Treatment with a selective androgen receptor modulator (RAD140) is linked to better cardiac function, with male-specific effects that are graded by interleukin-6

Abstract

Low testosterone levels are associated with impaired cardiac function and elevated circulating inflammatory markers, which are mechanistically related. Selective androgen receptor (AR) modulators (SARMs) can stimulate ARs and mimic testosterone. Whether SARMs affect cardiac structure or function and if circulating inflammatory markers grade these changes is unclear. We investigated the effects of a SARM (RAD140) on cardiac function and structure in older male and female mice and explored how systemic inflammatory markers relate. Older (23-month-old) male and female mice were treated with RAD140 (5 mg/kg/day) for 6-weeks. Left ventricular (LV) structure and function were measured with echocardiography. Serum cytokine levels were assessed with a multiplex assay, and LV gene expression was tested. RAD140 treated males and females had higher ejection fraction (+ 10.4%), stroke volume (+ 9.6 µL) and cardiac output (+ 4.5 mL/min), versus controls. However, RAD140 treated males had better markers of cardiac function, including myocardial strain (-5.8%) and isovolumic relaxation times (-4.7 ms) compared to controls, with lesser or opposite effects in females. Higher levels of interleukin-6 correlated with worse global circumferential strain in males, but not females. qPCR analysis for genes coding proteins related to cardiac function and structure revealed no differences, although AR mRNA was lower with RAD140 treatment in both sexes. No structural differences were observed. RAD140 treatment may enhance cardiac function in older male mice, but less so in older female mice. Increased circumferential strain in males related to lower interleukin-6 levels, supporting the AR’s connecting role between cardiac function and inflammatory markers.

Graphical abstract

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s11357-026-02151-9.

Introduction

Low levels of androgens (e.g., hypogonadism) negatively modulate cardiac function and are associated with cardiovascular diseases, including heart failure, in both males and females [1, 2]. The androgen receptor (AR) is expressed in cardiomyocytes, where it modulates myocardial hypertrophy, contractility, and extracellular matrix remodeling in both sexes [1, 3]. In both humans and rodents, age-related declines in circulating androgens are linked to reduced left ventricular ejection fraction and adverse ventricular remodeling [4, 5]. Such detrimental effects are related to a disruption of calcium handling and contractility in the myocardium [6]. Further, testosterone replacement in male and female rodent models helps to improve myocardial function [7, 8]. Low levels of testosterone have also been linked to impaired diastolic function, such that transitioning from contraction to relaxation and vice versa is impaired with chronically low testosterone [9]. Chronically low testosterone levels also promote a chronic and low-grade inflammatory state, marked by elevated systemic levels of interleukin-6 (IL-6) among other proinflammatory cytokines, which further impairs myocardial function and predisposes towards heart failure [10–12]. Testosterone replacement trials in older men and women have demonstrated variable effects on cardiac function and systemic inflammation, and have dose-dependent cardiovascular and hepatotoxic risks [4, 13–15]. Therefore, alternative approaches to harnessing the therapeutic power of the AR in the heart are worth exploring.

Selective androgen receptor modulators (SARMs) are non-steroidal ligands engineered to dissociate anabolic activity in muscle and bone from undesirable androgenic AR activation in prostate, liver, and other tissues [16, 17]. Preclinical studies demonstrate that compounds such as RAD140 and GTx-024 increase lean mass, preserve bone microarchitecture, and exhibit anti-tumor effects in males and females at doses that avoid supraphysiologic androgen peaks or prostate hypertrophy for males [18, 19]. Early-phase clinical trials report dose-dependent gains in muscle mass, improvements in physical performance, and osteoprotective benefits in older men and women, including cancer-cachectic patients, with a favorable safety profile at therapeutic doses [20–22]. However, little is known about their effects on the heart and any interaction with systemic inflammation. It is possible that the activation of the AR via a SARM could improve cardiac function by helping to augment key regulatory factors like calcium handling and contractility, which AR signaling facilitates [6, 23, 24]. On the other hand, the cardiac safety of SARMs remains mostly uncharacterized, and some clinical and preclinical evidence raises safety concerns by linking SARMs with acute heart failure and increased cardiac fibrosis, most notably in males [25–28]. Thus, research is needed to explore how SARMs affect the heart in both males and females, which will help elucidate whether they may have therapeutic value in the setting of heart disease relating to low androgen levels, or instead present cardiotoxic risks.

Given the important role of AR signaling in male and female hearts, the prevalence of androgen declines and cardiac dysfunction with aging, and the potential for both functional improvement and cardiotoxic risk, it is critical to define how SARMs influence myocardial health in these settings. To address this gap, we evaluated the effects of six-week administration of a SARM, RAD140 [19], on cardiac structure and function in older male and female C57Bl/6 mice using transthoracic echocardiography, systemic cytokine profiling, and mRNA quantification. We tested the hypotheses that RAD140 would: 1) enhance left ventricular performance, and 2) promote physiologic hypertrophy. How systemic inflammatory markers relate to these was also explored. This study provides foundational insight into the cardiac impacts of RAD140, informing its therapeutic potential and safety in the setting of aging.

