Chronic disuse and skeletal muscle structure in older adults: sex-specific differences and relationships to contractile function

Abstract

In older adults, we examined the effect of chronic muscle disuse on skeletal muscle structure at the tissue, cellular, organellar, and molecular levels and its relationship to muscle function. Volunteers with advanced-stage knee osteoarthritis (OA, n = 16) were recruited to reflect the effects of chronic lower extremity muscle disuse and compared with recreationally active controls (n = 15) without knee OA but similar in age, sex, and health status. In the OA group, quadriceps muscle and single-fiber cross-sectional area were reduced, with the largest reduction in myosin heavy chain IIA fibers. Myosin heavy chain IIAX fibers were more prevalent in the OA group, and their atrophy was sex-specific: men showed a reduction in cross-sectional area, and women showed no differences. Myofibrillar ultrastructure, myonuclear content, and mitochondrial content and morphology generally did not differ between groups, with the exception of sex-specific adaptations in subsarcolemmal (SS) mitochondria, which were driven by lower values in OA women. SS mitochondrial content was also differently related to cellular and molecular functional parameters by sex: greater SS mitochondrial content was associated with improved contractility in women but reduced function in men. Collectively, these results demonstrate sex-specific structural phenotypes at the cellular and organellar levels with chronic disuse in older adults, with novel associations between energetic and contractile systems.

aging and disease are frequently accompanied by reduced functional capacity and the development of physical disability (18). Reductions in physical activity with age are well documented (28, 60, 68) and promote chronic disease (31, 44, 53) and disability. As a result of their temporal coordination, it is difficult to disentangle the relative influence of age, disease, and physical inactivity on skeletal muscle biology and, in turn, the progression toward physical disability. Indeed, many of the muscle phenotypes that accompany aging and chronic disease, such as atrophy and contractile dysfunction, closely mimic those that accompany muscle disuse (1, 46). Whether developed gradually or rapidly following a clinical event, muscle disuse in chronic conditions, such as heart failure or chronic obstructive pulmonary disease, is characterized by its persistent nature, as numerous studies reveal reduced habitual activity levels in these populations (28, 60, 68). Our understanding of the effects of chronic disuse on skeletal muscle size and structure is limited, in part, by the fact that it would be unethical to experimentally impose muscle disuse on older adults for prolonged periods. Nonetheless, the important role played by chronic muscle disuse in mediating the functional sequelae of aging and age-related diseases makes such knowledge clinically relevant.

Atrophy is typically defined as reduced muscle size measured at the tissue level using a variety of noninvasive techniques. However, measurements at the whole muscle level may not adequately reflect adaptations in muscle size and structure at the cellular, subcellular, and molecular levels (15, 25, 39). Because skeletal muscle size and structure at these more basic anatomic levels are well-known regulators of function, characterization of muscle morphology with aging and disease at the cellular and subcellular levels may provide greater insight into functional phenotypes than assessments at the whole tissue level. Although numerous studies have characterized the effects of acute disuse on skeletal muscle structure in young adults (1, 46), only recently have studies been performed in older adults (16, 29, 35). It has been suggested that atrophy may be lessened in older adults (57), raising the intriguing possibility that structural adaptations to disuse may differ with age. Few studies, however, have evaluated the effects of chronic muscle disuse on muscle size, and none has extended measurements to muscle structure below the cellular level or evaluated the possibility that adaptations may differ by sex.

Our goal in this study was to define the effects of chronic disuse on structural characteristics of aged skeletal muscle from the whole tissue to the molecular level. We focused on structural features that are known to influence contractile and metabolic function and may contribute to diminished muscle function by evaluating whether subcellular structural adaptations with disuse explain variation in functional phenotypes described previously (12). We compared morphological differences in skeletal muscle between two groups of older men and women with different patterns of habitual physical activity: one group comprised individuals whose activity levels were limited by joint pain secondary to advanced-stage knee osteoarthritis (OA), and the other group consisted of habitually active older adults (Controls). This approach allowed us to recruit individuals with widely varying physical activity patterns but negligible differences in age, sex distribution, and metabolic, neuroendocrine, or cardiovascular health status. Finally, we evaluated the potential modulating role of sex, with the expectation that sex-specific structural adaptations may explain disuse-related differences in function between men and women (12).

METHODS

Subjects.

Sixteen older adults (8 men and 8 women) with symptomatic knee OA were recruited from the Adult Reconstruction Clinic of the Department of Orthopedics and the surrounding community. A subgroup of this cohort (n = 7), along with an additional three women with knee OA, was also used for analysis of myonuclear number. All participants self-reported receiving a clinical diagnosis of knee OA; seven of these individuals were recruited in close proximity to total knee arthroplasty surgery (bilateral or staged-bilateral in 3 volunteers and unilateral in 4 others). In volunteers entered into the study, symptomatic (6) and radiographic [Kellgren and Lawrence grade 3 or 4 (34)] advanced-stage knee OA was confirmed. Additionally, those in the OA group reported being inactive or participating in light-intensity activities, based on the Stanford Brief Activity Survey (59), which corresponds to activity levels in the 1–1.5 metabolic equivalent (MET) range. To eliminate the possible confounding effects of other chronic diseases or health conditions, volunteers were excluded if they had a history, clinical signs, or symptoms of diabetes, heart failure, pulmonary disease, thyroid disease, peripheral arterial disease, neurological or neuromuscular disease, or autoimmune disease; a current or past (within 10 yr) history of smoking; a current or past (within 10 yr) history of malignancy, excluding nonmelanoma skin cancer; or prior replacement of either knee. All volunteers had normal blood counts/chemistry and renal, liver, and thyroid function, based on standard blood tests. No participants were taking sex steroid replacement therapy (estrogen or estrogen-progestin therapy in women or androgen replacement in men), oral or inhaled corticosteroids, or any other medication that might affect muscle function. Four OA volunteers (21%, 2 women and 2 men) were on stable regimens of 3-hydroxy-3-methylglutaryl CoA reductase inhibitors (statins). Plasma creatine kinase levels were within the normative range in these volunteers, and none had symptoms or signs of statin-induced myopathy. We recently found that chronic, stable statin therapy does not affect skeletal muscle fiber size, mitochondrial morphology, or contractile function in patients without myalgia or elevated creatine kinase levels (unpublished observations), suggesting that inclusion of these individuals would not likely influence detection of effects of muscle disuse. Additionally, nine participants (47%, 4 men and 5 women) had hypertension and were on stable antihypertensive therapy, consisting of angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (56%), diuretics (11%), and adrenergic blocking agents (22%). Nine individuals (5 women and 4 men) were on nonsteroidal anti-inflammatory medications for their OA. None had received an intra-articular injection (hyaluronan or corticosteroid) for 6 mo prior to testing, and none had participated in a rehabilitation program for the 6 mo prior to testing.

