Transforming Growth Factor @ and Epidermal Growth Factor Alter Calcium Influx and Phosphatidylinositol Turnover in Rat- 1 Fibroblasts*

Transforming growth factor type /3 (TGFB) alters the cellular response to epidermal growth factor (EGF) for a variety of processes ranging from early transport activities and gene transcription to mitogenesis. In order to test the hypothesis that altered signal trans- duction mechanisms may mediate both the transforming effects of TGFB and the modulation of EGF-stimu- lated processes by TGFB, we have examined second messenger levels in response to growth factor treatment. The addition of EGF or prolonged treatment with TGFB increased the rate of 46Ca influx in serum-de-prived, confluent Rat- 1 cells, while the addition of EGF to TGFB-pretreated cells produced an additive increase in Ca2+ influx. The stimulation of Ca2+ influx by TGFB was only observed at incubation times greater than 1 h and was inhibited by inclusion of actinomycin D, suggesting that a newly transcribed gene product was required for the observed response to TGFB. Both EGF and TGFB displayed similar time and concentration dependencies for stimulation of Ca2+ influx and for accumulation of inositol trisphosphate (Ips). The increase in Ips accumulation in response to either EGF or TGFB required the presence of extracellular Ca2+, and the observed concentration dependencies were similar for the stimulation of phosphatidylinositol turnover and Ca2+ influx. The EGF- and TGFB-stimu-lated increases in Ca2+ influx could be blocked by co- balt, cadmium, and [ethylenebis(oxyethylenenitrilo)]


From the Oregon Health Sciences University, Portland, Oregon 97201
Transforming growth factor type / 3 (TGFB) alters the cellular response to epidermal growth factor (EGF) for a variety of processes ranging from early transport activities and gene transcription to mitogenesis. In order to test the hypothesis that altered signal transduction mechanisms may mediate both the transforming effects of TGFB and the modulation of EGF-stimulated processes by TGFB, we have examined second messenger levels in response to growth factor treatment. The addition of EGF or prolonged treatment with TGFB increased the rate of 46Ca influx in serum-deprived, confluent Rat-1 cells, while the addition of EGF to TGFB-pretreated cells produced an additive increase in Ca2+ influx. The stimulation of Ca2+ influx by TGFB was only observed at incubation times greater than 1 h and was inhibited by inclusion of actinomycin D, suggesting that a newly transcribed gene product was required for the observed response to TGFB. Both EGF and TGFB displayed similar time and concentration dependencies for stimulation of Ca2+ influx and for accumulation of inositol trisphosphate (Ips). The increase in Ips accumulation in response to either EGF or TGFB required the presence of extracellular Ca2+, and the observed concentration dependencies were similar for the stimulation of phosphatidylinositol turnover and Ca2+ influx. The EGF-and TGFB-stimulated increases in Ca2+ influx could be blocked by cobalt, cadmium, and [ethylenebis(oxyethylenenitrilo)] tetraacetic acid, but not by specific Ca2+ channel blockers such as nifedipine or verapamil, suggesting that these growth factors do not act via L-type voltagesensitive calcium channels. Those calcium blockers which inhibited Ca2+ influx also inhibited inositol phosphate release. These data, taken together, indicate that Ca" influx and inositol phosphate release are coupled in Rat-1 cells and suggest that influx of Ca2+ from the extracellular medium is responsible for the changes in Ips accumulation observed in response to both EGF and TGFB.
Epidermal growth factor (EGF)' provides a potent stimulus * This work was supported by United States Public Health Service Grants CA-39360 and CA-47404. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked ''advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed. for a variety of cell types and initiates a complex series of events culminating in mitogenesis (1). Transforming growth factor P (TGFP) can induce either positive or negative modulation of the mitogenic response to EGF, to promote either anchorage independence and proliferation (2, 3), or growth arrest and differentiation (4,5), depending on the cell type investigated. TGFB has been shown to modulate rapid responses to EGF, including glucose (6) and amino acid transport (7), phosphatidylinositol (PI) turnover (8), and gene transcription (9,10). The mechanisms by which TGFP exerts these various effects have not yet been elucidated.
