Clinical and laboratory evaluation of the safety of a bioartificial liver assist device for potential transmission of porcine endogenous retrovirus.

BACKGROUND: The potential risk of transmission of porcine endogenous retroviruses (PERV) from xenogeneic donors into humans has been widely debated. Because we were involved in a phase I/II clinical trial using a bioartificial liver support system (BLSS), we proceeded to evaluate the biosafety of this device. MATERIALS AND METHODS: The system being evaluated contains primary porcine hepatocytes freshly isolated from pathogen-free, purpose-raised herd. Isolated hepatocytes were installed in the shell, which is separated by a semipermeable membrane (100-kD nominal cutoff) from the lumen through which the patients' whole blood is circulated. Both before and at defined intervals posthemoperfusion, patients' blood was obtained for screening. Additionally, effluent collected from a clinical bioreactor was analyzed. The presence of viral particles was estimated by reverse transcriptase-polymerase chain reaction (RT-PCR) and RT assays. For the detection of pig genomic and mitochondrial DNA, sequence-specific PCR (SS-PCR) was used. Finally, the presence of infectious viral particles in the samples was ascertained by exposure to the PERV-susceptible human cell line HEK-293. RESULTS: PERV transcripts, RT activity, and infectious PERV particles were not detected in the luminal effluent of a bioreactor. Culture supernatant from untreated control or mitogen-treated porcine hepatocytes (cleared of cellular debris) also failed to infect HEK-293 cell lines. Finally, RT-PCR, SS-PCR, and PERV-specific RT assay detected no PERV infection in the blood samples obtained from five study patients both before and at various times post-hemoperfusion. CONCLUSION: Although longer patient follow-up is required and mandated to unequivocally establish the biosafety of this device and related bioartificial organ systems, these analyses support the conclusion that when used under standard operational conditions, the BLSS is safe.

Advances in bioartificial liver assist devices.

Rapid advances in development of bioartificial liver assist devices (BLADs) are exciting clinical interest in the application of BLAD technology for support of patients with acute liver failure. Four devices (Circe Biomedical HepatAssist, Vitagen ELAD, Gerlach BELS, and Excorp Medical BLSS) that rely on hepatocytes cultured in hollow-fiber membrane technology are currently in various stages of clinical evaluation. Several alternative approaches for culture and perfusion of hepatocytes have been evaluated in preclinical, large animal models of liver failure, or at a laboratory scale. Engineering design issues with respect to xenotransplantation, BLAD perfusion, hepatocyte functionality and culture maintenance, and ultimate distribution of a BLAD to a clinical site are delineated.

Novel bioartificial liver support system: preclinical evaluation.

Preclinical safety and efficacy evaluation of a novel bioartificial liver support system (BLSS) was conducted using a D-galactosamine canine liver failure model. The BLSS houses a suspension of porcine hepatocytes in a hollow fiber cartridge with the hepatocytes on one side of the membrane and whole blood flowing on the other. Porcine hepatocytes harvested by a collagenase digestion technique were infused into the hollow fiber cartridge and incubated for 16 to 24 hours prior to use. Fifteen purpose-bred male hounds, 1-3 years old, 25-30 kg, were administered a lethal dose, 1.5 g/kg, of D-galactosamine. The animals were divided into three treatment groups: (1b) no BLSS treatment (n = 6); (2b) BLSS treatment starting at 24-26 h post D-galactosamine (n = 5); and (2c) BLSS treatment starting at 16-18 h post D-galactosamine (n = 4). While maintained under isoflurane anesthesia, canine supportive care was guided by electrolyte and invasive physiologic monitoring consisting of arterial pressure, central venous pressure, extradural intracranial pressure (ICP), pulmonary artery pressure, urinary catheter, and end-tidal CO2. All animals were treated until death or death-equivalent (inability to sustain systolic blood pressure > 80 mmHg for 20 minutes despite massive fluid resuscitation and/or dopamine administration), or euthanized at 60 hours. All animals developed evidence of liver failure at 12-24 hours as evidenced by blood pressure lability, elevated ICP, marked hepatocellular enzyme elevation with microscopic massive hepatocyte necrosis and cerebral edema, elevated prothrombin time, and metabolic acidosis. Groups 2b and 2c marginally prolong survival compared with Group 1b (pairwise log rank censored survival time analysis, p = 0.096 and p = 0.064, respectively). Since survival times for Groups 2b and 2c are not significantly different (p = 0.694), the groups were combined for further statistical analysis. Survival times for the combined active treatment Groups 2b and 2c are significantly prolonged versus Group 1b (p = 0.047). These results suggest the novel BLSS reported here can have a significant impact on the course of liver failure in the D-galactosamine canine liver failure model. The BLSS is ready for Phase I safety evaluation in a clinical setting.

