Adult skeletal muscle undergoes adaptation in response to endurance exercise, including fast-to-slow fiber type transformation and enhanced angiogenesis. The purpose of this study was to determine the temporal and spatial changes in fiber type composition and capillary density in a mouse model of endurance training. Long-term voluntary running (4 wk) in C57BL/6 mice resulted in an approximately twofold increase in capillary density and capillary-to-fiber ratio in plantaris muscle as measured by indirect immunofluorescence with an antibody against the endothelial cell marker CD31 (466 ± 16 capillaries/mm2 and 0.95 ± 0.04 capillaries/fiber in sedentary control mice vs. 909 ± 55 capillaries/mm2 and 1.70 ± 0.04 capillaries/fiber in trained mice, respectively; P < 0.001). A significant increase in capillary-to-fiber ratio was present at day 7 with increased concentration of vascular endothelial growth factor (VEGF) in the muscle, before a significant increase in percentage of type IIa myofibers, suggesting that exercise-induced angiogenesis occurs first, followed by fiber type transformation. Further analysis with simultaneous staining of endothelial cells and isoforms of myosin heavy chains (MHCs) showed that the increase in capillary contact manifested transiently in type IIb + IId/x fibers at the time (day 7) of significant increase in total capillary density. These findings suggest that endurance training induces angiogenesis in a subpopulation of type IIb + IId/x fibers before switching to type IIa fibers.
- capillary density
- endothelial cells
- fiber type transformation
- vascular endothelial growth factor
adult mammalian skeletal muscle is composed of multiple fiber types with diverse functional capabilities. Tonically active muscles have mainly slow-twitch fibers with an oxidative phenotype, whereas phasically active muscles have a predominance of fast-twitch fibers with a glycolytic phenotype. Slow-twitch oxidative skeletal myofibers have greater capillary density than fast-twitch glycolytic myofibers (5, 15, 28, 38). One of the unique features of mammalian skeletal muscle is its remarkable ability to adapt to altered functional demands by changing these phenotypic profiles (8, 46). Various stimuli including innervation/neuromuscular activity, mechanical loading/unloading, physical activity, or even hormonal levels and aging can trigger alterations in various functional components, including contractile protein elements, proteins of oxidative phosphorylation, as well as the vasculature (22, 26, 27, 29, 41, 45). More recently, genetic engineering-mediated manipulation of expression of functional proteins has also been shown to induce phenotypic adaptation in skeletal muscle (23), providing mechanistic clues to the tightly regulated physiological structure in and around myofibers as an integrative unit for function.
The functional adaptability of muscle fiber type and capillarity has been most thoroughly explored in animal models of chronic motor nerve stimulation, in which a dramatic increase in capillary formation and an impressive fiber type transition have been reported (11, 17, 29, 37). After 4 wk of continuous electrical stimulation of the rabbit tibialis anterior muscle via the motor nerve, the number of capillaries per unit of cross-sectional area doubles (11, 17), and this predominantly fast-twitch muscle (only ∼6% slow-twitch type I fibers before stimulation) is transformed into a muscle containing almost entirely slow-twitch oxidative fibers (37). As a result, the previously fatigue-susceptible muscle becomes highly fatigue-resistant. The adaptive process begins early, with a significant increase in capillary density after only 2–4 days of stimulation (11, 16, 29). A reduction in fast-twitch glycolytic fibers (type IIb) and an accumulation of fast-twitch oxidative fibers (type IIa) appear to occur simultaneously, or shortly thereafter (between 2 and 12 days), whereas slow-twitch oxidative fibers do not become a major component of the tibialis anterior muscle until day 21 of stimulation (12, 29).
In humans and various animal models, endurance exercise, a more physiological model of increased muscle activity, also induces fast-to-slow fiber type switching and enhanced angiogenesis (25, 30–32, 39, 40, 44). In contrast to chronic motor nerve stimulation, less is known about the time course of muscular adaptations that occur during endurance exercise. Several studies have reported increased capillarity in skeletal muscle after long-term exercise training (3, 25, 30, 31, 33, 40). Other studies have described fiber type transformation after exercise (primarily type IIb to IIa), although the fiber type switching is less robust than the changes seen with chronic motor nerve stimulation (4, 14, 18, 47). It is possible that the relationship between the adaptations induced by endurance exercise is considerably different from that induced by chronic motor nerve stimulation. Understanding the temporal relationship between exercise-induced angiogenesis and other adaptive processes may provide valuable information regarding the functional dependence of one process on the other. If exercise-induced angiogenesis always precedes adaptation in the myofiber, i.e., mitochondrial biogenesis and switching of myosin heavy chain (MHC) isoform expression, it may suggest that angiogenesis plays a permissive role for, or a functional role in, promoting the later events.
