Revisiting liver transplant immunology: from the concept of immune engagement to the dualistic pathway paradigm.

Ever since the demonstration that allografts are rejected through immune reactions of the host, clinical therapies for organ allografts have relied on immune suppression to prevent these destructive events. A growing body of clinical and experimental data suggests that allografts elicit multiple, interactive immune responses. The result is not inevitably graft rejection, and "spontaneous" acceptance of fully allogeneic liver grafts occurs in rodents without immunosuppression. A spectrum of results range from spontaneous acceptance without immunosuppression to rejection with immunosuppression. The "dualistic pathway paradigm" aims to reconcile apparently conflicting observations in liver transplantation and proposes that: (1) immune engagement between the host and the allograft is instrumental in both rejection and acceptance; (2) there exist in all mammalian species congruent interactive pathways of immune activation whereby the fate of the allograft is determined by the quantitative results of these interactions; (3) the dualistic effect of immunosuppressive drugs on pathways of immune activation, conferring the capacity for favorable or unfavorable graft outcome should be investigated in experimental models in which organ allografts are spontaneously accepted. In conclusion the design of clinical strategies based on this research may contribute to protocols resulting in allograft acceptance without chronic immunosuppression.

Classical pathway complement destruction is not responsible for the loss of human erythrocytes during porcine liver perfusion.

BACKGROUND: Porcine livers perfused with human blood destroy 85% of human erythrocytes (red blood cells [RBC]) during prolonged extracorporeal perfusion, raising the possibility of a complement-mediated graft-versus-host effect. METHODS: Isolated porcine livers were perfused with fresh human blood. Plasma samples were analyzed for complement production by reverse CH50 analysis and porcine immunoglobulin class and specificity by enzyme-linked immunosorbent assay (ELISA) and flow cytometry. Anti-CD59 and anti-decay accelerating factor (DAF) monoclonal antibody were used to investigate whether human complement regulatory proteins inhibit porcine complement. RESULTS: After 64 hr of perfusion of porcine livers with human blood, mean complement activity in the perfusate was 95% of the starting value and increasing, whereas perfusion in the absence of a liver showed a falling complement activity of 28.7%. ELISA demonstrated porcine immunoglobulin (Ig) G and IgM in the xenoperfused human plasma. Whereas in a previous study flow cytometry demonstrated porcine antibodies specific for antigens on human T lymphocytes, in this study, anti-human RBC antibodies were not found. Xenoperfused human plasma did not lyse fresh human RBC. Human complement was consistently more efficient at lysing porcine RBC than was porcine complement at lysing human RBC, and human plasma inhibited the ability of porcine plasma to lyse human RBC, raising the possibility of cross-species complement regulation. Complement regulatory proteins on human RBC were blocked using mouse monoclonal anti-human CD59 and DAF. Blocking CD59, but not DAF, augmented lysis of human RBC by porcine complement. CONCLUSIONS: Human CD59 inhibits porcine complement. The production of porcine complement from xenoperfused porcine livers is unlikely to result in clinically significant injury mediated through the classical pathway of complement activation.

Steroid-free liver transplantation in children.

Although corticosteroids have been part of immunosuppressive regimens since the early days of transplantation, steroid avoidance could be beneficial. To test this hypothesis in paediatric liver transplantation, we compared liver-transplantation under steroid-free immunosuppression in 20 children, who received combined tacrolimus and basiliximab, with that under tacrolimus and steroids in 20 matched historical recipients as a historical control group. 12-month rejection-free survival was 75% in the tacrolimus-basiliximab group compared with 50% in the steroid group (p=0.05). Growth in the first year after transplantation was significantly better in the tacrolimus-basiliximab group than in the steroid group. Steroid avoidance was, therefore, not harmful to our patients, and combining tacrolimus with basiliximab as a steroid substitution seems a safe alternative to tacrolimus and steroid immunosuppression.

Maintenance triple immunosuppression with cyclosporin A, mycophenolate sodium and steroids allows prolonged survival of primate recipients of hDAF porcine renal xenografts.

