Ethylenecarbodiimide-Treated Splenocytes Carrying Male CD4 Epitopes Confer Hya Transplant Protection by inhibiting CD154 upregulation

Abstract

In humans and certain strains of laboratory mice, male tissue is recognized as non-self and destroyed by the female immune system via recognition of histocompatibility-Y chromosome-encoded antigens (Hya). Male tissue destruction is thought to be accomplished by cytotoxic T lymphocytes in a helper-dependent manner. We show that graft protection induced with the immunodominant Hya-encoded CD4 epitope (Dby) attached to female splenic leukocytes (Dby-SP) with the chemical cross-linker ethylene-carbodiimide (ECDI) significantly, and often indefinitely, prolongs the survival of male skin graft transplants in an antigen-specific manner. In contrast, treatments with the Hya CD8 epitopes (Uty- and Smcy-SP) failed to prolong graft survival. Dby-SP-tolerized CD4⁺ T cells fail to proliferate, secrete IFN-γ, or effectively prime a CD8 response in recipients of male grafts. Ag-SP treatment is associated with defective CD40-CD40L interactions as evidenced by the observation that CD4 cells from treated animals exhibit a defect in CD40L upregulation following in vitro antigen challenge. Furthermore, treatment with an agonistic anti-CD40 antibody at the time of transplantation abrogates protection from graft rejection. Interestingly, anti-CD40 treatment completely restores the function of Dby-specific CD4 cells, but not Uty- or Smcy-specific CD8 cells.

Introduction

Compared to depletion of T cells, specific inactivation of only select T cell populations is favorable for the treatment of immune-mediated conditions, such as autoimmunity and transplant rejection (1). The advantages of antigen-specific therapy lie primarily in avoiding the risk of opportunistic infections, neoplasia, and toxicity associated with current-generation immunosuppressants (2, 3). We have demonstrated that peptides attached to the surface of syngeneic splenic leukocytes (Ag-SP) with the chemical cross-linking agent ethylene-carbodiimide (ECDI)² effectively and safely induce antigen-specific immune tolerance (4). Ag-SP tolerance has been shown to prevent and treat the symptoms of experimental autoimmune encephalomyelitis (EAE) (5), Type-1 diabetes (T1D) (6), and allogeneic pancreatic islet transplant rejection (7) in the absence of non-specific immunosuppressive drugs. The involvement of clonal anergy, T_reg populations, co-stimulatory molecule blockade, and activation of negative co-stimulatory molecules such as PD-1 and CTLA-4 have been described as potential mechanisms of Ag-SP tolerance (6-9), but the precise effects on antigen-specific T cells following tolerogen encounter remain unclear.

Previous EAE studies induced tolerance using splenic leukocytes coupled to a variety of CD4 T cell epitopes (10), and although exquisite specificity for inactivating pathogenic CD4⁺ T cells was achieved resulting in effective disease therapy, the impact of Ag-SP therapy on CD8 T cells remained unclear. Conversely, CD8⁺ T cells play a prominent role in viral responses and allograft rejection, and Ag-SP tolerance can effectively control inflammation and/or tissue destruction in both models (7, 11). Our laboratory has suggested that the CD8 compartment can be functionally inactivated using Ag-SP for the prevention of allogeneic islet transplant rejection (7), but the complexity of that system has made it difficult to determine if tolerance is primarily induced directly in the CD8 compartment and/or if tolerant CD4⁺ T cells simply fail to prime a CD8 response.

Priming of CD8⁺ T cells can occur dependently or independently of CD4⁺ T cell help (12-17). Helper-independent CTL responses typically occur in the context of acute inflammatory reactions associated with signals mediated by recognition of Toll-like receptor ligands found in intracellular bacterial or viral pathogens (18-20). In both helper-dependent and independent responses, a determining outcome is the upregulation of costimulatory molecules on the surface of antigen-presenting cells. In the former case, this is accomplished by interactions between activated CD4⁺ T cells and antigen presenting cells (APCs), and in the latter case, APCs upregulate these molecules secondary to TLR ligand encounter (16, 17). In the absence of potent inflammatory cues, efficient CD8 responses are critically dependent upon CD4 T cells for acquisition of their effector function (16, 17). Several diffusible factors (IL-2 and IFN-γ) as well as surface-bound receptor-ligand pairs influence the priming of CTL responses by CD4⁺ T_H cells. Prominent among such receptor-ligand pairs within the TNF-TNFR superfamily are CD40 and its ligand CD154 [reviewed in (21)]. CD40 is constitutively expressed by B cells and the majority of APCs (22). CD40 ligation by CD154, expressed primarily by activated CD4⁺ T cells, results in an increase in expression of B7-family costimulatory molecules and pro-inflammatory cytokines (23, 24). Such “licensing” of APCs by T_H cells enables the priming of naïve CD8⁺ T cells and induction of productive and long-lasting cytolytic immune responses (13).

Using the well-characterized histocompatibility-Y antigen (Hya) model of chronic graft rejection (25), we show that Ag-SP ECDI-coupled to the immunodominant CD4 epitope (Dby) significantly prolongs skin graft survival, while Ag-SP coupled to the immunodominant CD8 epitopes (Uty and Smcy) (26) does not confer graft protection. Ag-SP carrying the CD4 epitope led to a failure in the priming of CTL responses to the Hya CD8 epitopes as determined by significantly suppressed CD8 activation and expansion, IFN-γ production and cytolytic activity. While others have previously observed the failure of Uty and Smcy-specific lysis in Dby-tolerant animals, CD4⁺ regulatory T cells were postulated to provide an active suppressive signal (27). Here, we report that secondary to encounter with ECDI-coupled cells, CD4⁺ T cells fail to prime a CTL response. We show that tolerized CD4⁺ T cells exhibit defective upregulation of CD154 (CD40L) and that cross-linking of CD40 in vivo with an agonistic CD40 mAb restores graft rejection in Dby-SP treated animals. Notably, treatment of protected animals with anti-CD40 resulted in a restoration of proliferation, IFN-γ secretion, and graft infiltration by CD4⁺ T cells, but not IFN-γ secretion, lytic activity, or graft infiltration by CD8⁺ CTLs. Collectively, these results indicate that targeting of graft-specific CD4⁺ T cells can suppress CD8⁺ cytolytic responses, resulting in significant protection of skin grafts. This protection was found to be dependent upon defective CD40 stimulation, and the restoration of graft rejection by anti-CD40 treatment of protected animals occurs via a CTL-independent mechanism.

