CD40Ig treatment results in allograft acceptance mediated by CD8⁺CD45RC^low T cells, IFN-γ, and indoleamine 2,3-dioxygenase

Treatment with CD40Ig results in indefinite allograft survival in a complete MHC-mismatched heart allograft model in the rat. Here we show that serial second, third, and fourth adoptive transfers of total splenocytes from CD40Ig-treated recipients into secondary recipients led to indefinite donor-specific allograft acceptance. Purification of splenocyte subpopulations from CD40Ig-treated recipients demonstrated that only the adoptively transferred CD8⁺CD45RC^low subset resulted in donor-specific long-term survival, whereas CD8⁺CD45RC^low T cells from naive animals did not. Accepted grafts displayed increased indoleamine 2,3-dioxygenase (IDO) expression restricted in the graft to ECs. Coculture of donor ECs with CD8⁺CD45RC^low T cells purified from CD40Ig-treated animals resulted in donor-specific IDO expression dependent on IFN-γ. Neutralization of IFN-γ or IDO triggered acute allograft rejection in both CD40Ig-treated and adoptively transferred recipients. This study demonstrates for what we believe to be the first time that interference in CD40–CD40 ligand (CD40-CD40L) interactions induces allospecific CD8⁺ Tregs that maintain allograft survival. CD8⁺CD45RC^low T cells act through IFN-γ production, which in turn induces IDO expression by graft ECs. Thus, donor alloantigen-specific CD8⁺ Tregs may promote local graft immune privilege through IDO expression.

Organ transplantation would benefit from immunotherapeutic strategies that enable the development of transplantation tolerance (1). Tolerance of donor antigens would avoid or reduce the administration of immunosuppressors and would simultaneously reduce the incidence of chronic graft dysfunction.

Over the past few years, Tregs of different phenotypes capable of suppressing immune responses have been identified (reviewed in ref. 2). Although several CD4⁺ populations have been described as being Tregs, CD8⁺ Tregs have also been identified in different models (3–14). One example of CD8⁺ Tregs is the non–cytokine-secreting, CD8⁺CD28^– alloantigen-specific Treg, which interacts directly with DCs or ECs and renders them tolerogenic (4, 11). Recently, a new natural regulatory CD8⁺CD45RC^lowFoxp3⁺ subset has also been characterized (3).

In certain cases, allogeneic tolerance can be maintained and perpetuated into subsequent recipients by multiple and successive adoptive transfer of CD4⁺ Tregs into naive hosts (1). This condition has been termed infectious tolerance and has until now only been described for CD4⁺ Tregs (1, 15, 16).

Administration of anti-CD40L (17) or anti-CD40 (18) mAbs as well as administration of CD40Ig (19) are among the most efficient treatments available to obtain long-term organ allograft survival in animal models, but the mechanisms are still poorly understood (reviewed in ref. 20). Anti-CD40L mAb has been shown to induce CD4⁺ Tregs (21) and deplete activated T cells through Fc-dependent mechanisms (22) or apoptosis (23). Interaction of DCs and T cells in the absence of CD40-CD40L interactions results in cross-tolerization of CD8⁺ T cells against cognate antigens (24–26). Anti-CD40L mAbs have also been shown to induce regulatory cells with a DC and/or NK phenotype in a model of autoimmune diabetes (27). CD40-CD40 ligand (CD40-CD40L) interactions can affect immune responses and inflammation not only through immune cell interactions but also by modulating the function of ECs and platelets (20).

We previously reported that administration of CD40Ig through gene transfer led to indefinite allograft survival in a complete MHC-mismatched heart allograft model in the rat (19). The aim of the present work is to analyze the in vivo mechanisms underlying graft acceptance.

Here we demonstrate that CD40Ig-induced graft acceptance is mediated by alloantigen-specific CD8⁺CD45RC^low Tregs that were able to impose inhibition of acute rejection to naive T cells when they were successively adoptively transferred up to 4 times. This is the first description of manipulation of CD40-CD40L interactions that resulted in the generation of CD8⁺ Treg cells. We also provide evidence that IFN-γ is produced by CD8⁺CD45RC^low T cells and that indoleamine 2,3-dioxygenase (IDO) plays a crucial role in suppression of graft rejection. Finally, we show that these CD8⁺ Tregs induce ECs from graft origin to produce IDO through IFN-γ.

CD40Ig-induced long-term survival is serially transferable to naive recipients. To assess whether tolerogenic cells were involved in the long-term allograft survival induced by CD40Ig treatment, we performed adoptive cell transfer experiments using splenocytes from CD40Ig-treated recipients. In agreement with our previously published results (19), recipients that received fully MHC-discordant LEW.1W heart allografts transduced with an adenovirus encoding CD40Ig (AdCD40Ig) exhibited long-term allograft survival (>120 days in 93% of recipients), whereas all recipients of grafts transduced with the control noncoding adenovirus dl324 (Addl324), rejected their grafts rapidly (8 ± 0.8 days). The first adoptive transfer of 50 × 10⁶ and 200 × 10⁶ splenocytes from CD40Ig-treated long-term survivor recipients (collected at day 120 after transplantation) to naive syngeneic recipients resulted in long-term acceptance of all LEW.1W cardiac allografts (Table 1). Grafts were rejected (6.5 ± 0.3 days) by recipients that did not receive splenocytes or that received splenocytes from control adenovirus-treated recipients (rejection 10 ± 2 days) (Table 1). Third-party Brown-Norway (BN) hearts transplanted into recipients receiving splenocytes from CD40Ig-treated long-term survivors with a LEW.1W heart were also promptly rejected (9 days) (Table 1). Whereas CD40Ig-treated recipients displayed high levels of circulating CD40Ig at day 10 after transplantation (229 ± 113 μg/ml), recipients of the first adoptive cell transfer had undetectable levels of CD40Ig (<0.02 ng/ml). Recipients with long-surviving grafts after adoptive cell transfer displayed a reduction by more than 90% of serum anti-LEW.1W alloantibodies (Supplemental Figure 1A; supplemental material available online with this article; doi:10.1172/JCI28801DS1) and a 51% reduction in splenocyte anti-donor CTL activity (Supplemental Figure 1B).

