Enhanced Oxygen Delivery Reverses Anaerobic Metabolic States in Prolonged Sandwich Rat Hepatocyte Culture

Experimental Cell Research 246, 221–232 (1999) Article ID excr.1998.4295, available online at http://www.idealibrary.com on

Enhanced Oxygen Delivery Reverses Anaerobic Metabolic States in Prolonged Sandwich Rat Hepatocyte Culture A. Bader,* ,1 N. Fru¨hauf,* M. Tiedge,† M. Drinkgern,† L. De Bartolo,‡ J. T. Borlak,* G. Steinhoff,* and A. Haverich* *Leibniz Laboratories of Biotechnology and Artificial Organs and †Abteilung fu¨r Klinische Biochemie, Medizinische Hochschule Hannover, 30659 Hannover, Germany; and ‡Research Institute on Membranes and Modelling of Chemical Reactors, (IRMERC), CNR, at Department of Chemical and Materials Engineering, University of Calabria, Arcavata di Rende (CS), Italy

It must be assumed that current petri dish primary hepatocyte culture models do not supply sufficient amounts of oxygen and thus cause anaerobic metabolism of the cells. This is contrary to the physiologic state of the cells. In vivo the liver is a highly vascularized organ with a rather high blood flow rate of a mixture of arterial and venous blood. The aim of the present study was to show the oxygen dependence of primary rat hepatocytes in long-term culture and to define appropriate conditions that could allow hepatocytes to maintain tissue specific functions in an aerobic environment. To this purpose matrix overlaid hepatocytes were either cultured on gas-permeable (fluorinated hydrocarbon films) or gas-impermeable (polystyrene) supports at 10% and 20% ambient oxygen concentration (v/v), respectively. Tissue-specific functions were assessed by studying albumin and urea secretion as well as xenobiotic metabolism. The mRNA expression and catalytic activities of the cytoprotective antioxidant enzymes mitochondrial manganese superoxide dismutase (MnSOD), cytosolic copper and zinc superoxide dismutase, peroxisomal catalase, and cytosolic glutathione peroxidase were investigated to assess intracellular responses to the defined variations in oxygen supply. Hepatocytes could successfully be maintained at aerobic conditions in long-term culture on gas-permeable PTFE films. At 50% (10%, v/v) of currently used oxygen levels lactate accumulation was prevented, a plateau-like albumin secretion reestablished, urea secretion improved, and xenobiotic metabolism proceeded at physiological rates. mRNA expression of cytoprotective enzymes responded to the pericellular availability of oxygen and was most pronounced in the case of MnSOD. However, the biggest stress factor for the hepatocytes still appeared to be the isolation procedure, as mRNA expression and catalytic activities were most elevated shortly thereafter. 1

To whom correspondence and reprint requests should be addressed at Leibniz Laboratories of Biotechnology and Artificial Organs, Medical School of Hannover, Podbielskistrasse 380, 30659 Hannover, Germany. Fax: 149 (511) 906 3569. E-mail: abader@ artificial-organs.de.

In conclusion, this study clearly shows the oxygen dependence of primary rat hepatocytes in long-term culture and indicates means to establish appropriate conditions for the aerobic culture of primary rat sandwich hepatocytes with full maintenance of function. The long-term culture of hepatocytes on oxygenating supports at in vivo-like oxygen tensions therefore appears to be more physiologic and beneficial for the cells. © 1999 Academic Press Key Words: oxygen; culture; albumin; urea; lactate; cytoprotective enzymes; monoxygenases; cytochromes; artificial liver.

INTRODUCTION

In primary hepatocyte cultures oxygen supply is a critical issue since it must be assumed that the highly oxygen demanding hepatocytes are generally maintained in petri dishes under oxygen-deficient culture conditions and are thus forced into anaerobic metabolic states [1]. In these models oxygen consumption does not only depend on hepatocellular uptake rates but is limited by culture medium thickness as well as ambient oxygen concentration. Despite this limitation hepatocytes generally tolerate hypoxia due to their extraordinary capacity to satisfy energy requirements by anaerobic glycolysis. This, however, results in an inefficient utilization of glucose since the conversion of glucose to lactate leads to the generation of 2 mol ATP/mol glucose compared to 38 mol ATP/mol glucose during oxidative phosphorylation [2]. As early as 1968 it was shown that the commonly used medium depths of 2–5 mm in petri dishes rapidly produced hypoxia, if hepatocytes would respire at their physiological rate [3]. Wallach and Sherwood [1] have shown by mathematical modeling that pericellular P O2 levels would approach 0 with a medium layer thickness of just 1 mm. Thus, to maintain physiological oxygen supply in cell culture is not a trivial issue, considering the fact that plastic walls and culture medium are efficient barriers of oxygen diffusion [4, 5].

221

0014-4827/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.

222

BADER ET AL.

The development of organotypical culture techniques therefore necessitates a detailed understanding of oxygen requirements of primary hepatocytes in vitro including knowledge of the hepatocellular threshold to oxygen toxicity. It must be noted that indeed aerobic metabolic states lay right in between too much or too little oxygen. Oxygen balance is thus readily inclined to reach either hypoxia or hyperoxia. Our own morphological observations in a three-dimensional coculture model of primary rat hepatocytes and sinusoidal cells indicated improved attachment and spreading of hepatocytes when cultured on gaspermeable membranes in normal air (160 mm Hg, 5%, CO 2) but also signs of oxygen toxicity after prolonged periods [6]. This model was utilizing hepatocytes cultured in a sandwich configuration, which reconstructs the matrix geometry of the liver in vitro. Three-dimensional matrix reconstruction had been shown to be beneficial for expression of a physiologic morphology and function of hepatocytes in vitro in a number of studies [7–14]. These developments had not been available in earlier studies using hepatocytes on membranes in which oxygen supply was set to room air with 5% CO 2 in short-term studies [15]. Short-term studies indicated the relevance of oxygen for attachment and spreading of cells and to this purpose elevations of the ambient P O2 up to 540 mm Hg for cultures on gasimpermeable supports were proposed [16]. The expectation of anaerobic metabolic states of hepatocytes according to the modeling of Wallach and Sherwood [1], and the presence of a second matrix layer covering the cells in the conventional sandwich hepatocyte model, gave us second thoughts about the utility of currently used oxygen concentrations in the air phase surrounding these cultures in standard petri dishes in a long-term situation. Our hypothesis was that oxygen demand would be highest during the attachment period and closer to physiologic levels in long-term culture. The aim of the present study was therefore to exactly define for the first time oxygen concentrations to achieve a physiologic aerobic long-term culture of sandwich hepatocytes. The advantage of the gas-permeable membrane support was in this context the independence from a superimposed culture medium height, which is rather variable. In this study lactate production was used as a key indicator for anaerobic glycolysis. mRNA expression and catalytic activities of cytoprotective enzymes were studied as markers for the intracellular response to controlled changes in oxygen supply. Albumin and urea secretion as well as the monoxygenase-dependent biotransformation of a xenobiotic (urapidil) were investigated to evaluate the dependence of tissue-specific functions in anaerobic and aerobic states.

