UTP Transactivates Epidermal Growth Factor Receptors and Promotes Cardiomyocyte Hypertrophy Despite Inhibiting Transcription of the Hypertrophic Marker Gene, Atrial Natriuretic Peptide*

In neonatal rat ventricular myocytes, activation of receptors that couple to the Gq family of heterotrimeric G proteins causes hypertrophic growth, together with expression of “hypertrophic marker” genes, such as atrial natriuretic peptide (ANP) and myosin light chain 2 (MLC2). As reported previously for other Gq-coupled receptors, stimulation of α1-adrenergic receptors with phenylephrine (50 μm) caused phosphorylation of epidermal growth factor (EGF) receptors as well as activation of ERK1/2, cellular growth, and ANP transcription. These responses depended on EGF receptor activation. In marked contrast, stimulation of Gq-coupled purinergic receptors with UTP caused EGF receptor phosphorylation, ERK1/2 activation, and cellular growth but minimal increases in ANP transcription. UTP inhibited phenylephrine-dependent transcription from ANP and MLC2 promoters but not transcription from myoglobin promoters or from AP-1 elements. Myocardin is a muscle-specific transcription enhancer that activates transcription from ANP and MLC2 promoters but not myoglobin promoters or AP-1 elements. UTP inhibited ANP and MLC2 responses to overexpressed myocardin but did not inhibit responses to c-Jun, GATA4, or serum response factor, all of which are active in nonmuscle cells. Thus, UTP inhibits transcriptional responses to phenylephrine only at cardiac-specific promoters, and this may involve the muscle-specific transcription enhancer, myocardin. These studies show that EGF receptor activation is necessary but not sufficient for ANP and MLC2 responses to activation of Gq-coupled receptors in ventricular myocytes, because inhibitory mechanisms can oppose such stimulation. ANP is a compensatory and protective factor in cardiac hypertrophy, and mechanisms that reduce its generation need to be defined.

Hearts respond to a range of pathological insults by increasing muscle mass, caused primarily by hypertrophic growth of the cardiomyocytes and by specific changes in the program of gene transcription. Hypertrophy in response to pressure overload involves heterotrimeric G proteins of the G q/11 class (1) and activation of G q is sufficient to cause hypertrophy in vivo (2), as well as in cell models (3). In keeping with this, activation of G q -coupled receptors including ␣ 1 -adrenergic receptors (␣ 1 -AR), 1 ET A receptors or AT 1 receptors on neonatal rat ventricular myocytes (NRVM) causes increased growth of the cardiomyocytes, as well as increasing transcription from "hypertrophic" marker genes, such as ANP and MLC2 (3). Signaling pathways associated with increased protein synthesis, and hence cell size, are known to be partly independent of those causing re-expression of ANP and MLC2, but both can be activated by G q (4). Thus, it might be expected that any receptor that activates G q would initiate hypertrophic growth and marker gene expression in NRVM. In agreement with this, we have recently reported that activation of G q -coupled purinergic receptors with UTP increases cardiomyocyte size and stimulates downstream MAP kinase activation, in an apparently similar manner to activation of ␣ 1 -AR with phenylephrine (PE) (5).
In addition to stimulating phospholipase C␤ (PLC␤) isoforms, G q -coupled receptors can induce transactivation of certain growth factor receptors (6). In NRVM, exogenous AT 1 receptors mediate phosphorylation of epidermal growth factor receptors (EGFR), and this appears to be critical for subsequent growth responses (7). Direct activation of G q by means of an activating toxin (from Pasteurella multocida) has similar effects, implying G q mediation (8). However, whether all G qcoupled receptors transactivate EGFR and whether EGFR activation is critical for growth or gene transcription are still not certain. Possibly of more importance, it is not known whether EGFR activation is sufficient for any of these responses and whether downstream responses to EGFR phosphorylation vary depending on the nature of the stimulus.
In the current studies, we have compared responses to stimulation of purinergic receptors with UTP to stimulation of ␣ 1 -AR with PE. Both factors caused phosphorylation of EGFR, and both activated ERK1/2 via mechanisms involving EGFR.
