Secretory patterns of growth hormone during basal periods in man

Secretory

Patterns

of Growth Hormone During Basal Periods in Man Mark W. Stolar and Gerhard Baumann

Human growth hormone (hGH) concentrations in plasma often fall to levels not detectable by RIA. These so-called basal levels prevail during the greater part of the day. Since hGH is involved in the homeostasis of several metabolic processes, it is important to examine its exact plasma concentration and secretory pattern during basal periods. We used an immunoadsorbent technique to extract hGH from large plasma samples to precisely measure basal hGH concentrations and their variation with time. Blood samples (20 mL) were drawn from 12 normal subjects in the fasted and rested state every 15 minutes over a three-hour period. Plasma hGH levels varied over three orders of magnitude (range, 34 to 60,000 pg/mL). During basal periods, episodes of secretory pulses, of moderate sustained secretion, and of complete secretory inactivity occurred. Women had significantly higher overall hGH levels as well as basal hGH levels than men, but no significant sex difference in the pulse frequency during basal periods could be detected in the limited time allotted for study. No convincing relationship was noted between variations in plasma glucose and the secretory pattern of hGH, or vice versa. We conclude that (I) hGH is secreted in an episodic fashion during basal periods. Conceptually, basal and stimulated hGH secretion may be viewed as extremes of a continuous spectrum of pituitary activity, (2) basal hGH levels are lower than heretofore appreciated, (3) the known tendency of women to higher hGH levels is also evident in the basal range, and (4) oscillations in plasma glucose do not affect the microsecretory pattern of hGH, nor are endogenous hGH pulses followed by acute changes in glycemia. Q 1986 by Grune & Stratton,

Inc.

GH levels in man vary over a wide range due to the episodic nature of pituitary secretion.‘.’ ThroughP out most of the day, plasma GH remains at basal levels, which are either undetectable or too low to be accurately measured by RIA. Because of this limitation, no precise knowledge exists about the minute-to-minute regulation of GH secretion or the concentrations to which tissues are exposed, except during the relatively infrequent secretory surges. GH secretion is believed to be regulated at the pituitary level by a dual control mechanism of two antagonistic principles, GH releasing hormone (GHRH) and somatostatin, but the line details of this interplay are presently unknown. llncertainty also prevails about the contribution of basal GH levels to metabolic homeostasis and somatic growth. Thus, it is desirable from both the standpoint of regulation of secretion and that of peripheral hormone action to gain knowledge about the exact GH concentration and its dynamic changes in the circulation during basal periods. Because of the relative insensitivity of RIA, achievement of this goal has been elusive. We have recently developed a technique that allows accurate measurement of plasma GH in the picomolar range.3 In the present study, we have applied this method to examine the dynamics of plasma GH and its relationship to plasma glucose during basal periods. LASMA

MATERIALS

Human

AND

METHODS

Suhject.c

lean subjects (six males and six females, aged 22 to 32. 92% to 109% of ideal body weight) participated in the study. In order to reproduce normally prevailing conditions. subjects were allowed normal activity prior to the experiment, but they avoided strenuous exercise None was taking any medication. They reported to the Clinical Research Center at 8 AM, having fasted since 10 PM the evening before. An indwelling intravenous catheter was placed in a forearm vem and kept open with a slow infusion of normal saline. The study was started after a one-hour rest period at 9 AM. The subjects remained supine during the study; they were allowed quiet mental activity, such as reading, but were not permitted to sleep. Their room remained well lit throughout the study. Blood samples Twelve normal

