Thymus-derived CD4+ CD25+ T regulatory cells (Tregs) are essential for the maintenance of self-tolerance. What critical factors and conditions are required for the extra-thymic development of Tregs remains an important question. In this study, we show that the anti-inflammatory extracellular matrix protein, thrombospondin-1, promoted the generation of human peripheral regulatory T cells through the ligation of one of its receptor, CD47. CD47 stimulation by mAb or a thrombospondin-1 peptide induced naive or memory CD4+CD25− T cells to become suppressive. The latter expressed increased amounts of CTLA-4, OX40, GITR, and Foxp3 and inhibited autologous Th0, Th1, and Th2 cells. Their regulatory activity was contact dependent, TGF-β independent, and partially circumvented by IL-2. This previously unknown mechanism to induce human peripheral Tregs in response to inflammation may participate to the limitation of collateral damage induced by exacerbated responses to self or foreign Ags and thus be relevant for therapeutic intervention in autoimmune diseases and transplantation.
Control of self-reactive T cells by naturally occurring CD4+CD25+ regulatory T cells (Tregs)5 mediates peripheral T cell tolerance and prevention of autoimmunity (1, 2). In vitro, Tregs are anergic and suppress the activation and proliferation of naive T cells in a non Ag-specific, but contact-dependent manner (3). In vivo proliferation of Tregs, together with their migration to inflammatory sites increases their potential function for the control of pathogen-specific effector T cells in sites of inflammation as well as tumor-specific T cells (4, 5, 6). Foxp3, a member of the forkhead family of transcription factors encoding the protein scurfin, is a master gene controlling the development and the function of Tregs (7, 8). Scurfy mice and humans with IPEX (Immune dysregulation, polyendocrinopathy, enteropathy) syndrome have a loss-of-function mutation of Foxp3, and develop severe autoimmune and inflammatory diseases (9). Adoptive transfer of Tregs from wild-type littermate into sf mice prevents development of the disease. Although there is no doubt that Tregs are generated within the thymus during normal T cell development, the peripheral conversion of CD4+CD25− T cells into regulatory T cells and their Ag-specific expansion has been recently demonstrated (10, 11). Such Tregs are distinct from the in vitro induced “adaptive” regulatory T cells that mainly include Tr1 and Th3 cells (1, 12, 13). Tr1 cells are Foxp3− (14) and differ from Tregs by their pattern of cytokine production (IL-10 and/or TGF-β, respectively) and their mode of suppression (in vitro contact-independent inhibitory function). An important question is, which of the endogenous receptors and/or cytokines convert extrathymic naive Foxp3− CD4+CD25− T cells into Tregs expressing Foxp3. Recent reports indicate that minute Ag dose turns on Foxp3 in murine naive CD4+CD25− T cells and converts them into regulatory T cells. TGF-β and IL-2 favor their peripheral conversion (15, 16, 17). Similarly, TGF-β induces human CD28- and TCR-stimulated CD4+CD25− T cells to acquire the function and phenotype of Tregs (18). In fact, TGF-β plays a critical role in the development, function, and survival of Tregs (19, 20, 21).
Thrombospondin-1 (TSP) is a potent anti-inflammatory molecule, which is rapidly and transiently released at high concentrations in injured and damaged tissues even in the absence of pathogen (22, 23). TSP, predominantly secreted by platelets and APC, inhibits angiogenesis, tumor cell growth and promotes apoptosis (24). Notably, TSP is composed of several structural domains that specifically interact with a range of cellular receptors differentially expressed by a variety of cells (24), which could explain its broad spectrum of biological activities. For instance, TSP exerts both stimulatory and inhibitory effects on T cells via two TSP receptors, i.e., α4β1 and CD47, respectively. However, intact TSP is predominantly an inhibitor of TCR signal transduction (25) and IL-12 responsiveness of naive and adult T cells (26, 27). Most relevant to the present study, TSP is the main activator of TGF-β (28). As such, TGF-β and TSP-null mice display a similar phenotype, persistent inflammation in multiple organs (28). Thus, TSP displays at least two essential properties required to dampen inflammation. It inhibits proinflammatory cytokine secretion by APC (29) and facilitates the removal of apoptotic cells by phagocytes (30).
