nature reviews immunology - september 2013.pdf-meta

This article is a tribute to Dr Leo Lefranois, Professor and Chairman of the Department of Immunology and Director of the Center for Integrated Immunology and Vaccine Research, University of Connecticut Health Center, Connecticut, USA. The editors of Nature Reviews Immunology are deeply saddened to hear of his unexpected passing and extend our condolences to his family and colleagues. His contribution to the field of immunology, in particular to our understanding of memory T cell responses, has been extensive and will be long remembered; this most recent study is testament to that.

T cells best known for their innate-like characteristics are not generally considered to be the first choice cells for a memory recall response. However, recent studies, including this paper by Leo Lefranois and colleagues, suggests that T cells in the intestinal tissues can in fact provide long-lasting memory responses, similarly to conventional T cells.

Research investigating the role of T cell responses in protection against Listeria spp. infection has been hampered by the absence of a mouse model that mimics natural infection. So the authors used a modified strain of Listeria monocytogenes that effi-ciently invades the mouse intestinal epithelium and that establishes a true enteric infection following oral inocu-lation. After oral infection, a large population of CD27CD44hi T cells appeared in the mesenteric lymph nodes (MLNs) and was still detect-able there 5 months later. Following secondary challenge with oral L. monocytogenes, this Tcell popula-tion rapidly expanded in the MLNs, blood and intestinal lamina propria, but not in the intestinal epithelium or peripheral lymph nodes (PLNs), which suggests that these cells are activated in the MLNs and that they migrate via the blood to the lamina propria.

Analysis of the specificity of this recall response showed that only the CD27CD44hi T cell subset, and not other T cell subsets, responded to L. monocytogenes; the CD27CD44hi T cell subset only expanded following oral infection not following intravenous infection and they did not respond to oral infec-tion with another intestinal bacterial pathogen. In addition, the kinetics of the Tcell secondary response were comparable to those of the antigen-specific T cell memory response. These observations support the idea that the T cell response is a context-dependent, pathogen-specific, bone fide memory response.

Further analysis confirmed that the CD27CD44hi T cell population that expands in the MLNs following infection with

L.monocytogenes was distinct to the T cell subsets in the PLNs. In particular, the MLN T cell population was unique in its ability to simultaneously produce high levels of both interferon- (IFN) and interleukin-17A (IL-17A). The small subset of MLN memory Tcells expressed either IFN or IL-17A 3 months after infection, and during the secondary response the mucosal MLN T cell subset upregulated both cytokines and became the main source of IL-17A.

So, does this T cell memory subset have a protective role in L.monocytogenes infection? Treatment of the mice with a monoclonal antibody specific for the Tcell receptor (TCR), which causes TCR internalization and which hinders the ability of the Tcells to respond to antigens, did not affect protection against L.monocytogenes. Furthermore, depletion of both CD4+ and CD8+ Tcells only resulted in a minimal loss of protection. However, when CD4+ and CD8+ Tcell depletion was combined with TCR internalization, the mice suffered a much greater loss of protection in L.monocytogenes infection.

So, the authors have identified mouse T cells that can form a stable memory population and that can provide protection in the intes-tinal mucosa in collaboration with Tcell memory.

Lucy Bird

M U C O S A L I M M U N O LO GY

What memories are made of

ORIGINAL RESEARCH PAPER Sheridan, B. S. et al. T cells exhibit multifunctional and protective memory in intestinal tissues. Immunity 39, 184195 (2013)FURTHER READING Vantourout, P. & Hayday, A. Six-of-the-best: unique contributions of T cells to immunology. Nature Rev. Immunol. 13, 88100 (2013)

mouse Tcells that can form a stable memory population and that can provide protection in the intestinal mucosa

Photodisc/GETTY

R E S E A R C H H I G H L I G H T S

NATURE REVIEWS | IMMUNOLOGY VOLUME 13 | SEPTEMBER 2013

Nature Reviews Immunology | AOP, published online 9 August 2013; doi:10.1038/nri3525

2013 Macmillan Publishers Limited. All rights reserved

The transcription factor PU.1 sup-ports myeloid cell lineage differentia-tion by establishing transcriptional networks that involve several feedback loops. Here, Rothenberg and colleagues describe a previously unappreciated mechanism of positive feedback regulation by PU.1 that is controlled by the cell cycle.

The authors used knock-in PU.1 reporter mice to screen PU.1 expres-sion levels during the differentiation of lineage-negative KIT+CD27+ fetal liver progenitor cells. Consistent with previous findings, these early multipotent haematopoietic pro-genitor cells expressed intermediate PU.1 levels, and they increased their expression of PU.1 as they differentiated into macrophages but decreased their expression of PU.1 as they differentiated into B cells. Remarkably, analysis of time-lapse microscopy movies showed that PU.1 synthesis rates did not change during the differentiation of myeloid progenitor cells into macrophages, but the length of the cell cycle increased. As the half-life of PU.1 is

longer than the duration of the cell cycle of haematopoietic progenitor cells, this cell cycle lengthening led to PU.1 accumulation. By contrast, the length of the cell cycle in B cells was comparable with the length of the cell cycle in progenitor cells, and PU.1 expression was downregulated through transcriptional regulation during Bcell differentiation.

So, is cell cycle control essential for PU.1 accumulation and myeloid cell differentiation? To test this, the authors increased the length of the cell cycle in progenitor cells by either transducing them with cyclin-dependent kinase (CDK) inhibitors or treating them with small-molecule cell cycle inhibitors. Cell cycle-arrested cells had higher levels of PU.1, as well as a higher output of myeloid cells, which required PU.1 activity as it was suppressed in cells that had been transduced with a PU.1 antagonist. Moreover, the transduction of progen-itor cells with exogenous PU.1 did not increase the synthesis rate of PU.1, but instead increased the length of the cell cycle, which led to PU.1 accumulation.

Consistent with these observations, the forced expression of exogenous PU.1 did not further activate the transcription of endogenous PU.1, but it decreased the transcription of cell cycle-promoting factors, including cyclin D2, CDC25A (also known as Mphase inducer phosphatase 1), MYB and MYC.

Thus, PU.1 levels are not only controlled through direct tran-scriptional feedback regulation (as seen in differentiating B cells) but also through cell cycle lengthening, which seems to result from, and to maintain, high PU.1 levels in devel-oping macrophages. Mathematical modelling confirmed that lymphoid and myeloid cell differentiation can depend on these two regulatory prin-ciples, and the authors suggest that transcription factors that have a long half-life might prolong the cell cycle to sustain a stable cell fate.

Maria Papatriantafyllou

H A E M ATO P O I E S I S

A long cell cycle for myeloid differentiation

ORIGINAL RESEARCH PAPER Kueh, H. Y. et al. Positive feedback between PU.1 and the cell cycle controls myeloid differentiation. Science 341, 670673 (2013)

transcription factors that have a long half-life might prolong the cell cycle to sustain a stable cell fate

NPG



Nature Reviews Immunology | AOP, published online 26 July 2013; doi:10.1038/nri3513


A new study from the group of J.Oriol Sunyer provides a fascinating insight into the evolutionary origins of mucosal immune defences. The authors show that the specialized adaptive immune mechanisms that are used to promote intestinal immu-nity in teleost fish are also used to protect their skin.

Early vertebrates such as teleost fish originated in aquatic environ-ments and their skin resembles a mucosal surface, as it comprises liv-ing epithelial cells and contains many

mucus-producing cells but lacks keratinization. The authors reasoned that the immune responses in the skin of teleost fish may resemble the responses seen in mucosal tissues, such as the gut.

A key hallmark of mucosal immunology in mammals is the pro-duction of polymeric IgA antibod-ies, which regulate microorganisms in a non-inflammatory manner. The authors previously found that intestinal immunity in teleost fish involves the production of IgT, which is analogous to mammalian IgA. Therefore, they examined whether IgT is associated with the skin of teleost fish. Indeed, IgT could be detected in the skin mucus of rainbow trout (Oncorhynchus mykiss), mainly in its polymeric form. Epithelial cells in trout skin also expressed the polymeric immu-noglobulin receptor, which is needed to transport immunoglobulins to mucosal surfaces. Higher levels of IgM than IgT were detected in both trout serum and skin mucus, but the ratio of IgT to IgM was 38-fold higher in the skin mucus compared with in the serum. Furthermore, ~60% of B cells isolated from trout skin were IgT-positive, which sug-gests that IgT has important func-tions at this site.

What might these functions be? The authors found that ~38% of trout skin bacteria were coated with IgT, whereas only ~12% were coated with IgM. In fact, more than 50% of the total IgT present in the skin mucus was found to be coating bacteria

(compared with 8% of total IgM). The authors suggest that this IgT is likely to be crucial for regulating the skin microbiota.

They further investigated whether IgT contributes to protec-tive immunity by infecting trout with the skin-tropic parasite Ichthyophtirius multifiliis, which causes white spot disease. Fish that survived I. multifiliis infection had a marked accumulation of IgT-positive B cells in the skin epidermis, but no accumulation of IgM-positive B cells. The characteristic white spots of the disease are caused by I.multifiliis trophonts, and the authors found that trophonts on the epidermis of infected trout stained with antibodies specific for IgT but not with those specific for IgM. Finally, they showed that IgT was the main parasite-specific immunoglobulin measurable in the skin mucus of trout that survived I.multifiliis infection, whereas IgM was the main parasite-specific immunoglobulin found in the serum.

These data show that adaptive immune responses in the skin of teleost fish share common features with mammalian mucosal immune responses. The authors suggest that the immunoglobulin responses that protect our mucosal surfaces are based on primordially conserved principles.

