Continued production of xenoimmune antibodies 6-8 years after clinical transplantation of fetal pig islet-like cell-clusters

Long-term IgG response to porcine Neu5Gc antigens without transmission of PERV in burn patients treated with porcine skin xenografts

Acellular materials of xenogenic origin are used worldwide as xenografts, and phase I trials of viable pig pancreatic islets are currently being performed. However, limited information is available on transmission of porcine endogenous retrovirus (PERV) after xenotransplantation and on the long-term immune response of recipients to xenoantigens. We analyzed the blood of burn patients who had received living pig-skin dressings for up to 8 wk for the presence of PERV as well as for the level and nature of their long term (maximum, 34 y) immune response against pig Ags. Although no evidence of PERV genomic material or anti-PERV Ab response was found, we observed a moderate increase in anti-αGal Abs and a high and sustained anti-non-αGal IgG response in those patients. Abs against the nonhuman sialic acid Neu5Gc constituted the anti-non-αGal response with the recognition pattern on a sialoglycan array differing from that of burn patients treated without pig skin. These data suggest that anti-Neu5Gc Abs represent a barrier for long-term acceptance of porcine xenografts. Because anti-Neu5Gc Abs can promote chronic inflammation, the long-term safety of living and acellular pig tissue implants in recipients warrants further evaluation.

Gal/non-Gal antigens in pig tissues and human non-Gal antibodies in the GalT-KO era

Our knowledge regarding Gal and non-Gal antigens in GalT-KO pig tissues can be summarized as α3Galactosyl-tranferase gene knock out eliminates the Galα3Galβ4GlcNAc-R antigen expression in pig tissues as well as anti-Gal antibody binding. Other Galα-terminating saccharides (e.g. iGb3 glycolipids and Galα2 determinants) may be present but have not been documented. α3Galactosyl-tranferase gene knock out slightly changes the carbohydrate antigen expression but no "new" antigens recognized by the human immune system have been found. Non-Gal antigens are both of protein and carbohydrate nature but their exact chemical structures are poorly defined. Regarding human non-Gal antibodies our knowledge is as Non-Gal antibodies exist naturally and increase in humans/non-human primate (NHP) receiving WT or GalT-KO pig grafts. Non-Gal antibodies with new antigen epitope recognition can be induced in humans/NHP after challenge by WT or GalT-KO pig grafts. Non-Gal antibodies react with both carbohydrates and proteins. Part of the protein reactivity is directed to glycoprotein carbohydrates chains. Non-Gal antibodies reacting with neuraminic acid terminated saccharides (both N-Acetyl and N-Glycoloyl variants) are present in humans/NHP. Anti-neuraminic acid antibodies are increased, as well as induced, after grafting pig organs into humans/NHP. Non-Gal antibodies does not cause hyperacute xenorejection but can be cytotoxic and cause xenoorgan damage. If humans sensitized to HLA antigens are at a higher risk of rejecting pig xenograft compared with non-sensitized individuals is not fully clarified. Clinical trials are needed to evaluate the relevance of non-Gal antigens/antibodies and for the xenofield to advance.

Correlation between insulin requirements and anti-galactose antibodies in patients with type 1 diabetes transplanted with neonatal pig islets

Pig xenografts represent an alternative source of organs for transplantation. Immunosuppression can prevent rejection, but involves high risk and cost. New anti-rejection techniques have been developed; however, results have not been successful. Few studies have reported xenoantibody levels in xenotransplanted patients with diabetes and no patients have reported a clinical correlation. We analysed anti-galactose (Gal) and other anti-pig antibody (APA) levels in xenotransplanted patients with type 1 diabetes and the relation to the clinical outcome. Twenty-three patients with type 1 diabetes were transplanted with porcine islets inside a device without immunosuppression. Demographic characteristics, insulin dose and xenoantibody levels at different periods were recorded. Anti-Gal and anti-pig antibodies were measured through indirect enzyme-linked immunosorbent assay (ELISA) and haemolytic anti-pig antibody assay. More than 50% were female; the mean current age, current diabetes duration, diabetes duration at xenotransplantation and time post-transplantation were: 20·8, 11, 5·5 and 5·7 years, respectively. Insulin doses remained with a mean reduction greater than 33% in more than 50% of the patients. The lowest anti-Gal antibody levels were related to the highest insulin dose reductions. This relationship could be explained by the device, Sertoli cells and accommodation process.

