The skeletal muscle capillary supply is an important determinant of maximum exercise capacity, and it is well known that endurance exercise training increases the muscle capillary supply. The muscle capillary supply and exercise-induced angiogenesis are regulated in part by vascular endothelial growth factor (VEGF). VEGF is produced by skeletal muscle cells and can be secreted into the circulation. We investigated whether there are differences in circulating plasma VEGF between sedentary individuals (Sed) and well-trained endurance athletes (ET) at rest or in response to acute exercise. Eight ET men (maximal oxygen consumption: 63.8 ± 2.3 ml·kg-1·min-1; maximum power output: 409.4 ± 13.3 W) and eight Sed men (maximal oxygen consumption: 36.3 ± 2.1 ml·kg-1·min-1; maximum power output: 234.4 ± 13.3 W) exercised for 1 h at 50% of maximum power output. Antecubital vein plasma was collected at rest and at 0, 2, and 4 h postexercise. Plasma VEGF was measured by ELISA analysis. Acute exercise significantly increased VEGF at 0 and 2 h postexercise in ET subjects but did not increase VEGF at any time point in Sed individuals. There was no difference in VEGF between ET and Sed subjects at any time point. When individual peak postexercise VEGF was analyzed, exercise did increase VEGF independent of training status. In conclusion, exercise can increase plasma VEGF in both ET athletes and Sed men; however, there is considerable variation in the individual time of the peak VEGF response.
- vascular endothelial growth factor
- soluble VEGF receptor Flt-1
aerobic exercise training results in several skeletal muscle adaptations, including the development of new capillaries, known as angiogenesis (2, 8). Exercise-induced angiogenesis increases the capillary surface area available for the diffusion of O2 and decreases the diffusional distance of O2 to the mitochondria, ultimately contributing to an increase in aerobic capacity (36). Vascular endothelial growth factor (VEGF) is a 45-kDa secretable basic heparin-binding homodimeric glycoprotein (14). VEGF is essential for embryonic vasculogenesis and angiogenesis (9, 13), maintenance of the capillary supply in normal skeletal muscle (34), and exercise-induced increases in skeletal muscle vascular density (1).
In sedentary individuals, acute exercise increases VEGF mRNA in human skeletal muscle (15, 20, 30). In contrast, skeletal muscle VEGF protein is reduced immediately after acute exercise, suggesting that VEGF is released by skeletal muscle (15). After exercise training, resting levels of skeletal muscle VEGF mRNA and protein are increased (18, 19), whereas the acute exercise-induced increase in VEGF mRNA is attenuated (29).
As a secretable protein, VEGF can be released into the circulation. Within 30 min of one-legged knee extension exercise, interstitial VEGF protein increases in human skeletal muscle of physically active individuals (22). Short-term exercise training can lower resting arterial plasma VEGF (19); however, circulating plasma VEGF is not different in older control and hypertensive patients when analyzed by activity levels (12). Plasma VEGF is higher in hypertensive (3), hyperlipidemic (6), atherosclerotic (7), and congestive heart failure patients (10). Interestingly, lipid-lowering medications reduce VEGF in hyperlipidemic patients (6). It has been suggested that elevated VEGF levels may be involved in the pathogenesis of heart disease (3). Thus an additional benefit to chronic exercise may be to lower VEGF and, therefore, slow the progression of heart disease.
In physically active men, femoral vein plasma VEGF protein is increased after 3 h of two-legged kicking ergometry, but arterial plasma VEGF remains unchanged, suggesting that circulating VEGF may not be increased in response to acute isolated muscle exercise (21). Circulating serum VEGF has been shown to decrease after a high-altitude marathon run (17) and increase immediately after a moderate-altitude ultramarathon run (32). Thus the effects of acute and chronic exercise on circulating VEGF are equivocal.
In the present report, we investigated whether a difference exists in plasma VEGF protein between well-trained endurance athletes (ET) and sedentary individuals (Sed) at rest and in response to acute exercise. We found that plasma VEGF is increased by acute systemic exercise at 0 and 2 h postexercise in ET only. There was no difference in VEGF between ET and Sed at any time point, due in part perhaps to large interindividual variability in VEGF. Analysis of peak postexercise plasma VEGF protein found that both ET and Sed significantly increased plasma VEGF protein, suggesting that acute systemic exercise does increase circulating VEGF levels in both Sed and ET; however, individual variability of the time course of the VEGF response may limit the ability to observe this increase.
