Induction of c-fos Proto-oncogene in Mesangial Cells by Cadmium*

Cadmium is mitogenic under some circumstances and has been shown to cause accumulation of transcripts for several proto-oncogenes in a variety of cells, but the mechanism(s) remain to be delineated. Here we show that CdCl2 causes an increase in c-fos mRNA within 30 min of exposure of mesangial cells. At 10 μm Cd2+, this increase persists for at least 8 h in both rat and human cells. The half-life of c-fos mRNA is the same whether it accumulates following 4 h of treatment with Cd2+ or is induced transiently by phorbol ester. Cycloheximide, which stabilizes the transcript, causes a synergistic increase when administered with CdCl2. Nuclear run-on analysis confirms that Cd2+ causes transcriptional activation of the c-fos gene. Calmodulin and Ca2+/calmodulin-dependent kinase, and classical protein kinase C (PKC) isoforms represent two Ca2+-dependent signaling pathways that can lead to induction of c-fos, and Cd2+ has been shown to activate both calmodulin and PKC in vitro, possibly by virtue of the similar ionic radii of Cd2+ and Ca2+. Therefore, we investigated the effect of Cd2+ on these pathways in vivo. 10 μm CdCl2 did not increase total PKC activity or Ca2+/calmodulin-dependent kinase II activity and inhibited the latter at higher concentrations, ruling out either pathway in the Cd2+-dependent induction of c-fos. However, Cd2+ did lead to a sustained activation of the Erk family mitogen-activated protein kinases (MAPK) that correlated with induction of c-fos. A specific inhibitor of the MAPK kinases, PD98059, partially inhibited the induction of c-fos by Cd2+. We conclude that Cd2+ induces c-fos at least in part by causing a sustained activation of MAPK independent of its ability to activate PKC and calmodulin in vitro.

Cadmium is a carcinogen in animals (1,2) and has been classified as a category 1 carcinogen in humans by the International Agency for Research on Cancer (3). Oral CdCl 2 induces malignant prostatic tumors in rats (2), and myoblasts treated with Cd 2ϩ become transformed and give rise to malignant sarcomas when injected into nude mice (4). At least in part these effects would seem to be a result of Cd 2ϩ -induced DNA damage. Single strand DNA breaks accumulate in cultured testicular cells treated with CdCl 2 (5). An enhanced rate of mutation in response to a number of chemical mutagens and UV light has been found in the presence of cadmium, supporting the view that Cd 2ϩ adversely affects DNA repair (6). Cadmium has been shown to inhibit DNA polymerase ␤, decrease transcription of O 6 -methylguanine-DNA methyl transferase, increase the UV mutation rate in competent but not repair-deficient xeroderma pigmentosum cells, and decrease the removal of pyrimidine dimers (for review, see Ref. 6).
When quiescent cells enter the cell cycle, expression of immediate early response genes is necessary for progression through G 1 and subsequent proliferation (7). Many of these are protooncogenes that encode nuclear transcription factors and determine subsequent expression of other genes. For example, Fos and Jun protein heterodimers constitute the AP-1 transcription factor that leads to transcriptional activation of many genes (8). Cadmium causes accumulation of transcripts of c-jun and c-myc in myoblasts (9); c-fos and egr-1 in fibroblasts (10); c-fos, c-jun, and c-myc in normal rat kidney fibroblasts (11); and c-fos, c-jun, c-myc, and egr-1 in LLC-PK1 proximal tubular cells (12). Thus, Cd 2ϩ has the potential to act as a mitogenic stimulus in some cells. Whereas DNA damage is associated with tumor initiation, effects on oncogene expression and mitogenesis are more likely to be associated with promotion and tumor progression (13).
On the other hand, Cd 2ϩ can induce apoptosis in isolated T lymphocytes (14) and cultured LLC-PK1 cells (12) and lead to apoptotic cell damage in canine proximal tubules (15) and rat testicular tissue (16). Genetic damage triggers apoptosis by a p53-mediated pathway, and a current concept of the action of p53 is that it arrests the cell cycle in late G 1 in response to damage, while DNA repair occurs (17,18). If repair is unsuccessful within a certain period of time, apoptosis will be initiated. Thus, Cd 2ϩ may initiate apoptosis by causing DNA damage and inhibiting DNA repair. Alternatively, mitogenic stimulation by Cd 2ϩ against a background of genetic damage may circumvent apoptosis and lead to escape and proliferation of cells destined for transformation. Therefore, Cd 2ϩ may be an interesting agent to explore the factors determining the balance between mitogenesis and apoptosis.
