Comparison with GH Deficiency

M Maccario, MR Valetto, P Savio, G Aimaretti, C Baffoni, M Procopio, S Grottoli, SE Oleandri, E Arvat and E Ghigo

Division of Endocrinology, Department of Internal Medicine, University of Turin, Italy

OBJECTIVE: To evaluate the maximal secretory capacity of somatotrope cells in obesity and to compare it with that in hypopituitaric patients with GH de®ciency. DESIGN: Stimulation with GHRH. (1 mg/kg iv) combined with arginine (ARG, 0.5 g/kg iv), which strongly potentiates the GH response to the neurohormone, likely inhibiting hypothalamic somatostatin. The reproducibility of the GH response to GHRH ARG was evaluated in a second session. SUBJECTS: Forty-®ve patients with simple obesity (OB 11 male and 34 female, age 40.5 Æ 1.8 y, BMI 38.8 Æ 1.1 kg/m2), 49 patients with hypopituitarism (GHD, 23 male and 26 female, 43.6 Æ 2.4 y, 24.7 Æ 0.7 kg/m2) and 44 normal young volunteers (NS, 25 male and 19 female, 33.8 Æ 1.0 y, 21.6 Æ 0.3 kg/m2) were studied. MEASUREMENTS: GH levels were assayed by IRMA method, basally at 7 60 and 0 min, and than every 15 min up to 120 min. Basal IGF±I levels were assayed by RIA method, after acid-ethanol extraction. RESULTS: IGF±I levels in OB were lower (P < 0.005) than those in NS but higher (P < 0.005) than those in GHD. Mean peak GH response to GHRH ARG in OB was clearly lower than that in NS (P < 0.005) and higher (P < 0.005) than that in GHD. Sixty±percent OB and 100% GHD showed peak GH responses lower than the minimum normal limit in NS (16.5 mg/l) while 4% OB and only 53% GHD with GH responses lower than 3 mg/l, the limit under which GH replacement therapy of severe de®ciency is allowed. Good intraindividual reproducibility of the GH response to GHRH arginine test was present in all groups (OB: r 0.78, P < 0.0001; GHD: r 0.57, P < 0.003; NS: r 0.74, P < 0.0001;. CONCLUSIONS: The maximal secretory capacity of somatotrope cells is clearly less than normal in the obese but still more than is seen in GHD subjects. However, in about 50% of obese patients, the pituitary GH releasable pool overlaps with that of hypopituitaric patients with GH de®ciency. Thus, even when the maximal secretory capacity of somatotrope cells is evaluated by a potent and reproducible provocative tests such as GHRH arginine, overweight has to be taken in a great account as the cause of severely impaired GH response in patients with suspected GH de®ciency.

Keywords: GH; obesity; GH de®ciency; GHRH arginine

Introduction The extent of GH insuf®ciency in obesity is not clari®ed and a comparison between the impairment of somatotrope secretion in obese and in hypopituitaric It is well known that GH secretion is reduced in patients with organic GH de®ciency has never been obesity, in which a clear cut reduction of both spon- done. taneous and stimulated GH release has been demon- On the other hand, the diagnosis of GH de®ciency strated.1±5 Interestingly, GH insuf®ciency in obesity is in adulthood is still an open problem.25,26 In fact, in acquired and normal somatotrope function is restored this latter condition IGF-I and IGFBP-3 assays as well by weight loss.2,6,7 as the study of spontaneous GH secretion are not The pathogenesis of GH insuf®ciency in obesity is reliable diagnostic tools25±27 while insulin-induced still unclear. Hypothalamic somatostatinergic hyper- hypoglycemia has been proposed as the golden stan- activity and/or GHRH hypoactivity has been hypothe- dard test.26 More recently, testing with GHRH com- sized3,8±11 though a pituitary impairment due to bined with arginine, which likely acts via inhibition of metabolic alterations can not be ruled out.12±18 In hypothalamic somatostatin release and potentiates the fact, hyperinsulinism and elevated free fatty acid GH response to the neurohormone,28,29 has been levels connote obesity19,20 and these alterations may proposed as the best way to explore the maximal play a major role in impairing somatotrope func- secretory capacity of somatotrope