| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From the Clinical Research Centers |
Division of Endocrinology, Department of Internal Medicine, National Science Foundation Center for Biological Timing, University of Virginia Health Sciences Center (J.D.V.), Charlottesville, Virginia 22908; the Endocrine Section, Medical Service, Salem Veterans Affairs Medical Center (A.I.), Salem, Virginia 24153; the Department of Pathology, Pennsylvania State University Medical School (L.M.D.), Hershey, Pennsylvania 17033-0850; and Geriatrics Medicine, Hunter Holmes McGuire Veterans Affairs Medical Center (T.M.), Richmond, Virginia 23249
Address all correspondence and requests for reprints to: Dr. J. D. Veldhuis, Division of Endocrinology, Department of Internal Medicine, Box 202, University of Virginia Health Sciences Center, Charlottesville, Virginia 22908. E-mail: jdv{at}virginia.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Although pulsatile FSH release is less visibly evident and analytically tractable than pulsatile LH secretion (9, 16), recent in vitro and in vivo experiments document directly both pulsatile and basal (nonpulsatile or constitutive) components of FSH secretion (5, 7, 17, 18, 19). The mechanisms that control the pulsatile vs. basal modes of FSH release probably differ, as several glycoprotein hormones, such as inhibin and activin, can selectively suppress or stimulate basal FSH secretion over hours (2), whereas GnRH evokes a rapid pulse of FSH secretion over minutes mimicking the burst-like mode of LH release (1).
To date, no clinical studies to our knowledge have explored the pathophysiological regulation of basal vs. pulsatile FSH release in healthy aging. This deficiency results in part from the analytical challenge of simultaneously quantifying basal and pulsatile gonadotropin secretion, especially without prior knowledge of relevant hormone-specific elimination rates (20). To the latter end, we recently performed bolus and steady state iv infusions of highly purified human FSH in hypopituitary men (followed by RIA, immunoradiometric assay (IRMA), and bioassay of the serum FSH disappearance curves) to calculate directly FSH’s biexponential kinetics (21). Given such half-life estimates, we could here apply half-life constrained multiparameter deconvolution analysis to evaluate the individual and joint contributions of basal and pulsatile FSH secretion to the selective FSH elevation observed in healthy older men (22, 23).
We hypothesized, first, that the monotropic rise in serum FSH concentrations in normal, unmedicated, ambulatory, and clinically eugonadal older men arises mechanistically from a preponderant increase in basal FSH secretion, with an unchanged (or reduced) young adult pulsatile mode of FSH release. Secondly, we postulated that the orderliness of FSH release patterns as quantified by approximate entropy is age independent. We tested these ideas by frequent (2.5-min) blood sampling overnight in 11 young and 8 older healthy men, followed by FSH chemiluminescence assay.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Eleven healthy young men, aged 21–34 yr, and eight older men, aged 62–72 yr, were recruited from the University of Virginia, Richmond, and Salem communities for study in the General Clinical Research Center after provision of written informed consent approved by the institutional review board. Each volunteer first underwent 1 night of adaptation to the study unit, followed by overnight blood sampling at 2.5-min intervals from a forearm vein. A long venous catheter was used to sample from an adjacent room to avoid disturbing the patient during sleep. Volunteers were healthy, ambulatory, and unmedicated men without acute or chronic illness, weight changes (2 kg within 10 days), or recent (within 2 weeks) transmeridian (more than three time zones) travel. There was no evidence by screening laboratory tests, physical examination, or medical history of hepatic, renal, hematological, metabolic, or endocrine diseases. Baseline serum concentrations of immunoreactive LH, FSH, PRL, GH, TSH, T4, resin T3 binding, testosterone, and estradiol were all unremarkable for age (12, 24).
Assays
Serum concentrations of FSH (or LH) were assayed in each sample
(or a pool) via an automated random access chemiluminescence-based
immunoassay (180, Chiron Corp., East Walpole, MA) using
WHO Second International Reference Preparation 94/632 (or 80/552).
