Organ Slice Viability Extended for Pathway Characterization: An in Vitro Model to Investigate Fibrosis

Alison E. M. Vickers*,1, Muriel Saulnier*, Elba Cruz*, Marjolijn T. Merema{dagger}, Kristine Rose*, Philip Bentley* and Peter Olinga{dagger}

* Novartis Pharmaceuticals Corporation, One Health Plaza, E. Hanover, New Jersey 07936 and {dagger} University of Groningen, Groningen, The Netherlands

Received June 24, 2004; accepted September 16, 2004


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Liver slice viability is extended to 96 h for rat, expanding the use of this in vitro model for studying mechanisms of injury and repair, including pathways of fibrosis. The contributing factors to increased organ slice survival consist of the use of a preservation solution for liver perfusion and slice preparation, obtaining rats that are within the weight range of 250–325 g, placing a cellulose filter atop the titanium mesh roller-insert to support the slice, and maintaining the slices in an optimized culture medium which is replaced daily. The liver slices remain metabolically active, synthesizing adenosine triphosphate (ATP), glutathione, and glycogen, and exhibit preserved organelle integrity and slice morphology. Slice preparation results in 2-cut surfaces which likely triggers a repair and regenerative response. The fibrogenic pathways are evident by the activation of stellate cells, the proliferation of myofibroblast-like cells, and an increased collagen deposition by 48 h. Markers indicative of activated stellate cells, {alpha}-smooth muscle actin, collagen 1a1, desmin, and HSP47 are substantiated by real time-PCR. Increased staining of {alpha}-smooth muscle actin initially around the vessels and by 72–96 h in the tissue is accompanied by increased collagen staining. Microarray gene expression revealed extracellular matrix changes with the up-regulation of cytoskeleton, filaments, collagens, and actin genes; and the down-regulation of genes linked with lipid metabolism. The improvements in extending liver slice survival, in conjunction with its three-dimensional multi-cellular complexity, increases the application of this in vitro model for investigating pathways of injury and repair, and fibrosis.

Key Words: liver slices; fibrosis in vitro model.


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Precision-cut organ slices represent an in vitro model that mimics closely the multi-cellular complexity and extra-cellular interactions of the intact organ. The architectural composition of the various cell types is retained, as well as the cell-matrix and cell-cell interactions. Through the application of this model to various organs and species, mechanistic pathways leading to organ injury can be investigated in vitro. Furthermore, the identification of biologically relevant markers of animal and human response to drug-induced injury can be compared.

The potential of organ slices as an in vitro model for predicting and investigating clinically relevant effects of candidate pharmaceuticals has been limited by the short survival time of organ slices in culture (Lerche-Langrand and Toutain, 2000Go). Preserving organ slice integrity and viability has been shown to be most successful by culturing the slices in a dynamic roller culture system. In spite of this, soft tissue such as rodent liver is more difficult to culture than the more dense tissue such as liver of larger animals or human (Fisher et al., 2001Go). Currently, as the culture time extends past 24 h, the organ slice begins to lose cells at the contact site of the slice with the screen insert, and thereby reduces organ slice architectural stability and viability. Extension of organ slice survival increases the feasibility of investigating drug induced effects at more physiologically relevant concentrations and after multiple dosing (Pfaller et al., 2003). Additionally, recovering high quality RNA over longer culture times allows for the investigation of cellular pathways of organ injury and recovery via gene expression technologies to address questions about mechanisms in extended cultures.

The liver has an extraordinary capacity to regenerate and recover from mechanical or chemical injury. Repair pathways are generally manifested by an increased deposition of collagen. The activation of liver cell populations, particularly stellate cells, contributes to the induction of fibrogenic pathways and increased deposition of extra-cellular matrix. Furthermore, the tissue response to injury may include aspects of both repair (fibrosis) and regeneration with the outcome determined by the extent of injury and the cross-talk and orchestration between the different cell types and the extracellular matrix.

This study describes a marked improvement in the methodology of preparing and culturing rat liver slices up to 96 h. Tissue integrity and viability are maintained, and high quality RNA is isolated for microarray analysis. Pathways of repair are demonstrated to be active. Organ slices demonstrate to be a valuable in vitro model for investigative toxicology studies, particularly for studying mechanism of injury and repair because of the multi-cellular and extracellular matrix complexity of the model.


    MATERIALS AND METHODS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Male Wistar rats (Wistar Hannover, HsdBrl:WH) about 10 weeks of age were obtained from Charles River Breeding Laboratories (Wilmington, MA) were kept on a 12 h light-dark cycle and allowed free access to food and water. Aerrane (isoflurane) was obtained from Amersham Life Sciences (Arlington Heights, IL). Dulbecco's modified Eagle's medium (DMEM), dialyzed fetal bovine serum, Glutamax I, and antibiotic-antimycotic solution were purchased from Gibco BRL (Gaithersburg, MD). Insulin, EGF, and corticosterone were obtained from Sigma (St. Louis, MO). Viaspan (Belzer-University Wisconsin (UW) cold storage solution) was obtained from DuPont Pharma (DE). The bovine IgG standard was obtained from Bio-Rad (Hercules, CA). Mixed cellulose-ester IMMOBILIN-NC filters (HATF, 0.45 µm surfactant and triton free, autoclavable) to support the slice were purchased from Millipore (Bedford, MA).

Organ slices. Male Wistar rats (250–275 g) are anesthetized intraperitoneally with a combination of ketamine (40–80 mg/kg, 100–200 µl) and xylazine (10 mg/kg, 25 µl). Heparin (200 IU in 200 µl saline) is then injected into the tail vein and the animal is placed into a sterile biosafety hood where all further work is done aseptically. Perfusion of the liver with a preservation solution, clears the organ of blood, and cools the organ to preserve function (Delriviere et al., 1996Go). The abdominal area of the rat is disinfected with a gauze sprayed with 70% alcohol, and opened by a cruciform incision of the skin and the musculature, which is then retracted to expose the liver. The portal vein is freed from attachments and a cannula (25-guage) connected to a pulsatile pump is inserted to flush the liver with 100 ml of cold UW solution containing 3 mM glutathione, 2 mM glutamax, 0.03 mM L-ascorbate, 2 mM sodium pyruvate, 10–7 M insulin, 10–6 M corticosterone, heparin (200 IU), antibiotics/antimycotic (10,000 U/ml penicillin G, 10,000 µg/ml streptomycin, and 25 µg/ml amphotericine B in 0.85% saline) at pH 7.4. The flow rate is increased to about 70 ml/min (~10 ml/g liver). The vena cava and aorta are snipped to allow perfusate out-flow, and the diaphragm is cut. After complete flushing of the organ, the liver is freed and immediately placed in cold UW solution for coring and slicing. Tissue cores (8 mm3) are prepared using a tissue coring press (Alabama Research and Development, AL) while keeping the organ moist with UW solution. The cores are sliced (200 ± 20 µm thick) in cold, oxygenated UW or V-7 preservation solution using a Vitron tissue slicer (Tucson, AZ) as previously described (Fisher et al., 1996Go).

