Exercise-Induced Expression of Vascular Endothelial Growth Factor mRNA in Rat Skeletal Muscle is Dependent on Fibre Type

Abstract

In this study, we quantified the expression of the vascular endothelial growth factor (VEGF) gene in individual muscle fibres at the end of a single 90 min run of 20−25 m min⁻¹, at 10 % incline. In addition, we evaluated the co-ordinated expression of several hypoxia-sensitive genes, including the ORP-150 gene. Individual fibres were taken from rat plantaris muscle, either at the end of a single bout of exercise or at rest, and classified as Type I, IIa, IIx or IIb, according to the expression of myosin heavy chain (MHC) isoforms. VEGF mRNA levels increased by 90 % in exercising whole plantaris in comparison with those in control muscle (P < 0.001), while the VEGF protein content increased by 72 % (P < 0.05). Using real-time PCR analysis, an accurate and reproducible method for quantification of mRNA levels, a marked rise in VEGF transcript levels was observed at the end of exercise in individual myofibres (P < 0.05), providing the first direct evidence that VEGF transcripts increase in muscle cells after a single bout of exercise. This exercise-induced increase in VEGF transcript levels was specifically observed in type IIb myofibres, which are predominantly glycolytic and more susceptible to local hypoxia than oxidative myofibres such as type I or IIa fibres (110 %, P < 0.05). Moreover, treadmill exercise increased the expression of two hypoxia-sensitive genes. The levels of mRNA for Flt-1, a VEGF-specific receptor, and those for ORP-150, a chaperone essential for the secretion of mature VEGF, increased in whole plantaris muscles (108 and 92 %, respectively, P < 0.05). Taken together, these findings are consistent with the suggestion that hypoxia could be one of the mechanisms involved in exercise-induced capillary growth.

Endurance training induces marked changes in biochemical and structural properties of skeletal muscles (Booth & Thomasson, 1991), including a significant increase in skeletal muscle capillarisation characterised by an elevated capillary density and capillary to fibre ratio (Andersen, 1975; Brodal et al. 1977; Hudlicka et al. 1992). This skeletal muscle vascular network development results mainly in capillary growth or angiogenesis (Risau, 1997). Increased capillarisation in endurance training is thought to be triggered by local mechanical and metabolic factors which may trigger the release of diffusible factors known as angiogenic growth factors (Folkman & Shin, 1992; Hudlicka et al. 1992). Vascular endothelial growth factor (VEGF) is probably the most studied because of its critical role in the regulation of angiogenesis that occurs during physiological and pathological events such as embryonic development, wound healing, ischaemia or tumour growth (Folkman & Shin, 1992; Ferrara, 2001). VEGF is a 45-46 kDa heparin-binding glycoprotein which acts as a highly mitogenic factor for vascular endothelial cells and a strong vascular permeability factor. Several previous studies examined the specific role of VEGF in training-induced capillary growth and showed that exercise induced increased expression of VEGF mRNA in the skeletal muscle of both humans and animals (Breen et al. 1996; Gustafson & Kraus, 2001). The training-induced increase in the transcription of the VEGF gene is commonly associated with an increase in the VEGF protein content in the skeletal muscle of both humans and animals (Asano et al. 1998; Gustafson et al. 1999; Amaral et al. 2001). This up-regulation of the VEGF gene seemed to be related to the intensity of exercise and was markedly increased in the hypoxic environment (Breen et al. 1996). Reduced oxygen tension within skeletal muscle has thus been suggested as a possible primary stimulus for exercise-induced angiogenesis (Breen et al. 1996; Roberts et al. 1997).

Using in situ hybridisation, VEGF mRNA has been shown in skeletal myofibres at the end of a single bout of exercise (Breen et al. 1996; Brutsaert et al. 2002). In contrast, VEGF immunostaining showed that the protein was only found in the extracellular matrix between the myocytes (Annex et al. 1998). Taken together, these findings suggest that muscle fibres are the source of VEGF. Because in situ hybridisation is not a technique designed to provide a quantitative measure of mRNA signal strength, exercise-induced VEGF mRNA expression in individual skeletal fibres needs to be examined. Moreover, skeletal muscles contain at least four types of fibre ranging from slow-twitch predominantly oxidative fibres (type I) to fast-twitch predominantly oxidative, intermediate oxidative or low oxidative fibres (types IIa, IIx and IIb, respectively). Muscle fibres are distributed amongst motor units and it is well accepted that during muscle contraction, motor units are recruited in an orderly manner, type I fibres being the first recruited, followed by type IIa, type IIx and IIb (Delp & Duan, 1996). In light of the recruitment pattern of muscle fibres during exercise, we sought to investigate whether the exercise-induced increase in VEGF gene expression in muscle fibres was fibre-type specific.