Yvonne Bordon

E VO L U T I O N

A gutsy defence of the skin

ORIGINAL RESEARCH PAPER Xu, Z. et al. Teleost skin, an ancient mucosal surface that elicits gut-like immune responses. Proc. Natl Acad. Sci. USA 110, 1309713102 (2013)

IgT is shown surrounding an Ichthyophtirius multifiliis trophont on the skin epidermis of an infected rainbow trout. The I. multifiliis trophont is stained in magenta, IgT is stained in green and the cell nuclei are stained with DAPI (blue). Scale bar represents 20 m. Image courtesy of D. G. Atria and J. O. Sunyer, School of Veterinary Medicine, University of Pennsylvania, USA.





Successful immunity to hepatitisB virus (HBV) infection depends on age: in most cases, the infection of adults results in viral clearance which depends on a potent, diverse adaptive immune response but in infants the virus usually persists. Publicover etal. now show that lymphocyte organization and immune priming in the liver, which are influenced by the age-dependent production of CXC-chemokine ligand 13 (CXCL13) by hepatic macrophages, contribute to HBV clearance in adults.

This group previously devel-oped a transgenic mouse model of HBV infection that mimics key aspects of human HBV clearance or persistence. This model involves the expression of HBV transgenes that encode HBV antigens or intact virus in the livers of mice lacking an adaptive immune system. The transfer of wild-type adult immune cells to adult transgenic mice results in an effective immune response and disease kinetics similar to those observed in humans who clear HBV infection. By contrast, the transfer of adult immune cells to young transgenic mice results in a restricted immune response and disease kinetics similar to those of infants who have persistent HBV infection.

Using this mouse model, the authors showed that priming of an effective HBV-specific adaptive immune response first occurs in the liver, followed by the lymph nodes and the spleen. Thus, is there a defect in priming in the liver of young mice that results in viral per-sistence? Hepatic clusters of mac-rophages, dendritic cells, B cells and T cells were shown to form in adult transgenic mice following immune cell transfer. By contrast, the number of such clusters was greatly reduced in the livers of young mice. These mice also had lower numbers of IgG+ cells and IgG+ clusters compared with adult mice, which suggests that there is reduced B cell differentiation and antibody class switching in the liver.

Macrophage depletion in adult transgenic mice using clodronate liposomes before immune cell transfer resulted in an altered hepatic lymphocyte organization that was similar to young mice. Furthermore, these adult mice did not develop a robust HBV-specific adaptive immune response. Hepatic macrophages from either transgenic or wild-type mice, but not macrophages from other tis-sues, were shown to increase their expression of CXCL13 with age. Adult recipients of immune cells that lacked the expression of the CXCL13 receptor, CXC-chemokine receptor5 (CXCR5), developed an ineffective immune response to HBV, had fewer IgG+ B cell clusters in the liver and had lower numbers of isotype-switched B cells com-pared with recipients of wild-type immune cells.

Finally, the authors showed that CXCL13 expression increases in an age-dependent manner in the human liver and is increased in the blood of adults who have cleared HBV infection.

Taken together, these data suggest that the age-dependent increase in CXCL13 production by hepatic macrophages contributes to the liver lymphocyte organization and the immune priming that facilitates effective immunity to HBV in adults.

Olive Leavy

A N T I V I R A L I M M U N I T Y

A mature way of controlling HBV

ORIGINAL RESEARCH PAPER Publicover, J. et al. Age-dependent hepatic lymphoid organization directs successful immunity to hepatitis B. J. Clin. Invest. http://dx.doi.org/10.1172/JCI68182 (2013)GETTY





A U TO I M M U N I T Y

SCARF1 helps clean up the deadA failure of the immune system to remove apoptotic cells can lead to autoimmune disease, such as systemic lupus erythematosus (SLE). This study identifies a key scavenger receptor that is used by phagocytes to recognize and to engulf apoptotic cells. Similar to the Caenorhabditis elegans homologue CED-1, SCARF1 (scavenger receptor class F member 1) was found to bind to apoptotic cells but not to live cells. This interaction did not involve direct binding to known eat-me signals such as phosphatidylserine, but rather involved specific recognition of the complement component C1q, which bound phosphatidylserine on dying cells. Dendritic cells, and to a lesser extent macrophages and endothelial cells, from Scarf1/ mice had impaired uptake of apoptotic cells. Consequently, dying cells accumulated in Scarf1/ mice and the mice developed a lupus-like disease, which was associated with autoantibody production, nephritis and dermatitis. ORIGINAL RESEARCH PAPER Ramirez-Ortiz, Z. G. et al. The scavenger receptor SCARF1 mediates the clearance of apoptotic cells and prevents autoimmunity. Nature Immunol. http://dx.doi.org/10.1038/ni.2670 (2013)

N E U R O I M M U N O LO GY

Linking immune and emotional healthIndirect evidence suggesting a link between immune function and mood disorders is supported by the anxiety-like behaviour that is characteristic of recombination-activating gene 1 (Rag1)/ mice. This study showed that the presence of CD4+ Tcells but not of CD8+ T cells in Rag1/ OT-II-transgenic mice but not in Rag1/ OT-I-transgenic mice reverts the increased digging and marble-burying behaviours of Rag1/ mice. Transient depletion or reconstitution of CD4+ or CD8+ T cells did not affect these activites, which indicates that life-long immunodeficient conditions are required to affect behaviour. There were no differences in systemic factors or in brain anatomy that could be an explanation for the altered emotional behaviour. Whole-brain microarray analysis showed that Rag1/ OT-II mice have a genetic fingerprint more similar to wild-type mice than to Rag1/ mice. Nine main signalling pathways (including genes involved in various neuropsychological conditions) were significantly altered in Rag1/ mice compared with wild-type mice.ORIGINAL RESEARCH PAPER Rattazzi, L. et al. CD4+ but not CD8+ T cells revert the impaired emotional behavior of immunocompromised RAG-1-deficient mice. Trans. Psych. 3, e280 (2013)

I M M U N E R E G U L AT I O N

Long non-coding RNAs in the immune systemStudies from the past few years have shown a role for long non-coding RNAs (lncRNAs) in regulating a range of physiological processes. Two studies now report a role for lncRNAs in the immune system. Rapicavoli et al. describe the induction of Lethe, a pseudogene lncRNA, by tumour necrosis factor and interleukin-1. Lethe negatively regulates nuclear factor-B signalling by binding directly to RELA. Lethe expression decreases with age, which might be associated with a decreased ability to control the inflammatory response. Carpenter et al. describe the induction of lincRNA-Cox2 downstream of Toll-like receptor signalling, which mediates the activation and repression of distinct sets of immune target genes. Transcriptional repression involves the interaction of lincRNA-Cox2 with heterogeneous nuclear ribonucleoproteins.ORIGINAL RESEARCH PAPERS Rapicavoli, N. A. et al. A mammalian pseudogene lncRNA at the interface of inflammation and anti-inflammatory therapeutics. eLIFE 2, e00762 (2013)| Carpenter, S. et al. A long noncoding RNA mediates both activation and repression of immune response genes. Science http://dx.doi.org/10.1126/science.1240925 (2013)

IN BRIEF




The stability of the regulatory T (TReg) cell population is crucial for health, as low TReg cell numbers correlate with autoimmune or inflamma-tory diseases, whereas cancer and immuno suppression have been associated with increased TReg cell numbers. A recent study by Gray, Liston and colleagues sheds light on the dynamic mechanisms that regulate TReg cell homeostasis.

First, the authors observed that the TReg cell compartment is very dynamic in the steady state, with TReg cells proliferating more frequently than conventional Tcells. Second, they used an intricate transgenic mouse system to study the response of the forkhead boxP3 (FOXP3)-expressing TReg cell population to perturbations in the number of TReg cells. The Foxp3 gene is located on each of the two X chromosomes in females, one of which is stochastically inactivated in somatic cells. So, in heterozygous Foxp3THY1.1 Foxp3DTR female mice, 50% of TReg cells express the congenic marker THY1.1 on their cell surface and the rest express the human diphtheria toxin receptor (DTR). Treatment of these mice with diphtheria toxin resulted in the deletion of half of the total TReg cell population, which was followed by a sixfold increase in the number of

THY1.1-positive TReg cells through enhanced proliferation and survival. This homeostatic expansion of the TReg cell population depended on co-stimulatory signalling and increased interleukin-2 (IL-2) expression by conventional T cells.

The homeostatic expansion of the TReg cell population in response to partial TReg cell deletion was followed by a contraction phase, during which the excess of newly generated TReg cells was removed through the intrinsic pathway of apoptosis, which involves BCL-2 antagonist/killer (BAK), BCL-2-associated X protein (BAX) and BCL-2-interacting mediator of cell death (BIM; also known as BCL-2L11). But how is this apoptotic pathway so dynamically regulated during the expansion and the contraction phases?

The deletion of the anti-apoptotic proteins B cell lymphoma2 (BCL-2) or BCL-XL (also known as BCL-2L1) from haematopoietic cells and TReg cells, respectively, did not affect TReg cell homeostasis. By contrast, the constitutive or the induced TReg cell-specific deletion of myeloid cell leukaemia sequence1 (MCL1; also known as BCL-2L3) resulted in an inflammatory pathol-ogy, as a result of a 60% decrease in

the number of peripheral TReg cells. Notably, MCL1 expression was twofold higher in TReg cells than in conventional Tcells in the steady state, and IL-2 induced a further increase in the expression of MCL1 in TReg cells following partial ablation of the TReg cell population. By contrast, BIM which can bind with high affinity to MCL1 and can inhibit its anti-apoptotic function was implicated in the initiation of apoptosis during the contraction phase, as its TReg cell-specific deletion resulted in an increase in the size of the TReg cell population.

