Estrogen Inhibits Cardiomyocyte Hypertrophy in Vitro

Isolation and Culture of Rat Neonatal Cardiomyocytes

Myocytes were isolated from the hearts of 1-to-3-day-old rats or pregnant female rats (Charles River Laboratories) using a cardiomyocyte isolation kit (Worthington), according to the manufacturer’s instructions. The cells were incubated in Dulbecco’s modified Eagle’s medium/F-12 supplemented with 10% fetal bovine serum, 1× ITS (insulin/transferrin/selenium) (Sigma) antibiotic and antimycotic, and 10 μ g/ml fibronectin (to aid adherence).

Measurement of Protein Synthesis, Protein Secretion, and Cell Area

After 24 h in media without serum, the cells were treated with either 100 nmol/liter AngII or 10 nmol/liter ET-1 or no treatment, in the presence or absence of 10 nmol/liter E₂. New protein synthesis, a marker of hypertrophy, was determined using [³H]leucine uptake. The cells were cultured as above with the addition of 1 μCi/ml [³H]leucine for 48 h. After 48 h, the cells were rinsed in Dulbecco’s modified Eagle’s medium/F-12 and lysed with Trizol. The protein and RNA were extracted per the manufacturer’s instructions. The protein fraction was counted in a scintillation counter, and the mean ± S.E. of the counts was plotted. Differences between treatments were compared for combined experimental data, and the results were analyzed by analysis of variance plus Schefe’s test (p < 0.05 was significant). ANP protein secretion into the cell incubation medium was determined after 24 h of incubation with AngII ± E₂ or E₂ alone. The medium was concentrated, and Western blot using ANP antisera (Phoenix Pharmaceutical) was performed. Cell surface area was determined under various conditions after culturing on coverslips and imaging by confocal microscopy. This was done at 40–50% confluency of the cells. Surface area was quantified by imaging to the complete boundary of 30 individual cells/condition using ImageJ software (National Institutes of Health). From this, a mean ± S.E. value was calculated for the surface area in each condition. The planar aspect and deviations of individual cells in this respect (z axis) were determined and adjusted.

RT-PCR

Analysis of gene expression in the cardiomyocytes was investigated using RT-PCR. PCR was run with primer sets for hypertrophic genes, common markers of cardiac hypertrophy in vivo and in vitro, or MCIP1. GAPDH was used as a standard. The primer sequences were as follows: skeletal muscle actin forward primer, AAGGACCTGTATGCCAACAA, and reverse primer, GAGAGGATTCGTCGTCCTGA; MCIP1, GACTGGAGCTTCATTGACTGCGAGA and AAGGAACCTACAGCCTCTTGGAAAG; GAPDH, AGCCACATCGCTCAGAACAC and GAGGCATTGCTGATGATCTTG; ANP, CGTGCCCCGACCCACGCCAGCATGGGCTCC and GGCTCCGAGGGCCAGCGAGCAGAGCCCTCA; and BNP, GGGCGCTCCTGCTCCTGCTCTTC and ACACCTGTGGGACGGGGGCTCTC.

Calcineurin Activity and NF-AT Transactivation

Phosphatase activity was determined using a BioMol kit per the manufacturer’s instructions. Cells were treated as above for 24 h and lysed, and the proteins were isolated. Calcineurin activity was determined by running duplicates of the samples in the presence of EGTA and subtracting this value from the total phosphatase activity. All samples were split into 6 wells (3 for total phosphatase and 3 for relative calcineurin), and the experiments were carried out three times.

Activity of the NF-AT transcription factor was determined by transfecting cardiomyocytes with a luciferase reporter linked to a promoter with three NF-AT sites (a gift from Dr. Chris Hughes, University of California, Irvine), using Lipofectamine Plus (Invitrogen). A control vector, TK-luc, contained the promoter sequence without NF-AT sites. Transfection efficiency was determined by co-transfecting a GFP construct. Cells were lysed, added to (200 μl) detection buffer, and immediately read on a luminometer. The luciferase values indicated are the mean of six determinations/condition.

NF-AT translocation from the cytoplasm to the nucleus was determined over time by Western blot. Cell fractions were obtained by centrifugation as described previously (9).