We previously reported that CD47 ligation induces naive T cell anergy (31). In this study, we provide evidence that interaction between CD47 on activated CD4+CD25− human T cells and CD47-specific TSP peptide promoted the generation of CD4+Foxp3+ T cells that suppressed the proliferation and cytokine production of autologous Tho, Th1, and Th2 cells via a contact-dependent mechanism. The induction of extra-thymic regulatory T cells in response to TSP represents a third mechanism whereby TSP may interfere with the inflammatory process, thus identifying CD47/TSP as potential targets for the development of anti-inflammatory drugs.
Materials and Methods
Anti-CD3 mAb (UCHT-1) was provided by Dr. C. Hivroz (Institut Curie, Paris, France). The CD32/B7.1 double-transfected mouse L fibroblasts have been described previously (31). The 4N1K peptide (KRFYVVMWWKK) corresponding to the C-terminal domain of TSP that selectively binds CD47 and the mutant 4NGG peptide (KRFYGGMWKK) were obtained from Genosys. The 4N1K sequence is highly conserved in all TSP family members (32). CD47 mAb, (clone B6H12, mouse IgG1, BioSource International) was used in soluble form. CFSE was purchased from Molecular Probes. Human soluble TGF-βRII was obtained from R&D Systems.
Cell preparation and culture conditions
The procedure of blood collection from human volunteers has been reviewed and approved by CHUM ethic and research committees. Umbilical cord blood mononuclear cells (CBMC) and PBMC from normal healthy volunteers were isolated by density-gradient centrifugation of heparinized blood using Lymphoprep (Nycomed). Naive CD4+CD25+ and CD4+CD25− T cells were selected from CBMC by depletion of CD8+ cells (EasySep kit; StemCell Technologies) followed by negative selection using CD25 microbeads (Miltneyi Biotec). For adult naive CD45RA+CD4+CD25− T cells, we performed an additional depletion using CD45RO microbeads. Memory CD4+ T cells were obtained by CD8 depletion (StemCell Technologies) followed by negative selection with CD25 and CD45RA microbeads. Purity was determined to be >97% CD4+CD25− by FACS analysis.
The differentiation of naive T cells into Th0, Th1, and Th2 cells was performed as described previously (31). Briefly, CBMC were cultured with PHA (1 μg/ml) in neutral conditions or in the presence of IL-12 (60 pM) and anti-IL-4 mAb (1 μg/ml) or a combination of IL-4 (20 ng/ml) and anti-IL-12 (5 μg/ml), respectively, together with soluble CD47mAb (10 μg/ml) and soluble 4N1K or 4NGG (50 μg/ml). After 3 days, cells (>98% CD3-positive cells) were washed and expanded for 9–12 days in cultured medium supplemented with IL-2. To assess cell division during expansion in IL-2, cells were labeled with CFSE at the end of the primary cultures. Cells were then washed and restimulated at 2 × 105 cells per ml with anti-CD3 (200 ng/ml) immobilized on mitomycin C-treated double CD32/B7.1 L-transfectants (2 × 105 cells per ml). Purified naive CD4+CD25− T cells were isolated from CBMC and primed with anti-CD3 (1 μg/ml) in the presence of IL-1 (10 ng/ml) and TNF-α (25 ng/ml). After 3 days, cells were expanded in medium supplemented with IL-2 and restimulated as described above. Cell proliferation was determined by adding 1 μCi [3H]thymidine (Amersham Biosciences) during 6 h of 2 days of culture. IFN-γ and IL-10 were measured by ELISA after 48-h restimulation as described (31). For TGF-β detection, T cells were restimulated for 72 h in synthetic HB101 serum-free medium (Irvine Scientific).
