Serum-free collagen sandwich cultures of adult rat hepatocytes maintain liver-like properties long term: a valuable model for in vitro toxicity and drug-drug interaction studies.

Cultures of primary hepatocytes from various species, including human, are used in several applications during pre-clinical drug development. Their use is however limited by cell survival and conservation of liver-specific functions in vitro. The differentiation status of hepatocytes in culture strongly depends on medium formulation and the extracellular matrix environment. We incubated primary rat hepatocytes for 10 days on collagen monolayer and in collagen sandwich cultures with or without serum. Restoration of polygonal cell shape and formation of functional bile canaliculi-like structures was stable only in serum-free sandwich cultures. Variations in general cell viability, as judged by the cellular ATP content, LDH release or apoptosis, were less pronounced between alternative cultures. The intracellular glutathione content was preserved close to in vivo levels especially in serum-free sandwich cultures. Basal activities of cytochrome P450 enzymes (P450) varied strongly between cultures. There was a minor effect on CYP1A but CYP2B activity was only detectable in the serum-free sandwich culture after 3 days and beyond. CYP2C activity was slightly elevated in both sandwich cultures, whereas CYP3A showed increased levels in both serum-free cultures. Inducibility of these P450s was fully maintained over time in serum-free collagen sandwich only. Gene expression was largely constant over time in serum-free sandwich cultures that was closest to liver. This liver-like property was supported by protein profiling results. Taken together, the serum-free collagen sandwich culture of primary rat hepatocytes maintained liver-like features over 10 days and is therefore a suitable model for long-term toxicity and drug-drug interaction studies.

Prevalidation of potential protein biomarkers in toxicology using iTRAQ reagent technology.

Today, toxicoproteomics still relies mainly on 2-DE followed by MS for detection and identification of proteins, which might characterize a certain state of disease, indicate toxicity or even predict carcinogenicity. We utilized the classical 2-DE/MS approach for the evaluation of early protein biomarkers which are predictive for chemically induced hepatocarcinogenesis in rats. We were able to identify statistically significantly deregulated proteins in N-nitrosomorpholine exposed rat liver tissue. Based on literature data, biological relevance in the early molecular process of hepatocarcinogenicity could be suggested for most of these potential biomarkers. However, in order to ensure reliable results and to create the prerequisites necessary for integration in routine toxicology studies in the future, these protein expression patterns need to be prevalidated using independent technology platforms. In the current study, we evaluated the usefulness of iTRAQ reagent technology (Applied Biosystems, Framingham, USA), a recently introduced MS-based protein quantitation method, for verification of the 2-DE/MS biomarkers. In summary, the regulation of 26 2-DE/MS derived protein biomarkers could be verified. Proteins like HSP 90-beta, annexin A5, ketohexokinase, N-hydroxyarylamine sulfotransferase, ornithine aminotransferase, and adenosine kinase showed highly comparable fold changes using both proteomic quantitation strategies. In addition, iTRAQ analysis delivered further potential biomarkers with biological relevance to the processes of hepatocarcinogenicity: e.g. placental form of glutathione S-transferase (GST-P), carbonic anhydrase, and aflatoxin B1 aldehyde reductase. Our results show both the usefulness of iTRAQ reagent technology for biomarker prevalidation as well as for identification of further potential marker proteins, which are indicative for liver hepatocarcinogenicity.

Comparison of software tools to improve the detection of carcinogen induced changes in the rat liver proteome by analyzing SELDI-TOF-MS spectra.

A common animal model of chemical hepatocarcinogenesis was used to demonstrate the potential identification of carcinogenicity related protein signatures/biomarkers. Therefore, an animal study in which rats were treated with the known liver carcinogen N-nitrosomorpholine (NNM) or the corresponding vehicle was evaluated. Histopathological investigation as well as SELDI-TOF-MS analysis was performed. SELDI-TOF-MS is an affinity-based mass spectrometry method in which subsets of proteins from biological samples are selectively adsorbed to a chemically modified surface. The proteins are subsequently analyzed with respect to their mass-charge ratios (m/z) by a time of flight (TOF) mass spectrometry (MS) approach. As data preprocessing of SELDI-TOF-MS spectra is essential, baseline correction, normalization, peak detection, and alignment of raw spectra were performed using either the Ciphergen ProteinChip Software 3.1 or functions implemented in the library PROcess of the BioConductor Project. Baseline correction and normalization algorithms of both tools lead to comparable results, whereas results after peak detection and alignment steps differed. Variability between technical and biological replicates was investigated. A linear mixed model with factors experimental group and time point was applied for each protein peak, taking into account the different correlation structure of technical and biological replicates. Alternatively, only median intensity values of technical replicates were used. Results of both models were similar and correlated well with those of the histopathological evaluation of the study. In conclusion, statistical analyses lead to comparable results, whereas parameter settings for preprocessing proved to be crucial.