Thus, TReg cells are a highly dynamic population, the expansion and the contraction of which is determined by the relative amounts of MCL1 and BIM. Moreover, the authors propose that the unique dependence of TReg cells on an induced cytokine (IL-2), rather than on a constitutively expressed cytokine (as in the case of B cells and conventional T cells), links TReg cell homeostasis with quantitative and functional sufficiency.

Maria Papatriantafyllou

R E G U L ATO RY T C E L L S

Keeping the numbers steady

ORIGINAL RESEARCH PAPER Pierson, W. et al. Antiapoptotic Mcl-1 is critical for the survival and niche-filling capacity of Foxp3+ regulatory T cells. Nature Immunol. http://dx.doi.org/10.1038/ni.2649 (2013)

TReg cells are a highly dynamic population, the expansion and the contraction of which is determined by the relative amounts of MCL1 and BIM

NPG



Nature Reviews Immunology | AOP, published online 26 July 2013; doi:10.1038/nri3514


Transforming growth factor- (TGF) supports the expansion of regulatory T (TReg) cell populations, but a new report now suggests that excessive TGF receptor (TGFR) signalling may lead to dysregulated TReg

cell activity and may underlie a diverse range of allergic diseases in humans.

Most studies investigating TGF have relied on genetically engineered mouse models. However, patients with LoeysDietz syndrome (LDS) have naturally occurring mutations in the genes encoding the TGFR and provide a unique opportunity to

investigate the functions of TGF in humans. Frischmeyer-Guerrerioet al. characterized a group of 58 patients with LDS, who have heterozygous mutations that affect either TGFBR1 or TGFBR2. They found that 31% of these patients had food allergies (compared with 26% of the general population) and that 45% had been diagnosed with asthma, 48% with allergic rhinitis and 38% with eczema. In addition, two thirds of the patients reported gastrointestinal complaints and 10% of patients had been diagnosed with eosinophilic gastrointestinal diseases, which have a prevalence of approximately 0.05% in the general population.

Consistent with the allergic phenotypes seen in the patients with LDS, these patients had significantly increased eosinophil counts, elevated levels of IgE and higher levels of the Thelper 2-type cytokines interleukin-5 (IL-5), IL-13 and CC-chemokine ligand 2. By contrast, their overall white blood cell counts and their serum levels of non-IgE antibody isotypes and 21 other cytokines were normal.

To assess whether these disease phenotypes arose from defects in immune tolerance, the authors exam-ined TReg cells from the patients with LDS. Forkhead box P3 (FOXP3)+ TReg cells from patients with LDS showed no functional defect in

invitro suppression assays. However, compared with non-allergic control individuals, patients with LDS had increased numbers of FOXP3+ TReg cells in the peripheral blood and, surprisingly, these cells expressed IL-13. In response to culture with increasing doses of TGF, naive Tcells from patients with LDS and from control individuals showed a similar upregulation of FOXP3 expression, but FOXP3+IL-13+ T cells only accumulated in the cultures of cells from patients with LDS.

The authors found that Tcells from patients with LDS had increased phosphorylation of the sig-nalling proteins SMAD2 and SMAD3 in response to TGF, which suggests that TGFR signalling is enhanced, rather than repressed, in these individuals. Notably, patients with allergic diseases, but not with LDS, also showed increased frequencies of IL-13+ TReg cells.

These findings suggest that, despite its role in supporting the expansion of TReg cell populations, excessive activation of the TGF sig-nalling pathway may promote allergic disease by skewing TReg cell functions.

Yvonne Bordon

A S T H M A A N D A L L E R GY

TGF too much of a good thing?

ORIGINAL RESEARCH PAPER Frischmeyer-Guerrerio, P. A. et al. TGF receptor mutations impose a strong predisposition for human allergic disease. Sci. Transl. Med. 5, 195ra94 (2013)

excessive activation of the TGF signalling pathway may promote allergic disease by skewing TReg cell functions

GETTY





New research published in Nature Medicine has led to the identification of a mechanism that is responsible for an increasingly recognized food allergy-related disease known as eosinophilic oesophagitis (EoE). Patients with EoE generally show a narrowing of the oesophagus, which leads to difficulty swallowing and food impaction. The study suggests that food allergy-associated inflammation in the oesophagus is driven by increased levels of the cytokine thymic stromal lympho poietin (TSLP) that promote exagger-ated basophil responses.

The recent finding that a gain-of-function polymor-phism in the gene encoding TSLP is associated with the development of EoE led the authors to investigate the role of this cytokine in the disease. They generated a disease model in which mice were first sensitized to a food allergen ovalbumin (OVA) through the skin, using the vitaminD analogue MC903 or tape-stripping to generate a lesion, and were then orally challenged with OVA. Similar to the disease in humans, mice that were sensitized and challenged with OVA developed inflammation and eosinophilia in the oesophagus. This EoE-like disease was associated with increased TSLP expression in the skin and oesophageal tissues, as well as with high levels of mRNA encoding

T helper 2-type cytokines and a basophil-specific protease. Further analysis showed that 30% of mice with the disease had impacted food in the oesophagus at the time of killing, whereas food impaction was never

observed in control mice. To explore the involvement of TSLP in this mouse

model of EoE, the authors first studied

TSLP receptor (TSLPR)-deficient mice. Unlike wild-type mice, Tslpr/ mice did not develop oesophageal eosinophilia follow-

ing sensitization and challenge with OVA.

In further support of a necessary and sufficient

role for TSLP, it was shown that EoE-like disease developed

after oral challenge when wild-type mice were sensitized to OVA in the presence of recombinant TSLP, but not when they were treated with OVA alone. Interestingly, although the EoE-like disease was associated with high levels of OVA-specific IgE, IgE-deficient and IgE-sufficient mice developed equivalent diseases, which suggests that the disease is IgE independent. This is consistent with reports that indicate that IgE-targeted therapies fail to ameliorate EoE in most patients.

TSLP is also known to promote basophil responses, which sug-gests that basophils are involved in EoE. Consistent with this idea, the

depletion of basophils during the sensitization phase, using either diphtheria toxin-based genetic depletion or the basophil-depleting CD200R3-specific antibody, limited the development of EoE-like disease. Moreover, targeting the TSLPbasophil pathway with a neutralizing TSLP-specific antibody or a CD200R3-specific antibody could effectively treat mice with established EoE-like disease, as indicated by decreased oesophageal eosinophilia and improved oeso-phageal function compared with mice treated with a control antibody.

Consistent with the idea of targeting the TSLPbasophil axis to treat human disease, the authors observed increased TSLP expres-sion and basophil frequencies in oesophageal biopsy samples from paediatric and adult patients with active EoE, compared with individuals without EoE or with inactive EoE. In addition, higher basophil frequencies were present in individuals who were homozygous or heterozygous for the risk-associated TSLP polymorphism. Taken together, these data suggest that patients with the TSLP risk allele have a predisposition to TSLP overexpression and the associated basophilia that might increase their likelihood of developing EoE after an encounter with trigger antigens.

Lucy Bird

A S T H M A A N D A L L E R GY

Basophils make food hard to swallow

ORIGINAL RESEARCH PAPER Noti, M. et al. Thymic stromal lymphopoietin-elicited basophil responses promote eosinophilic esophagitis. Nature Med. 19, 10051013 (2013)

targeting the TSLPbasophil pathway ... could effectively treat mice with established EoE-like disease

GETTY/Photodisc Images





Scavenger receptors were first defined by Goldstein and Brown in 1979 (REFS1,2). Scavenger activity was associ-ated with the ability of certain membrane receptors to bind to and to internalize oxidized low-density lipo-protein (oxLDL). Scavenger receptors were thought to recognize specific epitopes generated by oxidation of native LDL, hence enabling the differentiation between unaltered endogenous self molecules and modified self molecules3. Altered lipoproteins challenge normal homeostasis; indeed, oxLDL has been convincingly implicated in the pathogenesis of atherosclerosis47. For this reason, modified lipids and proteins are identified as danger-associated molecular patterns (DAMPs)8,9.

In recent years additional members of the scavenger receptor family have been identified and more has been learned about their properties1014. It is now appreciated that the range of ligands that they recognize is extremely diverse and includes unmodified endogenous proteins and lipoproteins, as well as a number of conserved micro-bial structures, such as bacterial lipopolysaccharide (LPS) and lipoteichoic acid (LTA)1517. To account for this wide range of scavenger receptor ligands, Witztum9 suggested that epitopes generated by peroxidation of endogenous proteins or lipoproteins resemble microbial structures. In view of the expanding number of cognate ligands, the definition of a scavenger receptor has been broad-ened to include not only the recognition of modified self molecules (which are a subset of DAMPs) but also

the recognition of several exogenous (that is, non-self) pathogen-associated molecular patterns (PAMPs). As such, scavenger receptors are considered to be a subclass of the membrane-bound pattern recognition receptors (PRRs)15,1820.

The scavenger receptors are structurally very hetero-geneous. They are subdivided into classes and, although members of each class share structural features, there is little or no homology among classes (FIG.1). The amalga-mation of the scavenger receptors into a superfamily is mostly due to their shared functional properties. Overall, scavenger receptors identify and remove unwanted enti-ties, through the recognition of modified self molecules (for example, apoptotic cells, mineral-laden debris or damaged proteins) or through the recognition of non-self molecules (for example, microorganisms or foreign particles)16,2027. Removal is often carried out by simple endocytosis but might entail more complex processes, such as macropinocytosis or phagocytosis, which both require elaborate signal transduction. Other emerging roles of these multifunctional receptors include cellular adhesion2830 and antigen presentation31.

