Nature Reviews Immunology 11, 155 (March 2011)

from the editors

E

xcessive inflammatory responses can be horribly destructive. Unsurprisingly, therefore, immune responses are dampened down and fine-tuned in a multitude of ways to minimize their damaging effects. Four articles in this issue provide examples of such

immune regulation on four different levels. Wolfgang Junger (page 201) explains that cells take it upon themselves

to control their behaviour to extracellular cues through autocrine feedback loops involving purinergic receptors. ATP or adenosine released by immune cells following stimulation feeds back on the cells through activating or suppressive purinergic receptors that fine-tune cell functions and activation. Luke O’Neill and colleagues (page 163) suggest that microRNAs acting on

▶ Cover: ‘Fine‑tuning’ by Simon Bradbrook,

inspired by the Reviews on pages 163 and 201.

mRNAs may be as important as transcription factors in controlling the protein content of a cell and therefore in regulating cell responses. They focus on the roles of microRNAs in the regulation of Toll-like receptor signalling and during the switch from a pro-inflammatory response to the resolution of inflammation. A fine balance between protective immunity and inflammatory pathology is also achieved through synergistic and antagonistic signalling crosstalk

LuCy bird

between innate immune receptors. The Review on page 187 illustrates

kirsty Minton

how pathogens can undermine this fine-tuning for their benefit through various mechanisms; for example, by co-opting inhibitory receptors, inducing immunosuppressive mediators, promoting safe pathogen uptake and oLive LeAvy

disrupting cooperative receptor interactions involved in host protection.

yvonne bordon

Finally, in an Essay on page 221, Polly Matzinger and Tirumalai Kamala propose that immune responses are regulated at the level of the tissues in which an insult occurs. They suggest that tissue cells are responsible for tailoring the effector class of an immune response to suit the tissue, thereby protecting it from the destructive consequences of inflammatory responses.

MAriA pApAtriAntAfyLLou

Editorial officEs

editoriAL AssistAnts: Laura Corns,

produCtion direCtor:

tokyo [email protected]

Ella Lines

web produCtion MAnAger:

Yvonne Strong

The Macmillan Building, 4 Crinan Street, London N1 9XW, UK Tel: +44 (0)20 7843 3620; Fax: +44 (0)20 7843 3629 Chief editor: Lucy Bird senior editors: Kirsty Minton, Olive Leavy AssoCiAte editors: Yvonne Bordon, Maria Papatriantafyllou Copy editor: Isabel Woodman senior Copy editors: Man Tsuey Tse, Lucie Wootton Copy editing MAnAger: Lewis Packwood Art editor : Simon Bradbrook Art ControLLer: Susanne Harris senior Art editors: Vicky Summersby, Patrick Morgan, Kirsten Lee

Deborah Anthony

Dan Pollock

Chiyoda Building 5F, 2‑37‑1 Ichigayatamachi, Shinjuku‑ku, Tokyo 162–0843, Japan Tel: +81 3 3267 8751; Fax: +81 3 3267 8746 AsiA–pACifiC pubLisher: Antoine E. Bocquet MAnAger: Koichi Nakamura

London [email protected]

MAnAging produCtion editor: Judith Shadwell

senior produCtion editor: Simon Fenwick produCtion ControLLer: Natalie Smith

MArketing MAnAgers: Tim Redding, Leah Rodriguez

direCtor, web pubLishing: heAd of web produCtion: Alexander Thurrell

ManagEMEnt officEs London [email protected]

The Macmillan Building, 4 Crinan Street, London N1 9XW, UK Tel: +44 (0)20 7833 4000; Fax: +44 (0)20 7843 4596/7 offiCe MAnAger: Laura Firman pubLisher: James Butcher

MAnAging direCtor:

new york [email protected]

AsiA–pACifiC sALes direCtor:

Nature Publishing Group, 75 Varick Street, 9th floor, New York, NY 10013–1917, USA Tel: +1 212 726 9200; Fax: +1 212 696 9006

Kate Yoneyama

Howard Ratner

New Delhi 110 002, India Tel/Fax: +91 11 2324 4186

Chief teChnoLogy offiCer:

heAd of internAL systeMs deveLopMent:

senior MArketing MAnAger: Peter Yoshihara

MArketing/produCtion MAnAger: Takesh Murakami

indiA 5A/12 Ansari Road, Daryganj, sALes And MArketing MAnAger, indiA:

Steven Inchcoombe

Anthony Barrera

heAd of softwAre serviCes:

Harpal Singh Gill

Philip Campbell

Luigi Squillante

Copyright © 2011 Nature Publishing Group Research Highlight images courtesy of Getty Images unless otherwise credited. Printed in Wales by Cambrian Printers on acid‑free paper

editor-in-Chief, nAture pubLiCAtions: AssoCiAte direCtors: Jenny Henderson, Dominic Pettit

editoriAL produCtion direCtor: James McQuat

gLobAL heAd of Advertising And CoMMerCiAL direCtor: Dean Sanderson heAd of nAture reseArCh & reviews MArketing: Sara Girard

naturE rEviEws | Immunology

volumE 11 | march 2011 | 155 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights Nature Reviews Immunology | AOP, published online 18 February 2011; doi:10.1038/nri2948

muCOSAL ImmuNOLOGy

Acid attack Immune responses in the intestine must be tightly controlled to prevent harmful attacks against commensal organisms, food proteins and self antigens. Previous studies have shown that dendritic cells (DCs) contribute to immune homeostasis in the intestine by constitutively generating high levels of retinoic acid, which, together with transforming growth factor-β (TGFβ), promotes the development of regulatory T (TReg) cells. However, a recent paper in Nature exposes another side to retinoic acid, showing that this vitamin A metabolite can team up with interleukin-15 (IL-15) to drive pro-inflammatory responses against food antigens. Patients with coeliac disease (who develop intestinal pathology owing to aberrant immune responses to dietary gluten) have increased levels of IL-15 in the gut mucosa, so DePaolo et al. explored whether this cytokine could affect the ability of intestinal DCs to promote TReg cell conversion. Compared with untreated controls, IL-15-treated intestinal DCs were less efficient in driving TReg cell differentiation in vitro and, unexpectedly, the addition of retinoic acid to the IL-15-treated DCs further limited their capacity to induce TReg cells. The authors extended these findings in vivo by using IL-15-transgenic mice that overexpress IL-15 in the intestinal lamina propria and

mesenteric lymph nodes (MLNs). Feeding wild-type mice ovalbumin (OVA) led to the appearance of OVA-specific TReg cells in MLNs, but significantly fewer OVA-specific TReg cells were detected in the MLNs of OVA-fed IL-15-transgenic mice. Moreover, TReg cell induction was further inhibited if IL-15-transgenic mice were fed retinoic acid together with OVA. Using splenic DCs, which do not constitutively produce retinoic acid, the authors found that retinoic acid synergized with IL-15 to promote DC-mediated production of the proinflammatory cytokines IL-12p70 and IL-23. In keeping with this, intestinal lamina propria cells from OVAfed IL-15-transgenic mice, but not from wild-type controls, produced interferon-γ (IFNγ) in response to stimulation with OVA in vitro. This antigen-specific production of IFNγ was enhanced if the IL-15-trangenic mice were fed retinoic acid along with OVA. In a series of further experiments, the authors showed that retinoic acid could synergize with IL-15 and either IL-12p70 or IL-6 to promote the development of a T helper 1 (TH1)- or TH17-type response, respectively. Exploration of the signalling mechanisms involved showed that retinoic acid and IL-15 synergized

NATURE REVIEwS | Immunology

to promote pro-inflammatory DC responses in a JUN N-terminal kinase (JNK)-dependent manner. Finally, DePaolo et al. addressed a question that has long puzzled immunologists. Although approximately 40% of the population express HLA-DQ2 and HLA-DQ8 (the MHC molecules that are associated with coeliac disease), only 2% of these individuals develop this disease. So what promotes the breakdown of tolerance to dietary gluten in this minority? The authors crossed IL-15-transgenic mice with HLA-DQ8-transgenic mice and found that the double-transgenic mice developed several hallmarks of early-stage coeliac disease in response to gluten feeding. These features included the development of gluten-specific antibodies and gluten-specific TH1 cells. Notably, the gluten-specific TH1 cell response was enhanced by retinoic acid, suggesting that in the presence of IL-15, retinoic acid can contribute to the breakdown of tolerance to food antigens. This study suggests that retinoic acid is not simply an agent of tolerance, but synergizes with other cytokines present in the intestinal environment to potentiate both pro- and anti-inflammatory DC responses. In addition, the authors propose that in the HLA-DQ8 and IL-15 double-transgenic mice, immunologists may finally have a physiologically relevant mouse model of coeliac disease. Yvonne Bordon

ORIGINAL RESEARCH PAPER DePaolo, R. W. et al. Co-adjuvant effects of retinoic acid and IL-15 induce inflammatory immunity to dietary antigens. Nature 10 Feb 2011 (doi:10.1038/nature09849)

VOLUME 11 | MARCH 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights

MacMIllan South afRIca

Nature Reviews Immunology | AOP, published online 18 February 2011; doi:10.1038/nri2945

tumOuR ImmuNOLOGy

Early exposure is inflammatory Sunburn during childhood has been associated with an increased risk of melanoma development, and in a recent Nature paper, M. Raza Zaidi, Glenn Merlino and colleagues have shown that changes in the immune response in the skin of young mice after exposure to ultraviolet B (UVB) light could help to promote melanoma development and progression. Melanocytes (pigment-producing cells) constitute approximately 1% of skin cells, making their biology and behaviour difficult to study in vivo. To overcome this, the authors generated a transgenic mouse model in which the expression of green fluorescent protein (GFP) in melanocytes could be switched on and off. These mice were then exposed to either UVB or UVA wavelengths at postnatal day 1,

immuno­ histochemical analyses of 27 human melanoma samples indicated that 19 contained substantial numbers of macrophages that expressed IFNγ

and skin samples were examined. UVB but not UVA triggered melanocyte activation (increased proliferation and migration to the epidermis), and this response lasted for at least 10 days. To further understand this, the authors isolated GFP+ melanocytes 1 day and 6 days after exposure to UVB and carried out an expression microarray. Although stress response genes showed increased expression 1 day after UVB exposure as expected, a subset of genes indicative of a response to interferon-γ (IFNγ) showed a delayed response to UVB at 6 days. To verify the involvement of IFNγ, the authors blocked both type I (IFNα and IFNβ) and type II (IFNγ) responses in young mice ahead of UVB exposure, and found that inhibition of the IFNγ response blocked melanocyte activation. So, which cells in the skin produce IFNγ after exposure to UVB? Immunohistochemical analyses of the skin from young mice 6 days after UVB exposure showed that the majority of immune cells in the skin were CD11b+F4/80+GR1– macrophages, and approximately 28% of these cells expressed IFNγ. The microarray data showed that the UVB-exposed melanocytes expressed ligands that bind the CC-chemokine receptors CCR2 and CCR5, and both of these receptors were expressed by the infiltrating macrophages. However, only CCR2-deficient mice showed a defect in the recruitment of F4/80+ macrophages after exposure to UVB. Moreover, the expression of CCL8, a CCR2 ligand, was substantially increased in

NATURE REVIEWS | Immunology

activated melanocytes at a time that corresponded to the arrival of the CCR2-expressing macrophages. So, are CCR2 and CCL8 important to melanoma development? Ectopic expression of CCL8 in mouse UV-induced melanoma cells promoted macrophage recruitment, and activated macrophages increased the growth rate of UV-melanoma cells in syngeneic mice, primarily by inhibiting apoptosis of the melanoma cells. Mice treated with IFNγ-specific antibodies showed reduced melanoma growth and, importantly, immunohistochemical analyses of 27 human melanoma samples indicated that 19 contained substantial numbers of macrophages that expressed IFNγ. On the basis of these findings, and the fact that the activated melanocytes showed an increased expression of genes that are involved in evasion of the immune response, the authors propose that sunburn in young skin enables melanocytes to evade immune-mediated elimination. Thus, these cells may survive in a state of equilibrium with the immune system prior to the development of a clinically apparent melanoma (an example of the immunoediting paradigm of tumour development). Whether patients with malignant melanoma might benefit from anti-IFNγ therapies needs to be addressed.

Nicola McCarthy Chief Editor, Nature Reviews Cancer

ORIGINAL RESEARCH PAPER Zaidi, M. R. et al. Interferon-γ links ultraviolet radiation to melanomagenesis in mice. Nature 469, 548–553 (2011)

VOLUME 11 | MARCH 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights

in brief R E G U L AT O RY T C E L L S

Killer cell Ig-like receptor (KIR) 3DL1 downregulation enhances inhibition of type 1 diabetes by autoantigen-specific regulatory T cells Qin, H. et al. Proc. Natl Acad. Sci. USA 108, 2016–2021 (2011)

Killer cell immunoglobulin-like receptors (KIRs) are expressed by natural killer (NK) cells. This study identified the inhibitory receptor KIR3DL1 on a subset of autoantigen-specific CD4+FOXP3– regulatory T cells (referred to as 2D2 TReg cells) in non-obese diabetic (NOD) mice, a mouse model of type 1 diabetes. Knockdown of KIR3DL1 expression resulted in increased interleukin-10 secretion by the 2D2 TReg cells, thereby enhancing their suppressive function. Moreover, adoptive transfer of KIR3DL1-deficient 2D2 TReg cells into NOD mice delayed or prevented the onset of type 1 diabetes. Thus, the authors suggest that KIR expression is not restricted to NK cells and that KIRs may negatively regulate the function of TReg cells in autoimmunity. T CELLS

CD4+ T cell help and innate-derived IL-27 induce Blimp-1-dependent IL-10 production by antiviral CTLs Sun, J. et al. Nature Immunol. 6 Feb 2011 (doi:10.1038/ni.1996)

Recent studies indicate that production of the anti-inflammatory cytokine interleukin-10 (IL-10) by cytotoxic T lymphocytes (CTLs) is important for counter-regulating their pro-inflammatory activity during infection. This paper now identifies the signals that drive IL-10 production by CTLs. Experiments with gene-deficient mice revealed that IL-27 (but not IL-12, IL-21, IL-35 or interferon-γ) was required for the induction of IL-10-producing CTLs following influenza virus infection. Also necessary were CD4+ T cells, which ‘helped’ CTL-mediated IL-10 production through local IL-2 production. In vivo experiments confirmed the requirement for IL-2 receptor signalling in IL-10-producing lung CTLs, and also found that innate cells were the source of IL-27. Finally, the findings that CTLs lacking the transcriptional regulator B lymphocyte-induced maturation protein 1 (BLIMP1; also known as PRDM1) failed to produce IL-10 in the presence of IL-2 and IL-27 and that BLIMP1-deficient mice showed enhanced pulmonary inflammation during infection support an important physiological role for BLIMP1-dependent IL-10 production by antiviral CTLs. i N N AT E i M M U N i T Y

Erythropoietin contrastingly affects bacterial infection and experimental colitis by inhibiting nuclear factor-κBinducible immune pathways Nairz, M. et al. Immunity 34, 61–74 (2011)

Erythropoietin (EPO) regulates erythrocyte production. However, EPO receptors (EPORs) are also expressed on non-erythroid cells, including immune cells. This study identifies an anti-inflammatory effect of EPO-induced signals on macrophages. Nairz et al. report that EPO suppresses the production of pro-inflammatory mediators by dampening nitric oxide synthase 2 (Nos2) transcription in activated macrophages. Mechanistically, Janus kinase 2 (JAK2) activation in response to EPO treatment inhibits the phosphorylation of NF-κB inhibitor-α (IκBα), thereby reducing nuclear factor-κB (NF-κB) transcriptional activity. This anti-inflammatory effect of EPO administration leads to impaired clearance of Salmonella Typhimurium by macrophages. Similarly, EPO inhibits pro-inflammatory gene expression in lamina propria myeloid cells, thereby reducing the severity of experimental colitis. Thus, the authors suggest that EPO and EPO derivatives could be used in the treatment of pathological inflammation.

nature reviews | Immunology

volume 11 | march 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights

M AC R O P H AG E S

Iron macrophages Macrophages contribute to wound healing through the processes of inflammation, matrix deposition and tissue remodelling. However, local accumulation of activated macrophages and macrophagederived pro-inflammatory factors can also lead to chronic inflammation, which impairs tissue repair. Recent research sheds light on the local factors that influence the contribution of macrophages to wound healing or chronic inflammation. Sindrilaru et al. first studied the phenotype of human macrophages from chronic venous leg ulcers (CVUs) and investigated the involvement of microenvironmental factors in macrophage-mediated pathogenesis. CVUs are known to contain high levels of iron and, interestingly, the authors found that macrophages in CVUs express high levels of the haemoglobin scavenger receptor CD163, which mediates the endocytosis of haemoglobin– haptoglobin complexes, leading to the accumulation of iron in intracellular stores. In addition, this CD163hi population was shown to

express molecules associated with the M1 macrophage subset, such as tumour necrosis factor (TNF), interleukin-12, CC-chemokine receptor 2 and inducible nitric oxide synthase. The authors hypothesized that the increased iron deposition in CVUs induces an unrestrained pro-inflammatory M1 macrophage population, thereby contributing to chronic inflammation in the lesions. To support this hypothesis they established a mouse model of CVU pathogenesis, in which repeated treatment of mice with iron dextran led to iron accumulation in macrophages. Indeed, iron-loaded mouse macrophages exhibited the same pro-inflammatory M1 phenotype as human CVU macrophages. High TNF production by this macrophage population was shown to be responsible for impaired wound healing, as TNF neutralization disrupted an autocrine feedback loop that maintains macrophages in an activated proinflammatory state and rescued the wound healing response. Similarly,

NATURe ReVIewS | Immunology

specific macrophage depletion dampened inflammation and improved tissue repair. Finally, Sindrilaru et al. clarified the mechanism by which unrestrained pro-inflammatory M1 macrophages cause chronic wounds. They showed that iron-induced reactive oxygen and nitrogen species are produced in CVUs and in the skin of iron-loaded mice in a TNF- and macrophagedependent manner, where they cause oxidative DNA damage and enhance protein nitration. Moreover, the production of reactive oxygen species by iron-loaded macrophages correlates with increased levels of the ageing markers phosphorylated histone H2AX (γH2AX) and INK4A in fibroblasts from CVUs and skin lesions. Therefore, the authors suggest that iron-loaded macrophages induce DNA damage and senescence in skinresident fibroblasts, thereby impairing their capacity for tissue repair. So this study describes how iron present in a chronic inflammatory microenvironment promotes TNF-mediated pro-inflammatory macrophage activation, which results in senescence in skin fibroblasts and restrains wound healing. Based on this research, novel targets for the treatment of CVUs and other chronic inflammatory diseases are revealed. Maria Papatriantafyllou ORIGINAL RESEARCH PAPER Sindrilaru, A. et al. An unrestrained proinflammatory M1 macrophage population induced by iron impairs wound healing in humans and mice. J. Clin. Invest. 7 Feb 2011 (doi:10.1172/JCI44490)

VolUMe 11 | MARCH 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights Nature Reviews Immunology | AOP, published online 18 February 2011; doi:10.1038/nri2949

ImmuNE RESPONSES

IL‑7 goes antiviral It is well appreciated that inter‑ leukin‑7 (IL‑7) has essential roles in lymphocyte development and homeo‑ stasis. A new study in Cell shows that IL‑7 also has potent pro‑immune functions during viral infection. Many chronic viral infections, such as those established in individu‑ als infected with HIV or hepatitis C virus (HCV), are characterized by an upregulation of inhibitory immune pathways. Pellegrini et al. hypothesized that IL‑7, which overcomes inhibitory mechanisms to promote homeostatic immune cell expansion, might also have the ability to overcome the inhibitory networks that develop during chronic viral infection. To explore this, they thera‑ peutically administered IL‑7 to mice with established chronic lymphocytic choriomeningitis virus (LCMV) clone 13 infection. Although control animals that received PBS were unable to clear LCMV clone 13 even several months after the initial infec‑ tion, mice that received IL‑7 therapy cleared LCMV clone 13 from the spleen, liver and other chronic viral

reservoirs shortly after the comple‑ tion of 3 weeks of treatment. IL‑7 seemed to promote antiviral immunity by increasing the numbers of LCMV‑specific CD8+ T cells, as well as by enhancing the effector func‑ tions of these cells. Treatment with IL‑7 also led to a marked expansion of B cells and non‑LCMV‑specific T cells, but cell depletion experi‑ ments showed that only CD4+ and CD8+ T cells were essential for the promotion of viral clearance in mice in response to IL‑7 therapy. Using transgenic reporter mice, the authors showed that IL‑7 expanded non‑ virus‑specific naive T cell populations during LCMV infection by promoting increased export of T cells from the thymus. Interestingly, although the total number of regulatory T (TReg) cells also increased in response to IL‑7 treatment, the proportion of TReg cells within the total T cell population was reduced in IL‑7‑treated mice. However, conditional depletion of TReg cells did not affect viral loads or lead to greater immunopathology in either of the treatment groups.

NATURe ReVIewS | Immunology

The authors next investigated cytokine responses and found that the levels of several inflam‑ matory cytokines, including IL‑6, IL‑17 and interferon‑γ, were dramatically increased in the serum of IL‑7‑treated mice. Notably, this elevated pro‑inflammatory cytokine response appeared to promote antiviral immunity without induc‑ ing increased immunopathology in infected tissues. This seemed to be due to IL‑7‑mediated upregula‑ tion of the cytoprotective cytokine IL‑22; indeed, antibody‑mediated blockade of IL‑22 in LCMV‑infected IL‑7‑treated mice caused significant hepatitis in these animals. Finally, having observed such a marked increase in inflammatory cytokine production, the authors asked whether IL‑7 might repress inhibitors of cytokine signalling. In support of this, they showed that T cells aberrantly upregulate suppres‑ sor of cytokine signalling 3 (SOCS3) during chronic LCMV infection, but IL‑7 treatment led to reduced expression of SOCS3 by T cells. Further experiments suggested that IL‑7 may repress SOCS3 indirectly by downregulating forkhead box O (FOXO) transcription factors and, strikingly, when mice with a T cell‑specific deletion of Socs3 were infected with LCMV, they were able to clear infection without any notable immunopathology, in a similar manner to the IL‑7‑treated mice. These findings indicate that IL‑7 promotes antiviral immunity by suppressing T cell expression of inhibitory molecules, such as SOCS3. The efficacy of therapeutic IL‑7 in this study is particularly exciting, as it suggests that IL‑7 therapy could be used for managing HIV or HCV infections in humans.

Yvonne Bordon

ORIGINAL RESEARCH PAPER Pellegrini, M. et al. IL‑7 engages multiple mechanisms to overcome chronic viral infection and limit organ pathology. Cell 3 Feb 2011 (doi:10.1016/j.cell.2011.01.011)

VOLUMe 11 | MARCH 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights

In the news KEEPING HIV OUT HIV infection paralyses the host’s immune system, resulting in a fatal susceptibility to infections. According to a recent study, it is not only HIV-positive individuals who are unprotected against infections. Christine E. Jones et al. report that uninfected infants who were prenatally exposed to HIV display lower levels of antibodies against infectious agents, such as Haemophilus influenzae type b, than unexposed infants (JAMA, 9 Feb 2011). The low antibody levels were attributed to the defective maternal immune system, as Jones reports that “once the HIV-exposed, uninfected babies received their routine vaccinations, they had antibody levels similar to, or higher than, HIV unexposed infants” (Bloomberg, 8 Feb 2011). Therefore, the researchers suggest that it is now imperative “to establish whether babies exposed prenatally to HIV could be better protected against infections through earlier vaccination, or through vaccine shots given to mothers” (Bloomberg, 8 Feb 2011). Also this month, our progress in the battle against HIV infection is reinforced by the findings of French scientists. Bomsel et al. report that a vaccine comprised of envelope subunit gp41 antigens contained in virosomes blocks the entry of the virus at mucosal sites before primary infection takes place (Immunity, 10 Feb 2011). “Our results clearly challenge the paradigm that mucosal protection requires significantly high levels of antibodies with virus neutralizing capacity in the blood”, concludes Bomsel (ScienceDaily, 10 Feb 2011). The vaccine induced IgA and IgG production in the mucosa, and this could stop viral entry. Interestingly, the authors suggest that “these findings may help to explain why a small population of highly exposed, but HIV-negative, women who exhibit gp41-specific IgA in their vaginal secretions are protected from infection” (ScienceDaily, 10 Feb 2011). Maria Papatriantafyllou

nature reviews | Immunology

volume 11 | march 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights

in brief E VO L U T I O N

A thymus candidate in lampreys Bajoghli, B. et al. Nature 470, 90–94 (2011)

The adaptive immune system of lampreys consists of lymphocyte-like cells that express somatically diversified antigen receptors, termed variable lymphocyte receptors (VLRs). Distinct forms of these receptors are expressed by different lymphocyte lineages: T-like lymphocytes express VLRA genes and B-like lymphocytes express VLRB genes. Boehm and colleagues now show that lampreys also contain thymus-like structures in which lymphocytes undergo VLRA assembly and expression. These structures, termed thymoids, were found in the gill basket of lamprey larvae and were shown to express forkhead box N1 (FOXN1) and delta-like B (DLL-B), which is reminiscent of thymic epithelial cells in jawed vertebrates. Lymphocytes expressing cytosine deaminase 1 (CDA1; which encodes the T-like lymphocyte-specific mediator of gene conversion in lampreys) were found in close proximity to DLL-B+ epithelial cells. Furthermore, non-functional VLRA gene rearrangement was found to only occur in the thymoids, suggesting that the thymoids are the sites of T-like lymphocyte development in lampreys. B CELLS

IL‑33 activates B1 cells and exacerbates contact sensitivity Komai-Koma, M. et al. J. Immunol. 186, 2584–2591 (2011)

Interleukin-33 (IL-33) is a new member of the IL-1 cytokine family that is produced by various tissue stromal cells. Here, the authors provide evidence for a role for IL-33 in the activation of B1 cells. First, the IL-33 receptor subunit ST2 (also known as IL-1RL1) was found to be expressed by activated B1 but not B2 cells. Second, intraperitoneal injection of IL-33 led to a marked increase in B1 but not B2 cells, together with higher levels of the B1 cell products IgM, IL-5 and IL-13. This effect was shown to depend on ST2 expression, as it was abolished in ST2-deficient mice. IL-33-induced B1 cell proliferation also required IL-5, which was probably produced by T cells or mast cells in response to IL-33. Last, physiological relevance for the effect of IL-33 on B1 cells was supported by the finding that ST2-deficient mice developed less severe hapten-induced contact sensitivity (which is known to involve B1 cell-derived IgM), and IL-33 treatment exacerbated the sensitivity response in wild-type mice in a B1 cell-dependent manner. R E G U L AT O RY T C E L L S

Aire‑dependent production of XCL1 mediates medullary accumulation of thymic dendritic cells and contributes to regulatory T cell development Lei, Y. et al. J. Exp. Med. 7 Feb 2011 (doi:10.1084/jem.20102327)

In this study, the authors assessed the factors involved in the accumulation of dendritic cells (DCs) in the thymic medulla and their contribution to self-tolerance. By screening for chemokine receptor expression by isolated thymic DCs and analysing various chemokine-deficient mice, the authors identified chemokine XC receptor 1 (XCR1; expressed by thymic DCs) and its ligand XCL1 (expressed by medullary thymic epithelial cells) as having important roles in the accumulation of thymic DCs. XCL1-deficient mice had impaired development of natural regulatory T cells, and thymocytes from these mice promoted inflammatory lesions in the lacrimal glands of nude mice (which lack mature T cells), suggesting a breakdown in self-tolerance in XCL1-deficient mice. Finally, thymic DC accumulation and the generation of natural regulatory T cells were diminished in autoimmune regulator (AIRE)-deficient mice owing to defective XCL1 expression.

nature reviews | Immunology

volume 11 | march 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights Nature Reviews Immunology | AOP, published online 11 February 2011; doi:10.1038/nri2942

t CELLS

Plastic TH17 cells Interleukin‑17 (IL‑17)‑producing T helper cells (TH17 cells) are a unique subset of effector CD4+ T cells. However, TH17 cells that are generated in vitro have remarkable plasticity and can produce the TH1 cell‑associated cytokine interferon‑γ (IFNγ) and/ or may lose the ability to produce IL‑17. It remains unclear whether TH17 cell plasticity also exists in vivo. Using mice in which the fate of cells that have expressed IL‑17A can be permanently traced, Stockinger and colleagues show that TH17 cells are plastic in vivo and that the functional fate of these cells is shaped by the inflammatory environment. In the fate‑reporter mice gener‑ ated for this study, cells that express IL‑17A are indelibly marked with enhanced yellow fluorescence protein (eYFP). Following the induc‑ tion of experimental autoimmune encephalomyelitis (EAE; a chronic inflammatory disease of the central nervous system) in these mice, the authors found that the numbers of eYFP+ TH17 cells in the

draining lymph nodes and spinal cord increased with disease progression, as expected. However, about half of these cells had stopped producing IL‑17A (referred to as ex‑TH17 cells). Conversely, the expression of IFNγ by eYFP+CD4+ T cells increased over time. In fact, almost all of the IFNγ‑ producing CD4+ T cells in the spinal cord were ex‑TH17 cells. So, TH17 cells give rise to both IL‑17+IFNγ– and IL‑17–IFNγ+ T cells during EAE. Further analysis revealed that eYFP+ T cells progressed from being mainly IL‑17+IFNγ– to being IL‑17+IFNγ+ and finally to express‑ ing IFNγ alone during the course of EAE. Moreover, only eYFP+ (and not eYFP–) T cells produced additional pro‑inflammatory cytokines (such as GM‑CSF (granulocyte–macrophage colony stimulating factor) and TNF (tumour necrosis factor)). Of note, IFNγ‑producing ex‑TH17 cells could be distinguished from eYFP–IFNγ+ TH1 cells by the expression of aryl hydrocarbon receptor and IL‑1 receptor type 1. IL‑23 is thought to contribute to TH17 cell plasticity in vitro, and the authors found that eYFP+ T cells isolated from IL‑23p19‑deficient fate‑reporter mice with EAE lacked IFNγ expression. Furthermore, loss of IL‑23p19 prevented the upregula‑ tion of T‑bet (a transcription factor that is essential for IFNγ produc‑ tion) by eYFP+ T cells. So, IL‑23 is required for the switch in cytokine production by TH17 cells from IL‑17A to IFNγ during chronic inflammation.

But does this switch occur during acute inflammatory responses that resolve quickly? Cutaneous infection of fate‑reporter mice with Candida albicans hyphae, which leads to an acute infection that is rapidly cleared, resulted in the accumula‑ tion of IL‑17‑producing TH17 cells (after an early IL‑17A response mediated by γδ T cells). IFNγ was also produced during this acute infection, but in contrast to the chronic inflammatory setting, none of the IFNγ‑expressing CD4+ T cells were eYFP+. Furthermore, by the time the pathogen was cleared, most eYFP+ T cells had stopped produc‑ ing IL‑17A and the absolute number of T cells in the skin was greatly reduced. Analysis of local antigen‑ presenting cells showed that they expressed high levels of the mRNA encoding the anti‑inflammatory cytokine IL‑10 and low levels of Il23 mRNA. So, the data show that during chronic inflammation TH17 cells deviate to express additional pro‑ inflammatory cytokines, whereas during an acute resolving inflam‑ matory response TH17 cells shut off IL‑17A production and rapidly disappear. In addition, IL‑23 seems to have an important role in driving the plasticity of TH17 cells during chronic inflammation.

Olive Leavy

ORIGINAL RESEARCH PAPER Hirota, K. et al. Fate mapping of IL‑17‑producing T cells in inflammatory responses. Nature Immunol. 30 Jan 2011 (doi:10.1038/ni.1993)

STUDIO 8

NATURE REvIEwS | Immunology

vOLUME 11 | MARCH 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReSeaRch highlightS

B CELLS

Short- and long-term memory

JOH N FO XX IMAGES

In a recent study in Science, Marc Jenkins and colleagues identify two populations of memory B cells that differ in terms of frequency, lifespan and activation potential. In a classical primary immune response, naive IgM+ B cells specific for the target antigen enter a germinal centre, where they undergo affinity maturation and immunoglobulin class switching. The cells then emerge as immunoglobulin-secreting plasmablasts or as surface-switchedimmunoglobulin (swIg)+ memory B cells, which generate plasmablasts after subsequent exposure to antigen. However, there is also evidence for the existence of IgM+ memory B cells, the derivation and function of which are unclear. To comprehensively analyse all memory B cell populations in normal mice, the authors tracked the fate of phycoerythrin-specific B cells using a new antigenbased enrichment method. Naive C57BL/6 mice had ~20,000 phycoerythrinspecific IgM+CD38+GL7– B cells in the spleen and lymph nodes. After subcutaneous injection with

phycoerythrin plus adjuvant, the number of phycoerythrin-specific B cells increased markedly to a peak of ~106 cells by day 13 and then declined to a stable population of ~150,000 cells. At day 8 after antigen injection, the phycoerythrin-specific B cells could be divided into IgM+ and swIg+ populations, both of which contained plasmablasts with high intracellular immunoglobulin concentrations, CD38–GL7+ germinal centre cells and CD38+GL7– memory cells. The number of IgM+ and swIg+ plasmablasts and germinal centre B cells peaked around day 13 and then declined, but the memory B cells had different dynamics according to their immunoglobulin phenotype. swIg+ memory B cells peaked on day 30 at ~60,000 cells and declined to very low numbers by day 450, whereas IgM+ memory B cells peaked on day 10 at ~120,000 cells and the population then remained stable until at least day 450. Adoptive transfer studies showed that both memory B cell populations in isolation could mount a memory response in recipient mice. However, different results were obtained when both memory B cell populations were present in mice that had been previously primed with phycoerythrin. On day 320 after priming, the mice contained 100,000 IgM+ and 2,000 swIg+ antigen-specific memory B cells. When these mice were rechallenged, the number of swIg+ memory cells increased 150-fold to generate 25,000 swIg+ plasmablasts (but no germinal centre

NATure revIeWs | Immunology

cells), whereas the number of IgM+ cells increased less than twofold. The poor secondary response of IgM+ memory B cells was attributed to the presence during the challenge of phycoerythrin-specific antibody (produced by swIg+ plasmablasts), which was shown to inhibit germinal centre formation by both naive B cells and IgM+ memory B cells. These results illustrate some analogies between how we think about humoral and neurological memories. In neuroscience, short-term memory is the capacity for holding a small amount of information in an active, readily available state for a short period of time. This is akin to the small number of short-lived but readily activated swIg+ memory B cells. By contrast, long-term memories can be retained for much longer, akin to the IgM+ memory B cells. The authors speculate that IgM+ memory B cells do not contribute to the secondary response until the immunoglobulin levels produced by swIg+ memory B cells have declined. As long-term memory is subject to fading, several recalls of memory might be required to prevent forgetting. In a similar manner, IgM+ memory B cells can form germinal centres to repopulate the reservoir of humoral immune memory.

Kirsty Minton

ORIGINAL RESEARCH PAPER Pape, K. A. et al. Different B cell populations mediate early and late memory during an endogenous immune response. Science 10 Feb 2011 (doi:10.1126/ science.1201730)

vOLuMe 11 | MArCH 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights

vA C C I N E S

Early and late protection from TB

The one approved vaccine against tuberculosis (known as the Mycobacterium bovis bacillus Calmette–Guérin (BCG) vaccine) and those currently in clinical trials are designed to protect against initial infection, as they incorporate Mycobacterium tuberculosis antigens that are expressed early in the disease process. However, they do not prevent the establishment of latent persistent infection or the reactivation of clinical disease — a major need both for infected patients and for reducing further transmission. Reporting in Nature Medicine, Aagaard and colleagues show that a vaccine containing antigens that are expressed in the early and late stages of tuberculosis protects against

late-stage infection in both pre- and post-exposure mouse models. The vaccine (termed H56) comprises a fusion protein (Ag85B–ESAT6–Rv2660c) and a cationic adjuvant (CAF01). Ag58B and ESAT6 are well known M. tuberculosis antigens that are secreted early in infection and have previously been shown to provide protective immunity. Rv2660c was identified in this study as one of the factors expressed at constant levels throughout infection. The authors proposed that simultaneous vaccination with these three antigens could produce multistage effects, enabling the immune system to mount a response to both the early and late phases of tuberculosis.

nATuRE REvIEwS | Immunology

In the early stage of infection (<4 weeks after exposure), the mycobacterial loads in the lungs of H56-vaccinated mice were similar to those of mice vaccinated with the BCG vaccine. However, in the later stage of infection (24 weeks after exposure), H56-vaccinated mice had significantly lower numbers of bacilli than BCG-vaccinated mice. Moreover, H56 was an effective booster to BCG, as mice that received the booster had significantly lower bacterial loads 24 weeks after infection than mice vaccinated with BCG alone. Importantly, in two mouse models of latent tuberculosis, H56 provided significant protection against reactivation of the disease. Detailed examination showed that H56 promoted the generation of antigenspecific, polyfunctional CD4+ T cells in the lungs. These T cells, which were shown to express interferon-γ, interleukin-2 and tumour necrosis factor, are thought to be important for the quality and endurance of the immune response. Together, these data show that immunization with an antigen that is expressed during the late stages of tuberculosis (together with earlystage antigens) enhances containment and prevents reactivation of the disease. The Statens Serum Institut, Denmark, have reported that H56 will enter clinical trials in South Africa in March 2011.

Man Tsuey Tse Locum Assistant Editor, Nature Reviews Drug Discovery

ORIGINAL RESEARCH PAPER Aagaard, C. et al. A multistage tuberculosis vaccine that confers efficient protection before and after exposure. Nature Med. 17, 189–194 (2011)

voluME 11 | MARCH 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights Nature Reviews Immunology | AOP, published online 11 February 2011; doi:10.1038/nri2943

I N f L A m m At I O N

ROSy outlook for TRAPS patients A paper published in the Journal of Experimental Medicine has furthered our growing understanding of the role of mitochondria, and the reactive oxygen species (ROS) that they produce, in inflammation. This study shows that mitochondrial ROS enhance pro-inflammatory cytokine production through the regulation of the mitogen-activated protein kinase (MAPK) pathway. ROS are important in host defence against pathogens but have also been implicated in inflammatory diseases, although the exact mechanisms are still unclear. Recently, it has been suggested that ROS activate the NLRP3 (NOD-, LRR- and pyrin domain-containing 3) inflammasome, although whether this is a direct or indirect effect is still under debate. In addition, ROS modulate various other signalling pathways and block the dephosphorylation of MAPKs.

To examine the role of ROS in inflammatory responses, the authors used cells from patients with TNF receptor-associated periodic syndrome (TRAPS). TRAPS is an autoinflammatory disease associated with enhanced innate immune responsiveness and abnormal intracellular trafficking of tumour necrosis factor receptor 1 (TNFR1; also known as TNFRSF1A) owing to missense mutations in TNFR1. Previous studies have shown that the enhanced inflammation in TRAPS is linked to increased activation of the MAPKs p38 and JUN N-terminal kinase (JNK), and that peripheral blood mononuclear cells (PBMCs) from patients with TRAPS overproduce TNF and interleukin-6 (IL-6) in response to low-dose lipopolysaccharide (LPS). Here, the authors found that baseline levels of ROS were increased in monocytes and neutrophils from patients with TRAPS. Neutralization of ROS decreased LPS-induced IL-6 and TNF production by PBMCs from patients with TRAPS but also by PBMCs from healthy donors. This neutralization also reduced the sustained JNK and p38 phosphorylation in cells with TRAPS-associated TNFR1 mutations. These data suggest that ROS have a role in normal

LPS-induced cytokine production and in the enhanced innate immune responsiveness associated with TRAPS through the maintenance of MAPK phosphorylation. Further studies showed that the source of the ROS responsible for these effects in both normal and TNFR1-mutant cells was mitochondria and not NADPH oxidases, and that the effect of ROS on the production of IL-6 and TNF was independent of the NLRP3 inflammasome. Furthermore, PBMCs with TRAPS-associated TNFR1 mutations had increased mitochondrial oxidative phosphorylation and ROS production. Finally, specific elimination of mitochondrial ROS reduced LPS-induced inflammatory cytokine production by both mutant and wildtype cells, confirming the importance of mitochondrial ROS for the production of pro-inflammatory cytokines. Therefore, reducing the levels of mitochondrial ROS may be a potential therapeutic strategy for patients with TRAPS.

Olive Leavy

ORIGINAL RESEARCH PAPER Bulua, A. C. et al. Mitochondrial reactive oxygen species promote production of proinflammatory cytokines and are elevated in TNFR1‑associated periodic syndrome (TRAPS). J. Exp. Med. 31 Jan 2011 (doi:10.1084/ jem.20102049)

TTY GE

NATURe RevIewS | Immunology

vOLUMe 11 | MARCH 2011 © 2011 Macmillan Publishers Limited. All rights reserved

ReseaRch highlights

I N F L A M M AT I O N

An innate signal for CNS disease Multiple sclerosis and the animal model experimental autoimmune encephalomyelitis (EAE) are thought to develop in genetically predisposed individuals after an environmental trigger activates myelin-specific T cells in the central nervous system (CNS). A recent study published in Immunity suggests that peptidoglycan (PGN), a component of bacterial cell walls, activates antigenpresenting cells (APCs) in the CNS to provide this trigger. PGN is recognized by the pattern recognition receptors nucleotidebinding oligomerization domain protein 1 (NOD1), NOD2 and Toll-like receptor 2 (TLR2), and

it initiates a pro-inflammatory signalling cascade downstream of the NOD proteins that is mediated by receptor-interacting protein 2 (RIP2; encoded by Ripk2). To investigate the role of PGN in neuroinflammation, the authors analysed the development of EAE in mice lacking the PGN sensors or the downstream signalling adaptor protein. The incidence of disease after EAE induction was similar in wild-type and gene-deficient (Nod1–/–, Nod2–/–, Tlr2–/– or Ripk2–/–) mice. However, the disease severity was reduced in all of the mutant mice, with Ripk2–/– mice developing the mildest disease.

NATuRE REvIEwS | Immunology

Experiments to determine the cause of this protection from disease progression in the gene-deficient mice revealed that it was not due to a T cell-intrinsic defect, as no differences in the development, proliferation, differentiation or CNS infiltration of myelin-specific T cells were observed between wild-type and mutant mice shortly after disease induction. However, 17 days after disease induction, fewer CD4+ T cells were observed in the CNS of Nod1–/–, Nod2–/– and Ripk2–/– mice than in the CNS of wild-type mice. In parallel with this decrease in T cell numbers, the mutant mice had fewer CD11c+ dendritic cells (DCs) in the CNS, with the most marked reduction occurring in Ripk2–/– mice. Ripk2–/– DCs did not show defective migration but instead seemed to be less activated (as determined by levels of MHC class II expression) than wild-type DCs. Moreover, Ripk2–/– DCs had an impaired ability to drive naive T cell differentiation into T helper 17 cells in vitro. Finally, evidence from bone marrow chimeras and adoptive transfer experiments confirmed that EAE progression depends on RIP2 activation in peripheral APCs, which migrate to the CNS and promote the reactivation and expansion of effector T cells.

Lucy Bird

ORIGINAL RESEARCH PAPER Shaw, P. J. et al. Signaling via the RIP2 adaptor protein in central nervous system-infiltrating dendritic cells promotes inflammation and autoimmunity. Immunity 34, 75–84 (2011)

vOLuME 11 | MARCH 2011 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS MicroRNAs: the fine-tuners of Toll-like receptor signalling Luke A. O’Neill*, Frederick J. Sheedy‡ and Claire E. McCoy§

Abstract | Toll-like receptor (TLR) signalling must be tightly regulated to avoid excessive inflammation and to allow for tissue repair and the return to homeostasis after infection and tissue injury. MicroRNAs (miRNAs) have emerged as important controllers of TLR signalling. Several miRNAs are induced by TLR activation in innate immune cells and these and other miRNAs target the 3ʹ untranslated regions of mRNAs encoding components of the TLR signalling system. miRNAs are also proving to be an important link between the innate and adaptive immune systems, and their dysregulation might have a role in the pathogenesis of inflammatory diseases.

Toll-like receptors (TLRs). A family of pattern recognition receptors that detect conserved microbial components during infection and initiate an inflammatory response. They are commonly expressed by cells of the immune system including macrophages and dendritic cells, as well as other sentinel cells such as epithelial cells. TLRs have also been implicated in the recognition of endogenous danger signals that are present in the body during disease.

*School of Biochemistry and Immunology, Trinity College Dublin, Ireland. ‡ School of Medicine, New York University Langone Medical Center, 550 First Avenue, New York 10016, USA. § Monash Institute of Medical Research, Monash University, Clayton, Victoria 3168, Australia. Correspondence to L.A.O’N. e‑mail: [email protected] doi:10.1038/nri2957 Published online 18 February 2011

It is well established that Toll-like receptors (TLRs) have important roles in detecting pathogens and in initiating inflammatory responses that subsequently prime specific adaptive immune responses during infection1. It has also been recognized that dysregulation of this process is a hallmark of inflammatory and autoimmune diseases2. It is therefore important that TLR signalling pathways are tightly regulated. Although many mechanisms for the negative regulation of TLR signalling have been described3, the induction of an anti-inflammatory response and the processes by which inflammation is resolved remain incompletely understood. MicroRNAs (miRNAs) have recently emerged as important regulators of gene expression. They were formerly thought to mainly repress the translation of target mRNAs, but it has recently been shown that the main function of miRNAs in mammalian systems is to decrease target mRNA levels4. Each highly conserved mammalian miRNA probably targets several hundred distinct mRNAs5, so it is probable that most mRNAs are controlled by miRNAs to some extent. miRNAs could therefore be as important as transcription factors in controlling the protein content of a cell. The expression of miRNAs is highly regulated and they are therefore well placed to function as immunomodulators. It is not surprising that given the dynamic nature of miRNAs, they are involved in regulating the components of TLR signalling and innate immune pathways. The signalling molecules that comprise each TLR signalling pathway are regulated by numerous mechanisms, including physical interactions, conformational changes, phosphorylation, ubiquitylation and proteasomemediated degradation6. A more energy-efficient way to regulate the activity of TLR signalling molecules could

be to destabilize the mRNA molecules that encode them. miRNA-mediated control of the expression of these signalling molecules — through either mRNA decay or translational inhibition — might not be as rapid as control through proteasomal degradation; however, this might be an advantage during infection, as miRNAmediated control of mRNA levels allows for a strong initial immune response that is gradually dampened down. However, it must be noted that for miRNAmediated targeting of the mRNAs encoding TLR signalling molecules to be effective, the signalling protein itself must also be targeted by a complementary method: either, it must be sufficiently unstable such that by the time it has passed one half-life no newly synthesized protein is available to take its place, or it must also be targeted for degradation by separate or complementary signalling mechanisms. The end result of the TLR signalling pathways is the activation of pro-inflammatory transcription factors that enhance the transcription of RNA polymerase IIsensitive genes such as those encoding cytokines, chemokines and antimicrobial enzymes. Because miRNAs are also transcribed by RNA polymerase II7,8, it stands to reason that miRNAs themselves are targets of TLR signalling pathways. In this Review, we examine how miRNAs regulate TLR signalling. We consider those miRNAs that are induced by TLR signalling and speculate that these miRNAs regulate the strength, location and timing of TLR responses. miRNAs might provide a link between innate and adaptive immune signalling pathways and they might also have a role in controlling the switch from a strong early pro-inflammatory response to the resolution phase of the inflammatory process.

NATuRE REvIEWs | Immunology

voLumE 11 | mARcH 2011 | 163 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS MicroRNAs (miRNAs). Small (18–22 nucleotide) RNA molecules that regulate gene expression by binding to the 3ʹ untranslated regions of specific mRNAs. They are derived from larger precursor and primary transcript molecules and are themselves transcriptionally regulated in a manner similar to mRNAs.

Induction of miRNAs by TLR signalling many studies have addressed the hypothesis that TLR signalling can modulate miRNA expression using various profiling techniques. Although a subset of miRNAs has emerged as strong targets of TLR signalling, subtle differences in miRNA expression profiles have been observed depending on the TLR stimulus used, treatment time, technology used and, importantly, the cell type. TABLE 1 summarizes the results of these profiling experiments and lists those miRNAs that have been confirmed to be regulated by TLR signalling in independent studies. multiple miRNAs are induced in innate immune cells, with a consensus emerging that miR-155, miR-146 and miR-21 are particularly ubiquitous. There is also evidence that the expression of certain miRNAs can decrease following TLR activation. similar to other TLR-responsive genes, miRNAs can be classified as early or late response genes: some miRNAs (for example, miR-155) are highly induced 2 hours after treatment, whereas other miRNAs (for example, miR-21) are induced at later times. The expression of most TLR-responsive miRNAs described so far depends on nuclear factor-κB (NF-κB) activity. In all cases of miRNA

induction by TLR activation that have been described so far, the transcription of miRNA primary transcripts is upregulated, although it remains possible that the processing of miRNA precursors could also be upregulated by TLR signalling 9. similar to other TLRresponsive genes, it is also important that the induction of TLR-responsive miRNAs is regulated and mechanisms are now being discovered that negatively regulate miRNA induction by TLR signalling; for example, miR-155 in particular is subject to negative regulation by interleukin-10 (IL-10)10. Less is known about how TLR signalling can decrease miRNA expression — this could be through transcriptional repression or through post-transcriptional mechanisms that destabilize miRNA transcripts, and these areas are being actively investigated. Despite the wealth of information regarding miRNA induction, there has been a tendency in the field of miRNA biology to document changes in miRNA levels without effectively analysing the functional consequences of these changes. New data are beginning to emerge from studies analysing the consequences of changes in miRNA expression for TLR biology that provide insights into the control of TLR signalling by miRNAs.

Table 1 | miRNAs regulated by TLRs miRnA

TlRs

Signalling molecules Cell type

other miRnA inducers

Refs

miR-155*

TLR2, TLR3, TLR4, TLR9

MYD88, TRIF, JNK, AP1, NF-κB, KSRP

BMDMs, THP1 cells, monocytes, macrophages, DCs‡, B cells, TReg cells

Helicobacter pylori, KSHV, EBV, oxidized LDL, TNF, IFNβ

9,10,22,25,27,29, 40,41,50,60,61, 124–127

miR-146

TLR2, TLR3, TLR4, TLR5

MYD88, NF-κB

THP1 cells, macrophages, BMDMs, T cells

EBV, VSV, RIG-I, TNF, IL-1

25,26,27,41,61,62, 74,77,80,82,124,126

miR-132

TLR4, TLR9

ND

THP1 cells, human monocytes and macrophages, BMDMs, splenocytes

KSHV

miR-21

TLR4

MYD88, TRIF, NF-κB

Inflamed lung tissue, RAW264.7 cells, BMDMs, B cells, H69 cholangiocytes

Cryptosporidium parvum, EBV (LMP1)

miR-223

TLR4

ND

Inflamed lung tissue, DCs

ND

27,128

miR-147

TLR2, TLR3, TLR4

MYD88, TRIF, NF-κB, IRF3

BMDMs, RAW264.7 cells, THP1 cells, alveolar macrophages

ND

80

miR-9

TLR2, TLR4, TLR7–TLR8

MYD88, NF-κB

Human monocytes and granulocytes

IL-1

41

miR-125b

TLR4

NF-κB

H69 cholangiocytes, rheumatoid arthritis synovial fibroblasts, LPS-tolerized THP1 cells

Cryptosporidium parvum

let-7e

TLR4

AKT1

Peritoneal macrophages

ND

22

miR-27b

TLR4

NF-κB

Human macrophages

ND

44

Upregulated

9,25,41,45,61

62,76,126, 128–132

32,57,131

Downregulated miR-125b§

TLR4

NF-κB, AKT1

Splenocytes, BMDMs, DCs

ND

let-7i||

TLR4

NF-κB, C/EBPβ

H69 cholangiocytes

Cryptosporidium parvum

23,48,133

22,27,50

miR-98

TLR4

ND

H69 cholangiocytes

Cryptosporidium parvum

134

*Expression inhibited by IL-10 and AKT. ‡High basal level of expression in this cell type. §The differences in miR-125b expression after TLR activation are currently unknown. ||Expression inhibited by the miRNA lin-8. AP1, activator protein 1; BMDM, bone marrow-derived macrophage; C/EBPβ, CCAAT/enhancer-binding protein-β; DC, dendritic cell; EBV, Epstein–Barr virus; IFNβ, interferon-β; IL, interleukin; IRF3, IFN-regulatory factor 3; JNK, JUN N-terminal kinase; KSHV, Kaposi’s sarcoma-associated herpesvirus; KSRP, KH-type splicing regulatory protein; LDL, low-density lipoprotein; LMP1, latent membrane protein 1; LPS, lipopolysaccharide; miRNA, microRNA; MYD88, myeloid differentiation primary-response protein 88; ND, not determined; NF-κB, nuclear factor-κB; RIG-I, retinoic acid-inducible gene I; TLR, Toll-like receptor; TNF, tumour necrosis factor; TRIF, TIR domain-containing adaptor protein inducing IFNβ; TReg cell, regulatory T cell; VSV, vesicular stomatitis virus.

164 | mARcH 2011 | voLumE 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

Nuclear factor-κB (NF-κB). A highly pro-inflammatory transcription factor that is activated by many stimuli, including TLR activation. NF-κB complexes are held inactive in the cytoplasm by inhibitor of NF-κB (IκB) proteins. Degradation and removal of IκB is a common NF-κB-activating process and TLR signalling pathways converge on this mechanism. NF-κB-responsive genes include those encoding cytokines, chemokines and antimicrobial enzymes.

3ʹ untranslated region

(3ʹ UTR). The RNA sequence found 3ʹ (downstream) of the stop codon in the open reading frame of a mRNA before the poly(A) tail sequence. 3ʹ UTR sequences vary in length and nucleotide content. It is now recognized that 3ʹ UTR sequences contain regulatory RNA sequences that determine the translation efficiency and stability of the mRNA, including miRNA target sites.

Foam cells Macrophages that localize at sites of early vascular inflammation and that subsequently ingest oxidized low-density lipoprotein and slowly become overloaded with lipids. Foam cells eventually die and attract more macrophages, further propagating inflammation in blood vessels.

Targeting of TLR signalling pathways by miRNAs Targeting TLR expression. The most obvious point at which to manipulate the TLR signalling pathway is at the level of receptor expression. Because TLR signalling induces a strong pro-inflammatory response, the expression of these receptors is restricted to certain cell lineages, including macrophages, dendritic cells (Dcs) and B and T cells11–14. Furthermore, the expression of particular TLRs is restricted to specific cell types to adapt these cells for particular functions15. In addition, the expression of particular miRNAs in mammals has been shown to be limited to particular cell types, which indicates that miRNAs might have a role in controlling cell differentiation and cell-specific functions16,17. An attractive hypothesis is that the differences in TLR distribution between different immune cell types could be the result of differential miRNA expression. However, so far, there is little evidence that TLRs themselves are directly targeted by miRNAs. Bioinformatic analysis of the 3ʹ untranslated regions (3ʹ UTRs) of human TLR mRNAs using the prediction program Targetscan shows that TLR-encoding genes have very few highly conserved target sites for miRNAs (F.J.s., c.E.m. and L.A.o’N., unpublished observations). These observations might indicate that TLR genes are constitutively expressed and that any regulation occurs at the transcriptional or post-translational level. It is also possible that post-transcriptional control of TLRs might be species specific, as a result of differential selective pressures that have affected the evolution of mammalian immune systems18,19. miRNA-mediated control might add to this diversity. If this is the case, it is entirely possible that TLR genes can be targeted by miRNAs through weaker, non-conserved sites. A recent study 20 that used a refined bioinformatic algorithm to predict active miRNA target sites in the 3ʹ uTRs of TLR and related genes showed that the myeloid-specific miRNA miR-223 (REF. 17) — which has an important role in granulopoiesis17,21 — is a strong candidate for regulating both TLR4 and TLR3 expression; TLR3 has been shown to be expressed at low mRNA levels in granulocytes12 possibly owing to increased levels of miR-223 in these cells. This implies that in resting cells, miRNA activity might regulate the potential of innate immune cells to respond to TLR activators. The mRNA encoding TLR4 is regulated by members of the let-7 miRNA family. In mouse peritoneal macrophages, the induction of let-7e expression decreases cell surface expression of TLR4, the mRNA of which contains a let-7 target site22. Furthermore, transfection of macrophages with antisense miRNA to let-7e leads to an increased lipopolysaccharide (LPs)-induced cytokine response22. The mRNA encoding TLR4 can also be targeted by other isoforms of the let-7 family, such as let-7i. Downregulation of let-7i expression was shown to increase TLR4 expression by human cholangiocytes (biliary epithelial cells) after Cryptosporidium parvum infection or LPs treatment23. The differential regulation of TLR4 expression in these cell types (cholangiocytes compared with peritoneal macrophages) by members

of the let-7 family might be due to the fact that epithelial cells would need to be activated and to enhance their pro-inflammatory properties during infection, whereas macrophages need to constantly survey the environment for bacteria and therefore require constitutive expression of TLR4. The expression of TLR4 by macrophages would need to be turned off at later times after ligand sensing to avoid excessive inflammation during infection. This regulation of TLR expression by miRNAs might also explain the differences observed in the induction patterns of miRNAs by TLRs in different cell types. TLR2 mRNA has been shown to be regulated by miR-105, the expression of which is higher in oral keratinocytes derived from patients who respond weakly to TLR2 activation with low levels of cytokine induction, presumably owing to decreased TLR2 expression24; this indicates that there might be a reciprocal relationship between TLR2 signalling and miR-105 expression. Although these data point to the regulation of certain TLRs by miRNAs, the absence of data in support of other TLRs being targeted by miRNAs underscores the importance of constitutive TLR expression. As discussed below, rather than shutting down the TLR signalling pathway completely by eliminating receptor expression, the trend for miRNA activity seems to be to tone down TLR activity through targeting key intracellular signalling proteins. Targeting TLR signalling proteins. The molecular targets of miR-146 are IL-1R-associated kinase 1 (IRAK1) and TNFR-associated factor 6 (TRAF6)25. These proteins are important components of the myeloid differentiation primary-response protein 88 (mYD88)-dependent pathway for NF-κB activation downstream of TLR2, TLR4, TLR5, TLR7–TLR8 and TLR9, which are the same TLRs that induce expression of miR-146 in the THP1 cell line. It was postulated that miR-146 can negatively regulate the mYD88–NF-κB signalling pathway after bacterial infection25. Recently, IRAK2, a kinase that is required for the persistence of NF-κB activation, has also been shown to be targeted by miR-146 (REF. 26), although the relevance of this observation for TLR signalling remains unclear. There is mounting evidence that miR-155 can negatively regulate TLR signalling pathways by targeting key signalling proteins. Inhibition of miR-155 in Dcs resulted in upregulated expression of components of the p38 mitogen-activated protein kinase (mAPK) pathway 27. TAK1-binding protein 2 (TAB2), a signalling molecule downstream of TRAF6 that activates mAPK kinases, was confirmed as a direct target of miR-155 (REF. 27). mYD88 has also been identified as a target of miR-155 in studies of miR-155 induction by Helicobacter pylori 28. Furthermore, mYD88 is a target of miR-155 in foam cells, which induce the expression of miR-155 when loaded with oxidized low-density lipoprotein29. Another Toll/IL-1R domain-containing adaptor protein, mYD88 adaptor-like protein (mAL; also known as TIRAP), which functions as a bridging adaptor for TLR2and TLR4-mediated mYD88-dependent signalling, has

NATuRE REvIEWs | Immunology

voLumE 11 | mARcH 2011 | 165 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS emerged as a target of miR-145 (REF. 30). It remains to be determined whether the expression of miR-145 is also regulated during TLR2 or TLR4 signalling. However, it is known that mAL undergoes proteasomal degradation following TLR2 and TLR4 stimulation31. Therefore, perhaps an additional level of control of mAL expression exists through miR-145. Table 2 | Verified targets of miRNAs in TLR signalling Target mRnA

miRnA(s)

Refs

Receptors TLR4

miR-223, let-7i, let-7e

TLR3

miR-223

20,22,23 20

TLR2

miR-105

24

Signalling molecules MYD88

miR-155

28,29

MAL

miR-145

30

IRAK1

miR-146

25

IRAK2

miR-146

26

TRAF6

miR-146

25

BTK

miR-348

32

TAB2

miR-155

27

IKKα

miR-223

37

IKKβ

miR-199

38

IKKε

miR-155

39,40

Transcription factors NF-κB1

miR-9

41

FOXP3

miR-155

35

C/EBPβ

miR-155

42,43

PPARγ

miR-27b

44

p300

miR-132

45

IL-6

let-7

48

TNF

miR-16, miR-125b, miR-155, miR-221, miR-579, miR-369-3

Cytokines 50,56,57

IL-10

miR-106, miR-466l

IL-12p35

miR-21

51,58 52

ACHE

miR-132

61

PDCD4

miR-21

SHIP1

miR-155

66,67

SOCS1

miR-155

82

Regulators 62

ACHE, acetylcholinesterase; BTK, Bruton’s tyrosine kinase; C/EBPβ, CCAAT/enhancer-binding protein-β; FOXP3, forkhead box P3; IL, interleukin; IKK, inhibitor of NF-κB kinase; IRAK, IL-1R-associated kinase; MAL, MYD88 adaptor-like protein; miRNA, microRNA; MYD88, myeloid differentiation primary-response protein 88; NF-κB1, nuclear factor-κB1; PDCD4, programmed cell death 4; PPARγ, peroxisome proliferator-activated receptor-γ; SHIP1, Src homology 2 (SH2) domain-containing inositol-5ʹ-phosphatase 1; SOCS1, suppressor of cytokine signalling 1; TAB2, TAK1-binding protein 2; TLR, Toll-like receptor; TNF, tumour necrosis factor; TRAF6, TNFR-associated factor 6.

