Intra-arterial delivery of a recombinant adenovirus does not increase gene transfer to tumor cells in a rat model of metastatic colorectal carcinoma.

Hepatic artery infusion of adenoviral vectors has been shown to increase transduction of certain hepatocellular malignancies in preclinical studies. In addition, clinical trials have begun evaluating the efficacy of gene transfer of cytotoxic genes to metastatic colorectal tumors through hepatic artery infusion. Here we evaluate the extent of gene expression and therapeutic effect following various routes of administration of recombinant adenovirus in a rat model of metastatic colorectal carcinoma. We administered adenovirus (AdCMVlacZ) to rats with established colorectal metastases through infusion into the hepatic artery, intravenous infusion, or direct injection into a tumor. Intravenous administration resulted in transduction of hepatocytes, but not tumor cells. Hepatic arterial administration failed to substantially increase transduction of tumor cells. In addition, ligation of the hepatic artery following infusion of adenovirus or the addition of lipiodol infusion had no effect on the transduction of tumor cells. We administered AdCMVp53 by direct injection into tumors, intravenous administration, or hepatic artery infusion to evaluate the delivery of a therapeutic gene. Direct injection of AdCMVp53 into established hepatic colorectal metastases resulted in a therapeutic response in comparison with both hepatic arterial and intravenous infusion of vector. These preclinical studies fail to support a strategy of infusion through the hepatic artery of recombinant adenovirus targeting tumor cells in the treatment of colorectal cancer liver metastases.

Systemic IFN-beta gene therapy results in long-term survival in mice with established colorectal liver metastases.

Most patients succumbing to colorectal cancer fail with liver-predominant metastases. To make a clinical impact in this disease, a systemic or whole-liver therapy may be required, whereas most cancer gene therapy approaches are limited in their ability to treat beyond local disease. As a preclinical model for cancer gene therapy, recombinant adenovirus containing the human IFN-beta (hIFN-beta) cDNA was delivered systemically in nude mouse xenograft models of human colorectal cancer liver metastases. The vector targeted hepatocytes that produced high levels of hIFN-beta in the liver, resulting in a profound apoptotic response in the tumors and significant tumor regression. hIFN-beta gene therapy not only resulted in improved survival and long-term cure in a micrometastatic model, but provided similar benefits in a clinically relevant gross disease model. A similar recombinant adenovirus containing the murine IFN-beta (mIFN-beta) cDNA also resulted in a therapeutic response and improved survival in syngeneic mouse models of colorectal cancer liver metastases. Depletion studies demonstrate a contribution of natural killer cells to this therapeutic response. The toxicity of an adenoviral vector expressing murine IFN-beta in a syngeneic model is also presented. These encouraging results warrant further investigation of the use of cancer gene therapy for targeting metastatic disease.

The caspase 9 inhibitor Z-LEHD-FMK protects human liver cells while permitting death of cancer cells exposed to tumor necrosis factor-related apoptosis-inducing ligand.

Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) is a potent inducer of apoptosis of transformed and cancer cells but not of most normal cells. Recent studies have revealed an unforeseen toxicity of TRAIL toward normal human hepatocytes, thereby bringing into question the safety of systemic administration of TRAIL in humans with cancer. We found that SW480 colon adenocarcinoma, or H460 non-small cell lung cancer cell lines, which are sensitive to TRAIL, were not protected by the caspase 9 inhibitor Z-LEHD-FMK from TRAIL-induced apoptosis. However, a human colon cancer cell line HCT116 and a human embryonic kidney cell line 293, which are sensitive to TRAIL, were protected by Z-LEHD-FMK from TRAIL-mediated death. Both HCT116 and SW480 cells were protected from TRAIL by the caspase 8 inhibitor Z-IETD-FMK, dominant-negative FADD and cellular FLIP-s and interestingly both cell lines displayed caspase 9 cleavage to a similar extent after TRAIL exposure. We confirmed that normal human liver cells are sensitive to TRAIL. Moreover, we found that normal human liver cells could be protected from TRAIL-induced apoptosis by simultaneous exposure to Z-LEHD-FMK. A similar brief exposure to TRAIL plus Z-LEHD-FMK inhibited colony growth of SW480 but not HCT116 cells. Because some cancer cell lines are not protected from TRAIL-mediated killing by Z-LEHD-FMK, we believe that a brief period of caspase 9 inhibition during TRAIL administration may widen the therapeutic window and allow cancer cell killing while protecting normal liver cells. This strategy could be further developed in the effort to advance TRAIL into clinical trials.