Unexpected immunoresponse to Gal and APA antigens in diabetic type 1 patients receiving neonatal pig islets after 6 years

Cotransplantation of porcine islets and Sertoli cells into preimplanted subcutaneous devices improve metabolic control in type 1 diabetic patients, and survive grafted for more than 4 years. We report here, further assessment of the endocrine and porcine nature of the surviving cells and the immune responses elicited toward Gal alpha(1,3)-Gal beta(1,4)-GlcNAc (Gal) antigen in patients who received a second and third transplants. No immunosuppressive drugs were administered. We were able to immunostain insulin- and glucagon-positive cells in all biopsies of patients and Sertoli cell markers in 60.9% of biopsies. Additionally, all biopsies tested, amplified the porcine COII gene. Patients demonstrated an increase in antipig antibodies in response to the first transplant with a decreasing response toward the second and third transplants. In all transplants, the IgG levels promptly returned to basal values after 3-4 months. The long-term survival of porcine cells and the reduced humoral immune response to multiple transplants indicate a form of tolerance. We have not been able to find CD25-positive cells, indicating that it is probably an immune accommodation of the graft.

The xenoantibody response and immunoglobulin gene expression profile of cynomolgus monkeys transplanted with hDAF-transgenic porcine hearts

BACKGROUND

Recent work has indicated a role for anti-Gal alpha 1-3Gal (Gal) and anti-non-Gal xenoantibodies in the primate humoral rejection response against human-decay accelerating factor (hDAF) transgenic pig organs. Our laboratory has shown that anti-porcine xenograft antibodies in humans and non-human primates are encoded by a small number of germline IgV(H) progenitors. In this study, we extended our analysis to identify the IgV(H) genes encoding xenoantibodies in immunosuppressed cynomolgus monkeys (Macaca fascicularis) transplanted with hDAF-transgenic pig organs.

METHODS

Three immunosuppressed monkeys underwent heterotopic heart transplantation with hDAF porcine heart xenografts. Two of three animals were given GAS914, a poly-L-lysine derivative shown to bind to anti-Gal xenoantibodies and neutralize them. One animal rejected its heart at post-operative day (POD) 39; a second animal rejected the transplanted heart at POD 78. The third monkey was euthanized on POD 36 but the heart was not rejected. Peripheral blood leukocytes (PBL) and serum were obtained from each animal before and at multiple time points after transplantation. We analyzed the immune response by enzyme-linked immunosorbent assay (ELISA) to confirm whether anti-Gal or anti-non-Gal xenoantibodies were induced after graft placement. Immunoglobulin heavy-chain gene (V(H)) cDNA libraries were then produced and screened. We generated soluble single-chain antibodies (scFv) to establish the binding specificity of the cloned immunoglobulin genes.

RESULTS

Despite immunosuppression, which included the use of the polymer GAS914, the two animals that rejected their hearts showed elevated levels of cytotoxic anti-pig red blood cell (RBC) antibodies and anti-pig aortic endothelial cell (PAEC) antibodies. The monkey that did not reject its graft showed a decline in serum anti-RBC, anti-PAEC, and anti-Gal xenoantibodies when compared with pre-transplant levels. A V(H)3 family gene with a high level of sequence similarity to an allele of V(H)3-11, designated V(H)3-11(cyno), was expressed at elevated levels in the monkey that was not given GAS914 and whose graft was not rejected until POD 78. IgM but not IgG xenoantibodies directed at N-acetyl lactosamine (a precursor of the Gal epitope) were also induced in this animal. We produced soluble scFv from this new gene to determine whether this antibody could bind to the Gal carbohydrate, and demonstrated that this protein was capable of blocking the binding of human serum xenoantibody to Gal oligosaccharide, as had previously been shown with human V(H)3-11 scFv.

CONCLUSIONS

DAF-transgenic organs transplanted into cynomolgus monkeys induce anti-Gal and anti-non-Gal xenoantibody responses mediated by both IgM and IgG xenoantibodies. Anti-non-Gal xenoantibodies are induced at high levels in animals treated with GAS914. Antibodies that bind to the Gal carbohydrate and to N-acetyl lactosamine are induced in the absence of GAS914 treatment. The animal whose heart remained beating for 78 days demonstrated increased usage of an antibody encoded by a germline progenitor that is structurally related, but distinct from IGHV311. This antibody binds to the Gal carbohydrate but does not induce the rapid rejection of the xenograft when expressed at high levels as early as day 8 post-transplantation.