To date, the best results in life-supporting pig-to-primate renal xenotransplantation have been obtained in recipients exposed to long-term immunosuppression with cyclophosphamide. As this agent is frequently associated with side-effects, we have explored the potential of a mycophenolate sodium-based maintenance immunosuppression in this model. Human decay-accelerating factor (hDAF) transgenic kidneys were transplanted into splenectomized and bilaterally nephrectomized cynomolgus monkeys immunosuppressed with mycophenolate sodium, cyclosporin A and steroids, and exposed to a brief induction course with cyclophosphamide (up to four doses). After transplantation, the primates were monitored daily for biochemical and haematological evaluations and for the measurements of haemolytic anti-pig antibodies (APA). A detailed histological analysis of each explanted graft was also performed. All the animals showed very poor initial graft function but survived for up to 51 days. In contrast to our previous studies in xenograft recipients on long-term immunosuppression with cyclophosphamide, minimal or no circulating xeno-directed antibodies, as measured by the evaluation of APA titres, were detected in this series although some degree of acute humoral rejection was observed in all the explanted grafts and was the primary cause of graft failure. Furthermore, in addition to areas of humorally mediated graft damage, we have observed for the first time areas with exclusive and prominent infiltration by CD2+ and CD8+ mononuclear cells presenting patterns compatible with tubulitis, glomerulitis and arteritis, which we have called acute cellular xenograft rejection (ACXR). In addition, CD68+ infiltrating macrophages and CD20+ B-cells were also present. This study demonstrates that a triple maintenance immunosuppression with mycophenolate sodium, cyclosporin A and steroids is a viable alternative to a cyclophosphamide-based immunosuppression to obtain prolonged survival of porcine organs transplanted into primates. However, a more stringent control of antibody forming cells remains essential to further extend the survival of xenografts in this model. In addition, the use of the immunosuppressive regimen reported here in the primate is associated with the occurrence of a new category of cell-mediated xenograft injury (ACXR) whose significance has yet to be clarified.

Potential implications of ABO blood group for vascular rejection in pig to human kidney xenotransplantation.

BACKGROUND: A substantial hurdle for successful xenotransplantation is to negate the effect of xenoreactive natural antibodies [mainly Galalpha1-3Galbeta1-4GlcNAc (alpha-Gal) specific] that cause hyperacute xenograft rejection. Galalpha1-3Gal molecules (alpha-Gal) have close structural homology with human ABO blood groups and therefore an individual's blood group might influence the formation of alpha-Gal specific antibodies. Genetic heterogeneity controlling alpha-Gal specific antibody formation could have important implications for future pig to human xenotransplantation clinical trials. We have investigated the relationship between ABO blood group and immunoglobulin M (IgM) and immunoglobulin G (IgG) alpha-Gal specific antibody titres in sera obtained from renal dialysis patients and healthy blood donors. METHODS: Serially diluted sera (n = 166) obtained from renal dialysis patients awaiting kidney transplantation (n = 116) and healthy blood donors (n = 50) were tested for IgM and IgG alpha-Gal antibodies using an enzyme-linked immunosorbent assay (ELISA) specific for alpha-Gal. The study cohort comprised 62, 48, 36 and 20 sera obtained from blood group O, A, B and AB individuals, respectively. Reciprocal alpha-Gal specific antibody titres were calculated from ELISA titration curves and stratified by individual blood group. RESULTS: No significant heterogeneity was found in IgM alpha-Gal specific antibody titres across ABO blood groups. In contrast, marked heterogeneity was observed in IgG alpha-Gal specific antibody titres when stratified by blood group. IgG alpha-Gal specific antibody titres were higher in sera obtained from blood group O renal dialysis patients [median titre 40, interquartile range (IQR) 14 to 72], compared with blood group A (median titre 18, IQR 7 to 54, P = 0.05), blood group B (median titre 6, IQR 0 to 15, P < 0.001) and blood group AB patients (median titre 3.5, IQR 0 to 16, P = 0.002). A similar correlation was found for IgG alpha-Gal specific antibody titres in sera obtained from healthy blood donors with median titres of 20 (IQR 12 to 34), 37 (10 to 91), 9 (0 to 20), and 5.5 (0 to 12) in blood groups O, A, B and AB individuals, respectively. There was a strong interrelationship between alpha-Gal specific antibody class and blood group, with both IgM and IgG alpha-Gal specific antibodies found in 84% of the blood group O sera, 73% of blood group A sera, 50% of blood group B sera and 40% of blood group AB sera (P < 0.001). In a subgroup of 39 renal dialysis patients, IgM and IgG alpha-Gal specific antibody titres were measured in two serum samples obtained at different time-points (median time interval 581 days, range 42 to 4414), and showed a high degree of stability (correlation coefficient 0.88 and 0.90 for IgM and IgG, respectively). CONCLUSION: IgG alpha-Gal specific antibody titres are significantly higher in the sera of blood group O and A renal dialysis patients and healthy individuals compared with blood groups B and AB. These data indicate that future clinical trials of pig to human xenotransplantation may be more problematic for non-blood group B patients who are likely to have high levels of IgG alpha-Gal specific antibodies that are associated with acute vascular rejection.

