Responses of Catecholestrogen Metabolism to Acute Graded Exercise ...

0021-972X/97/$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright © 1997 by The Endocrine Society

Vol. 82, No. 10 Printed in U.S.A.

Responses of Catecholestrogen Metabolism to Acute Graded Exercise in Normal Menstruating Women before and after Training* ´ E, PETER BALL, BA ¨ RBEL SEIDLITZ, GERRIT VAN KRANENBURG, CARL DE CRE PETER GEURTEN, AND HANS A. KEIZER Interuniversity Project on Reproductive Endocrinology in Women and Exercise, the Department of Applied and Experimental Reproductive Endocrinology (C.D.C.), The Institute for GynecoEndocrinological Research, B-3000 Leuven 3, Belgium; the Department of Biochemical and Clinical Endocrinology (P.B., B.S.), Medical University of Lu¨beck, D-23538 Lu¨beck, Germany; and the Department of Movement Sciences (G.V.K., P.G., H.A.K.), Faculty of Health Sciences, University of Maastricht, NL-6200 MD Maastricht, The Netherlands ABSTRACT It has been hypothesized that exercise-related hypo-estrogenemia occurs as a consequence of increased competition of catecholestrogens (CE) for catechol-O-methyltransferase (COMT). This may result in higher norepinephrine (NE) concentrations, which could interfere with normal gonadotropin pulsatility. The present study investigates the effects of training on CE responses to acute exercise stress. Nine untrained eumenorrheic women (mean percentage of body fat 6SD: 24.8 6 3.1%) volunteered for an intensive 5-day training program. Resting, submaximal, and maximal (tmax) exercise plasma CE, estrogen, and catecholamine responses were determined pre- and post training in both the follicular (FPh) and luteal phase (LPh). Acute exercise stress increased total primary estrogens (E) but had little effect on total 2-hydroxyestrogens (2-OHE) and 2hydroxyestrogen-monomethylethers (2-MeOE) (5 O-methylated CE after competition for catechol-O-methyltransferase). This pattern was not significantly changed by training. However, posttraining LPh

mean (6SE) plasma E, 2-OHE, and 2-MeOE concentrations were significantly lower (P , 0.05) at each exercise intensity (for 2-OHE: 332 6 47 vs. 422 6 57 pg/mL at tmax; for 2-MeOE: 317 6 26 vs. 354 6 34 pg/mL at tmax). Training produced opposite effects on 2-OHE:E ratios (an estimation of CE formation) during acute exercise in the FPh (reduction) and LPh (increase). The 2-MeOE:2-OHE ratio (an estimation of CE activity) showed significantly higher values at tmax in both menstrual phases after training (FPh: 111%; LPh: 123%; P , 0.05). After training, NE values were significantly higher (P , 0.05). The major findings of this study were that: training lowers absolute concentrations of plasma estrogens and CE; the acute exercise challenge altered plasma estrogens but had little effect on CE; estimation of the formation and activity of CE suggests that formation and O-methylation of CE proportionately increases. These findings may be of importance for NE-mediated effects on gonadotropin release. (J Clin Endocrinol Metab 82: 3342–3348, 1997)

M

norepinephrine (NE) concentrations, which might interfere with normal gonadotropin pulsatility and potentiate a status of hypoestrogenemia. A relationship between exercise and CE was first explored by Russel et al. (2, 4). These authors observed that in highly trained oligomenorrheic swimmers, there was a decrease in plasma LH, prolactin, and estradiol and a concomitant increase in circulating 2-hydroxyestrogens. After this, for more than a decade, no new research was proposed on the subject. Recently, however, several studies were launched in an attempt to avoid some of the major shortcomings in experimental design, methodology, and sample collection that had hindered interpreting the previous results. By using a prospective design, it was demonstrated that in untrained eumenorrheic women, acute exercise did not provoke any changes in either plasma 2hydroxyestrogens or their methylated products (5). In a subsequent study, it was observed that training simultaneously led to reduced resting plasma levels of estrogens and CE but, nevertheless, also to a proportionately increased formation and methylation (6). Therefore, it seemed that the original hypothesis, which had claimed that training increases 2-hydroxyestrogen levels, needed to be corrected. Apparently, training mainly increases the proportion of 2-hydroxyestrogen for-

ANY STUDIES have shown that there exists a strong association between physical exercise and altered reproductive axis functioning (for a review, see Ref. 1). Attempts to explain these phenomena have focused on nutritional intake, body composition, stress, and gonadotropin pulsatility. Intense physical exercise leads to hypoestrogenemia and may be accompanied by relatively high circulating plasma levels of certain catecholestrogens (CE) (2). CE are the 2- or 4-hydroxylated metabolites of estrone and estradiol, which have the pharmacological properties of both estrogens and catecholamines (CA) and are formed from estrogens in the brain and peripherally. It has been hypothesized that exercise-induced higher CE concentrations might increasingly compete for the CA-inactivating enzyme, catechol-Omethyltransferase (COMT) (3). This could result in higher