Nuclear Run-on and mRNA Stability

Nuclear run-on studies were performed as follows. Nuclei were isolated, followed by sucrose gradient centrifugation. The nuclei were transferred to 2× reaction buffer (10 mol/liter Tris-Cl, 5 mmol/liter MgCl₂, 0.3 mol/liter KCl) with 0.5 mmol/liter nucleotides, and 0.3 μmol/liter radiolabeled [³²P]UTP (>3000 μCi/mmol/liter). The radiolabeled RNA was isolated to probe a nylon blot with equal amounts of cDNA for GAPDH and MCIP1. Hybridizations were performed in Ultrahyb (Ambion) at 42 °C overnight and then washed twice in 2× SSC at 65 °C.

Stability assays were performed on MCIP1 mRNA by incubating the cardiomyocytes with/without E₂ for 6 h and then blocking transcription after 6 h with 50 nmol/liter 5,6-dichlorobenzimidazole (DRB). Total RNA was isolated over 8 h after the addition of the DRB. RT-PCR using primers specific to MCIP1 was used to determine the degree of mRNA stability. GAPDH primers were used as an internal control. The intensities of the MCIP1 bands were normalized to GAPDH and plotted for each of the time points.

RNA Interference

siRNA molecules were designed using Qiagen’s website and transfected into cardiomyocytes with Oligofectamine. A fluorescein isothiocyanate-labeled nonsense siRNA molecule controlled specificity. Transfected cells were recovered for 24 h and treated with hypertrophic agents with/without estrogen as before. Knockdown in gene expression was identified after isolating the RNA and performing RT-PCR (MCIP1) or protein blots (AKT) 48 h after expression of the siRNAs. Preliminary time courses of protein knockdown were carried out, leading to selection of 48 h of post-transfection for experiments (data not shown).

Kinase Activity and Nitric Oxide Assays

ERK and AKT activities were determined at the times shown by methods published previously (8, 9). PKC activity was determined by immunoprecipitating PKC protein with antibody H-300, 24 h after incubation of the cardiomyocytes with AngII ± E₂. This antibody against the C terminus of the kinase protein (Santa Cruz Biotechnology) recognized most PKC family members. Immunoprecipitates were used for an in vitro tube assay, utilizing histone-H1 as substrate for PKC activity. Kinase proteins were shown below the activities, as controls. Nitric oxide (NO) was measured as the conversion of [³H]arginine to citrulline by the cardiomyocyte, as we described previously (10).

Presence of ER in Cardiomyocytes

Isoforms of ER were determined in whole cells by confocal microscopy, using techniques we published previously (9). The presence of ER isoforms at the plasma membrane of neonatal cardiomyocytes was determined by Western blot, after separating membrane proteins by native gel electrophoresis, as we described previously (11). Antibodies for ERα and ERβ, respectively, were from Santa Cruz Biotechnology and Zymed Laboratories Inc., and the cardiac myosin heavy chain antibody was from Abcam.

Estrogen Inhibits AngII or ET-1-induced Cardiomyocyte Hypertrophy

Treatment of cardiomyocytes with the hypertrophic peptides AngII or ET-1 stimulated incorporation of leucine into newly formed protein (Fig. 1A, left). Co-incubation with 10 nm E₂ caused as much as an 80% loss of the stimulated protein synthesis. The inhibition by sex steroid occurred in a dose-related fashion, at a concentration as low as 1 nm E₂, and was substantially reversed by ICI182780, an ER antagonist (Fig. 1A, right). We also determined the expression of a classical hypertrophic marker, skeletal muscle actin (Fig. 1B). RT-PCR showed a clear increase in actin mRNA levels in cells treated with AngII or ET-1. However, in cells treated with either peptide plus E₂, the increases in mRNA expression were significantly smaller.

Cardiomyocyte cell size defines hypertrophy in vitro. We found that the mean cell area was substantially stimulated by the hypertrophic peptides and was reduced ~65% in E₂-treated cells (Fig. 1C). ET-1 induced an increase in cell size and actin staining compared with control, and both parameters were reversed by concomitant E₂ exposure (Fig. 1D). E₂ alone had no effect on these parameters (data not shown). Because similar results were observed regardless of the hypertrophic stimulus (either AngII or ET-1), and because AngII induces hypertrophy in part through ET-1 (12), many subsequent studies were performed with one of the two stimuli.

