Porcine thymus grafts support robust murine and human thymopoiesis, generating a diverse T cell repertoire that is deleted of donor and host-reactive cells, achieving specific xenograft tolerance. Positive selection is mediated exclusively by the xenogeneic thymic MHC. Although thymectomized, T cell-depleted normal mice usually remain healthy following xenogeneic thymic transplantation, thymus-grafted congenitally athymic mice frequently develop multiorgan autoimmunity. We investigated the etiology of this syndrome by adoptively transferring lymphocyte populations from fetal pig thymus-grafted BALB/c nude mice to secondary BALB/c nude recipients. Fetal pig thymus-grafted nude mice generated normal numbers of CD25+Foxp3+CD4 T cells, but these cells lacked the capacity to block autoimmunity. Moreover, thymocytes and peripheral CD4+CD25− cells from fetal pig thymus-grafted nude mice, but not those from normal mice, induced autoimmunity in nude recipients. Injection of thymic epithelial cells from normal BALB/c mice into fetal pig thymus grafts reduced autoimmunity and enhanced regulatory function of splenocytes. Our data implicate abnormalities in postthymic maturation, expansion, and/or survival of T cells positively selected by a xenogeneic MHC, as well as incomplete intrathymic deletion of thymocytes recognizing host tissue-specific Ags, in autoimmune pathogenesis. Regulatory cell function is enhanced and negative selection of host-specific thymocytes may potentially also be improved by coimplantation of recipient thymic epithelial cells in the thymus xenograft.
We previously demonstrated that donor-specific xenograft tolerance can be achieved in thymectomized (ATX),4 T cell-depleted mice by grafting fetal pig thymus and liver tissue (FP THY/LIV) under the kidney capsule (1, 2). In this xenogeneic pig-to-mouse model, mouse CD4+ T cells repopulated the periphery of T cell-depleted ATX mice after grafting with FP THY/LIV. These repopulated mouse CD4+ cells were tolerant to xenogeneic donor Ags, as indicated by specific nonresponsiveness to donor xenoantigens in MLRs and long-term acceptance of donor MHC-matched xenogeneic pig skin grafts, with the ability to reject third-party skin grafts (1, 2). In addition, these repopulating mouse CD4+ T cells were immunologically functional (3). Therefore, xenogeneic thymic transplantation provides a promising approach to achieving xenograft tolerance.
Our previous studies have demonstrated that intrathymic clonal deletion is one of the major mechanisms of tolerance to host and donor Ags in thymic xenografts (1, 2, 4). The negative selection of host-reactive thymocytes correlates with the presence of host MHC class IIhigh cells with dendritic cell morphology in the porcine thymic grafts (2). Although both donor pig and host mouse MHC molecules participate in the negative selection of mouse thymocytes in FP THY-grafted ATX mice, positive selection appears to be mediated only by pig MHC, with no demonstrable contribution from the host mouse MHC (5, 6).
Despite clear evidence, by analysis both of superantigen-reactive Vβ and of a transgenic TCR with known host reactivity, that host APC participate in negative selection in FP thymic grafts, we previously reported that a small percentage of FP THY/LIV-grafted ATX B6 mice (∼10%) and a markedly higher percentage of FP THY/LIV-grafted (FPG) nude mice (∼60%) develop an autoimmune disease. The disease manifests clinically as a wasting syndrome with multiorgan infiltration by recipient CD4 cells and can be induced by adoptive transfer into syngeneic nude mice of CD4+ T cells, which are essential for disease development (4). Cotransfer of normal syngeneic splenocytes prevented the occurrence of autoimmune disease in secondary BALB/c nude recipients of splenocytes from FPG nude mice (4), suggesting a possible failure of regulatory function in FPG nude mice. We have now used an adoptive transfer approach to address the roles of regulatory cells and effector cell selection in mediating this phenomenon.
Materials and Methods
Female BALB/c (H2d) and BALB/c nude mice were purchased from Charles River Laboratories. All mice were maintained in a specific pathogen-free facility and were housed in microisolator cages containing autoclaved feed, bedding, and acidified water. Second trimester (gestational age 60–75 days, estimated by observed estrus or mating and confirmed by ultrasound examination of the fetuses) partially inbred swine leukocyte Ag Massachusetts General Hospital (MGH) miniature swine fetuses were used as donors of porcine thymic and liver tissue. Massachusetts General Hospital miniature swine have been bred to homozygosity for the swine leukocyte Ag complex (7). Animal handling and care were in accordance with the American Association for the Accreditation of Laboratory Animal Care and institutional guidelines.