Independent studies demonstrated a linear correlation between FSH (or
LH) concentrations measured in this manner and duplicate values
determined independently by IRMA (r 
0.899; n = 18;
P < 10-5) (25). The
within-assay coefficients of variation (CVs) were less than 6.5%, and
the between-assay CVs less than 9.0%. The former was used to calculate
within-sample SD estimates for deconvolution
analysis. All samples from an individual were analyzed together to
eliminate interassay variability.
Serum inhibin B (Inhibin-B Dimer Assay Kit, Serotec, Kidlington, UK), total and bioavailable testosterone, LH, GH, TSH, dehydroepiandrosterone sulfate, cortisol, T4, insulin-like growth factor I (IGF-I), and IGF-binding protein-3 concentrations were assayed in overnight pools in each volunteer via commercially available kits using enzyme-linked immunosorbent assay (ELISA), RIA, or IRMA, as described previously (12, 24, 26, 27, 28). Estradiol was assayed by a commercially available double antibody RIA (Third-Generation DSL-39100, Diagostics Systems Laboratories, Inc., Webster, TX) with less than 7% cross-reactivity with estrone and less than 0.45% with less potent estrogens. Sensitivity was 0.6 pg/mL with the lowest standard at 1.5 pg/mL, linearity to 150 pg/mL, and an ED50 of 20 pg/mL. The range in normal men is reported as undetectable to 44 pg/mL (manufacturer’s data; n = 57 men). The intra- and interassay CVs were 2.4–6.5% and 3.7–9.9%, respectively, with 85–109% recovery (added estradiol, 11–120 pg/mL).
Deconvolution analysis
Deconvolution analysis with an a priori half-life constraint was applied to estimate pulsatile FSH secretion from the overnight serum FSH concentration time series, using techniques described previously (22, 23, 29, 30). We estimated basal FSH secretion and concurrently the number, duration, mass, amplitude, and frequency of statistically significant FSH secretory bursts (P < 0.05 vs. zero amplitudes by joint statistical confidence intervals) using previously published two-component FSH elimination kinetics (21). In particular, the mean rapid phase (initial) half-life of FSH was 110 min, the mean slower component was 620 min, and their fractional amplitude (of the slower component taken as a ratio of total disappearance) was 0.39 (21). FSH kinetics were assumed to be invariant throughout the sampling interval and independent of age. Deconvolution measures included the mass of FSH secreted per burst (integral of the computed secretory event), amplitude of the FSH secretory burst (maximal rate of FSH secretion attained within a release episode), half-duration (time elapsed in minutes at half-maximal secretory burst amplitude), frequency (number of FSH secretory pulses observed per sampling session), and interpulse interval (time in minutes separating the center of consecutive FSH secretory bursts). Basal FSH secretory rates are expressed as the amount (international units) of FSH secreted per unit distribution volume (liters) per unit time (minutes). Total FSH production overnight represents the sum of pulsatile (mean pulse mass multiplied by FSH burst number) and basal FSH secretion (mean FSH secretory rate multiplied by duration of sampling interval).
Approximate entropy (ApEn)
ApEn is a family of scale-independent statistics used to assess the orderliness or serial regularity of hormone patterns on a sample by sample basis; hence, ApEn quantitates variability that is subordinate to pulsatile or circadian rhythms. Any particular ApEn statistic is a single, finite, nonnegative, real number assigned as an ensemble value to a hormone profile, with larger values corresponding to greater relative randomness of the serial measures. ApEn measures technically the logarithmic likelihood that short runs of data patterns that are similar remain similar on the next incremental comparison (31). Two principal input parameters, namely m and r, are fixed to compute ApEn from vector sequences constructed from the observed data, where m represents the window length of consecutive hormone measurement pattern, and r is the tolerance for testing subpattern recurrence. To maintain scale invariance, r is typically fixed as a percentage or fraction of the between-sample SD of each time series, e.g. 20%, and m is fixed as a value of 1 or 2 denoting consecutive data vectors of length one or two data points. For the FSH series analyzed here, we calculated ApEn values with r = 0.2 and m = 1, and hence, the designation, ApEn (1, 20%). This provides an appropriate statistic for assessing subpattern reproducibility in data series of this length (14, 32).
