seems likely that the rosettes are dynamic structures in which loops of variable sizes undergo frequent folding and unfolding 34,75. In pro‑B cells, the locus contracts into a single condensed domain, in which the more distal VH gene segments have moved substantially so that all the VH gene segments are approximately equidistant from the DH–JH region. Interestingly, despite its greater overall condensation, the Igh locus in pro‑B cells exhibits greater conformational flexibility than in pre-pro‑B cells 75. Recombinationally active loci are therefore proposed to adopt a condensed but dynamic chromatin architecture that ensures that both proximal and distal V gene segments have a roughly equal chance of encountering and recombining with the (D)J segment 34,75. The idea that regulation of locus architecture has a key role in controlling V gene segment usage received further support from a recent study of the Tcra–Tcrd locus76. In pro‑T cells, in which all of the V gene segments must have the chance to undergo recombination during Tcrd locus assembly, both the 3ʹ and 5ʹ halves of the Vδ region are contracted and presumably in close proximity to the Dδ–Jδ region. However, in pre‑T cells, in which initial Tcra rearrangements preferentially involve 3ʹ Vα gene segments, the 3ʹ portion of the locus is contracted but the 5ʹ portion is not 76. Recombination of 5ʹ Vα gene segments would therefore be unlikely to occur until after an initial 3ʹ Vα-to‑Jα recombination event has deleted a portion of the locus and brought the 5ʹ Vα segments closer to the Jα region. In summary, large scale movement and structural reorganization of antigen receptor loci provide important layers of regulation, in addition to that of RSS accessibility, for the proper orchestration of V(D)J recombination.
Targeting of RAG proteins How do the RAG proteins find their way to accessible RSSs? Is it a passive process that relies on their sequencespecific DNA binding properties, or is it facilitated by other interactions? And where in the large antigen receptor loci do the RAG proteins bind to initiate V(D)J recombination? Do they bind uniformly to the many available RSSs, or is binding restricted to a particular portion of each locus? Recent advances have begun to answer these fundamental questions.
RAG2: reader of the histone code. The first major advance in this area was triggered by the observations that the C‑terminal domain of RAG2 contains a non-canonical plant homeodomain (PHD) finger 77,78 and that this region of RAG2 can interact with histones79. Two groups then made the seminal discovery that the RAG2 PHD finger specifically recognizes H3K4me3 and that this interaction is important for efficient V(D)J recombination 80,81. RAG2 mutations that abrogate this interaction substantially reduce the efficiency of V(D)J recombination80,81, as does a reduction in H3K4me3 levels80. Furthermore, mutation of a single critical residue within the RAG2 PHD finger (tryptophan 453) is associated with Omenn syndrome, which is a rare condition characterized by severe immunodeficiency that is strongly linked to inefficient V(D)J recombination82. Because H3K4me3 is highly enriched near the 5ʹ end of transcription units, the PHD finger of RAG2 should facilitate recruitment of RAG to RSSs located close to active promoters. Interestingly, recognition of H3K4me3 by RAG2 does more than simply enhance RAG binding; recent biochemical experiments indicate that it also enhances the catalytic activity of RAG83,84. Focal RAG binding in recombination centres. It is surprising that V(D)J recombination has been studied intensively for many years without an understanding of where the RAG proteins bind within antigen receptor loci. The recent development of a RAG chromatin immunoprecipitation (ChIP) assay has allowed for substantial insights into this problem85. In the Igh, Igk, Tcrb and Tcra loci, both RAG1 and RAG2 were found to bind specifically to a small region containing J gene segments (and the D gene segments that lie very close to the J region in Igh and Tcrb) (FIG. 1). In no case was binding detected at V or more distal D gene segments. These focal regions of RAG binding, referred to as recombination centres , display the features one would expect for accessible RSSs — namely, substantial germline transcription and high levels of histone H3 acetylation, H3K4me3 and RNA polymerase II occupancy. The dynamics of RAG binding closely mirror the developmental stage- and lineage-specific pattern of V(D)J recombination events, with two interesting exceptions. The binding of RAG to Igh or Tcrb is robust in pre‑B and pre‑T cells, respectively, cells in which V‑to-DJ recombination is suppressed and one allele is associated with pericentric heterochromatin63,68. Hence, allelic exclusion of Igh in pre‑B cells or Tcrb in pre‑T cells does not involve the exclusion of RAG from these loci, and association with pericentric heterochromatin might be compatible with RAG binding. It remains unclear how recombination of VH or Vβ gene segments that are proximal to a recombination centre is prevented in pre‑B and pre‑T cells, respectively. Interestingly, each RAG protein can recapitulate its highly specific pattern of binding in the absence of the other, with one exception noted below 85. How might this occur? For RAG1, specific recruitment
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 257 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS into recombination centres in the absence of RAG2 is probably mediated, at least in part, by direct RSS recognition. This suggestion is supported by experiments demonstrating that both the RSS and the nonamerbinding domain of RAG1 are required for the binding of RAG1 to a chromosomal recombination substrate85. Furthermore, RAG1 by itself has been shown to bind relatively tightly and with a slow dissociation kinetics to the 12RSS in vitro86. Hence, it appears that RAG1 has substantial RSS-specific binding activity in vivo in the absence of RAG2. For RAG2, the specific pattern of binding observed in the absence of RAG1 is probably due to the recognition of H3K4me3 by the RAG2 PHD finger. In support of this, a genome wide ChIP analysis revealed that RAG2 is recruited to thousands of active regions of the genome, in a pattern essentially identical to that of H3K4me3 (REF. 85). Intriguingly, several promoter regions in genes that do not encode antigen receptors showed substantial RAG2 binding but were found to lack detectable RAG1 binding. This indicates that developing lymphocytes contain a pool of RAG2 that is not associated with RAG1, although how widespread this phenomenon is remains to be determined. The tight link between accessibility and transcription leads to the strong prediction that binding of RAG depends on transcription and on transcriptional control elements. Indeed, the deletion of enhancer elements results in a global loss of RAG1 binding at the Tcrb and Tcra loci, and promoter deletion causes a local defect in RAG1 binding in the region downstream of the promoter 87. Importantly, blocking transcriptional elongation in the Jα cluster results in defective RAG1 binding immediately downstream of the blockade. Hence, promoters, enhancers and transcription elongation have a crucial role in creating the environment needed for RAG binding 87. The recombination centre model. These findings have led to the proposal that the complex series of events that occur during V(D)J recombination are orchestrated within recombination centres85. The recombination centre model (FIG. 3) begins with the creation of a focal region of highly accessible RSSs characterized by elevated levels of germline transcription, histone acetylation and H3K4me3. These conditions are ideal for the recruitment of RAG1 and RAG2, which probably engage both H3K4me3 and the RSSs within the recombination centre. This is followed by capture of a partner RSS to form the paired complex, and this in turn leads to DNA cleavage and recombination. When the partner is a distal V gene segment, architectural reorganization (such as looping or contraction) of the locus is probably required to bring the V gene segment into the proximity of the recombination centre. Our current simplistic conception is that V gene segments form a ‘cloud’ around the recombination centre and that the dynamic formation, movement and breakdown of chromatin loops containing the V gene segments gives each gene segment a reasonable chance of encountering the recombination centre.
Unanswered questions and future directions Pathways of RAG binding. The finding that each RAG protein can independently find its way into recombination centres suggests that there are at least three different pathways of RAG recruitment: one, binding of RAG1 to the RSS followed by recruitment of RAG2 to the bound RAG1; two, binding of RAG2 to H3K4me3 followed by recruitment of RAG1 to the bound RAG2; and three, recruitment of a preformed RAG1–RAG2 complex via interactions with the RSS and H3K4me3. The first pathway might be important in cycling cells. In the S, G2 and M phases of the cell cycle, RAG2 levels are very low and RAG1 would be predicted to bind accessible RSSs within recombination centres. Then, as cells re-enter the G1 phase and RAG2 levels rise, RAG2 could be recruited to the pre-bound RAG1. The second pathway might occur at the large number of sites outside of antigen receptor loci that lack a strong binding site for RAG1 but contain substantial levels of H3K4me3. The third pathway might dominate in cells expressing substantial levels of both RAG1 and RAG2, as the two proteins interact with one another and the resulting complex binds to the RSS with greater specificity and affinity than RAG1 alone6. It is appealing to think that the RAG1– RAG2 complex, with its ability to interact with both the RSS and H3K4me3, would be recruited more rapidly and efficiently into recombination centres than either RAG protein alone. The Igh locus appears to provide an example of this, as binding of RAG1 to the Igh recombination centre is markedly reduced in the absence of RAG2 (REF. 85). It is not clear why this was observed at Igh but not at the other antigen receptor loci examined85. It will be important to determine the extent to which the two RAG proteins exist separately from one another in vivo and how dynamic their association is. It will also be important to determine the extent to which the association of RAG1 with chromatin is influenced by its recently reported ability to interact with and ubiquitylate histone H3 (REF. 88). This information, together with additional genome-wide binding studies, should shed considerable light on the relative importance of the three pathways of RAG recruitment. Properties of recombination centres. Many of the basic features of recombination centres remain to be characterized. All of the recombination centres defined thus far contain multiple J gene segments (and associated RSSs) within a 2–10 kb region, but it is not known whether the relatively close spacing of the RSSs or even the presence of more than one RSS in the recombination centre contributes to RAG recruitment. It also remains unclear how many RAG molecules are present within a recombination centre, how many RSSs within a recombination centre are engaged by RAG at any one time in any particular cell and how dynamic the RAG–RSS interaction is. Regulatory potential of recombination centres. A particularly attractive feature of recombination centres is their potential to regulate both the DNA cleavage and the repair phases of V(D)J recombination. Because initial binding of RAG appears to be restricted to the J and proximal D
258 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Cryptic RSS A region of DNA that resembles a true recombination signal sequence (RSS) in some of its functionally important sequence features but does not lie adjacent to an antigen receptor gene segment.
6ETCNQEWU
gene segments of antigen receptor loci, V and distal D gene segments have a reduced ability to engage in undesirable V(D)J recombination events with one another, with RSSs in another antigen receptor locus, or with cryptic RSSs or alternative DNA structures89 in genes that do not encode antigen receptors. This initial binding preference should also limit the incidence of single strand nicking by RAG at V and distal D RSSs. RAG-generated nicks have been demonstrated to stimulate homologous recombination and thus could contribute to genome instability 90. Simultaneous engagement of multiple RSSs in a single recombination centre, if it occurs, would raise the possibility that more than one recombination event occurs on
455
8α
6'#
455
an allele at one time. This is neither necessary for gene assembly nor desirable from the standpoint of genome stability. So, do deterministic mechanisms exist to prevent this, or is it made unlikely by virtue of inefficient capture of a partner RSS? Similarly, how do developing lymphocytes avoid initiating V(D)J recombination on both alleles of the Igh, Igk or Tcrb loci simultaneously and thus prevent the risk of a violation of allelic exclusion? Interestingly, it was recently observed that both the Igh and Igk loci exhibit a phenomenon of developmentally regulated pairing of homologous chromosomes, and that this is reduced in the absence of RAG1 (REF. 64). Other data has suggested that homologue pairing and
,α
'α
%QQRGTCVKQPDGVYGGP6'#CPF'α
RGTJCRUD[NQQRKPI )GTONKPGVTCPUETKRVKQP
*-OG
40#RQN[OGTCUG++
4GEQODKPCVKQPEGPVTG
4#) 4#)
4#) 4#)
4#)
4#)
4#) 4#)
4#)
#EGV[NCVKQP *KUVQPGVCKN
4#)
4#)
Figure 3 | The recombination centre model. In the germline, prior to the initiation of recombination, the mouse T cell receptor α-chain (Tcra) locus is transcriptionally silent. Then, the T early α (TEA) promoter and the enhancer Eα cooperate (perhaps through a physical interaction mediated by looping of the DNA) to initiate germline transcription from TEA. This creates a domain marked by high levels of RNA polymerase II, histone acetylation and trimethylated histone H3 lysine 4 (H3K4me3). Thereafter, recruitment of the recombination activating gene (RAG) proteins occurs. Several mechanisms for this recruitment have been proposed (not shown): RAG1 recruitment to a recombination signal sequence (RSS) followed by RAG2; RAG2 recruitment to H3K4me3 followed by RAG1, or recruitment of the RAG1–RAG2 complex to an RSS or to H3K4me3. Large scale chromatin reorganization brings variable (V) gene segments into close proximity with the recombination centre. One V gene segment is stably captured in the paired complex with a RAG-bound joining (J) gene segment. The RAG proteins introduce double strand breaks adjacent to the V and J gene segments, completing the first phase of V(D)J recombination. RAG1–RAG2 complexes bound to RSSs probably also engage H3K4me3, which stimulates their cleavage activity83,84.
5VCDNGECRVWTGQHQPG8αIGPGUGIOGPV
4#)
.CTIGUECNGEJTQOCVKPTGQTICPK\CVKQP
4#) 4#)
4#) 4#)
&QWDNGUVTCPFDTGCMU
0CVWTG4GXKGYU^+OOWPQNQI[
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 259 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS the DNA damage response protein ataxia telangiectasia mutated (ATM) are involved in ensuring RAG-mediated cleavage of only one allele at a time64. One speculative possibility is that homologue pairing is mediated by a physical interaction between the recombination centres on the two alleles; this interaction could then facilitate communication between the two recombination centres so as to prevent initiation of V(D)J recombination simultaneously on the two alleles. It is attractive to think that the recombination centre serves as a hub, or ‘V(D)J recombination factory’91, within which both the DNA cleavage and DNA repair phases of the reaction are orchestrated. The other factors (proteins or RNA) found within recombination centres are not known, although it is likely that RNA polymerase II, other components of the transcription apparatus, histone modifying enzymes and chromatin remodelling complexes are present before DNA cleavage takes place. After cleavage, factors involved in the DNA damage response (such as ATM) and DNA repair (such as NHEJ proteins) are probably recruited into the recombination centre, and RAG might have an active role in this recruitment. DNA repair during V(D)J recombination is channelled into the NHEJ pathway by the RAG proteins90,92,93, and RAG1 is able to interact with the NHEJ factors Ku70 (also known as XRCC6) and Ku80 (also known as XRCC5)94 and with the DNA damage response factor MDC1 (mediator of DNA damage checkpoint protein 1) (G. Coster, D. Chen, D.G. Schatz and M. Goldberg, unpublished observations). An important open question is whether any DNA damage response or repair factors are recruited into recombination centres before DNA cleavage.
Homologous recombination Genetic recombination that occurs between regions of DNA with long stretches of homology. This occurs with a low frequency in somatic cells and at a much higher frequency in germ cells.
Steps leading to the formation of the paired complex. The locations of the recombination centres identified by ChIP analysis show no correlation with the presence of a particular type of RSS85, as JH and Jκ gene segments are flanked by 23RSSs, whereas Jβ and Jα gene segments are flanked by 12RSSs. It is therefore puzzling that an earlier study detected RAG-mediated nicks at 12RSSs but not 23RSSs in all antigen receptor loci examined, including Igk, leading to a ‘12RSS‑first’ model for paired complex formation during V(D)J recombination95. Why were nicks observed at Vκ but not Jκ gene segments when RAG binding was detected at Jκ but not Vκ gene segments? One potential explanation is that the ChIP analysis failed to detect binding to Vκ, perhaps because the RAG proteins are bound to only a small subset of the many (~140) Vκ gene segments in any one cell. This possibility cannot be ruled out, although it does not readily explain the absence of nicks at the highly accessible Jκ gene segments. In addition, as discussed previously85, the 12RSS‑first model cannot easily accommodate the RAG binding patterns observed at the Igh locus. As an alternative, we propose a ‘nonamer-first’ RSS recognition model that invokes asymmetry in the events that occur during RAG recognition of 12RSSs and 23RSSs in vivo. In this speculative model (FIG. 4), the initial step in RSS recognition is the freeing of a loop of RSS DNA from the surface of a nucleosome
by a chromatin remodelling complex 55, followed by binding of RAG1 (and RAG2 if present) and perhaps HMGB1 to the nonamer (FIG. 4b,c). The nonamer is the most likely site of initial DNA binding by RAG for two reasons. First, the nonamer serves as the stable ‘anchor’ for RAG on the RSS6, and second, the synapsis of two RSSs in solution requires RAG1, HMGB1 and the nonamer but not RAG2 or the heptamer 96. In the case of a 12RSS, nonamer binding is followed by heptamer engagement and nicking. However, in the case of a 23RSS, in which the heptamer and nonamer are farther apart on the face of the nucleosome, RAG anchored on the nonamer might have greater difficulty engaging the heptamer, so nicking might not occur as efficiently. The model therefore predicts that in a recombination centre containing 23RSSs (for example, at Igk), the predominant presynaptic species is a complex in which only the 23RSS nonamer is bound by RAG (FIG. 4c). We propose that synapsis of this 23RSS nonamer-only complex with a 12RSS allows for rapid 12RSS nicking (FIG. 4d,e) and slower nicking of the 23RSS (FIG. 4f). Once both RSSs are nicked, hairpin formation occurs rapidly at both RSSs97 (FIG. 4g), with the result that few nicked 23RSSs would accumulate within the cell. The nonamer-first model predicts that if a 23RSS were completely freed from its inhibitory interactions with a nucleosome, then nicking should occur before synapsis and perhaps be detectable. Consistent with this, a recent study observed nicking at the Dβ1 23RSS and not at Jβ 12RSSs98. Given that germline transcription from PDβ1 initiates within the Dβ1 23RSS99 and that Dβ1 resides in a long internucleosomal interval27, it is plausible that the Dβ1 23RSS is fully accessible for RAG binding and is readily nicked. Overall, current evidence suggests that V(D)J recombination need not initiate with binding of RAG to a 12RSS. Ectopic RAG binding and genome instability. The mouse and human genomes contain millions of cryptic RSS sequences, as well as many sequences capable of forming alternative (non‑B form) DNA structures5,100,101. Current evidence indicates that meaningful levels of RAG-mediated cleavage occur at some of these DNA sequences and that this contributes to lymphomaassociated genome alterations, such as deletion of the SCL–SIL locus in T cell acute lymphoblastic leukaemia and translocations of B cell lymphoma 2 (BCL2) to the IGH locus in follicular lymphoma1–3. The finding that RAG2 binds to thousands of transcriptionally active sites outside of antigen receptor genes raises several questions concerning the specificity and fidelity of V(D)J recombination. To what extent is RAG1 also recruited to these sites? If such recruitment occurs, is it mediated primarily through interactions with RAG2 or with cryptic RSSs, or by other mechanisms (such as recognition of alternative DNA structures86)? Do the RAG proteins interact with components of the transcriptional apparatus, and might the RAG2–H3K4me3 interaction influence histone modifications, chromatin structure or transcription? Most importantly, how often do the RAG proteins create nicks and double strand breaks at ectopic sites in the genome,
260 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS C 0QPCOGT
D
E
F 455
455 4#) 4#) */)$
0WENGQUQOG
*GRVCOGT
0KEM
G
H
0KEM
I
J (CUV
*CKTRKP HQTOCVKQP
*CKTRKP
*CKTRKP
Figure 4 | The nonamer-first model for RSS recognition. a | A schematic diagram of a 23 recombination signal sequence (23RSS) wrapped around a nucleosome, with the nonamer and heptamer indicated. b | Chromatin remodelling creates a loop of DNA, which might contain the nonamer, free of the nucleosome surface. 0CVWTG4GXKGYU^+OOWPQNQI[ c | A RAG1–RAG2–HMGB1 complex binds the 23RSS nonamer but fails to engage the heptamer. d | The nonamer-only 23RSS complex synapses with the 12RSS, and engagement of the 12RSS heptamer allows for nicking. e | The nicked 12RSS resides in a relatively stable synaptic complex, but the heptamer of the 23RSS is not engaged. f | Eventually, the heptamer of the 23RSS is freed from the nucleosome surface and can be engaged by the RAG complex. Nicking of the 23RSS can now occur. g | The paired complex in which both RSSs are nicked rapidly undergoes hairpin formation at both RSSs. h | Hairpin formation yields the doubly cleaved product. This pathway is proposed to be important in situations where the recombination centre contains 23RSSs (for example, in the Igk locus). HMGB1, high-mobility group protein B1; RAG, recombination activating gene.
what mechanisms exist to suppress such ectopic DNA damage and under what circumstances do the RAG proteins circumvent these controls? A comprehensive cataloguing of sites of RAG binding and of RAG-mediated instability will begin to answer some of these questions. Interestingly, both RAG and activation-induced cytidine deaminase (AID), which is crucial for immunoglobulin gene somatic hypermutation and class-switch recombination, are recruited to active regions of chromatin85,102 and to altered DNA structures2,89,103. Furthermore, it has been suggested that the combined action of RAG and AID can lead to DNA nicks and breaks and to genome instability 3. Hence, a convergence of the mechanisms that recruit these two dangerous enzyme systems to chromatin might increase the risk of genome instability in B cells and help explain why the majority of human lymphoid malignancies originate from B cells3.
Conclusions V(D)J recombination is essential for the development of the adaptive immune system in humans and most other vertebrate species. The reaction is capable of generating an extraordinarily diverse array of antigen
receptors, but this diversity comes at the cost of occasional genome instability and lymphoid malignancies. Recent advances have revealed important new aspects of how the RAG proteins are guided to appropriate target sites and have suggested new pathways that might explain mistargeting. These discoveries have revealed that complex processes related to nuclear dynamics and higher order chromatin architecture are layered on top of mechanisms that create gene segment accessibility so as to achieve lineage-specific and developmentally appropriate patterns of recombination. The RAG proteins are intimately involved in the regulation of V(D)J recombination through their ability to read a component of the histone code and to localize in focal recombination centres in antigen receptor loci. But some of these same properties of the RAG proteins might well contribute to their mistargeting to loci that do not encode antigen receptors and, ultimately, to the creation of ectopic DNA breaks and genome instability. Future experiments will probably reveal additional layers in the regulation of V(D)J recombination, as well as new regulatory functions inherent in RAG1 and RAG2.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 261 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS 1. 2.
3.
4.
5. 6. 7. 8.
9. 10. 11. 12. 13. 14.
15. 16. 17. 18.
19.
20.
21. 22.
23.
Mills, K. D., Ferguson, D. O. & Alt, F. W. The role of DNA breaks in genomic instability and tumorigenesis. Immunol. Rev. 194, 77–95 (2003). Lieber, M. R., Yu, K. F. & Raghavan, S. C. Roles of nonhomologous DNA end joining, V(D)J recombination, and class switch recombination in chromosomal translocations. DNA Repair 5, 1234–1245 (2006). Tsai, A. G. et al. Human chromosomal translocations at CpG sites and a theoretical basis for their lineage and stage specificity. Cell 135, 1130–1142 (2008). This study proposed, and provided support for, the fascinating idea that RAG and AID collaborate to create DNA nicks and breaks at methylated CpG sequences. Yancopoulos, G. D. & Alt, F. W. Developmentally controlled and tissue-specific expression of unrearranged VH gene segments. Cell 40, 271–281 (1985). This classic study provided the first evidence for, and proposed, the accessibility model for the control of V(D)J recombination. Lewis, S. M. The mechanism of V(D)J joining: lessons from molecular, immunological, and comparative analyses. Adv. Immunol. 56, 27–150 (1994). Swanson, P. C. The bounty of RAGs: recombination signal complexes and reaction outcomes. Immunol. Rev. 200, 90–114 (2004). Fugmann, S. D. & Schatz, D. G. Identification of basic residues in RAG2 critical for DNA binding by the RAG1–RAG2 complex. Mol. Cell 8, 899–910 (2001). Mundy, C. L., Patenge, N., Matthews, A. G. W. & Oettinger, M. A. Assembly of the RAG1/RAG2 synaptic complex. Mol. Cell. Biol. 22, 69–77 (2002). Jones, J. M. & Gellert, M. Ordered assembly of the V(D)J synaptic complex ensures accurate recombination. EMBO J. 21, 4162–4171 (2002). Gellert, M. V(D)J recombination: RAG proteins, repair factors, and regulation. Annu. Rev. Biochem. 71, 101–132 (2002). Cobb, R. M., Oestreich, K. J., Osipovich, O. A. & Oltz, E. M. Accessibility control of V(D)J recombination. Adv. Immunol. 91, 45–109 (2006). Hesslein, D. G. & Schatz, D. G. Factors and forces controlling V(D)J recombination. Adv. Immunol. 78, 169–232 (2001). Yancopoulos, G. D. & Alt, F. W. Regulation of the assembly and expression of variable-region genes. Annu. Rev. Immunol. 4, 339–368 (1986). Jung, D., Giallourakis, C., Mostoslavsky, R. & Alt, F. W. Mechanism and control of V(D)J recombination at the immunoglobulin heavy chain locus. Annu. Rev. Immunol. 24, 541–570 (2006). Goldmit, M. & Bergman, Y. Monoallelic gene expression: a repertoire of recurrent themes. Immunol. Rev. 200, 197–214 (2004). Krangel, M. S. T cell development: better living through chromatin. Nature Immunol. 8, 687–694 (2007). Schlissel, M. S. Regulating antigen-receptor gene assembly. Nature Rev. Immunol. 3, 890–899 (2003). Stanhope-Baker, P., Hudson, K. M., Shaffer, A. L., Constantinescu, A. & Schlissel, M. S. Cell type-specific chromatin structure determines the targeting of V(D)J recombinase activity in vitro. Cell 85, 887–897 (1996). This classic study demonstrated that chromatin structure is a critical determinant of DNA cutting by RAG. Kwon, J., Imbalzano, A. N., Matthews, A. & Oettinger, M. A. Accessibility of nucleosomal DNA to V(D)J cleavage is modulated by RSS positioning and HMG1. Mol. Cell 2, 829–839 (1998). Golding, A., Chandler, S., Ballestar, E., Wolffe, A. P. & Schlissel, M. S. Nucleosome structure completely inhibits in vitro cleavage by the V(D)J recombinase. EMBO J. 18, 3712–3723 (1999). McBlane, F. & Boyes, J. Stimulation of V(D)J recombination by histone acetylation. Curr. Biol. 10, 483–486 (2000). Kwon, J., Morshead, K. B., Guyon, J. R., Kingston, R. E. & Oettinger, M. A. Histone acetylation and hSWI/SNF remodeling act in concert to stimulate V(D)J cleavage of nucleosomal DNA. Mol. Cell 6, 1037–1048 (2000). Patenge, N., Elkin, S. K. & Oettinger, M. A. ATPdependent remodeling by SWI/SNF and ISWI proteins stimulates V(D)J cleavage of 5 S arrays. J. Biol. Chem. 279, 35360–35367 (2004).
24. Nightingale, K. P. et al. Acetylation increases access of remodelling complexes to their nucleosome targets to enhance initiation of V(D)J recombination. Nucl. Acids Res. 35, 6311–6321 (2007). 25. Du, H., Ishii, H., Pazin, M. J. & Sen, R. Activation of 12/23‑RSS‑dependent RAG cleavage by hSWI/SNF complex in the absence of transcription. Mol. Cell 31, 641–649 (2008). 26. Baumann, M., Mamais, A., McBlane, F., Xiao, H. & Boyes, J. Regulation of V(D)J recombination by nucleosome positioning at recombination signal sequences. EMBO J. 22, 5197–5207 (2003). 27. Kondilis-Mangum, H. D. et al. Transcription-dependent mobilization of nucleosomes at accessible TCR gene segments in vivo. J. Immunol. 184, 6970–6977 (2010). 28. Kaplan, N. et al. The DNA-encoded nucleosome organization of a eukaryotic genome. Nature 458, 362–366 (2009). 29. Segal, E. & Widom, J. What controls nucleosome positions? Trends Genet. 25, 335–343 (2009). 30. Zhang, Y. et al. Intrinsic histone–DNA interactions are not the major determinant of nucleosome positions in vivo. Nature Struct. Mol. Biol. 16, 847–852 (2009). 31. Segal, E. & Widom, J. Poly(dA:dT) tracts: major determinants of nucleosome organization. Curr. Opin. Struct. Biol. 19, 65–71 (2009). 32. Chowdhury, D. & Sen, R. Regulation of immunoglobulin heavy-chain gene rearrangements. Immunol. Rev. 200, 182–196 (2004). 33. Krangel, M. S. Mechanics of T cell receptor gene rearrangement. Curr. Opin. Immunol. 21, 133–139 (2009). 34. Jhunjhunwala, S., van Zelm, M. C., Peak, M. M. & Murre, C. Chromatin architecture and the generation of antigen receptor diversity. Cell 138, 435–448 (2009). 35. Schlissel, M. S. Regulation of activation and recombination of the murine Igκ locus. Immunol. Rev. 200, 215–223 (2004). 36. Abarrategui, I. & Krangel, M. S. Regulation of T cell receptor‑α gene recombination by transcription. Nature Immunol. 7, 1109–1115 (2006). This study provided the first direct support for the long held idea that transcriptional elongation is important for creating accessibility for V(D)J recombination. 37. Abarrategui, I. & Krangel, M. S. Noncoding transcription controls downstream promoters to regulate T‑cell receptor α recombination. EMBO J. 26, 4380–4390 (2007). 38. Sikes, M. L., Meade, A., Tripathi, R., Krangel, M. S. & Oltz, E. M. Regulation of V(D)J recombination: a dominant role for promoter positioning in gene segment accessibility. Proc. Natl Acad. Sci. USA 99, 12309–12314 (2002). 39. Fernex, C., Capone, M. & Ferrier, P. The V(D)J recombinational and transcriptional activities of the immunoglobulin heavy-chain intronic enhancer can be mediated through distinct protein-binding sites in a transgenic substrate. Mol. Cell. Biol. 15, 3217–3226 (1995). 40. Bolland, D. J. et al. Antisense intergenic transcription in V(D)J recombination. Nature Immunol. 5, 630–637 (2004). 41. Bolland, D. J. et al. Antisense intergenic transcription precedes Igh D‑to‑J recombination and is controlled by the intronic enhancer Eμ. Mol. Cell. Biol. 27, 5523–5533 (2007). 42. Chakraborty, T. et al. Repeat organization and epigenetic regulation of the DH‑Cμ domain of the immunoglobulin heavy-chain gene locus. Mol. Cell 27, 842–850 (2007). 43. Osipovich, O. A., Subrahmanyam, R., Pierce, S., Sen, R. & Oltz, E. M. Cutting edge: SWI/SNF mediates antisense Igh transcription and locus-wide accessibility in B cell precursors. J. Immunol. 183, 1509–1513 (2009). 44. Giallourakis, C. C. et al. Elements between the IgH variable (V) and diversity (D) clusters influence antisense transcription and lineage-specific V(D)J recombination. Proc. Natl Acad. Sci. USA 107, 22207–22212 (2010). 45. Oettinger, M. A. How to keep V(D)J recombination under control. Immunol. Rev. 200, 165–181 (2004). 46. Chakraborty, T. et al. A 220‑nucleotide deletion of the intronic enhancer reveals an epigenetic hierarchy in immunoglobulin heavy chain locus activation. J. Exp. Med. 206, 1019–1027 (2009).
