|Home | About | Journals | Submit | Contact Us | Français|
Clinical transplantation for the treatment of end-stage organ disease is limited by a shortage of donor organs. Successful xenotransplantation could immediately overcome this limitation. The development of homozygous α1,3-galactosyltransferase knockout (GalT-KO) pigs removed hyperacute rejection as the major immunologic hurdle to xenotransplantation. Nevertheless, GalT-KO organs stimulate robust immunologic responses that are not prevented by immunosuppressive drugs. Murine studies show that recipient thymopoiesis in thymic xenografts induces xenotolerance. We transplanted life-supporting composite thymokidneys prepared in GalT-KO miniature swine to baboons in an attempt to induce tolerance in a pre-clinical xenotransplant model. Here, we report the results of 7 xenogenic thymokidney transplants using a steroid-free immunosuppressive regimen that eliminated whole body irradiation in all but 1 recipient. The regimen resulted in average recipient survival of over 50 days. This was associated with donor-specific unresponsiveness in vitro and early baboon thymopoiesis in the porcine thymus tissue of these grafts, suggesting the development of T cell tolerance. The kidney grafts had no signs of cellular infiltration or deposition of IgG, and no grafts were lost due to rejection. These results show that xenogeneic thymus transplantation can support early human thymopoiesis, which in turn may induce T cell tolerance to solid organ xenografts.
Despite recent advances in the field of transplantation, there remains a large discrepancy between the number of patients that could benefit from transplantation and the number of available donor organs. Recently, techniques for reprogramming adult cells by gene transduction to produce pluriopotent stem cells have renewed interest in the possibility of tissue regeneration for organ repair while avoiding ethical issues associated with embryonic cell use(1). A recent report on the generation of rat hearts from stem cells noted that these structures expressed only 2% of the functional capacity of a normal rat heart(2). Such results indicate that this technology needs significant development before preclinical testing. Therefore, xenotransplantation of solid organs remains at the forefront of the search for a solution to the organ shortage.
The pig is generally considered the most suitable donor species for xenotransplantation(3). Previously, the major obstacle preventing successful transplantation of porcine organs into primates was the existence of natural antibodies against a terminal saccharide epitope, galactose-α1,3-galactose (Gal), produced by α1,3-galactosyltransferase (GalT)(4). While this gene is functional in pigs, it is not functional in Old World monkeys and humans. These antibodies comprise over 70% of preformed primate anti-pig antibody activity(3). Transplants performed across this barrier result in antibody-mediated graft rejection. The generation of homozygous GalT knockout (GalT-KO) pigs in 2002(5–7) removed this barrier to solid organ xenotransplantation. Utilizing GalT-KO pigs as donors, we and others reported studies of life-supporting kidney xenografts in baboons receiving chronic immunosuppression(8,9). While these grafts did not undergo hyperacute rejection, outcomes were not significantly better than those achieved using kidneys from pigs overexpressing human decay accelerating factor (hDAF)(10–12), indicating that additional strategies are required to achieve the clinical application of xenotransplantation.
We have shown that xenogeneic T cell responses between pigs and humans are at least as strong as allogeneic responses in vitro(13,14). Because T cells enhance B cell and NK cell activity in vivo, xenogeneic cellular responses are generally stronger than allogeneic cellular responses. In vivo, we have shown cellular infiltrates with antibody deposits in rejected GalT-KO kidneys(8). Others have reported rapid and complete rejection of GalT-KO kidneys with the development of high levels of induced non-Gal anti-Ab, including IgG, which is likely due to T cell activation(9). All these results support the importance of using a T cell tolerance strategy in xenotransplantation.
