When transplanted into type 1a diabetic recipients, islet allografts are subject both to conventional allograft immunity and, presumably, to recurrent autoimmune (islet-specific) pathogenesis. Importantly, CD4 T cells play a central role both in islet allograft rejection and in autoimmune disease recurrence leading to the destruction of syngeneic islet transplants in diabetic NOD mice. However, it is unclear how NOD host MHC class II (I-Ag7)-restricted, autoreactive CD4 T cells may also contribute to the recognition of allogeneic islet grafts that express disparate MHC class II molecules. We hypothesized that islet-specific CD4 T cells can target MHC-mismatched islet allografts for destruction via the “indirect” (host APC-dependent) pathway of Ag recognition. To test this hypothesis, we determined whether NOD-derived, islet-specific CD4 T cells (BDC-2.5 TCR transgenic cells) could damage MHC-mismatched islets in vivo independent of conventional allograft immunity. Results demonstrate that BDC-2.5 CD4 T cells can vigorously destroy MHC class II-disparate islet allografts established in NOD.scid recipients. Tissue injury is tissue-specific in that BDC-2.5 T cells destroy donor-type islet, but not thyroid allografts established in the same NOD.scid recipient. Furthermore, BDC-2.5 CD4 T cells acutely destroy MHC class II-deficient islet allografts in vivo, indicating that autoimmune pathogenesis can be completely independent of donor MHC class II expression. Taken together, these findings indicate that MHC-mismatched islet allografts can be vulnerable to autoimmune pathogenesis triggered by autoreactive CD4 T cells, presumably through indirect autoantigen recognition in vivo.
The vulnerability of islets to immune injury continues to be one of the major barriers to successful transplantation. Islets transplanted into type 1a diabetics not only face alloimmune rejection, but also the threat of recurrent autoimmunity. The recurrence of autoimmune pathogenesis to pancreas transplants from identical twins is well documented (1, 2, 3), signifying that in patients with autoimmune disease, there is no advantage to transplanting perfectly matched tissue because reactivation of the destructive autoimmune response is possible. Because islet autoreactive T cells should be restricted to a complex of islet-associated Ag presented in self-MHC, it has been suggested that transplanting allografts bearing a poor MHC match between the donor and the recipient might reduce the risk of recurrent autoimmune diabetes following transplantation (2). Unfortunately, clinical studies have indicated that pancreatic allografts are also susceptible to autoimmune disease recurrence (3, 4, 5, 6). Taken together these studies have established the following important points. First, complete MHC mismatching does not protect islet allografts from disease recurrence. Second, the immunosuppression typically used in organ transplantation slows, but does not prevent, recurrent autoimmune pathogenesis. Finally, these findings suggest that recurrent autoimmunity following transplantation is a clinically relevant occurrence in type 1a diabetic individuals. Therefore, if islet transplantation into type 1a diabetic patients is to be successful, the mechanisms by which the autoimmune response recognizes and targets transplanted islets for destruction must be understood, so that therapies may be designed to ensure the survival of the transplant.
In the NOD mouse, the spontaneous onset of diabetes requires the participation of both CD4 (7, 8, 9) and CD8 (10, 11, 12) T cells. However, under various conditions, either CD4 (13, 14, 15) or CD8 (16, 17, 18) islet-specific T cells can independently initiate the onset of diabetes in the NOD recipient. Transplantation studies have demonstrated that islet graft destruction, resulting from the alloresponse (19), xenoresponse (20), or autoimmune response (21) is a CD4 T cell-dependent process. These findings indicate that CD4 T cells play a central role in coordinating graft injury in various models of islet transplant immunity. However, other studies have suggested a role for CD8 T cells in allograft rejection in NOD mice (22). Thus, the precise role of different T cell subsets in the autoimmune-mediated destruction of transplanted islets remains controversial. Because allograft immunity and potential autoimmunity occur simultaneously in autoimmune recipients, there is limited understanding concerning the extent of actual autoimmune (islet-specific) targeting of islet allografts and the mechanisms by which autoreactive CD4 T cells contribute to islet graft destruction. Thus, the possibility exists that even if the alloresponse to an islet transplant can be prevented, a recurrence of the islet-specific autoimmune response may result in the destruction of the transplant (23).
