Clinical 600
Reduction of Baclofen-, but not Sodium Valproate-Induced Growth Hormone Release in Type I Diabetic Men V. Coiro1, L. Capretti3, L. Bianconi2, A. Castelli4, L. Cerri5, G. Roberti5, A. Marcato , R. Volpi1 and P. Chiodera2 Cattedra di Clinica Medica Generate, Universita di Parma, Parma Cattedra di Endocrinologia, Universita di Parma, Parma 3 Divisione di Medicina Generate, Ospedale di Codogno, Codogno 4 Divisione di Medicina Generate, Ospedale di Piacenza, Piacenza 5 Servizio di Medicina Nucleare, Ospedale di Piacenza, Piacenza, Italy
Summary The effects of sodium valproate (a drug enhancing endogenous gamma aminobutyric acid (GABA)ergic activity) and of the GABA analog baclofen (a GABA B receptor agonist) on serum G H levels was tested in 8 type I diabetic men and 8 normal controls. Sodium valproate (800 mg) or baclofen (10 mg) were given by mouth at 08.30 h on the experimental days. Control tests with a placebo were performed on different occasions. Basal G H levels were similar in controls and diabetic patients. Sodium valproate induced a 7 fold increase in serum G H concentrations in both groups. In contrast, baclofen-induced G H rise was significantly higher in normal controls (mean peak was 3.4 times higher than baseline) than in diabetic patients (mean peak was only 2.1 times higher than basal value). Serum G H levels did not change after placebo administration in any groups. These data suggest the presence of diabetes-induced alterations of a GABAergic pathway mediated by B receptors in the control of G H secretion. Alternatively, the data might indicate a change in diabetic men of other baclofen-sensitive neurotransmissions, different from GABA. Key words Baclofen — Sodium Valproate — Growth Hormone — Diabetes Mellitus
Introduction A variety of studies provide evidence of metabolic and neurochemical changes in the central nervous system of patients affected by diabetes mellitus (Mooradian 1988). Particularly, alterations in the neuroendocrine control of pituitary function have been described in insulin-dependent diabetics. In fact, different hormonal systems show modifications of their normal secretory pattern in basal conditions and in response to provocative stimuli (Mooradian 1988). In the present study, we focused our attention on the gamma-aminobutyric acid (GABA)-ergic control of Horm.metab.Res.23(1991)600-604 © Georg Thieme Verlag Stuttgart • New York
G H secretion in diabetes mellitus. GABA is widely distributed in the hypothalamus and pituitary (Okada, Nitsch-Hassler, Kim, Bak and Hassler 1971; Grandison and Guidotti 1979; Tappaz 1984) and appears to be involved in the regulation of G H secretion (Tuomisto and Mannisto 1985; Mutter 1987). Since GABA synthesis (from its metabolic precursor glutamic acid) and catabolism are affected by insulin-dependent diabetes mellitus (IDDM) (Nayeemunnisa and Nagaraj 1977; Jefferson and Neely 1982), we wondered whether functional alterations of the GABAergic system were present in diabetic patients. Administration of GABA and of several GABAergic compounds has been shown to stimulate basal G H secretion both in humans and animals (Tuomisto and Mannisto 1985; Mutter 1987). In light of these observations, we evaluated the G H response to sodium valproate, a drug which enhances endogenous GABAergic tonus through inhibition of GABA transaminase (Godin, Heiner, Mark and Mandel 1969) and baclofen, a selective agonist of GABA B receptors (Bowery, Hill, Hudson, Doble, Middlemiss, Shaw and Turnbull 1980; Harrison, Lange and Barker 1988), in men with type I diabetes mellitus. Subjects and Methods Eight insulin dependent type I diabetic men (30-43 years old) gave their informed consent to participate in this study. All were within 12% of their ideal body weight. Diabetes had had a sudden onset in all subjects, with a duration ranging from 9 to 28 years at the time of the study. From the onset of their disease the patients had been treated with insulin and at the time of this study all of them were hospitalized for adjustment of insulin therapy in order to achieve improved metabolic control. None of the patients had clinical features or laboratory evidence of ketosis or any signs of associated endocrine, renal, hepatic or other intercurrent disease. Subjects with no signs of retinopathy were chosen to participate in this study. Experiments were started after adjustment of insulin therapy (dose 20-45 IU/day, mean 30.5 IU/day of the intermediate duration monocomponent insulin plus short acting monocomponent insulin, given together once or twice daily), which remained unchanged during the period of this study. The last injection of insulin before each test was given on the day preceding each experimental day. As indices of control of the metabolic status of these patients, blood glucose levels were measured at 07.00, 11.00 and 17.00 h on the days preceding the tests; for each patient a mean value for blood glucose was obtained by averaging these three determinations. Received: 16 Oct. 1990
Accepted: 23 June 1991
Downloaded by: Universite Laval. Copyrighted material.
