Determinants of GH-releasing hormone and GH-releasing peptide synergy in men

Abstract

Age, sex steroids, and abdominal-visceral fat (AVF) jointly affect pulsatile growth hormone (GH) secretion. Pulsatile GH secretion in turn is controlled by GH-releasing hormone (GHRH), GH-releasing peptide (GHRP), and somatostatin. Marked stimulation of pulsatile GH secretion is achieved via GHRH-GHRP synergy. Nonetheless, how key modulators of GH secretion, such as age, sex steroids, and body mass index, modify GHRH-GHRP synergy is not known. The present strategy was to 1) infuse GHRH and GHRP-2 simultaneously to evoke synergy and 2) downregulate the gonadal axis with leuprolide and then restore placebo (Pl) or testosterone (T) to clamp the sex steroid milieu. Forty-seven men [18–74 yr of age, T = 7–1,950 ng/dl, estradiol (E₂) = 5–79 pg/ml, insulin-like growth factor (IGF)-I = 115–817 μg/l, AVF = 11–349 cm²] were studied. GHRH-GHRP synergy correlated negatively with age and AVF (both P < 0.001) and positively with IGF-I (P < 0.001) and IGF-binding protein (IGFBP)-3 (P = 0.031). Unstimulated basal (nonpulsatile) GH secretion correlated positively with T (P = 0.015) and E₂ (P = 0.004) concentrations. Fasting pulsatile GH secretion varied negatively with age (P = 0.017) and positively with IGF-I (P = 0.002) and IGFBP-3 (P = 0.001). By stepwise forward-selection multivariate analyses, AVF, IGF-I, and IGFBP-3 together explained 60% of the variability in GHRH-GHRP synergy (P < 0.001), E₂ accounted for 17% of the variability in basal GH secretion (P = 0.007), and IGF-I explained 20% of the variability in fasting pulsatile GH secretion (P = 0.002). In conclusion, a paradigm examining GHRH-GHRP synergy under a sex steroid clamp reveals highly selective control of basal, pulsatile, and synergistic peptide-driven GH secretion by AVF, E₂, and IGF-I in healthy men.

pulsatile secretion accounts for >85% of total daily growth hormone (GH) production (20). Burst-like GH release conveys significant physiological information to target tissues, including insulin-like growth factor I (IGF-I) production (22, 40, 41). Conversely, attenuation of GH pulse-dependent STAT5b signaling lowers IGF-I concentrations and causes growth failure (11, 20, 40). From a mechanistic perspective, a triad of pivotal peptides controls the size of GH secretory bursts: two amplifying signals, GH-releasing hormone (GHRH) and GH-releasing peptide (GHRP), and an inhibitory signal, somatostatin (SS) (4, 6, 16, 29, 34). Factors such as sex steroids, age, body composition, IGF-I, and GH feedback, thyroxine, glucocorticoids, free fatty acids, cytokines, neurotransmitters, exercise, and glycemia can modulate pulsatile GH secretion putatively by altering the release and/or actions of GHRH, ghrelin/GHRP, and SS (20, 31, 40). How such factors influence GHRH-GHRP synergy is not known.

The majority of investigations in humans and animals have evaluated the individual actions of GHRH, GHRP, or SS on GH secretion (3, 21). Some analyses have combined l-arginine with GHRH or GHRP, since l-arginine can antagonize GH feedback-induced hypothalamic SS outflow (2, 10, 19, 23, 33). A few studies have used simultaneous GHRH and GHRP infusions to evaluate GH reserve in potentially hypopituitary patients (4, 24, 36, 38). Transgenic, biomathematical, pharmacological, electrophysiological, and clinical models establish that GHRH and ghrelin/GHRP stimulate GH secretion synergistically (13, 16, 28, 34, 40). However, very little is known about the physiological factors that modulate GHRH-GHRP synergy in healthy individuals (36, 38). The present study evaluates modulation of two-peptide synergy by age, sex steroids, body composition, IGF-I, and IGF-binding protein (IGFBP)-3 concentrations during experimental sex steroid depletion and repletion in healthy men.