Eight- to 12-wk-old BALB/c nude mice were transplanted with a miniature swine FP THY/LIV fragment, each ∼2 mm3 in size, under the kidney capsule (8). All surgical interventions were performed with i.p. injection of ketamine (0.08 mg/g) and xylazine (0.012 mg/g) in combination with inhaled methoxyflurane (Pitman-Moore) to maintain stage III anesthesia.
Eleven to 12 wk after transplantation of FP THY/LIV, splenocytes were harvested from FP THY/LIV-grafted BALB/c (FPG) nude mice or control normal BALB/c mice. BALB/c nude mice that served as adoptive recipients received 3 Gy total body irradiation and i.v. injections of 2 × 107 splenocytes from FPG mice (GSPL) or normal BALB/c mice (nSPL). In cotransfer experiments, the indicated cell populations were transferred to BALB/c nude mice along with GSPL. The body weights of BALB/c nude recipients were followed weekly.
CD4+ or CD8+ T cells were purified from splenocytes by positive selection using mouse CD4 or CD8 Microbeads (Miltenyi Biotec) according to the manufacturer’s instructions. The purity of the resulting CD4+ and CD8+ T cells was >90 and >95%, respectively. To purify CD4+CD25+ and CD4+CD25− T cells, CD4+ T cells were isolated from spleen by negative selection using the mouse CD4+ T Cell Isolation Kit (Miltenyi Biotec) according to the manufacturer’s instructions. CD4+-enriched cells were then stained with FITC-labeled anti-CD4 mAb (RM4-5) and PE-labeled anti-CD25 mAb (PC61) and sorted on a MoFlo (DakoCytomation). The purities of CD4+CD25+ and CD4+CD25− T cells were >90 and >99%, respectively.
Cell staining, mAbs, and flow cytometry (FCM)
Mice were tail bled at regular intervals posttransplant to obtain peripheral blood lymphocytes, which were prepared with Histopaque gradient 1077 (Sigma-Aldrich). Splenocyte and lymph node (LN) cell suspensions were prepared by standard techniques. Levels of T cell reconstitution in the FPG mice that underwent transplantation and the phenotype of cells were determined by multicolor flow cytometric analysis using various combinations of the following mAbs: anti-mouse CD4 (RM4-5), CD8 (53-6.7), CD25 (PC61), CD69 (H1.2F3), CD44 (IM7), CD62L (MEL-14), CD45RB (16A), CD45 (30-F11), and isotype control mAbs (all purchased from BD Pharmingen). Nonviable cells were excluded from analysis by gating out lower forward scatter and high propidium iodide-retaining cells. For intracellular staining, cells were washed twice with wash buffer and surface Ags were stained with Abs. The cells were then fixed and permeabilized with the Cytofix/Perm kit (BD Pharmingen), followed by incubation with anti-mouse CTLA-4 (UC10-4F10-11; BD Pharmingen) or Foxp3 (FJK-16s; eBioscience) for 30 min at 4°C. Cells were washed twice in Perm/Wash Buffer (BD Pharmingen) and resuspended in PBS before analysis. All samples were acquired using a FACScan, FACSCalibur, or LSR-II cytometer (BD Biosciences) and data analyses were performed using WinList (Verity Software House) and FlowJo (Tree Star). The percentage of cells staining with a particular reagent or reagents was determined by subtracting the percentage of cells staining nonspecifically with the isotype control mAb from those staining in the same dot plot region with the anti-mouse mAbs.
Isolation of thymic epithelial cells and implantation into FP THY grafts
Thymi of 7-wk-old BALB/c mice were finely minced and rinsed with RPMI 1640 medium. The tissue fragments were gently stirred in RPMI 1640 medium for 30 min at room temperature to release the bulk of free thymocytes. The remaining tissue fragments were digested by collagenase/dispase/deoxyribonuclease in four incubations of 30 min each and pooled. To remove hematopoietic cells from the digested cell suspension, CD45+ cells were depleted using anti-mouse CD45 microbeads (Miltenyi Biotec) according to the manufacturer’s instructions. Less than 5% of these cells were positive for CD45. Freshly isolated thymic epithelial cells were injected into FP THY grafts at the time of initial transplantation and/or 5 wk later.