Methods

Study design

Study protocols were approved by the Dalhousie University Committee on Laboratory Animals and performed within the guidelines of the Canadian Council on Animal Care. C57Bl/6 mice were purchased from Charles River (St. Constant, Quebec) and aged in Dalhousie’s animal care facility with access to food and water ad libitum. For these experiments, mice aged 22–24 months (mean age = 23.0 months; n = 21 males; 15 females) were chosen as they model the age in which androgen declines tend to occur in humans [29, 30] and in this strain of mice [31]. Mice were split into two groups: Vehicle-treated or SARM-treated. Mice in the SARM group were hand-fed 100 µL of a cherry syrup mixture every day for 6-weeks. The syrup for the RAD140 group included 0.1% dimethyl sulfoxide (DMSO) as a vehicle for RAD140, dosed at 5 mg/kg/day. The drug dosing and duration regimen was selected as it has been demonstrated to have physiologic effects on skeletal muscle at this dose [19, 32]. Mice in the control group were fed the same volume of syrup with 0.1% DMSO but without RAD140.

Echocardiography was completed after the 6-weeks of treatment, and mice were then euthanized (200 mg/kg sodium pentobarbital and heparin i.p.). Prior to harvesting tissue, mice were weighed. Blood used for serum cytokine quantification was harvested following euthanasia by snipping the aorta and collecting blood directly from the heart. Blood was stored at room temperature for 15-min and then centrifuged at 1500 g for 15-min. Ventricular tissue was separated into two ventricular sections containing approximately equal parts right and left ventricle, also split evenly between the base and apical ends, and frozen at −80 °C for later use. Tibia length was recorded to normalize heart weights. During the 6-week study period, three male mice (two in the control group and one from the RAD140 group) and three female mice (one in the control group and two in the RAD140 group) were euthanized due to humane endpoints, with no specific disease patterns noted. One female mouse in the RAD140 group was close to study completion and was included in the final analysis. Mice were excluded from analyses using pairwise-deletion based on data availability (i.e., all available data were used for the analyses). Not all mice underwent echocardiography prior to euthanasia due to logistical constraints, leading to sample size differences between outcomes.

Echocardiography

Prior to imaging, mice were anaesthetized using 3% isoflurane infused into oxygen gas exhausting at 0.8 L/min. Anaesthesia was then maintained using 1–2% isoflurane with the same oxygen output. Anaesthetized mice were positioned supine on a multiaxial platform to enable easy repositioning during imaging, with their paws attached to the platform using surgical tape. Paws were laid over four recording electrodes to monitor the animal’s heart rate during imaging. The platform was heated and enabled mice to be kept at physiologic temperature (approximately 36.5 °C measured via rectal probe). After securing mice to the platform, depilating cream was then used to remove fur prior to applying transmission gel. Mice were kept supine throughout imaging, and their hearts were imaged transthoracically with a 30 MHz ultrasound probe using a Vevo 2100 electrocardiography machine (Fujifilm VisualSonics Inc.).

Brightness mode (B-mode) was used to correctly position the probe in the short-axis view of the heart, followed by Motion mode (M-mode) in the short axis that provided the images for assessing structural and functional parameters during systole and diastole, including left ventricular circumferential strain and mass. Long axis views were also recorded in B-mode for left ventricular longitudinal strain analysis. Apical four-chamber views were used to assess diastolic function using pulse-wave Doppler, with measurements taken adjacent to the mitral valve. After imaging, mice were allowed to wake from anaesthesia alone in a clean, standard housing container with bedding, a house, food, and water. Half of the recovery cage was positioned on top of a warm water bottle to maintain body temperature and improve comfort. Strain analysis and estimations of cardiac structure and function were calculated using the default settings on the Vevo Lab software (version 5.6.1). Echocardiography analysis was performed in a blinded fashion, such that all image analyses were performed by a researcher unaware of treatment allocation.

RNA isolation and quantification

RNA isolation was performed using a column-based fatty and fibrous RNA isolation kit (Bio-Rad; ref# 732–6830). Frozen tissue samples were transferred to 1.5 mL tubes and homogenized using a ceramic bead homogenizer (Omni International) in 1 mL of PureZOL RNA isolation reagent (Bio-Rad, ref# 1708891). After 5-min of incubation, 0.2 mL chloroform was added and another 5-min incubation was performed. Samples were then spun at 12000 g for 15-min at 4 °C before the aqueous phase was removed and placed into a fresh tube. An equal volume of 70% ethanol as the aqueous phase was mixed into the fresh tube and driven through an RNA-binding column by centrifuging at 10000 g for 30-s. The column was washed with a low-stringency wash solution (added and spun at 10000 g for 30-s) before a DNase I solution was added for 15-min to break down genomic DNA. One high-stringency wash solution was followed by a low-stringency wash, spun in the same manner. RNA samples were eluted and quantified using a NanoDrop Lite (Thermo Scientific).

