or Growth Hormone (GH)

5 downloads 0 Views 4MB Size Report
Catherine Kern for their technical assistance, and Dr. Robert Abbott for statistical advice. ... Russell-Jones DL, Weissberger AJ, Bowes SB, et al. 1993 The effects ...
0021-972X/97/$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright © 1997 by The Endocrine Society

Vol. 82, No. 11 Printed in U.S.A.

Short-Term Modulation of the Androgen Milieu Alters Pulsatile, But Not Exercise- or Growth Hormone (GH)Releasing Hormone-Stimulated GH Secretion in Healthy Men: Impact of Gonadal Steroid and GH Secretory Changes on Metabolic Outcomes* DAVID A. FRYBURG, ARTHUR WELTMAN, LINDA A. JAHN, JUDY Y. WELTMAN, EUGENE SAMOJLIK, RAYMOND L. HINTZ, AND JOHANNES D. VELDHUIS Division of Endocrinology and Metabolism, Department of Internal Medicine, and the General Clinical Research Center (D.A.F., A.W., L.A.J., J.Y.W., J.D.V.); National Science Foundation Center for Biological Timing (J.D.V.); and the Department of Human Services, University of Virginia Health Sciences Center (A.W.), Charlottesville, Virginia 22903; Endocrine Laboratory, Newark Beth Israel Medical Center, University of Medicine and Dentistry-New Jersey Medical School (E.S.), Newark, New Jersey 07112; and the Division of Endocrinology, Department of Pediatrics, Stanford University Medical Center (R.L.H.), Palo Alto, California 94305 ABSTRACT Gonadal steroids are known to alter GH secretion as well as tissue metabolism. The present study was designed to examine the effects of short term (2- to 3-week) alterations in gonadal steroids on basal pulsatile (nonstimulated) and exercise- and GH-releasing hormone-stimulated GH secretion, urinary nitrogen excretion, and basal and exercisestimulated oxygen consumption. Two protocols were conducted, which reflect a total of 18 separate studies. In the first paradigm, 5 healthy young men were each studied in a double blind, randomized manner during 3 different gonadal hormone manipulations, in which serum testosterone was varied from hypogonadal (induced by leuprolide) to eugonadal (saline injections) to high levels (testosterone enanthate, 3 mg/kgzweek, im). There was a washout period of 8 weeks between treatments. In the second protocol, 3 of the original subjects were studied after 2 weeks of treatment with stanozolol (0.1 mg/kgzday). Two to 3 weeks after the desired changes in serum testosterone, each subject was admitted to the General Clinical Research Center for study. The hypogonadal state (serum testosterone, 33 ng/dL) increased urinary nitrogen loss (by 34%; P , 0.005) and decreased basal metabolic rate (by 12%; P , 0.02) compared with the eugonadal state (testosterone, 796 ng/dL). High dose testosterone (1609 ng/dL) further decreased urinary nitrogen loss over the eugonadal state (by 16%; P , 0.05). Stanozolol yielded the lowest urinary nitrogen excretion of all (P , 0.03). Like urinary nitrogen, the basal metabolic rate showed the greatest change between the hypogonadal and eugonadal states (12%; P , 0.02), with a lesser change during high dose testosterone treat-

ment (4%). Analogously, end-exercise oxygen consumption rose by 11% between the hypogonadal and eugonadal states (P , 0.05). Between the hypogonadal and eugonadal states, no significant changes in pulsatile (nonstimulated), exercise-stimulated, or GRFstimulated GH secretion or serum insulin-like growth factor I concentrations were observed. Raising testosterone to supraphysiological levels increased pulsatile GH secretion by 62% over that with leuprolide and by 22% over that with saline (P , 0.05). High dose testosterone treatment also increased serum insulin-like growth factor I concentrations by 21% and 34% over those during the eugonadal and hypogonadal states, respectively (P , 0.01). Testosterone did not affect either exercise- or GRF-stimulated GH secretion. In protocol 2, stanozolol did not affect any parameter of GH secretion. To examine the interaction between GH secretion and testosterone on urinary nitrogen excretion and basal metabolic rate, a one-way analysis of covariance was undertaken. Statistical examination of GH production as the covariate and testosterone (by tertile) as the interactive factor demonstrated significant relationships between serum testosterone levels and either urinary nitrogen (P , 0.02) or basal metabolic rate (P , 0.01), but not GH secretion (P 5 NS). In summary, these results demonstrate that short term modulation of the androgen milieu affects metabolic outcome without necessitating changes in GH secretion. These results have significance for both normal physiology and for the treatment of hypogonadal GH-deficient patients. (J Clin Endocrinol Metab 82: 3710 –3719, 1997)

T

muscle) mass and loss of fat mass (1– 4). In some experimental circumstances, androgenic steroids also increase the basal metabolic rate (BMR) (5, 6). Despite these well demonstrated effects, the metabolic actions of gonadal steroids have been intertwined with and potentially confounded by the effects of GH, as gonadal steroids stimulate GH secretion (7–12) and increase circulating insulin-like growth factor I (IGF-I) concentrations (13). Indeed, inasmuch as GH promotes both protein anabolism (14 –18) and increases BMR (19), the metabolic actions of gonadal steroids may be partly ascribable to their indirect effects on GH secretion. Previous studies of the actions of testosterone on the GHIGF axis have shown that testosterone treatment in either

HE ANABOLIC properties of androgenic steroids have long been recognized. In both animals and humans, these hormones promote accrual of lean body (principally Received December 11, 1996. Revision received February 21, 1997. Rerevision received July 16, 1997. Accepted July 23, 1997. Address all correspondence and requests for reprints to: David A. Fryburg, M.D., Clinical Research, Pfizer Central Research, Eastern Point Road, Groton, Connecticut 06340. E-mail:david_a_fryburg@groton. pfizer.com. * This work was supported by USPHS Grants AR-01881, DK-38578, AG-147991 and RR-00847 to the University of Virginia General Clinical Research Center; the National Science Foundation Center for Biological Timing; and research grants from the Muscular Dystrophy Association and Eli Lilly, Inc.