Slices are individually loaded onto a mixed cellulose-ester IMMOBILON-NC filter (HATF, 0.45 µm) which rests on the titanium-screen roller (Vitron Inc., Tucson, AZ). The filters are either purchased as discs (13 mm in diameter) or cut into rectangular shapes (20 x 10 mm) and autoclaved, submerged in dH2O at 120°C for 25 min, and then placed into the titanium rollers prior to slice loading. Each roller is placed into a sterile 20-ml glass vial with 1.7 ml culture medium, Dulbecco's Modified Eagle Medium (DMEM 25 mM glucose, 4 mM glutamine, and HEPES), containing 2 mM sodium pyruvate, 100 nM glucagon, 30 nM insulin, 1 µM corticosterone, 1 nM EGF, 10,000 U/ml penicillin G, 10,000 µg/ml streptomycin and 25 µg/ml amphotericine B in 0.85% saline, and 5% dialyzed fetal bovine serum which has been previously heat inactivated for 1 h at 56°C at pH 7.4. The vials are capped, lined with Teflon inserts, and a 1 mm hole allows for gas exchange. The vials are placed into a Heraeus incubator (Kendro Laboratory Products, Newtown, CT) and maintained at 37°C, 85% O2, 5% CO2, and 80% humidity for the length of the study. After a preincubation period of 60–90 min, the medium is replaced and subsequently replaced every 24 h. Liver samples are collected at tissue harvest (Thar), after slicing but before incubation (Tpre), after the pre-incubation period and start of experiment (0 h), and at 24 hr intervals, 24, 48, 72, and 96 h for slice function, morphology, and molecular biology data collection.

Various screen inserts were evaluated including the titanium-screen (Medical grade titanium, Vitron Inc., Tucson, AZ) and a teflon-screen (Fluortex ETFE monofilament fabrics, SEFAR America, Inc., NY). Varying mesh sizes of the screens were evaluated. The titanium screens were 35 x 35 mm or 80 x 80, and the teflon screens (9 micron mesh opening and 45–47% open area, 1000/45 and 590/47). The slice cultures were maintained for 48 h as described above.

Functional assays. Organ slice viability was assessed by determining slice ATP and glutathione (GSH) content. At the time point each slice was weighed and homogenized in 10% trichloroacetic acid (TCA) with a Powergen 125 (Fisher Scientific) at room temperature. The samples were snap frozen in liquid nitrogen and stored at –76°C until analysis. The liver slice homogenates were thawed, placed on ice, and centrifuged for 10 min at 11,000 x g at 4°C. The resultant slice homogenate supernatant was used for measurement of intracellular ATP and GSH.

For ATP determination the slice homogenate supernatant 4.0 µl was added directly to a white Microlite 96 well flat bottomed plate (Dynex Technologies; Chantilly, VA) containing 6.0 µl of 0.5 M Tris-1 mM EDTA buffer (pH 8.9). Luminescence was recorded using a Victor 2, 1420 Multilabel counter equipped with an injector. The instrument injected 100 µl of luciferin/luciferase reagent from the ATP determination kit from molecular probes, shook the mixture for 1 s and then read the luminescence. The value from the blank wells was subtracted from the sample values and compared to a standard curve of ATP (0–100 µM). Data is presented as nmoles ATP/mg slice wet weight.

GSH content was determined from the slice homogenate supernatant (50 µl) transferred to a 96 well microtiter plate and 200 µl of Ellman's reagent (39.6 mg dithiobis-nitrobenzoic acid/10 ml EtOH diluted 1:10 with 0.5 M Tris-1 mM EDTA buffer (pH 8.9) added. The absorbance was determined at 405 nm using a SpectraMax 340 plate reader from Molecular Devices (Sunnyvale, CA) and the values compared to a standard curve of reduced glutathione (0–250 µM). Data is presented as nmoles GSH/mg slice wet weight.

Liver slice caspase 9 and 3 activities were measured following slice disruption in 500 µl of cold cell lysis buffer and centrifugation at 3000 x g for 5 min at 4°C. An aliquot (50 µl) was added to a microtiter plate and combined with 50 µl reaction buffer (prepared with 10 µl of fresh DDT per ml of 2x reaction buffer) and 5 µl of fluorogenic peptide, DEVD-AFC for caspase 3 and LEHD-AFC for caspase 9 (R&D Systems, Minneapolis, MN). Plates were incubated at 37°C for 1–2 h and then inserted into a Wallac Victor tray (Wallac Victor 1420 Multilabel Counter) and read at an excitation of 405 nm and an emission of 510 nm every 10 min for 70 min. The results are expressed as fluorescence/min/µg protein.

The total protein content of each homogenate was determined using a modified Bradford protocol (Bradford, 1976Go) according to the manufacturer's recommendations (Sigma, St. Louis, MO). Tissue homogenate was diluted 1:40 with 0.1 N NaOH and an aliquot (10 µl, ~10 ng protein) was added to a microtiter plate to which 200 µl of undiluted Bradford reagent (Sigma) was added. After mixing, the absorbance at 595 nm was measured on a SpectraMax 340 plate reader and compared to a standard curve of bovine IgG (0–500 µg/ml).

