Adiponectin is expressed by skeletal muscle fibers and influences muscle phenotype and function

Abstract

Adiponectin (Ad) is linked to various disease states and mediates antidiabetic and anti-inflammatory effects. While it was originally thought that Ad expression was limited to adipocytes, we demonstrate here that Ad is expressed in mouse skeletal muscles and within differentiated L6 myotubes, as assessed by RT-PCR, Western blot, and immunohistochemical analyses. Serial muscle sections stained for fiber type, lipid content, and Ad revealed that muscle fibers with elevated intramyocellular Ad expression were consistently type IIA and IID fibers with detectably higher intramyocellular lipid (IMCL) content. To determine the effect of Ad on muscle phenotype and function, we used an Ad-null [knockout (KO)] mouse model. Body mass increased significantly in 24-wk-old KO mice [+5.5 ± 3% relative to wild-type mice (WT)], with no change in muscle mass observed. IMCL content was significantly increased (+75.1 ± 25%), whereas epididymal fat mass, although elevated, was not different in the KO mice compared with WT (+35.1 ± 23%; P = 0.16). Fiber-type composition was unaltered, although type IIB fiber area was increased in KO mice (+25.5 ± 6%). In situ muscle stimulation revealed lower peak tetanic forces in KO mice relative to WT (−47.5 ± 6%), with no change in low-frequency fatigue rates. These data demonstrate that the absence of Ad expression causes contractile dysfunction and phenotypical changes in skeletal muscle. Furthermore, we demonstrate that Ad is expressed in skeletal muscle and that its intramyocellular localization is associated with elevated IMCL, particularly in type IIA/D fibers.

obese and type 2 diabetic patients display altered endocrine profiles, in particular, the factors secreted by adipose tissue (adipokines). One such adipokine, adiponectin (Ad), has recently been the subject of intense study, and alterations in Ad expression levels are now known to play a crucial role in the progression of various disease states (7, 9). Specifically, Ad expression is inversely correlated with adiposity; thus, obese patients have decreased Ad levels in circulation (1). It is likely that this Ad decrement contributes to disease progression, because strong correlations exist between circulating Ad and various components of the metabolic syndrome (12, 13, 19), and Ad-deficient animals tend to exhibit mild to moderate insulin resistance (10, 15). Furthermore, Ad administration in both wild-type (WT) or diabetic animal models enhances insulin sensitivity and corrects metabolic abnormalities, at least in part via increasing β-oxidation of lipids in skeletal muscle and liver (2, 26–28, 31).

Ad is found in the blood serum, with concentration ranging from 1 to 17 μg/ml (1), existing in several forms: low-molecular-weight trimers, medium-molecular-weight hexamers, and high-molecular-weight multimers (7, 24, 25). It was originally thought that Ad expression and secretion were limited to adipocytes; however, recent evidence has demonstrated its expression in other cells types and tissues. For example, cardiomyocytes derived from the atria (20) and the ventricles (5) express Ad, as does cardiac muscle in healthy and streptozotocin-induced diabetic rats (8). Interestingly, there are some recent reports indicating that Ad may be expressed in skeletal muscle. Specifically, Ad expression has been detected in C₂C₁₂ myotubes (3, 4) and in rodent skeletal muscle tissue (3, 4, 11, 30). However, these findings are contrasted by immunohistochemical (IHC) studies showing Ad localization in the perimysium and epimysium, but not within the muscle fiber proper (4, 30).

Previous studies have found increased lipid oxidation in skeletal muscle in response to Ad administration (19, 26–28, 31). Ad binding to its receptors activates downstream AMP-activated protein kinase and peroxisome proliferator-activated receptor-α pathways, leading to increased β-oxidation of lipids (28, 31). While it has been demonstrated that a reduced circulating Ad level occurs in obesity (1), and that white adipose tissue mass is increased in Ad-deficient animal models (29), to the best of our knowledge, no studies have examined the role of Ad on intramyocellular lipid (IMCL) content. In fact, very little is known about the effects that Ad exerts on skeletal muscle phenotype, or for that matter, function (i.e., contractility). Thus, in the present study we used Ad-null [knockout (KO)] mice (14) to determine the effect of Ad on several measures of skeletal muscle phenotype such as IMCL, fiber type composition, fiber area/size, and capillary density. We proposed that any change in skeletal muscle phenotype would change the functional capacity of the muscle and assessed this using an in situ electrical stimulation protocol.

