Research highlights
T- c e l l a c t i vat i o n
…two studies shed light on the exact location of primary CD8+ T‑cell activation in lymph nodes and on memory CD8+ T‑cell reactivation in peripheral tissues.
Where the action’s at T-cell activation in response to local infection occurs within the draining lymph nodes, possibly in deep areas of the lymph node adjacent to high endothelial venules. The activated T cells then migrate to the peripheral tissue, where they fight the infection. Following clearance of the pathogen, many T cells are retained in the peripheral tissue as memory cells but these non-lymphoid sites are viewed simply as locations of T-cell effector function. Now, two studies shed light on the exact location of primary
CD8+ T-cell activation in lymph nodes and on memory CD8+ T-cell reactivation in peripheral tissues. Using multiphoton-based intravital microscopy and a mouse model system, Hickman et al. set out to determine the exact location of antiviral CD8+ T-cell priming in draining inguinal lymph nodes following infection with either vaccinia virus or vesicular stomatitis virus. They found that within 2–3 hours of virus inoculation, infected cells could be easily visualized in a nearly continuous sheet just below the lymph node subcapsular sinus (SCS). In uninfected mice, transferred antigen-specific CD8+ T cells were located in the T-cell regions of the lymph node as expected, but following infection the CD8+ T cells relocated to the peripheral region of the lymph node just under the SCS. By contrast, antigen non-specific CD8+ T cells remained in the T-cell regions following infection. Further analysis showed that CD8+ T-cell redistribution required interactions with dendritic cells (DCs) bearing cognate antigen and that are located at the periphery of the lymph node. The antigen-specific CD8+ T cells formed stable interactions with the DCs and full activation of the CD8+ T cells occurred within the first 12 hours following infection. So, the data show that CD8+ T-cell activation
nature reviews | immunology
occurs at the periphery of the lymph node following viral infection and not adjacent to high endothelial venules as previously thought. Wakim et al. looked at how memory CD8+ T cells are reactivated in peripheral tissue during secondary challenge. They did this by transplanting sensory dorsal root ganglia from mice previously infected with herpes simplex virus (HSV) to naive recipient mice. The dorsal root ganglia contain virus-specific T cells from the initial infection and also latent HSV, which is reactivated by the transplant procedure. The authors observed a local re-expansion of HSV-specific CD8+ T cells following transplantation, which required local active HSV infection. This expansion was dependent on the recruitment of recipient antigen-presenting cells, most probably DCs, to the transplanted dorsal root ganglia and on the expansion of transplant-resident CD4+ helper T cells by the recruited DCs. So, this study shows that antigen encounter in peripheral tissues can result in the direct reactivation of memory CD8+ T cells, following DC infiltration and CD4+ helper T-cell expansion, as happens in lymph nodes in response to primary HSV infection. Olive Leavy ORIGINAL RESEARCH PAPERS Hickman, H. D. et al. Direct priming of antiviral CD8+ T cells in the peripheral interfollicular region of lymph nodes. Nature Immunol. 13 January 2008 (doi:10.1038/ni1557) | Wakim, L. M. et al. Dendritic cell-induced memory T cell activation in nonlymphoid tissues. Science 319, 198–202 (2008)
volume 8 | february 2008 © 2008 Nature Publishing Group
Research highlights
tumour immunology
TGFb betrays immunity for tumorigenesis The role of transforming growth factor‑β (TGFβ) in interactions between the immune system and tumour cells is controversial, with various studies indicating that this cytokine can function as either a tumour suppressor or a tumour promoter depending on the context. Here, Yang et al. propose that the recruitment of immature myeloid cells to the tumour can determine the switch between an antitum‑ origenic and a pro-tumorigenic TGFβ‑containing microenvironment. Using a mouse model of breast carcinoma — in which mammary epithelial cells express the polyoma virus middle T antigen (PyVmT) under the control of the mouse mammary tumour virus promoter — Yang and colleagues confirmed earlier results that showed that deletion of the gene encoding the type II TGFβ receptor (Tgfbr2) to abrogate TGFβ signalling resulted in increased pulmonary metastasis. In this case, TGFβ is, therefore, acting as a tumour suppressor. The tumours of PyVmT-expressing Tgfbr2-knockout mice contained a significantly higher number of Gr1+CD11b+ immature myeloid cells (also known as myeloid suppressor cells) than tumours from PyVmTexpressing control mice, and these Gr1+CD11b+ cells were found mainly at the invasive front of the tumours. To investigate the role of these Gr1+CD11b+ cells in tumorigenesis, Yang et al. co-injected Gr1+CD11b+ cells from tumour-bearing mice with
mouse mammary tumour cells into the mammary fat pad of BALB/c mice, and a significant increase was observed in lung metastasis compared with the injection of mouse mammary tumour cells alone or with Gr1+CD11b+ cells from normal spleens. In both in vitro and in vivo invasion assays, the addition of Gr1+CD11b+ cells from tumourbearing mice increased the invasion of surrounding tissues by tumour cells, and this promotion of tumour metastasis was shown to be through the production of matrix metalloproteinases (MMPs). The increased recruitment of Gr1+CD11b+ cells to tumours with Tgfbr2 deletion was shown to result from an increased production of CXC-chemokine ligand 5 (CXCL5) by the tumour cells. Blocking the interaction between CXCL5 and CXC-chemokine receptor 2 (CXCR2) on Gr1+CD11b+ cells inhibited their recruitment both in vitro and in vivo. Also, tumour-infiltrating Gr1+CD11b+ cells expressed higher levels of CXCR4 than splenic Gr1+CD11b+ cells from tumour-bearing mice, and Gr1+CD11b+ cells isolated from tumour tissues migrated in response to the CXCR4 ligand CXCL12 in vitro. Surprisingly, the mammary carcinomas with Tgfbr2 deletion contained significantly higher levels of TGFβ1 than control PyVmT tumours, which the authors suggest is the result of the increased infiltration of Gr1+CD11b+ cells. Gr1+CD11b+ cells from the spleens of tumour-
nature reviews | immunology
bearing mice were shown to produce more TGFβ1 than Gr1+CD11b+ cells from the spleens of normal mice. The authors therefore propose a model in which autologous TGFβ sig‑ nalling by mammary epithelial cells is tumour suppressive. Defective TGFβ signalling in tumour cells allows the recruitment of Gr1+CD11b+ cells to the tumour through the increased production of CXCL5 by transformed epithelial cells and the increased expression of CXCR4 by Gr1+CD11b+ cells. The recruited Gr1+CD11b+ cells then produce TGFβ1 and MMPs, thereby promoting tumour invasion and metastasis.
Kirsty Minton
ORIGINAL RESEARCH PAPER Yang, L. et al. Abrogation of TGFβ signalling in mammary carcinomas recruits Gr-1+CD11b+ myeloid cells that promote metastasis. Cancer Cell 13, 23–35 (2008)
volume 8 | february 2008 © 2008 Nature Publishing Group
Research highlights
In the news HIV: the ultimate hijacker With only 9 genes of its own encoding just 15 proteins, HIV relies on hijacking host-cell proteins for its survival and propagation. Researchers have now revealed how extensive this is, by identifying more than 200 host proteins not previously known to be exploited by HIV (Science, 10 January 2008). “Many could provide the basis for new drugs”, said Stephen Elledge of Harvard Medical School, Boston, USA, who led the study (Reuters, 11 January 2008). Elledge and colleagues identified these proteins using a library of small interfering RNAs specific for more than 21,000 human genes, blocking their expression one at a time and seeing the effect that had on the virus’s ability to infect new cells. Although the 273 human proteins identified in the screen may not be the complete list, Abraham Brass, a co-author of the study, is confident that “the majority of the ones we found are highly likely to play a role in HIV propagation” (New York Times, 11 January 2008). Most drugs currently used to treat HIV infection target the virus directly, but “these therapeutics all suffer the same problem, which is that you can get resistance” as a result of rapid viral mutation (ScienceDaily, 11 January 2008). Elledge suggests that “If you could inhibit a host protein that is necessary for HIV progression, HIV would have to develop a new ability … that’s why host proteins could make tremendously good drug targets.” (Nature, 10 January 2008.) However, Martin Hirsch, also at Harvard Medical School, and others have expressed concern that drugs that interact with human proteins could have serious side effects. Elledge counters that, “Perturbing one [protein] may not have a profound effect on a cell, but it may on HIV.” (Science, 11 January 2008.) Lucy Bird
nature reviews | immunology
volume 8 | february 2008 © 2008 Nature Publishing Group
Research highlights Nature Reviews Immunology | AOP, published online 21 January 2008; doi:10.1038/nri2259
I n n at e I M M U N I T Y
NF-κB is not alone A new nuclear factor has been discovered that is required for innate immune responses and is highly conserved in mice and flies. The new molecule, Akirin, functions in parallel with the transcription factor NF-κB (nuclear factor-κB) downstream of the immune deficiency (IMD) pathway in Drosophila melanogaster, and has an essential role downstream of the Toll-like receptor (TLR), tumour-necrosis factor (TNF) and interleukin-1β (IL-1β) signalling pathways that lead to the production of IL-6 in mice. Following immune challenge, D. melanogaster generates antimicrobial peptides as a defence
response. The transcription of genes that encode these peptides is mediated by two pathways, the Toll pathway (which has parallels with the mammalian TLR signalling pathways) and the IMD pathway (which has similarities with the mammalian TNF signalling pathway). To identify new components of the IMD pathway, the authors carried out a high-throughput genome-wide RNAi screen of D. melanogaster and identified Akirin, which was found to have two highly conserved homologues in mammals (Akirin-1 and Akirin-2). Expression of Akirin proteins was shown to be ubiquitous in flies and humans, and strictly nuclear. Further analysis in vitro and in vivo revealed that Akirin is involved in the IMD pathway but not in the Toll pathway. More specifically Akirin was shown to function downstream of IMD and, in keeping with its nuclear localization, downstream of or at the same level as Relish (a member of the NF-κB family). In view of these results, could the Akirin homologues have a similar function in mice? Akirin-1knockout mice developed with no obvious phenotype, but Akirin-2 was shown to be essential for embryogenesis. In response to stimulation with TLR ligands, IL-1β or TNF, Akirin-1-deficient mouse embryonic fibroblasts (MEFs)
nature reviews | immunology
produced similar amounts of IL-6 to their wild-type counterparts. By contrast, Akirin-2-deficient MEFs produced significantly lower amounts of IL-6 in response to the same stimuli. Further analysis showed that Akirin-2 was crucial for the regulation of only a specific set of lipopolysaccharide (LPS)- and IL-1β-inducible genes. The degradation of IκBα (inhibitor of NF-κB α) and the DNA-binding activity of NF-κB after stimulation with IL-1β and LPS were not affected in Akirin-2-deficient cells, indicating that, similar to D. melanogaster Akirin, Akirin-2 acts in parallel with or downstream of NF-κB in the regulation of TLR- and IL-1βinducible gene expression. This study has identified a previously unknown nuclear factor that, together with or downstream of NF-κB, can regulate innate immune responses. Further studies will be needed to determine precisely how Akirin proteins control gene expression.
Marta Tufet
ORIGINAL RESEARCH PAPER Goto, A. et al. Akirins are highly conserved nuclear proteins required for NF-κB-dependent gene expression in Drosophila and mice. Nature Immunol. 9, 97–104 (2008) FURTHER READING Ferrandon, D., Imler, J.-L., Hetru, C. & Hoffmann, J. A. The Drosophila systemic immune response: sensing and signalling during bacterial and fungal infections. Nature Rev. Immunol. 7, 862–874 (2007)
february 2008 © 2008 Nature Publishing Group
Research highlights
t u m o u r i m m u n ot h e r a p y
Little pig, little pig, let me come in With all its huffing and puffing, the nursery rhyme wolf could not blow down the little pigs’ house made of bricks. The wall of endothelial cells surrounding tumour-associated blood vessels can be equally impenetrable, preventing T cells from exiting the blood and destroying the tumour. A recent study in Nature Medicine identifies a new mechanism by which the tumour endothelium blocks
T-cell homing and describes how this can be overcome to enhance the efficacy of tumour immunotherapy. To understand how the tumour endothelium might regulate T-cell homing, the authors compared gene expression by tumour endothelial cells (TECs) from human ovarian cancers with and without tumour-infiltrating lymphocytes (TILs). Numerous genes were found to be differentially expressed; of particular interest to the authors was endothelin B receptor (ETBR), which was expressed at higher levels by TECs from tumours lacking TILs than those with TILs. The ligand for ETBR, endothelin‑1, is known to be produced at high levels by ovarian cancer cells and has a crucial role in regulating vascular homeostasis and permeability. Indeed, in an in vitro assay, the presence of endothelin-1 inhibited T-cell adhesion to endothelial cells (a step required for lymphocytes to migrate out of the bloodstream). However, the effect of endothelin-1 on adhesion could be neutralized by addition of the specific ETBR inhibitor BQ-788. Further studies showed that the endothelin‑1– ETBR interaction suppressed T-cell adhesion by inhibiting the upregulation of expression and clustering of intercellular adhesion molecule 1 (ICAM1) on the endothelium and that nitric-oxide release was required for this effect. Accordingly, the presence of BQ-788 restored ICAM1 expression and reduced nitric-oxide release, thereby facilitating T-cell adhesion to endothelial cells in vitro.
nature reviews | immunology
Next, the authors tested whether ETBR blockade by BQ-788 might have a beneficial effect on T-cell homing in mouse models of cancer and improve the efficacy of otherwise ineffective tumour immunotherapy protocols. In the two mouse models studied, systemic antitumour T-cell responses were induced by tumour vaccines (a tumour-cell-based vaccine and a DNA vaccine), but this did not delay the growth of inoculated tumour cells or prolong survival of the mice. However, when the vaccinated mice were treated with BQ-788, tumour growth was delayed and mouse survival increased. This beneficial effect of ETBR blockade by BQ-788 was shown to be the result of increased homing of tumour-specific T cells to the tumour and not due to the induction of higher numbers or activation of these cells. So, overriding the tumour-induced barrier function of the endothelium with pharmacological inhibitors, such as BQ-788, could prove to be a useful strategy to ensure that tumourspecific T cells that are induced in immunotherapeutic protocols gain sufficient access to the tumour. Lucy Bird ORIGINAL RESEARCH PAPER Buckanovich, R. J. et al. Endothelin B receptor mediates the endothelial barrier to T cell homing to tumors and disables immune therapy. Nature Med. 14, 28–36 (2008) FURTHER READING Kershaw, M. H. Opening the gateway to tumors. Nature Med. 14, 13–14 (2008)
volume 8 | february 2008 © 2008 Nature Publishing Group
Research highlights
in brief B-CELL RESPONSES
microRNA-155 regulates the generation of immunoglobulin class-switched plasma cells. Vigorito, E. et al. Immunity 27, 847–859 (2007)
MicroRNAs (miRNAs) are small RNA molecules that regulate gene expression by targeting specific mRNAs. Previous studies have shown that high-level expression of miR-155 is associated with B-cell malignancies and that germline deletion of miR-155 results in defective T- and B-cell immunity. Using chimeric mice in which only the B cells lack miR-155, this study is the first to show the B-cell-autonomous effects of this miRNA. Deficiency of miR-155 resulted in a defective class-switched antibody response as a result of a decreased number of class-switched plasma cells. Affinity maturation and memory B-cell responses were also impaired. The authors show that mRNA encoding the transcription factor PU.1 is a direct target of miR-155 and that overexpression of PU.1 recapitulates the phenotype caused by deletion of miR-155. Therefore, analysis of PU.1 target genes and other potential targets of miR-155 might provide new insights into B-cell differentiation. I N F L A M M AT I O N
Muramyl dipeptide activation of nucleotide-binding oligomerization domain 2 protects mice from experimental colitis. Watanabe, T. et al. J. Clin. Invest. 10 January 2008 (doi:10.1172/JCI33145)
A subpopulation of patients with Crohn’s disease have mutations in the gene encoding nucleotide-binding oligomerization domain 2 (NOD2). One potential mechanism to explain this association is that NOD2 negatively regulates Toll-like receptor 2 (TLR2) signalling and that loss-of-function mutations in NOD2 lead to an increased inflammatory response to the intestinal microflora through TLR2. In support of this model, this study reports that pre-stimulation of wild-type mice, but not of those with mutated NOD2, with bacterial muramyl dipeptide can protect against chemically induced colitis. The protective effect was associated with the suppression of pro-inflammatory cytokine production through TLR3, TLR4, TLR5 and TLR9 pathways, in addition to the TLR2 pathway, as a result of the induction of interferon-regulatory factor 4 (IRF4), which is a negative regulator of TLR signalling. T- cell acti vation
A scaffold protein, AHNAK1, is required for calcium signaling during T cell activation. Matza, D. et al. Immunity 10 January 2008 (doi:10.1016/ j.immuni.2007.11.020)
Calcium signalling has a crucial role in T-cell activation and proliferation, and the calcium response involves the influx of calcium through protein channels in the plasma membrane. A role for the scaffold protein AHNAK1 in calcium signalling has been previously suggested and the expression of AHNAK1 is linked with cell proliferation. This study shows that AHNAK1 is expressed by peripheral T cells and that proliferation and interleukin-2 production following T-cell receptor (TCR) stimulation is impaired in AHNAK1-deficient CD4+ T cells. In addition, AHNAK1 is required for TCR-induced calcium signalling in CD4+ T cells through its regulation of the expression at the plasma membrane of the L-type calcium 1 channels. So, this study identifies AHNAK1 as a new factor in T-cell calcium signalling.
nature reviews | immunology
volume 8 | february 2008 © 2008 Nature Publishing Group
Research highlights
in brief I N F L A M M AT I O N
Defective tryptophan catabolism underlies inflammation in mouse chronic granulomatous disease. Romani, L. et al. Nature 451, 211–215 (2008)
Chronic granulomatous disease (CGD) is characterized by defective NADPH oxidase activity and an inability to generate reactive oxygen species, leading to recurrent microbial infections. Using a mouse model involving infection with Aspergillus fumigatus (which is commonly seen in patients with CGD), Romani et al. propose a new mechanism to explain the recurrent infection and exaggerated inflammation in CGD. They show that a defect in the superoxide-dependent conversion of tryptophan to kynurenine in mice with CGD leads to unrestrained reactivity of a subset of γδ T cells that produce interleukin-17, to defective regulatory T-cell activity and to acute inflammatory lung injury. Replacement therapy with natural kynurenine reversed the inflammatory phenotype, raising the possibility that this might help to control CGD in humans. m u c o s al imm u nolo g y
Developmental switch of intestinal antimicrobial peptide expression. Ménard, S. et al. J. Exp. Med. 7 January 2008 (doi:10.1084/jem.20071022)
Antimicrobial peptides produced by Paneth cells have an important role in protection against intestinal infections and in maintaining enteric homeostasis. But Paneth cells do not develop until after the neonatal period (~2 weeks in mice) and the innate defence effector molecules that protect the intestine in neonates are ill defined. Ménard et al. show that CRAMP (cathelicidin-related antimicrobial peptide), but not Panethcell-derived known antimicrobial peptides, is constitutively expressed by intestinal epithelial cells (IECs) specifically in neonates. Downregulation of CRAMP expression between 14 and 21 days after birth was accompanied by increased proliferation of IECs and the formation of intestinal crypts — developmental changes associated with the postnatal period. In addition, CRAMP was shown to have protective activity against enteric commensal and pathogenic microorganisms. So, this study describes a switch in enteric antimicrobial-peptide expression during the neonatal and postnatal development. CY T O K I N E S
Interleukin 17-producing T helper cells and interleukin 17 orchestrate autoreactive germinal center development in autoimmune BXD2 mice. Hsu, H.-C. et al. Nature Immunol. 23 December 2007 (doi:10.1038/ni1552)
Best known for its role in stimulating inflammatory responses, interleukin-17 (IL-17) is now linked to the induction of autoreactive humoral responses. The inbred mouse strain BXD2 develops spontaneous arthritis driven primarily by pathogenic autoantibodies. Using these mice, Hsu et al. show that autoantibody production depends on the production of IL-17 by T helper cells, which promotes the formation of germinal centres (GCs). In vitro, IL-17 was shown to inhibit B-cell chemotaxis by upregulating negative regulators of chemokine receptor signalling, and this is thought to stabilize B-cell–T-cell interactions that are required for GC development. Accordingly, IL-17-receptor-deficient BXD2 mice had fewer GCs and lower serum autoantibody levels than wild-type BXD2 mice. So, these data indicate another way in which IL-17 is involved in autoimmunity.
nature reviews | immunology
volume 8 | month 2008 © 2008 Nature Publishing Group
Research highlights Nature Reviews Immunology | AOP, published online 21January 2008 doi:10.1038/nri2258
p h ag o cy to s i s
Autophagy lends a hand structures did not form around the TLR-induced phagosomes associated with autophagy components. So what might be the functional significance of the recruitment of autophagy machinery to phagosomes? The authors found that during the uptake of beads alone, the fusion of phagosomes with lysosomes (as indicated by phagosome acidification) was limited. By contrast, beads coupled with TLR ligands induced more rapid and extensive acidification of the phagosomes, indicating that the TLRinduced association of autophagy components promotes phagosome– lysosome fusion. Accordingly, knockdown of ATG5 expression inhibited this effect. Moreover, the live yeast Saccharomyces cerevisiae
nature reviews | immunology
survived better when engulfed by ATG7-deficient macrophages than by wild-type macrophages, suggesting that recruitment of elements of the autophagy pathway has an impact on the survival of engulfed microorganisms. So, by recruiting help from components of the autophagy pathway, TLR signalling may enhance phagosome maturation and thereby pathogen defence. Lucy Bird ORIGINAL RESEARCH PAPER Sanjuan, M. A. et al. Toll-like receptor signalling in macrophages links the autophagy pathway to phagocytosis. Nature 450, 1253–1257 (2007) FURTHER READING Levine, B. & Deretic, V. Unveiling the roles of autophagy in innate and adaptive immunity. Nature Rev. Immunol. 7, 767–777 (2007)
BANANASTOCK
…Toll-like receptor (TLR) signalling usurps components that are traditionally associated with autophagy to increase the efficiency of phagocytosis.
A new study in Nature shows that Toll-like receptor (TLR) signalling usurps components that are traditionally associated with autophagy to increase the efficiency of phagocytosis, thereby providing a previously unappreciated link between these two ancient microbial defence mechanisms. Autophagy is a process by which cells dispose of unwanted cytosolic components or microorganisms. A key step in this process is the conjugation of the autophagy marker LC3 to the lipid phosphatidylethanolamine, forming LC3 aggregates on autophagic membranes. Using a fluorescently tagged form of LC3, Sanjuan et al. observed that LC3 was also recruited to the membranes of phagosomes that had engulfed latex beads associated with various TLR ligands. Importantly, LC3 recruitment to such phagosomes only occurred if the latex particle engaged TLRs while it was being internalized and therefore did not occur on internalization of latex beads alone or in TLR2deficient macrophages that had ingested the TLR2 ligand zymosan. Other components of the autophagy process — namely ATG5 (autophagy-related 5) and ATG7 — were shown to be required for TLR-induced LC3 recruitment to phagosomes. And similar to autophagy, LC3 aggregation on phagosomes was preceded by the recruitment of beclin-1 and phosphoinositide 3-kinase activity. However, unlike conventional autophagosomes, double-membrane
february 2008 © 2008 Nature Publishing Group
Research highlights
thymocyte development
E proteins: manning the checkpoints
… new roles for E proteins at distinct stages of thymocyte development.
The generation of mature T cells involves the transition of thymocytes through a series of developmental checkpoints. Numerous factors are involved in thymocyte development, but the precise mechanisms that govern the transition from one developmental stage to the next are not completely understood. Now, two papers published in Immunity identify new roles for E proteins at distinct stages of thymocyte development. The transition from the double negative 3 (DN3) to the double positive (DP) stage of development requires a signal through a complex of the T-cell-receptor (TCR) β-chain and the pre-TCR α-chain (together known as the pre-TCR). The generation of a functional TCR β-chain in DN3 cells requires TCRβ gene rearrangement, which is regulated by the accessibility of the target chromatin. To ensure that only TCRs of a single specificity are expressed by each cell (a phenomenon known as allelic exclusion), monoallelic activation of the locus must occur, followed by a negative-feedback signal that prevents continued rearrangement.
But what is the molecular mechanism behind this phenomenon? Agata et al. assessed the role of E proteins in this process, as E proteins have previously been shown to have multiple roles in DN3 cells. There are four E proteins (E12, E47, E2-2 and HEB), two of which, E12 and E47, are splice variants of the same gene (Tcfe2a; also known as E2A). The authors showed that partial or complete loss of E47 in DN3 thymocytes resulted in a substantial or almost-complete loss of chromatin accessibility at the TCRβ locus, as determined by Vβ germline transcription and of histone 3 acetylation. TCRβ gene rearrangement was also perturbed in an E47 dosage-sensitive manner. So, E47 promotes TCRβ gene rearrangement in a dosagesensitive manner through the direct regulation of chromatin accessibility. But how is this gene rearrangement then suppressed to achieve allelic exclusion? The authors showed that binding of E47 to the TCRβ locus and subsequent histone acetylation were decreased following pre-TCR signalling. So, following TCRβ gene rearrangement, the pre-TCR that is formed transmits a negative-feedback signal that directly dissociates the E protein from the chromatin, thereby reducing chromatin accessibility and preventing continued gene rearrangement. Finally, the authors showed that enforced expression of E47 disrupts allelic exclusion by overriding the negative-feedback signal. In a separate study, Jones and Zhuang assessed the role of E proteins
nature reviews | immunology
in the transition between the DP and single positive (SP) stages, which is controlled by TCR-mediated positive selection. They generated DP thymocytes that were deficient in both HEB and E2A and found that the development of CD8+ T cells from these DP cells occurred in the absence of a functional TCR, suggesting that HEB and E2A are required to maintain cells at the DP stage. Further analysis of the mutant DP cells showed that many DP-associated genes were downregulated in these cells, whereas SP-associated genes were already being upregulated. Previous studies have shown that HEB and E2A are downregulated following positive selection, so the authors suggest that both HEB and E2A are required to maintain the DP-cell phenotype and prevent early transition to the CD8+ T-cell lineage until positive selection can occur. These papers describe a crucial role for E proteins in maintaining two key checkpoints in thymocyte development — the transition from DN3 to DP cells and from DP to CD8 SP cells.
Olive Leavy
ORIGINAL RESEARCH PAPERS Agata, Y. et al. Regulation of T cell receptor β gene rearrangements and allelic exclusion by the helix-loop-helix protein, E47. Immunity 27, 871–884 (2007) | Jones, M. E. & Zhuang, E. Acquisition of a functional T cell receptor during T lymphocyte development is enforced by HEB and E2A transcription factors. Immunity 27, 860–870 (2007) FUTHER READING Rothenberg, E. V., Moore, J. E. & Yui, M. A. Launching the T-cell-lineage developmental programme. Nature Rev. Immunol. 8, 9–21 (2008)
volume 8 | february 2008 © 2008 Nature Publishing Group
Research highlights Nature Reviews Immunology | AOP, published online 21 January 2008; doi:10.1038/nri2255
I n n at e i m m u n i t y
TAMing inflammation Toll-like receptors (TLRs) are crucial for detecting infection and activating the innate and adaptive immune systems. However, sustained TLR stimulation can result in chronic inflammation and is associated with the development of certain autoimmune diseases. Now, Lemke and colleagues have identified a new signalling feedback loop through the TAM (TYRO3, AXL and MER) receptor family, which broadly inhibits both TLR signalling pathways and TLR-induced cytokine signalling pathways. They suggest that this could provide a mechanism through which the immune system self-regulates. The TAM receptor family comprises three receptor tyrosine kinases — TYRO3, AXL and MER — that can be activated by binding to two closely related ligands, growth-arrest-specific 6 (GAS6) and protein S. The authors showed that stimulation of dendritic cells (DCs) from mice that were deficient in all three TAM receptors with various TLR ligands resulted in the production of elevated levels of proinflammatory cytokines compared with wild-type DCs. Similarly, preincubation of wild-type DCs with either GAS6 or protein S inhibited TLR-induced cytokine production. The authors next showed that the activation of TAM receptors blocked
several TLR-induced signalling events, including nuclear factor-κB (NF-κB) activation. Furthermore, suppressor of cytokine signalling 1 (SOCS1) and SOCS3, which can inhibit TLR signalling as well as cytokine signalling, were both induced in response to TAM-receptor stimulation in a signal transducer and activator of transcription 1 (STAT1)-dependent manner, and STAT1 activation was necessary for the inhibition of TLR-induced cytokine production. STAT1 is commonly associated with signalling downstream of the type I interferon receptor (IFNAR). Further analysis showed that, following TAM-receptor activation, DCs deficient in IFNAR failed to activate STAT1, did not upregulate SOCS protein expression and produced normal levels of cytokines in response to TLR signalling. In addition, they found that signalling through IFNAR upregulated AXL expression and that it associated with the R1 chain of IFNAR. This observation suggests that the TAM receptors can hijack IFNAR-induced STAT1 activation to suppress TLR-induced cytokine production through SOCS proteins. These findings reveal a new selfregulating cycle in the innate immune system, whereby TLR‑induced
activation of the IFNAR–STAT1 pathway is used to induce the expression of the TAM receptors, which then co-opt the same pathway to initiate an intrinsic negative feedback loop that inhibits both TLR- and cytokine-driven responses. These findings have important implications for the resolution of inflammation, and inhibitors of the TAM receptor signalling pathway could provide attractive candidates for improved vaccine adjuvants. Jennifer Middleton Intern, Nature Publishing Group Original Research Paper Rothlin, C. V. et al. TAM receptors are pleiotropic inhibitors of the innate immune response. Nature 131, 1124–1136 (2007)
Erratum
Innate immunity: TAMing inflammation Jennifer Middleton Nature Reviews Immunology 8, 93 (2008)
The original research paper for this highlight was published in Cell 131, 1124–1136 (2007) and not in Nature as stated.
nature reviews | immunology
february 2008 © 2008 Nature Publishing Group
Research highlights
dendritic cells
New DCs found deep in the skin …three studies have identified a new population of langerin+ DCs that reside in the dermis of the skin and that arise from a different developmental pathway compared with LCs.
Langerhans cells (LCs) are a subset of dendritic cells (DCs) that express the C-type lectin langerin (also known as CD207) and reside as immature cells in the epidermis of the skin. Following interaction with pathogens, LCs mature and migrate via lymphatics to the T-cell-rich areas of draining lymph nodes (DLNs). Now, three studies have identified a new population of langerin+ DCs that reside in the dermis of the skin and that arise from a different developmental pathway compared with LCs. Previous studies have shown that epidermal LCs are radioresistant cells and therefore remain host-cell-derived following irradiation in the generation of bone-marrow chimaeras. By contrast, all other DC populations are radiosensitive and are donor-cellderived in bone-marrow chimaeras. Poulin et al. generated chimeric mice by irradiating recipient mice and reconstituting them with
bone-marrow cells from mice in which the expression of enhanced green fluorescent protein (EGFP) was under the control of the langerin promoter. Unexpectedly, when they analysed cells from cutaneous DLNs in the chimeric mice, in addition to the anticipated population of EGFP– (that is, of host origin) radio resistant cells derived from migratory epidermal LCs, they also identified a population of EGFP+ donor-cellderived DCs (that is, the host cells were radiosensitive). This new type of DC, which is langerin+, resides in the dermis, has a rapid turnover and, following ablation, repopulates the dermis and the DLNs at a much faster rate than the LCs. So, whereas LCs are langerin+ radioresistant cells that reside in the epidermis, this new type of DC subset, which the authors refer to as langerin+ dermal DCs, is radiosensitive. Ginhoux and colleagues used the same mouse model system, but in this case taking advantage of the fact that in addition to EGFP the mice also express the diphtheria toxin receptor under the control of the langerin promoter. In these mice diphtheria toxin administration efficiently eliminated all langerin+ cells. After diphtheria toxin administration, langerin+ DCs were shown to repopulate the cutaneous DLNs (these cells were previously thought to be trafficked LC-derived cells) and the dermis long before the reappearance of LCs in the epidermis, suggesting the existence of a population of non-LC-derived langerin+ DCs. Further work showed that these langerin+ DCs are derived from bonemarrow precursors that can seed
nature reviews | immunology
both the dermis and the DLNs. The authors also showed that langerin+ DC precursors depend on endothelialcell selectins and CC-chemokine receptor 2 (CCR2) to seed the dermis and on CCR7 to migrate to the cutaneous DLNs. Both langerin+ DCs and LCs were shown to patrol the skin and convey antigenic information to cells in the DLNs. Similar to the other studies, Bursch et al. identified a new population of langerin+ DCs in the dermis of the skin, which was also present in other tissues, particularly in the lungs. In addition, they tested the function of these cells in skin immune responses by performing contact hypersensitivity (CHS) assays in mice in which langerin+ cells had been ablated using treatment with diphtheria toxin. In these mice, dermal langerin+ DCs reappear in the dermis within 3 days of treatment with diphtheria toxin, whereas LCs were not detectable in the epidermis until a much later time point. This difference in repopulation kinetics allowed the authors to specifically assess the role of dermal langerin+ DCs in CHS responses. They found that dermal langerin+ DCs are necessary for optimal CHS responses and can promote these responses in the absence of epidermal LCs. Taken together, these three studies identify a new population of dermal langerin+ DCs and they help clarify some of the confusing results derived from previous studies when the existence of these cells was unappreciated. Elaine Bell ORIGINAL RESEARCH PAPERS Poulin, L. F. et al. The dermis contains langerin+ dendritic cells that develop and function independently of epidermal Langerhans cells. J. Exp. Med. 204, 3119–3131 (2007) | Ginhoux, F. et al. Bloodderived dermal langerin+ dendritic cells survey the skin in the steady state. J. Exp. Med. 204, 3133–3146 (2007) | Bursch, L. S. et al. Identification of a novel population of Langerin+ dendritic cells. J. Exp. Med. 204, 3147–3156 (2007)
february 2008 © 2008 Nature Publishing Group
REVIEWS Evolving views on the genealogy of B cells Robert S. Welner, Rosana Pelayo and Paul W. Kincade
Abstract | Many fundamental concepts about immune system development have changed substantially in the past few years, and rapid advances with animal models are presenting prospects for further discovery. However, continued progress requires a clearer understanding of the relationships between haematopoietic stem cells and the progenitors that replenish each type of lymphocyte pool. Blood-cell formation has traditionally been described in terms of discrete developmental branch points, and a single route is given for each major cell type. As we discuss in this Review, recent findings suggest that the process of B‑cell formation is much more dynamic. Microarray analysis A technique for measuring gene transcription. It involves hybridization of fluorescently labelled cDNA prepared from a cell or tissue of interest with glass slides or other surfaces dotted with thousands of oligonucleotides or cDNA, ideally representing all expressed genes in the species.
Multiplex PCR analysis A variation on conventional PCR assays, in which more than one pair of primers is used to simultaneously amplify target DNA sequences.
Terminal deoxynucleotidyltransferase (TdT). An enzyme expressed during lymphocyte development that inserts nucleotides into the variable regions of T‑cell receptor and immunoglobulin genes, to create junctional diversity. Immunobiology and Cancer Program, Oklahoma Medical Research Foundation 825 NE 13th Street, Oklahoma City, Oklahoma 73104, USA. Correspondence to P.W.K. e‑mail:
[email protected] doi:10.1038/nri2234 Published online 21 January 2008
The pace of discovery and lack of standardized nomen‑ clature are challenges for cartographers of immune sys‑ tem development. Nevertheless, exciting animal models and marker systems have now been developed that facilitate the identification of important milestones in haematopoiesis. Classic textbook diagrams have depicted the haematopoietic process as a series of abrupt, binary fate decisions; however, irreversible commitment to the B‑cell lineage is gradually acquired over many transi‑ tions through phenotypically distinguishable subsets of bone-marrow cells. Moreover, lymphoid progenitors can assume alternative fates and reverse course in response to certain environmental signals. Consequently, it has been difficult to find meaningful terms for early lym‑ phopoietic cells, and lymphopoiesis must be viewed in a fundamentally new way. Here, we discuss early changes that set the progeny of haematopoietic stem cells (HSCs) on a path to becoming B cells. We focus on the ways in which investigators can track these events and we highlight several remaining uncertainties.
Early lymphoid progenitors It has long been a goal of immunologists to identify and characterize the ‘earliest’ lymphoid progenitors in the adult bone marrow, but there have been conflict‑ ing claims about the properties of such cells. Microarray, quantitative RT‑PCR and multiplex PCR analyses have generated a wealth of information about patterns of early gene expression. For example, many genes com‑ monly associated with T and B cells, such as those encoding CD3δ, pre‑T-cell receptor (TCR) α‑chain (pTCRα), the pre‑B-cell receptor (BCR) surrogate light chains VpreB and Vλ5 and paired box protein 5
nature reviews | immunology
(PAX5), are not active in HSCs1,2. However, stem cells do contain detectable mRNA transcripts that corre‑ spond to genes that are expressed in multiple other haematopoietic lineages. This may reflect the general availability of developmental options or chromatin sta‑ tus at key genetic loci and does not necessarily presage the fate of individual cells3. Indeed, this phenomenon has been referred to as ‘priming’ or ‘promiscuity’ of lymphoid and myeloid progenitors1. Almost all early cell progenitors that express lymphoid-related genes are also positive for granulocyte/macrophage-related transcripts2. Given this background, lymphopoietic cells might be called ‘lymphoid specified’ when they express detectable amounts of proteins that are normally restricted to lym‑ phocytes. Park and Osmond first used the expression of terminal deoxynucleotidyltransferase (TdT), an enzyme that diversifies antibody combining sites, as a defining feature of what they referred to as early pro‑B cells4. A series of subsequent studies confirmed the use of this marker to define B‑, natural killer (NK)- and T‑lymphoid-specified progenitors, which our research group has termed early lymphoid progenitors (ELPs)5,6. However, only polyclonal antibodies are available to detect TdT, which is present at quite low levels and can only be visualized in fixed cells. Fortunately, genereporter mice and characteristic patterns of cell-surface antigen expression now make it possible to isolate highly enriched populations of lymphoid-specified cells. Although not irrevocably committed to the lym‑ phoid lineage, ELPs appear to have a high probability of becoming lymphocytes and can do so experimentally with remarkable efficiency. volume 8 | february 2008 | 95
© 2008 Nature Publishing Group
REVIEWS subset (termed the LSK subset), which represents approximately 0.1% of the total cell pool in the bone marrow of C57BL/6 mice. The loss of potential for self renewal, the defining characteristic of stem cells, and life-long replenishment of all types of blood cell coincides with the expression of FMS-related tyrosine kinase 3 (FLT3; also known as FLK2)7,8. Although not universally accepted, increased FLT3 levels appear to mark the initial, gradual segregation of haematopoieticcell fates9–11. That is, cells of the LSK subset that upregu‑ late FLT3 expression have greatly reduced potential for the generation of megakaryocytes and erythrocytes9. Important supporting evidence was obtained with the generation of a transgenic mouse line that expressed a
The stem cell to ELP transition Long-term repopulating HSCs (LT-HSCs), short-term repopulating HSCs (ST-HSCs) and various subsets of multipotent progenitors (MPPs) have distinctive prop‑ erties that are routinely used for their identification and cell sorting (TABLES 1,2) (see supplementary information S1 (tables) for a fully referenced version). Some markers are only useful for particular strains and ages of mice, and others are influenced by inflammatory cytokines. However, a consensus is developing that the expression of two classes of cell-surface antigens reflects progression of HSCs to ELPs (FIG. 1). Stem cells and all early progenitors are found within the lineage (LIN)– stem-cell antigen 1 (SCA1)+ KIThi
Gene-reporter mice Genetically engineered knock-in mice that allow tracking of the expression of a gene of interest. This is achieved by replacing the gene of interest (or part of it) by a sequence that encodes a reporter molecule, such as green fluorescent protein (GFP). When the promoter region of the gene of interest is activated, GFP is expressed and living cells can be visualized by flow cytometry.
Table 1 | Haematopoietic progenitors with B-cell lineage potential* Progenitor
Definition‡
Additional characteristics§
Lineage potential
Mouse bone-marrow cells LT-HSC
LIN–SCA1hiKIThiFLT3–THY1.1lowCD34–CD150+
Self-renewal capacity, VCAM1+CD44+ IL-7Rα–CD27–CD48–Ccr9–
Multipotent
ST-HSC
LIN–SCA1hiKIThiFLT3–/lowTHY1.1lowCD34+
Poor self-renewal capacity, VCAM1+IL‑7Rα– CD4lowMAC1low
Multipotent
MPP
LIN–SCA1hiKIThiFLT3low/+THY1.1– CD27+IL‑7Rα–
No self-renewal capacity, VCAM1+/–CD44+ IL-7Rα–CD62L+/–Ccr9–/low
Multipotent
LMPP
LIN–SCA1+KIThiFLT3hiTHY1.1–VCAM1–
No self-renewal capacity, CD34+,Tpor–/lowEpor– Gcsfr+Pu.1+Gata1–ScllowIl7rα+Rag1+ sterile Igh (GFP+ in IkarosGFP reporter mice)
T, B and myeloid cell
ELP
LIN–SCA1+KIThiFLT3hiTHY1.1–VCAM1– (GFP+ in Rag1GFP knock-in mice)
Oestrogen-sensitive, CD44+CD27+IL‑7Rα–TdT+/– human IGHM+DHJH+Rag1+/–Il7rα+
B, T and NK cell, pDC and IKDC (weak myeloid cell)
Pro-lymphocyte
LIN–SCA1lowKITlowFLT3+ (GFP+ in Rag1GFP knock-in mice)
Oestrogen-sensitive, CD27+IL‑7Rα+/–DHJH+
B, T and NK cell
CLP
LIN–SCA1lowKITlowFLT3+IL-7Rα–
CD44+THY1.1–Ccr9+
B, T and NK cell and DC (weak myeloid cell)
CLP
LIN–SCA1lowKITlowIL-7Rα–CD24–/low CD43lowCD93hiB220–
CD44+
B and T cell (weak myeloid)
CLP2
LIN–SCA1+KIT–IL-7Rα+CD19–B220+ (human CD25+ in Ptcra reporter mice)
CD4+/–CD24lowCD43+CD44hiCD93lowLY6C–/+ FLT3+Ccr9+
T and B cell
EPLM
B220+KITlowCD19–NK1.1–
IL-7Rα+FLT3+
Fraction A (pre-pro-B)
LIN KIT IL-7Rα LY6C CD19 CD24 CD43lowCD93hiB220+
Rag1/Rag2 Pax5 E12 E47 Ebf1 Cd79a Cd79b Tdt+DHJH+
B and T cell
Fraction B (early pro-B)
KITlowIL-7Rα+CD19+CD24+CD43+CD93hi B220+
CD4–CD25–Rag1/Rag2+Pax5+E12+E47+Ebf1+ Cd79a+Cd79b+Tdt+DHJH+
B cell
B-1-cell precursor
CD19+CD45RA+B220–
LIN–SCA1low/–KITlow/–IL-7Rα+CD4low/–CD24+ CD43+CD93+
B-1 and myeloid cell
–
low
+
–
–
–
+
T, B and myeloid cell +
+
+
+
+
+
Human bone-marrow cells HSC
LIN–CD34+CD38–
B, NK and myeloid cell and DC
CLP
LIN–CD10+CD34+CD45RA+
CD38+THY1–HLA-DR+RAG1/RAG2+EBF1+PAX5+
B, T and NK cell, DC (weak myeloid cell)
BMP
CD19+CD34+CXCR4–
CD133–CD90–
B and myeloid cell
*For a fully referenced version of this table, see supplementary information S1 (tables). ‡Antibodies to mouse lineage markers include TER119, CD11B, GR1, CD19 or B220, CD8α, CD3, NK1.1 or DX-5. For inclusion of CLP2 cells and Fraction A, B220 is omitted. THY1.1 marker can be used only in PL congenic mice. Human lineage markers include glycophorin A, CD14, CD19, CD56, CD3 and CD8. §Information in italics refers to gene transcripts and DNA rearrangements. BMP, B-cell/ myeloid common progenitor; CCR9, CC-chemokine receptor 9; CLP, common lymphoid progenitor; DC, dendritic cell; EBF1, early B-cell factor 1; ELP, early lymphoid progenitor; EPLM, early progenitor with lymphoid and myeloid potential; Epor, erythropoietin receptor; FLT3, FMS-related tyrosine kinase 3; GATA, GATA-binding protein; G-CSFR, granulocyte colony-stimulating factor receptor; GFP, green fluorescent protein; HSC, haematopoietic stem cell; IgH, immunoglobulin heavy chain; IGHM, heavy chain of IgM; IKDC, interferon-producing killer DC; IL-7R, interleukin-7 receptor; LIN, lineage; LMPP, lymphoid-primed multipotent progenitor; LT-HSC, long-term repopulating HSC; MAC1, macrophage receptor 1; MPP, multipotent progenitor; NK, natural killer; PAX5, paired box gene 5; pDC, plasmacytoid DC; Ptcra, pre-TCR α-chain; RAG, recombination-activating gene; SCA1, stem-cell antigen 1; SCL, stem-cell leukaemia factor; ST-HSC, short-term repopulating HSC; TdT, terminal deoxynucleotidyltransferase; Tpor, thrombopoietin receptor; VCAM1, vascular cell-adhesion molecule 1.
96 | february 2008 | volume 8
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS Sterile IgH RNA (Sterile immunoglobulin heavychain RNA). A transcript of an unrearranged IgH locus, which does not produce a functional protein. Its presence is thought to reflect the accessibility of the locus for recombination.
reporter element under the control of the promoter for the transcription factor Ikaros (also known as IKZF1)12. One third of the LSK cells in these mice expressed the reporter gene and most of these also expressed uniformly high levels of FLT3 and had little potential to develop into the megakaryocyte or erythrocyte lineages. Another study analysed the THY1.1–/low LSK subset and the downregulation of expression of vascular cell-adhesion molecule 1 (VCAM1)13. Again, FLT3hi cells were efficient granulocyte and/or lymphoid progenitors, but had little ability to produce platelets and erythrocytes. ELPs that can be identified according to their expres‑ sion of recombination-activating gene 1 (RAG1) rep‑ resent approximately 15–18% of those LSK cells that express the highest levels of FLT3 (Refs 5,6) and that are also referred to as lymphoid-primed multipotent progenitors (LMPPs)9. LMPPs are potent progenitors, and elegant single-cell analyses have revealed that at
least 38% of these cells have the combined potential to develop into T cells, B cells and myeloid cells2. A smaller number of LMPPs is restricted to the lymphoid line‑ age (that is, B cells and/or T cells), whereas more than 60% of LMPPs yielded NK‑lineage cells. Approximately 31% of LMPPs express lymphoid transcripts, such as RAG1, sterile IgH (immunoglobulin heavy chain) and interleukin‑7 receptor α‑chain (IL-7Rα; also known as CD127), but individual cells are not uniformly posi‑ tive. This is in agreement with the original description of ELPs, that is, single cells that are positive for TdT, a human IgHm (heavy chain of IgM) transgene or both5. A subsequent study used mice in which the gene encod‑ ing green fluorescent protein (GFP) had been integrated into the Rag1 locus (termed Rag1GFP knock-in mice), such that GFP expression correlates with endogenous RAG1 expression 6, and found that individual ELPs expressed RAG1, TdT or both. Thus, lymphoid-lineage
Table 2 | Practical indicators for early events in B-cell lymphopoiesis* Marker‡
Use
Caveats
SCA1 (Ly6A)
Excludes myeloid-cell-biased progenitors; expressed by stem cells and MPPs
Strain dependent and upregulated in response to interferons
KIT (CD117)
Levels of expression determine immaturity of cell populations
Cells can re-acquire or upregulate this marker in culture
THY1.1 (CD90.1)
Enriches for LT-HSCs and ST-HSCs; lost at the MPP stage
Can only be used in PL strain of congenic mice
CD150 (SLAM)
Specifies LT-HSCs if used in conjunction with CD48– and CD41–
None
VCAM1 (CD106)
Expressed by HSCs and most MPPs; negative population of LSK cells enriches for lymphoid-cell-biased progenitors
Expressed by bone-marrow endothelial and stromal cells
FLT3 (FLK2, CD135)
Upregulated early by ST-HSCs; 25% of LSK cells expressing highest levels appear to be lymphoid-cell biased
Continuous range from low to high density with no clear separation within the LSK-cell population
CD34
In contrast to progenitors, adult HSCs express small amounts or none
Developmental age dependency; expressed by endothelial and stromal cells
CD27
Member of the TNF receptor family expressed by many progenitors, but LT-HSCs are negative
None
TdT
An early, lymphocyte-specific protein with functional consequences; marks a population of LSK cells referred to as ELPs
Not very abundant, intracellular protein; no monoclonal antibody available
AA4.1 (CD93)
Highly expressed on CLPs and some LSK cells; useful early B-cell lineage antibody in adults; highly expressed by fetal HSCs
493 monoclonal antibody recognizes a different epitope of the same molecule; there is controversy concerning expression within LSK subset for T- and B-cell progenitors; highly expressed by most adult BALB/c LSK cells but not B6 LSK cells (R. Hardy, personal communication)
CD24
Expressed on CLPs and levels distinguish fractions within the B-cell lineage
Not all monoclonal antibodies are equivalent; 30F used in the Hardy scheme; tendency to aggregate cells
L-selectin (CD62L)
Expressed by T-cell-lineage-biased and lymphoid DC progenitors
Marks a population of cells with reduced B-cell-lineage potential that needs to be further refined
RAG1GFP
GFP knock-in used to isolate pro-lymphocytes and some ELPs
Recognizes only a subpopulation of ELPs
RAG2GFP
NG-BAC transgenic reporter for RAG2-expressing cells
Some primitive GFP+ cells do not express endogenous RAG2
IkarosGFP
GFP+ LSK cells correlate with the FLT3hi lymphoid-biased progenitors
Some endogenous Ikaros+ cells do not express the GFP reporter
IL-7Rα (CD127)
Transcripts identified in the FLT3hi LSK subset; protein defines CLPs
Fetal mouse and adult human lymphoid progenitors are not totally IL-7 dependent
*For a fully referenced version of this table, see supplementary information S1 (tables). ‡Bone-marrow populations are typically lineage depleted using antibodies against: CD3, CD8, CD11C, CD19 and DX5 or NK1.1, B220, TER119, GR1 and MAC1. Fetal progenitor cells express low levels of MAC1. Additionally, CD4 may be passively acquired, has a variegated pattern of expression on lymphoid progenitors and should be avoided. CD19 is a later and ‘almost’ definitive marker for the B-cell lineage. Alternative protein names are provided in parentheses. CLP, common lymphoid progenitor; DC, dendritic cell; ELP, early lymphoid progenitor; FLT3, FMS-related tyrosine kinase 3; GFP, green fluorescent protein; HSC, haematopoietic stem cell; IL-7Rα, interleukin-7 receptor α-chain; LSK, LIN–SCA1+KIThi ; LT-HSC, long-term repopulating HSC; MPP, multipotent progenitor; RAG, recombination-activating gene; SCA1, stem-cell antigen 1; ST-HSC, short-term repopulating HSC; TdT, terminal deoxynucleotidyltransferase; TNF, tumour-necrosis factor; VCAM1, vascular cell-adhesion molecule 1.
nature reviews | immunology
volume 8 | february 2008 | 97 © 2008 Nature Publishing Group
REVIEWS
Lymphoid transcripts (such as Rag1 and Tdt)
ELPs ETPs and T cells LMPPs
B and NK cells pDCs NK-cell-like IKDCs
LT-HSCs Conventional DCs
MPPs ST-HSCs VCAM1
FLT3
Figure 1 | Founders of the immune system emerge gradually from an extremely rare subset in the bone marrow. Stem cells and all early cellNature progenitors found in a Reviewsare | Immunology rare bone-marrow subset that is lineage (LIN)– stem-cell antigen 1 (SCA1)+KIThi (known as the LSK subset). Within this subset, the progressive acquisition of expression of FMSrelated tyrosine kinase 3 (FLT3) occurs in synchrony with loss of expression of vascular cell-adhesion molecule 1 (VCAM1) and the initiation of the expression of lymphoid-cell associated genes. According to the levels of expression of FLT3 and VCAM1 together with other parameters, LSK cells can be subdivided into three groups: long-term repopulating haematopoietic stem cells (LT-HSCs), short-term repopulating HSCs (ST‑HSCs) and multipotent progenitors (MPPs). However, the transition between these cell types represents a continuum. Within the LSK subset, cells with the highest amount of FLT3 expression (representing 25% of the total LSK subset) have been referred to as lymphoid-primed multipotent progenitors (LMPPs) and approximately one third of these express transcripts for lymphoid genes, such as Rag1 (recombination-activating gene 1) and Tdt (terminal deoxynucleotidyltransferase). Within the LMPP subset early lymphoid progenitors (ELPs) are cells that express the products of these genes and are assumed to be lymphoid-cell specified. They are the most efficient cells in terms of their ability to replenish B and T cells. In ELPs, non-lymphoid-cell potential is decreased, but not completely lost, and their progeny can generate dendritic cells (DCs) under some circumstances. Plasmacytoid dendritic cells (pDCs) are the closest relatives of B cells, whereas conventional DCs are considered to be myeloid. Interferon-producing killer dendritic cells (IKDCs) are very similar to conventional natural killer (NK) cells, but may arise independently. ETP, early T-cell-lineage progenitor.
specification occurs in FLT3hi cells of the LSK subset, but it does not involve synchronous expression of all lymphoid genes. VCAM1 expression is emerging as another valuable marker for early events in lymphoid-lineage specifica‑ tion. An initial discovery was that it was constitutively expressed by bone-marrow stromal cells and that it medi‑ ated attachment of very late antigen 4 (VLA4)+ lymphoid progenitors in long-term cultures14. Microarray screens later revealed that it was also expressed by HSCs but was downregulated upon differentiation of these cells11,15. Important experiments by Lai and Kondo showed that whereas all THY1.1low ST‑HSCs and LT‑HSCs express VCAM1, the FLT3hi cells (that is, LMPPs) can be sub‑ divided according to the expression of this marker. That is, VCAM1– LSK cells represent a subset of LMPPs (FIG. 1) that do not have the potential to generate eryth‑ rocytes but that can rapidly generate B and T cells11,13,16. Clonal analysis revealed these cells to be a mixture of myeloid-restricted, myeloid- and lymphoid-restricted and lymphoid-restricted progenitors. Using Rag1GFP 98 | february 2008 | volume 8
knock-in mice, approximately 82% of cells with these characteristics (FLT3hiVCAM1–) would be designated ELPs16. Therefore, increased expression of FLT3 and loss of VCAM1 expression allow identification of the earliest lymphoid-specified progenitors in the LSK fraction of the bone marrow (FIG. 1; TABLE 1). Several other markers are useful in tracking the progres‑ sion of HSCs into the lymphoid lineages (TABLES 1,2). For example, transient expression of CD4, and the integrins MAC1 (also known as integrin-αM) and integrin‑α2 (also known as CD49b) have been used as markers of ST‑HSCs17,18. HSCs in THY1.1 congenic but not con‑ ventional mice are contained within the 8% of LSK cells that display low levels of THY1.1. Approximately 40% of these THY1.1low LSK cells express CD150 and, when used together with an additional marker (CD48), this defines CD150+CD48– LT‑HSCs19. CD34 is expressed by fetal and neonatal HSCs but they stop expressing it by 10 weeks after birth and it is only re-expressed on stem-cell activation20,21. Most LT‑HSCs are CD27–, whereas ELPs are among the 90% of LSK cells that display this marker5,6,22. AA4.1 (also known as CD93) is expressed by fetal but not adult HSCs 23,24 and has been extensively used to study B‑cell-lineage differen‑ tiation in adult bone marrow25,26. However, the ELPs identified in Rag1GFP knock-in mice were AA4.1–/low (R.S.W., unpublished observations), and a substantial number of T‑cell-lineage precursors might lack this marker27. The expression of L‑selectin (also known as CD62L) is upregulated by most RAG1+ ELPs, and this marker has been used to enrich subsets of early progenitors28–31. So, we conclude that FLT3 and VCAM1 are particu‑ larly useful markers for subdividing cells within the rare LSK subset of the bone marrow. When these markers are exploited in combination with other markers and reporter mice, it is possible to isolate and extensively study the founders of the immune system.
Transcription factors enable and support Much has been learned about the genes required for lymphoid-lineage specification and commitment, and it is interesting to consider those that are responsible for the changes described above. Although spheres of influ‑ ence are becoming reasonably defined for Ikaros, PU.1, EBF1 (early B‑cell factor 1), E2A and PAX5 (paired box protein 5), there is now evidence that additional tran‑ scription factors contribute to lymphoid-lineage speci‑ fication and commitment. Early haematopoietic cells in Ikaros-deficient mice fail to express FLT3 (Refs 32,33). Although FLT3 is therefore not available as a marker for the identification of ELPs in these mice, the subset of LSK cells that would normally include them has been imaged using a recently generated Ikaros reporter mouse model12. Importantly, Ikaros-deficient animals lacked ELPs defined according to RAG1 or IL‑7Rα expression, and LSK cells from these mice had a reduced potential to generate T‑lineage cells but had no B-cell‑lineage differentiation potential. These and other studies establish that Ikaros is an enabler of the earliest known events in lymphopoiesis. It will be exciting to understand www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS
B‑2 cells IgMlowIgDhiMAC1–B220hiCD23+ cells that originate from bone marrow and are distributed to mucosal and systemic immune compartments for the continuous secretion of antibodies with high affinity and fine specificity.