RNA analysis. Liver slices were snap frozen in a 96 well collection tube rack (RNeasy 96 kit, Qiagen Inc.), containing two (3 mm diameter) acid washed glass beads, at the time of collection. Just prior to RNA isolation 500 µl of RNA lysis buffer (guanidinium isothiocyanate and 1% ß-mercaptoethanol) was added to the slices. Slices were disrupted by vigorous (3 x 1 min at 30 Hz) shaking in a Qiagen Mixer Mill. Homogenate was cleared of unbroken cells and cellular debris by spinning through a QIAfilter 96 plate for 5 min at 3000 x g. If necessary an additional spin for 5 min at 4000 x g was carried out. An aliquot of homogenate (10 µl) was removed from the flow through of each sample and assayed for total protein content. For additional clarification, homogenates were subjected to proteinase digestion (10 µl of > 60 mAU/ml proteinase K) for 20 min at 55°C. Following proteinase digestion, 500 µl of 70% EtOH was added to the remaining homogenate to promote binding to the RNeasy 96 column. Samples were mixed by inversion, transferred to the RNeasy 96-column plate, and centrifuged at 6000 x g for 2 min, and the flow through discarded. The column was washed with 700 µl of RW1 wash buffer and the filter bound RNA samples were subjected to on-column DNase treatment (70 µl of DNase reaction buffer + 10 µl of DNase at 2.7 units/µl; Qiagen) for 15 min at room temperature to remove contaminating genomic DNA. Extensive washing with RNeasy wash buffers followed, and the purified RNA was eluted 2x with 50 µl of RNase free ddH2O. Yield and purity of RNA was determined by measurement of the optical density (OD), of a 1/50 dilution of the samples, at 260 and 280 nm using a SpectraMax Plus microplate spectrophotometer (Molecular Devices). An OD260 of 1 corresponds to a 40 µg/ml RNA solution. An OD260/280 ratio of ~1.7–2.0 signifies an RNA sample pure enough for use in expression analysis. To assess the integrity of the 18 s and 28 s rRNA, and hence that of the total RNA sample, an aliquot of total RNA (1 µg) was run on a 1.5% agarose gel and stained with the fluorescent RNA stain, Sybr Green II (Molecular Probes). RNA was visualized on a UV light box.

Double stranded cDNA was synthesized from full-length mRNA (5 µg) using Superscript Choice System (Gibco-BRL, Life Technologies, Rockville, MD). Following synthesis, the cDNA was purified by phenol/chlorophorm extraction and ethanol percipitation and then transcribed in vitro using Enzo BioArray High Yield RNA Transcript Labeling Kit (ENZO, Farmingdale, NY) to form biotin labeled cRNA. The cRNA (20 µg cRNA; ≥0.6 µg/µl) was fragmented and hybridized (12–15 µg cRNA, ≥0.05 µg/µl) overnight to the rat genome U34A array, ~7000 full length sequence and ~1000 ESTs (Affymetrix, Santa Clara, CA) for 16 h at 45°C. The arrays were washed and stained using the GeneChip Fluidics station. After washing, the arrays were scanned twice with the Gene Array scanner (Affymetrix). The data (.DAT file) was captured using the Affymetrix GeneChip Laboratory Information Management System (LIMS). The LIMS database is connected to a UNIX Sun Solaris server through a network filing system that allows for the average intensities for all probes cells (.CEL file) to be downloaded into an Oracle database (NPGN). Raw data was converted to expression levels using a target intensity of 150.

Real-time quantitative PCR (RT-PCR). RNA representing individual slices for each time point was used to verify the expression of stellate cell specific genes including {alpha}-smooth muscle actin, collagen 1a1, desmin, Hsp 47, and {alpha}-ß-crystallin. The primers for the genes, were designed using Primer Express v 1.5 (Applied Biosystems, Foster City, CA) and prepared by Applied Biosystems Assay on Design. The primers and probe, labeled with the 5' reporter dye 6-carboxy-fluorescein (FAM) and the 3' minor groove binder non-fluorescent quencher dye (MGBNFQ), were purchased from Applied Biosystems. The endogenous reference used was actin RNA. Optimal primer concentrations were determined by maintaining the RNA quantity while varying the forward and reverse primer concentrations. Validation of the probe was performed by varying the quantity of RNA and plotting the log of the RNA quantity versus {Delta}CT. Initially, cDNA was synthesized from 3 µg RNA using a Reverse Transcription System (Promega, Madison, WI). Reverse transcriptase polymerase chain reaction was performed for 10' at 25°C and 60' at 45°C, and the reaction stopped by heating 5' at 95°C. cDNA (1.25 µl, 50 ng total RNA) was used in real-time PCR reactions with 2X master mix without UNG (includes AmpliTaq Gold DNA polymerase, dNTPs with dUTP, the passive reference ROX, and optimized buffer components), TaqMan probe, and forward and reverse primers. The amplification consisted of 40 cycles of melting (15 min at 95°C) and annealing/extension (1 min at 60°C). Results were calculated using the Comparative Ct method described in User Bulletin #2 (Applied Biosystems). Data is presented as fold change over the vehicle treated samples ± SD. The following primers, listed in the 5' to 3' direction, were used: {alpha}-smooth muscle actin (X06801) forward = AGCTCTGGTGTGTGACAATGG, probe = CCGCCTTACAGAGCC, reverse = GGAGCATCATCACCAGCAAAG; collagen 1a1 (Z78279) forward = CCCACCGGCCCTACTG, probe = CCTCCTGGCTTCCCTG, reverse = GACCAGCTTCACCCTTAGCA; desmin (NM_022531) forward = TCCAACTGAGAGAAGAAGCAGAGA, probe = CTTCCCGCCATGCCAC, reverse = CCAGAGTGGCTGCATCCA; Hsp 74 (M69246) forward = AGACGAGTTGTAGAGTCCAAGAGT, probe = CTTCCCGCCATGCCAC, reverse = ACCCATGTGTCTCAGGAACCT; {alpha}-B-crystallin (NM_012935) forward = TTGGAGTCTGACCTCTTCTCTACAG, probe = CACTTCCCTGAGCCCC, reverse = AGGGTGGCCGAAGGTAGAA; actin (NM_031144) forward = CGAGGCCCAGAGCAAGAG, probe = CTGACCCTGAAGTACCC, reverse = TTGGTTACAATGCCGTGTTCAATG.

Morphology. Liver slices were either gently shaken from the insert in warmed (37°C) medium or maintained on the on the insert. Each slice was laid flat in a 6-well culture dish. Any remaining medium was removed and 2–3 drops of 10% neutral buffered formalin were dropped on top of the slice. After 10 min, additional formalin was added on top of the slice. After 1 h of fixing in formalin, the slice was rinsed in 80% ethanol, placed between foam inserts in a histological cassette, and stored in 80% ethanol at 4°C until paraffin embedding. The parrafin-embedded slices were cut at 5 µm thickness and processed for routine hematoxylin-eosin, PAS (periodic acid Schiff-glycogen-specific), and Masson-trichrome staining, and light microscopic evaluation.