Received October 22, 1996. Revision received March 25, 1997. Rerevision received June 18, 1997. Accepted June 27, 1997. Address all correspondence and requests for reprints to: C. De Creé, M.D., Department of Applied and Experimental Reproductive Endocrinology, The Institute for Gyneco-Endocrinological Research, P.O. Box 134, B-3000 LEUVEN 3 (Belgium). * Presented in part at the 44th Annual Meeting of the American College of Sports Medicine, Denver, Colorado, May 28 –31, 1997 (Abstract B-34).

3342

CATECHOLESTROGENS AND EXERCISE

mation from ordinary E and the proportion of O-methylation. Thus, in euestrogenic women, vigorous training will induce a drop in estrogens but also in CE and methylated CE; but, the percent drop in estrogens will outweigh the drop in CE and monomethylethers, suggesting a proportional increase in CE metabolism. Apparently, it is only when the subject becomes hypoestrogenemic and low estrogen levels have stabilized, that not only the proportional formation and methylation of 2-hydroxy CE will increase, but also the absolute levels, as indeed observed by Russel et al. The present study certainly does not attempt to resolve the debate between those who support or reject a biological role for CE. We also will not focus on any CE other than the 2-hydroxyestrogens. The purpose of this paper is limited to investigating whether, after a brief period of exhaustive training, plasma levels of 2-hydroxyestrogens in normal menstruating women respond any differently to acute exercise, when compared with pretraining conditions. Based on previous findings, our hypothesis is that in normal menstruating women, after training, the acute exercise will produce a proportionately higher formation and O-methylation of 2-hydroxyestrogens, which will not necessarily be accompanied by an increase in absolute plasma levels. It is also hypothesized that such increases will be most prominent during the luteal phase (LPh), because this is the phase when the female reproductive hormone system is most susceptible to exercise effects (1). Materials and Methods Experimental subjects Nine healthy eumenorrheic untrained women (mean age 6 sd: 20.4 yr 6 1.3, range 18 –22 yr; age of menarche: 13.2 6 1.3 yr; cycle length: 28.9 6 2.9 days) volunteered to participate in this study. Their mean (6sd) physical characteristics were: height: 172 6 4.7 cm, body mass: 61.4 6 4.8 kg, percentage of body fat (estimated according to Ref. 7): 24.8 6 3.1%. All subjects were nulliparous Caucasian women. Each individual gave her written consent, and procedures were carried out in accordance with the Declaration of Helsinki. Subjects then had to fill in a questionnaire to provide information concerning their training and gynecological history. It was a requirement that each individual have a regular ovulatory cycle occurring each 26 –32 days, with a documented LH surge in the months immediately before and after the experiments. In addition, they were not allowed to take oral contraceptives and had to be free of any disorder that could interfere with hormone metabolism. Each subject kept a menstrual diary. Ovulation was ascertained by plasma progesterone analysis (data not shown). The LH surge was determined in fasting morning urine using Ovusticks (Samenwerkende Apothekers, The Netherlands). Nutrient and daily caloric intake did not differ among the subjects when menstrual phases or pre- and posttraining cycles were compared. The subjects acted as their own controls. The same subjects have participated previously in another trial (3, 4).

Experimental design Each subject was exercise-tested four times: two standardized pretraining incremental exercise tests [one in the follicular phase (FPh) on a day between days 7–10 and one in the LPh, 10 –12 days after ovulation, i.e. on a day between days 23–25], and two posttraining exercise tests, again in each menstrual phase and on the same cycle days as the pretraining tests. After apparatus and gas calibration, the subject was placed on an electrically braked cycle ergometer (Lode b.v., Groningen, The Netherlands) and started cycling at time zero (t0). A detailed description of the incremental exercise test can be found elsewhere (5). Briefly, it

3343

consisted of a 2-min warm-up period at 25 Watt (W), followed by 4 min cycling at 50 W. Exercise intensity was increased by bouts of 50 W, each 4-min interval, up to 150 W. After having cycled for 4 min at 150 W, exercise intensity was lowered for 2 min to 50 W. At this submaximal level of approximately 60% maximal oxygen consumption (V˙O2max) (plasma lactate concentration of 2.0 –3.0 mmol/L), named tsubmax, blood samples were collected. Exercise intensity was then raised again for 1 min at 150 W. Each following minute, the exercise intensity was increased until total exhaustion or l00% V˙O2max, labeled tmax. Exercise intensity was lowered to 50 W during recuperation time. During the test, the subject breathed through a breathing mask. The subject’s expired gas was continuously sampled with a breath-by-breath method (Oxycon b, Mijnhardt-Ja¨ger b.v., Bunnik, The Netherlands) to determine metabolic and ventilatory variables. Recorded variables were: ventilation, oxygen consumption (V˙O2), and carbon dioxide expiration.

