Network Analysis of Primary Hepatocyte Dedifferentiation Using a Shotgun Proteomics Approach Cliff Rowe,† Christopher E. P. Goldring,*,† Neil R. Kitteringham,† Rosalind E. Jenkins,† Brian S. Lane,‡ Christopher Sanderson,§ Victoria Elliott,† Vivien Platt,† Peter Metcalfe,† and B. Kevin Park† MRC Centre for Drug Safety Science, Department of Pharmacology & Therapeutics, Sherrington Building, Ashton Street, The University of Liverpool, Liverpool L69 3GE, United Kingdom, School of Infection and Host Defense, Department of Medical Microbiology, The University of Liverpool, Liverpool L69 3BX, United Kingdom, and School of Biomedical Sciences, Department of Physiology, Crown Street, The University of Liverpool, Liverpool L69 3BX, United Kingdom Received February 25, 2010
The liver is the major site of xenobiotic metabolism and detoxification. Primary cultures of hepatocytes are a vital tool in the development of new therapeutic agents but their utility is hindered by the rapid loss of phenotype. Hepatocytes cultured in a sandwich of extracellular matrix protein maintain better hepatic function compared with cells cultured as a monolayer but a wide-ranging proteomics study of the differences in cultures has never been performed. We characterize the changing phenotype of rat hepatocytes in primary culture using iTRAQ proteomics and systems biology network analysis of the identified, significantly regulated, proteins. A total of 754 unique proteins were identified from 4 independent experiments. Of these, 413 proteins were common to at least 3 experiments and 328 proteins were identified in all experiments. Both culture systems displayed altered expression of many common proteins. Network analysis showed that the primary functions of these proteins were in metabolic pathways, immune responses and cytoskeleton remodelling. Monolayer cultures uniquely regulate proteins mapping to pathways of oxidative stress and cell migration, whereas sandwich culture affected translation regulation and apoptosis pathways. These experiments provide a detailed proteomics data set to direct further work into maintaining hepatic phenotype using cultured primary hepatocytes and stem cell derived hepatocyte-like cells. Keywords: hepatocytes • proteomics • cytochrome P450 • extracellular matrix • xenobiotic • drug metabolism
Introduction Primary cultures of mammalian hepatocytes are a vital tool in the development of new therapeutic agents.1 The liver can be a major determinant of the therapeutic, toxicological or physiological effects of potential new drugs. Hepatic metabolism may variously lead to a drug’s conversion to an active form,2 its reduced or abolished efficacy,3 or the formation of chemically reactive and toxic species.4 As such, the effects that a candidate drug has on the liver, and the effect the liver has on the candidate drug, may be major factors determining the continuation or termination of its development. * To whom correspondence should be addressed. MRC Centre for Drug Safety Science, Department of Pharmacology & Therapeutics, Sherrington Building, Ashton Street, The University of Liverpool, L69 3GE. E-mail:
[email protected]. Telephone: +44 (0)151 794 5979. Fax: +44 (0)151 794 5540. † MRC Centre for Drug Safety Science, Department of Pharmacology & Therapeutics, The University of Liverpool. ‡ School of Infection and Host Defense, Department of Medical Microbiology, The University of Liverpool. § School of Biomedical Sciences, Department of Physiology, The University of Liverpool.