262 | APRIL 2011 | VOLUME 11
47. Osipovich, O. et al. Targeted inhibition of V(D)J recombination by a histone methyltransferase. Nature Immunol. 5, 309–316 (2004). 48. Morshead, K. B., Ciccone, D. N., Taverna, S. D., Allis, C. D. & Oettinger, M. A. Antigen receptor loci poised for V(D)J rearrangement are broadly associated with BRG1 and flanked by peaks of histone H3 dimethylated at lysine 4. Proc. Natl Acad. Sci. USA 100, 11577–11582 (2003). 49. Johnson, K. et al. B cell-specific loss of histone 3 lysine 9 methylation in the VH locus depends on Pax5. Nature Immunol. 5, 853–861 (2004). 50. Smale, S. T. The establishment and maintenance of lymphocyte identity through gene silencing. Nature Immunol. 4, 607–615 (2003). 51. Workman, J. L. Nucleosome displacement in transcription. Genes Dev. 20, 2009–2017 (2006). 52. Li, B., Carey, M. & Workman, J. L. The role of chromatin during transcription. Cell 128, 707–719 (2007). 53. Barski, A. et al. High-resolution profiling of histone methylations in the human genome. Cell 129, 823–837 (2007). 54. Shogren-Knaak, M. et al. Histone H4‑K16 acetylation controls chromatin structure and protein interactions. Science 311, 844–847 (2006). 55. Zhang, Y. et al. DNA translocation and loop formation mechanism of chromatin remodeling by SWI/SNF and RSC. Mol. Cell 24, 559–568 (2006). 56. Wang, X. et al. Regulation of Tcrb recombination ordering by c‑Fos‑dependent RAG deposition. Nature Immunol. 9, 794–801 (2008). 57. Zhang, Z. X. et al. Transcription factor Pax5 (BSAP) transactivates the RAG-mediated VH‑to-DJH rearrangement of immunoglobulin genes. Nature Immunol. 7, 616–624 (2006). 58. Hesslein, D. G. et al. Pax5 is required for recombination of transcribed, acetylated, 5ʹ IgH V gene segments. Genes Dev. 17, 37–42 (2003). 59. Jackson, A., Kondilis, H. D., Khor, B., Sleckman, B. P. & Krangel, M. S. Regulation of T cell receptor β allelic exclusion at a level beyond accessibility. Nature Immunol. 6, 189–197 (2005). 60. Kosak, S. T. & Groudine, M. Form follows function: the genomic organization of cellular differentiation. Genes Dev. 18, 1371–1384 (2004). 61. Reddy, K. L., Zullo, J. M., Bertolino, E. & Singh, H. Transcriptional repression mediated by repositioning of genes to the nuclear lamina. Nature 452, 243–247 (2008). 62. Kosak, S. T. et al. Subnuclear compartmentalization of immunoglobulin loci during lymphocyte development. Science 296, 158–162 (2002). This paper provided the first evidence that V(D)J recombination is associated with contraction and movement away from the nuclear periphery of the recombining locus. 63. Skok, J. A. et al. Reversible contraction by looping of the Tcra and Tcrb loci in rearranging thymocytes. Nature Immunol. 8, 378–387 (2007). 64. Hewitt, S. L. et al. RAG‑1 and ATM coordinate monoallelic recombination and nuclear positioning of immunoglobulin loci. Nature Immunol. 10, 655–664 (2009). 65. Goldmit, M. et al. Epigenetic ontogeny of the Igk locus during B cell development. Nature Immunol. 6, 198–203 (2005). 66. Schlimgen, R. J., Reddy, K. L., Singh, H. & Krangel, M. S. Initiation of allelic exclusion by stochastic interaction of Tcrb alleles with repressive nuclear compartments. Nature Immunol. 9, 802–809 (2008). 67. Hewitt, S. L. et al. Association between the Igk and Igh immunoglobulin loci mediated by the 3ʹ Igk enhancer induces ‘decontraction’ of the Igh locus in pre‑B cells. Nature Immunol. 9, 396–404 (2008). 68. Roldan, E. et al. Locus ‘decontraction’ and centromeric recruitment contribute to allelic exclusion of the immunoglobulin heavy-chain gene. Nature Immunol. 6, 31–41 (2005). 69. Jia, J., Kondo, M. & Zhuang, Y. Germline transcription from T‑cell receptor Vβ gene is uncoupled from allelic exclusion. EMBO J. 26, 2387–2399 (2007). 70. Singh, N., Bergman, Y., Cedar, H. & Chess, A. Biallelic germline transcription at the κ immunoglobulin locus. J. Exp. Med. 197, 743–750 (2003). 71. Amin, R. H. et al. Biallelic, ubiquitous transcription from the distal germline Igκ locus promoter during B cell development. Proc. Natl Acad. Sci. USA 106, 522–527 (2009).
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS 72. Fitzsimmons, S. P., Bernstein, R. M., Max, E. E., Skok, J. A. & Shapiro, M. A. Dynamic changes in accessibility, nuclear positioning, recombination, and transcription at the Igκ locus. J. Immunol. 179, 5264–5273 (2007). 73. Sayegh, C., Jhunjhunwala, S., Riblet, R. & Murre, C. Visualization of looping involving the immunoglobulin heavy-chain locus in developing B cells. Genes Dev. 19, 322–327 (2005). 74. Fuxa, M. et al. Pax5 induces V‑to‑DJ rearrangements and locus contraction of the immunoglobulin heavychain gene. Genes Dev. 18, 411–422 (2004). 75. Jhunjhunwala, S. et al. The 3D structure of the immunoglobulin heavy-chain locus: implications for long-range genomic interactions. Cell 133, 265–279 (2008). This landmark study provided the highest resolution picture to date of the higher order chromatin architecture and dynamics of an antigen receptor locus. 76. Shih, H. Y. & Krangel, M. S. Distinct contracted conformations of the Tcra/Tcrd locus during Tcra and Tcrd recombination. J. Exp. Med. 207, 1835–1841 (2010). 77. Callebaut, I. & Mornon, J. P. The V(D)J recombination activating protein RAG2 consists of a six-bladed propeller and a PHD fingerlike domain, as revealed by sequence analysis. Cell. Mol. Life Sci. 54, 880–891 (1998). 78. Elkin, S. K. et al. A PHD finger motif in the C terminus of RAG2 modulates recombination activity. J. Biol. Chem. 280, 28701–28710 (2005). 79. West, K. L. et al. A direct interaction between the RAG2 C terminus and the core histones is required for efficient V(D)J recombination. Immunity 23, 203–212 (2005). 80. Matthews, A. G. et al. RAG2 PHD finger couples histone H3 lysine 4 trimethylation with V(D)J recombination. Nature 450, 1106–1110 (2007). This study provided a high resolution structure for the PHD finger of RAG2 bound to H3K4me3, as well as evidence for the biological importance of this interaction. 81. Liu, Y., Subrahmanyam, R., Chakraborty, T., Sen, R. & Desiderio, S. A plant homeodomain in RAG‑2 that binds hypermethylated lysine 4 of histone H3 is necessary for efficient antigen‑receptor‑gene rearrangement. Immunity 27, 561–571 (2007). This study demonstrated that the RAG2 PHD finger binds H3K4me3 and provided evidence for the biological importance of this interaction. 82. Gomez, C. A. et al. Mutations in conserved regions of the predicted RAG2 kelch repeats block initiation of V(D)J recombination and result in primary immunodeficiencies. Mol. Cell. Biol. 20, 5653–5664 (2000). 83. Shimazaki, N., Tsai, A. G. & Lieber, M. R. H3K4me3 stimulates the V(D)J RAG complex for both nicking and hairpinning in trans in addition to tethering in cis: implications for translocations. Mol. Cell 34, 535–544 (2009). 84. Grundy, G. J., Yang, W. & Gellert, M. Autoinhibition of DNA cleavage mediated by RAG1 and RAG2 is overcome by an epigenetic signal in V(D)J recombination. Proc. Natl Acad. Sci. USA 107, 22487–22492 (2010). 85. Ji, Y. et al. The in vivo pattern of binding of RAG1 and RAG2 to antigen receptor loci. Cell 141, 419–431 (2010). This study demonstrated that the binding of RAG1 and RAG2 to antigen receptor loci is developmentally regulated, lineage restricted and focused on small, highly active regions of chromatin referred to as recombination centres. 86. Zhao, S., Gwyn, L. M., De, P. & Rodgers, K. K. A non‑sequence‑specific DNA binding mode of RAG1
is inhibited by RAG2. J. Mol. Biol. 387, 744–758 (2009). 87. Ji, Y. et al. Promoters, enhancers, and transcription target RAG1 binding during V(D)J recombination. J. Exp. Med. 207, 2809–2816 (2010). 88. Grazini, U. et al. The RING domain of RAG1 ubiquitylates histone H3: a novel activity in chromatinmediated regulation of V(D)J joining. Mol. Cell 37, 282–293 (2010). 89. Raghavan, S. C., Swanson, P. C., Wu, X., Hsieh, C. L. & Lieber, M. R. A non‑B‑DNA structure at the Bcl‑2 major breakpoint region is cleaved by the RAG complex. Nature 428, 88–93 (2004). 90. Lee, G. S., Neiditch, M. B., Salus, S. S. & Roth, D. B. RAG proteins shepherd double-strand breaks to a specific pathway, suppressing error-prone repair, but RAG nicking initiates homologous recombination. Cell 117, 171–184 (2004). 91. Matthews, A. G. W. & Oettinger, M. A. RAG: a recombinase diversified. Nature Immunol. 10, 817–821 (2009). 92. Corneo, B. et al. Rag mutations reveal robust alternative end joining. Nature 449, 483–486 (2007). 93. Cui, X. & Meek, K. Linking double-stranded DNA breaks to the recombination activating gene complex directs repair to the nonhomologous end-joining pathway. Proc. Natl Acad. Sci. USA 104, 17046–17051 (2007). 94. Raval, P., Kriatchko, A. N., Kumar, S. & Swanson, P. C. Evidence for Ku70/Ku80 association with full-length RAG1. Nucl. Acids Res. 36, 2060–2072 (2008). 95. Curry, J. D., Geier, J. K. & Schlissel, M. S. Singlestrand recombination signal sequence nicks in vivo: evidence for a capture model of synapsis. Nature Immunol. 6, 1272–1279 (2005). 96. Yin, F. F. et al. Structure of the RAG1 nonamer binding domain with DNA reveals a dimer that mediates DNA synapsis. Nature Struct. Mol. Biol. 16, 499–508 (2009). 97. Yu, K. F. & Lieber, M. R. Mechanistic basis for coding end sequence effects in the initiation of V(D)J recombination. Mol. Cell. Biol. 19, 8094–8102 (1999). 98. Franchini, D. M., Benoukraf, T., Jaeger, S., Ferrier, P. & Payet-Bornet, D. Initiation of V(D)J recombination by Dβ-associated recombination signal sequences: a critical control point in TCRβ gene assembly. PLoS ONE 4, e4575 (2009). 99. Sikes, M. L., Gomez, R. J., Song, J. & Oltz, E. M. A developmental stage-specific promoter directs germline transcription of DβJβ gene segments in precursor T lymphocytes. J. Immunol. 161, 1399–1405 (1998). 100. Merelli, I. et al. RSSsite: a reference database and prediction tool for the identification of cryptic recombination signal sequences in human and murine genomes. Nucl. Acids Res. 38, W262–W267 (2010). 101. Zhao, J., Bacolla, A., Wang, G. & Vasquez, K. M. Non‑B DNA structure-induced genetic instability and evolution. Cell. Mol. Life Sci. 67, 43–62 (2010). 102. Yamane, A. et al. Deep-sequencing identification of the genomic targets of the cytidine deaminase AID and its cofactor RPA in B lymphocytes. Nature Immunol. 12, 62–69 (2010). 103. Staszewski, O. et al. Activation-induced cytidine deaminase induces reproducible DNA breaks at many non-Ig loci in activated B cells. Mol. Cell 41, 232–242 (2011). 104. Leu, T. M. & Schatz, D. G. rag‑1 and rag‑2 are components of a high‑molecular‑weight complex, and association of rag‑2 with this complex is rag‑1 dependent. Mol. Cell. Biol. 15, 5657–5670 (1995).
NATURE REVIEWS | IMMUNOLOGY
105. Lin, W.‑C. & Desiderio, S. Cell cycle regulation of V(D)J recombination-activating protein RAG‑2. Proc. Natl Acad. Sci. USA 91, 2733–2737 (1994). 106. Lee, J. & Desiderio, S. Cyclin A/CDK2 regulates V(D)J recombination by coordinating RAG‑2 accumulation and DNA repair. Immunity 11, 771–781 (1999). 107. Fugmann, S. D., Lee, A. I., Shockett, P. E., Villey, I. J. & Schatz, D. G. The RAG proteins and V(D)J recombination: complexes, ends, and transposition. Annu. Rev. Immunol. 18, 495–527 (2000). 108. De, P. & Rodgers, K. K. Putting the pieces together: identification and characterization of structural domains in the V(D)J recombination protein RAG1. Immunol. Rev. 200, 70–82 (2004). 109. Mathieu, N., Hempel, W. M., Spicuglia, S., Verthuy, C. & Ferrier, P. Chromatin remodeling by the T cell receptor (TCR)‑β gene enhancer during early T cell development: implications for the control of TCR‑β locus recombination. J. Exp. Med. 192, 625–636 (2000). 110. Bouvier, G. et al. Deletion of the mouse T‑cell receptor β gene enhancer blocks α‑β T‑cell development. Proc. Natl Acad. Sci. USA 93, 7877–7881 (1996). 111. Bories, J. C., Demengeot, J., Davidson, L. & Alt, F. W. Gene-targeted deletion and replacement mutations of the T‑cell receptor β‑chain enhancer: the role of enhancer elements in controlling V(D)J recombination accessibility. Proc. Natl Acad. Sci. USA 93, 7871–7876 (1996). 112. Oestreich, K. J. et al. Regulation of TCR β gene assembly by a promoter/enhancer holocomplex. Immunity 24, 381–391 (2006). 113. Spicuglia, S. et al. Promoter activation by enhancerdependent and -independent loading of activator and coactivator complexes. Mol. Cell 10, 1479–1487 (2002). 114. Whitehurst, C. E., Chattopadhyay, S. & Chen, J. Control of V(D)J recombinational accessibility of the Dβ1 gene segment at the TCRβ locus by a germline promoter. Immunity 10, 313–322 (1999). 115. Whitehurst, C. E., Schlissel, M. S. & Chen, J. Deletion of germline promoter PDβ1 from the TCRβ locus causes hypermethylation that impairs Dβ1 recombination by multiple mechanisms. Immunity 13, 703–714 (2000). 116. McMillan, R. E. & Sikes, M. L. Differential activation of dual promoters alters Dβ2 germline transcription during thymocyte development. J. Immunol. 180, 3218–3228 (2008). 117. McMillan, R. E. & Sikes, M. L. Promoter activity 5’ of Dβ2 is coordinated by E47, Runx1, and GATA‑3. Mol. Immunol. 46, 3009–3017 (2009). 118. Osipovich, O. et al. Essential function for SWI-SNF chromatin-remodeling complexes in the promoterdirected assembly of Tcrb genes. Nature Immunol. 8, 809–816 (2007).
Acknowledgements
The authors wish to thank E. Oltz, G. Teng, K. Shetty and J. Banerjee for comments on the manuscript. We apologize that not all of the relevant literature could be cited owing to space constraints.
Competing interests statement
The authors declare no competing financial interests.
FURTHER INFORMATION David G. Schatz’s homepage: http://medicine.yale.edu/ immuno/people/david_schatz.profile ALL LINKS ARE ACTIVE IN THE ONLINE PDF
VOLUME 11 | APRIL 2011 | 263 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS
Platelets and the immune continuum John W. Semple*‡§, Joseph E. Italiano Jr||¶ and John Freedman*‡
Abstract | Platelets are anucleate cells that are crucial mediators of haemostasis. Most immunologists probably don’t think about platelets every day, and may even consider these cells to be ‘nuisances’ in certain in vitro studies. However, it is becoming increasingly clear that platelets have inflammatory functions and can influence both innate and adaptive immune responses. Here, we discuss the mechanisms by which platelets contribute to immunity: these small cells are more immunologically savvy than we once thought. Megakaryocyte A giant multinucleated cell of the bone marrow that gives rise to platelets.
Thrombocytopenia Any disorder in which there is an abnormally low number of platelets (below 150,000 platelets per microlitre of blood). Thrombocytopenia is caused either by increased breakdown of platelets (in the blood, spleen or liver) or by low production of platelets.
*Toronto Platelet Immunobiology Group, Keenan Research Centre, Li Ka Shing Knowledge Institute, St. Michael’s Hospital, 30 Bond Street, Toronto, Ontario, Canada, M5B 1W8. ‡ Departments of Pharmacology, Medicine and Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada. § Canadian Blood Services, Toronto, Ontario, Canada. || Division of Hematology, Brigham and Women’s Hospital, Boston, Massachusetts, USA. ¶ Vascular Biology Program, Children’s Hospital, Harvard Medical School, Boston, Massachusetts, USA. Correspondence to J.W.S. e-mail:
[email protected] doi:10.1038/nri2956
In 1862, Giulio Bizzozero, the so-called ‘father of the platelet’, described a novel ‘morphological element’ in the blood with important roles in haemorrhage and thrombosis1. Platelets (as these elements became known) are small (approximately 2 μm in diameter), circulating, anucleate cells that are derived from megakaryocytes within the bone marrow. Their primary physiological role is to sense damaged vessel endothelium and accumulate at the site of the vessel injury, where they initiate blood clotting to block the circulatory leak. Platelets circulate at high-shear rates and are activated following binding to the collagen substratum or to other extra cellular matrix proteins that are exposed during vascular injury. Stable adhesion to collagen promotes the release of many soluble mediators from the platelet’s intracellular stores, leading to further platelet recruitment and activation. These events are regulated by complex interactions involving several families of molecules, including various selectins, integrins, lipids and cytokines (FIG. 1). Together with leukocytes and red blood cells, the activated platelets create a thrombus that arrests blood loss2,3. Probably the oldest recognized link between platelets and the immune system was the observation that platelets are targets of both autoimmune attack (in the case of immune thrombocytopenia) and alloimmune attack (which causes transfusion refractoriness, fetal or neonatal alloimmune thrombocytopenia and posttransfusion purpura) (BOX 1). However, there has been a gradual realization that platelets have important roles in modulating innate and adaptive immune responses. For example, platelets have been shown to have roles in the initiation of inflammation, angiogenesis, atherosclerosis, lymphatic development and tumour growth4. How platelets multitask and perform such diverse immune-related functions is still an enigma, but various lines of evidence suggest that many distinct facets of platelet biology are important. These include the unique origin and structure of platelets, their expression of immunomodulatory
molecules and cytokines and their ability to interact with various cells of the immune system. In this Review, we outline some of the key properties of platelets that enable them to contribute to immunity. We argue that these tiny fragments are perhaps as important for host immunity as they are for haemostasis.
Platelet origin, formation and structure The capacity of platelets to regulate immunity is suggested by their unique production and structural characteristics. There are approximately one trillion platelets circulating in the blood of an adult human, and because the lifespan of an individual platelet is only 8–10 days, 100 billion new platelets must be produced daily from bone marrow megakaryocytes in order to maintain normal platelet counts (150–400 × 109 platelets per litre of blood)5,6. In mice, platelet counts are generally much higher than in humans, with levels in the range of 900–1600 × 109 platelets per litre of blood. Owing to their extremely high numbers and their ability to release inflammatory mediators, platelets are uniquely positioned to perform several sentinel tasks and to quickly communicate with the cells of the immune system. Interestingly, like erythrocytes, platelets are confined to the circulation and do not enter the lymphatics, so they primarily interact with leukocyte populations in the spleen or liver. Whether this attribute is important in how platelets may modulate the immune response is unknown. Anucleate platelets are only found in mammals. Lower vertebrates, such as birds, reptiles, amphibians and fish, possess nucleated thrombocytes that carry out haemostatic functions7. Invertebrate species do not have platelets per se, but contain circulating nucleated cells in their haemolymph termed haemocytes7–9; these are similar in structure and function to vertebrate macrophages and express Toll-like receptors (TLRs), mediate phago cytosis and secrete antimicrobial peptides8,9. Invertebrate haemocytes, however, also mediate wound healing and
264 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS 2NCVGNGVUCFJGTGVQUKVGQHXCUEWNCTKPLWT[ 0GWVTQRJKN
4GF DNQQFEGNN
#EVKXCVKQPQH EQCIWNCVKQP ECUECFG
6JTQODQZCPG# 'PFQVJGNKCN EGNN
2NCVGNGV #&2
$CUGOGPV OGODTCPG
*CGOQUVCVKERNWIHQTOCVKQP
2NCVGNGVCIITGICVKQPCPFCEVKXCVKQP
6JTQODKP
(KDTKP
%QNNCIGPCPF '%/RTQVGKPU 5OQQVJOWUENGEGNN
Figure 1 | Platelet functions in haemostasis. The platelet’s primary physiological role is thought to be in haemostasis. 0CVWTG4GXKGYU^+OOWPQNQI[ In the first step of this process, a vascular injury exposes collagen and basement membrane proteins that allow the platelets to adhere to the substratum. The adherent platelets then aggregate and release platelet activation mediators, such as ADP and thromboxane A2. Following activation, the platelets produce thrombin, which catalyses the initiation of the coagulation cascade that eventually generates a mesh-like fibrin deposition. This structure contracts to form a tightly packed haemostatic plug that arrests blood leakage. ECM, extracellular matrix.
can induce haemolymph coagulation and clotting at sites of tissue injury or exoskeletal disruption8,9. These analogous primordial haematological responses may point to an evolutionary link between innate immune responses and platelet function. Perhaps, during evolution, the haemostatic and immune functions of haemocytes diverged and platelets and leukocytes were created. Theoretically, this could explain why platelets have many immune-related functions, that is, some of the early protective functions of haemocytes have lingered and still persist in the platelets of mammals.
Immune thrombocytopenia An abnormal drop in platelet numbers caused by the presence of platelet-specific antibodies. Platelet production may be normal or impaired, as the disorder can also be caused by antibodies directed against megakaryocytes. The term includes autoimmune, alloimmune and certain drug-induced thrombocytopenias.
Toll-like receptors A family of pattern recognition receptors that have essential roles in innate immunity. They are a class of single-membrane-spanning receptors that have the ability to recognize structurally conserved molecules from bacteria. Engagement of Toll-like receptors activates the immune response.
Platelet production. The production of platelets requires a complex progression of events that culminates with a single megakaryocyte releasing thousands of platelets into the circulation10,11. Megakaryocytes produce platelets by rearranging their cytoplasm into long extensions called proplatelets, which resemble strings of beads. These distinctive cellular appendages are built on a scaffold of motile microtubules that directs their formation11. The expansive network of elaborately branched and highly elongated proplatelets serves as an assembly line for platelet production. Although the megakaryocytes reside in the bone marrow, their proplatelets extend through junctions in the endothelial lining of blood vessels, from where they are released into the circulation and undergo further fragmentation into individual platelets. Recent work has described the mechanics of proplatelet maturation into individual platelets in the blood12.
by a well-defined and highly specialized cytoskeleton. This elaborate system of struts and girders preserves the shape and integrity of the platelet as it encounters highshear forces in the circulation. The three major cytoskeletal components of the resting platelet are the marginal microtubule coil, the spectrin-based membrane skeleton, and the actin-based cytoskeleton. One of the most distinguishing features of the resting platelet is its marginal microtubule coil, the primary function of which is to maintain the disc shape of the platelet. When platelets encounter a damaged vessel wall, they are activated and undergo a dramatic change of shape from smooth discs to spiny spheres. This process is initiated by influxes of calcium, which promote the formation of finger-like filopodia and pseudopods. The actin cytoskeleton provides the force for this rapid shape change13. During this reaction, the number of receptors for adhesive and clotting proteins in the platelet membrane increases, and activated platelets attract other platelets, which clump together and ultimately form a plug that seals off the vascular leak (FIG. 1). Furthermore, these processes lead to the expression of a plethora of cytokines, chemokines and cell surface molecules that not only initiate and perpetuate haemo stasis, but also serve to alert the immune system and induce leukocyte recruitment to the injured tissue. For example, activated platelets express P‑selectin and can therefore promote lymphocyte rolling and adhesion on high endothelial venules14.
Platelet structure. Blood platelets circulate as tiny oval discs (FIG. 2). The surface of the platelet plasma membrane is generally smooth except for periodic invaginations that delineate the entrances to the open canalicular system, a complex network of intertwining membrane tubes that permeate the platelet’s cytoplasm. Because of this system of folded membranes, platelets not only have an enormous surface area but can readily take up proteins and molecules and then re-release them on activation. The lentiform shape of the resting platelet is maintained
Platelet granules. Many of the non-haemostatic functions of platelets may result from their capacity to store a number of biologically active molecules in intracellular granules. These molecules can then be released into the circulation or translocated to the platelet surface during platelet activation. Platelets have three major types of storage granules: α-granules, dense granules and lysosomes. α-granules are the most abundant type of granule, with 40–80 per platelet, and they derive their protein content by a combination of endocytosis and
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 265 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Box 1 | Autoimmune and alloimmune platelet disorders Immune thrombocytopenia (ITP) is an autoimmune disorder in which platelets are opsonized by platelet-specific IgG autoantibodies. This leads to the premature destruction of platelets through Fc receptor-mediated phagocytosis by macrophages in the spleen90–92 (see the figure, part a). ITP has the distinction of being the first identified example of an immune-mediated disease that can be passively transferred by human plasma. In 1951, William Harrington infused himself with plasma from an ITP patient93; his platelet counts plummeted within hours and he required hospitalization. Subsequently, Evans identified the plasma factor involved as antibody94. Recent research has identified two other forms of platelet destruction in patients with ITP, mediated by CD8+ cytotoxic T lymphocytes (CTLs) and reactive oxygen species (ROS)95,96. Alloimmune platelet attack was observed when platelet components were first produced for transfusion in the 1950s. Allogeneic platelet transfusions frequently result in the development of MHC-specific alloantibodies, which target the platelets in subsequent transfusions and induce a state of platelet refractoriness. This prevents the recipient from deriving benefit from the transfusion97 (see the figure, part b). C+OOWPGVJTQODQE[VQRGPKC D6TCPUHWUKQPTGHTCEVQTKPGUU Platelets express a set of human platelet antigens (HPAs), and their diversity arises as a result of point mutations throughout the two major platelet glycoproteins, GPIIb–IIIa (also known as αIIbβ3 integrin) and GPIb–IX98. Fetal and neonatal alloimmune thrombocytopenia occurs when an HPA-negative mother carries an HPA-positive fetus and becomes 2TGUGPEGQHCWVQTGCEVKXG &GXGNQROGPVQHVTCPUHWUKQP immunized against the HPA antigens CPVKDQFKGUCPF%6.UNGCFUVQ KPFWEGF/*%URGEKȮECNNQCPVK expressed by the fetus when fetal RGTKRJGTCNRNCVGNGVFGUVTWEVKQP DQFKGUCPFUWDUGSWGPV blood mixes with the mother’s blood99. CPFOGICMCT[QE[VGKPJKDKVKQP VTCPUHWUKQPTGHTCEVQTKPGUU The mother’s HPA-specific IgG alloantibodies can then cross the E(GVCNCPFPGQPCVCNCNNQKOOWPG F2QUVVTCPUHWUKQPRWTRWTC placenta and destroy fetal, and in some VJTQODQE[VQRGPKC instances neonatal, platelets (see the figure, part c). *2#s Post-transfusion purpura is a rare condition in which a transfusion recipient becomes alloimmunized against the transfused platelets (usually owing to HPA-specific immune *2# recognition) and, a few weeks after the transfusion, develops thrombocytopenia. This condition is similar to ITP in that the &GXGNQROGPVQHVTCPUHWUKQP antibodies can destroy the recipient’s &GXGNQROGPVQH*2#URGEKȮE KPFWEGFCNNQCPVKDQFKGUCPF 100 own HPA-negative platelets (see the CNNQCPVKDQFKGUCPFUWDUGSWGPVHGVCN UWDUGSWGPVTGEKRKGPV figure, part d). CPFPGQPCVCNVJTQODQE[VQRGPKC VJTQODQE[VQRGPKC 0CVWTG4GXKGYU^+OOWPQNQI[
biosynthesis. The list of proteins housed in α-granules is extensive and includes coagulation factors, chemokines, adhesive proteins, mitogenic factors and regulators of angiogenesis. Although it was previously thought that these various molecules were packaged indiscriminately into one consistent population of α-granules that underwent simultaneous release, recent work suggests that platelets contain heterogeneous populations of α-granules that undergo differential patterns of release during platelet activation15–17. Sehgal and Storrie have identified two classes of α-granule: one that contains fibrinogen and another that contains von Willebrand factor 16. It is tempting to speculate that platelets may also differentially package pro-inflammatory and antiinflammatory molecules into distinct granule subpopulations. If so, it would indicate that platelets may store immunomodulatory substances in a specific manner in order to respond to different types of tissue damage.
Platelet-synthesized mediators. Platelets are not only warehouses that stockpile bioactive mediators, they also have the capacity to synthesize molecules. Recent evidence suggests that platelets contain a substantial amount of mRNA that is packaged during platelet formation from megakaryocytes. Furthermore, platelets contain all the translational machinery necessary to generate their own proteins during haemostatic and inflammatory events18–21. The notion of an anucleate cell being able to synthesize its own proteins is somewhat perplexing when one considers that the platelet already has all the preformed molecules necessary to mediate haemostasis. Nascent protein production by platelets may indicate that these cells have other as yet unappreciated functions in host defence. With the advent of platelet proteomic analyses, some of these hidden immune talents of platelets are being clarified. For example, proteomic analysis has demonstrated that platelets have the ability to secrete
266 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS /CTIKPCNOKETQVWDWNGDCPF
/KVQEJQPFTKQP
αITCPWNGU
1RGPECPKEWNCTU[UVGO
Figure 2 | The structure of a platelet. The figure shows the subcellular organization of a resting platelet viewed 0CVWTG4GXKGYU^+OOWPQNQI[ by thin-section electron microscopy. The marginal microtubule band encircles the cytoplasm of the platelet, maintaining its discoid shape. The α-granules constitute the majority of the storage granules, interspersed with dense granules, mitochondria, peroxisomes and lysosomes. The open caninicular system is formed by invaginations of the plasma membrane and is a complex network of interwinding membrane tubes that permeate the platelet’s cytoplasm.
more than 300 different proteins following activation with thrombin22, some of which (such as interleukin‑1 (IL‑1), TLRs and CD154 (also known as CD40L)) are clearly involved in processes other than blood clotting. Therefore, understanding the full range of platelet functions is becoming increasingly complex. The bioactive mediators and adhesive proteins expressed by activated platelets facilitate both homotypic interactions between platelets and heterotypic interactions between platelets and different immune cell populations. For example, activated platelets mediate the adhesion of neutrophils to the endothelium and also upregulate their pro-inflammatory functions. In addition, the interaction of platelets with dendritic cells (DCs) directs the DCs to sites of tissue injury and stimulates the release of inflammatory chemokines and cytokines by these DCs. Thus, platelets are pivotal mediators of cellular communication during inflammatory responses (FIG. 3). Furthermore, although platelets express relatively few integral membrane MHC molecules, they can adsorb soluble MHC class I molecules from the plasma, and this can have important clinical implications (BOX 2).
Tissue factor The major cellular initiator of clot formation. It is essential for generating thrombin from the zymogen prothrombin.
