Relationship between intermuscular adipose tissue infiltration and myostatin before and after aerobic exercise training

Abstract

Intermuscular adipose tissue (IMAT) is associated with impaired skeletal muscle contractile and metabolic function. Myostatin and downstream signaling proteins such as cyclin-dependent kinase 2 (CDK2) contribute to the regulation of adipose and skeletal muscle mass in cell culture and animals models, but this relationship remains incompletely understood in humans. The purpose of this study was to determine if the infiltration of IMAT was associated with skeletal muscle myostatin and downstream proteins before and after 12 wk of aerobic exercise training (AET) in healthy older women (OW; 69 ± 2 yr), older men (OM; 74 ± 3 yr), and young men (YM; 20 ± 1 yr). We found that the infiltration of IMAT was correlated with myostatin and phosphorylated CDK2 at tyrosine 15 [P-CDK2(Tyr15)]. IMAT infiltration was greater in the older subjects and was associated with lower skeletal muscle function and exercise capacity. After 12 wk of AET, there was no change in body weight. Myostatin and P-CDK2(Tyr15) were both decreased after AET, and the reduction in myostatin was associated with decreased IMAT infiltration. The decrease in myostatin and IMAT occurred concomitantly with increased exercise capacity, skeletal muscle size, and function after AET. These findings demonstrate that the reduction in IMAT infiltration after AET in weight stable individuals was accompanied by improvements in skeletal muscle function and exercise capacity. Moreover, the association between myostatin and IMAT was present in the untrained state and in response to exercise training, strengthening the potential regulatory role of myostatin on IMAT.

INTRODUCTION

Exercise capacity and lower extremity muscle power production are predictors for mobility, morbidity, and mortality (19, 44). With increasing age the loss of skeletal muscle size is associated with lower physical function. However, the deficit in exercise capacity and muscle power production persists even when normalized to lean mass or skeletal muscle cross-sectional area (CSA) (16). Therefore, other tissues and/or systems must also contribute to the decline in functional capacity and muscle power production in sedentary adults, especially with increasing age.

Intermuscular adipose tissue (IMAT) is located under the skeletal muscle fascia and in between muscle groups (13). In older adults, increased IMAT is negatively associated with lower extremity muscle performance and increased risk of mobility limitations (6, 7, 9, 51). To account for differences in muscle size among groups, conditions, or disease states, the infiltration of IMAT is often expressed relative to skeletal muscle CSA (5, 28). The infiltration of IMAT is not specific to the elderly, as young individuals rendered physically inactive by unilateral limb suspension demonstrated a significant increase in in both thigh and calf IMAT infiltration (35). These data suggest the infiltration of IMAT may be related to prolonged physical inactivity.

Conversely, greater levels of physical activity were associated with less IMAT infiltration (54) and have been shown to prevent further infiltration of IMAT (15). However, high levels of IMAT infiltration were accompanied with decreased quadriceps activation (60) and blunted skeletal muscle hypertrophy after resistance exercise (38). Previous studies in sedentary overweight/obese adults have shown that weight loss can decrease IMAT (12, 26). However, maintaining body weight is important with increasing age as weight loss is believed to place older adults at greater risk for sarcopenia, frailty, and mortality by accelerating skeletal muscle loss and/or limiting skeletal muscle hypertrophy after exercise (18). Therefore, it remains unknown if 1) aerobic exercise training can decrease IMAT independent of weight loss in normal weight individuals and 2) if a decrease in IMAT infiltration is accompanied with improvements in function (7).

The molecular and cellular signaling cascades that are associated with IMAT infiltration are also not well defined. Several factors secreted from skeletal muscle, such as cytokines and myostatin, may act on both skeletal muscle and noncontractile tissues in an autocrine, paracrine, and/or endocrine fashions (48). Myostatin, a member of the transforming growth factor-β (TGF-β) superfamily, has been proposed to be involved in the cross talk between skeletal muscle and surrounding adipose tissue (10, 58). Inhibition of myostatin signaling specifically in skeletal muscle but not adipose tissue has been shown to decrease fat mass and increase skeletal muscle mass in transgenic mice (17). Additional evidence indicates myostatin may promote adipogenesis and inhibit myogenesis in vitro by decreasing the activity of proteins like cyclin-dependent kinase 2 (CDK2) that favor cell cycle progression and increasing the activity of the cell cycle inhibitors (p21 and p53) (2, 10, 29). Collectively, these data imply that skeletal muscle myostatin can regulate both skeletal muscle and adjacent adipose tissue depots. However, the relationship between myostatin signaling proteins and IMAT infiltration has not been examined in humans.

