Endocrinology and Metabolism

Voluntary running, skeletal muscle gene expression, and signaling inversely regulated by orchidectomy and testosterone replacement

Chikwendu Ibebunjo, John K. Eash, Christine Li, QiCheng Ma, David J. Glass


Declines in skeletal muscle size and strength, often seen with chronic wasting diseases, prolonged or high-dose glucocorticoid therapy, and the natural aging process in mammals, are usually associated with reduced physical activity and testosterone levels. However, it is not clear whether the decline in testosterone and activity are causally related. Using a mouse model, we found that removal of endogenous testosterone by orchidectomy results in an almost complete cessation in voluntary wheel running but only a small decline in muscle mass. Testosterone replacement restored running behavior and muscle mass to normal levels. Orchidectomy also suppressed the IGF-I/Akt pathway, activated the atrophy-inducing E3 ligases MuRF1 and MAFBx, and suppressed several energy metabolism pathways, and all of these effects were reversed by testosterone replacement. The study also delineated a distinct, previously unidentified set of genes that is inversely regulated by orchidectomy and testosterone treatment. These data demonstrate the necessity of testosterone for both speed and endurance of voluntary wheel running in mice and suggest a potential mechanism for declined activity in humans where androgens are deficient.

  • andropause
  • sarcopenia
  • aging
  • pathways

although testosterone and other anabolic steroids have been in existence for more than half a century (reviewed in Refs. 11 and 18), only comparatively recently have they been approved for medical use as replacement therapy in conditions with deficiency or absence of endogenous testosterone and as adjunctive therapy to promote weight gain in conditions with involuntary weight loss (18). In addition, anabolic steroids are commonly abused by adolescents and young adults to boost physical appearance (muscle mass) and performance (muscle strength and power) and by participants of endurance sports such as cycling and marathons to enhance endurance and hasten recovery between intense physical efforts (16). Yet evidence to support these nonmedical uses of anabolic steroids is sparse, inconclusive, and often anecdotal, and the underlying mechanisms for any such effects not fully elucidated.

Testosterone is thought to produce its ergogenic (defined as “performance enhancing”) effects through both a canonical and a noncanonical pathway. In the canonical pathway, the anabolic steroid binds to the androgen receptor (AR) to initiate a conformational change and subsequent association with the AR transcriptional coactivator β-catenin. The complex of agonist/AR/β-catenin translocates into the nucleus and binds to AR-responsive elements located in the promoter of specific genes, including genes in the IGF-I (10), Wnt (34), and myostatin (Mstn) (25) pathways, to increase transcription (33). Noncanonically, androgen agonists have been shown to directly activate Akt (21), leading to multiple downstream effects.

Activation of Akt leads to phosphorylation/activation of downstream molecules, including mTOR, p70S6, and S6, all resulting in an increase in protein synthesis (13). Furthermore, activation of Akt leads to phosphorylation and thus inhibition of the family of FoxO (also called “forkhead”) transcription factors, which are required for upregulation of the muscle-specific E3 ubiquitin ligases muscle-specific RING finger 1 (MuRF1) and muscle atrophy F-box (MAFbx; also called atrogin-1) (31, 35). These E3 ubiquitin ligases induce the proteasome-mediated degradation of particular protein substrates, with MuRF1 activation resulting in the selective breakdown of myosin heavy chain (5) and the rest of the thick filament of the sarcomere (6). Indeed, MuRF1 and MAFBx have been shown to be induced in at least 13 distinct models of skeletal muscle atrophy in both rodents and humans (13). Other mechanisms have also been implicated for the effects of anabolic steroids, including 1) acting directly on mesenchymal stem cells and satellite cells to promote commitment into the myogenic lineage and on preadipocytes to inhibit their differentiation into the adipogenic lineage (17) and/or 2) interacting with the glucocorticoid receptor to inhibit catabolic signaling of glucocorticoids (9, 44). The net effect is an increase in protein synthesis and a decrease in protein degradation, leading to enhanced muscle fiber cross-sectional area and mass (4).

In contrast to the muscle mass and strength effects, surprisingly little is known of the underlying mechanisms of anabolic steroids on physical endurance and recovery between intense bouts of exercise. In young male rats that have had free access to a running wheel for 4 wk, treatment with 0.5 mg of nandrolone phenylpropionate every other day for an additional 4 wk increased submaximal running endurance on a treadmill by 41% but did not alter total distance run, maximum running speed, or muscle oxidative enzyme activity compared with saline-treated controls (38). In another study in male rats, the combination of nandrolone decanoate and voluntary wheel running increased the relative expression of myosin heavy chain (MyHC) type 1 and decreased MyHC type 2a proteins in the soleus muscle, whereas either treatment alone or androgen depletion by castration did not alter MyHC composition of the soleus or extensor digitorum longus muscles (28). Although running endurance was not assessed in this study, these authors (28) speculated that the increase in MyHC type 1 in the soleus muscle would explain the increase in submaximal running endurance reported by Van Zyl et al. (38). In humans, a double-blinded study found no differences in endurance performance, assessed by a standardized treadmill test or by the profile of several serum biomarkers of recovery, between healthy young men treated with 12 doses of testosterone undecanoate, 19-norandrostenedione, or the mannitol placebo over a 1-mo period (2). Thus, it is not clear whether anabolic steroids are beneficial for improving physical endurance or whether any changes are explained by changes in muscle fiber metabolism.

Therefore, the aims of this study were to establish 1) whether testosterone is critical for physical performance and endurance as assessed by voluntary wheel running capacity in a mouse model of testosterone depletion and replacement, 2) the morphological (muscle mass and fiber cross-sectional area) changes, and 3) the transcriptional and molecular signaling mechanisms associated with testosterone depletion and replacement in adult skeletal muscle.



Adult male C57BL/6NTac mice (Taconic), ∼4 mo old, were used for these studies. The mice were acclimated to the facility for ∼7 days and then randomized by body weight into 16 groups of six mice each. They were housed individually with environmental enrichment in a temperature- (72°F) and humidity-controlled (43%) room and maintained on a 12:12-h light-dark cycle (lights on from 6 AM to 6 PM). All mice were fed a standard rodent diet (no. 5053, PicoLab Rodent Diet 20; LabDiets) containing 62% carbohydrate, 23% protein, and 13% fat. Food and water were provided ad libitum. All animal procedures were approved by the Institutional Animal Care and Use Committee of the Novartis Institutes for Biomedical Research and were in compliance with the Animal Welfare Act Regulations 9 CFR Parts 1, 2, and 3 and US regulations (Guidelines for the Care and Use of Laboratory Animals, 1995).

