SEARCH

SEARCH BY CITATION

Keywords:

  • flagellin;
  • human;
  • inflammatory bowel disease;
  • Toll-like receptor;
  • T regulatory cell

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Acknowledgements
  5. References

Two related chronic inflammatory diseases, Crohn’s disease and ulcerative colitis, are together often referred to as inflammatory bowel disease (IBD). Current treatment options are not curative, and patients face lifelong therapy and debilitation. IBD is thought to be the product of a combination of genetic and environmental factors that result in the abnormal regulation of immune responses. Experimental models have demonstrated that normal CD4+ T-regulatory (Treg) cell responses and commensal bacteria are required for the maintenance of gut immune homeostasis. Recent evidence that CD4+ T cells express Toll-like receptors (TLRs) and respond directly to TLR ligands, suggests that signals from commensal bacteria may directly affect T-cell responses in the gut. In this review, we focus on evidence that defects in Treg cells may underlie IBD in humans. In addition, we discuss evidence that direct signaling via TLRs to T cells can affect IBD and that T-cell-dependent responses to bacterial proteins, such as flagellin, are central to the aetiology of this disease.


Abbreviations:
CBir

C3H/HeJBir

CD

Crohn’s disease

CpG

cytosine-phosphate-guanine

FOXP3

forkhead box P3

IBD

inflammatory bowel disease

IFN-γ

interferon-γ

IL

interleukin

LPS

lipopolysaccharide

TCR

T-cell receptor

TGF-β

transforming growth factor-β

Th

T helper

TLR

Toll-like receptor

Treg

T regulatory

UC

ulcerative colitis

Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Acknowledgements
  5. References

Crohn’s disease (CD) and ulcerative colitis (UC), known collectively as inflammatory bowel disease (IBD), are autoimmune-like disorders characterized by chronic, idiopathic inflammation of the intestinal mucosal tissue, which causes a range of symptoms including abdominal pain, severe diarrhoea, rectal bleeding and wasting.1,2 IBD tends to emerge in late childhood, occurs primarily in immunocompetent individuals and is most prevalent in westernized regions of the world.2,3 CD and UC are distinguished by the tissues affected: CD can affect any region of the gastrointestinal tract in a discontinuous and transmural manner, whereas pathology in UC is restricted to the surface mucosa of the colon, in particular the rectum.1,2 Current treatment regimens, including anti-inflammatory and immunosuppressive agents, are not curative and only reduce the degree of intestinal inflammation associated with disease.1

Although the aetiology of IBD is unclear, both CD and UC are believed to be T-cell-driven processes, with inflammation resulting from inappropriate cytokine production by subsets of CD4+ T-helper (Th) cells.1 Specifically, CD is associated with Th1 and Th17 cytokine profiles, whereas UC is typically associated with Th2 cells.1,2 It has been suggested that IBD results from defects in the T-cell-mediated regulatory processes that would normally prevent and/or terminate inflammatory responses; compelling data from mouse models support this hypothesis (reviewed in ref.4). In this review, we will focus on evidence that such defects in CD4+ CD25+ forkhead box P3 (FOXP3)+ T-regulatory (Treg) cell function may underlie IBD in humans, and discuss evidence that altered T-cell-dependent responses to bacterial proteins may be central to its aetiology.

Role of Treg cells in human IBD

The gut is an immunologically unique organ that must retain the ability to mount an adaptive response to pathogens while maintaining tolerance to dietary antigens and commensal bacteria.2 Much experimental evidence indicates that Treg cells expressing FOXP3 and/or interleukin (IL)-10 have a fundamental role in maintaining gut immune homeostasis.5,6 Pioneering experiments by Powrie et al. demonstrated, more than 15 years ago, that transfer of naive CD4+ T cells into immunodeficient mice in the absence of Treg cells results in colitis that is dependent on the presence of commensal bacteria.7–9 Remarkably, adoptive transfer of Treg cells not only prevents disease, but also, acting through both IL-10 and transforming growth factor--β (TGF-β)-dependent and -independent mechanisms, halts the progression of established disease and reverses pathology.5,10 In addition, Treg cells suppress not only T-cell-dependent colitis, but also intestinal inflammation triggered by bacteria.11 Together, these data provide a reason to believe that methods used to boost and/or restore Treg-cell numbers and function may be effective in the treatment of established IBD in humans.