Finally, Bruton’s tyrosine kinase (BTK) participates in the TLR4, TLR7–TLR8 and TLR9 signalling pathways to NF-κB activation, and Btk mRNA is a target of miR-348 (REF. 32), which is strongly induced by LPs treatment of rheumatoid arthritis synovial fibroblasts. It remains to be determined whether miR-348mediated regulation of Btk mRNA also occurs in macrophages and Dcs. What is interesting with regard to the TLR signalling molecules that are targeted by miRNAs is that relatively few proteins have been confirmed as direct targets of miRNAs (specifically, mYD88, mAL, IRAK1, IRAK2, TRAF6, BTK and TAB2; TABLE 2). However, these proteins are components of several TLR signalling pathways, which indicates that once one TLR is triggered, miRNAmediated targeting of common signalling proteins could silence signalling through multiple TLRs. Because most pathogens can engage several TLRs, miRNAs could help to avoid excessive pro-inflammatory responses after pathogen exposure by shutting down several TLR pathways. It is probable that miRNAs work together with multiple other mechanisms to control the expression of TLR signalling components. The combination of these mechanisms could result in timely and appropriate toning down and termination of the pro-inflammatory response (FIG. 1). Targeting transcription factors. The targeting of transcription factors by miRNAs would have a global impact on TLR-induced gene expression. many studies have highlighted the fact that miRNA-mediated targeting of transcription factors is an important aspect of miRNA function33,34. For example, a miRNA that is induced by a particular signalling pathway can often feedback to inhibit the transcription factor that is required for its induction34. Higher basal levels of miRNAs in certain cell types might function as important epigenetic switches required for the functional maintenance of the cell type35,36. For example, forkhead box P3 (FoXP3), a transcription factor that is required for the maintenance of regulatory T cells, was shown to drive the high level of miR-155 expression found in these cells; miR-155 then feeds back and targets FoXP3 to decrease its expression35. more generally, transcription factors are usually expressed at low levels in cells, which might be the result of strict control by miRNAs. Evidence is emerging that the pro-inflammatory transcription factors activated by TLR signalling are targeted directly by miRNAs. NF-κB activity is mainly controlled by inhibitor of NF-κB kinases (IKKs). IKKα was recently shown to be targeted by a subset of miRNAs including miR-223 (REF. 37) and IKKβ is targeted by miR-199 in human ovarian cancer cells 38. However, an effect of these miRNAs on TLR signalling was not directly examined in these studies. Analysis of both miR-155 and its Kaposi’s sarcoma-associated herpesvirus (KsHv) homologue has identified IKKε as a potential target, which supports the notion that miR-155 can negatively regulate innate immune signalling 39,40. more recently, the TLRresponsive miRNA miR-9 was shown to directly target

166 | mARcH 2011 | voLumE 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS NFKB1 mRNA41. NF-κB1 is cleaved to form the NF-κB p50 subunit, which has an important role in transactivation of the NF-κB p65 subunit. Therefore, the finding that miR-9 can target the NFKB1 gene identifies a key control point for TLR signalling. other transcription factors downstream of TLR signalling have been identified as miRNA targets. The transcriptional co-repressor ccAAT/enhancerbinding protein-β (c/EBPβ) has been identified as a target of miR-155 by the analysis of B cells that constitutively express miR-155 or mice that had received antisense miR-155 (REFS 42,43). The targeting of c/EBPβ by miR-155 was shown in these studies to decrease expression levels of granulocyte colony-stimulating factor (G-csF), and possibly IL-6, in splenocytes. Peroxisome proliferator-activated receptor-γ (PPARγ) has antiinflammatory effects and its expression is decreased after LPs treatment. This was shown to be the result of NF-κB-dependent induction of miR-27b, which directly targets Pparg mRNA; the inhibition of miR-27b resulted in increased PPARγ expression and blunted LPs-induced secretion of tumour necrosis factor (TNF)44. Finally, the transcriptional co-activator p300, which is often associated with cAmP-responsive element-binding protein (cREB) and is required for the induction of antiviral genes, was shown to be a direct target of miR-132 in KsHv-infected lymphatic endothelial cells45. multiple transcription factors can therefore be controlled by miRNAs, providing a direct mechanism to control the transcription of TLR-responsive genes. Targeting cytokine mRNAs. Bioinformatic analysis has identified several miRNA-binding sites in cytokine- and chemokine-encoding mRNAs46,47 (FIG. 2a). of note, IL6 mRNA contains a binding site for let-7 (REF. 48); given the fact that let-7 family members can be negatively regulated by TLR signalling and NF-κB activation48,49, this could potentially contribute to the increased IL-6 expression observed following TLR stimulation, although this has not been examined directly. similarly, the 3ʹ uTR of TNF mRNA contains a binding site for the LPs-downregulated miRNA miR-125b, which indicates a mechanism by which TLR signalling might stabilize TNF expression50. IL10 mRNA contains binding sites for 8 miRNAs in its 3ʹ uTR and overexpression of two of these — miR-106a and miR-106b — resulted in decreased IL-10 protein expression in a human Burkitt’s lymphoma Raji cell line51. However, a role for TLR signalling in the induction of miR-106a and miR-106b expression was not explored in this study. mRNA encoding the IL-12p35 subunit contains a target site for miR-21 (REF. 52), as confirmed by reporter assays in which the 3ʹ uTR of the gene encoding p35 was linked to the luciferase gene, although the extent to which this might contribute to TLR responses remains undetermined. Although evidence for the direct targeting of cytokine mRNAs by miRNAs is limited, it is increasingly apparent that miRNAs can function together with RNA-binding proteins to regulate mRNA expression through the Au-rich elements (AREs) that are found in numerous cytokine-encoding mRNAs (FIG. 2a). For example,

6.4 /#. /;& -NKPMGF WDKSWKVKP

$6+4#- +4#-

+4#- 64#( 6#-

6#$

6#$

+--γ +--α +--β

+κ$ OK40#

p65 p50 NF-κB NF-κB p65 p50

6.4KPFWEGFIGPGU NF-κB p65 p50

2TKOK40#

Figure 1 | miRnAs function together with other mechanisms to control the expression of TlR signalling 0CVWTG4GXKGYU^+OOWPQNQI[ components. The canonical Toll-like receptor 4 (TLR4) signalling pathway uses the adaptor molecules myeloid differentiation primary-response protein 88 (MYD88) and MYD88 adaptor-like protein (MAL) to propagate nuclear factor-κB (NF-κB)-dependent gene transcription, the products of which are required for initiating an inflammatory response. However, it is crucial that mechanisms exist to switch this pathway off to prevent over-amplification of this signal. One such mechanism is mediated by K48-linked ubiquitylation, which targets specific TLR signalling components for degradation by the proteasome. Another mechanism is mediated by TLR-induced microRNAs (miRNAs), several of which are regulated by NF-κB. miRNAs bind to the 3ʹ untranslated region of specific mRNA target sequences to inhibit the de novo synthesis of gene products, such as cytokines and components of the TLR signalling pathways. The TLR signalling molecules that are currently known to be targeted by miRNAs include MYD88, MAL, Bruton’s tyrosine kinase (BTK), IL-1R-associated kinase 1 (IRAK1), IRAK2, TNFR-associated factor 6 (TRAF6), TAK1-binding protein 2 (TAB2), inhibitor of NF-κB (IκB) kinase-α (IKKα) and IKKβ. In this manner, TLR-induced signals can tailor the levels of protein in the cell upon infection. Pri-miRNA, primary miRNA.

TNF and IL10 mRNAs both contain long AREs that are targeted by the RNA-binding protein tristetraprolin (TTP), which has a key role in mRNA destabilization downstream of TLR signalling 53–55. miR-16 cooperates with TTP to mediate TNF destabilization, although this

NATuRE REvIEWs | Immunology

voLumE 11 | mARcH 2011 | 167 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS C

D

NGV

+.4

+.O40# OK4 OK4 OK4 60(O40#

/;&

OK4 OK4D

(:4 #)1

OK4 662

+4#-

64#(

6#$

OK4C OK4

OK4 2&%& +.O40#

OK4N

+κ$

OK4

p65 p50

662

OK4

NF-κB p65 p50

+.FGRGPFGPVIGPGU

Figure 2 | miRnA-mediated control of cytokines induced by TlRs and Il-1 signalling. a |The microRNA (miRNA) let-7 has been shown to directly target and destabilize interleukin-6 (IL6) mRNA. miR-155 might have a role in the stabilization of tumour necrosis factor (TNF) mRNA. miR-369-3 mediates TNF stabilization only under conditions of serum starvation, which was shown to depend on the recruitment of fragile-X mental retardation-related protein 1 (FXR1) and 0CVWTG4GXKGYU^+OOWPQNQI[ argonaute 2 (AGO2). By contrast, miR-221 in association with tristetraprolin (TTP) can accelerate TNF destabilization in lipopolysaccharide (LPS)-tolerized cells. miR-16, miR-125b and miR-579 also have an important role in TNF destabilization, and miR-16 is thought to mediate its effect through association with TTP. miR-21 has been shown to positively regulate IL-10 expression through its targeted repression of the mRNA encoding programmed cell death 4 (PDCD4). miR-446l can stabilize IL10 mRNA by competing with TTP for association with the AU-rich elements (AREs) in IL10. miR-106 has been shown to directly target the 3ʹ untranslated region of IL10, leading to gene repression. b | Overexpression of miR-146a, miR-155 and miR-9 modulates IL-1 receptor (IL-1R) signalling, which might be the result of direct targeting of IL-1R-associated kinase 1 (IRAK1) and TNFR-associated factor 6 (TRAF6) by miR-146a, the targeting of TAK1-binding protein 2 (TAB2) by miR-155 and/or the targeting of the p50 subunit of nuclear factor-κB (NF-κB) by miR-9. IL-6-induced signalling can be modulated through the targeted repression of suppressor of cytokine signalling 1 (SOCS1) by miR-155 (not shown). IκB, inhibitor of NF-κB; MYD88, myeloid differentiation primary-response protein 88.

LPS tolerance A transient state of hyporesponsiveness to subsequent stimulation with lipopolysaccharide (LPS) after TLR activation.

miRNA has not yet been shown to be TLR responsive56. more relevant was a recent study 57 that showed that miR-221, miR-579 and miR-125b are expressed following the induction of a state of LPS tolerance, during which TNF mRNA is degraded. miR-221 was found to associate with TTP and to accelerate TNF mRNA decay, and miR-579 and miR-125b seemed to block TNF translation, possibly through recruitment of the translational inhibitor TIAR. However, it should also be noted that some of these effects could be mediated directly by the upregulation of miR-125b expression, which might destabilize TNF by direct binding 50. conversely, miRNAs can also compete with RNAbinding proteins to protect mRNA from destabilization. For example, miR-466l contains a seed region that is complementary to the canonical ARE ‘AuuuA’ sequence. Transfection of LPs-stimulated RAW264.7 macrophages with miR-466l resulted in the upregulation of Il10 mRNA and protein expression by competing with TTP for binding to the ARE sequence in Il10 mRNA, which protected the mRNA from TTP-mediated degradation58. In addition to the recruitment of specific RNA-binding proteins, mRNA stability can also depend on environmental factors. For example, miR-369-3, which associates directly with the ARE in TNF mRNA by base-pairing, could mediate translational activation of TNF only under

conditions of serum starvation, and this activation depended on recruitment of the RNA-binding proteins fragile-X mental retardation-related protein 1 (FXR1) and argonaute 2 (AGo2)59. By contrast, miR-369-3 could repress TNF when the cells were actively proliferating59. It has not been yet been investigated whether the activation of TLRs could affect the ability of a miRNA to degrade or stabilize mRNA sequences, but this possibility warrants further investigation. Although a direct binding site for miR-155 in the TNF mRNA has not been identified, miR-155 might be required for its stabilization, as miR-155-deficient B cells fail to produce TNF60. Furthermore, a role for miR-155 in TNF mRNA stabilization has been shown in HEK293 cells and miR-155-transgenic mice have increased levels of circulating TNF after LPs injection50. Although an exact mechanism has yet to be elucidated, it is possible that RNA-binding proteins might be involved. It will be interesting to determine whether other TLR-responsive cytokines are regulated by miRNAs. Targeting TLR signalling regulators. In addition to the insights discussed above, studies of the miRNAs that are induced by and control TLR signalling have provided us with a tool to uncover new molecules that have important and unexpected roles in TLR signalling pathways.

168 | mARcH 2011 | voLumE 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

Cholinergic anti-inflammatory pathway This pathway fine-tunes cytokine production during inflammation in a highly regulated and reflexive manner. Interaction of acetylcholine with the α7-nicotinic acetylcholine receptor expressed by macrophages results in the suppression of pro-inflammatory cytokine production. The main component of this pathway is the vagus nerve of the parasympathetic branch of the autonomic nervous system.

Luciferase reporter assay A method to measure the transcriptional response. This assay uses a regulatory sequence from a gene of interest fused to the gene that encodes luciferase to determine the effect of the regulatory sequence on gene expression. It is commonly used to determine promoter sequences and transcription factor-binding sites, but can also be used to determine miRNA targeting through the fusion of a 3ʹ UTR sequence containing miRNA target sites to the luciferase gene.

studies of miR-132 have identified that its target, the gene encoding acetylcholinesterase (AcHE), is a key regulator of TLR signalling and have provided links between TLRs and neuroinflammation. AcHE hydrolyses acetylcholine (an important component of the cholinergic anti-inflammatory pathway that is released by efferent vagus nerve fibres) and acetylcholine can block NF-κB nuclear translocation in macrophages and thus attenuate TLR-induced innate immune responses. so, an increase in miR-132 levels in response to TLR stimulation will result in the repression of AcHE and increased acetylcholine-mediated negative regulation of TLR-induced signals. Indeed, LPs-treated macrophages from mice in which the 3ʹ uTR of Ache mRNA has been deleted and therefore cannot bind miR-132 overproduce IL-6, IL-12 and TNF61. In our studies of the induction of miR-21 by LPs, we identified the tumour suppressor protein programmed cell death 4 (PDcD4) as a key target that is downregulated by miR-21 during TLR4 signalling in macrophages62. PDcD4 functions as an inhibitor of translation by inhibiting eukaryotic translation-initiation factor 4F (EIF4F)63,64, which is required for the initiation of translation at the 5ʹ uTR of mRNA sequences. Therefore, mRNAs with complex 5ʹ uTRs — which are longer, have a higher Gc content and contain the potential for secondary RNA structures — are more sensitive to EIF4F inhibition65 and therefore will also be more sensitive to the levels of PDcD4 in a cell. Interestingly, it has been proposed that mRNAs encoding growth factors and cytokines can be considered as complex mRNAs65 and would therefore be susceptible to PDcD4 activity. We found that the production of IL-10, which is increased at later time points after LPs treatment, depended on the levels of both miR-21 and PDcD4 (REF. 62). Although we have yet to show that PDcD4 inhibits IL10 mRNA translation directly at the 5ʹ uTR, it is clear that miR-21 promotes an anti-inflammatory response by increasing IL-10 production, which highlights a previously unappreciated level of control for IL-10 expression and might explain the differences in cytokine production downstream of many TLRs. src homology 2 (sH2) domain-containing inositol-5ʹphosphatase 1 (sHIP1) has been well characterized as a primary target of miR-155 (REF. 66). many studies have shown that increased miR-155 expression in response to LPs stimulation or pathogen infection in macrophages can lead to decreased sHIP1 expression10,43,66,67. This might be an important mechanism that is required for the propagation of a pro-inflammatory response as it has previously been shown that sHIP1 can function as a negative regulator of TLR-induced responses68–70. We showed that the inhibition of miR-155 expression by IL-10 led to an increase in sHIP1 expression, identifying a new target for IL-10-mediated anti-inflammatory responses10. Furthermore, AKT has been shown to negatively regulate miR-155, which indicates that a negative feedback loop might exist whereby sustained AKT expression can switch off miR-155 expression, allowing sHIP1 levels to increase and subsequently inhibit the AKT signalling pathway 22. We propose that the initial increase in miR-155 expression

in response to TLR4 activation downregulates expression of the negative regulator sHIP1, allowing TLR4 signalling to proceed. Later in the response, IL-10 is induced in response to the increased level of miR-21, which decreases the expression of PDcD4, an inhibitor of IL10 translation. IL-10 then feeds back on the pathway to decrease miR-155 levels, thereby restoring sHIP1 levels and limiting TLR4 signalling (FIG. 3).

Fine-tuning of TLR signalling by miRNAs Although there are accumulating data about the expression, induction and mRNA targeting of miRNAs in TLR signalling, there are currently surprisingly few studies that illustrate the global biological impact of this mRNA targeting by miRNAs and the full extent to which miRNAs actually control innate immune responses (BOX 1). many studies have illustrated the induction of a particular miRNA and used 3ʹ uTR luciferase reporter assays to identify targets of the miRNA within the TLR pathways, but few studies have actually inhibited this targeting process and assessed the effects on the immune response. Therefore, it is difficult to determine whether an individual miRNA has an important pro-inflammatory or anti-inflammatory role in TLR signalling. Influencing inflammatory cytokine signalling. The examination of cytokine expression when TLR-induced miRNAs are overexpressed or inhibited, as well as studies of miRNA-deficient mice, has helped to determine

6.4

OK4

OK4

5*+2

2&%&

↑+.

+.O40#

Figure 3 | Fine-tuning of TlR4 signalling by miR-155 and miR-21. Recent evidence0CVWTG4GXKGYU^+OOWPQNQI[ indicates that the microRNAs (miRNAs) miR-155 and miR-21 are important for the regulation of Toll-like receptor 4 (TLR4) signalling. TLR4 signalling increases miR-155 levels, which leads to the degradation of Src homology 2 (SH2) domain-containing inositol-5ʹ-phosphatase 1 (SHIP1), a negative regulator of TLR4 signalling; thereby, this process increases TLR4 signalling. However, TLR4 activation also increases the level of miR-21, which targets the mRNA encoding programmed cell death 4 (PDCD4), leading to increased production of interleukin-10 (IL-10) (as PDCD4 is an inhibitor of IL10 translation). IL-10 then feeds back on the pathway and specifically inhibits the induction of miR-155. This, in turn, leads to an increase in SHIP1 and the inhibition of TLR4 signalling. This process might be important for the inhibitory effect of IL-10 on TLR signalling and could be relevant for lipopolysaccharide (LPS) tolerance or for the resolution of TLR4-induced inflammatory responses.

NATuRE REvIEWs | Immunology

voLumE 11 | mARcH 2011 | 169 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Box 1 | Identification of miRNA targets The identification of microRNA (miRNA) targets is highly problematic. First, bioinformatic analysis is used to predict target mRNAs through sequence analysis of their 3ʹ untranslated regions (3ʹ UTRs), but many of these mRNAs might not even be expressed under the same conditions as the miRNA of interest. Second, once a correct mRNA target has been identified, the effects of the miRNA can be subtle, at least at the protein level. Twofold changes in mRNA levels are typically observed by microarray analysis after endogenous miRNA knockdown122. Changes in mRNA levels can be even more difficult to observe when you consider the problems of transfection, overexpression and immunodetection. Compounding this, at least for the study of Toll-like receptors (TLRs), can be the strong effects induced by TLR activation, which might be difficult to regulate by a single miRNA. Many studies use 3ʹ UTR luciferase reporter assays to confirm that a particular mRNA is targeted by a particular miRNA. However, true functional targets of miRNAs need to be confirmed through the use of antisense or overexpression assays both in vitro and in vivo. Antisense technology might be more physiologically relevant, as it inhibits the expression of endogenous miRNAs, compared with overexpression studies, which risk off-target effects. Target protection using morpholino-modified oligonucleotides can be used to confirm the specificity of a particular miRNA–mRNA interaction. However, ultimately, it will be the generation of genetically modified mice that are deficient for, or overexpress, candidate miRNAs or that have mutated miRNA target sites in specific genes that will prove most informative. It must also be pointed out that the concept of a ‘miRNome’ has emerged123, whereby the concentration of miRNAs in most immune cells, including dendritic cells and macrophages, was analysed. As only a few miRNAs were enriched in cells and a minimum threshold amount must be reached for miRNAs to repress target mRNAs124, the abundance of miRNAs and their ratios in the entire miRNome of a specific cell or tissue might be extremely important for their functions. Therefore, it is possible that the miRNAs that are most abundantly expressed and those that are highly induced in innate immune cells are of importance. Equally, the expression and abundance of mRNA targets in these cells will have a role in determining the outcome and must be considered.

Morpholino-modified oligonucleotide A nucleic acid analogue in which the base and phosphate linkages structurally differ from regular DNA or RNA. They are commonly 25 nucleotides in length and they function by blocking access of RNA-binding proteins or RNAs to target sites in mRNAs to which they are antisense. They can be used to protect a mRNA from miRNA activity by targeting the morpholino-modified oligonucleotide to a miRNA target site in a specific mRNA.

Langerhans cells Professional antigen-presenting dendritic cells that are localized in the skin epidermis.

whether specific miRNAs are pro-inflammatory or anti-inflammatory. Enforced expression of miR-146a decreases the expression of various pro-inflammatory chemokines and cytokines, such as cXc-chemokine ligand 8 (cXcL8) and cc-chemokine ligand 5 (ccL5) by epithelial cells71, IL-6 and cXcL8 by fibroblasts and TNF by osteoarthritic tissue after IL-1 stimulation72,73. miR-146a has also been shown to decrease the expression of TNF, IL-1β and IL-6 by THP1 monocytes during LPs tolerance74; the expression of IL-2 by activated T cells75; and the production of type I interferons (IFNs) by TLR7-stimulated peripheral blood mononuclear cells and by Epstein–Barr virus- or vesicular stomatitis virusinfected cell lines26,76,77. These effects are probably mediated by the targeted repression of IRAK1 and TRAF6 by miR-146a. These data, together with the observation that miR-146a-deficient mice develop an exaggerated pro-inflammatory response when exposed to LPs (mark Boldin and David Baltimore, personal communication), strongly support a role for miR-146a as a negative regulator of TLR4 signalling in vivo. Interestingly, miR-146a has also recently been shown to have a crucial role in mediating protective tolerance in intestinal epithelial cells of neonatal mice78. Furthermore, Langerhans cells constitutively express high levels of miR-146a compared with monocytes and interstitial Dcs and therefore have decreased sensitivity to TLR activation as a result of IRAK1 and TRAF6 repression by miR-146a; this renders Langerhans cells tolerant to inappropriate activation by commensal bacteria79.

similarly, enforced expression of other TLR-induced miRNAs, such as miR-9, miR-132 and miR-147, has been shown to alter cytokine production profiles, which indicates that these miRNAs also function to negatively regulate pro-inflammatory responses. overexpression of miR-9 in human primary chondrocytes decreased IL-1-induced expression of TNF and matrix metalloprotease 13 (REF. 73) . miR-132 functions to inhibit the TLR-induced production of pro-inflammatory cytokines61 and KsHv-mediated induction of miR-132 in lymphatic endothelial cells impaired expression of the antiviral genes IFNβ and IFN-stimulated gene of 15 kDa (IsG15), as well as IL-1β and IL-6 (REF. 45). This indicates that viruses can induce host miRNAs to their advantage by manipulating the expression of host cytokines. overexpression of miR-147 attenuates TLR2-, TLR3- and TLR4-induced production of TNF and IL-6, as well as of miR-147 itself 80. LPs challenge of miR-223-deficient mice results in exaggerated lung inflammation, which indicates that miR-223 might also target proteins to control and limit excess inflammation21. The role of miR-155 in TLR responses is more complex. on the one hand, inhibition or overexpression of miR-155 has shown that miR-155 negatively regulates the expression of cytokines and chemokines such as IL-1 and cXcL8 (REFS 27,28,40,81). on the other hand, miR-155 is required for the expression of TNF, IL-6, IL-23 and type I IFNs22,43,50,60,66,82–84, an effect that is probably mediated by the targeted suppression of suppressor of cytokine signalling 1 (socs1) and/or sHIP1, two negative regulators of cytokine-mediated and TLR signalling pathways22,82,85. The differences in cytokine regulation by miR-155 require careful consideration. They might simply be due to experimental design (for example, comparing overexpression and gene knockout studies), the cell types used and/or the time of miRNA induction that was analysed. From studies carried out in mouse models and the requirement for miR-155 in priming an adaptive immune response (see below), it seems that miR-155 might promote pro-inflammatory responses rather than being inhibitory. miR-155 might therefore function as a ‘brake’ or a molecular rheostat that represses the overactivation of a pro-inflammatory response without completely suppressing it. Another aspect of TLR-induced miRNAs is the feedforward and regulatory effects they can have on signalling pathways downstream of cytokine receptors (FIG. 2b). However, this has been reviewed elsewhere86 and is not discussed in detail here. Priming the adaptive immune response. Altered cytokine production profiles will obviously affect the ability of the innate immune system to prime adaptive immune responses. However, few studies have directly addressed the effect of miRNAs on this process. Indirect evidence from miRNA-deficient mice indicates that this might be an important feature of the TLR-induced miRNA response, particularly for miR-155; miR-155-deficient mice have global immune defects characterized by decreased Dc function and defective B and T cell responses60,87.

170 | mARcH 2011 | voLumE 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

Class switching The somatic recombination process by which immunoglobulin isotypes are switched from IgM to IgG, IgA or IgE.

Experimental autoimmune encephalomyelitis (EAE). An animal model of human multiple sclerosis. EAE develops in susceptible rodents and primates after immunization with antigens derived from the central nervous system.

Germinal centre A lymphoid structure that arises within B cell follicles after immunization with, or exposure to, a T celldependent antigen. It is specialized for facilitating the development of high-affinity, long-lived plasma cells and memory B cells.

Exosomes Small lipid bilayer vesicles that are released from dendritic cells and other cells. They are composed of cell membranes or are derived from the membranes of intracellular vesicles. They might contain peptide–MHC complexes and directly interact with antigen-specific lymphocytes, or they might be taken up by other antigen-presenting cells.

Resolvin D1 A lipid mediator that is induced in the resolution phase following acute inflammation. Resolvins are synthesized from the essential omega-3 fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA).

specifically, the failure of miR-155-deficient mice to acquire protective immunity after Salmonella typhimurium infection was found to be owing to a failure of Dcs to efficiently activate T cells; however, this observation was not examined further 87. The adaptive immune response in these mice was also found to be skewed towards a T helper 2 (TH2)-type immune response, which indicates that miR-155 is normally required for a TH1-type response and for the polarization of T cells towards a pro-inflammatory phenotype 87. In a later study 88, immunization of miR-155-deficient mice with dinitrophenylated LPs, a T cell-independent antigen, resulted in defective B cell class switching. However, it was not determined whether this effect was the result of a defective innate immune response. Induction of miR-155 expression upon Dc maturation by LPs resulted in the downregulation of expression of Dc-specific intercellular adhesion molecule-3-grabbing non-integrin (Dc-sIGN), which limited pathogen uptake by these cells; the authors postulated that this effect could have an important role in balancing T cell polarization89. Recently, it has been shown that miR-155 is required to promote TH17 cell polarization83. TH17 cells are potent drivers of experimental autoimmune encephalomyelitis (EAE), and mYD88 deficiency has been shown to protect against EAE in mouse models through the decreased induction of IL-17 (REF. 90). Interestingly, deficiency of miR-155 also protected against EAE83. using microarray analysis, miR-155-deficient Dcs were shown to express lower levels of the TH17 cell-inducing cytokines IL-6 and IL-23 than wild-type Dcs; it was postulated that miR-155 has an important role in promoting a pro-inflammatory response mediated by Dcs to prime TH17-type adaptive immune responses. It is therefore crucial that tissue-specific and/or inducible miRNA-deficient mice are generated to fully understand the role of miR-155 and other TLR-inducible miRNAs in determining the type of immune response. The generation of mice with mutations in the 3ʹ uTR of specific mRNA targets will also help to decipher the specific role of miRNAs. For example, mice with a mutated miR-155 target site in the 3ʹ uTR of the mRNA encoding activation-induced cytidine deaminase (AID), the key antibody class switching enzyme in B cells, have allowed investigators to specify the role of miR-155 in B cell responses. In this case, the induction of miR-155 in B cells following stimulation with LPs and IL-4 is crucial for controlling the germinal centre response and therefore the appropriate stimulation of humoral immune responses91,92. It is interesting that IL-10, which is an important regulator of both innate and adaptive immune responses, can inhibit LPs-induced expression of miR-155 (REF. 10). IL-10 is known to dampen the innate immune response by downregulating TLR-induced pro-inflammatory gene expression in macrophages and Dcs after pathogen infection and to inhibit the proliferation of and cytokine production by cD4+ T cells. By contrast, IL-10 seems to have a stimulatory effect on B cells, resulting in enhanced proliferation, differentiation and class-switch recombination93,94. The obvious differences in the effects

of IL-10 on these cells might be partly owing to its inhibitory effect on miR-155 (REF. 10). An increase in the level of the miR-155 target sHIP1 was shown to contribute to some of the anti-inflammatory effects of IL-10 on macrophages10, whereas an increase in the expression of other miR-155 targets might explain some of the respective regulatory and stimulatory effects of IL-10 on B and T cells, an area which we are actively pursuing. However, it remains possible that TLR-inducible miRNAs have a more direct effect on the adaptive immune response that is not solely dependent on cytokine expression. For example, miRNAs have been shown to be contained in exosomes95–98. Exosomes are membrane vesicles released by various cell types, including immune and tumour-derived cells, that are often present in biological fluids99,100. They contain common sets of molecules such as chaperone proteins, cytoskeletal proteins and proteins involved in transport and fusion, as well as cell typespecific molecules that reflect the cell of origin and the target cell. For example, exosomes derived from Dcs can transfer mHc class II molecules to recipient Dcs and neighbouring cells, in addition to priming T cell responses101–103. It is therefore possible that exosomes derived from antigen-presenting cells might contain miRNAs that have an important role in priming the adaptive immune response, an area that is worthy of further investigation. Together, these observations indicate that miRNAs derived from innate immune cells might be important regulators of adaptive immune responses through the modulation of cytokine expression and in a cell autonomous manner through exosomes.

miRNAs and resolution of inflammation It is clear from the above data that the miRNAs induced by TLRs have an important role in regulating the immune response. For example, an initial pro-inflammatory response might be mediated by the early TLR-induced expression of miR-155, which could limit and repress the expression of key negative regulators of TLR signalling, thereby promoting cytokine expression and adaptive immune responses. Furthermore, let-7i and miR-125b are downregulated early in the immune response to allow for the production of the pro-inflammatory mediators TLR4 and TNF, respectively. The induction of miR-146 expression later in the response has an important function as a negative regulator of innate immune responses, feeding back to turn off TLR signalling pathways. Following this, miR-21 is induced to promote the anti-inflammatory response and facilitate the production of IL-10. IL-10 can then feed back to inhibit the expression of miR-155, thereby further contributing to the anti-inflammatory effect. It is thought that both miR-9 and miR-147 contribute to the negative regulation of innate immune responses at later times by targeting NF-κB and affecting cytokine production. These data and the more recent finding that miRNAs, including miR-21 and miR-146b, are induced by the anti-inflammatory lipid mediator resolvin D1 in human macrophages104 imply that miRNAs are key players in the resolution of inflammation by mediating the removal of components of the TLR signalling pathways (FIG. 4).

NATuRE REvIEWs | Immunology

voLumE 11 | mARcH 2011 | 171 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS +OOGFKCVGGCTN[TGURQPUG

r4GRTGUUKQPQHPGICVKXGTGIWNCVQTU

 5*+251%5 r2TQOQVKQPQHE[VQMKPGGZRTGUUKQP r+PKVKCVKQPQHCFCRVKXGKOOWPGTGURQPUG NGVK OK4D

OK4

2TQKPȯCOOCVQT[TGURQPUG 'CTN[TGURQPUG r4GRTGUUKQPQH

6.4UKIPCNNKPI EQORQPGPVU  64#(+4#-

r0GICVKXGTGIWNCVKQP

QHCPVKXKTCNTGURQPUG  VJTQWIJR

OK4C OK4

r4GRTGUUKQP

QH0(κ$R

OK4

0GICVKXGTGIWNCVKQPQHKPȯCOOCVKQP .CVGTGURQPUG r4GRTGUUKQPQH

E[VQMKPGRTQFWEVKQP

OK4

r4GRTGUUKQPQH2&%& r+PFWEVKQPQH+.

OK4

OK4

#PVKKPȯCOOCVQT[TGURQPUG

Figure 4 | The role of miRnAs in the resolution of inflammation. Toll-like receptor (TLR) activation seems to result in the sequential induction of important microRNAs (miRNAs) that can control the strength and longevity of an inflammatory response. TLR signalling 0CVWTG4GXKGYU^+OOWPQNQI[ strongly induces miR-155, which has been proposed to enhance the pro-inflammatory response through the inhibition of negative regulators, such as Src homology 2 (SH2) domain-containing inositol-5ʹ-phosphatase 1 (SHIP1) and suppressor of cytokine signalling 1 (SOCS1), the promotion of cytokine expression and the subsequent induction of an adaptive immune response. The downregulation of let-7i and miR-125b miRNAs (indicated by dashed lines) could contribute to this effect owing to a lack of repression of their respective targets TLR4 and tumour necrosis factor (TNF). Later, the induction of miR-146a and miR-9 resolves the pro-inflammatory response by targeting TNFR-associated factor 6 (TRAF6) and IL-1R-associated kinase 1 (IRAK1), which are key components of TLR signalling pathways, and the nuclear factor-κB (NF-κB) subunit p50, respectively. miR-132 expression has also been shown to be increased by TLR signalling and would limit the antiviral response by targeting p300. At later times, miR-21 is induced and increases IL-10 expression by repressing programmed cell death 4 (PDCD4), which might contribute to the induction of an anti-inflammatory response through the inhibition of miR-155. miR-147 also promotes an anti-inflammatory response by repressing cytokine production.

miRNAs in disease Although miRNAs are required for the induction and the resolution of innate immune responses, it is clear that if they themselves are not controlled, they can contribute to inflammatory disorders and the progression of cancer. most evidence for this has come from profiling studies that investigate whether aberrant expression of particular miRNAs correlates with particular disease states, such as ulcerative colitis105. The aberrant expression of many miRNAs has been found to overlap between inflammation and cancer, in support of the age-old link that has been thought to exist between these diseases. As there are several recent reviews on this topic106–114, we only discuss a few key points.

multiple miRNA profiling studies have been carried out in patients with rheumatoid and osteoarthriits, psoriasis and atopic eczema, for which increased expression of the TLR-induced miRNAs miR-21, miR-132, miR-146a and miR-155 is commonly detected73,81,115–119. By contrast, miRNA profiling of samples from patients with systemic lupus erythematosus has detected a lack of miR-146a120. It is possible that differences in cytokine expression and NF-κB activity might contribute to the varying levels of miR-146 expression found in inflammatory diseases, owing to the inhibition of IRAK1 and TRAF6 by miR-146a. It is possible that the constant low-grade inflammation associated with inflammatory diseases leads to dysregulated miRNA expression, which in turn contributes to disease pathogenesis. In the case of the proinflammatory miRNA miR-155, it is easy to see how increased expression of this miRNA could result in the inappropriate activation of inflammatory pathways. However, for the other miRNAs that are upregulated in these diseases (miR-21, miR-132 and miR-146a), which have strong anti-inflammatory effects, it is more difficult to appreciate how their upregulation might contribute to the excessive inflammation characterized by these disorders. Again, the context-specific mRNA expression profiles in diseased cells must be considered. It might be that different miRNA targets are expressed in these cells or that the respective miRNA targets are no longer expressed. Alternatively, the anti-inflammatory effects of these miRNAs might result in disease because the cells are unable to mount appropriate immune responses. This could be a possible explanation for the development of cancer, as the TLR-inducible miRNAs have also been found to be upregulated in several types of cancer 106. Again, there is mounting evidence in support of TLR activation in cancer 107 and the dysregulation of miRNA expression is likely to have a role in this pathogenic process. Finally, another interesting aspect of the role of miRNAs in disease is how mutation or deletion of the 3ʹ uTR of mRNAs that are normally targeted by miRNAs could contribute to disease. For example, a polymorphism in the 3ʹ uTR of the mRNA encoding IRAK1 (a target of miR-146a) is associated with susceptibility to rheumatoid arthritis 121. similarly, loss of miRNA expression has been shown to contribute to disease. For example, loss of miR-145 and miR-146a transcripts within the 5q locus identified these miRNAs as key mediators of 5q syndrome, a haematopoietic malignancy that progresses to acute myeloid leukaemia30. Loss of non-coding transcripts in the 5q locus identified that the genes encoding both miR-145 and miR-146a were absent in this disease. As a result, the upregulation of mRNA targets for these miRNAs — such as the mRNAs encoding mAL, IRAK1 and TRAF6 — could be contributing to the disease. This finding emphasizes how the dysregulation in haematopoeitic cells of miRNAs that target key innate immune processes can cause leukaemia, emphasizing further the link between innate immunity and cancer.

172 | mARcH 2011 | voLumE 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Together, these studies are of huge therapeutic relevance as they not only provide useful diagnostic information but also help to define which miRNAs might be most relevant for therapeutic intervention in particular disease states.

Future directions The study of miRNAs has opened up many new areas in the regulation of TLR signalling. The targeting of key proteins in the TLR signalling pathways highlights the important role of miRNAs and also illustrates new negative feedback loops that can control the outcome of TLR responses — through quantitatively regulating cytokine production but also, qualitatively, by switching from proinflammatory to anti-inflammatory responses and by the induction of adaptive immune responses in a timely and orchestrated manner. Although many data have been obtained on the induction of miRNAs by TLRs, true 1. 2. 3.

4.

5. 6.

7.

8. 9.

10.

11.

12.

13.

14.

O’Neill, L. A. How Toll-like receptors signal: what we know and what we don’t know. Curr. Opin. Immunol. 18, 3–9 (2006). Cook, D. N., Pisetsky, D. S. & Schwartz, D. A. Toll-like receptors in the pathogenesis of human disease. Nature Immunol. 5, 975–979 (2004). Liew, F. Y., Xu, D., Brint, E. K. & O’Neill, L. A. Negative regulation of Toll-like receptor-mediated immune responses. Nature Rev. Immunol. 5, 446–458 (2005). Guo, H., Ingolia, N. T., Weissman, J. S. & Bartel, D. P. Mammalian microRNAs predominantly act to decrease target mRNA levels. Nature 466, 835–840 (2010). An in-depth study showing that miRNAs function at the post-transcriptional level mainly by decreasing mRNA levels rather than by inhibiting translation. Baek, D. et al. The impact of microRNAs on protein output. Nature 455, 64–71 (2008). Carpenter, S. & O’Neill, L. A. Recent insights into the structure of Toll-like receptors and posttranslational modifications of their associated signalling proteins. Biochem. J. 422, 1–10 (2009). 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). Lee, Y. et al. MicroRNA genes are transcribed by RNA polymerase II. EMBO J. 23, 4051–4060 (2004). Ruggiero, T. et al. LPS induces KH-type splicing regulatory protein-dependent processing of microRNA-155 precursors in macrophages. FASEB J. 23, 2898–2908 (2009). McCoy, C. E. et al. IL-10 inhibits miR-155 induction by Toll-like receptors. J. Biol. Chem. 285, 20492–20498 (2010). This is the first demonstration of modulation of a TLR-induced miRNA (miR-155) by IL-10, an effect that might be important for the anti-inflammatory functions of IL-10. Rock, F. L., Hardiman, G., Timans, J. C., Kastelein, R. A. & Bazan, J. F. A family of human receptors structurally related to Drosophila Toll. Proc. Natl Acad. Sci. USA 95, 588–593 (1998). Muzio, M. et al. Differential expression and regulation of Toll-like receptors (TLR) in human leukocytes: selective expression of TLR3 in dendritic cells. J. Immunol. 164, 5998–6004 (2000). Hornung, V. et al. Quantitative expression of Toll-like receptor 1–10 mRNA in cellular subsets of human peripheral blood mononuclear cells and sensitivity to CpG oligodeoxynucleotides. J. Immunol. 168, 4531–4537 (2002). Zarember, K. A. & Godowski, P. J. Tissue expression of human Toll-like receptors and differential regulation of Toll-like receptor mRNAs in leukocytes in response to microbes, their products and cytokines. J. Immunol. 168, 554–561 (2002).

functional data showing the exact effects of miRNAs on TLR responses are still required. These studies have also revealed novel targets for TLR signalling and have illustrated new processes that are activated downstream of TLRs. Importantly, they have highlighted the importance of post-transcriptional control of the inflammmatory response. The finding that several miRNAs work together with RNA-binding proteins and regulate the translation of TLR-responsive mRNAs in different ways opens up an exciting new avenue in TLR research. These extra levels of previously unappreciated gene regulation provide new targets for the therapeutic manipulation of these key innate immune signalling pathways. Given the modest fine-tuning and cell type-specific effects of TLRresponsive miRNAs, these might provide attractive drug targets. If this proves to be the case, then one must hope that the application of such RNA-based therapies does not get lost in translation.

15. Jarrossay, D., Napolitani, G., Colonna, M., Sallusto, F. & Lanzavecchia, A. Specialization and complementarity in microbial molecule recognition by human myeloid and plasmacytoid dendritic cells. Eur. J. Immunol. 31, 3388–3393 (2001). 16. Lagos-Quintana, M., Rauhut, R., Lendeckel, W. & Tuschl, T. Identification of novel genes coding for small expressed RNAs. Science 294, 853–858 (2001). 17. Chen, C. Z., Li, L., Lodish, H. F. & Bartel, D. P. MicroRNAs modulate hematopoietic lineage differentiation. Science 303, 83–86 (2004). The first study to investigate the specific role of miRNAs in haematopoiesis and lineage differentiation. 18. Mestas, J. & Hughes, C. C. Of mice and not men: differences between mouse and human immunology. J. Immunol. 172, 2731–2738 (2004). 19. Palsson-McDermott, E. M. et al. TAG, a splice variant of the adaptor TRAM, negatively regulates the adaptor MyD88-independent TLR4 pathway. Nature Immunol. 10, 579–586 (2009). 20. Heikham, R. & Shankar, R. Flanking region sequence information to refine microRNA target predictions. J. Biosci. 35, 105–118 (2010). 21. Johnnidis, J. B. et al. Regulation of progenitor cell proliferation and granulocyte function by microRNA-223. Nature 451, 1125–1129 (2008). This study reports an important role for miR-223 in the regulation of granulocytes and macrophages. 22. Androulidaki, A. et al. The kinase Akt1 controls macrophage response to lipopolysaccharide by regulating microRNAs. Immunity 31, 220–231 (2009). 23. 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). 24. Benakanakere, M. R. et al. Modulation of TLR2 protein expression by miR-105 in human oral keratinocytes. J. Biol. Chem. 284, 23107–23115 (2009). 25. 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). The first study to profile the miRNAs that are induced by TLR signalling and to propose that miRNAs function as negative regulators by targeting key proteins in the TLR signalling pathways. 26. Hou, J. et al. MicroRNA-146a feedback inhibits RIG-I-dependent type I IFN production in macrophages by targeting TRAF6, IRAK1, and IRAK2. J. Immunol. 183, 2150–2158 (2009). 27. Ceppi, M. et al. MicroRNA-155 modulates the interleukin-1 signaling pathway in activated human monocyte-derived dendritic cells. Proc. Natl Acad. Sci. USA 106, 2735–2740 (2009).

NATuRE REvIEWs | Immunology

28. Tang, B. et al. Identification of MyD88 as a novel target of miR-155, involved in negative regulation of Helicobacter pylori-induced inflammation. FEBS Lett. 584, 1481–1486 (2010). 29. Huang, R. S., Hu, G. Q., Lin, B., Lin, Z. Y. & Sun, C. C. MicroRNA-155 silencing enhances inflammatory response and lipid uptake in oxidized low-density lipoprotein-stimulated human THP-1 macrophages. J. Investig. Med. 51, 961–967 (2010). 30. Starczynowski, D. T. et al. Identification of miR-145 and miR-146a as mediators of the 5q-syndrome phenotype. Nature Med. 16, 49–58 (2010). The first study to show directly that chromosomal deletion of miRNAs can be associated with a particular disease: in this case, 5q myelodysplastic syndrome. 31. Mansell, A. et al. Suppressor of cytokine signaling 1 negatively regulates Toll-like receptor signaling by mediating Mal degradation. Nature Immunol. 7, 148–155 (2006). 32. Alsaleh, G. et al. Bruton’s tyrosine kinase is involved in miR-346-related regulation of IL-18 release by lipopolysaccharide-activated rheumatoid fibroblastlike synoviocytes. J. Immunol. 182, 5088–5097 (2009). 33. Chen, K. & Rajewsky, N. The evolution of gene regulation by transcription factors and microRNAs. Nature Rev. Genet. 8, 93–103 (2007). 34. Martinez, N. J. & Walhout, A. J. The interplay between transcription factors and microRNAs in genome-scale regulatory networks. Bioessays 31, 435–445 (2009). 35. Kohlhaas, S. et al. Cutting edge: the Foxp3 target miR-155 contributes to the development of regulatory T cells. J. Immunol. 182, 2578–2582 (2009). 36. Iliopoulos, D., Jaeger, S. A., Hirsch, H. A., Bulyk, M. L. & Struhl, K. STAT3 activation of miR-21 and miR-181b-1 via PTEN and CYLD are part of the epigenetic switch linking inflammation to cancer. Mol. Cell 39, 493–506 (2010). 37. Li, T. et al. MicroRNAs modulate the noncanonical transcription factor NF-κB pathway by regulating expression of the kinase IKKα during macrophage differentiation. Nature Immunol. 11, 799–805 (2010). 38. Chen, R. et al. Regulation of IKKβ by miR-199a affects NF-κB activity in ovarian cancer cells. Oncogene 27, 4712–4723 (2008). 39. Gottwein, E. et al. A viral microRNA functions as an orthologue of cellular miR-155. Nature 450, 1096–1099 (2007). One of the first studies to show that certain viral miRNAs are homologous to cellular miRNAs, providing another mechanism by which viruses can manipulate the host immune response. 40. Xiao, B. et al. Induction of microRNA-155 during Helicobacter pylori infection and its negative regulatory role in the inflammatory response. J. Infect. Dis. 200, 916–925 (2009).

voLumE 11 | mARcH 2011 | 173 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 41. Bazzoni, F. et al. Induction and regulatory function of miR-9 in human monocytes and neutrophils exposed to proinflammatory signals. Proc. Natl Acad. Sci. USA 106, 5282–5287 (2009). 42. Worm, J. et al. Silencing of microRNA-155 in mice during acute inflammatory response leads to derepression of c/ebpβ and down-regulation of G-CSF. Nucleic Acids Res. 37, 5784–5792 (2009). 43. Costinean, S. et al. Src homology 2 domain-containing inositol-5-phosphatase and CCAAT enhancer-binding protein-β are targeted by miR-155 in B cells of Eμ-MiR‑155 transgenic mice. Blood 114, 1374–1382 (2009). 44. Jennewein, C., von Knethen, A., Schmid, T. & Brune, B. MicroRNA-27b contributes to lipopolysaccharidemediated peroxisome proliferator-activated receptor-γ (PPARγ) mRNA destabilization. J. Biol. Chem. 285, 11846–11853 (2010). 45. Lagos, D. et al. miR-132 regulates antiviral innate immunity through suppression of the p300 transcriptional co-activator. Nature Cell Biol. 12, 513–519 (2010). 46. Asirvatham, A. J., Gregorie, C. J., Hu, Z., Magner, W. J. & Tomasi, T. B. MicroRNA targets in immune genes and the Dicer/Argonaute and ARE machinery components. Mol. Immunol. 45, 1995–2006 (2008). 47. Asirvatham, A. J., Magner, W. J. & Tomasi, T. B. miRNA regulation of cytokine genes. Cytokine 45, 58–69 (2009). 48. Iliopoulos, D., Hirsch, H. A. & Struhl, K. An epigenetic switch involving NF-κB, Lin28, Let-7 microRNA, and IL6 links inflammation to cell transformation. Cell 139, 693–706 (2009). 49. Viswanathan, S. R., Daley, G. Q. & Gregory, R. I. Selective blockade of microRNA processing by Lin28. Science 320, 97–100 (2008). 50. Tili, E. et al. Modulation of miR-155 and miR-125b levels following lipopolysaccharide/TNF-α stimulation and their possible roles in regulating the response to endotoxin shock. J. Immunol. 179, 5082–5089 (2007). 51. Sharma, A. et al. Posttranscriptional regulation of interleukin-10 expression by hsa-miR-106a. Proc. Natl Acad. Sci. USA 106, 5761–5766 (2009). 52. Lu, T. X., Munitz, A. & Rothenberg, M. E. MicroRNA-21 is up-regulated in allergic airway inflammation and regulates IL-12p35 expression. J. Immunol. 182, 4994–5002 (2009). 53. Carballo, E., Lai, W. S. & Blackshear, P. J. Feedback inhibition of macrophage tumor necrosis factor-α production by tristetraprolin. Science 281, 1001–1005 (1998). 54. Lai, W. S. et al. Evidence that tristetraprolin binds to AU-rich elements and promotes the deadenylation and destabilization of tumor necrosis factor-α mRNA. Mol. Cell Biol. 19, 4311–4323 (1999). 55. Stoecklin, G. et al. Genome-wide analysis identifies interleukin-10 mRNA as target of tristetraprolin. J. Biol. Chem. 283, 11689–11699 (2008). 56. Jing, Q. et al. Involvement of microRNA in AU-rich element-mediated mRNA instability. Cell 120, 623–634 (2005). 57. El Gazzar, M. & McCall, C. E. MicroRNAs distinguish translational from transcriptional silencing during endotoxin tolerance. J. Biol. Chem. 285, 20940–20951 (2010). 58. Ma, F. et al. MicroRNA-466l upregulates IL-10 expression in TLR-triggered macrophages by antagonizing RNA-binding protein tristetraprolinmediated IL-10 mRNA degradation. J. Immunol. 184, 6053–6059 (2010). 59. Vasudevan, S. & Steitz, J. A. AU-rich-elementmediated upregulation of translation by FXR1 and Argonaute 2. Cell 128, 1105–1118 (2007). 60. Thai, T. H. et al. Regulation of the germinal center response by microRNA-155. Science 316, 604–608 (2007). 61. Shaked, I. et al. MicroRNA-132 potentiates cholinergic anti-inflammatory signaling by targeting acetylcholinesterase. Immunity 31, 965–973 (2009). 62. Sheedy, F. J. et al. Negative regulation of TLR4 via targeting of the proinflammatory tumor suppressor PDCD4 by the microRNA miR-21. Nature Immunol. 11, 141–147 (2010). 63. Yang, H. S. et al. The transformation suppressor Pdcd4 is a novel eukaryotic translation initiation factor 4A binding protein that inhibits translation. Mol. Cell Biol. 23, 26–37 (2003). 64. Loh, P. G. et al. Structural basis for translational inhibition by the tumour suppressor Pdcd4. EMBO J. 28, 274–285 (2009).

65. Koromilas, A. E., Lazaris-Karatzas, A. & Sonenberg, N. mRNAs containing extensive secondary structure in their 5ʹ non-coding region translate efficiently in cells overexpressing initiation factor eIF-4E. EMBO J. 11, 4153–4158 (1992). 66. O’Connell, R. M., Chaudhuri, A. A., Rao, D. S. & Baltimore, D. Inositol phosphatase SHIP1 is a primary target of miR-155. Proc. Natl Acad. Sci. USA 106, 7113–7118 (2009). This paper identifies SHIP1 as an important target of miR-155 in TLR signalling. 67. Cremer, T. J. et al. MiR-155 induction by F. novicida but not the virulent F. tularensis results in SHIP downregulation and enhanced pro-inflammatory cytokine response. PLoS One 4, e8508 (2009). 68. An, H. et al. Src homology 2 domain-containing inositol-5-phosphatase 1 (SHIP1) negatively regulates TLR4-mediated LPS response primarily through a phosphatase activity- and PI-3K-independent mechanism. Blood 105, 4685–4692 (2005). 69. Gabhann, J. N. et al. Absence of SHIP-1 results in constitutive phosphorylation of tank-binding kinase 1 and enhanced TLR3-dependent IFN-β production. J. Immunol. 184, 2314–2320 (2010). 70. Sly, L. M., Rauh, M. J., Kalesnikoff, J., Buchse, T. & Krystal, G. SHIP, SHIP2 and PTEN activities are regulated in vivo by modulation of their protein levels: SHIP is up-regulated in macrophages and mast cells by lipopolysaccharide. Exp. Hematol. 31, 1170–1181 (2003). 71. Perry, M. M. et al. Rapid changes in microRNA-146a expression negatively regulate the IL-1β-induced inflammatory response in human lung alveolar epithelial cells. J. Immunol. 180, 5689–5698 (2008). 72. Bhaumik, D. et al. MicroRNAs miR-146a/b negatively modulate the senescence-associated inflammatory mediators IL-6 and IL-8. Aging 1, 402–411 (2009). 73. Jones, S. W. et al. The identification of differentially expressed microRNA in osteoarthritic tissue that modulate the production of TNF-α and MMP13. Osteoarthritis Cartilage 17, 464–472 (2009). 74. Nahid, M. A., Pauley, K. M., Satoh, M. & Chan, E. K. miR-146a is critical for endotoxin-induced tolerance: implication in innate immunity. J. Biol. Chem. 284, 34590–34599 (2009). 75. Curtale, G. et al. An emerging player in the adaptive immune response: microRNA-146a is a modulator of IL-2 expression and activation-induced cell death in T lymphocytes. Blood 115, 265–273 (2010). 76. Cameron, J. E. et al. Epstein–Barr virus latent membrane protein 1 induces cellular microRNA miR-146a, a modulator of lymphocyte signaling pathways. J. Virol. 82, 1946–1958 (2008). 77. Motsch, N., Pfuhl, T., Mrazek, J., Barth, S. & Grasser, F. A. Epstein–Barr virus-encoded latent membrane protein 1 (LMP1) induces the expression of the cellular microRNA miR-146a. RNA Biol. 4, 131–137 (2007). 78. Chassin, C. et al. miR-146a mediates protective innate immune tolerance in the neonate intestine. Cell Host Microbe 8, 358–368 (2010). 79. Jurkin, J. et al. miR-146a is differentially expressed by myeloid dendritic cell subsets and desensitizes cells to TLR2-dependent activation. J. Immunol. 184, 4955–4965 (2010). 80. Liu, G. et al. miR-147, a microRNA that is induced upon Toll-like receptor stimulation, regulates murine macrophage inflammatory responses. Proc. Natl Acad. Sci. USA 106, 15819–15824 (2009). 81. Stanczyk, J. et al. Altered expression of microRNA in synovial fibroblasts and synovial tissue in rheumatoid arthritis. Arthritis Rheum. 58, 1001–1009 (2008). One of the first studies to analyse miRNAs in the context of rheumatoid arthritis. 82. Wang, P. et al. Inducible microRNA-155 feedback promotes type I IFN signaling in antiviral innate immunity by targeting suppressor of cytokine signaling 1. J. Immunol. 185, 6226–6233 (2010). 83. O’Connell, R. M. et al. MicroRNA-155 promotes autoimmune inflammation by enhancing inflammatory T cell development. Immunity 33, 607–619 (2010). 84. Zhou, H. et al. miR-155 and its star-form partner miR-155* cooperatively regulate type I interferon production by human plasmacytoid dendritic cells. Blood 116, 5885–5894 (2010). 85. Yoshimura, A., Naka, T. & Kubo, M. SOCS proteins, cytokine signalling and immune regulation. Nature Rev. Immunol. 7, 454–465 (2007). 86. McCoy, C. E. The role of miRNAs in cytokine signalling. Front. Biosci. (in the press).

174 | mARcH 2011 | voLumE 11

87. Rodriguez, A. et al. Requirement of bic/ microRNA-155 for normal immune function. Science 316, 608–611 (2007). The first study to analyse a role for miR-155 in the immune system using miR-155-deficient mice. 88. Vigorito, E. et al. microRNA-155 regulates the generation of immunoglobulin class-switched plasma cells. Immunity 27, 847–859 (2007). 89. Martinez-Nunez, R. T., Louafi, F., Friedmann, P. S. & Sanchez-Elsner, T. MicroRNA-155 modulates the pathogen binding ability of dendritic cells (DCs) by down-regulation of DC-specific intercellular adhesion molecule-3 grabbing non-integrin (DC-SIGN). J. Biol. Chem. 284, 16334–16342 (2009). 90. Marta, M., Andersson, A., Isaksson, M., Kampe, O. & Lobell, A. Unexpected regulatory roles of TLR4 and TLR9 in experimental autoimmune encephalomyelitis. Eur. J. Immunol. 38, 565–575 (2008). 91. Dorsett, Y. et al. MicroRNA-155 suppresses activation-induced cytidine deaminase-mediated Myc‑Igh translocation. Immunity 28, 630–638 (2008). 92. Teng, G. et al. MicroRNA-155 is a negative regulator of activation-induced cytidine deaminase. Immunity 28, 621–629 (2008). Through genetic mutation of the miRNA binding site, references 91 and 92 describe the direct effect of miR-155 on one of its target mRNAs. 93. Asadullah, K., Sterry, W. & Volk, H. D. Interleukin-10 therapy — review of a new approach. Pharmacol. Rev. 55, 241–269 (2003). 94. O’Garra, A., Barrat, F. J., Castro, A. G., Vicari, A. & Hawrylowicz, C. Strategies for use of IL-10 or its antagonists in human disease. Immunol. Rev. 223, 114–131 (2008). 95. Hunter, M. P. et al. Detection of microRNA expression in human peripheral blood microvesicles. PLoS One 3, e3694 (2008). 96. Luo, S. S. et al. Human villous trophoblasts express and secrete placenta-specific microRNAs into maternal circulation via exosomes. Biol. Reprod. 81, 717–729 (2009). 97. Michael, A. et al. Exosomes from human saliva as a source of microRNA biomarkers. Oral Dis. 16, 34–38 (2010). 98. Valadi, H. et al. Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nature Cell Biol. 9, 654–659 (2007). An important paper showing that miRNAs in exosomes can mediate effects on neighbouring cells in an autocrine manner. 99. Camussi, G., Deregibus, M. C., Bruno, S., Cantaluppi, V. & Biancone, L. Exosomes/microvesicles as a mechanism of cell-to-cell communication. Kidney Int. 78, 838–848 (2010). 100. van Niel, G., Porto-Carreiro, I., Simoes, S. & Raposo, G. Exosomes: a common pathway for a specialized function. J. Biochem. 140, 13–21 (2006). 101. Matsumoto, K. et al. Exosomes secreted from monocyte-derived dendritic cells support in vitro naive CD4+ T cell survival through NF-κB activation. Cell. Immunol. 231, 20–29 (2004). 102. Raposo, G. et al. B lymphocytes secrete antigenpresenting vesicles. J. Exp. Med. 183, 1161–1172 (1996). 103. Thery, C. et al. Indirect activation of naive CD4+ T cells by dendritic cell-derived exosomes. Nature Immunol. 3, 1156–1162 (2002). 104. Recchiuti, A., Krishnamoorthy, S., Fredman, G., Chiang, N. & Serhan, C. N. MicroRNAs in resolution of acute inflammation: identification of novel resolvin D1–miRNA circuits. FASEB J. 18 Oct 2010 (doi:10.1096/fj.10-169599). 105. Wu, F. et al. MicroRNAs are differentially expressed in ulcerative colitis and alter expression of macrophage inflammatory peptide-2α. Gastroenterology 135, 1624–1635 (2008). 106. Calin, G. A. & Croce, C. M. MicroRNA signatures in human cancers. Nature Rev. Cancer 6, 857–866 (2006). 107. Chen, R., Alvero, A. B., Silasi, D. A., Steffensen, K. D. & Mor, G. Cancers take their Toll — the function and regulation of Toll-like receptors in cancer cells. Oncogene 27, 225–233 (2008). 108. Hussain, S. P. & Harris, C. C. Inflammation and cancer: an ancient link with novel potentials. Int. J. Cancer 121, 2373–2380 (2007). 109. Kanwar, J. R., Mahidhara, G. & Kanwar, R. K. MicroRNA in human cancer and chronic inflammatory diseases. Front. Biosci. 2, 1113–1126 (2010).

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 110. Pauley, K. M., Cha, S. & Chan, E. K. MicroRNA in autoimmunity and autoimmune diseases. J. Autoimmun. 32, 189–194 (2009). 111. Oglesby, I. K., McElvaney, N. G. & Greene, C. M. MicroRNAs in inflammatory lung disease — master regulators or target practice? Respir. Res. 11, 148 (2010). 112. Fasseu, M. et al. Identification of restricted subsets of mature microRNA abnormally expressed in inactive colonic mucosa of patients with inflammatory bowel disease. PLoS One 5, e13160 (2010). 113. Iborra, M., Bernuzzi, F., Invernizzi, P. & Danese, S. MicroRNAs in autoimmunity and inflammatory bowel disease: crucial regulators in immune response. Autoimmun. Rev. 11 Jul 2010 (doi:10.1016/ j.autrev.2010.07.002). 114. Wu, F. et al. Identification of microRNAs associated with ileal and colonic Crohn’s disease. Inflamm. Bowel Dis. 16, 1729–1738 (2010). 115. Nakasa, T. et al. Expression of microRNA-146 in rheumatoid arthritis synovial tissue. Arthritis Rheum. 58, 1284–1292 (2008). 116. Murata, K. et al. Plasma and synovial fluid microRNAs as potential biomarkers of rheumatoid arthritis and osteoarthritis. Arthritis Res. Ther. 12, R86 (2010). 117. Stanczyk, J. et al. Altered expression of miR-203 in rheumatoid arthritis synovial fibroblasts and its role in fibroblast activation. Arthritis Rheum. 27 Oct 2010 (doi:10.1002/art.30115). 118. Iliopoulos, D., Malizos, K. N., Oikonomou, P. & Tsezou, A. Integrative microRNA and proteomic approaches identify novel osteoarthritis genes and their collaborative metabolic and inflammatory networks. PLoS ONE 3, e3740 (2008). 119. Sonkoly, E. et al. MicroRNAs: novel regulators involved in the pathogenesis of psoriasis? PLoS ONE 2, e610 (2007).

120. Dai, Y. et al. Microarray analysis of microRNA expression in peripheral blood cells of systemic lupus erythematosus patients. Lupus 16, 939–946 (2007). 121. Chatzikyriakidou, A., Voulgari, P. V., Georgiou, I. & Drosos, A. A. A polymorphism in the 3ʹ-UTR of interleukin-1 receptor-associated kinase (IRAK1), a target gene of miR-146a, is associated with rheumatoid arthritis susceptibility. Joint Bone Spine 77, 411–413 (2010). 122. Selbach, M. et al. Widespread changes in protein synthesis induced by microRNAs. Nature 455, 58–63 (2008). This paper shows that each miRNA can repress the production of proteins on a global scale, although this repression is relatively mild. 123. Kuchen, S. et al. Regulation of microRNA expression and abundance during lymphopoiesis. Immunity 32, 828–839 (2010). 124. Brown, B. D. et al. Endogenous microRNA can be broadly exploited to regulate transgene expression according to tissue, lineage and differentiation state. Nature Biotechnol. 25, 1457–1467 (2007). 125. 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). 126. Landgraf, P. et al. A mammalian microRNA expression atlas based on small RNA library sequencing. Cell 129, 1401–1414 (2007). This study provides an in-depth analysis of miRNA libraries from multiple organs and cell types. 127. Yin, Q. et al. MicroRNA-155 is an Epstein–Barr virusinduced gene that modulates Epstein–Barr virusregulated gene expression pathways. J. Virol. 82, 5295–5306 (2008). 128. Moschos, S. A. et al. Expression profiling in vivo demonstrates rapid changes in lung microRNA levels following lipopolysaccharide-induced inflammation but not in the anti-inflammatory action of glucocorticoids. BMC Genomics 8, 240 (2007).