The present findings indicate that skeletal muscle fibers produce Ad and, more specifically, the intramyocellular Ad levels are primarily associated with type IIA/D fibers containing elevated IMCL. We also demonstrate for the first time that adult Ad-KO mice display an elevated body mass, an accumulation of IMCL, an increase in type IIB fiber area, and a reduction in peak contractile force of muscle compared with WT littermates.

METHODS

Cell lines and animals.

Low-passage (<15) proliferating L6 skeletal muscle cells (a kind gift from Dr. Amira Klip, The Hospital for Sick Children, Toronto, Canada) were expanded in culture [α-modified minimal essential medium (αMEM) supplemented with 10% (vol/vol) FBS and 1% streptomycin/amphotericin B] until appropriate numbers for experimentation were achieved. Cells were then seeded at high density (∼80% confluence) and exposed to differentiating media [αMEM supplemented with 2% (vol/vol) FBS and 1% streptomycin/amphotericin B] to promote myotube formation. Myotubes were harvested 7 days following exposure to differentiating media.

Ad-KO mice were generated as previously described (14). Ad-KO and WT mice were housed in standard cages and had access to standard diet and water ad libitum. The animal room was maintained at a constant temperature of 21°C, a constant humidity of 50%, and a 12:12-h light-dark cycle. All experiments were approved by the Animal Care Committee at York University and were conducted in accordance with guidelines set forth by the Canadian Council for Animal Care.

Detection of adiponectin mRNA.

TRIzol (Invitrogen, Burlington, Canada) reagent was used according to the manufacturer's instructions to collect total RNA from L6 skeletal muscle cells and murine skeletal muscle. RNA integrity was confirmed by ethidium bromide staining of agarose gels, as well as by optical density (OD) absorption ratio OD₂₆₀/OD₂₈₀ nm. Total RNA (1 μg) was reverse transcribed with Superscript II RNAse H-reverse transcriptase using random primers and oligo(dT) (Invitrogen). PCR analyses were carried out using 400 ng of cDNA with 500 nM of both sense (5′-GCAGAGATGGCACTCCTGGA-3′) and antisense (3′-CCCTTCAGCTCCTGTCATTCC-5′) primers. cDNA was amplified on an Eppendorf thermocycler using the following protocol: initial denaturation step of 95°C for 15 min followed by 35 cycles of 1) 95°C for 30 s, 2) 65°C for 30 s, and 3) 72°C for 30 s, and then a final extension at 72°C for 5 min.

To confirm that it was, in fact, Ad mRNA being detected, total DNA from mouse and rat skeletal muscle was extracted and amplified using a DNA extraction kit according to the manufacturer's instructions (XNATR, REDExtract-N-Amp Tissue PCR, Sigma-Aldrich, Oakville, Canada). The PCR product obtained was sequenced and confirmed to be Ad.

Western blot analysis for adiponectin.

Lysates were prepared for immunoblot analysis from L6 myotubes grown in six-well plates in nonheating, nonreducing condition as previously described (6). Before being loaded onto SDS-PAGE gels, the samples were diluted 1:1 (vol/vol) with 2× Laemmli sample buffer [62.5 mM Tris·HCl (pH 6.8), 2% (wt/vol) SDS, and 0.01% (wt/vol) bromophenol blue]. Equal amounts of proteins (30 μg) were resolved by SDS-PAGE (6% and 12%) and then transferred to polyvinylidene difluoride membranes (Bio-Rad, Burlington, ON, Canada). Membranes were probed with rabbit polyclonal primary antibody against full-length Ad followed by horseradish peroxidase-conjugated secondary antibody (anti-rabbit at 1:10,000 dilution) as described previously (14). Presence of Ad was detected by the enhanced chemiluminescence method.

