Incidence of hyperacute rejection in pig-to-primate transplantation using organs from hDAF-transgenic donors

Abstract

Background.  

Cynomolgus monkeys or baboons received under immunosuppression kidney or heart grafts from pigs transgenic for human decay-accelerating factor (hDAF) or from control pigs. Hyperacute rejection (HAR) is often difficult to differentiate from nonimmunological causes of organ or recipient dysfunction (NIC), and therefore, a thorough pathology review of all cases with 0–4 days survival (inclusive) was conducted.

Methods.  

Pathology slides were blinded and together with limited clinical data reviewed by two pathologists. After unblinding, data were compared with the original diagnosis made during the course of the program, and a final diagnosis was reached considering the complete clinical dataset.

Results.  

Life-supporting kidney transplantation was performed in 245 cynomolgus monkeys (234 hDAF, 11 controls), of which 102 cases had 0–4 day survival. None of the hDAF cases showed HAR, whereas this occurred in 27% of controls (P <10⁻⁶). Heterotopic heart transplantation was performed in 65 monkeys (57 hDAF, 8 controls), of which 41 cases had 0–4 day survival. HAR was observed in 7% of hDAF cases and in 57% of controls (P =0.002). Heterotopic heart transplantation in baboons was performed in 33 animals (28 hDAF, 5 controls), of which 15 cases had 0–4 day survival. HAR was observed in 11% of hDAF cases and in 20% of controls. Sixteen baboons were subjected to orthotopic heart transplantation, all from hDAF donors, out of which eight survived 0–4 days. The incidence of HAR was 6%.

Conclusions.  

In the largest series of pig-to-primate solid organ transplants performed thus far, the presence of the hDAF transgene fully prevents HAR of cynomolgus monkey kidney transplants and partially inhibits HAR of heart grafts in cynomolgus monkeys or baboons. The incidence of HAR in control grafts is significantly higher.

Based on their physiology and size porcine solid organs have been proposed as suitable organs for xenotransplantation. However, a major hurdle in achieving survival of such organs in large animal transplantation models including nonhuman primate recipients is antibody-mediated rejection. Like human beings, nonhuman primates have naturally existing anti-pig antibodies, which in combination with complement can cause the immediate destruction of the graft after reperfusion, so-called hyperacute rejection (HAR). Removal of these antibodies and/or inhibition of complement activation allows HAR to be managed, but subsequently, antibody-mediated and/or cellular rejection can occur. During the last decade, pigs transgenic for human regulators of complement activation have been developed, including human decay-accelerating factor (hDAF), membrane cofactor protein (CD46), and protectin (CD59). Organs from such animals are generally not hyperacutely rejected after transplantation in a nonhuman primate ^(1–8).

We have established a line of transgenic Large White/Landrace pigs transgenic for hDAF. A large transplantation program has been conducted using hearts or kidneys from these animals as donor organs and cynomolgus monkeys or baboons as recipients. In the majority of these studies induction immunosuppression was given using cyclophosphamide (Endoxana, ASTA Medica, Cambridge, UK; 2–4 intravenous or oral doses during the initial 6-day period), cyclosporine administered twice daily either orally or intramuscularly (Neoral, Novartis Pharma AG, Basel, Switzerland), and steroids administered either intramuscularly (methylprednisolone, Solu-Medrone, Pharmacia & Upjohn, Milton Keynes, UK) or orally (prednisolone, Prednesol, Glaxo Wellcome, Uxbridge, UK), with or without splenectomy at transplantation. Maintenance immunosuppression included oral cyclosporine, a tapering dose of steroids, and a third immunosuppressive agent. These agents included cyclophosphamide, mycophenolate mofetil (MMF, Cellcept, Roche AG, Basel, Switzerland), mycophenolate sodium (Novartis), leflunomide, and rapamycin or the rapamycin derivative RAD (Novartis). In a few studies, drugs or biologicals that specifically target antibody-mediated rejection were used, i.e., for complement inhibition, cobra venom factor or soluble complement receptor 1, and for natural antibody neutralization, soluble glycoconjugate. Results of orthotopic ⁽³⁾ and heterotopic ⁽⁴⁾ heart transplantation in baboons and kidney transplantation in cynomolgus monkeys ^(5,7) have been published previously. Noteworthy, the presence or absence of the transgene in hDAF and control donors was confirmed in each individual case by DNA slot blot analysis using a random-labeled ³²P hDAF cDNA probe and by immunohistochemistry using a monoclonal anti-hDAF antibody (BRIC 216, IBGRL Research Products, Elstree, UK) ⁽⁹⁾.