Cadmium-related effects on oncogene expression may play a critical role in deciding cell transformation or removal. For example, c-fos expression is part of a mitogenic response that is required for cell proliferation (8). On the other hand, numerous authors have described an association between c-fos expression and apoptosis (19), and overexpression of c-fos in transfected fibroblasts increased the apoptotic response to serum deprivation 10-fold (20). The outcome may in part relate to the time and amplitude of expression and sustained levels of c-fos mRNA appear to favor apoptosis over proliferation. The present study was undertaken to examine the response of c-fos to Cd 2ϩ in smooth muscle-like mesangial cells, and the role of two major Ca 2ϩ -dependent signal transduction pathways involved in c-fos induction.
[␣-32 P]dCTP (specific activity 3000 Ci/mmol) was from NEN Life Science Products, [␥-32 P]ATP (specific activity 4500 Ci/mmol) was from ICN (Costa Mesa, CA), and [␣-32 P]UTP (Ͼ3000 Ci/mmol) was from Amersham. A random primer DNA labeling kit was obtained from Boehringer Mannheim. TRIzol reagent for RNA isolation and yeast tRNA were from Life Technologies, Inc. RNAguard and RNase-free DNase 1 were products of Pharmacia Biotech Inc. A rat c-fos cDNA was obtained from T. Cruz (University of Toronto), and a cDNA for mouse 18 S rRNA was a gift from J. Koropatnick (University of Western Ontario). Affinity purified polyclonal rabbit anti-rat Erk-2 antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). Protein A-Sepharose was from Pharmacia. Phosphocellulose paper circles (P81, 2 cm) were from Whatman, and Hybond-N nylon membrane was from Amersham.
Cell Culture-Rat mesangial cells (RMC) were cultured from outgrowths of glomeruli harvested by sieving from kidneys of male Wistar rats, and characterized as described previously (21,22). They were maintained in 10-cm Petri dishes in a 5% CO 2 environment at 37°C, in RPMI 1640 medium with penicillin, streptomycin, and 10% FBS, passaged by trypsinization, and used between the 5th and 15th passages for all experiments reported here. Quiescence was induced by exchanging medium on cells at 60 -80% confluence (approximately 5 ϫ 10 4 cells/cm 2 ) with medium containing 0.4% FBS followed by 48 h of incubation. Human mesangial cells (HMC) were cultured from glomeruli obtained from the uninvolved part of a kidney resected for renal cell carcinoma, maintained in an identical manner to the RMC cultures, and used between passages 5 and 10.
RNA Isolation and Northern Blotting-Total RNA was isolated using TRIzol reagent, as described by Chomczynski and Mackey (23). Equal amounts of RNA (ca. 10 g) were denatured by the method of Gong (24), separated by electrophoresis on agarose-formaldehyde gels, and transferred to Hybond-N nylon membrane for hybridization with a c-fos cDNA that was labeled with [␣-32 P]dCTP by the random primer method. Levels of mRNA were quantitated by densitometry of the Northern blot autoradiographs and normalized to 18 S rRNA after probing with labeled cDNA to rat 18 S rRNA.