Experimental protocol.

Ad-KO and WT mice (n = 3 and n = 4 in Ad-KO and WT groups, respectively) were killed at 24 wk of age and were handled similarly in all aspects of the experiment. Body weight was measured biweekly from 20 wk of age until the time of death. Mice were sedated with a ketamine/xylazine intraperitoneal injection (ketamine: 150 mg/kg; xylazine: 10 mg/kg) before surgery. The gastrocnemius/plantaris/soleus (GPS) complex was isolated from its distal insertion, keeping nerve and vasculature intact, and secured to a Grass FTO3 force-displacement transducer by a small metal hook fastened to the calcaneal tendon. The leg was secured by screws tightened onto the medial and lateral condyles of the femur. Using a Grass S44B Stimulator, the muscle was stimulated through twin platinum electrodes that were applied directly to the surface of the muscle. Optimal voltage was determined by generating single-twitch contractions at increasing voltages until no increase in single-twitch force production was observed. Optimal muscle length was determined in a similar manner. Specifically, muscle length was manipulated and single-twitch force production was observed. The length and voltage at which a single twitch produced the greatest force were used throughout the stimulation protocol. The direct muscle stimulation protocol consisted of determination of the force that could be achieved during sequential increases in stimulation frequency followed by a low-frequency stimulation period of 2 min to determine rate of fatigue. With respect to force determination, this was achieved using a repeated stimulation of the muscle for ∼1 s separated by 4 s of rest; the stimulation frequency began at 2 Hz and was increased in 2-Hz increments until no further increase in force was observed. The peak force was produced at an average frequency of 22.3 ± 1.5 Hz. Following 2 min of rest, the muscle was stimulated at optimal voltage with twitch contractions at 2 Hz for 2 min to assess low-frequency fatigue rate. This fatigue protocol was chosen because it was of sufficiently low metabolic demand (O₂ supply and oxidative energy demand could be met) as to allow us to ascertain if there may be differences in the functionality of the contractile and/or calcium handling proteins. Muscle temperature was monitored and maintained at ∼35–37°C, while the muscle was kept hydrated with saline. Data were collected through an AD Instruments Bridge Amp and Powerlab 4/30 and were analyzed with Chart5 PowerLab software for Windows. Following the fatigue protocol, the tissues were immediately collected and either snap frozen or mounted on cork using mounting medium and quick frozen in isopentane cooled by liquid nitrogen. Samples were maintained at −80°C until analyzed.

Histochemical analyses.

Tibialis anterior 10-μm cross sections were cut in a ThermoShandon Cryotome and picked up on untreated glass slides. Metachromatic fiber type staining was performed as per Ogilvie and Feeback (17). Type I, IIA, IIB, and IID fibers appear dark blue, white/pink, purple, and blue, respectively. IMCL content was determined by staining muscle sections overnight at 32°C with a solution of propylene glycol saturated with Oil Red O (O9755, Sigma Aldrich). To determine muscle capillary density, the Lead ATPase stain was performed as per Rosenblatt et al. (21), which turns capillaries a dark brown.

Immunohistochemical analyses.

