INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

INSULIN-LIKE GROWTH FACTOR-I (IGF-I), a 7.5-kDa polypeptide, plays an important role in the regulation of somatic growth, metabolism, and cellular proliferation, differentiation, and survival. IGF-I circulates in association with specific binding proteins (BPs). To date, six IGFBPs have been purified from biological fluids and their cDNAs were cloned (30); other potential IGFBPs have been identified based on sequence homologies (19). Most IGF-I (>80%) circulates in a 150-kDa, high-affinity complex, which also contains IGFBP-3 and an acid-labile subunit. Whereas incorporation of IGFs into this complex may limit their acute effects on metabolism, there is evidence that association with factors that prolong the survival of IGFs may enhance their long-term effects on growth (16). IGF-I also circulates bound to lower-molecular-weight IGFBPs, including IGFBP-1, a 30-kDa protein, which is produced largely in the liver and is thought to be the major short-term modulator of IGF-I bioavailability (23). Less than 1% of IGF-I is thought to circulate in a free or rapidly dissociable state and is thought to be readily available to mediate effects on target tissues through an endocrine mechanism, similar to the situation with steroid and thyroid hormones.

Athough it is well documented that circulating levels of IGF-I are regulated by nutrition, insulin, and growth hormone (GH), evidence is accumulating that exercise is potentially another important regulator of IGF-I levels. Serum total IGF-I concentrations increase with endurance and strength types of exercise (2, 6, 14, 21). It has been observed that maximal oxygen uptake and physical activity are correlated to resting plasma levels of IGF-I (25). Similarly, two investigations have shown (26, 27) that endurance training results in elevated resting levels of serum total IGF-I in healthy young and old individuals. Several studies have found serum levels of IGFBP-1 to increase with exercise (2, 18, 20, 32), presumably reflecting a reduction in insulin secretion and high levels of stress hormones, but this finding is not universal. Moreover, changes in serum IGFBP-3 levels acutely following exercise also are equivocal (20, 29). Furthermore, there is minimal and conflicting evidence regarding the effects of regularly performed exercise on resting levels of either IGF-I or IGFBPs. One report found no effect of exercise on IGFBP-1 and -3 levels in elderly individuals (26), whereas another observed increased IGFBP-1 levels in middle-aged men (15). In view of these observations, along with the fact that training effects on free IGF-I levels have not been examined, a purpose of this study was to determine whether the intense and prolonged training, as performed by competitive swimmers, would produce adaptive increases in serum total and free IGF-I concentrations. Determinations of serum IGFBP-1 and IGFBP-3 also were made because of their important relationships with IGF-I action. Furthermore, since the athletes were studied over ~6 mo of training, the experimental design allowed for determination of these variables in different phases of training.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects, training programs, and blood sampling. Fourteen members of the Northwestern University men's swimming team (designated as Team 1M) volunteered for study. Nine female members and five male members of the University of Illinois at Chicago swimming team (designated as Team 2F and Team 2M, respectively) also volunteered for study. All subjects were between the ages of 18 and 22 yr. They gave written informed consent, and the project was previously approved by the Human Subjects Committees at both universities.

All training programs progressively increased training volume (distance swam) over the first 4 mo, followed by a leveling off and then a taper phase (Table 1). All of the teams exercised 5-6 days/wk. Team 1M began their training 3 wk later (near the end of September) than Teams 2F and 2M. Their regimen included days in which they exercised twice, three to four times a week. This included dry-land resistance training. After 1 mo, Team 1M was swimming twice a day, generally for three sessions per week. Over the last month of training, the (first) morning session emphasized stroke work. Teams 2F and 2M performed the same training regimen together. After 6 wk of training (middle of October), they began to swim twice per day for two sessions per week until the last month of training, when they continued to exercise once per day. The highest volumes of training for both teams occured between the 2-mo and 4-mo measurements for all teams. Tapering for the end-of-season competition began in the last month of training for all teams. This reduced training lasted ~4 wk.