Among the earliest detectable events following stimulation by a number of mitogens is the release of inositol phosphates (11-13) and an increase in intracellular levels of free Ca2+ (14). The role of PI turnover in the stimulation of cellular events by EGF is unclear. Stimulation of PI turnover by EGF has not been observed in mouse BALB/c 3T3 cells (15, 16), but this cell line does not exhibit a full mitogenic response to EGF (16). EGF has been shown to increase PI turnover in other cell types, such as hepatocytes (17), and in the A431 human carcinoma cell line (18, 19), in which EGF is inhibitory to proliferation (20). We have shown that EGF stimulates PI turnover in Rat-1 cells (8), a nontransformed fibroblastic cell line which responds to stimulation with a complete round of DNA synthesis in the absence of serum and other polypeptides (21). Therefore, Rat-1 cells provide an excellent system for the study of early events in EGF-stimulated mitogenesis. Although the major action of EGF is thought to be exerted via stimulation of the tyrosine kinase activity of the occupied EGF receptor (22), modulation of inositol phosphate release may also play a role in the total response to EGF. Furthermore, the tyrosine kinase activity of the EGF receptor may facilitate the phosphorylation of PI during the PI cycle (23,24).
Although many current studies have characterized the long term effects of TGFP on cell growth, little is known concerning the early molecular responses to TGFP treatment. We have demonstrated that treatment of serum-deprived Rat-1 cells with TGFP stimulates an increase in cellular IP3 levels which is maintained for several hours (8). This is the first report of alterations of second messenger levels by TGFP; furthermore, the sustained response to TGFP treatment is distinctly different than the transient alterations in IPS levels produced by other growth factors.
Previously, we reported that treatment of serum-deprived Rat-1 cells with TGFP markedly increased the accumulation of IP3 observed following a subsequent exposure to EGF, such that levels approximating those observed following serum stimulation were attained. Under these conditions EGF also stimulated an increase in intracellular [Ca"], as monitored by fura-2 fluorescence. However, the increase in intracellular  (17) found that addition of EGTA had no effect on the initial increase in intracellular Ca2+ seen in response to EGF. Others have postulated that a Ca2+-buffering artifact could explain the inability of EGF to increase intracellular Ca2+ levels in Ca2+-free medium (19).
In the present paper we have examined the relationship between Ca2+ influx and the changes in PI metabolism observed following treatment with either EGF or TGFP, in order to determine whether changes in Ca2+ influx in response to these growth factors are causally related to changes in PI metabolism.

EXPERIMENTAL PROCEDURES
Materials-TGFP was prepared from human blood platelets by the method of Assoian et al. (27) and was then purified to greater than 98% homogeneity by two rounds of reverse-phase HPLC in acetonitrile, 0.1% trifluoroacetic acid, using a Du Pont Zorbax Bio Series Protein Plus analytical column. TGFP eluted at 34% acetonitrile. EGF was prepared from mouse submaxillary glands by the method of Savage and Cohen (28) and was further purified to homogeneity as described by us (29).
Inositol Phosphate Assay-Rat-1 cells grown to confluency on 10cm plates were serum deprived for 36 h in Dulbecco's modified Eagle's medium (DMEM) containing 2 pCi/ml [3H]myo-inositol (American Radiolabeled Chemicals). Following removal of the labeling medium the cells were incubated for an additional 12 h in serum-free DMEM.
For the experiments utilizing Cd2+ the cells were rinsed and incubated in phosphate-free EMEM for 4 h prior to the assay, in order to prevent the formation of cadmium phosphate precipitates. Prior to harvesting, all cells were exposed to 100 mM LiCl for 20 min. The dose-response curves for the effect of Li+ on IP3 accumulation are shown in Fig. 1. Raising the concentration of LiCl in the assay medium had no significant effect on IP3 accumulation in nonstimu-a\ ,TGFB + ECF. Co Cells were preincubated with the indicated concentration of LiCl for 15 min. The preincubation medium was then replaced with EMEM containing 45Ca (10 pCi/ml) and the indicated concentration of LiC1, in the presence of (0) the TGFP vehicle, or (0) 10 ng/ml TGFp plus 100 ng/ml EGF. Each point indicates the mean k S.D., n = 3 (IP3 release) or n = 5 ('%a influx). Error bars have been included only when they exceed the size of the symbols. lated cells but increased the accumulation of IP3 in growth factorstimulated cells in a linear fashion. IP4 accumulation paralleled IP3 (data not shown). All inositol phosphate release experiments were therefore performed at 100 mM LiCl in order to observe maximal changes in the levels of IPS and IP4 stimulated by TGFP and EGF.
