Influence of extracorporeal porcine liver perfusion on nonhuman primates: Minimizing hemolysis improves subsequent survival

Influence of Extracorporeal Porcine Liver Perfusion on Nonhuman Primates: Minimizing Hemolysis Improves Subsequent Survival Ryuta Nishitai, Iwao Ikai, Hiroaki Terajima, Akiyoshi Kanazawa, Osamu Takeyama, Takehiko Uesugi, Hiroshi Okabe, Nagato Katsura, Takakazu Matsushita, Satoshi Yamanokuchi, Koichi Matsuo, Shinichi Sugimoto, Tomohiro Shiotani, and Yoshio Yamaoka The aim of this study is to detect and analyze risk factors of direct cross-circulation between porcine liver and nonhuman primates before a clinical application of extracorporeal liver perfusion (ECLP) as a liver-assist method. Porcine livers were perfused with baboon blood in an ECLP system. Six healthy baboons were directly connected to the ECLP system with continuous prostaglandin E1 administration. Cross-circulation was terminated in the following circumstances: (1) hepatic arterial or portal perfusion pressures elevated to 200 or 60 mm Hg, respectively; (2) massive exudative bleeding from the graft surface; or (3) bile output decreased to less than 5 ␮L/h/g of liver weight. In case 1, cross-circulation was continued for 10 hours. Severe macroscopic hemolysis occurred, and serum hemoglobin (s-Hb) concentration reached a peak of 47 mg/dL. The baboon died of acute renal failure 2 days later. Histological study of the perfused porcine liver showed marked microthrombi formation. In 3 of the later 5 cases, cross-circulation was discontinued when mild macroscopic hemolysis was observed. The duration of the 5 cross-circulations was maximally 6 hours (mean, 4.4 ⴞ 1.2 [SD] hours). Mean s-Hb concentration in the 5 cases was elevated to 14.8 ⴞ 5.8 mg/dL at the end of crosscirculation and decreased to the baseline level within 24 hours. These 5 baboons survived without organ dysfunction or immunologic disturbance. When severe hemolysis is avoided, direct cross-circulation using the ECLP system can be achieved without serious complications in nonhuman primates. (Liver Transpl 2001;7:615-622.)

E

xtracorporeal liver perfusion (ECLP) is a temporary liver-assist method. First reported by Otto et al,1 it was studied extensively during the 1960s to 1980s.2-5 ECLP was found to improve the neurological status of patients with acute hepatic failure by facilitating the removal of such toxic substances as ammonia and bilirubin.6 Since the 1980s, orthotopic liver transplantation has become the established treatment of choice for severe acute hepatic failure, and ECLP was not studied for a time. However, a chronic shortage of human livers available for transplantation and the possibility that the need for transplantation might be averted encouraged renewed interest in liver-assist methods. Thus, ECLP now is being reevaluated as a therapeutic modality of intervention in patients with acute hepatic failure.7-9 However, to date, there have

been no preclinical studies that evaluated the safety of ECLP in nonhuman primates. We developed a new ECLP system using a whole porcine liver as a liver-assist device. The administration of prostaglandin E1, which ameliorates injury mediated by xenogeneic humoral factors, has enabled the xenoperfused porcine liver to be kept viable for 9 hours in a closed circuit.10-13 The metabolic capacity of this ECLP system has been pharmacokinetically analyzed using galactose-, lidocaine-, and ammonia-loading tests and documented to maintain functional capacity similar to that of in situ porcine livers for up to 9 hours.14 Before an application of the ECLP system in the clinical arena, a preclinical assessment using nonhuman primates is a prerequisite. Because symptoms of acute hepatic failure vary, it is difficult to distinguish potential complications of ECLP from complications of the underlying liver disease. In this preclinical study, healthy baboons were used to investigate the influences of the ECLP system on primates.

Materials and Methods Animals and Liver Graft Procurement Six specific pathogen-free female pigs weighing 8 to 15 kg were obtained from Laboproducts Inc (Osaka, Japan). The pigs were fasted for 24 hours before the experiments, but

From the Department of Gastroenterological Surgery, Kyoto University Graduate School of Medicine, Kyoto, Japan. Supported in part by grants no. 09557105, 10044270, 10357010, and 11470244 from the Scientific Research Fund of the Ministry of Education, Japan, and grant no. JSPS-RFTF 96I00204 from the Research for the Future of Japan Society for the Promotion of Science. Address reprint requests to Ryuta Nishitai, MD, Department of Gastroenterological Surgery, Kyoto University Graduate School of Medicine, 54 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto, 606-8507, Japan. Telephone: 81-75-751-3242; FAX: 81-75-751-3669; E-mail: rnishi@ kuhp.kyoto-u.ac.jp Copyright © 2001 by the American Association for the Study of Liver Diseases 1527-6465/01/0707-0006$35.00/0 doi:10.1053/jlts.2001.25362