However, ␣ 1 -AR activation caused increased transcription of ANP, whereas UTP was relatively ineffective, even though it increased cell growth. Recent studies have presented evidence that ANP and the related peptides BNP and CNP have beneficial effects on the hypertrophying cardiomyocytes by limiting cellular growth responses (9). Thus, the finding that purinergic receptors initiate cardiomyocyte growth without substantial increases in ANP transcription suggests the possibility of a more malignant type of hypertrophy (10). The current manuscript investigates the reason for this difference in ANP response to ␣ 1 -AR agonists versus purinergic stimulation.

EXPERIMENTAL PROCEDURES
Culture of Neonatal Cardiomyocytes-NRVM cultures were prepared from 1-3-day-old Sprague-Dawley rat pups as previously described (11). The cells were preplated twice for 30 min each to remove nonmyocytes and left to attach for 18 h in DMEM, 10% fetal calf serum, 0.1 mM bromodeoxyuridine, 50 units/ml penicillin G, and 50 g/ml streptomycin sulfate onto uncoated dishes. The medium was then replaced with a defined serum-free medium consisting of DMEM, 10 g/ml human insulin, 10 g/ml bovine apo-transferrin, 0.1 mM bromodeoxyuridine, 50 units/ml penicillin G, 50 g/ml streptomycin sulfate, and 125 g/ml fungizone. Bromodeoxyuridine was omitted after 3 days.
Measurement of NRVM Hypertrophy-NRVM (400 cells/mm 2 ) were treated with 100 M UTP or 50 M PE (plus 1 M propranolol) with fresh additions of UTP every 8 h. After 48 h the cells were harvested into buffer containing 50 mM Tris-HCl, 100 mM NaCl, 2 mM EDTA, and 2 mM EGTA, pH 7.7, and the samples were assayed for protein (12) and DNA (13).
EGF Receptor Phosphorylation-NRVM plated on 9-cm dishes were washed with ice-cold saline and harvested by rocking in 800 l of lysis buffer, followed by centrifugation to remove particulate matter. Equal quantities of protein (1.5-3 mg) were diluted to 2.5 ml with Dulbecco's phosphate-buffered saline (JRH Biosciences), pH 7.0, and immunoprecipitated overnight with anti-EGFR antibody (1/200) at 4°C. Immunocomplexes were harvested using protein A-Sepharose, separated by SDS-PAGE on 10% gels, and probed with antibodies to phosphotyrosine and EGFR. Similar protocols were used for ErbB2 and ErbB4 phosphorylation.
Measurement of PLC Activity-NRVM were labeled with [ 3 H]inositol (5 Ci/ml), washed with unlabeled medium, and pretreated with 10 mM LiCl in DMEM for 10 min prior to the addition of agonists. [ 3 H]Inositol phosphates (InsPs) were extracted with ice-cold 5% trichloroacetic acid, 2.5 mM EDTA, 5 mM phytic acid, and the supernatants were subsequently treated with a 1:1 mixture of 1,1,2-trichlorotrifluoroethane:trin-octylamine to remove remaining trichloroacetic acid. The aqueous phase containing [ 3 H]InsPs was subjected to chromatography on Dowex 1 columns for total [ 3 H]InsP measurement.
Treatment of Data-The differences between treatment groups were assessed by one-way analysis of variance with Tukey's test for multiple comparisons and accepted as statistically significant at p Ͻ 0.05. Unless otherwise noted, the results shown are from representative experiments performed in triplicate, which were performed in independent NRVM preparations at least three times.
Materials-Fetal calf serum used during NRVM isolation was specially selected for low endotoxin and obtained from the Commonwealth Serum Laboratories (Parkville, Australia). DMEM, Hepes, and other materials for the preparation of cell culture solutions and media were cell culture grade, obtained from Sigma, and dissolved in milliQ H 2 O. Other reagents were obtained from Sigma or BDH/AnalaR and were of analytical reagent grade. The antibodies were from the following sources: phospho-ERK1/2, ERK1/2, and ErbB2 were from Cell Signaling Technology; ErbB1 and phosphotyrosine (clone 4G10) were from Upstate Biotechnology, Inc.; and ErbB4 was from Santa Cruz.

