Soothing touch of CD31 protects endothelium during cellular immune responses

CD31 is a member of the Ig gene superfamily that is present on the surface of platelets and all leukocytes (1). CD31 is also highly expressed on endothelial cells, where it represents a major constituent of the intercellular junction (1). Based on this cellular distribution and similarities in the cDNA sequence coding for extracellular domains, CD31 has initially been considered as a member “among others” of the large family of cell adhesion molecules (1). A striking illustration of this lies in the first name that the molecule was given: platelet endothelial cell adhesion molecule-1 (PECAM-1). However, the demonstration that the CD31 cytoplasmic tail contains consensus sequences typical of immunoreceptor tyrosine-based inhibitory motifs (ITIMs) has led to a reconsideration of CD31 biological functions (1). ITIM indeed contains tyrosine residues, the phosphorylation of which creates specific docking sites for Src-homology 2 domain-containing intracellular protein-tyrosine phosphatases. These catalytic enzymes, once localized to their cytoplasmic anchors and activated, are then able to inhibit tyrosine kinase-mediated signaling, and thus cellular activation (1). Building on these data, Cheung et al. (2) present in PNAS another advance in the understanding of the regulatory roles of CD31. Their study demonstrates that CD31 signaling through ITIM is necessary to prevent inflammatory-induced endothelial cell death.

Maintenance of vascular integrity during effector immune responses is a major challenge in the cooperation between the immune and the vascular systems. Endothelial cells are indeed essential to prevent the activation of the coagulation cascade, which would result in ischemic necrosis of the tissue. However, immune effectors that are activated in secondary lymphoid organs must interact with the endothelium to transit out of the bloodstream and access the antigen-rich sites, where the immune response is needed. Constitutive resistance of endothelial cells to cell death induced by inflammatory insults is therefore critical for “healthy” immune responses. This “immune privilege” of endothelial cells is well illustrated in transplantation. Because the donor is from the same species but genetically different, the recipient’s adaptive immune system recognizes donor-specific alloantigens expressed by the graft. Immune response that develops against alloantigens leads in turn to the destruction of the transplanted organ, a process named “rejection.” Interestingly, during rejection, recipients’ effector T cells cross the donor capillary endothelium without killing the endothelial cells (Fig. 1A) (3). In contrast, graft epithelial cells are highly susceptible to T-cell–mediated destruction (Fig. 1A).

Fig. 1.

T-cell– and antibody-mediated rejection of kidney allograft. (A) Typical features of T-cell–mediated rejection include infiltration of graft interstitium by recipient's mononuclear cells, which attack donor's tubular epithelial cells (black arrows). Despite the fact that trafficking of recipient's mononuclear cells into the interstitium require interactions with graft endothelial cells, the latter remain unharmed both in the macrovessels and the microvessels. (B) Pattern of expression of CD31 in a kidney allograft undergoing T-cell–mediated rejection. CD31 is detected on recipient's mononuclear cells, and donor's endothelial cells, from both the microvessels and the macrovessels. (C) Typical features of vascular chronic antibody-mediated rejection, also known as “graft arteriosclerosis.” Black arrow shows the thickness of the neointima. (D) The binding of alloantibodies to graft endothelial cells triggers classical complement pathway, which results in C4d deposition in graft microvessels and macrovessels. (Scale bar: 100 µm.) Abbreviations: a, artery; g, glomeruli; ptc, peritubular capillaries; t, tubules.

One mechanism by which endothelial cells may escape the destructive action of cytotoxic T cells is related to antigen presentation. The outcome of an interaction between cytotoxic T lymphocytes and target cells is partly determined by the amount of antigen presented by the target cells. Although endothelial cells are endowed with antigen-presenting abilities, they express lower levels of immunodominant epitope–MHC complexes on their surface than leukocytes or epithelial cells (4). By presenting a different antigenic repertoire compared with other cell types, endothelial cells may escape an attack by effector cytotoxic T lymphocytes, which have been educated by professional antigen-presenting cells and are therefore specific for immunodominant epitopes. This mechanism is, however, unlikely to be the only one involved because several studies have previously documented that triggering endothelial cell death receptors does not efficiently activate the extrinsic pathway of apoptosis (5).