Training protocol Subjects underwent standardized lab training on a cycle ergometer daily for 5 days immediately preceding the second FPh test and the second LPh test. Each day, training started with 6 min cycling at 100 W. Subsequently, intensity was increased every 2-min interval by 25 W, until the individual’s maximum of that day was reached. The purpose of this maximal training exercise was to establish the individual’s maximal physical working capacity (MPWC), to design her individualized interval training program. After a 10-min recovery period at 50% MPWC, the subject switched to interval training, starting with 2-min series at 90% of the individual’s MPWC, followed by 2-min recovery periods at 50%. When the individual failed to continue the 90% MPWC exercise bouts (i.e. when the number of revolutions decreased below 60 rpm), the intensity of the intervals was lowered by 10%. The same rationale was applied until the intensity of the intervals had to be reduced to 80% and 70%, respectively and the subject reached complete exhaustion, at which point training was stopped. This was usually close to 90 min of training. A more detailed description of the training schedule and standardization procedures can be found elsewhere (6). We selected this type of training because earlier studies from our laboratory had shown that this type of exercise was able to deplete muscle glycogen stores almost completely and to provoke a disturbance of normal gonadotropin pulsatility (8). These and other authors showed that it was the training intensity (even if short-term), rather than the duration of the exercise, that was responsible for interfering with normal hypothalamic gonadotropin release (8, 9).

Blood sample collection, preparation, and storage All samples were obtained according to the pattern based on both the basal body temperature (BBT) and the Ovustick method on a day between days 7–10 of the FPh and on a day between days 23–25 of the LPh. Half an hour before each test, the subject received a venous indwelling 21-gauge Teflon Quick catheter (Travenol laboratories, Deerfield, IL) in the antecubital vein for blood collection for total protein and hormone determination. A three-way stopcock was attached to the indwelling cannula, and patency was assured by injecting 1 mL sterile saline at fixed intervals. Blood was drawn during rest (t0) and during tsubmax and tmax in disposable syringes (20 mL). Blood samples were collected in precooled lyophilized ethylenediamine tetra-acetate glass tubes stored in an ice bath at 2 C. Blood samples for assays were immediately centrifuged at 2000 3 g for 10 min at 4 C. Blood for determination of 2-hydroxyestrogens and their monomethylethers was prepared as follows: After centrifugation, as described above, 2 mL plasma were pipetted in 5-mL polyethylene Eppendorf cups, in which 1 mL of a 3% aqueous ascorbic acid solution had been added to prevent oxidative decomposition. The samples were then instantly frozen, using liquid nitrogen, and stored at 280 C until assaying, as described above. The plasma for the determination of all other plasma hormones was pipetted in 3-mL Eppendorf cups (without ascorbic acid) and immediately frozen and stored, as described above.

Hormone analyses Because the free or unconjugated plasma concentrations of CE are often close to the minimal detectable concentration of normal assaying

3344

´ E ET AL. DE CRE

procedures, for the sake of the accuracy and validity, we determined the unconjugated and conjugated fractions. This was done in a single-assay procedure, with the conjugated CE being measured after hot acid hydrolysis. The various steps of this relatively complex RIA have been described previously in detail (5). The results for the CE are given as total 2-hydroxyestrogens (2-OHE), which consist of the sum of unconjugated 2-hydroxyestrone (2-OHE1) and 2-hydroxyestradiol (2-OHE2), without 16a- or any other CE metabolites. Similarly, the values for 2-hydroxyestrogen-monomethylethers (2-MeOE), as mentioned in this study, indicate the sum of the total conjugated and unconjugated fractions of both 2-hydroxyestrone-monomethylethers or methoxyestrone (2-MeOE1) and 2-hydroxyestradiol-monomethylethers or methoxyestradiol (2-MeOE2). Despite the fact that most literature on exercise and estrogens has only measured unconjugated plasma estradiol (E2), we felt we had to determine primary estrogens as total estrogens (E), if we wanted to derive any information on the formation of total CE. Thus, E represent the sum of conjugated 1 unconjugated estrone (E1) and E2. Similarly to the CE, E did not contain any 16a- or any other estrogen metabolites. E was determined in one specific RIA. The intraassay variability for all steroid and nonsteroid hormone analyses was below 10%. The interassay variability was less than 17.2% for 2-OHE and 2-MeOE, and 13.3% for E. The minimal detectable concentration was approximately 6 pg/mL for all CE assays. Because CE competitively bind to COMT, the enzyme that decomposes CA, it was decided to measure NE, epinephrine (EPI), and dopamine (DA). All CA were assessed by electrochemical determination, after separation, using HPLC. The procedure and the equipment used have been previously described in detail (5). The respective interassay coefficients of variation were 4.5% for NE, 11.3% for EPI, and 11.8% for DA. Plasma LH concentrations were determined using an immunoradiometric assay purchased from Serono (Geneva, Switzerland). The cross-reactivity with hCG, FSH, and TSH was # 1.6, and the sensitivity was 0.4 mIU/mL. Plasma progesterone (P4) was determined after ether extraction, using a commercially available kit from Radio-Isotopen Service (Wu¨rlingen, Switzerland; inter-assay coefficient of variation: 7.2%; minimal detectable concentration: 0.13 ng/mL). Hematocrit was measured by the microcentrifuge method and total protein by the Biuret method (data not shown).