3710

TESTOSTERONE MODULATION OF GH SECRETION AND METABOLIC OUTCOMES

hypogonadal or eugonadal subjects increases GH secretion and serum IGF-I concentrations. Analogously, suppression of gonadal steroid levels in children with precocious puberty decreases GH secretion (20). In many previous studies, the manipulation of gonadal steroids lasted 12 weeks or more before outcome measurements (21, 22). Given this background, the present study was designed to test several hypotheses regarding the interactions between gonadal steroids and GH secretion in healthy young men; namely, 1) short term (2- to 3-week) variations in serum testosterone levels (from low to high) modulate 24-h, exercise-stimulated, and GH-releasing hormone (GHRH)-stimulated GH secretion in healthy young men; 2) gonadal steroids regulate BMR and urinary nitrogen excretion via their complex interaction with GH; and 3) stanozolol, a nonaromatizable anabolic steroid, does not reproduce all of the actions of testosterone. Subjects and Methods Subjects Five healthy young adult men (aged, 21–26) were recruited for this study. All subjects were within 20% of ideal body weight for height, and none was taking medication other than that prescribed for the study. Before participation, each subject carefully reviewed and agreed to the protocol and consent form, which was approved by both the human investigation committee and the General Clinical Research Center (GCRC) advisory committee of the University of Virginia.

Protocol Two protocols are described. The goal of protocol 1 was to examine the effect of variations in serum testosterone levels on GH secretory dynamics as well as on other selected relevant hormonal and metabolic parameters. To do so, each subject was studied three times under conditions of low, normal, and high testosterone in a randomized, double blind fashion. Randomization assignment and drug preparation were conducted by the investigational pharmacist of the University of Virginia. Figure 1 illustrates the paradigm employed for protocol 1. To induce a hypogonadal state, each subject received leuprolide acetate (7.5 mg, im depot; Lupron, Abbott Laboratories, North Chicago, IL) at week 1. Subsequently, each subject returned weekly to the GCRC to receive saline injections. Blood samples for gonadal steroid hormone measurements were acquired at each weekly visit. On week 5, the subject received the last injection and was admitted to the GCRC for the in-patient portion of the protocol (see below). For the eugonadal study, each subject was treated only with saline for the 5 weeks and then admitted to the GCRC. For the high testosterone treatment arm, each subject was given im injections of saline for weeks 1 and 2, followed by im injections of testosterone enanthate (3 mg/kg)

FIG. 1. Schedule of out-patient injections for preparation of subjects. To create the hypogonadal state, leuprolide was injected on the first day, followed by saline injections. For the eugonadal state, only saline was given throughout the study. To make the duration of gonadal steroid manipulation equal in duration in the high and low testosterone states, saline was injected for the first 2 weeks, followed by 3 weeks of testosterone injections. Subjects were admitted the day of the last injection.

3711

at weeks 3, 4, and 5. The rationale for this approach was to be sure that the duration of hypogonadism was equal to that of hyperandrogenemia. That is, it is well known that leuprolide first induces an increase in gonadal steroid production before suppressing gonadal steroid production (23). Based upon preliminary experiments undertaken for this project (n 5 8 subjects) as well as published data, the testosterone injections were limited to 3 weeks, which matched the time the subjects were hypogonadal. For the entire course of the study, each subject was counseled to remain on a stable exercise regimen and weight-maintaining diet. The same in-patient protocol was performed under each of the (above) three conditions. Generally, the sequence of testing included exercise-stimulated GH secretion (day 1), 24-h nonstimulated GH secretion (day 2), and GHRH-stimulated GH secretion (day 3). Specifically, each subject was admitted to the GCRC the afternoon of day 0, before beginning the in-patient sampling paradigm. Each subject received the same weight maintenance diet during the GCRC stay. Sleep/wake times were 2300/0630 h, respectively. Except for underwater weighing, each subject was ad libitum activity only at the GCRC. At the time of admission (day 0), the last testosterone/saline injection was given, and an iv catheter was inserted for the withdrawal of blood samples. On day 1, after an overnight fast and bed rest, blood samples were withdrawn every 5 min for GH measurements for 1 h (0700 – 0800 h) before the subject exercised from 0800 – 0830 h on a cycle ergometer at an exercise intensity associated with a blood lactate concentration of 2.5 mmol/L. This exercise intensity has been previously shown to stimulate GH secretion (24). The blood lactate and VO2 peak response to exercise were predetermined in each subject in the eugonadal state during the recruitment phase of the study. With the subsequent three admissions, each subject exerted the same power output. Samples for GH secretion were withdrawn during the exercise bout (every 5 min) as well as for 3 h after completion of the exercise (every 10 min). After the subject had rested and eaten, later that afternoon, determination of body density by hydrostatic weighing (25) with residual lung volume measurement were performed. The computational procedure of Brozek et al. (26) was used to convert body density to percent body fat, fat mass, and fat-free mass. The following morning at 0600 h, indirect calorimetry was performed over 30 min (Delta Trac, Sensor Medics, Anaheim, CA) after the subject awoke and lay in bed in a temperature-stable environment for approximately 1 h. At 0800 h, quantification of 24-h GH secretion was initiated. Samples for GH determination were collected every 10 min through 0750 h on day 3. After the last GH sample at 0750 h on day 3, each subject received an iv bolus of GRF at a dose of 1 mg/kg (Serono). Sampling for GH continued every 10 min for the next 4 h. After the post-GRF sampling was completed, the iv lines were removed, and the subject was discharged. During the admission, blood samples were also collected and pooled on day 2 for total testosterone and estradiol (hourly), total IGF-I, and IGF-binding protein-3 (IGFBP-3) measurements (every 6 h). Samples for sex hormone-binding globulin (SHBG), dihydrotestosterone (DHT), free testosterone, and free estradiol determinations were also obtained from pooled day 2 plasma. Urine for nitrogen and creatinine were collected

3712

FRYBURG ET AL.

JCE & M • 1997 Vol 82 • No 11

in two 24-h aliquots from days 1–3 while the subjects were consuming a defined constant diet that was provided during each admission. There was an 8-week washout period between each of these three treatment protocols. Given the results observed in the first protocol, 3 of the subjects were also studied in a second protocol after treatment with stanozolol (0.1 mg/kgzday, orally) for the same length of time that testosterone was administered. After this treatment, each subject underwent the same protocol. This report reflects a total of 18 studies in these 5 subjects.

Assays All samples were processed in batch by subject (i.e. all samples from each of the three treatment states were run in the same assay on the same day). GH was assayed with a commercially available immunoradiometric assay (Nichols Institute, San Juan Capistrano, CA) that has a lower limit of detection of 0.2 mg/L. Total plasma IGF-I concentrations were measured by RIA after acid chromatographic separation (27). IGFBP-3 was determined using a two-site noncompetitive immunoradiometric assay (Diagnostic Systems Laboratory, Webster, TX). Total and free serum testosterone and estradiol concentrations and serum DHT concentrations were measured using previously described methods (28, 29). Urinary nitrogen was determined by the Kjeldahl method.