Liver Transplantation, Vol 7, No 7 ( July), 2001: pp 615-622

615

616

Nishitai et al

were allowed water ad libitum. Livers were harvested as described previously.11,14 Briefly, the livers were completely dissected from surrounding tissues and procured during in situ rapid flushing with 1,500 mL of cold (4°C) heparinized Ringer’s lactate solution through both the portal vein (PV) and hepatic artery (HA). Bile ducts were cannulated to accurately monitor bile production. Porcine livers weighed 292 ⫾ 39 g (mean ⫾ SD). Seven healthy baboons (Papio anubis; age, 3 to 13 years; Japan Monkey Centre, Aichi, Japan) weighing 11.5 to 20.5 kg (15.6 ⫾ 3.7 kg) were fed a standard commercial diet and allowed free access to food and water before experiments. One baboon was assigned to be a blood donor. Blood compatibility between the donor baboon and other animals was confirmed through cross-matching. Care of animals and all procedures were performed in accordance with the Guideline for Animal Care and Experiments of Kyoto University (Kyoto, Japan).

ECLP The porcine liver was floated in a 37°C colloidal solution to simulate the position in a prone pig. The perfusate was composed of fresh baboon blood diluted with Ringer’s lactate solution containing low-molecular-weight dextran 40 to give a hematocrit of 30%. The reservoir of the ECLP system was filled with perfusate and connected to the liver, which was perfused through both the PV and HA with separate roller pumps (CCS-20; JMS, Tokyo, Japan). Blood flow through the liver was maintained at a constant rate of 1 mL/min/g of liver, with total HA flow adjusted to be approximately 25% of total PV flow. Blood was returned to the reservoir through the suprahepatic inferior vena cava by a hydrostatic pressure gradient of 10 cm H2O. Partial oxygen pressures of the PV and HA were maintained between 45 and 60 mm Hg and 100 to 150 mm Hg, respectively, using separate membrane oxygenators with heat exchangers (Menox AL-2000; Kuraray, Osaka, Japan). The temperature of the perfusate was kept at 37°C throughout cross-circulation. Hydroxycortisone (100 mg/L), ampicillin (500 mg/L), and low-molecular-weight heparin (5 U/mL; dalteparin sodium; Kissei Pharmaceutical Co Ltd, Tokyo, Japan) were added to the perfusate before perfusion. Prostaglandin E1 (Ono Co Ltd, Osaka, Japan) was administered continuously at 2 ng/min/g of liver weight during cross-circulation. The activated coagulation time of the perfusate was maintained between 200 and 250 seconds by the intermittent administration of low-molecular-weight heparin. PV and HA pressures, exudate from the graft surface, and bile production were monitored as markers of liver graft viability. Baboons were intubated orotracheally under sedation with the intramuscular administration of ketamine hydrochloride (5 mg/kg) and xylazine (1 mg/kg; Bayer, Tokyo, Japan). General anesthesia consisted of nitrogen oxide and sevoflurane (Maruishi Pharmaceutical, Osaka, Japan), with an intermittent intravenous injection of ketamine hydrochloride and vecuronium bromide. A large-bore intravenous coaxial double lumen catheter (Blood Access UK-Catheter Kit 11.5 F; Unitika, Hyogo, Japan) was placed in the inferior vena

cava through the femoral vein and connected to the reservoir of the ECLP system. The flow rate of cross-circulation between the baboon and reservoir was adjusted to 50 mL/min by synchronized roller pumps (CCS-20; JMS). Arterial pressure, partial carbon dioxide pressure of the expired gas, and oxygen saturation level were monitored and kept within the physiological range. Urine output was also monitored through an indwelling urinary catheter.

Experimental Design We previously investigated the viability of extracorporeally perfused liver grafts by using a closed circuit.10,11,14 Based on our previous results, cross-circulation was terminated when any of the following conditions were observed: (1) HA perfusion pressure of 200 mm Hg or greater, (2) PV perfusion pressure of 60 mm Hg or greater, (3) massive exudative bleeding from the surface of the graft, or (4) bile output of 5 ␮L/h/g of liver weight or less.

Complete Blood Count, Biochemical Analysis, and Complement Activity Blood was sampled at various times for determination of complete blood count, lactate dehydrogenase (LDH), alanine aminotransferase (ALT), and creatinine values by an autoanalyzer. Serum hemoglobin (s-Hb) was measured by colorimetry. Total complement activity was evaluated by an assay of complement hemolytic activity and expressed as hemolytic complement units (CH50). Levels of complement factors C3 and C5 were measured using turbidimetric immunoassay and nephelometry, respectively.

Phenotype Analysis of Baboon Peripheral Lymphocytes Baboon peripheral-blood mononuclear cells (PBMCs) were isolated by Ficoll-Paque (Pharmacia Biotech; Uppsala, Sweden) density gradient centrifugation. Cells were washed twice and stained with a combination of the following phycoerythrin- or fluorescein isothiocyanate–labeled antihuman monoclonal antibodies (all from Becton-Dickinson, Mountain View, CA) for 30 minutes at 4°C: CD4 (helper/inducer T lymphocytes), CD8 (suppressor/cytotoxic T lymphocytes), CD20 (B lymphocytes), CD23 (B-lymphocyte differentiation antigen), CD25 (interleukin-2 receptor), and CD69 (early activation marker of lymphocytes). Cells were washed twice, and cell phenotypes were analyzed by flow cytometry using FACSCalibur (Becton-Dickinson).