UTP and PE Cause Growth of NRVM, but Only PE Substantially Increases Transcription of ANP-
We have recently reported that UTP causes hypertrophic growth of NRVM via G qcoupled P 2 Y-purinergic receptors (5). In the current study we examined this hypertrophic response with a view to defining critical signaling pathways. NRVM were treated with the purinergic receptor agonist UTP (100 M) or the ␣ 1 -AR agonist PE (50 M, plus 1 M propranolol), for 48h. Total cell lysates were prepared and assayed for DNA and protein. As shown in Fig. 1, UTP and PE caused similar increases in the protein/DNA ratio. Furthermore, the EGFR kinase inhibitor AG-1478 (5 M) reduced responses to both agonists, implying involvement of transactivated EGFR.
We next compared the abilities of UTP and PE to induce ANP transcription, a well established marker for cardiomyocyte hypertrophy. NRVM were treated with UTP (100 M) or shown are the amounts of protein/well relative to DNA g/g (means Ϯ S.E., n ϭ 6). **, p Ͻ 0.01 relative to no additions (NA). † †, p Ͻ 0.01 relative to no inhibitor. The experiment was performed three times with similar results. PE (50 M, plus 1 M propranolol) for 48 h. RNA was extracted, and ANP mRNA was quantified by RNase protection. PE caused a substantial increase in ANP transcription, but in contrast, UTP caused only a small increase when added alone and inhibited the ANP mRNA responses to PE when the two agonists were added together (Fig. 2).
PE and UTP Phosphorylate EGF Receptors-Sustained intense activation of G q in NRVM, either by stimulating overexpressed AT 1 receptors (7) or by treatment with an activating toxin (recombinant P. multocida toxin, rPMT) (8), causes phosphorylation of EGF receptors (ErbB1, EGFR) via a transactivation process and initiates downstream hypertrophic signaling pathways. Transactivation of EGFR by endogenously expressed receptors is less well documented. AG-1478 had inhibited growth responses to both UTP and PE, and we next measured EGFR phosphorylation in response to activation of endogenously expressed purinergic receptors with UTP or ␣ 1 -AR with PE.
NRVM were treated with UTP (100 M) or PE (50 M, plus 1 M propranolol) for 10 min. The lysates were prepared, immunoprecipitated with anti-EGFR antibody, subjected to SDS-PAGE, transferred, and probed with anti-phosphotyrosine antibodies. PE and UTP both caused transactivation of EGFR (ErbB1), as shown by increased tyrosine phosphorylation (Fig.  3). Neither UTP nor PE caused detectable phosphorylation of ErbB2 or ErbB4, even though these responded to neuregulin (NRG␤2). Thus, purinergic receptors and ␣ 1 -AR transactivate ErbB family members in an apparently similar manner. UTP and PE Activate ERK1/2 via EGFR Transactivation-We next examined whether downstream signaling from these two receptor classes required EGFR transactivation. NRVM were pretreated with AG-1478 (5 M) for 10 min and then UTP (100 M) or PE (50 M plus 1 M propranolol) was added for a further 10 min. The cells were lysed, and the extracts were subjected to SDS-PAGE, transferred, and probed with antibodies to phosphorylated ERK1/2. UTP and PE both caused activation of ERK1/2, and responses to both agonists were substantially inhibited by AG-1478 (Fig. 4). We have previously reported that neither of these agonists caused detectable activation of JNKs or p38 MAPKs in these cells under the conditions of our experiments (5).
