Stimulation of In Vitro Rat Hepatocyte Proliferation by Conditioned Medium Obtained from an Immortalized Macrophage Cell Line

Stimulation of In Vitro Rat Hepatocyte Proliferation by Conditioned Medium Obtained from an Immortalized Macrophage Cell Line

Toxicology in Vitro 13 (1999) 475±481 www.elsevier.com/locate/toxinvit Stimulation of In Vitro Rat Hepatocyte Proliferation by Conditioned Medium Obt...

212KB Sizes 0 Downloads 37 Views

Toxicology in Vitro 13 (1999) 475±481 www.elsevier.com/locate/toxinvit

Stimulation of In Vitro Rat Hepatocyte Proliferation by Conditioned Medium Obtained from an Immortalized Macrophage Cell Line E. PIATTI1, M. RIGHI2, L. MARABINI1, S. RADICE1 and E. CHIESARA1*

1 Department of Pharmacology ``E. Trabucchi'', School of Medicine, University of Milan, Via Vanvitelli, 32, 20129 Milan and 2CNR-Center of Cellular and Molecular Pharmacology, University of Milan, Via Vanvitelli, 32, 20129 Milan, Italy

(Accepted 4 October 1998) AbstractÐThe hepatomitogenic e€ect of conditioned medium (CDM), obtained from the N-11 mouse macrophage cell line was analysed in rat hepatocyte primary cultures. CDM concentrations from 0.01% to 100% were used and the stimulating action in terms of mitotic index (MI) was evaluated. A clear mitogenic e€ect was observed only with concentrations higher than 10% with peak e€ects around 60%. Further increase in CDM concentrations resulted in an MI decrease, and at 100% CDM the e€ect was totally abolished. Tests addressed to identify the presence of hepatocyte growth factor (HGF) yielded negative results. In order to identify the mitogenic factor(s) involved, we tested CDM obtained after lipopolysaccharide (LPS) stimulation of N-11 cells. Comparison of the results obtained with untreated or LPS stimulated CDMs suggested that macrophage activation does not a€ect the release of hepatomitogenic activity. To further characterize this macrophage-derived activity, we checked whether CDM could interact with the mitogenic e€ects of epidermal growth factor (EGF). CDM (10 or 50%) showed no stimulatory e€ect to hepatocytes cultured in the presence of a maximally stimulatory concentration of EGF. Conversely, both CDM concentrations were able to increase the MI of hepatocyte cultures treated with a suboptimal dose of EGF. These results suggest that macrophages release factor(s) which interact, in hepatocytes, with the EGF signal transduction mechanisms, or with the EGF receptor itself. # 1999 Elsevier Science Ltd. All rights reserved Keywords: growth factor; immune activation; proliferation; rat hepatocytes. Abbreviations: BSA = bovine serum albumin; CDM = conditioned medium; EGF = epidermal growth factor; FCS = foetal calf serum; HB-EGF = heparin-binding EGF-like growth factor; HGF = hepatocyte growth factor; IL-1 = interleukin 1; IL-6 = interleukin 6; LPS = lipopolysaccharide; MI = mitotic index; PKC = protein kinase C; TGF-b = tumour growth factor-b; TNF-a = tumour necrosis factor-a; WME = Williams' E medium.

INTRODUCTION

Regeneration of the liver, after a loss of liver mass by either surgical or chemicals means, is an example of one of the more dramatic processes of controlled cell growth. Hepatocyte primary cultures have been used in attempts to identify the growth stimulator for the regeneration process because, on the basis of studies with the quiescent, di€erentiated hepatocytes in serum-free cell culture systems, which retain many of the functions of hepatocytes in vivo and respond to various hormones, interference from secondary e€ects could be excluded. By these means, *Corresponding author.

