Development of a collagen gel sandwich hepatocyte bioreactor for detecting hepatotoxicity of drugs and chemicals

Abstract

Understanding the hepatotoxicity of drugs and chemicals is essential for progress in academic research, medical science and in the development of new pharmaceuticals. Studying hepatotoxicity in vitro is a challenging task because hepatocytes, the metabolically active cells of the liver, are very difficult to maintain in culture. After just 24 hours, the cells detach from the plate and die, and even if they survive they usually do not express the metabolic functions which they have in vivo. It has been observed by others that culturing hepatocytes between two layers of collagen type I maintains in vivo-like morphology and also many drug metabolizing enzymes for weeks. In spite of the research examining drug metabolism in collagen sandwiches, there are very few studies evaluating this system for investigating hepatotoxicity. We cultured primary rat hepatocytes in the collagen sandwich configuration and our goal was to optimize this system for long-term studies and to examine toxicity of a variety of hepatotoxins. By measuring secretions of urea and albumin, and P4501A activity, we determined the optimal cell density to be 50,000 cells/cm2. We also evaluated the need for epidermal growth factor (EGF) in our cultures, by comparing urea and albumin secretions in cultures grown with and without EGF. The cultures without EGF had significantly less secretion of both urea and albumin just two days after plating. Therefore, we decided to include EGF in the medium. The toxins we examined were aflatoxin B1, acetaminophen, carbon tetrachloride, N-methyl-N'-nitro-N-nitrosoguanidine (MNNG), methyl methane sulfonate (MMS), cadmium, vinyl acetate and dimethylformamide (DMF). The cells were sensitive to aflatoxin B1, MMS, MNNG and cadmium. However, they were

(cont.) immune to acetaminophen, carbon tetrachloride, vinyl acetate and DMF. Our Western Blots showed that CYP1A, 2B and 3A were maintained in the culture for a week, but CYP2E1 was lost gradually over time. CYP2E1 is also the primary metabolic enzyme for acetaminophen, carbon tetrachloride and DMF. Thus, it is possible that the lack of toxicity is due to the loss of the enzyme responsible for the metabolism of these compounds. Immunity to vinyl acetate suggests that carboxylesterase is also lost in culture, since this enzyme is the one which converts vinyl acetate to acetaldehyde. The metabolism of acetaminophen was also examined with liquid chromatography and mass spectrometry. Liquid chromatography showed that acetaminophen is metabolized primarily to the sulfate and glucuronide metabolites. In order to investigate whether the glutathione adduct was formed, we synthesized the adduct and determined its retention time with liquid chromatography and its fragmentation pattern with mass spectrometry. We isolated the fraction with the same retention time from the medium of acetaminophen-treated cells, and showed that it contains a peak with the same mass to charge ratio and fragmentation pattern as the glutathione adduct. We also examined the conditioned medium from the hepatocytes to investigate the secreted protein profile, which could potentially be used to find toxicity biomarkers. We were able to remove most of the albumin from the medium using an immuno-affinity column containing anti-albumin antibodies bound to protein A-agarose beads ...