The purpose of this study was to determine if 1) IMAT infiltration is associated with skeletal muscle myostatin signaling proteins; 2) there is an association between the change in IMAT and myostatin signaling proteins after 12 wk of aerobic exercise training (AET); and 3) a decrease in IMAT infiltration is accompanied by improved function, independent of weight loss. We chose to study young men (YM; 20 ± 1 yr; n = 7), older men (OM; 74 ± 3 yr; n = 6), and older women (OW; 69 ± 2 yr; n = 9) to specifically target normal weight individuals with increasing levels of IMAT infiltration.

METHODS

Subjects

Young men (YM; 20 ± 1 yr; n = 7), older men (OM; 74 ± 3 yr; n = 6), and older women (OW; 69 ± 2 yr; n = 9) were recruited from the local community. Before participation, all subjects provided written informed consent approved by the Institutional Review Board at Ball State University. YM provided a detailed medical and physical activity history while OM and OW completed a thorough physical examination that included blood chemistry profile, pulmonary function, and resting and exercise electrocardiograms. Subjects were excluded based on the following criteria 1) body mass index (BMI) ≥28 kg/m²; 2) type 1 or type 2 diabetes; 3) uncontrolled hypertension; 4) active cancer, cancer in remission, or having received treatment for any form of cancer in the previous 5 yr; 5) coronary artery disease; 6), cardiovascular disease (e.g., peripheral arterial disease, peripheral vascular disease); 7) abnormal thyroid function; 8) engaged in regular aerobic or resistance exercise more than one time per week for 20 min or longer during the previous year; 9) chronic and/or regular nonsteroidal anti-inflammatory drug consumption; and 10) any condition that presents a limitation to exercise training. Four OM were on cholesterol-lowering medications (i.e., statins), three were on blood pressure medications (non-β-blocker), and five were on medications for prostate health.

General Study Design

Eligible subjects underwent a series of baseline measurements for the determination of aerobic capacity using a graded exercise test on a cycle ergometer, thigh skeletal muscle and IMAT CSA assessed via magnetic resonance imaging (MRI), whole muscle power production of the knee extensor muscle group, and a muscle biopsy for the assessment of myostatin signaling proteins. Upon completion of the baseline assessments, subjects completed 12 wk of supervised aerobic exercise training performed on a stationary, upright cycle ergometer. The details of this training program have been described previously (20, 33). Briefly, during the first 7 wk, the participants progressively increased frequency (3–4 days per week), duration (20–40 min), and intensity (60–75% heart rate reserve), with the last 5 wk of exercise performed 4 days per week, 45 min per session, and at 80% heart rate reserve. All participants had a 100% exercise adherence. Following the training intervention, all subjects repeated the testing procedures that occurred at baseline.

Experimental Procedures

Aerobic capacity/graded exercise test.

Subjects performed graded exercise tests on a cycle ergometer for the assessment of maximal aerobic capacity (V̇o_2max) and peak exercise workloads (W) before and after the 12-wk aerobic training intervention as previously described (20, 22). Peak exercise workload was defined as the highest workload achieved during the V̇o_2max test.

Magnetic resonance imaging.