MATERIALS AND METHODS Chemicals. Unless otherwise stated all chemicals were of highest purity. Collagenase, penicillin, and streptomycin were obtained from Biochrom, Berlin, Germany, whereas all other cell culture reagents, as listed below, were purchased from Gibco, Weinheim, Germany. Percoll was from Pharmacia, Heidelberg, Germany. Glucagon was from Novo, Mainz, Germany, and insulin was from Hoechst, Frankfurt, Germany. H 2O 2 was bought from Sigma, Deisenhofen, Germany. Guanidine thiocyanate was obtained from Fluka (Neu-Ulm, Germany). All other molecular biology reagents were from Merck (Darmstadt, Germany). Isolation of rat hepatocytes. Cells were isolated from female Lewis rats (200 –250 g) by a modification of the two-step method described by Seglen [17]. Rats were anesthetized with ether. Following midline incision and cannulation of the portal vein, the liver was perfused in situ with 400 ml Ca 21-free Krebs ringer buffer (KRB) for 10 min followed by perfusion of 200 ml KRB containing 21.5 U (100 mg/200 ml) collagenase and 478 mM CaCl 2 for 8 min. The collagenase perfusate was not recirculated. All perfusates were equilibrated with 95% oxygen and 5% CO 2 at 37°C. Following perfusion of the liver, the liver capsule was gently removed and the dissolved liver tissue was washed three times with ice-cold KRB after filtration through a nylon mesh (100 mm). The parenchymal cell suspension was purified using a Percoll gradient centrifugation. The Percoll gradient was prepared by adding 1.2 ml of Hanks’ balanced salt solution to 10.8 ml Percoll. This was followed by addition of 12.5 ml washing buffer containing one-fourth of the cell pellet obtained from the previous steps. After gentle mixing of the resulting cell suspension, cells were centrifuged again at 50g for 5 min at 4°C. The remaining cell pellet was used for the following experiments. Viability of the cells ranged between 85 and 98% as assessed by the trypan blue exclusion test. On average the yield from one isolation procedure was 3–5 3 10 8 cells. Cell culture technique. Hepatocytes were cultured either on supports made of polystyrene (P) not permeable for gases or in petri dishes with a homogeneous oxygen-permeable flurocarbon (PTFE) film (M) at the bottom. Hepatocytes were seeded at a density of 2 3 10 6 cells at 20% ambient oxygen concentration for all groups, to ensure equal attachment conditions within the P and within the M groups, respectively. Following 24 h in culture four groups were prepared each consisting of cultures at 10 or 20% ambient oxygen concentration (P10, P20, M10, and M20), respectively. Primary rat hepatocytes were always embedded within two layers of collagen in a modification [9] of the method described previously by Hall et al. [18] and Dunn et al. [19]. Rat tail collagen was prepared according to the method of Elsdale and Bard [20]. A final concentration of 1.5 mg/ml collagen was used for coating petri dishes (60 mm diameter, Greiner, Frickenhausen, Germany). Four hours following seeding and attachment of the cells, culture medium was removed along with nonadherent cells and a second layer of 1 ml liquid and ice-cold collagen was pipetted on top of the hepatocytes. After gelation of this second matrix layer, a sandwich configuration with two layers of hydrated collagen gel was obtained to simulate the in vivo bipolar hepatocellular enclosure within the matrix of the space of Disse. Within 30 min after gelation of the second layer of matrix, culture medium was applied. The pH of the collagen was adjusted to 7.4 using a 103 Dulbecco’s modified Eagle medium concentrate, which was diluted with the collagen solution at a ratio of 1:10. Culture media were supplemented with 10% (v/v) fetal bovine serum, 9.6 mg/ml prednisolone, 0.014 mg/ml glucagon, and 0.16 U/ml insulin. Penicillin (200 U/ml) and 200 mg/ml streptomycin were also added. The culture medium was replaced with fresh medium every 24 h. Cell numbers during the cultivation periods were assessed by analysis of four representative spots on the petri dishes, using predefined observation grids on a Televal inverse microscope from Zeiss, Ger-