In light of their functional versatility and their selectiv-ity for a wide range of ligands (FIG.2; see Supplementary information S1,S2 (table, figure)), it is not surprising that scavenger receptors are involved in both the maintenance of homeostasis and in the pathogenesis of various dis-eases. Similarly to other PRRs15, scavenger receptors have

1Cell Biology Program, Hospital for Sick Children, Toronto, Ontario M5G 1X8, Canada.2Department of Biochemistry, University of Toronto, 555 University Avenue, Toronto, Ontario M5G 1X8, Canada.3Keenan Research Centre, Li Ka Shing Knowledge Institute, St. Michaels Hospital, 30 Bond Street, Toronto, Ontario M5B 1W8, Canada.*These authors contributed equally to this work.Correspondence to S.G.e-mail: [email protected]:10.1038/nri3515Published online 9 August 2013

Scavenger receptors in homeostasis and immunityJohnathan Canton1*, Dante Neculai1* and Sergio Grinstein1,2,3

Abstract | Scavenger receptors were originally identified by their ability to recognize and to remove modified lipoproteins; however, it is now appreciated that they carry out a striking range of functions, including pathogen clearance, lipid transport, the transport of cargo within the cell and even functioning as taste receptors. The large repertoire of ligands recognized by scavenger receptors and their broad range of functions are not only due to the wide range of receptors that constitute this family but also to their ability to partner with various coreceptors. The ability of individual scavenger receptors to associate with different coreceptors makes their responsiveness extremely versatile. This Review highlights recent insights into the structural features that determine the function of scavenger receptors and the emerging role that these receptors have in immune responses, notably in macrophage polarization and in the pathogenesis of diseases such as atherosclerosis and Alzheimers disease.

REVIEWS

NATURE REVIEWS | IMMUNOLOGY VOLUME 13 | SEPTEMBER 2013 | 621


Class A

Class B

Class D

Class E

Class F

Class G

Class H

Class I

Nature Reviews | Immunology

Scavenger receptor proteins

SR-A1

SR-B1

SREC1

SREC2

SR-PSOX

FEEL1

FEEL2

CD163

CD163L1

CD5

CD6

LIMP2

CD36

CD68

LOX1

MARCO

CSR1

SRCL

SCARA5

SRCR

Domain

LINK

EGF

EGF-like

EGF-laminin

Mucin

Transmembrane domain

FAS1

Collagen

Lectin

CD36

LAMP

CXC-chemokine

Coiled-coil

Danger-associated molecular patterns(DAMPs). Molecules that are released in association with tissue damage or injury; they promote inflammation and tissue repair by triggering pattern-recognition receptors. DAMPs can be released from the degraded stroma (for example, hyaluronan), from the cell nucleus (for example, high-mobility group box 1 protein) and from the cell cytosol (for example, ATP, uric acid, S100 molecules and heat-shock proteins).

Pathogen-associated molecular patterns(PAMPs). Conserved microbial structures that are recognized by innate receptors, including Toll-like receptors.

Pattern recognition receptors(PRRs). Host receptors (such as Toll-like receptors) that are able to sense pathogen-associated molecular patterns and to initiate signalling cascades (often involving the activation of nuclear factor-B) that lead to an innate immune response.

a central role in innate immunity, and their promiscuous affinity for modified lipids and pathogens might be the link between altered metabolism and inflammation20,3237. These recent findings and the rapid, continuing growth in the identification of members of the scavenger recep-tor family, provided the motivation for this Review. In this Review we restrict the discussion to the mammalian scav-enger receptors (for invertebrate scavenger receptors the reader is referred to other reviews3840).

Structural features of scavenger receptorsScavenger receptor classes. On the basis of sequence align-ments and protein domain architecture, Krieger18 proposed in 1997 that scavenger receptors should be subdivided into

six classes, designated A to F18. However, because func-tional considerations such as the types of modified LDL that were recognized by the receptors consider-ably influenced this classification, the resulting groups often include a range of structural determinants. Thus, as depicted in FIG.1, classA scavenger receptors contain a collagen domain, and might also have a type A scavenger receptor cysteine-rich (SRCR) domain or a C-type lectin (CLEC) domain; classB scavenger receptors contain a CD36 domain; classD scavenger receptors contain mucin-like and lysosome-associated membrane glycoprotein (LAMP) domains41; class E scavenger receptors only have a CLEC domain; and class F scavenger receptors are rich in epidermal growth factor (EGF) and EGF-like domains.

Figure 1 | Domain architecture of scavenger receptors. Mammalian membraneassociated scavenger receptors are multidomain proteins that are separated into eight classes. Scavenger receptors show more than 14 different characteristic protein domains that are identified in the figure inset. The combinations and permutations of domains give rise to a considerable diversity among classes. Note that although the majority of scavenger receptors are singlespan membrane proteins, members of class B (including CD36, SRB1 and lysosomal integral membrane protein 2 (LIMP2)) have two transmembrane domains. Despite the diversity in protein domain architecture it is striking that, with the exception of scavenger receptor expressed by endothelial cells 1 (SREC1; (also known as SCARF1) and SREC2, all other groups have very short cytoplasmic tails (shown in red). Furthermore, the cytoplasmic tails do not possess any identifiable protein domains or motifs. For descriptions of domain abbreviations and functions, see the SMART website. The class C scavenger receptor is not listed as it is only present in Drosophila melanogaster. CLEC, Ctype lectin; CSR1, cellular stress response protein (also known as SCARA3); EGF, epidermal growth factor; EGFlaminin, laminintype EGFlike; FAS1, fasciclin 1; FEEL1, fasciclin EGFlike laminintype EGFlike and link domaincontaining scavenger receptor 1 (also known as stabilin 1 and CLEVER1); LAMP, lysosomeassociated membrane glycoprotein; LOX1, lectinlike oxidized LDL receptor 1; MARCO, macrophage receptor with collagenous structure (also known as SCARA2 and SRA2); SRPSOX, scavenger receptor for phosphatidylserine and oxidized lowdensity lipoprotein (also known as CXCL16); SCARA5, scavenger receptor class A member 5; SRCL, scavenger receptor with Ctype lectin (also known as SCARA4 and CLP1); SRCR, scavenger receptor cysteinerich domain.

R E V I E W S

622 | SEPTEMBER 2013 | VOLUME 13 www.nature.com/reviews/immunol



Self ligandsLipoproteinsNative proteinsModied proteins LipidsDead cells or debris

Microbial ligandsGram-positive bacteriaGram-negative bacteria

Self ligandsNative proteinsModied proteins

Self ligandsNative proteinsModied proteinsLipidsDead cells or debris

Microbial ligandsGram-positive bacteriaGram-negative bacteriaVirusesFungi

Microbial ligandsGram-positive bacteriaGram-negative bacteria

Cell membrane

CD36 LOX1SREC1

The subsequent realization that scavenger receptors also participate in pathogen binding and clearance made it necessary to revise and to expand the original classi-fication. As a result, proteins such as CD163 that lack the ability to bind to modified LDL are now classified as scavenger receptors14,42. In 2005, two additional classes, G and H, were added to the scavenger receptor family to accommodate the new members43. The only recep-tor in class G has a CXC-chemokine domain44. Class H scavenger receptors, like class F receptors, have multiple EGF and EGF-like domains, but they can also have fasci-clin 1 (FAS1) and LINK (hyaluronan-binding) domains. Moreover, recent publications suggest the existence of three additional classes of scavenger receptors, which are characterized by hepatitis A virus cellular receptor1 (HAVCR1; also known as KIM1 and TIM1)13, the P2X purinoceptor 7 (REF.12) and CD163 (REFS14,42) (together with CD6 (REF.45) and CD5 (REF.46)). Unlike the well-established scavenger receptors, which are abundant in myeloid cells, HAVCR1 is highly expressed in the proxi-mal tubular epithelium, particularly in response to kid-ney injury47. The ectodomain of HAVCR1, which belongs to the immunoglobulin superfamily13, binds to and medi-ates the internalization of oxLDL47. P2X7, which has been recognized as a purinoceptor for a long time, was recently described to also function as a phagocytic receptor, facili-tating the uptake of non-opsonized particles and bacte-ria48,49, as well as apoptotic cells12. So far, no consensus has been reached as to whether CD5, CD6, CD163, HAVCR1 and P2X7 merit inclusion in the scavenger receptor fam-ily. If these receptors were to eventually be included, their unique structural features would require the creation of three novel classes of scavenger receptor, potentially designated I, J and K. ClassI receptors would contain a type B SRCR domain, class J receptors would contain a mucin-like and an immunoglobulin domain, and classK receptors would contain a purinergic receptordomain.

Structure determines function of scavenger receptors. The domain architecture of scavenger receptors raises two puzzling questions. Firstly, how is the remarkable functional overlap of the different types of scavenger

receptors (see Supplementary information S1,S2 (table, figure)) achieved, despite their lack of structural com-monality? Secondly, what confers scavenger properties to these receptors, considering that most of their con-stituent domains are not unique but are in fact shared by a multitude of other proteins with divergent activities? For instance, EGF domains can be found in 488 dif-ferent human proteins and CLEC domains in 169 oth-ers, but only a handful of these proteins have scavenger properties. In all likelihood, subtle differences in the sequence of each domain and in their arrangement in the three-dimensional structure of the protein (and pos-sibly in multimolecular complexes) will determine their functional selectivity. Detailed structural information will clearly be required to unravel the basis of scavenger receptor selectivity and function.