NATuRE REvIEWs | Immunology

129. Gantier, M. P. New perspectives in microRNA regulation of innate immunity. J. Interferon Cytokine Res. 30, 283–289 (2010). 130. Zhou, R., Hu, G., Gong, A. Y. & Chen, X. M. Binding of NF-κB p65 subunit to the promoter elements is involved in LPS-induced transactivation of miRNA genes in human biliary epithelial cells. Nucleic Acids Res. 38, 3222–3232 (2010). 131. Zhou, R. et al. NF-κB p65-dependent transactivation of miRNA genes following Cryptosporidium parvum infection stimulates epithelial cell immune responses. PLoS Pathog. 5, e1000681 (2009). 132. Cameron, J. E. et al. Epstein–Barr virus growth/ latency III program alters cellular microRNA expression. Virology 382, 257–266 (2008). 133. O’Hara, S. P. et al. NFκB p50-CCAAT/enhancerbinding protein-β (C/EBPβ)-mediated transcriptional repression of microRNA let-7i following microbial infection. J. Biol. Chem. 285, 216–225 (2010). 134. Hu, G. et al. MicroRNA-98 and let-7 confer cholangiocyte expression of cytokine-inducible Src homology 2-containing protein in response to microbial challenge. J. Immunol. 183, 1617–1624 (2009).

Acknowledgements

The authors would like to thank their respective funding bodies: L.A.O’N. is supported by Science Foundation Ireland, F.J.S. was supported by the Irish Research Council for Science, Engineering & Technology and C.E.M. is supported by a Health Research Board Ireland/Marie Curie fellowship.

Competing interests statement

The authors declare no competing financial interests.

FURTHER INFORMATION TargetScan: http://www.targetscan.org/ All lInkS ARe ACTIve In The onlIne pdF

voLumE 11 | mARcH 2011 | 175 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

DCs and NK cells: critical effectors in the immune response to HIV‑1 Marcus Altfeld*, Lena Fadda*, Davor Frleta‡ and Nina Bhardwaj‡

Abstract | Dendritic cells (DCs) and natural killer (NK) cells have central roles in antiviral immunity by shaping the quality of the adaptive immune response to viruses and by mediating direct antiviral activity. HIV‑1 infection is characterized by a severe dysregulation of the antiviral immune response that starts during early infection. This Review describes recent insights into how HIV‑1 infection affects DC and NK cell function, and the roles of these innate immune cells in HIV‑1 pathogenesis. The importance of understanding DC and NK cell crosstalk during HIV infection for the development of effective antiviral strategies is also discussed.

*Ragon Institute of MGH, MIT and Harvard, Massachusetts General Hospital, Charlestown, Massachusetts 02129, USA. ‡ New York University Cancer Institute, NYU Langone Medical Center, New York, New York 10016, USA. Correspondence to M.A. and N.B. e-mails: [email protected]; [email protected] All authors contributed equally to this work. doi:10.1038/nri2935

Despite some recent progress in HIV vaccine strategies, HIV infection remains a worldwide problem. Growing evidence indicates that the immune system is dysregu‑ lated during HIV infection, and this compromises the efficacy of potential immune therapies for HIV. Humoral immune responses occur late in HIV‑1 infection, with neutralizing antibodies appearing 3 months or more after initial infection1. T cell responses are elicited around 1–2 weeks after infection, but are largely ineffective owing to the early emergence of antigen escape variants of HIV1. Recent reports have highlighted the dual role of the innate immune system in early viral control and in con‑ tributing to disease pathology. Dendritic cells (DCs) and natural killer (NK) cells are vital mediators of the innate immune system and promote the development of adap‑ tive immune responses. DCs are crucial for activating and conditioning virus‑specific T cells, a process that is largely influenced by the preceding innate immune response. NK cells impede the early spread of viruses by producing cytokines and directly killing infected cells. HIV vaccine strategies that use DCs, through either in vitro manipula‑ tion of DCs isolated from patients or in vivo targeting of DC subsets, are currently being investigated. The success of these approaches will depend on a proper understand‑ ing of how DC biology is affected by HIV‑1 infection. NK cells may be crucial for early control of HIV infection and can have important roles in editing the function of DCs, thereby affecting the ability of DCs to prime antiviral effector T cells. This Review focuses on the roles of these two innate cell types during HIV‑1 infection.

DCs bridge innate and adaptive immunity Human DCs are rare potent antigen‑presenting cells that can be generally divided into myeloid CD11c+ ‘con‑ ventional’ DCs (cDCs) or plasmacytoid DCs (pDCs)2

(Table 1).

Both subsets specialize in detecting viruses and initiating innate and adaptive immune responses that lead to viral elimination or control. DCs express several receptors for recognizing viruses3, including pat‑ tern recognition receptors (PRRs) such as the Toll‑like receptors (TLRs) and C‑type lectins. DCs detect viruses in peripheral tissue sites and, following activation and viral uptake, migrate to draining lymph nodes, where they trigger adaptive immune responses and promote NK cell activation4 (Table 1). Activated cDCs produce cytokines such as interleukin‑12 (IL‑12), IL‑15 and IL‑18. IL‑12 is crucial for cDCs to induce T helper 1 (TH1) cell responses, which subsequently promote the potent cyto‑ toxic T lymphocyte (CTL) responses that are necessary for clearing virus‑infected cells5. Both IL‑12 and IL‑15 produced by cDCs can activate NK cells4 (Table 1). pDCs produce more type I interferons (IFNs) in response to HIV than any other cell type in the body, and stimulate cDCs in a bystander fashion as well as directly activating NK cells6. In the following section, we describe recent observations that have been made concerning DC biol‑ ogy and function during HIV‑1 infection. Specifically, we focus on how DCs bind and recognize HIV virions, and how DCs are in turn modulated by the virus, ultimately leading to their dysregulation in vivo. The implications of these events underlie the need to develop vaccine strate‑ gies that enhance the ability of DCs to prime potent T cell responses that can subvert the potential inhibitory effects of HIV on the immune system.

Binding and recognition of HIV by DCs HIV entry receptors. Binding and internalization of HIV by DCs is mediated by the various HIV entry recep‑ tors (which include CD4, CC‑chemokine receptor 5 (CCR5) and CXC‑chemokine receptor 4 (CXCR4)).

176 | mARCH 2011 | VoLume 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Table 1 | DC subsets and their role in HIV infection Conventional DCs (HLA-DR+CD11c+) Langerhans cells Dermal DCs Location

Epidermis, gut lumen

Blood DCs

CD103–

CD103+

BDCA3–

BDCA3+

Dermis

Dermis

Blood, secondary lymphoid organs

Blood

Plasmacytoid DCs (HLA-DR+CD123+)

Refs

• Blood • Secondary lymphoid organs • Peripheral tissues (for example, skin

2,7, 9,10, 15,27

and lungs)

C-type lectin Langerin expression

DC‑SIGN, DEC205

Langerin, DEC205

DEC205, DCIR

CLEC9A

BDCA2

Role in HIV infection

Internalize HIV into degradative Birbeck granules

Bind to HIV and transmit the virus to T cells in draining lymph nodes

Unknown

Unknown

Unknown

• Produce type I IFNs, which inhibit

9,10, 43,45, 53

TLR expression

TLR2,3,5*

TLR2,3,4,5*

TLR2,3,4,5*

TLR2,3,4,7,8

TLR2,3,8

TLR7,9

2, 15,27

Cytokine production

IL‑12, IL‑15, IL‑23, IL‑6, TNF, IL‑1β

IFNα, IFNβ, IL‑6, TNF

2,4–6

Function

• Priming of antigen‑specific CD4+ T cells, CD8+ T cells and B cells • NK cell activation through IL‑12

• Induction of TReg cells • Induction of plasma cells • NK cell activation through type I IFNs

2,4–6

Pathology during HIV infection

• Reduced frequency in peripheral blood • Conflicting reports of functionality or dysfunction

• Reduced frequency in peripheral

25, 26,28, 31–37

viral replication and induce bystander T cell death • Induce TReg cells through HIV‑induced IDO expression • Recruit T cells to sites of HIV infection (and thus facilitate viral spread) by producing chemokines such as CCL5

blood • Conflicting reports of functionality or dysfunction

BDCA, blood DC antigen; CCL5, CC‑chemokine ligand 5 (also known as RANTES); CLEC, C‑type lectin; DC, dendritic cell; DCIR, DC immunoreceptor (also known as CLEC4A); DC‑SIGN, DC‑specific ICAM3‑grabbing non‑integrin; IDO, indoleamine 2,3‑dioxygenase; IFN, interferon; IL, interleukin; NK, natural killer; TLR, Toll‑like receptor; TNF, tumour necrosis factor; TReg, regulatory T. *There have been confounding reports regarding TLR distribution on Langerhans cells and dermal DCs.

Viral synapse a polarized synaptic contact point for cell to cell transmission of viruses. These contacts can occur between DCs and CD4+ T cells.

Clusterin a ubiquitously expressed glycoprotein that functions as an extracellular chaperone. It is widely expressed in various types of cancer.

Birbeck granule a membrane-bound structure that is found in the cytoplasm of langerhans cells. These granules are rod- or tennis racket-shaped with a central linear density. Their formation is induced by langerin, an endocytic C-type lectin receptor that is specific to langerhans cells.

other chemokine receptors, including CCR3, CCR8, CCR9 and CXCR6 (Refs 7,8), have also been suggested to participate in HIV‑1 entry. In addition, cDCs express C‑type lectins that bind to HIV, including DC‑specific ICAm3‑grabbing non‑integrin (DC‑SIGN), langerin (also known as CLeC4K or CD207) and DC immu‑ noreceptor (DCIR; also known as CLeC4A)7,9 (Table 1). expression of these receptors varies based on cDC sub‑ type, localization and activation state. Although pDCs express C‑type lectins, such as blood DC antigen 2 (BDCA2; also known as CLeC4C), binding of HIV to pDCs is generally mediated by interaction of the HIV gp120 envelope protein with CD4 (Ref. 10) (Table 1). The identities of the receptors involved in virus–cDC interactions can affect virus fate. Binding of HIV to DC‑SIGN through its gp120 envelope protein leads to internalization of the virus into DC early endosomal compartments, where it is not degraded. This may enable DCs to deliver intact HIV virions to draining lymph nodes, where DC–T cell interactions could pro‑ mote infection of T cells through viral synapses and/ or cell–cell fusion. However, the contribution of this mode of transmission has been called into question, as the transfer of HIV‑1 to T cells can occur independ‑ ently of DC‑SIGN11. Conversely, another report has highlighted how interaction of HIV with DC‑SIGN may be important for transmission of HIV. Clusterin in semen is glycosylated differently to clusterin in blood

and specifically blocks the HIV–DC‑SIGN interaction, thereby preventing HIV transmission to CD4+ T cells. The different efficacies with which these two forms of clusterin block HIV–DC‑SIGN interactions may explain why sexual exposure to HIV is less likely to result in infection than blood‑borne exposure12. DCIR was also recently found to bind to HIV virions and, in a similar way to DC‑SIGN, promotes transmission of infectious virus to CD4+ T cells13. The C‑type lectin langerin, which can bind and inter‑ nalize HIV, is expressed by epidermal Langerhans cells, and migratory Langerhans cells are believed to enhance sexually transmitted HIV infection by promoting the spread of the virus to T cells in draining lymph nodes. In contrast to DC‑SIGN‑mediated internalization, HIV internalized through langerin is trafficked to birbeck granules, where the virus is rapidly degraded14; this sug‑ gests that Langerhans cells also function as a barrier to HIV infection. However, high viral concentrations enable langerin‑independent internalization of HIV by Langerhans cells, resulting in the transport of intact HIV to T cells by migratory Langerhans cells. In accord‑ ance with this, others have shown that in vitro‑derived Langerhans cells can transmit infectious HIV to T cells15. Thus, the fate of HIV varies depending on the particular cDC receptors used to bind the virus, and this empha‑ sizes how different DC subsets can either contribute to or hinder HIV infection.

NATuRe ReVIewS | ImmunoLogy

VoLume 11 | mARCH 2011 | 177 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Innate recognition of HIV by DCs. whereas binding and internalization of HIV is mediated by a variety of DC‑expressed surface receptors, activation of DCs by HIV is mainly thought to involve intracellular mem‑ bers of the TLR family. Human pDCs express TLR7 and TLR9 (Ref. 16), and recognition of HIV‑1 single‑ stranded RNA by TLR7 leads to pDC production of type I IFNs and other inflammatory cytokines. The ini‑ tial gp120–CD4 interaction promotes endocytosis of HIV‑1 by pDCs and the subsequent activation of TLR7 by viral RNA17. Genomic RNA from HIV‑1 contains several immunostimulatory Gu‑rich sequences, which activate the TLR7 pathway 18. Following the detection of HIV by this route, pDCs rapidly secrete high levels of IFNα owing to their constitutive expression of IFN regulatory factor 7 (IRF7)19. However, HIV does not induce the maturation of cDCs by activating TLRs17. This is surprising consider‑ ing that blood cDCs express TLR7 and TLR8, both of which can be activated by HIV‑derived single‑stranded RNA. one possible explanation is low viral replica‑ tion within cDCs; only 1–3% of cultured DCs become infected in vitro, and circulating blood DCs isolated from patients with HIV do not appear to be infected with the virus20,21. A recent study using a pseudotyped virus that lacked HIV envelope protein showed that when this strain of HIV‑1 carries the vpx gene, infection of DCs promotes their maturation and production of type I IFNs, and facilitates antiviral T cell immunity. This response is mediated by interaction of newly synthesized HIV‑1 cap‑ sid with cellular cyclophilin A (also known as PPIase A) and activation of the type I IFN‑inducing transcription factor IRF3 through an unknown cytoplasmic sen‑ sor 22. Thus, it is possible that there are too few virions within HIV‑1‑exposed cDCs to trigger TLR signalling, or that vpx is essential for this interaction. However, although viral replication in pDCs is also low, pDCs rapidly respond to HIV‑1 through TLR7. Therefore, another explanation is that HIV interaction with C‑type lectins may abrogate subsequent TLR responsiveness in cDCs. HIV, in a similar way to Mycobacterium spp., has been implicated in inhibiting TLR stimulation of cDCs through its interactions with DC‑SIGN23 and this is discussed in more detail below.

Autophagy an evolutionarily conserved process in which acidic double-membrane vacuoles sequester intracellular contents (such as damaged organelles and macromolecules) and target them for degradation, through fusion to secondary lysosomes.

Can TLR signalling promote HIV replication in DCs? De novo replication of integrated HIV‑1 in imma‑ ture cDCs can be initiated by TLR8‑ and DC‑SIGN‑ mediated signal transduction events. HIV is targeted to TLR8‑containing endosomal compartments by DC‑SIGN, and this initiates the transcription of inte‑ grated HIV‑1 DNA through TLR8 activation and sub‑ sequent nuclear factor‑κB (NF‑κB) signalling 24. The binding of gp120 to DC‑SIGN leads to RAF1‑mediated phosphorylation of the p65 subunit of NF‑κB, thereby permitting transcription elongation of nascent HIV‑1 transcripts and productive transcription within infected DCs24. Thus, ligation of both TLR8 and DC‑SIGN is required for the induction of signal transduction path‑ ways that promote the synthesis of complete viral tran‑ scripts from integrated proviral DNA. Co‑infection with

Candida albicans or Mycobacterium tuberculosis also triggers RAF1‑dependent phosphorylation, suggesting that co‑infection could enhance the productive tran‑ scription of HIV24. Although this in vitro work argues for a role of TLR8 in promoting HIV replication in cDCs, it is unclear whether this is a physiological mechanism for productive infection in cDCs, as replication within cDCs is very low. However, it raises the issue of whether the use of TLR8 agonists as immunotherapeutic adju‑ vants might increase productive replication within cDCs containing integrated HIV‑1. Autophagy and HIV recognition. autophagy leads to the delivery of cytosolic components (such as signalling mol‑ ecules and damaged organelles) to lysosomal compart‑ ments, which contain PRRs25. Following HIV‑1 infection, fusion between endosomes and autophagosomes is inhibited, thereby preventing autophagy‑mediated viral degradation26. This inhibition occurs through activation of mammalian target of rapamycin (mToR), a serine/ threonine kinase that is a negative regulator of autophagy. In accordance with inhibition of autophagy, HIV‑1 rep‑ lication in DCs leads to decreased cathepsin activity in these cells, possibly owing to a blockade of lysosomal fusion, which is necessary for cathepsin activation. This results in enhanced survival of HIV‑1 in phagosomes and decreased presentation of viral antigens27–29. In contrast to the aforementioned study, which showed that HIV can activate TLR8, HIV‑mediated inhi‑ bition of autophagy was shown by others to inhibit DC responsiveness to the classical TLR4 and TLR8 agonists lipopolysaccharide (LPS) and single‑stranded RNA26. one explanation for this discrepancy may be that acti‑ vation of TLR8 by HIV precedes autophagy inhibition. Alternatively, selective blockade of autophagosomes by HIV may depend on the particular HIV receptor used for viral entry. Clearly, additional studies are necessary to understand how HIV‑1 can modulate cDC responsive‑ ness to TLR agonists. It will also be important to deter‑ mine how HIV affects different blood cDC subtypes, such as the newly described BDCA3+ (also known as CD141+) blood DC subset30.

Dysregulation of DCs in HIV‑1 infection Decreased DC frequency. HIV‑1 infection leads to a progressive reduction in blood DC numbers (both cDCs and pDCs), which correlates with increasing plasma virus load and disease progression31. This decrease in DC frequency occurs early during acute HIV infection and is sustained in the later stages of infection20. Patients with chronic HIV infection also have fewer blood DCs compared with uninfected controls20,31. Direct infection of DCs has been suggested to be responsible for decreased DC frequency during HIV infection31. Several studies have evaluated whether DCs from the blood of patients with chronic HIV‑1 infection are infected with HIV, with some showing evidence of infection but others suggesting that no infection occurs21,32. Given the low level of HIV‑infected DCs, it is unlikely that direct infection can adequately explain the decreased DC frequency in the blood of patients with HIV.

178 | mARCH 2011 | VoLume 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Instead, the decline in blood DC frequency during viral infection may be due to indirect mechanisms. Aberrant IFNα production during HIV infection may impede the differentiation of cDCs from monocytes33 or other precursors. Also, DCs from patients with HIV show increased apoptosis compared with DCs from uninfected individuals34. A third possibility is that decreased DC frequen‑ cies in the blood may be due to the redistribution of DCs to secondary lymphoid organs. Indeed, several groups have reported an increase in pDCs in the lymph nodes35,36 and spleens37 of patients with HIV. The over‑ all functional capacity of DCs found in the secondary lymphoid organs of patients with HIV remains to be adequately evaluated. Finke et al.38 reported that cDC numbers returned to normal levels following antiretroviral therapy (ART). Another recent study showed no effect of primary HIV‑1 infection on cDC levels, and although pDC and CD4+ T cell numbers were reduced, they were restored following ART39. The differential restoration of cDC and pDC num‑ bers may be due to multiple components of the immune system going awry beyond repair in the chronic phase of infection. Altogether, these data warrant a closer scrutiny of the functional status of DC subsets during the immune recovery associated with ART. Altered cDC function. Although there is an overall consensus that blood DC numbers decline during HIV infection, it is still unclear whether DCs from patients with HIV are functionally impaired. As indicated previ‑ ously, HIV interaction with DC‑SIGN might negatively affect TLR‑induced activation of cDCs. Accordingly, HIV‑1 infection interferes with the maturation of blood cDCs and monocyte‑derived DCs in response to in vitro activation stimuli28,40. other studies have reported that cDCs are not func‑ tionally defective following HIV exposure20,21, showing that ex vivo isolated cDCs from patients with acute HIV infection may even be hyperresponsive to TLR stimula‑ tion20 and produce increased levels of IL‑12, IL‑6, TNF and CCL3 (also known as mIP1α) compared with cDCs from uninfected controls. By contrast, compared to pDCs isolated from uninfected controls, pDCs iso‑ lated from patients with acute HIV infection can show decreased IFNα production during the early stages of infection, but increased IFNα production at later stages of infection20. These discordant results may result from the limita‑ tions of analysing isolated DCs ex vivo. modulation of DC function in vivo may not necessarily be a result of direct HIV exposure, but could instead be due to factors that are produced by the host in response to infection. moreover, depending on the stage of HIV infection, DCs may be affected by the microenviron‑ ment. For example, during chronic HIV infection, monocytes have been shown to upregulate programmed cell death protein 1 (PD1) and produce IL‑10, possibly in response to the high levels of gut‑derived micro‑ bial products that are present in the plasma of these patients41. Circulating LPS and other pathogen‑derived

factors have been implicated in promoting the chronic immune activation that is observed during HIV infec‑ tion, either by direct activation of DCs or indirectly through stimulation of cells such as macrophages 42. Such stimuli may lead to partial maturation of cDCs in vivo, making these cells tolerant to subsequent stimuli ex vivo. Accordingly, several reports have indicated that cDCs from chronically infected patients are less effi‑ cient at stimulating T cell activation than cells from uninfected individuals31,40. understanding exactly how cDCs are functionally impaired during HIV infection is crucial for effective HIV vaccine design. This is especially true if functional impairment of DCs results from bystander mechanisms during HIV infection, as these mechanisms are likely to impede most DC‑targeting approaches, including the reintroduction of ex vivo‑manipulated DCs into patients. Altered pDC function. As with cDC function, it is unclear whether pDC function is impaired during HIV infection. Several groups have argued that type I IFN production by blood pDCs is attenuated during both acute and chronic HIV infection43,44, whereas others have indicated that circulating pDCs in patients with HIV viraemia show a normal type I IFN response45. In the former case, it is possible that circulating pDCs were activated in vivo by HIV before isolation and were, therefore, hyporesponsive to a secondary stimulus delivered ex vivo. Consistent with this, activation mark‑ ers are upregulated by blood pDCs during acute HIV infection, suggesting that these pDCs have responded to the virus in vivo20. Furthermore, IFN‑producing pDCs accumulate in the draining lymph nodes of patients with HIV35,36, although these pDCs are prone to apoptosis34,36. Nonetheless, evidence is emerging that pDCs have a cru‑ cial role in the immune dysregulation that leads to the development of AIDS. pDCs that are exposed to HIV upregulate the enzyme indoleamine 2,3‑dioxygenase (IDo), which metabolizes tryptophan to kynurenine, and this upreg‑ ulation is dependent on the gp120–CD4 interaction. Through expression of IDo, pDCs promote the differ‑ entiation of naive T cells into regulatory T (TReg) cells, which suppress effector T cell activation46. It remains to be elucidated whether increased frequencies of TReg cells prevent effective HIV‑specific adaptive immune responses and exacerbate HIV‑associated pathology, or whether TReg cells increase as a result of chronic immune activation in an effort to limit associated immuno‑ pathology. By promoting TReg cell responses, pDCs may be responsible for the decrease in TH17 cells that occurs during HIV infection47; this decrease is thought to lead to the loss of gut integrity and microbial translocation in patients with HIV42.

Response to HIV‑1 in humans: an overkill? uncontrolled and persistent inflammation contributes to the pathology that is associated with HIV infection. IFNα‑producing pDCs can contribute to the T cell loss that is characteristic of patients with AIDS by upreg‑ ulating CD4+ T cell expression of the pro‑apoptotic

NATuRe ReVIewS | ImmunoLogy

VoLume 11 | mARCH 2011 | 179 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS *+8

/QPQE[VGU *+8WRVCMGVJTQWIJ NCPIGTKPNGCFUVQ FGITCFCVKQPQHXKTWU

.CPIGTKP

&GITCFGF *+8

&%5+)0

%&

+.

8KTCN40# +.RTQFWEGFD[ 64GIEGNNUCPFOQPQE[VGU KPJKDKVU&%OCVWTCVKQP

*+8WRVCMGD[ &%5+)0DNQEMU E&%OCVWTCVKQP 'PFQUQOG

+PVCEV*+8 E&%

+ORCKTGFE&%CEVKXCVKQPKPJKDKVU CFCRVKXGKOOWPKV[VQ*+8

+.

64GIEGNN

R&%

R&%URTQOQVG 64GIEGNN FKȭGTGPVKCVKQP D[GZRTGUUKPI+&1

64GIEGNNUKPJKDKV6*EGNN HWPEVKQPUNGCFKPIVQKPETGCUGF OKETQDKCNVTCPUNQECVKQP

6.4 %JGOQMKPGU 6[RG+ +(0U 2TQOQVGKOOWPG CEVKXCVKQPCPF 6EGNNCRQRVQUKU

4GETWKV6EGNNU CPFHCEKNKVCVG HWTVJGTKPHGEVKQP

Figure 1 | cDCs and pDCs during HIV infection. Conventional dendritic cells (cDCs) are responsible for initiating antiviral adaptive immunity. Evidence indicates that cDC function is impaired during HIV‑1 infection. This is either due to direct viral interactions with cDCs (for example, interaction with DC‑SIGN (DC‑specific ICAM3‑grabbing non‑integrin) or 0CVWTG4GXKGYU^+OOWPQNQI[ blockade of autophagy) or a result of indirect mechanisms, such as the production of interleukin‑10 (IL‑10) by monocytes during infection. This cDC dysfunction could contribute to a lack of effective antiviral adaptive immunity. By contrast, plasmacytoid DCs (pDCs) are mediators of innate immunity. pDCs activated by HIV‑1 produce type I interferons (IFNs), which, in addition to inhibiting viral replication, may contribute to bystander CD4+ T cell death. Furthermore, evidence shows that pDCs produce T cell‑attracting chemokines, which may facilitate viral spread by providing a source of new T cells for HIV to infect. Finally, HIV‑exposed pDCs prime regulatory T (TReg) cells, which could impair cDC function and block effector T cell activity, further blunting adaptive immunity. Thus, pDCs can promote deleterious immunopathology during HIV‑1 infection. IDO, indoleamine 2,3‑dioxygenase; TH17, T helper 17; TLR7, Toll‑like receptor 7.

molecules TNF‑related apoptosis‑inducing ligand (TRAIL; also known as TNFSF10), death receptor 5 (DR5; also known as TNFRSF10B), CD95 (also known as FAS) and CD95 ligand (CD95L; also known as FASL)48,49. This may lead to the apoptosis of uninfected CD4+ T cells and promote generalized CD4+ T cell loss. Interestingly, HIV‑infected women are almost two times more likely to develop AIDS than HIV‑infected men with similar viral loads, and this may be because acti‑ vated pDCs from HIV‑infected women produce more type I IFNs than pDCs from HIV‑infected men50. Studies of non‑human primate models of simian immunodeficiency virus (SIV) infection support the idea that pDCs have pathological roles. Comparing SIV infection in African green monkeys (Cercopithecus aethiops), sooty mangabeys (Cercocebus atys) and macaques (Macaca sp.), it was shown that SIV infection in macaques progresses to AIDS at a rate similar to that in HIV‑1‑infected humans51. African green monkeys and sooty mangabeys, both of which are natural hosts for SIV, do not develop AIDS despite showing high levels of virus replication52. Attenuated type I IFN responses are believed to contribute to the lack of SIV‑induced immune pathology in sooty mangabeys53. Gene pro‑ filing has shown that in both non‑pathogenic models (African green monkeys and sooty mangabeys) and pathogenic models (macaques) of SIV infection there is a comparable IFN‑inducible response to acute SIV infec‑ tion, but only in pathogenic models is this IFN response maintained during chronic infection54,55.

Additional data from non‑human primate mod‑ els suggests that pDCs can facilitate HIV infection. In the vaginal mucosa of macaques, an initial founder population of T cells becomes infected with SIV but the recruitment of additional target T cells is required for viral spread. It was shown that an early influx of acti‑ vated pDCs occurs at the initial site of SIV infection and these cells subsequently recruit T cells by secreting chemokines such as CCL5 (also known as RANTeS)56. As mentioned above, HIV‑exposed pDCs can pro‑ mote the development of TReg cells, and these may pre‑ vent the induction of effector T cells by inhibiting cDC activation. Thus pDCs may promote the widespread dysregulation of DC function that occurs during HIV infection (fIG. 1). Furthermore, as DCs can promote NK cell activation, DC dysfunction may contribute to dys‑ regulated NK cell activity during HIV infection. This is discussed in more detail below.

NK cells in HIV‑1 infection NK cells and their receptors. NK cells promote anti‑ viral and antitumour immunity 57 by producing pro‑ inflammatory cytokines and by lysing infected or trans‑ formed cells. In addition, NK cells interact with T cells and DCs to shape the magnitude and quality of adaptive immune responses58,59. To date, no specific NK cell recep‑ tors that directly recognize HIV‑1‑infected cells have been identified, and the NK cell response to HIV‑1‑infected cells appears to be regulated by the balance of inhibi‑ tory and activating signals delivered to NK cells by these

180 | mARCH 2011 | VoLume 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Perforin a component of cytolytic granules that participates in the permeabilization of plasma membranes, allowing granzymes and other cytotoxic components to enter target cells.

Granzyme A a member of a family of serine proteinases that are found, primarily, in the cytoplasmic granules of cytotoxic T lymphocytes and natural killer cells. Granzymes enter target cells through perforin pores, then cleave and activate intracellular caspases to initiate target cell apoptosis.

infected cells. Below, we discuss the mechanisms by which NK cell receptors determine NK cell function in response to viral infections, with a particular focus on HIV‑1. NK cell responses are induced against cells that lack expression of the mHC class I molecules that bind NK cell inhibitory receptors60 (known as ‘missing self ’ recognition) or by cells that overexpress ligands for NK cell activating recep‑ tors. In addition to their role in controlling NK cell effec‑ tor functions, cognate interactions between inhibitory receptors and their ligands during NK cell development confer both self‑tolerance and functionality to NK cells, a process termed ‘licensing’ (bOX 1). The most frequently studied inhibitory receptors include the highly polymorphic killer cell immuno‑ globulin‑like receptors (KIRs), which are specific for clas‑ sical mHC class I molecules, and the non‑polymorphic CD94–NKG2A receptor that recognizes the non‑classical mHC molecule HLA‑e. NK cell activating receptors include NKG2D, activating members of the KIR family, the natural cytotoxicity receptors (NCRs) and the Fcγ receptor CD16. A comprehensive list of inhibitory and activating receptors expressed on NK cells is beyond the scope of this Review and is summarized elsewhere61. most NK cells in the peripheral circulation (approxi‑ mately 90%) are CD56low, constitutively produce high numbers of cytolytic granules and have the capacity to spontaneously lyse target cells in the absence of prior sensitization. CD56low NK cells express abundant levels of KIRs, C‑type lectins and NCRs, and are thought to be functionally and phenotypically mature cells. The remain‑ ing 10% of circulating NK cells are CD56hi and are often referred to as the ‘immunoregulatory’ NK cell subset as they are poorly cytotoxic, but following stimulation they secrete large amounts of pro‑inflammatory cytokines.

Box 1 | NK cell licensing Natural killer (NK) cell inhibitory receptors for self MHC class I molecules have important roles in controlling the NK cell response to potential target cells. More recently, it has been demonstrated that these receptors ensure NK cell tolerance towards self. Studies in humans and mice have shown that during development NK cells undergo an education process, in which MHC class I‑specific inhibitory receptors are involved in the calibration of NK cell effector functions97,99. Engagement of these receptors with MHC class I molecules provides a positive signal to NK cells that leads to licensing of fully competent mature peripheral NK cells. These mature NK cells can sense and lyse autologous cells that have downregulated their expression of MHC class I molecules97. NK cells that fail to undergo this education process, owing to a lack of MHC‑specific inhibitory receptor engagement, are unlicensed and have a reduced (but not absent) functional activity compared with licensed NK cells97–99. Thus, two major types of self‑tolerant NK cells exist with regard to MHC class I: licensed NK cells, which are self‑tolerant because they express inhibitory receptors for self; and unlicensed NK cells, which are also self‑tolerant as they are not functionally competent. NK cell licensing does not seem to be an ‘on–off’ switch, but rather a dynamic and quantitative process125,126. The strength of the interaction between inhibitory MHC class I‑specific receptors and their cognate MHC class I molecules appears to be proportional to the level of functional responsiveness. NK cells licensed on stronger inhibitory receptor–MHC interactions respond with increased strength and frequency than NK cells licensed on weaker inhibitory signals125. Indeed, a hierarchy of NK cell responses to MHC class I‑negative cells has been observed that is proportional to the number of different killer cell immunoglobulin‑like receptors (KIRs) for self‑HLA molecules that are expressed by the NK cell. Certain KIR and HLA alleles are also associated with more responsive NK cells100,127, suggesting that KIR–HLA combinations can differentially affect NK cell effector capacities.

CD56hi NK cells express high levels of C‑type lectins and NCRs, but have low‑level or no expression of KIRs. Although originally thought to be distinct cell lineages, it is now believed that these two NK cell subsets represent different stages of NK cell maturation, with CD56hi NK cells being less mature than CD56low NK cells62,63. In addi‑ tion, a third population of NK cells has been described, consisting of CD56– cells that accumulate during chronic viral infections64–66. These CD56– NK cells express a simi‑ lar receptor profile to CD56low NK cells, but are poorly cytotoxic and do not secrete cytokines. However, cord blood CD3–CD56– NK cells appear to be functionally competent67, suggesting differences in the differentiation status of these phenotypically similar NK cell subsets. NK cells and the antiviral response. Several mechanisms have been proposed for NK cell‑mediated recognition of HIV‑1‑infected cells (fIG. 2). NK cells may be able to detect HIV‑infected cells either directly through receptor‑ mediated interactions that have not yet been identified or indirectly following antibody‑mediated cross‑linking of CD16 (an Fc receptor for IgG). NK cell responses to HIV‑1‑derived peptides have also been observed, although whether these responses are mediated through CD16 or other NK cell receptors remains to be elucidated68,69. Thus, identification of receptors that mediate NK cell recogni‑ tion of HIV‑1 remains an important research topic. A number of studies have suggested that HIV‑1 uses specific strategies to evade NK cells. The HIV protein Nef (negative factor) is known to selectively downregu‑ late expression of HLA‑A and HLA‑B, but not HLA‑C or HLA‑e, in infected cells. This allows HIV to evade CTL responses, which are largely directed against HLA‑A‑ and HLA‑B‑restricted epitopes, and prevent the killing of HIV‑infected cells by NK cells, which express inhibi‑ tory receptors that bind to HLA‑C and/or HLA‑e70–72. Furthermore, Nef can impair NK cell activity by down‑ regulating the expression of NKG2D ligands (specifically mICA (mHC class I polypeptide‑related sequence A), uLBP1 (also known as N2DL1) and uLBP2 (also known as N2DL2)) on infected cells73. Altered NK cell phenotype and function during HIV‑1 infection. HIV‑1 infection is associated with a functional impairment of NK cells that is evident early after infec‑ tion and continues during disease progression. There is an inverse correlation between viraemia and NK cell‑ mediated suppression of HIV‑1 replication, and this may be attributed to several possible mechanisms74. First, peripheral NK cells in HIV‑1‑infected individuals have decreased intracellular stores of perforin and granzyme a, and this may account for the decreased cytotoxic capac‑ ity of NK cells75. Persistent HIV‑1 viraemia has also been shown to result in aberrant expression of several inhibitory and activating NK cell surface receptors. Lower expression of NCRs is associated with the decreased in vitro cytotox‑ icity of NK cells from HIV‑infected individuals76, and this may affect NK cell‑mediated clearance of virus‑infected cells in vivo. However, in other studies, NCR‑mediated NK cell activation has been suggested to contribute to pathology by promoting the loss of uninfected CD4+

NATuRe ReVIewS | ImmunoLogy

VoLume 11 | mARCH 2011 | 181 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS C

0-)&.

0-)& 0-EGNN

D

-+4&5

*+8KPHGEVGF %& 6EGNN

#EVKXCVKPITGEGRVQTsNKICPF KPVGTCEVKQPRTQOQVGU 0-EGNNCEVKXCVKQP

*+8KPHGEVGF %& 6EGNN

*.#FGRGPFGPVTGEQIPKVKQP D[CPCEVKXCVKPI-+4 RTQOQVGU0-EGNNCEVKXCVKQP

2GRVKFG

0-EGNN *.#$Y+

E

*.#$Y+

-+4&.

2GRVKFG 0-EGNN .QUUQH TGEQIPKVKQP

F

*+8KPHGEVGF %& 6EGNN

.QUUQHTGEQIPKVKQP D[CPKPJKDKVQT[-+4OC[ NGCFVQ0-EGNNCEVKXCVKQP

#PVKDQF[DKPFKPI VQ*+8KPHGEVGFEGNN *+8KPHGEVGF %& 6EGNN

0-EGNN

%TQUUNKPMKPIQH%& RTQOQVGUCPVKDQF[ FGRGPFGPVEGNNE[VQVQZKEKV[

%&

Figure 2 | nK cell-mediated recognition of HIV-1-infected cells. To date, it is not fully understood how natural killer (NK) cells recognize HIV‑1‑infected cells, and different mechanisms have been proposed. The expression on infected cells of NK cell activating 0CVWTG4GXKGYU^+OOWPQNQI[ receptor ligands, such as NKG2D ligands, results in the direct activation of NKG2D+ NK cells and target cell lysis (a). Changes in the epitopes presented by HLA class I molecules might allow the engagement of activating killer cell immunoglobulin‑like receptors (KIRs), and promote NK cell activation (b). Similarly, changes in HLA class I‑presented epitopes on HIV‑1‑infected cells can result in the disruption of the binding of inhibitory KIRs, leading to NK cell activation (c). Finally, antibodies binding to HIV‑1‑infected cells can cross‑link the Fcγ receptor CD16 and activate CD16+ NK cells (d).

T cells during HIV infection77. Therefore, the role of NCRs in controlling or contributing to HIV‑1 pathology is currently not understood and further investigation is required. Second, a significant proportion (15–25%) of NK cells in the peripheral circulation of HIV‑1‑infected individuals express some markers of activated NK cells (for example, HLA‑DR and CD69) but not others (for example, CD25 and NKp44 (also known as NCR2))78. This suggests that there is an incomplete activation of NK cells in these individuals, and this may be due to chronic stimulation resulting in NK cell exhaustion and anergy. In addition, although overall NK cell numbers are unchanged during HIV infection, infected individuals have fewer CD3–CD56+ NK cells and show expansion of the functionally anergic CD3–CD56– NK cell subset64,65. HIV‑1 infection has also been associated with an expansion of NK cell populations that express the acti‑ vating NKG2C receptor and decreased expression of the inhibitory NKG2A receptor (predominantly on CD56low NK cells)79. This may alter the balance of NK cell responses towards activation. NKG2C is functional in HIV‑positive individuals80; however, it is not known

how NKG2C functions in HIV infection in the context of diminished NCR‑mediated responses. An increase in cir‑ culating NKG2C+ NK cells in healthy individuals corre‑ lates with cytomegalovirus (CmV)‑seropositive status79, and it is possible that the expansion of NKG2C+ NK cell populations during HIV‑1 infection may be a result of CmV reactivation in HIV‑1‑infected individuals. Indeed, perturbation of the NK cell compartment during HIV‑1 infection may have serious consequences for protection from opportunistic infections. Finally, NK cells that express CD4 and CXCR4 can be infected with HIV‑1 in vitro, resulting in altered func‑ tion81. This suggests that CD4+ NK cells may be a reservoir for HIV‑1 in vivo, and further investigation is required to explore this possibility.

KIRs and HLA molecules in HIV infection early studies showed that the control of HIV replication is associated with the expression of certain HLA‑B alle‑ les that have the Bw4 epitope82, including HLA‑B*57 and HLA‑B*27. These molecules serve as ligands for the NK cell inhibitory receptor KIR3DL1 and as putative ligands for the NK cell activating receptor KIR3DS1 (fIG. 2). KIR3DS1 and HLA‑Bw4‑80I. martin et al.83 first dem‑ onstrated the impact of KIRs and HLA molecules on the outcome of HIV infection. The first study to dem‑ onstrate a link between expression of KIR and HLA molecules and the outcome of HIV infection showed that individuals expressing KIR3DS1 in conjunction with HLA‑Bw4‑80I alleles (HLA‑Bw4 alleles that have an isoleucine residue at position 80) have a slower pro‑ gression to AIDS than individuals that express either one or neither of these molecules. These findings were sup‑ ported by studies demonstrating that KIR3DS1+ NK cells preferentially expand during primary HIV‑1 infection in HLA‑Bw4‑80I+ subjects84 and efficiently suppress HIV‑1 replication in HLA‑Bw4‑80I+ target cells85. Furthermore, KIR3DS1 expression has been associated with enhanced NK cell function during primary HIV‑1 infection86, and the proportion of individuals homozygous for KIR3DS1 was found to be higher in a cohort of HIV‑1‑exposed but persistently uninfected individuals than in a group with primary HIV‑1 infection87. However, studies aimed at demonstrating a direct interaction between KIR3DS1 and HLA‑Bw4‑80I have not shown such interactions88,89. Two hypotheses are proposed in the literature. First, KIR3DS1 may recog‑ nize HLA‑Bw4‑80I in the presence of HLA‑presented viral peptides or stress‑induced self peptides90. Indeed HIV‑1 infection has been shown to uniquely alter the presentation of host‑encoded peptides by mHC class I molecules91. Second, recognition of HLA‑Bw4‑80I by KIR3DS1 might require an additional cellular pro‑ tein that is expressed during HIV‑1 infection. This hypothesis is supported by evidence from murine cyto‑ megalovirus (mCmV) infection, in which the NK cell activating receptor Ly49P was shown to interact with the mouse mHC molecule H2‑Dk and mediated recogni‑ tion of mCmV‑infected cells only in the presence of an additional, as‑yet‑undefined protein92,93.

182 | mARCH 2011 | VoLume 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS KIR3DL1 and HLA‑Bw4‑80I. Certain alleles of the inhibi‑ tory receptor KIR3DL1 have been shown to provide pro‑ tection against HIV‑1 disease progression94. KIR3DL1 is highly polymorphic, and different alleles result in different levels of KIR3DL1 protein expression on NK cells95,96. A genetic study on a large cohort of HIV‑1‑infected indi‑ viduals revealed that individuals co‑expressing high levels of KIR3DL1 and HLA‑Bw4‑80I had slower HIV‑1 disease progression than individuals expressing low levels of both KIR3DL1 and HLA‑Bw4‑80I94. Furthermore, KIR3DL1+ NK cell populations are expanded in HIV‑positive individuals who express HLA‑Bw4‑80I, highlighting the importance of this receptor–ligand interaction in HIV‑1 infection84. These data showing a protective effect for an inhibi‑ tory KIR in HIV infection seem to contradict previous reports, which suggested that NK cell activation is pro‑ tective during HIV infection. However, this discrepancy could be explained by the interactions of KIRs with HLA molecules during NK cell development. As described in bOX 1, developing NK cells undergo a licensing proc‑ ess, in which engagement of inhibitory KIRs by HLA molecules arms NK cells with functional capacity 97–99. The strength of interaction between inhibitory KIRs and HLA molecules at this licensing step proportion‑ ately enhances later functional capacity. Indeed, studies have shown that KIR3DL1+ NK cells from individuals homozygous for HLA‑Bw4 respond more potently to mHC class I‑negative cells than NK cells from individu‑ als expressing only one copy or no copies of HLA‑Bw4 (Ref. 100). KIR3DL1hi NK cells licensed on strong inhibi‑ tory signals would therefore respond more strongly to Nef‑mediated downregulation of HLA‑Bw4 than KIR3DL1low NK cells. Interestingly, protection against HIV‑1 disease progression was also observed in indi‑ viduals with the KIR3DL1*004 allotype of KIR3DL1 that is not expressed on the cell surface, but detected intra‑ cellularly 94. The mechanism behind this protection is unknown but suggests a role for intracellular KIR–HLA interactions in modulating antiviral immunity. NK cell licensing might also explain the protective effect of a single nucleotide polymorphism (SNP) that was recently identified in a genome‑wide association study 101. In this study, a dimorphism 35 kb upstream of the HLA‑C locus that is associated with higher levels of HLA‑C mRNA and surface expression was shown to be associated with better control of HIV‑1 viraemia101,102. It is possible that during NK cell development, the inter‑ action of NK cells expressing inhibitory KIR2D receptors with cells expressing high levels of HLA‑C (the ligand for KIR2D receptors) may promote the functional capacity of these NK cells. In summary, recent genetic and functional stud‑ ies have shown a protective effect for certain KIRs and HLA molecules during HIV‑1 disease progres‑ sion. The mechanisms underlying protection are thought to involve increased NK cell activity against HIV‑1‑infected cells; however, no data currently exist regarding the role of NK cells during HIV‑1 infection in vivo. A better understanding of the functions of NK cells during HIV‑1 infection will be required in order to

harness these important innate effector cells in vaccine‑ induced responses, particularly in the context of the recent description of NK cell‑mediated recall responses in mice103,104.

DC–NK cell crosstalk in HIV infection In addition to their own antiviral functions, NK cells can regulate antiviral immunity by modulating DC func‑ tion58,105–110. Crosstalk between NK cells and DCs results in activation of both cell types, with DCs upregulating NK cell effector functions and NK cells inducing further mat‑ uration of DCs (fIG. 3). Fernandez et al.58 initially showed that DCs can promote NK cell activity against tumours in vivo, and several subsequent studies have explored DC‑mediated ‘priming’ of NK cells4,111. Both cytokine production and cell–cell interactions have been found to be involved in this process. Cytokines produced by cDCs, such as IL‑12 and IL‑18, can promote NK cell production of IFNγ in vitro4,106,108,111, and pDC production of type I IFNs, as well as cell–cell contact, is required to promote NK cell proliferation and cytotoxicity 58,105,107,108 (fIG. 3). Activated NK cells can boost ongoing adaptive responses by producing IFNγ, which promotes TH1 cell polarization105,109. Reciprocally, NK cell‑mediated activation of DCs has been shown to promote the differen‑ tiation of DCs that are more capable of inducing efficient CTL responses110. Furthermore, NK cell‑mediated lysis of virus‑infected cells can provide a source of apoptotic bodies for uptake by DCs; this promotes DC maturation and the presentation of viral antigens to T cells. NK cells can also kill immature DCs in a process referred to as ‘DC editing’111. In vitro studies have dem‑ onstrated that a low ratio of NK cells to immature DCs promotes DC maturation, whereas a higher NK cell to immature DC ratio can result in NK cell‑mediated kill‑ ing of DCs110. This editing process is dependent on NK cell expression of NKp30 (also known as NCR3)111, and seems to be mediated by NK cell subsets that express CD94–NKG2A but lack inhibitory KIRs112, although other receptors and mechanisms may be involved. Recent evidence suggests that the crosstalk between NK cells and DCs is disrupted during HIV‑1 infection. cDCs from HIV‑1‑infected individuals show reduced secretion of IL‑12, IL‑15 and IL‑18, resulting in decreased NK cell activation. In addition, activation of NK cells by pDCs is impaired during HIV infection because NK cells are less responsive to type I IFNs113. Furthermore, the anergic CD56– NK cells that accumulate during pro‑ gressive HIV‑1 infection do not produce IFNγ and TNF following stimulation with mHC‑devoid target cells, decreasing their ability to promote DC activation64,65. NK cell‑mediated DC editing is also severely compro‑ mised during progressive HIV‑1 infection, and NK cells from individuals with chronic HIV‑1 infection show a decreased ability to kill immature DCs114. The defect appears to be largely due to an increase in the proportion of CD56– NK cells with impaired NKp30 function115. The precise mechanisms by which HIV‑1 impairs NK cell and DC crosstalk remain to be fully elucidated. effects of HIV‑expressed Nef and Tat (transactivator of transcrip‑ tion) on DCs and NK cells have been described116,117.

NATuRe ReVIewS | ImmunoLogy

VoLume 11 | mARCH 2011 | 183 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS #EVKXCVGFE&%URTQFWEG RTQKPȯCOOCVQT[E[VQMKPGU VJCVUVKOWNCVG0-EGNNU

R&%URTQFWEGV[RG++(0UCPF RTQOQVG0-EGNNRTQNKHGTCVKQP CPFE[VQVQZKEKV[ 6[RG++(0U

+.+.+.

E&%

E&%s0-EGNNETQUUVCNM KUCVVGPWCVGFFWTKPI *+8KPHGEVKQP

0-EGNN

&WTKPI*+8KPHGEVKQP 0-EGNNUCTGNGUU TGURQPUKXGVQV[RG++(0U

R&%

+(0γ 0-EGNNURTQOQVG CFCRVKXGKOOWPKV[D[ MKNNKPIKOOCVWTG&%U

#EVKXCVGF0-EGNNURTQFWEG +(0γCPFRTQOQVG6*V[RG KOOWPKV[

&WTKPI*+8KPHGEVKQP 0-EGNNUJCXGCFGETGCUGF CDKNKV[VQMKNNKOOCVWTG&%U

+OOCVWTG&%

Figure 3 | DC–nK cell crosstalk. Dendritic cells (DCs) are activated by HIV‑1 and secrete pro‑inflammatory cytokines, including interleukin‑12 (IL‑12), IL‑15 and type I interferons (IFNs), that stimulate natural killer (NK) cells. Activated NK cells 0CVWTG4GXKGYU^+OOWPQNQI[ secrete IFNγ, which promotes DC maturation and T helper 1 (TH1)‑type immune responses. Furthermore, NK cells can eliminate immature DCs, and this editing process promotes the induction of adaptive T cell immunity. HIV‑1‑induced functional impairment of NK cells, as well as of DCs, interferes with this crosstalk. cDC, conventional DC; pDC, plasmacytoid DC.

HIV‑1‑infected DCs become resistant to NK cell‑ mediated lysis, and this is associated with the upregulation of cell death inhibitors in infected DCs. These inhibitors are upregulated by the high‑mobility group box 1 pro‑ tein (HmGB1) and protect HIV‑1‑infected DCs from TRAIL‑dependent apoptosis118. Furthermore, crosstalk with NK cells was found to promote viral replication in HIV‑1‑infected DCs, but this could be prevented by blocking HmGB1 activity 119. In addition, increased pro‑ duction of IL‑10 during HIV‑1 infection can protect immature DCs from NK cell‑mediated lysis, resulting in accumulation of partially mature, poorly immunogenic DCs in the lymph nodes of infected individuals120. overall, the dysregulated crosstalk that occurs between NK cells and DCs during progressive HIV‑1 infection appears to be a consequence of the impairment of both DC and NK cell functions (fIG. 3). Given the crucial role of DCs in determining the quality of the adaptive immune response to infections, this compromised editing of DC function by NK cells during HIV‑1 infection might have significant consequences for the antiviral T and B cell responses. In order to develop interventions aimed at enhancing immunity to HIV‑1, further research is required to better understand the molecular mechanisms involved in NK cell–DC crosstalk and how this crosstalk is disrupted during infection with HIV‑1.

Harnessing DCs and NK cells for vaccination Although HIV‑1 infection is considered to be a chronic condition, immune dysregulation can occur early in infec‑ tion and affect subsequent disease progression. Given the complexity of HIV infection, it remains unclear how the innate immune system can be harnessed to induce

effective T and B cell‑mediated protective immunity. understanding how DCs and NK cells are affected during HIV infection may provide new targets for vaccine design or even therapeutic modulation of disease. Administration of myeloid DCs that have been pre‑treated with inactivated HIV enhances immune control of HIV in patients121, sug‑ gesting that functionally intact antigen‑presenting cells are required to limit viral replication. Adjuvants and vaccine vectors that target cDCs and pDCs simultaneously could promote adaptive immunity and limit TReg cell induction in order to control virus entry at mucosal sites, as well as systemically 122,123. For example, the approach used in a recent vaccine trial in Thailand (which used an HIV envelope immunogen and a canarypox virus–HIV vector) could be improved by using DC‑targeted adjuvants and superior vectors in a ‘prime–boost strategy’. Given the speed with which HIV gains entry into cells, innate defences need to be rapidly mobilized. DCs could be targeted to activate specific NK cells and promote their cytolytic functions. epidemiological data have indi‑ cated that certain NK receptors, such as KIR3DS1, are important for controlling HIV disease83, and in mouse models, NK cells can develop into protective virus‑spe‑ cific memory cells103,124. Therefore, the development of DC‑based vaccine strategies that elicit HIV‑specific NK cell responses and stimulate the production of memory cells may be crucial for the success of future vaccines. Finally, pDCs are emerging as a population with impor‑ tant roles in contributing to HIV‑induced pathology. As such, vaccine strategies that aim to promote cDC and NK cell responses during HIV infection would have to be balanced to prevent any deleterious consequences of immune activation.

184 | mARCH 2011 | VoLume 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 1.

2. 3. 4.

5. 6.

7.

8.

9. 10.

11.

12.

13.

14. 15.

16.

17.

18.

19. 20.

21.

22.

McMichael, A. J., Borrow, P., Tomaras, G. D., Goonetilleke, N. & Haynes, B. F. The immune response during acute HIV‑1 infection: clues for vaccine development. Nature Rev. Immunol. 10, 11–23 (2010). Liu, Y. J. Dendritic cell subsets and lineages, and their functions in innate and adaptive immunity. Cell 106, 259–262 (2001). Rescigno, M. & Borrow, P. The host‑pathogen interaction: new themes from dendritic cell biology. Cell 106, 267–270 (2001). Ferlazzo, G. et al. Distinct roles of IL‑12 and IL‑15 in human natural killer cell activation by dendritic cells from secondary lymphoid organs. Proc. Natl Acad. Sci. USA 101, 16606–16611 (2004). Lanzavecchia, A. & Sallusto, F. Regulation of T cell immunity by dendritic cells. Cell 106, 263–266 (2001). Romagnani, C. et al. Activation of human NK cells by plasmacytoid dendritic cells and its modulation by CD4+ T helper cells and CD4+CD25hi T regulatory cells. Eur. J. Immunol. 35, 2452–2458 (2005). Wu, L. & KewalRamani, V. N. Dendritic‑cell interactions with HIV: infection and viral dissemination. Nature Rev. Immunol. 6, 859–868 (2006). Ignatius, R. et al. The immunodeficiency virus coreceptor, Bonzo/STRL33/TYMSTR, is expressed by macaque and human skin‑ and blood‑derived dendritic cells. AIDS Res. Hum. Retroviruses 16, 1055–1059 (2000). Geijtenbeek, T. B. et al. DC‑SIGN, a dendritic cell‑ specific HIV‑1‑binding protein that enhances trans‑ infection of T cells. Cell 100, 587–597 (2000). Schmidt, B., Ashlock, B. M., Foster, H., Fujimura, S. H. & Levy, J. A. HIV‑infected cells are major inducers of plasmacytoid dendritic cell interferon production, maturation, and migration. Virology 343, 256–266 (2005). Boggiano, C., Manel, N. & Littman, D. R. Dendritic cell‑mediated trans‑enhancement of human immunodeficiency virus type 1 infectivity is independent of DC‑SIGN. J. Virol. 81, 2519–2523 (2007). Sabatte, J. et al. Human seminal plasma abrogates the capture and transmission of human immunodeficiency virus type 1 to CD4+ T cells mediated by DC‑SIGN. J. Virol. 81, 13723–13734 (2007). Lambert, A. A., Gilbert, C., Richard, M., Beaulieu, A. D. & Tremblay, M. J. The C‑type lectin surface receptor DCIR acts as a new attachment factor for HIV‑1 in dendritic cells and contributes to trans‑ and cis‑ infection pathways. Blood 112, 1299–1307 (2008). de Witte, L. et al. Langerin is a natural barrier to HIV‑1 transmission by Langerhans cells. Nature Med. 13, 367–371 (2007). Fahrbach, K. M. et al. Activated CD34‑derived Langerhans cells mediate transinfection with human immunodeficiency virus. J. Virol. 81, 6858–6868 (2007). Kadowaki, N. et al. Subsets of human dendritic cell precursors express different Toll‑like receptors and respond to different microbial antigens. J. Exp. Med. 194, 863–869 (2001). Beignon, A. S. et al. Endocytosis of HIV‑1 activates plasmacytoid dendritic cells via Toll‑like receptor–viral RNA interactions. J. Clin. Invest. 115, 3265–3275 (2005). This study showed how HIV stimulates pDCs for type I IFN production through TLR7 signalling. Meier, A. et al. MyD88‑dependent immune activation mediated by human immunodeficiency virus type 1‑encoded Toll‑like receptor ligands. J. Virol. 81, 8180–8191 (2007). Honda, K. et al. IRF‑7 is the master regulator of type‑I interferon‑dependent immune responses. Nature 434, 772–777 (2005). Sabado, R. L. et al. Evidence of dysregulation of dendritic cells in primary HIV infection. Blood 116, 3839–3852 (2010). This study examined the changes that occur in cDC and pDC frequency and function over the course of early HIV infection. Smed‑Sorensen, A. et al. Differential susceptibility to human immunodeficiency virus type 1 infection of myeloid and plasmacytoid dendritic cells. J. Virol. 79, 8861–8869 (2005). Manel, N. et al. A cryptic sensor for HIV‑1 activates antiviral innate immunity in dendritic cells. Nature 467, 214–217 (2010).

23. Geijtenbeek, T. B. et al. Mycobacteria target DC‑SIGN to suppress dendritic cell function. J. Exp. Med. 197, 7–17 (2003). 24. Gringhuis, S. I. et al. HIV‑1 exploits innate signaling by TLR8 and DC‑SIGN for productive infection of dendritic cells. Nature Immunol. 11, 419–426 (2010). This study showed that productive HIV infection in cDCs is induced by TLR8 signalling, which initiates viral RNA transcription, and DC‑SIGN signalling, which promotes transcript elongation. 25. Deretic, V. Multiple regulatory and effector roles of autophagy in immunity. Curr. Opin. Immunol. 21, 53–62 (2009). 26. Blanchet, F. P. et al. Human immunodeficiency virus‑1 inhibition of immunoamphisomes in dendritic cells impairs early innate and adaptive immune responses. Immunity 32, 654–669 (2010). 27. Granelli‑Piperno, A., Shimeliovich, I., Pack, M., Trumpfheller, C. & Steinman, R. M. HIV‑1 selectively infects a subset of nonmaturing BDCA1‑positive dendritic cells in human blood. J. Immunol. 176, 991–998 (2006). 28. Granelli‑Piperno, A., Golebiowska, A., Trumpfheller, C., Siegal, F. P. & Steinman, R. M. HIV‑1‑infected monocyte‑derived dendritic cells do not undergo maturation but can elicit IL‑10 production and T cell regulation. Proc. Natl Acad. Sci. USA 101, 7669–7674 (2004). 29. Harman, A. N. et al. HIV‑1‑infected dendritic cells show 2 phases of gene expression changes, with lysosomal enzyme activity decreased during the second phase. Blood 114, 85–94 (2009). 30. Poulin, L. F. et al. Characterization of human DNGR‑1+ BDCA3+ leukocytes as putative equivalents of mouse CD8α+ dendritic cells. J. Exp. Med. 207, 1261–1271 (2010). 31. Donaghy, H., Gazzard, B., Gotch, F. & Patterson, S. Dysfunction and infection of freshly isolated blood myeloid and plasmacytoid dendritic cells in patients infected with HIV‑1. Blood 101, 4505–4511 (2003). 32. Cameron, P. U. et al. Preferential infection of dendritic cells during human immunodeficiency virus type 1 infection of blood leukocytes. J. Virol. 81, 2297–2306 (2007). 33. Kodama, A. et al. Impairment of in vitro generation of monocyte‑derived human dendritic cells by inactivated human immunodeficiency virus‑1: involvement of type I interferon produced from plasmacytoid dendritc cells. Hum. Immunol. 71, 541–550 (2010). 34. Meera, S., Madhuri, T., Manisha, G. & Ramesh, P. Irreversible loss of pDCs by apoptosis during early HIV infection may be a critical determinant of immune dysfunction. Viral Immunol. 23, 241–249 (2010). 35. Dillon, S. M. et al. Plasmacytoid and myeloid dendritic cells with a partial activation phenotype accumulate in lymphoid tissue during asymptomatic chronic HIV‑1 infection. J. Acquir. Immune Defic. Syndr. 48, 1–12 (2008). 36. Lehmann, C. et al. Plasmacytoid dendritic cells accumulate and secrete interferon alpha in lymph nodes of HIV‑1 patients. PLoS ONE 5, e11110 (2010). 37. Nascimbeni, M. et al. Plasmacytoid dendritic cells accumulate in spleens from chronically HIV‑infected patients but barely participate in interferon‑α expression. Blood 113, 6112–6119 (2009). 38. Finke, J. S., Shodell, M., Shah, K., Siegal, F. P. & Steinman, R. M. Dendritic cell numbers in the blood of HIV‑1 infected patients before and after changes in antiretroviral therapy. J. Clin. Immunol. 24, 647–652 (2004). 39. Killian, M. S., Fujimura, S. H., Hecht, F. M. & Levy, J. A. Similar changes in plasmacytoid dendritic cell and CD4 T‑cell counts during primary HIV‑1 infection and treatment. AIDS 20, 1247–1252 (2006). 40. Martinson, J. A. et al. Dendritic cells from HIV‑1 infected individuals are less responsive to Toll‑like receptor (TLR) ligands. Cell. Immunol. 250, 75–84 (2007). 41. Said, E. A. et al. Programmed death‑1‑induced interleukin‑10 production by monocytes impairs CD4+ T cell activation during HIV infection. Nature Med. 16, 452–459 (2010). 42. Brenchley, J. M. et al. Microbial translocation is a cause of systemic immune activation in chronic HIV infection. Nature Med. 12, 1365–1371 (2006). 43. Kamga, I. et al. Type I interferon production is profoundly and transiently impaired in primary HIV‑1 infection. J. Infect. Dis. 192, 303–310 (2005).