To examine the slices by electron microscopy modified Karnovsky's-fixed tissues were further processed and postfixed in 0.1 M sodium cacodylate-buffered 1% osmium tetroxide, dehydrated, embedded in epon Embed (Epon) 812. Semi-thin sections (1 mm) were stained with toluidine blue to select relevant areas. Ultra-thin sections from these areas (60–90 nm) were stained with uranyl acetate and lead citrate for electron microscopy with a Zeiss electron microscopy (EM)-902 microscope (Carl Zeiss, Oberkochen, Germany).

Liver slice collagen was detected by the Masson trichrome stain (Luna, 1968Go). To detect glycogen a PAS (periodic acid shift) stain followed by a diastase resistance test was performed on the slices (Luna, 1968Go). Immunohistochemical localization of {alpha}-smooth muscle actin ({alpha}-SMA) was performed on liver slices exposed to increasing culture times (Beljaars et al., 2000Go). Sections were stained according to standard indirect immunoperoxidase methods. Antigen retrieval was performed by incubating the sections in 0.01 M Tris/HCl (pH = 9) overnight at 80°C, 30 min at room temperature and then washed with PBS. Endogenous peroxidase was blocked by a 20 min incubation with PBS containing 0.075% H2O2. The primary antibody {alpha}-SMA (Sigma) was diluted in PBS (1:500), and the secondary antibody RAM-PO (Dako 1:50) was diluted in PBS containing 5% normal rat serum.

Statistics. For the biochemical assays a one-way ANOVA followed by two sided Dunnett's multiple comparisons test was employed with Graphpad Prism Software version 3.1 (Graphpad Software, Inc.; San Diego, CA).

Statistical analysis of the Affymetrix gene expression data was performed using PartekPro 5.0 (Partek Inc., St. Charles, MO), comparing the 24, 48, 72, and 96 h data to the time 0 data. There were four replicate samples per time point from two independent experiments. Statistical significant genes were designated as those with a p ≤ 0.05 based on one-way ANOVA and a randomization experiment of 200 experiments to correct for multiple comparisons.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Rat liver slices maintained on the mixed cellulose ester filters retain slice integrity and an even surface on the contact side of the slice to the filter. Both the titanium and the teflon screens cause a loss of cells at the contact point of the slice with the screen. With longer incubation times, 48 h, the extent of the cell loss can lead to a complete break in the slice longitudinally, which appears as a cut in the planar view (Fig. 1).



View larger version (100K):
[in this window]
[in a new window]
 
FIG. 1. Schematic of the dynamic organ slice culture system and slice morphology (cross-section and planar view) after culturing the slices on a titanium roller insert only (H&E, 20x) versus a filter covered titanium roller insert (H&E, 10x).

 
Slice Function
Liver slice ATP and GSH levels are important indicators of slice viability. Both ATP and GSH are synthesized by the slices over the duration of the cultures. The slice ATP levels recover compared to the tissue levels at harvest following organ perfusion and decline to that level at 72–96 h of culture. Significant increases in liver slice GSH levels occur compared to both the harvest of the slices (Tpre) and following the pre-incubation and start of the experiment (Fig. 2). The medium is replaced every 24 h providing an adequate supply of precursors.



View larger version (12K):
[in this window]
[in a new window]
 
FIG. 2. ATP and GSH liver slice levels at the times of slice harvest (Tpre), following a pre-incubation period with a medium change (0), and for the culture times 24–96 h. Statistical values represent four slices per time point compared to time 0 of multiple independent experiments.

 
Rat liver slice wet weight and total protein content decreases with time in culture. The greatest change in slice wet weight, ~30% decrease, occurs within the first 24 h of incubation. Screen type (titanium or teflon) or the placement of a mixed cellulose ester filter on the screen does not affect this initial decrease of slice wet weight. Following the initial decrease in slice weight, the mixed cellulose ester filters maintained slice wet weight throughout the 96 h culture period. The slices also attached to the filters, which is an indication of healthy metabolically active tissue.

Caspase 9 and caspase 3 activities, markers of apoptosis, remained relatively constant over the duration of the 96 h culture period. The findings suggest that the culture conditions did not adversely affect apoptosis (data not shown).

Morphology
Light microscopic evaluation of H&E stained slices show that the liver slices maintain its normal architecture for the 96 h. Electron microscopic evaluation confirms that the cells are rich in mitochondria and rough endoplasmic reticulum, and that the nuclear and organelle integrity are maintained throughout the culture (Fig. 3). Multifocal areas of coagulative necrosis exist at the periphery of the slice, suggesting damage incurred from the coring of the tissue with a stainless steel coring tool. Foci of single or groups of cell necrosis was evident around the central veins and randomly dispersed throughout the parenchyma.



View larger version (226K):
[in this window]
[in a new window]
 
FIG. 3. Organelle integrity of the rat liver slices is maintained as shown by the electron micrograph (3000x) of 96 h slices.

 
The liver slices are metabolically active, synthesizing glycogen and storing lipids from the nutrients of the culture medium. A stain for glycogen, PAS, followed by a diastase digestion, confirmed hepatocyte glycogen storage. Slice glycogen content increased with culture time pericentral and centrilobular areas. The presence of lipid vacuoles within the cytoplasm of hepatocytes was confirmed by EM at 72–96 h culture times (data not shown). Glycogen accumulation by the slices was minimized by varying the amount of glucagon, insulin and corticosterone in the culture medium. The final conditions are listed in the materials and methods.

Tissue repair is evident by 48 h of culture as characterized by the deposition of collagen, and a proliferative response. Mitotic activity is exhibited by biliary cells and hepatocytes, in conjunction with PCNA staining to confirm the proliferative status of the slices. Electron microscopy revealed the proliferation of spindle shaped cells in the portal spaces around immature epithelial type cells forming canaliculi (Fig. 4). The increased deposition of collagen around vessels and within the tissue was demonstrated with a masson trichrome stain for collagen (Fig. 5). The appearance of myofibroblasts within the liver slice was shown by an increased {alpha}-smooth muscle actin staining with culture time (Fig. 6).



View larger version (105K):
[in this window]
[in a new window]
 
FIG. 4. Neobile ductile formation in rat liver slice portal spaces at 72 h (H&E 400x, EM 3000x).

 


View larger version (91K):
[in this window]
[in a new window]
 
FIG. 5. Collagen detection by masson trichrome staining of rat liver slices at harvest (20x), and at 72 h (10x) of incubation in culture.

 


View larger version (151K):
[in this window]
[in a new window]
 
FIG. 6. Immunohistochemical localization of {alpha}-smooth muscle actin in rat liver slices at 24, 48. and 72 h (200x) and at 96 h (400x) of culture.