The VEGF gene is known to be up-regulated by hypoxic stress, especially through the activation of the nuclear transcription factor, hypoxia-inducible factor-1 (HIF-1) (Forsythe et al. 1996). During hypoxia, HIF-1 is stabilised, translocated to the nucleus, and able to bind onto hypoxia-responsive element (HRE), a short sequence located in the promoter of several genes, including the VEGF gene. The interaction between HIF-1 and HRE induces an up-regulation of the transcriptional activity of the VEGF gene. The in vivo bioactivity of VEGF requires post-translational processing (Walter et al. 1996), and the optimal function of vascular chaperones is essential for secretion of mature VEGF. Recently, Ozawa et al. (2001) purified and cloned the gene for the oxygen-regulated protein-150 (ORP-150), a hypoxia-inducible endoplasmic reticulum protein which shows a binding affinity for VEGF. ORP-150 acts as a molecular chaperone under hypoxia to facilitate VEGF intracellular transport and processing between the endoplasmic reticulum and the Golgi apparatus (Ozawa et al. 2001). Although the increased expression of VEGF during prolonged exercise is now a well-known fact, the associated changes in the expression of oxygen-regulated chaperones such as ORP-150 has never been examined. The bioactivity of VEGF is also related to the presence of two specific receptors, Flt-1 and Flk-1/KDR (Gerber et al. 1997). An HRE sequence has been identified in the promoter region of the Flt-1 gene, and this gene has been shown to be up-regulated by hypoxia, although this is not the case for the Flk-1 gene. Increased capillarisation in endurance training is thought to be triggered by local mechanical and metabolic factors including hypoxia (Hudlicka et al. 1992). It was thus hypothesised that several hypoxia-regulated genes, including the ORP-150 and Flt-1 genes, were up-regulated in a co-ordinated manner in whole skeletal muscle at the end of a single bout of exercise.

The main objectives of the present study were to: (1) quantify the level of VEGF mRNA expression in individual muscle fibres at the end of prolonged exercise; (2) examine the fibre-specific responses of VEGF expression; and (3) evaluate the co-ordinated expression of two hypoxia-sensitive genes (ORP-150 and Flt-1) within muscle, in response to a single bout of exercise. The results demonstrate for the first time that the skeletal myofibre itself is a true source of production of VEGF mRNA. Using real-time RT-PCR analysis, we showed that VEGF production by single myofibres was fibre-type specific, mainly restricted to fibres expressing type IIb MHC, the fastest MHC isoform. The exercise-induced increase in VEGF mRNA levels was also associated with an increased expression in whole muscle of ORP-150 mRNA, an oxygen-dependent regulator of VEGF protein elaboration.

METHODS

Animals

Female Wistar strain albino rats (weight, 240-260 g; n = 12) were randomly assigned to two experimental groups: (1) active rats (n = 6); and (2) control rats (n = 6). All animals were housed four per cage with a 12-12 h light-dark cycle, at an ambient temperature of 22 ± 2 °C. They had free access to food and water. Experiments were performed under the control of the ethics committee of the Centre de Recherche du Service Santé des Armées.

Experimental design

All rats were first familiarised with a rodent treadmill running activity at a moderate level (10 m min⁻¹, 0 % gradient, 10-15 min) each day for 5 days before the exhaustive exercise session. All animals were then rested for 3 days in their cages. Thereafter, exercising rats ran at 20-25 m min⁻¹, 10 % gradient for 90 min. Mild electric shocks were used sparingly to motivate the animals to run. Exhaustion was defined as the time when animals were unable to keep in pace with the treadmill for up to 1 min, and when they remained on their backs when placed in this position. Active rats were compared to control animals kept at rest on the treadmill for 90 min. Immediately after the end of exercise, animals were anaesthetised with pentobarbital (70 mg (kg body weight)⁻¹, I.P.; Sanofi Santé Animal, Libourne, France), and both plantaris muscles were removed. One was rapidly frozen in liquid nitrogen and stored at −80 °C until required for further analysis of whole muscle tissue. The second was used for fibre dissections and immediately put into cooled Krebs-Henseleit solution containing (mM): NaCl 118, NaHCO₃ 25, KCl 4.7, KH₂PO₄ 1.2, MgSO₄.7H₂O 1.2. Animals were then killed by removal of the heart.