Memory CD4+CD25− T cells (0.5 × 106 cells/ml) were activated with plate-bound anti-CD3 mAb (10 μg/ml) alone or with soluble anti-CD28 (0.5 μg/ml) (BD Pharmingen) in the presence or absence of CD47 mAb (10 μg/ml) in 96-well flat-bottom microplates. After 4 days of priming, cells were washed and expanded (1 × 106 cells/ml) in 24-well plate for 7 days in the presence of anti-CD3 mAb (0.1 μg/ml) immobilized on CD32 L-transfectants (8 × 104 cells/ml) and IL-2 (200 IU/ml). At day 4, 1 ml of complete medium supplemented with IL-2 (400 IU/ml) was added to the culture. T cells were collected and re-stimulated weekly. After three rounds of expansion, memory T cells were rested for 48 h in complete medium with IL-2 before examining their regulatory activity.
Tas (1 × 105 cells per ml) generated from CBMC, CD25−CBMC, and umbilical naive CD4+CD25− T cells were cocultured with autologous Th0, Th1, or Th2 cells at a 1:1 ratio unless otherwise indicated. Cultures were stimulated for 48–72 h with anti-CD3 (50 ng/ml) immobilized on mitomycin C-treated double CD32/B7.1 L-transfectants fibroblasts (1–2 × 105 cells per ml). To determine their mode of suppression, Tas and Th0 or Th1 were physically separated using Transwell culture system (BD Biosciences). In some experiments, the regulatory function of Tas was assessed as follows: freshly isolated purified CD4+ responder T cells (1 × 105 per ml) were CFSE labeled and cocultured with autologous unlabelled purified CD4+ T cells or allogeneic expanded naive or memory regulatory T cells at a 3:1 ratio.
Flow cytometry analysis
Cells were stained with FITC-, PE-, PerCP-conjugated CD4, CD25, and CD28 mAb; granzyme A; Fas; FasL; biotinylated CD47; and OX40-conjugated mAbs followed by allophycocyanin-streptavidin (BD Biosciences), allophycocyanin-granzyme B (Caltag), PE-coupled GITR mAb, PE-TGF-β chicken Ab (R&D Systems) or isotype-matched control Abs (BD Biosciences). Cells were analyzed using a FACScalibur (BD Biosciences). For intracytoplasmic staining, cells were fixed with 1% paraformaldehyde for 10 min at room temperature, washed, and stained in permeabilizing buffer (PBS containing 0.03% saponin) with PE-conjugated CTLA-4 (BD Biosciences). To detect intracytoplasmic Foxp3 expression, cells were first fixed and then stained with PE-anti-human Foxp3 mAb (eBiosciences) in permabilizing buffer for 30 min at 4°C.
Foxp3 mRNA analysis
RNA was extracted using an RNeasy Mini kit (Qiagen), and 1 μg of each sample was reversed transcribed using SuperScript (Invitrogen Life Technologies). For real-time quantitative PCR (TaqMan), message levels were quantified using the ABI PRISM 7700 sequence detector system (BD Applied Biosystems). The quantitative PCR was performed using TaqMan universal PCR master mix (BD Applied Biosystems) according to the manufacturer′s protocol. Relative expression of each sample was determined by normalizing expression of each target to GAPDH, and then comparing this value to the normalized expression in a reference sample to calculate a fold-change value. The values for samples were calculated by ddCT method. Foxp3 quantitative analysis was performed using Assay on Demand and GAPDH using predeveloped TaqMan assay reagent target kits (Applied Biosystems).