JCE & M • 1997 Vol 82 • No 10

the FPh, and from 250 6 31 W to 258 6 25 W in the LPh). Pretraining VO2max was lower (P , 0.05) in the LPh, compared with the FPh (41.5 6 4.1 vs. 45.7 6 4.5 mL•kg21•min21, respectively). After training, the values were 42.5 6 4.1 (FPh) and 43.5 6 3.4 mL•kg21•min21 (LPh). Hormonal responses to exercise and training

Hormonal responses to exercise are expressed both as absolute values (Fig. 1) and as percent changes from pretraining resting values (Table 1). The posttraining plasma levels of E and 2-OHE were significantly reduced (P , 0.05) at all exercise intensities. These reductions were most pronounced during the LPh. Primary estrogen levels at maximal exercise intensity showed a percent increase from resting levels, which was significant during each of the posttraining trials (P , 0.05). Percent changes in plasma CE during acute exercise were not significant. The magnitude of the overall changes in plasma concentrations of E and 2-OHE during acute exercise was not significantly altered by training in either of the cycle phases measured. In spite of posttraining plasma concentrations of E (tsubmax: 256%, tmax: 245%) and 2-OHE (tsubmax: 224%, tmax: 221%), which were significantly

Statistical analysis Anthropometric data, before and after training, were compared using paired t tests. To obtain a rough estimation of CE formation, we calculated the ratio of the end product to its precursor, i.e. the 2-OHE:E ratio. Furthermore, we computed the ratio of O-methylated CE to nonmethylated CE (2-MeOE:2-OHE), as a measure of CE activity. Hormonal data were adjusted for changes in plasma volume during acute exercise (10). Comparisons of hormonal responses, between different menstrual phases, were made by a two-way mixed-model ANOVA for groups with repeated measures. The level of significance for individual contrasts within each of the subjects was adjusted to limit experimental error rate to a maximum of 0.05. The ratios of the steroid responses were evaluated by analyzing the magnitude of absolute and relative changes during submaximal and maximal exercise intensity, compared with resting values. Phase, group, phase group (interactive), and subject effects were evaluated independently by maximum likelihood procedures. Correlations were evaluated by ANOVA. Accounting for individual variability across phases did not alter the correlation. Descriptive data are presented as means 6sd. Comparisons between group means are given as means 6 se.

Results Physiological characteristics

Menstrual cycles in all subjects were ovulatory both before and after training, although after training the plasma levels of LH and P4 were lower (data not shown). Menstrual cycle length was not significantly different after training. After having participated in an intensive 5-day training program, there was a trend for mean MPWC of the subjects to be higher (from 241 6 40 W to 250 6 23 W in

FIG. 1. Plasma responses of total estrogens (E) and CE (2-OHE & 2-MeOE) to acute incremental exercise in different menstrual phases before and after a 5-day exhaustive training program. Data are means 6 SE. *, Responses significantly different from resting values; or F, from pretraining values (P , 0.05).


3345

TABLE 1. Plasma responses of estrogens and catecholestrogens during exercise before and after a 5-day exhaustive training program. Results are expressed as percent changes from rest in the FPh and LPh Exercise intensity Hormone

E (%) Before After 2-OHE (%) Before After 2-MeOE (%) Before After 2-OHE:E (%) Before After 2-MeOE:2-OHE (%) Before After