Table 1

Cardiac allograft survival after serial adoptive cell transfers

Second adoptive cell transfers of 50 × 10⁶ to 200 × 10⁶ splenocytes from CD40Ig-treated recipients into naive animals also resulted in long-term acceptance of allogeneic hearts in all recipients (Table 1). All subsequent (third and fourth) adoptive cell transfers into naive animals performed with 50 × 10⁶ splenocytes recovered 120 days after transplantation and adoptive cell transfer resulted in long-term cardiac survival (Table 1).

These results demonstrate that CD40Ig-treated recipients contain regulatory cells capable of transferring donor-specific transplantation tolerance, defined as acceptance of a second first-party graft and rejection of a third-party one. This transplantation tolerance was infectious and acted through dominant suppression of antiallogeneic immune responses.

Transfer of graft acceptance is dependent on CD8⁺ T cells. Whereas transfer of total splenocytes from CD40Ig-treated recipients uniformly transferred tolerance, T cell or CD8 depletion from the splenocytes abrogated the tolerance transfer (rejection at 14.5 ± 3.5 and 11 ± 0.7 days, respectively) (Figure 1A). In contrast, transfer of CD4-depleted or NK-depleted splenocytes resulted in indefinite graft survival (Figure 1A). To confirm these results, we adoptively transferred positively selected CD40Ig T cells, CD40Ig CD4⁺ T cells, and CD40Ig CD8⁺ T cells. Adoptive transfer of positively selected T cells and CD8⁺ T cells resulted in indefinite allograft survival in all recipients, whereas adoptive transfer of CD4⁺ T cells led to allograft rejection 15 days (Figure 1B). CD8⁺ T cells were routinely more than 97% TCRαβ⁺ and CD8α⁺ and less than 2% CD4⁺ or CD161⁺ (data not shown).

Figure 1

CD8⁺ T cells mediate transfer of transplantation tolerance after CD40Ig treatment. Cells from control rats that had rejected their grafts or from CD40Ig-treated recipients were injected i.v. the day of transplantation into LEW.1A recipients that received LEW.1W heart transplants (day 0) and that were sublethally irradiated (4.5 Gy, at day –1). (A) Total splenocytes (50 × 10⁶; n = 6) or splenocytes depleted of T cells (n = 2), CD4⁺ cells (n = 2), CD8⁺ cells (n = 4), or NK cells (n = 2) were injected. (B) Splenocytes (50 × 10⁶; n = 6), purified T cells (n = 3), CD8⁺ T cells (n = 4), or CD4⁺ T cells (n = 2) were injected. Graft survival was assessed by abdominal palpation of cardiac beating. (C) Grafts were either untreated (n = 9; thick black line), treated with a control mAb (3G8) (n = 6, gray line), transduced with Addl324 (5 × 10¹⁰ IP; n = 9; dotted line), transduced with AdCD40Ig (n = 27, black line with open circles), or transduced with AdCD40Ig and injected with a depleting anti-CD8 mAb (OX8) (n = 12; black line with open triangles) or an anti-MHC class I mAb (OX18) (n = 3; black line with filled diamonds). A group of control animals received the control mAb (3G8) in addition to CD40Ig (n = 6; black line with open diamonds). **P < 0.01, ***P < 0.0001 versus animals injected with AdCD40Ig or AdCD40Ig and 3G8.

Moreover, in vivo depletion of CD8⁺ T cells resulted in allograft rejection in 50% of the recipients (n = 12, P < 0.0001 versus CD40Ig plus control mAb) (Figure 1C), confirming an important role for these cells in the establishment of tolerogenic mechanisms during CD40Ig treatment. This treatment induced a permanent and marked reduction in peripheral blood CD8α⁺ and CD8β⁺ cells as shown by flow cytometry analyses performed every 2 weeks (Supplemental Figure 2) and by histological analysis of rejected cardiac grafts that lacked infiltrating CD8α⁺ and CD8β⁺ cells (data not shown). Administration to CD40Ig-treated recipients of an anti-MHC class I monomorphic mAb abolished long-term allograft survival, indicating that recognition of MHC class I antigens by CD8⁺ cells was implicated in the effect of CD40Ig (Figure 1C). Together, these data demonstrate that CD40Ig-induced donor-specific allograft acceptance is mediated by CD8 T cells.

CD8⁺CD45RC^low T cells are the key regulators induced by CD40Ig treatment. It has recently been shown that the level of CD45RC expression on rat CD8⁺ T cells distinguishes 2 CD8⁺ subsets that differ in their cytokine profiles and function (3). The CD8⁺CD45RC^low subset has been shown to contain Tregs that inhibit graft versus host disease and mixed lymphocyte reactions. To evaluate the contribution of CD8⁺CD45RC^low T cells in CD8⁺-mediated tolerance after CD40Ig treatment, CD8⁺CD45RC^low T cells and CD8⁺CD45RC^high T cells were sorted to a purity of greater than 99% (Figure 2A) and adoptively transferred into naive graft recipients (Figure 2B). Adoptive transfer of 2.5 × 10⁶ CD8⁺CD45RC^low T cells from CD40Ig-treated donors into syngeneic recipients resulted in indefinite allograft survival, whereas transfer of the same number of CD8⁺CD45RC^high T cells failed to inhibit acute rejection (Figure 2B). Adoptive transfer of CD8⁺CD45RC^low T cells from naive nongrafted animals in the same conditions did not inhibit acute rejection (Figure 2B).

Figure 2

CD8⁺CD45RC^low T cells mediate transfer of transplantation tolerance after CD40Ig treatment. (A) CD8⁺CD45RC^low and CD8⁺CD45RC^high T cells were sorted to greater than 99% purity from spleens. (B) Cells were purified from naive, CD40Ig-treated, or adoptively transferred recipients 120 days after transplantation. A total of 2.5 × 10⁶ cells of each cell population was injected i.v. the day of transplantation of LEW.1W hearts into naive LEW.1A recipients sublethally irradiated (4.5 Gy) at day –1. Graft survival was assessed by abdominal palpation of cardiac beating. Black line, CD8⁺CD45RC^low T cells from CD40Ig-treated animals (n = 4). Dotted line, CD8⁺CD45RC^low T cells from adoptively transferred animals (n = 2). Gray line, CD8⁺CD45RC^high T cells from CD40Ig-treated animals (n = 4). Dashed gray line, CD8⁺CD45RC^low T cells from naive animals (n = 4). P < 0.01 for CD8⁺CD45RC^low from CD40Ig-treated animals versus CD8⁺CD45RC^high and CD8⁺CD45RC^low from naive animals.