Histology and immunofluorescence staining
Various organs were fixed with 10% formalin and processed for H&E staining. FP THY/LIV grafts and normal BALB/c murine and porcine thymic tissues were cryosectioned and fixed with cold acetone for 10 min. After blocking with 1% BSA for 20 min, tissue sections were incubated with pretitrated primary rat anti-mouse I-A/I-E Ab (2G9; BD Biosciences) and rabbit anti-cytokeratin Ab (Z622; DakoCytomation) for 45 min at room temperature, washed, and incubated with FITC-conjugated goat anti-rat IgG and Texas Red-conjugated goat anti-rabbit IgG (Vector Laboratories). After washing with PBS, the coverslips were mounted onto slides using DakoCytomation fluorescent mounting medium.
MLR and suppression assay
To compare the suppressive capacity of CD4+CD25+ T cells from FPG nude mice and normal BALB/c mice, we performed MLR coculture assays. CD4+CD25− and CD4+CD25+ T cells were isolated from splenocytes by MACS using the mouse CD4+ T Cell Isolation Kit and CD25 microbeads (Miltenyi Biotec) according to the manufacturer’s instructions. Pooled CD4+CD25− T cells from normal BALB/c mice (8 × 104/well) were incubated with irradiated (30 Gy) C57BL/6 splenocytes (8 × 104/well) and CD4+CD25+ T cells isolated from individual normal BALB/c or FPG nude mice at various ratios in U-bottom 96-well plates in RPMI 1640 medium supplemented with 15% CPSR-3 (Sigma-Aldrich) with 2 mM l-glutamine, 0.1 mM nonessential amino acids (Life Technologies), 1 mM sodium pyruvate, 10 U/ml penicillin, 10 μg/ml streptomycin, 1% HEPES buffer, and 1 × 10−5 M 2-ME (Sigma-Aldrich) at 37°C in 5% CO2 for 5 days. Cells were harvested after 16 h of incubation with 1 μCi of [3H]thymidine. [3H]Thymidine incorporation was measured by a beta counter. Data are expressed as stimulation index (cpm of stimulated culture/cpm of unstimulated culture). Unstimulated control cultures were the same responder cells incubated in medium alone.
All data are presented as the mean ± SEM. Statistical analyses were performed using Student’s t test with Welch’s correction; one-way ANOVA and two-way repeated measure ANOVA (for analyses of weight change) were used to compare groups with GraphPad Prism software. A p < 0.05 was considered to be statistically significant.
Reconstitution of mouse T cells and phenotypic characteristics of CD4+CD25+ T cells in FPG BALB/c nude mice
To confirm T cell reconstitution and assess the level of regulatory T cell (Treg) development in BALB/c nude mice grafted with FP THY/LIV (FPG nude mice), we evaluated peripheral reconstitution of mouse T cells by FCM. Increased numbers of CD4+ cells and of CD4+CD25+ cells were detected in spleens of FPG nude mice compared with those from normal BALB/c and BALB/c nude controls 11 wk after FP THY/LIV transplantation (Fig. 1 A). The percentages of CD25+ cells among CD4 cells were similar between the two groups (data not shown). Thus, FP THY/LIV transplantation allows excellent peripheral reconstitution of mouse CD4+ and CD4+CD25+ T cells in BALB/c nude mice.
Although constitutive CD25 expression is a marker for THY-derived regulatory CD4+ T cells (Treg), CD25 is also expressed on activated T cells. To distinguish whether CD4+CD25+ T cells in FPG nude mice were activated T cells or Treg, the expression of Foxp3, a unique transcription factor in Treg, was analyzed in CD4+CD25+ cells of FPG nude mice by intracellular staining. As shown in Fig. 1,B, CD4+CD25+ cells in FPG nude mice were predominantly Foxp3+. Almost all CD4+CD25+ cells expressed Foxp3 (Fig. 1,C) and the intensities of Foxp3 in CD4+CD25+ cells of FPG nude mice were similar to those in normal BALB/c mice (Fig. 1 B). Significantly increased proportions of CD4+ cells expressing CTLA-4, another characteristic marker for Treg, were detected among CD4+CD25+ cells from FPG nude mice (71.8 ± 5.2%) compared with normal BALB/c controls (63.3 ± 6.5%, p < 0.05). These data show that CD4+CD25+ T cells of FPG nude mice express typical Treg markers.