1
GABAergic Control of GH in WDM
Horm. metab. Res. 23 (1991)
601
Table 1 Clinical and biochemical characteristics of the diabetic patients (n = 8). Age (years) 36.2 ±2.7
Diabetes duration (years)
Insulin dose (IU/24 h)
Average blood glucose (mmol/L)
Average urinary glucose (g/day)
HbAic%
17.4 ±2.8
30.5 ±2.8
8.9 ±0.5
24.7 ±2.5
9.6 ±0.5
On the experimental days, blood samples were also taken for measurement of glycosylated hemoglobin (HbAic); blood glucose and free testosterone (fT), dihydrotestosterone (DHT) and 17 beta estradiol (E2) levels. Furthermore, 24 h urinary glucose excretion was collected at the time of the tests. Eight normal men (26-42 years old) within 10% of their normal body weight, without any signs of endocrine disease or family history of diabetes mellitus were studied as controls. All normal and diabetic subjects were tested three times; once with sodium valproate, another time with baclofen and on another occasion with a placebo. Tests were carried out in random order with an interval of at least 6 days. Tests followed a similar procedure. At 08.00 h of the experimental day, an antecubital vein was cannulated with subjects lying recumbent and fasting since the previous evening. The cannula was kept patent by a slow infusion of normal saline (NaCl 0.9%) and was used for blood sampling. Baseline blood samples were withdrawn 30 min later, just before (time 0) the administration of 800 mg of sodium valproate (Depakin, Sigma Tau), 10 mg of baclofen (Lioresal, Ciba Geigy) or placebo p. o. Drugs or placebo were given at 08.30 h. Further blood specimens were taken at times 30, 60, 90 and 120 min. All samples were used for the measurement of serum GH concentrations. GH was measured with specific RIA {Schalch and Parker 1964). The lower limit of sensitivity for GH was 0.5 ng/ml. The intra-assay coefficient of variation was 3.8% and the inter-assay coefficient of variation was 8.5 % for GH. Free T, DHT, E2, glucose and HbAic concentrations were measured in the samples taken at time 0 of all experiments. Circulating fT(McCann and Kirkish 1985), DHT {Parker, Ellegood and Mahesh 1974) and E2 (Hertogh, Thomas, Bietlot and Vanderheyden 1975) were measured with specific RIA. The sensitivity of the RIA was 0.15 pg/ml for fT, 7.8 pg/ml for DHT and 10 pg/ml for E2 respectively. Intra-assay and inter-assay coefficients of variation were respectively 4.1 % and 6.0% for Free T, 7.2% and 14.0% for DHT and 4.7% and 8.2% for E2. All specimens
from a single subject were analyzed in duplicate in the same assay. HbAic was assayed by high pressure liquid chromatography after hemolysis of the red blood cells (Koenig and Cerami 1975; Koenig, Arajo and Cerami 1976) using reagents obtained from Bio Rad Laboratories (Richmond, CA) (normal range 4.15%-9.08%). Blood glucose was measured with an IL 918 autoanalyzer (Instrumentation Laboratory, Milan, Italy), using a glucose oxidase peroxidase procedure. Data were analyzed with paired and unpaired t-test, analysis of variance or linear correlation coefficient, as appropriate. Data are reported as mean ± SE. Results The indices of control of the metabolic status in the diabetic patients are reported in Table 1. Blood circulating concentrations of sex steroids (mean of three basal determinations) in diabetic patients (fT = 26.2 ± 5.4 pg/ml; D H T = 60.413.4 pg/ml; E2 = 17.5 ± 2 . 8 pg/ml) were similar to those observed in normal controls (fT = 28.0 + 5.0 pg/ml; D H T = 62.4 ± 4 . 0 pg/ml; E 2 = 16.2 ±2.6 pg/ml). Normal controls and diabetics showed similar basal concentrations of G H in all tests (Figs. 1 and 2). Administration of placebo did not modify circulating G H levels (data not shown). Sodium
valproate
test
Serum G H concentrations rose sharply in both groups after sodium valproate administration reaching mean peak levels 7 times higher than baseline at 90 min (in both groups: p < 0.001 vs baseline). The G H response was not significantly different between the two groups (Fig. 1).
Downloaded by: Universite Laval. Copyrighted material.
Data are expressed as means ± SE. The number of diabetic patients is shown in parentheses.
V. Coiro, L. Capretti, L. Bianconi, A. Castelli, L. Cerri, G. Roberti, A. Marcato etal.
Baclofen test GH also increased after treatment with baclofen in both groups, reaching peak levels at 90 min after drug ingestion (p < 0.001 in controls, p < 0.01 in diabetics vs baseline). However, the mean GH peak was 3.4 times higher than baseline in the normal controls, but only 2.1 times higher than baseline in diabetics. The GH response was significantly higher in the normal controls than in diabetic subjects (F = 9.92, p < 0.02) (Fig. 2). When baclofen-induced peak GH responses in diabetic subjects were considered together with the duration of diabetes mellitus or the indices of their metabolic status (blood glucose or HbAi c values at time 0 of the baclofen test), no significant correlations were found. No side effects were observed after sodium valproate or baclofen administration.
ent manifestation of IDDM of long lasting duration. However, the lack of correlation between peak GH responses to baclofen and indices of metabolic status in diabetics argues against this hypothesis.