Hormone secretion analysis Deconvolution analysis. A multiparameter deconvolution technique was used to estimate specific measures of GH secretion from the exercise, GHRH, and 24-h serum GH concentration profiles (30). For preliminary estimates of secretory burst number, amplitude, and timing, a waveform-independent deconvolution methodology (Pulse2) was used, which assumed zero basal secretion and a nominal endogenous GH half-life of 14 –18 min (31, 32). The multiparameter deconvolution methodology (DECONV) was then applied using these initial estimates, as described previously (33). We estimated the following specific measures of GH secretion: secretory burst frequency (number of statistically significant secretory pulses per 24 h), amplitude (maximal rate of calculated GH secretion attained within a release episode), mass (integral of the calculated secretory pulse or the amount of hormone secreted per burst/ unit distribution volume), the half-duration (minutes), and the endogenous GH half-life (minutes). The daily pulsatile GH production rate was computed as the product of the mean GH secretory burst mass and frequency. Statistical analysis. The randomization code was broken after all subjects’ samples had been processed and deconvolved. Statistical comparisons for all five subjects for the leuprolide-, saline-, and testosterone-treated states were then made via ANOVA for repeated measures, and post-hoc comparisons were made using Duncan’s test. Interactions between gonadal steroid modulation and GH secretion on BMR and urinary nitrogen excretion were tested with one-way analysis of covariance (ANCOVA). As indicated, paired comparisons were made to contrast the effects of stanozolol with those of the other gonadal steroid hormone states. Statistical analysis was conducted with True Epistat statistical software (Epistat Services, Richardson, TX).

Results Time course for the manipulation of serum testosterone concentrations

Figure 2 depicts the time course of total serum testosterone levels during the preparation period for each subject, as determined by single morning testosterone samples. The week 1 data point summarizes the initial testosterone value for each subject at the beginning of the leuprolide, saline, or testosterone treatment protocols. At week 1, subjects started each arm of the study with approximately the same total serum testosterone. For the leuprolide-treated subjects, total testosterone levels rose slightly at week 2 and then declined by week 3, approaching nearly undetectable levels by week

FIG. 2. Time course of serum testosterone concentrations during the leuprolide, saline, and testosterone treatments.

4. During treatment with saline alone, testosterone concentrations remained fairly stable. Testosterone treatment, initiated after the week 3 sample was withdrawn, rose to approximately 1500 ng/dL and remained at that level for the remainder of the study. As depicted in Fig. 2, the duration of the hypogonadal period approximated that of the high testosterone period. Serum concentrations of total and free testosterone, estradiol, DHT, and SHBG during the GCRC admission

Figure 3 summarizes and contrasts the resulting serum concentrations of total and free testosterone and estradiol as well as DHT and SHBG. As evident in the figure, the gonadal steroid manipulations and stanozolol treatment elicited significant alterations in total testosterone concentrations comparable to those presented in Fig. 2 (Fig. 3a; by ANOVA, P , 0.0001). Parallel changes were observed in total estradiol (Fig. 3b; by ANOVA, P , 0.0001) as well as free testosterone (Fig. 3c; by ANOVA, P , 0.0001), free estradiol (Fig. 3d; P , 0.01), and DHT (Fig. 3e; P , 0.0001). Post-hoc comparisons are also indicated in Fig. 3. In addition to the expected shifts in total and free hormone levels, several observations are important to emphasize. First, with regard to the effects of stanozolol, total testosterone and DHT decreased to levels below those observed in the salinetreated group (total testosterone: P , 0.06, stanozolol vs. saline; DHT: P , 0.03, stanozolol vs. saline), changes that were associated with the marked decrease in SHBG to undetectable levels (P , 0.001, stanozolol vs. saline). Second, despite the decrease in total testosterone and SHBG concentrations, free testosterone levels remained unchanged as did total and free estradiol levels compared with those during saline treatment (Fig. 3). Finally, despite fairly parallel alterations in total and free hormone levels, the relationship between total testosterone and estradiol and free testosterone and estradiol diverged during this study. That is, across treatments, the total testosterone/total estradiol ratio increased from 24 6 1 (leuprolide) to 180 6 40 (saline; P , 0.001), but did not change further with testosterone treatment (189 6 19). In contrast, the free testosterone/free estradiol ratio increased from 8 6 1 (leuprolide) to 130 6 15 (saline) and further to 235 6 19 (by ANOVA, P , 0.0001; P , 0.0001, leuprolide vs. saline; P , 0.001, saline vs. testosterone).

TESTOSTERONE MODULATION OF GH SECRETION AND METABOLIC OUTCOMES

3713

FIG. 3. Effects of gonadal steroid manipulation on serum total and free testosterone and estradiol, DHT, and SHBG concentrations. All data are expressed as the mean 6 SEM. ANOVA results are indicated in the upper corner of each figure. Post-hoc analysis is indicated by superscript letters. Letters that differ from one another denote significantly different means at P , 0.05.

Metabolic responses to gonadal steroid manipulation: urinary nitrogen excretion/BMR/body composition

Leuprolide’s suppression of gonadal steroid levels resulted in the most marked urinary excretion of nitrogen to 8.66 6 1.47 g nitrogen/g creatinine, which diminished in a step-wise manner to 6.42 6 1.64 (saline) and 5.37 6 0.73 (testosterone). As shown in the upper panel of Fig. 4, these changes were statistically different from one another (P , 0.005, by ANOVA; leuprolide vs. saline, P , 0.01; saline vs. testosterone, P , 0.05; leuprolide vs. testosterone, P , 0.001). Compared with subjects receiving stanozolol (4.6 6 0.7 g nitrogen/g creatinine), these changes were also significant (P , 0.002, by ANOVA; P , 0.02, stanozolol vs. saline or leuprolide). Compared with the effects of testosterone, stanozolol yielded a slight, but statistically insignificant, improvement in nitrogen loss (P , 0.08). Fat-free mass also tended to increase in this short time frame [65.8 6 10.1 kg (leuprolide) vs. 66.9 6 10.5 kg (saline) vs. 67.9 6 10.8 kg (testosterone)], but not significantly (by ANOVA, P 5 0.22). Like urinary nitrogen loss, modulation of gonadal steroid levels also affected oxygen consumption and calculated BMR (Fig. 4, middle panel). As observed for changes in urinary nitrogen loss between the hypogonadal to the eugonadal state, VO2 increased by 11% from 236 6 29 (leuprolide) to 262 6 25 (saline) and to 272 6 23 (testosterone) and 265 6 30 (stanozolol) mL O2/min (P , 0.01, by ANOVA; P , 0.02, leuprolide vs. saline or testosterone; P 5 NS, saline vs. testosterone). Similar conclusions were made for the BMR, which increased from 1628 6 90 (leuprolide) to 1818 6 78