Methods and Materials

Mice

Age-matched male and female C57BL/6 mice were purchased from the Jackson Laboratory (Bar Harbor, ME). Homozygous OT-II mice were purchased from the Jackson Laboratory and bred in-house. Mice were housed in the Center for Comparative Medicine in sterile micro-isolator cages with ad libitum access to water and chow.

Tolerance induction

Splenic leukocytes were coupled to antigens as previously described in the presence of 30 mg/ml ethylene-carbodiimide (ECDI) and 1mg/ml peptide (4). For tolerance induction to male antigens in C57BL/6 females, 10⁸ antigen-coupled splenocytes (Ag-SP) were administered i.v. seven days and again three hours prior to engraftment. C57BL/6 females received Ag-SP coupled to either the CD4 epitope found in male antigen (Dby) or coupled to two CD8 epitopes (Uty and Smcy) added at equimolar ratios. Tolerance in OT-II mice was induced by i.v. injection of 10⁸ OVA_323-339-SP 7d prior to assay. Peptides (Dby – NAGFNSNRANSSRSS, Uty – WMHHNMDLI, Smcy – KCSRNRQYL, and OVA_323-339 – ISQAVHAAHAEINEAGR) were obtained from Genemed Synthesis, Inc. (San Antonio, TX).

Skin Grafting

Orthotopic split-thickness tail skin grafting was performed as previously described (28). The grafts were scored by daily visual inspection for edema, pigment loss, and hair loss. Rejection was defined by complete hair loss and greater than 80% pigment loss. Differences in survival times were tested for significance by log-rank tests.

T cell recall assays

Animals were sacrificed at the indicated times post-transplantation. Single-cell suspensions of spleen and draining lymph nodes were explanted into 96-well plates and challenged with Dby, Uty, Smcy, or OVA_323-339 (irrelevant antigen control) peptides (5 to 0.05μM). Anti-CD3 (2C11) stimulation was included as a control for proliferation and cytokine secretion. Cultures were grown in RPMI-1640 supplemented with 10% heat-inactivated FBS, 1% penicillin-streptomycin, 50μM β-mercaptoethanol, and 25mM HEPES buffer for 72h. Proliferation was measured by pulsing with 1μCi of [³H]-thymidine at 48h and harvesting 24h later. Culture supernatants were collected at 72h and cytokine secretion measured by ELISA using anti- IFN-γ clones XMG1.2 (capture) and biotin-R4-6A2 (detection) (eBioscience San Diego, CA), streptavidin-HRP, and enzyme substrate (BioFX, Owing Mills, MD). OVA_323-339 responses in naïve or treated OT-II mice were measured similarly.

In vivo CTL assays

Target cells were labeled with 5μM and 0.5μM concentrations of carboxyfluorescein diacetate (CFDA-SE) (Invitrogen/Molecular Probes, Carlsbad, CA) at room temperature for 8 minutes and quenched in the presence of 20% heat-inactivated FBS for 5 min, allowing distinct identification by flow cytometry (11). Fluorescent labeled cells were counted and mixed at a 1:1 ratio prior to i.v. injection. Targets were loaded with antigens (5μM) for 90 min at 37°C and cytometric analyses of recipient spleens was performed 6 days following i.v. transfer.

Antibodies and Flow Cytometry

Allophycocyanin-conjugated H-2D^b tetramers specific for Uty and Smcy T cell receptors were obtained from Dr. Amy Stout (NIH tetramer facility). Other primary conjugates used in this study include CD44-PE/Cy7 CD62L-APC/Alexa750, CD69-FITC CD127-Biotin, CD40-FITC CD154-PE (CD40L) CD8-eFluor605, CD4-APC, and CD3-PacificBlue or PerCP (eBioscience, San Diego, CA). Live-dead discrimination was performed by VID exclusion (Invitrogen). Detection of biotinylated reagents was accomplished using streptavidin-PE/Cy7 (eBioscience). For analyses of cultured cells, dead cells were excluded from analysis using Violet Dead Cell Stain (Invitrogen/Molecular Probes, Carlsbad, CA). Flow cytometric analyses were carried out using a Becton-Dickinson FACSCanto II (Franklin Lakes, NJ). Data were collected and analyzed using FACSDiVa software. Anti-CD40 (FGK45.5) and isotype control rIgG2a were purchased from Miltenyi Biotec (Auburn, CA).

Skin graft histology

Mouse tails were fixed in 4% paraformaldehyde in PBS overnight at 4°C followed by infiltration with 30% sucrose in PBS overnight at 4 °C. Sections containing grafts were frozen in cryomolds in O.C.T. on dry ice and cryostat sectioned at 10um. Prior to labeling, sections were air dried at RT for at least 30min., rinsed in DH2O to remove O.C.T. and fixed in −20 °C acetone for 10min. Sections were dried at RT for 10min., washed in PBS 3×5min., blocked in 5% normal donkey serum in PBS + 0.1% triton x-100 (PBS+) for 60min. prior to incubation with primary antibodies in PBS+ overnight at 4 °C. After washing in PBS 3×10min., sections were incubated in secondary antibodies in PBS+ for 1hr. at room temperature, washed in PBS 3× 10min., incubated in DAPI for nuclear staining 5min., washed 3×5min. in PBS then coverslipped using hard setting Vectamount (Vector Laboratories, Burlingame, CA). For biotinylated primaries, DAKO peroxidase block was used for 30min. followed by Vector streptavidin/biotin block prior to PBS+ blocking step. Streptavidin/tyramide system (PerkinElmer Life Sciences, Downers Grove, IL) was used to visualize biotinylated primary antibodies. Antibodies used include: anti-mouse CD4 (eBioscience, San Diego, CA) anti-mouse CD8, (Santa Cruz Biotechnology, Santa Cruz, CA) and anti-mouse F4/80, (Invitrogen, Carlsbad, CA). Images were taken on a Leica DM 5000B microscope using ImagePro software.

