Epidermal growth factor (EGF) and somatomedin C regulate G1 progression in competent BALB/c-3T3 cells

The activity in platelet-poor plasma that allowed density-arrested BALB/c-3T3 cells rendered competent by a transient exposure to platelet-derived growth factor (PDGF) to traverse G1 and enter the S phase has been termed progression activity. Epidermal growth factor (EGF) and somatomedin C-supplemented medium was shown to be capable of replacing the progression activity of 5% platelet-poor plasma (PPP) for competent density-inhibited BALB/c-3T3 cells. Exposure of competent cells to medium supplemented with EGF and somatomedin C reduced the 12 h minimum G1 lag time found in plasma-supplemented medium by 2 h. It is suggested that the reduction in the minimum time required for progression through G1 is due to the availability of free, unbound somatomedin C. Complete G1 traverse required both EGF and somatomedin C; however, the traverse of the last 6 h of G1 and entry into the S phase required only somatomedin C. Though EGF and somatomedin C could replace the G1 phase progression activity of plasma, medium supplemented with EGF and somatomedin C did not support complete cell cycle traverse or growth of sparse cultures of BALB/c-3T3 cells.     

Platelet-derived growth factor stimulation of [3H]-glucosamine incorporation in density-arrested BALB/c-3T3 cells

G0/G1 traverse in density-arrested BALB/c-3T3 cells is controlled by multiple serum-derived growth factors. Platelet-derived growth factor (PDGF) initiates a proliferative response, whereas factors present in plasma facilitate progression through G0/G1. In the absence of competence formation, progression factors are unable to stimulate cell cycle traverse. We have identified the stimulation of a biochemical process specific to competence formation in BALB/c-3T3 cells. PDGF treated BALB/c-3T3 cells incorporated 5-10-fold more [3H]-glucosamine (GlcN) into acid-insoluble material as compared to platelet-poor plasma (PPP) treated cultures. Increased GlcN incorporation occurred in density-arrested BALB/c-3T3 cells in response to treatment with other competence factors, fibroblast growth factor, and Ca3 (PO4)2 and was not due to cell-cycle traverse. Stimulation of [3H]-GlcN incorporation by PDGF was time dependent, and increased incorporation of [3H]-GlcN into protein required de novo protein synthesis. Several mechanisms through which PDGF could increase GlcN incorporation into cellular material were examined. Results of these studies suggest an increase in the cellular capacity to glycosylate proteins is a response to or a part of competence formation.     

Control of the Balb/c-3T3 cell cycle by nutrients and serum factors: analysis using platelet-derived growth factor and platelet-poor plasma.

Much controversy regarding the relationship between nutrients and serum in regulation of cell growth can be reconciled by recognizing that serum contains multiple factors which regulate different events in the cell cycle. Serum was fractionated into a platelet-derived growth factor (PDGF), which induces cells to become competent to synthesize DNA, and plasma which allows competent cells to traverse G0/G1 and enter the S phase. Nutrients are not required for the cellular response to PDGF; however amino acids are required for plasma to promote the entry of PDGF-treated, competent cells into S phase. The nutrient independent, PDGF-modulated, growth regulatory event (competence) is located 12 hours prior to the G1/S phase boundary in quiescent, density-arrested Balb/c-3T3 cells. The nutrient dependent, plasma-modulated event is located six hours prior to the G1/S phase boundary and corresponds in concentration of amino acids required for DNA synthesis. Infection of density-arrested Balb/c3T3 cells with SV40 overrides both the nutrient independent and the nutrient dependent growth regulatory events.     

Elevated intracellular concentrations of cyclic AMP inhibited serum-stimulated, density-arrested BALB/c-3T3 cells in mid G1

The stimulation of DNA synthesis in quiescent, density-arrested BALB/c-3T3 cells by platelet-derived growth factor in plasma-supplemented medium was inhibited by the presence of isobutylmethylxanthine (IBMX) and cholera toxin, although neither IBMX or cholera toxin when used alone inhibited the stimulation of DNA synthesis. The cells were reversibly inhibited in mid G1 at a point 6 hr prior to the initiation of DNA synthesis. The inhibition of cell cycle traverse was associated with a 10-15 fold increase in cellular cyclic AMP concentration over basal levels. The reversal of this inhibition by removal of IBMX was correlated with a dramatic decrease in cyclic AMP levels. The traverse of G1 and the initiation of DNA synthesis after release from the cholera toxin and IBMX inhibition was dependent on the presence of plasma in the medium. Either somatomedin C (10-20 ng/ml) or insulin (10(-6)-10(-5) M) completely replaced the plasma requirement for late G1 progression and entry into S phase. Once the inhibited cells were released from the IBMX and cholera toxin block a subsequent increase in cyclic AMP did not prevent entry into S phase. The presence of cholera toxin alone inhibited the stimulation of human dermal fibroblasts. The elevation of intracellular cyclic AMP levels in the human dermal fibroblasts by cholera toxin was two to three fold greater than that found in the BALB/c-3T3 cells in the presence of cholera toxin and the IBMX.     