We have previously shown that co-transplantation of porcine thymus tissue as a vascularized graft can induce tolerance across full allogeneic barriers to kidneys and hearts in a miniature swine model(15–17). Moreover, xenogeneic tolerance can be achieved via thymic xenotransplantation in mice and humanized mouse models(18–20). When this strategy was applied in the pig-to-baboon model using hDAF composite thymus and kidney (thymokidney) grafts, the kidneys were rejected due to anti-Gal antibodies by day 29, although there was clear evidence for T cell unresponsiveness in vitro(21). Using GalT-KO donors, which eliminate humoral rejection by anti-Gal antibodies, a survival advantage was conferred by the vascularized thymus graft, resulting in life-supporting renal xenograft survival up to 83 days with normal creatinine levels(8). However, the initial immunosuppressive regimens used to facilitate the induction of tolerance by the thymus grafts included administration of steroids and whole body irradiation (WBI) and were associated with a high incidence of early post-operative complications, mainly infectious in nature, which decreased average survival to 34 days(8). In order to develop a more clinically applicable regimen, we have now eliminated whole body irradiation and corticosteroids from the treatment protocol.
Here, we report that the modified protocol resulted in an average survival greater than 50 days, and no graft was lost due to rejection. Additionally, for the first time we were able to demonstrate that the porcine thymus tissue supported early baboon thymopoiesis. This finding was associated with donor-specific unresponsiveness in vitro. Histological analysis revealed no signs of cellular infiltration or deposition of IgG. These results suggest that vascularized thymus tissue transplantation may be able to induce tolerance across the xenogeneic barrier and represent a step toward clinical application of this strategy.
All animals were cared for according to the guidelines of the Massachusetts General Hospital Institutional Animal Care and Use Committee. Recipient baboons (Papio hamadryus, n=7) weighing 7 to 12 kg were purchased from Mannheimer Foundation, Homestead, FL. Xenogeneic organs were obtained from by GalT-KO miniature swine (n=7) weighing 9 to 27 kg. The generation of these animals has been published previously(6). All pigs were produced from either GalT-KO homozygous × heterozygous or homozygous × homozygous breeding (22).
Seven thymokidneys were prepared by implantation of autologous thymic tissue under the kidney capsule as previously described (23).
All operations were performed under general anesthesia. Seven baboons were thymectomized 3 weeks prior to thymokidney transplantation. Indwelling plastic catheters (Saint Gobain Performance Plastics, Reading, PA) were inserted into the carotid artery and the internal and external jugular veins 7 days prior to thymokidney transplant and maintained on a flexible tether and primate jacket (Lomir Biomedical, Malone, NY). All animals were splenectomized immediately prior to transplantation of the thymokidney, but during the same procedure. The thymokidney transplantation procedure is identical to that of a normal orthotopic renal transplantation (21). All anastomoses were completed in less than 40 minutes, and all kidney grafts produced urine within 5 minutes of revascularization.
The immunosuppressive regimens used are shown in detail in Table 1.
Renal graft biopsies were performed for creatinine levels > 3.0 ng/dl or at the time of necropsy. Since data in our pig model demonstrated that a biopsy of the thymus interferes with the induction of tolerance to renal transplants (24), thymus histology was obtained only at completion of the experimental period. Tissue samples were fixed in 1% formaldehyde, embedded in paraffin, and subsequently sectioned. Tissues were stained using either hematoxylin and eosin (H&E), or periodic acid-Schiff (PAS). Coded samples were examined by light microscopy, and rejection was diagnosed according to a standardized grading system (25). Immunohistochemistry was performed on frozen sections using polyclonal antibodies reactive to primate CD3 (polyclonal rabbit anti-human CD3. A0452, Dako, Denmark), primate CD4 (mouse anti-human CD4 mAb. 1F6, Zymed, CA), primate CD8 (mouse anti-human CD8, BD Pharmingen), cytokeratin (mouse anti-human cytokeratin mAb. AE1/AE3, Dako, Denmark), CMV (mouse anti-CMV mAb, DDG9 + CCH2, Dako, Denmark) IgM (polyclonal rabbit anti-human IgM, Dako, Denmark ), IgG (polyclonal rabbit anti-human IgG, Dako, Denmark), complement C3 (polyclonal rabbit anti-human C3c, Dako, Denmark), and complement C5b-9 (mouse anti-human C5b-9 mAb, aE11, Dako, Denmark).
