Impact of Thrombocytopenia on Survival of Baboons with Genetically Modified Pig Liver Transplants: Clinical Relevance

Authors


* Corresponding author: David K. Cooper, cooperdk@upmc.edu

Abstract

A lack of deceased human donor livers leads to a significant mortality in patients with acute-on-chronic or acute (fulminant) liver failure or with primary nonfunction of an allograft. Genetically engineered pigs could provide livers that might bridge the patient to allotransplantation. Orthotopic liver transplantation in baboons using livers from α1,3-galactosyltransferase gene-knockout (GTKO) pigs (n = 2) or from GTKO pigs transgenic for CD46 (n = 8) were carried out with a clinically acceptable immunosuppressive regimen. Six of 10 baboons survived for 4–7 days. In all cases, liver function was adequate, as evidenced by tests of detoxification, protein synthesis, complement activity and coagulation parameters. The major problem that prevented more prolonged survival beyond 7 days was a profound thrombocytopenia that developed within 1 h after reperfusion, ultimately resulting in spontaneous hemorrhage at various sites. We postulate that this is associated with the expression of tissue factor on platelets after contact with pig endothelium, resulting in platelet and platelet-peripheral blood mononuclear cell(s) aggregation and deposition of aggregates in the liver graft, though we were unable to confirm this conclusively. If this problem can be resolved, we would anticipate that a pig liver could provide a period during which a patient in liver failure could be successfully bridged to allotransplantation.

Abbreviations 
EC

endothelial cells

GTKO

α1,3-galactosyltransferase gene-knockout

MLR

mixed lymphocyte reaction

PBMC

peripheral blood mononuclear cell(s)

TF

tissue factor

Tx

transplantation

WT

wild-type

Introduction

Despite the increasing use of livers from living and extended criteria deceased donors, the demand for liver transplantation (Tx) continues to grow (1). The lack of available livers is more important in patients with acute-on-chronic or acute liver failure or in the presence of primary allograft nonfunction, when clinical deterioration may develop rapidly (2–6).

A readily available animal source of organs and tissues for clinical Tx (cross-species Tx or xenoTx) could provide livers in an emergency. Pigs are currently considered to be the preferred source animal species for clinical xenoTx (7,8). If pig organs could be transplanted successfully into human patients, the advantages would be numerous. In particular, they would be available at any time, and might therefore at least provide a ‘bridge’ to alloTx.

There are few reports of xenogeneic liver Tx in large animal models (reviewed in Hara et al. [9]), particularly in the relevant pig-to-nonhuman primate model (10–12), but a further exploration in this model would appear timely. Recent advances in the genetic engineering of pigs have greatly increased the protection of pig vascular endothelial cells (EC) against the primate antibody-mediated immune response (13). In this respect, deletion of the gene for α1,3-galactosyltransferase (GTKO pigs) (14) has provided pigs that do not express the Galα1,3Gal oligosaccharide epitopes that are the major antigenic target for primate anti-pig antibodies (15,16). The transgenic expression of a human complement-regulatory protein, such as CD55 (17) or CD46 (18), provides further protection to pigs cells against the primate immune response (13). These advances have significantly prolonged pig heart (19,20) and kidney (21,22) graft survival in immunosuppressed nonhuman primates, and suggest that similar results could be obtained following pig liver xenoTx.

We here report liver xenoTx in baboons using genetically engineered pigs as the source of livers and a clinically applicable immunosuppressive regimen. The aim of the study was to explore pig livers as a means of ‘bridging’ patients for a period of a few days until a suitable allograft became available.

Materials and Methods

Animals

Baboons (Papio anubis, University of Oklahoma Health Sciences Center, Oklahoma City, OK), weighing 7–12 kg of either gender and of known ABO blood type, were recipients of one allograft and xenografts (Table 1). Wild-type (genetically unmodified, WT) Landrace/large white pigs (Country View Farm, Schellsburg, PA, USA) were recipients of allografts and the source of one xenograft (Table 1). GTKO (14) or GTKO pigs transgenic for CD46 (18) (GTKO/CD46) (Revivicor, Blacksburg, VA, USA) were sources of xenografts (Table 1). All pigs were of blood type O (nonA), weighed 3–12 kg, and were of either gender.

Table 1.  Recipient and donor information, graft and recipient survival, and causes of graft failure or recipient death
RecipientDonorDonor pig typeDonor weight (kg)Recipient weight (kg)Donor blood typeRecipient blood typeSpxGraft survival (days)Complications/Outcome
  1. aElectively euthanized according to IACUC protocol.

  2. B, baboon; GTKO, α1,3-galactosyltransferase gene-knockout; GTKO/CD46, α1,3-galactosyltransferase gene-knockout/membrane cofactor protein; hetero, heterozygous; homo, homozygous; IS, immunosuppressive therapy; P, pig; PNF, primary nonfunction; OR, operating room; Spx, splenectomy; WT, wild-type (unmodified).