Murine mRNA for the beta-subunit of integrin is increased in BALB/c-3T3 cells entering the G1 phase from the G0 state

The amount of murine mRNA for the beta-subunit of integrin is enriched 6-fold when BALB/c-3T3 cells traverse from the G0 state to the G1 phase, whereas it remains at the basal level when the cells are growing continuously. The peak of its appearance is at 10 h after serum stimulation. The increase in integrin mRNA at a specific point in cell proliferation may be correlated with growth-signal transduction.     

Intracellular univalent cations and the regulation of the BALB/c-3T3 cell cycle

Addition of serum to density-arrested BALB/c-3T3 cells causes a rapid increase in uptake of Na+ and K+, followed 12 h later by the onset of DNA synthesis. We explored the role of intracellular univalent cation concentrations in the regulation of BALB/c-3T3 cell growth by serum growth factors. As cells grew to confluence, intracellular Na+ and K+ concentrations ([Na+]i and [K+]i) fell from 40 and 180 to 15 and 90 mmol/liter, respectively. Stimulation of growth of density-inhibited cells by the addition of serum growth factors increased [Na]i by 30% and [K+]i by 13-25% in early G0/G1, resulting in an increase in total univalent cation concentration. Addition of ouabain to stimulated cells resulted in a concentration-dependent steady decrease in [K+]i and increase in [Na+]i. Ouabain (100 microM) decreased [K+]i to approximately 60 mmol/liter by 12 h, and also prevented the serum- stimulated increase in 86Rb+ uptake. However, 100 microM ouabain did not inhibit DNA synthesis. A time-course experiment was done to determine the effect of 100 microM ouabain on [K+]i throughout G0/G1 and S phase. The addition of serum growth factors to density-inhibited cells stimulated equal rates of entry into the S phase in the presence or absence of 100 microM ouabain. However, in the presence of ouabain, there was a decrease in [K+]i. Therefore, an increase in [K+]i is not required for entry into S phase; serum growth factors do not regulate cell growth by altering [K+]i. The significance of increased total univalent cation concentration is discussed.     

Identification of a BALB/c-3T3 cell protein modulated by platelet-derived growth factor.

The platelet-derived growth factor (PDGF) stimulates density-arrested BALB/c-3T3 cells to synthesize a protein (pII; Mr, 35,000) that is constitutively synthesized by spontaneously transformed BALB/c-3T3 (ST2-3T3) cells which do not require PDGF for growth. Antisera against a major excreted protein family (MEP) of retrovirus-transformed cells quantitatively precipitated cellular pII. PDGF-stimulated pII has the same molecular weight, a similar charge, and similar antigenic determinants as authentic MEP isolated from ST2-3T3 or retrovirus-transformed cells. MEP represented about 2% of the nonnuclear proteins synthesized by ST2-3T3 cells and 0.3 to 0.6% of the proteins synthesized by PDGF-treated BALB/c-3T3 cells, a three- to sixfold increase over the background. In BALB/c-3T3 cells, less PDGF was required for pII (MEP) synthesis than for DNA synthesis. PDGF induced a selective increase in pII (MEP) within 40 min. Such preferential synthesis was inhibited by brief treatment with actinomycin D, suggesting a requirement for newly formed RNA. The constitutive synthesis of pII (MEP) by ST2-3T3 cells was not inhibited by actinomycin D. Five spontaneously or chemical carcinogen-transformed tumorigenic BALB/c-3T3 cell lines were studied; they neither required PDGF for growth nor responded to it. These cell lines became arrested at confluence with a G1 DNA content. Each of these independently isolated lines synthesized pII (MEP) constitutively. Thus, the synthesis of pII (MEP) may be required, but is not sufficient, for PDGF-modulated DNA synthesis.     

