JOURNAL OF CELLULAR PHYSIOLOGY 150:559-567 (1992)
Epidermal Growth Factor Stimulates cAMP Accumulation in Cultured Rat Cardiac Myocytes Y l M l N G YU, BlPlN C. NAIR, AND TARUN 6.PATEL* Department of Pharmacology, The Health Science Center, University of Tennessee, Memphis, Tennessee 38 163 We have previously shown that epidermal growth factor (EGF) augments CAMP accumulation in the heart and stimulates cardiac adenylyl cyclase via a G protein mediated mechanism (Nair et al., 1989). More recently, employing an antibody against the carboxy-terminus decapeptide of G,, we have demonstrated that G,, mediates the effects of EGF on cardiac adenylyl cyclase (Nair et al., 1990). Since the heart comprises of a variety of cell types, the purpose of the studies presented here was to determine whether or not the effects of EGF on adenylyl cyclase were mediated in cardiac myocytes or noncardiomyocytes. Therefore, cultures of ventricular cardiomyocytes and noncardiomyocytes from neonatal rat hearts were established and characterized. Apart from the differences in cellular morphology, cardiomyocytes but not the noncardiomyocytes employed in our studies expressed the a-and P-myosin heavy chain (MHC) mRNA and the P-MHC protein. Additionally, as described previously, treatment of cardiomyocytes with thyroid hormone increased a-MHC mRNA and decreased the expression of P-MHC mRNA, indicating that the cardiomyocytes employed in our studies were responding in a physiologically relevant manner. EGF in a time-dependent manner increased cAMP accumulation in the cardiomyocytes but not in noncardiomyocytes. Maximum and half-maximum effects were observed at 100 n M and 2 n M concentrations of EGF, respectively. As determined by the presence of immunoreactive EGF receptors and tyrosine phosphorylation of the 170 kDa protein in membranes of cardiomyocytes and noncardiomyocytes, both the cell populations contained functional ECF receptors. Therefore, the differential effects of EGF on cAMP accumulation in the two cell populations appear to be due to differential coupling of the EGF receptors to the adenylyl cyclase system rather than the absence of EGF receptors in noncardiomyocytes. Consistent with our previous findings in isolated membranes and perfused rat hearts, EGF-elicited increase in cAMP accumulation in cardiomyocytes did not involve activation of P-adrenoreceptors and was abolished by prior treatment of cells with cholera toxin. Overall, our findings demonstrate that EGF-elicited increase in cAMP accumulation in the heart is the reflection of changes in cAMP content of cardiomyocytes and not noncardiomyocytes.
Epidermal growth factor (EGF), a 53 amino acid polypeptide elicits a wide spectrum of biological actions in a number of tissues and cell types (see Carpenter 1979, 1987 for reviews). Rabkin et al. (1987) have described a novel effect of EGF in cultured chick embryo cardiac myocytes. These authors demonstrated that EGF produces a chronotropic effect in cultured ventricular myocytes from chick embryos. Maximal increase in the beating rate of ventricular myocytes was observed 15 min after the addition of EGF and the response was sustained over a period of 1.5 h (Rabkin et al., 1987). Because cAMP is well established to produce chronotropic effects in the heart (see, e.g., Hofmann et al., 1987) and since the mechanism(s) by which EGF stimulated the beating of cardiac myocytes remained unknown, we proposed the hypothesis that EGF increased the accumulation of cAMP in the heart. To date, our studies 0 1992 WILEY-LISS, INC.
have shown that EGF increases cAMP accumulation in the perfused rat heart and this increase in cellular accumulation of cAMP is the result of stimulation of the cardiac adenylyl cyclase activity (Nair et al., 1989). The stimulation of cardiac adenylyl cyclase and increase in tissue cAMP content were independent of stimulation of 6-adrenoreceptors or activation of protein kinase C (Nair e t al., 1989). However, like P-adrenoreceptor-mediated activation of adenylyl cyclase, EGF stimulates cardiac adenylyl cyclase via a G protein (Nair et al., 1989). More recently, employing a n antibody (CS1) directed against the carboxy-terminus decapeptide
Received August 22,1991; accepted October 3, 1991.
*To whom reprint requestsicorrespondence should be addressed.
560
YU ET AL.
pernatant through a 23-gauge needle into a syringe and filtering through a Nitex (mesh 50) sieve. The density of cells in the filtrate was adjusted to 2 x lo6 cells/ml of CMRL 1066 medium containing 10% serum components. Four million cells (2 ml) were plated in 35 mm dishes (Falcon 3001) and placed in a 37°C humidified incubator equilibrated with 95% air/5% CO,; 18 h after plating the cells, the medium was changed to remove unattached cells. Under these conditions, cardiomyocytes, which represented > 85% of the cell population, beat in a synchronous fashion 48 h after plating. For the purposes of culturing nonmyocyte cells, after removing the myocyte-enriched supernatant from the plates in the preattachment step, 2 ml of CMRL 1066 medium containing 5% horse serum and 5% fetal calf serum were added to each of the 35 mm dishes. Under these conditions, nonmyocytes divided and were confluMATERIALS AND METHODS ent 3 days after plating. To ensure minimum contamiCultures of neonatal rat myocardial and nation of noncardiomyocytes by cardiomyocytes, the nonmyocardial cells procedure of Orlowski and Lingrel (1990) was emThe protocol employed was derived from the methods ployed. Essentially, cells were harvested with a 0.05% described by Harary e t al. (1974) and Libby (1984). trypsin, 0.02% EDTA solution, and replated a t a 1:2 Essentially, hearts from 20 to 40 Sprague Dawley rat split in serum supplemented medium. By this method pups (1-4 days old) were aseptically excised, and atria noncardiomyocytes that had been subjected to two pasand surrounding tissues were trimmed. The remain- sages did not allow any significant growth of cardiac ing heart tissue, mainly ventricles, was washed with myocytes (< 5% of total cell population). Following the Hanks-HEPES medium, pH 7.4, containing the follow- second passage, the nonmyocytes were treated identiing ingredients at the concentrations indicated: NaCl cally to the cultured myocytes. (140 mM), KC1 (5.4 mM), MgC1, (0.8 mM), NaH,PO, (0.44 mM), Na2HP0, (0.34 mM), HEPES (20 mM), Incubations of cells NaHCO, (4.2 mM), and glucose (5.5 mM). Digestion Twenty-four h prior to experimentation, medium in was initiated by the addition of a mixture of Viokase (0.1% w/v) and DNAse (0.002% w/v) dissolved in the the culture dishes was replaced with CMRL 1066 meHanks-HEPES medium. The mixture of heart tissue dium containing 0.5% horse serum and 0.5% fetal calf and the enzymes was stirred a t 200 rpm in a humidified serum. Subsequent to the 24-h period of incubation in incubator (37°C) for 10 min. Following agitation, the medium containing 1%serum components, the cells mixture was allowed to stand so that the heart mince were incubated for a period of 2 h in serum-free CMRL settled to the bottom of the flask. The supernatant from 1066 medium supplemented with 0.5% (w/v) fat-free the first digestion step was decanted and discarded. A bovine serum albumin and the inhibitor of cAMP phosfresh batch of the Viokase/DNAse mixture was then phodiesterases, 3-isobutyl-1-methylxanthine(IBMX, added (1 mllheart) and the digestion procedure re- 100 pM). Cells were then exposed to EGF and other test peated. The supernatant from the second and subse- agents a t the concentrations noted for a period of 10 quent six more digests was carefully removed and min unless indicated otherwise. The incubations were placed in 50 ml sterile conical tubes containing a n terminated by aspirating the medium, adding 1ml of 1 equal volume of CMRL 1066 medium supplemented N NaOH to each dish and freezing the cells on dry ice. with 5% horse serum and 5% fetal calf serum on ice. Determinations of cAMP content Following filtration through a Nitex (mesh 80) sieve, cardiac myocytes and nonmyocytes in the supernatant After thawing the frozen cells, cAMP was extracted were harvested by centrifugation (850g for 10 min). The into NaOH by scraping the cells and agitating the mixcell pellets were suspended in CMRL 1066 medium sup- ture a t room temperature for 1h. The NaOH extracts plemented with 10% serum components (5% horse se- were diluted 1:lO with 50 mM sodium acetate buffer, rum, 5% fetal calf serum) and cell density was adjusted pH 6.4, and 50 pl aliquots of each sample were assayed to 3 x lo6 cells/ml and 2 ml of the cell suspension was in duplicate for cAMP content employing the radioplaced in 35 mm Falcon 3001 dishes. The cells were left immunoassay technique described by Brooker et al. to attach for a period of 1.5 h a t 37°C in a humidified (1979). incubator equilibrated with 95% air and 5% CO,. This time is sufficient to allow attachment of nonmyocyte Protein determinations cells, but the cardiac myocytes remain unattached, and therefore, the supernatant from this preattachment The NaOH extracts of cells from the procedure destep is enriched with cardiac myocytes (Blonde1 et al., scribed above were assayed for protein content by the 1971; Harary e t al., 1974; Libby, 1984; Polinger, 1970, method of Bradford (1976) employing bovine serum alWenzel, 1970). The supernatant from this preattach- bumin a s standard. The protein standards contained ment step was combined and any cell aggregates that NaOH a t a concentration equivalent to that present in may be present were dissociated by aspirating the su- assay mixtures of cell extracts.