ALLOTRANSPLANTATION
 P40307P40107WT14.1 15.4OOno>3Electively euthanizeda No IS
 P40207P40407WT10.6 13.6OOno>2Electively euthanizeda No IS
 B3408B16407-7.411.3ABABno>30Electively euthanizeda
XENOTRANSPLANTATION
 B16907P3008WT12.5 12OBno<1Electively euthanizeda No IS
 B3308P8308GTKO11.6 10.1OBno<1Size-mismatch, euthanized in OR
 B3508P14408GTKO/CD46 hetero6.5 7.2OAByes<1Size-mismatch, euthanized in OR
 B3108P8208GTKO6.110OABno6Sudden death due to severe bleeding in abdomen and lungs
 B3208P14508GTKO/CD46 hetero6.710.1OAyes4Very high tacrolimus levels. Died from sepsis
 B7708P21708GTKO/CD46 homo3.4 9.4OAByes7Severe bleeding in abdomen and lungs. Euthanized
 B7808P22108GTKO/CD46 hetero6.710.1OAByes6Severe bleeding in abdomen. Euthanized
 B7908P26708GTKO/CD46 hetero3.3 8.6OByes<1Donor treated with clodronate liposomes. Euthanized due to PNF of the liver
 B8108P26608GTKO/CD46 homo3.9 9.2OByes1Donor treated with clodronate liposomes. Euthanized due to PNF of the liver
 B18508P3909GTKO/CD46 hetero6.011.5OByes5Severe bleeding in abdomen and lungs. Euthanized
 B18908P4009GTKO/CD46 hetero5.810.0OByes6Sudden death due to severe bleeding in abdomen, small intestine, pericardial cavity, and myocardium

All animal care was in accordance with the Principles of Laboratory Animal Care formulated by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH publication No. 86-23, revised 1985). Protocols were approved by the University of Pittsburgh Institutional Animal Care and Use Committee (IACUC# 0706493).

Surgical procedures

Standard techniques were used. (See subsection ‘Surgical Procedures’ in Supporting Information.) Donor liver ischemia time was <2 h.

Immunosuppressive and supportive therapy

Pig allotransplantation (n = 2):  No immunosuppressive therapy was given. Analgesia and cefazolin prophylaxis were administered (Table 2).

Table 2.  Induction and maintenance immunosuppressive therapy and supportive therapy
TherapyDoseDuration
  1. Immunosuppressive therapy consisted of induction with thymoglobulin (Genzyme, Cambridge, MA) and maintenance with tacrolimus (Astellas Pharma US, Deerfield, IL), mycophenolate mofetil (MMF; Roche, Basel, Switzerland) and methylprednisolone. In 3 baboons (B7908, B8108 and B18908) cyclophosphamide (Baxter, Deerfield, IL) replaced thymoglobulin. All baboons received prophylactic intravenous antibiotic and ganciclovir therapy throughout the post-Tx period of follow-up.

Induction therapy
 Antithymocyte globulin5–10 mg/kg i.v.Day –3 and –1 (if T cell count >500 cells/mm3)
 Cyclophosphamide40 mg/kg i.v.Day –2 (only in B7908, B8108, and B18908)
20 mg/kg i.v.Day –1 (only in B7908, B8108, and B18908)
 Cobra venom factor1–3 mg i.v.Days –1 to 1 (only in B18508)
Maintenance therapy
 Tacrolimus0.05–0.1 mg/kg × 2/day i.m.From day –3 (to maintain 12 h blood trough levels of 10–15 ng/mL)
 Mycophenolate mofetil110 mg/kg/day i.v. continuous infusionFrom day –4 (to maintain blood levels of 3–5 μg/mL)
 Methylprednisolone10 mg/kg i.v.Tapering from day 0
Supportive therapy
 Famotidine0.25 mg/kg × 2/day i.v.From day –7
 Cefazolin25 mg/kg × 2 day i.v.From day –7
 Heparin100 U/kg i.v.Day 0 (only in B18508 and B18908)
 Ganciclovir5 mg/kg/day i.v.From day –3
 Dopamine2–3 μg/kg/min i.v.Days 0–7
 Prostacycline20 ng/kg/min i.v.Days 0–7
 Imipenem500 mg × 2/day i.v.From day 0
 Amikacin150 mg /day i.v.From day 0
 Ketorolac0.5 mg/kg/day i.v.Days –1 to 1 (in B7708 and B7808)
Days –1 to 6 (in B18508 and B18908)
 Vitamin K5 mg/day i.v.From day 0 on alternate days (only in B18508 and B18908)
Treatment of donor pig
 Clodronate liposomes90 mg/kg i.v.Day –2 (only in donor pig of B7908)
Day –2 and –1 (only in donor pig of B8108)

Baboon allotransplantation and xenotransplantation:  A clinically applicable immunosuppressive regimen (Table 2) was administered to all baboons except the one that received a WT pig liver (B16907) (Table 1). Some received ketorolac and/or heparin in an effort to prevent thrombocytopenia (Tables 1 and 2). In one case (B18508), cobra venom factor was administered for 3 days in order to deplete complement activity (Table 2). Clodronate liposomes (Cl2MDP) (generously provided by Dr. Robin Pierson of the University of Maryland) were used with the aim of depleting macrophages (including Kupffer cells) in two donor pigs on days −2 ± day-1 (Table 2).