B‑1 cells IgMhiIgDlowMAC1+B220low CD23– cells that are dominant in the peritoneal and pleural cavities. Their precursors develop in the fetal liver and omentum, and in adult mice, the size of the B‑1-cell population is kept constant owing to the self-renewing capacity of these cells. B‑1 cells recognize self components, as well as common bacterial antigens, and they secrete antibodies that tend to have low affinity and broad specificity.
whether Ikaros directly controls the expression of FLT3, RAG1, IL‑7Rα and other essential genes in ELPs. The related transcription factor Aiolos silences such genes at appropriate, later stages of differentiation34, indicating that members of this family both initiate and sustain lymphopoiesis of the B-cell lineage. Mouse embryos with deletions of the transcription factor PU.1 (which is encoded by the gene Sfpi1) are non-viable, and their haematopoietic cells are severely compromised in their ability to make conventional B‑2 cells. Similar to Ikaros-deficient mice, FLT3 expres‑ sion is not upregulated by developing haematopoietic cells, and a series of downstream B-cell‑lineage genes are not expressed35. By contrast, PU.1 expression is not required for generation of the specialized B‑1-cell subset of lymphocytes, and its downregulation can even be a determinant of B‑1-cell production36. Contrary to what was once believed, the levels of PU.1 expression do not determine lymphoid- versus myeloid-lineage choices, although high concentrations of this factor can suppress lymphopoiesis37. Clearly, PU.1 has important, but incom‑ pletely understood roles in early lymphopoietic cells. Many B‑cell-lineage genes are direct or indirect targets of the basic helix-loop-helix (bHLH) family of transcription factors. These include EBF1, PAX5, λ5, Igα chain (also known as CD79a and MB‑1) and RAG1 (Refs 38,39). One such bHLH transcription factor is E2A (which is comprised of E47 and E12); its deletion com‑ pletely prevents B‑cell production and immunoglobulin gene recombination in the B-cell lineage40. However, bone marrow from E2A- or E47-deficient mice contained cells corresponding to the earliest stages of lymphopoiesis (that is, cells with an LSK phenotype that have mRNA transcripts encoding FLT3 and IL‑7Rα), reduced tran‑ scripts for Igβ chain (also known as CD79b and B29) and normal competence for producing T cells41,42. Moreover, MPP cell lines with selective deficiency for B-cell lymphopoiesis have been isolated from E2A-deficient bone marrow and expanded in culture with cytokines38. Therefore, E2A is dispensable for ELP formation but important for subsequent differentiation events. Activities of E2A and other members of the bHLH family are negatively regulated by transcriptional repressors, such as inhibitor of DNA binding 1 (ID1), ID2, ID3 and ID4 (Ref. 40). In the haematopoietic system, the transcription factor EBF1 is expressed almost exclusively in B‑lineage cells43. Moreover, the Ebf1 gene is active in early lymphopoietic cells sorted from fetal and adult Rag1GFP knock-in mice6,44. Artificial overexpression of EBF1 in haematopoietic cells was sufficient to direct their differentiation to the B-cell lineage, demonstrating the potent instructive ability of this transcription factor45. Similar to E2A-deficient mice, Ebf1–/– mice have ELPs defined by their potential to dif‑ ferentiate into the T-cell lineage, their CD27 expression and high FLT3 expression, as well as their expression of transcripts encoding TdT, E2A, sterile IgH and IL‑7Rα (Refs 35,46). Finally, the importance of EBF1 is shown by the fact that cell progenitors from Ebf1–/– mice show no immunoglobulin gene recombination and do not activate a series of essential genes.
nature reviews | immunology
CD19 is one of several direct targets of the transcrip‑ tion factor PAX5, and its expression almost perfectly coincides with commitment to the B‑cell lineage (see later). Furthermore, PAX5 is essential for maintaining B‑cell-lineage characteristics and simultaneously sup‑ pressing the differentiation to other haematopoietic lineages47. Pax5 mRNA is not detectable in ELPs from Rag1GFP knock-in mice or in LSK FLT3hi LMPPs2,6. Furthermore, experiments with PAX5 reporter mice indicate that its expression is only homogeneous in more mature CD19+ cells in the lineage48. Although less well studied, other transcription fac‑ tors also contribute to B-cell lymphopoiesis. For exam‑ ple, progenitors taken from fetal B‑cell lymphoma 11a (BCL‑11a)-deficient mice generated few B cells and were said to lack transcripts encoding EBF1, PAX5, IL‑7Rα and CD19 (Ref. 49), indicating that BCL‑11a is important at an early stage. Targeting of the proto-oncogene Pbx1 (pre‑Bcell leukaemia transcription factor 1) is also embryonic lethal, but HSCs from fetal liver were found to be very compromised with respect to B-cell lymphopoiesis50. Conditional targeting of the same gene in adults showed that it has no role beyond the CD19+ stage. Finally, cells recovered from fetuses lacking forkhead box P1 (FOXP1) generated few B cells in adoptive transfer assays51. There were cell progenitors in fetal liver that could be expanded with cytokines, and patterns of gene expression suggested that B-cell‑lineage development was not totally blocked. Nevertheless, further investigation is needed to see how these three transcription factors contribute to the formation or differentiation of ELPs. A sequence of events was suggested by the fact that PU.1 directly regulates the genes encoding EBF1 and IL‑7Rα (Refs 35,52). Accordingly, progression to the B-cell lineage by PU.1-deficient cell progenitors was made possible in culture with the artificial introduc‑ tion of Ebf1 and, to a lesser extent, Il7ra, but not Pax5. Experiments with a tumour-cell line demonstrated that the E2A component E12 also regulates EBF1 expression53. Furthermore, B‑cell-lineage defects were most severe when E2A heterozygous mice were crossed with EBF1 heterozygous mice54. Although EBF1 is an important initiator of B-cell lymphopoiesis, it is noteworthy that its levels are considerably augmented at later stages of devel‑ opment through IL‑7Rα- and STAT5 (signal transducer and activator of transcription 5)-mediated signals55. Also, one of two Ebf1 promoters is positively responsive to PAX5 (Ref. 56). This establishes a feed forward situation whereby progression and maintenance of committed progeni‑ tors can be sustained. Such complex inter-relationships have been discussed in recent reviews39,57. To summarize, mechanisms responsible for the early suppression of megakaryocytic and erythroid develop‑ ment, for the generation of FLT3hiVCAM1– LSK cells and for the initiation of lymphopoiesis in this rare bonemarrow subset are not completely understood. However, Ikaros has now been shown to be a transcription factor that is absolutely required for the earliest events of lym‑ phopoiesis, and the progression of some cells towards the T-cell lineage in Ikaros-deficient mice indicates that a degree of fate determination might be occurring even volume 8 | february 2008 | 99
© 2008 Nature Publishing Group
REVIEWS at this early stage. Investigation of the functions of PU.1, E2A and PAX5 has provided excellent examples of how transcription factors can cooperate and cross-compete in complex ways.
Osteoclasts Multinucleated macrophagerelated cells that absorb bone.
Angiopoietins A family of proteins that promote the formation of new blood vessels.
Early environmental cues Other important issues in lymphopoiesis relate to the nurturing environments for HSCs and the nature of extracellular signals that regulate the earliest events. The pool of subendosteal osteoblasts (cells that line the inner bone surface) has long attracted attention as a potential ‘niche’ to support early lympho-haematopoiesis, but most HSCs are more centrally located in the bone marrow and near perivascular stromal cells19. It will be a chal‑ lenge to define such niches because cells with the same mesenchymal origin may assume different functions and produce factors that vary in response to their location. For example, stromal cell lines commonly used to sup‑ port lympho-haematopoiesis are multipotent and can be induced in culture to become osteoblasts or fat cells. Lymphoid progenitors are scattered throughout the bone marrow, and the earliest ones preferentially interact with VCAM1+ CXC-chemokine ligand 12 (CXCL12)producing reticular cells58,59. Although it is clear that CXCL12 is important for attracting and retaining lymphoid progenitors in the bone marrow, the nature of other important factors produced by reticular cells is undefined. Distinctly different from reticular cells, VCAM1+ stromal cells produce IL‑7, which is needed to support the survival and expansion of progenitors, as well as for immunoglobulin gene recombination. A possibly comparable situation has been found in human bone marrow, where a series of lymphoid progenitors were aligned in close association with VCAM1+/–CD10+ stromal cells60. In addition to these components of the bone marrow, knockout studies indicate that endothelial cells and osteoclasts can also contribute somehow to haematopoiesis61,62. Also, HSCs develop, reside and/or traffic in other organs63,64. Given the complexity and uncertainty about cells that comprise niches in the bone marrow, it is surprising that progress is being made in the identification of molecules they produce that control the survival, mitotic activity and differentiation of haematopoietic cells65,66. Notch ligands, stem-cell factor (SCF), thrombopoietin, IL‑3, IL‑11, IL‑6, insulin-like growth factor 2, fibroblast growth factor and WNT3a have all been used to support modest expansion of LT‑HSCs without loss of self-renewal potential67–70. However, gene-targeting experiments have yet to dem‑ onstrate that any of these molecules or their receptors are absolutely essential for lympho-haematopoiesis. The LSK subset is defined by high levels of expression of the transmembrane protein tyrosine-kinase receptor KIT, and ELPs are responsive to its corresponding ligand, SCF. Although it is clear that this pair of molecules maintains HSC homeostasis, an important role in B-cell lymphopoiesis for KIT–SCF associated signals is only obvious many weeks after birth71. As discussed above, upregulation of the structurally related protein tyrosine kinase FLT3 parallels the formation of ELPs, however, targeting this receptor protein kinase has only modest
100 | february 2008 | volume 8
effects on B-cell‑lineage development72. This discrepancy could be explained by the potential presence of another, as yet unknown, receptor for the corresponding cytokine FLT3 ligand (FLT3L), as more substantial deficiencies in B‑cell precursors have been found in B-cell precur‑ sors from mice that lack this cytokine73,74. Nevertheless, numbers and functions of HSCs in FLT3L-deficient mice were normal, and they could generate T cells74. So, no single HSC-expressed or MPP-expressed cytokine receptor has yet been identified that is absolutely and selectively needed for generating the base components of the immune system. Although simultaneous deletion of Kit and Flt3 has demonstrated that these functionally related receptors cooperate72, it will be important to know if they induce the expression of Ikaros, which is required for the initiation of B-cell lymphopoiesis. Impressively high B‑cell-lineage cloning efficiencies were reported when ELP-enriched LMPPs were cul‑ tured under relatively simple conditions, that is, serumfree cultures containing SCF, FLT3L and IL‑7 (Ref. 8). Although clonal expansion is even higher in stromalcell co-cultures2,75, it is clear that these three factors act sequentially and cooperatively to support HSCs, MPPs and the early stages of B-cell lymphopoiesis. In addition to these positive regulators, there are many factors that constrain clonal expansion and lineage pro‑ gression at this early stage. Nearly all adult bone-marrow HSCs, and most ELPs spend a considerable amount of time in a quiescent, G0 state of cell-cycle arrest76–78. This status may be important for controlling population sizes and the life-long integrity of cells that replenish the immune system. Some mechanisms by which this regulation may occur have been proposed. Angiopoietins and osteopontin are thought to regulate stem-cell num‑ bers79–81, but information is lacking about their effects on very early lymphoid cells. The formation of ELPs from HSCs is blocked by IL‑6, and progenitors seem to be re-directed to a myeloid-cell fate by this cytokine75,82. Interestingly, IL‑6 has no negative influence and in fact enhances the proliferation of ELPs. Stage-specific action may also be a feature of prostaglandins, which although long known to induce apoptosis in B-cell‑lineage pro‑ genitors, they have been recently described as stimulants for HSC self renewal83,84. There is also substantial evi‑ dence to suggest that steroid hormones selectively inhibit B-cell lymphopoiesis under steady-state conditions85. More controversial are the functions associated with members of the complex WNT family of molecules. For example, there are conflicting reports concerning the physiological importance of WNT3a in maintaining HSCs69,86,87. One recent report concludes that although WNT signalling is at least needed for normal T-cell lym‑ phopoiesis, it is accomplished without dependence on its canonical signalling partners β‑ or γ‑catenin88. However, artificial stimulation of the canonical signalling pathway frequently used by WNT3a dramatically inhibits, and even reverses lineage progression. That is, it caused lymphoid- or myeloid-restricted progenitors to become multipotent89,90. Furthermore, WNT5a appears to limit the expansion of B‑lineage cells at a later stage91, and other WNT-related molecules merit additional study. www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS Box 1 | B-cell lymphopoiesis in humans Although broad concepts in lymphopoiesis may be extrapolated from experimental animals to humans, it is clear that there are significant species differences. Long-term cultures and mouse xenotransplantation models are being used to test the self-renewal capacity and/or differentiation potential of early human progenitor cells. The majority of early precursor cells express CD34 but lack lineage markers, and initial expression of CD38 correlates with lineage commitment. Human equivalents of mouse early lymphoid progenitors (ELPs) are still being characterized, but CD7+CD34+CD38– and CD7+CD34+CD45RAhi cells in umbilical cord blood generate no myeloid-cell progeny and contain clonal progenitors of B or natural killer (NK) cells and T or NK cells, respectively. Although the CD7+CD34+CD45RAhi cells can produce all three lymphoid lineages, they are skewed towards T or NK cells and have been speculated to be thymus-seeding progenitors. Moreover, these cells are rare in adult bone marrow, and their relationship to the B-cell lineage requires further study. The LIN–CD10+CD34+ candidate common lymphoid progenitors (CLPs) in human bone marrow are predominately B‑cell-lineage progenitors and they also have the potential to differentiate into T cells, dendritic cells and NK cells (TABLE 1). Furthermore, gene expression profiles of LIN–CD10+CD34+ are consistent with affilitation to the B‑cell lineage (that is, they express genes that encode RAG1 (recombination-activating gene 1), RAG2, IL‑7Rα (interleukin-7 receptor α-chain), EBF1 (early B-cell factor 1), PAX5 (paired box protein 5), immunoglobulin β-chain and the surrogate light chain VpreB). Although less information is available on lymphopoiesis in humans than in mice, subsequent changes that result in CD19 expression are being documented, and B‑cell precursors are responsive to, but not dependent on IL‑7.
OP9-DLL1 system Stromal cells derived from osteopetrotic OP/OP mice (OP9 cells) are useful in cocultures initiated with myeloid progenitor cells because they do not produce macrophage colony-stimulating factor 1 that can cause excessive generation of macrophages and prevent the development of lymphoid cells. When stably transduced to overexpress the Notch ligand delta-like-ligand 1 (DLL1), these cells can support a surprising degree of progression in the T‑cell lineage. Adherent layers of OP9-DLL1 stromal cells function like an artificial thymus in culture when lymphoid progenitors are added.
Plasmacytoid dendritic cells (pDCs). A subset of DCs the microscopic appearance of which resembles plasmablasts. In humans, these DCs can be derived from lineage-negative stem cells in peripheral blood and are the main producers of type I interferons (IFNs) in response to virus infections. Recent studies have identified subsets of type‑I-IFNproducing DCs in mice, which are identified by expression of B220, Ly6C and other markers.
One multifaceted effector of lymphopoiesis is Notch. Transient ligation of Notch receptors inhibits B-cell lymphopoiesis, whereas sustained stimulation of the Notch signalling pathway is required to support pro‑ gression in the T-cell lineage92. This critical determinant of T‑cell versus B‑cell lineage differentiation has been extensively studied, and many distinctly different types of bone-marrow cell can be affected by Notch signals (see discussion below). Of particular interest is the fact that the sensitivity of lymphoid progenitors to Notch signals is regulated in multiple ways. For example, the transcriptional repressor leukaemia/lymphoma related factor (LRF) normally blocks proximal Notch signalling events in bone-marrow cells and the increased sensitivity of LRF-deficient progenitors causes differentiation into the T‑cell lineage rather than to the B‑cell lineage93. As another level of control, B‑lineage cells are unique in that Notch signals result in degradation of the essential transcription factor E2A94. Thus, an impressive number of extracellular molecules has been identified that support initiation of and progres‑ sion within the B‑lymphoid lineage. None is uniquely and totally required, and their actions are opposed by a series of other factors. For example, certain disease circum‑ stances can divert the effects of IL‑6 to an alternative HSC differentiation process; later we discuss how pathogen products can do the same. Experiments with the WNT family of molecules suggest that at least some early steps in B-cell lymphopoiesis might be reversible.
Succeeding stages of lymphopoiesis The progeny of ELPs are known by a variety of names, but they represent a series of progenitors that are progres‑ sively more likely to become lymphocytes (TABLES 1,2). In a landmark paper, Kondo and colleagues described LSK IL‑7Rα+ common lymphoid progenitors (CLPs) in
nature reviews | immunology
mouse bone marrow, and showed that at least some indi‑ vidual cells with these characteristics had the potential to differentiate into both T cells and B cells95. The same fraction also generated NK cells when transplanted, but no signs of non-lymphoid differentiation were observed at the time. Subsequent studies established that CLPs represent major intermediates in the B‑cell- and NK-cell‑ lineage pathways but are not totally dedicated to those fates (reviewed in Ref. 58). For example, several investiga‑ tors have found that highly purified CLPs can transiently produce small numbers of dendritic cells (DCs) and myeloid cells26,29,96. Importantly, there is near consensus that CLPs do not contribute substantially to T‑cell-lineage development92. Although T‑cell-lineage lymphocytes can be generated when IL‑7Rα+ progenitors are injected into the thymus or cultured with the stromal-cell line OP9 that expresses the Notch ligand delta-like ligand 1 (DLL1; a process known as the OP9-DLL1 system), they are unlikely to do so under normal circumstances29,95,97,98. Investigators should carefully specify cell-separation conditions, and it may be time for an international work‑ shop to consider standardized terminology. For example, our laboratory uses the term CLP to refer to cells with the exact phenotype that was originally described by Kondo and colleagues. However, Allman and colleagues use CLP to refer to LIN–IL-7Rα+AA4.1+SCA1low/– bonemarrow cells99, whereas Hardy and colleagues now refer to LIN–IL-7Rα+KIT+CD24lowCD43lowAA4.1hi cells as CLPs26. Nonetheless, there is agreement that acquisition of IL‑7R expression by LIN– cells with uniquely low lev‑ els of KIT expression represents an important milestone in B‑cell production. Important parallel studies are also being conducted with human cells (Box 1). Most subsequent differentiation steps are well rep‑ resented in the extremely useful scheme developed by Hardy and colleagues. They classified subsets of B‑cell progenitors into multiple fractions: Fraction A (which includes pre-pro‑B cells), Fraction B (which includes pro‑B cells) and Fraction C (which includes large pre‑B cells)25. Furthermore, it has been refined over the years, exploiting new advances in flow cytometry and the availability of new markers. The earliest progenitors of the original series of the subsets defined by Hardy and colleagues (B220+CD24low; Fraction A) is likely to be downstream of ELPs and CLPs, and it includes a surpris‑ ing variety of cells in lymphoid, NK‑cell, plasmacytoid DC (pDC) and other DC lineages 100. As one example, Fraction A includes B220+KITlowCD19–NK1.1– bonemarrow cells referred to by Rolink and colleagues as early progenitors with lymphoid and myeloid poten‑ tial (EPLM)96. With the exception of B220 expression, these EPLMs are very similar to CLPs. Importantly, some EPLMs have been shown to generate both T and B cells, and they retain some residual potential to generate macrophages in cultures with stromal cells. Treatment of mice with FLT3L results in the expansion of the number of EPLMs at the expense of B cells, suggesting they might have been directed to DC fates101. Another population within Fraction A is thought to have combined B‑cell and T‑cell-lineage potential, and has been designated as the CLP2 subset102. It has so far only been identified volume 8 | february 2008 | 101
© 2008 Nature Publishing Group
REVIEWS in transgenic pTCRα reporter mice and differs from EPLMs in that it expresses low levels of KIT (TABLE 2). The acquisition of CD122 (the β‑chain for the IL‑2 and IL‑15 receptor) by Fraction A cells signals commit‑ ment to the NK‑cell lineage and is followed by NK1.1 and DX5 expression58,103,104. Additionally, Ly6C+ and/or CD4+ subsets of Fraction A represent DC lineages31,99,105. B220+Ly6C+CD11c+CD19–CD43+DX5– pDCs in the bone marrow are closely related to B cells and can be Stem cells and multipotent progenitors: LIN–SCA1+KIThi
Early thymic progenitors: LIN–SCA1+KIThiFLT3+/–TdT+
LT-HSCs CD62L+
ETPs TdT+ ELPs RAG1+ T cell
NK-cell-like IKDC
Rag1++ RAG1
CLPs
B220–/low CD19+ pDC
B-1-cell progenitors
IL-7Rα+
B220+ CD19+ KIT+
NK cell Pre-proB cells Pro-lymphocytes: LIN–SCA1+/lowKITlowFLT3+
B-1 cell
pro-B cells
B-2 cell
Fraction A: B220+CD19–CD24low
Figure 2 | Progression through the earliest stages of lymphopoiesis is gradual, but bone-marrow cells can be categorized on the basisNature of expression various Reviews | of Immunology markers. Stem cells, including long-term repopulating haematopoietic stem cells (LT-HSCs), and multipotent progenitors in the bone marrow are present within the lineage (LIN)– stem-cell antigen 1 (SCA1)+KIThi subset (termed the LSK subset). A subset of these progress to become early lymphoid progenitors (ELPs), which express one or more proteins associated with lymphoid-cell lineages, such as RAG1 (recombinationactivating gene 1) and TdT (terminal deoxynucleotidyltransferase). The transcription factors PU.1 and Ikaros (not shown), together with signals from the cytokine receptors KIT and FMS-related tyrosine kinase 3 (FLT3), are needed to maintain normal numbers of ELPs. Closely related ELPs with partially overlapping properties can generate the precursors for B cells, T cells, natural killer (NK) cells, interferon-producing killer dendritic cells (IKDCs) and plasmacytoid DCs (pDCs). However, most B and NK cells are generated from a pro-lymphocyte subset containing common lymphoid progenitors (CLPs). Although CLPs and other categories of progenitors, such as CLP2 cells found in certain reporter mice (not shown), can generate T cells under experimental conditions, they are unlikely to represent a major source of early thymic progenitors (ETPs). The transcription factors early B‑cell factor (EBF), E2A and paired box protein 5 (PAX5) (not shown) function in a cooperative way with signals from the interleukin‑7 receptor α-chain (IL-7Rα) to support transit through the CLP and pre-pro‑B-cell stages. Note that a category long designated as Fraction A is now divided into subsets that include NK‑ and pDC-lineage cells, in addition to pre-pro‑B cells. Pre-pro‑B cells give rise to CD19+ pro‑B cells that seem to be firmly committed to becoming conventional B‑2 cells. A recently described category of B220–CD19+ progenitors are restricted to the production of B‑1 cells, but their origin is unknown. 102 | february 2008 | volume 8
subdivided into at least two stable subsets31. One subset is particularly related to the B-cell lineage with respect to the expression of RAG1, EBF1, Igα and BCL‑11a. Although this type of pDC contains Pax5 transcripts, it lacks cell-surface CD19 expression and is present in Pax5–/– mice. Similar to B cells, the development of pDCs is blocked by signalling through the Notch pathway31,106. Hardy and colleagues further narrowed the definition of B220+CD24low Fraction A to Ly6C– AA4.1hiCD43lowKITlowIL-7Rα+ cells, again differing from CLPs only with respect to B220 expression26. This revised definition now excludes most cells with a potential to differentiate into the myeloid-cell lineage, although the fraction still contains some cells that, in culture, can be induced to differentiate into the T‑cell lineage. One important conclusion from these and related studies is that acquisition of B220 expression alone does not signal restriction to the B-cell lineage. The cardinal trait of cells in Hardy’s Fraction B is the expression of CD19, and this event is totally dependent on its targeting by PAX5 (Ref. 100). There are many other PAX5-dependent targets, and an important role of this transcription factor is to inhibit the ability of B‑celllineage progenitors to develop into other lineages 47. This was demonstrated by conditional gene targeting experiments that showed that extensively self-renewing, multipotent cell lines could be established from Pax5knockout mice107. Similar to EPLMs, these cell lines were B220+ but otherwise similar to CLPs. It is remarkable that mature B cells can be induced to become T cells just by depriving them of PAX5, and it could be possible that PAX5 may become downregulated in some malignancies associated with the B-cell lineage47. CD19 expression is likely to occur after the upregu‑ lation of B220 by most B‑cell progenitors. However, Dorshkind and colleagues described rare (0.1–0.4% of nucleated bone-marrow cells) CD45RA+B220–CD19+ cells that developed into lymphocytes exclusively in the B220 lowIgM hiIgD lowCD43 +CD23 –MAC1 +CD5 +/– B‑1a-cell and B‑1b-cell subsets108,109. These B‑1-cellrestricted progenitors were IL‑7Rα + and clonally expanded in the presence of IL‑7. Although some of them also produced small numbers of macrophages in culture, they were poorly responsive to myeloid-cell growth and differentiation factors. Expression of B220 is dependent on PU.1 (a transcription factor encoded by Sfpi1), and it is therefore interesting that condi‑ tional deletion of Sfpi1 in CD19+ progenitors biases their differentiation towards B‑1 cells35,36. Production of B‑1 cells and conventional B‑2 cells may normally be supported by thymic stromal lymphopoietin (TSLP) and IL‑7, respectively110; however, it is interesting to note that these two important cytokines are redundant when artificially made available at appropriate concen‑ trations111. B‑1-cell progenitors are unable to produce T cells, NK cells or DCs, and are predominantly but not exclusively present during fetal life109. Given the apparent importance of B‑1 cells in innate immunity, the relationship of these progenitors to those cells that are responsible for generating conventional B cells merits further study. www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS MPPs
Pathogen or TLR ligands
ELP HSCs
LT-HSC
ST-HSC
LMPPs
FLT3lowVCAM1+
Prolymphocytes
Fraction A
CLP
Pre-pro-B cell
FLT3hiVCAM1–
Pathogen or TLR ligands
Pathogen or TLR ligands
Pro-B cell
B cell
NK cell
CMP NK-cell-like IKDC Monocyte
cDC pDC
Figure 3 | Early stages of B-cell lymphopoiesis can be influenced by microbial and viral products. Stem cells and Nature Reviews | Immunology progenitors of lymphoid and myeloid cells express functional Toll-like receptors (TLRs) and associated molecules. This allows recognition of endogenous TLR ligands that might be released from damaged tissues, as well as substances made by pathogens. Normally quiescent haematopoietic stem cells (HSCs) are driven into the cell cycle and their differentiation can be promoted by exposure to the TLR ligand lipopolysaccharide, and myeloid progenitors can mature even in the absence of growth and differentiation factors. Common myeloid progenitors (CMPs) normally derive from vascular celladhesion molecule 1 (VCAM1)+ multipotent progenitors (MPPs) under the influence of cytokines. The ability of common lymphoid progenitors (CLPs) to generate B cells is curtailed by TLR2, TLR4 or TLR9 ligands, whereas the production of several categories of dendritic cells (DCs) is enhanced. We have found that the re-direction of progenitors also occurs during viral infection. cDC, conventional DC; ELP, early lymphoid progenitor; FLT3, FMS-related tyrosine kinase 3; IKDC, interferon-producing killer DC; LMPP, lymphoid-primed multipotent progenitor; LT‑HSC, long-term repopulating HSC; NK, natural killer; pDC, plasmacytoid DC; ST‑HSC, short-term repopulating HSC.
Other cell-lineage differentiation decisions are similarly incremental and NK‑like cells represent an example. The original Fraction A described by Hardy and colleagues includes recently described B220+Ly6C– DX5+NK1.1+CD11c+ interferon-producing killer DCs (IKDCs) that appeared to have hybrid NK‑cell and DC properties112,113. Although they closely resemble NK cells with respect to IFN production, dependence on the ID2 transcriptional repressor, IL‑15 responsiveness and cyto‑ toxic activities, IKDCs are most efficiently derived from early progenitors30,114. By contrast, most conventional NK cells appear to originate from CLPs. Although there is intense debate concerning which bone-marrow progenitors replenish the adult thymus, there is agreement that many distinctly different cell types present in the bone marrow and blood can be induced to develop into T cells under experimental conditions92. Among these, ELPs share many properties with early T‑cell-lineage progenitors (ETPs), expand at least 20,000 fold when transplanted and produce dou‑ ble positive (CD4+CD8+) thymocytes for 12 weeks16,29,92 (R.S.W., unpublished observations). Equivalents to the adult bone-marrow MPPs, ELPs and CLPs have been found in the blood, and they retain the ability to make both T‑ and B‑lineage cells92,115. However, this does not exclude the possibility that a small subset of these cells is T‑cell-lineage restricted and likely to colonize the thymus. Also, targeting of Notch or LRF disrupts nature reviews | immunology
B-cell lymphopoiesis in the bone marrow while sup‑ porting T-cell‑lineage progression93,116. Importantly, commitment to the T‑cell lineage is not complete, and early thymocytes retain the option of producing some NK‑, DC‑ and myeloid-lineage cells92,117. Reminiscent of the situation with NK cells and DCs, at least some steps in T‑cell-lineage differentiation can occur outside the thymus118. It will be important to learn whether cells made in different sites and ways are functionally specialized and stable. In summary, loss of myeloid potential is gradual and not coincident with changes in any one cell-surface marker (FIG. 2). ELPs have a low ability to clonally expand with factors that induce myeloid-cell differentiation, whereas CLPs and several related progenitors produce small numbers of macrophages. This residual potential is finally extinguished in most, but not all, CD19+ cells. Many bone-marrow subsets can generate T cells, but ELPs are the best candidates for replenishing the thymus under normal conditions.
Stem cells express Toll-like receptors Antigen-specific receptors are expressed by maturing lymphocytes and used as checkpoints to choose cells that will be most effective for the adaptive immune system, whereas cells of the innate immune system recognize bacterial and viral products through Toll-like receptors (TLRs). Haematopoietic progenitors were long thought to volume 8 | february 2008 | 103
© 2008 Nature Publishing Group
REVIEWS be incapable of discrimination between self and non-self, but we now know that they express functional TLRs119. Moreover, exposure of highly purified stem cells to TLR2 or TLR4 ligands drives them to proliferate and encour‑ ages their differentiation, whereas myeloid progenitors rapidly generate macrophages with similar stimulation (FIG. 3). CLPs are also responsive to TLR ligands, and more recent studies revealed that TLR9 can mediate their re-direction to DC lineages during viral infection (R.S.W., R.P. and P.W.K., unpublished observations). These findings show that commitment is not irreversible at the pro-lymphocyte/CLP stage and progenitors are still responsive to environmental cues.
Concluding remarks The appreciation that lymphoid progenitors are flexible begs several important questions. For example, is there a survival advantage in rapidly boosting the number of innate immune effectors during life-threatening infec‑ tions? Do DCs derived from lymphoid progenitors and those that are close relatives of B cells (such as pDCs)
1. 2.
3. 4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
Miyamoto, T. et al. Myeloid or lymphoid promiscuity as a critical step in hematopoietic lineage commitment. Dev. Cell 3, 137–147 (2002). Mansson, R. et al. Molecular evidence for hierarchical transcriptional lineage priming in fetal and adult stem cells and multipotent progenitors. Immunity 26, 407–419 (2007). This is the latest in a series of papers concerning ELPs enriched by the selection of LSK cells with the highest levels of FLT3 expression. Hu, M. et al. Multilineage gene expression precedes commitment in the hematopoietic system. Genes Dev. 11, 774–785 (1997). Park, Y.‑H. & Osmond, D. G. Dynamics of early B lymphocyte precursor cells in mouse bone marrow: proliferation of cells containing terminal deoxynucleotidyl transferase. Eur. J. Immunol. 19, 2139–2144 (1989). Medina, K. L. et al. Identification of very early lymphoid precursors in bone marrow and their regulation by estrogen. Nature Immunol. 2, 718–724 (2001). Igarashi, H., Gregory, S. C., Yokota, T., Sakaguchi, N. & Kincade, P. W. Transcription from the RAG1 locus marks the earliest lymphocyte progenitors in bone marrow. Immunity 17, 117–130 (2002). Christensen, J. L. & Weissman, I. L. Flk‑2 is a marker in hematopoietic stem cell differentiation: a simple method to isolate long-term stem cells. Proc. Natl Acad. Sci. USA 98, 14541–14546 (2001). Adolfsson, J. et al. Upregulation of Flt3 expression within the bone marrow Lin–Sca1+c-kit+ stem cell compartment is accompanied by loss of self-renewal capacity. Immunity 15, 659–669 (2001). Adolfsson, J. et al. Identification of Flt3+ lymphomyeloid stem cells lacking erythro-megakaryocytic potential a revised road map for adult blood lineage commitment. Cell 121, 295–306 (2005). Forsberg, E. C., Serwold, T., Kogan, S., Weissman, I. L. & Passegue, E. New evidence supporting megakaryocyte-erythrocyte potential of flk2/flt3+ multipotent hematopoietic progenitors. Cell 126, 415–426 (2006). Lai, A. Y., Lin, S. M. & Kondo, M. Heterogeneity of Flt3-expressing multipotent progenitors in mouse bone marrow. J. Immunol. 175, 5016–5023 (2005). Yoshida, T., Ng, S. Y., Zúñiga-Pflücker, J. C. & Georgopoulos, K. Early hematopoietic lineage restrictions directed by Ikaros. Nature Immunol. 7, 382–391 (2006). This report describes a useful new reporter mouse that helps to identify ELPs and their dependence on Ikaros. Lai, A. Y. & Kondo, M. Asymmetrical lymphoid and myeloid lineage commitment in multipotent
14.
15. 16.
17.
18.
19.
20. 21. 22. 23.
24. 25.
26.
have unique functions? Is repeated stimulation with TLR ligands during chronic infections detrimental to HSCs and could this contribute to their premature ageing? Does expression of these receptors provide an explanation for why HSCs are present outside the bone marrow? There are many examples of B‑lineage cells being converted to other fates experimentally, but do normal B cells become unstable and cause disease as is thought to be the case for T cells120? This smorgasbord of important issues represents a delightful challenge, as do the underlying molecular mechanisms. A degree of consensus is being reached about steps involved in the progression from HSCs to progenitors that are destined to become B cells, and there is a grow‑ ing appreciation that it is a gradual process. We are far from understanding how it responds to challenges, but use of fate mapping and infectious disease models is certain to provide important insights. Furthermore, it is already possible to see how manipulation and repro‑ gramming of haematopoietic cells might someday be used to repair and restore immune defences.
hematopoietic progenitors. J. Exp. Med. 203, 1867–1873 (2006). Miyake, K. et al. A VCAM-like adhesion molecule on murine bone marrow stromal cells mediates binding of lymphocyte precursors in culture. J. Cell Biol. 114, 557–565 (1991). Forsberg, E. C. et al. Differential expression of novel potential regulators in hematopoietic stem cells. PLoS Genet. 1, e28 (2005). Lai, A. Y. & Kondo, M. Identification of a bone marrow precursor of the earliest thymocytes in adult mouse. Proc. Natl Acad. Sci. USA 104, 6311–6316 (2007). References 13 and 16 demonstrate that downregulation of VCAM1 expression is an important early event in the progression from HSCs to lymphoid progenitors. Morrison, S. J. & Weissman, I. L. The long-term repopulating subset of hematopoietic stem cells is deterministic and isolatable by phenotype. Immunity 1, 661–673 (1994). Wagers, A. J. & Weissman, I. L. Differential expression of α2 integrin separates long-term and short-term reconstituting Lin-/loThy1.1loc-kit+Sca‑1+ hematopoietic stem cells. Stem Cells 24, 1087–1094 (2006). Kiel, M. J. et al. SLAM family receptors distinguish hematopoietic stem and progenitor cells and reveal endothelial niches for stem cells. Cell 121, 1109–1121 (2005). This study shows that CD150 can be used in conjunction with other markers to identify LT‑HSCs and even visualize them in situ. Matsuoka, S. et al. CD34 expression on long-term repopulating hematopoietic stem cells changes during developmental stages. Blood 97, 419–425 (2001). Sato, T., Laver, J. H. & Ogawa, M. Reversible expression of CD34 by murine hematopoietic stem cells. Blood 94, 2548–2554 (1999). Wiesmann, A. et al. Expression of CD27 on murine hematopoietic stem and progenitor cells. Immunity 12, 193–199 (2000). Orlic, D., Fischer, R., Nishikawa, S.‑I., Nienhuis, A. W. & Bodine, D. M. Purification and characterization of heterogeneous pluripotent hematopoietic stem cell populations expressing high levels of c‑kit receptor. Blood 82, 762–770 (1993). Venezia, T. A. et al. Molecular signatures of proliferation and quiescence in hematopoietic stem cells. PLoS Biol. 2, e301 (2004). Hardy, R. R., Carmack, C. E., Shinton, S. A., Kemp, J. D. & Hayakawa, K. Resolution and characterization of pro‑B and pre‑pro‑B cell stages in normal mouse bone marrow. J. Exp. Med. 173, 1213–1225 (1991). Rumfelt, L. L., Zhou, Y., Rowley, B. M., Shinton, S. A. & Hardy, R. R. Lineage specification and plasticity in CD19– early B cell precursors. J. Exp. Med. 203, 675–687 (2006).
104 | february 2008 | volume 8
27. Mojica, M. P. et al. Phenotypic distinction and functional characterization of pro‑B cells in adult mouse bone marrow. J. Immunol. 166, 3042–3051 (2001). 28. Perry, S. S. et al. L‑selectin defines a bone marrow analogue to the thymic early T‑cell-lineage progenitor. Blood 103, 2990–2996 (2004). 29. Perry, S. S., Welner, R. S., Kouro, T., Kincade, P. W. & Sun, X.‑H. Primitive lymphoid progenitors in bone marrow with T lineage reconstituting potential. J. Immunol. 177, 2880–2887 (2006). 30. Welner, R. S. et al. Interferon-producing killer dendritic cells arise via a unique differentiation pathway from primitive c‑KitHiCD62L+ lymphoid progenitors. Blood 109, 4825–4831 (2007). 31. Pelayo, R. et al. Derivation of two categories of plasmacytoid dendritic cells in murine bone marrow. Blood 105, 4407–4415 (2005). 32. Wang, J. H. et al. Selective defects in the development of the fetal and adult lymphoid system in mice with an Ikaros null mutation. Immunity 5, 537–549 (1996). 33. Nichogiannopoulou, A., Trevisan, M., Neben, S., Friedrich, C. & Georgopoulos, K. Defects in hematopoietic stem cell activity in Ikaros mutant mice. J. Exp. Med. 190, 1201–1214 (2000). 34. Thompson, E. C. et al. Ikaros DNA-binding proteins as integral components of B cell developmental‑stage‑ specific regulatory circuits. Immunity 26, 335–344 (2007). 35. Medina, K. L. et al. Assembling a gene regulatory network for specification of the B cell fate. Dev. Cell 7, 607–617 (2004). This impressive study demonstrates how transcription factors can work sequentially and cooperatively to initiate lymphopoiesis. 36. Ye, M., Ermakova, O. & Graf, T. PU.1 is not strictly required for B cell development and its absence induces a B‑2 to B‑1 cell switch. J. Exp. Med. 202, 1411–1422 (2005). 37. DeKoter, R. P., Kamath, M. B. & Houston, I. B. Analysis of concentration-dependent functions of PU.1 in hematopoiesis using mouse models. Blood Cells Mol. Dis. 39, 316–320 (2007). 38. Ikawa, T., Kawamoto, H., Wright, L. Y. & Murre, C. Long-term cultured E2A-deficient hematopoietic progenitor cells are pluripotent. Immunity 20, 349–360 (2004). 39. Nutt, S. L. & Kee, B. L. The transcriptional regulation of B cell lineage commitment. Immunity 26, 715–725 (2007). 40. Sun, X. H. Multitasking of helix‑loop‑helix proteins in lymphopoiesis. Adv. Immunol. 84, 43–77 (2004). 41. Bain, G. et al. Both E12 and E47 allow commitment to the B cell lineage. Immunity 6, 145–154 (1997). 42. Borghesi, L. et al. E47 is required for V(D)J recombinase activity in common lymphoid progenitors. J. Exp. Med. 202, 1669–1677 (2005).
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS 43. Hagman, J. & Lukin, K. Transcription factors drive B cell development. Curr. Opin. Immunol. 18, 127–134 (2006). 44. Yokota, T. et al. Unique properties of fetal lymphoid progenitors identified according to RAG1 gene expression. Immunity 19, 365–375 (2003). 45. Zhang, Z., Cotta, C. V., Stephan, R. P., deGuzman, C. G. & Klug, C. A. Enforced expression of EBF in hematopoietic stem cells restricts lymphopoiesis to the B cell lineage. EMBO J. 22, 4759–4769 (2003). 46. Lin, H. & Grosschedl, R. Failure of B‑cell differentiation in mice lacking the transcription factor EBF. Nature 376, 263–267 (1995). 47. Cobaleda, C., Jochum, W. & Busslinger, M. Conversion of mature B cells into T cells by dedifferentiation to uncommitted progenitors. Nature 449, 473–477 (2007). This paper describes how even mature B cells have a remarkable ability to lose differentiated properties and regain potential for generating multiple lineages when deprived of PAX5. 48. Fuxa, M. & Busslinger, M. Reporter gene insertions reveal a strictly B lymphoid-specific expression pattern of Pax5 in support of its B cell identity function. J. Immunol. 178, 8222–8228 (2007). 49. Liu, P. et al. Bcl11a is essential for normal lymphoid development. Nature Immunol. 4, 525–532 (2003). 50. Sanyal, M. et al. B cell development fails in the absence of the Pbx1 proto-oncogene. Blood 109, 4191–4199 (2007). 51. Hu, H. et al. Foxp1 is an essential transcriptional regulator of B cell development. Nature Immunol. 7, 819–826 (2006). 52. DeKoter, R. P., Lee, H. J. & Singh, H. PU.1 regulates expression of the interleukin‑7 receptor in lymphoid progenitors. Immunity 16, 297–309 (2002). 53. Kee, B. L. & Murre, C. Induction of early B cell factor (EBF) and multiple B lineage genes by the basic helix‑loop‑helix transcription factor E12. J. Exp. Med. 188, 699–713 (1998). 54. O’Riordan, M. & Grosschedl, R. Coordinate regulation of B cell differentiation by the transcription factors EBF and E2A. Immunity 11, 21–31 (1999). 55. Kikuchi, K., Lai, A. Y., Hsu, C. L. & Kondo, M. IL‑7 receptor signalling is necessary for stage transition in adult B cell development through upregulation of EBF. J. Exp. Med. 201, 1197–1203 (2005). 56. Roessler, S. et al. Distinct promoters mediate the regulation of Ebf1 gene expression by interleukin‑7 and Pax5. Mol. Cell Biol. 27, 579–594 (2007). This paper provides an interesting example of how one of two Ebf1 promoters supports feed-forward regulation of PAX5 expression and lineage progression. 57. Ye, M. & Graf, T. Early decisions in lymphoid development. Curr. Opin. Immunol. 19, 123–128 (2007). 58. Hirose, J. et al. A developing picture of lymphopoiesis in bone marrow. Immunol. Rev. 189, 28–40 (2002). 59. Tokoyoda, K., Egawa, T., Sugiyama, T., Choi, B. I. & Nagasawa, T. Cellular niches controlling B lymphocyte behaviour within bone marrow during development. Immunity 20, 707–718 (2004). 60. Torlakovic, E., Tenstad, E., Funderud, S. & Rian, E. CD10+ stromal cells form B‑lymphocyte maturation niches in the human bone marrow. J. Pathol. 205, 311–317 (2005). 61. Yao, L., Yokota, T., Xia, L., Kincade, P. W. & McEver, R. P. Bone marrow dysfunction in mice lacking the cytokine receptor gp130 in endothelial cells. Blood 106, 4093–4101 (2005). 62. Walkley, C. R., Shea, J. M., Sims, N. A., Purton, L. E. & Orkin, S. H. Rb regulates interactions between hematopoietic stem cells and their bone marrow microenvironment. Cell 129, 1081–1095 (2007). 63. Taniguchi, H., Toyoshima, T., Fukao, K. & Nakauchi, H. Presence of hematopoietic stem cells in the adult liver. Nature Med. 2, 198–203 (1996). 64. Bhattacharya, D., Rossi, D. J., Bryder, D. & Weissman, I. L. Purified hematopoietic stem cell engraftment of rare niches corrects severe lymphoid deficiencies without host conditioning. J. Exp. Med. 203, 73–85 (2006). 65. Wilson, A. & Trumpp, A. Bone-marrow haematopoietic‑stem‑cell niches. Nature Rev. Immunol. 6, 93–106 (2006). 66. Suda, T., Arai, F. & Hirao, A. Hematopoietic stem cells and their niche. Trends Immunol. 26, 426–433 (2005). 67. Varnum-Finney, B., Brashem-Stein, C. & Bernstein, I. D. Combined effects of Notch signalling and cytokines
68. 69. 70. 71.
72.
73.
74.
75. 76.
77.
78. 79. 80. 81. 82.
83.
84. 85.
86.
87.
88.
89.
90. 91.
induce a multiple log increase in precursors with lymphoid and myeloid reconstituting ability. Blood 101, 1784–1789 (2003). Zhang, C. C. & Lodish, H. F. Murine hematopoietic stem cells change their surface phenotype during ex vivo expansion. Blood 105, 4314–4320 (2005). Reya, T. et al. A role for Wnt signalling in self-renewal of haematopoietic stem cells. Nature 423, 409–414 (2003). Sauvageau, G., Iscove, N. N. & Humphries, R. K. In vitro and in vivo expansion of hematopoietic stem cells. Oncogene 23, 7223–7232 (2004). Waskow, C., Paul, S., Haller, C., Gassmann, M. & Rodewald, H. R. Viable c‑KitW/W mutants reveal pivotal role for c‑kit in the maintenance of lymphopoiesis. Immunity 17, 277–288 (2002). Mackarehtschian, K. et al. Targeted disruption of the flk2/flt3 gene leads to deficiencies in primitive hematopoietic progenitors. Immunity 3, 147–161 (1995). McKenna, H. J. et al. Mice lacking flt3 ligand have deficient hematopoiesis affecting hematopoietic progenitor cells, dendritic cells, and natural killer cells. Blood 95, 3489–3497 (2000). Sitnicka, E. et al. Key role of flt3 ligand in regulation of the common lymphoid progenitor but not in maintenance of the hematopoietic stem cell pool. Immunity 17, 463–472 (2002). Maeda, K. et al. IL‑6 blocks a discrete early step in lymphopoiesis. Blood 106, 879–885 (2005). Passegue, E., Wagers, A. J., Giuriato, S., Anderson, W. C. & Weissman, I. L. Global analysis of proliferation and cell cycle gene expression in the regulation of hematopoietic stem and progenitor cell fates. J. Exp. Med. 202, 1599–1611 (2005). Cheshier, S. H., Morrison, S. J., Liao, X. & Weissman, I. L. In vivo proliferation and cell cycle kinetics of long-term self-renewing hematopoietic stem cells. Proc. Natl Acad. Sci. USA 96, 3120–3125 (1999). Pelayo, R. et al. Cell cycle quiescence of early lymphoid progenitors in adult bone marrow. Stem Cells 24, 2703–2713 (2006). Arai, F. et al. Tie2/angiopoietin‑1 signalling regulates hematopoietic stem cell quiescence in the bone marrow niche. Cell 118, 149–161 (2004). Zhang, C. C. et al. Angiopoietin-like proteins stimulate ex vivo expansion of hematopoietic stem cells. Nature Med. 12, 240–245 (2006). Stier, S. et al. Osteopontin is a hematopoietic stem cell niche component that negatively regulates stem cell pool size. J. Exp. Med. 201, 1781–1791 (2005). Nakamura, K. et al. Src homology 2‑containing 5‑inositol phosphatase (SHIP) suppresses an early stage of lymphoid cell development through elevated interleukin‑6 production by myeloid cells in bone marrow. J. Exp. Med. 199, 243–254 (2004). Kouro, T., Medina, K. L., Oritani, K. & Kincade, P. W. Characteristics of early murine B lymphocyte precursors and their direct sensitivity to negative regulators. Blood 97, 2708–2715 (2001). North, T. E. et al. Prostaglandin E2 regulates vertebrate haematopoietic stem cell homeostasis. Nature 447, 1007–1011 (2007). Kincade, P. W. et al. Lymphoid lineage cells in adult murine bone marrow diverge from those of other blood cells at an early, hormone-sensitive stage. Semin. Immunol. 14, 385–394 (2002). Koch, U. et al. Simultaneous loss of β- and γ-catenin does not perturb hematopoiesis or lymphopoiesis. Blood 13 September 2007 (doi:10.1182/blood-200707-099754). Nemeth, M. J., Topol, L., Anderson, S. M., Yang, Y. & Bodine, D. M. Wnt5a inhibits canonical Wnt signalling in hematopoietic stem cells and enhances repopulation. Proc. Natl Acad. Sci. USA 104, 15436–15441 (2007). Jeannet, G. et al. Long-term, multilineage hematopoiesis occurs in the combined absence of β‑catenin and γ-catenin. Blood 28 September 2007 (doi:10.1182/blood-2007-07-102558). Baba, Y., Garrett, K. P. & Kincade, P. W. Constitutively active β‑catenin confers multilineage differentiation potential on lymphoid and myeloid progenitors. Immunity 23, 599–609 (2005). Baba, Y. et al. Constitutively active β-catenin promotes expansion of multipotent hematopoietic progenitors in culture. J. Immunol 177, 2294–2303 (2006). Liang, H. et al. Wnt5a inhibits B cell proliferation and functions as a tumor usuppressor in hematopoietic tissue. Cancer Cell 4, 349–360 (2003).
nature reviews | immunology
92. Bhandoola, A., Von Boehmer, H., Petrie, H. T. & Zúñiga-Pflücker, J. C. Commitment and developmental potential of extrathymic and intrathymic T cell precursors: plenty to choose from. Immunity 26, 678–689 (2007). 93. Maeda, T. et al. Regulation of B versus T lymphoid lineage fate decision by the proto-oncogene LRF. Science 316, 860–866 (2007). 94. Nie, L., Xu, M., Vladimirova, A. & Sun, X. H. Notchinduced E2A ubiquitination and degradation are controlled by MAP kinase activities. EMBO J. 22, 5780–5792 (2003). 95. Kondo, M., Weissman, I. L. & Akashi, K. Identification of clonogenic common lymphoid progenitors in mouse bone marrow. Cell 91, 661–672 (1997). 96. Balciunaite, G., Ceredig, R., Massa, S. & Rolink, A. G. A B220+CD117+CD19+/– hematopoietic progenitor with potent lymphoid and myeloid developmental potential. Eur. J. Immunol. 35, 2019–2030 (2005). 97. Allman, D. et al. Thymopoiesis independent of common lymphoid progenitors. Nature Immunol. 4, 168–174 (2003). 98. Huang, J. et al. Propensity of adult lymphoid progenitors to progress to DN2/3 stage thymocytes with Notch receptor ligation. J. Immunol. 175, 4858–4865 (2005). 99. Izon, D. et al. A common pathway for dendritic cell and early B cell development. J. Immunol. 167, 1387–1392 (2001). 100. Hardy, R. R., Kincade, P. W. & Dorshkind, K. The protean nature of cells in the B lymphocyte lineage. Immunity 26, 703–714 (2007). 101. Ceredig, R., Rauch, M., Balciunaite, G. & Rolink, A. G. Increasing Flt3L availability alters composition of a novel bone marrow lymphoid progenitor compartment. Blood 108, 1216–1222 (2006). 102. Martin, C. H. et al. Efficient thymic immigration of B220+ lymphoid-restricted bone marrow cells with T precursor potential. Nature Immunol. 4, 866–873 (2003). 103. Rosmaraki, E. E. et al. Identification of committed NK cell progenitors in adult murine bone marrow. Eur. J. Immunol. 31, 1900–1909 (2001). 104. Kouro, T., Kumar, V. & Kincade, P. W. Relationships between early B‑ and NK‑lineage lymphocyte precursors in bone marrow. Blood 100, 3672–3680 (2002). 105. Nikolic, T., Dingjan, G. M., Leenen, P. J. & Hendriks, R. W. A subfraction of B220+ cells in murine bone marrow and spleen does not belong to the B cell lineage but has dendritic cell characteristics. Eur. J. Immunol. 32, 686–692 (2002). 106. Dontje, W. et al. Delta‑like1‑induced Notch1 signalling regulates the human plasmacytoid dendritic cell versus T‑cell lineage decision through control of GATA‑3 and Spi‑B. Blood 107, 2446–2452 (2006). 107. Nutt, S. L., Heavey, B., Rolink, A. G. & Busslinger, M. Commitment to the B‑lymphoid lineage depends on the transcription factor Pax5. Nature 401, 556–562 (1999). 108. Montecino-Rodriguez, E., Leathers, H. & Dorshkind, K. Bipotential B‑macrophage progenitors are present in adult bone marrow. Nature Immunol. 2, 83–88 (2001). 109. Montecino-Rodriguez, E., Leathers, H. & Dorshkind, K. Identification of a B‑1 B cell-specified progenitor. Nature Immunol. 7, 293–301 (2006). References 108 and 109 describe the identification of progenitors restricted to the B‑1‑cell lineage in fetal and adult tissues, which represents a breakthrough in the understanding of the origin of these specialized B cells. 110. Dias, S., Silva, H. Jr, Cumano, A. & Vieira, P. Interleukin‑7 is necessary to maintain the B cell potential in common lymphoid progenitors. J. Exp. Med. 201, 971–979 (2005). 111. Chappaz, S., Flueck, L., Farr, A. G., Rolink, A. G. & Finke, D. Increased TSLP availability restores T and B cell compartments in adult IL‑7 deficient mice. Blood 110, 3862–3870 (2007). 112. Chan, C. W. et al. Interferon-producing killer dendritic cells provide a link between innate and adaptive immunity. Nature Med. 12, 207–213 (2006). 113. Taieb, J. et al. A novel dendritic cell subset involved in tumor immunosurveillance. Nature Med. 12, 214–219 (2006). 114. Spits, H. & Lanier, L. L. Natural killer or dendritic: what’s in a name? Immunity 26, 11–16 (2007). 115. Umland, O., Mwangi, W. N., Anderson, B. M., Walker, J. C. & Petrie, H. T. The blood contains multiple distinct progenitor populations with clonogenic B and T lineage potential. J. Immunol. 178, 4147–4152 (2007).
volume 8 | february 2008 | 105 © 2008 Nature Publishing Group
REVIEWS 116. Pui, J. C. et al. Notch1 expression in early lymphopoiesis influences B versus T lineage determination. Immunity 11, 299–308 (1999). 117. Heinzel, K., Benz, C., Martins, V. C., Haidl, I. D. & Bleul, C. C. Bone marrow-derived hematopoietic precursors commit to the T cell lineage only after arrival in the thymic microenvironment. J. Immunol. 178, 858–868 (2007). 118. Lambolez, F. et al. The thymus exports long-lived fully committed T cell precursors that can colonize primary lymphoid organs. Nature Immunol. 7, 76–82 (2006). 119. Nagai, Y. et al. Toll-like receptors on hematopoietic progenitor cells stimulate innate immune system replenishment. Immunity 24, 801–812 (2006).
Reference 119 describes how the identification of functional TLRs on stem and progenitor cells shows how these cells may directly contribute to host defence mechanisms. 120. Meresse, B. et al. Reprogramming of CTLs into natural killer-like cells in celiac disease. J. Exp. Med. 203, 1343–1355 (2006).
Acknowledgements
Our work is supported by grants from the National Institutes of Health, USA (AI20069 and AI058162). We are grateful for the helpful comments from our colleagues and especially K. Garrett. We apologize for not being able to cite many relevant primary papers and reviews because of space limitations.
106 | february 2008 | volume 8
DATABASES Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene EBF1 | FLT3 | Ikaros | PAX5 | RAG1 | SCF | TdT | VCAM1
FURTHER INFORMATION Paul Kincade’s homepage: http://www.omrf.ouhsc.edu/ OMRF/Research/14/KincadeP.asp
SUPPLEMENTARY INFORMATION See online article: s1 (tables)
All links are active in the online pdf
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS
Shaping and reshaping CD8+ T‑cell memory John T. Harty* and Vladimir P. Badovinac‡
Abstract | The ability to develop and sustain populations of memory T cells after infection or immunization is a hallmark of the adaptive immune response and a basis for protective vaccination against infectious disease. Technical advances that allow direct ex vivo identification and characterization of antigen-specific CD8+ T cells at various stages of the response to infection or vaccination in mouse models have fuelled efforts to characterize the factors that control memory CD8+ T‑cell generation. Here, we dissect the input signals that shape the characteristics of the memory CD8+ T‑cell response and discuss how manipulation of these signals has the potential to reshape CD8+ T‑cell memory and improve the efficacy of vaccination.