Baclofen stimulates GH secretion in men, but not in women (Monteleone, Maj, Iovino and Steardo 1988); therefore, an altered sex steroid milieu in our diabetics could be supposed to be responsible for the reduced GH response to this drug. However, this possibility was excluded by the finding of normal circulating levels of androgens and estrogens in our patients. It is possible that in contrast with sodium valproate, baclofen-induced GH release is not mediated by a GABAergic pathway, but rather by stimulation of other receptors. For example, the alpha 2 adrenoceptors are sensitive to baclofen (Fung, Swarbrick and Fillenz 1985) and play a stimulatory role in the control of GH secretion (Tuomisto and Mannisto 1985; Mutter 1987). Therefore, the reduced effect of Discussion baclofen might be attributed to the wellknown adrenergic disThe administration of either sodium valproate order affecting type I diabetics (Mooradian 1988; Christensen or baclofen stimulates GH secretion in man. Both drugs freely 1974; Robertson, Halter and Porte Jr. 1976; Bolli, Compagcross the blood brain barrier (Naik, Guidotti and Costa 1976; nucci, Cartechini, De Feo, Santeusanio, Puxeddu and Brunetti Elias, Valente, Szekeres and Grossman 1982; Racagni, Apud, 1979). However, this hypothesis is unlikely, because the actiCocchi, Locatelli and Mutter 1982); however, their actions in vation of alpha 2 receptors by the selective adrenergic agent the central nervous system appear to be exerted through differ- clonidine is known to induce higher GH responses in diabetics ent mechanisms. In fact, sodium valproate enhances the endo- than in normal subjects (Speroni, Ceda, Capretti and Volenti genous GABAergic tone through inhibition of GABA trans1983). Alternatively, if GABAergic mechanisms mediate the aminase (Godin et al. 1969), whereas baclofen shows both a actions of both sodium valproate and baclofen, we must supnumber of effects different from GABA (Kelly and Moore pose that two different GABAergic pathways stimulate GH 1978; Sawynok and La Bella 1981) and GABA like actions secretion, one mediated by GABA B receptors and unmasked (Bowery 1982). Baclofen does not bind GABA A receptors, by baclofen; the other mediated by the activation of GABA A and thus the GABA like effects of the drug have been at- receptors, and thus sensitive to sodium valproate. The former, tributed to its ability to stimulate GABA B receptors (Bowery et but not the latter would be altered in diabetes mellitus. In al. 1980; Harrison, Lange and Barker 1988). agreement with this hypothesis, Martin, Massol, Pichat and Puech (1988) have demonstrated a decrease in central GABA In the present study, we have found a reduced B receptor binding sites in diabetic rats, without alterations of GH response to baclofen in diabetic patients, who were nor- GABA A receptor density. However, endogenous GABA remally responsive to sodium valproate. The different effect of leased by sodium valproate should be active also on GABA B diabetes mellitus on sodium valproate- and baclofen-induced receptors (Harvey 1976; Hill and Bowery 1981), and thus, a GH secretion suggests an impairment of the mechanisms un- diminution of sodium valproate-induced GH release was exderlying baclofen, but not sodium valproate action in diabet- pected in diabetic patients. The lack of changes in the GH reics. It is possible that the lower GH response to baclofen in dia- sponse to sodium valproate might suggest the presence of combetics was a reflection of high blood glucose levels or a persist-
Downloaded by: Universite Laval. Copyrighted material.
602 Horm. metab. Res. 23 (1991)
GABAergic Control of GH in IDDM
Horm. metab. Res. 23 (1991)
603
Fung, S. C, M. J. Swarbrick, M. Fillenz: Effect of baclofen on in vitro noradrenaline release from rat hippocampus and cerebellum: an action at an alpha-2-adrenoceptor. Neurochem. Int. 7: 155-163 (1985) Gamse, R., D. E. Vaccaro, G. Gamse, M. Di Pace, T. 0. Fox, S. E. Leeman: Release of immunoreactive somatostatin from hypothalamic cells in culture: inhibition by y-aminobutyric acid. Proc.Natl. Acad. Sci. 77:5552-5556(1980) Godin, Y., L. Heiner, J. Mark, P. Mandel: Effects of di-N-propylacetate, an anticonvulsant compound on GABA metabolism. J. Neurochem. 16:869-873(1969) Grandison, L., A. Guidotti: -Yaminobutyric acid receptor function in rat anterior pituitary: evidence for control of prolactin release. Endocrinology 105:754-759(1979) Harrison, N. L., G. D. Lange, J. F. Barker: (-)-Baclofen activates presynaptic GABAB receptors on GABAergic inhibitory neurons from embryonic rat hippocampus. Neurosci. Lett. 85: 105—109 Interestingly, whereas a direct effect of GABA (1988) at the pituitary level is unlikely (Vijayan and Mc Cann 1978; Harvey, P. K. P.: Some aspects of the neurochemistry of Epilim. In: Fiok, Acs, Makara and Erdo 1984); a GABAergic involvement Clinical and pharmacological aspects of sodium valproate in the control of both G H R H {Murakami, Kato, Kabayama, (Epilim) in the treatment of epilepsy; Legg N. J., Ed., Tunbridge Tojo, Inoue and Imura 1985) and somatostatin (Gamse, VacWelles(1976), pp. 130-135 Hertogh, R., K. Thomas, Y. Bietlot, I. Vanderheyden: Plasma levels of caro, Gamse, Di Pace, Fox and Leeman 1980; Takahara, unconjugated estrone, estradiol and estriol and of HCS Yunoki, Hosogi, Yakushiji, Kageyama and Ofuji 1980; Stryker, throughout pregnancy in normal women. J. Clin. Endocrinol. Conlin and Reichlin 1986) release has been found in the rat. Metab. 