NATuRe ReVIewS | ImmunoLogy

44. Sachdeva, N., Asthana, V., Brewer, T. H., Garcia, D. & Asthana, D. Impaired restoration of plasmacytoid dendritic cells in HIV‑1‑infected patients with poor CD4 T cell reconstitution is associated with decrease in capacity to produce IFN‑α but not proinflammatory cytokines. J. Immunol. 181, 2887–2897 (2008). 45. Lehmann, C. et al. Increased interferon alpha expression in circulating plasmacytoid dendritic cells of HIV‑1‑infected patients. J. Acquir. Immune Defic. Syndr. 48, 522–530 (2008). 46. Manches, O. et al. HIV‑activated human plasmacytoid DCs induce Tregs through an indoleamine 2,3‑dioxygenase‑dependent mechanism. J. Clin. Invest. 118, 3431–3439 (2008). This paper showed that HIV‑stimulated pDCs induce TReg cell differentiation through upregulation of IDO. 47. Favre, D. et al. Tryptophan catabolism by indoleamine 2,3‑dioxygenase 1 alters the balance of TH17 to regulatory T cells in HIV disease. Sci. Transl. Med. 2, 32ra36 (2010). 48. Herbeuval, J. P. et al. Regulation of TNF‑related apoptosis‑inducing ligand on primary CD4+ T cells by HIV‑1: role of type I IFN‑producing plasmacytoid dendritic cells. Proc. Natl Acad. Sci. USA 102, 13974–13979 (2005). 49. Herbeuval, J. P. et al. Differential expression of IFN‑α and TRAIL/DR5 in lymphoid tissue of progressor versus nonprogressor HIV‑1‑infected patients. Proc. Natl Acad. Sci. USA 103, 7000–7005 (2006). 50. Meier, A. et al. Sex differences in the Toll‑like receptor‑ mediated response of plasmacytoid dendritic cells to HIV‑1. Nature Med. 15, 955–959 (2009). 51. McMichael, A. J. HIV vaccines. Annu. Rev. Immunol. 24, 227–255 (2006). 52. Manches, O. & Bhardwaj, N. Resolution of immune activation defines nonpathogenic SIV infection. J. Clin. Invest. 119, 3512–3515 (2009). 53. Mandl, J. N. et al. Divergent TLR7 and TLR9 signaling and type I interferon production distinguish pathogenic and nonpathogenic AIDS virus infections. Nature Med. 14, 1077–1087 (2008). 54. Bosinger, S. E. et al. Global genomic analysis reveals rapid control of a robust innate response in SIV‑ infected sooty mangabeys. J. Clin. Invest. 119, 3556–3572 (2009). 55. Jacquelin, B. et al. Nonpathogenic SIV infection of African green monkeys induces a strong but rapidly controlled type I IFN response. J. Clin. Invest. 119, 3544–3555 (2009). 56. Li, Q. et al. Glycerol monolaurate prevents mucosal SIV transmission. Nature 458, 1034–1038 (2009). 57. Biron, C. A., Nguyen, K. B., Pien, G. C., Cousens, L. P. & Salazar‑Mather, T. P. Natural killer cells in antiviral defense: function and regulation by innate cytokines. Annu. Rev. Immunol. 17, 189–220 (1999). 58. Fernandez, N. C. et al. Dendritic cells directly trigger NK cell functions: cross‑talk relevant in innate anti‑tumor immune responses in vivo. Nature Med. 5, 405–411 (1999). 59. Mailliard, R. B. et al. Dendritic cells mediate NK cell help for Th1 and CTL responses: two‑signal requirement for the induction of NK cell helper function. J. Immunol. 171, 2366–2373 (2003). 60. Karre, K., Ljunggren, H. G., Piontek, G. & Kiessling, R. Selective rejection of H–2‑deficient lymphoma variants suggests alternative immune defence strategy. Nature 319, 675–678 (1986). 61. Lanier, L. L. Up on the tightrope: natural killer cell activation and inhibition. Nature Immunol. 9, 495–502 (2008). 62. Romagnani, C. et al. CD56brightCD16– killer Ig‑like receptor– NK cells display longer telomeres and acquire features of CD56dim NK cells upon activation. J. Immunol. 178, 4947–4955 (2007). 63. Yu, J. et al. CD94 surface density identifies a functional intermediary between the CD56bright and CD56dim human NK‑cell subsets. Blood 115, 274– 281 (2010). 64. Mavilio, D. et al. Characterization of CD56–/CD16+ natural killer (NK) cells: a highly dysfunctional NK subset expanded in HIV‑infected viremic individuals. Proc. Natl Acad. Sci. USA 102, 2886–2891 (2005). 65. Alter, G. et al. Sequential deregulation of NK cell subset distribution and function starting in acute HIV‑1 infection. Blood 106, 3366–3369 (2005). 66. Gonzalez, V. D. et al. Expansion of functionally skewed CD56‑negative NK cells in chronic hepatitis C virus infection: correlation with outcome of pegylated IFN‑α and ribavirin treatment. J. Immunol. 183, 6612–6618 (2009).

VoLume 11 | mARCH 2011 | 185 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 67. Lu, X. et al. CD16+ CD56– NK cells in the peripheral blood of cord blood transplant recipients: a unique subset of NK cells possibly associated with graft‑versus‑ leukemia effect. Eur. J. Haematol. 81, 18–25 (2008). 68. Stratov, I., Chung, A. & Kent, S. J. Robust NK cell‑ mediated human immunodeficiency virus (HIV)‑ specific antibody‑dependent responses in HIV‑infected subjects. J. Virol. 82, 5450–5459 (2008). 69. Tiemessen, C. T. et al. Cutting edge: unusual NK cell responses to HIV‑1 peptides are associated with protection against maternal‑infant transmission of HIV‑1. J. Immunol. 182, 5914–5918 (2009). 70. Collins, K. L., Chen, B. K., Kalams, S. A., Walker, B. D. & Baltimore, D. HIV‑1 Nef protein protects infected primary cells against killing by cytotoxic T lymphocytes. Nature 391, 397–401 (1998). 71. Cohen, G. B. et al. The selective downregulation of class I major histocompatibility complex proteins by HIV‑1 protects HIV‑infected cells from NK cells. Immunity 10, 661–671 (1999). 72. Bonaparte, M. I. & Barker, E. Killing of human immunodeficiency virus‑infected primary T‑cell blasts by autologous natural killer cells is dependent on the ability of the virus to alter the expression of major histocompatibility complex class I molecules. Blood 104, 2087–2094 (2004). 73. Cerboni, C. et al. Human immunodeficiency virus 1 Nef protein downmodulates the ligands of the activating receptor NKG2D and inhibits natural killer cell‑mediated cytotoxicity. J. Gen. Virol. 88, 242–250 (2007). 74. Kottilil, S. et al. Innate immunity in human immunodeficiency virus infection: effect of viremia on natural killer cell function. J. Infect. Dis. 187, 1038–1045 (2003). 75. Portales, P. et al. Interferon‑α restores HIV‑induced alteration of natural killer cell perforin expression in vivo. AIDS 17, 495–504 (2003). 76. De Maria, A. et al. The impaired NK cell cytolytic function in viremic HIV‑1 infection is associated with a reduced surface expression of natural cytotoxicity receptors (NKp46, NKp30 and NKp44). Eur. J. Immunol. 33, 2410–2418 (2003). 77. Vieillard, V., Strominger, J. L. & Debre, P. NK cytotoxicity against CD4+ T cells during HIV‑1 infection: a gp41 peptide induces the expression of an NKp44 ligand. Proc. Natl Acad. Sci. USA 102, 10981–10986 (2005). 78. Fogli, M. et al. Significant NK cell activation associated with decreased cytolytic function in peripheral blood of HIV‑1‑infected patients. Eur. J. Immunol. 34, 2313–2321 (2004). 79. Mela, C. M. et al. Switch from inhibitory to activating NKG2 receptor expression in HIV‑1 infection: lack of reversion with highly active antiretroviral therapy. AIDS 19, 1761–1769 (2005). 80. Guma, M. et al. Imprint of human cytomegalovirus infection on the NK cell receptor repertoire. Blood 104, 3664–3671 (2004). 81. Bernstein, H. B. et al. CD4+ NK cells can be productively infected with HIV, leading to downregulation of CD4 expression and changes in function. Virology 387, 59–66 (2009). 82. Flores‑Villanueva, P. O. et al. Control of HIV‑1 viremia and protection from AIDS are associated with HLA‑Bw4 homozygosity. Proc. Natl Acad. Sci. USA 98, 5140–5145 (2001). 83. Martin, M. P. et al. Epistatic interaction between KIR3DS1 and HLA‑B delays the progression to AIDS. Nature Genet. 31, 429–434 (2002). This study first demonstrated the protective effect of KIR–HLA combinations in HIV‑1 disease progression and showed that individuals expressing KIR3DS1 in conjunction with its putative ligand (HLA‑Bw4‑80I) have a slower progression to AIDS compared with individuals lacking either or both receptor–ligand pairs. 84. Alter, G. et al. HLA class I subtype‑dependent expansion of KIR3DS1+ and KIR3DL1+ NK cells during acute human immunodeficiency virus type 1 infection. J. Virol. 83, 6798–6805 (2009). 85. Alter, G. et al. Differential natural killer cell‑mediated inhibition of HIV‑1 replication based on distinct KIR/ HLA subtypes. J. Exp. Med. 204, 3027–3036 (2007). 86. Long, B. R. et al. Conferral of enhanced natural killer cell function by KIR3DS1 in early human immunodeficiency virus type 1 infection. J. Virol. 82, 4785–4792 (2008). 87. Boulet, S. et al. Increased proportion of KIR3DS1 homozygotes in HIV‑exposed uninfected individuals. AIDS 22, 595–599 (2008).

88. Gillespie, G. M. et al. Lack of KIR3DS1 binding to MHC class I Bw4 tetramers in complex with CD8+ T cell epitopes. AIDS Res. Hum. Retroviruses 23, 451–455 (2007). 89. Carr, W. H. et al. Cutting edge: KIR3DS1, a gene implicated in resistance to progression to AIDS, encodes a DAP12‑associated receptor expressed on NK cells that triggers NK cell activation. J. Immunol. 178, 647–651 (2007). 90. Altfeld, M. & Goulder, P. ‘Unleashed’ natural killers hinder HIV. Nature Genet. 39, 708–710 (2007). 91. Hickman, H. D. et al. Cutting edge: class I presentation of host peptides following HIV infection. J. Immunol. 171, 22–26 (2003). 92. Lee, S. H. et al. Susceptibility to mouse cytomegalovirus is associated with deletion of an activating natural killer cell receptor of the C‑type lectin superfamily. Nature Genet. 28, 42–45 (2001). 93. Kielczewska, A. et al. Ly49P recognition of cytomegalovirus‑infected cells expressing H2‑Dk and CMV‑encoded m04 correlates with the NK cell antiviral response. J. Exp. Med. 206, 515–523 (2009). 94. Martin, M. P. et al. Innate partnership of HLA‑B and KIR3DL1 subtypes against HIV‑1. Nature Genet. 39, 733–740 (2007). This large genetic study showed that individuals with a KIR3DL1hiHLA‑Bw4‑80I+ phenotype had slower HIV‑1 disease progression than individuals with a KIR3DL1lowHLA‑Bw4‑80I+ phenotype, indicating that KIR–HLA interactions can enhance the protective effect of HLA‑Bw4 alleles. 95. Yawata, M. et al. Roles for HLA and KIR polymorphisms in natural killer cell repertoire selection and modulation of effector function. J. Exp. Med. 203, 633–645 (2006). 96. Thomas, R. et al. Novel KIR3DL1 alleles and their expression levels on NK cells: convergent evolution of KIR3DL1 phenotype variation? J. Immunol. 180, 6743–6750 (2008). 97. Kim, S. et al. Licensing of natural killer cells by host major histocompatibility complex class I molecules. Nature 436, 709–713 (2005). 98. Fernandez, N. C. et al. A subset of natural killer cells achieves self‑tolerance without expressing inhibitory receptors specific for self‑MHC molecules. Blood 105, 4416–4423 (2005). 99. Anfossi, N. et al. Human NK cell education by inhibitory receptors for MHC class I. Immunity 25, 331–342 (2006). 100. Kim, S. et al. HLA alleles determine differences in human natural killer cell responsiveness and potency. Proc. Natl Acad. Sci. USA 105, 3053–3058 (2008). 101. Fellay, J. et al. A whole‑genome association study of major determinants for host control of HIV‑1. Science 317, 944–947 (2007). This genome‑wide association study showed that a protective SNP associated with higher HLA‑C expression correlates with better control of HIV‑1 viraemia. 102. Thomas, R. et al. HLA‑C cell surface expression and control of HIV/AIDS correlate with a variant upstream of HLA‑C. Nature Genet. 41, 1290–1294 (2009). 103. Sun, J. C., Beilke, J. N. & Lanier, L. L. Adaptive immune features of natural killer cells. Nature 457, 557–561 (2009). 104. O’Leary, J. G., Goodarzi, M., Drayton, D. L. & von Andrian, U. H. T cell‑ and B cell‑independent adaptive immunity mediated by natural killer cells. Nature Immunol. 7, 507–516 (2006). 105. Kikuchi, T. et al. Dendritic cells pulsed with live and dead Legionella pneumophila elicit distinct immune responses. J. Immunol. 172, 1727–1734 (2004). 106. Gerosa, F. et al. Reciprocal activating interaction between natural killer cells and dendritic cells. J. Exp. Med. 195, 327–333 (2002). 107. Gerosa, F. et al. The reciprocal interaction of NK cells with plasmacytoid or myeloid dendritic cells profoundly affects innate resistance functions. J. Immunol. 174, 727–734 (2005). 108. Borg, C. et al. NK cell activation by dendritic cells (DCs) requires the formation of a synapse leading to IL‑12 polarization in DCs. Blood 104, 3267–3275 (2004). 109. Martin‑Fontecha, A. et al. Induced recruitment of NK cells to lymph nodes provides IFN‑γ for TH1 priming. Nature Immunol. 5, 1260–1265 (2004). 110. Piccioli, D., Sbrana, S., Melandri, E. & Valiante, N. M. Contact‑dependent stimulation and inhibition of

186 | mARCH 2011 | VoLume 11

dendritic cells by natural killer cells. J. Exp. Med. 195, 335–341 (2002). 111. Ferlazzo, G. et al. Human dendritic cells activate resting natural killer (NK) cells and are recognized via the NKp30 receptor by activated NK cells. J. Exp. Med. 195, 343–351 (2002). 112. Della Chiesa, M. et al. The natural killer cell‑mediated killing of autologous dendritic cells is confined to a cell subset expressing CD94/NKG2A, but lacking inhibitory killer Ig‑like receptors. Eur. J. Immunol. 33, 1657–1666 (2003). 113. Reitano, K. N. et al. Defective plasmacytoid dendritic cell–NK cell cross‑talk in HIV infection. AIDS Res. Hum. Retroviruses 25, 1029–1037 (2009). 114. Tasca, S. et al. Escape of monocyte‑derived dendritic cells of HIV‑1 infected individuals from natural killer cell‑mediated lysis. AIDS 17, 2291–2298 (2003). 115. Mavilio, D. et al. Characterization of the defective interaction between a subset of natural killer cells and dendritic cells in HIV‑1 infection. J. Exp. Med. 203, 2339–2350 (2006). This paper shows that NK cell‑mediated DC editing is severely impaired in progressive HIV‑1 infection, and this appears to be due to an increased accumulation of CD56– NK cells with impaired NKp30 function. 116. Quaranta, M. G. et al. HIV‑1 Nef impairs the dynamic of DC/NK crosstalk: different outcome of CD56dim and CD56bright NK cell subsets. FASEB J. 21, 2323–2334 (2007). 117. Poggi, A. et al. NK cell activation by dendritic cells is dependent on LFA‑1‑mediated induction of calcium‑ calmodulin kinase II: inhibition by HIV‑1 Tat C‑terminal domain. J. Immunol. 168, 95–101 (2002). 118. Melki, M. T., Saidi, H., Dufour, A., Olivo‑Marin, J. C. & Gougeon, M. L. Escape of HIV‑1‑infected dendritic cells from TRAIL‑mediated NK cell cytotoxicity during NK‑DC cross‑talk—a pivotal role of HMGB1. PLoS Pathog. 6, e1000862 (2010). 119. Saidi, H., Melki, M. T. & Gougeon, M. L. HMGB1‑dependent triggering of HIV‑1 replication and persistence in dendritic cells as a consequence of NK‑DC cross‑talk. PLoS ONE 3, e3601 (2008). 120. Alter, G. et al. IL‑10 induces aberrant deletion of dendritic cells by natural killer cells in the context of HIV infection. J. Clin. Invest. 120, 1905–1913 (2010). This work showed that IL‑10 secreted in response to HIV‑1 renders immature DCs resistant to NK cell‑mediated killing, resulting in the accumulation of poorly immunogenic DCs in the lymph nodes of infected patients. 121. Lu, W., Arraes, L. C., Ferreira, W. T. & Andrieu, J. M. Therapeutic dendritic‑cell vaccine for chronic HIV‑1 infection. Nature Med. 10, 1359–1365 (2004). 122. Hansen, S. G. et al. Effector memory T cell responses are associated with protection of rhesus monkeys from mucosal simian immunodeficiency virus challenge. Nature Med. 15, 293–299 (2009). 123. Wyand, M. S. et al. Protection by live, attenuated simian immunodeficiency virus against heterologous challenge. J. Virol. 73, 8356–8363 (1999). 124. Paust, S. et al. Critical role for the chemokine receptor CXCR6 in NK cell‑mediated antigen‑specific memory of haptens and viruses. Nature Immunol. 11, 1127–1135 (2010). 125. Brodin, P., Lakshmikanth, T., Johansson, S., Karre, K. & Hoglund, P. The strength of inhibitory input during education quantitatively tunes the functional responsiveness of individual natural killer cells. Blood 113, 2434–2441 (2009). 126. Yu, J. et al. Hierarchy of the human natural killer cell response is determined by class and quantity of inhibitory receptors for self‑HLA‑B and HLA‑C ligands. J. Immunol. 179, 5977–5989 (2007). 127. Yawata, M. et al. MHC class I‑specific inhibitory receptors and their ligands structure diverse human NK‑cell repertoires toward a balance of missing self‑ response. Blood 112, 2369–2380 (2008).

Acknowledgements

This work was supported by grants from the US National Institutes of Health (to N.B. and M.A.), the Doris Duke Charitable Foundation (to M.A.), the Bill & Melinda Gates Foundation (to N.B. and M.A.) and the Phillip T. and Susan M. Ragon Foundation (to M.A. and L.F.).

Competing interests statement

The authors declare competing financial interests: see Web version for details.

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

Microbial manipulation of receptor crosstalk in innate immunity George Hajishengallis* and John D. Lambris‡

Abstract | In the arms race of host–microbe co-evolution, successful microbial pathogens have evolved ingenious ways to evade host immune responses. In this Review, we focus on ‘crosstalk manipulation’ — the microbial strategies that instigate, subvert or disrupt the molecular signalling crosstalk between receptors of the innate immune system. This proactive interference undermines host defences and contributes to microbial adaptive fitness and persistent infections. Understanding how pathogens exploit host receptor crosstalk mechanisms and infiltrate the host signalling network is essential for developing interventions to redirect the host response and achieve protective immunity. Pattern recognition receptor (PRR). A host receptor that can sense pathogen-associated molecular patterns and initiate signalling cascades that lead to an innate immune response. PRRs can be membrane bound (such as Toll-like receptors) or soluble cytoplasmic receptors (such as NOD-like receptors).

*University of Louisville Department of Microbiology and Immunology, Oral Health and Systemic Disease Research Group, 501 South Preston Street, Louisville, Kentucky 40292, USA. ‡ University of Pennsylvania School of Medicine, Department of Pathology and Laboratory Medicine, 422 Curie Boulevard, Philadelphia, Pennsylvania 19104, USA. e-mails: [email protected]; [email protected] doi:10.1038/nri2918

Front-line defence cells, such as neutrophils, macrophages and dendritic cells (DCs), detect invading pathogens through germline-encoded pattern recognition receptors (PRRs). Soluble and membrane-bound PRRs alert the mammalian immune system through both extracellular and intracellular activation cascades, such as the complement and Toll-like receptor (TLR) pathways, respectively. The aim is to elicit innate antimicrobial and inflammatory responses and initiate adaptive immunity for the control or elimination of infection1,2. PRRs recognize relatively invariant microbial structures, often referred to as pathogen-associated molecular patterns (PAMPs), which are shared by related groups of microorganisms3. Different PRRs generally recognize distinct PAMPs, a concept that is best illustrated by the diverse ligand specificities of TLRs, which are the prototypical and best characterized PRR family 1,4. The broad but distinct specificities of the PRRs and their ability to form functional multi-receptor complexes in lipid rafts5,6 allow for the generation of large combinatorial repertoires. This further diversifies the recognition and signalling capacities of cooperating PRRs and, at least in principle, enables the host to detect almost any type of infection, discriminate between different pathogens and mount a context-relevant immune response. Sentinel cells receive a variety of ‘input messages’ from their environment, including those communicated by pathogen-sensing PRRs. The cell needs to appropriately process and integrate this information, which is relayed intracellularly through nonlinear signalling cascades. A systematic analysis of intracellular signalling crosstalk has shown that a large number of pathways converge on a relatively limited set of interaction mechanisms, which include both synergistic and antagonistic

interactions7. Synergistic pathways greatly increase the sensitivity of detection, in that several individually weak stimuli can combine to elicit a vigorous cellular response. Antagonistic pathways increase the specificity of the host response by restraining it and preventing collateral tissue damage. Signalling crosstalk is therefore important for the normal function of the immune system; it can synergistically activate host defences to clear infections or it can antagonistically dampen unwarranted host responses2,8–11. Two characteristic examples are the cooperation of TLR2 with the C-type lectin dectin 1 (also known as CLEC7A) to stimulate antifungal immunity 12 and the homeostatic suppression of TLR-induced pro-inflammatory responses by the glucocorticoid and adenosine receptors13,14. Briefly stated, coordinated signalling crosstalk can help to maintain a fine balance between protective immunity and inflammatory pathology. However, chronic infections and disease can ensue when bacterial, viral or eukaryotic parasitic pathogens successfully evade, neutralize or subvert immune detection, signal transduction or effector killing functions15–17. Microbial pathogens that disable host defences preferentially target the innate immune system16. In part, this is because the innate defences of the host are the first to be encountered by pathogens. In addition, by subverting innate immunity, pathogens can undermine the overall host defence system, given the instructive role of innate immune mechanisms in the development of the adaptive immune response3. One way in which pathogens could undermine host immunity to promote their adaptive fitness is through the manipulation of crosstalk interactions between innate immune receptors. Indeed, despite the physiological

nATuRE REvIEwS | Immunology

vOLuME 11 | MARCH 2011 | 187 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Box 1 | PRRs and microbial virulence proteins

Pathogen-associated molecular pattern (PAMP). A conserved molecular pattern that is found in pathogens but not mammalian cells. Examples include terminally mannosylated and polymannosylated compounds, which bind the mannose receptor, and various microbial products, such as bacterial lipopolysaccharides, hypomethylated DNA, flagellin and double-stranded RNA, which bind Toll-like receptors.

Lipid raft A membrane microdomain rich in cholesterol, sphingolipids and glycosylphosphatidylinositol-anchored proteins. These domains partition receptors for various cellular signalling and trafficking processes.

To exploit or subvert the functions of pattern recognition receptors (PRRs), pathogens use virulence factors rather than pathogen-associated molecular patterns (PAMPs), which are evolutionarily selected targets of pattern recognition. Unlike PAMPs, virulence proteins are characterized by mutability and lack of conservation, and they can readily respond to environmental changes. Therefore, virulence proteins can contribute to the adaptive fitness of pathogens and are unlikely to have been targeted for pattern recognition in the course of host–microorganism co-evolution3. However, the converse notion, that virulence proteins might have evolved to bind and possibly exploit PRRs, is suggested by documented examples of microbial protein interactions with Toll-like receptors (TLRs) and other innate immune receptors72,116,122,128,129. A characteristic example involves LcrV, a virulence protein of pathogenic Yersinia spp. that induces TLR2-mediated immunosuppression122. The TLR2-interacting epitope of LcrV maps to a 19-amino-acid amino-terminal sequence that is divergent from that of Pseudomonas aeruginosa PcrV, an LcrV homologue that does not bind TLR2. Moreover, a point mutation in the TLR2-binding epitope of LcrV abrogates its capacity to interact with TLR2 and decreases the virulence of the organism122. The plasticity of virulence proteins enables pathogens to manipulate the crosstalk interactions of host receptor signalling through various mechanisms. Molecular mimicry of host ligands (or counter-receptors) by microbial structures can activate inhibitory receptors24,50 (see the figure, part a). Virulence enzymes can convert host molecules into active agonists or ligands to manipulate modulatory receptors31,74 (see the figure, part b). Microbial proteins can function as host-receptor mimetics to block functional interactions of cooperative host receptors43 (see the figure, part c).

0CVWTCN NKICPFU +PJKDKVQT[ TGEGRVQT

The process by which integrins (such as complement receptor 3) become activated (assume a high-affinity binding state) through intracellular signalling initiated by other receptors, such as anaphylatoxin receptors or Toll-like receptors. By contrast, outside-in signalling refers to intracellular signalling initiated by the activated and ligated integrins.

Immunoreceptor tyrosine-based inhibitory motif (ITIM). A structural motif containing a tyrosine residue that is found in the cytoplasmic tails of several inhibitory receptors, such as Fcγ receptor IIB and paired immunoglobulinlike receptor B (PIRB). The consensus six-amino-acid ITIM sequence is (I/V/L/S)XYXX(L/V), in which X denotes any amino acid. Ligand-induced clustering of these inhibitory receptors results in tyrosine phosphorylation, often by SRC-family tyrosine kinases, which provides a docking site for the recruitment of cytoplasmic phosphatases.

*QUVTGEGRVQT NKICPFOKOGVKE

#EVKXCVKPI TGEGRVQT

*QOGQUVCVKE TGIWNCVKQPQH EGNNCEVKXCVKQP

Inside-out signalling

E

D

C

*QUV UWDUVTCVG /KETQQTICPKUO

'P\[OCVKE IGPGTCVKQPQH PCVWTCNNKICPFD[ OKETQQTICPKUO

2#/2U 4

+PJKDKVKQPQH 2#/2KPFWEGF EGNNCEVKXCVKQP

4

/QFWNCVKQPQH 4UKIPCNNKPI

significance of innate receptor crosstalk for immunity and homeostasis, undesirable outcomes can arise when these same mechanisms come under the control of pathogens. At least in principle, pathogens could induce antagonistic pathways, leading to immune suppression. Furthermore, pathogens might induce synergistic interactions to skew the host immune response away from protective immunity — for example, by promoting a T helper 2 (TH2) cell response when protective immunity is mediated by TH1 cells. A growing body of literature indicates that diverse pathogens use specific virulence factors (BOX 1) to exploit mechanisms of PRR cooperation, either at the receptor level or during downstream signalling (see Supplementary information S1 (table)). In this Review, we focus on the microbial strategies that instigate, subvert or disrupt innate immune receptor crosstalk, thereby contributing to microbial adaptive fitness and persisting infections. These mechanisms are referred to as ‘crosstalk manipulation’. we do not cover general immune evasion strategies of pathogens16,18–20 or the other ingenious tactics used by pathogens to interfere with intracellular signalling pathways through the direct targeting of signalling intermediates (such as the inactivation of signalling molecules through cleavage or dephosphorylation by virulence proteins)21–23; these topics have been extensively covered in excellent recent reviews. The objective of this Review is to summarize

/KETQDKCN 4OKOGVKE

4

0CVWTCN NKICPFU

4

%QQRGTCVKXG EGNNCEVKXCVKQP

0QEQQRGTCVKXG EGNNCEVKXCVKQP

and discuss the relevant literature identifying virulence 0CVWTG4GXKGYU^+OOWPQNQI[ factors and hijacked receptors that mediate ‘crosstalk manipulation’, and to organize distinct microbial tactics into common themes. we discuss the ways in which pathogens can: co-opt host inhibitory receptors (which are normally involved in homeostatic crosstalk), sometimes by expressing host-ligand mimetics24–28 (FIG. 1); instigate crosstalk pathways for the synergistic induction of immunosuppressive mediators, such as interleukin-10 (IL-10)29,30 (FIG. 2) or cyclic AMP31,32 (FIG. 3); induce insideout signalling to transactivate safe uptake pathways (which are intended for apoptotic cells)33–36 (FIG. 4); selectively inhibit TH1 cell-mediated immunity by capitalizing on complement–TLR regulatory crosstalk37–39 (FIG. 5); exploit TLR–TLR cross-inhibition40,41; and disrupt the functional receptor interactions that are required for cooperative protective signalling 42–44 (TABLE 1). Through these diverse mechanisms, pathogens hack into the host receptor crosstalk network to dysregulate the innate immune system for their own benefit.

Co-option of host inhibitory receptors A distinct set of host inhibitory immune receptors signal through immunoreceptor tyrosine-based inhibitory motifs (ITIMs), which recruit phosphatases such as SH2 domain-containing protein tyrosine phosphatase 1 (SHP1; also known as PTPn6), SHP2 (also known

188 | MARCH 2011 | vOLuME 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

Immunoreceptor tyrosine-based activation motif (ITAM). A structural motif containing two tyrosine residues that is found in the cytoplasmic tails of several signalling adaptor molecules. The motif has the form YXX(L/I) X6-12YXX(L/I), in which X denotes any amino acid. The tyrosine residues are targets for phosphorylation by SRC-family protein tyrosine kinases and subsequent binding of proteins that contain SRC homology 2 (SH2) domains, such as spleen tyrosine kinase (SYK).

Oxidative burst The process in phagocytic cells by which molecular oxygen is reduced by the NADPH oxidase system to produce reactive oxygen species, such as hydrogen peroxide and hydroxyl radicals. These are toxic oxidants that can destroy targeted microorganisms (for example, in the phagosome lumen).

Extracellular DNA trap Often referred to by the acronym NET (neutrophil extracellular trap). Upon activation (for example, through Toll-like or Fcγ receptors), neutrophils release nuclear content such as chromatin (DNA, histones and other proteins). This forms a web-like scaffold for the exposure of released antimicrobial molecules at high local concentrations, resulting in the trapping and extracellular killing of bacteria.

as PTPn11) or SH2 domain-containing inositol-5phosphatase (SHIP). These phosphatases attenuate signalling induced by juxtaposed receptors by dephosphorylating signalling intermediates8. ITIM-bearing receptors often, but not exclusively, interact with and inhibit immunoreceptor tyrosine-based activation motif (ITAM)-coupled receptors, such as the Fcγ receptors (FcγRs) and triggering receptor expressed on myeloid cells 1 (TREM1)45. ITAMs are found in the cytoplasmic domains of certain transmembrane adaptors (such as the FcR common γ-chain (FcRγ) and DAP12 (also known as TYROBP)) and generally mediate activating signals through the activation of spleen tyrosine kinase (SYk). However, ITAM-mediated cell activation requires highavidity ligation of the ITAM-coupled receptors, whereas low-avidity (or tonic) ligation of these receptors generates inhibitory signals mediated by SHP1 (REFS 46,47). An ITAM that is functioning in an inhibitory mode is referred to as an ITAMi47. Several microorganisms exploit ITIM-bearing or ITAMi-coupled receptors, which they can co-ligate with targeted receptors24–27. The resulting juxtaposition of ITIM-bearing or ITAMi-coupled receptors with the targeted receptors (such as TLRs or phagocytic receptors) allows the induction of inhibitory crosstalk that suppresses cellular activation and/or phagocytosis. For example, Moraxella catarrhalis and Neisseria meningitidis use virulence proteins (uspA1 and Opa, respectively) to bind and activate carcinoembryonic antigen-related cell adhesion molecule 1 (CEACAM1) on respiratory epithelia, resulting in inhibition of TLR2 signalling. The potential significance of this inhibitory effect is underscored by the heavy involvement of TLR2 in the inflammatory and antimicrobial responses of pulmonary epithelial cells. At the molecular level, TLR2 associates with CEACAM1 following their co-activation by these respiratory pathogens, although the microbial TLR2 ligands involved have not been identified26. Subsequently, phosphorylation of the cytoplasmic ITIM of activated CEACAM1 recruits SHP1, which limits phosphorylation of the p85α regulatory subunit of phosphoinositide 3-kinase (PI3k) that is recruited to the TLR2 cytoplasmic domain26 (FIG. 1a). This in turn inhibits the activation of the PI3k–AkT pathway and decreases the production of IL-8 and granulocyte–macrophage colony-stimulating factor, which regulate the development, mobilization and activation of granulocytes in response to respiratory infections. These data indicate a mechanism of immune evasion through CEACAM1 exploitation, which might be shared by other CEACAM1-interacting pathogens such as Neisseria gonorrhoeae and enteropathogenic Escherichia coli and Salmonella spp. strains. However, the lack of appropriate animal models (M. catarrhalis and N. meningitidis infections are specific to humans) prevents in vivo confirmation of the model. Group B Streptococcus (GBS) can bind to sialic acidbinding immunoglobulin-like lectins (Siglecs) on leukocytes in a sialic acid-dependent or -independent manner, depending on the GBS serotype. For example, the sialylated capsular polysaccharide of serotype III GBS mimics host sialylated glycans and binds Siglec-9 (REF. 24), and

the cell wall-anchored β-protein (also known as Bac) of serotype Ia and III GBS strains binds Siglec-5 (REF. 25). GBS engagement of ITIM-bearing Siglec-5 or Siglec-9 activates inhibitory SHP2-dependent signals that interfere with cellular activation (FIG. 1b). This leads to the inhibition of several leukocyte antimicrobial functions, including phagocytosis, induction of the oxidative burst and formation of extracellular DNA traps, and this allows GBS to escape killing by monocytes and/or neutrophils24,25. The receptor(s) that crosstalk with Siglecs in this immune evasion mechanism have not been identified, but they could be TLRs, as Siglec-E (the mouse orthologue of Siglec-9) downregulates TLR4 signalling through SHP2 (REF. 48). An evasion strategy analogous to that of GBS might be used by Staphylococcus aureus in mouse macrophages, in which the TLR2-induced inflammatory response is counteracted by co-activation of another ITIM-containing receptor, the murine paired immunoglobulin-like receptor B (PIRB)49 (FIG. 1c). under physiological conditions, PIRB regulates the activation of the ITAM-coupled PIRA. The identity of the PIRB-interacting ligand(s) of S. aureus is uncertain, although polyanionic molecules, such as dextran sulphate and polyinosinic acid, competitively inhibit the binding of this bacterium to PIRB49. PIRB has human orthologues — Ig-like transcript 2 (ILT2; also known as LIR1) and ILT5 (also known as LIR3) — that are also bound by S. aureus and other bacteria49. whether the ILT receptors (also known as leukocyte immunoglobulinlike receptors (LIRs)) in humans are exploited by bacterial pathogens remains to be investigated. However, ILT2 is used by human cytomegalovirus (HCMv) to suppress natural killer (nk) cell-mediated cytolysis, through SHP1-dependent crosstalk signals that interfere with activating nk cell receptors28,50 (FIG. 1d). Intriguingly, activation of ILT2 and phosphorylation of its ITIM are initiated by interaction with HCMv uL18, a viral glycoprotein that not only mimics but also outcompetes MHC class I molecules, as it binds ILT2 with >1,000-fold higher affinity 50. Another virus, HIv-1, binds DC immunoreceptor (DCIR; also known as CLEC4A), an ITIM-containing C-type lectin, and this interaction promotes HIv-1 infection of DCs51. Through signalling crosstalk, endocytosed DCIR inhibits the production of TLR8-induced IL-12 and TLR9-induced interferon-α (IFnα) in conventional and plasmacytoid DCs, respectively 52,53 (FIG. 1e). Therefore, this pathway might contribute to immune evasion by HIv-1 and other DCIR-binding viruses. E. coli evades phagocytic killing by inducing ITAMi signals through FcγRIII (also known as CD16) and its ITAM-bearing signalling adaptor, FcRγ, which crosstalk with the class A scavenger receptor macrophage receptor with collagenous structure (MARCO)27 (FIG. 1f). Specifically, E. coli binds FcγRIII directly in a non-opsonic manner (that is, without antibody), and this low-avidity interaction induces FcRγ phosphorylation, followed by SHP1 recruitment. SHP1 dephosphorylates PI3k, which is thereby unable to support MARCO-mediated phagocytosis of E. coli 27. Mice deficient in FcγRIII or FcRγ have increased survival rates in models of sepsis induced by caecal ligation and puncture, and this is attributed, in part, to their enhanced ability to clear E. coli 27.

nATuRE REvIEwS | Immunology

vOLuME 11 | MARCH 2011 | 189 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS C

D

/ECVCTTJCNKU 0OGPKPIKVKFKU

5KINGE QT5KINGE %'#%#/

6.4

+6+/ 5*2

E

)TQWR$ 5VTGRVQEQEEWU

5CWTGWU 2+4$

6.4!

+6+/

+6+/ 5*2

2+-

6.4

5*2 2+-s#-6 RCVJYC[!

#-6 1ZKFCVKXGDWTUV %[VQMKPGKPFWEVKQP $CEVGTKCNMKNNKPI

+PȯCOOCVQT[ TGURQPUG

F

0-EGNN

G

H 'EQNK

#EVKXCVKQP



+PȯCOOCVQT[ TGURQPUG

'PFQUQOG 5*2 +6+/

(Eγ4+++

/#4%1

8KTWU &%+4

%&

0-)%

5*2

+.6

(E4γ 5;-

+6+/ 6.4

5*2 QT5*2

6.4 *.#'

+6#/K

2+2JCIQE[VQUKU

7. +(0α *%/8KPHGEVGFEGNN

60( +.

Figure 1 | Inhibition of cell activation by pathogen-ligated ITIm-bearing or ITAmi-coupled receptors. a | Moraxella catarrhalis and Neisseria meningitidis use specific virulence proteins to activate carcinoembryonic antigen-related cell 0CVWTG4GXKGYU^+OOWPQNQI[ adhesion molecule 1 (CEACAM1), which co-associates with and inhibits Toll-like receptor 2 (TLR2) signalling. The underlying crosstalk involves phosphorylation of the CEACAM1 immunoreceptor tyrosine-based inhibitory motif (ITIM), which recruits SH2 domain-containing protein tyrosine phosphatase 1 (SHP1; also known as PTPN6); this suppresses the phosphorylation of phosphoinositide 3-kinase (PI3K) and downstream activation of the AKT-mediated pro-inflammatory pathway. b | Serotypes of Group B Streptococcus (GBS) bind sialic acid-binding immunoglobulin-like lectins (Siglecs), either through molecular mimicry of host sialylated glycans or through a cell wall-anchored protein. The activation of ITIM-bearing Siglec-5 or Siglec-9 by GBS activates inhibitory SHP2 (also known as PTPN11)-dependent signals that interfere with TLR-mediated cellular activation and antimicrobial functions. c | Staphylococcus aureus uses the ITIM-containing paired immunoglobulinlike receptor B (PIRB) to crosstalk with and inhibit the TLR2-induced inflammatory response, possibly by inhibiting the PI3K–AKT pathway. d | Human cytomegalovirus (HCMV) expresses an MHC class I homologue, UL18, which interacts with immunoglobulin-like transcript 2 (ILT2; also known as LIR1) and activates ITIM-dependent and SHP1-mediated signalling. This inhibits natural killer (NK) cell activating receptors, such as the NK group 2, member C (NKG2C)–CD94 complex, and interferes with NK cell-mediated cytolysis of the HCMV-infected cell. e | Upon activation by viruses, the ITIM-bearing DC immunoreceptor (DCIR; also known as CLEC4A) becomes internalized into endosomes and inhibits endosomal TLR signalling — specifically, it inhibits production of TLR8-induced interleukin-12 (IL-12) and TLR9-induced interferon-α (IFNα) in conventional and plasmacytoid dendritic cells, respectively. f | Escherichia coli evades macrophage receptor with collagenous structure (MARCO)-dependent phagocytic killing through inhibitory crosstalk with Fcγ receptor III (FcγRIII; also known as CD16). Specifically, non-opsonized E. coli binds with low affinity to FcγRIII and induces partial phosphorylation of the FcR common γ-chain (FcRγ) ITAM (ITAMi), leading to weak mobilization of spleen tyrosine kinase (SyK) but strong recruitment of SHP1. SHP1 dephosphorylates PI3K and impairs MARCO-dependent phagocytosis. TNF, tumour necrosis factor.

It is conceivable that at least some of these inhibitory receptors could be used by the immune system to limit and control unnecessary inflammatory responses, as is the case for commensal Gram-negative bacteria in the gut 54. In this context, the interaction of a receptor such as PIRB with a commensal organism might avoid a vigorous host immune response, an outcome

that is beneficial for both the host and the microorganism. Although it might sound counterintuitive to suggest that commensal and pathogenic microorganisms share immune evasion mechanisms, the latter express additional factors, such as invasins and/or toxins, that enable them to breach epithelial barriers and invade host cells.

190 | MARCH 2011 | vOLuME 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Induction of immunosuppressive mediators Many pathogens capitalize on the immunosuppressive properties of IL-10 and cAMP to undermine aspects of innate immune defence55–58. Although IL-10- and cAMP-dependent signalling pathways have important regulatory functions in maintaining homeostasis of the immune system, excessive and sustained production of these mediators impairs the killing capacity of phagocytes. The functions that are suppressed by IL-10 and cAMP include the production of reactive oxygen or nitrogen intermediates and pro-inflammatory cytokines, and phagolysosomal fusion and acidification. However, differences between the effects of IL-10 and cAMP do exist; for example, IL-10 additionally inhibits the induction of co-stimulatory molecule expression by antigenpresenting cells, whereas cAMP inhibits degranulation of neutrophils more consistently than IL-10 (REFS 55, 58–60). Moreover, cAMP upregulates rather than inhibits IL-6 production32,58. Because cAMP is generated through enzymatic action (by adenylyl cyclase), its levels can increase markedly within minutes of stimulation and, in fact, cAMP subsequently enhances transcription of the IL10 gene61. Some microorganisms have genetically encoded mechanisms for regulating cAMP or IL-10 production. Specifically, bacteria such as Bordetella pertussis and Bacillus anthracis express their own adenylyl cyclase enzyme for unregulated production of cAMP56,57, and certain viruses (such as Epstein–Barr virus, orf virus and equine herpesvirus 2) encode their own viral homologue of IL-10 (REF. 55). Alternatively, other pathogens can elicit IL-10 or cAMP production through the induction of synergistic crosstalk pathways29–32,62,63 (see Supplementary information S1 (table)). Induction of IL‑10. Several clinically important human pathogens (including Mycobacterium tuberculosis, Mycobacterium leprae, Helicobacter pylori, Candida albicans, measles virus and HIv-1) induce crosstalk between TLRs and the C-type lectin DC-SIGn (DC-specific ICAM3-grabbing non-integrin; also known as CD209), and this leads to high levels of IL-10 production by DCs29,62,63 (FIG. 2a). For example, mycobacterial mannosylated lipoarabinomannan (ManLAM) binds DC-SIGn and induces a complex signalling pathway that activates the serine/threonine kinase RAF1 (FIG. 2a). RAF1, in turn, induces phosphorylation of the p65 subunit of nuclear factor-κB (nF-κB) on Ser276 and subsequent acetylation on several lysines. The acetylation of p65 requires the activation of both DC-SIGn and TLR signalling, and is mediated by the related acetyltransferases CREB-binding protein and p300, which are recruited to p65 by binding to its phosphorylated Ser276 residue. This acetylation allows nF-κB to mediate prolonged and enhanced transcription of IL10; the underlying mechanism involves the enhanced DnA-binding affinity and transcriptional activity of acetylated p65 and its prolonged presence in the nucleus29. The same pathway was later shown to induce enhanced transcription of the IL12A and IL12B genes64. The upregulation of both TH1 cell-promoting and TH1 cell-repressing cytokines

(IL-12 and IL-10, respectively) is consistent with earlier observations that mycobacteria induce IL-10-producing TH cells without a TH1 or TH2 bias64,65. The interaction of mycobacteria with DC-SIGn has also been associated with impaired or intermediate-stage maturation of DCs62,66, although it is not clear whether this represents immune evasion or a host mechanism to decrease inflammatory pathology. By contrast, the interaction of H. pylori with DC-SIGn through fucose-containing lipopolysaccharide (LPS) Lewis antigens leads to increased IL-10 and decreased IL-12 production and, eventually, to the inhibition of TH1 cell development 63,64 (FIG. 2b). The two pathways for IL-10 production through DC-SIGn (FIG. 2a,b) are therefore divergent. Indeed, in contrast to mannose-containing DC-SIGn ligands (such as mycobacterial ManLAM, fungal mannan and HIv-1 gp120), the binding of fucose-containing DC-SIGn ligands (such as Lewis X of H. pylori) proactively excludes RAF1 and other select components from the DC-SIGn signalling complex and thereby modifies downstream signal transduction64. The spirochaete Borrelia burgdorferi (the causative agent of Lyme disease) induces DC-SIGn crosstalk with TLR2, although the DC-SIGn ligand is contributed by its tick vector (Ixodes scapularis)67 (FIG. 2c). Specifically, TLR2 and DC-SIGn are activated, respectively, by B. burgdorferi lipoproteins and I. scapularis salivary protein Salp15, which is captured by outer surface protein C of the bacterium. Although the composition of the DC-SIGn signalling complex for B. burgdorferi was not reported in the same detail as that for H. pylori (described in the study discussed above64), Salp15 contains mannose structures and induces RAF1 signalling 67. However, in this case, RAF1 activation does not lead to acetylation of the nF-κB p65 subunit, as described for ManLAM. It is thought that because Salp15 can also bind CD4, this receptor might participate in the DC-SIGn–TLR2 signalling crosstalk and alter the RAF1-dependent signalling pathway (FIG. 2c). In this pathway, Salp15-induced RAF1 activation stimulates MAPk/ERk kinase (MEk) signalling, which promotes IL6 and TNF mRnA decay and impairs nucleosome remodelling and, hence, transcriptional activation at the IL12A promoter. The same pathway does not destabilize IL10 mRnA; in fact, coactivation of DCs with B. burgdorferi and Salp15 synergistically enhances IL-10 production67. As IL-12 inhibits IL-10 production, it is possible that the observed downregulation of IL-12 expression could contribute to the increased IL-10 levels. In terms of biological relevance, the consequences of Salp15 utilization by B. burgdorferi include inhibition of TLR-dependent maturation of DCs and their capacity to activate T cells, which is advantageous for both the arthropod vector and the bacterial pathogen67. In neutrophils, mycobacteria use another C-type lectin (potentially CLEC5A), which is coupled to the ITAM-containing adaptor protein DAP12, to induce SYk-dependent crosstalk with the TLR2–MYD88 (myeloid differentiation primary response protein 88) pathway. This crosstalk synergistically upregulates

nATuRE REvIEwS | Immunology

vOLuME 11 | MARCH 2011 | 191 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS C

/[EQDCEVGTKC %CPFKFCCNDKECPU *+8 /GCUNGUXKTWU

D

E

F 5CNR

*R[NQTK $DWTIFQTHGTK

&%5+)0 6.4

6.4 4#5 .#4) 4*1# 4#( .52 %0- -54

/;&

%&

6.4

&%5+)0

!

4#5

/;&

.52 /;&

/[EQDCEVGTKC

%.'%# 

6.4

+6#/ 5;-

4#(

/;&

p50 p65 /'-

R0(κ$ RJQURJQT[NCVKQP

NF-κB #-6

R

60(CPF+. R0(κ$ CEGV[NCVKQP

↑+. ↓+.

+.# !

↑+.+.#CPF+.$

IGPGVTCPUETKRVKQP

+ORCKTGFQT KPVGTOGFKCVG &%OCVWTCVKQP

7PDKCUGF 6*EGNN FKȭGTGPVKCVKQP

↑+.

↑+.

6*EGNNRQNCTK\CVKQP

&%OCVWTCVKQP

+PȯCOOCVKQP

↑/[EQDCEVGTKCN

RGTUKUVGPEG

Figure 2 | Pathogen-induced host receptor crosstalk to stimulate Il-10 production. a | The indicated pathogens express mannose-containing ligands that bind DC-specific ICAM3-grabbing non-integrin (DC-SIGN; also known as 0CVWTG4GXKGYU^+OOWPQNQI[ CD209) and induce crosstalk with Toll-like receptors (TLRs) through RAF1. Induction of RAF1 signalling involves the participation of the LSP1–KSR1–CNK scaffolding complex and upstream activators (LARG, RAS and RHOA), and this pathway then mediates the phosphorylation and acetylation of TLR-activated nuclear factor-κB (NF-κB) p65 subunit. This results in increased transcription of the interleukin-10 (IL10), IL12A and IL12B genes owing to the enhanced DNA-binding affinity and transcriptional activity of acetylated p65. b | Helicobacter pylori binds DC-SIGN through fucose-containing lipopolysaccharide Lewis antigens and activates leukocyte-specific protein 1 (LSP1)-dependent (but RAF1-independent) signalling, leading to increased IL-10 production, decreased IL-12 production and the inhibition of T helper 1 (TH1) cell development. c | Borrelia burgdorferi uses the salivary protein Salp15 of its tick vector to induce DC-SIGN–TLR crosstalk. Here, DC-SIGN-induced RAF1 signalling does not lead to p65 acetylation but stimulates MAPK/ERK kinase (MEK) signalling, which promotes IL6 and tumour necrosis factor (TNF) mRNA decay and impairs nucleosome remodeling at the IL12A promoter. Conversely, IL10 mRNA is not destabilized but, rather, IL-10 production is synergistically increased, and this leads to inhibition of dendritic cell (DC) maturation. This divergent RAF1 pathway might be attributed to Salp15 binding to CD4, which may participate in the crosstalk. d | In neutrophils, mycobacteria interact with a C-type lectin (possibly CLEC5A) linked to immunoreceptor tyrosine-based activation motif (ITAM)-bearing DAP12. This interaction induces spleen tyrosine kinase (SyK)-dependent crosstalk with the TLR2–MyD88 (myeloid differentiation primary response protein 88) pathway, and this synergistically upregulates IL-10 production through sustained phosphorylation of AKT and p38 mitogen-activated protein kinase. This decreases lung inflammation but increases the persistence of a high mycobacterial burden in a mouse lung infection model. CNK, connector enhancer of KSR; KSR1, kinase suppressor of RAS1; LARG, leukaemia-associated RHO guanine nucleotide exchange factor (also known as RhoGEF12).

G protein-coupled receptors (GPCRs). Also known as seventransmembrane-domain receptors, this large group of receptors can bind a diverse set of molecules (such as chemokines, complement anaphylatoxins, hormones and neurotransmitters) and can induce intracellular signalling by coupling to heterotrimeric GTP-regulated signalling proteins.

IL-10 production through rapid and sustained phosphorylation of two kinases: AkT and p38 mitogen-activated protein kinase (p38 MAPk)30 (FIG. 2d). In a chronic infection model in mice, neutrophil-derived IL-10 decreased lung inflammation but contributed to the persistence of a high mycobacterial burden30. Induction of cAMP. Porphyromonas gingivalis is a periodontal pathogen that is also implicated in systemic conditions such as atherosclerosis and rheumatoid arthritis68,69. This Gram-negative bacterium uses an array of virulence factors to evade immune elimination and chronically persist in human hosts70. Recent evidence indicates that P. gingivalis achieves this partly

by subverting immune receptor signalling crosstalk31,32 (FIG. 3). Specifically, P. gingivalis induces the recruitment and co-association in macrophage lipid rafts of TLR2 and two G protein-coupled receptors (GPCRs) — CXCchemokine receptor 4 (CXCR4) and the complement C5a receptor (C5aR) — leading to the induction of high and sustained levels of cAMP31,32. P. gingivalis activates TLR2 through its surface fimbriae and lipoproteins. notably, it does not rely on immunological means for C5aR activation. Indeed, the bacterium can generate C5a through its own C5 convertase-like enzymatic activity, mediated by Arg-specific cysteine proteinases (the RgpA and RgpB gingipains) 31. In addition, P. gingivalis can directly activate CXCR4 through its

192 | MARCH 2011 | vOLuME 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS %

2IKPIKXCNKU

%C

%4

%&

%;6

4#%

%:%4

%C4

2+-

6.4

)αK

6.4s 6.4 '4-CPFQT '4- E#/2

%C

+4( 2-# +.

K015

$CEVGTKCN MKNNKPI

)5-β +(0γ

Figure 3 | Integration of subversive crosstalk pathways, leading to inhibition of pathogen killing. Porphyromonas gingivalis interacts with Toll-like receptor 2 (TLR2) 0CVWTG4GXKGYU^+OOWPQNQI[ (specifically with the CD14–TLR2–TLR1 complex) and TLR4. The latter receptor is blocked by the bacterium’s atypical lipopolysaccharide (which functions as a TLR4 antagonist) and so cannot induce protective responses. The TLR2 response is proactively modified through crosstalk with other receptors that are regulated by P. gingivalis. For example, P. gingivalis controls the C5a receptor (C5aR) using bacterial Arg-specific cysteine proteinases, which cleave C5 to release biologically active C5a. This C5a stimulates intracellular Ca2+ signalling, which synergistically enhances the otherwise weak cyclic AMP (cAMP) response induced by TLR2 activation alone. Maximal cAMP induction requires the participation of CXC-chemokine receptor 4 (CXCR4), which is activated directly by the pathogen’s fimbriae and associates with both TLR2 and C5aR in lipid rafts. The ensuing activation of cAMP-dependent protein kinase (PKA) inactivates glycogen synthase kinase-3β (GSK3β) and impairs the inducible nitric oxide synthase (iNOS)-dependent killing of the pathogen in macrophages. An additional pathway induced downstream of TLR2 is an inside-out signalling pathway, mediated by RAC1, phosphoinositide 3-kinase (PI3K) and cytohesin 1 (CyT1), that transactivates complement receptor 3 (CR3; also known as αMβ2 integrin or CD11b–CD18). Activated CR3 binds P. gingivalis and induces extracellular signal-regulated kinase 1 (ERK1) and/or ERK2 signalling, which in turn selectively downregulates the expression of the interleukin-12 (IL-12) p35 and p40 subunits through the suppression of interferon regulatory factor 1 (IRF1). This inhibitory ERK pathway is also activated downstream of C5aR (not shown here for clarity; see FIG. 5). Decreased production of bioactive IL-12 and, secondarily, of interferon-γ (IFNγ), leads to impaired immune clearance of P. gingivalis.

Anaphylatoxins The pro-inflammatory fragments C3a and C5a that are generated during the activation of the complement system. They mediate various inflammatory responses through their corresponding G protein-coupled receptors, such as chemotaxis, oxidative burst and histamine release (from mast cells), but they (in particular, C5a) can also regulate other innate immune components (such as TLRs) through crosstalk signalling pathways.

surface fimbriae (albeit using different epitopes from those mediating TLR2 binding 71,72), without requiring CXCL12 as a ligand32. Recognition of P. gingivalis by TLR2 alone induces a weak cAMP response, whereas activation of CXCR4 or C5aR alone fails to induce cAMP. Strikingly, however, P. gingivalis-stimulated TLR2 cooperates with activated C5aR and CXCR4 to synergistically increase cAMP production. This, in turn, greatly increases cAMP-dependent protein kinase (PkA) signalling, which inactivates glycogen synthase kinase-3β and, hence, impairs inducible nitric oxide synthase (inOS)-dependent killing of bacteria in vitro and in vivo31,32 (FIG. 3). The C5aR–TLR2 crosstalk depends on Gαi-coupled C5aR signalling and the mobilization of intracellular calcium31, which potentiates concurrent cAMP signalling and, hence, PkA activation7. Although the C5aR–TLR2 and CXCR4–TLR2

crosstalk pathways can proceed independently of each other, maximal cAMP induction requires the cooperation of all three receptors31. P. gingivalis interacts with at least one other TLR, TLR4, although the ligands involved are atypical LPS molecules that only weakly activate TLR4 (in the case of LPS with a 5-acyl monophosphate lipid A structure) or even antagonize TLR4 (4-acyl monophosphate lipid A)73 (FIG. 3). The Gs protein-coupled A2A and A2B adenosine receptors (A2AR and A2BR) respond to extracellular adenosine and increase intracellular cAMP levels. Activated A2AR (which has a higher affinity than A2BR for adenosine) crosstalks with and inhibits TLR-induced inflammatory pathways14. Intriguingly, S. aureus expresses cell wall-associated adenosine synthase A (AdsA), which converts adenosine monophosphate to adenosine74. The pathogen exploits the immunosuppressive properties of the adenosine it generates to ‘disable’ phagocytes in the blood and escape immune clearance. AdsA-deficient mutants of S. aureus have a survival disadvantage in the blood that can be reversed by the addition of exogenous adenosine. The authors of this paper also identified another ten species of Gram-positive bacteria (such as B. anthracis, Clostridium perfringens and Listeria monocytogenes) that express homologues of the adenosine synthase domain of AdsA. A study of B. anthracis showed that it also uses AdsA to escape phagocytic clearance, which suggests that additional AdsA-expressing bacteria share this evasion mechanism74. Given that A2AR signalling inhibits TH1 and TH17 cell development, while promoting the generation of adaptive regulatory T cells75, AdsA-expressing pathogens might also be able to manipulate T cell-mediated immunity.

Inside-out signalling Complement receptor 3 (CR3; also known as αMβ2 integrin or CD11b–CD18) is a versatile β2 integrin that binds multiple ligands or counter-receptors (such as the complement component iC3b and intercellular adhesion molecule 1 (ICAM1)) and contributes to the phagocytosis of apoptotic cells, leukocyte trafficking and the regulation of cytokine production2. Its adhesive activity is tightly regulated; whereas CR3 has a low-affinity conformation in resting cells, a rapid and transient shift to a high-affinity state can be triggered through inside-out signalling by chemokine or anaphylatoxin receptors 76. TLRs can also induce insideout signalling for CR3 activation, as originally shown for TLR2 (REFS 77,78) and recently confirmed for TLR4 (REF. 79). The TLR2 inside-out signalling pathway proceeds through RAC1, PI3k and cytohesin 1 (REFS 77,78,80) (FIG. 4). In contrast to TLR4, however, the TLR2 inside-out pathway does not depend on MYD88 (REF. 81). This is because PI3k can be recruited directly to the TLR2 cytoplasmic tail, which contains PI3k-binding motifs that are absent from this region of TLR4 (REF. 82). In terms of physiological significance, the ability of pathogen-sensing receptors, such as the TLRs, to transactivate CR3 might contribute to leukocyte recruitment to sites of infection80.

nATuRE REvIEwS | Immunology

vOLuME 11 | MARCH 2011 | 193 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS $RGTVWUUKU 'HCGECNKU

2IKPIKXCNKU $CPVJTCEKU /[EQDCEVGTKC %&

%4

6.4s6.4

α8β

%;6 4#% 2+-

2+-

+PUKFGQWV UKIPCNNKPI

+PUKFGQWV UKIPCNNKPI

%&

'PJCPEGF RGTUKUVGPEG

Figure 4 | Pathogen-induced transactivation of CR3-mediated internalization. Certain bacteria (such 0CVWTG4GXKGYU^+OOWPQNQI[ as Porphyromonas gingivalis, Mycobacterium tuberculosis and Bacillus anthracis) bind CD14 and induce Toll-like receptor 2 (TLR2)–TLR1 inside-out signalling for activating and binding complement receptor 3 (CR3; also known as αMβ2 integrin or CD11b–CD18), which leads to a relatively ‘safe’ uptake of these organisms by macrophages. The signalling pathway that activates the high-affinity state of CR3 is mediated by RAC1, phosphoinositide 3-kinase (PI3K) and cytohesin 1 (CyT1). Enterococcus faecalis and Bordetella pertussis stimulate their uptake by CR3 through an alternative inside-out signalling pathway. This mechanism is activated by the interaction of these bacteria with a receptor complex comprising αVβ3 integrin and CD47, and is dependent on PI3K signalling. Similarly, CR3-mediated uptake of these bacteria prevents their intracellular killing and promotes their persistence in the mammalian host.

Although CR3 is a phagocytic receptor, it is not linked to vigorous microbicidal mechanisms such as those activated by FcγR-mediated phagocytosis83,84 and, under certain conditions, CR3-derived phagosomes do not fuse with lysosomes85. This is possibly related to the physiological role of CR3 in the uptake of apoptotic cells, which are not normally recognized as a danger that warrants a strong host immune response2. not surprisingly, therefore, the TLR2–CR3 crosstalk pathway is a target of immune subversion by several pathogens. P. gingivalis, M. tuberculosis and B. anthracis use, respectively, their fimbriae, lipoarabinomannan and the BclA glycoprotein to interact with the CD14–TLR2 receptor complex and induce TLR2-mediated transactivation of CR3 (REFS 33,77,78) (FIG. 4). This mechanism allows these bacteria to hijack the phagocytic functions of CR3 for a relatively safe ‘outside-in’ entry into macrophages. The subversive effect of P. gingivalis is evident from the finding that CR3-deficient macrophages are superior to wildtype controls at intracellular killing of this pathogen34,86. Moreover, compared with CR3-deficient mice, wildtype mice have increased susceptibility to infection with B. anthracis spores; this is attributed to the ‘safe’ storage of the spores in macrophages after uptake by CR3, and their carriage to sites of spore germination and bacterial growth33. Similarly, the ability of M. tuberculosis to survive in macrophages might depend, at least in part, on the

stimulation of TLR2-induced CR3-mediated phagocytosis78, although an additional step involves the recruitment of coronin 1A to CR3-derived phagosomes, preventing their fusion with lysosomes85. Two other organisms, Enterococcus faecalis and B. pertussis, activate CR3-mediated phagocytosis through alternative inside-out signalling pathways. Specifically, the ‘aggregation substance’ glycoprotein of E. faecalis and the filamentous haemagglutinin of B. pertussis interact with a signalling complex comprising the αvβ3 integrin and the integrin-associated protein CD47, leading to CR3 transactivation in macrophages and neutrophils36,87 (FIG. 4). CR3-mediated uptake of E. faecalis by phagocytes does not induce the oxidative burst and thus promotes the survival of the bacterium35,88. B. pertussis takes advantage of CR3-mediated phagocytosis to escape immune clearance in vivo89; the pathogen is readily cleared if it is phagocytosed through FcγRIII89, which, unlike CR3, is coupled to potent microbicidal mechanisms83,84. Interestingly, at least in two cases, TLR2-mediated inside-out signalling is activated by the same virulence proteins (namely, P. gingivalis fimbriae34 and B. pertussis filamentous haemagglutinin36) that bind transactivated CR3 for non-opsonic phagocytosis. Given the role of CR3 in the phagocytosis of iC3b-coated apoptotic cells, the pathogens that target CR3 as a ‘preferred’ portal of entry (FIG. 4) might have evolved to co-opt a homeostatic, anti-inflammatory mechanism to evade the innate immune system.

