EGF- and CPA-induced mitogenic stimuli are differentially down-regulated by TGF-beta1 in cultured rat hepatocytes

EGF- and CPA-induced mitogenic stimuli are differentially

down-regulated by TGF-

β1 in cultured rat hepatocytes

Reto Fasciati

1

_{and Peter Maier}

2

_{be the crucial step in clonal selection and promotion of}

preneoplastic cells towards cancer. More likely, endogenous

Institute of Toxicology, Swiss Federal Institute of Technology and

and exogenous mitogens affect different cell populations

University of Zu¨rich, CH-8603 Schwerzenbach, Switzerland

within a tissue, populations which differ in their response to

1_{Present address: Amplimmun AG, Via Gravulaina, Ch-7523 Madulain,}

negative growth stimuli or in their vulnerability to acquiring

Switzerland

mutations. In rodents, the liver was found to be particularly

2_{To whom correspondence should be addressed}

sensitive towards non-genotoxic carcinogens (5). Therefore,

Down-regulation of the mitogenic activity of the rodent liver

_{we addressed the question whether cultured rat hepatocytes,}

carcinogen cyproterone acetate (CPA) and of epidermal

_{stimulated by an exogenous or an endogenous mitogen, respond}

growth factor (EGF) were compared in cultured rat

_{differentially to negative growth control initiated by}

trans-hepatocytes. Both hepatomitogens produce an increase in

_{forming growth factor}

_β

_{1 (TGF-}

_β

_{1*). This was carried out}

the expression of proliferating cell nuclear antigen (PCNA)

_{with two rodent liver mitogens, the carcinogenic and synthetic}

and in [

3

_{H]thymidine incorporation in a dose-dependent}

steroid cyproterone acetate (CPA) and the endogenous peptide

manner. In combination, the two mitogens induced an

_{epidermal growth factor (EGF).}

additive mitogenic response. Concomitant exposure to the

_{CPA is found in oral anti-acnegenic contraceptives and is}

growth inhibitory cytokine transforming growth factor

β1

_{used to treat prostate carcinoma, male hypersexualism and}

(TGF-

β1) resulted in a differential dose-dependent down-

_{signs of androgenization in women. In a rat liver foci bioassay,}

regulation of PCNA-expressing cells. The correspond-

_{CPA induced altered foci in female rats (6) but not in male}

ing down-regulation of CPA-induced PCNA expression

_{rats (7). At lower doses CPA increased benign liver proliferation}

required a 3- to 5-fold higher TGF-

β1 concentration than

_{(8) and at higher levels hepatic tumours in rats and mice, both}

for EGF-induced expression. In contrast, CPA-exposed

_{males and females (9). The mitogenic activity of CPA in rat}

hepatocytes become vulnerable to and EGF-exposed cells

_{liver was demonstrated in vivo (10) and in vitro (11), whereas}

protected against the apoptosis-inducing activity of TGF-

_{in human hepatocytes the mitotic activity was found only in}

β1 (.0.1 ng/ml). Under culture conditions that mimicked

_{the presence of growth factors (12). The synthetic steroid}

a pericentral-equivalent microenvironment (low oxygen

_{exerted no mutagenic activity in tests with bacteria and}

tension, low glucagon concentration), PCNA expression

_{mammalian cells (13,14), but recent findings demonstrated an}

was 3-fold lower and CPA-specific resistance was no longer

_{induction of DNA repair synthesis (15,16). Reports about its}

detectable. It is concluded that EGF and CPA induce their

clastogenic activity in hepatocytes were inconclusive (17).

growth stimuli preferentially in the periportal area of the

_{Highly persistent DNA adducts of the metabolite 3}

_α

-hydroxy-liver but in different hepatocyte sub-populations, which

cyproterone acetate were found (18,19), particularly in female

differ in their down-regulation of premitotic events by

rat liver (20). Together, the data imply that CPA is a complete

TGF-

β1. At low TGF-β1 concentrations, EGF-stimulated

carcinogen in female rodent liver at high doses, whereas in

cells shift back into a resting cell cycle phase, whereas

male liver it is predominantly the mitogenic stimulus that

CPA-treated hepatocytes are eliminated by apoptosis at

contributes to tumourigenesis.

higher TGF-

β1 concentrations.

Negative growth control in hepatocytes can be initiated

by low levels of TGF-

β

1 (21). This cytokine is a potent

inhibitor of cell growth in parenchymal liver cells as well as

Introduction

in rat and human hepatocytes (22,23). In addition, TGF-

β

1 is

Cell proliferation is an important promoting factor in tumour

one of the most effective endogenous growth inhibitors after

induction and progression in rodents (1,2). Chemically induced

_{partial hepatectomy (24). It is expressed in most liver cells,}

cell cycle progression can be stimulated directly by interactions

_{including hepatocytes (25), and acts via paracrine and autocrine}

with growth stimulating factors (3) or indirectly by regenerative

_mechanisms.

growth as a consequence of chemically induced cell necrosis

_{In order to obtain in vitro information which adequately}

(4).

_{represents the situation in liver, cultured hepatocytes should}

Endogenously controlled high rates of cell proliferation are

_{be growth arrested and should preserve or re-acquire}

tissue-well tolerated in many tissues, whereas exogenous chemicals

_{equivalent metabolic competence. This was achieved by the}

that induce a mitogenic stimuli often act as non-genotoxic

_{use of a tissue-specific extracellular matrix (26,27), high cell}

carcinogens. This suggests that mitogenesis per se cannot

_{densities and a serum-free culture medium supplemented with}

low insulin concentrations and a protease inhibitor (28). Under

*Abbreviations: TGF-β1, transforming growth factor-β1; CPA, cyproterone

these conditions, within 2 days in culture hepatocytes express

acetate; EGF, epidermal growth factor; PCNA, proliferating cell nuclear

_{their liver-specific metabolic competence (29), they respond}

antigen; CMF, crude membrane fraction; PBS, phosphate-buffered saline;

_{to tissue-equivalent oxygen tensions (29), to glucoregulatory}

DMSO, dimethylsulphoxide; BSA, bovine serum albumin; TdR, thymidine;

hormones (29) and to the apoptotic stimuli of cytokines (30).