Assay of Antiporcine Antibodies Titers of antiporcine immunoglobulin M (IgM) and IgG in baboon plasma samples were determined by flow cytometry using porcine PBMCs as targets. Porcine PBMCs were collected by centrifugation using Ficoll-Paque concentration gradient. Aliquots of 0.5 ⫻ 106 were cultured for 30 minutes at 4°C with 100 ␮L of 1:10 diluted baboon plasma samples. After a secondary staining with phycoerythrin-conjugated mouse antihuman IgM monoclonal antibody or fluorescein

Extracorporeal Porcine Liver Perfusion

isothiocyanate– conjugated mouse antihuman IgG monoclonal antibody (PharMingen Int, Fujisawa Pharmaceutical Co Ltd, Tokyo, Japan), stained cells were examined by flow cytometry. Results are expressed as relative mean channel fluorescence.

Porcine-Derived Antibody Concentrations of porcine-derived IgG and IgM antibodies in baboon blood samples were determined by a sandwich enzyme-linked immunosorbent assay (ELISA). Affinity-purified antiporcine IgG or IgM goat antibody (Kirkegaard & Perry Laboratories, Gaithersburg, MD) was used to coat the ELISA plate. Serially diluted solutions of purified normal porcine IgG (Jackson ImmunoResearch Laboratories Inc, West Grove, PA) and IgM (Biogenesis Inc, Kingston, NH) acted as controls. Baboon plasma samples and control solutions were added to the ELISA plate and incubated at room temperature for 1 hour. The plate was washed 3 times with phosphate-buffered saline. Purified goat antibody directed against porcine IgG or IgM was conjugated to alkaline phosphatase (Kirkegaard & Perry Laboratories) and applied as the secondary antibody. The plate was washed again with phosphate-buffered saline. After a 15-minute reaction with p-nitrophenyl phosphate alkaline phosphatase substrate (Sigma-Aldrich, St Louis, MO), the absorbance at 405 nm was determined by spectrophotometry.

Histological Studies Tissue samples of perfused porcine livers were obtained immediately after the cessation of cross-circulation, fixed with 4% formaldehyde, and embedded in paraffin. An autopsy was performed if a baboon died. Biopsy specimens were stained with hematoxylin and eosin, and fibrin deposition was determined with phosphotungstic acid hematoxylin staining.

In Vitro Hemolytic Assay An in vitro hemolytic assay was used to examine whether porcine antibody could induce hemolysis of baboon red blood cells (RBCs). Washed RBCs and normal serum were collected from the donor baboon. RBCs (2.5 ⫻ 108) were incubated for 20 minutes at 37°C with 300-␮L samples of either normal porcine serum (RBC-xeno), normal baboon serum (RBCallo), or plasma from a baboon that had undergone a crosscirculation (RBC-exp) in which complement activity was completely blocked by 10 mmol/L of EDTA. Sensitized RBCs were then washed and incubated with 300 ␮L of normal baboon serum as a source of complement for 1 hour at 37°C. The s-Hb concentration of each supernatant was determined by colorimetry.

Statistical Analysis Data are expressed as mean ⫾ SD. The significance of alterations of parameters at varying times with respect to baseline levels was statistically analyzed by a paired t-test with

617

Dunnett adjustment. Statistical significance is defined as P less than .05.

Results Direct Cross-Circulation and Postoperative State of Baboons In case 1, cross-circulation was maintained for 10 hours. s-Hb concentrations increased gradually, reaching a peak of 47 mg/dL (Fig. 1A). Cross-circulation was discontinued after severe macroscopic hemolysis was observed. The baboon developed oliguric renal failure and died 2 days later. Although the source of s-Hb included porcine passenger RBCs, as well as baboon RBCs, the porcine liver was washed out with abundant Ringer’s lactate solution; thus, passenger porcine RBCs were supposed to be few. Based on this result, all subsequent cross-circulations were terminated when mild macroscopic hemolysis was observed in addition to the previously outlined criteria. The duration of cross-circulation for cases 2 to 6 was 4.4 ⫾ 1.2 hours. Mild macroscopic hemolysis occurred in 3 of the 5 experiments, but exudative bleeding from the graft surface was not observed and bile production was kept constant at 14.6 ⫾ 4.3 ␮L/h/g of liver. Normal urine output was maintained during the observation period. The baboons rapidly recovered from anesthesia and did not require medication thereafter. They regained a normal appetite within 3 days and have survived for more than 1 year (Table 1). The general status of baboons in cases 2 to 6 was evaluated by complete blood count, s-Hb level, and biochemical analysis (Table 2). The animals experienced severe anemia during cross-circulation, but hematocrits recovered to baseline levels at 1 month after cross-circulation. s-Hb level reached 14.8 ⫾ 5.8 mg/dL at the end of cross-circulation, but returned to baseline level within 24 hours. Hemolysis was not observed after the discontinuation of cross-circulation. At the end of cross-circulation, there was a 6- to 7-fold elevation in serum LDH level compared with baseline values, which was similar to the s-Hb level elevation. LDH levels normalized over 3 days. The platelet count decreased significantly during perfusion, but recovered to baseline levels within 3 days. Creatinine levels remained within the normal range both during and after cross-circulation. ALT level did not increase during cross-circulation, but was elevated to 88.4 ⫾ 48.2 IU/L at 24 hours, indicating slight hepatocellular damage, but rapidly normalized. Long-term observations showed that direct cross-circulation with the porcine liver did not induce serious hepatic or renal injury.