2658 Journal of Proteome Research 2010, 9, 2658–2668 Published on Web 04/08/2010
Human primary hepatocytes can offer an excellent model of the liver hepatocyte, maintaining many functions in the short term, but their utility is hindered by loss of liver phenotype with time5 and by the poor availability and predictability in the supply of fresh human cells. The phase I metabolic enzymes, particularly the cytochrome P450 family of membranebound hemoproteins, are of particular relevance to drug metabolism and hepatotoxicity studies. Despite real advances in culture techniques, these proteins are poorly maintained in hepatocytes cultured for several days, and preparations must generally be used immediately where possible. While species differences in drug metabolism and susceptibility to some hepatotoxic compounds can occur,6 single cell suspensions and adherent cultures of rat hepatocytes remain a useful tool in drug development programs. They are used in initial toxicity screens of candidate compounds and can help to reduce the number of compounds progressing to in vivo toxicity studies, and also decrease the number of experimental animals needed. However, rodent hepatocyte cultures also have a propensity to dedifferentiate. Many factors affect the phenotype of cultured rat hepatocytes and various studies have 10.1021/pr1001687
2010 American Chemical Society
Proteomic Analysis of Rat Primary Hepatocyte Dedifferentiation 7–9
been devoted to establishing optimal techniques for culture. Hepatocytes cultured as a monolayer spread and flatten and take on a fibroblast-like appearance, while those cultured in a sandwich of extracellular matrix proteins reform structures and morphologies resembling those found in liver and remain viable and functional for more prolonged time periods.10 Genetranscript array-based studies have illustrated the important role extracellular matrix plays in maintaining a more liver-like gene expression profile,11,12 but the question of how this relates to the proteome has not been fully addressed. Global proteomics technologies now make such studies possible and a recent study has demonstrated the large divergence in the phenotype of a mouse liver cell line away from a mature mouse hepatocyte phenotype,13 confirming the inherent limitations of cell lines as liver models. However, a temporal, shotgun proteomics analysis of the diverging phenotype of differentially cultured primary hepatocytes has not been performed to date. In this study, we use quantitative (iTRAQ) proteomics, a rigorous statistical analysis in the R computing environment, and network analysis (Metacore GeneGo) to characterize the temporal changes in the proteins constituting membrane preparations and map these changes to cellular networks. The data sets quantitatively describe, on a large scale, the temporal changes in individual proteins that occur when hepatocytes are cultured under minimal conditions for 3 days and the effects that Matrigel sandwich culture has on this phenotype. We show that the abundance of proteins with function in metabolic, immune response, and cytoskeleton remodelling pathways were changed in cultures in the presence or absence of extracellular matrix. Cultures without extracellular matrix overlay displayed uniquely altered proteins with functions in pathways of oxidative stress and cell migration, whereas maintenance of the cells in sandwich culture influenced translation regulation and apoptosis pathways. These findings generate new testable hypotheses to explore why the phenotype changes in culture and may uncover potential new targets for manipulation in order to better maintain the hepatocyte phenotype in culture.
Experimental Section Hepatocyte Isolation and Culture. Culture dishes were either purchased precoated with rat tail collagen I (Becton Dickinson, Oxford, U.K.), or were coated in-house with Matrigel (Becton Dickinson, Oxford, U.K.). Matrigel, diluted to 1 mg/ mL in ice-cold Williams’ E Medium (WEM) (Sigma, Poole, U.K.), was spread evenly across the plate (6 µg/cm2) using a sterile glass spreader and left to dry at room temperature for at least 1 h before use. Male Wistar rats weighing approximately 250 g (Charles River, Manston, U.K.) were housed in controlled environmental conditions of 12 h light/dark cycle with free access to food and water. Rats were administered 200 µL of pentoject by intraperitoneal injection and, once deeply anesthetized, a laparotomy was performed. The hepatic portal vein was cannulated using a 20 g × 48 mm vialon catheter (Becton Dickinson, U.K.), tied in place with cotton thread. Hanks balanced salt solution (HBSS) (Sigma, Poole, U.K.) without calcium and magnesium was perfused through the liver at 37 °C for 10 min while the liver was excised. Once free of the carcass, 100 mL of complete HBSS containing 50 mg of collagenase A (Roche, Welwyn Garden City, U.K.) was perfused through the liver with recirculation until it was digested. The cells were dispersed into 100 mL of WEM (Sigma, Poole, U.K.) and filtered through a 125
research articles
