Sex-Specific Effects of the β3-Adrenergic Receptor Agonist Mirabegron in a Murine Two-Hit Heart Failure Model with Preserved Ejection Fraction

1. Introduction

Heart failure with preserved ejection fraction (HFpEF) is a multifactorial syndrome usually resulting from various non-addressed risk factors, such as hypertension, obesity and diabetes, hypertension, kidney diseases, atrial fibrillation, tobacco, age, and the female sex (menopause) [1].

HFpEF is associated with high morbidity and mortality (reaching up to 80% and 50%, respectively, within a 5-year after diagnosis) and represents more than 50% of HF-related hospitalizations [2].

Several preclinical models of HFpEF have been developed, notably the L-NAME (nitric oxide synthase inhibition) + high-fat diet (HFD) mouse model, introduced by Schiattarella et al. [3], which initiated the trend of combining “hits” to induce the syndrome in animals. [4,5,6,7] Most of these new models involve an approach targeting major HFpEF risk factors such as hypertension and metabolic alterations. [7] These animal models all display cardiac hypertrophy (CH), concentric left ventricle (LV) remodelling, diastolic dysfunction, myocardial fibrosis and preserved LV ejection fraction (EF). Pulmonary congestion, exercise intolerance, and elevated natriuretic and inflammatory markers are typically present. Some models were characterized in both sexes, as females often show a less severe phenotype.[7]

The HFpEF model in C57Bl/6J mice that we recently developed (the MHS model) is also based on the combination of hypertension (Angiotensin II [AngII] continuous infusion) and metabolic alterations induced by a high-fat diet (HFD) for 28 days. [8] This model is reproducible and induces the expected HFpEF features such as CH, LV concentric remodelling, LA enlargement and decreased exercise capacity in both male and female mice. This combination of stress induces HFpEF in older females, irrespective of their hormonal status (ovariectomized or not). [8,9] Old males (20 months), however, tend to evolve toward HFrEF. [9]

HFpEF has long been recognized as a multifaceted syndrome affecting multiple organs. It is estimated that around 80% of HFpEF patients are obese, suggesting that the adipose tissue may be a significant contributing factor in the development of this condition. Obesity is associated with a chronic inflammatory state that can exacerbate the severity of other HFpEF risk factors, such as aging, hypertension, kidney disease, and atrial fibrillation, among others.

We observed in the MHS model that the circulatory microRNA (miR) profile was strongly perturbed. The adipose tissue, along with the liver, is a significant source of miRs found in the circulatory system [10]. More recently, we reported that cold housing exacerbated the HFpEF phenotype in mice [11]. Thermoneutrality in mice is estimated to be around 29-32°C. At these temperatures, thermogenesis, which involves the activation of the brown adipose tissue (BAT), is at its minimum. Animals housed at 30°C had a less severe HFpEF phenotype compared to those at standard room temperature (22°C). Housing at 10°C resulted in mice exposed to the MHS in a loss of EF.

β3-adrenergic receptor (β3-AR) activation stimulates the production of factors by the BAT that are believed to protect the cardiovascular system in pathological situations, such as heart failure. [12,13]

Here, instead of using cold to induce BAT activation, we relied on a specific β3-adrenergic receptor (β3-AR) agonist, mirabegron (Mira). Mira has been shown to dampen cardiac remodelling in male mice treated with AngII [14,15]. Here, we demonstrate that Mira had sex-specific effects on the cardiac response to the MHS, with a greater benefit observed in male mice compared to females.

2. Materials and Methods

Animals

C57BL/6J male and female 7-week-old mice were purchased from the Jackson Laboratory (Bar Harbor, ME, USA). Mice were housed (3-4 mice/cage) on a 12-hour light-12-hour dark cycle with free access to food and water. The protocol was approved by the Université Laval's animal protection committee and followed the recommendations of the Canadian Council on Laboratory Animal Care (#2023-1249). This study was conducted in accordance with the ARRIVE guidelines. Mice were randomly assigned to the various experimental groups: controls, MHS, Mirabegron (Mira), and MHS + Mira. The experimental design is schematized in Figure 1A.