To determine the localization of Ad within muscle cross sections, we performed serial IHC staining for dystrophin, a protein localized to the muscle membrane, followed by Ad. Tibialis anterior sections from WT and Ad-KO mice were cut for IHC analysis of muscle Ad content, were immediately fixed in ice-cold 4% paraformaldehyde for 5 min, and were then thoroughly rinsed in phosphate-buffered saline. Sections were blocked in 10% normal horse serum for 50 min (NHS; S-2000, Vector Laboratories, Burlington, Canada), incubated in mouse monoclonal anti-dystrophin primary antibody for 1 h (1:100 in 1.5% NHS; ab7164-100, AbCam, Cambridge, MA), then incubated in horse anti-mouse fluorescein isothiocyanate secondary antibody for 1 h (1:100 in 1.5% NHS; FI-2080, Vector Laboratories). Sections were subsequently blocked in 10% normal goat serum for 50 min (NGS; S-1000, Vector Laboratories), incubated in rabbit polyclonal Ad primary antibody for 1 h (1:100 in 1.5% NGS), incubated in biotin goat anti-rabbit secondary antibody for 10 min (1:100 in 1.5% NGS; BA-1000 Vector Laboratories), and then incubated in streptavidin Texas Red for 10 min (1:100 in 1.5% NGS; SA-5006, Vector Laboratories). Nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) in the dark for 2 min (1:1,000; D-8417, Sigma-Aldrich). Negative control samples were exposed to all conditions except primary antibodies, where they were incubated with 1.5% serum for the appropriate period of time. In all cases, negative controls displayed an absence of signal.

Microscopy and data analysis.

Images were acquired with a Nikon Eclipse 90i microscope, Nikon Mercury Lamp fluorescent bulb, and Q-Imaging MicroPublisher 3.3 RTV camera with Q-Capture software for Windows XP. IHC images were taken with the appropriate filter sets for fluorescein, Texas Red, and DAPI and were then superimposed using Adobe Photoshop CS2 to determine Ad localization. Fiber-type percentage was determined by manually counting each fiber type (an average of 473 fibers were counted per muscle section; n = 3 for both groups). Fiber size was determined for each fiber type (an average of 89 fibers were counted per muscle section; n = 3 for both groups). Capillary-per-fiber ratio was determined by manually counting the total number of capillaries and dividing by the number of fibers counted (>500 fibers were counted per muscle section; n = 3 for both groups). Intensity of Oil Red O staining of IMCL droplets was assessed with Simple PCI 6 software for Windows XP, a method essentially the same as that used by Malenfant et al. (16) (30 fibers were counted per muscle section; n = 3 for both groups). Serial sections from the GPS complex of three WT mice were stained for IMCL (Oil Red O), fiber type (metachromatic), and Ad expression (IHC), respectively. More than 600 total fibers were analyzed for fiber type, IMCL, and intramyocellular Ad expression intensity. Images were converted to grayscale to determine optical density with Scion Imaging software. The grayscale is evaluated on a range of 0 (completely black) to 255 (completely white). The x- and y-axes of Fig. 2B represent the optical density quantified for IMCL and Ad within this grayscale range (0 to 255). To determine a relationship between muscle fiber type, IMCL, and intramyocellular Ad expression, we undertook two specific analyses. First, a correlation analysis was undertaken to determine a relationship between IMCL and intramyocellular Ad (within and irrespective of fiber types). Second, a plot of IMCL versus intramyocellular Ad per fiber type was generated, and the 50th percentiles for IMCL and intramyocellular Ad were determined and applied as quartiles (see Fig. 2B). This allowed for the classification of each fiber into one of four categories: 1) low Ad, low IMCL (221 fibers); 2) high Ad, low IMCL (84 fibers); 3) low Ad, high IMCL (84 fibers); and 4) high Ad, high IMCL (222 fibers). In total, 65 type I, 277 type IIA, 126 type IIB, and 143 type IID fibers were counted.

Fig. 2.

Ad expression appears related to fiber type and intramyocellular lipid (IMCL) density. A: serial sections of WT mouse gastrocnemius/plantaris/soleus (GPS) stained for fiber type (left), lipid content (middle), and Ad via immunohistochemical staining (right; Ad appears red, and nuclei appear blue). Ad expression was observed surrounding all muscle fibers, whereas an elevated intramyocellular Ad expression was detected in some fibers. Type IIA, IIB, and IID fibers are labeled A, B, and D, respectively. Type I fibers are not displayed. B: quantification of optical density of WT mouse TA sections stained for IMCL and Ad. Increased Ad expression was detected predominantly in type IIA and IID fibers displaying relatively greater IMCL content while not detected in fibers with relatively lower IMCL content. Intramyocellular Ad expression was generally lower in type IIB and type I fibers. The x- and y-axes represent the optical density quantified for IMCL and Ad within the grayscale range (0 to 255). R² = 0.28, P < 0.0001. C: fibers counted in each quadrant were totaled and are expressed as either the percentage of each fiber type found within each quadrant (top table) or the quadrant distribution within each fiber type (bottom table).