Comparison of allogeneic and xenogeneic in vitro T-cell proliferative responses in sensitized patients awaiting kidney transplantation.

Highly sensitized patients awaiting kidney transplantation may be potential candidates for future clinical trials using pig organ donors. Because of crossreactivity between human leucocyte antigens (HLA) and swine leucocyte antigens (SLA), such patients might have heightened T-cell responses to porcine xenoantigens. We determined whether lymphocytes from allo-sensitized patients displayed secondary (cyclosporine resistant) T-cell proliferative responses against porcine xenoantigens. Lymphocytes from six non-sensitized, seven sensitized [immunoglobulin G (IgG) panel reactive antibodies (PRA) 11 to 84%], 14 highly sensitized patients (IgG PRA > 84%) and 12 healthy individuals were tested [in the presence and absence of Cyclosporin A (CsA)] to determine their proliferative response to human (allogeneic) and to porcine (xenogeneic) stimulator cells. Lymphocytes from all study groups showed a strong proliferative response to allogeneic and xenogeneic stimulator cells with no significant difference between non-sensitized and sensitized individuals. Addition of CsA (100 and 500 ng/ml) inhibited (>90%) proliferation of lymphocytes from all non-sensitized patients to both allogeneic and xenogeneic stimulators. CsA was less effective at inhibiting proliferation of lymphocytes from sensitized patients and highly sensitized patients to allogeneic stimulators [29% (n=21) and 50% (n=42) respectively were resistant to CsA inhibition (100 ng/ml)]. In contrast, cyclosporine inhibited proliferation of lymphocytes from the majority of sensitized and highly sensitized patients to xenogeneic stimulator cells [14% (n=21) and 14% (n=42) respectively were resistant to CsA inhibition (100 ng/ml)]. HLA sensitized patients awaiting renal transplantation display cyclosporine resistant proliferative T-cell responses to allogeneic stimulators but proliferative responses to xenogeneic stimulators are more amenable to suppression by CsA. This finding suggests that humoral sensitisation to HLA antigens is not necessarily indicative of a heightened in vitro T-cell response to SLA antigens.

Removal of anti-Galalpha1,3Gal xenoantibodies with an injectable polymer.