 
Gene Expression
RNA isolated from liver slices maintained over four days of culture was suitable for micro-array hybridization and analysis. The RNA yield was at least 10 µg/mg total protein, and the integrity was intact with all samples exhibiting clear, tight bands of 18 and 28 s rRNA upon visual inspection using fluorescent agarose gel electrophoresis. RNA purity, as determined by the 260/280 ratio of the samples, was within the desired range of 1.7–2.0.

Tissue fibrosis involves an increased production of extracellular matrix and increased cell proliferation regulated by factors and cytokines. Many genes linked with cell matrix, transcription and growth factors, protein synthesis, inflammation and cellular metabolism, were altered with time in culture.

Genes associated with the extra-cellular matrix were up-regulated with time in culture in the rat liver slices (Table 1). The increased gene expression of collagens, fibronectin Fn, {alpha}-smooth muscle actin, vimentin, and tropomyosin 1{alpha}, suggest stellate cell activation. MMP-14 is a membrane metalloproteinase present in parenchymal and non-parenchymal liver cells. TIMP-1, a tissue inhibitor of matrix metalloproteinases, is predominantly present in stellate cells and myofibroblasts (Bataller and Brenner, 2001Go; Nieto et al., 2001Go; Okuyama et al., 2002Go). The activation of stellate cells was verified by Real-Time-PCR of the increased expression of stellate cell marker genes (Fig. 7).


View this table:
[in this window]
[in a new window]
 
TABLE 1 Rat Liver Slice Gene Expression Changes (U34A Array) at 24–96 h of Culture Relative to Time 0

 


View larger version (14K):
[in this window]
[in a new window]
 
FIG. 7. Stellate cell marker genes in rat liver slices measured by real-time quantitative PCR.

 
The regulation of genes associated with growth factors (TGF-ß, insulin-like growth factor), and proinflammatory cytokines (IL-1, IL-6) are associated with tissue fibrosis and are altered (down-regulated) in this study. The significance of the direction of the gene changes is not known and may be related to the dynamics and kinetics of the gene expression changes in the in vitro system. Liver enriched transcription factors, functionally required for differentiation and metabolism, such as HNF-4 and HNF-6 are repressed (Schrem et al., 2002Go). HNF-4 is involved in the regulation of genes in diverse metabolic pathways including glucose, cholesterol, and fatty acid metabolism. HNF-6 can inhibit the glucocorticoid-induced stimulation of genes involved with liver glucose metabolism, 6-phosphofructo-2-kinase. The slices exhibited a repression in genes involved in multiple metabolic pathways including carbohydrate metabolism, cholesterol and lipid synthesis. The repression of several carbohydrate genes including glucose-6-phosphatase gene expression can in part be explained by the components of the culture medium, including ~25 mM glucose. In agreement with the result of glutathione levels increased and maintained in culture, at the molecular level there was evidence for glutathione biosynthesis throughout the culture time course.


    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Rat liver slice viability and morphology is optimized to increase the utility of the model for characterizing liver injury. The contributing factors to increasing slice survival include (1) the use of a preservation solution to both perfuse the organ in situ and during slice preparation, (2) use of a cellulose insert atop the titanium mesh to support the slice during culture, and (3) use of an optimized culture medium.

Slices maintained on a screen insert of titanium, Teflon, or stainless steel, and with various mesh sizes, lose tissue integrity at the contact point of the slice with the screen within 24 h of culture. In this study improvements in slice preparation and culturing contribute to preserving tissue quality, stability, and slice survival. For culturing, the slices are placed onto a support, a mixed cellulose-ester filter, which rests on the screen of the roller insert to prevent damage induced by direct contact, and is a simple and fast alternative to coating screens with either collagen or matrigel. The culture medium is replaced daily so that compound exposure can mimic a once-a-day treatment.

The liver slices are metabolically active and morphologically healthy, which is indicated by the recovery and maintenance of cellular energy levels (ATP), replenishment of slice antioxidant GSH status, and synthesis of glycogen from the culture medium nutrients. Both insulin and hydrocortisone, components of the culture medium, have been shown to contribute to the homeostasis of glutathione (Lu et al., 1992Go). The morphological features representative of metabolically active tissue included numerous mitochondria, abundant rough endoplasmic reticulum, and maintenance of the nucleus and organelle integrity over the culture period. The attachment of the slices to the cellulose-ester filter is also an indication of viable, metabolically active tissue, similar to hepatocytes binding to various substrata or extra cellular matrix components and exhibiting an enhanced expression of tissue specific functions (Bissell et al., 1990).

Aspects of organ harvesting and maintenance that have been shown to be beneficial for organ transplantation can be adapted to the organ slice methodology. In this study, the rat weight is kept within a range of 250–325 g to ensure a reasonable size of the liver and vessels while minimizing the fat content (Delriviere et al., 1996Go). Additionally, the animals are not fasted because it is deleterious for organ preservation (Adam et al., 1992Go). A transplant quality preservation solution like UW is used to perfuse the liver and to prepare the tissue cores (Collins and Wicomb, 1992Go; Sumimoto and Kamada, 1990Go). The slices are then prepared in a preservation solution to maintain organ viability, either UW or V-7 (Fisher et al., 1996Go). Optimal liver slice function and a balance of glycogen formation and gluconeogenesis was achieved by optimizing the medium glucoregulatory hormones insulin, glucagon and hydrocortisone (Bissel et al., 1973Go, 1978Go; Dich et al., 1988Go; Ohno and Maier, 1994Go). Pyruvate was included in the medium because of its protective role in ischemia reperfusion injury (Sileri et al., 2001Go). EGF was included to support hepatocyte, biliary epithelium and connective tissue formation (Michalopoulos et al., 2001Go). Other media and supplementations known to support hepatocytes in culture could also be suitable (Behrsing et al., 2003Go).

Many cell types in the liver can synthesize extracellular matrix proteins, however, it is the hepatic stellate cells, perivascular mesenchymal cells representing 5–8% of total liver cells, that are key in the wound-healing, fibrogenic response. The characteristic biological consequence is the increased production, deposition and remodeling of extracellular matrix, i.e., laminin and collagen. Hepatic stellate cells produce growth factors and cytokines, in addition to extracellular matrix components, and are pivotal in the regenerative response to injury either from tissue resection or to an inflammatory response triggered by chemical injury to hepatocytes. The induction and outcome of repair processes, fibrosis, and regeneration, depends on the cross-talk and orchestration between the different cell types and the release of chemical mediators; while the liver cells continue to perform the essential functions needed for homeostasis including glucose regulation, synthesis of many blood proteins, and bile secretion (Bataller and Brenner, 2001Go; Friedman and Arthur, 2002Go; Michalopoulos and De Frances, 1997Go). The preparation of organ slices is the result of 2-cut surfaces. This technique could initiate an injury and repair response, as well as trigger a regenerative response. The concurrence of cell proliferation and increased collagen deposition suggests an overall repair process. Following surgical resection, the induction of fibrogenic pathways is regarded as a scarring response to injury (Bataller and Brenner, 2001Go).

