| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Department of Psychiatry and Sleep Laboratory, Erasme Hospital (J.M., P.L., M.K.), Center for the Study of Biological Rhythms (R.L., G.C.), Laboratory of Experimental Medicine (G.C.), Université Libre de Bruxelles, B-1070 Brussels, Belgium; and the Department of Medicine, University of Chicago (E.V.C.), Chicago, Illinois 60637
Address all correspondence and requests for reprints to: Julien Mendlewicz, M.D., Ph.D., Department of Psychiatry, Erasme Hospital, Université Libre de Bruxelles, 808 route de Lennik, B-1070 Brussels, Belgium. E-mail: jmendlew{at}ulb.ac.be
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
In recent years, growing interest has focused on the importance of genetic factors in the regulation of neuroendocrine systems (7, 8), but few studies are available on the genetic regulation of the somatotropic axis (9, 10, 11). Recent molecular studies have localized the human GH gene cluster on chromosome 17 (12) and identified the human GHRH receptor gene (13). Several genetic syndromes relating short stature and GH insufficiency have been described (14). GH has a major effect on height, and the estimated heritability of height reaches 92% (10).
The present study was designed to determine the relative contributions of genetic and nongenetic factors on individual differences in GH secretion during waking and sleep. The 24-h profile of plasma GH was obtained at 15-min intervals in a sample of 10 monozygotic and 9 dizygotic twin pairs. Sleep was polygraphically monitored, and pulsatile GH secretion was estimated by mathematical deconvolution. A statistical method specifically designed for the analysis of twin studies was used for the identification of genetic and environmental influences on GH secretion and sleep.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Male twins were selected from the Twin Register of the University of Antwerp (Antwerp, Belgium) and the Twin Register of the Vrije Universiteit Brussel (Brussels, Belgium) after a careful medical evaluation. Criteria for eligibility included normal health and the absence of personal or family history of endocrine or psychiatric disorder. As GH secretion is markedly influenced by sleep (3, 4, 5), only subjects with regular sleep-wake schedules and absence of sleep complaints were included. Twin pairs in which one member was taking drugs, was a shift worker, or had traveled across time zones during the past 3 months were excluded. The protocol was approved by the institutional review board. Informed consent was obtained from all subjects and from their parents if they were under 18 yr of age. The volunteers were paid for their participation in the study. The twins were classified as mono- (MZ) and dizygotic (DZ) after analysis of different genetic markers, including HLA and ABO, Rh, MnSs, Kk, Lea, Leb, Fya, Fyb, Jka, Jkb, and P1 blood groups. Other details of subject recruitment have been previously reported (8, 15).
A total of 10 MZ pairs (aged 16–30 yr; mean, 21.4 yr) and 9 DZ pairs (aged 20–34 yr; mean, 25.2 yr) were included in the present study. All subjects had reached stage Tanner V of sexual development. In the MZ group, 8 of the 10 twin pairs were living together. In the DZ group, 4 of the 9 twin pairs were cohabiting. All subjects were living in the same geographical area (within a 50-km radius from the investigation center).
Experimental protocol
All investigations were performed in the Sleep Laboratory of the Department of Psychiatry, Erasme Hospital, Université Libre de Bruxelles (Brussels, Belgium). Both members of each pair were studied simultaneously in separate rooms located in the immediate vicinity of each other. On admission, all subjects had a physical examination, including measurements of body weight and height, and routine laboratory tests. All were found to be in normal health. After 1 night of habituation, sleep was polygraphically recorded during 4 consecutive nights. On the day preceding the last night of recording, a catheter was inserted into a forearm vein between 1200–1400 h. Blood sampling for GH determinations was started 1 h after catheter insertion, and blood samples were obtained at 15-min intervals for 25 h. Data collected during the first hour of sampling were discarded to avoid artifactual effects related to the venipuncture stress. During the night, the catheter was connected to plastic tubing extending into an adjacent room, and sampling was thus performed without disturbing the subject. The iv line was kept patent with a slow drip (10 mL/h) of heparinized saline (750 IU heparin in 0.9 g NaCl/dL). All subjects were ambulatory during the day and were fed the standard hospital diet (breakfast at 0800 h, lunch at 1230 h, dinner at 1900 h). Daytime naps were prevented. The subjects were asked to retire around 2230 h and were allowed to wake up spontaneously in the morning. During bedtime hours, the lights were turned off.
GH assay
Duplicate determinations of plasma GH concentrations were performed using a polyclonal antibody RIA with a lower limit of sensitivity of 0.4 µg/mL (16). All samples from the same twin pair were analyzed in the same assay. The intraassay coefficient of variation averaged 9% in the range 0.4–2.0 µg/L, 6% in the range 2.0–5.0 µg/L, and 5% above 5.0 µg/L. The interassay coefficient of variation averaged 15%.
