Clinical Endocrinology (1991) 35, 353-360

The interaction of growth hormone releasing hormone and somatostatin in the generation of a GH pulse in man P. C. Hindmarsh, C. E. Brain, 1. C. A. F. Robinson*, D. R. Matthewst and C. G. D. Brook Endocrine Unit, The Middlesex Hospital, London, *The Division of Neurophysiology and Neuropharrnacology, National Institute for Medical Research, London NW7 1AA and t T h e Diabetes Research Laboratories, Radcliffe Infirmary, Woodstock Road, Oxford OX2 6HE, UK (Received 15 March 1991; returned for revision 21 May 1997; finally revised 3 June 1991; accepted 11 June 1991)

Summary OBJECTIVE To study the regulation of the growth hormone (OH) response to growth hormone releasing hormone (GHRH) in the presence or absence of somatostatin pretreatment. DESIGN Seven healthy adult male volunteers of normal height and weight and aged between 19 and 29 years underwent four separate studies (each containing three cycles In one day) in random order. The studies were separated from each other by at least a week. On day 1, three consecutive cycles (between 0800 and 2000 hours) consisted each of a saline infusion for 3 hours which was stopped prior to a bolus Injection of saline and followed by 60 minutes of more intensive blood sampling. On day 2, the bolus injections were of GHRH given after saline infusion. On days 3 and 4 somatostatin infusions were administered instead of saline over the 3-hour periods followed by bolus Injections of saline or GHRH respectively. In all studies, samples were collected for the measurement of serum OH concentration at 15-minute intervals from time 0 to 180 minutes and then at 5-minute Intervals for a further 60 minutes, this cycle being repeated three times. MEASUREMENTS Serum GH concentrations were analysed by rerlal array averaging. RESULTS Prompt release of OH was observedin response to GHRHgiven against a saline background(day 2, cycle 1) (mean at 60 minutes 49.2f14.7 mU/I) but the responses observed during the second and third cycles were attenuated (mean at 60 minutes 17.2f 4.0mU/I; P= 0.025). GH release between somatostatin infusions (somatostatin

Correspondence: Professor C. G. D. Brook,Endocrine Unit, The Middlesex Hospital, Mortimer Street, London WIN 8AA, UK.

withdrawal; day 3) occurred twice as often as that observed during saline infusions (62% day 3: 29% day 1). The response, although qualitatively simllar to that induced by GHRH, was reduced In amplitude and the time of onset variable (5-45 minutes). Onday 4,the administrationof GHRH as a bolus Injection combined with somatostatin withdrawal led to consistent and repeatable OH responses (mean at 60 minutes, cycle 1, 39.7f 10.8 mull; cycles 2 and 3, 37.4f 9.4 mU/i) which were similar to those observed with GHRH alone (day 2, cycle 1) (mean 39.7f10.8mU/i) (P=NS). CONCLUSIONS These data suggest that endogenous somatostatin secretion is important in determining the ability qf the somatotroph to respond to repeated growth hormone releasing hormone stimulation and that for regular GH pulse generation a close Interplay between growth hormone releasing hormone and somatostatin is required.

In most species, growth hormone (GH) is secreted in a pulsatile fashion. The pattern of G H secretion is important for optimal body growth in humans and rodents (Hindmarsh et al., 1987; Albertsson-Wikland & Rosberg, 1988; Jansson et al., 1982; Clark etal., 1985; Clark & Robinson, 1985a)and is important in the rodent for regulation of hepatic enzyme function (Mode et al., 1982; Jeffrey et al., 1990). GH release from the anterior pituitary is under the control of at least two peptides, somatostatin (SS) and growth hormone releasing hormone (GHRH), as well as a number of other factors, such as nutrients, thyroid hormones and steroids. In the rodent it is likely that G H secretory patterns reflect a balance between SS and GHRH release (Tannenbaum & Ling, 1984; Plotsky & Vale, 1985). Abrupt withdrawal of SS in the rat invariably leads to rebound secretion of G H and this effect is more pronounced in uiuo than in uitro (Kraicer et al., 1986; Weiss et al., 1987). The rebound GH secretion that follows SS withdrawal in viuo appears to be mediated by GHRH (Clark et al., 1988; Miki et al., 1988; Sugihara et al., 1989). Further, pretreatment of rats with SS enhances GH responsiveness to GHRH challenge (Clayton & Bailey, 1987; Stachura et al., 1988). SS, therefore, may play an important role not only in trough generation but also in the maintenance of GH surges by optimizing somatotroph responses to successive pulses of GHRH (Sato et at., 1990). Indirect evidence for a similar episodic secretion of SS in man comes from the observation 353