Proton MRI of the thigh was measured before and after the 12-wk aerobic intervention as previously described (20, 22). Scans were obtained after 1 h of supine rest to avoid the influence of fluid shifts (4). Subjects were positioned with an adjustable foot restrain for fixation of joint angles and thus muscle lengths. A standard of 1% CuSO₄ was placed along the length of the leg, such that it appeared in the field of view of all images to eliminate bias in viewing images, resulting from day-to-day variations in the magnetic field. MRIs were coded and transferred to a personal computer in a blinded fashion. IMAT, subcutaneous adipose tissue (SAT), and skeletal muscle CSA were determined by National Institutes of Health software Medical Image Processing, Analysis and Visualization (MIPAV) following the detailed procedures previously published by Buford et al. (6). Briefly, to correct for varying shading caused by gradient-driven eddy currents, a nonparametric, nonuniform intensity normalization (N3) algorithm was used within the MIPAV program. Subsequently, the bone and SAT were removed from the image by creating a region of interest around these tissues. Once the bone and SAT were removed, muscle and IMAT were segmented using an automated, objective pixel clustering function using fuzzy-c-means. This allows pixels to be clustered to an indicated class (e.g., skeletal muscle, IMAT, and background) via a membership function that bins pixel intensities to a membership value of 0 or 1. If the pixel intensity is close to the median of the indicated class, then it receives a score of 1. Pixel intensities far from the median receive values close to zero. This approach is employed to remove subjectivity of identifying muscle and adipose tissue pixels. We analyzed a midthigh image that approximated the location of the skeletal muscle biopsy sample five times to achieve a coefficient of variation <1% as we have previously performed for measurements of the quadriceps femoris CSA (20, 22, 30).

Skeletal muscle power production.

Peak power of the knee extensor muscle group was assessed using an inertial ergometer connected to a strain gauge load cell and potentiometer interfaced with a personal computer (20, 33). After multiple orientation sessions for familiarization, subjects performed two to three sessions separated by more than 2 days. After 10 min warm-up on a stationary bicycle, subjects completed two submaximal repetitions followed by three maximal attempts with 3-min rest between sets. The maximal dynamic power output during the concentric contraction was recorded throughout the range of motion.

Muscle biopsy procedure.

Percutaneous muscle biopsy samples were obtained from the vastus lateralis muscle. Muscle samples were dissected free of visible adipose and connective tissue and frozen in liquid nitrogen. Posttraining muscle biopsy samples were obtained 48 h after the last exercise session.

Western blotting.

Western blotting was performed using methods previously described (23, 30, 32). Proteins (20 μg) were resolved on a 4–12% gel via electrophoresis for 75 min at 100 V and transferred to PVDF for 2 h at 100 mA per gel. Membranes were blocked with 5% BSA-Tris-buffered saline with 0.1% Tween-20 (TBST) for 60 min and incubated overnight in primary antibodies with gentle agitation at 4°C. Primary antibodies were purchased from Cell Signaling [CDK2 2546; P-CDK2(Thr160), 2561; p53 2527; acetyl-p53(Lys382) 2525; SAPK/JNK 9552; P-SAPK/JNK(Thr183/Tyr185) 9251)], Abcam [P-CDK2 (Tyr15) Ab76146; P-CDK2(Thr14) Ab68265], or Millipore (Myostatin AB3239]. After 3× 5 min TBS with rinses, blots were incubated with anti-rabbit HRP conjugated-secondary antibody (7074; Cell Signaling), rinsed 3× 10 min of Tris-buffered saline with 0.1% Tween-20, and exposed to an enhanced chemiluminescent substrate. Digital images were captured using a chemiluminescent imaging system (FluorChem SP). Verification of equal protein loading was accomplished by Ponceau staining. Molecular weights of immunodetected proteins were confirmed by molecular weight markers SeeBlue 2 and Magic Mark (Invitrogen). Each subjects pre- and posttraining samples were analyzed on the same blot and samples were loaded in alternating pattern of OW, OM, and YM to control for interassay variability.

Statistics

Data are presented as means ± SE, and a two-tailed significance was set at P < 0.05. Data were analyzed using a two-way (group × time) ANOVA with repeated measures for time with a Greenhouse-Geisser correction for violation of sphercity when present. The absolute change from pre- to posttraining between groups was analyzed utilizing a one-way ANOVA. A Bonferroni post hoc test was completed when applicable. Pearson’s correlation coefficient (r) was used to determine associations between dependent variables. SPSS was used for ANOVAs, and GraphPad Prism 7.0 was used for the Pearson’s correlation and least squares regression.

RESULTS

Body Weight and BMI

Body weight was different between groups (Table 1), but no significant differences were detected for BMI. By design, body weight and BMI were not significantly different after 12 wk of AET (Table 1).

Table 1.