Experimental design.

The experimental groups are summarized in Table 1. To deplete baseline endogenous testosterone levels, mice in groups 2 and 4–16 were orchidectomized (ORX) 7 days before start of treatment or voluntary wheel running. Mice in groups 1 and 3 served as controls for the effects of ORX at the start (day 0) and end (day 28) of the treatments (testosterone and/or voluntary wheel running), respectively.

View this table:
Table 1.

Description of treatment groups

To investigate the effects of testosterone on muscle mass and signaling in sedentary (S) mice, mice in groups 5–10 that had previously been ORX for 7 days were injected subcutaneously with 10 mg·kg−1·day−1 of testosterone (T) (no. T1500; Sigma Aldrich) in corn oil (no. 901414; MP Biomedical) or the same volume of corn oil vehicle (V) for 3, 7, or 28 days. Body weight and food consumption were measured at weekly intervals beginning 14 days before start of dosing (the later-designated day 0) until and including the last day of study. Daily food consumption was calculated as the difference between the amount of food provided and the amount of food left divided by the number of days in between.

To investigate the effects of testosterone on voluntary wheel running (R), mice in groups 11–16 that had previously been ORX for 7 days were introduced to a voluntary running wheel model (no. 80820; Lafayette Instruments, Lafayette, IN) and concurrently dosed with either T in corn oil at 10 mg·kg−1·day−1 or an equal volume of the corn oil V by subcutaneous injection for 3, 7, or 28 days. Body weight and food consumption were also measured as described above for the S mice. Access to the running wheel was interrupted daily for ∼1 h (∼1–2 PM) to allow for dosing and weighing of the mice and food and/or cleaning of the cage. Wheel revolutions were recorded daily, and the distance run per day as well as per hour was calculated from the wheel circumference and revolutions.

Tissue weights.

At the end of the treatment period, euthanasia was performed by carbon dioxide asphyxiation, and blood was collected by cardiac puncture into serum separator tubes (Sarstedt). The following tissues were rapidly dissected out and weighed: skeletal muscles from both right and left fore- (biceps and triceps brachii) and hindlimbs (tibialis anterior, gastrocnemius-soleus-plantaris complex, and quadriceps), bulbocavernosus and levator ani muscle complex (BLA), heart, liver, spleen, and white (perirenal) and brown adipose pads. The left triceps brachii muscle was snap-frozen in 2-methylbutane precooled in liquid nitrogen for histological analysis, whereas other tissues were snap-frozen directly in liquid nitrogen. Because it was the most responsive to T depletion and replacement of all the appendicular skeletal muscles studied, the triceps brachii muscle was subsequently analyzed for fiber cross-sectional area and gene and protein expression.

Blood analysis.

Serum samples on days 3 and 7 from sedentary vehicle-treated (SD3V and SD7V), running vehicle-treated (RD3V and RD7V), and running testosterone-treated (RD3T and RD7T) mice were analyzed for creatine kinase levels using the Olympus CK-Nac reagent on an Olympus Analyzer model AU400E (Olympus, Center Valley, NJ).

Muscle fiber histomorphometry.

Serial sections, 8 μm thick, were cut from the frozen left triceps brachii muscles of mice in all day 28 groups (i.e., D28-INT, D28-ORX, SD28V, SD28T, RD28V, and RD28T) and 1) stained with hematoxylin and eosin (H & E) to visualize the structure of the myofibers, 2) stained for succinate dehydrogenase (SDH) activity for fiber oxidative capacity, or 3) coimmunostained for laminin and either MyHC type 1 or MyHC type 2A to determine fiber type composition and fiber cross-sectional area. For the dual laminin and MyHC staining, the sections were blocked in 10% goat serum for 30 min at room temperature and incubated with a mixture of an anti-laminin IgG (diluted 1:400) antibody produced in rabbit (Sigma-Aldrich) and an anti-MyHC type 1 (ATCC no. HB283) or type 2A (ATCC no. HB277) antibody, washed, incubated with a polymer rabbit-on-rodent horseradish peroxidase secondary antibody, washed, and incubated with the 3,3′-diaminobenzidine tetrahydrochloride substrate (Vector Laboratories) per the manufacturer's protocol. Images of the entire tissue section were acquired using Scanscope (Aperio, Vista, CA), and 1) H & E-stained sections were visualized for fiber morphology, 2) the proportion of the section was occupied by fibers with weak, medium, or strong SDH reactivity, and 3) the mean and frequency distribution of the cross-sectional area of all fibers in the section were determined automatically using a custom software, Astoria version 3.0, developed at Novartis Institutes for Biomedical Research. More than 3,500 fibers in each section were measured automatically by this method.

RNA analyses.

The triceps brachii muscles from sedentary or voluntary wheel-running mice treated with vehicle or testosterone for 3, 7, or 28 days were analyzed for gene expression changes. The right triceps brachii muscle was pulverized under liquid nitrogen, and one-half was used for total RNA extraction using the RNeasy Fibrous Tissue Midi Kit per the manufacturer's protocol (Qiagen, Valencia, CA). RNA concentration and purity (A260/A280 ratios >1.9) were assessed using a NanoDrop-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE), and RNA integrity was assured using an Agilent (Santa Clara, CA) 2100 Bioanalyzer.

Quantitative real-time PCR.