LONG-TERM CULTURE OF HEPATOCYTES ON OXYGENATING SUPPORTS many. Counting was performed on days 0, 1, 3, 7, and 14 in culture in parallel to the analytical studies. For all consecutive studies hepatocytes were divided in two groups consisting of matrix overlaid hepatocytes cultured at either 10% or 20% of ambient oxygen concentration (v/v). These groups were again subdivided into two subgroups of cells cultured either on standard gas-impermeable polystyrene (P) petri dishes (Greiner, Germany) or on gas-permeable (M) PTFE films (Petriperm petri dishes, Hereaus, Osterode, Germany). Thus the resulting four individual subgroups were defined as P10, P20, M10, and M20 respectively. All cultures were performed at 37°C and 5% CO 2. Albumin secretion studies. Albumin secretion was determined immunochemically with an ELISA technique [7], using supernatants collected on a daily basis for 2 weeks. Cell culture supernatants were collected every 24 h up to 14 days in culture and stored at 4°C for further analysis. Chromatographically purified rat albumin and the monoclonal antibody for rat albumin were purchased from Cappel (Durham, NC). Ninety-six-well plates (Nunc, Wiesbaden, Germany) were coated with albumin and left overnight at 4°C. The coating buffer contained 1 mg/ml albumin [7]. After washing the plate four times, 50 ml of cell culture supernatant was added to the wells and incubated with 50 ml of anti-rat albumin antibody conjugated to horseradish peroxidase. After 24 h at 4°C, substrate buffer containing O-phenylenediamine dihydrochloride and H 2O 2 was added for 6 min. The reaction was stopped with 100 ml of 8 N H 2SO 4. Absorbance was measured at 490 nm using a Dynatech M5000 spectrophotometer. The standard curve ranged between 1 and 100 mg albumin/ml. Urea secretion studies. Colorimetric analysis of urea was performed using a Vitras 950 IRL autochemical diagnostics from Johnson and Johnson. Cell culture supernatants were collected every 24 h up to 14 days in culture and analyzed daily. The experiments were repeated three times with cells from three different isolations. Data are shown in millimoles according to international SI classifications. Lactate measurements. Lactate was measured in the supernatant of cultures maintained at 10 and 20% ambient oxygen concentration either on gas-permeable films or on standard polystyrene petri dishes on day 1, 3, 7, and 14 of culture. Lactate concentration was determined with an automated analyzer according to the manufacturing recommendations (Precision Instruments, U.S.A.). The experiments were repeated three times using cells from three individual isolations. mRNA expression of antioxidant enzymes. Restriction enzymes, SP6/T7 transcription kits, and digoxigenin (DIG) nucleic acid detection kits were purchased from Boehringer (Mannheim, Germany). Hybridization transfer films Hybond N were supplied by Amersham (Braunschweig, Germany). The cDNAs coding for rat CuZn superoxide dismutase (CuZnSOD), rat mitochondrial Mn superoxide dismutase (MnSOD), and rat glutathione peroxidase were kindly provided by Dr. Ye-Shih Ho (Detroit, MI). The human (cross-reacting with rat) catalase and actin cDNA were obtained from the American Tissue Culture Collection (Rockville, MD). mRNA expression of MnSOD, CuZnSOD, glutathione peroxidase, catalase, and actin was measured by Northern blotting of P and M cultures. Cultures were maintained at 10% or 20% ambient oxygen concentration (v/v) for 2 weeks. Each group consisted of 40 dishes. Cells were harvested for mRNA extraction on day 0 (3 h following seeding) and on days 1, 3, 7, and 14 in culture. Freshly isolated cells served as controls. Total RNA was extracted according to Chomczynski and Sacchi [21]. Five micrograms of total RNA was loaded per lane and mRNA levels were compared to ribosomal bands and visualized in an ethidium bromidestained agarose gel. Total RNA was separated on formamide/formaldehyde agarose gels and transferred to nylon films. Hybridization was done as described previously [22] using DIG-labeled probes coding for CuZnSOD, MnSOD, glutathione peroxidase, and catalase. Bands were visualized by chemiluminescence on autoradiography films and were quantified densitometrically using NIH Image 1.52 analysis program.

223

Antioxidant enzyme activities. Activities of total MnSOD, CuZnSOD, glutathione peroxidase, and catalase were measured in cultures maintained at 10 or 20% ambient oxygen concentration (v/v) up to 2 weeks. Each group consisted of six dishes. Experiments were repeated three times. Cells were harvested and pooled on day 0 (3 h following seeding) and on days 1, 3, 7, and 14 in culture. Freshly isolated cells served as controls. Cell cultures were homogenized in 50 mM potassium phosphate buffer (pH 7.8). The homogenates were sonicated on ice for 1 min in 15-s bursts at 90 W with a Braun-Sonic 125 sonifier. Thereafter the homogenates were centrifuged at 35,000g and 4°C for 40 min and the supernatant was stored at 220°C until measurement. Superoxide dismutase activities were measured in a photometric assay according to [23] using xanthine oxidase as a source of O 22 and NBT. The activities of the tissues were plotted against a standard curve of purified superoxide dismutase from bovine liver. Addition of 5 mM NaCN to the samples specifically inhibited CuZnSOD. Subtraction of SOD activity after NaCN treatment from total SOD activity gave the MnSOD activity of the sample. One unit of activity was defined as the amount of superoxide dismutase protein which gives a half-maximal inhibition of NBT reduction. Catalase activity was measured by ultraviolet spectroscopy monitoring the decomposition of hydrogen peroxide at 240 nm [24]. One unit of catalase activity was defined as micromoles of hydrogen peroxide consumed per minute at 25°C. Glutathione peroxidase activity was measured in a photometric assay at 37°C using glutathione reductase and NADPH in a coupled reaction [25]. The decrease of NADPH absorbance was monitored at 340 nm. Activities were calculated against a glutathione peroxidase standard according to [26] and expressed as units per milligram of protein. Biotransformation studies with urapidil. Urapidil was added at days 1, 3, 7, and 14 in culture at a concentration of 2.5 mg/ml for a period of 24 h. At the end of the incubation period the cell culture supernatant and the cell matrix phase were aspirated from the petri dishes. All samples were immediately frozen and kept at 220°C. Fifteen cultures were prepared for each individual day. Six cultures served as controls and did not receive urapidil. Urapidil and its metabolites were analyzed by high-performance liquid chromatography according to the method of Zech et al. [27]. This method employed a better extraction method when compared to the liquid– liquid extraction method previously described [28]. Urapidil and metabolites M2 and M3 were detected with a lower limit of quantification of 5 ng/ml, and for M1 a lower limit of quantification was determined at 15 ng/ml. The chemical structure of urapidil metabolites was previously reported [29] and synthetic standards of metabolites were synthesized by Dr. Pru¨sse, Research Laboratories of Byk Gulden, Konstanz, Germany. Chemically synthesized standards of urapidil metabolites were used for identification and quantification purposes. Phase II metabolites are not detected by this method.