Information regarding scavenger receptor structure is currently fairly scant. To our knowledge no single scavenger receptor has been fully characterized and only the X-ray or nuclear magnetic resonance (NMR) struc-tures of a few isolated domains from selected receptors have been obtained. The domains that have been charac-terized include type A and typeB SRCRs50,51, CLEC 5254, EGF, lysosome membrane protein 2 (LIMP2), LAMP55, FAS1, LINK and P2X4 domains. Nevertheless a pattern is beginning to emerge. FIGURE3 shows both cartoon and surface representations of the ligand-binding domains of macrophage receptor with collagenous structure (MARCO; also known as SCARA2 and SR-A2)50 and lectin-like oxidized lectin-like oxidized LDL receptor1 (LOX1; also known as OLR1 and SCARE1)52, which highlights their electrostatic potential. Although struc-turally unrelated, the surfaces that are engaged in ligand binding share a high degree of similarity in terms of shape and charge distribution, displaying clusters of cationic residues that are generally centrally located, bounded by anionic patches. This singular electro-static profile might explain the preference of scavenger receptors for polyanionic ligands, which accounts for the functional overlap of ostensibly dissimilar domains. Accordingly, mutating the arginine residues that form the cationic patch on the surface of the SRCR domain of

Figure 2 | Scavenger receptors and their ligands: functional overlap. Diagrammatic representation of the binding specificity of scavenger receptors for self or alteredself ligands (yellow boxes), and for nonself ligands (green boxes). This figure simplifies the information listed in Supplementary information S1,S2 (table, figure). The figure highlights the broad ligand specificity and the functional overlap of three representative scavenger receptors: CD36, scavenger receptor expressed by endothelial cells 1 (SREC1) and lectinlike oxidized LDL receptor 1 (LOX1).

R E V I E W S




SRCR domain (MARCO) C-type lectin domain (LOX1)

MARCO (R431A, R433A, R466A or R468A) impaired the ability of this protein to bind to acetylated LDL (acLDL)50. Similarly, mutations that reduced the posi-tive charge on the surface of LOX1 (R208N, R229N or R248N) inhibited acLDL binding and uptake52, and the residues K164 and K166 were shown to be important for oxLDL binding by CD36 (also known as platelet glycoprotein4)56. These residues are predicted to be part of a cationic patch on the surface of CD36, as deduced by structural homology modelling (D.N., S.G., R. Collins, S. Dhe-Paganon, P. Loppnau, J. Peters, J. C. Pizarro, J. Plumb, M. Ravichandran, P. Saftig, M. Schwake, A. Seitova, W. S. Trimble and F. Zunke, unpublished observations). Extending this hypothesis further, we suggest that a set of conserved arginine resi-dues in the chemokine domain of scavenger receptor for phosphatidylserine and oxidized low-density lipo-protein (SR-PSOX; also known as CXCL16 a class G scavenger receptor) are exposed on its ligand-binding surface. This comes from the observation that charge-neutralizing mutations of these residues (R59A, R67A and R73A) preclude the binding of oxLDL and bacteria to SR-PSOX57.

What seem to be paradoxical observations in the field can also be explained when considered in the context of this electrostatic patch model. A striking example is provided by the SRCR domain, which was shown to mediate binding of bacteria, LPS and modified LDL by MARCO50,58,59 (FIG.3). By contrast, the related SRCR domain of SR-A1 (also known as SCARA1 and MSR1) is not involved in ligand recognition, but instead medi-ates interactions with other membrane proteins. In

SR-A1 it is the collagen domain that is responsible for ligand binding. This conundrum can be resolved when comparing the electrostatic map of the SRCR domain of MARCO with the homology model inferred for the SRCR of SR-A1. The positive arginine patch that is present on the surface of MARCO (FIG.3) is absent in the case of SR-A1 (REF.50).

The electrostatic patch model helps to explain the preference of scavenger receptors for polyanionic ligands; however, the precise structural determinants of the ligands themselves are less clear. This is prob-ably due to the large scavenger receptor ligand reper-toire and the scarcity of structural information about ligandreceptor complexes. One exception is the oxLDLCD36 interaction. Oxidized lipids (which are a major constituent in oxLDL) are the moieties that are recognized by CD36 (REFS6064). Oxidation of the acyl chain of phosphatidylcholine generates a terminal -hydroxy-,-unsaturated carbonyl group that adopts a unique conformation, protruding into the aqueous phase where it becomes accessible to the receptor27,65,66. Phosphatidylserine has been reported to become oxi-dized in a similar way, functioning as an effective ligand for CD36 on the surface of apoptotic cells67.

Two other structural features of the scavenger receptor family deserve mentioning. Firstly, with few exceptions (for example, scavenger receptor expressed by endothelial cells 1 (SREC1; also known as SCARF1) and SREC2 (also known SCARF2,)10, the scavenger receptors have only short cytosolic tails that lack dis-cernible signalling motifs. This feature is discussed in more detail below in the context of scavenger receptor function. Secondly, the propensity of scavenger recep-tors to oligomerize is also noteworthy43,50,52,58,6870. This increases the avidity of binding, thus oligomerization of scavenger receptors might favour the binding of large, multivalent ligands such as modified lipoproteins and bacteria50,69.

Functional features of scavenger receptorsScavenger receptors have been attributed an impressively broad range of functions and are thought to be involved in complex events such as phagocytosis, antigen pres-entation and the clearance of apoptotic cells. Therefore, it is not surprising that scavenger receptors have been shown to activate a range of diverse signalling pathways.

The exact mechanisms by which scavenger recep-tors convey signals following ligand binding remain unclear, not least because few if any of the recep-tors have discernible signalling motifs or domains. A typical case is that of CD36. Similarly to other class B scavenger receptors, CD36 has two transmembrane domains and both its amino terminus and its car-boxyl terminus are cytoplasmic. As the N terminus is particularly short (only seven residues in length), the C-terminal tail is thought to be the site of signal transduction71. Indeed, this region has been shown to associate with SRC family kinases, including FYN, YES and LYN7174. Of note, the C-terminal region of CD36 contains a CXCX5K motif, which is also found in the cytosolic tails of the Tcell co-receptors CD4

Figure 3 | Structural features of the ligand-binding site of scavenger receptors. Scavenger receptor domains of known structure are compared (the same symbols are used as in FIG.1). Shown are the cartoon representation (middle) and the electrostatic potential (bottom) of the putative ligandbinding surface of the dimeric scavenger receptor cysteinerich (SRCR) domain of macrophage receptor with collagenous structure (MARCO) and the Ctype lectin (CLEC) domain of lectinlike oxidized LDL receptor 1 (LOX1). The red patches indicate the regions of most negative electrostatic potential, whereas the blue patches show the regions of most positive electrostatic potential. Notice the shape and the charge similarity of the receptors shown, despite their differences in primary sequence.

R E V I E W S



oxLDL

RAC

LYN SYK MYD88 FYN

TSP1

IRAK TRAF6

p38 JNK

LYN

FYN

Paxillin

PYK2

p130CAS

IKKMyosin II

p44 or p42

PPAR

9-HODE and13-HODE

VAV1FAK1GTP


CD36 ?O2

Fibrillar -amyloid

O2

TLR2TLR6

Staphylococcusaureus

NADPHoxidase

Actinpolymerization

Pro-inammatorycytokines

Pro-apoptoticresponse

Gene transcription PPAR CD36

Cellpolarity

CD47

6

1

p50 p65

NF-B

and CD8, and which functions as a docking site for SRC family kinases75,76. However, it was convincingly shown that in CD36 the CXCX5K motif is not a major docking site for FYN and LYN71. Therefore, to the best of our knowledge, the exact nature of the interaction between CD36 one of the most extensively studied scavenger receptors and the SRC family kinases remains unclear. The signalling function of CD36 has also been linked to the activation of mitogen-activated protein kinases (MAPKs). The specific MAPKs that are engaged by CD36 vary depending on the cellular context and on the nature of the ligand; for example, in cells derived from the vascular endothelium, the p38 MAPKs are activated by CD36 following bind-ing of thrombo spondin 1 (REF.75); MAPK/ERK kinase kinase 2 (MEKK2; also known as MAP3K2), Jun N-terminal kinase 1 (JNK1; also known as MAPK8) and JNK2 (also known as MAPK9) are activated in macrophages in response to oxLDL71; and MAPK p44 and p42 are activated in response to -amyloid binding in both microglia and macrophages77,78 (FIG.4).