The RNA samples were reverse-transcribed to cDNA using a high-capacity cDNA kit (Applied Biosystems, Foster City, CA), and the cDNA was subjected to real-time quantitative PCR using an ABI Prism 7900 sequence detection system and gene-specific primers for insulin-like growth factor I (IGF-I; Mm00439561_m1), IGF-binding protein-2 (IGFBP-2; Mm00492632_m1), IGFBP-3 (Mm00515156_m1), IGFBP-5 (Mm00516037_m1), AR (Mm00442688_m1), MAFbx (Mm00499518_m1), MuRF1 (Mm01185221_m1), Mstn (Mm00440328_m1), medium-chain acyl-coenzyme A dehydrogenase (MCAD; Mm00431611_m1), isocitrate dehydrogenase 3α (IDH3A; Mm00499674_m1), ATP synthase β-subunit (ATP5B; Mm00443967_g1), peroxisome proliferative activated receptor-γ coactivator-1α (PPARGC1A or PGC-1α), MyHC type 1 (Myh7b_Type_I, Mm00600544_m1), MyHC type 2a (Myh2, Mm00454991_m1), MyHC type 2b (Myh4; Mm01332518_m1), MyHC type 2x/d [Myh2d(x); Mm01332489_m1], myogenin (Myog; Mm00446194_m1), MyoD1 (Mm00440387_m1), Myf5 (Mm00435124_m1), and TATA-binding protein (Mm01277045_m1) purchased from Applied Biosystems. The transcript levels were normalized to that of TATA-binding protein in the same preparation, and the fold change relative to the non-ORX mice on day 0 was calculated as 2−ΔΔCT.

Microarray analysis.

To identify other genes altered by orchidectomy, testosterone treatment alone, voluntary wheel running alone, or testosterone and voluntary wheel running combined, aliquots of 2–4 μg of each RNA sample were subjected to microarray analysis using the GeneChip Mouse Genome 430 2.0 Array (Affymetrix, Santa Clara, CA). The qualities of raw intensity files were checked using the multivariate outlier detection method MVOutlier. Probe sets that were absent in all samples were removed. The 87 (out of 91) chips that passed quality control were normalized using the GC-RMA method. The moderated t-test implemented within the Bioconductor package was used to evaluate individual probe sets for differences between any two groups. Upregulated and downregulated genes were given positive and negative moderated t values, respectively. To reduce the false-positive rate inherent when thousands of null hypotheses are tested simultaneously, the false discovery rate (FDR) multiple testing correction method was applied to the data to obtain the FDR q value, a measurement of the rate of false discovery from the distribution of the P value. The signal was considered to be significantly altered (up- or downregulated) only if the absolute value of fold change was ≥1.5 and the q value was ≤0.05. Additionally, the data were analyzed by the gene set enrichment analysis (GSEA) method (36), with significance set at a q value <0.10 to identify pathways significantly enriched by an experimental condition (e.g., ORX, testosterone, or voluntary wheel running or the combinations).

Western blotting.

The other half of the pulverized right triceps brachii muscle (from sedentary or voluntary wheel-running mice after 3, 7, or 28 days of treatment with vehicle or testosterone) was homogenized in a buffer (containing 50 mM NaF, 1 mM Na3VO4, 2 M β-glycerophosphate, 10 mM Na4P2O4, 1% Triton, and 1 tablet protease inhibitor in 10 ml of PBS) using a TissueLyser II per the manufacturer's protocol (Qiagen). Muscle samples obtained from a separate study in which mice were injected intramuscularly with insulin to induce IGF-I signaling were processed in parallel and loaded with the test samples on each gel to facilitate comparison of results across gels. The homogenates were centrifuged for 20 min at 4°C (14,000 rpm), and the protein concentration of the collected supernatant was determined using a BCA protein assay kit (Fisher Scientific). Samples were diluted in NuPAGE LDS sample buffer with 1 M DTT, denatured for 5 min at 95°C, and 20 or 30 μg protein of each sample per lane of a 4–12% NuPAGE gel was loaded, electrophoresed at 150 V for 90 min, and transferred onto nitrocellulose membranes using an iBlot device (Invitrogen). Membranes were blocked in Tris-buffered saline (TBS) with 5% (wt/vol) nonfat carnation milk for 1 h at room temperature and incubated overnight at 4°C with primary antibody in TBS containing 5% BSA and 0.05% Tween 20. The primary antibodies used were against phospho-Akt (Thr308), phospho-Akt (Ser473), total Akt, phospho-p70S6 kinase (Thr389), phospho-p70S6 kinase (Thr421/Ser424), total p70S6 kinase, phospho-GSK-3α/β (Ser21/9), total GSK-3α, phospho-IGF-I receptor (IGF-IR)β (Tyr1316), or total IGF-IRβ, and all anti-rabbit antibodies were from Cell Signaling Technology (Danvers, MA). The membranes were then incubated with secondary antibody (anti-rabbit IgG, horseradish peroxidase linked from Cell Signaling Technology) diluted 1:2,000 in TBS containing 5% nonfat carnation milk and 0.05% Tween 20 for 1 h at room temperature and washed in TBS, and immunoreactivity was visualized using either the SuperSignal West Pico or femto rabbit IgG detection kit (Thermo Fisher Scientific, Rockford, IL); the membranes were exposed on Kodak film as well as directly on an Epichemi Darkroom system (UVP, Upland, CA). The intensities of the bands on a membrane, including the reference sample, were quantified using the Labworks Analysis Software from UVP, and the intensity of each sample band relative to the reference sample band on the same membrane was calculated to facilitate comparison of band intensities across membranes.

Statistical analysis.

Data are presented as means ± SE. Statistical analysis involved comparison of group means of ORX vs. intact (INT) mice or V- vs. T-treated mice at each of the day 3, 7, or 28 time points by one-way analysis of variance followed by Bonferroni's multiple comparison tests using GraphPad Prism 5. Differences were considered significant at P ≤ 0.05.


Testosterone is required for voluntary wheel running.