Single fibre dissection

A small bundle of muscle fibres was excised from the whole muscle under a low-power stereomicroscope using sharp-ended tweezers. Single fibres were precisely isolated and separated transversely into two halves. Each half-fibre was then separated again into two quarters: MHC determination was carried out using one half of each half-fibre, whilst the other half was reserved for further mRNA analyses. The same operation was repeated until 40-60 single fibres were isolated from each muscle, which required approximately 4 h.

Analysis of MHC protein

Single fibres were subjected to analysis of MHC isoforms using an SDS-PAGE electrophoretic method adapted from Talmadge & Roy (1993). Myosin was extracted from the two quarter parts of a single fibre in 20 μl of a solution containing (mM): NaCl 300, NaH₂PO₄ 100, Na₂HPO₄ 50, Na₄P₂O₇ 10, MgCl₂.6H₂O 1, EDTA 10 and 2-β-mercaptoethanol 1.4; pH 6.5. After incubation for 24 h at 4 °C, the half-fibre was entirely digested and the 20 μl mixture was then diluted with 20 μl glycerol. Extracts were stored at −20 °C until required for the separation process. Electrophoresis was performed using a Mini Protean II system (Biorad, Marnes-la-Coquette, France). Separating gel solution contained 30 % glycerol, 8 % acrylamide-bis (50 : 1), 0.2 M Tris, 0.1 M glycine and 0.4 % sodium dodecyl sulfate (SDS). Stacking gel was composed of 30 % glycerol, 4 % acrylamide-bis (50 : 1), 70 mM Tris, 4 mM EDTA and 0.4 % SDS. Then 10 μl myofibril samples were denatured using 10 μl buffer containing 5 % 2-β-mercaptoethanol, 100 mM Tris base, 5 % glycerol, 4 % SDS and bromophenol blue, for 3 min at ≈100 °C. Myofibrillar homogenates were loaded onto vertical gels, whilst two lanes were loaded with protein extract from a control plantaris muscle known to contain the four adult MHC isoforms. Gels were run at constant voltage (72 V) for 31 h and then silver stained (Agbulut et al. 1996). The MHC protein isoform bands were scanned using a densitometer system equipped with an integrator (GS-700, Biorad). The MHC isoforms expressed by single fibres were identified by comparing them with bands of myosin extracts from control plantaris muscle.

Western blot analysis

Samples (60-80 mg) from each plantaris were homogenised in an extraction buffer containing (mM): sucrose 300, Hepes 20, sodium azide 1, with 0.5 trypsin inhibitor units (TIU) per sample aprotinin; pH 7.4). Concentrations were determined by Bradford assay and 15 μg total protein was then separated by electrophoresis on a 10 % SDS-PAGE gel. A standardised amount of protein (15 μg) prepared from a plantaris extract was deposited on each gel to act as an internal standard for comparison across blots. The gels were transferred overnight to a nitrocellulose membrane (Hybond C-extra RPN 2020E, Amersham, Orsay, France), which was blocked for 2 h at 4 °C in 5 % non-fat dried milk diluted in Tris-buffered saline (Sigma, St Quentin Fallavier, France). Blots were incubated for 2 h at 4 °C with a primary polyclonal anti-VEGF antibody (sc-507, Santa Cruz, Heidelberg, Germany). Washed blots were then incubated for 1.5 h at 4 °C with a secondary horseradish peroxidase-conjugated antibody (sc-2004, Santa Cruz). Washed blots were subjected to the ECL Western blotting detection reagents kit (RPN 2209, Amersham) to detect the immunoreactive signal by chemiluminescence, and the membranes were then exposed to X-ray film (Hyperfilm ECL RPN 3103K, Amersham) for 4 min. The relative VEGF expression was determined by the VEGF band intensity/internal standard band intensity ratio.

Because of insufficient data in the literature on VEGF protein analysis using Western blotting in rat and especially in muscle, we decided to validate our biochemical method. Molecular mass markers were commonly used to determine the respective molecular masses of the labelled bands and an immunoreactive band was regularly detected at ≈46 kDa. Human recombinant VEGF (rhVEGF, 15 ng; 293-RV-010, R&D Systems Europe, Abington, UK) was also loaded on a gel to ensure the specificity of the immunoreactive signal. As expected, the band corresponding to rhVEGF migrated slightly further than heart protein extracts, probably because rhVEGF was not glycosylated (Fig. 1A). An increased amount of protein extracts was also loaded to verify that antibody-antigen complexes were within the linear response and not saturated (Fig. 1A).