Central to this issue is the question of how a host MHC-restricted autoreactive CD4 T cell can recognize MHC-disparate islet allografts. Because optimal T cell activation requires both TCR engagement and costimulation (24), there are two conceptually distinct pathways by which CD4 T cells may recognize graft Ag: the “direct pathway” and the “indirect pathway” (25). In the direct pathway, recipient T cells recognize graft-derived Ags directly on the surface of graft-derived APCs capable of providing costimulation. This recognition requires TCR engagement of foreign MHC-peptide complexes on the surface of donor cells. In the indirect pathway, CD4 T cells respond to recipient APCs that capture, process, and present graft Ags in the context of recipient class II MHC in the presence of costimulation. Importantly, CD4 T cells activated via the indirect pathway should be restricted to recipient, not donor, class II MHC molecules. The indirect pathway is of particular importance when considering autoreactive T cells, which must have originally recognized their autoantigen in the context of self-MHC. Because indirect autoreactive CD4 T cells should be incapable of direct TCR-mediated engagement of MHC-disparate islet transplants, we hypothesize that autoreactive CD4 T cells target islet allografts for destruction via the indirect (host APC-dependent) pathway of Ag recognition.
In this study, we evaluate the nature of recognition involved in the response of an autoreactive CD4 T cell to MHC-disparate islet grafts in the absence of concomitant allograft immunity. These studies demonstrate that islet-specific CD4 T cells rapidly initiate tissue-specific islet allograft destruction that is independent of donor MHC class II expression, indicating that recognition of islet Ag by islet-specific CD4 T cells occurs via the indirect recognition pathway.
Materials and Methods
BDC-2.5 TCR transgenic NOD mice (26) were bred to NOD.scid mice (NOD.CB17-Prkdcscid/J; The Jackson Laboratory) to generate BDC-2.5 TCR transgenic NOD mice homozygous for the scid mutation (BDC-2.5 scid). BDC-2.5 scid mice that became diabetic before 35 days of age were used as splenocyte (SC)3 donors in adoptive transfer experiments.
The following mice were obtained from The Jackson Laboratory and used as islet donors in islet transplantation experiments: C57BL/6J (H-2b), MHC class II-deficient C57BL/6 (B6;129S-H2dlAbl-Eα), NOD.scid (H-2g7), and NOD CIITA−/− (NOD.129S2(B6)-C2tatm1Ccum/FlvJ). Adult male (8–10 wk old) NOD.scid mice (H-2g7) or spontaneously diabetic NOD/BDC female mice were used as transplant recipients. Adult (8–10 wk old) male NOD.scid mice were used as adoptive transfer recipients. The following mice were obtained from The Jackson Laboratory and used as lymphocyte donors in MLRs: B6.H-2g7 (B6.NOD-(D17Mit21-D17Mit10)/LtJ;H-2g7), NOD.B10 (NOD.B10Sn-H2b/J;H-2b), NOD CIITA−/− (NOD.129S2(B6)-C2tatm1Ccum/FlvJ), and C57BL/6J (H-2b). All mice were bred and maintained at the Barbara Davis Center Rodent Facility (Denver, CO) under specific pathogen free conditions. All experiments were performed in compliance with guidelines established by the Animal Care and Use Committee of the University of Colorado Health Sciences Center.
Islet isolation and transplantation
NOD.scid islet graft recipients were rendered diabetic by the i.v. injection of 180–200 mg/kg streptozotocin (SZ; Calbiochem). Diabetes was determined by the daily monitoring of blood glucose levels with a Precision blood glucose meter (MediSense). Animals were considered diabetic after two consecutive blood glucose measurements ≥17 mM. Onset of diabetes was dated from the first consecutive reading. Diabetic mice were transplanted with islets within 48–72 h of the onset of diabetes. Transplants consisting of 450–500 islets were grafted under the left kidney capsule.
Islets were isolated from the pancreata of anesthetized adult mice by a collagenase digestion protocol modified from Gotoh et al. (27). In brief, pancreata were distended by direct injection of 4 ml of 2.5 mg/ml collagenase (type V, C-9263; Sigma-Aldrich) into the common pancreatic duct. The pancreata were then surgically removed and further digested in 1 ml of 2.5 mg/ml collagenase for 10–11 min at 37°C. Islet digests were vigorously shaken, washed three times in HBSS/HEPES, 10 μg/ml DNase, 0.1% BSA, 1% penicillin-streptomycin, and strained through a 500-μm metal mesh filter to separate the islets from the surrounding undigested tissue. These crude islet digests were further purified by centrifugation on a Histopaque density gradient (1.119; Sigma-Aldrich) for 20 min at 800 × g. The media Histopaque interface was removed, washed in HBSS/HEPES, 10% FBS, 1% penicillin-streptomycin, and islets were handpicked into aliquots of 450–500 for transplantation.