Many groups have undertaken association studies in patients with IBD to define whether changes in Treg cells can be correlated with disease, with a particular focus on defining differences between circulating cells and gut tissue cells. The vast majority of these studies have analyzed CD4+ CD25+ FOXP3+ Treg cells as they can be tracked and are isolated relatively easily. Early studies found that patients with CD or UC have decreased numbers of peripheral suppressive CD4+ CD45ROCD25+ FOXP3+ Treg cells during active disease, suggesting that the severity of disease is inversely correlated with the frequency of peripheral Treg cells.12–15 In contrast, several studies have found that the lamina propria, mesenteric lymph nodes and intestinal inflamed mucosa of patients with either CD or UC actually contain increased numbers of Treg cells compared with healthy controls.5,13,15,16 These data are consistent with findings in other gut-inflammatory conditions, such as diverticulitis, pseudomembranous colitis and cytomegalovirus-induced colitis, which show a similar correlation between an increase in FOXP3+ Treg-cell numbers and disease severity.13 Combined with similar data from other autoimmune diseases,17,18 these findings support the hypothesis that Treg cells traffic to sites of inflammation in an attempt to restore immune homeostasis. Consistent with this notion, a recent study found that the frequency of FOXP3+ cells in the colorectal mucosa decreased following treatment with granulocyte and monocyte adsorptive apheresis and correlated with clinical response.19

Notably, in vitro functional analysis of Treg cells from the peripheral blood or intestinal mucosal tissue of IBD patients revealed that, despite numerical changes, they maintain normal cell-contact-dependent, cytokine-independent suppressive capacity, even against pathogenic T-effector cells derived from the inflamed mucosa.20 Furthermore, mucosa-derived Treg cells from IBD patients effectively suppress both proliferation and cytokine production, in contrast to Treg cells from rheumatoid arthritis patients that are reported to specifically lack the capacity to suppress cytokine production.16,17 These findings from IBD also appear to contrast with a report describing a clear defect in suppressive capacity of cells from patients with multiple sclerosis, although this deficiency was dependent on the strength of T-cell-receptor (TCR) activation,21 which is known to have a major effect on in vitro suppression assays.22 As studies of Treg cells from human IBD mucosal biopsies have not examined suppressive capacity with varying strengths of T-cell stimulation, it remains possible that subtle functional defects, possibly only in antigen-specific responses, in these cells have been overlooked.

Problems with assessing Treg-cell numbers and function in human inflammatory diseases

Considering the extensive evidence from mouse models, and data from other human autoimmune diseases, it is somewhat surprising that no defect in the numbers or in vitro function of tissue-derived Treg cells from IBD patients has been reported. It is important to note, however, that the in vitro assays of suppressive function of such cells are performed outside their inflamed environment of origin, and, in the absence of the in vivo cytokine milieu, may not accurately reflect their in vivo activity. An additional consideration is that in humans, in contrast to mice, FOXP3 is expressed by non-suppressive activated T-effector cells following stimulation, so simple enumeration of FOXP3+ cells is unlikely to assess Treg cell number or function accurately.23–27 On the other hand, recent data suggest that activated CD4+ T-effector cells which accumulate within inflamed joints do not frequently undergo the demethylation required to express activation-induced FOXP3.28 If this holds true in the inflamed gut, then activation-induced expression of FOXP3 may not be a major confounding factor in the enumeration of Treg cells in the gut. Finally, recent evidence from murine studies shows that CD4+ CD25+ FOXP3+ Treg cells can remain FOXP3+ yet display a Th17 cell phenotype when cultured in the presence of IL-6.29,30 If human Treg cells have a similar capacity to remain FOXP3+ while acquiring the properties of inflammatory Th17 cells in a pro-inflammatory environment, this would also have a significant impact on the interpretation of many studies that enumerate FOXP3+ cells. Clearly, ex vivo analyses of human Treg cells must be accompanied by rigorous functional assays and an extensive phenotypic analysis, including expression of CD127 and intracellular cytokine staining for IL-2 and interferon-γ (IFN-γ), which are not suppressed by activation-induced FOXP3.23

Toll-like receptor engagement affects CD4+ T-effector and Treg cells

The importance of Toll-like receptor (TLR)–ligand interactions in dendritic cell biology, and the resulting effects on adaptive immunity, are well established.31 More recently, however, it has been recognized that TLRs are expressed on CD4+ T cells, and that direct engagement of TLRs on T cells also affects adaptive immunity (reviewed in refs32–34). In the context of IBD, the role of TLRs on CD4+ T cells is of particular interest because commensal bacteria in the gut represent a large source of potential TLR ligands, and loss of integrity of the gut epithelium in IBD patients35 could result in abnormal exposure of T cells to bacteria-derived innate immune signals, and thereby promote disease.