Nuclear Run-on Assay-Nuclei were isolated from RMC by the method of Celano et al. (25). Seven 10-cm Petri dishes of cells, per experimental point, were lysed in buffer A (20 mM Tris-HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl 2 , 0.1% Nonidet P-40) and homogenized in a sterile Dounce homogenizer. Nuclei were pelleted at 1000 ϫ g for 10 min. The nuclear pellet was resuspended in 20 l of buffer B (36.5% glycerol, 12 mM Tris/HEPES, pH 7.0, 4 mM dithiothreitol, 50 mM (NH 4 )SO 2 , 10 mM MgCl 2 , 6 mM MnCl 2 , 10 mM NaF, 18 M creatine phosphate, 32 g/ml creatine phosphokinase, 300 mM KCl, and 0.1 mM EDTA). To start nuclear transcription, 1 mM each of ATP, GTP, and CTP, 100 units of RNAguard, and 200 Ci of [␣-32 P]UTP were added to the nuclear suspension. The mixture was incubated at 30°C for 30 min. The reaction was terminated by digestion with 100 units of RNase-free DNase 1 and 10% (v/v) 10 mM CaCl 2 at 37°C for 15 min. The sample was subsequently treated with 10% (v/v) 10 ϫ SET buffer (5% SDS, 50 mM EDTA, 100 mM Tris-HCl, pH 7.4) and 10% (v/v) 0.4 mg/ml proteinase K. Carrier yeast tRNA (100 g) was added, and the sample was incubated at 37°C for 30 min. After the incubation, 1 ml of TRIzol was added for RNA isolation according to the manufacturer's protocol, except that three additional cycles of ammonium acetate/ethanol precipitation were performed to remove free [ 32 P]UTP. Purified 32 P-labeled RNA was hybridized to Hybond nylon membrane with immobilized slots containing 1 g of c-fos cDNA and 18 S RNA cDNA inserts. The hybridization was carried out at 42°C for 3 days and washed at high stringency (2 ϫ SSC ϩ 0.1% SDS at 65°C for 1 h; 2 ϫ SSC ϩ 0.5% SDS at 22°C for 1 h) prior to exposure to x-ray film.
PKC Assay-Cells were plated in 96-well tissue culture plates at 2 ϫ 10 4 cells per well, and after 16 h the FBS content of the medium was decreased to 0.4% for 48 h. The cells were then treated with fresh medium with or without CdCl 2 for various times, and incubated for 10 min at 30°C in 40 l of reaction solution (137 mM NaCl, 5.4 mM KCl, 10 mM MgCl2, 0.3 mM Na 2 HPO 4 , 0.4 mM KH 2 PO 4 , 25 mM ␤-glycerophosphate, 5.5 mM D-glucose, 5 mM EGTA, 1 mM CaCl 2 , 20 mM HEPES, 100 M ATP, 50 g/ml digitonin, pH 7.2) containing 120 g/ml of the peptide VRKRTLRRL and 10 Ci/ml [␥-32 P]ATP, as described (26). Each reaction was terminated by adding 20 l of ice cold 25% (w/v) trichloroacetic acid on ice and spotted onto phosphocellulose circles prewashed sequentially with water, buffer, and 75 mM phosphoric acid. After standing for 15 min at room temperature, the circles were washed (4 min each with gentle shaking) three times in 75 mM phosphoric acid and once in 2.75 mM sodium phosphate, pH 7.5, before liquid scintillation counting.
MAPK Assay-Quiescent cells in 10-cm Petri dishes were treated with CdCl 2 or TPA for various times before being scraped into 800 L of lysis buffer (50 mM Tris-HCl, pH 7.4, 1% (v/v) Nonidet P-40, 0.25% (w/v) sodium deoxycholate, 150 mM NaCl, 5 mM EGTA, protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 g/ml aprotinin, 1 g/ml leupeptin), and phosphatase inhibitors (1 mM Na 3 VO 4 , 1 mM NaF)). Cells were sonicated for 5 s and centrifuged at 100,000 ϫ g for 15 min. Aliquots of cytosol containing 500 g of protein were precleared by adding 1 g of normal rabbit IgG with 20 L of protein A-Sepharose and centrifuging in a microcentrifuge for 5 min. The supernatant was incubated with 2 g of anti-Erk-2 antibody (3 h, 4°C) and immunoprecipitates were recovered by incubation (2 h) with a 50% slurry of protein A-Sepharose. MAPK (mitogen-activated protein kinase) activity in the immunoprecipitates was determined by phosphorylation of myelin basic protein. Immunoprecipitates were mixed with 20 mM HEPES buffer, pH 7.4, containing 10 mM MgCl 2 , 2 mM MnCl 2 , 0.5 mM EGTA, 10 mM NaF, 0.5 mM Na 3 VO 4 , 1 mM dithiothreitol, 0.5 mg/ml myelin basic protein, 2 M AMP-dependent protein kinase inhibitor, 50 M ATP, and 5 Ci of [␥-32 P]ATP. After incubation at 30°C for 30 min, reaction mixtures were added to an equal volume of 2 ϫ sample buffer and subjected to 16% SDS-polyacrylamide gel electrophoresis, according to Laemmli (27). Incorporation of 32 P into myelin basic protein was determined by autoradiography and densitometry.