Subversive complement–TLR crosstalk Complement and TLRs are rapidly activated in response to infection, and common PAMPs (such as LPS and CpG DnA) function as both complement activators and TLR ligands2. In fact, the early innate immune response is shaped, to a large extent, by bidirectional crosstalk between the two systems11. For example, activation of the complement anaphylatoxin receptors (the GPCRs C3aR and C5aR) synergistically enhances TLR-induced production of pro-inflammatory and antimicrobial mediators2,90,91. The signalling pathways involved in complement–TLR4 crosstalk converge at the level of MAPks, specifically extracellular signal-regulated kinase 1 (ERk1), ERk2 and Jun n-terminal kinase (Jnk)90. This synergy could potentially enhance innate resistance to infection, and similar crosstalk effects explain, at least in part, why pharmacological inhibition of C5aR protects against sepsis that is induced by high doses of LPS. In a reciprocal reinforcing manner, TLR activation induces the expression of complement components and/or receptors2. Moreover, TLR signalling decreases the desensitization of GPCRs by downregulating the expression of GPCR kinases that induce receptor phosphorylation and internalization92. This TLR activity would be expected to prolong the activation of C3aR and C5aR in response to infection. Regulation of IL‑12 production. The crosstalk between anaphylatoxin receptors (particularly C5aR) and TLRs also has specific antagonistic effects, at least in macrophages, that selectively alter the induction of

194 | MARCH 2011 | vOLuME 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS IL-12-family cytokines. The underlying mechanism of C5aR–TLR crosstalk, which depends on ERk1 and/or ERk2 and PI3k signalling, suppresses the activation of interferon regulatory factor 1 (IRF1) and IRF8. These factors regulate the expression of IL-12 and related cytokines such as IL-23 (REFS 37,93). The C5aR–ERk– IRF1 pathway preferentially inhibits IL-12p70 production, whereas the C5aR–PI3k–IRF8 pathway mainly downregulates the production of IL-23 (REF. 37). Similar inhibitory effects on IL-12 induction are seen when other complement receptors (including gC1qR, CD46 and CR3) are co-activated with TLR4 or TLR2 in mouse macrophages or human monocytes39,94,95. The activation of these complement receptors by their natural ligands (such as C3b and C5a), which are produced during the complement cascade or by non-complement host enzymes (such as thrombin and kallikreins)2, might be a regulatory mechanism to attenuate T cell-mediated inflammation, given the important role of IL-12 in TH1 $CEVGTKWO 2HCNEKRCTWO KPHGEVGF GT[VJTQE[VG

#RQRVQVKE EGNN

8KTWU

K%D

$CEVGTKWO

%

%D %S

6.4

%&

%&

%C %C4

%4

/CPPQUG TGEGRVQT

'4-CPFQT '4-

'4-CPFQT '4-

+4(

+4(

%&

I%S4

2+-

2QUV VTCPUETKRVKQPCN OGEJCPKUO

+.

6*EGNNOGFKCVGFKOOWPKV[

Figure 5 | Selective inhibition of TlR-induced Il-12 production by pathogeninstigated PRR crosstalk. The crosstalk between Toll-like receptors (TLRs) and anaphylatoxin receptors (particularly C5a receptor (C5aR)) or0CVWTG4GXKGYU^+OOWPQNQI[ other complement receptors (such as complement receptor 3 (CR3; also known as αMβ2 integrin or CD11b–CD18), gC1q receptor (gC1qR) and CD46) selectively inhibits the induction of interleukin-12 (IL-12) production. Relatively little is known regarding the pathways that mediate this selective inhibition; signalling molecules that have been implicated, such as extracellular signal-regulated kinase 1 (ERK1), ERK2 and phosphoinositide 3-kinase (PI3K), are shown downstream of the corresponding receptors. At least for ERK1 and ERK2, the selectivity of IL-12 inhibition is attributed to the suppression of a crucial transcription factor, interferon regulatory factor 1 (IRF1). Post-transcriptional mechanisms might also contribute to IL-12 inhibition. Activation of the complement receptors by their natural ligands might have a homeostatic function, and this is also a possibility for other innate immune receptors (such as CD36, mannose receptor and CD150 (also known as SLAM)) that share the ability to downregulate IL-12 production. However, these same receptors can be activated by bacterial, viral or parasitic pathogens, which can thereby downregulate TLR-induced IL-12 production to interfere with host defences (such as the inhibition of T helper 1 (TH1) cell-mediated immunity). Although microbial molecules that function as ligands for C5aR have been described, this receptor can also come under pathogen control through the enzymatic generation of high levels of C5a by microbial C5 convertase-like enzymes. P. falciparum, Plasmodium falciparum; PRR, pattern recognition receptor.

cell differentiation and activation96. For example, inhibition of IL-12p70 production by C5aR–TLR4 crosstalk can suppress TH1 cell-mediated pathology. Moreover, decreased IL-12p70 production following the interaction of CR3 on macrophages with iC3b-coated apoptotic cells might prevent unwarranted inflammation and T H1 cell activation during the phagocytosis of apoptotic cells2. Conversely, C1q deficiency in humans and mice causes inflammatory autoimmune pathology 97, although it is uncertain whether, and to what extent, this results from a lack of C1q–gC1qR-mediated homeostatic regulation of T cells through crosstalk with TLR signalling pathways in antigen-presenting cells. The significance of these antagonistic crosstalk interactions becomes more evident in the context of microbial pathogenesis, in which complement receptors seem to modify TLR signalling and skew the TH cell response in a manner that interferes with protective immunity 37–39. Certain non-complement innate immune receptors are also implicated in the selective inhibition of TLRinduced IL-12 production98, and examples of pathogens that exploit complement and non-complement receptors are given below (FIG. 5). Leishmania major, an intracellular parasite of macrophages, seems to benefit from complement activation and C5aR-induced inhibition of TH1 cell-mediated immunity 37. This conclusion is based on the finding that BALB/c mice, which are normally susceptible to cutaneous leishmaniasis, acquire TH1 cell-dependent resistance to L. major infection when C5aR is genetically ablated37. Microbial C5 convertase-like enzymes (such as the gingipains of P. gingivalis), which generate C5a, have been implicated in the selective downregulation of IL-12 production and the inhibition of IL-12-dependent immune clearance in vivo99. Similar evasive strategies, involving alternative complement receptors, are used by other pathogens. These include measles virus, human herpesvirus 6 and adenovirus (groups B and D), which all interact with CD46 through specific virulence proteins or ligands39,100–103 (see Supplementary information S1 (table)). Measles virus also inhibits TLR4-induced IL-12 production in DCs, although in this case the inhibitory signal is delivered by CD150 (also known as SLAM)104. In both cases, however, the virus uses its haemagglutinin to bind CD150 or CD46. Some other pathogens (such as N. gonorrhoeae, N. meningitidis and Group A Streptococcus)105 also use CD46 as a receptor, thus raising the possibility that these organisms also inhibit IL-12 production. Hepatitis C virus uses its core protein to bind gC1qR on macrophages or DCs and thereby inhibit IL-12 production and TH1 cell-mediated immunity; this is thought to be an important mechanism whereby the virus can establish persistent infections38. This evasion mechanism might be shared by other pathogens, such as L. monocytogenes and S. aureus, which can also interact with gC1qR using specific virulence proteins106,107. Mycobacteria can downregulate TLR4-induced IL-12 production through the mannose receptor, although this mechanism has broader anti-inflammatory effects108,109.

nATuRE REvIEwS | Immunology

vOLuME 11 | MARCH 2011 | 195 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Several microorganisms express virulence proteins that interact directly (in a non-opsonic manner) with CR3, although specific CR3-mediated inhibition of IL-12 production has been documented for only a few cases. These examples include Histoplasma capsulatum, B. pertussis and P. gingivalis 95,114–116, which use specific virulence proteins to inhibit IL-12 production (see Supplementary information S1 (table)). CR3-dependent immune subversion was confirmed in vivo for P. gingivalis, which uses its fimbriae to bind CR3, activate ERk1 and ERk2 signalling, and thereby inhibit TLR2-induced IL-12 production116 (FIG. 3). This allows P. gingivalis to survive in wild-type but not CR3-deficient mice (or normal mice in which CR3 is pharmacologically inhibited), as mice lacking CR3 activity produce higher levels of IL-12 and, secondarily, IFnγ. However, the host-protective effect of CR3 inhibition can be reversed by antibody-mediated neutralization of IL-12 (REF. 116).

K015

↑#4)

%'$2β 7PKPHGEVGF OCETQRJCIG

56#6

+. /[EQDCEVGTKC

+. )%5(

6.4

6.4

/;&

+PHGEVGF OCETQRJCIG

56#6 %'$2β ↑#4)

%'$2β

K015

↑+.+.)%5(

Figure 6 | myD88-dependent arginase induction prevents nitric oxide production in both infected and uninfected macrophages. The activation of myeloid differentiation primary response protein 88 (MyD88) signalling by 0CVWTG4GXKGYU^+OOWPQNQI[ mycobacteria (at least in part through Toll-like receptor 2 (TLR2)) induces CCAAT/ enhancer-binding protein-β (C/EBPβ)-mediated induction of interleukin-6 (IL-6), IL-10 and granulocyte colony-stimulating factor (G-CSF) production. These signal transducer and activator of transcription 3 (STAT3)-activating cytokines function in both autocrine and paracrine manners to induce arginase 1 (ARG1) expression, which is partially dependent on C/EBPβ. The ARG1 that is produced can inhibit inducible nitric oxide synthase (iNOS) activity through competition for their common substrate, arginine. The MyD88-dependent pathway for arginase production was shown to confer a survival benefit for mycobacteria in vivo and is thought to counteract pathways that activate nitric oxide production, such as TLR4 signalling.

Plasmodium falciparum selectively inhibits IL-12 production and suppresses DC maturation and T cell activation through interactions with the scavenger receptor CD36. Such interactions with CD36 are mediated by P. falciparum erythrocyte membrane protein 1 (PfEMP1), which is expressed on infected erythrocytes 110,111 and also interacts with TLR2 (REF. 112) . Although PfEMP1-modulated DCs secrete high levels of IL-10, their inability to produce IL-12 is an IL-10-independent effect 110,111. More recently, PfEMP1 was also implicated in specific suppression of the early induction of IFnγ production, although this involves a CD36-independent mechanism113. Moreover, the role of CD36 in malarial pathogenesis is complex and multifactorial, as suggested by a report that CD36-deficient mice have defective clearance of the parasite112.

TLR–TLR interplay The capacity of TLR signalling pathways for crossregulation9 could potentially be exploited by certain pathogens. This could be achieved through the induction of conflicting signals by distinct pathogen-expressed TLR ligands, and recent papers lend support to this concept. For example, M. tuberculosis expresses lipoproteins and glycolipids that function in an inhibitory mode through TLR2 to downregulate TLR9 signalling pathways. This decreases the production of IFnα and IFnβ in response to bacterial CpG DnA and impairs antigen cross-presentation on MHC class I molecules40. This crosstalk mechanism might explain why IFnα and IFnβ are not important factors in host immunity to mycobacteria40. Hepatitis C virus uses its core protein to activate TLR2-mediated production of inhibitory cytokines (such as IL-10) by human monocytes, and this suppresses TLR9-induced IFnα production by plasmacytoid DCs41. This mechanism involves transcellular crosstalk, as human plasmacytoid DCs express TLR9 but not TLR2. By contrast, mouse myeloid DCs, which were used in the study of M. tuberculosis40, express both TLR2 and TLR9; in this system, direct intracellular TLR2–TLR9 signalling crosstalk is mainly responsible for TLR2-mediated inhibition of TLR9-induced antibacterial immunity. M. tuberculosis and Toxoplasma gondii were shown to promote their survival and ability to cause disease in mouse models through MYD88-dependent induction of macrophage arginase 1 (ARG1), which inhibits nitric oxide production by macrophages by competing with inOS for the common substrate, arginine117. MYD88-mediated ARG1 expression in mycobacteria-infected macrophages depends, in part, on TLR2 activation, whereas strong induction of nitric oxide production is mediated by TLR4 activation117. In a follow-up paper, the same group showed that the TLR–MYD88-mediated expression of ARG1 is not a direct result of MYD88 signalling but is controlled by the MYD88-dependent production of IL-6, IL-10 and granulocyte colony-stimulating factor, which mediate their effects through signal transducer and activator of transcription 3 (STAT3)118 (FIG. 6). The implication of these findings is that ARG1 can be induced in a paracrine

196 | MARCH 2011 | vOLuME 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Table 1 | Microbial disruption of cooperative interactions between innate immune receptors Pathogen

Virulence molecules Crosstalking and their targets receptors R1

R2

Coxiella burnetii

Uncertain; possible involvement of smooth-type LPS, which is thought to target CD47

αVβ3–CD47

Group A Streptococcus

Mac, a CD11b mimetic, binds FcγRIII

FcγRIII

Filarial nematodes

Secreted glycoprotein TLR4 ES-62 forms a complex with TLR4

Cell type

Cellular response mechanism and outcome of without interference disruption

Ref

CR3

Monocytes

αVβ3–CD47-induced inside-out signalling activates CR3 (lectin site)-mediated phagocytosis, leading to intracellular killing*

The pathogen is taken up by αVβ3, leading to intracellular survival; the mechanism is unclear, but perhaps smooth-type LPS interferes with the co-signalling function of CD47

42

CR3

Neutrophils

Opsonophagocytosis, oxidative burst and killing

Mac blocks FcγRIII–CR3 interactions for outside-in signalling, thereby inhibiting the neutrophil antimicrobial response

43

FcεRI

Mast cells

FcεRI-mediated mast cell degranulation

The sequestration and degradation of PKCα, which is required for the coupling of FcεRI to PLD, results in the inhibition of mast cell activation

44

CR3, complement receptor 3 (also known as αMβ2 integrin or CD11b–CD18); FcγRIII, Fcγ receptor III (also known as CD16); FcεRI, Fcε receptor I; LPS, lipopolysaccharide; PKCα, protein kinase Cα; PLD, phospholipase D; R, receptor; TLR4, Toll-like receptor 4. *Refers to avirulent C. burnetii, which expresses rough-type, rather than smooth-type, LPS.

manner; therefore, mycobacteria can ‘instruct’ both infected and uninfected macrophages to decrease nitric oxide production, thereby rendering these cells permissive to the mycobacterial intracellular lifestyle. Interestingly, ARG1 can also be expressed by ‘alternatively activated’ macrophages through a STAT6-dependent pathway in the context of TH2 cell-mediated immunity 118. This might affect the inOS-dependent killing of pathogens such as Francisella tularensis, which induces the production of IL-4 and IL-13 and thus activates macrophages through the alternative pathway 119. C. albicans expresses ligands for both TLR2 and TLR4. whereas TLR4 signalling confers protection against infection 120, TLR2 signalling suppresses the capacity of macrophages to kill C. albicans and promotes host susceptibility to invasive candidiasis121. The immunosuppressive effect of TLR2 is mediated through the induction of high levels of IL-10 (REF. 121). Although a direct, cell-intrinsic TLR2–TLR4 inhibitory crosstalk pathway has not yet been identified, C. albicans seems to use TLR2 signalling to counteract potential TLR4-dependent immunity. Pathogenic Yersinia spp. also induce TLR2-dependent IL-10 production and cause immunosuppression by means of the secreted virulence protein Lcrv122. An intriguing question is how pathogen-stimulated TLR2 signalling can induce immunosuppressive levels of IL-10, given that several TLR2 ligands (including synthetic lipopeptides) induce an overall pro-inflammatory response. It is plausible that complex microbial structures activate TLR2 in tandem with functionally associated co-receptor(s), such as certain C-type lectins29,30,64, and that IL-10 production is actually induced by the resulting receptor crosstalk between TLR2 and a particular co-receptor. This notion is consistent with the inability of mycobacteria to induce high levels of IL-10 in neutrophils through the TLR–MYD88 pathway unless they activate a C-type lectin–SYk-dependent pathway in parallel30.

Disruption of cooperative receptor interactions Pathogens have evolved tactics to disrupt productive cooperation between certain innate immune receptors (TABLE 1). Integrins are an important target in this regard, owing to their capacity to engage in dynamic physical and/or functional interactions with several other receptors in lipid rafts123. Indeed, Coxiella burnetii impairs the crosstalk between the αvβ3 integrin– CD47 signalling complex and CR3 that is required for activation of CR3 (REF. 42). This prevents CR3-mediated uptake and post-phagocytic killing of this bacterium by monocytes. The exact mechanism is unclear, although it is thought that the smooth-type LPS of virulent C. burnetii interferes with CD47 signalling functions, which are not disrupted when monocytes are exposed to avirulent C. burnetii expressing rough-type LPS42. The avoidance of CR3-mediated phagocytosis by C. burnetii contrasts with other bacteria that voluntarily activate this uptake pathway (FIG. 4). However, the phagocytosis of C. burnetii is mediated by the carboxyterminal lectin sites of CR3 which, unlike the aminoterminal I domain used by P. gingivalis, E. faecalis and other CR3-exploitative bacteria, are linked to induction of the oxidative burst 124,125. Group A Streptococcus secretes a CD11b mimetic, Mac, that binds FcγRIII and blocks its productive interaction with CR3. The disruption of this functional association inhibits cooperative outside-in signalling and impairs opsonophagocytosis, the oxidative burst and bacterial killing 43. Pathogens may also have developed ways to disrupt productive interactions between non-integrin receptors. TLRs are functionally linked to FcRs, and this has important implications for immunity and inflammatory pathology 8,126. Filarial nematodes express a secreted glycoprotein, aminopeptidase ES-62, that forms a complex with TLR4, leading to the sequestration and degradation of protein kinase Cα. This disrupts the coupling of FcεRI to phospholipase D and thus prevents mast cell

nATuRE REvIEwS | Immunology

vOLuME 11 | MARCH 2011 | 197 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS degranulation44. It has not been specifically addressed whether this mechanism is used by the parasite to evade mast cell-mediated immunity, although ES-62 could be exploited as a potential therapeutic in allergy. Several viruses and parasites encode soluble molecules that mimic host receptors. Such molecules include the myxoma virus M-T7 glycoprotein, which scavenges IFnγ and sequesters C-, CC- and CXC-chemokines, and the Schistosoma mansoni chemokine-binding protein, which binds CC-, CXC- and CX3C-chemokines127. These and other decoy receptors contribute to immune evasion by preventing the interaction of cytokines or chemokines with their signalling receptors127. It is conceivable — although not yet specifically addressed — that such decoys disrupt crosstalk interactions between the affected chemokine receptors and PRRs.

Concluding remarks and future perspectives Receptor crosstalk in the innate immune system is crucial to coordinate microorganism-sensing signals and allow the host to tailor an appropriate immune response. However, many pathogens subvert these functions, often by taking control of host regulatory receptors. This can be achieved by mimicking host ligands or counter-receptors24,28,43, or the host enzymes that generate such ligands31,74. Moreover, pathogens

Kawai, T. & Akira, S. The role of pattern-recognition receptors in innate immunity: update on Toll-like receptors. Nature Immunol. 11, 373–384 (2010). 2. Ricklin, D., Hajishengallis, G., Yang, K. & Lambris, J. D. Complement: a key system for immune surveillance and homeostasis. Nature Immunol. 11, 785–797 (2010). A comprehensive review of the complement system, summarizing recent and emerging evidence that complement engages in reciprocal crosstalk interactions with other immune and physiological systems (such as TLRs and coagulation), aiming to fine-tune the host response to infection and other insults. 3. Medzhitov, R. Toll-like receptors and innate immunity. Nature Rev. Immunol. 1, 135–145 (2001). 4. Beutler, B. A. TLRs and innate immunity. Blood 113, 1399–1407 (2009). 5. Triantafilou, K., Triantafilou, M. & Dedrick, R. L. A CD14-independent LPS receptor cluster. Nature Immunol. 2, 338–345 (2001). 6. Hajishengallis, G. et al. Differential interactions of fimbriae and lipopolysaccharide from Porphyromonas gingivalis with the Toll-like receptor 2-centred pattern recognition apparatus. Cell. Microbiol. 8, 1557–1570 (2006). 7. Natarajan, M., Lin, K. M., Hsueh, R. C., Sternweis, P. C. & Ranganathan, R. A global analysis of cross-talk in a mammalian cellular signalling network. Nature Cell Biol. 8, 571–580 (2006). A systematic analysis of intracellular signalling crosstalk, showing that a large number of signalling pathways converge on a relatively limited set of interaction mechanisms, including both synergistic and antagonistic interactions. 8. Ivashkiv, L. B. Cross-regulation of signaling by ITAMassociated receptors. Nature Immunol. 10, 340–347 (2009). 9. Lee, M. S. & Kim, Y. J. Signaling pathways downstream of pattern-recognition receptors and their cross talk. Annu. Rev. Biochem. 76, 447–480 (2007). 10. Zak, D. E. & Aderem, A. Systems biology of innate immunity. Immunol. Rev. 227, 264–282 (2009). 11. Hajishengallis, G. & Lambris, J. D. Crosstalk pathways between Toll-like receptors and the complement system. Trends Immunol. 31, 154–163 (2010). 1.

can ‘voluntarily’ interact with TLRs (or other innate immune receptors) through virulence molecules that have evolved to recognize such receptors25,32,122,128,129. It is clear that these interactions do not represent pattern recognition but rather involve subversive action that promotes the adaptive fitness of the pathogens. Future research may identify additional pathogen virulence factors and hijacked host receptors, particularly among those receptors with regulatory roles. For example, it would provide a significant survival advantage if a pathogen has evolved to interact with regulatory Toll/IL-1 receptor domain-containing transmembrane proteins (such as single immunoglobulin IL-1-related receptor (SIGIRR) and IL-1 receptor-like 1 (IL1RL1; also known as ST2)), which are involved in negative crosstalk with the IL-1 receptor and/or TLRs9. Some of the evasion mechanisms discussed in this Review need to be substantiated in appropriate animal models, which are not always available. Identification of the mechanisms of ‘crosstalk manipulation’, through continued research and improved animal models, is essential in order to develop approaches that counteract immune subversion by pathogens. Indeed, antagonistic blockade of hijacked host receptors or neutralization of the virulence factors involved might offer promising options for controlling infection and associated immunopathology.

12. Goodridge, H. S. & Underhill, D. M. Fungal recognition by TLR2 and dectin-1. Handb. Exp. Pharmacol. 183, 87–109 (2008). 13. Ogawa, S. et al. Molecular determinants of crosstalk between nuclear receptors and Toll-like receptors. Cell 122, 707–721 (2005). 14. Lukashev, D., Ohta, A., Apasov, S., Chen, J. F. & Sitkovsky, M. Cutting edge: physiologic attenuation of proinflammatory transcription by the Gs proteincoupled A2A adenosine receptor in vivo. J. Immunol. 173, 21–24 (2004). 15. Lambris, J. D., Ricklin, D. & Geisbrecht, B. V. Complement evasion by human pathogens. Nature Rev. Microbiol. 6, 132–142 (2008). 16. Finlay, B. B. & McFadden, G. Anti-immunology: evasion of the host immune system by bacterial and viral pathogens. Cell 124, 767–782 (2006). 17. Bowie, A. G. & Unterholzner, L. Viral evasion and subversion of pattern-recognition receptor signalling. Nature Rev. Immunol. 8, 911–922 (2008). 18. Diacovich, L. & Gorvel, J. P. Bacterial manipulation of innate immunity to promote infection. Nature Rev. Microbiol. 8, 117–128 (2010). 19. Flannagan, R. S., Cosio, G. & Grinstein, S. Antimicrobial mechanisms of phagocytes and bacterial evasion strategies. Nature Rev. Microbiol. 7, 355–366 (2009). 20. Sansonetti, P. J. & Di Santo, J. P. Debugging how bacteria manipulate the immune response. Immunity 26, 149–161 (2007). 21. Brodsky, I. E. & Medzhitov, R. Targeting of immune signalling networks by bacterial pathogens. Nature Cell Biol. 11, 521–526 (2009). 22. Bhavsar, A. P., Guttman, J. A. & Finlay, B. B. Manipulation of host-cell pathways by bacterial pathogens. Nature 449, 827–834 (2007). 23. Roy, C. R. & Mocarski, E. S. Pathogen subversion of cell-intrinsic innate immunity. Nature Immunol. 8, 1179–1187 (2007). 24. Carlin, A. F. et al. Molecular mimicry of host sialylated glycans allows a bacterial pathogen to engage neutrophil Siglec-9 and dampen the innate immune response. Blood 113, 3333–3336 (2009). 25. Carlin, A. F. et al. Group B Streptococcus suppression of phagocyte functions by protein-mediated engagement of human Siglec-5. J. Exp. Med. 206, 1691–1699 (2009).

198 | MARCH 2011 | vOLuME 11

26. Slevogt, H. et al. CEACAM1 inhibits Toll-like receptor 2-triggered antibacterial responses of human pulmonary epithelial cells. Nature Immunol. 9, 1270–1278 (2008). 27. Pinheiro da Silva, F. et al. CD16 promotes Escherichia coli sepsis through an FcRγ inhibitory pathway that prevents phagocytosis and facilitates inflammation. Nature Med. 13, 1368–1374 (2007). This paper provided the first demonstration that ITAMi can be exploited by bacteria, in this case to induce a crosstalk that inhibits phagocytosis, leading to uncontrolled E. coli infection and sepsis. 28. Reyburn, H. T. et al. The class I MHC homologue of human cytomegalovirus inhibits attack by natural killer cells. Nature 386, 514–517 (1997). 29. Gringhuis, S. I. et al. C-type lectin DC-SIGN modulates Toll-like receptor signaling via Raf-1 kinase-dependent acetylation of transcription factor NF-κB. Immunity 26, 605–616 (2007). 30. Zhang, X., Majlessi, L., Deriaud, E., Leclerc, C. & Lo-Man, R. Coactivation of Syk kinase and MyD88 adaptor protein pathways by bacteria promotes regulatory properties of neutrophils. Immunity 31, 761–771 (2009). References 29 and 30 identified TLR co-receptors (C-type lectins) and elucidated the underlying signalling pathways that explain how inflammatory TLRs can induce high levels of the anti-inflammatory cytokine IL-10. 31. Wang, M. et al. Microbial hijacking of complement– Toll-like receptor crosstalk. Sci. Signal. 3, ra11 (2010). This study was the first to show that anaphylatoxin generation by a virulence enzyme is exploited by the pathogen to induce subversive complement–TLR crosstalk that impairs protective immunity. 32. Hajishengallis, G., Wang, M., Liang, S., Triantafilou, M. & Triantafilou, K. Pathogen induction of CXCR4/ TLR2 cross-talk impairs host defense function. Proc. Natl Acad. Sci. USA 105, 13532–13537 (2008). 33. Oliva, C., Turnbough, C. L. Jr & Kearney, J. F. CD14–Mac-1 interactions in Bacillus anthracis spore internalization by macrophages. Proc. Natl Acad. Sci. USA 106, 13957–13962 (2009).

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 34. Wang, M. et al. Fimbrial proteins of Porphyromonas gingivalis mediate in vivo virulence and exploit TLR2 and complement receptor 3 to persist in macrophages. J. Immunol. 179, 2349–2358 (2007). 35. Sumuth, S. D. et al. Aggregation substance promotes adherence, phagocytosis, and intracellular survival of Enterococcus faecalis within human macrophages and suppresses respiratory burst. Infect. Immun. 68, 4900–4906 (2000). 36. Ishibashi, Y., Claus, S. & Relman, D. A. Bordetella pertussis filamentous hemagglutinin interacts with a leukocyte signal transduction complex and stimulates bacterial adherence to monocyte CR3 (CD11b/CD18). J. Exp. Med. 180, 1225–1233 (1994). 37. Hawlisch, H. et al. C5a negatively regulates Toll-like receptor 4-induced immune responses. Immunity 22, 415–426 (2005). 38. Waggoner, S. N., Hall, C. H. & Hahn, Y. S. HCV core protein interaction with gC1q receptor inhibits Th1 differentiation of CD4+ T cells via suppression of dendritic cell IL-12 production. J. Leukoc. Biol. 82, 1407–1419 (2007). 39. Karp, C. L. et al. Mechanism of suppression of cellmediated immunity by measles virus. Science 273, 228–231 (1996). The first paper to show that a complement receptor (CD46) can downregulate IL-12 induction by TLR agonists, even though the concept of mammalian TLRs had not yet been established. 40. Simmons, D. P. et al. Mycobacterium tuberculosis and TLR2 agonists inhibit induction of type I IFN and class I MHC antigen cross processing by TLR9. J. Immunol. 185, 2405–2415 (2010). 41. Dolganiuc, A. et al. Hepatitis C virus (HCV) core protein-induced, monocyte-mediated mechanisms of reduced IFN-α and plasmacytoid dendritic cell loss in chronic HCV infection. J. Immunol. 177, 6758–6768 (2006). 42. Capo, C. et al. Subversion of monocyte functions by Coxiella burnetii: impairment of the cross-talk between αvβ3 integrin and CR3. J. Immunol. 163, 6078–6085 (1999). 43. Lei, B. et al. Evasion of human innate and acquired immunity by a bacterial homolog of CD11b that inhibits opsonophagocytosis. Nature Med. 7, 1298–1305 (2001). This study presents a novel concept according to which a virulence protein that mimics an innate immune receptor disrupts its cooperative interactions with another receptor, resulting in inhibition of crucial antimicrobial responses. 44. Melendez, A. J. et al. Inhibition of FcεRI-mediated mast cell responses by ES-62, a product of parasitic filarial nematodes. Nature Med. 13, 1375–1381 (2007). 45. Turnbull, I. R. & Colonna, M. Activating and inhibitory functions of DAP12. Nature Rev. Immunol. 7, 155–161 (2007). 46. Wang, L. et al. Indirect inhibition of Toll-like receptor and type I interferon responses by ITAM-coupled receptors and integrins. Immunity 32, 518–530 (2010). 47. Pinheiro da Silva, F., Aloulou, M., Benhamou, M. & Monteiro, R. C. Inhibitory ITAMs: a matter of life and death. Trends Immunol. 29, 366–373 (2008). 48. Boyd, C. R. et al. Siglec-E is up-regulated and phosphorylated following lipopolysaccharide stimulation in order to limit TLR-driven cytokine production. J. Immunol. 183, 7703–7709 (2009). 49. Nakayama, M. et al. Paired Ig-like receptors bind to bacteria and shape TLR-mediated cytokine production. J. Immunol. 178, 4250–4259 (2007). 50. Chapman, T. L., Heikeman, A. P. & Bjorkman, P. J. The inhibitory receptor LIR-1 uses a common binding interaction to recognize class I MHC molecules and the viral homolog UL18. Immunity 11, 603–613 (1999). 51. Lambert, A. A., Gilbert, C., Richard, M., Beaulieu, A. D. & Tremblay, M. J. The C-type lectin surface receptor DCIR acts as a new attachment factor for HIV-1 in dendritic cells and contributes to trans- and cisinfection pathways. Blood 112, 1299–1307 (2008). 52. Meyer-Wentrup, F. et al. DCIR is endocytosed into human dendritic cells and inhibits TLR8-mediated cytokine production. J. Leukoc. Biol. 85, 518–525 (2009). 53. Meyer-Wentrup, F. et al. Targeting DCIR on human plasmacytoid dendritic cells results in antigen presentation and inhibits IFN-α production. Blood 111, 4245–4253 (2008).

54. Cerf-Bensussan, N. & Gaboriau-Routhiau, V. The immune system and the gut microbiota: friends or foes? Nature Rev. Immunol. 10, 735–744 (2010). 55. Redpath, S., Ghazal, P. & Gascoigne, N. R. Hijacking and exploitation of IL-10 by intracellular pathogens. Trends Microbiol. 9, 86–92 (2001). 56. Confer, D. & Eaton, J. Phagocyte impotence caused by an invasive bacterial adenylate cyclase. Science 217, 948–950 (1982). 57. Turk, B. E. Manipulation of host signalling pathways by anthrax toxins. Biochem. J. 402, 405–417 (2007). 58. Vojtova, J., Kamanova, J. & Sebo, P. Bordetella adenylate cyclase toxin: a swift saboteur of host defense. Curr. Opin. Microbiol. 9, 69–75 (2006). 59. Conti, P. et al. IL-10, an inflammatory/inhibitory cytokine, but not always. Immunol. Lett. 86, 123–129 (2003). 60. Lavelle, E. C. et al. Effects of cholera toxin on innate and adaptive immunity and its application as an immunomodulatory agent. J. Leukoc. Biol. 75, 756–763 (2004). 61. Brenner, S. et al. cAMP-induced interleukin-10 promoter activation depends on CCAAT/enhancerbinding protein expression and monocytic differentiation. J. Biol. Chem. 278, 5597–5604 (2003). 62. Geijtenbeek, T. B. et al. Mycobacteria target DC-SIGN to suppress dendritic cell function. J. Exp. Med. 197, 7–17 (2003). 63. Bergman, M. P. et al. Helicobacter pylori modulates the T helper cell 1/T helper cell 2 balance through phase-variable interaction between lipopolysaccharide and DC-SIGN. J. Exp. Med. 200, 979–990 (2004). 64. Gringhuis, S. I., den Dunnen, J., Litjens, M., van der Vlist, M. & Geijtenbeek, T. B. Carbohydratespecific signaling through the DC-SIGN signalosome tailors immunity to Mycobacterium tuberculosis, HIV-1 and Helicobacter pylori. Nature Immunol. 10, 1081–1088 (2009). This paper characterizes the composition of the DC-SIGN signalling complex, which depends on the sugar specificity of the activating DC-SIGN ligand; in turn, this determines the nature of the induced signalling pathway. 65. Madura Larsen, J. et al. BCG stimulated dendritic cells induce an interleukin-10 producing T-cell population with no T helper 1 or T helper 2 bias in vitro. Immunology 121, 276–282 (2007). 66. Dulphy, N. et al. Intermediate maturation of Mycobacterium tuberculosis LAM-activated human dendritic cells. Cell. Microbiol. 9, 1412–1425 (2007). 67. Hovius, J. W. et al. Salp15 binding to DC-SIGN inhibits cytokine expression by impairing both nucleosome remodeling and mRNA stabilization. PLoS Pathog. 4, e31 (2008). This paper presents the interesting concept of subversive receptor crosstalk being mediated by a pathogen and its arthropod vector, which contribute distinct agonists for TLRs and DC-SIGN, respectively. 68. Lundberg, K., Wegner, N., Yucel-Lindberg, T. & Venables, P. J. Periodontitis in RA — the citrullinated enolase connection. Nature Rev. Rheumatol. 6, 727–730 (2010). 69. Pihlstrom, B. L., Michalowicz, B. S. & Johnson, N. W. Periodontal diseases. Lancet 366, 1809–1820 (2005). 70. Hajishengallis, G. Porphyromonas gingivalis–host interactions: open war or intelligent guerilla tactics? Microbes Infect. 11, 637–645 (2009). 71. Pierce, D. L. et al. Host adhesive activities and virulence of novel fimbrial proteins of Porphyromonas gingivalis. Infect. Immun. 77, 3294–3301 (2009). 72. Hajishengallis, G., Ratti, P. & Harokopakis, E. Peptide mapping of bacterial fimbrial epitopes interacting with pattern recognition receptors. J. Biol. Chem. 280, 38902–38913 (2005). 73. Darveau, R. P. Periodontitis: a polymicrobial disruption of host homeostasis. Nature Rev. Microbiol. 8, 481–490 (2010). 74. Thammavongsa, V., Kern, J. W., Missiakas, D. M. & Schneewind, O. Staphylococcus aureus synthesizes adenosine to escape host immune responses. J. Exp. Med. 206, 2417–2427 (2009). 75. Zarek, P. E. et al. A2A receptor signaling promotes peripheral tolerance by inducing T-cell anergy and the generation of adaptive regulatory T cells. Blood 111, 251–259 (2008). 76. Abram, C. L. & Lowell, C. A. The ins and outs of leukocyte integrin signaling. Annu. Rev. Immunol. 27, 339–362 (2009).

nATuRE REvIEwS | Immunology

77. Harokopakis, E. & Hajishengallis, G. Integrin activation by bacterial fimbriae through a pathway involving CD14, Toll-like receptor 2, and phosphatidylinositol-3-kinase. Eur. J. Immunol. 35, 1201–1210 (2005). 78. Sendide, K. et al. Cross-talk between CD14 and complement receptor 3 promotes phagocytosis of mycobacteria: regulation by phosphatidylinositol 3-kinase and cytohesin-1. J. Immunol. 174, 4210–4219 (2005). References 77 and 78 were the first to describe TLR signalling pathways that transactivate integrins through inside-out signalling. 79. Han, C. et al. Integrin CD11b negatively regulates TLR-triggered inflammatory responses by activating Syk and promoting degradation of MyD88 and TRIF via Cbl-b. Nature Immunol. 11, 734–742 (2010). 80. Harokopakis, E., Albzreh, M. H., Martin, M. H. & Hajishengallis, G. TLR2 transmodulates monocyte adhesion and transmigration via Rac1- and PI3K-mediated inside-out signaling in response to Porphyromonas gingivalis fimbriae. J. Immunol. 176, 7645–7656 (2006). 81. Hajishengallis, G., Wang, M. & Liang, S. Induction of distinct TLR2-mediated proinflammatory and proadhesive signaling pathways in response to Porphyromonas gingivalis fimbriae. J. Immunol. 182, 6690–6696 (2009). 82. Arbibe, L. et al. Toll-like receptor 2-mediated NF-κB activation requires a Rac1-dependent pathway. Nature Immunol. 1, 533–540 (2000). 83. Lowell, C. A. Rewiring phagocytic signal transduction. Immunity 24, 243–245 (2006). 84. Caron, E. & Hall, A. Identification of two distinct mechanisms of phagocytosis controlled by different Rho GTPases. Science 282, 1717–1721 (1998). 85. Gatfield, J. & Pieters, J. Essential role for cholesterol in entry of mycobacteria into macrophages. Science 288, 1647–1650 (2000). 86. Hajishengallis, G., Wang, M., Harokopakis, E., Triantafilou, M. & Triantafilou, K. Porphyromonas gingivalis fimbriae proactively modulate β2 integrin adhesive activity and promote binding to and internalization by macrophages. Infect. Immun. 74, 5658–5666 (2006). 87. Vanek, N. N. et al. Enterococcus faecalis aggregation substance promotes opsonin-independent binding to human neutrophils via a complement receptor type 3-mediated mechanism. FEMS Immunol. Med. Microbiol. 26, 49–60 (1999). 88. Rakita, R. M. et al. Enterococcus faecalis bearing aggregation substance is resistant to killing by human neutrophils despite phagocytosis and neutrophil activation. Infect. Immun. 67, 6067–6075 (1999). 89. Hellwig, S. M. et al. Targeting to Fcγ receptors, but not CR3 (CD11b/CD18), increases clearance of Bordetella pertussis. J. Infect. Dis. 183, 871–879 (2001). An elegant in vivo demonstration of the differential outcomes of FcγR- and CR3-mediated phagocytosis, with the latter being relatively ineffective in bacterial clearance; this explains why several pathogens proactively induce their uptake by CR3. 90. Zhang, X. et al. Regulation of Toll-like receptormediated inflammatory response by complement in vivo. Blood 110, 228–236 (2007). 91. Lappegård, K. T. et al. Human genetic deficiencies reveal the roles of complement in the inflammatory network: lessons from nature. Proc. Natl Acad. Sci. USA 106, 15861–15866 (2009). 92. Fan, J. & Malik, A. B. Toll-like receptor-4 (TLR4) signaling augments chemokine-induced neutrophil migration by modulating cell surface expression of chemokine receptors. Nature Med. 9, 315–321 (2003). 93. la Sala, A., Gadina, M. & Kelsall, B. L. Gi-protein-dependent inhibition of IL-12 production is mediated by activation of the phosphatidylinositol 3-kinase-protein 3 kinase B/Akt pathway and JNK. J. Immunol. 175, 2994–2999 (2005). 94. Waggoner, S. N., Cruise, M. W., Kassel, R. & Hahn, Y. S. gC1q receptor ligation selectively downregulates human IL-12 production through activation of the phosphoinositide 3-kinase pathway. J. Immunol. 175, 4706–4714 (2005). 95. Marth, T. & Kelsall, B. L. Regulation of interleukin-12 by complement receptor 3 signaling. J. Exp. Med. 185, 1987–1995 (1997). 96. Goriely, S., Neurath, M. F. & Goldman, M. How microorganisms tip the balance between interleukin-12 family members. Nature Rev. Immunol. 8, 81–86 (2008).

vOLuME 11 | MARCH 2011 | 199 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 97. Manderson, A. P., Botto, M. & Walport, M. J. The role of complement in the development of systemic lupus erythematosus. Annu. Rev. Immunol. 22, 431–456 (2004). 98. Sutterwala, F. S., Noel, G. J., Clynes, R. & Mosser, D. M. Selective suppression of interleukin-12 induction after macrophage receptor ligation. J. Exp. Med. 185, 1977–1985 (1997). 99. Liang, S. et al. The C5a receptor impairs IL-12-dependent clearance of Porphyromonas gingivalis and is required for induction of periodontal bone loss. J. Immunol. 186, 869–877 (2011). 100. Smith, A. et al. Selective suppression of IL-12 production by human herpesvirus 6. Blood 102, 2877–2884 (2003). 101. Santoro, F. et al. Interaction of glycoprotein H of human herpesvirus 6 with the cellular receptor CD46. J. Biol. Chem. 278, 25964–25969 (2003). 102. Iacobelli-Martinez, M., Nepomuceno, R. R., Connolly, J. & Nemerow, G. R. CD46-utilizing adenoviruses inhibit C/EBPβ-dependent expression of proinflammatory cytokines. J. Virol. 79, 11259–11268 (2005). 103. Gaggar, A., Shayakhmetov, D. M. & Lieber, A. CD46 is a cellular receptor for group B adenoviruses. Nature Med. 9, 1408–1412 (2003). 104. Hahm, B., Cho, J. H. & Oldstone, M. B. Measles virus–dendritic cell interaction via SLAM inhibits innate immunity: selective signaling through TLR4 but not other TLRs mediates suppression of IL-12 synthesis. Virology 358, 251–257 (2007). 105. Gill, D. B. & Atkinson, J. P. CD46 in Neisseria pathogenesis. Trends Mol. Med. 10, 459–465 (2004). 106. Braun, L., Ghebrehiwet, B. & Cossart, P. gC1q-R/p32, a C1q-binding protein, is a receptor for the InlB invasion protein of Listeria monocytogenes. EMBO J. 19, 1458–1466 (2000). 107. Nguyen, T., Ghebrehiwet, B. & Peerschke, E. I. B. Staphylococcus aureus protein A recognizes platelet gC1qR/p33: a novel mechanism for staphylococcal interactions with platelets. Infect. Immun. 68, 2061–2068 (2000). 108. Nigou, J., Zelle-Rieser, C., Gilleron, M., Thurnher, M. & Puzo, G. Mannosylated lipoarabinomannans inhibit IL-12 production by human dendritic cells: evidence for a negative signal delivered through the mannose receptor. J. Immunol. 166, 7477–7485 (2001). 109. Chieppa, M. et al. Cross-linking of the mannose receptor on monocyte-derived dendritic cells activates an anti-inflammatory immunosuppressive program. J. Immunol. 171, 4552–4560 (2003). 110. Urban, B. C., Willcox, N. & Roberts, D. J. A role for CD36 in the regulation of dendritic cell function. Proc. Natl Acad. Sci. USA 98, 8750–8755 (2001).

111. Urban, B. C. et al. Plasmodium falciparum-infected erythrocytes modulate the maturation of dendritic cells. Nature 400, 73–77 (1999). 112. Patel, S. N. et al. Disruption of CD36 impairs cytokine response to Plasmodium falciparum glycosylphosphatidylinositol and confers susceptibility to severe and fatal malaria in vivo. J. Immunol. 178, 3954–3961 (2007). 113. D’Ombrain, M. C. et al. Plasmodium falciparum erythrocyte membrane protein-1 specifically suppresses early production of host interferon-γ. Cell Host Microbe 2, 130–138 (2007). 114. Long, K. H., Gomez, F. J., Morris, R. E. & Newman, S. L. Identification of heat shock protein 60 as the ligand on Histoplasma capsulatum that mediates binding to CD18 receptors on human macrophages. J. Immunol. 170, 487–494 (2003). 115. McGuirk, P. & Mills, K. H. Direct anti-inflammatory effect of a bacterial virulence factor: IL-10-dependent suppression of IL-12 production by filamentous hemagglutinin from Bordetella pertussis. Eur. J. Immunol. 30, 415–422 (2000). 116. Hajishengallis, G., Shakhatreh, M.-A. K., Wang, M. & Liang, S. Complement receptor 3 blockade promotes IL-12-mediated clearance of Porphyromonas gingivalis and negates its virulence in vivo. J. Immunol. 179, 2359–2367 (2007). 117. El Kasmi, K. C. et al. Toll-like receptor-induced arginase 1 in macrophages thwarts effective immunity against intracellular pathogens. Nature Immunol. 9, 1399–1406 (2008). 118. Qualls, J. E. et al. Arginine usage in mycobacteriainfected macrophages depends on autocrine– paracrine cytokine signaling. Sci. Signal. 3, ra62 (2010). 119. Shirey, K. A., Cole, L. E., Keegan, A. D. & Vogel, S. N. Francisella tularensis live vaccine strain induces macrophage alternative activation as a survival mechanism. J. Immunol. 181, 4159–4167 (2008). 120. Netea, M. G. et al. The role of Toll-like receptor (TLR) 2 and TLR4 in the host defense against disseminated candidiasis. J. Infect. Dis. 185, 1483–1489 (2002). 121. Netea, M. G. et al. Toll-like receptor 2 suppresses immunity against Candida albicans through induction of IL-10 and regulatory T cells. J. Immunol. 172, 3712–3718 (2004). 122. Sing, A. et al. A hypervariable N-terminal region of Yersinia LcrV determines Toll-like receptor 2-mediated IL-10 induction and mouse virulence. Proc. Natl Acad. Sci. USA 102, 16049–16054 (2005). This study provides molecular evidence that virulence proteins might have evolved to interact with and exploit TLRs; this exploitative interaction contrasts with the concept of ‘pattern recognition’, which aims to promote innate

200 | MARCH 2011 | vOLuME 11

immunity through TLR-mediated recognition of conserved microbial structures. 123. Petty, H. R., Worth, R. G. & Todd, R. F. Interactions of integrins with their partner proteins in leukocyte membranes. Immunol. Res. 25, 75–95 (2002). 124. Le Cabec, V., Cols, C. & Maridonneau-Parini, I. Nonopsonic phagocytosis of zymosan and Mycobacterium kansasii by CR3 (CD11b/CD18) involves distinct molecular determinants and is or is not coupled with NADPH oxidase activation. Infect. Immun. 68, 4736–4745 (2000). 125. Xia, Y. et al. Function of the lectin domain of Mac-1/complement receptor type 3 (CD11b/CD18) in regulating neutrophil adhesion. J. Immunol. 169, 6417–6426 (2002). 126. Rittirsch, D. et al. Cross-talk between TLR4 and FcγReceptorIII (CD16) pathways. PLoS Pathog. 5, e1000464 (2009). 127. Mantovani, A., Bonecchi, R. & Locati, M. Tuning inflammation and immunity by chemokine sequestration: decoys and more. Nature Rev. Immunol. 6, 907–918 (2006). 128. Pathak, S. K. et al. Direct extracellular interaction between the early secreted antigen ESAT-6 of Mycobacterium tuberculosis and TLR2 inhibits TLR signaling in macrophages. Nature Immunol. 8, 610–618 (2007). 129. Postma, B. et al. Chemotaxis inhibitory protein of Staphylococcus aureus binds specifically to the C5a and formylated peptide receptor. J. Immunol. 172, 6994–7001 (2004).

Acknowledgements

The authors regret that several important studies could only be cited indirectly through comprehensive reviews, owing to space and reference number limitations. Work in the authors’ laboratories is supported by US Public Health Service Grants DE015254, DE017138, DE018292 and DE021580 (to G.H.) and CA112162, AI68730, AI30040, AI72106, EB3968 and GM62134 (to J.D.L.).

Competing interests statement

The authors declare competing financial interests: see Web version for details.

FURTHER INFORMATION George Hajishengallis’s homepage: www.louisville.edu/ dental/ohr/faculty-staff/george-hajishengallis.html John D. Lambris’s homepage: www.lambris.com

SUPPLEMENTARY INFORMATION See online article: S1 (table) All lInkS ARe ACTIVe In The onlIne PDf

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

Immune cell regulation by autocrine purinergic signalling Wolfgang G. Junger*‡

Abstract | Stimulation of almost all mammalian cell types leads to the release of cellular ATP and autocrine feedback through a diverse array of purinergic receptors. Depending on the types of purinergic receptors that are involved, autocrine signalling can promote or inhibit cell activation and fine-tune functional responses. Recent work has shown that autocrine signalling is an important checkpoint in immune cell activation and allows immune cells to adjust their functional responses based on the extracellular cues provided by their environment. This Review focuses on the roles of autocrine purinergic signalling in the regulation of both innate and adaptive immune responses and discusses the potential of targeting purinergic receptors for treating immune-mediated disease. Immune synapse A large junctional structure that is formed at the cell surface between a T cell and an antigen-presenting cell. It is also known as the supramolecular activation cluster. Important molecules that are involved in T cell activation — including the T cell receptor, numerous signal-transduction molecules and molecular adaptors — accumulate in an orderly manner at this site. Immune synapses are now known to also form between other types of immune cells; for example, between dendritic cells and natural killer cells.

*Department of Surgery, Beth Israel Deaconess Medical Center, Harvard Medical School, 330 Brookline Avenue, Boston, Massachusetts 02215, USA. ‡ Ludwig Boltzmann Institute for Traumatology, Vienna, A‑1200 Austria. e‑mail: [email protected] doi:10.1038/nri2938 Published online 18 February 2011

Cellular ATP serves as an energy carrier that drives virtually all cell functions. Therefore, the discovery that intact cells can release a portion of their cellular ATP came as a surprise to most researchers1. Over the last 2 decades, a total of 19 different purinergic receptor subtypes that can recognize extracellular ATP and adenosine have been cloned and characterized2. These receptors include eight P2Y receptor subtypes, seven P2X receptor subtypes, and four P1 (adenosine) receptor subtypes. In addition, several families of ectonucleotidases that hydrolyse ATP to ADP, AMP and adenosine have been found3. Distinct sets of these purinergic receptors and ectonucleotidases are expressed on the cell surface of different mammalian cell types, where they regulate cell activation through cell-type specific purinergic signalling systems4,5. Controlled ATP release from intact cells was first discovered in neurons, which release ATP into neuronal synapses. Since then, many aspects of purinergic signalling in neurons have been elucidated 6. Additional work revealed that similar purinergic signalling mechanisms regulate key aspects of many other physiological processes, including activation of the different cell types of the immune system7. For example, T cell activation induces the release of ATP through pannexin 1 channels that translocate with P2X receptors to the immune synapse, where they promote calcium influx and cell activation through autocrine purinergic signalling 8–11. Neutrophils release ATP in response to chemotactic mediators, and autocrine signalling through purinergic receptors regulates the

chemotaxis of these cells12. Activation of purinergic receptors in immune cells can elicit either positive or negative feedback mechanisms and thus can tightly regulate immune responses. In addition to the autocrine feedback mechanisms that regulate the function of healthy immune cells, purinergic receptors allow immune cells to recognize the ATP that is released from damaged or stressed host cells. Thus, the purinergic signalling systems of immune cells serve an important function in the recognition of ‘danger’ signals, and ATP that is released by stressed cells acts as a ‘ find-me signal’ that guides phagocytes to inflammatory sites and promotes clearance of damaged and apoptotic cells 13. Purinergic signalling is also crucial for the activation of inflammasomes and the subsequent release of cytokines, such as interleukin-1β (IL-1β), in response to damage-associated molecular patterns (DAMPs) and pathogen-associated molecular patterns (PAMPs)14. Several excellent review articles have been published that describe in detail the mechanisms by which mammalian cells release ATP15, the pharmacological and structural properties of the different purinergic receptors16–17 and ectonucleotidases18, and the multiple roles for paracrine purinergic signalling in regulating a wide range of physiological processes, including immune cell functions5,7. This Review will therefore mainly focus on the roles of autocrine purinergic signalling systems in immune cell activation (FIG. 1) and how these purinergic systems integrate extracellular cues, such as danger signals from inflamed tissues.

NATuRe RevIewS | Immunology

vOLuMe 11 | MARCh 2011 | 201 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 'ZVTCEGNNWNCT

#FGPQUKPG TGWRVCMG

2UKIPCNNKPI #62 J[FTQN[UKU

#62TGNGCUG

#/2 J[FTQN[UKU

#&2

2UKIPCNNKPI

#FGPQUKPG J[FTQN[UKU

#FGPQUKPG

#62

+PQUKPG

#/2

H/.2 %GNN OGODTCPG

+PVTCEGNNWNCT

(24

2CPPGZKP

2:

2;

'062&U

Find-me signal A signal emitted by dying cells to promote the recruitment of scavenger cells, which clear the apoptotic cell body.

Inflammasome A large multiprotein complex comprising an NLR (NOD-like receptor), the adaptor protein ASC (apoptosis-associated speck-like protein containing a CARD; also known as PYCARD) and pro-caspase 1. The assembly of the inflammasome leads to the activation of caspase 1, which cleaves pro-interleukin-1β (pro-IL-1β) and pro-IL-18 to generate the active pro-inflammatory cytokines.

Connexin and pannexin hemichannels Channels in the cell membrane formed from either connexin or pannexin molecules. Two hemichannels on adjacent cells can interact to form a gap junction, which allows intercellular communication by enabling the exchange of cytosolic molecules between the adjoining cells. In individual cells, connexin and pannexin hemichannels can facilitate the release of cellular ATP into the extracellular space.

Inside-out signalling The process by which intracellular signalling mechanisms result in the activation of cell surface receptors, such as integrins. By contrast, outside-in signalling is the process by which ligation of a cell surface receptor activates signalling pathways inside the cell.

%60U '06U

2

#FGPQUKPG FGCOKPCUG

#FGPQUKPG TGE[ENKPI

#62 3WCNKVCVKXG KPHQTOCVKQP

%&

3WCPVKVCVKXG KPHQTOCVKQP (WPEVKQPCN TGURQPUG

Figure 1 | Components of autocrine purinergic signalling systems. A schematic of the key elements involved in 0CVWTG4GXKGYU^+OOWPQNQI[ purinergic signalling. Receptor activation (for example, stimulation of formyl-peptide receptors (FPRs) by N-formylmethionyl-leucyl-phenylalanine (fMLP)) leads to the opening of pannexin 1 hemichannels and the release of ATP from the cell. Released ATP promotes the autocrine activation of P2 receptors. Ectonucleotidases (such as ectonucleoside triphosphate diphosphohydrolases (ENTPDs) and ecto-5’-nucleotidase (also known as CD73)) promote hydrolysis of ATP and the formation of adenosine, which activates P1 receptors. Adenosine is ‘neutralized’ by adenosine deaminase, which converts adenosine to inosine, or recycled through concentrative nucleoside transporters (CNTs) or equilibrative nucleoside transporters (ENTs). Autocrine purinergic signalling provides cells with quantitative cues that define how and to what extent the cells respond to the qualitative cues that they perceive through specific surface receptors that detect pathogens, antigens, chemokines and cytokines. Autocrine purinergic mechanisms allow localized stimulation because signalling molecules can be confined to specific regions, and different spatio-temporal combinations of signalling components can be arranged to meet specific functional requirements. Moreover, these autocrine feedback processes can be influenced in a paracrine fashion by purinergic receptor ligands that are generated by other cells, for example, in infected or damaged tissues.

Components of purinergic signalling ATP release. Immune cells recognize ATP that is released from damaged tissues and dying cells as a danger signal that elicits a variety of inflammatory responses19–21. In addition to damaged cells, intact cells (including immune cells themselves) can release ATP under normal physiological conditions. ATP release from intact cells was first observed in neuronal cells, which use vesicular transport to release ATP into the cleft of chemical synapses22. Non-neuronal cell types can also release ATP through vesicular transport 5; however a number of additional mechanisms have been reported. These mechanisms include release through stretch-activated channels, voltage-dependent anion channels, P2X7 receptors (a purinergic receptor subtype involved in opening large pores in the cell membrane), and connexin and pannexin hemichannels15. Pannexin 1 hemichannels were recently found to promote ATP release from several immune cell types. Similarly to connexin hemichannels, pannexin hemichannels are thought to form gap junctions between adjacent cells and enable rapid intercellular communication. In individual cells, including immune cells, these hemichannels can release ATP. The pannexin family has three members (pannexin 1–3) and these proteins

are distant relatives of the larger connexin family, which has 21 members23–25. In mouse neutrophils, ATP release appears to occur through connexin 43 hemichannels26. human neutrophils release ATP through pannexin 1 channels in response to stimulation of formyl-peptide receptors (FPRs), Fcγ receptors, IL-8 receptors, complement C5a receptors and leukotriene B4 receptors27. Pannexin 1 also facilitates ATP release from T cells after activation of the T cell receptor (TCR) and the CD28 costimulatory receptor, or following exposure of T cells to osmotic stress9,11. The surface expression of pannexin 1 is highly dynamic, and activation of a T cell or neutrophil causes translocation and accumulation of pannexin 1 at the T cell immune synapse or the leading edge of the polarized neutrophil, respectively 11,27. Extracellular ATP metabolism. Following its release into the extracellular space, ATP is rapidly hydrolysed in a stepwise manner to ADP, AMP and adenosine by ectonucleotidases18. This process terminates P2 receptor activation, prevents receptor desensitization and sets the stage for rapid and repeated signalling through purinergic inside-out signalling via purinergic receptors. Four families of ectonucleotidases that hydrolyse extracellular ATP have

202 | MARCh 2011 | vOLuMe 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS been identified in mammalian cells. These ectoenzymes include the ectonucleoside triphosphate diphosphohydrolase (eNTPD) family, the ectonucleotide pyrophosphatase/phosphodiesterase (eNPP) family, the alkaline phosphatase family and ecto-5ʹ-nucleotidase (also known as CD73)3,18,28. The eNTPD family comprises seven different isoforms (eNTPD1–6 and eNTPD8), of which eNTPD1–3 and eNTPD8 are known to hydrolyse ATP or ADP to AMP with different substrate preferences. The three members of the eNPP family (eNPP1–3) hydrolyse ATP to AMP and pyrophosphate. Finally, the four members of the alkaline phosphatase family (intestinal, placental, germ cell and tissue-nonspecific alkaline phosphatase) can hydrolyse ATP, ADP and AMP to adenosine, and the sole ecto-5ʹ-nucleotidase isoform converts AMP to adenosine3,18. ectonucleotidases are found on the surfaces of virtually all mammalian cell types with distinct distribution patterns that define cell-specific local purinergic ligand environments. Adenosine turnover and cellular reuptake. extracellular adenosine, the ligand of P1 receptors, can be removed either by adenosine deaminase, which converts adenosine to inosine, or by cellular reuptake through concentrative nucleoside transporters (CNTs) or equilibrative nucleoside transporters (eNTs)29,30. In addition to these mechanisms, adenosine kinase can convert extracellular adenosine back to AMP18. Two different adenosine deaminase isoforms, ADA and CeCR1 (also known as ADA2), are found on the surface of many cell types, including leukocytes31–35. Both isoforms promote co-stimulatory signalling at the immune synapse, resulting in increased production of T helper 1 (Th1)-type cytokines and increased T cell proliferation36,37. The importance of ADA is further underscored by the fact that a genetic disorder resulting in ADA deficiency causes severe combined immunodeficiency (SCID) — a condition characterized by impaired T and B cell development and an increased risk of infection — owing to excessive A2A adenosine receptor activation36,38,39. Adenosine reuptake is accomplished by two structurally unrelated families of nucleoside transporters: the four equilibrative nucleoside transporters (eNT1–4; also known as SLC29A1–4) and the three concentrative nucleoside transporters (CNT1–3; also known as SLC28A1–3)29,30,39. The best characterized eNT family members, eNT1 and eNT2, are inhibited by widely used drugs such as dipyridamole, dilazep and draflazine40. eNTs are involved in the activation of T cells41, B cells42 and macrophages43. Although nucleoside transporters have been linked with lymphocyte proliferation44, little is known about the underlying mechanisms. however, the ubiquitous expression of nucleoside transporters in different tissues suggests that they may ‘recharge’ purinergic signalling systems by recycling adenosine. Purinergic receptors. The most widely studied elements of purinergic signalling systems are the purinergic receptors that respond to extracellular ATP, adenosine and related nucleotides. Purinergic receptors are divided into three major families based on their pharmacological and

structural properties2 (TABLE 1). P2X receptors function as ATP-gated ion channels that facilitate the influx of extracellular cations, including calcium ions. Although all the P2X receptors respond only to ATP, they do so with different ligand affinities and ligand-induced receptor desensitization kinetics45. Among P2X receptors, the P2X7 subtype has a special role, as micromolar ATP concentrations elicit its ion channel function, whereas higher ATP concentrations (in the millimolar range) induce P2X7 to form large conductance pores that are involved in apoptosis45. P2Y receptors are G protein-coupled receptors (GPCRs) that recognize ATP and several other nucleotides, including ADP, uTP, uDP and uDP-glucose. These receptors can bind to a number of different G proteins (TABLE 1). P1 receptors are also GPCRs and recognize adenosine. A2A and A2B receptors couple to stimulatory G proteins (Gs) and typically suppress cell responses by upregulating intracellular cyclic AMP levels. By contrast, A1 and A3 receptors couple to Gi/o or Gq/11 proteins and promote cell activation. Purinergic signalling systems. Different cell types express distinct sets of the purinergic signalling components described above, and this allows the formation of customized purinergic signalling complexes (FIG. 1). These complexes permit immune cells to mount unique responses to ATP, adenosine and the other ligands described in TABLE 1. The role of purinergic signalling as a paracrine mechanism of intercellular communication among immune cells has been widely recognized7. however, recent findings show that purinergic signalling in immune cells has additional roles that are less well known. ATP released from damaged or stressed cells serves as an important danger signal that induces specific immune responses, and ATP can be released from immune cells themselves in response to normal cell stimulation. This triggers autocrine purinergic feedback mechanisms that are essential regulators of immune cell responses. Crosstalk between paracrine signals and autocrine purinergic signalling mechanisms makes it possible for immune cells to adapt their responses to the extracellular cues generated by the surrounding cells in different tissues.