Fatigue index was calculated as the mean contractile force of the last five contractions divided by the mean contractile force of the first five contractions during the 2-min fatigue protocol. All statistical analyses were performed with GraphPad Prism 5 software for Windows XP. Differences between WT and Ad-KO mice in Oil Red O IMCL stains, capillary-per-fiber ratio (muscle capillarity), body weight, muscle mass and length, epididymal fat mass, peak force, and fatigue index were determined with an unpaired t-test. Fiber size, fiber type percentage, and 2-min fatigue curves were analyzed with a two-way ANOVA and Bonferroni posttest. All data presented are means ± SE.

RESULTS

Ad expression in skeletal muscle.

Ad mRNA was detected in L6 myotubes and in WT mouse adipose tissue (Fig. 1A). The presence of Ad protein in L6 myotubes was confirmed using Western blot analysis, which displayed detectable Ad at all molecular weights (Fig. 1B). To further confirm that muscle was capable of producing Ad and to determine if there was a fiber-type expression profile of Ad, PCR analysis was performed on rat and mouse extensor digitorum longus (EDL) and soleus skeletal muscles, which are muscles primarily composed of type II and type I fibers, respectively. As can be seen in Fig. 1C, there was no detectable difference in Ad expression between muscles at the mRNA level. To ensure that it was, in fact, Ad being detected, a PCR gene product from the soleus and EDL samples was sequenced and confirmed to be Ad. Western analyses further supported the expression of Ad in WT mouse muscle, whereas, not surprisingly, Ad-KO mouse muscle homogenates showed a lack of Ad protein expression (Fig. 1B). In addition to the recombinant Ad used as a positive control (first lane, Fig. 1B), WT subcutaneous adipose tissue and WT liver were run and resulted in a positive and negative blot, respectively (data not shown). The detection of Ad protein expression in skeletal muscle was further verified using IHC (Fig. 1D). Costaining of muscle with Ad and dystrophin, to label the plasma membrane, allowed us to localize the Ad expression. What we found was a gradient of Ad expression with high levels surrounding the muscle fibers and lower levels within the muscle fiber. Using the same camera exposure time, negative controls showed no signal.

Fig. 1.

Detection of adiponectin (Ad) mRNA and protein within skeletal muscle in vitro and in vivo. A: PCR detection of Ad mRNA in L6 myotubes and adipose tissue. Ad mRNA was detected in L6 myotubes (lane 3) and in adipose tissue (lane 4) from C57BL6 mice. β-Actin was used for control lanes (L6 myotube in lane 1 and adipose tissue in lane 2). B: Western blot detection of Ad in L6 myotubes and in soleus (Sol) muscle of wild-type (WT) and Ad knockout (KO) mice compared with recombinant full-length (f) Ad. Samples were run on a 6% gel. C: PCR detection of Ad mRNA in rodent skeletal muscle. Ad mRNA was detected in extensor digitorum longus (EDL) and Sol muscles of Wistar rats and C57BL6 mice. D: immunohistochemical staining of Ad in the tibialis anterior (TA) of WT mice, with low (top) and high (bottom) magnification. Dystrophin staining, used to identify fiber membranes (left), appears green; Ad expression (middle) appears red; an overlay of these two images is at right.

