Regression from pathological hypertrophy in mice is sexually dimorphic and stimulus specific

Abstract

Pathological cardiac hypertrophy is associated with increased morbidity and mortality. Understanding the mechanisms whereby pathological cardiac growth can be reversed could be of therapeutic value. Here, we show that pathways leading to regression of pathological cardiac hypertrophy are strongly dependent on the hypertrophic trigger and are significantly modified by sex. Two pathological stimuli causing hypertrophy via distinct pathways were administered to male and female mice: angiotensin II (ANG II) or isoproterenol (Iso). Stimuli were removed after 7 days of treatment, and left ventricles (LVs) were studied at 1, 4, and 7 days. ANG II-treated females did not show regression after stimulus removal. Iso-treated males showed rapid LV hypertrophy regression. Somewhat surprisingly, RNAseq analysis at day 1 after removal of triggers revealed only 45 differentially regulated genes in common among all the groups, demonstrating distinct responses. Ingenuity pathway analysis predicted strong downregulation of the TGFβ1 pathway in all groups except for ANG II-treated females. Consistently, we found significant downregulation of Smad signaling after stimulus removal including in ANG II-treated females. In addition, the ERK1/2 pathway was significantly reduced in the groups showing regression. Finally, protein degradation pathways were significantly activated only in Iso-treated males 1 day after stimulus removal. Our data indicate that TGFβ1 downregulation may play a role in the regression of pathological cardiac hypertrophy via downregulation of the ERK1/2 pathway and activation of autophagy and proteasome activity in Iso-treated males. This work highlights that the reversal of pathological hypertrophy does not use universal signaling pathways and that sex potently modifies this process.

NEW & NOTEWORTHY Pathological cardiac hypertrophy is a major risk factor for mortality and is thought to be largely irreversible in many individuals. Although cardiac hypertrophy itself has been studied extensively, very little is understood about its regression. It is important that we have a better understanding of mechanisms leading to regression, why this process is not reversible in some individuals and that sex differences need to be considered when contemplating therapies.

INTRODUCTION

Cardiac hypertrophy is a major risk factor for mortality and can be caused by high blood pressure, diabetes, mutations in sarcomeric proteins, and aortic valve stenosis (1–4). In mammals, this hypertrophy is the result of increased cardiomyocyte size, leading to thickening of the left ventricular (LV) walls and a decrease in the volume of the LV chamber. If cardiac hypertrophy persists for an extended time, there can be many maladaptive changes to the myocardium including cell death, fibrosis, and lengthening and thinning of the cardiomyocytes, ultimately leading to cardiac dilation and, potentially, heart failure (5). In the early stages of cardiac hypertrophy, the increase in cardiomyocyte size is compensatory to normalize ventricular wall stress. With early treatment of the underlying cause, cardiac hypertrophy can be reversed. However, there are varying degrees of regression of LV hypertrophy in patients with hypertension based on the treatment they receive (6). Angiotensin II (ANG II) receptor antagonists, calcium channel antagonists, and angiotensin-converting enzyme inhibitors can significantly decrease LV mass index by 13%, 11%, and 10%, respectively (6). However, diuretics and β-adrenergic receptor blockers did not affect LV mass index (6). In addition, weight loss and decreased sodium intake led to similar decreases in LV mass index compared with patients who received antihypertensive drug treatments (7). Following aortic valve replacement for aortic stenosis, patients experienced LV mass reductions ranging from 17% to 31%, and these changes were strongly associated with the severity of LV hypertrophy before surgery (4). Overall, the longer the heart experienced a pathological stress and the increased workload, the more maladaptive changes occurred, and hypertrophy was less likely to be reversible (5, 8). There are notable cases in which regression of chronic pathological cardiac hypertrophy occurred more readily, including bariatric surgery in which significant decreases in LV mass ranged from 21% to 30% (9–12) or a left ventricular assist device (LVAD) in which significant decreases in LV mass ranged from 28% to 41% (13, 14). Although many mechanisms that lead to cardiac hypertrophy are known, very little is understood about regression of cardiac hypertrophy.

Models for regression of cardiac hypertrophy include a debanding from transverse aortic constriction (TAC) (15–19), or the removal of a hypertrophic agonist such as ANG II or Iso (20–22). Each of these models demonstrated regression after ∼7 days, and there was a return to baseline of pathological gene expression markers such as atrial natriuretic factor (ANF) (18, 20, 22), brain natriuretic peptide (BNP) (17), collagen 1 A (17), and a normalization of the functional response such as ejection fraction (17, 19) and cardiac output (19, 21). However, each of these studies only reported a single time point after which regression was already complete. In addition, although one group reported a possible state of irreversible hypertrophy with 1 wk of debanding after chronic TAC (6 wk), there was still a reversal of pathological gene expression including, Myh7 and Acta-1 (17). However, the rate of regression, and the many molecular mechanisms that occur during the process of regression, are still unknown. A better understanding of the immediate responses and the progression of regression, along with a model of incomplete hypertrophy regression, will be important to understand why some patients experience regression and others do not. Finally, sex differences in regression have not been adequately addressed.