Statistics
Statistical comparisons were made by a two-tailed unpaired Student’s t test for mean serum FSH concentrations, and the Wilcoxon (rank sum) unpaired nonparametric test for specific deconvolution measures due to their non-Gaussian distributions. Secretory measures are reported as the mean ± SEM (median). Linear correlations were determined by Pearson’s coefficient. P < 0.05 was considered statistically significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
for three
young and three older men, serum FSH concentrations overnight were
visually pulsatile and superimposed upon apparently time- invariant
baseline release. FSH secretion profiles calculated by deconvolution
analysis using (previously published) biexponential FSH kinetics are
illustrated in Fig. 1B.
|
, mean overnight serum
FSH concentrations were approximately 2-fold higher in the group of
older men compared to those in their young counterparts
(P = 0.0066). Integrated values were comparably age
dependent (not shown). In contrast, overnight (pooled) serum TSH,
estradiol, inhibin B, and GH concentrations were age invariant (Table 1). Serum IGF-I, IGF-binding protein-1,
T3, and dehydroepiandrosterone sulfate
concentrations were lower in older men, whereas that of cortisol was
higher (Table 1). LH concentrations averaged
as follows: older, 3.8 ± 0.41 IU/L; young, 3.4 ± 0.33 IU/L
(P = NS). Figure 2 depicts the corresponding individual
mean LH values in all 19 volunteers.
|
|
.
Bioavailable, but not total, testosterone showed a significant age
contrast (P < 0.01). In a chemiluminescent assay of
total serum testosterone (26), values averaged 523 ± 40 (young)
and 425 ± 48 ng/dL (older; P = NS); and, in
another RIA (27), values averaged 470 ± 39 (young) and 397
± 36 ng/dL (older; P = NS).
|
The age contrast in FSH secretory burst amplitude/mass was specific, as the half-duration of FSH secretory bursts was similar in both age groups [14 ± 2.1 (12) min in young and 9.3 ± 1.9 (10) min in older subjects], as was the mean number of FSH secretory bursts overnight [5.0 ± 0.51 (5) in young and 5.3 ± 0.53 (5) in older men; P = NS]. The FSH intersecretory pulse interval also did not differ with age, averaging in young 81 ± 6.2 (72) vs. in older men 70 ± 8 (67) min.
Basal FSH secretory rates were elevated 2-fold in older individuals at 0.0042 ± 0.0009 (0.0031) vs. 0.0021 ± 0.0004 (0.0016) IU/L·min in young volunteers (P < 0.01).
Overnight basal, pulsatile, and total (summed basal and pulsatile) FSH
secretion values in young and older men are compared in Fig. 4. For each measure, older men (compared to
their young counterparts) secreted more FSH (P = 0.040
to P < 0.01). Expressed as a percentage of the total
FSH secretion, the pulsatile component was similar at 75 ± 4.4%
(75) in young vs. 73 ± 4.9% (79) in older
subjects.
|
gives mean FSH ApEn values in the two groups of older and young
men, which were not significantly different.
