Human hepatocytes both in suspension and in primary culture are increasingly used for biochemical, pharmacological, and toxicological research. Cultured hepatocytes retain typical differentiated hepatic functions; however, a major drawback of this in vitro model is the relative instability of P450 and other liver-specific enzymes along time in culture (Gómez-Lechón et al., 1990, 2004). Despite certain limitations, the metabolic competence of cultured human hepatocytes, in addition to ethical and economic considerations, are key reasons for considering this experimental model as a valuable tool for anticipating the metabolism of new drugs in the early stages of development and for investigating possible hepatic side effects of drugs on specific liver functions.

Since the introduction of two-step collagenase perfusion to isolate hepatocytes from human liver, the methods for obtaining high yields of viable hepatocytes have been extensively studied and improved (Gómez-Lechón et al., 1990; Dorko et al., 1994; David et al., 1998). A major limitation of using human liver for xenobiotic metabolism studies is ensuring a regular supply of adequate amounts of the tissue required to sustain a research program. Obviously, nonpathologic liver tissue is needed for hepatocyte obtention, which has greatly hindered the widespread use of human hepatocytes. Healthy human liver is only occasionally available for experimental purposes, and it is still too scarce to support the high demand derived from the use of in vitro assays during drug development. This scarcity is increased by the competing demands for human cells for clinical use, such as hepatocyte transplantation programs or artificial liver systems. The major sources of hepatic tissue are livers that are discarded for transplantation, tissue from split-liver transplantation, waste material from partial therapeutic hepatectomy, or small sized surgical biopsies. Knowledge of the suitability of liver samples from different origins as sources of viable and metabolically competent human hepatocytes for drug metabolism research has become of great interest. However, up to now, information about the influence of the liver source on functionality of cultured hepatocytes has been scarce (Olinga et al., 1998a,b; Serralta et al., 2003).

The expansion of the liver transplantation programs has contributed to the increasing availability of human liver tissue for research purposes. Donor livers that cannot be implanted for different reasons can be used for hepatocyte isolation (Donini et al., 2001). In contrast to tissue samples from other sources, livers from organ donors are perfused in situ with a cold preservation solution to avoid warm ischemia, and the tissue is usually transported and maintained under these conditions for several hours until hepatocyte isolation. Cold ischemia is a risk factor for organ function (Muhlbacher et al., 1999), and it could be also considered as a possible factor involved in the efficiency of the isolation procedure and the metabolic competence of cultured cells. There are some concerns about the quality of hepatocytes obtained from liver samples stored in the University of Wisconsin (UW) solution, the most commonly used solution to preserve livers for transplantation (Quintana et al., 2003). It was reported that hepatocytes isolated from fresh liver maintain better functionality than hepatocytes obtained from organ donor liver after hypothermic storage for several hours in UW solution (Vons et al., 1990; David et al., 1998). However, other studies reported that human liver, isolated hepatocytes, or liver slices can be efficiently stored in cold solutions for several hours with minor losses of functionality (Caraceni et al., 1994; Dorko et al., 1994; Olinga et al., 1998a,b; Spinelli et al., 2002). Celsior solution (CS), previously utilized for heart and lungs, was recently proposed for kidney and liver preservation (Muhlbacher et al., 1999). Preliminary results show that the new CS could be a valid alternative to UW preservation solution in multiorgan harvesting, including the liver (Faenza et al., 2001; Cavallari et al., 2003; Janssen et al., 2003).

The aim of our study was to assess the efficacy of CS in cold preservation of human liver tissue for hepatocyte isolation. For this, we comparatively evaluated the oxidative drug-metabolizing capacity of primary human hepatocytes prepared from unused liver grafts perfused and preserved in CS or from other sources of liver tissue. Furthermore, the response of cultured hepatocytes to model cytochrome P450 (P450) inducers was also examined.