PBLs were prepared from whole blood as previously described (21).
MLR assays were performed as previously described (13).
CTL assays were performed by plating responder and stimulator (irradiated at 2,500 cGy) PBMCs at 2×105 cells per U-bottomed well. Media was supplemented with hu-IL-15 to a final concentration of 5 U/ml for 50% of the effectors. 2×106 target PBLs/ml were blasted with 0.002 mg/ml PHA (Difco) for 7 days at 37°C. Targets were split on day two with media supplemented with hu-IL-2 to a final concentration of 100U/ml. Targets were labeled with 51Cr and incubated at 37°C for 4 hours with effector cultures at a concentration of 4×103 target cells per well. Supernatants were harvested using the Skatron Supernatant Collection System (Skatron, Sterling, VA), and 51Cr release was measured using a gamma counter.
The presence of cytotoxic non-Gal antibodies was determined by complement mediated cytotoxicity assays, the details of which have been previously reported (26).
PBMC were isolated from recipient baboon whole blood. DNA was isolated using standard techniques. Quantitative PCR assays were performed as previously described (27) and bCMV copies were normalized to copies of baboon CCR5.
The presence of baboon CD3, CD4, CD8 and CD20 on the surface of lymphocytes and cells in the donor thymus was determined by direct flow cytometry using a Becton Dickinson FACScan microfluorimeter (Sunnyvale, CA). Juvenile naïve thymus tissue was obtained during recipient thymectomy and thymocytes used as a control for normal expression of CD antigens. The following antibodies reactive to primate CD antigens were utilized: CD3 (polyclonal rabbit anti-human CD3. A0452, Dako, Denmark), CD4 (mouse anti-human CD4 mAb. 1F6, Zymed, CA), CD8 (mouse anti-human CD8, BD Pharmingen), and CD20 (mouse anti-human CD20, BD Pharmigen). Staining was performed with 1×106 of PBMC or thymocytes resuspended in 100 µl of HBSS (Life Technologies, Grand Island, NY) containing 0.1% bovine serum albumin (BSA) and 0.05% NaN3. Cells were incubated for 30 minutes at 4°C in the dark with saturating concentrations of conjugated antibodies. After three washes, cells were analyzed by flow cytometry using propidium iodide gating to exclude dead cells. FITC-conjugated anti-mouse IgG1 and PE-conjugated anti-mouse IgG2a were used as controls for specific binding.
Cells from animal tissues were obtained from biopsy samples or isolated PBMC. The thymus portion was identified by gross inspection and a portion of the kidney immediately deep to the thymus was isolated as a control for sjTRECs. Tissue was homogenized mechanically using a filter membrane, and DNA was extracted using a DNeasy Qiagen tissue kit. DNA concentration was determined via Nanodrop spectrophotometer by measuring optical density at 260nm. All samples were checked twice for DNA concentration before QPCR reaction, and fell within 5% error margin.
Primers were derived for QPCR reaction from the 1170 bp band of the baboon sjTREC, with the expected PCR product containing the recombination signal sites of the δ-rec and ψJα regions. In a total volume of 25 ul, each reaction included 800 nM of ψJα forward primer: 5’- GGTGTCTCTGTCAACAAAGTTGATGC -3’, with 800 nM of δ-rec reverse primer: 5’- ATGACAAGTTCAGCCCTCCATGTC -3’, with 200 nM of labeled probe: 5’-/56-FAM/ CCCTGTCTGCTCTTCATTCACCATTCTCACGAG / 36-TAMSp/ -3’.
Other reagents included 0.25 ul Hot start Taq polymerase (Qiagen), 2 ul 10mM dNTPs, 150 ng of DNA template, 2.5 ul of 10× Hot start Taq buffer. The cycling conditions were as follows: 95 C × 15 minutes, followed by 50 cycles of the steps: 94 C × 30 s, 55 C × 60 s, 72 C × 30 s. All QPCR reactions and quantitative analyses were performed using a Strategene M×3005 machine.