EXPERIENCE IN THE PIG-TO-NONHUMAN PRIMATE SOLID ORGAN TRANSPLANTATION PROGRAM

We learned that some basic principles in xenograft rejection do not unequivocally apply to pig-to-primate xenotransplantation models.

• First, a control nontransgenic porcine graft upon transplantation into a nonhuman primate is not invariably subject to hyperacute rejection, i.e., the immediate blackening upon reperfusion with instant destruction of the vasculature and parenchyma. Not only could transplants show a delayed occurrence of “hyperacute rejection” but also there was prolonged survival in a number of control grafts under immunosuppressive treatment, i.e., 5 of 11 control pig kidney grafts in cynomolgus monkeys survived with life-supporting function for up to 6, 7, 29, 30, and 30 days, respectively. For heterotopic heart transplantation in cynomolgus monkeys, one of seven cases survived with a beating graft until day 42 after transplantation, and for heterotopic heart transplantation into baboons, this was observed in three of five cases surviving with a beating heart until 5, 8, and 10 days after transplantation, respectively.
• Second, some transplants from hDAF-transgenic donors were subject to hyperacute rejection, as detailed below.
• Third, we learned that the differentiation between rejection and other causes of graft dysfunction in the immediate posttransplant period can be quite complicated. This in particular applies to kidney transplantation. Biopsies in the first days after transplantation could not be taken, and graft function could only be assessed by indirect parameters like urine production and serum creatinine values; these parameters do not provide any possibility to differentiate between rejection and other causes of graft dysfunction and are not predictive at all for organ function after possible recovery of, e.g., ischemia/reperfusion injury. It is well established that such nonimmunological causes of graft dysfunction are more likely to occur in pediatric transplant situations ^(10,11), a situation comparable with nonhuman primate xenotransplantation in small animals like cynomolgus monkeys. To match organ size between donor and recipient, organs from very young pigs were often used, and such organs are particularly vulnerable to damage and could show immediate posttransplant dysfunction with a histologic picture resembling HAR. A steady increase in serum creatinine during the first posttransplant days is therefore not predictive of organ damage due to rejection or delayed graft function, and in such cases it can be difficult in graft histopathology to differentiate between different causes of graft damage. To illustrate this point, infiltration by polymorphonuclear granulocytes alone does not differentiate between antibody-mediated rejection and ischemia/reperfusion injury. We therefore introduced immunohistochemistry for complement and immunoglobulin deposits, considered to be major components in antibody-mediated rejection, in the proper pathological assessment of a grafted organ (detailed below, Table 1).
• Finally, the introduction of a transgene into graft donors, like the hDAF transgene in our studies, opens the possibility of new histologic entities in immunologic fundamentals of xenograft rejection. In a few cases in which it was possible to obtain renal graft biopsies in the first hour after reperfusion, we were able to demonstrate focal damage ascribed to a possible antibody-mediated rejection (vascular damage associated with focal hemorrhage and infiltration by polymorphonuclear granulocytes), which is generally considered to be a nonreversible condition of graft destruction. However, in the case of a kidney from an hDAF-transgenic donor, this turned out to be reversible as demonstrated by normal histology in biopsies taken later after transplantation (G. Pino-Chavez, unpublished data).