Voltage polarity relay--optimal control of electrochemical urea oxidation.

Voltage polarity relay (VPR) is shown to optimize the urea oxidation rate and urea current utilization under constant current conditions in direct electrochemical urea oxidation. Direct electrochemical urea oxidation is characterized by reversible deactivation of the working electrode due to oxidation products remaining on the surface and the requirement that the working electrode potential remain below about 1.1 V relative to Ag/AgCl in order to prevent undesirable secondary electrochemical oxidations. The VPR method monitors the potential of the working electrode relative to a suitable reference and changes system polarity when the upper potential set limit is reached. Thus, what was the working electrode becomes the counter electrode and vice versa. Since urea oxidation products are desorbed from the counter electrode when its potential drops below about -0.6 V relative to Ag/AgCl, alternating electrode functions between working and counter provides cyclic electrode regeneration and continuous urea oxidation. VPR is believed to optimize constant current control for any electrochemical system that exhibits behavior similar to direct electrochemical urea oxidation.

Ammonia transport across hydrophobic membranes. Application to dialysate regeneration.

Removal of ammonia from a recirculating dialysate buffer in a portable hemodialysis application can be achieved by countercurrent, gas phase ammonia transfer across a hydrophobic membrane into an acid solution. Ammonia transfer fluxes as high as 0.076 mumol/sec/m2 have been achieved using a Sarns Turbo Membrane Oxygenator (Sarns-3M, Ann Arbor, MI) with a 1.9 m2 membrane surface area (0.145 mumol/sec actual rate). A simple physical model based upon ammonia desorption at the gas-dialysate buffer interface in a membrane pore, ammonia diffusion through the gas filled pore, and subsequent ammonia absorption at the gas/acid interface side of the pore quantitatively describes the experimental data. The ammonia transfer rate is most dependent upon dialysate buffer pH (higher pH promoting transfer rate) and ammonia concentration in the dialysate buffer (higher concentrations promoting transfer rate). A 500 fold improvement in transfer rate, however, will be required for clinical application.

Zeolitic ammonium ion exchange for portable hemodialysis dialysate regeneration.

Ammonia removal from a recirculating dialysate stream is a major challenge in developing a truly portable, regenerable hemodialysis system. Three zeolites, type F, type W, and clinoptilolite, were found to have good ammonia ion exchange capacity with linear equilibrium ion exchange coefficients of 0.908, 0.488, and 0.075 L/g, respectively. The linear equilibrium ion exchange coefficient relates dialysate ammonia concentration (mumol/L) to the amount of ammonia absorbed by zeolite (mumol/g) at equilibrium. Ammonia uptake by zeolite powders was fast, with equilibrium reached within 15 sec. Zeolite ammonia ion exchange and regeneration through multiple cycles was studied using an ion exchange column containing clinoptilolite pellets. Zeolite ion exchange capability was regenerated by flushing the column with 2 mol/L sodium chloride after an ion exchange run. The column maintained ammonia ion exchange capacity through six ion exchange/regeneration cycles, demonstrating multiple dialysis use possibilities. Atomic absorption spectroscopy of the column effluent showed no detectible (< 1 part per million) Si or Al leached from the zeolite.