Figure 5 highlights the negative linear
correlations between basal FSH secretion rates and serum estradiol or
bioavailable testosterone, but not inhibin B, concentrations. In
contrast, FSH burst mass did not show any correlation with estradiol,
bioavailable testosterone, or inhibin B. Mean serum FSH concentrations
correlated negatively with serum estradiol (r = -0.59;
P = 0.020).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The mean (overnight) serum FSH concentration was approximately 2-fold higher in the older (than young) men studied here. This measurement by chemiluminescence assay agrees with most earlier clinical reports of elevated circulating immunoreactive FSH concentrations in older individuals (2), but contrasts with diminished circulating FSH levels in the older male rodent (35, 36). Thus, in the human, either primary gonadoprival states or older age (e.g. in both men and peri- and postmenopausal women) will evoke elevated serum FSH concentrations, whether assessed by chemiluminescence (present data), RIA, IRMA, or bioassay (9, 11, 12, 37, 38, 39, 40). Gonadotropin changes in the healthy older men studied here were monotropic, as FSH rose, but there was no associated increase in the mean serum LH concentration. Other hormone concentrations were altered in the expected manner in healthy, unmedicated, ambulatory, unstressed older males, but how or whether such changes contribute to isolated FSH hypersecretion is not known.
By deconvolution analysis, we quantitated a 2-fold amplification of the basal FSH secretory rate overnight in older men, assuming identical FSH kinetics in both age groups. This age-related augmentation could reflect partial withdrawal of gonadal negative feedback control of FSH release (13, 41). For example, the presently measured decrease in bioavailable (but not total) serum testosterone concentrations in the older subjects might plausibly account for the rise in basal FSH secretion. Indeed, linear correlation analysis (n = 19 men) disclosed a significant negative relationship between serum bioavailable testosterone concentrations and basal FSH secretion rates. In addition, augmented FSH secretion rates in older men might be mediated by reduced (intra-)pituitary actions of follistatin and/or inhibin, and/or greater intrapituitary drive of gonadotrope cell FSH biosynthesis, e.g. by activin (2). The last speculation would be consistent with activin’s ability to stimulate FSH release in the rat in vitro and in the monkey in vivo (4). Other regulatory factors might also modulate the synthesis of FSH (3, 4, 17). However, independently of the (various) biochemical inputs to gonadotrope cells, our inference of a 2-fold heightened basal FSH secretion rate in older men indicates for the first time to our knowledge that age can control the nonpulsatile mode of pituitary FSH secretion.
Sustained pulsatile secretion of FSH (and LH) requires an intermittent, rather than continuous, GnRH pulse stimulus (1, 2, 42, 43). Conversely, GnRH antagonist infusion will significantly reduce serum FSH (and LH) concentrations, blunt FSH secretory pulse amplitude, and abolish the acute surge-like release of FSH triggered by combined estrogen and progesterone administration in older women (8, 44, 45, 46, 47). Down-regulation experiments with GnRH agonistic analogs also support a role for episodic endogenous GnRH release in maintaining the pulsatile discharge of FSH (2). In contrast to GnRH’s (feedforward) drive of FSH production, negative feedback by gonadal sex steroid hormones will oppose GnRH-stimulated FSH release in the sheep, rat, and human (2). For example, a pulsatile GnRH infusion in men fails to stimulate progressive serum FSH increases when serum (aromatizable) androgen concentrations rise (48). Conversely, a continuous iv infusion of estradiol suppresses FSH secretory burst mass and inhibits FSH release stimulated by an (exogenous) iv bolus of GnRH (10, 11, 12). Thus, FSH secretion appears to require relevant stimulatory signals, such as intermittent GnRH and/or other agonists (feedforward drive) and, conversely, is susceptible to negative modulation by inhibitory feedback inputs. In accordance with the latter idea, we report here that serum bioavailable testosterone (an aromatizable androgen) and estradiol concentrations both correlate negatively with basal FSH secretion rates. Of interest, neither serum sex steroid nor inhibin B concentrations correlated with FSH secretory pulse mass. Thus, a plausible speculation is that aging results in accentuated net gonadotrope cell basal FSH release, on the one hand, and attenuated feedback inhibition of FSH release by estradiol and bioavailable testosterone, on the other hand. In contrast, serum inhibin B concentrations measured by ELISA were similar in both age cohorts, perhaps reflecting the excellent health of the older volunteers studied here, the specificity of the ELISA used, the number of subjects studied, and/or our 6- to 8-h overnight pooling of sera for specific inhibin B assay.