cells and to distin- tion.14,17,18,21±24 guish between normal subjects and GH de®cient patients.30 At present, there is consensus about the fact that rhGH replacement therapy is allowed only Correspondence: Dr E Ghigo, Division of Endocrinology, Ospedale Molinette, C.so Dogliotti 14, 10126 Torino, Italy. for GHD patients presenting with GH peak response Received 16 April 1996, revised 5 June 1996; accepted 18 lower than 3 mg/l after provocative stimulation.31 September 1996 GH reserve in obesity M Maccario et al 28 Based on the foregoing, the aim of the present study ity of the assay was 0.1 mg/l. The inter- and intra- was to verify the extent of GH insuf®ciency in obesity assay coef®cients of variation were 10.1±15.7% and evaluating the maximal secretory capacity of somato- 7.6±15.5%, respectively. trope by testing with GHRH arginine; the results in The secretory responses are expressed as absolute obese patients were compared with those in patients values (mg/l). Results are expressed as mean Æ s.e.m. with organic GH de®ciency and in normal controls. and individual, absolute levels. Statistical analysis To strengthen the reliability of the results, the repro- of the data was carried out using non parametric ducibility of the GH response to GHRH arginine Kruskal±Wallis or Wilcoxon test. was also studied in all groups. Results Methods Basal GH levels in OB and GHD were similar (1.1 Æ 0.3 vs 0.4 Æ 0.1 mg/l) and lower (P < 0.0005) Forty-®ve patients with simple obesity (OB, 11 male than those in NS (2.3 Æ 1.7 mg/l). and 34 female, 40.5 Æ 1.8 y, BMI 38.8 Æ 1.1 kg/m2), IGF±I levels in OB (132.4 Æ 8.4 mg/l) were higher 49 patients with hypopituitarism (GHD, 23 male and (P < 0.005) than those in GHD (61.6 Æ 6.0 mg/l) and 26 female, 43.6 Æ 2.4 y, 24.7 Æ 0.7 kg/m2) and 44 lower (P < 0.005) than those in NS (174 Æ 5.7 mg/l). normal subjects (NS, 25 male and 19 female, IGF±I levels showed a weak negative relationship 33.8 Æ 1.0 y, 21.6 Æ 0.3 kg/m2) were studied. In GHD with BMI in OB (r 70.51, P < 0.0001), but not in 41 out of 49 patients had multiple anterior pituitary GHD and NS. de®ciency and were under appropriate replacement In NS, the mean peak GH response to ARG therapy (L±thyroxine, cortisone, testosterone or oes- GHRH was 64.5 Æ 5.4 mg/l. On the other hand, the GH tradiol, medroxyprogesterone). The remaining 8 GHD responses to this test in GHD are almost absent had isolated GH de®ciency demonstrated by classical (3.1 Æ 0.4 mg/l, P < 0.005 vs NS) (Figure 1). stimuli (namely, demonstrated in childhood and con- In OB, the GH response to ARG GHRH test ®rmed in adulthood) and the anterior pituitary func- (18.4 Æ 2.2 mg/l, P < 0.00005) than that in NS and tion other than GH secretion was normal. In OB, higher (P < 0.005) than that in GHD (Figure 1). The normal anterior pituitary function other than somato- GH response to ARG GHRH was not related to sex trope secretion, had been shown by normal serum or BMI in any of the groups while in OB it did not TSH and thyroid hormone, 24 h urinary cortisol and depend on regional fat distribution serum gonadal steroid levels. Normal menstrual The reproducibility of the GH response to bleeding or normal libido and potency had been ARG GHRH test was good in all groups both reported. The existence of other abnormalities was between subjects (GH peak: 1 vs 2 session in NS excluded. No patient had diabetes or was under 70.4 Æ 10.1 vs 76.2 Æ 10.7 mg/l; GHD: 2.3 Æ 0.5 vs pharmacological treatment. All subjects were eating 2.2 Æ 0.6 mg/l; OB: 21.3 Æ 5.7 vs 21.2 Æ 4.9 mg/l) and an unrestricted diet and their body weight had within±subjects (r 0.74, P < 0.0001; r 0.57, remained unchanged for 2 months before the study. P < 0.003 and r 0.78, P < 0.001, respectively) After informed consent, all subjects underwent a (Figures 2 and 3). test with GHRH (1 mg/kg as iv bolus at 0 min) Analyzing the individual GH peak responses to arginine (ARG, arginine hydrochloride, 0.5 g/kg ARG GHRH in the different