Thus, CD40Ig treatment functionally modified CD8⁺CD45RC^low T cells to induce indefinite allograft survival, since CD8⁺CD45RC^high T cells from the same animals or CD8⁺CD45RC^low T cells from naive ones did not inhibit acute rejection.

Analysis of CD8⁺CD45RC^low T cell cytokine profile, Foxp3 expression, and phenotype. Analysis of IFN-γ production by intracellular staining of activated splenocytes showed that CD40Ig treatment or adoptive transfer of splenocytes significantly increased the frequency of IFN-γ–producing cells compared with that found in naive rats (14.1% ± 1.7% and 24.9% ± 1.4% versus 5.1% ± 1.2%, P < 0.05 and P < 0.0001, respectively) and untreated recipients (8.9% ± 1.1%) (Figure 3A). Moreover, a significantly higher percentage of IFN-γ–producing cells was detected in the CD8⁺CD45RC^low T cell population of CD40Ig-treated rats compared with the CD8⁺CD45RC^high population (36.4% ± 11.6% versus 7.1% ± 4.4%, P = 0.001). IL-10, TGF-β, IL-13, and IL-4 were either low or barely detectable (data not shown).

Figure 3

Expression of IFN-γ and regulatory molecules by CD8⁺CD45RC^low T cells. (A) Intracytoplasmic analysis of IFN-γ production in total splenocytes and in CD8⁺CD45RC^low and CD8⁺CD45RC^high cells from CD40Ig-treated or adoptively transferred animals. Cells were harvested, stimulated for 7 hours using phorbol myristate acetate and ionomycin, and analyzed with a FITC-labeled anti–IFN-γ mAb. Results are expressed as percent of positive cells. Each symbol indicates a value for an individual animals. Lines indicate mean ± SD. ATS, adoptively transferred. *P < 0.05, **P < 0.01, and ***P < 0.001. (B) Quantitative analysis of transcript accumulation was performed after isolation in CD8⁺CD45RC^low or CD8⁺CD45RC^high T cells isolated from CD40Ig-treated recipients or in CD8⁺CD45RC^low cells from naive animals. Results are expressed in arbitrary units of molecules normalized to HPRT ± SD. GITR, glucocorticoid-induced tumor necrosis factor receptor family–related gene. *P < 0.05, **P < 0.01, and ***P < 0.001.

Analysis of mRNA expression by CD8⁺CD45RC^low T cells from CD40Ig-treated animals revealed significantly higher amounts of Foxp3, glucocorticoid-induced tumor necrosis factor receptor family–related gene (GITR), IFN-γ, IL-13, IL-10, TGF-β, and IL-2 as compared with amounts found in CD8⁺CD45RC^high T cells, whereas CTLA4, heme oxygenase-1 (HO-1), IL-4, CD28, and perforin were expressed at comparable amounts. IDO was undetectable in both groups (Figure 3B). CD8⁺CD45RC^low T cells from CD40Ig-treated animals versus naive animals revealed significantly higher amounts of GITR, TGF-β, IL-2, and CD25, whereas CD28 was decreased (Figure 3B).

The majority of CD8⁺CD45RC^low T cells from naive, Addl324-treated, and AdCD40Ig-treated recipients were negative for CD4 (95%), CD28 (90%), OX40 (93%), CD103 (90%), CD161 (80%), CD11b (90%), CD11c (97%), MHC class II (80%), and CD25 (90%), whereas they were positive for CD62L (40%), CD8β (95%), and TCRαβ (96%) (Supplemental Figure 3 and data not shown).

These data demonstrate that CD40Ig-induced allograft acceptance is mediated by a CD8⁺CD45RC^low T cell subset characterized by production of IFN-γ and regulatory cytokines that expresses several markers that characterize regulatory cells.

Selected usage of TCR Vβ11 family in CD8⁺CD45RC^low T cells from CD40Ig-treated animals. The length distribution of the TCR β-chain complementary determining region 3 (CDR3) hypervariable region is gaussian in naive T cells and is altered in activated T cell responses. Qualitative analysis of the TCR Vβ families’ repertoire of spleen CD8⁺CD45RC^low T cells from CD40Ig-treated recipients showed a significantly altered distribution of the Vβ11 family (a mean alteration ± SD of 46.2% ± 14.7%), with a predominant peak of the same CDR3 length in all animals (Figure 4) and a normal gaussian-type distribution for all other Vβ families (data not shown). Compared with naive animals, Vβ11 family transcript accumulation was increased 11-fold in spleen CD8⁺CD45RC^low T cells in CD40Ig-treated animals. As controls, spleen CD8⁺CD45RC^low T cells from naive animals or CD8⁺CD45RC^high T cells from CD40Ig-treated animals showed unaltered gaussian-type Vβ11 family distribution (Figure 4) (9.85% ± 1% and 11.4% ± 1.5%, respectively; P < 0.01 compared with CD8⁺CD45RC^low T cells from CD40Ig-treated animals).

Figure 4

Analysis of the TCR repertoire in CD8⁺CD45RC^low T cells. CD8⁺CD45RC^low or CD8⁺CD45RC^high T cells were purified from spleen of CD40Ig-treated or naive animals. A qualitative and quantitative analysis of the TCR Vβ transcriptome was performed as described in Methods. Each histogram represents a qualitative immunoscope analysis of the Vβ11 family repertoire of 1 animal. The histograms display the intensity of fluorescence in arbitrary units as a function of runoff CDR3 Vβ11 length in nucleotides. The CDR3 length distribution was unaltered (gaussian) in CD8⁺CD45RC^low from all naive animals or CD8⁺CD45RC^high from CD40Ig-treated recipients and altered (non-gaussian) in spleen CD8⁺CD45RC^low T cells from all CD40Ig-treated recipients with a predominant CDR3 of always the same length (27 nucleotides, arrows).

That CD8⁺CD45RC^low Tregs from CD40Ig-treated recipients express a selected TCR Vβ11 gene family repertoire demonstrates an induced immune response and suggests public TCR responses to common donor antigen(s).