In vivo selection of hepatocytes transduced with adeno-associated viral vectors.

A murine model for hereditary tyrosinemia Type I (HTI) was evaluated for in vivo gene therapy with adeno-associated viral (AAV) vectors expressing the enzyme fumarylacetoacetate hydrolase. Transduction of a limited number of hepatocytes was accomplished following infusion of vector into the portal circulation. Corrected hepatocytes were expanded in vivo by withdrawing a drug which prevents the accumulation of toxic metabolites. The liver was eventually repopulated with hepatocytes harboring a functional and apparently integrated AAV provirus. Recipient animals regained normal liver function and architecture and the underlying metabolic derangements were normalized. After 9 months, vector-treated animals showed benign hepatomas, whereas in untreated animals areas of marked dysplasia were present within hepatomas.

Augmentation of innate host defense by expression of a cathelicidin antimicrobial peptide.

Antimicrobial peptides, such as defensins or cathelicidins, are effector substances of the innate immune system and are thought to have antimicrobial properties that contribute to host defense. The evidence that vertebrate antimicrobial peptides contribute to innate immunity in vivo is based on their expression pattern and in vitro activity against microorganisms. The goal of this study was to investigate whether the overexpression of an antimicrobial peptide results in augmented protection against bacterial infection. C57BL/6 mice were given an adenovirus vector containing the cDNA for LL-37/hCAP-18, a human cathelicidin antimicrobial peptide. Mice treated with intratracheal LL-37/hCAP-18 vector had a lower bacterial load and a smaller inflammatory response than did untreated mice following pulmonary challenge with Pseudomonas aeruginosa PAO1. Systemic expression of LL-37/hCAP-18 after intravenous injection of recombinant adenovirus resulted in improved survival rates following intravenous injection of lipopolysaccharide with galactosamine or Escherichia coli CP9. In conclusion, the data demonstrate that expression of an antimicrobial peptide by gene transfer results in augmentation of the innate immune response, providing support for the hypothesis that vertebrate antimicrobial peptides protect against microorganisms in vivo.

Fas-Fas ligand interactions play a major role in effector functions of cytotoxic T lymphocytes after adenovirus vector-mediated gene transfer.

Adenovirus vectors transduce liver hepatocytes with extreme efficiency; however, transgene expression is eliminated within 2 weeks. Extinction of transgene expression has been attributed to infiltrating cytotoxic T lymphocytes (CTLs) in the liver in a process that resembles a number of human diseases, including viral and autoimmune hepatitis. In this study we investigated the role of Fas-Fas ligand interactions in killing of vector-transduced hepatocytes in vitro and in vivo. Intrahepatic lymphocytes (IHLs) isolated from livers of mice administered adenovirus vector demonstrated cytolytic activity against vector-infected primary hepatocytes. The in vitro CTL activity of the IHLs involving both CD4+ and CD8+ T cells was MHC class I restricted and could be blocked by soluble Fas-IgG. Adoptive transfer of IHLs from immune-competent mice immunized with Ad-lacZ into Ragl-deficient mice previously infused with Ad-lacZ resulted in rapid elimination of beta-galactosidase-transduced hepatocytes. Transfer of these cells into Fas-deficient mice (B6-lpr) failed to eliminate lacZ expression; likewise IHLs from immunized FasL-deficient mice (B6-gld) failed to eliminate lacZ expression in Rag1-deficient mice. Finally, in vivo administration of soluble Fas-IgG abrogated the ability of Ad-lacZ-primed IHLs to eliminate transgene expression. These studies establish an essential role for Fas-Fas ligand interactions in the mechanism of elimination of adenoviral vector-mediated transgene expression in the liver.

Adeno-associated virus as a vector for liver-directed gene therapy.

Factors relevant to the successful application of adeno-associated virus (AAV) vectors for liver-directed gene therapy were evaluated. Vectors with different promoters driving expression of human alpha-1-antitrypsin (alpha-1AT) were injected into the portal circulation of immunodeficient mice. alpha-1AT expression was stable but dependent on the promoter. Southern analysis of liver DNA revealed approximately 0.1 to 2.0 provirus copies/diploid genome in presumed head-to-tail concatamers. In situ hybridization and immunohistochemical analysis revealed expression in approximately 5% of hepatocytes clustered in the pericentral region. These results support the use of AAV as a vector for diseases treatable by targeting of hepatocytes.