Sex differences in cardiac diseases have been observed for many years now, with some differences observed in the rates and extent of hypertrophy and/or regression. For example, with aortic stenosis, although women experienced cardiac hypertrophy more often, and to a greater extent, they also regressed from hypertrophy faster than men (23). In addition, males had higher expression of collagen I and III and matrix metalloproteinase-2 genes due to aortic stenosis, which may be a factor contributing to the slower rates of regression in males (23, 24). In the studies of LVAD placement, although some of the patients receiving an LVAD were female, most were males (13, 14, 25). Although sex was accounted for in the demographics, sexes were combined when analyzing LV mass differences and molecular changes. Similarly, patients receiving bariatric surgery were primarily female, but sexes were combined when analyzing results. However, there was one study that showed that male sex was independently associated with an increase in the 1-yr mortality rate postbariatric surgery (26). In rodent models of regression of hypertrophy, debanding from TAC or removal of the hypertrophic agonist (ANG II or Iso), were carried out with either only male rodents (16, 17, 20–22), the sex was not stated (18), or the results of the sexes were combined (15). We are just beginning to understand the many molecular mechanisms that underlie sex differences, even at baseline (27). Therefore, it is essential to define sexually dimorphic cardiac differences, both at baseline, and in response to stress.

Here, we compare cardiac hypertrophy and regression induced by two different pathological hypertrophic stimuli, ANG II or Iso in male and female mice. In addition, we compare the sexes at baseline for many cardiac parameters and find significant differences. Although both agonists caused hypertrophy, the extent of regression of hypertrophy was distinct between the sexes and the agonists. There were differences in transcriptome, activation of signaling pathways, extracellular matrix composition, and protein degradation pathways depending on the hypertrophic stimulus and/or biological sex. Furthermore, there were very few genes regulated in common among the groups.

METHODS

Animals and Treatments

All animal treatments were approved by the Institutional Animal Care and Use Committee at the University of Colorado Boulder (protocol no. 2351) and were in accord with the National Institutes of Health (NIH) guidelines. Wild-type, 10- to 12-wk-old C57Bl/6 male and female mice (Jackson Laboratories) were fed ad libitum standard rodent chow and housed in a 12-h:12-h light/dark cycle. Mice were treated with ANG II (2.88 mg/kg/day) or Iso (30 mg/kg/day) for 7 days (Fig. 1A). ANG II was diluted in sterile saline. Iso was prepared in 1 µM ascorbic acid, diluted in sterile saline. ANG II and Iso were released through osmotic minipumps (Alzet model 2001). Mice were anesthetized with Iso (3%) via spontaneous inhalation. Surgical procedures were performed on a 37°C recirculated heating pad. The analgesic buprenorphine was used at 1 mg/kg. To study regression, the osmotic pumps were removed after 7 days of ANG II/Iso; the same surgical procedure was used as placement of the minipump. Mice were euthanized by first anesthetizing with isoflurane (3%) via spontaneous inhalation; then, the heart was removed. Mice were euthanized at either 7 days of ANG II/Iso, indicating peak hypertrophy, or at post-stimulus days 1, 4, and 7 (P1, P4, and P7, respectively) (Fig. 1A). Hypertrophy was determined as the ratio of the mass of LV + septum over tibia length (LV/TL). There were vehicle controls (1 µM ascorbic acid in sterile saline for Iso or sterile saline alone for ANG II) at each time point. Hearts were dissected, and the left ventricle was weighed then flash-frozen in liquid nitrogen. Tissue was placed at −80°C until further analysis.

Figure 1.

Pathological cardiac hypertrophy and regression depend on the hypertrophic trigger and are modulated by sex. A: experimental set-up. Mice were treated with vehicle control, ANG II, or Iso for 7 days, administered through an osmotic pump. Pumps were removed and regression was studied at various time points. B: LV/TL in males and females compared with vehicle control group. n = 4–8/group. Means ± SE. One-way ANOVA post hoc-uncorrected Fisher’s LSD. *P < 0.05, ***P < 0.001, ****P < 0.0001 significance. C: plots of gene expression measured by RNA sequencing comparing postremoval day 1 (P1) and hypertrophy in males and females treated with ANG II (top) and Iso (bottom). D: Venn diagram showing common differentially expressed genes among all groups. E: biological functions identified using ingenuity pathway analysis (IPA) on the 45 differentially expressed genes in common to all groups. ANG II, angiotensin II; Iso, isoproterenol; LV/TL, left ventricle weight/tibia length; P, post hypertrophy day.