LDH, lactate dehydrogenase; PC cultures, pericentral-equivalent cultures; PP

scraped off in 5% trichloracetic acid and transferred to glass fibre filters in a

mitogenic stimuli of EGF and CPA are differentially

down-cell harvester. Non-precipitable DNA was removed by washing with distilled

regulated by TGF-

β

1. Three possible differences were

water. The3_{H radioactivity on each filter was determined in a scintillation}

addressed: (i) a different dose response towards TGF-

β

1;

_{counter (Irga-Save scintillation cocktail; Beckmann LS-6000LL).}

(ii) a difference in down-regulation by apoptosis; (iii) the

_{Measurement of lactate dehydrogenase (LDH) release and total protein content}

sensitivity of this down-regulation to factors characteristic of

LDH release was analysed spectrophotometrically (COBAS Fara

auto-different regions of the liver lobules, namely oxygen tension

_{analyser) using a commercially available kit (Boehringer-Mannheim,}

and ratio of the glucoregulatory hormones (insulin/glucagon).

Mannheim, Germany). Protein content was determined by the method of

Lowry et al. (34) using BSA as standard.

Apoptotic events

Materials and methods

Apoptotic cells (condensed and fragmented nuclei) were identified 48 h after

Chemicals and biologicals _{the beginning of treatment in Hoechst dye 339342 stained hepatocytes as}

Chemicals and biochemicals were of reagent grade and purchased from reported earlier (30). Cells with condensed nuclei were rarely seen. A Sigma (St Louis, MO). Cell culture products were purchased from Gibco fragmented nucleus was defined as (i) one or more brightly fluorescent (Paisley, UK) and from Seromed (Berlin, Germany). Porcine collagen type I (condensed) chromatin pieces within an area equivalent to a nucleus surrounded was from Pentapharm (Basel, Switzerland). Human recombinant TGF-β1 was by cytoplasmic material (apoptotic bodies) or (ii) a cluster of chromatin pieces a generous gift of Dr K.Frei (Institute for Clinical Immunology, University within a cell without a nucleus.

Hospital of the University of Zurich, Switzerland). Human recombinant EGF

Data analysis

was purchased from Becton Dickinson (Bedford, MA).

The mean value and the standard error of the mean (SEM) were calculated Antibodies were purchased from the following sources: monoclonal mouse

for each parameter measured. Significance of differences between means was anti-proliferating cell nuclear antigen (PCNA) antibody (PC-10) from DAKO

calculated using Student’s t-test for paired or unpaired samples as appropriate. (Glostrup, Denmark); Cy3™-conjugated AffiniPure F(ab’)2 fragment goat

The level of significance chosen was P, 0.05. The description of the dose– anti-mouse IgG 1 IgM (H1L) from Milan Analytica AG (La Roche,

response curves and calculation of the IC5 and IC50values were performed Switzerland).

according to the model described by Bruinink (35).

Isolation, cultivation and treatment of hepatocytes

Hepatocytes were isolated from male Sprague–Dawley rats (random bred,

Results

240–270 g body wt, ZUR/SIV; Institute of Laboratory Animals, University

of Zurich, Switzerland) by two-step in situ collagenase perfusion as described

_{Constitutive PCNA expression and DNA synthesis in untreated}

previously (31). The isolated hepatocytes were washed and concentrated by

cells

three centrifugation steps (2 min at 20 g and 4°C) in serum-free William’s E

medium supplemented with 10 mM HEPES, 2 mM L-glutamine, 10 nM

_{Three hours after seeding, ~40 PCNA-positive nuclei/mm}

2

insulin, 1µg/ml aprotinin, 100 IU/ml penicillin/streptomycin and cultured in

_{were detectable. This corresponds to ~2% of the originally}

the same medium without HEPES but supplemented instead with 100 nM

seeded hepatocytes. This spontaneous PCNA expression

dexamethasone, 10 nM glucagon and 30 nM SeCl2. Hepatocytes were seeded

decreased over culture time down to

,0.1% after 96 h. In

on Costar 24-well plates (2 cm2/well) at a density of ~1.53105 cells/cm2

under 5% CO2/air. The influence of oxygen tension and glucoregulatory

contrast, [

3

H]TdR incorporation was not markedly changed

hormone ratio were investigated in gas-permeable Teflon membrane dishes

_{during the 4 day culture period; on day 2 a slight but}

(Petriperm, 20 cm2_{, 2310}5 _cells/cm2_{; Heraeus AG, Zurich, Switzerland)}

insignificant increase was observed. In cultures under low,

cultured under 13 or 4% O2. All culture dishes were precoated with a mixture

periportal-equivalent oxygen tension (4% O

2

) and low glucagon

of liver crude membrane fraction (CMF) and porcine collagen type I [6µg

rat liver CMF and 60 ng COL in 71µl/cm2phosphate-buffered saline (PBS),

concentrations (1 nM), this down-regulation of PCNA within

overnight precoating at 4°C] as described earlier (27,28,31).

_{48 h was effective in a similar way.}

In all the experiments the medium was changed 4 and 20 h after seeding and

CPA and EGF increase PCNA expression and DNA synthesis

then daily. After 2 days in culture, hepatocytes were exposed to the mitogens

and/or to TGF-β1 for 48 h. Stock solutions of CPA were freshly prepared in

_{CPA, up to a concentration of 6.25}

_µ

_{M, dose-dependently}

dimethylsulphoxide (DMSO) and diluted with culture medium (final DMSO

_{increases the number of PCNA-expressing nuclei within 48 h.}

concentration,0.1%). Human recombinant TGF-β1 (106_{U/mg; Genzyme,}

Above 12

µ

M CPA, the response plateaued; 12

µ

M CPA

Cambridge, MA) and the human recombinant EGF were freshly prepared in

increased the control value 3.5-fold and 25

µ

M CPA only

0.1% bovine serum albumin (Sigma; initial fractionation by cold alcohol

precipitation) in PBS and diluted up to 103-fold with culture medium.