In vitro morphological and functional characterization of isolated porcine hepatocytes for extracorporeal liver support: bile acid uptake and conjugation.

Recently, researchers have focused on the use of bioartificial liver (BAL) to support patients with fulminant hepatic failure (FHF). We have developed a cell-based BAL, consisting of porcine hepatocytes in a hollow-fiber bioreactor. To better characterize BAL metabolic functions in vitro, bioreactors were inoculated with 48-h-cultured, microcarrier-attached hepatocytes and perifused with recirculating human plasma that contained either 1 microCi of [24-14C] plasma-enriched cholate or 1 microCi of [24-14C] plasma-enriched taurocholate. Bile acids were sampled hourly and separated into four fractions (unconjugated, glycoconjugated, tauroconjugated, and sulfated) for radioactivity determination. Following 3 h perifusion, the glycoconjugated and sulfated bile acid fractions in the bioreactor extrafiber space were significantly elevated when compared to the recirculating plasma. During perifusion with taurocholate-enriched plasma, a relative decrease in the tauroconjugated fraction and an increase in the glycoconjugated fraction were observed. Cholate was accumulated by hepatocytes to a level threefold lower than taurocholate; however, a significant proportion of radioactivity (<25%) was detected in the glycoconjugated fraction. Ultrastructural examination of microcarrier-attached hepatocytes illustrated that the features typical of metabolically active liver cells were maintained. Our data demonstrate the ability of BAL to clear bile acids from the circulation, to accumulate cholate and taurocholate, and to conjugate a substantial amount of cholic acid.

Long-term correction of albumin levels in the Nagase analbuminemic rat: repopulation of the liver by transplanted normal hepatocytes under a regeneration response.

Numerous studies have reported successful transplantation of hepatocytes with demonstration of function. However, none have shown long-term correction of a liver-related metabolic defect. Male Nagase analbuminemic rats, immunosuppressed with cyclosporin-A, were transplanted with normal hepatocytes (2 x 10(7) cells/rat) isolated from allogeneic male Sprague-Dawley rat donors. Hepatocytes were selectively transplanted via the portal vein tributary into the posterior liver lobes of Nagase analbuminemic rats. Following 2 wk, to allow engraftment, selected transplanted rats (Group I) were reoperated and the portal venous branch supplying the anterior liver lobes was permanently ligated, resulting in their atrophy and induction of regeneration in the residual transplant-bearing lobes. Control rats consisted of: Group II-transplanted with normal hepatocytes without portal branch ligation; Group III-transplanted with analbuminemic hepatocytes with portal branch ligation; and Group IV-nontransplanted analbuminemic rats with portal branch ligation. The experimental period extended to 3 mo posttransplantation. All rats transplanted with normal hepatocytes demonstrated a significant elevation in serum albumin levels (ELISA). Group I rats had dramatic elevations in serum albumin to near normal levels (1.78 +/- 0.20 g/dl), and maintained these levels until the end of the experiment. Albumin levels in Group II rats reached 0.26 +/- 0.07 g/dl (p < 0.001), whereas Group III and IV rats showed no changes in serum albumin levels throughout the experiment. Immunohistology of liver tissue obtained from Group I rats, demonstrated large numbers (22.6 +/- 7.5%) of albumin-positive hepatocytes populating the recipient liver. This is the first report of near-total and sustained correction of a genetic defect in liver function in an experimental animal model following allogeneic hepatocyte transplantation.

Automated liver cell processing facilitates large scale isolation and purification of porcine hepatocytes.

An automated method for large scale isolation and purification of porcine hepatocytes is described. Liver cells were harvested by a two-step portal vein perfusion with ethylenediaminetetraacetate and collagenase. Hepatocyte purification was carried out using either a standard manual processing method (Procedure A) or an automated processing method using a filtration chamber and a programmable cell washer (Procedure B). Both methods produced high cell yields (Procedure A: 1.30 +/- 0.55 x 10(10) viable hepatocytes/liver; Procedure B: 1.38 +/- 0.32 x 10(10) viable hepatocytes/liver) and viability (Procedure A: 89 +/- 6.5%; Procedure B: 92 +/- 3.9%). Hepatocyte purity was significantly greater after Procedure B than after Procedure A (93.1 +/- 3.1% versus 83.1 +/- 3%, p < 0.01). Isolated hepatocytes by either method were morphologically intact, as demonstrated by transmission electron microscopy showing integrity of plasma membranes and intracellular organelles. Cultured hepatocytes isolated by either method were functionally intact, although those isolated by Procedure A showed significantly lower activity of microsomal 7-ethoxycoumarin-O-deethylase activity (p < 0.05) and mitochondrial succinate dehydrogenase activity (p < 0.01). In conclusion, use of the automated hepatocyte processing method resulted in efficient large scale preparation of porcine hepatocytes, with higher purity and greater retention of differentiated liver metabolic functions, and was found to be less time consuming and less labor intensive.