RESULTS

Morphology P10, P20, M10, and M20 cultures were kept for 14 days in culture (Figs. 1A–1D). Light microscopic evaluation shows that all cultures had formed polyhedral cells with surrounding biliary canaliculi. In P10 cultures, however, disintegrating cells were noted and in P20 and M20 cultures a subconfluent arrangement was observed. Only M10 cultures maintained an intact monolayer of polyhedric cells with distinct pericellular biliary zones even 14 days postseeding (in culture) (Fig. 1C).

224

BADER ET AL.

FIG. 1. Light microscopy of P10 (A), P20 (B), M10 (C), and M20 (D) cultures. Bar, 100 mm. M10 cultures are shown as confluent cultures.

Lactate Formation

Albumin Secretion

Lactate production was measured in the supernatant of matrix overlaid hepatocyte on days 1, 3, 7, and 14. When hepatocytes were cultured on gas-impermeable supports (P) high lactate concentrations were measured at the end of the 24-h incubation intervals. P10 cultures had highest lactate concentrations until the end of the experiment (Fig. 2). Lactate levels ranged between 0.24 6 0.01 (day 3) and 0.27 6 0.02 mg/ml at day 14. Lactate levels for M10 and M20 cultures ranged from 0 6 0.01 to 0.03 6 0.02 mg/ml.

Standard hepatocyte cultures on gas-impermeable supports at 20% ambient oxygen supply (P20) served as controls. For these cultures a typical increase in albumin secretion was observed following the first 4 –5 days in culture (Fig. 3A). The cell-specific albumin secretion rate at 1 week in culture corresponded to 120 –144 pg/cell/24 h with P20 cultures. With P10 cultures at a 10% ambient oxygen concentration albumin secretion did not increase (Fig. 3A). Also M20 cultures at 20% oxygen (v/v) did not produce the typical albumin

LONG-TERM CULTURE OF HEPATOCYTES ON OXYGENATING SUPPORTS

225

FIG. 1—Continued

secretion pattern as seen with P20 cultures (Fig. 3B). Matrix overlaid hepatocytes at 10% oxygen (v/v) cultured on films (M10) increased the albumin secretion similar to the P20 cultures. Urea Secretion Oxygen dependence of urea secretion is shown in Figs. 4A and 4B. Rat hepatocytes cultured on gasimpermeable supports at 10% ambient oxygen supply (P10) had lowest urea secretory levels over the whole

culture period and the steepest drop in secretory rates if compared with the initial rate (34 6 7%). P20 cultures generally expressed higher secretory rates compared to P10 cultures (Fig. 4A) but urea secretion still dropped below 3 mmol/liter within 4 days in culture (62 6 6%). Membrane cultures, receiving 10% or 20% oxygen up to 14 days in culture, showed best maintenance and highest level of urea secretory capacity over the full culture period, while no major difference was observed among M10 (76.7%) and M20 (83.4%; Fig.

226

BADER ET AL.

compound (6-{3-[4-(o-hydroxyphenyl)piperazinyl]-propylamino}-1,3 dimethymethyluracil), and an uracil–Ndemethylated compound (6-{3-[4-(o-methoxyphenyl)piperazinyl]-propylamino}-1-methyluracil). The production of these metabolites among the various cultures was compared. P10 cultures had the highest parent compound concentrations following a 24-h incubation at substrate concentrations of 2.5 mg/ml on day 3 and day 7 in culture (Fig. 5A). In consequence, the metabolite concentrations were lowest in extracts of these cultures on days 3 and 7. P20 cultures and the gas-permeable film cultures (M10, M20) were similar among each other (Figs. 5B–5D). In all four groups the production of metabolites increased continuously with proceeding culture time. The ratio of the metabolites was similar under the different conditions. mRNA Expression of Cytoprotective Enzymes FIG. 2. Lactate concentrations in the supernatant of P10, P20, M10, and M20 cultures. Supernatant was replaced with fresh culture medium every 24 h. Cells were maintained in culture for 2 weeks. Data are shown as means 6 standard error. The mean of six measurements for each time point is given. Cells originated from at least three different isolations.

4B). Supernatant concentrations never dropped below 3 mmol/liter. Biotransformation of Urapidil Urapidil was metabolized to at least three metabolites and included a hydroxylated product (6-{3-[4-(omethoxy-p-hydroxyphenyl)piperazinyl]-propylamino}1,3-dimethymethyluracil), and an O-demethylated

The expression of cytoprotective enzymes was studied in liver tissue, in freshly isolated hepatocytes, and in the P10, P20, M10, and M20 groups following seeding on collagen-coated surfaces with gas-impermeable (P) and gas-permeable (M) surfaces and on days 1, 3, 7, and 14 in culture. Attachment of cells was permitted at identical ambient oxygen concentrations (20% v/v; Table 1). The mRNA expression of constitutively expressed actin served as a control. Actin mRNA expression did not differ among the various culture conditions. However, its expression was increased with time in culture (Fig. 6A). The level of mRNA expression of cytoprotective enzymes in freshly isolated cells was generally higher than in native liver extracts (Table 1) and the

FIG. 3. Albumin concentrations in the supernatant of sandwich rat hepatocytes cultured for 2 weeks at 10 or 20% ambient oxygen concentration (v/v) on gas-impermeable (A) or gas-permeable (B) supports. Data are shown as means 6 standard error of the mean of at least six cultures from three different isolations.