Metrbol~sm. ‘:oI

35

No 9 (September),

1986:

pp 883-888

(20 mL) were drawn through the indwelling catheter every 15 minutes for a three-hour period. Samples were placed into chilled heparinized test tubes and centrifuged at 4 “C. Plasma samples were frozen (-20 “C) and thawed once before analysis. The study had been approved by the Northwestern University and Northwestern Memorial Hospital Human Research Committees. All subjects gave informed consent. Assays Human GH (hGH) was determined in a preliminary fashion in all plasma samples by direct double antibody RIA.J Samples containing less than 2.5 ng/mL hGH were then subjected to immunoextracted radioimmunoassay (IERIA). This technique has been described in detail e1sewhere.j Briefly, hGH was extracted from 10 to 13 mL plasma by immunoadsorbent chromatography on an anti-hCH column. A small amount of “‘I-labeled hGH (0.2 to 0.8 ng) had been added to the plasma samples to estimate recovery. The columnbound hGH was eluted with 4 mol/L potassium thiocyanate. Plasma extracts were then assayed for hGH by RIA, using ‘2s1-hGH as a tracer. The concentration of hGH in the original plasma sample was calculated, taking the volume of plasma extracted. the recovery figures, and exogenously added “‘I-hGH into account. The critical technical considerations have been presented previously.’ Overall recovery of hGH in the immunoextraction procedure was 56.2% + 1.32% (mean + SEM). Each plasma extract was assayed in triplicate. After preliminary determination of approximate hGH concentration, all samples were assayed in the same assay run. using appropriate volumes to insure reading in the midrange of the standard curve (B/B, = 0.35 to 0.65). [(B/B,) is the ratio between the amount of tracer bound to antibody at a given ligand concentration (B) and the amount of tracer bound in the absence of unlabeled ligand (B,)]. -From the Center for Endocrinolog.v. Metabolism and Nufrition. Department of Medicine,Northwestern Universit.v Medical School, Chicago. Supported b_vGrant Nos. AM27047, AMO7169. KRO5370. and RR48 from the National Instifutes of Health. Address reprinr requesrs to G. Bnumann. MI). 303 Easr Chicago ilve, Chicago, IL 6061 I c I986 b_vGrune & Stratton. Im,. 0026 -0495/86/3509-00l8%03.00/0

883

884

STOLAR AND SAUMANN

Glucose was measured in unextracted plasma samples by a glucose oxidase method, using a Beckman glucose analyzer (Beckman Instruments, Palo Alto, Calif). The intra-assay coefficients of variation (CV) for hGH were 5.2% in the range above 2.5 ng/mL, and 8.6% in IERIA at an average level of 187 pg/mL. Interassay CV for hGH were 12.8% in the high range and 14.9% in low range by IERIA, and 2% for glucose.

Statistical Analysis Mean overall GH levels in a given subject were determined by averaging all values from that subject. Mean basal GH levels were similarly determined by averaging all values occurring below a certain level arbitrarily designated as the upper limit of “basal” (ie, 5, 2, or 1 ng/mL). Comparisons between the sexes were made by analysis of variance. Potential differences between the sexes were also assessed by determining the frequency of basal values (ie, those below a certain limit) and analysis by contingency table and x2 with correction for continuity. The presence of a secretory pulse was defined according to the method of Santen and Bardin,’ adapted for the variability of our assay. Thus, an increment of 30% (or 2 x the interassay CV) of a value over the preceding one was considered to exceed assay noise and represent a true spike. The appropriateness of this analysis was

Materials Dr U.J. Lewis (La Jolla, Calif) kindly provided hGH (lot no. 306-13-3), which was used as an RIA standard and for radioiodination. A lactoperoxidase method was used to prepare ‘2SI-hGH (specific activity, 73 &i/rg) and “‘I-hGH (specific activity, 37 to 62 pCi/pg)! Commercially available materials were as previously listed.’

IC 822” 5

2

I 0.5

02

d

r

24y

4:,1 I

IO d23y 5

2

I g

05

0.2

IO 825~ 5

0.5

2

0.2

I 0.5

0.2

0.1

005

0.02

CLOCK

TIME

Fig 1. Plasma hGH and glucose concentrations in six normal men. Note the logarithmic ordinate for hGH.