Purinergic signalling and immune responses Immune cells have developed highly sensitive receptor systems that allow them to execute their many roles in immune surveillance and host defence. To perform these tasks effectively, neutrophils, for example, must detect trace amounts of the chemoattractants that help them locate and migrate to sites of infection and inflammation46. T cells are able to perform the astonishingly difficult task of recognizing individual peptide–MhC complexes with great selectivity and sensitivity 47. Although it is clear that these extraordinary feats require sophisticated signalamplification systems, the nature of these amplification mechanisms has remained unclear. Recent findings have shown that stimulation of neutrophils and T cells leads to the rapid release of ATP, which triggers autocrine purinergic feedback loops that amplify the weak stimuli these cells receive during cell activation.

NATuRe RevIewS | Immunology

vOLuMe 11 | MARCh 2011 | 203 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Table 1 | Purinergic receptors and downstream signalling Receptor ligands5,7 ligand binding affinities EC50 (μm)7

Desensitization main downstream Expression in different immune cell types7 rate or signalling events16 neutro- mono- macro- Dendritic T cells B cells85 nK g-protein phils12 cytes phages cells cells86 16,45 coupling

P2X ATP receptors P2X1

ATP

0.05–1

<1 s

Ca2+ and Na+ influx

+

+

+

+

+

+

+

P2X2

ATP

1–30

>20 s

Ca influx

ND

ND

ND

ND

ND

+

ND

P2X3

ATP

0.3–1

<1 s

Cation influx

ND

ND

ND

ND

ND

+

+

P2X4

ATP

1–10

>20 s

Ca influx

+

+

+

+

+

+

+

P2X5

ATP

1–10

>20 s

Ion influx

+

+

+

+

+

+

ND

P2X6

ATP

1–12

-

Ion influx

ND

ND

ND

ND

ND

+

+

P2X7

ATP

>100

>20 s

Cation influx and pore formation

+

+

+

+

+

+

+

2+

2+

P2Y nucleotide receptors* P2Y1

ADP

8

Gq/11

PLCβ activation

+

+

+

+

+

+

+

P2Y2

ATP, UTP

0.1 (ATP), 0.2 (UTP)

Gq/11, Gi/o

PLCβ activation, cAMP inhibition

+

+

+

+

+

+

+

P2Y4

UTP (ATP, UDP)

2.5

Gq/11, Gi/o

PLCβ activation, cAMP inhibition

ND

+

+

+

+

+

ND

P2Y6

UDP, UTP

0.3 (UDP), Gq/11 6 (UTP)

PLCβ activation

+

+

+

+

+

+

ND

P2Y11

ATP

17

Gs, Gq/11

cAMP production, PLCβ activation

ND

+

+

+

+

+

ND

P2Y12

ADP

0.07

Gi/o

cAMP inhibition

+

+

+

ND

+

+

ND

P2Y13

ADP, ATP

0.06 Gi/o (ADP), 0.26 (ATP)

cAMP inhibition

+

+

ND

+

+

+

ND

P2Y14

UDPglucose

0.1–0.5

Gq/11

PLCβ activation

+

ND

ND

+

+

+

+

P1 adenosine receptors A1

adenosine 0.2–0.5

Gi/o

cAMP inhibition

+

+

+

+

ND

ND

ND

A2A

adenosine 0.6–0.9

Gs

cAMP production

+

+

+

+

+

+

+

A2B

adenosine 16–64

Gs

cAMP production

+

+

+

+

+

ND

+

A3

adenosine 0.2–0.5

Gi/o, Gq/11

cAMP inhibition, InsP3 generation

+

+

+

+

+

ND

+

Purinergic receptors are divided into the P1, P2X and P2Y subfamilies that respond to different nucleotide and nucleoside ligands with a range of binding affinities2,7,16,45. cAMP, cyclic AMP; EC50, half-maximum effective concentration; InsP3, inositol-1,4,5-trisphosphate; ND, not detected; NK, natural killer; PLCβ, phospholipase Cβ. *In addition to the P2Y receptors shown in this table, genes encoding the related receptors P2Y3, P2Y5, P2Y7, P2Y8, P2Y9, P2Y10 were cloned but found not to respond to purinergic ligands, except for P2Y3, which is an avian analogue of the mammalian P2Y6 receptor.

Regulation of neutrophil and macrophage chemotaxis. Stimulation of chemotaxis receptors induces rapid ATP release from neutrophils, and this is necessary to promote neutrophil recognition of chemotactic gradients and guide these cells to infected or inflamed tissues12. A new method that allows direct imaging of ATP release from live cells has revealed that ATP is released in a polarized manner, resulting in the accumulation of extracellular ATP near the cell surface closest to the source of chemoattractants. The released ATP can activate multiple adjacent P2Y2 receptors to provide autocrine signal amplification, which greatly increases the intracellular signals generated by chemotactic stimuli (FIG. 2). Inhibiting purinergic signalling by blocking ATP release or P2Y2 receptors, or by expediting the removal of extracellular ATP, impairs chemotaxis. Similarly,

overwhelming the endogenous autocrine purinergic systems by adding excessive exogenous ATP also blocks chemotaxis12. Thus, altering ATP concentrations in the extracellular environment can modulate chemotaxis by interfering with the autocrine signalling systems of neutrophils. Chemotaxis is a complex process that involves gradient sensing, cell polarization and directed migration. Gradient sensing is the process by which a cell detects chemoattractants and recognizes differences in chemoattractant concentrations in its extracellular environment. During cell polarization, a cell assumes an elongated shape that aligns within the chemotactic gradient field. At the same time, specific receptors and signalling molecules translocate to the leading and trailing edges of the cell. Finally, directed migration is the actual forward

204 | MARCh 2011 | vOLuMe 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

H/.2

2CPPGZKP #62 J[FTQN[UKU D['062&

(24 '062& #62 # 2;

5KIPCNCORNKȮECVKQP )TCFKGPVUGPUKPI #

#62

#62

5KIPCNNKPIEQORQPGPVU VTCPUNQECVGVQNGCFKPI GFIGQHEGNN

5KIPCNCORNKȮECVKQP

(QTYCTF OKITCVKQP

(WTVJGTUKIPCN CORNKȮECVKQP

#FGPQUKPG #EVKXCVKQPQH #TGEGRVQTU RTQOQVGUEGNN OKITCVKQP

0GWVTQRJKN

'062& +PKVKCNEGNNCEVKXCVKQP CPFITCFKGPVUGPUKPI

2QNCTK\CVKQP

(QTYCTFOKITCVKQP

Figure 2 | Purinergic signal amplification regulates neutrophil chemotaxis. Neutrophils are activated by bacterial formylated peptides, such as N-formyl-methionyl-leucyl-phenylalanine (fMLP), through formyl-peptide receptors (FPRs). 0CVWTG4GXKGYU^+OOWPQNQI[ This induces the release of ATP through pannexin 1 hemichannels and thus results in the activation of P2Y2 receptors, which provide signal amplification and facilitate the sensing of chemotactic gradients (left). These signalling mechanisms allow the cell to polarize within the chemotactic gradient field (centre) and induce the translocation of pannexin 1, ectonucleoside triphosphate diphosphohydrolase 1 (ENTPD1, also known as CD39) and A3 receptors to the leading edge of the cell. Accumulation of these purinergic signalling components focuses ATP release and the formation of adenosine at the leading edge and results in autocrine activation of A3 receptors, which promote cell migration towards the source of fMLP (right). Preliminary evidence suggests that negative autocrine feedback through A2A adenosine receptors (which block cell activation and remain uniformly distributed across the cell surface) supports chemotaxis by promoting the retraction of the trailing edge of the cell (W.G.J., unpublished observations) (not depicted in the figure).

movement of a cell within a chemoattractant gradient field. Recent studies have shown that autocrine purinergic feedback mechanisms control several aspects of these different processes. Initially, stimulation of a chemotaxis receptor induces ATP release and the activation of multiple adjacent P2Y2 receptors. This process can be thought of as an inside-out signalling event, in which ATP serves as a second messenger that amplifies chemotaxis signals through nearby P2Y2 receptors. During cell polarization, purinergic signalling molecules (such as pannexin 1, eNTPD1 and A3 adenosine receptors) translocate to the leading edge of the cell, where membrane ruffling and pseudopod protrusion lead to the release of additional ATP, which is converted to adenosine by eNTPD1 (REFS 12,27,48). Finally, adenosine activates A3 receptors that are concentrated at the leading edge, and this additional purinergic feedback loop promotes migration and forward movement of neutrophils towards the source of chemoattractants49 (FIG. 2). exogenous ATP and adenosine have long been known to modulate neutrophil responses50–56. The discovery of the autocrine purinergic signalling systems described above sheds light on the underlying mechanisms by which these compounds regulate chemotaxis. Similar autocrine signalling systems were recently identified in macrophages and these regulate chemotaxis through P2Y2 and A3 receptors, as well as through P2Y12, A2A and A2B receptors57. Interestingly, A2A receptors also have a role in the chemotaxis of microglial cells, controlling uropod

retraction in these cells58. In addition to the GPCR type of purinergic receptors that are involved in these responses12,57,59, recent work by Lecut et al.60 has shown that the structurally distinct P2X1 receptor can regulate the chemotaxis of neutrophils through the activation of RhO family kinases. Regulation of chemotaxis in other immune cells. Purinergic signalling may also be involved in the chemotaxis of other immune cell types. ATP itself, when released from injured cells, has been reported to promote chemotactic and phagocytic responses in microglia through P2Y12 and P2Y13 receptors, and P2Y6 receptors, respectively 61–66. Similar observations were reported for mast cells, monocytes, dendritic cells and eosinophils67–70. however, a drawback of many of these studies is that chemotaxis was not observed directly (that is, under a microscope), but was instead monitored indirectly using transwell assay systems, which do not differentiate between chemotaxis and chemokinesis. Thus, it is possible that ATP increases random migration of some of these cell types rather than promoting chemotaxis. Similarly, purinergic receptor knockout or silencing experiments, although extremely useful, can yield ambiguous results that must be considered carefully because loss of chemotaxis does not necessarily indicate that a specific purinergic receptor is a true chemotaxis receptor. Instead, the receptor may be required for autocrine purinergic signalling that regulates chemotaxis in

NATuRe RevIewS | Immunology

vOLuMe 11 | MARCh 2011 | 205 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Inflammatory site

Resting

Gradient sensing

Dying cell Polarization

Chemotaxis Chemokinesis, degranulation, phagocytosis

Neutrophil

ATP Chemoattractant Chemotaxis speed

Chemokinesis

Figure 3 | Regulation of phagocyte chemotaxis by autocrine and paracrine purinergic signalling mechanisms. Nature Reviews | Immunology Phagocytes, such as neutrophils, monocytes and macrophages, require autocrine purinergic signalling to detect and migrate in response to chemoattractants and danger signals issuing from inflamed and infected sites. Long-range chemotactic signals, such as N-formyl-methionyl-leucyl-phenylalanine (fMLP) and interleukin-8 (IL-8) promote the recruitment of phagocytes to inflamed tissues. ATP that is released from dying cells at the site of inflammation is short lived and can, therefore, only serve as a short-range signal to neighbouring phagocytes. The ATP released from the dying cells promotes their uptake by phagocytes by interfering with autocrine purinergic signalling mechanisms (and thus trapping phagocytes within the vicinity) and by upregulating phagocytosis and other phagocyte killing mechanisms, resulting in clearance of the dying cells.

response to other stimuli, for example DAMPs or PAMPs that are released at inflammatory sites. ATP and other danger signals may also induce sentinel cells to secrete chemokines, such as IL-8, that can recruit leukocytes to inflammatory sites71–74. Such interconnected mechanisms, including ATP release, inflammasome activation, IL-8 secretion and FPR-mediated chemotaxis, were recently shown to regulate recruitment of neutrophils to sites of inflammation75. however, the authors concluded that ATP released from inflammatory sites does not serve as a chemoattractant. This is in contrast to other reports suggesting that apoptotic cells release ATP through pannexin 1 and that ATP itself acts as a powerful ‘find-me’ signal that recruits monocytes13,76. It remains to be seen whether this response involves additional chemotactic mediators that are released from apoptotic cells along with ATP. Given the ubiquitous nature of ATP release by healthy cells, chemotaxis towards ATP would seem counterproductive under most physiological circumstances. Thus, it seems more likely that released ATP regulates chemotaxis rather than inducing it (FIG. 3). Signal amplification in T cells. T cells recognize antigens through their TCRs, which localize to the immune synapse and physically interact with peptides that are presented on MhC molecules by antigen-presenting cells (APCs). A surprisingly low number of peptide–MhC complexes are required for T cell activation77, and it is unclear how such weak signals are able to provide sufficiently strong stimulatory signals to activate T cells during their brief encounter with APCs. In order to respond with high sensitivity and selectivity to their antigens, it is clear that T cells must possess mechanisms that permit signal amplification78. Several models of TCR signalling

have been proposed47,78. The serial-triggering model assumes that a single peptide–MhC complex can activate multiple TCRs in fast succession. Kinetic proofreading models may explain how T cells can either be activated or remain in a resting state based on very small differences in the antigen-binding affinities of their TCR for the different antigens they encounter. Cooperative models propose that multiple TCR complexes cooperate through physical contact or second messengers, allowing T cells to distinguish rare high-affinity antigens from the high background signal generated by the abundance of low-affinity antigens. It has also been proposed that mechanical forces exerted on T cells and accessory cells as a result of TCR engagement are crucial for T cell activation and antigen recognition79. Recent evidence suggests that purinergic signalling at the immune synapse could serve as a signal amplification mechanism needed for antigen recognition. A number of studies have shown that T cells can release ATP in response to various extracellular stimuli, suggesting that purinergic signalling may indeed play an active role in T cell activation8,80,81. T cells express many members of the P2X, P2Y and P1 receptor families, as well as the eNTPD1 ectonucleotidase82–84. Recent findings by Schenk et al.9 have revealed that TCR stimulation triggers the release of cellular ATP through pannexin 1 channels. Pannexin 1 translocates to the immune synapse, where it releases ATP and promotes T cell activation, suggesting that autocrine purinergic feedback processes, similar to those identified in neutrophils, also regulate T cell activation9–11. however, although neutrophils seem to amplify stimulatory signals using P2Y2 receptors, purinergic signal amplification in T cells occurs through P2X1, P2X4 and P2X7 receptors.

206 | MARCh 2011 | vOLuMe 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 4QNGHQT2TGEGRVQTGZRTGUUKQP CPF#62TGNGCUGD[#2%!

#2%

%& %&

/*% ENCUU++

2CPPGZKP

2:

%C

2GRVKFG

+OOWPGU[PCRUG

2:

2CPPGZKP 6EGNN

%&

'PXKTQPOGPVCN #62

2:

6%4 #62

6%4CPF%& UVKOWNCVKQPRTQXKFGU CEVKXCVQT[UKIPCN

2CTCETKPG2:TGEGRVQT UKIPCNNKPIRTQXKFGU GPXKTQPOGPVCNUKIPCNU

2:

#WVQETKPGRWTKPGTIKE UKIPCNNKPIRTQXKFGU UKIPCNCORNKȮECVKQP

'PVT[QH%C XKC 2:TGEGRVQTU RTQOQVGU0(#6 CEVKXCVKQP

5KIPCNKPVGITCVKQP (WPEVKQPCNTGURQPUGU +.O40#VTCPUETKRVKQP 6EGNNRTQNKHGTCVKQP

Figure 4 | Purinergic signalling in T cell activation. Antigen recognition by T cells involves the formation of an immune synapse between a T cell and an antigen-presenting cell (APC). The immune synapse contains a large number of signalling molecules that are required for T cell activation, including T cell receptors (TCRs), MHC molecules, 0CVWTG4GXKGYU^+OOWPQNQI[ co-stimulatory receptors and the purinergic signalling receptors P2X1, P2X4 and P2X7. In response to TCR and CD28 stimulation, pannexin 1, P2X1 receptors and P2X4 receptors translocate to the immune synapse. ATP released through pannexin 1 promotes autocrine signalling via the P2X receptors. Confinement of ATP in the immune synapse results in a powerful autocrine feedback mechanism that facilitates the signal amplification required for antigen recognition. P2 receptors expressed and ATP released by APCs may also have important roles in regulating the antigen recognition process. NFAT, nuclear factor of activated T cells.

P2X receptors function as calcium channels, and autocrine activation of these receptors facilitates calcium influx and downstream signalling, leading to IL2 transcription11. Interestingly, two of the three P2X receptor subtypes that are involved in this autocrine signalling process, P2X1 and P2X4, translocate along with pannexin 1 to the immune synapse within minutes of TCR stimulation11. This increases the effectiveness of purinergic signalling by concentrating key signalling elements within the immune synapse. Although P2X7 receptors are found at the immune synapse, they do not translocate in response to TCR stimulation and the majority of P2X7 receptors remain uniformly distributed across the cell surface. This suggests that different P2X receptor subtypes have different functions during the various stages of T cell activation, and that P2X7 receptors may allow T cells to remain responsive to ATP generated by surrounding tissues and cells that are not directly involved in antigen presentation. Similarly to the relocalization of signalling molecules to the leading edge of polarized neutrophils, the translocation of pannexin 1 and P2X receptors to the T cell immune synapse promotes strong positive purinergic feedback

mechanisms, which are further amplified in the confined space of the synaptic cleft (FIG. 4). under such conditions, it is feasible that ATP released in response to the engagement of a single TCR complex could indeed result in the activation of a sufficiently large number of P2X receptors to elicit powerful and selective T cell responses. Regulation of other immune cells. Although there is no direct proof, ample indirect evidence suggests that autocrine purinergic signalling could also regulate the activation of many other immune cell types in addition to those discussed above. Most immune cells express at least some of the different P1 and P2 receptors and are able to release ATP7,85–88. B cells, for example, release ATP, and extracellular ATP promotes B cell proliferation87,88. In addition to controlling migratory responses in dendritic cells, monocytes, macrophages and eosinophils, extracellular ATP has been shown to regulate the migration of NK cells89. Because virtually all mammalian cell types possess purinergic receptors and most, if not all, cells release ATP in response to cell stimulation, there is little doubt that autocrine purinergic mechanisms regulate the functional responses of many of these cell

NATuRe RevIewS | Immunology

vOLuMe 11 | MARCh 2011 | 207 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS types. Thus, autocrine purinergic signalling seems to be a recurring theme in mammalian cell activation. In support of this conclusion, recent work using human embryonic kidney (heK) cells has revealed that activation of non-immune cells through various adrenergic receptors also requires autocrine signalling through P1 and P2 receptors90. Future work in this area will undoubtedly provide exciting new insights into the roles of autocrine purinergic responses in immune cell activation.

ATP as a danger signal Necrotic and apoptotic cells release ATP, which can serve as a find-me signal that attracts monocytes to phagocytose and remove dead or dying cells13,20. This process involves the activation of P2Y2 receptors on monocytes, which are thought to follow ATP gradients to find their way to apoptotic cells. however, because of the short half-life of ATP in most tissues, it is uncertain whether this process alone could attract phagocytes over long distances20,82. It is more likely that damaged and stressed cells release additional find-me signals that are involved in the recruitment of phagocytes75,91. Recently, formylated peptides, which are released from the mitochondria of damaged cells, have been shown to induce neutrophil activation and chemotaxis in response to severe trauma and sterile inflammation75,92. It is likely that such chemotactic mediators, together with chemokines and ATP, orchestrate the complex processes that guide phagocytes during their long-range approach and final encounter with target cells at inflammatory foci75. This may also be true for pathological situations, such as those that follow severe trauma or cancer therapy, in which ATP release and danger signals emitted by damaged cells stimulate host immune responses through Toll-like receptors (TLRs) and NOD-like receptors (NLRs). Interestingly, recent work has shown that extracellular ATP can promote NLR-mediated inflammasome assembly. Activation of the NOD-, LRR- and pyrin domaincontaining 3 (NLRP3; also known as NALP3) inflammasome has been shown to involve ATP release through pannexin 1 and purinergic signalling through P2X7 receptors. Inflammasome activation triggers innate immune defences by inducing the maturation of pro-inflammatory cytokines, such as IL-1β, in a caspase 1-dependent manner21,71,93. Activation of the NLRP3 inflammasome has also been observed in response to tumour cell destruction after cancer therapy 20. This process releases ATP and DAMPs, which stimulate P2X7 receptors and the NLRP3 inflammasome of dendritic cells, facilitating IL-1β production and promoting immunity against tumours94. The exact mechanisms that link purinergic signalling to NLRP3 inflammasome activation remain to be determined. however, it is intriguing to speculate that autocrine purinergic signalling events could be involved in this and other immune cell responses to DAMPs and PAMPs. Inhibition of immune responses Through the various mechanisms described above, extracellular ATP promotes immune cell activation and pro-inflammatory responses, particularly following the acute release of ATP (for example, in response

to cell stimulation, stress or tissue damage) and when extracellular ATP is present at high concentrations. however, ATP can also have anti-inflammatory effects, especially when extracellular ATP is generated chronically and at low concentrations19,95,96. Such dual effects of ATP have been observed in dendritic cells, monocytes, CD4+ T cells and neutrophils69,97–101. The overall response elicited by each cell type depends on the particular P2 receptor subtypes that are stimulated on the cell surface. P2 receptor desensitization and internalization in response to chronic or high extracellular ATP concentrations and the activation of suppressive P1 and P2 receptors (such as A2A or P2Y11) are additional mechanisms that may affect the responses of different immune cell types to ATP. T cell suppression. A2 adenosine receptors have a major role in suppressing immune responses. The complex network of ectonucleotidases that regulates the ligand availability for P1 and P2 receptors in different tissues has a central role in defining the immune responses in these tissues7,18. eNTPD1, which converts ATP to AMP (the precursor of adenosine), and ecto-5ʹ-nucleotidase, which generates adenosine from AMP, are particularly important for the balance between the pro- and antiinflammatory effects of released cellular ATP28. This can be observed in eNTPD1-deficient mice, which develop more severe injury-induced inflammation than wildtype mice, apparently owing to reduced ATP hydrolysis, adenosine formation and A2A receptor activation102. In addition, ecto-5ʹ-nucleotidase has been shown to have important roles in cancer immunity. Several recent studies have demonstrated that cancer cells express high levels of ecto-5ʹ-nucleotidase and thus promote the generation of adenosine, which impairs the functions of antitumour CD8+ T cells through A2A receptormediated signalling 104. Similarly, A2A receptor-driven purinergic signalling events attenuate allograft rejection in animal models by impairing T cell responses105–106. eNTPD1 and ecto-5ʹ-nucleotidase are also expressed on the surface of regulatory T (TReg) cells, and this allows these cells to convert extracellular ATP to adenosine and suppress effector cells through A2A receptormediated signalling 107. TReg cell expression of eNTPD1 has also recently been shown to impair NK cell activity and promote metastatic growth in a mouse tumour model108. Thus, the functions of effector T cells can be suppressed by autocrine or paracrine purinergic signalling loops that involve the hydrolysis of extracellular ATP released from TReg cells, other immune cells or tumour cells. Although much evidence suggests that A2A receptors have a central role in the downregulation of T cell functions, the extent to which P2Y receptors, and particularly the Gs-coupled P2Y11 receptors, can also contribute to this process remains to be seen. Inhibition of phagocyte function. Neutrophils express the A2A and, to a lesser extent, A2B adenosine receptors, which are Gs-coupled P1 receptors. In polarized neutrophils, A2A receptors remain uniformly distributed across the cell surface12, and this suggests that these

208 | MARCh 2011 | vOLuMe 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS receptors may provide a global inhibitory signal that is an important requirement for efficient chemotaxis46. Neutrophils rapidly convert extracellular ATP to adenosine12,48, which can bind to A2A receptors to provide a powerful counterpoint to the stimulatory effects of P2Y2 and A3 receptors. Furthermore, the suppressive actions of A2A receptors provide an important mechanism for limiting key neutrophil functions, including the oxidative burst, degranulation and phagocytosis12,48,51,109. Because of their predominant role in downregulating the effector functions of neutrophils and other immune cells, A2A receptors have been extensively investigated as targets for anti-inflammatory therapies to treat patients with asthma, arthritis and other acute and chronic inflammatory diseases110.

Therapeutic targeting of purinergic receptors Purinergic receptors are attractive therapeutic targets for treating a range of diseases because they are accessible at the cell surface. however, subtype-specific agonists and antagonists are currently available only for P1 receptors and a few of the 15 P2 receptor subtypes. Because of this limitation and the incomplete understanding of the varied roles of the different purinergic signalling events in physiological and pathophysiological processes, only a small number of purinergic drugs are currently in clinical use or being tested in clinical trials. A notable exception are platelet-targeting drugs, such as clopidogrel (Plavix; Bristol–Myers Squibb/Sanofi–Aventis), which act as P2Y12 receptor antagonists. These drugs prevent platelet aggregation and are used to treat vascular ischaemic events in patients with atherosclerosis

and acute coronary syndrome and are thus among the most widely used pharmaceuticals today 111. Although the pharmacological actions of these drugs are thought to be platelet specific, it is likely that immune cells expressing P2Y12 receptors are also affected (TABLE 1). One of the most common side effects of clopidogrel is severe neutropaenia, which may be a reflection of such off-target effects of the drug 112. Compared to the P2 receptors, a much wider selection of powerful and specific drugs is available for the modulation of adenosine receptors113–115. Many of these drugs are derivatives of caffeine and have been extensively tested and shown to be effective in treating acute and chronic inflammatory diseases in preclinical trials115. Currently, several drugs are being tested in clinical trials for their efficacy in chronic obstructive pulmonary disease, rheumatoid arthritis and other inflammatory diseases115. In preclinical trials, several pharmacological approaches have been used to increase the availability of extracellular adenosine, including blocking adenosine reuptake (for example, with dipyridamole), blocking adenosine conversion to inosine (with ADA inhibitors, such as pentostatin and erythro-9-(2-hydroxy-3-nonyl) adenine (ehNA)), or increasing the release of adenosine from cells (for example, by low-dose treatment with methotrexate116,117). These approaches have shown strong anti-inflammatory effects in various experimental models, including ischaemia–reperfusion injury, sepsis and airway inflammation115,116. Several studies have shown that the A3 receptor agonist IB-MeCA (N6-(3-iodobenzyl)-5ʹ-(N-methylcarbamoyl)adenosine) can ameliorate arthritis symptoms in mouse

Box 1 | Remaining questions and future research directions Future research focusing on the specific roles of the different purinergic signalling components in different cell types is likely to yield exciting information. How do the multiple members of the P1 and P2 receptor families, ectonucleotidases and other components assemble to form autocrine purinergic signalling systems that can shape distinct and appropriate immune responses? Which defects of these complex purinergic signalling systems result in immune disorders and which defects can be compensated for; for example, by substitution of a mutated purinergic receptor with another isoform or splice variant? Although pannexin 1 appears to be a major player in facilitating ATP release from T cells, neutrophils and monocytes, it is still unclear how other ATP release mechanisms (such as vesicular transport) regulate ATP release and how all the possible release mechanisms control cell responses such as antigen recognition, gradient sensing and chemotaxis. These questions above and a number of other open issues, such as the spatio-temporal aspect of ATP release and the dynamics of ATP hydrolysis, require new analytical tools to assess ATP release. For example, it will be important to develop new imaging methods that are more sensitive and technically less challenging15,121 and combine these techniques with assays that allow us to monitor the release of ATP and other nucleotides in situ. For example, Di Virgilio and co-workers recently developed a technique that enables real-time, whole-body imaging of ATP release in mice122,123. Another important question that would benefit from such advanced ATP imaging techniques is how purinergic signalling facilitates intercellular communication among immune cells; for example, between dendritic cells and T cells during antigen presentation. Are some purinergic signalling systems at the immune synapse designed to specifically coordinate intercellular communication, while others serve to amplify activation, like the purinergic systems in chemical and electrical neuronal synapses? More sophisticated pharmaceutical tools must also be developed to specifically and selectively block different purinergic signalling events. This requires selective inhibitors of release channels, such as pannexin 1, and more selective drugs to target the different P2 receptors, ectonucleotidases, adenosine deaminases and nucleoside transporters. The purinergic signalling field is still in its infancy and there is much still to be learned before we fully understand the functions of the individual purinergic signalling components. For instance, the roles of the multiple splice variants of the P2X7 receptor are unclear. Another interesting future direction for research is the investigation of a potentially more universal role for autocrine purinergic signalling in cell activation. What types of cells and receptor classes require purinergic signalling mechanisms for signal amplification? Is purinergic inside-out signalling linked with calcium outside-in signalling and, if so, what mechanisms are involved?

NATuRe RevIewS | Immunology

vOLuMe 11 | MARCh 2011 | 209 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS models118,119. IB-MeCA has also been tested with positive results in a Phase II trial in patients with rheumatoid arthritis120, suggesting that stimulation of A3 receptors attenuates the inflammatory processes involved in rheumatoid arthritis, possibly by A3 receptor desensitization or internalization. The complexity of the processes by which autocrine and paracrine purinergic signalling mechanisms can control immune cell functions has been a major hurdle in successfully targeting these mechanisms in inflammatory diseases and immune disorders. The rapidly growing knowledge of the purinergic signalling events that regulate immune responses provides new incentives and insights that should foster the development of new drugs and therapeutic approaches to treat major diseases such as chronic obstructive pulmonary disease, ischaemia– reperfusion injury, arthritis, sepsis, inflammatory bowel diseases, allergic diseases and asthma.

Concluding perspectives ATP release and autocrine purinergic signalling feedback regulate numerous cellular immune responses through inside-out signalling, and constitute a powerful amplification mechanism to increase the sensitivity 1.

2.

3.

4. 5. 6.

7.

8.

9.

Burnstock, G., Fredholm, B. B., North, R. A. & Verkhratsky, A. The birth and postnatal development of purinergic signalling. Acta Physiol. 199, 93–147 (2010). This is a comprehensive review by the pioneers in the field who describe the development and growth of the purinergic signalling field from its beginning, with the initial discovery of ATP, to its current widespread scope that encompasses a wide range of research fields. Ralevic, V. & Burnstock, G. Receptors for purines and pyrimidines. Pharmacol. Rev. 50, 413–492 (1998). This review offers an in-depth view of the molecular and pharmacological properties of the known purinergic receptors. Zimmermann, H. Extracellular metabolism of ATP and other nucleotides. Naunyn Schmiedebergs Arch. Pharmacol. 362, 299–309 (2000). This review by a leader in the field provides detailed information on the various ectoenzyme families that facilitate the hydrolysis of extracellular ATP and related nucleotides. Burnstock, G. Unresolved issues and controversies in purinergic signalling. J. Physiol. 586, 3307–3312 (2008). Corriden, R. & Insel, P. A. Basal release of ATP: an autocrine-paracrine mechanism for cell regulation. Sci. Signal. 3, re1 (2010). Abbracchio, M. P., Burnstock, G., Verkhratsky, A. & Zimmermann, H. Purinergic signalling in the nervous system: an overview. Trends Neurosci. 32, 19–29 (2009). Bours, M. J., Swennen, E. L., Di Virgilio, F., Cronstein, B. N. & Dagnelie, P. C. Adenosine 5ʹ-triphosphate and adenosine as endogenous signaling molecules in immunity and inflammation. Pharmacol. Ther. 112, 358–404 (2006). This comprehensive review article provides a summary of the evidence that purinergic signalling systems are involved in the regulation of virtually all immune cell populations. Filippini, A., Taffs, R. E. & Sitkovsky, M. V. Extracellular ATP in T lymphocyte activation: possible role in effector functions. Proc. Natl Acad. Sci. USA 87, 8267–8271 (1990). This work first demonstrated ATP release from stimulated T cells. Schenk, U. et al. Purinergic control of T cell activation by ATP released through pannexin-1 hemichannels. Sci. Signal. 1, ra6 (2008). This paper elucidated the important role of pannexin 1 in facilitating ATP release, which

10. 11.

12.

13.

14. 15.

16.

17. 18.

and selectivity of immune cells to activating stimuli. Cells need these autocrine amplification mechanisms to tailor their functional responses to various extracellular cues. These autocrine mechanisms act as checkpoints that either enhance or inhibit immune cell activation depending on which purinergic receptors participate in signalling. This concept allows for multifaceted intercellular signalling through purinergic interactions among neighbouring cells and suggests many exciting future directions for research, such as investigating how groups of immune cells use their individual purinergic signalling systems to shape the collective immune response at sites of infection. It will also be very interesting and important to determine how the nature of the surrounding tissue affects the autocrine purinergic signalling systems of individual cells; for example, in neutrophils migrating though different tissues in order to reach their target destination. A number of questions remain to be answered before we can fully understand the diverse roles of autocrine purinergic signalling in immune cell regulation (BOX 1); a better appreciation of these issues will be important for developing new immunomodulatory therapies that target purinergic receptors.

drives the autocrine purinergic signalling events that are required for T cell activation. Yip, L. et al. Autocrine regulation of T-cell activation by ATP release and P2X7 receptors. FASEB J. 23, 1685–1693 (2009). Woehrle, T. et al. Pannexin-1 hemichannel-mediated ATP release together with P2X1 and P2X4 receptors regulate T cell activation at the immune synapse. Blood 116, 3475–3484 (2010). This paper showed that pannexin 1 and two key P2X receptors translocate to the immune synapse and contribute to calcium influx in response to TCR stimulation. Chen, Y. et al. ATP release guides neutrophil chemotaxis via P2Y2 and A3 receptors. Science 314, 1792–1795 (2006). This paper was the first report to indicate that neutrophil chemotaxis is regulated by autocrine purinergic signalling. Elliott, M. R. et al. Nucleotides released by apoptotic cells act as a find-me signal to promote phagocytic clearance. Nature 461, 282–286 (2009). This paper showed that ATP released from apoptotic cells serves as a find-me signal and induces the recruitment of monocytes. Trautmann, A. Extracellular ATP in the immune system: more than just a “danger signal”. Sci. Signal. 2, pe6 (2009). Praetorius, H. A. & Leipziger, J. ATP release from non-excitable cells. Purinergic Signal. 5, 433–446 (2009). This review article provides a thorough overview of the various ATP release mechanisms that are involved in purinergic signalling as well as the methods available to assess ATP release. Abbracchio. et al. International Union of Pharmacology LVIII: update on the P2Y G proteincoupled nucleotide receptors: from molecular mechanisms and pathophysiology to therapy. Pharmacol. Rev. 58, 281–341 (2006). Burnstock, G. Pathophysiology and therapeutic potential of purinergic signaling. Pharmacol. Rev. 58, 58–86 (2006). Yegutkin, G. G. Nucleotide- and nucleoside-converting ectoenzymes: important modulators of purinergic signalling cascade. Biochim. Biophys. Acta 1783, 673–694 (2008). This review article provides a comprehensive overview of the complex enzymatic systems that regulate extracellular concentrations of the ligands of purinergic receptors.

210 | MARCh 2011 | vOLuMe 11

19. Di Virgilio, F. Purinergic mechanism in the immune system: a signal of danger for dendritic cells. Purinergic Signal. 1, 205–209 (2005). 20. Ravichandran, K. S. Find-me and eat-me signals in apoptotic cell clearance: progress and conundrums. J. Exp. Med. 207, 1807–1817 (2010). 21. Schroder, K. & Tschopp, J. The inflammasomes. Cell 140, 821–832 (2010). 22. Burnstock, G. Purinergic signalling and disorders of the central nervous system. Nature Rev. Drug Discov. 7, 575–590 (2008). 23. Dahl, G. & Locovei, S. Pannexin: to gap or not to gap, is that a question? IUBMB Life 58, 409–419 (2006). 24. Laird, D. W. Life cycle of connexins in health and disease. Biochem. J. 394, 527–543 (2006). 25. MacVicar, B. A. & Thompson, R. J. Non-junction functions of pannexin-1 channels. Trends Neurosci. 33, 93–102 (2010). 26. Eltzschig, H. K. et al. ATP release from activated neutrophils occurs via connexin 43 and modulates adenosine-dependent endothelial cell function. Circ. Res. 99, 1100–1108 (2006). 27. Chen, Y. et al. Purinergic signaling: a fundamental mechanism in neutrophil activation. Sci. Signal. 3, ra45 (2010). This paper shows that autocrine purinergic signalling is an essential event in neutrophil activation. 28. Beldi, G. et al. The role of purinergic signaling in the liver and in transplantation: effects of extracellular nucleotides on hepatic graft vascular injury, rejection and metabolism. Front. Biosci. 13, 2588–2603 (2008). 29. Gray, J. H., Owen, R. P. & Giacomini, K. M. The concentrative nucleoside transporter family, SLC28. Pflugers Arch. 447, 728–734 (2004). 30. Baldwin, S. A. et al. The equilibrative nucleoside transporter family, SLC29. Pflugers Arch. 447, 735–743 (2004). 31. Franco, R., Valenzuela, A., Lluis, C. & Blanco, J. Enzymatic and extraenzymatic role of ecto-adenosine deaminase in lymphocytes. Immunol. Rev. 161, 27–42 (1998). 32. Cristalli, G. et al. Adenosine deaminase: functional implications and different classes of inhibitors. Med. Res. Rev. 21, 105–128 (2001). 33. Kameoka, J., Tanaka, T., Nojima, Y., Schlossman, S. F. & Morimoto, C. Direct association of adenosine deaminase with a T cell activation antigen, CD26. Science 261, 466–469 (1993). 34. Ciruela, F. et al. Adenosine deaminase affects ligandinduced signalling by interacting with cell surface adenosine receptors. FEBS Lett. 380, 219–223 (1996).

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 35. Herrera, C. et al. Adenosine A2B receptors behave as an alternative anchoring protein for cell surface adenosine deaminase in lymphocytes and cultured cells. Mol. Pharmacol. 59, 127–134 (2001). 36. Franco, R., Pacheco, R., Gatell, J. M., Gallart, T. & Lluis, C. Enzymatic and extraenzymatic role of adenosine deaminase 1 in T-cell–dendritic cell contacts and in alterations of the immune function. Crit. Rev. Immunol. 27, 495–509 (2007). 37. Zavialov, A. V. et al. Human adenosine deaminase 2 induces differentiation of monocytes into macrophages and stimulates proliferation of T helper cells and macrophages. J. Leukoc. Biol. 88, 279–290 (2010). 38. Apasov, S. G. & Sitkovsky, M. V. The extracellular versus intracellular mechanisms of inhibition of TCRtriggered activation in thymocytes by adenosine under conditions of inhibited adenosine deaminase. Int. Immunol. 11, 179–189 (1999). 39. Apasov, S. G., Blackburn, M. R., Kellems, R. E., Smith, P. T. & Sitkovsky, M. V. Adenosine deaminase deficiency increases thymic apoptosis and causes defective T cell receptor signaling. J. Clin. Invest. 108, 131–141 (2001). 39. Molina-Arcas, M., Casado, F. J. & Pastor-Anglada, M. Nucleoside transporter proteins. Curr. Vasc. Pharmacol. 7, 426–434 (2009). 40. King, A. E., Ackley, M. A., Cass, C. E., Young, J. D. & Baldwin, S. A. Nucleoside transporters: from scavengers to novel therapeutic targets. Trends Pharmacol. Sci. 27, 416–425 (2006). 41. Kichenin, K., Pignede, G., Fudalej, F. & Seman, M. CD3 activation induces concentrative nucleoside transport in human T lymphocytes. Eur. J. Immunol. 30, 366–370 (2000). 42. Soler, C. et al. Regulation of nucleoside transport by lipopolysaccharide, phorbol esters, and tumor necrosis factor-α in human B-lymphocytes. J. Biol. Chem. 273, 26939–26945 (1998). 43. Soler, C. et al. Macrophages require different nucleoside transport systems for proliferation and activation. FASEB J. 15, 1979–1988 (2001). 44. Smith, C. L., Pilarski, L. M., Egerton, M. L. & Wiley, J. S. Nucleoside transport and proliferative rate in human thymocytes and lymphocytes. Blood 74, 2038–2042 (1989). 45. Jarvis, M. F. & Khakh, B. S. ATP-gated P2X cationchannels. Neuropharmacology 56, 208–215 (2009). 46. Janetopoulos, C. & Firtel, R. A. Directional sensing during chemotaxis. FEBS Lett. 582, 2075–2085 (2008). This article provides an overview of the current knowledge of the signalling requirements that are needed for efficient chemotaxis. 47. Choudhuri, K. & van der Merwe, P. A. Molecular mechanisms involved in T cell receptor triggering. Semin. Immunol. 19, 255–261 (2007). 48. Corriden, R. et al. Ecto-nucleoside triphosphate diphosphohydrolase 1 (E-NTPDase1/CD39) regulates neutrophil chemotaxis by hydrolyzing released ATP to adenosine. J. Biol. Chem. 283, 28480–28486 (2008). 49. Junger, W. G. Purinergic regulation of neutrophil chemotaxis. Cell. Mol. Life Sci. 65, 2528–2540 (2008). 50. Cronstein, B. N. Adenosine, an endogenous antiinflammatory agent. J. Appl. Physiol. 76, 5–13 (1994). 51. Fredholm, B. B. Purines and neutrophil leukocytes. Gen. Pharmacol. 28, 345–350 (1997). 52. Ward, P. A., Cunningham, T. W., McCulloch, K. K. & Johnson, K. J. Regulatory effects of adenosine and adenine nucleotides on oxygen radical responses of neutrophils. Lab. Invest. 58, 438–447 (1988). 53. Zhang, Y., Palmblad, J. & Fredholm, B. B. Biphasic effect of ATP on neutrophil functions mediated by P2U and adenosine A2A receptors. Biochem. Pharmacol. 51, 957–965 (1996). 54. de la Harpe, J. & Nathan, C. F. Adenosine regulates the respiratory burst of cytokine-triggered human neutrophils adherent to biologic surfaces. J. Immunol. 143, 596–602 (1989). 55. Meshki, J., Tuluc, F., Bredetean, O., Ding, Z. & Kunapuli, S. P. Molecular mechanism of nucleotideinduced primary granule release in human neutrophils: role for the P2Y2 receptor. Am. J. Physiol. Cell Physiol. 286, C264–C271 (2004). 56. Verghese, M. W., Kneisler, T. B. & Boucheron, J. A. P2U agonists induce chemotaxis and actin polymerization in human neutrophils and differentiated HL60 cells. J. Biol. Chem. 271, 15597–15601 (1996).

57. Kronlage, M. et al. Autocrine purinergic receptor signaling is essential for macrophage chemotaxis. Sci. Signal. 3, ra55 (2010). This report demonstrates that macrophage chemotaxis is regulated by autocrine purinergic signalling. 58. Orr, A. G., Orr, A. L., Li, X. J., Gross, R. E. & Traynelis, S. F. Adenosine A2A receptor mediates microglial process retraction. Nature Neurosci. 12, 872–878 (2009). 59. Kukulski, F. et al. Extracellular ATP and P2 receptors are required for IL-8 to induce neutrophil migration. Cytokine 46, 166–170 (2009). 60. Lecut, C. et al. P2X1 ion channels promote neutrophil chemotaxis through Rho kinase activation. J. Immunol. 183, 2801–2909 (2009). 61. Honda, S. et al. Extracellular ATP or ADP induce chemotaxis of cultured microglia through Gi/o-coupled P2Y receptors. J. Neurosci. 21, 1975–1982 (2001). This is the first report demonstrating chemotaxis of microglia towards ATP in an ATP concentration gradient field. 62. Inoue, K. Purinergic systems in microglia. Cell. Mol. Life Sci. 65, 3074–3080 (2008). 63. Davalos, D. et al. ATP mediates rapid microglial response to local brain injury in vivo. Nature Neurosci. 8, 752–758 (2005). 64. Nimmerjahn, A., Kirchhoff, F. & Helmchen, F. Resting microglial cells are highly dynamic surveillants of brain parenchyma in vivo. Science 308, 1314–1318 (2005). 65. Haynes, S. E. et al. The P2Y12 receptor regulates microglial activation by extracellular nucleotides. Nature Neurosci. 9, 1512–1519 (2006). 66. Koizumi, S. et al. UDP acting at P2Y6 receptors is a mediator of microglial phagocytosis. Nature 446, 1091–1095 (2007). 67. McCloskey, M. A., Fan, Y. & Luther, S. Chemotaxis of rat mast cells toward adenine nucleotides. J. Immunol. 163, 970–977 (1999). 68. Liu, Q. H. et al. Expression and a role of functionally coupled P2Y receptors in human dendritic cells. FEBS Lett. 445, 402–408 (1999). 69. Schnurr, M. et al. ATP gradients inhibit the migratory capacity of specific human dendritic cell types: implications for P2Y11 receptor signaling. Blood 102, 613–620 (2003). 70. Müller, T. et al. The purinergic receptor P2Y2 receptor mediates chemotaxis of dendritic cells and eosinophils in allergic lung inflammation. Allergy 65, 1545–1553 (2010). 71. Piccini, A. et al. ATP is released by monocytes stimulated with pathogen-sensing receptor ligands and induces IL-1β and IL-18 secretion in an autocrine way. Proc. Natl Acad. Sci. USA 105, 8067–8072 (2008). 72. Kukulski, F. et al. Endothelial P2Y2 receptor regulates LPS-induced neutrophil transendothelial migration in vitro. Mol. Immunol. 47, 991–999 (2010). 73. Ben Yebdri, F., Kukulski, F., Tremblay, A. & Sévigny, J. Concomitant activation of P2Y2 and P2Y6 receptors on monocytes is required for TLR1/2-induced neutrophil migration by regulating IL-8 secretion. Eur. J. Immunol. 39, 2885–2894 (2009). 74. Hyman, M. C. et al. Self-regulation of inflammatory cell trafficking in mice by the leukocyte surface apyrase CD39. J. Clin. Invest. 119, 1136–1149 (2009). 75. McDonald, B. et al. Intravascular danger signals guide neutrophils to sites of sterile inflammation. Science 330, 362–366 (2010). This article demonstrates that chemotaxis of neutrophils to inflammatory sites involves a complex network of short-distance and long-distance signals, but shows that ATP released at inflammatory sites is not directly involved in neutrophil chemotaxis. 76. Chekeni, F. B. et al. Pannexin 1 channels mediate ‘find-me’ signal release and membrane permeability during apoptosis. Nature 467, 863–867 (2010). This paper describes the mechanisms by which apoptotic cells release cellular ATP that serves as a danger signal. 77. Valitutti, S., Müller, S., Cella, M., Padovan, E. & Lanzavecchia, A. Serial triggering of many T-cell receptors by a few peptide–MHC complexes. Nature 375, 148–151 (1995). 78. van der Merwe, P. A. The TCR triggering puzzle. Immunity 14, 665–668 (2001). 79. Ma, Z. & Finkel, T. H. T cell receptor triggering by force. Trends Immunol. 31, 1–6 (2010).

NATuRe RevIewS | Immunology

80. 81.

82. 83.

84.

85.

86. 87. 88.

89.

90.

91. 92. 93.

94.

95. 96. 97.

98. 99.

100.

101.

This review article presents evidence that mechanical forces, which are known to induce ATP release from many cell types, may have an important role in signal amplification at the immune synapse. Canaday, D. H. et al. ATP and control of intracellular growth of mycobacteria by T cells. Infect. Immun. 70, 6456–6459 (2002). Into, T., Okada, K., Inoue, N., Yasuda, M. & Shibata, K. Extracellular ATP regulates cell death of lymphocytes and monocytes induced by membrane-bound lipoproteins of Mycoplasma fermentans and Mycoplasma salivarium. Microbiol. Immunol. 46, 667–675 (2002). Di Virgilio, F. et al. Nucleotide receptors: an emerging family of regulatory molecules in blood cells. Blood 97, 587–600 (2001). Wang, L., Jacobsen, S. E., Bengtsson, A. & Erlinge, D. P2 receptor mRNA expression profiles in human lymphocytes, monocytes and CD34+ stem and progenitor cells. BMC Immunol. 5, 16 (2004). Leal, D. B. et al. Characterization of NTPDase (NTPDase1; ectoapyrase; ecto-diphosphohydrolase; CD39; EC3.6.1.5) activity in human lymphocytes. Biochim. Biophys. Acta 1721, 9–15 (2005). Lee, D. H., Park, K. S., Kong, I. D., Kim, J. W. & Han, B. G. Expression of P2 receptors in human B cells and Epstein-Barr virus-transformed lymphoblastoid cell lines. BMC Immunol. 7, 22 (2006). Beldi, G. et al. Deletion of CD39 on natural killer cells attenuates hepatic ischemia/reperfusion injury in mice. Hepatology 51, 1702–1711 (2010). Padeh, S., Cohen, A. & Roifman, C. M. ATP-induced activation of human B lymphocytes via P2-purinoceptors. J. Immunol. 146, 1626–1632 (1991). Sakowicz-Burkiewicz, M., Kocbuch, K., Grden, M., Szutowicz, A. & Pawelczyk, T. Adenosine 5ʹ-triphosphate is the predominant source of peripheral adenosine in human B lymphoblasts. J. Physiol. Pharmacol. 61, 491–499 (2010). Gorini, S. et al. ATP secreted by endothelial cells blocks CX3CL1-elicited natural killer cell chemotaxis and cytotoxicity via P2Y11 receptor activation. Blood 116, 4492–4500 (2010). Sumi, Y. et al. Adrenergic receptor activation involves ATP release and feedback through purinergic receptors. Am. J. Physiol. Cell Physiol. 299, C1118–C1126 (2010). Peter, C., Wesselborg, S. & Lauber, K. Molecular suicide notes: last call from apoptosing cells. J. Mol. Cell. Biol. 2, 78–80 (2010). Zhang, Q. et al. Circulating mitochondrial DAMPs cause inflammatory responses to injury. Nature 464, 104–107 (2010). Franchi, L., Warner, N., Viani, K. & Nuñez, G. Function of Nod-like receptors in microbial recognition and host defense. Immunol. Rev. 227, 106–128 (2009). Ghiringhelli, F. et al. Activation of the NLRP3 inflammasome in dendritic cells induces IL-1β-dependent adaptive immunity against tumors. Nature Med. 15, 1170–1178 (2009). Boeynaems, J. M. & Communi, D. Modulation of inflammation by extracellular nucleotides. J. Invest. Dermatol. 126, 943–944 (2006). Di Virgilio, F., Boeynaems, J. M. & Robson, S. C. Extracellular nucleotides as negative modulators of immunity. Curr. Opin. Pharmacol. 9, 507–713 (2009). Kaufmann, A. et al. “Host tissue damage” signal ATP promotes non-directional migration and negatively regulates Toll-like receptor signaling in human monocytes. J. Biol. Chem. 280, 32459–32467 (2005). Duhant, X. et al. Extracellular adenine nucleotides inhibit the activation of human CD4+ T lymphocytes. J. Immunol. 169, 15–21 (2002). Chen, Y., Shukla, A., Namiki, S., Insel, P. A. & Junger, W. G. A putative osmoreceptor system that controls neutrophil function through the release of ATP, its conversion to adenosine, and activation of A2 adenosine and P2 receptors. J. Leukoc. Biol. 76, 245–253 (2004). Chen, Y., Hashiguchi, N., Yip, L. & Junger, W. G. Hypertonic saline enhances neutrophil elastase release through activation of P2 and A3 receptors. Am. J. Physiol. Cell Physiol. 290, C1051–C1059 (2006). Yip, L. et al. Hypertonic stress regulates T-cell function by the opposing actions of extracellular adenosine triphosphate and adenosine. Shock 27, 242–250 (2007).

vOLuMe 11 | MARCh 2011 | 211 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 102. Mizumoto, N. et al. CD39 is the dominant Langerhans cell-associated ecto-NTPDase: modulatory roles in inflammation and immune responsiveness. Nature Med. 8, 358–365 (2002). 103. Stagg, J. & Smyth, M. J. Extracellular adenosine triphosphate and adenosine in cancer. Oncogene 29, 5346–5358 (2010). 104. Otha, A. et al. A2A adenosine receptor protects tumors from antitumor T cells. Proc. Natl Acad. Sci. USA 103, 13132–13137 (2006). 105. Ohtsuka, T. et al. Ecto-5ʹ-nucleotidase (CD73) attenuates allograft airway rejection through adenosine 2A receptor stimulation. J. Immunol. 185, 1321–1329 (2010). 106. Sevigny, C. P. et al. Activation of adenosine 2A receptors attenuates allograft rejection and alloantigen recognition. J. Immunol. 178, 4240–4249 (2007). 107. Deaglio, S. et al. Adenosine generation catalyzed by CD39 and CD73 expressed on regulatory T cells mediates immune suppression. J. Exp. Med. 204, 1257–1265 (2007). 108. Sun, X. et al. CD39/ENTPD1 expression by CD4+Foxp3+ regulatory T cells promotes hepatic metastatic tumor growth in mice. Gastroenterology 139,1030–1040 (2010). 109. Inoue, Y., Chen, Y., Hirsh, M. I., Yip, L. & Junger, W. G. A3 and P2Y2 receptors control the recruitment of neutrophils to the lungs in a mouse model of sepsis. Shock 30, 173–177 (2008). 110. Fredholm, B. B. Adenosine receptors as drug targets. Exp. Cell Res. 316, 1284–1288 (2010). 111. Price, M. J. Bedside evaluation of thienopyridine antiplatelet therapy. Circulation 119, 2625–2632 (2009). 112. Kam, P. C. & Nethery, C. M. The thienopyridine derivatives (platelet adenosine diphosphate receptor

antagonists), pharmacology and clinical developments. Anaesthesia 58, 28–35 (2003). 113. Jacobson, K. A. & Gao, Z. G. Adenosine receptors as therapeutic targets. Nature Rev. Drug Discov. 5, 247–264 (2006). 114. Baraldi, P. G., Tabrizi, M. A., Gessi, S. & Borea, P. A. Adenosine receptor antagonists: translating medicinal chemistry and pharmacology into clinical utility. Chem. Rev. 108, 238–263 (2008). 115. Haskó, G., Linden, J., Cronstein, B. & Pacher, P. Adenosine receptors: therapeutic aspects for inflammatory and immune diseases. Nature Rev. Drug Discov. 7, 759–770 (2008). This review article summarizes the roles of P1 receptors in immune disorders and discusses possible therapeutic strategies for targeting these receptors. 116. Cronstein, B. N. Low-dose methotrexate: a mainstay in the treatment of rheumatoid arthritis. Pharmacol. Rev. 57, 163–172 (2005). 117. Cronstein, B. N., Naime, D. & Ostad, E. The antiinflammatory mechanism of methotrexate. Increased adenosine release at inflamed sites diminishes leukocyte accumulation in an in vivo model of inflammation. J. Clin. Invest. 92, 2675–2682 (1993). 118. Baharav, E. et al. Antiinflammatory effect of A3 adenosine receptor agonists in murine autoimmune arthritis models. J. Rheumatol. 32, 469–476 (2005). 119. Bar-Yehuda, S. et al. The anti-inflammatory effect of A3 adenosine receptor agonists: a novel targeted therapy for rheumatoid arthritis. Expert Opin. Investig. Drugs 16, 1601–1613 (2007). 120. Silverman, M. H. et al. Clinical evidence for utilization of the A3 adenosine receptor as a target to treat rheumatoid arthritis: data from a phase II clinical trial. J. Rheumatol. 35, 41–48 (2008).

212 | MARCh 2011 | vOLuMe 11

121. Corriden, R., Insel, P. A. & Junger, W. G. A novel method using fluorescence microscopy for real-time assessment of ATP release from individual cells. Am. J. Physiol. Cell Physiol. 293, C1420–C1425 (2007). 122. Pellegatti, P., Falzoni, S., Pinton, P., Rizzuto, R. & Di Virgilio, F. A novel recombinant plasma membrane-targeted luciferase reveals a new pathway for ATP secretion. Mol. Biol. Cell 16, 3659–3665 (2005). 123. Pellegatti, P. et al. Increased level of extracellular ATP at tumor sites: in vivo imaging with plasma membrane luciferase. PLoS ONE 3, e2599 (2008).

Acknowledgements

I acknowledge with great appreciation the work of my many colleagues in this field, even though much of their important work could not be cited here. I also thank my co-workers and colleagues. Special gratitude goes to Y. Chen, L. Yip, T. Woehrle, Y. Inoue, Y. Sumi and N. Hashiguchi, and to my close collaborators S. Robson and P. Insel. I also acknowledge the major funding sources that have supported the work in my laboratory: US National Institutes of Health grants GM-51477, GM-60475, AI-072287, AI-080582 and Congressionally Directed Medical Research Programs grant PR043034.

Competing interests statement

The author declares no competing financial interests.

FURTHER INFORMATION Wolfgang g. Junger’s homepage: http://www.bidmc.org/Research/Departments/Surgery/ TraumaSurgery/Jungerlaboratory.aspx All lInKs ARE ACTIvE In ThE onlInE PDf

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

Emerging inflammasome effector mechanisms Mohamed Lamkanfi

Abstract | Caspase 1 activation by inflammasome complexes in response to pathogenassociated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs) induces the maturation and secretion of the pro-inflammatory cytokines interleukin-1β (IL-1β) and IL-18. Recent reports have begun to identify additional inflammasome effector mechanisms that proceed independently of IL-1β and IL-18. These include the induction of pyroptotic cell death, the restriction of bacterial replication, the activation of lipid metabolic pathways for cell repair and the secretion of DAMPs and leaderless cytokines. These non-canonical functions of caspase 1 illustrate the diverse mechanisms by which inflammasomes might contribute to innate immunity, repair responses and host defence. Cryopyrinopathies A spectrum of hereditary autoinflammatory diseases that are caused by mutations in the gene encoding NLR family, pyrin domaincontaining 3 (NLRP3) that trigger continuous activation of the NLRP3 inflammasome. Based on the severity and spectrum of the symptoms — which can include urticarial skin rashes, prolonged episodes of fever, sensorineural hearing loss, headaches, cognitive deficits and renal amyloidosis — these diseases are classified as familial cold autoinflammatory syndrome, Muckle–Wells syndrome or chronic infantile neurological cutaneous articular syndrome.

Department of Biochemistry, Ghent University, and VIB Department of Medical Protein Research, Albert Baertsoenkaai 3, B‑9000 Ghent, Belgium. e‑mail: Mohamed.Lamkanfi@ VIB‑UGent.be doi:10.1038/nri2936

Inflammasomes (BOX 1) are emerging as key regulators of the innate immune response, and the activity of these multiprotein complexes has been linked to inflamma‑ tory bowel diseases1–5, vitiligo6, gouty arthritis7, type 1 and type 2 diabetes8,9, and less common autoinflamma‑ tory disorders that are collectively referred to as cryopyrinopathies10,11. Inflammasome complexes are thought to be assembled around members of the NOD-like receptor (NLR) or HIN‑200 protein families 12 (FIG. 1). These receptors are thought to detect microbial pathogenassociated molecular patterns (PAMPs) and endogenous damage-associated molecular patterns (DAMPs) in intra‑ cellular compartments, similar to the role of mammalian Toll‑like receptors (TLRs) at the cell surface and within endosomes13. Although it is incompletely understood how NLRs detect microbial ligands and DAMPs14,15, it is evident that inflammasome assembly results in the activation of caspase 1 (BOX 2). This evolutionarily con‑ served cysteine protease is mainly known for its role in the maturation of the pro‑inflammatory cytokines interleukin‑1β (IL‑1β) and IL‑18 (ReFs 16–19). IL‑1β and IL‑18 are related cytokines that are pro‑ duced as cytosolic precursors and usually require caspase 1‑mediated cleavage for full activation and secretion16–19. However, additional proteases, including caspase 8, myeloblastin (also known as proteinase 3) and granzyme A, have been shown to convert pro‑IL‑1β into a biologically active cytokine in several established mouse models of human disease20–23. This indicates that caspase 1 is not always required for the maturation of IL‑1β, and such redundancy in controlling IL‑1β matura‑ tion might safeguard the host immune response against

viral and bacterial pathogens that target caspase 1 activa‑ tion in inflammasomes24. Indeed, IL‑1β and IL‑18 were recognized early on for their ability to cause a wide variety of biological effects associated with infection, inflamma‑ tion and autoimmunity 25. IL‑1β regulates systemic and local responses to infection, injury and immunological challenge by generating fever, activating lymphocytes and promoting leukocyte transmigration into sites of injury or infection25. Although IL‑18 lacks the pyrogenic activity of IL‑1β, it induces interferon‑γ (IFNγ) production by activated T cells and natural killer cells in the presence of IL‑12, thereby contributing to T helper 1 (TH1) cell polarization25,26. In the absence of IL‑12, IL‑18 can pro‑ mote TH2 cell responses through the production of TH2 cell cytokines such as IL‑4, IL‑5 and IL‑10 (ReFs 26–28). More recently, IL‑18 has also been implicated in driving TH17 cell responses because it synergizes with IL‑23 to induce IL‑17 production by already committed TH17 cells29,30. Thus, IL‑1β and IL‑18 are important inflammasome effector molecules, as illustrated by the marked response to therapy with IL‑1 inhibitors found in patients with cryopyrinopathies, who have increased inflammasome activation31,32. However, not all inflammasome functions can be abrogated by neutralization of IL‑1β and IL‑18. For exam‑ ple, caspase 1‑deficient mice are resistant to lipopoly‑ saccharide (LPS)‑induced shock, whereas mice lacking both IL‑1β and IL‑18 are susceptible33. Moreover, a recent study showed that IL‑1β and IL‑18 are not required for caspase 1‑mediated clearance of several bacterial patho‑ gens (namely, modified Salmonella enterica subsp. enterica serovar Typhimurium strains that constitutively express

NATuRe RevIewS | Immunology

voLuMe 11 | MARcH 2011 | 213 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Box 1 | Inflammasomes Inflammasomes are intracellular multiprotein complexes that mediate the proximity-induced autoactivation of caspase 1. Inflammasome-mediated caspase 1 activation has been shown to occur in macrophages, dendritic cells, epithelial cells and possibly other cell types during bacterial, viral, fungal and parasitic infections. Inflammasomes are activated in response to stimulation with damage-associated molecular patterns (DAMPs), such as uric acid and ATP, and upon exposure to crystalline substances, such as monosodium urate, silica and asbestos particles12,24. The molecular composition of inflammasome complexes is stimulus dependent, with certain members of the NOD-like receptor (NLR) and HIN-200 receptor families functioning as the activating platform in these complexes. Genetic studies in mice indicate the existence of at least four types of inflammasome (FIG. 1). Three of these contain NLR proteins, namely NLR family, pyrin domain-containing 1B (NLRP1B), NLR family, CARD-containing 4 (NLRC4) and NLRP3. The fourth type of inflammasome contains the HIN-200 protein absent in melanoma 2 (AIM2)12. The bipartite adaptor protein ASC (apoptosis-associated speck-like protein containing a CARD; also known as PYCARD) probably has a key role in inflammasome assembly and caspase 1 activation by bridging the interaction between NLRs or HIN-200 proteins and caspase 1, although the precise role of ASC in the activation of the NLRP1B and NLRC4 inflammasomes is debated. NLRP1B and NLRC4 contain a caspase recruitment domain (CARD) at their carboxyl and amino termini, respectively (unlike AIM2 and NLRP3, which have a pyrin domain) and can therefore interact directly with caspase 1 when overexpressed, without requiring ASC. However, evidence of a role for ASC in the activation of the endogenous NLRC4 inflammasome is provided by the observation that robust caspase 1 activation and the production of interleukin-1β (IL-1β) and IL-18 are markedly decreased in ASC-deficient macrophages infected with viral or bacterial pathogens, or exposed to a variety of DAMPs and crystalline substances50,63,88.

flagellin, Legionella pneumophila and Burkholderia thailandensis)34. Similarly, caspase 1‑deficient mice are more susceptible to infection with Francisella tularensis than mice lacking both IL‑1β and IL‑18 (ReF. 35), and this indicates that additional caspase 1‑dependent mecha‑ nisms might contribute to the control of infection. In this regard, several recent publications have begun to characterize a range of new inflammasome functions and effector molecules that seem to operate independently of IL‑1β and IL‑18. In this article, I review these emerg‑ ing non‑canonical inflammasome effector mechanisms and attempt to illustrate how they might contribute to immune responses.