4-fold. EGF at 1 ng/ml increased PCNA expression almost

PCNA analysis and measurement of Hoechst dye fluorescence

2.5-fold; at 10 ng/ml the increase was 7-fold (Figure 1A).

Hepatocytes were immunostained for PCNA, an auxiliary protein of DNA

[

3

_{H]TdR incorporation was induced dose-dependently up to}

polymerase δ, by a method modified from Kurki et al. (32). Briefly, the

_{a concentration of 12}

µ

_{M CPA. A doubling of this dose led}

hepatocytes were fixed and permeabilized as follows: 15 min at 4°C in 1%

_{to a decrease in DNA synthesis. EGF, given in concentrations}

formalin in PBS containing 20µg/mlα-lysophosphatidylcholine, then 15 min

of 10 ng/ml, increased [

3

_{H]TdR incorporation 3- to 4-fold}

in ice-cold methanol and subsequently 15 min in 0.1% NP-40, both at 4°C.

(Figure 3).

Non-specific binding sites were blocked in 1% BSA in PBS (30 min, room

temperature). Hepatocytes were then exposed to the anti-PCNA mAb (PC-10,

_TGF-

_β

_{1 down-regulates PCNA-expressing hepatocytes}

1:100, 1% BSA in PBS) or to mouse ascites fluid (1:100). PCNA-expressing

EGF-induced PCNA expression was down-regulated by

con-cells were stained with the second antibody (Cy3-conjugated, 1:250) and

hepatocyte nuclei with Hoechst 339342 (Calbiochem; 2 µg/ml in PBS).

_{centrations of 0.04 ng/ml TGF-}

β

_{1 and higher, whereas}

CPA-PCNA-positive nuclei were identified by fluorescence microscopy (Ex/Em

_{induced stimulation was unaffected up to 0.1 ng/ml TGF-}

_β

₁

340/460 nm) and the total DNA content by area measurements per well or

(Figure 1A). Between 0.1 and 1 ng/ml, TGF-

β

1 inhibited

dish (Cytofluor™ 2350 multiplate scanner, Millipore, Bedford, MA; Ex/Em

both CPA- and EGF-mediated mitogenic activity in a

dose-340/460 nm).

dependent manner. The IC

5

and IC

50

values of TGF-

β

1 were

Measurement of DNA synthesis

0.012 and 0.124 ng/ml for EGF- and 0.065 and 0.305 ng/ml for

[3H]Thymidine (TdR) incorporation of the hepatocytes was determined

CPA-stimulated hepatocytes respectively. Among hepatocytes

according to Vintermyr et al. (33) with minor modifications. After the 15 h

pulse (1µCi/well or plate [3_{H]TdR, 33–48 h after addition of the mitogens)}

_{exposed concomitantly to EGF and CPA, the mitogenic}

the cells were fixed in ethanol (30 min) and DNA content determined

_{response was additive or even slightly synergistic, whereas the}

(propidium iodide, Cytofluor™ 235, Ex/Em530/620 nm). Subsequently, the

_{sensitivity toward TGF-}

β

_{1 was intermediate between the IC}

5

monolayers were rinsed (ice-cold PBS) and fixed (5% aqueous trichloracetic

and IC

50

values of TGF-

β

1 for the two mitogens alone

acid, ice-cold 70% ethanol, 10 min each). RNA and bulk protein were

Fig. 3. Inhibition of EGF- (10 ng/ml) and CPA (12µM)-induced [3H]TdR

Fig. 1. Dose-dependent down-regulation of EGF- (10 ng/ml), CPA- (12µM)

incorporation (15 h pulse) by low and high levels of TGF-β1. CPA or EGF (A) and EGF1CPA-induced PCNA expression (B) by TGF-β1 (mean of

were administred together with TGF-β1 from days 2–4 in culture. Mean of three independent experiments6 SD). CPA or EGF were administred

three independent experiments6 SD. together with TGF-β1 from days 3–4 in culture. PCNA-positive nuclei were

counted with a microscope as described in Materials and methods.

Fig. 4. Dose-dependent induction of apoptosis (fragmented nuclei) by

TGF-β1 and stimulation or inhibition by concomitant exposure to CPA or EGF respectively (A). (B) Simultaneous exposure to CPA and EGF. Measurements were performed after 4 days in culture. Values represent the means6 SD of three independent experiments.

Fig. 2. Down-regulation of PCNA expression after exposure to low and

high concentrations of the hepatomitogens (mean of three independent experiments6 SD) by low (A) and high (B) levels of TGF-β1. Values

4), as measured by fluorescence microscopy of Hoechst dye

represent relative inhibition when compared with cultures without TGF-β1.

33

9342 stained nuclei (30). After exposure to 1 ng/ml TGF-

β

1,

the CPA-exposed hepatocytes showed a 3-fold higher incidence

When exposed to 0.1 ng/ml TGF-

β

1, the relative inhibition

of fragmented nuclei (Figure 4A) and the EGF-exposed cells a

of PCNA nuclei was the same in control cells as in those

50% lower incidence (Figure 4B) than cells treated with

TGF-treated with 1 or 10 ng/ml EGF, namely ~40% (Figure 2A).