Phenotypic and functional characterization of human T cell clones indirectly activated against adult pig islet cells

BACKGROUND

Xenotransplanted patients produce xenospecific IgG1 antibodies directed against epitopes other than Galalpha1,3Gal. IgG1 antibody production is believed to be dependent upon T cell help. Therefore, as a natural continuation of our work aimed at characterizing the xenoimmune antibody response against pig islet cells, we have also examined the T cell response. T cell reactivity against islet cells is believed to result from indirect antigen presentation, and our in vitro study was designed to mimic the response in vivo. The main purpose of this study was to characterize the phenotype, the immunological specificity and the functional capacity of indirectly activated T cell clones, reactive against pig islet cell antigens.

MATERIALS AND METHODS

Human T cell clones, activated against pig islet cells in the presence of autologous antigen-presenting cells, were produced from limiting dilutions of bulk cultures. Clonality was investigated by T cell receptor Vbeta (TcRVbeta) expression analysis. Clonal specificity was studied in proliferation assays using different pig cells as stimulators. ELISpot experiments were performed to detect cytokine production patterns. The cytotoxic capacity of the clones was assessed using standard cell-mediated lysis tests and different porcine and human target cells. Several long-term bulk cultures of human lymphocytes, indirectly activated against pig islet cells, maintained for up to 60 days, were used as a control for possible bias in the selection of the clones.

RESULTS

Nineteen CD4+ TcRValphabeta+ T cell clones were recovered. No activation of natural killer T cells or gammadelta-T cells was recorded. There was no bias in the TcRVbeta-usage. The immunological specificity differed between clones; some were specifically reactive against pig islet cell antigens, while others were reactive with antigens present on a variety of pig cells. All clones produced a broad spectrum of cytokines, e.g. interferon (IFN)-gamma, tumor necrosis factor (TNF)-alpha, interleukin (IL)5, IL10 and IL13, with no evidence of bias for a particular phenotype. None of the T cell clones were cytotoxic against pig islet cells, but two clones were cytotoxic against pig phytohemagglutinin (PHA)-blasts.

CONCLUSION

The analysis of several, indirectly activated, human CD4+ T cell clones shows that the response against pig islet cells is heterogeneous both with regard to immunological specificity and functional characteristics. This heterogeneity was further confirmed by analysis of the long-term bulk cultures of human lymphocytes, indirectly activated against pig islet cells.

Similarities in the immunoglobulin response and VH gene usage in rhesus monkeys and humans exposed to porcine hepatocytes

Background

The use of porcine cells and organs as a source of xenografts for human patients would vastly increase the donor pool; however, both humans and Old World primates vigorously reject pig tissues due to xenoantibodies that react with the polysaccharide galactose α (1,3) galactose (αGal) present on the surface of many porcine cells. We previously examined the xenoantibody response in patients exposed to porcine hepatocytes via treatment(s) with bioartficial liver devices (BALs), composed of porcine cells in a support matrix. We determined that xenoantibodies in BAL-treated patients are predominantly directed at porcine αGal carbohydrate epitopes, and are encoded by a small number of germline heavy chain variable region (V_H) immunoglobulin genes. The studies described in this manuscript were designed to identify whether the xenoantibody responses and the IgV_H genes encoding antibodies to porcine hepatocytes in non-human primates used as preclinical models are similar to those in humans. Adult non-immunosuppressed rhesus monkeys (Macaca mulatta) were injected intra-portally with porcine hepatocytes or heterotopically transplanted with a porcine liver lobe. Peripheral blood leukocytes and serum were obtained prior to and at multiple time points after exposure, and the immune response was characterized, using ELISA to evaluate the levels and specificities of circulating xenoantibodies, and the production of cDNA libraries to determine the genes used by B cells to encode those antibodies.

Results

Xenoantibodies produced following exposure to isolated hepatocytes and solid organ liver grafts were predominantly encoded by genes in the V_H3 family, with a minor contribution from the V_H4 family. Immunoglobulin heavy-chain gene (V_H) cDNA library screening and gene sequencing of IgM libraries identified the genes as most closely-related to the IGHV3-11 and IGHV4-59 germline progenitors. One of the genes most similar to IGHV3-11, V_H3-11^cyno, has not been previously identified, and encodes xenoantibodies at later time points post-transplant. Sequencing of IgG clones revealed increased usage of the monkey germline progenitor most similar to human IGHV3-11 and the onset of mutations.

Conclusion

The small number of IGV_H genes encoding xenoantibodies to porcine hepatocytes in non-human primates and humans is highly conserved. Rhesus monkeys are an appropriate preclinical model for testing novel reagents such as those developed using structure-based drug design to target and deplete antibodies to porcine xenografts.