The failure to identify bonafide signalling domains and the context-dependent variability of the down-stream effectors activated by CD36 can both be recon-ciled by a single model. It seems probable that CD36 and in all probability most scavenger receptors function as components of heteromultimeric signal-ling complexes known as signalosomes (FIG.4). Indeed, CD36 has been shown to form complexes not only with SRC family kinases but also with a striking range of transmembrane proteins that include Toll-like recep-tor 2 (TLR2), TLR4 and TLR6, 1 integrin, 2 integrin, 5 integrin and tetraspanins, such as CD9 and CD81 (REFS7982). The promiscuity that has been reported for CD36 might be typical of the entire scavenger receptor family. We suggest that at least some of these associ-ated proteins function as co-receptors, which renders the scavenger receptors necessary but not sufficient to initiate signal transduction. It is currently unclear if the association of scavenger receptors with the ancillary molecules is constitutive and stable, or whether this occurs only in response to exogenous ligands. The idea

Figure 4 | Scavenger receptors engage multiple intracellular signalling pathways. Scavenger receptor signalling can result in very different outcomes depending on the ligand that is engaged and the cellular context. This is exemplified by CD36, which has been studied in some detail. CD36 can form complexes with integrins (for example, 61 and other 1 and 2 integrins), Tolllike receptors (TLRs) and other molecules, including the tetraspanins CD9 and CD81. The presence of specific ligands probably determines the nature of the complex formed. In most instances, the engagement of CD36 causes the activation of SRC family tyrosine kinases, such as FYN and/or LYN. Following oxidized lowdensity lipoprotein oxLDL binding, prolonged activation of focal adhesion kinase 1 (FAK1), together with the VAV1mediated activation of RAC and the inhibition of nonmuscle myosin II, result in actin polymerization, increased cell spreading and loss of cell polarity. RAC also stimulates the NADPH oxidase. Activating ligands for peroxisome proliferatoractivated receptor (PPAR), such as 9hydroxyoctadecadienoic acid (9HODE) and 13HODE, are also delivered to the cell following oxLDL binding to CD36, which results in the stimulation of PPAR, increasing the expression CD36. In response to other ligands, including amyloid and thrombospondin 1, CD36 activates mitogenactivated protein kinase (MAPK) family serine/threonine kinases, such as p44, p42, p38, JunN-terminal kinase (JNK) and the tyrosine kinase proline-rich tyrosine kinase 2 (PYK2), and recruits the adaptor proteins p130CAS (also known as BCAR1) and paxillin. These ligands induce actin rearrangement and stimulate the production of proinflammatory cytokines and of proapoptotic signals. CD36 can also partner with TLR complexes in response to pathogen ligands, which signal the production of proinflammatory cytokines through a myeloid differentiation primaryresponse protein 88 (MYD88) and nuclear factorB (NFB)dependent pathway. The question mark indicates an as yet uncharacterised coreceptor that has been proposed to cooperate with CD36 to mediate oxLDL binding. IKK, IB kinase; IRAK, IL1 receptorassociated kinase; SYK, spleen tyrosine kinase; TRAF6, TNF receptorassociated factor 6; TSP1, testisspecific protein 1.

R E V I E W S




Macrophage

M1 macrophage

Normalfunction

Pathologicalfunction

TLR2TLR6

TLR4TLR6

CD163

IL-12 and TNF

CD36

MERTK

SR-A1Staphylococcusaureus

M1 polarization

Pro-inammatory response and pathogen clearance

Neuron deathand damage

Anti-inammatory response and wound healing

Alzheimers disease

Foam cellformation

M2 polarization

M2 macrophage

Haemoglobinhaptoglobin complex

TWEAK sequestration

IL-10Apoptotic cell

Neuron

Sterile inammation

-amyloid

IL-1, NO and ROS

Atherosclerosis

Foam cell formationLipids

oxLDL

SR-A1

that the association only forms in response to exogenous ligands would confer flexibility to the system, allowing cells endowed with a finite number of scavenger recep-tors to tune and to maximize their responses to a range of ligands.

A particular receptor may form various types of com-plexes with different co-receptors, not only in different cell types but also in a single cell type. This is best exem-plified by the class A scavenger receptor SR-A1, which

partners with tyrosine protein kinase MER (MERTK) to form a functional complex that enables apoptotic cell uptake83 (FIG.5). The association with SR-A1 was shown to be essential for optimal phosphorylation of MERTK and for the subsequent signalling events such as phospholipase C2 phosphorylation and activa-tion that are required for apoptotic cell clearance83. On the other hand, SR-A1 interacts with TLR4 in the presence of LPS84. In macrophages, this association is

Figure 5 | Scavenger receptors contribute to the functional phenotype of polarized macrophages. Macrophages can polarize into M1 (also known as classically activated) and M2 (also known as alternatively activated) macrophages that have distinct functional phenotypes. The expression of several scavenger receptors, including SRA1 and CD163, is increased in M2 macrophages. The increased expression of SRA1 and CD163 contributes to the prototypical M2 functions: apoptotic cell clearance, sequestration of the inflammatory cytokine TNFrelated weak inducer of apoptosis (TWEAK), clearance of haemoglobinhaptoglobin complexes at sites of tissue damage and the subsequent production of antiinflammatory cytokines. By contrast, the expression of CD36 in M1 macrophages contributes to their characteristic phenotype by complexing with Tolllike receptors (TLRs) to potentiate the production of inflammatory cytokines. The differential expression of scavenger receptors in polarized cells contributes to various pathologies, including Alzheimers disease and atherosclerosis. The increased expression of CD36 and SRA1 on M2 macrophages can result in the accelerated uptake of modified lowdensity lipoprotein (LDL) and in the intracellular accumulation of cholesterol, thus contributing to the formation of foam cells. Conversely, engagement of CD36TLR4TLR6 receptor complexes in M1 macrophages (or microglia) results in sterile inflammation and consequent damage to local tissues at sites of amyloid accumulation. IL, interleukin; MERTK, tyrosine protein kinase MER; NO, nitric oxide; ROS, reactive oxygen species; TNF, tumour necrosis factor.

R E V I E W S



Immunoreceptor tyrosine- based activation motif(ITAM). A structural motif containing a tyrosine residue that is found in the cytoplasmic tails of several signalling molecules. The consensus sequence consists of TyrXXLeu or TyrXXIle. The tyrosine is a target for phosphorylation by SRC tyrosine kinases and for the subsequent binding of proteins containing SRC homology 2 domains.

required for efficient activation of the nuclear factor-B (NF-B) pathway by LPS84 (see REF.85 for a conflicting perspective). Strikingly, the engagement of SR-A1 can therefore produce either a pro- or anti-inflammatory response depending on the nature of the co-receptor. Asimilar dichotomous behaviour has been described for CD36. This receptor induces inflammatory reactions in response to LTA or diacylated lipoproteins when in a complex with the TLR2TLR6 heterodimer, and also in response to oxLDL or fibrillar -amyloid when in a complex with the TLR4TLR6 heterodimer79,86 (FIG.5). By contrast, the CD36-mediated internalization of Plasmodium falciparum-infected erythrocytes does not induce the production of pro-inflammatory cytokines87 and this is also likely to be the case for CD36-mediated ingestion of apoptotic bodies. Although this behav-iour has been shown for the class A and B receptors, it remains to be determined whether this is a general feature of the scavenger receptorfamily.

Scavenger receptors also rely on the formation of multimolecular complexes to achieve their ligand-internalization function82. It was recently shown that CD36 is bridged to the immunoreceptor tyrosine-based activation motif (ITAM)-containing high-affinity immunoglobulin- receptor subunit- (FcR) by a complex consisting of 1 integrins and/or 2 integrins, CD9 and CD81 (REF.82). By incorporating FcR this multimolecular signalling complex can engage spleen tyrosine kinase (SYK), which possesses tandem SRC homology 2 (SH2) domains ideally spaced to engage the phosphorylated tyrosines of the ITAM motif, thereby mediating the internalization of CD36-bound ligands. Importantly, the ability to internalize ligands is not limited to CD36 but extends to other scav-enger receptors, including SR-A1 and MARCO. Internalization can alter the mode of signalling or terminate it, and can also have metabolic functions, for instance by delivering modified lipoproteins to lysosomes. As in the case of CD36, receptors lacking identifiable endocytosis determinants might depend on their association with ancillary signalosome molecules for their internalization.

A scavenger receptor as a JackofalltradesDespite their name, scavenger receptors are involved in more than just scavenging. They have been shown to carry out several functions, including function-ing as lipid transporters, as chaperones that transport other cellular proteins to their destination and even as chemokines44,57,88. Various types of lipids have been reported to be transported by scavenger receptors: cholesterol esters are delivered to steroidogenic tissues and to liver cells by SR-B1, which is a non-endocytic high-density lipoprotein (HDL) receptor8993, whereas fatty acids are taken up by a variety of cells via CD36 (REFS36,93,94). In both instances, lipid transfer may occur via a hollow section or tunnel connecting the ligand-binding surface of class B scavenger receptors to the exofacial leaflet of the membrane bilayer. This tunnel, recently uncovered by crystallographic deter-minations of the structure of these receptors, might

be equivalent to the fatty acid-binding pocket that was previously proposed to exist on the exofacial domain of CD36 (REF.36). Interestingly, the same pocket or tun-nel may have a role in the gustatory perception of fatty acids. CD36, which is abundant in the lingual papillae, has been implicated in the ability to taste fats; indeed, individuals carrying the single nucleotide polymor-phism rs1761667 G allele, which is a common CD36 variant, show greater oral sensitivity to fat than individ-uals carrying the A allele, which causes lower expression of CD36 (REFS9597). Thus the same protein might be responsible for promoting excessive lipid ingestion, for clearing the modified species that are generated when lipoproteins circulate in excess and for the formation of foam cells and atherosclerotic plaques (see below).

CD36 has also been implicated in the formation of cytokine-induced multinucleated giant cells98. Multinucleated giant cells are present in granuloma-tous conditions such as tuberculosis and the for-eign-body reaction to implanted materials, in which they restrict intercellular spreading of mycobacteria and might be involved in implant rejection, respec-tively99,100. Although the detailed mechanism respon-sible for these effects remains unclear, it has been suggested that multinucleated cells arise from the inter-action of CD36 with phosphatidylserine on the surface of neighbouring cells98.