*Department of Microbiology and Interdisciplinary Graduate Program in Immunology, University of Iowa, Iowa City, Iowa, USA. ‡ Department of Pathology, University of Iowa, Iowa City, Iowa, USA. Correspondence to J.T.H. e‑mail:
[email protected] doi:10.1038/nri2251
Thymic selection shapes the naive T‑cell repertoire to express T‑cell receptors (TCRs) that exhibit strong reac‑ tivity to pathogen peptides displayed by self MHC class I molecules1. Because of the need to ensure the capacity to respond to the enormous diversity in the microbial universe, naive T cells that can recognize specific patho‑ gen epitopes are infrequent in the total T‑cell population (ranging from tens to hundreds of cells per mouse)2–6. This results in a repertoire that can respond to most pathogens; however, substantial proliferative expansion in the numbers of pathogen-specific T cells is required to combat infection. By virtue of a defined set of homing receptors (such as CD62L; also known as SELL) and chemokine recep‑ tors, such as CC‑chemokine receptor 7 (CCR7), naive CD8+ T cells circulate between the blood and second‑ ary lymphoid organs, where they survey the antigens displayed by dendritic cells (DCs). As a consequence of infection or immunization, DCs acquire foreign antigens and differentiate into mature antigen-presenting cells (APCs) that express co-stimulatory molecules, produce cytokines and migrate to secondary lymphoid organs to initiate responses from the rare, pathogen-specific CD8+ T cells7,8. Imaging studies suggest that initial transient interactions between CD8+ T cells and DCs eventually result in stable conjugate formation, lasting for hours9. Recent data suggest that signals received during this interaction — including antigen (signal 1), co-stimulation (signal 2), and pro-inflammatory cytokines (signal 3) — programme the responding CD8+ T cells in a fashion that will eventually lead to the formation of memory T‑cell populations10–12. In response to these activating signals,
nature reviews | immunology
the pathogen-specific CD8+ T cells embark on a prolif‑ erative expansion in numbers, with one precursor giving rise to more than 10,000 daughter cells (>13 divisions) over the course of 5–8 days (FIG. 1)11,13. This proliferation is accompanied by differentiation into a population of ‘effector’ CD8+ T cells that leave the secondary lymphoid organs to survey the body for signs of infection. When effector T cells encounter infected cells they manifest antimicrobial functions such as cytolysis and the ability to rapidly produce interferon‑γ (IFNγ), with some cells also producing tumour-necrosis factor (TNF). Importantly, whether the T‑cell response is success‑ ful in eliminating the pathogen or not, the numbers of pathogen-specific CD8+ T cells undergo a precipitous decline in all organs during the ‘cell death’ or ‘contraction’ phase11,13. In general, contraction eliminates 90–95% of pathogen-specific effector CD8+ T cells, presumably to maintain flexibility to respond against new pathogens and to limit immunopathology, which could occur with the presence in the host of large numbers of activated T cells that fail to clear an infection. Although contrac‑ tion is substantial, it is incomplete and some pathogenspecific T cells survive in all tissues14,15. Therefore, the number of pathogen-specific CD8+ T cells that survive the contraction phase exceeds the number of naive pre‑ cursors specific for that pathogen antigen (FIG. 1). Interestingly, the number of memory CD8+ T cells present at the end of the contraction phase can, at least in laboratory mice, be maintained for essentially the life of the host16. Although the numbers of memory CD8+ T cells are stable, maintenance of memory CD8+ T‑cell numbers after acute infection or vaccination is a dynamic volume 8 | february 2008 | 107
© 2008 Nature Publishing Group
REVIEWS process that involves cytokine-driven proliferation that is balanced by cell death (a process known as memory turnover)17–20. In addition, phenotypic and functional properties of memory T‑cell populations change with time and also may be regulated by the tissue of residence13,21–24. Whereas early studies suggested that the expression of homing molecules could be used to differ‑ entiate functional T‑cell subsets (that is, CD62LlowCCR7low effector memory T cells versus CD62LhiCCR7hi central memory T cells)15,25–27, more recent studies suggest that cell-surface phenotype often fails to ascribe functional attributes to these memory CD8+ T‑cell phenotypes13. In this Review, we will discuss recent data addressing how input signals shape antigen-specific CD8+ T‑cell memory generation and how these signals may be manipulated to improve the efficacy of vaccines.
Endogenous versus transgenic T cells The advent of MHC class I tetramers and peptide-stimulated intracellular cytokine staining (ICS), which allow the direct ex vivo identification of antigen-specific CD8+ T cells, in conjunction with the application of TCR-transgenic adoptive transfer approaches, has stimulated an explo‑ sion of research on the CD8+ T‑cell response to infection or vaccination28–33. Although each approach has limita‑ tions (tetramers, for example, address only phenotype and lack sensitivity when evaluating rare events because of background staining, and ICS for one or two cytokines might not identify all relevant T cells), tetramers and ICS allow the analysis of the endogenous T‑cell response to a
Antigen non-specific T cell DC
MHC class I tetramers A soluble tetrahedral complex artificially generated by using a fluorochrome-coupled avidin to join four biotinylated MHC class I molecules with a peptide of interest and β2-microglobulin. The resulting MHC class I tetramer can be used as a reagent to identify antigen-specific CD8+ T-cell populations.
Intracellular cytokine staining (ICS). A tool used for the analysis of the ability of T cells to produce a cytokine in response to a specific stimulus, such as peptide stimulation. In this assay, the usual cytokine secretion pathway is paralyzed and intracellular accumulation of the cytokine is monitored by antibody staining and FACS analysis.
b
Activation
Antigen-specific naive CD8+ T cell
Expansion
c
infection or immunization, which initiates from low numbers of naive precursor T cells. One major strength of the TCR-transgenic adop‑ tive transfer approach is the ability to analyse specific gene products in T‑cell biology, by directly comparing wild-type and gene-deficient TCR-transgenic T cells responding to infection or immunization after stimula‑ tion in the same animal. Based on initial conditions used with the TCR-transgenic adoptive transfer approach32, most published studies transferred 106 or more TCRtransgenic T cells into naive mice. These large numbers of transferred cells were considered a strength of the TCR-transgenic adoptive transfer models, because this approach allows for the identification of the transferred cells before and early after infection or vaccination. However, this experimental design creates a situation in which the antigen-specific T cells are present at 1,000– 100,000-fold higher numbers than the corresponding endogenous naive precursor T‑cell frequency. Over the past several years, studies compared specific aspects of the T‑cell response of mice containing high or titrating numbers of TCR-transgenic T cells. Collectively, these studies suggested that several features of the T‑cell response were altered at high input TCR-transgenic T‑cell numbers34–37. In agreement with these studies, we recently showed that essentially all facets, such as pro‑ liferative expansion, phenotype and function (FIG. 2), of the CD8+ T‑cell response during the generation of CD8+ T-cell memory were altered when the numbers of TCRtransgenic T cells transferred were sufficiently high to
Contraction
d
Early memory
e
Late memory
Apoptotic cell CD62L
CD127 Memory turnover
IFNγ-producing CD8+ T cell IFNγ- and TNF-producing CD8+ T cell IFNγ-, TNF- and IL-2-producing CD8+ T cell
Figure 1 | CD8+ T‑cell response to infection and/or vaccination. a | The initial activation of rare, antigen-specfic CD8+ Nature Reviews | Immunology T cells occurs shortly after infection or vaccination, as a result of their encounter with mature dendritic cells (DCs) that express antigenic peptides and co-stimulatory molecules. b | Activation is followed by the proliferative expansion in numbers and differentiation of the antigen-specific CD8+ T cells into primary effector populations. These exhibit both functional and phenotypic differences from their naive precursors. For example, most effector CD8+ T cells downregulate the expression of CD127 (the IL‑7 receptor α‑chain) and CD62L, which are molecules that are uniformly expressed by naive CD8+ T cells, and acquire the ability to produce effector cytokines such as interferon‑γ (IFNγ) and to lesser extent tumour-necrosis factor (TNF), as well as the ability to perform cytolysis. c | Most (90–95%) of the primary effector CD8+ T cells generated during the expansion phase die during the ensuing programmed contraction (cell death) phase. d | The surviving CD8+ T cells initiate the early memory pool and slowly undergo additional phenotypic changes, such as the upregulation of CD127 and CD62L expression, as well as functional changes such as TNF production by all IFNγ-producing cells with some cells also able to produce interleukin‑2 (IL‑2) following antigen stimulation. These changes are consistent with late memory CD8+ T‑cell populations. e | Although numerically stable, CD8+ T‑cell memory maintenance is a dynamic process that includes balanced proliferation (memory turnover) and cell death that is dependent on homeostatic cytokines (for example, IL‑7 and IL‑15).
108 | february 2008 | volume 8
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS prevent the endogenous T‑cell response to the same epitope38. Surprisingly, for at least one popular TCRtransgenic model — the OT‑1 mouse model, in which the T cells are specific for an ovalbumin-derived a
Expansion
Activation
Contraction
High number of TCR-transgenic CD8+ T cells Effector TCR-transgenic Antigen Apoptotic CD8+ T cell non-specific T cell cell
Memory Memory TCR-transgenic CD8+ T cell
DC
Naive TCR-transgenic CD8+ T cell Low number of TCR-transgenic CD8+ T cells
Naive endogenous antigen-specific CD8+ T cell Effector endogenous CD8+ T cell
Memory endogenous CD8+ T cell
b
% CD62L
Number of antigenspecific CD8+ T cells
100
LOD
0
Time
Time
100 % IL-2
% CD127
100
50
0
50
Time
50
0
Time
High number of TCR-transgenic CD8+ T cells Low number of TCR-transgenic CD8+ T cells or endogenous CD8+ T cells
Figure 2 | Influence of TCR-transgenic precursor frequency on important attributes of the CD8+ T‑cell response to infection. a | Transfer high numbers of Nature of Reviews | Immunology T‑cell receptor (TCR)-transgenic CD8+ T cells results in competition for antigenpresenting cells (APCs), and perhaps other resources between TCR-transgenic T cells, and inhibits the priming of the endogenous repertoire that is specific for the same epitope. Competition is abrogated when physiological numbers (low numbers) of TCRtransgenic T cells are transferred; an experimental setting that also permits a robust endogenous response. b | High number of TCR-transgenic CD8+ T cells present at the time of infection will proliferate less, reach peak numbers earlier and show decreased efficiency of memory CD8+ T‑cell generation compared with endogenous and/or TCRtransgenic CD8+ T cells seeded at low numbers. In addition, expression of surface molecules (accelerated upregulation of CD62L and CD127 (the IL‑7 receptor α‑chain)) and effector functions (increased antigen-stimulated interleukin‑2 (IL‑2) production and decreased expression of granzyme B (not shown) at the peak of the conventional expansion phase) are substantially altered when the initial TCR-transgenic CD8+ T‑cell precursor frequency is high enough to inhibit the endogenous CD8+ T‑cell response to the same epitope. DC, dendritic cell; LOD, limit of detection. nature reviews | immunology
peptide 39 — transfer of fewer than 500 T cells was required to permit a concurrent robust endogenous CD8+ T‑cell response and allow the TCR-transgenic T cells to mimic the endogenous CD8+ T‑cell response to infection with Listeria monocytogenes38. Importantly, the specific number of TCR-transgenic T cells that can be used to mimic the endogenous response may be influenced by several factors — including the affinity of the TCR for antigen, the number of endogenous pre‑ cursor cells and the pathogen or vaccination method used — that must be empirically determined for each model38. Together, these data might have an impact on the interpretation of a large fraction of the existing published studies and also of recent imaging studies of early T‑cell–DC interactions, which rely on the transfer of large numbers of TCR-transgenic T cells40. The good news is, that due to the robust proliferative response of CD8+ T cells responding to infection or immuniza‑ tion, many studies with even very low numbers of input TCR-transgenic T cells (~3–6 T cells) are feasible38. This experimental design allows for the powerful com‑ parisons between wild-type and gene-deficient TCRtransgenic T cells responding in the same host, but creates a context that approximates normal physiology and also allows direct comparison with the endogenous T‑cell response of the host. Although the data cited above show that non-physi‑ ological input numbers of TCR-transgenic T cells result in immune responses that fail to mimic the endogenous response, it would be remiss not to point out the poten‑ tial value of experiments that compare T‑cell responses (that is proliferation and differentiation) from titrated input numbers of TCR-transgenic T cells. If controlled for other variables such as inflammation, studies with titrated numbers of input TCR-transgenic T cells may reveal important characteristics of the T‑cell response that are regulated by the strength of signals (specifi‑ cally signal 1 and signal 2, which depend on cell–cell contact) between the responding T cells and the APC. Experiments can then be designed to alter these param‑ eters (through alteration of antigen load or regulation of specific co-stimulatory molecule expression) to determine the impact of these factors on T‑cell responses from physio logical precursor numbers. However, given the dramatic impact that input numbers of TCR-transgenic T cells have on the characteristics of the immune response, it should not be surprising if some cherished immunological paradigms are challenged by data generated from experi‑ ments carried out at physiological TCR-transgenic T‑cell input numbers.
CD4+ T‑cell help A brief interaction with antigen in vitro or in vivo ‘pro‑ grammes’ CD8+ T cells with the capacity to differentiate into memory T cells13,41–44. However, it has become clear that a number of other signals in addition to antigen determine whether effective CD8+ T‑cell memory is generated and maintained in vivo. One long-standing observation is the requirement for CD4+ T‑cell ‘help’ for CD8+ T‑cell memory45. Early data using non-infectious immunization regimens suggested that CD4+ T‑cell volume 8 | february 2008 | 109
© 2008 Nature Publishing Group
REVIEWS
OT‑I mouse model Mice inwhich the CD8 T cells have a transgenic rearranged T-cell receptor (Vα2, Vβ5) that recognizes a peptide (OVA257–264) derived from chicken albumin in the context of H2-Kb. It can be used to study the response of CD8+ T cells to antigen.
T-bet A member of the T‑box family of transcription factors. It is a master switch in the development of T helper 1 (TH1)-cell responses, through its ability to regulate expression of the interleukin‑12 receptor, inhibit signals that promote TH2-cell development and promote the production of interferon‑γ.
help consisted of CD40 ligand (CD40L; also known as CD154)-mediated signals that ‘conditioned’ DCs to become effective stimulators of naive CD8+ T cells45–48. These results were extended by the observation that, in the absence of CD4+ T cells during priming, CD8+ T cells exhibited heightened expression of the death receptor TRAIL (tumour-necrosis factor (TNF)-related apoptosis-inducing ligand), and underwent apoptosis when restimulated49. Importantly, TRAIL-deficient CD8+ T cells exhibited normal secondary expansion when gen‑ erated in the absence of CD4+ T cells49. This observation was consistent for both non-infectious and lymphocytic choriomeningitis virus (LCMV) infection models, but with the caveat that the secondary expansion potential in both situations was measured in vitro49. Together, these and other studies suggested that CD4+ T‑cell help func‑ tioned early and altered the programme of the CD8+ T‑cell response, in part by regulating TRAIL expression49–54. By contrast, adoptive transfer studies of TCRtransgenic T cells in the LCMV infection model showed that, although CD4+ T cells were not essential for a robust primary expansion of the CD8+ T‑cell response, they were essential for maintaining memory CD8+ T‑cell numbers and functional attributes, such as memory turn‑ over, CD127 (the interleukin‑7 (IL‑7) receptor α‑chain) expression and the capacity for vigorous proliferation in response to rechallenge54,55. These data argued that CD4+ T‑cell help occurs during the maintenance of CD8+ T‑cell memory populations. Recently, studies of LCMV infec‑ tion models show that, although a deficiency in TRAIL expression rescues memory CD8+ T‑cell function in the absence of CD4+ T cells for up to 60 days post infection, eventually the memory CD8+ T cells lose their functional integrity56. This suggests that CD4+ T cells may help (at least in some situations) to set the initial programme of the responding CD8+ T cells and also help to maintain the functionality of the memory populations. Although it seems that the role of CD4+ T cells may differ follow‑ ing infection or immunization with non-replicating antigen45, there is sufficient disparity in results between infection models57,58 to suggest that the biology of the pathogen may also influence the nature or requirement for CD4+ T‑cell help. These and many other studies provide strong evi‑ dence that CD4+ T cells provide essential services for the optimal generation of CD8+ T‑cell memory; however, the precise nature of these signals remains elusive59. One early notion was that CD4+ T‑cell-derived IL‑2 might be an important factor for the optimal generation of CD8+ T‑cell memory. However, experiments with IL‑2deficient mice showed that this cytokine is not essential for apparently normal expansion and memory CD8+ T‑cell generation60–62. In an interesting twist, recent data demonstrate normal expansion, contraction and memory generation, but aberrant secondary responses by CD8+ T cells that lack expression of the high-affinity chain of the IL‑2 receptor (CD25)61. The functionality of CD25-deficient memory CD8+ T cells could be restored by injection of a complex of IL‑2 and an IL‑2-specific antibody that allows signalling through the low affinity IL‑2 receptor (CD122)61,63. Importantly, this treatment
110 | february 2008 | volume 8
only restored memory CD8+ T‑cell function if admin‑ istered at the time of immunization but not at the time of rechallenge61. Therefore, IL‑2 seems to be an essential input signal of the CD8+ T‑cell response programme that leads to the effective generation of CD8+ T‑cell memory. At this time, the cell types that produce the relevant IL‑2 for CD8+ T‑cell memory have not been identified and, specifically, it is not known if IL‑2 can replace CD4 + T‑cell help for memory CD8+ T‑cell generation. Finally, specific reasons why ‘unhelped’ CD8+ T cells lack functionality remain to be defined. In this regard, recent studies demonstrate that most unhelped CD8+ T cells exhibit relatively high T‑bet expression64; a pat‑ tern which is also exhibited by so-called short-lived effector CD8+ T cells64–66. Importantly, deletion of T‑bet in T cells restored their ability to develop into central memory CD8+ T cells (CD62LhiCCR7hi) and function even in the absence of CD4+ T cells64. On the basis of these data, future studies to examine the relationship between T‑bet and TRAIL expression in unhelped CD8+ T cells may be informative. In summary, although it is clear that CD4+ T cells have a crucial role in generating optimal CD8+ T‑cell memory, the precise signals and pathways that constitute this help remain unclear and current studies have only scratched the surface with regard to how CD4+ T‑cell help regulates the molecules that control the behaviour of memory CD8+ T cells. Therefore, this area of research should remain fruitful in the foreseeable future.
Memory CD8+ T‑cell maintenance One early controversy on the role of antigens in T‑cell memory has been settled by a large body of evidence showing that antigen persistence is not required for the maintenance of memory CD8+ T cells67 (FIG. 3a). Under circumstances where the pathogen and antigen are cleared (acute infection), memory CD8+ T cells are maintained by signals through receptors that contain the common cytokine receptor γ‑chain (γc; such as receptors for IL‑7 and IL‑15). IL‑7 and IL‑15 appear to deliver sur‑ vival and proliferative signals to memory CD8+ T cells, respectively17–20. As the numbers of antigen-specific CD8+ memory T cells do not change appreciably with time, this relatively slow cytokine-driven ‘basal’ proliferation must be accompanied by an equivalent rate of cell death (memory turnover). Understanding how this process is regulated is a fertile area of unexplored immunology, with potential relevance to vaccination. By contrast, persistent infections often cause sub‑ stantial alterations in the phenotype and functionality of memory CD8+ T cells68,69 (FIG. 3b). This may result in the deletion of CD8+ T cells of certain specificities or the exhaustion of their effector functions, without deletion70. Recent progress in this area demonstrates that the expression of programmed cell death 1 (PD1) by memory CD8 + T cells in persistently infected mice maintains their decreased functionality. Indeed, blocking PD1 enhances the clearance of persistent LCMV infection and improves the functionality of the memory CD8+ T cells71. In addition, other recent studies suggest that IL‑10 production during chronic www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS infection limits memory CD8+ T‑cell antiviral activity72,73. Importantly, memory CD8+ T cells derived from per‑ sistently infected mice fail to survive when transferred to an antigen-free environment 74. Maintenance of these T cells requires the interaction with viral pep‑ tide and cannot be accomplished by the presence of IL‑7 and IL‑15 or other inflammatory cytokines74–77. Consistent with this requirement for antigen, memory CD8+ T‑cell maintenance during persistent infection is associated with extensive and rapid proliferation of these cells, rather than the slow memory turnover observed after acute infection74,76. It is reasonable to assume that this extensive proliferation will result in the eventual deletion of antigen-specific CD8+ T cells, as other studies suggest that these cells can only under‑ take a finite number of divisions78. Indeed, this is true for CD8+ T cells specific for some, but not all epitopes after persistent infection with LCMV70,79. In the con‑ text of recent studies, it will be of interest to determine whether blocking PD1 also decreases the cycling of memory CD8 + T cells during chronic infection. Possible explanations for the failure to delete all virusspecific CD8+ T cells may include the relative amounts of specific epitopes produced or, perhaps, the rate at a Acute infection
Memory maintenance
Number of antigenspecific CD8+ T cells
CD62L
LOD Time
b Persistent infection
CD127
Antigen non-specific T cell
Number of antigenspecific CD8+ T cells
PD1
LOD Time
CD8+ T-cell response to antigen 1 CD8+ T-cell response to antigen 2 Infection
Memory turnover
Antigen
Antigen-driven memory turnover IFNγ-producing CD8+ T cell IFNγ- and TNF-producing CD8+ T cell IFNγ-, TNF- and IL-2-producing CD8+ T cell
Figure 3 | Memory CD8+ T‑cell maintenance after acute or chronic infection. Nature Reviews | Immunology a | After the clearance of pathogen and antigens following acute infection, memory CD8+ T cells are stably maintained by signals through the interleukin‑7 (IL-7) receptor and IL‑15 receptor delivered by IL‑7 (survival) and IL‑15 (proliferation). The phenotype of the CD8+ memory T‑cell populations evolves towards a canonical memory profile and the functionality of the cells continues to improve with time. b | Persistent infections lead to substantial changes in the phenotype and function of memory CD8+ T‑cell responses. Continued exposure to antigenic peptides and not homeostatic cytokines force CD8+ T cells to undergo extensive proliferation that leads to declining numbers and/or deletion of antigen-specific CD8+ T cells. In addition, memory CD8+ T cells in chronically infected hosts gradually loose their ability to produce cytokines (IL‑2, then tumournecrosis factor (TNF), then interferon‑γ (IFNγ)) and never downregulate their expression of programmed cell death 1 (PD1). The rate of functional impairment and deletion of memory CD8+ T‑cell populations seem to be epitope specific during chronic infections. LOD, limit of detection. nature reviews | immunology
which new antigen-specific CD8+ T cells are produced during chronic infection. This notion is supported by studies showing that the activation of newly produced CD8+ T cells contributes to the persistence of memory and the heterogeneity of memory phenotype during persistent polyomavirus infection80. In summary, the requirements for memory CD8+ T‑cell maintenance and the functional characteristics of the memory populations are dramatically different after persistent compared with acute infection. One goal for the future will be to understand whether it is the host or the pathogen that benefits from the exhaustion of antigenspecific CD8+ T cells during chronic infection. In this regard, it may be instructive that mice lacking CD8+ T cells fail to clear LCMV infection and undergo a CD4+ T‑celldependent wasting disease after intracranial infection81. These data suggest that the immune alterations resulting from the interplay between the persistent virus and the CD8+ T cell may ultimately benefit both parties.
Inflammation and memory T-cell differentiation Early experiments revealed that not only peptide–MHC complexes (signal 1), but also co-stimulatory signals (signal 2) were required to activate naive T cells82,83. Consistent with this, adjuvants that increased T‑cell responses in vivo also stimulated co-stimulatory mol‑ ecule expression by APCs7,8. The connection between adjuvants and innate immune activation through Toll-like receptors (TLRs) and other innate patternrecognition receptors (PRRs) provided a mechanism for the regulation of co-stimulatory molecule expres‑ sion by APCs7,8. However, these studies revealed that activation of innate immune cells through PRRs also induced the production of pro-inflammatory cytokines, such as type I IFNs (produced by plasmacytoid DCs), IL‑12 (produced by DCs, macrophages and other cells) and IFNγ (produced by natural killer (NK) cells, NKT cells and previously activated CD8 + T cells)10. Similarly, infections evoke various combinations of these cytokines, the composition of which probably depends on the specific PRRs that are activated by each pathogen10. At least one role of these pro-inflammatory cytokines appears to be in the delivery of ‘maturation’ signals to DCs. However, recent studies demonstrate that pro-inflammatory cytokines act directly on responding CD8+ T cells to affect crucial aspects of memory generation (signal 3). Expansion. Early in vitro studies showed that the addition of IL‑12 or type I IFNs to T‑cell cultures enhanced the proliferation and survival of the activated T cells. These studies led to the concept that signal 3 was important for optimal CD8+ T‑cell responses12,84–86. Over the last several years, the notion that pro-inflammatory cytokines act directly on responding CD8+ T cells has been elegantly vetted by experiments that show a reduced expansion of T cells deficient in various cytokine receptors (for example, for type I IFNs, IL‑12 or IFNγ) compared with wild-type TCR-transgenic T cells responding to infection in the same wild-type recipients87–91 (FIG. 4a). Importantly, the absence of these pro-inflammatory volume 8 | february 2008 | 111
© 2008 Nature Publishing Group
REVIEWS cytokine receptors does not compromise proliferation of these gene-deficient TCR-transgenic T cells, but instead it decreases their survival rates, which results in progres‑ sively fewer gene-deficient CD8+ T cells compared with wild-type CD8+ T cells during the expansion phase87–91. Although the molecular details remain to be fully understood, it is thought that signals mediated by these pro-inflammatory cytokines may increase the expression of pro-survival molecules such as B‑cell lymphoma 3 (BCL-3) (Refs 12,92) or potentially alter the balance of pro-apoptotic and anti-apoptotic BCL-2-family members to promote T‑cell survival during proliferation.
Memory CD8+ T cell
Interestingly, however, the nature of the relevant signal 3 appears to depend on the pathogen that evokes the response10,93,94. For example, deletion of the gene encoding the type I IFN receptor (IFNAR) has a major deleterious effect on the expansion of TCR-transgenic T cells after LCMV infection, whereas the same TCRtransgenic T cells that lack the IL‑12 receptor or IFNγ receptor (IFNGR) exhibit more moderate defects in this model. In addition, the lack of IFNAR expression has a less dramatic impact on CD8+ T‑cell expansion after infection with L. monocytogenes or vaccinia virus when compared with the major decrease in CD8+ T‑cell expansion of these cells after LCMV infection87,89,94,95. Although each of these pathogens evokes the expres‑ sion of type I IFNs, the magnitude and kinetics of the response are dependent on the pathogenic organism89,96. In addition, L. monocytogenes and vaccinia virus induce robust early IL‑12 and IFNγ production, whereas these cytokines are induced with delayed kinetics after LCMV infection10. Therefore, the specific PRRs that are activated by each pathogen may dictate which pro-inflammatory cytokines are selected to serve as the signal 3 for CD8+ T‑cell survival during the expansion phase. This notion may serve to explain the seemingly contradictory finding that mice in which all cells lack the expression of IFNAR exhibit relatively normal expansion in the primary CD8+ T‑cell response to LCMV infection96,97. In this scenario, the absence of IFNAR could potentially shift the balance of cytokine production after LCMV infection to favour the use of a substitute signal 3 (possibly IL‑12 or IFNγ). Alternatively, the failure to clear the infection in IFNARdeficient mice alters the duration of antigen-persistence, and potentially the proliferative response of CD8+ T cells.
Figure 4 | Inflammatory cytokines and CD8+ T‑cell homeostasis after infection. Nature Reviews | Immunology a | Antigenic peptides presented by mature (co-stimulatory molecule and cytokine expressing) dendritic cells (DCs) to naive CD8+ T cells trigger their proliferative expansion and differentiation into effector CD8+ T cells. During the early stages of the response multiple cytokines (such as interleukin‑12 (IL‑12), type I interferons (IFNs) and IFNγ) can serve as the signal 3 that is required for survival during proliferation. The precise nature of signal 3 depends on the specific pathogen used to elicit the response suggesting redundancy and complexity in the choice of which pro-inflammatory cytokines serve as signal 3 for CD8+ T‑cell survival during the proliferative expansion. b | 5–10% of CD8+ T cells detected at the peak of the expansion survive the contraction phase and initiate the memory CD8+ T‑cell pool. The contraction phase is diminished in the absence of inflammation (for example the absence of IFNγ) or when the balance and activation state of pro- and anti-apoptotic B‑cell lymphoma 2 (BCL-2) family members expressed by CD8+ T cells is altered (for example in the absence of BCL-2-interacting mediator of cell death (BIM)). Diminished contraction results in increased numbers of memory CD8+ T cells.
Contraction. In most models of infection or vaccination in which CD8+ T‑cell responses are engendered from the endogenous repertoire or low numbers of transferred TCR-transgenic T cells, the expansion phase peaks at days 5–8 after activation. Proliferation is followed by a transi‑ tion to the contraction phase, which is characterized by a decline in numbers of antigen-specific CD8+ T cells in all tissues (FIG. 4b). The onset of the contraction phase often correlates with the clearance of acute infection, therefore, the textbook notion is that the T cells somehow sense the clearance of infection and then ‘downsize’ to preserve the flexibility of the immune system to respond against new pathogens. However, experimental manipu‑ lations (such as antibiotic treatment beginning 2 days after infection with L. monocytogenes) that truncate the infection had little impact on the onset or magnitude of CD8+ T‑cell contraction41. In addition, the onset and ini‑ tial kinetics of antigen-specific CD8+ T‑cell contraction were similar after acute virus infections to that following chronic or persistent infection41,74,98, although the persist‑ ent and chronic infections do alter long-term CD8+ T‑cell memory69,99–101. Although it is not completely clear that CD8+ T‑cell contraction is similarly regulated following acute and persistent infection, these data suggest that contraction is ‘programmed’ (independently of pathogen or antigen clearance) into the responding CD8+ T‑cell population by early signals after infection41.
a
Early expansion
Activation
Plus signal 3 Naive antigen-specific CD8+ T cell
IL-12, IFNγ and type I IFNs
Naive CD8+ T cell Minus signal 3
Peak expansion
Effector CD8+ T cell
Without one or more cytokines (such as IL-12, IFNγ or type I IFN)
Apoptotic cell
b
Peak expansion
Contraction With specific cytokines (such as IFNγ)
Memory
Without specific cytokines (such as IFNγ) or pro-apoptotic BIM
112 | february 2008 | volume 8
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS
B-cell-lymphoma‑2 family (BCL‑2 family). A family of proteins that contain at least one BCL‑2-homology domain (BH). The family is classified into three groups: antiapoptotic multidomain proteins, such as BCL‑2, BCL-XL and MCL1 (myeloid-cell leukaemia sequence 1), which contain four BHs; proapoptotic multidomain proteins, such as BAX (BCL‑2associated X protein) and BAK (BCL‑2 antagonist/killer), which contain three BHs; and a proapoptotic subfamily of proteins that contain only one BH, the BH3-only proteins.
Activation-induced cell death (AICD). A form of regulated cell death that is induced during lymphocyte activation.
P14 mice Transgenic mice of which the T cells recognize a peptide derived from lymphocytic choriomeningitis virus in the context of H2-Db.
Prime–boost vaccination When a single application of a vaccine is insufficient, repeated immunizations are administered using the same vaccine preparation (homologous prime boost) or using different vaccine preparations (heterologous prime boost) to sequentially stimulate a better immune response. Prior exposure to one strain causes an antibody T-cell response to shared epitopes following exposure to a second strain.
The molecular mechanism(s) controlling contrac‑ tion of the antigen-specific CD8+ T‑cell population are incompletely understood. Early in vitro studies showed that effector, but not memory CD8+ T cells exhibited heightened sensitivity to death after antigen stimulation (a process known as activation-induced cell death (AICD))11,13. However, AICD appears to require the CD95- and TNFmediated cell-death pathways, whereas in vivo studies showed normal contraction of CD8+ T cells that lack both of these molecules102,103. Major current models involve alterations in the balance and activation state of pro- and anti-apoptotic BCL-2 family members expressed by CD8+ T cells104–108. Indeed, deletion of the BCL-2 family member BIM (BCL‑2-interacting mediator of cell death) severely compromises the deletion of superantigenactivated T cells in vivo105 and results in reduced contrac‑ tion of antigen-specific CD8+ T cells after herpesvirus infection106. However, the antigen-specific CD8+ T‑cell population of certain specificities eventually contract in BIM-deficient mice infected with LCMV, suggesting that additional pathways contribute to the death of previously activated CD8+ T cells109–111. Importantly, several studies suggest that inflammation and specifically IFNγ and IL‑12 can influence the con‑ traction phase (FIG. 4b). Initial studies demonstrated that CD8+ T cells failed to contract in BALB/c IFNγ-deficient mice after infection with an attenuated (actA-deficient) strain of L. monocytogenes that was otherwise cleared with similar kinetics in wild-type and immunocompromised hosts112. Reduced contraction is also observed in BALB/c IFNγ-deficient mice and B6 IFNγ-deficient mice after LCMV infection112,113, although recent reverse-transcrip‑ tion PCR analyses that show persistence of LCMV in B6 IFNγ-deficient mice complicates this interpretation114. Importantly, after infection with L. monocytogenes, both CD8+ and CD4+ T cells exhibit abnormal contraction in IFNγ-deficient or IFNGR-deficient mice that were treated with antibiotics to eliminate the infection at day 4. So, persistent infection does not explain how IFNγ regulates CD8+ (and CD4+) T‑cell contraction after infection with L. monocytogenes115,116. Finally, L. monocytogenes infec‑ tion of ampicillin-pretreated wild-type mice results in the generation of CD8+ memory T cells in the absence of contraction115. However, simultaneous induction of proinflammatory cytokines, including IFNγ, by injection of CpG-containing DNA (a TLR9 agonist) induced contrac‑ tion of the responding wild-type CD8+ T cells without altering their expansion115. Other studies demonstrate that antigen-specific CD8+ T cells receive IFNγ signals within 12 hours after L. monocytogenes infection117. Although the initial thinking was that pro-inflammatory cytokines were regulating contraction by signalling directly to CD8+ T cells117,118, recent studies suggests that IFNγ may also act indirectly through cells other than T cells to influence CD8+ T‑cell contraction90,113. Finally, the absence of the IL‑12 receptor on antigen-specific T cells not only reduced expansion but also reduced contraction after infection119. Together, these studies are consistent with the notion that pro-inflammatory cytokines contribute important signals to regulate the contraction phase, perhaps by regulating BCL-2-family members.
nature reviews | immunology
Recent studies also suggested that pro-inflammatory signals, specifically IL‑12, control the amount of T‑bet expression in the responding CD8+ T cells66. In this model, relative T‑bet expression is thought to differentiate short-lived effector CD8+ T cells (T-bethi) from memory precursor effector CD8+ T cells (T-betlow) and thus deter‑ mines which cells survive contraction. Although these data showed a clear correlation between inflammation, T‑bet expression and the phenotype of the responding CD8+ T cells, the link between T‑bet and the contraction of the CD8+ T‑cell response is less clear, as T‑bet-deficient TCR-transgenic CD8+ T cells from P14 mice still contract by more than 80% from the peak of the response follow‑ ing LCMV infection66 and the LCMV-specific CD8+ T‑cell response appears to undergo normal contraction in T‑bet-deficient mice64. Therefore, studies to determine the roles pro-inflammatory cytokines have in regulating contraction will continue to be of high priority. Despite these advances, none of the current models provides an adequate explanation for the ‘counting’ mechanisms that ensure the survival of a precise fraction (5–10%) of the responding CD8+ T cells to seed the initial memory pool. In this regard, recent data demonstrate that naive CD8+ T cells undergo an initial asymmetric division after interaction with antigen-laden APCs120, suggesting the intriguing notion that segregation of those cells with the potential to give rise to long-term memory T cells versus short-lived effector‑T-cell populations may be a very early event in the CD8+ T‑cell response. It will be of great interest to determine how pro-inflammatory cytokines influence the asymmetric first division of the CD8+ T‑cell response.
Rate of memory CD8+ T‑cell generation Defining when ‘memory’ is established after infection or vaccination remains a thorny issue for the field, which complicates the comparison between different model systems and is of great importance in the application of prime–boost vaccination. Perhaps the only uniformly accepted characteristic of memory CD8+ T cells is their extensive heterogeneity in phenotype13,21–23,54. Although there is no single phenotypic marker that can be used to unequivocally identify a memory T cell, multi parameter flow cytometric approaches in conjunction with tetramer- or ICS-based identification of antigenspecific CD8+ T cells provides a means to characterize how the phenotype of CD8+ T cells evolves with time following infection and how this phenotype corre‑ lates with functional attributes. One example of this approach is the division of memory CD8+ T cells into CD62LlowCCR7low effector memory and CD62LhiCCR7hi central memory subsets based primarily on the expres‑ sion of cell-surface receptors that control homing to lymph nodes15,25–27. However, it is clear that responding CD8+ T‑cell populations modulate the expression of a number of phenotypic and functional markers as they progress to the memory phenotype (BOX 1). Importantly, the changes in expression of each marker are not nec‑ essarily coordinate and are quite dramatically depend‑ ent on the initial numbers of naive T‑cell precursors seeded in TCR-transgenic adoptive transfer assays38. volume 8 | february 2008 | 113
© 2008 Nature Publishing Group
REVIEWS In addition, the phenotypic changes associated with the memory phenotype occur relatively rapidly after some infections (for example, within ~1–2 months after L. monocytogenes infection) and more slowly after other infections (for example, >6 months after LCMV infection)27,121. Whether the changes observed with time in the phenotype of antigen-specific CD8+ T‑cell popu‑ lations reflect changes in gene expression within indi‑ vidual cells or outgrowth of specific populations, owing to different rates of basal proliferation, remains an area of controversy in the field27,36, and both mechanisms may ultimately contribute to the phenotypic evolution of the memory CD8+ T‑cell population. An alternative way to define memory CD8+ T‑cell populations is on the basis of their behaviour follow‑ ing restimulation. In one definition13, effective memory CD8 + T cells are capable of vigorous secondary expansion in response to rechallenge or booster immu‑ nization that leads to an increased number of second‑ ary memory CD8+ T cells compared with that of primary memory CD8+ T cells. This boosted number of second‑ ary memory CD8+ T cells will often be associated with increased protection11,13. Using this definition, 40–60 days after acute infection or vaccination with inflam‑ matory adjuvants is generally sufficient for antigenspecific CD8+ T‑cell populations to acquire effective memory characteristics. Interestingly, this time frame is consistent with evidence that global gene expression patterns in memory CD8+ T‑cell populations appear to stabilize around 40 days post infection122. However, this functional definition does not correlate particularly well with the rate that T‑cell populations convert from an effector memory phenotype to a central memory phenotype after many acute infections. Together, these studies suggest the possibility that the rate at which antigen-specific CD8+ T‑cell popula‑ tions acquire characteristics of the memory phenotype might not be constant but rather dependent on the nature of the pathogen or vaccination strategy used to elicit the response (FIG. 5). Direct support for this notion stems from experiments showing that immu‑ nization with peptide-coated mature DCs dramatically accelerates the rate at which antigen-specific CD8 + T cells acquire both the phenotypic 123,124 and func‑ tional characteristics of memory T cells, including the ‘boostability’ of late memory T cells123. Importantly, the early memory phenotype and functional attributes Box 1 | Phenotype of CD8+ T‑cell populations Most, but not all, of the CD8+ T cells at the peak of the expansion phase of the response that follows acute infection have an effector cell phenotype that can be approximated as CD27midCD44hiCD62LlowCD127lowCCR7lowKLRG1hi (killer-cell lectin-like receptor subfamily G, member 1) GZMBhi (granzyme B), with the population exhibiting a functional profile of high interferon‑γ (IFNγ), medium tumour-necrosis factor (TNF), and no interleukin‑2 (IL‑2) production after in vitro antigen stimulation13,21‑23,54. Expression of these phenotypic and functional attributes in the population of antigen-specific CD8+ T cells changes with time after clearance of infection, with most T cells at late time points exhibiting a CD27hi CD44hiCD62LhiCD127hiCCR7hiKLRG1lowGZMB– phenotype, with high IFNγ and TNF and medium IL‑2 production13,21–23,54.
114 | february 2008 | volume 8
of DC‑stimulated CD8+ T cells reverted to effector (that is, non-boostable) status if the immunization with peptide-coated DCs was accompanied with a pro-inflammatory stimulus 123 (FIG. 5c). Interestingly, the CD8 + T cells stimulated by DC immunization underwent apparently normal contraction, suggest‑ ing that sufficient signals were present in this vacci‑ nation scheme to programme the contraction phase of the T‑cell response even in an environment with reduced inflammation. Consistent with these notions, CD8+ T cells that respond to LCMV infection in IL‑12receptor-deficient mice119 or T‑bet-deficient mice64 exhibit rapid acquisition of memory characteristics. Together, these studies suggest that the early inflammatory milieu dictates the rate at which CD8+ T cells acquire memory characteristics and that this process does not occur via changes in the absolute number of antigen-specific CD8+ T cells nor does it dramatically influence the dynamics of the response. Under this scenario, the quantity or quality of pro-inflammatory cytokines determine whether CD8+ T cells exhibit sustained effector phenotype and func‑ tion to combat the active infection or rapidly progress to memory status. In summary, pro-inflammatory cytokines can influ‑ ence the proliferative expansion, the programme of contraction and the rate of memory differentiation of responding CD8+ T cells. These data underscore how the adaptive immune system has learned to use signals from the innate immune system to shape the CD8+ T‑cell response to deal with specific pathogens. In addition to inflammatory cytokines, it is clear that varying TCRtransgenic T‑cell precursor frequency (and therefore competition for signals 1, 2 and 3) can alter the differen‑ tiation of memory T‑cell populations36,38 and that these populations can change their phenotype and functional attributes depending on their location24. Therefore, the timing or intensity of multiple input signals can regulate key characteristics of memory development after infection or immunization. In addition, these data highlight an important unre‑ solved question in the biology of memory T cells — what is the immunological ‘rationale’ for the extensive hetero geneity in phenotype and function in memory T‑cell populations? That is, even populations of memory T cells that are specific for the same epitope, generated from endogenous precursors or physiological numbers of TCR-transgenic T cells, display a diverse repertoire of phenotypic and functional attributes and this hetero geneity persists for long periods of time after the initial stimulation. Our working hypothesis is that the hetero‑ geneity in memory T‑cell phenotype and function main‑ tains subsets of T cells with heightened responsiveness and facilitates the broadest tissue coverage to provide the host with the greatest capacity to deal with recurrent infections or with reservoirs of microorganisms that may persist in specific tissues. This notion may be consistent with the more activated phenotype of memory T‑cell populations that are observed during full-blown chronic infections70. By contrast, memory T‑cell populations will become more homogenous with time following clear‑ ance of acute infections. However, even in this situation www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS Activation
a LCMV infection Antigen non-specific T cell
Progression to memory Antigenspecific T cell
Effector CD8+ T cell Memory CD8+ T cell
Mature DC
Apoptotic cell
Naive DC
Inflammation
107
106
106
105
b L. monocytogenes infection
‘Normal’ progression to memory (pathogen dependent)
Unmanipulated
Inflammation
Antibiotic post-treatment Accelerated progression to memory 106
Inflammation
105
Antibiotic pre-treatment Rapid memory generation without contraction
?
105
Inflammation
105
c Peptide–DC vaccination Immunization with in vitro peptide-coated mature DCs
Rapid memory generation with contraction 106
105
Inflammation
Figure 5 | Inflammation regulates the progression of CD8+ T cells to memory. Acute infection unmanipulated Nature of Reviews | Immunology hosts generates an antigen-specific CD8+ T‑cell response that will undergo relatively slow conversion from effector to memory cell phenotype and function. In that scenario, inflammatory cytokines participate in the maturation of dendritic cells (DCs) to facilitate priming of naive CD8+ T cells, and also deliver inflammatory signal 3 to primed CD8+ T cells for maximal proliferative expansion. a, b | The kinetics of the CD8+ T‑cell responses are pathogen dependent as CD8+ T cells with the same antigen specificity will undergo different amounts of proliferative expansion and acquire memory characteristics at different rates depending on whether they are triggered by viruses (for example lymphocytic choriomeningitis virus (LCMV)) or bacteria (such as Listeria monocytogenes). However, in both models of infection a substantial interval (weeks for L. monocytogenes-induced and months for LCMV-induced CD8+ T‑cell responses) is required for the conversion from effector to memory CD8+ T cells. b | Decreasing the inflammatory response after L. monocytogenes infection and priming of CD8+ T cells (by the post-infection administration of antibiotics) does not change the overall kinetics of CD8+ T-cell responses (that is, the magnitude of expansion, contraction and memory CD8+ T‑cell numbers), but accelerates the acquisition of memory characteristics. Blocking inflammation and bacterial infection throughout L. monocytogenes infection (by pre-infection administration of antibiotics) impairs the magnitude of expansion and also reduces contraction. This scenario also accelerates the transition from effector to memory CD8+ T cells. c | CD8+ T‑cell responses evoked by in vitro matured peptide-coated DC vaccination are primed in an environment of globally decreased inflammation. Immunization with in vitro activated DCs results in robust expansion (that is, an apparently normal signal 3) and normal contraction of antigen-specific CD8+ T‑cell responses. However, the responding CD8+ T cells do not receive the full complement of inflammatory cytokines such as that produced during infection, and conversion to the memory phenotype and function are accelerated. Accelerated memory phenotype is also observed in T‑bet-deficient or interleukin‑12-receptor (IL‑12R)-deficient CD8+ T‑cell responses or when responses are generated from high-input numbers of T‑cell receptor (TCR)-transgenic T cells38,64,66,119. The numbers in the boxes represent approximate numbers of antigen-specific CD8+ T cells at various stages of the response.
nature reviews | immunology
volume 8 | february 2008 | 115 © 2008 Nature Publishing Group
REVIEWS the continued trafficking of memory T cells through tissues might cause a subset of these cells to undergo phenotypic and functional changes that preserve hetero‑ geneity and maintain immunity. Given our incomplete understanding, precise dissection of the signals that control heterogeneity in memory T‑cell populations remains an important goal. As detailed below, at least some of the important input signals in memory T‑cell generation are amendable to manipulations that could improve vaccination.
Manipulating memory to improve vaccination In many cases, increasing the numbers of memory CD8+ T cells correlates with increased protection from reinfection11,13. As memory CD8+ T-cell numbers are generally proportional to the peak in expansion 125, one approach to improve vaccination is to manipulate signals 1, 2 and 3, to ensure the most robust expansion of antigen-specific CD8+ T cells and thereby lead to the highest number of memory T cells. This strategy has and continues to receive extensive attention, with the development of strong adjuvants as a major goal to improve vaccination. In essence, the goal of such approaches is to devise vaccination regimens that most closely approximate the conditions that occur during infection. However, as discussed, a highly inflamed environment will pro‑ mote contraction and also result in the slow acquisi‑ tion of memory characteristics by the responding CD8+ T cells115,121,123. Under certain circumstances, such as IFNγ deficiency, contraction of antigen-specific CD8+ T cells can be limited, resulting in increased numbers of memory CD8+ T cells112,113. Importantly, these increased memory CD8 + T‑cell numbers in IFNγ-deficient mice provided substantial immunity against virulent L. monocytogenes infection126. In addition, limiting the exposure of antigen-specific CD8+ T cells to proinflammatory cytokines (for example, vaccination with peptide-coated DCs) can accelerate the acquisition of memory characteristics and the ability to respond to booster immunizations within days after the initial immunization121,123. Together, these data suggest that judicious interference with pro-inflammatory cytokine signalling to the responding CD8+ T cells has the poten‑ tial to shorten the time interval required for boosting the numbers of antigen-specific CD8+ T cells, thereby gen‑ erating protective numbers of such cells faster than can be achieved using infectious agents or strong adjuvants in the initial immunization. The challenge is how to generate mature DCs, capable of delivering signals 1, 2 and 3, while preventing the CD8+ T cells from receiving non-essential pro-inflammatory cytokine signals that will delay the acquisition of memory characteristics. Primary versus secondary CD8+ T‑cell memory Finally, almost all research in the field of CD8+ memory T cells has focused on the behaviour of these cells after primary immunization. However, many successful vac‑ cination regimens, including those in humans, require booster immunizations to engender the most robust protective memory T-cell response127. Importantly, 116 | february 2008 | volume 8
recent studies demonstrate that the characteristics of secondary memory CD8 + T cells are quite differ‑ ent when compared with the primary CD8 + T‑cell populations from which they are derived (FIG. 6). For example, secondary memory CD8 + T cells undergo protracted contraction, show substantially delayed upregulation of central memory T‑cell characteris‑ tics, and fail to undergo basal proliferation compared with primary memory CD8+ T cells (either effector memory or central memory T cells) examined at the same time point after infection41,78,128–131. Interestingly, secondary memory CD8+ T cells exhibited sustained granzyme B expression and provided better protection than primary memory CD8+ T cells against challenge with L. monocytogenes78,130. An important issue for the future, which could influence vaccine design, is to determine whether secondary memory CD8 + T cells will exhibit enhanced protection against all or only a subset of pathogens. Why would the number of antigen encounters control the behaviour of memory T‑cell populations? It seems apparent that maintaining heightened effec‑ tor functions (such as granzyme B production) would be beneficial in areas where recurrent infections with the same pathogen are common. In addition, delayed upregulation of CD62L may improve immunity under some circumstances, by ensuring that a large pool of secondary memory CD8+ T cells are localized to tertiary tissues, where they may immediately defend against local reinfection. Although memory CD8 + T cells exhibited robust proliferative responses to secondary and tertiary stimulation, in one study, the CD8+ T‑cell response began to decline by the third restimulation, suggesting that memory CD8+ T cells are able to undergo an extensive but a finite number of divisions78. Importantly, the failure of secondary memory CD8 + T cells to upregulate CD62L also prevents these populations from entering lymph nodes. Therefore, secondary memory CD8 + T‑cell populations are more permissive than CD62Lexpressing primary memory CD8+ T‑cell populations in allowing the priming of new naive CD8 + T cells that will give rise to replacement memory popula‑ tions 130. Finally, it remains to be determined at the molecular level why primary and secondary memory CD8 + T‑cell populations exhibit different charac‑ teristics. Based on the knowledge that high versus low input numbers of naive TCR-transgenic T cells result in dramatic differences in both proliferation and differentiation, we speculate that the increased number of divisions undertaken by secondary compared with primary memory CD8+ T cells may be an important parameter in determining how the number of antigen exposures regulates the characteris‑ tics of each memory population. How the properties of secondary memory CD8+ T cells are influenced by inflammation and/or CD4+ T‑cell help, and whether these properties can be manipulated in a similar man‑ ner to the primary memory CD8+ T‑cell response are important questions with direct relevance to vaccines based on a prime–boost strategy. www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS a
b
Number of antigen-specific CD8+ T cells
Secondary effector response Primary effector response
Secondary memory response
Primary memory response
CD62L LOD
1st challenge
Primary memory CD8+ T cells IL-2-producing memory CD8+ T cell
CD8+ memory T cell
Memory turnover
Secondary memory CD8+ T cells
2nd challenge Time ? ‘Early’ memory (months)
‘Late’ memory (years)
Figure 6 | Primary and secondary CD8+ T‑cell memory. a | Kinetics of the CD8+ T‑cell response to the initial (primary memory response) and subsequent (secondary memory response) infections. Booster infection or vaccination can increase Nature Reviews | Immunology secondary memory CD8+ T‑cell numbers compared with primary stimulation. b | The number of antigen encounters has a dramatic impact on the characteristics of the resulting CD8+ T‑cell memory populations. For example, most memory CD8+ T cells (induced after primary immunization) will convert into CD62L+CD8+ T cells weeks (with Listeria monocytogenes) or months (with lymphocytic choriomeningitis virus (LCMV)) after primary infection. A fraction (~30–40%) of these stable primary memory populations will produce interleukin‑2 (IL‑2) following antigen stimulation. Following secondary challenge primary memory CD8+ T cells will rapidly expand and become secondary effector CD8+ T cells (secondary immunization) that will go through a prolonged contraction phase to establish secondary memory CD8+ T cells at numbers that are higher than those obtained following primary infection. Secondary memory CD8+ T cells show a substantially decreased rate of memory turnover, fail to upregulate CD62L expression and produce IL-2. In a limited number of reports to date secondary memory CD8+ T cells need years after secondary challenge to become predominately CD62Lhi T cells. LOD, limit of detection.
Concluding remarks The upshot of these and many other studies that we could not discuss owing to space limitations is that vari‑ ous signals, acting in concert, contribute to the ultimate characteristics of memory CD8 + T‑cell populations elicited by infection or vaccination. Interestingly, many of these signals such as pro-inflammatory cytokines originate from the innate immune system and depend on activation of specific PRRs to identify the nature of the pathogen. Therefore, different pathogens will elicit innate immune responses with both common and unique features, which in turn will ‘tailor’ the adaptive memory CD8+ T‑cell response. In this regard, future studies to determine the differences in the memory phenotype that is generated after infections with diverse pathogens, in concert with detailed analyses of concurrent innate immune responses, will be essential to determine the range of heterogeneity in CD8+ T‑cell memory populations. Major challenges for the future include identification of the precise input signals that
1. 2. 3. 4.
Jameson, S. C. Maintaining the norm: T‑cell homeostasis. Nature Rev. Immunol. 2, 547–556 (2002). Blattman, J. N. et al. Estimating the precursor frequency of naive antigen-specific CD8 T cells. J. Exp. Med. 195, 657–664 (2002). Casrouge, A. et al. Size estimate of the αβTCR repertoire of naive mouse splenocytes. J. Immunol. 164, 5782–5787 (2000). Kedzierska, K. et al. Quantification of repertoire diversity of influenza-specific epitopes with
5.
6.
shape CD8+ T‑cell memory and determination of the molecular basis for how the responding CD8+ T cells decode the myriad of signals encountered during immune responses to generate effective memory. This information will provide the crucial platform to design interventions to manipulate specific input signals and enhance vaccination by improving the numerical or qualitative aspects of the resulting memory CD8+ T‑cell populations. Importantly, many of the signals that shape CD8+ T‑cell memory have the potential to affect other aspects of the adaptive (B cells and CD4+ T cells) or innate (NK cells, macrophages, neutrophils and DCs) immune responses. Therefore, future endeavours should incorporate a holistic approach to enhance vaccination. Finally, detailed analyses to identify how and why mul‑ tiple antigen exposures influence the characteristics of memory CD8+ T‑cell responses remains a largely untapped but potentially fertile area of experimentation with important implications for human vaccines based on prime–boost strategies.
predominant public or private TCR usage. J. Immunol. 177, 6705–6712 (2006). Pewe, L. L., Netland, J. M., Heard, S. B. & Perlman, S. Very diverse CD8 T cell clonotypic responses after virus infections. J. Immunol. 172, 3151–3156 (2004). Moon, J. J. et al. Naive CD4+ T cell frequency varies for different epitopes and predicts repertoire diversity and response magnitude. Immunity 27, 203–213 (2007).
nature reviews | immunology
7. 8. 9.
