Evaluation of a Novel Renewable Hepatic Cell Model for Prediction of Clinical CYP3A4 Induction Using a Correlation-Based Relative Induction Score Approach

Metabolism enzyme induction-mediated drug-drug interactions need to be carefully characterized in vitro for drug candidates to predict in vivo safety risk and therapeutic efficiency. Currently, both the Food and Drug Administration and European Medicines Agency recommend using primary human hepatocytes as the gold standard in vitro test system for studying the induction potential of candidate drugs on cytochrome P450 (CYP), CYP3A4, CYP1A2, and CYP2B6. However, primary human hepatocytes are known to bear inherent limitations such as limited supply and large lot-to-lot variations, which result in an experimental burden to qualify new lots. To overcome these shortcomings, a renewable source of human hepatocytes (i.e., Corning HepatoCells) was developed from primary human hepatocytes and was evaluated for in vitro CYP3A4 induction using methods well established by the pharmaceutical industry. HepatoCells have shown mature hepatocyte-like morphology and demonstrated primary hepatocyte-like response to prototypical inducers of all three CYP enzymes with excellent consistency. Importantly, HepatoCells retain a phenobarbital-responsive nuclear translocation of human constitutive androstane receptor from the cytoplasm, characteristic to primary hepatocytes. To validate HepatoCells as a useful tool to predict potential clinical relevant CYP3A4 induction, we tested three different lots of HepatoCells with a group of clinical strong, moderate/weak CYP3A4 inducers, and noninducers. A relative induction score calibration curve-based approach was used for prediction. HepatoCells showed accurate prediction comparable to primary human hepatocytes. Together, these results demonstrate that Corning HepatoCells is a reliable in vitro model for drug-drug interaction studies during the early phase of drug testing.

Methods for advanced hepatocyte cell culture in microwells utilizing air bubbles

Flat, two-dimensional (2D) cell culture substrates are simple to use but offer little control over cell morphologies and behavior. In this article, we present a number of novel and unique methods for advanced cell culture in microwells utilizing air bubbles as a way to seed cells in order to provide substantial control over cellular microenvironments and organization to achieve specific cell-based applications. These cell culture methods enable controlled formation of stable air bubbles in the microwells that spontaneously formed when polar solvents such as cell culture media are loaded. The presence of air bubbles (air bubble masking) enables highly controllable cell patterning and organization of seeded cells as well as cell co-culture in microwells. In addition, these cell culture methods are simple to use and implement, yet versatile, and have the potential to provide a wide range of microenvironments to improve in vivo-like behavior for a number of cell types and applications. The air bubble masking technique can also be used to produce a micron thick layer of collagen film suspended on top of the microwells. These collagen film enclosed microwells could provide an easy way for high throughput drug screening and cytotoxicity assays as different drug compounds could be pre-loaded and dried in selected microwells and then released during cell culture.

Microstructured multi-well plate for three-dimensional packed cell seeding and hepatocyte cell culture

In this article, we present a microstructured multi-well plate for enabling three-dimensional (3D) high density seeding and culture of cells through the use of a standard laboratory centrifuge to promote and maintain 3D tissue-like cellular morphology and cell-specific functionality in vitro without the addition of animal derived or synthetic matrices or coagulants. Each well has microfeatures on the bottom that are comprised of a series of ditches/open microchannels. The dimensions of the microchannels promote and maintain 3D tissue-like cellular morphology and cell-specific functionality in vitro. After cell seeding with a standard pipette, the microstructured multi-well plates were centrifuged to tightly pack cells inside the ditches in order to enhance cell-cell interactions and induce formation of 3D cellular structures during cell culture. Cell-cell interactions were optimized based on cell packing by considering dimensions of the ditches/open microchannels, orientation of the microstructured multi-well plate during centrifugation, cell seeding density, and the centrifugal force and time. With the optimized cell packing conditions, we demonstrated that after 7 days of cell culture, primary human hepatocytes adhered tightly together to form cord-like structures that resembled 3D tissue-like cellular architecture. Importantly, cell membrane polarity was restored without the addition of animal derived or synthetic matrices or coagulants.

Perfusion-based microfluidic device for three-dimensional dynamic primary human hepatocyte cell culture in the absence of biological or synthetic matrices or coagulants

We describe a perfusion-based microfluidic device for three-dimensional (3D) dynamic primary human hepatocyte cell culture. The microfluidic device was used to promote and maintain 3D tissue-like cellular morphology and cell-specific functionality of primary human hepatocytes by restoring membrane polarity and hepatocyte transport function in vitro without the addition of biological or synthetic matrices or coagulants. A unique feature of our dynamic cell culture device is the creation of a microenvironment, without the addition of biological or synthetic matrices or coagulants, that promotes the 3D organization of hepatocytes into cord-like structures that exhibit functional membrane polarity as evidenced by the expression of gap junctions and the formation of an extended, functionally active, bile canalicular network.

Force spectroscopy of hepatocytic extracellular matrix components

We present atomic force microscopy and force spectroscopy data of live hepatocytes (HEPG2/C3A liver cell line) grown in Eagle's Minimum Essential Medium, a complex solution of salts and amino acids commonly used for cell culture. Contact-mode imaging and force spectroscopy of this system allowed correlation of cell morphology and extracellular matrix (ECM) properties with substrate properties. Force spectroscopy analysis of cellular "footprints" indicated that the cells secrete large polymers (e.g., 3.5mum contour length and estimated MW 1000kDa) onto their substrate surface. Although definitive identification of the polymers has not yet been achieved, fluorescent-labeled antibody staining has specified the presence of ECM proteins such as collagen and laminin in the cellular footprints. The stretched polymers appear to be much larger than single molecules of known ECM components, such as collagen and heparan sulfate proteoglycan, thus suggesting that the cells create larger entangled, macromolecular structures from smaller components. There is strong evidence which suggests that the composition of the ECM is greatly influenced by the hydrophobicity of the substrate surface, with preferential production and/or adsorption of larger macromolecules on hydrophobic surfaces.