(saline) to 1898 6 104 (testosterone) and 1817 6 130 (stanozolol) Cal/24 h (P , 0.01, by ANOVA; P , 0.02, leuprolide vs. saline or testosterone; P 5 NS, saline vs. testosterone). This increase was still present when energy expenditure values were normalized for fat-free mass (not shown). In addition to affecting basal oxygen consumption, modulation of testosterone levels also altered oxygen consumption at the end of exercise from 2.05 6 0.83 (leuprolide) to 2.32 6 0.83 (saline) to 2.47 6 0.99 (testosterone) L/min (P , 0.05, by ANOVA; P , 0.05, testosterone vs. leuprolide; P , 0.08, saline vs. testosterone). The response in the stanozolol group was indistinguishable from either saline or testosterone, but was significantly greater than observed in the same three subjects treated with leuprolide (2.11 6 0.32; P , 0.05). GH secretory dynamics

Unstimulated GH secretion. Figure 5 summarizes the changes in 24-h nonstimulated GH secretory dynamics in response to gonadal steroid manipulation. In response to altering testosterone levels, the mean serum GH concentration rose from 1.55 6 0.28 (leuprolide) to 1.92 6 0.38 (saline) to 2.67 6 0.38 mg/L (testosterone) (by ANOVA, P , 0.01; leuprolide or saline vs. testosterone, P , 0.03). Correspondingly, the area under the 24-h GH concentration curve rose from 2234 6 811 (leuprolide) to 2780 6 1146 (saline) to 4578 6 2629 (testosterone) mg/Lzmin (by ANOVA, P 5 0.05; leuprolide or saline vs. testosterone, P , 0.05; leuprolide vs. saline, P 5 NS). In contrast to testosterone treatment, stanozolol did not affect

3714

FRYBURG ET AL.

JCE & M • 1997 Vol 82 • No 11

FIG. 4. Effect of gonadal steroid manipulation on urinary nitrogen excretion, basal oxygen consumption, and exercise-induced oxygen consumption. All data are expressed as the mean 6 SEM. ANOVA results are indicated in the upper corner of each figure. Post-hoc analysis is indicated by superscript analysis. Letters that differ from one another denote significantly different means at P , 0.05.

the mean GH concentration (2.06 6 0.05 mg/L) or integrated GH release (2962 6 79 mg/Lzmin). Similar conclusions are obtained using deconvolution analysis. That is, short term reduction in serum gonadal steroids did not substantially alter GH production rates, whereas short term testosterone treatment elevated GH production by 22% over that with saline treatment and 62% over that with leuprolide treatment (leuprolide, 62 6 26; saline, 88 6 43; testosterone, 107 6 39 mg/Lzday (by ANOVA, P , 0.05; leuprolide vs. testosterone, P , 0.05; saline vs. testosterone, or leuprolide, P 5 NS). As observed for integrated GH, stanozolol treatment did not alter GH production rates (85 6 26 mg/L/day). The gradient in mean integrated GH concentrations or estimated production rates from leuprolide to testosterone was not due to either a change in the number of GH bursts per 24 h or the estimated half-life of GH (Table 1). Although the mass of GH secreted per burst did appear to increase (Table 1), this change was not statistically significant due to a large change during leuprolide treatment in one subject (P 5 0.17). Exclusion of this subject disclosed a significant increase in mass of GH secreted per burst (leu-

FIG. 5. Effect of gonadal steroid manipulation on mean 24-h serum GH concentrations, integrated serum GH concentration, and daily GH production rate, as estimated by deconvolution analysis. All data are expressed as the mean 6 SEM. ANOVA results are indicated in the upper left corner of each figure. Post-hoc analysis is indicated by superscript letters. Letters that differ from one another indicate significantly different means at P , 0.05.

prolide, 5.8 6 0.7; saline, 7.5 6 1.1; testosterone, 10.8 6 1.3 mg/L; P 5 0.02, by ANOVA). Stimulated GH secretion. Table 2 summarizes the measures of exercise and GHRH-stimulated GH secretion during the four different treatment states. All subjects achieved the same exercise intensity during each test (data not shown). As can be seen in Table 2, modulation of the gonadal steroid environment did not alter exercise-stimulated GH secretion. Similarly, modulation of gonadal steroids did not alter GHRHstimulated GH release. Serum IGF-I concentrations

As observed for 24-h GH secretion, serum total IGF-I concentrations were also altered by short term variations in gonadal steroid levels (Fig. 4). Between leuprolide and saline

TESTOSTERONE MODULATION OF GH SECRETION AND METABOLIC OUTCOMES

treatment, IGF-I concentrations did not change (314 6 24 to 348 6 16 mg/L), but increased significantly with testosterone treatment (421 6 29 mg/L; P , 0.01). Like 24-h GH secretion, stanozolol treatment did not affect total IGF-I concentrations (259 6 43 mg/L). IGFBP-3 concentrations, in contrast, did not change in leuprolide (3295 6 164 ng/mL) vs. saline (3287 6 124 ng/mL) vs. testosterone (3293 6 198 ng/mL) treatment conditions. Stanozolol, however, decreased IGFBP-3 to 2393 6 168 ng/mL (P , 0.01 vs. all other treatments). Contributions of gonadal steroids to GH secretion, BMR, and urinary nitrogen excretion

As one of the goals of this experiment was to determine the relative contributions of GH and gonadal steroids (in the setting of changing GH secretory responses to altered gonadal steroids), one statistical approach to this query is to examine outcome measures using one-way ANCOVA. In this instance, GH production is the covariate, and separation of testosterone levels by tertiles provides the interactive factors. This analysis reveals that when organized by tertile of testosterone (i.e. low, normal, or high), accounting for the variability due to GH production yielded a significant relation between urinary nitrogen and testosterone (P , 0.02). A similar ANCOVA with urinary nitrogen excretion as the outcome variable disclosed the same significant relationship between testosterone and BMR (P 5 0.01). To test the consistency of the ANCOVA strategy, if testosterone was made the covariate and GH production was the factor, no significant relationship was observed for either BMR or urinary nitrogen excretion and GH secretion (P 5 NS). Figures 6 and 7 graphically depict the ANCOVA statistical results. In both of these figures, the x- and z-axes segregate data by tertile of testosterone and 24-h pulsatile GH secretion. The first tertile for either indicates the lower third of measured values, and the third tertile indicates the highest. Thus, there are a total of nine cells in each graph. The value TABLE 1. Effect of gonadal steroid manipulation on number of GH secretory bursts, mass per burst, and estimated GH half-life

Bursts Mass/burst Half-life

Leuprolide

Saline

Testosterone

Stanozolol

864 10 6 9 22 6 1

10 6 2 963 23 6 6

963 12 6 4 23 6 6

12 6 1 762 26 6 7

Data are expressed as the mean 6 SEM. Variables: bursts (number of GH secretory events per 24 h; mass per burst), mean integrated area under each of the calculated secretory bursts (micrograms of GH/L distribution volume); and half-life (minutes).