To determine the effects of androgen depletion and replacement on voluntary wheel running, male mice were depleted of androgens by orchidectomy and 7 days after orchidectomy were administered either exogenous T solubilized in corn oil V or the V alone (negative control) and housed in a cage without (S) or with (R) free access to a mouse running wheel, and the volitional running activity of each mouse was monitored for 3, 7, or 28 days. The treatment groups are delineated in Table 1. In the first 3 days on the running wheel, the distance run by each ORX mouse decreased progressively (on average, 6,000, 4,000, and 3,500 m on days 1, 2, and 3, respectively) irrespective of whether they received T or V (Fig. 1A). Thereafter, the distance run per day by ORX mice receiving V continued to decrease to <1,000 m, whereas the distance run by ORX mice supplemented with T increased progressively to >8,000 m/day (Fig. 1A). Analysis of the data in smaller time bins revealed some interesting patterns. On day 1, both the V- and T-treated groups started to run at ∼6 PM (that is, at start of the dark cycle), reached peak speeds of 400–500 m/h by 7 PM, and sustained this speed until ∼5 AM before decelerating (Fig. 1B). On day 3, both groups also started running at ∼6 PM, reached a peak speed of ∼600 m/h by 8 PM, and immediately started to decelerate (Fig. 1C). By days 4–28, the running profile of mice receiving T diverged from that of mice receiving only the vehicle as a negative control. Whereas acceleration in the first 30–90 min of running was comparable, V-treated mice peaked at 300–600 m/h and then started to decelerate, and mice on T continued to accelerate to a peak speed of 1,000–1,200 m/h and sustained a speed within 80% of this peak for ∼6 h before decelerating (Fig. 1, D–G). This running pattern resulted in a gradual increase in total distance run by the T-treated mice such that, by days 7–28 after the initiation of running, T-treated mice were running ∼3–6 times the distance run by V-treated ORX mice (Fig. 1A) as a result of both a greater duration and speed of running during the evening hours (Fig. 1, D–G).

Fig. 1.

The running profile of orchidectomized (ORX) or intact (INT; non-ORX) mice treated with vehicle (V) or testosterone (T). Mean distance run/day (blocks indicate days further analyzed as distance run/h; A) and distance run/h on days 1 (B), 3 (C), 7 (D), 14 (E), 21 (F), and 28 (G) by ORX mice treated with corn oil V (gray line) or 10 mg·kg−1·day−1 sc T (black line). H: INT mice (black line) were compared with ORX mice treated with V from days 1 to 14, followed by treatment with T from days 15 to 28 (gray line) (n = 5–6).

To further confirm that the decreased running in the vehicle-treated ORX mice was indeed due to orchidectomy-induced T depletion, in a subsequent smaller study, we compared the running profile of intact (sham-ORX) mice treated with the V from days 1 to 28 with ORX mice treated with V for the first 14 days, followed by treatment with T from days 15 to 28. The results (Fig. 1H) demonstrated that the running profile of sham-ORX intact mice is almost identical to that of ORX mice treated with T, with both characterized by a slight decrease followed by progressive increase in distance run per day (presumably reflective of a training effect) such that, from about day 5, the distance run per day was three to five times greater compared with ORX V-treated mice (compare the running profile of the T-treated group in Fig. 1A and the INT + V group in Fig. 1H). In comparison, distance run per day by ORX mice receiving the vehicle did not increase during the 14 days they were treated with V but rapidly increased to the levels in INT mice within 3 days of the start of treatment with T (Fig. 1H). These data confirmed the pivotal role of T in determining volitional wheel running activity.

Body and organ weights and food consumption.

Body weight declined transiently but insignificantly in ORX mice (Supplemental Fig. S1A; Supplemental Material can be found online at the AJP-Endocrinology and Metabolism web site), and food intake declined in all mice on day 0 (i.e., by 7 days after anesthesia with or without ORX), which was presumably related to anesthesia (Supplemental Fig. S1B). Treating ORX mice with T beginning on day 0 (i.e., 7 days after ORX) increased the body weight of sedentary (but not of voluntary wheel running) mice by days 14–28 compared with V treatment (Supplemental Fig. S1A). Interestingly, the increase in body weight in the sedentary T-treated mice occurred without an increase in food consumption, whereas food intake was increased in the voluntarily running T-treated mice, although their body weight did not increase (Supplemental Fig. S1B).

ORX and testosterone inversely perturb muscle weights and muscle fiber cross-sectional area.

The weight of the BLA is very sensitive to changes in T levels and was therefore used to verify the effectiveness of ORX to deplete endogenous T and the appropriateness of the dose of T used for supplementation (Fig. 2A). On days 7 and 35 after ORX (i.e., days 0 and 28 of treatment), the weights of the BLA muscle were 24 and 67% less than in INT mice, respectively (Fig. 2A). In both sedentary and running mice, T treatment for 7 or 28 days significantly increased the weights of the BLA muscles toward the levels in INT mice (Fig. 2A).

Fig. 2.

Mean wet weights of the bulbocavernosus and levator ani (BLA) muscle complex (A), triceps brachii (Br) muscle (B), pooled foreleg (FL; triceps Br and biceps Br) (C) and pooled hindleg (HL; tibialis anterior, soleus, plantaris, gastrocnemius, and quadriceps) muscles (D) of the triceps Br muscles, and serum creatine kinase (CK) levels (E) and serum CK levels normalized to distance run in the preceding 24 h (F) of INT or ORX sedentary (S) or voluntarily running (R) mice treated with corn oil V or T for 3, 7, or 28 days. Data are means ± SE (n = 5–6). *P < 0.05, **P < 0.01, and ***P < 0.001 between INT vs. ORX and V- vs. T-treated mice at the same time point.

The effects of testosterone depletion and replacement on weight-bearing muscles were also assessed (Fig. 2, B–D). There was a tendency for appendicular skeletal muscles to be smaller in ORX than in intact mice, but the differences were significant only for the pooled (combined weight of the biceps and triceps brachii) muscles of the forelimb (Fig. 2C). Conversely, T treatment in sedentary mice tended to increase the weight of all appendicular skeletal muscles, with the increase being significant for the tibialis anterior (Supplemental Fig. S1C) and the triceps brachii muscles on day 28 (Fig. 2B). In voluntary wheel-running mice, although there was a similar trend for muscle weight to be greater after testosterone compared with vehicle treatment, the differences were significant only for the bicep brachii muscle on day 7 (Supplemental Fig. S1F) and the pooled muscles of the forelimb on day 28 (Fig. 2C). A similar, statistically insignificant, trend was present for the individual and pooled muscles of the hindlimb (Supplemental Fig. S1, C–F and Fig. 2D).

Serum levels of creatine kinase tended to be greater in V- compared with T-treated running mice on days 3 and 7, but the differences were not significant, presumably because of interanimal variations (Fig. 2E). However, given that testosterone-treated mice were running two- to threefold the distance run by vehicle-treated mice, serum creatine kinase levels per distance run were significantly less on day 7 in testosterone-treated mice than in vehicle-treated mice (Fig. 2F).