Table 1: Diagnostic criteria in pig-to-primate organ transplantation

DEFINITION OF REJECTION TYPES

Based on our experience using organs from hDAF-transgenic donors, it became evident that histopathological evaluation alone of a xenograft cannot give an unequivocal diagnosis of graft dysfunction. We defined three basic rejection types based on graft histology and clinical parameters, which were designated as HAR, acute humoral xenograft rejection (AHXR), and acute cellular xenograft rejection (ACXR). Diagnostic criteria for these rejection types, which proved adequate in diagnosing rejection of hDAF-transgenic organs, are listed in Table 1. The histopathologic picture of HAR does not differ essentially from that of AHXR, but rather the difference between these two rejection types is based on the time after transplantation that rejection occurs, and observations whether the graft has been functioning during the first posttransplant day (i.e., urine production in case of a kidney xenograft). Related to the temporal difference, it cannot be completely ruled out that antibodies induced after transplantation play a role in AHXR, whereas HAR is solely associated with the activity of pre-existing natural antibodies. We grouped cases without any sign of rejection in one category: so-called nonimmunological cause of graft/recipient dysfunction (NIC). This category includes a variety of underlying causes like surgical/technical failures, drug toxicity, and early posttransplant infection. Because our program had its main focus on xenograft rejection and differentiation between various causes of NIC can be quite difficult, we did not aim for further differentiation within the NIC category.

DESIGN OF THE RETROSPECTIVE PATHOLOGY REVIEW

To ensure that a proper diagnosis was made in each individual case, we performed a retrospective pathology review of all transplants with survival between 0 and 4 days, inclusive. This time period was selected because grafts in our experience with minimal or no function, in particular kidney grafts, could have been left in the recipient before the animal was terminated (see reasoning outlined above). The review was performed according to the following protocol:

• After selection of all cases with a survival between 0 and 4 days, inclusive, all pathology slides were blinded, so that there was no bias with respect to previously known animal or histology code; slides selected were the paraffin section stained with hematoxylin and eosin and frozen tissue sections stained for immunoglobulin (Ig)G (clone A5711, Dako), IgM (clone AF6, Immunotech), complement C3 (clone HAV3-4, Dako), and complement C5b-9 (clone AE11, Dako).
• A limited set of clinical data was extracted, which included survival, a description of the anastomoses at necropsy, and for kidney grafts the presence or absence of urine output and terminal serum creatinine values. This data sheet did not include the transgenic status of the graft (i.e., whether originating from a hDAF-transgenic or control organ).
• Two pathologists independently reviewed the slides and clinical information and prepared a diagnosis. After collection of the data and processing by a third study member, cases with discrepant findings were re-evaluated in a meeting with both pathologists, in which a consensus diagnosis was reached. In several cases a consensus was reached after additional clinical information was given upon request by the pathologists. This review was completed within a 2-week study period.
• Subsequently, the cases were decoded and each diagnosis was compared with the original diagnosis made during the course of the program. In the discussion of these results, it became evident that discrepancies between original diagnosis and the diagnosis reached in the review could be ascribed to insufficient clinical data presented to the pathologists. Therefore, all cases with a discrepancy were re-evaluated by the transplant team, taking into consideration the full clinical dataset. In this team meeting, the final diagnosis was prepared.

This review was performed for all transplants performed in our institution between 1995 and 2000, which included 245 life-supporting kidney transplants in bilaterally nephrectomized cynomolgus monkeys (234 organs from hDAF-transgenic donors, 11 organs from control donors), 65 heterotopic heart transplants in cynomolgus monkeys (57 hDAF, 8 controls), 33 heterotopic heart transplants in baboons (28 hDAF, 5 controls), and 16 life-supporting orthotopic heart transplants in baboons (all from hDAF-transgenic donors). The review of cases with 4 days survival or less comprised 102 cynomolgus monkey kidney transplants, 41 cynomolgus monkey heart transplants, 15 baboon heterotopic heart transplants, and 8 baboon orthotopic heart transplants. In all cases, consensus was reached by the pathologists, including 51 kidney cases and 22 heart cases in which there was a discrepancy with the initial diagnosis (NIC, HAR, AHXR). After comparison of the pathologists’ diagnosis with the original diagnosis, the transplant team re-examined 42 kidney cases and 31 heart cases to reach consensus on a final diagnosis.

RESULTS

Original and Final Diagnosis

Data on the original diagnosis made during the course of the program and final diagnosis after review are presented in Table 2. ACXR is not listed in this table, because it was only observed in one kidney transplant accompanying AHXR.