β

1 alone. CPA and EGF alone had no effect on the spontaneous

In contrast, at this low TGF-

β

1 dose 3

µ

M and 12

µ

M

CPA-level of apoptotic cells. In cells exposed concomitantly to EGF

treated cells showed a further stimulation of PCNA nuclei

and CPA the higher sensitivity to TGF-

β

1-induced apoptotic

(20 and 10% respectively). At a higher dose of TGF-

β

1

processes was abolished (Figure 4B).

(0.5 ng/ml) all cells showed inhibition of PCNA expression,

LDH release after 48 h exposure to TGF-

β

1, in combination

but much more pronounced in EGF-treated, in EGF

1CPA-with aberrant nuclei, reflects the late phase of apoptotic cells

treated or in control cells (60–80% inhibition) than in

CPA-(30). Extracellular LDH activity was not significantly increased

exposed cells (35%) (Figure 2B).

after exposure to the mitogens (Figure 5). With increasing

TGF-[

3

_{H]TdR incorporation was not differentially affected: both}

β

1 concentration, in parallel with the incidence of fragmented

CPA- and EGF-induced [

3

_{H]TdR incorporation was reduced}

nuclei, extracellular LDH activity increased; it was higher in the

by one third at 0.1 ng/ml and returned to control levels at

CPA- (Figure 5A) and lower in the EGF-exposed hepatocytes

1 ng/ml TGF-

β

1. In untreated hepatocytes, after exposure to

(Figure 5B) than in the corresponding cultures without mitogen

0.1 ng/ml TGF-

β

1 [

3

_{H]TdR incorporation was doubled, but}

treatment.

the difference was not significant (Figure 3).

Differential inhibitory activity of TGF-β1 on PCNA expression

Differential susceptibility of CPA- and EGF-exposed cells

is abolished in a pericentral equivalent microenvironment

towards TGF-

β

1-induced apoptosis

The experimental conditions used in the experiments described

Exposure of hepatocytes to TGF-

β

1 concentrations

.0.1 ng/ml

approxi-of CPA-stimulated hepatocytes to down-regulation by low levels

of TGF-

β

1.

Discussion

Investigations on growth stimuli or impaired negative growth

control in cell cultures might be greatly affected by the

proliferative capacity and the metabolic state of the cells. In

the present study, at the beginning of the exposure to the

mitogens the hepatocytes expressed liver-specific functions

(28,29) and a low spontaneous mitogenic activity comparable

with that found in the intact organ.

Negative growth control and liver-specific metabolism

Fig. 5. Dose-dependent induction of LDH release by TGF-β1 and

stimulation (A) or inhibition (B) by concomitant exposure to the

_{It was shown that during liver perfusion and the following}

hepatomitogens. Values represent the means6 SD of three independent

tissue dissociation, growth stimuli similar to those observed

experiments.

after partial hepatectomy are induced in hepatocytes (37). This

entry into the cell cycle after liver perfusion might be the

reason for the initially high levels of PCNA expression in our

hepatocytes. However, only a small proportion of these

PCNA-positive cells enter S phase, as indicated by the only slightly

increased DNA synthesis rate within 48 h after seeding. Thus,

enzymic dissociation shifts hepatocytes from G

0

into middle

to late G

1

, but not into S phase. Accordingly, the majority of

PCNA-expressing hepatocytes in our cultures represent cells

arrested in mid to late G

1

phase (38).

The efficient spontaneous down-regulation of this

‘com-pensatory’ growth stimulus within 2 days in our cultures is

most likely the consequence of the high cell density

(down-regulation of immediate early genes; 39,40) and the culture

conditions chosen. Expression of liver-specific metabolism

seems to be linked to active negative growth control. This is

concluded from the fact that the loss of P450 mRNAs after

hepatocyte isolation parallels an increase in proliferative

activ-ity (37) and an increase in xenobiotic metabolism after 2–7

days in culture (41) parallels the decrease in PCNA expression.

Mitogenic effects of EGF and CPA

The present experiments demonstrate that EGF and CPA

shift quiescent hepatocytes with well-preserved liver-specific

functions (29,41) into G

1

and some into S phase. No serum,

and in the case of CPA no co-administration of EGF, was

required, in contrast to a previous report (42), in which PCNA

Fig. 6. Induction of PCNA expression in PC cultures (1 nM glucagon,

expression was induced by EGF but not by CPA. A

well-incubation atmosphere 4% O2) and down-regulation by TGF-β1. The

functioning liver metabolism might be of specific importance

reference culture (10 nM glucagon, incubation atmosphere 13% O2)

corresponds to 24-well cultures. Simultaneous exposure to CPA, EGF and

for CPA because this chemical affects microsomal

mono-TGF-β1 from days 3–4 in culture. Values represent the means6 SD of

_{oxygenase activity (43) and, prior to DNA binding, requires}

three independent experiments.

metabolic activation (20), most likely via a hydroxysteroid

sulfotransferase (44). Whether metabolic activation is also

required for the mitogenic activity of CPA remains to be

mate periportal conditions (36). Pericentral-equivalent cultures

(PC cultures) can be mimicked by culturing cells at low glucagon

elucidated.

Effective CPA concentrations in our in vitro

investiga-concentrations (1 nM) and low oxygen tension. This can be

achieved by the use of Teflon membrane culture dishes kept at

tions are in the range of those found in the serum of rats

treated with the synthetic hormone (45). It is noteworthy that

4% O

2

(28). As a corresponding reference, hepatocytes were

cultured in parallel in Teflon membrane culture dishes under

PCNA expression plateaus at CPA concentrations which induce

maximal DNA synthesis and remains at higher concentrations

13% O

2

in the same medium as used in the 24-well dishes (PP

cultures). The oxygen tension is then comparable to that under air

where DNA synthesis is inhibited. Clearly, only a limited

number of hepatocytes are able to respond to the mitogenic

in conventional culture dishes. Accordingly, differential

down-regulation was again detectable (Figure 6). However, two find-

stimulus of CPA. The additive, partially synergistic response

after combined treatment with CPA and EGF (Figure 1B)