3715

displayed as a column in each cell (y-axis) reflects the mean of urinary nitrogen/creatinine excretion or BMR measurements of subjects’ data sorted by these tertiles; this value is displayed on the y-axis. Note that there is one cell (testosterone tertile 1, GH production tertile 3) into which no subject’s data segregated. Figure 6 displays the corresponding ANCOVA for urinary nitrogen excretion. Viewing the data from the lowest to highest testosterone tertile within the lowest or middle GH secretion tertile suggests that increasing testosterone, even when GH secretion is low, substantially decreases urinary nitrogen loss. Conversely, viewing the data from the lowest to highest GH tertile does not yield the same visual conclusion as that for testosterone. Figure 7 depicts a similar finding, i.e. serum testosterone, not GH production, is the key determinant of the increase in BMR. Discussion

Several new observations are made in the present study. First, between the hypo- and eugonadal states, no changes in either 24-h GH secretion or serum IGF-I concentration were observed, in contrast to the increases seen with high testosterone concentrations. Second, ambient gonadal steroid concentrations in any of the three treatment states did not affect maximal GH responses to either GHRH or exercise. Third, despite the lack of changes in GH secretion between the hypo- and eugonadal states, significant changes in urinary nitrogen excretion, BMR, and end-exercise oxygen consumption were observed. Fourth, stanozolol (a potent nonaromatizable androgen) exerted significant metabolic effects, comparable to the effects of high testosterone, without substantially altering either GH secretion or IGF-I levels, although conclusions in this treatment state are limited by sample size. Application of statistical techniques to separate the relative contributions of testosterone and GH secretion disclosed that enhanced GH production per se had little effect on metabolic outcome, whereas ambient gonadal steroid levels were highly correlated with the changes in either urinary nitrogen loss or BMR. These observations suggest that it is the variation in gonadal steroids in the setting of an intact GH secretory axis that determines tissue metabolic outcome. Much work has preceded the present study demonstrating a complex relationship between gonadal steroids and GH secretion (7, 10, 13, 20, 34 –38), which has been well reviewed previously (11, 38). Briefly, contemporary thought on this topic states that testosterone, probably through peripheral conversion (via aromatization) to estradiol, positively regu-

TABLE 2. Effect of gonadal steroid manipulation on exercise and GHRH-stimulated GH secretion Leuprolide

Exercise Mean GH GH area GH production GHRH Mean GH GH area GH production

Saline

Testosterone

Stanozolol

2.1 6 1.9 1159 6 1003 28 6 25

1.2 6 0.6 652 6 310 22 6 12

3.0 6 1.2 1602 6 677 48 6 22

1.7 6 1.3 985 6 829 18 6 13

7.5 6 2.9 1867 6 717 52 6 22

4.5 6 3.8 1189 6 909 33 6 19

8.7 6 1.9 2037 6 384 68 6 18

4.0 6 1.5 988 6 379 29 6 18

Data expressed as the mean 6 SEM. Variables: mean GH (micrograms per L), GH area (micrograms per L min), and GH production (mass, micrograms per L released after secretagogue).

3716

FRYBURG ET AL.

JCE & M • 1997 Vol 82 • No 11

FIG. 6. Triaxial display of urinary nitrogen excretion (normalized for creatinine) segregated by tertiles of serum testosterone concentrations and 24-h pulsatile GH production rate. Tertiles run from low (51) to high (53) hormone levels. Movement from low to high serum testosterone tertiles discloses significant changes in urinary nitrogen excretion, whereas a similar pattern is not observed between the low and high GH production tertiles (by ANCOVA).

FIG. 7. Triaxial display of BMR segregated by tertiles of serum testosterone concentrations and 24-h pulsatile GH production. Tertiles run from low (51) to high (53) hormone levels. Movement from low to high serum testosterone tertiles discloses significant changes in BMR, whereas a similar pattern is not observed between the low and high GH production tertiles (by ANCOVA).

lates GH secretion in hypogonadal and eugonadal humans as well as in animals. Testosterone administration to either prepubertal boys or men increases serum IGF-I concentrations (13, 39), changes that can be antagonized by tamoxifen (36). Indeed, estradiol increases GH secretion in both humans and animals (10, 35, 37), particularly if it is administered orally. Transdermal estrogen, in contrast to oral preparations, may not always increase GH secretion putatively because this route of administration does not decrease serum IGF-I concentrations and, by autoregulation, stimulate GH release (10, 37). Only when transdermal estradiol doses are high does this route of administration cause an increase in GH secretion, and even then IGF-I levels are decreased (10, 37). In the present study in healthy young men, no significant change in GH secretion was detected between the short term (2- to 3-week) hypogonadal and eugonadal states, whereas high dose testosterone (of a magnitude that might be abused by athletes) significantly increased both GH secretion and circulating IGF-I concentrations. The lack of a significant

difference in GH secretion between the hypo- and eugonadal states in the present study may be due to either the short period of the manipulation (between 2–3 weeks) and/or the fact that these men had been previously eugonadal, in contrast to studies in hypogonadal men or prepubertal boys. Despite changes in unprovoked GH release, neither GHRH nor exercise-stimulated GH secretion was altered by modulation of the gonadal steroid environment. These observations are consistent with most, but not all, published data (40 – 42) and suggest that either these stimuli are not affected over the short term by gonadal steroids or that these maneuvers are potent enough to override a more subtle effect of testosterone/estradiol variation. To define possible changes induced by sex steroids in the sensitivity of the GH-IGF axis to secretagogues, lower doses of GHRH or less strenuous exercise regimens would probably be required, such as those performed by Dawson-Hughes and colleagues (8). Since Kochakian’s early observations that gonadectomy increases nitrogen (protein) loss, and testosterone replace-