Real-time quantitative polymerase chain reaction (RT-qPCR)

Complementary DNA (cDNA) was synthesized using a cDNA kit (Bio-Rad; ref# 17326880). One µg of RNA template was added to a super-mix containing reverse transcriptase, repeating deoxythymidine oligonucleotide primers, and random hexamer primers. Samples were then primed for 5-min at 25 °C and reverse-transcribed for 20-mintues at 45 °C, before undergoing an inactivating step (1-min at 95 °C). Samples were stored at −20 °C until real-time quantitative polymerase chain reaction (RT-qPCR) assays were performed.

RT-qPCR was carried out using SYBR Green chemistry (Bio-Rad, catalog #1725271). The SYBR and cDNA samples were kept on ice prior to reaction setup. For each well (total volume 20 µL), the master mix consisted of 10 µL SYBR, 2 µL primer mix (forward and reverse), 2 µL cDNA template, and 6 µL nuclease-free water, for a final primer concentration of 500 nM. Samples were sealed with a transparent film and amplified using a Bio-Rad CFX96 thermal cycler. Thermal cycling conditions included an initial denaturation and polymerase activation phase at 95 °C for 5 min, followed by 40 amplification cycles of 10 s at 95 °C and 20 s at 60 °C. To ensure accurate normalization, six potential reference genes were evaluated for expression stability using M-values, and GAPDH and HPRT were selected based on their consistent expression across treatment groups in ventricular tissue (M-value < 0.5). Non-template controls were included to monitor for potential contamination or nonspecific amplification. Primer sequences for all targets are detailed in Supplementary Table 1.

Circulating cytokine measurements

A Luminex sandwich immunoassay measured 23 analytes using terminal blood samples at a 1:4 dilution (15 µL serum per sample) as per the manufacturer’s instructions and read using a MagPix multiplex system (BioRad). Four-parameter logistic standard curves were created from which cytokine concentrations were interpolated. Samples were run in duplicate and averaged before analysis or curve fitting. A blank well was measured and subtracted from all readings to remove background signals. The serum cytokine values used in this study were published previously in this paper [32] and used for different outcomes.

Statistics

All statistics were performed on GraphPad Prism 10.4.2 (La Jolla, CA) for Windows, GraphPad Software, Boston, Massachusetts USA, www.graphpad.com, and IBM SPSS Statistics V28.0. Echocardiography measures were all assessed using a 2-way ANOVA (time × treatment). Šídák post hoc tests were used to examine subsets within the ANOVAs and the restricted maximum likelihood method. Correlational analysis between echocardiographic measures and cytokines is described in the Supplementary Methods. The 2^−ΔΔCt method was used to analyze relative changes in mRNA expression in the SARM-treated mice normalized to vehicle-treated mice separated by sex [33]. All statistical tests were two-tailed and considered significant at α ≤ 0.05.

Results

Effects of RAD140 on left ventricular structure and function

Figure 1 shows representative M-mode echocardiography images from all groups from the short axis perspective, highlighting left ventricular diameters and wall thickness across the cardiac cycle.

Fig. 1.

Representative M-mode echocardiography traces of the left ventricle using a short axis view. Note: a = anterior wall during systole; b = anterior wall during diastole; c = ventricular diameter during systole; d = ventricular diameter during diastole; e = posterior wall during systole; and f = posterior wall during diastole

Cardiac structure: Fig. 2 visualizes how there were no clear differences in left ventricular structure between RAD140 treated males or females compared to vehicle treated controls, as captured in the short axis. This included analyses of estimated left ventricular mass, anterior and posterior wall thicknesses during end systole and end diastole, and ventricular diameters during diastole and systole. This is corroborated by the ex-vivo raw heart weights, which did not differ significantly between the RAD140 and control groups (Table 1). This finding held true after normalizing to body weight and tibia length measured post-euthanasia.

Fig. 2.

Left ventricular diameter (panels A, B), thickness (panels C-F), and mass (panel G) were not different in older male and female mice with or without 6-weeks of treatment with RAD140. Measurements were made using transthoracic echocardiography and tested with a 2-factor (sex × treatment) analysis of variance. N(control + RAD140) = 18 (9 + 9) males, 9 (4 + 5) females. A horizontal line represents the mean

Table 1.

Ex-vivo heart ventricular weights did not differ between RAD140 and vehicle treated male or female mice

Variable	Group	Males		Females
		Average	SD	Average	SD
Heart ventricles (mg)	Vehicle	170	30	172	32
	RAD140	155	24	165	19
	p-value	0.279		0.647
Tibia (mm)	Vehiclea	18.9	0.2	18.8	0.3
	RAD140	19.2	0.4	19.1	0.3
	p-value	0.099		0.185
Body weight (g)	Vehicle	39.2	6.5	34.4	6.4
	RAD140	37.6	6.2	35.1	6.4
	p-value	0.612		0.844
Heart ventricles × tibia−1	Vehiclea	8.85	1.62	9.17	1.57
(mg × mm−1)	RAD140	8.10	1.27	8.65	0.51
	p-value	0.340		0.540
Heart ventricles × body weight−1	Vehicle	4.43	0.85	5.21	1.57
(mg × g−1)	RAD140	4.21	0.80	4.76	0.51
	p-value	0.598		0.513