Mice were randomly assigned to the various experimental groups (n = 7-8 per group). Experienced technicians monitored the mice's health and behaviour daily during the protocol, and the animals were weighed weekly. Three intact male mice died in the MHS group.

Metabolic and hypertensive stress (MHS). As described before, mice were implanted with an osmotic minipump (Alzet #1004) providing a continuous infusion of angiotensin II (AngII; 1,5mg/kg/day) (Sigma) for 28 days and fed with a high-fat diet (HFD: 60% calories; Research Diets Cat. #D12492). [8] Mirabegron (Toronto Research Chemicals, Toronto, Canada) treatment. Mice were implanted with an osmotic minipump (Alzet #1004) that provided a continuous infusion of mirabegron (dissolved in DMSO and PEG400 [Vehicle]; 2 mg/kg/day) (Sigma) for 28 days. Experienced technicians monitored the mice's health and behaviour daily during the protocol, and the animals were weighed weekly.

Echocardiography

Echocardiography was performed under isoflurane anesthesia as described previously (8, 16).

RNA Isolation and Quantitative Real-Time Polymerase Chain Reaction

As described previously, total RNA was extracted from LV tissue (8). Quantitative RT-PCR was used to quantify LV gene expression in at least six animals per group. Cyclophilin A (Ppia) served as the control “housekeeping” gene for studying cardiac genes, while RPL13 was used for studying brown fat genes. The primers used are listed in Table 1.

Myocardial Fibrosis Evaluation

Myocardial samples were fixed, sliced in serial sections (10μm thick), and stained with Picrosirius Red to assess the percentage of interstitial fibrosis. The formula used to calculate the percentage of interstitial fibrosis was: ((% Fibrosis) / (% Fibrosis + % Tissue)) × 100 [17]. Perivascular fibrosis was not included in the estimation of interstitial fibrosis.

Cardiomyocyte Cross-Sectional Area

The cardiomyocyte cross-sectional area (CSA) was visualized with immunofluorescent wheat germ agglutinin-FITC (Sigma) staining as previously described [17].

BAT Histology

BAT samples were preserved in a 4% paraformaldehyde solution for 24 hours and transferred to 70% ethanol. The specimens were paraffin-embedded, cut into 4-μm sections and stained with hematoxylin and eosin.

Statistical Analysis

All data are expressed as mean ± standard error of the mean (SEM). Outliers were removed using the ROUT test with a Q of 1% with Prism. Intergroup comparisons were conducted using the Student's T-test using GraphPad Prism 10.6 (GraphPad Software Inc., La Jolla, CA, USA). Comparisons among more than two groups were analyzed using a two-way ANOVA followed by Fisher's Least Significant Difference (LSD) test. P<0.05 was considered statistically significant.

3. Results

3.1. Mirabegron, a β3-Adrenergic Receptor Agonist, Has Sex-Specific Effects on Body and Cardiac Growth in Young C57BL/6 Mice

As illustrated in Figure 1A, C57Bl6/J male and female mice received AngII and the high-fat diet (HFD) for four weeks (MHS) alongside or not with mirabegron for 28 days. Control mice receiving the vehicle or mirabegron (Mira) alone were studied in parallel.

Indexed heart weight (for tibial length) was lowered in animals treated with mirabegron (Figure 1B). MHS mice receiving Mira also had a smaller heart, although the stress caused cardiac hypertrophy. Body weight remained unchanged in mice receiving Mira, but body growth (as measured by tibial length) was reduced in males, not in females (Figures 1C and 1D).

The MHS has been shown to lead to left atrial (LA) enlargement [8]. We observed this in mice of both sexes. However, LA were smaller in MHS males receiving Mira. This was not the case for females. (Figure 1E). Overall, we did not observe differences in the levels of cardiac hypertrophy (compared to the respective control groups) in mice treated with Mira (Figure 1F) compared to those not treated.