Figure 1. Quantitative densitometry of VEGF protein in whole muscle and individual myofibres.

A, representative set of Western blots of VEGF protein in rat plantaris muscle. VEGF protein was detected after 10 % SDS-PAGE. a, b and c, samples from plantaris muscle corresponding to 35, 25, and 15 μg of skeletal muscle protein, respectively. d, sample from left ventricle of hypoxic rats intended to produce intense immunoreactive signals needed for determination of the standard curves (15 μg of total protein). e, sample of human recombinant VEGF (15 ng). B, representative set of Western blots of VEGF protein in plantaris muscles taken from control and active rats. C, detection of VEGF protein in pooled myofibres. The immunoreactive band was compared with that obtained in whole muscle. D, quantitative densitometry of VEGF protein in whole plantaris muscle. VEGF is quantified by measuring both the optical density and the surface area of the immunoreactive bands. Data are means ± s.e.m. * Significantly different from control non-exercised plantaris muscles, P < 0.05.

Because of an insufficient amount of total proteins in single isolated fibres, VEGF protein expression was not measurable in half-fibres. Consequently, 30 individual fibres were pooled, whatever their phenotype, and were analysed, as previously described for whole muscle, in order to obtain a qualitative detection of VEGF protein (Fig. 1C).

Total RNA isolation

Total RNA was isolated from samples using RNABle reagent (Eurobio, Les Ulis, France), and following the manufacturer's instructions. Whole muscle samples (approximately 15 mg each) and each individual fibre were extracted using 1 ml and 50 μl of RNABle, respectively. Total RNA was precipitated from the aqueous phase with isopropanol, then washed with 100 % ethanol, and finally suspended in a small volume of ribonuclease-free water (Eurobio). The volumes of total RNA solution obtained for whole muscle and individual fibre were 50 and 15 μl, respectively. Optical density (OD) measurements were made at 260 and 280 nm in order to determine RNA concentration and purity.

Design and synthesis of primers for PCR

All primers used in this study were synthesised at Eurogentec (Saraing, Belgium) and designed with the MacVector software (Accelrys, Orsay, France). Selected forward (F) and reverse (R) primers for VEGF, Flt-1, Flk-1, HIF-1α, ORP-150, cyclophylin-A (CycA) and platelet endothelial cell adhesion molecule (PECAM) were as follows:

VEGF F: 5′ ATCATGCGGATCAAACCTCACC 3′;

VEGF R: 5′ GGTCTGCATTCACATCTGCTATGC 3′ generating an 80 bp DNA fragment;

Flt-1 F: 5′ CGACACTCTTTTGGCTCCTTCTAAC 3′;

Flt-1 R: 5′ TGACAGGTAGTCCGTCTTTACTTCG 3′ generating an 83 bp DNA fragment;

Flk-1 F: 5′ GTACCAAACCATGCTGGATTGC 3′;

Flk-1 R: 5′ CTTGCAGGAGATTTCCCAAGTG 3′ generating a 92 bp DNA fragment;

HIF-1α F: 5′ ATGACCACTGCTAAGGCATCAGC 3′;

HIF-1α R: 5′ AGGTTAAGGCTCCTTGGATGAGC 3′ generating a 119 bp DNA fragment;

ORP-150 F: 5′ ACGCTCCGTTATTTCCAGCAC 3′;

ORP-150 R: 5′ TGCCTCTGTGGGTCAACATTG 3′ generating a 110 bp DNA fragment;

PECAM F: 5′ TTTCAGCAAGATTGCCGAGGAGAGG 3′;

PECAM R: 5′ TTTGGAGAGCATTTCGCACACCTGG 3′ generating a 118 bp DNA fragment;

cycA F: 5′ AGCATGTGGTCTTTGGGAAGGTG 3′;

cycA R: 5′ CTTCTTGCTGGTCTTGCCATTCC 3′ generating a 92 bp DNA fragment.

Reverse transcription reaction

Reverse transcription (RT) was carried out in a 40μl final volume with 4 μg total RNA for whole muscle and 15 μl RNA solution for half-fibres. RT was performed using 300 U M-MLV reverse transcriptase (Promega, Madison, WI, USA) according to the manufacturer's instructions, with 30 u recombinant RNasin ribonuclease inhibitor (Promega, Charbonnières, France) and 10 ng Oligo(dT)₁₅ Primer (Promega).