We performed 7 thymokidney transplants using an immunosuppressive regimen designed to facilitate the induction of tolerance across the xenogeneic barrier. Previous regimens utilized in our laboratory included WBI and steroids during the induction period. We hypothesized that these interventions were associated with the high rate of infectious complications in the perioperative period. Additionally, WBI would likely suppress the production of T cell precursors by the bone marrow that would migrate to the pig thymus and undergo positive and negative selection, resulting in pig-specific T cell tolerance. In order to decrease the rate of perioperative infections and accelerate the development of thymopoiesis, steroids were eliminated from the protocol for all animals and whole body irradiation was eliminated from the protocol in 6 of 7 animals.
The average survival of animals in this series was 51 days (Fig 1), which compares favorably with the average survival of 34 days reported in our previous series. When we exclude one animal that died of a drug reaction to LoCD2b at POD 18, with normal creatinine (0.5 mg/dl) and histology, the average survival was greater than 56 days. No graft was lost due to rejection and no cellular infiltrates were identified on histology (Table 2). One animal died from CMV with invasion of the kidney, intestine and lung on POD 28 (Fig 2). In terms of technical complications, one animal, B187, required operative repair of the ureterovesicular anastomosis on POD 8, which was likely due to humoral rejection of the distal ureter by non-Gal NAb. Otherwise, there were no complications related to surgery.
Serum creatinine levels for the animals in this series are shown in Figure 1b. In the animals that developed significant proteinuria and edema, the creatinine levels fluctuated, but this was not associated with the development of cellular infiltrates or IgG deposition on histology (see next section). In animals that did not develop significant proteinuria, serum creatinine levels remained relatively stable.
The hematologic parameters of the animals in this series are shown in Figure 3. Following xenotransplantation, the white blood cell counts, hematocrit levels and platelet counts remained stable with the exception of the animal that developed a clinically significant CMV infection. In this animal, the platelet levels fell concurrently with a rise in baboon CMV DNA copies in the peripheral blood (Fig 2g). Importantly, no porcine virus transmission, including porcine CMV, was detected in the peripheral blood of this, or any other, animal, demonstrating that the infection was of host origin and not due to pig-derived infection (28).
Histology of several kidney grafts is shown in Figure 4. Light microscopy showed that although the general architecture of each graft was preserved, there were varying degrees of glomerulopathy and minimal thrombotic microangiopathy (Fig 4a,b,c,d). The degree of glomerulopathy and thrombotic microagiopathy seemed to be related to the degree of proteinuria, with the animals that had clinical edema demonstrating a greater degree of glomerulopathy on biopsy. Cellular infiltrates in the renal parenchymea were not seen even in the grafts from animals with proteinuria.
Immunohistochemical analysis of the grafts showed that all transplants had deposition of anti-pig non-Gal IgM in the kidney and ureter at early time points, with less deposition at late time points (Fig 4e, f). Decreased levels of anti-pig IgM over time suggested that the anti-non-Gal deposits were due to preformed antibodies. Small deposits of complement were seen in the grafts and they were associated with the deposition of IgM antibodies. There was no detectible IgG deposition in any graft. The lack of IgG deposition suggests a lack of T cell sensitization, which would cause class switching from IgM to IgG via the cognate interaction. Moreover, since the animals that demonstrated clinically significant edema with fluctuating creatinine levels did not have evidence of T cell infiltrates or IgG deposition, it appears that the changes in creatinine levels were due to fluctuations in intravascular volume status resulting from proteinuria and not from immunological rejection.
In order to confirm the histological findings showing a lack of sensitization to the xenogeneic thymokidneys, we analyzed the sera of the recipients pre and post transplantation for evidence of cytotoxic anti-pig antibodies. The recipients used in this study possessed preformed antibodies targeting non-Gal antigens, although at a level that did not produce hyperacute or delayed antibody mediated rejection. Recipient serum displayed approximately 11% cytotoxicty to GalT-KO PMBC pre-transplant due to preformed non-Gal antibodies circulating in the recipients’ serum. After transplantation, this level of cytotoxic activity decreased to near background levels. At the time of sacrifice, the levels of circulating cytotoxic antibodies remained below pre-transplant levels, indicating a lack of sensitization to non-Gal antigens (Fig 5a). This is consistent with the immunohistochemical analysis of the grafts showing only minimal IgM and no IgG deposition of anti-pig antibodies on the kidney grafts at the time of sacrifice.