Following experimental treatment as indicated, the cells were rinsed twice with ice-cold phosphate-buffered saline and extracted with 1 M formic acid. The formic acid extract was diluted to 0.1 M formic acid, 0.2 M ammonium formate by addition of NH4OH. IP3 and IP4 were then eluted from Dowex-1 ion exchange columns as previously described (8,30). The identity of the eluted fractions was confirmed by HPLC analysis, using a Partasil 10 SAX column (31).
"Ca Influx-Rat-1 cells grown to confluency in 24-well plates were rinsed and incubated in serum-free DMEM overnight. For the experiments utilizing Cd2+, the cells were rinsed and incubated in phosphate-free EMEM for 4 h prior to the assay, in order to prevent the formation of cadmium phosphate precipitates. The assay was initiated by aspirating the DMEM and adding EMEM containing 10 pCi/ml 'Ta (Du Pont-New England Nuclear), in the presence or absence of growth factor. Unless otherwise indicated, the assay medium contained 1 mM unlabeled CaC12. The assay was terminated at the indicated times by aspirating the labeled medium and rinsing the cells 5 X with ice-cold phosphate-buffered saline containing 25 mM MgClz to displace %a bound extracellularly. The cells were lysed in 0.5 M HCl and the 45Ca content of the lysate was determined by liquid scintillation spectrometry. Specific 46Ca influx was determined at each point by subtracting the nonspecific '%a uptake observed in the presence of a 20-fold excess of CaClz. In contrast to the inositol phosphate accumulation assays, the "Ca" influx experiments were performed in the absence of LiCl. Although addition of LiCl reduced both nonstimulated and growth factor-stimulated Caz+ influx, TGFp and EGF had comparable qualitative effects in the presence or absence of Li+ (Fig. 1). 45Ca2' influx experiments were therefore performed in the absence of LiCl in order to observe the maximum level of stimulation.
Bioassay of TGFP Actiuity on Rat-l Celk-The biological responsiveness of Rat-1 cells to TGFO was determined by measuring the acidification of the culture medium in cells exposed to TGFP. Rat-1 cells plated on 35-mm dishes were grown in DMEM plus 10% calf serum, in the presence or absence of TGFP. TGFp concentrations ranged from 0.001 ng/ml to 20 ng/ml, and equivalent volumes of the acetonitrile/trifluoroacetic acid vehicle were added to all plates. After 3 days the medium was transferred to 12 X 75-mm glass tubes, and the pH of duplicate samples was determined.
Measurement of Increases in Cell Number in Response to TGFP-Rat-1 cells were seeded in 35-mm dishes at an initial density of 10' cells/plate, as determined by measurement of cell number using automated cell counter (Coulter Electronics, Hialeah, Fl). Triplicate dishes were harvested by trypsinization, and cell numbers were determined at intervals.

RESULTS
Growth Effects of TGFp on Rat-1 Cells-Prolonged treatment of serum-deprived Rat-1 cells with TGFP resulted in phenotypic changes typical of transformation, including acquisition of a transformed morphology, rapid acidification of the culture medium, and stimulation of growth in confluent monolayers. The effect of TGFP on growth of Rat-1 cell monolayers is shown in Fig. 2. When TGFp was added to exponentially growing Rat-1 cells, there was an initial delay in the rate of proliferation. However, unlike the vehicletreated Rat-1 cells which stopped proliferating and maintained a stable population density for at least 5 days, the TGFP-treated cells continued to proliferate until they eventually detached from the culture dishes.