Monitoring of recipient pigs and baboons

Blood cell counts, chemistry (including tests of hepatic and renal function), and coagulation parameters were measured daily by standard methods (University of Pittsburgh Medical Center Central Laboratory, Presbyterian Hospital, Pittsburgh, PA). Tacrolimus levels were measured daily and MMF levels at least ×3/week. Blood cultures were performed at least weekly and whenever infection was suspected clinically. Methods of immunological monitoring have been described previously by our group (13). Flow cytometry to monitor T and B cell numbers was performed pre-Tx (to determine the dose of thymoglobulin) and post-Tx. Anti-non-Gal antibodies were monitored by flow cytometry (13,23). Serum cytotoxicity was determined using GTKO and GTKO/CD46 pig peripheral blood mononuclear cells (PBMC) as target cells (24). The mixed lymphocyte reaction (MLR) was carried out pre-Tx (before any immunosuppression was started) in all baboons and post-Tx in one case (25). Total complement activity was measured by the CH50 test (26).

Histopathology and immunohistopathology of tissues

Biopsies were obtained from the donor liver before excision, approximately 2 h after reperfusion before abdominal closure, and at the time of euthanasia or death. Tissues were fixed in 10% formalin and embedded in paraffin. Four micron (4 μm) sections were stained with hematoxylin and eosin for light microscopy. At necropsy, all major recipient organs were examined and biopsies taken. Immunohistochemical staining for IgM, IgG and C3 was performed (27,28).

For immunofluorescence, liver biopsies (taken at the same time intervals as aforementioned) were stored at –80°C until processed, as previously described using FITC-conjugated antibodies to IgG (Dako-F0202, Dako, Denmark), IgM (Dako-0203, Dako) and C3 (Dako-F0201, Dako) (29). (Further details are in subsection ‘Histopathology and Immunohistopathology of Tissues’ of Supporting Information.)

Experimental groups

A total of 14 liver transplants were performed (Table 1), of which 3 were allotransplants (Group 1; pig n = 2, baboon n = 1) and 11 were xenotransplants (Group 2; WT-pig-to-baboon n = 1, GTKO or GTKO/CD46 pig-to-baboon n = 10). The major aim of alloTx was to develop the surgical technique in pigs and baboons and, in the baboon, to determine that the immunosuppressive regimen was successful in preventing rejection of an allograft.

Results

Graft and recipient survival and outcome

Group 1: Allotransplantation:  Two nonimmunosuppressed WT pigs with allografts recovered well from the operative procedure and were electively euthanized 2 and 3 days after liver Tx. The response to ischemia/reperfusion was different from that seen after human liver Tx in that only AST (but not ALT) increased temporarily (not shown). Because no immunosuppressive therapy had been administered, both livers showed mild mononuclear portal and perivenular inflammation with mild centrilobular congestion, representing the earliest phases of acute cellular rejection (not shown).

A single baboon with an allograft remained well and was electively euthanized on post-Tx day 30. The immunosuppressive regimen suppressed the CD3+, CD4+, CD8+ and CD20+ lymphocyte counts throughout the 30-day period (not shown). The response to ischemia/reperfusion was similar to that seen after human liver Tx in that AST and ALT increased temporarily, but had returned to normal values within 3 days (not shown). Liver function then remained excellent, indicating the efficacy of the immunosuppressive regimen. Microscopy of liver biopsies taken at 2 h (before closure of the abdomen), 23 days (by percutaneous needle biopsy) and at necropsy (on day 30) indicated no features of rejection (not shown).

Group 2: Xenotransplantation:  A single WT pig-to-baboon transplant (with no immunosuppressive therapy) was followed for 5 h, at which time the baboon was euthanized. Blood was drawn at intervals and liver biopsies were taken hourly. Liver function showed an increase in AST, but not in ALT, ALP, GGT, total and direct bilirubin. Total protein and albumin levels fell gradually and significantly within 5 h (total protein 5.8 vs. 2.0 g/dL, albumin 3.2 vs. 0.9 g/dL pre and post-Tx, respectively. Changes in platelet count were as in the immunosuppressed baboons (see further). Biopsies taken at hourly intervals indicated the steady development of hyperacute rejection, with patches of interstitial hemorrhage and edema, a polymorphonuclear cellular infiltrate and intravascular thrombi (not shown).