H & E staining of sections of the biceps brachii muscles of all treatment groups on day 28 revealed no qualitative differences in muscle fiber morphology between groups (Fig. 3, A–D). Interestingly, evaluation of sections for SDH activity and MyHC type 2A immunoreactivity revealed that the combination of voluntary running and testosterone treatment increased the percent composition of fibers with strong SDH reactivity, reduced the composition of fibers with weak SDH reactivity (Fig. 3, E–H and M), and increased the percent composition of fibers expressing MyHC 2A (Fig. 3, I–L and N). Histomorphometric analysis indicated that mean fiber cross-section area was reduced, although not significantly so, and the frequency distribution of fiber cross-sectional area shifted to the left in ORX compared with intact mice (P > 0.05). In sedentary ORX mice, testosterone tended to increase mean fiber cross-sectional area and to shift the frequency distribution curve to the right relative to vehicle treatment. Interestingly, voluntary wheel running alone or in combination with testosterone replacement did not significantly further alter either the mean or frequency distribution of fiber cross-sectional area (Fig. 3, O and P).

Fig. 3.

Cryosections of the triceps Br muscle from ORX S mice treated with V (SD28V; A, E, and I) or T (SD28T; B, F, and J) or ORX running mice treated with V (RD28V; C, G, and K) or T (RD28T; D, H, and L) stained with hematoxylin and eosin (left) for succinic dehydrogenase activity (SDH; middle) or for myosin heavy chain (MyHC) type 2A (right). The %fibers in each section with weak, medium, or strong SDH reaction (*P < 0.05, ***P < 0.001 vs. SD28V; M), the %fibers positive for MyHC type 2A (*P < 0.05 vs. all other groups; N), and the mean (O) and frequency distribution (P) of the cross-sectional area (CSA) of fibers in each section were quantified. Data are means ± SE (n = 5–6).

Quantitative PCR analysis demonstrates regulation of atrophy- and hypertrophy-relevant genes in muscle.

To elucidate the direct transcriptional effects of testosterone on skeletal muscles, RNA samples from triceps brachii muscles of sedentary or voluntary wheel-running mice treated with vehicle or testosterone for 3, 7, or 28 days (n = 5–6 mice in each of the 16 treatment groups; Table 1) were analyzed by quantitative real-time PCR for changes in expression of select candidate genes known to be relevant to skeletal muscle atrophy and hypertrophy (13). These analyses revealed significant and consistently reciprocal changes after ORX vs. testosterone replacement for several genes (Fig. 4). The expression of the muscle-specific E3 ligases MuRF1 and MAFBx, genes associated with muscle atrophy, was upregulated three- to fourfold by ORX relative to intact mice on both days 0 and 28 and reduced significantly by testosterone therapy on days 3–7 (Fig. 4, A and B). In contrast, the expression of IGF-I, IGF-IR, and the AR transcripts, genes associated with muscle hypertrophy, was reduced (significantly only for IGF-I) by orchidectomy and elevated greater than fivefold by testosterone treatment in sedentary mice and to a comparable or lesser extent in voluntary wheel-running mice, albeit not at all time points (Fig. 4, C–E). Surprisingly, the expression of Mstn, a negative regulator of muscle mass, was also reduced by orchidectomy and elevated by testosterone in sedentary but not in voluntary wheel-running mice (Fig. 4F). Since mice treated with testosterone ran faster and for longer durations, and mitochondrial biogenesis is associated with endurance, we analyzed the muscle samples for expression of candidate mitochondrial genes. Orchidectomy downregulated [significant for uncoupling protein 1 (UCP1) only] and testosterone upregulated UCP1, UCP3, and PGC-1α transcripts, but this reciprocal relationship was not consistently apparent for the other transcripts evaluated, including transcripts for MyHC types 1, 2A, 2B, and 2X, IDH3A, ATP5B, and MCAD (Supplemental Fig. S2).

Fig. 4.

Quantitative analysis of mRNA from the triceps Br muscle for candidate genes. Transcripts for muscle-specific RING finger-1 (MuRF1; A), muscle atrophy F-box (MAFbx; B), IGF-I (C), IGF-I receptor (IGF-IR; D), androgen receptor (AR; E), and myostatin (Mstn; F) were reciprocally regulated by ORX and T replacement. Data are means ± SE (n = 4–6). **P < 0.01 and ***P < 0.001 between INT vs. ORX or V- vs. T-treated mice at the same time point.

Identification of genes uniquely and inversely regulated by orchidectomy and testosterone as markers of testosterone status.

To identify novel genes uniquely regulated by testosterone depletion and replacement, the same triceps brachii muscle RNA samples were subjected to microarray analysis. Of the 45,101 probe sets on the Mouse Genome 430 2.0 chip (representing >39,000 transcripts), 33,454 (∼74%) probe sets had present calls from at least one sample. A comparison of the gene expression profiles of the intact non-ORX vs. the ORX mice revealed that, by day 0 of the study (i.e., 7 days after orchidectomy), a total of 2,455 genes were significantly perturbed ≥1.5-fold by ORX, with 130 upregulated and 2,325 downregulated. Corresponding numbers on day 28 (i.e., 35 days after orchidectomy) were 2,306 genes perturbed ≥1.5-fold, with 192 genes upregulated and 2,114 genes downregulated compared with age-matched intact, non-ORX controls. Figure 5A is a heat map of a subset of the regulated genes in intact mice (D28-INT) and sedentary ORX mice treated for 3, 7, or 28 days with vehicle (SD3V, SD7V, or SD28V) or testosteorone (SD3T, SD7T, and SD28T) showing that testosterone replacement restored the gene expression profile in ORX mice toward the pattern in intact mice.

Fig. 5.

Genes reciprocally regulated by ORX and T replacement. Heat maps of gene expression profile after ORX in S mice treated with V (SD3V, SD7V, and SD28V) or T (SD3T, SD7T, and SD28T) compared with INT non-ORX mice (D28-INT; A) and of genes reciprocally regulated by ORX vs. T replacement at all time points (green represents downregulated genes and red upregulated genes; B). Quantitative PCR confirmation of the reciprocal regulation of transcripts for Slc2a3 (C), Kcng4 (D), advillin (Avil; E), and calcium-sensing receptor (CASR; F) by ORX and T replacement. Data are means ± SE (n = 4–6). **P < 0.01 and ***P < 0.001 between INT vs. ORX or V- vs. T-treated mice at the same time point.