Induction of human regulatory T cells by CD47/TSP interaction
Our initial studies indicated that CD47 ligation on naive T cells promotes the development of anergic T cells that produce very low amounts of cytokines including IL-2, IL-4, IL-5, IL-13, and IFN-γ (31). In this study, we show that the newly generated T cells, hereafter named Tas (for T anergic/suppressor cells), displayed regulatory activity (Fig. 1). We started with a well-established model known to induce Th0, Th1, and Th2 cells from cord blood mononuclear cells (CBMC), an abundant source of unmanipulated naive T cells. CBMC were stimulated with PHA in neutral conditions in the absence or presence of irrelevant control (Th0) or CD47 mAb (Tas); after 3 days, >98% of the cells in culture were CD3+ T cells (31). The cultured T cells were CFSE labeled and expanded in IL-2, and cell division was examined at day 5 (Fig. 1,A). Both CD47 mAb-treated CD4+ and CD8+ T cells (Tas) underwent less cell division than did Th0 cells. In fact, Tas expanded four times less and displayed significant lower proliferative response than did Th0 to CD3/CD28 costimulation (Fig. 1,B), indicating that CD47 ligation induced a general T cell hyporesponsiveness even after the delivery of two signals. Tas secreted less IL-10 but similar amounts of TGF-β, compared with Th0 (Fig. 1,C). They suppressed in a dose-dependent manner the proliferation and production of IFN-γ by autologous Th0 in response to low and high CD3 stimulation (Fig. 1, D and F). The mean percentage reduction in Th0 proliferation and IFN-γ production in presence of Tas in five experiments was 65 and 73%, respectively.
TSP, the natural ligand of CD47 is a multifunctional molecule containing several structurally and functionally distinct domains (24). Because T cells express a second TSP receptor, α4β1, we studied the 4N1K peptide that corresponds to the C-terminal domain of TSP and selectively binds CD47 as ligand. CBMC were stimulated with PHA in the presence of 4N1K (Tas) or irrelevant control peptide 4NNG (Th0), washed and expanded with IL-2. CD47/4N1K interactions on primary naive cells promoted the generation of regulatory T cells (Th0 + Tas). The 4NNG-treated cells lacked suppressive function (2 × Th0 cells) (Fig. 1 G). These data strongly suggested that CD47 mAb, as reported previously, exert an agonistic activity mimicking the natural CD47 ligand, TSP.
Contact-dependent and TGF-β-independent regulatory function of Tas
Effector T cells are refractory to the suppressive activity of Tregs in the presence of IL-4 (33). In this study, we showed that Tas inhibited IL-4 and IFN-γ production as well as cell proliferation of autologous Th2 and Th1 cells, respectively (Fig. 2, A and B). We next attempted to explore the mechanisms implicated in Tas regulatory function and found that the suppression of autologous T cells was contact dependent (Fig. 2 B). Accordingly, Tas did not suppress the proliferation and cytokine production when separated from their targets by a semipermeable membrane.
Murine Tregs regulatory function is contact dependent and exogenous IL-2 abrogates the suppression of proliferation but not cytokine production of T effector cells in a cocultures system (34). IL-2 circumvented Tas inhibition of Th0 cell proliferation and IFN-γ production. However, Il-2 strongly augmented IFN-γ production and Th0 + Tas cells still produced significantly less IFN-γ than did Th0 cells (Fig. 2 C).
Some studies report that Tregs exert their suppressive function through endogenous TGF-β, expressed as a membrane-bound or secreted molecule. Also, TSP is the main activator of TGF-β through its type1 repeat and not its C-terminal domain (28). In this study, we showed that Th0 and Tas produced similar amounts of TGF-β (Fig. 1,C). It was still relevant to examine whether the CD47-induced regulatory function was mediated by active TGF-β. The inhibitory function of Tas was not overcome by soluble TGF-βIIR, pointing to a TGF-β-independent mode of suppression (Fig. 2 D). Observations indicated that Tas regulatory function was not altered by the addition of IL-1 or IL-6, known to abrogate the suppressive function of murine Tregs (35), and that neutralization of TGF-β and IL-10 in the induction phase (CBMC primary cultures) did not preclude suppressive function of the expanded Tas (P. Grimbert and M. Rubio, unpublished observations). Thus, the development of Tas following CD47/4N1K did not require the presence of the TGF-β and IL-10, known to convert naive T cells into adaptive Tregs and/or Tr1.