METHODS

Subjects.

Volunteers provided written informed consent approved by the Mayo Institutional Review Board and reviewed by the US Food and Drug Administration under investigator-initiated new drug numbers for GHRP-2 and GHRH. Exclusion criteria were exposure to psychotropic or neuroactive drugs within five biological half-lives; body mass index (BMI) <18 and >32.5 kg/m²; anemia (hemoglobin <12.8 g/dl); recent or active drug or alcohol abuse, psychosis, depression, mania, or anxiety state; acute or chronic organ system disease; use of testosterone (T), other anabolic steroids, or glucocorticoids; endocrinopathy, other than primary thyroid failure treated with l-thyroxine replacement; night-shift work or recent transmeridian travel (exceeding 3 time zones within 7 days of study); acute weight change (loss or gain of >2 kg in 6 wk); allergy to administered peptides; and unwillingness to provide written informed consent. Each subject had an unremarkable medical history and physical examination and normal screening laboratory tests of hepatic, renal, endocrine, metabolic, and hematological function. The men reported normal sexual development and function.

Visceral fat mass.

Intra-abdominal-visceral fat (AVF) mass was estimated by single-slice abdominal CT scan at L₃–L₄, exactly as reported previously (35).

Protocol.

The study design was parallel cohort, double blind, and prospectively randomized. Twenty-four healthy young men and 23 healthy older men each received two consecutive intramuscular injections of depot leuprolide acetate (3.75 mg, 3 wk apart) to deplete systemic T and estradiol (E₂) concentrations. Beginning on the day of the second leuprolide injection, volunteers were given saline (n = 13 young, n = 12 older) or a pharmacological dose of 200 mg of T enanthate (n = 11 young, n = 11 older) intramuscularly weekly in three doses (designated days 0, 7, and 14) in prospective, randomized double-blind order. The dual-secretagogue infusion was scheduled during days 12–16. Each participant was studied in the Clinical Translational Research Unit (CRU) after a standardized meal and an overnight fast.

Subjects were admitted to the CRU before 1700 and stayed overnight. To limit nutritional confounds, a constant meal (vegetarian or nonvegetarian; 8 kcal/kg, distributed as 50% carbohydrate, 20% protein, and 30% fat) was given in the CRU at 1800 on the night before the study. Volunteers then fasted and abstained from alcohol and caffeine overnight until the end of the infusion on the next day.

In the CRU, catheters were placed in contralateral forearm veins at 0700 to allow simultaneous infusion of secretagogues and blood sampling every 10 min for a total of 6 h beginning at 0800. Sampling encompassed a 3-h baseline and a 3-h stimulation interval. The latter comprised combined continuous intravenous infusions of 1 μg·kg⁻¹·h⁻¹ GHRH (GEREF, Serono, Norwalk, MA) and 1 μg·kg⁻¹·h⁻¹ GHRP-2 (Takeda Pharmaceuticals, Deerfield, IL), which are maximally stimulatory in adults (3, 38).

Blood was also withdrawn at 0800 for later assay of serum E₂, T, luteinizing hormone (LH), follicle-stimulating hormone (FSH), IGF-I, IGFBP-1, IGFBP-3, albumin, and sex hormone-binding globulin (SHBG) concentrations. Lunch was provided after sampling before discharge from the CRU.

Hormone assays.

Serum GH concentrations were determined in duplicate by automated ultrasensitive two-site immunoenzymatic chemiluminescence assay performed on the DxI automated system (Beckman Instruments, Chaska, MN). Interassay coefficients of variation (CVs) were 6.1, 4.3, 5.0, and 4.8% at 0.46, 3.0, 7.2, and 13.6 μg/l, respectively. Intra-assay CVs were 4.7, 3.5, and 3.2% at 0.37, 2.5, and 14.8 μg/l, respectively. The lowest detectable GH concentration at 95% confidence is 0.008 μg/l determined by processing a 6-point calibration curve, 5 quality controls, and 10 replicates of zero calibrator in multiple assays.