Protein Isolation and Western Blot Analysis

LV tissue was homogenized in urea buffer, consisting of 8 M urea, 2 M thiourea, 50 mM Tris (pH 6.8), 75 mM DTT, 3% SDS, and 0.05% bromophenol blue. Protein concentration was determined using Pierce 600 Protein Assay Reagent (Thermo Fisher Scientific, 22660) with the Pierce Ionic Detergent Compatibility Reagent (Thermo Fisher Scientific, 22663). Proteins were run on 4%–12% Bis-Tris gels and transferred to nitrocellulose membranes. Membranes were blocked with 5% BSA in Tris-buffered saline-Tween 20 (TBST) (TBS + 0.1% Tween) for 1 h at room temperature. Primary antibodies were incubated overnight in 5% BSA (TBST) at 4°C. Secondary antibodies were incubated for 1 h at room temperature. Membranes were imaged using ECL reagent (Perkin-Elmer, NEL104001EA). Quantification was determined using ImageQuant. All primary antibodies were purchased through Cell Signaling Technology and used at a 1:1,000 dilution: p-SMAD2 (3108), p-Akt (4058), p-p38 (4511), p-ERK1/2 (9101), LC3 (2775), except α-vinculin, purchased from Sigma (V9131). Secondary antibodies horseradish peroxidase (HRP)-conjugated anti-mouse (Jackson ImmunoResearch) or anti-rabbit (Cell Signaling Technology) were used at 1:5,000 dilutions.

Proteasome Activity Assay

Left ventricle tissue was homogenized in proteasome buffer, consisting of 50 mM HEPES, 20 mM KCl, 5 mM MgCl₂, and 1 mM DTT. Samples were centrifuged at 10,000 g for 30 min at 4°C. The supernatant was placed in a new tube, and the protein concentration was determined using Pierce BCA Protein Assay (Thermo Fisher Scientific, 23227). Each reaction contained 15-µg protein in a final volume of 230 µL proteasome buffer. In addition, each sample had a complimentary reaction that contained the proteasome inhibitor MG132 (20 nM). Sample (230 µL) was prepared on a 96-well white flat bottom plate. Fluorescent substrate (10 µL) was added; Suc-LLVY-AMC was used to measure chymotrypsin activity (18 µM; Enzo Life Sciences BML-P802). The samples were kept on ice until this point. The reaction was started by placing the plate in the plate reader at 37°C, and fluorescence was measured every 3 min for 60 min (excitation, 360 nm; and emission, 460 nm). Proteasome activity was determined by calculating the change in fluorescence; this value was then subtracted from the change in fluorescence from the complimentary reaction containing MG132. Each sample was run in triplicate.

RNA Isolation

LV tissue was homogenized in Tri Reagent (Molecular Research Center, TR118). Chloroform was added and incubated at room temperature for 15 min, then centrifuged at 12,000 g for 15 min at 4°C. The aqueous layer was removed and placed in a new tube. Isopropanol was added and incubated at room temperature for 15 min, then centrifuged at 12,000 g for 15 min at 4°C. The supernatant was removed, and the pellet was washed with 70% ethanol, then centrifuged at 7,500 g for 5 min at 4°C. The supernatant was removed, and the RNA was resuspended in water.

RNAseq Preparation and Analysis

RNA samples were submitted to Novogene for library preparation, by PolyA selection, and sequencing. All samples had a sequencing depth of at least 20 million 150-bp paired-end reads. All differential expressed gene (DEG) analyses were carried out in R (v. 3.5.0) with Bioconductor (3.10) package edgeR (3.10.2). Reads were removed from analysis if the expression was less than 0.5 counts/million. Data were fit using glmRTFit function. P values were adjusted by the Benjamini–Hochberg method to control the false discovery rate (FDR) at 0.05. Once exclusively DEGs were identified, we used QIAGEN ingenuity pathway analysis (IPA) software for identifying predicted functional pathways and upstream regulators.

cDNA Preparation and Quantitative Real-Time PCR

RNA was reverse transcribed using SuperScript III reverse transcriptase (Invitrogen, 18080044) and the protocol was followed according to the manufacturer’s instructions. cDNA was diluted to 1 µg/µL in water. Each qPCR reaction contained 4 µg cDNA + SYBR Green PCR Master Mix (Invitrogen 4309155) + 12.5 µM primer set. Thermocycler settings were determined using SYBR Green PCR Master Mix protocol. ΔΔCt was calculated using 18S as a normalizer (see Table 1).

Table 1.

Primers

Primer	Forward	Reverse
18S	GCCGCTAGAGGTGAAATTCTTG	CTTTCGCTCTGGTCCGTCTT
Col1a1	TACCGCTGGAGAACCTGGAA	GGGACCTTGTACACCACGTT
Postn	AAGTTTGTTCGTGGCAGCAC	TGTTTCTCCACCTCCTGTGG

Hydroxyproline Assay

A hydroxyproline assay kit was used (Sigma-Aldrich; MAK008) and followed according to the manufacturer’s instructions. In brief, 10 mg of tissue was homogenized and then hydrolyzed in 12 M HCl. Each sample (30 µL) was transferred in duplicate to a 96-well plate and allowed to dry in a 60°C oven for 60 min. Chloramine-T/oxidation buffer was added to each sample and standard well, followed by the diluted p-dimethylaminobenzaldehyde (DMAB) reagent. Samples were incubated for 90 min at 60°C. Absorbance was measured at 560 nm.