Active controls (8 men and 7 women) were selected to match OA participants for age and sex. Controls were healthy and free from disease or medications that could affect muscle size/function and were recruited using inclusion/exclusion criteria identical to those enumerated above for knee OA volunteers, with notable exceptions. Controls did not have symptoms consistent with knee OA (6) or radiographic evidence of significant knee OA (Kellgren and Lawrence grade >2) and self-reported (via Stanford Brief Activity Survey) being recreationally active and participating in moderate- to very heavy-intensity activities (59). This provides a control population with activity levels in the moderate to high range, as reference data on MET level of activities overestimate energy requirements by ∼35% in older adults (54) but exclude sedentary individuals or those who engage in light activities. Additionally, we did not recruit individuals actively training for athletic competition. Five individuals (33%, 3 men and 2 women) were on stable regimens of statins. Plasma creatine kinase levels were within the normative range in these volunteers, and none had symptoms or signs of statin-induced myopathy. Additionally, six controls (40%, 4 men and 2 women) had hypertension and were on stable antihypertensive therapy, consisting of angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (67%), diuretics (33%), and adrenergic blocking agents (17%). Thus, controls were reasonably matched to the OA group for the frequency of various medications. Data on skeletal muscle contractile function at the tissue, cellular, and molecular level and single muscle fiber size from fibers evaluated for mechanical parameters in OA volunteers and controls (n = 31) have recently been published (12). Physical characteristics are reproduced to provide necessary descriptive information, and cellular/molecular functional data are used to examine how structural adaptations relate to variation in functional parameters. Written informed consent was obtained from each volunteer prior to participation in the study. All protocols were approved by the Committees on Human Research at the University of Vermont.

Experimental protocol.

Eligibility was determined during a screening visit, at which time a medical history was obtained, a physical examination was performed, blood samples were taken, and bilateral whole muscle strength was tested [isometric at 55°, as described elsewhere (64)], the latter to determine which leg would be studied, as well as to familiarize volunteers with the strength-testing procedure. These strength data were used to determine the weaker and the stronger leg. For OA participants, the leg with the lower peak isometric torque was studied throughout the rest of testing (i.e., subsequent strength testing, muscle biopsy, and tissue composition); for controls, the leg with the higher peak isometric torque was studied. The rationale for this decision was to further dichotomize the population for skeletal muscle use/function, in addition to self-reported activity level selection criteria described above. After the screening visit, each volunteer was outfitted with an accelerometer (Caltrac, County Technology, Gray Mills, WI), which was to be worn for 10 days to assess free-living, weight-bearing activity levels. Volunteers who met entry criteria and were enrolled underwent a more comprehensive battery of whole muscle strength testing during an outpatient visit. At least 5 days later, muscle tissue was obtained via percutaneous biopsy of the vastus lateralis (VL), and body composition was assessed.

Whole body and tissue morphology.

Body mass was measured on a digital scale (ScaleTronix, Wheaton, IL), and regional body composition was determined by dual-energy X-ray absorptiometry (DEXA; GE Lunar Prodigy, Madison, WI). Thigh fat-free tissue mass, a proxy of skeletal muscle mass (26), was measured using the region-of-interest (ROI) option of the software. For this ROI, the distal cut point was made at the femoral condyles. The length of the femur was then measured from the most inferior aspect of the medial condyle to the most superior aspect of the greater trochanter, and the proximal cut point was made parallel to the femoral condyles at 60% of this length. This proximal cut point was chosen, instead of the more common demarcation used to assess leg fat-free mass via a diagonal line bisecting the femoral neck (26), to eliminate contribution to the fat-free tissue mass estimate from gluteal musculature. Moreover, this cut point permits assessment of a large proportion of the major quadriceps muscles (39, 41). Each thigh was analyzed separately to permit assessment of the study and nonstudy legs. Additionally, quadriceps femoris muscle cross-sectional area (CSA) of both legs was measured using computed tomography (CT), as described elsewhere (11). Briefly, CT image data were analyzed (ImageJ v1.44, National Institutes of Health, Bethesda, MD) by tracing a ROI around the quadriceps muscle group, and all pixels outside this ROI were eliminated. Pixels within the ROI were differentiated between muscle and fat on the basis of radio density, measured in Hounsfield units (0–100 for muscle and −190 to −30 for fat), and the sum of pixels corresponding to each tissue type was used to quantify their CSA. Each leg was analyzed separately to allow comparison of study and nonstudy legs. Midthigh muscle CSA was not assessed in five OA patients.

Muscle biopsy processing.

Biopsy of the vastus lateralis muscle was performed as described elsewhere (62). For single-fiber CSA assessment, tissue was placed immediately into cold (4°C) dissecting solution [20 mM N,N-bis(2-hydroxyethyl)-2-aminoethanesulfonic acid (BES), 5 mM ethylene glycol-bis(2-aminoethylether)-N,N,N,N-tetraacetic acid (EGTA), 5 mM MgATP, 1 mM free Mg²⁺, 1 mM dithiothreitol, and 0.25 mM inorganic phosphate, with an ionic strength of 175 meq, pH 7.0, and pCa 8]. Muscle fiber bundles were carefully dissected, tied to glass rods at slightly stretched lengths, and chemically skinned, as described elsewhere (43), with long-term storage at −20°C in storage solution with 50% (vol/vol) glycerol until measurements, which were performed within 3 wk. Tissue not processed in this manner was frozen in liquid N₂ and stored at −80°C until analysis. Two additional portions of muscle tissue were collected for morphological analysis. One sample was collected and fixed in glutaraldehyde-paraformaldehyde for electron microscopy, as described elsewhere (43). The other sample was collected from a subgroup of the volunteers (n = 16, 6 control and 10 OA) and frozen in embedding medium (OCT, Sakura, Torrance, CA) in isopentane cooled with liquid N₂ for immunohistochemistry.