A 63 kDa protein is secreted from BALB/c-3T3 cells entering the G1 phase from the G0 state

A 63 kDa protein is detectable in the culture fluid of mouse BALB/c-3T3 cells traversing from the G0 state to the G1 phase, whereas it is undetectable in the culture fluid of quiescent or growing BALB/c-3T3 cells. Secretion of the protein is maximal at 10 h after serum addition. G0-specific ts mutant cells (rat tsJT60) also secrete the 63 kDa protein only when the quiescent cells are stimulated by serum addition at permissive temperature. These facts indicate that the 63 kDa protein is secreted only from cells traversing from the G0 state to the G1 phase.     

Effect of ouabain on growth regulation by serum components in Balb/c-3T3 cells: inhibition of entry into S phase by decreased protein synthesis

The effect of inhibition of the cell membrane Na+-K+ pump on the Balb/c-3T3 cell growth cycle was studied. Inhibition of the Na+-K+ pump resulted in a dose-dependent reduction of intracellular K+ concentration ((K+)i). However, inhibition of protein synthesis in Go/G1 and of subsequent entry into S phase occurred only after (K+)i fell below a critical threshold (50-60 mmoles/liter). Thus, when the (K+)i falls below a critical threshold, protein synthesis is inhibited, preventing cells from entering the S phase. The platelet-derived growth factor (PDGF) induces cells to become "competent" to traverse the cell cycle; the platelet-poor plasma component of serum allows competent cells to progress through G0/G1 and enter S phase. Inhibition of the Na+-K+ pump did not prevent the induction of competence by PDGF, but it did reversibly inhibit plasma-mediated events in early G0/G1. Similarly, cycloheximide inhibited plasma-mediated events but did not prevent PDGF-induced competence. Thus, protein synthesis may not be required for induction of competence; alternatively, the induction of the competent state may occur in these cells after removal of PDGF and protein synthesis inhibitor. Protein synthesis is required for subsequent plasma-mediated events in G0/G1.     

Cell cycle-dependent gene expression in V point-arrested BALB/c-3T3 cells

Density-arrested BALB/c-3T3 cells stimulated to proliferate in an amino acid-deficient medium arrest in mid-G₁ at a point termed the V point. Cells released from V point arrest require 6 hr to traverse late G₁ and enter S phase. As data presented here show that mRNA synthesis is needed for 2–3 hr after release of cells from the V point, after which inhibition of mRNA synthesis does not prevent entry into S phase, we used this mid-G₁ arrest protocol to analyze gene expression in late G₁. We found that although stimulation of cells in amino acid-deficient medium did not inhibit the induction of genes expressed in early G₁, genes normally expressed in late G₁ were expressed only after release from the V point. The expression of late G₁ genes in cells released from the V point was temporally similar, in respect to G₁ location, as was seen in stimulation of quiescent G_o cells. As this protocol effectively divides gene expression into early (pre-V point) and late (post-V point) categories, it should be useful in studies of growth factor-modulated events that regulate traverse of late G₁ and commitment to DNA synthesis. In addition, we used c-myb antisense oligonucleotides to show that c-myb expression, which occurs in late G₁, is required for BALB/c-3T3 fibroblasts to traverse late G₁ and initiate DNA synthesis. © 1993 Wiley-Liss, Inc.     

Induction of DNA synthesis in dog thyrocytes in primary culture: synergistic effects of thyrotropin and cyclic AMP with epidermal growth factor and insulin

We have investigated the growth effects of thyrotropin (TSH) (mimicked by forskolin and acting through cyclic AMP), epidermal growth factor (EGF), serum (10%) and insulin on quiescent dog thyroid epithelial cells in primary culture in a serum-free defined medium. These cells were previously shown to retain the capacity to express major thyroid differentiation markers. In the presence of insulin and after a similar prereplicative phase of 18 +/- 2h, TSH, EGF, and serum promoted DNA synthesis in such quiescent cells only a minority of which had proliferated in vitro before stimulation. The combination of these factors induced more than 90% of the cells to enter S phase within 48 h and near exponetial proliferation. Analysis of the cell cycle parameters of the stimulated cells revealed that the G1 period duration was similar to the length of the prereplicative phase of quiescent thyroid cells; this might indicate that they were in fact in an early G1 stage rather than in G0 prior to stimulation. TSH and EGF action depended on or was potentiated by insulin. Strikingly, nanomolar concentrations of insulin were sufficient to support stimulation of DNA synthesis by TSH, while micromolar concentrations of insulin were required for the action of EGF. This suggests that insulin supported the action of TSH by acting on its own high affinity receptors, whereas its effect on EGF action would be related to its somatomedinlike effects at high supraphysiological concentrations. Insulin stimulated the progression in the prereplicative phase initiated by TSH or forskolin. In addition, in some primary cultures TSH must act together with insulin to stimulate early events of the prereplicative phase. In the presence of insulin, EGF, and forskolin, an adenylate cyclase activator, markedly synergized to induce DNA synthesis. Addition of forskolin 24 h after EGF or EGF 24 h after forskolin also resulted in amplification of the growth response but with a lag equal to the prereplicative period observed with the single compound. This indicates that events induced by the second factor can no longer be integrated during the prereplicative phase set by the first factor. These findings demonstrate the importance of synergistic cooperation between hormones and growth factors for the induction of DNA synthesis in epithelial thyroid cells and support the proposal that essentially different mitogenic pathways--cyclic AMP-dependent or independent--may coexist in one cell.     