On the basis of the results in Fig. 1D, we were interested to determine whether there was a relationship between fiber type, IMCL, and Ad expression. To resolve this, we examined serial sections of WT mouse GPS stained for fiber type, IMCL, and Ad expression (Fig. 2A). By plotting IMCL versus intramyocellular Ad expression and applying quartiles, we demonstrated that ∼85% of all fibers deemed to have high intramyocellular Ad were identified as type IIA or IID. Furthermore, >72% of all fibers with elevated intramyocellular Ad also demonstrated elevated IMCL. Of the fibers that displayed high intramyocellular Ad and high IMCL, 94% were type IIA/D, with type I and IIB contributing the remainder equally (Fig. 2B). It is important to note that not all type IIA/D fibers with high lipid density expressed higher levels of Ad.

Ad-KO muscle phenotype and function.

Body weight was not significantly different between the Ad-KO and WT mice until 22 wk of age, at which point the Ad-KO mice were significantly heavier than their WT littermates. This difference in body mass continued until the time of death at 24 wk of age (Fig. 3A). No change in the mass of the GPS muscle group (Fig. 3B) was noted, even when GPS weight was related to change in body mass. No change in muscle length was observed between the two groups (WT: 1.28 ± 0.08 cm vs. KO: 1.27 ± 0.05 cm). Epidydymal fat pad mass in Ad-KO mice was not statistically different compared with WT (+35.1 ± 23%, P = 0.16; Fig. 3C).

Fig. 3.

Changes in body weight, fat, and muscle mass within Ad-null (KO) and WT mice. A: body mass was measured biweekly from 20 wk of age until the time of harvest at 24 wk of age. GPS (B) and epididymal fat mass (C) from WT and Ad-KO mice were assessed at the time of death (24 wk of age). n = 4 in both WT and KO groups. *Significant difference between WT and KO groups (P < 0.05).

To obtain a complete representation of the effects of Ad disruption on skeletal muscle, we examined the functional capacity of the GPS muscle group in WT and Ad-KO mice. Although a significant reduction in peak tetanic force (relative to GPS mass) was detected in Ad-KO muscle compared with WT muscle (Fig. 4A), no change was observed in rate of muscle fatigue during a 2-min low-frequency stimulation protocol (Fig. 4B). Likewise, fatigue index was not different between groups (Fig. 4C). Given the differences observed in peak tetanic force, it was prudent to examine the skeletal muscle for changes in fiber type composition and size. Metachromatic staining of fiber type (Fig. 5A) did not reveal any significant differences in the overall fiber type composition of the tibialis anterior (Fig. 5B, top); however, when we quantified fiber area, a significant increase in area was detected in the Ad-KO type IIB fibers compared with that of WT muscle (Fig. 5B, bottom).

Fig. 4.

Muscle function in the presence and absence of Ad. A: peak force, measured in grams, relative to GPS weight. Individual data points are plotted. *Significant difference in relative peak force between WT and KO groups (P < 0.05). B: contractile force during 2-min fatigue protocol relative to peak force. C: fatigue index from 2-min fatigue protocol. n = 4 in WT group, and n = 3 in KO group.

Fig. 5.

Muscle fiber type composition, area, and capillarity of Ad-KO mice. A: metachromatic fiber type stain of TA from WT and KO mice. Fiber types IIA, IIB, and IID (labeled A, B, and D, respectively) are indicated. B: fiber type percentage (top) and average area of each fiber type (bottom) in WT and KO mice. Two-way ANOVA revealed a main effect for Ad KO; post hoc analysis revealed a significant difference in fiber area between WT and KO groups (*P < 0.05). C: lead ATPase capillary stain of TA from WT and KO mice. D: capillary-to-fiber ratio. Capillaries are indicated by arrows. n = 3 in both WT and KO groups.

Because previous work has demonstrated that Ad is capable of mediating angiogenesis (22), we sought to determine if there was a difference in muscle capillary density in the absence of adiponectin. Lead ATPase staining of WT and Ad-KO tibialis anterior muscle did not demonstrate any difference in muscle capillary density (Fig. 5, C and D).