Estrogen Stimulates ANP and BNP Expression

Several investigators have reported that E₂ stimulates ANP gene expression, correlating to sex steroid prevention of hypertrophy (13, 14). It is also well known that ANP and BNP expression in the heart is up-regulated by a variety of hypertrophic factors, and the natriuretic peptides are anti-hypertrophic (15). Thus, natriuretic peptide stimulation may be a compensatory response that limits the extent of cardiac enlargement, because severe hypertrophy is ultimately maladaptive. We therefore asked whether E₂ stimulates ANP and BNP, and what role this may have on the effects of E₂.

We first determined the effects of AngII, E₂, or both on ANP and BNP gene expression at 24 h. We found that individually, AngII and E₂ each stimulated the gene expression of ANP and BNP (Fig. 1E, left). We also found that AngII and E₂ each stimulated ANP and BNP secretion into the incubation medium (Fig. 1E, right). We then asked whether the ability of E₂ to stimulate ANP production impacted the anti-hypertrophy effects of the sex steroid. We determined leucine incorporation in the setting of AngII, E₂, or both, with or without ANP antibody added to the culture medium.

As seen in Fig. 1A, right, E₂ inhibited AngII-induced protein synthesis. This was partially prevented by ANP anti- body, at a concentration of antisera that fully blocked the anti-hypertrophic effect of the exogenous added ANP, 100 nm (Fig. 1A, right). ANP antibody also slightly augmented AngII-induced leucine incorporation.

Estrogen Prevents Calcineurin Activation

Cardiomyocyte hypertrophy often involves calcineurin, inducing NF-AT translocation to the nucleus (16). We found that the calcineurin inhibitor, FK506, caused a significant decrease in ET-1-stimulated leucine incorporation (Fig. 2A). These results suggest that hypertrophic effects of ET-1 are mediated in part through calcineurin.

To investigate a possible interaction between calcineurin and E₂, we performed assays to determine overall phosphatase activity and calcineurin activity specifically. Cardiomyocytes treated with ET-1 had large increases in both total phosphatase and calcineurin activity (Fig. 2B). In cells treated with ET-1 and E₂, there was a small reduction in overall phosphatase activity and a large reduction in calcineurin activity. This suggests that the E₂ effect is relatively specific to this phosphatase and not to overall activity stimulated by ET-1. The ER antagonist, ICI182780, inhibited the effect of E₂ (Fig. 2B) but had no action by itself (the latter data is not shown). This precludes a non-ER action of E₂.

Downstream events in the calcineurin pathway were analyzed by investigating the localization and activity of the transcription factors, NF-ATc3 and NF-ATc4. We first determined that NF-ATc3 translocated to the nucleus maximally after 60 min of exposure to AngII, and the effect persisted for an additional hour (Fig. 2C, left). The cytoplasmic pool of this transcription factor was much more abundant and did not appreciably diminish. This suggests that a small but important subpopulation of NF-AT participates in the hypertrophic response. We also determined that E₂ significantly blocks the AngII-induced increase in NF-ATc3 translocation to the nucleus (Fig. 2C, right). Similar results were found for NF-ATc4 (data not shown).

These results suggested that E₂ might block AngII-induced NFAT transcriptional activity. To determine this, cardiomyocytes were transfected with a 3xNF-AT-luc plasmid (which has three NF-AT-binding sites linked to the luciferase gene) or the promoter-less vector and treated with AngII and/or E₂. AngII-treated cells demonstrated increased luciferase activity, compared with cells expressing the promoter-less vector or to cells transfected with the NF-AT reporter, but were not exposed to angiotensin (control) (Fig. 2D). No significant change in activity was observed when cells were treated with E₂ alone. In cells treated with both AngII and E₂, ~75% of the increased luciferase activity observed in the AngII alone cells was lost. Thus, E₂ prevents the stimulation of NF-AT transcriptional activity, related in part to preventing nuclear translocation of the transcription factor.

MCIP1 Is Induced by Estrogen to Down-regulate Calcineurin Activity

How might E₂ oppose the calcineurin-regulated hypertrophic pathway? Perhaps this involves the calcineurin inhibitor MCIP1 (17, 18). To determine this, we carried out RT-PCR. Cardiomyocytes treated with AngII or ET-1 showed low basal levels of MCIP1 expression, insignificantly above base line (Fig. 3A). However, in E₂-treated cells, abundant MCIP1 levels were detected, and AngII or ET-1 did not affect this. The PI3K inhibitor LY294002 abrogated the increased MCIP1 expression in cardiomyocytes because of estrogen. These results extend our previous findings in endothelial cells (8) and suggest that this estrogen action in cardiomyocytes occurs in part through PI3K.