Differential patterns of reaction of human natural antibodies to pig hepatocytes and vascular endothelium.

We have recently conducted a series of experiments to characterize the pattern of reaction of human natural antibodies (NA) with individual pig liver cells. Pooled normal human serum (PHS) was incubated with cultured pig hepatocytes (HEP), aortic endothelial cells (AEC), and portal endothelial cells (PEC), and the reaction of NA to different cell types was measured by antibody-mediated cytotoxic (MTT assay), antibody binding (ELISA), and flow cytometric analysis. The human NA displayed a differential pattern of binding with hepatocytes exhibiting a more limited expression of xenoantigen expression than either aortic or portal endothelial cells. These differences in reaction patterns were also noted for Western blot analysis of individual cell membrane extracts. Preincubation of the pig cells with anti-pig MHC antibodies did not inhibit the binding of human IgM natural antibodies to the pig cells. Comparison of the pattern of NA absorption following the use of bioartificial liver support in patients with acute hepatic failure demonstrated limited ability of pig hepatocytes to absorb substantial amounts of NA. These studies indicate that pig hepatocytes are less vulnerable to NA cytotoxicity than pig vascular endothelial cells and that pig vascular endothelial cells express xenoantigens that are unique and not found on hepatocytes.

Repeated intraportal hepatocyte transplantation in analbuminemic rats.

The optimal site for implantation of isolated hepatocytes has not been established. We have developed a novel technique which allows repeated infusion of hepatocytes into the portal system via an indwelling catheter. Seven Nagase Analbuminemic rats (NAR) underwent single intraportal infusion of 2 x 10(7) isolated normal albumin-producing rat hepatocytes. Another seven NAR rats underwent placement of indwelling catheters into the portal venous system via the gastroduodenal vein. Each of them received six batches of 5 x 10(6) normal albumin producing hepatocytes. Seven control NAR rats were infused repeatedly (intraportally) with saline only. Plasma albumin (ELISA) showed significant increase in experimental animals and was more pronounced (p < 0.05) in rats transplanted repeatedly than in those given a single dose of cells. Immunohistochemical staining of the liver sections confirmed the presence of transplanted albumin producing hepatocytes. Rats transplanted with a single large batch of isolated hepatocytes showed liver tissue damage, whereas those subjected to repeated cell infusions had normal liver histology. We have developed a novel intraportal transplantation method which allows successful engraftment of a large number of isolated hepatocytes.

Early clinical experience with a hybrid bioartificial liver.

BACKGROUND: Severe liver failure is associated with high mortality. Orthotopic liver transplantation (OLT) is the only effective therapeutic modality; there is a need for a 'bridge' system to support patients until an organ becomes available. METHODS: A bioartificial liver (BAL) was used to treat 10 patients with severe liver failure. A plasmapheresis system was used to pump patient plasma through a module with porcine hepatocytes. Each treatment lasted 6-7 h. RESULTS: All patients tolerated the procedure(s) well. Eight patients underwent OLT following BAL treatment(s). There were two late deaths after recovery from liver failure. Five patients with increased intracranial pressure (ICP) and decerebration had ICP normalization, increased cerebral perfusion pressure and full neurologic recovery after OLT. There was improvement in the level of encephalopathy and a significant decrease in serum ammonia after BAL treatment(s). CONCLUSIONS: BAL treatment is safe and beneficial and can be successfully used as a 'bridge' to transplantation.

Carboxyfluorescein (CFSE) labelling of hepatocytes for short-term localization following intraportal transplantation.