Materials and Methods

Chemicals. Collagenase and β-glucuronidase/arylsulfatase were obtained from Roche Diagnostics (Mannheim, Germany); 7-methoxyresorufin was obtained from Molecular Probes Europe BV (Leiden, The Netherlands); coumarin, chlorzoxazone, diclofenac, 7-ethoxycoumarin, 7-hydroxycoumarin, resorufin, testosterone, and androstenedione were purchased from Sigma-Aldrich (St. Louis, MO); 4′-hydroxydiclofenac, (S)-(+)-mephenytoin, 4′-hydroxymephenytoin, 6-hydroxychlorzoxazone, 2β-, 6β-, 16α-, 16β-, and 11β-hydroxytestosterone (OHT), paclitaxel, and 6-hydroxy paclitaxel were supplied by Ultrafine (Manchester, UK); a sample of 15β-OHT (G.D. Searle and Co., Skokie, IL) was kindly supplied by Dr. B. Blaauboer (Utrecht, The Netherlands); culture media (Ham F-12/Williams) and newborn calf serum were obtained from Invitrogen (Carlsbad, CA); and all other reagents used in this study were of analytical grade.

Human Liver Tissues. Livers from cadaveric donors were perfused with CS. Perfusion of the livers was performed by cannulation of aorta and porta, and 2 liters of CS was used for both aortic and portal perfusion. Perfusion pressure was always constant. Eleven CS-flushed liver grafts were used. Samples of liver tissue (2-4 g) were obtained at the time of bank surgery from the liver grafts. Liver tissue was always obtained from the anterior surface of segment III of Couinaud. Samples were kept in preservation solution at 4°C until hepatocyte isolation. The sample was discarded when the isolation procedure could not be performed within 36 h after perfusion of the liver (cold ischemia time over 36 h).

Morphologically normal liver samples obtained in the course of surgical liver resections (SLRs) (cholecystectomy, hydatic cyst, colon carcinoma, gastric carcinoma) were also used for hepatocyte isolation. In this new group, a total of 20 SLR biopsies were obtained, in conformity with the rules of the Hospital's Ethics Committee and after informed consent from patients. These samples were immediately processed after extraction without perfusion with preservation solutions.

All liver samples were from donors who were not suspected of harboring any infectious disease and tested negative for human immunodeficiency virus and hepatitis. No underlying liver pathology was present in any of the cases. For cell harvesting, all the samples were obtained from an area with anatomic configuration (adequate venous caliber and Glisson capsule surrounding three of the stereofaces) that allowed adequate perfusion of the tissue and isolation of hepatocytes.

Isolation and Culture of Hepatocytes. Human hepatocytes were isolated using a two-step perfusion technique and cultured as described (Gómez-Lechón et al., 1990). Cellular viability was assessed by the dye exclusion test with 0.4% trypan blue in saline. Hepatocytes were seeded on fibronectin-coated plastic dishes (3.5-μg/cm²) at a density of 8 × 10⁴ viable cells/cm² and cultured in Ham's F-12/Williams (1:1) medium supplemented with 2% newborn calf serum, 50 mU/ml penicillin, 50 μg/ml streptomycin, 0.1% bovine serum albumin, 10^-8 M insulin, 25 μg/ml transferrin, 0.1 μM sodium selenite, 65.5 μM ethanolamine, 7.2 μM linoleic acid, 17.5 mM glucose, 6.14 mM ascorbic acid, and 0.64 mM N-ω-nitro-l-arginine methyl ester. The medium was changed 1 h later to remove unattached hepatocytes. By 24 h, the cells were shifted to serum-free medium and 10^-8 M dexamethasone was added. Thereafter, the medium was renewed daily.

Evaluation of Metabolic Competence of Human Hepatocytes. After 24 h in culture, P450 activities were measured by incubating intact hepatocyte monolayers at 37°C with specific substrates (Table 1). Reactions were stopped by aspirating the incubation medium from plates, and medium samples were then incubated with β-glucuronidase and arylsulfatase for 2 h at 37°C (Donato et al., 1993). Metabolites formed during oxidation of P450 substrates and released into medium of incubation were quantified fluorometrically or by HPLC analysis (Table 1). Total cellular protein was quantified as described (Lowry et al., 1951).