There are now several reports characterizing the expression profile and function of TLRs on T-cell subsets, and these are summarized in Table 1. In most cases TLR engagement results in enhanced survival and accelerated induction of T-effector-cell responses, with the data relating to TLR-2, -3, -5 and -9 being the most consistent and conclusive. For example, stimulation of TLR-2 with bacterial lipoproteins/lipopeptides induces cytokine production and the proliferation of both murine36–38 and human CD4+ T cells.23,37,39 Stimulation of TLR-3, the receptor for viral and synthetic double-stranded RNA, enhances the survival, but not the proliferation, of activated murine CD4+ T cells.40 Flagellin, the only known ligand for TLR-5,41,42 enhances TCR-stimulated proliferation and cytokine production from human T-effector cells.39,43,44 Similarly, activation of TLR-7/-8, the receptor for viral single-stranded RNAs and nucleosides, stimulates memory CD4+ T cells to increase IFN-γ production.39 Finally, activation of TLR-9, the ligand for unmethylated cytosine-phosphate-guanine (CpG) motifs from bacteria, directly co-stimulates murine CD4+ T cells.45

Table 1.   Summary of the expression pattern and function of Toll-like receptor (TLR)1–10 on mouse and human CD4+ T-cell subsets
 HumanMouse
T-effector cellsT-regulatory cellsT-effector cellsT-regulatory cells
ExpressionBiologyExpressionBiologyExpressionBiologyExpressionBiology
  1. T-regulatory refers to cells defined on the basis of CD25 and/or forkhead box P3 (FOXP3) expression. T effector refers to cells that were either total CD4+ T cells, or CD4+ T cells that were depleted of T-regulatory cells. ND, not determined as of July 2008.

TLR-1Expressed, mRNA39,90,91NDNDNDExpressed, mRNA55,58NDExpressed mRNA55,92ND
TLR-2Expressed, mRNA;39,90,91 expressed in memory cells;37 not expressed mRNA and protein;93Increased proliferation and cytokine production in memory cells23,37,39Not expressed, mRNA and protein93Hsp60 enhanced suppression49Expressed, mRNA and protein36,48,55Increased proliferation any cytokine production,36,37,48 specific effect on Th1, not Th2 cells38Expressed, mRNA55,94PAM3Cys increased proliferation, transient decrease in FOXP3 expression and suppression,47,48
TLR-3Expressed mRNA39,90No biological effect detected39NDNDExpressed, mRNA and protein36,40,55Enhanced survival and proliferation36,40Expressed, mRNA55ND
TLR-4Expressed mRNA,39,90 not expressed mRNA and protein43,93No biological effect detected37,39,43Expressed mRNA43, not expressed protein43,93No biological effect detected43Expressed, mRNA and protein36,55Enhanced proliferation36Expressed, mRNA55Increased proliferation,55 no biological effect detected37,47
TLR-5Expressed; mRNA and protein39,43,90,91,93Enhanced proliferation and cytokine production39,43,44Expressed, mRNA and protein43,93Enhanced FOXP3 expression and suppressive capacity43Expressed, mRNA;40,55 not expressed, mRNA58 or protein36NDExpressed, mRNA55ND
TLR-6Expressed mRNA,90 not expressed mRNA and protein39,93NDExpressed, mRNA and protein93NDExpressed, mRNA55,58NDExpressed, mRNA55ND
TLR-7Expressed mRNA39,90,91Enhanced proliferation39Not Expressed mRNA50NDExpressed, mRNA55,58NDExpressed. mRNA55ND
TLR-8Expressed mRNA,90 not expressed, protein39,93No biological effect detected50Expressed mRNA protein50,93Reduced suppressive capacity50Expressed, mRNA55,58NDExpressed, mRNA55ND
TLR-9Expressed, mRNA39,90,91Expression down-regulated upon infection91Not Expressed, mRNA50NDExpressed, mRNA and protein36,40,95Increased proliferation36,40,51 and Bcl-x(L) expression95Not expressed, mRNA55Blocks suppressive function;51 necessary for suppression in a lupus model52
TLR-10Expressed, mRNA,90,91 not expressed, mRNA and protein39,93NDExpressed mRNA and protein93Expression regulated by FOXP393NDNDNDND