To identify genes that are responsive to testosterone per se, we filtered the ∼2,400 genes regulated ≥1.5-fold by orchidectomy to identify genes that were regulated in the opposite directions and to at least a comparable magnitude by testosterone treatment in sedentary mice. By these criteria, of the 192 genes upregulated by orchidectomy on day 28 in sedentary mice, 10 (5.2%), 11 (5.7%), and 17 (8.9%) were downregulated by testosterone replacement for 3, 7, and 28 days, respectively. Similarly, of the 2,114 genes downregulated by orchidectomy, 25 (1.2%), 261 (12.3%), and 261 (12.3%) were upregulated by testosterone replacement for 3, 7, and 28 days, respectively. To further increase the fidelity of the genes identified as markers of androgen signaling, we filtered the data for genes that were regulated ≥1.5-fold in one direction at all time points (days 0 and 28) after ORX and ≥1.5-fold in the opposite direction at all time points (days 3, 7, and 28) after testosterone treatment in sedentary mice. This most stringent analysis identified a total of nine genes that were regulated (7 that were downregulated and 2 that were upregulated) at both days 0 and 28 after orchidectomy and reciprocally regulated (up- and downregulated, respectively) at all time points (days 3, 7, and 28 inclusive) after testosterone treatment in sedentary mice. Figure 5B shows a heat map of the nine genes, two of which are nonannotated. The fold change and associated q values for these transcripts are summarized in Table 2. These microarray findings were subsequently confirmed by quantitative PCR using gene-specific primers for slc2a3 (solute carrier 2, the facilitated glucose transporter, GLUT3), slc38a4 (solute carrier 38, member 4, a sodium-coupled amino acid transporter; data not shown), Kcng4 (potassium voltage-gated channel, subfamily G, member 4), advillin, and CASR (the calcium-sensing receptor) (Fig. 5, C–F).

View this table:
Table 2.

The FC and q values for transcripts reciprocally regulated at all time points by orchidectomy and testosterone treatment in sedentary mice

ORX suppresses and testosterone activates the IGF-I/Akt pathway.

Consistent with the changes in IGF-I transcript expression, Western blot analysis of the triceps brachii muscle indicated that IGF-I signaling was suppressed by orchidectomy and enhanced by testosterone. The level of phospho-Akt (both Thr308 and Ser473) compared with total Akt (which was unchanged) was significantly less in ORX than in intact mice and less in sedentary or wheel-running mice treated with vehicle compared with testosterone-treated mice (Fig. 6, A–C). Similar results were obtained for the ratio of phospho-p70 (residues Thr389 and Thr421/Ser424) to total p70 (Fig. 6, D–F). In comparison, phospho-GSK-3α/β, total GSK-3β, phospho-IGF-IRβ, and total IGF-IRβ were not perturbed by orchidectomy or testosterone treatment (Supplemental Fig. S3).

Fig. 6.

Effects of orchidectomy and T replacement on phosphorylation of Akt and p70. A: protein extracts from the triceps Br muscle were blotted for phospho-Akt (Thr308), phospho-Akt (Ser473), and total Akt. B and C: quantification of the intensity of the bands for phospho-Akt(308) and phospho-Akt(473) normalized to total Akt and expressed relative to the D0-INT group. D: proteins from the triceps Br muscle were blotted for phospho-p70(389) and phospho-p70(424) residues. E and F: quantification of the intensity of the bands for phospho-p70(389) and phospho-p70(424) normalized to total p70 and expressed as fold change from the D0-INT group. Data are means ± SE (n = 3–6). *P < 0.05, **P < 0.01, ***P < 0.001 between INT vs. ORX and V- vs. T-treated mice at the same time point.

ORX suppresses and testosterone activates energy metabolism pathways.

Since the mechanism for testosterone-induced changes in muscle mass and function might involve minor changes in multiple genes in a given pathway rather than large changes in a few genes, the data were analyzed by the GSEA method (36) to identify pathways significantly enriched by the experimental treatments. This analysis identified several pathways to be significantly enriched by the various manipulations. The top 10 up- and downregulated pathways on day 28, ranked by the −log10 q values, are summarized in Supplemental Fig. S4, A and B. Two major patterns (clusters) of regulation were observed. In the more common and striking patterns, ORX downregulated the pathways (involved with metabolism/oxidative phosphorylation, ubiquinone, vitamin, carbohydrate, aminoacid, lipid), testosterone replacement alone restored them to baseline levels, voluntary running alone had little or no additional effect, and voluntary running plus testosterone further significantly upregulated the pathways. The changes in expression of the probe sets in these pathways after the combination of voluntary running and testosterone treatment vs. vehicle treatment in sedentary mice are summarized in Supplemental Fig. S4A. Of note, relative to vehicle-treated sedentary mice, the combination of voluntary running and testosterone replacement upregulated ∼95 and 82% of the probe sets in the oxidative phosphorylation and ubiquinone pathways, respectively, perturbing especially genes in complex I and to a lesser extent complexes III, IV, and V of the mitochondrial electron transport chain (Supplemental Fig. S4C). In the second cluster (comprising pathways involved with a wide variety of processes, including cytoskeleton remodeling, cell cycle, immune response, and proteolysis), ORX, testosterone alone, or voluntary running alone minimally regulated the pathways, but the combination of testosterone plus voluntary running significantly downregulated them. Surprisingly, there was no cluster of pathways that was significantly upregulated by orchidectomy and downregulated by voluntary running alone or with testosterone treatment.


Loss of skeletal muscle mass, strength, and performance is a common complication of physical inactivity, chronic wasting diseases, prolonged or high-dose glucocorticoid therapy, and the natural aging process in mammals. Also common to most of these conditions are declines in both physical activity and endogenous androgen levels. Yet it is not clear whether a decline in androgen levels is causally related to the declines in physical activity and muscle mass and performance in these disease states.