Blood samples were obtained from an antecubital vein. All subjects reported for blood sampling in the morning after an overnight fast. For Team 1M, samples were obtained between ~6:00 and 9:00 AM. For Teams 2F and 2M, fasting samples were obtained between 6:00 and 11:00 AM. Initial samples were obtained at the start of training. Second samples were taken after ~2 mo of training for each team. Third samples were obtained after ~4 mo of training. This 2- to 4-mo time point represented an interval with the highest training volume completed by all teams. A fourth sample was obtained only from Team 1M, when they were in the middle of their taper on the next to last week of the season. Final samples were obtained ~1 wk after the last competition for Team 1M and after 11-13 days (for 1 subject after 1 wk) after the last competition for Teams 2F and 2M.

Measurement of total and free serum IGF-I. Total IGF-I was measured by a nonextraction ELISA obtained from Diagnostic Systems Laboratories (DSL), Webster, TX. Samples were incubated in microtitration wells coated with monoclonal antibody to IGF-I. After incubation and washing, the wells were treated with another anti-IGF-I monoclonal antibody conjugated to horseradish peroxidase. The wells were washed, and substrate tetramethylbenzidine was added. The product of the reaction, after acidification, was measured spectrophotometrically at 450 nm. No cross-reactivity with IGF-II, insulin, or GH was detected. The intra-assay coefficient of variation (CV) was 8.6%. Sensitivity according to DSL was 1.3 pmol/l (0.01 ng/ml).

Free IGF-I was measured by a direct-assay kit provided by DSL. This is a two-site immunoradiometric assay similar to that described by Miles et al. (24). The conditions differed from those described for total IGF-I above in that plasma-unbound and readily dissociable IGF-I was captured by the first antibody under conditions in which plasma-bound IGF-I was not dissociated from the plasma BPs. The antibody-bound IGF-I was then detected by using a radiolabeled second antibody directed against a second epitope on IGF-I. The specificity was as described for total IGF-I. The intra-assay CV was 10.5%. Sensitivity according to DSL was 3.9 pmol/l (0.03 ng/ml).

Serum IGFBP-1 and IGFBP-3 determinations. The principle of each of these ELISAs is as described for total IGF-I, with spectrophotometric measurement as the end point. Materials were obtained from DSL. No cross-reactivities with IGF-I, IGF-II, human GH, insulin, IGFBPs 2-6 (for IGFBP-1), or IGFBP-1 (for IGFBP-3) have been found by the supplier. Intra-assay precision was 5.1% for IGFBP-1 and 9.5% for IGFBP-3. Sensitivity according to DSL was 55 pmol/l (0.25 ng/ml) for IGFBP-1 and 1.4 pmol/l (0.04 ng/ml) for IGFBP-3.

Measurement of serum cortisol. Cortisol was measured by a direct assay after 1/100 dilution in 0.1 M citrate buffer, pH 4.0 (7). Antiserum produced in this laboratory cross-reacts 17.45% with 11-deoxycortisol, 0.2% with cortisol phosphate, and <0.1% with dexamethasone, progesterone, estradiol, and testosterone. [1,2-³H]cortisol for the assay was obtained from New England Nuclear, DuPont, Boston, MA. Antibody-bound ligand was separated by dextran-coated charcoal. The sensitivity of the assay was 22 nmol/l (8 ng/ml). The intra- and interassay CVs for cortisol in recent assays were 8 and 11%, respectively.

Skinfold and girth analyses. Skinfold thicknesses were measured at five sites. They were shoulders, upper arm, forearm, thigh, and calf. Girth (circumference) measurements were performed at these same sites. These sites were considered to be potentially sensitive to any muscularity changes during the course of training. They were analyzed individually and as a summation of the respective measurements.