Preformed and elicited Ab's against the Galalpha1,3Gal terminating carbohydrate chains (alphaGal Ab's) are the primary cause of hyperacute and acute vascular xenograft rejection in pig-to-primate transplantation. alphaGal Ab's are produced by long-lived Ab-producing cells that are not susceptible to pharmacological immunosuppression. We reasoned that antigen-specific elimination of alphaGal Ab's might be achieved in vivo by systemic administration of nonimmunogenic polyvalent alphaGal structures with high avidity for alphaGal Ab's. We devised GAS914, a soluble trisaccharide-polylysine conjugate of approximately 500 kDa that effectively competes for alphaGal binding by alphaGal IgM (IC(50), 43 nM) and IgG (IC(50), 28 nM) in vitro. Injections of GAS914 in cynomolgus monkeys, at the dose of 1 mg/kg, resulted in the immediate decrease of more than 90% of circulating alphaGal Ab's and serum anti-pig cytotoxicity. In baboons, repeated injections of GAS914 effectively reduced both circulating alphaGal Ab's and cytotoxicity over several months. Studies with [(14)C]GAS914 in rhesus monkeys and Gal(-/-) mice indicate that GAS914 binds to circulating alphaGal Ab's and that the complex is quickly metabolized by the liver and excreted by the kidney. Remarkably, posttreatment alphaGal Ab titers never exceeded pretreatment levels and no sensitization to either alphaGal or the polylysine backbone has been observed. Furthermore there was no apparent acute or chronic toxicity associated with GAS914 treatment in primates. We conclude that GAS914 may be used therapeutically for the specific removal of alphaGal Ab's.

Sensitisation to swine leukocyte antigens in patients with broadly reactive HLA specific antibodies.

We have previously shown that IgG HLA specific antibodies in the sera of highly sensitised patients awaiting renal transplantation can cross-react with swine leukocyte antigens (SLA). In this study we determined the frequency of patient serum IgG HLA specific antibody binding to a porcine lymphocyte panel and the likelihood of locating a cross-match negative pig donor for sensitised patients. Serum samples (n = 82) were obtained from 35 sensitised [current IgG panel reactive antibodies (PRA) > 10%] and seven nonsensitised patients awaiting renal transplantation at Addenbrooke's Hospital, Cambridge, UK. Fifty sera had IgG HLA specific PRA of 11-84%, 20 had IgG PRA of >84% and 12 had 0% PRA (negative controls). Sera were absorbed with porcine erythrocytes to remove xenoreactive natural antibodies and tested for cross-reactive IgG HLA specific antibody binding by flow cytometry against a panel of porcine lymphocytes obtained from 23 human decay accelerating factor (hDAF) transgenic pigs. A total of 1,884 cross-match combinations were tested and 369 (20%) gave a positive porcine lymphocyte cross-match. For sera from sensitised patients with IgG PRA (11-64%), only 6 of 805 (0.75%) cross-match tests were positive. In contrast, for sera from patients with high IgG PRA (>64%), 363 of 805 (45%) cross-match tests were positive (p < 0.0001). There was no difference in the frequency of positive cross-matches between patient sera with IgG PRA 65-84% and highly sensitised patient sera with IgG PRA 85-100% [156/345 (45%) vs. 207/460 (45%)]. This study demonstrates that only patient sera with broadly reactive IgG HLA specific PRA (>64%) cross-react with porcine lymphocytes. If future clinical trials of xenotransplantation are undertaken, it may be of value to select a cross-match-negative pig organ donor for such patients.

Porcine livers perfused with human blood mount a graft-versus-"host" reaction.

BACKGROUND: A major focus of xenotransplantation research is the interaction between human immune effector mechanisms and porcine tissues. We present evidence that a transplanted porcine organ might also mount a significant immune response to a human recipient. METHODS: Isolated porcine livers were perfused with fresh human blood. Plasma samples were analyzed for complement production by reverse CH50 analysis. ELISA was used to determine the amount and class of porcine immunoglobulin in human blood after xenoperfusion. Flow cytometry was used to determine the specificity and class of porcine immunoglobulin in human blood after xenoperfusion and to determine whether porcine immunoglobulins had bound to the human lymphocytes in the blood perfusing the porcine livers. Electron microscopy was used to evaluate the interaction of porcine Kupffer cells and human erythrocytes. RESULTS: Over the course of 72 hr of extracorporeal perfusion of porcine livers with human blood, the hematocrit fell progressively to as low as 2.5% of starting values, a phenomenon not seen in experiments using porcine blood. We have demonstrated both porcine complement and immunoglobulin in the human blood after xenoperfusion. The porcine antibodies in the human blood have specificity for human lymphocyte antigens. In fact, with increasing duration of perfusion, 40% of the xenoperfusions showed increasing titers of porcine antibodies with specificity for human lymphocyte antigens suggesting a response by primed porcine lymphocytes. However, the majority of erythrocytes are extracted directly by Kupffer cells in the liver. CONCLUSIONS: These data demonstrate the ability of porcine livers to generate both a humoral and cellular graft versus host response to human cells.