LONG-TERM CULTURE OF HEPATOCYTES ON OXYGENATING SUPPORTS

227

FIG. 4. Urea concentrations in the supernatant of sandwich rat hepatocytes cultured for 2 weeks at 10 or 20% ambient oxygen concentration (v/v) on gas-impermeable (A) or gas-permeable (B) supports. Data are shown as means 6 standard error of the mean of at least three cultures from three different isolations.

mRNA levels determined in freshly isolated cells were defined as the 100% value (Table 1). The mRNA expression of the mitochondrial enzyme MnSOD was increased when compared with that of CuZnSOD, catalase (CAT), and glutathione peroxidase (GPX) (Table 1, Fig. 6B). The mRNA expression of MnSOD surpassed that of freshly isolated cells in both M10 and M20 cultures and 24 h postseeding the levels of expression were close to 200% and remained at this level throughout the study period. The MnSOD mRNA expression was higher in M20 than M10 groups and, similarly, P20 cultures tended to express higher levels of MnSOD mRNA when compared with the P10 group. The order of expression was M20 . M10 . P20 . P10. P20 cultures expressed MnSOD mRNA at approximately 89 –90% of freshly isolated cells at days 7 and 14 in culture. In contrast, the MnSOD mRNA expression in P10 cultures was reduced to 50% at day 14 in comparison with freshly isolated cells (Table 1). The mRNA expression of the cytosolic CuZnSOD enzyme was similar to that of MnSOD. However, differences between the various groups were not as evident as noted for MnSOD. After 2 weeks in culture the CuZnSOD mRNA expression was higher in M20 cultures than in the remaining groups (M20 . M10 . P20 . P10, Table 1, Fig. 6C), but did not surpass that of freshly isolated cells. Indeed the highest expression was observed in freshly isolated cells and during the attachment periods for gas-permeable film and plastic cultures, as shown in Table 1. Gas-permeable film cultures had higher mRNA expression at the end of the attachment period. Thereafter CuZnSOD mRNA expression was reduced to 37% in P10 cultures and to 58% in P20 cultures. CuZnSOD mRNA expression in

M10 cultures was 67 and 77% in M20 cultures at day 14 postseeding. Expression of catalase mRNA was highest in freshly isolated cells and 3 h following seeding and attachment of cells. Initially the expression of catalase was similar between both P and M cultures (Table 1, Fig. 6D), but declined within 24 h. Catalase mRNA expression was higher with increasing oxygen availability and followed the order M20 . M10 . P20 . P10). At the end of the cell attachment period the mRNA expression of GPX in the M group was comparable to that of freshly isolated cells. With P cultures a decline to 86% was noted (Table 1, Fig. 6E). Thereafter, a continuous decline in GPX expression occurred in all groups with GPX levels ranging between 23 and 35% on day 7. The glutathione peroxidase mRNA expression was comparable among the P10, P20, M10, and M20 cultures. Antioxidant Activities Activities of cytoprotective enzymes were studied using the same conditions for the same type of groups as done for mRNA studies. Activities were expressed in units per milligram of cell protein. Total SOD activity was highest in M20 groups during the course of culture. This was similar for the mRNA expression found in this study. Comparable differences among the P10, P20, M10, and M20 groups for GPX and CAT activity during the first and second week in culture were not found, which again corresponded to the lack of striking differences for these two enzymes with respect to mRNA expression. However, as shown in Table 2, CAT activity was highest at the end of the attachment phase and in freshly isolated cells. Initially the expression of

228

BADER ET AL.

FIG. 5. Biotransformation of urapidil in primary rat hepatocytes cultured on gas-impermeable supports (A,B polystyrene) at 10% (P10) and 20% (P20) and on gas-permeable supports (C,D, PTFE gas-permeable film, M10, M20) in a sandwich configuration. Urapidil and its metabolites were analyzed by HPLC. Data indicate the mean 6 SEM of 12 cultures for each time point originated from three isolations. Metabolite production is shown as the sum of three metabolites. M1 is a hydroxylated product (6-{3-[4-(o-methoxy-p-hydroxyphenyl)piperazinyl]-propylamino}-1,3-dimethyluracil), M2 is an O-demethylated compound (6-{3-[4-(o-hydroxyphenyl)piperazinyl]-propylamino}-1,3-dimethylmethyluracil), and M3 is a uracil–N-demethylated compound (6-{3-[4-(o-methoxyphenyl)-piperazinyl]-propylamino}-1-methyluracil).

catalase was similar between both P and M cultures (Table 2), but declined within 24 h. This reflects also the observed results found for mRNA expression of this enzyme. DISCUSSION

In this study we defined aerobic culture conditions for primary rat sandwich hepatocytes and showed the oxygen dependence of tissue-specific functions. Only the cultures maintained on gas-permeable films at an ambient oxygen concentration of 10% (v/v) exclusively produced high amounts of albumin, urea, and drug metabolites over the entire study period (14 days) and simultaneously did not accumulate lactate. The aim of the study was to achieve a practical definition of required ambient oxygen supply conditions for the long-term aerobic culture of primary hepatocytes in a sandwich configuration. In living cells cytoprotective enzymes are considered to be a sensitive intracellular defense system against oxygen-induced damage, as cytoprotective enzymes in-

activate free radicals and prevent unscheduled oxidations. In particular, superoxide radical anions are converted by a family of metal enzymes, termed superoxide dismutases. In eukaryotic cells two types of enzymes are expressed, e.g., the cytosolic copper- and zinc-containing superoxide dismutase (CuZnSOD) and a manganese-containing mitochondrial superoxide dismutase (MnSOD). The superoxide radicals generated for instance by xanthine oxidase are rapidly converted into O 2 and H 2O 2. CAT, a tetrameric hemoprotein, catalyzes the conversion of H 2O 2. GPX is a selen-containing tetrameric protein, which catalyzes the reduction of hydrogen peroxide by oxidation of glutathione. The latter two enzymes keep the hydrogen peroxide concentration within the range of 10 27 and 10 29 mol/ liter [30] as oxygen radicals and other reactive species can destroy proteins and nucleic acids by oxidation [31]. The expression levels of cytoprotective enzymes were used as markers to avoid a selection of oxidizing and thus potentially toxic culture conditions. Freshly isolated cells had raised mRNA levels of