ings differed in PC cultures (Figure 6): (i) EGF and CPA were

both much less effective (EGF 1/2 and CPA 1/4 of ‘periportal’

indicates that CPA stimulates additional cells that are not

responding to EGF. A similar additive response was reported

values); (ii) CPA-stimulated hepatocytes lost their resistance to

0.1 ng/ml TGF-

β

1. The metabolic status of the hepatocytes,

for the induction of DNA synthesis (12). The two mitogens

therefore act via different pathways and as a consequence

which is differentially affected by the two culture conditions

The differential mitoinhibitory activity of TGF-

β

1

of TGF-

β

1 and by that they become resistant to

apoptosis-inducing levels of TGF-

β

1. In contrast, CPA-stimulated

hepato-TGF-

β

1 is a key factor in pre-replicative negative growth

cytes cannot re-enter the resting cell cycle phase initiated by

control. After partial hepatectomy, TGF-

β

1 mRNA is increased

low levels of TGF-

β

1, but become a target for higher

TGF-in non-parenchymal liver cells before the major wave of

β

1 levels and escape via apoptosis. In favour of the second

DNA synthesis has been initiated (46). Accordingly, the

interpretation are the additive/synergistic induction of

PCNA-mitoinhibitory activity of TGF-

β

1 acts in cells at or prior to

expressing cells (Figure1B), the fraction of cells responding

the G

1

/S cell cycle phase transition (47).

to 0.1 ng/ml TGF-

β

1 after combined exposure to CPA and

Normal hepatocytes release TGF-

β

by an autocrine

EGF (Figure 2A) and, as discussed in the following section,

mechanism in its latent form (25) and are unable to activate

the influence of an altered metabolism as induced under PC

TGF-

β

complexes in vitro (24). This and the dose–response

culture conditions.

relationship seen in our study confirms that the added activated

Regiospecific response

recombinant TGF-

β

1 antagonizes endogenous and exogeneous

growth signals in rat hepatocytes. The TGF-

β

1 concentrations

Theoretically, liver lobules can be divided into a proliferative/

used were physiologically relevant. For the corresponding

_{mitogenic (periportal) and a functional/apoptotic (pericentral)}

down-regulation of CPA-exposed cells a 3- (IC

50

) to 5-fold

compartment (50–52). Hepatocyte cell populations within liver

(IC

5

) higher TGF-

β

1 concentration is required when compared

lobule regions might differ in their degree of differentiation

with EGF-stimulated cells (Figure 1). This differential down-

and their metabolic competence. Reduction of glucagon

con-regulation can be observed at low or high concentrations of

_{centration and of oxygen tension shift cellular metabolism}

the two mitogens (Figure 2A). This demonstrates that the

towards that described for the pericentral region (28). The

resistance is related to the type of mitogen and not to the

_{reduced activity of the two mitogens and the loss of resistance}

intensity of the mitogenic response. The absence of a differen-

to low levels of TGF-

β

1 in PC cultures (Figure 6) correspond

tial response in cells that enter DNA synthesis (Figure 3)

well with the reported increased sensitivity to TGF-

β

1 under

suggests that down-regulation of PCNA expression occurs

_{these culture conditions (30). It also corresponds well with}

prior to the G

1

/S phase transition.

in vivo studies in which the mitotic activity of EGF (53) and

of CPA (8,54) was mainly found in the periportal area,

Down-regulation of CPA-stimulated hepatocytes via an

including down-regulation by apoptosis after withdrawal of

apoptotic pathway

CPA (54). The present in vitro findings support data from an

It is well established that high levels of TGF-

β

1 induce

animal study in which apoptosis induced by CPA withdrawal

apoptosis in vivo (48) and in vitro (30,49). This was also

did not occur in DNA synthesizing cells (54).

found in our study at 48 h culture time at concentrations above

The loss of the differential response in PC cultures can be

0.1 ng/ml TGF-

β

1. The differential dose-dependent increase

_{attributed either to low glucagon concentrations or low oxygen}

in fragmented nuclei (Figure 4A) and LDH release (Figure 5A

tension. The interaction of glucagon with mitogenic processes

and B) indicate that CPA-treated hepatocytes are sensitized to

_{is controversial and seems to depend on the exposure levels}

and EGF-treated ones protected against the apoptotic activity

and cell density chosen. High levels of glucagon (10 nM), as

of TGF-

β

1. The differential down-regulation can also be seen

_{used in our 24-well or PP cultures, might even be mitoinhibitory}

at higher TGF-

β

1 levels (Figure 4B). In EGF-exposed cells

(55). No interaction was detectable between TGF-

β

1 and

the IC

50

(0.124 ng/ml) of TGF-

β

1 for PCNA down-

_{glucagon in rat hepatocyte cultures (47). However, glucagon}

regulation induced no apoptosis, whereas the IC

50

for CPA

_{concentrations chosen affect basal metabolism in the}

hepato-(0.305 ng/ml) significantly increased the number of

frag-cytes (28) and low oxygen tension, as reported earlier (56,57),

mented nuclei (Figure 4A). This fits well with an in vitro study

_{modulates chemically induced growth stimuli.}

in which CPA increased the apoptotic activity of TGF-

β

1 in

In conclusion, the experiments demonstrate differences

cultured rat hepatocytes (42). The protective effect of EGF is

_{between the down-regulation of an endogenous (EGF) and an}

stronger than the apoptotic stimulus of CPA, as indicated by

exogenous (CPA) mitogenic stimulus by TGF-

β

1. The exposure

the additive response after simultaneous exposure to CPA

_{levels of the growth inhibitory cytokine and the metabolic}

and EGF (Figure 4B). This indicates that the EGF-mediated

state of the cultured hepatocytes are crucial. It seems that

protection against apoptosis precedes sensitization and is also

endogenous and exogenous growth stimuli affect different

effective in CPA-stimulated hepatocytes.