Figure 2 shows cardiac morphological and functional changes induced by the MHS in mice treated or not with Mira. As illustrated in Figures 2A-C, the MHS, as expected, induced LV concentric remodelling, characterized by increased left ventricle (LV) wall thickness and a decreased LV diastolic diameter, resulting in an increased LV relative wall thickness or RWT (LV wall thickness/EDD). In mice treated with Mira, LV walls were thicker in males, EDD was smaller, and RWT was thus increased compared to male animals not receiving Mira. In females, the effects of Mira were not present, except for a slight reduction of EDD.

Table 2 summarizes additional echo measurements in our mice. Diastolic interventricular septum thickness (IVSd) and end-systolic diameter (ESD) were reduced in MHS+Mira males compared to the MHS-only group. This was not the case for females. Using bidimensional echocardiography LV imaging, we calculated volumes and ejection fraction using Simpson’s method. As illustrated in Figures 2D and 2E, end-diastolic (EDV) and end-systolic (ESV) LV volumes were reduced in male animals treated with Mira but not in females. Interestingly, this resulted in a higher ejection fraction (EF) in males. Cardiac output (CO) remained stable in all groups (Table 2).

We measured diastolic function parameters using echocardiography (Table 3). We had observed in previous studies that four weeks of MHS was not sufficient to induce pulmonary congestion, but prolonging the stress up to eight weeks increased lung water content (Figure 3A). Here, we observed that Mira treatment increased lung water content in both control and MHS mice. This water content was not higher in MHS mice, however. Liver weight was increased in males by Mira but not in females (Figure 3B). The MHS reduced liver weight in all female groups, but only in males treated with Mira.

We then measured the mRNA levels of six genes associated with pathological hypertrophy. Gene expression of the atrial natriuretic peptide (Nppa) was, as expected, increased in MHS mice not receiving Mira. In females treated with Mira, MHS caused a similar increase in Nppa, but this increase was almost completely obliterated in males (Figure 3C). Brain natriuretic peptide (Nppb) gene expression was increased by the MHS in both male and female mice, but at a significantly lower level in females (Figure 3D). We had observed in the past that collagens (Col1a and Col3a) mRNA levels were minimally or not modulated in females after 28 days of MHS, although myocardial fibrosis content was increased [8,9,17]. We confirmed this previous observation, and Mira had no additional effects. In males, mRNA levels for these two genes were increased by MHS; however, Mira completely blocked this effect (Figures 3E and 3F). We also measured genes associated with myocardial fibroblast activation, specifically periostin (Postn) and Thrombospondin 4 (Thbs4). In animals receiving the vehicle, both gene mRNA levels were increased. This was blunted in males receiving Mira, but not in females. (Figures 3G and 3H)

We then studied the scapular brown adipose tissue (BAT) depot of our mice. As illustrated in Figure 4A, Mira treatment led to an increase in the weight of BAT in control mice. This increase was more elevated in males than in females. Interestingly, the MHS partially reversed the effects of Mira in BAT. This can be visualized in Figure 4B, which shows representative pictures of the BAT depot from control Mira-treated female mice (top) and after MHS (bottom). The MHS BAT was also darker. This can be attributed to the lower fat content in brown adipocytes (Figure 4C) and smaller cell size, as illustrated in Figure 4D. To estimate the adipocyte size, we counted the number of nuclei in a microscope field. Except for vehicle-treated MHS males, all MHS groups had more nuclei and thus smaller brown adipocyte cell size.

Activation of the β3-adrenergic receptor (β3-AR) by Mira is expected to lead to an increase in thermogenesis and the expression of the mitochondrial uncoupling protein 1 (Ucp1) gene. As illustrated in Figure 4E, Mira treatment resulted in the expected rise in Ucp1 mRNA levels in the BAT (Figure 4E). We also measured the β3-AR gene expression levels (Adrb3) in the BAT. Mira treatment did not affect these levels in control animals and MHS females. In MHS males, Adrb3 gene expression was increased over control groups irrespective of Mira treatment (Figure 4F).