In both young and older men, the nonaromatizable androgen,
5
-dihydrotestosterone, infused continuously iv over 3 days
suppresses FSH secretory burst frequency, but not FSH pulse mass (10, 11). Conversely, antiandrogen treatment often increases gonadotropin
(LH) pulse frequency (49, 50). Thus, we reason that older men’s
enhanced FSH secretory pulse mass with no attendant rise in burst
frequency is probably not due to selective withdrawal of androgen
receptor-mediated negative feedback. Rather, as estrogen reduces FSH
pulse mass without altering FSH peak frequency (2), we speculate that
the low bioavailable testosterone concentrations measured in older men
indirectly raise FSH secretory burst mass by limiting substrate
availability for intrapituitary estrogen formation.
Whereas the frequency of low amplitude LH pulses tends to rise (24, 26, 51), and conversely, that of higher amplitude LH peaks tends to fall (52) in older men, the present results show that overnight FSH pulse number is not higher in elderly individuals. The neuroendocrine mechanism underlying this discrepancy between LH and FSH pulse frequencies in aging men is not known. One plausible conjecture is that aging results in a partial loss of GnRH-FSH pulse coupling, perhaps analogously to reduced GnRH-LH pulse linkage in uremia (53). A second hypothesis is an age-related disparity in feedback and/or feedforward factors that regulate FSH (vs. LH) release, e.g. preferential (net) augmentation of gonadotrope cell FSH synthesis, release, or storage, which thereby would enlarge FSH secretory burst mass in older men. Consistent with this view of enhanced availability of pituitary FSH is the consistently heightened release of FSH after near-maximally effective single dose GnRH injection in older men (37, 39, 54, 55) and a recent report of elevated activin A in older (albeit premenopausal) women (56).
Direct hypothalamo-pituitary venous blood sampling in the horse and sheep have recently documented significant basal as well as pulsatile modes of FSH release (5, 18, 57). Whereas similar invasive blood sampling in the human is not ethically practicable, our finding of amplified FSH secretory burst mass (and amplitude) in older individuals would make FSH pulse detection in the peripheral blood relatively more reliable in older than in young subjects. Thus, our analyses should not be vulnerable to a preferential type II statistical error in FSH pulse detection in older (compared with young) individuals.
As appraised via the approximate entropy statistic, the serial orderliness of LH release deteriorates significantly in older men, whether evaluated by 10-min blood sampling over 24 h (14) or by 2.5-min blood sampling overnight (58). In contrast, the present intensive sampling schedule demonstrates the lack of any erosion of pattern regularity of FSH release with aging. One plausible clinical hypothesis to explicate this age-related contrast in the relative orderliness of LH and FSH release is that aging (or one of its covariates) differentially attenuates the GnRH-driven subcellular signaling, synthesis, packaging, and/or secretion of LH (vs. FSH) by gonadotrope cells (17), thus eliciting more disorderly patterns of LH (but not FSH) release in older subjects. Alternatively, FSH release is stabilized by other (less age-dependent) factors.
Whether blood removal rates for FSH are altered in hypopituitary individuals (22) or healthy older men is not known definitively. This is a plausible conjecture, as multiple FSH glycosylation products with potentially different half-lives exist (59, 60). Heterogeneity of FSH molecules is evident in heterologous kinetic assays [e.g. human serum FSH extracts injected into hypophysectomized mice (61)] as well as by deconvolution analysis of FSH release stimulated by GnRH pulses at different stages of the normal menstrual cycle (62). In the case of LH, reduction of gonadal sex steroid negative feedback promotes the accumulation in blood of more acidic LH isoforms, which typically exhibit more prolonged in vivo half-lives due to their greater posttranslational sialic acid composition (60, 63). An unexpected (and as yet unreported) increase in the endogenous FSH half-life in older men would require less FSH secretion per unit distribution volume to achieve equivalently elevated serum FSH concentrations in older individuals, which would thus tend to falsely mimic enhanced basal FSH secretion analytically (20). Against this postulate is the observation that endogenous FSH half-lives calculated by deconvolution analysis after GnRH-stimulated FSH release in older men are not increased (12, 32, 54). On the other hand, as the mass of FSH secreted per burst is computed as an incremental rise above basal secretion, any hypothetically prolonged FSH half-life in older individuals would not account for their higher calculated FSH secretory burst mass.