Identification of the V genes encoding xenoantibodies in non-immunosuppressed rhesus monkeys

The major immunological barrier that prevents the use of wild-type pig xenografts as an alternative source of organs for human xenotransplantation is antibody-mediated rejection. In this study, we identify the immunoglobulin variable region heavy (IgV(H)) chain genes encoding xenoantibodies to porcine heart and fetal porcine islet xenografts in non-immunosuppressed rhesus monkeys. We sought to compare the IgV(H) genes encoding xenoantibodies to porcine islets and solid organ xenografts. The immunoglobulin M (IgM) and IgG xenoantibody response was analysed by enzyme-linked immunosorbent assay and cDNA libraries from peripheral blood lymphocytes were prepared and sequenced. The relative frequency of IgV(H) gene usage was established by colony filter hybridization. Induced xenoantibodies were encoded by the IGHV3-11 germline progenitor, the same germline gene that encodes xenoantibodies in humans mounting active xenoantibody responses. The immune response to pig xenografts presented as solid organs or isolated cells is mediated by identical IgV(H) genes in rhesus monkeys. These animals represent a clinically relevant model to identify the immunological basis of pig-to-human xenograft rejection.

Biological activity of pig islet-cell reactive IgG antibodies in xenotransplanted diabetic patients

BACKGROUND

The IgG antibody response in type I diabetic patients, transplanted with fetal pig islet-like cell-clusters, was investigated using purified immunoglobulin G (IgG) fractions from sera collected 7 to 9 yr after transplantation. From our earlier studies, we knew that the immunological specificities of xenoreactive IgG1 and IgG2 antibodies are different, and that IgG1 antibodies, in contrast to the IgG2 population, are mainly directed against non-Galalpha1,3Gal epitopes. In this study our goal was to establish whether xenoreactive IgG1 and IgG2 antibodies react with pig islet cells and, if so, to identify the target cell type, the biological function as well as the specificity of such antibodies for islet cell antigens. Sera from xenotransplanted patients were compared with those of patients with diabetes, selected for high titres of islet-cell specific autoantibodies.

METHODS

IgG antibody fractions from patient sera were purified on a protein G column. Surface expression of target antigens was studied using flow cytometry as well as immunofluorescence microscopy. The biological function of islet-cell reactive sera was tested using antibody dependent cellular cytotoxicity with both xenogeneic adult pig islet cells and allogeneic human islet cells as targets. Antibody specificity was assessed using 2D Western blots with both fetal and adult pig islet as well as human islet cell antigenic preparations.

RESULTS

Some of the diabetic patients, who have been transplanted with xenogeneic fetal pig islet cells, continue to produce xenospecific IgG1 and IgG2 antibodies for 7 to 9 yr post-transplantation. A separate analysis of IgG1 and IgG2 antibodies showed that IgG1 antibodies react with pig islet beta cells, whereas IgG2 antibodies mainly react with non-endocrine pig cells. Such antibodies are xenospecific, as they were found to mediate antibody-dependent cellular cytotoxicity of adult pig, but not human islet target cells. The reverse was true for antibodies from non-transplanted diabetic patients with high titres of autoantibodies against beta cells. Fluorescence analysis as well as 2D gel Western blots revealed that the reactivity was variable between patient samples, indicating that target antigens for non-Galalpha1,3Gal-specific antibodies are heterogeneous.

CONCLUSION

Thus, xenotransplantation of diabetic patients induces islet-beta cell reactive xenospecific IgG1 antibodies, which are biologically active and can mediate antibody-dependent cellular cytotoxicity of pig islet cells.

Evidence that tilapia islets do not express alpha-(1,3)gal: implications for islet xenotransplantation

BACKGROUND

Cell therapy for diabetes using teleost fish islet tissue has emerged as an intriguing alternative to the use of islet tissue from mammalian pancreases. The islet tissue, called Brockman bodies (BBs), is anatomically distinct from the pancreatic exocrine tissue and can be easily identified and isolated. Islets harvested from Nile tilapia (Oreochromis niloticus), when transplanted into streptozotocin-diabetic nude mice, produce long-term normoglycemia and achieve mammalian-like glucose tolerance profiles. We asked whether tilapia express the alpha-(1,3)gal epitope, the immunodominant target of human xenogeneic responses.