354

Clinical Endocrinology (1991) 35

P. C . Hindmarsh et a/.

that the serum GH concentration response to GHRH is dependent on the GH secretory status at the time of GHRH administration (Devesa et al., 1989; Suri et al., 1990). The role of SS in the regulation of pulsatile GH release in man is far from clear. Repetitive administration of GHRH over short periods (1-3 hours) leads to a diminished GH response in normal subjects (Losa et al., 1985; Suri et al., 1991) but continuous administration of GHRH for periods of 7 and 14 days augments GH pulsatility in both adults and short children (Vance et al., 1985; Rochiccioli et al., 1986; Hulse et al., 1986; Brain et al., 1988; Ross et al., 1990). Although transient pituitary desensitization by chronic GHRH stimulation has been observed in rodents in uitro (Bilezikjian et al., 1986) and in vivo (Arsenijevic et af., 1987), this does not appear to be a major problem in man as shown by the growth and G H secretion of children receiving continuous GHRH treatment (Brain el al., 1990), the manifestation of acromegaly by patients with GHRH secreting tumours (Thorner et al., 1984) and the augmentation by continuous GHRH infusion of the already elevated G H concentrations in patients with acromegaly (Gelato et al., 1990). It has been suggested that one reason for the efficacy of continuous GHRH to promote GH secretion and growth lies in the preservation of SS pulsatility. There is in-vitro evidence that SS pulsatility is one of the factors that prevents desensitization (Simard et al., 1987; Clayton & Bailey, 1987) but in-vivo evidence is lacking. We have, therefore, investigated the interrelationship of GHRH and SS in the generation of GH secretory episodes paying particular attention to the role of intermittent SS exposure in regularizing the GH response to GHRH. Methods

Subjects

Seven adult healthy male volunteers of normal height and weight aged between 19 and 29 years took part in the study. The protocol was approved by the Ethics Committee of the Middlesex Hospital and informed consent was obtained from each individual prior to commencement of the studies.

Day I Saline Day 2 GHRH

Saline GHRH

Saline GHRH

Day 3 Saline Day 4 GHRH

Saline GHRH

Saline GHRH

I Clock time

Fig. 1 Protocol design. Infusion of saline (days 1 and 2) or SS( I -

14) (days 3 and 4) took place between 0800 and 1100 h, I200 and 1500 h, and 1600 and 1900 h. Bolus injections of either saline (days 1 and 3) or GHRH(I-Z9)NH2(days 2 and 4) took place at 1100, 1500 and 1900 h. Blood samples for estimation of serum GH concentration were drawn at 15-minute intervals between 0800 and 1100 h, 1200 and 1500 h, and 1600 and 1900 h and at 5minute intervals between 1100 and 1200 h, 1500 and 1600 h, and 1900 and 2000 h.

concentrations. Immediately after this sample was drawn, an infusion of physiological normal saline was commenced through a Graseby pump over 3 hours (1 ml/h) on days 1 and 2. On days 3 and 4, somatostatin (1-14) (Ferring Pharmaceuticals, Malmo, Sweden) was added to the saline infused and delivered at a rate of 50-100 pg/mz body surface area/h. Blood samples were then drawn at 15-minuteintervals for the next 3 hours. At 1100 h the infusion was stopped: on days 1 and 3 an intravenous bolus injection of physiological normal saline (1 ml) was administered and on days 2 and 4 an intravenous bolus injection of GHRH( 1-29)NH2 (KabiPharmacia, Stockholm, Sweden) was given in a dose of 100 pg. Samples were then drawn at 5-minute intervals for the next 60 minutes and the cycles repeated. The studies were performed in random order and separated from each other by at least 1 week. All samples were spun, separated and stored at -20°C prior to analysis in the same assay. Assays

Protocol

Following an overnight fast, an indwelling intravenous cannula was inserted in the forearm at 0700 h for infusion and drug administration purposes and a second indwelling intravenous cannula inserted in the other forearm for blood sampling. A repetitive balanced design was used and is illustrated in Fig. 1. At 0800 h a blood sample was drawn for measurement of serum GH, insulin and blood glucose

Serum GH concentrations were measured using an immunoradiometric assay (Hybritech Tandem-R hGH Kit (Hybritech, Europe). The interassay coefficient of variation was 9.6% and the intra-assay coefficients of variation were 6.3, 3.5, 3.2 and 3.0% at serum GH concentrations of 4-9, 30.5, 26.3 and 51.3 mU/1 respectively. The sensitivity of the assay was 0.5 mu/]. Serum insulin concentration was measured by a double