40:93-101 (1975) These different GABAergic actions might be exerted through Hill, D. R., N. G. Bowery: 3H-baclofen and 3H-GABA bind to bicuculdifferent mechanisms, which might be alternatively sensitive line-insensitive GABAB sites in rat brain. Nature 290: 149—152 to baclofen or sodium valproate stimulation. At present, it is (1981) uncertain whether sodium valproate or baclofen selectively Jefferson, L. J., J. R. Neely: Intermediary metabolism. In: Diabetes Mellitus and obesity; Brodoff B. N. and Bleicher S. J., Eds., Willimodify the release of GHRH or somatostatin, or whether both ams and Wilkins (1982), pp. 3 -24 drugs are active on both neuroregulatory hormones. Kelly, P., K. Moore: Dopamine concentrations in the rat brain following injections into substantia nigra of baclofen, y-amino-butyric In conclusion, the present study shows reducacid, Y-hydroxybutyric acid, apomorphine and amphetamine. tion of baclofen-, but not sodium valproate-stimulated G H Neuropharmacology 17:169-174(1978) secretion in type I diabetics. These findings may be explained Koenig, R. J., D. C. Arajo, A. Cerami: Increased hemoglobin Ai c in by alterations of a GABAergic pathway mediated by B recepdiabetic mice. Diabetes 25:1 — 5(1976) Koenig, R. J, A. Cerami: Synthesis of hemoglobin Aic in normal and tors or by modifications of other baclofen-sensitive neudiabetic mice: potential model of basement membrane thickening. rotransmissions. Proc.Natl. Acad. Sci. 72:3687-3791 (1975) Martin, P., J. Massol, P. Pichat, A. J. Puech: Decreased central GABA Acknowledgements B receptor binding sites in diabetic rats. Neuropsychobiology 19: 146-148(1988) This study was in part supported by a Ministero PubMc Cann, D., L. Kirkish: Evaluation of free testosterone in serum. J. blica Istruzione grant and by Bayer Diagnostics S. P. A. Cavenago BriClin. Immunol. 8:234-236(1985) anza, Italy. Monteleone, P., M. Maj, M. Iovino, L. Steardo: Evidence for a sex difference in the basal growth hormone response to GABAergic stimulation in humans. Acta Endocrinol. 119:353—357 (1988) References Mooradian, A. D.: Diabetic complications of the central nervous system. Endocr. Rev. 9:346-356 (1988) Anden, N. E., H. Wachtel: Biochemical effects of baclofen (parachlorMutter, E. E.: Neural control of somatotropic function. Physiol. Rev. phenyl-GABA) on the dopamine and noradrenaline in the rat 67:962-1053(1987) brain. Acta Pharmacol. Toxicol. 40:310- 320 (1977) Bolli, G., P. Compagnucci, M. G. Cartechini, P. De Feo, F. Santeusanio, Murakami, Y., Y. Kato, Y. Kabayama, K. Tojo, T. Inoue, H. Imura: Involvement of growth-releasing-factor in growth hormone secreA. Puxeddu, P. Brunetti: Urinary excretion and plasma levels of tion induced by gamma-aminobutyric acid in conscious rats. Ennorepinephrine and epinephrine during diabetic ketoacidosis. docrinology 117:787-789(1985) ActaDiabetol.Lat. 16:157-167(1979) Naik, S. R., A. Guidotti, E. Costa: Central GABA receptor agonists: a Bowery, N. G.: Baclofen: 10 years on. TIPS 3:400-403 (1982) comparison of muscimol and baclofen. Neuropharmacology 15: Bowery, N. G., D. R. Hill, A. L. Hudson, A. Doble, D. N. Middlemiss, J. 479-484(1976) Shaw, M. J. Turnbull: H(—) Baclofen decreases neurotransmitter Nayeemunnisa, J. S. Nagaraj: Neurochemical correlates of alloxan release in the mammalian CNS by an action at a novel GABA rediabetes: gamma aminobutyric acid of amphibian brain. Exceptor. Nature 283:92-94 (1980) perientia33:1186-1187(1977) Christensen, N. J.: Plasma norepinephrine and epinephrine in unOkada, Y, C. Nitsch-Hassler, J. S. Kim, I. J. Bak, R. Hassler: Role of treated diabetics during fasting and after insulin administration. Y-aminobutyricacid (GABA) in the extrapyramidal motor system. Diabetes23:1-8(1974) I. Regional distribution of GABA in rabbit, rat, guinea pig, and Dutar, P., R. A. Nicoll: A physiological role of GABAB receptors in the baboon CNS. Expl. Brain Res. 13:514-518(1971) central nervous system. Nature 332:156-158 (1988) Elias, A. N., L. J. Valenta, A. V. Szekeres, M. K. Grossman: Regulatory Parker, C. R. Jr., J. O. Ellegood, V. B. Mahesh: Comparative evaluation of antisera for 5a-dihydrotestosterone radioimmunoassay. J. role of gamma-aminobutyric acid in pituitary hormone secretion. Steroid Biochem. 6:9-13(1974) Psychoneuroendocrinology 7:15—30 (1982) Racagni, G., J. A. Apud, D. Cocchi, V. Locatelli, E. E. Mutter: GABAerFiok, J, Z. Acs, G. B. Makara, S. L. Erdo: Site of y-aminobutyric acid gic control of anterior pituitary hormone secretion. Life Sci. 31: (GABA)-mediated inhibition of growth hormone secretion in the 823-838(1982) rat. Neuroendocrinology 39:510-516(1984)
Downloaded by: Universite Laval. Copyrighted material.