Following CD40Ig treatment, long-surviving grafts showed strong upregulation of regulatory molecule transcripts and accumulation of CD8⁺IFN-γ⁺ cells. Quantitative RT-PCR analysis showed that, compared with syngeneic grafts, long-surviving grafts from CD40Ig-treated recipients displayed a 50- to 100-fold increase in Foxp3 and IDO mRNA, a 5- to 10-fold increase in PIR-B (the rat homolog of human ILT4), CTLA4, HO-1, a trend toward an increase in IFN-γ (P = 0.052), and the TCR β-chain, perforin, and IL-2. No significant differences were observed for TGF-β, IL-6, or IL-10 (Figure 5A). Long-surviving grafts from adoptively transferred recipients showed 20- and 100-fold increased accumulation of IFN-γ and IDO, respectively (Supplemental Figure 4), indicating that mRNA accumulation for regulatory molecules was a specific tolerogenic phenomenon independent of adenoviral gene transfer into the graft.

Figure 5

Grafts accepted long term after CD40Ig treatment showed strong accumulation of transcripts for regulatory molecules. Heart grafts from syngeneic donors or from allogeneic donors were harvested 120 days after transplantation. Recipients from allogeneic donors were treated with CD40Ig. (A) Heart total RNA was analyzed using real-time quantitative RT-PCR. Symbols indicate values for individual animals in arbitrary units. *P < 0.05, and **P < 0.01 versus syngeneic animals. (B) Confocal analysis of allogeneic or syngeneic (inserts) grafts using anti-CD8α (red) or anti–IFN-γ (green) antibodies. Merge analysis shows IFN-γ expression almost exclusively restricted to CD8⁺ cells (arrows), whereas some cells are IFN-γ^– (arrowheads). Identical results were obtained with 3 grafts from CD40Ig-treated animals and from adoptively transferred grafts. Original magnification, ×400.

Immunohistology analysis showed the presence of abundant CD8⁺ cells expressing IFN-γ in CD40Ig long-surviving grafts but not in syngeneic grafts (Figure 5B). Thus, long-surviving grafts display CD8⁺ IFN-γ⁺ cells and increased expression of molecules with tolerogenic or antiinflammatory activities, such as PIR-B, HO-1 and IDO, that may create local tolerogenic mechanisms.

IFN-γ and IDO mediated the effects of CD40Ig and CD8⁺CD45RC^low T cells. To analyze the contribution of IFN-γ to CD40Ig-induced donor-specific allograft acceptance mediated by CD8⁺CD45RC^low T cells, we treated recipients with a neutralizing anti–IFN-γ mAb from the day of transplantation and CD8⁺CD45RC^low transfer. This treatment abrogated long-term allograft survival in 3 of 4 recipients (Figure 6A). The remaining cardiac graft displayed reduced beating strength (from +++ to +). Administration of anti–IFN-γ alone did not modify the kinetics of acute rejection (data not shown).

Figure 6

IFN-γ and IDO mediate the effect of CD40Ig and of CD8⁺CD45RC^low T cells. (A) Graft survival after 1-MT or anti–IFN-γ mAb administration beginning the day of transplantation (d0) to CD40Ig-treated animals or to animals receiving adoptive transfers at 120 days after transplantation (d120). Heartbeat of nonrejected grafts after 1-MT or anti–IFN-γ administration was + compared with +++ in CD40Ig-treated or adoptively transferred recipients. Arrow indicates adoptive transfer and beginning of treatment with 1-MT. (B) Representative Western blot analysis of IDO expression in cardiac grafts from syngeneic, CD40Ig-treated, and adoptively transferred recipients at day 120. Histograms show quantification of IDO normalized to tubulin in 3–4 samples per group. **P < 0.01, syngeneic versus adoptively transferred grafts. (C) Confocal analysis of grafts from adoptively transferred or syngeneic recipients using anti-IDO (green) and anti-CD31 (red) antibodies. Merge analysis shows IDO expression almost exclusively restricted to ECs. The inserts represent staining using an irrelevant mAb (3G8) or rabbit serum. Identical results were obtained with 2 other grafts from adoptively transferred animals and 2 grafts from CD40Ig-treated animals. Original magnification, ×400. (D) Kynurenine in cultured supernatants. CD8⁺CD45RC^low or CD8⁺CD45RC^high T cells from CD40Ig-treated recipients (n = 2 in each group) were cultured for 2 days with ECs from LEW.1W animals in the presence of anti–IFN-γ or control mAb (3G8) at 20 μg/ml. Levels are expressed as μM ± SD. *P < 0.05. (E) Quantitative RT-PCR analysis of IDO mRNA in ECs cultured for 2 days with CD8⁺CD45RC^low or CD8⁺CD45RC^high T cells from CD40Ig-treated recipients (n = 3 in each group). **P < 0.01 for CD8⁺CD45RC^low versus CD8⁺CD45RC^high.

Since IFN-γ is a major stimulus for IDO expression (28) and IDO was among the genes with the highest expression in the graft, we evaluated the role of IDO by pharmacologically blocking its activity with 1-methyl tryptophan (1-MT). Administration of 1-MT from the day of transplantation together with CD40Ig treatment at day 0 or 120 after adoptive transfer of CD8⁺CD45RC^low T cells resulted in rapid graft rejection in the majority of treated animals (Figure 6A).

Western blot analysis showed higher amounts of IDO in grafts from CD40Ig-treated and adoptively transferred recipients compared with the amounts shown in syngeneic grafts (Figure 6B). Higher amounts of IDO were also found in spleen of transplanted rats treated with adoptively transferred cells compared with spleen of rats with rejected grafts (Supplemental Figure 5). Immunohistologic analysis showed that ECs were the main cellular source of IDO within the grafts (Figure 6C).

To prove that Tregs generated by CD40Ig treatment could induce ECs to produce IDO, we performed in vitro coculture of donor ECs with CD8⁺CD45RC^low or CD8⁺CD45RC^high T cells. IDO activity, as measured by kynurenine accumulation in the culture supernatant, was detected when ECs were cocultured with CD8⁺CD45RC^low but not CD8⁺CD45RC^high T cells (Figure 6D). Furthermore, induction of IDO expression by ECs was dependent on IFN-γ production by CD8⁺CD45RC^low T cells, since kynurenine production was abrogated upon neutralization of IFN-γ with an mAb (Figure 6D). Analysis of IDO mRNA accumulation in ECs confirmed induction of IDO expression by recombinant IFN-γ used as a positive control (data not shown) as well as by CD8⁺CD45RC^low but not CD8⁺CD45RC^high T cells (Figure 6E). CD8⁺CD45RC^low T cells from CD40Ig-treated animals only induced IDO expression in ECs from graft donor LEW.1W origin but not from syngeneic LEW.1A ECs (data not shown).

