Swimming is superior to running in inducing physiological cardiac hypertrophy and enhancing myocardial performance

Abstract

Aerobic exercise training (AET) can induce cardiac hypertrophy, but the specific adaptive response for different types of AET remains unclear. We evaluated nonsingular cardiac remodeling in rats through running (RT) and swimming (ST) training at approximately 75% of VO₂max. Male Wistar rats (8–10 weeks old; ~ 250 g) were divided into untrained (UT), RT, and ST groups. The RT and ST were performed five days a week, once daily for 60 min for eight weeks. Cardiopulmonary fitness was assessed by measuring maximal oxygen consumption and swimming time to exhaustion. Echocardiography evaluated left ventricular parameters, while myocardial mechanics were assessed through the papillary muscle. Histology and Western blotting were performed to evaluate cardiomyocyte size and proteins modulating phosphoinositide 3-kinase (PI3K_110α)/AKT1 signaling. Real-time PCR was used to assess the expression of genes and microRNAs involved in myocardial hypertrophy. Both AET protocols enhanced cardiopulmonary capacity, but only the ST group showed increased myocardial mass, cardiomyocyte growth, and LV cavity size, along with greater tension and papillary muscle shortening velocity. A more pronounced alteration in gene expression pattern for proteins modulating PI3K_110α/AKT1 signaling was found in the ST group than in the RT group. A similar difference was also found for microRNA 1, 21, 27a, 124, and 144 expressions. ST is more effective than RT in inducing cardiac hypertrophy and enhancing contractility, linked to the PTEN-AKT-S6K1 pathway and increased expressions of microRNAs 1, 21, 27a, 124, and 144. Thus, ST is superior to RT for inducing physiological cardiac hypertrophy.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36818-2.

Introduction

Aerobic exercise training (AET) can induce cardiac hypertrophy (CH)¹. In contrast to pathological CH, AET induces an increase in cardiac mass with preserved or enhanced myocardial function². Different modalities of AET have been reported to induce CH (e.g., running, cycling, swimming, rowing, and cross-country skiing)³, and different types of AET may induce a nonunique pattern of cardiac remodeling. A previous study revealed that triathletes and cyclists had greater heart mass than marathon runners⁴. In addition, the athletes did not develop a homogeneous hypertrophic phenotype but rather an eccentric or concentric pattern of left ventricular (LV) remodeling.

In animal models, running (RT) and swimming (ST) are two popular exercise forms to analyze cardiac hypertrophy^5,6. Schaible and Scheuer⁷ pioneered the comparison of the hypertrophic pattern between two AET modalities and reported that ST was superior to RT in increasing cardiac mass and LV performance. To the best of our knowledge, no studies have directly investigated, in a head-to-head manner, the molecular mechanisms that may account for the distinct patterns of CH observed between RT and ST. In this context, although both RT and ST have been well documented to improve myocardial inotropism^8,9, this has not been analyzed in a direct comparison within the same experimental setting. Consequently, this study examined cardiac remodeling in rats undergoing RT compared to those undergoing ST regimens.

Experiments were performed to examine LV structural and functional changes and in vitro myocardial inotropism. We also evaluated the differential gene and protein expression profiles of the components of the phosphoinositide 3-kinase (PI3K_110α)/AKT_p110α signaling pathway, which is known to be involved in AET-induced CH^10,11. Moreover, we evaluated the differential profile of MicroRNAs (miRNAs) induced by RT and ST. Although several studies have examined miRNAs expression regulated by AET^12,13, less is known about their expression patterns directly compared between RT and ST. A single study by Liu et al.¹⁴ documented the differential expression of several miRNAs in mice subjected to voluntary wheel running or ST. Here, we aimed at the comparative expression of miR‑1, miR‑21, miR‑27a, miR‑124, miR‑143, and miR‑144 in the myocardium of rats subjected to RT and ST. These miRNAs play pivotal roles in angiogenesis, the regulation of cardiomyocyte growth and contractility, the modulation of inflammatory pathways, vascular remodeling, and cytoskeletal organization, as well as in oxidative stress responses and cardioprotection. Notably, their expression has been shown to be modulated by exercise training, highlighting their potential involvement in exercise-induced cardiac remodeling and adaptation^15,16. Therefore, this study provides new insights into how different modalities of AET distinctly regulate CH and myocardial function. By directly comparing RT and ST under the same experimental conditions, we demonstrate that ST is more effective than RT in promoting CH and enhancing inotropic performance, and we link these adaptations to differential regulation of the PTEN-AKT-S6K1 signaling pathway and specific miRNAs.