Immune mediators expressed by platelets Cytokines and chemokines. Many molecules stored in platelet granules, such as platelet-derived growth factor, ADP and thromboxane A2, are required for haemostatic functions. For example, platelet-derived growth factor has been shown to be instrumental in initiating and regulating wound healing at sites of inflammation23. However, platelet granules also contain various
pro-inflammatory and anti-inflammatory cytokines and chemokines that have no clear role in haemo stasis (TABLE 1). For example, platelets contain the largest amount of transforming growth factor-β (TGFβ) in the body 24. Although the role of this potent immuno suppressive factor in platelet-mediated haemostasis is unclear, circulating platelets appear to be important for regulating blood levels of TGFβ. Patients with immune thrombocytopenia have low levels of circulating TGFβ, but following therapy to restore normal platelet counts, their TGFβ levels recover 25,26. Of interest, several recent reports have shown that patients with immune thrombocytopenia have deficiencies in CD4+CD25+FOXP3+ regulatory T (TReg) cells. However, therapies that increase platelet counts (such as intravenous immunoglobulins, dexamethasone, rituximab (Rituxan/Mabthera; Genentech/Roche/Biogen Idec) or thrombopoietin) restore TReg cell numbers and functions in these individuals27–31. Whether this is related to changes in TGFβ levels is still unclear, but given that TReg cell differentiation is dependent on TGFβ, it is possible that platelet-stored TGFβ contributes to TReg cell homeostasis. The functions of the other cytokines and chemokines stored by platelets are not yet understood but this is an active area of research CD40 and CD154. CD40 and CD154 are a receptor– ligand pair with a central role in promoting interactions between lymphocytes and antigen-presenting cells, and they have been proposed to have important functions during thrombotic disease32. CD154 is a transmembrane protein that is expressed primarily by CD4+ T cells. However, in 1998, CD154 was also shown to be expressed by activated platelets33. It is now known that platelet-expressed CD154 can interact with CD40 on endothelial cells to induce endothelial cell upregulation of intercellular adhesion molecule 1 (ICAM1) and vascular cell adhesion molecule 1 (VCAM1) and release of CC‑chemokine ligand 2 (CCL2), thereby promoting leukocyte recruitment to inflammatory sites34. In addition, activated platelets release soluble CD154, which can interact with vascular cells (including endothelial cells) and induce the upregulation of E‑selectin and P‑selectin and the release of IL‑6 and tissue factor35,36. In fact, most of the circulating soluble CD154 in human plasma is generated from activated platelets, and levels of soluble CD154 may be an indicator of the degree of platelet activation within the host 37. The finding that CD154 is expressed by platelets has opened an interesting new avenue of research concerning how platelets can affect adaptive immunity. For example, it has recently been recognized that in some instances, platelet-derived CD154 can support B cell differentiation and immunoglobulin class-switching 38,39. During viral infections, transfer of wild-type platelets to CD154‑deficient mice led to increased production of virus-specific IgG. Although this increase in IgG levels was transient, the animals were protected from subsequent viral infection, suggesting that platelet-derived CD154 has the ability to promote protective adaptive immune responses40.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 267 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS C
2NCVGNGV
$CEVGTKC
D
6.4
2NCVGNGV
#EVKXCVGFRNCVGNGVU RTQFWEGE[VQMKPGU CPFEJGOQMKPGUCPF RTQOQVGPGWVTQRJKN CEVKXCVKQP
E
#EVKXCVGFRNCVGNGVURTQOQVG &%CEVKXCVKQPYJKEJKPETGCUGU CPVKIGPRTGUGPVCVKQPVQ6EGNNU
%& %& &% 0GWVTQRJKN 6JTQODQEKFKPU 2NCVGNGVGZRTGUUGF UGNGEVKPURTQOQVG PGWVTQRJKNVGVJGTKPI
6TCRRKPI QHDCEVGTKC D[RNCVGNGVU
%[VQMKPGUCPF EJGOQMKPGU
6%4
5GNGEVKPNKICPF 5GNGEVKP (KDTKP
&COCIGF GPFQVJGNKWO
$CEVGTKCNFGUVTWEVKQP
/*%
6EGNN
#EVKXCVKQPQHCFCRVKXG KOOWPGTGURQPUGU
0GWVTQRJKNCEVKXCVKQP
Figure 3 | Three typical examples of how activated platelets can mediate cell–cell interactions and affect innate immune responses. a | Platelet Toll-like receptor (TLR) expression enables activated platelets to bind and capture bacteria. Subsequently, the platelets may directly kill the bacteria by producing thrombocidins or by aggregating around the bacteria and ‘trapping’ them for elimination by professional phagocytes. b | It is now clear0CVWTG4GXKGYU^+OOWPQNQI[ that platelets can also heterotypically interact with a wide variety of cells, including leukocytes. Activated platelets promote neutrophil tethering and activation through the expression of selectins, CD154 (also known as CD40L) and inflammatory cytokines and chemokines. c | Similarly, activated platelets can also promote the activation of monocytes and dendritic cells (DCs), particularly through CD40–CD154 interactions. This leads to increased antigen presentation to T cells and enhances adaptive immune responses. TCR, T cell receptor.
Platelet-derived CD154 was also found to augment CD8+ T cell responses, both in vitro and in vivo, and to promote protective T cell responses following infection with Listeria monocytogenes41,42. Therefore, it appears that CD154 expression by platelets may have an important role in linking innate and adaptive immune responses and promoting protective immunity. On the other hand, it has also been shown that platelet-expressed CD154 can interfere with DC maturation43. CD154 from either resting or activated platelets was found to reduce the production of the pro-inflammatory cytokines IL‑12p70 and tumour necrosis factor (TNF) by DCs and to augment DC-mediated production of the anti-inflammatory cytokine IL‑10 (REF. 43). This suggests that platelets can also influence adaptive immunity through their effects on DCs42. Taken together, these studies strongly suggest that platelet-expressed CD154 is involved in regulating adaptive immune responses. Platelet expression of TLRs. The TLRs are a family of pattern recognition receptors that are expressed by professional phagocytes, such as neutrophils, macrophages and DCs, and promote immune activation in response to conserved molecular motifs expressed by pathogens44,45. To date, TLRs have been detected on platelets from birds46, cattle47,48, mice and humans49–57. Both mouse and human platelets were initially shown to
express TLR1, TLR2 and TLR3, and it was suggested that platelet-expressed TLRs act as a bridge between platelets and inflammatory responses49,50. Subsequently, many investigators have confirmed the expression of TLR1–9 by both human and mouse platelets and have shown that some platelet-expressed TLRs are functional and can modulate sepsis-induced thrombocytopenia and TNF production in vivo51–57. Specifically, lipopolysaccharide (LPS)-induced thrombocytopenia is dependent on platelet expression of TLR4 (REFS 51,53), and recent evidence suggests that platelets can actively bind circulating bacteria and their products and present these to neutrophils and cells of the reticuloendothelial system53–57. Sepsis is associated with marked neutrophil activation, primarily in the pulmonary and hepatic microcirculation, and this can lead to multiple organ dysfunction syndrome (MODS)58. Bacteria activate neutrophils mainly through TLRs, and this causes the neutrophils to become more structurally rigid and trapped in the lungs59. Subsequent to neutrophil recruitment, platelet numbers increase within the lungs of septic hosts60. Related to this, Clark et al.57 described a novel TLR4‑mediated platelet–neutrophil interaction that leads to bacterial ‘trapping’. Trapping was dependent on TLR4+ platelets binding LPS and adhering to neutrophils, resulting in neutrophil activation61. Plateletbound LPS stimulated neutrophil degranulation and
268 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Box 2 | Platelets and MHC molecules Like most non-professional antigen-presenting cells (APCs), platelets are devoid of MHC class II molecules. However, they do contain significant levels of MHC class I molecules on their surface. Most of these molecules are MHC class I heavy chains that have been cleaved from other cells and adsorbed by the platelets from the plasma, and it has been estimated that 80,000–120,000 MHC class I heavy chains are associated with each platelet101. Not only can these denatured MHC class I molecules on platelets be recognized by MHC-specific alloantibodies, they can also be shed from the platelet surface during storage in blood banks, although the clinical consequences of a transfusion of &GPCVWTGF soluble MHC class I molecules is unknown. At the /*%ENCUU+ *.#OQNGEWNGUKPFWEG OQNGEWNGU CNNQCPVKDQF[TGURQPUG recipient T cell level, allogeneic platelet MHC class I DWVPQ%6.TGURQPUG molecules appear to be incapable of inducing cytotoxic T lymphocyte (CTL) responses in vitro102. This lack of CTL induction in vitro correlates with the in vivo observations that allogeneic platelet transfusions can )2++Ds+++C mediate the so-called ‘transfusion effect’, in which they 2NCVGNGV prolong the survival of donor-matched skin allografts103. These immunomodulatory effects appear to be unique to the denatured platelet MHC class I molecules, as *WOCPRNCVGNGVCPVKIGPU other intact integral platelet membrane proteins, such KPFWEGCNNQCPVKDQF[CPF as platelet glycoprotein IIIa (GPIIIa; also known as β3 %6.TGURQPUGU integrin), can stimulate productive CTL responses104 )2+Ds+: (see the figure). 0CVWTG4GXKGYU^+OOWPQNQI[
Neutrophil extracellular traps Networks of extracellular fibres, produced by neutrophils, that catch and help kill pathogens. The extracellular fibril matrix is composed of decondensed chromatin.
the release of DNA, which formed structures resembling neutrophil extracellular traps (NETs)62 that extended downstream from the platelet–neutrophil aggregates (FIG. 4). In vivo imaging showed that the liver sinusoids and lung capillaries were the primary sites of NET formation and bacterial trapping 61. However, although the formation of NETs may have a beneficial effect for the host in isolating and preventing the spread of invading bacteria, platelet-induced activation of neutrophils can also promote injury to the host. For example, activated neutrophils produce reactive oxygen species that damage the pulmonary endothelium. There is also evidence to suggest that thrombocytes from lower vertebrates can be directly activated by LPS through TLR4, in vitro and in vivo. For example, chicken thrombocytes express TLR4 and upregulate mRNAs for IL‑6 and cyclooxygenase 2 (COX2; also known as PGHS2) following treatment with LPS46. This upregulation of mRNA expression was dependent on the activation of mitogen-activated protein kinase and nuclear factor-κB (NF-κB) signalling pathways in the thrombo cytes, suggesting that, at least in birds, thrombocytes contribute to innate immunity by upregulating cytokines and other immunomodulators in response to microbial products46. In addition, thrombocytopenic mice produce less TNF in response to LPS, but transfusion of TLR4‑expressing platelets enhanced TNF production following treatment with LPS51,53. These results suggest that platelets can respond to infectious agents, and this has fuelled speculation that these anucleate cells behave as circulating sentinels that have the ability to quickly alert immune cells, such as neutrophils, following host infection53–57. There are several other unique aspects of platelet TLR expression that deserve a mention. For example, platelets express all the components of the LPS receptor signalling complex (including TLR4, CD14, MD2 (also known as LY96) and myeloid differentiation primary
response protein 88 (MYD88)), and the ability of LPS to activate platelets is abolished in MYD88‑deficient mice or following treatment with TLR4‑specific blocking antibodies57. Furthermore, human platelets can bind to the O157:H7 type of LPS expressed by enterohaemorrhagic Escherichia coli (EHEC) through TLR4. Such binding of LPS was detected on platelets from 12 out of 14 children with EHEC-associated haemolytic uraemic syndrome (HUS) but not on platelets derived from EHEC-infected children who did not develop HUS, suggesting that platelet–LPS interactions may contribute to platelet destruction during HUS56. These findings indicate that platelets have the ability to bind bacterial molecules and that this type of interaction may be involved in activating the innate immune system during sepsis. Therefore, it is important to understand how platelets use TLRs to interact with neutrophils and other leukocytes, as this process may be involved in promoting immunopathological disease. Platelets and TREM1. Another way in which platelets may promote innate immunity is through interaction with triggering receptor expressed on myeloid cells 1 (TREM1). TREM1 belongs to the V‑type immunoglobulin superfamily and is primarily expressed by neutrophils and monocytes63. Expression of TREM1 increases after stimulation with microbial products and, in synergy with TLR ligands, stimulation of TREM1 activates several granulocyte effector responses, including the respiratory burst, phagocytosis and IL‑8 release. Although the identity of the natural TREM1 ligand is unknown, Haselmayer et al.64 recently demonstrated that a ligand for TREM1 is expressed by human platelets and showed that recombinant soluble TREM1 binds specifically to platelets. It was also shown that when neutrophils and platelets interact in the presence of LPS, TREM1 activation increases neutrophil-mediated production of reactive oxygen species and secretion of IL‑8 (REF 64).
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 269 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Table 1 | Platelet-associated molecules with immune functions Molecule
Location within platelet and mechanism of release
Activation state
Functions
Cellular targets
Histamine
Synthesized, location unknown; released during platelet activation
Preformed
Promotes allergic-type hypersensitivities
Endothelial cells, monocytes, PMNs, NK cells, T cells, B cells, eosinophils
Serotonin (also known as 5-HT)
Dense granules; released during platelet activation
Preformed
Actions on CNS; promotes coagulation and T cell activation
Monocytes, macrophages, platelets
Thromboxane A2
Plasma membrane
Synthesized
Promotes inflammation and Platelets, macrophages, coagulation T cells
PAF
Plasma membrane
Synthesized
Promotes inflammation
Platelets, PMNs, monocytes, macrophages
PDGF
α-granules; released during platelet activation
Preformed
Promotes wound healing
Monocytes, macrophages, T cells
TGFβ
α-granules; released during platelet activation
Preformed
Growth inhibition; immunosuppression
Monocytes, macrophages, T cells, B cells
CXCL7 (also known as NAP2)
Not known
Cleavage product of preformed precursor
Chemokine
PMNs
CXCL4 (also known as PF4)
α-granules; released during platelet activation
Preformed
Chemokine
PMNs, platelets
CXCL1 (also known as GROα)
α-granules; released during platelet activation
Preformed
Chemokine
PMNs
CXCL5 (also known as ENA78)
α-granules; released during platelet activation
Preformed
Chemokine
PMNs
CCL5 (also known as RANTES)
α-granules; released during platelet activation
Preformed
Chemokine
Monocytes, eosinophils, basophils, NK cells, T cells, DCs, platelets
CCL3 (also known as MIP1α)
α-granules; released during platelet activation
Preformed
Chemokine
Monocytes, eosinophils, basophils, NK cells, DCs
Preformed
Chemokine
Monocytes, basophils, NK cells, DCs
CCL7 (also known as MCP3) Not known IL‑1β (and IL-1 precursor protein)
Not known
Synthesized
T cell activation; multiple effects
Monocytes, macrophages, DCs, T cells
HMGB1
Not known
Preformed
Inflammatory gene regulation
Macrophages, PMNs, endothelial cells
Thrombocidins 1 and 2
α-granules; released during platelet activation
Preformed
Antibacterial peptides
Bacterial cells
CD40
Plasma membrane
Preformed
Co-stimulation; endothelial interactions
T cells
CD154
α-granules and plasma membrane; released and cleaved
Preformed
Co-stimulation; endothelial interactions
B cells, DCs, macrophages, monocytes, endothelial cells
TLR1, TLR2, TLR3, TLR5, TLR6, TLR7
Plasma membrane
Preformed
Pathogen detection
PMNs, DCs, macrophages, monocytes, platelets
TLR4
Plasma membrane
Preformed
Pathogen detection
PMNs, DCs, macrophages, monocytes
TLR9
Unknown internal site; upregulated on plasma membrane during platelet activation
Preformed
Pathogen detection
PMNs, DCs, macrophages, monocytes
TREM1 ligand
Plasma membrane; expression increased during platelet activation
Preformed
Pathogen detection
PMNs, DCs, macrophages, monocytes
CCL, CC-chemokine ligand; CXCL, CXC-chemokine ligand; DC, dendritic cell; HMGB1, high mobility group protein B1; IL‑1β, interleukin‑1β; NK, natural killer; PAF, platelet-activating factor; PDGF, platelet-derived growth factor; PMN, polymorphonuclear leukocyte; TGFβ, transforming growth factor-β; TLR, Toll-like receptor; TREM1, triggering receptor expressed on myeloid cells 1. The table is adapted and expanded from REF. 105.
270 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS .25RTQOQVGUPGWVTQRJKNCFJGUKQP CPFGPFQVJGNKCNEGNNCEVKXCVKQP
)TCOPGICVKXG DCEVGTKC .25 .25
#EVKXCVGFPGWVTQRJKNU IGPGTCVG415VJCV FCOCIGGPFQVJGNKWO
#EVKXCVGF PGWVTQRJKN
6.4
2NCVGNGV 64'/
415 6.4GZRTGUUKPIRNCVGNGVU RTGUGPV.25VQPGWVTQRJKNU CPFRTQOQVGVJGKTCEVKXCVKQP
0GWVTQRJKN 2NCVGNGVURTQOQVGPGWVTQRJKN HQTOCVKQPQH0'6UYJKEJ VTCRHTGGDCEVGTKC 'PFQVJGNKCNEGNN
Figure 4 | TLR4+ platelets mediate pulmonary neutrophil activation during sepsis. Lipopolysaccharide (LPS) from Gram-negative bacteria partially activates the pulmonary 0CVWTG4GXKGYU^+OOWPQNQI[ endothelium, causing it to upregulate the expression of adhesion molecules. This allows neutrophils to adhere to the endothelium and to become activated. Toll-like receptor 4 (TLR4)-expressing platelets present bound LPS to adherent neutrophils and stimulate them to generate reactive oxygen species (ROS) that damage the pulmonary endothelium, and this causes capillary leakage. The platelet-mediated presentation of LPS also directs the neutrophils to extrude their DNA and form neutrophil extracellular traps (NETs) that stretch downstream and can trap free bacteria. TREM1, triggering receptor expressed on myeloid cells 1.
Perhaps related to this, it was also recently demonstrated that TREM-like transcript 1 (TLT1), an orphan receptor only expressed in the α-granules of platelets and megakaryocytes, was markedly increased in the plasma of patients with sepsis and correlated significantly with disseminated intravascular coagulation in these patients65. Collectively, these observations suggest that TREM1 ligands and the TREM-like family may promote the interaction of neutrophils and platelets during inflammatory responses and that this receptor system could be targeted to suppress potentially harmful platelet-induced inflammatory responses during sepsis.
Platelets in immunity and immunopathology Platelet functions during microbial infection. Antibacterial proteins, such as defensins, are important effector molecules of the innate immune system that have been detected in many species79. The vast majority of these proteins are of a cationic nature and this is thought to be crucial in allowing them to bind and disrupt bacterial membranes79. Interestingly, platelets store such antibacterial proteins within their α-granules; collectively, these molecules are referred to as thrombocidins80. Several of the thrombocidins have been characterized and it appears that at least some of them, such as thrombocidin 1 and thrombocidin 2, are related to the CXC-chemokine family 81. Both thrombocidin 1 and
thrombocidin 2 are lethal to a wide range of bacterial species, including Bacillus subtilis, E. coli, Staphylococcus aureus and Lactococcus lactis 81. Platelet-derived thrombocidins have also been shown to kill fungi, such as Cryptococcus neoformans 81. Although microbial killing by these platelet-derived molecules is rapid, it does not appear to be dependent on pore formation81. The fact that platelets contain such antibacterial proteins that can be released upon activation strengthens the notion that platelets may have direct roles in protecting the host from infection. It has recently been suggested that human platelets can drive pathological events during infection with Plasmodium spp. parasites (the causative agents of malaria) by promoting sequestration of infected red blood cells in the cerebral vasculature82. However, another study demonstrated that human platelets can directly kill cultured red blood cells that are infected with Plasmodium falciparum parasites80, and this killing was shown to be inhibited by platelet inhibitors, such as aspirin83. Furthermore, red blood cells from thrombocytopenic mice were found to be highly susceptible to infection with Plasmodium chabaudi 83. It appears that platelets can bind to infected red cells and induce red blood cell destruction and intracellular parasite killing. These results suggest that platelets can have protective effects during erythrocytic malarial infections that are distinct from their pathological roles in cerebral malaria. With regard to their functions in responses to viruses, platelets have been shown to promote cytotoxic T lymphocyte (CTL)-mediated antiviral immune responses during infection with hepatitis B virus42. It appears that platelet activation is necessary for the accumulation of virus-specific CTLs at the site of hepatic inflammation; however, this also promotes immunopathology in the liver 42. Taken together, the data suggest that platelets can promote pathogen killing both directly and indirectly, enabling host protection against the offending pathogen. However, such activity by platelets can also drive detrimental immunopathological responses during certain infections. Atherosclerosis. The idea that platelets can have pathological pro-inflammatory functions initially developed from studies of atherosclerosis development 66. Inflammatory responses can substantially alter the procoagulant and anticoagulant properties of the endothelium, and this leads to enhanced communication between leukocytes, endothelial cells and platelets. These interactions are primarily mediated by plateletexpressed P‑selectin, and mice that lack P‑selectin expression have significant delays in atherosclerotic lesion formation compared with control animals in two mouse models of atherosclerosis (low-density lipoprotein receptor (LDLR)-deficient or apolipoprotein E (APOE)-deficient mice)67,68. In addition, when platelets adhere to inflamed atherosclerotic endothelium and become activated they secrete pro-inflammatory mediators, including CD154, IL‑1β and CCL5 (also known as RANTES), which promote activation of the endothelium and enhance monocyte recruitment 69.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 271 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS It is now well established that platelets directly contribute to the progression of atherosclerosis70,71. Langer et al.72 demonstrated that platelets promote the adhesion of DCs to injured carotid arteries in mice, and experiments in vitro confirmed that the recruitment of DCs by platelets depended on DC expression of αMβ2 integrin (also known as CD11b–CD18 and MAC1) and platelet expression of junctional adhesion molecule C (JAMC); this interaction led to DC activation and phagocytosis of platelets72. These types of plateletdependent processes not only show that platelets can promote the recruitment of immune cells such as DCs, but also show that they can do so in a manner that may enhance the progression of atherosclerotic lesions. Recently, it was shown that activated platelets expressing CD154 may promote atherogenesis and inflammation by inhibiting the recruitment of TReg cells (FIG 5). By transfusing platelets from CD154‑deficient mice into APOE-deficient recipients, Lievens et al.73 elegantly showed that expression of CD154 by platelets accelerates both the initial and advanced stages of atherosclerosis73. The enhanced atherosclerosis response was associated with a disruption of T cell homeostasis and deficient recruitment of TReg cells into atherogenic
'PFQVJGNKCN EGNN 64GIEGNN
2NCVGNGVUKPJKDKV 64GIEGNNTGETWKVOGPV
2NCVGNGV 2NCVGNGVUCEVKXCVG GPFQVJGNKCNEGNNUCPF RTQOQVGKPȯCOOCVKQP
6)(βHTQO64GIEGNNURTQOQVGU CPVKKPȯCOOCVQT[TGURQPUGU VJCVUVCDKNK\GVJGVJTQODWU
'ODQNK 64GIEGNN 4WRVWTG
6)(β
2NCSWG
r+PETGCUGFVJTQODWUUVCDKNKV[ r4GFWEGFVJTQODWUHTCIKNKV[
r+PETGCUGFVJTQODWUUK\G r+PETGCUGFVJTQODWUHTCIKNKV[
5OQQVJOWUENGEGNNU
#EVKXCVKQP
Figure 5 | Platelet-derived CD154 has an integral role in atherogenic plaque production. Activated platelets express and secrete CD154 (also known as CD40L), which can activate the endothelium and modulate a variety0CVWTG4GXKGYU^+OOWPQNQI[ of cellular processes. This culminates in a pro-inflammatory response that increases thrombus formation but reduces the stability of the thrombus, leading to rupture and the production of emboli. One mechanism by which platelet-derived CD154 can promote atherogenesis is by inhibiting the migration of regulatory T (TReg) cells to the site of the thrombus. This prevents the establishment of an anti-inflammatory milieu that would normally stabilize the thrombus. TGFβ, transforming growth factor-β.
plaques73. Overall, these findings indicate that platelets can, through various receptor–ligand interactions, promote inflammatory responses that contribute to atherosclerosis progression. Sepsis. Thrombocytopenia is common in critically ill patients with sepsis and is associated with increased mortality. During sepsis, both leukocytes and platelets become activated, and this contributes to the initiation of disseminated intravascular coagulation and eventual multisystem organ failure, as blood flow and oxygen delivery are decreased and activation of inflammatory cytokine networks is enhanced74. Furthermore, increased P‑selectin expression by activated platelets in patients with sepsis is associated with increased formation of platelet microparticles, which express surface receptors that enable them to interact with leukocytes75,76. When activated, platelets release membrane-derived microparticles that are approximately 0.1 to 1.0 μm in diameter; in humans, these microparticles express various molecules that are also found on platelets, such as P‑selectin and platelet glycoprotein IIb–IIIa (also known as αIIbβ3 integrin) 77. Platelet-derived microp articles readily adhere to various cell types and can activate endothelial cells, leukocytes and other platelets. Perhaps more importantly, these microparticles can act as messengers and deliver signals through soluble mediators, such as CCL5 (REF 78). Rheumatoid arthritis. Rheumatoid arthritis is a chronic inflammatory and autoimmune disorder that typically affects the synovial joints of the hands and feet, leading to painful swelling and bone erosion84. Both the innate and adaptive arms of the immune system contribute to the development of the disease, and early studies suggested that platelets and platelet-derived microparticles can accumulate in the joints of patients with rheumatoid arthritis85, although it was not known whether this contributed to the disease. Boilard et al.86 recently demonstrated that platelets may be essential for the progression of inflammatory arthritis. Using a combination of pharmacological and genetic methods this study identified platelet glyco protein VI (the predominant collagen receptor expressed by platelets) and the Fc receptor common γ-chain as major activators for platelet microparticle production in patients with rheumatoid arthritis86. Microparticle shedding appeared to be stimulated by fibroblast-like cells that line the joint cavity. In support of these observations, clinical studies have demonstrated that synovial fluids from patients with rheumatoid arthritis contain platelet microparticles, whereas fluids from patients with osteoarthritis do not. Platelet microparticles derived from the joint fluid of patients with rheumatoid arthritis could activate fibroblast-like synoviocytes to release the inflammatory cytokine IL‑1. Microparticle-associated IL‑1 appears to function as a key player in intensifying the inflammation associated with rheumatoid arthritis. For example, IL‑1 can induce joint synoviocytes to produce other pro-inflammatory cytokines and chemokines (such as
272 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS IL‑8), which promote neutrophil recruitment and activation within the joint 86. Taken together, these results suggest that platelets and their microparticles may have an important role in promoting the joint pathology observed in patients with inflammatory arthritides, such as rheumatoid arthritis. Transfusion-related acute lung injury (TRALI). TRALI is defined as a non-cardiogenic pulmonary oedema disorder that typically occurs within 6 hours following a transfusion of plasma-containing blood products87. The significance of TRALI in transfusion medicine has burgeoned recently, as it has been ranked as one of the leading causes of transfusion-related fatalities87. Studies of the roles of platelets have suggested that these cells may be involved in the pathogenesis of TRALI. For example, ligation of TLRs on platelets can lead to the release of CD154 (REF 88), and Khan et al.89 have suggested that soluble CD154 present in platelet concentrates stored for transfusion purposes may promote activation of pulmonary neutrophils through CD40 engagement, thereby
1.
2. 3. 4. 5. 6. 7. 8.
9.
10. 11. 12. 13. 14.
15.
16.
17.
Bizzozero, G. Sur un nouvel èlèment morphologique du sang chez les mammiferes et son importance dans la thrombose et dans la coagulation. Arch. Ital. Biol. 1, 1–5 (1882). Gawaz, M. et al. Platelets in inflammation and atherogenesis. J. Clin. Invest. 115, 3378–3384 (2005). Nieswandt, B. et al. Integrins in platelet activation. J. Thromb. Haemost. 7 (Suppl. 1), 206–209 (2009). Smyth, S. S. et al. Platelet functions beyond hemostasis. J. Thromb. Haemost. 7, 1759–1766 (2009). Kaushansky, K. Lineage-specific hematopoietic growth factors. N. Engl. J. Med. 354, 2034–2045 (2006). Italiano, J. E. Jr & Hartwig, J. H. in Platelets 2nd edn (eds Michelson, A. D. & Coller, B. S.) 23–44 (Elsevier, Amsterdam, 2007). Levin, J. in Platelets 2nd edn (eds Michelson, A. D. & Coller, B. S.) 3–22 (Elsevier, Amsterdam, 2007). Hose, J. E., Martin, G. G. & Gerard, A. S. A decapod hemocyte classification scheme integrating morphology, cytochemistry, and function. Biol. Bull. 178, 33–45 (1990). Götz, P. & Boman, H. G. in Comprehensive Insect Physiology, Biochemistry and Pharmacology (eds Kerkut, G. A. & Gilbert, L. I.) 453–485 (Pergamon, Oxford, 1985). Thon, J. N. & Italiano, J. E. Jr. Platelet formation. Semin. Hematol. 47, 220–226 (2010). Patel, S. R. et al. Differential roles of microtubule assembly and sliding in proplatelet formation by megakaryocytes. Blood. 106, 4076–4085 (2005). Thon, J. N. et al. Cytoskeletal mechanics of proplatelet maturation and platelet release. J. Cell Biol. 191, 861–874 (2010). Hartwig, J. Mechanisms of actin rearrangements mediating platelet activation. J. Cell Biol. 118, 1421–1442 (1992). Diacovo, T. G., Puri, K. D., Warnock, R. A., Springer, T. A. & von Andrian, U. H. Platelet-mediated lymphocyte delivery to high endothelial venules. Science 273, 252–255 (1996). Italiano, J. E. Jr et al. Angiogenesis is regulated by a novel mechanism: pro- and antiangiogenic proteins are organized into separate platelet α granules and differentially released. Blood. 111, 1227–1233 (2008). Sehgal, S. & Storrie, B. Evidence that differential packaging of the major platelet granule proteins von Willebrand factor and fibrinogen can support their differential release. J. Thromb. Haemost. 5, 2009–2016 (2007). White, G. C. & Rompietti, R. Platelet secretion: indiscriminately spewed forth or highly orchestrated? J. Thromb. Haemost. 5, 2006–2008 (2007).
contributing to TRALI. Future study of the roles of platelets in this disorder will be important for improving the efficacy and safety of blood transfusions.
Concluding remarks Platelets are the main mediators of haemostasis and thrombin generation, but it is becoming apparent that, like leukocytes, they have multiple functions in innate and adaptive immunity. It appears that these cellular fragments express and secrete many pro-inflammatory molecules that serve to initiate and modulate immune responses. Exciting recent studies on platelet-expressed TLRs and their functions have led to a new understanding of the role of platelets in infectious processes, and other recent investigations have suggested that platelets may promote autoimmunity. Further research is still required, however, to clarify the exact roles that platelets have in modulating the immune system. Platelets have gone from being regarded as ‘red cell dust’ a mere fifty years ago to being viewed as integral players in inflammatory processes and immunity.
18. Wicki, A. N. et al. Isolation and characterization of human blood platelet mRNA and construction of a cDNA library in λgt11. Confirmation of the platelet derivation by identification of GPIb coding mRNA and cloning of a GPIb coding cDNA insert. Thromb. Haemost. 61, 448–453 (1989). 19. Shashkin, P. N., Brown, G. T., Ghosh, A., Marathe, G. K. & McIntyre, T. M. Lipopolysaccharide is a direct agonist for platelet RNA splicing. J. Immunol. 181, 3495–3502 (2008). 20. Lindemann, S. & Gawaz, M. The active platelet: translation and protein synthesis in an anucleate cell. Semin. Thromb. Hemost. 33, 144–150 (2007). 21. Denis, M. M. et al. Escaping the nuclear confines: signal-dependent pre-mRNA splicing in anucleate platelets. Cell 122, 379–391 (2005). 22. Coppinger, J. A. Moderation of the platelet releasate response by aspirin. Blood 109, 4786–4792 (2007). 23. Mazzucco, L., Borzini, P. & Gope, R. Platelet-derived factors involved in tissue repair — from signal to function. Transfus. Med. Rev. 24, 218–234 (2010). 24. Assoian, R. K., Komoriya, A., Meyers, C. A., Miller, D. M. & Sporn, M. B. Transforming growth factor‑β in human platelets: identification of a major storage site, purification, and characterization. J. Biol. Chem. 258, 7155–7160 (1983). 25. Andersson, P. O., Stockelberg, D., Jacobsson, S. & Wadenvik, H. A transforming growth factor-β1‑ mediated bystander immune suppression could be associated with remission of chronic idiopathic thrombocytopenic purpura. Ann. Hematol. 79, 507–513 (2000). 26. Andersson, P. O., Olsson, A. & Wadenvik, H. Reduced transforming growth factor-β1 production by mononuclear cells from patients with active chronic idiopathic thrombocytopenic purpura. Br. J. Haematol. 116, 862–867 (2002). 27. Ling, Y., Cao, X., Yu, Z. & Ruan, C. Circulating dendritic cells subsets and CD4+Foxp3+ regulatory T cells in adult patients with chronic ITP before and after treatment with high-dose dexamethasome. Eur. J. Haematol. 79, 310–316 (2007). 28. Liu, B. et al. Abnormality of CD4+CD25+ regulatory T cells in idiopathic thrombocytopenic purpura. Eur. J. Haematol. 78, 139–143 (2007). 29. Stasi, R. et al. Analysis of regulatory T cell changes in patients with idiopathic thrombocytopenic purpura receiving B‑cell depleting therapy with rituximab. Blood 112, 1147–1150 (2008). 30. Olsson, B., Ridell, B., Carlsson, L., Jacobsson, S. & Wadenvik, H. Recruitment of T cells into bone marrow of ITP patients possibly due to elevated expression of VLA‑4 and CX3CR1. Blood 12, 1078–1084 (2008).