Before transplantation, the islets were drawn up into PE-50 silastic tubing (BD Biosciences) using a hand-machined islet micrometer (generously provided by Dr. R. Rajotte, University of Alberta, Edmonton, Alberta, Canada) and concentrated by centrifugation of the tubing. Islets were transplanted under the left renal capsule by exposing the left kidney through a flank incision and pushing the kidney through the incision. With the aid of a dissecting microscope, the kidney capsule was cut with a needle, the tubing was inserted under the kidney capsule, and the islets were delivered under the kidney capsule with the micrometer. Normoglycemia was reestablished within 24–72 h of successful transplantation. Graft function was determined by monitoring blood glucose with a Precision blood glucose meter (MediSense). Animals were considered normoglycemic after two consecutive measurements <10 mM. Islets were allowed to engraft for at least 14 days to permit vascularization of the graft by host vascular endothelium before the introduction of BDC-2.5 scid SC. In control animals, graft-dependent euglycemia was confirmed by nephrectomy of engrafted kidneys 60 days after transplant.
CD4 T cell and CD8 T cell depletion of NOD mice
CD4+ or CD8+ cells were extensively depleted with anti-CD4 (GK1.5; 10 mg/kg) or anti-CD8 (116--13.1; 10 mg/kg), respectively, administered on days −7, −2, 0, 2, 7, 14, and 28 relative to islet transplant (treatment was terminated if acute graft rejection occurred during the treatment period). Depletion of CD4 or CD8 T cells was confirmed by flow cytometric analysis of NOD recipients using the following Ab pairs: anti-CD4 (RM4-4; BD Pharmingen) and anti-CD8 (53-6.7; BD Pharmingen). T cells were depleted to <2% of initial numbers in peripheral blood and levels remained depressed for 14 days beyond the last dose of mAb.
BDC-2.5 NOD lymph node and SC were harvested and depleted of RBCs by separation on the density separation medium Lympholyte-M (Cedarlane Laboratories). Lymphocytes were then enriched for CD4 T cells using Cellect CD4 cell immunocolumns (Cedarlane Laboratories) according to the manufacturer’s instructions. The purity of the responders was determined by flow cytometry. MLR cultures were then established by combining 2 × 105 BDC-2.5 column-enriched CD4 responders cells with 3 × 105 gamma-irradiated (2500 rad) SC stimulators with or without gamma-irradiated C57BL/6 islet Ag in quadruplicate 200-μl cultures in 96-well flat-bottom tissue culture plates. Cells were cultured in Eagle’s MEM containing 10% FBS, 10−5 M 2-ME, 0.1 mM MEM nonessential amino acids, and antibiotics and incubated at 37°C in 10% CO2 in air. Proliferative responses were determined by pulsing the culture wells with 1 μCi of tritiated thymidine for 6 h on days 2, 3, and 4 of culture. Cultures were harvested onto glass fiber filters (Wallac) and incorporated tritiated thymidine was quantified on a Wallac beta emission counter.
In tissue specificity controls, a sample of thyroid from the same donor as the islet graft was cotransplanted beneath the kidney capsule. In brief, both thyroid lobes were removed with sterile technique from a donor animal and placed in a petri dish containing 1× HBSS immediately before transplantation. Each recipient animal received one thyroid lobe beneath the capsule on the opposite end of the kidney as the islet graft. The presence or absence of infiltrate in the transplanted thyroid tissue was determined histologically following nephrectomy of the grafted kidney or sacrifice of the recipient.
Transfer of BDC-2.5 scid SCs
Single cell suspensions of SC from BDC-2.5 scid were prepared in HBSS using glass homogenizers. RBC were removed from spleen cell preparations by separation on the density separation medium Lympholyte-M (Cedarlane Laboratories). Mononuclear cells were counted and mononuclear cell viability was determined by trypan blue (Sigma-Aldrich) exclusion using a hemocytometer and phase contrast microscopy.
In experiments designed to ascertain the capacity of BDC-2.5 SC to transfer diabetes, NOD.scid mice were injected i.p. with either 106, 105, 5 × 104, 104, or 103 BDC-2.5 scid SC in 300 μl of HBSS.