Results

Treatment with splenocytes ECDI-coupled with the Hya-specific MHC class II-restricted Dby epitope promotes long-term survival of C57BL/6 male skin grafts

Females of mouse strains with the H2^b haplotype [e.g., C57BL/6 (B6)] generate strong cellular immune responses against Hya-disparate tissue, as measured by allograft rejection kinetics, and by the development of Hya-specific DTH and CTL responses (25). Naïve B6 females were treated i.v. with either ECDI-fixed B6 male splenocytes, or with Dby-SP (syngeneic female splenic leukocytes ECDI coupled with the Hya CD4 epitope), or with Uty/Smcy-SP (female splenocytes ECDI coupled with the Hya CD8 epitopes Uty and Smcy) on days -7 and 0 relative to engraftment with male tail skin. Male grafts survived for significantly longer times (p < 0.001) on Dby-SP treated animals (median = 77 days) than on untreated animals (median = 21 days) or Uty/Smcy-SP treated animals (median = 27 days) (Figure 1A). Antigen specificity of Ag-SP therapy was tested by engrafting treated and non-treated females with skin from 3^rd party female C57BL/10 donors. Both non-protected and Dby-SP-treated females rejected B10 grafts with equivalent kinetics (median = 17 days, Figure 1E) indicating that Dby-SP specifically regulates Hya expressing grafts. Furthermore, protection of male B6 grafts is not observed in animals treated with splenocytes coupled to an irrelevant antigen. OVA_323-339-SP treated recipients rejected male grafts at a median of 22 days. Female B6 control grafts were not rejected by any treatment group and survived indefinitely (data not shown).

Figure 1. Hya-specific MHC class II, but not MHC class I, restricted Hya epitopes ECDI coupled to B6 female-derived splenocytes promote long-term, antigen-specific survival of male skin grafts on B6 female recipients.

(A) Naïve C57BL/6 females were treated i.v. with syngeneic female splenic leukocytes ECDI-linked to the CD4 epitope (Dby) or the CD8 Hya epitopes (Uty and Smcy) or not tolerized on days -7 and 0 relative to engraftment with male tail skin grafts (and female control grafts – data not shown). Untreated control mice were included as a baseline for rejection time. Graft survival was monitored visually for 100 days. Male skin graft survival was significantly prolonged (*p<0.001) in female recipients treated with female Dby-SP (filled circles) compared to both untreated (open circles) and Uty/Smcy-SP treated (filled triangles) female recipients. (B) DTH responses of female B6 mice to ear challenge with 10μg soluble Dby were determined at 10, 20 and 60 days post transplantation in naïve, untreated and Dby-SP treated female B6 mice. Ear swelling responses in naïve C57BL/6 mice served as the baseline. Ear swelling responses in Dby-SP treated mice were significantly (*p<0.05) less than those in non-tolerized controls. In vitro recall responses of splenic T cells from untreated and Dby-SP tolerized mice were determined 10 days following transplantation upon stimulation with anti-CD3 (clone 2C11-positive control), OVA_323-339 (negative control), and Dby peptides by [³H]-thymidine incorporation (C) and IFN-γ secretion (D). Proliferative and IFN-γ responses were significantly (*p <0.01) suppressed in Dby-SP tolerized animals upon challenge with either the CD4 (Dby) or CD8 (Uty and Smcy) Hya epitopes. Proliferative responses to Dby could be restored to control levels by the addition of 25 U/ml exogenous IL-2. (E) Antigen-specificity of the induction and effector stages of the regulatory effect was tested by treating female B6 mice i.v. with syngeneic female splenocytes ECDI-coupled with either OVA_323-339 or the Dby peptide at days -7 and 0 relative to engraftment with both male B6 skin and female skin from a C57BL/10 donor. Neither Dby-SP (open circles) nor OVA_323-339-SP (open triangles) treatment significantly prolonged B10 skin graft survival. However, Dby-SP (closed circles), but not OVA_323-339-SP (closed triangles) significantly protected (*p<0.001) male B6 grafts from rejection. Five recipients were included in each group. All data shown represent at least 3 independent experiments.

Dby-SP treatment suppressed Dby-specific CD4 recall responses as determined both by in vivo DTH (examined longitudinally, Figure 1B), and in vitro proliferation (Figure 1C) and IFN-γ production (Figure 1D), assessed 14 days post transplant. These observations are consistent with our previous work demonstrating that Ag-SP tolerance results in antigen-specific decreases in effector CD4 Th1/17 cell responses (1, 4). Confirming the previous work of Jenkins and Schwartz, addition of 25 units/ml of exogenous IL-2 to recall cultures restored [³H]-thymidine uptake by T cells from Dby-SP-treated recipients (Figure 1C) indicating that anergy plays a significant role in the Dby-SP-induced unresponsiveness.

IFN-γ responses to the Hya CD8 epitopes, Uty and Smcy, were suppressed in Dby-SP tolerant animals (Figure 1D). Collectively, these findings indicate that Ag-SP treatment targeting the immunodominant Hya CD4 epitope (Dby) is necessary and sufficient to prolong the survival of male tissue grafts, and that tolerance to the CD4 epitope results in priming failure of Hya-specific CD8 T cell responses. CD8⁺ populations in Uty/Smcy-SP-treated animals also exhibit diminished functional responses. This is discussed in further detail below.