PE Increases Transcription of ANP via G q and EGFR Transactivation-Given the failure of UTP to induce a robust ANP transcriptional response, even though UTP couples to G q and transactivates EGFR, we next examined the ANP response to PE and its dependence on G q and EGFR. To do this, we transfected NRVM with an ANP-luciferase construct, together with other plasmids expressing G␣ q or a dominant negative EGFR (EGFR-DN) and measured luciferase activity as an index of ANP gene transcription. PE (50 M plus 1 M propranolol) was added, and the cells were harvested after a further 20 h. Overexpression of G␣ q increased ANP transcription and markedly enhanced responses to PE. PE responses were inhibited by overexpression of EGFR-DN or by the addition of 5 M AG-1478 (Fig. 5). Neither the EGFR-DN nor AG-1478 inhibited ANP responses to overexpression of MKK7, the immediate activator of JNK (Fig. 5, inset), confirming that the inhibition by EGFR-DN and AG-1478 was not nonspecific nor a consequence of cytotoxicity. The experiments show that the ANP response to PE involves G q and EGFR, even though UTP, which also activates G q and EGFR, is relatively ineffective in increasing ANP transcription and inhibits responses to PE. Similar results were obtained when MLC2 transcription was measured using an MLC-luciferase construct (data not shown). UTP Inhibits ANP and MLC Transcription when Activated by PE or G q , but Not Early Signaling Responses-We next investigated the UTP inhibition of PE-activated transcription from ANP and MLC2 promoters. NRVM were transfected with ANP-luciferase or MLC2-luciferase together with G␣ q or blank vector and subsequently stimulated with 50 M PE plus 1 M propranolol, and then ANP and MLC2 responses were measured after 20 h. UTP inhibited ANP and MLC2 responses to PE, G␣ q , and PE plus G␣ q , as shown in Fig. 6.
Such inhibition of PE responses might reflect cross-desenstization of ␣ 1 -AR by purinergic receptors. To address this possibility, we examined effects of UTP on immediate signaling responses to PE, PLC activation, and ERK1/2 phosphorylation. NRVM were labeled with [ 3 H]inositol, pretreated with LiCl, and then stimulated for 20 min with 50 M PE (plus 1 M propranolol), with 100 M UTP, or with the two agonists together. InsPs were extracted and quantified. Both PE and UTP activated PLC, as demonstrated by increased [ 3 H]InsPs and the two agonists produced an additive response when added together (Fig. 7, upper panel). UTP also did not inhibit the ERK1/2 response to PE (Fig. 7, lower panel). Thus, UTP exerts an inhibitory effect on ANP and MLC2 responses but does not inhibit signaling responses proximal to the receptor. This argues against cross-desensitization of ␣ 1 -ARs by purinergic receptor activation. The lack of inhibition of proximal events, as well as the finding that UTP inhibited ANP and MLC2 responses even in the presence of overexpressed G␣ q , also shows that the purinergic inhibition of PE transcriptional responses is not due to competition for available G q .
UTP Does Not Inhibit Responses from an NFAT-sensitive Myoglobin Promoter or from AP-1 Response Elements-PE activates promoters other than ANP and MLC2 in cardiomyocytes, and we next examined whether UTP inhibited all of these responses. Similar experiments to those described above were performed using myoglobin (Mb)-luciferase or 2xTREluciferase, constructs responsive to NFAT or AP-1 elements, respectively. NRVM were transfected with Mb-luciferase or 2xTRE-luciferase, subsequently treated with 50 M PE (plus 1 M propranolol) or with 100 M UTP and luciferase activity measured after 20 h. Both PE and UTP caused increased transcription from Mb-luciferase and 2xTRE-luciferase reporter constructs, and the response to UTP was quantitatively similar to the PE response. Furthermore, UTP did not inhibit responses to PE, in contrast to findings with ANP-and MLC2luciferase constructs (Fig. 8). Thus, the inhibitory effect of UTP is restricted to cardiac-specific promoters, even though NFAT and AP-1 elements are activated by hypertrophic stimuli such as PE (20,21).