(RMHLRQYELL) of the (Y subunit of G, we have demonstrated that G, is the G protein involved in mediating the actions of EGF on cardiac adenylyl cyclase (Nair et al., 1990). Although our findings that EGF increases cAMP accumulation in rat hearts by stimulating adenylyl cyclase activity (Nair et al., 1989, 1990) can explain the chronotropic effects of EGF in ventricular myocytes (Rabkin et al., 1987), since the heart is a heterogeneous tissue comprising a variety of cell types, to date, the cell type(s) in which EGF increases cAMP accumulation remain to be defined. Therefore, studies presented in this communication were performed to determine whether or not EGF increased cAMP content in both cardiomyocytes and nonmyocytes or selectively altered the cyclic nucleotide content in only one cell population.
EGF INCREASES cAMP IN CARDIAC MYOCYTES
Isolation of total cellular RNA and Northern analysis Total cellular RNA was isolated from cultures of cardiac myocytes and nonmyocytes by the method of Chomczynski (1987) as modified by Cinna/Biotecx laboratories International Inc. (RNAzol method). After checking the purity of the RNA (ratio of Absorbance a t 260 nmlabsorbance a t 280 nm > 1.9) and following quantitation, 10 pg of glyoxal-denatured total cellular RNA was fractionated on 1.2% agarose gels. The fractionated RNA on agarose gels was blotted onto nitrocellulose membranes by capillary transfer and fixed on the nitrocellulose membranes by UV-crosslinking. Following prehybridization in the presence of 50% formamide, 100 pg/ml of RNAse-free t-RNA (E. coli), 0.1% SDS, 5 x Denhardt's, 6~ SSC (sodium chloride, sodium citrate, pH 7.0), the Northern blots were hybridized in the same medium, for a period of 20 h, with end-labelled (by T, kinase reaction) 30 oligonucleotide cDNAs complementary to the a-and 6-MHC mRNAs. Following two washes a t room temperature for 15 min and one wash at 50°C for 15 min in medium containing 3 X SSC, 0.1% SDS, and 0.1% sodium pyrophosphate, the blots were exposed to Kodak X-OMAT film. When poly (A)+ mRNA was isolated from total cellular RNA, hybond-mAP (messenger affinity paper) was employed and the manufacturer's (Amersham Corp.) directions were followed. TM
Western blotting Total cell proteins (75 pg) were separated on 6% polyacrylamide gels employing the method of Laemmli (1970). For the detection of P-myosin heavy chain (pMHC) protein, following electrophoretic transfer of the proteins onto nitrocellulose, the Western blots were developed a s described by Gierschik et al. (1985) employing the monoclonal anti-P-myosin heavy chain antibody (CCM-52) as the primary antibody and antimouse IgG conjugated to horseradish peroxidase as the secondary antibody. For estimations of molecular mass of proteins on immunoblots, biotinylated molecular weight markers were employed and these were visualized simultaneously with the proteins recognized by the CCM-52 antibody by avidin horseradish peroxidase conjugate using 3,3'-diaminobenzidine and H,O, as substrates. In experiments concerning the monitoring of phosphotyrosine containing proteins (see below), the nitrocellulose blots were incubated with the monoclonal anti-phosphotyrosine antibody (PY20) as the primary antibody for 1 h and after washing with phosphate-buffered saline (PBS) + 0.3% (v/v) Triton X-100, the nitrocellulose membranes were exposed to goat antimouse horseradish peroxidase conjugate as the secondary antibody. After washing the blots with PBS + 0.3% Triton X-100, the bands were visualized on Kodak X-Omat film, employing the ECL Western blotting detection system (Amersham) according to the instructions provided by the supplier. To determine molecular weights of the various proteins on Western blots, in all experiments, lanes containing molecular weight standards blotted onto nitrocellulose membranes were stained separately with coomassie blue.
561
Assay of functional EGF receptors This assay was based upon the premise that activation of EGF receptor kinase by EGF would cause tyrosine phosphorylation of membrane proteins. Therefore, in membranes derived from cardiomyocytes and nonmyocytes, EGF-elicited tyrosine phosphorylation of the 170 kDa protein was monitored. Briefly, membrane protein (150 pg/50 pl assay) was incubated for 1h a t 0°C in a medium (pH 7.41, which contained the following at the indicated final concentrations: P-glycerophosphate, 20 mM; glycerol 2.5% (v/v); NaC1, 70 mM; Triton X-100,0.025% (v/v); sodium ortho-vanadate, 100 pM; p-nitrophenyl phosphate, 20 mM; ATP, 200 pM; creatine kinase, 1mg/ml; and phosphocreatine, 12 mM. Incubations were performed in the presence and absence of EGF. Reactions were terminated by the addition of 50 p1 of 2X Laemmli sample buffer (Laemmli, 1970) and heating a t 100°C for 3 mins. Proteins in the samples were separated on 7.5% polyacrylamide gels according to the procedure of Laemmli (1970) and transferred to nitrocellulose membranes for Western blotting according to the method of Burnette (1981). Immunoprecipitation of EGF receptors Whole cell proteins (50 pg) were solubilized in 300 pl of buffer containing 50 mM P-glycerophosphate, 70 mM NaC1, 1% (v/v)Triton X-100, 10%(vlv) glycerol, 200 pM PMSF. The samples were then treated with 15 pl of goat antimouse IgG-agarose for 30 min a t 4°C. Following centrifugation, 1 pg of monoclonal anti-EGF receptor antibody was added to the supernatants and the mixture was incubated overnight at 4°C. After addition of 15 p1 of goat antimouse IgG-agarose and incubation for 1h a t 4"C, the samples were centrifuged and the resulting pellets washed three times with the aforementioned solubilization buffer. The final pellets were resuspended in 50 p1 of Laemmli sample medium (Laemmli, 1970) and subjected to Western blot analysis as described above using the anti-EGF receptor antibody. MATERIALS EGF was purchased from Biomedical Technologies. Viokase (pancreatic lipase) was purchased from A. H. Robbins, and DNAse was obtained from Worthington Biochemicals. CMRL 1066 tissue culture medium, horse serum, and fetal calf serum were obtained from Gibco Laboratories. The Nitex mesh 80 and 50 sieves were purchased from Tetco Corp. cAMP antibodies were obtained from Kew Scientific. Goat antimouse IgG-agarose was purchased from Sigma Chemical Co. The 30 oligonucleotide cDNAs complementary to the aand P-MHC mRNAs were the generous gift from Drs. John Orlowski and Jerry B. Lingrel (University of Cincinnati College of Medicine). The monoclonal anti-pmyosin heavy chain antibody (CCM-52)was generously provided by Dr. William A. Clark, J r . (Northwestern University Medical School). Hybond-mAP (messenger affinity paper), which is manufactured by Orgenics Ltd., and monoclonal anti-EGF receptor antibody (EGF-R1) were purchased from Amersham Corp. All other agents and materials were of the highest quality commercially available.