Effects of sodium butyrate and acidic fibroblast growth factor on TDC32300 cultured cells

BACKGROUND

Many studies have demonstrated that transition duct cells (TDC) are facultative liver stem cells. Our laboratory established TDC32300 cell lines with hepatic progenitor markers. The authors proposed that cell culture using sodium butyrate (NaBut) and acidic fibroblast growth factors (aFGF) may support the differentiation of TDC32300 cells along the hepatic lineage.

METHODS

TDC32300 cells were cultured in four different conditions 1) STON media alone; 2) STON with NaBut in 3 different concentrations, 1 mM, 3.75 mM and 5 mM; 3) STON with aFGF; and 4) STON with aFGF and dexamethasone. After day 5, the cultured cells were fixed and stained with monoclonal antibodies to rat liver antigens and anti-proliferating nuclear antigen (PCNA).

RESULTS

Proliferation of TDC32300 cells cultured in the high concentration of NaBut (3.75 and 5 mM) was inhibited. This phenomenon was confirmed by the reduction in cell number and decrease in PCNA expression. Irrespective of the concentration, NaBut did not alter the phenotype of the TDC32300 cultured cells. aFGF with or without dexamethasone also did not alter the phenotypic characteristic of TDC32300 cells.

CONCLUSION

TDC32300 cells may not be the hepatic progenitors or that their differentiation may require other culture conditions.

White matter injury after cerebral ischemia in ovine fetuses

The effects of cerebral ischemia on white matter changes in ovine fetuses were examined after exposure to bilateral carotid artery occlusion. Fetal sheep were exposed to 30 min of ischemia followed by 48 (I/R-48, n = 8) or 72 (I/R-72, n = 10) h of reperfusion or control sham treatment (control, n = 4). Serial coronal sections stained with Luxol fast blue/hematoxylin and eosin were scored for white matter, cerebral cortical, and hippocampal lesions. All areas received graded pathologic scores of 0 to 5, reflecting the degree of injury where 0 = 0%, 1 = 1% to 25%, 2 = 26% to 50%, 3 = 51% to 75%, 4 = 76% to 95%, and 5 = 96% to 100% of the area damaged. Dual-label immunofluorescence using antibodies against glial fibrillary acidic protein (GFAP) and myelin basic protein (MBP) were used to characterize white matter lesions. Basic fibroblast growth factor (FGF-2) was measured in the frontal cortex by ELISA. Results of the pathologic scores showed that the white matter of the I/R-72 (2.74 +/- 0.53, mean +/- SEM) was more (p < 0.05) damaged when compared with the control (0.80 +/- 0.33) group. Cortical lesions were greater (p < 0.05) in the I/R-48 (2.12 +/- 0.35) than the control (0.93 +/- 0.09) group. White matter lesions were characterized by reactive GFAP-positive astrocytes and a loss of MBP in oligodendrocytes. The ratio of MBP to GFAP decreased (p < 0.05) as a function of ischemia, indicative of a proportionally greater loss of MBP than GFAP. FGF-2 concentrations were higher (p < 0.05) in the I/R-72 than the control group and there was a direct correlation between the pathologic scores (PS) and FGF-2 concentrations (FGF-2 = e((1.6 PS-0.90)) + 743, n = 17, r = 0.73, p < 0.001). We conclude that carotid artery occlusion results in quantifiable white matter lesions that are associated with a loss of MBP from myelin, and that FGF-2, a purported mediator of recovery from brain injury in adult subjects, increases in concentration in proportion to the severity of brain damage in the fetus.

Developmental variability in expression and regulation of inducible nitric oxide synthase in rat intestine

Inducible nitric oxide synthase (iNOS) may be a key mediator of intestinal injury, which varies with developmental age. One member of the mitogen-activated protein kinase (MAPK) family, p38, is involved in the lipopolysaccharide (LPS)-mediated iNOS induction. The involvement of p38 MAPK in basal and LPS-induced iNOS expression was examined in the rat intestine at two developmental ages. Neonatal (4 days postnatal) and adolescent (15 days postnatal) rats were injected with LPS (5 microg/g ip), a selective p38 inhibitor (SB 203580), or both. Tissue was removed after 4 h and 6 h for mRNA and protein analysis. iNOS mRNA and protein were markedly upregulated in the adolescent female following LPS exposure, whereas males had an attenuated response. Neonates had a minimal response. SB 203580 suppressed LPS-induced iNOS mRNA and protein in the ileum, more so in females than in males. Adolescent ileal p38 activation was constitutively high and nonresponsive to LPS. Basal and post-LPS p38 phosphorylation was low in neonatal ileum. We conclude that ileal iNOS expression is developmentally regulated and influenced by gender and that p38 is permissive for LPS effect. The developmental regulation of p38 may contribute to age-dependent variations of intestinal injury.