Results

Cardiopulmonary fitness

All groups had an increase in body weight during follow-up (Fig. 1A). The training program spanned a total of nine weeks, including an initial exercise adaptation period, with sessions lasting 60 min per day, five days per week, at an intensity of approximately 75% of VO₂max. The VO₂max of the RT and ST groups significantly increased by more than 5%, whereas that of the untrained group decreased (Fig. 1B). In addition, the trained rats swam longer during the exhaustive test, and the two exercise protocols were equally effective in increasing muscle citrate synthase activity (Fig. 1C-D). Exercise training did not affect the resting heart rate (Fig. 1E).

Fig. 1.

(A) Body weight, (B) maximal oxygen uptake (VO_2max, animals performed the exercise test for 14–16 min to reach VO₂max and (C) swimming test (N = 16 per experimental group). (D) citrate synthase activity (N = 8 per experimental group); (E) resting heart rate (N: UT = 6; RT = 8; ST = 7). ^#p < 0.05 vs. the UT group.

Effects of AET on LV structure and function

Swimming training induced an eccentric hypertrophy phenotype, as evidenced by increased LV diameters (Table 1). As assessed by echocardiography or hemodynamics, exercise training did not alter LV systolic or diastolic properties. Figure 2A-B shows that the ST group had significantly greater heart weight/body mass values and LV wet weight/body mass than the other groups. AET did not change right ventricular mass (Fig. 2C). A greater transverse fiber diameter and nuclear volume of cardiomyocytes accompanied the increase in myocardial mass (Fig. 2D-E). Compared with those in the UT group, the biometric and histologic findings of the rats in the RT group were similar.

Table 1.

LV morphology and function. Echocardiographic (N = 16 per experimental group) and hemodynamic (N: UT = 14; RT = 16; ST = 16) analysis of the left ventricle (LV). BM, body mass; CO, cardiac output; CI, cardiac index; HR, heart rate; LVEDD, LV end-diastolic diameter; LVEDP, LV end-diastolic pressure; LVESD, LV end-systolic diameter; LVSP, LV peak systolic pressure; SF, shortening fraction; SV, stroke volume; SVI, stroke volume index; SW, systolic work; +dP/dt, maximal positive time derivatives of the LV developed pressure; -dP/dt, maximal negative time derivatives of the LV developed pressure. ^#p < 0.05 vs. UT. ^&p < 0.05 vs. RT group.

	UT	RT	ST
Echocardiography
LVEDD (mm)	6.5 ± 1	6.9 ± 0.5	7.3 ± 0.4
LVESD (mm)	3.7 ± 0.4	3.5 ± 0.4	4.2 ± 0.3&
LVEDD/BM (mm/g)	0.022 ± 0.007	0.022 ± 0.005	0.029 ± 0.007#&
LVESD/BM (mm/g)	0.010 ± 0.002	0.011 ± 0.003	0.015 ± 0.004#
SF (%)	46 ± 5	44 ± 2	48 ± 3
E-wave (cm2)	0.66 ± 0.11	0.66 ± 0.04	0.65 ± 0.06
A-wave (cm2)	0.31 ± 0.04	0.37 ± 0.07	0.35 ± 0.07
E/A ratio	2 ± 0.3	1.9 ± 0.3	1.8 ± 0.2
Hemodinamics
HR (bpm)	321 ± 45	316 ± 27	312 ± 32
PLVSP	129 ± 7	127 ± 10	119 ± 10
LVEDP (mmHg)	6 ± 1	6 ± 1	7 ± 1
+dP/dt (mmHg/s)	8609 ± 890	8881 ± 1526	9801 ± 2166
-dP/dt (mmHg/s)	6976 ± 2054	7293 ± 1622	6411 ± 1453
CO (mL/min)	60 ± 17	60 ± 7	50 ± 6
CI (mL/min/kg)	147 ± 42	150 ± 19	133 ± 7
SV (mL/bpm)	0.19 ± 0.05	0.20 ± 0.02	0.16 ± 0.02
SVI (mL/kg/bpm)	0.47 ± 0.13	0.47 ± 0.08	0.43 ± 0.05
SW (g·m/min)	0.32 ± 0.10	0.30 ± 0.06	0.25 ± 0.04

Fig. 2.

Effects of exercise training on biometric parameters (N = 16 per experimental group) and histological markers (N = 7 per experimental group) of myocardial hypertrophy. (A) HW/BW, heart weight/body weight; (B) LV/BW, left ventricular weight/body weight; (C): RV/BW, right ventricular weight/body weight; (D) FW, fiber width (a total of 20 cells were analyzed per animal); (E) NV, nuclear volume of cardiomyocytes (a total of 40 nuclei were analyzed per animal). Representative light micrographs of myocardial section were stained with haematoxylin–eosin. Bars in the micrographs indicate fiber size. Arrows indicate the cardiomyocyte nucleus. ^#p < 0.05 vs. UT. ^&p < 0.05 vs. RT group. Scale bar = 50 μm.