Clinical Endocrinology (1991)35

GHRH and somatostatin

t n

2ot \

l o t

OL,

355

I

,

I

I

I

I

,

I

I

I

,

I

1

E

Fig. 2 The role of GHRH and somatostatin in the generation of GH pulses. Representative serum GH concentration profile from a 22-year-old male. a, Day I. 3-hourly saline infusions coupled with bolus injections of saline. b, Day 2, 3-hourly saline infusions coupled with bolus injections of 100 pg of GHRH(I-29)NHz. c, Day 3, 3-hourly SS(I14) infusions coupled with bolus injections of SS(1-14). d, Day 4, 3-hourly SS(1-14) infusions coupled with bolus injections of 100 pg GHRH(I-29)NH2. Arrows indicate times of bolus injections. Bars indicate duration of infusions.

r I

6

1

1000

antibody radioimmunoassay procedure (Morgan & Lazarow, 1963). Intra and inter-assay coefficients of variation were 3.9 and 9.6% respectively at concentrations between 20 and 25 mU/l. Blood glucose concentrations were measured using the glucose oxidase method (YSI 23 M glucose analyser: Yellow Springs, Ohio). The intra-assay coefficients of variation were 0.8 and 0.9%at blood glucose concentrations of 4.0 and 8.0 mmol/l respectively. Statistics The serum GH, insulin and blood glucose concentrations measured during the infusion off/bolus injection cycles which lasted 60 minutes were studied using the technique of standard array averaging (Dawson, 1954; Lang ef al., 1982). Individual serum GH concentration cycles were defined from the point of administration of the intravenous bolus/ withdrawal ofinfusion. Serum GH concentration cycleswere defined and used as the event to average the simultaneous changes in serum insulin and blood glucose concentrations. Thus, other factors affecting blood glucose and serum insulin concentrations occurring randomly with respect to the GH cycles would be averaged out. This technique allows small

I500

2000 h

1

1

1000

1

1

1

1

1

1500

1

1

1

1

1

2000 h

Clock time

but consistent responses to stimulation to be detected amongst large spontaneous activity. In the case of day 3 studies the cycles were also defined from the time point of each individual’s GH response because of the variable timing of its onset relative to SS(1-14) withdrawal. Three cycles for each study day from the seven individuals were analysed. Cycle correlation was then performed to determine the time relationship between serum GH concentrations and changes in blood glucose and serum insulin concentration. The data are presented as cycle 1 (i.e. stimulus against a basal background) and the combinations of cycles 2 and 3 (i.e. stimulus against a background of previous stimulation) for each of the study days. The data are plotted as the mean f SEM. Where the error bars do not overlap, P was less than 0.025. Student’s t-test was used to compare values at different time points. Results

Figure 2 shows the changes in serum GH concentrations during the three cycles on the 4 days of the study in one individual. The data for all seven individuals are shown in Fig. 3 in which cycle 1 data alone and the combined data from cycles 2 and 3 are compared.

356

Clinical Endocrinology (1991) 35

P. C. Hindmarsh el a / .

10 r-

-

d

0

I60

40 0

0

u 10

20

30

40

50

60

Time (min)

Fig. 3 a, Serum GH; b, insulin; and c, blood glucose concentrations during 0 , cycle 1 and 0,cycles 2 and 3 in seven adult males. a l x l , Day 1, withdrawal of saline infusion and bolus injection of saline. a2x2, Day 2, withdrawal of saline infusion and bolus injection of GHRH(I-29)NH2. a3-c3, Day 3, withdrawal of SS(1-14) infusion and bolus injection of saline. a4-24, Day 4, withdrawal of SS(1-14) infusion and bolus injection of GHRH(I-29)NH2. Data shown as meanf SEM.

Day 1 . Intermittent saline infusion with saline bolus injections

On the control day there were no significant differences between the serum GH concentrations measured during cycle 1 and cycles 2 and 3. For example, the mean serum GH concentrations at 30 and 60 minutes during cycle 1 were 3.5 f 2.6 and 3.9 1a9 mU/1 respectively and during cycles 2 and 3 , l . l k0.2 and 1.4f0.6mU/1 respectively. Spontaneous GH pulses were detected in six of the 21 60-minute study periods (29%). Blood glucose concentrations did not change throughout and serum insulin concentrations remained

within a narrow range (cycle 1,9-7-14.3 mU/I: cycles 2 and 3, 18.9-22.9 mU/1). The higher values during cycles 2 and 3 probably reflect the light snacks taken at lunch and tea-time.