t0

tsubmax

tmax

FPh

LPh

FPh

LPh

FPh

LPh

100 73 6 3.5b

100 47 6 2.7b

99 6 12.7 78 6 3.6b

105 6 11.2 49 6 2.6b

112 6 10.9a 87 6 7.6b

119 6 20.0 55 6 2.9a,b

100 96 6 9.5

100 76 6 7.2b

103 6 9.7 90 6 8.6b

100 6 9.1 76 6 6.9b

106 6 9.8 92 6 9.5b

100 6 9.7 79 6 8.0b

100 97 6 9.5

100 91 6 5.7b

100 6 12.1 101 6 9.1

100 6 10.3 92 6 7.9b

107 6 20.6 98 6 15.8

104 6 10.8 93 6 8.3b

100 106 6 7.5

100 118 6 10.8b

106 6 12.5 94 6 8.8a,b

94 6 9.2 118 6 9.2b

106 6 11.3 82 6 7.5a,b

82 6 9.2b 106 6 10.8a,b

100 111 6 22.2

100 128 6 20.0b

101 6 10.3 119 6 19.3b

106 6 13.0 119 6 14.0b

101 6 11.9 112 6 16.7b

103 6 12.5 126 6 19.5b

N 5 9; pretraining hormone concentration at rest for each menstrual phase 5 100%; values are means 6 SE. P , 0.05; a, Acute exercise-induced change from rest, significantly different; b, posttraining percent change significantly different from pretraining value at the same exercise intensity.

TABLE 2. Plasma responses of LH and catecholamines during exercise before and after a 5-day exhaustive training program Exercise intensity Hormone

LH (mIU/mL) Before After NE (pg/mL) Before After Epi (pg/mL) Before After DA (pg/mL) Before After

t0

tsubmax

FPh

LPh

3.6 6 0.4 4.4 6 0.3

4.8 6 1.0 2.9 6 0.3a

178 6 17 253 6 83

tmax

FPh

LPh

FPh

LPh

167 6 49 237 6 25a

763 6 261b 746 6 80b

338 6 93b 509 6 91a,b

1,411 6 289b 2,812 6 680a,b

2,218 6 508b 2,881 6 417a,b

22 6 3 39 6 6

44 6 8 42 6 6a

77 6 32b 80 6 11b

66 6 13b 62 6 8b

293 6 100b 301 6 196b

49 6 16 17 6 4a

,10 21 6 6a

36 6 8 32 6 6b

,10 34 6 8a

52 6 12 103 6 39a,b

454 6 127a,b 337 6 71b 29 6 12b 30 6 7

Results are expressed as means 6 SE. Pretraining values are reprinted from De Creé et al. (5, 6) and appear by permission of the American Physiological Society. P , 0.05. a Posttraining value significantly different from pretraining value at the same exercise intensity. b Acute exercise-induced change from rest significantly different.

reduced (P , 0.05) during the LPh, the O-methylated plasma CE concentrations only decreased very modestly (tsubmax: 28%, tmax: 27%). Both before and after training, the 2-OHE:E ratios at maximal exercise intensity were lower than resting values during the same cycle phase. Before training, the 2-OHE:E ratio was always lower during the LPh than during the FPh, whereas after training, FPh ratios were always lower than those of the LPh. CE activity, as derived from the 2-MeOE:2-OHE ratios, significantly (P , 0.05) increased in response to training. Plasma NE and EPI levels progressively rose in response to incremental exercise (Table 2). Increases were significant both at submaximal (P , 0.05) and maximal (P , 0.001) exercise intensity in either phase. Acute exercise-induced responses in plasma NE were higher after training (2, 811 6 680 vs. 1, 411 6 289 pg/mL for the FPh). We were unable to

identify a consistent picture for DA levels, although after training, there was a trend for plasma levels to increase proportionally to the exercise intensity, but only in the FPh. Discussion Training effects on performance during acute exercise

Before training, the V˙O2max was significantly lower during the LPh, when compared with the FPh. This is in line with many other studies (11), although some authors have reported no cycle phase effects (12–15), or higher V˙O2max values, in the LPh (16). Lower V˙O2max values in the LPh have generally been attributed to the higher progesterone levels, which are postulated to relatively increase ventilation for a given exercise intensity. After training, this cycle phase difference in VO2max values disappeared.

3346

´ E ET AL. DE CRE

Training effects on primary E responses to acute exercise

Interindividual differences for E concentrations are much higher than for the unconjugated E2 and E1 fractions. Similar to previously documented E2 responses (12, 16 –18), plasma E levels increase with incremental exercise, and this in both menstrual phases, before and after training. Increasing plasma estrogen concentrations during incremental exercise are assumed to be caused by an approximately halved hepatic MCR and splanchnic flow (19). Most studies that have measured estrogens during exercise after physical training have also found that the percent increases in the concentration of female sex hormones during a certain absolute exercise intensity, will be lower than before training (16, 20). This finding is confirmed by the present study, as far as exercise during the LPh is concerned. Training effects on CE levels during acute exercise

No reports on the effects of training on 2-hydroxy CE during acute exercise have been published before. Our results show that acute exercise stress exerts little effect on plasma 2-OHE and 2-MeOE levels. This was not any different after a period of brief, exhaustive training. Plasma concentrations of both 2-OHE and 2-MeOE were reduced at all intensities after training, but reductions in 2-MeOE (210%) were smaller than for 2-OHE (225%). Although, this finding can be interpreted as a decrease in both levels and O-methylation of CE, possibly it could also mean that the proportion of CE that is O-methylated, increased by approximately 15%. Training effects on CE formation and activity during acute exercise