Real-time PCR

The PCR reactions were carried out in a 20μl final volume with the LC Fast Start DNA Master SYBR Green kit (Roche Diagnostics, Meylan, France) using 4 mM MgCl₂, 0.4 μM of each primer (final concentration) and 0.01 μl or 2 μl of cDNA for whole muscle and half-fibre, respectively. Quantitative PCRs were performed using LightCycler (Roche Diagnostics) for 50 cycles of 95 °C for 20 s, 65 °C for 5 s and 72 °C for 8 s. Results were analysed with RelQuant software (Roche Diagnostics). The expression of the CycA gene, a housekeeping gene, was measured and used as an internal standard to normalise the mRNA levels of target genes. The CycA gene was used as an internal control because CycA mRNA levels were unaffected by ageing or exercise (Alway et al. 2002; Murphy et al. 2003). Moreover, we previously checked that the expression of the CycA gene was not tissue specific. Furthermore, we confirmed that the VEGF mRNA levels were unaffected by the 4 h isolation procedure (Table 1).

Table 1.

Quantitative analysis of mRNA signals for VEGF in individuals fibers from plantairs muscles of exercised (active, n = 6) and resting rats (control, n = 6), according to the time of myofibre isolation

	Beginning of the dissection period	Middle of the dissection period	End of the dissection period
Control rats	0.21 ± 0.06	0.30 ± 0.04	0.22 ± 0.12
Active rats	0.43 ± 0.12	0.54 ± 0.21	0.30 ± 0.08

Values are means ± s.e.m. mRNA levels are expressed as fluorescence (μg RNA)⁻¹ (mg wet weight)⁻¹.

Purity of single fibres

Before quantitative analysis of mRNA expression, we needed to check the purity of all half-fibres and confirm the absence of endothelial cell material. PECAM is known to be expressed by endothelial cells and was chosen as a negative marker. After a PCR amplification step, melting curves were determined for CycA and PECAM using RelQuant software. In case of a positive signal for PECAM, the sample was excluded from the study. To be considered pure and valid, myofibres needed to be both positive for CycA and negative for PECAM mRNA signals.

Standard curves and resulting expression for whole muscle

Measurement of amplified mRNA uses the crossing point (CP), that is, the cycle number of amplification at which the signal is detected above the background and is in the exponential phase. The CPs were between 21 and 35 cycles of amplification for VEGF, 27 and 38 for Flt-1, 24 and 37 for Flk-1, 23 and 37 for ORP-150, 21 and 35 for HIF-1α, and 17 and 33 for CycA.

All amplified cDNAs for our genes of interest were quantified and normalised against a housekeeping gene, as previously described by Giulietti et al. (2001). VEGF, Flt-1, Flk-1, HIF-1α and ORP-150 mRNA levels were expressed as relative copy number normalised against CycA mRNA using RelQuant software. This was achieved by constructing a standard curve from serially diluted (1 to 10 000) positive control for VEGF, Flt-1, Flk-1, HIF-1α, ORP-150 and CycA, that is, cDNA obtained from a whole plantaris sample after similar RNA extraction and reverse transcription, as with other whole muscles of the study.

Standard curves and resulting expression of single fibres

The crossing points for VEGF and CycA were between 33 and 39 cycles of amplification. A signal detected around 39 cycles was near the limit of detection of the LightCycler instrument. For these reasons, PCRs were carried out with no dilution of the material from single fibres. One difficulty was to determine the standard curve because it was essential to obtain a positive control matrix with the same chemical environment as that of half-fibre cDNA. To meet this objective, we used a cDNA solution obtained by pooling an equal volume of each half-fibre. Because half-fibre cDNAs were not diluted for PCR, the positive control pool was specifically enriched with a VEGF cDNA fragment (Genbank AF215725: nt 40-392) to begin the standard curve at least three cycles before the mean crossing point of the half-fibre pool alone.

Statistical analyses

All data were presented as means ± s.e.m. Student's t test was used to determine differences between post-exercise VEGF mRNA and protein, ORP-150, HIF-1α and VEGF receptor mRNA levels in whole muscle. A two-way ANOVA (fibre type × exercise) was used to examine the fibre-type specificity of the VEGF mRNA response to exercise. When appropriate, differences between groups were tested with a Newman-Keuls post hoc test. Statistical significance was accepted at P < 0.05.