µm gauze into a sterile container. The cells were washed three times by centrifugation at 100g for 2 min and an aliquot was taken for counting. Cell suspensions (0.5 × 106/mL) were prepared in WEM supplemented with 100 U/mL penicillin and 100 µg/mL streptomycin, 2 mM L glutamine (Sigma, Poole, U.K.), 10% fetal calf serum (Lonza biologics, Slough, U.K.), and 1 µg/mL bovine insulin (Sigma, Poole, U.K.) and seeded at 70 000 cells/cm2. The seeded plates were incubated at 37 °C in a humidified atmosphere containing 5% CO2 for 3 h before the medium was exchanged for serum- and insulin-free WEM. For Matrigel sandwich cultures, the replacement medium was ice-cold and contained 0.25 mg/mL Matrigel. Protein Preparation. At each time point, hepatocytes were harvested from three 100 mm plates into ice-cold phosphatebuffered saline (PBS) and cell pellets were obtained by centrifugation (1 min, 1000g, 4 °C). Pellets were resuspended in 0.5 mL of PBS and sonicated with 10 2-s pulses. The homogenates were centrifuged at 9000g for 20 min and the supernatants centrifuged at 100 000g for 1 h. The resulting pellets (crude microsomes/membrane fractions) were resuspended in 50 µL of triethylammonium bicarbonate, 0.1% SDS, and the protein concentration were determined using Bradford assay reagents (Sigma, Poole, U.K.). Western Blotting for Cytochrome P450. Proteins were denatured and reduced by heating to 80 °C for 10 min in Laemmli sample buffer (Sigma, Poole, U.K.) and 10 µg of each sample was resolved through a 10% polyacrylamide gel (Biorad, Hemel Hempstead, U.K.) and transferred to a nitrocellulose membrane (GE Healthcare, Slough, U.K.). After transfer, the nitrocellulose membrane was blocked overnight in Trisbuffered saline, containing 0.1% Tween 20, (Sigma, Poole, U.K.) at 4 °C. CYP1A2 and CYP2E1 antibodies (Abcam, Cambridge, U.K.) were used according to the product insert. CYP3A2 antibody was a kind gift from by Dr. R. Edwards (Imperial College, London, U.K.). Membranes were incubated with horseradish peroxidise-conjugatedanti-mouse(CYP1A2),anti-rabbit(CYP2E1), or anti-sheep (CYP3A2) IgG antibody and detected using ECL detection reagents and ECL hyperfilm (GE Healthcare, Slough, U.K.). Band volumes were measured by densitometry. RNA Extraction and RT2 Profiler PCR Array. Total RNA was extracted from approximately 1 × 106 cells using Trizol reagent (Invitrogen, Paisley, U.K.) according to the manufacturer’s instructions. The RNA was quantified using a nanodrop spectrophotometer and 1 µg was DNase-treated and reversetranscribed using the reagents supplied with the rat drug metabolism SA Biosciences RT2 profiler PCR array (TebuBio, Peterborough, U.K.). Total RNA (1 µg) was added to 2 µL of 5× genomic DNA elimination buffer and the volume adjusted to 10 µL. The mixture was incubated at 42 °C for 5 min and immediately placed on ice for 1 min. A reverse transcription (RT) cocktail was made up of 4 µL of buffer 3, 1 µL of primer and external control mix, 2 µL of RT enzyme mix 3, and 3 µL of RNase free water per reaction. A total of 10 µL of the RT cocktail was added to each 10 µL Genomic DNA elimination mixture for a final volume of 20 µL. The reaction was carried out at 42 °C for 15 min and immediately stopped by heating to 95 °C for 5 min. Then, 91 µL of nuclease-free water was added to each reaction and the diluted cDNA was held on ice for immediate use or stored at -20 °C. The real-time PCR reactions were set up according to the SA Biosciences RT2 profiler PCR array instructions; 102 µL of Journal of Proteome Research • Vol. 9, No. 5, 2010 2659
research articles
Rowe et al.
Figure 1. Hepatocytes differ in morphology under different culture conditions. Hepatocytes were seeded from the same cell suspension onto culture dishes precoated with either collagen I (A) or Matrigel (B). Images were taken using an inverted light microscope at 3 h (before Matrigel overlay for sandwich cultures), 24, 48, and 72 h after seeding following each media change. Scale bar ) 100 µm.
diluted cDNA was added to 1350 µL of SYBR green mix and 1248 µL of nuclease-free water. Then, 25 µL of the experimental cocktail was loaded into each well on the supplied 96-well plate, containing gene specific primers, and the reaction and carried out on an ABI 7000 machine. iTRAQ Analysis. The labeling procedure was undertaken according to the manufacturer’s instructions (Applied Biosystems, Warrington, U.K.). A total of 75 µg of protein from each sample, in a volume of 20 µL of triethylammonium bicarbonate, was reduced with 2 µL of tris(2-carboxyethyl)phosphine at 60 °C for 1 h. Cysteine sulfhydryls were blocked by addition of 1 µL of methyl methanethiosulfonate at room temperature for 10 min. Proteins were digested by addition of 10 µL of reconstituted trypsin and incubation at 37 °C overnight (12-16 h). Isopropanol was added to each iTRAQ reagent, which was then transferred to the appropriate sample tube. Proteins derived from freshly- isolated cells were labeled with iTRAQ tag 113; hepatocytes cultured on collagen I for 24, 48, or 72 h were labeled with tags 114, 115, or 116, respectively. Proteins from hepatocytes cultured as a Matrigel sandwich for 24, 48, 72, or 168 h were labeled with tags 117, 118, 119, and 121 respectively. Following incubation for 2 h at room temperature, the tagged proteins were combined into a fresh tube and made up to a final volume of 5 mL in cation exchange mobile phase (10 mM KH2PO4/25% acetonitrile (ACN) pH < 3). The pH was adjusted to