E₂ and T were measured by tandem liquid chromatography-ion spray mass spectrometry. For E₂, intra-assay CVs were 18, 3.8, and 7.2% at 3.6, 40, and 297 pg/ml (multiply by 3.68 for pmol/l). Interassay CVs were 8.1, 4.7, and 4.9% at 16, 31, and 119 pg/ml, respectively. For T, the analytic range is 7–2,000 ng/dl (multiply by 0.0347 for nmol/l) for a 0.1-ml volume. Intra-assay CVs were 3.3, 2.8, 2.2, and 2.0% at T concentrations of 16, 64, 184, and 927 ng/dl, respectively; corresponding interassay CVs were 5.1, 3.8, 3.7, and 2.8%.

IGFBP-1, IGFBP-3, and total IGF-I concentrations were quantified by immunoradiometric assays (Diagnostic Systems Laboratories, Webster, TX). Intra-assay CVs for IGF-I were 3.4, 3, and 1.5% at 9.4, 55, and 264 μg/l, respectively, and interassay CVs were 9 and 6.2% at 64 and 157 μg/l.

Statistical analysis.

One-way ANOVA was used to compare baseline (unstimulated) hormone concentrations in the four groups. Two-way ANOVA (2 × 2 factorial design) was used to examine the individual and interactive effects of normal vs. low T or E₂ (2 independent factors) and age (young vs. older) on the summed mass of GH secreted in pulses over the 3 h after secretagogue infusion. Post hoc contrasts were made via Tukey's honestly significantly difference test (15). Linear regression analysis was applied to examine the relationship between GH secretion and age, AVF, and IGF-I, IGFBP-1, and IGFBP-3 concentrations.

Values are means ± SE. Experiment-wise P < 0.05 was construed as statistically significant.

Statistical power.

Data from 18 studies in hypogonadal or normal men (n = 149 subjects total) indicate that parenteral T supplementation increases mean GH concentrations by a weighted-mean effect size [standard deviations (SDs)] of 1.8 (7, 8, 12, 17, 18, 25, 30, 32, 37, 39). In women, the effect size of age on GHRH/GHRP-2 stimulation was 1.2 SDs (38). Power analysis assumed that T or E₂ depletion and age in men mediate effects of similar magnitudes. If comparisons are made by two-way ANOVA at P < 0.05, then analysis of data from 45 subjects would achieve >90% power to detect the smaller effect size.

Deconvolution analysis.

GH concentration-time series were analyzed using a recently developed automated deconvolution method, which was verified mathematically by direct statistical proof and validated empirically by hypothalamopituitary sampling and simulated pulsatile time series (9). First, the Matlab-based algorithm detrends the data and normalizes concentrations to the unit interval [0,1] (27). Second, the program creates multiple sets of potential pulse times via an incremental smoothing process (a nonlinear adaptation of the heat-diffusion equation). Third, a maximum-likelihood estimation method calculates all secretion and elimination parameters for each of the multiple candidate pulse-time sets. Deconvolution parameters comprise basal secretion (β₀), two half-lives (α₁, α₂), secretory-burst mass (η₀, η₁), random effects on burst mass (σ_A), procedural/measurement error (σ_ε), and a three-parameter flexible gamma secretory-burst waveform (β₁, β₂, β₃). The fast half-life of GH was represented as 3.5 min, constituting 37% of the decay amplitude, and the slow half-life as 20.8 min (14). The Akaike information criterion (1) is used to select the optimal pulse-time set from the multiple candidate sets. Interpulse intervals are then described by a two-parameter Weibull process (a more general form of a Poisson process), which defines the frequency (number of bursts per sampling period, λ of a Weibull distribution) and the regularity of interpulse intervals (unitless γ of Weibull). Other parameters are basal and pulsatile secretion rates (concentration units/session), mass secreted per burst (concentration units), and waveform mode (time delay to maximal secretion after burst onset).