Hya-specific CD8 cells display a naïve phenotype in Dby-SP tolerant animals

To further investigate the effects of CD4 tolerance on the development of CD8 effector responses, we measured the activation and lytic capacity of CD8 T cells specific for Uty and Smcy in the spleens and draining lymph nodes of control and Dby-SP tolerized male graft recipients. H-2D^b tetramers identifying the Uty and Smcy-specific TCRs were used to identify expression of the activation marker CD44 on Hya-specific CD8⁺ populations. Compared to naïve, non-engrafted B6 females, non-tolerant graft recipients contain a distinct population of activated (CD44⁺) Hya-specific CD8⁺ T cells (Figure 2). Dby-SP treatment significantly inhibited the expansion and activation of Hya-specific CTLs, as evidenced by a diminished number and proportion of tetramer⁺CD44⁺ cells in both spleens and graft draining lymph nodes of tolerant animals. Figure 2 displays data collected at day 14 post-transplantation. Similar results were observed 10, 20, or 40 days post-transplantation (data not shown) although day 14 was the peak in Uty- and Smcy-specific CD8 cells in rejecting controls. This finding also coincides with an observed diminution of in vitro proliferative responses against Uty and Smcy following Dby-SP treatment (not shown). To assess the functional lytic capacity of Hya-specific CD8⁺ T cells, we conducted in vivo cytolysis assays using a variety of target pairs differentially labeled with CFSE. Specific targets consisted of male splenocytes, female splenocytes pulsed with CD8 Hya epitopes (Uty and Smcy), or female splenocytes pulsed with all three Hya epitopes. These targets were paired with reference targets consisting of female splenocytes, female splenocytes pulsed with an irrelevant CD8 epitope (TMEV VP2), or female splenocytes pulsed with an irrelevant CD8 epitope (VP2) and an irrelevant CD4 epitope (OVA_323-339), respectively. Consistent with the phenotypic analyses, CD8⁺ cells in Dby-tolerant animals failed to lyse male splenocytes or female splenocytes presenting male antigen (Figure 3). The results from both the tetramer analyses and lysis assays indicate that CD4⁺ T_H cells in Dby-SP-tolerant animals fail to prime effective Hya-specific CTL responses.

Figure 2. Dby-SP-induced nonresponsivenss of Hya-specific CD4⁺ T cells leads to failed priming Hya-specific CD8⁺ T cells specific for the Smcy and Uty epitopes.

The activation frequency (CD44⁺) of Hya Uty- and Smcy-epitope-specific CD8⁺ T cells was determined using MHC Class-I tetramers. (A) Live, single CD8⁺ cells were identified. CD44 expression on Uty-specific CD8⁺ T cells in spleens (B-D) and draining LNs (H-J), and on Smcy-specific CD8⁺ T cells in spleens (E-G) and draining LNs (K-M) of naïve, non-tolerized (No Rx) and i.v. Dby-SP tolerized female B6 mice receiving B6 male tail skin grafts 14 d previously are shown. The total number of tetramer⁺ events is listed on each dot plot. Percentages of activated (CD44⁺) graft-specific CD8 cells, which appear in the top region of the tetramer⁺ box are also listed. Data are representative of three independent experiments.

Figure 3. Dby-SP-induced nonresponsivenss of Hya-specific CD4⁺ T cells results in diminished cytolytic activity of Hya-specific CD8⁺ T cells specific for the Smcy and Uty epitopes.

B6 females were ungrafted (naïve), untreated (No Rx) or i.v. tolerized with Dby-SP prior to engraftment with male skin. 14 days later, in vivo Hya-specific cytolytic activity was determined. Peptide-loaded targets were administered (one specific target and one reference target) and were discernable by differential CFSE labeling. In panels A-D, whole male splenocytes (specific) vs. female splenocytes (reference) were used; in panels E-H, female splenocytes were pulsed with a combination of the Hya Uty and Smcy peptides (specific) or TMEV VP2 (reference); in panels I-J, female splenocytes were pulsed with the combination of the Hya Uty, Smcy, and Dby peptides (specific) or TMEV VP2 and OVA_323-339 (reference). White bars on the graphs (Panels D, H, and L) represent calculated lysis in non-treated graft recipients, black bars represent lysis observed in Dby-SP-treated recipients, and grey bars represent lysis observed in unmanipulated naïve female controls. In vivo cytolytic responses in Dby-SP tolerized mice were significantly (*p<0.01) lower than those in non-tolerized controls. Data are representative of 3 independent experiments.

Ag-SP treatment decreases CD154 upregulation by T_H cells

Due to the failed priming of the antigen-specific CD8 compartment observed in animals tolerized against the immunodominant Hya-specific Dby CD4 epitope, we reasoned that there is possibly a defect in one or more of the mechanisms utilized by CD4⁺ T_H cells to prime Hya-specific CTLs. Previous studies have demonstrated defective IL-2 secretion by T_H cells secondary to Ag-SP treatment (29, 30), and several studies, including the present (Figure 1), have shown a defect in IFN-γ synthesis. Upstream of both of these cytokines is the involvement of CD40/CD154 - a TNF-family receptor-ligand pair that is critical for T cell co-stimulation, licensing of APCs, and Th-dependent activation of CD8⁺ T cells (24). CD40 ligation on APCs by CD154 expressed by activated CD4⁺ T cells increases APC expression of B7-family costimulatory molecules and pro-inflammatory cytokines, enabling the differentiation of naïve CD8⁺ T cells to functional cytolytic effectors [reviewed in (16, 17, 24)]. We therefore measured the ability of CD4⁺ T cells to upregulate surface CD154 (CD40L) following treatment with Ag-SP. OTII TCR transgenic mice were left untreated (naïve), or tolerized via the i.v. injection of 10⁸ OVA_323-339-SP or 10⁸ MBP_84-104-SP. Seven days later, spleens were removed and live CD4⁺ cells were analyzed for CD154 surface expression immediately upon explant and at serial time points (6, 12, 24, 48, and 72h) post stimulation with OVA_323-339 in vitro. Peak expression was observed 6 h post-stimulation. Upon antigen encounter, a significantly lower frequency of CD4⁺CD154⁺ T cells was detected in cultures from OVA_323-339-SP tolerized in comparison to naïve mice or mice tolerized to MBP_84-104-SP, suggesting that Ag-SP encounter results in suboptimal activation and a decreased ability to present CD154 to other leukocytes (Figure 4).