Statistical Analysis

Statistical differences were determined between two groups using the Student’s two-tailed t test. Between multiple groups, one-way ANOVA was performed followed by uncorrected Fisher’s LSD post hoc test. Outliers were determined using the Grubbs test with an α < 0.05. P values of < 0.05 were considered significant.

RESULTS

Pathological Cardiac Hypertrophy and Regression

To determine whether a pathological response to different stimuli would show similar patterns of hypertrophy and regression, we treated male and female C57/Bl6 mice with an activator of the renin-aldosterone-angiotensin system (RAAS), ANG II, or the β-adrenergic agonist, Iso, to induce a pathological cardiac remodeling and hypertrophy. After 7 days, treatments were removed and hearts were analyzed at different time points to investigate regression of cardiac hypertrophy (Fig. 1A). Male mice treated with ANG II (2.88 mg/kg/day) experienced an increase of 25.6% in LV/TL (Fig. 1B, Supplemental Fig. S1A; all Supplemental material is available at https://doi.org/10.6084/m9.figshare.19314041.v1) and female mice experienced an increase of 32.7% in LV/TL (Fig. 1B, Supplemental Fig. S1B), but these sex differences were not statistically significant. Iso treatment (30 mg/kg/day) induced a 33.8% increase in LV/TL in males (Fig. 1B, Supplemental Fig. S1A). Female mice showed a 28.5% increase in LV/TL (Fig. 1B, Supplemental Fig. S1B). Although this male-female difference was not statistically significant, it corroborates an earlier study that showed that females have a more modest hypertrophic response to Iso treatment (28).

After the removal of hypertrophic stimuli, mice showed significantly different rates of regression. After ANG II was removed, neither males nor females showed any significant regression at P1, and although male hearts regressed significantly at P4 (Fig. 1B, Supplemental Fig. S1, A and B), the LV weights of both males and females, remained significantly larger (∼20%) than the vehicle controls at P7, demonstrating incomplete regression. Iso male mice regressed faster than all other groups (Fig. 1A). Regression was significant already after day 1 (P1) of Iso removal in male mice and they showed complete regression after 4 days (P4) when compared with the vehicle control (Fig. 1A). After Iso was removed, female mice showed significant, but incomplete regression immediately at P1, similar to males (Fig. 1B). However, Iso females exhibited a slower regression and only showed complete regression at P7 (Fig. 1B, Supplemental Fig. S1). This is the first report to investigate the immediate response and the rates of regression with two different models, including one that is irreversible at the time points studied here, in addition to examining biological sex as a variable.

RNAseq Revealed Sex- and Trigger-Dependent Gene Expression Responses to Removal of the Hypertrophic Stimulus

Considering the slow and incomplete regression from ANG II and the fast regression response of male and female mice after Iso treatment, we investigated the cardiac transcriptomes of all experimental groups after hypertrophy and at P1 of regression. We performed RNAseq and assessed gene expression differences elicited by the removal of hypertrophic stimuli (Fig. 1C). After ANG II removal, males showed 831 differentially expressed genes, whereas females showed 2,063 differentially regulated genes (Fig. 1D). In Iso males, 1,152 genes were differentially expressed when comparing P1 with hypertrophy, whereas Iso females showed only 340 differentially expressed genes (Fig. 1D). This result was unexpected considering that the removal of ANG II induced little regression of cardiac hypertrophy in either male or female mice. Moreover, these large responses of the transcriptome after removal of the hypertrophic trigger suggest that the heart is responding strongly in ANG II-treated mice, although not culminating in regression.

Of note, only 45 differentially expressed genes were common among all regression groups (Fig. 1D, Supplemental Fig. S2), suggesting that the removal of hypertrophic stimulus triggers largely distinct responses. IPA canonical pathways analysis of the 45 common genes identified two enriched pathways both related to cell cycle control (Fig. 1F). However, the vast majority of common genes did not cluster into enriched pathways. Thus, our data indicate distinct transcriptional profiles associated with the removal of hypertrophic stimuli that depend on the nature of the hypertrophic trigger and biological sex.

Fibrotic Signaling Is Inactivated During Regression of Cardiac Hypertrophy But Fibrosis Increases