METHODS AND RESULTS

Immunostaining with the alpha-(1,3)gal-specific IB4 lectin on tilapia BB, liver, heart, spleen, and head kidney was negative, as was staining with murine anti-alpha-gal-specific monoclonal antibodies. Absence of alpha-gal-specific binding of IB4 or murine anti-gal mAbs to dispersed BBs was confirmed by fluorescent-activated cell sorter analysis. Tilapia BB cell membranes failed to reduce binding of anti-alpha-(1,3)gal-specific mAb in an enzyme-linked immunosorbent assay (ELISA) inhibition assay, while porcine and murine tissue lysates did. Tilapia BB cell lysates were shown to be devoid of alpha-1,3 galactosyltransferase activity by ELISA. Transplantation of tilapia BBs into diabetic alpha-gal knockout (gal KO) mice was not associated with accelerated xenograft rejection when compared with wild type control recipients (mean survival time 6.5 days vs. 7.2 days). Tilapia BBs failed to induce a rise in anti-gal IgG and IgM titers in gal KO mice, while the transplant of wild type mouse islets into gal KO mice caused a significant rise in anti-gal IgG and IgM antibodies.

CONCLUSIONS

We conclude that tilapia BBs are devoid of alpha-gal expression, and may offer an alternative to swine as a donor species for islet xenotransplantation.

Expression of CTLA-4 in nonhuman primate lymphocytes and its use as a potential target for specific immunotoxin-mediated apoptosis: results of in vitro studies

T-cell-mediated immunoregulation is one of the main mechanisms implicated in induction and maintenance of transplantation tolerance. In this regard, deletion or modulation of xeno/alloantigen-specific T cells, as well as blocking of their interactions with other cell populations, are currently being pursued for tolerance induction in humans as well as nonhuman primates. In order to investigate whether cytotoxic T-lymphocyte antigen-4 (CTLA-4) may represent a suitable target for a T cell depletion approach in nonhuman primate models, we analysed CTLA-4 expression in peripheral blood mononuclear cells (PBMCs) from nonhuman primates and the potential role of two anti-CTLA-4 saporin-conjugated immunotoxins. The analysis was performed in PBMCs from 8 cynomolgus monkeys from Philippines and from Mauritius both at protein level by flow cytometry and at transcriptional level by RT-PCR. In addition, the apoptotic role of the immunotoxins was investigated. The results showed that CTLA-4 was expressed at variable levels depending on the origin of the cynomolgus monkeys and the resting or activated cell condition. CTLA-4 was not expressed on resting Mauritius PBMCs and showed a lower up-regulation upon PMA/PHA activation compared to the Philippines PBMCs that expressed CTLA-4 also before activation. Two CTLA-4 RNA transcripts (672 and 550 bp) were detected with levels variations after cell stimulation. Two anti-CTLA-4 immunotoxins induced in vitro apoptosis of activated PBMCs from both sources of cynomolgus monkeys. This is the first report that documents CTLA-4 expression both at protein and transcriptional level by nonhuman primate PBMCs and provides novel perspectives of xeno/allograft rejection immunotherapy based on CTLA-4 targeting.

Porcine endothelium activated by anti-α-GAL antibody binding mediates increased human neutrophil adhesion under flow

BACKGROUND

Neutrophils participate in acute vascular rejection (AVR) of organ xenografts. Induced antibodies (Abs), including anti-Galalpha1,3Gal (alpha-Gal) Abs, have been suggested to cause AVR. We investigated the adhesion of naive human neutrophils to porcine aortic endothelial cells (PAECs) stimulated with anti-alpha-Gal Abs under conditions of flow. In addition, the ability of human neutrophils to adhere to human and porcine endothelium under static and flow conditions was evaluated.

METHODS AND RESULTS

In a flow-adhesion assay, a significant increase in adhesion of human neutrophils to PAECs, but not to human aortic endothelial cells (HAECs), was detected 6 hours after anti-alpha-Gal Ab-binding. After Ab stimulation, PAECs expressed CD62E and increased levels of CD106, indicating an activated endothelial cell (EC) phenotype. In a migration assay, supernatants from Ab-stimulated PAECs induced migration of human neutrophils, which was partially blocked by anti-porcine (p) interleukin (IL)-8 Abs and an antagonist to platelet-activating factor (PAF). In static and flow-adhesion assays, no difference in adhesion of human neutrophils to unstimulated or tumor necrosis factor (TNF)-alpha-stimulated HAECs and PAECs could be detected.

CONCLUSIONS

Our data suggest that anti-alpha-Gal Abs play an important role in the initiation of AVR by mediating adhesion and recruitment of neutrophils within an organ xenograft. In contrast with previous investigations, our data argues against a differential recognition of PAECs and HAECs by human neutrophils. Thus, to prevent AVR and accomplish long-term xenograft survival, it will be important to remove anti-alpha-Gal Abs before and after pig-to-human transplantation.