Cell cycle dependency of murine cytomegalovirus replication in synchronized 3T3 cells.

Synchronized murine 3T3 cells have been used to investigate the possible dependency of murine cytomegalovirus replication upon the cell cycle. The normal latent period of 12 h characteristic of asynchronous 3T3 cells was protracted to more than 24 h after an early G1 infection in synchronous cells. In this case viral progeny were not detected until after the initiation of the host S-phase. Cells maintained in the G1 phase did not replicate virus. This failure could not be explained by a decrease in virus penetration but was apparently due to a requirement for an event associated with the host S-phase. Thymidine-induced inhibition of cell cycle traverse also blocked virus replication. Viral DNA synthesis did not initiate until after the initiation of host DNA. In contrast, herpes simplex virus type 1 replicated in 3T3 cells independently of the cell cycle.     

Mitogenic response to epidermal growth factor (EGF) modulated by platelet-derived growth factor in cultured fibroblasts

The mitotic effects of epidermal growth factor (EGF) were investigated in two cultured fibroblast lines, BALB/c-3T3 and C3H 10T1/2 cells. EGF (30 ng/ml) added to quiescent 3T3 cells in medium containing either platelet-poor plasma or 10(-5) M insulin caused only minimal increases in the percentage of cells stimulated to initiate DNA synthesis. In contrast, EGF acted synergistically with either insulin or plasma to stimulate DNA synthesis in quiescent cultures of 10T1/2 cells, although the maximum effects of EGF were measured at concentrations several-fold greater than those found in either serum or plasma. In either 3T3 or 10T1/2 cells a transient preexposure to platelet-derived growth factor (PDGF) caused over a 10-fold increase in the sensitivity to the mitogenic effects of EGF. It is therefore possible that a primary action of PDGF is to increase the sensitivity of fibroblasts to EGF, independent of whether EGF alone is found to be mitogenic.     

Cell cycle arrest of heat-inactivated ts 2 BALB/C-3T3 mouse fibroblasts, temperature-sensitive for DNA synthesis

The ts 2 derivative of BALB/c-3T3 mouse fibroblasts is a cell division cycle (cdc) mutant. Upon expression of the heat-sensitive defect, ts 2 cells arrest late in G1 at, or very near the G1/S traverse. This conclusion derives from three kinds of experiment. In the first the cells were brought to different stages of the cell cycle by physiological manipulation, or with specific anti-metabolites. They were then released from the resulting blocks, and their subsequent cell-cycle progression, at the permissive- and non-permissive temperature (npt), was followed. The second experiment was an execution point analysis. In the third, premature chromosome condensation was performed between metaphase HeLa cells and temperature-blocked ts 2 cells. The resulting prematurely-condensed chromosomes were largely of the morphotype of very late G1 cells. The ts 2 cells are prevented from expressing their defect by temporary incubation at 38.5 degrees C in the G0, non-cycling state and by prior arrest in early S phase, imposed by hydroxyurea treatment. Such prevention is not allowed ts 2 cells incubated at the npt in the absence of isoleucine, a procedure which brings cells to mid-G1 arrest.     