The ratio of the end product divided by its precursor gives an estimation of the proportion formed, as suggested previously (5). Acute exercise stress had little effect on 2-OHE:E and 2-MeOE:2-OHE ratios before or after training. However, the O-methylation of CE apparently increases in response to training, particularly in the LPh, as suggested by both absolute and percent increases in 2-MeOE:2-OHE ratios. It must be admitted, though, that the use of these ratios to estimate formation and methylation is not without problems. For example, we do not have any firm proof that changes in these ratios are not a nonspecific effect of changes in estrogen levels. We also do not know whether the hydroxylation and methylation reactions are not relatively saturated at a higher substrate concentration, in which case, everything that lowers the substrate concentration would increase the ratio of product to substrate. This is a substantial problem, and introducing radiolabeled compounds might offer additional information on in vivo CE formation and methylation. Hypothesis 1: effects of training on CE formation

The hypothesis that training facilitates C-2 hydroxylation, largely depends on the results of a single study. Russel and co-authors (2) determined resting levels of plasma (unconjugated) 2-OHE in well-trained oligomenorrheic swimmers. Although, their results must be interpreted with caution (blood treatment procedures, no use of ascorbic acid to prevent oxidative decomposition of CE, large variations in age

JCE & M • 1997 Vol 82 • No 10

of the samples, etc.), these authors found significantly increased resting plasma concentrations of 2-OHE in response to (a much longer period of) intense physical training. Our results apparently are difficult to reconcile with their findings, because we found reduced levels of 2-OHE. However, do reduced plasma levels necessarily mean that formation of the product is impaired? Not if, for example, there is either a concomitant increase in MCR, or less available precursor. To help us decide if the results from both studies are indeed irreconcilable, the actual CE formation and O-methylation are crucial. From a close examination of our results, we learn that the magnitude of the posttraining decreases in plasma 2-OHE or 2-MeOE was generally smaller than that of their precursor. This might mean that, in spite of lower posttraining plasma levels, there is still a proportionate increase in both formation and methylation of CE. What about the previous studies? Although no other authors have reported on formation and activity of 2-hydroxy CE in association with exercise, the second study by Russel et al. (4) provided enough data to allow calculation of the 2-OHE:E ratios. In controls, the ratio was 0.13 6 0.03; in swimmers training 55 km/week, 0.77 6 0.30; and in swimmers doing 90 km/week, 1.20 6 0.21. So, clearly the formation of 2-hydroxy CE seems to increase in response to training. This led us to make the following suggestion for a modification of the initial hypothesis: Based on our own and previous findings, we speculate that, perhaps, training indeed induces an increase in C-2 hydroxylation. But the absolute concentrations of CE possibly may only increase in the presence of a substantially low level of precursor (estrogen), as was the case in the hypoestrogenemic subjects included in the study by Russel et al. (4). We speculate that, as long as the subject is euestrogenemic, a higher formation of CE may not be accompanied by increases in absolute levels of CE. If this indeed is true, then the results of both studies are perfectly reconcilable. Training effects on CA responses to acute exercise

There is a consensus that sympathoadrenal activity, as indicated by NE and EPI levels, proportionately increases with the intensity of the exercise (16, 21–23). Increases in NE during exercise have been found to be closely associated to oxygen availability; the higher the oxygen saturation, the lower plasma NE concentrations (22). It is suspected that changes in secretion and metabolism of reproductive hormones, as well as substrate use, are modulated also by the sympathoadrenal system. It has been demonstrated that training rapidly reduces acute NE responses to a given absolute exercise intensity [in men (22, 23), although similarly detailed studies are not available in women]. Nevertheless, a higher V˙O2max and maximal exercise intensity, as a result of training, still may be accompanied by a higher circulating NE level in the end. This is suggested also by our results. Hypothesis 2: effects of training on CE formation & menstruation

Although still poorly understood, CE may play a physiological role in the control of the menstrual cycle through