Treatment of Cultures. Human hepatocytes were exposed to inducers 24 h after plating. Phenobarbital (PB) was prepared as an aqueous solution and added directly to cultures at a final concentration of 1 mM. 3-Methylcholanthrene (MC) and rifampicin (RIF) were dissolved in dimethyl sulfoxide and added to the culture medium to give a final concentration of 2 μM and 50 μM, respectively. The dimethyl sulfoxide concentration in culture medium never exceeded 0.1% (v/v). Control cultures were treated with the same concentration of the respective solvent.

Isolation, Quantification, and Purification of Total RNA from Cell Cultures. Total RNA was extracted from 6-cm-diameter culture plates of human hepatocytes using TRIzol, following the supplier's recommendations. The amount of purified RNA was estimated by RiboGreen fluorescence, and its purity was determined by the absorbance ratio at 260:280 nm. RNA was incubated for 15 min at 23°C with DNase I (1 unit/μg) according to the supplier's recommendations, followed by thermal inactivation of the enzyme (10 min at 65°C) in the presence of 2.5 mM EDTA and rapid cooling to 4°C.

Measurement of mRNAs by RT-PCR. The reverse transcription (RT) reaction mixture consisted of 1 μg of total RNA, which was reverse transcribed in 20 μl of reverse transcriptase buffer, 10 mM dithiothreitol, 500 μM deoxynucleotides, 3 μM oligo d(T)14 primer, 60 U RNase, and 250 U reverse transcriptase. The reaction was allowed to proceed for 60 min at 42°C, followed by 5 min of heating at 95°C and then rapid cooling on ice. The cDNA was stored at -20°C until use. The polymerase chain reaction (PCR) was conducted in semiautomatic equipment (LightCycler; Roche Diagnostics). The conditions used for the quantitative PCR, the specific primers, and the strategy of quantification were as previously described (Pérez et al., 2003).

Statistical Analysis. Data are expressed as mean ± S.D. A nonparametric Mann-Whitney U test was used to determine whether there is any evidence of a difference between CS and SLR groups. Student's t test was used to compare hepatocytes treated with inducers with respect to the corresponding control (untreated cells).

Results

P450 Expression in Cultured Hepatocytes from CS-Perfused Liver Grafts. Hepatocytes were isolated from 11 different human liver grafts perfused and preserved in CS as described under Materials and Methods. By 24 h of culture, functional competence of the cultures was systematically evaluated in all human hepatocyte preparations by measuring P450-dependent oxidations. To this end, ECOD activity was quantified as representative of total P450 activity, and 6β-OHT was evaluated as a selective probe for CYP3A4, the major P450 enzyme in human liver. For both P450 activities, variability was observed among individual donors. This variability, similar to that observed in human hepatocyte cultures prepared from other sources of liver tissue, is illustrated in Fig. 1. As an average, human hepatocyte preparations included in the CS group showed significantly lower ECOD activity values than did hepatocytes from SLR samples (9.2 ± 5.2 versus 17.4 ± 6.4 pmol/mg · min), whereas no significant differences were observed in 6β-OHT activity (29.7 ± 21.0 versus 41.8 ± 26.6 pmol/mg · min).

In six human hepatocyte cultures from the CS group, an exhaustive analysis of P450 enzymes was performed. The expression of major hepatic P450 involved in drug metabolism was examined by measuring specific mRNA contents and catalytic activities. Levels of mRNAs corresponding to 10 P450 enzymes, namely, CYP1A1, 1A2, 2A6, 2B6, 2C9, 2C19, 2D6, 2E1, 3A4, and 3A5, were quantified in 24-h-old cultured hepatocytes (Table 2). Measurable levels of all P450 mRNAs studied were found in the six cultures. P450 mRNA variation among cell preparations was small (<5-fold variation) for CYP2E1 and CYP2B6, whereas CYP1A1 showed the highest variability (>400-fold variation). The most abundant P450 mRNAs were those of CYP2C9, CYP2E1, and CYP3A4, representing about 37%, 22%, and 24%, respectively, of total P450 mRNA content. In contrast, CYP1A1, CYP1A2, and CYP2A6 mRNA levels represented <1% of the total in all hepatocyte preparations. In general, hepatocytes from the CS group showed lower P450 mRNA content than did hepatocytes obtained from other sources. However, the relative distribution of individual P450 mRNAs was similar in both groups (Table 2).