pensatory mechanisms. On the other hand, it is possible that the G H releasing effect of sodium valproate is mediated only by GABA A receptors and thus that it is not affected by GABA B receptor changes. This hypothesis requires a complete separation between GABA A and GABA B pathways and is supported by the observation that GABA A and GABA B sites have different distributions in the central nervous system (Dutar and Nicoll 1988). G H secretion from the pituitary is under the dual regulation of the inhibitory peptide somatostatin and the stimulatory hormone GHRH (Tuomisto and Mannisto 1985; Mutter 1987). Therefore, at least two hypothalamic neuroanatomic pathways may be supposed in the control of G H secretion.
Horm. metab. Res. 23 (1991)
V. Coiro, L. Capretti, L. Bianconi, A. Castelli, L. Cerri, G. Roberti, A. Marcato et al.
Robertson, R. P., Halter, J. B., D. J. Porte Jr.: A role for alpha-adrenergic receptors in abnormal insulin secretion in diabetes mellitus. J. Clin. Invest. 57:791-795(1976) Sawynok, J., F. S. La Bella: GAB A and baclofen potentiate the K + evoked release of methionine-enkephalin from rat striatal slices. Eur. J.Pharmacol. 70:103-110(1981) Schalch, D. S., M. L. Parker: A sensitive double antibody immunoassay for human growth hormone in plasma. Nature 203:1141 —1145 (1964) Speroni, G., G. P. Ceda, L. Capretti, G. Volenti: Clonidine and GH secretion in insulin dependent diabetes (ISS). Horm. Met. Res. 15: 46-47(1983) Stryker, T. D., T. Conlin, S. Reichlin: Influence of a benzodiazepine, midazolam, and gamma-aminobutyric acid (GABA) on basal somatostatin secretion from cerebral and diencephalic neurons in dispersed cell culture. Brain Res. 362:339-343 (1986) Takahara, J., S. Yunoki, H. Hosogi, W. Yakushiji, J. Kageyama, T. Ofuji: Concomitant increase in serum growth hormone and hypothalamic somatostatin in rats after injection of gamma-
aminobutyric acid, aminooxyacetic acid or gamma-hydroxybutyricacid. Endocrinology 106:343-347 (1980) Tappaz, M. L.: GABA and anterior pituitary function: anatomical data. Psychoneuroendocrinology 9: 85—95 (1984) Tuomisto, J., P. Mannisto: Neurotransmitter regulation of anterior pituitary hormones. Pharmacol. Rev. 37:249-332(1985) Vijayan, E., S. M. Mc Cann: Effects of intraventricular injection of gamma-aminobutyric acid (GABA) on plasma growth hormone and thyrotropin in conscious ovariectomized rats. Endocrinology 103:1888-1893(1978) Requests for reprints should be addressed to: Vittorio Coiro, M. D. Cattedra di Clinica Medica Generale e Terapia Medica Universita di Parma Via Gramsci l4 1-43100 Parma (Italy)
Downloaded by: Universite Laval. Copyrighted material.