various mechanisms (for a review, see Ref. 3). Unfortunately, the activities of CE seem very complex and literature findings inconsistent. A dozen studies, mostly dating back to the 1980’s, suggested that CE mediate the release of LH (3, 24, 25). Recent investigations have revealed that such intracerebral effects of CE seem to involve NE. Reznikov and Rozenko (26) were able to demonstrate that peripheral administration of CE in rats leads to a significant increase in intrahypothalamic NE content. Several authors have shown that LH secretion is directly mediated by NE (27–29). These findings strengthen the speculations of a physiological role for CE in the control of the menstrual cycle. As known, the inactivation of CE and CA is catalyzed by the same enzyme, COMT. However, the affinity of COMT for CE is higher than for NE. We postulate that in both cases, eu- and hypoestrogenemic exercising women, there is a proportionately higher O-methylation of CE, which may lead to an increase in NE concentrations and possibly interfere with NE-mediated processes, such as gonadotropin release. Certainly, this hypothesis remains susceptible to criticism. First, there is abundant evidence in favor of the classical theory, assuming that exerciseinduced hypoestrogenemia is a consequence of gonadotropin dysfunction (30) and not vice versa. However, it also must be avowed that scientists have mainly investigated the classical pathway, rather than the reverse mechanism. Second, a series of studies by Merriam and collaborators (31–33) addressed two extremely important points that are relevant to the central hypothesis of the current paper. These authors argued that peripheral concentrations of CE are too low, and their MCR too high, to exert their effects centrally to affect gonadotropin secretion (31). However, one cannot rule out the possibility that during exercise, there may be processes going on in the brain that are parallel to those that we have observed in the periphery. Furthermore, there is evidence that administration of CE, results in a specific decrease in LH in both men (32) and women (33). These studies support a possible role for CE in the mediation of gonadotropin release, irrespective of whether CE function as estrogens in a negative-feedback mode or via noradrenergic modulation of GnRH, as suggested in our hypothesis. Finally, because of the limitations of peripheral measurements for making inferences about intracerebral processes, we advocate additional investigations, using infusions or antagonists, to further elucidate the role of CE in the etiology of exercise-related hypoestrogenemia. Conclusion

In conclusion, our findings suggest that in previously untrained eumenorrheic women, training leads to decreased plasma levels of estrogens, 2-OHE, and 2-MeOE. This reduction is most pronounced during the LPh. It, however, is not certain that, despite these lower plasma levels, there is also a reduced formation and activity of CE. As far as can be derived from the end product to precursor ratio, we believe that training, irrespective of plasma levels, increases the proportion of CE formed, as well as their activity (O-methyl-

3347

ation). These findings may be of significance for NE-mediated effects on gonadotropin release. Acknowledgments We are indebted to all women who kindly participated in this study. We gratefully acknowledge the help of Alex Vermeulen, M.D., Ph.D., of the Department of Endocrinology and Metabolism of the University of Ghent, and Michel Ostyn, M.D., of the Institute of Physical Education of the Catholic University of Leuven, Belgium, for their scientific advice as to the experimental design and preparation of this study. The authors also gratefully acknowledge the technical assistance of Kerstin Mannheimer (Medical University of Lu¨beck); Yvonne Janssen, M.B.; and Monique Van Der Heyden, M.B. (University of Maastricht). Furthermore, we express our gratitude to Alicja Bialkowska, M.Eng., for the computer tables and graphs; Kristin Hibler, M.A. (Department of Speech Communication, University of Washington, Seattle, WA), and Gary Hibler, Ph.D. (Portland, OR), for the proofreading of the manuscript and for their critical comments.