CaMK II Assay-Cells in 10-cm Petri dishes were treated with CdCl 2 for various times and lysed by several freeze-thaw cycles in buffer containing 50 mM HEPES, 50 mM sodium pyrophosphate, 1 mM EGTA, 25 mM NaF, 1 mM Na 3 VO 4 , 1 mM dithiothreitol, 0.5% (v/v) Nonidet P-40, 1 g/ml aprotinin, 1 g/ml leupeptin, 1 g/ml pepstatin, and 1 mM phenylmethylsulfonyl fluoride. After brief sonication (3 ϫ 5 s) and centrifugation (10 min at 17,000 ϫ g), the protein content of the supernatant was determined by the method of Peterson (28). Fractions of the supernatant containing 5 g of protein were incubated with 10 volumes of Ca 2ϩ /calmodulin-dependent protein kinase (CaMK) assay buffer at 30°C for 3 min. The assay buffer for autonomous activity was 50 mM HEPES, pH 7.5, containing 10 mM MgCl2, 0.1 mM ATP, 10 M autocamtide-2, 1 mM EGTA, and [␥-32 P]ATP (5 Ci/ml). For total CaMK II activity, EGTA was replaced with 3 mM CaCl 2 and 1 mM calmodulin. The reaction was stopped by adding trichloroacetic acid to 5% (w/v) and after centrifugation the supernatant was spotted onto prewashed phosphocellulose filter paper circles for scintillation counting after extensive washing with 75 mM H 3 PO 4 .

RESULTS
When RMC are serum-starved (0.4% FBS) for 48 h they become quiescent, reverting to G o , and then when stimulated to re-enter the cell cycle with serum or phorbol ester they begin to express c-fos mRNA within 15 to 30 min (29). To test if Cd 2ϩ could mimic this mitogenic response, quiescent cells were treated with CdCl 2 at 1 or 10 M. Both concentrations caused an appearance of c-fos mRNA in 30 min (Fig. 1). The higher concentration caused a sustained increase that plateaued at 2 h and remained elevated for at least 8 h. HMC showed similar behavior to RMC with respect to accumulation of c-fos mRNA in response to 10 M Cd 2ϩ . The higher concentration of Cd 2ϩ may initiate additional mechanisms affecting c-fos mRNA, or a larger initial response to the higher concentration could trigger a self-sustaining response. Alternatively, 10 M Cd 2ϩ may cause inhibition of mechanisms (e.g. phosphatase activity) normally limiting activation of c-fos induction. In keeping with the latter view, TPA (50 ng/ml), an activator of PKC and potent inducer of c-fos in these cells, caused a transient appearance of c-fos mRNA that peaked at 30 -60 min and rapidly declined thereafter (Fig. 1).
Because c-fos mRNA is rapidly degraded, changes in its stability can rapidly affect its concentration. Therefore, we considered whether some nonspecific effect of Cd 2ϩ such as inhibition of protein synthesis might be affecting mRNA at this level, as was observed for example, with Zn 2ϩ in fibroblasts (30). 10 M Zn 2ϩ did not cause an accumulation of c-fos mRNA above control levels in RMC (Fig. 2). Cycloheximide, an efficient inhibitor of protein synthesis, increased c-fos mRNA to a level comparable to Cd 2ϩ at 4 h. However, Cd 2ϩ and cycloheximide together had a strongly synergistic effect, increasing the relative amount of c-fos mRNA about 5-fold above either treatment alone in both RMC and HMC (Fig. 2). This suggested that Cd 2ϩ might be inducing c-fos transcription, whereas cycloheximide stabilized the newly synthesized message.
