| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Pediatric Biochemistry (P.M.C.), Royal Hospital for Sick Children, Edinburgh EH9 1LF, United Kingdom; Section of Child Life and Health (P.M.C., A.E.M.E., C.J.H.K.), Department of Reproductive and Developmental Sciences, University of Edinburgh EH9 1UW, Edinburgh, United Kingdom; Department of Biochemistry (A.M.W.), Royal Infirmary, Glasgow G4 OSF, United Kingdom; and School of Biological and Molecular Sciences (N.P.G.)., Oxford Brookes University, Oxford OX3 0BP, United Kingdom
Address all correspondence and requests for reprints to: Dr. P. M. Crofton, Department of Pediatric Biochemistry, Royal Hospital for Sick Children, Sciennes Road, Edinburgh EH9 1LF, United Kingdom. E-mail: patricia.crofton{at}luht.scot.nhs.uk.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
-0.85, P < 0.01). Only LH nocturnal mean and amplitude, respectively, contributed independently to prediction of testosterone and inhibin B nocturnal means, explaining 71 and 65% of their variability. We conclude that both testosterone and inhibin B are related to nocturnal LH release in peripubertal boys but over different time scales. |
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In adult males, inhibin B is produced by Sertoli cells under the influence of FSH and has a role in feedback regulation of FSH (12). In boys, inhibin B is high during infancy, decreases gradually to a nadir in midchildhood, and then increases progressively before pubertal onset, reaching a new peak in early puberty (13, 14, 15). We and others have demonstrated that, in prepubertal boys and boys in early puberty, inhibin B does not have any discernible daytime relationship with FSH, although during mid- to late puberty, an inverse relationship develops (13, 15). During late prepuberty and early puberty, inhibin B correlates positively with LH and testosterone, although these relationships disappear by midpuberty (13, 15).
In adult men, inhibin B has been shown to follow a circadian rhythm, peaking in the early morning and reaching a nadir in the late afternoon, followed by gradually increasing nocturnal concentrations (16). In these adult men, temporal interaction was observed with testosterone concentrations. However, there is no information on any circadian rhythm in inhibin B in children. Its relationship, if any, with nocturnal pulses of LH and FSH, or with nocturnal testosterone secretion, is unknown.
Our aims in this study were to investigate nocturnal changes in testosterone and inhibin B in pre- and peripubertal boys and establish the interrelationships (if any) among nocturnal inhibin B, testosterone, and the pulsatile release of LH and/or FSH in boys as they approach clinical puberty.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Stored samples from 20 short normal boys aged between 6.0 and 16.2 yr who were enrolled in long-term prospective clinical trials of growth-promoting treatments (17) were selected for this study. Subjects were excluded if they had evidence of systemic or endocrine disease or any genital disorder such as cryptorchidism or hypospadias. All boys had normal hormonal responses to a combined insulin hypoglycemia/GnRH/TRH test. As part of the original studies, initial evaluation before treatment commenced included an assessment of spontaneous overnight GH secretion (see below). Only boys with normal overnight GH profiles were included in this study. As part of the original study, pubertal assessments were made at baseline and prospectively at three monthly intervals from enrollment. Genital stage was assessed according to Tanner (18). Testicular volume was determined using a Prader orchidometer (19) and was expressed as the mean of the right and left testis. For a boy who was prepubertal at enrollment, the date of onset of physical puberty was defined as the first clinic visit at which he was observed to be at genital stage 2 (G2). Full informed consent was obtained from both children and their parents, and the original studies were approved by the local ethics committee.