Following genetically modified pig-to-baboon xenoTx (n = 10), 4 recipients survived <1 day; the causes of termination of the experiment are listed in Table 1. The major problem in two cases (B3308, B3508) was size-mismatch between donor liver and recipient abdomen, resulting in our inability to close the abdomen without compressing the transplanted liver and causing hemodynamic compromise (Table 3). From this experience, we determined that a liver from a pig of the same weight as the recipient baboon was significantly larger (up to almost 60% larger) than the baboon liver.

Table 3.  Baboon and pig liver size-matching
Recipient baboonBaboon weight (kg)Baboon liver weight (g)Baboon liver weight as% of baboon weightPigPig weight (kg)Pig liver weight (blood-free) (g)Pig liver weight as% of pig weightPig liver weight as% of baboon weightBaboon: Pig weight ratiodBaboon: pig liver weight ratioeGraft survival (days)bPig liver weight at necropsy (g)
  1. aEuthanized because of size-mismatch, bmean graft survival, excluding those that survived <1 day, cpig liver weight at necropsy was significantly greater than pre-Tx (p < 0.05). As the pig livers were weighed immediately after excision from the donor when blood-free (following perfusion with UW solution), but the native baboon livers were weighed after hepatectomy (and were not blood-free), this would indicate that a pig liver may be as much as double the size of that from a comparable-weight baboon. We subsequently used smaller pigs as donors (e.g. P21708 = 3.4 kg vs. B7708 = 9.4 kg), dBaboon weight is considered 1, eBaboon liver weight is considered 1.

B330810.12282.26%P830811.6 4423.81%4.38%a1:1.14a1:1.94a<1496
B3508 7.21482.06%P144086.52423.72%3.36%a1:0.90a1:1.64a<1266
B310810.02122.12%P82086.12083.41%2.08%1:0.611:0.986326
B320810.12001.98%P145086.72483.70%2.45%1:0.661:1.244329
B7708 9.42482.64%P217083.41384.06%1.47%1:0.361:0.567236
B780810.11941.92%P221086.72533.78%2.94%1:0.661:1.306371
B7908 8.61641.91%P267083.31003.03%1.16%1:0.381:0.61<1146
B8108 9.22582.80%P266083.91634.18%1.77%1:0.421:0.631193
B1850811.52502.17%P39096.01913.18%1.66%1:0.521:0.765308
B1890810.02542.54%P40095.81813.12%1.81%1:0.581:0.716354
Mean9.3 ± 1.1216 ± 392.31% 6.0 ± 2.4217 ± 93c3.61%2.31%  5.7 ± 0.9b303 ± 94c

In two other cases (B7908, B8108), in an effort to deplete macrophages, including Kupffer cells, we had treated the donor pig with clodronate liposomes on the day before liver Tx (Table 2). Although the pigs remained healthy with normal liver function, after Tx both baboons experienced primary hepatic nonfunction, necessitating euthanasia. The sequence of events suggested that clodronate, when combined with an ischemic period during the Tx procedure, was toxic to the liver.

Six baboons survived for 4, 5, 6, 6, 6 and 7 days, respectively (Table 1). All 6 recovered well after Tx. In one (B3208), very high levels of tacrolimus (>50 ng/mL) were associated with the development of Gram-negative sepsis (Pseudomonas orzyhabitans) and death on day 4. In subsequent cases, we discontinued tacrolimus for at least 24 h after liver Tx until there was evidence of good hepatic function. The other 5 baboons remained clinically well but required transfusion of baboon washed red blood cells to maintain hematoctrit. They were euthanized or died from spontaneous bleeding into the peritoneal, thoracic and/or pericardial cavities, small intestine, lungs and/or myocardium, which resulted from profound thrombocytopenia (see further).

Minor modifications made to the immunosuppressive regimen, for example replacement of ATG by cyclophosphamide, or the addition of cobra venom factor, did not appear to influence outcome.

Rapid development of thrombocytopenia

All 11 baboons that received xenografts experienced the rapid development of severe thrombocytopenia, with platelet counts <50,000/mm3 within 5 h after liver reperfusion, except in one case where this level was not reached until 18 h (Figure 1A). Studies in 2 baboons indicated that the reduction in platelet count occurred within the first hour (Figure 1B). Splenectomy performed in the recipient before reperfusion of the liver graft did not influence outcome. Thrombocytopenia occurred rapidly even in the 2 baboons that received livers from pigs treated with clodronate liposomes, suggesting, though not proving, that this was not related to phagocytosis of platelets by Kupffer cells. When intraoperative heparin (100 units/kg) was administered (B18508, B18908), thrombocytopenia developed less rapidly, with platelet counts of 90–100,000/mm3 on postoperative day 1 (B18508, B18908), but this level was not maintained.

Figure 1.