Figure 4. Ag-SP treatment inhibits CD154 upregulation on antigen-specific CD4⁺ T cells upon antigen recall.

B6 OVA_323-339-specific OTII TCR transgenic mice were untreated (naïve) (A&D) or injected i.v. with 10⁸ OVA_323-339-SP (B&E), or 10⁸ MBP_84-104-SP (C&F). Seven days later, splenocytes from these animals were harvested, stained and analyzed for surface expression of CD154 immediately upon explant and again at 6 hours following in vitro re-stimulation with 5μM OVA_323-339. Maximal CD154 expression was observed at 6 h post-culture. FMO (G) = fluorescence minus one control. Data are representative of at least 3 independent experiments. 10,000 CD4⁺ events appear in each dot plot.

CD40 stimulation abrogates the tolerogenic effect of Dby-SP in vivo and in vitro

In order to assess the functional significance of the observed defect in CD40L upregulation, we determined the effect of administration of an agonist mAb specific for CD40 (clone FGK45.5) to Dby-SP-tolerized graft recipients. This clone is known to bypass the requirement for CD4 help by stimulating the upregulation of costimulatory markers on APCs (31). B6 females were tolerized with Dby-SP on days -7 and 0, engrafted with male tail skin on day 0, and were treated i.p. with 100μg of FGK45.5 or control rat IgG2a 24 h after engraftment. Compared to Dby-SP-treated B6 females receiving isotype control antibody, which were significantly protected from rejection of male tail skin grafts (median = 78 days), Dby-SP-treated animals receiving FGK45.5 rejected male grafts (median = 28 days) similarly to non-tolerized controls (median = 19 days) (Figure 5A), indicating that CD40 ligation overcomes the protection afforded by Ag-SP treatment. FG45.5-induced in vivo stimulation of CD40 also reversed the suppression of Dby-specific in vitro proliferation (Figure 5B) and IFN-γ secretion (Figure 5C).

Figure 5. Dby-SP-induced protection of Hya skin grafts is reversed by CD40 cross-linking.

(A) 5-7 untreated (No Rx) and Dby-SP tolerized female B6 mice received male tail skin grafts on day 0. 24 hours later, separate groups of treated mice were injected i.p. with 100μg of IgG2a isotype control antibody or with the agonistic anti-CD40 monoclonal antibody, FGK45.5. Graft survival was monitored by visual inspection for 100 days. The prolonged survival of male skin grafts observed in Dby-SP tolerized recipients treated with isotype control antibody was completely reversed by FGK45.5 treatment, *p<0.001. 10 days following transplantation and antibody treatment, recall responses of splenic T cells from the various treatment groups to in vitro stimulation with the Dby peptide were determined by [³H]-thymidine incorporation (B) and IFN-γ secretion (C). Suppressed proliferative and IFN-γ responses in the Dby-SP tolerized animals were significantly reversed by treatment with the FGK45.5, *p <0.01. Data is representative of at least 3 independent experiments.

Cytometric analyses of Hya tetramer⁺ CD8⁺ cells show that FGK45.5 administration to Dby-SP-tolerized recipients, in comparison to control IgG2a treatment, resulted in a significant enhancement of the frequency and numbers of activated graft-specific Uty- and Smcy-specific T cells (CD8⁺tetramer⁺CD44⁺) reaching levels similar those observed in non-tolerized controls (Figure 6A-G). However, IFN-γ secretion in response to in vitro recall challenge with either the Uty or Smcy CD8 epitopes was not restored following FGK45.5 treatment of tolerized graft recipients (Figure 6H), nor was in vivo CTL function restored in tolerant recipients following CD40 ligation (Figure 6I). Collectively, these data indicate that CD40 ligation overcomes functional Ag-SP-induced graft protection by reversing unresponsiveness in the Hya-specific CD4⁺ compartment. Although anti-CD40 treatment resulted in recovery of control numbers of activated (CD44⁺) Hya-specific CD8⁺ T cells, the effector function of these cells as determined by IFN-γ production and lytic capacity was not restored.

Figure 6. FGK45.5-mediated reversal of Dby-SP-induced protection of Hya skin grafts is not associated with restoration of Hya-specific CD8 T cell IFN-γ production or CTL activity.

Untreated (No Rx) and Dby-SP tolerized female B6 mice received male tail skin grafts on day 0. 24 hours later, separate groups of Dby-SP treated mice were injected i.p. with 100 μg of IgG2a isotype control or with the agonistic anti-CD40 monoclonal antibody, FGK45.5. 10 days post-transplantation, spleens from these treated mice were analyzed for the frequency of activated (CD44⁺), Hya Uty- (A-C) and Smcy-epitope-specific CD8⁺ T cells (D-F) using MHC Class-I tetramers. (G) The total numbers of activated Hya-specific CD8⁺ T cells are plotted. Similar results were observed upon analysis of LN CD8⁺ T cells (not shown). The in vitro INF-γ recall response of splenocytes from the various treatment groups to recall stimulation with the Hya CD4 and CD8 epitopes was determined (H), as was the in vivo lytic activity of to targets pulsed with a combination of the Hya Uty and Smcy CD8 epitopes (target) or TMEV VP2 (reference) (I). Interestingly, the IFN-γ response was restored to stimulation with the CD4 Dby epitope, but neither IFN-γ production by nor the lytic function of CD8 T cells specific for the Uty or Smcy epitopes were restored by CD40 ligation. Data represent 2 independent experiments.