Growth and metabolic activity of immortalized porcine hepatocytes in extracorporeal hollow-fiber liver assist devices

The development of a cell based extracorporeal liver assist device offers a promising clinical approach to bridge individuals suffering from acute liver failure to transplant. However, a major drawback of the existing technology is the lack of a continuous supply of well differentiated hepatocytes. Although some investigators have used primary porcine cells, this approach demands costly, labor-intensive isolation procedures and yields cells with inconsistent detoxification capacity. The limitations of primary cells led us to develop the HepLiu immortalized porcine hepatocyte cell line for use in liver assist devices (LADs). HepLiu cells are nontumorigenic and exhibit multiple hepatic detoxification functions including diazepam and acetaminophen metabolism. To investigate the suitability of HepLiu cells for artificial liver support, morphology, as well as xenobiotic metabolism, was studied in perfused polysulfone hollow-fiber LADs. HepLiu cells were cultured in the intercapillary space of a prototype LAD, and the metabolism of diazepam, acetaminophen, and 7-ethoxycoumarin was evaluated over 25 days in culture. Our results indicated that HepLiu cells proliferated rapidly following inoculation of the LAD until Day 10 when proliferation appeared to cease. Ultrastructural analysis demonstrated that HepLiu cells retained many of the features of primary hepatocytes including desmosomes that sealed bile canalicular-like structures and junctional complexes (intermediate, gap junctions) that appeared concentrated in the paracanalicular areas. Unlike primary porcine hepatocytes, HepLiu cells retained drug metabolic function throughout the 25 day culture period. Diazepam metabolism by HepLiu cells was consistently higher than that of primary cells. Acetaminophen metabolism persisted throughout the 25 day period albeit at a much lower level than the primary cells exhibited on Days 1 or 2. In conclusion, we have shown that HepLiu cells proliferate to occupy the intercapillary space of perfused hollow-fiber LADs following inoculation, and retain their metabolic capacity for Phase I and Phase II detoxification reactions in perfusion culture. Our findings suggest that HepLiu cells may provide an alternative to primary porcine hepatocytes as the cellular component of bioartificial liver support systems.

Liver stem cells: a potential source of hepatocytes for the treatment of human liver disease

Severe liver injury often leads to the proliferation of oval cells, which differentiate along hepatocytic and biliary lineages. Because oval cells proliferate only when hepatocyte replication is impaired, they are considered to be the progeny of facultative liver stem cells (FLSCs). Identification and isolation of FLSCs has been hampered by the lack of markers that delineate these bipotential progenitors. We hypothesized that transition ductal cells are FLSCs because they are located in a unique anatomical niche sharing tight junctions with a neighboring hepatocyte and another terminal ductular cell. Alternatively, it has been proposed recently that bone marrow-derived stem cells are FLSCs since these cells differentiate along the hepatic lineage following colonization of the liver. The intent of this review is to provide insight into the nature and origin of liver stem cells and to explore the possibility that stem cell technology may lead to the development of clinical modalities for the treatment of human liver disease.

The critical role of Shc in insulin-like growth factor-I-mediated mitogenesis and differentiation in 3T3-L1 preadipocytes

Insulin-like growth factor-I (IGF-I) stimulates mitogenesis in proliferating preadipocytes, but when cells reach confluence and become growth arrested, IGF-I stimulates differentiation into adipocytes. IGF-I induces signaling pathways that involve IGF-I receptor-mediated tyrosine phosphorylation of Shc and insulin receptor substrate 1 (IRS-1). Either of these adaptor proteins can lead to activation of the three-kinase cascade ending in activation of the extracellular signal-regulated kinase 1 and -2 (ERK-1 and -2) mitogen-activated protein kinases (MAPKs). Several lines of evidence suggest that activation of MAPK inhibits 3T3-L1 preadipocyte differentiation. We have shown that IGF-I stimulation of MAPK activity is lost as 3T3-L1 preadipocytes begin to differentiate. This change in MAPK signaling coincides with loss of IGF-I-mediated Shc, but not IRS-1, tyrosine phosphorylation. We hypothesized that down-regulation of MAPK via loss of proximal signaling through Shc is an early component in the IGF-I switch from mitogenesis to differentiation in 3T3-L1 preadipocytes. Treatment of subconfluent cells with the MEK inhibitor PD098059 inhibited both IGF-I-activation of MAPK as well as 3H-thymidine incorporation. PD098059, in the presence of differentiation-inducing media, accelerated differentiation in subconfluent cells as measured by expression of adipocyte protein-2 (aP-2), peroxisome proliferator-activated receptor gamma (PPARgamma) and lipoprotein lipase (LPL). Transient transfection of subconfluent cells with Shc-Y317F, a dominant-negative mutant, attenuated IGF-I-mediated MAPK activation, inhibited DNA synthesis, and accelerated expression of differentiation markers aP-2, PPARgamma, and LPL. We conclude that signaling through Shc to MAPK plays a critical role in mediating IGF-I-stimulated 3T3-L1 mitogenesis. Our results suggest that loss of the ability of IGF-I to activate Shc signaling to MAPK may be an early component of adipogenesis in 3T3-L1 cells.