AET improves myocardial mechanics

There were marked differences in myocardial performance among the three groups. Swimming training resulted in muscles that developed more force (Fig. 3A). Furthermore, +dT/dt and -dT/dt were significantly greater in the ST group than in the UT and RT groups (Fig. 3B-C). Running training had a small effect on the inotropic properties of the myocardium, resulting in an increase in + dT/dt compared with UT animals. None of the exercise modalities changed resting tension (Fig. 3D).

Fig. 3.

In vitro papillary muscle performance (N: UT = 14; RT = 16; ST = 16). (A) Peak developed tension (DT); maximal (B, +dT/dt) positive and (C, −dT/dt) negative time derivative of the DT; (D) resting tension (RT). ^#p < 0.05 vs. UT. ^&p < 0.05 vs. the RT group.

Effects of exercise training on the expression of Akt pathway genes and MicroRNAs

We observed that trained mice presented increased mRNA levels of PI3K_110α and AKT1 (Fig. 4). However, the PTEN, TSC2, mTOR, S6K1, and 4EBP1 levels were decreased in both exercise groups. A significant reduction in GSK-3β was observed only in the RT group.

Fig. 4.

Myocardial gene expression (N = 8 per experimental group). (A) 110 kDa catalytic subunit of phosphoinositide 3-kinase (PI3K_110α); (B) phosphatase and tensin homolog (PTEN); (C) serine/threonine kinase 1 (AKT); (D) TSC complex subunit 2 (TSC2); (E) mammalian target of rapamycin (mTOR); (F) glycogen synthase kinase 3 beta (GSK-3β); (G) ribosomal protein S6 kinase beta-1 (P70S6); (H) eukaryotic translation initiation factor 4E binding protein 11 (4EBP1). ^#p < 0.001 vs. UT. ^&p < 0.05 vs. RT group.

Cardiovascular microRNAs are differentially expressed with exercise training and are involved in cardiac remodeling¹⁷. Compared with the UT protocol, both AET protocols significantly increased the expression of microRNAs 1, 21, 27a, 124, 143, and 144 (Fig. 5). Notably, swimming training upregulated microRNAs 1, 21, 27a, 124 and 144 in the myocardium compared to UT and RT groups.

Fig. 5.

Myocardial miRNA expression (N = 8 per experimental group). (A) miRNA-1; (B) miRNA-21; (C) miRNA-27a; (D) miRNA-124; (E) miRNA-143; (F) miRNA-144. ^#p < 0.0001 vs. UT. ^&p < 0.05 vs. RT group.

The AET mode promotes the differential expression of Akt signaling

Serum IGF-1 expression was not affected by RT or ST (Fig. 6). The AET mode resulted in different expression parameters of Akt pathway proteins (Fig. 7). While PTEN levels were significantly reduced in the swimming group, PI3K_110α was upregulated in the RT and ST groups (Fig. 7). Downstream targets of PI3K/Akt, mTOR, and GSK-3β were examined and may be critical regulators of exercise-induced CH¹⁸. Figure 7 shows significant myocardial downregulation of TSC2, mTOR, pmTOR, pmTOR/mTOR ratio, pGSK-3β, p70S6, and p4EBP1 in swim-trained rats. Running training decreased only the levels of TSC2, mTOR and pGSK-3β. Both AET protocols increased the GSK-3β/pGSK-3β ratio, but only ST promoted a higher pP70S6/P70S6 ratio.

Fig. 6.

Serum insulin-like growth factor type I (IGF-1) concentrations in the UT, RT, and ST groups (N = 10 per experimental group).

Fig. 7.

Myocardial protein expression of the Akt pathway was evaluated by immunoblotting (N = 5–8 per experimental group). (A) Phosphatase and tensin homolog (PTEN); (B) 110 kDa catalytic subunit of phosphoinositide 3-kinase (PI3K_110α); (C) serine/threonine kinase 1 (AKT); (D) phospho-AKT (pAKT); (E) pAKT/AKT ratio; (F) TSC complex subunit 2 (TSC2); (G) phospho-TSC2 (pTSC2); (H) pTSC2/TSC2 ratio; (I) mammalian target of rapamycin (mTOR); (J) phospho-mTOR (pmTOR); (K) pmTOR/mTOR ratio; (L) glycogen synthase kinase 3 beta (GSK-3β); (M) phospho-GSK-3β (pGSK-3β); (N) GSK-3β/pGSK-3β ratio; (O) ribosomal protein S6 kinase beta-1 (P70S6); (P) phospho-S6K1 (pP70S6); (Q) pP70S6/P70S6 ratio; (R) eukaryotic translation initiation factor 4E binding protein 11 (4EBP1); (S) phospho-4EBP1 (p4EBP1); (T) p4EBP1/4EBP1 ratio. ^#p < 0.05 vs. UT. ^&p < 0.05 vs. RT group.