In vitro assays were performed to determine if the recipients had evidence of donor-specific unresponsiveness. At the time the assays were performed, all experimental animals were still receiving immunosuppression. Although there was no response to stimulation with pig cells, MLR assays showed general hyporesponsiveness to allogeneic baboon stimulation (data not shown), consistent with therapeutic levels of immunosuppression, including anti-CD154 mAb, which is a potent inhibitor of CD4 T cells(29,30). In contrast, CML assays that assess cytotoxic responses mediated largely by CD8 T cells showed donor-specific unresponsiveness at days 49 and 78 in two of seven baboons, while the others were generally hyporesponsive to both donor and third-party allogeneic PBLs (Fig 5b). Although these animals remained on immunosuppression, this result suggested that thymus co-transplantation may have induced a state of T cell tolerance, abrogating cellular and minimizing humoral immune responses.
Induction of xenogeneic tolerance in pig to mouse or humanized mouse models is associated with mouse or human thymopoiesis, respectively, in the porcine grafts (18–20). We therefore examined whether the porcine thymus grafts were supporting the maturation of baboon T cells in the recipients. The thymus portion of the thymokidney grafts demonstrated typical thymic architecture with H&E staining, including Hassell’s bodies surrounded by lymphocytes (Fig 6a). Immunohistochemistry revealed that these areas corresponded to sites with viable porcine thymic epithelial cells (Figure 6b). Further analysis of 3 thymus grafts revealed that the morphologically lymphocytic cells in the porcine thymus tissue expressed baboon CD4, but not appreciable amounts of baboon CD3 (Fig 6c, d), characteristic of T cells undergoing thymopoiesis. In the other grafts, no CD4 or CD3 positive cells were seen, indicating that in these animals thymopoiesis had not progressed beyond the triple negative CD3/CD4/CD8 stage. However, it also demonstrated that the morphologically lymphocytic cells seen in the thymus portion of these grafts were not mature graft infiltrating CD3 T cells that migrated to this site from the peripheral blood.
In order confirm that individual cells displayed multiple CD antigens indicating thymopoiesis, we performed a more detailed analysis of the morphologically lymphoid cells in the thymic tissue of one of the recipients (sacrificed at Day 49) by isolating these cells and performing FACS analysis. A population of baboon cells in the thymus were found to be baboon CD4+/CD8+/CD3low, with expression of these markers similar to those found in a naïve baboon thymus (Fig 6e, f). There was no expression of porcine CD4, CD8 or CD3, confirming that these cells were of baboon origin. FACs analysis of the peripheral blood of this animal showed no cells that were CD4+/CD8+/CD3low, indicating that this phenotype was unique to the lymphocytic cells in the porcine thymus tissue. As a final confirmation, we also performed qPCR to determine if the cells had detectible numbers of T cell receptor excision circles (TRECs), indicating that proper genotypic rearrangement of the T cell receptor was occurring. The thymic tissue had detectible numbers of baboon TRECs, while the renal tissue immediately adjacent to the thymic tissue had undetectable levels of TRECs, showing the rearrangement was specific to cells in the porcine thymus tissue (Fig 6g). TREC assays performed with pig-specific primers did not detect any active porcine thymopoiesis, consistent with the absence of porcine cells in the thymus tissue found with FACS analysis.
This modified regimen extended the mean survival of discordant life-supporting renal xenografts to greater than 50 days without rejection. To our knowledge, this is the longest reported survival of life-supporting solid xenografts and the only strategy that avoids all adaptive rejection processes for such long periods of time. While the use of donors null for Gal expression was essential for the survival of these grafts, the results of this series are significantly different from the published results of other groups that used life-supporting GalT-KO kidneys without co-transplantation of donor thymus tissue, where maximum survival was 16 days (9).