Platelet counts in baboons after pig liver transplantation. (A) Throughout the postoperative period (n = 11). (B) During the first 5 h after reperfusion (n = 2). In several cases, the platelet count fell to <10 000/mm3 within the first 5 h.

As coagulation parameters remained within the near-normal range in all 6 baboons that survived for 4–7 days (see further), it seemed clear that it was the thrombocytopenia that was the major cause of the spontaneous internal hemorrhage. Baboon platelets were not available to us for transfusion and the transfusion of uncrossmatched human platelets (B7708 and B7808) did not increase the platelet count even temporarily. The transfusion of fresh whole baboon blood resulted in only transient increases in platelet counts (not shown). Sustained thrombocytopenia was followed by sudden and repeated falls in hematocrit (not shown), most likely from spontaneous internal bleeding that began in some baboons within 24 h of the development of thrombocytopenia and that responded to the transfusion of baboon washed red blood cells or fresh whole blood.

Liver function and coagulation in baboons that survived 4–7 days

In baboons that survived 4–7 days, pig liver function and coagulation parameters remained within the normal or near-normal range. These data will be reported fully elsewhere. In summary, liver transaminases (AST, ALT) increased temporarily from ischemia/reperfusion injury (Figure 2A). Increases in AST, GGT and ALP were observed in some baboons terminally. Increases in total and direct bilirubin suggested an intrahepatic cholestatic injury or abnormally viscous bile (Figure 2B). (At necropsy, the bile duct anastomoses were entirely patent, but viscous bile was present in the hepatic and bile ducts.) Total protein and albumin levels fell within a few hours to levels that are normal for pigs, but could be maintained at the levels normal for baboons by the continuous i.v. infusion of human albumin.

Figure 2.

Figure 2.

Four selected parameters of hepatic function in baboons after pig liver transplantation (n = 11). (A) AST, (B) Direct bilirubin, (C) INR, (D) Fibrinogen.

Figure 2.

Figure 2.

Four selected parameters of hepatic function in baboons after pig liver transplantation (n = 11). (A) AST, (B) Direct bilirubin, (C) INR, (D) Fibrinogen.

Complement activity remained high, suggesting sustained activation. Since the half-life of complement fractions is <24 h, the complement activity post-Tx was likely to have been through activation of complement produced by the pig liver. When cobra venom factor was administered to one baboon (B18508) on days –1, 0 and 1, CH50 remained between 11 and 13% for 4 days but this did not influence the outcome.

Despite apparently adequate production of coagulation factors (not shown), resulting in normal or near-normal coagulation parameters (Figure 2C, D) (though the INR was on occasions >1.2 and rose to 2.0 in 2 baboons, and fibrinogen fell after repeated internal hemorrhage in most baboons), 5 of the 6 baboons died or were euthanized following major spontaneous internal hemorrhage (associated with the severe thrombocytopenia). Our observations were that deterioration in coagulation parameters was secondary to the development of thrombocytopenia.

Immunologic monitoring in baboons that survived 4–7 days

White blood cell counts:  After the administration of thymoglobulin (days –3 and –1) or of cyclophosphamide (days –1 and 0), the total lymphocyte, T and B cell counts fell markedly and remained low in the majority of baboons until the termination of the experiments; the reduction in these counts was comparable to that seen in the baboon with a liver allograft (not shown). The white blood cell and monocyte counts generally remained normal or below normal, although there was some fluctuation in the monocyte counts (not shown).

Anti-nonGalα1,3Gal antibodies:  These were measured in 4 of the 6 baboons (B7708, B7808, B18508, B18908). There was a slight reduction in anti-nonGal IgM on day 1 (possibly associated with some binding of IgM to the transplanted pig organ), but no further change (not shown). IgG levels were very low throughout the experiments. There was no increase in IgM or IgG that may have indicated T cell-dependent sensitization to the pig graft (although this is unlikely to have developed within the 7 days of follow-up).

Serum complement-dependent cytotoxicity assay:  The serum complement-dependent cytotoxicity assay was performed in 4 baboons (B7708, B7808, B18508, B18908) at 2 time-points (pre-Tx and day of euthanasia/death). Pre-Tx lysis of PBMCs from GTKO pigs was 8–33%; this lysis decreased to 0–5% after Tx. When PBMCs from GTKO/CD46 pigs were used as target cells, there was 0% lysis pre and post-Tx (not shown) (likely due to constitutive expression of CD46).

Mixed lymphocyte reaction (MLR):  Although an MLR was performed in all baboons pre-Tx, and showed a brisk response to all stimulator cells (allo, WT, GTKO and GTKO/CD46 pig PBMCs), it was carried out post-Tx in only one case, in which almost complete suppression of the response to all stimulator cells was demonstrated.