Expression of MHC subunits

The α- and β-MHC subunits are closely related to the myocardial contractile velocity¹³. As shown in Fig. 8, AET did not significantly alter β-MHC mRNA or protein expression. Cardiac α-MHC mRNA levels increased with both AET protocols, but their protein levels were similar to those of untrained controls.

Fig. 8.

Gene (A-B) and protein (C-D) expression of the beta (β-MHC) and alpha (α-MHC) myosin heavy chains in the myocardium (N = 4–6 per experimental group). ^#p < 0.001 vs. UT group.

Discussion

Our study demonstrates for the first time that swimming training not only induces more robust cardiac hypertrophy than running but also enhances myocardial contractility through distinct regulation of the PI3K/Akt, mTOR, and GSK-3β pathways and specific miRNAs. These direct head-to-head comparisons reveal modality-specific molecular and functional adaptations, providing new insights into the mechanisms underlying exercise-specific cardiac remodeling.

In this study, swimming training led to improved functional fitness, as measured by VO₂max and swimming endurance test, compared to running training. These findings contrast those of clinical studies indicating that RT and ST result in minimal performance transfer from one modality to another^19,20. A similar adaptive response to the two aerobic modalities may relate to the skeletal muscle demands during exercise. Humans are bipedal, so lower limb muscles are more engaged during running, whereas swimming involves greater upper limb action²⁰. However, considering that rats are quadrupedal, running and swimming training require a substantial engagement of forelimb and hindlimb muscles²¹.

The citrate synthase activity in skeletal muscle has been used as a marker of cellular oxidative capacity and mitochondrial density following exercise training²². Data from this study confirm that muscle citrate synthase activity is highly responsive to AET and illustrate that the adaptive response is independent of training modality²³. Thus, greater citrate synthase activity has been reported for both exercise modalities and should be considered as potential outcome in future investigations involving RT and ST protocols in rats.

Treadmill running and swimming training did not induce functional adaptations in the LV. Similar LV echocardiographic and hemodynamic findings were reported in rats that ran for a total of 13 weeks or swam for eight weeks^9,24. The lack of changes in the LV functional pattern at baseline between the experimental groups does not exclude the possibility of adaptations to AET in assessing cardiac reserve. For example, hemodynamic stress has been advocated as a usual approach to identify overt cardiac dysfunction in several experimental models²⁵. Therefore, our findings of higher VO_2max may result from improved cardiac performance during exercise.

Previous data have shown improved myocardial inotropism without improvement in LV performance after AET⁹. Therefore, we evaluated myocardial contractile capacity in the papillary muscle preparation, and our results are consistent with an improved contractile state after AET. ST was the only AET mode that resulted in an improvement in DT, which was accompanied by an increase in cardiomyocyte width. This is important because cell morphometry is a determinant of cardiomyocyte tension development²⁶. MHC-α and MHC-β have significantly different enzymatic properties, with α having a higher rate of contraction^27,28. However, since increased gene expression of α-MHC did not translate into increased protein expression in our experiments, the increase in + dT/dt induced by AET cannot be explained by differences in the expression of MHC isoforms. The phosphorylation of sarcomeric myosin regulatory light chain (RLC) and myosin binding protein C (cMyBP-C) proteins may increase the maximal rate of force development in cardiomyocytes^29–31. Considering that AET promotes more significant phosphorylation of RLC and cMyBP-C^32,33, one might expect both AETs to have resulted in greater + dT/dt in papillary muscles. The -dT/dt ratio is a marker of myocardial relaxation that increased only in the ST group. The mechanisms involved in the myocardial relaxation effect of exercise are not fully understood. Still, the more significant phosphorylation of phospholamban, implying greater removal of cytoplasmic calcium to the sarcoplasmic reticulum through calcium-ATPase channels, may play an important role³⁴.

The LV echocardiographic and histomorphometric findings are consistent with the literature, which indicates that ST increases the myocardial mass³⁵. These findings corroborate the results of the study by Schaible and Scheuer⁷, who reported that ST was more effective in increasing cardiac mass than RT. In contrast to our study, however, Schaible and Scheuer⁷ did report increased cardiac mass from baseline with RT. These differences can be attributed to the higher weekly training volume in the study by Schaible and Scheuer⁷, where the animals underwent two 75-minute training sessions per day. Additionally, other investigators using exercise durations similar to those used in our study reported either an increase in myocardial mass^36,37 or no effect of RT^23,38,39. Furthermore, the increase in the LV diastolic/systolic diameter-to-BW ratio in our study shows that the ST mode primarily induced volume overload-induced cardiac eccentric hypertrophy with predominant longitudinal myocyte growth⁴⁰. Although myocyte length was not measured in the present study, we observed an increase in myocyte width and nuclear volume, paralleling changes in cardiac and LV/BW ratios. This finding is consistent with the idea that ET-induced eccentric hypertrophy may result in proportional increases in myocyte width and length⁴¹.