To determine whether there is a causal relationship between testosterone levels and physical activity, in this study, male mice were ORX to deplete endogenous testosterone and assessed for voluntary wheel-running activity in comparison first with testosterone-supplemented ORX mice and later with intact, sham-operated mice. In all three groups, the first 3 days of running were associated with progressive decline in distance run per day, presumably reflecting fatigue or soreness associated with initiation of running. Thereafter, however, the distance run increased progressively to a peak of ∼10,000 m/day in the intact and testosterone-treated ORX mice (presumably a “training effect”) but continued to decline to below ∼3,000 m/day in ORX mice treated with vehicle. The running patterns described over the 28-day period (the decline in distance run over the initial 3 days) as well as the difference in running patterns between testosterone- and vehicle-treated ORX mice by days 4–28 have not been reported previously, presumably because in previous studies the animals were either acclimated to the running wheels for several days before monitoring began or the data for several (from 2 to 10) days were binned together (8). Indeed, the importance of analyzing activity data in small bins was previously highlighted by the finding that, whereas mdx and wild-type control mice are indistinguishable with regard to their voluntary wheel running ability when assessed by total distance run per day (24-h bins), analysis of the data in smaller bins (5-min bins) revealed an intermittent running pattern for mdx mice quite distinct from a continuous running pattern for wild-type mice (15). By analyzing the running data in smaller bins in this study, we were able to demonstrate that testosterone administration to ORX mice significantly increased both speed and duration of running. It was shown previously that continuous testosterone administration prevented orchidectomy-induced atrophy of fast (extensor digitorum longus) and slow (soleus) muscles and increased resistance to evoked tetanic fatigue only in the soleus muscle compared with vehicle treatment in mice (1). Similarly, in noncastrated weight-lifting rats, administration of the anabolic steroid nandrolone decanoate did not change the maximum weight a rat could lift but significantly increased the number of times the weight could be lifted (thus, the total weight lifted) and reduced the increase in serum creatine kinase levels after the bout of weight lifting, suggesting, respectively, an increase in endurance and decrease in muscle injury by anabolic steroids (37).

Unlike the rapid and dramatic changes in voluntary wheel running after testosterone depletion and replacement, the changes in muscle morphology (muscle mass and fiber cross-sectional area) were slower in onset and modest in appendicular skeletal muscles. However, muscle mass changes were rapid and dramatic in the BLA muscle complex. These findings are consistent with previous reports on the relative sensitivities of appendicular and levator ani muscles to androgens (1) and the relative inability of voluntary wheel running to induced hypertrophy, especially in fast-twitch muscles (19). To help elucidate potential mechanisms for the effects of testosterone on skeletal muscle, we analyzed the triceps brachii muscle from sedentary or wheel running mice that had been treated with vehicle or testosterone for 3, 7, or 28 days for candidate genes known to be inversely regulated during muscle anabolism and catabolism (22) as well as by microarray analysis to identify novel genes responsive to testosterone per se by virtue of their being regulated in one direction by orchidectomy and regulated in the opposite direction and to at least a similar magnitude upon testosterone replacement for 3, 7, and 28 days. We analyzed the triceps brachii rather than the more testosterone-responsive BLA muscle because the latter, although a striated muscle, is not typical or representative of the more clinically relevant appendicular skeletal muscles for which anabolic steroids are used. However, the marked sensitivity of BLA muscle weight to changes in androgen status was used, in lieu of serum testosterone levels being measured, to confirm the adequacy of testosterone depletion by orchidectomy and the appropriateness of the testosterone replacement dose. The duration of the study was limited to 35 days after orchidectomy because muscle atrophy after denervation or orchidectomy almost reaches a plateau during this period (1, 32).

Candidate gene analysis confirmed the reciprocal regulation of MuRF1 and MAFbx by orchidectomy (upregulated) and testosterone treatment (back downregulated), consistent with increased and reduced catabolism, respectively (22), thus confirming a role for testosterone in regulating muscle protein catabolism. In contrast to MuRF1 and MAFBx, the expression of Mstn, a negative regulator of muscle mass, was downregulated after orchidectomy and restored toward intact control levels by testosterone replacement. This finding is consistent with a report of reduced Mstn mRNA in muscles of chronic spinal cord-injured patients (24) but contrary to the report that Mstn mRNA tends to be reduced with resistance training and increased with detraining in humans (20) and reports of increased levels of Mstn-immunoreactive protein in serum and muscle of subjects with muscle wasting (14, 20, 30, 42, 43). However, it is noteworthy that the protein bands identified as Mstn in some of these studies have been shown to also be present in muscles from Mstn knockout mice (20). Nonetheless, a downregulation of Mstn mRNA after orchidectomy-induced muscle atrophy and upregulation after testosterone restoration of muscle mass will be consistent with the hypothesis that Mstn is a chalone that negatively regulates muscle mass to status quo (23, 24, 26). The apparently counteracting changes in atrogenes (MuRF1 and MAFBx) on the one hand and Mstn on the other might have thus contributed to the rather slow and modest changes in muscle mass after orchidectomy or testosterone treatment. On the other hand, the expression of IGF-I, IGF-IR, and AR mRNA was downregulated by orchidectomy (significantly only for IGF-I) and upregulated by testosterone replacement, in agreement with previous reports in humans (10, 27, 39). Furthermore, Western blot analysis of these same muscle samples showed that phosphorylation of Akt and p70 was downregulated after orchidectomy and back upregulated to or above intact control levels after testosterone treatment.

Although the increase in IGF-I signaling and decrease in expression of atrogenes after testosterone treatment would explain the restoration of muscle mass in ORX mice, they do not explain the combination of increased speed and duration of running after testosterone replacement. Although the rapidity of the changes in running behavior after orchidectomy vs. testosterone treatment might suggest a centrally mediated mechanism, we nonetheless explored whether peripheral muscle mechanisms also contribute. The findings that 1) expression of UCP1, UCP3, and PGC-1α mRNA, although not of IDH3A, ATP5B, and MCAD mRNA, was consistently and reciprocally regulated by orchidectomy and testosterone and 2) the combination of voluntary running and testosterone replacement was associated with an increase in fiber SDH (oxidative) capacity and type 2A fiber composition would suggest that these probably contributed, even if only partially, to the changes in running behavior after orchidectomy or testosterone replacement. Another potential explanation for the changes in running behavior may have to do with the reported wound healing effects of testosterone (10, 43). Indeed, relative to sedentary ORX mice treated with the vehicle, serum creatine kinase levels on days 3 and 7 of running tended to be greater in vehicle-treated mice and less in testosterone-treated mice despite the latter mice running two- to threefold more, consistent with the report that anabolic steroids increased endurance and reduced serum creatine kinase levels in weight-lifting rats (42).