Subject characteristics

	Young Men (n = 7)		Older Men (n = 6)		Older Women (n = 9)		P Values
	Pre	Post	Pre	Post	Pre	Post	Time	Group	Interaction
Body weight, kg	84.6 ± 4.5	84.0 ± 4.0	82.3 ± 4.3	81.5 ± 2.8	68.0 ± 3.1	67.3 ± 4.1	P = 0.24	P = 0.01	P = 0.99
BMI, kg/m2	26.4 ± 1.6	26.2 ± 1.5	26.1 ± 1.4	25.8 ± 1.1	25.1 ± 1.7	24.9 ± 1.4	P = 0.20	P = 0.81	P = 0.94
Peak exercise workload, W	261 ± 12	313 ± 13	143 ± 6	172 ± 9	87 ± 5	122 ± 6	P < 0.01	P < 0.01	P = 0.02
Skeletal muscle power, W	925 ± 68	955 ± 73	509 ± 45	522 ± 53	273 ± 49	328 ± 50	P = 0.02	P < 0.01	P = 0.57
SAT CSA, cm2	65 ± 11	56 ± 8	39 ± 4	39 ± 3	92 ± 15	88 ± 13	P = 0.06	P = 0.01	P = 0.25
Thigh skeletal muscle CSA, cm2	145 ± 6	150 ± 6	118 ± 6	123 ± 6	77 ± 5	83 ± 5	P < 0.01	P < 0.01	P = 0.97
Normalized peak workload, W/cm2	1.81 ± 0.07	2.10 ± 0.08	1.22 ± 0.08	1.41 ± 0.09	1.13 ± 0.05	1.50 ± 0.06	P < 0.01	P < 0.01	P = 0.05
Normalized skeletal muscle power production, W/cm2	6.38 ± 0.40	6.37 ± 0.37	4.31 ± 0.24	4.25 ± 0.25	3.41 ± 0.47	3.89 ± 0.46	P = 0.14	P < 0.01	P = 0.05

Values are means ± SE. BMI, body mass index; CSA, cross-sectional area; Pre, baseline; Post, after training; SAT, subcutaneous adipose tissue. Analysis was performed by a two-way repeated measures ANOVA with a Bonferroni post hoc test. A Greenhouse-Giesser correction was implemented. P < 0.05, main effect for time, group, and/or interaction.

Peak Exercise Workload

Significant main effects (P < 0.05) for group, time, and interaction were present for peak exercise workload and when peak exercise workload was expressed relative to thigh skeletal muscle CSA (normalized peak exercise workload) (Table 1). The increase in peak exercise workload after 12 wk of AET was greater in YM versus OM (P = 0.02). Conversely, the increase in normalized peak exercise workload after 12 wk of AET was greater in the OW vs. OM (P = 0.05).

Skeletal Muscle Size and Power Production

Significant main effects (P < 0.05) for group and time were present for thigh skeletal muscle CSA and skeletal muscle power production (Table 1). Thigh skeletal muscle CSA and knee-extensor power production were lower in the older groups compared with the YM. Thigh skeletal muscle CSA and maximal muscle power production were increased after 12 wk of AET (Table 1).

Significant main effects (P < 0.05) for group and time were also present when skeletal muscle power production was normalized to thigh skeletal muscle CSA (normalized maximal muscle power production). After 12 wk of AET there was a nonsignficant increase (P = 0.14) in normalized maximal muscle power production that was driven by a within group difference (P < 0.01) in OW.

IMAT

Before AET, IMAT infiltration was inversely correlated with normalized peak exercise workload (Fig. 1A) and normalized whole muscle knee extensor power production (Fig. 1B). Signficant (P < 0.05) main effects for group, time, and interaction were present for IMAT infiltration (Fig. 1C). Signficant (P < 0.05) main effects for group and time were present for IMAT CSA (Fig. 1D).

Fig. 1.

A and B: intermuscular adipose tissue (IMAT) infiltration was inversely correlated to normalized peak exercise workload (A) and normalized maximal skeletal muscle power production (B). C–E: IMAT infiltration, the ratio of IMAT cross-sectional area (CSA) cm²/thigh skeletal muscle CSA cm², was different among groups (C) and decreased with aerobic exercise training (AET; D) and IMAT CSA was greater in older men (OM) and older women (OW) compared with younger men (YM) and decreased after AET (E). The decrease in IMAT infiltration after AET was different among groups with the OW having a greater decrease in IMAT infiltration compared with YM (P = 0.01). F: the decrease in IMAT CSA was not significantly different among groups. G: representative MR images of the thigh. Subcutaneous adipose tissue and the femur have been removed and segmentation of IMAT (white) and skeletal muscle (black) is presented. Pearson’s correlation coefficient (r) and least squares regression (r²) were performed to determine associations between variables. A two-way ANOVA was performed to determine effects for group, time, and group by time interaction. A Bonferroni post hoc was performed with a Greenhouse-Geisser correction. A one-way ANOVA was used to determine differences in the absolute change from pre- to posttraining among groups; young men (YM; n = 7), older men (OM; n = 6), and older women (OW; n = 9). ^aP < 0.05 vs. YM; #P < 0.05 main effect for time.