We investigated the expression of several miRNAs previously reported to be modulated by AET. An elevated levels of miRNA-1 was detected in the trained animals, as previously described by other researches⁴², although this finding is not consistent across the literature⁴³. We also observed increased expression of miRNAs 21, 27a, and 144, consistent with studies showing similar changes in running or swimming models^44–46. In contrast, our results for miRNAs 124 and 143 diverge from prior studies reporting reduced expression in trained rodents^16,47,48. These discrepancies may be related to differences in exercise intensity and duration. For example, no changes in miRNA-143 were reported after low-intensity, moderate-volume training (60 min, 5 days/week, 10 weeks)⁴⁶, whereas reduced expression was found in female rats subjected to swimming with 5% overload (60 min, 5 days/week, 8 weeks)⁴⁶.

A greater expression of miRNA-1, 21, 27a, 124 and miRNA-143 was detected in the ST group than in the RT group. These miRNAs have been implicated in the regulation of angiotensin II. An increase in miRNA-27a leads to decreased expression of the gene encoding the enzyme that converts angiotensin I to angiotensin II, whereas increased miRNA-143 decreases the expression of ACE2, which converts angiotensin II to angiotensin 1–7¹. Thus, it is possible that the higher levels of miRNA-27a/143 after AET may be associated with lower levels of angiotensin II and angiotensin 1–7. Nevertheless, angiotensin II has not been implicated in physiological cardiac hypertrophy⁴⁸. Using a similar ST intensity to the present study, other investigators demonstrated lower levels of miRNA-27a, which was associated with a reduction in angiotensin II⁴⁶. On the other hand, reports are showing a trend toward a reduction in the expression of miRNA-143 accompanied by an increase in ACE2 and angiotensin 1–7 after ST⁴⁶. Finally, activation of the angiotensin 1–7/Mas receptor axis does not seem to affect ST-induced cardiac hypertrophy but prevents the detrimental deposition of extracellular matrix proteins in the LV⁴⁹.

ST has been shown to upregulate miRNA-21 and miRNA-144, where they can inhibit PTEN expression^16,50,51. Interestingly, in this study, despite the increased expression of miRNA-21 and miRNA-144 in the ST group, the PTEN content was lower than that in the RT group; however, PTEN expression was reduced only in the ST group. PTEN has the opposite effect on PI3K_110α, causing the dephosphorylation of phosphatidylinositol-3,4,5-trisphosphate (PIP3) to phosphatidylinositol-4,5-bisphosphate (PIP2)⁵². The activation of PI3K_110α leads to the phosphorylation of AKT by PIP3⁵³. The increase in PI3K_110α expression after AET is well known⁵⁴; therefore, both AET modalities increase the expression of PI3K_110α and pPI3K_110α. Here, the increased expression of PI3K_110α occurred despite the increased expression of miRNA-124, an inhibitor of PI3K_110α expression⁵⁵. Despite the increased expression of PI3K_110α in the ST and RT groups, only the ST group presented the highest phosphorylation of AKT, which may be due to the decreased dephosphorylation of PIP3 by PTEN in the ST group. AKT phosphorylation can inhibit GSK-3β and TSC2, which, when activated, inhibit protein synthesis⁵⁴.

Several growth factors are involved in the activation of PI3K_110α to induce physiological cardiac hypertrophy, including IGF-1, insulin, thyroid hormone, vascular endothelial growth factor, and neuregulin 1². We did not observe changes in the serum concentration of IGF-1 with RT or ST, suggesting that other pathways may be responsible for greater PI3K_110α activation after AET.

In this study, although no changes were observed in the basal content of GSK-3β and TSC2, both AET protocols promoted a reduced phosphorylation of the GSK-3β at the Ser9 residue. This finding may indicate an increased GSK-3β activity, as suggested by the GSK-3β/pGSK-3β ratio. GSK-3β activation induces the export of GATA-4 from the nucleus and prevents the transcription of genes associated with hypertrophic signaling⁵⁶. The increased expression of miRNA-1 can decrease GATA-4 and Mef2a levels, which inhibits nuclear factor activation of activated T cells (NFAT)⁵⁷. Here, the expression of miRNA-1 was increased in the ST group. The increase in miRNA-1 and GSK-3β activity with exercise suggests a lower level of GATA-4 in the nucleus. However, this matter needs to be investigated because we did not examine the translation of GATA-4. Phosphorylated TSC2 inhibits mTOR1, a serine/threonine kinase that regulates many cellular processes (e.g., activation of translation and transcription via phosphorylation of S6K1 and 4EBP1, respectively)⁵⁸.