Data are mean ± standard deviation (SD). N(control + RAD140) = 18 (9^a + 9) males, 12 (6 + 6) females. Independent t-tests between Vehicle and RAD140 treated mice are presented for each sex. All values were measured acutely following euthanasia. ^aTwo male vehicle-treated mice did not have tibia length collected and were thus omitted from these analyses

Cardiac output and related measures: By contrast, there were clear differences between RAD140 treated mice and vehicle controls in measures of systolic function as measured in the short axis (Fig. 3). Notably, the direction of differences was the same for male and female mice. Heart rate remained unchanged in a pooled analysis (p = 0.874), but stroke volume was significantly higher in RAD140 treated mice in a pooled analysis (p = 0.013), as was cardiac output (p = 0.021). These changes were underpinned by a higher fractional shortening and ejection fraction for SARM treated mice in a pooled analysis (p = 0.016; Fig. 3D & E), which had a significant post hoc test for the males only (p = 0.050 and 0.042, respectively).

Fig. 3.

Cardiac output and stroke volume, but not heart rate, tended to be higher in RAD140 treated older male and female mice compared to vehicle controls (panels A-C). Similar differences were observed in fractional shortening and ejection fraction, although post hoc analysis was statistically significant for males but not females (panels D, E). Measurements were taken using transthoracic echocardiography tested with a 2-factor (sex × treatment) analysis of variance. + Denotes a significant effect of RAD140 treatment * Denotes a significant difference according to a Šídák's multiple comparisons test relative to baseline within the group. N(control + RAD140) = 18 (9 + 9) males, 9 (4 + 5) females. A horizontal line represents the mean

Diastolic function and transition times: Diastolic function was assessed by evaluating the flow of blood into the left ventricle immediately adjacent to the mitral valve during the early (E) and atrial (A) filling stages along with their ratio. Figure 4 visualizes representative images for this analysis, captured during the apical four-chamber view. An impaired heart may have difficulty filling, leading to a reduced E wave and subsequent lower E/A ratio. No significant differences were observed in these parameters (Fig. 5; p > 0.05 for all). Further, the ratio between E and E’ was evaluated (E’ is the velocity of the mitral valve annulus, which moves slower when the heart has difficulty relaxing). No effect of treatment was noted in E/E’ (p = 0.861; Fig. 5D).

Fig. 4.

Representative images of Doppler echocardiography adjacent to the mitral valve, as measured in the apical four-chamber view. Note: A = atrial wave; E = early wave; IVCT = isovolumic contraction time; and IVRT = isovolumic relaxation time

Fig. 5.

Left ventricular diastolic functional velocity and transition time parameters assayed using transthoracic echocardiography did not differ between older male and female mice with or without 6-weeks of treatment with RAD140 (panels A-D). Isovolumic relaxation time (IVRT) was shorter in RAD140 treated male mice, but not females (panel E). Differences in isovolumic contraction time (IVCT) between the RAD140 group and vehicle group were opposite between the sexes, with treated males tending to have longer times and treated females shorter (panel F). Parameters were tested with a 2-factor (sex × treatment) analysis of variance. × Denotes a significant (sex × treatment) interaction. * Denotes a significant difference according to a Šídák's multiple comparisons test relative to baseline within the group. Note: IVCT = isovolumic contraction time; IVRT = isovolumic relaxation time; PWD = pulse wave doppler; PWT = pulse wave tissue. N(control + RAD140) = 18 (9 + 9) males, 9 (4 + 5) females. A horizontal line represents the mean

However, the time spent transitioning from contraction into relaxation, termed the isovolumic relaxation time (IVRT), was shorter in RAD140 treated males but not females (Figs. 5E). A post hoc test revealed that the IVRT was significantly lower in males treated with RAD140 compared to male controls (p = 0.014), which trended in the opposite direction for females (treatment × sex p = 0.012). On the other hand, isovolumic contraction time (IVCT), which is the transition from relaxation to contraction, differed by sex and was longer in male mice treated with RAD140 compared to male controls, but was shortened in females (treatment × sex p = 0.022; Fig. 5F).

Left ventricular strain analysis: Global longitudinal strain (GLS) and global circumferential strain (GCS) were measured, which denotes the relative movement of the left ventricle in the longitudinal and circumferential direction each beat, respectively. Strain is in negative units because it denotes the shortening of the heart muscle, and larger negative values indicate greater myocardial movement (i.e., and improvement in function). Male SARM treated mice tended to have increased GLS (e.g., more negative) versus controls, whereas the opposite was observed in females, but this did not reach statistical significance (interaction p = 0.07; Fig. 6A). However, the difference in GCS was significantly affected by treatment and sex (sex × treatment p = 0.050; Fig. 6B), with male SARM treated mice having increased strain compared to controls in a post hoc analysis (p = 0.042). This was further reflected in the fractional area change (FAC) measured in the parasternal short axis, which denotes the change in the two-dimensional area of the left ventricle between diastole and systole, relative to the area at the end of diastole. Males treated with RAD140 had increased FAC compared to their controls (p = 0.044; Fig. 6C) and effect directions differed between the sexes (interaction p = 0.050).