NATURE REVIEWS | IMMUNOLOGY
31. Yu, J. et al. Defective circulating CD25 regulatory T cells in patients with chronic immune thrombocytopenic purpura. Blood 112, 1325–1328 (2008). 32. Grewal, I. S. & Flavell, R. A. CD40 and CD154 in cell-mediated immunity. Annu. Rev. Immunol. 16, 111–135 (1998). 33. Henn, V. et al. CD40 ligand on activated platelets triggers an inflammatory reaction of endothelial cells. Nature 391, 591–594 (1998). 34. Andre, P., Nannizzi-Alaimo, L., Prasad, S. K. & Phillips, D. R. Platelet-derived CD40L: the switchhitting player of cardiovascular disease. Circulation 106, 896–899 (2002). 35. Hammwöhner, M. et al. Platelet expression of CD40/ CD40 ligand and its relation to inflammatory markers and adhesion molecules in patients with atrial fibrillation. Exp. Biol. Med. 232, 581–589 (2007). 36. Anand, S. X., Viles-Gonzalez, J. F. & Badimon, J. J. Membrane-associated CD40L and sCD40L in atherothrombotic disease. Thromb. Haemost. 90, 377–384 (2003). 37. Henn, V., Steinbach, S., Büchner, K., Presek, P. & Kroczek, R. A. The inflammatory action of CD40 ligand (CD154) expressed on activated human platelets is temporally limited by coexpressed CD40. Blood 98, 1047–1054 (2001). 38. von Hundelshausen, P. & Weber, C. Platelets as immune cells: bridging inflammation and cardiovascular disease. Circ. Res. 100, 27–40 (2007). 39. Elzey, B. D. et al. Platelet-mediated modulation of adaptive immunity. A communication link between innate and adaptive immune compartments. Immunity 19, 9–19 (2003). 40. Sprague, D. L. et al. Platelet-mediated modulation of adaptive immunity: unique delivery of CD154 signal by platelet-derived membrane vesicles. Blood 111, 5028–5036 (2008). 41. Elzey, B. D. et al. Platelet-derived CD154 enables T‑cell priming and protection against Listeria monocytogenes challenge. Blood 111, 3684–3691 (2008). 42. Iannacone, M. et al. Platelets mediate cytotoxic T lymphocyte-induced liver damage. Nature Med. 11, 1167–1169 (2005). 43. Kissel, K. et al. Human platelets target dendritic cell differentiation and production of proinflammatory cytokines. Transfusion 46, 818–827 (2006). 44. Janeway, C. A. Jr. The immune system evolved to discriminate infectious non-self from non-infectious self. Immunol. Today 13, 11–15 (1992). 45. Janeway, C. A. Jr & Medzhitov, R. Innate immune recognition. Annu. Rev. Immunol. 20, 197–216 (2002).
VOLUME 11 | APRIL 2011 | 273 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS 46. Scott, T. & Owens, M. D. Thrombocytes respond to lipopolysaccharide through Toll-like receptor‑4, and MAP kinase and NF‑κB pathways leading to expression of interleukin‑6 and cyclooxygenase‑2 with production of prostaglandin E2. Mol. Immunol. 45, 1001–1008 (2008). 47. Kuckleburg, C. J., Tiwari, R. & Czuprynski, C. J. Endothelial cell apoptosis induced by bacteriaactivated platelets requires caspase‑8 and ‑9 and generation of reactive oxygen species. Thromb. Haemost. 99, 363–372 (2008). 48. Kuckleburg, C. J., McClenahan, D. J. & Czuprynski, C. J. Platelet activation by Histophilus somni and its lipooligosaccharide induces endothelial cell proinflammatory responses and platelet internalization. Shock 29, 189–196 (2008). 49. Aslam, R., Freedman, J. & Semple, J. W. Murine platelets express Toll-like receptor 2: a potential regulator of innate and adaptive immunity. Platelets 15, 267–269 (2004). 50. Shiraki, R. et al. Expression of Toll-like receptors on human platelets. Thromb. Res. 113, 379–385 (2004). 51. Andonegui, G. et al. Platelets express functional Toll-like receptor‑4. Blood 106, 2417–2423 (2005). 52. Cognasse, F. et al. Evidence of Toll-like receptor molecules on human platelets. Immunol. Cell Biol. 88, 196–198 (2005). 53. Aslam, R. et al. Platelet Toll-like receptor expression modulates lipopolysaccharide-induced thrombocytopenia and tumor necrosis factor‑α production in vivo. Blood 107, 637–641 (2006). 54. Semple, J. W., Aslam, R., Kim, M., Speck, E. R. & Freedman, J. Platelet-bound lipopolysaccharide enhances Fc receptor-mediated phagocytosis of IgG opsonized platelets. Blood 109, 4803–4805 (2007). 55. Patrignani, P. et al. Reduced thromboxane biosynthesis in carriers of Toll-like receptor 4 polymorphisms in vivo. Blood 107, 3572–3574 (2006). 56. Ståhl, A. L. et al. Lipopolysaccharide from enterohemorrhagic Escherichia coli binds to platelets via TLR4 and CD62 and is detected on circulating platelets in patients with hemolytic uremic syndrome. Blood 108, 167–176 (2006). 57. Zhang, G. et al. Lipopolysaccharide stimulates platelet secretion and potentiates platelet aggregation via TLR4/MyD88 and the cGMP-dependent protein kinase pathway. J. Immunol. 182, 7997–8004 (2009). 58. Welbourn, C. R. & Young, Y. Endotoxin, septic shock and acute lung injury: neutrophils, macrophages and inflammatory mediators. Br. J. Surg. 79, 998–1003 (1992). 59. McClenahan, D. J., Evanson, O. A., Walcheck, B. K. & Weiss, D. J. Association among filamentous actin content, CD11b expression, and membrane deformability in stimulated and unstimulated bovine neutrophils. Amer. J. Vet. Res. 61, 380–386 (2000). 60. Ma, A. C. & Kubes, P. Platelets, neutrophils, and neutrophil extracellular traps (NETs) in sepsis. J. Thromb. Haemost. 6, 415–420 (2008). 61. Clark, S. R. et al. Platelet TLR4 activates neutrophil extracellular traps to ensnare bacteria in septic blood. Nature Med. 13, 463–469 (2007). This elegant study provided evidence that platelets act as sensors of bacterial infection by demonstrating that activation of platelet TLR4 stimulates the binding of activated platelets to neutrophils, leading to the release of NETs to catch and kill bacteria. 62. Brinkmann, V. et al. Neutrophil extracellular traps kill bacteria. Science 303, 1532–1535 (2004). 63. Klesney-Tait, J., Turnbull, I. R. & Colonna, M. The TREM receptor family and signal integration. Nature Immunol. 7, 1266–1273 (2006). 64. Haselmayer, P., Grosse-Hovest, L., von Landenberg, P., Schild, H. & Radsak, M. P. TREM‑1 ligand expression on platelets enhances neutrophil activation. Blood 110, 1029–1035 (2007). In this article, the authors demonstrated that platelets express the ligand for TREM1, and showed that the binding of this molecule to neutrophil TREM1 augments LPS-induced neutrophil activation, and ultimately the innate inflammatory response.
65. Washington, A. V. et al. TREM-like transcript‑1 protects against inflammation-associated hemorrhage by facilitating platelet aggregation in mice and humans. J. Clin. Invest. 119, 1489–1501 (2009). 66. Ross, R. Atherosclerosis: an inflammatory disease. N. Engl. J. Med. 340, 115–126 (1999). 67. Dong, Z. M., Brown, A. A. & Wagner, D. D. Prominent role of P‑selectin in the development of advanced atherosclerosis in ApoE-deficient mice. Circulation 101, 2290–2295 (2000). 68. Burger, P. C. & Wagner, D. D. Platelet P‑selectin facilitates atherosclerotic lesion development. Blood 101, 2661–2666 (2003). 69. Lindemann, S. et al. Activated platelets mediate inflammatory signaling by regulated interleukin 1β synthesis. J. Cell Biol. 154, 485–490 (2001). 70. Langer, H. F. & Gawaz, M. P. Platelet-vessel wall interactions in atherosclerotic disease. Thromb. Haemost. 99, 480–486 (2008). 71. Davi, G. & Patrono, C. Platelet activation and atherothrombosis. N. Engl. J. Med. 357, 2482–2494 (2007). 72. Langer, H. F. et al. Platelets recruit human dendritic cells via Mac‑1/JAM‑C interaction and modulate dendritic cell function in vitro. Arterioscler. Thromb. Vasc. Biol. 27, 1463–1470 (2007). 73. Lievens, D. et al. Platelet CD40L mediates thrombotic and inflammatory processes in atherosclerosis. Blood 116, 4317–4327 (2010). 74. Langer, H. F., Baughman, R. P., Lower, E. E., Flessa, H. C. & Tollerud, D. J. Thrombocytopenia in the intensive care unit. Chest 104, 1243–1247 (1993). 75. Gawaz, M., Dickfeld, T., Bogner, C., Fateh-Moghadam, S. & Neumann, F. J. Platelet function in septic multiple organ dysfunction syndrome. Intensive Care Med. 23, 379–385 (1997). 76. Jacoby, R. C. et al. Platelet activation and function after trauma. J. Trauma 51, 639–647 (2001). 77. Ogura, H. et al. Activated platelets enhance microparticle formation and platelet–leukocyte interaction in severe trauma and sepsis. J. Trauma 50, 801–809 (2001). 78. Mause, S. F., von Hundelshausen, P., Zernecke, A., Koenen, R. R. & Weber, C. Platelet microparticles: a transcellular delivery system for RANTES promoting monocyte recruitment on endothelium. Arterioscler. Thromb. Vasc. Biol. 25, 1512–1518 (2005). 79. Boman, H. G. Peptide antibiotics and their role in innate immunity. Annu. Rev. Immunol. 13, 61–92 (1995). 80. Yeaman, M. R., Puentes, S. M., Norman, D. C. & Bayer, A. S. Partial characterization and staphylocidal activity of thrombin-induced platelet microbicidal protein. Infect. Immun. 60, 1202–1209 (1992). 81. Krijgsveld, J. et al. Thrombocidins, microbicidal proteins from human blood platelets, are C‑terminal deletion products of CXC chemokines. J. Biol. Chem. 275, 20374–20381 (2000). 82. Cox, D. & McConkey, S. The role of platelets in the pathogenesis of cerebral malaria. Cell. Mol. Life Sci. 67, 557–568 (2010). 83. McMorran, B. J. et al. Platelets kill intraerythrocytic malarial parasites and mediate survival to infection. Science 323, 797–800 (2009). This elegant study demonstrated that platelets can kill parasites in vitro and in mice, and found that platelet-deficient mice are more likely to die from malaria than mice with normal platelet counts. In addition, it showed that a single dose of aspirin may interfere with platelets sufficiently to prevent their killing power. 84. Choy, E. H. & Panayi, G. S. Cytokine pathways and joint inflammation in rheumatoid arthritis. N. Engl. J. Med. 344, 907–916 (2001). 85. Knijff-Dutmer, E. A., Koerts, J., Nieuwland, R., Kalsbeek-Batenburg, E. M. & van de Laar, M. A. Elevated levels of platelet microparticles are associated with disease activity in rheumatoid arthritis. Arthritis Rheum. 46, 1498–1503 (2002). 86. Boilard, E. et al. Platelets amplify inflammation in arthritis via collagen-dependent microparticle production. Science 327, 580–583 (2010). This interesting study described an essential role for platelet-derived microparticles in the development of the autoimmune disease rheumatoid arthritis, and identified the collagen receptor glycoprotein VI as a key stimulator of microparticle generation.
274 | APRIL 2011 | VOLUME 11
87. Looney, M. R., Gropper, M. A. & Matthay, M. A. Transfusion-related acute lung injury: a review. Chest 126, 249–258 (2004). 88. Cognasse, F., Lafarge, S., Chavarin, P., Acquart, S. & Garraud, O. Lipopolysaccharide induces sCD40L release through human platelets TLR4, but not TLR2 and TLR9. Intensive Care Med. 33, 382–384 (2007). 89. Khan, S. Y. et al. Soluble CD40 ligand accumulates in stored blood components, primes neutrophils through CD40, and is a potential cofactor in the development of transfusion-related acute lung injury. Blood 108, 2455–2462 (2006). 90. Cines, D. B., Bussel, J. B., Liebman, H. A. & Luning Prak, E. T. The ITP syndrome: pathogenic and clinical diversity. Blood 113, 6511–6521 (2009). 91. Semple, J. W., Provan, D., Garvey, M. B. & Freedman, J. Recent progress in understanding the pathogenesis of immune thrombocytopenia (ITP). Curr. Opin. Haematol. 17, 590–595 (2010). 92. Semple, J. W. Infections, antigen presenting cells, T cells and immune tolerance: their role in the pathogenesis of immune thrombocytopenia (ITP). Hematol. Oncol. Clin. North Am. 23, 1177–1192 (2009). 93. Harrington, W. J., Minnich, V., Hollingsworth, J. W. & Moore, C. V. Demonstration of a thrombocytopenic factor in the blood of patients with thrombocytopenic purpura. J. Lab. Clin. Med. 38, 1–10 (1951). 94. Evans, R. S., Takahashi, K., Duane, R. T., Payne, R. & Liu, C. Primary thrombocytopenic purpura and acquired hemolytic anemia; evidence for a common etiology. Arch. Intern. Med. 87, 48–65 (1951). 95. Nardi, M., Tomlinson, S., Greco, M. & Karpatkin, S. Complement-independent, peroxide-induced antibody lysis of platelets in HIV‑1‑related immune thrombocytopenia. Cell 106, 551–561 (2000). 96. Olsson, B. et al. T‑cell‑mediated cytotoxicity toward platelets in chronic idiopathic thrombocytopenic purpura. Nature Med. 9, 1123–1124 (2003). 97. Delaflor-Weiss, E. & Mintz, P. D. The evaluation and management of platelet refractoriness and alloimmunization. Transfus. Med. Rev. 14, 180–196 (2000). 98. Landau, M. & Rosenberg, N. Molecular insight into human platelet antigens: structural and evolutionary conservation analyses offer new perspective to immunogenic disorders. Transfusion 51, 558–569 (2011). 99. Kaplan C. Neonatal alloimmune thrombocytopenia: a 50‑year story. Immunohematology 23, 9–13 (2007). 100. Shtalrid, M. et al. Post-transfusion purpura: a challenging diagnosis. Isr. Med. Assoc. J. 8, 672–674 (2006). 101. Kao, K. J., Cook, D. J. & Scornik, J. C. Quantitative analysis of platelet surface HLA by W6/32 anti-HLA monoclonal antibody. Blood 68, 627–632 (1986). 102. Gouttefangeas, C. et al. Thrombocyte HLA molecules retain nonrenewable endogenous peptides of megakaryocyte lineage and do not stimulate direct allocytotoxicity in vitro. Blood 95, 3168–3175 (2000). 103. Aslam, R., Speck, E. R., Kim, M., Freedman, J. & Semple, J. W. Transfusion-related immunomodulation (TRIM) by platelets is dependent on their expression of MHC class I molecules and is independent of leukocytes. Transfusion 48, 1778–1786 (2008). 104. Chow, L. et al. A novel mouse model demonstrating both antibody- and T cell-mediated thrombocytopenia: differential response to therapy. Blood 115, 1247–1253 (2010). 105. Weyrich, A. S. & Zimmerman, G. A. Platelets: signaling cells in the immune continuum. Trends Immunol. 25, 489–495 (2004).
Acknowledgements
The authors thank A. H. Lazarus, H. Ni and V. Leytin for their helpful discussions and advice.
Competing interests statement
The authors declare no competing financial interests.
FURTHER INFORMATION John W. Semple’s homepage: http://www.angelfire.com/ut/johnsnotes/index.html ALL LINKS ARE ACTIVE IN THE ONLINE PDF
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS
Immunity to fungal infections Luigina Romani
Abstract | Fungal diseases represent an important paradigm in immunology, as they can result from either a lack of recognition by the immune system or overactivation of the inflammatory response. Research in this field is entering an exciting period of transition from studying the molecular and cellular bases of fungal virulence to determining the cellular and molecular mechanisms that maintain immune homeostasis with fungi. The fine line between these two research areas is central to our understanding of tissue homeostasis and its possible breakdown in fungal infections and diseases. Recent insights into immune responses to fungi suggest that functionally distinct mechanisms have evolved to achieve optimal host−fungus interactions in mammals.
Yeast A unicellular form of a fungus, consisting of oval or spherical cells, usually about 3 to 5 μm in diameter, that reproduce asexually by a process termed blastoconidia formation (budding) or by fission.
Spore An asexual or sexual reproductive element of a fungus.
Department of Experimental Medicine and Biochemical Sciences, Microbiology Section, University of Perugia, Via del Giochetto, 06122 Perugia, Italy. e-mail:
[email protected] doi:10.1038/nri2939 Published online 11 March 2011
Fungi are heterotrophic eukaryotes that are traditionally and morphologically classified into yeast and filamentous forms. The study of these eukaryotes has been motivated by their unique and fascinating biology, their many useful products (including wine, cheese and antibiotics), their use as experimental systems for basic biology and their importance as animal and plant pathogens. Most fungi are ubiquitous in the environment, and humans are exposed by inhaling spores or small yeast cells. Examples of common fungi include Aspergillus fumigatus, Cryptococcus neoformans and the thermally dimorphic fungi (Histoplasma capsulatum, Blastomyces dermatitidis, Paracoccidioides brasiliensis, Coccidioides immitis, Penicillium marneffei and Sporothrix schenckii) (BOX 1). Fungi are very proficient at sensing their surroundings and responding to cues that promote their survival in changing environments. As a result, they can interact with plants, animals or humans in multiple ways, establishing symbiotic, commensal, latent or pathogenic relationships. Their ability to colonize almost every niche within the human body involves specific reprogramming events that enable them to adapt to environmental conditions, fight for nutrient acquisition and deal with or even exploit ‘stresses’ generated by host defence mechanisms1–3. Genomic and transcriptome-based approaches have revealed a link between fungal metabolism, morpho genesis and the response to stress in adaptation to the host environment2. Such adaptations can enhance pathogen virulence but can also provide opportunities for potential therapeutic targets4. Fungi are associated with a wide spectrum of diseases in humans and animals, ranging from acute self-limiting pulmonary manifestations and cutaneous lesions in immunocompetent individuals to inflammatory diseases
and severe life-threatening infections in immunocompromised patients (BOX 1). As the population of immunosuppressed individuals has increased (secondarily to the increased prevalence of cancer, chemotherapy, organ transplantation and autoimmune diseases), so has the incidence of fungal diseases5,6. Furthermore, it has been anticipated that global warming will bring new fungal diseases for mammals7. Many fungal species (including Pneumocystis jiroveci 8 and commensal fungi, such as Malassezia spp. and Candida albicans) have co-evolved with their mammalian hosts over millions of years. This suggests the existence of complex mechanisms of immune surveillance in the host and of sophisticated fungal strategies to antagonize immunity. The immune system does not remain ignorant of commensal or ubiquitous fungi, and so a fine balance between pro- and anti-inflammatory signals is required to maintain a stable host–fungus relationship, the disruption of which can have pathological consequences (BOX 2). In this Review, I explain that the host immune response to fungi comprises two main components — resistance (the ability to limit fungal burden) and tolerance (the ability to limit the host damage caused by the immune response or other mechanisms). Both strategies are evolutionarily conserved in plants and vertebrates9, and understanding the interplay between them may allow us to define how fungi have adapted to the mammalian immune system and to translate this knowledge into new medical practices.
Recognition of fungi by the innate immune system PAMPs. Innate immune mechanisms are used by the host to respond to a range of fungal pathogens in a rapid and conserved manner. The constitutive mechanisms of innate defence are present at sites of continuous
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 275 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Box 1 | Major fungal pathogens, their habitats and associated diseases Respiratory allergy Asthma is common in the developed world and is increasing in frequency. Studies have linked worsening asthma with exposure to species of Aspergillus, Alternaria, Cladosporium and Penicillium. Mould sensitivity has been associated with increased asthma severity, and increased hospital and intensive care admissions and death in adults, and with increased bronchial reactivity in children. Skin diseases The skin can be a point of entry for fungal infections when the epithelial barrier is breached, or it can be a site for disseminated, systemic fungal diseases. Although Malassezia yeasts are a part of the normal microbiota, they have been associated with a number of diseases affecting the human skin, such as pityriasis versicolor, folliculitis, seborrhoeic dermatitis and dandruff, atopic dermatitis and psoriasis. Chronic mucocutaneous candidiasis, a primary immune deficiency presenting as an inability to clear yeasts, is an intractable manifestation of Candida albicans infection. Recurrent vulvovaginal candidiasis (VVC) VVC is a widespread mucosal infection, caused by saprophytic and opportunistic yeasts belonging to the Candida genus, that can affect up to 75% of women of child-bearing age. There are several predisposing factors, including antibiotic and oral contraceptive usage, hormone replacement therapy, pregnancy and uncontrolled diabetes mellitus. Despite therapeutic advances, VVC remains a common problem worldwide, with a high associated cost and a high concern for drug resistance. Inflammatory bowel disease (IBD) The gastrointestinal tract uses a system of tolerance and controlled inflammation to limit the response to dietary or pathogen-derived antigens in the gut. Mucosal homeostasis arises from a highly dynamic balance between host protective immunity and regulatory mechanisms. When this complex system breaks down in a genetically predisposed individual, the resulting immune response may lead to IBD. Antibodies against Saccharomyces cerevisiae are present in a subgroup of patients with Crohn’s disease and correlate with C. albicans colonization. These findings suggest that altered sensing of C. albicans colonization could contribute to aberrant immune responses in IBD. Invasive fungal diseases (IFDs) IFDs are nosocomial and device-related infections that occur in patients with haematological disorders or following solid organ or haematopoietic stem cell transplantation. It has been suggested that the agricultural use of fungicides may have contributed to drug resistance in people with life-threatening IFDs. IFDs have also been reported in patients who are not at high risk, such as patients with H1N1 influenza virus or Mycobacterium tuberculosis infection and those receiving tumour necrosis factor-targeted therapy.
Organism
Habitat
Aspergillus spp.
Soil; decaying organic materials; indoor air environments
• Aspergilloma • Acute and chronic pneumonias • Cerebral aspergillosis • Allergy, ABPA, SAFS
Pneumocystis spp.
No known environmental habitat; person-to-person transmission
• Pneumonia • COPD
Cryptococcus spp.
Environment, in association with decaying materials and trees
• Pneumonia • Meningitis • Disseminated disease
Candida spp.
Commensal of human gastrointestinal tract and vagina
• Disseminated infections • Mucocutaneous infections (oropharyngeal, skin and nail infections) • Vaginitis
Malassezia spp.
Commensal of human skin
• Cutaneous infections (pityriasis versicolor, seborrhoeic dermatitis) • Allergic atopic eczema
Blastomyces dermatitidis
Soil, in association with decaying wood
• Acute and chronic pneumonias • Skin lesions • Disseminated disease
Coccidioides immitis
Alkaline soil
• Self-limited influenza-like syndrome • Pneumonia • Disseminated disease
Histoplasma capsulatum
Soil contaminated with bird or bat guano
• Self-limited influenza-like syndrome • Acute and chronic pneumonias • Disseminated disease
Paracoccidioides Soil and digestive brasiliensis tract of some animals
Image*
Disease
• Asymptomatic • Acute and chronic pneumonias • Disseminate disease • Cutaneous lesions
ABPA, allergic bronchopulmonary aspergillosis; COPD, chronic obstructive pulmonary disease; SAFS, severe asthma with fungal sensitisation. *Images courtesy of www. doctorfungus.org © (2005).
interaction with fungi and include the barrier function of the skin and the mucosal epithelial cell surfaces of the respiratory, gastrointestinal and genitourinary tracts10. Microbial antagonism, defensins, collectins and the complement system also provide constitutive defence mechanisms and opsonic recognition of fungi. For example, complement receptor 3 (CR3; a heterodimer of CD11b and CD18) recognizes complement deposited on β‑(1,6)glucans on the fungus surface (FIG. 1). Moreover, host cells
express pattern recognition receptors (PRRs) — such as Toll-like receptors (TLRs), C‑type lectin receptors (CLRs) and the galectin family proteins11–13 — that sense pathogenassociated molecular patterns (PAMPs) in fungi (FIG. 1). PRRs on phagocytes initiate downstream intracellular events that promote the activation of the immune system and the clearance of fungi, with the specific immune response generated depending on the cell type involved. Monocytes, macrophages and neutrophils, as well as some
276 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Box 2 | Fungal infections and diseases: from immunity to immunopathology
Toll-like receptors (TLRs). A family of membrane-spanning proteins that recognize pathogen-associated molecular patterns (which are shared by various microorganisms), as well as damaged host cell components. TLRs signal to the host that a microbial pathogen is present or that tissue damage has occurred. They are characterized by an ectodomain that has varying numbers of leucine-rich repeat motifs and a cytoplasmic Toll/ IL‑1 receptor (TIR) domain that recruits adaptors, such as the myeloid differentiation primary response protein 88 (MYD88) and TIR domain-containing adaptor protein inducing IFNβ (TRIF; also known as TICAM1).
C-type lectin receptors (CLRs). A large family of proteins that have one or more carbohydrate-recognition domains. CLRs exist as transmembrane and soluble proteins, and include the mannose receptor, dectin 1, dectin 2 and DC‑SIGN, as well as soluble molecules, such as the complement-activating mannose-binding lectins, which are involved in antifungal immunity.
Inflammasome A large multiprotein complex that contains certain NOD-like receptors, RIG‑I-like receptors and IFI200 proteins, the adaptor protein apoptosisassociated speck-like protein containing a CARD (ASC; also known as PYCARD) and pro-caspase 1. Assembly of the inflammasome leads to the activation of caspase 1, which cleaves pro-interleukin‑1β (pro-IL‑1β) and pro-IL‑18 to generate the active cytokines.
The bipolar nature of the inflammatory process in fungal infections Bidirectional influences between infection and immune-related pathology have been known to exist in chronic mucocutaneous candidiasis (CMC) and chronic disseminated candidiasis (CDC). Although occasionally associated with autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy (a condition of dysfunctional T cell activity), CMC encompasses a variety of clinical disorders, in which the inability to clear Candida albicans yeasts results in yeast persistence in recurring lesions of the skin, nails and mucous membranes125. Patients with CMC often develop endocrine and inflammatory disorders, which suggests that immune responses are dysregulated. CDC is typically observed during neutrophil recovery in patients with acute leukaemia and requires protracted antifungal therapy. However, the efficacy of adjuvant corticosteroid therapy in these patients supports the pathophysiological hypothesis that CDC is a fungus-related immune reconstitution inflammatory syndrome (IRIS; see below)135. These observations highlight the truly bipolar nature of the inflammatory process in infection, at least during infection with specific fungi. Early inflammation is beneficial in containing the infection, but an uncontrolled inflammatory response is detrimental and may eventually oppose disease eradication. This condition is exemplified in mice with chronic granulomatous disease (CGD), in which an intrinsic, genetically determined failure to control inflammation to sterile fungal components determines the animals’ inability to resolve an infection with Aspergillus fumigatus136. One major implication of these findings is that, at least in specific clinical settings, a state of chronic or intractable fungal disease is the result of an exaggerated inflammatory response that probably compromises the host’s ability to cope with infecting fungi, and not of an ‘intrinsic’ susceptibility to infection. Thus, fungal diseases represent an important paradigm in immunology, as they can result from either a lack of recognition or an overactivation of the inflammatory response. The immune reconstitution inflammatory syndrome Clinically severe fungal infections occur in patients with IRIS, a disorder that is characterized by local and systemic inflammatory reactions that can result in quiescent or latent infections, which manifest as opportunistic mycoses137. IRIS responses are also found in otherwise immunocompetent individuals and are probably associated with disease severity in paracoccidioidomycosis, blastomycosis or Malassezia folliculitis. Thus, the conceptual principles of IRIS underscore the adverse effects of an overzealous and dysregulated immune response on the resolution of fungal infections and support a role for immunotherapies that are tailored to augment protective immunity.
cells that are normally non-phagocytic (such as epithelial and endothelial cells)14, mostly contribute to the antifungal innate immune response through phagocytosis and direct pathogen killing. By contrast, uptake of fungi by dendritic cells (DCs) induces DC maturation and this promotes the differentiation of naive T cells into effector T helper (TH) cell subtypes. To achieve optimal activation of antigen-specific adaptive immune responses, it is first necessary to activate the pathogen-detection mechanisms of the innate immune system. However, PRR activation is a double-edged sword, as PRRs might, paradoxically, promote some infections and cause tissue damage. Not surprisingly, therefore, fungi exploit PRRs to divert and subvert host immune responses in order to survive and eventually replicate (as discussed later). The fungal cell wall varies in composition depending on the morphotype, growth stage and environment of the fungal species, and is the main source of PAMPs that are recognized by PRRs on mammalian cells15. The three major cell wall components, found in all medically important fungi, are: β‑glucans (which are polymers of glucose), especially β‑(1,3)-glucans with varying numbers of β‑(1,6) branches; chitin (which is a polymer of N‑acetylglucosamine); and mannans (which are chains of several hundred mannose molecules that are added to fungal proteins via N‑ or O‑linkages). β‑(1,2)-linked oligomannosides are also PAMPs, and these molecules are recognized by galectin 3, which allows phagocytes to discriminate between pathogenic and non-pathogenic yeasts11 (FIG. 1). During the course of a fungal infection, multiple host PRRs are likely to be stimulated by fungal PAMPs in different combinations depending on the fungal species and on the host cell types. Therefore, the final
immune response will depend not only on the relative degree of stimulation of the individual receptors but also on the level of receptor cooperativity and the cellular localization. CLRs. CLRs are central for fungal recognition and for the induction of the innate and adaptive immune responses, and individuals with genetic deficiencies in CLRs are highly susceptible to fungal infections (TABLE 1). CLR family members include dectin 1 (also known as CLEC7A), dectin 2 (also known as CLEC6A), mincle (also known as CLEC4E), DC‑specific ICAM3‑grabbing non-integrin (DC-SIGN), the mannose receptor (also known as macrophage mannose receptor 1), langerin (also known as CLEC4K) and mannose-binding lectin16. Dectin 1 is the main PRR that recognizes β‑glucans and, following ligation, it induces the production of pro- and antiinflammatory cytokines and chemokines16 (FIG. 1). This is achieved through the activation of two distinct signalling pathways downstream of dectin 1, the spleen tyrosine kinase (SYK)–caspase recruitment domaincontaining protein 9 (CARD9) pathway and the RAF pathway. These pathways act synergistically and, through cross-regulatory mechanisms, induce and fine-tune canonical and non-canonical nuclear factor-κB (NF-κB) activation and cytokine gene expression17. The SYK–CARD9 pathway also activates the NLRP3 (NOD‑, LRR- and pyrin domain-containing 3) inflammasome, which results in proteolytic activation of the pro-inflammatory cytokines interleukin‑1β (IL‑1β) and IL‑18 by caspase 1. Both human (TABLE 1) and mouse studies show that genetic deficiencies of dectin 1 (REFS 18–20) and CARD9 (REFS 21,22) are associated with susceptibility to fungal infections.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 277 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS 2#/2U 1NKPMGFOCPPCP 40# &0# 6.4 6.4 6.4
&%5+)0 6.4 6.4 4#5
/;&
αINWECP %JKVKP
0NKPMGFOCPPCP βINWECP
64+(
/KPENG
&GEVKP
&GEVKP (E4γ
(E4γ
2 5;- 2 2
5;- 2
/CPPQUG TGEGRVQT
βINWECP
%DQT%F EQCVGFβINWECP $#&
βOCPPQUKFGU
%&
)CNGEVKP
%4
!
!
!
!
!
!
(WPICNMKNNKPI CPFE[VQMKPGU
2JCIQE[VQUKU CPFE[VQMKPGU
2JCIQE[VQUKU EJGOQMKPGU CPFE[VQMKPGU
2
415 4#( R
'4-
%#4& $%.
R R %CPQPKECNQT PQPECPQPKECN0(κ$
0.42 +PȯCOOCUQOG
/#.6
2TQ ECURCUG
R R +4( %[VQRNCUO
%CPQPKECNQT PQPECPQPKECN0(κ$
#5%
2TQ+.β 2TQ+.
+.β +.