groups of patients infused iv from 0 to 30 min). and assuming 16.5 mg/l as the minimum normal GH In 14 OB, 15 GHD, and 18 NS the test was repeated peak (3rd percentile of the GH response in normal in another session, at least three days apart. All tests subjects), as expected, all GHD (100%) had GH peaks began at 08.30 am after an overnight fasting and lower than this limit, according to their GH de®ciency 30 min after an indwelling catheter had been inserted (Figure 4). The peak GH response to ARG GHRH in a cubital vein kept patent by slow saline infusion. was frequently lower than the normal limit in OB (27/ Blood samples were taken basally, at 760 and 0 min 45, 60%), thus indicating their somatotrope insuf®- and then every 15 min from 0 to 120 min. All ciency (Fig. 4). samples from an individual subject were analyzed Assuming a value lower than 3 mg/l as the limit together. indicating severe somatotrope impairment, this was Serum GH levels were measured in duplicate by present in only 26 out of 49 (53%) GHD. Even 2 immunoradiometric assay (hGH±CTK, Sorin, Italy). OB (4%) showed GH peak lower than this limit The sensitivity of the assay was 0.1 mg/l. The inter- (Figure 4). and intra±assay coef®cients of variation were 4.9± 6.5% and 1.5±2.9% respectively. Plasma IGF±I levels were measured in duplicate by radioimmuno assay Discussion (Nichols Institute Diagnostics, San Juan, Capistrano, USA). To avoid interference by binding proteins, all The results of the present study demonstrate that the samples were treated with acid ethanol. The sensitiv- maximal secretory capacity of somatotrope cells is GH reserve in obesity M Maccario et al 29

Figure 1 Mean ( s.e.m.) peak and area under curve (AUC) GH responses to ARG GHRH in normal subjects, obese subjects and patients with organic GH de®ciency.

Figure 2 Between-subject variability in the GH response to ARG GHRH in normal subjects, patients with organic GH de®ciency and obese patients. clearly less than normal in obesity but still more than The pathogenesis of GH insuf®ciency in obesity is is seen in organic GH de®ciency. However, in about still unclear. A hypothalamic pathogenesis for reduced 50% of obese patients, the pituitary GH releasable somatotrope activity in obese patients has been pro- pool overlaps with that of hypopituitaric patients with posed GH de®ciency. Noteworthy, the GH response to A somatostatinergic hyperactivity has been sug- GHRH arginine test shows good intra-individual gested by some authors10 but it is not supported by reproducibility, both in obese and GHD patients, as other data.3,5,34 Also our present data showing that well as in normal subjects. GHRH combined with a substance likely acting via It is widely accepted that in obesity there is a clear inhibition of hypothalamic somatostatin, such as argi- reduction of both spontaneous and stimulated GH nine,28,29 fails to elicit a normal GH response in secretion1±5,11,32 while IGF-I levels have been found obesity, agree with other data38 against the existence high, normal or low.15,33±37 Our present data con®rm of an exaggerated somatostatinergic activity. that somatotrope function is reduced in this condition On the other hand, also the existence of GHRH and, as IGF-I levels were found low as well, indicate hypoactivity has been proposed in obesity;8,39,40 that a true reduction in the activity of GH/IGF-I axis is this, in turn, could cause reduced GH synthesis and, present. The fact that IGF-I levels in obese patients thus, explain the reduction in the maximal secre- were still higher than those in GHD may be due to the tory capacity of somatotrope cells and present hyperinsulinism that connotes obesity20 and it is data.8,41 known to positively in¯uence IGF-I synthesis.37 Other data, however, do not support the existence GH reserve in obesity M Maccario et al 30

Figure 3 Within-subject variability in the GH response to ARG GHRH in normal subjects, patients with organic GH de®ciency and obese patients.