562
YiJ ET AL.
Fig. 1. Photomicrographs of cultured ventricular cardiomyocytes (A) and noncardiomyocytes (B) isolated from neonatal rat hearts. Cardiomyocytes (A) were isolated and cultured as described in Materials and Methods. Cells represented above were in culture for 5 days in CMRL 1066 medium supplemented with 5% horse serum and 5% fetal calf
serum; magnification employed was 400 X . Noncardiomyocytes (B) were in culture for 4 days in CMRL 1066 medium supplemented with 5% horse serum and 5% fetal calf serum; magnification employed was 200~.
RESULTS AND DISCUSSION In order to determine whether EGF-mediated increase in CAMPaccumulation in the heart (Nair et al., 1989) is the result of a n increase in this cyclic nucleotide in cardiomyocytes andlor noncardiomyocytes, the initial aim of our studies was to establish and characterize cultures of rat cardiomyocytes and noncardiomyocytes. The myocyte and nonmyocyte populations from rat hearts are readily distinguished by their morphol-
ogy. The cardiac myocytes (Fig. 1A) are elongated in shape, contain myofibrillar structures, and beat in synchrony. In contrast, the nonmyocytes are flatter, polygonal in shape, lack the contractile elements, and therefore, do not beat (Fig. 1B). These characteristics are similar to those documented by others (see, e.g., Harary et al., 1974; Libby, 1984; Orlowski and Lingrel, 1990). In addition, to further characterize the differences between cardiomyocytes and nonmyocytes and to
EGF INCREASES CAMPIN CARDIAC MYOCYTES
563
Fig. 2. Differential expression of p-MHC (A) and a-MHC mRNA (B) levels in cardiomyocytes and noncardiomyocytes. Total RNA (10 pg each) from cardiomyocytes (MI and noncardiomyocytes (NM) was isolated and fractionated a s described under Materials and Methods.
Following transfer onto nitrocellulose membranes the RNAs were hybridized with 32P-labeled synthetic oligonucleotide probes complementary to the a-MHC mRNA (B) and p-MHC mRNA (A). The Northern blots were exposed to Kodak X-OMAT film for 2 days.
demonstrate that the cardiac myocytes cultured under the conditions of our studies respond in a physiologically relevant manner, the presence and modulation by thyroid hormone of a- and P-myosin heavy chain (MHC) mRNA levels was investigated. The data presented in Figure 2 demonstrate that the total RNA from cardiomyocytes contained transcripts for both aand p-MHC, whereas the total RNA from nonmyocytes did not contain mRNA for a-and P-MHC. The short (30 oligonucleotide long) cDNAs complementary to the aand p-MHC mRNA also nonspecifically hybridized with the 28s ribosomal RNA in both myocytes and nonmyocytes (Fig. 2). Although these data are identical to those previously described by Orlowski and Lingrel (1990), to avoid nonspecific hybridization in further experiments (see below) with a- and @-MHC cDNAs, poly(A)+ mRNA isolated from cells was employed. The presence and absence of myosin heavy chain transcripts in cardiomyocytes and noncardiomyocytes, respectively, were also manifested in the expression of the corresponding proteins. For instance, as monitored by immunoblotting, cardiomyocytes expressed the P-myosin heavy chain protein (Fig. 3, lanes 1 and 2), whereas this protein was not expressed in noncardiomyocytes (Fig. 3, lanes 3 and 4). Moreover, to demonstrate that cardiac myocytes used in our studies are similar to those employed by other groups and responding in a physiologically relevant manner the regulation of a-and P-MHC mRNAs by thyroid hormone was investigated. Three days after plating, cultures of cardiomyocytes were treated for 4 days with either 1 pM thyroid hormone (T3) or vehicle (NaOH). The medium was changed every 48 h. Five micrograms of the poly A+ mRNA isolated from these cells was subjected to Northern analysis employing the 30 oligonucleotide probes complementary to the a- or P-MHC mRNAs. Figure 4 shows that as compared with controls, thyroid hormone treatment increased the expression of a-MHC mRNA, and decreased the P-MHC mRNA levels (Fig. 4).It is noteworthy that the same blot was probed with
Fig. 3. Differential expression of p-MHC protein in cardiomyocytes and noncardiomyocytes. Total cellular protein (75 pg each) from the two types of cells after 5 days in culture was separated on 6%polyacrylamide gels by SDS-PAGE. The proteins were electrophoretically transferred onto nitrocellulose and Western blots were developed with monoclonal antibody to p-MHC as the primary antibody. Lanes 1and 2 represent proteins from cardiac myocytes; lanes 3 and 4 show proteins from nonmyocytes. The single protein recognized by the p-MHC antibody (CCM-52) in cardiomyocyte protein is the 200 kDa P-MHC.
YU ET AL.
564
Nortlicrn Blot of Poly A+ mRNA from Cardiac Myocytes
T3
c
T3
II-MHC
a -MHC+
Fig. 4. Antithetic regulation of a- and p-MHC mRNA by thyroid hormone treatment in cardiomyocytes. Poly(A)' mRNA from cardiac myocytes, which were either treated or not treated with thyroid hormone (T,)was fractionated on 1.2% agarose gels and after transfer onto nitrocellulose Northern analysis with cDNAs complementary to
the a-MHC and p-MHC mRNAs was performed as described in Materials and Methods. Cardiac myocytes were exposed to either T, (1FM) or vehicle (C) for 4 days. The same blot was hybridized first with the a-MHC cDNA (left) and after stripping the a-MHC cDNA, with the p-MHC cDNA probe.
Time Course of cAMP accumulation i+n
S 0
.-E
0
5
10
15
20
Minutes After Addition of EGF (100 nM) Fig. 5. Time course of EGF-elicited cAMP accumulation in cultured cardiomyocytes. Cells were cultured as described in Materials and Methods. Two hs prior to experimentation cells were placed in medium devoid of serum. Following addition of EGF (100 nM) the reactions were terminated a t the designated time points by the addition of 1N NaOH and freezing the cells. Data presented are mean 2 S.E.M. (n = 4)ofpercent stimulation over control values (12.1 t 0.6 pmolimg protein) from one of three separate experiments. *p < 0.001 as compared with control at time zero; unpaired Students t-test analysis.