RESULTS

Mean ages of the young and older cohorts were 24 ± 0.72 yr (n = 24) and 60 ± 1.6 yr (n = 23), respectively. The age range overall was 18–74 yr. BMI and AVF were higher in older subjects (Table 1). The overall range of BMI was 20–34 kg/m², and that of AVF was 11–349 cm². Comparison of baseline (pre-leuprolide) endocrine data in the young and older cohorts disclosed similar concentrations of total T, LH, thyroid-stimulating hormone, and IGFBP-1. Older men had lower IGF-I, IGFBP-3, albumin, prolactin, bioavailable T, and free T and higher SHBG and FSH concentrations than young men.

Table 1.

Baseline hormone data

	Preintervention Measurements				P
		Young		Older
	Pl (n =13)	T (n =11)	Pl (n =12)	T (n =11)
Age, yr	23±1.1*	24±0.99*	57±1.7†	64±2.4†	< 0.001
BMI, kg/m2	25±0.82	24±0.86	27±1.0	26±0.84	0.087
IGFBP-1, μg/l	14±2.7	16±3.2	22±3.8	15±2.2	0.242
IGFBP-3, μg/l	4,732±222*†	4,876±283*	3,994±156†‡	3,710±175‡	< 0.001
IGF-I, μg/l	434±35*	420±27*	242±24†	264±37†	< 0.001
SHBG, nmol/l	23±3.3*	27±3.5*†	36±4†	40±4.4†	0.002
Albumin, g/dl	4.7±0.053*	4.7±0.089*†	4.3±0.11†‡	4.2±0.065‡	< 0.001
LH, IU/l	3.9±0.45	3.2±0.39	3.3±0.51	3.5±0.75	0.657
FSH, IU/l	3.5±0.37*	3.3±0.63*	5.0±0.80*†	7.0±1.0†	0.003
Prolactin, μg/l	9.0±0.92*†	12±2.1*	7.0±0.56†	7.9±0.69*†	0.040
E2, pg/ml	31±2.2	25±1.5	25±2.8	25±3.3	0.149
T, ng/dl
Total	595±45	548±44	533±70	503±51	0.258
Bioavailable	263±29*	194±17*	130±14†	107±11†	< 0.001
Free	20±2.2*	15±1.2*†	11±1.1†‡	9.0±0.86‡	< 0.001
TSH, mIU/l	2.5±0.40	1.6±0.18	2.6±0.42	2.8±0.52	0.063

Values are means ± SE. BMI, body mass index; IGFBP, insulin-like growth factor (IGF)-binding protein; SHBG, sex hormone-binding globulin; LH, luteinizing hormone; FSH, follicle-stimulating hormone; E₂, estradiol (multiply by 3.68 to convert to pmol/l); T, testosterone (multiply by 0.0347 to convert to nmol/l); TSH, thyroid-stimulating hormone. P values were determined by ANOVA. Means in any given row with unshared (unique) symbols ( *, †, ‡) differ significantly by Tukey's post hoc test: * differs from †, but neither * nor † differs from *‡.

Figure 1A presents endocrine measurements obtained after leuprolide administration with restoration of placebo (Pl) vs. T in young and older individuals. Age per se did not influence T, E₂, or LH concentrations when separately assessed in the hypogonadal and eugonadal contexts. The overall range of T was 7–1,950 ng/dl (all subjects), and of that of E₂ was 5–79 pg/ml. FSH was suppressed less after T in older than young men. IGF-I and IGFBP-3 concentrations were lower in older than younger individuals, whether Pl or T was restored (Fig. 1B). For the combined groups, IGF-I ranged from 115 to 817 μg/dl and IGFBP-3 from 2,705 to 5,822 μg/l. Higher prolactin concentrations were induced by T than by Pl supplementation in both age groups. Age elevated IGFBP-1 and SHBG concentrations only in the low-T/E₂ milieu. Albumin was lower in older men receiving leuprolide with T than in any other group.

Fig. 1.

Concentrations of sex steroids and gonadotropins (top) and extragonadal hormones (bottom) in 47 men stratified by age (young and older) and randomly assigned to 3 wk of placebo (Pl) or testosterone (T) repletion after administration of the gonadotropin-releasing hormone agonist (leuprolide). LH, luteinizing hormone; FSH, follicle-stimulating hormone; IGF-I, insulin-like growth factor I; IGFBP, IGF-binding protein; SHBG, sex hormone-binding globulin. Values are means ± SE. P values were determined by ANOVA. Different (unshared) superscripts (A, B, C) denote significantly different means by Tukey's post hoc test.