Subjects. Eight male Sed and eight male ET volunteered to participate in the study, in accordance with the University & Medical Center Institutional Review Board. All ET regularly competed in bicycle, triathlon, or duathlon racing. All subjects were healthy nonsmokers, with no history of cardiopulmonary disease, and completed a Physical Activity Readiness Questionnaire before the commencement of the study. The physical characteristics of the subjects are listed in Table 1.
Maximal O2 consumption. Maximal O2 consumption (V̇o2 max) and maximum power output were measured on an electronically braked cycle ergometer (Lode, Excaliber Sport, Groningen, The Netherlands). Minute ventilation, O2 uptake, and CO2 production were continuously monitored via open-circuit spirometry (True Max 2400, Parvo Medics, Salt Lake City, UT). Heart rate was measured continuously (Accurex Plus, Polar Electro, Woodbury, NY). The test began with a 5-min warm-up at 75 W for Sed and 150 W for ET. After the warm-up, the workload was increased 25 W every minute until volitional fatigue. Subjects were verbally encouraged to continue for as long as possible. The criterion used to assess V̇o2 max included 1) a heart rate in excess of 90% of age predicted maximum (220 - age), 2) a respiratory exchange ratio of ≥1.10, and 3) identification of a plateau (≤150 ml increase) in O2 uptake despite a further increase in workload. In all tests, at least two of three criteria were met.
Body composition. Body density was determined via hydrostatic weighing. Residual volume was measured by O2 dilution (38). Body fat percentage was determined from body density based on the two-compartment model (33).
Experimental design. At least 5 days after the determination of V̇o2 max, subjects reported to the Human Performance Laboratory between 0700 and 0900 after a 12-h fast. Blood was sampled from a catheter inserted in an antecubital vein. The catheter was routinely flushed with saline to maintain patency of the catheter throughout the procedure. Fasted plasma was collected at rest for the determination of VEGF, the soluble VEGF receptor Flt-1 (sFlt-1), glucose, and insulin. For comparison between our subjects and previous reports on VEGF in exercise-trained subjects (17, 32), fasted serum was also collected for the determination of VEGF. All subjects performed and completed a 1-h cycle ergometer exercise bout at 50% of maximum power output. At 0, 2, and 4 h postexercise, plasma samples were collected for the measurement of VEGF. The 0-h sample was collected within 5 min after the completion of the exercise bout. Subjects completely consumed a 650-kcal standardized snack after the first postexercise blood draw, but within 45 min of the cessation of exercise, and were allowed water ad libitum throughout the study. Plasma was separated from venous samples obtained in the presence of EDTA and frozen at -80°C until analysis. Serum was allowed to clot for 30 min before separation and frozen at -80°C until analysis. VEGF is a heparin-binding protein (14); thus heparin was not used as an anticoagulant in the present study.
VEGF and sFlt-1. VEGF was determined in duplicate by high-sensitivity ELISA according to the manufacturer's instructions (R&D Systems, Minneapolis, MN). The sensitivity of the VEGF kit is 9.0 pg/ml. According to the manufacturer, significantly high levels of sFlt-1 can interfere with VEGF measurements. Thus plasma sFlt-1 was determined in duplicate by high-sensitivity ELISA according to the manufacturer's instructions (R&D Systems) to determine whether sFlt-1 would interfere with the VEGF measurement. The sensitivity of the sFlt-1 kit is 5.01 pg/ml.
Insulin and glucose. Hypoglycemia, induced by a hyperinsulinemic clamp, increases serum VEGF protein in young, healthy men (11). To investigate whether there is a relationship between resting VEGF and insulin sensitivity, plasma insulin and glucose were determined by an immobilized enzyme biosensor for glucose (YSI 2300 STAT Plus Glucose and Lactate Analyzer, Yellow Springs Instruments, Yellow Springs, OH) and by a chemiluminescent enzyme immunoassay for insulin (Access, Beckman Coulter, Brea, CA) according to the manufacturer's instructions. The homeostasis model assessment (HOMA) allows for the quantitative assessment of insulin action by using fasting concentrations of glucose and insulin according to the formula
where FPI and FPG are the fasting insulin (in μU/ml) and glucose concentrations (in mmol/l), respectively. The formula is based on an array of predicted glucose and insulin values that would be expected for many potential combinations of insulin action (27).