604
Reduction of Baclofen-, but not Sodium Valproate-Induced Growth Hormone Release in Type I Diabetic Men V. Coiro1, L. Capretti3, L. Bianconi2, A. Castelli4, L. Cerri5, G. Roberti5, A. Marcato , R. Volpi1 and P. Chiodera2 Cattedra di Clinica Medica Generate, Universita di Parma, Parma Cattedra di Endocrinologia, Universita di Parma, Parma 3 Divisione di Medicina Generate, Ospedale di Codogno, Codogno 4 Divisione di Medicina Generate, Ospedale di Piacenza, Piacenza 5 Servizio di Medicina Nucleare, Ospedale di Piacenza, Piacenza, Italy
Summary The effects of sodium valproate (a drug enhancing endogenous gamma aminobutyric acid (GABA)ergic activity) and of the GABA analog baclofen (a GABA B receptor agonist) on serum G H levels was tested in 8 type I diabetic men and 8 normal controls. Sodium valproate (800 mg) or baclofen (10 mg) were given by mouth at 08.30 h on the experimental days. Control tests with a placebo were performed on different occasions. Basal G H levels were similar in controls and diabetic patients. Sodium valproate induced a 7 fold increase in serum G H concentrations in both groups. In contrast, baclofen-induced G H rise was significantly higher in normal controls (mean peak was 3.4 times higher than baseline) than in diabetic patients (mean peak was only 2.1 times higher than basal value). Serum G H levels did not change after placebo administration in any groups. These data suggest the presence of diabetes-induced alterations of a GABAergic pathway mediated by B receptors in the control of G H secretion. Alternatively, the data might indicate a change in diabetic men of other baclofen-sensitive neurotransmissions, different from GABA. Key words Baclofen — Sodium Valproate — Growth Hormone — Diabetes Mellitus
Introduction A variety of studies provide evidence of metabolic and neurochemical changes in the central nervous system of patients affected by diabetes mellitus (Mooradian 1988). Particularly, alterations in the neuroendocrine control of pituitary function have been described in insulin-dependent diabetics. In fact, different hormonal systems show modifications of their normal secretory pattern in basal conditions and in response to provocative stimuli (Mooradian 1988). In the present study, we focused our attention on the gamma-aminobutyric acid (GABA)-ergic control of Horm.metab.Res.23(1991)600-604 © Georg Thieme Verlag Stuttgart • New York
G H secretion in diabetes mellitus. GABA is widely distributed in the hypothalamus and pituitary (Okada, Nitsch-Hassler, Kim, Bak and Hassler 1971; Grandison and Guidotti 1979; Tappaz 1984) and appears to be involved in the regulation of G H secretion (Tuomisto and Mannisto 1985; Mutter 1987). Since GABA synthesis (from its metabolic precursor glutamic acid) and catabolism are affected by insulin-dependent diabetes mellitus (IDDM) (Nayeemunnisa and Nagaraj 1977; Jefferson and Neely 1982), we wondered whether functional alterations of the GABAergic system were present in diabetic patients. Administration of GABA and of several GABAergic compounds has been shown to stimulate basal G H secretion both in humans and animals (Tuomisto and Mannisto 1985; Mutter 1987). In light of these observations, we evaluated the G H response to sodium valproate, a drug which enhances endogenous GABAergic tonus through inhibition of GABA transaminase (Godin, Heiner, Mark and Mandel 1969) and baclofen, a selective agonist of GABA B receptors (Bowery, Hill, Hudson, Doble, Middlemiss, Shaw and Turnbull 1980; Harrison, Lange and Barker 1988), in men with type I diabetes mellitus. Subjects and Methods Eight insulin dependent type I diabetic men (30-43 years old) gave their informed consent to participate in this study. All were within 12% of their ideal body weight. Diabetes had had a sudden onset in all subjects, with a duration ranging from 9 to 28 years at the time of the study. From the onset of their disease the patients had been treated with insulin and at the time of this study all of them were hospitalized for adjustment of insulin therapy in order to achieve improved metabolic control. None of the patients had clinical features or laboratory evidence of ketosis or any signs of associated endocrine, renal, hepatic or other intercurrent disease. Subjects with no signs of retinopathy were chosen to participate in this study. Experiments were started after adjustment of insulin therapy (dose 20-45 IU/day, mean 30.5 IU/day of the intermediate duration monocomponent insulin plus short acting monocomponent insulin, given together once or twice daily), which remained unchanged during the period of this study. The last injection of insulin before each test was given on the day preceding each experimental day. As indices of control of the metabolic status of these patients, blood glucose levels were measured at 07.00, 11.00 and 17.00 h on the days preceding the tests; for each patient a mean value for blood glucose was obtained by averaging these three determinations. Received: 16 Oct. 1990
Accepted: 23 June 1991
Downloaded by: Universite Laval. Copyrighted material.