BASAL GROWTH HORMONE SECRETION

verified by visual inspection of the profiles. The limited time of observation, as well as the wide variability in pulse amplitude, pulse frequency, pulse shape, and baseline precluded the use of sophisticated time series analysis techniques. The degree of relation between GH and glucose levels, and vice versa, was assessed by correlation analysis of the two coincident variables, GH and corresponding glucose 15 and 30 minutes later, and glucose and corresponding GH 15 and 30 minutes later. The rationale for the time shifts in the correlation analyses was that a GH spike may be followed by a fall in glucose, and that a change in

different pattern. Despite expected basal conditions of fasting and relative rest, wide fluctuations in hGH levels occurred in all subjects. Levels of hGH were clearly in the readable range by IERIA in all plasma samples, the minimum detectable level being approximately 25 pg/mL. Many of the changes in plasma hGH levels apparent in this study would not have been identified by conventional RIA. During the limited period of observation, imposed by the need for large blood samples at each time point, there was no evidence for periodicity of hGH secretion. Rather, secretory bursts, periods of moderate secretion, and episodes of total secretory inactivity were interspersed with each other seemingly at random. The latter were suggested by several instances in which plasma hGH levels declined with a half-life of about 20 minutes, which approximates the known

glucose may be followed by a change in GH secretion (see “Discussion”). RESULTS

The time-related changes of plasma hGH concentration are depicted in Figs 1 and 2. Each individual showed a

100

100

? 22Y

50

9

50

024~

t .

20

20

IO

5

2

I

0.5

0.2

= E \ tCD

IO

5

c

I o-

5

;

Fig 2. Plasma hGH and glucose concentrations in six normal women of reproductive age. Note the logarithmic ordinate for hGH.

2

9

IO

II

I2

9

CLOCK TIME

0

12

886

STOLAR AND BAUMANN Table 1. Mean Overall and Basal Plasma hGH Levels in Six Normal Adult Women and Six Normal Adult Men Difference* (men Y women)

P Value for

Women All observations

Men

6.58

k 1.18

values below 5 ng/mL

2.28

+ 0.21

values below 2 ng/mL

0.993

+ 0.124

0.645

k 0.081

to.05

values below 1 ng/mL

0.486

f 0.050

0.331

k 0.041

CO.05

3.54

k 0.85

10.05

Basal hGH 1.20 k 0.15

-Co.005

Values are given as ng/mL (mean +- SEM). Because of the uncertain definition of what constitutes basal levels, the averages for three arbitrarily defined basal ranges are given.

l8ased on analysis of variance.

rate of hGH from the circulation6~’ In the basal state (defined as hGH t2 ng/mL), 72% of the increases in hGH exceeded assay noise, indicating true secretory events. The relative incidence of such events with respect to the duration of the basal state was similar in men and women (ie, 15/46 data points [33%] v 8/30 data points [27%], respectively). When the basal state was defined as hGH tl ng/mL, similar relationships held. In that case, 71% of hGH rises qualified for true increments, and 12/38 data points (32%) represented pulses in men, v 3/l 5 (20%) in women. A definition of the basal state as t5 ng/mL again yielded similar percentages. The differences in pulse frequency between men and women were not significant in any case (P > 0.5). The average overall hGH concentration was significantly higher in women than in men (Table 1). Likewise, the mean basal hGH concentration was significantly higher in women regardless of whether basal was defined as ~5 ng/mL, ~2 ng/mL, or ~1 ng/mL (Table 1). The tendency for males to have lower hGH levels was also reflected in the greater frequency of low values encountered in men (Table 2). Plasma glucose remained within the normal fasting range in all subjects throughout the study period. Scattergram analysis of the correlation between hGH and glucose indicated that no convincing relationship existed between hGH and glucose, or vice versa. This was true for coincident values and for values temporally shifted by 15 and 30 minutes.

disappearance

DISCUSSION

The present study extends our previous observations3 to examine in detail the secretory pattern of hGH during periods of relative rest, commonly called the basal state. The use of an ultrasensitive assay, combined with a rapid samTable 2. Frequency of Basal Plasma hGH Levels in Six Normal Adult Women and Six Normal Adult Men

Men

W0men

P Valuefor Difference* (men v women)

Basal hGH values below 5 ng/mL

51 (66%)

65 (84%)

<0.025

25 (32%)

50 (64%)