Therefore, IDO is necessary for both the early induction of long-term allograft survival by CD40Ig and for the maintenance phase of tolerance mediated by CD8⁺CD45RC^low T cells. Infectious tolerance by CD8⁺CD45RC^low T cells was also dependent on IFN-γ, and CD8⁺CD45RC^low T cells induced IDO in ECs through IFN-γ production.

IDO overexpression in the graft but not at a distant site inhibits allograft rejection. Since IDO expression in the graft was associated with prolonged survival, we generated a recombinant adenovirus coding for IDO and tested the effect of IDO overexpression in the graft or at a distant site such as the hind limb. When IDO gene transfer was performed in the graft, survival was significantly prolonged, with 45% of the grafts surviving more than 90 days, whereas survival was not significantly prolonged when gene transfer was performed in the hind limb (Figure 7). Although 1-MT blocks the effect of CTLA4Ig treatment in a model of islet allotransplantation in mice (29), administration of 1-MT to CTLA4Ig-treated rats did not result in graft rejection (Supplemental Figure 6), indicating that the effects on CD40Ig-treated animals were specific. These results confirmed that IDO overexpression in the graft is sufficient to induce permanent allograft survival, and although IDO gene transfer in the graft was less efficient than transfer of CD8⁺CD45RC^low T cells, this can be explained by a transient expression of IDO when using adenoviruses.

Figure 7

IDO gene transfer into the graft but not at a distant site results in long-term allograft survival. AdIDO or Addl324 (2 × 10¹⁰ IP) were injected into the cardiac grafts or intramuscularly into the hind limb the day of transplantation. Rejection was defined when beating of the transplanted heart could no longer be detected. P < 0.001, AdIDO graft versus Addl324 and AdIDO i.m.

The major findings of the current study were that long-term allograft survival induced by CD40Ig was dependent on CD8⁺CD45RC^low Tregs, that IFN-γ and IDO were necessary for long-term allograft survival, and that interaction between CD8⁺CD45RC^low T cells and graft ECs possibly underlines these mechanisms. The allograft acceptance induced by CD40Ig treatment was obtained in a major MHC-mismatched cardiac allograft model, was able to be adoptively transferred by CD8⁺CD45RC^low T cells into euthymic transplanted recipients, and was donor specific.

Infectious allogeneic tolerance was first demonstrated in thymectomized mice treated with anti-CD4 and anti-CD8 mAbs, and tolerance was serially transferred using splenocytes or CD4⁺ but not CD8⁺ cells (1). The same group showed that anti-CD40L and anti-CD8 mAb treatment also resulted in infectious tolerance mediated by CD4⁺ T cells (30). Infectious tolerance by CD4⁺ T cells has also been described in rats following donor-specific blood transfusion (DST) (16) and anti-CD4 mAb treatment (15). In all these models, only CD4⁺ T cells transferred tolerance or were much more potent than CD8⁺ T cells (16). Thus, our model of CD40Ig treatment is the first in which infectious tolerance is mediated by CD8⁺ T cells. Furthermore, the potency of these cells is high, since as few as 50 × 10⁶ splenocytes or 2.5 × 10⁶ CD8⁺CD45RC^low T cells were able to inhibit acute rejection in an MHC-incompatible strain combination for at least 4 generations without loss of efficacy. It is unclear why CD4⁺ Tregs were absent in our model but predominate in mouse models using anti-CD40L mAbs. Some differences in these models, such as different molecules (CD40Ig versus anti-CD40L), the continuous presence of CD40Ig versus only inductive treatment with anti-CD40L, and the species used may explain this discrepancy. CD40Ig also induced indefinite survival of allogeneic livers (31), but the role of CD4⁺ or CD8⁺ Tregs was not analyzed. Anti-CD40L mAbs have been proposed to act at least in part through depletion of activated CD4⁺ T cells (22, 23). Absolute numbers of CD4⁺ T cells were decreased in CD40Ig-treated recipients compared with numbers of CD4⁺ T cells in grafted untreated or adoptively transferred recipients but were comparable to those of naive animals (Supplemental Figure 7A). CD8⁺, CD8⁺CD45RC^low, and CD4⁺CD25^high T cells were not modified in CD40Ig-treated animals versus naive or control Addl324-treated recipients, whereas CD8⁺CD45RC^low T cells were increased in adoptively transferred recipients versus naive or CD40Ig-treated recipients. CD40Ig also had a differential effect in vitro in CD4⁺ and CD8⁺ T cells, since it showed higher inhibition of MLR responses by CD4⁺CD25^– naive T cells than CD8⁺CD45RC^high or CD8⁺CD45RC^low cells (Supplemental Figure 7B). These results suggest a specific differential effect of CD40Ig on CD4⁺ T cells compared with CD8⁺ T cells in grafted animals. This could be related to higher levels of CD40L expression on rat CD4⁺ versus CD8⁺ T cells (data not shown). These results also suggest that defective alloantigen presentation due to blockade of CD40-CD40L interactions and decrease of CD4⁺ T cells may participate in the generation of CD8⁺CD45RC^low cells.