The relationship between differentiation and proliferation in late gestation fetal rat hepatocytes

Hepatocyte proliferation and differentiation occur simultaneously during late mammalian gestation. We hypothesized that regulation of hepatocyte growth and differentiation would be coordinated in late gestation fetal hepatocyte cultures such that proliferation would be most active in a population of less well-differentiated cells. Cultured fetal hepatocytes (embryonic d 19 and 21; E19 and E21) were studied using double staining immunofluorescent microscopy. Differentiation was assessed as staining for alpha-fetoprotein (AFP), three markers of enzymic differentiation (glucokinase [GK], phosphoenolpyruvate carboxykinase [PEPCK], and carbamoyl phosphate synthase [CPS]), and a hepatocyte cell-cell adhesion molecule (C-CAM). Proliferation was assessed using immunocytochemical detection of proliferating cell nuclear antigen (PCNA) or 5-bromo-2'-deoxy-uridine (BrdU) incorporation into DNA. Fetal hepatocyte cultures consisted of a heterogeneous population of cells, slightly more than half of which were proliferative under defined, growth factor-free conditions. These cultures were heterogeneous for AFP expression. There was no correlation between the expression of AFP and PCNA or AFP and S-phase entry (BrdU staining) during the first 48 h in culture. Similar results were obtained in staining for the enzymic differentiation markers and C-CAM. In addition, the differentiation status of cultured fetal hepatocytes was unrelated to a presumed indicator of mature growth regulation, mitogenic responsiveness to transforming growth factor alpha (TGFalpha), and hepatocyte growth factor (HGF). Finally, absence of any correlation between proliferation and differentiated phenotype was supported by in vivo studies using staining for PCNA, AFP, CPS, and PEPCK in liver sections. These results indicate that the developmental program governing differentiation of late gestation fetal rat hepatocytes is independent from mechanisms controlling proliferation.

Evaluation of the differentiation potential of WB-F344 rat liver epithelial stem-like cells in vivo. Differentiation to hepatocytes after transplantation into dipeptidylpeptidase-IV-deficient rat liver

After intrahepatic transplantation into livers of adult syngeneic German-strain Fischer 344 rats that are deficient for the bile canalicular enzyme dipeptidyl peptidase IV (DPP-IV), cultured WB-F344 rat liver epithelial cells (without exogenous marker genes) integrate into hepatic plates and differentiate into hepatocyte-like cells that are morphologically and functionally indistinguishable from mature hepatocytes. In this model system, the differentiated progeny of transplanted WB-F344 cells are identified among the DPP-IV-negative host hepatocytes by their expression of bile canalicular DPP-IV enzyme activity. DPP-IV-positive hepatocyte-like cells also expressed other markers of hepatocytic differentiation, including albumin, transferrin, and alpha-1-antitrypsin, suggesting that the progeny of transplanted WB-F344 cells express a complete hepatocyte differentiation program. These results complement our previous studies indicating WB-F344 cells can serve as stem-like precursor cells for differentiated hepatocytes and strengthen the suggestion that WB-F344 rat liver epithelial cells represent the cultured counterpart of liver stem-like hepatocyte progenitor cells present in the normal adult rat liver.

Modulation of mitogen-independent hepatocyte proliferation during the perinatal period in the rat

Late gestation fetal rat hepatocytes can proliferate under defined in vitro conditions in the absence of added mitogens. However, this capacity declines with advancing gestational age of the fetus from which the hepatocytes are derived. The present studies were undertaken to investigate this change in fetal hepatocyte growth regulation. Examination of E19 fetal hepatocyte primary cultures using immunocytochemistry for 5-bromo-2'-deoxyuridine (BrdU) incorporation showed that approximately 80% of these cells traverse S-phase of the cell cycle over the first 48 h in culture. Similarly, 65% of E19 hepatocytes maintained in culture under defined mitogen-free conditions for 24 h showed nuclear expression of proliferating cell nuclear antigen (PCNA). These in vitro findings correlated with a high level of immunoreactive PCNA in immunofluorescent analyses of E19 liver. In contrast, E21 (term) liver showed little immunoreactive PCNA. The in vivo finding was recapitulated by in vitro studies showing that E21 hepatocytes had low levels of BrdU incorporation during the first day in culture and were PCNA negative shortly after isolation. However, within 12 h of plating, E21 hepatocytes showed cytoplasmic staining for PCNA. Although maintained under mitogen-free conditions, PCNA expression progressed synchronously to a nucleolar staining pattern at 24 to 48 h in culture followed by intense, diffuse nuclear staining at 60 h which disappeared by 72 h. This apparently synchronous cell cycle progression was confirmed by studies showing peak BrdU incorporation on the third day in culture. Whereas DNA synthesis by both E19 and E21 hepatocytes was potentiated by transforming growth factor alpha (TGF alpha), considerable mitogen-independent DNA synthesis was seen in hepatocytes from both gestational ages. These results may indicate that fetal hepatocytes come under the influence of an exogenous, in vivo growth inhibitory factor as term approaches and that this effect is relieved when term fetal hepatocytes are cultured.

Putative liver progenitor cells: conditions for long-term survival in culture

Oval cells proliferate extensively in the livers of animals exposed to oncogenic insults, are bipotent and are believed to be related to the so far unidentified liver stem cell. In normal liver, cells antigenically related to oval cells and expressing liver and epithelial markers are considered to be liver progenitor cells. We isolated, by fluorescence-activated cell sorting or magnetic bead sorting, cells expressing the oval cell antigens OC.2 or OC.3 from the liver of normal newborn or day 12 embryonal age rats. Magnetic bead sorting of positive cells was as efficient as fluorescence-activated cell sorting. A two-chamber culture system was devised in which cells were plated onto transwell filters coated with type IV collagen and cultured in a serum-free Ham's F12 medium supplemented with free fatty acids and bovine serum albumin. Under these conditions, cells remained viable for up to 6 weeks and their antigenic phenotype was unchanged throughout. Approximately 30% of sorted cells expressed epithelial and/or liver-specific markers. Growth factors mitogenic for epithelial cells and hepatocytes did not elicit cell proliferation. These results provide an important background for further studies designed to determine the biological significance of OC.2+ and OC.3+ cells in normal liver, to test the liver stem cell hypothesis and to develop protocols for the expansion in vitro of normal liver progenitors.