618

Nishitai et al

Table 1. Duration of Cross-Circulation, Reasons for Discontinuation, and Clinical Courses of Baboons Case No.

Perfusion Duration (h)

Reason for Discontinuation of Cross-Circulation

Organ Dysfunction of Baboon

Prognosis of Baboon

1 2 3 4 5 6

10 2.5 3.5 5 6 5

Severe macroscopic hemolysis and anuria (baboon side) Arterial perfusion pressure ⬎200 mm Hg (graft side) Portal perfusion pressure ⬎60 mm Hg (graft side) Mild macroscopic hemolysis (baboon side) Mild macroscopic hemolysis (baboon side) Mild macroscopic hemolysis (baboon side)

Acute renal failure Nothing particular Nothing particular Nothing particular Nothing particular Nothing particular

Dead (⬉2 d) Alive (⬎1 yr) Alive (⬎1 yr) Alive (⬎1 yr) Alive (⬎1 yr) Alive (⬎1 yr)

Immunologic Evaluation of Baboons Changes in immunologic components are listed in Table 3. CH50 and C5 did not show significant changes. C3 level decreased transiently at the end of cross-circulation, recovering to the normal range within 24 hours. These data suggest that although complement components of the host animal may have been activated by contact with the extracorporeal circuit, including the porcine liver graft, complement consumption was insufficient to reduce the total complement cascade activity. The CD4⫹/CD8⫹ ratio of T cells and CD25 on both CD4⫹ and CD8⫹ T cells were stable for 6 months. CD69 on CD4⫹ and CD8⫹ T cells and CD69 or CD23 on B cells (CD20⫹) did not show a remarkable change up to 6 months (data not shown). Antipig IgM and IgG levels decreased during perfusion and remained low for 3 days. After a remarkable increase at 1 week, they gradually decreased to baseline levels by 1 year (Table 3). Very low levels of porcine-derived IgG and IgM

antibodies were detectable in baboon blood both during and after cross-circulation (Table 4). The levels gradually decreased and became undetectable after 1 to 4 weeks. In comparison to IgG/IgM concentrations present in control baboon serum (mean, 975 ⫾ 173 and 112 ⫾ 22 mg/dL, respectively; n ⫽ 6), porcinederived IgG and IgM accounted for approximately 0.1% and 2% of total serum IgG and IgM at the end of cross-circulation, respectively. Histological Studies Hematoxylin and eosin staining of the porcine liver from case 1 showed periportal edema, neutrophil adhesion to the endothelium of the portal vein and sinusoids, and hemorrhage in Glisson’s capsule (Fig. 2A). In addition, focal necrosis of hepatocytes was also evident. Conversely, the only histopathologic changes found in the porcine liver of case 5, which underwent the longest perfusion among cases 2 to 6, were mild periportal edema and neutrophil attachment to the endothelial surface (Fig. 2B). Porcine liver grafts of the remaining

Table 2. Laboratory Data of Cases 2 to 6 Before and After Cross-Circulation

Time

Hb (g/dL)

s-Hb (mg/dL)

Platelets (⫻104/␮L)

LDH (IU/L)

ALT (IU/L)

Creatinine (mg/dL)

Baseline* End point† 24 h 3d 1 wk 1 mo 3 mo 1 yr

13.6 ⫾ 1.8 9.0 ⫾ 1.9‡ 7.3 ⫾ 1.7§ 7.3 ⫾ 2.2§ 8.2 ⫾ 0.9§ 12.2 ⫾ 1.4 12.2 ⫾ 0.4 11.7 ⫾ 1.3

1.8 ⫾ 0.7 14.8 ⫾ 5.8§ 1.6 ⫾ 1.0 1.8 ⫾ 0.4 2.6 ⫾ 0.8 2.4 ⫾ 0.8 1.8 ⫾ 1.2 3.0 ⫾ 3.5

27.5 ⫾ 11.7 7.7 ⫾ 3.9‡ 7.6 ⫾ 1.6‡ 11.6 ⫾ 4.3 25.4 ⫾ 8.3 28.6 ⫾ 7.3 26.8 ⫾ 13.7 27.3 ⫾ 6.8

329 ⫾ 66 2,232 ⫾ 2,389‡ 1,834 ⫾ 1,116 911 ⫾ 378 726 ⫾ 263 314 ⫾ 63 353 ⫾ 88 363 ⫾ 88