A feature of exercise-induced angiogenesis that has recently been appreciated is the heterogeneity of the response among muscles of different fiber types to physiological stimuli (13, 15, 25, 40, 48). The existing evidence is consistent with the notion that fast-twitch skeletal muscle displays phenotypic changes more readily than slow-twitch muscle (48). The apparent determining factor is the muscle phenotypic profile, but other factors such as type, intensity, and duration of the exercise regime may also contribute to the differential responsiveness (20). Because most of the previous comparisons were between different groups of muscles or different regions of the same muscle, it is impossible to know whether the differences are due to a difference in neuromuscular recruitment or a difference in fiber type composition. Thus it is desirable to define the kinetics and the interrelationship of the myofiber adaptation within individual muscles between different subtypes of myofibers.
Angiogenic growth factors such as vascular endothelial growth factor (VEGF) may play important roles in skeletal muscle adaptation after increased contractile activity. Fiber type-specific differential expression of angiogenic growth factor expression has been investigated. For example, mRNA expression of multiple angiogenic growth factors and related receptors demonstrate fiber type specificity (30). Of particular interest is the finding that treadmill running induces significant increases in VEGF mRNA and protein, with the increased VEGF mRNA being specific to type IIb myofibers (7). However, it is not known whether the fiber type-specific changes in the mRNAs of angiogenic factors lead to a fiber type-specific increase in capillary density.
The purpose of this study was to advance our understanding of the temporal and spatial features of endurance exercise-induced angiogenesis and fiber type switching in skeletal muscle. We used a mouse model of endurance training because recent developments in mouse genetics and molecular biology provide useful tools to elucidate cellular and molecular mechanisms of skeletal muscle adaptation.
MATERIALS AND METHODS
Male C57BL/6J mice (8 wk of age) were obtained commercially (The Jackson Laboratory), housed in temperature-controlled quarters (21°C) with a 12:12-h light-dark cycle, and provided with water and chow (Purina) ad libitum. For endurance training, the mice were housed individually in cages (13 × 13 × 30 cm) equipped with running wheels (11 cm in diameter) and were allowed to run voluntarily. The voluntary running cages were modified from a previous model (1) by replacing the mechanical switch on the running wheel with a magnetic reed switch (Radio Shack) connected to the Dataquest Acquisition and Analysis System (Data Sciences International). The mice ran for 1, 3, 7, 14, or 28 days (n = 6 for each time point). A group of mice housed individually in running cages with locked running wheels comprised a sedentary group (n = 6). Separate groups of mice that either were sedentary or voluntarily ran for 3 or 7 days (n = 5 for each time point) were used for quantification of VEGF protein. All mice were killed, and the plantaris muscles were harvested for tissue sectioning 24 h after the last bout of running. All procedures involving these animals conformed to the guidelines for use of laboratory animals published by the U.S. Department of Health and Human Services and were approved by the Duke University Animal Use Committee.