RESULTS

Exercise-induced changes in VEGF mRNA and protein in whole muscle

As previously reported, an increase in VEGF mRNA expression occurred in skeletal muscle in response to acute exercise. VEGF mRNA levels increased by 90 % in exercising whole plantaris in comparison with control muscle (P < 0.001) (Fig. 2A). VEGF protein was quantified in whole muscle using Western blot analysis (Fig. 1A and B). As previously shown, VEGF was detected as an immunoreactive band at ≈46 kDa (Annex et al. 1998). The VEGF protein content of plantaris muscles from active rats increased by 72 % in comparison with sedentary animals (P < 0.05) (Fig. 1D).

Figure 2. Quantitative analysis of VEGF gene expression by real-time RT-PCR.

VEGF gene expression in whole plantaris muscle (A) and in individual myofibres (B) from control non-exercised (n = 6) and active rats (n = 6). Data are means ± s.e.m. * Significantly different from control values, P < 0.05, ** Significantly different from control values, P < 0.001.

Purity of single fibres

Each myofibre was divided transversely into four parts, two for fibre-type determination, the other two for gene expression analysis. Before VEGF mRNA quantification, each individual fibre was tested for both PECAM and CycA mRNA expression. A negative signal for PECAM mRNA proved the lack of contamination by endothelial cell material, while a positive expression of CycA gene confirmed the presence of biological material after dissection, RNA extraction and reverse transcription. After all these validation analyses, 90 % of single myofibres taken from plantaris muscles of control and active rats were found to be both negative for PECAM and positive for CycA mRNA expression. These pure myofibres were then identified by determining their MHC composition. Hybrid fibres expressing several MHC isoforms were excluded from the study; only those expressing one MHC isoform were selected and distributed between the control (n = 57) or active (n = 55) groups.

Expression of VEGF mRNA and protein within single fibres

A marked rise in VEGF transcript levels was observed at the end of exhaustive treadmill exercise in pure myofibres (P < 0.05) (Fig. 2B). The presence of VEGF protein in pure and pooled isolated fibres was verified. As was the case for whole muscle, an immunoreactive band was detected for pooled half-fibres, whatever their phenotype (Fig. 1C). This finding demonstrates for the first time that in addition to its well-known production by endothelial cells and to its presence in the extracellular matrix, VEGF protein is present within pure myofibres.

Fibre-type specificity of VEGF mRNA expression

Subpopulations of myofibres were identified by determining their MHC composition. A polyacrylamide gel electrophoresis procedure was used to determine the MHC content of each myofibre and to identify types I, IIa, IIx and IIb fibres. Figure 3 shows VEGF mRNA expression in slow and fast fibres of plantaris muscle from control and active animals. Acute treadmill exercise did not significantly change the VEGF mRNA levels in slow (type I) and fast type IIa and type IIx fibres. In contrast, higher levels of VEGF mRNA were found in fast type IIb fibres of exercised animals than in resting animals (110 %, P < 0.05).

Figure 3. Quantitative analysis of VEGF gene expression by real-time RT-PCR in single myofibres from control non-exercised and active rats.

Individual fibres were negative for PECAM mRNA signal and were categorised into four types, type I, IIa, IIx and IIb, according to their MHC isoform content (Delp & Duan, 1996). Data are means ± s.e.m. * Significantly different from control values, P < 0.05.

However, VEGF transcripts were not detected in all selected myofibres and some fibres probably did not express the VEGF gene. Table 2 shows the distribution of VEGF-positive pure myofibres in control and exercised rats. As expected, the percentage of fibres positive for VEGF signal was higher in active than in control animals. Moreover, the percentage of positive myofibres was higher in type IIb fibres than in slow, type IIa or type IIx fibres, in both control and active rats.

Table 2.

Distribution of individual myofibres positive for VEGF mRNA according to their fibre type inplantaris muscles of control (n = 6) and active rats (n = 6)

	Fibre type	Total number of fibres examinated	Fibres positive for VEGF
Control	type I	10	60%
	type IIa	11	64%
	type IIx	26	54%
	type IIb	40	76%
	all types	87	65%
Active	type I	7	71%
	type IIa	9	78%
	type IIx	24	50%
	type IIb	33	94%
	all types	73	75%

Individual myofibres were positive for cyclophilin-A, negative for platelet for platelet endothelial cell adhesion molecule (PECAM), and could be either positive or negative for VEGF (see Methods and Fig.2).

Flt-1 and Flk-1 mRNA expression

Treadmill exercise significantly increased the Flt-1 mRNA levels by 108 % in whole plantaris muscles (P < 0.05). In contrast, a single bout of exercise had no detectable effect on Flk-1 mRNA expression estimated just after completion of the treadmill running exercise (Fig. 4).