Simultaneous infusion of GHRH and GHRP-2 evoked marked time-dependent release of GH (Fig. 2A). Peak GH concentrations differed significantly among cohorts (P = 0.005 by ANOVA), with maximal values in young men given leuprolide with T (P = 0.006 vs. older men given leuprolide + Pl and P = 0.008 vs. older men given leuprolide + T; Table 2). Peak GH concentrations did not differ among the young men given leuprolide with Pl, young men given leuprolide with T, and older men given leuprolide with T. Combined-peptide stimulation of pulsatile GH secretion as assessed by deconvolution analysis yielded the same descending order of responses (P = 0.001 by ANOVA; Fig. 2B). Unstimulated pulsatile GH secretion was 5.9-fold higher in the young than the older men given leuprolide with T (P = 0.045 ANOVA). Basal (nonpulsatile) GH secretion and secretory-burst shape (mode) did not differ among the four study groups (Table 2).

Fig. 2.

A: time course of growth hormone (GH) concentrations monitored every 10 min for 3 h during saline infusion followed by 3 h of simultaneous infusion of GH-releasing hormone (GHRH) and GH-releasing peptide (GHRP)-2 (both at 1 μg·kg⁻¹·h⁻¹). B: pulsatile GH secretion estimated in the 4 groups of volunteers receiving combined GHRH-GHRP-2 infusion. Values are means ± SE. P values reflect ANOVA and Tukey's post hoc comparisons.

Table 2.

ANOVA of pulsatile GH secretion under GHRH-GHRP-2 drive

End Point	Young		Older		P
	Pl (n =13)	T (n =11)	Pl (n =12)	T (n =11)
Peak GH concn, μg/l	64±9.0*‡	80±14*	34±5.1†	55±13†‡	0.005
GH secretion, μg·l−1·3 h−1
Pulsatile	253±43*‡	293±30*	113±15†	182±45†‡	0.001
Basal	3.8±1.5	3.1±1.3	2.2±0.55	3.9±0.85	0.48
Secretory-burst mode, min	15±1.4	15±0.67	13±1.5	16±2.4	0.79

Values are means ± SE. GH, growth hormone; GHRH, GH-releasing hormone; GHRP, GH-releasing peptide. P values were determined by ANOVA. Different (unshared) symbols ( *, †, ‡) within a row denote significant post hoc differences by Tukey's test.

Regression analysis was used to evaluate modulation of GHRH-GHRP-2 synergy (dependent variable) by AVF, age, and sex steroid concentrations (independent variables). The effect of continuous combined GHRH-GHRP-2 stimulation on pulsatile GH secretion was explained univariately by age (negatively, R² = 0.24, P < 0.001), AVF (negatively, R² = 0.40, P < 0.01), IGF-I (positively, R² = 0.39, P < 0.001), and IGFBP-3 (positively but weakly, R² = 0.10, P = 0.031; Fig. 3A). Bivariate regression analysis revealed that AVF and IGF-I jointly predict pulsatile GH secretion during GHRH-GHRP-2 drive (overall R² = 0.54, P < 0.0001; Fig. 3B). Individual partial R² contributions were highly significant (P = 0.0006 for AVF and P = 0.0013 for IGF-I). On the other hand, the positive individual effects of IGFBP-3 and IGF-I were accounted for fully by IGF-I in bivariate regression (R² = 0.41, P < 0.001). Analogously, when age and IGF-I were allowable independent variables, IGF-I emerged as the dominant predictor of GHRH-GHRP-2 synergy (R² = 0.47, P < 0.001). When age or IGFBP-3 was paired with T as possible joint determinants of GHRH-GHRP-stimulated GH secretion, age and IGFBP-3 alone were the respective primary correlates (both P < 0.001, R² = 0.44 and R² = 0.26).

Fig. 3.