Single muscle fiber morphology.

Single-fiber CSA was measured on segments of chemically skinned muscle fibers (∼3 mm long) mechanically isolated specifically for morphological analyses, as described previously (43). Briefly, single fibers were isolated from bundles of muscle tissue that were chemically skinned. Fibers chosen for evaluation were structurally intact and without gross morphological abnormalities (e.g., tears or structural inhomogeneity due to damage/stretching during isolation), similar to muscle fibers used for mechanical analyses. There is no bias for fibers of a particular size in general, although we tend not to assess <40-μm-diameter fibers, as these fibers can be susceptible to damage during measurements (e.g., moving between the air-liquid interface). Our selection criteria would potentially bias our estimates of fiber CSA in the current study away from observing atrophy in the OA group, if one assumes that OA participants would have a greater number of small-diameter fibers. After isolation, single fibers were attached to aluminum T clips, mounted between two hooks under a compound light microscope in dissecting solution (see above), and pulled taut. Top and side diameters were measured (the latter using a right-angled, mirrored prism) to the nearest 0.10 μm at 250-μm intervals along the length of the fiber at ×100 magnification using a digital filar eyepiece micrometer (Lasico, Los Angeles, CA) and used to estimate average CSA, with the assumption of an elliptical cross section. At the completion of measurements, each single fiber was placed in gel loading buffer (2% sodium dodecyl sulfate, 62.5 mM Tris, 10% glycerol, 0.001% bromophenol blue, and 5% β-mercaptoethanol, pH 6.8) for analysis of myosin heavy chain (MHC) expression, as described elsewhere (43). Compared with measurement of CSA via immunohistochemistry, this method has the following advantages: 1) it permits assessment of the specific MHC isoform expression of each fiber, with clear delineation of hybrid fibers, and 2) it obtains an estimate of CSA that is the integrated average along the length of the muscle fiber to account for any regional heterogeneity in muscle fiber size (10). Parenthetically, we previously described CSA data for a separate group of fibers isolated for mechanical analyses (12), but fewer fibers were analyzed and assessment of CSA was less complete (e.g., 3–5 vs. ∼10–12 measurements/fiber) than in the current study.

Chemical skinning is known to cause 20% swelling of the muscle fiber in lower vertebrates (20, 21), but its effects on human fibers have, to our knowledge, not been assessed. To address this issue, using procedures similar to those described in our early studies (20, 21), we performed a substudy of a separate group of eight volunteers (2 young and 6 older adults) to directly measure the degree of swelling in human muscle fibers. Briefly, on the day of the biopsy, a small bundle of muscle fibers was dissected from the biopsy and placed directly in mineral oil to preserve the in vivo fiber diameter. While in oil, single fibers were dissected from the bundle, aluminum T clips were fastened to each end, and the fiber was mounted on two elevated hooks immersed in mineral oil in a glass-bottomed chamber. Sarcomere length was set to 2.65 μm, and diameter was measured in triplicate as a proxy of in vivo diameter. Thereafter, the oil was replaced with skinning solution [20 mM BES, 5 mM EGTA, 5 mM MgATP, 1 mM free Mg²⁺, 1 mM dithiothreitol, 0.25 mM inorganic phosphate, and 1% (vol/vol) Triton X-100, with an ionic strength of 175 meq, pH 7.0, and pCa 8]. After 10 min of skinning, at which time the lattice reached a stable diameter, diameter was measured again, in triplicate. The degree of fiber swelling was estimated as the difference between the skinned and in vivo diameters relative to in vivo diameter.

MHC isoform expression and fiber type distribution.

MHC isoform distribution was measured in tissue homogenates by gel electrophoresis, as described previously (43, 61). Gels were loaded per unit protein content (RC DC protein assay, Bio-Rad), with bovine serum albumin (BSA) used as a standard. To account for the fact that an individual muscle fiber can express more than one MHC isoform (i.e., hybrid fibers), we evaluated group differences in the fiber type composition (I, IIA, IIX, I/IIA, IIAX, and I/IIA/IIX) of muscle by calculating the fractional contribution of each fiber type to all the single muscle fibers dissected for these volunteers for mechanical [893 fibers (12)] and morphological (622 fibers) analyses (total n = 1,515, 49 ± 1 fibers/patient). Fiber type data are expressed as a fraction of the total number of fibers evaluated in each individual (%) and total CSA (%CSA). For %CSA, the sum of CSAs for fibers of a given type were summed and expressed as a percentage of the sum of all fiber CSAs for each participant.

Ultrastructural measurements.

Electron microscopy measurements were conducted on intact (i.e., unskinned) skeletal muscle fiber bundles to assess myofibrillar ultrastructure, including myofibrillar fractional area and A-band length, as described elsewhere (43). Additionally, intermyofibrillar (IMF) and subsarcolemmal (SS) mitochondrial fractional area, average area, and number were measured. For IMF mitochondria, procedures were identical to those described previously (11, 63); i.e., an Intuos graphics pad (Wacom, Vancouver, WA) was used to highlight mitochondria. For mitochondria in the SS region, defined by the sarcolemma and the first layer of myofibrils, the procedure was similar to that used in our previous study (11). SS mitochondrial data are expressed as number per length of visible sarcolemma, average size, total mitochondrial area relative to length of visible sarcolemma, and fraction of nonnuclear, SS mitochondrial space. SS mitochondrial data were not obtained for one OA woman because of a considerable amount of damage to the sarcolemma in her sample.

Myonuclear number.