Stimulation of TLRs can also directly influence the function of Treg cells.46 For example, using mice deficient for TLR-2 or MyD88, an intracellular adaptor protein known to be essential for many TLR-mediated effects, it was shown that stimulation of TLR-2 on Treg cells, in the context of IL-2 and TCR triggering, induces proliferation and a transient loss of FOXP3 expression and suppressive capacity.47,48 In apparent direct contrast, exposure of human Treg cells to endogenous heat-shock protein 60, which also stimulates TLR-2, enhances their suppressive activity.49 TLR-5 is also expressed by human Treg cells at levels similar to those on CD4+ T-effector cells, and stimulation with its ligand, flagellin, enhances FOXP3 expression and suppressive capacity.43 In contrast to the effects of TLR-5, triggering TLR-8 on human Treg cells reverses suppression, while CD4+ T-effector cells remain unaffected.50 Similarly, stimulating TLR-9 acts directly on Treg cells to block suppression.51 In contrast, Tregs from tlr-9−/− mice are impaired in their ability to suppress an in vitro assay of autoreactivity in a lupus model, suggesting that TLR-9 may provide a positive signal to Treg cells in this context.52 An as-yet unanswered question is whether Treg cells differ from T-effector cells in terms of whether they are receptive to TLR signals independently from TCR engagement, or at very low concentrations of antigens.

There is significant controversy regarding the role of direct signalling via TLR-4 to T cells. In some cases lipopolysaccharide (LPS) (recognized by TLR-4) was reported to stimulate the activation of protein kinase C in CD3+ T cells53 and perforin production in a subset of CD4+ T cells in patients with ankylosing spondylitis.54 Moreover, Fukata et al.36 found that LPS-stimulated proliferation of T-effector cells was dependent on Myd88. In contrast, others found no effect of LPS on CD4+ T-effector cells.37,43,47 Whether or not LPS/TLR-4 interactions influence Treg cells is also controversial. Initial reports showed that LPS directly enhances suppression by murine Treg cells through TLR-4;55 however, subsequent studies found the expression level of TLR-4 on murine Treg cells to be low.40,47,48 With respect to human Treg cells, two studies found that these cells neither express TLR-4 nor respond to LPS,37,43 whereas another report suggests that Treg cells do express TLR-4 and that LPS stimulation results in decreased FOXP3 expression.56 The mostly likely explanation for these varying results is differences in the purity of the LPS used (commercial sources are often contaminated with TLR-2 ligands) as well as the purity of the cells assayed, which must be rigorously purified to ensure no contaminating antigen-presenting cells could be activated in response to LPS and thereby provide co-stimulation and cytokines.

Role for altered TLRs in IBD

TLRs are expressed on a large variety of immune cells and play a major role in initiating immune responses. In the steady state the gut is a rich source of TLR ligands from commensal bacteria, and during disease a variety of endogenous TLR ligands may also be released in the process of inflammation and tissue destruction. TLRs thus have dual roles in IBD: they are necessary for maintaining tolerance and eliminating pathogenic microorganisms, but they can also amplify inappropriate immune responses that ultimately cause chronic inflammation. The vast majority of studies have considered the role of TLRs in IBD in terms of their effects on innate immune cells, such as dendritic cells and macrophages (reviewed in ref.57). Recently, however, it has been demonstrated that specific ligation of TLRs on T cells can also influence the development of colitis. Pathogenic T-effector cells from Myd88-deficient mice have a significant defect in Th17 cell differentiation in vivo and do not induce colitis when transferred into immunodeficient mice.36,58 Moreover, myd88/− Treg cells are less able to suppress the adoptive transfer of colitis,36 and treatment of mice with a TLR-9 ligand induces the development CD4+ Treg cells that are protective in the T-cell transfer model of colitis.59 These results suggest that T-cell-specific stimulation of TLRs, presumably by commensal bacteria, has a significant role in the development of IBD.