AET did not alter mTOR or 4EBP1 activity but resulted in increased S6K1 activity in the ST group. Thus, the PTEN-AKT-S6K1 pathway appears to be involved in ST-induced cardiac hypertrophy. It is difficult to understand why there was no increase in mTOR activation in the ST group, since mTOR stabilizes S6K1⁵⁹ and is, therefore, a key protein in the PTEN-AKT-S6K1 pathway.

Conclusions

The present study revealed that ST induced cardiac hypertrophy and improved myocardial inotropism more efficiently than RT. Myocardial hypertrophy in ST is associated with the downstream PTEN-AKT-S6K1 pathway and increased expression of miRNAs 1, 21, 27a, 124, and 144. Our results suggest that ST is superior to RT in inducing physiological cardiac hypertrophy.

Methods

Animals

This study was conducted in accordance with the “Guide for the Care and Use of Laboratory Animals,” which was prepared by the Institute of Laboratory Animal Resources and published by the National Institutes of Health (NIH Publication No. 86 − 23, revised 1996). It also followed the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. The experimental protocol received approval from the Institutional Research Ethics Committee of the Federal University of São Paulo (no. 9209020321). Male Wistar rats (8–10 weeks old; ~ 250 g) from the central animal facility of the Federal University of Sao Paulo were randomly assigned to one of the following experimental groups: UT, untrained rats (n = 24); ST, swimming training (n = 24); RT, running training (n = 24).

AET protocols

AET was performed as described by Schaible and Scheuer⁷. Rats swimming with no load attached to their tails were assumed to exercise at approximately 75% of their maximal oxygen consumption (VO₂max)^60,61. The rats swam in a tank (132 cm in diameter and 80 cm deep) filled with tap water heated to 32–34 °C via a feedback-controlled electric heating coil⁶². The water was maintained in turbulence to provide continuous swimming behavior and not to float. First, the rats were acclimated to the pool for six consecutive days, with exercise sessions limited to 10 min on the first day and increasing by 10 min each day. Then, ST rats swam five days a week, once daily for 60 min for eight weeks. The RT rats ran on a motorized treadmill (CL4002, Caloi, São Paulo, Brazil) for the same duration. In the adaptation phase, the animals ran for 10 min at a speed of 10 m/min and a gradient of 2%. Each day, the exercise period was increased by 10 min, with the speed and elevation increasing by 2 m/min and 2%, respectively. Then, the rats ran five times per week (60 min per day) at a speed of 20 m/min at an 8%, which would achieve a similar percentage of VO₂max⁷. Animals with “stop and run” behavior were excluded from the study.

VO₂max determination

The VO₂max was assessed via a motorized treadmill coupled with a gas analyzer (Panlab, Harvard Bioscience Company, MA, USA). Before the exercise test, animals were acclimated to running as previously described by our group⁶². Each rat underwent a 2-minute warm-up period at 25 cm/s, and the running speed was increased by nine cm/s every 2 min until exhaustion. A steady state of oxygen uptake with progressive increases in running speed and a respiratory exchange ratio of ≥ 1.05 were considered to define VO₂max. Analyses were performed at baseline and the end of the experimental protocol.

Exercise capacity swimming test

Exercise capacity was assessed via an exhaustive swim test⁶². A load equivalent to 10% of body weight was placed around the waist of each rat, and the animals were individually observed to determine the time required for exhaustion. Exhaustion was defined as the point at which the rats could not swim to the surface for ten seconds, and investigators would intervene after that time.

Echocardiography

Twenty-four hours after the last exercise session, the rats were anesthetized (ketamine plus xylazine mixture (50 and 10 mg/kg), i.p.), and a transthoracic echocardiogram was performed with a 12 MHz transducer (Sonos-5500, Hewlett-Packard, MA, USA)⁹. The LV end-diastolic diameter (LVEDD) and LV end-systolic diameter (LVESD) were measured. Systolic function was analyzed via the three transverse planes’ shortening fraction (SF). Pulsed-wave Doppler analysis of mitral valve inflow provided flow velocity data to determine diastolic function parameters (E waves, A waves, and the E/A ratio).

LV hemodynamics

LV hemodynamic studies were performed 24 h after echocardiography. The rats received urethane (1.2 g/kg i.p.) and were intubated and ventilated (rodent ventilator, model 683, Harvard Apparatus, Holliston, MA, USA). A Millar SPR-320-F-gauge catheter tip micromanometer (Millar Instruments, Houston, TX, USA) was inserted into the LV cavity through the right carotid artery^9,63. Moreover, an ultrasound flow probe (Transonic System Inc., Ithaca, NY, USA) was positioned in the ascending aorta⁶⁴. Measurements of the LV parameters included cardiac output (CO), cardiac index (CI), end-diastolic pressure (LVDP), heart rate (HR), maximal positive (+ dP/dt), and negative (-dP/dt) time derivatives of developed pressure, stroke volume (SV), stroke volume index (SVI), systolic pressure (LVSP), and systolic work (SW).