Our 20 study subjects were selected from the original study on the basis of chronological age, pubertal status at the time of investigation, and (for prepubertal boys) the time before subsequent onset of physical puberty. Group 1 comprised five prepubertal boys with familial short stature but no significant growth delay, who were younger than 8 yr at the time of investigation. Group 2 comprised five prepubertal boys with familial short stature but no significant growth delay, older than 8 yr but evaluated 1.5 yr or more before the onset of physical puberty. Group 3 comprised five prepubertal boys with familial short stature but no significant growth delay, evaluated 1.0 yr or less before the onset of physical puberty. All boys in groups 1–3 had testicular volumes of 2 ml. Among the 10 boys in groups 2 and 3 for whom their age at subsequent onset of physical puberty was known, after the initial evaluation, six were subsequently treated with GH (Norditropin, NovoNordisk, Copenhagen, Denmark, in a dose of 15–24 IU/m2·wk, divided into daily sc injections) and four with oxandralone (2.5 mg/d orally). All entered puberty at the normal time and progressed through it at the expected rate. Group 4 comprised five boys aged 14–16 yr, in early puberty (G2) at the time of investigation, with a median testicular volume of 4 ml (range 4–6 ml). This group had constitutional delay of growth and puberty with a mean bone age delay of 2.4 yr (range 1.4–3.6 yr).
Blood sampling protocol
For the original study, baseline evaluation of spontaneous overnight GH secretion was assessed by withdrawal of venous blood through an indwelling cannula, kept patent with heparinized saline (heparin 400 U/ml physiological saline). Separate aliquots were collected every 20 min from 2000 h to 0800 h. During sleep (lights out at 2100–2300 h, awakened at 0800 h), blood sampling was carried out remotely in an adjoining room via an extension cannula through an aperture in the interconnecting wall. Plasma was separated immediately and stored at -20 C until assay of GH. After GH assay for the original study, the remaining plasma was refrozen until assay of LH, FSH, testosterone, and inhibin B for this study. Samples were stored for a maximum of 10 yr and underwent a maximum of two freeze-thaw cycles before assay of LH, FSH, and inhibin B, and a maximum of three freeze-thaw cycles before assay of testosterone. We have found no differences in LH, FSH, testosterone, and inhibin B between samples stored under these conditions and samples stored at -70 C for much shorter periods without freeze-thaw cycles, matched for pubertal stage (our unpublished observations). The concentrations of LH and testosterone reported in this study are similar to those previously reported (8, 9). All samples from each subject were processed in a single analytical assay to minimize analytical variation.
Measurements
LH and FSH were measured by DELFIA time-resolved immunofluorescence assays (Wallac Oy, Turku, Finland), standardized against the second international reference preparations 80/552 and 78/549 respectively. Assay sensitivities were 0.15 U/liter and 0.25 U/liter for LH and FSH, respectively. All samples were assayed in duplicate. Quality control samples included in each run had within-assay coefficients of variation of 2.4% at 6.4 U/liter and 2.5% at 11.3 U/liter for LH and 5.6% at 5.1 U/liter and 2.2% at 10.3 U/liter for FSH.
Testosterone was measured by automated immunoassay on the Centaur ADVIA analyser (Bayer PLC, Newby, Berkshire, UK). Assay sensitivity was 0.5 nmol/liter. Within-assay coefficients of variation were 9.4% at 1.7 nmol/liter, 6.6% at 2.7 nmol/liter, and 3.7% at 6.8 nmol/liter.
Inhibin B was measured in duplicate by double-antibody ELISA, as described (20). Immunopurified inhibin B (provided by the author N.P.G.) was used as a standard, diluted serially in fetal calf serum; this material had been calibrated against recombinant dimeric inhibin B (Genentech, San Francisco, CA). Assay sensitivity was 8 ng/liter. Within-assay coefficients of variation were 7.4% at 88 ng/liter and 8.4% at 233 ng/liter.
Pulse analysis
This was applied to within-subject overnight profiles to provide an objective evaluation of LH, FSH, and testosterone pulses. We used the threshold method (21) as modified by Baird et al. (22) and validated (8) against a computerized algorithm based on the Pulsar method of testing for significant excursions from a baseline (23, 24). According to this algorithm, the criteria for a pulse are as follows: 1) the hormone concentration in two consecutive samples is greater than the mean of the two previous samples, and 2) the increase in at least one of the peak samples is greater than 3 times the within-assay SD at the prepulse hormone concentration. To apply the second of these criteria to LH and FSH pulses, the SD of within-assay duplicate measurements of each hormone was calculated at the appropriate hormonal level, based on all plasma samples from the study subjects. For testosterone, the within-assay SD at the appropriate concentration was used (see above). Inhibin B did not show a pulsatile pattern. After identification of a pulse, the prepulse nadir was identified as the lowest hormone concentration occurring up to 40 min (LH or FSH) or 60 min (testosterone) before the peak of the pulse. The pulse amplitude was calculated as the pulse peak minus the preceding nadir.