_hepatocyte

_populations.

_{CPA-stimulated}

_cells

_are

_less

Taken together, the results suggest that EGF-stimulated

susceptible to down-regulation by low levels of TGF-

β

1

PCNA-expressing hepatocytes are down-regulated by a shift

_{than are EGF-exposed cells. Down-regulation in CPA-exposed}

from a prereplicative into a resting cell cycle phase, whereas

hepatocytes, in contrast to EGF-exposed hepatocytes, occurs

CPA-stimulated cells are more resistant to this pathway and

_{preferentially via apoptosis inducible at higher TGF-}

_β

_{1 levels.}

are eliminated by apoptosis. Two different modes of action

Therefore, at low (e.g. tissue equivalent) TGF-

β

1 levels,

CPA-are possible. Firstly, EGF-exposed hepatocytes CPA-are arrested in

stimulated hepatocytes have less chance of re-entering the

mid to late G

1

phase and thus are sensitive to the mitoinhibitory

_{resting cell cycle phase compared with EGF-stimulated cells}

activity of low levels of TGF-

β

1, whereas CPA-exposed

_{and have a higher risk of acquiring and fixing a spontaneous}

hepatocytes proceed further into the cell cycle (e.g. late G

1

_{premutagenic lesion into a mutation which could contribute to}

phase), at which point they lose their responsiveness to low

_{tumour formation.}

levels of TGF-

β

1 but become more vulnerable to the

apoptosis-inducing signal at higher concentrations. Alternatively, EGF

_{Acknowledgements}

and CPA induce PCNA expression in different, as yet

unidenti-This research project was funded by Interpharma, Basle, Switzerland. We

fied or uncharacterized, hepatocyte sub-populations. EGF-

_{thank Prof. Dr K.Frei of the Section of Clinical Immunology, University}

stimulated cells are able to switch from PCNA expression into

_{Hospital of Zu¨rich, Switzerland and Prof. Dr D.R.Dietrich, University of}

Konstanz, Germany, for helpful suggestions and discussions.

23. Petersen,B., Yee,C.J., Bowen,W., Zarnegar,R. and Michalopoulos,G.K.

References

(1994) Distinct morphological and mitoinhibitory effects induced by TGF-1. Pitot,H.C. and Sirica,A.E. (1980) The stages of initiation and promotion β1, HGF and EGF on mouse, rat and human hepatocytes. Cell Biol.

in hepatocarcinogenesis. Biochim. Biophys. Acta, 605, 191–205. Toxicol., 10, 219–230.

2. Schulte-Hermann,R., Ohde,G., Schuppler,J. and Trosiener-Timmerman,I. 24. Jakowlew,S.B., Mead,J.E., Danielpour,D., Wu,J., Roberts,A.B. and (1981) Enhanced proliferation of putative preneoplastic cells in rat liver Fausto,N. (1991) Transforming growth factor-β isoforms in rat liver following treatment with the tumor promoters phenobarbital, hexachloro- regeneration: messenger RNA expression and activation of latent TGFβ. cyclohexane, steroid compounds and nafenopin. Cancer Res., 41, 2556– _{Cell Regulat., 2, 535–548.}

2561. 25. Bissell,D.M., Wang,S.S., Jarnagin,W.R. and Roll,F.J. (1995) Cell-specific 3. Cunningham,M.L., Foley,J., Maronpot,R.R. and Matthews,H.B. (1991) _{expression of transforming growth factor βin rate liver. Evidence for} Correlation of hepatocellular proliferation with hepatocarcinogenicity _{autocrine regulation of hepatocyte proliferation. J. Clin. Invest., 96,} induced by the mutagenic noncarcinogen:carcinogen pair 2,6- and 2,4- _447–455.

diaminotoluene. Toxicol. Appl. Pharmacol., 107, 562–567. _{26. Saad,B., Schawalder,H.P. and Maier,P. (1993) Crude liver membrane} 4. Butterworth,B.E., Popp,J.A., Conolly,R.B. and Goldsworthy,T.L. (1992) _{fractions maintain liver specific functions in long term, serum free rat}

Chemically induced cell proliferation in carcinogenesis. IARC Sci. Publ., _{hepatocyte cultures. In Vitro Cell Dev. Biol., 29A, 32–40.}

116, 279–305. _{27. Saad,B., Scholl,F.A., Thomas,H., Schawalder,H-.P., Streit,V., Waechter,F.}

5. Ashby,J. and Tennant,R.W. (1985) Chemical structure, Salmonella _{and Maier,P. (1993) Crude liver membrane fractions and extracellular} mutagenicity and carcinogenicity as indicators of genotoxic carcinogenesis _{matrix components as substrate regulate differentially the preservation and} among 222 chemicals tested in rodents by the U.S. NCI/NTP. Mutat. Res., _{inducibility of cytochrome P-450 isoenzymes in cultured rat hepatocytes.}

204, 17–115. _{Eur. J. Biochem., 213, 805–814.}

6. Deml,E., Schwarz,L.R. and Oesterle,D. (1993) Initiation of enzyme altered

28. Ohno,K. and Maier,P. (1994) Cultured rat hepatocytes adapt their cellular foci by the synthetic steroid cyproterone acetate in rat liver foci bioassay.

glycolytic activity and adenylate energy status to tissue oxygen tension:

Carcinogenesis, 14, 1229–1231.

influences of extracellular matrix components, insulin and glucagon. J. 7. Schuppler,J., Damme,J. and Schulte-Herman,R. (1983) Assay of some

Cell. Physiol., 160, 358–366.

endogenous and synthetic sex steroids for tumor-initiating activity in rat

29. Maier,P., Saad,B. and Schawalder,H.P. (1994) Periportal- and centrilobular-liver using the Solt–Farber system. Carcinogenesis, 4, 239–241.