26.4 ⫾ 9.7 36.2 ⫾ 22.6 88.4 ⫾ 48.2§ 57.0 ⫾ 30.8 36.0 ⫾ 13.6 40.2 ⫾ 22.4 44.8 ⫾ 27.6 35.2 ⫾ 13.0

0.88 ⫾ 0.13 0.90 ⫾ 0.13 0.94 ⫾ 0.15 0.80 ⫾ 0.06 0.82 ⫾ 0.12 0.92 ⫾ 0.19 0.96 ⫾ 0.27 1.06 ⫾ 0.22

NOTE. Values expressed as mean ⫾ SD (n ⫽ 5). *Baseline sample was collected shortly before cross-circulation. †End point sample was collected immediately after cross-circulation was terminated. ‡P ⬍ .05. §P ⬍ .01.

619

Extracorporeal Porcine Liver Perfusion

Table 3. Complement and Xenoreactive Antibody Levels of Cases 2 to 6 After Cross-Circulation

Time

CH50 (U/mL)

C3 (mg/dL)

C5 (mg/dL)

Antiporcine IgM*

Antiporcine IgG*

Baseline End point 24 h 3d 1 wk 1 mo 3 mo 1 yr

16.6 ⫾ 2.3 15.3 ⫾ 3.5 15.9 ⫾ 3.8 19.0 ⫾ 2.5 19.2 ⫾ 1.1 18.3 ⫾ 4.1 20.4 ⫾ 3.4 18.0 ⫾ 4.2

48.4 ⫾ 4.5 29.2 ⫾ 6.4† 34.2 ⫾ 6.6 61.2 ⫾ 5.6 66.0 ⫾ 9.7 57.4 ⫾ 13.2 60.4 ⫾ 14.4 60.6 ⫾ 7.6

7.4 ⫾ 0.7 5.0 ⫾ 1.2 5.8 ⫾ 1.6 7.7 ⫾ 2.9 9.2 ⫾ 2.3 6.9 ⫾ 1.2 7.7 ⫾ 1.2 8.0 ⫾ 0.9

19.7 ⫾ 11.0 ND 4.6 ⫾ 2.7† 9.8 ⫾ 5.9† 52.7 ⫾ 23.5† 42.6 ⫾ 26.7 31.1 ⫾ 16.1 14.8 ⫾ 6.4

3.1 ⫾ 3.6 ND 1.9 ⫾ 1.0 3.9 ⫾ 5.3 19.8 ⫾ 26.9 25.6 ⫾ 20.5 9.3 ⫾ 9.1 1.7 ⫾ 0.2

NOTE. Values expressed as mean ⫾ SD (n ⫽ 5). Abbreviation: ND, not determined. *Expressed as relative mean channel fluorescence. †P ⬍ .05.

cases also showed correspondingly mild histopathologic findings (data not shown). Phosphotungstic acid hematoxylin staining of the porcine liver from case 1 showed the formation of microthrombi accompanying fibrin net along sinusoids and in small vessels in Glisson’s capsule (Fig. 3). However, cases 2 to 6 did not show significant evidence of intravascular thrombosis. Light microscopic examination of the baboon’s kidneys from case 1 showed scattered thrombi in glomerular capillaries and diffuse acute tubular necrosis (Fig. 1B), whereas such other organs as lungs and heart were not injured. Renal biopsies of the baboons in cases 2 to 6 performed 1 year after cross-circulation were normal.

uate whether porcine antibody-mediated intravascular hemolysis of baboon RBCs was present. RBCs of a normal baboon incubated with normal baboon’s serum should remain free from xenogeneic antibodies. Conversely, RBCs incubated with either porcine serum or plasma samples obtained from baboons at the end point of cross-circulation may be expected to be sensitized with complement-fixing porcine antibaboon antibodies that would induce cell lysis in the subsequent incubation with normal baboon serum. Mean s-Hb concentrations after hemolytic reactions using RBC-exp, RBC-allo, and RBC-xeno were 9.2 ⫾ 3.5, 7.0 ⫾ 0.0, and 12.3 ⫾ 5.2 mg/dL, respectively. No significant difference was found among the 3 values.

In Vitro Hemolytic Assay The detection of porcine IgG and IgM in the plasma of baboons during cross-circulation prompted us to eval-

Table 4. Porcine IgG and IgM Antibody Levels in Plasma Samples of Cases 1 to 6 IgG (mg/dL)

IgM (mg/dL)

Time

Case 1

Case 2-6

Case 1

Case 2-6

Baseline End point 24 h 3d 1 wk 1 mo 3 mo

0.0 1.1 0.8

0.0 ⫾ 0.0 0.6 ⫾ 0.2 0.5 ⫾ 0.1 0.3 ⫾ 0.1 0.3 ⫾ 0.1 0.1 ⫾ 0.1 0.0 ⫾ 0.0

0.0 2.3 1.2

0.0 ⫾ 0.0 3.9 ⫾ 1.0 1.4 ⫾ 0.4 0.4 ⫾ 0.2 0.1 ⫾ 0.1 0.0 ⫾ 0.0 0.0 ⫾ 0.0

NOTE. Values for Cases 2 to 6 are expressed as mean ⫾ SD (n ⫽ 5).