Determination of GH secretory rates
Significant pulses of GH secretion were identified using a modification of the computer algorithm ULTRA (17). The threshold for significance of a pulse was set at twice the intraassay coefficient of variation in the relevant concentration range. For each significant pulse, the amount of GH secreted was calculated by deconvolution based on a one-compartment model for GH clearance and variable individual half-lives, as previously described (5). The half-life was adjusted for each pair of subjects in the previously reported physiological range of 15–21 min (18) by an iterative process designed to minimize the number of negative secretory rates. On the average, the half-disappearance time was 18.4 ± 2.3 min (mean ± SD). A volume of distribution of 7% of body weight was used in these calculations. The SD of the error associated with each estimated secretory rate was calculated following the theory of error propagation and under the assumption of normally distributed errors on plasma levels. The duration of a secretory pulse was defined as the time interval separating the preceding and following troughs. In each individual profile, the level of baseline secretion was estimated as the secretory rate necessary to maintain the baseline GH concentrations during the interpulse intervals. For each significant pulse, pulsatile GH secretion was calculated by subtracting the baseline secretion from the total secretion. The amount of pulsatile GH secretion over a given time interval was determined by summing the amounts of pulsatile secretion in each of the significant pulses occurring during that time interval. If a pulse overlapped two time intervals, the amount of GH secreted was divided proportionally.
Sleep recording and analysis
The polygraphic recordings of sleep were visually scored at 20-s intervals in stages wake, I, II, III, IV, and rapid eye movement (REM) according to standardized criteria (19). Sleep onset and morning awakening were defined, respectively, as the first and last 20-s intervals, scored II, III, IV, or REM. The sleep period was defined as the time interval separating sleep onset from morning awakening. SW stages were defined as the sum of stages III and IV. The twins included in the present study were a subset of a larger sample of 26 pairs of twins for whom a detailed analysis of genetic effects on sleep parameters has been previously reported (15). As in adult normal men, nocturnal GH secretion is influenced strongly by sleep, we repeated for the subset of subjects included in the present study the analysis of genetic effects for the durations of stages wake, I+II, III+IV (SW sleep), and REM.
Statistical analysis of genetic variance
The twin values for each of the analyzed parameters were submitted to the TWINAN method for analysis of twin data developed by Christian et al. (20, 21, 22, 23). This method is applicable only if mean values for the investigated parameter are not significantly different between MZ and DZ twins. Therefore, for each parameter, possible differences between both groups were tested using a two-tailed unpaired t test. When applied to small twin samples such as those in the present study, this analysis is relatively sensitive to outlying values. Therefore, for each parameter, the outlying values were identified using a two-tailed test (24, 25) with a significance level of 0.05. When a significant outlier(s) was identified, the analysis was repeated after excluding the outlying pair(s). A genetic effect on the variability of a given parameter was considered significant if the appropriate estimate of genetic variance was significant (P < 0.05) and if the result was not critically dependent on the inclusion of significant outlier(s).
Possible differences in the MZ/DZ variances of each parameter were tested using an F test for significance of differences. When the variances were similar in both groups, as indicated by an arbitrary cut-off significance level of F test with P > 0.20, the within-pair estimate of genetic variance (GWT) was calculated according to the formula GWT = MWDZ - MWMZ, with MWDZ and MWMZ being the mean squares for within-pair variation in DZ and MZ twins, respectively. The significance of GWT was tested by a one-tailed F = MWDZ/MWMZ. As this estimate may be biased if the variances in the MZ and DZ groups are not similar, another estimate, the among-component estimate of genetic variance GCT, was used when the F test had a P < 0.20, according to the formula GCT = (GWT + GAT)/2, with GAT = MAMZ - MADZ, MAMZ and MADZ being the mean squares for among-pair variation in MZ and DZ twins, respectively. The significance of GCT was tested by a two-tailed F' = (MAMZ + MWDZ)/(MADZ + MWMZ).
Among the outputs of the TWINAN analysis are the intraclass correlation coefficients, which estimate for the investigated parameter the similarity in the MZ and DZ twin pairs. The existence of a significant genetic effect is reflected typically in significant intraclass correlations in both the MZ and DZ groups, with a higher intraclass correlation in the MZ than in the DZ group. If, however, the intraclass correlation is significant in the MZ group but not in the DZ group, the existence of a higher environmental covariance in the MZ group, rather than of a genetic effect, could be suspected, because a genetic effect should be reflected in a significant level of correlation among the DZ twins, who are related genetically as siblings. It is also possible that relatively small DZ intraclass correlations may reflect the presence of gene interactions (22, 26). To examine the first possibility, the TWINAN method includes a test to exclude the existence of a MZ-DZ difference in environmental covariance. In the present study, differences in environmental covariance could be associated with the higher proportion of twins living together in the MZ group. Therefore, a two-factor ANOVA of the within-pair differences in the parameter under consideration, using zygosity and cohabitational status (i.e. living together or not) as cofactors, was also performed.