In experiments designed to assess the ability of BDC-2.5 CD4 cells to target and destroy islet grafts, diabetic NOD.scid mice were transplanted with islet grafts. After a minimum of 14 days, NOD.scid mice bearing functioning islet grafts (blood glucose <10 mM) were challenged i.p. with 105 or 106 freshly isolated SC from BDC-2.5 scid mice. Controls for the diabetogenicity of the SC consisted of 8- to 10-wk-old NOD.scid mice challenged i.p. with 105 or 106 freshly isolated BDC-2.5 SC from the same SC preparations.
In experiments designed to establish that the transfer of diabetes is dependent upon CD4 T cells present in SCs, NOD.scid mice were injected i.p. with 105 BDC-2.5 SC on day 0. CD4+ cells were depleted with a short course of anti-CD4 mAb (GK1.5) therapy. A total of 10 mg/kg GK1.5 (ATCC no. TIB-207; American Type and Culture Collection) was administered i.p. on the opposite side of the abdomen on days 0, 2, and 7. Controls for the diabetogenicity of the SC consisted of NOD.scid mice injected i.p. with 105 BDC-2.5 SC and treated with 1 mg of rat IgG (I-4131; Sigma-Aldrich) on days 0, 3, and 6.
In all experiments, blood glucose was monitored daily following challenge and animals were considered diabetic after two consecutive measurements >17 mM. Diabetic mice were sacrificed, and their pancreas or graft was removed for histology.
The purity of freshly isolated SC from BDC-2.5 scid mice for use in adoptive transfers and lymph node cells from BDC-2.5 NOD mice for use in MLR, before and after CD4+ T cell enrichment, was assessed by flow cytometry. Cells were stained with the following panel of Abs (BD Pharmingen) for phenotyping by flow cytometry: 1) FITC-coupled hamster anti-TCR (clone H57-597), 2) PE-coupled rat anti-CD4 (clone RM4-4), 3) FITC-coupled rat anti-Vβ4 (clone KT4), and 4) FITC-coupled rat anti-CD19 (clone 1D3). Frequency determinations were calculated from single-parameter and double-parameter fluorescence histograms on a BD Biosciences FACSCalibur after gating on viable lymphocytes. CellQuest software (BD Biosciences) was used to analyze flow cytometry data.
Both transplanted NOD.scid and control NOD.scid mice that received BDC-2.5 scid SC were sacrificed upon the development of diabetes (two consecutive blood glucose values ≥17 mM). In transplanted control animals, engrafted kidneys were nephrectomized 60 days after transplant. At the conclusion of each study, the relevant tissues were removed and fixed in 10% (v/v) formalin in aqueous phosphate buffer and embedded in paraffin. Paraffin-embedded tissues were cut into 20-μm sections. Paraffin sections were stained with H&E (Harris; Fischer Scientific) to determine the extent of mononuclear infiltration. The presence of insulin and glucagon in grafts was determined by immunoperoxidase staining of parallel tissue sections. Immunoperoxidase staining used guinea pig anti-insulin antiserum (DAKO) and a level 2 multispecies ultra streptavidin peroxidase detection kit (Signet Laboratories) followed by a diaminobenzidine chromogen (Signet Laboratories) as a substrate. Sections stained for insulin were counterstained with Gill’s hematoxylin (Fischer Scientific). The degree of mononuclear cell infiltration and islet tissue damage was determined for each section. Light microscopy and photography was performed using a Leitz Orthoplan (Leica Microsystems) and a Spot Digital Imaging System (Diagnostic Instruments).
Mann-Whitney U nonparametric analysis and Fisher exact tests using the commercially available software GraphPad Instat version 3 were used to determine significance of diabetes onset in adoptive transfer studies.