Heath, W. R. & Carbone, F. R. Cross-presentation, dendritic cells, tolerance and immunity. Annu. Rev. Immunol. 19, 47–64 (2001). Steinman, R. M., Hawiger, D. & Nussenzweig, M. C. Tolerogenic dendritic cells. Annu. Rev. Immunol. 21, 685–711 (2003). Germain, R. N., Miller, M. J., Dustin, M. L. & Nussenzweig, M. C. Dynamic imaging of the immune system: progress, pitfalls and promise. Nature Rev. Immunol. 6, 497–507 (2006).
volume 8 | february 2008 | 117 © 2008 Nature Publishing Group
REVIEWS 10. Haring, J. S., Badovinac, V. P. & Harty, J. T. Inflaming the CD8+ T cell response. Immunity 25, 19–29 (2006). 11. Kaech, S. M., Wherry, E. J. & Ahmed, R. Effector and memory T‑cell differentiation: implications for vaccine development. Nature Rev. Immunol. 2, 251–262 (2002). 12. Mescher, M. F. et al. Signals required for programming effector and memory development by CD8+ T cells. Immunol. Rev. 211, 81–92 (2006). 13. Badovinac, V. P. & Harty, J. T. Programming, demarcating, and manipulating CD8+ T‑cell memory. Immunol. Rev. 211, 67–80 (2006). 14. Marshall, D. R. et al. Measuring the diaspora for virus-specific CD8+ T cells. Proc. Natl Acad. Sci. USA 98, 6313–6318 (2001). 15. Masopust, D., Vezys, V., Marzo, A. L. & Lefrancois, L. Preferential localization of effector memory cells in nonlymphoid tissue. Science 291, 2413–2417 (2001). 16. Homann, D., Teyton, L. & Oldstone, M. B. Differential regulation of antiviral T‑cell immunity results in stable CD8+ but declining CD4+ T‑cell memory. Nature Med. 7, 913–919 (2001). 17. Becker, T. C. et al. Interleukin 15 is required for proliferative renewal of virus-specific memory CD8 T cells. J. Exp. Med. 195, 1541–1548 (2002). 18. Goldrath, A. W. et al. Cytokine requirements for acute and Basal homeostatic proliferation of naive and memory CD8+ T cells. J. Exp. Med. 195, 1515–1522 (2002). 19. Schluns, K. S., Kieper, W. C., Jameson, S. C. & Lefrancois, L. Interleukin‑7 mediates the homeostasis of naive and memory CD8 T cells in vivo. Nature Immunol. 1, 426–432 (2000). 20. Tan, J. T. et al. Interleukin (IL)‑15 and IL‑7 jointly regulate homeostatic proliferation of memory phenotype CD8+ cells but are not required for memory phenotype CD4+ cells. J. Exp. Med. 195, 1523–1532 (2002). 21. Hikono, H. et al. T‑cell memory and recall responses to respiratory virus infections. Immunol. Rev. 211, 119–132 (2006). 22. Kedzierska, K., La Gruta, N. L., Turner, S. J. & Doherty, P. C. Establishment and recall of CD8+ T‑cell memory in a model of localized transient infection. Immunol. Rev. 211, 133–145 (2006). 23. Lefrancois, L. Development, trafficking, and function of memory T‑cell subsets. Immunol. Rev. 211, 93–103 (2006). 24. Marzo, A. L., Yagita, H. & Lefrancois, L. Cutting edge: migration to nonlymphoid tissues results in functional conversion of central to effector memory CD8 T cells. J. Immunol. 179, 36–40 (2007). 25. 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). 26. Weninger, W., Manjunath, N. & von Andrian, U. H. Migration and differentiation of CD8+ T cells. Immunol. Rev. 186, 221–233 (2002). 27. Wherry, E. J. et al. Lineage relationship and protective immunity of memory CD8 T cell subsets. Nature Immunol. 4, 225–234 (2003). 28. Altman, J. D. et al. Phenotypic analysis of antigenspecific T lymphocytes. Science 274, 94–96 (1996). 29. Badovinac, V. P. & Harty, J. T. Intracellular staining for TNF and IFN‑γ detects different frequencies of antigenspecific CD8+ T cells. J. Immunol. Methods 238, 107–117 (2000). 30. Busch, D. H., Pilip, I. M., Vijh, S. & Pamer, E. G. Coordinate regulation of complex T cell populations responding to bacterial infection. Immunity 8, 353– 362 (1998). 31. Butz, E. A. & Bevan, M. J. Massive expansion of antigen-specific CD8+ T cells during an acute virus infection. Immunity 8, 167–175 (1998). 32. Kearney, E. R., Pape, K. A., Loh, D. Y. & Jenkins, M. K. Visualization of peptide-specific T cell immunity and peripheral tolerance induction in vivo. Immunity 1, 327–339 (1994). 33. Murali-Krishna, K. et al. Counting antigen-specific CD8 T cells: a reevaluation of bystander activation during viral infection. Immunity 8, 177–187 (1998). 34. Hataye, J., Moon, J. J., Khoruts, A., Reilly, C. & Jenkins, M. K. Naive and memory CD4+ T cell survival controlled by clonal abundance. Science 312, 114–116 (2006). This study shows that memory CD4+ T‑cell responses are influenced by the number of TCRtransgenic CD4+ T cells that are initially transferred prior to immunization.
35. Kemp, R. A., Powell, T. J., Dwyer, D. W. & Dutton, R. W. Cutting edge: regulation of CD8+ T cell effector population size. J. Immunol. 173, 2923–2927 (2004). 36. Marzo, A. L. et al. Initial T cell frequency dictates memory CD8+ T cell lineage commitment. Nature Immunol. 6, 793–799 (2005). This study together with reference 38 shows that all phases of the CD8+ T‑cell response to infection are influenced by the number of TCR-transgenic CD8+ T cells initially transferred prior to infection. 37. Probst, H. C., Dumrese, T. & van den Broek, M. F. Cutting edge: competition for APC by CTLs of different specificities is not functionally important during induction of antiviral responses. J. Immunol. 168, 5387–5391 (2002). 38. Badovinac, V. P., Haring, J. S. & Harty, J. T. Initial T cell receptor transgenic cell precursor frequency dictates critical aspects of the CD8+ T cell response to infection. Immunity 26, 827–841 (2007). 39. Hogquist, K. A. et al. T cell receptor antagonist peptides induce positive selection. Cell 76, 17–27 (1994). 40. Reiner, S. L., Sallusto, F. & Lanzavecchia, A. Division of labor with a workforce of one: challenges in specifying effector and memory T cell fate. Science 317, 622–625 (2007). 41. Badovinac, V. P., Porter, B. B. & Harty, J. T. Programmed contraction of CD8+ T cells after infection. Nature Immunol. 3, 619–626 (2002). 42. Kaech, S. M. & Ahmed, R. Memory CD8+ T cell differentiation: initial antigen encounter triggers a developmental program in naive cells. Nature Immunol. 2, 415–422 (2001). 43. Mercado, R. et al. Early programming of T cell populations responding to bacterial infection. J. Immunol. 165, 6833–6839 (2000). 44. van Stipdonk, M. J., Lemmens, E. E. & Schoenberger, S. P. Naive CTLs require a single brief period of antigenic stimulation for clonal expansion and differentiation. Nature Immunol. 2, 423–429 (2001). References 41–44 show that the initial antigenencounter is sufficient to programme CD8+ T‑cell expansion, contraction and memory CD8+ T‑cell generation. 45. Bevan, M. J. Helping the CD8+ T‑cell response. Nature Rev. Immunol. 4, 595–602 (2004). 46. Bennett, S. R. et al. Help for cytotoxic‑T‑cell responses is mediated by CD40 signalling. Nature 393, 478–480 (1998). 47. Ridge, J. P., Di Rosa, F. & Matzinger, P. A conditioned dendritic cell can be a temporal bridge between a CD4+ T‑helper and a T‑killer cell. Nature 393, 474–478 (1998). 48. Schoenberger, S. P., Toes, R. E., van der Voort, E. I., Offringa, R. & Melief, C. J. T‑cell help for cytotoxic T lymphocytes is mediated by CD40–CD40L interactions. Nature 393, 480–483 (1998). 49. Janssen, E. M. et al. CD4+ T‑cell help controls CD8+ T‑cell memory via TRAIL-mediated activation-induced cell death. Nature 434, 88–93 (2005). 50. Bourgeois, C., Rocha, B. & Tanchot, C. A role for CD40 expression on CD8+ T cells in the generation of CD8+ T cell memory. Science 297, 2060–2063 (2002). 51. Janssen, E. M. et al. CD4+ T cells are required for secondary expansion and memory in CD8+ T lymphocytes. Nature 421, 852–856 (2003). 52. Shedlock, D. J. & Shen, H. Requirement for CD4 T cell help in generating functional CD8 T cell memory. Science 300, 337–339 (2003). 53. Sun, J. C. & Bevan, M. J. Defective CD8 T cell memory following acute infection without CD4 T cell help. Science 300, 339–342 (2003). 54. Williams, M. A., Holmes, B. J., Sun, J. C. & Bevan, M. J. Developing and maintaining protective CD8+ memory T cells. Immunol. Rev. 211, 146–153 (2006). 55. Sun, J. C., Williams, M. A. & Bevan, M. J. CD4+ T cells are required for the maintenance, not programming, of memory CD8+ T cells after acute infection. Nature Immunol. 5, 927–933 (2004). 56. Badovinac, V. P., Messingham, K. A., Griffith, T. S. & Harty, J. T. TRAIL deficiency delays, but does not prevent, erosion in the quality of “helpless” memory CD8 T cells. J. Immunol. 177, 999–1006 (2006). 57. Marzo, A. L. et al. Fully functional memory CD8 T cells in the absence of CD4 T cells. J. Immunol. 173, 969–975 (2004). 58. Ramsburg, E. A., Publicover, J. M., Coppock, D. & Rose, J. K. Requirement for CD4 T cell help in
118 | february 2008 | volume 8
59. 60. 61.
62.
63.
64.
65. 66.
67. 68. 69.
70. 71.
72. 73. 74.
75.
76.
77.
78.
79.
maintenance of memory CD8 T cell responses is epitope dependent. J. Immunol. 178, 6350–6358 (2007). Khanolkar, A., Badovinac, V. P. & Harty, J. T. CD8 T cell memory development: CD4 T cell help is appreciated. Immunol. Res. 39, 94–104 (2007). D’Souza, W. N. & Lefrancois, L. IL‑2 is not required for the initiation of CD8 T cell cycling but sustains expansion. J. Immunol. 171, 5727–5735 (2003). Williams, M. A., Tyznik, A. J. & Bevan, M. J. Interleukin‑2 signals during priming are required for secondary expansion of CD8+ memory T cells. Nature 441, 890–893 (2006). Yu, A. et al. Efficient induction of primary and secondary T cell-dependent immune responses in vivo in the absence of functional IL‑2 and IL‑15 receptors. J. Immunol. 170, 236–242 (2003). Boyman, O., Kovar, M., Rubinstein, M. P., Surh, C. D. & Sprent, J. Selective stimulation of T cell subsets with antibody-cytokine immune complexes. Science 311, 1924–1927 (2006). Intlekofer, A. M. et al. Requirement for T‑bet in the aberrant differentiation of unhelped memory CD8+ T cells. J. Exp. Med. 204, 2015–2021 (2007). This paper together with reference 66 identifies a key role for T‑bet in controlling the fate determination of CD8+ T cells that are responding to infection. Hamilton, S. E. & Jameson, S. C. CD8+ T cell differentiation: choosing a path through T‑bet. Immunity 27, 180–182 (2007). Joshi, N. S. et al. Inflammation directs memory precursor and short-lived effector CD8+ T cell fates via the graded expression of T‑bet transcription factor. Immunity 27, 281–295 (2007). Murali-Krishna, K. et al. Persistence of memory CD8 T cells in MHC class I‑deficient mice. Science 286, 1377–1381 (1999). Klenerman, P. & Hill, A. T cells and viral persistence: lessons from diverse infections. Nature Immunol. 6, 873–879 (2005). 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). Wherry, E. J. & Ahmed, R. Memory CD8 T‑cell differentiation during viral infection. J. Virol. 78, 5535–5545 (2004). Barber, D. L. et al. Restoring function in exhausted CD8 T cells during chronic viral infection. Nature 439, 682–687 (2006). This paper reports the use of PD1–PD1 ligand blockade therapy to enhance the numbers and functional status of LCMV-specific CD8+ T cells and facilitate the clearance of chronic LCMV infection. Brooks, D. G. et al. Interleukin‑10 determines viral clearance or persistence in vivo. Nature Med. 12, 1301–1309 (2006). Ejrnaes, M. et al. Resolution of a chronic viral infection after interleukin‑10 receptor blockade. J. Exp. Med. 203, 2461–2472 (2006). 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). Fuller, M. J. et al. Cutting edge: emergence of CD127high functionally competent memory T cells is compromised by high viral loads and inadequate T cell help. J. Immunol. 174, 5926–5930 (2005). 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). Lang, K. S. et al. Inverse correlation between IL‑7 receptor expression and CD8 T cell exhaustion during persistent antigen stimulation. Eur. J. Immunol. 35, 738–745 (2005). Masopust, D., Ha, S. J., Vezys, V. & Ahmed, R. Stimulation history dictates memory CD8 T cell phenotype: implications for prime-boost vaccination. J. Immunol. 177, 831–839 (2006). This paper together with references 128–131 describe the fundamental phenotypic and functional differences of CD8+ T cells that are responding to multiple rounds of antigenic stimulations compared with primary CD8+ T cells. Moskophidis, D., Lechner, F., Pircher, H. & Zinkernagel, R. M. Virus persistence in acutely infected immunocompetent mice by exhaustion of
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS
80.
81.
82. 83. 84. 85.
86.
87.
88.
89.
90.
91.
92. 93.
94.
95.
96.
97.
antiviral cytotoxic effector T cells. Nature 362, 758– 761 (1993). Vezys, V. et al. Continuous recruitment of naive T cells contributes to heterogeneity of antiviral CD8 T cells during persistent infection. J. Exp. Med. 203, 2263– 2269 (2006). Zajac, A. J., Quinn, D. G., Cohen, P. L. & Frelinger, J. A. Fas-dependent CD4+ cytotoxic T‑cell‑mediated pathogenesis during virus infection. Proc. Natl Acad. Sci. USA 93, 14730–14735 (1996). Greenwald, R. J., Freeman, G. J. & Sharpe, A. H. The B7 family revisited. Annu. Rev. Immunol. 23, 515–548 (2005). Watts, T. H. TNF/TNFR family members in costimulation of T cell responses. Annu. Rev. Immunol. 23, 23–68 (2005). Curtsinger, J. M. et al. Inflammatory cytokines provide a third signal for activation of naive CD4+ and CD8+ T cells. J. Immunol. 162, 3256–3562 (1999). Curtsinger, J. M., Valenzuela, J. O., Agarwal, P., Lins, D. & Mescher, M. F. Type I IFNs provide a third signal to CD8 T cells to stimulate clonal expansion and differentiation. J. Immunol. 174, 4465–4469 (2005). Schmidt, C. S. & Mescher, M. F. Adjuvant effect of IL‑12: conversion of peptide antigen administration from tolerizing to immunizing for CD8+ T cells in vivo. J. Immunol. 163, 2561–2567 (1999). Aichele, P. et al. CD8 T cells specific for lymphocytic choriomeningitis virus require type I IFN receptor for clonal expansion. J. Immunol. 176, 4525–4529 (2006). Curtsinger, J. M., Johnson, C. M. & Mescher, M. F. CD8 T cell clonal expansion and development of effector function require prolonged exposure to antigen, costimulation, and signal 3 cytokine. J. Immunol. 171, 5165–5171 (2003). Kolumam, G. A., Thomas, S., Thompson, L. J., Sprent, J. & Murali-Krishna, K. Type I interferons act directly on CD8 T cells to allow clonal expansion and memory formation in response to viral infection. J. Exp. Med. 202, 637–650 (2005). Sercan, O., Hammerling, G. J., Arnold, B. & Schuler, T. Innate immune cells contribute to the IFN‑γ‑dependent regulation of antigen-specific CD8+ T cell homeostasis. J. Immunol. 176, 735–739 (2006). Whitmire, J. K., Tan, J. T. & Whitton, J. L. Interferongamma acts directly on CD8+ T cells to increase their abundance during virus infection. J. Exp. Med. 201, 1053–1059 (2005). Mitchell, T. C. et al. Immunological adjuvants promote activated T cell survival via induction of Bcl‑3. Nature Immunol. 2, 397–402 (2001). Thomas, S., Kolumam, G. A. & Murali-Krishna, K. Antigen presentation by nonhemopoietic cells amplifies clonal expansion of effector CD8 T cells in a pathogen-specific manner. J. Immunol. 178, 5802– 5811 (2007). Way, S. S., Havenar-Daughton, C., Kolumam, G. A., Orgun, N. N. & Murali-Krishna, K. IL‑12 and type‑I IFN synergize for IFN‑γ production by CD4 T cells, whereas neither are required for IFN‑γ production by CD8 T cells after Listeria monocytogenes infection. J. Immunol. 178, 4498–4505 (2007). Thompson, L. J., Kolumam, G. A., Thomas, S. & Murali-Krishna, K. Innate inflammatory signals induced by various pathogens differentially dictate the IFN‑I dependence of CD8 T cells for clonal expansion and memory formation. J. Immunol. 177, 1746– 1754 (2006). Cousens, L. P. et al. Two roads diverged: interferon α/β- and interleukin 12-mediated pathways in promoting T cell interferon γ responses during viral infection. J. Exp. Med. 189, 1315–1328 (1999). Ou, R., Zhou, S., Huang, L. & Moskophidis, D. Critical role for α/β and γ interferons in persistence of lymphocytic choriomeningitis virus by clonal
exhaustion of cytotoxic T cells. J. Virol. 75, 8407–8423 (2001). 98. Fuller, M. J., Khanolkar, A., Tebo, A. E. & Zajac, A. J. Maintenance, loss, and resurgence of T cell responses during acute, protracted, and chronic viral infections. J. Immunol. 172, 4204–4214 (2004). 99. Fuller, M. J. & Zajac, A. J. Ablation of CD8 and CD4 T cell responses by high viral loads. J. Immunol. 170, 477–486 (2003). 100. Gallimore, A. et al. Induction and exhaustion of lymphocytic choriomeningitis virus-specific cytotoxic T lymphocytes visualized using soluble tetrameric major histocompatibility complex class I‑peptide complexes. J. Exp. Med. 187, 1383–1393 (1998). 101. Zhou, S., Ou, R., Huang, L. & Moskophidis, D. Critical role for perforin‑, Fas/FasL‑, and TNFR1-mediated cytotoxic pathways in down-regulation of antigenspecific T cells during persistent viral infection. J. Virol. 76, 829–840 (2002). 102. Nguyen, L. T. et al. TNF receptor 1 (TNFR1) and CD95 are not required for T cell deletion after virus infection but contribute to peptide-induced deletion under limited conditions. Eur. J. Immunol. 30, 683–688 (2000). 103. Reich, A., Korner, H., Sedgwick, J. D. & Pircher, H. Immune down-regulation and peripheral deletion of CD8 T cells does not require TNF receptor-ligand interactions nor CD95 (Fas, APO‑1). Eur. J. Immunol. 30, 678–682 (2000). 104. Grayson, J. M., Zajac, A. J., Altman, J. D. & Ahmed, R. Cutting edge: increased expression of Bcl‑2 in antigenspecific memory CD8+ T cells. J. Immunol. 164, 3950–3954 (2000). 105. Hildeman, D. A. et al. Activated T cell death in vivo mediated by proapoptotic bcl‑2 family member bim. Immunity 16, 759–767 (2002). 106. Pellegrini, M., Belz, G., Bouillet, P. & Strasser, A. Shutdown of an acute T cell immune response to viral infection is mediated by the proapoptotic Bcl‑2 homology 3‑only protein Bim. Proc. Natl Acad. Sci. USA 100, 14175–14180 (2003). 107. Hildeman, D., Jorgensen, T., Kappler, J. & Marrack, P. Apoptosis and the homeostatic control of immune responses. Curr. Opin. Immunol. 19, 516–521 (2007). 108. Plas, D. R., Rathmell, J. C. & Thompson, C. B. Homeostatic control of lymphocyte survival: potential origins and implications. Nature Immunol. 3, 515–521 (2002). 109. 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). 110. Wojciechowski, S. et al. Bim mediates apoptosis of CD127lo effector T cells and limits T cell memory. Eur. J. Immunol. 36, 1694–1706 (2006). 111. Wojciechowski, S. et al. Bim/Bcl‑2 balance is critical for maintaining naive and memory T cell homeostasis. J. Exp. Med. 204, 1665–1675 (2007). 112. Badovinac, V. P., Tvinnereim, A. R. & Harty, J. T. Regulation of antigen-specific CD8+ T cell homeostasis by perforin and interferon‑γ. Science 290, 1354–1358 (2000). 113. Tewari, K., Nakayama, Y. & Suresh, M. Role of direct effects of IFN‑γ on T cells in the regulation of CD8 T cell homeostasis. J. Immunol. 179, 2115–2125 (2007). 114. Whitmire, J. K., Eam, B., Benning, N. & Whitton, J. L. Direct interferon‑γ signaling dramatically enhances CD4+ and CD8+ T cell memory. J. Immunol. 179, 1190–1197 (2007). 115. Badovinac, V. P., Porter, B. B. & Harty, J. T. CD8+ T cell contraction is controlled by early inflammation. Nature Immunol. 5, 809–817 (2004). 116. Haring, J. S. & Harty, J. T. Aberrant contraction of antigen-specific CD4 T cells after infection in the absence of γ interferon or its receptor. Infect. Immun. 74, 6252–6263 (2006).
nature reviews | immunology
117. Haring, J. S., Corbin, G. A. & Harty, J. T. Dynamic regulation of IFN‑γ signaling in antigen-specific CD8+ T cells responding to infection. J. Immunol. 174, 6791–6802 (2005). 118. Refaeli, Y., Van Parijs, L., Alexander, S. I. & Abbas, A. K. Interferon‑γ is required for activation-induced death of T lymphocytes. J. Exp. Med. 196, 999–1005 (2002). 119. Pearce, E. L. & Shen, H. Generation of CD8 T cell memory is regulated by IL‑12. J. Immunol. 179, 2074–2081 (2007). 120. Chang, J. T. et al. Asymmetric T lymphocyte division in the initiation of adaptive immune responses. Science 315, 1687–1691 (2007). This paper suggests that a single T cell can promote diverse cell fates through an asymmetric initial cell division. 121. Badovinac, V. P. & Harty, J. T. Manipulating the rate of memory CD8+ T cell generation after acute infection. J. Immunol. 179, 53–63 (2007). 122. Kaech, S. M., Hemby, S., Kersh, E. & Ahmed, R. Molecular and functional profiling of memory CD8 T cell differentiation. Cell 111, 837–851 (2002). 123. Badovinac, V. P., Messingham, K. A., Jabbari, A., Haring, J. S. & Harty, J. T. Accelerated CD8+ T‑cell memory and prime-boost response after dendritic-cell vaccination. Nature Med. 11, 748–756 (2005). This report describes how inflammatory signals influence the rate of memory CD8+ T‑cell responses following peptide–DC vaccination. 124. Lacombe, M. H., Hardy, M. P., Rooney, J. & Labrecque, N. IL‑7 receptor expression levels do not identify CD8+ memory T lymphocyte precursors following peptide immunization. J. Immunol. 175, 4400–4407 (2005). 125. Hou, S., Hyland, L., Ryan, K. W., Portner, A. & Doherty, P. C. Virus-specific CD8+ T‑cell memory determined by clonal burst size. Nature 369, 652– 654 (1994). 126. Harty, J. T. & Bevan, M. J. Specific immunity to Listeria monocytogenes in the absence of IFN‑γ. Immunity 3, 109–117 (1995). 127. Woodland, D. L. Jump-starting the immune system: prime-boosting comes of age. Trends Immunol. 25, 98–104 (2004). 128. Badovinac, V. P., Messingham, K. A., Hamilton, S. E. & Harty, J. T. Regulation of CD8+ T cells undergoing primary and secondary responses to infection in the same host. J. Immunol. 170, 4933–4942 (2003). 129. Grayson, J. M., Harrington, L. E., Lanier, J. G., Wherry, E. J. & Ahmed, R. Differential sensitivity of naive and memory CD8+ T cells to apoptosis in vivo. J. Immunol. 169, 3760–3770 (2002). 130. Jabbari, A. & Harty, J. T. Secondary memory CD8+ T cells are more protective but slower to acquire a central-memory phenotype. J. Exp. Med. 203, 919–932 (2006). 131. Unsoeld, H. & Pircher, H. Complex memory T‑cell phenotypes revealed by coexpression of CD62L and CCR7. J. Virol. 79, 4510–4513 (2005).
Acknowledgements
We wish to thank Dr S. Perlman for critical reading of this manuscript. We apologize to those colleagues whose work we could not cite owing to space limitations. Work in the Harty laboratory is supported by grants from the National Institutes of Health, USA.
DATABASES Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene CCR7 | CD25 | CD62L | CD122 | CD127 | IFNγ | IL‑2 | IL‑7 | IL‑12 | IL‑15 | PD1 | T‑bet | TRAIL All links are active in the online pdf
volume 8 | february 2008 | 119 © 2008 Nature Publishing Group
REVIEWS
Micromanagement of the immune system by microRNAs Harvey F. Lodish*, Beiyan Zhou*, Gwen Liu‡ and Chang-Zheng Chen‡
Abstract | MicroRNAs (miRNAs) are an abundant class of evolutionarily conserved small non-coding RNAs that are thought to control gene expression by targeting mRNAs for degradation or translational repression. Emerging evidence suggests that miRNA-mediated gene regulation represents a fundamental layer of genetic programmes at the posttranscriptional level and has diverse functional roles in animals. Here, we provide an overview of the mechanisms by which miRNAs regulate gene expression, with specific focus on the role of miRNAs in regulating the development of immune cells and in modulating innate and adaptive immune responses. Forward genetics A classical genetic analysis approach that proceeds from phenotype to genotype by positional cloning or candidategene analysis.
*Whitehead Institute for Biomedical Research and Department of Biology, Massachusetts Institute of Technology, Nine Cambridge Center, Cambridge, Massachusetts 02142, USA. ‡ Department of Microbiology and Immunology, Baxter Laboratory of Genetic Pharmacology, Stanford University School of Medicine, 269 Campus Drive, CCSR 3205, Stanford, California 94305‑5175, USA. Correspondence to C.‑Z.C. and H.F.L. e‑mails:
[email protected];
[email protected] doi:10.1038/nri2252 Published online 21 January 2008
Mammals have evolved complex genetic programmes that regulate the development and function of immune cells, and enable the immune system to mount specific responses against invading foreign pathogens while maintaining tolerance to self. Aberrant regulation of the immune system leads to various immune-related pathological disorders, such as autoimmune diseases and leukaemias. Previous studies have revealed a wealth of knowledge about the complex transcriptional programmes that regulate the expression and function of various growth factors, cellsurface receptors, intracellular signalling molecules and transcription factors, which have a central role in maintaining the equilibrium and function of vertebrate immune systems. However, little is known about the role of post-transcriptional regulation of gene expression in the development and function of immune cells. Interestingly, recent discoveries have revealed the existence of potentially widespread regulatory mechanisms that function at a post-transcriptional level and that have crucial roles in animal development. The mediators of these processes are known as microRNAs (miRNAs) — an abundant class of endogenous small non-coding RNAs of approximately 22 nucleotides in length that can regulate gene expression post-transcriptionally by affecting the degradation and translation of target mRNAs (reviewed in REFS 1–3). This Review focuses on the mechanisms by which miRNAs regulate gene expression and provides an overview of the important and diverse roles of individual miRNAs in the vertebrate immune system and in host–viral interactions.
Presence of abundant miRNA genes in animals The discovery of lin‑4 and let‑7, which are the prototype miRNA genes in Caenorhabditis elegans, and their
120 | february 2008 | volume 8
functional characterization led to the suggestion that animal genomes might also contain lin‑4- and let‑7-like non-coding RNA genes that control gene expression at post-transcriptional levels4–8. The lin‑4 and let‑7 genes each produces a mature miRNA of ~22 nucleotides in length that is processed from a stem-loop precursor miRNA (pre-miRNA) of ~60 nucleotides in length. It is thought that lin-4 and let‑7 mature miRNAs can form imperfect Watson–Crick base pairs at multiple sites within the 3′ untranslated region (UTR) of their cognate mRNA targets, the lin‑14 and lin‑41 mRNAs, respectively, to repress their expression at the post-transcriptional level. So far, only a few miRNA genes have been identified using forward genetics approaches, including the lin‑4, let‑7 and lsy‑6 genes in C. elegans, and the Bantam and mir‑14 genes in Drosophila melanogaster4,5,9–11. A large number of endogenous mature lin‑4- and let‑7-like small RNAs were identified through deliberate cloning of small RNAs from multiple animal species or through computational predictions12–20. These computational predictions suggested that the human genome might encode as few as one thousand pre-miRNAs to as many as tens of thousands pre-miRNAs18,19,21. This figure range is much higher than the number of miRNA genes that have been identified to date by cloning20,22,23. Interestingly, some functional miRNA genes (such as lsy‑6) seem to produce extremely low levels of mature miRNA, as indicated by high-throughput sequencing analyses23, which suggests that high levels of mature miRNAs might not be necessary for the efficient function of some miRNA genes. If there are many lsy‑6-like miRNA genes that produce low levels of mature miRNA, but yet still function efficiently, then www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS it is possible that the number of miRNA genes could be on the high side of current estimates. It is also important to note that known miRNAs can be classified into large families that comprise members that encode closely related mature miRNA sequences, often differing only by one or two nucleo tides. For example, in the human genome there are 11 let‑7 miRNA genes that produce 8 types of slightly different mature let‑7 miRNA (FIG. 1a). The product of one of the human let‑7 miRNA genes let-7i regulates Toll-like receptor 4 (TLR4) expression and contributes
to cholangiocyte (human biliary epithelial cells) immune responses against Cryptosporidium parvum infection24. Interestingly, the conservation of miRNA gene family members is often limited to the mature miRNA sequences of ~22-nucleotide in length but not to the sequences flanking the pre-miRNAs and pre-miRNA stem-loops (FIG. 1a,b) 25. These observ ations suggest that some miRNA genes evolved by gene duplication and by varying the pre-miRNA flanking and stem-loop sequences 22,26. Finally, many pre-miRNA stem-loops are clustered in the genome
a Mature let-7 miRNAs
b Pre-let-7 miRNAs
Caenorhabditis elegans
Caenorhabditis elegans G
1 UGAGGUAGUAGGUUGUAUAGUU 22
let-7
G
Homo sapiens let-7a-1 let-7a-2 let-7a-3 let-7b let-7c let-7d let-7e let-7f-1 let-7f-2 let-7g let-7i
Homo sapiens
UGAGGUAGUAGGUUGUAUAGUU UGAGGUAGUAGGUUGUAUAGUU UGAGGUAGUAGGUUGUAUAGUU UGAGGUAGUAGGUUGUGUGGUU UGAGGUAGUAGGUUGUAUGGUU AGAGGUAGUAGGUUGCAUAGUUGAGGUAGGAGGUUGUAUAGUUGAGGUAGUAGAUUGUAUAGUU UGAGGUAGUAGAUUGUAUAGUU UGAGGUAGUAGUUUGUACAGUUGAGGUAGUAGUUUGUGCUGU-
c Primary miRNA transcripts
m7G
--- AAUA U UCCGGUGAGGUAG AGGUUGUAUAGUUU GG U ||||||||||||| |||||||||||||| || AGGCCAUUCCAUC UUUAACGUAUCAAG CC U AG U UGG ACCA GA
UG || AC
let-7
AAAAAn
U GU UUAGGGUCACAC UGGGA GAG AGUAGGUUGUAUAGUU C let-7a-1 ||||| ||| |||||||||||||||| C AUCCU UUC UCAUCUAACAUAUCAA A UG UAGAGGGUCACC UU G U UAGAA UA A AGG GAG UAG AGGUUGUAUAGUU U C U let-7a-2 ||| ||| ||| ||||||||||||| | | C UCC UUC AUC UCCGACAUGUCAA A G A -U G C --UAG GG A U GU --------U GGG GAG AGUAGGUUGUAUAGUU UGGGGC let-7a-3 ||| ||| |||||||||||||||| |||||| C UCC UUC UCAUCUAACAUAUCAA GUCCCG U UG UAGGGUAUC U
let-7b
U UCAGGGCAGUGAUG CGGGG GAGGUAGUAGGUUGUGUGGUU U ||||| ||||||||||||||||||||| GUCCC UUCCGUCAUCCAACAUAUCAA U UAGAAGGCUCCCCG
let-7c
A UU G U UA G UA A GC UCCGGG GAG UAG AGGUUGUAUGGUU GA U C C || |||||| ||| ||| ||||||||||||| || | | C CG AGGUUC UUC AUC UCCAACAUGUCAA UU A G C CU G U -- G GG U
let-7d
A C UUUAGGGCAGGGAUU CCUAGGA GAGGUAGUAGGUUG AUAGU U ||||||| |||||||||||||| ||||| GGAUUCU UUCCGUCGUCCAGC UAUCA U A AUGGAGGAACACCCG
let-7e
C CU G UGAGGAGGAC CC GGG GAG UAGGAGGUUGUAUAGU A C || ||| ||| |||||||||||||||| GG CCC UUC AUCCUCCGGCAUAUCA C A CU G CUAGAGGAAC
let-7f-1
A UG --------U UCAG G AGGUAGUAGAUUGUAUAGUUGU GGGGUAG G |||| | |||||||||||||||||||||| ||||||| A AGUC C UCCGUUAUCUAACAUAUCAAUA UCCCAUU U - CU GAGGACUUG U U GU UUAGGGUCAUAC GUGGGA GAG AGUAGAUUGUAUAGUU C |||||| ||| |||||||||||||||| CACCCU UUC UCAUCUGACAUAUCAA C G UG UAGAGGUUCUAC
U
let-7f-2
m7G
AAAAAn let-7g
U A UGAGG -A A A GGC GAGGUAGU GUUUGUACAGUU GUCU UG UACC C ||| |||||||| |||||||||||| |||| || |||| CCG UUCCGUCA CGGACAUGUCAA UAGA AC AUGG C A C ----GG C
let-7i
U U -------- U UGU UGGC GAGGUAGUAGUUUGUGC GUU GG CGGGU G |||| ||||||||||||||||| ||| || ||||| A AUCG UUCCGUCAUCGAACGCG CAA UC GCCCG C U UAGAGGUG UUA
A
Figure 1 | MicroRNAs (miRNAs) and miRNA genes. a | Sequence alignment of Caenorhabditis elegans and Homo sapiens mature let‑7 miRNA family members. Of interest to this article, it has been recently shown that let‑7i Nature Reviews Immunology regulates Toll-like receptor 4 expression and contributes to cholangiocyte (human biliary epithelial cells)| immune 24 responses against Cryptosporidium parvum infection . It is not yet known if miRNA genes that produce nearly identical mature miRNAs, such as the let‑7 miRNA family members, have the same activity in biological assays. b | Predicted stem-loop structures containing the C. elegans and H. sapiens mature let‑7 miRNAs family members. Mature miRNA sequences are indicated in red. c | Representation of primary miRNA transcripts with one (top) or two (bottom) precursor miRNA(s). nature reviews | immunology
volume 8 | february 2008 | 121 © 2008 Nature Publishing Group
REVIEWS RNA-induced silencing complex (RISC). A multi-protein small interfering RNA (siRNA) complex that binds short antisense RNA strands and guides the cleavage of target RNAs. This complex is thought to be important for posttranscriptional gene regulation by siRNAs and microRNAs.
miRNA gene
(FIG. 1c) and may be co-transcribed and processed into
multiple mature miRNAs25. However, it has not been tested whether miRNA genes that produce nearly identical mature miRNAs, such as the let-7 family of miRNAs (FIG. 1), have the same activity in biological assays.
The biogenesis of mature miRNAs Mature miRNAs are produced from long primary transcripts that contain the pre-miRNA, through a series of endonucleolytic maturation steps 27 (FIG. 2). Some primary miRNA transcripts are made by RNA polymerase II and have 5′ caps and 3′ poly(A) tails28,29, whereas others are transcribed by RNA polymerase III (REF. 30). In the canonical pathway of miRNA biogenesis, long primary miRNA transcripts are processed into
Primary miRNA transcript
Primary miRNA transcription
m7G Drosha DGCR8
AAAAAn
Nucleus RAN–GTP Exportin-5
Pre-miRNA Cytoplasm
Dicer TRBP
miRNA RISC
m7GpppG
Ribosome
Translational repression
AAAAn m7GppppG Target mRNA
AAAAn
Destabilize mRNA target(s) m7GpppG AAAAn
Figure 2 | MicroRNA biogenesis and function in animal cells. Animal genomes have specific genes that encode microRNAs (miRNAs). miRNA genes encode long primary Nature Reviews | Immunology mRNA transcripts that in turn produce mature miRNAs through a series of endonucleolytic maturation steps that are thought to be essential for the production of functional miRNAs. Primary miRNA transcripts are processed into precursor miRNA (pre-miRNA) stem-loops of ~60 nucleotides in length by the nuclear RNase III enzyme Drosha. The pre-miRNA is then actively transported to the cytoplasm by Exportin-5 in a RAS-related nuclear protein–guanosine triphosphate (GTP)-dependent manner and further processed into a ~21-nucleotide duplex. The final step of miRNA maturation is the selective loading of the functional strand of the small RNA duplex onto the RNAinduced silencing complex (RISC). Mature miRNAs then guide the RNA-induced silencing complex to cognate target genes and repress target gene expression by either destabilizing target mRNAs or repressing their translation. DGRC8 (DiGeorge syndrome critical region gene 8) and TRBP (TAR (HIV) RNA binding protein 2) are double-stranded RNA-binding proteins that facilitate mature miRNA biogenesis by Drosha and Dicer RNase III emzymes, respectively33,135–138. 122 | february 2008 | volume 8
corresponding pre-miRNA stem-loops of ~60 nucleotides in length by the nuclear-specific ‘microprocessor’ complex that is comprised of the RNase III enzyme Drosha and its partner DGCR8 (DiGeorge syndrome critical region gene 8; known as Pasha in D. melanogaster)31–34. In D. melanogaster, however, some intronic miRNA precursors, termed ‘Mitrons’, are processed in the nucleus by the usual RNA splicing machinery and not by the Drosha endonuclease35,36. In either case, the premiRNAs are then actively transported to the cytoplasm by Exportin-5 in a RAS-related nuclear protein–guanosine triphosphate (RAN–GTP)-dependent manner and are further processed into ~22-nucleotide duplexes by the cytoplasmic RNase III enzyme Dicer37–42. The functional miRNA strand is then selectively loaded into the RNA-induced silencing complex (RISC)43,44. As indicated by miRNA cloning analyses22, the RISC-loading process is often asymmetric in that a small RNA (~22 nucleo tides) corresponding to only one side of the miRNA stem-loop precursor is preferentially incorporated, whereas the complementary strand (miR*) may be degraded. The information for the sequential processing, maturation and RISC loading of miRNAs is likely to be encoded in the sequences of the primary and premiRNAs. Based on this assumption, when expressing an miRNA for experimental purposes, it is probably necessary to include the corresponding genomic flanking sequences of that miRNA to ensure proper miRNA processing and maturation45–49. Mature miRNA expression may be regulated at various stages of biogenesis. The relative ratio of mature miRNA to pre-miRNA can change depending on the tissue, which indicates that regulation occurs at the post-transcriptional level50,51, although it is unclear what the biological relevance and mechanisms might be for such regulation.
Mechanisms of action of miRNAs The end result of miRNA-mediated gene regulation is clearly a reduction in the total amount of target protein that is produced; however, the fate of most mRNAs that are targeted by miRNAs remains unclear. In some cases, miRNAs have been shown to repress target gene expression at the translational level, with mRNA levels remaining constant and the level of the encoded protein declining, whereas in other cases, miRNAs repress target gene expression by triggering the degradation of target mRNAs (reviewed in Ref. 52). For example, using DNA microarray analysis, Lim et al. found that transfected miRNAs induce the degradation of a large number of mRNAs that contain putative miRNA binding sites53. Nevertheless, it is still unclear how miRNAs mediate target gene repression. However, a number of mechanisms are emerging by which miRNAs can inhibit translation54,55. miRNAs can inhibit either the initiation or the post-initiation (elongation) stages of protein translation. Early studies in C. elegans showed that lin‑4 miRNA represses the lin‑14 and lin‑28 mRNAs at a step following the initiation of mRNA translation56,57. Moreover, polysome profile analyses of human cells showed that endogenous miRNAs are associated with translating polysomes58,59. This analyses www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS
Microarray analysis A technique for measuring the transcription of genes. It involves hybridization of fluorescently labelled cDNA prepared from a cell or tissue of interest with glass slides or other surfaces dotted with thousands of oligonucleotides or cDNA, ideally representing all expressed genes in the species.
Polysome profile Polysomes (or polyribosomes) are a cluster of ribosomes that are attached along the length of a single molecule of mRNA. Polysomes read this mRNA simultaneously, helping to synthesize the same protein at different spots on the mRNA. A polysome profile refers to the distribution of polysomes as determined by gradient centrifugation of cytoplasmic extracts. The method is used to study the association of mRNAs with ribosomes.
m G cap 7
The 7‑methylguanosine that is linked by a triphosphate bridge to the first transcribed nucleotide at the 5′ end of eukaryotic mRNA. Recognition of the m7G cap by the capbinding protein eIF4E is the initiation step of capdependent translation.
Argonaute (AGO) proteins A large family of ~95 KDa proteins that contain conserved PAZ (piwi, argonaut and zwille) and PIWI domains and are involved in post-transcriptional gene silencing. Mammals have four AGO family members (AGO1, AGO2, AGO3 and AGO4), each of which might be a component of an RNAinduced silencing complex.
Seed nucleotides This term refers to the seven nucleotides found at the 5′ region of an miRNA (nucleotides 2–8). Many computational target prediciton programmes require an exact Watson–Crick complementary match between the target sites and the seed nucleotides of a mature miRNA.
suggest that miRNAs may inhibit ribosome movement along mRNAs, indicating that translational repression by miRNAs might be due to the inhibition of translation initiation or the inhibition of elongation or the enhancement of ribosomal drop-off from translating polysomes. Mounting evidence suggests that miRNAs may inhibit target gene expression by blocking translation initiation. Pillai et al. showed that target gene repression by let-7a miRNA occurs at the initiation stage of mRNA translation possibly by interfering with the recognition of the m7G cap by the cap-binding protein60. Similarly, miRNAs were shown to inhibit the initiation of target mRNA translation in in vitro cell-free translational repression assays61–63. Supporting this possibility, Kiriakidou et al. have identified a cap-binding-protein-like motif in human argonaute 2 (AGO2) proteins that is essential for miRNA-mediated target gene repression in human cells; this suggests that AGO2 might inhibit translation initiation by binding to the m7G cap on a target mRNA, thereby impeding the binding of the cap-binding protein eIF4E (eukaryotic translation initiation factor 4E)64. These seemingly conflicting mechanisms of target gene repression by miRNAs may on the one hand reflect the different experimental systems used in various studies — that is, the types of miRNA, the reporters, the cell lines or the cell lysates that were used. On the other hand, these studies suggest the possibility of more than one mechanism by which miRNAs can repress target gene expression.
Regulatory targets of miRNA genes It is evident that target-gene identification holds the key to deciphering the molecular mechanisms by which miRNA genes exert their biological functions in animals. Despite the progress in computational target-gene prediction, finding the functionally relevant targets of an miRNA remains a difficult task. Many of the fundamental assumptions and principles of miRNA and target-gene interactions are derived from the early genetic studies carried out on the lin‑4 and lin‑7 miRNAs and their interactions with cognate target genes. Based on these early analyses, it is thought that miRNAs regulate the expression of target genes by forming base-pair interactions with them, that miRNAs generally form imperfect base pairs with target sites located in the UTRs of their target mRNAs, that multiple miRNA binding sites in target UTRs are required for efficient regulation, and that these sites are evolutionarily conserved4–6,56. These principles provide the experimental foundation that has been used for the computational identification of miRNA target genes. Several groups have postulated that miRNA target sites comprise a ‘core sequence’ that forms perfect or near-perfect base pairs with seven or eight bases near the 5′ end of an miRNA (known as the seed nucleotides)65–68, and independently developed computational algorithms to predict miRNA target genes on the basis of this principle (reviewed in Ref. 69). It is estimated that 30% to 92% of human genes are regulated by miRNAs19,70. Recently, computational programmes have been developed for target-gene identification without recourse to have revealed several general principles regarding miRNAs
nature reviews | immunology
and target-gene regulation: first, each miRNA can potentially regulate a large number of protein-coding genes; second, many miRNAs potentially act in combination to regulate the same target gene(s); and third, predicted miRNA target genes are not restricted to a particular functional category or biological pathway, but rather are involved in a wide variety of biological processes (reviewed in Ref. 69). The finding that a majority of protein-coding genes in D. melanogaster and mammals are potentially regulated by miRNAs is astonishing19,70,72–75 and suggests that miRNA-mediated gene expression may have a widespread influence on the protein composition of animal cells2,53,76,77. However, available miRNA target-gene prediction algorithms are likely to have high false-positive and false-negative rates. There is limited overlap between the target genes that have been predicted by various algorithms69, owing in part to our limited understanding of the principles that govern miRNA-mediated target-gene regulation. The general principles that have been derived from the few genetically validated miRNA targets are unlikely to be all-encompassing. In fact, among the small group of genetically validated target mRNAs there are clear exceptions to these computational principles. Target mRNAs are repressed as efficiently by miRNA-binding sites in the 5′ UTR as by those in the 3′ UTR78, which suggests that miRNAs could repress target genes through binding sites other than those within the 3′ UTR. However, to date, computational programmes have not been developed to search for sites other than those within the 3′ UTRs. Moreover, only a single, canonical miRNA target site is required for the regulation of the cog‑1 and hairy mRNAs by the lys-6 and mir-7 miRNAs, respectively, in C. elegans9,67,79. Interestingly, perfect pairing of the seed nucleotides is not required for cog‑1 regulation by the lys-6 miRNA80. Moreover, the RNA secondary structure and RNA sequences outside of the mature miRNA binding regions have potent effects on target recognition 81–83, indicating a previously unappreciated complexity of miRNA interactions with their target mRNAs. Finally, computational predictions can only suggest potential physical interactions between miRNAs and mRNA targets, and thus many target genes predicted to date may not be biologically relevant as specific miRNAs and their predicted target genes may never be expressed in the same cell.
miRNAs and lineage differentiation As a fundamental layer of post-transcriptional gene regulation, it is not surprising that miRNA genes have diverse and crucial roles in mammals (reviewed in Ref. 3). By inactivating the essential protein components of the miRNA and small interfering RNA (siRNA) machinery, such as Dicer and AGO2, researchers have revealed that this system has a crucial role in the development of worms, flies, fish and mice (reviewed in Ref. 84). More importantly, the targeted deletions of individual miRNA genes in mice have indicated their essential role in the development and function of the volume 8 | february 2008 | 123
© 2008 Nature Publishing Group
REVIEWS Small interfering RNAs (siRNAs). A class of doublestranded RNAs (dsRNAs) of ~21 nucleotides in length, generated from long dsRNAs. siRNAs silence gene expression by promoting the cleavage of perfectly matched mRNAs. siRNAs can also be generated by in vitro synthesis and can be used to ‘knockdown’ (that is, to silence the expression of) a specific gene.
Chromatin immunoprecipitation (ChIP). The use of antibodies specific for transcription factors to precipitate nucleic-acid sequences from chromatin for amplification.
cardiac and immune systems85–88. Altered expression levels of miRNAs have been reported in various haemato poietic disorders, such as leukaemias and lymphomas. In fact, distinct signatures of a group of miRNAs are correlated with specific types of leukaemias89,90. The importance of miRNA genes in haematopoiesis is further supported by studies showing that dysregulation of miRNA gene expression in haematopoietic-cell lineages seems to contribute to cancer pathogenesis (reviewed in REFS 91,92). Using either miRNA cloning or microarray analyses, researchers have identified groups of miRNAs that are differentially or abundantly expressed by specific haematopoietic tissues1,20,93–96. Neilson and colleagues have characterized miRNA expression profiles in T cells at various stages of their development by direct miRNA cloning and sequencing. They have shown that in addition to changes in the levels of individual miRNAs between distinct T‑cell developmental stages, the global miRNA levels vary dramatically in parallel with changes in the translational capacity of the cell96. More interestingly, miRNA profiling in naive, effector and memory CD8+ T cells has revealed that a few highly expressed miRNAs are dynamically regulated during antigenspecific T‑cell differentiation97. Finally, genome-wide chromatin immunoprecipitation (ChIP) analyses have revealed that the expression of forkhead box P3 (FOXP3), a transcription factor that is required for the development
and function of regulatory T cells, may directly control the expression of the miRNA miR‑155, implicating miR‑155 in regulatory T‑cell formation or function98. These studies on the expression of miRNAs, which show a dynamic expression pattern relative to the various stages of development, set a solid foundation for further examination of the role of miRNAs in T‑cell development in the thymus and peripheral lymphoid organs. Many miRNAs that are also differentially regulated in haematopoietic-cell lineages have important roles in modulating the development and function of immune cells and host–pathogen interactions (TABLE 1). One of the first miRNAs that was shown to have a role in the development of vertebrate immune cells was miR‑181a; this miRNA is highly expressed by cells in the thymus and is expressed at lower levels by cells in the heart, lymph nodes and bone marrow45,99,100. In the bone marrow, miR‑181a is expressed at higher levels by B220+ B cells than by CD3+ T cells45. Specifically, miR‑181a expression in bone-marrow-derived B cells decreases during B‑cell maturation from the pro‑B-cell to pre‑B‑cell stage of development100. In addition, ectopic expression of miR‑181a in enriched haematopoietic stem and progenitor cells (HSPCs) resulted in an increase in the percentage of CD19+ B cells and a decrease in the percentage of CD8+ T cells in short-term mouse bone-marrow reconstitution assays45, demonstrating that lineage-specific miRNAs might have a role in regulating lymphocyte development.
Table 1 | RNAs that have an important role in the development and function of immune cells or in host–pathogen interactions miRNA
Function
Validated immune targets
Refs
Host miRNA miR‑32
Limits PFV1 replication by targeting the viral genome
PFV1 and ORF2
miR‑155
Required for T‑cell differentiation, germinal centre B‑cell responses and responses to bacterial and viral infection
MAF
86, 87
116
miR‑150
Regulates B‑cell development and T‑cell activation by targeting transcription factors
Myb
100, 104
miR‑181a
Regulates B‑cell and T‑cell development and modulates T‑cell sensitivity to antigens by controlling the expression of multiple phosphatases in the TCR signalling pathway
DUSP5, DUSP6, SHP2, PTPN22, BCL‑2 and CD69
45, 99
miR‑146
Response to bacterial infection as part of TLR–NF-κB signalling
IRAK1 and TRAF6
114
miR‑132
Response to bacterial infection and involved in CREB signalling
ND
114
miR‑122
Facilitates the replication of HCV through interactions at the 5′ UTR of the viral RNA
HCV RNA
118
miR-BART2
An EBV-encoded miRNA that is up-regulated during the lytic stage, targeting EBV BALF5
EBV BALF5
122
SV40
Accumulates at late stages of infection, targeting early viral mRNAs and reduces susceptibility to CTLs
Viral large and small T antigens
124
miR-LAT
An HSV1-encoded miRNA that inhibits host-cell apoptosis through reducing TGFβ and SMAD3 expression to maintain viral latency
TGFβ and SMAD3
126
miR-UL112
An HCMV-encoded miRNA that interferes with NK‑cell function by targeting the host gene MICB.
MICB
127
Viral miRNA
BALF5, BamHI L fragment 5; BCL-2, B‑cell lymphoma 2; CREB, cAMP-responsive-element-binding protein; CTL, cytotoxic T lymphocyte; DUSP, dual-specificity protein phosphatase; EBV, Epstein–Barr virus; HCMV, human cytomegalovirus; HCV, hepatitis C virus; HSV1, herpes simplex virus‑1; IRAK, interleukin‑1 receptorassociated kinase; MICB, MHC-class‑I-polypeptide-related sequence B; miRNA, microRNA; ND, not determined; NF‑κB, nuclear factor-κB; NK, natural killer; ORF2, open reading frame 2; PFV1, primate foamy virus type 1; PTPN22, protein tyrosine phosphatase, non-receptor type 22; SHP2, SH2-domain-containing protein tyrosine phosphatase 2; SMAD3, mothers against decapentaplegic homologue 3; TCR, T‑cell receptor; TGFβ, transforming growth factor‑β; TLR, Toll-like receptor; TRAF, TNF receptor-associated factor; UTR, untranslated region.
124 | february 2008 | volume 8
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS
B-1 cells IgMhiIgDlowMac1+B220lowCD23cells that are dominant in the peritoneal and pleural cavities. Their precursors develop in the fetal liver and omentum, and in adult mice, the size of the B-1cell population is kept constant owing to the self-renewing capacity of these cells. B-1 cells recognize self components, as well as common bacterial antigens, and they secrete antibodies that tend to have low affinity and broad specificity.