signi®cant advances in the identi®cation of hepatocyte mitogens have been made during the past 20 years, but as yet the speci®c mechanism(s) that translate(s) the loss of functioning hepatocytes into a proliferative stimulus remains to be determined. Primary culture adult rat hepatocytes can be induced into DNA synthesis and cell division by insulin and EGF in a serum-free medium at low cell density, that is, one-third or one-quarter of that at cell con¯uency and collagen substrata (Kost et al., 1991; Mc Gowan et al., 1981). Recent in vivo and in vitro studies have shown that a high number of factors are involved in hepatocyte growth control, including complete mitogens [epidermol growth factor (EGF), tumour growth factor-b (TGF-b), hep-

0887-2333/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved. Printed in Great Britain PII: S0887-2333(99)00003-X

476

E. Piatti et al.

atocyte growth factor (HGF), tumour necrosis factor-a (TNF-a), heparin-binding EGF-like growth factor (HB-EGF)], comitogens (norepinephrine, insulin) and growth inhibitors [tumour growth factor-b (TGF-b), interleukin 1 (IL-1), interleukin 6 (IL-6)] (Koch et al., 1990; Michalopoulos et al., 1990). Moreover, there are indications of metabolic cooperation between Kup€er cells and hepatocytes which may be important for the regulation of normal and pathologic liver function. Kuppfer cells are known to secrete IL-1 and other cytokines such as TGF-b and growth factors such as HGF (Lindroos et al., 1992). However, the e€ect of the Kup€er cells in the regulation of hepatocyte proliferation in vitro is not clear. Whereas previous reports have indicated a reduction of hepatocyte proliferation (Nakamura et al., 1988), more recent studies have shown the presence of a heat-labile factor in conditioned medium from primary cultures of Kup€er cells which stimulates DNA synthesis. According to Meyer (1991), Kup€er cells might therefore regulate liver regeneration in both directions. This possibility of control could prove also to be very important with respect to the immune reactivity of Kup€er cells, suggesting that immune activation might induce a change in the pattern of hepatostimulating factors released by these cells. Unfortunately, the study of this issue is hampered by the diculty of preparing large numbers of pure Kup€er cell cultures in a reproducible way. Thus, the use of immortalized cell lines was proposed as approach to the unravelling of this problem. In the absence of rat Kup€er cell lines the closest available was from mouse macrophages immortalized by activated myc oncogenes. These cells have been extensively characterized to show that the alterations introduced in the process of cell immortalization do not a€ect the general reactivity of the cell. In particular, they show the expected phenotype and express macrophage functions (Righi et al., 1989). In addition, these cells maintain their immunoresponsiveness since they can be induced to secrete cytokines and cytotoxic intermediates after stimulation with immunomodulators (Pirami et al., 1991). In this paper we have used the microglial cell line N-11, derived from mouse brain macrophages, to examine the role of macrophages on hepatocyte proliferation in vitro and to determine whether activation might in¯uence this activity.

MATERIALS AND METHODS

Isolation of rat liver parenchymal cells Parenchymal hepatocytes were isolated from Sprague±Dawley male albino rats (200±250 g), using the two-step collagenase perfusion technique according to Seglen (1976) with minor modi®cations. Preparations were used that exhibited