0WENGWU &GHGPUKPURJCIQE[VQUKU EJGOQMKPGUE[VQMKPGU CPF+&1
&GHGPUKPURJCIQE[VQUKU EJGOQMKPGUCPFE[VQMKPGU
Figure 1 | Signalling pathways in innate recognition of fungi. Pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs) that are present during fungal infections are recognized by pattern recognition receptors (PRRs). The major PRRs are Toll-like receptors (TLRs); C‑type lectin receptors (CLRs; such as dectin 1 (also known as CLEC7A), dectin 2 (also known as CLEC6A), DC-specific ICAM3‑grabbing non-integrin (DC-SIGN), mincle and the mannose receptor); galectin family proteins (such as galectin 3) and receptor for advanced glycation end-products (RAGE). TLRs and CLRs activate multiple intracellular pathways upon binding to specific fungal PAMPs, including β‑glucans (especially β‑(1,3)-glucans with varying numbers of β‑(1,6) branches), chitin, mannans linked to proteins through N‑ or O‑linkages, β‑(1,2)‑linked oligomannosides and fungal nucleic acids. These signals activate canonical or non-canonical nuclear factor-κB (NF-κB) and the NOD‑, LRR- and pyrin domain-containing 3 (NLRP3) inflammasome, and this culminates in the production of defensins, chemokines, cytokines, reactive oxygen species (ROS) and indoleamine 2,3‑dioxygenase (IDO). Complement receptor 3 (CR3) and members of the scavenger receptor family (such as CD36) mediate recognition of β‑glucans and the fungal adhesin BAD1 (Blastomyces adhesion 1). After TLR activation, protease-activated receptors (PARs) sense proteolytic virulence factors and tissue injury and contribute to fungal recognition through a dual sensor system. In addition, the alarmin S100B, through the spatio-temporal integration of signals from TLRs and RAGE, allows the immune system to discriminate between pathogen-derived and endogenous danger signals. By forming complexes with various TLR2 ligands, S100B inhibits TLR2 through a paracrine epithelial cell- and neutrophil-mediated regulatory circuit, and this accounts for its anti-inflammatory activity. However, the ability of S100B to bind nucleic acids results in the activation of intracellular TLRs that signal through TIR domain-containing adaptor protein inducing IFNβ (TRIF; also known as TICAM1) and this eventually resolves damage-associated inflammation through transcriptional downregulation of S100B gene expression. ASC; apoptosisassociated speck-like protein containing a CARD; BCL‑10, B cell lymphoma 10; CARD9, caspase recruitment domain-containing protein 9; ERK, extracellular signal-regulated kinase; FcRγ, Fc receptor γ‑chain; IL, interleukin; IRF3, IFN-regulatory factor 3; MALT1, mucosa-associated lymphoid tissue lymphoma translocation protein 1; MYD88, myeloid differentiation primary response protein 88; SYK, spleen tyrosine kinase.
2#/2U
/2U
(WPICNQTJQUVRTQVGCUGU 6.4
6.4
5$ 5$
6.4
4#)'
2#4
64+(
/;&
/;& 5$
'4-
0WENGKE CEKF
R 6.4
R R %CPQPKECNQT PQPECPQPKECN0(κ$
64+(
6.4
R R %CPQPKECNQT PQPECPQPKECN0(κ$
%[VQMKPGU +PȯCOOCVKQP
278 | APRIL 2011 | VOLUME 11
%[VQMKPGU +PȯCOOCVKQP
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
0CVWTG4GXKGYU^+OOWPQNQI[
REVIEWS
Allergic bronchopulmonary aspergillosis (ABPA). A condition that is characterized by an exaggerated airway inflammation (hypersensitivity response) to Aspergillus spp. (most commonly Aspergillus fumigatus). It occurs most often in patients with asthma or cystic fibrosis.
Protease-activated receptors (PARs). A family of four G protein-coupled receptors. Proteolytic cleavage within the extracellular amino terminus exposes a tethered ligand domain, which activates the receptors to initiate multiple signalling cascades. Many proteases that activate PARs are produced during tissue damage, and PARs make important contributions to tissue responses to injury, including haemostasis, repair, cell survival, inflammation and pain.
NOD-like receptors (NLRs). A family of cytosolic proteins that recognize pathogen-associated molecular patterns and endogenous ligands. The recognition of ligands induces a signalling cascade leading to activation of nuclear factor-κB, or the inflammasome, to produce pro-inflammatory cytokines. NLRs are also involved in signalling for cell death.
Dectin 2 recognizes high-mannose structures that are common to many fungi and binds hyphal forms with higher affinity than yeast forms. Dectin 2 selectively pairs with the Fc receptor γ‑chain (FcRγ) to induce pro-inflammatory cytokine and leukotriene release (FIG. 1). Dectin 2‑deficient mice are highly susceptible to infection with C. albicans but not with C. neoformans 23,24; however, the underlying reasons for the different susceptibilities are unclear. Mincle, which is mainly expressed by macrophages, is also an FcRγ-associated activating receptor. It senses damaged cells, recognizes Malassezia spp.25 and C. albicans 26 and, similarly to dectin 1, induces NF‑κB-mediated inflammatory responses through SYK–CARD9 signalling. The mannose receptor and DC‑SIGN recognize branched N‑linked mannans, and both receptors can direct mannosylated fungal antigens into the DC endocytic pathway 27,28, thereby promoting antigen processing and presentation to T cells. Indeed, the mannose receptor has been shown to be involved in the promotion of antifungal TH17 cell responses29. The mannose receptor also has affinity for α‑glucans and chitin, whereas langerin, which is selectively expressed by Langerhans cells, mainly recognizes sulphated and mannosylated glycans30. Although the mannose receptor is involved in the phagocytosis of unopsonized Candida yeasts31, deficiency of this receptor does not confer susceptibility to C. albicans infection as it does to C. neoformans infection32. Consistent with the lack of classical signalling motifs within the cytoplasmic tail, the mannose receptor induces the production of inflammatory cytokines in collaboration with TLRs, dectin 1 and peroxisome proliferator activated receptor‑γ33. TLRs. TLR2, TLR4 and TLR9 are the main TLRs that are involved in sensing fungal components, such as zymosan, phospholipomannan, O‑linked mannans and fungal DNA12. Although studies have shown that mice lacking the TLR signalling adaptor myeloid differentiation primary response protein 88 (MYD88) are highly susceptible to infections with various fungi13, the physiological roles of individual TLRs in fungal infections are still unclear. In general, the contribution of individual TLRs may vary depending on the fungal species, fungal morphotypes, route of infection and receptor cooperativity. Nevertheless, human studies have shown that a polymorphism in TLR4 (Asp229Gly) is associated with increased susceptibility to pulmonary aspergillosis34–36 and bloodstream candidiasis37, and that a polymorphism in the promoter of TLR9 (T‑1237C) is associated with allergic bronchopulmonary aspergillosis (ABPA)36 (TABLE 1). Similarly to CLRs, such as the mannose receptor and DC‑SIGN, TLRs facilitate the presentation of fungal antigens by DCs and tailor T cell responses. This is consistent with the role of TLRs in controlling microbial antigen processing and presentation during the simultaneous phagocytosis of self and non-self components38. During inflammation, host and fungal proteases trigger the activation of protease-activated receptors (PARs), a family of G protein-coupled receptors39. The stimulation of TLRs by fungi unmasks the divergent roles of PAR1
and PAR2 in downstream signalling and inflammation. After fungal recognition by TLRs, PARs become activated to sense proteolytic virulence factors and tissue injury, to mediate pro-inflammatory (PAR1) or antiinflammatory (PAR2) responses and to modulate the activity of TLRs (FIG. 1). Thus, TLRs regulate PAR signalling and vice versa39. A similar model of dual sensing of fungal PAMPs and virulence factors has been observed in Drosophila and plants. In Drosophila, a fungal protease used by the entomopathogenic fungus Beauveria bassiana to digest the cuticle has been shown to activate the Toll pathway by inducing the maturation of Persephone into an active protease40. In plants, an indirect mode of non-self recognition — through the perception of host destruction — has been identified, in which secreted proteins that are produced by biotrophic fungi both trigger and suppress host defence41. NLRs. Although cytoplasmic receptors for fungi have yet to be described, the NOD-like receptors (NLRs) are implicated in sensing fungi and, once activated, these receptors induce the production of IL‑1β and IL‑18 through the formation of inflammasomes12,42,43. Mice lacking IL‑1 receptor type I (IL‑1RI) signalling, IL‑18 or caspase 1 have disparate patterns of susceptibility to fungal infections12; however, mice lacking NLRP3 consistently show enhanced susceptibility to candidiasis44,45. Consistent with an association between the NLRP3 inflammasome and several autoinflammatory conditions, and also with epithelial cell protection in the gut 46, defective NLRP3 activation increases C. albicans colonization in the gut and exacerbates inflammation in Crohn’s disease47. This illustrates how a commensal organism such as C. albicans can become pathogenic in certain contexts. DAMPs. Mammalian PRRs recognize not only PAMPs but also damaged host cell components, such as nucleic acids and alarmins, collectively known as damageassociated molecular patterns (DAMPs)48. Despite the identification of specific signalling pathways that negatively regulate responses to either PAMPs or DAMPs48, the unexpected convergence of the molecular pathways responsible for the recognition of PAMPs and DAMPs raised the question of whether and how the host immune system discriminates between these two types of molecular patterns. The relative contributions of PAMPs and DAMPs to inflammation, immune homeostasis and mechanisms of repair during infection were also unclear. However, a mechanism has recently been described that allows the host to discriminate between PAMP- and DAMP-induced immune responses; the alarmin S100B coordinates this process via the spatiotemporal integration of signals from TLRs and the receptor for advanced glycation end-products (RAGE)49. By sequential binding to fungus-derived TLR2 ligands and nucleic acids, S100B first inhibits TLR2‑induced inflammation during fungal pneumonia and then subsequently activates intracellular TLR3 and TLR9 to induce its own transcriptional downregulation (FIG. 1). Thus, the crosstalk between RAGE and TLRs represents a regulatory circuit in infection, whereby an endogenous danger signal
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 279 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS Table 1 | Major single nucleotide polymorphisms associated with susceptibility to fungal infections and diseases Gene
SNPs or haplotypes
SNP effect
Disease
Outcome
Refs
CARD9
Q295X
Low numbers of TH17 cells
Chronic mucocutaneous candidiasis
Susceptibility
21
CXCL10
+11101C/+1642G/ −1101A
Reduced chemokine production by DCs exposed to Aspergillus fumigatus
Invasive aspergillosis
Susceptibility
138
DECTIN1
Y223S
Reduced zymosan-binding capacity and IFNγ production
Oropharyngeal candidiasis
Resistance
139
Y238X
Decreased cell surface expression, β‑glucan-binding capacity and impaired cytokine production
Chronic mucocutaneous candidiasis, Candida albicans colonization and invasive aspergillosis
Susceptibility
DEFB1
−44G
Unknown
C. albicans carriage
Resistance
142
IFNG
+874TT
Increased levels of IFNγ
ABPA, CCPA
Susceptibility
143
IL1RN IL1A IL1B
VNTR2/−889C/−511T
Increased levels of C‑reactive protein
Invasive aspergillosis
Susceptibility
144
IL4
−1098T/−589C/−33C
Unknown
Chronic disseminated candidiasis
Susceptibility
145
−589T
Increased levels of vaginal IL-4 and reduced levels of nitric oxide and MBL
Recurrent VVC
Susceptibility
146
18,140, 141
−589T
Reduced levels of IL-4
Paracoccidioidomycosis
Susceptibility
147
IL4R
I75V
Upregulation of CD23 expression
ABPA
Susceptibility
148
IL10
−1082AA −1082A/−819C/−592C
Reduced levels of IL-10
Invasive aspergillosis
Resistance
−1082A
Reduced levels of IL-10
CCPA
Susceptibility
143
−1082GG
Increased levels of serum IL-10
ABPA, A. fumigatus colonization
Susceptibility
151
IL15
+13689A
Increased levels of IL-15
ABPA, CCPA
Susceptibility
143
IL23R
R381Q
Impaired production of IL-17A
Invasive aspergillosis
Resistance
152
MASP2
D105G
Impaired MBL function
Invasive aspergillosis
Susceptibility
153
MBL2
O/O, A/O
Impaired MBL activity
Invasive aspergillosis, CCPA, VVC
Susceptibility
153,154
LXA/O
Reduced levels of circulating MBL
Invasive aspergillosis
Susceptibility
153
+1011A
Elevated plasma MBL levels and high peripheral blood eosinophilia
ABPA
Susceptibility
155
NLRP3
Length polymorphism (allele 7)
Impaired production of IL-1β
Recurrent VVC
Susceptibility
156
PLG
D472N
Predicted to enhance plasminogen binding to A. fumigatus
Invasive aspergillosis
Susceptibility
157
SFTPA2
A91P, R94R
Increased levels of total IgE and eosinophilia
ABPA, CCPA
Susceptibility
158,159
TGFB1
+869C
Decreased levels of TGFβ
CCPA
Susceptibility
143
TLR1
R80T, N248S
Unknown
Invasive aspergillosis
Susceptibility
160
TLR4
D299G/T399I
Predicted to impair the ligand−binding domain
Invasive aspergillosis, A. fumigatus colonization, CCPA, C. albicans systemic infections
Susceptibility
34–37
TLR6
S249P
Unknown
Invasive aspergillosis
Susceptibility
160
TLR9
T−1237C
Increased NF‑κB binding affinity
ABPA
Susceptibility
36
TNF
−308G
Decreased levels of TNF
ABPA, CCPA
Susceptibility
143
TNFR1
+36G, −609T
Decreased levels of TNFR1 mRNA
Invasive aspergillosis
Susceptibility
161
TNFR2
VNTR at −322
Unknown
Invasive aspergillosis
Susceptibility
162
149,150
ABPA, allergic bronchopulmonary aspergillosis; CARD9, caspase-recruitment domain family, member 9; CCPA, chronic cavitary pulmonary aspergillosis; CXCL, CXC-chemokine ligand; DC, dendritic cell; DEFB1, β‑defensin 1; IFN, interferon; IL, interleukin; IL1RN, IL-1 receptor antagonist; MASP2, mannan-binding lectin serine protease 2; MBL, mannose-binding lectin; NF‑κB, nuclear factor-κB; NLRP3, NOD‑, LRR- and pyrin domain-containing 3; PLG, plasminogen; SFTPA2, surfactant protein A2; SNP, single nucleotide polymorphism; TGF, transforming growth factor; TLR, Toll-like receptor; TH17, T helper 17; TNF, tumour necrosis factor; TNFR, TNF receptor; VNTR, variable-number tandem repeat; VVC, vulvovaginal candidiasis.
280 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS protects the host against pathogen-induced inflammation and a nucleic acid-sensing mechanism terminates the inflammation induced by the endogenous danger signal. This raises the intriguing possibility that the host may have developed mechanisms to ameliorate the response to DAMPs via PAMPs.
Hyphae In moulds, spores germinate to produce branching filaments called hyphae, which are 2–10 μm in diameter and which may form a mass of intertwining strands called a mycelium.
Hydrophobins A family of small, moderately hydrophobic proteins that are characterized by the conserved spacing of eight cysteine residues. Hydrophobins are present on the surface of many fungal conidia, and are responsible for the rodlet configuration of the outer conidial layer.
Delayed-type hypersensitivity response A cellular immune response to antigen that develops over a period of ~24–72 hours. The response is characterized by the infiltration of T cells and monocytes and depends on the production of T helper 1type cytokines.
Paracoccidioidomycosis A chronic granulomatous disease involving the lungs, skin, mucous membranes, lymph nodes and internal organs that is caused by Paracoccidioides brasiliensis. Symptoms include skin ulcers, adenitis and pain owing to abdominal organ involvement.
Fungal evasion of inflammation Fungi produce several factors that are potent regulators of the host inflammatory response50,51. By masking or subverting the host detection systems, fungi may avoid inflammation, and this contributes to fungal adaptation and opportunism52,53. As mentioned, the fungal cell wall is a dynamic structure that is continuously changing throughout the fungus cell cycle and during morphological transition. For example, β‑(1,3)-glucans are exposed in the bud scar of C. albicans yeasts but are masked on hyphae, thus favouring fungal escape from recognition by dectin 1. Similarly, α‑(1,3)-glucans, which are associated with virulence in B. dermatitidis, H. capsulatum and P. brasiliensis, block innate immune recognition of β‑glucans by dectin 1 (REF. 54). In addition, A. fumigatus conidia are covered by hydrophobins and melanin that prevent immune recognition55, whereas P. jiroveci evades immunosurveillance by changing the expression of major surface glycoproteins56. Also, many fungi exploit CR3 to dampen the inflammatory response and allow intracellular fungal parasitism10. The most extreme example of evasion of innate immune recognition is mediated by the capsule of C. neoformans, which completely covers the fungal cell wall and prevents recognition by PRRs and the induction of inflammation57. In addition, C. neoformans yeast can escape from macrophages through an expulsive mechanism that does not kill the host cell and avoids inflammation58. The so called ‘Trojan horse’ model suggests that replication within, lateral transfer between and eventual expulsion of yeasts from macrophages might explain how C. neoformans establishes latency and spreads in the host without triggering inflammation59. By continually activating the PRR system, it is possible that fungi contribute to inflammatory processes and promote autoimmunity. Indeed, dectin 1 and fungal β‑glucans have been implicated in the induction of autoimmune arthritis60 and psoriasis61, and zymosan has been linked with the induction of experimental autoimmune encephalomyelitis62. T cell responses to fungi In higher organisms, innate sensing mechanisms are hard-wired to activate distinct CD4+ TH cells that have protective and non-protective functions against fungi (FIG. 2). This suggests that the adaptive immune system has co-evolved with ubiquitous or commensal fungi. DCs are uniquely adept at decoding the fungus-associated information. A whole-genome transcriptional analysis of fungus-stimulated DCs indicated the presence of a specific transcriptional programme that governs the recognition of fungi63. The ability of a given DC subset to respond with different activating programmes and to activate distinct intracellular signalling pathways following the ligation of different PRRs64,65 confers unexpected plasticity to the
DC system and contributes to shaping T cell responses in infection66,67 and following vaccination68 (FIG. 2). The capacity of DCs to initiate different adaptive antifungal immune responses also depends on specialization and cooperation between DC subsets66,67. Inflammatory DCs initiate antifungal TH17 and TH2 cell responses in vivo through signalling pathways involving the TLR adaptor MYD88, whereas tolerogenic DCs activate TH1 and regulatory T (TReg) cell differentiation programmes through mechanisms that involve the signalling adaptor TRIF (TIR domain-containing adaptor protein inducing IFNβ; also known as TICAM1). In addition, signal transducer and activator of transcription 3 (STAT3), which affects the balance between canonical and non-canonical activation of NF‑κB and thus the expression of the enzyme indole amine 2,3‑dioxygenase (IDO), has a key contribution to DC plasticity and functional specialization. The multiple, functionally distinct, receptor signalling pathways in DCs ultimately affect the balance between CD4+ effector T cells and TReg cells and thus are likely to be exploited by fungi to enable them to establish commensalism or infection. Although epithelial cells are not professional antigenpresenting cells, they may have important immunological roles, as they express PRRs. Following the stimulation of these receptors, epithelial cells can initiate and amplify TH2 cell responses, via thymic stromal lymphopoietin, IL‑25 and IL‑33 (REF. 69), and provide the machinery required for the induction of T cell tolerance70 (FIG. 2). TH1 cells. A dominant TH1 cell response correlates with protective immunity against fungi10,71–73 and effective fungal vaccines74. TH1 cell activation is determined by the DC response to the combination of TLR and CLR signals provided by fungi (FIG. 2). Through the production of the signature cytokine IFNγ and the provision of help for the production of opsonizing antibodies, TH1 cells are instrumental in the optimal activation of phagocytes at sites of infection. Therefore, the failure of T cells to deliver activating signals to effector phagocytes may predispose patients to overwhelming infections, limit the therapeutic efficacy of antifungal agents and antibodies and favour the persistence of fungi10. Adaptive immune responses to commensal or dimorphic fungi occur in immunocompetent individuals (as indicated by a positive delayed-type hypersensitivity (DTH) response) and correlate with protection and a favourable prognosis. In the case of C. neoformans infection, the high prevalence of antibodies to cryptococcal antigens in normal individuals suggests that primary infection is followed by fungal growth restriction and concomitant immunity 10. Indeed, direct inhibition of T cell proliferation by fungal polysaccharides may underlie the defective cellular immunity of patients with persistent cryptococcal infections75. Studies of patients with polar forms of paracoccidioidomycosis have shown an association between TH1 cell reactivity and asymptomatic and mild forms of the infection, whereas TH2 cell responses are associated with severe disease and disease relapse. Thus, the finding that oestradiol favours TH1‑type immune responses may explain why paracoccidioidomycosis is at least ten times more frequent in men than in women76.
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 281 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS C&GPFTKVKEEGNN
/4
&%5+)0 .CPIGTKP
&GEVKP /KPENG (E4γ
+.
5;%#4&
&GEVKP
%[UVGKP[N NGWMQVTKGPGU
4#(
#URGTIKNNWUEQPKFKC
6.4
6.4 6.4
/;&
64+( 6[RG++(0 +. +. +&1
8.#
#URGTIKNNWU
+. +.β +.
+.
#URGTIKNNWUJ[RJCG $NCUVQO[EGU
6.4
%CPFKFC[GCUVU
/;& +.
+.
%CPFKFCJ[RJCG +. +. +. +.β
%QEEKFKQKFGU %T[RVQEQEEWU >OCVQRJ[VGU (WPICNCNNGTIGPU *KUVQRNCUOC
6*EGNN
6*EGNN
6*EGNN
+.#+.( +.
+(0γ r2TQOQVKQPQHHWPICN
ENGCTCPEG r+PȯCOOCVKQP
+. +.
r4GETWKVOGPVQHPGWVTQRJKNU r&GHGPUKPU r+PȯCOOCVKQP
• Inhibition of fungal clearance r#NNGTI[
D/CETQRJCIG
/CNCUUG\KC 2CTCEQEEKFKQKFGU
6)(β +.
2PGWOQE[UVKU 5CEEJCTQO[EGU
r.KOKVGFRCVJQNQI[ r+OOWPQUWRRTGUUKQP
5RQTQVJTKZ
E'RKVJGNKCNEGNN
/KPENG %4
64GIEGNN
6.4 ):/4
6.4
/4 +4(
6.4 &GEVKP
5;KPȯCOOCUQOG 6[RG++(0
+.
6*EGNN
6*EGNN
+.β +.
/4 FGEVKP %CPQPKECNQT PQPECPQPKECN0(κ$
)CNGEVKP +. 6)(β
64GIEGNN
+.
+.
6*EGNN
%%.
+. +&1
6*EGNN
+. +.
64GIEGNN
Figure 2 | CD4+ T cell subsets in fungal infections. The figure shows how different antigen-presenting cells stimulate the differentiation of CD4+ T helper (TH) cells and regulatory T (TReg) cells in response to fungi, depicting the transcription factors 0CVWTG4GXKGYU^+OOWPQNQI[ involved, the cytokines produced and the possible effector and regulatory functions induced. Through the production of distinct sets of cytokines and other mediators, T cells can act as immune effectors and as master regulators of the inflammatory and effector responses of innate cells. The ability of dendritic cells (a), macrophages (b) and epithelial cells (c) to respond to fungi with flexible intracellular signalling pathways that reflect the different pattern recognition receptor– pathogen-associated molecular pattern combinations confers unexpected plasticity on the system and contributes to shaping T cell responses. The multiple, functionally distinct signalling pathways in antigen-presenting cells ultimately affect the local TH cell/TReg cell balance, and are likely to be exploited by fungi to allow commensalism or opportunism. CARD9, caspase recruitment domain-containing protein 9; CCL3, CC‑chemokine ligand 3; CR3, complement receptor 3; DC-SIGN, DC-specific ICAM3‑grabbing non-integrin; FcRγ, Fc receptor γ‑chain; GXMR, receptor(s) for the Cryptococcus capsular component glucuronoxylomannan; IDO, indoleamine 2,3‑dioxygenase; IFN, interferon; IL, interleukin; IRF3, IFN-regulatory factor 3; MR, mannose receptor; MYD88, myeloid differentiation primary response protein 88; NF‑κB, nuclear factor-κB; SYK, spleen tyrosine kinase; TGFβ, transforming growth factor‑β; TLR, Toll-like receptor; TRIF, TIR domain-containing adaptor protein inducing IFNβ (also known as TICAM1); VLA5, very late antigen 5.
282 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS T H2 cells. IL‑4 and IL‑13 provide the most potent proximal signals for the commitment of naive T cells to the TH2 cell lineage, which, by dampening protective TH1 cell responses and promoting the alternative pathway of macrophage activation, favours fungal infections, fungus-associated allergic responses and disease relapse77–79. Accordingly, limiting IL‑4 production restores antifungal resistance80. In atopic subjects and neonates, the suppressed DTH response to fungi is associated with elevated levels of antifungal IgE, IgA and IgG. In patients with cystic fibrosis, heightened TH2 cell reactivity is associated with ABPA81; however, TH2 cell-dependent humoral immune responses may afford some protection82, in part by promoting TH1 cell responses83,84 and also by altering the intra cellular trafficking of fungi within macrophages85 and fungal gene expression86. A mechanism whereby serum IgM with specificity for conserved fungal antigens bridges innate and adaptive immune responses against fungal organisms has also recently been described87. The efficacy of certain vaccines that elicit the production of protective antibodies indicates that antibody responses can make a decisive contribution to host defence against medically important fungi74,88,89. TH17 cells. Although TH1 cell responses are central to host protection against fungi, it is also clear that patients with genetic defects in the IL‑12, IL‑23 and IFNγ pathways do not have increased susceptibility to most infectious agents, including fungi22, with few exceptions90. Indeed, genetic deficiencies have indicated a role for the dectin 1–CARD9, STAT3 and TH17 cell pathways in protection against fungal infections6,22. TH17 cells have an important function in the host response against extracellular pathogens, but they are also associated with the pathogenesis of many autoimmune and allergic disorders. TH17 cell activation occurs in fungal infections23,44,66,67,70,73,91–101, mainly through the SYK–CARD9, MYD88 and mannose receptor signalling pathways in DCs and macrophages (FIG. 2). It is inhibited by the RAF and TRIF–type I IFN pathways, suggesting that mechanisms of activation and inhibition of TH17 cells are present downstream of both CLRs and TLRs. TH17 cells are present in the fungus-specific T cell memory repertoire in humans102–104 and mediate vaccineinduced protection in mice105. However, host defence against A. fumigatus relies on TH1 cell responses rather than TH17 cell responses104, and patients with chronic mucocutaneous candidiasis (with or without autosomal dominant hyper-IgE syndrome) have defective TH17 and also TH1 cell responses106. This could be explained by the notion that TH17 cells, although found early during the initiation of an immune response, are involved in a broad range of both TH1- and TH2‑type responses. Indeed, a role for TH17 cells in supporting TH1 cell responses has been shown in experimental mucosal candidiasis99,107. In addition, preliminary evidence shows that in experimental aspergillosis, increased TH2 cell responses and fungal allergy were observed in conditions of defective IL‑17A receptor (IL‑17RA) signalling (L.R., A. De Luca, T. Zelante and R.G. Iannitti, unpublished observations).
These findings point to an important regulatory function for the TH17 cell pathway in promoting TH1‑type immune responses and restraining TH2‑type responses, and also explain the immunological findings observed in patients with chronic mucocutaneous candidiasis and autosomal dominant hyperIgE syndrome. In terms of effector functions, although the ability of IL‑17A to mobilize neutrophils and induce the production of defensins greatly contributes to the prompt and efficient control of an infection at different body sites, conflicting results have been obtained regarding whether the IL‑17A–IL‑17RA pathway is essential24,105,107,108 or not 91,99,109 during infection. This suggests that the activity of this pathway may depend on the stage and site of infection, and is probably influenced by environmental stimuli that induce cells to produce TH17 cell-associated cytokines, including IL‑22 (see below). It is intriguing that TH17 cell responses are downregulated by C. albicans 110, and failure of this downregulation may eventually result in chronic inflammation and impair the resolution of the infection93,111. The mechanisms that link inflammation to chronic infection may involve a failure to restrain inflammation following IL‑17A‑dependent neutrophil recruitment, thereby preventing optimal protection and favouring fungal persistence. Thus, the TH17 cell pathway could be involved in the immunopathogenesis of chronic fungal diseases, in which persistent fungal antigens may promote immune dysregulation. This may occur in patients with autoimmune polyendocrine syndrome type 1 and in the mouse model of this disorder (autoimmune regulator (AIRE)-deficient mice), in which excessive TH17‑type responses to fungi have been observed112.
Balancing resistance and tolerance to fungi The role of TReg cells. During a fungal infection, the immune response must eliminate the fungus while limiting collateral damage to tissues and restoring a homeostatic environment. Several clinical observations suggest an inverse relationship between IFNγ and IL‑10 production in patients with fungal infections10. High levels of IL‑10, which negatively affect IFNγ production, are detected in chronic candidal diseases, in the severe forms of endemic mycoses and in neutropenic patients with aspergillosis, and thus have been linked to susceptibility to fungal infections113. However, given its prominent effect on the resolution of inflammation, IL‑10 production may be a consequence, rather than a cause, of the infection113. This predicts that, in the case of chronic fungal infections that are dominated by non-resolving inflammation, IL‑10 acts as a homeostatic host-driven response to keep inflammation under control. TReg cells with antiinflammatory activity have been described in fungal infections of both mice and humans. In experimental fungal infections, both inflammation and immune tolerance in the respiratory or gastrointestinal mucosa were shown to be controlled by the coordinated activation of different TReg cell subsets. However, as TReg cell responses may limit the efficacy of protective immune responses, the consequence of TReg cell activity is reduced damage to the host but also fungal persistence113 and, eventually,
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 283 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS C
&KGVCT[VT[RVQRJCP (WPICNDKQVC
/KETQDKQVC
1s
#*4NKICPFU #PVKOKETQDKCN RGRVKFGU
)WV NWOGP
56#6
'RKVJGNKCN EGNN +&1 +&1
+.
&%
.-[PWTGPKPG
414γV 0-Rs %& 6EGNNU
↑414γV
414γV 0-R #*4 KPPCVGN[ORJQKFEGNNU
6*EGNN
+.