In summary, we have investigated the neuroendocrine mechanisms underlying the monotropic elevation in serum FSH concentrations in healthy older men. The older men studied here overnight exhibited young adult serum concentrations of LH, total testosterone, estradiol, and inhibin B, but reduced serum bioavailable testosterone. Deconvolution analysis with earlier published biexponential FSH kinetics revealed a doubling of mean serum FSH concentrations in older men, which originated mechanistically from their nearly 2-fold elevated basal (nonpulsatile) FSH secretion rate combined with a 50% rise in FSH secretory burst mass. These dual hypersecretory features in aging men were specific, as no differences were apparent in FSH secretory pulse duration, interpulse interval, or frequency. The bipartite mechanisms of FSH hypersecretion suggest an age-related augmentation of both time-invariant basal (putatively only minimally GnRH-dependent) and pulsatile (presumptively largely GnRH-driven) FSH release. These new observations on the mechanisms of heightened FSH release in older men contrast with earlier appraisals of LH secretory dynamics, which often show a decline in LH secretory burst mass (and amplitude) with a reciprocal rise in LH pulse frequency in aging. Using the ApEn statistic, we further note that the quantifiable orderliness of sample to sample FSH (unlike LH) release patterns remains unaltered in older men. We conclude that reproductive aging in men provides a physiological paradigm of dissociated LH and FSH regulation in three primary respects: 1) preferential elevation of basal FSH secretion rates, 2) opposite changes in FSH (increased) and LH (decreased) secretory pulse amplitude (mass), and 3i) divergent age-related relative preservation of the orderliness of FSH (vs. LH) release. The exact hypothalamo-pituitary-gonadal feedforward and feedback control mechanisms that govern such age-related distinctions in the regulated output of the two primary gonadotropins will require further clinical study.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received March 26, 1999.
Revised May 25, 1999.
Accepted July 8, 1999.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-subunits by pituitary cells from old male rats. Am J Physiol.
249:E145–E151.
This article has been cited by other articles:
 |
|
 |
|
  P. Y. Liu, P. Y. Takahashi, P. D. Roebuck, and J. D. Veldhuis Age or Factors Associated with Aging Attenuate Testosterone's Concentration-Dependent Enhancement of the Regularity of Luteinizing Hormone Secretion in Healthy Men J. Clin. Endocrinol. Metab., October 1, 2006; 91(10): 4077 - 4084. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Iranmanesh and J. D. Veldhuis Combined Inhibition of Types I and II 5 {alpha}-Reductase Selectively Augments the Basal (Nonpulsatile) Mode of Testosterone Secretion in Young Men J. Clin. Endocrinol. Metab., July 1, 2005; 90(7): 4232 - 4237. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Matsumoto Andropause: Clinical Implications of the Decline in Serum Testosterone Levels With Aging in Men J. Gerontol. A Biol. Sci. Med. Sci., February 1, 2002; 57(2): M76 - 99. [Full Text]
|
|
|
|
|
|
J. A. Schnorr, M. J. Bray, and J. D. Veldhuis Aromatization Mediates Testosterone's Short-Term Feedback Restraint of 24-Hour Endogenously Driven and Acute Exogenous Gonadotropin-Releasing Hormone-Stimulated Luteinizing Hormone and Follicle-Stimulating Hormone Secretion in Young Men J. Clin. Endocrinol. Metab., June 1, 2001; 86(6): 2600 - 2606. [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 |