Myonuclear number was quantified by immunohistochemistry of frozen tissue in a subgroup of volunteers (n = 6 control and 10 OA). Briefly, frozen sections (6 μm) were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 10 min. Slides were then incubated in 0.1% Sudan Black in 70% methanol for 15 min to reduce autofluorescence and then in 1% BSA-0.1% Triton X-100 in PBS for 15 min. Sections were blocked in 10% normal goat serum in PBS-1% BSA for 60 min and incubated overnight at 4°C in polyclonal rabbit anti-laminin antibody (1:100 dilution; ab11575, Abcam) and a mouse anti-MHC I antibody (1:250 dilution; A4.951-c, Developmental Studies Hybridoma Bank, University of Iowa) in 1% BSA-0.1% Triton X-100 in PBS. After the overnight incubation, slides were incubated for 60 min in goat anti-rabbit (1:500 dilution; Alexa Fluor 488, Invitrogen) and goat anti-mouse (1:500 dilution; Alexa Fluor 555, Invitrogen) secondary antibodies in 1% BSA-0.1% Triton X-100 in PBS and subsequently stained for myonuclei via application of 4′,6-diamidino-2-phenylindole (1:2,000; Invitrogen) for 10 min. Multiple 5-min wash steps were performed with PBS between most incubations and after 4′,6-diamidino-2-phenylindole staining, and slides were maintained in a light-blocking, humidified chamber throughout staining to prevent photobleaching. At the completion of staining, mounting medium (Citifluor AF1, Electron Microscopy Sciences) was applied, and a coverslip was mounted over the sections. Visualization and imaging were performed using an Olympus BX50 microscope, with digital images acquired at ×20 magnification. CSA measurements for MHC I and II fibers were performed using image analysis software (Metamorph version 7.7.9.0, Molecular Devices, Sunnyvale, CA), and nuclei contained within the laminin border were counted for each fiber. Although the myonuclei count will include muscle satellite cells, these cells are a small fraction (∼2.5%) of the nuclei under the laminin border (49, 57).

Single-fiber contractile function and myosin-actin cross-bridge mechanics/kinetics.

Single-fiber contractile function measurements were characterized by measurement of maximally Ca²⁺-activated (pCa 4.5) isometric tension and shortening velocity and power output from isotonic load clamps performed at 15°C, as described previously for these volunteers (12), and myosin-actin cross-bridge mechanics/kinetics using sinusoidal length perturbation analysis performed at 25°C, as described previously in detail (42). Myosin-actin cross-bridge mechanics/kinetics include the rate of myosin force production, which reflects the rate of transition between weakly and strongly bound states (70), myosin attachment time (t_on), which reflects the average time that myosin is strongly bound to actin (48), and B, which represents the magnitude of the B process, which results from the fitting of complex modulus data and is proportional to the number and stiffness of strongly bound cross bridges. These data have previously been reported in detail (12) and are used here only to interrogate the relationship between structural parameters and cellular/molecular function.

Statistics.

Differences in means between groups were determined using unpaired t-tests and analysis of variance, with the latter including sex as an additional fixed effect to interrogate whether the effects of chronic disuse differ between men and women. In cases where multiple observations were used to characterize a single participant (e.g., single-fiber CSA and mitochondrial content/morphology), a linear mixed model was used to evaluate group and group × sex effects, as described elsewhere (12). This approach included a random effect to account for multiple observations clustered within each volunteer (i.e., numerous muscle fibers evaluated per volunteer). However, the general linear model does not artificially inflate sample size; i.e., comparisons of group and group × sex effects are based on the number of participants, not the number of fibers examined. Associations between measures were determined using Pearson's correlation coefficients for normally distributed data. For data that were not normally distributed (Shapiro-Wilk test), data transformations (e.g., log₁₀) were applied in an attempt to achieve normality. For those variables that did not achieve normality with transformation (e.g., SS mitochondrial data), Spearman's rank correlation coefficients were used. For variables with multiple observations within an individual, data were collapsed to a single mean value for correlation analyses. Values are means ± SE. Statistical analyses were performed using IBM SPSS Statistics (version 20.0, IBM, Armonk, NY) and SAS (version 9.3, SAS Institute, Cary, NC).

RESULTS

Physical characteristics.

Subject characteristics are listed in Table 1 by group assignment and sex. There were no differences between groups in age or body size and no effect of sex to modify any of these parameters. There was, however, a group × sex effect (P < 0.01) for body mass index, reflecting a lower body mass index in control women. By design, physical activity level was lower in the OA than the control group (P < 0.01), and sex did not modify these group differences.

Table 1.

Subject characteristics

	Total		Male		Female
	Control	OA	Control	OA	Control	OA
Age, yr	67.5 ± 1.4	71.2 ± 1.5	68.0 ± 2.1	70.6 ± 2.0	67.0 ± 2.2	72.1 ± 2.6
Height, cm	168 ± 2	166 ± 3	174 ± 2	175 ± 1	162 ± 1	158 ± 3
Body mass, kg	69.8 ± 4.1	70.3 ± 2.6	81.3 ± 4.5	77.2 ± 3.0	56.8 ± 2.2	63.5 ± 2.4
Body mass index, kg/m2	24.5 ± 1.0	25.3 ± 0.6	26.9 ± 1.1	25.1 ± 0.9	21.7 ± 1.0	25.6 ± 0.8*

Values are means ± SE. OA, osteoarthritis.

^*P < 0.01, group × sex effect.

Thigh tissue composition.

Quadriceps muscle CSA and thigh fat-free tissue mass are shown in Fig. 1. There were no differences between study and nonstudy legs in quadriceps muscle CSA or thigh fat-free tissue mass between the control and OA groups, suggesting that muscle atrophy in the OA group is generalized to both limbs. Quadriceps muscle CSA was significantly lower in the OA than control group for the study leg (P < 0.01) and the nonstudy leg (P = 0.03), whereas differences in thigh fat-free tissue mass were not significant for either leg. Sex did not modify group variation in quadriceps muscle CSA or thigh fat-free tissue mass.

Fig. 1.

Thigh tissue composition measured by computed tomography [quadriceps (Quad) cross-sectional area (CSA); A] and dual-energy X-ray absorptiometry [fat-free tissue mass (FFM), B] in study (S) and nonstudy (NS) legs of controls and individuals with knee osteoarthritis (OA). Values are means ± SE. *P < 0.05, **P < 0.01 for group effect.

Single muscle fiber morphology.