Proximity-induced autoactivation A process in which two or more initiator caspases are brought sufficiently close to induce their autocatalytic activation. This process is thought to occur in large cytosolic protein complexes to which caspase zymogens are recruited by means of homotypic interactions between the caspase recruitment domain (CARD) or death effector domain (DeD) motifs in their pro-domains and several bipartite adaptor molecules.

Unconventional protein secretion Secretory proteins usually contain amino‑terminal or internal signal peptides that target them to the translo‑ cation apparatus of the endoplasmic reticulum (eR)36,37. From the eR lumen, such proteins are transported to the Golgi apparatus and then to the extracellular space in Golgi‑derived secretory vesicles that fuse with the plasma membrane37. This pathway of protein export is known as the ‘eR–Golgi’ or ‘classical’ secretory pathway. However, cytoplasmic and nuclear proteins that lack an eR‑targeting signal peptide can exit cells through eR‑ and Golgi‑independent pathways38. For example, mature IL‑1β was shown to be secreted independently of the eR and the Golgi apparatus more than 20 years ago39. The number of proteins that have been shown to be released by unconventional protein secretion has grown to more than 20, including fibroblast growth factor 2 (FGF2), the lectins galectin 1 and galectin 3, and possibly the DAMP high‑mobility group box 1 (HMGB1)38. After their release into the extracellular space, these effectors can enhance

inflammatory, cell survival and repair responses through activation of cell surface receptors, such as FGF recep‑ tor 1, the IL‑1 and IL‑18 receptors and the receptor for advanced glycation end‑products (RAGe). Although the molecular mechanism by which IL‑1β, IL‑18 and other proteins that lack signal peptides are secreted remains obscure, several models have been proposed to explain the release of such ‘leaderless pro‑ teins’ in microvesicles that are shed from the plasma membrane, or in secretory lysosomes or exosomes38. Interestingly, adherent monocytes from caspase 1‑ deficient mice and peritoneal macrophages from mice lacking two inflammasome components — namely, NLR family, pyrin domain‑containing 3 (NLRP3) and apoptosis‑associated speck‑like protein containing a cARD (ASc; also known as PYcARD) — not only failed to secrete IL‑1β and IL‑18 after LPS stimulation16,18,19,40, but were also partially defective in the secretion of the leaderless cytokine IL‑1α17,40. unlike IL‑1β and IL‑18, IL‑1α does not undergo caspase 1‑mediated cleavage26. This might indicate that caspase 1 modulates IL‑1α secretion indirectly by regulating the secretory path‑ way of this cytokine, and may point to a broader role for caspase 1 in regulating unconventional protein secretion. Indeed, pharmacological inhibition, RNA interference (RNAi)‑mediated downregulation and targeted deletion of caspase 1 were all recently shown to block the secre‑ tion of IL‑1β, IL‑1α and FGF2 by macrophages, uvA‑ irradiated fibroblasts and uvB‑irradiated keratinocytes, respectively 41. In addition, caspase 1 activation by either the NLRP3 inflammasome or the NLR family, cARD‑ containing 4 (NLRc4) inflammasome was required for secretion of the nuclear DAMP HMGB1 from activated and infected macrophages33. Because the enzymatic activity of caspase 1 is required for the secretion of each of these leaderless proteins33,41, caspase 1 might medi‑ ate the proteolytic activation of a secretion apparatus of unknown identity. In this context, the small GTPase RAB39A was recently characterized as a caspase 1 sub‑ strate that is involved in the secretion of IL‑1β from LPS‑activated human THP‑1 monocytes42. However, it remains to be determined how caspase 1‑mediated cleav‑ age affects RAB39A function and whether RAB39A has a role in the secretion of additional leaderless proteins. An alternative mechanism by which caspase 1 might promote the release of leaderless proteins such as HMGB1 involves the induction of a specialized caspase 1‑mediated cell death programme in activated immune cells, which is often referred to as pyroptosis (see below). elucidation of the mechanism awaits the characterization of the molecular components mediating caspase 1‑dependent unconventional protein secretion and pyroptosis.

Pyroptosis Most caspases (BOX 2) — caspases 2, 3 and 6–10 — are implicated in the induction and execution of apopto‑ sis43,44. This form of programmed cell death is respon‑ sible for organ shaping during embryonic development and for preserving homeostasis in adult organisms. Typical morphological features of apoptosis include plasma membrane blebbing, condensation of the nucleus,

214 | MARcH 2011 | voLuMe 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

$CPVJTCEKU NGVJCNVQZKP

56[RJKOWTKWO 2CGTWIKPQUC .RPGWOQRJKNC 5ƓGZPGTK

78KTTCFKCVKQP /KETQDKCN2#/2U 'PFQIGPQWU&#/2U %T[UVCNU r+QPȯWZ! r4GCEVKXGQZ[IGP

URGEKGU!

!

r.[UQUQOCN

%[VQUQNKE FU&0# 0.42

0.42

#+/

#+/

2;&

2;&

2;&

2;&

%#4& 0.4% !

2;&

!

2;&

0.42+$ %#4&

%#4& 0.4%

0.42+$ %#4&

RTQVGCUGU!

.44U

&0#XKTWUGU (VWNCTGPUKU .OQPQE[VQIGPGU

%#4&

%#4&

%#4&

%#4&

#5%

%#4&

#EVKXG ECURCUG

%#4&

2TQECURCUG

Figure 1 | Inflammasomes: composition and stimuli. The NOD-like receptor (NLR) proteins NLR family, pyrin domain-containing 1B (NLRP1B), NLR family, CARD-containing 4 (NLRC4) and NLRP3, and the HIN-200 protein absent in melanoma 2 (AIM2) assemble inflammasomes in a stimulus-specific manner12. NLRP1B recognizes the cytosolic presence of the Bacillus anthracis lethal toxin64. The NLRC4 inflammasome is assembled after detection of bacterial flagellin or the basal body rod component of the bacterial type III and type IV secretion systems of Salmonella enterica subsp. enterica serovar Typhimurium, Pseudomonas aeruginosa, Legionella pneumophila and Shigella flexneri12,58,62. NLRP3 is activated when macrophages are exposed to UV irradiation, microbial pathogenassociated molecular patterns (PAMPs), endogenous damage-associated molecular patterns (DAMPs) such as ATP, or crystals such as monosodium urate, silica and asbestos. Recognition of these PAMPs, DAMPs and crystals is thought to involve the detection of a common secondary messenger, such as K+ fluxes, reactive oxygen species or lysosomal proteases14,15. By contrast, AIM2 directly binds double-stranded DNA (dsDNA) in the cytosol to induce caspase 1 activation in cells infected with Francisella tularensis, Listeria monocytogenes or DNA viruses such as cytomegalovirus and vaccinia virus68–70,72–74. The adaptor protein ASC (apoptosis-associated speck-like protein containing a CARD; also known as PYCARD) is probably required for the full activation of all inflammasome complexes, although its role in the NLRP1B and NLRC4 inflammasomes is still debated45. CARD, caspase recruitment domain; LRR, leucine-rich repeat. PYD, pyrin domain.

DNA fragmentation and general shrinkage of the cell volume45. Apoptotic cells usually fail to elicit inflam‑ matory responses because the cytoplasmic content is shielded from the extracellular environment by packag‑ ing in ‘apoptotic bodies’ (FIG. 2). These apoptotic bodies are membrane‑bound cell fragments that are rapidly phagocytosed in vivo by neighbouring cells and resident phagocytes46,47. The integrity of the envelope surround‑ ing apoptotic bodies is usually preserved until after they have been engulfed by professional antigen‑presenting cells or neighbouring cells to prevent accidental spilling

of the intracellular content into the extracellular milieu47. Apoptotic cells can also prevent the accidental induction of inflammation by inactivating the immunostimulatory activity of the DAMP HMGB1 through the oxidation of residue cys106 (ReF. 48). Finally, the uptake of apoptotic bodies by macrophages and dendritic cells was recently proposed to actively suppress antigen presentation and the secretion of inflammatory cytokines by these cells49. unlike most caspases, caspase 1 is not involved in the induction of apoptosis. Instead, caspase 1 activation in neurons, macrophages and dendritic cells drives a spe‑ cialized form of cell death known as pyroptosis34,50–53. caspase 1‑dependent cell death was first reported to occur in macrophages infected with Shigella flexneri, the causative agent of bacillary dysentery 54,55. Pyroptosis was subsequently observed in macrophages and dendritic cells infected with the facultative intracellular pathogens S. Typhimurium, Pseudomonas aeruginosa and L. pneumophila45,56,57. each of these pathogens induces caspase 1 activation through the NLRc4 inflammasome (FIG. 1). These pathogens are recognized when bacterial flagel‑ lin is transferred through specialized bacterial type III and type IV secretion systems into the host cell cytosol or upon detection of the basal body rod component of the S. flexneri or P. aeruginosa type III secretion appara‑ tus58–63. The induction of pyroptosis is not restricted to the NLRc4 inflammasome, as Bacillus anthracis infec‑ tion induces the pyroptotic cell death of mouse macro‑ phages through the NLRP1B inflammasome64,65. This occurs when the anthrax metalloprotease lethal factor gains access to the cytosol of susceptible macrophages64. Notably, lethal factor‑induced pyroptosis was shown to confer resistance to infection with B. anthracis spores in vivo65, highlighting the importance of pyroptosis for host defence against pathogens. The pyroptotic cell death of macrophages infected with Staphylococcus aureus requires activation of the NLRP3 inflammasome66,67. Although the precise mechanism is still debated, acti‑ vation of this inflammasome could proceed through several mutually non‑exclusive mechanisms, including K+ efflux, the generation of reactive oxygen species, lyso‑ somal destabilization and the translocation of microbial ligands into the host cytosol14,15. Finally, infections with the bacterial pathogens Listeria monocytogenes and F. tularensis induce pyroptosis upon their recognition in the host cell cytosol by the absent in melanoma 2 (AIM2) inflammasome68–74. The role of caspase 1 in the immune response to F. tularensis, the causative agent of tularaemia, is illustrated by the observation that caspase 1‑deficient mice have increased susceptibility to infection with this pathogen75. Notably, mice lacking the canonical inflammasome substrates IL‑1β and IL‑18 are less susceptible to infection with F. tularensis than caspase 1‑deficient mice 35, and this indicates that additional caspase 1‑dependent mechanisms, such as pyroptosis, might make an important contribution to the control of F. tularensis infection. Indeed, although caspase 1 activity is required for pyroptosis51,76, this form of cell death proceeds independently of IL‑1β and IL‑18 (ReF. 56). A recent study established the cru‑ cial in vivo role of pyroptosis in clearing a modified

NATuRe RevIewS | Immunology

voLuMe 11 | MARcH 2011 | 215 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS Box 2 | Caspases Caspases are an evolutionarily conserved family of metazoan cysteine proteases with 11 representatives in humans: caspases 1–10 and caspase 14. Caspases cleave various cellular substrates after aspartic acid residues and have essential roles in apoptosis, inflammation, cell proliferation and cell differentiation89. For example, caspase 3mediated cleavage of mammalian STE20-like kinase 1 (MST1; also known as STK4) was shown to be crucial for skeletal muscle differentiation, and caspase 8-mediated cleavage of the long splice variant of cellular FLICE-like inhibitory protein (cFLIPL) regulates the balance between apoptosis induction and nuclear factor-κB (NF-κB)-driven T cell proliferation90. All caspases are synthesized as zymogens consisting of an amino-terminal pro-domain of variable length and a carboxy-terminal protease domain. Caspases can be subdivided according to the length of their pro-domain (see the figure). Initiator caspases (such as caspases 1 and 8) have large pro-domains containing homotypic protein–protein interaction motifs of the death domain superfamily, specifically either a caspase recruitment domain (CARD) or a death effector domain (DED). These interaction motifs allow the recruitment of pro-caspases into multiprotein complexes (such as the inflammasomes) by homotypic interactions with adaptor molecules such as ASC (apoptosis-associated speck-like protein containing a CARD; also known as PYCARD). Within these complexes, pro-caspases undergo the conformational changes and/or autoprocessing required for their activation91. By contrast, the effector caspases (caspases 3, 6, 7 and 14) have short pro-domains of only a few amino acids and they lack any homotypic interaction motifs. These caspases require proteolytic maturation by the initiator caspases or other proteases to achieve maximum enzymatic activity. Unlike the ‘true’ caspases listed above, human caspase 12 is devoid of enzymatic activity because crucial catalytic residues have been mutated. In addition, most people express a truncated form of caspase 12 that resembles the human CARD-only proteins CARD17 (also known as INCA), CARD18 (also known as ICEBERG) and CARD16 (also known as COP or pseudo-ICE)92. #OKPQ VGTOKPWU

+PKVKCVQT ECURCUGU

%CTDQZ[N VGTOKPWU

2TQFQOCKP

2TQVGCUGFQOCKP

%CURCUGUCPF

%#4&

%CURCUGFQOCKP

%CURCUGUCPF

&'&

%CURCUGFQOCKP

'ȭGEVQT ECURCUGU %CURCUGUCPF

NOD-like receptor (NLR). The human NLR family comprises 22 members. They share a domain organization that usually includes an amino-terminal caspase recruitment domain (CARD) or pyrin domain (PYD), followed by an intermediary nucleotide-binding oligomerization domain (NOD) and carboxy-terminal leucine-rich repeat motifs. NLRs are thought to survey the host cytosol and intracellular compartments for pathogenand damage-associated molecular patterns to activate signalling pathways that contribute to the host innate immune response.

%CURCUGFQOCKP

S. Typhimurium strain that constitutively expressed 0CVWTG4GXKGYU^+OOWPQNQI[ bacterial flagellin34. The pyroptotic cell death of infected macrophages exposed intracellular bacteria to extracel‑ lular immune surveillance and allowed their destruction by neutrophils. Pyroptosis also conferred protection against the bacterial pathogens L. pneumophila and B. thailandensis, which establishes this inflammasome effector mechanism as a crucial component of host defence against intracellular bacterial pathogens34. Although the molecular mechanisms controlling pyroptosis are still poorly defined, this cell death mode differs morphologically from apoptosis in that pores with a diameter of 1–2 nm appear in the plasma membrane of pyroptotic cells at early time points52. The resulting ion fluxes could explain some of the hallmarks of pyroptotic cells, including cytoplasmic swelling, osmotic lysis and the release of the intracellular content into the extracel‑ lular milieu77 (FIG. 2). In addition to eliminating infected immune cells, this process can enhance innate and adap‑ tive immune responses by exposing microbial antigens to surveillance by the immune system. Together with the fact that caspase 1 activation is linked with the produc‑ tion of mature IL‑1β and IL‑18 and the release of leader‑ less cytokines and DAMPs (such as HMGB1 and S100

proteins), these characteristics are thought to render pyroptosis (and associated inflammasome functions) an inherently pro‑inflammatory cell death mode. Despite the different immunological outcomes of apoptosis and pyroptosis (FIG. 2), apoptotic and pyrop‑ totic cells share several prominent morphological and biochemical features. Nuclear condensation and oligo‑ nucleosomal DNA fragmentation are observed in both cell death modes19,46,51,52,76. Moreover, the DNA damage sensor poly(ADP‑ribose) polymerase 1 (PARP1) is proc‑ essed into an 89 kDa fragment during both apoptosis and pyroptosis46,78. under homeostatic conditions, PARP1 participates in genomic DNA repair and DNA replica‑ tion by catalysing the synthesis of poly(ADP‑ribose) in a process that consumes NAD+ and the ATP energy stores of the cell. Thus, cleavage of PARP1 during both apop‑ tosis and pyroptosis indicates that PARP1 inactivation might be a general strategy used by cells undergoing pro‑ grammed cell death, possibly to preserve energy stores in order to allow for proper dismantling of the cell. Finally, maturation of caspase 3 and caspase 7 is observed during both apoptosis and pyroptosis51,78,79, although pyroptosis‑ associated DNA fragmentation, PARP1 processing and plasma membrane permeabilization are not affected in macrophages lacking either of these executioner caspases51,55,78. This might not come as a complete surprise given that caspase 3 and caspase 7 are partially redundant and that deletion of both caspases was necessary for pro‑ tection against apoptosis80. Nevertheless, S. Typhimurium has been shown to induce pyroptosis in macrophages lacking both caspase 3 and caspase 7 (ReF. 45), and this confirms that these executioner caspases are not neces‑ sary for pyroptosis. Thus, although apoptosis and pyrop‑ tosis share the morphological and biochemical features described above (FIG. 2), the signalling pathways involved are distinct. Despite recent progress in characterizing the molecular features of pyroptotic cell death and its role in host defence against bacterial pathogens, much remains to be learned about the signalling mechanism by which caspase 1 induces pyroptosis.

Inhibition of glycolysis In an attempt to identify new caspase 1 substrates that could explain the phenotype of pyroptotic cells, a caspase 1 digestome analysis was carried out, and this identified several crucial enzymes of bioenergetic pathways as potential caspase 1 targets81. Biochemical studies confirmed that the glycolysis enzymes fructose‑ bisphosphate aldolase, glyceraldehyde‑3‑phosphate dehy‑ drogenase, α‑enolase and pyruvate kinase can be cleaved by recombinant caspase 1. These enzymes operate in a metabolic pathway that replenishes cellular ATP energy stores through the conversion of glucose to pyruvate. caspase 1‑mediated processing of glyceraldehyde‑3‑ phosphate dehydrogenase inhibited its enzymatic activ‑ ity 81, and this indicated that caspase 1 might decrease the metabolic rate of infected cells. To address whether caspase 1‑mediated processing of these metabolic enzymes occurred during infection, peritoneal macro‑ phages of wild‑type and caspase 1‑deficient mice were infected with S. Typhimurium. As expected, aldolase was

216 | MARcH 2011 | voLuMe 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 2[TQRVQUKU Pathogen-associated molecular pattern (PAMP). A conserved pathogen molecule that is usually essential for microbial survival, and that contains either nucleic acid structures that are unique to microorganisms or cell wall components (such as lipopolysaccharide and flagellin) that are not found in mammalian cells. PAMPs are ligands for receptors of the host innate immune system.

Damage-associated molecular pattern (DAMP). A molecule that is produced or released from host cells upon cellular stress, damage or non-physiological cell death. DAMPs are also referred to as ‘alarmins’ and are thought to be responsible for the initiation and perpetuation of inflammatory responses and tissue repair under non-infectious (sterile) conditions. examples include high-mobility group box 1 (HMGB1), ATP, uric acid and heat-shock proteins.

Unconventional protein secretion The secretion of cytoplasmic and nuclear proteins into the extracellular space through an incompletely understood mechanism that does not require the translocation apparatus of the classical endoplasmic reticulum (eR)–Golgi secretion pathway. Proteins that are secreted through this route include interleukin-1α (IL-1α), IL-1β, IL-18, fibroblast growth factor 2, galectin 1, galectin 3 and possibly high-mobility group box 1.

Pyroptosis A specialized form of programmed cell death that requires caspase 1 activity. It is characterized by cytoplasmic swelling, early plasma membrane rupture, nuclear condensation and internucleosomal DNA fragmentation. The cytoplasmic content is released into the extracellular space, and this is thought to augment inflammatory and repair responses. Pyroptosis occurs in myeloid cells infected with pathogenic bacteria, and it might affect cells of the central nervous system and the cardiovascular system under ischaemic conditions.

#RQRVQUKU /GODTCPG TWRVWTG

/GODTCPG UYGNNKPI /GODTCPG XGUKENGHQTOCVKQP

2TGUGTXCVKQP QHOGODTCPG KPVGITKV[

%CURCUGCPF CEVKXCVKQP

&0# HTCIOGPVCVKQP

0WENGCT EQPFGPUCVKQP

+.β 4GNGCUGQH KPVTCEGNNWNCT EQPVGPVU

%CURCUGCPF CEVKXCVKQP

%[VQUQNKE UYGNNKPI

&0# HTCIOGPVCVKQP

*/)$

#RQRVQVKEDQF[ HQTOCVKQP

%[VQUQNKE UJTKPMCIG 0WENGCT EQPFGPUCVKQP 2#42ENGCXCIG

2#42ENGCXCIG

+. 2TQKPȯCOOCVQT[ EGNNFGCVJ

+OOWPQNQIKECNN[ UKNGPVEGNNFGCVJ

Figure 2 | main features of pyroptosis. The molecular mechanisms underlying pyroptosis are still poorly defined. Morphologically, pyroptotic cells are characterized by the early loss of plasma membrane integrity77, and this is accompanied 0CVWTG4GXKGYU^+OOWPQNQI[ by the shedding of membrane vesicles93. Pyroptotic and apoptotic cells share several prominent features (shown in blue boxes), including nuclear condensation and internucleosomal DNA fragmentation, cleavage of the DNA damage repair enzyme poly(ADP-ribose) polymerase 1 (PARP1) and activation of the executioner caspases caspase 3 and caspase 7 (ReF. 45). However, the volume of the cytoplasmic compartment of pyroptotic cells increases, whereas apoptosis is characterized by general shrinkage of the cell volume. Together with the role of caspase 1 in cytokine maturation and unconventional protein secretion, the release of the cytoplasmic content into the extracellular milieu during pyroptosis is thought to render this form of cell death inherently pro-inflammatory45. By contrast, apoptosis is usually considered to be immunologically silent because the cytoplasmic content is packaged in apoptotic bodies and these membrane-bound cell fragments are rapidly phagocytosed in vivo by neighbouring cells and resident phagocytes47. HMGB1, high mobility group box 1; IL, interleukin.

processed in infected wild‑type macrophages, but not in those lacking caspase 1 (ReF. 81). concurrently, the rate of glycolysis in the caspase 1‑deficient cells was markedly higher, further supporting an inhibitory role for caspase 1 in the regulation of glycolysis. Because myeloid cells are highly dependent on glycolysis for ATP production82, caspase 1‑mediated inactivation of glycolysis enzymes might restrict intracellular pathogen replication by quickly depleting energy sources and by preparing infected host cells to undergo pyroptosis. However, it remains to be determined whether and to what extent caspase 1‑ mediated inactivation of glycolysis enzymes contributes to protection against bacterial pathogenicity in vivo.

Cell survival Another mechanism by which caspase 1 might protect host cells is by repairing the damage caused by bacte‑ rial pore‑forming toxins that are released by pathogenic bacteria83. The pores formed by these toxins can range significantly in diameter, depending on the nature of the toxin. For example, S. aureus α‑toxin and Aeromonas hydrophila aerolysin typically produce holes in the host cell membrane with a diameter of only 2 nm, whereas Streptococcus pneumoniae and L. monocytogenes pro‑ duce toxins that can create perforations of up to 50 nm in diameter 84. consequently, the latter toxins allow trans‑ location of large proteins across the plasma membrane, whereas S. aureus α‑toxin and A. hydrophila aerolysin

only render the plasma membrane permeable to small inorganic ions84. Nevertheless, depending on the concen‑ tration of these toxins and the targeted cell type, the dam‑ age elicited by these toxins can range from irreparable cell destruction to temporal perforations that can quickly be resealed by the host cell’s dedicated repair machinery. Repairing bacterial toxin‑induced damage to the plasma membrane requires the activation of lipid bio‑ genesis pathways, the molecular machinery of which is controlled by two transcription factors known as sterol regulatory element‑binding protein 1 (SReBP1) and SReBP2 (ReF. 84). Interestingly, activation of SReBP1 and SReBP2 is controlled by the NLRP3 and NLRc4 inflammasomes in fibroblasts that have been intoxicated with S. aureus α‑toxin or A. hydrophila aerolysin, or infected with live Aeromonas trota bacteria83. This path‑ way is thought to promote host cell survival because inhi‑ bition or transcriptional downregulation of SReBP1 and SReBP2 enhanced cell death responses83. Inflammasome‑ induced activation of SReBPs is thought to be indirect, but the caspase 1 substrates that drive this pathway remain to be found. It would be interesting to determine whether caspase 1 activates SReBPs in macrophages infected with live A. hydrophila, because these bacteria were recently shown to activate caspase 1 through the NLRP3, but not the NLRc4, inflammasome85. This might reveal differential signalling mechanisms induced by the three A. hydrophila cytotoxins85.

NATuRe RevIewS | Immunology

voLuMe 11 | MARcH 2011 | 217 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS

Glycolysis A metabolic pathway that generates the cellular high-energy store ATP by oxidizing glucose to pyruvate. In eukaryotic cells, pyruvate is further oxidized into CO2 and H2O in a process known as ‘aerobic respiration’. This results in a net yield of 36–38 molecules of ATP per metabolized molecule of glucose.

Autophagosome A double-membrane-bound vesicle that is used by eukaryotic cells to target protein aggregates, damaged organelles and invading microorganisms for digestion by lysosomal hydrolases. This catabolic process allows recycling of cellular components and is thought to contribute to cell death, cell survival during starvation, cellular differentiation and host defence against infectious agents.

#EVKXG ECURCUG

5GETGVKQPQTTGNGCUGQH NGCFGTNGUUE[VQMKPGU CPF&#/2U

#EVKXCVKQP QH54'$2U

5GETGVKQPQH ITQYVJHCEVQTU

.KRKF OGODTCPG DKQIGPGUKU

+PȯCOOCVKQP

the NLRP3 and NLRc4 inflammasomes in inducing proteolytic maturation of caspase 7 in activated immune cells that were exposed to ATP and nigericin or infected with live S. Typhimurium51. In contrast to caspase 7, the activation of caspase 3 was not affected in caspase 1‑ deficient macrophages51, and this indicates that the activation of caspase 3 and caspase 7 is differentially reg‑ ulated during inflammation. These observations identi‑ fied an alternative mechanism by which inflammasomes might control bacterial infections. Indeed, caspase 7 activation downstream of the NLRc4 inflammasome was subsequently reported in macrophages infected with L. pneumophila 79. Moreover, caspase 7‑deficient macrophages are less capable of restricting intracellular L. pneumophila replication, possibly owing to defects in the fusion of bacteria‑containing phagosomes with lysosomes and the delayed induction of macrophage cell death79. Importantly, caspase 1 and caspase 7 regu‑ late L. pneumophila growth in the lungs of orally infected mice61,79, demonstrating the importance of this inflamma‑ some effector pathway in host defence against this bac‑ terial pathogen. However, it remains to be determined

%#4&

Two of at least six specialized secretion systems by which Gram-negative pathogens can deliver virulence factors into eukaryotic host cells. Pathogenic bacteria such as Shigella, Salmonella, Yersinia, Chlamydia and Pseudomonas spp. all make use of a type III secretion system to infect host cells and to modulate signalling pathways. By contrast, pathogens such as Helicobacter pylori, Legionella pneumophila and Bordetella pertussis make use of a type IV secretion system for the horizontal transfer of plasmid DNA containing antibiotic resistance genes and to inject effector proteins into eukaryotic host cells.

Caspase 7 activation Activation of the NLRc4 inflammasome was recently shown to restrict the intracellular replication of L. pneumophila, the causative agent of a severe form of bacte‑ rial pneumonia known as Legionnaires’ disease61,86,87. Inflammasome activation in resistant mouse strains results in the rapid caspase 1‑dependent delivery of L. pneumophila to lysosomes, where the bacteria are degraded. By contrast, defective inflammasome activa‑ tion in Nlrc4–/– and Casp1–/– mice allows bacterial repli‑ cation in specialized intracellular vesicles that resemble autophagosomes61,87. Notably, inflammasome‑mediated restriction of L. pneumophila replication proceeds independently of IL‑1β and IL‑18 (ReFs 61,87), but the caspase 1 substrates that are responsible for this proc‑ ess remain unclear. A proteome‑wide screen for new caspase 1 targets identified caspase 7, an effector caspase, as a direct substrate of caspase 1, and biochemical studies confirmed that caspase 7 is cleaved by caspase 1 after the canonical activation sites Asp23 and Asp198 (ReF. 51). Importantly, studies in macrophages from Nlrp3–/–, Nlrc4–/–, Asc–/– or Casp1–/– mice confirmed the role of

%#4&

Type III and type IV secretion

%NGCXCIGQH 2#42CPF IN[EQN[UKUGP\[OGU

%NGCXCIGQHCU [GVWPKFGPVKȮGF UWDUVTCVGU

2TQECURCUG

%JCPIGUKPEGNNWNCT GPGTI[UQWTEGU

/GODTCPG RGTOGCDKNK\CVKQP CPF&0# HTCIOGPVCVKQP

%CURCUG

4GRCKTCPFJGCNKPI

2[TQRVQUKU

4GOQXCNQH KPVTCEGNNWNCT TGRNKECVKQPPKEJGU

+OOWPG UWTXGKNNCPEG QHOKETQDKCN EQORQPGPVU

4GUVTKEVKQPQH .GIKQPGNNC TGRNKECVKQPD[ VCTIGVKPIVQ N[UQUQOGU

Figure 3 | Caspase 1 effector mechanisms. Pathogen invasion of macrophages and dendritic cells triggers the assembly of inflammasome complexes and caspase 1 activation. Active caspase 1 induces inflammation by mediating the 0CVWTG4GXKGYU^+OOWPQNQI[ extracellular secretion or release of leaderless cytokines such as interleukin-1β (IL-1β), IL-18 and IL-1α, and possibly damage-associated molecular patterns (DAMPs) such as high mobility group box 1 (HMGB1), through an as yet unknown mechanism. Caspase 1 also promotes repair and healing responses by inducing lipid membrane biogenesis through the activation of sterol regulatory element binding proteins (SREBPs) and through the secretion or release of growth factors such as fibroblast growth factor 2. The latter contributes to repair through ligation of cell surface receptors on target cells. Caspase 1 cleaves poly(ADP-ribose) polymerase 1 (PARP1) and glycolysis enzymes, possibly to prepare host cells to undergo pyroptosis. This specialized cell death programme removes intracellular niches for microbial replication and eliminates infected immune cells. Moreover, it might help to tune immune responses by releasing microbial components into the extracellular milieu, where they can be detected by the immune system. It is probable that caspase 1 cleaves additional as yet unidentified substrates that are responsible for early membrane permeabilization and oligonucleosomal DNA fragmentation during pyroptosis. Finally, inflammasome-mediated activation of caspase 7 (an effector caspase) restricts bacterial replication in Legionella-infected macrophages by targeting the infectious agent to lysosomes. CARD, caspase recruitment domain. 218 | MARcH 2011 | voLuMe 11

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS whether inflammasome‑mediated activation of caspase 7 also restricts the replication of S. Typhimurium and other bacterial pathogens, and whether this inflammasome effector pathway is activated during viral infection.

Conclusions and perspectives It has become evident in recent years that inflamma‑ somes have important roles in innate immune signalling and host defence. In particular, our knowledge of how inflammasome complexes of distinct composition are assembled in a stimulus‑dependent manner has grown significantly. Now, the different effector mechanisms (in addition to IL‑1β and IL‑18 secretion) by which inflam‑ masomes might contribute to immunity and host defence are also starting to emerge. As described above, recent studies have highlighted a range of new inflammasome functions and effector mechanisms (FIG. 3). caspase 1 has been shown to control the secretion of leaderless cytokines and proteins such as IL‑1α and FGF2, as well as the release of endogenous DAMPs such as HMGB1. Moreover, excessive caspase 1 activation in damaged neu‑ rons and infected myeloid cells induces pyroptotic cell death. Furthermore, caspase 1 dampens the metabolic rate of infected cells by cleaving key enzymes of the glyco‑ lysis pathway and regulates lipid metabolic pathways for cell repair. Finally, activation of the executioner caspase 7 downstream of inflammasomes contributes to restriction of Legionella replication in infected macrophages. Interestingly, most of these emerging functions of caspase 1 seem to operate independently of the canoni‑ cal substrates IL‑1β and IL‑18, and this indicates that inflammasomes contribute to innate immune responses in a variety of ways. Indeed, these effector mechanisms probably function together to mount a fast and effec‑ tive innate immune response against the pathogen and

Villani, A. C. et al. Common variants in the NLRP3 region contribute to Crohn’s disease susceptibility. Nature Genet. 41, 71–76 (2009). 2. Zaki, M. H. et al. The NLRP3 inflammasome protects against loss of epithelial integrity and mortality during experimental colitis. Immunity 32, 379–391 (2010). 3. Allen, I. C. et al. The NLRP3 inflammasome functions as a negative regulator of tumorigenesis during colitisassociated cancer. J. Exp. Med. 207, 1045–1056 (2010). 4. Bauer, C. et al. Colitis induced in mice with dextran sulfate sodium (DSS) is mediated by the NLRP3 inflammasome. Gut 59, 1192–1199 (2010). 5. Dupaul-Chicoine, J. et al. Control of intestinal homeostasis, colitis, and colitis-associated colorectal cancer by the inflammatory caspases. Immunity 32, 367–378 (2010). 6. Jin, Y. et al. NALP1 in vitiligo-associated multiple autoimmune disease. N. Engl. J. Med. 356, 1216–1225 (2007). 7. Martinon, F., Petrilli, V., Mayor, A., Tardivel, A. & Tschopp, J. Gout-associated uric acid crystals activate the NALP3 inflammasome. Nature 440, 237–241 (2006). 8. Magitta, N. F. et al. A coding polymorphism in NALP1 confers risk for autoimmune Addison’s disease and type 1 diabetes. Genes Immun. 10, 120–124 (2009). 9. Larsen, C. M. et al. Interleukin-1-receptor antagonist in type 2 diabetes mellitus. N. Engl. J. Med. 356, 1517–1526 (2007). 10. Agostini, L. et al. NALP3 forms an IL-1β-processing inflammasome with increased activity in Muckle–Wells autoinflammatory disorder. Immunity 20, 319–325 (2004). 1.

11.

12. 13. 14.

15.

16.

17.

18.

19.

20.

might instruct the adaptive immune system regarding the systemic risk posed by the invading microbial agent. At an early phase of infection, activation of caspase 7 downstream of the inflammasome might halt pathogen replication in intracellular replication niches, while at the same time inflammasome‑mediated activation of SReBPs could repair pathogen‑induced damage to the plasma membrane. when bacterial (or viral) loads fur‑ ther increase, inflammasomes can induce pyroptosis and instruct macrophages and dendritic cells to initi‑ ate the unconventional secretion or passive release of pro‑inflammatory cytokines, growth factors, DAMPs and microbial antigens to alert the immune system of an imminent threat and to initiate repair responses. In this context, the crucial role of pyroptosis in host defence was recently demonstrated in mice infected with the Gram‑negative bacterial pathogens L. pneumophila and B. thailandensis 34. Pyroptosis also confers resistance against Gram‑positive pathogens such as B. anthracis in vivo65, but its role in clearing viral and fungal infections remains to be determined. In addi‑ tion, much remains to be learned regarding exactly how caspase 1 induces pyroptosis and how inflammasomes regulate the unconventional secretion of leaderless pro‑ teins. what are the crucial components of the molecular machinery driving the secretion of leaderless proteins and how are they regulated by inflammasomes? Are unconventional protein secretion and pyroptosis intrin‑ sically linked or can they be uncoupled? what is the relative contribution of each of these inflammasome effector mechanisms to innate and adaptive immune responses during microbial infections? Answering these and other questions will undoubtedly illuminate intriguing new mechanisms by which inflammasomes contribute to host defence and immunity.

Hoffman, H. M., Mueller, J. L., Broide, D. H., Wanderer, A. A. & Kolodner, R. D. Mutation of a new gene encoding a putative pyrin-like protein causes familial cold autoinflammatory syndrome and Muckle–Wells syndrome. Nature Genet. 29, 301–305 (2001). Lamkanfi, M. & Dixit, V. M. The inflammasomes. PLoS Pathog. 5, e1000510 (2009). Kawai, T. & Akira, S. TLR signaling. Cell Death Differ. 13, 816–825 (2006). Tschopp, J. & Schroder, K. NLRP3 inflammasome activation: the convergence of multiple signalling pathways on ROS production? Nature Rev. Immunol. 10, 210–215 (2010). Lamkanfi, M. & Dixit, V. M. Inflammasomes: guardians of cytosolic sanctity. Immunol. Rev. 227, 95–105 (2009). Gu, Y. et al. Activation of interferon-γ inducing factor mediated by interleukin-1β converting enzyme. Science 275, 206–209 (1997). Kuida, K. et al. Altered cytokine export and apoptosis in mice deficient in interleukin-1β converting enzyme. Science 267, 2000–2003 (1995). Ghayur, T. et al. Caspase-1 processes IFN-γ-inducing factor and regulates LPS-induced IFN-γ production. Nature 386, 619–623 (1997). Li, P. et al. Mice deficient in IL-1β-converting enzyme are defective in production of mature IL-1β and resistant to endotoxic shock. Cell 80, 401–411 (1995). Joosten, L. A. et al. Inflammatory arthritis in caspase 1 gene-deficient mice: contribution of proteinase 3 to caspase 1-independent production of bioactive interleukin-1β. Arthritis Rheum. 60, 3651–3662 (2009).

NATuRe RevIewS | Immunology

21. Irmler, M. et al. Granzyme A is an interleukin 1βconverting enzyme. J. Exp. Med. 181, 1917–1922 (1995). 22. Maelfait, J. et al. Stimulation of Toll-like receptor 3 and 4 induces interleukin-1β maturation by caspase-8. J. Exp. Med. 205, 1967–1973 (2008). 23. Mayer-Barber, K. D. et al. Caspase-1 independent IL-1β production is critical for host resistance to Mycobacterium tuberculosis and does not require TLR signaling in vivo. J. Immunol. 184, 3326–3330 (2010). 24. Kanneganti, T. D. Central roles of NLRs and inflammasomes in viral infection. Nature Rev. Immunol. 10, 688–698 (2010). 25. Sims, J. E. & Smith, D. E. The IL-1 family: regulators of immunity. Nature Rev. Immunol. 10, 89–102 (2010). 26. Dinarello, C. A. Immunological and inflammatory functions of the interleukin-1 family. Annu. Rev. Immunol. 27, 519–550 (2009). 27. Hoshino, T. et al. Cutting edge: IL-18-transgenic mice: in vivo evidence of a broad role for IL-18 in modulating immune function. J. Immunol. 166, 7014–7018 (2001). 28. Nakanishi, K., Yoshimoto, T., Tsutsui, H. & Okamura, H. Interleukin-18 regulates both Th1 and Th2 responses. Annu. Rev. Immunol. 19, 423–474 (2001). 29. 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). 30. Harrington, L. E. et al. Interleukin 17-producing CD4+ effector T cells develop via a lineage distinct from the T helper type 1 and 2 lineages. Nature Immunol. 6, 1123–1132 (2005).

voLuMe 11 | MARcH 2011 | 219 © 2011 Macmillan Publishers Limited. All rights reserved

REVIEWS 31. Hoffman, H. M. et al. Efficacy and safety of rilonacept (interleukin-1 trap) in patients with cryopyrinassociated periodic syndromes: results from two sequential placebo-controlled studies. Arthritis Rheum. 58, 2443–2452 (2008). 32. Lachmann, H. J. et al. Use of canakinumab in the cryopyrin-associated periodic syndrome. N. Engl. J. Med. 360, 2416–2425 (2009). 33. Lamkanfi, M. et al. Inflammasome-dependent release of the alarmin HMGB1 in endotoxemia. J. Immunol. 185, 4385–4392 (2010). This paper describes the role of the NLRP3 and NLRC4 inflammasomes in extracellular release of HMGB1. 34. Miao, E. A. et al. Caspase-1-induced pyroptosis is an innate immune effector mechanism against intracellular bacteria. Nature Immunol. 11, 1136–1142 (2010). This paper describes the crucial role of caspase 1‑mediated cell death as a host defence mechanism against bacterial pathogens. 35. Henry, T. & Monack, D. M. Activation of the inflammasome upon Francisella tularensis infection: interplay of innate immune pathways and virulence factors. Cell. Microbiol. 9, 2543–2551 (2007). This paper describes the increased susceptibility of caspase 1‑deficient mice to infection with Francisella tularensis compared with mice lacking IL‑1β and IL‑18. 36. Trombetta, E. S. & Parodi, A. J. Quality control and protein folding in the secretory pathway. Annu. Rev. Cell. Dev. Biol. 19, 649–676 (2003). 37. Lee, M. C., Miller, E. A., Goldberg, J., Orci, L. & Schekman, R. Bi-directional protein transport between the ER and Golgi. Annu. Rev. Cell. Dev. Biol. 20, 87–123 (2004). 38. Nickel, W. & Rabouille, C. Mechanisms of regulated unconventional protein secretion. Nature Rev. Mol. Cell Biol. 10, 148–155 (2009). 39. Rubartelli, A., Cozzolino, F., Talio, M. & Sitia, R. A novel secretory pathway for interleukin-1β, a protein lacking a signal sequence. EMBO J. 9, 1503–1510 (1990). 40. Sutterwala, F. S. et al. Critical role for NALP3/CIAS1/ Cryopyrin in innate and adaptive immunity through its regulation of caspase-1. Immunity 24, 317–327 (2006). 41. Keller, M., Ruegg, A., Werner, S. & Beer, H. D. Active caspase-1 is a regulator of unconventional protein secretion. Cell 132, 818–831 (2008). This paper characterizes the role of caspase 1 in the secretion of IL‑1α and FGF2. 42. Becker, C. E., Creagh, E. M. & O’Neill, L. A. Rab39a binds caspase-1 and is required for caspase-1-dependent interleukin-1β secretion. J. Biol. Chem. 284, 34531–34537 (2009). 43. Salvesen, G. S. & Riedl, S. J. Caspase mechanisms. Adv. Exp. Med. Biol. 615, 13–23 (2008). 44. Riedl, S. J. & Shi, Y. Molecular mechanisms of caspase regulation during apoptosis. Nature Rev. Mol. Cell Biol. 5, 897–907 (2004). 45. Lamkanfi, M. & Dixit, V. M. Manipulation of host cell death pathways during microbial infections. Cell Host Microbe 8, 44–54 (2010). 46. Strasser, A., O’Connor, L. & Dixit, V. M. Apoptosis signaling. Annu. Rev. Biochem. 69, 217–245 (2000). 47. Taylor, R. C., Cullen, S. P. & Martin, S. J. Apoptosis: controlled demolition at the cellular level. Nature Rev. Mol. Cell Biol. 9, 231–241 (2008). 48. Kazama, H. et al. Induction of immunological tolerance by apoptotic cells requires caspasedependent oxidation of high-mobility group box-1 protein. Immunity 29, 21–32 (2008). 49. Savill, J., Dransfield, I., Gregory, C. & Haslett, C. A blast from the past: clearance of apoptotic cells regulates immune responses. Nature Rev. Immunol. 2, 965–975 (2002). 50. Mariathasan, S. et al. Differential activation of the inflammasome by caspase-1 adaptors ASC and Ipaf. Nature 430, 213–218 (2004). 51. Lamkanfi, M. et al. Targeted peptidecentric proteomics reveals caspase-7 as a substrate of the caspase-1 inflammasomes. Mol. Cell. Proteomics 7, 2350–2363 (2008). This paper describes caspase 7 as a downstream effector of the NLRP3 and NLRC4 inflammasomes. 52. Fink, S. L. & Cookson, B. T. Caspase-1-dependent pore formation during pyroptosis leads to osmotic lysis of infected host macrophages. Cell. Microbiol. 8, 1812–1825 (2006).

53. Lamkanfi, M. et al. Glyburide inhibits the Cryopyrin/ Nalp3 inflammasome. J. Cell Biol. 187, 61–70 (2009). 54. Chen, Y., Smith, M. R., Thirumalai, K. & Zychlinsky, A. A bacterial invasin induces macrophage apoptosis by binding directly to ICE. EMBO J. 15, 3853–3860 (1996). 55. Hilbi, H. et al. Shigella-induced apoptosis is dependent on caspase-1 which binds to IpaB. J. Biol. Chem. 273, 32895–32900 (1998). 56. Monack, D. M., Detweiler, C. S. & Falkow, S. Salmonella pathogenicity island 2-dependent macrophage death is mediated in part by the host cysteine protease caspase-1. Cell. Microbiol. 3, 825–837 (2001). This paper reports that pyroptosis proceeds independently of IL‑1β and IL‑18. 57. van der Velden, A. W., Velasquez, M. & Starnbach, M. N. Salmonella rapidly kill dendritic cells via a caspase-1-dependent mechanism. J. Immunol. 171, 6742–6749 (2003). 58. Suzuki, T. et al. Differential regulation of caspase-1 activation, pyroptosis, and autophagy via Ipaf and ASC in Shigella-infected macrophages. PLoS Pathog. 3, e111 (2007). 59. Miao, E. A. et al. Cytoplasmic flagellin activates caspase-1 and secretion of interleukin 1β via Ipaf. Nature Immunol. 7, 569–575 (2006). 60. Franchi, L. et al. Cytosolic flagellin requires Ipaf for activation of caspase-1 and interleukin 1β in salmonella-infected macrophages. Nature Immunol. 7, 576–582 (2006). 61. Amer, A. et al. Regulation of Legionella phagosome maturation and infection through flagellin and host Ipaf. J. Biol. Chem. 281, 35217–35223 (2006). 62. Miao, E. A. et al. Innate immune detection of the type III secretion apparatus through the NLRC4 inflammasome. Proc. Natl Acad. Sci. USA 107, 3076–3080 (2010). 63. Sutterwala, F. S. et al. Immune recognition of Pseudomonas aeruginosa mediated by the IPAF/ NLRC4 inflammasome. J. Exp. Med. 204, 3235–3245 (2007). 64. Boyden, E. D. & Dietrich, W. F. Nalp1b controls mouse macrophage susceptibility to anthrax lethal toxin. Nature Genet. 38, 240–244 (2006). 65. Terra, J. K. et al. Cutting edge: resistance to Bacillus anthracis infection mediated by a lethal toxin sensitive allele of Nalp1b/Nlrp1b. J. Immunol. 184, 17–20 (2010). 66. Munoz-Planillo, R., Franchi, L., Miller, L. S. & Nunez, G. A critical role for hemolysins and bacterial lipoproteins in Staphylococcus aureus-induced activation of the Nlrp3 inflammasome. J. Immunol. 183, 3942–3948 (2009). 67. Mariathasan, S. et al. Cryopyrin activates the inflammasome in response to toxins and ATP. Nature 440, 228–232 (2006). 68. Sauer, J. D. et al. Listeria monocytogenes triggers AIM2-mediated pyroptosis upon infrequent bacteriolysis in the macrophage cytosol. Cell Host Microbe 7, 412–419 (2010). 69. Fernandes-Alnemri, T. et al. The AIM2 inflammasome is critical for innate immunity to Francisella tularensis. Nature Immunol. 11, 385–393 (2010). 70. Rathinam, V. A. et al. The AIM2 inflammasome is essential for host defense against cytosolic bacteria and DNA viruses. Nature Immunol. 11, 395–402 (2010). 71. Jones, J. W. et al. Absent in melanoma 2 is required for innate immune recognition of Francisella tularensis. Proc. Natl Acad. Sci. USA 107, 9771–9776 (2010). 72. Wu, J., Fernandes-Alnemri, T. & Alnemri, E. S. Involvement of the AIM2, NLRC4, and NLRP3 inflammasomes in caspase-1 activation by Listeria monocytogenes. J. Clin. Immunol. 30, 693–702 (2010). 73. Warren, S. E. et al. Cutting edge: cytosolic bacterial DNA activates the inflammasome via Aim2. J. Immunol. 185, 818–821 (2010). 74. Tsuchiya, K. et al. Involvement of absent in melanoma 2 in inflammasome activation in macrophages infected with Listeria monocytogenes. J. Immunol. 185, 1186–1195 (2010). 75. Mariathasan, S., Weiss, D. S., Dixit, V. M. & Monack, D. M. Innate immunity against Francisella tularensis is dependent on the ASC/caspase-1 axis. J. Exp. Med. 202, 1043–1049 (2005). 76. Monack, D. M., Raupach, B., Hromockyj, A. E. & Falkow, S. Salmonella typhimurium invasion induces

220 | MARcH 2011 | voLuMe 11

77.

78.

79.

80.

81.

82. 83.

84.

85.

86.

87.

88.

89.

90.

91. 92.

93.

apoptosis in infected macrophages. Proc. Natl Acad. Sci. USA 93, 9833–9838 (1996). Fernandes-Alnemri, T. et al. The pyroptosome: a supramolecular assembly of ASC dimers mediating inflammatory cell death via caspase-1 activation. Cell Death Differ. 14, 1590–1604 (2007). Malireddi, R. K., Ippagunta, S., Lamkanfi, M. & Kanneganti, T. D. Cutting edge: proteolytic inactivation of poly(ADP-ribose) polymerase 1 by the Nlrp3 and Nlrc4 inflammasomes. J. Immunol. 185, 3127–3130 (2010). This paper describes the inflammasome‑dependent cleavage of PARP1 during pyroptosis. Akhter, A. et al. Caspase-7 activation by the Nlrc4/Ipaf inflammasome restricts Legionella pneumophila infection. PLoS Pathog. 5, e1000361 (2009). This paper describes the role of caspase 7 activation by the NLRC4 inflammasome in restricting Legionella replication. Lakhani, S. A. et al. Caspases 3 and 7: key mediators of mitochondrial events of apoptosis. Science 311, 847–851 (2006). Shao, W., Yeretssian, G., Doiron, K., Hussain, S. N. & Saleh, M. The caspase-1 digestome identifies the glycolysis pathway as a target during infection and septic shock. J. Biol. Chem. 282, 36321–36329 (2007). This paper describes glycolysis enzymes as substrates of caspase 1. Cramer, T. et al. HIF-1α is essential for myeloid cellmediated inflammation. Cell 112, 645–657 (2003). Gurcel, L., Abrami, L., Girardin, S., Tschopp, J. & van der Goot, F. G. Caspase-1 activation of lipid metabolic pathways in response to bacterial poreforming toxins promotes cell survival. Cell 126, 1135–1145 (2006). This paper reports that inflammasomes activate lipid metabolic pathways. Gonzalez, M. R., Bischofberger, M., Pernot, L., van der Goot, F. G. & Freche, B. Bacterial pore-forming toxins: the (w)hole story? Cell. Mol. Life Sci. 65, 493–507 (2008). McCoy, A. J. et al. Cytotoxins of the human pathogen Aeromonas hydrophila trigger, via the NLRP3 inflammasome, caspase-1 activation in macrophages. Eur. J. Immunol. 40, 2797–2803 (2010). Lamkanfi, M. et al. The Nod-like receptor family member Naip5/Birc1e restricts Legionella pneumophila growth independently of caspase-1 activation. J. Immunol. 178, 8022–8027 (2007). Zamboni, D. S. et al. The Birc1e cytosolic patternrecognition receptor contributes to the detection and control of Legionella pneumophila infection. Nature Immunol. 7, 318–325 (2006). Broz, P. et al. Redundant roles for inflammasome receptors NLRP3 and NLRC4 in host defense against Salmonella. J. Exp. Med. 207, 1745–1755 (2010). Fischer, U., Janicke, R. U. & Schulze-Osthoff, K. Many cuts to ruin: a comprehensive update of caspase substrates. Cell Death Differ. 10, 76–100 (2003). Lamkanfi, M., Festjens, N., Declercq, W., Vanden Berghe, T. & Vandenabeele, P. Caspases in cell survival, proliferation and differentiation. Cell Death Differ. 14, 44–55 (2007). Boatright, K. M. et al. A unified model for apical caspase activation. Mol. Cell 11, 529–541 (2003). Martinon, F. & Tschopp, J. Inflammatory caspases and inflammasomes: master switches of inflammation. Cell Death Differ. 14, 10–22 (2007). Pizzirani, C. et al. Stimulation of P2 receptors causes release of IL-1β-loaded microvesicles from human dendritic cells. Blood 109, 3856–3864 (2007).

Acknowledgements This work was supported by European Union Framework Program 7 (Marie Curie grant 256432) and by the Fund for Scientific Research – Flanders.

Competing interests statement The author declares no competing financial interests.

FURTHER INFORMATION Mohamed Lamkanfi’s homepage: http://www.vib.be/en/ research/scientists/Pages/Mohamed-Lamkanfi-Lab.aspx All lInks Are ACtIve In the onlIne pdf

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

PErsPEcTIvEs E S S AY

Tissue-based class control: the other side of tolerance Polly Matzinger and Tirumalai Kamala

Abstract | In this Essay, we offer a new perspective on how immune responses are regulated. We do not cover how they are turned on and off, but focus instead on the second major aspect of an immune response: the control of effector class. Although it is generally thought that the class of an immune response is tailored to fit the invading pathogen, we suggest here that it is primarily tailored to fit the tissue in which the response occurs. To this end, we cover such topics as the nature of T helper (TH) cell subsets (current and yet to be discovered), the nature of privileged sites, the difference between oral tolerance and oral vaccination, why the route of immunization matters, whether the TH1‑type response is really the immune system’s primary defense, and whether there might be a different role for some regulatory T cells. When confronted with a potential threat, the immune system faces two decisions: first, whether to respond or not, and second, what kind of response to make. The first decision has fascinated immunologists for decades, sparking several theories and much experimental work, most of which rests on the assumption that the immune system responds to ‘foreign’ antigens, and that maintaining self tolerance is a matter of controlling autoreactive T and B cells. But dealing with autoreactive lymphocytes is only half of the problem. Even in the absence of any autoreactivity, the wrong immune effector class can completely destroy a tissue1–5. The control of effector class, however, has had little theoretical input. Students generally learn that the immune system matches the effector class to the pathogen that it is fighting (for example, making IgE against worms, and cytotoxic T lymphocytes (CTLs) against viruses and intracellular bacteria). But it is not easy to see how the immune system could discriminate between worms, viruses or intracellular versus extracellular bacteria, as T cell receptors bind peptide– MHC complexes; B cell receptors bind small epitopes on proteins, carbohydrates and lipids that are present in most living

organisms; and the ‘innate’ receptors, such as the Toll-like receptors (TLRs) and NODlike receptors (NLRs), are so promiscuous that they don’t distinguish between ligands from different phyla, or between pathogenderived and self-derived alarm signals6–9. Although current data suggest that TLR5 and NOD2 (nucleotide-binding oligomerization domain protein 2) may be fairly specific, it would be difficult to use even these receptors to make decisions about effector class. NOD2 does not distinguish between intracellular and extracellular bacteria, as it binds muramyl dipeptide, a component of almost all bacterial cell walls10. Similarly, TLR5 binds the flagella of both intracellular pathogens (such as Listeria monocytogenes) and extracellular pathogens (such as Escherichia coli and Pseudomonas aeruginosa)11, whereas it does not bind the flagella of several other important pathogens (including Helicobacter pylori, Staphylococcus aureus and Campylobacter jejuni)11. So, what controls the effector class of an immune response? The idea that it might be the tissues, rather than the immune system, has grown slowly over the 13 years that we have been studying immunity from the perspective of the danger model12,13. Initially,

NATuRE REvIEWS | Immunology

the model did not offer any clues as to how one effector class might be chosen over another, as it was designed to cover only the immune system’s first decision (whether to respond or not). It proposed that perturbed tissues initiate immune responses by sending alarm signals that activate local antigen-presenting cells (APCs), whereas healthy tissues display their own antigens or allow ‘resting’ APCs to display those antigens to induce peripheral tolerance. In effect, this model suggested that turning immune responses on or off was the prerogative of the tissues. It takes only a small step to suggest that tissues may also control the effector class, such that the class of an immune response is tailored to the tissue in which it occurs, rather than to the invading pathogen. The basics of this idea were outlined in two earlier articles12,14. In this Essay, we describe the idea more fully, suggesting mechanisms by which tissues could carry out this function, describing some wellknown immunological phenomena in light of this view, and pointing out the possibility that a complete definition of the immune system should perhaps include every tissue in the body. Class and T helper cell subsets Let us start by defining what we mean by immune effector class. Although the term “class” has historically been used to define different antibody isotypes (such as IgM and IgG), we prefer a definition that also includes the participating cells. Thus, each effector class combines a particular set of helper cells and the antibodies and effector cells that they promote. Currently, three main subclasses are generally accepted. TH1-type responses consist of T cells that produce interleukin-2 (IL-2), interferon-γ (IFNγ) and tumour necrosis factor (TNF), as well as B cells that make complement-fixing IgG antibodies, CTLs, activated natural killer (NK) cells and macrophages that produce free oxygen radicals. TH2-type responses consist of T cells that produce IL-4, IL-5, IL-13 and IL-10, B cells making IgE and IgG1, macrophages that express arginase, and the influx of eosinophils. TH17-type responses consist of T cells that produce IL-17 and the influx of neutrophils. vOLuME 11 | MARCH 2011 | 221

© 2011 Macmillan Publishers Limited. All rights reserved

PersPectives However, there have long been clues that this is an oversimplification. Human T cells, for example, often show non-classical cytokine expression patterns, and it took a long time to persuade researchers working with human cells to adopt the TH1- and TH2-type classifications. In mice, Kelso15 found that micromanipulated single T cells do not stably make TH1- or TH2-type cytokine ‘cassettes’. She suggested that the TH1- and TH2-type patterns are only the extremes of a multidimensional continuum; that individual T cells normally make only a small number of stochastically produced cytokines; and that populations of T cells can diversify to produce any number of different cytokine combinations. More recently, O’Shea and Paul16 proposed a somewhat similar scenario, but these views are far from universal. For example, Pulendran’s and Oppenheim’s groups found that dendritic cells (DCs) stimulated by Porphyromonas gingivalisderived lipopolysaccharide (LPS)17 or by eosinophil-derived neurotoxin18 induce TH cells that secrete IL-5 but not IL-4 (the signature TH2-type cytokine). In addition, Prussin and colleagues19 found two different subsets of TH cells in atopic patients, some of which produce IL-5 but not IL-4. Nevertheless, rather than postulate the existence of a TH cell that did not fit into the standard categories, all of these authors called these IL-5-producing cells “TH2 cells”. We believe it is time to follow Kelso’s lead and stop forcing various kinds of immune responses into a few common categories. Although the TH1/TH2 paradigm has been useful in establishing the concept that different sorts of TH cells promote different classes of

response, it has limited our ability to recognize the potentially enormous diversity of immune responses. If we were to stop consigning TH cells to a small group of numbered subsets, but instead name them by what they produce (as is done for TH17 cells) or by the responses they promote (as is done for follicular helper T cells), we would uncover the possibility that there are a large number of differentiation paths that TH cells can take. We would suggest that each particular effector cell (such as each type of B cell, CTL, NK cell, macrophage, eosinophil, neutrophil and basophil) is controlled by a particular set of secreted and membranebound signals (from TH cells and from other sources) and can be combined with any other effector cell to make a wide variety of carefully tailored immune responses. Given the existence of such a variety of effectors, and the TH cells that facilitate them, what determines the ultimate effector combination that arises in any particular immune response? We were drawn to the possibility that this is the responsibility of the tissues by two old immunological phenomena: immune-privileged sites and oral tolerance. Immune-privileged sites Immune-privileged sites are organs in which allogeneic transplants are not rejected. Neonatal hearts, for example, are rejected when transplanted under the skin or kidney capsules of adult recipients20, but survive indefinitely if placed into the anterior chamber of the eye21, the brain21, the testes22 or the hamster cheek pouch23. These observations are generally interpreted as evidence that ‘privileged’ sites exclude24, disable25 or suppress26 immune cells.

Table 1 | Cytokines that tailor immune effector class in the eye and gut Cytokine

DTH or TH1-type response

TReg cell induction

IgA production

TGFβ

↓140

↑141

↑29,30

vIP

↓142

↑143

↑31,32

αMsH

↓

↑

?

TGFβ

↓140

↑141

↑29,30

vIP

↓

143

↑

↑31,32

TsLP

↓

↑146

↑147‡

vitamin A (retinoic acid)

↓148,149

↑119,150,151

↑57

In the eye

144

145

In the gut* 142

DTH, delayed‑type hypersensitivity; αMsH, α‑melanocyte stimulating hormone; TGFβ, transforming growth factor‑β; Treg, regulatory T; TsLP, thymic stromal lymphopoietin; vIP, vasoactive intestinal peptide. *There are other molecules secreted by gut epithelial cells (such as APrIL (a proliferation‑inducing ligand; also known as TNFsF13) and BAFF (B cell‑activating factor; also known as TNFsF13B) that are not as well studied but which are likely also to influence the class of the immune response. ‡TsLP promotes IgA indirectly by promoting APrIL and interleukin‑10 production.