1
GABAergic Control of GH in WDM
Horm. metab. Res. 23 (1991)
601
Table 1 Clinical and biochemical characteristics of the diabetic patients (n = 8). Age (years) 36.2 ±2.7
Diabetes duration (years)
Insulin dose (IU/24 h)
Average blood glucose (mmol/L)
Average urinary glucose (g/day)
HbAic%
17.4 ±2.8
30.5 ±2.8
8.9 ±0.5
24.7 ±2.5
9.6 ±0.5
On the experimental days, blood samples were also taken for measurement of glycosylated hemoglobin (HbAic); blood glucose and free testosterone (fT), dihydrotestosterone (DHT) and 17 beta estradiol (E2) levels. Furthermore, 24 h urinary glucose excretion was collected at the time of the tests. Eight normal men (26-42 years old) within 10% of their normal body weight, without any signs of endocrine disease or family history of diabetes mellitus were studied as controls. All normal and diabetic subjects were tested three times; once with sodium valproate, another time with baclofen and on another occasion with a placebo. Tests were carried out in random order with an interval of at least 6 days. Tests followed a similar procedure. At 08.00 h of the experimental day, an antecubital vein was cannulated with subjects lying recumbent and fasting since the previous evening. The cannula was kept patent by a slow infusion of normal saline (NaCl 0.9%) and was used for blood sampling. Baseline blood samples were withdrawn 30 min later, just before (time 0) the administration of 800 mg of sodium valproate (Depakin, Sigma Tau), 10 mg of baclofen (Lioresal, Ciba Geigy) or placebo p. o. Drugs or placebo were given at 08.30 h. Further blood specimens were taken at times 30, 60, 90 and 120 min. All samples were used for the measurement of serum GH concentrations. GH was measured with specific RIA {Schalch and Parker 1964). The lower limit of sensitivity for GH was 0.5 ng/ml. The intra-assay coefficient of variation was 3.8% and the inter-assay coefficient of variation was 8.5 % for GH. Free T, DHT, E2, glucose and HbAic concentrations were measured in the samples taken at time 0 of all experiments. Circulating fT(McCann and Kirkish 1985), DHT {Parker, Ellegood and Mahesh 1974) and E2 (Hertogh, Thomas, Bietlot and Vanderheyden 1975) were measured with specific RIA. The sensitivity of the RIA was 0.15 pg/ml for fT, 7.8 pg/ml for DHT and 10 pg/ml for E2 respectively. Intra-assay and inter-assay coefficients of variation were respectively 4.1 % and 6.0% for Free T, 7.2% and 14.0% for DHT and 4.7% and 8.2% for E2. All specimens
from a single subject were analyzed in duplicate in the same assay. HbAic was assayed by high pressure liquid chromatography after hemolysis of the red blood cells (Koenig and Cerami 1975; Koenig, Arajo and Cerami 1976) using reagents obtained from Bio Rad Laboratories (Richmond, CA) (normal range 4.15%-9.08%). Blood glucose was measured with an IL 918 autoanalyzer (Instrumentation Laboratory, Milan, Italy), using a glucose oxidase peroxidase procedure. Data were analyzed with paired and unpaired t-test, analysis of variance or linear correlation coefficient, as appropriate. Data are reported as mean ± SE. Results The indices of control of the metabolic status in the diabetic patients are reported in Table 1. Blood circulating concentrations of sex steroids (mean of three basal determinations) in diabetic patients (fT = 26.2 ± 5.4 pg/ml; D H T = 60.413.4 pg/ml; E2 = 17.5 ± 2 . 8 pg/ml) were similar to those observed in normal controls (fT = 28.0 + 5.0 pg/ml; D H T = 62.4 ± 4 . 0 pg/ml; E 2 = 16.2 ±2.6 pg/ml). Normal controls and diabetics showed similar basal concentrations of G H in all tests (Figs. 1 and 2). Administration of placebo did not modify circulating G H levels (data not shown). Sodium
valproate
test
Serum G H concentrations rose sharply in both groups after sodium valproate administration reaching mean peak levels 7 times higher than baseline at 90 min (in both groups: p < 0.001 vs baseline). The G H response was not significantly different between the two groups (Fig. 1).
Downloaded by: Universite Laval. Copyrighted material.
Data are expressed as means ± SE. The number of diabetic patients is shown in parentheses.
V. Coiro, L. Capretti, L. Bianconi, A. Castelli, L. Cerri, G. Roberti, A. Marcato etal.
Baclofen test GH also increased after treatment with baclofen in both groups, reaching peak levels at 90 min after drug ingestion (p < 0.001 in controls, p < 0.01 in diabetics vs baseline). However, the mean GH peak was 3.4 times higher than baseline in the normal controls, but only 2.1 times higher than baseline in diabetics. The GH response was significantly higher in the normal controls than in diabetic subjects (F = 9.92, p < 0.02) (Fig. 2). When baclofen-induced peak GH responses in diabetic subjects were considered together with the duration of diabetes mellitus or the indices of their metabolic status (blood glucose or HbAi c values at time 0 of the baclofen test), no significant correlations were found. No side effects were observed after sodium valproate or baclofen administration.
ent manifestation of IDDM of long lasting duration. However, the lack of correlation between peak GH responses to baclofen and indices of metabolic status in diabetics argues against this hypothesis.