14 (18%)

36 (46%)


values below 2 ng/mL values below 1 ng/mL

Values are given as the number are the percentages *Based

on x2.

of total

of observations.

observations

in each

Values

in parentheses

sex (n = 78).

pling sequence, documents that basal hGH concentrations are highly variable, can change rapidly, and that hGH levels may fall to values much lower than previously appreciated. All hGH levels in the present study were clearly detectable and in the range judged to yield accurate results (ie, >25 pg/mL3). (The slightly greater assay sensitivity than that previously reported3 resulted from judicious selection of sample volumes for assay.) Our data indicate that the pulsatile nature of pituitary secretion and its apparent randomicity persist in the basal state. Together with our previous study, conducted over a 24-hour period with twohour samples,3 the present data permit a composite sketch of the circadian secretory pattern of GH, which for the first time includes the basal periods. The data are still by necessity limited because of the large amounts of blood required for assay. We restricted the blood loss in our study to a maximum of 300 mL in order to minimize stress from hypovolemia. Therefore, the observation period was insufficient to draw strong inferences about long-term periodicity of hGH secretion. We deliberately chose ambulatory subjects rather than subjects who had remained at bed rest before study because the ambulatory state is more closely representative of physiologic conditions. Morning hGH levels after an overnight fast are considered to be relatively quiescent.* However, despite a one-hour rest period before initiation of the study, five subjects had hGH levels above 5 ng/mL at the start of the study. All but one of these subjects displayed cessation of secretion followed by a return to the basal state. The women had significantly higher overall mean hGH levels than the men. This is in agreement with several,‘-” but not all”-‘3 previous reports. This sex difference, when observed, has been attributed to the effect of estrogens on hGH secretion. It is well documented that the plasma hGH response to provocative stimuli is higher in women, that the response is greatest during the middle of the menstrual cycle, and that in males the response can be enhanced by administering estrooccurring hGH gens.i4,” Reported data on spontaneously levels are conflicting, although the majority of studies favor the view that overall levels tend to be higher in females.8-‘3 With regard to basal hGH levels, it is presently not clear whether a sex difference exists. One study dealing with this question concluded that basal levels are similar in men and women.’ However, because of the aforementioned assay limitations in the basal range, no reliable data exist to clarify this point. Our findings clearly show that basal hGH levels are significantly higher in females regardless of how basal is defined. A similar sex difference in basal GH levels has been

887

BASAL GROWTH HORMONE SECRETION

reported for the rat.i6 Thus, it appears that women have higher plasma hGH concentrations at all levels of pituitary activity, provided that a sufficiently sensitive assay is employed. Since metabolic clearance of hGH is largely independent of its plasma concentration,“V’2 the fluctuations in plasma levels are primarily a function of pituitary secretion. The secretion of hGH by somatotrophs is believed to be under dual control by GHRH and somatostatin, as well as modulated by other factors, such as somatomedins.” Present knowledge recognizes hGH secretion as episodic, with major surges occurring during sleep, at random times, and postprandially.‘,2 In normal adults, major secretory peaks occur infrequently, with hGH levels being basal during the greater part of the day.‘,’ The boundary between basal and stimulated hGH levels, although conventionally placed at 5 ng/ mL, is ill-defined. Our data indicate that in the basal state, periods of moderately active secretion alternate with periods of cessation of secretion. Thus, the episodic nature of hGH secretion appears to persist at a reduced amplitude in the basal state. It follows from these considerations that the distinction between basal and stimulated hGH levels may be an artificial one largely derived from the insensitivity of RIAs. Our data suggest that plasma hGH levels should be viewed rather as a continuum of values ranging over several orders of magnitude, with very high levels occurring only infrequently. A similar concept of widely fluctuating GH levels is well established for the rat.‘* Since glucose serves as an important mediator of GH release, we examined the possibility that plasma glucose changes play a role in the oscillations of GH secretion during basal periods. However, we found no relationship between changes in glucose and those in basal hGH levels. It appears, therefore, that the minor fluctuations of blood glucose observed in the fasted state are insufficient to substantially influence hGH secretion. A similar conclusion has been reached for the postabsorptive state by others.19 GH is also an important modulator of glucose metabolism. Both anti-insulin activity and properties facilitating glucose utilization (so-called insulinlike activity) are inherent in GH.2a-23 While the physiologic significance of GH as an insulin-antagonist is well accepted, the early insulin-like action of GH has long been considered a pharmacologic phenomenon. This effect could only be demonstrated with