Statistical analysis
For LH, FSH, and testosterone, results below the assay sensitivity were expressed as the detection limit. For inhibin B all results were within assay range. Data were expressed as mean and range unless otherwise stated. For each subject, we computed his mean nocturnal hormone concentration, equivalent to the area under the curve because sampling was at the same 20-min intervals for each subject. The group mean of the within-subject overnight mean values was then calculated for each of groups 1–4, respectively. Additionally, for testosterone, its nocturnal increment was calculated for each subject by subtracting the concentration at 2000 h from that at 0800 h, followed by calculation of the group mean. Comparisons among groups were by Kruskall-Wallace ANOVA because, although data conformed approximately to a gaussian distribution, variances differed among groups 1–4. Comparisons between variables were by Spearman rank correlation analysis, with correction for ties. Statistical significance was accepted at P < 0.05.
To evaluate nocturnal variation of inhibin B and testosterone, data series for each subject was first smoothed by the method of running means, whereby the hormonal concentration at each time point was expressed as the mean for that time point and for the preceding and subsequent time points. The mean concentration at each time point was then calculated for each group (1, 2, 3, 4).
Because all puberty-related hormones show strong correlations, we used multiple regression analysis to identify which hormonal variables(s) contributed independently to the prediction of either testosterone (nocturnal mean and increment) or inhibin B (at 2000 h and nocturnal mean). We used forward stepwise regression, with reiterative evaluation to remove noncontributory variables. For testosterone, the variables introduced to the model were LH nocturnal mean, LH mean pulse amplitude, FSH nocturnal mean, FSH mean pulse amplitude, inhibin B at 2000 h, and inhibin B nocturnal mean. For inhibin B, the variables introduced to the model were LH nocturnal mean, LH mean pulse amplitude, FSH nocturnal mean, FSH mean pulse amplitude, testosterone nocturnal mean, and testosterone nocturnal increment. The proportion of the variability in the dependent variable predicted by the model was estimated using the adjusted r2 to compensate for the number of variables entered.
Auto- and cross-correlation analyses were used to investigate any temporal couplings between hormonal time series (25). The auto- or cross-correlation coefficient rk measures the correlation between two values a distance (lag) k time units apart. For autocorrelation analysis, the serial nocturnal inhibin B values obtained in each subject were correlated with an exact copy at progressively increasing lag times to determine whether there was evidence of episodic release. Cross-correlation analysis is particularly suited to assessing whether pulsatile release of different hormones is temporally related. By progressively lagging one data array with respect to the other, the lag of the peak rk value can be assessed, indicating the phase difference between the arrays. After calculating the rk value for each time lag k (in 20-min steps from -300 min to +300 min) for each subject, the rk values from different subjects were pooled using Fisher’s Z transformation to calculate the mean population rk at each time lag k.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Nocturnal LH pulse frequency increased during peripuberty (groups 3 and 4, Table 1
), and the median clock time for the first LH pulse was 2400 h in both groups. Nocturnal mean LH and FSH increased progressively from groups 2 to 4, and there was a trend toward increasing LH pulse amplitude.
|
). FSH nocturnal mean and pulse amplitude were less strongly related to time before pubertal onset (rs -0.77 and -0.71, P < 0.05).
|
Testosterone at 0800 h and its nocturnal increment and mean were all higher in group 4 boys than in the prepubertal groups (Table 1
). There was evidence of a downward trend in testosterone during the early part of the night in all groups, reaching a nadir around 2300–2400 h (Fig. 2). Thereafter, there was a progressive increase in testosterone in groups 3 and 4, compared with little or none in the younger groups, with group 4 boys showing by far the greatest increase. The younger group 1 and 2 boys had no detectable testosterone pulses. Two group 3 boys had a single discrete testosterone pulse whereas four of five boys in group 4 had three to six nocturnal pulses of testosterone (Table 1). The single group 4 boy with no detectable testosterone pulses (and the lowest testosterone nocturnal mean) also had the lowest LH nocturnal mean and pulse amplitude in the group.