Average CSA data from measurements taken along the length of chemically skinned single muscle fiber segments (average fiber length = 3,028 ± 18 μm) and frequency histograms are shown in Fig. 2. Because MHC I/IIA, I/IIA/IIX, and IIX fibers were relatively infrequent (each <3% of all fibers evaluated across groups), we performed statistical analyses on and present data for the three most prevalent isoforms (I, IIA, and IIAX). Cellular level estimates suggest that muscle atrophy is present in the OA group and that atrophy differs by fiber type and by sex. Lower muscle fiber CSA was observed for MHC I (−18%, P = 0.01), IIA (−28%, P = 0.06), and IIAX (−17%, P < 0.05) fibers (Fig. 2B). Whereas sex did not modify group effects in MHC I or IIA fibers, there was a significant group × sex effect (P = 0.02) for MHC IIAX fibers (Fig. 2C). Pair-wise comparisons showed that average CSA for MHC IIAX fibers did not differ between control and OA women but was significantly reduced in OA men vs. controls (P < 0.01). MHC IIAX CSA was greater for control men than both groups of women (P < 0.01 for both). CSA data for MHC IIA and IIAX fibers excluded results from one male control volunteer, for whom 10 of 14 MHC IIA and IIAX fibers were statistical outliers (>3 SD from the mean) and data for the remaining 4 MHC II-type fibers were >2 SD from the mean.

Fig. 2.

A: top-view (d₁) and side-view (d₂) diameters at multiple points (250-μm intervals) along the length of chemically skinned muscle fiber segments. B: single-fiber CSA of myosin heavy chain (MHC) I, IIA, and IIAX fiber types in control and OA groups. C: group × sex effect (P = 0.02) for MHC IIAX fibers. Number of fibers is shown at the base of each bar for group average data. Values are means ± SE. *P < 0.05. **P ≤ 0.01. †P = 0.06. D–F: frequency histograms (x-axis in bins of 1,000 μm²) for the 3 most prevalent fiber types.

Muscle fiber segments undergo swelling (∼20%) upon chemical skinning (21). Because the degree of swelling has not been reported in human fibers, we measured changes in fiber diameter upon chemical skinning (n = 8 volunteers and 45 fibers). We observed an ∼20% increase in muscle fiber diameter following chemical skinning (121.6 ± 1.3%), which agrees with previous work from our laboratories in amphibians and lower-order mammals (20, 21). We would not expect differences in swelling between the OA and control group, as the main osmolytes in skinned fibers are myofilament proteins, and our prior work (12) shows that myosin and actin, the two most abundant myofilament proteins, do not differ between groups. More importantly, we found that smaller fibers experienced greater swelling upon skinning (r = −0.578, P < 0.01), as others have observed in the setting of genetic models of fiber hypertrophy (7). This suggests that the extent of atrophy would, if anything, be underestimated in the smaller fibers from OA participants and strengthens our confidence that size differences in skinned single fibers indeed reflect in vivo differences in muscle fiber size.

MHC isoform expression in tissue homogenates and single-fiber MHC fiber type distribution.

We also determined the admixture of fiber types in tissue samples, as this could impact muscle function (23). Using muscle tissue homogenates, we found no group differences in the relative expression of MHC I, IIA, or IIX via gel electrophoresis and no group × sex effects (Fig. 3A). In contrast, there were group differences when we examined the relative expression of different fiber types as a fraction of fiber CSA (%CSA; Fig. 3B) or as a fraction (%) of total number of fibers. Differences in fiber type distribution (%CSA) between the control and OA group, respectively, are as follows, in ascending order of abundance: MHC I (60 ± 6% and 51 ± 6%, n = 15 control and 16 OA) > IIA (27 ± 5% and 21 ± 4%, n = 15 control and 15 OA) > IIAX (11 ± 3% and 24 ± 4%, n = 13 control and 15 OA, P < 0.03) > I/IIA (4.2 ± 0.7% and 3.8 ± 0.3%, n = 11 control and 8 OA) > I/IIA/IIX (2.6 ± 0.2% and 5.5 ± 2.7%, n = 2 control and 6 OA) > IIX (0% and 12 ± 4%, n = 0 control and 4 OA). Data for the three most prevalent isoforms are shown in Fig. 3B. Results were similar when data were expressed as a fraction (%) of the total number of fibers evaluated in each individual [MHC I (58 ± 5% and 50 ± 6%, n = 15 control and 16 OA) > IIA (28 ± 5% and 22 ± 4%, n = 15 control and 15 OA) > IIAX (11 ± 3% and 24 ± 4%, P < 0.02, n = 13 control and 15 OA) > I/IIA (5.0 ± 0.7% and 4.3 ± 0.5%, n = 11 control and 8/OA) > I/IIA/IIX (3.7 ± 0.03% and 6.6 ± 3.2%, n = 2 control and 6 OA) > IIX (0% and 10 ± 2.4%, n = 0 control and 4 OA)], with differences between groups being significant only in the distribution of MHC IIAX (P < 0.02). Thus the OA group expresses a greater proportion of MHC IIAX fibers than controls. Variation in %CSA and fraction (%) of the total number of fibers evaluated in each individual does not summate to 100% across isoforms, because not all fiber types were expressed in all OA participants and controls. Thus the fractional contribution of various isoforms can vary across volunteers for each fiber type classification. The small n for MHC IIX and I/II hybrid isoforms (I/IIA/IIX and I/IIA) and the fact that they cumulatively represent only 5.7% (87 of 1,515) of all fibers assessed for the MHC isoform explain their exclusion from Fig. 3.

Fig. 3.

A and B: myosin isoform (I, IIA, and IIX) distribution in tissue homogenates and isolated single muscle fiber segments (MHC I, IIA, and IIAX) in control and OA groups. Data for single fibers are expressed as a fraction of total fiber CSA, with CSA for all fibers within each fiber type for each individual being summed and expressed relative to the summed total CSA from all fibers within that individual. C: sex × fiber type effect for MHC I and IIAX fibers in men and women, with bars representing both control and OA men and control and OA women. Number of fibers is shown at the base of each bar. Values are means ± SE. *P < 0.05. **P < 0.01.

When sex was considered in the model, we found sex effects for MHC IIAX and I fibers, with greater expression of MHC IIAX fibers (P < 0.01 for sex effect for both %CSA and fraction of the total number of fibers evaluated in each individual) and reduced expression of MHC I fibers (P < 0.05 and P = 0.06 for %CSA and fraction of the total number of fibers evaluated in each individual, respectively) in men, with no sex effects in MHC IIA and no group × sex effects for any fiber type. Thus there was also greater relative expression of MHC IIAX and less expression of MHC I in men than women in general (Fig. 3C), in keeping with prior reports (24).

Myofilament ultrastructure.