Since Ad is known to increase β-oxidation of lipids within skeletal muscle, we examined IMCL content using Oil Red O staining. There was a significantly greater IMCL content in Ad-KO compared with WT muscle (+75.1 ± 25%), with this difference observed in all fibers regardless of fiber type (Fig. 6).

Fig. 6.

IMCL concentration within control and Ad-KO mice. A: Oil Red O stain for IMCL in WT and KO mice. B: average fat area as a percentage of fiber area per individual fiber measured. C: total fat area relative to total fiber area measured. n = 3 in both WT and KO groups. *Significant difference between WT and KO groups (P < 0.05).

DISCUSSION

In the present study we demonstrate, both in vitro and in vivo, that skeletal muscle cells produce adiponectin and this expression appears to be associated with type IIA and IID fibers containing elevated levels of IMCL. We also demonstrate for the first time that Ad-KO mice at 6 mo of age display an elevated body mass, an accumulation of IMCL, an increase in type IIB fiber size, and a reduction in peak contractile force of muscle compared with WT littermates. Knowing that Ad expression is inversely correlated with adiposity and that obese patients display a decrease in circulating Ad levels (1, 7), it is noteworthy that the changes in muscle phenotype in the absence of adiponectin in the present study are in agreement with studies on obese human skeletal muscle (16).

Previous studies have demonstrated that the presence of Ad prevents the accumulation of lipid within cells. For instance, Ouchi et al. (18) found that exposure to physiological concentrations of recombinant human Ad suppressed the accumulation of lipid droplets in human monocyte-derived macrophages in vitro. Furthermore, skeletal muscle triglyceride content in obese and lipoatrophic mice was found to decrease modestly with supplementation of either full-length or globular Ad (29). Consistent with these findings, the present study demonstrates an increase of lipid content directly within the myofiber in Ad-KO mice. Whether this increase in IMCL is the result of depressed β-oxidation resultant from the absence of adiponectin (28, 29, 31) or is due to elevated transport of free fatty acids into the muscle fiber has yet to be fully elucidated.

It was originally proposed that Ad was expressed and secreted solely by adipose tissue. However, recent studies indicate that Ad expression is detectable within cardiomyocytes and skeletal muscle tissue (3, 4, 11, 30). Studies by Yang et al. (30) and Delaigle et al. (4) using IHC techniques to detect Ad in skeletal muscle tissue sections suggested that adiponectin was localized to the endomysium, epimysium, and perimysium but not within the myofiber. In the current study, Ad localization within the myofiber was verified by costaining with the plasma membrane protein, dystrophin. Our findings, unlike those of others (4, 30), illustrate that Ad can be detected within the myofiber proper as well as the surrounding endomysium, likely the result of circulating Ad. The expression of Ad within the muscle fiber was nonuniform, with the highest levels detectable near the plasma membrane. This is consistent with Ad being transcribed and translated within the muscle cell and then secreted into the interstitium. This conclusion is supported by our results showing that L6 myotubes, in vitro, are capable of expressing Ad mRNA and protein.