%& 6EGNN
↑(1:2
64GIEGNN
+.# 4GUKUVCPEG %QPVTQNQHHWPICNDWTFGP
6QNGTCPEG %QPVTQNQHKPȯCOOCVKQP
D
$GPGȮEKCNKPȯCOOCVKQP &GVTKOGPVCNKPȯCOOCVKQP
2TQVGEVKXGKOOWPKV[ %QOOGPUCNKUO +OOWPQTGIWNCVKQP
(WPICNRGTUKUVGPEG +OOWPQUWRRTGUUKQP
Figure 3 | Resistance and tolerance to fungi and the regulation of these processes. a | The tryptophan metabolism pathway is exploited by the mammalian host and by 0CVWTG4GXKGYU^+OOWPQNQI[ commensals to increase fitness in response to fungal infection through the processes of resistance and tolerance. Infection of the gut with Candida albicans leads to the production of interleukin‑22 (IL‑22) by CD3–NKp46+RORγt+AHR+ innate lymphoid cells, through a mechanism involving aryl hydrocarbon receptor (AHR) ligands and IL‑23. IL‑22 then acts on epithelial cells, leading to the activation of signal transducer and activator of transcription 3 (STAT3) and, together with IL‑17A produced by NKp46–RORγt+ T cells, to the production of antimicrobial peptides. Various indole derivatives, which are generated through the conversion of dietary tryptophan by commensal intestinal microorganisms, act as endogenous ligands for AHR and thereby contribute to IL‑22 production. Fungus-induced activation of tryptophan catabolism by indoleamine 2,3‑dioxygenase (IDO) expressed by dendritic cells (DCs) and epithelial cells leads to the production of immunologically active compounds that induce the transcription of forkhead box P3 (FOXP3) and suppress the transcription of retinoic acid receptor-related orphan receptor-γt (RORγt) in T cells, resulting in the generation of regulatory T (TReg) cells. b | These findings support a model in which the AHR–IL‑22 axis, together with the IL‑17A–T helper 17 (TH17) cell pathway, control initial fungal growth (that is, resistance) and epithelial cell homeostasis. By contrast, the exploitation of the interferon-γ–IDO axis for functional specialization of antifungal regulatory mechanisms (that is, tolerance) may have allowed the fungal microbiota to evolve with the mammalian immune system, survive in conditions of inflammation and prevent dysregulated immune responses. The balance between resistance and tolerance to fungi may accommodate the spectrum of host–fungus relationships, ranging from protection and immunopathology to fungal persistence and immunosuppression.
immunosuppression114 (FIG. 2). Thus, by controlling the quality and magnitude of innate and adaptive effector responses, TReg cells may be responsible for a spectrum of outcomes, ranging from protective tolerance (defined as a host response that ensures survival of the host through a trade-off between sterilizing immune responses and their negative regulation, which limits pathogen elimination) to overt immunosuppression. Furthermore, this suggests that the interactions between fungi and the host immune system may determine whether a fungus is defined as a commensal or as a pathogen, and this status may change continuously. The contribution of fungi. It is not surprising that many of the strategies that mammalian hosts have developed to coexist peacefully with their microbiota can be hijacked or manipulated by commensals to ensure their own survival. Manipulation of the regulatory network of the host by the fungal microbiota is one such mechanism to ensure fungal survival113. C. albicans and the fungal product zymosan have been shown to activate a tolerogenic programme in gut macrophages115 and DCs67,98, resulting in the activation of TReg cell-dependent immune tolerance. In normal skin, Malassezia spp. fungi downregulate inflammation by inducing TGFβ1 and IL‑10 production, and thus are able to establish themselves as commensals. By contrast, in atopic dermatitis and psoriasis, the interaction of fungi with the defective skin barrier promotes epithelial hyperproliferation, inflammatory cell recruitment and disease exacerbation116. In the case of C. albicans, besides behaving as a commensal, this fungus can also actively promote tolerance, leading to the amelioration of gut inflammation67. Thus, similarly to symbionts, the fungal microbiota may actively contribute to the balance between inflammation and tolerance at mucosal surfaces, as well as at distant sites, to benefit both the host and the fungus. Moreover, host regulatory responses may contribute to the transition of fungi from symbionts to pathobionts. In this scenario, it is clinically important to distinguish between conditions in which yeasts are a cause (that is, required for disease), a trigger (that is, not required, but may favour disease progression) or a sign (that is, pathogenicity is promoted by a host failure) of non-resolving inflammation and associated clinical manifestations. Tryptophan metabolism. A reciprocal relationship has been described between the development of forkhead box P3 (FOXP3)+ TReg cells and effector TH17 cells in fungal infections96,117. IDO is a metabolic enzyme that has been shown to affect the TReg/TH17 cell balance during fungal infections, resulting in the suppression of inflammation and the promotion of protective tolerance (FIG. 3). Initially identified in infection because of its antimicrobial activity (through tryptophan starvation of intracellular parasites), IDO is now widely recognized as a suppressor of acute inflammatory responses and a regulator of mammalian immune homeostasis118. Unsurprisingly therefore, the induction of IDO by microorganisms may be an evasion mechanism that
284 | APRIL 2011 | VOLUME 11
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS allows them to establish commensalism or chronic infection. Through their capacity to induce TReg cells and inhibit TH17 cell development, IDO-expressing cells and kynurenines (molecules produced by IDO) may have unexpected potential in the control of inflammation and allergy in fungal infections119.
Symbiont An intestinal microorganism that contributes to host nutrition and fitness through a mutualistic, beneficial interaction.
Pathobiont A microbial symbiont that can cause diseases as a consequence of the perturbation of intestinal homeostasis.
Dysbiosis Alteration of the symbiont microbial community.
The AHR–IL‑22 pathway. Recent evidence indicates that IL‑22, a member of the IL‑10 cytokine family, has a crucial role in innate immune defence and mucosal protection from damage120. IL‑22 is produced by various cell types, including innate lymphoid cells that express natural killer (NK) cell markers (such as NKp46 (also known as NCR1)), NKT cells, lymphoid tissue-inducer cells, TH1 cells and TH17 cells. It regulates intestinal homeostasis and wound healing by activating STAT3 in epithelial cells121. A recent study has shown that IL‑22 is required for the control of C. albicans growth at mucosal sites in the absence of TH1 and TH17 cells99. In this study, IL‑22 produced by NKp46+ innate lymphoid cells expressing the aryl hydrocarbon receptor (AHR) was found to directly target intestinal epithelial cells. This resulted in the induction of STAT3 phosphorylation in the epithelial cells and the release of S100A8 and S100A9 peptides, which are known to have antifungal activity and anti-inflammatory effects (FIG. 3). Consistent with this role for IL‑22, patients with autosomal dominant hyper-IgE syndrome owing to dominant-negative mutations of STAT3 have a defective TH17 cell response to C. albicans 122, and this is probably amplified by compromised IL‑22‑induced effects on STAT3‑mutant epithelial cells. Vaginal epithelial cells also produce S100A8 and S100A9 following interaction with C. albicans 123, suggesting the possible involvement of IL‑22 in vaginal candidiasis. In addition, naturally occurring IL‑22‑producing cells are highly enriched at mucosal sites, where continuous exposure to fungi occurs, and IL‑22‑expressing CD4+ memory T cells specific for C. albicans are present in healthy individuals124 but are lacking in patients with chronic mucocutaneous candidiasis125. So, IL‑22 production in the mucosa may be a primitive mechanism of resistance against fungi under conditions of limited inflammation. Various indole derivatives, which are generated from dietary tryptophan by commensal intestinal microorganisms 126, act as endogenous ligands for AHR and mediate IL‑22 production127,128. This suggests that the tryptophan metabolism pathway could be exploited by commensals and the host to increase fitness in response to fungi, through the induction of resistance and tolerance. These findings support a model (FIG. 3) in which the AHR–IL‑22 axis, in conjunction with the IL‑17–TH17 cell pathway, controls initial fungal growth (that is, host resistance) and epithelial cell homeostasis, through primitive antifungal effector mechanisms, such as the release of defensins and antimicrobial peptides. By contrast, the exploitation of the IFNγ–IDO axis for antifungal regulatory mechanisms (which promote tolerance) may have allowed the fungal microbiota to evolve with the mammalian immune system, survive in conditions of inflammation and prevent dysregulated immune responses129. The two
pathways, although non-redundant, are reciprocally regulated and compensate for each other in the relative absence of either one99. Accordingly, commensaldriven mucosal responses are upregulated in animals that lack IDO130, and IL‑22 responses are upregulated when adaptive immune responses are defective99.This may have led us to underestimate the role of IL‑22 in mucosal candidiasis in IL‑22‑deficient, but otherwise immunocompetent, mice107,131. The model also explains the increased susceptibility to certain fungal infections following antibiotic-induced dysbiosis.
Translating basic research into clinical practices New antifungal drugs. The past decades have brought important progress to the development of more effective and safe antifungal agents88. However, medical treatments that increase host resistance, such as antibiotics, place selective pressures on pathogens. As tolerance mechanisms are not expected to exert the same selective pressure on pathogens, new drugs that target tolerance pathways could provide therapies to which pathogens will not develop resistance. Immune therapies. Breakthroughs in our understanding of how mucosal homeostasis is established, maintained or disrupted during fungal exposure and/or colonization should help to guide the development of new therapeutics that target specific inflammatory or metabolic end points. For example, limiting inflammation — through PRR agonism or antagonism — to stimulate a protective immune response to fungi should pave the way for the rational design of novel immunomodulatory therapies. Combination therapy with such immunomodulatory agents, including antibody-based immunotherapy 88, will probably maximize antifungal effects and reduce immune-mediated pathology and damage. Tryptophan metabolites are also possible targets for therapy as they simultaneously activate antifungal resistance and limit overzealous inflammatory host responses117. Vaccines. With the exception of a killed spherule vaccine against coccidioidomycosis, no fungal vaccine trials have ever been carried out. However, the level of our understanding of fungus−host interactions has progressed to the point at which vaccines against fungi and fungal diseases may become a reality 132,133. Indeed, the screening of signalling pathways in DCs using a systems biology approach could be exploited for the development of chimeric vaccines that can target resistance and tolerance in fungal infections. Functional genomics. It is now clear that genetic variants of molecules involved in the innate recognition of fungi may account, in part, for the inherited differences in human susceptibility to fungal infections35,134. Although the analysis of the genetic traits that modulate susceptibility to fungal infections is complex, it may allow the identification of genetic markers for fungal diseases that occur in high-risk patients. Understanding which patients are at highest risk of developing a lifethreatening infection is at present a major unmet need,
NATURE REVIEWS | IMMUNOLOGY
VOLUME 11 | APRIL 2011 | 285 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS and genetic markers will probably assist in risk assessment. TABLE 1 summarizes the single nucleotide polymorphisms (SNPs) of human innate immune genes that have been linked with susceptibility to fungal infections and diseases.
Concluding remarks and future directions There are several challenging issues in the field of medical mycology and infection-related immunological disorders. These include the control of inflammation leading to tolerance, the molecular bases of immune regulation and dysregulation, and the way in which commensal but opportunistic fungi can switch from a ‘friendly’ relationship with the host to a pathological relationship by evading or subverting host inflammation. If fungi either
1. 2. 3. 4.
5. 6.
7. 8.
9.
10. 11.
12.
13.
14.
15. 16. 17. 18.
Hube, B. Fungal adaptation to the host environment. Curr. Opin. Microbiol. 12, 347–349 (2009). Brown, A. J., Odds, F. C. & Gow, N. A. Infection-related gene expression in Candida albicans. Curr. Opin. Microbiol. 10, 307–313 (2007). Cooney, N. M. & Klein, B. S. Fungal adaptation to the mammalian host: it is a new world, after all. Curr. Opin. Microbiol. 11, 511–516 (2008). Richie, D. L. et al. A role for the unfolded protein response (UPR) in virulence and antifungal susceptibility in Aspergillus fumigatus. PLoS Pathog. 5, e1000258 (2009). Pappas, P. G. Opportunistic fungi: a view to the future. Am. J. Med. Sci. 340, 253–257 (2010). Vinh, D. C., Sugui, J. A., Hsu, A. P., Freeman, A. F. & Holland, S. M. Invasive fungal disease in autosomaldominant hyper-IgE syndrome. J. Allergy Clin. Immunol. 125, 1389–1390 (2010). Garcia-Solache, M. A. & Casadevall, A. Global warming will bring new fungal diseases for mammals. MBio 1, e00061‑10 (2010). Cushion, M. T. et al. Transcriptome of Pneumocystis carinii during fulminate infection: carbohydrate metabolism and the concept of a compatible parasite. PLoS ONE 2, e423 (2007). Schneider, D. S. & Ayres, J. S. Two ways to survive infection: what resistance and tolerance can teach us about treating infectious diseases. Nature Rev. Immunol. 8, 889–895 (2008). Romani, L. Immunity to fungal infections. Nature Rev. Immunol. 4, 1–23 (2004). Jouault, T. et al. Host responses to a versatile commensal: PAMPs and PRRs interplay leading to tolerance or infection by Candida albicans. Cell. Microbiol. 11, 1007–1015 (2009). van de Veerdonk, F. L., Kullberg, B. J., van der Meer, J. W., Gow, N. A. & Netea, M. G. Host–microbe interactions: innate pattern recognition of fungal pathogens. Curr. Opin. Microbiol. 11, 305–312 (2008). Bourgeois, C., Majer, O., Frohner, I. E., Tierney, L. & Kuchler, K. Fungal attacks on mammalian hosts: pathogen elimination requires sensing and tasting. Curr. Opin. Microbiol. 13, 401–408 (2010). Liu, M. et al. The endothelial cell receptor GRP78 is required for mucormycosis pathogenesis in diabetic mice. J. Clin. Invest. 120, 1914–1924 (2010). This paper describes a unique susceptibility of patients with diabetic ketoacidosis to mucormycosis and provides a foundation for the development of new therapeutic interventions aimed at targeting the receptor 78 kDa glucose-regulated protein in endothelial cells. Latge, J. P. Tasting the fungal cell wall. Cell. Microbiol. 12, 863–872 (2010). Brown, G. D. Innate antifungal immunity: the key role of phagocytes. Annu. Rev. Immunol. 29, 1–21 (2011). Geijtenbeek, T. B. & Gringhuis, S. I. Signalling through C‑type lectin receptors: shaping immune responses. Nature Rev. Immunol. 9, 465–479 (2009). Ferwerda, B. et al. Human dectin‑1 deficiency and mucocutaneous fungal infections. N. Engl. J. Med. 361, 1760–1767 (2009).
19. 20.
21.
22. 23. 24.
25. 26.
27.
28. 29.
30.
31. 32.
33.
prevent or trigger excessive and deleterious inflammatory responses, this raises the question of whether collateral damage and inflammatory diseases are provoked only by pathogenic fungi or if commensals can also serve as perpetrators. A related question is how and whether the fungal microbiota contributes to the regulation of inflammation in health and disease. Challenging existing paradigms via a multidisciplinary approach in the fields of fungal pathology and immunopathology, functional genomics, proteomics and bioinformatics will probably lead towards the discovery of ‘commensal signatures’ for the fungal biota and the development of multi-pronged therapeutic approaches for mucosal and systemic fungal diseases (see the website of the European Project ALLFUN for further information).
The first study to describe the association between the Tyr238X mutation in dectin 1 and mucocutaneous candidiasis owing to defective IL‑17, IL‑6 and TNF production. Saijo, S. et al. Dectin‑1 is required for host defense against Pneumocystis carinii but not against Candida albicans. Nature Immunol. 8, 39–46 (2007). Taylor, P. R. et al. Dectin‑1 is required for β-glucan recognition and control of fungal infection. Nature Immunol. 8, 31–38 (2007). References 19 and 20 were the first two studies to indicate a function for β‑glucan recognition by dectin 1 in antifungal immunity and demonstrate that a non-TLR signalling pathway is required for the induction of protective immune responses. Glocker, E. O. et al. A homozygous CARD9 mutation in a family with susceptibility to fungal infections. N. Engl. J. Med. 361, 1727–1735 (2009). The first study to show an association between susceptibility to chronic mucocutaneous candidiasis and homozygous mutations in CARD9. Puel, A. et al. Inborn errors of mucocutaneous immunity to Candida albicans in humans: a role for IL‑17 cytokines? Curr. Opin. Immunol. 22, 467–474 (2010). Robinson, M. J. et al. Dectin‑2 is a Syk-coupled pattern recognition receptor crucial for Th17 responses to fungal infection. J. Exp. Med. 206, 2037–2051 (2009). Saijo, S. et al. Dectin‑2 recognition of α-mannans and induction of Th17 cell differentiation is essential for host defense against Candida albicans. Immunity 32, 681–691 (2010). Yamasaki, S. et al. C‑type lectin Mincle is an activating receptor for pathogenic fungus, Malassezia. Proc. Natl Acad. Sci. USA 106, 1897–1902 (2009). Wells, C. A. et al. The macrophage-inducible C‑type lectin, mincle, is an essential component of the innate immune response to Candida albicans. J. Immunol. 180, 7404–7413 (2008). Lam, J. S., Huang, H. & Levitz, S. M. Effect of differential N‑linked and O‑linked mannosylation on recognition of fungal antigens by dendritic cells. PLoS ONE 2, e1009 (2007). Cambi, A. et al. Dendritic cell interaction with Candida albicans critically depends on N‑linked mannan. J. Biol. Chem. 283, 20590–20599 (2008). van de Veerdonk, F. L. et al. The macrophage mannose receptor induces IL‑17 in response to Candida albicans. Cell Host Microbe 5, 329–340 (2009). This study identifies the macrophage mannose receptor as the specific PRR that triggers TH17 cell responses to C. albicans in humans. Tateno, H. et al. Dual specificity of Langerin to sulfated and mannosylated glycans via a single C‑type carbohydrate recognition domain. J. Biol. Chem. 285, 6390–6400 (2010). Heinsbroek, S. E. et al. Stage-specific sampling by pattern recognition receptors during Candida albicans phagocytosis. PLoS Pathog. 4, e1000218 (2008). Dan, J. M., Kelly, R. M., Lee, C. K. & Levitz, S. M. Role of the mannose receptor in a murine model of Cryptococcus neoformans infection. Infect. Immun. 76, 2362–2367 (2008). Gales, A. et al. PPARγ controls dectin‑1 expression required for host antifungal defense against Candida albicans. PLoS Pathog. 6, e1000714 (2010).
286 | APRIL 2011 | VOLUME 11
34. Bochud, P. Y. et al. Toll-like receptor 4 polymorphisms and aspergillosis in stem-cell transplantation. N. Engl. J. Med. 359, 1766–1777 (2008). This report provides strong evidence for the association between a donor TLR4 polymorphism and the risk of invasive aspergillosis among recipients of haematopoietic cell transplants from unrelated donors. 35. Carvalho, A. et al. Polymorphisms in Toll-like receptor genes and susceptibility to infections in allogeneic stem cell transplantation. Exp. Hematol. 37, 1022–1029 (2009). 36. Carvalho, A. et al. Polymorphisms in Toll-like receptor genes and susceptibility to pulmonary aspergillosis. J. Infect. Dis. 197, 618–621 (2008). 37. Van der Graaf, C. A. et al. Toll-like receptor 4 Asp299Gly/Thr399Ile polymorphisms are a risk factor for Candida bloodstream infection. Eur. Cytokine Netw. 17, 29–34 (2006). 38. Blander, J. M. & Medzhitov, R. On regulation of phagosome maturation and antigen presentation. Nature Immunol. 7, 1029–1035 (2006). 39. Moretti, S. et al. The contribution of PARs to inflammation and immunity to fungi. Mucosal Immunol. 1, 156–168 (2008). This study shows that recognition of fungi occurs through a dual sensor system involving crosstalk between TLRs that sense fungal PAMPs and PARs that sense fungal virulence. 40. Gottar, M. et al. Dual detection of fungal infections in Drosophila via recognition of glucans and sensing of virulence factors. Cell 127, 1425–1437 (2006). 41. Ellis, J. G., Dodds, P. N. & Lawrence, G. J. The role of secreted proteins in diseases of plants caused by rust, powdery mildew and smut fungi. Curr. Opin. Microbiol. 10, 326–331 (2007). 42. Joly, S. et al. Cutting edge: Candida albicans hyphae formation triggers activation of the Nlrp3 inflammasome. J. Immunol. 183, 3578–3581 (2009). 43. Said-Sadier, N., Padilla, E., Langsley, G. & Ojcius, D. M. Aspergillus fumigatus stimulates the NLRP3 inflammasome through a pathway requiring ROS production and the Syk tyrosine kinase. PLoS ONE 5, e10008 (2010). 44. Gross, O. et al. Card9 controls a non-TLR signalling pathway for innate anti-fungal immunity. Nature 442, 651–656 (2006). This study defines a new innate immune pathway involving CARD9 as a key transducer of dectin 1 signalling and mediator of innate antifungal immunity. 45. Gross, O. et al. Syk kinase signalling couples to the Nlrp3 inflammasome for anti-fungal host defence. Nature 459, 433–436 (2009). This study describes the molecular basis for IL‑1β production in candidiasis and identifies a crucial function for the NLRP3 inflammasome in antifungal host defence in vivo. 46. Zaki, M. H. et al. The NLRP3 inflammasome protects against loss of epithelial integrity and mortality during experimental colitis. Immunity 32, 379–391 (2010). 47. Rehaume, L. M., Jouault, T. & Chamaillard, M. Lessons from the inflammasome: a molecular sentry linking Candida and Crohn’s disease. Trends Immunol. 31, 171–175 (2010).
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS 48. Bianchi, M. E. DAMPs, PAMPs and alarmins: all we need to know about danger. J. Leukoc. Biol. 81, 1–5 (2007). 49. Sorci, G. et al. The danger signal S100B integrates pathogen- and danger-sensing pathways to restrain inflammation. PLoS Pathog. (in the press). 50. Tsitsigiannis, D. I. & Keller, N. P. Oxylipins as developmental and host–fungal communication signals. Trends Microbiol. 15, 109–118 (2007). 51. Haas-Stapleton, E. J. et al. Candida albicans modulates host defense by biosynthesizing the pro-resolving mediator resolvin E1. PLoS ONE 2, e1316 (2007). 52. Rappleye, C. A. & Goldman, W. E. Fungal stealth technology. Trends Immunol. 29, 18–24 (2008). 53. Levitz, S. M. Innate recognition of fungal cell walls. PLoS Pathog. 6, e1000758 (2010). 54. Rappleye, C. A., Eissenberg, L. G. & Goldman, W. E. Histoplasma capsulatum α‑(1,3)-glucan blocks innate immune recognition by the β-glucan receptor. Proc. Natl Acad. Sci. USA 104, 1366–1370 (2007). This work highlights that α‑(1,3)-glucans contribute to the pathogenesis of histoplasmosis by concealing immunostimulatory β‑glucans from dectin 1. 55. Aimanianda, V. et al. Surface hydrophobin prevents immune recognition of airborne fungal spores. Nature 460, 1117–1121 (2009). This work shows that surface hydrophobins on dormant airborne fungal conidia mask their recognition by the immune system and prevent immune responses. 56. Stringer, J. R. Antigenic variation in Pneumocystis. J. Eukaryot. Microbiol. 54, 8–13 (2007). 57. Vecchiarelli, A. Fungal capsular polysaccharide and T‑cell suppression: the hidden nature of poor immunogenicity. Crit. Rev. Immunol. 27, 547–557 (2007). 58. Alvarez, M. & Casadevall, A. Phagosome extrusion and host-cell survival after Cryptococcus neoformans phagocytosis by macrophages. Curr. Biol. 16, 2161–2165 (2006). 59. Charlier, C. et al. Evidence of a role for monocytes in dissemination and brain invasion by Cryptococcus neoformans. Infect. Immun. 77, 120–127 (2009). 60. Yoshitomi, H. et al. A role for fungal β-glucans and their receptor Dectin‑1 in the induction of autoimmune arthritis in genetically susceptible mice. J. Exp. Med. 201, 949–960 (2005). 61. de Koning, H. D. et al. A comprehensive analysis of pattern recognition receptors in normal and inflamed human epidermis: upregulation of Dectin‑1 in psoriasis. J. Invest. Dermatol. 130, 2611–2620 (2010). 62. Manicassamy, S. et al. Toll-like receptor 2‑dependent induction of vitamin A-metabolizing enzymes in dendritic cells promotes T regulatory responses and inhibits autoimmunity. Nature Med. 15, 401–409 (2009). 63. Rizzetto, L. & Cavalieri, D. A systems biology approach to the mutual interaction between yeast and the immune system. Immunobiology 215, 762–769 (2010). 64. Huang, H. et al. Distinct patterns of dendritic cell cytokine release stimulated by fungal β-glucans and Toll-like receptor agonists. Infect. Immun. 77, 1774–1781 (2009). 65. Dan, J. M., Wang, J. P., Lee, C. K. & Levitz, S. M. Cooperative stimulation of dendritic cells by Cryptococcus neoformans mannoproteins and CpG oligodeoxynucleotides. PLoS ONE 3, e2046 (2008). This study provides a rationale for combining mannosylated antigens with TLR ligands in the design of immune modulators and vaccines against fungi. 66. Bonifazi, P. et al. Intranasally delivered siRNA targeting PI3K/Akt/mTOR inflammatory pathways protects from aspergillosis. Mucosal Immunol. 3, 193–205 (2010). 67. Bonifazi, P. et al. Balancing inflammation and tolerance in vivo through dendritic cells by the commensal Candida albicans. Mucosal Immunol. 2, 362–374 (2009). 68. Ersland, K., Wuthrich, M. & Klein, B. S. Dynamic interplay among monocyte-derived, dermal, and resident lymph node dendritic cells during the generation of vaccine immunity to fungi. Cell Host Microbe 7, 474–487 (2010). 69. Paul, W. E. & Zhu, J. How are TH2‑type immune responses initiated and amplified? Nature Rev. Immunol. 10, 225–235 (2010). 70. de Luca, A. et al. Non-hematopoietic cells contribute to protective tolerance to Aspergillus fumigatus via a TRIF pathway converging on IDO. Cell. Mol. Immunol. 7, 459–470 (2010).
71. de Oliveira, L. L., Coltri, K. C., Cardoso, C. R., Roque-Barreira, M. C. & Panunto-Castelo, A. T helper 1‑inducing adjuvant protects against experimental paracoccidioidomycosis. PLoS Negl. Trop. Dis. 2, e183 (2008). 72. Nesbit, L., Johnson, S. M., Pappagianis, D. & Ampel, N. M. Polyfunctional T lymphocytes are in the peripheral blood of donors naturally immune to coccidioidomycosis and are not induced by dendritic cells. Infect. Immun. 78, 309–315 (2010). 73. Zhang, Y. et al. Robust Th1 and Th17 immunity supports pulmonary clearance but cannot prevent systemic dissemination of highly virulent Cryptococcus neoformans H99. Am. J. Pathol. 175, 2489–2500 (2009). 74. Spellberg, B. et al. Antibody titer threshold predicts anti-candidal vaccine efficacy even though the mechanism of protection is induction of cell-mediated immunity. J. Infect. Dis. 197, 967–971 (2008). 75. Yauch, L. E., Lam, J. S. & Levitz, S. M. Direct inhibition of T‑cell responses by the Cryptococcus capsular polysaccharide glucuronoxylomannan. PLoS Pathog. 2, e120 (2006). 76. Pinzan, C. F., Ruas, L. P., Casabona-Fortunato, A. S., Carvalho, F. C. & Roque-Barreira, M. C. Immunological basis for the gender differences in murine Paracoccidioides brasiliensis infection. PLoS ONE 5, e10757 (2010). 77. Bozzi, A. et al. Analysis of memory T cells in the human paracoccidioidomycosis before and during chemotherapy treatment. Immunol. Lett. 114, 23–30 (2007). 78. Brasch, J. Pathogenesis of tinea. J. Dtsch Dermatol. Ges. 8, 780–786 (2010). 79. Muller, U. et al. IL‑13 induces disease-promoting type 2 cytokines, alternatively activated macrophages and allergic inflammation during pulmonary infection of mice with Cryptococcus neoformans. J. Immunol. 179, 5367–5377 (2007). 80. Szymczak, W. A. & Deepe, G. S. Jr. The CCL7–CCL2– CCR2 axis regulates IL‑4 production in lungs and fungal immunity. J. Immunol. 183, 1964–1974 (2009). 81. Kreindler, J. L. et al. Vitamin D3 attenuates Th2 responses to Aspergillus fumigatus mounted by CD4+ T cells from cystic fibrosis patients with allergic bronchopulmonary aspergillosis. J. Clin. Invest. 120, 3242–3254 (2010). This study suggests that the induction of TReg cells by vitamin D3 may prevent or treat ABPA in patients with cystic fibrosis. 82. Kling, H. M. et al. Relationship of Pneumocystis jiroveci humoral immunity to prevention of colonization and chronic obstructive pulmonary disease in a primate model of HIV infection. Infect. Immun. 78, 4320–4330 (2010). 83. Guimaraes, A. J., Frases, S., Gomez, F. J., ZancopeOliveira, R. M. & Nosanchuk, J. D. Monoclonal antibodies to heat shock protein 60 alter the pathogenesis of Histoplasma capsulatum. Infect. Immun. 77, 1357–1367 (2009). 84. Subramaniam, K. S. et al. The absence of serum IgM enhances the susceptibility of mice to pulmonary challenge with Cryptococcus neoformans. J. Immunol. 184, 5755–5767 (2010). 85. Shi, L. et al. A monoclonal antibody to Histoplasma capsulatum alters the intracellular fate of the fungus in murine macrophages. Eukaryot. Cell 7, 1109–1117 (2008). 86. McClelland, E. E., Nicola, A. M., Prados-Rosales, R. & Casadevall, A. Ab binding alters gene expression in Cryptococcus neoformans and directly modulates fungal metabolism. J. Clin. Invest. 120, 1355–1361 (2010). This paper describes a new mode of action for antibody-mediated immunity based on the modulation of microbial gene expression and metabolism. 87. Rapaka, R. R. et al. Conserved natural IgM antibodies mediate innate and adaptive immunity against the opportunistic fungus Pneumocystis murina. J. Exp. Med. 207, 2907–2919 (2010). 88. Ostrosky-Zeichner, L., Casadevall, A., Galgiani, J. N., Odds, F. C. & Rex, J. H. An insight into the antifungal pipeline: selected new molecules and beyond. Nature Rev. Drug Discov. 9, 719–727 (2010). 89. Casadevall, A. & Pirofski, L. A. A reappraisal of humoral immunity based on mechanisms of antibodymediated protection against intracellular pathogens. Adv. Immunol. 91, 1–44 (2006).
NATURE REVIEWS | IMMUNOLOGY
90. Vinh, D. C., Masannat, F., Dzioba, R. B., Galgiani, J. N. & Holland, S. M. Refractory disseminated coccidioidomycosis and mycobacteriosis in interferon-γ receptor 1 deficiency. Clin. Infect. Dis. 49, e62–e65 (2009). 91. Hardison, S. E., Wozniak, K. L., Kolls, J. K. & Wormley, F. L. Jr. Interleukin‑17 is not required for classical macrophage activation in a pulmonary mouse model of Cryptococcus neoformans infection. Infect. Immun. 78, 5341–5351 (2010). 92. Murdock, B. J. et al. Coevolution of TH1, TH2, and TH17 responses during repeated pulmonary exposure to Aspergillus fumigatus conidia. Infect. Immun. 79, 125–135 (2010). 93. Zelante, T. et al. IL‑23 and the Th17 pathway promote inflammation and impair antifungal immune resistance. Eur. J. Immunol. 37, 2695–2706 (2007). 94. Chamilos, G. et al. Generation of IL‑23 producing dendritic cells (DCs) by airborne fungi regulates fungal pathogenicity via the induction of TH‑17 responses. PLoS ONE 5, e12955 (2010). 95. Deepe, G. S. Jr & Gibbons, R. S. Interleukins 17 and 23 influence the host response to Histoplasma capsulatum. J. Infect. Dis. 200, 142–151 (2009). 96. Kroetz, D. N. & Deepe, G. S. Jr. CCR5 dictates the equilibrium of proinflammatory IL‑17+ and regulatory Foxp3+ T cells in fungal infection. J. Immunol. 184, 5224–5231 (2010). 97. Armstrong-James, D. P. et al. Impaired interferon-γ responses, increased interleukin‑17 expression, and a tumor necrosis factor-α transcriptional program in invasive aspergillosis. J. Infect. Dis. 200, 1341–1351 (2009). 98. De Luca, A. et al. Functional yet balanced reactivity to Candida albicans requires TRIF, MyD88, and IDOdependent inhibition of Rorc. J. Immunol. 179, 5999–6008 (2007). 99. De Luca, A. et al. IL‑22 defines a novel immune pathway of antifungal resistance. Mucosal Immunol. 3, 361–373 (2010). This work defines a staged response in mucosal candidiasis, involving an early IL‑22‑dominated response that provides antifungal resistance, followed by a TH1 or TReg cell response that prevents fungal dissemination and provides immunological memory. 100. Kleinschek, M. A. et al. IL‑23 enhances the inflammatory cell response in Cryptococcus neoformans infection and induces a cytokine pattern distinct from IL‑12. J. Immunol. 176, 1098–1106 (2006). 101. Leibundgut-Landmann, S. et al. Syk- and CARD9‑dependent coupling of innate immunity to the induction of T helper cells that produce interleukin 17. Nature Immunol. 8, 630–638 (2007). 102. Acosta-Rodriguez, E. V. et al. Surface phenotype and antigenic specificity of human interleukin 17‑ producing T helper memory cells. Nature Immunol. 8, 639–646 (2007). 103. Bozza, S. et al. Immune sensing of Aspergillus fumigatus proteins, glycolipids, and polysaccharides and the impact on Th immunity and vaccination. J. Immunol. 183, 2407–2414 (2009). In contrast to reference 54, this study shows that A. fumigatus α‑(1,3)-glucans induce immune activation in aspergillosis. Therefore, evasion of this recognition may contribute to the pathogenic potential of dimorphic fungal pathogens but not opportunistic fungi. 104. Chai, L. Y. et al. Anti-Aspergillus human host defence relies on type 1 T helper (Th1), rather than type 17 T helper (Th17), cellular immunity. Immunology 130, 46–54 (2010). 105. Wuthrich, M. et al. Vaccine-induced protection against 3 systemic mycoses endemic to North America requires Th17 cells in mice. J. Clin. Invest. 121, 554–568 (2010). 106. Ryan, K. R. et al. Impaired dendritic cell maturation and cytokine production in patients with chronic mucocutanous candidiasis with or without APECED. Clin. Exp. Immunol. 154, 406–414 (2008). 107. Conti, H. R. et al. Th17 cells and IL‑17 receptor signaling are essential for mucosal host defense against oral candidiasis. J. Exp. Med. 206, 299–311 (2009). This study shows the importance of the TH17 cell pathway in the control of oropharyngeal candidiasis. The work in reference 99 describes that TH17 cells are not essential in gastric candidiasis. These studies raise the question as to whether the activity of the TH17 cell subset is compartmentalized.