The activation of normally quiescent hepatic stellate cells as well as portal fibroblasts by either mechanical or chemical-induced injury causes a proliferative and phenotypic transformation into cells with myofibroblast-like features, which are related to smooth muscle cells, and express enhanced levels of {alpha}-smooth muscle actin (Kinnman et al., 2003Go; Schuppan et al., 2001Go). The myofibroblastic conversion of peribiliary cells in the portal tracts following a bile duct ligation is distinct from hepatic stellate cells, as shown by the in vitro emergence of myofibroblasts from bile duct preparations devoid of hepatic stellate cells. Cellular markers of myofibroblastic phenotype overlap for hepatic stellate cells and the peribiliary cells with {alpha}-smooth muscle actin and {alpha}1-collagen, while desmin and IL-6 levels are greater in the hepatic stellate cells (Kinnman et al., 2003Go). Further characterization of hepatic stellate cell specific markers include the cytoskeleton protein desmin, as well as markers of activated stellate cells, heat shock proteins HSP47 and {alpha}-B-crystallin (Geerts, 2001Go). Rat hepatic stellate cells activated in vitro demonstrate increased collagen synthesis and cell proliferation (Sato et al., 1995Go).

In this study, the morphological features of repair in the liver slices initially occur in cells surrounding the periportal regions and then appear within the slice. The activation of hepatic stellate cells and myofibroblastic peribiliary cells is prominent at 48 h based on the time-course of increased gene expression of cell specific markers (desmin, HSP47, collagen, and {alpha}-smooth muscle actin) and the deposition of collagen and {alpha}-smooth muscle actin proteins. Cultured stellate cells exhibit activation within hours after isolation, as measured by the increased gene and protein expression of {alpha}-B-crystallin (Cassiman et al., 2001Go; Lang et al., 2000Go).

Gene expression changes reflecting cell-extracellular matrix interactions are prominent in this study. The extracellular matrix provides cells with positional information and a scaffold for adhesion, migration, and proliferation, and consists of collagens, glycoproteins, proteoglycans, and glycosaminoglycans. Increased expression of {alpha}-smooth muscle actin, microtubule cytoskeleton genes, and several types of intermediate filaments including vimentin and desmin are evident and are known to respond to the stellate cell activation process (Sato et al., 2003Go). Proteomic analysis of activated rat stellate cells by 2-D gels revealed increased protein levels of fibril forming collagens (type 1 and III), and cytoskeletal elements including contractile proteins ({alpha}-smooth muscle actin, tropomyosin), microtubules ({alpha}-tubulin), and filaments (vimentin) (Kristensen et al., 2000Go). Moreover, stellate cells are a major source of TIMP-1 and TIMP-2 (tissue inhibitors of metalloproteinases), which inhibit the activity of metalloproteinases, such as MMP-14, MMP-2, and the degradation of collagens (Friedman and Arthur, 2002Go). The extracellular matrix participates in fibrosis through the binding of various molecules including growth factors, cytokines, matrix metalloproteinases, and processing enzymes (Schuppan et al., 2001Go). The interaction of tissue stellate cells and fibroblasts with the extracellular matrix make these cells subject to systemic factors like ischemia, endothelin, thrombin, which are related or non-related to inflammation (Atamas, 2002Go).

The process of liver regeneration is also linked with increased TIMP-1, MMP-2, and MMP-9 activities (Kim et al., 2000Go). Additionally, insulin-like growth factor and TIMP-1 may act as survival factors triggering an anti-apoptotic effect (Iredale, 2001Go). Liver regeneration models have provided insight into the proliferative time course of the liver cell types (Michalopoulos et al., 1997Go). Parenchymal cell proliferation forming avascular islands is followed by sinusoidal endothelial cells and stellate cell proliferation, which migrate into the parenchymal clusters, followed by the formation of new vascular branches and perisinusoidal space. Key signalling molecules include growth factors (HGF and EGF), cyokines (TGF{alpha}, IL-6, TNF{alpha}) as well as insulin and norepinephrine. Hepatic stellate cells likely have a pivotal role in liver regeneration including both parenchymal cell proliferation and ECM remodeling since these cells are a major producer of growth factors, cytokines, ECM components and MMPs (Sato et al., 2003Go).

The down regulation of genes linked with lipid metabolism, and some of the drug metabolizing enzymes has been reported in a rat liver regeneration model and in rat hepatocyte cultures (Fukuhara et al., 2003Go; Runge et al., 2000Go). In particular, decreases in several of the hepatic apolipoproteins, acyl-CoA synthetase, and CYP3A may be associated with an increase of an inflammatory or acute phase response (Fukuhara et al., 2003Go; Starkel et al., 2000Go). Furthermore, there is an impairment of lecithin:cholesterol acyltransferase activity, and the metabolism of steroid hormones by delta-4-3-ketosteroid 5-beta-reductase (Fukuhara et al., 2003Go; Takeuchi et al., 1976Go).

In models of liver regeneration, chemical induced liver injury, or bile-duct ligation, repair pathways are initiated that include fibrosis compounded by inflammation, proliferation, metabolism, and cell death. The effects of modulating the fibrosis pathway have in part been investigated with PDGF stimulators and inhibitors, and by following the recovery phase after carbon tetrachloride exposure (Jiang et al., 2004Go; Kinnman et al., 2000Go, 2003Go). In all of these in vivo models the gene expression profile depends on the origin of the injury and the cell types affected by the induction of fibrosis. Therefore, it is difficult to compare the exact gene expression patterns, however the most common genes expressed in these in vivo models include {alpha}-smooth muscle actin, procollagen, HSP47, desmin and {alpha}-B-crystallin, which are genes expressed by activated stellate cells. In future organ slice studies, the elucidation of mechanism(s) such as injury due to ischemia and reperfusion, the high glucose content of the culture medium, or bile accumulation, on stellate cell activation and increased extra-cellular matrix production will be studied (Paradis et al., 2001Go). Recent studies in rat liver slices demonstrate that the addition of carbon tetrachloride, a fibrotic agent in vivo, leads to activation of the stellate cells within the slice (Olinga et al., 2002Go). The activation of stellate cells in liver slices by various means allows for future investigations in the role of stellate cells in the pathogenesis of tissue fibrosis.