222 | MARCH 2011 | vOLuME 11

There is, however, a world of difference between data and interpretations. The interpretation that immunity cannot occur in privileged sites makes little evolutionary sense. These tissues are wet, warm and full of nutrients. Without protection by the immune system, wouldn’t they promptly be exploited by pathogens? Luckily, other interpretations exist. Streilein’s work on the eye reveals a much more interesting picture. The eye is a complex tissue, containing delicate specialized cells that cannot survive a full-blown TH1-type or delayed-type hypersensitivity (DTH) response. It protects itself by a process Streilein called “anterior chamber-associated immune deviation”27, in which cells lining the anterior chamber secrete cytokines — transforming growth factor-β (TGFβ), vasoactive intestinal peptide (vIP) and α-melanocyte stimulating hormone (αMSH) — that suppress TH1-type and DTH responses and increase the activity of regulatory T (TReg) cells28 (Table 1). Although this appears to be immune suppression, a closer look suggests something different. TGFβ, vIP and αMSH all promote IgA production29–32, but IgA cannot reject a transplant. So if we measure ocular immunity only by transplant rejection, and ignore the perfectly functional IgA response, we call it “tolerance”, “deviation”, “suppression” or “regulation”. But this is none of those things. It is simply a class of response that protects the eye without destroying it. The cells of the eye also express FAS ligand (also known as CD95L), which can trigger T cells through surface FAS (also known as CD95) to die by apoptosis25. Although initially interpreted as evidence that lymphocytes entering the eye are eliminated, newer evidence that TH1 cells express more FAS than TH2 cells33 suggests that the eye ‘chooses’ what kinds of TH cells it allows or excludes. We predict that the other privileged sites also exert similar control over the local class of immune response. Oral tolerance and oral vaccination Oral tolerance has been extensively studied in experimental autoimmune encephalomyelitis. This is an experimental system in which an animal that is immunized with a strong adjuvant plus a brain-derived protein, such as myelin basic protein (MBP), acquires an autoimmune disease mediated by TH1-type, TH17-type and DTH responses34 that somewhat resembles multiple sclerosis in humans. Feeding the animal MBP prior to MBP immunization prevents disease and reduces T cell responses35,36. The T cells no longer proliferate in response to MBP or make IFNγ www.nature.com/reviews/immunol

© 2011 Macmillan Publishers Limited. All rights reserved

PersPectives or TNF and, when transferred into another mouse, these T cells suppress the recipient’s autoimmune response. The interpretation of these results, that orally administered antigen generates tolerance and TReg cells, has spurred a novel treatment for allergy known as sublingual immunotherapy 37, in which patients apply small amounts of allergen sublingually each day. But what about oral vaccination, which can elicit protective immunity against poliovirus in humans38 and against rabies virus in raccoons and coyotes39? What is the difference between oral vaccination and oral tolerance? In many cases, very little. Oral administration of antigen in mice elicits at least three different kinds of response. When given in large doses, it can induce systemic deletional tolerance40,41 (presumably because enough antigen leaks into the circulation to be presented by resting DCs), as well as a local IgA response42. The addition of various adjuvants converts this into a systemic response that includes CTLs and IgG43. Lower doses of antigen can lead to the activation of TH3 cells that produce IL-4, IL-5, IL-10 and TGFβ and promote the production of IgA42,44. Because the TH3 cells can suppress inflammatory mechanisms through their secretion of IL-10 and TGFβ, and because the presence of IgA is rarely assessed, this type of response is also often labelled as tolerance. But this is not tolerance. It is simply a switch to an immune response that is appropriate for the intestinal environment. How does the gut promote the IgA response? Intestinal epithelial cells express TLRs45 and secrete cytokines46 (Table 1). Some of these cytokines (including TGFβ and vIP) are similar to those produced in the eye, others (such as thymic stromal lymphopoietin (TSLP)) are unique to the gut, and some factors (such as vitamins A and D) are acquired in the diet and modified for use. All of these factors shape the immune response (Table 1). For example, TGFβ, TSLP, vitamin A and vitamin D have been shown to suppress TH1-type and DTH responses and promote the production of IgA. Thus, the gut seems to promote the antibody subclass (IgA) that is locally most useful, while simultaneously preventing destructive TH1-type and DTH responses. Tissue-appropriate immunity Why would a tissue suppress DTH and TH1-type responses? Why not make many different effector classes to ensure pathogen clearance? The reason is that these mechanisms are terribly destructive. TNF and IFNγ induce cell death47,48; IL-17 recruits

Box 1 | Class control by organs, tissues or regions? Is it organs or tissues that control the immune response? On the one hand, one could argue that the function of each organ dictates that it should encourage particular types of response and discourage others. But are organs homogeneous in their needs? For example, the skin has a barrier function that may require certain types of immune functions, but the dermis and epidermis are not the same. They have different populations of antigen-presenting cells (APCs) and might promote different immune response classes. The jejunum and ileum, although comprising parts of the small intestine, have their own subsets of microflora and may need somewhat different types of immune protection. Even within these intestinal regions, the epithelial microenvironments are different. In the villus crypts, where most epithelial cell division occurs and where bacterial infection would be particularly hazardous, Paneth cells produce large amounts of antimicrobial peptides. Further up the villus, the epithelial cells take on more of their own protection, expressing Toll-like receptors and producing different types of bactericidal molecules. One could argue, then, that each cell type has the ability to produce immune protective and immune-modulating signals, and that this implies that control of the immune response lies at the level of the cell. However, the same cell type might behave differently in different organs. For example, the vascular endothelium is unlikely to be the same in the lungs, liver, heart, skin and kidneys. Will it communicate differently in those different sites with the cells of the immune system? At the moment we don’t have answers to these questions. We don’t know what comprises a minimum tissue ‘unit’ that communicates with the immune system. For that reason, we use the word “tissue” to define a local mixture of tissue cells that communicate with each other and with the bone marrow-derived cells that constitute the rest of the immune system. In some cases, a local tissue might also communicate systemically with other tissues to help define the initiation, longevity and effector class of an immune response.

neutrophils; CTLs and NK cells kill target cells directly or through antibody-dependent cell-mediated cytotoxicity 49,50; macrophages release oxygen radicals51; and complement drills holes in cell membranes52. These are devastating weapons! The eye and gut are not alone in being susceptible to damage by these powerful responses. Each organ is made of an intricate combination of tissues (bOX 1), precisely tuned to perform particular functions that can easily be compromised by destructive effector mechanisms. For example, strong TH1-type responses can destroy the placenta1,53, pancreatic islets3, skin1, eye4, brain54 and small intestine2. So most tissues are likely to have mechanisms to avoid such destruction while promoting appropriate local immunity. How might tissues communicate their preferences to local and circulating cells of the immune system? First, the local tissue can produce (or modify) cytokines, chemokines and other communicating molecules (for example, antimicrobial peptides such as LL37 (Ref. 55); neuroactive molecules such as vIP and noradrenaline56; or vitamins such as vitamins A57 and D58). These factors can influence tissue-resident APCs to promote a certain effector class (or classes) while discouraging others. They can also affect the entry and exit of innate and adaptive immune cells, and govern what these cells do in the local environment. For example, fluid from the eye’s anterior chamber can induce peritoneal macrophages to suppress

NATuRE REvIEWS | Immunology

DTH responses59 and enhance TH2-type responses60. In the gut, TGFβ acts as a switch factor that induces B cells to produce IgA61, and vitamin D helps to recruit TH2 cells rather than TH1 cells by promoting the production of CC-chemokine ligand 22 (CCL22)62, which recruits CC-chemokine receptor 4 (CCR4)+ TH2 cells63. In the skin, locally produced vitamin D can suppress local DTH responses64. Second, many tissues have resident T cells that respond to stress-induced self molecules rather than foreign antigens. The gut has intraepithelial αβ T cells65 and γδ T cells66, as well as mucosa-associated invariant T (MAIT) cells67. These cells respond to the stress-induced self molecules RAE1 (retinoic acid early-inducible protein 1) and MR1 (MHC class I related) in mice, and to MICA (MHC class I polypeptide-related sequence A), MICB and MR1 in humans68. Similarly, up to 40% of liver-resident T cells are NKT cells that recognize several lipid molecules presented by the stress-induced antigen-presenting glycoprotein CD1d69. Mouse and cattle skin contains dendritic epidermal γδ T cells (DETCs), which respond to RAE1 (Ref. 70) and help to heal the skin by making keratinocyte growth factor 71, IL-2 and IFNγ72. The purpose of these tissueresident cells seems to be to survey the tissues they reside in for signs of stress and to maintain the health of these tissues. The cells may achieve this by making molecules that are important for healing and/or by producing cytokines (and probably vOLuME 11 | MARCH 2011 | 223

© 2011 Macmillan Publishers Limited. All rights reserved

PersPectives chemokines, endogenous danger signals and neuropeptides) that promote the class of immune response appropriate for that tissue at that time. Third, neuronal signals may contribute to tissue-specific class control. Neuropeptides can influence effector class73, and many leukocytes express neuropeptide receptors that were previously thought to be restricted to the nervous system74–78. Overall, tissues seem to have multiple ways of communicating with the immune system and promoting appropriate local immune functions. We assume that each tissue has a particular set of effectors that it prefers. IgA, for example, is appropriate for the eye79, the gut and other mucosal surfaces, but may not be the right effector for the brain, which would have its own preferred response class. The three phases of an immune response If TH1-type, TH17-type and DTH responses are so harmful, why have they evolved? One possibility is that tissues that regenerate easily (such as the skin and liver) can tolerate the damage, whereas others (including the eye, brain and pancreatic islets) might not. Thus, a DTH response in a non-regenerative tissue might be a case of the right action in the wrong place. Another possibility is that DTH is the immunological counterpart of frostbite. On exposure to cold, capillaries dilate to keep extremities warm, but with intense cold or long exposure the capillaries constrict, causing loss of extremities while preserving core temperature. Perhaps the immune system’s response is similarly biphasic. It first uses less destructive immune mechanisms to deal with adversity in a way that maintains the health of all tissues. But if that doesn’t work, it switches to a second phase of destructive responses, sacrificing some tissues to preserve life. In fact, there could be three phases. In phase one, the tissue summons innate cells for clean up and repair. If this isn’t sufficient, it recruits the adaptive immune system, which attempts to clear the problem with a locally tailored effector class. And if that doesn’t work, the tissue brings in the highly destructive TH1-type, TH17-type and DTH responses. DTH responses might destroy the eye while clearing an ocular infection80, but they ensure that the individual survives. In the gut, DTH responses may cause temporary flattening of gut villi and result in diarrhoea, but villi regenerate quickly. And, although a TH1-type response may cause a fetus to abort while clearing an infection, it saves the mother’s life.

Nevertheless, such destruction is a tremendous price to pay for good health and should only be summoned when necessary. How does a tissue determine when to switch from one class of immune response to another? Controlling the switch When considering what might control the switch between the phases of an immune response, we begin with three assumptions. The first is that tissues ‘educate’ resident APCs (and incoming T and B cells) to promote certain types of immune response and suppress others81–86. Although it has been suggested that there are many subsets of DCs because of the need to distinguish many different pathogens16, these subsets might also correlate with the different kinds of tissues that need protection (for example, Langerhans cells can be distinguished from dermal DCs in unperturbed skin). Furthermore, DCs remember their origins. DCs that migrate from the gut to mesenteric lymph nodes induce T cells to express guthoming receptors (such as CCR9)87 and produce cytokines (including IL-4, IL-10 and TGFβ) that suppress TH1-type responses and promote IgA production41,61,81,88, whereas DCs that leave damaged skin induce T cells to express skin-homing receptors (namely α4β1 integrin, CCR4 and CCR10)89,90. We predict that a thorough analysis will show that each tissue imparts specific instructions that result in different populations of TH cells promoting different combinations of effectors, each appropriate for that particular injured or infected tissue. For example, three different TH cell subsets protect three different organs: CD62L+ T cells protect islets from diabetes, CD25+ T cells protect the stomach from gastritis, and CD45RBlow T cells protect the gut from colitis91. The instructions that tissues pass to T cells might be communicated through tissue-derived molecules, which could influence local or newly entering T cells directly or which could be transported to the T cells by APCs. Alternatively, the signals might be produced by APCs as a consequence of the education that they received from their local tissue. If antigen recognition delivers ‘signal one’ and co-stimulation delivers ‘signal two’, then tissue-influenced signals governing the ensuing effector class could be called ‘signal three’. The second assumption we make is that when signal three is absent, the default action is a TH1-type or DTH response. If a tissue becomes so damaged that it cannot deliver APC-educating signals, the backup response

224 | MARCH 2011 | vOLuME 11

is probably necessary. This fits with the common findings that T cells stimulated by DCs generated without tissue-derived signals (such as DCs generated in plastic dishes) or by CD3-specific antibody tend to produce TH0- or TH1-type responses, whereas T cells stimulated by DCs from mesenteric lymph nodes or Peyer’s patches produce IL-4, IL-10 and TGFβ88,92. Thus, tissue-derived signals serve as checkpoints to prevent the destructive default backup response. Third, we assume that APCs carrying tissue instructions to draining lymph nodes survive for a while and are then replaced by new APCs. Their experimentally deduced functional lifespans range from 3 days93 to 3 weeks94. Given these assumptions, there are at least two non-exclusive possibilities for how tissues might manage the switch from locally tailored immune responses to destructive backup responses: one based on time, the other on signal strength. Time-dependent class switch. When activated by exogenous or endogenous6,95,96 alarm signals, tissue-resident APCs migrate to the draining lymph nodes and are replaced by new APC precursors. If the immune Figure 1 | A model for tissue-based class control of immune responses. a | resting tis‑ sues educate local antigen‑presenting cells (APcs). b | Following an insult (such as an injury or infection), the APcs leave the tissue to stimulate naive T cells to make tissue‑educated responses. c | If the innate immune response clears the infec‑ tion (or injured tissue), the tissue heals and edu‑ cates newly arriving APcs. An adaptive immune response is not needed and ceases. d | If the innate immune response does not stop the infection, then tissue‑educated adaptive immune responses are initiated. If these clear the pathogen, then the tissue heals. e | If the tissue‑educated adaptive immune response cannot resolve the infection, then a second wave of newly entering APcs will be activated in a local tissue environment that now contains more extensive damage. The new APcs may be properly educated or they may not be (because the high level of damage would result in fewer signals from the tissue). If not, they will leave the tissue and stimulate the emergency backup response. f | If the backup response clears the pathogen, then the tissue heals, but with some scarring or fibrosis occurring. g | If the initial insult is severe, the local APcs leave the tissue without receiving a complete education. This could be because the severely damaged tissues cannot provide the right signals or because the tissue provides signals that override the original educa‑ tion. These APcs launch the immediate backup response. DTH, delayed‑type hypersensitivity; Tcr, T cell receptor; TH1, T helper 1. www.nature.com/reviews/immunol

© 2011 Macmillan Publishers Limited. All rights reserved

PersPectives E6KUUWGJGCNGF

+HKPUWNVKUUVQRRGFD[ KPPCVGKOOWPGTGURQPUG

'FWECVKPI UKIPCN

6JGVKUUWGKUJGCNGFD[GKVJGT KPPCVGQTCFCRVKXGKOOWPG TGURQPUGU0GY#2%UTGRNCEG VJGQPGUVJCVNGȎCPFDGEQOG GFWECVGF

D(KTUVFGEKUKQPKPKVKCVGCVKUUWGGFWECVGFTGURQPUG

+HKPUWNVKUUVQRRGFD[ CFCRVKXGKOOWPGTGURQPUG

CPFDGIKPVJGTGRCKTRTQEGUU

+PLWTGFQTKPHGEVGF VKUUWGEGNN

F5GEQPFFGEKUKQPVKUUWGGFWECVGFCFCRVKXG

KOOWPGTGURQPUGDGIKPUVQENGCTVJGKPHGEVKQP

+PVGTOGFKCVG KPUWNV 0GWVTQRJKN #EVKXCVKPI CPF GFWECVKPI!UKIPCN #2%

6KUUWGGFWECVGF GȭGEVQT6EGNN

/CETQRJCIG 5KIPCN 'FWECVGF #2%

/KPKOCNQT KPVGTOGFKCVG KPUWNV EQOOQP

/*% 6%4

5KIPCN

C4GUVKPIVKUUWG 6KUUWGEGNN

#2%

'FWECVKPI UKIPCN

#PVKIGP

+HKPUWNV EQPVKPWGU

#PVKDQF[

0GY #2% #EVKXCVKPI UKIPCN

0CKXG 6EGNN

7PGFWECVGF #2%

&TCKPKPI N[ORJPQFG

+PEQOKPIGFWECVGFGȭGEVQT6EGNNUCPFCPVKDQFKGUWUG VKUUWGGFWECVGFTGURQPUGUVQDGIKPENGCTKPIVJGKPHGEVKQP 0GY#2%UOKITCVGKPVQVJGKPLWTGFVKUUWGCPFCTG CEVKXCVGFVQNGCXGYKVJQWVIGVVKPICEQORNGVGGFWECVKQP

+PPCVGKOOWPGEGNNUCTTKXGVQUVCTVVJGTGRCKTRTQEGUU 'FWECVGF#2%UOKITCVGVQFTCKPKPIN[ORJPQFGUFGNKXGT EQUVKOWNCVQT[UKIPCNU UKIPCNCPFGFWECVKPIUKIPCNU

UKIPCNVJCVKPFWEGPCKXG6EGNNUVQOCMGVKUUWG GFWECVGFGȭGEVQT6EGNNUUQOGQHYJKEJOKITCVGVQ KPLWTGFQTKPHGEVGFVKUUWGUQVJGTUVQ$EGNNHQNNKENGU

+HKPUWNV EQPVKPWGU

4GURQPUGUKPXQNXKPI VKUUWGGFWECVGF#2%U 4GURQPUGUKPXQNXKPI WPGFWECVGF#2%U

I+PKVKCVGVJGKOOGFKCVGDCEMWRTGURQPUG

G6JKTFFGEKUKQPKPKVKCVGVJGGOGTIGPE[DCEMWRTGURQPUG

4GUVKPIVKUUWGURTQFWEGUKIPCNU VJCVGFWECVGTGUKFGPV#2%U

#PVKIGP

#EVKXCVKPI CPF QXGTTKFKPI!UKIPCN

5GXGTG KPUWNV TCTG

5KIPCN

#PVKDQF[

#EVKXCVKPI UKIPCN

6*EGNN 5KIPCN

7PGFWECVGF #2%

6*EGNN

7PGFWECVGF #2% 0QUKIPCN

&TCKPKPI N[ORJPQFG

#2%UNGCXGVJGKPLWTGFVKUUWGYKVJQWVTGEGKXKPIC EQORNGVGGFWECVKQP QTJCXKPIJCFVJGKTGFWECVKQP QXGTTKFFGP!CPFUVKOWNCVGVJGGOGTIGPE[6*V[RGQT &6*DCEMWRTGURQPUG'ȭGEVQT6EGNNUCPFEQORNGOGPV ȮZKPICPVKDQFKGUIQVQVJGVKUUWGCPFENGCTVJGKPHGEVKQP

0QUKIPCN

&TCKPKPI N[ORJPQFG

#PGYYCXGQH#2%UNGCXGUVJGKPLWTGFVKUUWGYKVJQWVTGEGKXKPIC EQORNGVGGFWECVKQPCPFVJGUGEGNNUUVKOWNCVGVJGGOGTIGPE[6* V[RGQT&6*DCEMWRTGURQPUG'ȭGEVQT6EGNNUCPFEQORNGOGPV ȮZKPICPVKDQFKGUIQVQVJGVKUUWGCPFENGCTVJGKPHGEVKQP

+PUWNVKUUVQRRGF D[DCEMWRTGURQPUG

+PUWNVKUUVQRRGFD[ DCEMWRTGURQPUG

H6KUUWGJGCNGFYKVJUECTTKPI 5ECT 'FWECVKPI UKIPCN

+PEQOKPIDCEMWRGȭGEVQTUCPF CPVKDQFKGUJCXGENGCTGFVJG KPHGEVKQPDWVCNUQETGCVGFOQTG FCOCIG6JGVKUUWGJGCNUYKVJ ȮDTQUKUQTUECTTKPI0GY#2%UIGV GFWECVKPIUKIPCNUHTQOVJG TGOCKPKPIVKUUWGEGNNU 0CVWTG4GXKGYU^+OOWPQNQI[

NATuRE REvIEWS | Immunology

vOLuME 11 | MARCH 2011 | 225 © 2011 Macmillan Publishers Limited. All rights reserved

PersPectives response elicited by the original wave of tissue-resident APCs clears the infection, the new APC precursors that migrate to the healing or healed tissue will become the new resident tissue-educated APCs. If, however, the infection persists, this second wave of APCs may be activated before they are educated. When these uneducated APCs migrate to the draining lymphoid tissue, they will elicit the default backup TH1-type or DTH response. In this scenario, the backup mechanism occurs only because the initial tissue-oriented response has failed (fIG. 1). Damage- or stimulus-dependent class switch. under conditions of severe infection97, cell stress or destruction8, or in the presence of noxious adjuvants28, the local tissue may need to promote a strong backup response right from the start or may simply have difficulty sending educating signals. APCs migrating from the tissue without having had the proper final set of instructions will, therefore, induce the backup response. Later, when most of the infection has been cleared and the healing tissue is once again able to send educating signals, the response will revert to the tissue-oriented response98. This might account for the occasional late appearance of less noxious responses99 (fIG. 1). These are not mutually exclusive scenarios. There might be situations in which memory T cells influence APCs after they arrive at the draining lymph node and counteract the tissue-derived instructions that the APCs arrived with. For example, T cells induced by oral immunization can enter popliteal lymph nodes and re-educate DCs to promote TH2- or TH3-type responses rather than TH1-type responses83,92. There might be times when the signals that APCs receive through their innate receptors are so strong that they overcome the tissue-educating signals. There might also be positive or negative feedback loops. For example, DCs that promote TH1-type or DTH responses make IL-12, which stimulates NK cells. In turn, activated NK cells can rapidly kill activated DCs100, ensuring that a TH1-type or DTH response doesn’t last long. And when the damage gets too great, the resulting hypoxia can downregulate the destructive immune response and switch it to another class101. What about the pathogens? One could argue that the idea that tissues promote their own particular types of response, irrespective of the infecting agent, goes against a wealth of data supporting

the textbook consensus that the immune system tailors its responses to the pathogen (by initiating CTL responses to viruses and intracellular bacteria, and TH2-type responses to extracellular bacteria and worms)16,102. However, the evidence is not as clear as the textbooks would like to us to believe. First, the view that the immune system mobilizes CTLs to clear viruses and IgE to eliminate worms is rather oversimplified. Most viruses elicit antibody responses, and the isotypes vary with the infection site. Measles and rubella viruses, for example, inhabit the skin103 and elicit mostly IgG1, IgG3 and IgG4 (Ref. 104), but rotavirus and influenza virus, which inhabit the gut and the lungs, respectively, induce strong IgA responses105 (which can be more important than CTLs for protection from reinfection106). unfortunately these local tissue responses are often missed, either because they simply aren’t measured (for example, TGFβ and secretory IgA are often left out of standard tests because they are difficult to measure) or because the relevant B and T cells home to distinctly local sites using specific chemokine receptors and adhesion molecules107,108, and discharge their antibodies and cytokines into local secretions rather than serum109.

Tissues are not simply passive recipients of immune protection What about the idea that TH2-type responses are the best mechanism to clear helminths? In mice infected with schistosomes, treatment with antibody specific for IgE actually causes a reduction in worm burden110, and worm numbers do not differ between IL-4-deficient and wild-type mice111. Furthermore, IL-4-deficient mice have lower burdens of Onchocerca microfilariae and greater resistance to reinfection than wild-type mice112. In human studies, people living in Schistosoma mansoni endemic areas who remain clinically uninfected make strong IFNγ responses to worm antigens, whereas those who maintain chronic low worm burdens make IL-4 (Refs 113,114). This suggests that an IFNγ-associated response clears the worms and prevents reinfection, whereas TH2 cell functions instead allow the worms to establish low-level colonization and produce more worms without doing serious damage. One could ask,

226 | MARCH 2011 | vOLuME 11

therefore, if the IgE responses reported against worms might be due to instructions from the worm, rather than decisions by the immune system. When studying immunity to a pathogen, it is not always clear whose agenda we are studying. viruses are known to have evolved all sorts of mechanisms to subvert or modulate immune responses115, and we should not expect less from pathogens with larger and more complex genomes116. Their survival depends on it and they have had evolutionary time to devise the mechanisms. bOX 2 gives a few examples of the strategies used by parasites and worms to modulate immune responses in their favour. The influence of history If tissues and their resident immune cells tend to promote immune response classes other than the TH1-type or DTH response, why have immunologists thought of the TH1-type response as the ‘normal’ response for so long? This may be a historical accident. Before we had in vitro culture systems and other laboratory tools, there were two main ways of measuring cellular (as opposed to humoral) immunity: the tuberculosis skin DTH test and graft rejection. Later, as we developed in vitro correlates of such cellular immunity, we looked for components that were part of these responses. For example, we generated assays for CTLs and NK cells, for macrophages that produce oxygen radicals, for T cells that produce TNF and IFNγ, and for antibodies that fix complement. In other words, we measured things that kill! If anyone wondered why immune responses should be so destructive, they probably assumed that this is the only way to fight pathogens and that the collateral damage is simply the price we pay. Because what we think influences what we do, we also geared our model systems to generate these responses. We found culture conditions to promote them in vitro and adjuvants that elicit them in vivo (is it any surprise that our ‘best’ adjuvant for cellular immunity, complete Freund’s adjuvant, contains mycobacteria?). We named the CD4+ T cells that enhance them TH1 cells, and when we discovered that IgE production and allergy seem to be driven by a different kind of helper cell, we named them TH2 cells and called their suppressive effect on TH1 cell responses ‘immune deviation’. When we discovered that orally administered antigen elicits TH3 cell responses, which involve TH cells making TGFβ and IL-10 and promoting IgA production, we www.nature.com/reviews/immunol

© 2011 Macmillan Publishers Limited. All rights reserved

PersPectives Box 2 | Class control by pathogens Viruses are known to have a plethora of mechanisms to influence and/or avoid the immune system115. Other organisms have not been as extensively studied, but some data are beginning to emerge. Some pathogens might choose to reside in tissues that not only offer good shelter and nutrition, but that also promote an effector class that is unable to effectively clear that parasite. Pathogens can also generate their own immune-influencing signals to exploit host defense strategies to their advantage. Leishmania parasites are an example. Sandflies taking a blood meal on mammalian skin induce a typical wound healing response that summons ‘alternatively activated’ macrophages that express high levels of arginase130. This is precisely what Leishmania parasites need: readily available macrophages that they can infect to start the next stage of their life cycle. How does the macrophage-associated arginase help in this process? Catabolism of arginine involves two enzymes that compete with each other. Inducible nitric oxide synthase (iNOS) generates nitric oxide, which effectively kills intracellular Leishmania, whereas arginase catabolizes arginine to ornithine, a precursor of essential nutrients for Leishmania131. Leishmania parasites influence events even more in their favour by inducing the sandfly to secrete promastigote secretory gel (PSG), a potent inducer of arginase-expressing macrophages132. The more parasites harboured by the sandfly, the more PSG in its midgut and the more it deposits in the skin when it bites. Thus, the process by which Leishmania are deposited into host skin during a natural infection induces precisely the responses that are most likely to maximize Leishmania survival and propagation, namely the migration to the site of infection of macrophages that synthesize the nutritional compounds that Leishmania require. Mycobacteria also have immune-subverting effects. Recent studies133 on early events in the establishment of mycobacterial infections have reversed long-standing assumptions about the purpose of the granuloma, one of the most ancient host defense strategies by which multicellular organisms wall off infectious agents and prevent their spread through the body134. Surprisingly, granulomas form rapidly during infection with virulent mycobacteria and have greater levels of macrophage recruitment, motility and apoptosis than those that form during infection with non-virulent mycobacteria (which are poorly formed and result in attenuated infections). The result of the accelerated granuloma formation with virulent mycobacteria is early dissemination of the infection, through the release of infected macrophages from the primary granulomas and production of secondary granulomas at distal sites133. Thus, the virulent mycobacterium converts an evolutionarily ancient form of host defense into a convenient and pliant tool that enables its survival and more efficient propagation. Bordetella bronchiseptica135, Schistosoma mansoni136–138 and Fasciola hepatica9 express molecules that promote T helper 2 cell responses, whereas Lewis antigen-expressing Helicobacter pylori139 specifically block T helper 1 cell differentiation, presumably because such strategies work to their advantage. These strategies are likely to be only the tip of the iceberg, and we will find many more as we study the relationships between pathogens, commensals, symbionts and their hosts.

mostly ignored them, or focused instead on their ability to suppress the TH1-type inflammatory response and called them TR1 cells or TReg cells117–120. Although TReg cells were originally thought to be autoreactive T cells that are educated in the thymus121 to become suppressors122,123, there are now known to be several subsets. There are natural and induced TReg cells, thymic and peripheral TReg cells and self-reactive and non-selfreactive TReg cells. There are TReg cells induced by culture in the absence of costimulatory molecules and those induced in the presence of co-inhibitory molecules or other molecules such as TGFβ, vitamin D, retinoic acid or antibodies specific for CD40 ligand. There are TReg cells that make TGFβ or IL-10 or neither. And there have been endless discussions about which ones are the ‘real’ TReg cells124. We would suggest that many of these TReg cells are actually misunderstood memory helper T cells. In an unimmunized animal, they

will have been stimulated mostly by antigens entering through the nose, mouth and intestine. To help to promote appropriate responses for these mucosal tissues they may also need to suppress other classes of response. It would not be surprising, therefore, to discover that many CD25+ memory TH cells from a normal mouse can suppress IFNγ or TNF production, graft rejection or CTL activity. One might ask why our postulated helper activity of tissue-specific TReg cells has not been observed. The main reason is that TReg cells are not often tested for helper functions. More than 99% of the work on TReg cells uses assays that measure only their ability to suppress TH1-type functions125. If we don’t measure the mucosal memory response, we will mistakenly call this tolerance, suppression or regulation. In fact, in two studies in which intestinal antibody production was measured, forkhead box P3 (FOxP3)+ cells were shown to become helper cells for the production of IgA126,127.

NATuRE REvIEWS | Immunology

We are all guilty of this. Out of all of the ways to evaluate immune responses, each laboratory tends to use only a small subset. So when we do something (such as feed antigen to animals, or add cytokines or antibodies to in vitro cultures) and notice that the immune response we are measuring decreases, we call it tolerance. Then when we transfer cells from that animal to another, or add them to other cells in vitro, and find that the recipient also fails to respond in our limited set of assays, we call it suppression. But we don’t notice that, unobserved, another type of response is increasing. If, for example, we measured TH2 cells only by their ability to suppress TH1-type responses, we would miss their capacity to promote IgE and IgG1 production and their role in allergy and asthma. This is not really suppression, but a switch of effector class. Perhaps, if we were to begin measuring the helper functions of TReg cells, we might find a new range of tools to help us modulate immunity. We suggest that it is time to discard the idea that the destructive TH1-type or DTH response is the immune system’s core mechanism. When someone tells us that the TH1-type response is the “natural” response in their experimental system, we need to ask, “what adjuvant are you using, what dose of antigen are you using, where are you injecting it and what other responses are you missing?” A few final thoughts Although the body must clear some pathogens, or at least keep them in their place, our commensal microflora must also be maintained and our tissues kept healthy in the process. To accomplish this, the immune system does not use a set of rigidly defined TH1, TH2 or TH17 cells, but a wide variety of TH cells that respond to signals from their environment to mount a carefully balanced response to adversity. Some of these cells are tissue resident and some circulate. Some have genetically defined invariant T cell receptors, whereas others are somatically generated in each individual. Each of these TH cells associates with various groups of B cells, macrophages, cytotoxic cells, neutrophils, eosinophils, basophils and tissue cells to tailor the response to the local milieu and the pathogen. In fact, we should perhaps redefine the immune system to include every tissue in the body. Tissues are not simply passive recipients of immune protection, but are active participants in their own defense. They express TLRs128,129. They produce antimicrobial peptides and antiviral cytokines, vOLuME 11 | MARCH 2011 | 227

© 2011 Macmillan Publishers Limited. All rights reserved

PersPectives such as type I IFNs. They produce ‘eat me’ signals to bring in scavenger cells, alarm signals to activate local APCs, class-influencing signals to modulate local immune responses and chemokines to recruit cells for repair, remodelling and immunity. Finally, they may potentially also transmit ‘health’ signals to send away all of these cells when they are no longer needed. To fully understand these complex interactions we will need to step back, have another look, start using assays that measure a wider array of immune functions, and embrace the complexity that we find. Polly Matzinger and Tirumalai Kamala are at the Ghost Lab, Laboratory of Cellular and Molecular Immunology, T‑Cell Tolerance and Memory Section, National Institute of Allergy and Infectious Diseases, National Institutes of Health, 9000 Rockville Pike, Bethesda, Maryland 20892, USA. Correspondence to P.M. e‑mail: [email protected] doi:10.1038/nri2940 1.

2.

3. 4.

5.

6.

7. 8. 9.

10. 11. 12. 13. 14. 15. 16. 17.

Streilein, J. W. Pathologic lesions of GVH disease in hamsters: antigenic target versus ‘innocent bystander’. Prog. Exp. Tumor Res. 16, 396–408 (1972). Elson, C. O., Reilly, R. W. & Rosenberg, I. H. Small intestinal injury in the graft versus host reaction: an innocent bystander phenomenon. Gastroenterology 72, 886–889 (1977). Krook, H. et al. A distinct Th1 immune response precedes the described Th2 response in islet xenograft rejection. Diabetes 51, 79–86 (2002). Knisely, T. L., Luckenbach, M. W., Fischer, B. J. & Niederkorn, J. Y. Destructive and nondestructive patterns of immune rejection of syngeneic intraocular tumors. J. Immunol. 138, 4515–4523 (1987). Abram, M. et al. Effects of pregnancy-associated Listeria monocytogenes infection: necrotizing hepatitis due to impaired maternal immune response and significantly increased abortion rate. Virchows Arch. 441, 368–379 (2002). Seong, S. Y. & Matzinger, P. Hydrophobicity: an ancient damage-associated molecular pattern that initiates innate immune responses. Nature Rev. Immunol. 4, 469–478 (2004). Kawai, T. & Akira, S. The role of pattern-recognition receptors in innate immunity: update on Toll-like receptors. Nature Immunol. 11, 373–384 (2010). Zhang, Q. et al. Circulating mitochondrial DAMPs cause inflammatory responses to injury. Nature 464, 104–107 (2010). Dowling, D. J. et al. Major secretory antigens of the helminth Fasciola hepatica activate a suppressive dendritic cell phenotype that attenuates Th17 cells but fails to activate Th2 immune responses. Infect. Immun. 78, 793–801 (2010). Koch, A. L. Bacterial wall as target for attack: past, present, and future research. Clin. Microbiol. Rev. 16, 673–687 (2003). Andersen-Nissen, E. et al. Evasion of Toll-like receptor 5 by flagellated bacteria. Proc. Natl Acad. Sci. USA 102, 9247–9252 (2005). Matzinger, P. Tolerance, danger, and the extended family. Annu. Rev. Immunol. 12, 991–1045 (1994). Matzinger, P. The danger model: a renewed sense of self. Science 296, 301–305 (2002). Matzinger, P. Friendly and dangerous signals: is the tissue in control? Nature Immunol. 8, 11–13 (2007). Kelso, A. Th1 and Th2 subsets: paradigms lost? Immunol. Today 16, 374–379 (1995). O’Shea, J. J. & Paul, W. E. Mechanisms underlying lineage commitment and plasticity of helper CD4+ T cells. Science 327, 1098–1102 (2010). Pulendran, B. et al. Lipopolysaccharides from distinct pathogens induce different classes of immune responses in vivo. J. Immunol. 167, 5067–5076 (2001).

18. Yang, D. et al. Eosinophil-derived neurotoxin acts as an alarmin to activate the TLR2–MyD88 signal pathway in dendritic cells and enhances Th2 immune responses. J. Exp. Med. 205, 79–90 (2008). 19. Prussin, C., Lee, J. & Foster, B. Eosinophilic gastrointestinal disease and peanut allergy are alternatively associated with IL-5+ and IL-5– TH2 responses. J. Allergy Clin. Immunol. 124, 1326–1332 (2009). 20. Hume, D. M. & Egdahl, R. H. Progressive destruction of renal homografts isolated from the regional lymphatics of the host. Surgery 38, 194–214 (1955). 21. Medawar, P. B. Immunity to homologous grafted skin; the fate of skin homografts transplanted to the brain, to subcutaneous tissue, and to the anterior chamber of the eye. Br. J. Exp. Pathol. 29, 58–69 (1948). 22. Barker, C. F. & Billingham, R. E. Immunologically privileged sites. Adv. Immunol. 25, 1–54 (1977). 23. Barker, C. F. & Billingham, R. E. The lymphatic status of hamster cheek pouch tissue in relation to its properties as a graft and as a graft site. J. Exp. Med. 133, 620–639 (1971). 24. Fijak, M. & Meinhardt, A. The testis in immune privilege. Immunol. Rev. 213, 66–81 (2006). 25. Ferguson, T. A. & Griffith, T. S. A vision of cell death: Fas ligand and immune privilege 10 years later. Immunol. Rev. 213, 228–238 (2006). 26. Caspi, R. R. Ocular autoimmunity: the price of privilege? Immunol. Rev. 213, 23–35 (2006). 27. Streilein, J. W. Ocular immune privilege: therapeutic opportunities from an experiment of nature. Nature Rev. Immunol. 3, 879–889 (2003). 28. Niederkorn, J. Y. & Stein-Streilein, J. History and physiology of immune privilege. Ocul. Immunol. Inflamm. 18, 19–23 (2010). 29. Coffman, R. L., Lebman, D. A. & Shrader, B. Transforming growth factor β specifically enhances IgA production by lipopolysaccharide-stimulated murine B lymphocytes. J. Exp. Med. 170, 1039–1044 (1989). 30. Sonoda, E. et al. Transforming growth factor β induces IgA production and acts additively with interleukin 5 for IgA production. J. Exp. Med. 170, 1415–1420 (1989). 31. Boirivant, M. et al. Vasoactive intestinal polypeptide modulates the in vitro immunoglobulin A production by intestinal lamina propria lymphocytes. Gastroenterology 106, 576–582 (1994). 32. Kimata, H. & Fujimoto, M. Vasoactive intestinal peptide specifically induces human IgA1 and IgA2 production. Eur. J. Immunol. 24, 2262–2265 (1994). 33. Fang, Y., Yu, S., Ellis, J. S., Sharav, T. & Braley-Mullen, H. Comparison of sensitivity of Th1, Th2, and Th17 cells to Fas-mediated apoptosis. J. Leukoc. Biol. 87, 1019–1028 (2010). 34. Shinohara, M. L., Kim, J. H., Garcia, V. A. & Cantor, H. Engagement of the type I interferon receptor on dendritic cells inhibits T helper 17 cell development: role of intracellular osteopontin. Immunity 29, 68–78 (2008). 35. Khoury, S. J., Hancock, W. W. & Weiner, H. L. Oral tolerance to myelin basic protein and natural recovery from experimental autoimmune encephalomyelitis are associated with downregulation of inflammatory cytokines and differential upregulation of transforming growth factor β, interleukin 4, and prostaglandin E expression in the brain. J. Exp. Med. 176, 1355–1364 (1992). 36. Miller, A., Lider, O., Roberts, A. B., Sporn, M. B. & Weiner, H. L. Suppressor T cells generated by oral tolerization to myelin basic protein suppress both in vitro and in vivo immune responses by the release of transforming growth factor β after antigen-specific triggering. Proc. Natl Acad. Sci. USA 89, 421–425 (1992). 37. Scadding, G. & Durham, S. Mechanisms of sublingual immunotherapy. J. Asthma 46, 322–334 (2009). 38. Salk, D. & Salk, J. Vaccinology of poliomyelitis. Vaccine 2, 59–74 (1984). 39. Sterner, R. T., Meltzer, M. I., Shwiff, S. A. & Slate, D. Tactics and economics of wildlife oral rabies vaccination, Canada and the United States. Emerg. Infect. Dis. 15, 1176–1184 (2009). 40. Chen, Y. et al. Peripheral deletion of antigen-reactive T cells in oral tolerance. Nature 376, 177–180 (1995). 41. Alpan, O., Rudomen, G. & Matzinger, P. The role of dendritic cells, B cells, and M cells in gut-oriented immune responses. J. Immunol. 166, 4843–4852 (2001).

228 | MARCH 2011 | vOLuME 11

42. Mestecky, J., Russell, M. W. & Elson, C. O. Perspectives on mucosal vaccines: is mucosal tolerance a barrier? J. Immunol. 179, 5633–5638 (2007). 43. Maloy, K. J., Donachie, A. M., O’Hagan, D. T. & Mowat, A. M. Induction of mucosal and systemic immune responses by immunization with ovalbumin entrapped in poly(lactide-co-glycolide) microparticles. Immunology 81, 661–667 (1994). 44. Weiner, H. L. Induction and mechanism of action of transforming growth factor-β-secreting Th3 regulatory cells. Immunol. Rev. 182, 207–214 (2001). 45. Abreu, M. T. Toll-like receptor signalling in the intestinal epithelium: how bacterial recognition shapes intestinal function. Nature Rev. Immunol. 10, 131–144 (2010). 46. Kaplan, H. J. & Niederkorn, J. Y. Regional immunity and immune privilege. Chem. Immunol. Allergy 92, 11–26 (2007). 47. Boehm, U., Klamp, T., Groot, M. & Howard, J. C. Cellular responses to interferon-γ. Annu. Rev. Immunol. 15, 749–795 (1997). 48. Tracey, K. J. & Cerami, A. Tumor necrosis factor: a pleiotropic cytokine and therapeutic target. Annu. Rev. Med. 45, 491–503 (1994). 49. Kagi, D., Ledermann, B., Burki, K., Zinkernagel, R. M. & Hengartner, H. Molecular mechanisms of lymphocyte-mediated cytotoxicity and their role in immunological protection and pathogenesis in vivo. Annu. Rev. Immunol. 14, 207–232 (1996). 50. Mukaida, N. Antibody-dependent cell-mediated cytotoxicity (ADCC). Nippon Rinsho 57 (Suppl.), 571–573 (1999). 51. MacMicking, J., Xie, Q. W. & Nathan, C. Nitric oxide and macrophage function. Annu. Rev. Immunol. 15, 323–350 (1997). 52. Carroll, M. C. The role of complement and complement receptors in induction and regulation of immunity. Annu. Rev. Immunol. 16, 545–568 (1998). 53. Chaouat, G. et al. TH1/TH2 paradigm in pregnancy: paradigm lost? Int. Arch. Allergy Immunol. 134, 93–119 (2004). 54. Linthicum, D. S., Mackay, I. R. & Carnegie, P. R. Measurement of cell-mediated inflammation in experimental murine autoimmune encephalomyelitis by radioisotopic labeling. J. Immunol. 123, 1799–1805 (1979). 55. Gilliet, M. & Lande, R. Antimicrobial peptides and self-DNA in autoimmune skin inflammation. Curr. Opin. Immunol. 20, 401–407 (2008). 56. Nijhuis, L. E., Olivier, B. J. & de Jonge, W. J. Neurogenic regulation of dendritic cells in the intestine. Biochem. Pharmacol. 80, 2002–2008 (2010). 57. Tokuyama, H. & Tokuyama, Y. Retinoids enhance IgA production by lipopolysaccharide-stimulated murine spleen cells. Cell. Immunol. 150, 353–363 (1993). 58. Baeke, F., Takiishi, T., Korf, H., Gysemans, C. & Mathieu, C. Vitamin D: modulator of the immune system. Curr. Opin. Pharmacol. 10, 482–496 (2010). 59. Wilbanks, G. A., Mammolenti, M. & Streilein, J. W. Studies on the induction of anterior chamberassociated immune deviation (ACAID) III. Induction of ACAID depends upon intraocular transforming growth factor-β. Eur. J. Immunol. 22, 165–173 (1992). 60. Kosiewicz, M. M., Alard, P. & Streilein, J. W. Alterations in cytokine production following intraocular injection of soluble protein antigen: impairment in IFN-γ and induction of TGF-β and IL-4 production. J. Immunol. 161, 5382–5390 (1998). 61. Stavnezer, J. & Kang, J. The surprising discovery that TGFβ specifically induces the IgA class switch. J. Immunol. 182, 5–7 (2009). 62. Penna, G. et al. 1,25-Dihydroxyvitamin D3 selectively modulates tolerogenic properties in myeloid but not plasmacytoid dendritic cells. J. Immunol. 178, 145–153 (2007). 63. Imai, T. et al. Selective recruitment of CCR4-bearing Th2 cells toward antigen-presenting cells by the CC chemokines thymus and activation-regulated chemokine and macrophage-derived chemokine. Int. Immunol. 11, 81–88 (1999). 64. Lemire, J. M. & Adams, J. S. 1,25-Dihydroxyvitamin D3 inhibits the passive transfer of cellular immunity by a myelin basic protein-specific T cell clone. J. Bone Miner. Res. 7, 171–177 (1992). 65. Cheroutre, H. & Lambolez, F. The thymus chapter in the life of gut-specific intra epithelial lymphocytes. Curr. Opin. Immunol. 20, 185–191 (2008).

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved

PersPectives 66. Groh, V., Steinle, A., Bauer, S. & Spies, T. Recognition of stress-induced MHC molecules by intestinal epithelial γδ T cells. Science 279, 1737–1740 (1998). 67. Treiner, E. et al. Selection of evolutionarily conserved mucosal-associated invariant T cells by MR1. Nature 422, 164–169 (2003). 68. Hue, S. et al. A direct role for NKG2D/MICA interaction in villous atrophy during celiac disease. Immunity 21, 367–377 (2004). 69. Bendelac, A., Savage, P. B. & Teyton, L. The biology of NKT cells. Annu. Rev. Immunol. 25, 297–336 (2007). 70. Jameson, J. & Havran, W. L. Skin γδ T-cell functions in homeostasis and wound healing. Immunol. Rev. 215, 114–122 (2007). 71. Boismenu, R. & Havran, W. L. Modulation of epithelial cell growth by intraepithelial γδ T cells. Science 266, 1253–1255 (1994). 72. Boismenu, R., Hobbs, M. V., Boullier, S. & Havran, W. L. Molecular and cellular biology of dendritic epidermal T cells. Semin. Immunol. 8, 323–331 (1996). 73. Braun, A., Wiebe, P., Pfeufer, A., Gessner, R. & Renz, H. Differential modulation of human immunoglobulin isotype production by the neuropeptides substance P, NKA and NKB. J. Neuroimmunol. 97, 43–50 (1999). 74. Dunzendorfer, S. & Wiedermann, C. J. Neuropeptides and the immune system: focus on dendritic cells. Crit. Rev. Immunol. 21, 523–557 (2001). 75. Ho, W. Z., Lai, J. P., Zhu, X. H., Uvaydova, M. & Douglas, S. D. Human monocytes and macrophages express substance P and neurokinin-1 receptor. J. Immunol. 159, 5654–5660 (1997). 76. Lai, J. P., Douglas, S. D. & Ho, W. Z. Human lymphocytes express substance P and its receptor. J. Neuroimmunol. 86, 80–86 (1998). 77. Qian, B. F., Zhou, G. Q., Hammarstrom, M. L. & Danielsson, A. Both substance P and its receptor are expressed in mouse intestinal T lymphocytes. Neuroendocrinology 73, 358–368 (2001). 78. Shibata, M., Hisajima, T., Nakano, M., Goris, R. C. & Funakoshi, K. Morphological relationships between peptidergic nerve fibers and immunoglobulin Aproducing lymphocytes in the mouse intestine. Brain Behav. Immun. 22, 158–166 (2008). 79. Meek, B., Speijer, D., de Jong, P. T., de Smet, M. D. & Peek, R. The ocular humoral immune response in health and disease. Prog. Retin. Eye Res. 22, 391–415 (2003). 80. Lausch, R. N., Monteiro, C., Kleinschrodt, W. R. & Oakes, J. E. Superiority of antibody versus delayed hypersensitivity in clearance of HSV-1 from eye. Invest. Ophthalmol. Vis. Sci. 28, 565–570 (1987). 81. Everson, M. P. et al. Dendritic cells from different tissues induce production of different T cell cytokine profiles. J. Leukoc. Biol. 59, 494–498 (1996). 82. Bode, U. et al. Dendritic cell subsets in lymph nodes are characterized by the specific draining area and influence the phenotype and fate of primed T cells. Immunology 123, 480–490 (2008). 83. Iliev, I. D., Mileti, E., Matteoli, G., Chieppa, M. & Rescigno, M. Intestinal epithelial cells promote colitisprotective regulatory T-cell differentiation through dendritic cell conditioning. Mucosal Immunol. 2, 340–350 (2009). 84. Rimoldi, M. et al. Intestinal immune homeostasis is regulated by the crosstalk between epithelial cells and dendritic cells. Nature Immunol. 6, 507–514 (2005). 85. Xia, S. et al. Hepatic microenvironment programs hematopoietic progenitor differentiation into regulatory dendritic cells, maintaining liver tolerance. Blood 112, 3175–3185 (2008). 86. Zeuthen, L. H., Fink, L. N. & Frokiaer, H. Epithelial cells prime the immune response to an array of gutderived commensals towards a tolerogenic phenotype through distinct actions of thymic stromal lymphopoietin and transforming growth factor-β. Immunology 123, 197–208 (2008). 87. Kunkel, E. J. et al. Lymphocyte CC chemokine receptor 9 and epithelial thymus-expressed chemokine (TECK) expression distinguish the small intestinal immune compartment: epithelial expression of tissue-specific chemokines as an organizing principle in regional immunity. J. Exp. Med. 192, 761–768 (2000). 88. Iwasaki, A. & Kelsall, B. L. Freshly isolated Peyer’s patch, but not spleen, dendritic cells produce interleukin 10 and induce the differentiation of T helper type 2 cells. J. Exp. Med. 190, 229–239 (1999).

89. Mora, J. R. et al. Reciprocal and dynamic control of CD8 T cell homing by dendritic cells from skin- and gut-associated lymphoid tissues. J. Exp. Med. 201, 303–316 (2005). 90. Calzascia, T. et al. Homing phenotypes of tumorspecific CD8 T cells are predetermined at the tumor site by crosspresenting APCs. Immunity 22, 175–184 (2005). 91. Alyanakian, M. A. et al. Diversity of regulatory CD4+ T cells controlling distinct organ-specific autoimmune diseases. Proc. Natl Acad. Sci. USA 100, 15806–15811 (2003). 92. Alpan, O., Bachelder, E., Isil, E., Arnheiter, H. & Matzinger, P. ‘Educated’ dendritic cells act as messengers from memory to naive T helper cells. Nature Immunol. 5, 615–622 (2004). 93. De Smedt, T. et al. Antigen-specific T lymphocytes regulate lipopolysaccharide-induced apoptosis of dendritic cells in vivo. J. Immunol. 161, 4476–4479 (1998). 94. Helft, J. et al. Antigen-specific T-T interactions regulate CD4 T-cell expansion. Blood 112, 1249–1258 (2008). 95. Martinon, F., Mayor, A. & Tschopp, J. The inflammasomes: guardians of the body. Annu. Rev. Immunol. 27, 229–265 (2009). 96. Yu, L., Wang, L. & Chen, S. Endogenous Toll-like receptor ligands and their biological significance. J. Cell. Mol. Med. 14, 2592–2603 (2010). 97. 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). 98. Belkaid, Y. The role of CD4+CD25+ regulatory T cells in Leishmania infection. Expert Opin. Biol. Ther. 3, 875–885 (2003). 99. Jankovic, D. et al. Conventional T-bet+Foxp3– Th1 cells are the major source of host-protective regulatory IL-10 during intracellular protozoan infection. J. Exp. Med. 204, 273–283 (2007). 100. Shah, P. D., Gilbertson, S. M. & Rowley, D. A. Dendritic cells that have interacted with antigen are targets for natural killer cells. J. Exp. Med. 162, 625–636 (1985). 101. Sitkovsky, M. & Lukashev, D. Regulation of immune cells by local-tissue oxygen tension: HIF1α and adenosine receptors. Nature Rev. Immunol. 5, 712–721 (2005). 102. Pulendran, B., Palucka, K. & Banchereau, J. Sensing pathogens and tuning immune responses. Science 293, 253–256 (2001). 103. Takahashi, H. et al. Detection and comparison of viral antigens in measles and rubella rashes. Clin. Infect. Dis. 22, 36–39 (1996). 104. Isa, M. B. et al. Comparison of immunoglobulin G subclass profiles induced by measles virus in vaccinated and naturally infected individuals. Clin. Diagn. Lab. Immunol. 9, 693–697 (2002). 105. Gonzalez, R., Franco, M., Sarmiento, L., Romero, M. & Schael, I. P. Serum IgA levels induced by rotavirus natural infection, but not following immunization with the RRV-TV vaccine (Rotashield), correlate with protection. J. Med. Virol. 76, 608–612 (2005). 106. Liew, F. Y., Russell, S. M., Appleyard, G., Brand, C. M. & Beale, J. Cross-protection in mice infected with influenza A virus by the respiratory route is correlated with local IgA antibody rather than serum antibody or cytotoxic T cell reactivity. Eur. J. Immunol. 14, 350–356 (1984). 107. Mora, J. R. et al. Generation of gut-homing IgAsecreting B cells by intestinal dendritic cells. Science 314, 1157–1160 (2006). 108. Fagarasan, S. & Honjo, T. Intestinal IgA synthesis: regulation of front-line body defences. Nature Rev. Immunol. 3, 63–72 (2003). 109. Brandtzaeg, P. Induction of secretory immunity and memory at mucosal surfaces. Vaccine 25, 5467–5484 (2007). 110. Amiri, P. et al. Anti-immunoglobulin E treatment decreases worm burden and egg production in Schistosoma mansoni-infected normal and interferon γ knockout mice. J. Exp. Med. 180, 43–51 (1994). 111. King, C. L., Malhotra, I. & Jia, X. Schistosoma mansoni: protective immunity in IL-4-deficient mice. Exp. Parasitol. 84, 245–252 (1996). 112. Hogarth, P. J., Folkard, S. G., Taylor, M. J. & Bianco, A. E. Accelerated clearance of Onchocerca microfilariae and resistance to reinfection in interleukin-4 gene knockout mice. Parasite Immunol. 17, 653–657 (1995).

NATuRE REvIEWS | Immunology

113. Brito, C. F., Caldas, I. R., Coura Filho, P., Correa-Oliveira, R. & Oliveira, S. C. CD4+ T cells of schistosomiasis naturally resistant individuals living in an endemic area produce interferon-γ and tumour necrosis factor-α in response to the recombinant 14 kDa Schistosoma mansoni fatty acid-binding protein. Scand. J. Immunol. 51, 595–601 (2000). 114. Viana, I. R. et al. Interferon-γ production by peripheral blood mononuclear cells from residents of an area endemic for Schistosoma mansoni. Trans. R. Soc. Trop. Med. Hyg. 88, 466–470 (1994). 115. Tortorella, D., Gewurz, B. E., Furman, M. H., Schust, D. J. & Ploegh, H. L. Viral subversion of the immune system. Annu. Rev. Immunol. 18, 861–926 (2000). 116. Terrazas, C. A., Terrazas, L. I. & Gomez-Garcia, L. Modulation of dendritic cell responses by parasites: a common strategy to survive. J. Biomed. Biotechnol. 2010, 357106 (2010). 117. Thorstenson, K. M. & Khoruts, A. Generation of anergic and potentially immunoregulatory CD25+CD4 T cells in vivo after induction of peripheral tolerance with intravenous or oral antigen. J. Immunol. 167, 188–195 (2001). 118. Zhang, X., Izikson, L., Liu, L. & Weiner, H. L. Activation of CD25+CD4+ regulatory T cells by oral antigen administration. J. Immunol. 167, 4245–4253 (2001). 119. Coombes, J. L. et al. A functionally specialized population of mucosal CD103+ DCs induces Foxp3+ regulatory T cells via a TGF-β and retinoic aciddependent mechanism. J. Exp. Med. 204, 1757–1764 (2007). 120. Sun, J. B., Raghavan, S., Sjoling, A., Lundin, S. & Holmgren, J. Oral tolerance induction with antigen conjugated to cholera toxin B subunit generates both Foxp3+CD25+ and Foxp3–CD25– CD4+ regulatory T cells. J. Immunol. 177, 7634–7644 (2006). 121. 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). 122. Sakaguchi, S., Yamaguchi, T., Nomura, T. & Ono, M. Regulatory T cells and immune tolerance. Cell 133, 775–787 (2008). 123. Shevach, E. M. Mechanisms of Foxp3+ T regulatory cell-mediated suppression. Immunity 30, 636–645 (2009). 124. Shevach, E. M. From vanilla to 28 flavors: multiple varieties of T regulatory cells. Immunity 25, 195–201 (2006). 125. Mempel, T. R. et al. Regulatory T cells reversibly suppress cytotoxic T cell function independent of effector differentiation. Immunity 25, 129–141 (2006). 126. Tsuji, M. et al. Preferential generation of follicular B helper T cells from Foxp3+ T cells in gut Peyer’s patches. Science 323, 1488–1492 (2009). 127. Cong, Y., Feng, T., Fujihashi, K., Schoeb, T. R. & Elson, C. O. A dominant, coordinated T regulatory cell–IgA response to the intestinal microbiota. Proc. Natl Acad. Sci. USA 106, 19256–19261 (2009). 128. Zarember, K. A. & Godowski, P. J. Tissue expression of human Toll-like receptors and differential regulation of Toll-like receptor mRNAs in leukocytes in response to microbes, their products, and cytokines. J. Immunol. 168, 554–561 (2002). 129. Hammad, H. et al. House dust mite allergen induces asthma via Toll-like receptor 4 triggering of airway structural cells. Nature Med. 15, 410–416 (2009). 130. Shearer, J. D., Richards, J. R., Mills, C. D. & Caldwell, M. D. Differential regulation of macrophage arginine metabolism: a proposed role in wound healing. Am. J. Physiol. 272, E181–E190 (1997). 131. Kropf, P. et al. Arginase and polyamine synthesis are key factors in the regulation of experimental leishmaniasis in vivo. FASEB J. 19, 1000–1002 (2005). 132. Rogers, M. et al. Proteophosophoglycans regurgitated by Leishmania-infected sand flies target the L-arginine metabolism of host macrophages to promote parasite survival. PLoS Pathog. 5, e1000555 (2009). 133. Davis, J. M. & Ramakrishnan, L. The role of the granuloma in expansion and dissemination of early tuberculous infection. Cell 136, 37–49 (2009). 134. Ulrichs, T. & Kaufmann, S. H. New insights into the function of granulomas in human tuberculosis. J. Pathol. 208, 261–269 (2006). 135. Siciliano, N. A., Skinner, J. A. & Yuk, M. H. Bordetella bronchiseptica modulates macrophage phenotype leading to the inhibition of CD4+ T cell proliferation and the initiation of a Th17 immune response. J. Immunol. 177, 7131–7138 (2006).

vOLuME 11 | MARCH 2011 | 229 © 2011 Macmillan Publishers Limited. All rights reserved

PersPectives 136. Everts, B. et al. Omega-1, a glycoprotein secreted by Schistosoma mansoni eggs, drives Th2 responses. J. Exp. Med. 206, 1673–1680 (2009). 137. Steinfelder, S. et al. The major component in schistosome eggs responsible for conditioning dendritic cells for Th2 polarization is a T2 ribonuclease (omega-1). J. Exp. Med. 206, 1681–1690 (2009). 138. van der Kleij, D. et al. A novel host-parasite lipid cross-talk. Schistosomal lyso-phosphatidylserine activates Toll-like receptor 2 and affects immune polarization. J. Biol. Chem. 277, 48122–48129 (2002). 139. Bergman, M. P. et al. Helicobacter pylori modulates the T helper cell 1/T helper cell 2 balance through phase-variable interaction between lipopolysaccharide and DC-SIGN. J. Exp. Med. 200, 979–990 (2004). 140. Meade, R. et al. Transforming growth factor-beta 1 inhibits murine immediate and delayed type hypersensitivity. J. Immunol. 149, 521–528 (1992). 141. Bommireddy, R. & Doetschman, T. TGFβ1 and Treg cells: alliance for tolerance. Trends Mol. Med. 13, 492–501 (2007). 142. Goetzl, E. J. et al. Enhanced delayed-type hypersensitivity and diminished immediate-type hypersensitivity in mice lacking the inducible VPAC2

receptor for vasoactive intestinal peptide. Proc. Natl Acad. Sci. USA 98, 13854–13859 (2001). 143. Delgado, M., Chorny, A., Gonzalez-Rey, E. & Ganea, D. Vasoactive intestinal peptide generates CD4+CD25+ regulatory T cells in vivo. J. Leukoc. Biol. 78, 1327–1338 (2005). 144. Taylor, A. W., Yee, D. G., Nishida, T. & Namba, K. Neuropeptide regulation of immunity. The immunosuppressive activity of alpha-melanocytestimulating hormone (α-MSH). Ann. NY Acad. Sci. 917, 239–247 (2000). 145. Taylor, A. W. Modulation of regulatory T cell immunity by the neuropeptide alpha-melanocyte stimulating hormone. Cell. Mol. Biol. 49, 143–149 (2003). 146. Mazzucchelli, R. et al. Development of regulatory T cells requires IL-7Rα stimulation by IL-7 or TSLP. Blood 112, 3283–3292 (2008). 147. He, B. et al. Intestinal bacteria trigger T cellindependent immunoglobulin A2 class switching by inducing epithelial-cell secretion of the cytokine APRIL. Immunity 26, 812–826 (2007). 148. Wiedermann, U. et al. Vitamin A deficiency increases inflammatory responses. Scand. J. Immunol. 44, 578–584 (1996). 149. Cantorna, M. T., Nashold, F. E. & Hayes, C. E. Vitamin A deficiency results in a priming environment

230 | MARCH 2011 | vOLuME 11

conducive for Th1 cell development. Eur. J. Immunol. 25, 1673–1679 (1995). 150. Kang, S. G., Lim, H. W., Andrisani, O. M., Broxmeyer, H. E. & Kim, C. H. Vitamin A metabolites induce guthoming FoxP3+ regulatory T cells. J. Immunol. 179, 3724–3733 (2007). 151. Sun, C. M. et al. Small intestine lamina propria dendritic cells promote de novo generation of Foxp3 T reg cells via retinoic acid. J. Exp. Med. 204, 1775–1785 (2007).

Acknowledgements

We thank A. Bendelac, T. Honjo, B. Jabri, Y. Rosenberg, F. Di Rosa, R. Schwartz, N. Singh, the ‘ghosts’ (K. Abdi, A. PerezDiez, A. Morgun and N. Shulzhenko) and especially P. Chappert for commenting on the manuscript. T.K. would like to express special thanks to D. Usharauli for his encouragement and support during the preparation of this Essay. We apologize to the authors whose work we didn’t cite owing to lack of space. This work was supported by the intramural program of the National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, Maryland, USA.

Competing interests statement

The authors declare no competing financial interests.

www.nature.com/reviews/immunol © 2011 Macmillan Publishers Limited. All rights reserved