When a significant genetic effect was evidenced, the heritability estimate was calculated according to the formula: heritability = (intraclass correlation coefficient for MZ twins - intraclass correlation coefficient for DZ twins) x 2.
Unless otherwise indicated, all group values are reported as the mean ± SEM.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
. DZ twins
were slightly, but significantly, older than MZ twins, and their body
mass indexes (BMIs) was significantly higher. Therefore, possible
effects of age and BMI on the various GH and sleep parameters were
tested. As expected, the duration of SW sleep was inversely related to
age (P < 0.03), and all GH parameters were inversely
related to both age and BMI (P < 0.10 at least).
Therefore, GH values were adjusted by multiple linear regression
analysis for the effect of age and BMI, and SW sleep values were
adjusted by simple linear regression analysis for the effect of age.
For all these parameters, the TWINAN analysis was repeated on adjusted
data, and a genetic effect was considered significant if both the
initial analysis of original data and the repeated analysis of adjusted
data yielded consistently significant results.
|
GH secretion
All individual GH profiles exhibited the typical pattern of normal young men, with stable low levels abruptly interrupted by secretory pulses (5). As expected (5), a sleep-onset pulse associated with the first phase of SW sleep was present in all but one profile and constituted in most cases the major secretory episode. Overall, nearly two thirds of the total 24-h GH secretion occurred during the sleep period, which lasted, on the average, 7 h and 50 min. However, no significant relationship was observed between the individual amounts of sleep GH secretion and corresponding durations of SW sleep.
Representative 24-h plasma GH profiles in two MZ and two DZ twin pairs
are shown in Fig. 1. Visual examination
suggests a greater similarity of those profiles in MZ than in DZ twins
during the wake period and, to a lesser extent, during the sleep
period. A summary of the results of the genetic variance analysis is
given in Table 2.
|
|
(bottom). A significant
difference in variance between the two groups was present
(P = 0.02), primarily because of the contribution of a
single 24-yr-old twin pair with very high values in the DZ sample. This
pair failed to meet our criteria of significance for outlying values.
When the entire sample was analyzed, the intraclass coefficients of
correlation were high and similar in both groups (0.76 in MZ twins,
0.69 in DZ twins), and the applicable estimate of genetic variance did
not provide evidence for a genetic effect (P = 0.94).
Similar results were obtained after adjustment of GH values for the
effects of age and BMI (P = 0.94). Although these
results would suggest that GH secretion during sleep is more dependent
on environmental than genetic factors, caution must be exerted because
this conclusion rests entirely on the inclusion of the single DZ twin
pair with unusually higher sleep-associated GH secretion. Indeed, when
this pair is excluded, the difference in variance becomes
nonsignificant (albeit with a probability level of 0.14,
i.e. lower than the arbitrary cut-off point of 0.20), the
intraclass coefficients of correlations are 0.76 in MZ twins and 0.55
in DZ twins, and the within-pair estimate of genetic variance is
significant (P < 0.025). In view of these
considerations, the present data do not allow exclusion of a genetic
influence on GH secretion during the sleep period.
|
(top). The
variances were similar in both groups, and no outlier was detected. The
intraclass coefficient of correlation was higher in MZ twins (0.87)
than in DZ twins (0.50). The estimate of genetic variance was
significant (P < 0.05), and the heritability estimate
reached 0.74. The possibility of MZ-DZ differences in environmental
covariance could be excluded at a P = 0.06 level.
Similar conclusions were reached after adjustment of GH values for the
effects of age and BMI (genetic variance significant at
P = 0.09; heritability estimate of 0.67). Thus, the
present data provide consistent evidence for a major genetic influence
on daytime GH secretion. The 24-h pulsatile GH secretion was similar in MZ and DZ twins (573 ± 77 vs. 610 ± 98 µg; P = NS). The variances were similar in both groups, and no outlier was detected. The intraclass coefficient of correlation was somewhat higher in MZ twins (0.87) than in DZ twins (0.74), and the estimate of genetic variance was significant (P = 0.05), but the heritability estimate was only 0.27. The possibility of MZ-DZ differences in environmental covariance could be excluded (P < 0.005). Similar conclusions were reached after adjustment of GH values for the effects of age and BMI (genetic variance significant at P = 0.09; heritability estimate of 0.18). Thus, those data indicate that the 24-h GH secretion is partially influenced by genetic factors.