Correlation of expression of connexin mRNA isoforms with degree of cellular differentiation

Examination of rat hepatic cell lines has revealed a correlation between the differentiated state of the cells and the gap junctional proteins, or connexins, they express. The cell lines RLC (Gershenson et al, 1970) and FTO.2B (Killary et al, 1984) were examined and compared to primary adult hepatocytes for expression of fetal and adult hepatic antigens under various tissue culture conditions. Maximal expression of fetal antigens was observed in cells grown in serum-supplemented medium, on either tissue culture plastic or type IV collagen. Maximal expression of adult specific antigens was seen in cells grown in a hormonally defined medium containing heparin, on type I or type IV collagen. The cell line RLC strongly expressed fetal antigens, while FTO.2B expressed both fetal and adult antigens. These cell lines and another poorly differentiated hepatic cell line, WB-F344 (Tsao et al., 1984) were used to assess the developmental profile of mRNAs encoding isoforms of gap junctions: connexins 26, 32, and 43. The cell lines each transcribed mRNAs of all three connexins, as determined by transcriptional elongation analysis. By contrast, only certain of the connexin mRNAs could be detected in specific cell lines by Northern analysis: RLC expressed only connexin 43 mRNA; WB-F344 expressed connexin 32 and 43 mRNAs. Selection among the connexin mRNAs appears to occur post-transcriptionally. Culture of the cell lines in hormonally defined medium vs. serum supplemented medium did not affect the patterns of connexin mRNA abundance. When the cell lines were cultured in hormonally defined medium containing heparin, however, the level of connexin mRNAs did vary: Connexin 26 mRNA increased in WB-F344 cells, and connexins 32 and 43 mRNAs increased in FTO.2B, but connexin 43 mRNA decreased in WB-F344 and RLC. The abundance of connexin mRNAs also varied when the cell lines were analyzed at different cell densities: connexin 43 mRNA increased with cell density in RLC and WB-F344, and connexin 26 mRNA peaked at an intermediate density and fell at higher cell densities in WB-F344. The differences in connexin mRNA expression among cell lines characteristic of different stages of hepatic differentiation, and the differences in regulation of connexin mRNAs in the hepatic cell lines, suggest distinct biological roles of the highly homologous proteins. Moreover, connexin gene expression may be a marker of hepatic development: as hepatocytes differentiate the proportions of connexin 43 then 26 mRNAs decrease while that of connexin 32 mRNA increases.

Delineation of antigenic pathways of ethionine-induced liver cancer in the rat

Although oval cell proliferation is observed prior to the appearance of hepatic nodules and hepatocellular carcinomas during ethionine-induced liver carcinogenesis in the rat, the role of these presumptive hepatic stem cells during the neoplastic process remains controversial. In order to investigate this question, we have used a panel of monoclonal antibodies against antigens associated with normal hepatocytes, oval cells and transplantable hepatocellular carcinomas (THC) to trace antigenic pathways leading to liver cancer. Male ACI rats were fed a choline-deficient diet containing 0.1% DL-ethionine for 4, 16 or 30 weeks. Immunocytochemical analysis of frozen liver sections revealed a subpopulation of hepatic nodules (7/52), carcinomas (8/15) and lung metastases (3/5) containing populations of cells expressing both oval cell, hepatocyte and neoplastic markers. Carcinomas expressing oval cell markers often appeared as a mosaic of well-defined patches composed of phenotypically distinct cells. Many of the phenotypes expressed closely mimicked patterns of expression observed in fetal and neonatal liver. THC derived from primary tumors positive for oval cell antigens (4/5) continued to express these markers. Northern blot analysis and immunocytochemical analysis revealed that 4/5 primary hepatocellular carcinomas (PHC) and THC expressed alpha-fetoprotein (AFP) and albumin transcripts and contained subpopulations expressing AFP together with hepatocyte and oval antigens. In contrast, a well-differentiated PHC and its corresponding THC lacked AFP mRNA and oval cell antigens but showed strong expression of both hepatocyte and neoplastic markers. These results demonstrate that a subpopulation of malignant and metastatic hepatocellular carcinomas are comprised of cells expressing multiple oval cell markers in this model system.

Isolation, propagation, and characterization of rat liver serosal mesothelial cells

Although rat liver epithelial cell (RLEC) lines have been developed by a number of laboratories, the identity of the clonogenic nonparenchymal progenitors is unknown. To provide insight into the derivation of RLEC, we immunoisolated serosal liver mesothelial cells (LMC) and bile duct epithelial cells and attempted to propagate each epithelial cell population using culture conditions routinely employed to establish RLEC lines. Briefly, the selective reactivity of LMC with two bile duct cell surface markers, OC.2 and BD.2, was exploited to develop an immunocytochemical technique to isolate LMC. Livers were collagenase dissociated, the mesothelial capsule was "peeled" and digested with pronase to destroy contaminating hepatocytes, and rare biliary ductal epithelial cells were immunodepleted using OC.2. LMC were subsequently isolated by selective binding to magnetic beads adsorbed with BD.2 and cultured in supplemented Waymouths 752/1 media containing 10% fetal calf serum. Proliferating BD.2+ LMC rapidly formed epithelial-like monolayers that could be continuously subcultured after trypsinization. In contrast, attempts to establish cell lines from purified OC.2+ bile duct epithelial cells were unsuccessful. Results from reverse transcriptase polymerase chain reaction analysis confirmed that LMC expressed Wilms' tumor transcripts, a lineage marker for mesodermally-derived cells. In summary, our findings clearly demonstrate that LMC can be continuously propagated using culture conditions routinely employed to establish RLEC lines, an observation that supports the contention that some RLEC lines may be derived from LMC.