Anti-pig antibody levels in naïve baboons and cynomolgus monkeys.

Anti-pig antibodies (APA) were analysed in serum from 28 naïve wild-caught baboons (originating from Kenya) and 31 naïve captive-bred cynomolgus monkeys (13 from the Philippines and 18 from Mauritius), using a haemolytic assay with pig erythrocytes (APA), flow cytometry on the porcine lymphoma T-cell cell line L35, and enzyme linked immunosorbent assay (ELISA) using alpha-Gal type II and type VI antigen. This was extended in baboon samples by the evaluation in two laboratories (Imutran, Cambridge, UK and Immerge, Boston, USA), and by antibody absorption using either immobilized alpha-Gal type II or alpha-Gal type VI. Anti-porcine antibodies were demonstrated in all assays with substantial variability within and between the three non-human primate groups. Immunoglobulin (Ig)M antibody levels tended to be similar to or higher than those in a pooled normal human standard serum while IgG levels tended to be lower. Highest antibody levels were recorded in Mauritius cynomolgus monkeys. There were statistically significant correlations between assays for IgM or IgG class anti-Gal antibodies using either alpha-Gal type II or alpha-Gal type VI as antigen, both for different assays and two laboratories involved. Also, significant correlations were observed between the anti-Gal and L35 binding assays. Baboon sera before and after absorption to immobilized alpha-Gal type II or type VI were analysed for anti-Gal type VI or type II antibody: levels were almost undetectable indicating that most anti-Gal antibodies react to epitopes shared between alpha-Gal type II and type VI oligosaccharides. Finally, the relation between APA and outcome of porcine heart xenotransplantation in cynomolgus monkeys and baboons showed no apparent relation between pre-transplant APA levels and the occurrence of hyperacute rejection (HAR) when compared with non-immunological cause of organ/recipient dysfunction or acute humoral xenograft rejection during the first 4 days post-transplantation or survival exceeding 4 days post-transplantation.

Serum anti-pig antibodies as potential indicators of acute humoral xenograft rejection in pig-to-cynomolgus monkey kidney transplantation.

BACKGROUND: Hyperacute rejection of solid organ pig xenografts in nonhuman primates has been overcome by using donors transgenic for human complement regulatory proteins, but grafts are still susceptible to humoral (antibody-mediated) rejection. We investigated whether circulating xenoreactive antibodies are a useful indicator of this xenograft rejection. METHODS: Five assays were employed in a retrospective analysis on 20 selected cynomolgus monkey recipients of renal xenografts transgenic for human decay-accelerating factor, with survival between 4 and 60 days. The assays included hemolytic and hemagglutination assays and the measurement of immunoglobulin (Ig)G and IgM binding to porcine endothelial cells and leukocytes, and to the Gal alpha 1-3Gal trisaccharide (Gal) antigen. To assess non-Gal-directed antibodies, sera were absorbed with a Gal-coated resin. A predictive value was defined as an increase in antibody levels before a decline in graft failure (>20% increase in creatinine levels) and humoral rejection in graft pathology. RESULTS: Data on hemolytic anti-pig antibody correlated with those on IgM antibody to endothelial cells, leukocytes, and Gal. In absorbed sera IgM and IgG antibody to endothelial cells and leukocytes correlated with each other, indicative for an elicited antibody response to non-Gal antigens. Sixteen animals showed humoral rejection, and in all but two animals one or more assays was considered of predictive value. On the other hand, increased antibody levels were noted in two animals without signs of rejection in graft pathology and in two cases with cellular xenograft rejection. CONCLUSIONS: It is recommended to use multiple assays (preferably hemolytic, anti-Gal, and anti-endothelial cell) to be able to fully monitor the peripheral antibody responses in pig-to-primate xenograft recipients.