Statistics. The data were analyzed by using repeated-measures analysis of variance using the SYSTAT program of Wilkinson (36). Adjusted values were calculated for eight missing data points (out of 126 total) when samples were not obtained. These values represented only one missed time point for eight different subjects. Following significant F ratios, Duncan's multiple-range post hoc tests were used for evaluation of individual means. Statistical significance was set at the 95% level of confidence.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Total body mass, skinfold and girth analyses, and performance. Total body mass remained constant throughout all phases of training in all three groups (Table 2). The summation of skinfolds remained constant for Team 1M. Team 2F had skinfold reductions of 13 and 18% by the last two measurement points. Team 2M had a 25% reduction in total skinfolds by 2 mo of training and maintained this reduction throughout the remaining phases of training (Table 2). When girth measurements were evaluated individually, the same pattern was observed at each site. Therefore, the average summation of girths is reported; it remained unchanged in all teams at all phases of training.

Based on differences in times from early season until the last competition, Teams 1M, 2F, and 2M swam 7.0, 3.1, and 5.7% faster, respectively.

Total IGF-I. For Team 1M, there was a progressive increase in total serum IGF-I levels for the first two measurement points (up to 4 mo of training) (Fig. 1). Values were significantly (P < 0.05) increased by 76% above levels from the start of training and were maintained at these higher levels throughout the end of training.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   A: resting total serum concentrations of insulin-like growth factor I (IGF-I) at various phases of training. Values are means ± SE. Statistically higher differences: Team 1M (n = 14 men) at 4 mo, 5 mo, and at end of season vs. start of training; Team 2F (n = 9 women) at 4 mo and end of season vs. 2 mo; P < 0.05. B: resting free serum concentrations of IGF-I at various phases of training. Values are means ± SE. Statistically higher differences: Team 1M at 2, 4, and 5 mo vs. start of training; Teams 2M (n = 5 men) and 2F at all training and at end of season points vs. start of training; P < 0.05. See text for further description of teams.

Compared with starting levels, total IGF-I levels for Team 2M also tended to rise with training, but these changes did not achieve statistical significance (Fig. 1). Team 2M represented a small sample size, and, although the results were not significant by ANOVA, it should be pointed out that all five subjects had higher total IGF-I levels after 2 and 4 mo of training than they had at the start of training. Individual improvements varied between 3 and 141%. These effects were significant when nonparametric statistics (sign test) were used.

For Team 2F, there were significantly higher (68%) concentrations of total IGF-I at 4 mo of training and at the end of training, compared with the 2-mo training point (Fig. 1). The female response did not follow the pattern of increased IGF-I levels observed in men for the first 2 mo of training. When all the swim teams were pooled, data analysis showed that there were significantly higher total IGF-I levels after both 2 and 4 mo of training, compared with those at the start of training.

Free IGF-I. The levels of free IGF-I for Team 1M were significantly increased by 77-102% at all training measurements, compared with those observed at the start of training (Fig. 1). Values returned to pretraining levels 7 days after competitive training was over. Similarly for Teams 2F and 2M, free IGF-I concentrations were more than twofold higher after 2 mo of training and were further elevated (four- to sevenfold) at 4 mo and after training had ended, compared with levels at the beginning of training (Fig. 1). A pooled analysis of all the teams indicated that free IGF-I levels at all training points were significantly higher than at the beginning of training.

IGFBP-3. For Teams 1M and 2M, the levels of immunoreactive IGFBP-3 were significantly higher (30 and 97%, respectively) by 4 mo than at the start of training and stayed at these higher levels throughout further training and the end of training (Fig. 2). The differences in the amount of increase between teams primarily reflect the lower initial baseline levels for Team 2M. Team 2F had higher concentrations of IGFBP-3 at both training (29%, 2 mo; 45%, 4 mo) and end of training (53%) points, compared with values at the start of training.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   A: resting serum concentrations of IGF-binding protein-3 (IGFBP-3) at various phases of training. Values are means ± SE. Statistically higher differences: Teams 1M and 2M at 4 and 5 mo (Team 1M) and at end of season vs. start of training; Team 2F at all training and at end of training points vs. start of training; P < 0.05. B: resting molar ratios of total IGF-I/IGFBP-3 at various phases of training. Values are means ± SE. Statistically lower differences: Team 2F at 2 mo vs. start of training; P < 0.05.