CD8⁺ Tregs have been described in several models (3–14). In allotransplantation models, CD8⁺CD28^– Tregs were induced by in vitro allogeneic stimulation of human cells (11) or by DST treatment in rats (4). Induction of the immunoreceptor tyrosine-based inhibitory motifs–containing motifs ILT3 and ILT4 on human DCs (11) or of PIR-B on graft rat ECs (4) have been described. The CD8⁺Foxp3⁺ cells induced by DST did not produce IFN-γ and were able to transfer tolerance in only 60% of primary recipients (4). Thus, the CD8⁺ Tregs described in the present study could be a counterpart to the human CD8⁺CD28^– cells but are different from the ones described in DST-treated rats. CD8⁺ T cells tolerogenic for alloantigens have also been described following oral exposure to alloantigens (5) or to DST treatment (13, 14, 16), but their phenotype or cytokine production was not analyzed further. Administration of agonistic anti-CD137 mAbs to mice to suppress autoimmune arthritis through the generation CD8⁺CD11c⁺ Tregs capable of producing IFN-γ was recently described (8). DCs from these mice expressed IDO, and the tolerogenic effect of anti-CD137 mAb was abolished by treatment with anti–IFN-γ or 1-MT. Thus, agonistic anti-CD137 and CD40Ig seem to generate similar mechanisms of tolerance acting through CD8⁺ T cells, IFN-γ, and IDO. It remains to be determined which of the 2 costimulatory pathways is upstream of the other in each model and whether certain phenotypic differences truly distinguish the different cell populations (CD8⁺CD45RC^low T cells are CD11c^–). Finally, a population of naturally occurring CD8⁺CD45RC^low Tregs was recently described in the rat (3). In contrast to the lack of effect of these natural CD8⁺CD45RC^low cells in our cardiac transplantation model, CD40Ig-treated recipients showed CD8⁺CD45RC^low cells able to suppress graft rejection. This indicated that blockade of CD40 signals may have enabled an increase in suppressive function and/or expansion of alloantigen-specific T cells within naturally occurring CD8⁺CD45RC^low T cells, as described for natural versus induced CD4⁺ Tregs in the absence of CD40 signals (26) or following other treatments (32, 33). The presence of a restricted TCR Vβ11 usage by CD8⁺CD45RC^low T cells from CD40Ig-treated recipients and not naive animals suggests the expansion of an antigen-specific population of cells. As previously reported for natural and induced CD4⁺ Tregs (33), transcript levels of several cytokines and markers (IL-10, IFN-γ, and Foxp3) were not significantly different in CD8⁺CD45RC^low T cells from CD40Ig-treated compared with naive animals. Nevertheless, for other molecules analyzed (GITR, TGF-β, IL-2, CD25, and CD28), there were significant differences, and a DNA microarray analysis has shown 127 genes differentially regulated between the 2 populations (B. Le Mauff and I. Anegon, unpublished observations).

Although induction of tolerance to alloantigens in vivo using anti-CD40L (34), or CD25⁺CD4⁺ Tregs (35) has also been shown to require IFN-γ, the role in tolerance of downstream molecules regulated by IFN-γ, such as IDO, was not demonstrated. IFN-γ may act as a double-edged sword, since it may not only inhibit acute rejection but also promote chronic rejection (36). Thus, the production of IFN-γ might explain why most graft tolerance–promoting strategies, including those resulting in infectious tolerance, such as administration of anti-CD40L, result in simultaneous efficient inhibition of acute rejection and development of chronic rejection lesions (37).

Long-surviving grafts showed increased expression of mRNA for PIR-B, HO-1, and IDO. PIR-B is homologous to human ILT4, which is expressed by DCs and ECs and inhibits alloreactive T cell activation (4). HO-1 is a tissue-protective gene which was shown to be necessary for tolerance induction in anti-CD40L and DST treatment (38), but its inhibition in our model by in vivo administration of tin-protoporphyrin did not abrogate CD40Ig-induced tolerance (data not shown).

IDO degrades tryptophan, which can be further degraded through the kynurenine pathway (28), and its activity can induce T cell apoptosis or anergy (28, 39). IDO can be produced by several cell types including DCs and ECs, and IFN-γ is a major stimulus of its functional expression (28, 40). Specific inhibition of IDO using 1-MT has been shown to abrogate tolerance in a number of models (8, 28, 29, 41). In all of these studies, IDO production was analyzed in DCs but not in other cell types, despite the fact that ECs also express IDO (40). Our results demonstrate, for what we believe to be the first time, that IDO has a pivotal role in long-term allograft survival following manipulation of CD40-CD40L interactions. The present results also demonstrate for what we believe to be the first time that IDO can be produced by graft ECs and that Tregs can induce the expression of IDO in ECs. Nevertheless, since we have also documented increased IDO expression in spleens of CD40Ig-treated animals, the relative role of graft ECs versus spleen cells producing IDO requires further study.

The activation of CD8⁺CD45RC^low T cells in CD40Ig-treated animals is donor alloantigen–specific, since BN third-party grafts but not first-donor LEW.1W grafts were rejected upon adoptive transfer. Furthermore, Vβ TCR repertoire alterations also indicate that CD8⁺CD45RC^low T cell responses were selected and likely directed against donor MHC class I antigens, since anti-MHC class I mAb administration abrogated long-term graft survival. Further experiments are needed to define whether CD8⁺CD45RC^low T cells recognize MHC class Ia antigens or Ib molecules, as previously described for CD8⁺ Tregs in mice models of autoimmunity (6). Although triggering of CD8⁺CD45RC^low T cells was donor alloantigen specific, production of IFN-γ and IDO could theoretically result in the nonspecific inhibition of other concomitant immune responses. Nevertheless, we have previously shown that CD40Ig-treated recipients show normal anticognate immune responses after CD40Ig gene transfer at early and late time points after gene transfer (19). It is possible that activation of CD8⁺CD45RC^low T cells by graft ECs localizes the effects of IFN-γ and IDO to the graft. The localization and/or activation of Tregs in target tissues has been previously described in transplant (42) and autoimmune diabetes (43) models. As recently speculated, Tregs may act by inducing an immune privileged site through mechanisms similar to those occurring in natural immune privileged sites such as for IDO in the placenta (44). Alternatively, IFN-γ is directionally secreted into the immunological synapse, whereas other cytokines show multidirectional secretion (45). This could provide another mechanism for a limited effect even in lymphoid secondary organs.

DC–T cell interaction in the absence of CD40L-CD40 results in cross-tolerization of CD8⁺ T cells, whereas DC stimulation through CD40 results in T cell priming (24–26). When the role of IDO was analyzed, it was shown to be essential for cross-tolerization of CD8⁺ T cells by DCs (25). Thus, it is possible that in our model, CD40 blockade resulted in alloantigen presentation by DCs and/or ECs to CD8⁺ T cells in a tolerogenic manner through IDO expression.

In summary, we have provided in vivo evidence of a role for a CD8⁺CD45RC^low T cell population with tolerogenic properties following CD40Ig treatment. We have demonstrated that CD8⁺CD45RC^low T cells are capable of infectious allogeneic tolerance through dominant suppression by expression of IFN-γ, which in turn induces IDO expression.

View Supplemental data

We thank the Vector Core of the University Hospital of Nantes, which is supported by the Association Française Contre les Myopathies, for producing the adenoviral vectors. We thank Emanuel Merieau, Bernard Martinet, Helga Smit, and Séverine Remy for technical assistance. We thank the Fondation Progreffe for financial support and Crédit Agricole for the donation of the FACSAria.