UTP Selectively Inhibits ANP and MLC Responses to Myocardin-The data had shown that UTP selectively inhibited ANP and MLC2 responses to PE but not events close to the receptors. This pointed to a UTP inhibited step beyond immediate signaling responses. The failure of UTP to inhibit PE responses at myoglobin promoters or AP-1 elements suggested selective inhibition of cardiac-specific promoters. Myocardin is a newly described transcription enhancer that confers muscle specificity on the ubiquitous transcription factor, SRF (17). We next investigated whether UTP inhibited ANP and MLC2 responses to overexpression of myocardin. NRVM were transfected with ANP-or MLC-luciferase constructs together with plasmids expressing myocardin. UTP (100 M) was added, and luciferase activity was measured after 20 h. Myocardin caused substantial activation of ANP transcription at all concentrations tested, with very high activation observed at the highest concentration of added DNA, as reported previously (17). UTP inhibited ANP transcription at all of these myocardin doses (Fig. 9A). For comparison, the experiments were also performed with overexpressed GATA4, c-Jun ϩ JNK, and SRF, transcription factors not restricted to muscle-specific genes. All of the transcription factors increased ANP transcription, but in contrast to myocardin, the responses were not inhibited by UTP (Fig. 9B). Responses to myocardin were synergistic with the PE response, pointing to a role for myocardin in the ANP response to PE (Fig. 9C).
Similar experiments were performed using MLC-luciferase. As with ANP-luciferase, GATA4, c-Jun ϩ JNK, SRF, and myocardin all activated MLC2 transcription, but only the response to myocardin was inhibited by UTP. As described previously (17), the maximal MLC2 response to myocardin was lower than the ANP response (Figs. 9A and 10A). Myocardin substantially increased PE-induced MLC2 transcription (Fig. 10B).
Myocardin has been reported to activate muscle-specific promoters exclusively (17). In the current study we confirmed that myocardin did not activate Mb-luciferase or 2xTRE-luciferase constructs. NRVM were transfected with Mb-luciferase or 2xTRE-luciferase, 50 M PE together with 1 M propranolol was added, and luciferase activity measured after 20 h. Over-expression of myocardin did not increase transcription from either promoter and did not enhance the activation by PE. In contrast, Mb-luciferase expression was increased by the other transcription factors tested, and PE further enhanced activation (Fig. 11).
Thus UTP inhibits PE-induced transcription of ANP and of MLC2, but transcription from myoglobin promoters or from AP-1 elements was not inhibited. UTP selectively inhibited ANP and MLC2 responses to myocardin, and myocardin in turn selectively activates ANP and MLC2 promoter elements. Given that the PE response is enhanced by myocardin, it is possible that the suppression of PE induced ANP and MLC2 transcription by UTP is related to its effect on the musclespecific transcription enhancer, myocardin.

DISCUSSION
Activation of G q by appropriate receptor agonists, by G␣ q overexpression, or by treatment with the G q -activating toxin from P. multocida induces hypertrophic growth of NRVM that includes increases in cell size as well as re-expression of fetal genes (8,22). Both PE and UTP activate G q -coupled receptors stimulate PLC␤ subtypes, transactivate EGFRs, and phosphorylate ERK1/2 (23) (Figs. 3 and 4). Although other G q -coupled receptors have been shown to transactivate EGFR in NRVM (7), this is the first direct demonstration of transactivation by endogenous ␣ 1 -AR and P 2 Y receptors. Both agonists also cause substantial increases in cardiomyocyte size ( Fig. 1) (5). But, in marked contrast, only PE caused a robust ANP response (Fig.  2), even though our data showed that the ANP response to PE involves G q and EGFR activation (Fig. 5).
Subsequent experiments showed that UTP inhibited ANP responses to PE, consistent with the relative ineffectiveness of UTP as a transcriptional activator of ANP expression (Figs. 2 and 6). We also showed that overexpression of G␣ q did not reduce the UTP inhibition; rather UTP inhibited the G␣ q activation (Fig. 5). This ruled out competition for G q as a mechanism of inhibition of PE responses by UTP. Moreover, UTP did not inhibit immediate responses to PE such as EGFR phosphorylation or ERK1/2 activation, as would be expected from competition for available G q (Fig. 8). Thus, the UTP-inhibited step is subsequent to early signaling responses associated with ANP transcription. Inhibition by UTP was selective for the two cardiac-specific promoters regulating ANP and MLC2 transcription. When experiments were performed using luciferase constructs incorporating AP-1 sites or the myoglobin promoter that includes NFAT-responsive elements, UTP, like PE, was stimulatory (Fig. 7). Furthermore, UTP stimulation of AP-1 elements was additive with rather than inhibitory to the PE response. Thus, the UTP inhibition is selective for particular promoters, apparently those that are cardiac-specific.