A: linear regression of unstimulated (saline-infused; top) or GHRH-GHRP-2-stimulated (bottom) pulsatile GH secretion on age [young (Y) and older (O)]. Data point enclosed in a box was a statistical outlier at P < 0.001 based on Studentized residuals. B: joint linear regression of GHRH- and GHRP-2-stimulated pulsatile GH secretion (vertical axis) on IGF-I concentrations (horizontal axis) and abdominal-visceral fat (AVF; oblique axis) in 47 men. Bivariate R² and P values, along with partial P values and slopes for the individual covariates IGF-I and AVF, are shown.

On the basis of the outcomes of univariate and bivariate regression, stepwise forward-selection multivariate analyses was applied. This approach indicated that AVF (negatively), IGF-I (positively), and IGFBP-3 (negatively) together explained 59.8% of the variability in GHRH-GHRP synergy (P < 0.0001).

Basal GH secretion correlated positively with total T (R² = 0.14, P = 0.015) and total E₂ (R² = 0.20 P = 0.004) by univariate analysis. AVF, age, IGF-I, and IGFBP-3 were not significant determinants. Bivariate regression disclosed that age (P = 0.007) and IGF-I (P = 0.034) were joint predictors of basal GH release (R² = 0.18, P = 0.021). Stepwise forward-selection multivariate regression identified E₂ as the principal predictor of basal GH output (multivariate R² = 0.17, P = 0.007).

According to univariate regression, pulsatile GH secretion during saline infusion was related negatively to age (R² = 0.15 P = 0.017) and positively to IGF-I (R² = 0.20, P = 0.002) and IGFBP-3 (R² = 0.21, P = 0.001) concentrations. Bivariate regression analysis revealed that age negatively (P = 0.003) and T positively (P = 0.053) together predicted 20% of the variability in fasting pulsatile GH secretion (overall P = 0.0077; Fig. 4). Age and E₂ together acted similarly (R² = 0.19, P = 0.014; age P = 0.006, E₂ P = 0.060), but the E₂ trend was not significant. Multivariate analyses showed that IGF-I was the primary statistical determinant of fasting pulsatile GH secretion (R² = 0.20, P = 0.002).

Fig. 4.

Conjoint negative effect of age (oblique axis) and positive effect of T concentrations (horizontal axis) on unstimulated (saline-infused) pulsatile GH secretion (vertical axis) in 47 healthy men. See Fig. 3 legend for format.

DISCUSSION

Salient aspects of this work are 1) introduction of an experimental paradigm of T/E₂ depletion and repletion in young and older individuals to dissect the separate and interrelated effects of age and sex steroids on GH secretion, 2) combined continuous intravenous infusion of maximally stimulatory doses of GHRH and GHRP-2 over 3 h to appraise secretagogue synergy, 3) inclusion of men with a wide range of body compositions, so as to estimate the concomitant influence of AVF, and 4) use of a validated deconvolution method to separately quantify pulsatile and basal GH secretion. Under these conditions, stepwise forward-selection multivariate analysis revealed that 1) AVF (negatively), IGF-I (positively), and IGFBP-3 (positively) jointly determine pulsatile GH secretion stimulated by GHRH-GHRP-2, accounting for 60% of total response variability; 2) E₂ concentrations constitute a positive correlate of basal (nonpulsatile) GH secretion independently of age, AVF, and IGF-I, IGFBP-3, and T concentrations, explaining 17% of interindividual variations in this measure; and 3) IGF-I concentrations positively predict fasting pulsatile GH secretion during saline infusion, representing 20% of intersubject variance. These outcomes define distinct physiological modulators of basal, fasting pulsatile, and synergistically GHRH-GHRP-driven pulsatile GH secretion.

Two-way ANOVA disclosed that age and T/E₂ availability influence the magnitude of GHRH-GHRP-2 synergy (overall P < 0.001 with no statistical interaction). In the low-T/E₂ milieu, the magnitude of dual-peptide synergy in older men was 45% of that in young men (P = 0.017). The age-related deficit was less striking in the replaced-T/E₂ milieu (older/young response percentage 62%, P = 0.055). The lack of an age × T/E₂ interaction at good (>90%) statistical power could indicate that young and older men respond similarly to a change in sex steroid milieu. This postulate would be consistent with dose-response studies showing age-comparable effects of T on skeletal muscle (5). GHRH-GHRP synergy is putatively important physiologically when GHRH and ghrelin are jointly available in high concentrations (40) and when SS outflow is low (20).