TESTOSTERONE MODULATION OF GH SECRETION AND METABOLIC OUTCOMES

ment attenuates that loss or net catabolism (1), studies in both animals and humans have demonstrated that androgens or anabolic steroids increase body weight and lean body mass and often decrease fat mass (1, 3, 5, 21, 26, 34, 43– 45). The anabolic action on lean body mass was largely ascribable to an increase in muscle and was observed in hypogonadal male and female animals or eugonadal men given high dose testosterone or anabolic steroids (1, 3, 5, 21, 26, 34, 43– 45). Although some studies suggested that androgens increased protein mass via decreases in proteolysis (46), most investigations have reported that androgens augment muscle mass through increases in muscle protein synthesis without affecting degradation (5, 21, 43, 45, 47– 49). As a result, an earlier concept that androgens decrease glucocorticoidmediated increases in muscle protein degradation (46) has been discounted. Like androgens, in both adult animals and humans, GH increases fat-free mass and reduces corporal fat mass. In a parallel manner to androgens, GH promotes net protein accretion in humans through increases in both whole body (15, 16) and muscle protein synthesis (14, 16). Thus, both GH and androgens individually promote protein anabolism through increasing protein synthesis, which, if sustained for a sufficiently long period of time without commensurate rises in protein degradation, translates into increased muscle mass. Given the similarity in the mechanism of protein accrual and the effect of gonadal steroids on GH secretion, it has been postulated that tissue growth associated with heightened gonadal steroid levels was probably due to GH (11). In this interpretation, the overall growth acceleration of puberty, therefore, would depend upon gonadal steroids altering GH secretion, which, in turn, promoted linear growth and the accrual of protein (particularly muscle) mass. Interpretation of the direct and/or interactive effects of androgens on protein anabolism is confounded by the aforementioned intricate physiological relationship between gonadal steroids and GH. Separation of the actions of these two hormones required the use of the hypophysectomized animal model, in which the individual and combined effects of gonadal steroids and GH could be evaluated. In contrast to the early work of Scow and Hagan (50), studies in rats, lambs, and humans support the idea that androgens either directly promote growth or augment GH-stimulated growth. Klindt et al. (44) observed that in hypophysectomized rats, testosterone by itself is a weak, but demonstrable, anabolic agent. Dose-response analysis of GH-testosterone interactions demonstrated that the provision of a fixed dose of testosterone significantly amplified the GH-stimulated weight gain, particularly at low doses of GH (44). Interestingly, at the highest doses of GH, supplementation with testosterone did not yield a significantly greater gain in weight than that in the nonsupplemented group. This last point may explain the lack of synergy between testosterone and GH in the studies of Scow and Hagan (50). In humans, Attie et al. observed that in children with precocious puberty and GH deficiency, premature endogenous production of gonadal steroids per se augmented linear growth without GH treatment (51). Similarly, hypophysectomized and castrated prepubertal lambs grew in response

3717

to testosterone alone (52). In fact, gonadal steroids could be additive with GH in promoting growth (51). The data from the present study are entirely consistent with these observations and reemphasize the need to modify the view that gonadal steroids are passive actors in the growth process. Instead of using models in which pituitary function has been experimentally or pathologically eliminated, we investigated this issue by examining some simple metabolic end points in the context of controlling for GH secretion. Although more sophisticated metabolic end points (such as amino acid tracer kinetics) would have been desirable, blood sampling and the complexity of the present study precluded the use of additional techniques. Despite the limitations of urinary nitrogen collections (53), the decline in urinary nitrogen excretion under blinded and randomized conditions discloses a critical piece of information, that net protein anabolism has been stimulated. The parallel response in BMR, independently measured from urinary nitrogen excretion, provides additional support for the interactive nature of gonadal steroids and GH at the level of tissue metabolism. What is also unique to this experiment is the observation that end-exercise oxygen consumption was markedly enhanced by 19% from the low to high testosterone states. At present, there are no other endocrine agents known to increase end-exercise oxygen consumption. This increase may be a reflection of the augmented basal oxygen consumption. The results of the present study strongly suggest that gonadal steroids are potent modulators of the observed metabolic response and are more highly correlated with these responses than GH secretion. This is not to imply that GH is not necessary for these changes. Rather, based upon a large body of previous experimental work, it is likely that GH is necessary for the full anabolic expression of androgens (as observed in vitro) (54, 55). Thus, the present studies reorient perspective on the relative contributions by androgens and GH to this outcome. The concept that androgens alter the anabolic response to GH was also suggested by Malhotra et al., who studied the effects of oxandrolone on 24-h GH secretion and linear growth in boys at various stages of puberty (56). These investigators examined the relationship of the boys’ growth rates to GH secretion, stage of puberty, and treatment with oxandrolone. They found that oxandrolone did not affect GH secretion, and that in a multivariate regression model, only serum testosterone level and treatment with oxandrolone (not GH secretion) were correlated with the observed growth. Thus, for both growing adolescents as well as fully developed (adult) men, manipulation of androgens can influence metabolic outcome without substantive alterations in GH secretion. A corollary observation has also been made in women entering an exercise training program. Those with the highest basal androgen levels garnered the greatest response to the exercise program (57). In summary, short term variation in the gonadal steroid environment alters 24-h, but not stimulated, GH secretion. By contrast, stanozolol in a smaller group of subjects does not affect GH secretion. In contrast to the relative lack of effect on GH secretion, these large changes in gonadal steroids substantially alter urinary nitrogen balance, BMR, and ex-

3718

FRYBURG ET AL.

ercise-induced oxygen consumption. These responses were statistically independent of the observed variation in GH secretion. Taken together with the results of earlier studies, these observations suggest that gonadal steroids probably modulate the target tissue response to GH. The cellular mechanism(s) accounting for such an interaction is as yet unknown.

21. 22.

23.

Acknowledgments

24.

The authors thank the nurses and staff of the GCRC of the University of Virginia for their excellent care of the subjects, Ms. Ginger Bauler and Catherine Kern for their technical assistance, and Dr. Robert Abbott for statistical advice.

25. 26.