Interference of glucose sensing by amino acids.

Interference by membrane permeable substances on nonspecific electrodes is a major problem in glucose sensing. Alanine, lysine, phenylalanine, and cystine were chosen for study to gain insight into this problem. These compounds represent the classes of mono-amino aliphatic, di-amino aliphatic, aromatic, and sulfur containing amino acids, respectively. Cyclic voltammetry experiments were performed using a Pt electrode (1.77 mm2). The reductive current of glucose at -0.750 V versus Ag/AgCl was measured with increasing concentrations of interfering substances in Krebs-Ringer phosphate buffer (pH 7.4) at 37 degrees C. Experimental results have shown that these amino acids have an inhibitory effect on the glucose signal. An important finding was that the interferences from phenylalanine and cystine were more pronounced than those of lysine and alanine. An initial drop in the glucose signal was seen at less than 2.0 mg/dl of alanine or lysine and at less than 0.5 mg/dl of phenylalanine or cystine. Additional increase in the concentrations of interfering substance did not cause further appreciable signal reduction. The results confirm that glucose sensing using a non-specific electrode is possible in fluids containing interfering substances such as amino acids.

A microchip glucose sensor.

A major problem in development of a glucose sensor for use in an implantable artificial pancreas is the lack of reproducibility in signals from sensor to sensor. Each glucose sensor fabricated with currently used methods has a unique response to varying levels of glucose concentration and thus needs to be individually calibrated before use. We have adapted microchip manufacturing techniques for the fabrication of electrochemically based glucose sensors with standardized and reproducible function. Scanning electron microscopic study of the resulting electrode surfaces shows them to be smooth and featureless at all levels of magnification. X-ray diffraction analysis of the electrodes indicates preferential exposure of the [1,1,1] crystal interface. Cyclic voltammetry evaluation of initial sensor response to varying glucose concentrations shows excellent sensor to sensor reproducibility for all sensors made with the same underlayment. Sensors made with titanium underlayment appear to be more differentiated and thus more sensitive to variations in glucose concentration than are sensors with chromium underlayment. Although the initial response of microchip glucose sensors appears to be standardized and reproducible, additional development of an appropriate electrical insulation material is required before long-term study of signal stability is feasible.

Safety observations in phase I clinical evaluation of the Excorp Medical Bioartificial Liver Support System after the first four patients.

A Phase I clinical safety evaluation of the Excorp Medical, Inc, Bioartificial Liver Support System (BLSS) is in progress. Inclusion criteria are patients with acute liver failure of any etiology, presenting with encephalopathy deteriorating beyond Parson's Grade 2. The BLSS consists of a blood pump, heat exchanger to control blood temperature, oxygenator to control oxygenation and pH, bioreactor, and associated pressure and flow alarm systems. Patient liver support is provided by 70-100 g of porcine liver cells housed in the hollow fiber bioreactor. A single support period evaluation consists of 12 hour extracorporeal perfusion with the BLSS sandwiched between 12 hours of pre (baseline) and 12 hours of post support monitoring. Blood chemistries and hematologies are obtained every 6 hours during monitoring periods and every 4 hours during perfusion. Physiologic parameters are monitored continuously. The patient may receive a second treatment at the discretion of the clinical physician. Preliminary evaluation of safety considerations after enrollment of the first four patients (F, 41, acetaminophen induced, two support periods; M, 50, Wilson's disease, one support period; F, 53, acute alcoholic hepatitis, two support periods; F, 24, chemotherapy induced, one support period) is presented. All patients tolerated the extracorporeal perfusion well. All patients presented with hypoglycemia at the start of perfusion, treatable by IV dextrose. Transient hypotension at the start of perfusion responded to an IV fluid bolus. Only the second patient required heparin anticoagulation. No serious or unexpected adverse events were noted. Moderate biochemical response to support was noted in all patients. Completion of the Phase I safety evaluation is required to fully characterize the safety of the BLSS.