While intramyocellular Ad expression was present in all muscle fibers examined, there appeared to be a fiber-type distribution to its expression (see Fig. 1D). To further assess this, we undertook serial sections of WT mouse GPS and stained for fiber type, IMCL, and adiponectin. Analysis was then undertaken to determine the IMCL density and intramyocellular Ad expression intensity for each fiber type. It was determined that intramyocellular Ad expression was distinctly elevated in the fast-twitch oxidative type IIA and IID fibers, whereas type I and IIB fibers, which are slow-twitch oxidative and fast-twitch glycolytic, respectively, rarely demonstrated this elevated intramyocellular Ad expression. Interestingly, the presence of elevated intramyocellular Ad expression was correlated with the IMCL density (R² = 0.28; P < 0.0001). To more clearly demonstrate this phenomenon, all data were plotted, and 50th percentile quadrants are displayed (Fig. 2B). This plot clearly illustrates that most fibers demonstrating an elevated intramyocellular Ad also display an increase in IMCL density (>73%). What was very surprising was the finding that of the fibers demonstrating an elevated intramyocellular Ad and IMCL, >94% were identified as type IIA or type IID. The reasons underlying this observation are unclear, especially in light of our findings of increased IMCL in the Ad-KO muscle. It may be that muscle fibers increase their own Ad expression in an autocrine response to increased lipid deposition. An increase in local Ad expression may be an important mechanism for increasing β-oxidation within the fiber to limit IMCL levels. The hypothesis that skeletal muscle Ad expression is increased by IMCL content is supported by the finding that endurance exercise training is associated with both elevations in IMCL levels and Ad receptor expression along with increases in skeletal muscle β-oxidation (23). Although there is currently no evidence to suggest this, we must concede that it is also possible that the higher levels of intramyocellular Ad in the aforementioned fibers may represent circulating Ad being internalized upon binding to its receptor and not the production of Ad by the muscle per se. Regardless, the reasons underlying this potential Ad-IMCL relationship finding clearly demand further investigation.

Muscle function was assessed using in situ electrical stimulation of the mouse lower limb skeletal muscle. Peak tetanic force was ∼50% lower in Ad-KO mice without any change to the rate of fatigue at low-frequency stimulation. To explain this force decrement, we investigated relevant indices of muscle phenotype such as fiber type composition and muscle capillarity. However, none of these variables were different between Ad-KO and WT mice. A recent report demonstrated that skeletal muscle protein degradation was increased in a low-Ad mouse model fed a high-fat diet and that protein degradation in C2C12 cells was rescued with Ad administration (32). Suppression of proteolysis was mediated by an Ad-induced increase in IRS-1 and Akt activity, which inhibited E3 ubiquitin ligase (32). In the current study, while muscle mass did not change, an increase in IMCL content was detected. It is conceivable that in the absence of adequate Ad signaling there was an increase in muscle proteolysis, coupled with a suppression of β-oxidation or increased lipid uptake, resulting in similar muscle masses between Ad-KO and WT mice. Taken together, this would result in a reduction in muscle contractile protein content and, therefore, contractility, but would likely have little effect on short-duration, low-frequency fatigue rates, as assessed in the present study. Future studies may consider using a more demanding fatigue protocol to determine if these muscles are more susceptible to high-frequency fatigue.

The increase in overall type IIB fiber size that was observed is consistent with the findings in obese human skeletal muscle (16). The fact that the Ad-KO muscle generated lower peak tetanic forces may be explained by the greater IMCL content in Ad-KO muscle increasing the volume within each muscle fiber.

In summary, the present study is the first to demonstrate the presence of Ad within skeletal muscle fibers and that this expression is correlated with IMCL content. We further demonstrate the functional consequences of Ad disruption in vivo on skeletal muscle. We demonstrate that Ad-KO mice accumulate IMCL above levels observed in WT littermates and demonstrate impaired skeletal muscle peak force production. Other measures of muscle phenotype, such as fiber type and capillarity, did not significantly change in the absence of Ad; however, type IIB fibers were significantly larger in area compared with control. These novel findings further emphasize the role that Ad plays in metabolic control within skeletal muscle and, in particular, its role in maintaining healthy muscle function. The consistency of our results with those observed in obese skeletal muscle reinforce the view that skeletal muscle defects occurring because of the lack of adiponectin may be an important contributing factor in disease pathophysiology of obesity and type 2 diabetes.

GRANTS

The authors thank the following agencies for support: Sick Kids Foundation (to T. J. Hawke), Natural Science and Engineering Research Council of Canada (to T. J. Hawke and M. C. Riddell), Canadian Institutes of Health Research (to T. J. Hawke and G. Sweeney), Canadian Foundations for Innovation/Ontario Innovation Trust (to T. J. Hawke and M. C. Riddell), and National Institutes of Health (to L. Chan). V. Vu is the recipient of a Doctoral Student Research Award from the Canadian Diabetes Association.