It was observed that, in addition to producing higher cell densities, TGFp induced the acidification of the culture medium more rapidly than in control plates. Medium acidification has been reported for a large number of transformed cells and is generally considered to be a marker for the transformed phenotype (32). The ability of TGFP to induce acidification of the medium over a wide range of concentrations is shown in Fig. 3. Thus, in contrast to the cytostatic effects of TGFP tially growing Rat-1 cells were exposed to concentrations of TGFp ranging from 0.001-20 ng/ml for 3 days. Acidification of the culture medium from duplicate plates was determined as described under "Experimental Procedures. on some epithelial cell types (4, 5), in Rat-1 cells TGFP acted to induce a transformed phenotype which was characterized by cell overgrowth and medium acidification. EGF and TGFP were each individually capable of producing anchorage-independent growth of Rat-1 cells in soft agar (data not shown). Because of the inherent difficulties in gathering reproducible measurements of anchorage-independent growth in soft agar, determination of medium acidification proved to be a more reliable and quantifiable index of TGFP activity.
Effects of TGFP on Ca2+ Influx-Our previous work has demonstrated that the presence of extracellular Ca2+ is required for EGF-induced changes in intracellular Ca2+ in Rat-1 cells previously exposed to TGFP (8). Ca2+ influx was measured in Rat-1 cells following exposure to EGF and TGFp in order to correlate changes in Ca2+ fluxes with changes in inositol phosphate levels. Treatment of serum deprived, confluent Rat-1 cells with 10 ng/ml TGFP for 4 h increased the rate of 45Ca influx by !&fold, from 2.57 f 0.17 to 6.04 f 0.14 pmol/min/106 cells (Fig. 4). Addition of 100 ng/ml EGF also produced an immediate increase in the rate of Caz+ influx by more than 2-fold (to 8.12 & 0.32 pmol/min/106 cells), while addition of EGF to TGFP-pretreated cells resulted in an additive increase in the rate of 45Ca influx (Fig. 4). Ca2+ influx in response to either growth factor alone, or to the combined factors, was approximately linear over the first 5 min of exposure to 45Ca. In the 45Ca uptake experiments, a concentration of 100 ng/ml EGF was used in order to optimize the initial rate measurements by increasing the rate of association of EGF with the EGF receptor.
The stimulation of Ca2+ influx by TGFP shown in Fig. 4 was measured following a preincubation in TGFp for 4 h, an exposure time previously determined to be maximal for TGFP stimulation of inositol phosphate release (8). We determined the temporal characteristics of the increase in Ca2+ influx seen in response to TGFP, and the effects of actinomycin D on that response (Fig. 5A). Serum-deprived confluent Rat-1 cells were incubated with 10 ng/ml TGFP, or the vehicular control, for the indicated times prior to determination of the rate of 45Ca influx. Incubation of cells with TGFP for periods under 1 h had no effect on Ca2+ influx. A significant increase in the rate of Ca2+ influx was observed by 2 h, and the rate of influx was maximal by 4 h. After 4 h the influx rate remained at a constant level, which was approximately 2-fold higher than control values (Fig. 5A). Coincubation with actinomycin D (10 pg/ml) blocked the TGFP-stimulated increase in 45Ca influx, indicating that an actively transcribed gene product was required for the stimulation of TGFP. Addition of actinomycin D to cells pretreated with TGFP for 4 h resulted in a rapid decline in the rate of Ca2+ influx, with a half-life of approximately 2 h (Fig. 5A). Incubation with actinomycin D, but not the ethanol vehicle, also produced a reduction of 45Ca influx in control cells. In order to demonstrate that the observed effects of actinomycin D were through blockade of a TGFP-induced gene product, rather than a nonspecific effect on the 45Ca influx mechanism, we examined the effects of actinomycin D on EGF-stimulated Ca2+ influx (Fig. 5B). While incubation with actinomycin D for 4 h completely blocked the TGFP-stimulated increase in Ca2+ influx, it had no effect on Ca2+ influx stimulated by EGF.