Cotransfer of normal CD4+ or CD8+ splenocytes can prevent autoimmune disease in adoptive recipients
We have previously reported that normal BALB/c splenocytes (nSPL) can suppress the ability of splenocytes from FPG nude mice (GSPL) to induce autoimmune disease in secondary BALB/c nude recipients (4). To identify splenocyte populations with this suppressive activity, we cotransferred various cell subsets of nSPL along with GSPL to lightly irradiated (3 Gy) BALB/c nude mice. Cotransfer of 107 CD4+ nSPL completely suppressed the weight loss and tissue injury induced by 2 × 107 GSPL in secondary BALB/c nude recipients. In contrast to recipients of GSPL alone, secondary recipients of a mixture of GSPL and CD4+ nSPL showed increasing weight and no clinical signs of autoimmune disease, similar to recipients of 2 × 107 nSPL alone (data not shown). Cotransfer of 107 CD8+ nSPL also protected secondary BALB/c nude recipients from any clinical or histologic evidence of autoimmune disease induced by 2 × 107 GSPL (data not shown). These results are consistent with the ability of both subsets of syngeneic T cells to suppress autoimmunity induced by lymphocytes from nude rats receiving hamster THY xenografts (9).
To quantitatively compare the suppressive activity of CD4+ and CD8+ nSPL, we cotransferred various doses of CD4+ or CD8+ nSPL. As shown in Fig. 2, cotransfer of 107 or 3 × 106 CD4+ nSPL suppressed the autoimmunity induced by 2 × 107 GSPL in secondary BALB/c nude recipients to a similar extent. Fig. 2 also illustrates that the potency of CD8+ nSPL as suppressors of the disease was lower than that of CD4+ nSPL, since cotransfer of 3 × 106 CD8+ nSPL in the same experiment failed to suppress autoimmunity, while 107 CD8+ nSPL again completely suppressed the disease.
The phenotype of CD4+ cells in spleen and lymph nodes of nude mice 7 wk after adoptive transfer of GSPL was consistent with a high level of activation and possibly lymphopenia-driven expansion. CD4+ cells in both spleen and lymph nodes of recipients of GSPL alone showed significantly increased levels of CD69 and CD44 and decreased CD62L and CD45RB expression compared with those of recipients of nSPL (data not shown). Cotransfer of CD4+ nSPL with GSPL led to significantly reduced levels of CD69 and CD44, with increased proportions of CD62L+ and reduced proportions of CD45RBlow cells compared with secondary recipients of GSPL alone (data not shown). These effects of nSPL CD4 cells may reflect reduced activation of GSPL-derived CD4 cells and/or dilution by nonactivated, nonexpanding nSPL-derived CD4 cells. However, cotransfer of CD8+ nSPL led to reduced CD69 expression on GSPL-derived CD4 cells, indicating that they suppressed GSPL CD4 cell activation. nSPL-derived CD8 cells did not, however, preserve CD62L or CD45RB expression on GSPL CD4+ cells (data not shown).
Cotransfer of normal CD4+CD25+ or CD4+CD25− splenocytes can inhibit autoimmune disease in secondary recipients
To identify subpopulations of CD4+ nSPL with suppressive activity, we fractionated cotransferred CD4+ nSPL populations. As shown in Fig. 3,A, cotransfer of 9 × 105 CD4+CD25+ nSPL completely suppressed the autoimmunity induced by 2 × 107 GSPL in secondary BALB/c nude recipients. This suppression was dependent on the number of CD4+CD25+ nSPL cotransferred, as shown in Fig. 3 B.
Cotransfer of 5 × 106 CD4+CD25− nSPL also initially suppressed the autoimmunity induced by 2 × 107 GSPL (Fig. 3,A). However, as shown in Fig. 3 B, CD4+CD25− nSPL were less potent suppressors than the CD25+ subset, as 4 × 106 CD25− nSPL reduced the late weight loss in adoptive recipients less markedly than approximately one-tenth that number of CD4+CD25+ nSPL.