cDNAs for the a-MHC and P-MHC mRNAs. The antithetic regulation of a- and p-MHC mRNAs by thyroid hormone observed in our studies is similar to that demonstrated by others in whole hearts (Chizzonite and Zak, 1984; Lompre et al., 19841, and cultured cells (see, e.g., Gustafson et al., 1987; Orlowski and Lingrel, 19901, indicating that cardiomyocytes employed in the studies described here respond to normal physiological stimuli, and therefore, are adequate as a n experimental model. The next series of experiments were performed to investigate the effects of EGF on cAMP accumulation
in cardiac myocytes and to establish the temporal changes in this cyclic nucleotide following the addition of EGF. The data in Figure 5 demonstrate that 6 min after the addition of EGF (100 nM) to cultures of cardiomyocytes, significant elevation in cellular cAMP could be detected. Maximal stimulation of cAMP accumulation in response to EGF was observed a t the 10min time point (Fig. 5) and the increase in cellular content of the cyclic nucleotide was sustained over the remaining 10-min period of incubation. Since EGF-elicited increase in cellular cAMP content was maximal and sustained after the initial 10 min, in all subsequent experiments, incubations were terminated after cells had been exposed to EGF for 10 min. EGF increased cellular cAMP accumulation in cardiomyocytes in a concentration dependent manner. Hence, half-maximal and maximal stimulation of cAMP accumulation in cardiomyocytes was observed at EGF concentrations of 2 nM and 100 nM, respectively (Fig. 6). In contrast, in cultures of noncardiomyocytes EGF at concentrations as high as 1 pM did not stimulate accumulation of cAMP (Fig. 6). These data, therefore, demonstrate that EGF stimulates cAMP accumulation in cardiomyocytes but not in noncardiomyocytes. The inability of EGF to increase cellular accumulation of cAMP in noncardiomyocytes was not a reflection of a defect(s1 in the adenylyl cyclase andlor its regulatory system, since the P-adrenoreceptor agonist isoproterenol increased cAMP content in these cells by 2.6-fold (c.f. control = 16.5 t 1(n = 4) pmol/mg protein vs. 43.1 4.3 (n = 4) pmol/mg protein in presence of 100 nM isoproterenol). The lack of a n effect of EGF on cAMP accumulation in noncardiomyocytes may result from either the lack of EGF receptors on this cell type or the lack of coupling between the EGF receptors and adenylyl cyclase system. To investigate the presence of functional EGF receptors on membranes of noncardiomyocytes, the abil-
*
EGF INCREASES cAMP IN CARDIAC MYOCYTES Cardiomyocytes
565
Non-cardiomyocytes h
h
.-ai
._ ai
0
0
!-
U
n
9
9
E
E
-3
i
L
9
9
EGF (nM) Fig. 6. Concentration dependence of EGF-mediated increase in cellular cAMP accumulation in cultured ventricular cardiomyocytes (left) and noncardiomyocytes (right).After 4 days in culture, the cells were placed in medium containing 0.5%horse serum and 0.5% fetal calf serum. Twenty-two hs later the cells were placed in serum-free medium and after 2 hours challenged with different concentrations of
EGF (nM) EGF for 10 mins. Control incubations received the vehicle but no EGF. Reactions were terminated as described in legend to Figure 5 and cAMP measured as described in Materials and Methods. Data presented are mean S.E.M. (n = 4)from a representative of three similar experiments. #p < 0.05, *p < 0.001 as compared with controls in the absence of EGF; unpaired Students t-test analysis.
*
ity of EGF to phosphorylate membrane proteins on the EGF-elicited increase in cAMP accumulation in isotyrosine residues was investigated. This approach is lated perfused hearts or EGF-mediated activation of based upon the premise that apart from binding the adenylyl cyclase (Nair et al., 1989), experiments with ligand, a functional EGF receptor would also activate propranolol(100 nM) demonstrated that the P-adrenothe intrinsic protein tyrosine kinase activity and phos- receptor antagonist did not alter the actions of EGF on phorylate membrane proteins. Figure 7A shows that adenylyl cyclase (data not shown). although several proteins were phosphorylated on tyThe effects of EGF on cAMP accumulation in cardiorosine residues, the most prominent increase in ty- myocytes are smaller than those reported by us in inrosine phosphorylation in response to EGF was associ- tact perfused rat hearts (Nair et al., 1989). However, ated with a 170 kDa protein. Since EGF increased this difference may be a reflection of the fact that for tyrosine phosphorylation of the 170 kDa protein in several days, cardiomyocytes are exposed to serum membranes isolated from both noncardiomyocytes as components enriched with various growth factors, well as cardiomyocytes (Fig. 7A), it is inferred that which may decrease EGF responsiveness. Nonetheless, functional EGF receptors are present in both cardiac our finding that EGF stimulates cAMP accumulation myocytes and noncardiomyocytes. The contention that in cardiomyocytes but not in noncardiomyocytes sugEGF receptors are present in both cell types is further gests that cAMP mediates the chronotropic effects of supported by the finding that as monitored by Western EGF in cultured ventricular myocytes (Rabkin et al., analysis, the anti-EGF receptor antibody was able to 1987). Moreover, since in the experiments described immunoprecipitate EGF receptor protein (170 kDa) above, the phosphodiesterase inhibitor, IBMX was from both cardiomyocytes and noncardiomyocytes present, the data strongly suggest that EGF-mediated (Fig. 7B). stimulation of adenylyl cyclase is responsible for the Previously we demonstrated that EGF stimulates increase in cellular cAMP and this contention is consiscardiac adenylyl cyclase via a G protein (5),namely, G, tent with our previous reports (Nair et al., 1989, 1990) (Nair et al., 1990). Therefore, to investigate at the cel- concerning direct measurements of cardiac adenylyl cylular level the role of G, in EGF-elicited stimulation of clase. Interestingly, despite the ability of EGF to stimcAMP content in cardiomyocytes, experiments were ulate tyrosine phosphorylation in both cardiomyocytes performed with cholera toxin. Cholera toxin ADP-ribo- as well as noncardiomyocytes and the presence of imsylates and maximally activates G, (Gillman, 1984) munoreactive EGF receptors on membranes of both cell such that its activity cannot be further modulated by types, the effects of EGF on cellular cAMP accumulaagonists. The data in Table 1 show that cholera toxin tion were restricted to cardiomyocytes. Thus it appears treatment of cardiomyocytes increased cellular cAMP that the coupling between the EGF receptors and adecontent by twofold and abolished the ability of EGF to nylyl cyclase second messenger system requires the stimulate adenylyl cyclase. One explanation for these participation of component(s), which is(are) present in findings is that when adenylyl cyclase is maximally some cell types b-ut not in others. In this respect i t is stimulated by cholera toxin, EGF cannot further modu- noteworthy that EGF decreases glucagon- and proslate its activity. However, since employing a n antibody taglandin El-stimulated cAMP accumulation in hepaagainst G,, we have previously shown that G, mediates tocytes and fibroblastic cells, respectively, without althe effects of EGF on adenylyl cyclase (Nair et al., tering basal cAMP levels (Anderson e t al., 1979; Bosch 19901, these findings support the notion that G, medi- et al., 1986). Whether or not the attenuation of agonist ates the effects of EGF on adenylyl cyclase. Similarly, elevated cAMP accumulation in hepatocytes and fibroconsistent with our previous demonstration that blasts by EGF is due to modulations of adenylyl cyclase P-adrenoreceptor activation is not involved in either or cAMP phosphodiesterase(s) is currently unknown.
566
YU ET AL.
TABLE 1. Effect of cholera toxin treatment of cardiomyocytes on the abilitv of EGF to increase cAMP accumulation1 ~
Control CTX cAMP accumulation pmol/ 10 min/mg protein
Condition
+
No addition EGF (100 nM)
14.3 1.3 22.4 5 1.2*
+
31.2 29.2
+ 4.5* + 4.1 (n.s.)