References 1. Kochakian CD. 1937 Testosterone and testosterone acetate and the protein and energy metabolism of castrate dogs. Endocrinology. 21:750 –755. 2. Friedl KE, Dettori JR, Hannan CJ, Patience TH, Plymate SR. 1991 Comparison of the effects of high dose testosterone and 19-nortestosterone to a replacement dose of testosterone on strength and body composition in normal men. J Steroid Biochem Mol Biol. 40:607– 612. 3. Forbes GB, Porta CR, Herr BE, Griggs RC. 1992 Sequence of changes in body composition induced by testosterone and reversal of changes after drug is stopped. JAMA. 267:397–399. 4. Bhasin S, Storer T, Berman N, et al. 1996 The effects of supraphysiologic doses of testosterone on muscle size and strength in normal men. N Engl J Med. 335:1–7. 5. Lobley E, Connell A, Buchan V, Skene PA, Fletcher JM. 1987 Administration of testosterone to wether lambs:effects on protein and energy metabolism and growth hormone status. J Endocrinol. 115:439 – 445. 6. Welle S, Jozefowicz R, Forbes G, Griggs RC. 1992 Effect of testosterone on metabolic rate and body composition in normal men and men with muscular dystrophy. J Clin Endocrinol Metab. 74:332–335. 7. Illig R, Prader A. 1970 Effect of testosterone on growth hormone secretion in patients with anorchia and delayed puberty. J Clin Endocrinol Metab. 30:615– 618. 8. Dawson-Hughes B, Stern D, Goldman J, Reichlin S. 1986 Regulation of growth hormone and somatomedin-C secretion in postmenopausal women: effect of physiological estrogen replacement. J Clin Endocrinol Metab. 63:424 – 432. 9. Ho KY, Evans WS, Blizzard RM, et al. 1987 Effects of sex and age on the 24-hour profile of growth hormone secretion in man: importance of endogenous estradiol concentrations. J Clin Endocrinol Metab. 64:51–58. 10. Friend KE, Hartman ML, Pezzoli SS, Clasey JL, Thorner MO. 1996 Both oral and transdermal estrogen replacement increase GH release in postmenopausal women–a clinical research center study. J Clin Endocrinol Metab. 81:2250 –2256. 11. Kerrigan JR, Rogol AD. 1992 The impact of gonadal steroid hormone action on growth hormone secretion during childhood and adolescence. Endocr Rev. 13:281–298. 12. Mauras NM, Blizzard RM, Link K, Johnson ML, Rogol AD, Veldhuis JD. 1987 Augmentation of growth hormone secretion during puberty: evidence for a pulse amplitude-modulated phenomenon. J Clin Endocrinol Metab. 64:596 – 601. 13. Hobbs CJ, Plymate SR, Rosen CJ, Adler RA. 1993 Testosterone administration increases insulin-like growth factor-I levels in normal men. J Clin Endocrinol Metab. 77:776 –779. 14. Fryburg DA, Gelfand RA, Barrett EJ. 1991 Growth hormone acutely stimulates muscle protein synthesis in normal humans. Am J Physiol. 260:E499 –E504. 15. Horber FF, Haymond MW. 1990 Human growth hormone prevents the protein catabolic side effects of prednisone in humans. J Clin Invest. 86:265–272. 16. Wolf RF, Heslin MJ, Newman E, Pearlstone DB, Gonnonne A, Brennan MF. 1992 Growth hormone and insulin combine to improve whole-body and skeletal muscle protein kinetics. Surgery. 112:284 –291. 17. Russell-Jones DL, Weissberger AJ, Bowes SB, et al. 1993 The effects of growth hormone on protein metabolism in adult growth hormone-deficient patients. Clin Endocrinol (Oxf). 38:427– 431. 18. Cuneo RC, Salomon F, Wiles CM, Hesp R, P. Sonksen. 1991 Growth hormone treatment in growth hormone-deficient adults. Effects on muscle mass and strength. J Appl Physiol. 70:688 – 694. 19. Salomon F, Cuneo R, Hesp R, Sonksen P. 1989 The effects of treatment with recombinant human growth hormone on body composition and metabolism in adults with growth hormone deficiency. N Engl J Med. 321:1797–1803. 20. DiMartino-Nardi J, Wu R, Fishman K, Saenger P. 1991 The effect of long-

27. 28. 29. 30. 31. 32. 33.

34. 35. 36. 37.

38. 39. 40. 41.

42. 43. 44. 45. 46. 47.