greater than 80% viability, measured by trypan blue exclusion. Cell culture Rat hepatocytes were plated on collagen-coated slides in 60 mm diameter dishes at subcon¯uent density (35000 cells/cm2) and cultured in Williams E medium (WEM) supplemented with 0.2% bovine serum albumin (BSA), 6  10ÿ6 mM insulin, 10ÿ7 M dexamethasone and 50 mg gentamicin/ml (Sigma). After 2 hr incubation at 378C in 5% of a CO2 atmosphere and in the presence of 5% foetal calf serum (Gibco), the medium was replaced in order to remove dead cells. After a period of incubation of 24 hr to permit functional recovery of isolated hepatocytes, the cells were treated for 48 hr with EGF or conditional medium (CDM). The N-11 cell line, derived from primary mouse microglial cells (Righi M., 1989) was cultured in RPMI 1640 supplemented with 10% foetal calf serum (FCS), penicillin (100 IU/ml) and streptomycin (100 mg/ml), by seeding 105 cells in 60 mm diameter petri dishes; medium changes being made every 2 days. After reaching con¯uency at about 4 days after seeding, cells were subcultured by trypsinization. Preparation of conditioned medium In order to obtain the CDM, N-11 cells at about 70% con¯uency were used on the second day after seeding. As WEM causes a reduced growth rate for these cells, the usual N-11 maintaining medium was used for this purpose, then conditioned RPMI was mixed with WEM in di€erent proportions in the hepatocyte cultures. Conditioned medium was obtained 12 hr after the addition of fresh RPMI containing a reduced concentration of FCS (2%) to N-11 cells in the logarithmic growth phase. In order to assess the e€ect of the treatment of macrophage cells with LPS from Escherichia coli 026:B6 (Sigma), the substance was added to the culture medium at a ®nal concentration of 1 mg/ml and left for about 12 hr. The medium was centrifuged at high speed (100,000 g for 30 min) in order to eliminate cellular debris and viruses. The supernatant was then sterilized by ultra®ltration and stored in aliquots at ÿ808C. Treatments The stimulating action of EGF was evaluated at doses from 10 to 50 ng/ml in complete culture medium WME (containing 0.2% BSA, 6 nM insulin, 100 nM dexamethasone, 50 mg/ml gentamicin) by an increase in the mitotic index (MI). We tested the stimulating action of the N-11 mouse macrophage cell line conditioned medium at di€erent concentrations with complete WME medium ranging from 0.01% to 100% of CDM. Cultures were either incubated with CDM obtained from N-11 cells, CDM obtained from N-11 cells treated with LPS for 12 hr

Stimulation of rat hepatocytes by macrophage conditioned medium

Fig. 1. (a) E€ect of CDM treatment on the proliferation of primary culture rat hepatocytes seeded at the density of 3.5  104 cells/cm2 measured as cumulative percentage of cells blocked in metaphase. CDM was obtained from cultured N-11 cells; 0W represents the mitotic index of parenchymal cells cultured in WME medium; 0R represents the mitotic index of parenchymal cells cultured in RPMI medium. Values shown are means 2SD from four separate experiments. The di€erence between control, 0.78 20.16 and 0.8620.02 using WME(0W) or RPMI (0R), respectively, and treatments are statistically signi®cant as evaluated by contingency tables (P < 0.01). (b). E€ect of EGF on the mitotic index increase in rat hepatocytes; values are mean2 SD from three separate experiments. The di€erence between control and treatments are statistically signi®cant as evaluated by contingency tables (P < 0.01).

477

478

E. Piatti et al.

or with CDM plus EGF. The treatments were started 20 hr after cell attachment to allow a functional recovery of the cells; after 48 hr of treatment the culture was treated with colcemid at a concentration of 0.4 mg/ml for at least 3 hr to arrest the cells in C-metaphase. Evaluation of mitotic index and statistical analysis At the end of the culture treatment, cells were treated with a hypotonic solution (0.56% KC1) for 7 min, ®xed with methanol±acetic acid (3:1) and stained with 3% Giemsa solution. Afterwards, for the evaluation of mitotic index (metaphase number/ 1000 total cells) the slides were collected for microscopic examination and coded for analysis to avoid operator bias. Experimental groups were compared using contingency tables; signi®cant di€erences were set at P < 0.05.

RESULTS

E€ect of N-11 conditioned medium on the proliferation of quiescent hepatocytes The induction of hepatocyte proliferation due to addition of CDM derived from overnight cultures