‘Cardiomyocytes weretreated either with or withnuticontrol)cholera toxin (CTX, 15 ng/ml for 15 h). Experimental conditions were similar to those for experiments in Figures 5 and 6 except that the cholera toxin was added 9 h after placing cells in serum-freemedium. Thecells were then challenged with 100 nM EGF for IOmin and cAMP content was measured as described in Materials and Methods. *Valuespresentedaremean S.E.M. offourdeterminations;p
Epidermal Growth Factor Stimulates cAMP Accumulation in Cultured Rat Cardiac Myocytes Y l M l N G YU, BlPlN C. NAIR, AND TARUN 6.PATEL* Department of Pharmacology, The Health Science Center, University of Tennessee, Memphis, Tennessee 38 163 We have previously shown that epidermal growth factor (EGF) augments CAMP accumulation in the heart and stimulates cardiac adenylyl cyclase via a G protein mediated mechanism (Nair et al., 1989). More recently, employing an antibody against the carboxy-terminus decapeptide of G,, we have demonstrated that G,, mediates the effects of EGF on cardiac adenylyl cyclase (Nair et al., 1990). Since the heart comprises of a variety of cell types, the purpose of the studies presented here was to determine whether or not the effects of EGF on adenylyl cyclase were mediated in cardiac myocytes or noncardiomyocytes. Therefore, cultures of ventricular cardiomyocytes and noncardiomyocytes from neonatal rat hearts were established and characterized. Apart from the differences in cellular morphology, cardiomyocytes but not the noncardiomyocytes employed in our studies expressed the a-and P-myosin heavy chain (MHC) mRNA and the P-MHC protein. Additionally, as described previously, treatment of cardiomyocytes with thyroid hormone increased a-MHC mRNA and decreased the expression of P-MHC mRNA, indicating that the cardiomyocytes employed in our studies were responding in a physiologically relevant manner. EGF in a time-dependent manner increased cAMP accumulation in the cardiomyocytes but not in noncardiomyocytes. Maximum and half-maximum effects were observed at 100 n M and 2 n M concentrations of EGF, respectively. As determined by the presence of immunoreactive EGF receptors and tyrosine phosphorylation of the 170 kDa protein in membranes of cardiomyocytes and noncardiomyocytes, both the cell populations contained functional ECF receptors. Therefore, the differential effects of EGF on cAMP accumulation in the two cell populations appear to be due to differential coupling of the EGF receptors to the adenylyl cyclase system rather than the absence of EGF receptors in noncardiomyocytes. Consistent with our previous findings in isolated membranes and perfused rat hearts, EGF-elicited increase in cAMP accumulation in cardiomyocytes did not involve activation of P-adrenoreceptors and was abolished by prior treatment of cells with cholera toxin. Overall, our findings demonstrate that EGF-elicited increase in cAMP accumulation in the heart is the reflection of changes in cAMP content of cardiomyocytes and not noncardiomyocytes.
Epidermal growth factor (EGF), a 53 amino acid polypeptide elicits a wide spectrum of biological actions in a number of tissues and cell types (see Carpenter 1979, 1987 for reviews). Rabkin et al. (1987) have described a novel effect of EGF in cultured chick embryo cardiac myocytes. These authors demonstrated that EGF produces a chronotropic effect in cultured ventricular myocytes from chick embryos. Maximal increase in the beating rate of ventricular myocytes was observed 15 min after the addition of EGF and the response was sustained over a period of 1.5 h (Rabkin et al., 1987). Because cAMP is well established to produce chronotropic effects in the heart (see, e.g., Hofmann et al., 1987) and since the mechanism(s) by which EGF stimulated the beating of cardiac myocytes remained unknown, we proposed the hypothesis that EGF increased the accumulation of cAMP in the heart. To date, our studies 0 1992 WILEY-LISS, INC.
have shown that EGF increases cAMP accumulation in the perfused rat heart and this increase in cellular accumulation of cAMP is the result of stimulation of the cardiac adenylyl cyclase activity (Nair et al., 1989). The stimulation of cardiac adenylyl cyclase and increase in tissue cAMP content were independent of stimulation of 6-adrenoreceptors or activation of protein kinase C (Nair e t al., 1989). However, like P-adrenoreceptor-mediated activation of adenylyl cyclase, EGF stimulates cardiac adenylyl cyclase via a G protein (Nair et al., 1989). More recently, employing a n antibody (CS1) directed against the carboxy-terminus decapeptide
Received August 22,1991; accepted October 3, 1991.
*To whom reprint requestsicorrespondence should be addressed.
560
YU ET AL.
pernatant through a 23-gauge needle into a syringe and filtering through a Nitex (mesh 50) sieve. The density of cells in the filtrate was adjusted to 2 x lo6 cells/ml of CMRL 1066 medium containing 10% serum components. Four million cells (2 ml) were plated in 35 mm dishes (Falcon 3001) and placed in a 37°C humidified incubator equilibrated with 95% air/5% CO,; 18 h after plating the cells, the medium was changed to remove unattached cells. Under these conditions, cardiomyocytes, which represented > 85% of the cell population, beat in a synchronous fashion 48 h after plating. For the purposes of culturing nonmyocyte cells, after removing the myocyte-enriched supernatant from the plates in the preattachment step, 2 ml of CMRL 1066 medium containing 5% horse serum and 5% fetal calf serum were added to each of the 35 mm dishes. Under these conditions, nonmyocytes divided and were confluMATERIALS AND METHODS ent 3 days after plating. To ensure minimum contamiCultures of neonatal rat myocardial and nation of noncardiomyocytes by cardiomyocytes, the nonmyocardial cells procedure of Orlowski and Lingrel (1990) was emThe protocol employed was derived from the methods ployed. Essentially, cells were harvested with a 0.05% described by Harary e t al. (1974) and Libby (1984). trypsin, 0.02% EDTA solution, and replated a t a 1:2 Essentially, hearts from 20 to 40 Sprague Dawley rat split in serum supplemented medium. By this method pups (1-4 days old) were aseptically excised, and atria noncardiomyocytes that had been subjected to two pasand surrounding tissues were trimmed. The remain- sages did not allow any significant growth of cardiac ing heart tissue, mainly ventricles, was washed with myocytes (< 5% of total cell population). Following the Hanks-HEPES medium, pH 7.4, containing the follow- second passage, the nonmyocytes were treated identiing ingredients at the concentrations indicated: NaCl cally to the cultured myocytes. (140 mM), KC1 (5.4 mM), MgC1, (0.8 mM), NaH,PO, (0.44 mM), Na2HP0, (0.34 mM), HEPES (20 mM), Incubations of cells NaHCO, (4.2 mM), and glucose (5.5 mM). Digestion Twenty-four h prior to experimentation, medium in was initiated by the addition of a mixture of Viokase (0.1% w/v) and DNAse (0.002% w/v) dissolved in the the culture dishes was replaced with CMRL 1066 meHanks-HEPES medium. The mixture of heart tissue dium containing 0.5% horse serum and 0.5% fetal calf and the enzymes was stirred a t 200 rpm in a humidified serum. Subsequent to the 24-h period of incubation in incubator (37°C) for 10 min. Following agitation, the medium containing 1%serum components, the cells mixture was allowed to stand so that the heart mince were incubated for a period of 2 h in serum-free CMRL settled to the bottom of the flask. The supernatant from 1066 medium supplemented with 0.5% (w/v) fat-free the first digestion step was decanted and discarded. A bovine serum albumin and the inhibitor of cAMP phosfresh batch of the Viokase/DNAse mixture was then phodiesterases, 3-isobutyl-1-methylxanthine(IBMX, added (1 mllheart) and the digestion procedure re- 100 pM). Cells were then exposed to EGF and other test peated. The supernatant from the second and subse- agents a t the concentrations noted for a period of 10 quent six more digests was carefully removed and min unless indicated otherwise. The incubations were placed in 50 ml sterile conical tubes containing a n terminated by aspirating the medium, adding 1ml of 1 equal volume of CMRL 1066 medium supplemented N NaOH to each dish and freezing the cells on dry ice. with 5% horse serum and 5% fetal calf serum on ice. Determinations of cAMP content Following filtration through a Nitex (mesh 80) sieve, cardiac myocytes and nonmyocytes in the supernatant After thawing the frozen cells, cAMP was extracted were harvested by centrifugation (850g for 10 min). The into NaOH by scraping the cells and agitating the mixcell pellets were suspended in CMRL 1066 medium sup- ture a t room temperature for 1h. The NaOH extracts plemented with 10% serum components (5% horse se- were diluted 1:lO with 50 mM sodium acetate buffer, rum, 5% fetal calf serum) and cell density was adjusted pH 6.4, and 50 pl aliquots of each sample were assayed to 3 x lo6 cells/ml and 2 ml of the cell suspension was in duplicate for cAMP content employing the radioplaced in 35 mm Falcon 3001 dishes. The cells were left immunoassay technique described by Brooker et al. to attach for a period of 1.5 h a t 37°C in a humidified (1979). incubator equilibrated with 95% air and 5% CO,. This time is sufficient to allow attachment of nonmyocyte Protein determinations cells, but the cardiac myocytes remain unattached, and therefore, the supernatant from this preattachment The NaOH extracts of cells from the procedure destep is enriched with cardiac myocytes (Blonde1 et al., scribed above were assayed for protein content by the 1971; Harary e t al., 1974; Libby, 1984; Polinger, 1970, method of Bradford (1976) employing bovine serum alWenzel, 1970). The supernatant from this preattach- bumin a s standard. The protein standards contained ment step was combined and any cell aggregates that NaOH a t a concentration equivalent to that present in may be present were dissociated by aspirating the su- assay mixtures of cell extracts.