Table 2: Original diagnosis made during the course of the program and final diagnosis after pathology review of cases with 0–4 days survival, inclusive^a

For the kidney transplants, there was agreement between the original and the final diagnosis in 84 cases (61 with NIC, 3 with HAR, 15 with AHXR, and 5 classified as not evaluable). For 11 cases the final diagnosis was NIC, whereas the original diagnosis was HAR, AHXR, or not evaluable. Five cases originally diagnosed as NIC received a final diagnosis of AHXR or not evaluable. One case originally diagnosed as AHXR was finally diagnosed as not evaluable, whereas the converse was observed for another case. Out of the 41 cases reviewed from the cynomolgus heart transplant program, the original diagnosis was confirmed by the final diagnosis in 33 cases. The final diagnosis of NIC was reached in three cases originally diagnosed as HAR or not evaluable, whereas in five cases the original diagnosis of NIC was revised to either HAR or AHXR. In 13 of the 15 cases from the baboon heterotopic heart program, the final diagnosis was the same as the original one; 1 case originally diagnosed as NIC received the final diagnosis of HAR, and 1 case originally diagnosed as HAR was finally diagnosed as AHXR. Seven of eight cases from the orthotopic heart transplant program had the same final diagnosis as the original diagnosis, and in one case the original diagnosis of AHXR was changed to NIC.

Incidence of NIC and HAR

Using the results of the final diagnoses, we evaluated the total transplantation program for its incidence of NIC and HAR. In this evaluation, animals with survival exceeding 4 days were included. In the vast majority of the cases with >4 day survival, the diagnosis made during the course of the program was either deterioration in clinical condition related to drug side effects, infection, or rejection (AHXR or ACXR). Hence, because the diagnosis in cases with >4 day survival was not taken into review, the diagnosis of NIC or AHXR in this evaluation regards only the 4-day time period included to the review presented above. All cases in which the final diagnosis was not evaluable were excluded from this analysis. The incidence of NIC was calculated for the combined groups of hDAF transplants and control transplants and that of HAR and AHXR separately for hDAF transplants and controls (data are shown in Table 3).

The total number of evaluable kidney transplants was 238, including 227 hDAF transplants and 11 controls. There were 72 cases with a final diagnosis of NIC, hence the incidence of NIC was 30%. Three out of 11 control transplants (27%) had the final diagnosis of HAR compared with 0 (0%) of 227 hDAF transplants. This difference was highly statistically significant (P <10⁻⁶, chi-square analysis). Twenty hDAF transplants had AHXR during the first 4 days posttransplantation, whereas this was not observed in any of the controls.

Similar observations were recorded for the series of heterotopic heart transplants in cynomolgus monkeys, which included 55 evaluable hDAF transplants and 7 evaluable controls. Twenty-nine cases were diagnosed as NIC during the first 4 days posttransplantation, i.e., an incidence of 47%. Four hDAF transplants had the final diagnosis of HAR (7%) compared with four of seven transplants (57%) in the control group; the difference was statistically significant (P =0.002). The incidence of AHXR in the two groups (one hDAF transplant, zero in the controls) was not significant.

The series of heterotopic heart transplants in baboons included 28 hDAF transplants and 5 controls. The incidence of NIC in this series was 21%. The incidence of HAR was lower in the hDAF group (three cases, 11%) than in the controls (one case, 20%). Orthotopic life-supporting heart transplantation was performed in 16 baboons, in all cases using hDAF organs. The incidence of NIC was 44%, HAR was 6%, and 50% of animals survived for more than 4 days. We previously reported on 10 of the 16 transplants in the series of orthotopic transplants ^(3,4) and 3 transplants in heterotopic heart transplants ⁽⁴⁾; for 1 heterotopic and 6 orthotopic transplants with survival between 0 and 4 days, the original diagnosis was confirmed in the present review.