of N-11 cells was evaluated by light microscope analysis of MI in rat hepatocyte primary cultures. Figure 1, panel a, shows the metaphase frequency in cultured rat hepatocytes when treated for 48 hr in the presence of di€erent CDM amounts. The addition of CDM was able to cause a dose-dependent hepatocyte proliferation increase with maximal e€ect observed at 60% CDM. Concentrations of 5%, 10%, 50% and 80% CDM produced a proliferation increase as assessed statistically by the chi-square test (P < 0.01). Conversely, the proliferative e€ect was abolished at very high concentrations and for cells grown in CDM (100%) the values were similar to those observed with the negative controls 0W and 0R). The pattern of activity of CDM was similar to that observed with EGF, a known hepatomitogenic factor (Fig. 1, panel b). In both tests the increase of MI in respect to basal values was comparable, being about sixfold for CDM (60%) and 8.2-fold for EGF (40 ng/ml). Analysis of very low CDM concentrations (from 0.01 to 1%) con®rmed the dose-dependent e€ect, showing a signi®cant di€erence with respect to the controls only with 1% CDM (data not shown). To clarify whether the hepatomitogenic e€ect might be due to cyotokines or indeed factors induced by immune activation, CDM derived from

Fig. 2. E€ect on hepatocyte proliferation when CDM is obtained from N-11 cells grown in the absence or presence of LPS (1 mg/ml); 0W represents the mitotic index of parenchymal cells cultured in WME medium; 0R represents the mitotic index of parenchymal cells cultured in RPMI medium. Control values without LPS (hatched bars) are 0.85 20.41 and 0.862 0.02 using WME (0W) or RPMI (0R), respectively; control values with LPS (stippled bars) are 0.932 0.52 and 0.87 20.19, using WME (0W) or RPMI (0R) respectively. Values shown are means2SD from four separate experiments. The di€erence between control values are statistically signi®cant as evaluated by contingency tables (P < 0.01).

Stimulation of rat hepatocytes by macrophage conditioned medium

LPS-treated cultures was tested. Because we were unable to prevent the carry-over of LPS from microglial cultures, LPS was added to obtain a ®nal concentration of 1 mg/ml, both in control or stimulated hepatocyte cultures. Figure 2 reports the comparison between the stimulating action on rat hepatocytes of CDM obtained from cultures treated with LPS for 12 hours. As it can be seen, use of CDMs from LPS-treated or untreated cultures did not cause marked di€erences in MI apart from a slight decrease at the dose of 50%. E€ect of N-11 conditioned medium on the proliferation of EGF-stimulated hepatocytes In order to further characterize the action of the CDM, its e€ects were tested on the hepatomitogenic activity of EGF. When hepatocytes were stimulated with EGF at 40 ng/ml, that is, the dose that stimulates the maximal response in our culture conditions, no further increases were seen by the addition of either dose of CDM, 10% or 50%. Conversely, when EGF was added at the dose that gave about one-half maximal response (10 ng/ml), an increase in the MI was observed with the addition of either 10% or 50% CDM. The di€erence between control and treated samples was signi®cant

479

(chi-square test) both in the case of treatment with CDM alone or CDM plus 10 ng EGF/ml (Fig. 3). DISCUSSION

The aim of this study was to gather further information about the ability of adult rat hepatocytes in primary culture to proliferate by studying the mitogenic e€ect of a conditioned medium obtained from macrophage derived cells, the N-11 cell line, since it is well known that macrophages play a key role in liver regeneration in vivo (Katsumoto et al., 1989). In fact, Kup€er cells have been implicated in the regulation of hepatocyte proliferation, but contradictory results on a stimulatory as well as inhibitory action have been reported (Michalopoulos, 1990; West et al., 1989). Studies on hepatocytes in vitro have also added weight to this hypothesis; some authors identi®ed in Kup€er cell media some of the most important mediators in¯uencing hepatocyte proliferation. However, the results are not unequivocal. According to Meyer et al. (1991), the presence of TGFb in the conditioned medium obtained from isolated and cultured Kup€er cells is responsible for the growth inhibitory action of EGF stimulated rat hepatocytes isolated from the same animal. On the other hand, Katsumoto (1989)

Fig. 3. E€ect on hepatocyte proliferation when CDM is co-administered with EGF; values are mean 2SD from three separate experiments. *, ** = signi®cantly di€erent from control (P < 0.05; P < 0.01, respectively) using the chi-square test.