(RMHLRQYELL) of the (Y subunit of G, we have demonstrated that G, is the G protein involved in mediating the actions of EGF on cardiac adenylyl cyclase (Nair et al., 1990). Although our findings that EGF increases cAMP accumulation in rat hearts by stimulating adenylyl cyclase activity (Nair et al., 1989, 1990) can explain the chronotropic effects of EGF in ventricular myocytes (Rabkin et al., 1987), since the heart is a heterogeneous tissue comprising a variety of cell types, to date, the cell type(s) in which EGF increases cAMP accumulation remain to be defined. Therefore, studies presented in this communication were performed to determine whether or not EGF increased cAMP content in both cardiomyocytes and nonmyocytes or selectively altered the cyclic nucleotide content in only one cell population.
EGF INCREASES cAMP IN CARDIAC MYOCYTES
Isolation of total cellular RNA and Northern analysis Total cellular RNA was isolated from cultures of cardiac myocytes and nonmyocytes by the method of Chomczynski (1987) as modified by Cinna/Biotecx laboratories International Inc. (RNAzol method). After checking the purity of the RNA (ratio of Absorbance a t 260 nmlabsorbance a t 280 nm > 1.9) and following quantitation, 10 pg of glyoxal-denatured total cellular RNA was fractionated on 1.2% agarose gels. The fractionated RNA on agarose gels was blotted onto nitrocellulose membranes by capillary transfer and fixed on the nitrocellulose membranes by UV-crosslinking. Following prehybridization in the presence of 50% formamide, 100 pg/ml of RNAse-free t-RNA (E. coli), 0.1% SDS, 5 x Denhardt's, 6~ SSC (sodium chloride, sodium citrate, pH 7.0), the Northern blots were hybridized in the same medium, for a period of 20 h, with end-labelled (by T, kinase reaction) 30 oligonucleotide cDNAs complementary to the a-and 6-MHC mRNAs. Following two washes a t room temperature for 15 min and one wash at 50°C for 15 min in medium containing 3 X SSC, 0.1% SDS, and 0.1% sodium pyrophosphate, the blots were exposed to Kodak X-OMAT film. When poly (A)+ mRNA was isolated from total cellular RNA, hybond-mAP (messenger affinity paper) was employed and the manufacturer's (Amersham Corp.) directions were followed. TM
Western blotting Total cell proteins (75 pg) were separated on 6% polyacrylamide gels employing the method of Laemmli (1970). For the detection of P-myosin heavy chain (pMHC) protein, following electrophoretic transfer of the proteins onto nitrocellulose, the Western blots were developed a s described by Gierschik et al. (1985) employing the monoclonal anti-P-myosin heavy chain antibody (CCM-52) as the primary antibody and antimouse IgG conjugated to horseradish peroxidase as the secondary antibody. For estimations of molecular mass of proteins on immunoblots, biotinylated molecular weight markers were employed and these were visualized simultaneously with the proteins recognized by the CCM-52 antibody by avidin horseradish peroxidase conjugate using 3,3'-diaminobenzidine and H,O, as substrates. In experiments concerning the monitoring of phosphotyrosine containing proteins (see below), the nitrocellulose blots were incubated with the monoclonal anti-phosphotyrosine antibody (PY20) as the primary antibody for 1 h and after washing with phosphate-buffered saline (PBS) + 0.3% (v/v) Triton X-100, the nitrocellulose membranes were exposed to goat antimouse horseradish peroxidase conjugate as the secondary antibody. After washing the blots with PBS + 0.3% Triton X-100, the bands were visualized on Kodak X-Omat film, employing the ECL Western blotting detection system (Amersham) according to the instructions provided by the supplier. To determine molecular weights of the various proteins on Western blots, in all experiments, lanes containing molecular weight standards blotted onto nitrocellulose membranes were stained separately with coomassie blue.
561
Assay of functional EGF receptors This assay was based upon the premise that activation of EGF receptor kinase by EGF would cause tyrosine phosphorylation of membrane proteins. Therefore, in membranes derived from cardiomyocytes and nonmyocytes, EGF-elicited tyrosine phosphorylation of the 170 kDa protein was monitored. Briefly, membrane protein (150 pg/50 pl assay) was incubated for 1h a t 0°C in a medium (pH 7.41, which contained the following at the indicated final concentrations: P-glycerophosphate, 20 mM; glycerol 2.5% (v/v); NaC1, 70 mM; Triton X-100,0.025% (v/v); sodium ortho-vanadate, 100 pM; p-nitrophenyl phosphate, 20 mM; ATP, 200 pM; creatine kinase, 1mg/ml; and phosphocreatine, 12 mM. Incubations were performed in the presence and absence of EGF. Reactions were terminated by the addition of 50 p1 of 2X Laemmli sample buffer (Laemmli, 1970) and heating a t 100°C for 3 mins. Proteins in the samples were separated on 7.5% polyacrylamide gels according to the procedure of Laemmli (1970) and transferred to nitrocellulose membranes for Western blotting according to the method of Burnette (1981). Immunoprecipitation of EGF receptors Whole cell proteins (50 pg) were solubilized in 300 pl of buffer containing 50 mM P-glycerophosphate, 70 mM NaC1, 1% (v/v)Triton X-100, 10%(vlv) glycerol, 200 pM PMSF. The samples were then treated with 15 pl of goat antimouse IgG-agarose for 30 min a t 4°C. Following centrifugation, 1 pg of monoclonal anti-EGF receptor antibody was added to the supernatants and the mixture was incubated overnight at 4°C. After addition of 15 p1 of goat antimouse IgG-agarose and incubation for 1h a t 4"C, the samples were centrifuged and the resulting pellets washed three times with the aforementioned solubilization buffer. The final pellets were resuspended in 50 p1 of Laemmli sample medium (Laemmli, 1970) and subjected to Western blot analysis as described above using the anti-EGF receptor antibody. MATERIALS EGF was purchased from Biomedical Technologies. Viokase (pancreatic lipase) was purchased from A. H. Robbins, and DNAse was obtained from Worthington Biochemicals. CMRL 1066 tissue culture medium, horse serum, and fetal calf serum were obtained from Gibco Laboratories. The Nitex mesh 80 and 50 sieves were purchased from Tetco Corp. cAMP antibodies were obtained from Kew Scientific. Goat antimouse IgG-agarose was purchased from Sigma Chemical Co. The 30 oligonucleotide cDNAs complementary to the aand P-MHC mRNAs were the generous gift from Drs. John Orlowski and Jerry B. Lingrel (University of Cincinnati College of Medicine). The monoclonal anti-pmyosin heavy chain antibody (CCM-52)was generously provided by Dr. William A. Clark, J r . (Northwestern University Medical School). Hybond-mAP (messenger affinity paper), which is manufactured by Orgenics Ltd., and monoclonal anti-EGF receptor antibody (EGF-R1) were purchased from Amersham Corp. All other agents and materials were of the highest quality commercially available.