Immunohistochemistry techniques were used for fiber type determination and capillary density analysis. Briefly, plantaris muscles were harvested and saturated in 30% sucrose-phosphate-buffered saline solution (PBS) for ∼2 h, placed in optimal cutting temperature (OCT) tissue-freezing medium (Miles Pharmaceuticals, West Haven, CT), and frozen in liquid nitrogen-cooled isopentane. Frozen sections (6 μm) were cut in a cryostat on microscope slides (Superfrost Plus; Fisher Scientific, Pittsburgh, PA). Slides were allowed to come to room temperature, fixed in 4% paraformaldehyde-PBS for 10 min at 4°C, and permeabilized with 0.3% Triton X-100-PBS for 10 min at 4°C. Blocking solution [5% normal goat serum (NGS)-PBS] was applied for 30 min at room temperature, followed by incubation with MHC I antibody (BF-F8) diluted 1:100 in 5% NGS-PBS at 4°C overnight. Three consecutive washes with PBS for 5 min each were followed by sequential incubation with cyanine Cy5-conjugated goat anti-mouse IgG secondary antibody (1:50) at room temperature for 30 min. The muscle sections were then washed three times with PBS, fixed in 4% paraformaldehyde for 2 min at 4°C, and blocked with 5% NGS-PBS for 30 min. Sequential staining, as described above, was then performed with MHC IIa antibody (SC-71, 1:100) followed by rhodamine red-X-conjugated goat anti-mouse IgG and antibody against endothelial cells (rat anti-mouse monoclonal CD-31, 1:25; Serotec) followed by fluorescein-conjugated goat anti-rat IgG. Images were captured under the confocal microscope (Olympus).
After immunohistochemical staining, the percentage of type I fibers (cyanine Cy5 stained), type IIa fibers (rhodamine red-X stained), and type IIb + IId/x (unstained) fibers was determined. The entire plantaris muscle cross section was analyzed at ×20 magnification, with care taken to ensure comparable cross section locations. Similarly, the capillary density was determined by counting the total number of capillaries (fluorescein stained) on the entire muscle cross section at ×20 magnification, and results are expressed as capillaries per unit area of muscle tissue (a hemocytometer was used to standardize area measurements). To further ensure that the analysis of capillary density was not subject to error from muscle atrophy or interstitial edema, capillary density was also determined by dividing the number of capillaries by the number of muscle fibers to yield the capillary-to-fiber ratio (4). To determine capillary contacting, the number of capillaries in contact with a muscle fiber, capillaries around each muscle fiber type (I, IIa, IIb + IId/x) were counted (total of >600 fibers were counted for each type to calculate the means). This was done on sedentary and 3-day, 7-day, and 28-day running samples. Type IIa fiber cross-sectional areas were measured by using Scion Image software (n > 20 for each section). The staining, photography, and image analysis were performed by a single operator with no knowledge of the coding system.
Endothelial alkaline phosphatase staining.
Capillary density was also assessed by endogenous endothelial alkaline phosphatase staining on frozen sections as previously described (42). Briefly, slides were fixed in 4% paraformaldehyde and incubated with nitro blue tetrazolium (NBT)-5-bromo-4-chloro-3-indolyl phosphate p-toluidine salt (BCIP) (GIBCO-BRL; Grand Island, NY) for 1 h, followed by eosin staining. Capillaries appear dark blue against a red background. Capillary density was measured by counting six random high-power (×40 magnification) fields or a minimum of 200 fibers on an inverted light microscope. Photographs were taken with an Optometrics analog camera and Adobe Premier Version 5.1, and these images were analyzed with an NIH Image analysis system. For the reasons described above, the capillary density was again expressed both as capillaries per unit area and capillaries per muscle fiber. As was done for the immunofluorescence, the staining, photography, and image analysis were performed by a single operator with no knowledge of the coding system.
Quantification of VEGF protein.
Plantaris muscles were harvested from each mouse. For protein extraction, muscle samples were weighed (14–16 mg) and then homogenized with the use of a Pro 200 model polytron (Monroe, CT) in 400 μl of homogenization buffer containing 10 mM Tris (pH 7.4) and 100 mM NaCl. The suspension was then centrifuged twice at 8,000 g at 4°C for 15 min. The protein content of the supernatant was determined by Bradford assay (9). Total VEGF concentration in the plantaris muscle supernatant was determined with a solid-state enzyme-linked immunosorbent assay (ELISA) system with a mouse Quantikine VEGF ELISA Kit (R & D Systems, Minneapolis, MN), as previously described (21). Assay sensitivity is <3.0 pg/ml. Results are expressed in picograms per milligram of muscle based on standard recombinant protein curves. All techniques and materials used in these experiments were in accordance with the provided protocols.
All data are presented as means ± SE. Comparisons were performed between the sedentary and endurance-trained mice by unpaired Student's t-test and among different time points during voluntary running by ANOVA followed by the Tukey test. P < 0.05 was accepted as statistically significant.
Long-term voluntary running induces significant increase in capillary density in plantaris muscle.