Defective accumulation in lymphopenic hosts of T cells generated in FP THY/LIV grafts
The ability of CD4+ nSPL, including CD25− cells, as well as CD8+ nSPL to suppress T cell activation and autoimmune disease induced by GSPL raised the possibility that their suppressive function might be due in part to their ability to respond to lymphopenia-driven stimuli and hence “compete for resources” (10). To address this possibility, we compared the accumulation of T cells derived from FPG nude mice to those from normal BALB/c mice following adoptive transfer into BALB/c nude mice. As shown in Fig. 4,A, transfer of 3 × 105 CD4+CD25− GSPL led to expansion in secondary recipients, but there was significantly reduced accumulation in peripheral blood at 3 and 6 wk (p < 0.05 and p < 0.01, respectively; Fig. 4,A) compared with recipients of similar cell populations from nSPL. No significant differences were detected at 6 wk in splenic or LN T cell accumulation in recipients of 3 × 105 CD4+CD25− GSPL vs those receiving 3 × 105 CD4+CD25− nSPL. When 3 × 106 CD4+CD25− GSPL cells were transferred, significantly reduced CD4 T cell accumulation was detected in blood at 3 and 6 wk (p < 0.001), and in both spleen and LNs at 6 wk (p < 0.01) compared with recipients of 3 × 106 CD4+CD25− nSPL (Fig. 4 A). These data are consistent with a defect in lymphopenia-driven expansion or survival among CD4+CD25− cells generated in a xenogeneic thymic graft.
Since CD4+CD25+ Treg have been shown to require self- class II MHC for their maintenance in the periphery (11), we performed similar comparisons following adoptive transfer of 3 × 105 CD4+CD25+ nSPL or GSPL. As shown in Fig. 4 B, although evidence for expansion of the injected cells was observed in both groups, the accumulation of these cells in blood, spleen, and lymph nodes at 6 wk was also significantly reduced in recipients of CD4+CD25+ GSPL vs those receiving CD4+CD25+ nSPL (p < 0.001, p < 0.05, p < 0.05, respectively). These data are consistent with a defect in lymphopenia-driven expansion and/or survival among Treg generated in a xenogeneic thymic graft.
Abnormalities in both CD4 effector cells and Treg developing in FP THY/LIV grafts
To further analyze the etiology of autoimmunity in secondary BALB/c nude mouse recipients of GSPL, we asked whether CD25+ cells or CD25− cells from these animals are responsible for inducing autoimmunity in secondary recipients. As shown in Fig. 5,A, adoptive transfer of 3 × 105 CD4+CD25+ GSPL or nSPL did not cause clinical evidence of autoimmune disease in secondary BALB/c nude recipients. In contrast, transfer of 3 × 105 or 3 × 106 CD4+CD25− GSPL led to rapid weight loss (Fig. 5,A) and other signs of autoimmune disease (data not shown) in secondary BALB/c nude mouse recipients. Transfer of up to 3 × 106 CD4+CD25− nSPL did not cause autoimmunity in secondary BALB/c nude recipients (Fig. 5 A). Thus, CD25− and not CD25+CD4 T cells generated in porcine THY xenografts have an increased tendency to induce autoimmunity compared with those that develop in a normal mouse THY.
Although cotransfer of 5 × 105 CD4+CD25+ nSPL completely suppressed the weight loss (Fig. 5 B) and all clinical evidence of autoimmunity in BALB/c nude mice receiving 2 × 107 GSPL (data not shown), similar numbers of CD4+CD25+ GSPL did not suppress autoimmunity. Thus, Treg in spleens of FPG nude mice were defective in the ability to suppress autoimmune disease.
Phenotypic analyses and in vitro functional analyses of Treg from GSPL were consistent with this result. As shown in Fig. 6,A, CD4+CD25+ T cells from FPG nude mice showed a marked defect, compared with those from normal BALB/c mice, in their ability to suppress alloresponses of syngeneic BALB/c CD4+CD25− T cells. The phenotype of Treg from FPG mice was also abnormal, with increased CD44, increased CD45RB, and decreased CD62L expression compared with Treg from normal BALB/c mice (Fig. 6, B and C).
Because CD25+“natural” Treg develop intrathymically, we next assessed the ability of thymocytes derived from FPG nude mice to suppress autoimmunity in the adoptive transfer model. As shown in Fig. 7,A, cotransfer of 2 × 107 thymocytes containing 7.4 × 104 or 10.2 × 104 CD4+CD8−CD25+ cells from FP THY grafts harvested 5 or 11 wk after implantation, respectively, not only failed to suppress autoimmunity induced by GSPL, but actually accelerated autoimmunity. Adoptive transfer of thymocytes (2 × 107) from FPG nude mice alone also caused autoimmunity. In contrast, similar numbers of thymocytes from normal BALB/c mice (normal thymocytes) completely suppressed the autoimmunity induced by GSPL and did not cause autoimmunity on their own (Fig. 7 B). These data demonstrate that selection in a porcine thymic xenograft results in an intrinsic tendency of T cells to cause autoimmune disease.