229

LONG-TERM CULTURE OF HEPATOCYTES ON OXYGENATING SUPPORTS

TABLE 1 Northern Blot Analysis of Cytoprotective Enzyme Expression mRNA expression levels (%) Time/enzyme MnSOD Day 0 Day 0 (3 h) Day 1 Day 3 Day 7 Day 14 CuZnSOD Day 0 Day 0 (3 h) Day 1 Day 3 Day 7 Day 14 CAT Day 0 Day 0 (3 h) Day 1 Day 3 Day 7 Day 14 GPX Day 0 Day 0 (3 h) Day 1 Day 3 Day 7 Day 14

Liver

FC

93 6 1

100

86 6 6

62 6 21

69 6 3

P10

P20

M10

M20

100 6 15 76 6 11 77 6 9 51 6 6

51 6 6 121 6 17 66 6 4 90 6 6 89 6 15

65 6 6 68 6 7 60 6 5 37 6 10

67 6 12 53 6 1 80 6 5 68 6 3 58 6 8

58 6 63 6 70 6 67 6

8 4 7 4

87 6 15 62 6 4 70 6 5 65 6 5 77 6 2

86 12 6 14 6 13 6

4 4 5 4

74 6 12 66 3 21 6 7 13 6 3 22 6 5

21 6 22 6 28 6 34 6

3 6 7 1

87 6 16 6 28 6 39 6 29 6

5 4 5 1 2

45 6 49 6 25 6 13 6

6 9 5 5

86 6 6 61 6 8 70 6 10 35 6 11 13 6 4

45 6 40 6 24 6 15 6

7 3 3 2

102 6 46 6 52 6 23 6 16 6

5 5 4 3 4

80 6 20 164 6 15 185 6 15 180 6 15

156 6 15 199 6 17 215 6 9 231 6 9 235 6 19

100

100

100

Note. Data are normalized with reference to freshly isolated cells (100%). P10 is identical to P20, and M10 is identical to M20 during the attachment phase since all groups attached at 20% oxygen (v/v). Normal liver was obtained from separate animals. Data are shown as means 6 SEM of at least three to six experiments. Image analysis was performed using NIH Image 1.53 software.

cytoprotective enzymes when compared with native liver tissue. This increase in mRNA expression may be attributed to the trauma of cell isolation. Thereafter the expression declined well below this initial value. In contrast the mRNA expression of MnSOD was above the level of freshly isolated cells. Indeed an up to fivefold difference in mRNA expression could be observed when P10 cultures were compared with M20 cultures at day 14 in culture. The mRNA expression of actin did not differ among P and M cultures, as expected for this constitutively expressed gene. This study shows that the culture of primary rat hepatocytes on gas-permeable supports led to the establishment of an aerobic primary hepatocyte culture model thereby abolishing the unwanted and unphysiologic accumulation of lactate. As shown for the mRNA expression of MnSOD, CuZnSOD, CAT, and GPX, these cytoprotective enzymes were accordingly modified in culture responding to the improved oxygen availability. Catalytic activities tended to be raised for total SOD in M20 cultures. This reflects the increase in MnSOD mRNA in these cultures.

Our findings exemplify that only cultures on gaspermeable PTFE films which were kept at reduced ambient oxygen concentrations (M10) had a typical plateau-like albumin secretory pattern and high monooxygenase activity without evidence of an in vitro accumulation of lactate. In contrast M20 cultures had signs of oxygen toxicity, as judged by a reduced production of albumin, which was most pronounced during the second week in culture. The decrease in tissuespecific functions in M20 cultures was paralleled by highest levels of SOD transcriptory rates and catalytic activities, which may indicate oxidative stress. P10 cultures had the lowest monooxygenase activity and the lowest albumin, lowest urea, and highest lactate levels. The model drug urapidil, which was used in this study, had been shown previously to be metabolized in a species-dependent manner in short-term cultures (3 days) of rat and human sandwich hepatocytes [9]. Catalytic activity of urapidil appeared to be a less sensitive parameter than albumin and urea secretion to hypoxic or hyperoxic conditions, since it was still relatively well maintained in all P and M groups over

230

BADER ET AL.

FIG. 6. mRNA expression of actin (A) and cytoprotective enzyme expression (B–E) in long-term cultures of sandwiched rat hepatocytes. 10 mg total RNA was loaded on each lane. I, gas-permeable films; lane 1, freshly isolated cells; lane 2, 3 h at 20% oxygen (v/v); lane 3, M10 at 1 day; lane 4, M20 at 1 day; lane 5, M10 at 3 days; lane 6, M20 at 3 days; lane 7, M10 at 7 days; lane 8, M20 at 7 days; lane 9, M10 at 14 days; lane 10, M20 at 14 days. II, gas-impermeable support; lane 1, freshly isolated cells; lane 2, 3 h at 20% oxygen (v/v); lane 3, P10 at 1 day; lane 4, P20 at 1 day; lane 5, P10 at 3 days; lane 6, P20 at 3 days; lane 7, P10 at 7 days; lane 8, P20 at 7 days; lane 9, P10 at 14 days; lane 10, P20 at 14 days.

14 days. Among these four groups P10 groups had the lowest metabolite formation rates. Yet only M10 cultures maintained a morphological appearance of hepatocytes similar to their in vivo appearance as shown by polyhedric cell shapes and maintenance of cell polarity with distinct bile canaliculi formation. This condition was maintained in a fully confluent arrangement over the whole study period. As shown previously for PTFE petri dishes, the actual P O2 concentration in the supernatant corresponds

to the incubator atmosphere [4]. In the present study this would equate to a P O2 at the gas-permeable film– collagen interface of approximately 160 mm Hg (M20) and of 80 mm Hg (corresponding to periportal conditions) for the M10 cultures, respectively. The P O2 of hepatocyte supernatants of gas-impermeable P20 dishes was previously reported to be 60 mm Hg and as low as 40 mm Hg for P10 cultures [4], which would correspond to a more perivenous situation. It is important to note that oxygen participates in