JCE & M • 1997 Vol 82 • No 11

acting analog of luteinizing hormone-releasing hormone on growth hormone ecretory dynamics in children with precocious puberty. J Clin Endocrinol Metab. 73:902–906. Griggs RC, William K, Jozefowicz RF, Herr BE, Forbes G, Halliday D. 1989 Effect of testosterone on muscle mass and muscle protein synthesis. J Appl Physiol. 66:498 –503. Liu L, Merriam GR, Sherins RJ. 1987 Chronic sex steroid exposure increases mean plasma growth hormone concentration and pulse amplitude in men with isolated hypogonadatrophic hypogonadism. J Clin Endocrinol Metab. 64:651– 656. Sharifi R, Soloway M, Group LS. 1990 Clinical study of leuprolide depot formulation in the treatment of advanced prostate cancer. J Urol. 143:68 –71. Weltman A, Weltman JY, Schurrer R, Evans WS, Veldhuis JD, Rogol AD. 1992 Endurance training amplifies the pulsatile release of growth hormone: effects of training intensity. J Appl Physiol. 72:2188 –2196. Wilmore JH. 1969 The use of actual, predicted and constant residual volumes in assessment of body composition by underwater weighing. Med Sci Sport Exer. 1:87–90. Brozek J, Grande F, Anderson JT, Keys A. 1963 Densitometric analysis of body composition: revision of some quantitative assumptions. Ann NY Acad Sci. 110:113–140. Powell DR, Rosenfeld RG, Baker BK, Hintz RL. 1986 Serum somatomedin levels in adults with chronic renal failure. J Clin Endocrinol Metab. 63:1186 –1192. Samojlik E, Kirschner MA, Silber D, Schneider G, Ertel NH. 1984 Elevated production and metabolic clearance rates of androgens in morbidly obese women. J Clin Endocrinol Metab. 59:949 –953. Samojlik E, Kirschner MA, Ribot S, Szmal NE. 1992 Changes in the hypothalamic-pituitary-gonadal axis in man after cadaveric kidney transplantation and cyclosporine therapy. J Androl. 13:332–336. Veldhuis JD, Johnson ML. 1992 Deconvolution analysis of hormone data. Methods Enzymol. 210:539 –575. Veldhuis JD, Johnson ML. 1995 Specific methodological approaches to selected contemporary issues in deconvolution analysis of pulsatile neuroendocrine data. Methods Neurosci. 28:25–92. Johnson ML, Veldhuis JD. 1995 Evolution of deconvolution analysis as a hormone pulse detection method. Methods Neurosci. 28:1–24. Veldhuis JD, Liem AY, South S, et al. 1995 Differential impact of age, sexsteroid hormones, and obesity on basal vs. pulsatile growth hormone secretion in men as assessed in an ultrasensitive chemiluminescence assay. J Clin Endocrinol Metab. 80:3209 –3222. Zachmann E, Prader A. 1970 Anabolic and androgenic effect of testosterone in sexually immature boys and its dependency on growth hormone. J Clin Endocrinol Metab. 30:85–95. Rosenfield RL, Furlanetto R. 1985 Physiologic testosterone or estradiol induction of puberty increases plasma somatomedin-C. Pediatrics. 107:415– 417. Weissberger AJ, Ho KKY. 1993 Activation of the somatotrophic axis by testosterone in adult males: evidence for the role of aromatization. J Clin Endocrinol Metab. 76:1407–1412. Bellantoni MF, Vittone J, Campfield AT, Bass KM, Harman SM, Blackman MR. 1996 Effects of oral vs. transdermal estrogen on the growth hormone/ insulin-like growth factor I axis in younger and older postmenopausal women–a clinical research center study. J Clin Endocrinol Metab. 81:2848 –2853. Veldhuis JD. 1995 The neuroendocrine regulation and implications of pulsatile GH secretion: gender effects. Endocrinologist. 5:198 –213. Link K, Blizzard RM, Evans WS, Kaiser DL, Parker MW, Rogol AD. 1986 The effects of androgens on the pulsatile release and the 24-h mean concentration of GH in peripubertal males. J Clin Endocrinol Metab. 62:159 –164. Devesa J, Lois N, Arce V, Diaz MJ, Lima L, Tresguerres JAF. 1991 The role of sexual steroids in the modulation of growth hormone (GH) secretion in humans. J Steroid Biochem Mol Biol. 40:165–173. Evans WS, Borges JCC, Vance ML, et al. 1984 Effects of human pancreatic growth hormone-releasing factor-40 on serum growth hormone, prolactin, LH, FSH, and somatomedin-C concentrations in normal women throughout the menstrual cycle. J Clin Endocrinol Metab. 59:1006 –1010. Lima L, Arce V, Lois N, et al. 1989 Growth hormone responsiveness to GHRH in normal adults is not affected by short-term gonadal blockade. Acta Endocrinol (Copenh). 120:31–36. Choo JJ, Emery PW, Rothwell NJ. 1991 Dose-dependent effects of an anabolic steroid, nandrolone phenylpropionate (Durabolin), on body composition and muscle protein metabolism in female rats. Ann Nutr Metab. 35:141–147. Klindt J, Ford JJ, MacDonald GJ. 1990 Synergism of testosterone propionate with growth hormone in promoting growth of hypophysectomized rats:effect of sexual differentiation. J Endocrinol. 127:249 –256. Martinez JA, Buttery PJ, Pearson JT. 1984 The mode of action of anabolic agents: the effect of testosterone on muscle protein metabolism in the female rat. Br J Nutr. 52:515–521. Mayer M, Rosen F. 1977 Interaction of glucocorticoids and androgens with skeletal muscle. Metabolism. 26:937–962. Menschikowski M, Jung K, Junghans P, Petzke KJ, Albrecht V. 1988 The

TESTOSTERONE MODULATION OF GH SECRETION AND METABOLIC OUTCOMES

48. 49.

50. 51.

52.

influence of a steroid hormone and of physical exercise on protein metabolism in rats. Exp Clin Endocrinol. 92:341–348. Powers ML, Florini JR. 1975 A direct effect of testosterone on muscle cells in tissue culture. Endocrinology. 97:1043–1047. Santidrian S, Moreyra M, Munro HN, Young VR. 1982 Effect of testosterone on the rate of myofibrillar protein breakdown in castrated and adrenalectomized male rats measured by the urinary excretion of 3-methylhistidine. Metabolism. 31:1200 –1205. Scow RO, Hagan SN. 1965 Effect of testosterone propionate and growth hormone on growth and chemical composition of muscle and other tissues in hypophysectomized male rats. Endocrinology. 77:852– 858. Attie KM, Ramirez-Nelson R, Conte FA, Kaplan SL, Grumbach MM. 1990 The pubertal growth spurt in eight patients with true precocious puberty and growth hormone deficiency:evidence for a direct role of sex steroids. J Clin Endocrinol Metab. 71:975–983. Young JR, Mesiano S, Hintz R, et al. 1989 Growth hormone and testosterone can independently stimulate the growth of hypophysectomized prepubertal

53. 54. 55. 56.

57.

3719

lambs without any alteration in circulating concentrations of insulin-like growth factors. J Endocrinol. 121:563–570. Kopple JD. 1987 Uses and limitations of the balance technique. J Paren Enteral Nutr. 11:79S– 85S. Thompson SH, Boxhorn LK, Kong W, Allen RE. 1989 Trenbolone alters the responsiveness of skeletal muscle satellite cells to fibroblast growth factor and insulin-like growth factor-I. Endocrinology. 124:2110 –2117. Yang S, Xu X, Bjorntorp P, Eden S. 1995 Additive effects of growth hormone and testosterone on lipolysis in adipocytes of hypophysectomized rats. J Endocrinol. 147:147–152. Malhotra A, Poon E, Tse W-Y, Pringle PJ, Hindmarsh PC, Brook CGD. 1993 The effects of oxandrolone on the growth hormone and gonadal axes in boys with constitutional delay of growth and puberty. Clin Endocrinol (Oxf). 38:393–398. Hakkinen K, Pakarinen A, Kallinen M. 1992 Neuromuscular adaptations and serum hormones in women during short-term intensive strength training. Eur J Appl Physiol. 64:106 –111.

AMERICAN ASSOCIATION FOR CANCER RESEARCH GERTRUDE B. ELION CANCER RESEARCH AWARD Supported by a Grant from Glaxo Wellcome Oncology The purpose of the AACR’s Gertrude B. Elion Cancer Research Award is to foster meritorious basic, clinical, or translational research in the U.S. or Canada by a non-tenured scientist at the level of Assistant Professor. Tenured faculty, federal government employees, and employees of private industry are not eligible for this award. Terms: The one-year award includes a $30,000 grant plus travel to the AACR Annual Meeting to accept the award. Candidates must be nominated by a member of the AACR and submit a detailed application. Deadline: December 15, 1997 For application information contact: Jenny Anne Horst-Martz American Association for Cancer Research Public Ledger Building, Suite 826 150 South Independence Mall West Philadelphia, PA 19106-3483 215-440-9300; Fax: 215-440-9372 E-mail: [email protected].