Microarray analysis of RNA samples from the triceps brachii muscles also identified other novel genes that are responsive to testosterone per se and that might contribute to explain the phenotypes observed. Of the ∼2,300 genes perturbed ≥1.5-fold in skeletal muscle by ORX, <0.5% were completely inversely regulated by testosterone replacement at all time points, and four of these were in the solute carrier (SLC) family of genes. The latter might suggest that a conserved AR-responsive promoter element dominates the expression control of this family. SLC2, -15, -30, and -38 were all downregulated by orchidectomy and upregulated by testosterone replacement. Of note, Slc2a3 is GLUT3, a major glucose transporter that was found originally in fetal skeletal muscle but is expressed predominantly in the brain and neural tissues and is reported to be upregulated by IGF-I (7). A decrease in glucose transport could contribute to the decline in running capacity after orchidectomy, making it of interest to determine in future studies whether an increase in expression of GLUT3 is sufficient to revert, at least in part, the loss of activity. Kcng4 is the voltage-gated calcium channel Kv6.4. A decrease in its expression after orchidectomy suggests the possibility of a direct effect of androgen loss on muscle excitability and thus a decline in the excitability of muscles in settings of hypogonadism. In addition to these downregulated genes, CASR, the extracellular calcium-sensing receptor, was upregulated, as was advillin. CASR, a plasma membrane G protein-coupled receptor, is expressed primarily in the parathyroid gland and renal tubules and controls calcium homeostasis through regulating release of parathyroid hormone and renal cation handling and might thus influence muscle contractility. The sum of these regulations may constitute a signature for androgenic status. It is noteworthy that the candidate genes confirmed to be reciprocally regulated by orchidectomy and testosterone replacement (e.g., MuRF1, MAFBx, IGF-I) were not identified by the microarray analysis, presumably because of the very stringent filtering criteria; the fold changes for MuRF1 and MAFbx ranged from −1.0 to −1.4 after orchidectomy and +1.4 to +1.7 after testosterone replacement. Also of note is that previous investigations of genes regulated by orchidectomy and androgen (dihydrotestosterone) replacement did not identify the testosterone uniquely responsive genes reported here presumably, at least in part, because of differences in methodology (12).

Because the mechanism for the ergogenic effects of testosterone might involve minor changes in multiple genes in a given pathway rather than large changes in a few genes, we further interrogated the microarray data by GSEA to identify pathways regulated in a reciprocal manner by orchidectomy and testosterone treatment at all time points. Interestingly, the pathways most significantly perturbed (downregulated) by orchidectomy were those involved in energy metabolism, notably the oxidative phosphorylation (OXPHOS) and ubiquinone pathways, and these were restored to or above baseline levels by the combination of voluntary wheel running and testosterone replacement. It is noteworthy that, 1) compared with vehicle-treated sedentary mice, in voluntarily running mice dosed with testosterone ∼95 and 82% of the probe sets in the OXPHOS and ubiquinone pathways, respectively, were at or above baseline levels and 2) the probe sets represented genes in complexes I, III, IV, and V of the mitochondrial electron transport chain. The regulation of these pathways by orchidectomy and testosterone replacement in this study is consistent with a previous report of decreased expression of mitochondrial genes in the OXPHOS pathway in men with low testosterone levels and low maximal aerobic capacity (29). The changes in the OXPHOS and ubiquinone pathways are also consistent with the increase in fibers with strong SDH activity as well as the increase in composition of type 2A fibers and altogether may have contributed, at least in part, to the increase in running after testosterone replacement. However, given the rapidity of onset of changes in voluntary wheel running behavior after testosterone depletion and replacement, most apparent from Fig. 1H, it is also possible that the impetus to run is centrally mediated. This is conceivable given that 1) ARs are widely distributed in the brain, 2) anabolic steroids strongly reinforce social behaviors, including sexual drive, competition, and aggression in both humans and animals (40, 41), and 3) such androgen-induced increases in drive, aggression, and competition could lead to increased physical activity (e.g., running or weight lifting) and, as a consequence, to increased muscle mass and strength (3) as well as anecdotal improvements in vitality and physical endurance. Irrespective of the underlying mechanisms, however, the present findings of depletion and restoration of voluntary wheel running after orchidectomy and testosterone treatment, respectively, are consistent with a causal relationship between testosterone levels and voluntary wheel running.

In summary, the present results indicate that testosterone plays a critical role in modulating voluntary wheel running capacity (both speed and duration of running, hence endurance) and a modest role in regulating appendicular skeletal muscle mass in mice. The study identified novel genes that were reciprocally regulated by orchidectomy and testosterone replacement and elucidated some potential mechanisms for the effects of testosterone, including upregulation of genes and pathways involved in energy metabolism (most notably, oxidative phosphorylation), downregulation of the atrogenes necessary for muscle atrophy, and upregulation of IGF-I signaling. To the extent that a decline in testosterone levels is a common accompaniment of many conditions with reductions in physical activity, muscle mass, and muscle performance, these findings support the utility of testosterone-like agents to manage the decline in muscle performance observed in these conditions. Second, to the extent that the effects of testosterone on voluntary wheel running can be extended to human physical activity and are centrally mediated, even if only in part, the present findings would suggest that therapies that cross the blood-brain barrier and mimic testosterone-mediated signaling in the brain might be advantageous over those that do not. These remain areas to be understood to aid in the development of effective and safe anabolic agents to treat the muscle dysfunction of disease and aging.


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


We thank the Muscle Diseases Group at the Novartis Institutes for Biomedical Research (NIBR) for their enthusiastic support along with the rest of the NIBR community, in particular Mark Fishman, Brian Richardson, and Andrew Mackenzie. We also thank Nicole Hartman (NIBR Translational Sciences) for microarray analysis of the samples and Jeffrey Brown (NIBR, Cardiovascular and Metabolic Diseases) for the serum creatine kinase analysis.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
View Abstract