Fig. 6.

Global longitudinal strain (GLS) was not statistically significantly affected in older male mice treated with RAD140 compared to their vehicle controls (panel A). Global circumferential strain (GCS) and fractional area change (FAC) as measured from the short axis during transthoracic echocardiography was greater (i.e., more movement, improved function) in male mice, while treated female mice tended to have the opposite relationship (panels B, C). Parameters were tested with a 2-factor (sex × treatment) analysis of variance. × Denotes a significant (sex × treatment) interaction. * Denotes a significant difference according to a Šídák's multiple comparisons test relative to baseline within the group. Note: GLS = Global longitudinal strain; GCS = Global circumferential strain; and FAC = Fractional area change. N(control + RAD140) = 18 (9 + 9) males, 9 (4 + 5) females. A horizontal line represents the mean

Myocardial strain was then separated into epicardial and endocardial sections for a deeper analysis. The magnitude of tissue velocity, strain, and strain rate were assessed in three myocardial perspectives (radial, longitudinal, and circumferential). Radial strain denotes the thickening of tissue during each beat, which happens as the muscle contracts, while the other two strains are described above. SARM treated males had increased tissue velocity and strain in the endocardium, and strain in the epicardium, in the short axis versus control mice (p < 0.05 for all; Supplementary Fig. 1). There were no significant differences between treated and control female mice in the short axis. In the longitudinal axis, RAD140 treated males and females differed in their response (Supplementary Fig. 2), with male mice showing greater epicardial longitudinal strain and strain rates versus controls, whereas the opposite was seen in females (sex × treatment p = 0.024 and 0.020, respectively). Endocardial circumferential tissue velocity, strain, and strain rate was again greater in SARM treated males versus controls (p < 0.05 for all; Supplementary Fig. 3), but no differences were observed in females.

Overall, myocardial strain was in general greater in RAD140 treated males versus controls, especially in the endocardium, but such effects were not seen or were in the opposite direction for females.

Myocardial strain and circulating cytokines: Because impaired myocardial strain can relate to higher levels of chronic inflammation [34, 35], GCS and GLS were correlated with circulating cytokines measured previously from this cohort of mice [32]. In male mice, higher levels of the pro-inflammatory cytokines IL-1α, IL-6, and MCP-1 related to worse (lower) GCS (p < 0.05 for all; Supplementary Table 2; Supplementary Fig. 4). The correlation between CGS with IL-6 (Fig. 7A) was notable, given that RAD140 treatment was related to lower level of IL-6 in males [32]. On the other hand, IL-6 and GLS in males were unrelated (Fig. 7B). No relationships were found between any cytokine and GCS or GLS in female mice (Supplementary Tables 2 & 3; Supplementary Fig. 5), nor were any relationships found between GLS and cytokines in males or females (Supplementary Table 3).

Fig. 7.

Higher levels of interleukin-6 (IL-6) correlated with a lower (i.e., less movement) amount of circumferential strain in male mice (Panel A; p = 0.016). This was not true for longitudinal strain (Panel B; p = 0.920). Data presented on a log scale and analyzed using a pooled Spearman correlation analysis. N(control + RAD140) = 18 (9 + 9) males

Effects of RAD140 on mRNA expression in ventricular tissue

Androgens are known to influence calcium handling and the expression of different myosin heavy chains in the heart [6]. Given the differences in cardiac function seen in RAD140 treated mice, genes encoding proteins relating to calcium handling were measured from ventricular samples. These included the sarcoendoplasmic reticulum calcium ATPase-2 (atp2a2), L-type calcium channel (cacna1c), calsequestrin (casq2), phospholamban (pln), and the ryanodine receptor type-2 (ryr2). mRNA levels of the cardiac signalling proteins β₁-adrenergic receptor (adrb1) and calcium/calmodulin dependent protein kinase kinase 2 (camkk2) were also measured, alongside genes for the AR (ar), two types of collagen (col1a1 and col3a1), and two types of myosin heavy chains common in the heart (myh6 and myh7). Only the expression levels of mRNA encoding for the AR differed after SARM treatment, with treated mice of both sexes having lower expression compared to vehicle treated mice (p = 0.015; Fig. 8). The other genes were not significantly affected (p > 0.05).

Fig. 8.

Gene expression levels were not different between RAD140 and control male and female groups (panels A, C-L) except for the androgen receptor (ar) gene levels, which were lower in male and female RAD140 treated mice compared to their vehicle controls (panel B). mRNA was isolated from ventricular heart tissue and analyzed relative to reference gene averages within each mouse (each circle). mRNA expression levels were tested with a 2-factor (sex × treatment) analysis of variance. + Denotes a significant effect of RAD140 treatment. N(control + RAD140) = 19 (10 + 9) males, 13 (6 + 7) females. A horizontal line represents the mean