VOLUME 11 | APRIL 2011 | 287 © 2011 Macmillan Publishers Limited. All rights reserved
REVIEWS 108. Huang, W., Na, L., Fidel, P. L. & Schwarzenberger, P. Requirement of interleukin‑17A for systemic antiCandida albicans host defense in mice. J. Infect. Dis. 190, 624–631 (2004). 109. Lin, L. et al. Th1‑Th17 cells mediate protective adaptive immunity against Staphylococcus aureus and Candida albicans infection in mice. PLoS Pathog. 5, e1000703 (2009). 110. Cheng, S. C. et al. Candida albicans dampens host defense by downregulating IL‑17 production. J. Immunol. 185, 2450–2457 (2010). 111. Loures, F. V., Pina, A., Felonato, M. & Calich, V. L. TLR2 is a negative regulator of Th17 cells and tissue pathology in a pulmonary model of fungal infection. J. Immunol. 183, 1279–1290 (2009). 112. Ahlgren, K. M. et al. Increased IL‑17A secretion in response to Candida albicans in autoimmune polyendocrine syndrome type 1 and its animal model. Eur. J. Immunol. 41, 235–245 (2011). 113. Romani, L. & Puccetti, P. Protective tolerance to fungi: the role of IL‑10 and tryptophan catabolism. Trends Microbiol. 14, 183–189 (2006). 114. Ferreira, M. C., de Oliveira, R. T., da Silva, R. M., Blotta, M. H. & Mamoni, R. L. Involvement of regulatory T cells in the immunosuppression characteristic of patients with paracoccidioidomycosis. Infect. Immun. 78, 4392–4401 (2010). 115. Dillon, S. et al. Yeast zymosan, a stimulus for TLR2 and dectin‑1, induces regulatory antigen-presenting cells and immunological tolerance. J. Clin. Invest. 116, 916–928 (2006). 116. Baker, B. S. The role of microorganisms in atopic dermatitis. Clin. Exp. Immunol. 144, 1–9 (2006). 117. Romani, L., Zelante, T., De Luca, A., Fallarino, F. & Puccetti, P. IL‑17 and therapeutic kynurenines in pathogenic inflammation to fungi. J. Immunol. 180, 5157–5162 (2008). 118. Zelante, T., Fallarino, F., Bistoni, F., Puccetti, P. & Romani, L. Indoleamine 2,3‑dioxygenase in infection: the paradox of an evasive strategy that benefits the host. Microbes Infect. 11, 133–141 (2009). 119. Grohmann, U. et al. Reverse signaling through GITR ligand enables dexamethasone to activate IDO in allergy. Nature Med. 13, 579–586 (2007). This paper indicates that the induction of IDO expression could be an important mechanism underlying the anti-inflammatory action of corticosteroids in fungal allergy. 120. Wolk, K. & Sabat, R. Interleukin‑22: a novel T‑ and NK‑cell derived cytokine that regulates the biology of tissue cells. Cytokine Growth Factor Rev. 17, 367–380 (2006). 121. Pickert, G. et al. STAT3 links IL‑22 signaling in intestinal epithelial cells to mucosal wound healing. J. Exp. Med. 206, 1465–1472 (2009). 122. Milner, J. D. et al. Impaired TH17 cell differentiation in subjects with autosomal dominant hyper-IgE syndrome. Nature 452, 773–776 (2008). This work reports that TH17 cell responses to C. albicans are defective in patients with autosomal dominant hyper-IgE syndrome. 123. Yano, J., Lilly, E., Barousse, M. & Fidel, P. L. Jr. Epithelial cell-derived S100 calcium-binding proteins as key mediators in the hallmark acute neutrophil response during Candida vaginitis. Infect. Immun. 78, 5126–5137 (2010). 124. Liu, Y. et al. Memory IL‑22‑producing CD4+ T cells specific for Candida albicans are present in humans. Eur. J. Immunol. 39, 1472–1479 (2009). 125. Eyerich, K., Eyerich, S., Hiller, J., Behrendt, H. & Traidl-Hoffmann, C. Chronic mucocutaneous candidiasis, from bench to bedside. Eur. J. Dermatol. 20, 260–265 (2010). 126. Bjeldanes, L. F., Kim, J. Y., Grose, K. R., Bartholomew, J. C. & Bradfield, C. A. Aromatic hydrocarbon responsiveness-receptor agonists generated from indole‑3‑carbinol in vitro and in vivo: comparisons with 2,3,7,8‑tetrachlorodibenzo‑pdioxin. Proc. Natl Acad. Sci. USA 88, 9543–9547 (1991). 127. Trifari, S., Kaplan, C. D., Tran, E. H., Crellin, N. K. & Spits, H. Identification of a human helper T cell population that has abundant production of interleukin 22 and is distinct from TH17, TH1 and TH2 cells. Nature Immunol. 10, 864–871 (2009). 128. Heath-Pagliuso, S. et al. Activation of the Ah receptor by tryptophan and tryptophan metabolites. Biochemistry 37, 11508–11515 (1998). 129. Zelante, T., De Luca, A., D’Angelo, C., Moretti, S. & Romani, L. IL‑17/Th17 in anti-fungal immunity: what’s new? Eur. J. Immunol. 39, 645–648 (2009).
130. Harrington, L. et al. Deficiency of indoleamine 2,3‑dioxygenase enhances commensal-induced antibody responses and protects against Citrobacter rodentium-induced colitis. Infect. Immun. 76, 3045–3053 (2008). 131. Kagami, S., Rizzo, H. L., Kurtz, S. E., Miller, L. S. & Blauvelt, A. IL‑23 and IL‑17A, but not IL‑12 and IL‑22, are required for optimal skin host defense against Candida albicans. J. Immunol. 185, 5453–5462 (2010). 132. Cassone, A. Fungal vaccines: real progress from real challenges. Lancet Infect. Dis. 8, 114–124 (2008). 133. Cutler, J. E., Deepe, G. S. Jr & Klein, B. S. Advances in combating fungal diseases: vaccines on the threshold. Nature Rev. Microbiol. 5, 13–28 (2007). References 132 and 133 are comprehensive reviews summarizing how the elucidation of the mechanisms of protective immunity against fungal diseases has renewed interest in the development of vaccines against the mycoses. 134. Mezger, M., Einsele, H. & Loeffler, J. Genetic susceptibility to infections with Aspergillus fumigatus. Crit. Rev. Microbiol. 36, 168–177 (2010). 135. Legrand, F. et al. Adjuvant corticosteroid therapy for chronic disseminated candidiasis. Clin. Infect. Dis. 46, 696–702 (2008). 136. Romani, L. et al. Defective tryptophan catabolism underlies inflammation in mouse chronic granulomatous disease. Nature 451, 211–215 (2008). The first direct demonstration of the causal link between defective tryptophan catabolism and susceptibility to fungal infections owing to uncontrolled inflammatory responses. 137. Singh, N. & Perfect, J. R. Immune reconstitution syndrome and exacerbation of infections after pregnancy. Clin. Infect. Dis. 45, 1192–1199 (2007). A comprehensive overview of clinical conditions in which immunological recovery and an imbalance characterized by either suboptimal or excessive immune responses can also be harmful to the host by adversely affecting the resolution of infection. 138. Mezger, M. et al. Polymorphisms in the chemokine (C‑X‑C motif) ligand 10 are associated with invasive aspergillosis after allogeneic stem-cell transplantation and influence CXCL10 expression in monocyte-derived dendritic cells. Blood 111, 534–536 (2008). 139. Plantinga, T. S. et al. Genetic variation of innate immune genes in HIV-infected african patients with or without oropharyngeal candidiasis. J. Acquir. Immune Defic. Syndr. 55, 87–94 (2010). 140. Cunha, C. et al. Dectin‑1 Y238X polymorphism associates with susceptibility to invasive aspergillosis in hematopoietic transplantation through impairment of both recipient- and donor-dependent mechanisms of antifungal immunity. Blood 116, 5394–5402 (2010). 141. Plantinga, T. S. et al. Early stop polymorphism in human DECTIN‑1 is associated with increased Candida colonization in hematopoietic stem cell transplant recipients. Clin. Infect. Dis. 49, 724–732 (2009). 142. Jurevic, R. J., Bai, M., Chadwick, R. B., White, T. C. & Dale, B. A. Single-nucleotide polymorphisms (SNPs) in human β-defensin 1: high-throughput SNP assays and association with Candida carriage in type I diabetics and nondiabetic controls. J. Clin. Microbiol. 41, 90–96 (2003). 143. Sambatakou, H., Pravica, V., Hutchinson, I. V. & Denning, D. W. Cytokine profiling of pulmonary aspergillosis. Int. J. Immunogenet. 33, 297–302 (2006). 144. Sainz, J., Perez, E., Gomez-Lopera, S. & Jurado, M. IL1 gene cluster polymorphisms and its haplotypes may predict the risk to develop invasive pulmonary aspergillosis and modulate C‑reactive protein level. J. Clin. Immunol. 28, 473–485 (2008). 145. Choi, E. H. et al. Association between chronic disseminated candidiasis in adult acute leukemia and common IL4 promoter haplotypes. J. Infect. Dis. 187, 1153–1156 (2003). 146. Babula, O. et al. Frequency of interleukin‑4 (IL‑4) ‑589 gene polymorphism and vaginal concentrations of IL‑4, nitric oxide, and mannose-binding lectin in women with recurrent vulvovaginal candidiasis. Clin. Infect. Dis. 40, 1258–1262 (2005). 147. Bozzi, A., Reis, B. S., Pereira, P. P., Pedroso, E. P. & Goes, A. M. Interferon-γ and interleukin‑4 single nucleotide gene polymorphisms in paracoccidioidomycosis. Cytokine 48, 212–217 (2009).
288 | APRIL 2011 | VOLUME 11
148. Knutsen, A. P., Kariuki, B., Consolino, J. D. & Warrier, M. R. IL‑4 alpha chain receptor (IL‑4Rα) polymorphisms in allergic bronchopulmonary aspergillosis. Clin. Mol. Allergy 4, 3 (2006). 149. Sainz, J. et al. Interleukin‑10 promoter polymorphism as risk factor to develop invasive pulmonary aspergillosis. Immunol. Lett. 109, 76–82 (2007). 150. Seo, K. W. et al. Protective role of interleukin‑10 promoter gene polymorphism in the pathogenesis of invasive pulmonary aspergillosis after allogeneic stem cell transplantation. Bone Marrow Transplant. 36, 1089–1095 (2005). 151. Brouard, J. et al. Influence of interleukin‑10 on Aspergillus fumigatus infection in patients with cystic fibrosis. J. Infect. Dis. 191, 1988–1991 (2005). 152. Carvalho, A. et al. Prognostic significance of genetic variants in the IL‑23/Th17 pathway for the outcome of T cell-depleted allogeneic stem cell transplantation. Bone Marrow Transplant. 45, 1645–1652 (2010). 153. Granell, M. et al. Mannan-binding lectin pathway deficiencies and invasive fungal infections following allogeneic stem cell transplantation. Exp. Hematol. 34, 1435–1441 (2006). 154. Donders, G. G., Babula, O., Bellen, G., Linhares, I. M. & Witkin, S. S. Mannose-binding lectin gene polymorphism and resistance to therapy in women with recurrent vulvovaginal candidiasis. BJOG 115, 1225–1231 (2008). 155. Kaur, S. et al. Elevated levels of mannan-binding lectin (MBL) and eosinophilia in patients of bronchial asthma with allergic rhinitis and allergic bronchopulmonary aspergillosis associate with a novel intronic polymorphism in MBL. Clin. Exp. Immunol. 143, 414–419 (2006). 156. Lev-Sagie, A. et al. Polymorphism in a gene coding for the inflammasome component NALP3 and recurrent vulvovaginal candidiasis in women with vulvar vestibulitis syndrome. Am. J. Obstet. Gynecol. 200, 303.e1–303.e6 (2009). 157. Zaas, A. K. et al. Plasminogen alleles influence susceptibility to invasive aspergillosis. PLoS Genet. 4, e1000101 (2008). 158. Saxena, S., Madan, T., Shah, A., Muralidhar, K. & Sarma, P. U. Association of polymorphisms in the collagen region of SP‑A2 with increased levels of total IgE antibodies and eosinophilia in patients with allergic bronchopulmonary aspergillosis. J. Allergy Clin. Immunol. 111, 1001–1007 (2003). 159. Vaid, M. et al. Distinct alleles of mannose-binding lectin (MBL) and surfactant proteins A (SP‑A) in patients with chronic cavitary pulmonary aspergillosis and allergic bronchopulmonary aspergillosis. Clin. Chem. Lab. Med. 45, 183–186 (2007). 160. Kesh, S. et al. TLR1 and TLR6 polymorphisms are associated with susceptibility to invasive aspergillosis after allogeneic stem cell transplantation. Ann. NY Acad. Sci. 1062, 95–103 (2005). 161. Sainz, J. et al. TNFR1 mRNA expression level and TNFR1 gene polymorphisms are predictive markers for susceptibility to develop invasive pulmonary aspergillosis. Int. J. Immunopathol. Pharmacol. 23, 423–436 (2010). 162. Sainz, J. et al. Variable number of tandem repeats of TNF receptor type 2 promoter as genetic biomarker of susceptibility to develop invasive pulmonary aspergillosis. Hum. Immunol. 68, 41–50 (2007).
Acknowledgements
I thank the large number of researchers who have contributed to this field and whose work was not cited or was cited through the review articles of others because of space limitations. This work is supported by the EU Specific Targeted Research Projects SYBARIS (FP7‑Health‑2009‑single-stage, contract number 242220) and ALLFUN (FP7‑Health‑2010‑singlestage, contract number 260338) and by the Fondazione per la Ricerca sulla Fibrosi Cistica (project number FFC21/2010). I also thank C. Massi Benedetti of the University of Perugia for editorial assistance and my numerous collaborators for their dedicated work in my laboratory.
Competing interests statement
The author declares no competing financial interests.
FURTHER INFORMATION European Project ALLFUN: http://www.altaweb.eu/allfun ALL LINKS ARE ACTIVE IN THE ONLINE PDF
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES OPINION
Are senescence and exhaustion intertwined or unrelated processes that compromise immunity? Arne N. Akbar and Sian M. Henson
Abstract | Can the immune system be reactivated continuously throughout the lifetime of an organism or is there a finite point at which repeated antigenic challenge leads to the loss of lymphocyte function or the cells themselves or both? Replicative senescence and exhaustion are processes that control T cell proliferative activity and function; however, there is considerable confusion over the relationship between these two intrinsic cellular control mechanisms. In this Opinion article, we compare the molecular regulation of senescence and exhaustion in T cells. Available data suggest that both processes are regulated independently of each other and that it may be safer to block exhaustion than senescence to enhance immunity. Repeated antigenic stimulation throughout life may compromise antigen-specific T cells in two ways. First, they may become functionally exhausted and lose essential functional activity that is necessary for immune protection1–4. Second, repeated T cell stimulation can lead to a loss of the replicative capacity of some antigenspecific T cell populations, as a result of telomere erosion and/or unrepaired DNA damage (a process known as replicative senescence)5–7. The terms senescence and exhaustion are often used interchangeably when referring to highly differentiated T cells8–10, but it is not clear what these terms mean from a mechanistic standpoint and whether both processes ultimately lead to decreased immune function. Although there has been substantial progress in identifying the mechanisms that regulate both processes, they are normally investigated independently of each other and it is unclear whether exhausted cells are also senescent and vice versa. An exciting finding is that both exhaustion and senescence are not passive events but are controlled by active signalling processes1,11,12. The key issue is whether blocking pathways
that maintain either the exhausted or the senescent state, or both, can boost immune function, especially in older individuals13. Senescence Cell senescence has been recognized as a biological process since the 1960s14, when fibroblasts were shown to undergo growth arrest after extensive replication. This process was subsequently shown to be due to erosion of telomeres (which are repeating hexameric sequences of nucleotides at the ends of linear chromosomes), and is known as telomere-dependent senescence6,15. In the absence of compensatory factors, telomeres shorten by about 50–100 bases during each replicative cycle, until a critical point at which the exposed DNA end of the telo mere is recognized as a DNA double strand break15. This recruits a complex of proteins that are involved with DNA repair, initiating a process commonly referred to as the DNA damage response (DDR)11,16,17 (FIG. 1). The development of telomere-dependent senescence is delayed if cells can upregulate the expression of the enzyme telomerase, which adds telomeres to the ends of chromosomes15. Many cells with a high proliferative
NATURE REVIEWS | IMMUNOLOGY
rate in vivo, such as germ cells, epithelial cells and T cells, can upregulate telomerase expression5,15. Furthermore, many cancer cells have high levels of telomerase activity, which gives them unlimited proliferative capacity 15,16. Cellular senescence can also occur when DNA is damaged by other means that are independent of telomere erosion, and this is known as telomere-independent senescence16,18. Such telomere-independent mechanisms include damage by reactive oxygen species or ionizing radiation, chromatin perturbation and activation of p53 and stress pathways in response to growth factor deprivation16,18. Most of these situations induce a DDR that is virtually identical to that triggered by critical telo mere erosion and, unless the DNA can be repaired, the cells undergo growth arrest. Most of the work on DDR-triggered signalling pathways that lead to proliferative arrest has been performed in fibroblasts. Nevertheless, as the replicative capacity of human T cells can be regulated by telomeredependent senescence19,20 in a similar manner to that of fibroblasts, the DDR may be similar in both cell types (A.N.A. and T. von Zglinicki, unpublished observations). Although telomere-independent senescence has also been described in T cells, much less information is available on this process21,22, and therefore it will not be discussed further here. There are also considerable differences in the regulation of senescence in rodents and humans23,24. One example is the fact that mice have considerably longer telomeres than humans but have a much shorter lifespan23,24. There are three main phases in the development of cellular senescence (FIG. 1). The first is the induction of this process by telomere-dependent or telomereindependent mechanisms. This is followed by the DDR, which is associated with temp orary growth arrest. However, if the DNA damage cannot be repaired, the signalling processes that prevent cell cycling become fixed and a permanent senescent state ensues16,17 (phase 3). It is currently unclear whether senescence-related DNA damage foci in T cells contain the same clusters of proteins as those found in fibroblasts, and whether the signalling pathways that VOLUME 11 | APRIL 2011 | 289
© 2011 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES 2JCUG
2JCUG
2JCUG
&0#FCOCIGTGURQPUG r1ZKFCVKXGUVTGUUTGURQPUG r&0#FCOCIGHQEK r#6/CPF#64UKIPCNNKPI
+PFWEVKQP r6GNQOGTGUJQTVGPKPI r0QPVGNQOGTKE &0#FCOCIG s415 s+QPK\KPITCFKCVKQP s78 r)TQYVJHCEVQT FGRTKXCVKQP
)TQYVJCTTGUV r5GPGUEGPEG &0#FCOCIGTGURQPUG
#6/
*#:
2
42#
0$5 4#& 4#& 4(% /4' *75 4#& 4#& #6/
#64
R
%*-
#64
R
%*-
/-
R %&% %GNNE[ENKPI
Figure 1 | The three phases of senescence induction. The onset of senescence occurs in three phases. The first is induction, which can be triggered by multiple stimuli, including telomere shortening, nontelomeric DNA damage (for example, by reactive oxygen species (ROS), ionizing radiation or ultra 0CVWTG4GXKGYU^+OOWPQNQI[ violet (UV) light) and the removal of growth factors from cell culture media. These stresses trigger the second phase, a DNA damage response (DDR) that may be initiated at telomeric or non-telomeric DNA loci, and this inhibits cell cycling until the DNA is repaired. The DDR involves the activation of sensor kinases, such as ataxia telangiectasia mutated (ATM) and ATM and RAD3 related (ATR), and the formation of DNA damage foci containing various proteins, such as histone H2AX, replication protein A (RPA; which recruits ATR) and RAD family members. Activation of ATM and ATR by the DRR results in activation of the downstream transducer kinases checkpoint kinase 1 homologue (CHK1) and CHK2, activation of checkpoint proteins, such as p53 and the cyclin-dependent kinase inhibitor p21 (also known as CIP1), and inhibition of M-phase inducer phosphatase (CDC25). Signalling through the mitogen-activated protein kinase p38 may also be important for the development of senescence. If the damaged DNA is repaired, the DNA damage foci disintegrate, the DDR ceases and the cell resumes cell cycling17. However, if the DNA damage is not repaired immediately, this signalling pathway can maintain growth arrest for long periods (7–10 days) and, during this time, inactivation of DNA checkpoint kinases can restore cell cycle progression17. However, prolonged DDR signalling ultimately leads to phase 3 and irreversible senescence. This mechanism has mainly been demonstrated in fibroblasts. MK2, MAPK-activated protein kinase 2; MRE11, meiotic recombination 11 homologue; NBS1, Nijmegen breakage syndrome 1 (also known as nibrin); RFC, replication factor C.
induce growth arrest are identical in both cell types. However, there are known similarities, including signalling through p53 (REF. 25) and the mitogen-activated protein kinase p38 (REF. 26) and the involvement of the cyclin-dependent kinase inhibitors p16 (also known as INK4A)27 and p21 (also known as CIP1)25,27, which have a central role in growth arrest in senescent cells13,28. The data indicating that both telomeredependent and telomere-independent senescence are associated with DNA damage and growth arrest has led to the hypothesis that this process is an anticancer mechanism. In addition, the local control of T cell senescence in tissue microenvironments where active immune responses are occurring may allow for local control of latent or reactivating infections. This mechanism would inhibit the excessive proliferation of memory T cells and thus would not incur the risk of inflammation29.
Which T cells have characteristics of senescence? The induction of senescence in human primary T cells may limit the long-term persistence of specific immune responses during ageing 13,30,31. This is because thymic involution dictates that the maintenance of existing memory T cell pools throughout life has to involve periodic episodes of proliferation rather than the continuous input of new naive T cells32,33. Thus, persistent viruses may induce senescence in specific T cell populations that are induced to proliferate repeatedly throughout life9,31. Several studies have attempted to define senescent human T cells by investigating T cell populations that have been repeatedly activated and cultured for long periods in vitro34,35. However, these cultured cells have been selected for their ability to persist after long-term culture. It is not surprising, therefore, that they appear to be more resistant to apoptosis34 than newly isolated highly
290 | APRIL 2011 | VOLUME 11
differentiated T cells36. Thus, the extrapo lation of data obtained from long-term cultured T cells and ex vivo-derived populations has to be performed with caution. Human T cells at early and late stages of differentiation can be identified by their expression of various cell surface markers37,38 (TABLE 1). Multiple combinations of cell surface markers can be used to discriminate between T cells, and the relative expression levels of CD27 and CD28 can be used to distinguish undifferentiated from differentiated T cells9,20,39. Undifferentiated CD4+ T cells express both CD27 and CD28, but after repeated activation they first lose CD27 expression (CD27–CD28+; inter mediate differentiation) and then CD28 expression, eventually giving rise to a highly differentiated CD27–CD28– population of cells9,20,39. Undifferentiated and differentiated CD8+ T cells are also CD27+CD28+ and CD27–CD28–, respectively, but these cells lose CD28 expression before CD27 expression during differentiation32,38. As T cells differentiate from a CD27+CD28+ to a CD27–CD28– phenotype, their telomeres shorten and telomerase activity is progressively lost (TABLE 1). The loss of telomerase activity in primary human T cells is associated with altered phosphorylation of telomerase reverse transcriptase (TERT; the catalytic component of telomerase) by the kinase AKT and altered translocation of TERT to the nucleus5,6,20. However, the transduction of T cells with TERT prevents excessive telo mere loss and delays the development of replicative senescence after activation19,20,30,40. Nevertheless, TERT-transduced T cells are still susceptible to telomere-independent senescence, which possibly results from unrepaired cumulative DNA damage that ultimately leads to growth arrest21. In fibroblasts, telomere-dependent senescence does not lead to cell death, and senescent cells are viable and metabolically active and can persist in a nonproliferative state in culture for long periods41. Long-term cultured T cells that are close to senescence are also resistant to cell death in vitro34. However, as discussed above, these cells have been selected for their ability to survive in vitro and may not be representative of freshly isolated highly differentiated T cells, which exhibit many characteristics of senescence but are prone to apoptosis9,20. This indicates that freshly isolated CD4+ and CD8+ T cells that are highly differentiated and thus have lost expression of CD27 and CD28 (CD27–CD28– cells) may be at an earlier stage of senescence than fully senescent www.nature.com/reviews/immunol
© 2011 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES fibroblasts9,20 or long-term cultured T cells34. Although susceptible to apoptosis, it is possible that these CD27–CD28– T cells may persist in vivo in the presence of appropriate survival signals42. This has functional implications for immunity during ageing 43 and during chronic infections, which are associated with the accumulation of large numbers of CD27–CD28– T cells9,10,39,44–47. Are highly differentiated ‘pre-senescent’ T cells functional? The short answer to this question is yes. Highly differentiated T cells that are close to senescence (either CD27–CD28– cells or similar populations identified by other markers) (TABLE 1) can secrete high levels of cytokines such as interferon‑γ (IFNγ) and tumour necrosis factor (TNF). Moreover, highly differentiated CD8+ T cells also express high levels of granzyme B and perforin, indicating that they have the potential to mediate high cytotoxic activity 9,20,38,44,45. In addition, individual T cells within the highly differentiated CD27–CD28– population can simultan eously carry out multiple functions (such as secretion of interleukin‑2 (IL‑2), IFNγ and TNF, and expression of CD40 ligand) to the same extent as less differentiated memory T cell populations48. However, these cells have a reduced capacity to replicate after activation9,20. Interestingly, previous studies that described the susceptibility of effector T cells to apoptosis may not have considered that cell death may have been initiated by senescence-related signalling and not by the induction of apoptosis per se49. Can senescence-related growth arrest be reversed? Growth arrest in fibroblasts that are at an early stage of senescence can be reversed by blocking key mediators (such as p53 (REFS 12, 50), p38 (REFS 12, 51) and p21 (REFS 25, 27)), DDR proteins (such as ataxia telangiectasia mutated (ATM)) or the cell cycle arrest protein checkpoint kinase 2 homologue (CHK2)16,18. Prolonged DDR signalling ultimately leads to irreversible senescence, which may involve continuous signalling through feedback loops12. However, there are two caveats to consider. First, although blocking senescence pathways may lead to enhanced proliferative activity, the T cells may not remain functional. For example, p38 signalling is involved in the induction of senescence26 but is also required for certain T cell functions, such as cytokine secretion52. Therefore, enhancing the proliferation of pre-senescent T cells by p38 blockade may reduce the secretion of IFNγ and TNF51. The second caveat is that because the
Table 1 | Characteristics of differentiation, senescence and exhaustion Characteristic
Early Intermediate Late differentiation differentiation differentiation
References
Differentiation markers CD45RA
+++
+/–
+/–
9,13,21,37–39,48
CD27
+++
+/–
–
9,13,21,37–39,48
CD28
+++
+/–
–
9,13,21,37–39,48
CCR7
+++
++
–
37–39,48
CD57
+
++
+++
13,37 13,88
Functional characteristics BCL-2
+++
++
+
AKT*
+++
++
–
Cytotoxicity (granzyme B and perforin)
+
++
+++
3,19,21,37,59
Proliferation
+++
++
+/–
3,19,21,59
IL-2
+++
+
–
IFNγ
+
++
+++
3,13,21,37
TNF
+
++
+++
3,13,39
21,82
3,13,21,37,39
Senescence characteristics Telomere length
+++
++
+
6,13,15,19–21
Telomerase
+++
++
–
5,13,15,19–21
KLRG1
+
++
+++
PD1
+
+++
++
CTLA4
++
+++
+
TIM3
–
+
++
69,71–74
LAG3
+
+
+
4,60,69,74
BIM
+
++
+++
55,74,85,86,88
BLIMP1
–
+
+++
55,74,87,88
8,82–84
Exhaustion markers 1–4,62,65,69,74,75 4,62,69,70,74,75
BCL-2, B cell lymphoma 2; BIM, BCL‑2-interacting mediator of cell death; BLIMP1, B lymphocyte-induced maturation protein 1; CCR7, CC-chemokine receptor 7; CTLA4, cytotoxic T lymphocyte antigen 4; IFNγ, interferon-γ; IL-2, interleukin-2, KLRG1, killer cell lectin-like receptor subfamily G, member 1; LAG3, lymphocyte activation gene 3; PD1, programmed cell death 1; TIM3, T cell immunoglobulin domain and mucin domain protein 3; TNF, tumour necrosis factor. *Phosphorylation on Ser473.
DDR and p53 pathways are considered to be a first line of defence against cancer 28, blocking these proliferative checkpoints is associated with a risk of malignancy owing to the propagation of T cells with DNA damage. Exhaustion Exhaustion is characterized by the progressive loss of T cell function, leading to the deletion of the exhausted cell. Functional exhaustion develops when there is a high antigenic load, and it was first described in mice during chronic infection with lympho cytic choriomeningitis virus (LCMV), when virus-specific CD8+ T cells were found to lack effector functions53. Non-functional antigen-specific CD8+ T cells have also been observed during infection with simian
NATURE REVIEWS | IMMUNOLOGY
immunodeficiency virus (SIV)54, HIV4, hepatitis B virus (HBV)46, hepatitis C virus (HCV)55,56 and human T lymphotropic virus 1 (HTLV1)57. In addition, nonfunctional T cells have been observed in patients with a high tumour burden58. A feature of functional exhaustion is that it affects many antiviral properties of both mouse and human CD8+ T cells, and loss of distinct T cell functions occurs in a hierarchical manner 59. IL‑2 production and robust proliferation are the first functions to be lost, whereas TNF production is lost later. Cytotoxic activity is also lacking in exhausted human CD8+ T cells. However, although previous studies have shown that exhausted T cells in mice have defective cytotoxicity 1,59, a recent study has suggested VOLUME 11 | APRIL 2011 | 291
© 2011 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES limited defects in cytotoxicity in these cells60. At a severe stage of exhaustion, IFNγ production is eventually compromised59 and, ultimately, exhausted T cells are deleted if the high antigenic load persists59. Exhaustion can also occur in CD4+ T cells in both mice61 and humans4. The current consensus is that functional exhaustion is a way of limiting the magnitude of effector T cell responses. Although this may safeguard against autoimmune responses, it may also compromise effective immunity against persistent infectious agents and tumours62. Which T cells have characteristics of exhaustion? Exhausted CD8+ T cells in both mice and humans display phenotypic markers that are typically associated with effector memory T cell populations37,59. During LCMV infection in mice, exhausted CD8+ T cells express low levels of L‑selectin (also known as CD62L), CC‑chemokine receptor 7 (CCR7), IL‑7 receptor and IL‑15 receptor 63,64. The loss of CD28 expression is a feature of exhausted human CD8+ T cells, although it is not clear whether these cells are at an intermediate or late stage of differentiation (TABLE 1). Exhausted CD4+ T cells also have an effector memory T cell phenotype61.