DISCUSSION

This thorough review was initiated to best differentiate NIC from rejection. The differentiation between these diagnostic entities can be complex. This is illustrated by the fact that most changes from the original to the final diagnosis (Table 2) involved NIC either in the original or the final diagnosis. The incidence of NIC in the four models used was high, with 21% in heterotopic heart transplants in baboons, 30% in kidney transplants in cynomolgus monkeys, 47% in heterotopic heart transplants in cynomolgus monkeys, and 50% in orthotopic heart transplants in baboons. This high incidence is in part explained by the complex nature of the transplantation procedure, including aspects like the vulnerability to dysfunction due to the small (pediatric) size of the organ and in part by the lack of tolerance to the relatively aggressive immunosuppressive regimen. From the surgical/technical point of view, heterotopic heart transplantation in baboons is the easiest to perform, which explains the lowest incidence of NIC in this model (21%). The more than twofold higher incidence of NIC in heterotopic heart transplantation in cynomolgus monkeys is explained by the smaller size of heart grafts and recipients used in this model. Drug treatment side effects could be a contributing factor to the differences between the models in outcome of NIC. However, we do not consider this a major factor, because almost all animals received a similar aggressive induction treatment comprising cyclophosphamide, cyclosporine, and steroids. Because the dose levels were variable both within and between various study protocols and because individual animals responded differently to drug exposure, it is not possible to unequivocally relate the difference in NIC between models to drug side effects.

The pathology review was performed without knowledge of the transgenic status of the graft, and the relationship between graft outcome and hDAF status was therefore examined first after unblinding of the results. This analysis revealed some remarkable features. First, the hypothesis that kidneys or hearts from control pigs are invariably subject to HAR and that the expression of hDAF allows HAR to be managed, seems not to be as clear cut as first expected. The HAR rate for hDAF organs was 0% in kidney transplantation and between 6% and 11% in the various heart transplant models (see Table 3). This was statistically significantly lower than that of control transplants in both cynomolgus models. In this interpretation, it is noteworthy that the control groups were small compared with the groups of hDAF transplants, so that the estimated incidence of HAR is likely to be most accurate in the hDAF groups. Also, specific treatments to neutralize antibodies and/or complement were in general not given. We therefore conclude that the expression of hDAF completely prevents HAR of porcine kidney grafts in cynomolgus monkeys and partially inhibits HAR in porcine heart transplants in cynomolgus monkeys and baboons.

Table 3: Incidence of NIC and HAR

There is no unequivocal explanation for our observation that some of the hDAF heart grafts were subject to HAR, whereas this was not the case for any of the 227 evaluable transgenic kidney grafts. It should be noted in this respect that the transgenic status of the grafts was confirmed in each individual animal and graft. Also, in a separate analysis we were unable to relate the occurrence of HAR of either control or hDAF-transgenic organs to the level of pretransplant antibody as assessed by a hemolytic anti-pig antibody assay ⁽¹²⁾. Taking the data from this assay together for all cynomolgus monkeys and baboons, as well as for the transplantation models applied, the median values (ranges within brackets) for NIC, HAR, AHXR, and survival >4 days, respectively, were as follows: for control organs: 1700 (1590–1800), 1730 (1550–1810), 1440 (single case), and 930 (280–940), respectively; and for hDAF-transgenic organs: 1200 (190–2410), 1670 (1080–2980), 1440 (940–2210), and 1400 (160–2770), respectively (the value in a pooled normal human serum is 1000).

In conclusion, the present results of a thorough review of the largest series of pig-to-primate solid organ transplants conducted thus far confirms that kidneys from hDAF-transgenic pigs transplanted into cynomolgus monkeys subjected to conventional immunosuppression are not subject to HAR and that the incidence of HAR in hearts from hDAF pigs transplanted in either cynomolgus monkeys or baboons ranges between 6% and 11%. This incidence is significantly lower than that observed in non-hDAF controls. However, HAR is not invariably observed in control organs after transplantation into a nonhuman primate despite the fact that neutralization/inactivation of natural antibody or complement is not included in the immunosuppressive treatment.

Acknowledgments.

The authors thank Dr. D.K.C. Cooper (Transplantation Biology Research Center, Massachusetts General Hospital, Boston, MA) and Dr. K. Paradis (formerly Director of Clinical Research, Imutran Ltd) for valuable comments on the manuscript.