Stimulation of DNA synthesis in quiescent rabbit chondrocytes in culture by limited exposure to somatomedin-like growth factors

Cartilage-derived factor (CDF), extracted from fetal bovine cartilage, and multiplication-stimulating activity (MSA) stimulated DNA synthesis in quiescent rabbit costal chondrocytes in culture under serum-free conditions. As described previously, when added in the presence of fibroblast growth factor (FGF) or epidermal growth factor (EGF) a somatomedin-like growth factor, CDF or MSA, synergistically stimulated DNA synthesis in the cultured chondrocytes. The present study showed that exposure of the cells to MSA or CDF for only the initial 5 h was sufficient for transmission of their full stimulatory effect. Furthermore, the limited exposure did not alter the time course of stimulation of DNA synthesis: [3H]thymidine incorporation into DNA began to increase after 16 h and reached a maximum after 24 h. In contrast to the somatomedin-like growth factors, FGF and EGF were required continuously in the culture medium during traverse of the entire G1 phase for stimulation of DNA synthesis, and the mitogenic effects of FGF and EGF in cultured chondrocytes were stronger than those of CDF and MSA. Synergistic stimulation of DNA synthesis by CDF or MSA in the presence of FGF or EGF could be observed as long as FGF or EGF was continuously present, even when CDF or MSA was withdrawn after the first 5 h of culture. These findings suggest that, in contrast to FGF and EGF, somatomedin-like growth factors affect an early distinct stage in the G1 phase of chondrocytes.     

Cellular ras activity is required for passage through multiple points of the G0/G1 phase in BALB/c 3T3 cells.

Microinjection experiments demonstrated a requirement for cellular ras activity late in G1. In this study, we used two separate methods to identify an additional requirement for cellular ras activity early in the G0/G1 phase of the cell cycle. Quiescent BALB/c cells were injected with anti-ras antibody prior to stimulation with serum. The cells would therefore be inhibited in progression through the cell cycle at the earliest point requiring ras function. Alternatively, cells were inhibited in late G1 as in previous studies by injecting anti-ras several hours after serum addition to quiescent cells. The injected cultures were then treated with chemical cell cycle inhibitors known to function in mid-G1. Cells injected with anti-ras prior to serum stimulation were retained at a point of ras requirement prior to the execution point of the chemical inhibitor, while cells injected 3 to 5 h after serum stimulation were retained at a point of ras requirement downstream of the execution point of the chemical inhibitor. To confirm these results, quiescent BALB/c cells were injected with anti-ras antibody prior to or several hours following serum addition. In this case, however, second injections of oncogenic ras or adenoviral E1A protein were performed to overcome the inhibitory effects of the anti-ras antibody. Cells injected prior to serum addition were clearly inhibited at an early point of Ras requirement since they required 5 or 6 h longer to enter S phase than cells injected with anti-ras antibody after serum addition.     

Resumption of cell cycle in BALB/c-3T3 fibroblasts arrested by polyamine depletion: relation with "competence" gene expression

Serum deprivation arrests BALB/c-3T3 fibroblasts (clone A31) in G0 phase, where resumption of the cell division cycle can be induced by addition of serum or of specific growth factors in a defined sequence: PDGF (inducer of a state of "competence," characterized by the expression of a family of genes including c-myc), epidermal growth factor EGF and IGF1 (Leof et al., 1982, 1983). When exponentially growing A31 cells are placed for greater than or equal to 2 days in a medium containing the alpha-difluoromethylornithine (alpha DFMO), an irreversible inhibitor of ornithine decarboxylase, they become arrested in G1 phase as a consequence of polyamine depletion (Medrano et al., 1983). In the alpha DFMO-arrested cells, addition of putrescine (60 microM) in a culture medium containing 6% fetal calf serum (FCS), but not in serum-free medium, is sufficient to induce G1 progression and entry into S phase (as determined by 3H-thymidine incorporation). The level of "competence" mRNAs is high in alpha DFMO-arrested cells. After addition of putrescine in FCS-containing medium, these mRNAs continue to be present for at least 3 h. A large proportion of alpha DFMO-arrested cells can be induced to progress to S phase by insulin (1 microM, acting via IGF1 receptor) plus putrescine in a serum-free medium (greater than or equal to 50% of FCS effect). In this case, the levels of "competence" mRNAs become low or undetectable within 3 h, EGF (10 nM) plus insulin had only slightly greater effect than insulin alone on the progression of alpha DFMO-arrested cells. When the alpha DFMO-arrested cells are subsequently incubated during 3 days in a low-serum-containing medium (0.25% FCS), they do not replenish their supply of polyamines, and then continue to express the c-myc gene. The recruitment of the polyamine-depleted, serum-deprived cells into the cell division cycle does not require PDGF and can be induced by addition of EGF and insulin plus putrescine. These data indicate that alpha DFMO arrests majority of the cells at a point situated beyond the PDGF- and EGF-dependent portion of G1 phase. During the subsequent serum deprivation, the alpha DFMO-arrested cells remain "competent" (PDGF-independent), probably as a consequence of their continued expression of c-myc (and possibly other PDGF-inducible genes).     