Activated phenotype and Foxp3 expression by Tas
We next examined the phenotype of Tas and compared it to that of Th0 cells. As expected from activated T cells, CD4+ Th0 cells expressed the three conventional markers of Tregs, i.e., GITR, CTLA-4, and OX40 (Fig. 3, A and C). However, the expression of these three molecules was significantly up-regulated in CD4+ Tas. Tas did not express significant amounts of membrane TGF-β, latency-associated peptide of TGF-β nor TSP (data not shown), further corroborating with their TGF-β-independent suppressive function. They expressed similar quantities of CD25, CD28, and CD47 to Th0 cells. Because the suppression of Th0 by Tas was contact dependent, we examined the expression of mediators of cytotoxicity. CD4+ Tas expressed similar amounts of Fas and granzyme A to Th0 and no significant amount of FasL and granzyme B (Fig. 3 B).
Foxp3 is considered to be the hallmark of Tregs, and we found an increased proportion of CD4+CD25+Foxp3+ cells in Tas, compared with Th0 (Fig. 3,D). In keeping with recent observations, Th0 were comprised of a mixture of CD4+CD25+Foxp3− effectors and CD4+CD25+Foxp3+ T cells (11). We confirmed the up-regulation of Foxp3 expression by quantifying mRNA in Th0 and Tas using real-time PCR (Fig. 3 D). Tas displayed a 3-fold increase in Foxp3 mRNA, compared with Th0. Also, Foxp3 mRNA expression was higher in freshly isolated and highly purified CD4+CD25+ than CD4+CD25− cord blood T cells, confirming that Foxp3 was indeed expressed in naive CD4+CD25+ T cells in humans (36, 37).
Conversion of naive CD4+CD25− naive and memory T into regulatory T cells by CD47 ligation
Thus, one could argue that CD47 ligation exclusively expanded the proliferation of preexisting naive CD4+CD25+ Tregs present in CBMC. As a first approach, we used CBMC depleted in CD25+ cells (CBMC CD25−) and showed that, in the absence of naive CD4+CD25+ Tregs, CD47 ligation promoted the development of regulatory Tas (Fig. 4 A). These data suggest that induction of Tas occurred by engaging CD47 on CD45RA+CD25− T cells and did not require the presence of naive CD4+CD25+Foxp3+ Tregs in the starting cultures. However, we did not exclude the possibility that CD47/TSP interactions might as well favor the expansion of thymus-derived Tregs.
So far, we showed that ligation of the TSP receptor on unfractionated CBMC promoted the generation of regulatory T cells enriched in GITR+OX40+CTLA-4+CD4+ T cells and CD4+Foxp3+ T cells. We also observed an increased proportion of Fas+ granzymeA+granzymeB+ CD8+ T cells in the CD47 mAb-treated cell preparations (Fig. 3,B). In fact, the CD47-induced enriched Tas population displayed no cytolytic function on autologous target cells or K562 leukemic cell lines (data not shown), whereas the CD4+ T cells isolated at the end of Tas expansion in IL-2 (the proportion ranges from 28 to 70% in different blood samples), displayed a similar suppressive function to the nonseparated Tas cell preparation (Fig. 4 B). Notwithstanding, these data did not rule out a requirement for CD8+ T cells at the initial phase of CD4+ Tas induction.