562
YiJ ET AL.
Fig. 1. Photomicrographs of cultured ventricular cardiomyocytes (A) and noncardiomyocytes (B) isolated from neonatal rat hearts. Cardiomyocytes (A) were isolated and cultured as described in Materials and Methods. Cells represented above were in culture for 5 days in CMRL 1066 medium supplemented with 5% horse serum and 5% fetal calf
serum; magnification employed was 400 X . Noncardiomyocytes (B) were in culture for 4 days in CMRL 1066 medium supplemented with 5% horse serum and 5% fetal calf serum; magnification employed was 200~.
RESULTS AND DISCUSSION In order to determine whether EGF-mediated increase in CAMPaccumulation in the heart (Nair et al., 1989) is the result of a n increase in this cyclic nucleotide in cardiomyocytes andlor noncardiomyocytes, the initial aim of our studies was to establish and characterize cultures of rat cardiomyocytes and noncardiomyocytes. The myocyte and nonmyocyte populations from rat hearts are readily distinguished by their morphol-
ogy. The cardiac myocytes (Fig. 1A) are elongated in shape, contain myofibrillar structures, and beat in synchrony. In contrast, the nonmyocytes are flatter, polygonal in shape, lack the contractile elements, and therefore, do not beat (Fig. 1B). These characteristics are similar to those documented by others (see, e.g., Harary et al., 1974; Libby, 1984; Orlowski and Lingrel, 1990). In addition, to further characterize the differences between cardiomyocytes and nonmyocytes and to
EGF INCREASES CAMPIN CARDIAC MYOCYTES
563
Fig. 2. Differential expression of p-MHC (A) and a-MHC mRNA (B) levels in cardiomyocytes and noncardiomyocytes. Total RNA (10 pg each) from cardiomyocytes (MI and noncardiomyocytes (NM) was isolated and fractionated a s described under Materials and Methods.
Following transfer onto nitrocellulose membranes the RNAs were hybridized with 32P-labeled synthetic oligonucleotide probes complementary to the a-MHC mRNA (B) and p-MHC mRNA (A). The Northern blots were exposed to Kodak X-OMAT film for 2 days.
demonstrate that the cardiac myocytes cultured under the conditions of our studies respond in a physiologically relevant manner, the presence and modulation by thyroid hormone of a- and P-myosin heavy chain (MHC) mRNA levels was investigated. The data presented in Figure 2 demonstrate that the total RNA from cardiomyocytes contained transcripts for both aand p-MHC, whereas the total RNA from nonmyocytes did not contain mRNA for a-and P-MHC. The short (30 oligonucleotide long) cDNAs complementary to the aand p-MHC mRNA also nonspecifically hybridized with the 28s ribosomal RNA in both myocytes and nonmyocytes (Fig. 2). Although these data are identical to those previously described by Orlowski and Lingrel (1990), to avoid nonspecific hybridization in further experiments (see below) with a- and @-MHC cDNAs, poly(A)+ mRNA isolated from cells was employed. The presence and absence of myosin heavy chain transcripts in cardiomyocytes and noncardiomyocytes, respectively, were also manifested in the expression of the corresponding proteins. For instance, as monitored by immunoblotting, cardiomyocytes expressed the P-myosin heavy chain protein (Fig. 3, lanes 1 and 2), whereas this protein was not expressed in noncardiomyocytes (Fig. 3, lanes 3 and 4). Moreover, to demonstrate that cardiac myocytes used in our studies are similar to those employed by other groups and responding in a physiologically relevant manner the regulation of a-and P-MHC mRNAs by thyroid hormone was investigated. Three days after plating, cultures of cardiomyocytes were treated for 4 days with either 1 pM thyroid hormone (T3) or vehicle (NaOH). The medium was changed every 48 h. Five micrograms of the poly A+ mRNA isolated from these cells was subjected to Northern analysis employing the 30 oligonucleotide probes complementary to the a- or P-MHC mRNAs. Figure 4 shows that as compared with controls, thyroid hormone treatment increased the expression of a-MHC mRNA, and decreased the P-MHC mRNA levels (Fig. 4).It is noteworthy that the same blot was probed with
Fig. 3. Differential expression of p-MHC protein in cardiomyocytes and noncardiomyocytes. Total cellular protein (75 pg each) from the two types of cells after 5 days in culture was separated on 6%polyacrylamide gels by SDS-PAGE. The proteins were electrophoretically transferred onto nitrocellulose and Western blots were developed with monoclonal antibody to p-MHC as the primary antibody. Lanes 1and 2 represent proteins from cardiac myocytes; lanes 3 and 4 show proteins from nonmyocytes. The single protein recognized by the p-MHC antibody (CCM-52) in cardiomyocyte protein is the 200 kDa P-MHC.
YU ET AL.
564
Nortlicrn Blot of Poly A+ mRNA from Cardiac Myocytes
T3
c
T3
II-MHC
a -MHC+
Fig. 4. Antithetic regulation of a- and p-MHC mRNA by thyroid hormone treatment in cardiomyocytes. Poly(A)' mRNA from cardiac myocytes, which were either treated or not treated with thyroid hormone (T,)was fractionated on 1.2% agarose gels and after transfer onto nitrocellulose Northern analysis with cDNAs complementary to
the a-MHC and p-MHC mRNAs was performed as described in Materials and Methods. Cardiac myocytes were exposed to either T, (1FM) or vehicle (C) for 4 days. The same blot was hybridized first with the a-MHC cDNA (left) and after stripping the a-MHC cDNA, with the p-MHC cDNA probe.
Time Course of cAMP accumulation i+n
S 0
.-E
0
5
10
15
20
Minutes After Addition of EGF (100 nM) Fig. 5. Time course of EGF-elicited cAMP accumulation in cultured cardiomyocytes. Cells were cultured as described in Materials and Methods. Two hs prior to experimentation cells were placed in medium devoid of serum. Following addition of EGF (100 nM) the reactions were terminated a t the designated time points by the addition of 1N NaOH and freezing the cells. Data presented are mean 2 S.E.M. (n = 4)ofpercent stimulation over control values (12.1 t 0.6 pmolimg protein) from one of three separate experiments. *p < 0.001 as compared with control at time zero; unpaired Students t-test analysis.