Recurrent autoimmune pathogenesis to pancreatic islet isografts is CD4 T cell-dependent, but does not require islet class II MHC
Although multiple studies indicate that both CD4 and CD8 T cells are necessary for disease onset in NOD mice, the role of these T cell subsets in the recurrence of disease in islet transplants has been less clear. To verify the role of the CD4 and CD8 T cell subsets in graft destruction due to recurrent autoimmunity, spontaneously diabetic NOD mice were depleted of CD4 or CD8 T cells by mAb therapy and then grafted with syngeneic NOD islets. Graft function was determined by monitoring blood glucose levels. Islet isografts transplanted into untreated animals and grafted animals receiving depleting anti-CD8 mAb therapy demonstrated a high failure to engraft. Furthermore, functioning islet grafts in these groups exhibited rapid disease recurrence (Fig. 1 A). However, CD4 depletion of diabetic NOD mice prevented acute disease recurrence (p < 0.0019 relative to untreated controls) with two of nine grafts surviving the entire 100-day course of the experiment. Graft destruction coincided with the return of CD4 T cells to the periphery (DNS). These results demonstrate that disease recurrence to syngeneic islet grafts is significantly CD4 T cell-dependent and relatively CD8 T cell-independent. As such, these findings are consistent with our previous results in which allogeneic islet graft destruction in spontaneously diabetic NOD mice was CD4 T cell-dependent and CD8 T cell-independent (21).
Although disease recurrence in syngeneic islet grafts is CD4 T cell-dependent, it is not clear whether this response requires MHC class II expression by the NOD islet donor. To address this question, spontaneously diabetic NOD mice were grafted with islets from NOD MHC CIITA-deficient (CIITA−/−) islets. CIITA−/− mice lack conventional MHC class II expression (28, 29). Both NOD.scid control and NOD CIITA−/− islets exhibited equivalent acute recurrence of autoimmune pathogenesis (Fig. 1 B), consistent with another recent report (30). These results indicate that donor MHC class II expression is not required for acute disease recurrence and suggest that CD4 T cell involvement in this process could involve indirect (host MHC-restricted) autoantigen recognition.
BDC-2.5 TCR transgenic T cells as a model of indirect islet-specific T cell recognition in vitro and in vivo
To explore the potential role of “indirect” autoreactive CD4 T cells in recurrent autoimmune pathogenesis to islet transplants, we next set out to characterize a model of acute disease recurrence using a defined population of autoreactive, islet-specific CD4 T cells from the BDC-2.5 TCR transgenic mouse. The BDC-2.5 transgenic mouse bears transgenic TCR genes from an islet-specific NOD-derived CD4 T cell clone (26). This transgenic model yields a population of diabetogenic T cells specific for an undefined islet granule-associated Ag in the context of NOD class II MHC (I-Ag7) (31, 32). Importantly, when bred onto the NOD.scid background (33), the resulting BDC-2.5 scid mice become spontaneously diabetic at 3–5 wk of age, and SCs vigorously transfer disease to immune-deficient NOD mice. Thus, SCs from diabetic BDC-2.5 scid mice were used as a source of monospecific autoreactive, islet-specific CD4 T cells in a model of acute disease recurrence.
To confirm the specificity of BDC-2.5 CD4 T cells, in vitro cultures were established to determine whether proliferation by BDC-2.5 transgenic CD4 T cells to islet Ag requires the NOD class II MHC I-Ag7. BDC-2.5 transgenic CD4 T cells were column-enriched and cultured with gamma-irradiated APCs in the presence or absence of islet Ag. As shown in Fig. 2,A, BDC-2.5 CD4 T cells responded vigorously to islet Ag in the presence of NOD (I-Ag7) and B6.H-2g7 (I-Ag7) APCs. Importantly, the response to islet Ags required presentation by APCs in that there was undetectable reactivity to either islet Ag alone or APCs alone. Furthermore, BDC-2.5 CD4 T cells were unable to proliferate to islet Ag in the presence of class II MHCnull NOD CIITA−/− APCs, indicating that BDC-2.5 CD4 T cells require MHC class II islet Ag presentation by APCs. Previous studies have demonstrated that the BDC-2.5 clone reacts to an islet-associated Ag expressed by all mouse strains tested (34). This property also is exhibited by BDC-2.5 scid T cells that can respond to allogeneic B6 islets presented by NOD APC (Fig. 2 B). Importantly, BDC-2.5 scid T cells do not demonstrate reactivity to B6 APCs alone, indicating that that this clone is not fortuitously cross-reactive to the H-2b haplotype. We also find that TCR transgenic BDC-2.5 scid CD4 T cells do not transfer disease to B6 (H-2b) scid recipients (0 of 6), further illustrating that this clone is unable to react to islet Ags in the context of I-Ab in vivo. Taken together, these data indicate that the BDC-2.5 scid CD4 T cell can respond to either syngeneic or allogeneic islet Ag presented in the context of NOD APCs. As such, BDC-2.5 scid T cells are a relevant example of autoreactive CD4 T cells with indirect reactivity to donor islet Ags.