of impaired GHRH activity5 while there is evidence indicating that hyperinsulinism and/or elevated free fatty acid levels, which are usually elevated in obe- sity15,19,20 could play a major role in causing hyposo- matotropinism acting via the hypothalamus or, more likely, at the pituitary level.12,13,17,18,21,23,40 Con®rm- ing the metabolic pathogenesis of GH de®ciency in obesity there is the well known evidence that spontaneous and stimulated GH secretion are restored by weight loss2,6 and that a clear inverse relationship exists between GH secretion and body mass.2 From a diagnostic point of view, it has to be considered that the diagnosis of GH de®ciency in adulthood is still a matter of debate though there is clear evidence for usefulness of rhGH replacement.25 The assay of IGF-I and IGFBP-3 as well as the evaluation of spontaneous GH secretion have been Figure 4 Individual GH peaks after ARG GHRH in obese found unreliable to distinguish between normal sub- patients (OB) and in patients with organic GH de®ciency (GHD). 25,30 Dotted line at 16.5 mg/l indicates lowest normal peak GH jects and GHD patients. Insulin-induced hypogly- response.30 Dotted line at 3.0 mg/l indicates the peak GH response cemia has been indicated as the golden standard test under which severe GHD is shown and rhGH replacement is by some authors25 but not by others.29,42,43 On the allowed.31 other hand, testing with GHRH arginine, which strongly potentiates the GH-releasing effect of the neurohormone allowing for the evaluation of the Noteworthy, in agreement with other studies with maximal secretory capacity of somatotrope cells,29 classical provocative stimuli,1,3±5 even when the max- has been found reliable in distinguishing between imal secretory capacity of somatotrope cells is eval- normal and GHD subjects.30 The results of the present uated by a potent and reproducible provocative test study con®rm that GHRH arginine is a potent sti- such as GHRH arginine, overweight has to be taken mulatory test of the maximal secretory capacity of in a great account as the cause of severely impaired somatotrope cells and, noteworthy, demonstrate that it GH response in patients with suspected GH de®- has good intraindividual reproducibility thus strength- ciency. In fact, our study demonstrates that the reduc- ening its reliability. The minimum normal limit of GH tion of the GH reserve in obesity may be so marked peak response to this test (16.5 mg/l) is higher than that 50% of obese patients with otherwise normal those arbitrarily assigned for other provocative stimuli pituitary function may have such an important impair- and allows to clearly show that patients with organic ment of the GH response to GHRH arginine that it GHD have subnormal GH response distinguishing overlaps with that seen in hypopituitaric patients with them from normal subjects. Interestingly, assuming a GH de®ciency. GH peak response lower than 3 mg/l, as evidence for Thus, there is an overlap between the maximal severe somatotrope impairment which needs rhGH secretory capacity of somatotrope cells in hypopitui- replacement therapy,31 only 60% of GHD patients taric and obese patients which are organic and func- were below this limit. This evidence implies that in tional models of GH de®ciency, respectively. Of a large percentage of adult patients with GHD, the course, we are not proposing obese patients for impairment of somatotrope function is partial and either routinary assessment of the GH reserve and/or does not need replacement therapy.44 MR scans of the hypothalamic-pituitary area or for GH reserve in obesity M Maccario et al 31 treatment with rhGH. In fact, there is evidence show- nine, or hypoglycemia in normal and obese subjects. J Clin ing that GH secretion in obesity is restored by weight Endocrinol Metab 1990; 70: 1361±1990. loss.2,6 Moreover, clinical data and IGF-I levels are 12 Casanueva FF et al. Free fatty acids block growth hormone (GH) releasing hormone stimulated GH secretion in man able to distinguish between obesity and hypopituitar- directly at the pituitary. J Clin Endocrinol Metab 1987; 65: ism.45 634±642. In all, our present data show that the