We therefore determined whether direct CD47 engagement on purified naive CD4+CD25− T cells converted them into regulatory T cells. Indeed, TSP inhibits dendritic cell (DC) maturation and T cell activation and thus may exert its effect either on DC and/or on T cells to induce CBMC-derived Tas (25, 38). To this end, purified CD45RA+CD4+CD25− T cells, which did not express Foxp3 mRNA (Fig. 3,D), were stimulated with anti-CD3 in the presence of CD47 mAb or 4N1K peptide. Naive T cells absolutely require a costimulatory signal for their initial activation but CD28 costimulation prevents CD47-induced anergy in naive T cells (31). We thus supplemented the CD3-activated naive T cells with IL1-β and TNF-α (39), a context mimicking inflammatory conditions whereby TSP is abundantly produced in vivo. CD47 ligation on activated purified naive CD4+CD25− T cells promoted the development of Tas that suppressed IFN-γ production of autologous Th0 cells and expressed increased amounts of Foxp3 when compared with Th0 generated in the presence of control mAb or 4NNG peptide (Fig. 4,C). To ascertain the suppressive function of Tas on the target cells, we independently investigated the proliferation of the latter, using CFSE-labeled flow cytometry. Tas generated from adult naive CD4+CD25− T cells significantly inhibited cell division of allogeneic adult CD4+ T cells as measured by CFSE dilution (Fig. 4,D). Finally, CD47 ligation induced adult CD3-stimulated CD4+CD25− memory T to become suppressor cells (Fig. 4E). However, in this particular case, memory T cells required repetitive stimulation with anti-CD3 and IL-2 to become efficient regulatory T cells. Again, Tas suppressive function was contact dependent under these experimental conditions (data not shown). In humans and mice (11, 40), CD28 allows the peripheral conversion of CD4+CD25− T cells into regulatory T cells. In this study, we confirmed these observations (Fig. 4 E, right). However, CD28 costimulation at priming impaired CD47-induced Tas development from memory T cells, in agreement with the impairment of anergy induction in naive T cells (31).
Taken together, our data demonstrate that interaction between a CD47-specific TSP peptide and CD47 on human naive (cord blood or adult) and memory CD4+CD25− T cells promoted the generation of regulatory T cells that were enriched in CD4+Foxp3+ and exerted their suppressive function in a contact-dependent manner.
Our present findings identified TSP, a protein rapidly and transiently expressed at high levels in inflamed and damaged tissues, as an inducer of human peripheral regulatory T cells. The latter were hyporesponsive to CD3 and CD28 costimulation, produced small quantities of cytokines and suppressed the proliferation and cytokine production by autologous Th0, Th1, and Th2 cells in a contact-dependent manner. These adaptive Tregs expressed an activated phenotype as well as increased amounts of Foxp3. As such, the CD47-induced Tregs resemble naturally occurring thymus-derived Tregs. However, Tas exhibited distinct features: they expressed similar amount of CD25 to Th0 and displayed neither granzyme B expression nor cytolytic function. The induction of adaptive Tregs by TSP during an inflammatory process is reminiscent of recent findings showing that CD46 ligation by the C3b complement fragment induces the development of cytotoxic Tr1. The latter express substantial amounts of granzyme B (41, 42). Also, Tr1/IL-10-induced Tregs lack Foxp3 expression and do not require cell-cell contact to exert their regulatory function in vitro (14, 43). Tas thus differed from CD46 mAb- or IL-10-induced Tregs by phenotype, cytokine profile and mode of suppression (13, 44). TSP is produced by immature and mature DC (29). Immature DC are described as tolerogenic (45) in that they induce T cells that secrete large quantities of IL-10 and suppress the proliferation of autologous freshly isolated PBMC in a contact-dependent manner (44). Interestingly, PGE2 is a potent inducer of TSP in DC (29) and induces Foxp3 on human CD4+ T cells that display contact-independent regulatory function (46).