Activity of six relevant drug-metabolizing P450 enzymes were also determined by the use of selective substrates. As expected, interindividual variability was also found, with MROD (CYP1A2) activity showing the highest variations between donors (Table 3). Similar to that previously observed for mRNA levels, comparison of individual activity values found in cultured hepatocytes prepared from CS-perfused liver grafts or livers from other sources revealed a comparable P450 enzyme activity pattern in both groups.

Testosterone Metabolism Profiles by Primary Human Hepatocytes. The oxidative metabolism of testosterone was analyzed to estimate drug-metabolizing capacity of the cells. Testosterone is metabolized in a regioselective manner by several P450 enzymes and can be used as a multienzymatic substrate to investigate the activity of multiple P450s simultaneously. After 24-h incubation of cultured cells with testosterone, 6β-OHT (CYP3A4) and androstenedione (CYP2C19) were the most actively formed in all hepatocyte preparations (data not shown). Other metabolites identified, ranked according to formation rates, were 2β-OHT (CYP3A4), 15β-OHT (CYP3A4), 16α-OHT, and 16β-OHT (CYP2C9). Comparison of the testosterone metabolite profile of hepatocytes from CS-preserved liver grafts with those of cells isolated from other liver sources confirmed the high similarities in metabolic capability existing between both groups (Fig. 2). Independent of the total testosterone oxidation rates, which were about 30% lower in the CS group (83 ± 47 versus 119 ± 55), no appreciable differences in the relative formation of each metabolite were found.

P450 Inducibility in Human Hepatocyte Cultures. The ability of primary human hepatocytes obtained from liver grafts preserved in CS to respond to P450 induction was examined by treating cultured cells with prototypical enzyme inducers. After 24 h in standard culture conditions, hepatocytes were exposed to 2 μM MC, 1 mM PB, or 50 μM RIF and levels of specific mRNA from the major hepatic P450s were quantified 48 h later.

For all the P450s examined, variability was observed among individual donors in the magnitude of the response to model inducers. This variability is illustrated in Fig. 3 for the effect of MC on CYP1A2 mRNA, PB on CYP2B6 mRNA, and RIF on CYP3A4 mRNA levels. Figure 3A shows CYP1A2 mRNA content of control or MC-treated human hepatocyte cultures from six different liver graft donors preserved with CS. Overall, CYP1A2 mRNA levels following exposure to MC ranged from 40- to 225-fold over untreated cells. Compared with the effect in hepatocytes prepared from SLR samples, a similar response was found in both groups (about 100-fold increases over control). Increases of CYP2B6 and CYP3A4 mRNA levels after treatment with PB or RIF, respectively, were also observed in hepatocyte cultures (Fig. 3, B and C). The averaged response in hepatocytes from the CS group was comparable to that of cultures from SLRs for CYP2B6 mRNA but lower for CYP3A4 mRNA. The inducers also produced increases of CYP1A2 and CYP3A4 activities (Fig. 4). For all the hepatocyte cultures from the CS group, MC was a strong inducer of MROD (CYP1A2) activity (25-fold over control), and both PB and RIF increased 6β-OHT (CYP3A4) activity (about 3-fold over control).