When the total IGF-I/IGFBP-3 results were plotted as molar ratios (Fig. 2), there were no significant increases at any of the training points. In fact, the ratio for Team 2F was significantly lower after 2 mo of training than at the start of training.

IGFBP-1. There was only one significant change in IGFBP-1 levels among all teams at the various training points. Team 1M concentrations declined by 45% when measured at the end of training, compared with concentrations at 4 mo of training (Fig. 3).

View larger version (15K): [in this window] [in a new window]   Fig. 3.   Resting serum concentrations of IGF-binding protein-1 (IGFBP-1) at various phases of training. Values are means ± SE. Statistically lower differences: Team 1M at end of training vs. at 4 mo; P < 0.05.

Cortisol. Serum levels of cortisol were relatively stable throughout the first 4 mo of training (3 measurements) for all teams, with values ranging from 547 ± 54 to 615 ± 33 nmol/l for Team 1M, from 450 ± 86 to 505 ± 67 nmol/l for Team 2M, and from 657 ± 66 to 809 ± 110 nmol/l for Team 2F. By the end of training, all teams showed markedly elevated levels of serum cortisol, with Teams 1M (1,134 ± 165 nmol/l), 2M (834 ± 124 nmol/l), and 2F (1,013 ± 74 nmol/l) higher by 94, 70, and 25%, respectively, than serum cortisol levels seen at the beginning of training. The smaller percentage increase found for Team 2F members was due to their high initial starting levels (809 ± 110 nmol/l).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Two previous studies have evaluated the effects of training on resting total IGF-I concentrations (26, 27). Poehlman et al. (26) observed that cycling training of low-to-moderate intensity for 8 wk partially reverses the age-related decrease in IGF-I concentrations observed in older individuals. More recently, Roelen et al. (27) found in young subjects that more intense cycling training twice a day for 2 wk produced a 37% increase in resting total levels of IGF-I. Our results, obtained with young swimmers, showed a positive effect of training on serum levels of total IGF-I. The most striking effects were found for Team 1M, with increases of 70-80% over the last three measurement points. However, the fact that after 2 mo of training only modest changes in total IGF-I had occurred (women had even lower values) in all groups suggests that a relatively long course of training is required to realize peak adaptive increases in total IGF-I with this type of training. Furthermore, these data show that the elevated total IGF-I concentrations persist even with marked reductions in training volume.

Additionally, the total IGF-I responses in the female swimmers did not follow the pattern of increased IGF-I levels observed in men for the first 2 mo of training. Supporting this fact is the finding that the total IGF-I/IGFBP-3 ratio for Team 2F was significantly lower after 2 mo of training than at the start of training. Interestingly, this is a potentially important issue in that there may be a difference in the early response to exercise between men and women. Studies appear warranted to examine these short-term effects to further determine whether gender-specific responses to training exist.

One of the major new observations of this work are the marked increases in free IGF-I levels that accompanied training in all three teams. However, there were certain inconsistencies. Team 1M was the only group to demonstrate increased levels in both total and free IGF-I throughout training. Another was the variable responses in free IGF-I levels at the end of training, when free IGF-I levels were severalfold higher for Teams 2M and 2F than were those for Team 1M. When taken together, these discrepancies might be related to the different training regimens, as Team 1M had a higher volume of swim training and more dry-land work than had Teams 2F and 2M. At the last measurement point, training had completely ended for Teams 2F and 2M, whereas at least some of the members of Team 1M continued to swim at a reduced level. Consequently, the immediate postseason activities of the subjects may have varied over a wide range. Nevertheless, it is not presently known how both the differences in training and the end-of-season activities would elicit diversified responses.