480

E. Piatti et al.

reported that the conditioned medium derived from Kup€er cells was able to stimulate hepatocyte proliferation. In this paper, the hepatomitogenic e€ect of medium conditioned from an immortalized mouse macrophage cell line (N-11) is described. The advantage of these macrophage-derived cells is that, unlike Kup€er cells, they are able to replicate inde®nitely in culture without alteration of their immune characteristics and speci®cities (Righi, 1989). The stimulation of cell proliferation at di€erent concentrations of CDM obtained from N-11 cells was therefore studied in hepatocyte cultures. The CDM appeared to contain growth stimulatory as well as inhibitory activity, albeit at di€ering concentrations. This was inferred by the bell shape of the curve obtained using di€erent concentrations of CDM. This result has to be attributed to an inhibitory e€ect due to the presence of increasing amounts of the conditioned medium, since hepatocyte cultures grown in RPMI supplemented with BSA and hormones do not show a reduction in MI when compared with hepatocytes grown in WME. Some authors (Nakamura et al., 1988; Satoh and Yamazaki, 1992) have reported an inhibitory action by IL-1 on cultured cells. The in vivo situation is quite di€erent, because Koga and Ogasawara (1991), demonstrated an increase in MI after administration of IL-1 to rats; however, it is well known that other factors with an endocrine e€ect are able to play a role in the modulation of liver proliferation during regeneration. In fact, hepatocyte proliferation might be regulated by factors produced by hepatocytes that act through an autocrine mechanism (like TGF-a); whereas HGF, TGF-b and TNF-a are made in non-parenchymal cells and act by paracrine e€ect (Steer, 1995). HGF and EGF can also act by both endocrine and autocrine mechanisms (Labreque, 1994). In addition, the e€ect of a given growth factor can be modi®ed by the co-administration of other growth factors. This would explain the di€erences observed between in vitro and in vivo experiments. For example, the ability of TGF-b to inhibit or promote cell proliferation appears to be dependent on the target cell types, cell density used as well as the concomitant presence of other growth factors (Nakamura et al., 1985, Petersen et al., 1994). When TGFb and EGF were added simultaneously to G0 arrested cells, in murine ®broblast, induction of DNA synthesis was enhanced compared with that of EGF alone. Conversely, when TGF-b was added either 2 hr before or after EGF stimulation, it actually inhibited induction of DNA synthesis (Fukami et al., 1995). In order to clarify whether immune activation might contribute to the release of growth factor(s), we evaluated the hepatomitogenic activity of CDM from immune activated N-11 cells. If a slight and non-signi®cant decrease in some samples is disre-

garded, the hepatomitogenic activity of CDM was not a€ected by stimulation of N-11 cells with lipopolysaccharide (LPS), a powerful macrophage activator. LPS usually induces N-11 cells to secrete a number of reactive intermediates or in¯ammatory cytokines (IL-1a, TNF-a), but apparently this did not increase or impair the hepatomitogenic activity of CDM. The absence of modi®cations might thus indicate that release of the mitogen(s) is constitutive and independent of immune activation. This suggests that cytokines may not play a major role in the hepatomitogenic activity observed in CDM (Fausto et al., 1995). In this respect it is important to emphasize the time limit of our system. Actually, Beyer and Theologides (1993) have shown that the hepatomitogenic action of TNF-a appears only after 3 days of treatment, and this period is not reached in our experimental protocol. Apart from cytokines, other factors are important for hepatocyte proliferation. Among them the hepatocyte growth factor (HGF), also known as scatter factor (Zarnegar and Michaelopoulos, 1989). As HGF is known to be produced by macrophages (Lindroos et al., 1991; Noij et al., 1990), its possible presence in N-11 cultures appeared to be free of scatter factor activity when tested on a suitable indicator cell line (MDCK). Similarly, HGF transcripts in N-11 cells on hybrydization with HGF speci®c probe were not observed (data not shown). As a further approach to identify the factor involved, the possibility that CDM might synergize with EGF was investigated. At the dose of EGF that caused a maximal proliferation, the simultaneous presence of CDM did not gave signi®cant changes to EGF-induced proliferation. On the other hand, when EGF was used at a lower concentration (10 ng/ml), that producing half-maximal e€ect, the presence of CDM gave rise to a further increase in the extent of the proliferation, leading to the notion that both stimuli may involve activation of the same receptor or other common signal transduction mechanisms. Macrophages are not known to synthetize EGF, but in lung pathologic processes at the sites of active ®brosis (Leslie et al., 1997). Generally they do give rise to other members of the same family, such as HB-EGF and TGF-a, which may interact with the EGF receptor (Higashayama et al., 1991). The HB-EGF factor appears to play an important role in hepatocyte proliferation although its e€ects suggest a general role in wound healing (Michalopoulos and DeFrances, 1997) more than a speci®c e€ect on liver regeneration. We were not able to recognize the presence of HB-EGF in CDM, but it is extremely interesting to note that N-11 cells seemed to display a constitutive PKClike activity in comparison to another macrophage cell line (N-13) immortalized by the same oncogene. As expression and release of HB-EGF appears to be enhanced by activation of protein kinase