Renewed interest in the transplantation of isolated hepatocytes into the liver as a potential therapy for liver disease has stimulated the development of methods for the identification of donor cells within the recipient organ. We describe a method for cellular tagging and in vivo identification of intraportally transplanted hepatocytes using an intracellular fluorescent dye, 5(6)-carboxyfluorescein diacetate, succinimidyl-ester (CFSE). Rat and porcine hepatocytes were isolated and labelled with CFSE. The optimal conditions for labelling consisted of a buffered saline suspension of hepatocytes (5 x 10(6) cells/mL) in 20.0 microM CFSE incubated for 15 min at 37 degrees C. In vitro, labelled hepatocytes were cultured either on fibronectin-coated chamber slides or in culture flasks. Cultures were evaluated in situ by fluorescence photomicrography or by fluorescence-activated cell sorting (FACS) after cell detachment. Cell viability was assessed serially and cultured, labelled hepatocytes retained the dye for up to 3 wk (last day of study). CFSE did not effect hepatocyte viability and there was no evidence of intercellular diffusion of the dye. In vivo, syngeneic Lewis rats underwent selective portal vein infusion of freshly isolated, labelled hepatocytes (2.0 x 10(7) cells/2.0 mL saline/animal) into the posterior liver lobes. All recipients were sacrificed 48 h and 96 h later and their livers examined. Transplanted hepatocytes were identified by fluorescence microscopy in tissue sections and by FACS following collagenase digestion of the liver tissue. CFSE persisted in a population of viable, engrafted hepatocytes. FACS analysis demonstrated that 9 +/- 3% of the hepatocytes in the posterior liver lobes were labelled 48 and 96 h after transplantation. At 96 h following transplantation, multiple engrafted hepatocytes could be observed by fluorescence microscopy around the central veins. CFSE labelling allows for both in vitro identification and in vivo localization of donor hepatocytes. Furthermore, it appears to be more stable and specific for labelling hepatocytes than other tested dyes (especially DiI).

A bioartificial liver to treat severe acute liver failure.

OBJECTIVE: To test the safety and efficacy of a bioartificial liver support system in patients with severe acute liver failure. SUMMARY BACKGROUND DATA: The authors developed a bioartificial liver using porcine hepatocytes. The system was tested in vitro and shown to have differentiated liver functions (cytochrome P450 activity, synthesis of liver-specific proteins, bilirubin synthesis, and conjugation). When tested in vivo in experimental animals with liver failure, it gave substantial metabolic and hemodynamic support. METHODS: Seven patients with severe acute liver failure received a double lumen catheter in the saphenous vein; blood was removed, plasma was separated and perfused through a cartridge containing 4 to 6 x 10(9) porcine hepatocytes, and plasma and blood cells were reconstituted and reinfused. Each treatment lasted 6 to 7 hours. RESULTS: All patients tolerated the procedure(s) well, with neurologic improvement, decreased intracranial pressure (23.0 +/- 2.3 to 7.8 +/- 1.7 mm Hg; p < 0.005) associated with an increase in cerebral perfusion pressure, decreased plasma ammonia (163.3 +/- 21.3 to 112.2 +/- 9.8 microMoles/L; p < 0.01), and increased encephalopathy index (0.60 +/- 0.17 to 1.24 +/- 0.22; p < 0.03). All patients survived, had a liver transplant, and were discharged from the hospital. CONCLUSIONS: This bioartificial liver is safe and serves as an effective "bridge" to liver transplant in some patients.

Isolated hepatocytes in a bioartificial liver: A single group view and experience.

Despite recent advances in medical supportive therapy, patients with severe fulminant hepatic failure (FHF) have mortality rate approaching 90%. Investigators have attempted to improve survival by using various extracorporeal liver support systems loaded with sorbents and liver tissue preparations. None of them succeeded in gaining clinical acceptance and orthotopic liver transplantation (OLT) remains a primary therapeutic option for patients with FHF. In this study, authors discuss the systems which utilize isolated hepatocytes. Most of these devices were tested in vitro and in animals with chemically and surgically induced liver failure. In some studies, signficant levels of detoxification and liver functions were achieved. The authors describe their own hepatocyte-based artificial liver (BAL). It is based on plasma perfusion through a hollow-fiber module seeded with matrix-anchored porcine hepatocytes. The BAL was used 14 times to treat 9 patients with acute liver failure. On 10 occasions, a charcoal column was included in the plasma circuit. Each treatment lasted 7 +/- 1 h. All procedures were tolerated well and 8 patients (including 6 patients with FHF) underwent OLT. Five patients with increased intracranial pressure (ICP) and evidence of decerebration had normalization of ICP and enjoyed full neurologic recovery after OLT. Laboratory data showed evidence for bilirubin conjugation, decrease in blood ammonia, maintenance of low lactic acid levels, and increase in the ration between the branched chain and aromatic amino acids. No allergic reactions to xenogeneic hepatocytes were observed. The authors conclude that BAL treatment with porcine hepatocytes appears to be safe and can help maintain patients alive and neurologically intact until a liver becomes available for transplantation. (c) 1994 John Wiley & Sons, Inc.