pituitary GH 13 Alvarez CV et al. Evidence for a direct pituitary inhibition by releasable pool in obesity is impaired to an extent free fatty acids of in vivo growth hormone responses to growth frequently overlapping with that in organic GH de®- hormone-releasing hormone in the rat. Neuroendocrinology 1991; 53: 185±189. ciency. This evidence further points to take into great 14 Quabbe HJ et al. Hypothalamic modulation of growth hor- account the adiposity before concluding that an hypo- mone secretion in the rhesus monkey: evidence from intracer- pituitaric patient is GH de®cient. ebroventricular infusions of glucose, free fatty acids and ketone bodies. J Clin Endocrinol Metab 1991; 73: 765±770. 15 Bjorntop P. Regional obesity. In: Bjorntop P, Brodoff BN Acknowledgements (eds). Obesity Philadelphia Lippincott Company: Philadelphia; This study was supported by MURST (Ministero pp 579±586. dell'Universita' e della Ricerca Scienti®ca e Tecno- 16 Procopio M et al. Short-term fasting in obesity fails to restore the blunted GH responsiveness to GH-releasing hormone logica) and FSMEM (Fondazione Studio Malattie alone or combined with arginine. Clinical Endocrinology Endocrino-Metaboliche). 1995; 43: 665±669. The authors wish to thank Prof. F. Camanni for 17 Maccario M et al. Effects of Acipimox, an antilipolytic drug, revising the manuscript and A. Barberis, A. Bertagna, on the growth hormone (GH). Response to GH-releasing M. Taliano for their skilful technical assistance. hormone alone or combined with arginine in obesity. Meta- bolism 1996; 45: 1±6. 18 Lee EJ et al. Acipimox potentiates growth hormone (GH) References response to GH-releasing hormone with or without pyridos- tigmine by lowering serum free fatty acid in normal and obese subjects. J Clin Endocrinol Metab 1995; 80: 2495±2498. 1 Glass AR, Burman KD, Daahms WT, Boehm TM. Endocrine 19 Lillioya S et al. Free fatty acid metabolism and obesity in function in human obesity. Metabolism 1981; 130: 89±104. name: in vivo-vitro comparison. Metabolism 1986; 35: 505± 2 Williams T et al. Impaired growth hormone responses to 514. growth hormone-releasing factor in obesity. N Engl J Med 20 Polonsky KS et al. Quantitative study of insulin secretion and 1984; 311: 1043±1407. clearance in normal and obese subjects. J Clin Invest 1988; 81: 3 Cordido F, Casanueva FF, Dieguez C. Cholinergic receptor 435±441. activation by piridostigmine restores growth hormone (GH) 21 Melmed S. Insulin suppresses growth hormone secretion in rat responsiveness to GH-releasing hormone administration in pituitary cells. J Clin Invest 1984; 73: 1425±1433. obese subjects: evidence for hypothalamic somatostatinergic 22 Isaacs RE, Gardner DG, Baxter JD. Insulin regulation of rats participation in the blunted GH release of obesity. J Clin growth hormone gene expression. Endocrinology 1987; 120: Endocrinol Metab 1989; 68: 290±293. 2022±2028. 4 Veldhuis JD et al. Dual defects in pulsatile growth hormone secretion and clearances subserve the hyposomatropism of 23 Imaki T et al. The effect of glucose and free fatty acids on obesity in man. J Clin Endocrinol Metab 1991; 72: 51±59 . growth hormone (GH)±releasing factor-mediated GH secre- 5 Ghigo E et al. Arginine potentiates but does not restore the tion in rats. Endocrinology 1986; 118: 2390±2394. blunted growth hormone response to growth hormone-releas- 24 Penalva A et al. Role of cholinergic muscarinic pathways on ing hormone in obesity. Metabolism 1992; 5: 560±563. the free fatty acid inhibition of GH responses to GHRH in 6 Tanaka K et al. Very-low-calorie diet-induced weight reduc- normal men. Clin Endocrinol (Oxford) 1990; 33: 171±176. tion reverses impaired growth hormone secretion response to 25 deBoer H et al. Clinical aspects of growth hormone de®ciency growth hormone-releasing hormone, arginine and L-DOPA in in adults. Endocr Rev 1995; 16: 63±86. obesity. Metabolism 1990; 39: 892±896. 