We evaluated whether muscle atrophy in OA volunteers extends to adaptations in myofilament ultrastructure. However, no differences in myofilament fractional area were observed between the OA and control groups, and there was no effect of sex to modify variation between groups. There was a trend (P = 0.08) for a group × sex effect, with myofibrillar fractional area being higher in OA women than control women but no differences between groups in men. Finally, we found no group differences in A-band length and no modifying effect of sex.

Mitochondrial content and morphology.

To examine whether reduced physical activity and atrophy associated with knee OA was accompanied by changes in mitochondrial content or morphology, we evaluated IMF (Fig. 4) and SS (Fig. 5) populations of mitochondria. No differences in IMF mitochondrial fractional area, average size, or number per unit area were found between groups, nor were there any group × sex effects. However, when examining IMF mitochondrial parameters in correlation analyses, we found a relationship between IMF mitochondrial fractional area and physical activity level (r = 0.409, P = 0.02; Fig. 4D).

Fig. 4.

Intermyofibrillar (IMF) mitochondrial (mito) fractional area (A), average size (B), and number (C) in control and OA groups and the relationship between IMF mitochondrial fractional area and log₁₀ physical activity level (D). Group average data are means ± SE.

Fig. 5.

Subsarcolemmal (SS) mitochondrial number per sarcolemmal length (A), content per unit sarolemmal length (B), average (ave) size (C), and fractional area per unit SS space (D) in control and OA groups, with groups split by sex. Lines across all data denote group × sex effects; lines over each sex denote differences between control and OA groups. Measures of SS mitochondria for 1 OA woman are not included. Values are means ± SE. *P ≤ 0.05. **P < 0.01.

Similarly, for SS mitochondria, we found no differences in any index between the OA and control groups. However, we did observe group × sex effects for these parameters. SS mitochondrial number (P < 0.01; Fig. 5A) and content per length of sarcolemma (P < 0.01; Fig. 5B) and per fraction of SS space (P = 0.05; Fig. 5D) showed group × sex effects, while there was a trend (P = 0.08) toward a group × sex effect for average mitochondrial size (Fig. 5C). Pair-wise comparisons showed that the lower mitochondrial area per length of sarcolemma was due primarily to fewer mitochondria in OA women (both P < 0.01). However, pair-wise comparisons showed sex differences between control men and women, with higher SS mitochondrial content per sarcolemmal length (P < 0.01), but not per area of SS space, in control women than men, due to a greater number of mitochondria (P < 0.05). As an aside, because SS mitochondria variables were nonnormally distributed, we also analyzed ranked data via the mixed model and found similar group × sex effects [P < 0.05 to P < 0.01, with a trend (P = 0.06) for average size]. Thus the modifying effect of sex on OA is set on the background of sex differences in this mitochondrial population in active, older adults.

Myonuclear number.

To examine whether reduced physical activity and atrophy in OA were accompanied by loss of myonuclei, we measured myonuclear number in MHC I and II fibers in tissue sections via immunohistochemistry in a subset of volunteers (Fig. 6). However, the number of myonuclei per fiber did not differ between groups for MHC I or II fibers. Moreover, no sex × group effects were noted.

Fig. 6.

Myonuclear number per fiber for MHC I (B) and II (D) fibers in control and OA groups. A and C: representative images from control and OA volunteers, with laminin stained green, MHC I stained red, and nuclei stained blue [DAPI (4′,6-diamidino-2-phenylindole)]. Scale bar = 50 μm.

Relationship between mitochondrial content and cellular/molecular functional parameters.

Because our prior studies examining healthy, sedentary younger and older adults and older adults with disease (12, 61) showed relationships between mitochondrial content and molecular function, we explored associations of SS mitochondrial content and morphology to cellular and molecular contractile parameters in MHC I and IIA fibers (Table 2). We focused our analyses on SS mitochondria, as this population showed sex-specific differences (Fig. 5), similar to cellular/molecular contractile parameters reported previously in these volunteers (12). To provide a brief overview of sex-specific differences, we showed that disuse had a greater effect to reduce single muscle fiber function in women than men. In MHC I fibers, t_on was prolonged in OA women (P = 0.01) compared with all other groups, indicating slowed cross-bridge kinetics. In MHC IIA fibers, there were group × sex effects for muscle power output (P < 0.05) and velocity (V_max, P = 0.07), with OA men showing increases in both. At the molecular level, these group × sex effects were explained by faster cross-bridge kinetics (reduced t_on and increased myosin rate of force production, P = 0.06 for both) in OA men.

Table 2.

Relationships between SS mitochondrial abundance and cellular and molecular indexes of contractile function

	MHC I	MHC IIA
	ton	Vmax	Myosin RFP	Po	B
Men
SS area		−0.654* (13)		−0.478‡ (13)	−0.700† (15)
SS number		−0.670* (13)
Women
SS area	−0.538* (14)	0.624* (10)	0.508‡ (14)
SS number	−0.670† (14)	0.745† (10)	0.565* (14)

Values reflect correlation coefficients; only significant relationships are shown. Number of subjects for each correlation are indicated in parentheses. Sample sizes for myosin heavy chain (MHC) IIA fibers varied, because MHC IIA fibers were not evident in all subjects analyzed via isotonic load clamps [shortening velocity (V_max) and tension at 15°C (P_o)] and/or sinusoidal analysis [myosin rate of force production (RFP) and index reflecting number and stiffness of strongly attached myosin-actin cross bridges (B)] and because measures of subsarcolemmal (SS) mitochondria for 1 woman with OA are not included. t_on, Myosin attachment time.

^*P ≤ 0.05;

^†P ≤ 0.01;

^‡P ≤ 0.10.