equivalent oxygen tension affect liver specific functions and xenobiotic 8. Schulte-Hermann,R., Hoffman,V., Parzefall,W., Kallenbach,M., Gerhardt,

metabolism in long-term rat hepatocyte cultures. Toxicol. In Vitro, 8, A. and Schuppler,J. (1980) Adaptive response of rat liver to the gestagen

423–435. and anti-androgen cyproterone acetate and other inducers. II. Induction of

30. Ohno,K., Amman,P., Fasciati,R. and Maier,P. (1995) Transforming growth growth. Chem.-Biol. Interactions, 31, 287–300.

factor β1 preferentially induces apoptotic cell death in rat hepatocytes 9. Schuppler,J. and Gu¨nzel,P. (1979) Liver tumors and steroid hormones in

cultured under pericentral-equivalent conditions. Toxicol. Appl.

rats and mice. Arch. Toxicol., 2, 181–195.

Pharmacol., 132, 277–236.

10. Schulte-Hermann,R., Ochs,H., Bursch,W. and Parzefall,W. (1988)

31. Maier,P., Saad,B. and Ohno,K. (1995) New approaches for the preservation Quantitative structure–activity studies on effects of sixteen different

of metabolic zonation in rat hepatocyte cultures. In Goldberg,A.M. and steroids on growth and monooxygenase of rat liver. Cancer Res., 48,

van Zutphen,L.F.M. (eds), The World Congress on Alternatives and Animal 2462–2468.

Use in the Life Sciences: Education, Research, Testing. Alternative Methods

11. Parzefall,W., Monschau,P. and Schulte-Hermann,R. (1989) Induction by

in Toxicology and the Life Sciences. Mary Ann Liebert, New York, Vol.

cyproterone acetate of DNA synthesis and mitosis in primary cultures of

11, pp. 213–219. adult rat hepatocytes in serum free medium. Arch. Toxicol., 63, 456–461.

12. Parzefall,W., Erber,E., Sedivy,R. and Schulte-Hermann,R. (1991) Testing 32. Kurki,P., Ogata,K. and Tan,E.M. (1988) Monoclonal antibodies to for induction of DNA synthesis in human hepatocyte primary cultures by proliferating cell nuclear antigen (PCNA)/cyclin as probes for proliferating rat liver tumor promoters. Cancer Res., 51, 1143–1147. cells by immunofluorescence microscopy and flow cytometry. J. Immunol. 13. Lang,R. and Redmann,U. (1979) Non-mutagenicity of some sex hormones Methods, 109, 49–59.

in the Ames’s Salmonella microsome mutagenicity test. Mutat. Res., 67, 33. Vintermyr,O.K., Boe,R., Bruland,T., Houge,G. and Døskeland,S.O. (1993) 361–365. Elevated cAMP gives short-term inhibition and long-term stimulation of 14. Lang,R. and Reimann,R. (1993) Studies for a genotoxic potential of some hepatocyte DNA replication: roles of the cAMP-dependent protein kinase

endogenous and exogeneous sex steroids. I. Communication: examination _{subunits. J. Cell. Physiol., 156, 160–170.}

for the induction of gene mutations using Ames Salmonella/microsome _{34. Lowry,O.H., Rosebrough,N.J., Farr,A.L. and Randall,R.J. (1951) Protein} test and HGPRT test in V 79 cells. Environ. Mol. Mutagen., 21, 272–304. _{measurement with the Folin phenol reagent. J. Biol. Chem., 193, 265–275.} 15. Neumann,I., Thierau,D., Andrae,U., Greim,H. and Schwarz,L.R. (1992) _{35. Bruinink,A. (1992) Serum-free monolayer cultures of fetal chick brain} Cyproterone acetate induces DNA damage in cultured rat hepatocytes _{and retina: immunoassays of developmental markers, mathematical data} and preferentially stimulates DNA synthesis inγ-glutamyltranspeptidase- _{analysis, and establishment of optimal culture conditions. In Zbinden,G.} positive cells. Carcinogenesis, 13, 373–378. _{(ed.), The Brain in Bits and Pieces. MTC, Zollikon, Switzerland, pp. 40–46.} 16. Martelli,A., Mattioli,F., Fazio,S., Andrae,U. and Brambilla,G. (1995) DNA _{36. Holzer,C. and Maier,P. (1987) Maintenance of periportal and pericentral} repair synthesis and DNA fragmentation in primary cultures of human _{oxygen tensions in primary rat hepatocyte: influence on cellular DNA and} and rat hepatocytes exposed to cyproterone actetate. Carcinogenesis, 16, _{protein content monitored by flow cytometry. J. Cell. Physiol., 133,}

1265–1269. _297–304.

17. Kasper,P., Tegethoff,K. and Mueller,L. (1995) In vitro mutagenicity studies

37. Padgham,C.R., Boyle,C.C., Wang,X.J., Raleigh,S.M., Wright,M.C. and on cyproterone acetate using female rat hepatocytes for metabolic activation

Paine,A.J. (1993) Alteration of transcription factor mRNAs during the and as indicator cells. Carcinogenesis, 16, 2309–2314.

isolation and culture of rat hepatocytes suggests the activation of a 18. Kerdar,R.S., Baumann,A., Brudny-Klo¨ppel,M., Biere,H., Blode,H. and

proliferative mode underlies their de-differentiation. Biochem. Biophys. Kuhnz,W. (1995) Identification of 3α-hydroxycyproterone acetate as a

Res. Commun., 197, 599–605.

metabolite of cyproterone acetate in the bile of female rats and the potential

38. Dietrich,D.R. (1993) Toxicological and patholocgical applications of of this and other already known or putative metabolites to form DNA

proliferating cell nuclear antigen (PCNA), a novel endogenous marker for adducts in vitro. Carcinogenesis, 16, 1835–1841.

cell proliferation. Crit. Rev. Toxicol., 23, 77–109. 19. Werner,S., Topinka,J., Wolff,T. and Schwarz,L.R. (1995) Accumulation

39. Kumatori,A., Nakamura,T. and Ichihara,A. (1991) Cell density dependent and persistence of DNA adducts of the synthetic steroid cyproterone

expression of the c-myc gene in primary cultured rat hepatocytes. Biochem. acetate in rat liver. Carcinogenesis, 16, 2369–2372.