Stimulation of rat hepatocytes by macrophage conditioned medium

C(PKC), it will now be interesting to investigate whether other cell lines display, in comparison to N-11, a similar or a decreased hepatomitogenic activity. This hypothesis is supported by the ®nding of Briers et al. (1994) that a similarly immortalized murine macrophage cell line (MMGT1) is able to secrete EGF or EGF-like proteins after activation of PKC by LPS treatment. A future step in our work will be to identify and quantify the factor(s) responsible for the proliferation induction. This experimental approach would contribute to a better understanding of the complex interactions between macrophages and hepatocytes on liver regeneration. In fact, the isolated parenchymal cell culture system has been extremely valuable in identifying an increasing number of stimulatory and inhibitory substances as well as the initial steps in their mechanism of action, thus providing an extremely attractive model for studying of cellular growth control. AcknowledgementsÐWe are grateful to Professor Comoglio, who kindly supervised the analyses performed in the Torino University laboratories to check for the presence of HGF in the conditioned medium obtained from N-11 cells. REFERENCES

Beyer H. S. and Theologides A. (1993) Tumor necrosis factor is a direct hepatocyte mitogen in the rat. Biochemistry and Molecular Biology International 29, 1± 4. Briers T. W., Desmaretz C. and Vanmechelen E. (1994) Generation and characterization of mouse microglial cell lines. Journal of Neuroimmunology 52, 153±164. Fausto N., Laird A. D. and Webber E. M. (1995) Liver regeneration. 2. Role of growth factors and cytokines in hepatic regeneration. FASEB Journal 9, 1527±1536. Fukami J., Tsuji K., Ueno A. and Ide T. (1995) Transforming growth factor-b 1 has both promoting and inhibiting e€ects on induction of DNA synthesis in human ®broblasts. Experimental Cell Research 216, 107±112. Higashiyama S., Abraham J. A., Miller J., Fiddes J. C. and Klagsbrun M. (1991) A heparin-binding growth factor secreted by macrophage-like cells that is related to EGF. Science 251, 936±939. Katsumoto F., Miyazaki K. and Nakayama F. (1989) Stimulation of DNA synthesis in heaptocytes by Kup€er cells after partial hepatectomy. Hepatology 9, 405±410. Koch K. S., Lu X. P., Brenner D. A., Fey G. H., Martinez-Conde A. and Le€ert H. L. (1990) Mitogens and hepatocyte growth control in vivo and in vitro. In Vitro Cellular and Developmental Biology 26, 1011±1023. Koga M. and Ogasawara H. (1991) Induction of hepatocyte mitosis in intact adult rat by interleukin-1 and interleukin-6. Life Sciences 49, 1263±1270. Kost D. P. and Michalopoulos G. K. (1991) E€ect of 2% dimethyl sulfoxide on the mitogenic properties of epidermal growth factor in primary hepatocyte culture. Journal of Cellular Physiology 147, 274±280. LaBrecque D. (1994) Liver regeneration: a picture emerges from the puzzle. American Journal of Gastroenterology 89, S86±S96.