Phenotypic heterogeneity within clonogenic ductal cell populations isolated from normal adult rat liver

Oval cells represent a heterogeneous population composed of ductal cells, transitional cells with characteristics of both hepatocytes and bile ductal cells, and bipotential stem cells capable of differentiation along a biliary or hepatocytic lineage. In an attempt to define markers that would distinguish between individual cell types within the oval cell population, a number of investigators have utilized hybridoma technology to produce cell type-specific monoclonal antibodies. Several of these have proved to be of value in delineating lineage relationships during fetal development and carcinogenesis in the adult liver. Most recently, monoclonal antibodies specific for OC2 and OC3, two oval cell antigens identified in our laboratory, have been used in combination with magnetic beads or a fluorescence-activated cell sorter to isolate antigenically defined subpopulations from adult and fetal rat liver. Using OC2-positive fetal liver cells as an immunogen, we have produced a monoclonal antibody identifying a bile ductal antigen, designated BD1, that is differentially expressed by oval cells and normal ductal cells. This antigen shows a heterogeneous pattern of reactivity that defines three distinct cell populations in regenerating rat liver: a BD1-negative, [3H]thymidine-labeled cell population thought to contain hepatic stem cells; a BD1-positive, thymidine-negative population of terminally differentiated ductal cells; and a BD1-positive, [3H]thymidine-positive population of mature ductal cells. Analysis of BD1 expression in vitro on continuous lines of bile duct epithelial cells (BDEC) demonstrated that BD1 was rapidly increased in late G1 and lost during G2/M. High passage cultures of BDEC and primary cultures of oval cells expressed low or undetectable levels of BD1 and high passage BDEC failed to express BD1 when arrested in late G1. Taken together, these results suggested that oval cells and high passage BDEC might share a subtle defect in cell cycle regulation marked by an inability to upregulate the expression of BD1.

Characterization of a mature bile duct antigen expressed on a subpopulation of biliary ductular cells but absent from oval cells

Hybridomas were produced with immune spleen cells generated by immunization of Balb/c mice with oval cell antigen (OC.2)-positive 17-day fetal liver cells isolated on antibody-coated magnetic beads. A primary screen by indirect immunofluorescence on frozen sections of fetal and adult liver identified several hybridomas secreting antibodies reactive with bile ducts and oval cells. One of these recognized an epitope, designated BD1, which was expressed on intrahepatic bile ducts in 16-day fetal and adult rat liver and in liver from rats fed a choline-deficient diet containing ethionine and rats treated with 2-acetylaminofluorine, but was absent from morphologically defined oval cells induced by a choline-deficient diet containing ethionine or by 2-acetylaminofluorine. Double-labeling immunofluorescence analysis with monoclonal antibody BD1 and OV6, a monoclonal antibody that reacts uniformly with all bile ducts and oval cells, revealed that BD1 expression on the intrahepatic bile ductular cells of normal adult rat liver was heterogeneous with a major ductal cell population (60% to 70%) expressing high levels of BD1 and a minor ductal cell population (30% to 40%) displaying undetectable or low levels of BD1 expression. Analysis of fetal liver demonstrated the presence of BD1-positive cells at day 16 of gestation on duct-like structures in contact with portal mesenchyme, an observation suggesting that expression of BD1 was associated with commitment of hepatoblasts to a ductular lineage. Taken together, our findings suggest that oval cells may be derived from an antigenically distinct subpopulation of bile ductal cells.

Derivation of ductlike cell lines from a transplantable acinar cell carcinoma of the rat pancreas

Two cell lines were derived from a transplantable acinar cell carcinoma that had been established from a primary carcinoma of the pancreas in an azaserine-treated Lewis rat. The cultured tumor cells initially produced amylase, but production of exocrine enzymes ceased after 1-2 weeks in culture. The cultured cells were tumorigenic in Lewis rats, and one line produced solid tumors composed of ductlike structures surrounded by dense fibrous tissue. The second cell line produced partially solid and partially cystic tumors with a mixed phenotype of squamous, mucinous, and glandular areas when it grew in vivo following regrafting. Both cell lines lost structural and immunohistochemical acinar cell markers while acquiring duct cell markers during culture and regrafting. These studies provide strong support for the hypothesis that ductlike carcinomas can arise from neoplastic pancreatic acinar cells in rats.

Long-term culture and characteristics of normal rat liver bile duct epithelial cells

BACKGROUND

A method was established for isolation and long-term culture of bile duct epithelial cells (BDEC) of normal adult rat liver that does not require the preparation of highly purified BDEC.

METHODS

After dissociation of the liver parenchyma by collagenase perfusion, the liver remnant containing the intact biliary tree was minced into small fragments, embedded in a rat tail collagen gel, and cultured for 6 days in hormonally defined serum-free Dulbecco's Modified Eagle Medium/F12 medium (SFDM). BDEC cultures were subsequently subcultured and maintained on rat tail collagen gels in SFDM medium containing 5 mumol/L forskolin and 5%-10% Nu Serum IV (Collaborative Research, Bedford, MA).