Address correspondence to: Ignacio Anegon, ITERT-INSERM U643, CHU de Nantes, 30 Boulevard Jean Monnet, 44093 Nantes, France. Phone: 33-240-08-74-15; Fax: 33-240-08-74-11; E-mail: Ignacio.Anegon@univ-nantes.fr.

Nonstandard abbreviations used: AdCD40Ig, adenovirus encoding CD40Ig; Addl324, adenovirus dl324 (noncoding); BN, Brown-Norway; CD40L, CD40 ligand; CDR3, complementary determining region 3; DST, donor-specific blood transfusion; HO-1, heme oxygenase-1; IDO, indoleamine 2,3-dioxygenase; IP, infectious particles; 1-MT, 1-methyl tryptophan.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J. Clin. Invest.117:1096–1106 (2007). doi:10.1172/JCI28801

See the related article at Yin and yang interplay of IFN-γ in inflammation and autoimmune disease.

1. Waldmann, H., Cobbold, S. 2004. Exploiting tolerance processes in transplantation. Science. 305:209-212.
   View this article via: PubMed Google Scholar
2. Wood, K.J., Sakaguchi, S. 2003. Regulatory T cells in transplantation tolerance. Nat. Rev. Immunol. 3:199-210.
   View this article via: PubMed Google Scholar
3. Xystrakis, E., et al. 2004. Identification of a novel natural regulatory CD8 T-cell subset and analysis of its mechanism of regulation. Blood. 104:3294-3301.
   View this article via: PubMed Google Scholar
4. Liu, J., et al. 2004. Rat CD8+ FOXP3+ T suppressor cells mediate tolerance to allogeneic heart transplants, inducing PIR-B in APC and rendering the graft invulnerable to rejection. Transpl. Immunol. 13:239-247.
   View this article via: PubMed Google Scholar
5. Zhou, J., Carr, R.I., Liwski, R.S., Stadnyk, A.W., Lee, T.D. 2001. Oral exposure to alloantigen generates intragraft CD8+ regulatory cells. J. Immunol. 167:107-113.
   View this article via: PubMed Google Scholar
6. Jiang, H., Chess, L. 2000. The specific regulation of immune responses by CD8+ T cells restricted by the MHC class Ib molecule, Qa-1. Annu. Rev. Immunol. 18:185-216.
   View this article via: PubMed Google Scholar
7. Dhodapkar, M.V., Steinman, R.M. 2002. Antigen-bearing immature dendritic cells induce peptide-specific CD8(+) regulatory T cells in vivo in humans. Blood. 100:174-177.
   View this article via: PubMed Google Scholar
8. Seo, S.K., et al. 2004. 4-1BB-mediated immunotherapy of rheumatoid arthritis. Nat. Med. 10:1088-1094.
   View this article via: PubMed Google Scholar
9. Bisikirska, B., Colgan, J., Luban, J., Bluestone, J.A., Herold, K.C. 2005. TCR stimulation with modified anti-CD3 mAb expands CD8⁺ T cell population and induces CD8⁺CD25⁺ Tregs. J. Clin. Invest. 115:2904-2913.
   View this article via: JCI PubMed Google Scholar
10. Balashov, K.E., Khoury, S.J., Hafler, D.A., Weiner, H.L. 1995. Inhibition of T cell responses by activated human CD8+ T cells is mediated by interferon-γ and is defective in chronic progressive multiple sclerosis. J. Clin. Invest. 95:2711-2719.
    View this article via: JCI PubMed Google Scholar
11. Chang, C., et al. 2002. Tolerization of dendritic cells by Ts cells: the crucial role of inhibitory ILT3 and ILT4. Nat. Immunol. 3:237-243.
    View this article via: PubMed Google Scholar
12. Najafian, N., et al. 2003. Regulatory functions of CD8⁺CD28^– T cells in an autoimmune disease model. J. Clin. Invest. 112:1037-1048.
    View this article via: JCI PubMed Google Scholar
13. Douillard, P., et al. 1996. Donor-specific blood transfusion induced tolerance in adult rats with a dominant TcR-Vb rearrangement in heart allografts. J. Immunol. 157:1250-1260.
    View this article via: PubMed Google Scholar
14. Oluwole, S.F., Ng, A.K., Reemtsma, K., Hardy, M.A. 1989. The mechanism of the induction of immunologic unresponsiveness to rat cardiac allografts by recipient pretreatment with donor lymphocyte subsets. Transplantation. 48:281-288.
    View this article via: PubMed Google Scholar
15. Onodera, K.L., et al. 1999. Distinct tolerance pathways in sensitized allograft recipients after selective blockade of activation signal 1 or signal 2. Transplantation. 68:288-293.
    View this article via: PubMed Google Scholar
16. Kataoka, M., Margenthaler, J.A., Ku, G., Flye, M.W. 2003. Development of infectious tolerance after donor-specific transfusion and rat heart transplantation. J. Immunol. 171:204-211.
    View this article via: PubMed Google Scholar
17. Larsen, C.P., et al. 1996. Long-term acceptance of skin and cardiac allografts after blocking CD40 and CD28 pathways. Nature. 381:434-438.
    View this article via: PubMed Google Scholar
18. Adams, A.B., et al. 2005. Development of a chimeric anti-CD40 monoclonal antibody that synergizes with LEA29Y to prolong islet allograft survival. J. Immunol. 174:542-550.
    View this article via: PubMed Google Scholar
19. Guillot, C., et al. 2002. Prolonged blockade of CD40-CD40 ligand interactions by gene transfer of CD40Ig results in long-term heart allograft survival and donor-specific hyporesponsiveness, but does not prevent chronic rejection. J. Immunol. 168:1600-1609.
    View this article via: PubMed Google Scholar
20. Quezada, S.A., Jarvinen, L.Z., Lind, E.F., Noelle, R.J. 2004. CD40/CD154 interactions at the interface of tolerance and immunity. Annu. Rev. Immunol. 22:307-328.
    View this article via: PubMed Google Scholar
21. Honey, K., Cobbold, S., Waldmann, H. 1999. CD40 ligand blockade induces CD4+ T cell tolerance and linked suppression. J. Immunol. 163:4805-4810.
    View this article via: PubMed Google Scholar
22. Monk, N.J., et al. 2003. Fc-dependent depletion of activated T cells occurs through CD40L-specific antibody rather than costimulation blockade. Nat. Med. 9:1275-1280.
    View this article via: PubMed Google Scholar