In this study, we provided evidence that CD47 ligation converted naive or memory CD4+CD25− T cells into CD4+Foxp3+ T cells. CD4+CD25+Foxp3+ regulatory T cells are thymus-derived and now designated endogenous Tregs or natural Tregs, but increasing evidence in mice and humans indicate that CD4+CD25+Foxp3-expressing T cells may be very efficiently induced outside the thymus. The induction of peripheral adaptive Tregs is often proven to be TGF-β- mediated (15, 18, 47). In that regard, Foxp3+ Tregs are rapidly induced following blood-stage infection with malaria sporozoite, and their presence is associated with a burst of TGF-β production, decreased proinflammatory cytokine production, and Ag-specific immune responses (48). Because TSP is an important activator of TGF-β, it was relevant to demonstrate that TSP (i.e., 4N1K peptide) might promote the generation of CD4+ Tregs via its direct interaction with CD47 and Tas suppressive function was TGF-β independent.
Selective expression of the Foxp3 gene in naturally occurring Tregs during thymic development is not sufficient to protect otherwise Foxp3-null mice from developing disease (9, 49). This suggests that continued peripheral expression of Foxp3 is required to control disease. Again, TGF-β participates to the maintenance of peripheral Foxp3 expression (21), and the present results suggest that intact TSP may directly or indirectly contribute to this process in vivo. Nevertheless, while Foxp3 expression is definitely required to confer suppressive activity in murine Tregs (7), it may not be sufficient for the function of human Tregs isolated from healthy donors or patients suffering from multiple sclerosis or autoimmune polyglandular syndrome type II (50, 51). The latter express similar amounts of Foxp3, CTLA-4, and TGF-β to Tregs from control donors.
TSP is endowed with potent anti-inflammatory activities at steady state and under inflammatory conditions. In vivo, using a cell-based gene therapy in a mouse model of rheumatoid arthritis, TSP acts by suppressing the IL-1β and enhancing the TGF-β production (52). In this study, we propose that TSP that is produced in the lymph node (mainly by mature DC) may convert naive CD4+CD25− T cells into Tregs that negatively control the inflammatory processes either in lymphoid organs and/or in sites of inflammation. Increased recruitment of naturally occurring Tregs combined with peripheral conversion of naive T cells appear to occur in situ when exacerbated inflammation takes place, for instance in joints of rheumatoid arthritis patients (53). However, the molecules involved in Tas migration to tissues remain to be determined. The CD103 Ag, predominantly expressed by CD25− and CD25+ murine Tregs with suppressive properties (54), is induced by TGF-β (55) and involved in the retention of Tregs to the inflammatory sites (56). CCR4 and CCR8 also are expressed on human Tregs (57). Observations indicate that Tas, like freshly isolated human Tregs (58), do not express significant levels of CD103 and displayed similar amounts of CCR7, CCR4, and CCR8 to Th0 (P. Grimbert, S. Bouguermouh, and M. Rubio, unpublished observations). Finally, TSP, also secreted at peripheral sites in response to inflammation, might further dampen the inflammatory process by down-regulating DC maturation via at least three receptors: CD47, CD51, and CD36 (38), as well as facilitating the engulfment of damaged tissues and cells.
Therefore, it can be hypothesized that TSP, akin to TGF-β and IL-10, when released in the context of an inflammatory response, promotes the generation of peripheral regulatory T cells that keep in check the inflammatory process and avoid collateral damage induced by self or foreign Ags. After renal allograft, TSP, TFG-β, and Foxp3 mRNA are increased (23, 59). Surprisingly, the latter are significantly correlated with acute rejection, and high levels of Foxp3 transcripts predict an acute rejection reversal. Finally, ex vivo induction and expansion of human suppressor T cells by CD47 mAb, as recently proposed for CD45RO/RB mAb (60), may open avenues to design strategies for the treatment of graft rejection, inflammatory and autoimmune diseases.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the Canadian Institute for Health and Research (MOP-4490). Philippe Grimbert was sponsored by the French Society of Transplantation.
Abbreviations used in this paper: Treg, regulatory T cell; TSP, thrombospondin-1; CBMC, cord blood mononuclear cell; DC, dendritic cell.