either very large doses of GH or with more physiologic doses in a preexistent GH-deficient state.“.24 On the other hand, small increases in glucose clearance have recently been demonstrated in normal man during the first 1% hours of an infusion with physiologic hGH amounts.21 In view of these facts, we postulated that a possible insulinlike effect of endogenous GH would be most clearly observed (1) after a period of relative GH deficiency (ie, low preexistent plasma level) and (2) before counterregulatory events could take effect. Our study provided an opportunity to examine this hypothesis as it documented both relative GH deficiency and relative GH repletion, and rapid transitions between these two states. Despite several instances of abrupt hGH increases IO-fold to lOO-fold over baseline values, we observed no subsequent changes in plasma glucose. Thus, the insulinlike effect of endogenous hGH appears to be of insufficient magnitude to acutely affect glycemia in vivo. A fundamental issue is whether basal levels of hGH are biologically relevant and whether their pusatile nature is necessary for normal physiology. In this context it is also important to recognize that in the basal state, plasma hGH consists in part of immunoreactive hGH fragments.25 The threshold value for hGH action in man is not well defined. Studies of certain metabolic activities of GH suggest threshold levels of 5 to 10 ng/mL.26 The majority of patients with active acromegaly also have hGH levels above that range. On the other hand, occasionally acromegalic patients show clinical manifestations of GH excess and elevated somatomedin-C concentrations when plasma hGH levels are consistently below 5 ng/mL2’ (and Baumann G, unpublished observations). In vitro studies have demonstrated that an hGH concentration of 0.5 ng/mL induces down regulation of GH receptors on cultured lymphocytes.28.29 According to the present study, secretory spikes of this magnitude are relatively common during basal periods. Thus, the question whether basal periods contribute to overall growth cannot be unequivocally answered at present. Threshold studies for the various metabolic activities as well as the long-term actions of GH with continuous and pulsatile hGH administration will be required to address this issue. ACKNOWLEDGMENT We thank Dr Mark Molitch for a critical review of the manuscript and Connie Mora for assistance in its preparation.

REFERENCES JW, Roffwarg HP. Boyar RM, et al: Age-related changes in the twenty-four hour spontaneous secretion of growth hormone. J Clin Endocrinol Metab 3.5:665-670, 1972 1. Finkelstein

2. Parker DC, Rossman LG. Kripke DF, et al: Rhythmicities human growth hormone concentrations in plasma, in Krieger (ed): Endocrine Rhythms, New York, Raven, 1979, pp 143-173

in DT

3. Drobny EC, Amburn K, Baumann G: Circadian variation of basal growth hormone in man. J Clin Endocrinol Metab 57524528. 1983 4. Rogol AD, Chrambach A: Radioiodinated human and amniotic fluid prolactins with preserved molecular Endocrinology 97:406-417. 1975 5. Santen

RJ,

Bardin

CW:

Episodic

luteinizing

pituitary integrity. hormone

secretion in man. Pulse analysis, clinical interpretation, physiologic mechanisms. J Clin Invest S2:2617-2628, 1973 6. Parker ML, Utiger RD, Daughaday WH: Studies on human growth hormone. II. The physiological disposition and metabolic fate of human growth hormone in man. J Clin Invest 41:262-268, 1962 7. Refetoff S, Sijnksen PH: Disappearance rate of endogenous and exogenous human growth hormone in man. J Clin Endocrinol Metab 30:386-392, 1970 8. Unger RH, Eisentraut AM, Madison LL, et al: Fasting levels of growth hormone in men and women. Nature 205:804-805, 1965 9. Frantz AG, Rabkin MT: Effects of estrogen and sex difference on secretion of human growth hormone. J Clin Endocrinoi Metab 25:147&1480, 1965