Statistics. Student's t-test was used to test for differences in subject characteristics and differences in plasma VEGF protein concentrations between Sed and ET at each time point, sFlt-1, glucose, insulin, and HOMA. Repeated-measures ANOVA (time postexercise) was used to test for differences in plasma VEGF within Sed and ET. A 2 × 2 mixed-plot factorial repeated-measures ANOVA was performed to analyze for differences in resting and peak postexercise plasma VEGF concentrations between the groups and to analyze for differences in resting plasma and serum VEGF between groups. After a significant F-ratio, a Fisher's paired least significant difference post hoc test was used to determine differences within a group between conditions. One sample at 2 h and one sample at 4 h were lost from ET and not replaced. The alpha was established at P ≤ 0.05 for all statistics.
The circulating plasma VEGF response to acute exercise for Sed and ET is shown in Fig. 1. VEGF was increased by exercise in ET at 0 and 2 h postexercise (rest: 32.2 ± 6.5; 0 h: 45.3 ± 7.6; 2 h: 45.2 ± 10.0; and 4 h: 36.1 ± 7.6 pg/ml). There was no increase in VEGF in Sed at any time point (rest: 41.3 ± 6.4; 0 h: 44.7 ± 7.3; 2 h: 48.8 ± 8.1; and 4 h: 48.0 ± 11.3 pg/ml). There was no difference between Sed and ET at any time point. During this analysis, we observed considerable variation in VEGF between subjects. Individual VEGF responses for Sed and ET are shown in Fig. 2. It is clear that the interindividual differences in VEGF are relatively large. Because of the large individual variances in the VEGF response to acute exercise (Fig. 2), we analyzed peak postexercise VEGF between Sed and ET. Peak postexercise VEGF responses are shown in Fig. 3. Exercise increased peak postexercise VEGF independent of training status (Sed: rest = 41.3 ± 6.4; exercise = 60.0 ± 11.0 pg/ml; and ET: rest = 32.2 ± 6.5; exercise = 49.8 ± 8.1 pg/ml). Resting plasma sFlt-1 protein concentrations are shown in Fig. 4. No difference was found between Sed and ET (41.8 ± 5.3 and 35.7 ± 2.6 pg/ml, respectively). The level of sFlt-1 found in either Sed or ET was considerably lower than that reported to interfere with the measurement of VEGF by the manufacturer (1,250 pg/ml).
Previous reports in trained subjects have shown equivocal results on the effect of exercise on VEGF. To determine whether these differences are the result of differences in blood source, we analyzed resting VEGF from both plasma and serum. Figure 5 demonstrates that, as expected, VEGF is much greater in serum than in plasma. Resting VEGF was approximately sixfold greater in serum compared with plasma samples (Sed: plasma = 41.3 ± 6.3, serum = 221.0 ± 28.4; and ET: plasma = 32.1 ± 6.5, serum = 221.0 ± 28.4 pg/ml) (Fig. 5A). There was no effect of training status on the difference between plasma and serum VEGF. A significant positive correlation (r = 0.68; P = 0.004) between serum and plasma VEGF was observed, suggesting that individuals with high plasma VEGF also have high serum VEGF (Fig. 5B).
Values for glucose, insulin, and HOMA and correlation values with resting VEGF are presented in Table 2. There were no differences between Sed and ET for glucose, insulin, or HOMA, although the differences in insulin (P = 0.07) and HOMA (P = 0.08) for a two-tailed test did approach statistical significance. No correlations were found between resting plasma VEGF and glucose, insulin, or HOMA.
In the present study, circulating plasma VEGF increased at 0 and 2 h postexercise in ET, whereas no increase was observed in Sed. However, when peak postexercise responses were analyzed, VEGF was significantly elevated in both Sed and ET. These results suggest that circulating VEGF is elevated in response to traditional systemic exercise in both Sed and ET, although there is considerable interindividual variation in the circulating VEGF response to exercise.