To further address the idea that Cd 2ϩ and cycloheximide affected c-fos mRNA by different mechanisms, the rate of decay of the message was measured in actinomycin D-treated cells (Fig. 3). Subconfluent RMC were first treated with Cd 2ϩ for 4 h or TPA for 30 min, and then RNA synthesis was stopped with actinomycin D (5 g/ml). Whether c-fos mRNA was induced by TPA or allowed to accumulate in the presence of Cd 2ϩ , it was degraded at a comparable rate. However, cycloheximide prevented any detectable degradation over 90 min (Fig. 3). Because c-fos mRNA turnover is fully blocked by cycloheximide, the large synergistic effect of Cd 2ϩ plus cycloheximide on the mRNA level cannot be due to further inhibition of degradation by Cd 2ϩ but must result from induction by Cd 2ϩ together with prevention of turnover by cycloheximide. To demonstrate transcriptional activation of c-fos by Cd 2ϩ , a nuclear run-on assay was performed. In four separate experiments, labeled transcripts hybridizing to c-fos cDNA in nuclear extracts of cells treated with 10 M CdCl 2 for 4 h exceeded hybridization from extracts of control cells by 1.8 -4.9-fold (mean 3.4-fold) when corrected for hybridization to 18 S cDNA. A typical slot blot is shown in Fig. 4.
We have previously shown that heparin inhibits induction of c-fos in RMC through at least two major pathways, one operating through activation of MAPK and another independent of MAPK and dependent on intracellular Ca 2ϩ (29). Therefore, to determine whether these signaling pathways were involved in the response to Cd 2ϩ , we determined whether heparin affected c-fos mRNA levels in Cd 2ϩ -treated cells. In both RMC and HMC (Fig. 2), heparin (1 g/ml) caused a diminution in the increase in c-fos mRNA observed with Cd 2ϩ treatment. This was not due to inhibition of Cd 2ϩ uptake. The rate of 109 Cd 2ϩ uptake by RMC was unaffected by heparin at concentrations used in these experiments (data not shown), as predicted from polyelectrolyte binding theory in the presence of millimolar Ca 2ϩ and Mg 2ϩ .
Several kinase inhibitors were used to examine the possible involvement of kinase cascades in the response to Cd 2ϩ at 4 h. None of the inhibitors alone caused a significant increase in c-fos mRNA above basal levels (data not shown). Genistein, an inhibitor of tyrosine kinases, caused a significant decrease (Table I). Genistein had no effect on the uptake of Cd 2ϩ by the FIG. 2. Effects of multiple agents on c-fos mRNA levels. Quiescent mesangial cells were treated for 4 h with 10 M CdCl 2 in serumfree medium with heparin (Hep, 1 g/ml), cycloheximide (Cyc, 50 g/ ml), or ZnCl 2 (10 M), or with medium alone as indicated. Autoradiograms of Northern blots probed for c-fos mRNA are shown. The experiment was performed once with human cells (HMC) and three times with rat cells (RMC). The histogram shows the signal intensities for RMC (mean Ϯ S.D.), corrected for loading by probing for 18 S rRNA, and expressed relative to the value for Cd 2ϩ treatment alone taken as 100%.
FIG. 1. Increased c-fos mRNA in Cd 2؉ -treated mesangial cells. Quiescent rat (RMC) or human (HMC) mesangial cells were treated with 1 or 10 M CdCl 2 in serum-free medium for the indicated periods of time before isolation of total RNA for Northern blotting with a cDNA encoding c-fos. Blots were also probed for 18 S rRNA to correct for loading. Alternatively, cells were treated with phorbol ester (TPA, 50 ng/ml). a, Autoradiograms of the c-fos blots; b, histograms of the intensity ratios of c-fos to 18 S signals determined by laser densitometry for RMC treated with 1 (Ⅺ) or 10 M (f) CdCl 2 ; c, histograms of c-fos mRNA:18 S rRNA ratios for RMC treated with TPA (Ⅺ) and HMC treated with 10 M CdCl 2 (f). Each blot is from a single experiment. The results with CdCl 2 are implicitly repeated under several different protocols in subsequent figures and Table I. cells (data not shown). At concentrations that inhibited ionomycin-dependent CaMK-mediated induction of c-fos but did not affect induction in response to serum stimulation, 2 neither the general CaMK inhibitor, KT 5926, nor the CaMK II-specific inhibitor, KN-93, caused any significant diminution in the response of c-fos mRNA to Cd 2ϩ . The PKC inhibitor, calphostin C, was also without effect. However, staurosporine, a relatively nonspecific inhibitor of several kinases including PKC (32) and CaMK (33), decreased c-fos induction by about 70%. Therefore, to confirm that neither PKC nor CaMK were involved in the response to Cd 2ϩ , we measured the activity of PKC and CaMK directly.