Interestingly, miR‑181a, at least when considered alone rather than in combination with other miRNAs, appears to function as a lymphoid-lineage modulator rather than as a developmental switch, as the increase in B-lineage cells did not completely block the differentiation of lymphoid and myeloid cell types. This may differ from the effects exerted by lineage-specific transcription factors or oncogenes, which, when ectopically expressed, can completely shut down the differentiation of one cell lineage101. Not surprisingly, other miRNAs with distinct expression profiles in haematopoietic or lymphoid cell types were shown to have distinct roles in immune-cell development. For example, miR‑223, a myeloid-specific miRNA, the expression of which might be controlled by the myeloid-specific transcription factors PU.1 and members of the C/EBP (CCAAT/enhancer-binding protein) family, seems to have an important role in regulating granulopoiesis45,102,103. Recent studies also revealed that miR-150, an miRNA that is specifically expressed by mature lymphocytes, has a key role in B‑cell differentiation93,100,104. In contrast to miR‑181a, the expression of miR‑150 increases during B‑cell maturation in the bone marrow and T‑cell maturation in the thymus, but decreases rapidly when naive T cells differentiate into T helper 1 (TH1) or TH2 cells93,100. When ectopically expressed in HSPCs, miR‑150 blocks B‑cell development at the transition from the pro‑B-cell to pre‑B-cell developmental stage, leading to severe defects in the production of mature B cells. Xiao and colleagues further demonstrated the importance of miR‑150 in B‑cell formation using gain- and loss-of-function mouse models104. miR-150-knockout mice have an approximately twofold increase in the number of splenic B-1 cells, but have no apparent defect in the development of other lymphoid-derived T‑ and B‑cell types. A slight delay in T‑cell development was observed in mice that express an miR-150 transgene early in life, but, by 18 weeks of age, they had dramatically impaired B‑cell development with normal T‑cell levels, which is consistent with the phenotype observed by Zhou and colleagues100,104. According to target validation using a luciferase reporter assay and the correlation between Myb and miR‑150 expression levels in wild-type mice compared with miR‑150-deficient mice, Myb might be a crucial target of miR‑150. Intriguingly, although it is known that Myb has a key role in both B‑ and T‑cell development, overexpression or deletion of miR‑150 in mice affects only the development of B cells, but not T cells104. It is also of interest that deletion of miR‑150 in mice results in a significant increase in the production of ‘natural’ antibodies in the blood, and that miR‑150 expression in B cells is often rapidly downregulated following B‑cell activation with IgM-specific antibodies, CpG-containing DNA or lipopolysaccharide104. These findings suggest that miR‑150 has a role in regulating B‑cell development but might also have a role in adaptive or innate immune responses that is yet to be fully explored. However, it is clear that miRNAs are components of the molecular circuitry that controls lymphocyte development, but that the characterization of the role of miRNAs in the development of immune cells is still in its infancy.
nature reviews | immunology
miRNAs and adaptive immune responses In addition to regulating haematopoietic-cell lineage differentiation, miRNAs also have an important role in modulating adaptive immune responses in mice86,87,99. The bic locus, a common retroviral integration site in avian leukosis virus (ALV)-induced B‑cell lymphomas, encodes the bic non-coding RNA that may collaborate with the oncogene Myc to control cell growth105. More recently it was found that the bic non-coding RNA is the primary transcript for miR‑155 and therefore is now referred to as bic/mir-155 (Refs 106,107). This non-coding RNA transcript and the mature miR‑155 contained within it are overexpressed in human B‑cell lymphomas, including diffuse large B‑cell lymphomas, Hodgkin lymphomas, and Burkitt lymphomas107–109. Mice with a bic/mir‑155 transgene, the expression of which is targeted to B cells, develop B‑cell malignancy110. Interestingly, mice deficient in the bic/mir‑155 gene are viable and fertile but have profound defects in their protective immune responses86,87. Mice carrying mutated bic/mir‑155 alleles are less responsive to immunizations and are not protected from virulent Salmonella typhimurium infection after immunization with a nonvirulent aroA mutant strain of the bacteria. Using both gain-of-function and loss-of-function analyses, Thai et al. elegantly demonstrated that miR‑155 might control the formation and response of germinal-centre B cells in part by controlling cytokine production87. Consistent with the role of miR‑155 in B‑ and T‑cell responses, the expression of miR‑155 is upregulated in B and T cells following their activation, and the deletion of the bic/mir‑155 gene in mice leads to pleiotropic defects in the function of B cells, T cells and dendritic cells86. Significantly, in these miR-155-deficient mice, less IgMswitched antigen-specific antibodies are produced by activated B cells compared with control cells and the weak production of interleukin‑2 (IL‑2), IL‑4 and interferon‑γ (IFNγ) by activated T cells that is observed in response to immunization indicates impaired B‑ and T‑cell-mediated immune responses, possibly due to a biased differentiation toward TH2 cells compared with TH1 cells86,87. These studies establish a crucial role for miR‑155 in the adaptive immune response. Further characterization of the function of miR‑155 function in various immune cell populations and the identification of functionally relevant target genes will help elucidate the molecular and cellular mechanisms by which miR‑155 regulates B‑ and T‑cell responses and protective immunity. In another study, Li et al. demonstrated that miR‑181a can function to modulate the strength and threshold of T‑cell receptor (TCR) signalling, thereby influencing T‑cell sensitivity to antigens99 (FIG. 3). In this sense miR‑181a acts like a rheostat, or dimmer switch, in that it modulates cell pathways rather than simply turning them on or off. Ectopic miR‑181a expression in mature T cells augments their sensitivity to peptide antigens, whereas inhibition of miR‑181a in immature T cells reduces sensitivity and impairs both positive and negative selection. Moreover, the quantitative regulation of TCR signalling strength by miR‑181a can lead to a change in the TCR signalling threshold and can enable volume 8 | february 2008 | 125
© 2008 Nature Publishing Group
REVIEWS mature T cells to recognize antagonists as agonists, demonstrating the intricate relationship between quantitative regulation and the switch-like response in immune cells. More importantly, miR‑181a controls the TCR signalling threshold and strength in part by simultaneously dampening the expression of multiple phosphatases that are negative regulators of distinct steps of the TCR signalling cascade, such as SHP2 (SH2 (SRC homology 2)-domaincontaining protein tyrosine phosphatase 2), PTPN22 (protein tyrosine phosphatase, non-receptor type 22), DUSP5 (dual-specificity protein phosphatase 5) and DUSP6. Multi-target regulation by miR‑181a is required for fine-tuning T‑cell sensitivity, as knocking down individual phosphatases by specific short-hairpin RNAs a Low miR-181a expression in mature effector T cells
Peptide–MHC complexes
DUSP6
DUSP5
PTPN22
SHP2
Expression level
Signal input
miR-181a Phosphatases or other negative regulators of TCR signalling
10 High resistance 5
0
Low resistance
Signal output Reduced or inhibited proliferation and cytokine production
b High miR-181a expression in immature T cells
Peptide–MHC complexes
DUSP6
DUSP5
PTPN22
10 High resistance SHP2
miR-181a
Expression level
Signal input
Phosphatases or other negative regulators of TCR signalling
5
0
Low resistance
Signal output Enhanced proliferation and cytokine production
Figure 3 | Modulation of the antigen sensitivity of T cells by the miR‑181a dimmer Nature(miR‑181a) Reviews | Immunology switch (or rheostat). a | In mature effector T cells, microRNA‑181a is expressed at low levels and therefore the signalling rheostat is tuned to high resistance. That is, by decreasing miR‑181a expression and de-repressing the negative signals that are controlled by the negative regulators of T‑cell receptor (TCR) signalling (such as SH2 (SRC homology 2)-domain-containing protein tyrosine phosphatase 2 (SHP2), protein tyrosine phosphatase, non-receptor type 22 (PTPN22), dual-specificity protein phosphatase 5 (DUSP5) and DUSP6) the signal output, such as T‑cell proliferation and cytokine secretion induced by peptide–MHC complexes, is dramatically reduced or completely turned off. b | By contrast, in immature T cells, such as double-positive thymocytes, miR‑181a is expressed at high levels and thus the rheostat is tuned to low resistance. That is, by increasing miRNA expression and repressing the negative signals that are controlled by the negative regulators of TCR signalling, the signal output from the identical stimulation is dramatically enhanced. The expression of miR‑181a is regulated during T‑cell development and maturation, and the level of miR‑181a expression correlates with T‑cell sensitivity to antigens, suggesting that miR‑181a might act as an intrinsic rheostat or dimmer switch to tune T‑cell sensitivity to antigens during T‑cell development and maturation. Other miRNA ‘rheostats’ might be used in different cell types and for different receptors when the miRNA and corresponding targets coexist. 126 | february 2008 | volume 8
(shRNAs) cannot recapitulate the effects of miR‑181a overexpression, whereas restoring the expression of single phosphatases does reduce and/or abrogate the effects of miR‑181a overexpression. These findings reveal that the T‑cell threshold and sensitivity to antigens is controlled by a network of genes at distinct stages of TCR signalling, and provide solid evidence as to how multi-target regulation by miRNA genes is used to carry out such a task. In addition, the dynamic regulation of miR‑181a expression during T‑cell development and maturation seems to correlate with changes in T‑cell sensitivity to antigens99,111. Inhibition of miR‑181a expression in immature T cells reduces antigen sensitivity and impairs both positive and negative selection. The sensitivity of T cells to antigen is intrinsically regulated to ensure the proper development of T‑cell specificity and sensitivity to foreign antigens while avoiding self recognition. In immature CD4+CD8+ double-positive thymocytes, low-affinity antigenic peptides that are unable to activate mature effector T cells are sufficient to induce strong activation and clonal deletion112, whereas antagonists that are normally inhibitory to effector T cells can induce positive selection113. However, little is known about the intrinsic molecular programmes that control this. Therefore, miR‑181a might in part contribute to the fine tuning and intrinsic regulation of T‑cell sensitivity to antigens, thus, having a central role in the development and maintenance of tolerance and immunity. Of note, miR‑181a, which is dynamically regulated during T‑cell and B‑cell maturation, may also have an important role in other immune cell types96,99,100. As many of the phosphatase targets regulated by miR‑181a are also expressed in distinct T‑ and B‑cell subsets, one intriguing possibility is that miR‑181a could modulate different receptor signalling pathways by controlling the expression of similar target phosphatase genes (FIG. 3). Indeed, we have found that miR‑181a promotes early T‑cell development in the thymus through potentiating pre-TCR and Notch signalling (T. K. Mao and C.‑Z.C., unpublished observations). However, miR‑181a is likely to have other roles in T‑cell function. The ectopic expression of miR‑181a influences the co-stimulatory pathway through presently unknown targets99. miR‑181a may also regulate the expression of anti-apoptotic proteins, such as B‑cell lymphoma 2 (BCL‑2) and the cell surface regulator CD69 (Ref. 96). These findings suggest that further characterization of the molecular networks controlled by miR‑181a would probably yield additional insights about other possible functions for miR‑181a in adaptive immune responses.
miRNAs and innate immune responses miRNAs may also have crucial roles in regulating the innate immune response, the first line of defence that relies on phagocytes, such as granulocytes and macrophages. TLRs have a major role in the recognition of invading bacteria, viruses and other pathogens, and in the initiation of the innate immune response. Triggered by ligand binding, TLR-mediated signalling starts with the recruitment of adaptor proteins to the receptor, followed by stimulation www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS of a protein kinase cascade that consequently activates transcription factors, such as activator protein 1 (AP1) and nuclear factor-κB (NF-κB), and ultimately resulting in immune gene expression. Recently, three miRNAs, miR‑146a, miR‑132 and miR‑155, were found to be regulated in response to immune-cell stimulation by endotoxins114. When monocytes were treated with the endotoxin lipopolysaccharide (LPS) to mimic signalling by a bacterial infection, the expression levels of these three miRNAs were dramatically increased. Interestingly, induction of miR‑146a expression can only be triggered by the TLRs that reside on the cell surface and recognize bacterial constituents, but not by the intracellular TLRs that mainly sense viral nucleic acids, suggesting that miR‑146a may respond to bacterial invasion rather than viral infection. In addition, the inducible expression of miR‑155, an miRNA that (as discussed before) is important for B‑cell function and is associated with B‑cell developmental disorders, was observed in both bacterial and viral infection events as a result of TLR3-induced tumour-necrosis factor (TNF) autocrine signalling114,115. These findings clearly indicate that miRNAs may be integral components of innate immune responses, and that further characterization of the role of miRNAs will probably reveal novel insights into the molecular mechanisms by which miRNAs regulate innate immune responses. Emerging evidence also suggests that miRNAmediated gene regulation may serve as a defence mechanism against viral infections in vertebrate cells, thereby
Primary miRNA transcript m7G
Drosha DGRC8
AAAAAn
Pre-miRNA
RISC–miR-32 complex
Tas
5' UTR
Tas
3' UTR
RAN–GTP Exportin 5
Dicer TRBP
RISC–miR-122 complex 5' UTR HCV RNA
PFV1 RNA PFV1 replication
HCV replication
Figure 4 | MicroRNAs in host–pathogen interaction. Host microRNAs (miRNAs) may interfere with viral infections by interacting with viral genomes. In human cells, after infection of the retrovirus primate foamy virus type 1 (PFV1), host cell miR‑32 binds to Nature Reviews | Immunology partial complementary sites in the viral 3′ untranslated region (UTR) that is shared by five PFV1 mRNAs, leading to reduced viral replication. To repress host miRNA machinery, PFV1 produces Tas, a virus-encoded RNA interference suppressor protein that inhibits host miRNA function by targeting RNA-interfering silencing complex (RISC), to counteract the host antiviral response. By contrast, hepatitis C virus (HCV) can exploit miRNA-mediated gene regulation to facilitate its replication by recruiting a liver-cell specific miRNA, miR‑122, to bind the viral RNA at the 5′ UTR. DGRC8, DiGeorge syndrome critical region gene 8; GTP, guanosine triphosphate; TRBP, TAR (HIV) RNA binding protein 2. nature reviews | immunology
providing another layer to the innate immune response. Certain host miRNAs appear to have evolved to regulate viral infection. Human miR‑32 contributes to the repression of the replication of the retrovirus primate foamy virus type 1 (PFV1) in cultured human cells by binding to partially complementary sites in the 3′ UTRs of five different PFV1-derived mRNAs116. The downregulation of these five viral target genes by miR‑32 slows PFV1 replication. This study identified a human cellular miRNA that has an antiviral function, and suggests a possibly broad impact of miRNA-mediated gene regulation on viral infection. Further supporting this idea, Pedersen et al. have recently shown that the IFNsignalling system, the key defence mechanism against viral infection in mammalian cells, works in concert with miRNAs to control viral infection117. They have shown that IFNβ can induce the expression of numerous cellular miRNAs. Specifically, eight of the IFNβ-induced miRNAs (miR‑1, miR‑30, miR‑128, miR‑196, miR‑296, miR‑351, miR‑431 and miR‑448) have seed nucleotides that form near-perfect base pair matches with the HCV genome and may contribute to the antiviral effects of IFNβ against HCV. Interestingly, IFNβ also induces the downregulation of the expression of miR‑122, an miRNA that has been shown to be essential for HCV replication in liver cells118. These findings provide clear evidence that cellular miRNAs are integrated components of the mammalian innate immune response, and present another dimension to innate immunity that is based on direct interactions between viral and host-encoded nucleic acids. Not surprisingly, such a gene regulatory mechanism may also be exploited by viruses to facilitate their infection (FIG. 4)119–121. Many viruses have been found to encode miRNAs that regulate both viral and host mRNAs (reviewed in ref. 121). Viruses that encode such miRNAs include the Epstein–Barr virus (EBV), Kaposi’s sarcoma-associated herpes virus (KSHV) and cytomegalovirus (CMV) 122–125. In cells latently infected with EBV, different miRNAs are expressed during different stages, suggesting that viral miRNAs contribute to the regulation and maintenance of viral latency122,123. miR-LAT is an miRNA encoded by herpes simplex virus‑1 (HSV1) that maintains host cell latency and inhibits cell apoptosis by reducing transforming growth factor‑β (TGFβ) and SMAD3 (mothers against decapentaplegic homologue 3) expression in the host cell, thereby interfering with TGFβ-dependent signalling and preventing host cell death126. Finally, a human cytomegalovirus (HCMV)-encoded miRNA, miR-UL112, represses the expression of MHC-class‑Ipolypeptide-related sequence B (MICB). MICB is a stress-induced ligand of the natural-killer cell (NK cell)-activating receptor NKG2D (natural-killer group 2, member D), which is required for NK‑cellmediated killing of virus-infected cells127. These results demonstrate that HCMV evades immune surveillance by targeting a cellular mRNA with a virally encoded miRNA, and suggest that viruses use miRNAs not only to regulate their own life cycles but also to evade host immune surveillance. volume 8 | february 2008 | 127
© 2008 Nature Publishing Group
REVIEWS RNA interference A mechanism for RNA-guided regulation of gene silencing in which double-stranded RNA inhibits the expression of genes with complementary nucleotide sequences.
In an interesting twist, albeit with an opposite outcome, miR‑122, which is highly abundant and specifically expressed by liver cells, actually facilitates the replication of the hepatitis C virus (HCV) through interactions at the 5′ UTR of the HCV RNA118. These results demonstrate that HCV has evolved to exploit miRNAmediated gene regulation to facilitate its replication through an as yet unknown mechanism. As noted above, animal miRNAs have so far only been shown to act at the post-transcriptional level to repress gene expression. It is interesting to note that most viral miRNAs identified so far have no substantial homology to one another or to any known animal miRNAs. Moreover, miRNAs are only found in DNA viruses and not in RNA viruses or retroviruses123. It is also intriguing that no viral-encoded siRNAs have been identified in virusinfected cells122,123. In addition to encoding miRNAs that regulate their own genes and host genes, certain viruses encode silencing suppressor proteins that counteract miRNA or siRNA-mediated immunity. It is thought that PFV1 has such a mechanism to suppress miR‑32 function to allow it to successfully infect cells. Indeed PFV1 encodes the silencing suppressor Tas that can interfere with the miR‑32-mediated downregulation of its mRNA in a nonspecific manner 116. Similarly, HIV‑1 uses Tat, one of its transcriptional activators, as an miRNA-silencing suppressor that interferes with Dicer function and prevents the processing of doublestranded RNA (dsRNA) into siRNA128,129. Interestingly, a Tat-deficient HIV‑1 strain does not spread effectively in human cells, perhaps due in part to its inability to suppress RNA interference in the host cell, although this hypothesis needs to be further investigated. These studies depict yet another defensive factor that viruses use to evade host miRNA-based immunity and optimize their ability to infect and replicate. Collectively, the discovery that both viruses and hosts use miRNAs for their own advantages introduces a new level of gene regulation that modulates pathogen–host interactions.
Perspectives These studies provide initial evidence that many haematopoietic-cell-specific or lineage-specific miRNAs have a key role in regulating the development of different types of immune cells. However, much remains to be learned before we will be able to integrate the posttranscriptional genetic programmes that are controlled by miRNA genes into the genetic circuitry that maintains the homeostasis of the immune system. Many growth factors and transcription factors have key roles in regulating cell-lineage determination, cell proliferation and differentiation, and selection checkpoints during lymphopoiesis130,131. However, the quantitative regulation,
1.
2.
Bartel, D. P. MicroRNAs: genomics, biogenesis, mechanism, and function. Cell 116, 281–297 (2004). A comprehensive review that describes the genomics, biogenesis and mechanism of action of miRNAs. Bartel, D. P. & Chen, C. Z. Micromanagers of gene expression: the potentially widespread influence of
3. 4.
which determines the size of the stem-cell or progenitorcell pools, cell-cycle progression, the rate of cell division, and the kinetics and timing and/or order of lineage differentiation, remains largely unknown. Such dynamic and finely tuned processes are likely to be regulated by quantitative gene regulatory mechanisms. Compared with other gene regulatory mechanisms, such as chromatin modification and transcriptional controls, miRNA-mediated gene regulation occurs at the step directly before protein synthesis, and thus may be more suited for the fine-tuning of gene expression and quantitative regulation2. More generally, miRNAs provide a mechanism for managing gene expression, allowing mRNAs to be translated at one stage of differentiation and then shut off at a slightly later stage. It has been proposed that, at the molecular level, certain miRNA genes function as rheostats by precisely regulating the quantitative levels of protein synthesis and thus many other biological processes. The degree of gene repression achieved by an miRNA may be quantitatively dependent on the number of target sites on the target UTR, the degree of pairing to the miRNA, the levels of miRNA expressed, the presence of other miRNAs that regulate the same target gene and the cleavage of the target RNAs2. Furthermore, these early studies clearly demonstrate that miRNAs have a central role in modulating immune responses. As a fundamental type of genetic programme, it is likely that miRNA-mediated post-transcriptional gene expression has important roles in many aspects of adaptive and innate immune responses. Discovering the biological functions of miRNA genes and the molecular networks controlled by these non-coding RNA genes would probably fill some of the gaps in our existing knowledge and possibly reveal novel concepts about the regulation of the immune system. Given the crucial role that miRNA genes could have in modulating immune responses, it is likely that any dysregulation of miRNA expression may contribute to the pathogenesis of certain autoimmune diseases132–134. The complete impact of miRNAs on the immune system has yet to be fully understood and many interesting questions remain to be answered before we can unravel all of the functions of miRNAs and integrate miRNAmediated gene regulation with other cellular genetic regulatory networks. Further characterization of miRNAmediated gene regulation during haematopoiesis should shed light on the homeostatic control of haematopoietic stem-cell self-renewal and lineage differentiation, and also provide insights as to how aberrant miRNA regulation may contribute to autoimmune diseases, leukaemias and other blood-related pathological disorders. Finally, one could envision in the future the development of novel therapeutics to target miRNAs to modulate immune responses.
metazoan microRNAs. Nature Rev. Genet. 5, 396–400 (2004). Bushati, N. & Cohen, S. M. microRNA functions. Annu. Rev. Cell Dev. Biol. 23, 175–205 (2007). Lee, R. C., Feinbaum, R. L. & Ambros, V. The C. elegans heterochronic gene lin‑4 encodes small RNAs with antisense complementarity to lin‑14. Cell 75, 843–854 (1993).
128 | february 2008 | volume 8
5.
Reference 4 is a hallmark paper that describes the identification of the lin‑4 miRNA gene, the first example of an miRNA gene, and shows that it has a crucial role in regulating the timing of lineage differentiation in C. elegans. Wightman, B., Ha, I. & Ruvkun, G. Posttranscriptional regulation of the heterochronic gene lin‑14 by lin‑4 mediates temporal pattern
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS
6. 7.
8.
9. 10.
11.
12.
13. 14. 15. 16. 17. 18. 19.
20. 21. 22. 23. 24.
25. 26.
27. 28. 29.
30. 31. 32.
formation in C. elegans. Cell 75, 855–862 (1993). Reference 5 is a hallmark paper showing that the heterochronic gene lin‑14 is regulated by lin‑4 miRNA at the post-transcriptional level in C. elegans. Reinhart, B. J. et al. The 21 nucleotide let‑7 RNA regulates developmental timing in Caenorhabditis elegans. Nature 403, 901–906 (2000). Slack, F. J. et al. The lin‑41 RBCC gene acts in the C. elegans heterochronic pathway between the let‑7 regulatory RNA and the LIN‑29 transcription factor. Mol. Cell 5, 659–669 (2000). Pasquinelli, A. E. et al. Conservation across animal phylogeny of the sequence and temporal regulation of the 21 nucleotide let‑7 heterochronic regulatory RNA. Nature 408, 86–89 (2000). Johnston, R. J. & Hobert, O. A microRNA controlling left/right neuronal asymmetry in Caenorhabditis elegans. Nature 426, 845–849 (2003). Brennecke, J., Hipfner, D. R., Stark, A., Russell, R. B. & Cohen, S. M. bantam encodes a developmentally regulated microRNA that controls cell proliferation and regulates the proapoptotic gene hid in Drosophila. Cell 113, 25–36 (2003). Xu, P., Vernooy, S. Y., Guo, M. & Hay, B. A. The Drosophila microRNA Mir‑14 suppresses cell death and is required for normal fat metabolism. Curr. Biol. 13, 790–795 (2003). Lau, N. C., Lim, L. P., Weinstein, E. G. & Bartel, D. P. An abundant class of tiny RNAs with probable regulatory roles in Caenorhabditis elegans. Science 294, 858–862 (2001). Lee, R. C. & Ambros, V. An extensive class of small RNAs in Caenorhabditis elegans. Science 294, 862–864 (2001). Lagos-Quintana, M., Rauhut, R., Lendeckel, W. & Tuschl, T. Identification of novel genes coding for small expressed RNAs. Science 294, 853–858 (2001). Lagos-Quintana, M. et al. Identification of tissuespecific microRNAs from mouse. Curr. Biol. 12, 735–739 (2002). Lim, L. P., Glasner, M. E., Yekta, S., Burge, C. B. & Bartel, D. P. Vertebrate microRNA genes. Science 299, 1540 (2003). Lim, L. P. et al. The microRNAs of Caenorhabditis elegans. Genes Dev.17, 991–1008 (2003). Bentwich, I. et al. Identification of hundreds of conserved and nonconserved human microRNAs. Nature Genet. 35, 766–770 (2005). Miranda, K. C. et al. A pattern-based method for the identification of microRNA binding sites and their corresponding heteroduplexes. Cell 126, 1203–1217 (2006). Landgraf, P. et al. A mammalian microRNA expression atlas based on small RNA library sequencing. Cell 129, 1401–1414 (2007). Berezikov, E. et al. Phylogenetic shadowing and computational identification of human microRNA genes. Cell 120, 21–24 (2005). Griffiths-Jones, S. The microRNA registry. Nucleic Acids Res. 32, D109–D111 (2004). Ruby, J. G. et al. Large-scale sequencing reveals 21URNAs and additional microRNAs and endogenous siRNAs in C. elegans. Cell 127, 1193–1207 (2006). Chen, X. M., Splinter, P. L., O’Hara S, P. & Larusso, N. F. A cellular micro-RNA, let-7i, regulates Toll-like receptor 4 expression and contributes to cholangiocyte immune responses against Cryptosporidium parvum infection. J. Biol. Chem. 282, 28929–28938 (2007). Altuvia, Y. et al. Clustering and conservation patterns of human microRNAs. Nucleic Acids Res. 33, 2697–2706 (2005). Megraw, M., Sethupathy, P., Corda, B. & Hatzigeorgiou, A. G. miRGen: a database for the study of animal microRNA genomic organization and function. Nucleic Acids Res. 35, D149–D155 (2007). Kim, V. N. MicroRNA biogenesis: coordinated cropping and dicing. Nature Rev. Mol. Cell Biol. 6, 376–385 (2005). Lee, Y. et al. MicroRNA genes are transcribed by RNA polymerase II. EMBO J. 23, 4051–4060 (2004). Cai, X., Hagedorn, C. H. & Cullen, B. R. Human microRNAs are processed from capped, polyadenylated transcripts that can also function as mRNAs. RNA 10, 1957–1966 (2004). Borchert, G. M., Lanier, W. & Davidson, B. L. RNA polymerase III transcribes human microRNAs. Nature Struct. Mol. Biol. 13, 1097–1101 (2006). Lee, Y. et al. The nuclear RNase III Drosha initiates microRNA processing. Nature 425, 415–419 (2003). Denli, A. M., Tops, B. B., Plasterk, R. H., Ketting, R. F. & Hannon, G. J. Processing of primary microRNAs by
33.
34.
35.
36. 37.
38.
39.
40.
41.
42. 43.
44. 45.
46. 47.
48.
49. 50. 51. 52. 53. 54. 55. 56.
57.
the microprocessor complex. Nature 432, 231–235 (2004). Landthaler, M., Yalcin, A. & Tuschl, T. The human DiGeorge syndrome critical region gene 8 and its D. melanogaster homolog are required for miRNA biogenesis. Curr. Biol. 14, 2162–2167 (2004). Han, M. H., Goud, S., Song, L. & Fedoroff, N. The Arabidopsis double-stranded RNA-binding protein HYL1 plays a role in microRNA-mediated gene regulation. Proc. Natl Acad. Sci. USA 101, 1093–1098 (2004). Okamura, K., Hagen, J. W., Duan, H., Tyler, D. M. & Lai, E. C. The mirtron pathway generates microRNAclass regulatory RNAs in Drosophila. Cell 130, 89–100 (2007). Ruby, J. G., Jan, C. H. & Bartel, D. P. Intronic microRNA precursors that bypass Drosha processing. Nature 448, 83–86 (2007). Hammond, S. C., Bernstein, E., Beach, D. & Hannon, G. J. An RNA-directed nuclease mediates posttranscriptional gene silencing in Drosophila cells. Nature 404, 293–296 (2000). Bernstein, E., Caudy, A. A., Hammond, S. M. & Hannon, G. J. Role for a bidentate ribonuclease in the initiation step of RNA interference. Nature 409, 295–296 (2001). Lee, Y., Jeon, K., Lee, J. T., Kim, S. & Kim, V. N. MicroRNA maturation: stepwise processing and subcellular localization. EMBO J. 21, 4663–4670 (2002). Yi, R., Qin, Y., Macara, I. G. & Cullen, B. R. Exportin‑5 mediates the nuclear export of pre-microRNAs and short hairpin RNAs. Genes Dev. 17, 3011–3016 (2003). Bohnsack, M. T., Czaplinski, K. & Gorlich, D. Exportin 5 is a RanGTP-dependent dsRNA-binding protein that mediates nuclear export of pre-miRNAs. RNA 10, 185–191 (2004). Lund, E., Guttinger, S., Calado, A., Dahlberg, J. E. & Kutay, U. Nuclear export of microRNA precursors. Science 303, 95–98 (2004). Lingel, A., Simon, B., Izaurralde, E. & Sattler, M. Structure and nucleic-acid binding of the Drosophila Argonaute 2 PAZ domain. Nature 426, 465–469 (2003). Liu, J. et al. Argonaute2 is the catalytic engine of mammalian RNAi. Science 305, 1437–1441 (2004). Chen, C. Z., Li, L., Lodish, H. F. & Bartel, D. P. MicroRNAs modulate hematopoietic lineage differentiation. Science 303, 83–86 (2004). This paper describes the identification of miRNAs that are differentially expressed during haematopoietic-lineage differentiation and shows that miR‑181a can modulate B‑ and T‑cell differentiation, providing the first example of miRNA function in vertebrate cells. Zeng, Y. & Cullen, B. R. Structural requirements for premicroRNA binding and nuclear export by Exportin 5. Nucleic Acids Res. 32, 4776–4785 (2004). Zeng, Y. & Cullen, B. R. Efficient processing of primary microRNA hairpins by Drosha requires flanking nonstructured RNA sequences. J. Biol. Chem. 280, 27595–27603 (2005). Zeng, Y., Yi, R. & Cullen, B. R. Recognition and cleavage of primary microRNA precursors by the nuclear processing enzyme Drosha. EMBO J. 24, 138–148 (2005). Zeng, Y., Cai, X. & Cullen, B. R. Use of RNA polymerase II to transcribe artificial microRNAs. Methods Enzymol. 392, 371–380 (2005). Wulczyn, F. G. et al. Post-transcriptional regulation of the let‑7 microRNA during neural cell specification. FASEB J. 21, 415–426 (2007). Thomson, J. M. et al. Extensive post-transcriptional regulation of microRNAs and its implications for cancer. Genes Dev. 20, 2202–2207 (2006). Nilsen, T. W. Mechanisms of microRNA-mediated gene regulation in animal cells. Trends Genet. 23, 243–249 (2007). Lim, L. P. et al. Microarray analysis shows that some microRNAs downregulate large numbers of target mRNAs. Nature 433, 769–773 (2005). Pillai, R. S., Bhattacharyya, S. N. & Filipowicz, W. Repression of protein synthesis by miRNAs: how many mechanisms? Trends Cell. Biol. 17, 118–126 (2007). Jackson, R. J. & Standart, N. How do microRNAs regulate gene expression? Sci. STKE 367, re1 (2007). Olsen, P. H. & Ambros, V. The lin‑4 regulatory RNA controls developmental timing in Caenorhabditis elegans by blocking LIN‑14 protein synthesis after the initiation of translation. Dev. Biol. 216, 671–680 (1999). Seggerson, K., Tang, L. & Moss, E. G. Two genetic circuits repress the Caenorhabditis elegans
nature reviews | immunology
58.
59.
60. 61.
62. 63. 64. 65. 66.
67. 68. 69.
70.
71. 72. 73. 74. 75. 76. 77.
78.
79.
80.
81.
82. 83.
heterochronic gene lin‑28 after translation initiation. Dev. Biol. 243, 215–225 (2002). Maroney, P. A., Yu, Y., Fisher, J. & Nilsen, T. W. Evidence that microRNAs are associated with translating messenger RNAs in human cells. Nature Struct. Mol. Biol. 13, 1102–1107 (2006). Nottrott, S., Simard, M. J. & Richter, J. D. Human let7a miRNA blocks protein production on actively translating polyribosomes. Nature Struct. Mol. Biol. 13, 1108–1114 (2006). Pillai, R. S. et al. Inhibition of translational initiation by Let‑7 MicroRNA in human cells. Science 309, 1573–1576 (2005). Wakiyama, M., Takimoto, K., Ohara, O. & Yokoyama, S. Let‑7 microRNA-mediated mRNA deadenylation and translational repression in a mammalian cell-free system. Genes Dev. 21, 1857–1862 (2007). Mathonnet, G. et al. MicroRNA inhibition of translation initiation in vitro by targeting the cap-binding complex eIF4F. Science 317, 1764–1767 (2007). Thermann, R. & Hentze, M. W. Drosophila miR2 induces pseudo-polysomes and inhibits translation initiation. Nature 447, 875–878 (2007). Kiriakidou, M. et al. An mRNA m7G cap binding-like motif within human Ago2 represses translation. Cell 129, 1141–1151 (2007). Lai, E. C. MicroRNAs are complementary to 3′UTR motifs that mediate negative post-transcriptional regulation. Nature Genet. 30, 363–364 (2002). Lewis, B. P., Shih, I. H., Jones-Rhoades, M. W., Bartel, D. P. & Burge, C. B. Prediction of mammalian microRNA targets. Cell 115, 787–798 (2003). This is one of the first reports on computational prediction of miRNA target genes. It shows that seed nucleotides, which are the 5′ 2–8 nucleotides of a mature miRNA, are crucial for computational target gene prediction. Stark, A., Brennecke, J., Russell, R. B. & Cohen, S. M. Identification of Drosophila microRNA targets. PLoS Biol. 1, E60 (2003). O’Donnell, K. A., Wentzel, E. A., Zeller, K. I., Dang, C. V. & Mendell, J. T. c‑Myc‑regulated microRNAs modulate E2F1 expression. Nature 435, 839–843 (2005). Rajewsky, N. microRNA target predictions in animals. Nature Genet. 38, S8–S13 (2006). This is a comprehensive review that describes the computational methods for miRNA target gene prediction. Lewis, B. P., Burge, C. B. & Bartel, D. P. Conserved seed pairing, often flanked by adenosines, indicates that thousands of human genes are microRNA targets. Cell 120, 15–20 (2005). Grimson, A. et al. MicroRNA targeting specificity in mammals: determinants beyond seed pairing. Mol. Cell 27, 91–105 (2007). Enright, A. J. et al. MicroRNA targets in Drosophila. Genome Biol. 5, R1 (2003). John, B. et al. Human microRNA targets. PLoS Biol. 2, e363 (2004). Rajewsky, N. & Socci, N. D. Computational identification of microRNA targets. Dev. Biol. 267, 529–535 (2004). Krek, A. et al. Combinatorial microRNA target predictions. Nature Genet. 37, 495–500 (2005). Farh, K. K. et al. The widespread impact of mammalian microRNAs on mRNA repression and evolution. Science 310, 1817–1821 (2005). Stark, A., Brennecke, J., Bushati, N., Russell, R. B. & Cohen, S. M. Animal microRNAs confer robustness to gene expression and have a significant impact on 3′UTR evolution. Cell 123, 1133–1146 (2005). Lytle, J. R., Yario, T. A. & Steitz, J. A. Target mRNAs are repressed as efficiently by microRNA-binding sites in the 5′ UTR as in the 3′ UTR. Proc. Natl Acad. Sci. USA 104, 9667–9672 (2007). Chang, S., Johnston, R. J. Jr, Frokjaer-Jensen, C., Lockery, S. & Hobert, O. MicroRNAs act sequentially and asymmetrically to control chemosensory laterality in the nematode. Nature 430, 785–789 (2004). Didiano, D. & Hobert, O. Perfect seed pairing is not a generally reliable predictor for miRNA-target interactions. Nature Struct. Mol. Biol. 13, 849–851 (2006). Vella, M. C., Choi, E. Y., Lin, S. Y., Reinert, K. & Slack, F. J. The C. elegans microRNA let‑7 binds to imperfect let‑7 complementary sites from the lin‑41 3′UTR. Genes Dev. 18, 132–137 (2004). Vella, M. C., Reinert, K. & Slack, F. J. Architecture of a validated microRNA::target interaction. Chem. Biol. 11, 1619–1623 (2004). Long, D. et al. Potent effect of target structure on microRNA function. Nature Struct. Mol. Biol. 14, 287–294 (2007).
volume 8 | february 2008 | 129 © 2008 Nature Publishing Group
REVIEWS 84. Plasterk, R. H. Micro RNAs in animal development. Cell 124, 877–881 (2006). 85. Zhao, Y. et al. Dysregulation of cardiogenesis, cardiac conduction, and cell cycle in mice lacking miRNA‑1‑2. Cell 129, 303–317 (2007). 86. Rodriguez, A. et al. Requirement of bic/microRNA‑155 for normal immune function. Science 316, 608–611 (2007). This report shows that mice deficient in miR‑155 are immunodeficient, and have defects in the function of B and T cells, as well as dendritic cells. 87. Thai, T. H. et al. Regulation of the germinal center response by microRNA‑155. Science 316, 604–608 (2007). Using gain‑of‑function and lost‑of‑function analyses in mice, this study demonstrates an important role for miRNA‑155 in the differentiation of T helper cells and the establishment of germinal centres. 88. van Rooij, E. et al. Control of stress-dependent cardiac growth and gene expression by a microRNA. Science 316, 575–579 (2007). 89. Calin, G. A. et al. MicroRNA profiling reveals distinct signatures in B cell chronic lymphocytic leukemias. Proc. Natl Acad. Sci. USA 101, 11755–11760 (2004). 90. Lu, J. et al. MicroRNA expression profiles classify human cancers. Nature 435, 834–838 (2005). 91. Garzon, R., Fabbri, M., Cimmino, A., Calin, G. A. & Croce, C. M. MicroRNA expression and function in cancer. Trends Mol. Med. 12, 580–587 (2006). 92. Esquela-Kerscher, A. & Slack, F. J. Oncomirs — microRNAs with a role in cancer. Nature Rev. Cancer 6, 259–269 (2006). 93. Monticelli, S. et al. MicroRNA profiling of the murine hematopoietic system. Genome Biol. 6, R71 (2005). 94. Ramkissoon, S. H. et al. Hematopoietic-specific microRNA expression in human cells. Leuk. Res. 30, 643–647 (2006). 95. Choong, M. L., Yang, H. H. & McNiece, I. MicroRNA expression profiling during human cord blood-derived CD34 cell erythropoiesis. Exp. Hematol. 35, 551–564 (2007). 96. Neilson, J. R., Zheng, G. X., Burge, C. B. & Sharp, P. A. Dynamic regulation of miRNA expression in ordered stages of cellular development. Genes Dev. 21, 578–589 (2007). This report describes the systematic cloning of miRNAs from purified T‑cell populations, and should be a useful resource for studying miRNAs that may have roles during T-cell development. 97. Manocha, M., Shankar, P., Sharp, P. A. & Manjunath, N. miRNA profiling of naive, effector, and memory CD8 T cells. PLoS ONE 2, e1020 (2007). 98. Zheng, Y. et al. Genome-wide analysis of Foxp3 target genes in developing and mature regulatory T cells. Nature 445, 936–940 (2007). 99. Li, Q. J. et al. miR‑181a is an intrinsic modulator of T cell sensitivity and selection. Cell 129, 147–161 (2007). This study shows that miR‑181a can function as an antigen sensitivity rheostat to modulate T‑cell sensitivity to antigens during T‑cell development and maturation by downregulating the expression of multiple phosphatases in the TCR signalling pathway. 100. Zhou, B., Wang, S., Mayr, C., Bartel, D. P. & Lodish, H. F. miR‑150, a microRNA expressed in mature B and T cells, blocks early B cell development when expressed prematurely. Proc. Natl Acad. Sci. USA 104, 7080–7085 (2007). 101. Daley, G. Q., Van Etten, R. A. & Baltimore, D. Induction of chronic myelogenous leukemia in mice by the P210bcr/abl gene of the Philadelphia chromosome. Science 247, 824–830. (1990). 102. Fazi, F. et al. A minicircuitry comprised of microRNA‑223 and transcription factors NFI‑A and C/ EBPα regulates human granulopoiesis. Cell 123, 819–831 (2005). 103. Fukao, T. et al. An evolutionarily conserved mechanism for microRNA‑223 expression revealed by microRNA gene profiling. Cell 129, 617–631 (2007). 104. Xiao, C. et al. MiR‑150 controls B cell differentiation by targeting the transcription factor c‑Myb. Cell 131, 146–159 (2007).
Using both gain- and loss‑of‑function analyses in mice, this study demonstrates the important role of miR‑150 in B‑cell development through targeting Myb. 105. Tam, W., Ben-Yehuda, D. & Hayward, W. S. bic, a novel gene activated by proviral insertions in avian leukosis virus-induced lymphomas, is likely to function through its noncoding RNA. Mol. Cell. Biol. 17, 1490–1502 (1997). 106. Eis, P. S. et al. Accumulation of miR‑155 and BIC RNA in human B cell lymphomas. Proc. Natl Acad. Sci. USA 102, 3627–3632 (2005). 107. Kluiver, J. et al. BIC and miR‑155 are highly expressed in Hodgkin, primary mediastinal and diffuse large B cell lymphomas. J. Pathol. 207, 243–249 (2005). 108. Metzler, M., Wilda, M., Busch, K., Viehmann, S. & Borkhardt, A. High expression of precursor microRNA‑155/BIC RNA in children with Burkitt lymphoma. Genes Chromosomes Cancer 39, 167–169 (2004). 109. Tam, W. & Dahlberg, J. E. miR‑155/BIC as an oncogenic microRNA. Genes Chromosomes Cancer 45, 211–212 (2006). 110. Costinean, S. et al. Pre‑B cell proliferation and lymphoblastic leukemia/high-grade lymphoma in Eµ-miR155 transgenic mice. Proc. Natl Acad. Sci. USA 103, 7024–7029 (2006). 111. Davey, G. M. et al. Preselection thymocytes are more sensitive to T cell receptor stimulation than mature T cells. J. Exp. Med. 188, 1867–1874 (1998). 112. Pircher, H., Rohrer, U. H., Moskophidis, D., Zinkernagel, R. M. & Hengartner, H. Lower receptor avidity required for thymic clonal deletion than for effector T‑cell function. Nature 351, 482–485 (1991). 113. Hogquist, K. A., Jameson, S. C. & Bevan, M. J. The ligand for positive selection of T lymphocytes in the thymus. Curr. Opin. Immunol. 6, 273–278 (1994). 114. Taganov, K. D., Boldin, M. P., Chang, K. J. & Baltimore, D. NF‑κB‑dependent induction of microRNA miR‑146, an inhibitor targeted to signaling proteins of innate immune responses. Proc. Natl Acad. Sci. USA 103, 12481–12486 (2006). 115. O’Connell, R. M., Taganov, K. D., Boldin, M. P., Cheng, G. & Baltimore, D. MicroRNA‑155 is induced during the macrophage inflammatory response. Proc. Natl Acad. Sci. USA 104, 1604–1609 (2007). 116. Lecellier, C. H. et al. A cellular microRNA mediates antiviral defense in human cells. Science 308, 557–560 (2005). This paper demonstrates that the cellular miRNA miR‑32 can effectively limit replication of PFV1, and a suppressor protein encoded by the virus can counteract the repressive effect of miR‑32 in a plant system. 117. Pedersen, I. M. et al. Interferon modulation of cellular microRNAs as an antiviral mechanism. Nature 449, 919–921 (2007). This report shows that the IFN signalling system, the key defence mechanism against viral infection in mammalian cells, works in concert with miRNAs to control viral infection. 118. Jopling, C. L., Yi, M., Lancaster, A. M., Lemon, S. M. & Sarnow, P. Modulation of hepatitis C virus RNA abundance by a liver-specific microRNA. Science 309, 1577–1581 (2005). This paper shows that the host miRNA miR‑122 could be used by HCV to potentiate HCV replication. 119. Lu, R. et al. Animal virus replication and RNAimediated antiviral silencing in Caenorhabditis elegans. Nature 436, 1040–1043 (2005). 120. Li, H., Li, W. X. & Ding, S. W. Induction and suppression of RNA silencing by an animal virus. Science 296, 1319–1321 (2002). This study shows that flock house virus (FHV) encodes an RNAi suppressor protein, B2, which is required for the FHV infection of D. melanogaster host cells, indicating the importance of RNA interference pathway in antiviral defence in flies. 121. Sarnow, P., Jopling, C. L., Norman, K. L., Schutz, S. & Wehner, K. A. MicroRNAs: expression, avoidance and subversion by vertebrate viruses. Nature Rev. Microbiol. 4, 651–659 (2006).
130 | february 2008 | volume 8
122. Pfeffer, S. et al. Identification of virus-encoded microRNAs. Science 304, 734–736 (2004). This is the first study to show that DNA viruses encode miRNAs. 123. Pfeffer, S. et al. Identification of microRNAs of the herpesvirus family. Nature Methods 2, 269–276 (2005). 124. Sullivan, C. S., Grundhoff, A. T., Tevethia, S., Pipas, J. M. & Ganem, D. SV40-encoded microRNAs regulate viral gene expression and reduce susceptibility to cytotoxic T cells. Nature 435, 682–686 (2005). This report demonstrates that virus-encoded miRNAs could regulate SV40-encoded mRNAs and facilitate viral infection. 125. Cai, X. et al. Kaposi’s sarcoma-associated herpesvirus expresses an array of viral microRNAs in latently infected cells. Proc. Natl Acad. Sci. USA 102, 5570–5575 (2005). 126. Gupta, A., Gartner, J. J., Sethupathy, P., Hatzigeorgiou, A. G. & Fraser, N. W. Anti-apoptotic function of a microRNA encoded by the HSV‑1 latency-associated transcript. Nature 442, 82–85 (2006). 127. Stern-Ginossar, N. et al. Host immune system gene targeting by a viral miRNA. Science 317, 376–381 (2007). 128. Triboulet, R. et al. Suppression of microRNA-silencing pathway by HIV‑1 during virus replication. Science 315, 1579–1582 (2007). 129. Bennasser, Y. & Jeang, K. T. HIV‑1 Tat interaction with Dicer: requirement for RNA. Retrovirology 3, 95 (2006). 130. Metcalf, D. & Nicola, NA. The hematopoietic colonystimulating factors. From biology to clinical applications. (Cambridge Univ. Press, 1995). 131. Shivdasani, R. A. & Orkin, S. H. The transcriptional control of hematopoiesis. Blood 87, 4025–4039 (1996). 132. Zheng, Y. et al. Genome-wide analysis of Foxp3 target genes in developing and mature regulatory T cells. Nature 445, 936–940 (2007). 133. Raveche, E. S. et al. Abnormal microRNA‑16 locus with synteny to human 13q14 linked to CLL in NZB mice. Blood 109, 5079–5086 (2007). 134. Lian, S. et al. GW bodies, microRNAs and the cell cycle. Cell Cycle 5, 242–245 (2006). 135. Gregory, R. I. et al. The microprocessor complex mediates the genesis of microRNAs. Nature 432, 235–240 (2004). 136. Gregory, R. I., Chendrimada, T. P., Cooch, N. & Shiekhattar, R. Human RISC couples microRNA biogenesis and posttranscriptional gene silencing. Cell 123, 631–640 (2005). 137. Haase, A. D. et al. TRBP, a regulator of cellular PKR and HIV‑1 virus expression, interacts with Dicer and functions in RNA silencing. EMBO Rep. 6, 961–967 (2005). 138. Chendrimada, T. P. et al. TRBP recruits the Dicer complex to Ago2 for microRNA processing and gene silencing. Nature 436, 740–744 (2005).
Acknowledgements
We thank members of the Chen and Lodish laboratories and also V. Ambros for his helpful comments on this manuscript. This research on miRNAs is supported by grants 1R01HL081612‑01 to C.‑Z. C. and 5R01DK068348 to H.F.L.
DATABASES Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene DUSP5 | DUSP6 | miR‑146a | miR‑150 | miR‑155 | miR‑181a | PTPN22 | SHP2
FURTHER INFORMATION miRBase targets: http://microrna.sanger.ac.uk/targets/v4/ miRNA registry: http://microrna.sanger.ac.uk/sequences/ PicTar: http://pictar.bio.nyu.edu/ TargetScan: http://www.targetscan.org/ All links are active in the online pdf
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS
Phagocytosis and comparative innate immunity: learning on the fly Lynda M. Stuart* and R. Alan Ezekowitz §
Abstract | Phagocytosis, the engulfment of material by cells, is a highly conserved process that arose before the development of multicellularity. Phagocytes have a key role in embryogenesis and also guard the portals of potential pathogen entry. They discriminate between diverse particles through the array of receptors expressed on their surface. In higher species, arguably the most sophisticated function of phagocytes is the processing and presentation of antigens derived from internalized material to stimulate lymphocytes and long-lived specific immunity. Central to these processes is the generation of a phagosome, the organelle that forms around internalized material. As we discuss in this Review, over the past two decades important insights into phagocytosis have been gleaned from studies in the model organism Drosophila melanogaster. Phagocytes Cells such as macrophages and neutrophils that internalize particulate material in an actindependent manner. Derived from the greek phagein (to eat) and kytos (cell).
Apoptotic cell A cell that has died by a genetically regulated programme of cell death. It is characterized by cell shrinkage, chromatin condensation, cellmembrane blebbing and DNA fragmentation. Eventually, the cell breaks up into many membrane-bound apoptotic bodies, which are phagocytosed by neighbouring cells.
*Developmental Immunology, Department of Pediatrics, Massachusetts General Hospital, Harvard Medical School, 55 Fruit Street, Boston, Massachusetts 02144, USA. § Merck Research Laboratories, Lincoln Avenue, Rahway, New Jersey 07065, USA. Correspondence to L.M.S. e‑mail:
[email protected] doi:10.1038/nri2240
Phagocytosis is the process by which particles are recognized, bound to the surface of cells and internalized into a ‘phagosome’, the organelle that forms around the engulfed material (FIG. 1). This carefully orchestrated cascade of events begins with the engagement of phagocytic receptors that activate numerous signalling pathways. These signals coordinate an orderly progression of changes, including the rearrangement of the cytoskeleton that guides the circumferential movement of membrane to internalize the bound particles. Phagocytosis is an essential cellular process, performed by unicellular organisms and many cell types found in metazoans. In simple organisms, such as the slime mould Dictyostelium discoideum, the motile phagocytes patrol the cellular aggregates (slugs) and have a role in host defence1. In higher organisms, engulfment is a particular feature of professional phagocytes, such as macrophages and neutrophils. These cells demonstrate a remarkable ability to internalize material and are able to engulf particles larger than their own surface area. Two main classes of targets are removed by phagocytosis: microorganisms and ‘altered self ’ particles, exemplified by apoptotic cells. Engulfment of apoptotic cells is prominent during tissue remodelling and embryogenesis when excess cells undergo programmed cell death and removal. An illustrative example of the crucial role for macrophages in development is in holometabolous insects, such as the fruit fly (Drosophila melanogaster), in which haemocytes have a crucial role in morphogenesis (BOX 1). Apoptosis begins by late stage 10 of embryogenesis and the removal of these effete cells by phagocytes that appear at the beginning of late
nature reviews | immunology
stage 11 of embryogenesis is essential for development2,3. Likewise, in mice, macrophages populate remodelling tissues early during embryogenesis and remove dying cells4,5. During pathogen invasion, phagocytosis is also the cornerstone of the early innate immune response and host-defence mechanisms of many species6. For unicellular organisms, such as amoeba, bacteria internalized from the extracellular milieu are an essential nutrient source7. However, the intracellular growth of these internalized microorganisms must be limited to prevent them overwhelming the host, and it is likely therefore that pathogen sensing has evolved along with the most fundamental role for phagocytosis — that is, nutrient acquisition. In higher species, phagocytes also function as important antigen-presenting cells and are required to prime effective adaptive immunity. The central role of phagocytes in host defence is emphasized by observations that many pathogen virulence factors directly manipulate the process of engulfment or functions of the phagosome. Despite progress in understanding the complexity of phagocytosis since its first description by Metchnikoff a century ago (BOX 2), our knowledge remains incomplete. The study of phagocytosis is complicated by the partial redundancy of key components and it appears that many cell-surface receptors and components of the phagocytic machinery often have multiple overlapping functions. This high level of redundancy may reflect the essential role of phagocytosis and the need for alternative back-up mechanisms to remove dying cells during development, and invading pathogens during infectious challenge. Furthermore, examining the relative contribution of volume 8 | february 2008 | 131
© 2008 Nature Publishing Group
REVIEWS Pseudopod tips Particle
Pseudopod base Early phagosome
Phagocytic cup Endosome
Exocyst Lysosome
Phagolysosome
Destruction of particle or pathogen killing
Figure 1 | Phagocytosis delivers the bound particle from the cell surface into the phagosome. Receptor ligation triggers an orderly progression of cellular changes leading to rearrangement of the actin cytoskeleton and membrane | Immunology remodelling. Once formed, the phagosome undergoes maturation by fission and limited fusionNature eventsReviews with endosomes and lysosomes to generate the acidic and hydrolytic environment of the mature phagolysosome. During phagocytosis, the membrane undergoes areas of marked curvature, in the base of the phagocytic cup and at the origins and tips of extending pseudopods. Studies of fly cells have suggested that this membrane curvature may be stabilized by coat proteins, including clathrin and the coatamer protein complex. To internalize large or numerous particles, phagocytes must recruit additional membrane from intracellular stores, such as endosomes and the endoplasmic reticulum. During this process the exocyst, an octodimeric complex that associates with Drosophila melanogaster phagosomes, is suggested to tether the recruited endosomes to the base of the phagocytic cup. The exocyst and the coatamer protein complex may also have roles in the docking of vesicles with the phagosome during maturation.
Holometabolous insects Insects, such as Drosophila melanogaster, that have a pupal stage during which they undergo complete reorganization and metamorphosis.
Haemocytes Cells found within the haemolymph of an insect that are equivalent to the blood cells in vertebrates.
RNA interference The silencing of gene expression by the introduction of double-stranded RNAs that trigger the specific degradation of a homologous target mRNA and often subsequently decrease production of the encoded protein.
individual components of this system is difficult as primary mammalian phagocytes are not readily amenable to genetic manipulations, such as cDNA overexpression or knockdown of expression of candidate receptors by RNA interference (RNAi). Investigators have therefore relied on classic cell biology and microscopy techniques, as well as exploring the role of individual receptors that have been overexpressed in heterologous, non-professional phagocytes such as COS cells8,9 or by using cells isolated from knockout mice. These approaches have provided important insights into Fc receptor (FcR)- and complementmediated phagocytosis and signalling10,11. However, the limitations of the mammalian system have hampered the study of phagocytosis in intact organisms and have prevented the use of more exploratory and unbiased approaches. Therefore, building on the work of developmental biologists, some researchers in the field of innate immunity and macrophage cell biology have turned to genetically tractable model organisms that have been used successfully to study host defence pathways and the complex cell biology of phagocytosis. In this regard, work in both D. discoideum and Caenorhabditis elegans has been instructive but is beyond the scope of this Review (for recent reviews, see Refs 12,13) and we focus instead on recent work in the fruit fly, D. melanogaster. Professional circulating phagocytes were identified in D. melanogaster several years ago (BOX 1) and their identification suggested that this genetically tractable system might be well suited for the study of professional phagocytes. Over the past decade the crucial components of engulfment in these cells have been defined. The initial approaches used the amenability of D. melanogaster to genetic manipulation in vitro, ex vivo and in vivo14,15, and recently an in vivo screen has been used to identify a new E3-ubiquitin-ligase-dependent mechanism of uptake of apoptotic cells16. However, although informative, these in vivo screens are also labour intensive. Therefore, to take advantage of the emerging power of RNAi, an in vitro system was developed using a variant of the well-known
132 | february 2008 | volume 8
embryonic S2 (Schneider line 2) cell line15,17. These cells demonstrate properties similar to mammalian macrophages and efficiently ingest particles and bacteria in a temperature- and actin-dependent manner. In addition, S2 cells adopt a morphology after phagocytosis that is identical to that observed by primary larval plasmatocytes and reminiscent of the typical electron micrographs of professional phagocytes18. This in vitro system has been readily amenable to high-throughput screening with RNAi and has established S2 cells as a powerful platform for RNAi screens19. Subsequently, S2 cells have been used in many RNAi screens to identify molecules involved in host interactions with pathogens, such as Escherichia coli, Staphylococcus aureus19–21, Mycobacterium fortuitum22, Candida albicans 23 and Listeria monocytogenes 24,25 (reviewed in ref. 26). Moreover, much of this work has been carried out in parallel with studies revealing a more extensive understanding of the innate immune defence system in this organism27,28. In addition, the experimental tractability of D. melanogaster has facilitated the testing of the role of phagocytosis, and some of the phagocytic machinery, in host defence in vivo29–34. Using various models of pathogenesis, it is now clear that phagocytes are required to regulate some, but not all, bacterial infections. Here we discuss these advances and place recent findings in the broader context of mammalian phagocytosis. In addition, we examine how findings from this model system have provided insights into the evolutionary origins of immunology, adding value to this approach.
A phagosome-centric view of engulfment After internalization, particles are delivered into a de novogenerated organelle, the phagosome. This simple definition of this organelle belies its complexity and importance in immune defence. Phagosomes have a central role in both innate and adaptive immunity. It has become evident that the nature and composition of each phagosome is determined by factors such as the cell-surface receptors www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS Opsonin A soluble molecule that, when bound to a particle, enhances uptake of the particle by phagocytosis.
Pattern-recognition receptors Receptors that recognize molecular patterns often associated with pathogens, such as lipopolysaccharide found in the bacterial cell walls of Gram-negative microorganisms. They also recognize molecular patterns expressed by non-pathogenic bacteria.
engaged by the particle, the mechanism of entry used, the origins of the contributing membrane and the nature of the cargo35,36. In most circumstances, the nascent organelle undergoes a process termed ‘maturation’ by fusion and limited fission events with endosomes and lysosomes to generate the mature phagolysosome37 (FIG. 1). The mature phagosome is highly hydrolytic, being able to limit the replication of bacteria and in many cases can kill the intern alized microorganism. However, the fate of the newly formed phagosome can be actively subverted by certain pathogens that have developed mechanisms to hijack or escape the natural maturation process38. Therefore, each phagosome is uniquely tailored for, and by, its cargo. For these reasons phagosomes are heterogeneous, quasi-stable and highly complex organelles, making the study of the underlying cell biology challenging. Recent work using proteomics and computational modelling of D. melanogaster has made a significant advance in this area and provided a novel framework that accommodates the emerging complexity of the phagosome21. Building on previous studies in which the protein content of mammalian phagosomes was determined35, proteomic analysis of generic latex-beadcontaining phagosomes from D. melanogaster S2 cells has revealed many similarities between the phagosomes from these evolutionarily distant species. It is evident that, even in this simpler organism, the phagosome is a highly complex organelle (FIG. 2). Approximately 600 D. melanogaster proteins were identified to be associated with phagosomes, 70% of which had mammalian orthologues, validating this as a model system for mammalian phagocytosis. Using a new computational approach, it was predicted that hundreds of protein–protein interactions exist within the phagosome, which contain numerous subcellular biomodules or complexes including the vacuolar-ATPase (V-ATPase), the exocyst and the coatamer protein complex21. Furthermore, it identified the potential for numerous signalling pathways that originate from this subcellular structure such as the nuclear factor-κB (NF-κB) and mitogen-activated protein kinase (MAPK) signalling pathways21 (FIG. 2).
Box 1 | Drosophila melanogaster development and phagocyte function D. melanogaster development is complex and comprises embryonic, three larval, pupal and adult stages. The entire developmental cycle from fertilization to adulthood is completed in 9 days. Embryonic development takes 20 hours and is divided into 17 stages that are defined by precise morphological events. It includes gastrulation, neurogenesis and generation of parasegments, and is terminated by larval hatch (see The Interactive Fly website). At the end of the three larval stages, pupariation takes place. Over the ensuing 4 days, metamorphosis occurs entailing widespread tissue remodelling, resorbtion of larval tissues and growth of adult organs. The adult fly hatches on day 9. Phagocytic blood cells or haemocytes can first be observed at the end of embryonic stage 11, shortly after the onset of apoptosis2. The first studies of insect haemocytes were done by Rizki and Rizki in the 1960s; they showed the phagocytic capacity and the integral role that haemocytes have in defence against microorganisms100. Similar to mammals, D. melanogaster haematopoiesis occurs in two distinct waves at two different sites — in the case of D. melanogaster, one embryonic, the other one larval. Both lineages of haemocytes persist to the adult stage. Depending on the developmental stage, the predominant function of D. melanogaster phagocytes may either be the removal of effete cells and larval tissues (embryos and pupae) or in pathogen surveillance and clearance (larvae and adults).
nature reviews | immunology
It is evident therefore, that there is amazing complexity to this organelle. Thus, although phagocytosis might be considered simply as a mechanism of waste disposal, an alternative view is that its function is to generate the phagosome, an organelle that is required for the effective induction of host defence and other important homeostatic processes. We discuss what has been learned from D. melanogaster about how phagosomes form and the multiple roles this has in the cell biology of phagocytes.
Cell-surface recognition Phagocytosis is initiated by the ligation of cell-surface receptors that either directly bind to the particle or to opsonins that are deposited on the particle’s surface. Both forward and reverse genetics, as well as RNAi screens, in D. melanogaster cells have been successfully used to identify potential phagocytic machinery19–25. Some of the receptors have direct mammalian orthologues and demonstrate conservation of function, whereas others appear to be unique to insects (TABLE 1). These data indicate that there are four main classes of molecules involved in recognition: complement-like opsonins, scavenger receptors, a newly emerging family of epidermal growth factor (EGF)-like-repeat-containing receptors, and a highly variant receptor and opsonin, Down syndrome cell-adhesion molecule (DSCAM). Complement-like opsonization. Complement components have multiple functions in mammalian host defence, including chemoattraction, formation of the membrane-attack complex and opsonization. The bestcharacterized opsonins in D. melanogaster are a group of proteins that are related to mammalian α2‑macroglobulin and C3, the thioester-containing proteins (TEPs). Similar to mammalian acute-phase proteins, the transcription and secretion of TEPs is upregulated after bacterial challenge39. Insight into the function of D. melanogaster TEPs was derived from an S2-cell RNAi screen, in which a member of the TEP family, macroglobulin-related protein (MCR; also known as TEPVI), was found to bind and increase phagocytosis of C. albicans23. Other TEPs also bind E. coli (TEPII) and S. aureus (TEPIII) and enhance phagocytosis23. In addition, an in vivo RNAi strategy confirmed the role of the TEPs in phagocytosis in the mosquito Anopheles gambiae40, indicating that other insects also use these opsonins. However, the functional equivalent of complement receptors in D. melanogaster remains to be defined. Scavenger receptors. The scavenger receptors are structurally unrelated multi-ligand receptors that are defined by their ability to bind polyanionic ligands. They bind a wide range of pathogens and modified self ligands, and have emerged as important patternrecognition receptors (PRRs)41 in many species. Some mammalian scavenger receptors have additional roles, acting not only as phagocytic receptors but also in host defence42; these include CD14 and CD36, which regulate Toll-like receptor 4 (TLR4) and TLR2 signalling, respectively20,43,44. Supporting their importance in host volume 8 | february 2008 | 133
© 2008 Nature Publishing Group
REVIEWS Box 2 | The discovery of phagocytes and phagocytosis Elie Metchnikoff was a Russian embryologist whose seminal work on phagocytosis and ‘physiological inflammation’ was inspired by the observation of these processes in a simple metazoan, the sea-star larva97. Metchnikoff observed the accumulation of cells he termed phagocytes around a rose thorn that he used to induce an inflammatory insult in the larva. This observation was the nidus from which his theories of cellular immunity were generated and for which he won a Nobel Prize in 1908. Importantly, his work emphasized the value of comparative embryology and established the validity of studying phagocytic engulfment in model organisms.