No relationships were found between mechanic/kinetics variables and SS mitochondrial morphology when control and OA participants were pooled (data not shown). However, when examined separately in each sex, relationships were found. In men, MHC IIA shortening velocity was negatively related to SS mitochondrial area and number (r = −0.654, P = 0.02 and r = −0.670, P = 0.01, respectively). Moreover, in MHC IIA fibers, there was a negative relationship between SS mitochondrial area and the number/stiffness of strongly bound cross bridges (i.e., B; r = −0.700, P < 0.01), which could promote a reduction in tension, as suggested by a trend toward a negative relationship between SS mitochondrial area and maximal Ca²⁺-activated tension (r = −0.478, P = 0.10). In contrast, in women, MHC IIA unloaded shortening velocity was positively associated with SS mitochondrial area and number (r = 0.624, P = 0.05 and r = 0.754, P = 0.01), which was explained at the molecular level by the fact that greater SS mitochondrial area and number were associated with increased myosin-actin cross-bridge kinetics, as evidenced by increasing myosin rate of force production (r = 0.508, P = 0.06 and r = 0.565, P = 0.04, respectively). In MHC I fibers in women, there was a similar relationship with increasing cross-bridge kinetics, with greater SS mitochondrial abundance (area and number) correlating negatively with t_on (r = −0.538, P < 0.05 and r = −0.670, P < 0.01, respectively). Note that the directionality of the correlation is due to the fact that shorter t_on indicates faster cross-bridge kinetics.

These reciprocal relationships between SS mitochondria and cellular/molecular functional parameters in men and women prompted us to evaluate whether IMF mitochondria showed similar sex-specific relationships to contractile parameters. In the pooled cohort, positive associations were found between IMF mitochondrial fractional area and tension in MHC I (r = 0.36, P = 0.04) and IIA (r = 0.43, P = 0.02) fibers. This relationship was explained at the molecular level by a positive association between IMF mitochondria and an increased number/stiffness of strongly bound cross bridges (i.e., B) in MHC I (r = 0.40, P = 0.03) and IIA (r = 0.33, P = 0.07) fibers (i.e., more strongly bound cross bridges yield greater tension). Interestingly, there were sex-specific differences in these associations. In men, IMF mitochondrial fractional area was correlated with MHC I and IIA tension (r = 0.629, P < 0.01 and r = 0.545, P < 0.05, respectively). This was explained at the molecular level in MHC I fibers by positive correlations between IMF mitochondrial fractional area and the number/stiffness of strongly bound cross bridges (i.e., B; r = 0.536, P < 0.05) and increased t_on (r = 0.731, P < 0.01). In MHC IIA fibers, greater IMF average mitochondrial area was associated with increased number/stiffness of strongly bound cross bridges (r = 0.628, P = 0.01). In contrast, no correlations were found between IMF mitochondrial parameters and cellular/molecular functional measures in women.

DISCUSSION

Our study revealed atrophy of skeletal muscle with chronic disuse in older adults at the tissue level, which was explained at the cellular level by reductions in muscle fiber size. The OA group was found to have a greater relative expression of MHC IIAX fibers, and atrophy in these fibers was apparent only in men. In contrast to differences at the cellular level, myofibrillar ultrastructure, myonuclei content, and mitochondrial content and morphology generally did not differ between groups, with the exception of sex-specific adaptations in SS mitochondria, which were driven by differences in women. Interestingly, we found that the greater abundance of SS mitochondria was positively associated with cellular/molecular contractile function in women but was negatively associated in men. These relationships are notable, as they parallel sex-specific and, in some cases, reciprocal adaptations in cellular and molecular function in response to disuse in this cohort (12), suggesting that different cellular and organellar skeletal muscle structural adaptations in men and women may alter their trajectories toward disability (32).

Quadriceps CSA was reduced in older adults with OA (Fig. 1A), as might be expected with long-standing muscle disuse. As an aside, discordant findings between DEXA and CT imaging may relate to limitations of DEXA to discriminate these differences in older adults (15, 39). Whole muscle CSA deficits (−19%) were paralleled by reductions in fiber CSA, which were evident in the three most prevalent fiber types examined (I, IIA, and IIAX). In fact, the average relative reduction in CSA in these fiber types agreed well with reductions at the whole muscle level (−21%). Reductions in CSA were greatest in MHC IIA fibers (Fig. 2B). This finding differs from other animal (5) and human (46) models of disuse, which show that disuse-related atrophy is greater in slow- than fast-twitch muscle fibers. However, it is important to point out that the type of disuse in our cohort is distinguished from these acute models in that it is more chronic (i.e., years) and less severe (i.e., OA participants are ambulatory but simply performed less weight-bearing activity) (Table 1). Differences between results might also be explained by the fact that OA participants and controls in the current study are older, and it is generally believed that MHC II fibers are more sensitive to deterioration with age (17). In support of our results, other reports suggest greater atrophy of MHC II fibers with chronic disuse associated with OA (45, 50, 56, 58). However, the histochemical technique (i.e., myosin ATPase staining) used in these prior studies does not permit delineation of pure or hybrid fiber subtypes. In this context, our study provides the first rigorous determination of the fiber type specificity of atrophy in chronic muscle disuse in older adults.

Our study is also the first to consider the modifying effects of sex on fiber size in response to muscle disuse (45, 50, 56, 58). In MHC IIAX fibers, we observed pronounced atrophy in OA men (Fig. 2), and no differences between OA women and controls. Moreover, both groups of women were similar to OA men, suggesting that MHC IIAX fiber CSA is differently regulated by sex independent of muscle use. In a previous study on age-related differences in muscle morphology, in a separate cohort of older men and women, we described qualitative differences in muscle atrophy in MHC II-expressing fibers (Fig. 2, C and D) (11), but we did not differentiate between MHC II subtypes. Interestingly, studies that have considered the effects of sex on acute disuse in younger volunteers have not found fiber type-specific atrophy in general or in MHC IIAX fibers specifically (65, 66, 69), suggesting that sex differences in MHC IIAX fiber atrophy in response to disuse may be unique to older adults.

Selective reduction in MHC IIAX fiber CSA with disuse in men (Fig. 2C) could further reduce whole muscle function, owing to greater power output in these fibers (8) and their contribution to whole muscle dynamic function (23, 30). However, MHC IIAX fiber CSA was disproportionately greater in control men than all other groups, and these fibers were more prevalent in men in general. In other words, the sex-related differences in MHC IIAX fiber expression may counterbalance atrophy to maintain whole muscle function. Along these same lines, the isoform shift toward a faster phenotype in the OA group in general (Fig. 3) is in keeping with adaptations in other disuse models (5, 46) and may serve to offset reductions in whole muscle power output due to disuse-related reductions in force production secondary to atrophy. While speculative, the notion that alterations in isoform expression, either in a sex- or disuse-specific manner, counteract the functional effects of cellular atrophy highlights the potentially complex interplay between structural and functional adaptations to muscle disuse in older adults, ranging from muscle fiber type admixture to cellular/molecular function. Considering the importance of muscle power output to whole muscle/body function (51), sex-specific variation in adaptations in fiber size and/or fiber type composition in response to disuse may influence the trajectory toward disability differently in men and women (32).