Correlation of Ca2+ Influx with Production of Inositol Phosphates-We next determined the dose-response characteristics for the TGFP-induced increase in Ca2+ influx and inositol phosphate accumulation (Fig. 6). The concentration of TGFP required to produce half-maximal stimulation of 45Ca influx was 1.5 ng/ml, while the half-maximal dose for inositol phosphate release was 2.0 ng/ml. Both dose-response measure- preincubation with TGFB followed by EGF for 5 min. TGFp was added at 10 ng/ml concentration in acetonitrile/trifluoroacetic acid buffer (1 pl/ml). Control cells received the acetonitrile/trifluoroacetic acid vehicle. EGF was added at 100 ng/ml. Actinomycin D in 50% ethanol was added at 10 pg/ml concentration. All culture media were adjusted to 0.1% ethanol. 45Ca influx was measured for 5 min at each time point as described under "Experimental Procedures." Each point represents the mean f S.D., n = 4. Error bars have been included only when they exceed the size of the symbols. ments were obtained following a 4-h incubation with TGFP.
Maximal 45Ca influx (200% control) was obtained a t 30 ng/ ml TGFP, while maximal inositol phosphate accumulation (350% control) was obtained a t 10 ng/ml TGFP. Accumulation of inositol tetrakisphosphate (IF',) paralleled the increase in IP3 at all points, although the total accumulation in the IP4 peak was lower. There was a good correlation between the concentrations of TGFP which altered second messenger levels and those which induced the acidification of the growth medium (Fig. 3).
Parallel experiments were performed t o determine the doseresponse characteristics for stimulation of 45Ca influx and accumulation of inositol phosphates by EGF (Fig. 7). In contrast to the close similarity between the doses of TGFP required for stimulation of Ca2+ influx and P I turnover (Fig.  6), the concentration dependencies for stimulation of Ca2+ uptake and IP3 release by EGF were not equivalent. The halfmaximal dose for stimulation of 45Ca influx was 15 ng/ml EGF, and the dose response was nearly linear up to the maximal dose tested, 100 ng/ml. The half-maximal dose of EGF for stimulation of IPS was 4 ng/ml, and near maximal  stimulation (approximately 300% control) was observed a t 10 ng/ml EGF. Accumulation of IP4 was also stimulated approximately 2-fold by addition of EGF.
Whether considered in terms of Ca2+ influx or PI metabolism, the dose-response for TGFP stimulation was unaffected by cotreatment with EGF. Similarly, addition of TGFB had no effect on the dose-response characteristics for the stimulation of Ca2+ influx and PI metabolism by EGF (data not shown).
Dependence of Second Messenger Alterations on Extracellular [Ca"]-We next examined the effects of varying extracellular [Ca"] on the EGF-and TGFP-stimulated responses, in order t o determine if the entry of Ca2+ across the plasma membrane was essential for a n increase in inositol phosphate release (Fig. 8). The rate of Ca2+ influx in response to either EGF or TGFP was found to be highly dependent on the concentration of extracellular calcium. Maximal stimulation of the rate of Ca2+ influx in response to either EGF or TGFp was obtained at physiological extracellular Ca2+ concentrations (1-2 mM). Only minimal Ca2+ influxes were observed a t 0.1 mM extracellular Ca2+ (Fig. 8).
A similar dependence on extracellular Ca2+ was found for and Phosphatidylinositol Turnover a " 30  added to the appropriate plates at this time. The assay was initiated by aspirating the incubation medium and adding EMEM containing 10 pCi/ml '5Ca and the indicated concentration of CaC12. The cells were treated as follows: 0, no addition; 0, EGF, 100 ng/ml; A, TGFP 10 ng/ml; A, TGFP 10 ng/ml plus EGF 100 ng/ml. 45Ca uptake was terminated after 3 min and "Ca2+ influx content was determined as described under "Experimental Procedures." Each point represents the mean t S.D., n = 3. Error bars have been included only when they exceed the size of the symbols.  (0,O) was added to the appropriate plates for a 4-h preincubation prior to harvesting. The cells were treated as follows: 0, control cells, no added stimulus; 0, EGF, 10 ng/ml, added for 5 min; A, TGFB 10 ng/ml for 4 h; A, cells were pretreated with TGFP for 4 h, followed by EGF 10 ng/ml for 5 min. IP3 accumulation was determined as described under "Experimental Procedures." Each point represents the average and error from two plates. Error burs have been included only when they exceed the size of the symbols. the production of IP3 in response to EGF and TGFP (Fig. 9). Thus, the increases in Ca2+ influx and in IPS accumulation observed in response to either EGF or TGFp were markedly dependent on external Ca2+ in Rat-1 cells. The reduction in IP3 accumulation observed following lowering of extracellular Ca2+ to 0.1 mM suggests that high levels of Ca2+ influx may be a prerequisite for the growth factor-mediated stimulation of PI metabolism.