We first sought to determine whether endurance exercise results in enhanced angiogenesis in active skeletal muscle. The average daily running distance for the training group was 10.2 ± 0.7 km. This long-term voluntary running resulted in significant increases in indexes of capillary density in plantaris muscle as determined by indirect immunofluorescence with an antibody specific for the endothelial marker CD31 (Fig. 1A and Table 1). The capillary density increased from 466 ± 16 capillaries/mm2 in sedentary mice to 909 ± 55 capillaries/mm2 in trained mice (P < 0.001). The capillary-to-fiber ratio increased from 0.95 ± 0.04 in sedentary mice to 1.70 ± 0.04 in trained mice (P < 0.001). Similar results (1.16 ± 0.08 capillaries/fiber in sedentary mice vs. 1.79 ± 0.08 capillaries/fiber in trained mice; P < 0.001) were obtained when the endogenous alkaline phosphatase staining assay was used (not shown).
Voluntary running induces significant increase in percentage of type IIa myofibers in plantaris muscle.
To assess changes in fiber type composition in plantaris muscle, we performed indirect immunofluorescence using antibodies specific for two MHC isoforms (I and IIa). The percentage of type IIa myofibers in plantaris muscle increased significantly from 15.3 ± 1.1% in the sedentary mice to 31.7 ± 1.9% in the trained mice (P < 0.001; Fig. 1B and Table 1). Conversely, the percentage of type IIb + IId/x myofibers decreased significantly from 83.0 ± 1.0% in the sedentary mice to 68.5 ± 1.8% in the trained mice (P < 0.001). The percentage of type I myofibers was low in plantaris muscle and remained unchanged (1.7 ± 0.2% in sedentary mice vs. 1.8 ± 0.2% in trained mice, P > 0.05; Table 1). There were also no significant differences in the average cross-sectional area of type IIa fibers between sedentary control and trained mice (1,059.7 ± 116.4 and 1,053.1 ± 122.6 μm2, respectively; P > 0.05).
Increased capillary density occurs before fiber type switching.
We performed analyses for capillary density and fiber type composition in plantaris muscle during a time course (1, 3, 7, 14, and 28 days) of voluntary running to gain a better understanding of the temporal changes of capillary density and fiber type composition. The average running distance increased steadily and reached a plateau around day 8 (Fig. 2A). The increase in capillary-to-fiber ratio became statistically significant at day 7, where the number of capillaries per fiber increased from 1.01 ± 0.04 at day 3 to 1.44 ± 0.08 at day 7 (P < 0.001), suggesting that active angiogenesis occurred between day 3 and day 7 (Fig. 2B). On the other hand, a significant change in fiber type occurred after 14 days of running, where the percentage of type IIa fibers increased from 21.2 ± 1.2% at day 7 to 32.7 ± 1.0% at day 14 (P < 0.001), suggesting that the period between these time points is when the majority of fiber type switching occurred (Fig. 2C).
Increased VEGF protein expression precedes increased capillary density.
Because the increase in capillary density became significant after day 7, we sought to determine whether the enhanced angiogenesis was associated with an increase of VEGF protein in the muscle. VEGF protein concentration in plantaris muscle was 6.3 ± 1.3, 11.5 ± 2.0, and 10.3 ± 1.4 pg/mg soluble muscle protein in the sedentary and 3-day and 7-day trained mice, respectively (P < 0.05 for both day 3 and day 7 mice vs. sedentary mice; Fig. 3). Because capillary density did not change significantly until day 7, the increase in VEGF protein concentration in plantaris muscle preceded the increase in capillary density.
Transient increase in capillary contacting around type IIb + IId/x myofibers precedes fiber type switching.