Discussion

In this study, six weeks of RAD140 administration in 23-month-old C57Bl/6 mice was related to higher levels of left ventricular performance in males, evidenced by higher fractional shortening, FAC, cardiac output, and improved myocardial strain metrics. While RAD140 treated female mice were similar to males regarding differences in stroke volume, ejection fraction, and cardiac output, opposite differences were observed in myocardial strain, IVRT, and IVCT. Structural parameters such as ventricular wall thickness and mass were unchanged across both sexes, and the diastolic indices E/A and E/E′ remained unaffected. Incorporating circulating cytokines into the analysis revealed that lower IL-6 levels in RAD140 treated males correlated with better circumferential strain, suggesting an anti-inflammatory component to the functional improvements. RAD140 treatment did not relate to differences in the transcription of 11 genes involved in calcium handling, extracellular matrix, adrenergic signaling, and myosin heavy chain composition, although a downregulation of the androgen receptor transcript was observed. Overall, our results suggest that RAD140 may enhance cardiac performance in aged mice, especially males, while some effects on the heart in females may be deleterious.

RAD140 positively but differentially affected cardiac function in males and females

Because low circulating testosterone is related to worse cardiac function in humans (e.g., [5]) and in mice (e.g. [9],), the effects of RAD140 treatment on cardiac structure and function were evaluated using echocardiography. A notable finding from this study was the association between RAD140 treatment and enhanced parameters of cardiac function in older but otherwise healthy mice. Treated male mice had greater (better) myocardial strains compared to the vehicle controls. This may have particular importance into older age, as frailty in humans has been related to worse myocardial strain and lower ejection fraction in both men and women [36]. The increased ejection fraction and myocardial strain and strain rate aligns with what is known about AR-mediated effects on cardiomyocyte calcium handling and contraction. In isolated ventricular myocytes from male rats, gonadectomy reduced Ca²⁺ release and delayed reuptake, which was reversed with testosterone supplementation [23]. It has also been shown that testosterone helps increase adrenergic-mediated inotropic and lusitropic effects in rat hearts [24]. Clinically, similar androgen-driven effects on systolic performance have been reported in human males. One study randomized cancer patients to receive 100 mg of testosterone enanthate or placebo for seven weeks and reported an ejection fraction increase of 6.2% in the testosterone group compared to a 1.8% decline in placebo controls [37]. However, testosterone replacement therapy in the setting of heart failure for older men did not yield any increase in ejection fraction in two smaller clinical trials [38, 39]. The increased left ventricular function also contrasts with well documented evidence that prolonged anabolic steroid use (or during substance use disorder) is associated with impaired systolic function [40]. Given that androgen deprivation leads to similar outcomes [5], it is clear that an optimal dose of androgens or SARMs is required to optimize function.

The pooled analysis including males and females suggested that ejection fraction and cardiac output increased across both sexes, but understanding the impact specifically on females is more difficult. In a clinical trial of women with heart failure, transdermal testosterone treatment did not lead to any change in ejection fraction [41]. Left ventricular ejection fraction is determined by both myocardial strain and wall thickness [42]. Strain was increased in the male RAD140 group versus control, but groups were not substantially different in female mice. It is likely that small differences in systolic wall thicknesses (anterior and posterior) and ventricular diameters, with little change in diastolic thicknesses or diameter, led to the observed difference in cardiac output. This is because SV is directly determined by these parameters and is elevated in female mice on RAD140 versus vehicle treated controls, while HR was similar. Nonetheless, an increased sample size is needed in future trials to adequately power a study to elucidate the effects of RAD140 on female cardiac function.

The favorable quickening in IVRT in treated males was also notable. Lower testosterone levels have been shown to associate with longer IVRTs and IVCTs in older male mice with low testosterone [9]. This has also been shown in male rats, whereby gonadectomy delayed the transitions between left ventricular contraction and relaxation [43]. While it was not investigated in the current study, it is possible that RAD140 helped to shorten calcium transients in the myocardium by upregulating the phosphorylation of phospholamban, thereby slowing the sarco(endo)plasmic reticulum Ca²⁺-ATPase [44]. This would facilitate myocardial relaxation and shorten IVRT. Regarding the transition to contraction, there was no quickening of IVCT. The anaesthetized state may influence this finding, as isoflurane suppresses myocardial contractility [45] and the adrenergic potentiation of androgens would likely be absent without any exogenous stimulation [24]. A dobutamine challenge in RAD140 treated mice, for example, would be informative to assay the effects on transitioning from relaxation to contraction.

These results as a whole, which suggest that RAD140 improved cardiac function, must be interpreted alongside clinical evidence suggesting that SARM treatment may be detrimental or toxic to the heart under some circumstances. One case report suggested that RAD140 use in a young male contributed to left ventricular fibrosis, a lowered ventricular ejection fraction, and global hypokinesis that reversed following SARM discontinuation and appropriate pharmacotherapy [26]. This aligns with other clinical case reports of myocardial injury relating to SARMs [25, 26, 27, [28]. Preclinical research has also reported that primary cardiac fibroblasts isolated from male rat hearts treated with Ostarine, another SARM, increased proteins associated with fibrosis and reduced cell viability [25]. Notably, these effects were absent in cardiac fibroblasts derived from female rats [25], which may help explain the sex differences we observed.