Biophys. Res. Commun., 178, 480–485.

20. Topinka,J., Andrae,U., Schwarz,L.R. and Wolff,T. (1993) Cyproterone

40. Shimbara,N., Takashina,M., Sato,C., Iizuka,M., Kobayashi,S., Tanaka,K. acetate generates DNA adducts in rat liver and in primary rat hepatocyte

and Icjihara,A. (1992) c-myc expression is down-regulated by cell–cell cultures. Carcinogenesis, 14, 423–427.

and cell–extracellular matrix contacts in normal hepatocytes, but not in 21. Fausto,N. (1991) Multifunctional roles for transforming growth factorβ1.

hepatoma cells. Biochem. Biophys. Res. Commun., 184, 825–831.

Lab. Invest., 65, 497–498.

41. Maier,P., Schawalder,H.P., Milosevic,N., Bouis,P., Thomas,H. and Ohno,K. 22. Nakamura,T., Tomita,Y., Hirai,R., Yamaoka,K., Kaji,K. and Ichihara,A.

(1997) Insulin and glucagon downregulate specifically the phenobarbital (1985) Inhibitory effect of transforming growth factorsβin DNA synthesis

inducible CYP2B1/2 in cultured rat liver cells. Biochem. Pharmacol., of rat hepatocytes in primary culture, Biochem. Biophys. Res. Commun.,

42. Oberhammer,F.A. and Qin Hong-Min (1995) Effect of three tumour promoters on the stability of hepatocyte cultures and apoptosis after transforming growth factor-β1. Carcinogenesis, 16, 1363–1371. 43. Schulte-Hermann,R. and Parzefall,W. (1980) Adaptive response of rat liver

to the gestagen and anti-androgen cyproterone acetate and other inducers. I. Induction of drug metabolizing enzymes. Chem.-Biol. Interactions., 31, 279–286.

44 Schwarz,L.R., Werner,S., Topinka,J. Andrae,U., Neumann,I. and Wolff,T. (1995) The liver as origin and target of reactive intermediates exemplified by the progesterone derivative, cyproterone acetate. In Snyder,R. et al. (eds), Biological Reactive Intermediates V. Plenum Press, New York, NY, 243–251.

45. Bursch,W., Du¨sterberg,B., Schulte-Hermann,R. (1986) Growth, regression and cell death in rat liver as related to tissue levels of the hepatomitogen cyproterone acetate. Arch. Toxicol., 59, 221–227.

46. Braun,L., Mead,L.E., Panzica,M., Mikumo,R., Bell,G.I. and Fausto,N. (1988) Transforming growth factor β mRNA increases during liver regeneration: a possible paracrine mechanism of growth regulation. Proc.

Natl Acad. Sci. USA, 85, 1539–1543.

47. Thoresen,G.H., Refsnes,M. and Christoffersen,T. (1992) Inhibition of hepatocyte DNA synthesis by transforming growth factorβ1 and cyclic AMP: effect immediately before the G1/S border. Cancer Res., 52, 3598–3603.

48. Bursch,W., Oberhammer,F., Jirtle,R.L., Askari,M., Sedivy,R., Grasl-Kraupp,B., Purchio,A.F. and Schulte-Hermann,R. (1993) Transforming growth factor-β1 as a signal for induction of cell death by apoptosis. Br.

J. Cancer, 67, 531–536.

49. Oberhammer,F.A., Pavelka,M., Sharma,S., Tiefenbacher,R., Purchio,A.F., Bursch,W. and Schulte-Hermann,R. (1992) Induction of apoptosis in cultured hepatocytes and in regressing liver by transforming growth factor

β1. Proc. Natl Acad. Sci. USA, 89, 5408–5412.

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

51. Reid,L.M., Fiorino,A.S., Sigal,S.H. and Holst,P.A. (1992) Extracellular matrix gradients in the space of disse: relevance to liver biology.

Hepatology, 15, 1198–1203.

52. Arber,A., Zajicek,G. and Ariel,I. (1988) The streaming liver II. Hepatocyte life history. Liver, 8, 80–87.

53. Gebhardt,R. and Jonitza,D. (1991) Different proliferative response of periportal and perivenous hepatocytes to EGF. Biochem. Biophys. Res.

Commun., 181, 1201–1207.

54. Roberts,R.A., Soames,A.R., Gill,J.H., James,N.H. and Wheeldon,E.B. (1995) Non-genotoxic hepatocarcinogens stimulate DNA synthesis and their withdrawal induces apoptosis, but in different hepatocyte populations.

Carcinogenesis, 16,1693–1698.

55. Parzefall,W., Erber,E., Kainzbauer,E. and Schulte-Hermann,R. (1996) Effects of insulin, glucagon, and triiodothyronine on DNA synthesis in rat hepatocyte primary cultures induced by liver tumor promoters and EGF. Toxicol. In Vitro, 10, 183–193.

56. Maier,P. and Schawalder,H.P. (1993) Physiological oxygen tension modulates the chemically induced mitogenic reponse of cultured rat hepatocytes. J. Cell. Physiol., 156, 119–129.

57. Duivenvoorden,W.M.C. and Maier,P. (1994) Nongenotoxic carcinogens shift cultured rat hepatocytes into G1 cell cycle phase: influence of tissue oxygen tension on cells with different ploidy. Eur. J. Cell Biol., 64, 368–375.

Received on March 29, 1996; revised on October 16, 1996; accepted on January 14, 1997