Figure 4. Quantitative densitometry for the ratios of Flt-1/cycA and Flk-1/cycA mRNA.

Quantitative densitometry for Flt-1/cycA (A) and Flk-1/cycA (B) mRNA ratio in whole plantaris muscle of rats at rest (control, n = 6) and following 90 min of treadmill exercise (active, n = 6). Gene expression was examined by real-time RT-PCR. Data are means ± s.e.m. * Significantly different from control values, P < 0.05.

ORP-150 and HIF-1α mRNA expression

Figure 5 shows ORP-150 and HIF-1α mRNA expression in whole plantaris muscles of control and exercised rats. No significant change was shown in HIF-1α mRNA level, whilst acute running exercise led to a 92 % increase in expression of ORP-150 mRNA (P < 0.05).

Figure 5. Quantitative densitometry for the ratios of ORP-150/cycA and HIF-1α/cycA mRNA.

Quantitative densitometry for ORP-150/cycA and HIF-1α/cycA mRNA ratios in whole plantaris muscle of rats at rest (control, n = 6) and following 90 min of treadmill exercise (active, n = 6). Gene expression was examined by real-time RT-PCR. Data are means ± s.e.m. * Significantly different from control values, P < 0.05.

DISCUSSION

In the present study we sought to quantify the fibre-type specificity of VEGF mRNA production at the end of a single bout of exercise in rats and to examine the co-ordinated expression of genes both regulated by hypoxia and involved in the regulation of expression or activity of the VEGF gene. In essence, we found that (1) skeletal muscle fibres alone, independently of endothelial cells, were a true source of production for VEGF mRNA; (2) levels of VEGF mRNA were mainly increased in fast type IIb fibres; and (3) levels of ORP-150 (an oxygen-dependent regulator of VEGF elaboration) mRNA increased at the end of a single bout of exercise in whole muscle.

Consistent with many previous studies (Breen et al. 1996; Asano et al. 1998; Gustafsson et al. 1999; Gavin et al. 2000), the present results show that levels of VEGF mRNA and protein both increased in whole skeletal muscle after a single bout of exercise, by 90 and 72 %, respectively. This finding suggests that the increase in VEGF transcript and protein are quantitatively related. It has recently been shown that the exercise-induced increase in VEGF mRNA observed in the initial stages of training gradually decreased over the course of a training programme (Lloyd et al. 2003). It is thus likely that the marked increase in VEGF mRNA and protein observed in the present study would be attenuated as training progresses, in accordance with the notion that VEGF production is under negative feedback (Richardson et al. 2000).

Using in situ hybridisation, a non-quantitative method, to detect mRNA signals, previous studies suggested that the predominant location of VEGF mRNA after exercise was within muscle fibres (Breen et al. 1996; Brutsaert et al. 2002; Lloyd et al. 2003). Real-time PCR analysis used in the present study provides the first direct evidence that levels of VEGF mRNA increased in muscle cells at the end of a single bout of exercise. This novel approach provided a very accurate and reproducible quantification of mRNA levels and, unlike other semi-quantitative PCR methods, it did not require post-PCR sample handling, thus preventing product contamination (Heid et al. 1996; Gibson et al. 1996). One of the main findings of this study was that levels of VEGF mRNA in pure myofibres from active rats were 88 % higher than in control animals, whilst levels of VEGF mRNA increased by a similar magnitude in whole muscle. The increased percentage of myofibres positive for VEGF mRNA at the end of exhaustive exercise, compared with those observed at rest, provided additional evidence that the exercise-induced increase in VEGF mRNA levels happened predominantly in muscle cells, but also in other cells such as endothelial cells.

Relevant to this finding is the possibility of fibre-type specificity in the expression of the VEGF gene during exercise. The regional differences in expression of VEGF mRNA after exercise have been recently investigated. Only modest and non-significant regional differences were shown using in situ hybridisation in gastrocnemius muscle after a single bout of treadmill running (Brutsaert et al. 2002). Using a reliable method to quantify VEGF gene expression, our results clearly show that levels of VEGF mRNA were twice as high in fast type IIb fibres at the end of exhaustive exercise as at rest. The apparent discrepancy with the previous study could be related to a larger involvement of plantaris muscle during running or to a more accurate and reliable technique specifically designed to provide quantitative measures of levels of VEGF mRNA within individual fibres. In contrast to in situ hybridisation, in which results may be related to differences in cell types other than myofibres, such as endothelial cells (Brutsaert et al. 2002), the technique used in the present study clearly separated VEGF expression by individual fibres and by endothelial cells.