3.2. Mirabegron Effects on Cardiomyocyte and Extracellular Matrix Remodelling After MHS

The C57BL/6J mouse heart does not express the β3-adrenergic receptor gene. From the previously published LV bulk RNASeq data [8; GEO Accession Number: GSE240171], we illustrated in Figure 5A the LV gene expression of the β1, β2, and β3-AR genes. The normalized counts of each gene for control and MHS mice (three males and three females for each group) show that the β3-AR gene expression levels (Adrb3) in the LV are undetectable compared to the other two genes encoding these receptors. Interestingly, the levels of the β1-AR gene (Adrb1) were lower in MHS animals, but those of the β2 form (Adrb2) remained stable.

Cardiomyocyte size was estimated on LV sections stained with fluorescent wheat germ agglutinin (WGA). The MHS increased cardiomyocyte cross-sectional area (CSA) as expected in males treated with or without Mira. The CSA was smaller in males receiving Mira. In females, treatment with Mira increased the cardiomyocyte in control females, and the MHS did not lead to a further rise in this parameter (Figures 5B and 5C).

The MHS led to increased myocardial fibrotic content in males and females (Figures 5D and 5E). This rise in fibrosis tended to be reduced in males treated with Mira (p = 0.077).

4. Discussion

The adipose tissue is now regarded as an endocrine organ that secretes many factors, part of a complex network of interorgan crosstalk [18,19]. If this is largely recognized for the white adipose tissue, this is becoming increasingly accepted that BAT actively participates in this crosstalk. The BAT has long been regarded mainly as a thermogenic tissue, but its secretory activity is now recognized. “Batokines” are molecules secreted by the BAT, and several pieces of evidence support that they provide several benefits to other systems, including the cardiovascular system. [20,21]

In this study, we aimed to activate the BAT using a selective β3-AR agonist, mirabegron and investigate if it would benefit the heart in a murine HFpEF model. For several parameters measured, this was the case, but mainly in males. Male mice receiving Mira had after MHS, reduced body weight, shorter tibial length, lower indexed heart weight, and less atrial enlargement. Females treated with Mira maintained their body weight, and tibial length remained unchanged. Left atrial enlargement was similar after MHS in mice treated with Mira compared to females not receiving Mira. Still, females treated with Mira had smaller hearts.

When analyzed by echocardiography, control Mira-treated male hearts had thicker LV walls, smaller LV chamber and increased EF. Females did not show this response to Mira. However, stroke volume and cardiac output were lowered by Mira treatment in females, but not in males. Mira did not modulate the echo diastolic parameters in males, but in females, the A and A’ waves were lower, suggesting the atrial contraction may have been affected.

We observed that mice express almost undetectable LV levels of the Adrb3 gene in both control and MHS animals. As expected, the β1-AR and β2-AR were the primary forms expressed in the left ventricle of the mice. The expression of the β1-AR form was reduced in HFpEF animals, suggesting downregulation that may be due to an excess of circulating catecholamines. The mRNA levels of the β3-AR mRNA were too low for detection in all mice. This indicates that mirabegron effects were most likely the result of its action on extra-cardiac tissues, notably the BAT, but not only. The β3-AR is also present in white adipose tissue (WAT) depots, and its activation can lead to beiging [22,23]. In the WAT, the β3-AR is present in beige or “brite” adipocytes (brown adipocytes located in the WAT), and its activation results in the activation of thermogenesis in the WAT [24]. This beiging phenomenon likely changes the amount or nature of factors usually released by the WAT. It is thus expected that Mira may have modulated factors secreted by the WAT in our mice, which in turn could have affected the heart.

Evidence suggests that the β3-adrenergic receptor (β3-AR) is upregulated in the failing heart in both human and animal models. [25,26,27] How do we reconcile our observations with these previous reports? Firstly, in the normal heart, β3-AR is low and possibly, the nature of the pathological stress may influence its up-regulation. Second, in previous studies, upregulation was observed in end-stage ischemic or dilated cardiomyopathies, whereas here, it is observed in a model of HFpEF.