The infiltration of IMAT was lower after AET (P < 0.05, Fig. 1C), and the decrease was more pronounced (P < 0.01) in OW compared with YM (Fig. 1E). IMAT CSA was also decreased (P < 0.05, main effect for time, Fig. 1D) after 12 wk of AET (OW: −0.79 ± 0.38; OM: −0.77 ± 0.51; YM: −0.19 ± 0.68 cm²; Fig. 1F). Representative MRIs of IMAT and skeletal muscle are shown in Fig. 1G.

Myostatin and Downstream Signaling Proteins

The activity of CDK2 is regulated by both myostatin and the phosphorylation status of its two inhibitory sites [tyrosine 15 (Tyr15) and threonine 14 (Thr14)] and one activation site [threonine 160 (Thr160)]. We found that myostatin protein content (Fig. 2A) and P-CDK2(Tyr15) (r = 0.45, P < 0.05, data not shown) were positively correlated with IMAT infiltration. Total CDK2 protein content and the ratio of Tyr15 phosphorylated to total CKD2 were not related to IMAT infiltration. Myostatin (P = 0.09, Fig. 2B) and P-CDK2(Tyr15) (P < 0.01) were different between groups while there were no significant differences between the ratio of phosphorylated (Tyr15) to total CKD2. There were no significant differences among groups nor associations to IMAT infiltration for other select proteins downstream of myostatin such as total, acetylated, or the ratio of acetylated to total p53; total, phosphorylated (Thr160 and Tyr14), or the ratio of phosphorylated to total CDK2 (Thr160 and Tyr14); or total, phosphorylated, or the ratio of phosphorylated to total SAPK/JNK (data not shown).

Fig. 2.

A: myostatin was associated with intermuscular adipose tissue (IMAT) infiltration. B: myostatin tended (P = 0.09) to be different between groups and decreased (P < 0.05, main effect for time) after aerobic exercise training (AET). C: the change in myostatin was not statistically different among groups. D: the decrease in myostatin was associated with the decrease in IMAT infiltration. E and F: representative Western blots (E) and Ponceau S staining (F). Pearson’s correlation coefficient (r) and least squares regression (r²) were performed to determine associations between variables. AU: arbitrary units. A two-way ANOVA was performed to determine effects for group, time, and group by time interaction. A Bonferroni post hoc was performed with a Greenhouse-Geisser correction. A one-way ANOVA was used to determine differences in the absolute change from pre- to posttraining among groups; young men (YM; n = 7), older men (OM; n = 6), and older women (OW, n = 9). #P < 0.05 main effect for time.

After 12 wk of AET, myostatin (Fig. 2C), P-CDK2(Tyr15), and the ratio of phosphorylated Tyr15 to total CDK2 were decreased (P < 0.05). The decrease in P-CDK2(Tyr15) was greater in OW vs. YM. The decrease in myostatin was correlated to the decrease in IMAT infiltration (r = 0.41, P = 0.05; Fig. 2D) while the decrease in P-CDK2(Tyr15) and the ratio of phosphorylated Tyr15 to total CDK2 was not. For the change in myostatin, one participant was a high-responder and had a large decrease in myostatin. Upon removal of this participant, there tended to be a relationship (P = 0.10; r = 0.30) between the change in myostatin and the change in IMAT infiltration. P-CDK2(Thr160) was increased (P < 0.05, data not shown) after 12 wk of AET but not when expressed relative to total CDK2. Representative blots for proteins of interest and onceau S staining are shown in Fig. 2, E and F, respectively. There were no significant differences in total, acetylated, or the ratio of acetylated to total p53; total, phosphorylated (Tyr14), or the ratio of phosphorylated to total CDK2 (Tyr14); and total, phosphorylated ,or the ratio of phosphorylated to total SAPK/JNK after 12 wk of AET (data not shown).