Fast type IIb fibres are rarely recruited in rats housed in standard laboratory cages. During running, increased involvement of the type IIb fibres occurs with increased workload (i.e. treadmill elevation) and exercise duration (Terjung, 1976; Dudley et al. 1982). Moreover, acute exhaustive exercise in unaccustomed rats probably increases the contribution of fast type IIb fibres, as suggested by the responses of some metabolic enzymes in muscles predominantly composed of type IIb fibres (Lawler et al. 1993). It is known that type IIb fibres have low oxidative potential and are the largest fibres (Delp & Duan, 1996). Marked regional variations in fibre size and capillarity have been shown within the same skeletal muscle, and fast-twitch muscles such as plantaris show higher maximal diffusion distances of oxygen in regions predominantly composed of type IIb fibres (Deveci et al. 2001). When fast-fatiguable motor units are recruited during muscle contraction, local ischaemic areas are thus expected to occur within type IIb fibres (Richardson et al. 1995). Previous experiments supported the hypothesis that decreases in intracellular P_O2 could stimulate VEGF expression after exercise (Breen et al. 1996; Roberts et al. 1997; Richardson et al. 2001). Our results showing an increase in VEGF mRNA in response to exercise in fast glycolytic type IIb fibres are consistent with the hypothesis that local P_O2 could be involved in VEGF gene expression during exercise.

Our results failed to show any change in levels of HIF-1α mRNA in whole muscle after exercise, largely because HIF-1 oxygen regulation is mediated through stabilisation and translocation to the nucleus of HIF-1α, one of the two components of HIF-1, and not from events occurring at the level of gene expression (Kallio et al. 1998). The up-regulation of VEGF mRNA observed in rats exposed to moderate ambient hypoxia, with no change in HIF-1α protein expression, supports this hypothesis (Tang et al. 2000). So, the lack of alteration in levels of HIF-1α mRNA shown in the present study does not contradict the theory of a cellular hypoxic stress occurring during intensive exercise.

We studied the response to exercise of several genes that are known to be regulated by hypoxia. A 92 % increase in levels of ORP-150 mRNA occurred in whole muscle after a single bout of exercise. ORP-150 facilitates the intracellular transport of VEGF and the increased mRNA levels for this protein could contribute to ensuring the bioactivity of VEGF protein (Ozawa et al. 2001). It has been previously reported that the transcription of the two VEGF receptor genes was regulated differently by hypoxia (Gerber et al. 1997). It was shown that hypoxia up-regulates the expression of the Flt-1 gene in endothelial cells in vitro, while Flk-1/KDR mRNA levels were unchanged. This differential transcriptional regulation of the two receptors by hypoxia was related to an HRE located in the promoter region of the Flt-1 gene (Gerber et al. 1997). The present results support previous studies showing an exercise-induced increase in Flt-1 mRNA levels, whilst no significant change was detected in Flk-1 mRNA expression (Gavin & Wagner, 2002). Taken together, the increased levels in whole muscle of ORP-150 and Flt-1 mRNA, two hypoxia-inducible genes, are consistent with the suggestion that hypoxia could be one of the factors involved in exercise-induced angiogenesis. Moreover, these findings are consistent with the observation that levels of VEGF mRNA increased mainly in type IIb fibres, which are more sensitive to hypoxia than high-oxidative fibres such as type I or type IIa fibres. However, several stimuli other than local hypoxia or metabolic factors related to high cellular energy requirements may be involved in the exercise-induced up-regulation of the VEGF gene. While increased blood flow did not account for the increased VEGF mRNA expression at the end of exercise (Roca et al. 1998), other biomechanical events related to increased load bearing within muscle may play a role in modulating VEGF expression (Gustafsson & Kraus 2001).

In conclusion, the present study is the first to clearly demonstrate that within whole muscle, myofibres are a source of production of VEGF mRNA at the end of a single bout of exercise. Using a quantitative approach, we have demonstrated that the exercise-induced increase in VEGF mRNA occurred in fast type IIb fibres, which are large in size and predominantly glycolytic. These findings, together with the increased levels of mRNA transcripts of two hypoxia-inducible genes, ORP-150 and Flt-1, are consistent with the suggestion that inadequate tissue cell oxygenation may be one of the triggers of exercise-induced capillary growth.