Discussion This study was performed to evaluate the safety of direct cross-circulation between porcine livers and nonhuman primates using our ECLP system. Use of the ECLP system for 10 hours resulted in severe macroscopic hemolysis and irreversible acute renal failure. Conversely, when cross-circulation was performed for less than 6 hours, hemolysis was mild and kidneys and liver did not sustain significant injury. A temporary decrease and the subsequent increase in antipig IgM and IgG levels indicate that xenoreactive natural antibodies were absorbed by xenogeneic ECLP, but production of antipig antibodies was induced by exposure to xenoantigens. Activation markers on T cells or B cells did not show a remarkable change, suggesting that xenogeneic ECLPs have no serious long-term influence on the cellular immunity of patients. The most significant risk of the procedure appears to

620

Nishitai et al

Figure 1. (A) s-Hb concentration during cross-circulation of case 1. (B) Pathological findings of the baboon kidney from case 1 showing scattered intravascular thrombi and severe acute tubular necrosis. (Hematoxylin and eosin staining; original magnification ⴛ400.)

be hemolysis, reported previously in xenogeneic perfusion models of the kidney, lung, and liver.11,15-17 The mechanisms of hemolysis in organ xenoperfusion have not yet been fully elucidated. Two factors may be impli-

Figure 3. Pathological findings of the liver graft from case 1 after 10 hours of cross-circulation showing microthrombi along sinusoids and in small vessels. (Phosphotungsic acid hematoxylin staining; original magnification ⴛ400.)

Figure 2. (A) Pathological findings of the liver graft from case 1 after 10 hours of cross-circulation showing periportal edema, neutrophil attachment to portal and endothelial surfaces, and hemorrhage in Glisson’s capsule. (B) Pathological findings of the liver graft from case 5 after 6 hours of cross-circulation showing mild periportal edema and neutrophil attachment to portal and endothelial surfaces. (Hematoxylin and eosin staining; original magnification ⴛ400.)

cated in hemolysis during cross-circulation with xenogeneic ECLP. The first is mechanical hemolysis induced by the circuit itself, and the second is hemolysis related to xenogeneic immunologic reactions. It is possible that the extracorporeal circuit, including roller pumps and membrane oxygenators, could mechanically damage erythrocytes.18,19 However, in our previous study in which fresh human blood was perfused for 9 hours in the ECLP circuit without the porcine liver, s-Hb levels were low. Although membrane physiological characteristics of baboon RBCs and human RBCs are not exactly the same, there are several common characteristics, such as the major antigen, size, and shape.20,21 Therefore, it was suggested that the cir-

Extracorporeal Porcine Liver Perfusion

cuit did not significantly contribute to the hemolysis evident during the cross-circulation of case 1.11 Complement component C1q is known to crossreact with antibodies derived from other mammalian species.22,23 Porcine IgG and IgM detected in the plasma of baboons during cross-circulation may include antibodies against baboon RBC antigens. Therefore, baboon erythrocytes coated with porcine IgG or IgM would be predicted to fix baboon C1q, resulting in activation of the complement cascade in baboon serum, thus inducing cell lysis. However, the in vitro hemolytic assay used in this study showed that baboon erythrocytes incubated with porcine antibodies in normal porcine serum or baboon plasma samples showed a very mild degree of lysis when incubated with baboon complement. This suggested that the low level of porcine antibodies detected in the baboons’ plasma did not contribute significantly to hemolysis during cross-circulation. In addition, these data indicate that porcine serum does not contain a significant level of complement-fixing baboon RBC-binding antibodies. Therefore, hemolysis, which occurred during and after direct cross-circulation, was not humorally mediated by porcine antibodies, although there is still a possibility that immune complex-bound RBCs may be susceptible to reticuloendothelial systemic destruction in the baboon. However, there is another possible cause for the hemolysis. Perfusion of a porcine liver with xenogeneic blood may cause endothelial injury as a result of humoral immunologic reactions. Endothelial damage and dysfunction may then lead to the deposition of fibrin and formation of platelet thrombi.7,24 In case 1, fibrin deposition along endothelial surfaces was found in the perfused porcine liver graft. This fibrin might directly damage baboon erythrocytes, resulting in intravascular hemolysis.25 Conversely, fibrin deposition and intravascular thrombus formation was not detected in the livers from cases 2 to 6, although they showed evidence of mild humoral injury. The platelet count dramatically decreased immediately after connection to the ECLP system, indicating significant trapping of baboon platelets in the porcine liver, presumably as a result of endothelial cell injury and secondary coagulation. A similar drastic decrease in the platelet count was reported during the passage of human blood through an extracorporeally perfused porcine kidney, which resulted in thrombus formation occluding glomerular and peritubular capillaries.26 Severe platelet depletion and a rapid reduction in coagulation factors were also observed immediately after reperfusion of pig livers transplanted into baboons, which culminated in disseminated intravascular coagulation.27 It is reported