The finding that UTP selectively inhibited PE responses at ANP and MLC promoters suggested the likelihood of an effect mediated by transcription factors specific for those promoters. Myocardin is a newly described transcription enhancer that acts selectively on ANP and MLC2 promoters (17) (Figs. 9 -11). Myocardin associates with SRF and confers muscle specificity on the activity of SRF, which, on its own, is functional in nonmuscle cells. Many of the other transcription factors known to bind the ANP promoter, including GATA4 and c-Jun, also activate promoters that are not muscle-specific (24,25). In the current studies, we found that UTP treatment caused a selective inhibition of myocardin-induced ANP and MLC2 transcription. Responses to c-Jun, GATA4, or SRF were not inhibited. This myocardin selectivity potentially explains the selective inhibition of PE-induced transcription from ANP and MLC2 but not from myoglobin or AP-1 elements. The mechanism by which myocardin activity is regulated is not currently known, and a role in PE-mediated responses remains to be established. However, there are a number of factors that argue in favor of myocardin as the target of UTP inhibition of PE-induced ANP and MLC2 transcription. First, UTP treatment depressed transcriptional responses to myocardin. Second, PE responses were inhibited only at myocardin-sensitive promoters. Third, the PE response was markedly enhanced by overexpressed myocardin. Taken together these findings suggest that PE activation of ANP and MLC2 promoters involves myocardin, among other factors, and that UTP treatment reduces PE responses by interfering with myocardin functioning at these promoters. However, it must be stressed that direct evidence is lacking and that it is also possible that there are two separate and independent inhibitory actions of UTP one targeting myocardin and the other some other factor acting beyond immediate signaling responses to PE. It is also likely that other musclespecific factors are involved in the responses described.
Cardiomyocytes express a number of different purinergic receptors of which the P 2 Y2 and P 2 Y4 receptors might be expected to respond robustly to UTP (26). In acute experiments, responses to UTP were mimicked by ATP (5). Thus, ATP enhanced PLC activity and caused EGFR phosphorylation and ERK1/2 activation, but because ATP is cytotoxic to cardiomyocytes with prolonged exposure (5), it is impossible to evaluate effects on transcriptional responses. Similarly, we found suramin, an antagonist at P 2 Y1, P 2 Y2, and P 2 Y4 receptors (26), to be toxic with chronic treatment. Therefore, it has not been possible to define the receptors responsible for the inhibitory action of UTP on ANP and MLv2 transcription.
Pathological insults including chronically heightened blood pressure, myocardial infarction, valve disease, and some congenital defects lead to hypertrophic growth of cardiomyocytes (27,28). In most cases this is associated with increased expression and release of ANP from the ventricular myocytes. Increased plasma ANP would be expected to increase sodium excretion and to cause dilatation of the vasculature, thereby reducing afterload on the heart (29,30). In addition, a recent study has suggested direct beneficial effects of ANP on the cardiomyocytes themselves. Cardiac targeted knockout of ANP receptor type C lead to a worsened prognosis following pressure overload in vivo (9). Thus, a hypertrophic response that induces cellular growth without increased ANP expression might be expected to be particularly detrimental to the heart (10). It is not obvious how UTP would cause sustained myocardial responses in vivo, unless there is substantial local release from the heart under some circumstances. ATP, but not UTP, is released from cardiac sympathetic nerves during stimulation (31) and might cause chronic activation of purinergic receptors under some conditions. It is possible that the cytotoxic effect of ATP, observed in cell culture, is due to degradation products accumulating in the culture medium and that in vivo repeated release and clearance of ATP would allow for sustained growth responses, but this remains to be established. In any case, we have described an inhibition of ANP transcriptional responses by UTP, possibly involving the newly described transcription factor, myocardin. Further studies are needed to establish exactly how myocardin responses are regulated, whether myocar- din is involved in ␣ 1 -AR-mediated responses, and how inhibition is induced by chronic exposure to UTP.