The authors of the other prior study attributed the short survival time to the presence of anti-non-Gal antibodies in the recipients at the time of transplantation (9). The animals used in that study had less than 20% cytotoxicity to GalT-KO PBMC before the transplant, which is similar to the levels in our study. One major difference, however, was that the animals in our study did not develop sensitization to donor antigens, as demonstrated by the lack of anti-pig IgG on immunohistochemistry and in vitro assays. The animals in the prior study (9), however, had increased levels of anti-donor antibodies at the time of rejection, suggesting sensitization of the recipient to the graft. We believe that the presence of donor thymus tissue contributed to donor-specific T cell unresponsiveness in our model, which accounts for the disparate outcomes between our results and those reported previously (9).
Studies in our research center first showed that thymus transplantation could induce porcine-specific tolerance to donor porcine skin grafts in a pig thymus to mouse transplant model(19). It was further demonstrated that murine APC migrated to the porcine thymus and together with porcine APC residing in the thymus tissue mediated intrathymic negative selection of self and xenoreactive T cells (31). Positive selection in the thymus was found to be mediated only by porcine MHC(32). These two phenomena lead to specific tolerance to porcine skin grafts in the recipients(18). To ensure that the porcine thymus would support human thymopoiesis, these studies were extended to a humanized mouse model where the murine recipient’s immune system was reconstituted with human lymphocytes. We demonstrated that the porcine thymus supported human T cell development that resulted in a diverse human T cell repertoire with specific tolerance to porcine antigens(20,33).
Here, we report that the porcine thymus grafts support recipient thymopoiesis in a large animal pig to baboon model. This finding was associated with the lack of an elicited anti-porcine antibody response to the donor and also with the development of donor-specific unresponsiveness in vitro. This tolerance may involve the induction of baboon thymopoiesis in the pig thymus, as in the mouse and humanized mouse models(18–20). The inclusion of the donor thymus transplantation in our series, with the abrogation of T cell responses, could explain the disparate outcomes between this series and the previous report of isolated GalT-KO kidney transplantation. However, despite this advance, late complications remain that must be addressed before this strategy can be applied in the clinic.
One unresolved problem in this kidney model is the proteinuria associated with glomerulopathy that develops following transplantation. Although the proteinuria is unlikely due to rejection and is manageable clinically by administration of albumin, heavy proteinuria is an indicator of kidney damage at the glomerular level. In the clinic, transplantation glomerulopathy and proteinuria have been associated with both anti-ABO and anti-HLA antibodies(34). Other studies indicate that antibodies targeting aminopeptidases on podocytes contribute to proteinuria across xenogeneic barriers(35). Since all animals in this report had preformed antibodies at the time of xenotransplantation, and since the animals with clinically significant proteinuria developed symptoms in the first ten days, it seems possible that the preformed antibodies could be involved in causing the glomerular damage that leads to proteinuria. Our immunohistologic data support the possibility that the deposition of preformed non-Gal antibodies at the time of xenotransplantation promotes glomerulopathy and proteinuria, similar to ABO incompatible allotransplantation. This problem may be alleviated by pre-adsorption of non-Gal antibodies or by adding transgenes that could protect the glomeruli from complement-mediated damage, such as CD55, and thrombotic microangiopathy, such as CD39 and thrombomodulin. In order to achieve further improvement toward clinical application of this strategy, we are now collaborating with other centers to add these and other transgenes to the GalT-KO animals.
The results of this study suggest that, using these modified animals, human thymopoiesis in the porcine thymus may lead to tolerance induction across the xenogeneic barrier.
The authors would like to thank Fujisawa Healthcare, Inc (Deerfield, IL) for generously providing FK506. This work was supported by the following grant: NIH Program Project 5PO1-A145897, and Adam Griesemer received assistance from the AST/JDRF fellowship grant.
This work was supported by grants from the National Institutes of Health (PO1 AI045897-07 and U01 AI066331) and Adam Griesemer received support from the AST/JDRF Fellowship