N-ethyl‑N-nitrosourea (ENU) mutagenesis An alkylating agent and potent mutagen that when given to mice efficiently induces random mutations.
defence, scavenger-receptor families are expanded in many species, particularly those that lack classic adaptive immunity. As an example, only three CD36-like scavenger receptors exist in humans but there are ten in D. melanogaster. It is likely that this expansion is driven by the ability of these receptors to bind pathogens and hence expansion increases the repertoire of receptors for bacterial recognition and provides a strong survival advantage. The class B scavenger receptors, CD36 and the CD36like protein known as Croquemort, are noteworthy examples of multifunctional receptors in mammals and flies, respectively. In mammals, CD36, in association with integrins αvβ3-integrin and thrombospondin, mediates the engulfment of apoptotic cells45. Similarly, Croquemort is a receptor for apoptotic cells in D. melanogaster14,46. Apoptosis is prominent in the embryo2 and Croquemortdeficient flies demonstrate persistence of uncleared dying cells, a defect that is rescued by transgenic expression of Croquemort. So, Croquemort appears to be a member of an evolutionarily conserved, CD36-like family of receptors for apoptotic cells and was the first example to validate the use of D. melanogaster as a model to study mammalian phagocytic receptors. Croquemort is also expressed in adult flies when apoptosis is not prominent, an observation that raised the possibility that it has additional functions beyond the clearance of dying cells. Consistent with a role as a multi-ligand receptor, it is now known that Croquemort binds S. aureus, and this led to the identification of mammalian CD36 as a phagocytic receptor for S. aureus20. Contemporaneous to this discovery, an N‑ethyl‑N-nitrosourea (ENU) mutagenesis strategy in mice identified a non-sense mutation in Cd36 (oblivious) that resulted in defective host defence against S. aureus44. CD36, and presumably Croquemort, recognize S. aureus through diacylated lipopeptides that are found on the bacterial cell surface. Interestingly, not only does CD36 act as a phagocytic receptor but it also cooperates with TLR2 and TLR6 to increase signalling in response to lipoteichoic acid (LTA) and diacylated lipopeptides in mammals20,44. It remains to be determined whether Croquemort functions similarly in Toll signalling in D. melanogaster. Another class B scavenger receptor, Peste, which was identified by an RNAi screen, is a receptor for bacteria in D. melanogaster. It can bind M. fortuitum22, suggesting that mammalian class B scavenger receptors may be involved in the recognition of mycobacteria. Although not yet tested in vivo, this possibility is supported by
134 | february 2008 | volume 8
overexpression experiments in which the class B scavenger receptors SR‑BI and/or SR‑BII, but not CD36, conferred the ability to bind M. fortuitum in vitro22. In contrast to these two examples of functionally conserved class B scavenger receptors, D. melanogaster has a structurally unrelated insect-specific class of scavenger receptors, SR‑C (scavenger receptor class C)47. SR‑CI is a type I membrane protein that contains domains that are related to complement control proteins and mucins, and binds bacteria17. It belongs to a small gene family of four proteins, two of which are predicted to be secreted proteins. Consistent with a role in host defence, SR‑CI expression is upregulated in fly larva after exposure to bacteria48 and a high level of naturally occurring polymorphism exists in Sr-CI that has been associated with varying levels of resistance to bacterial infection49. Unlike SR‑CI, SR‑CII is not thought to have a role in immune defence49. Receptors with multiple EGF-like repeats. Accumulating evidence from numerous species indicates the existence of a new family of PRRs that use EGF-like repeats to recognize diverse ligands30,50 (FIG. 3). Members of this family are particularly well defined in D. melanogaster, although they are found in many insect species, in C. elegans and in mammals, in which they might have similar functions50. A D. melanogaster protein, Eater, is perhaps one of the best characterized proteins of this newly emerging family of EGF-like-repeat-containing PRRs. It is a type I membrane protein that contains 32 characteristic EGF-like repeats in the extracellular domain. Eater acts as a bona fide PRR that directly binds microbial surfaces through its four N‑terminal EGF-like repeats30. Eater was identified as a putative target of Serpent, a D. melanogaster GATA transcription factor that had been found to be essential for bacterial phagocytosis by an RNAi screen19. Silencing Eater expression in S2 cells decreases bacterial binding and uptake, and macro phages from flies that lack Eater are impaired in their ability to phagocytose bacteria, leading to increased susceptibility to certain infections30. Eater therefore appears to be a major receptor for a broad range of pathogens in D. melanogaster. A similar D. melanogaster molecule, Nimrod C1, was identified as the protein target of a haemocyte-specific antibody50. It is a 90 kDa transmembrane protein with 10 EGF-like repeats that are similar to those found in Eater. RNAi-mediated silencing of Nimrod C1 expression decreases the uptake of bacteria, and overexpression of this protein increases the adherence of S2 cells to tissue-culture plates50, suggesting that it acts both as a phagocytic receptor and a potential adhesion molecule. The gene encoding Nimrod C1 is part of a cluster of 10 related Nimrod genes that are found in D. melanogaster. Similar proteins are also found in silk moths and the beetle Holotrichia diomphalia51, in which they serve as secreted opsonins that are involved in bacterial clearance. Orthologues also exist in flesh flies, mosquitoes and honeybees but not in mammals. www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS Draper, originally identified as being regulated by the transcriptional factor glial cells missing, is expressed on the surface of glial cells and D. melanogaster macro phages52. It is required for the removal of apoptotic cells in the central nervous system52 and embryonic macrophages and S2 cells53. Draper also contributes to axon pruning54 and the removal of severed axons in models of Wallerian degeneration55. Draper is the D. melanogaster orthologue of the C. elegans apoptoticcell receptor, CED-1 (Ref. 13), and similar to CED-1, Draper triggers engulfment through engagement of CED-6 by an NPXY (where X denotes any amino acid) motif on its intracellular tail56. This motif is conserved in the mammalian apoptotic-cell receptor CD91 (also known as LRP), which suggests that Draper, CED-1 and CD91 trigger similar engulfment machinery in C. elegans, D. melanogaster and mammals, respectively. However, CD91 and Draper differ in their extracellular domain and the closest mammalian homologues of Draper are multiple EGF-like domain 10 (MEGF10) and MEGF11 (Ref. 57). The essential role of Draper in altered-self recognition in D. melanogaster strongly suggests that MEGF10 and MEGF11 might also have a role in the recognition of dying cells in mammalian systems57.
DSCAM and receptor diversification. A key feature of immune receptors is the enormous repertoire required to accommodate a large and rapidly evolving pool of potential pathogens. In mammals, antigen receptor diversity is generated somatically through gene rearrangement and, in the case of B cells, is further increased by somatic hypermutation. By contrast, these classic features of adaptive immunity were thought to be absent in D. melanogaster, raising the question of how D. melanogaster generates the required repertoire to deal with emerging pathogens. The recent discovery of the role of an immunoglobulin superfamily member, DSCAM, in host defence in D. melanogaster might provide the answer. DSCAM was first described in neuronal development and, through homophilic and heterophilic interactions, determines the correct wiring of neurons58. To accomplish the necessary diversity to accommodate many different neurons, it is predicted that DSCAM may have more than 38,000 potential splice variants that are generated by combining constant and variable regions59. Interestingly, it is possible that D. melanogaster immune tissues (comprised of haemocytes and fat-body cells) can express more than 18,000 different extracellular domains of DSCAM60 (FIG. 4a), and soluble forms of DSCAM have also been identified in cell supernatants and in the haemolymph.
Pre-initiation complex Translation Signalling via G-proteins
Immune receptor signalling Nuclear pore complex DSCAM?
G-proteins
NF-κB
Toll? GPCR D. melanogaster phagosome
Chemical signal
ABC transporter
Cell survival or apoptosis
RAS Exocyst
Rab
Vesicle trafficking
GFR Integrin
Endosome V-ATPase
Scavenger receptor
MAPK JNK
PYK2 Cytoskeleton
CD42
The ordered rearrangement of variable regions of genes encoding antigen receptors that contributes to increase receptor diversity.
The process by which antigenactivated B cells in germinal centres mutate their rearranged immunoglobulin genes. The B cells are subsequently selected for those expressing the ‘best’ mutations on the basis of the ability of the surface immunoglobulin to bind antigen.
RAC
ARF
Gene rearrangement
Somatic hypermutation
p38 MAPK
Calpain
Chaperonin-containing T complex
Figure 2 | The phagosome is a multifunctional organelle. Proteomic analysis has identified more than 600 Nature Reviewsof | Immunology proteins that are associated with the phagosomes of Drosophila melanogaster. Bioinformatic analyses these proteins suggest that numerous protein–protein interactions and macromolecular complexes associate with this organelle. These include the vacuolar-ATPase (V-ATPase), the exocyst and the chaperonin-containing T complex. In addition, components of multiple signalling pathways are present in association with phagosomes, suggesting this as a point of initiation of signal transduction. These pathways are likely to be downstream of transmembrane proteins found in the phagosome such as G‑protein-coupled receptors (GPCRs), scavenger receptors, integrins, the Toll receptor and growth-factor receptors (GFRs). ABC, ATP-binding cassette; ARF, ADP-ribosylation factor; MAPK, mitogen-activated protein kinase; DSCAM, Down syndrome cell-adhesion molecule; NF-κB, nuclear factor-κB; PYK2, protein tyrosine kinase 2.
nature reviews | immunology
volume 8 | february 2008 | 135 © 2008 Nature Publishing Group
REVIEWS Table 1 | Cell-surface recognition receptors and opsonins Drosophila melanogaster receptors
Ligands
Related mammalian molecules
Refs
TEPs
TEPVI (MCR)
Candida albicans
Complement components
TEPIII
Staphylococcus aureus
TEPII
Escherichia coli
Scavenger receptors
Croquemort*
Apoptotic cells and S. aureus
CD36*
14, 20, 46
Peste*
S. aureus, Listeria monocytogenes
SR-BI*
22
SR-CI
S. aureus, E. coli and dsRNA
ND
17
EGF-like-repeat-containing Nimrods
Eater
S. aureus, L. monocytogenes, E. coli and dsRNA
30
Draper
Apoptotic cells, axon pruning and severed axons
MEGF10; MEGF11; CD91 (LRP); SREC; Stabilin 1 and Stabilin 2
Nimrod C1
E. coli and S. aureus
Others
PGRP-LC
E. coli
Mammalian PGRPs‡
PGRP-SA (soluble)
S. aureus
23
52–55 50 19
*Direct mammalian orthologues. ‡Soluble molecules with no role in phagocytosis. MCR, macroglobulin-complement related protein; LRP, lipoprotein receptorrelated protein; MEGF, multiple epidermal growth factor (EGF)-like domain; ND, not determined; PGRP, peptidoglycan-recognition protein; SR, scavenger receptor; TEP, thioester-containing protein; SREC, scavenger receptor expressed by endothelial cells.
Convergent evolution The development of similar characteristics in organisms that are unrelated as a consequence of how each adapts to a similar evolutionary pressure but which occurs after evolutionary divergence.
Pseudopod A transient protrusion from the cell during cell movement or to envelop a particle that is to be internalized.
DSCAM binds E. coli and potentially acts as both a phagocytic receptor and an opsonin60 (FIG. 4b). This suggestion is supported by the identification of DSCAM in the phagosome proteome21. The crystal structures of two DSCAM isoforms reveal each one with two distinct surface epitopes, one on either side of the receptor, which are generated by its hypervariable amino-acid residues. This configuration allows a given DSCAM isoform to form a homodimer (via a homophilic interaction through one hypervariable epitope) and retain the ability to recognize, opsonize and crosslink pathogens (via a second hypervariable epitope)61. DSCAM therefore has many similarities to antibodies and might have similar functions in D. melanogaster as immunoglobulins in mammals. Individual cells have been predicted to express between 15 and 50 DSCAM isoforms62, and data from the mosquito A. gambiae suggest that there is increased secretion of pathogen-specific isoforms after infectious challenge63. However, although these observations raise the possible preferential use of splice variants after infection, they have not yet been confirmed in studies of DSCAM in other species. Nonetheless, it is tempting to compare DSCAM and immunoglobulin receptors. DSCAM exists in a secreted form and an intriguing possibility is that DSCAM-secreting cells are functionally similar to mammalian immunoglobulin-secreting plasma cells (FIG. 4b). DSCAM provides a potential example of convergent evolution in the immune system of these two distant species and suggests that highly variable receptors (such as immunoglobulins, highly variable lymphocyte receptors (VLRs) of lamprey64 and DSCAM splice variants) are examples of ‘optimal design’ and are a desired feature found in many effective immune systems.
136 | february 2008 | volume 8
The machinery of uptake Over the past 20 years, an extensive body of work has defined the machinery that functions downstream of mammalian complement and of Fc receptors65–68 (reviewed in REFS 69–71), although much less is known about how other phagocytic receptors function. Over the past 3 years, genome-wide RNAi screening strategies in D. melanogaster S2 cells have facilitated the identification of the components of the machinery required for the uptake of E. coli, S. aureus, M. fortuitum, L. monocytogenes and C. albicans19–25. These studies have shown that much of the machinery triggered by complement receptors and FcRs in mammals is also used for pathogen uptake in D. melanogaster, an organism that does not express these receptors. In addition to validating many known regulators of uptake, these screens have also provided novel insights into the process of internalization. In particular, RNAi screens in D. melanogaster S2 cells have been the first to implicate the coatamer protein complex19,21–25 and the exocyst19,21,22,24 in bacterial uptake. The coatamer complex. During phagocytosis marked changes occur in the plasma membrane, which must bend significantly to form the phagocytic cup and during pseudopod extension (FIG. 1). Although such membrane remodelling is vital during engulfment, the molecular mechanisms that stabilize the areas of membrane curvature are poorly understood. Three well-recognized mechanisms exist for the generation of membrane curvature and vesicle budding, which are mediated by the proteincoat complexes clathrin, coat-protein complex I (COPI) and COPII (REF. 72), and it is likely that some or all of these contribute to maintain the curvature of the membrane that is required for phagocytosis. Clathrin‑mediated endocytic uptake is probably the best understood www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS and serves as a paradigm for how coated vesicles form. Clathrin acts to both concentrate adaptor proteins and cargo and, through polymerization into a lattice, stabilizes the budding membrane. Clathrin is implicated in the uptake of certain pathogens75. COPI and COPII are best known for their role in vesicle trafficking from the Golgi to the endoplasmic reticulum (ER) and from the ER to the Golgi, respectively73,74. Although, the possible involvement of the COPI and COPII in engulfment has not been extensively investigated in mammals, several D. melanogaster RNAi screens have implicated them in particle internalization19,21–25. The association of COPI and COPII with the D. melanogaster phagosome21 support a role for these proteins in phagocytosis and one can speculate that the COPI and COPII might help to establish the curvature that is required for membrane invagination (FIG. 1). An alternative, but not mutually exclusive, possibility is that COPI and COPII stabilize the curvature of the membrane of the vesicles (endosomes and lysosomes) with which phagosomes fuse during maturation76 (FIG. 1). The exocyst. The exocyst is an octodimeric complex consisting of the proteins Sec3, Sec5, Sec6, Sec8, Sec10, Sec15, Exo70 and Exo84, and was first identified as being assembled between the plasma membrane and secretory vesicles during neurotransmitter exocytosis77. This complex tethers the docking vesicle to the plasma membrane and hence facilitates membrane fusion mediated by soluble N‑ethylmaleimide-sensitive factor (NSF) attachment protein (SNAP) and soluble NSF attachment protein receptor (SNARE) complexes78. The exocyst is now known to be involved in diverse processes, such as the delivery of adhesion or signalling molecules to targeted sites on the plasma membrane and mobilizing membrane fragments from intracellular compartments to areas of rapid membrane remodelling, such as the leading edge of migratory cells79. Proteomic analysis of the D. melanogaster phagosome identified six of the eight known exocyst components, suggesting that the complex is also assembled on phagosomes, and this has been confirmed in mammalian cells21. Silencing of the exocyst components has been reported to decrease bacterial internalization and affect intracellular survival in several independent RNAi screens19,21,22,24, which supports the functional contribution of this complex to particle engulfment. The current model suggests that the exocyst is required to tether endosomes to the phagocytic cup21 or lysosomes to the maturing phagosome (FIG. 1), but it requires further validation.
Exocytosis The release of material contained within vesicles by fusion of the vesicles with the plasma membrane.
Nutrient acquisition and immune sensing The highly hydrolytic and digestive capacity of the phagosome is evident even in unicellular organisms, in which the phagosome is required to degrade bacteria and derive energy from the internalized microorganisms. To prevent them from being overwhelmed by the bacterial ‘meal’, the phagosome also destroys the internalized organisms, which suggests a primordial link between nutrient acquisition, pathogen destruction and innate
nature reviews | immunology
immune sensing (FIG. 5a,b). D. melanogaster has provided insight into how bacterial by-products are generated after phagocytosis and sensed by the immune system. Digestion of pathogens and PRR ligand generation. Recent work in D. melanogaster has suggested that exocytosis of bacterial-degradation products from the phagosome contributes to activation of the systemic innate immune response80. This has been suggested as a haemocyte-expressed lysosomal protein, Psidin, Other Cys- and CCXGacontaining domains EGF-like repeat (6 Cys) EGF-like repeat (8 Cys) EMI domain
Nimrod B1
Eater Nimrod A
LRP
Nimrod C1 Draper (CED-1)
Drosophila melanogaster
MEGF10
Holotrichia diomphalia
SREC SREC2 Homo sapiens
Figure 3 | An emerging superfamily of EGF-likeNature Reviewsand | Immunology repeat-containing phagocytic receptors opsonins. This figure depicts selected members of a superfamily of receptors, the extracellular domains of which contain an N‑terminus with a characteristic, cysteine-flanked CCXGa (where X denotes any amino acid and a denotes an aromatic amino acid) motif followed by multiple epidermal growth factor (EGF)-like repeats. The Emilin (EMI) domain is a dimerization domain101 that was first described in Caenorhabditis elegans CED‑1, a putative phagocytic receptor for apoptotic cells. Several members of this emerging superfamily of scavenger receptors have been implicated experimentally in the phagocytosis or clearance of microbial pathogens (Eater and Nimrod C1 in Drosophila melanogaster, lipopolysaccharide-recognition protein (LRP) in the large korean beetle, Holotrichia diomphalia) or of apoptotic cells (D. melanogaster Draper and Homo sapiens MEGF10 (multiple EGF-like domain 10), the respective homologues of C. elegans CED-1). Direct recognition of the microbial ligand lipopolysaccharide was demonstrated for LRP51 and for the four N‑terminal EGFlike repeats of Eater (C. Kocks and R.A.E., unpublished observations). SREC, scavenger receptor expressed by endothelial cells. volume 8 | february 2008 | 137
© 2008 Nature Publishing Group
REVIEWS was identified as contributing to systemic immune activation. One interpretation is that Psidin is required to generate bacterial degradation products that are then exocytosed from the phagolysosome into the haemolymph to amplify the innate immune response systemically by activating fat-body cells, the main source of circulating antimicrobial peptides. This a Dscam
DSCAM Immunoglobulin-like domain
b
Transmembrane domain
Fibronectin domain
DSCAM
BCR
D. melanogaster haemocyte
B cell
Escherichia coli
Antigen Plasma cell
Secreted DSCAM
Opsonization
DSCAM receptor?
Antibody
Opsonization
Uptake of pathogen FcR
Activated haemocytes
Mammalian macrophage
Mammalian B cell
Figure 4 | Insect DSCAM and mammalian immunoglobulins display common functional characteristics. a | Gene and protein structure of Drosophila Nature Reviewsmelanogaster | Immunology DSCAM (Down syndrome cell-adhesion molecule). Reminiscent of immunoglobulins, mutually exclusive alternative splicing occurs for exon clusters 4, 6, 9 and 17 to generate DSCAM receptor diversification with greater than 30,000 predicted splice variants, approximately 18,000 of which could be expressed by haemocytes and fat-body cells. A smaller 170 kDa form of DSCAM has been identified and is believed to represent the secreted form of the protein. Variation occurs in the second, third and seventh immunoglobulin domains. b | DSCAM and B-cell receptors (BCRs) in D. melanogaster and mammals, respectively, recognize foreign antigen and transmit signals to immune cells resulting in cell activation or cell survival. These signals also stimulate the production of secreted forms of DSCAM from activated D. melanogaster fat-body cells or haemocytes or, in mammals, immunoglobulin production by plasma cells. These molecules in turn act as opsonins, facilitating uptake through receptors on phagocytes. In mammals, antibody is recognized by Fc receptors (FcRs) but in D. melanogaster haemocytes it is unclear whether DSCAM-opsonized particles are recognized by a dedicated receptor or through homotypic interactions with membrane-bound DSCAM. In addition, both the BCR and DSCAM can act directly as endocytic or phagocytic receptors to internalize antigens or pathogens. 138 | february 2008 | volume 8
model suggests a potentially new mechanism for generating immunogenic ligands that are sensed by the innate immune system. Mammals also have a homologue of Psidin, and it will be of interest to determine whether phagosome-generated pathogen-associated degradation products also amplify systemic immune responses in mammals. PGRP-LE and cytosolic innate immune sensing. In addition to releasing bacterial derivatives back into the circulation, it is possible that the phagosome might also deliver them into the cytosol 81. In mammals, cytosolic bacterial fragments, such as muramyl dipeptide (MDP), are sensed by members of the nucleotidebinding domain, leucine-rich-repeat‑containing (NLR) family82–84, which also recognize intact intracellular pathogens or bacterial components that are introduced by the secretory apparatus associated with certain microorganisms. In D. melanogaster, a structurally unrelated member of the peptidoglycanrecognition protein (PGRP) family, PGRP-LE, in addition to functioning as an extracellular soluble receptor, can also function as an intracellular receptor that senses the bacterial derivative tracheal cytotoxin (TCT)85. How TCT accesses the cytosol is unknown and one possibility is that phagocytosis delivers the ligand to PGRP-LE . The functional equivalence between D. melanogaster PGRP-LE and mammalian NLRs suggest convergent evolution and it appears that these distantly related organisms have both evolved mechanisms to sense pathogen invasion into the cytosol. Interestingly, although a secreted soluble form of PGRP-LE exists, soluble NLRs have not been described. Secreted PGRP-LE can function in a cell-autonomous manner and may have a role as a co-receptor for PGRP-LC, which is expressed on the cell surface86, and hence function in a manner analogous to CD14 that cooperates with TLR4 to respond to lipopolysaccharide in mammals.
Lessons pertaining to adaptive immunity The adaptive immune response is characterized by the establishment of immunological memory made possible by clonal expansion of effector lymphocytes and somatic mutation and germline recombination of antigen receptors. Antigen presentation by antigenpresenting cells drives the clonal expansion of B and T cells and the retention of memory cells. Although D. melanogaster does not have a conventional adaptive immune response as defined by these characteristics (for a recent debate, see Ref. 87), evidence suggests that the immune system of D. melanogaster may have greater complexity than previously appreciated28 and that studies in flies might even provide some insights into mammalian adaptive immunity. Immune adaptation. Although the specializations of the immune system that define adaptive immunity were believed to be unique to jawed vertebrates, recent studies in lamprey have identified an alternative adaptive immune system64. However, studies in www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS a Nutrient acquisition by
b Phagocytosis and host defence
unicellular organisms
in mammals
Exocytosis Exocytosis
Nutrients Bacteria
TLR
Phagosome
NLR De novo protein synthesis
NF-κB
ER
Innate immune effectors
c Phagocytosis and antigen presentation
Nucleus
CD4+ T cell
CD8+ T cell TCR
MHC class II Antigen
Ub Ub Ub Immunoproteasome
Ub Ub Ub
TAP
Peptides
MHC class I
ER
Figure 5 | Nutrient acquisition is possibly the evolutionary origin of the link between phagocytosis and bacterial recognition. a | Phagocytosis is linked to Nature Reviews | Immunology nutrient acquisition by unicellular organisms. Internalized nutrients are degraded in a multistep process: enzymatic degradation; export of desired components from the phagosome into the cytosol; exported material may be further modified by proteasomal degradation; component amino acids are then used for protein synthesis; and undesired components are exported out of the cell by exocytosis. b | In higher species, phagocytosis has an essential role in innate immune sensing and degraded products of internalized bacteria may traffic and be modified as for nutrients. In host defence, the degraded products are sensed by pattern-recognition receptors (PRRs), such as Toll-like receptors (TLRs), in the phagosome lumen. Hypothetically, phagocytosis may also deliver ligands to cytosolic receptors, such as the nucleotide-binding domain, leucine-rich-repeatcontaining (NLR) family in mammals and the structurally unrelated peptidoglycanrecognition protein (PGRP-LE) in Drosophila melanogaster. c | In mammals, phagocytosis is essential for antigen presentation and adaptive immunity. Internalized antigens follow a similar path from the phagosome as nutrients. The antigenic peptides are then transported by transporter associated with antigen processing (TAP) either into the endoplasmic reticulum (ER) or back into the phagosome, and cross-presented on MHC class I molecules. D. melanogaster phagosomes contain much of the machinery required for antigen processing but lack the antigen-presentation machinery (that is, MHC and related molecules such as tapasin). These models suggest that the basic template of both innate immune sensing and antigen processing may be built on primordial functions of phagocytosis in nutrient absorption that existed before the acquisition of adaptive immunity. NF-κB, nuclear factor-κB; TCR, T-cell receptor; Ub, ubiquitin. nature reviews | immunology
invertebrates, such as D. melanogaster, have failed to conclusively demonstrate immune adaptation87. That said, it has been reported that resistance to certain pathogens can be transferred from mother to progeny in D. melanogaster, suggesting inheritable resistance88. In addition, recent data indicate that fruit flies may demonstrate immunological memory in response to certain pathogens89. This memory is sustained for the life of the flies and is specific, protecting only against the pathogen from the original infection. Importantly, this innate immune memory requires Toll but occurs independently of antimicrobial peptides, the key effectors induced by this pathway. Pertinent to this Review, the effector memory response requires phagocytes and can be blocked by paralysis of the phagocytic machinery. However, why memory is induced only by certain pathogens and the exact role of phagocytes in this context is unclear. Future work will need to define the molecular mechanism of this resistance before it can be conclusively said that D. melanogaster has the capacity for immune adaptation and whether this primitive adaptive immunity is the immunological forerunner of our own or whether it is independently evolved. Antigen presentation. In mammals, the phagosome is important for antigen presentation. Phagocytosed material is processed into antigenic peptides that are either loaded onto MHC class II molecules for presentation to CD4+ T cells or, by a currently poorly defined mechanism, access the endogenous pathway and are cross-presented on MHC class I molecules to activate CD8+ T cells. Intriguingly, the efficiency of cross-presentation is greatly increased in circumstances in which the antigen has been phagocytosed90,91. Although a subject of ongoing debate, it has been suggested that the contribution of the ER to the phagosome membrane facilitates this process, as machinery that is normally used for the dislocation of newly synthesized proteins from the ER is co-opted for phagosome-associated cross-presentation 92–95 (FIG. 5c). Analysis of D. melanogaster phagosomes also identified ER components and fly orthologues for many of the mammalian proteins required for antigen processing21. However, as this approach was identical to that used to isolate mammalian phagosomes and that originally implicated the ER as a source of phagosome membrane, we do not consider this independent verification. Nonetheless, this intriguing observation suggests that the basic machinery required for antigen processing might have been associated with phagocytosis before the evolutionary acquisition of adaptive immunity. If this is the case, it is likely to reflect a role for this machinery in nutrient acquisition (FIG. 5). Thus, the nidus for adaptive immunity may not simply have been the acquisition of recombinationactivating genes and immunoglobulin rearrangement but rather evolved in a modular way, with the early step of antigen processing made possible by building on an already existing function of phagocytosis in peptide and amino-acid absorption96. volume 8 | february 2008 | 139
© 2008 Nature Publishing Group
REVIEWS Concluding remarks and future directions Phagocytes have emerged as crucial guardians and central effector cells in host defence. More than a century ago, Metchnikoff ’s work emphasized the value of comparative embryology and established the validity of studying phagocytic engulfment in model organisms97. What is perhaps surprising is that similar model systems to those used by Metchnikoff in the nineteenth century have remained invaluable tools for the study of phagocytosis a 100 years later. These systems continue to provide insights into the complex cell biology of phagocytes and their importance in the evolutionary origins of immunity. We expect to see yet more advances from studies of D. melanogaster phagocytes by the application of cutting-edge technologies and systems-based approaches, 1. 2.
3. 4. 5.
6. 7.
8. 9.
10. 11. 12. 13.
14.
15.
Chen, G., Zhuchenko, O. & Kuspa, A. Immune-like phagocyte activity in the social amoeba. Science 317, 678–681 (2007). Tepass, U., Fessler, L. I., Aziz, A. & Hartenstein, V. Embryonic origin of hemocytes and their relationship to cell death in Drosophila. Development 120, 1829–1837 (1994). Zhou, L., Hashimi, H., Schwartz, L. M. & Nambu, J. R. Programmed cell death in the Drosophila central nervous system midline. Curr. Biol. 5, 784–790 (1995). Gordon, S. et al. Localization and function of tissue macrophages. Ciba Found. Symp. 118, 54–67 (1986). Hopkinson-Woolley, J., Hughes, D., Gordon, S. & Martin, P. Macrophage recruitment during limb development and wound healing in the embryonic and foetal mouse. J. Cell Sci. 107 (Pt 5), 1159–1167 (1994). Hoffmann, J. A., Kafatos, F. C., Janeway, C. A. & Ezekowitz, R. A. Phylogenetic perspectives in innate immunity. Science 284, 1313–1318 (1999). Vogel, G., Thilo, L., Schwarz, H. & Steinhart, R. Mechanism of phagocytosis in Dictyostelium discoideum: phagocytosis is mediated by different recognition sites as disclosed by mutants with altered phagocytotic properties. J. Cell Biol. 86, 456–465 (1980). Ezekowitz, R. A. et al. Uptake of Pneumocystis carinii mediated by the macrophage mannose receptor. Nature 351, 155–158 (1991). Kruskal, B. A., Sastry, K., Warner, A. B., Mathieu, C. E. & Ezekowitz, R. A. Phagocytic chimeric receptors require both transmembrane and cytoplasmic domains from the mannose receptor. J. Exp. Med. 176, 1673–1680 (1992). Aderem, A. & Underhill, D. M. Mechanisms of phagocytosis in macrophages. Annu. Rev. Immunol. 17, 593–623 (1999). Greenberg, S. & Grinstein, S. Phagocytosis and innate immunity. Curr. Opin. Immunol. 14, 136–145 (2002). Cardelli, J. Phagocytosis and macropinocytosis in Dictyostelium: phosphoinositide-based processes, biochemically distinct. Traffic 2, 311–320 (2001). Gumienny, T. L. & Hengartner, M. O. How the worm removes corpses: the nematode C. elegans as a model system to study engulfment. Cell Death Differ. 8, 564–568 (2001). Franc, N. C., Heitzler, P., Ezekowitz, R. A. & White, K. Requirement for croquemort in phagocytosis of apoptotic cells in Drosophila. Science 284, 1991–1994 (1999). This study shows that Croquemort is required for efficient phagocytosis of apopototic cell corpses in the D. melanogaster embryo in vivo and that its expression level is regulated by the amount of apoptosis in the embryo. Pearson, A. M. et al. Identification of cytoskeletal regulatory proteins required for efficient phagocytosis in Drosophila. Microbes Infect. 5, 815–824 (2003). This study describes the use of a genetic-deficiency screen in primary larval haemocytes ex vivo to identify several cytoskeletal proteins that are involved in phagocytosis. It demonstrates the use of D. melanogaster to address molecular mechanisms underlying bacterial phagocytosis and
including genomics and proteomics. D. melanogaster S2 cells will continue to be a preferred cell type for the high-throughput study of the complex cell biology of engulfment and to rapidly dissect the innate immune signalling pathways that are triggered by encounters with pathogens. In addition, using advanced imaging techniques in D. melanogaster embryos, it is now possible to obtain subcellular resolution that is sufficient to monitor organelle trafficking and to study phagosomes in vivo98,99. Understanding how the process of phagocytosis is tailored for and by its numerous and varied cargo remains a challenge. We anticipate that this model system will remain pivotal and, through its simplicity and genetic tractability, will continue to facilitate comparative innate immunity and provide new insights into the complex cell biology of phagocytosis and evolutionary origins of host defence.
also establishes the validity of using S2 cells for this purpose. 16. Silva, E., Au-Yeung, H. W., Van Goethem, E., Burden, J. & Franc, N. C. Requirement for a Drosophila E3ubiquitin ligase in phagocytosis of apoptotic cells. Immunity 27, 585–596 (2007). 17. Ramet, M. et al. Drosophila scavenger receptor CI is a pattern recognition receptor for bacteria. Immunity 15, 1027–1038 (2001). In this study, compounds related to scavengerreceptor ligands were used to inhibit phagocytosis of bacteria by S2 cells and led to the identification of a D. melanogaster scavenger receptor that is involved in this process. 18. Rabinowitz, S., Horstmann, H., Gordon, S. & Griffiths, G. Immunocytochemical characterization of the endocytic and phagolysosomal compartments in peritoneal macrophages. J. Cell Biol. 116, 95–112 (1992). 19. Ramet, M., Manfruelli, P., Pearson, A., Mathey-Prevot, B. & Ezekowitz, R. A. Functional genomic analysis of phagocytosis and identification of a Drosophila receptor for E. coli. Nature 416, 644–648 (2002). This reference describes the first high-throughput RNAi screen in S2 cells, identifying roles for the GATA transcription factor Serpent, COPI and COPII proteins and a member of the PGRP family (PGRP-LC) in bacterial phagocytosis in D. melanogaster. 20. Stuart, L. M. et al. Response to Staphylococcus aureus requires CD36-mediated phagocytosis triggered by the COOH-terminal cytoplasmic domain. J. Cell Biol. 170, 477–485 (2005). This study describes the use of a high-throughput RNAi screen in S2 cells that led to the investigation of the role of mammalian CD36 in the phagocytosis and response to S. aureus infection. 21. Stuart, L. M. et al. A systems biology analysis of the Drosophila phagosome. Nature 445, 95–101 (2007). This paper describes a systems-biology approach combining proteomics, bioinformatics and highthroughput RNAi that provided the first comprehensive analysis of phagosomes. It highlights novel and unexpected roles in phagocytosis for COPI and COPII and exocyst complexes. 22. Philips, J. A., Rubin, E. J. & Perrimon, N. Drosophila RNAi screen reveals CD36 family member required for mycobacterial infection. Science 309, 1251–1253 (2005). A genome-wide RNAi screen in S2 cells identified the D. melanogaster gene peste, a CD36 family member, and paved the way for the finding that CD36 family members serve as receptors for mycobacterial entry into human cells. 23. Stroschein-Stevenson, S. L., Foley, E., O’Farrell P, H. & Johnson, A. D. Identification of Drosophila gene products required for phagocytosis of Candida albicans. PLoS Biol. 4, e4 (2005). This study describes a genome-wide RNAi screen for evolutionarily conserved D. melanogaster genes involved in the recognition and phagocytosis of C. albicans that establishes various TEPs as microbial opsonins that can distinguish C. albicans (MCR), E. coli (TEPII) and S. aureus (TEPIII).
140 | february 2008 | volume 8
24. Agaisse, H. et al. Genome-wide RNAi screen for host factors required for intracellular bacterial infection. Science 309, 1248–1251 (2005). This study describes the use of a genome-wide RNAi screen in S2 cells to address commonalities and differences in host cell invasion and intracellular growth by bacteria with different cellular replication sites. 25. Cheng, L. W. et al. Use of RNA interference in Drosophila S2 cells to identify host pathways controlling compartmentalization of an intracellular pathogen. Proc. Natl Acad. Sci. USA 102, 13646–13651 (2005). In this study three genome-wide RNAi screens in S2 cells were used to elucidate host molecules involved in different facets of cell parasitism (entry, vacuolar escape and intracellular growth) by the bacterial pathogen L. monocytogenes. 26. Ayres, J. S. & Schneider, D. S. Genomic dissection of microbial pathogenesis in cultured Drosophila cells. Trends Microbiol. 14, 101–104 (2006). 27. Lemaitre, B. & Hoffmann, J. The host defense of Drosophila melanogaster. Annu. Rev. Immunol. 25, 697–743 (2007). 28. Brennan, C. A. & Anderson, K. V. Drosophila: the genetics of innate immune recognition and response. Annu. Rev. Immunol. 22, 457–483 (2004). 29. Elrod-Erickson, M., Mishra, S. & Schneider, D. Interactions between the cellular and humoral immune responses in Drosophila. Curr. Biol. 10, 781–784 (2000). In this study functional ablation of phagocytes by polystyrene beads was used to demonstrate that humoral defence mechanisms act together with phagocytosis to generate effective immune responses in D. melanogaster. 30. Kocks, C. et al. Eater, a transmembrane protein mediating phagocytosis of bacterial pathogens in Drosophila. Cell 123, 335–346 (2005). This study establishes that phagocytosis is a cellular host defence mechanism in D. melanogaster. Transcriptional profiling and highthroughput RNAi in S2 cells was used to identify a new type of EGF‑like‑repeat-containing receptor (Eater) that has a crucial role in the host defence against bacterial infections in vivo. 31. Williams, M. J., Wiklund, M. L., Wikman, S. & Hultmark, D. Rac1 signalling in the Drosophila larval cellular immune response. J. Cell Sci. 119, 2015–2024 (2006). 32. Williams, M. J., Ando, I. & Hultmark, D. Drosophila melanogaster Rac2 is necessary for a proper cellular immune response. Genes Cells 10, 813–823 (2005). 33. Avet-Rochex, A., Perrin, J., Bergeret, E. & Fauvarque, M. O. Rac2 is a major actor of Drosophila resistance to Pseudomonas aeruginosa acting in phagocytic cells. Genes Cells 12, 1193–1204 (2007). 34. Nehme, N. T. et al. A model of bacterial intestinal infections in Drosophila melanogaster. PLoS Pathog. 3, e173 (2007). 35. Garin, J. et al. The phagosome proteome: insight into phagosome functions. J. Cell Biol. 152, 165–180 (2001).
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS 36. Desjardins, M. ER‑mediated phagocytosis: a new membrane for new functions. Nature Rev. Immunol. 3, 280–291 (2003). 37. Desjardins, M., Huber, L. A., Parton, R. G. & Griffiths, G. Biogenesis of phagolysosomes proceeds through a sequential series of interactions with the endocytic apparatus. J. Cell Biol. 124, 677–688 (1994). 38. Meresse, S. et al. Controlling the maturation of pathogen-containing vacuoles: a matter of life and death. Nature Cell Biol. 1, E183–188 (1999). 39. Lagueux, M., Perrodou, E., Levashina, E. A., Capovilla, M. & Hoffmann, J. A. Constitutive expression of a complement-like protein in toll and JAK gain‑of‑function mutants of Drosophila. Proc. Natl Acad. Sci. USA 97, 11427–11432 (2000). 40. Moita, L. F. et al. In vivo identification of novel regulators and conserved pathways of phagocytosis in A. gambiae. Immunity 23, 65–73 (2005). 41. Janeway, C. A. Jr. Approaching the asymptote? Evolution and revolution in immunology. Cold Spring Harb. Symp. Quant. Biol. 54, 1–13 (1989). 42. Gordon, S. Pattern recognition receptors: doubling up for the innate immune response. Cell 111, 927–930 (2002). 43. Wright, S. D., Ramos, R. A., Tobias, P. S., Ulevitch, R. J. & Mathison, J. C. CD14, a receptor for complexes of lipopolysaccharide (LPS) and LPS binding protein. Science 249, 1431–1433 (1990). 44. Hoebe, K. et al. CD36 is a sensor of diacylglycerides. Nature 433, 523–527 (2005). 45. Savill, J., Hogg, N., Ren, Y. & Haslett, C. Thrombospondin cooperates with CD36 and the vitronectin receptor in macrophage recognition of neutrophils undergoing apoptosis. J. Clin. Invest. 90, 1513–1522 (1992). 46. Franc, N. C., Dimarcq, J. L., Lagueux, M., Hoffmann, J. & Ezekowitz, R. A. Croquemort, a novel Drosophila hemocyte/macrophage receptor that recognizes apoptotic cells. Immunity 4, 431–443 (1996). This paper describes the D. melanogaster receptor Croquemort, a scavenger receptor that is related to mammalian CD36, and shows that it binds to and mediates the uptake of apoptotic cells. 47. Pearson, A., Lux, A. & Krieger, M. Expression cloning of dSR-CI, a class C macrophage-specific scavenger receptor from Drosophila melanogaster. Proc. Natl Acad. Sci. USA 92, 4056–4060 (1995). 48. Irving, P., Troxler, L. & Hetru, C. Is innate enough? The innate immune response in Drosophila. C. R. Biol. 327, 557–570 (2004). 49. Lazzaro, B. P. Elevated polymorphism and divergence in the class C scavenger receptors of Drosophila melanogaster and D. simulans. Genetics 169, 2023–2034 (2005). 50. Kurucz, E. et al. Nimrod, a putative phagocytosis receptor with EGF repeats in Drosophila plasmatocytes. Curr. Biol. 17, 649–654 (2007). 51. Ju, J. S. et al. A novel 40-kDa protein containing six repeats of an epidermal growth factor-like domain functions as a pattern recognition protein for lipopolysaccharide. J. Immunol. 177, 1838–1845 (2006). 52. Freeman, M. R., Delrow, J., Kim, J., Johnson, E. & Doe, C. Q. Unwrapping glial biology: Gcm target genes regulating glial development, diversification, and function. Neuron 38, 567–580 (2003). 53. Manaka, J. et al. Draper-mediated and phosphatidylserine-independent phagocytosis of apoptotic cells by Drosophila hemocytes/macrophages. J. Biol. Chem. 279, 48466–48476 (2004). 54. Awasaki, T. et al. Essential role of the apoptotic cell engulfment genes draper and ced‑6 in programmed axon pruning during Drosophila metamorphosis. Neuron 50, 855–867 (2006). 55. MacDonald, J. M. et al. The Drosophila cell corpse engulfment receptor Draper mediates glial clearance of severed axons. Neuron 50, 869–881 (2006). 56. Su, H. P. et al. Interaction of CED‑6/GULP, an adapter protein involved in engulfment of apoptotic cells with CED‑1 and CD91/low density lipoprotein receptorrelated protein (LRP). J. Biol. Chem. 277, 11772–11779 (2002). 57. Hamon, Y. et al. Cooperation between engulfment receptors: the case of ABCA1 and MEGF10. PLoS ONE 1, e120 (2006). 58. Schmucker, D. & Flanagan, J. G. Generation of recognition diversity in the nervous system. Neuron 44, 219–222 (2004). 59. Schmucker, D. et al. Drosophila Dscam is an axon guidance receptor exhibiting extraordinary molecular diversity. Cell 101, 671–684 (2000).
60. Watson, F. L. et al. Extensive diversity of Igsuperfamily proteins in the immune system of insects. Science 309, 1874–1878 (2005). This paper describes the hypervariable, neuronal immunoglobulin-superfamily receptor DSCAM as being abundantly expressed in the immune tissues of D. melanogaster and implicates thousands of alternative splice forms as PRRs and opsonins in the phagocytosis of microorganisms. 61. Meijers, R. et al. Structural basis of Dscam isoform specificity. Nature 449, 487–491 (2007). 62. Neves, G., Zucker, J., Daly, M. & Chess, A. Stochastic yet biased expression of multiple Dscam splice variants by individual cells. Nature Genet. 36, 240–246 (2004). 63. Dong, Y., Taylor, H. E. & Dimopoulos, G. AgDscam, a hypervariable immunoglobulin domain-containing receptor of the Anopheles gambiae innate immune system. PLoS Biol. 4, e229 (2006). 64. Pancer, Z. et al. Somatic diversification of variable lymphocyte receptors in the agnathan sea lamprey. Nature 430, 174–180 (2004). 65. Caron, E. & Hall, A. Identification of two distinct mechanisms of phagocytosis controlled by different Rho GTPases. Science 282, 1717–1721 (1998). 66. Hall, A. B. et al. Requirements for Vav guanine nucleotide exchange factors and Rho GTPases in FcγRand complement-mediated phagocytosis. Immunity 24, 305–316 (2006). 67. Olazabal, I. M. et al. Rho-kinase and myosin-II control phagocytic cup formation during CR, but not FcγR, phagocytosis. Curr. Biol. 12, 1413–1418 (2002). 68. Colucci-Guyon, E. et al. A role for mammalian diaphanous-related formins in complement receptor (CR3)-mediated phagocytosis in macrophages. Curr. Biol. 15, 2007–2012 (2005). 69. Castellano, F., Chavrier, P. & Caron, E. Actin dynamics during phagocytosis. Semin. Immunol. 13, 347–355 (2001). 70. Etienne-Manneville, S. & Hall, A. Rho GTPases in cell biology. Nature 420, 629–635 (2002). 71. Swanson, J. A. & Hoppe, A. D. The coordination of signaling during Fc receptor-mediated phagocytosis. J. Leukoc. Biol. 76, 1093–1103 (2004). 72. Antonny, B. Membrane deformation by protein coats. Curr. Opin. Cell Biol 18, 386–394 (2006). 73. Gurkan, C., Stagg, S. M., Lapointe, P. & Balch, W. E. The COPII cage: unifying principles of vesicle coat assembly. Nature Rev. Mol. Cell Biol. 7, 727–738 (2006). 74. Lippincott-Schwartz, J. & Liu, W. Insights into COPI coat assembly and function in living cells. Trends Cell Biol. 16, e1–e4 (2006). 75. Veiga, E. & Cossart, P. The role of clathrin-dependent endocytosis in bacterial internalization. Trends Cell Biol. 16, 499–504 (2006). 76. Botelho, R. J., Hackam, D. J., Schreiber, A. D. & Grinstein, S. Role of COPI in phagosome maturation. J. Biol. Chem. 275, 15717–15727 (2000). 77. TerBush, D. R., Maurice, T., Roth, D. & Novick, P. The Exocyst is a multiprotein complex required for exocytosis in Saccharomyces cerevisiae. EMBO J. 15, 6483–6494 (1996). 78. Boyd, C., Hughes, T., Pypaert, M. & Novick, P. Vesicles carry most exocyst subunits to exocytic sites marked by the remaining two subunits, Sec3p and Exo70p. J. Cell Biol. 167, 889–901 (2004). 79. Clandinin, T. R. Surprising twists to exocyst function. Neuron 46, 164–166 (2005). 80. Brennan, C. A., Delaney, J. R., Schneider, D. S. & Anderson, K. V. Psidin is required in Drosophila blood cells for both phagocytic degradation and immune activation of the fat body. Curr. Biol. 17, 67–72 (2007). 81. Herskovits, A. A., Auerbuch, V. & Portnoy, D. A. Bacterial ligands generated in a phagosome are targets of the cytosolic innate immune system. PLoS Pathog. 3, e51 (2007). 82. Martinon, F., Burns, K. & Tschopp, J. The inflammasome: a molecular platform triggering activation of inflammatory caspases and processing of proIL-β. Mol. Cell 10, 417–426 (2002). 83. Ogura, Y. et al. Nod2, a Nod1/Apaf‑1 family member that is restricted to monocytes and activates NF‑κB. J. Biol. Chem. 276, 4812–4818 (2001). 84. Inohara, N., Ogura, Y., Chen, F. F., Muto, A. & Nunez, G. Human Nod1 confers responsiveness to bacterial lipopolysaccharides. J. Biol. Chem. 276, 2551–2554 (2001). 85. Kaneko, T. et al. PGRP-LC and PGRP-LE have essential yet distinct functions in the Drosophila immune
nature reviews | immunology
response to monomeric DAP-type peptidoglycan. Nature Immunol. 7, 715–723 (2006). 86. Takehana, A. et al. Peptidoglycan recognition protein (PGRP)-LE and PGRP-LC act synergistically in Drosophila immunity. EMBO J. 23, 4690–4700 (2004). 87. Hauton, C. & Smith, V. J. Adaptive immunity in invertebrates: a straw house without a mechanistic foundation. Bioessays 29, 1138–1146 (2007). 88. Boman, H. G., Nilsson, I. & Rasmuson, B. Inducible antibacterial defence system in Drosophila. Nature 237, 232–235 (1972). 89. Pham, L. N., Dionne, M. S., Shirasu-Hiza, M. & Schneider, D. S. A specific primed immune response in Drosophila is dependent on phagocytes. PLoS Pathog. 3, e26 (2007). 90. Kovacsovics-Bankowski, M. & Rock, K. L. A phagosome‑to‑cytosol pathway for exogenous antigens presented on MHC class I molecules. Science 267, 243–246 (1995). 91. Kovacsovics-Bankowski, M., Clark, K., Benacerraf, B. & Rock, K. L. Efficient major histocompatibility complex class I presentation of exogenous antigen upon phagocytosis by macrophages. Proc. Natl Acad. Sci. USA 90, 4942–4946 (1993). 92. Houde, M. et al. Phagosomes are competent organelles for antigen cross-presentation. Nature 425, 402–406 (2003). 93. Ackerman, A. L., Kyritsis, C., Tampe, R. & Cresswell, P. Early phagosomes in dendritic cells form a cellular compartment sufficient for cross presentation of exogenous antigens. Proc. Natl Acad. Sci. USA 100, 12889–12894 (2003). 94. Guermonprez, P. et al. ER‑phagosome fusion defines an MHC class I cross-presentation compartment in dendritic cells. Nature 425, 397–402 (2003). 95. Touret, N. et al. Quantitative and dynamic assessment of the contribution of the ER to phagosome formation. Cell 123, 157–170 (2005). 96. Desjardins, M., Houde, M. & Gagnon, E. Phagocytosis: the convoluted way from nutrition to adaptive immunity. Immunol. Rev. 207, 158–165 (2005). 97. Metchnikoff, I. I. On the present state of the question of immunity in infectious disease [online]
(1908). 98. Wood, W. et al. Wound healing recapitulates morphogenesis in Drosophila embryos. Nature Cell Biol. 4, 907–912 (2002). 99. Stramer, B. et al. Live imaging of wound inflammation in Drosophila embryos reveals key roles for small GTPases during in vivo cell migration. J. Cell Biol. 168, 567–573 (2005). 100. Rizki, T. M. & Rizki, R. M. The cellular defense system of Drosophila melanogaster. (eds King, R. C. & Akai, H.) 579–604 (Plenum Publishing Corporation, New York, 1984). An insightful book chapter with stunning electron micrographs describing D. melanogaster blood cells and their possible roles in pathogen defence. 101. Callebaut, I., Mignotte, V., Souchet, M. & Mornon, J. P. EMI domains are widespread and reveal the probable orthologs of the Caenorhabditis elegans CED‑1 protein. Biochem. Biophys. Res. Commun. 300, 619–623 (2003).
Acknowledgements
We thank members of the laboratory of Developmental Immunology, Massachusetts General Hospital, for continuous inspiration. Our particular thanks go to C. Kocks for help with the original figure 3 and for critical reading and helpful input into the manuscript. Finally, we apologize to our colleagues whose work we have been unable to cite or discuss owing to space constraints.
DATABASES Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene CD36 | CD91 | CED-1 | MEGF10 | MEGF11 FlyBase: http://flybase.bio.indiana.edu/ Croquemort | Draper | DSCAM | Eater | MCR | Nimrod C1 | Peste | PGRP-LE | SR‑CI
FURTHER INFORMATION Lynda M. Stuart’s homepage: http://ccib.mgh.harvard.edu/faculty-stuart.htm Laboratory of Developmental Immunology: http://www.mgh.harvard.edu/devimmunol/index.htm The Interactive Fly: http://www.sdbonline.org/fly/aimain/1aahome.htm All links are active in the online pdf
volume 8 | february 2008 | 141 © 2008 Nature Publishing Group
REVIEWS
Regulation of immunological homeostasis in the respiratory tract Patrick G. Holt*, Deborah H. Strickland*, Matthew E. Wikström* and Frode L. Jahnsen‡
Abstract | The respiratory tract has an approximate surface area of 70 m2 in adult humans, which is in virtually direct contact with the outside environment. It contains a uniquely rich vascular bed containing a large pool of marginated T cells, and harbours a layer of single-cellthick epithelial tissue through which re-oxygenation of blood must occur uninterrupted for survival. It is therefore not surprising that the respiratory tract is never more than a short step away from disaster. We have only a partial understanding of how immunological homeostasis is maintained in these tissues, but it is becoming clear that the immune system has evolved a range of specific mechanisms to deal with the unique problems encountered in this specialized microenvironment. Epithelial cells Cells that line all tissues and act to protect them by regulating or resisting the passage of exogenous matter.
Secretory goblet cells Mucus-secreting cells within the airway epithelium.
*Telethon Institute for Child Health Research, Centre for Child Health Research, The University of Western Australia, Perth 6009, Western Australia. ‡ LIIPAT, Centre for Immune Regulation, Institute of Pathology, University of Oslo and Division of Pathology, Rikshospitalet, N‑0027 Oslo, Norway. Correspondence to P.G.H. e-mail: [email protected] doi:10.1038/nri2236 Published online 21 January 2008
Respiratory epithelial-cell surfaces present a large frag ile interface with the external environment, which is exposed continuously to a broad range of antigens during respiration. Maintenance of local immunological homeo stasis and hence the integrity of these gas-exchange surfaces stretches the discriminatory powers of the immune system to their limits, as incoming antigens are dominated by highly immunogenic but harmless pro teins of plant and animal origin which, if they evoked efficient adaptive immune responses, would doom the host to premature death from chronic airway inflam mation. The balancing act that the respiratory mucosal immune system must perform involves discrimination of this background antigenic noise from the much rarer signals transmitted by pathogen-associated antigens. Not only must the immune system deal with this low signal-to-noise ratio problem, but having selected anti gens for T‑cell priming, it must tightly regulate ensuing T‑cell memory responses to minimize damage to local epithelial-cell surfaces. This is of particular importance in relation to alveolar gas-exchange surfaces, as these tissues contain the largest vascular bed in the body and function as a magnet for circulating memory T cells. The key mechanisms through which the local immune system performs this balancing act is the theme of this Review. In particular, we focus on the microanatomical organization of local elements of the innate and adap tive immune systems, the cellular dynamics of immune induction in this microenvironment and the roles of indi vidual cell types in fine tuning local immune responses. Important areas not addressed in this Review owing to
142 | february 2008 | volume 8
space limitations include the role of neuropeptides in local immunoregulation, control of wound healing and/or tis sue repair in response to chronic inflammation and the potential future impact of advances in stem-cell biology.
Organization of the immune system in the lungs The lungs can be divided into two functionally distinct compartments: the conducting airways overlayed by mucosal tissue, and the lung parenchyma, which comprises thin-walled alveoli that are specialized for gas exchange (FIG. 1). Distinct populations of immune cells reside in these adjacent areas, reflecting the dif fering functions of local tissues, as well as the differing levels of exposure to airborne antigens throughout the respiratory tree. Conducting airways. The respiratory epithelium of the airway mucosa comprises ciliated cells and secretory goblet cells that, together with locally produced secreted IgA, provide mechanisms for mucociliary clearance of inhaled antigens. The mucosa contains dense networks of dendritic cells (DCs) and macrophages that develop early in life 1. Within these dense networks, the DC pool is composed of both myeloid DCs and plasmacytoid DCs (pDCs), but the myeloid DC sub set generally dominates, particularly in the airway mucosa 2. Resident airway mucosal DCs (AMDCs) are specialized for immune surveillance by antigen acquisition but lack the ability to efficiently present antigen3. They are strategically positioned for antigen uptake both within and directly beneath the surface www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS Respiratory epithelium
Trachea
Intraepithelial DC
Plasma cell
Bronchus Airborne antigen
Conducting airways
Gland Macrophage T cell
Bronchus Alveolar duct
DC
a Alveolus
Lung parenchyma
Parenchymal DC Parenchymal T cell
Lung homing
Afferent lymphatics
d
Alveolar space
c
b
Naive T cell Afferent lymphatics Systemic distribution
Plasmacytoid dendritic cells (pDCs). A population of cells with a plasma-cell-like morphology that produce high levels of type I interferons after exposure to viruses. Human pDCs express high levels of CD123, the interleukin‑3 receptor α‑chain, and depend on interleukin‑3 as a growth factor.