Basal circulating VEGF regulation. We did not observe any difference in resting and circulating VEGF between Sed and ET (Fig. 1). In contrast, plasma arterial VEGF is lower after a 10-day exercise training program (19). Perhaps the discrepancy between studies can be explained by the difference in exercise training status. In the present investigation, ET had established a long-term pattern of aerobic training, exercising 6 days/wk for at least the previous 6 mo, which is in stark contrast to 10 days of aerobic training. Another plausible explanation could be the use of a repeated-measures design by Gustafsson et al. (19), which statistically excludes the intersubject variability. Whether a long-term aerobic training intervention would change resting plasma VEGF levels requires further study.
In the present study, resting circulating plasma VEGF was 9.1 pg/ml lower in ET than in Sed. We performed a sample size analysis of our data and found that, for a difference of 9.1 pg/ml, a standard deviation of 18.0 pg/ml, a power of 0.80, and an α of 0.05, it would require a sample size of 62 subjects/group. It could be questioned whether our inability to observe differences in resting VEGF may result from the difference in age between Sed and ET. It is currently believed that age does not affect resting serum or plasma VEGF protein (25, 37); thus the fact that ET were older than Sed would not be expected to influence our findings.
Circulating VEGF is increased in several cardiovascular diseases, including hypertension, hyperlipidemia, atherosclerosis, and congestive heart failure (3, 6, 7, 10). It has been suggested that elevated VEGF levels may contribute to heart disease (3). Lipid-lowering medications lower VEGF levels in hyperlipidemia (6). If VEGF contributes to the pathogenesis of heart disease, then treatments that can lower VEGF levels may prove beneficial in the treatment of heart disease. It is well known that exercise can lower the risk of developing heart disease (5). In the studies demonstrating an elevated VEGF (3, 6, 7, 10), mean plasma VEGF ranged from 130 to 400 pg/ml, whereas controls ranged from 56 to 78 pg/ml. In the present study, our sedentary subjects demonstrated a mean VEGF of 41 pg/ml; thus in asymptomatic individuals VEGF is already low. Whether an exercise training intervention can significantly lower VEGF in populations with elevated VEGF remains to be determined.
We had questioned whether differences in fasting blood glucose could explain interindividual differences in circulating VEGF because hypoglycemia increases circulating VEGF (11). We did not observe a correlation between fasting glucose and VEGF. This may be because, in the present investigation, glucose exceeded 80.0 mg/dl in all subjects. In the investigation from Dantz et al. (11), blood glucose was driven below 75.0 mg/dl, which resulted in an increase in circulating serum VEGF. Therefore, it is possible that a threshold (<75.0 mg/dl) must be exceeded before an associated increase in VEGF protein is observed.
Acute exercise and VEGF regulation. Circulating VEGF was increased at 0 and 2 h postexercise in ET but was not increased at any specific time point in Sed (Fig. 1). However, we did not observe any difference in VEGF between Sed and ET at any time point. VEGF is released into the femoral vein 1 h postexercise in physically active men during two-legged knee-extension exercise (21). Interestingly, arterial plasma VEGF is not increased postexercise, suggesting that, although skeletal muscle may release VEGF, tissue other than recovering muscle may uptake VEGF, and thus circulating VEGF may not reflect the increased release of VEGF from skeletal muscle (21). Nemet et al. (28) found that 10 min of unilateral wrist flexion exercise increased circulating serum VEGF. Interestingly, Gunga et al. (17) observed a decrease in serum VEGF immediately after a high-altitude marathon run, whereas Schobersberger et al. (32) observed an increase in serum VEGF immediately after a moderate-altitude ultramarathon run. Given the influence of altitude and the nature of marathon and ultramarathon running, we questioned whether traditional, systemic (cycle ergometry) exercise of moderate intensity and duration, which is typically employed in exercise training programs, increases circulating VEGF. On further inspection of the VEGF responses to exercise, we observed considerable interindividual variability in VEGF (Fig. 2). In the present investigation, an evaluation of each individual's VEGF time course revealed that six of eight Sed increased plasma VEGF protein at some time point after exercise (Fig. 2A), whereas all eight ET increased VEGF at some time point postexercise (Fig. 2B). Comparing peak postexercise vs. resting VEGF revealed that exercise increased VEGF independent of training status (Sed increased 45% and ET increased 55%) (Fig. 3).