DISCUSSION

We report a novel finding that the infiltration of IMAT is correlated to greater skeletal muscle myostatin and P-CDK2(Tyr15) in sedentary individuals and that the training-induced change in IMAT infiltration is accompanied by reductions in myostatin and P-CDK2(Tyr15). Furthermore, this is the first study to demonstrate that in the absence of significant weight loss, older adults can reduce IMAT following aerobic exercise training. The reduction in IMAT also occurs concomitantly with improved measures of function. These data suggest that IMAT infiltration is related to skeletal muscle myostatin and support future research to determine if targeting myostatin or downstream proteins can causally decrease IMAT infiltration and increase skeletal muscle function.

As expected, there was increased infiltration of IMAT in older vs. younger sedentary adults and the infiltration of IMAT was correlated to a decline in normalized skeletal muscle power production and peak exercise workload. These findings are congruent with previous investigations demonstrating that older adults and those with chronic diseases have greater IMAT infiltration and functional impairments (36). Lower skeletal muscle function may be due to less contractile tissue per unit volume. However, previous research has shown that IMAT infiltration is associated with impaired force generation due to altered pennation angle of muscle fiber bundles (14, 60). In addition to the potential alterations of skeletal muscle architectural properties by IMAT, communication between skeletal muscle and adipose tissues may occur via myokines, such as myostatin, and potentially lead to diminished function.

We show for the first time that skeletal muscle myostatin and CDK2 phosphorylation at Tyr15 were associated with the infiltration of IMAT. While myostatin is known to be a negative regulator of skeletal muscle growth by decreasing muscle protein synthesis, there is also evidence that myostatin may act as common regulator of both myogenic and adipogenic cells (2, 10). It is important to note that skeletal muscle samples contain postmitotic myofibers and an abundance of proliferating mesenchymal stem/stromal cells (MSCs). Myostatin and CDK2 Tyr15 phosphorylation inhibit the cell cycle progression from G1 Phase to S Phase, blocking DNA synthesis for myogenic proliferation (53). Conversely, myostatin and other members of the TGF-β superfamily have been shown to stimulate nonmyogenic MSCs such as pericytes and fibroadipogenic progenitor (FAP) cells leading to ectopic adipose deposition and fibrosis (11, 57, 61). Pericytes and FAP cells that contribute to fat accumulation and fibrosis are positive for platelet derived growth factor receptor α (PDGFRα⁺) (27, 55, 56). Myostatin signaling proteins, perhaps through PDGF, are associated with the infiltration of IMAT and we provide evidence that these select proteins within the myostatin signaling cascade are amenable to AET.

After AET, alterations in myostatin, P-CDK2(Tyr15), and P-CDK2(Thr160) occurred concomitantly with decreased IMAT infiltration and increased skeletal muscle size. These results are consistent with a cross-sectional study by Mikkelsen et al. (42), who demonstrated that trained runners had lower noncontractile tissue and myostatin mRNA expression than their sedentary counterparts. Work by our group and others has consistently shown acute and chronic aerobic exercise decreased myostatin mRNA expression (24, 30, 47). Deletion, inhibition, or natural mutations of myostatin can lead to a doubling in muscle mass and decreased fat mass (10, 17, 34, 40, 41). In lieu of genetic manipulation, we showed that AET can produce a more physiologically relevant decrease in myostatin mRNA (−49%) (30) and protein (~11%), with the latter being correlated to decreased IMAT infiltration. In addition to myostatin, increased P-CDK2(Thr160), an activation site, and decreased P-CDK2(Tyr15), an inhibitory site, may allow increased CD2K activity, DNA synthesis (i.e., cell proliferation) (46), and/or protein synthesis (21, 45, 50) after aerobic exercise training. Future studies are required to determine if myostatin and CDK2 may contribute to changes in IMAT infiltration and skeletal muscle size and function after aerobic exercise training by influencing MSC proliferation and/or post-mitotic muscle fiber protein synthesis.