To define the spatial feature of the enhanced angiogenesis, we performed triple-color indirect immunofluorescence for MHC I, MHC IIa, and CD31. The unstained muscle fibers were considered type IIb + IId/x. Fiber type-specific capillary contacting was determined by counting the capillaries around individual fibers of different types (Fig. 4A). Capillary contacts per fiber for type IIb + IId/x fibers increased from 1.59 ± 0.09 in sedentary mice to 2.06 ± 0.11 (P < 0.01) at day 7 of running, a time point before the majority of fiber type conversion, and decreased toward the sedentary control level (1.82 ± 0.08 capillaries/fiber; P > 0.05 vs. sedentary mice) at 28 days of running, a time point following the majority of fiber type conversion (Fig. 4B). In contrast, capillary contacts per fiber for type IIa fibers remained unchanged at day 7 (3.92 ± 0.12 capillaries/fiber in sedentary mice vs. 4.03 ± 0.10 capillaries/fiber at day 7; P > 0.05) and increased at day 28 (4.35 ± 0.10 capillaries/fiber; P < 0.05 vs. sedentary) (Fig. 4C). Thus increased capillary density in plantaris muscle during long-term voluntary running could be attributed to an early (around day 7) increase in capillary density in type IIb + IId/x fibers followed by a later (after 7 days) combinatory increase of capillary density in type IIa fibers and percentage of type IIa fibers.
Mammalian skeletal muscle has remarkable plasticity to respond to both physiological and pathological stimuli. This adaptability is achieved primarily through changes in contractile protein composition, oxidative and glycolytic metabolism, and vascularity, which in turn directly impact the exercise capacity of the muscle (22, 26, 27, 29, 41, 45). Exercise training is perhaps the most widely employed physiological stimulus to skeletal muscle, but the mechanisms that underlie skeletal muscle adaptation are not completely understood. The major finding of our study is that exercise-induced changes in skeletal muscle vascularity occur before a significant alteration in fiber type composition and that changes in VEGF protein occur before changes in capillary density. Moreover, we demonstrate fiber type specificity to the angiogenic response at the cellular level in skeletal muscle, with capillary density increasing specifically around type IIb + IId/x fibers before switching to type IIa fibers, a finding that has not been previously described.
Chronic motor nerve stimulation is an “extreme” model of increased contractile activity and perhaps more accurately represents a model of fiber type transformation. Studies using chronic motor nerve stimulation have clearly delineated the changes in capillary density and fiber type composition (11, 17, 29, 37). It is not known whether the temporal relationship between enhanced angiogenesis and fiber type switching is an inherently programmed event for skeletal muscle in response to increased contractile activity. In this study, we used a voluntary running model to examine the temporal and spatial changes in peripheral skeletal muscle induced by endurance exercise in mice. Consistent with a previous report in other species (30), capillary density in active skeletal muscle in mice showed a sigmoid increase during long-term endurance exercise (Fig. 2B). As has been reported for chronic motor nerve stimulation (29), this enhanced angiogenesis occurs before changes in contractile protein composition. Therefore, the sequential changes in vascularity and fiber type composition are recapitulated in the mouse model of endurance exercise.
The apparent temporal order of the changes in capillary density and fiber type composition in plantaris muscle during exercise training may be due to the differences in the inherent kinetics of the proteins and the cellular components involved. Contractile proteins, such as MHC, have a relatively long half-life (34) and, with the techniques used in this study, it may take a significantly longer time to show detectable changes in MHC than in endothelial cells. Functionally, it is possible that increased capillary density plays a permissive role for the ultimate phenotypic changes in the myofibers. However, these findings and the proposed scenario do not dispute the notion that an orchestrated signaling cascade through hard-wired circuitry built in the skeletal muscle at the onset of exercise training is directly involved in the highly coordinated adaptive processes of angiogenesis and fiber type switching.
This study is the first to demonstrate that VEGF protein in active skeletal muscle is increased in response to endurance exercise in mice. We show that VEGF protein remains elevated for at least 24 h after cessation of running when mice are allowed to run voluntarily for 3 or 7 days (Fig. 2D). This temporal pattern of VEGF protein expression is consistent with a long-standing hypothesis that induced VEGF mediates vasculature remodeling in response to endurance exercise. VEGF, among many angiogenic factors that are upregulated in contracting skeletal muscles and have potential functional roles in inducing angiogenesis (19, 24, 30, 35, 36), is most likely to contribute directly to exercise-induced angiogenesis for the following reasons. First, various types of muscle contraction, including voluntary running, induce a transient increase in VEGF mRNA (10, 24) and protein in skeletal muscle (2, 7, 13). Second, pharmacological inhibition of VEGF blocks exercise-induced VEGF protein expression and angiogenesis in rat skeletal muscle, which can be recapitulated by VEGF-neutralizing antibody (2). Finally, conditional deletion of VEGF in mouse skeletal muscle results in capillary regression (43). It remains to be determined unambiguously whether enhanced VEGF expression in response to exercise is required for enhanced angiogenesis in skeletal muscle.