When calmodulin is activated, one consequence is autophos-phorylation of CaMK II, with the development of autonomous (i.e. calmodulin-independent) CaMK II activity. Transfer of synchronized RMC to serum-free medium with or without 10 M Cd 2ϩ resulted in an apparent early increase in autonomous CaMK II activity within about 15 min that nevertheless failed to reach significance (data not shown). Thereafter, autonomous activity declined with time, irrespective of the presence of Cd 2ϩ , suggesting an effect of serum deprivation rather than Cd 2ϩ on basal CaMK II activity. Furthermore, 10 M Cd 2ϩ added to the assay reaction mixture in vitro had no effect on either autonomous or total CaMK II activity (data not shown). This decrease was paralleled by a decrease in total CaMK II activity, measured by activation with the excess Ca 2ϩ and calmodulin in vitro, suggesting decreased synthesis or increased degradation of CaMK II in serum-free conditions and indicating that activation of CaMK II by Cd 2ϩ is not the mechanism responsible for induction of c-fos by Cd 2ϩ . To further confirm this, the concentration dependence of autonomous CaMK II activity after 1 h treatment with Cd 2ϩ was measured (Fig. 5). After 1 h, when serum deprivation was beginning to cause a decrease in both autonomous and total CaMK II activity, Cd 2ϩ exposure caused a dose-dependent decrease in autonomous activity that became significant at 5 M. Under these conditions, Cd 2ϩ had no significant effect on total CaMK II activity, up to the highest concentration tested (20 M). When RMC growing in 10% FBS were transferred to serumfree medium, there was an initial decrease in total PKC activity, followed by a rebound to a variable and somewhat higher than control value over several hours. This pattern was unaffected by the presence of Cd 2ϩ (data not shown). The increase was insufficient to induce c-fos and does not account for the induction seen with Cd 2ϩ , for example in Fig. 1, because no increase in c-fos mRNA was detected on Northern blots from serum-free controls. To further rule out activation of PKC by Cd 2ϩ , cells were treated for 4 h with between 5 and 30 M CdCl 2 in serum-free conditions. PKC activity was not increased above the serum-free control. Furthermore, a similar treatment with Cd 2ϩ failed to augment the increase in PKC activity caused by treatment with TPA (50 ng/ml) during the last 30 min of the 4 h treatment period or to overcome inhibition of PKC by staurosporine (data not shown). In keeping with the failure of calphostin C to inhibit the induction of c-fos by Cd 2ϩ , PKC appears not to be involved in the process.
Although Cd 2ϩ did not activate PKC, it might activate 2   MAPK via other pathways. Therefore, we measured MAPK activity in RMC at different times after treating quiescent cells with 10 M Cd 2ϩ . Cadmium caused an increase in immunoprecipitable MAPK activity that peaked at 15 min and then declined but remained above control levels for several hours (Fig.  6). To determine whether a sustained increase in MAPK was responsible for the sustained increase in c-fos mRNA, we used the MEK-1/2-specific inhibitor PD98059. This flavone prevents the activation of MAPK by inhibiting the MAPK kinases, MEK-1 and MEK-2 (34). At 25 M, PD98059 caused a marked decrease in the level of c-fos mRNA induced by a 4 h treatment with CdCl 2 (Fig. 7). However, it did not prevent the induction completely, nor did higher concentrations show any greater effect.