Macroscopic findings at necropsy

After reperfusion, the liver showed a transient slightly mottled appearance in a minority of cases, though this was indistinguishable from that seen in allografts. At necropsy in the baboons that had survived for 4–7 days, all 6 showed skin petechial hemorrhages. There was free heavily blood-stained fluid or blood in the peritoneal cavity in all cases, the gastrointestinal tract (predominantly the small intestine) had blood within the lumen, (n = 3), the lungs showed blood in the alveolae and bronchi (n = 4) and there was a heavily blood-stained pericardial effusion and myocardial petechiae in one case, indicating spontaneous hemorrhage at these sites. The livers showed random patchy dark areas, suggesting injury, interspersed with extensive areas where the appearances were relatively normal (Figure 3A).

Figure 3.

Figure 3.

Macroscopic and microscopic assessment of pig liver xenografts. (A) Macroscopic appearance of the pig liver in B18908 at necropsy (day 6). This liver showed fewer dark areas than most other livers. (B) Histology of the pig liver in B7808 at necropsy (day 6) (H&E) from a macroscopically injured (dark) area showing areas of hemorrhagic necrosis interspersed within conserved liver structure (×100 magnification). (C) Same area at ×200. (D) From a macroscopically normal area, showing normal liver structure (×100). (E) Same area at ×200. (F) Confocal microscopy of a GTKO/CD46 pig liver (in B7808) at necropsy (day 6) after immunofluorescence staining for platelets (CD41; red), PBMCs (CD45; green) and nuclei (DAPI; blue) showing platelet and platelet-PBMC aggregates in the liver sinusoids (×1000). Immunofluorescence studies showed platelet and/or platelet-PBMC aggregates in the blood and within the liver grafts. In most cases, deposition of IgM, IgG and C3 was both minimal and patchy or occasionally absent (not shown).

Figure 3.

Figure 3.

Macroscopic and microscopic assessment of pig liver xenografts. (A) Macroscopic appearance of the pig liver in B18908 at necropsy (day 6). This liver showed fewer dark areas than most other livers. (B) Histology of the pig liver in B7808 at necropsy (day 6) (H&E) from a macroscopically injured (dark) area showing areas of hemorrhagic necrosis interspersed within conserved liver structure (×100 magnification). (C) Same area at ×200. (D) From a macroscopically normal area, showing normal liver structure (×100). (E) Same area at ×200. (F) Confocal microscopy of a GTKO/CD46 pig liver (in B7808) at necropsy (day 6) after immunofluorescence staining for platelets (CD41; red), PBMCs (CD45; green) and nuclei (DAPI; blue) showing platelet and platelet-PBMC aggregates in the liver sinusoids (×1000). Immunofluorescence studies showed platelet and/or platelet-PBMC aggregates in the blood and within the liver grafts. In most cases, deposition of IgM, IgG and C3 was both minimal and patchy or occasionally absent (not shown).

Figure 3.

Figure 3.

Macroscopic and microscopic assessment of pig liver xenografts. (A) Macroscopic appearance of the pig liver in B18908 at necropsy (day 6). This liver showed fewer dark areas than most other livers. (B) Histology of the pig liver in B7808 at necropsy (day 6) (H&E) from a macroscopically injured (dark) area showing areas of hemorrhagic necrosis interspersed within conserved liver structure (×100 magnification). (C) Same area at ×200. (D) From a macroscopically normal area, showing normal liver structure (×100). (E) Same area at ×200. (F) Confocal microscopy of a GTKO/CD46 pig liver (in B7808) at necropsy (day 6) after immunofluorescence staining for platelets (CD41; red), PBMCs (CD45; green) and nuclei (DAPI; blue) showing platelet and platelet-PBMC aggregates in the liver sinusoids (×1000). Immunofluorescence studies showed platelet and/or platelet-PBMC aggregates in the blood and within the liver grafts. In most cases, deposition of IgM, IgG and C3 was both minimal and patchy or occasionally absent (not shown).

Figure 3.

Figure 3.

Macroscopic and microscopic assessment of pig liver xenografts. (A) Macroscopic appearance of the pig liver in B18908 at necropsy (day 6). This liver showed fewer dark areas than most other livers. (B) Histology of the pig liver in B7808 at necropsy (day 6) (H&E) from a macroscopically injured (dark) area showing areas of hemorrhagic necrosis interspersed within conserved liver structure (×100 magnification). (C) Same area at ×200. (D) From a macroscopically normal area, showing normal liver structure (×100). (E) Same area at ×200. (F) Confocal microscopy of a GTKO/CD46 pig liver (in B7808) at necropsy (day 6) after immunofluorescence staining for platelets (CD41; red), PBMCs (CD45; green) and nuclei (DAPI; blue) showing platelet and platelet-PBMC aggregates in the liver sinusoids (×1000). Immunofluorescence studies showed platelet and/or platelet-PBMC aggregates in the blood and within the liver grafts. In most cases, deposition of IgM, IgG and C3 was both minimal and patchy or occasionally absent (not shown).