Our data also demonstrate another level of complexity in the angiogenic response. Specifically, we noted increased capillary contact in type IIb + IId/x fibers during the active phase of muscle adaptation and in type IIa fibers later when a steady state of fiber type composition was reached. These findings suggest that endurance exercise induces angiogenesis in type IIb + IId/x fibers before switching to type IIa fibers. On the basis of this information, we can hypothesize that coordinated skeletal muscle adaptation occurs in response to endurance exercise with an initial enhanced expression of angiogenic factors to mediate angiogenesis, which may permit or promote the later adaptive processes, such as fiber type switching.
The average capillary contacts in type IIb + IId/x increased moderately from 1.59 ± 0.09 capillary contacts/fiber in sedentary mice to 2.06 ± 0.11 capillary contacts/fiber (P < 0.01) at day 7 of running, a time when there was no discernible change in fiber type composition. However, this seemingly moderate change is consistent with a complete transformation of the capillary-to-fiber ratio to one equivalent to that of type IIa fibers after a careful calculation. Overall, 16.4% of the total myofibers switched from type IIb + IId/x (83%) to type IIa. If the increased capillary contact occurs in these 16.4% of fibers, there must be an increase of an average of 2.38 capillary contacts/fiber, which brings the actual capillary contacts to 3.97 capillary contacts/fiber, a value that is almost the same as in type IIa fibers (3.92 capillary contacts/fiber). Therefore, it is possible that an increase of capillary density in a subpopulation of type IIb + IId/x fibers occurs before fiber type switching. The schematic diagram in Fig. 5 depicts a proposed spatial and sequential order of changes in capillary density and fiber type composition in skeletal muscle during endurance training.
It is worth noting that the cross-sectional area of type IIa fibers remains unchanged after long-term voluntary running whereas there is a significant increase in the percentage of type IIa fibers (Table 1 and Fig. 2C). Technically, it was not feasible in the present study to distinguish the fibers that had changed from type IIb + IId/x from the type IIa fibers that were present before the training. If these fibers were truly identical in the cross-sectional area, it would suggest that a process of decreases in cross-sectional area occurs during fiber type transformation from type IIb + IId/x to type IIa because the fast-twitch glycolytic fibers are larger than fast-twitch oxidative fibers in rodents (6). This is also depicted in Fig. 5.
The delayed increase in capillary contact in type IIa fibers, along with the increased percentage of type IIa fibers, may collectively contribute to increased total capillary density observed in trained skeletal muscle. Previous observation of enhanced angiogenesis in predominantly slow-twitch soleus muscle (25) is consistent with the notion that capillary density in slow-twitch oxidative and fast-twitch oxidative fibers may increase further after long-term endurance training.
Because modulating exercise capacity has the potential to impact an enormous number of clinical conditions, therapeutic manipulations that are intended to alter the response to exercise would benefit from an advanced understanding of the temporal and spatial changes that occur in muscle in response to exercise. Although the mechanisms responsible for the fiber type specificity of the angiogenic responses are currently unknown and warrant further investigation, there are important implications of these findings. First, the differential adaptive response within a muscle among different fiber types has improved our understanding of the cellular mechanisms of heterogeneity in adaptability between glycolytic muscle and oxidative muscle. Second, our data suggest that angiogenesis is one of the early steps in skeletal muscle adaptation. Data from our investigation will also be critical for the refinement of studies that can be performed to inhibit the angiogenic response to exercise to directly test the mechanistic link between angiogenesis and skeletal muscle plasticity.
This work was supported by American Heart Association (AHA) Scientist Development Grant 0130261N (to Z. Yan) and in part by AHA Established Investigator Grant EI0140126N and a Department of Veterans Affairs Merit Review Grant (to B. H. Annex).
We are appreciative of the excellent technical support from C. Z. Ireland, A. M. Pippen, and M. Zhang. We are thankful to Dr. H. Tibbals at the University of Texas Southwestern Medical Center for assisting in the modification of the mouse running cage.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society