We then examined the pathogenicity of BDC-2.5 CD4 T cells in vivo. To determine an appropriate dose of autoreactive, islet-specific BDC-2.5 scid TCR transgenic CD4 T cells for use in recurrent autoimmune pathogenesis studies, graded numbers of SCs (103–106) from the BDC-2.5 scid were transferred into immune-deficient NOD.scid mice (Fig. 3,A). Mice receiving >104 BDC-2.5 scid SCs became rapidly diabetic, whereas mice receiving <104 BDC-2.5 scid SCs exhibited reduced incidence and delayed onset of diabetes. Histological evaluation of pancreata removed from diabetic mice revealed the presence of a massive mononuclear cell infiltrate simultaneous with the destruction of islet architecture and the loss of insulin-producing β cells (DNS). Flow cytometric analysis of the SC populations used in adoptive transfer experiments demonstrated that within the CD45+ gate, ∼70% of the cells are BDC-2.5 (CD4+Vβ4+) T cells, ∼5% of the cells are DX5+ NK cells, and the identity of the remaining cells (CD8−CD19−) is unknown. Finally, treatment with anti-CD4 mAb therapy prevented disease transfer to NOD.scid mice by BDC-2.5 scid SCs, indicating that islet destruction is dependent upon CD4+ cells contained within the SC inoculum (Fig. 3 B). Taken together, these results establish the BDC-2.5 transgenic T cell as an initiator of an aggressive CD4 T cell-dependent autoimmune pathogenesis independent of other lymphocyte subsets.
To determine whether BDC-2.5 scid transgenic T cells can mediate recurrent autoimmune pathogenesis to islet transplants, islet allografts were established in NOD.scid mice rendered diabetic by treatment with SZ (SZ-NOD.scid). To determine whether BDC-2.5 CD4 T cells are able to initiate the destruction of MHC-disparate islet transplants, NOD.scid mice bearing either NOD.scid isografts (H-2g7) or MHC-disparate C57BL/6 (B6; H-2b) allografts were challenged with BDC-2.5 scid SCs (Fig. 4). Parallel unmanipulated NOD.scid mice also received BDC-2.5 scid SCs in each experiment as an internal control for the pathogenicity of the BDC-2.5 scid SC inoculum. Both NOD.scid isografts and B6 allografts were capable of durable function in that grafts in unchallenged NOD.scid mice survived >60 days. Removal of graft-bearing kidneys from animals with long-term graft function caused a rapid reversion to hyperglycemia, indicating that maintenance of normoglycemia was graft-dependent. However, the transfer of BDC-2.5 T cells orchestrated the rapid destruction of both NOD.scid isografts (Fig. 4,A) and B6 allografts (Fig. 4 B) with equivalent kinetics. Thus, these findings indicate that BDC-2.5 transgenic T cells are capable of coordinating the destruction of both NOD.scid isografts and MHC-disparate B6 islet allografts.
BDC-2.5 CD4 T cells orchestrate the destruction of MHC-disparate islet allografts via autoimmune (islet-specific) recognition in vivo
Because these results clearly show that BDC-2.5 CD4 T cells are capable of eliciting the destruction of syngeneic and allogeneic islet grafts, it was important to verify that the destruction of allogeneic islets was due to autoimmune recognition and not because of a cross-reactive allogeneic response. To distinguish between these two pathways of recognition, diabetic NOD.scid mice received dual transplants of allogeneic islets and thyroid under opposite poles of the left kidney capsule (21). Once islet graft function was established, mice were challenged with BDC-2.5 scid SCs to determine whether graft recognition by the BDC-2.5 transgenic T cell is tissue-specific. Islet graft function was monitored by measuring blood glucose levels, and the kidney bearing the grafts was removed for histological analysis 1 day after the onset of hyperglycemia. The thyroid transplant served as a “sentinel” graft to distinguish islet-specific autoimmunity from conventional alloimmunity because thyroid grafts express allogeneic Ags, but presumably not the BDC-2.5 islet-specific Ag. Challenge of NOD.scid mice bearing dual B6 islet/thyroid transplants with BDC-2.5 scid SCs demonstrated that BDC-2.5 T cells are only able to trigger the destruction of islet tissue (Table I). The adjacent thyroid grafts were completely intact upon histological evaluation without discernible mononuclear cell accumulation (Fig. 5). Islet and thyroid grafts harvested from mice that received cotransplants, but were not reconstituted with a T cell population, displayed intact allogeneic islet and thyroid grafts for >60 days. Importantly, the histology of thyroid grafts in animals receiving BDC-2.5 T cells resembled the control grafts, whereas the corresponding islet grafts in these animals were completely destroyed.