In order to characterize the type of calcium channel responsible for the growth factor-mediated increases in cellular Ca2+ permeability, we examined the effects on Ca2+ influx of a number of specific and nonspecific Ca2+ channel antagonists. Inhibitors of voltage-sensitive Ca2+ channels, such as nifedipine or verapamil, neither reduced total Ca2+ influx nor altered the stimulation of Ca2+ influx by EGF or TGFp (Table  I). Parallel experiments measuring inositol phosphate accumulation demonstrated that neither nifedipine nor verapamil altered either the basal or growth factor-stimulated levels of IPJ (data not shown). Although these specific Ca2+ channel blockers effectively inhibited the Ca2+ influx driven by depolarization with 50 mM KC1 (Table I), they were ineffective against the growth factor-stimulated Ca2+ influx.
In contrast to these specific inhibitors of voltage-sensitive Ca2+ channels, two nonspecific Ca2+ channel antagonists, Co2+ and Cd", were highly effective inhibitors of both the Ca2+ influx and inositol phosphate release stimulated by EGF and TGFP (Fig. 10). Inclusion of 3 mM Co2+, which was previously shown to completely block the increase in intracellular free Ca2+ in response to combined EGF and TGFp treatment (8), also reduced the growth factor-stimulated Ca2+ influx to the levels seen in unstimulated cells (Fig. 1OA). IPS accumulation in response to EGF and TGFP was also inhibited by Co2+ (Fig. lOB), and the dose-response characteristics for inhibition of 45Ca influx and IPS accumulation were similar (data not shown). The presence of 1 mM Cd" in phosphate-free medium was also sufficient to completely inhibit the growth

TABLE I Effect of Ca2+ channel blockers on 'Ta" influx
Rat-1 cells grown on 24-well plates were serum-deprived overnight in DMEM. 4 h prior to the assay the cells were treated with TGFp (10 ng/ml) or the acetonitrile/trifluoroacetic acid vehicle. The assay was performed in EMEM containing 10 pCi/ml '5Ca plus 1 mM CaC12, in the presence or absence of stimulus as follows: basal, no added stimulus; TGFP, 4-h preincubation with 10 ng/ml TGFP; EGF, 100 ng/ml added at time 0; TGFP preincubated for 4 h with 10 ng/ml, followed by addition of 100 ng/ml EGF at time 0. High K+ denotes cells treated with EMEM plus 40 mM added KCl, in the presence of 10 pCi/ml '5Ca and 1 mM CaC12.
Nifedipine and verapamil were dissolved in ethanol, then diluted 1:lO in EMEM before addition to cells (0.05% ethanol final concentration). Both nifedipine and verapamil were added to wells for a 5-min preincubation before initiation of the 5 min "Ca uptake assay and were also present in the assay buffer. Each point represents the mean f S.D. for three separate wells.  Serum-derived Rat-1 cells were treated as described below. Cells treated with TGFp were incubated in the presence of 10 ng/ml TGFp for 4 h prior to the determination of either "Ca influx or IP3 accumulation. Those cells were not exposed to TGFp were exposed to the TGFp vehicle alone for the 4-h preincubation period. Each point represents the mean f S.D. for n = 5 (Ca2+ influx) or n = 3 (IP3 accumulation). Ca2+ influx was measured over a 5-min period as described under "Experimental Procedures." The 45Ca-labeling medium also included EGF, TGFB, Co2+ or Cd2+, as detailed below. The effect of Co2+ or Cd2+ on 45Ca2+ influx was determined in cells treated as described below, proceeding from left to right for panels A and C unstirn, no stimulus was added. Control 45Ca influx rates were 2. 3 mM Co2+ or 1 mM Cd2+ for 5 min. The medium was then replaced with fresh EMEM containing 45Ca, EGF (100 ng/ml), TGFO, and either Coz+ or Cd". IP3 accumulation in serum-deprived Rat-1 cells was determined as described under "Experimental Procedures." All cells were treated with 100 mM Li+ for 20 min prior to harvesting. Also added to the culture medium at either 5 or 10 min prior to harvesting were EGF, Co2+, or Cd2+, as detailed below. The effect of Coz+ or Cd2+ on induction of IPS accumulation was determined in cells treated as described below, proceeding from left to right for panels B and D. unstirn factor-mediated stimulation of both Ca2+ influx (Fig. 1OC) and PI turnover (Fig. 1OD). Phosphate-free EMEM was employed to prevent the formation of cadmium-phosphate precipitates. TGFp was significantly less effective in phosphatefree incubation medium than in normal DMEM (1 mM phosphate). Neither Co2+ nor Cd" significantly depressed the rate of 45Ca influx in unstimulated cells. Further evidence for a direct correlation between Ca2+ influx and PI metabolism was obtained when Rat-1 cells were exposed to the Ca2+ ionophore A23187. A 1 pg/ml dose of A23187 was sufficient to increase Ca2+ influx by 70%, and a 120% increase in IP3 accumulation was observed concomitantly (Table 11).