Downstream of PI3K, AKT activation may play a role in estrogen induction of MCIP1. To investigate this we first determined the ability of AngII, E₂, or both to modulate AKT activation (reflected by Ser-473 phosphorylation). We found that at 30 min of incubation, either E₂ or AngII stimulated AKT activation, and there was no significant effect from combining both (Fig. 3B). However, at 24 h the AngII effect continued; E₂ alone caused an insignificant activation of AKT, and the steroid significantly inhibited AngII-induced activation. The effects of E₂ on MCIP1 and many of the hypertrophic parameters were determined at 24 h, indicating it is unlikely that E₂ is acting through AKT. To test this further, we examined the effects of silencing AKT (via siRNA). In this setting, we also did not find a significant effect on E₂-stimulated MCIP1 (Fig. 3C). These results indicate that E₂-induced PI3K signaling to MCIP up-regulation does not occur through AKT.

We then investigated whether E₂ stimulates MCIP1 gene expression by transcriptional transactivation or by increasing the stability of the message. To determine this, we performed nuclear run-on and mRNA stability assays. Nuclei from cardiomyocytes treated with or without estrogen were isolated, and the new RNA was used to probe a blot of cDNA for GAPDH (control) or MCIP1. Blots hybridized with RNA from cardiomyocytes treated with E₂ had significantly increased expression of MCIP1, compared with control cells (Fig. 3D). GAPDH levels were similar from both pools of transcripts. These results clearly indicate that E₂ stimulates MCIP1 gene transcription.

To investigate stabilization of the transcript, the cells were treated with E₂ as before, but transcription was stopped at 6 h by blocking with 50 μg/ml DRB. At various times after addition of the transcriptional inhibitor, mRNA was isolated and prepared for RT-PCR (Fig. 3E). At the time of addition of the DRB, MCIP1 levels were higher in the estrogen-treated cells compared with the nontreated controls. With the passage of time, estrogen-treated cells maintained relatively high levels of MCIP1, whereas mRNA isolated from the nontreated cells had little detectable MCIP1 by 4 h from the point of addition of the DRB. Together, estrogen stimulates both the transcription of MCIP and the stability of the message.

MCIP1 Underlies E₂ Inhibition of Cardiac Hypertrophy

It is important to determine whether the anti-hypertrophic effects of E₂ were significantly related to stimulating MCIP1. To address this, we designed siRNA molecules and transfected these into cardiomyocytes. Antibodies to MCIP were not available to determine protein knockdown, and RT-PCR was performed on cells treated with control or MCIP1 siRNA (Fig. 4A). Compared with nontreated cells (Fig. 3) or cells transfected with a control GFP siRNA (Fig. 4A), MCIP1 siRNA caused a strong reduction in RNA levels. The MCIP1 siRNA had no significant effect on GAPDH transcripts.

We then evaluated interfering RNA for MCIP1 on cardiomyocyte protein synthesis. Leucine incorporation stimulated by AngII was no different in the setting of either control or MCIP siRNAs (Fig. 4B). E₂ significantly reduced leucine incorporation in cells incubated with both the steroid and AngII in the presence of control siRNA. However, MCIP siRNA prevented this anti-hypertrophic effect of E₂ by 50%, limited by our transfection efficiency of 60%. These results indicate that MCIP induction/function is important in the inhibition of hypertrophy by E₂.

To understand further the role of MCIP1, we treated cardiomyocytes with siRNA and analyzed them for calcineurin activity. In control siRNA-treated cells, E₂ substantially reduced AngII-stimulated calcineurin activity. However, in cells treated with siRNA molecules against MCIP1, the decrease in calcineurin activity because of E₂ was reversed. This indicates an important role of MCIP1 to mediate E₂-down-regulation of calcineurin activity (Fig. 4C) and related hypertrophy.