For each group, we then performed IPA canonical pathways analysis on differentially expressed genes between P1 and hypertrophy. Results of the analyses were compared among groups to identify shared and distinct signaling pathways (Fig. 2A, Supplemental Fig. S3). All groups exhibited similar downregulation of cell cycle and proliferation signaling at P1 except for female ANG II. Male Iso also showed a higher degree of downregulation of fibrosis signaling (identified as “hepatic fibrosis signaling”) and cardiac hypertrophy signaling when compared with all the other groups. IPA prediction of upstream regulators identified TGF-β1 inhibition in male Iso at P1 as the most significant regulation among all groups (Fig. 2B). In male Iso, 248 differentially expressed genes were associated with inhibition of TGF-β1, including numerous fibrotic genes downregulated at P1 when compared with hypertrophy (Fig. 2C). The formation of a fibrotic network in the myocardium may play a role in the ability to regress following a pathological stress. We, therefore, measured hydroxyproline content with the hypothesis that higher collagen levels would inhibit or reduce regression, especially in male and female ANG II that showed the least regression (Fig. 1B). Unexpectedly, collagen content was not increased in hypertrophy for any of the groups and showed a similar increasing trend after the removal of the hypertrophic trigger in all the groups. Hydroxyproline content was significantly increased by P4 in both ANG II and Iso males (Fig. 2D) than in vehicle control mice, whereas in female ANG II, the increase was significant only at P7. Collagen content in female Iso did not reach statistical significance by 7 days after removal of the trigger even though regression was complete at that time point. We predicted that the (re)activation of fibrotic genes during regression was the cause of increased fibrosis; therefore, we measured the expression of collagen1 a1 (Col1a1) and periostin (Postn), both of which are components of the extracellular space and contribute to fibrotic networks (29). Interestingly, both Col1a1 and Postn were increased in hypertrophy for all the groups (Supplemental Fig. S4) and progressively reduced with the removal of the hypertrophic trigger, with the exemption of female ANG II that maintained a higher expression level of both genes at P1. Overall, female Iso showed the smallest increase in expression of Col1a1 and Postn in hypertrophy and a significant downregulation of both genes at P1. The same group showed only a mild nonsignificant increase of hydroxyproline content at P7. Of note, hydroxyproline content and Col1a1 expression were higher at baseline in females than in males, with no difference in the expression of Postn (Supplemental Fig. S5A). These data together indicate that collagen content is sexually dysmorphic at baseline and the expression of fibrotic genes depends on the hypertrophic trigger although fibrosis can be manifested at later times even after the removal of the hypertrophic signal and during cardiac regression.

Figure 2.

Fibrotic signaling involved upon removal of hypertrophic trigger. A: heatmap showing the top canonical pathways identified by ingenuity pathway analysis (IPA) that are enriched at post-removal day 1 (P1) in male and female mice treated for 7 days with ANG II or Iso. Complete list in Supplemental Fig. S3. B: TGFβ1 in male Iso is the most significant upstream regulator predicted (P = 4.99e-61) compared with any other regulator in all groups. IPA predicts the inhibition of TGFβ1 regulator for male Iso (z score = −6.085). C: selected differentially expressed genes regulated by TGFβ1 in male Iso at P1 and normalized by hypertrophy. D: hydroxyproline content measured in all groups. Seven days of treatment with hypertrophic trigger did not induce an increase of collagen deposition (vehicle vs. hypertrophy). Significant increase of collagen deposition was observed after removal of hypertrophic stimuli when compared with hypertrophy. Means ± SE. One-way ANOVA post hoc-uncorrected Fisher’s LSD. *P < 0.05, **P < 0.01, ***P < 0.001 significance. n = 4–8/group. ANG II, angiotensin II; Iso, isoproterenol.

TGF-β1 Canonical and Noncanonical Pathways

TGF-β1 signals through Ser/Thr kinase receptors that activate a canonical pathway by phosphorylating Smad2/3 (small mother against decapentaplegic), and noncanonical pathways via PI3K/AKT, Ras/ERK, and MEKK/p38 (Fig. 3A). Phosphorylated Smad2/3 forms homomeric and heteromeric SMAD complexes that translocate to the nucleus and induce a fibrotic gene program (30, 31). We first evaluated the baseline levels for SMAD signaling and the TGF-β1 noncanonical pathways between males and females (Supplemental Fig. S5B). Activation of p-SMAD2/3 was not different between males and females, suggesting that the higher levels of Col1a1 observed in females was not the result of a specific signaling but rather the healthy baseline levels associated with female biological sex. Conversely, Akt and ERK1/2 signaling were both higher in females than in males (Supplemental Fig. S5B). There was no sex difference in p-p38 at baseline (Supplemental Fig. S5B). Significant activation of p-Smad2/3 was observed in all groups in hypertrophy when compared with vehicle, in good agreement with gene expression levels of Col1a1 and Postn (Supplemental Fig. S4, A and B). At P1, consistent with gene expression and RNAseq data showing downregulation of genes downstream of the TGF-β1/Smad pathway (Fig. 2C), p-Smad2/3 was significantly decreased in Iso-treated males and females. It was significantly decreased at P4 in ANG II-treated males and remained elevated in ANG II-treated females (Fig. 3B), also in agreement with Col1a1 and Postn gene expression. Overall, ANG II infusion for 7 days resulted in longer activation of Smad2/3 when compared with Iso infusion and extended beyond the removal of the hypertrophic trigger (Fig. 3B). Akt phosphorylation (Ser473) was significantly inactivated only in Iso females at P1, although it was not increased with Iso treatment when compared with vehicle control (Fig. 3C). p38 signaling (phosphorylation on Thr180 and Tyr182) was moderately but significantly reduced only in ANG II males at P1 when compared with hypertrophy, although it was not significantly different in hypertrophy when compared with vehicle control (Fig. 3D). Finally, ERK1/2 signaling (Thr202/Tyr204) was significantly activated in males treated with ANG II or Iso and significantly inactivated in both groups at P1 (Fig. 3E). At P1, Iso females also showed significant inactivation of ERK pathway, although they did not exhibit a significant activation of ERK in hypertrophy when compared with vehicle (Fig. 3E).