26 Hoffman DM et al. Diagnosis of growth hormone de®ciency 7 Kelijman M, Frohman LA. Enhanced growth hormone respon- in adults. Lancet 1994; 343: 1064±1068. siveness to GH-releasing hormone after dietary manipulation 27 Bates AS et al. Assessment of GH status in adults with GH in obese and non-obese subjects. J Clin Endocrinol Metab de®ciency using serum growth hormone, serum insulin-like 1988; 66: 498±494. growth factor 1, and urinary growth hormone excretion. Clin 8 Tannenbaum GS, Lapointe M, Gurd W, Finkelstein JA. Endocrinol 1995; 42: 425±430. Mechanisms of impaired growth hormone secretion in geneti- 28 Alba-Roth J, Muller AO, Schopohl J, von Werder K. Arginine cally obese Zucker rats: roles of growth hormone-releasing stimulates growth hormone secretion by suppressing endogen- factor and somatostatin. Endocrinology 1990; 127: 3087± ous somatostatin release. J Clin Endocrinol Metab 1988; 67: 3095. 1186±1189. 9 Ahmad I, Finklestein JA, Downs TR, Frohman LA. Obesity- 29 Ghigo E. Neurotransmitter control of growth hormone secre- associated decrease in growth hormone-releasing hormone tion. In: De la Cruz LF (ed). Regulation of Growth Hormone gene expression: a mechanism for reduced growth hormone and Somatic Growth Excerpta Medica, Amsterdam, 1992; pp mRNA levels in genetically obese Zucker rats. Neuroendocri- 103±136. nology 1993; 58: 332±337. 30 Ghigo E et al. New approach to the diagnosis of growth 10 Berelovitx M, Finkelstein JA, Thominet JL. Impaired growth hormone de®ciency in adults. Europ J Endocrinol 1996; 134: hormone secretion in spontaneous experimental obesity: a 352±356. result of increased somatostatin (SRIF) tone. [Abstract] 64th 31 Thorner MO et al. Diagnosis of growth hormone de®ciency in Annual Meeting of Endocrine Society, San Antonio, TX.1983; adults. J Clin Endocrinol Metab 1995; 80: 3097±3099. 84: A16. 32 Rasmussen MH et al. Lack of stimulation of 24±hour growth 11 Cordido F, Dieguez C, Casanueva FF. Effect of central hormone release by hypocaloric diet in obesity. J Clin Endo- cholinergic neurotrasmission enhancement by piridostigmine crinol Metab 1995; 80: 796±801. on the growth hormone secretion elicited by clonidine, argi- 33 Clemmons DR, Judson J Van WYR. Factors controlling blood GH reserve in obesity M Maccario et al 32 concentration of Somatomedin C. Clinics in Endocrinology weight and obese men: differential metabolic regulation and Metabolism 1984; 13: 113±143. during energy restriction. Metabolism 1995; 5: 605±610. 34 Loche S et al. The effect of galanin on baseline and GHRH- 41 Renier G, Gaudreau P, Deslauriers N, Brazeau P. In vitro and induced growth hormone secretion in obese children. Clinic in vivo growth hormone responsiveness to growth hormone- Endocrinol (Oxford) 1990; 33: 187±192. releasing factor in male and female Zucker rats. Neuroendo- 35 Marin P et al. Low concentration of insulin-like growth crinology 1989; 50: 454±459. factor±I in abdominal obesity. Inter J Obes 1993; 17: 83±89. 42 Baum HBA et al. Assessment of growth hormone secretion in 36 Minuto F et al. Spontaneous growth hormone and somatome- men with adult onset GH de®ciency compared with that in din-C/ insulin- like growth factor-I secretion in obese subjects normal men±A clinical research center study. J Clin Endocri- during puberty. J Endocrinol Invest 1988; 11: 489±495. nol Metab 1996; 81: 81±92. 37 Houston B, O'Neill IE. Insulin and growth hormone act 43 Hoech HC, Vestergaard P, Jakobsen PE, Laurgberg P. Test of synergistically to stimulate insulin-like growth factor produc- growth hormone secretion in adults: poor reproducibility of tion by cultured chicken hepatocytes. J Endocrinol 1991; 128: the insulin tolerance test. Eur J Endocrinol 1995; 133: 305± 389±393. 312. 38 Berelowitz M, Bruno JF, White JD. Regulation of hypotha- 44 Nicolson A, Toogood AA, Rahim A, Shalet SM. The pre- lamic neuropeptide expression by peripheral metabolism. valence of severe growth hormone de®ciency in adults who Trends Endocrinol Metab 1992; 3/4: 127±133. received growth hormone replacement in childhood. Clin 39 Ghigo E et al. Repetitive GHRH administration fails to Endocrinol 1996; 44: 311±316. increase the response to GHRH in obese subjects. Hormone 45 Thyssen JP, Ketelslegirs JM, Underwood LE. Multivarial Metabolic Res 1993; 25: 305±308. regulation of the Insulin-like Growth Factors. Endocr 40 Riedel M et al. Pulsatile growth hormone secretion in normal± Review 1994; 15: 80±101.