231

LONG-TERM CULTURE OF HEPATOCYTES ON OXYGENATING SUPPORTS

TABLE 2 Antioxidant Activity (U) per Milligram of Cell Protein Expression levels (U/mg cell protein) Time/enzyme SOD (total) Day 0 Day 0 (3 h) Day 1 Day 3 Day 7 Day 14 CAT Day 0 Day 0 (3 h) Day 1 Day 3 Day 7 Day 14 GPX Day 0 Day 0 (3 h) Day 1 Day 3 Day 7 Day 14

Liver

597

324

0.76

FC

P10

P20

M10

M20

594.8 671.0 793.0 879.4

396.4 671.0 772.0 793.0 899.8

996.3 920.1 840.0 772.7

447.3 793.0 1260.7 970.9 1138.7

124.3 58.2 57.5 94.2

319.7 115.3 73.8 53.5 75.0

142.9 83.2 44.1 71.7

394.4 130.0 59.5 58.1 73.6

655.8

253.0

0.54 0.52 0.61 0.57 0.66

0.40 0.67 0.66 0.5 0.24

0.64 0.41 0.72 0.28

0.41 0.63 0.52 0.32 0.47

Note. Normal liver was obtained from separate animals. Data are shown as means of three to six experiments. Samples were pooled for analysis.

hepatocellular differentiation [32–34]. Oxygen consumption is relevant for a variety of metabolic studies using primary hepatocytes in vitro [35, 36]. Alternative concepts to improve oxygenation of primary hepatocytes in vitro involve the use of oxygenating hollow fibers [37]. Combinations of hollow fibers with a flat plate device have been developed by culturing the cells on gas-impermeable supports and by arranging oxygenating hollow fibers in close proximity to the cells [38]. In conclusion, the use of gas-permeable films in a petri dish configuration permitted the establishment of well-defined and prolonged aerobic culture conditions for primary hepatocytes. To this purpose it was necessary to reduce ambient P O 2 to 10% (v/v), which consequently allowed high expression levels of tissue-specific functions in vitro. This confirms our hypothesis that oxygen requirements are high initially and must be reduced to physiologic levels in longterm cultures. Our findings are of general relevance for long-term hepatocyte cultures as current practice generally has neglected the aspect of oxygen supply in vitro and since petri dish systems are still the most widely used models. Our results may also be easily applied for the aerobic long-term culture of primary hepatocytes in specifically designed largescale bireactors [39].

We are grateful to the BMBF Germany for providing funding for this project.

REFERENCES 1.

Wallach, D. F., and Sherwood, P. (1976). Diffusion in tissue cultures on gas-permeable and impermeable supports. J. Theor. Biol. 56, 443– 458.

2.

Jensen, M. D. (1976). Diffusion in tissue culture on gas-permeable and impermeable supports. J. Theor. Biol. 56, 443– 458.

3.

McLimans, W. F., Blumenson, L. E., and Tunnah, K. V. (1968). Kinetics of gas diffusion in mamalian cell culture systems (IL theory). Biotech. Bioeng. 10, 741–763.

4.

Holzer, C., and Maier, P. (1987). Maintenance of periportal and pericentral oxygen tensions in primary rat hepatocyte cultures: Influence on cellular DNA and protein content monitored by flow cytometry. J. Cell. Physiol. 133, 297–304.

5.

Catapano, G., De Bartolo, L., Lombardi, C. P., and Drioli, E. (1996). The effect of oxygen transport resistances on the viability and functions of isolated rat hepatocytes. Int. J. Artif. Organs 19, 61–71.

6.

Bader, A., Knop, E., Kern, A., Boker, K., Fru¨hauf, N., Crome, O., Esselmann, H., Pape, C., Kempka, G., and Sewing, K.-F. (1996). 3-D coculture of hepatic sinusoidal cells with primary hepatocytes—Design of an organotypical model. Exp. Cell Res. 226, 223–233.

7.

Bader, A., Rinkes, I. H. B., Closs, I. E., Ryan, C. M., Toner, M., Cunningham, J. M., Tompkins, G. R., and Yarmush, M. L. (1992). A stable long-term hepatocyte culture system for studies of physiologic processes: Cytokine stimulation of the acute

232

8.

9.

10.

11.

12.

13.

14.

15.

16.

17. 18.

19.

20. 21.

22.

23.