RESULTS

Established BDEC lines continued to express ductal specific markers including gamma-glutamyl transpeptidase, cytokeratins 7 and 19, and a number of monoclonal antibody-defined bile duct antigens, such as OC.2, OC.3, and OV6.

CONCLUSIONS

The availability of a method to establish normal BDEC lines will allow further investigation of the function of bile duct cells and their role in normal liver differentiation and carcinogenesis.

Antigenic relationship between oval cells and a subpopulation of hepatic foci, nodules, and carcinomas induced by the "resistant hepatocyte" model system

Although proliferation of small ductular-like cells, designated oval cells, is often observed during the early stages of chemically induced hepatocarcinogenesis, their role during the carcinogenic process remains controversial. To investigate the possibility that oval cells may give rise to preneoplastic lesions that ultimately progress to hepatocellular carcinomas, we have carried out phenotypic analysis with a panel of monoclonal antibodies to determine if there is an antigenic relationship between oval cells and hepatic foci, nodules, and tumors induced by the resistant hepatocyte model system. In this model, rats are given a single dose (200 mg/kg) of diethylnitrosamine, followed by a brief exposure to 2-acetylaminofluorene and a partial hepatectomy. We found that approximately 10% of the early focal lesions observed 28 days after diethylnitrosamine expressed either one or both of the oval cell antigens designated OC.2 and OV-6. By 28 weeks after diethylnitrosamine, 16 of 16 hepatic nodules heterogeneously expressed OV-6 whereas 5-10% of the persistent nodules contained scattered small hepatocyte-like cells that expressed OC.2. Examination of resistant hepatocyte-induced primary hepatocellular carcinomas with an expanded panel of monoclonal antibodies demonstrated that most cells comprising 29 of 29 tumors expressed OV-6 and that 15-20% of the OV-6-positive tumors contained subpopulations of cells also expressing 3 additional oval cell antigens, OC.2, OC.3, and OV-1. All of the tumors examined expressed normal levels of the hepatocyte antigens, H.1 and HBD.1, and had dramatically reduced levels of H.2, H.4, and cell CAM 105 but showed elevated levels of the transferrin receptor, gamma-glutamyltranspeptidase, and the normal hepatocyte antigen, H.5. In conclusion, our findings demonstrate an antigenic relationship between oval cells and a subpopulation of hepatic foci, nodules, and tumors in the resistant hepatocyte model, suggesting that at least some primary tumors may be derived from oval cells in this model system.

A Fischer rat substrain deficient in dipeptidyl peptidase IV activity makes normal steady-state RNA levels and an altered protein. Use as a liver-cell transplantation model

Dipeptidyl peptidase IV (DPPIV) is a serine exoproteinase expressed at high levels in epithelial cells of kidney, liver and small intestine. Recently Watanabe, Kohima & Fujimoto [(1987) Experientia 43, 400-401] and Gossrau et al. [(1990) Histochem. J. 22, 172-173] reported that Fischer 344 rats are deficient in this enzyme. We have examined DPPIV expression in Fischer 344 rats available from U.S. and German suppliers and find that livers of the U.S. Fischer rats, in contrast with their German counterparts, express active DPPIV (D+). Northern analysis of liver RNA showed comparable levels of 3.4 kb and 5.6 kb DPPIV transcripts in both D+ rats from the U.S. and German (D-) rats. Monoclonal antibody (MAb) 236.3 to DPPIV immunoprecipitated at 150 kDa enzymically active (105 kDa, denatured) protein from surface-labelled D+ hepatocytes and reacted with canalicular and sinusoidal membranes (as shown by immunofluorescence microscopy). MAb 236.3 failed to immunoprecipitate a labelled peptide from D- cell extract or to stain D- liver sections. Polyclonal antibody (PAb) specific for DPPIV immunoprecipitated an enzymically active peptide from D+ hepatocyte extracts and a smaller, inactive peptide from D- hepatocyte extracts. Peptide maps of DPPIV immunoprecipitated from D+ extracts with MAb 236.3 and PAb were identical, but differed from that of the D- hepatocyte component recognized by PAb. The molecular basis of the DPPIV deficiency in the D- rats thus appears to be the translation of an enzymically inactive protein missing the epitope recognized by MAb 236.3. We have exploited these D- rats as hosts for syngeneic transplantation of liver cells from D+ Fischer rats. DPPIV expression is stable in the transplanted cells and allows them to be readily distinguished from the surrounding D- tissue.

Identification and characterization of a rat hepatic oncofetal membrane glycoprotein