DISCUSSION
The c-fos gene is well characterized, and its transcription is regulated by two DNA regulatory elements, the cAMP-response element and the serum-response element (SRE) (35,36). MAPK can be activated by PKC through Raf-1 and MEK (37), and it then phosphorylates the Ets-like transcription factor, Elk, that forms part of the transcription factor complex that binds to SRE (36, 38 -40). CaMK II phosphorylates another factor in this complex, the serum response factor (41,42), so activation of either PKC or CaMK II can induce c-fos through SRE. CaMK II can also phosphorylate the cAMP-response element-binding protein, CREB, and phosphorylation of CREB is itself sufficient to induce c-fos (43). It has recently been shown that c-fos transcription through the SRE can be activated by an increase in cytoplasmic Ca 2ϩ , whereas CREB is dependent upon intranuclear Ca 2ϩ (44). In view of earlier reports that Cd 2ϩ can activate both PKC and calmodulin (see below), we considered both as possible candidates for mediating induction of c-fos by Cd 2ϩ .
The ionic radii of Cd 2ϩ and Ca 2ϩ are very similar (0.97 and 0.99 Å, respectively (45)), and Cd 2ϩ can substitute for Ca 2ϩ in a number of Ca 2ϩ -binding proteins including calmodulin and Ca 2ϩ -dependent PKC isoforms. Calmodulin is a highly conserved and widely distributed peptide containing four Ca 2ϩbinding loops. Occupancy of these sites activates the Ca 2ϩcalmodulin complex to interact with a number of other proteins including CaMK II and myosin light chain kinase and in turn affect their activity (46). In vitro Cd 2ϩ can induce conformational changes in calmodulin and activate calmodulin-dependent phosphodiesterase (47)(48)(49), a property it shares with other metal ions with similar ionic radii (47). Activation occurs even in the presence of physiological concentrations of glutathione (50) but requires a minimum concentration of Cd 2ϩ of approximately 10 M to become significant (47,48). It has been postulated that increased activity of myosin light chain kinase contributes to the vascular toxicity of cadmium (51), although even higher Cd 2ϩ concentrations are required to activate myosin light chain kinase, another Ca 2ϩ -calmodulin partner, in vitro (51). Intracellular Cd 2ϩ concentrations in cadmium-exposed cells are in the range of 1 pM (52, 53) with higher amounts of Cd 2ϩ inducing and becoming sequestered by metallothioneins. Therefore, it is unlikely that Cd 2ϩ would directly activate calmodulin in vivo, although prevention of Cd 2ϩ -induced testicular necrosis in mice by a series of calmodulin inhibitors (54) suggests a role of altered calmodulin activity in Cd 2ϩ toxicity perhaps secondary to changes in Ca 2ϩ metabolism (55). One consequence of calmodulin activation is autophosphorylation of CaMK II with the development of autonomous CaMK II activity (56). The failure of Cd 2ϩ treatment to activate autonomous CaMK II activity in this study is consistent with insufficient intracellular Cd 2ϩ concentrations being attainable to activate calmodulin in cultured RMC. The "classical" isoforms (␣, ␤I, ␤II, ␥) of PKC are Ca 2ϩ -dependent (57). In in vitro assays with unfractionated PKC, Cd 2ϩ stimulated PKC activity at concentrations up to 100 M in low Ca 2ϩ medium and was inhibitory only at higher concentrations (51). Cd 2ϩ also enhanced the interaction of PKC with isolated rat liver nuclei (58) and increased histone S-III phosphorylation significantly at about 30 pM when added to an in vitro assay mixture (59). Therefore, in the intracellular environment, any effect of Cd 2ϩ on PKC should be stimulatory. Induction of c-fos in Cd 2ϩ -treated proximal tubule cells was partially inhibited by the PKC inhibitor, H-7, and those authors (12) attributed this induction to activation of PKC secondary to a Cd 2ϩ -mediated increase in [Ca 2ϩ ] i rather than to a direct effect of Cd 2ϩ . Although they did not measure [Ca 2ϩ ] i , Cd 2ϩ binding to an orphan receptor has been reported to cause mobilization of intracellular Ca 2ϩ stores in cultured fibroblasts (60). On the other hand, Epner and Herschman (10) showed that PKC down-regulation with phorbol ester pretreatment did not attenuate induction of several phorbol ester-inducible oncogenes, which include c-fos, and they concluded that PKC signaling was not involved in the mechanism. In RMC under our experimental conditions, 8 h of exposure to 10 M CdCl 2 caused an increase in [Ca 2ϩ ] i from 137 Ϯ 25 nM to only 259 Ϯ 59 nM (53), which is insufficient to induce c-fos. Nor did the present study show any effect of Cd 2ϩ on total PKC activity above that caused by a change of medium alone. Together with the failure of the PKC inhibitor, calphostin C, to inhibit c-fos induction, we consider these data to rule out either primary or secondary activation of PKC as underlying c-fos induction by Cd 2ϩ in RMC.