Figure 3.

TGFβ1 signaling. A: scheme showing the TGFβ1 signaling pathways via SMAD, AKT, ERK, and p38. B: Western blot of SMAD2/3 phosphorylation. C: Western blot of Akt phosphorylation. There was little regulation of Akt in response to ANG II or Iso in males or females. Protein quantifications were normalized to vinculin. D: Western blot of ERK1/2 was activated in male mice with ANG II or Iso, which decreased after stimulus removal. ERK1/2 was not activated in female mice with ANG II or Iso, but was decreased after the removal of Iso. E: Western blot of p38 phosphorylation. Protein quantifications were normalized to vinculin. Means ± SE. One-way ANOVA post hoc-uncorrected Fisher’s LSD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 significance; n = 4–8/group. Hypertrophy group compared with vehicle control; P1, P4, and P7 compared with hypertrophy. ANG II, angiotensin II; Iso, isoproterenol.

Protein Degradation Pathways Are Differentially Regulated by Sex and the Hypertrophic Trigger

Considering that during cardiac hypertrophy there is an increase in protein synthesis (32), we hypothesized that protein degradation pathways, namely, the ubiquitin-proteasome system and autophagy, would be regulated to promote regression and degrade proteins that accumulated during hypertrophy. Proteasome activity was determined by incubating lysates with the proteasome chymotrypsin-like substrate, Suc-LLVY-AMC (33). First, we observed that there were no differences in proteasome activity between males and females at baseline (Supplemental Fig. S5C). We found that proteasome activity increased with both ANG II and Iso in male mice at hypertrophy; however, the increase was larger in ANG II-treated mice (Fig. 4A). Although proteasome activity decreased immediately following the removal of ANG II, the removal of Iso resulted in a significant further increase, suggesting it may be involved in protein turnover in both hypertrophy and regression (Fig. 4A). By P4, the time point of complete regression, male Iso proteasome activity returned to baseline. In female mice, proteasome activity remained unchanged throughout hypertrophy and regression for both ANG II and Iso treatments (Fig. 4A). We measured autophagy by quantitation of the two forms of LC3. LC3 is a small protein that becomes lipidated in the growing autophagosome, and we measured the amount of the lipidated form (LC3II) relative to the unlipidated form (LC3I). At baseline, females had twofold more LC3II/LC3I, compared with males, indicating that they had higher autophagic activity (Supplemental Fig. S5C). Autophagy did not appear to be activated by ANG II-induced hypertrophy or during regression in male or female mice. Autophagy increased in males and females at P1 following the removal of Iso. Iso females showed a significant reduction of autophagy during hypertrophy when compared with vehicle controls, possibly to promote the accumulation of proteins. In addition, removal of Iso activated autophagy in females at P1 and remained higher at P4 and P7 when compared with the hypertrophy time point (Fig. 4B), but it was not different than baseline levels measured in vehicle controls. These results indicate that mice treated with Iso may regress at higher rates than ANG II-treated groups via induction of autophagy.

Figure 4.

Protein degradation pathways. A: measurements of proteasome. Proteasome increased in males in response to ANG II and Iso; then increased further in response to Iso removal but decreased after ANG II removal. Proteasome activity was unchanged in female mice. B: autophagy activity did not significantly change in male and female ANG II. Autophagy increased after the removal of Iso in male mice. In female mice, autophagy activity decreased with Iso; then increased after Iso removal. Means ± SE. One-way ANOVA post hoc-uncorrected Fisher’s LSD. *P < 0.05, **P < 0.01, ***P < 0.001 significance; n = 4–8/group. Hypertrophy group compared with vehicle control; P1, P4, and P7 compared with hypertrophy. ANG II, angiotensin II; Iso, isoproterenol.

DISCUSSION

Pathological cardiac hypertrophy is a risk factor for mortality and can be reversed with pharmacological treatment, weight loss, or surgery in some patients; however, not all patients respond to treatment by regressing their hypertrophy (1, 4, 6–8). The primary aim of our study was to determine whether sex and/or the nature of the pathological trigger impact regression differently. We induced hypertrophy in male and female mice with two different pathological agonists, ANG II and Iso, and studied gene expression programs and signaling pathways following the removal of the agonists. We showed that early after removal of pathological triggers, distinct gene expression profiles were elicited, and that regression was dependent both on sex and pathological trigger.