|
Inhibin B
Groups 1 and 2 had similar inhibin B nocturnal means, whereas the peripubertal boys in groups 3 and 4 had considerably higher means (Table 1). All groups had evidence of a downward trend in inhibin B concentrations during the night, with a nadir around 0500–0700 h, followed by a rise (Fig. 3). The decrease was most pronounced in groups 3 and 4, the groups with the highest inhibin B levels. The early morning rise was most pronounced in group 4, the boys in early puberty. Although we found no evidence of discrete pulses in inhibin B, there was some evidence of nonpulsatile episodic fluctuations in all groups with significant (P < 0.05) autocorrelations over a lag interval of 0–20 min for group 1, 0–40 min for group 2, 0–80 min for group 3, and 0–40 min for group 4.
|
).
|
Table 2 shows significant correlations among all hormonal variables for the 20 study subjects. Testosterone and inhibin B were found to be most highly correlated with LH mean and pulse amplitude.
|
Temporal coupling of hormones in peripubertal boys
Possible temporal interactions among hormones were explored using cross-correlation analysis in the eight peripubertal boys who had more than three nocturnal LH pulses. LH and testosterone were strongly cross-correlated, with a peak correlation at a time lag of +60 to +160 min, indicating that increases in testosterone followed pulses in LH (Fig. 5). This is illustrated for one individual group 4 boy in Fig. 6. The single subject with lower cross-correlations between LH and testosterone had the highest nocturnal mean levels of LH (3.61 U/liter) and testosterone (7.62 nmol/liter, 2.20 ng/ml) in the study and also the highest level of testosterone at 2000 h (5.40 nmol/liter, 1.56 ng/ml).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Although it has long been established that testosterone shows a nocturnal increase in peripubertal boys (3, 5, 8, 26), its secretory dynamics in adolescent boys have not hitherto been investigated. We found that testosterone did not increase uniformly through the night. It declined gradually in the early part of the night to reach a nadir at around 2300 h to midnight. Thereafter, it rose progressively to reach its highest values at around 0700–0800 h, with evidence of pulsatile secretion in the peripubertal boys. This postmidnight rise in testosterone therefore coincided with the nocturnal onset of pulsatile LH release. This is the first report to our knowledge of pulsatile testosterone secretion in peripubertal boys. In boys with more than three nocturnal LH pulses, strong cross-correlations were observed between LH and testosterone in seven of eight boys. Higher LH/testosterone cross-correlation coefficients were observed in our study, compared with 24-h time series analyses in a group of five adult men (25). This may be because of the more discrete nature of nocturnal LH and testosterone pulses in peripubertal boys, compared with adult men. The only boy in our study with a lower LH/testosterone peak cross-correlation was the boy who, hormonally, appeared to have progressed furthest in puberty with the highest nocturnal mean LH and testosterone and indirect evidence of antecedent daytime LH pulsatility (high levels of testosterone at 2000 h and LH pulses in the early evening). In our study on peripubertal boys, maximal cross-correlations occurred at time lags of 60–160 min, compared with 50–70 min in adult men (25).
There have been no previous published studies on inhibin B nocturnal secretion in children, although there has been one report of circadian variation of inhibin B in adult men (16). In adult men, peak values were observed in the early morning, with a nadir in the late afternoon, followed by gradually increasing nocturnal values (16). We did not have the opportunity (for ethical reasons) to measure inhibin B throughout the full 24-h period, but our observations indicate that its circadian rhythm in pre- and peripubertal boys almost certainly differs from that in adults. Particularly in the older peripubertal boys, there was a downward trend through the evening and night, reaching a nadir around 0500–0700 h, followed by an early morning rise. No discrete pulses were detected, but there was some evidence of episodic secretion, as previously noted for adult men (16).