In conclusion, rat liver slice viability, integrity and morphology has been extended out to 96 h. Tissue repair, manifested as fibrogenic pathways, is demonstrated by the increased deposition of collagen, increased staining of {alpha}-smooth muscle actin protein and increased expression of genes indicative of extra-cellular matrix. The liver slices possess the three-dimensional organ architecture and cellular interactions which provides this in vitro model as a means to investigate pathways of injury and repair, including fibrosis.


    ACKNOWLEDGMENTS
 
The authors thank April Pollack for her support with performing the slice cultures and functional assays, as well as Natalie Paladini and Fran Grossman for histology support.


    NOTES
 

1 To whom correspondence should be addressed at present address: Allergan, Inc., 2525 Dupont Drive, Irvine CA 92623. Fax: (714) 246-5850. E-mail: vickers_alison{at}allergan.com.


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
Adam, R., Astarcioglu, I., Gigou, M., Isaac, J., and Bismuth, H. (1992). The influence of the glycogen content of the donor graft function and survival in rat liver transplantation. Transplantation 54, 753–756.[ISI][Medline]

Atamas, S. P. (2002). Complex cytokine regulation of tissue fibrosis. Life Sci. 72, 631–643.[CrossRef][ISI][Medline]

Bataller, R., and Brenner, D. A. (2001). Hepatic stellate cells as a target for the treatment of liver fibrosis. Semin. Liver Dis. 21, 437–451.[CrossRef][ISI][Medline]

Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254.[CrossRef][ISI][Medline]

Behrsing, H. P., Vickers, A. E. M., and Tyson, C. A. (2003). Extended rat liver slice survival and stability monitored using clinical biomarkers. Biochem. Biophys. Res. Commun. 312, 209–213.[CrossRef][ISI][Medline]

Beljaars, L., Molema, G., Schuppan, D., Geerts, A., De Bleser, P. J., Weert, B., Meijer, D. K. F., and Poelstra, K. (2000). Successful targeting to rat hepatic stellate cells using albumin modified with cyclic peptides that recognize the collagen type VI receptor. J. Biol. Chem. 275, 12743–12751.[Abstract/Free Full Text]

Bissell, D. M., Hammaker, L. E., and Meyer, U. A. (1973). Parenchymal cells from adult rat liver in nonproliferating monolayer culture. I. Functional studies. J. Cell Biol. 59, 722–734.[Abstract/Free Full Text]

Bissell, D. M., Levine, G. A., and Bissel, M. J. (1978). Glucose metabolism by adult hepatocytes in primary culture and by cell lines from rat liver. Am. J. Physiol. 234, C122–C130.[ISI][Medline]

Bissel, D. M., Caron, J. M., Babiss, L. E., and Friedman, J. M. (1990). Transcriptional regulation of the albumin gene in cultured rat hepatocytes, role of basement-membrane matrix. Mol. Biol. Med. 7, 187–197.[ISI][Medline]

Cassiman, D., Roskams, T., van Pelt, J., Libbrecht, L., Aertsen, P., Crabbe, T., Vankelecom, H., and Denef, C. (2001). Alpha B-crystallin expression in human and rat hepatic stellate cells. J. Hepatol. 35, 200–207.[CrossRef][ISI][Medline]

Collins, G. M., and Wicomb, W. N. (1992). New organ preservation solutions. Kidney Int. 42, S-197–S-202.[ISI]

Delriviere, L., Gibbs, P., Kobayashi, E., Goto, S., Kamada, N., and Gianello, P. (1996). Detailed modified technique for safer harvesting and preparation of liver graft in the rat. Microsurgery 17, 690–696.[Medline]

Dich, J., Vind, C., Grunnet, N. (1988). Long-term culture of hepatocytes: Effect of hormones on enzyme activities and metabolic capacity. Hepatology 8, 39–45.[ISI][Medline]

Fisher, R. L., Hasal, S. J., Sanuik, J. T., Hasal, K. S., Gandolfi, A. J., and Brendel, K. (1996). Cold- and cryopreservation of dog liver and kidney slices. Cyrobiology 33, 163–171.[CrossRef]

Fisher, R. L., Ulreich, J. B., Nakazato, P. Z., and Brendel, K. (2001). Histological and biochemical evaluation of precision-cut liver slices. Toxicol. Methods 11, 59–79.[CrossRef][ISI]

Friedman, S. L., and Arthur, M. J. P. (2002). Reversing hepatic fibrosis. Sci. Med. July/August, 194–205.

Fukuhara, Y., Hirasawa, A., Li, X.-K., Kawasaki, M., Fujino, M., Funeshima, N., Katsuma, S., Shiojima, S., Yamada, M., Okuyama, T., Suzuki, S., Tsujimoto, G. (2003) Gene expression profile in the regenerating rat liver after partial hepatectomy. J. Hepatol. 38, 784–792.[CrossRef][ISI][Medline]

Geerts, A. (2001). History, heterogeneity, developmentatl biology, and functions of quiescent hepatic stellate cells. Semin. Liver Dis. 21, 311–335.[CrossRef][ISI][Medline]

Iredale, J. P. (2001). Hepatic stellate cell behavior during resolution of liver injury. Semin. Liver Dis. 21, 427–436.[CrossRef][ISI][Medline]

Jiang, Y., Liu, J., Waalkes, M., and Kang, Y. J. (2004). Changes in gene expression associated with carbon tetrachloride-induced liver fibrosis persist after cessation of dosing in mice. Toxicol. Sci. 79, 404–410.[Abstract/Free Full Text]

Kim, T.-H., Mars, W. M., Stolz, D. B., and Michalopoulos, G. K. (2000). Expression and activation of Pro-MMP-2 and Pro-MMP-9 during rat liver regeneration. Hepatology 30, 75–82.[CrossRef]

Kinnman, N., Hultcrantz, R., Barbu, V., Rey, C., Wendum, D., Poupon, R., and Housset, C. (2000). PDGF-mediated chemoattraction of hepatic stellate cells by bile duct segments in cholestatic liver injury. Lab. Invest. 80, 697–707.[ISI][Medline]

Kinnman, N., Francoz, C., Barbu, V., Wendum, D., Rey, C., Holtcrantz, R., Poupon, R., and Housset, C. (2003). The myofibroblastic conversion of peribiliary fibrogenic cells distinct from hepatic stellate cells is stimulated by platelet-derived growth factor during liver fibrogenesis. Lab. Invest. 83, 163–173.[ISI][Medline]