ANG II and Iso are known to act through different pathways converging on cardiac hypertrophy. ANG II is the main effector in the RAAS, and binds to ANG II receptors (AT₁R) (34). ANG II induces inflammation along with many pathological hypertrophic markers (NF-κB, TNFα, MAPK, and Akt signaling), which ultimately leads to an increase in blood pressure (34) and cardiac hypertrophy. Iso activates adrenergic receptors that are a class of G protein-coupled receptors on the cell surface that cause a canonical signaling cascade leading to increased concentration of Ca²⁺ in the cytosol (35). This increase ultimately results in faster contractions of the cardiomyocytes and mimics the increased workload observed in the disease state. Increased adrenergic signaling via catecholamines can also lead to high blood pressure and hypertrophy. In fact, patients with hypertension can be effectively treated with adrenergic receptor antagonists (36). Although Iso induces cardiac hypertrophy in the mouse strain (C57Bl/6), there are reports of less pathology (37) and little evidence of fibrosis (38, 39) compared with other mouse strains. In contrast, with ANG II, there has been evidence for pathological signaling and significant cardiac fibrosis (34, 40) in the C57Bl/6 background. A previous study in FVB mice compared the different rates of regression of heart weights between Iso and ANG II and showed that regression occurred from both stimuli (20). However, the timing and dosage of these experiments differ from our current study (20). In that work, ANG II was used at 200 ng/kg/min × 14 days, which is ∼15 times less than our dosage for twice as long, and Iso was used at 15 mg/kg/day × 7 days, which is half of our treatment. Regression was then observed at 7 days after Iso treatment and at 14 days after ANG II. In addition, the authors only investigated hypertrophy and regression in male mice. Our work provides the first comparison between male and female regression from pathological hypertrophy in either of these models.

In our work, hypertrophic responses were not significantly different between either agonist or sex (Fig. 1B, and Supplemental Fig. S1, A and B). However, regression showed both sex differences and agonist-specific differences (Fig. 1B, and Supplemental Fig. S1, A and B). In response to ANG II withdrawal, males experienced a significant regression at P4, whereas female LV weights remained higher and did not regress within the experimental window of 7 days (Fig. 1B, and Supplemental Fig. S1, A and B). After the withdrawal of Iso treatment, males completely regressed by 4 days, whereas it took females 7 days to completely regress (Fig. 1B, and Supplemental Fig. S1, A and B). A previous study showed that complete regression from ANG II-induced hypertrophy occurred after 7 days, but the mice were of different genetic backgrounds and the dose was lower than the one used in our study, resulting in <20% of cardiac hypertrophy (20). Our work also shows significant differences in the rate of regression comparing males and females.

Using RNAseq, we compared gene expression of LVs at the hypertrophy time point and at day 1 after removal of the trigger in Iso and ANG II males and females. The magnitude of total gene expression changes only 24 h after stimulus removal was much higher than we expected: 3,458 genes significantly regulated with very little commonality shared by all four groups. Only 45 differentially expressed genes were shared by all four groups (Fig. 2D) after removal of the hypertrophic trigger, of which only four or five genes clustered in cell cycle control and kinetochore metaphase signaling pathways, respectively (Fig. 1E). Also somewhat surprisingly, more than 2,000 genes were differentially expressed in females following the removal of ANG II (compared with ANG II hypertrophy) (Fig. 1, C and D) despite the fact that this group did not show significant regression (Fig. 1B, and Supplemental Fig. S1A). Analyzing the entire transcriptome of each group after removal of the hypertrophic trigger, we observed that male Iso showed higher downregulation of cardiac hypertrophy and fibrosis pathways (Fig. 2A), as well as the most significant downregulation of TGFβ1 (Fig. 2B). Although fibrosis was not previously associated with Iso treatment in mice (38, 39), we found remarkable downregulation of numerous fibrotic genes in our male Iso group (Fig. 2C). These results led us to conclude that less fibrotic gene signaling in male Iso was likely to contribute to the faster regression observed in this group compared with the others. In addition, previous studies showed that ANG II induces significant fibrosis (34, 40, 41), which we predicted was the underlying cause of irreversible hypertrophy observed in the ANG II groups. However, collagen content increased over time during regression in all groups and was significant 7 days after the removal of the trigger with the exception of female Iso. Gene expression of collagen (Col1a1) and periostin (Postn) was significantly decreased with the removal of hypertrophic triggers in all groups except for female ANG II. These results showed a temporal disconnection between a fibrotic gene program and fibrosis that warrants further investigation. Recently, a study found that hypertrophy induced by ANG II was associated with increased Tgfb1, Postn, and Col1a1 gene expression in both male and female mice that resulted in increased tissue fibrosis (41). These different results could be explained by the different experimental designs when compared with our work. McLellan et al. used ANG II infusion (1.5 mg/kg/day) for 14 days, whereas we administered ANG II (2.88 mg/kg/day) for 7 days. Of note, female hearts showed significantly higher levels of collagen deposition and Col1a1 gene expression than males at baseline (Supplemental Fig. S5A). Although the role of biological sex on cardiac function is still understudied (42), strong evidence indicates that sex dimorphism extends from whole heart function to myofibril mechanics (27) and exists in many, if not all, cardiac cell populations (43), including cardiac fibroblasts (44) that are largely responsible for collagen expression and deposition.