Euthanasia method and myocardial mechanics evaluation

Immediately after the hemodynamic study, the animals were euthanized with an overdose of urethane (4.8 g/kg, i.p.). The hearts were quickly removed. LV posterior papillary muscles were isolated and prepared as previously described⁶². The muscles were isometrically contracted for 15 min and stretched to the apices of their length-tension curves. Data were recorded via AcqKnowledge 3.5.7 software (Biopac Systems Inc., Goleta, CA, USA) to determine the maximum developed tension (DT), maximum rate of tension increase (+ dT/dt) and decrease (-dT/dt), and resting tension (RT).

Biometric and histomorphometric analysis

Following euthanasia, the hearts were rapidly removed and weighed. The LV was separated from the rest of the heart, weighed, and sectioned transversely at the midventricular level. The basal portion was fixed in a 10% formalin-buffered solution for light microscopy. Myocardial tissue was sectioned at a thickness of 7 μm and stained with hematoxylin-eosin. Histological images were visualized using an Olympus microscope at 40× magnification, and analyzed using Image Tool software 3.0. In longitudinal sections, cardiomyocytes with preserved nuclei were measured to determine nuclear volume³⁸. Cardiomyocytes with visible nuclei and intact membranes were selected for diameter measurement, which was obtained by manually tracing the mid-nuclear region¹⁶.

Citrate synthase assay

Soleus muscles were homogenized in phosphate buffer (50 mM sodium phosphate, 1 mM EDTA, and protease inhibitor cocktail; Sigma‒Aldrich, São Paulo, SP, Brazil) and centrifuged at 12,000×g and 4 °C for 15 min. The pellet was discarded, and the supernatant was used for the citrate synthase activity assay via a standard kit (Sigma‒Aldrich, MO, USA). The sample absorbance at 412 nm was monitored in 96-well plates with a SpectraMax M5 spectrophotometer (Molecular Devices, CA, USA).

Gene expression

LV samples (0.5 g) were homogenized in TRIzol^® reagent for RNA extraction and prepared to obtain cDNA as previously described in detail⁶⁵. Amplification and data acquisition were performed via Abi Prism 7500 Fast (Applied Biosystems, Waltham, MA) with the SYBR Green Core Reaction Kit (Applied Biosystems) to calculate the relative mRNA levels of each target gene: 110 kDa catalytic subunit of PI3K_110α, phosphatase and tensin homolog (PTEN), TSC complex subunit 2 (TSC2), AKT1, pAKT1, glycogen synthase kinase 3 beta (GSK-3β), mammalian target of rapamycin (mTOR), eukaryotic translation initiation factor 4E binding protein 1 (4EBP1), ribosomal protein S6 kinase beta-1 (P70S6), myosin heavy chain alpha (MHC-α) and myosin heavy chain beta (MHC-β). Differentially expressed genes were normalized to the expression level of the housekeeping genes GAPDH and 18 S subunit ribosomal RNA.

MiRNA profiling

The cDNA for microRNA analysis was synthesized from total RNA via specific primers according to the TaqMan microRNA assay⁶⁶. The 15 µL reactions obtained via the TaqMan miRNA Reverse Transcription Kit protocol (Applied Biosystems, CA, USA) were incubated in a thermal cycler (Applied Biosystems, CA, USA) for 30 min at 16 °C, 30 min at 42 °C, 5 min at 85 °C, and then held at 4 °C continuously. The 20 µl PCR mixture contained 10 µl of TaqMan Universal PCR Master Mix II, 1.33 µl of RT product, 7.67 µl of nuclease-free water, and 1 µl of primer and probe mixture from the TaqMan microRNA assay protocol for miRNA-1, −21, 27a, 27b, 124, 143, 144, and 145. Reactions were performed at 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 60 s. Samples were normalized to the small nuclear RNA U6 (RNU6) gene.

Enzyme-linked immunosorbent assay (ELISA)

The frozen serum was homogenized in phosphate-buffered saline plus a proteinase inhibitor cocktail (Sigma Chemical, St. Louis, MO, USA). The homogenates were subjected to ELISA via a specific commercial kit (R&D Systems, USA) to evaluate insulin-like growth factor 1 (IGF-1).