The results obtained after quantification of CYP1A1, 1A2, 2A6, 2B6, 2C9, 2C19, 2D6, 2E1, 3A4, and 3A5 mRNAs of human hepatocytes treated with MC, PB, or RIF are summarized in Fig. 5. Levels of specific mRNAs from six (for MC and RIF treatment) and five (for PB) individual cell cultures prepared from CS-preserved liver grafts were averaged and compared with those quantified in hepatocytes prepared from other sources of human liver samples. In both groups of hepatocyte cultures, treatment with MC caused a marked increase in CYP1A1 and CYP1A2 mRNA content, whereas minor or no effects were observed in the levels of the other P450 mRNAs. Exposure of human hepatocyte cultures from the CS group to PB markedly increased (>7-fold over control) CYP2A6, CYP2B6, CYP2C9, and CYP3A4 mRNA content (Fig. 5A). No significant changes in the mRNA levels of the other P450 enzymes examined were observed. A similar response to PB was found in hepatocyte cultures prepared from other sources of liver tissue (Fig. 5B). Comparable effects on mRNA levels of individual P450 enzymes were observed after treatment with RIF of hepatocytes from both groups of liver samples. Increases of several P450 mRNAs were observed, and the highest effect was produced on CYP3A4 mRNA (>10-fold over untreated cells). None of the model inducers studied had any significant effect on CYP2D6 or CYP2E1 mRNA content of hepatocytes, independently of the type of liver sample used for cell isolation.

Discussion

Different solutions have been used for liver graft preservation, and among them, CS has recently emerged as a multiorgan preservation solution (Faenza et al., 2001; Cavallari et al., 2003). In the present study, the suitability of liver grafts perfused and maintained in CS for preparing metabolically competent hepatocyte cultures has been examined. Similar to previous reports for other solutions (Thomas et al., 1995; Zeilinger et al., 2002; Serralta et al., 2003), notable differences in basal P450 expression were found between human hepatocytes prepared from CS-preserved liver grafts and cultures prepared from other sources of liver tissue. Compared with hepatocytes obtained from SLRs, reduced levels of specific mRNAs and catalytic activities of the major P450 enzymes were observed in hepatocyte cultures from the CS group (Tables 2 and 3). Because of this, a lower drug oxidative capability, monitored as testosterone oxidative metabolism, was also found (Fig. 2). However, no differences in the relative testosterone metabolic profile were observed, suggesting a similar pattern of P450 enzymes in both groups of hepatocyte cultures.

In vitro studies using human hepatocytes are usually interpreted without consideration of the type of sample used for cell harvesting. A general lower viability as well as yield of isolation procedure and functionality have been reported in hepatocytes prepared from organ donor livers as compared with cells isolated from SLRs (Vons et al., 1990; Thomas et al., 1995; Zeilinger et al., 2002; Serralta et al., 2003). A major difference in the procurement of both groups of liver samples is that livers from organ donors are perfused in situ with a cold preservation solution. A few previous reports have analyzed metabolic competence of human hepatocyte cultures obtained from liver grafts preserved in UW or Euro-Collins solutions (Thomas et al., 1995; Olinga et al., 1998a; Zeilinger et al., 2002; Serralta et al., 2003), but this is, to our knowledge, the first study in which P450 enzymes have been examined in hepatocytes from CS-flushed liver grafts. Hypothermic preservation of hepatocytes and subsequent normothermic culturing can induce necrosis and DNA fragmentation (Abrahamse et al., 2003). CS has been proved as effective in the protection of hepatocytes against cold preservation injury from ischemia and reper-fusion as UW and better than other preservation solutions (Straatsburg et al., 2002; Abrahamse et al., 2003; Janssen et al., 2003). Cold ischemia has been proposed as another possible factor involved in the efficiency of the isolation procedure and the metabolic competence of cultured cells. Cold ischemia time was <36 h for all the liver graft samples included in our study, and no correlation between P450 enzymes and the time of hypothermic preservation was found. These results are in agreement with previous observations showing that human liver, isolated hepatocytes, or liver slices can be efficiently cold-stored for several hours with minor losses of functionality (Caraceni et al., 1994; Dorko et al., 1994; Olinga et al., 1998b).