A major interest of our laboratory is to delineate the pathways leading to experimentally induced liver cancer in the rat. Although the cellular progenitors of primary hepatocellular carcinoma remain controversial, current findings suggest that proliferation of chemically initiated liver epithelial cells gives rise to hepatic nodules, a rare population of which eventually progress to carcinoma. Presently, the availability of cell surface markers that are closely associated with malignant progression is needed for the identification, isolation, and further characterization of these rare malignant cells. In this paper, we describe two new monoclonal antibodies (MAbs), MAb 324.5 and MAb 324.9, that recognize a novel oncofetal membrane glycoprotein, designated TuAg1. MAbs 324.5 and 324.9 were produced using three different transplantable hepatocellular carcinoma cell lines during immunization and screening. MAb 324.5 and MAb 324.9 were shown to be reactive with different epitopes on TuAg1 by competitive immunoprecipitation assays combined with results from immunodepletion analysis and one-dimensional V-8 peptide maps. TuAg1 showed variations in molecular weight from 78,000 to 92,000, and a marked heterogeneity in pI, with charge variants ranging between 4.3 and 6.0. The 324.5-epitope was not expressed at detectable levels in any adult normal tissues or during liver regeneration but was transiently expressed during fetal liver development as shown by indirect immunofluorescence analysis of frozen tissue sections. In contrast, the 324.9-epitope was observed on nerve fibers and ganglia and on sperm tails in the adult rat and also appeared independently of the 324.5-epitope during fetal development. Although normal hepatocytes did not express TuAg1, isolated hepatocytes became positive during the first 24 h of primary culture. Attempts to modulate the in vitro expression of TuAg1 were unsuccessful; however, TuAg1 was lost within 7 days following ectopic transplantation of cultured hepatocytes into the pancreas. During the carcinogenic process, TuAg1 was expressed by a rare population of hepatic nodules, by many primary liver tumors, and by all lung metastases and transplantable hepatocellular carcinomas examined to date. Taken together, these observations suggest that the in vivo constitutive expression of this novel oncofetal membrane antigen is closely associated with acquisition of the malignant phenotype during hepatocarcinogenesis.

Selective proliferation of chemically altered rat liver epithelial cells following hepatic transplantation

Although proliferation of oval cells is often observed during the early stages of chemical hepatocarcinogenesis, the role of these putative hepatic stem cells during the neoplastic process is unknown. In earlier studies our laboratory showed that feeding a choline-deficient (CD) diet containing 0.05% 2-acetylaminofluorene (CD-AAF) to rats produced three subpopulations of oval cells that antigenically resemble biliary duct cells, fetal liver cells, and transitional cells. In the present investigation we have employed a semiallogeneic transplantation protocol in order to study the fate of these nonparenchymal epithelial cells (NPEC) beyond the 4-week endpoint imposed by the lethality of CD-AAF diet. An enriched NPEC suspension containing gamma-glutamyl-transpeptidase (GGT)-positive oval cells (greater than 75%) was isolated from ACI rats maintained on CD-AAF diet for 3 weeks. The donor cells were transplanted via the portal vein into livers of male F1 progeny (LExACI) that had been fed a CD diet for 7 days prior to receiving a partial hepatectomy and the cell suspension. Host rats were then fed either a CD or choline-supplemented (CS) diet for 12 weeks and killed. Colonies of donor-derived cells identified in frozen sections by their lack of reactivity with ACI anti-LE alloantiserum in indirect immunofluorescence (IF) assays were only observed in rats continuously fed the CD diet. Histochemical analysis indicated that the donor-derived colonies expressed GGT, a preneoplastic marker for liver cancer. IF assays using MAbs previously shown to be capable of distinguishing between oval cells and mature hepatocytes indicated that the donor-derived colonies consisted of a mixture of cells with phenotypes resembling those of mature and immature hepatocytes rather than those of oval or ductal cells. Although the cellular origin of the GGT+ donor-derived colonies has not been unequivocally resolved, our results demonstrate that the livers of rats fed a CD-AAF diet contain a chemically altered call population that can be induced to proliferate by a CD diet. In contrast, a CD diet did not promote colonization when normal hepatocytes were employed as the donor cell population, suggesting that the GGT+ oval cells and not the few contaminating GGT- hepatocytes (1%) in the CD-AAF donor cell suspension were the preneoplastic precursors that gave rise to donor-derived colonies. This transplantation protocol will be useful to define the biological potential of chemically altered liver cells during carcinogenesis.

Characterization of a new monoclonal antibody to rat macrophages and Kupffer cells

We have characterized the cell and tissue binding specificity of a newly generated monoclonal antibody, Mab Ku-1, which shows selective reactivity with rat macrophages and Kupffer cells. The hybridoma secreting Mab Ku-1 was constructed by fusion of 8653 myeloma cells with spleen cells isolated from a mouse immunized with nonparenchymal liver cells coated with antihepatocyte antibodies. When binding was assessed by indirect immunofluorescence on frozen sections from normal liver tissue, Mab Ku-1 showed strong reactivity with Kupffer cells but was unreactive with hepatocytes, endothelial cells, bile ducts or lymphocytes. Both resident and activated macrophages bound Mab Ku-1. Reactivity in other tissues was compatible with specificity for macrophages. In the gut, scattered cells in the lamina propria were positive, whereas epithelial cells were negative. Individual cells in the lung were reactive. In the spleen, cells in the red pulp peripheral to germinal centers bound antibody. Reactivity of Mab Ku-1 to isolated Kupffer cells correlated with endogenous peroxidase activity. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of components immunoprecipitated by Mab Ku-1 from detergent lysates of Kupffer cells biosynthetically labeled with 35S-cysteine and 35S-methionine demonstrated that the reactive antigen was a peptide with an apparent molecular weight of 107 kD. This rat macrophage-reactive monoclonal antibody is a useful marker for identification of macrophage populations in tissue as well as in isolated cell populations.

Exposure of newborn rats to pharmacologically active compounds may permanently alter carcinogen metabolism

Administration of phenobarbital to mother rats during early lactation causes long-term, perhaps permanent, alteration of hepatic microsomal mixed-function oxidase activity and aflatoxin B1 adduct formation in the adult male offspring. These findings suggest that perinatal exposure to pharmacologically active compounds may be a determinant of cancer risk.