|Challenge .||Specificity .||No. Destroyed/Total .||.|
|.||.||Islets .||Thyroid .|
|BDC-2.5 scid SCs||Autoreactive||6/6||0/6|
|Diabetic NOD SCs||Autoreactive and alloreactive||7/7||7/7|
|Challenge .||Specificity .||No. Destroyed/Total .||.|
|.||.||Islets .||Thyroid .|
|BDC-2.5 scid SCs||Autoreactive||6/6||0/6|
|Diabetic NOD SCs||Autoreactive and alloreactive||7/7||7/7|
To confirm that thyroid tissue could be destroyed by an appropriate allogeneic response, a parallel group of cotransplanted mice received 5 × 106 SCs isolated from wild-type diabetic NOD mice. Because heterogeneous SCs from a diabetic NOD mouse contain T cells capable of mounting a response to both islet-associated autoantigens and alloantigens, both the islet and thyroid grafts should be destroyed. Indeed, both islet grafts and thyroid grafts in cotransplanted NOD.scid mice that received SCs from a diabetic NOD mouse were targeted for destruction (Table I and Fig. 5), confirming that the thyroid graft was indeed subject to allograft immunity. Taken together, these observations indicate that islet graft destruction mediated by BDC-2.5 T cells is an islet-specific event, and not due to a cross-reaction with alloantigens. These results are in agreement with similar studies using the insulin-specific PD-12-4.4 CD4 T cell clone (35).
BDC-2.5 T cell-coordinated destruction of islet transplants does not require class II MHC expression by the islet graft
The data presented thus far illustrate that BDC-2.5 T cells target islet allografts via a tissue-specific mechanism. However, the possibility exists that islet graft recognition by BDC-2.5 T cells may result from direct Ag recognition of the graft due to an unexpected cross-restriction of the BDC-2.5 TCR with allogeneic class II MHC bearing islet-associated Ags. To exclude this possibility, we tested whether class II MHC expression on islet grafts is necessary for their destruction by islet-specific CD4 T cells. Wild-type B6 control or class II MHC-deficient islet grafts (36) were established in chemical-induced diabetic NOD.scid mice. After graft function was established, grafted mice were challenged with BDC-2.5 scid SCs and graft function was followed by monitoring blood glucose levels. As shown in Fig. 6, transfer of BDC-2.5 scid SCs triggered the rapid destruction of both control B6 islet allografts and B6 MHC II−/− islets (p = 0.5665). The tempo of the response (hyperglycemia) was indistinguishable from parallel ungrafted NOD.scid mice receiving BDC-2.5 scid SCs. Also, both B6 wild-type and MHC class II−/− islet allografts functioned >60 days in unchallenged NOD.scid recipients.
Collectively, these results indicate that islet-specific CD4 T cells mediate the rapid tissue-specific destruction of MHC-disparate islet allografts. Graft destruction does not require class II MHC expression by any cells of the islet graft. Thus, recognition and destruction of islet β cells does not require direct TCR-mediated or “cognate” engagement of the islet transplant. As such, these results are consistent with a model of islet allograft destruction in which islet-specific CD4 cells recognize islet Ag via the host MHC-restricted indirect pathway.
A major shortcoming in the current understanding of the autoimmune pathogenesis of islet transplants is the lack of information regarding the pathway of Ag recognition involved (e.g., direct vs indirect). Previous efforts to examine the mechanism by which autoreactive T cells recognize MHC-disparate islet transplants and target them for destruction have been hindered by a lack of an experimental system capable of distinguishing between conventional allograft immunity and the tissue-specific autoimmune response, because both of these immune processes occur simultaneously. Consequently, very little is known about the ability of a defined autoreactive (islet-specific) T cell to contribute to the recognition and destruction of an MHC-disparate islet graft. Due to the CD4 T cell dependency of autoimmune disease recurrence in NOD mice (21), we chose to investigate the pathways of islet allograft recognition used by a defined population of islet-specific CD4 T cells.