In light of these findings and the observation that Uty/Smcy-SP treatment results in impaired lysis but normal graft rejection (Figure 1, and Figure 6I), we reasoned that Hya-disparate graft destruction can take place despite an impaired CD8 response. This was confirmed by examining histological sections of engrafted skin for infiltrating CD4⁺ and CD8⁺ cells. Male grafts on non-treated rejecting controls contain significant CD4⁺ and CD8⁺ infiltrate while grafts from Dby-SP-treated animals contain greatly reduced CD4 infiltrate and virtually no infiltrating CD8⁺ T cells, consistent with their prolonged survival. Conversely, grafts from animals receiving Dby-SP followed by an FGK45.5 treatment contained CD4⁺ cells but virtually no CD8⁺ cells. A similar observation was made in grafts from Uty/Smcy-treated animals, which were found to contain significant CD4⁺ infiltrate but reduced CD8⁺ infiltrate. This indicates that FGK45.5 treatment restores the ability of CD4⁺ T cells but not CD8⁺ T cells to infiltrate the graft of Dby-SP-treated mice. Together these data support the conclusion that Hya-disparate graft rejection can occur in the context of an impaired (following Uty/Smcy-SP treatment) or completely non-functioning (following Dby-SP + FGK45.5) CTL response.

The effects of FGK45.5 binding on the recipient APC populations were measured by cytometric analysis. Female B6 animals were transplanted with CD45.1 congenic skin 24 hours prior to intraperitoneal injection of FGK45.5 or control rat IgG. 72 hours following transplantation, spleens and lymph nodes were analyzed for APC activation, as determined by increased expression of the B7 family costimulatory molecule CD86. As expected, FGK45.5 treatment broadly activated APCs, as evidenced by the large increase in CD86⁺ cells (Figure 8D). Lineage phenotyping of the CD86⁺ population revealed that the primary activated APC population was B cells (75-80% of the CD86⁺ population, Figure 8E), although a significant expansion and activation of dendritic cells was also observed (Figure 8F).

Figure 8. FGK45.5 treatment results in increased numbers of activated B cells and dendritic cells.

C57BL/6 females were engrafted with skin from a CD45.1 congenic donor female. 24 hours following transplantation, recipient animals received 100μg of FGK45.5 or rat IgG2a via i.p. injection. 48 hours following antibody treatment (and 72h post-transplantation), spleens and lymph nodes were analyzed for the expression of the B7-family costimulatory molecule CD86 on various lineages of APCs. CD86 expression was upregulated on the cells of graft recipients treated with FGK45.5 (Panel D), but not those treated with isotype control (Panel C), nor on naïve control animals (B). Enumeration and phenotyping revealed that the majority of CD86 expressers are B cells (E) and dendritic cells (F). Three mice were included in each group.

Discussion

The present data demonstrate that minor Hya alloantigenic peptides attached to syngeneic leukocytes using ECDI confer dominant antigen-specific transplant protection dependent on alteration of CD40/CD154 signaling. Induction of graft protection was dependent on the administration of cells coupled with the dominant CD4 epitope (Dby), but not CD8 epitopes (Uty, Smcy) of male antigen. These findings are consistent with a previous study by Chai, et al. where transplant tolerance was induced using soluble Dby peptide, but not soluble Uty or Smcy (27). Paradoxically, a prior study from the same laboratory had reported that immature DCs (iDCs) loaded with the Uty peptide conferred tolerance to Hya-disparate skin grafts, while Dby-loaded iDCs sensitized female recipients and decreased male graft survival time (32). We found that Dby-coupled female splenocytes conferred significant protection to male skin grafts (Figure 1) confirming a previous report from our laboratory which obtained a similar outcome using large numbers of ECDI-fixed male splenocytes. While we observed some protection using an equivalent number of ECDI-fixed male splenocytes (median survival 39 days, data not shown), the protection was inferior to that conferred by Dby-SP. As supported by titration experiments in this and in our previous study [28], this is likely due to the increased amount of Dby antigen provided to the recipient in the Dby-SP treatment vs. the male splenocyte treatment.

The ability of ECDI-fixed, antigen-coupled splenocytes (Ag-SP) to induce peripheral tolerance has been documented in a number of CD4⁺ T cell-mediated immune disorders including EAE, Type 1 diabetes, and allogeneic islet transplantation (5-7, 10, 33). As was observed in those models, we found that treatment with Dby-SP significantly diminished DTH responses to Dby in vivo and reduced Dby-specific in vitro recall responses as assessed by proliferation and IFN-γ production, indicating profound unresponsiveness in the CD4 compartment. Further, Dby-SP tolerized mice failed to develop CTLs specific for the immunodominant Hya CD8 T cell epitopes. Our findings indicate that Ag-SP therapy can be successfully used to specifically control the rejection of mismatched tissue grafts with the proviso that the epitopes are known. Importantly, both CD4 and CD8 responses are diminished in the absence of broad-scale immunosuppressive agents and CD40/CD154 interactions are safely inhibited. We also find that complementing the observed defect in CD154 expression with anti-CD40 treatment restored graft rejection through CD4 re-activation, but not through CD8 activation. FGK45.5 has previously been reported to successfully and unsuccessfully prime CTL responses in the absence of CD4 T cells (31, 34).