Combining different approaches, Cheung et al. (2) demonstrate that exposure of endothelial cells to various inflammatory stimuli in vitro, including TNFα and cytotoxic T lymphocytes, resulted in the phosphorylation of the ITIMs of CD31, which counteracted the activation of the extrinsic pathway of apoptosis via the activation of the Erk/Akt pathway. Akt activation by CD31 signaling in endothelial cells prevented the localization of the forkhead transcription factor FoxO3 to the nucleus, thus inhibiting the transcription of the proapoptotic genes and derepressing expression of the antiapoptotic gene cFlar (2). This study therefore provides important novel insights on the molecular mechanisms whereby the vascular endothelium remains undamaged while interacting with effector immune cells migrating to the site of the response. Noteworthy, in contrast with endothelial cells, epithelial cells lack CD31 expression (Fig. 1B), which could account for their high susceptibility to T-cell–mediated destruction (Fig. 1A). In line with this hypothesis, the authors showed that CD31 gene transfer was sufficient to recapitulate the cytoprotective mechanisms in CD31-negative pancreatic beta cells, which became resistant to T-cell–mediated rejection when grafted in fully allogeneic recipients (2).

CD31 immunoregulatory role is not limited to endothelial cells. T cells and antigen-presenting cells both express CD31, and transhomophilic interactions between T cells and dendritic cells (DCs) play a crucial nonredundant role in defining the size of the ensuing immune response. The same group demonstrated a few years ago that triggering CD31 signaling in T cells during their priming, resulted in partial inhibition of proximal T-cell receptor signaling, which regulated primary clonal expansion, and acquisition of killing capacity or regulatory functions (6). Furthermore, an independent research group has recently reported that CD31 is also a key coinhibitory receptor on DCs. Indeed, the disruption of CD31 signaling favored the immunogenic maturation and migration of resident DCs to the draining lymph nodes (7). In contrast, triggering CD31 signaling during DC maturation resulted in reduced expression of costimulatory molecules and production of inflammatory cytokines, whereas the expression of regulatory cytokines TGF-β and IL-10 was increased (7).

Based on these findings, which demonstrate that CD31 signaling plays key nonredundant roles in pacifying cell–cell interactions at the vessel/blood interface under conditions of immunological stress, it was tempting to speculate that CD31 pathway could represent an attractive therapeutic target for immune-mediated diseases targeting the vasculature. Atherosclerosis, the principal cause of myocardial infarction and stroke, results from the infiltration of lipoproteins in the intimal space and the development of a pathogenic cellular immune response to modified lipoproteins within the arterial wall (8). Interestingly, it has been shown that genetic ablation of CD31 enhanced lesion formation in experimental atherosclerosis (9). Conversely, treating atherosclerosis-prone mice with a fusion protein composed of the extracellular portion of CD31 fused to the constant part of IgG1 (CD31Rg), reduced the size of atherosclerotic lesions (10). Furthermore, CD31Rg treatment correlated with more circulating regulatory T cells, blunted blood T-cell activation, and reduced T-cell infiltration within plaques (10).

So, is CD31 signaling the panacea for endothelial cells? Although graft endothelial cells remain relatively unharmed by recipient’s cytotoxic T cells, most of transplanted organs develop with time vascular chronic rejection lesions, also known as “graft arteriosclerosis” (Fig. 1C). Recipient’s immune response against the graft is indeed not limited to the generation of cellular effectors but also comprises the production of alloantibodies (11, 12). A seminal work by Russell, Colvin, and coworkers (13) demonstrated more than 20 y ago that the passive transfer of alloantibodies was sufficient to promote the development of graft arteriosclerosis. Binding of alloantibodies to directly accessible allogeneic targets expressed by graft endothelial cells triggers the classical complement pathway, as witnessed by the deposition of C4d in graft microvasculature (Fig. 1D). This leads to the release of adhesion molecules and cytokines that in turn recruit innate immune effectors (11, 12). Chronic vascular inflammation promotes the development of typical histological lesions and irreversible loss of graft function (11, 12). Interestingly, Cheung et al. (2) show in their study that CD31-mediated cytoprotection did not extend to mechanisms of cell death other than apoptosis: CD31 KO and CD31 WT endothelial cells were equally sensitive to antibody-mediated cell death.

Therefore, if the soothing touch of CD31 appears necessary to protect endothelial cells during their interaction with cellular effectors, the role of this pathway in the initiation and/or the effector phase of antibody-mediated vascular injuries is less evident and warrants further investigations.