How is exhaustion regulated? Exhausted T cells are subject to complex layers of negative regulation. This involves signalling through multiple inhibitory receptors that inhibit functional and proliferative responses2 (FIG. 2). The CD28 family member programmed cell death 1 (PD1) has been shown to be the most highly expressed inhibitory receptor on CD8+ T cells during chronic infection, and this receptor has a major role in regulating T cell exhaustion during infection1,2. PD1 expression is upregulated during chronic LCMV infection in mice1 and by virus-specific CD4+ and CD8+ T cells during HIV1 infection4,65,66. Increased expression of PD1 by T cells also occurs during HBV and HCV infections55,56,67. An important observation is that blockade of the PD1 signalling pathway can restore antigen-specific T cell responses (both proliferation and cytokine secretion) in LCMV-infected mice1 and in humans with chronic HIV4,65,66, HBV68 and HCV55,56 infections. Several other inhibitory receptors have also been shown to induce T cell unresponsiveness during chronic infections. These receptors include cytotoxic T lymphocyte antigen 4 (CTLA4)2,69,70, T cell
2JCUG
2JCUG
2JCUG
+PFWEVKQP r*KIJXKTCNQT VWOQWT CPVKIGPNQCF
&KȭGTGPVKCVKQP r7RTGIWNCVKQPQH KPJKDKVQT[TGEGRVQT GZRTGUUKQP
+PJKDKVQT[TGEGRVQTUKIPCNNKPI r)TQYVJCTTGUV r&GETGCUGFE[VQVQZKEECRCEKV[ r&GETGCUGFE[VQMKPGRTQFWEVKQP
#PVKIGP
%6.#
6+/ .#)
2&
2+5GT 2
2&
/*% ENCUU+
#-6
O614%
6JT 2
2&- (1:1
%& 6EGNN
R )s5RJCUGVTCPUKVKQP
Figure 2 | A hypothetical scheme for the induction of proliferative exhaustion by inhibitory 0CVWTG4GXKGYU^+OOWPQNQI[ receptor signalling. This scheme is based on available information regarding the programmed cell death 1 (PD1) signalling pathway81, but it is not clear how other inhibitory receptors signal. T cell exhaustion occurs in the presence of a high antigenic load, such as in chronic viral infection or in cancer. The T cell responds to this high antigenic load by differentiating into effector T cells in an attempt to clear the antigen. During this response, the T cell starts to express inhibitory receptors, such as PD1, cytotoxic T lymphocyte antigen 4 (CTLA4), T cell immunoglobulin domain and mucin domain protein 3 (TIM3) and lymphocyte activation gene 3 (LAG3). Signalling through these receptors causes growth arrest and the loss of cytokine production and cytotoxic ability. Inhibitory receptor signalling mediates growth arrest through inhibition of the AKT signalling pathway (as a result of inhibition of phospho inositide 3‑kinase (PI3K), the activity of which is required for AKT phosphorylation). This in turn lifts the block on forkhead box O (FOXO) transcription factors and activates the transcription of p27 (also known as KIP1), thereby preventing the transition from G1 to S phase in the cell cycle93–95. The mechanisms responsible for the other functional changes that occur during exhaustion are not known. mTORC, mTOR complex; PDK1, 3‑phosphoinositide-dependent protein kinase 1.
292 | APRIL 2011 | VOLUME 11
immunoglobulin domain and mucin domain protein 3 (TIM3)71–73 and lymphocyte activation gene 3 (LAG3)2,69 (FIG. 3). Exhausted CD8+ T cells can be divided into groups that express different numbers and combinations of inhibitory receptors. Using the LCMV mouse model of infection, the pattern of inhibitory receptor co-expression was shown to affect the functional quality of virus-specific CD8+ T cells during infection2,69. A similar integration of inhibitory receptor signalling in T cells has also been found in T cells from HIV-infected individuals70. Furthermore, blockade of both TIM3 and PD1 ligand 1 (PDL1) has been shown to be more effective in restoring antitumour immunity than targeting either pathway alone72,73. In addition, certain cytokines such as IL‑10 and transforming growth factor‑β (TGFβ), as well as regulatory T cells, may also contribute to the lack of T cell functionality during situations of high antigenic burden74. Although the expression of inhibitory receptors changes during T cell differentiation (TABLE 1), the expression of these receptors per se does not indicate that a T cell is exhausted. For example, human T cells that express PD1, CTLA4 and other inhibitory receptors can exhibit functional activity after activation75. Human T cells at intermediate stages of differentiation express the highest levels of PD1 and CTLA4, and these cells have relatively long telomeres and are therefore not, by definition, pre-senescent or end-stage differentiated T cells (TABLE 1). Gene profiling of the different T cell subsets has shown increased inhibitory receptor expression at the late stages of differentiation76–78, but microarray data generated from senescent and exhausted T cells have not been directly compared. Nevertheless, T cell populations at the late stages of differentiation express relatively high levels of these inhibitory receptors, indicating that there is the potential to control the function of T cells that are close to senescence by inhibitory receptor signalling. Signalling pathways that lead to exhaustion. Most of the available information regarding the mechanisms responsible for T cell exhaustion relates to signalling pathways that are triggered by individual inhibitory receptors. The best characterized of these receptors are CTLA4 and PD1 (REFS 75, 79). CTLA4 signalling has been discussed extensively elsewhere80 and so is not addressed here. PD1 signalling during T cell stimulation has been shown to inhibit www.nature.com/reviews/immunol
© 2011 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES phosphoinositide 3‑kinase (PI3K) activation through a mechanism that involves the recruitment of either SH2 domain-containing protein tyrosine phosphatase 1 (SHP1; also known as PTPN6) or SHP2 (also known as PTPN11) to the immunoreceptor tyrosine motif of PD1 (REF. 81). This in turn prevents the phosphorylation of AKT, which triggers a signalling cascade that regulates cell survival and cell cycling 81 (FIG. 2). Blockade of PD1 signalling has been shown to increase T cell proliferative responses in mice and humans1,62 and to increase the phosphory lation of AKT on Ser473 in highly differentiated CD8+ T cells in humans (S.M.H. and A.N.A., unpublished observations). This is different to the mechanism by which proliferation is inhibited in senescent T cells, in which DDR signalling via ATM and ATR (ATM and RAD3 related) leads to growth arrest (FIG. 1). Although many researchers consider killer cell lectin-like receptor subfamily G, member 1 (KLRG1) to be a senescenceassociated molecule82–84, blockade of KLRG1 signalling also reverses the proliferative dysfunction of highly differentiated human CD8+ T cells by inducing Ser473 phosphor ylation of AKT82. This suggests that signalling via KLRG1 actually regulates an exhaustion-related pathway, but in pre-senescent T cells. However, the possibility that KLRG1 also regulates additional pathways that are involved in senescence cannot be ruled out. It is not clear whether blocking the PI3K–AKT axis through PD1 and KLRG1 signalling also inhibits cytokine production and cytotoxicity in exhausted T cells. BCL‑2‑interacting mediator of cell death (BIM; also known as BCL2L11), a proapoptotic member of the B cell lymphoma 2 (BCL‑2) family, is upregulated by exhausted T cells during chronic viral infections in mice85 and humans86. B lymphocyteinduced maturation protein 1 (BLIMP1; also known as PRDM1), a transcriptional repressor that regulates terminal differentiation of B cells and CD8+ T cells, is also induced in exhausted T cells during chronic LCMV infection69,87. How BLIMP1 induces proliferative arrest is unclear, but alterations in the expression of the proproliferative factor inhibitor of DNA binding 3 (ID3) may be involved88. BLIMP1 also suppresses IL‑2 transcription88. Studies in the mouse LCMV model have suggested that alterations of cytotoxic effector gene expression in CD8+ T cells are regulated by the transcription factors T‑bet, eomesodermin and runt-related transcription factor 3
(RUNX3), whereas the loss of cytokine production may be regulated by nuclear factor of activated T cells (NFAT)88. However, it is not clear at present how these transcription factors are regulated in exhausted T cell populations. Although exhaustion and senescence signalling can both induce proliferative arrest, the proximal signalling events that lead to this process may be distinct (FIG. 3). Senescence or exhaustion blockade T cell function is reduced in patients with chronic viral infections or with malignant tumours4,55,56,65–67,72. This raises the question of whether it is possible to enhance immunity by blocking T cell senescence or exhaustion, or both. As senescence mechanisms are related to DNA damage, reversing the growth arrest by selective blockade carries a risk of malignancy. Nevertheless, some researchers consider that the benefits of re-inducing telomerase in pre-senescent T cells may outweigh the risks involved30,89. As T cells that are close to senescence are functional20,36, blocking senescence pathways may not be of added benefit to functionality. By contrast, selective blockade of functionally exhausted T cells may enhance their proliferative and functional activity without breaching signalling checkpoints for malignant transformation.
For ageing. Highly differentiated T cells that have short telomeres (pre-senescent cells) accumulate in older humans in vivo 9,20,21,30,32,43,90. However, these cells remain functional despite their reduced proliferative activity and do not exhibit characteristics of functional exhaustion36. Nevertheless, these T cells express multiple inhibitory receptors69, so it is possible that their function could be increased further by inhibitory receptor blockade. It is not clear whether blocking inhibitory receptors increases telomerase activity. Although one report has shown that telo merase was induced by PD1 blockade91, this was not assessed in highly differentiated T cell populations or in older humans. The immune response to vaccination against infectious diseases is decreased in older humans. Inhibitory receptor blockade may improve this response, and might also prove to be a possible therapy for tumours and persistent infections. The feasibility of manipulating multiple inhibitory pathways has been demonstrated recently through the combined targeting of the TIM3 and the
NATURE REVIEWS | IMMUNOLOGY
'ZJCWUVKQP +Phibitor[TGEGRVQT -.4) UKIPCNNKPI
2&
2
#-6
2+-
(1:1 R %[ENKP' %&- %[ENKP& %&- %&-
R R
%[ENKP# %&-
R %[ENKP$ %&-
R #6/ CPF#64
%[ENKP# %&-
5GPGUEGPEG &0#FCOCIGTGURQPUG
Figure 3 | A putative scheme suggesting that senescence and exhaustion pathways 0CVWTG4GXKGYU^+OOWPQNQI[ block T cell proliferation by distinct mechanisms. Different mechanisms are responsible for the cell cycle arrest observed in senescence and exhaustion. Senescence stimuli result in the activation of ataxia telangiectasia mutated (ATM) and ATM and RAD3 related (ATR), which stimulate effector components of the DNA damage response pathway, such as p53, the mitogen-activated protein kinase p38 and p16 (also known as INK4A). This causes G1 growth arrest by blocking the function of cyclins and cyclin-dependent kinases (CDKs). Much less is known about the molecular events that control exhaustion. However, data on signalling through programmed cell death 1 (PD1) suggest that this inhibitory receptor acts by preventing AKT phosphorylation, through the inhibition of phosphoinositide 3‑kinase (PI3K). This in turn lifts the block on forkhead box O (FOXO) transcription factors and activates the transcription of p27 (also known as KIP1), causing G1–S phase transition. The senescenceassociated inhibitory receptor killer cell lectin-like receptor subfamily G, member 1 (KLRG1) also mediates its inhibitory signals by preventing AKT phosphorylation (on Ser473), and this removes the block on p27 transcription, enabling the G1–S phase transition.
PD1 pathways. Dual blockade was found to be more effective than blockade of either pathway alone in controlling tumour growth in mice73 and in reversing the exhaustion of melanoma-specific T cells isolated from patients with advanced melanoma72. VOLUME 11 | APRIL 2011 | 293
© 2011 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES For persistent infections. Highly differentiated T cells with short telomeres accumulate in patients with chronic viral infections, such as Epstein–Barr virus, HIV, HBV and HCV infections47,61,86. In addition, as discussed above, the T cells from these patients also exhibit characteristics of exhaustion. This suggests that blocking either senescence or exhaustion signalling may enhance the efficacy of T cells in these individuals. In one study, a small molecule inhibitor that activates the extracellular signal-regulated kinase (ERK) mitogen-activated protein kinase pathway was shown to enhance the proliferation, telomerase activity and replicative lifespan of T cells from patients with HIV89. Interestingly, the enhanced telomerase activity was also associated with increased cytokine production89. However, it is not known whether this inhibitor acts directly on DDR-related senescence signalling pathways or whether it is effective on T cells at all stages of differentiation. It may also be possible to prevent replicative senescence by using genetic manipulation to reintroduce telomerase into T cells before they reach a critically senescent state21,92. Such telomerase-expressing nonsenescent CD8+ T cells do not have karyotypic abnormalities, retain functional activity and are dependent on antigen stimulation for proliferation in vitro92. However, it is not known whether these T cells eventually become susceptible to telomere-independent senescence and whether the reintroduction of telomerase increases the risk of malignancy by conferring unlimited proliferative potential to cells that may harbour damaged DNA. Blocking of PD1 has been shown to enhance the function of HIV-specific CD4+ and CD8+ T cells in patients4. Thus, inhibitory receptor blockade could increase the functional activity of virus-specific T cells, and this may also be useful for the treatment of other persistent viral infections such as HBV and HCV. The blockade of inhibitory receptors could also increase the ability of the cells to proliferate after activation, and this would be achieved by modulating pathways that are not related to a DDR response (FIG. 3). However, it is currently not clear which combination of receptors should be targeted or when, relative to the stage of disease, this manipulation should this be carried out. It also remains to be determined whether this intervention could lead to nonspecific inflammation. Nevertheless, it is possible that, with experience, the benefits of manipulating both senescence and exhaustion for boosting immunity may be greater than the risks in certain situations.
Concluding remarks A detailed assessment of data describing signalling pathways that regulate senescence and exhaustion has led to the surprising conclusion that they appear to be distinct processes. The fact that different signalling pathways are engaged indicates that selective blockade of either pathway is feasible. Our understanding of both of these signalling pathways in human T cells is incomplete, and further investigation may identify more subtle ways of manipulating the function of these cells to boost immunity. For example, very little is known about how telomereindependent senescence manifests itself during T cell differentiation and whether this process changes during ageing and persistent infection. It is also not clear how the integration of signalling pathways following the engagement of multiple inhibitory receptors leads to the inhibition of some, but not all, T cell functions. This highlights the need to understand how the expression of inhibitory receptor ligands is regulated in different tissues. The fact that early senescence and exhaustion in T cells are both reversible is cause for optimism, as this offers the opportunity for intervention, although safety issues and side effects relating to the possible development of malignancy and immunopathology have to be taken into consideration. However, as human life expectancy continues to increase, with an associated decrease in the quality of life through decreased immunity, some drastic measures may be necessary. Arne N. Akbar and Sian M. Henson are at the Division of Infection and Immunity, University College London, UK. Correspondence to A.N.A. e-mail:
[email protected]
9.
10. 11. 12. 13. 14. 15. 16. 17. 18. 19.
20.
21.
22.
23. 24. 25.
26. 27.
doi:10.1038/nri2959 28. 1. 2.
3.
4. 5. 6. 7. 8.
Barber, D. L. et al. Restoring function in exhausted CD8 T cells during chronic viral infection. Nature 439, 682–687 (2006). Blackburn, S. D. et al. Coregulation of CD8+ T cell exhaustion by multiple inhibitory receptors during chronic viral infection. Nature Immunol. 10, 29–37 (2009). Freeman, G. J., Wherry, E. J., Ahmed, R. & Sharpe, A. H. Reinvigorating exhausted HIV-specific T cells via PD‑1–PD‑1 ligand blockade. J. Exp. Med. 203, 2223–2227 (2006). Day, C. L. et al. PD‑1 expression on HIV-specific T cells is associated with T‑cell exhaustion and disease progression. Nature 443, 350–354 (2006). Akbar, A. N. & Vukmanovic-Stejic, M. Telomerase in T lymphocytes: use it and lose it? J. Immunol. 178, 6689–6694 (2007). Hodes, R. J., Hathcock, K. S. & Weng, N. P. Telomeres in T and B cells. Nature Rev. Immunol. 2, 699–706 (2002). Effros, R. B., Dagarag, M., Spaulding, C. & Man, J. The role of CD8+ T‑cell replicative senescence in human aging. Immunol. Rev. 205, 147–157 (2005). Voehringer, D. et al. Viral infections induce abundant numbers of senescent CD8 T cells. J. Immunol. 167, 4838–4843 (2001).
294 | APRIL 2011 | VOLUME 11
29.
30. 31. 32. 33. 34.
35. 36.
Fletcher, J. M. et al. Cytomegalovirus-specific CD4+ T cells in healthy carriers are continuously driven to replicative exhaustion. J. Immunol. 175, 8218–8225 (2005). Appay, V., Almeida, J. R., Sauce, D., Autran, B. & Papagno, L. Accelerated immune senescence and HIV‑1 infection. Exp. Gerontol. 42, 432–437 (2007). d’Adda di Fagagna, F. et al. A DNA damage checkpoint response in telomere-initiated senescence. Nature 426, 194–198 (2003). Passos, J. F. et al. Feedback between p21 and reactive oxygen production is necessary for cell senescence. Mol. Syst. Biol. 6, 347 (2010). Akbar, A. N., Beverley, P. C. & Salmon, M. Will telomere erosion lead to a loss of T‑cell memory? Nature Rev. Immunol. 4, 737–743 (2004). Hayflick, L. The limited in vitro lifetime of human diploid cell strains. Exp. Cell Res. 37, 614–636 (1965). Blackburn, E. H. Telomeres and telomerase: their mechanisms of action and the effects of altering their functions. FEBS Lett. 579, 859–862 (2005). Campisi, J. & d’Adda di Fagagna, F. Cellular senescence: when bad things happen to good cells. Nature Rev. Mol. Cell Biol. 8, 729–740 (2007). d’Adda di Fagagna, F. Living on a break: cellular senescence as a DNA-damage response. Nature Rev. Cancer 8, 512–522 (2008). Passos, J. F. & Von Zglinicki, T. Oxygen free radicals in cell senescence: are they signal transducers? Free Radic. Res. 40, 1277–1283 (2006). Rufer, N. et al. Ex vivo characterization of human CD8+ T subsets with distinct replicative history and partial effector functions. Blood 102, 1779–1787 (2003). Plunkett, F. J. et al. The loss of telomerase activity in highly differentiated CD8+CD28–CD27– T cells is associated with decreased Akt (Ser473) phosphorylation. J. Immunol. 178, 7710–7719 (2007). Plunkett, F. J. et al. The impact of telomere erosion on memory CD8+ T cells in patients with X‑linked lymphoproliferative syndrome. Mech. Ageing Dev. 126, 855–865 (2005). Migliaccio, M., Raj, K., Menzel, O. & Rufer, N. Mechanisms that limit the in vitro proliferative potential of human CD8+ T lymphocytes. J. Immunol. 174, 3335–3343 (2005). Akbar, A. N., Soares, M. V., Plunkett, F. J. & Salmon, M. Differential regulation of CD8+ T cell senescence in mice and men. Mech. Ageing Dev. 121, 69–76 (2000). Itahana, K., Campisi, J. & Dimri, G. P. Mechanisms of cellular senescence in human and mouse cells. Biogerontology 5, 1–10 (2004). Herbig, U., Jobling, W. A., Chen, B. P., Chen, D. J. & Sedivy, J. M. Telomere shortening triggers senescence of human cells through a pathway involving ATM, p53, and p21CIP1, but not p16INK4a. Mol. Cell 14, 501–513 (2004). Iwasa, H., Han, J. & Ishikawa, F. Mitogen-activated protein kinase p38 defines the common senescencesignalling pathway. Genes Cells 8, 131–144 (2003). Shiloh, Y. ATM and related protein kinases: safeguarding genome integrity. Nature Rev. Cancer 3, 155–168 (2003). Campisi, J. Senescent cells, tumor suppression, and organismal aging: good citizens, bad neighbors. Cell 120, 513–522 (2005). Reed, J. R. et al. Telomere erosion in memory T cells induced by telomerase inhibition at the site of antigenic challenge in vivo. J. Exp. Med. 199, 1433–1443 (2004). Effros, R. B. Replicative senescence of CD8 T cells: effect on human ageing. Exp. Gerontol. 39, 517–524 (2004). Nikolich-Žugich, J. Ageing and life-long maintenance of T‑cell subsets in the face of latent persistent infections. Nature Rev. Immunol. 8, 512–522 (2008). Akbar, A. N. & Fletcher, J. M. Memory T cell homeostasis and senescence during aging. Curr. Opin. Immunol. 17, 480–485 (2005). Goronzy, J. J., Lee, W. W. & Weyand, C. M. Aging and T‑cell diversity. Exp. Gerontol. 42, 400–406 (2007). Spaulding, C., Guo, W. & Effros, R. B. Resistance to apoptosis in human CD8+ T cells that reach replicative senescence after multiple rounds of antigen-specific proliferation. Exp. Gerontol. 34, 633–644 (1999). Pawelec, G., Wagner, W., Adibzadeh, M. & Engel, A. T cell immunosenescence in vitro and in vivo. Exp. Gerontol. 34, 419–429 (1999). Libri, V. et al. Cytomegalovirus infection induces the accumulation of short-lived, multifunctional
www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved
PERSPECTIVES
37. 38.
39. 40. 41. 42. 43. 44.
45.
46. 47.
48. 49.
50. 51.
52. 53. 54. 55. 56.
57.
58.
CD4+CD45RA+CD27– T cells: the potential involvement of interleukin‑7 in this process. Immunology 132, 326–339 (2011). Appay, V. et al. Memory CD8+ T cells vary in differentiation phenotype in different persistent virus infections. Nature Med. 8, 379–385 (2002). Sallusto, F., Geginat, J. & Lanzavecchia, A. Central memory and effector memory T cell subsets: function, generation, and maintenance. Annu. Rev. Immunol. 22, 745–763 (2004). Hamann, D. et al. Phenotypic and functional separation of memory and effector human CD8+ T cells. J. Exp. Med. 186, 1407–1418 (1997). Roth, A. et al. Telomerase levels control the lifespan of human T lymphocytes. Blood 102, 849–857 (2003). Faragher, R. G. & Kipling, D. How might replicative senescence contribute to human ageing? Bioessays 20, 985–991 (1998). Akbar, A. N. & Salmon, M. Cellular environments and apoptosis: tissue microenvironments control activated T‑cell death. Immunol. Today 18, 72–76 (1997). Pawelec, G. et al. Human immunosenescence: is it infectious? Immunol. Rev. 205, 257–268 (2005). Khan, N. et al. Cytomegalovirus seropositivity drives the CD8 T cell repertoire toward greater clonality in healthy elderly individuals. J. Immunol. 169, 1984–1992 (2002). Wills, M. R. et al. Identification of naive or antigenexperienced human CD8+ T cells by expression of costimulation and chemokine receptors: analysis of the human cytomegalovirus-specific CD8+ T cell response. J. Immunol. 168, 5455–5464 (2002). Das, A. et al. Functional skewing of the global CD8 T cell population in chronic hepatitis B virus infection. J. Exp. Med. 205, 2111–2124 (2008). Hoare, M. et al. CD4+ T‑lymphocyte telomere length is related to fibrosis stage, clinical outcome and treatment response in chronic hepatitis C virus infection. J. Hepatol. 53, 252–260 (2010). Koch, S. et al. Multiparameter flow cytometric analysis of CD4 and CD8 T cell subsets in young and old people. Immun. Ageing 5, 6 (2008). Mountz, J. D., Wu, J., Zhou, T. & Hsu, H. C. Cell death and longevity: implications of Fas-mediated apoptosis in T‑cell senescence. Immunol. Rev. 160, 19–30 (1997). Beausejour, C. M. et al. Reversal of human cellular senescence: roles of the p53 and p16 pathways. EMBO J. 22, 4212–4222 (2003). Davis, T. et al. Synthesis and in vivo activity of MK2 and MK2 substrate-selective p38αMAPK inhibitors in Werner syndrome cells. Bioorg. Med. Chem. Lett. 17, 6832–6835 (2007). Ono, K. & Han, J. The p38 signal transduction pathway: activation and function. Cell Signal. 12, 1–13 (2000). Zajac, A. J. et al. Viral immune evasion due to persistence of activated T cells without effector function. J. Exp. Med. 188, 2205–2213 (1998). Velu, V. et al. Enhancing SIV-specific immunity in vivo by PD‑1 blockade. Nature 458, 206–210 (2009). Urbani, S. et al. PD‑1 expression in acute hepatitis C virus (HCV) infection is associated with HCV-specific CD8 exhaustion. J. Virol. 80, 11398–11403 (2006). Nakamoto, N. et al. Synergistic reversal of intrahepatic HCV-specific CD8 T cell exhaustion by combined PD‑1/CTLA‑4 blockade. PLoS Pathog. 5, e1000313 (2009). Islam, S. A. et al. Persistence of human immunodeficiency virus type 1‑specific cytotoxic T‑lymphocyte clones in a subject with rapid disease progression. J. Virol. 75, 4907–4911 (2001). Lee, P. P. et al. Characterization of circulating T cells specific for tumor-associated antigens in melanoma patients. Nature Med. 5, 677–685 (1999).
59. Wherry, E. J., Blattman, J. N., Murali-Krishna, K., van der Most, R. & Ahmed, R. Viral persistence alters CD8 T‑cell immunodominance and tissue distribution and results in distinct stages of functional impairment. J. Virol. 77, 4911–4927 (2003). 60. Richter, K., Agnellini, P. & Oxenius, A. On the role of the inhibitory receptor LAG‑3 in acute and chronic LCMV infection. Int. Immunol. 22, 13–23 (2010). 61. Klenerman, P. & Hill, A. T cells and viral persistence: lessons from diverse infections. Nature Immunol. 6, 873–879 (2005). 62. Sharpe, A. H., Wherry, E. J., Ahmed, R. & Freeman, G. J. The function of programmed cell death 1 and its ligands in regulating autoimmunity and infection. Nature Immunol. 8, 239–245 (2007). 63. Wherry, E. J., Barber, D. L., Kaech, S. M., Blattman, J. N. & Ahmed, R. Antigen-independent memory CD8 T cells do not develop during chronic viral infection. Proc. Natl Acad. Sci. USA 101, 16004–16009 (2004). 64. Shin, H., Blackburn, S. D., Blattman, J. N. & Wherry, E. J. Viral antigen and extensive division maintain virus-specific CD8 T cells during chronic infection. J. Exp. Med. 204, 941–949 (2007). 65. Petrovas, C. et al. PD‑1 is a regulator of virus-specific CD8+ T cell survival in HIV infection. J. Exp. Med. 203, 2281–2292 (2006). 66. Trautmann, L. et al. Upregulation of PD‑1 expression on HIV-specific CD8+ T cells leads to reversible immune dysfunction. Nature Med. 12, 1198–1202 (2006). 67. Boni, C. et al. Characterization of hepatitis B virus (HBV)-specific T‑cell dysfunction in chronic HBV infection. J. Virol. 81, 4215–4225 (2007). 68. Boettler, T. et al. Expression of the interleukin‑7 receptor α chain (CD127) on virus-specific CD8+ T cells identifies functionally and phenotypically defined memory T cells during acute resolving hepatitis B virus infection. J. Virol. 80, 3532–3540 (2006). 69. Wherry, E. J. et al. Molecular signature of CD8+ T cell exhaustion during chronic viral infection. Immunity 27, 670–684 (2007). 70. Kaufmann, D. E. et al. Upregulation of CTLA‑4 by HIV-specific CD4+ T cells correlates with disease progression and defines a reversible immune dysfunction. Nature Immunol. 8, 1246–1254 (2007). 71. Jones, R. B. et al. Tim‑3 expression defines a novel population of dysfunctional T cells with highly elevated frequencies in progressive HIV‑1 infection. J. Exp. Med. 205, 2763–2779 (2008). 72. Fourcade, J. et al. Upregulation of Tim‑3 and PD‑1 expression is associated with tumor antigen-specific CD8+ T cell dysfunction in melanoma patients. J. Exp. Med. 207, 2175–2186 (2010). 73. Sakuishi, K. et al. Targeting Tim‑3 and PD‑1 pathways to reverse T cell exhaustion and restore anti-tumor immunity. J. Exp. Med. 207, 2187–2194 (2010). 74. Yi, J. S., Cox, M. A. & Zajac, A. J. T‑cell exhaustion: characteristics, causes and conversion. Immunology 129, 474–481 (2010). 75. Greenwald, R. J., Freeman, G. J. & Sharpe, A. H. The B7 family revisited. Annu. Rev. Immunol. 23, 515–548 (2005). 76. Czesnikiewicz-Guzik, M. et al. T cell subset-specific susceptibility to aging. Clin. Immunol. 127, 107–118 (2008). 77. Fann, M. et al. Gene expression characteristics of CD28null memory phenotype CD8+ T cells and its implication in T‑cell aging. Immunol. Rev. 205, 190–206 (2005). 78. Cao, J. N., Gollapudi, S., Sharman, E. H., Jia, Z. & Gupta, S. Age-related alterations of gene expression patterns in human CD8+ T cells. Aging Cell 9, 19–31 (2010). 79. Sharpe, A. H. & Freeman, G. J. The B7–CD28 superfamily. Nature Rev. Immunol. 2, 116–126 (2002).
NATURE REVIEWS | IMMUNOLOGY
80. Rudd, C. E., Taylor, A. & Schneider, H. CD28 and CTLA‑4 coreceptor expression and signal transduction. Immunol. Rev. 229, 12–26 (2009). 81. Parry, R. V. et al. CTLA‑4 and PD‑1 receptors inhibit T‑cell activation by distinct mechanisms. Mol. Cell. Biol. 25, 9543–9553 (2005). 82. Henson, S. M. et al. KLRG1 signaling induces defective Akt (ser473) phosphorylation and proliferative dysfunction of highly differentiated CD8+ T cells. Blood 113, 6619–6628 (2009). 83. Ouyang, Q. et al. Age-associated accumulation of CMV-specific CD8+ T cells expressing the inhibitory killer cell lectin-like receptor G1 (KLRG1). Exp. Gerontol. 38, 911–920 (2003). 84. Voehringer, D., Koschella, M. & Pircher, H. Lack of proliferative capacity of human effector and memory T cells expressing killer cell lectinlike receptor G1 (KLRG1). Blood 100, 3698–3702 (2002). 85. Grayson, J. M., Weant, A. E., Holbrook, B. C. & Hildeman, D. Role of Bim in regulating CD8+ T‑cell responses during chronic viral infection. J. Virol. 80, 8627–8638 (2006). 86. Lopes, A. R. et al. Bim-mediated deletion of antigenspecific CD8 T cells in patients unable to control HBV infection. J. Clin. Invest. 118, 1835–1845 (2008). 87. Shin, H. et al. A role for the transcriptional repressor Blimp‑1 in CD8+ T cell exhaustion during chronic viral infection. Immunity 31, 309–320 (2009). 88. Angelosanto, J. M. & Wherry, E. J. Transcription factor regulation of CD8+ T‑cell memory and exhaustion. Immunol. Rev. 236, 167–175 (2010). 89. Fauce, S. R. et al. Telomerase-based pharmacologic enhancement of antiviral function of human CD8+ T lymphocytes. J. Immunol. 181, 7400–7406 (2008). 90. van de Berg, P. J. et al. Cytomegalovirus infection reduces telomere length of the circulating T cell pool. J. Immunol. 184, 3417–3423 (2010). 91. Lichterfeld, M. et al. Telomerase activity of HIV‑1‑specific CD8+ T cells: constitutive up-regulation in controllers and selective increase by blockade of PD ligand 1 in progressors. Blood 112, 3679–3687 (2008). 92. Dagarag, M., Evazyan, T., Rao, N. & Effros, R. B. Genetic manipulation of telomerase in HIV-specific CD8+ T cells: enhanced antiviral functions accompany the increased proliferative potential and telomere length stabilization. J. Immunol. 173, 6303–6311 (2004). 93. Henson, S. M. & Akbar, A. N. KLRG1—more than a marker for T cell senescence. Age 31, 285–291 (2009). 94. Medema, R. H., Kops, G. J., Bos, J. L. & Burgering, B. M. AFX-like Forkhead transcription factors mediate cell-cycle regulation by Ras and PKB through p27kip1. Nature 404, 782–787 (2000). 95. Greenwald, R. J., Latchman, Y. E. & Sharpe, A. H. Negative co-receptors on lymphocytes. Curr. Opin. Immunol. 14, 391–396 (2002).
Acknowledgements
Work leading to this Review was funded by the British Biotechnology and Biological Sciences Research Council and the Wellcome Trust ViP Scheme. We also wish to thank numerous colleagues for extensive discussions that helped in the production of this article.
Competing interests statement
The authors declare no competing financial interests.
FURTHER INFORMATION Arne N. Akbar’s and Sian M. Henson’s homepage: http://www.ucl.ac.uk/infection-immunity ALL LINKS ARE ACTIVE IN THE ONLINE PDF
VOLUME 11 | APRIL 2011 | 295 © 2011 Macmillan Publishers Limited. All rights reserved