There is no general consensus of opinion on the mechanism(s) involved in the induction of proto-oncogenes by Cd 2ϩ . It has been suggested that metal regulatory elements like those that bind the Zn 2ϩ -dependent metal transcription factor, MTF-1, required for metallothionein induction, may be involved in the induction of early response phorbol ester-inducible oncogenes (10), but none have been identified in these genes to date. In fact, only metallothionein (61) and metalresponsive heme oxygenase (62) genes have been found to contain consensus metal regulatory element sequences, and to our knowledge those of heme oxygenase have not been shown to be functional. Zinc can cause accumulation of c-fos mRNA in fibroblasts by increasing mRNA stability secondary to a nonspecific effect on protein synthesis (30). Cadmium, under the conditions of our experiments, causes a depression in the rate of protein synthesis by RMC (53). This does not account directly for the transcriptional activation observed in the present study, but indirect effects on synthesis of inhibitory trans-acting factors cannot be ruled out. The c-fos gene contains a fos intragenic regulatory element (FIRE) in the first exon, to which binding of a repressive factor leads to transcriptional block. Removal of the factor rapidly allows continuation of transcription (63). The SRE itself may bind a labile protein factor that exerts negative regulation (64). Indeed, the "superinduction" of c-fos by a number of factors observed in the presence of cycloheximide has been attributed to the rapid degradation of such factors without replacement. However, cycloheximide clearly exerts an important effect on c-fos mRNA stability in this study. The strong synergistic effect of Cd 2ϩ and cycloheximide on c-fos mRNA levels argues against a significant contribution to the phenomenon from Cd 2ϩ -mediated suppression of protein synthesis but rather suggests an independent mechanism mediated through a signal transduction pathway. Whereas sustained elevation of MAPK activity appears to be a major contributor to induction of c-fos by Cd 2ϩ , the incomplete inhibition of induction by PD98059 and the partial inhibition by genistein and staurosporine indicate that additional pathways (mecha-nisms) also operate.
A notable result of this study is the sustained time of elevation of MAPK activity and c-fos expression in Cd 2ϩ -exposed cells. MAPK activation by serum is accompanied by measurable c-fos mRNA levels at 15 min that peak at 0.5-1 h and decline to nearly undetectable levels by 2 h (22). In keeping with a sustained increase in MAPK activity, 10 M Cd 2ϩ causes a prolonged increase in c-fos mRNA. At lower concentrations of Cd 2ϩ (1 M) (Fig. 1), the increase in c-fos mRNA appears more transient. Whereas MAPK activity is maximal at 1-2 min after stimulation of quiescent RMC with serum or TPA and declines thereafter (29), the increase in response to Cd 2ϩ reaches a maximum level sometime after 5 min but remains at 60% of the value at 15 min for at least a further 8 h (Fig. 6). This apparent delay in activation relative to serum or TPA likely reflects the rate of entry of Cd 2ϩ into the cells. We have previously found that exposure of RMC to 10 M CdCl 2 produces cytosolic concentrations of ionic Cd 2ϩ of about 1 pM (53), consistent with values reported in other cells (52). Although intracellular cadmium continues to accumulate (31), induction of metallothionein and its sequestration of Cd(II) prevents further increases in cytosolic [Cd 2ϩ ]. The early activation of MAPK and c-fos transcription suggests that this ionic pool is responsible. Persistence of these processes is consistent with kinetically inert Cd 2ϩ binding to an as yet unidentified effector target.