We followed up on the predicted TGFβ1 inhibition by assessing the activation status of both canonical and noncanonical TGFβ1 pathways (Fig. 3A). In line with the RNAseq analysis showing reduced expression of genes targeted by the TGFβ1/Smad pathway, we found that the Smad2/3 signaling pathway was inactivated in male and female Iso at P1 and in male ANG II at P4. It remained significantly activated in female ANG II (Fig. 3B). Noncanonical TGFβ1 pathways showed different levels of (in)activation among groups. There were no significant changes in p-Akt levels during Iso- or ANG II-induced hypertrophy in males. Female mice showed a very minimal decreased p-Akt levels during hypertrophy, with only a significant decrease of p-Akt in response to ANG II and the removal of ANG II (Fig. 3C). None of the groups showed activation of the p38 pathway but the removal of the hypertrophic trigger elicited a significant decrease in male ANG II only. Last, there was a significant increase in p-ERK1/2 during ANG II- or Iso-induced hypertrophy in males (Fig. 3D) and no significant changes in female mice due to ANG II or Iso. Immediately following the removal of the stimulus, all male mice and female Iso mice, experienced significant decreases in p-ERK1/2 (Fig. 3A). Activation of ERK1/2 has been observed in many pathological models (45) and inhibition of p-ERK1/2 has been shown to inhibit cardiac hypertrophy (46, 47). Our work provides additional insights into the ERK pathway by demonstrating its rapid inactivation following the removal of a pathological stimulus, similarly to p-Smad2/3, in the groups that show hypertrophic regression.

We hypothesized the involvement of protein degradation pathways to promote regression of cardiac hypertrophy. At baseline, proteasome activity was not different between males and females, whereas significantly higher autophagy was observed in female hearts than in males (Supplemental Fig. S5C). This result is consistent with a previous report claiming 50% higher LC3II levels in females than in males in C57Bl/6 (although the results were unpublished in the publication) (48). In males, proteasome activity increased with Iso treatment and further increased immediately after the removal of Iso (Fig. 4A). In contrast, ANG II resulted in a greater increase in proteasome activity compared with Iso; however, the activity level was significantly reduced immediately following the withdrawal of ANG II at P1. These results indicate that the proteasome may have an active role in regression from Iso-induced hypertrophy, but not following the removal of ANG II, in male mice. There were no notable changes in proteasome activity in female mice (ANG II or Iso) (Fig. 4A). Regarding the autophagy response to pathological hypertrophy, reports vary with some showing autophagy was increased (49), whereas others reported autophagy was decreased during early hypertrophy (50, 51), and then increased in failing hearts (51). However, each of these studies only reported on male rodents, and because autophagy is an important target for many pharmaceutical treatments, it is important to understand how autophagy is differently regulated in males and females in response to cardiac stressors. In our study, Iso did not induce autophagy in male mice, nor did ANG II in either sex. Female Iso showed instead decreased autophagy in hypertrophy (Fig. 4B). After the removal of Iso, autophagy increased in both males and females at P1 compared with the Iso hypertrophic state (Fig. 4B). Some reports indicate autophagy can be an indicator of health, as autophagy tends to decrease with age and poor health (52). Furthermore, when autophagy was promoted by caloric restriction, diastolic dysfunction was delayed in an aging rodent (53), whereas when autophagy was inhibited, cardiac function and structure declined (52). This could explain the faster regression observed in male and female Iso-treated mice after removal of the hypertrophic trigger when compared with ANG II groups. Male Iso showed increases of both autophagy and the proteasome activity, following stimulus removal, possibly resulting in the fastest regression rate among all groups. We posit protein degradation pathways could be a determining factor in regulating regression of cardiac hypertrophy.

In conclusion, complete regression of pathological cardiac hypertrophy occurred in the Iso model in both sexes, but the rate of regression was slower in females. Incomplete regression occurred in the ANG II model in both biological sexes. These differences could be due to alterations in signaling pathways, fibrotic gene expression, or protein degradation pathways that appear to be influenced by the hypertrophic trigger and by the biological sex (Fig. 5). Future studies will include probing more pathways to further understand regression, especially any sex differences that may affect how patients are treated and respond to antihypertensive treatments.

Figure 5.

Summary table of significant results. Arrows indicate P < 0.05. Red arrow indicates upregulation and green arrow downregulation. ANG II, angiotensin II; Iso, isoproterenol; LV/TL, left ventricle weight/tibia length.

SUPPLEMENTAL DATA

Supplemental figures may be found at https://doi.org/10.6084/m9.figshare.19314041.v1.

GRANTS

This work was supported by the Tom Marsico Endowment Fund and National Institute of General Medical Sciences Grant GM029090 (to L.A.L) and Human Frontiers Science Program Fellowship LT001449/2017‐L and American Heart Association Postdoctoral Fellowship 20POST3521111 (to C.C).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

D.L.M. and L.A.L. conceived and designed research; D.L.M., C.C., A.R.S., and C.D.O. performed experiments; D.L.M., C.C., and A.R.S. analyzed data; D.L.M. and C.C. interpreted results of experiments; C.C. prepared figures; C.C. drafted manuscript; D.L.M., C.C., A.R.S., and L.A.L. edited and revised manuscript; D.L.M., C.C., A.R.S., C.D.O., and L.A.L. approved final version of manuscript.