Multivariate analysis of primary factors determining GHRH-GHRP-2 synergy unveiled composite contributions by AVF (negative), IGF-I (positive), and IGFBP-3 (negative), with overall R² = 0.60 and each partial P ≤ 0.01. The relative strengths of different determinants can also be expressed by standardized regression coefficients [slope normalized to SD of the independent variable (42)]. Absolute values of standardized coefficients were (10⁻³ slope units) 10.3, 3.87, and 0.074, for AVF, IGF-I, and IGFBP-3, respectively. Whereas the contribution of IGF-I was positive, that of IGFBP-3 was negative. The latter finding could suggest that IGFBP-3 mutes pulsatile GH responses to GHRH and GHRP. In this regard, transgenetic silencing of IGFBP-3 expression can enhance linear growth, suggesting an inhibitory effect of the binding protein. Inhibition might be mediated by direct cellular effects of IGFBP-3 and/or by sequestration of otherwise unbound (free) IGF-I (40).

Univariate regression analysis identified E₂ (R² = 0.31) and T (R² = 0.14) as significant individual positive determinants of basal GH secretion. Bivariate regression showed that age and IGF-I concentrations were positive correlates. When age, AVF, T, E₂, IGF-I, and IGFBP-3 were allowed to enter into stepwise forward-selection multivariate regression, E₂ emerged as the dominant positive correlate of basal GH secretion (R² = 0.17). With the assumption of a GH distribution volume of 57 ml/kg in men (20), each 10 pg/ml increase in E₂ concentration would predict an increment of 26 μg/day of basal GH secretion. The positive correlation between E₂ and basal GH secretion might reflect direct estrogenic stimulation of somatotrope secretion, as recognized in pituitary cell cultures and heterotopic transplants (26).

By univariate assessment, age (negatively, R² = 0.13), IGF-I (positively, R² = 0.20), and IGFBP-3 (positively, R² = 0.21) predicted unstimulated pulsatile GH secretion. In bivariate analyses, the principal paired correlates were age (negative) and T (positive), with overall R² = 0.20 and P = 0.008. Each 100 ng/dl decrement in T or 5-yr increment in age was associated with respective decreases in unstimulated pulsatile GH secretion of 6.4 and 35 μg/day. According to multivariate regression analysis, IGF-I was the dominant determinant (R² = 0.20, P = 0.002), such that each 100 μg/l increment in IGF-I translated into 45 μg more GH secreted daily in pulses. The correspondence between IGF-I concentrations and pulsatile GH secretion may reflect the fact that GH pulses activate STAT5b signaling and that STAT5b mediates stimulation of IGF-I synthesis (11).

Unresolved questions include whether longer-term T/E₂ depletion would exert comparable or greater effects. In addition, how AVF, age, and sex steroids together influence feedback by GH and IGF-I remains to be defined. Paradigms comparable to that presented here should have utility in further dissecting multifactorial mechanisms that regulate GH secretion.

In summary, a paradigm of fixed maximal combined GHRH-GHRP stimulation and experimentally controlled T/E₂ milieus in healthy men reveals E₂-predominant regulation of basal (nonpulsatile) GH secretion; a strong positive association between IGF-I and fasting pulsatile GH secretion; and conjoint AVF (negative), IGFBP-3 (negative), and IGF-I (positive) determination of GHRH-GHRP synergy. These outcomes indicate the need for more extended physiological models of GH-axis regulation that incorporate individual and interactive effects of age, T, E₂, AVF, IGF-I, and IGFBP-3.

GRANTS

This work was supported in part by Center for Translational Science Activities Grant 1 UL 1 RR-024150 to the Mayo Clinic and Foundation from the National Center for Research Resources and National Institute on Aging Grant R01 AG-19695.