Correction of autoimmune defects by coimplantation of mouse thymic epithelial cells with porcine THY grafts
We hypothesized that the addition of murine, host-type thymic epithelial cells (mTEC) to the porcine thymic graft might prevent the development of autoimmunity by improving the negative selection of host-specific T cells and/or by promoting the positive selection of CD25+CD4+ regulatory cells. We evaluated the effect of injecting normal BALB/c mTEC obtained by collagenase/dispase/DNase digestion and negative selection (with anti-CD45 MACS beads) into porcine THY grafts at the time of implantation into BALB/c nude mice and/or 5 wk later (by laparotomy and injection into the graft, which had enlarged markedly by this time). As shown in Fig. 8,A, GSPL from FPG nude mice that received FP THY/LIV grafts into which normal BALB/c mTEC were injected on day 0 and at 5 wk induced significantly less wasting syndrome in BALB/c nude mouse adoptive recipients. Moreover, cotransfer of GSPL from recipients of FPG with mTEC partially suppressed the wasting syndrome induced by GSPL from mice grafted without normal BALB/c mTEC. Fig. 8,A also shows that injection of normal BALB/c mTEC at 5 wk, without the day 0 injection, did not reduce the ability of GSPL to induce wasting syndrome. Two-color immunohistochemical staining of the thymic grafts that were injected with mTEC, using anti-mouse class II mAb and an Ab specific for cytokeratin (Fig. S15), clearly showed the presence of mTEC in the grafts of animals in which these cells had been injected at day 0 and 5 wk (Fig. 8,B). However, the FP THY grafts injected with mTEC only at week 5 contained very few mTEC (Fig. 8 B).
Fetal porcine thymic grafts can efficiently support mouse and human thymopoiesis, resulting in the generation of a T cell repertoire that is specifically tolerant of the xenogeneic donor (1, 2, 12). Recent extension of this approach using vascularized porcine thymic grafts in a nonhuman primate model has allowed, for the first time, acceptance of life-supporting α-1,3-galactosyl transferase knockout porcine kidney grafts for months with no evidence of rejection (13).
Although normal mice that are thymectomized and T cell depleted usually display excellent health following T cell reconstitution from FP THY grafts, a high proportion of FP-grafted nude mice, which lack a native THY, eventually develop a multiorgan autoimmune syndrome characterized by murine CD4 cell infiltration of lung, liver, ovary, and intestine (4).
A similar phenomenon has been reported in nude rats receiving fetal hamster thymic xenografts, in which disease transfer to secondary recipients could also be blocked by CD4+ or CD8+ T cells from euthymic syngeneic donors (9). Normal numbers of cells with phenotypic characteristics of Treg were detected in the THY-grafted animals (9). In our pig-to-mouse model, we used Foxp3 staining to demonstrate that increased numbers of Treg repopulate the periphery of THY xenograft recipients compared with euthymic mice and that sorted Treg from these animals show defective suppressive activity in vitro and in vivo. Natural Treg are CD25+CD4+Foxp3+ T cells that are generated intrathymically (14, 15).
However, abnormal natural Treg are not the sole cause of autoimmunity, as we have demonstrated an abnormal propensity of sorted CD25− T cells from xenogeneic THY-grafted mice to cause autoimmunity, as well as abnormalities in the homeostasis of both effector and regulatory CD4+ T cells originating in xenogeneic THY grafts. Our data show that, although both CD4+ and CD8+ cells from euthymic syngeneic mice suppress the ability of adoptively transferred T cells from THY xenograft recipients to cause autoimmunity, the potency of the CD4 subset is greater than that of CD8 cells. Moreover, both CD25+ and CD25− CD4 cells could suppress disease, although the CD25+ subset showed more than 10-fold greater potency.
The increased susceptibility of FPG congenitally athymic nude mice to the autoimmune syndrome compared with immunocompetent mice may reflect the presence in the latter of Treg in the periphery before T cell depletion with mAbs. Some of these Treg may persist after conditioning, protecting the mice from autoimmunity following porcine thymic implantation, as BALB/c Treg have been shown to be relatively resistant to in vivo CD4 depletion with GK1.5 (16), the mAb used in our studies.