888

10. Thompson RG, Rodriguez A, Kowarski A, et al: Growth hormone: Metabolic clearance rates, integrated concentrations, and production rates in normal adults and the effect of prednisone. J Clin Invest 51:3193-3199, 1972 11. Taylor AL, Finster JL, Mintz DH: Metabolic clearance and production rates of human growth hormone. J Clin Invest 48:23492358, 1969 12. MacGillivray MH, Frohman LA, Doe J: Metabolic clearance and production rates of human growth hormone in subjects with normal and abnormal growth. J Clin Endocrinol Metab 30:632-638, 1970 13. Miller JD, Tannenbaum GS, Colle E, et al: Daytime pulsatile growth hormone secretion during childhood and adolescence. J Clin Endocrinol Metab 55:989-994, 1982 14. Merimee TJ, Fineberg SE: Studies of the sex based variation of human growth hormone secretion. J Clin Endocrinol Metab 33:896-902,197l 15. Merimee TJ, Rabinowitz D, Fineberg SE: Arginine initiated release of human growth hormone: Factors modifying the response in normal man. N Engl J Med 280:1434-1438,1969 16. Eden S: Age- and sex-related differences in episodic growth hormone secretion in the rat. Endocrinology 105:555-560, 1979 17. Berelowitz M, Szabo M, Frohman LA, et al: Somatomedin-C mediates growth hormone negative feedback by effects on both the hypothalamus and the pituitary. Science 2 12:1279-l 28 1, 1981 18. Tannenbaum GS, Martin JB: Evidence for an endogenous ultradian rhythm governing growth hormone secretion in the rat. Endocrinology 98:562-570, 1976 19. Tse TF, Clutter WE, Shah SD, et al: Neuroendocrine responses to glucose ingestion in man. Specificity, temporal relationships, and quantitative aspects. J Clin Invest 72:27&277, 1983 20. Merimee TJ, Rabin D: A survey of growth hormone secretion and action. Metabolism 22:1235-1251, 1973

STDLAR AND BAUMANN

21. MacGorman LR, Rizza RA, Gerich JE: Physiological concentrations of growth hormone exert insulin-like and insulin antagonistic effects on both hepatic and extrahepatic tissues in man. J Clin Endocrinol Metab 53:556-559, 1981 22. Bratusch-Marrain PR, Smith D, DeFronzo RA: The effect of growth hormone on glucose metabolism and insulin secretion in man. J Clin Endocrinol Metab 55:973-982, 1982 23. Sherwin RS, Schulman GA, Hendler R, et al: Effects of growth hormone on oral glucose tolerance and circulating metabolic fuels in man. Diabetologia 24:155-161, 1983 24. Goodman HM: Growth hormone and the metabolism of carbohydrate and lipid in adipose tissue. Ann NY Acad Sci 148:419-440,1968 25. Baumann, G, Stolar MW, Amburn K: Molecular forms of circulating growth hormone during spontaneous secretory episodes and in the basal state. J Clin Endocrinol Metab 60:1216-1220, 1985 26. Goodman HM: Growth hormone and the metabolism of carbohydrate and lipid in adipose tissue. Ann NY Acad Sci 148:419-440,1968 27. Moses AC, Molitch ME, Sawin CT, et al: Bromocriptine therapy in acromegaly: Use in patients resistant to conventional therapy and effect on serum somatomedin-C. J Clin Endocrinol Metab 53:752-758, 1981 28. Eastman RC, Lesniak MA, Roth J, et al: Regulation of receptor by homologous hormone enhances sensitivity and broadens scope of radio-receptor assay for human growth hormone. J Clin Endocrinol Metab 49:262-268, 1979 29. Rosenfeld RG, Hintz RL: Modulation of homologous receptor concentrations: A sensitive radioassay for human growth hormone in acromegalic, newborn and stimulated plasma. J Clin Endocrinol Metab 50:62-69, 1980