Histology and immunohistology of pig livers

Early features of hyperacute rejection (interstitial hemorrhage and edema, polymorphonuclear cellular infiltrate, intravascular thrombosis) were seen in the WT pig liver transplanted into the nonimmunosuppressed baboon (B16907) (not shown). Histology of the GTKO and GTKO/CD46 pig livers taken 1–2 h after reperfusion showed normal liver structure, with no features of hyperacute rejection (not shown); immunofluorescence staining demonstrated minimal or absent IgM, IgG and C3 deposition, confirming the lack of immunologic damage (not shown). Macroscopic and microscopic features of the livers at necropsy are shown in Figure 3.

Discussion

There is a significant mortality of patients with acute or acute-on-chronic liver failure and in those with primary nonfunction of an allograft (30). ‘Bridging’ of these patients has been attempted by hepatocyte Tx, bioartificial liver support and extracorporeal pig liver perfusion, but none of these approaches has convincingly improved patient survival (6,31–33). The orthotopic Tx of a genetically engineered pig liver could potentially provide successful bridging and offer better metabolic support than the above approaches; removal of a necrotic native liver might prevent further deterioration in the patient's clinical condition while awaiting alloTx.

There are few reports of xenogeneic liver Tx in large animal models (reviewed in Hara H et al. [9]) (See subsection ‘Xenogeneic Liver Transplantation in Large Animal Models’ in Supporting Information.) and only one clinical attempt at pig liver xenoTx (34) (see subsection ‘Clinical Pig Liver Transplantation’ in Supporting Information).

The genetic engineering of pigs holds out great prospects in protecting their organs from primate humoral and cellular immune responses (8,13). There are no previous published data reporting the survival of livers from GTKO pigs after Tx into nonhuman primates. Our main aim was not to obtain long-term graft or recipient survival, but to determine whether GTKO or GTKO/CD46 pig livers would support a baboon for a period of time sufficient—under clinical circumstances—to obtain an allograft, which, in most clinical situations, is within 7 days. Although the numbers are small, there appeared to be no difference in outcome between GTKO and GTKO/CD46 pig liver Tx. Following heart and kidney Tx, the addition of CD46 reduces the incidence of early graft failure (Azimzadeh et al., unpublished).

The one WT pig liver transplanted showed early features of hyperacute rejection. Using GTKO or GTKO/CD46 pig livers, we did not observe hyperacute rejection, confirmed on histological examination and immunofluorescence of biopsies taken after 1–2 h in which there was a virtual absence of IgM, IgG and C3 deposition on the vascular endothelium.

One of the early problems we met was that of size-mismatch between donor pig and recipient baboon (Tables 1 and 3) (see subsection ‘Size-Mismatch between Donor and Recipient’ in Supporting Information).

The major complication seen in all xenoTx experiments was the development of profound thrombocytopenia that began within 1 h after reperfusion and reached dangerously low levels in all but one case within 5 h (Figure 1). Immediate or early thrombocytopenia is not seen after Tx of pig hearts or kidneys into baboons under similar experimental conditions (19–22), though it is seen early after pig lung Tx (35). The reasons for these discrepancies are not yet clear. Thrombocytopenia was, however, documented by Ramirez et al. in the pig-to-baboon liver Tx model (10) and also by Starzl et al. after baboon-to-human liver Tx (36), but in neither case did the platelet numbers fall so low. Tector et al. also reported almost immediate thrombocytopenia following the Tx of dog livers in pigs (37).

In the present study, immunofluorescence staining indicated that platelet/PBMC aggregates developed in the liver sinusoids by post-Tx day 6 and 7 (Figure 3F), but were not seen in liver biopsies taken within 2 h of reperfusion, by which time thrombocytopenia had already developed. The exact cause of the rapid development of thrombocytopenia therefore remains uncertain. Thrombocytopenia has been reported after ABO-incompatible liver alloTx in association with thrombotic microangiopathy and consumptive coagulopathy (38), but we did not identify definitive features of these pathologies (see subsection ‘Thrombotic Microangiopathy’ in Supporting Information). The administration of clodronate liposomes in an effort to deplete macrophages in the donor pig before liver excision did not appear to prevent this rapid loss of platelets, suggesting, though not proving, that they were not being taken up by pig Kupffer cells, as might have been suspected from the studies by Rees et al. (39).

There were some late features of inadequate coagulation, for example INR > 2.0 and fibrinogen < 75 mg/dL, but our impression was that these were a consequence of repeated internal hemorrhages rather than being primary causative factors in the onset of bleeding. We cannot, however, completely exclude that coagulation inefficiencies were playing a role in the development of spontaneous bleeding. Although we administered heparin in two baboons in an effort to reduce platelet aggregation, we did not utilize specific anti-platelet agents, which might have been more effective, though we believe this is unlikely.