Additional evidence for the role of TLRs in IBD comes from studies of flagellin and TLR-5. TLR-5-deficient mice, originally described by Uematsu et al.,60 develop spontaneous colitis,61 suggesting that flagellin–TLR-5 interactions normally have a protective anti-inflammatory role. In contrast, in humans, a common polymorphism in TLR-5 that produces a dominant-negative receptor is protective against CD, but not UC, in people of Ashkenazi Jewish descent.62,63 The reason for this apparent contradiction is unclear, but could relate to differences between mice and humans in species of commensal bacteria, roles of the intracellular sensor of flagellin [the nucleotide-binding oligomerization domain (NOD)-like receptor known as Ipaf],64 effects of TLR-5 on CD4+ T cells and/or differences in the expression of TLR-5. Differences between mouse and humans in expression of TLR-5 may be of particular importance, as studies have shown that in mouse intestinal tissue only CD11chi CD11bhi lamina propria dendritic cells express TLR-5,65 whereas in humans, TLR-5 is expressed on a variety of cells including intestinal epithelial cells,42 dendritic cells,66 endothelial cells,67 CD4+ T-effector cells and Treg cells.43

In addition to TLR-5, polymorphisms in the TLR-1, -2, -4, -6 and -9 loci have been associated with IBD in humans,68–70 although the functional effects of these variants are not well defined. In terms of protein expression, TLR-3 is down-regulated, whereas TLR-2 and -4 are up-regulated, in the intestinal mucosa of patients suffering from active IBD.71,72In vitro evidence suggests that altered TLR-2 signalling is associated with disease because peripheral blood mononuclear cells isolated from patients with active CD or UC produce more TNF-α in response to TLR-2 stimulation than those of healthy controls.73 There is also genetic evidence that functional polymorphisms in the intracellular sensor for bacterial peptidoglycan, NOD2 (also known as CARD15), is linked to CD,2 but there is currently no evidence regarding a possible role for these receptors in T cells.

Role for bacterial antigens in IBD

There is strong evidence that abnormal regulation of the immune response to commensal bacteria is a driving force in IBD (reviewed in ref.74) Mouse models of IBD have shown that the development of colitis is dependent on the presence of gut flora because the disease is generally absent in mice raised in germ-free environments and is ameliorated by the administration of antibiotics75 in normal laboratory conditions. Moreover, in IL-10-deficient mice, the development of spontaneous colitis requires commensal flora and an intact TLR–MyD88 signaling pathway.76 More direct evidence for the antigenic role of commensal flora comes from studies of immune responses towards specific gut bacterial antigens. For example, spontaneously colitic C3H/HeJBir mice (CBir), generated by selective breeding of C3H/HeJ mice for the colitic phenotype, show increased B-cell and T-cell reactivity towards bacterial antigens,77 and transfer of CD4+ T cells from colitic CBir mice causes colitis in immunodeficient recipient mice.78,79 This response is specific for commensal flora because the relevant T cells do not respond to food antigens or faecal extracts from germ-free mice.80

A variety of clinical observations in IBD patients suggest that responses to commensal bacteria are also central to the development of human disease. Most notably, IBD lesions are preferentially found in the terminal ileum and colon, which harbour the highest concentrations of bacteria. In addition, altering commensal flora with antibiotics or probiotics can reduce the symptoms associated with disease, albeit transiently.81 Similar to colitic mice, human IBD patients show enhanced immune responsiveness to gut bacterial antigens and concurrent loss of tolerance.82,83

Some of the specific bacterial antigens driving disease in mouse models of colitis and human IBD have recently been identified. One key antigen that drives disease is flagellin, the major structural subunit of bacterial flagella (reviewed in ref.42). Initially, flagellin from commensal Clostridia species was found to be an antigen in the CBir mouse model of spontaneous colitis. Remarkably, when sera from CBir mice were used to probe a caecal bacterial phage display library, 25% of all proteins cloned were flagellins from commensal organisms.84 More detailed analysis revealed that CBir mice have antibody responses to the amino-terminal conserved domain of flagellin, and that flagellin-specific CD4+ T-effector cells can induce severe colitis.84 Remarkably, even in wild-type mice, chemically induced colitis is associated with the development of anti-flagellin immunoglobulin responses, indicating that flagellin can be a dominant antigen in both spontaneous and experimentally induced colitis.85,86 Immune responses to flagellin are also implicated in human IBD:84,87–89 subsets of patients with CD, but not with UC, have elevated antibody responses to CBir flagellin,84,88 as well as to a wider variety of flagellin subtypes.87,88 That flagellin is the only known human TLR ligand which is a protein, and can thus activate both the innate and adaptive immune systems in parallel, may be one reason why this protein is a dominant antigen in IBD.