Sleep parameters
For all sleep parameters, mean group values were similar in
MZ and DZ twins (Table 1). The results of the analysis of genetic
variance are given in Table 2. No significant outlier was detected for
any of the sleep parameters except for the duration of wake stages.
Neither genetic nor environmental effects could be evidenced for stages
wake, I and II, or REM. A significant genetic effect was found for the
duration of stages III and IV (SW sleep), with a higher intraclass
coefficient of correlation in MZ twins (0.81) than in DZ twins (0.43),
an estimated genetic variance significant at P < 0.03,
and a heritability estimate of 0.76. The possibility of MZ-DZ
differences in environmental covariance could be excluded at a
P < 0.10 level. Similar conclusions were reached after
adjustment of SW values for age (genetic variance significant at
P < 0.02; heritability estimate of 0.91).
Anthropometric parameters
Differences in mean weight and height values were not significant
(Table 1). The results of genetic variance analysis are shown in Table 2. No significant outlier was detected.
For weight, the variances were similar in both groups. The estimate of genetic variance was close to significance (P = 0.08), with a heritability estimate of 0.66, and the possibility of a difference in environmental covariance could be excluded (P = 0.04), indicating that weight is partially influenced by genetic factors.
For height, the intraclass coefficient of correlation was markedly higher in MZ twins than in DZ twins (0.96 vs. 0.48), the estimated genetic variance was significant at P < 0.01, and the heritability estimate reached 0.96. The possibility of a difference in environmental covariance could be excluded at a P = 0.07 level.
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The existence of a possible genetic determination of GH release patterns was first suggested in 1974 by Parker and Rossman (9), who observed an apparent similarity in plasma 24-h GH profiles in one pair of identical twins, but not in dizygotic twins. Indirect evidence consistent with a genetic influence on GH secretion has been reported. Using twin studies, a genetic component has been demonstrated on circulating levels of insulin-like growth factor I (IGF-I) (10, 11, 27), which are GH dependent (28), and on height (10), which is strongly correlated with IGF-I levels (27). Genetic studies have identified the GHRH receptor gene (13), the GH gene cluster (12), and the IGF-I gene (29), and genetic syndromes relating short stature and growth hormone insufficiency have been described (14). Interestingly, we have recently shown that daytime secretion of PRL, a member of the PRL-GH gene family (30), is also under partial genetic control (31).
The high estimated heritability evidenced in the present study for wake GH secretion in young adults is consistent with the observation that in children, the variation in IGF-I levels is almost completely under genetic control (27). In elderly adults, however, the contribution of genetic factors to the regulation of IGF-I concentrations is much lower (10, 11). Aging is associated with a progressive increase in somatostatinergic tone (32), which is likely to be at least in part responsible for large decreases in both GH and IGF-I secretions (33, 34, 35). The relative influence on GH secretion of environmental factors such as nutrition and physical exercise could increase with aging, leading to a modulation of the expression of genetic control.
Despite the fact that multiple nongenetic factors are known to markedly affect human GH secretion, including nutritional status, stress, levels of physical activity, and sleep quality, the present study clearly demonstrates that genetic factors play an important role in determining the amount of pulsatile GH secreted in normal young adults during the daytime and, to a lesser extent, during the entire 24-h cycle.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Received August 5, 1998.
Revised November 23, 1998.
Accepted December 11, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  E.A.M. Kuijper, C.B. Lambalk, D.I. Boomsma, S. van der Sluis, M.A. Blankenstein, E.J.C. de Geus, and D. Posthuma Heritability of reproductive hormones in adult male twins Hum. Reprod., August 1, 2007; 22(8): 2153 - 2159. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Wagner, K. Hemminki, E. Grzybowska, R. Klaes, B. Burwinkel, P. Bugert, R. K. Schmutzler, B. Wappenschmidt, D. Butkiewicz, J. Pamula, et al. Polymorphisms in genes involved in GH1 release and their association with breast cancer risk Carcinogenesis, September 1, 2006; 27(9): 1867 - 1875. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K Wagner, K Hemminki, E Israelsson, E Grzybowska, R Klaes, B Chen, D Butkiewicz, J Pamula, W Pekala, and A Forsti Association of polymorphisms and haplotypes in the human growth hormone 1 (GH1) gene with breast cancer Endocr. Relat. Cancer, December 1, 2005; 12(4): 917 - 928. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. L. Anderson and N. G. Anderson The Human Plasma Proteome: History, Character, and Diagnostic Prospects Mol. Cell. Proteomics, November 1, 2002; 1(11): 845 - 867. [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 |