Several scavenger receptors, particularly those of classB, have well-documented roles as chaperones. LIMP2, which is a member of the class B scavenger recep-tors, is essential for the delivery of -glucocerebrosidase from the endoplasmic reticulum (ER) to the lyso-somes101. Mutations that impair the association between LIMP2 and its cargo cause several neurodegenerative and renal diseases, such as myoclonic epilepsy and nephrotic syndrome101103. Similarly, the class G scav-enger receptor fasciclin EGF-like laminin-type EGF-like and link domain-containing scavenger receptor 1 (FEEL1; also known as stabilin 1) has been implicated in the intracellular sorting and lysosomal delivery of chitinase-like protein11; macrophages release FEEL1 by lysosomal secretion, thereby affecting inflammation and regulating apoptosis.

Certain scavenger receptors are susceptible to cleav-age by exofacial proteases, which results in the shed-ding of soluble products to the circulation. Soluble forms of SR-PSOX and of the classI receptors CD163, CD5 and CD6 have been detected in the plasma46,104,105. Remarkably, the proteolytic fragments released from the membrane carry out functions that markedly dif-fer from those of the precursor receptor. For instance, the soluble form of SR-PSOX is an interferon-regulated chemokine that stimulates CXC-chemokine receptor6 (CXCR6), which is expressed by activated Tcells and natural killer T cells88,106. CD163, which functions as an endocytic receptor for haptoglobinhaemoglobin com-plexes in its membrane-associated form107109, is also a substrate of proteases14. Its soluble extracellular domain retains the ability to associate with iron and can thereby inhibit the growth of bacterial pathogens. Moreover, soluble CD163, as well as fragments released from CD5

R E V I E W S



and CD6, are elevated in inflammation and in autoim-mune disease. Even though their specific function is unknown, the soluble forms of classI receptors have been suggested to be potentially useful biomarkers for various clinical conditions104,107,110.

SR-A1 was recently shown to prevent calcification of the vasculature and soft tissue. The formation of proteinmineral complexes, referred to as calciprotein particles, is a physiological mechanism that facilitates the clear-ance of calcium phosphate nanocrystals from the extra-cellular milieu in order to prevent their deposition and potentially pathological calcification. SR-A1-deficient macrophages have an impaired ability to bind and to internalize calciprotein particles26. Moreover, prolonged exposure of macrophages to calciprotein particles results in significant upregulation of SR-A1 (REF.111). Taken together, these observations suggest a key role for this receptor in calciprotein particle clearance.

Scavenger receptors, specifically SR-A1 and MARCO, also have a role in the maintenance of the microarchitec-ture and functionality of the marginal zone of the spleen. The depletion of these receptors results in aberrant distribution of splenic macrophages112, which, in turn, are required for Bcell retention in the marginal zone113.

In summary, scavenger receptors have important physiological roles inside cells, on their surface and in the circulation. This range of disparate functions empha-sizes the rather arbitrary consolidation of the scavenger receptors into a singlefamily.

Receptors and macrophage polarizationIn the physiological setting, macrophages respond to environmental stimuli, such as TLR agonists and sig-nals from activated lymphocytes, by assuming distinct functional phenotypes. It is generally accepted that there is a great deal of plasticity between their phenotypes and, depending on the combination of stimuli that they receive, macrophages can exist in various shades of acti-vation114. That said, a useful paradigm for understanding macrophage polarization has been to study the extremes of the activation range: that is, classically activated macrophages and alternatively activated macrophages (referred to as M1 and M2 macrophages, respectively). M1 macrophages are generally characterized as having an interleukin-12 (IL-12)hi IL-23hiIL-10low phenotype and are efficient producers of reactive oxygen species, nitro-gen intermediates and inflammatory cytokines, such as tumour necrosis factor- (TNF) and IL-6 (REF.115). M1 macrophages are considered to be essential participants in T helper 1 (TH1) cell responses and to have a potent microbicidal and tumoricidal capacity116. Conversely, M2 macrophages have an IL-12lowIL-23lowIL-10hi phe-notype and a variable capacity to produce pro-inflam-matory cytokines115. M2 macrophages are considered to have a central role in tissue repair and remodelling, in the resolution of inflammation, in apoptotic cell clearance and in the control of extracellular parasites116.

In recent years, increasing attention has been paid to the contribution of scavenger receptors to mac-rophage polarization. The expression of several scav-enger receptors, such as CD163, SR-A1 and CD36 is

markedly increased in M2 macrophages14,117121. Indeed, CD163 is a well-accepted marker of the M2 macrophage pheno type. Not only are some scavenger receptors more highly expressed in M2 cells than in M1 cells but also the presence of some receptors contributes to the polarization programme of these cells. Signals deliv-ered by CD36 and SR-A1 to the ER stress, JNK and peroxisome proliferator-activated receptor- (PPAR) pathways are seemingly necessary for the generation of the M2 phenotype117. The elevated expression of scav-enger receptors is congruent with the function of M2 cells in apoptotic cell clearance and in the suppression of inflammation. For instance, by increasing the sur-face expression of SR-A1, along with its co-receptor MERTK (FIG.5), M2 macrophages are better able to engulf apoptotic bodies83,122124; CD36 also contributes to this function125. Conversely, CD163 is instrumen-tal in promoting an anti-inflammatory phenotype in M2 macrophages. It can sequester and thus inactivate pro-inflammatory molecules such as TNF-related weak inducer of apoptosis (TWEAK)126, and attenuates hae-moglobin-associated damage that is a source of inflam-mation127,128. SR-A1 has similar anti-inflammatory effects in macrophages129,130 (FIG.5).

The preceding observations have led to the mis-conception that the entire family of scavenger recep-tors is upregulated in M2 macrophages and that scavenger receptors are anti-inflammatory in all cases. Neither of these conclusions is warranted. Although M2-polarizing factors, such as IL-4 and macrophage colony-stimulating factor (M-CSF), increase SR-A1 expression, they concomitantly decrease the expres-sion of another class A scavenger receptor, MARCO130. Conversely, M1-polarizing factors such as LPS and granulocyte/macrophage colony-stimulating factor (GM-CSF) increase the expression of MARCO, but decrease SR-A1 levels130. Moreover, recent studies indi-cate that the differential expression of scavenger recep-tors helps to define the functional phenotype of M1 and M2 macrophages. Accordingly, MARCO positively regulates pro-inflammatory cytokine production, whereas SR-A1 has the opposite effect130.

It is also worth noting that, because they function as part of complex signalling platforms, the context in which scavenger receptors are present is as important as their absolute level of expression. This is well illustrated by CD36: the net amount of this receptor increases in M2 macrophages, which suggests that it has an anti-inflammatory function; however, CD36 is also present in M1 cells in which it can interact with TLRs to produce pro-inflammatory cytokines in response to microbial ligands (FIG.5)86,131,132. Thus, the predominant function of CD36 may be determined by the type and the extent of expression of co-receptors. In this regard, it is relevant that TLR2 and TLR4 are preferentially expressed by M1macrophages133.

In summary, although scavenger receptors are more prominently expressed by M2 macrophages, they are not exclusive to this macrophage population and can con-tribute to pro-inflammatory macrophage responses in certain contexts.

R E V I E W S



Scavenger receptors and innate immunityIt is now abundantly clear that, in addition to scaveng-ing modified lipoproteins, many of the scavenger recep-tors have the ability to recognize conserved PAMPs on microbial surfaces. The role that scavenger receptors have in innate immunity, including the phagocytosis and the clearance of various microbial species, has been extensively reviewed in recent years16,17,134; therefore, in this section we will focus on some of the more recent advances concerning ligand specificity, on the interplay between scavenger receptors and other PRRs, and on the subversion of scavenger receptor function by pathogens.

As new information accumulates, the range of ligands recognized by scavenger receptors is becom-ing apparent. SR-A1, for example, has been shown to bind to the lipid A moiety of LPS (which is a feature of Gram-positive bacteria), LTA (which is expressed by Gram-negative bacteria) and bacterial CpG DNA134 (FIG.1; Supplementary information S1 (table)). As a result, SR-A1 can mediate the non-opsonic uptake of Neisseria meningitides, Listeria monocytogenes and Staphylococcus aureus24,135138. MARCO shares with SR-A1 the ability to recognize LPS, LTA and CpG DNA, and can also bind N.meningitides16. The shared structural features (FIG.2) and ligand specificities (FIG.1; Supplementary information S1 (table)) of SR-A1 and MARCO seem to suggest that these receptors are func-tionally redundant. However, a recent study showed that SR-A1 and MARCO recognize overlapping but distinct sets of endogenous and microbial ligands, including several N.meningitides surface proteins, which high-lights the distinct specificities of these two related scav-enger receptors139. Such small differences in selectivity might have evolved to increase the repertoire of innate immune recognition139. As we learn more about the specificity of other receptors, this idea may be applied to the entire scavenger receptorfamily.

Another general feature that is shared by several members of the scavenger receptor family is the ability to interact with and to influence signalling through other PRRs. Several recent studies have provided interesting examples of the interplay between scavenger receptors and TLRs. In some instances, a synergistic relationship exists between the two types of PRRs. A recent analy-sis showed that SR-A1 interacts with TLR4 to promote the phagocytosis of the Gram-negative bacterium Escherichia coli, whereas SR-A1 and TLR2 cooperate in the phagocytosis of the Gram-positive bacterium S.aureus140. In addition, SR-A1 potentiates the respon-siveness of PRRs that are located in endo membranes: by mediating pathogen internalization, SR-A1 enhances the inflammatory response mediated by TLR3 (REF.85). The functional cooperation between scavenger recep-tors and TLRs is not unique to SR-A1 and has been shown to occur for several other scavenger receptors. MARCO, for example, partners with TLR2 and CD14 in the recognition of the Mycobacterium tuberculosis glycolipid trehalose 6,6-dymycolate and is required for the optimal production of pro-inflammatory cytokines in response to this bacterial product141. Similarly, as discussed above, the class B scavenger receptor CD36

can form a functional complex with TLR2 and TLR6, which augments cytokine responses to S. aureus-derived LTA and which enhances the internalization of P.falciparum-infected erythrocytes79,86. Thus, a pattern is emerging: inflammatory ligands are recognized by both a scavenger receptor and another sensor PRR, such as a TLR. In this paradigm that was first appreciated by Mukhopadhyay etal.85, the scavenger receptors potenti-ate the function of the sensor PRRs, thereby augmenting the inflammatory response.