EGF-dependent growth inhibition in MDA-468 human breast cancer cells is characterized by late G1 arrest and altered gene expression.

The MDA-468 human breast cancer cell line displays the unusual phenomenon of growth inhibition in response to pharmacological concentrations of EGF. This study was initiated with the objective of elucidating the cellular mechanisms involved in EGF-induced growth inhibition. Following EGF treatment the percentage of MDA-468 cells in G1 phase increased, together with a concomitant depletion in S and G2/M phase populations, as revealed by flow cytometry of DNA content. The apparent G1 block in the cell cycle was confirmed by treating the cells with vinblastine. DNA synthesis was reduced to about 35% of that measured in control, untreated cells after 48 h of EGF treatment, as measured by the incorporation of [3H]thymidine. DNA synthesis returned to normal following the removal of EGF from the growth-arrested cells. In order to locate the EGF-induced event responsible for the G1 arrest more precisely, we examined the expression of certain cell cycle-dependent genes by Northern blot analysis. EGF treatment did not alter either the induction of the early G1 marker, c-myc, or the expression of the late G1 markers, proliferating cell nuclear antigen, and thymidine kinase. However, EGF-treated cells revealed down regulation of p53 and histone 3.2 expression, which are expressed at the G1/S boundary and in S phase, respectively. These results indicate that EGF-induced growth inhibition in MDA-468 human breast cancer cells is characterized by a reversible cell cycle block at the G1/S boundary.     

Temporal requirement for epidermal growth factor and insulin in the stimulation of hepatocyte DNA synthesis

Primary monolayer cultures of adult rat hepatocytes were used to study the temporal interaction of epidermal growth factor (EGF) and insulin in their stimulation of DNA synthesis. The hepatocytes were cultured both under defined conditions and with serum. EGF and insulin interacted synergistically. The entry into S phase (G1 exit) followed first-order kinetics both in untreated and hormone-stimulated cells. Addition of EGF and insulin at the time of plating did not alter the lag period before the DNA synthesis started (25-26 h), but the rate constant for the S phase entry increased five- to sixfold. Experiments where the time of hormone addition was varied indicated that insulin exerted its strongest effect at the time of plating, whereas the cells became more responsive to EGF after being cultured for up to 40-50 h. The responsiveness to EGF at these later stages required an early exposure of the hepatocytes to insulin. When the administration of EGF to insulin-pretreated hepatocytes was postponed for 44 h after plating in serum-free medium, the cellular sensitivity was increased as compared to EGF treatment at 0 h (a one-log shift of the dose-effect curve), the rate of S phase entry was more rapid, and the lag period for the onset of the EGF effect (i.e., shift of rate constant) was shortened (6-7 h vs. 26 h).     

Dual effects of glucagon and cyclic AMP on DNA synthesis in cultured rat hepatocytes: stimulatory regulation in early G1 and inhibition shortly before the S phase entry

Although several lines of evidence implicate cyclic AMP in the humoral control of liver growth, its precise role is still not clear. To explore further the role of cyclic AMP in hepatocyte proliferation, we have examined the effects of glucagon and other cyclic AMP-elevating agents on the DNA synthesis in primary cultures of adult rat hepatocytes, with particular focus on the temporal aspects. The cells were cultured in a serum-free, defined medium and treated with epidermal growth factor (EGF), insulin, and dexamethasone. Exposure of the hepatocytes to low concentrations (10 pM-1 nM) of glucagon in the early stages of culturing (usually within 6 h from plating) enhanced the initial rate of S phase entry without affecting the lag time from the plating to the onset of DNA synthesis, whereas higher concentrations inhibited it. In contrast, glucagon addition at later stages (24-45 h after plating) produced only the inhibition. Thus, if glucagon was added at a time when there was a continuous EGF/insulin-induced recruitment of cells to S phase, the rate of G1-S transition was markedly decreased within 1-3 h. This inhibitory effect occurred at low glucagon concentrations (ID50 less than 1 nM) and was mimicked by cholera toxin, forskolin, isobutyl methylxanthine, and 8-bromo cyclic AMP. The results indicate that cyclic AMP has dual effects on hepatocyte proliferation with a stimulatory modulation early in the prereplicative period (G0 or early G1), and a marked inhibition exerted immediately before the transition from G1 to S phase.