621

that loss of anticoagulant functions of the endothelium leads to platelet aggregation and subsequent fibrin formation, which may then lead to erythrocyte fragmentation or microhemolysis.25,28 Macchiarini et al15 reported severe hemolysis in relation to intravascular thrombi of fibrin and platelet in extracorporeal pig lung perfusion with human blood. Moreover, RBC fragmentation promotes thrombus formation.29 The macroscopic hemolysis observed during direct cross-circulation in case 1 therefore should be attributed to direct erythrocyte fragmentation resulting from humoral factor–mediated endothelial cell injury in the porcine liver graft. In some clinical ECLP trials between concordant species (baboon as a donor to human) or allo combination (human as a donor to human), liver grafts could be perfused for up to 72 hours6,8 and resulted in some improvements in the prognosis of patients with acute hepatic failure. To prolong the duration of cross-circulation between discordant species, it may be necessary to suppress humoral reactions, particularly those directed to the endothelial cell surface. Immunologic modulations, such as the use of liver grafts from transgenic pigs expressing human complement regulatory proteins on the endothelial cell surface or immunoadsorption of patients before cross-circulation, could be effective in maintaining critically important endothelial function of the isolated liver grafts.30,31 In addition, systemic inhibition of the complement cascade by the administration of soluble human complement-regulatory proteins may be another effective strategy to reduce humoral factor–mediated endothelial injury.24 Although both complement activity and absolute levels of individual complement components were maintained through the experiments in this study, complement activity in patients with acute hepatic failure may be low as a result of impaired synthesis of complement components and increased consumption of complements.7,32-35 Our ECLP system therefore might be maintained for a longer duration when clinically applied for the treatment of patients with acute hepatic failure.

Acknowledgment The authors thank Dr Chikuma Hamada of the Department of Pharmacoepidemiology of Kyoto University for concrete advice about statistical analysis, Dr Rei Takahashi of the Department of Pathology and Tumor Biology of Kyoto University for helpful discussions, and Dr Jeffrey L. Platt of Mayo Medical School for fruitful comments and reviews of the manuscript.

622

Nishitai et al

References 1. Otto JJ, Pender JC, Cleary JH, Sensenig DM, Welch S. The use of a donor liver in experimental animals with elevated blood ammonia. Surgery 1958;43:301-309. 2. Abouna GM, Kirkley JR, Hull CJ, Ashcroft T, Kerr DNS. Treatment of hepatic coma by extracorporeal pig-liver perfusion. Lancet 1969;1:64-68. 3. Eiseman B, Liem DS, Raffucci F. Heterologous liver perfusion in treatment of hepatic failure. Ann Surg 1965;162:329-345. 4. Gundermann KJ, Lie TS. Treatment of acute hepatic failure by hemoperfusion over baboon and human livers: Alphafetoprotein levels in serum. Artif Organs 1980;4:27-29. 5. Ozawa K, Kamiyama Y, Kimura K, Ukikusa M, Kono Y, Yamato T, et al. Clinical experience of postoperative hepatic failure treatment with pig or baboon liver cross-hemodialysis with an interposed membrane. Artif Organs 1982;6:433-446. 6. Abouna GM, Fisher LM, Still WJ, Hume DM. Acute hepatic coma successfully treated by extracorporeal baboon liver perfusions. BMJ 1972;1:23-25. 7. Collins BH, Chari RS, Magee JC, Harland RC, Lindman BJ, Logan JS, et al. Mechanisms of injury in porcine livers perfused with blood of patients with fulminant hepatic failure. Transplantation 1994;58:1162-1171. 8. Fox IJ, Langnas AN, Fristoe LW, Shaefer MS, Vogel JE, Antonson DL, et al. Successful application of extracorporeal liver perfusion: A technology whose time has come. Am J Gastroenterol 1993;88:1876-1881. 9. McLaughlin BE, Tosone CM, Custer LM, Mullon C. Overview of extracorporeal liver support systems and clinical results. Ann N Y Acad Sci 1999;875:310-325. 10. Terajima H, Shirakata Y, Yagi T, Mashima S, Shinohara H, Satoh S, et al. Long-duration xenogeneic extracorporeal pig liver perfusion with human blood. Transpl Int 1996;9(suppl 1):S388-391. 11. Terajima H, Shirakata Y, Yagi T, Mashima S, Shinohara H, Satoh S, et al. Successful long-term xenoperfusion of the pig liver: Continuous administration of prostaglandin E1 and insulin. Transplantation 1997;63:507-512. 12. Yagi T, Ikai I, Terajima H, Satoh S, Kanazawa A, Shinohara H, et al. The protective effects of prostaglandin E1 on sinusoidal endothelial cells in xenogeneic pig liver perfusion. J Surg Res 1997;73:28-34. 13. Yagi T, Satoh S, Kanazawa A, Uesugi T, Takeyama O, Yoneyama T, et al. Extracorporeal xenoperfusion of the pig liver as a liver assist device: Effects of prostaglandin E1 and soluble complement receptor type 1. Transplant Proc 1998;30:74-75. 14. Uesugi T, Ikai I, Yagi T, Satoh S, Kanazawa A, Yoneyama T, et al. Evaluation of ammonia and lidocaine clearance, and galactose elimination capacity of xenoperfused pig livers using a pharmacokinetic analysis. Transplantation 1999;68:209-214. 15. Macchiarini P, Mazmanian GM, Oriol R, de Montpreville V, Dulmet E, Fattal S, et al. Ex vivo lung model of pig-to-human hyperacute xenograft rejection. J Thorac Cardiovasc Surg 1997; 114:315-325. 16. Fiane AE, Videm V, Scholz T, Pedersen T, Karlsen H, Geiran O, et al. Ex vivo model for xenotransplantation. Transplant Proc 1995;27:3560-3561. 17. Buttery JE, Parbhoo SP, Baron DN. Plasma enzyme and isoenzyme changes during perfusion of the isolated pig liver. Clin Sci Mol Med 1974;46:105-112. 18. Moen O, Fosse E, Braten J, Andersson C, Hogasen K, Mollnes TE, et al. Differences in blood activation related to roller/centrifugal