23. Blair, P.J., et al. 2000. CD40 ligand (CD154) triggers a short-term CD4(+) T cell activation response that results in secretion of immunomodulatory cytokines and apoptosis. J. Exp. Med. 191:651-660.
    View this article via: PubMed Google Scholar
24. Albert, M., Jegathesan, M., Darnell, R. 2001. Dendritic cell maturation is required for the cross-tolerization of CD8+ cells. Nat. Immunol. 2:1010-1017.
    View this article via: PubMed Google Scholar
25. Grohmann, U., et al. 2001. CD40 ligation ablates the tolerogenic potential of lymphoid dendritic cells. J. Immunol. 166:277-283.
    View this article via: PubMed Google Scholar
26. Martin, E., O’Sullivan, B., Low, P., Thomas, R. 2003. Antigen-specific suppression of a primed immune response by dendritic cells mediated by regulatory T cells secreting interleukin-10. Immunity. 18:155-167.
    View this article via: PubMed Google Scholar
27. Homann, D., et al. 2002. CD40L blockade prevents autoimmune diabetes by induction of bitypic NK/DC regulatory cells. Immunity. 16:403-415.
    View this article via: PubMed Google Scholar
28. Mellor, A.L., Munn, D.H. 2004. IDO expression by dendritic cells: tolerance and tryptophan catabolism. Nat. Rev. Immunol. 4:762-774.
    View this article via: PubMed Google Scholar
29. Grohmann, U., et al. 2002. CTLA-4-Ig regulates tryptophan catabolism in vivo. Nat. Immunol. 3:1097-1101.
    View this article via: PubMed Google Scholar
30. Graca, L., Honey, K., Adams, E., Cobbold, S., Waldmann, H. 2000. Anti-CD154 therapeutic antibodies induce infectious transplantation tolerance. J. Immunol. 165:4783-4786.
    View this article via: PubMed Google Scholar
31. Nomura, M., et al. 2002. Induction of donor-specific tolerance by adenovirus-mediated CD40Ig gene therapy in rat liver transplantation. Transplantation. 73:1403-1410.
    View this article via: PubMed Google Scholar
32. Gregori, S., et al. 2001. Regulatory T cells induced by 1 alpha,25-dihydroxyvitamin D3 and mycophenolate mofetil treatment mediate transplantation tolerance. J. Immunol. 167:1945-1953.
    View this article via: PubMed Google Scholar
33. Chiffoleau, E., et al. 2002. Role for thymic and splenic regulatory CD4+ T cells induced by donor dendritic cells in allograft tolerance by LF15-0195 treatment. J. Immunol. 168:5058-5069.
    View this article via: PubMed Google Scholar
34. Konieczny, B.T., et al. 1998. IFN-gamma is critical for long-term allograft survival induced by blocking the CD28 and CD40 ligand T cell costimulation pathways. J. Immunol. 160:2059-2064.
    View this article via: PubMed Google Scholar
35. Sawitzki, B., et al. 2005. IFN-gamma production by alloantigen-reactive regulatory T cells is important for their regulatory function in vivo. J. Exp. Med. 201:1925-1935.
    View this article via: PubMed Google Scholar
36. Tellides, G., et al. 2000. Interferon-gamma elicits arteriosclerosis in the absence of leukocytes. Nature. 403:207-211.
    View this article via: PubMed Google Scholar
37. Shirasugi, N., et al. 2002. Prevention of chronic rejection in murine cardiac allografts: a comparison of chimerism- and nonchimerism-inducing costimulation blockade-based tolerance induction regimens. J. Immunol. 169:2677-2684.
    View this article via: PubMed Google Scholar
38. Yamashita, K., et al. 2006. Heme oxygenase-1 is essential for and promotes tolerance to transplanted organs. FASEB J. 20:776-778.
    View this article via: PubMed Google Scholar
39. Munn, D.H., et al. 2005. GCN2 kinase in T cells mediates proliferative arrest and anergy induction in response to indoleamine 2,3-dioxygenase. Immunity. 22:633-642.
    View this article via: PubMed Google Scholar
40. Beutelspacher, S.C., et al. 2006. Expression of indoleamine 2,3-dioxygenase (IDO) by endothelial cells: implications for the control of alloresponses. Am. J. Transplant. 6:1320-1330.
    View this article via: PubMed Google Scholar
41. Munn, D.H., Mellor, A.L. 2004. IDO and tolerance to tumors. Trends Mol. Med. 10:15-18.
    View this article via: PubMed Google Scholar
42. Lee, I., et al. 2005. Recruitment of Foxp3+ T regulatory cells mediating allograft tolerance depends on the CCR4 chemokine receptor. J. Exp. Med. 201:1037-1044.
    View this article via: PubMed Google Scholar
43. Chen, Z., Herman, A.E., Matos, M., Mathis, D., Benoist, C. 2005. Where CD4+CD25+ T reg cells impinge on autoimmune diabetes. J. Exp. Med. 202:1387-1397.
    View this article via: PubMed Google Scholar
44. Waldmann, H., et al. 2004. Regulatory T cells and organ transplantation. Semin. Immunol. 16:119-126.
    View this article via: PubMed Google Scholar
45. Huse, M., Lillemeier, B.F., Kuhns, M.S., Chen, D.S., Davis, M.M. 2006. T cells use two directionally distinct pathways for cytokine secretion. Nat. Immunol. 7:247-255.
    View this article via: PubMed Google Scholar
46. Braudeau, C., et al. 2004. Induction of long-term cardiac allograft survival by heme oxygenase-1 gene transfer. Gene Ther. 11:701-710.
    View this article via: PubMed Google Scholar
47. Hill, M., et al. 2005. Heme oxygenase-1 inhibits rat and human breast cancer cells proliferation: mutual cross-inhibition with indoleamine 2,3-dioxygenase. FASEB J. 19:1957-1968.
    View this article via: PubMed Google Scholar
48. Guillot, C., et al. 2000. Tolerance to cardiac allografts via local and systemic mechanisms after adenovirus-mediated CTLA4Ig expression. J. Immunol. 164:5258-5268.
    View this article via: PubMed Google Scholar
49. Haspot, F., et al. 2005. Anti-CD28 antibody-induced kidney allograft tolerance related to tryptophan degradation and TCR class II B7 regulatory cells. Am. J. Transplant. 5:2339-2348..
    View this article via: PubMed CrossRef Google Scholar