Interestingly, we observed similar increases in plasma VEGF between Sed and ET despite a twofold greater absolute workload during the 1-h exercise bout for ET (Fig. 3). It might be suggested from these results that the same absolute workload may produce differences in peak postexercise plasma VEGF. Gavin and Wagner (16) previously demonstrated a plateau effect in the skeletal muscle VEGF mRNA response to acute exercise, where VEGF mRNA was progressively increased from rest to workloads of 37.5 and 50% of maximum but was not further increased at a workload of 62.5%. Whether a similar plateau effect occurs in the circulating VEGF response to acute exercise remains to be investigated.
Recently, our laboratory has observed that VEGF protein is decreased in skeletal muscle immediately after exercise (15). This finding led us to question whether the disappearance of VEGF from the skeletal muscle may result in an increase in circulating VEGF. During the completion of the present work, Hoffner et al. (22) demonstrated that VEGF is released by skeletal muscle into the interstitial space. Combined with the work from Hiscock et al. (21) demonstrating that VEGF is increased in femoral vein postexercise, there is convincing evidence that VEGF is released by skeletal muscle during exercise. The present results demonstrating that VEGF is increased in the circulation in response to exercise are consistent with these findings.
Serum vs. plasma VEGF. Because reports on the effect of exercise have used both plasma (12, 21) and serum (17, 28, 32) for the measurement of VEGF, we questioned whether there were differences in plasma and serum between Sed and ET. We also had questioned whether individuals with high plasma VEGF demonstrate high serum VEGF. At rest, we found that serum VEGF was sixfold greater than plasma VEGF in both Sed and ET (Fig. 5A). This is in agreement with previous reports, which found a fivefold greater serum than plasma VEGF (24, 37). In addition, our subjects demonstrated serum and plasma VEGF values consistent with values of healthy subjects from previous reports (4, 37).
We did observe a significant correlation between serum and plasma VEGF protein (Fig. 5B; r = 0.68). This is consistent with other reports, which demonstrated significant correlations ranging from r = 0.56 to r = 0.75, where VEGF was measured from healthy controls and cancer patients (4, 37). Thus differences in VEGF with exercise should be detectable independent of the source of the VEGF measurement. It should be noted, however, that the use of serum for the quantification of circulating VEGF protein appears inappropriate because of the large contribution of platelet-derived VEGF (26). Platelets have been identified as significant transporters of VEGF (24, 31, 35). VEGF is released from platelets on platelet activation, and VEGF increases in serum as the time allowed for serum clotting increases (26). Therefore, serum collection time must be rigorously maintained to reduce additional experimental error in the measurement of blood-borne VEGF.
sFlt-1 regulation. sFlt-1 is a circulating high-affinity VEGF receptor that has been shown to specifically inhibit VEGF-induced mitogenesis in human umbilical vein endothelial cells (23). This suggests that sFlt-1 may act as a circulating VEGF antagonist in vivo. Recently, circulating plasma sFlt-1 was analyzed in hypertensives patients and controls (12). With the use of a standard Baecke physical activity questionnaire to classify older, healthy, and hypertensive patients into low and high physical activity categories, no difference in sFlt-1 was found between the groups. In the present investigation, we observed no difference in resting sFlt-1 protein between Sed and ET. Additionally, sFlt-1 was measured to determine whether levels exceeded a threshold of 1,250 pg/ml, which, according to the manufacturer, can interfere with VEGF quantification. Our results indicate that training status does not influence the level of sFlt-1 antagonism at rest and eliminates sFlt-1 as an extraneous variable in the measurement of VEGF protein.
In summary, we have demonstrated that circulating plasma VEGF does increase in response to acute exercise in both Sed and ET, although there is considerable interindividual variability. The use of either plasma or serum for the measurement of VEGF should yield similar conclusions on circulating VEGF. In addition, endurance exercise training does not alter circulating sFlt-1 during resting conditions. These results suggest that long-term aerobic training may not alter circulating VEGF regulation at rest or in response to acute exercise in young, asymptomatic individuals.
This study was supported in part by National Institute on Aging Grant AG-21891, an East Carolina University Faculty Senate grant, and a University of North Carolina Institute of Nutrition grant.
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