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29. 30.

31.

32. 33.

34.

35.

pumps and heparin-coated/uncoated surfaces in a cardiopulmonary bypass model circuit. Perfusion 1996;11:113-123. Hansbro SD, Sharpe DA, Catchpole R, Welsh KR, Munsch CM, McGoldrick JP, Kay PH. Haemolysis during cardiopulmonary bypass: An in vivo comparison of standard roller pumps, nonocclusive roller pumps and centrifugal pumps. Perfusion 1999;14:3-10. Galili U, Shohet SB, Kobrin E, Stults CL, Macher BA. Man, apes, and Old World monkeys differ from other mammals in the expression of alpha-galactosyl epitopes on nucleated cells. J Biol Chem 1988;263:17755-17762. Fomufod AK, Castro O, Slaughter LJ, Cothran LN, Hayes NR, Africano E. Massive sequestration of human sickle cells after transfusion to a baboon. J Med Primatol 1986;15:71-79. Kitamura H. Interspecies incompatibilities of complement factors and regulators. In: Rother K, Till GO, Hansch GM, eds. The Complement System, 2nd rev. ed. Heidelberg, Germany: Springer-Verlag; 1998:179-188. Tanaka S, Kitamura F, Suzuki T. Studies on the hemolytic activity of the classical and alternative pathway of complement in various animal species. Complement 1987;4:33-41. Satoh S, Terajima H, Yagi T, Kanazawa A, Shinohara H, Gomi T, et al. Humoral injury in porcine livers perfused with human whole blood. Transplantation 1997;64:1117-1123. Hattori A, Watanabe T, Izumi T. Scanning electron microscope study on hemostatic reaction. Mural thrombus after the removal of endothelium, with special references to platelet behavior, site of fibrin formation and microhemolysis. Arch Histol Jpn 1978; 41:205-227. Breimer ME, Bjorck S, Svalander CT, Bengtsson A, Rydberg L, Lie-Karlsen K, et al. Extracorporeal (“ex vivo”) connection of pig kidneys to humans. I. Clinical data and studies of platelet destruction. Xenotransplantation 1996;3:328-339. Majado MJ, Ramirez P, Minguela A, Hernandez Q, Gonzalez C, Loba M, et al. Evolution of blood coagulation factors and hemotherapeutic support in three pig-to-baboon orthotopic liver xenotransplants. Transplant Proc 1999;31:2622-2624. Dale J, Ohlsson K, Nordstoga K, Aasen AO. Intravascular hemolysis and ultrastructural changes of erythrocytes in lethal canine endotoxin shock. Eur Surg Res 1980;12:39-51. Hardaway RMD. Mechanism of traumatic shock. Surg Gynecol Obstet 1980;151:65-69. Pascher A, Poehlein C, Stangl M, Hoebel G, Thiery J, MuellerDerlich J, Hammer C. Application of immunoapheresis for delaying hyperacute rejection during isolated xenogeneic pig liver perfusion. Transplantation 1997;63:867-875. Pascher A, Poehlein C, Storck M, Prestel R, Mueller-Hoecker J, White DJ, et al. Immunopathological observations after xenogeneic liver perfusions using donor pigs transgenic for human decay-accelerating factor. Transplantation 1997;64:384-391. De la Tour B, Kazatchkine M. The complement system in hepatopathies. Ann Med Interne (Paris) 1985;136:222-224. Ellison RTD, Mason SR, Kohler PF, Curd JG, Reller LB. Meningococcemia and acquired complement deficiency. Association in patients with hepatic failure. Arch Intern Med 1986; 146:1539-1540. Irshad M, Acharya SK, Joshi YK, Tandon BN. Role of fibronectin and complement in immunopathogenesis of acute and subacute hepatic failure. J Gastroenterol Hepatol 1994;9:355-360. Larcher VF, Wyke RJ, Mowat AP, Williams R. Bacterial and fungal infection in children with fulminant hepatic failure: Possible role of opsonisation and complement deficiency. Gut 1982; 23:1037-1043.