The development of platelet/PBMC aggregates in the blood would appear to be important in the loss of free platelets. We failed to identify platelet or platelet-PBMC aggregates in the liver in biopsies taken 1 h after reperfusion, even though profound thrombocytopenia had developed by this time. Aggregation could possibly have been focal, or sinusoidal EC may have already engulfed the aggregates. Studies in our laboratory (that will be reported elsewhere) indicated that aggregation began immediately after reperfusion, and increased from 2% to 67% within days of liver Tx when measured by flow cytometry. Expression of tissue factor (TF) by baboon platelets and/or monocytes (and pig EC) may be an important factor (40,41). We postulate that, once recipient platelets (and monocytes) become activated by contact with pig EC, they express TF (40), initiating aggregation systemically and, particularly, in the liver sinusoids. EC expression of von Willibrand factor may play a role (Lin CC, unpublished). Prior splenectomy did not prevent loss of platelets. The severe thrombocytopenia resulted in episodes of spontaneous hemorrhage at various extra-hepatic sites and in the liver graft, necessitating euthanasia.

Microscopic examination of the liver grafts indicated patches of hemorrhage and necrosis with vascular thrombosis. We hypothesize that these were associated with aggregation of platelets in contact with the vascular EC, resulting in obstruction to the microcirculation. These random areas of pathology showed no clear pattern and did not conform to the underlying anatomy, for example they were not related to the portal triads.

Immunofluorescence staining for IgG, IgM and C3 at necropsy confirmed minimal or patchy immunoglobulin and virtually no complement deposition on the endothelium. Thus, although we cannot rule out a role for antibody in initiating activation of the donor endothelium, leading to activation of recipient platelets, our in vitro studies suggest that neither antibody nor complement is necessary for platelet TF to be expressed. Nevertheless, the absence of immunoglobulin deposition does not rule out the effect of antibody in ABO-incompatible liver allografts (42). Furthermore, immunoglobulin can be phagocytosed by the sinusoidal EC (43), complicating interpretation of the pathology.

Monitoring of pig liver function provided encouraging data. At least for the first few days, almost normal liver function was documented by liver enzymes and synthesis of pig proteins and pig coagulation factors (not shown), resulting in near-normal parameters of coagulation. Increases in GGT and ALP in two cases, together with increases in total and direct bilirubin levels, suggested bile stasis, confirmed by thick viscous bile in the ducts at necropsy unassociated with surgical obstruction. This finding correlates with similar observations by Starzl et al. in their baboon-to-human liver transplants (36). However, we cannot exclude duct-specific immunologic injury as a factor.

There was no increase in anti-nonGal IgM or IgG antibody levels or of complement-dependent cytotoxicity of pig cells, indicating no sensitization to the donor. In vitro studies in our laboratory (Hara H et al., unpublished) have demonstrated that PBMCs and vascular EC from GTKO/CD46 pigs are almost totally protected against human or baboon serum complement-dependent cytotoxicity, an observation confirmed in the present study. The single post-Tx MLR carried out indicated almost complete suppression of a cellular response. In any case, although the data are currently limited, sensitization to a pig xenograft does not appear to be detrimental to a subsequent allograft (44).

We conclude that, after orthotopic GTKO or GTKO/CD46 pig liver Tx in conventionally immunosuppressed baboons, there is adequate hepatic function, as evidenced by the majority of tests of liver function, complement activity, protein synthesis, coagulation factors and coagulation parameters. The major problem that prevented more prolonged survival beyond 7 days was the effect of the profound thrombocytopenia that developed within hours of reperfusion of the liver, ultimately resulting in spontaneous hemorrhage at various sites. We postulate that this is associated with the expression of TF on platelets after contact with pig endothelium, resulting in platelet and platelet-monocyte aggregation, and deposition of aggregates in the liver graft; phagocytosis within the liver may also play a role. If this problem can be resolved, we would anticipate that a pig liver could provide a period of support during which a patient could be successfully bridged to alloTx.

Acknowledgments

Burcin Ekser, MD, is a recipient of an American Society of Transplantation/European Society for Organ Transplantation Exchange Grant and of a Young Investigator Award from the American Transplant Congress, 2009. The authors thank Drs. Dirk J. van der Windt and Eefje M. Dons for surgical assistance, Drs. Richard Pierson, Vladimir Bogdanov and Dorian Haskard for providing essential agents for the study, and Diann Flunk-Flavin, Stacey Cashman, and Michael Nakon for excellent technical help with the operative procedures. Work on xenotransplantation in the Thomas E. Starzl Transplantation Institute of the University of Pittsburgh is supported in part by NIH grants # U01 AI068642 and # R21 A1074844, and by Sponsored Research Agreements between the University of Pittsburgh and Revivicor, Inc., Blacksburg, VA. The baboons were provided by the Oklahoma University Health Sciences Center, Division of Animal Resources, which is supported in part by NIH P40 sponsored grant RR012317-09.

Ancillary