cDNAs for the a-MHC and P-MHC mRNAs. The antithetic regulation of a- and p-MHC mRNAs by thyroid hormone observed in our studies is similar to that demonstrated by others in whole hearts (Chizzonite and Zak, 1984; Lompre et al., 19841, and cultured cells (see, e.g., Gustafson et al., 1987; Orlowski and Lingrel, 19901, indicating that cardiomyocytes employed in the studies described here respond to normal physiological stimuli, and therefore, are adequate as a n experimental model. The next series of experiments were performed to investigate the effects of EGF on cAMP accumulation
in cardiac myocytes and to establish the temporal changes in this cyclic nucleotide following the addition of EGF. The data in Figure 5 demonstrate that 6 min after the addition of EGF (100 nM) to cultures of cardiomyocytes, significant elevation in cellular cAMP could be detected. Maximal stimulation of cAMP accumulation in response to EGF was observed a t the 10min time point (Fig. 5) and the increase in cellular content of the cyclic nucleotide was sustained over the remaining 10-min period of incubation. Since EGF-elicited increase in cellular cAMP content was maximal and sustained after the initial 10 min, in all subsequent experiments, incubations were terminated after cells had been exposed to EGF for 10 min. EGF increased cellular cAMP accumulation in cardiomyocytes in a concentration dependent manner. Hence, half-maximal and maximal stimulation of cAMP accumulation in cardiomyocytes was observed at EGF concentrations of 2 nM and 100 nM, respectively (Fig. 6). In contrast, in cultures of noncardiomyocytes EGF at concentrations as high as 1 pM did not stimulate accumulation of cAMP (Fig. 6). These data, therefore, demonstrate that EGF stimulates cAMP accumulation in cardiomyocytes but not in noncardiomyocytes. The inability of EGF to increase cellular accumulation of cAMP in noncardiomyocytes was not a reflection of a defect(s1 in the adenylyl cyclase andlor its regulatory system, since the P-adrenoreceptor agonist isoproterenol increased cAMP content in these cells by 2.6-fold (c.f. control = 16.5 t 1(n = 4) pmol/mg protein vs. 43.1 4.3 (n = 4) pmol/mg protein in presence of 100 nM isoproterenol). The lack of a n effect of EGF on cAMP accumulation in noncardiomyocytes may result from either the lack of EGF receptors on this cell type or the lack of coupling between the EGF receptors and adenylyl cyclase system. To investigate the presence of functional EGF receptors on membranes of noncardiomyocytes, the abil-
*
EGF INCREASES cAMP IN CARDIAC MYOCYTES Cardiomyocytes
565
Non-cardiomyocytes h
h
.-ai
._ ai
0
0
!-
U
n
9
9
E
E
-3
i
L
9
9
EGF (nM) Fig. 6. Concentration dependence of EGF-mediated increase in cellular cAMP accumulation in cultured ventricular cardiomyocytes (left) and noncardiomyocytes (right).After 4 days in culture, the cells were placed in medium containing 0.5%horse serum and 0.5% fetal calf serum. Twenty-two hs later the cells were placed in serum-free medium and after 2 hours challenged with different concentrations of
EGF (nM) EGF for 10 mins. Control incubations received the vehicle but no EGF. Reactions were terminated as described in legend to Figure 5 and cAMP measured as described in Materials and Methods. Data presented are mean S.E.M. (n = 4)from a representative of three similar experiments. #p < 0.05, *p < 0.001 as compared with controls in the absence of EGF; unpaired Students t-test analysis.
*
ity of EGF to phosphorylate membrane proteins on the EGF-elicited increase in cAMP accumulation in isotyrosine residues was investigated. This approach is lated perfused hearts or EGF-mediated activation of based upon the premise that apart from binding the adenylyl cyclase (Nair et al., 1989), experiments with ligand, a functional EGF receptor would also activate propranolol(100 nM) demonstrated that the P-adrenothe intrinsic protein tyrosine kinase activity and phos- receptor antagonist did not alter the actions of EGF on phorylate membrane proteins. Figure 7A shows that adenylyl cyclase (data not shown). although several proteins were phosphorylated on tyThe effects of EGF on cAMP accumulation in cardiorosine residues, the most prominent increase in ty- myocytes are smaller than those reported by us in inrosine phosphorylation in response to EGF was associ- tact perfused rat hearts (Nair et al., 1989). However, ated with a 170 kDa protein. Since EGF increased this difference may be a reflection of the fact that for tyrosine phosphorylation of the 170 kDa protein in several days, cardiomyocytes are exposed to serum membranes isolated from both noncardiomyocytes as components enriched with various growth factors, well as cardiomyocytes (Fig. 7A), it is inferred that which may decrease EGF responsiveness. Nonetheless, functional EGF receptors are present in both cardiac our finding that EGF stimulates cAMP accumulation myocytes and noncardiomyocytes. The contention that in cardiomyocytes but not in noncardiomyocytes sugEGF receptors are present in both cell types is further gests that cAMP mediates the chronotropic effects of supported by the finding that as monitored by Western EGF in cultured ventricular myocytes (Rabkin et al., analysis, the anti-EGF receptor antibody was able to 1987). Moreover, since in the experiments described immunoprecipitate EGF receptor protein (170 kDa) above, the phosphodiesterase inhibitor, IBMX was from both cardiomyocytes and noncardiomyocytes present, the data strongly suggest that EGF-mediated (Fig. 7B). stimulation of adenylyl cyclase is responsible for the Previously we demonstrated that EGF stimulates increase in cellular cAMP and this contention is consiscardiac adenylyl cyclase via a G protein (5),namely, G, tent with our previous reports (Nair et al., 1989, 1990) (Nair et al., 1990). Therefore, to investigate at the cel- concerning direct measurements of cardiac adenylyl cylular level the role of G, in EGF-elicited stimulation of clase. Interestingly, despite the ability of EGF to stimcAMP content in cardiomyocytes, experiments were ulate tyrosine phosphorylation in both cardiomyocytes performed with cholera toxin. Cholera toxin ADP-ribo- as well as noncardiomyocytes and the presence of imsylates and maximally activates G, (Gillman, 1984) munoreactive EGF receptors on membranes of both cell such that its activity cannot be further modulated by types, the effects of EGF on cellular cAMP accumulaagonists. The data in Table 1 show that cholera toxin tion were restricted to cardiomyocytes. Thus it appears treatment of cardiomyocytes increased cellular cAMP that the coupling between the EGF receptors and adecontent by twofold and abolished the ability of EGF to nylyl cyclase second messenger system requires the stimulate adenylyl cyclase. One explanation for these participation of component(s), which is(are) present in findings is that when adenylyl cyclase is maximally some cell types b-ut not in others. In this respect i t is stimulated by cholera toxin, EGF cannot further modu- noteworthy that EGF decreases glucagon- and proslate its activity. However, since employing a n antibody taglandin El-stimulated cAMP accumulation in hepaagainst G,, we have previously shown that G, mediates tocytes and fibroblastic cells, respectively, without althe effects of EGF on adenylyl cyclase (Nair et al., tering basal cAMP levels (Anderson e t al., 1979; Bosch 19901, these findings support the notion that G, medi- et al., 1986). Whether or not the attenuation of agonist ates the effects of EGF on adenylyl cyclase. Similarly, elevated cAMP accumulation in hepatocytes and fibroconsistent with our previous demonstration that blasts by EGF is due to modulations of adenylyl cyclase P-adrenoreceptor activation is not involved in either or cAMP phosphodiesterase(s) is currently unknown.
566
YU ET AL.
TABLE 1. Effect of cholera toxin treatment of cardiomyocytes on the abilitv of EGF to increase cAMP accumulation1 ~
Control CTX cAMP accumulation pmol/ 10 min/mg protein
Condition
+
No addition EGF (100 nM)
14.3 1.3 22.4 5 1.2*
+
31.2 29.2
+ 4.5* + 4.1 (n.s.)
‘Cardiomyocytes weretreated either with or withnuticontrol)cholera toxin (CTX, 15 ng/ml for 15 h). Experimental conditions were similar to those for experiments in Figures 5 and 6 except that the cholera toxin was added 9 h after placing cells in serum-freemedium. Thecells were then challenged with 100 nM EGF for IOmin and cAMP content was measured as described in Materials and Methods. *Valuespresentedaremean S.E.M. offourdeterminations;p