As is often the case with innate immune receptors, scavenger receptors can be co-opted by pathogens to function in their infectious cycle. One well-studied example is the subversion of SR-B1 by hepatitis C virus, which uses this scavenger receptor as a co-receptor for entry into host cells. Functional complementation assays and the inhibitory effect of other SR-B1 ligands, such as oxLDL, showed that the lipid transfer activity of the receptor is essential for viral entry142145. In addition, SR-B1 is used by the intracellular pathogen Chlamydia trachomatis for survival in host cells. C.trachomatis, which resides in a membrane-bound intra cellular compartment termed the inclusion, has long been rec-ognized to depend on the acquisition of host-derived factors (including lipids) for survival in its intracellu-lar niche. One mechanism by which it acquires host-derived lipids is through the recruitment of SR-B1 to the inclusion membrane, where the lipid transfer activity of the scavenger receptor mediates the delivery of phos-phatidylcholine to the lumen of the inclusion. This role is crucial to the progression of infection: inhibition of SR-B1-mediated lipid transfer impairs the intracellular replication of C.trachomatis146.

Another class B receptor, LIMP2, has been identi-fied as the cellular receptor for enterovirus 71 (EV71), coxsackievirus 7 (CVA7), CVA14 and CVA16 entry into host cells147,148. In the case of EV71, LIMP2 not only functions as a receptor but also as a determinant of viral uncoating and therefore of infection efficiency149. Intriguingly, CD36 which has been implicated in the clearance of several bacterial and protozoan pathogens is co-opted by mycobacteria150152. The Drosophila melanogaster CD36 homologue Peste has been identi-fied as an important determinant of uptake of myco-bacteria into host cells153. In addition, CD36 deficiency results in reduced susceptibility to mycobacterial infection both invivo and invitro154; the mechanisms whereby CD36 improves mycobacterial survival are yet to be elucidated. The finding that pathogens have evolved mechanisms to subvert scavenger receptor function emphasizes the need for a clearer understand-ing of the roles that scavenger receptors have at the front line of hostpathogen interactions.

Scavenger receptors and diseaseConsidering the number of receptors that constitute the scavenger receptor family and the wide range of functions they carry out, the involvement of scavenger receptors in the pathogenesis of multiple diseases was anticipated. However, the extent and the mechanism of this involve-ment have not yet been fully appreciated because the

R E V I E W S



study of most scavenger receptors is in its infancy. Nevertheless, by participating in the recognition and the internalization of oxLDL23 and -amyloid155,156, and in the transport of fatty acids93, scavenger receptors have been implicated in diseases as diverse as athero-sclerosis4,5,157160, type2 diabetes mellitus94,161,162 and Alzheimers disease34,155,163,164. A brief overview of the involvement of scavenger receptors in these disorders, with a particular focus on CD36, is discussedbelow.

Scavenger receptors in atherosclerosis. Atherosclerosis is a chronic inflammatory disease characterized by a complex interplay between metabolic and immune pro-cesses, which may lead to the formation of vulnerable plaques165,166. The structural disruption of these plaques can cause atherothrombotic vascular disease, which is the most frequent cause of death in the industrialized world166. The pathogenesis of atherosclerosis is not yet fully understood, but a key event in the development of primary atherosclerotic plaques is the inability of mac-rophages to properly process modified lipoproteins, which results in the formation of foam cells. As SR-A1, MARCO, CD36, SR-B1, LOX1 and SR-PSOX can all recognize oxidation-specific epitopes of oxLDL, their role in atherosclerosis has been extensively investigated 4,44,158,167170. It has been unambiguously shown in in vitro studies that these receptors function as a major conduit for intracellular cholesterol accumulation171. However, when assessed invivo using gene-knockout strategies in hyperlipidaemic apolipoprotein E (ApoE)/ mice, the contribution of individual receptors (for example, of SR-A1) to atherosclerosis is much less clear3, prob-ably as a result of functional redundancy. Nevertheless the pro-atherogenic role of CD36 has been convinc-ingly shown172: by coupling to TLR4 and TLR6, CD36 can trigger a sterile inflammatory response, which induces NF-B activation when exposed to modified LDL173,174. Accordingly, genetic deletion of TLR4 or of the TLR signalling adaptor myeloid differentiation primary-response protein 88 (MYD88) attenuates atherosclerosis175,176.

Conversely, oxidized components of oxLDL, such as 9-hydroxyoctadecadienoic acid (9-HODE) and 13-HODE, are potent activators and ligands for PPAR, which is a transcription factor that is important in lipid metabolism177. Following activation, PPAR hetero-dimerizes with the retinoid X receptor and the newly formed complex binds directly to PPAR-response elements178,179. One such response element is found in the CD36 promoter, which causes increased CD36 expression. Thus, oxLDL has synergistic effects that might lead to ER stress180 and foam cell formation, which are early steps in atherogenesis. In addition, by stimulating CD36 on the surface of platelets181, oxLDL increases platelet reactivity and fosters a prothrombotic state182,183, which increases the risk of a cardiovascu-lar episode31,184187. Furthermore, phosphatidylserine exposed on the surface of microparticles released by shedding cells can bind to CD36, which renders the platelets more sensitive to activation and to aggrega-tion188. Microparticles are often generated at sites of

vascular injury and inflammation, which are areas of elevated risk for thrombus formation. The classE scav-enger receptor LOX1 is also expressed on platelets189, but in an activation-dependent manner. The inhibi-tion of LOX1 results in a dose-dependent reduction in agonist-induced platelet aggregation and activation190.

In contrast to CD36, SR-B1 was shown to have not only an anti-atherogenic effect191194 but also to inhibit platelet aggregation and thrombosis195,196. These effects occur in the liver, where SR-B1 mediates the transport of choles-terol from HDL to the hepatocyte89,197. In cholesterol-laden macrophages, HDL is loaded with cholesterol by reverse transport down its concentration gradient. The protec-tive role of SR-B1 is thought to reflect the net discharge of cholesterol from HDL to hepatocytes, which ultimately process the cholesterol for biliary excretion198,199. By indi-rectly removing cholesterol from macrophages and foam cells, SR-B1 reduces atherosclerosis90,200202,183. Accordingly, several recent reports have identified a strong associa-tion of SR-B1 polymorphisms with atherosclerosis and cardiovascular disease203205.

Other scavenger receptors also contribute to ath-erosclerosis: the deletion of SR-PSOX exacerbates atherosclerosis206 and MARCO expression is induced in mouse plaques207. However, at this stage their precise role and mode of action are unclear.

Scavenger receptors in type2 diabetes. Type2 diabetes mellitus is a metabolic disorder characterized by the accumulation of fatty acids and lipid metabolites that lead to alterations in insulin signalling, which causes the development of insulin resistance94,208. Scavenger recep-tors also have a role in this disease. CD36 is known to mediate fatty acid uptake in insulin-sensitive tissues such as adipocytes, skeletal muscle and cardiac mus-cle209211. Pharmacological experiments using trans-port inhibitors, as well as CD36 gene deletion studies, showed that nearly 70% of fatty acids are taken up by the heart via this transporter protein212. In animal models of insulin resistance, the increased rate of fatty acid transport into muscle correlated with an increase in levels of plasmalemmal CD36. Although fatty acid oxidation increases in these muscles, the primary fate of the fatty acids that have been taken up by CD36 is esterification209. The consequent accumulation of lipids is the primary cause of insulin resistance209,210. Similarly, fatty acid transport is markedly increased in skeletal muscles of obese humans and those with type2 diabetes, even though CD36 mRNA and protein are not altered211,213. The transport activity of CD36 could be regulated and such regulation might go awry in obe-sity and diabetes. Along these lines, common CD36 gene variants including the rs3211867, rs3211883, rs3211908 and rs1527483 polymorphisms asso-ciate with measures of obesity 214 and adiposity 215. Nevertheless, the literature regarding the association between common CD36 polymorphisms and insulin sensitivity remains controversial215220. It is also inter-esting that CD36-containing microparticles correlate with the development of diabetes, which potentially provides a biomarker for the disease221.

R E V I E W S



Platelet CD36 also enhances the risk of arterial thrombosis in individuals with diabetes. Advanced gly-cation end-products generated under the chronic hyper-glycaemic conditions associated with diabetes can bind to and activate platelet CD36, which accounts for at least some of the vascular complications associated with dia-betes222. These associations are an explanation for why polymorphisms that affect the expression of CD36 cor-relate with risk of developing thrombosis223. The class E scavenger receptor LOX1, which is expressed on platelets in an activation-dependent manner189, also influences the state of platelet activation190.

Scavenger receptors in Alzheimers disease. Alzheimers disease is characterized by a protracted inflammatory response driven by microglia, which are the central nervous system macrophages. The lesions found in the brains of patients with Alzheimers disease consist of senile plaques that contain -amyloid f

nature reviews immunology - september 2013.pdf-meta

Documents

t cell memory subset

t cell subsets

cd27cd44hi t cell subset

cd27cd44hi t cell population

mln t cell population

role of t cell responses

conventional t cells

mucosal mln t cell subset