On the other hand, a first-in-human clinical trial of another SARM, GSK2881078, assessed men and postmenopausal women for safety and reported no clinically meaningful changes in echocardiograms, electrocardiography, and cardiac biomarkers relating to toxicity [46]. This follows a 2025 systematic review of clinical trials involving SARMs that concluded there are “low rates and nonstatistically significant differences compared to placebo with cardiovascular, hepatotoxicity, prostate and haematologic [adverse events] at short to mid-term follow-ups” [47]. Overall, these findings support the hypothesis that at least some SARMs can be cardiotoxic, especially at high doses, with males potentially being affected to a greater degree than females. It is possible that our study was shorter or used a dose lower than required to reach a toxic level, but nonetheless our findings of enhanced performance must be considered alongside other evidence of toxicity or non-effect. The equivocal and relatively incomplete evidence across case reports and clinical trials underscores the need for more high-quality research involving SARMs to elucidate the cardiovascular benefits and risks of SARMs.

Systemic inflammatory markers, cardiac strain, and RAD140 treatment

The relationship between higher circulating cytokines in males, most notably IL-6, and reduced myocardial strain was also interesting and provides more evidence linking myocardial dysfunction to circulating cytokines. Low testosterone levels relate to higher IL-6 levels in men [48, 49] and IL-6 levels increase during androgen deprivation [50]. We previously described how RAD140 treated males, but not females had significantly lower levels of IL-6 than vehicle treated controls [32]. It is possible that RAD140 acted to suppress IL-6 production, although replication of these findings in a more controlled fashion is necessary to conclude this. A clinical trial investigating the impact of a different SARM, GTx-024 reported there was a non-significant decrease in the levels of IL-6 in the 1 mg per day treatment group [51]. However, little else has been reported regarding RAD140’s impact on circulating inflammatory markers, and testosterone supplementation does not clearly lower IL-6 levels [10]. Our investigations linking other circulating cytokine levels with myocardial strain yielded few other relationships (e.g., MCP-1 and IL1-α in males in the Supplemental data), and all three are considered to be pro-inflammatory. However, IL-6 was the only cytokine linked both with RAD140 treatment (lower levels with RAD140 treatment [32]) and changes in strain. These results, alongside the strong connection IL-6 has with frailty across humans and mice [52], further informs IL-6’s connection with age-related cardiac dysfunction and raises interesting questions about how IL-6 reduction may be beneficial in treating cardiovascular disease [53].

The connection between a heightened inflammatory state, including elevated IL-6, and worse myocardial performance is well known [54]. It may therefore not be surprising that better left ventricular function, measured by myocardial strain, was associated with lower IL-6 levels in our study. Higher IL-6 levels are linked with worse systolic function in healthy middle-aged men and women [34, 35] and are associated with poorer myocardial strain in children with multisystem inflammatory syndrome [55]. In men with chronic heart failure, higher testosterone levels were also associated with more favorable cardiac resynchronization therapy, which was paralleled by decreases in IL-6 levels [56]. Further, a knockout IL-6 mouse model was protected against a pressure-overload model of heart failure by inhibiting hypertrophy, fibrosis, and apoptosis [57]. In addition to IL-6, higher IL-1α and MCP-1, both of which were related to myocardial strain in this study, are linked with worse cardiac function [58, 59]. Neither of these cytokines were significantly different between RAD140 treated or vehicle groups of mice [32]. Whether RAD140 has sufficient immunomodulatory effects to disrupt the relationship between inflammation and cardiac dysfunction is not known, although the results from the present study suggest some linkage.

Study considerations and limitations

While our findings support RAD140’s potential to enhance cardiac performance in aged mice, they must be balanced against some emerging safety concerns and the constraints of our experimental design. The case reports and preclinical studies described earlier give some pause to the interpretation that RAD140 may facilitate improvements in the long-term, although a recent review of evidence from clinical trials does not raise the same intensity of concern [47]. For example, cardiac fibrosis has been linked to SARM use (e.g., [25, 26]), although no differences were noted in collagen gene expression in the present study. Testosterone supplementation has been shown to alleviate cardiac fibrosis in older mice [60], while other research has shown that testosterone replacement in castrated mice increased cardiac fibrosis relative to the gonadectomized group [61]. Such differential findings between testosterone, SARMs, and treatment context highlight the complexity and uniqueness of AR signaling across different ligands through variable cofactor recruitment [62]. Extracardiac toxicity or undesired effects of SARMs that indirectly lead to cardiovascular disease risk is another important safety consideration. Such outcomes include hypertension, brought about by AR activation and subsequent potentiation of the renin–angiotensin–aldosterone system [63]. On the other hand, evidence from nonhuman primates suggests SARMs may be able to mitigate other cardiovascular risk factors like dyslipidemia and excess fat mass [64]. Extending treatment duration beyond six weeks and incorporating established models of cardiac disease (e.g., pressure-overload or ischemia–reperfusion injury) will be essential to determine whether RAD140’s functional benefits persist or whether pathological remodeling emerges over longer dosing regimens.