The reduced capacity of both effector T cells and Treg from FPG mice to accumulate in T cell-deficient hosts may contribute to the autoimmunity in nude mice by failing to compete with activated autoreactive T cells for cytokines and other “resources.” A defect in lymphopenia-driven expansion could explain the reduced accumulation of adoptively transferred T cells from FPG mice compared with those from normal mice in secondary BALB/c nude recipients. The same MHC-peptide complexes responsible for positive selection may be required in the periphery to allow optimal T cell survival and expansion in a lymphopenic environment (17), and such complexes are not found in the periphery by T cells positively selected by a xenogeneic thymic MHC. We observed a slight increase in the decay of memory-type cells from FPG compared with syngeneic murine thymic grafts following graftectomy (18), consistent with a defect in homeostatic expansion.
The inability of thymocytes and peripheral CD4+CD25+ cells from FPG nude mice to suppress the autoimmunity induced by CD25−CD4+ T cells from FPG nude mice most likely reflects a reduced ability of the porcine thymic epithelium to positively select Treg that recognize murine MHC-peptide complexes in the periphery. The thymic epithelium plays an important role in the positive selection of Treg (19, 20, 21, 22). We have demonstrated that conventional CD4 T cells in FPG mice are positively selected by porcine MHC, with no measurable contribution from the murine MHC (5, 6) and the same is likely to apply to Treg. Implantation of normal BALB/c thymic epithelial cells in FP THY grafts enhanced the ability of GSPL to suppress autoimmunity, suggesting an approach to overcoming this defect. Additionally, our large animal studies utilize primarily vascularized thymic tissue rather than thymic fragments, and studies in the rat-to-mouse model suggest that vascularization can prevent autoimmunity in THY xenograft recipients, in association with enhanced Treg generation (23).
Treg require interactions with self-MHC in the periphery to become fully functional (11) and the same peptide-MHC complex in the periphery as that which led to positive selection in the THY is needed to promote lymphopenia-driven expansion (24). Consistent with a failure of Treg developing in THY xenografts to re-encounter the selecting ligands in the periphery, we observed less expansion of Treg transferred from FPG mice than those from normal mice in T cell-deficient recipients. Thus, Treg positively selected on porcine MHC-peptide complexes in the THY may fail to interact effectively with the murine MHC-peptide complexes in the periphery and hence fail to become fully functional and capable of suppressing autoimmunity. Moreover, reduced homeostatic expansion may result in failure of Treg to compete with autoreactive T cells with higher affinity for self-Ags transferred to nude mice, resulting in reduced suppression of autoimmunity (10).
Porcine THY xenografts are nevertheless capable of generating functional mouse Treg that suppress antidonor responses (25) and specifically prolong donor skin graft survival in an adoptive transfer model (Y. Zhao and M. Sykes, unpublished data). Sun et al. (26) reported that nude mice receiving neonatal porcine thymic grafts have Foxp3+ Treg in the periphery that suppress antipig responses but show defective suppression of alloresponses in vitro, consistent with our results. However, in contrast to our results, these cells suppressed autoimmunity following adoptive transfer to nude mice (26). This discrepancy most likely reflects the much greater ratio of Treg to CD25− CD4 cells transferred by Sun et al. (26) and the fact that the CD25− cells in their study were derived from normal BALB/c mice rather than FPG mice (see below). Together, these studies suggest that the defect in suppression of autoreactivity by Treg from FPG mice is relative and not absolute. Effective suppression of antipig responses with reduced suppression of antihost responses is consistent with an exclusive role of the xenogeneic thymic epithelium in positively selecting Treg. Although further studies are needed to address this hypothesis, the reduced (compared with normal syngeneic mice) capacity of Treg from FPG mice to suppress alloresponses may reflect the same phenomenon. Since the potency of alloresponses is thought to reflect the cross-reactivity with allogeneic MHC/peptides of T cells positively selected on self-MHC-peptide complexes, the same phenomenon may result in the presence of a relatively high frequency in the normal Treg repertoire of cells that suppress alloresponses. This situation would not prevail among Treg selected by a xenogeneic MHC. Additional effects of the xenogeneic disparities between pig and mouse may also contribute to the reduction in Treg function. The altered phenotype of Treg in FPG mice resembles that observed in cyclosporine-treated mice, in which regulatory function was also reduced (27). The reduced CD62L and increased CD45RB expression we observed is consistent with the reduced suppressive function, as the most potent suppression has been attributed to CD62L+ and CD45RBlow subsets of CD25+ CD4 T cells (28, 29, 30).