In humans, activation of the β3-adrenergic receptor is associated with endothelial relaxation and reduced myocardial contractility. [28] The cardiac levels of β3-AR have been shown to rise in patients with heart failure [25,26]. Again, the cardiac β3-AR levels are one to two orders of magnitude less than the β1-AR or the β2-AR. Overexpressing the human β3-AR in the mouse heart has been shown to protect it from pathological hypertrophy and abnormal fibrosis in response to neurohormonal stimulation [29]. Cardiac β3-AR expression also protects against ischemia-reperfusion injury in mice [30]. These benefits were shown to be mediated via the coupling of this receptor to nitric oxide and via antioxidant protective effects.

Mirabegron is primarily used in the treatment of overactive bladder syndrome [31]. Still, two clinical trials have been conducted in patients with heart failure to explore the potential benefits of β3-AR activation. The results of these trials (one conducted in NYHA functional class III and IV patients and the other in NYHA I and II patients) were neutral [32,33,34]. Another study in patients with combined pre- and post-capillary pulmonary hypertension associated with symptomatic HF also resulted in neutral outcomes [35].

The β3-AR is highly expressed in the BAT, and the use of specific agonists, such as mirabegron, has often been employed to stimulate this fat depot in rodents [36,37,38]. As mentioned, it is challenging to attribute the effects of Mira solely to factors generated by the BAT. One way to show this would be to remove or inactivate the BAT.

In a rat model of metabolic syndrome (Dahl-Salt/obese rats), they observed that salt-loading hypertension was associated with whitening of the BAT and that its surgical removal would attenuate the exacerbation of LV diastolic dysfunction, inflammation, and fibrosis. Ucp1 gene expression was reduced in the BAT of rats with metabolic syndrome, suggesting that normal BAT function is crucial for the cardiac response to pathological stress. [39]

In one of these studies, researchers induced pressure overload in mice by using abdominal aortic constriction and then kept them in a cold environment. Physical removal of interscapular BAT during cold exposure decreased thyroid hormone, T3, synthesis and mitigated pathological cardiac remodelling. However, mice harbouring a UCP1 KO and eliminating thermogenesis but preserving BAT integrity, they observed increased T3 synthesis and exacerbated pathological remodelling. This study suggests that there are differences between BAT ablation and inactivation in their roles in controlling thyroid hormone synthesis.[40]

BAT surgical removal was shown to block the benefits from exercise in a myocardial ischemic-reperfusion model in the mouse. They demonstrated that the BAT secretes small extracellular vesicles and that their miR content regulates cardiomyocyte survival and participates in exercise cardioprotection. [41]

Taken together, these studies show the interaction between the BAT and the heart. Others tried to identify factors other than miRs that could mediate the protective function of the BAT on the heart. The fibroblast growth factor 21 has been proposed as one of these factors.

A link between activation of the α2-AR/FGF21 axis in the BAT and the heart has been unveiled in DOCA-salt-treated mice (a hypertension model). Ablation of the α2-AR or of FGF21 in the BAT was shown to exacerbate the cardiac pathological phenotype, suggesting that the BAT was responsible for the production of cardioprotective factors such as FGF21. [42]

Another study investigating the effects of a α2-AR agonist (dexmedetomidine; DEX) in a mouse myocardial ischemia/reperfusion injury model also identified FGF21 as a cardioprotective factor. DEX increased circulating FGF21 levels, but BAT ablation countered the benefits of DEX treatment, suggesting that the FGF21 rise in concentration originated mainly from the BAT.[43]

In addition to FGF21, other batokines were shown to produce cardioprotective effects [44]. The batokine 12,13-dihydroxy-9Zoctadecenoic acid (12,13-diHOME) has been shown to improve cardiac hemodynamics by increasing cardiomyocyte contraction, relaxation, and mitochondrial respiration. Treatment with 12,13-diHOME blocks the effects of a high-fat diet on cardiac function and remodelling. [44]