Mass transfer in a hollow fiber device used as a bioartificial liver.

The transport of albumin, glucose and fluid in a hollow fiber bioartificial liver (BAL) was predicted by theory and measured experimentally. Results from the experiment were used in a three compartment transport model to estimate the transfiber fluid flux. Fluid convection driven by the transfiber pressure gradient transported solutes across the semi-permeable fibers of the BAL. Diffusion contributed significantly to the transfiber transport of both glucose and albumin. Modification of the BAL system hydraulic geometry significantly increased transfiber pressure gradient with corresponding increase in transfiber fluid flux. Transfiber solute transport was increased by osmotically driven transport which resulted from the revised flow geometry. Substantial delivery of hepatocyte synthesis products is possible with careful control of the physical parameters and operating conditions of the BAL.

Fibroblast transplantation in rats: transduction and function of foreign genes.

The purpose of this study was to develop a method of transduction of normal skin fibroblasts with a retroviral vector expressing the human factor IX (hF-IX) gene and transplantation of the transduced cells into rats. A retroviral vector containing the hF-IX gene, as well as a selectable marker (hygromycin B resistance; HB), was used to transduce cultured normal skin (rat and human) fibroblasts. Transduced fibroblasts were selected with HB; resistant clones were propagated and assessed for expression of active hF-IX in vitro. Transduced cells secreted significant amounts of hF-IX in vitro (327.9 +/- 14.2 ng/10(6) cells) which showed measurable clotting activity, indicating terminal processing of the gene-transcribed hF-IX product by the transduced fibroblasts. Integration of the proviral gene into rat and human fibroblast genomic DNA was confirmed by polymerase chain reaction. Transduced rat fibroblasts were attached to collagen-coated microcarriers and transplanted intraperitoneally into syngeneic recipients. Transplanted microcarrier-attached cells formed aggregates in the peritoneal cavity and exhibited positive immunohistochemical staining for hF-IX up to 8 weeks following transplantation, the time of termination of the experiment. These experimental results may have significant implications in developing strategies for future clinical therapy for hemophilia B.

Selective intraportal hepatocyte transplantation in analbuminemic and Gunn rats.

Although significant progress has been achieved in isolated hepatocyte transplantation, the optimal site of cell implantation has not yet been determined. We have developed a novel experimental method of intraportal hepatocyte transplantation that allows easy assessment of the morphology and function of transplanted hepatocytes. Donor hepatocytes were harvested from Sprague-Dawley rats by in situ EDTA/collagenase perfusion. Fifteen recipient Nagase analbuminemic rats (NAR) underwent cannulation of the gastroduodenal vein under ether anesthesia. Either the posterior or anterior liver lobes were selectively infused with cells by occluding the portal venous supply of the nontransplanted liver lobes. Normal donor hepatocytes (2 x 10(7)) suspended in normal saline were infused over 1 min (4 ml). Recipients were treated with cyclosporine for the duration of the experiment. Plasma albumin levels were determined by ELISA, before and at various intervals after transplantation. In NAR rats transplanted with normal hepatocytes, there was a significant (P < 0.003) and sustained (12 weeks) increase in plasma albumin levels. Control NAR rats transplanted with NAR hepatocytes (n = 8) showed no significant changes in plasma albumin levels. Similarly, normal Wistar hepatocytes were infused intraportally into the posterior lobes of Gunn rats (n = 4), which lack the ability to conjugate bilirubin. Pre- and posttransplantation bile was collected following bile duct cannulation. Bile analysis by HPLC, demonstrated a significant (P = 0.04) increase in the level of bilirubin conjugates following transplantation and a corresponding decrease in total serum bilirubin (P = 0.04). Our experimental data demonstrate that direct selective intraportal infusion of hepatocytes is an effective technique of hepatocyte transplantation in the rat.