Nocturnal inhibin B was similar in the youngest boys (group 1) and those boys who were more than 1.5 yr from physical puberty (group 2). However, boys in late prepuberty (group 3) had considerably higher levels and those in early puberty (group 4) higher levels still. Among boys for whom their age at subsequent pubertal onset was known, inhibin B nocturnal mean (like LH) was higher the closer a boy was to the onset of puberty.
Inhibin B was most highly correlated with LH pulse amplitude and nocturnal mean. Multiple regression analysis confirmed that LH pulse amplitude was the only independent predictor of inhibin B at 2000 h and its nocturnal mean. However, in contrast to the strong temporal interactions between LH and testosterone, cross-correlation analysis did not suggest any significant short-term temporal coupling between pulsatile release of LH and inhibin B levels within the 12 h period studied. The interaction is therefore likely to occur over a longer period. The lack of short-term temporal coupling between LH and inhibin B would be consistent with our observation that GnRH administration has no effect on inhibin B over a 60-min period (our unpublished observations). The strong relationship between LH pulse amplitude and inhibin B in our study is consistent with previous observations (based on daytime samples) of a positive relationship between single measurements of LH and inhibin B in boys in late prepuberty and early puberty (13) and with inhibin B increases in response to three weeks of intramuscular hCG treatment in prepubertal boys with cryptorchidism (27).
Once the effect of LH had been accounted for, testosterone did not have an independent role in predicting inhibin B variability. We, and others, have previously observed a positive relationship between inhibin B and testosterone in early pubertal boys (G2) but not during mid- to late puberty (13, 15). It is likely that daytime testosterone levels in these boys reflected LH pulsatile release during the previous night and may therefore have been a surrogate for nocturnal LH secretion rather than an independent regulator of inhibin B. In the present study, we observed no positive temporal coupling between nocturnal inhibin B and testosterone that might have indicated direct interaction at testicular level. Instead, there was a weak negative interaction over a wide span of lag times that almost certainly simply reflected their opposite time trends during the night. This contrasts with a previous report that, in adult men, testosterone had a significant influence on daytime levels of inhibin B with a lag of 30–60 min (16). However, the testicular site of production of inhibin B (currently a subject for controversy), hormonal control and, as we have now demonstrated, the circadian rhythm of inhibin B may all differ between children and adult men.
We, and others, have previously demonstrated that there is no relationship between inhibin B and FSH in pre- and peripubertal boys during the day (13, 15). Multiple regression analysis in this study also found that FSH did not contribute independently to the prediction of nocturnal inhibin B (although a subtle modulatory influence cannot be excluded). There was also no significant temporal interaction between FSH and inhibin B.
We conclude that, in peripubertal boys, testosterone declines gradually in the early part of the night, followed by a postmidnight increase reflecting the nocturnal onset of LH pulsatility. Most boys with nocturnal pulsatile LH release also had discrete testosterone pulses, and strong cross-correlations were observed between LH and testosterone. This is the first report of nocturnal pulsatile release of testosterone in response to LH nocturnal pulsatility in peripubertal boys. Nocturnal inhibin B levels were higher in peripubertal compared with younger boys and showed a downward trend overnight with no evidence of pulsatility and no evidence of short-term interactions with any other hormone studied. Inhibin B (like LH) was higher the closer a boy was to the onset of physical puberty. LH pulse amplitude was identified as the only independent predictor of nocturnal inhibin B. This study sheds further light on the nocturnal secretory dynamics of testosterone and inhibin B in pre- and peripubertal boys.
|
Footnotes |
|---|
Abbreviations: G2, Genital stage 2; rk, correlation of two data arrays at a lag time of k units apart.
Received May 21, 2003.
Accepted August 18, 2003.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  S. Ramaswamy Pubertal Augmentation in Juvenile Rhesus Monkey Testosterone Production Induced by Invariant Gonadotropin Stimulation Is Inhibited by Estrogen J. Clin. Endocrinol. Metab., October 1, 2005; 90(10): 5866 - 5875. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S Ramaswamy, C R Pohl, G R Marshall, and T M Plant A switch from continuous to episodic testicular testosterone release in response to pulsatile LH stimulation in juvenile rhesus monkeys (Macaca mulatta) J. Endocrinol., October 1, 2004; 183(1): 61 - 68. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