However, a lack of natural Treg specific for recipient Ags cannot explain the increased autoimmunity induced by CD25−CD4+ T cells and thymocytes from FPG mice. These data suggest a possible defect in the negative selection of host-reactive thymocytes in FP THY grafts. Although mouse MHC class II+ hematopoietic cells in porcine THY grafts contribute effectively to negative selection of murine thymocytes (5, 6), the thymic epithelium “ectopically” produces proteins that are otherwise produced only in specialized peripheral organs (31, 32, 33). FP THY grafts, whose epithelium only expresses porcine tissue-specific Ags, might therefore fail to delete mouse tissue-specific Ag-reactive T cells. Consistently, the successful engraftment of normal BALB/c mTEC in FP grafts reduced the capacity of thymocytes from FP grafts to induce autoimmunity.
In the fetal hamster to nude rat thymic transplantation model discussed above, autoimmunity could be prevented by coimplanting several types of syngeneic fetal epithelial tissues, including the THY, but not heart tissue, into the graft (9, 34). Although these results suggested that the addition of recipient epithelial cell Ags might prevent the autoimmunity, our study is the first to demonstrate that purified host-type thymic epithelial cells can ameliorate disease, persist long-term within the thymic xenograft, and allow the development of regulatory cells that can inhibit autoimmunity upon adoptive transfer.
Consistent with our observation that normal CD8 T cells suppressed autoimmune potential and CD4 cell activation from GSPL, a paucity of regulatory CD8 cells was implicated in autoimmunity in rat to nude mouse and hamster to nude rat thymic transplantation models (9, 35, 36, 37). CD8+ Treg have been implicated in a number of tolerance models (38, 39, 40, 41). Phenotypically mature, normal, and functional murine CD8 single-positive T cells are effectively generated in porcine thymic grafts, but these cells fail to repopulate the periphery (42). Fortunately, this defect is not observed for human T cells developing in porcine thymic grafts (12, 43).
Self-MHC-peptide complexes down-modulate the function of autoreactive T cells (11). It is possible that the absence in the periphery of FPG mice of the thymic (porcine) positively selecting MHC-peptide complexes leads to a failure of this modulation. Consistent with this possibility, increased levels of the CD69 activation marker were detected on adoptively transferred peripheral T cells from FPG nude mice compared with normal mice. Because CD69 is not up-regulated on T cells that expand due to lymphopenic stimuli (17, 44, 45, 46, 47), the up-regulation of this marker strongly suggests T cell activation.
Despite all of the possible ramifications of positive selection of thymocytes on a xenogeneic epithelium discussed above, the T cell repertoire that is positively selected in a xenogeneic THY is broad (48, 49), and this may allow sufficient cross-reactivity on host MHC to explain the robust responses to protein Ags and ability to clear opportunistic infection that is conferred by xenogeneic THY grafting (3). Our data suggest that positive selection of both effector cells and Treg recognizing recipient MHC may be enhanced by adding recipient thymic epithelial cells to the xenografts.
In summary, our data implicate several mechanisms in the propensity of T cells from FPG mice to induce autoimmunity in nude mice, including defects in positive selection and peripheral activation of host-specific Treg, defects in negative selection, defects in peripheral modulation of autoreactive T cells, and possible defects in lymphopenia-driven expansion and survival. All of these defects can be potentially overcome by coimplantation of recipient-type thymic epithelial cells. Further exploration of this approach will be of considerable importance for the clinical applicability of this promising approach to inducing xenograft tolerance.
We thank Drs. Kazuhiko Yamada and Yong-Guang Yang for helpful review of this manuscript and Kelly Walsh for expert assistance with its preparation. We also thank Dr. David H. Sachs and Scott Arn for providing porcine tissues and reagents, James Winter for assistance with porcine surgeries, and Guiling Zhao for technical assistance.
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 P01 AI045897 and P01 HL18646, Juvenile Diabetes Research Foundation Basic Science Grant 1-2007-723, and Immerge BT. Y.F. was supported in part by the Uehara Memorial Foundation. T.O. was supported in part by a Postdoctoral Fellowship for Research of Abroad from the Japan Society for the Promotion of Science.
Abbreviations used in this paper: ATX, thymectomized; FP, fetal pig; THY, thymus; FPG, FP THY-grafted; FP THY/LIV, FP THY and liver tissue; GSPL, splenocytes from FPG mice; nSPL, normal splenocytes; LN, lymph node; FCM, flow cytometry; Treg, regulatory T cell; mTEC, murine thymic epithelial cell.
The online version of this article contains supplemental material.