Further, the lack of standardized dosing regimens is an important barrier to understanding the effects of RAD140 and how our dosing of 5 mg/kg translates to potential effects in humans. Although this approximate dose was previously demonstrated to minimize prostate growth while producing anabolic effects in skeletal muscle using young male rats [19], a direct translation of that dose to humans is difficult to estimate. This is related to the fact that limited pharmacokinetic data for RAD140 is available, with evidence in humans suggesting the half-life is 45 h in postmenopausal women with metastatic breast cancer using doses of 50, 100, and 150 mg per day [65], with no readily available data for mice. The rate of metabolism in mice, however, is generally higher relative to body weight, due to a greater liver-to-body proportion and substantially more cytochrome P450 enzyme families [66, 67]. However, dose-dependent lean-mass growth in young nonhuman primates has been reported following 28-days of 0.01, 0.1, and 1 mg/kg RAD140 administration, although sample sizes (n = 3) were small and statistical significance was not reached [19]. Combined, it can be reasoned that doses as low as 0.01 mg/kg may be physiologically relevant in human health, although specific doses that may affect cardiac function are unelucidated. The first in-human trial reported low incidence rates of sinus tachycardia, atrial flutter, pericardial effusion, palpitations, and right ventricular failure, but stated that all except palpitations were assessed to be unrelated to RAD140.

Lastly, several considerations limit the generalizability of our results. First, we relied solely on transthoracic echocardiography at the end of the study for functional assessment. While this method is non-invasive and clinically relevant, additional functional information could be gleaned from electrophysiological or hemodynamic measurements that invasive catheterization and pressure–volume loop analysis would reveal [68]. The lack of baseline data also impairs any estimation of change in function over time. Second, we did not measure endogenous testosterone levels. Although mice of this age and strain likely have lowered testosterone levels at this age [31], direct confirmation would clarify whether RAD140’s actions were in a hypogonadal or eugonadal state. For example, lower circulating testosterone levels relate to impaired transition time in similarly aged mice [9] and impair cardiac calcium handling [69]. Knowing the testosterone levels in the Vehicle groups would help to inform whether RAD140 was replacing natural effects of testosterone or providing a unique contribution. Thus, it is a significant limitation to not have the endogenous testosterone concentrations. Third, our cohort size, particularly for females, was modest and thus limited statistical power to detect sex-dependent effects, despite several being identified. Fourth, the biologic significance of the differences in the AR gene is difficult to interpret without accompanying protein quantification. Further investigation into effects on protein expression would be valuable. Similarly, more detailed histological investigations, as well as expanding gene expression analysis to include TGBF1 and CCN2, and performing in-situ analyses would better explore any effects RAD140 on mechanisms of cardiac fibrosis. Lastly, we tested RAD140 in otherwise healthy mice, leaving unanswered questions about its efficacy and safety in the context of common cardiac comorbidities such as hypertension or diabetes. Future studies should expand sample sizes, include both sexes across age groups, confirm baseline hormonal status, and employ disease models to more fully characterize RAD140’s cardiac risk–benefit profile before considering translational applications.

Conclusion

In summary, six weeks of RAD140 administration in aged male C57BL/6 mice was associated with greater left ventricular systolic performance and a shortening of active relaxation (i.e., IVRT), which included improved myocardial strain that was graded by circulating IL-6 levels. Both female and male mice treated with RAD140 had higher ejection fraction and cardiac output versus their controls. Unlike males, females did not show higher strain or changes in transition times. In fact, there were opposite trends in these parameters for females versus males, which may be statistically identifiable with a higher sample size. There were no detected differences in mRNA levels of select genes relating to function or structure, indicating that alternative approaches are needed to understand any effects. The sex-specific outcomes underscore the complexity of SARM action in the aging heart across sexes and highlight several avenues for future work, including longer treatment durations, dose–response evaluations, inclusion of young and disease-model cohorts, and deeper mechanistic studies of AR signaling and inflammatory modulation. Such investigations will be essential to fully define the safety, efficacy, and translational potential of RAD140 and related SARMs as therapy for age-related cardiovascular disease.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization: Stefan S. Heinze and Susan E. Howlett; Methodology: All authors; Formal analysis and investigation: Stefan S. Heinze; Writing—original draft preparation: Stefan S. Heinze; Writing—review and editing: All authors; Funding acquisition: Susan E. Howlett; Resources: Susan E. Howlett; Supervision: Susan E. Howlett.

Funding

During this work SSH was supported by a Canada Graduate Scholarship – Doctoral Research Award (Canadian Institutes of Health Research); a Level II Izaak Walton Killam Predoctoral Scholarship; a Scotia Scholars Award (Nova Scotia Health Research Foundation), and a MacDonald Graduate Studentship (Dalhousie Medical Research Foundation). This work was supported by grants to SEH from the Canadian Institutes for Health Research to (PGT 155961) and the Heart and Stroke Foundation of Canada (grant number G-22–0031992). APY was supported by an Alexander Graham Bell Canada Graduate Scholarship from the Natural Sciences and Engineering Research Council of Canada.

Data Availability

Specific data may be made available upon reasonable request to the corresponding author.

Declarations

Conflict of interest

All authors declare no conflicts of interest.