We observed that FGK45 treatment of Dby-SP treated mice restored normal levels of graft rejection as well as graft infiltration, proliferation, and IFN-γ responses in the CD4 compartment, but not infiltration, lysis, or IFN-γ responses in the CD8 compartment. These findings are in agreement with pioneering work performed using FGK45, which collectively suggests that FGK45 effects on tumor and graft destruction are mediated primarily through CD4 T cells. Shepherd and Kerkvliet demonstrated that CTL responses against P815 tumor cells are not initiated in CD154^−/− mice, and that neither FGK45.5 treatment, nor the use of B7 over-expressing P815 cells restored CTL function, indicating that increased co-stimulation alone is insufficient for CTL priming in the absence of CD4 help/CD154(31). Likewise, FGK45 treatment only partially restores CTL function and allograft destruction in CD4^−/− mice. CD8-mediated graft destruction was again suggested to be independent of costimulation through the use of CD28-deficient recipients, indicating that while FGK45 treatment certainly increases B7 expression on APCs, it contributes minimally to direct priming of CTLs (34). Another study found that anti-tumor CTL responses could be primed in animals subjected to CD4 T cell depletion and FGK45 treatment, however, CD4 depletion was initiated 4 days following FGK45 treatment (35). It is therefore possible that, bolstered by FGK45 treatment, sufficient CD4 activation occurred to prime protective anti-tumor CTL responses prior to the depletion of T helper cells. Direct evidence for enhanced CD4 activation is supported by the finding that FGK45 treatment induced increases in antigen-specific CD4 T cell number, proliferation, IL-2, and IFN-γ secretion (36). Our observation of restored CD4, but not CD8, activity despite an increase in B7-family expression on recipient APCs is in agreement with the previous studies. FGK45-rescued CD4s may fail to provide a critical determinant for CTL priming. FGK45 (administered 24h following Ag-SP treatment and engraftment) cannot be affecting the input Ag-SP, as have shown that ECDI-fixed splenocytes under apoptotic death and are removed from the circulation and lymphoid organs within 12 h (33).

Cross-linking of peptide to the cell surfaces using ECDI rapidly induces apoptosis in Ag-SP (33), and i.v. administration of Ag-SP leads to the establishment of clonal anergy in T cells specific for the chemically-affixed peptide (1, 6, 37). The induction of apoptosis in Ag-SP is likely an important determinant in this therapeutic strategy, as apoptotic cells have been shown to be tolerogenic while cells dying by necrosis have been shown to be pro-inflammatory (38). More recent work exploring the mechanism of apoptotic cell tolerance determined that previously activated apoptotic expressing CD154 failed to induce tolerance (39) complimenting a previous study from the same group which showed that FGK45.5 treatment abrogated apoptotic cell-induced tolerance (40). Consistent with these findings, we found that Ag-SP encounter induces a defect in CD154 expression in the targeted T cell and that this defect is crucial to the protection of Hya-disparate grafts.

Although ECDI-catalyzed coupling antigens to donor leukocytes has been a tool for inducing antigen-specific immune tolerance in our laboratory for nearly 30 years (4), the mechanism(s) of suppression is not completely clear. Initial observations by Jenkins and Schwartz showed that antigen-specific Th1 clones stimulated by ECDI-fixed, antigen-coupled splenocytes failed to produce IL-2, resulting in failure of those T cells to mount a response during secondary encounter with the cognate antigen. However, these cells remained responsive to IL-2, and anergy could be prevented or reversed by supplementing IL-2 (9). Likewise, we showed that diminished proliferative responses of T cells from Dby-SP-treated male graft recipients in response to Dby re-challenge could be rescued by addition of exogenous IL-2 (Figure 1C). IL-2 is also a critical factor in CD4 priming of CD8 responses. In a different model of tolerance induction to Hya mismatched tissue, it was concluded that limited IL-2 production by Hya-specific CD4⁺ T cells was the critical helper dependent factor for the generation of an effective and long-lasting Hya-specific CD8 T cell response (27). CD8⁺ T cells primed in the absence of IL-2 fail to mount sufficient responses during secondary encounter with antigen (41). This may account for the apparently conflicting results obtained in the paper by James, et al. (32), in which Uty-pulsed iDCs conferred graft survival, but Dby-pulsed iDCs induced strong anti-graft responses. Although grafts were retained in Uty-iDC recipients, Uty-specific CD8⁺ T cells displayed an activated phenotype and secreted IFN-γ in recall assays. This could indicate that Uty-iDCs may have induced antigen-specific CD8 responses, but because the activation occurred in the absence of CD4-derived IL-2, a productive memory response was not induced. Despite the belief that iDCs present antigen in a tolerogenic fashion, the fact that they were administered subcutaneously (via the footpad) supports the idea that they actually functioned to immunize the animals in that study. In our experience with Ag-SP, the route of administration is a critical parameter, as subcutaneous injection stimulates DTH responses, while i.v. injection induces tolerance (4).

It has been proposed that clonal anergy is the major mechanism by which Ag-SP confer immune tolerance. The induction of clonal anergy is thought to occur when T cells receive “signal 1” (via the TCR) but not “signal 2” (via CD28) (reviewed in (42)). This control mechanism probably evolved to avoid inappropriate activation of T cells in the absence of a “danger signals” (in the form of a TLR ligand, for example) which induce APCs to upregulate expression of B7 family costimulatory molecules and secretion of pro-inflammatory cytokines. Ag-SP administration clearly does not provide a “danger signal” to APCs, as evidenced by the fact that LPS-induced APC activation abrogated the protective effect of Ag-SP therapy (8, 43). Induction and maintenance of T cell anergy following encounter with ECDI-fixed splenocytes is critically dependent upon a delicate balance of positive and negative co-stimulatory signals consistent with the demonstrated involvement of both CTLA-4/B7 and PD-1/PD-L1 interactions (6-8). In addition to TLR ligands, CD40/CD154 interactions are critical regulators of co-stimulatory molecule expression, cytokine secretion, and APC survival. Both mechanisms are known to regulate CD4 priming of CTL responses, and both abrogate Ag-SP-induced immune tolerance. The involvement of other TNFR family molecular pairs has not been determined in Ag-SP-induced immune tolerance. We expect perturbations to molecular pairs such as 41-BB/41-BBL and OX40-OX40L may also contribute to Ag-SP-induced hyporesponsiveness.