Western blotting

Frozen LV tissue (50 mg) was homogenized as previously described⁶², and 30 µg of homogenate was prepared for transfer to hydrophobic polyvinylidene membranes (Hybond-P, Amersham Biosciences; Piscataway, NJ, USA). The membranes were incubated overnight at 4 °C with the following rabbit primary antibodies (Cell Signaling, Massachusetts, USA): anti-PTEN (1:500), anti-PI3K_p110α (1:1000), anti-Akt1 (1: 5000), anti-phosphoAkt1_Ser473(1:1000), anti-GSK3β (1:1000), anti-phospho-GSK-3β_Ser9 (1:1000), anti-TSC2 (1:1000), anti-phosphoTSC2_Thr1462 (1:1000), anti-mTOR (1:2000), anti-phospho-mTOR_Ser2448 (1:1000), anti-S6k1 (1:1500), anti-phospho-S6k1_Thr389 (1: 2000), anti-4EBP1 (1:2000),and anti-phospho-4EBP1_Thr37/46 (1:1000). Moreover, mouse primary antibodies against MHC-α (1:2000, Sigma Aldrich, MO, USA), MHC-β (1:2000, Sigma Aldrich), and GAPDH (1:10000, Abcam, Cambridge, United Kingdom) were also used. The membranes were washed five times and incubated with horseradish peroxidase (HRP)-conjugated rabbit anti-mouse antibody (1:2000; Jackson Immuno, West Grove, PA) for 60 min. Bound antibody was detected using an enhanced chemiluminescence reagent for 1 min. The bands were visualized and digitized using the Amersham Imager 600 system (GE Health Care, Little Chalfont, Pittsburgh, PA, USA) and quantified using ImageQuant™ TL software. Identical amounts of protein were loaded into each well of the gel, GAPDH levels were used as loading controls and to normalize the data.

Statistical analysis

The data are presented as the means ± standard deviations of the means. The Shapiro‒Wilk test was used to evaluate Gaussian distributions. Levine’s test was applied to characterize the homogeneity of variances. Data with a normal distribution were analyzed using one-way ANOVA followed by the Newman‒Keuls post hoc test, except for body weight data, which were analyzed using two-way ANOVA with Bonferroni post hoc correction. Nonparametric data were analyzed using the Kruskal‒Wallis test complemented by Dunn’s test. All analyses were performed via Graph Pad Prism software (version 8.0.2; CA, USA). The level of significance adopted was p ≤ 0.05.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AET

Aerobic exercise training

CH

cardiac hypertrophy

LV

left ventricular

PI3K

phosphoinositide 3-kinase

AKT

protein kinase B

VO₂max

maximal oxygen consumption

EDTA

ethylenediaminetetraacetic acid

LVEDD

left ventricular end-diastolic diameter

LVESD

left ventricular end-systolic diameter

SF

shortening fraction

CO

cardiac output

CI

cardiac index

LVDP

left ventricular end-diastolic pressure

HR

heart rate

+dP/dt

maximal positive time derivatives of developed pressure

-dP/dt

maximal negative time derivatives of developed pressure

SV

stroke volume

SVI

stroke volume index

LVSP

left ventricular systolic pressure

SW

systolic work

FCSD

fiber cross-sectional diameter

HR

heart rate

HW/BW

heart weight/body weight

LV/BW

left ventricular weight/body weight

NV

nuclear volume of cardiomyocytes

RV/BW

right ventricular weight/body weight

DT

developed tension

+dT/dt

maximum rate of tension increase

-dT/dt

maximum rate of tension decrease

RT

resting tension

PTEN

phosphatase and tensin homolog

TSC2

TSC complex subunit 2

PI3K_110α

110 kDa catalytic subunit of PI3K

pAKT1

phosphorilated protein

GSK-3β

glycogen synthase kinase 3 beta

mTOR

mammalian target of rapamycin

4EBP1

eukaryotic translation initiation factor 4E binding protein 1

P70S6

ribosomal protein S6 kinase beta-1

MHC-α

myosin heavy chain alpha

MHC-β

myosin heavy chain beta

GAPDH

Glyceraldehyde 3-phosphate dehydrogenase

IGF-1

insulin-like growth factor 1

Author contributions

Concept and design: AY, DSB. Acquisition, analysis, or interpretation of data: ELA, LDS, MTDS, FLM, RSF. Drafting of the manuscript: AY, AJS. Critical manuscript revision for important intellectual content: LDS, JASJ, PJFT, ACCG, AJS. Statistical analysis: BLM. Supervision: AJS.

Funding

This work was supported by the São Paulo Research Foundation, FAPESP [grant numbers 2013/20011-7, 2014/08273-9, 15/11028-9]; the National Council for Scientific and Technological Development, CNPq [grant number 300199/2025-2].

Data availability

The datasets used and analyzed in the current study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval

The Institutional Research Ethics Committee of the Federal University of São Paulo reviewed and approved the study.