481

Leslie C. C., McCormick-Shannon K., Shannon J. M., Garrick B., Damm D., Abraham J. A. and Mason R. J. (1997) Heparin-binding EGF-like growth factor is a mitogen for rat alveolar type II cells. American Journal of Respiratory Cell and Molecular Biology 16, 379±387. Lindroos P., Tsai W. H., Zarnegar R. and Michalopoulos G. K. (1992) Plasma levels of HGF in rats trated with tumor promoters. Carcinogenesis 13, 139±141. McGowan J. A., Strain A. J. and Bucher N. L. (1981) DNA synthesis in primary cultures of adult rat hepatocytes in a de®ned medium: e€ects of epidermal growth factor, insulin, glucagon and cyclic-AMP. Journal of Cellular Physiology 108, 353±363. Meyer D. H., Bachem M. G. and Gressner A. M. (1991) Bidirectional e€ects of Kup€er cells on hepatocyte proliferation in vitro. FEBS Letters 283, 150±154. Michalopoulos G. K. (1990) Liver regeneration: molecular mechanisms of growth control. FASEB Journal 4, 176± 187. Michalopoulos G. K. and DeFrances M. C. (1997) Liver regeneration. Science 276, 60±66. Nakamura T., Arakaki R. and Ichihara A. (1988) Interleukin-1 b is a potent growth inhibitor of adult rat hepatocytes in primary culture. Experimental Cell Research 179, 488±497. Nakamura T., Tomita Y., Hirai R., Yamaoka K., Kaji K. and Ichihara A. (1988) Inhibitory e€ect of transforming growth factor-b on DNA synthesis of adult rat hepatocytes in primary culture. Biochemical and Biophysical Research Communications 133, 1042±1050. Noij S., Tashiro K., Koyama E., Nohno T., Ohyama K., Taniguchi S. and Nakamura T. (1990) Expression of hepatocyte growth factor gene in endothelial and Kup€er cells of damaged rat livers, as revealed by in situ hybridization. Biochemical and Biophysical Research Communications 173, 42±47. Petersen B., Yee C. J., Bowen W., Zarnegar R. and Michalopoulos G. K. (1994) Distinct morphological and mito-inhibitory e€ects induced by TGF-b 1, HGF and EGF on mouse, rat and human hepatocytes. Cell Biology and Toxicology 10, 219±230. Pirami L., Stockinger B., Corradin S. B., Sironi M., Sassano M., Valsasnini P., Righi M. and RicciardiCastagnoli P. (1991) Mouse macrophage clones immortalized by retroviruses are functionally heterogeneous. Proceedings of the National Academy of Sciences of the U.S.A. 88, 7543±7547. Righi M., Mori L., De Libero G., Sironi M., Biondi A., Mantovani A., Donini S. D. and Ricciardi-Castagnoli P. (1989) Monokine production by microglial cell clones. European Journal of Immunology 19, 1443±1448. Satoh M. and Yamazaki M. (1992) Tumor necrosis factor stimulates DNA synthesis of mouse hepatocytes in primary culture and is suppressed by transforming growth factor b and interleukin 6. Journal of Cellular Physiology 150, 134±139. Seglen P. O. (1976) Preparation of isolated rat liver cells. Methods in Cell Biology 13, 29±83. Steer C. J. (1995) Liver regeneration. FASEB Journal 9, 1396±1400. West M. A., Billiar T. R., Curran R. D., Hyland B. J. and Simmons R. L. (1989) Evidence that rat Kup€er cells stimulate and inhibit hepatocyte protein synthesis in vitro by di€erent mechanisms. Gastroenterology 96, 1572±1582. Zarnegar R. and Michalopoulos G. (1989) Puri®cation and biological characterization of human hepatopoietin A, a polypeptide growth factor for hepatocytes. Cancer Research 49, 3314±3320.