BADER ET AL. phase response in rat and human hepatocytes. Biotechnol. Prog. 8, 219 –225. Dunn, J. C., Tompkins, R. G., and Yarmush, M. L. (1992). Hepatocytes in collagen sandwich: Evidence for transcriptional and translational regulation. J. Cell Biol. 116, 1043–1053. Bader, A., Zech, K., Crome, O., Christians, U., Pichlmayr, R., and Sewing, K.-F. (1994). Use of organotypical cultures of primary hepatocytes to analyze drug biotransformation in man and animals. Xenobiotica 24, 623– 633. Knop, E., Bader, A., Bo¨ker, K., Pichlmayr, R., and Sewing, K. F. (1995). Normal ultrastructure of primary hepatocytes depends on bipolar attachment to extracellular matrix. Anat. Rec. 242, 337–349. LeCluyse, E. L., Audus, K. L., and Hochman, J. H. (1994). Formation of extensive canalicular networks by rat hepatocytes cultured in collagen-sandwich configuration. Am. J. Physiol. 266, C1764 –C1774. Bader, A., Knop, E., Bo¨ker, K. H. W., Crome, O., Fru¨hauf, N., Gonschior, A., Christians, U., Esselman, H., Pichlmayr, R., and Sewing, K. F. (1996). Tacrolimus (FK 506) metabolism in primary rat hepatocytes depends on extracellular matrix geometry. NS Arch. Pharmacol. 353, 461– 473. Bader, A., Reimer, P., Knop, E., Christians, U., Weissleder, R., Sewing, K.-F. (1995). An organotypical in vitro model of the liver parenchyma for uptake studies of diagnostic MR receptor agents. Magn. Res. Imaging. 13, 991–996. Bader, A., Knop, E., Bo¨ker, K., Fru¨hauf, N., Schu¨ttler, W., Oldhafer, K., Pichlmayr, R., and Sewing, K.-F. (1995). A novel bioreactor design for in vitro reconstruction of in vivo liver characteristics. Artif. Organs 19, 368 –374. Gerlach, J., Kloppel, K., Stoll, P., Vienken, J., and Muller, C. (1990). Gas supply across membranes in bioreactors for hepatocyte culture Artif. Organs 14, 328 –333. Rotem, A., Toner, M., Bhatia, S., Foy, B. D., Tompkins, R. G., and Yarmush, M. L. (1994). Oxygen is a factor determining in vitro tissue assembly: Effects on attachment and spreading of hepatocytes. Biotech. Bioeng. 43, 654 – 660. Seglen, P. O. (1976). Preparation of isolated rat liver cells. Methods Cell Biol. 13, 29 – 83. Hall, H. G., Farson, D. A., and Bissell, M. J. (1982). Lumen formation by epithelial cell lines in response to collagen overlay: a morphogenetic model in culture. Proc. Natl. Acad. Sci. USA 79, 4672– 4666. Dunn, J. C., Yarmush, M. L., Koebe, H. G., and Tompkins, R. G. (1989). Hepatocyte function and extracellular matrix geometry: Long-term culture in a sandwich configuration [published erratum appears in FASEB J. 1989, 3(7), 1873]. FASEB J. 3, 174 –177. Elsdale, T., and Bard, J. (1972). Collagen substrata for studies on cell behavior. J. Cell Biol. 54, 626 – 637. Chomczynski, P., and Sacchi, N. (1987). Single-step method of RNA isolation by acid guanidinium thiocyanate–phenol– chloroform extraction. Anal. Biochem. 162, 156 –159. Lenzen, S., Drinkgern, J., and Tiedge, M. (1996). Low cytoprotective enzyme expression in pancreatic islets compared with various other mouse tissues. Free Rad. Biol. Med. 20, in press. Oberley, L. W., and Spitz, D. R. (1984). Assay of superoxide dismutase activity in tumor tissue. Methods Enzymol. 105, 457– 464.

Received February 27, 1998 Revised version received August 3, 1998

24.

Claiborne, A. (1985). Catalase activity. In “Handbook of Methods for Oxygen Radical Research” (R. A. Grennwald, Ed.), pp. 283–284, CRC Press, Boca Raton, FL.

25.

Gu¨nzler, W. A., Kremers, H., and Flohe, L. (1974). An improved coupled test procedure for glutathione peroxidase (EC 1-11-19-) in blood. Z. Klin. Chem. Klin. Biochem. 12, 444 – 448.

26.

Cornelius, J. G., Luttge, B. G., and Peck, A. B. (1993). Antioxidant enzyme activities in IDD-prone and IDD-resistant mice: A comparative study. Free Rad. Biol. Med. 14, 409 – 420.

27.

Zech, K., and Huber, R. (1986). Determination of urapidil and its metabolites in human serum and urine: Comparison of liquid–liquid and fully automated liquid–solid extraction. J. Chromatogr. 353, 351–360.

28.

Zech, K., Eltze, M., Kilian, U., Sanders, K. H., and Kolassa, N. (1984). Biotransformation of urapidil: Metabolites in serum and urine and their biological activity in vitro and in vivo. Arch. Int. Pharmacol. Ther. 272, 180 –196.

29.

Sturm, E., and Zech, K. (1984). Biotransformation of urapidil: Isolation and identification of metabolites in mouse, rat, dog and man. Biomed. Mass Spectrom. 11, 211–216.

30.

Halliwell, B., and Gutteridge, J. M. (1995). Oxygen is poisenous—An introduction to oxygen toxicity and free radicals. In “Free Radicals in Biology and Medicine” (B. Halliwell, J. M. Gutteridge, Eds.), pp. 10 –19, Clarendon Press, Oxford.

31.

Kahl, R. (1994). Leber. In “Lehrbuch der Toxikologie” (H. Marquardt and S. G. Scha¨fer, Eds.), pp. 78 –197, Wissenschaftsverlag, Mannheim/Leipzig/Wien/Zu¨rich.

32.

Jungermann, K., and Thurman, R. G. (1992). Hepatocyte heterogeneity in the metabolism of carbohydrates. Enzyme 46, 33–58.

33.

Jungermann, K. (1992). Role of intralobular compartmentation in hepatic metabolism. Diabet. Metab. 18, 81– 86.

34.

Gebhardt, R. (1992). Metabolic zonation of the liver: Regulation and implications for liver function. Pharmacol. Ther. 53, 275– 354.

35.

Henly, D. C., and Berry, M. N. (1993). Effect of palmitate concentration on the relative contributions of the beta-oxidation pathway and citric acid cycle to total O 2 consumption of isolated rat hepatocytes. Biochim. Biophys. Acta 1175, 269 – 276.

36.

Henly, D. C., Phillips, J. W., and Berry, M. N. (1995). Stimulation of gluconeogenesis leads to an increased rate of beta-oxidation in hepatocytes from fasted diabetic but not from fasted normal rats. Biochem. Biophys. Acta 1244(1), 92–98.

37.

Gerlach, J., Schnoy, N., Smith, M. D., and Neuhaus, P. (1994). Hepatocytes culture between woven capillary networks: A microscopy study. Artif. Organs 18, 226 –230.

38.

Smith, M. D., Airdrie, I., Courtney, J. M., Cousins, R. B., Ekevall, E., Grant, M. H., and Gaylor, J. D. S. (1997). Development and characterization of a hybrid artificial liver with integral membrane oxygenation. In “Bioartificial Liver: The Critical Issues” (S. Muaraca, Ed.), pp. 27–35, CIC Edizioni Internazionali, Padova.

39.

Bader, A., De Bartolo, L., and Haverich, A. (1997). Initial evaluation of the performance of a scaled up flat membrane bioreactor (FMB) with pig liver cells. In “Bioartificial Liver: The Critical Issues” (S. Muraca, Ed.), pp. 36 – 41, CIC Edizioni Internazionali, Padova.