References 1. Keizer HA, Rogol AD. 1990 Physical exercise and menstrual cycle alterations. What are the mechanisms? Sports Med. 10:218 –235. 2. Russel JB, Mitchell DE, Musey PI. 1984 The role of b-endorphin and catechol estrogens on the hypothalamic-pituitary axis in female athletes. Fertil Steril. 42:690 – 695. 3. De Creé C. 1990 The possible involvement of endogenous opioid peptides and catecholestrogens in provoking menstrual irregularities in women athletes. Int J Sports Med. 11:329 –348. 4. Russel JB, Mitchell DE, Musey PI. 1984 The relationship of exercise to anovulatory cycles in female athletes: hormonal and physical characteristics. Obstet Gynecol. 63:452– 456. 5. De Creé C, Ball P, Seidlitz B, Van Kranenburg G, Geurten P, Keizer HA. Effects of a training program on resting plasma 2-hydroxycatecholestrogen levels in eumenorrheic women. J Appl Physiol. In press. 6. De Creé C, Ball P,Seidlitz B, van Kranenburg G, Geurten P, Keizer HA. Effects of a training program on resting plasma 2-hydroxycatecholestrogen levels in eumenorrheic women. J Appl Physiol. In press. 7. Durnin JV, Rahaman MH. 1967 The assessment of the amount of fat in human body from measurement of skinfold thickness. Br J Nutr. 21:681– 687. 8. Keizer HA, Platen P, Menheere PP. 1989 The hypothalamic/pituitary axis under exercise stress: the effects of aerobic and anaerobic training. In: Rogol A, Laron Z, eds. Hormones and sport. Serono symposium publications vol. 55. New York: Raven Press; 101–105. 9. Rogol AD, Weltman A, Weltman JY. 1992 Durability of the reproductive axis in eumenorrheic women during 1 yr of endurance training. J Appl Physiol. 72:1571–1580. 10. Dill DB, Costill DL. 1979 Calculation of percentage changes in volumes of blood, plasma, and red cells in dehydration. J Appl Physiol. 37:247–248. 11. Lebrun CM, McKenzie DC, Prior JC. 1995 Effects of menstrual cycle phase on athletic performance. Med Sci Sports Exerc. 27:437– 444. 12. Keizer HA, Kuipers H, De Haan J. 1987 Multiple hormonal responses to physical exercise in eumenorrheic trained and untrained women. Int J Sports Med. [Suppl]3:139 –150. 13. Bemben DA, Salm PC, Salm AJ. 1995 Ventilatory and blood lactate responses to maximal treadmill exercise during the menstrual cycle. J Sports Med Phys Fitness. 35:257–262. 14. De Souza MJ, Maguire MS, Rubin KR. 1990 Effects of menstrual phase and amenorrhea on exercise performance in runners. Med Sci Sports Exerc. 22:575–580. 15. Hall-Jurkowski JE, Jones NL, Toews CJ. 1981 Effects of menstrual cycle on blood lactate, O2 delivery, and performance during exercise. J Appl Physiol. 51:1493–1499. 16. Lavoie J-M, Dionne N, Helie R. 1987 Menstrual cycle phase dissociation of blood glucose homeostasis during exercise. J Appl Physiol. 62:1084 –1089. 17. Bunt JC, Bahr JM, Bemben DA. 1987 Comparison of estradiol and testosterone levels during and immediately following prolonged exercise in moderately active and trained males and females. Endocr Res. 13:157–172. 18. Jurkowski JE, Jones NL, Walker WC. 1978 Ovarian hormone response to exercise. J Appl Physiol. 44:109 –114. 19. Keizer HA, Poortman J, Bunnik GS. 1980 Influence of physical exercise on sex-hormone metabolism. J Appl Physiol. 48:765–769. 20. Keizer HA, Kuipers H, De Haan J. 1987 Effect of a 3-month endurance training program on metabolic and multiple hormonal responses to exercise. Int J Sports Med. [Suppl]3:150 –160. 21. Chin WN, Chang FE, Dodds WG. 1987 Acute effects of exercise on plasma catecholamines in sedentary and athletic women with normal and abnormal menses. Am J Obstet Gynecol. 157:938 –944.

3348

´ E ET AL. DE CRE

22. Galbo H. 1986 Autonomic neuroendocrine responses to exercise. Scand J Sports Sci. 8:3–17. 23. Mc Murray RG, Forsythe WA, Mar MH. 1987 Exercise intensity-related responses of b-endorphin and catecholamines. Med Sci Sports Exerc. 19:570 –574. 24. Ladosky W, Azambuja HC, Schneider HT. 1983 Possible mechanism of action of 2-hydroxylated estradiol on the possible feedback control for LH release in the rat. J Steroid Biochem. 19:639 – 644. 25. Adashi EY, Rakoff J, Divers W. 1980 The effect of acutely administered 2-hydroxyestrone on the release of gonadotropins and prolactin before and after estrogen priming in hypogonadal women. Obstet Gynecol Surv. 35:363–365. 26. Reznikov AG, Nosenko ND. 1995 Catecholamines in steroid-dependent brain development. J Steroid Biochem Mol Biol. 53:349 –353. 27. Barraclough CA, Wise PM. 1982 The role of catecholamines in the regulation of pituitary luteinizing hormone and follicle-stimulating hormone secretion. Endocr Rev. 3:91–119.

JCE & M • 1997 Vol 82 • No 10

28. Bergen H, Leung PC. 1986 Norepinephrine inhibition of pulsatile LH release: receptor specificity. Am J Physiol 250:E205–E211. 29. Kalra SP, Simpkins JW. 1981 Evidence for noradrenergic mediation of opioid effects on luteinizing hormone secretion. Endocrinology. 109:776 –782. 30. Veldhuis JD, Evans WS, Demers LM. 1985 Altered neuroendocrine regulation of gonadotropin secretion in women distance runners. J Clin Endocrinol Metab. 61:557–563. 31. Kono S, Merriam GR, Brandon DD. 1983 Radioimmunoassay and metabolic clearance rate of catecholestrogens, 2-hydroxyestrone and 2-hydroxyestradiol in man. J Steroid Biochem Mol Biol. 19:627– 633. 32. Merriam GR, Kono S, Keiser HR. 1981 Effect of catechol estrogen infusions upon gonadotropin and prolactin concentrations in men. J Clin Endocrinol Metab. 53:784 –789. 33. Merriam GR, Pfeiffer DG, Loriaux DL. 1983 Catechol estrogen and the control of gonadotropin and prolactin secretion in man. J Steroid Biochem Mol Biol. 19:619 – 625.