I1
Effect of Ca2' ionophore A23187 on 46Ca influx and IP3 accumulation '5Ca influx and inositol phosphate accumulation were determined as described under "Experimental Procedures." A23187 (1 pg/ml) or the MezSO solvent was added for a 5-min incubation period. Each value represents the mean * S.D. for n = 5 (45Ca influx) or n = 3 (IPa accumulation cells with TGFP resulted in a transient (5 min) elevation in inositol phosphate release, which was markedly increased, both in magnitude and duration, by pretreatment with TGFP. As expected from the large change in cellular IP3 levels, EGF treatment of cells exposed to TGFP produced an increase in cytosolic free Ca2+. The increase in intracellular-free Ca2+ stimulated by EGF and TGFP required the presence of extracellular Ca2+, and could be blocked by either EGTA or Co2+. This result suggested that the influx of extracellular Ca2+ was the initial event which led ultimately to increases in cytosolic free Ca2+ levels as intracellular Ca2+ stores were mobilized.
In this paper we have presented data which indicates that both EGF and TGFP stimulated Ca2+ influx into Rat-1 cells via channels which are not voltage dependent. The increase in the rate of Ca2+ influx seen in response to EGF and TGFP were tightly coupled to inositol phosphate release. The data suggest that the influx of Ca2+ may be the proximal stimulus for increased PI turnover. The evidence supporting this conclusion includes: ( a ) both 46Ca influx and PI turnover displayed similar dose responses and temporal responses to stimulation by either or both growth factors; ( b ) the stimulation of both Ca2+ influx and PI turnover by TGFB and EGF required the presence of extracellular calcium and showed similar Ca2+ concentration dependence; and (c) the use of divalent cations which block Ca2+ channels prevented both the Ca2+ influx and IPS accumulation stimulated by TGFP and EGF.
A wide variety of stimuli have been found to increase PI turnover, including neurotransmitters in the central nervous system (12), hormones and polypeptides in the liver (13), and mitogens effective on a number of cultured cell types (11, 14). The generally accepted mechanism for the release of inositol phosphates is that binding of Ca2+-mobilizing ligands to receptors is coupled to phospholipase C through a G-protein, thereby reducing the Ca2+ requirement for activation of phospholipase C to physiological levels (11, 33,34). Activated phospholipase C mediates the hydrolysis of phosphatidylinositol-4,5-bisphosphate to produce IP3. The resulting increase in IP3 produced by addition of those agents operating by Gprotein-mediated activation of phospholipase C is often very rapid, occurring within seconds. In contrast, the data we have presented here indicate that the PI response to EGF was relatively slow (2-5 min), while the PI response to TGFP took hours to develop, suggesting that these two growth factors triggered PI metabolism through a mechanism not involving rapid activation of phospholipase C.