The benefits of BAT function in cardiac pathological states have also been demonstrated using BAT implantation in the abdominal cavity of mice with LV pressure overload induced by transverse aortic constriction, or TAC. BAT implantation reduced the loss of EF in TAC mice. In a mouse model of aortic dissection/aneurysm, up-regulation of β3-AR in the perivascular adipose tissue. Activation of β3-AR by Mira significantly prevented AngII-induced AD/AA formation in mice. [45]

In our study, Mira treatment led to cardiac sex differences in both control and MHS mice. Assuming that Mira effects were mediated mainly via BAT activation of the β3-AR and secretion of batokines. The response of the heart to these batokines is thus sex-specific, suggesting that the nature or amount of these BAT factors differ between males and females. Our study was not designed to answer this question, and the paucity of studies on BAT physiology that investigated sex differences in mice leaves us with limited material to hypothesize about the reasons for these differences.

A recent study showed that activation of the BAT accelerated cardiac metabolic remodelling and reduced hypertrophic and fibrotic responses to pressure overload from TAC. [46] Interestingly, when animals were housed in a cold environment (16°C) to stimulate UCP1-dependent thermogenesis, this effect was not observed uniformly for males and females. At 16°C, females did not increase their UCP1 levels compared to room temperature (23°C). UCP1 levels were more elevated in females at 23°C than in males, suggesting that the cold stress from being housed at 23°C activated the BAT more in females.

We recently reported the effects of different housing temperatures on the HFpEF phenotype of MHS mice. [11] We did not observe significant sex differences in Ucp1 gene expression between male and female animals. Housing at room temperature and in a cold environment (10°C in our case) was associated with a similar rise in Ucp1 expression in both sexes. Unlike the study above, the MHS increased BAT Ucp1 gene expression; however, this is likely linked to the effects of Angiotensin II on the BAT via the type 2 Ang II receptor.

We used AngII in the MHS model to induce hypertension and as a direct cardiac hypertrophic factor. As such, the MHS induces browning of the BAT in mice, accompanied by elevated Ucp1 gene expression and decreased adipocyte size, suggesting the activation of this fat depot [11]. Although the high-fat diet in MHS is a known factor that induces obesity, its use in conjunction with AngII results in no body weight gain. [8, 9, 11, 17. The HFD without AngII for four weeks increased body weight in mice compared to a standard diet; however, AngII had the opposite effect in males. [8] Combining the two results in similar body weights in mice after four weeks, as per the MHS. The control of adiposity by angiotensin II is mediated by the Angiotensin II receptor type 2 (AT2R). AT2R KO results in increased obesity in mice fed an HFD, and its activation with a specific agonist (C21) blunts weight gain [47,48,49]. C21 can also lead to BAT activation (upregulation of Ucp1) and the browning of WAT in mice [49]. In the MHS model, the effects of AngII on the BAT seem to have overruled those of the high-fat diet.

Mirabegron is usually associated with a loss of body weight since activation of BAT thermogenesis reroutes energy consumption. [50] We observe this in males but not in females, who tend to maintain their body weight more consistently. Mira slowed body growth only in males. Interscapular BAT depot size was also increased more in males than in females.

5. Study Limitations

This study used a two-hit model to induce HFpEF that resembles human HFpEF. It is still not representative of the entire spectrum of HFpEF patients.

Mirabegron treatment was initiated simultaneously with the MHS, potentially slowing or preventing the development of HFpEF. This experimental design is not ideally suited to test the therapeutic potential of β3-AR activation in the context of an established heart disease.

The translation of our results into human terms is challenging. Still, they emphasize that activation of the BAT is an interesting avenue worth pursuing in the future and that studying animals of both sexes can provide essential insights.

Several parameters were not measured in this study, including BP, animal metabolic activity, or daily activity, which may have affected our results.

6. Conclusions

Our results indicate that β3-AR activation via the specific agonist can benefit mice with HFpEF. These benefits seem to target more male animals, and the reasons for this need further research efforts to be uncovered.