Kristensen, D. B., Kawada, N., Imamura, K., Miyamoto, Y., Tateno, C., Seki, S., Kuroki, T., and Yoshizato, K. (2000). Proteome analysis of rat hepatic stellate cells. Hepatology 32, 268–277.[CrossRef][ISI][Medline]

Lerche-Langrand, C., and Toutain, H. J. (2000). Precision-cut liver slices: Characteristics and use for in vitro pharmaco-toxicology. Toxicology 153, 221–253.[CrossRef][ISI][Medline]

Lang, A., Schrum, L. W., Schoonhoven, R., Tuvia, S., Solis-Herruzo, J. A., Tsukamoto, H., Brenner, D. A., and Rippe, R. A. (2000). Expression of small heat shock protein alpha-ß-crystallin is induced after hepatic stellate cell activation. Am. J. Physiol. Gastrointest. Liver Physiol. 279, G1333–G1342.[Abstract/Free Full Text]

Lu, S. C., Kuhlenkamp, J., and Kaplowitz, N. (1992). Insulin and glucocorticoid dependence of hepatic {gamma}-glutamylcysteine synthetase and glutathione synthesis in the rat. J. Clin. Invest. 90, 524–532.[ISI][Medline]

Luna, L. G. (Ed.) (1968). Manual of Histologic Methods of the Armed Forces Institute of Pathology, 3rd ed. McGraw-Hill New York, NY.

Michalopoulos, G. K., and DeFrances, M. C. (1997). Liver regeneration. Science 276, 60–66.[Abstract/Free Full Text]

Michalopoulos, G. K., Bowen, W. C., Mule, K., and Stolz, D. B. (2001). Histological organization in hepatocyte organoid cultures. Am. J. Pathol. 159, 1877–1887.[Abstract/Free Full Text]

Nieto, N., Dominguez-Rosales, J. A., Fontana, L., Salazar, A., Armendaria-Borunda, J., Greenwel, P., and Rojkind, M. (2001). Rat hepatic stellate cells contribute to the acute-phase response with increased expression of {alpha}1 (I) and {alpha}1 (IV) collagnes, tissue inhibitor of metalloproteinase-1, and matrix-metalloproteinase-2 messenger RNAs. Hepatology 33, 597–607.[CrossRef][ISI][Medline]

Ohno, K., and Maier, P. (1994). Cultured rat hepatocytes adapt their cellular glycolytic activity and adenylate energy status to tissue oxygen tension: Influences of extracellular matrix components, insulin and glucagon. J. Cell. Physiol. 160, 358–366.[ISI][Medline]

Okuyama, H., Shimahara, Y., and Kawada, N. (2002). The hepatic stellate cell in the post-genomic era. Histol. Histopathol. 17, 487–495.[ISI][Medline]

Olinga, P., Groothuis, G., Bezuijen, J., Merema, M., Draaisma, A., Meijer, D., and Friedman, S. (2002). A new model for analysis of stellate cell activation in a physiologic milieu using a liver slice system. Hepatology 36(Part 2 Suppl. S), 378.[CrossRef]

Paradis, V., Perleumuter, G., Bonvoust, F., Dargere, D., Parfait, B., Vidaud, M., Conti, M., Huet, S., Ba, N., Buffet, C., and Bedossa, P. (2001). High glucose and hyperinsulinemia stimulate connective tissue growth factor expression: A potential mechanism involved in progression to fibrosis in nonalcoholic steatohepatitis. Hepatology 34, 738–744.[CrossRef][ISI][Medline]

Pfaller, W., Balls, M., Clothier, R., Coecke, S., Dierickx, P., Ekwall, B., Hanley, B. A., Hartung, T., Prieto, P., Ryan, M. P., Schmuck, G., Sladowski, D., Vericat, J.-A., Wendel, A., Wolf, A., and Zimmer, J. (2001). Novel advanced in vitro methods for long-term toxicity testing. ATLA 29, 393–426.[Medline]

Runge, D., Runge, D. M., Jaeger, D., Lubecki, K. A., Stolz, D. B., Karathanasis, S., Kietzmann, T., Strom, S. C., Jungermann, K., Fleig, W. E., and Michalopoulos, G. K. (2000). Serum-free, long-term cultures of human hepatocytes: Maintenance of cell morphology, transcription factors, and liver specific functions. Biochem. Biophys. Res. Com. 269, 46–53.[CrossRef][ISI][Medline]

Sato, M., Suzuki, S., and Senoo, H. (2003). Hepatic stellate cells: Unique characteristics in cell biology and phenotype. Cell Structure Funct. 28, 105–112.[CrossRef][ISI]

Sato, T., Kato, R., and Tyson, C. A. (1995). Regulation of differentiated phenotype of rat hepatic lipocytes by retinoids in primary culture. Exp. Cell Res. 217, 72–83.[CrossRef][ISI][Medline]

Schrem, H., Klempnauer, J., and Borlak, J. (2002). Liver-enriched transcription factors in liver function and development. Part I: The hepatocyte nuclear factor network and liver specific gene expression. Pharmacol. Rev. 54, 129–158.[Abstract/Free Full Text]

Schuppan, D., Ruehl, M., Somasundaram, R., and Hahn, E. G. (2001). Matrix as a modulator of hepatic fibrogenesis. Semin. Liver Dis. 21, 351–372.[CrossRef][ISI][Medline]

Sileri, P., Schena, S., Morini, S., Rastellini, C., Pham, S., Benedetti, E., and Cicalese, L. (2001). Pyruvate inhibits hepatic ischemia-reperfusion injury in rats. Transplantation 72, 27–30.[CrossRef][ISI][Medline]

Starkel, P., Laurent, S., Petit, M., Van Den Berge, V., Lambotte, L., and Horsmans, Y. (2000). Early down-regulation of cytochromes P450 3A and 2E1 in the regenerating rat liver is not related to the loss of liver mass or the process of cellular proliferation. Liver 20, 405–410.[CrossRef][ISI][Medline]

Sumimoto, R., and Kamada, N. (1990). Lactobionate as the most important component in UW solution for liver preservation. Transplantation Proc. 22, 2198–2199.[ISI][Medline]

Takeuchi, N., Katayama, Y., Matsumiya, K., Uchida, K., and Yamamura, Y. (1976). Feed-back control of cholesterol synthesis in partially hepatecomized rats. Biochem. Biophys. Acta 21, 57–68.