Day 2. Intermittent saline infusion with GURU (I-29)NH2 bolus injections

Discontinuation of the saline infusion and bolus injection of GHRH( 1-29)NH2 produced an immediate (within 5 minutes) increase in serum GH concentrations in all three cycles (21/21, 100%). However, the first GH response was

Clinical Endocrinology (1991) 35

significantly greater than that to the second and third injections (at 60 minutes: cycle 1, 49.2 f 14.7 mU/I, cycles 2 and 3 17.2f4.0 mu/]; t=2.095, P=O.O25). Blood glucose and serum insulin concentrations were not significantly different from those observed on day 1 and similar increases in cycles 2 and 3 compared to cycle 1 were again observed. Day 3. Intermittent somatostatin (1-14) infusion with saline bolus injections

The time from switching off the SS(1-14) infusion to the rise of serum GH concentrations was variable between individuals (median 18 minutes, range 5-45). The serial array averaged data were, therefore, analysed in two ways, first as group data with a common event time point when the SS(114) infusion was switched off and, secondly,as individualized data centred on the actual time at which serum GH concentrations began to rise. The first method revealed that cessation of the first period of SS(1-14) infusion was accompanied by a slight increase in serum GH concentration but this was not significantly different from the control day (at 60 minutes: day 1 cycle I , 3.9f 1.9 mu/]; day 3 cycle 1, 4.7 f3 4 mu/]). However, GH released in cycles 2 and 3 on day 3 were significantly greater than that in cycle 1 (at 60 minutes: cycles 2 and 3, 15.8k4.3 mU/I, cycle 1, 4.7f3.8 mu/]; t=1.923, f < 0.05) and also significantly greater than values observed during cycles 2 and 3 on day 1 (at 60 minutes: day 3 cycles 2 and 3, 15.8f4.3 mU/I, day 1 cycles 2 and 3, 1.4k0.6 mu/]; t = 3.308, P=O.OOS). The serum G H concentration response to SS(1-14) withdrawal still occurred in 13 of the 21 studies (62%), approximately twice as frequently as that on day I . Analysis of the data individualized to the commencement of the rise of GH showed that the amplitude of this optimized cycle (cycles 2 and 3) was 15.9f4.6 mU/I compared with the response to GHRH (1-29)NHzduring cycles 2 and 3 on day 2 (mean 256f 7.8 mu/]; t = 1.07, P=NS). The shape of the serum GH response was similar. An increased blood glucose concentration was evident at the commencement of each cycle following cessation of the SS(1-14) infusion (cycle 1, 6.4f 1.0 mmol/l, cycles 2 and 3, 6.9 f0.6 mmol/l). Serum insulin concentrations were not significantly different from day I but rose significantly after SS(1-14) withdrawal, reaching a peak 15-20 minutes later. Because of this increase in insulin, blood glucose concentrations fell significantly below the initial value only 30-40 minutes into the cycle and below the values recorded on day 1, 40-45 minutes into the cycle after the peak of GH had occurred (Fig. 3). The blood glucose at the end ofcycle 1 was 3.4_+0.3mrnol/landattheendofcycles2and3was3.6f0.3 mmol/l. One individual reported symptomatic hypogly-

GHRH and somatostatin

357

caemia at the end of cycle 2 (blood glucose 2.1 mmol/l) and another at the end ofcycle 3 (blood glucose 2.0 mmol/l). Both were given glucose drinks.

Day 4. Intermittent SS(1-14) infusion with GHRH(1-29)NH2 bolus infections

GHRH( 1-29)NHl administration coupled with SS(1-14) withdrawal resulted in a G H response in all individuals during cycle I . The response was identical to that observed during day 2 cycle 1 (at 30 minutes: day 4,42.21t_14.8 mU/l, day 2, 42.7f 13.2 mu/]; at 60 minutes: day 4, 39.7f10.8 mu/], day 2, 49.2f 14.7 mu/]). However, the serum GH concentrations measured during cycles 2 and 3 on day 4 were significantly greater than during cycles 2 and 3 on day 2 (at 60 minutes: day 4, 37.4f9.4 mu/], day 2, 17.2f4.0 mu/]; r=2.00, P

The interaction of growth hormone releasing hormone and somatostatin in the generation of a GH pulse in man.

To study the regulation of the growth hormone (GH) response to growth hormone releasing hormone (GHRH) in the presence or absence of somatostatin pret...
741KB Sizes 0 Downloads 0 Views