The role of CD8⁺ T cells in the rejection of Hya-disparate tissue is not completely clear. Male skin grafts are accepted by both β2m-deficient and CD8 T cell-depleted B6 females (44-46), however attempts to specifically inactivate Hya-specific CD8⁺ populations suggest that graft survival and CTL activity often do not correlate. For instance, Uty-pulsed iDCs given subcutaneously were shown to prevent rejection of male grafts, but did not prevent Uty-specific CD8 cells from becoming activated and secreting IFN-γ (32). Our results (Figure 1) and others (27) noted that Uty/Smcy-tolerized mice rejected Hya-disparate grafts (Figure 1A) despite the fact that in our study, these cells had diminished ability to produce IFN-γ and mediate lysis in response to recall with the CD8 epitopes (data not shown). In addition, increased male graft survival observed following the i.v. injection of large numbers of viable male splenocytes correlated with diminished Hya-specific DTH, but not CTL, responses (28). These studies collectively indicate that CD8 effector functions are insufficient for male graft rejection, or can be complimented by other CD4-mediated effector mechanisms to reject Hya-disparate grafts in the absence of CTL activity. Consistent with the notion of a disconnect between graft rejection and CTL activity, we found that graft rejection observed in Dby-tolerized, FGK45.5-treated females is associated with a restoration of the Hya-specific CD4⁺ response (proliferation, IFN-γ secretion), but not the CD8⁺ response (lysis, or IFN-γ secretion). We also show that tetramer⁺ CD8 T cells express CD44 following anti-CD40 treatment, indicating that surface phenotype is not necessarily indicative of effector function, in agreement with previous observations (47). It is also possible that Ag-SP-induced Tregs (7) allow expansion and partial activation of CD8 populations, but regulate their effector functions (27, 47).

That CD4 T cells are necessary and sufficient for the rejection of allogeneic tissue has been well described in models of transplant across MHC class II disparate barriers. Waldmann and colleagues showed that CD4 T cells expressing a transgenic TCR specific for Hya on a RAG1 KO background was sufficient for rejection of Hya mismatched skin grafts. Effector molecules such as IFN-γ, FAS-FASL, and granzyme B have all been implicated in CD4 T cell mediated tissue rejection (48), of which IFN-γ may have the most significant role due to its ability to mobilize immune responses against the target tissue and modify the local milieu. The pleiotropic effects of IFN-γ include the inducible expression of MHC class II and costimulatory molecules on endothelium and tissue parenchyma, increased vascular permeability, the downstream recruitment of effector cell types, and enhanced activation of macrophages, monocytes, and NK cells (49). Recently, IFN-γ production by CD4 T cells was shown to be critical for the recruitment of cytotoxic CD8 T cells to target tissues via its ability to enhance local tissue expression of CXCL9 and CXCL10 (50). Interestingly, in our model we did not find a role for IFN-γ production in the recruitment of CD8 T cells, as FGK45.5 restored IFN-γ production in Dby-SP tolerized CD4 T cells, and yet we failed to see any significant presence of CD8 T cells in the rejecting grafts of Dby-SP tolerized recipients receiving FGK45.5. These results suggest that the impairment of the Hya-specific CD8 T cell response goes beyond a defect in their ability to traffic to the graft site, and that rejection is a CD4 T cell mediated event. As IFN-γ can induce tissue expression of MHC class II, allowing for direct recognition of Hya-expressing tissue by CD4’s, it is possible that Hya-specific T cells are mediating graft rejection by direct cytotoxicity, as was recently observed in a model of CD4 T cell mediated tumor immunity (51). Alternatively, as favored by the correlation between DTH responsiveness and chronic graft rejection, it is likely that the Hya-specific CD4’s may mediate graft rejection via an indirect pathway, through the activation of Hya-bearing, tissue resident APCs. Bone marrow transplantation across a class-II disparate barrier was capable of inducing 100% lethal graft-versus-host disease (GVHD) in a CD4 T cell-dependent manner, even when chimeric recipients (MHC class II knockouts reconstituted with WT APCs’) were used (52). Thus, CD4 T cells were able to mediate GVHD with wild-type kinetics when only the tissue resident APCs expressed the alloantigens. Propagation of tissue destruction and inflammation by macrophages and microglia has also been described in models of autoimmunity (53, 54).

Aside from a defect in IL-2 production, anergy is associated with the expression of several E3 ubiquitin ligases that function to dampen activation signals in T cells by targeting the relevant signaling molecules for proteosomal degradation. One such E3 ligase is GRAIL, which was found to mediate the ubiquitination of CD154 in experimentally anergized T cells (55). As T cells encountering Ag-SP exhibit defective CD154 upregulation after subsequent antigenic challenge, it is possible that Ag-SP tolerance may induce GRAIL expression leading to CD154 degradation. Conversely, Ag-SP treatment may induce a GRAIL-independent defect in CD154 expression associated with anergy induction. In support of the latter hypothesis, CD154 expression has been shown to be the critical factor in determining the diabetogenic potential of pancreas-specific CD4⁺ T cells - NOD/4.1/RAG2^−/−/CD154^+/+ mice develop T1D, while NOD/4.1/RAG2^−/−/CD154^−/− mice fail to develop disease even after treatment with agonistic anti-CD40 (56). Ongoing experiments are determining the involvement of GRAIL in regulating CD154 expression in Ag-SP-induced tolerance.

Figure 7. Reduced CD8 graft infiltrate does not correlate with graft survival.

B6 female mice received female (panels A & B) and male (C-J) skin grafts following treatment with Uty/Smcy-SP (C & D), nothing (E & F), Dby-SP + rIgG2a (G & H), or Dby-SP + FGK45.5 (I & J). Ten days after engraftment, histologic sections were prepared from graft-containing tail areas and stained for CD4 (B, D, F, H, J) or CD8 (A, C, E, G, I). Grafts from mice displaying rejection (i.e. Uty/Smcy-SP-treated, Dby-SP + FGK45.5-treated, and untreated groups) contain variable amounts of CD8 infiltrate and considerable CD4 infiltrate. Grafts that are retained (female control grafts and male grafts on Dby-SP + rIgG2a-treated females) contain very few infiltrating T cells.