In addition to factors directly derived from differences in the procurement and conditions of tissue preservation before cell dissociation, aspects related to intrinsic characteristics of donors must also be considered. Donor age, pathological status of the liver, and drug intake can influence the viability and functionality of human hepatocytes in primary culture. The observed P450 differences between hepatocytes from both groups cannot be attributed to age differences (ages of donor averaged 47 ± 20 and 55 ± 16 for liver grafts and SLR samples, respectively). The healthy/pathological status of the liver is likely a more relevant factor, as it is recognized that different liver diseases alter P450 expression and metabolic capacity for drugs that are metabolized by oxidative pathways (Bastien et al., 2000; Renton, 2001; Orlando et al., 2003). All liver samples used in our study were obtained from donors who were not suspected of harboring any infectious disease and tested negative for human HIV and viral hepatitis, and no malignant tissue pathology was present in any of the cases (confirmed by the pathology laboratory). Fat liver accumulation has been associated with a decrease in activity and enzyme expression of several P450s (Leclercq et al., 1998; Weltman et al., 1998). None of the liver samples included in the present study show severe steatosis, and only liver 1 presented moderate steatosis (10-30%, pathologist confirmation). Interestingly, hepatocyte cultures from these liver grafts showed the lowest P450 activity levels (Fig. 1).

Upon repeated administration, certain drugs can alter their own metabolism or that of other therapeutic agents by increasing the expression of P450 enzymes. Metabolic interactions due to enzyme induction are much less frequent than those caused by inhibition; however, their consequences can be clinically relevant. Most P450s involved in drug metabolism are inducible, including members of CYP1A, CYP2A, CYP2B, CYP2C, CYP2E, and CYP3A subfamilies. Nowadays, primary human hepatocytes and liver slices are the only in vitro models for global examination of inductive potential of new drugs. High variability in the responsiveness to inducers is found between different human hepatocyte preparations (LeCluyse 2000; Edwards et al., 2003; Madan et al., 2003). Genetics, drug treatment, smoking habits, or alcohol intake are possible factors responsible for these differences. Another important factor is the quality of the tissue and, subsequently, of the cell preparation. To control interindividual variability, known model inducers are included in each cell preparation as positive controls of induction. The response of P450 enzymes to prototypical inducers has been comparatively studied in hepatocytes from CS-preserved liver grafts and from SLR samples. A similar pattern of P450 induction, in terms of magnitude, reproducibility, and specificity of the response, was found. Independently of the source of liver tissue used for cell isolation, the specificity of the response of hepatocytes to P450 inducers was reproducible from one cell preparation to another. MC produced a selective inducer of CYP1A enzymes (Fig. 5). In contrast, PB and RIF produced effects on several P450s (Fig. 5). Both chemicals highly increased CYP2A6, 2B6, and 3A4 mRNA content, whereas lower effects were produced on CYP2C9, 2C19, or 3A5 or 1A1 mRNAs. These results are in agreement with those previously reported for specific P450 activities or apoprotein levels (Donato et al., 1995; Meunier et al., 2000; Edwards et al., 2003, Madan et al., 2003).

Our results suggest that hepatocyte cultures prepared from CS-flushed liver grafts can be applied to drug metabolism studies, despite their low P450 expression compared with hepatocytes from SLR samples. These findings become very interesting as human liver tissue has become increasingly available for scientific research, due to the expansion of liver transplantation programs. Moreover, the demand for the clinical use of human hepatocytes is increasing. Large quantities of healthy differentiated cells are needed for therapeutic applications (i.e., hepatocyte transplantation and bioartificial liver support systems). Viable human hepatocytes can be successfully isolated from different types of liver samples. However, livers considered unsuitable for transplantation or unused segments obtained from split/reduced grafts are the only possible sources of large pieces for massive hepatocyte isolation required for potential clinical applications. The availability of human livers for cell harvesting is still low, and methods are needed to increase the storage time of the tissue to extend its usage for both clinical and research purposes.

In conclusion, primary cultures of human hepatocytes prepared from liver grafts preserved in CS appear to be a suitable in vitro model for evaluating the metabolism of new drug entities and their potential effects as P450 inducers. This source of liver tissue renders a greater number of cells than small SLR samples, and as metabolically competent hepatocytes are obtained, they can be used for both research and clinical applications.