Lamina propria Loose connective tissue that is located immediately under the airway epithelium.
Mast cells A leukocyte population that secretes histamine and other inflammatory mediators on antibody crosslinking of its IgE receptors, and that is largely responsible for acute manifestations of the allergic response.
Plasma cells Antibody-secreting cells that are generated from antigenspecific B cells.
e
Efferent lymphatics
Draining lymph node
Blood vessel Type II epithelial cell Type I epithelial cell
Figure 1 | Antigen uptake and migratory patterns for immune induction in the lungs. Local immune cells in the two lung compartments showing capture of airborne antigens and subsequent recognition by T cells in the draining lymph nodes. Luminal antigens are sampled by dendritic cells (DCs) that are located within the surface epithelium of the bronchial mucosa (a) or in the alveoli (b). Antigen-bearing DCs upregulate CC‑chemokine receptor 7 and migrate through Nature Reviews | Immunology the afferent lymphatics to the draining lymph nodes and present antigenic peptides to naive antigen-specific T cells (c). Activated T cells proliferate and migrate through the efferent lymphatics and into the blood via the thoracic duct. Depending on their tissue-homing receptor profile, effector T cells will exit into the bronchial mucosa through postcapillary venules in the lamina propria or through the pulmonary capillaries in the lung parenchyma (d), or disseminate from the bloodstream throughout the peripheral immune system (for example, to other mucosal sites) (e).
epithelium, and extend protrusions into the airway lumen4, similar to what has been reported for DCs in intestinal tissue 5 (FIG. 2). This suggests that AMDCs can sample directly from the airway luminal sur face through the intact epithelium4. T cells are also found in relatively high numbers in the mucosa, both intraepithelially and within the underlying lamina propria (FIG. 3). As in the gut, most intraepithelial T cells express CD8, whereas CD4+ T cells are more frequently found in the lamina propria. Both subsets mainly have an effector- and/or memory-cell phenotype, as defined by their expression of CD45RO. The lamina propria also contains mast cells (BOX 1) and plasma cells (mainly producing polymeric IgA) and some scattered B cells. Aside from their central role in antibody production, it is possible that B cells also contribute to local antigen presentation, given the recent demonstration of such a function for B cells in the lymph nodes that drain the lungs6.
nature reviews | immunology
In addition to effector-cell populations, the airway mucosa also contains potential inductive sites known as bronchial-associated lymphoid tissue (BALT). Classical BALT comprises discrete lymphoid-cell aggregates under lying a specialized epithelium, analogous to tonsillar tis sue and Peyer’s patches. The presence of BALT differs between species and its importance in adult humans has been questioned, however, BALT is common in young children and mostly contains isolated lymphoid follicles (I. Heier, K. Malmström, A. S. Pelkonen, L. P. Malmberg, M. Kajosaari, M. Turpeinen, H. Lindahl, P. Brandtzaeg, F.L.J. and M. J. Mäkelä, unpublished observations). Importantly, the inducible BALT from mice that lack other organized lymphoid tissue can generate protective immunity against pathogens, such as influenza virus7. This suggests that BALT may have a significant role in local immunological homeostasis within the respiratory tract early in life when important elements within central lymphoid structures are functionally immature. volume 8 | february 2008 | 143
© 2008 Nature Publishing Group
REVIEWS Anergized cell A cell that is characterized by its weak response to normal stimuli. An anergized T cell is unable to produce large amounts of interleukin‑2 or proliferate vigorously when stimulated via CD3 or its T‑cell receptor.
Mucociliary elevator Upward transport of mucus stream from the lungs by ciliated epithelial cells.
Pattern-recognition receptors (PRRs). Host receptors (such as Toll-like receptors) that are able to sense pathogen-associated molecular patterns and initiate signalling cascades (involving activation of nuclear factor-κB) that lead to an innate immune response.
Parenchymal lung. Progressive branching of the bronchi gives rise to bronchioles that extend to alveolar ducts that further branch into blind-ended alveolar sacs (FIG. 1). The alveoli are separated by thin walls of inter stitium containing pulmonary capillaries that are in close contact with the alveolar space, and some stromal cells (FIG. 1). Immune cells in the lung parenchyma are located both above the alveolar epithelium in the terminal airways and in the underlying parenchyma. Under steady-state conditions, the leukocyte popula tion in the alveolar space is dominated by alveolar macrophages (more than 90% of the total cell popula tion), the remainder being mainly DCs and T cells. The lung parenchyma also contains scattered macro phages, DCs and T cells, as well as B cells and mast cells, but no plasma cells. In addition, large numbers of T cells are sequestered in the vascular bed of the lung parenchyma. The contribution of this large population of cells to local immunological homeostasis is unclear, but many appear to be ‘anergized’ and might represent end-stage post-activated memory cells en route for removal through the liver or the mucociliary elevator8.
Immune induction in the lung: cellular dynamics Respiratory mucosal surfaces, particularly in the upper conducting airways, are chronically exposed to myriad non-pathogenic environmental antigens. To protect against the potential immunopathological consequences of continuously responding to these ubiquitous stimuli, the local ‘default’ immune response takes the form of non-inflammatory, low-level T helper 2 (T H2)-cell immunity3 and/or a form of T‑cell-mediated immuno logical tolerance9,10. This response pattern reiterates that of the chronically stimulated gastric mucosa11. The a Small intestinal mucosa Intestinal epithelial cell
b Bronchial mucosa
Luminal antigen Tight junction
Airway epithelial cell
AMDC
Capillary vessel
Basement membrane
DC Draining lymph node
Draining lymph node
Figure 2 | A model for antigen sampling at mucosal sites. a | Resident dendritic cells (DCs) that are located beneath the microvasculature in the small intestinal extend Nature Reviewsvilli | Immunology cellular projections around the subepithelial vessels and up between the intestinal epithelial cells5. These cellular protrusions enable them to directly sample antigens on the luminal side. The DCs preserve the integrity of the epithelial-cell barrier by expressing tight-junction proteins. After antigen uptake the DCs upregulate CC‑chemokine receptor 7 (CCR7), which allows them to migrate to the draining lymph nodes. b | Analogous to intestinal resident DCs, resident airway mucosal DCs (AMDCs) that are situated within the epithelial compartment in the respiratory mucosa (both in rodents and humans) extend cellular projections into the airway lumen. Although direct evidence for the uptake of luminal antigens by AMDCs is still lacking, this model provides a plausible mechanism for continuous immune surveillance of intact respiratory mucosal surfaces.
144 | february 2008 | volume 8
mechanism(s) underlying this protective default are only partially understood, but appear to be under the control of local DC subsets3,10,12–14. Moreover, there is an additional level of backup control to regulate the inten sity of ‘breakthrough’ memory T‑cell responses, which evade this protective default, in the form of the potent T‑cell-inhibitory activity of lung macrophages8,15,16. This mechanism is particularly active in the lung paren chyma through the immunosuppressive functions of resident alveolar macrophages17,18, providing a last line of defence for alveolar gas-exchange surfaces against T‑cell-mediated inflammation. The identification of potentially pathogenic anti gens is facilitated through pattern-recognition receptors, such as Toll-like receptors (TLRs), that are expressed by the network of sentinel AMDCs (TABLE 1) , and results in effective bypassing of the T H2-cell default and the subsequent generation of effector memory responses against these antigens. It is of interest to note, however, that some (probably low) level of TLR stimulation of DCs in the airway mucosa appears to be obligatory to enable initial recognition per se of inert inhaled antigens19. Qualitative and quantitative aspects of the ensuing immune responses to such antigens (including tolerance induction) might be determined in part by the level of co-exposure to TLR-stimulatory agents, such as lipopolysaccharide (LPS)20, which are ubiquitous in ambient air. Tissue-specific lymphocyte homing. As noted above, surveillance of airway surfaces for these different classes of antigen is carried out primarily through the AMDC network, which continuously shuttles incoming antigens to regional lymph nodes (RLNs), which are the primary sites for the induction of immunological memory against inhaled antigens (FIG. 1). Following activation, subsequent to their migration to the RLNs, these DCs develop a potent capacity to prime naive T cells and initi ate immune responses. Moreover, the resulting memory T cells have an enhanced capacity for selective migration back to the original challenge site — the concept of tissuespecific homing21. Whereas naive T cells express adhesion molecules and chemokine receptors that restrict their migration principally (but not entirely22) to organized lymphoid tissue, activated memory T cells downregulate these lymphoid-tissue-homing receptors and upregu late tissue-specific adhesion molecules and chemokine receptors that target their migration to non-lymphoid tissues21. This imprinting of tissue-homing properties is best described for the gut and skin. Priming of T and B cells in Peyer’s patches and mesenteric lymph nodes preferentially induces the expression of α4β7-integrin and CC‑chemokine receptor 9 (CCR9)23,24, whereas T cells that are primed in peripheral lymph nodes upregulate cutaneous leukocyte antigen, CCR4 (REF. 25) and CCR10 (REF. 26) (FIG. 4). Importantly, recent results suggest that antigen-presenting DCs process and ‘interpret’ locally produced metabolites to programme tissue-specific lym phocyte homing. In the case of gut-associated lymphoid tissue (GALT), resident DCs metabolize vitamin A to retinoic acid, which stimulates α4β7-integrin and CCR9 www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS
Figure 3 | Distribution of DCs and T cells in the tracheal mucosa. Cryosection from a rat trachea (1 µm Nature Reviews | Immunology optical section) showing immunofluorescence staining of dendritic cells (DCs; MHC class II stained green) and T cells (CD2 stained red). Note that most of the T cells are in close contact with DCs. The dotted line represents the basement membrane of the surface epithelium. Magnification ×600.
expression by T cells23; and in the skin, local DCs use metabolites of vitamin D3 to programme T cells in RLNs for epidermotropism27. Moreover, DCs resident in GALT that metabolize vitamin A also induce the generation of gut-homing IgA-secreting B cells28.
Epidermotropism Movement towards the epidermis.
Lymphocyte homing to the lungs. Although there are studies that suggest that tissue-specific T‑cell homing is also operative in the respiratory tract, distinct hom ing phenotypes have yet to be defined29. Moreover, cell recruitment to the lungs is unique in that it has two sepa rate circulatory systems — the bronchial arteries from the systemic circulation that nourish the bronchial wall and the low-pressure pulmonary system that circulates through the lung parenchyma. The mucosa of the cen tral airways is part of the common mucosal-associated lymphoid tissue (MALT), a system that integrates the immune responses of distinct mucosal tissues. However, although MALTs from different sites share several prop erties (such as IgA responses), there are distinct tissuetrafficking patterns for both B and T cells that depend on their site of induction (FIG. 4). For example, plasma-cell precursors that are primed in respiratory-tract lym phoid tissues that home to the tracheal and bronchial mucosa express only low levels of the gut-homing mol ecules α4β7-integrin and CCR9, but express high levels of α4β1-integrin and CCR10 (REF. 24). Importantly, the counterparts of α4β1-integrin and CCR10, vascular celladhesion molecule 1 (VCAM1) and CC‑chemokine ligand 28 (CCL28), respectively, are constitutively expressed by airway mucosal endothelial cells24. Lung T cells also express a phenotype that is distinct from gut-homing T cells30. In the lung parenchyma, leukocytes migrate mainly through pulmonary capillaries. As a result, the concept of a coordinated multistep migratory process through the postcapillary venules that occurs in most tissues is not always required for extravasation in the lungs.
nature reviews | immunology
Although the recirculation of memory T cells through the lung parenchyma appears to be less tightly regulated than in the intestine, several lines of evidence suggest that the egress of T cells into the lung parenchyma is, in some cases, a regulated process. In human and mouse lungs, it has been shown that memory CD8+ T cells that are specific for respiratory viruses selec tively accumulate in the lung parenchyma 29. Under non-inflammatory conditions, endothelial cells in the pulmonary vasculature express intercellular adhesion molecule 1 (ICAM1) and P‑selectin, but not VCAM1. Galkina and co-workers 31 showed that preferential vascular retention and egress of effector CD8+ T cells to the normal mouse lung was mediated by lymphocyte function-associated antigen 1 (LFA1)–ICAM1 and CCR5–CCR5-ligand interactions. Also, inhibition of P‑selectin glycoprotein ligand 1 (PSGL1), which binds P‑selectin, reduced the number of T cells in the alveoli and lung parenchyma. Kallinich et al.32 have shown that CXC-chemokine receptor 6 (CXCR6) was strongly expressed by human T cells that were obtained by bronchoalveolar lavage. High levels of its ligand, CXC-chemokine ligand 16 (CXCL16), were also detected in the parenchyma, indicating that the CXCR6–CXCL16 pair could be involved in tissue-specific lung homing 33. Influenza-virusspecific CD8+ T cells that accumulate in the lung tissue of mice express the collagen-binding α 1β 1-integrin (also known as VLA1), and blocking this integrin led to a reduction in antigen-specific T cells, most dram atically in the airways34. Moreover, the level of LFA1 expression is markedly reduced on airway T cells com pared with parenchymal T cells35. These observations suggest that selective expression of integrins might be an important mechanism for the retention of T cells within the airways, at least in mice. In summary, the general immunological milieu in healthy lungs appears to be equivalent to that of a ‘low-level alert’ status, in which efficient surveil lance mechanisms are maintained for discrimination between trivial and potentially pathogenic antigens, whereas the capacity for local mobilization of strong immune effector mechanisms remains tightly con strained. This scenario limits the risk for damage to host lung tissues by unnecessary immune responses to non-pathogenic antigens or excessively aggressive memory T‑cell responses against pathogens. Various cell populations contribute to the overall maintenance of local immunological homeostasis within the respi ratory tract, and in some cases exhibit functions that appear to be specifically adapted to the unique local tis sue microenvironments. Particularly notable examples are discussed in the next section of this Review.
Specialized roles of individual cell types There are a number of individual cell types that have spe cialized roles in the fine tuning of immunological home ostasis in the lungs. In this section, we cover the most relevant cell types, although others, such as mast cells (BOX 1) and natural killer (NK)-cell populations (BOX 2) are also important. volume 8 | february 2008 | 145
© 2008 Nature Publishing Group
REVIEWS Protein-recall antigens Antigens to which individuals or experimental animals have been previously sensitized.
Inhalation tolerance The development of immunological tolerance to repeatedly inhaled antigen.
Lung DC populations: ‘gatekeepers’ of local adaptive immunity. Resident lung DCs share many phenotypic traits with DCs from other non-lymphoid tissues, and the typical profile includes high expression of MHC class II and CD205, together with low expression of CD8, CD40, CD80 and CD86 (REFS 3,36). In the steady state, they are specialized for antigen uptake and processing but lack the capacity for efficient antigen presentation, an attribute they normally develop only after migration to the RLNs3. However, in some circumstances, typified by atopic asthma in which lung DCs have a key role in the associated airway inflammation37, lung DCs can transiently develop potent antigen-presenting cell (APC) activity in situ (see below). Several distinct subpopulations of lung DCs have been identified38–40, but, until comparatively recently, the principal focus of research had been directed towards the myeloid DC subset, in particular, that found within the airway mucosa. A unique attribute of AMDC networks is their extremely high turnover rate in the steady state41 and their capacity for further rapid upregulation in the face of local inflammatory challenge3. Notably, DC precursor recruitment follow ing bacterial challenge exhibits kinetics that are com parable to neutrophils4,42, and only slightly less-rapid response patterns are evident following challenge with viruses and protein-recall antigens43. It is noteworthy that the recruitment of AMDCs at baseline and in response to bacterial stimuli is CCR1 and CCR5 dependent, whereas corresponding responses to virus or proteinrecall antigen use alternative chemokine ligand–receptor combinations43,44. In addition, experiments in CCR7deficient mice have demonstrated a key role for CCR7 in the baseline emigration of antigen-bearing lung DCs to RLNs13, further emphasizing the heterogeneity of chemokine usage patterns in relation to the fine control of local DC‑mediated immune surveillance; inhalation tolerance to a non-pathogenic antigen could not be induced in the CCR7-deficient mice13. It is pertinent to note that the current focus on these highly dynamic AMDC subsets has so far ignored the potential role of
Box 1 | Mast cells Mast cells are widely distributed in tissues but are most prominent near mucosal surfaces that are exposed to the environment, including the airway mucosa. Mast-cell phenotype, function and distribution can be modulated by many factors. Mast cells can promote innate antimicrobial immunity by activation through diverse mechanisms, including Toll-like receptors and receptors for complement and other inflammatory peptides85. Mast cells can be activated differentially by a range of stimuli to release, depending on the stimulus, a wide array of biologically active products (for example, tumour-necrosis factor (TNF), cathelicidins and preformed mediators such as histamine, prostaglandins, cytokines, chemokines, free radicals and growth factors), which have potent effects on other inflammatory cells85,86. Mast cells can also be phagocytic, and inter alia can have a role in neutrophil recruitment87, stimulation of IgE production by B cells, promotion of T‑cell proliferation88 and migration85, and potentially in the mediation of regulatory‑T-cell-mediated tolerance89. Many lines of evidence have also linked mast cells or their products (for example, histamine, TNF, prostaglandin E2 and thymic stromal lymphopoietin) to the regulation of dendritic-cell function and consequently to T‑cell polarization85,90. Most studies, however, have been performed in vitro, and the in vivo situation is unclear, but their proximity to alveolarmacrophage populations suggests the potential for important interaction effects.
146 | february 2008 | volume 8
a key subset (comprising 15% of AMDCs) of AMDCs with half-lives of more than 10 days41. There is increasing interest in the role of the pDC subset in the lungs, particularly in view of recent find ings that implicate these cells in tolerance induction to inhaled antigen12. A further indication of the potentially important functions of the pDC subset is their distinct pattern of TLR expression (TABLE 1) and accompanying high capacity to produce interferon-α (IFNα) in response to microbial stimuli45,46. However, unlike myeloid DCs, human pDCs have poor APC activity45 and there is no evidence for pDC migration out of lung tissues; rather, they behave as lymphocytes, entering lymphoid organs from the blood and responding locally to inflamma tion. Although resting pDCs are poor T‑cell stimulators, mature pDCs from mice47 and humans48 can present anti gen to CD4+ and CD8+ memory T cells in vitro. In vivo, T‑cell priming might be limited to CD8+ T cells in mice39, but pDCs might stimulate CD4+ T cells in some lymphoid organs49. These findings are collectively consistent with the perceived primary role of pDCs in antiviral immu nity, although their capacity for CD4+ T‑cell activation requires more detailed characterization. Lung epithelial cells: the first line of defence. Airway epithelial cells (AECs) have key roles in the regulation of lung homeostasis. As well as serving an important bar rier function in the exclusion of incoming environmental antigens, AECs secrete into the overlying fluid a range of regulatory and effector molecules that are involved in front-line defence against pathogens (TABLE 2). In addi tion, they have broad-ranging roles in the modulation of the activity of adjacent cell populations. Relevant functions include: modulation of airway smooth-muscle activity50; production of nitric oxide, which is a potent inhibitor of the functional activation of lung DCs15 and memory T cells16; amplification of host responses to microorganisms through the secretion of chemokines and cytokines51,52, in particular, type I and III IFNs53,54; production of mediators such as granulocyte colonystimulating factor (G-CSF), granulocyte/macrophage colony-stimulating factor (GM-CSF) and ICAM1, which enhance the local survival of recruited inflammatory cells55; regulation of respiratory-tract DC functions by growth factors such as GM‑CSF56; and modulation of airway B‑cell accumulation and/or activation57. In addi tion, the diverse array of receptors that are expressed by AECs enables them to respond dynamically to incom ing stimuli, including allergens, microorganisms and particulate material, and to thus adjust their mediatorproduction profiles to meet individual threats. Lung macrophages: multifaceted effector and modulatory roles. Macrophages have a central role in the maintenance of immunological homeostasis and host defence, and in the lungs the key population is composed of alveolar macrophages. In the steady state, the principal function of alveolar macrophages is phagocytosis and seques tration of antigen from the immune system to shield local tissues from the development of specific immune responses58. Alveolar macrophages have been shown to www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS take up most of the particulate material that is delivered intranasally, but they do not migrate to RLNs, nor are they considered to have a significant role in antigen presentation17,59. Alveolar macrophages also actively sup press the induction of adaptive immunity, as confirmed by studies showing that the use of clodronate-filled lipo somes to deplete alveolar macrophages in vivo rendered rodents susceptible to T‑cell-mediated inflammatory responses to otherwise harmless inhaled antigens18. The suppressive effects of alveolar macrophages were initially attributed to the direct suppression of T cells by nitric oxide and/or mediators such as interleukin‑10 (IL‑10), prostaglandins and transforming growth factor‑β (TGFβ). However, lung DCs clearly have enhanced APC function in alveolar-macrophage-depleted rodents, indicating that alveolar macrophages might suppress APC function rather than directly suppressing T cells60. More recently, it has been shown that alveolar-macrophage depletion results in increased numbers of DCs in alveolar spaces and increased uptake of particles by DCs, resulting in increased migration to RLNs59. These studies demonstrate a probable role for alveolar macrophages in the steadystate regulation of DC migration and localization. Further highlighting the complexities of these regulatory mechanisms, it has also been shown that under homeostatic conditions, alveolar macrophages closely adhere to AECs, inducing the TGFβ-dependent expression of αvβ6-integrin by AECs. The binding of latent TGFβ to αvβ6-integrin on AECs results in the local activation of TGFβ in close proximity to alveo lar macrophages. Activated TGFβ binds to alveolar macrophages and induces suppression of cytokine production61,62. Mice deficient in αvβ6-integrin have constitutively activated alveolar macrophages63; this mechanism seems to be unique to the lungs and exemplifies the specialized modulation of macrophage function by the local microenvironment to meet the needs of the tissue. Although resting alveolar macrophages are normally maintained in a quiescent state and produce small amounts of pro-inflammatory cytokines, they maintain the capacity to be activated in response to extrinsic and intrinsic stimuli, including cytokines, microorganisms and particulate material. The activation of alveolar mac rophages results in a shift in their functional capacity to that of efficient effector cells that participate in phago cytosis, killing64 and coordination of the innate immune response65. TLR ligation on alveolar macrophages results in their detachment from AECs, loss of αvβ6-integrin Table 1 | TLR expression on human lung DCs DC subset
mRNA detected
Protein detected Ligand stimulation shown
Myeloid DC1 (CD1c+)
TLR1, TLR2, TLR3, TLR4, TLR6, TLR8
TLR1, TLR2, TLR4
TLR2, TLR3, TLR4
38
Myeloid DC2 (BDCA3+)
TLR1, TLR2, TLR3, TLR4, TLR6, TLR8
TLR1, TLR2, TLR4
TLR2, TLR3, TLR4
38
TLR9
TLR7, TLR9
Plasmacytoid TLR1, TLR6, TLR7, DC (CD123+) TLR9
Refs
38,46
BDCA3, blood DC antigen 3; DC, dendritic cell; TLR, Toll-like receptor.
nature reviews | immunology
expression and, consequently, loss of the influence of the αvβ6-integrin–TGFβ axis over local homeostasis62. The turnover of alveolar-macrophage populations is normally extremely slow relative to the adjacent DC populations. In the steady state, alveolar macrophages are largely renewed by local precursor-cell proliferation, but during inflammation renewal essentially occurs via incoming monocytes and is regulated through the CCL2–CCR2 axis66. It is worth noting that the functional phenotype of recently recruited monocytes contrasts markedly with that of resident alveolar macrophages; in particular, the monocytes can function as effective APCs before their maturation into immunosuppressive alveo lar macrophages, which occurs over a period of days67. Their recruitment thus opens a transient ‘window’ for T‑cell responsiveness in the peripheral lung. Notably, this monocyte maturation process can be delayed by GM‑CSF, thus prolonging the window for the local induction of T‑cell immunity15. In addition, the immi grating monocytes invariably include immature DCs, which, under the influence of GM‑CSF, will also enhance T‑cell responsiveness during this period68. Regulatory T cells: sheet-anchoring local adaptive immunity. Forkhead box P3 (FOXP3) +CD4 +CD25 + regulatory T (TReg) cells are central to the control of peripheral T‑cell responses69 and use various medi ators for this purpose, including IL‑10 and/or TGFβ, and possibly nitric oxide70. In relation to the lung, they are believed to have key roles in protection against the inflammatory sequelae of airway infections, exempli fied by viral bronchiolitis, and in protection against the induction and expression of atopic disease. The process of inhalation tolerance, which protects against sensitization (and ensuing IgE production) to aero allergens is T‑cell mediated9, and recent evidence has implicated TReg cells as mediators of this form of tol erance10. The relative importance of antigen-specific versus nonspecific TReg cells in this and related contexts remains unresolved. TReg cells may also have a central role in controlling the local activation of allergen-specific TH cells that evade tolerance. We have demonstrated that the trig gering step in the asthma late-phase response involves cognate interactions between allergen-bearing AMDCs and incoming memory T cells within the airway mucosa, leading to transient upregulation of local APC activity followed by TH‑cell activation71. This proc ess and the ensuing airway hyperresponsiveness can be attenuated by TReg cells that are recruited into the airway mucosa following the first wave of inflamma tion that is triggered by allergen inhalation14, which is consistent with a role for these cells in limiting the intensity and duration of asthma exacerbations. Failure of this TReg-cell mechanism in atopic people who are sensitized to perennial aeroallergens would potentially lead to repeated cycles of TH2‑cell-associated inflam mation, thus stimulating the tissue repair and/or wound-healing responses, which are the harbingers of the airway-tissue remodelling that is the hallmark of chronic human asthma72. volume 8 | february 2008 | 147
© 2008 Nature Publishing Group
REVIEWS Endothelial cell
a Small
intestine
Blood α4β7-integrin IgA-secreting MADCAM1 cell
c Bronchial mucosa α4β1-integrin
VCAM1
CCL25 CCR10 CCL28
CCR9 MADCAM1
T cell
α4β1-integrin
VCAM1
CCL25 LFA1
CCR9
Respiratory epithelium
ICAM1
b Skin
d Lung parenchyma α4β1-integrin
VCAM1
LFA1
CCR4
PSGL1
CLA E-selectin
CCL17
CCR4
Type I epithelial cell Alveolar space
P-selectin
CCR10 CCR4
CCL17 CCL22 CXCL16
ICAM1
CCR5
CXCR6
↓ LFA1
CXCR6
CCL27
CCL5 iNKT cell
↑ VLA1
CCR4 CCR4
CCL17 CCL22
CCL22 CCL17
Type II epithelial cell
Figure 4 | Trafficking of lymphoid cells mediated by specific adhesion molecules and chemokine receptors. Reviews | Immunology Recruitment of lymphoid cells into target tissues requires specific chemokine recognition andNature adhesion-receptor engagement. a | IgA-secreting cells and T cells that are primed in Peyer’s patches and mesenteric lymph nodes preferentially express α4β7-integrin and CC‑chemokine receptor 9 (CCR9). Endothelial cells of postcapillary venules in the intestinal mucosa constitutively express ligands for α4β7-integrin and CCR9, namely mucosal addressin cell-adhesion molecule 1 (MADCAM1) and CC‑chemokine ligand 25 (CCL25), respectively, which allow lymphoid cells that are induced in intestinal lymphoid tissue to enter this mucosal effector site. b | T cells that are primed in peripheral lymph nodes lack intestinal homing receptors, but express cutaneous leukocyte antigen (CLA), α4β1-integrin, CCR4 and CCR10. These T cells are targeted to skin tissues that express ligands for these receptors (E-selectin, vascular cell-adhesion molecule 1 (VCAM1), CCL17 and CCL27, respectively) and cannot enter the intestinal mucosa. c | Distinct homing phenotypes for lymphoid-cell trafficking to the respiratory tract have yet to be defined, but recent information suggests that activated lymphoid cells that are induced by respiratory antigens express distinct phenotypes and lack intestinal-homing molecules. IgA-secreting cells that home to the bronchial mucosa express α4β1-integrin, whereas T cells express α4β1-integrin and lymphocyte function-associated antigen 1 (LFA1), which correspond with their counterparts VCAM1 and intercellular adhesion molecule 1 (ICAM1), respectively, that are constitutively expressed on the vessel wall in the bronchial mucosa. The α4β1-integrin–VCAM1 interaction is also crucial for lymphocyte homing to bronchial-associated lymphoid tissue, which is unique among secondary lymphoid tissues. IgA-secreting cells that traffic to the bronchial mucosa also express CCR10 and the ligand CCL28 is produced locally. The chemokine receptors that are involved in T‑cell homing to the central airways are less well defined. d | The expression of LFA1, P‑selectin glycoprotein ligand 1 (PSGL1) and CCR5 appear to be important for T cells to enter the lung parenchyma, whereas their further migration into the alveoli depends on expression of multiple chemokine receptors. Furthermore, their retention in the airways is associated with downregulation of LFA1 and upregulation of α1β1-integrin (also known as VLA1). Recruitment of invariant natural killer T (iNKT) cells to the lung parenchyma depends on CCR4, but the adhesion molecules involved in iNKT-cell-trafficking have not been defined.
Of additional interest, steroids that are used in the treatment of human asthma have recently been shown to promote TReg-cell activity in vitro and in vivo73, and this could explain some of the positive effects of inhaled steroids in asthmatics. Moreover, it has also recently been demonstrated that ongoing protection of airway tissues in sensitized animals via TReg cells relies absolutely 148 | february 2008 | volume 8
on continuation of allergen exposure, as withdrawal of stimulation leads to rapid decline in airway TReg-cell numbers and restores susceptibility to the TH2-celldependent airway-hyperresponsiveness-inducing effects of aeroallergen challenge14. Whether prolonged stimulation of TReg cells by the continuous therapeu tic administration of allergen can eventually exhaust www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
REVIEWS Box 2 | Natural-killer-cell populations Natural killer (NK) cells and invariant NKT (iNKT) cells have distinct roles in the maintenance of immunological homeostasis. NK‑cell deficiency is associated with an inability to control virus infections, and in the context of the lungs, NK cells have been implicated in host responses to Bordetella pertussis91, Mycobacterium tuberculosis92, influenza virus93 and respiratory syncytial virus94. The contribution of NK cells to controlling infections with these pathogens varies, but in most cases their influx closely follows the kinetics of infection. iNKT cells differ from NK cells in that they express CD3 and an invariant T‑cellreceptor α‑chain95. They respond to glycolipid antigens that are presented by CD1d, and produce large quantities of interferon‑γ and interleukin‑4 (Ref. 95), although the mechanisms that regulate their cytokine repertoire remain unclear96. The role of iNKT cells in host defence in the lung has largely been inferred from experiments in knockout mice that lack iNKT cells, which show increased susceptibility to a wide range of bacterial and viral pathogens and parasites97. It has recently been proposed that iNKT cells also have a role in asthma pathogenesis: experiments with iNKT-deficient mice show reduced responses to sensitization and challenge that can be largely reconstituted by the adoptive transfer of iNKT cells; and iNKT cells that are triggered by glycolipid antigens can mediate airway hyperresponsiveness98. The extent to which these findings can be extrapolated to human asthma remains unclear. A recent study reported substantial numbers of CD4+ iNKT cells in bronchoalveolar lavage from patients with asthma99, but others have failed to confirm these findings100. However, it is not yet clear how applicable these findings are to the broad spectrum of human asthma phenotypes, or how iNKT cells are activated by aeroallergen exposure.
relevant allergen-specific TH2-memory-cell populations remains to be established. Further investigation of this concept and the mechanisms responsible will ultimately guide the development of new therapeutic strategies. It is also noteworthy that in the setting of TLR‑induced activation of DCs, the suppressive role of TReg cells is circumvented, thus enabling the induction of appropri ate adaptive immune responses74,75. Reversal of TReg-cell suppression requires TLR-induced IL‑6 production, and IL‑1 can synergize with IL‑6 in this process74–76. In situ ations in which atopic asthma is acutely exacerbated by respiratory infections77, it is plausible that the infection and activation of DCs via pathogen-associated molecular patterns, and the ensuing derailment of TReg-cell control, could be central to triggering the exacerbation of asthma, and this issue merits more detailed investigation. Recent insights into effector T‑cell functions. For the past three decades, the TH1–TH2-cell paradigm has provided the basis for the systematic study of immune regulation. In the context of immunologically related lung diseases, these two sides of the yin–yang dichotomy of TH‑cell func tion have supplied much of the conceptual framework for studies on the pathogenesis of infectious and allergic res piratory diseases, and related drug development. However, it is becoming increasingly evident that TH‑cell functions are considerably more complex and heterogeneous than originally envisaged, and underlying concepts are rapidly being revised on several fronts. Two recent developments that are related to T‑cell function are of particular rel evance to immunoregulation in the lungs. The first relates to the potential key role of TH17 cells in disease pathogenesis75,76. Differentiation of this T‑cell population, which produces large amounts of IL‑17, IL‑22 and IL‑23, is thought to be induced by IL-6, IL-21 nature reviews | immunology
and TGFβ in mice75,76 and by the combination of IL‑6 and IL‑1β in humans78. IL‑23, originally thought to have a role in the differentiation of TH17 cells, is now regarded as being important for maintaining TH17-cell immune pathology. IL‑22 has been shown to have a crucial role in innate skin immunity by the induction of cyclic AMP and might also have an important role in respiratory-tract immunology. IL‑17 is an important cytokine that is involved in the mobilization and generation of neutrophils, and as such might have a pivotal role in pathogen clearance. In relation to the compartmentalization of TH‑cell functions, TH1-type cytokines strongly inhibit the development not only of TH2 cells but also of TH17 cells, and TH2-type cytokines (notably IL‑4) inhibit the development of both TH1 and TH17 cells75,76,79. The association of IL‑17 and neutro phils with more severe forms of immunoinflammatory disease79 leads one to ponder the importance of TH17 cells in the pathogenesis of atopic asthma, particularly as the only T cells that can produce IL‑6 appear to be TH2 cells. Interestingly, the differentiation of TH17 and TReg cells appears to be mutually exclusive 75,76. In mice, TGFβ induces the differentiation of both TH17 and TReg cells, however, IL‑6 inhibits TReg-cell differentiation and TH17 cells constitute the resultant population. It has been proposed that the reversal of TReg-cell suppres sive function that is induced by TLR activation may be a reflection of an increase in TH17-cell differentiation (involving IL‑6 and TGFβ) rather than a decrease in the ability for suppression76. However, a different scenario may be that TReg-cell populations (which have reversed suppressive capacity) also undergo proliferation, and once pro-inflammatory cytokine production has suffi ciently subsided, there would be an increased pool size of TReg cells readily available to limit immune pathology and act to re‑establish steady-state conditions75. An additional conceptual development that is relevant to this discussion relates to the role of AECs in driving the selection of disease-related TH‑cell phenotypes through the expression of potent T‑cell modulatory molecules. Notable examples are thymic stromal lymphopoietin, which drives the development of tumour-necrosis-fac tor-producing ‘inflammatory’ TH2 cells80, and the IL‑17 cytokine family member IL‑25, which drives overall TH2cell differentiation81. In this context, we have recently demonstrated by microarray analysis the differential expression of the IL‑25 receptor (known as IL‑17RB) in human allergen-specific memory TH2 cells from atopic individuals as part of a large transcriptome that contains multiple previously unrecognized TH2-cell‑associated genes82. These findings underscore the potential use of bioinformatics-driven approaches in the identification of covert effector pathways in respiratory inflammatory diseases, and this represents a potentially fruitful avenue for future research.
Conclusions Although substantial progress has been made in the elucidation of the basic mechanisms that underlie the fine control of immune responses to different classes volume 8 | february 2008 | 149
© 2008 Nature Publishing Group
REVIEWS Table 2 | First-line defence molecules produced by airway epithelial cells (AECs) AEC-secreted product
Action
Refs
Mucins
Host defence; bind infectious agents
85
Surfactant protein C
Maintenance of surfactant proteins; bind infectious agents
85
Surfactant protein A and surfactant protein D (collectins)
Opsonins for pathogen clearance; direct inhibition; activate other immune cellular functions
86
Complement and complement cleavage products
Promote phagocytosis; bridging of innate and adaptive immunity; resolution and repair
87
Antimicrobial peptides (defensins, cathelicidins, Direct antimicrobial action; effector molecules; activation histatins, lysozyme, lactoferrin, SLPI, Elafin, PLUNC of adaptive immunity and BPI)
88,89
BPI, bacterial permeability-increasing protein; PLUNC, palate, lung and nasal epithelium clones; SLPI, secretory leukoprotease inhibitor.
of inhaled antigens, subsequent translation of these largely experimental findings into the human disease arena remains an elusive goal. The archetypal example of this quandary is atopic asthma, the key manifestations of which can be readily replicated in animal models by deliberately bypassing the fundamental tolerance mecha nisms that limit sensitization to non-pathogenic airborne antigens, and can be readily controlled by a range of highly selective TH2-cell antagonists. The surging prevalence of respiratory inflammatory syndromes such as asthma over the past 2–3 decades and the disappointing clini cal efficacy of the latest generation of anti-inflammatory drugs developed to treat these diseases attests to the fact that we are still lacking a clear understanding of the key mechanisms that underlie immunoregulation in this
1.
2.
3.
4.
5.
6.
7. 8.
9.
Nelson, D. J., McMenamin, C., McWilliam, A. S., Brenan, M. & Holt, P. G. Development of the airway intraepithelial dendritic cell network in the rat from class II major histocompatibility (Ia)-negative precursors: differential regulation of Ia expression at different levels of the respiratory tract. J. Exp. Med. 179, 203–212 (1994). Jahnsen, F. L. et al. Rapid dendritic cell recruitment to the bronchial mucosa of patients with atopic asthma in response to local allergen challenge. Thorax 56, 823–826 (2001). Stumbles, P. A. et al. Resting respiratory tract dendritic cells preferentially stimulate T helper cell type 2 (Th2) responses and require obligatory cytokine signals for induction of Th1 immunity. J. Exp. Med. 188, 2019–2031 (1998). Jahnsen, F. L. et al. Accelerated antigen sampling and transport by airway mucosal dendritic cells following inhalation of a bacterial stimulus. J. Immunol. 177, 5861–5867 (2006). Rescigno, M. et al. Dendritic cells express tight junction proteins and penetrate gut epithelial monolayers to sample bacteria. Nature Immunol. 2, 361–367 (2001). Lund, F. E. et al. B cells are required for generation of protective effector and memory CD4 cells in response to pneumocystis lung infection. J. Immunol. 176, 6147–6154 (2006). Moyron-Quiroz, J. E. et al. Role of inducible bronchus associated lymphoid tissue (iBALT) in respiratory immunity. Nature Med. 10, 927–934 (2004). Strickland, D. H., Kees, U. R. & Holt, P. G. Regulation of T‑cell activation in the lung: isolated lung T‑cells exhibit surface phenotypic characteristics of recent activation including downmodulated TcR, but are locked into G0/G1 phase of the cell cycle. Immunology 87, 242–249 (1996). Holt, P. G. & Sedgwick, J. D. Suppression of IgE responses following antigen inhalation: a natural homeostatic mechanism which limits sensitization to aeroallergens. Immunol. Today 8, 14–15 (1987).
microenvironment, and as such, we are in need of new paradigms to drive drug discovery83. As well as the covert complexity of underlying effector mechanisms, part of the problem may be the limitations that are inherent in modelling what are, in reality, chronic inflammatory diseases in short-term experimental settings. In particular, emerging evidence suggests that primary sensitization to airborne antigens often occurs in infancy when immune functions are (developmentally) attenuated, but may not manifest as clinically relevant airway disease until later in life when immune effector mechanisms in the respiratory tract are fully competent84. Dealing with such subtleties of disease pathogenesis remains a major challenge to progress in this area.
10. Umetsu, D. T. & DeKruyff, R. H. The regulation of allergy and asthma. Immunol. Rev. 212, 238–255 (2006). 11. Weiner, H. L. et al. Oral tolerance: immunologic mechanisms and treatment of animal and human organ-specific autoimmune diseases by oral administration of autoantigens. Annu. Rev. Immunol. 12, 809–837 (1994). 12. de Heer, H. J. et al. Essential role of lung plasmacytoid dendritic cells in preventing asthmatic reactions to harmless inhaled antigen. J. Exp. Med. 200, 89–98 (2004). 13. Hintzen, G. et al. Induction of tolerance to innocuous inhaled antigen relies on a CCR7-dependent dendritic cell-mediated antigen transport to the bronchial lymph node. J. Immunol. 177, 7346–7354 (2006). 14. Strickland, D. H. et al. Reversal of airway hyperresponsiveness by induction of airway mucosal CD4+CD25+ regulatory T cells. J. Exp. Med. 203, 2649–2660 (2006). This study demonstrates how interactions between airway mucosal DCs and TReg cells control the intensity and duration of TH‑cell memory responses in the airways. 15. Bilyk, N. & Holt, P. G. Inhibition of the immunosuppressive activity of resident pulmonary alveolar macrophages by granulocyte/macrophage colony-stimulating factor. J. Exp. Med. 177, 1773–1777 (1993). 16. Eriksson, U. et al. Human bronchial epithelium controls TH2 responses by TH1-induced, nitric oxidemediated STAT5 dephosphorylation: implications for the pathogenesis of asthma. J. Immunol. 175, 2715–2720 (2005). 17. Lambrecht, B. N. Alveolar macrophage in the driver’s seat. Immunity 24, 366–368 (2006). A comprehensive review of the multifaceted immunomodulatory and effector roles of alveolar macrophages. 18. Thepen, T., Van Rooijen, N. & Kraal, G. Alveolar macrophage elimination in vivo is associated with an increase in pulmonary immune response in mice. J. Exp. Med. 170, 499–509 (1989).
150 | february 2008 | volume 8
19. Eisenbarth, S. C. et al. Lipopolysaccharide-enhanced, Toll-like receptor 4‑dependent T helper cell type 2 responses to inhaled antigen. J. Exp. Med. 196, 1645–1651 (2002). A landmark study demonstrating the role of environmental LPS in facilitating the recognition of, and host responses to, inhaled protein antigens. 20. Gereda, J. E. et al. Relation between house-dust endotoxin exposure, type 1 T‑cell development, and allergen sensitisation in infants at high risk of asthma. Lancet 355, 1680–1683 (2000). 21. Agace, W. W. Tissue-tropic effector T cells: generation and targeting opportunities. Nature Rev. Immunol. 6, 682–692 (2006). 22. Cose, S., Brammer, C., Khanna, K. M., Masopust, D. & Lefrancois, L. Evidence that a significant number of naive T cells enter non-lymphoid organs as part of a normal migratory pathway. Eur. J. Immunol. 36, 1423–1433 (2006). 23. Iwata, M. et al. Retinoic acid imprints gut-homing specificity on T cells. Immunity 21, 527–538 (2004). 24. Kunkel, E. J. & Butcher, E. C. Plasma-cell homing. Nature Rev. Immunol. 3, 822–829 (2003). 25. Campbell, J. J. et al. The chemokine receptor CCR4 in vascular recognition by cutaneous but not intestinal memory T cells. Nature 400, 776–780 (1999). 26. Homey, B. et al. CCL27-CCR10 interactions regulate T cell-mediated skin inflammation. Nature Med. 8, 157–165 (2002). 27. Sigmundsdottir, H. et al. DCs metabolize sunlightinduced vitamin D3 to ‘program’ T cell attraction to the epidermal chemokine CCL27. Nature Immunol. 8, 285–293 (2007). 28. Mora, J. R. et al. Generation of gut-homing IgAsecreting B cells by intestinal dendritic cells. Science 314, 1157–1160 (2006). 29. Kohlmeier, J. E. & Woodland, D. L. Memory T cell recruitment to the lung airways. Curr. Opin. Immunol. 18, 357–362 (2006). A review of control of the recruitment of recirculating memory TH cells in lung tissues.
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
REVIEWS 30. Campbell, J. J. et al. Expression of chemokine receptors by lung T cells from normal and asthmatic subjects. J. Immunol. 166, 2842–2848 (2001). 31. Galkina, E. et al. Preferential migration of effector CD8+ T cells into the interstitium of the normal lung. J. Clin. Invest. 115, 3473–3483 (2005). 32. Kallinich, T. et al. Chemokine-receptor expression on T cells in lung compartments of challenged asthmatic patients. Clin. Exp. Allergy. 35, 26–33 (2005). 33. Morgan, A. J. et al. Expression of CXCR6 and its ligand CXCL16 in the lung in health and disease. Clin. Exp. Allergy. 35, 1572–1580 (2005). 34. Ray, S. J. et al. The collagen binding α1β1 integrin VLA‑1 regulates CD8 T cell-mediated immune protection against heterologous influenza infection. Immunity 20, 167–179 (2004). 35. Ely, K. H., Cookenham, T., Roberts, A. D. & Woodland, D. L. Memory T cell populations in the lung airways are maintained by continual recruitment. J. Immunol. 176, 537–543 (2006). 36. von Garnier, C. et al. Anatomical location determines the distribution and function of dendritic cells and other APCs in the respiratory tract. J. Immunol. 175, 1609–1618 (2005). 37. Lambrecht, B. N., Salomon, B., Klatzmann, D. & Pauwels, R. A. Dendritic cells are required for the development of chronic eosinophilic airway inflammation in response to inhaled antigen in sensitized mice. J. Immunol. 160, 4090–4097 (1998). The first study to provide direct evidence for the central role of airway mucosal DCs in triggering TH‑cell memory responses to inhaled allergen. 38. Demedts, I. K., Brusselle, G. G., Vermaelen, K. Y. & Pauwels, R. A. Identification and characterization of human pulmonary dendritic cells. Am. J. Respir. Cell Mol. Biol. 32, 177–184 (2005). 39. Schlecht, G. et al. Murine plasmacytoid dendritic cells induce effector/memory CD8+ T‑cell responses in vivo after viral stimulation. Blood 104, 1808–1815 (2004). 40. Wikstrom, M. E. & Stumbles, P. A. Mouse respiratory tract dendritic cell subsets and the immunological fate of inhaled antigens. Immunol. Cell Biol. 85, 182–188 (2007). 41. Holt, P. G., Haining, S., Nelson, D. J. & Sedgwick, J. D. Origin and steady-state turnover of class II MHCbearing dendritic cells in the epithelium of the conducting airways. J. Immunol. 153, 256–261 (1994). 42. McWilliam, A. S., Nelson, D., Thomas, J. A. & Holt, P. G. Rapid dendritic cell recruitment is a hallmark of the acute inflammatory response at mucosal surfaces. J. Exp. Med. 179, 1331–1336 (1994). A milestone study demonstrating that rapid recruitment of immature DCs into the airway mucosa represents the ‘default’ response to local challenge with virtually all classes of airborne environmental stimuli. 43. Stumbles, P. A. et al. Regulation of dendritic cell recruitment into resting and inflamed airway epithelium: use of alternative chemokine receptors as a function of inducing stimulus. J. Immunol. 167, 228–234 (2001). 44. Robays, L. J. et al. Chemokine receptor CCR2 but not CCR5 or CCR6 mediates the increase in pulmonary dendritic cells during allergic airway inflammation. J. Immunol. 178, 5305–5311 (2007). 45. Demedts, I. K., Bracke, K. R., Maes, T., Joos, G. F. & Brusselle, G. G. Different roles for human lung dendritic cell subsets in pulmonary immune defense mechanisms. Am. J. Respir. Cell Mol. Biol. 35, 387–393 (2006). 46. Masten, B. J. et al. Characterization of myeloid and plasmacytoid dendritic cells in human lung. J. Immunol. 177, 7784–7793 (2006). 47. Krug, A. et al. Interferon-producing cells fail to induce proliferation of naive T cells but can promote expansion and T helper 1 differentiation of antigenexperienced unpolarized T cells. J. Exp. Med. 197, 899–906 (2003). 48. Fonteneau, J.‑F. et al. Activation of influenza virusspecific CD4+ and CD8+ T cells: a new role for plasmacytoid dendritic cells in adaptive immunity. Blood 101, 3520–3526 (2003). 49. Sapoznikov, A. et al. Organ-dependent in vivo priming of naive CD4+, but not CD8+, T cells by plasmacytoid dendritic cells. J. Exp. Med. 204, 1923–1933 (2007). 50. Howarth, P. H. et al. Epithelially derived endothelin and nitric oxide in asthma. Int. Arch. Allergy Immunol. 107, 228–230 (1995).
51. Mayer, A. K. et al. Differential recognition of TLRdependent microbial ligands in human bronchial epithelial cells. J. Immunol. 178, 3134–3142 (2007). 52. Pichavant, M. et al. Impact of bronchial epithelium on dendritic cell migration and function: modulation by the bacterial motif KpOmpA. J. Immunol. 177, 5912–5919 (2006). 53. Contoli, M. et al. Role of deficient type III interferon-λ production in asthma exacerbations. Nature Med. 12, 1023–1026 (2006). 54. Wark, P. A. B. et al. Asthmatic bronchial epithelial cells have a deficient innate immune response to infection with rhinovirus. J. Exp. Med. 201, 937–947 (2005). 55. Cox, G., Gauldie, J. & Jordana, M. Bronchial epithelial cell-derived cytokines (G-CSF and GM‑CSF) promote the survival of peripheral blood neutrophils in vitro. Am. J. Respir. Cell Mol. Biol. 7, 507–513 (1992). 56. Bleck, B., Tse, D. B., Jaspers, I., Curotto de Lafaille, M. A. & Reibman, J. Diesel exhaust particle-exposed human bronchial epithelial cells induce dendritic cell maturation. J. Immunol. 176, 7431–7437 (2006). 57. Kato, A., Truong-Tran, A. Q., Scott, A. L., Matsumoto, K. & Schleimer, R. P. Airway epithelial cells produce B cell-activating factor of TNF family by an IFN‑β‑dependent mechanism. J. Immunol. 177, 7164–7172 (2006). 58. MacLean, J. A. et al. Sequestration of inhaled particulate antigens by lung phagocytes. A mechanism for the effective inhibition of pulmonary cell-mediated immunity. Am. J. Pathol. 148, 657–666 (1996). 59. Jakubzick, C., Tacke, F., Llodra, J., van Rooijen, N. & Randolph, G. J. Modulation of dendritic cell trafficking to and from the airways. J. Immunol. 176, 3578–3584 (2006). 60. Holt, P. G. et al. Downregulation of the antigen presenting cell function(s) of pulmonary dendritic cells in vivo by resident alveolar macrophages. J. Exp. Med. 177, 397–407 (1993). 61. Munger, J. S. et al. The integrin αvβ6 binds and activates latent TGFβ1: a mechanism for regulating pulmonary inflammation and fibrosis. Cell 96, 319–328 (1999). 62. Takabayashi, K. et al. Induction of a homeostatic circuit in lung tissue by microbial compounds. Immunity 24, 475–487 (2006). 63. Morris, D. G. et al. Loss of integrin αvβ6-mediated TGF-β activation causes Mmp12-dependent emphysema. Nature 422, 169–173 (2003). 64. Gwinn, M. R. & Vallyathan, V. Respiratory burst: role in signal transduction in alveolar macrophages. J. Toxicol. Environ. Health B Crit. Rev. 9, 27–39 (2006). 65. Underhill, D. M. & Ozinsky, A. Phagocytosis of microbes: complexity in action. Annu. Rev. Immunol. 20, 825–852 (2002). 66. Winter, C. et al. Lung-specific overexpression of CC chemokine ligand (CCL) 2 enhances the host defense to Streptococcus pneumoniae infection in mice: role of the CCL2–CCR2 axis. J. Immunol. 178, 5828–5838 (2007). 67. Bilyk, N. & Holt, P. G. Cytokine modulation of the immunosuppressive phenotype of pulmonary alveolar macrophages via regulation of nitric oxide production. Immunology 86, 231–237 (1995). 68. Landsman, L., Varol, C. & Jung, S. Distinct differentiation potential of blood monocyte subsets in the lung. J. Immunol. 178, 2000–2007 (2007). 69. Banham, A. H., Powrie, F. M. & Suri-Payer, E. FOXP3+ regulatory T cells: current controversies and future perspectives. Eur. J. Immunol. 36, 2832–2836 (2006). 70. Chen, C., Lee, W. H., Zhong, L. & Liu, C. P. Regulatory T cells can mediate their function through the stimulation of APCs to produce immunosuppressive nitric oxide. J. Immunol. 176, 3449–3460 (2006). 71. Huh, J. C. et al. Bidirectional interactions between antigen-bearing respiratory tract dendritic cells (DCs) and T cells precede the late phase reaction in experimental asthma: DC activation occurs in the airway mucosa but not in the lung parenchyma. J. Exp. Med. 198, 19–30 (2003). 72. Busse, W. W. & Lemanske, R. F. Asthma. N. Engl. J. Med. 344, 350–362 (2001). 73. Hartl, D. et al. Quantitative and functional impairment of pulmonary CD4+CD25hi regulatory T cells in pediatric asthma. J. Allergy Clin. Immunol. 119, 1258–1266 (2007). 74. Pasare, C. & Medzhitov, R. Toll pathway-dependent blockade of CD4+CD25+ T cell-mediated suppression by dendritic cells. Science 299, 1033–1036 (2003).
nature reviews | immunology
75. Weaver, C. T., Harrington, L. E., Mangan, P. R., Gavrieli, M. & Murphy, K. M. Th17: an effector CD4 T cell lineage with regulatory T cell ties. Immunity 24, 677–688 (2006). 76. Chen, Z. & O’Shea, J. J. Regulation of IL-17 production in human lymphocytes. Cytokine 2 November 2007 (doi:10.1016/j.cyto.2007.09.009). 77. Johnston, S. L. et al. The relationship between upper respiratory infections and hospital admissions for asthma: a time-trend analysis. Am. J. Respir. Crit. Care Med. 154, 654–660 (1996). 78. Acosta-Rodriguez, E. V., Napolitani, G., Lanzavecchia, A. & Sallusto, F. Interleukins 1β and 6 but not transforming growth factor-β are essential for the differentiation of interleukin 17-producing human T helper cells. Nature Immunol. 8, 942–949 (2007). 79. Steinman, L. A brief history of TH17, the first major revision in the TH1/TH2 hypothesis of T cell-mediated tissue damage. Nature Med. 13, 139–145 (2007). A recent review on the evolving TH17-cell story, which is highly relevant to the pathogenesis of respiratory inflammatory diseases. 80. Liu, Y.‑J. et al. TSLP: an epithelial cell cytokine that regulates T cell differentiation by conditioning dendritic cell maturation. Annu. Rev. Immunol. 25, 193–219 (2007). A definitive review on the role of epithelial-cellderived thymic stromal lymphopoietin in programming DCs to drive the differentiation of specific subphenotype(s) of TH cells, in particular ‘inflammatory’ TH2 cells. 81. Angkasekwinai, P. et al. Interleukin 25 promotes the initiation of proallergic type 2 responses. J. Exp. Med. 204, 1509–1517 (2007). 82. Bosco, A. et al. Identification of novel Th2-associated genes in T memory responses to allergens. J. Immunol. 176, 4766–4777 (2006). 83. Holt, P. G. et al. Drug development strategies for asthma: in search of a new paradigm. Nature Immunol. 5, 695–698 (2004). 84. Holt, P. G., Upham, J. W. & Sly, P. D. Contemporaneous maturation of immunological and respiratory functions during early childhood: Implications for development of asthma prevention strategies. J. Allergy Clin. Immunol. 116, 16–24 (2005). 85. Galli, S. J. et al. Mast Cells as “tunable” effector and immunoregulatory cells: recent advances. Annu. Rev. Immunol. 23, 749–786 (2005). 86. Galli, S. J., Nakae, S. & Tsai, M. Mast cells in the development of adaptive immune responses. Nature Immunol. 6, 135–142 (2005). 87. Malaviya, R., Ikeda, T., Ross, E. & Abraham, S. N. Mast cell modulation of neutrophil influx and bacterial clearance at sites of infection through TNF-α. Nature 381, 77–80 (1996). 88. Nakae, S. et al. Mast cells enhance T cell activation: importance of mast cell costimulatory molecules and secreted TNF. J. Immunol. 176, 2238–2248 (2006). 89. Lu, L. F. et al. Mast cells are essential intermediaries in regulatory T‑cell tolerance. Nature 442, 997–1002 (2006). 90. Suto, H. et al. Mast cell-associated TNF promotes dendritic cell migration. J. Immunol. 176, 4102–4112 (2006). 91. Byrne, P., McGuirk, P., Todryk, S. & Mills, K. H. G. Depletion of NK cells results in disseminating lethal infection with Bordetella pertussis associated with a reduction of antigen-specific Th1 and enhancement of Th2, but not Tr1 cells. Eur. J. Immunol. 34, 2579–2588 (2004). 92. Junqueira-Kipnis, A. P. et al. NK cells respond to pulmonary Infection with Mycobacterium tuberculosis, but play a minimal role in protection. J. Immunol. 171, 6039–6045 (2003). 93. Gazit, R. et al. Lethal influenza infection in the absence of the natural killer cell receptor gene Ncr1. Nature Immunol. 7, 517–523 (2006). 94. Haynes, L. M. et al. Involvement of Toll-like receptor 4 in innate immunity to respiratory syncytial virus. J. Virol. 75, 10730–10737 (2001). 95. Kronenberg, M. Toward an understanding of NKT cell biology: progress and paradoxes. Annu. Rev. Immunol. 23, 877–900 (2005). 96. Godfrey, D. I. & Kronenberg, M. Going both ways: Immune regulation via CD1d-dependent NKT cells. J. Clin. Invest. 114, 1379–1388 (2004). 97. Tupin, E., Kinjo, Y. & Kronenberg, M. The unique role of natural killer T cells in the response to microorganisms. Nature Rev. Microbiol. 5, 405–417 (2007).
volume 8 | february 2008 | 151 © 2008 Nature Publishing Group
REVIEWS 98. Meyer, E. H. et al. Glycolipid activation of invariant T cell receptor+ NK T cells is sufficient to induce airway hyperreactivity independent of conventional CD4+ T cells. Proc. Natl Acad. Sci. USA 103, 2782–2787 (2006). 99. Akbari, O. et al. CD4+ invariant T‑cell‑receptor+ natural killer T cells in bronchial asthma. N. Engl. J. Med. 354, 1117–1129 (2006). 100. Vijayanand, P. et al. Invariant natural killer T cells in asthma and chronic obstructive pulmonary disease. N. Engl. J. Med. 356, 1410–1422 (2007). 101. Mulugeta, S. & Beers, M. F. Surfactant protein C: its unique properties and emerging immunomodulatory
role in the lung. Microbes Infect. 8, 2317–2323 (2006). 102. Wright, J. R. Immunoregulatory functions of surfactant proteins. Nature Rev. Immunol. 5, 58–68 (2005). 103. Hawlisch, H. & Kohl, J. Complement and Toll-like receptors: key regulators of adaptive immune responses. Mol. Immunol. 43, 13–21 (2006). 104. Blasi, F., Tarsia, P. & Aliberti, S. Strategic targets of essential host-pathogen interactions. Respiration 72, 9–25 (2005). 105. Zasloff, M. Antimicrobial peptides of multicellular organisms. Nature 415, 389–395 (2002).