Several studies have shown decreased IMAT after weight loss, as reviewed in Ref. 1, including exercise-induced weight loss in middle-aged, obese, men and women (12). However, in older adults who are not obese, such as the subjects in the current study, weight loss may blunt skeletal muscle hypertrophy and place individuals at a greater risk for skeletal muscle loss and frailty (3, 18). We show that 12 wk of AET decreased IMAT CSA and infiltration in weight stable participants and increased thigh skeletal muscle CSA, peak exercise capacity, and maximal skeletal muscle power production. Therefore, this study provides two clinically relevant findings that 1) AET decreased IMAT in weight-stable individuals and 2) a decline in IMAT infiltration was accompanied by improved skeletal muscle size and function. Emerging evidence suggests IMAT may also be reduced after resistance exercise (37); however, it appears those with greater IMAT levels had an impaired hypertrophic response after resistance exercise training (38). This may mitigate the potential beneficial effects of resistance exercise in obese or older individuals prone to high levels of IMAT infiltration and who are in the greatest need of decreased IMAT and improved skeletal muscle function. We found that despite presenting with the highest IMAT infiltration before training, OW achieved a similar absolute increase in thigh and quadriceps femoris (31) skeletal muscle size after AET compared with YM and OM. Collectively, these data suggest that AET is an effective exercise mode to decrease IMAT and increase exercise capacity, skeletal muscle size, and skeletal muscle power production in weight-stable individuals.

Limitations

We acknowledge the limitations of a relatively small sample size (n = 22); however, this cohort was recruited with the goal of studying participants across a continuum of IMAT infiltration. Without the inclusion of a young group of women, the direct influence of age cannot be fully ascertained. While each subject served as their own sedentary control before starting the study, we did not include a nonintervention control group. Since age and sedentary behavior are associated with increased IMAT infiltration, not including a nonintervention control group may have limited our understanding of the impact of exercise on decreasing and/or preventing the infiltration of IMAT during 12 wk of sedentary behavior.

Perspectives and Significance

The data in the current study demonstrate a novel correlation between IMAT infiltration and myostatin signaling proteins yet importantly do not imply causation. We believe these associations provide an important framework for future studies to determine if myostatin is obligatory in the regulation of IMAT infiltration with and without AET (Fig. 3). The repeated contractile activity of exercise elicits a profound influence on the surrounding skeletal muscle microenvironment and MSCs. Future studies should determine if a decline in myostatin signaling after exercise lowered IMAT infiltration by decreased adipogenic and fibrogenic potential of PDGFRα⁺ MSCs. We hypothesize that a reduction in the myostatin-PDGF signaling pathway may contribute, through unknown mechanisms to the diminished IMAT infiltration and increased muscle function. In support of this conjecture, inhibition of myostatin by a bioneturalizing antibody (59), TGF-β1 by a small molecule inhibitor (SB431542) (8), and PDGF signaling by Imatinib (49) has been shown to suppress fatty infiltration after anterior cruciate ligament or rotator cuff tear, respectively. However, nonselective inhibition of PDGFRα⁺ and PDGFRβ⁺ after 3–10 days of synergistic ablation inhibited skeletal muscle hypertrophy in young, healthy mice (52). The divergent effects of PDGRFα signaling that favor adipogenesis and fibrosis during damage, disease, and perhaps sedentary aging but promote regeneration during normal healthy conditions (27) appear to be mediated by polyadenylation of an intronic transcriptional variant of PDGRFα (43). Future research is warranted to determine if myostatin and PDGF signaling mediate the observed decrease in IMAT infiltration and accompanied increase in exercise capacity and skeletal muscle size and function.

Fig. 3.

A conceptual figure to provide a proposed framework for how myostatin may mediate intermuscular adipose tissue (IMAT) infiltration during sedentary aging and after aerobic exercise training. Future research is needed to test this hypothesis. MSC, mesenchymal stem cells; MPS, muscle protein synthesis; IMAT, intermuscular adipose tissue; CSA, cross-sectional area.

We demonstrate that increased IMAT infiltration at baseline and the reduction in IMAT infiltration after AET are related to skeletal muscle myostatin protein content. Furthermore, AET decreased IMAT infiltration with a concomitant increase in skeletal muscle power production and peak exercise workload. Overall, these adaptations to AET occurred without significant weight loss and may have a host of functional and metabolic health benefits for sedentary adults, especially with increasing age.

GRANTS

This work was supported by the National Institute on Aging Grant R15-AG-032127.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

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