The data available regarding the stimulation of Ca2+ influx by hormones and peptide mitogens is less comprehensive. Ca2+ influx appears to be essential for the maintenance but not for the initiation of sustained processes such as salivary secretion (35) and pancreatic enzyme secretion (36). Studies performed in nonsecretory cells demonstrate that intracellu-lar Ca2+ in many cell types shows a biphasic response to Ca2+mobilizing growth factors (11, 37). The initial increase does not require extracellular Ca2+ but instead involves IP3-mediated mobilization of intracellular Ca". In contrast, extracellular Ca2+ is required for the delayed and more prolonged secondary phase of Ca2+ elevation. In this schema the hydrolysis of PIPz is the initiating trigger which leads to the mobilization of intracellular Ca2+, and entry of Ca2+ across the plasma membrane is a later process, activated by a poorly defined mechanism. However, in those systems in which PI turnover and Ca2+ influx can be distinctly separated, IP3 release is not always the initiating event. Macara (38) has reported that activation of 45Ca influx in response to EGF is independent of PI turnover in A431 cells, and is therefore not a consequence of PIPz hydrolysis.
Receptor-operated Ca2+ channels have been postulated in several cell systems (39) but have been directly demonstrated in only a few instances. Patch clamp recordings indicate that ATP can directly gate Ca2+-permeable channels in smooth muscle (40). This receptor-activated Ca2+ channel operated independently of second messengers and required only that ATP bind to its receptor. Other mechanisms for growth factor-stimulated increases in Ca2+ permeability have been suggested, involving hydrolysis of PIP, and subsequent opening of plasma membrane Ca2+ channels by either IP3 (41), or IPI (42), or by Ca2+ released from intracellular stores in response to IP3 or IP4 (43). Our data indicate that accumulation of IP3 and IP4 was not responsible for the increase in Ca2+ influx, as LiC1-induced accumulation of IP3 and IP4 was not accompanied by an increase in Ca2+ influx. Since concentrations of Li+ which allow observation of growth factorstimulated IP3 accumulation caused a decrease in Ca2+ influx (Fig. l), the data presented in Figs. 4-10 argue that EGF and TGFP directly modulate Ca2+ influx, independent of changes in PI metabolism. However, these data cannot exclude the possibility that alternate second messenger systems may act to amplify or modulate the growth factor-stimulated changes in Ca2+ influx.
Direct activation of phospholipase C by elevated intracellular [Ca"] has been postulated in other cell systems, based on the observation that extracellular Ca2+ is required for changes in intracellular Ca2+ levels in several cell types (33). These reports have been largely discounted (33), on the basis that use of EGTA to reduce extracellular Ca2+ levels would subsequently lower intracellular [Ca"] below the level at which coupling by G-proteins is effective (34). Similarly, the observation that Ca2+ ionophores can stimulate IP3 release has been attributed to the nonphysiologically high levels of Ca2+ brought into the cell by ionophores. Our data indicate that even a moderate reduction in the concentration of extracellular Ca2+, without the addition of EGTA, is sufficient to markedly reduce the Ca2+ influx and accumulation of IP3 seen in response to stimulation by TGFP and EGF. Similarly, increasing the rate of Ca" influx even moderately with the ionophore A23187 is sufficient to double the release of IPS. These data suggest that phospholipase C is highly sensitive to the rate of Ca2+ influx.
It is clear that phospholipase C is a highly Ca2+-dependent enzyme. Early reports indicated that in vitro activation requires very high, nonphysiological concentrations of Ca2+ (34). Ryu et al. (44) recently purified a PI-specific phospholipase C. This enzyme shows a steep calcium dependence in the physiological range between to 10"j M Ca2+. An attractive hypothesis for growth factor action is that increased CaZ+ influx through the growth factor-operated channels increases the local concentration of Ca2+ sufficiently to increase phos-pholipase C activity. The activation of phospholipase C results

Phurmacol. 35,2447-2453
Biol. Chem. 261,723-727 and TGFP produces an additive increase in Ca2+ influx suggests that EGF and TGFP may interact with distinct populations of Ca2+ channels, or may influence distinct channelactivating mechanisms. The differing sensitivity of the EGF and TGFB effects to actinomycin D supports this conclusion and indicates that TGFP may exert its influence by promoting the transcription of a gene product which either subsequently modulates Ca2+ channel activity or is itself a Ca2+ channel. While this observation may not represent a universal explanation for the cellular mechanisms of TGFP action on regulation of gene transcription, it may play an important role in either the positive or negative modulations of cell function by TGFP in various cell types.