Regulation of Hypertrophic Signaling through ERK and PKC by E₂

It has been shown previously that AngII and ET-1 enact hypertrophic signaling in part through ERK and PKC activation (19, 20). Both the AngII and ET-1 receptors are G protein-coupled, and G_q-dependent signaling triggers the activation of ERK and PKC, resulting in cardiac hypertrophy (21–23). It is possible that E₂ down-regulates AngII-induced ERK and PKC activation in the cardiomyocytes.

We first asked whether ERK and PKC activity might be relevant to AngII-induced hypertrophy. As seen in Fig. 5A, the ability of AngII to stimulate leucine incorporation was significantly blocked by E₂, calphostin C (PKC inhibitor), or by PD98059 (ERK inhibitor). Thus, multiple signaling pathways contribute to the overall hypertrophic effects of AngII.

We then asked whether E₂ modulates AngII-induced kinase activity. We found that AngII significantly stimulated PKC activity, directed against histone H1 as substrate (Fig. 5B). This was prevented by E₂ and in turn reversed by ICI182780. We further found that FK506 and cyclosporin A each prevented AngII-induced PKC activity. This suggests a cross-talk from calcineurin to PKC activation. Similar results were found for ERK activity (Fig. 5C). Most interestingly, the ability of E₂ to inhibit AngII-induced ERK was only significant at 24 h, when hypertrophic parameters were determined in many of our studies. After only 5 min, both AngII and E₂ individually stimulated ERK kinase activity. Thus, the duration of E₂ exposure is critical to modulate kinase activity in a fashion consistent with its anti-hypertrophic effects.

Does the ability of E₂ to stimulate MCIP1 impact the regulation of PKC and ERK? To answer this, we introduced an siRNA to MCIP1 or GFP (control) into the cardiomyocytes, recovered the cells over 24 h, and then incubated the cells with AngII ± E₂ for 24 h. As shown in Fig. 5D, silencing MCIP1 did not significantly reverse the ability of AngII to stimulate or E₂ to inhibit either kinase activity. This suggests that E₂ acts independently of MCIP1 to down-regulate ERK and PKC activity.

It has been published previously that ANP inhibits ET-1-and AngII-induced ERK and PKC activation (24, 25). Therefore, we asked whether E₂-induced ANP secretion may underlie the inhibition of the two kinases. We found that ANP antibody significantly blocked the ability of E₂ to inhibit AngII-induced PKC or ERK activity (Fig. 5E). Thus, this aspect of the anti-hypertrophic actions of the sex steroid specifically depends on ANP production/secretion.

Finally, it has been shown that NO functions as an anti-hypertrophy factor (26, 27). Here, we demonstrate that NO production is stimulated by E₂, and generation is reduced by AngII (the latter probably mediated through the AT1 receptor). However, the ability of E₂ to stimulate NO is not mediated through MCIP, as shown using RNA interference to this protein (Table I).

Estrogen Receptors in Neonatal and Adult Cardiomyocytes

Our results indicate that E₂ acts through the ER to mediate its anti-hypertrophic actions, indicated specifically using the ER antagonist ICI182780. Furthermore, if our results in neonatal cardiomyocytes are applicable to adult heart, then ER should be present as well. To demonstrate this, we carried out confocal microscopy studies of cultured neonatal and adult cardiomyocytes (Fig. 6A). In both types of cells, we saw the presence of each ER isoform. The distribution was somewhat different in that ERα was predominantly localized in the nucleus, with some cytoplasmic/membrane presence. In contrast, ERβ was predominantly localized to the cytoplasmic/perinuclear areas of the cell, with evidence of membrane localization as well. Few ERβ receptors could be seen in the nucleus of the cardiomyocytes.

Because we showed that E₂ up-regulation of MCIP occurs through PI3K-dependent signaling, this implicates signaling from plasma membrane-localized ER (8). We therefore evaluated the presence of the membrane ER in the cardiomyocytes. ERα and ERβ homodimers were the predominant forms of the receptors in the plasma membrane, upon incubation of the cells with E₂ (Fig. 6B). In the absence of E₂, the monomer for each isoform predominated, and the presence of the ERα/ERβ heterodimer was also seen, augmented by the steroid ligand. Our results are very similar to the presence and forms of ER in membranes from endothelial cells (11). These data support the ability of E₂ to act through ER to oppose cardiac hypertrophy, in part mediated through membrane receptors.