Overall, the free IGF-I findings suggest that factors other than total IGF-I levels also play an important role in determining IGF bioavailability in vivo. Along similar lines, Skjaerbaek et al. (31) also observed a greater increase in serum free IGF-I relative to total IGF-I in response to 14 days of GH treatment. Because most IGFs circulate in association with IGFBP-3 as part of a high-molecular-weight complex, and since IGF-I has been thought to regulate levels of IGFBP-3 (4), we also measured levels of IGFBP-3 in the circulation of athletes during training. Levels of IGFBP-3 also increased, but the ratio of total IGF-I to levels of immunoreactive IGFBP-3 was not increased in any of the three groups, indicating that changes in free IGF-I levels cannot be accounted for by changes in the IGF-I/IGFBP-3 complex or in the relative amounts of IGFBP-3 in the circulation. Whereas this is not normally what would be expected from the total and free IGF-I data, at the same time, it is important to recognize that immunoassays measure levels of both intact and proteolytic cleaved fragments of IGFBP-3 (22) and that proteolytic alterations of IGFBP-3 by circulating proteases have been shown to reduce the stability of IGF/IGFBP-3 interactions and increase the bioavailability of IGFs in other settings (3, 5, 28). Schwarz et al. (29) have reported that IGFBP-3 protease activity is increased in short-term cycling exercise. Whether this increase reflects alterations in GH secretion (28) or other factors remains to be determined. Additional studies also are required to determine whether circulating IGFBP-3 protease activity and modifications of circulating forms of IGF-I and IGFBP-3 contribute to the sustained and consistent increase in IGF-I availability in two of the groups, which we observed in long-term training.

Changes in circulating levels of other IGFBPs also may contribute to changes in the availability of IGFs during training. IGFBP-1 is thought to be the major short-term modulator of IGF bioavailability (23) and is regulated by multiple factors, including insulin, glucocorticoids, cAMP agonists, and GH (33). Inverse relationships between serum levels of IGFBP-1 and free IGF-I without changes in levels of total IGF-I have been reported in obesity, overnight fasting, and after glucose ingestion (12, 13). Other researchers have found (18, 20, 32) that IGFBP-1 levels may increase after prolonged exercise. In the present study, increases in free IGF-I occurred without any decrease in the level of IGFBP-1, with the exception of a decline in Team 1M at the end of training. These results indicate that changes in free IGF-I levels are not due to a reduction in IGFBP-1 levels during long-term training. Studies to examine circulating levels of other IGFBPs may provide additional insight into specific mechanisms determining the effects of training on IGF bioavailability in young athletes.

A consistent finding from this work is that resting serum cortisol concentrations became elevated in all groups only at the last measurement, after the taper phase of training and the last competition had been over for 1-2 wk. It is possible that the effects of the high-intensity training and the stress of competition contributed to this phenomenon. Nevertheless, the reason as to why this effect is found at this point of training is unknown. Additionally, we observed that the women swimmers had higher serum cortisol levels than the men. Villanueva et al. (35) also demonstrated chronically elevated serum cortisol in women runners compared with sedentary women. This effect may be unique to training women.

One consideration emerging from these findings of elevated levels of total and free IGF-I is their physiological relevance to exercise-induced adaptations, particularly in skeletal and cardiac muscle. Several lines of available information point to IGF-I at least indirectly as a potential contributing factor in 1) the hypertrophy and/or maintenance of muscle mass associated with various types of exercise and functional overload (1, 8, 10); 2) glucoregulation after exercise training (17); and 3) improvement in cardiac contractile parameters in both rats and humans (9, 11, 34). Our own findings showed that two of the three teams did not have any changes in total body mass but had significant reductions in skinfold measurements. This reduction occurred with a maintenance of girth measurements in these same areas, which suggests indirectly an increase in muscle mass in these areas. The augmented levels of serum total and free IGF-I and IGFBP-3 support the possibility that IGF-I plays a role in these responses.

In summary, these results indicate that serum levels of total and free IGF-I and total IGFBP-3 can be increased with strenuous training and maintained with reduced training (tapering). The increases in free IGF-I are not accounted for by alterations in total levels of IGF-I or IGFBP-3. Other important determinants of free IGF-I appear to be operating. Among the candidates are other BPs and changes in the IGF-I/IGFBP-3 complex, including modifications of IGFBP-3 (such as increased proteolysis), which can affect the interstability of peptide interaction.