Preclinical models provide evidence that mitochondrial structural adaptations may contribute to disuse atrophy (52). Moreover, we previously described relationships between mitochondrial content/morphology and cellular/molecular contractile function (12, 61). Thus we evaluated whether disuse-related muscle atrophy was associated with variation in mitochondrial content and/or morphology by evaluating IMF and SS mitochondrial populations. Although no differences in IMF or SS mitochondrial content or morphology between groups were found, we observed a positive correlation between IMF abundance and physical activity (Fig. 4), in keeping with the well-documented relationship of muscle use and mitochondrial content/function (33, 37). Additionally, we noted effects of disuse on SS morphology when sex was included in the statistical model (Fig. 5), driven by higher SS mitochondrial content in control women compared to all other groups. Previous work from our laboratory and others in sedentary to moderately active older men and women found no sex differences in SS mitochondrial content (11, 13), suggesting that variation in activity level in older adults in the present study leads to differential adaptations in SS mitochondria in men and women. We are not aware of any studies examining sex-specific adaptations in SS mitochondria in response to altered muscle use in older adults or aged animals to compare with our current findings. Older men and women increase IMF mitochondrial content in response to endurance training (33), with greater improvements in oxidative enzyme function in the SS fraction (40), but sex differences in these responses were not analyzed. In older men, cast immobilization for 2 wk was associated with a greater reduction in SS than IMF mitochondria, similar to young men, but women were not studied (47). Thus our current results suggest novel effects of sex on the regulation of SS mitochondria in response to variations in muscle use in older adults.

As we previously demonstrated sex-specific cellular and molecular functional adaptations to chronic disuse in this population (12), with women showing a less functional phenotype, we further examined whether sex-specific variation in SS mitochondria might predict variation in function. Surprisingly, we found reciprocal relationships in men and women (Table 2). In men, a greater abundance of SS mitochondria was associated with deleterious contractile phenotypes; in women, a greater content or number of SS mitochondria was associated with more beneficial functional phenotypes. To our knowledge, this is the first demonstration of associations between SS mitochondrial content/morphology and contractile phenotypes, much less sex-specific reciprocal relationships. When these relationships are considered in the context of the reduction in SS mitochondria in OA women, it is tempting to hypothesize that the sex-specific response to disuse in SS mitochondrial content may contribute to their deficits in functional adaptation to disuse compared with men (12). Although we are cautious not to infer cause and effect from these correlations, their novelty deserves some discussion.

We initially expected more robust relationships with IMF mitochondria, owing to their greater abundance and our prior demonstration of relationships between IMF mitochondria and functional parameters (12, 61). Although connected to the IMF pool (14), SS mitochondria comprise only 20% of myocellular mitochondria and are located distal to most myofilaments. However, compared with the IMF pool, SS mitochondria are more responsive to the effects of muscle use/disuse (2, 3, 27, 36, 47). Moreover, because of their close proximity to myonuclei (14, 19), alterations in SS content and/or function that affect energy supply (40), reactive oxygen species production (55), and/or apoptotic stimuli (2, 3) may alter nuclear gene transcription in ways that modulate myofilament and/or mitochondrial function throughout the cell. This could occur through an effect of SS mitochondria to modulate myonuclear number, as animal models show that SS mitochondria produce more apoptotic stimuli in response to disuse (2), and it has been suggested that loss of myonuclei contributes to muscle atrophy (4). However, we did not observe differences in myonuclei content, in keeping with recent studies showing no loss of myonuclei with disuse (9, 67) or modulation of myonuclear number with variation in activity in older men (38). A more plausible mechanism may be via modulation of nuclear function. Indeed, recent studies have uncovered a novel mechanism whereby myonuclei coordinate mitochondrial gene transcription (22). Of course, it is equally plausible that, owing to their proximity to myonuclei, SS mitochondria are simply more robust biomarkers for the effects of disuse on the cell. Regardless of the physiological pathways underlying these associations, our results raise the intriguing possibility that sex-specific adaptations in mitochondrial biology may influence contractile function.

In summary, our findings have identified, for the first time, sex-specific alterations in skeletal muscle fiber size and fiber type distribution with disuse in older adults. Additionally, we report sex-specific adaptations in SS mitochondria and reciprocal relationships between mitochondrial abundance and contractile function in men and women that are compartment-specific. Unfortunately, we were unable to define the specific mechanisms underlying sex differences, and the cross-sectional nature of the study is a clear limitation. However, these studies highlight a novel link between energetic and contractile systems in the context of muscle use patterns in older adults and raise the intriguing possibility that disability may manifest in men and women differently in response to disuse via unique effects on, and interrelationships between, metabolic and contractile phenotypes.

GRANTS

This study was funded by National Institute on Aging Grant AG-033547 (to M. J. Toth), National Heart, Lung, and Blood Institute Institutional National Research Service Award HL-007647 (to D. M. Callahan), and National Institute on Aging Mentored Research Scientist Development Award AG-031303 (to M. S. Miller).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

D.M.C., T.W.T., M.S.M., S.B.H., H.S., and M.J.T. performed the experiments; D.M.C., T.W.T., M.S.M., S.B.H., H.S., N.C.C., J.R.S., P.D.S., B.D.B., and M.J.T. analyzed the data; D.M.C. and M.J.T. interpreted the results of the experiments; D.M.C., T.W.T., and M.J.T. prepared the figures; D.M.C., T.W.T., and M.J.T. drafted the manuscript; D.M.C., T.W.T., M.S.M., S.B.H., H.S., N.C.C., J.R.S., P.D.S., P.A.A., D.W.M., B.D.B., and M.J.T. edited and revised the manuscript; D.M.C., T.W.T., M.S.M., S.B.H., H.S., N.C.C., J.R.S., P.D.S., P.A.A., D.W.M., B.D.B., and M.J.T. approved the final version of the manuscript; P.A.A., B.D.B., and M.J.T. developed the concept and designed the research.