152 | february 2008 | volume 8
Acknowledgements
P.G.H. is supported by the National Health and Medical Research Foundation of Australia.
DATABASES Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene α1β1-integrin | α4β7-integrin | αvβ6-integrin | CCL28 | CCR1 | CCR4 | CCR5 | CCR9 | CCR10 | CXCL16 | CXCR6 | IL‑6 | IL‑17 | IL‑25 | TGFβ | VCAM1 All links are active in the online pdf
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
Perspectives opinion
The reverse stop-signal model for CTLA4 function Christopher E. Rudd
Abstract | Activation of the T‑cell co-receptor cytotoxic T‑lymphocyte antigen 4 (CTLA4) has a pivotal role in adjusting the threshold for T‑cell activation and in preventing autoimmunity and massive tissue infiltration by T cells. Although many mechanistic models have been postulated, no single model has yet accounted for its overall function. In this Opinion article, I outline the strengths and weaknesses of the current models, and present a new ‘reverse stop-signal model’ to account for CTLA4 function. Cytotoxic T‑lymphocyte antigen 4 (CTLA4) is a co-receptor expressed by activated, memory T cells and regulatory T cells1–6. It is a homodimer that can bind to the homodimeric ligands CD80 (also known as B7.1) and CD86 (also known as B7.2) expressed by antigen-presenting cells (APCs)7,8, although a preference for binding to CD80 has been reported9. At any one time, CTLA4 is only present on the surface of activated T cells at low levels; the majority of CTLA4 proteins are instead localized in intracellular compartments and are only transported to the cell surface in response to T‑cell receptor (TCR) ligation10,11. The level of transport is determined by the strength of the TCR-mediated signal and is facilitated by the chaperone adaptor TCR-interacting molecule (TRIM)12. Following binding to CD80 and/or CD86, CTLA4 is rapidly endo‑ cytosed, a process which is facilitated by the adaptor protein 2 (AP2) complex13–15. CTLA4 has the capacity to dampen proliferation and guard against autoimmu‑ nity, both of which are attributes that were most clearly illustrated by the phenotype of CTLA4-deficient mice. Waterhouse et al. and Tivol et al. first showed that these mice died by 3–4 weeks of birth as a result of massive lymphoadenopathy, tissue infiltra‑ tion and destruction of multiple organs16,17. The observed lymphoproliferation was later shown to be mediated by the co-stimulationdependent activation of CD4+ T cells18.
Whether the phenotype of Ctla4–/– mice is strictly autoimmune in nature and/or is a type of proliferative, inflammatory syndrome is still not clear. On the one hand, there is no evidence to date of TCR repertoire skewing19 or infiltration of site-specific organs by CD4+ T cells, which would otherwise point to the presence of a site-specific autoantigen. On the other hand, mutations in the CTLA4 gene locus have been mapped to several autoimmune disorders, such as autoimmune hypothy‑ roidism and type 1 diabetes20. Furthermore, reductions in CTLA4 expression induced by small interfering RNA (siRNA) can cause a more rapid onset of type 1 diabetes in mice21. These observations have led to the sug‑ gestion that CTLA4 raises the threshold for the TCR-induced activation of T cells, and consequently prevents responses to low-affinity self antigens. The basis for the extensive tissue infiltration observed in CTLA4-deficient mice is still not under‑ stood, but requires integrin and T‑cell motility and migration mediated by homing receptors. Increased T-cell migration may result from the hyperactivation of T cells through the TCR, and may facilitate the localization of T cells to sites where they can be activated by self antigen. This excessive tissue infiltration facilitates organ destruction, which is responsible for the eventual death of Ctla4–/– mice.
nature reviews | immunology
A major challenge over the past decade has been to uncover the underlying basis of the function of CTLA4. Understanding this basis should provide key information on the factors that induce autoimmunity. To date, no overarching, single mechanism has been described that explains the function of CTLA4. One question is whether the coreceptor can engage more than one pathway. Both cell-extrinsic events (that is, those involving factors that operate outside of the CTLA4-expressing cell, such as cytokines or other cells) and cell-intrinsic events (that is, regulatory effects mediated from within the CTLA4-expressing cell) have been reported to be involved in CTLA4 function, suggesting multiple pathways. If more than one pathway is involved, then the question remains: is one or more pathways responsible for the disease phenotype of Ctla4–/– mice and the normal function of CTLA4 in the immune response, or are some pathways linked to the preven‑ tion of autoimmunity, with other pathways having a role in regulating other aspects of the immune response? In this Opinion article, I outline the strengths and weaknesses of the current models of CTLA4 function and present a new ‘reverse stop-signal model’ to account for its function. By reversing the TCRinduced stop signal for T‑cell motility, this model proposes that CTLA4 limits the dwell time between T cells and APCs, thereby reducing the level of T‑cell activation. The reverse stop-signal model can potentially explain how CTLA4 regulates the activation threshold of T cells, anergy, autoimmunity, tissue infiltration and various T‑cell effector functions. Cell-extrinsic factors Bachmann and co-workers first showed that although the adoptive transfer of purified Ctla4–/– T cells into recombinationactivating gene 2 (RAG2)-deficient mice (which lack mature B and T cells) resulted in disease, the co-transfer of wild-type bone marrow with Ctla4–/– T cells pre‑ vented disease development and enhanced antigen responsiveness22,23. One possible explanation for these results is that the CTLA4+ T cells are simply more effective in competing for antigen that is presented volume 8 | february 2008 | 153
© 2008 Nature Publishing Group
Perspectives by APCs compared with Ctla4–/– T cells. Alternatively, there may be one or more extrinsic factors produced by CTLA4+ T cells that dampen the reactivity of Ctla4–/– T cells and prevent disease develop‑ ment. Although the identity of these pos‑ sible extrinsic components is not known, a role has been postulated for naturally occurring CD4+CD25+ regulatory T (TReg) cells, transforming growth factor‑β (TGFβ) and indoleamine 2,3-dioxygenase (IDO).
of autoimmunity (FIG. 1a), these mice eventu‑ ally succumb to disease25. Furthermore, normal numbers of functionally active FOXP3+ TReg cells have been identified in Ctla4–/– mice27. Although a causal connection between TReg cells and the Ctla4–/– phenotype cannot be excluded, current observations suggest that TReg cells can limit lymphopro‑ liferation, but that loss of TReg-cell function is not the cause of the autoimmune phenotype observed in Ctla4–/– mice.
TReg cells. TReg cells express CTLA4 and have well-documented suppressive effects in vari‑ ous autoimmune disorders24. The transcrip‑ tion factor forkhead box P3 (FOXP3) has a central role in the development of TReg cells, and ablation of the Scurfin gene, which encodes FOXP3, produces a phenotype similar to that observed in Ctla4–/– mice25,26. However, although overexpression of FOXP3 in Ctla4–/– mice can delay the onset
TGFβ. Similarly, TGFβ has been proposed as a factor that can suppress the excessive lymphoproliferation of Ctla4–/– cells (FIG. 1b). Similar to Ctla4–/– mice, TGFβ-deficient mice die by 3–4 weeks of age as a result of a multiorgan inflammatory syndrome28. TGFβ is produced by TReg cells and has been shown to suppress inflammatory bowel disease (IBD)29. Although CTLA4 ligation can increase TGFβ production by
Control of Ctla4–/– T cells
a
CTLA4 TCR
Delayed onset of disease
FOXP3+
Regulatory T cell
Ctla4–/– T cell
Negative regulation of an immune response by CTLA4
b
MHC class I and/or II TGFβ Inhibition of T-cell activation?
AntigenCD80 and/or presenting cell CD86
T cell
c IDO Inhibition of T-cell function
Figure 1 | Cell-extrinsic models of CTLA4 function: cell-extrinsic factors may contribute to the hyper-lymphoproliferation phenotype of CTLA4-deficient mice. a | Reviews Control| of cytotoxic Nature Immunology T-lymphocyte antigen 4 (Ctla4)–/– T cells by regulatory T cells. Transgenic expression of the transcription factor forkhead box P3 (FOXP3) delays the hyperproliferation and the onset of disease in Ctla4–/– mice; however, these mice eventually die from disease. b,c | Other models have implicated extrinsic factors in the negative regulation of the immune response by CTLA4. Possible dysregulation in the expression of transforming growth factor‑β (TGFβ) (b) or of indoleamine 2,3-dioxygenase (IDO) (c) has been pro‑ posed to contribute to the Ctla4–/– disease phenotype. Although CTLA4-specific antibodies may induce TGFβ expression, a neutralizing antibody to TGFβ has limited effects on CTLA4 function. Similarly, although IDO can be induced by CTLA4 engagement of CD80 and/or CD86 expressed by dendritic cells, Ido-knockout mice do not show a Ctla4–/– phenotype. Although these factors may operate to suppress the immune response in specific autoimmune diseases, a causative role in the development of Ctla4–/– phenotype is questionable.
154 | february 2008 | volume 8
some T‑cell populations28, it may not do so with CD4+ T cells30, which are the cells that are needed for the induction of disease progression in Ctla4–/– mice18. Furthermore, a neutralizing antibody to TGFβ cannot reverse the inhibition of T cells induced by CTLA4 crosslinking, and ligation of CTLA4 inhibits the proliferation of both wild-type and Tgfb–/– T cells30. Although a role for TGFβ occurs in specific circumstances29, a major role for TGFβ in the development of the disease phenotype in Ctla4–/– mice appears unlikely. IDO. Another potential cell-extrinsic factor of great interest has been the tryp‑ tophan degrading enzyme IDO31 (FIG. 1c). IDO is the rate-limiting enzyme of the tryptophan metabolism pathway and has immunoregulatory effects. APCs can be induced to produce IDO by intracellular pathogens, pro-inflammatory factors such as interferon‑γ (IFNγ), and through ligation of CD80 and/or CD86 by CTLA4. However, Ido-knockout mice do not have the same phenotype as Ctla4–/– mice32. So, although CTLA4-induced IDO production may be important for the suppression of T-cell func‑ tion in niches of the immune system, it is unlikely to account for the phenotype of the Ctla4–/– mice. Overall, although a cell-extrinsic ele‑ ment may regulate the disease phenotype of Ctla4–/– mice, the nature of this element is not fully understood. It has been mis‑ takenly assumed that because wild-type cells can produce an extrinsic factor that controls the behaviour of Ctla4–/– T cells, this factor must also be the cause of the disease phenotype. Cell-intrinsic factors Cell-intrinsic factors also govern the behaviour of CTLA4+ T cells. CTLA4 and CD3 antibody co-ligation inhibits cytokine production and proliferation of T cells33,34, whereas mutations in CTLA4 can interfere with its modulation of T‑cell function35,36. The mere transfer of specific cytoplasmic sequences from CD28 — which is a costimulatory molecule expressed by T cells that also binds CD80 and CD86 — to CTLA4 results in positive signalling by CTLA4 ligation37. Several models of cellintrinsic regulation have been proposed for CTLA4 function and these include competition between CTLA4 and CD28 for ligand binding, engagement of nega‑ tive signalling proteins by CTLA4 and the inhibition of lipid-raft and microcluster formation by CTLA4. www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
Perspectives CTLA4 competition with CD28 for binding to ligand. CD80 and CD86 bind to CTLA4 with 50–100-fold greater avidity than to CD28 (Refs 7,8). This finding led to the proposal that CTLA4 acts by simply outcompeting CD28 for its ligand (FIG. 2a). This would block CD28 signalling and dampen the T‑cell response to antigen. In support of this model, the expression of a truncated CTLA4 transgene that lacks its cytoplasmic domain (which is needed for signal transduction) retarded the onset of the disease in Ctla4–/– mice38. The expression of the transgene at even higher levels than endogenous CTLA4 expression resulted in lymphoadenopathy, but not in tissue infiltration. These observations indicated that the CTLA4 ectodomain can compete with CD28 for binding to CD80 and CD86 in retarding the onset of dis‑ ease, and that the cytoplasmic domain of CTLA4 regulates tissue infiltration, rather than lymphoproliferation. Overexpression of transgenic wild-type CTLA4 also inhibited T‑cell function in a CD28- and CD80-dependent manner38,39. However, these findings are complicated by the observation that CD28 expression is also needed for the development of the dis‑ ease phenotype in Ctla4–/– mice. Blockade of CD28 binding to its ligands prevents the dis‑ ease phenotype40. Similarly, the loss of CD28 or the expression of an inhibitory mutant of CD28 can prevent disease41. Whether this observation points to a special connection between CTLA4 and CD28, or is simply due to a requirement for CD28 co-stimulation in responses to self antigen is unclear. To date, no evidence has been forthcoming on the selective inhibition of CD28 or TCR signal‑ ling by CTLA4 (Refs 42,43). Although CTLA4 ligation can displace CD28 at the immuno‑ logical synapse44, both CD28-dependent and CD28-independent inhibition of T‑cell acti‑ vation by CTLA4 has been described39,45–47. Overall, the above findings fit with the model in which CTLA4 can compete to some degree with CD28 for ligand binding in dampening the responses of T cells to antigen. However, T cells that lack CTLA4 expression only elicit the full proliferative response that leads to the development of disease or heightened antigen responses in the presence of CD28 co-stimulation. Engaging negative signalling pathways. A second model proposes that CTLA4 can exert its inhibitory function by engaging negative signalling molecules that regulate TCR signalling. Phosphatases remove phosphate groups from proteins, often
Competition
a
MHC class I/II
TCR Competition for CD80 and/or CD86 reduces CD28 co-stimulation and delays onset of disease
APC
CD80 and/or CD86
CTLA4
T cell CD28
Regulation of signalling components
b SHP2 PP2A
Possible inhibition of immune response
c Upregulation of CBL-B raises TCR signalling threshold — possible immune regulation? ↑ CBL-B
Inhibition of lipid-raft and microcluster formation Lipid raft d Downregulation of lipid-raft formation inhibits immune response?
e
ZAP70 microcluster Disruption of ZAP70-containing microcluster formation impairs TCR activation
Figure 2 | Cell-intrinsic models of CTLA4 function: cell-intrinsic factors contribute to the | Immunology function of T cells from CTLA4-deficient mice. a | RepresentationNature of theReviews competition model. Cytotoxic T-lymphocyte antigen 4 (CTLA4) high-avidity binding to CD80 and/or CD86 can outcompete CD28 for binding for the same ligands and impair co-stimulation. In support of this model, overexpression of a truncated form of CTLA4 that lacks a cytoplasmic domain and is therefore unable to induce intracellular signals, can partially delay the onset of disease. b,c | Representation of the model supporting the regulation of signalling components. CTLA4 can bind to intracellular phosphatases (PTPases) SRC homology 2 (SH2)-domain-containing PTPase 2 (SHP2) and protein phosphatase 2A (PP2A) (b). These could act to inhibit the function of T-cell receptor (TCR) signalling targets by dephosphorylation. Likewise, CTLA4 may upregulate expression of the E3 ligase Casitas B‑lineage lymphoma B (CBL‑B) (c). However, although CBL‑B-deficient mice are prone to spontane‑ ous or antigen-induced autoimmunity, they fail to develop the lethal severity that is observed in CTLA4-deficient mice. d,e | Representation of the model whereby CTLA4 co-ligation with the TCR impairs the formation of lipid rafts and ζ‑chain-associated protein kinase of 70 kDa (ZAP‑70) micro‑ clusters. CTLA4 could impair signalling and raise the threshold needed for TCR signalling by a reduction in lipid-raft formation on the surface of T cells (d). This contrasts with the ability of CD28 to upregulate lipid-raft formation. These opposing effects of CD28 versus CTLA4 on lipid-raft forma‑ tion could fit a model in which co-receptors modulate activation by simply increasing or decreasing the availability of key TCR-signalling components in lipid rafts. Likewise, CTLA4 could impair TCR signalling by reducing ZAP70 microcluster formation (e). These ZAP70 microclusters are needed for the TCR-driven phosphotryosine activation cascade.
inhibiting their function, and indeed CTLA4 is thought to bind the protein tyrosine phos‑ phatases (PTPases) SHP2 (SRC homology 2 (SH2)-domain-containing PTPase 2)42,48,49 and PP2A (protein phosphatase 2A)50,51 (FIG. 2b). Antibody-induced co-ligation of
nature reviews | immunology
CTLA4 with the TCR can reduce tyrosine phosphorylation of linker for the activation of T cells (LAT) and inhibit the activation of the extracellular-signal-regulated kinases (ERKs)42,52,53. LAT is an adaptor protein that recruits other binding proteins, whereas volume 8 | february 2008 | 155
© 2008 Nature Publishing Group
Perspectives ERKs are key serine/threonine kinases that are required for the production of inter‑ leukin‑2 (IL‑2). However, a limitation of this model is that binding of CTLA4 to SHP2 must be indirect as it has been shown more recently that CTLA4 lacks SHP2-binding sites49, and a direct connection between SHP2 and the dephosphorylation of TCR substrates has not been made. Another limitation is that SHP2 has generally been found to be a positive regulator of ERKs54. Despite these limitations, it remains possible that SHP2 (or related SHP1) binding may dephosphorylate certain TCR signalling proteins and reduce TCR-induced T‑cell activation. By contrast, PP2A has a clear inhibitory effect on cell growth by inhibiting protein kinase B (PKB), a kinase that is needed for multiple functions in T cells. Although promising, this model is limited by the relatively low stoichiometry of PP2A binding to CTLA4 relative to its great abundance in cells, and by reports that show that PP2A binds to CD28 (Ref. 55). This has necessitated the proposal of models whereby PP2A dis‑ sociates from CD28 (but not CTLA4) upon TCR ligation, or in which PP2A operates to inhibit the inhibitory signals mediated by CTLA4 (Ref. 50). The exact function of PP2A in this array of opposing possibilities needs clarification. Lastly, CTLA4 ligation may increase the expression of the E3 ligase Casitas B‑lineage lymphoma B (CBL‑B), an E3 ligase of the ubiquitylation pathway that has been linked to autoimmunity56–58 (FIG. 2c). However, although CBL‑B-deficient mice are prone to spontaneous or antigen-induced autoimmu‑ nity, they fail to develop the lethal severity that is observed in Ctla4–/– mice57–59.
Inhibition of lipid-raft and microcluster formation. TCR ligation induces within seconds the formation of lipid rafts and microclusters that act as platforms for signalling in T cells. These lipid rafts or glycolipid-enriched microdomains (GEMs) are enriched in cholesterol and signalling proteins. CTLA4 can be compartmentalized within these lipid rafts50, but has been reported to block lipid-raft formation on the surface of naive T cells53,60 (FIG. 2d) and to reduce the presence of the TCR ζ‑chain in the lipid rafts61. By contrast, CD28 has the opposite effect by potently upregulating lipid-raft formation following TCR ligation62. A model that takes into account the opposing effects of CD28 and CTLA4 on the expression of lipid rafts proposes that these receptors differentially regulate T‑cell activation by simply increasing or decreasing the formation of lipid rafts and thereby the availability of key components in the lipid rafts that are needed for TCR signalling. The limitations of this model are that the effect of CTLA4 on pre-activated T cells or secondary responses is less pronounced than on naive T cells (C.E.R., unpublished observations), and that there is an ongoing debate as to the relevance of lipids rafts in TCR signalling. A similar model has been proposed on the basis that CTLA4 ligation has recently been shown to inhibit microcluster forma‑ tion on T cells (BOX 1). Microclusters of the TCR complex and signalling proteins form within seconds following TCR ligation and are the sites of tyrosine phosphorylation63,64. The kinase ζ‑chain-associated protein kinase of 70 kDa (ZAP70) and adaptors, such as LAT and SH2-domain-containing leukocyte protein of 76 kDa (SLP76) form microclus‑ ters63,64. Intriguingly, Schneider et al. showed
Box 1 | Microclusters and the immunological synapse T cells interact with antigen-bearing antigen-presenting cells (APCs) to form conjugates. The interface between the T cell and APC is termed the immunological synapse and involves the engagement of surface receptors, such as the T‑cell receptor (TCR), CD4, CD8 CD28 and cytotoxic T‑lymphocyte antigen 4 (CTLA4). Although initial work focused on an arrangement of the receptors that was termed the supramolecular activation cluster (SMAC), it has proved difficult to demonstrate a role for the SMAC in signalling as SMAC formation occurs after the initiation of tyrosine phosphorylation. Instead, modern imaging techniques have revealed the formation of higher-order macromolecular signalling complexes, termed microclusters. These clusters of 100 nm to 500 nm in diameter form within seconds following TCR ligation, they serve as sites of TCRinduced tyrosine phosphorylation and comprise receptors or signalling proteins (such as the protein tyrosine kinases ZAP70 (ζ-chain-associated protein kinase of 70 kDa)) or adaptor proteins (such as LAT (linker for activation of T cells), SLP76 (SH2-domain-containing leukocyte protein of 76 kDa) and GADS (GRB2-related adaptor protein)). TCR microclusters tend to form around the peripheral region of the contact area and then migrate towards a central region where they are eliminated or endocytosed. Adaptor proteins, such as SLP76 and GADS, are found within the same microclusters, whereas ZAP70 and SLP76 segregate into separate microclusters. These separate microclusters appear to interact transiently with each other. A single microcluster may be sufficient to induce increased amounts of intracellular calcium and T‑cell activation.
156 | february 2008 | volume 8
that co-ligation of the TCR and CTLA4 can effectively block the formation of ZAP70containing microclusters in T cells65 (FIG. 2e). This occurred concurrently with a loss of calcium mobilization that is needed for T‑cell proliferation65. Impaired microcluster formation could explain the reduced phos‑ phorylation of TCR substrates that has been observed without involving PTPases. Overall, the aforementioned studies demonstrate a negative effect of CTLA4 on events that are linked to T‑cell activation. The contentious issue concerns the degree to which they operate in different systems, and whether they reflect cause or effect. The ‘reverse stop-signal model’ for CTLA4 Given that previous models fail to fully account for the overall function of CTLA4, we considered the possibility that the func‑ tion of CTLA4 might be mediated by a completely different mechanism to the ones proposed in the previous section. Some of our earlier studies showed that CTLA4 can in fact engage positive signalling pathways. It shares with CD28 the ability to bind to the lipid kinase phosphoinositide 3‑kinase (PI3K)66 and can activate JUN kinase (JNK)52. PI3K activates lipids that recruit proteins to the inner surface of the plasma membrane, whereas JNK activates various transcription factors, such as JUN. In view of these results, the question arose as to whether the model of CTLA4 as a ‘negative signalling co-receptor’ was in fact correct. Alternatively, was it possible that CTLA4 could function by actively engaging a posi‑ tive event that could produce a negative out‑ come in terms of dampening the response to antigen? To address these questions we therefore set out to study the mechanisms underlying CTLA4 function from a different perspective. Using CTLA4-specific antibodies or CD80–immunoglobulin fusion proteins, we showed that CTLA4 engagement increased clustering of lymphocyte function‑associated antigen 1 (LFA1; also known as αLβ2integrin) on the surface of T cells and its binding to intercellular adhesion molecule 1 (ICAM1) expressed by APCs67. T-cell motil‑ ity is achieved by a combination of changes in the affinity and avidity of LFA1 for its ligand. These effects on LFA1 clustering were greater than those observed follow‑ ing activation of the TCR or CD28. Even classic TCR ‘inside-out’ signalling for LFA1 activation was partially dependent on the expression of CTLA4, as shown by impaired adhesion of Ctla4–/– T cells67. Consistent with this, two groups also independently showed www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
Perspectives that CTLA4 can activate the well-established activator of integrins70, the GTPase RAP1 (Refs 67,69).
Given the need for the activation of LFA1 in motility (BOX 2), we next showed that CTLA4 induced the rapid polarization of T cells70 and increased the motility of mouse and human T cells71. This observation in turn provided a clue to an alternate explana‑ tion for the ‘inhibitory effect’ of CTLA4 on T‑cell function. TCR ligation has long been known to slow or reduce T‑cell motil‑ ity; an event termed the ‘stop signal’72–75. This reduction in T-cell motility was first observed in vitro, and later in more sophis‑ ticated in vivo studies using laser-scanning excitation microscopy72–75. The key point is that reduced motility is needed for T cells to stably conjugate with APCs. Without the stop signal, T cells continue to move or remain tethered to the APC but are unable to form an extended interface with the APC and an immunological synapse. Stable immunological-synapse formation is needed for the scanning of peptide–MHC com‑ plexes, their engagement by the TCR and the induction of signalling cascades (FIG. 3a). In this context, we next made the key observation that CTLA4 ligation can effec‑ tively reverse or override the TCR-induced stop signal, such that the T cells continue to move as if they never encountered antigen65,71 (FIG. 3b). Two-photon micro‑ scopy analysis of selected CTLA4+ and CTLA4– T cells from DO11.10 transgenic mice demonstrated this event even more graphically in vivo where CTLA4+ T cells continued to move rapidly in the presence of peptide antigen in mesenteric lymph nodes at all time points examined (15–24 hours) following antigen injection. CTLA4– T cells slowed at time points 15–20 hours following antigen injection, as previously reported for mixed populations of T cells72–75. Similarly, CTLA4+ T cells engaged with A20 B cells or isolated dendritic cells (DCs) for shorter periods compared with CTLA4– cells in response to peptide antigen in vitro. This in turn correlated with reduced cytokine production and proliferation73. These findings were paralleled in another study in which a T‑cell hybridoma carrying a TCR specific for tetanus toxin peptide was transfected with CTLA4 (Ref. 65). The mere expression of CTLA4 prevented the cells from forming extended regions of contact with APCs. Although the T cells could remain tethered to the APCs, they showed shorter dwell times and following dissociation con‑ tinued to move on to another target. Those T cells that did remain tethered were often
Box 2 | Integrin adhesion and T‑cell motility T cells are highly motile, achieving speeds of 10–15 µm per minute. This fast motility is necessary for T cells to migrate along vessel surfaces to lymph nodes and sites of inflammation where they can become activated by forming conjugates with antigen-presenting cells (APCs). The adhesion and migration of lymphocytes across high endothelial venules is achieved by actin tread-milling and actin–myosin contractions and is mediated by cell-surface, activated integrins, which are heterodimers of covalently linked α and β chains. The key integrin expressed by T cells is lymphocyte function-associated antigen 1 (LFA1; also known as αLβ2-integrin)81,86 and it binds to intercellular adhesion molecule 1 (ICAM1), ICAM2 and ICAM3, which are expressed by APCs and by epithelial cells that line the blood vessels. Integrins expressed by resting T cells are inactive and become activated following T‑cell-receptor ligation or by chemokines through inside-out signalling. Integrin activation is achieved through two different mechanisms, by causing conformational changes in the integrin itself, which results in an increase in ligand affinity, or by integrin clustering, which results in increased ligand avidity. Although high-affinity LFA1mediated adhesion can result in cell immobilization, it is also required for increased cell motility.
seen dangling at the point of contact with the APCs. Concurrent with this impaired contact, the cells failed to form ZAP70-containing microclusters and to mobilize intracellular calcium65. Calcium mobilization is needed for the induction of many events in T‑cell activation, including the movement of the transcription factor NFAT (nuclear factor of activated T cells) into the nucleus for tran‑ scription of the Il2 gene. Overall, we propose that CTLA4 enhances motility, reverses the TCR ‘stop signal’ and acts as a gatekeeper of conjugation by reduc‑ ing the time of contact between T cells and APCs. Thus, the reverse stop-signal model is fundamentally different from previously proposed models and can potentially explain multiple aspects of CTLA4 function and the phenotype of the Ctla4–/– mouse. Potential implications First, at the most elementary level, the effects of CTLA4 on T-cell–APC dwell times could explain the basis for the restricted expression of CTLA4 on the surface of activated T cells. Naive T cells require a much longer contact time with APCs to become activated. Expression of CTLA4 by these T cells would inappropriately limit the contact time with the APC and prevent primary responses to antigen. By contrast, secondary responses and re-stimulation by antigen require shorter periods of contact for effective antigen recognition. In this manner, CTLA4 could act to fine-tune anti‑ gen recognition to allow for re-stimulation of activated and memory T cells in a more efficient manner. Second, the effect on T-cell–APC dwell times could explain previous findings that CTLA4 can raise the threshold needed for T‑cell activation. By limiting the interac‑ tion time between the T cell and the APC, and limiting the formation of an immuno‑ logical synapse with an enforced contact
nature reviews | immunology
area, CTLA4 could limit the number of TCR–MHC ligation events, microcluster for‑ mation and signalling. This could raise the threshold that the TCR needs to overcome to induce productive signalling. T cells express‑ ing CTLA4 would be expected to select for more avid TCR–peptide–MHC interactions with a capacity to generate more robust signals and be less reactive to low‑affinity self antigen. Third, alterations in the stop signal could contribute to the development of autoimmu‑ nity in Ctla4–/– mice. By limiting dwell times, CTLA4 would prevent responses to lower affinity ligands such as self antigen. In the absence of CTLA4, a subset of CD4+ T cells remain attached to APCs for extended periods (>6 hours)71. This extended contact time could allow TCR interactions with lowaffinity self antigens to result in tonic signal‑ ling and activation. This could be especially important in secondary responses in which less contact time is needed to induce proinflammatory cytokine production. Fourth, our findings could account for the requirement of CTLA4 for the induction of T‑cell anergy76,77. CTLA4 induced the same tethering between T cells and DCs that has been reported in tolerance induc‑ tion78,79. Tolerance also leads to a more rapid restoration of motility58,59, as is observed with CTLA4. The time limitation on the T‑cell–APC interaction and ligation of receptors by CTLA4 may lead to the genera‑ tion of partial-activation signals that can induce T‑cell anergy 80. Fifth, our proposed model could also account for the massive T‑cell tissue infil‑ tration that is observed in Ctla4–/– mice. Integrins are needed for trans-endothelial T‑cell migration and homing to peripheral compartments81,82. By altering integrin activation, CTLA4 would also be predicted to alter the migration of T cells to periph‑ eral tissues and sites of inflammation. volume 8 | february 2008 | 157
© 2008 Nature Publishing Group
Perspectives No CTLA4 ligation
a
CD28
CD80 and/or CD86
Stable, reinforced contact
Tethered
ZAP70 microcluster DC
Stop signal
Chemokines
↑ Ca2+
TCR
ICAM1
LFA1
Peptide–MHC complex
IL-2
CTLA4 ligation
b
Tethered, weak contact
CTLA4
No stop signal
↓ ↓ ↓
ZAP70 microcluster formation Ca2+ IL-2
Figure 3 | The ‘reverse stop-signal’ model. a | T cells normally make initial contact with antigen-presenting cells (APCs) owing to exposure to chem‑ okines this is followed by T‑cell receptor (TCR) ligation, which leads to ‘inside-out’ signalling that upregulates the expression of lymphocyte function-associated antigen 1 (LFA1) adhesion molecules and induces a ‘stop signal’ that allows for enforced contact of the T cell with the APC and efficient TCR ligation. This leads to ζ‑chain-associated protein kinase of 70 kDa (ZAP70) microcluster formation, calcium mobilization and eventually cytokine production (for example, interleukin‑2 (IL‑2)). b | Cytotoxic T‑lymphocyte antigen 4 (CTLA4) ligation reverses the stop signal, thereby interfering with the formation of the immunological syn‑ apse and enforced contact with APCs. Instead, although T cells remain tethered, they fail to make enforced contact with APCs and have shorter dwell times. This leads to a reduction in ZAP70 microcluster formation,
Lymphocytes use integrins for adhesion and transmigration across high endothelial venules. CTLA4-mediated integrin activa‑ tion might facilitate this process, allowing access of activated T cells to sites of inflam‑ mation. In lymph nodes, CTLA4+ T cells might engage APCs for shorter time periods followed by increased motility leading to a more frequent re-encounter with other APCs. This may result in less pronounced asymmetric cell division and faster egress from the lymph nodes. The major potential limiting factor of these effects could be the as yet established effects of chemokines on CTLA4-mediated function. It remains possible that chem‑ okines may counter the effects of CTLA4 on T cells, in which case the outcome will depend on the local concentration and type of chemokine(s) involved. In the context of the phenotype of Ctla4–/– mice, we recently reported that T cells from these diseased mice are actually resistant to the induction of a stop signal by CD3
Continued motility
calcium mobilization and IL‑2 production. This would reduce the effi‑ Natureneeded Reviewsfor | Immunology ciency of T‑cell activation, and raise the threshold productive TCR signalling. It could also account for CTLA4 involvement in the induc‑ tion of T‑cell anergy by reducing the efficacy of conjugation, whereas the prolonged interaction of certain CTLA4-deficient T cells could allow for reactivity to self antigen. By contrast, with the shorter dwell-times, CTLA4 might operate to optimize secondary effector responses. The reduced dwell time would provide enough time for effector T cells (that is, cyto‑ toxic T lymphocytes) to engage their target and move to the next target. This would enhance the efficiency of killing of tumour or virally infected cells by maximizing the number of cells killed within a given time frame. In lymph nodes, the more rapid dissociation from the APCs could also lead to more frequent re-encounters with other APCs and faster egress from the lymph node.
ligation83. This feature might have developed owing to chronic inflammation leading to the desensitization of the TCR complex. Alternatively, it could represent a population of autoreactive T cells that have become resistant to TCR-mediated regulation. In either case, this decoupling event would be expected to promote more extensive migra‑ tion of activate T cells into tissues, resulting in delocalized inflammatory responses. Last, our model provides an alternative explanation for enhanced antitumour killing by CTLA4-specific antibodies84. The exact nature of the action of CTLA4-specific anti‑ bodies in clinical trials is not clear, although it has been assumed that they block a nega‑ tive signal. Although some CTLA4-specific antibodies may block negative signals through CD80 and/or CD86, some CTLA4specific antibodies may simply reduce dwell times between cytolytic T cells (CTLs) and tumour cells. Given that the CTL killing of targets requires a shorter contact period than in the case of antigen presentation, CTLA4
158 | february 2008 | volume 8
ligation would provide just enough time for the CTL killing of the tumour target fol‑ lowed by a rapid move to the next target. In this way, CTLA4 ligation by antibody would enhance the efficiency of tumour killing by maximizing the number of targets killed within a given time frame. Overall, this may underscore a principal function for CTLA4 — the ‘fine tuning’ or optimization of sec‑ ondary effector responses. It may also oper‑ ate in other contexts, such as in the release of cytokines, with B cells and other cell types. With this controlled reduction in dwell times and increased motility, CTLA4 may have evolved to maximize the interactions of T cells with as many targets as possible over a given time frame. Future prospects Motility and migration have key roles in various biological processes, including neuronal differentiation and embryo‑ genesis. Surprisingly, little is understood regarding the role of these events in thymic www.nature.com/reviews/immunol
© 2008 Nature Publishing Group
Perspectives differentiation and mature T‑cell function. The effects of CTLA4 on motility, migra‑ tion and T‑cell–APC conjugation under‑ scores the general potential importance of cell–cell contact times in T‑cell immunity and in the development of immune pathologies. Outstanding questions concern the rela‑ tionship between our and previously reported models. The models are not mutually exclusive. For example, the reverse stop-signal model is compatible with the competition between CTLA4 and CD28 for binding to CD80 and/or CD86. CD28 engagement might promote the TCR stop signal and more stable contact with APCs, which would be reversed by CTLA4 binding. Likewise, the finding that the cytoplasmic domain of CTLA4 influences tissue infiltration38 is compatible with CTLA4 regulation of motil‑ ity. Similarly, CTLA4 could recruit PTPases such as PP2A in the reversal of the stop signal. Last, CTLA4 might generate signals that affect the stop signal without CD80 and/or CD86 binding and account for the ligandindependent function of CTLA4 (Ref. 85). Another question is the degree to which the reduction in ZAP70-containing micro‑ cluster formation is due to some signalling event, or due to altered adhesion and motil‑ ity. The reduction occurred concurrently with increased cell motility, such that the contact with a CD3-specific antibody on slides was less firm and stable65. This sug‑ gests, but does not prove, cause and effect. It is still possible that two parallel CTLA4 signalling processes occur; one that alters cell motility and another that inhibits TCRmediated signalling that leads to ZAP70containing microcluster formation. Further questions remain, such as what are the overall effects of CTLA4 on migration to compartments of the immune system and does the reversal of the stop signal operate equally in all T‑cell subsets? To date, our analysis has focused on CD4+ T cells. Nevertheless, the conjugation time was scored as an average within a population of CD4+ T cells with varying degrees of response from the mean. A small percent of CD4+ T cells failed to respond to the reversal of the stop signal. These cells may represent a subset of different T‑cell populations (such as TReg cells) or T cells in a different stage of the cell cycle. Also, CD8+ T cells tend to require longer interaction times with APCs than do CD4+ T cells. Does this mean that CTLA4 is less active on this subset of T cells? The last prediction is that CTLA4 should induce more frequent interactions
with APCs in the lymph nodes, but the consequences are still unclear. One pos‑ sibility is that TCR-activated T cells will exit or egress the lymph nodes more rapidly. In this context, it will be important to address how CTLA4-induced modula‑ tion of cell adhesion can be modified by cell‑extrinsic factors, such as chemokines. Christopher E. Rudd is at the Cell Signalling Section, Division of Immunology, Department of Pathology, Tennis Court Road, CB1 4QP Cambridge, UK; Cambridge Institute for Medical Research, Hills Road, CB2 0XY Cambridge, UK. e‑mail: [email protected] 1. 2. 3. 4. 5. 6. 7.
8. 9.
10. 11.
12. 13. 14.
15.
16. 17.
18.
19.
20.
Riley, J. L. & June, C. H. The CD28 family: a T‑cell rheostat for therapeutic control of T‑cell activation. Blood 105, 13–21 (2005). Chambers, C. A. et al. The role of CTLA‑4 in the regulation and initiation of T‑cell responses. Immunol. Rev. 153, 27–46 (1996). Abbas, A. K., Lohr, J., Knoechel, B. & Nagabhushanam, V. T cell tolerance and autoimmunity. Autoimmun. Rev. 3, 471–475 (2004). Greenwald, R. J., Latchman, Y. E. & Sharpe, A. H. Negative co-receptors on lymphocytes. Curr. Opin. Immunol. 14, 391–396 (2002). Rudd, C. E. & Schneider, H. Unifying concepts in CD28, ICOS and CTLA4 co-receptor signalling. Nature Rev. Immunol. 3, 544–556 (2003). Alegre, M. L., Frauwirth, K. A. & Thompson, C. B. T‑cell regulation by CD28 and CTLA‑4. Nature Rev. Immunol. 1, 220–228 (2001). Zhang, X., Schwartz, J. C., Almo, S. C. & Nathenson, S. G. Crystal structure of the receptor-binding domain of human B7–2: insights into organization and signaling. Proc. Natl Acad. Sci. USA 100, 2586–2591 (2003). Stamper, C. C. et al. Crystal structure of the B7–1/ CTLA‑4 complex that inhibits human immune responses. Nature 410, 608–611 (2001). van der Merwe, P. A., Bodian, D., Daenke, S., Linsley, P. & Davis, S. J. CD80 (B7–1) binds both CD28 and CTLA‑4 with a low affinity and very fast kinetics. J. Exp. Med. 185, 393–403 (1997). Linsley, P. S. et al. Intracellular trafficking of CTLA‑4 and focal localization towards sites of TCR engagement. Immunity 4, 535–543 (1996). Egen, J. G. & Allison, J. P. Cytotoxic T lymphocyte antigen‑4 accumulation in the immunological synapse is regulated by TCR signal strength. Immunity 16, 23–35 (2002). Valk, E. et al. T cell receptor-interacting molecule acts as a chaperone to modulate surface expression of the CTLA‑4 coreceptor. Immunity 25, 807–821 (2006). Zhang, Y. & Allison, J. P. Interaction of CTLA‑4 with AP50, a clathrin-coated pit adaptor protein. Proc. Natl Acad. Sci. USA 94, 9273–9278 (1997). Schneider, H. et al. Cytolytic T lymphocyte-associated antigen‑4 and the TCR zeta/CD3 complex, but not CD28, interact with clathrin adaptor complexes AP‑1 and AP‑2. J. Immunol. 163, 1868–1879 (1999). Shiratori, T. et al. Tyrosine phosphorylation controls internalization of CTLA‑4 by regulating its interaction with clathrin-associated adaptor complex AP‑2. Immunity 6, 583–589 (1997). Waterhouse, P. et al. Lymphoproliferative disorders with early lethality in mice deficient in Ctla‑4. Science 270, 985–988 (1995). Tivol, E. A. et al. Loss of CTLA‑4 leads to massive lymphoproliferation and fatal multiorgan tissue destruction, revealing a critical negative regulatory role of CTLA‑4. Immunity 3, 541–547 (1995). Chambers, C. A., Sullivan, T. J. & Allison, J. P. Lymphoproliferation in CTLA‑4‑deficient mice is mediated by costimulation-dependent activation of CD4+ T cells. Immunity 7, 885–895 (1997). Gozalo-Sanmillan, S., McNally, J. M., Lin, M. Y., Chambers, C. A. & Berg, L. J. Cutting edge: two distinct mechanisms lead to impaired T cell homeostasis in Janus kinase 3‑ and CTLA‑4‑deficient mice. J. Immunol. 166, 727–730 (2001). Ueda, H. et al. Association of the T‑cell regulatory gene CTLA4 with susceptibility to autoimmune disease. Nature 423, 506–511 (2003).
nature reviews | immunology
21. Chen, Z., Stockton, J., Mathis, D. & Benoist, C. Modeling CTLA4-linked autoimmunity with RNA interference in mice. Proc. Natl Acad. Sci. USA 103, 16400–16405 (2006). 22. Bachmann, M. F., Kohler, G., Ecabert, B., Mak, T. W. & Kopf, M. Cutting edge: lymphoproliferative disease in the absence of CTLA‑4 is not T cell autonomous. J. Immunol. 163, 1128–1131 (1999). 23. Bachmann, M. F. et al. Normal responsiveness of CTLA‑4‑deficient anti-viral cytotoxic T cells. J. Immunol. 160, 95–100 (1998). 24. Sakaguchi, S. & Powrie, F. Emerging challenges in regulatory T cell function and biology. Science 317, 627–629 (2007). 25. Khattri, R., Cox, T., Yasayko, S. A. & Ramsdell, F. An essential role for Scurfin in CD4+CD25+ T regulatory cells. Nature Immunol. 4, 337–342 (2003). 26. Fontenot, J. D., Gavin, M. A. & Rudensky, A. Y. Foxp3 programs the development and function of CD4+CD25+ regulatory T cells. Nature Immunol. 4, 330–336 (2003). 27. Read, S. et al. Blockade of CTLA‑4 on CD4+CD25+ regulatory T cells abrogates their function in vivo. J. Immunol. 177, 4376–4383 (2006). 28. Chen, W., Jin, W. & Wahl, S. M. Engagement of cytotoxic T lymphocyte-associated antigen 4 (CTLA‑4) induces transforming growth factor beta (TGF-β) production by murine CD4+ T cells. J. Exp. Med. 188, 1849–1857 (1998). 29. Izcue, A., Coombes, J. L. & Powrie, F. Regulatory T cells suppress systemic and mucosal immune activation to control intestinal inflammation. Immunol. Rev. 212, 256–271 (2006). 30. Sullivan, T. J. et al. Lack of a role for transforming growth factor-β in cytotoxic T lymphocyte antigen‑4‑mediated inhibition of T cell activation. Proc. Natl Acad. Sci. USA 98, 2587–2592 (2001). 31. Munn, D. H., Sharma, M. D. & Mellor, A. L. Ligation of B7–1/B7–2 by human CD4+ T cells triggers indoleamine 2,3-dioxygenase activity in dendritic cells. J. Immunol. 172, 4100–4110 (2004). 32. Mellor, A. L. & Munn, D. H. IDO expression by dendritic cells: tolerance and tryptophan catabolism. Nature Rev. Immunol. 4, 762–774 (2004). 33. Krummel, M. F. & Allison, J. P. CTLA‑4 engagement inhibits IL‑2 accumulation and cell cycle progression upon activation of resting T cells. J. Exp. Med. 183, 2533–2540 (1996). 34. Walunas, T. L. et al. CTLA‑4 can function as a negative regulator of T cell activation. Immunity 1, 405–413 (1994). 35. da Rocha Dias, S. & Rudd, C. E. CTLA‑4 blockade of antigen-induced cell death. Blood 97, 1134–1137 (2001). 36. Yi, L. A., Hajialiasgar, S. & Chuang, E. Tyrosinemediated inhibitory signals contribute to CTLA‑4 function in vivo. Int. Immunol. 16, 539–547 (2004). 37. Yin, L., Schneider, H. & Rudd, C. E. Short cytoplasmic SDYMNM segment of CD28 is sufficient to convert CTLA‑4 to a positive signaling receptor. J. Leukoc. Bio. 73, 178–182 (2003). 38. Masteller, E. L., Chuang, E., Mullen, A. C., Reiner, S. L. & Thompson, C. B. Structural analysis of CTLA‑4 function in vivo. J. Immunol. 164, 5319–5327 (2000). 39. Engelhardt, J. J., Sullivan, T. J. & Allison, J. P. CTLA‑4 overexpression inhibits T cell responses through a CD28–B7‑dependent mechanism. J. Immunol. 177, 1052–1061 (2006). 40. Tivol, E. A. et al. CTLA4Ig prevents lymphoproliferation and fatal multiorgan tissue destruction in CTLA‑4‑deficient mice. J. Immunol. 158, 5091–5094 (1997). 41. Tai, X., Van Laethem, F., Sharpe, A. H. & Singer, A. Induction of autoimmune disease in CTLA‑4–/– mice depends on a specific CD28 motif that is required for in vivo costimulation. Proc. Natl Acad. Sci. USA 104, 13756–13761 (2007). 42. Marengere, L. E. et al. Regulation of T cell receptor signaling by tyrosine phosphatase SYP association with CTLA‑4. Science 272, 1170–1173 (1996). 43. 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). 44. Pentcheva-Hoang, T., Egen, J. G., Wojnoonski, K. & Allison, J. P. B7–1 and B7–2 selectively recruit CTLA‑4 and CD28 to the immunological synapse. Immunity 21, 401–413 (2004). 45. Lin, H. et al. Cytotoxic T lymphocyte antigen 4 (CTLA4) blockade accelerates the acute rejection of cardiac allografts in CD28-deficient mice: CTLA4 can
volume 8 | february 2008 | 159 © 2008 Nature Publishing Group
Perspectives
46. 47.
48. 49.
50. 51. 52.
53.
54.
55. 56. 57. 58. 59. 60.
function independently of CD28. J. Exp. Med. 188, 199–204 (1998). Hwang, K. W. et al. Transgenic expression of CTLA‑4 controls lymphoproliferation in IL‑2‑deficient mice. J. Immunol. 173, 5415–5424 (2004). Fallarino, F., Fields, P. E. & Gajewski, T. F. B7–1 engagement of cytotoxic T lymphocyte antigen 4 inhibits T cell activation in the absence of CD28. J. Exp. Med. 188, 205–210 (1998). Lee, K. M. et al. Molecular basis of T cell inactivation by CTLA‑4. Science 282, 2263–2266 (1998). Schneider, H. & Rudd, C. E. Tyrosine phosphatase SHP‑2 binding to CTLA‑4: absence of direct YVKM/ YFIP motif recognition. Biochem. Biophys. Res. Commun. 269, 279–283 (2000). Teft, W. A., Kirchhof, M. G. & Madrenas, J. A molecular perspective of CTLA‑4 function. Annu. Rev. Immunol. 24, 65–97 (2006). Chuang, E. et al. The CD28 and CTLA‑4 receptors associate with the serine/threonine phosphatase PP2A. Immunity 13, 313–322 (2000). Schneider, H. et al. Cutting edge: CTLA‑4 (CD152) differentially regulates mitogen-activated protein kinases (extracellular signal-regulated kinase and c‑Jun N‑terminal kinase) in CD4+ T cells from receptor/ligand-deficient mice. J. Immunol. 169, 3475–3479 (2002). Martin, M., Schneider, H., Azouz, A. & Rudd, C. E. Cytotoxic T lymphocyte antigen 4 and CD28 modulate cell surface raft expression in their regulation of T cell function. J. Exp. Med. 194, 1675–1681 (2001). Frearson, J. A. & Alexander, D. R. The phosphotyrosine phosphatase SHP‑2 participates in a multimeric signaling complex and regulates T cell receptor (TCR) coupling to the Ras/mitogen-activated protein kinase (MAPK) pathway in Jurkat T cells. J. Exp. Med. 187, 1417–1426 (1998). Chambers, C. A. & Allison, J. P. Costimulatory regulation of T cell function. Curr. Opin. Cell Biol. 11, 203–210 (1999). Li, D. et al. Cutting Edge: Cbl-b: one of the key molecules tuning CD28- and CTLA-4-mediated T cell costimulation. J. Immunol. 173, 7135–7139 (2004). Chiang, Y. J. et al. Cbl‑b regulates the CD28 dependence of T‑cell activation. Nature 403, 216–220 (2000). Bachmaier, K. et al. Negative regulation of lymphocyte activation and autoimmunity by the molecular adaptor Cbl‑b. Nature 403, 211–216 (2000). Loeser, S. et al. Spontaneous tumor rejection by cbl‑b‑deficient CD8+ T cells. J. Exp. Med. 204, 879–891 (2007). Rudd, C. E., Martin, M. & Schneider, H. CTLA‑4 negative signaling via lipid rafts: A new perspective. Sci. STKE 2002, PE18 (2002).
61. Chikuma, S., Imboden, J. B. & Bluestone, J. A. Negative regulation of T cell receptor-lipid raft interaction by cytotoxic T lymphocyte-associated antigen 4. J. Exp. Med. 197, 129–135 (2003). 62. Viola, A., Schroeder, S., Sakakibara, Y. & Lanzavecchia, A. T lymphocyte costimulation mediated by reorganization of membrane microdomains. Science 283, 680–682 (1999). 63. Yokosuka, T. et al. Newly generated T cell receptor microclusters initiate and sustain T cell activation by recruitment of Zap70 and SLP‑76. Nature Immunol. 6, 1253–1262 (2005). 64. Bunnell, S. C. et al. Persistence of cooperatively stabilized signaling clusters drives T‑cell activation. Mol. Cell Biol. 26, 7155–7166 (2006). 65. Schneider, H., Smith, X., Liu, H., Bismuth, G. & Rudd, C. E. CTLA‑4 expression disrupts ZAP‑70 microcluster formation, T‑cell/APC conjugation and calcium mobilization. Eur. J. Immunol. 38, 40–47 (2007). 66. Schneider, H., Prasad, K. V., Shoelson, S. E. & Rudd, C. E. CTLA‑4 binding to the lipid kinase phosphatidylinositol 3‑kinase in T cells. J. Exp. Med. 181, 351–355 (1995). 67. Schneider, H., Valk, E., da Rocha Dias, S., Wei, B. & Rudd, C. E. CTLA‑4 up-regulation of lymphocyte function-associated antigen 1 adhesion and clustering as an alternate basis for coreceptor function. Proc. Natl Acad. Sci. USA 102, 12861–12866 (2005). 68. Katagiri, K. et al. Rap1 is a potent activation signal for leukocyte function-associated antigen 1 distinct from protein kinase C and phosphatidylinositol‑3‑OH kinase. Mol. Cell Biol. 20, 1956–1969 (2000). 69. Dillon, T. J., Carey, K. D., Wetzel, S. A., Parker, D. C. & Stork, P. J. Regulation of the small GTPase Rap1 and extracellular signal-regulated kinases by the costimulatory molecule CTLA‑4. Mol. Cell Biol. 25, 4117–4128 (2005). 70. Wei, B., da Rocha Dias, S., Wang, H. & Rudd, C. E. CTL-associated antigen‑4 ligation induces rapid T cell polarization that depends on phosphatidylinositol 3‑kinase, Vav‑1, Cdc42, and myosin light chain kinase. J. Immunol. 179, 400–408 (2007). 71. Schneider, H. et al. Reversal of the TCR stop signal by CTLA‑4. Science 313, 1972–1975 (2006). 72. Dustin, M. L., Bromley, S. K., Kan, Z., Peterson, D. A. & Unanue, E. R. Antigen receptor engagement delivers a stop signal to migrating T lymphocytes. Proc. Natl Acad. Sci. USA 94, 3909–3913 (1997). 73. Bousso, P. & Robey, E. Dynamics of CD8+ T cell priming by dendritic cells in intact lymph nodes. Nature Immunol. 4, 579–585 (2003). 74. Miller, M. J., Wei, S. H., Parker, I. & Cahalan, M. D. Two-photon imaging of lymphocyte motility and antigen response in intact lymph node. Science 296, 1869–1873 (2002).
160 | february 2008 | volume 8
75. Hugues, S. et al. Distinct T cell dynamics in lymph nodes during the induction of tolerance and immunity. Nature Immunol. 5, 1235–1242 (2004). 76. Greenwald, R. J., Boussiotis, V. A., Lorsbach, R. B., Abbas, A. K. & Sharpe, A. H. CTLA‑4 regulates induction of anergy in vivo. Immunity 14, 145–155 (2001). 77. Frauwirth, K. A., Alegre, M. L. & Thompson, C. B. CTLA‑4 is not required for induction of CD8+ T cell anergy in vivo. J. Immunol. 167, 4936–4941 (2001). 78. Shakhar, G. et al. Stable T cell-dendritic cell interactions precede the development of both tolerance and immunity in vivo. Nature Immunol. 6, 707–714 (2005). 79. Zinselmeyer, B. H. et al. In situ characterization of CD4+ T cell behavior in mucosal and systemic lymphoid tissues during the induction of oral priming and tolerance. J. Exp. Med. 201, 1815–1823 (2005). 80. Schwartz, R. H. A cell culture model for T lymphocyte clonal anergy. Science 248, 1349–1356 (1990). 81. Hogg, N., Laschinger, M., Giles, K. & McDowall, A. T‑cell integrins: more than just sticking points. J. Cell Sci. 116, 4695–4705 (2003). 82. Dustin, M. L., Bivona, T. G. & Philips, M. R. Membranes as messengers in T cell adhesion signaling. Nature Immunol. 5, 363–372 (2004). 83. Downey, J., Smith, A., Schneider, H., Hogg, N. & Rudd, C. E. TcR/CD3 mediated stop-signal is decoupled in T‑cells from CTLA‑4 deficient mice. Immunol. Lett. 115, 70–72 (2008). 84. Zang, X. & Allison, J. P. The b7 family and cancer therapy: costimulation and coinhibition. Clin. Cancer Res. 13, 5271–5279 (2007). 85. Chikuma, S., Abbas, A. K. & Bluestone, J. A. B7‑independent inhibition of T cells by CTLA‑4. J. Immunol. 175, 177–181 (2005). 86. Dustin, M. L. Stop and go traffic to tune T cell responses. Immunity 21, 305–314 (2004).
Acknowledgements
C.E.R is a Principal Research Fellow of the Wellcome Trust.
DATABASES Entrez Gene: http://www.ncbi.nlm.nih.gov/entrez/query. fcgi?db=gene CBL‑B | CD28 | CD80 | CD86 | CTLA4 | FOXP3 | IDO | LFA1 | PP2A | SHP2 | TGFβ | ZAP70
FURTHER INFORMATION Christopher E. Rudd’s homepage: http://www.path.cam.ac. uk/pages/rudd All links are active in the online pdf
www.nature.com/reviews/immunol © 2008 Nature Publishing Group
Erratum
Creating immune privilege: active local suppression that benefits friends, but protects foes Andrew L. Mellor and David H. Munn Nature Reviews Immunology 8, 74–80 (2008)
The authors declare competing financial interests. See web version for details.
© 2008 Nature Publishing Group