Concluding remarks and a model

It is clear that multiple defects in innate and adaptive immunity can initiate and sustain IBD. Here we have focused on evidence as to whether insufficient numbers or function of Treg cells could be a primary cause of IBD. Despite overwhelming data from animal models that alterations in the numbers or function of Treg cells cause colitis,4 it has been difficult to detect significant defects in peripheral or gut Treg cells in association with UC or CD in humans. These in vitro studies of immune regulation with human cells, however, have been limited by inadequate markers to track and isolate Treg cells, and an inability to replicate the in vivo inflammatory environment. There is therefore an urgent need to develop better assays with which to study the ex vivo Treg cells isolated from inflammatory environments and to identify markers that are specifically associated with function. If there is a real deficiency in Treg-cell function in human IBD, then cell therapy-based strategies similar to those that are effective in mice5,10 could be developed.

Of particular interest in the context of IBD is how TLR ligands and antigens derived from commensal bacteria may directly influence the balance and function of T-effector and Treg cells. Via its capacity to act as an antigen,84 an adjuvant both for T-cell and antibody responses85 and as a co-stimulatory molecule,43 the TLR-5 ligand flagellin appears to be central to this process (Fig. 1). We can speculate that the increased epithelial permeability associated with IBD could result in abnormal exposure of T cells in the lamina propria to flagellin from commensal bacteria. Flagellin could be phagocytosed and presented as an antigen by dendritic cells and, in parallel, stimulate TLR-5 on immune and non-immune cells. Evidence that mutations in TLR-5 may protect humans from CD62,63 suggests that pharmacological inhibition of this pathway may be therapeutically beneficial. Notably, because expression of TLR-5 in humans is much more widespread than in mice,85 parallel studies in mice and humans will be needed to define the clinical relevance of data from models of IBD.

image

Figure 1.  Model for how flagellin could influence the balance between T–effector (Teff) and T-regulatory (Treg) cells in inflammatory bowel disease (IBD). Increased epithelial permeability associated with IBD could result in abnormal exposure of cells in the lamina propria to flagellin from commensal bacteria. Flagellin could be phagocytosed and presented as an antigen by dendritic cells (DCs) to CD4+ T cells. In parallel, flagellin could stimulate Toll-like receptor 5 (TLR-5) on immune and non-immune cells to induce maturation, proliferation and cytokine production. Depending on the concentration, and on other inflammatory signals, flagellin could thereby alter the balance between Treg and Teff cells. Other TLR ligands from commensal bacteria may also directly stimulate TLRs on T cells and contribute to the overall effect. FL, flagellin; FOXP3, forkhead box P3; MHC, major histocompatibility complex; TCR, T-cell receptor.

Download figure to PowerPoint

With respect to flagellin acting as a TLR ligand on CD4+ T cells, low concentrations enhance the expression of FOXP3 and the suppressive capacity of Treg cells, whereas high concentrations stimulate T-effector cell function.43 Presumably, in active IBD, the latter situation would prevail and perpetuate the loss of normal Treg-cell function. Whether or not flagellin also has a role as an antigen in altering the balance of Treg and T-effector cells in this context remains to be defined. However, we speculate that the induction of antigen-specific tolerance to flagellin by promoting the development of flagellin-specific Treg cells may be an attractive approach to restoring intestinal regulatory responses and contribute to the treatment of IBD.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Acknowledgements
  5. References

The authors’ own work is partly supported by a new emerging team grant on IBD from the Canadian Institutes for Health Research (IIN 84037) and the Crohn’s and Colitis Foundation of Canada. M.E.H. holds a CIHR Graduate Masters Award and a CIHR/MSFHR Transplantation Training Program award, and G.H. holds an MSFHR postdoctoral fellowship. C.A.P. holds a Canada Research Chair in Regulatory Lymphocytes of the Immune System. T.S. holds a CIHR New Investigator Award. M.K.L. holds a Canada Research Chair in Transplantation and is a MSFHR Scholar. We thank Paul Orban for critical reading of the manuscript.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Acknowledgements
  5. References