To determine the extent of allograft targeting by islet-specific CD4 T cells, we studied the response of a defined islet-specific CD4 T cell population from the BDC-2.5 TCR transgenic NOD.scid mouse to islet allografts deficient for candidate molecules involved in islet recognition. Our results clearly demonstrate that the autoimmune mechanism can target islet allografts for destruction, and that the recognition and destructive processes involved in autoimmune disease recurrence do not require MHC class II expression by the transplant. Although the precise nature of the BDC-2.5 CD4 T cell-APC engagement remains to be demonstrated, these data strongly argue that islet-specific CD4 T cells recognize islet Ag via the indirect pathway of Ag recognition. These observations are consistent with the outcome of a previous study investigating the role of indirect recognition in the spontaneous onset of diabetes in a transgenic model of disease (37).
The observation that indirect, autoreactive CD4 T cells can mount an islet destructive response has important consequences for clinical transplantation. Within the field of transplantation, there is an ongoing controversy regarding whether the autoimmune repertoire can recognize MHC-disparate islet transplants. On the one hand, studies support an indirect mechanism of autoimmune pathogenesis to islet allografts (21). In contrast, other reports have also suggested that the disease component of islet graft rejection can only target the graft if it is MHC-matched to the recipient (22, 38). This latter hypothesis posits that autoreactive T cells should not be able to mount a TCR-mediated, destructive response to MHC-disparate islets because non-self MHC-Ag complexes should not possess sufficient homology to self-MHC-autoantigen complexes to permit recognition by the autoreactive T cell. Consequently, MHC-disparate grafts might actually be preferred for transplantation into type 1a diabetic patients because the immune response to the transplant would lack an autoimmune component. This hypothesis implies a mechanism of pathogenesis that requires a direct interaction between the islet-specific T cell and MHC-peptide complexes expressed on the surface of the islet. However, if indirect recognition is a major component of the autoimmune response to islet transplants, then MHC-disparate grafts should be vulnerable to autoimmune destruction if the donor islet bears an Ag, which when processed and presented by host APCs triggers autoreactive CD4 T cells.
Taken together, our results suggest the following model of BDC-2.5 CD4 T cell-orchestrated islet graft recognition and destruction involving indirect (host MHC-restricted) recognition. In this model, islet injury does not require an MHC-restricted engagement of the islet target cells. Rather, graft-derived cells/Ags, including islet autoantigens, are acquired by host APCs and then processed and presented on its surface in the context of recipient class II MHC. The conclusion that such indirect recognition of islet autoantigens is sufficient to initiate islet graft destruction clearly leads to the subsequent issue of the specific effector mechanisms responsible for tissue injury. Because a CD4 T cell TCR-mediated interaction with the islet target is not required for injury, our data imply that graft destruction in this model is the outcome of “bystander damage” resulting from local inflammation induced by a CD4 T cell/APC interaction at the graft site. Several proinflammatory cytokines (39, 40, 41) and/or mediators such as reactive oxygen species (42, 43, 44) have been implicated in contributing to such islet injury. One study has suggested that islet injury by BDC 2.5 T cells requires TNFR1 expression on the target islets (15), but our own similar studies have not suggested a requirement for TNFR expression for this process (T. Kupfer, unpublished observations). Our ongoing studies are investigating the role of candidate pathways of inflammatory injury in this model system. This injury appears to be quite complex and does not require a rate-limiting contribution of a number of major candidate inflammatory cytokines long associated with islet dysfunction (15), but may well involve oxygen free radical inflammatory injury (45, 46) as recently shown in the pathogenesis of BDC 2.5 autoreactive CD4 T cells in vivo (47).
In conclusion, our data indicate that autoreactive CD4 T cells target islet transplants via indirect Ag recognition regardless of the donor MHC haplotype. Thus, MHC-disparate islet allografts are subject to CD4 T cell-dependent autoimmune recognition that can be sufficient for overt graft failure.
We thank J. Beilke and N. Kuhl for help with islet isolation and M. Coulombe, H. C. Gelhaus, M. Sleater, J. Beilke, and S. Weber for critical evaluation of the manuscript.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grants RO1 DK AI 55333 (to R.G.G.) and the Diabetes and Endocrinology Research Center, University of Colorado Health Sciences Center DERC P30 DK57516.
Abbreviations used in this paper: SC, splenocyte; SZ, streptozotocin.