Baclofen stimulates GH secretion in men, but not in women (Monteleone, Maj, Iovino and Steardo 1988); therefore, an altered sex steroid milieu in our diabetics could be supposed to be responsible for the reduced GH response to this drug. However, this possibility was excluded by the finding of normal circulating levels of androgens and estrogens in our patients. It is possible that in contrast with sodium valproate, baclofen-induced GH release is not mediated by a GABAergic pathway, but rather by stimulation of other receptors. For example, the alpha 2 adrenoceptors are sensitive to baclofen (Fung, Swarbrick and Fillenz 1985) and play a stimulatory role in the control of GH secretion (Tuomisto and Mannisto 1985; Mutter 1987). Therefore, the reduced effect of Discussion baclofen might be attributed to the wellknown adrenergic disThe administration of either sodium valproate order affecting type I diabetics (Mooradian 1988; Christensen or baclofen stimulates GH secretion in man. Both drugs freely 1974; Robertson, Halter and Porte Jr. 1976; Bolli, Compagcross the blood brain barrier (Naik, Guidotti and Costa 1976; nucci, Cartechini, De Feo, Santeusanio, Puxeddu and Brunetti Elias, Valente, Szekeres and Grossman 1982; Racagni, Apud, 1979). However, this hypothesis is unlikely, because the actiCocchi, Locatelli and Mutter 1982); however, their actions in vation of alpha 2 receptors by the selective adrenergic agent the central nervous system appear to be exerted through differ- clonidine is known to induce higher GH responses in diabetics ent mechanisms. In fact, sodium valproate enhances the endo- than in normal subjects (Speroni, Ceda, Capretti and Volenti genous GABAergic tone through inhibition of GABA trans1983). Alternatively, if GABAergic mechanisms mediate the aminase (Godin et al. 1969), whereas baclofen shows both a actions of both sodium valproate and baclofen, we must supnumber of effects different from GABA (Kelly and Moore pose that two different GABAergic pathways stimulate GH 1978; Sawynok and La Bella 1981) and GABA like actions secretion, one mediated by GABA B receptors and unmasked (Bowery 1982). Baclofen does not bind GABA A receptors, by baclofen; the other mediated by the activation of GABA A and thus the GABA like effects of the drug have been at- receptors, and thus sensitive to sodium valproate. The former, tributed to its ability to stimulate GABA B receptors (Bowery et but not the latter would be altered in diabetes mellitus. In al. 1980; Harrison, Lange and Barker 1988). agreement with this hypothesis, Martin, Massol, Pichat and Puech (1988) have demonstrated a decrease in central GABA In the present study, we have found a reduced B receptor binding sites in diabetic rats, without alterations of GH response to baclofen in diabetic patients, who were nor- GABA A receptor density. However, endogenous GABA remally responsive to sodium valproate. The different effect of leased by sodium valproate should be active also on GABA B diabetes mellitus on sodium valproate- and baclofen-induced receptors (Harvey 1976; Hill and Bowery 1981), and thus, a GH secretion suggests an impairment of the mechanisms un- diminution of sodium valproate-induced GH release was exderlying baclofen, but not sodium valproate action in diabet- pected in diabetic patients. The lack of changes in the GH reics. It is possible that the lower GH response to baclofen in dia- sponse to sodium valproate might suggest the presence of combetics was a reflection of high blood glucose levels or a persist-
Downloaded by: Universite Laval. Copyrighted material.
602 Horm. metab. Res. 23 (1991)
GABAergic Control of GH in IDDM
Horm. metab. Res. 23 (1991)
603
Fung, S. C, M. J. Swarbrick, M. Fillenz: Effect of baclofen on in vitro noradrenaline release from rat hippocampus and cerebellum: an action at an alpha-2-adrenoceptor. Neurochem. Int. 7: 155-163 (1985) Gamse, R., D. E. Vaccaro, G. Gamse, M. Di Pace, T. 0. Fox, S. E. Leeman: Release of immunoreactive somatostatin from hypothalamic cells in culture: inhibition by y-aminobutyric acid. Proc.Natl. Acad. Sci. 77:5552-5556(1980) Godin, Y., L. Heiner, J. Mark, P. Mandel: Effects of di-N-propylacetate, an anticonvulsant compound on GABA metabolism. J. Neurochem. 16:869-873(1969) Grandison, L., A. Guidotti: -Yaminobutyric acid receptor function in rat anterior pituitary: evidence for control of prolactin release. Endocrinology 105:754-759(1979) Harrison, N. L., G. D. Lange, J. F. Barker: (-)-Baclofen activates presynaptic GABAB receptors on GABAergic inhibitory neurons from embryonic rat hippocampus. Neurosci. Lett. 85: 105—109 Interestingly, whereas a direct effect of GABA (1988) at the pituitary level is unlikely (Vijayan and Mc Cann 1978; Harvey, P. K. P.: Some aspects of the neurochemistry of Epilim. In: Fiok, Acs, Makara and Erdo 1984); a GABAergic involvement Clinical and pharmacological aspects of sodium valproate in the control of both G H R H {Murakami, Kato, Kabayama, (Epilim) in the treatment of epilepsy; Legg N. J., Ed., Tunbridge Tojo, Inoue and Imura 1985) and somatostatin (Gamse, VacWelles(1976), pp. 130-135 Hertogh, R., K. Thomas, Y. Bietlot, I. Vanderheyden: Plasma levels of caro, Gamse, Di Pace, Fox and Leeman 1980; Takahara, unconjugated estrone, estradiol and estriol and of HCS Yunoki, Hosogi, Yakushiji, Kageyama and Ofuji 1980; Stryker, throughout pregnancy in normal women. J. Clin. Endocrinol. Conlin and Reichlin 1986) release has been found in the rat. Metab. 40:93-101 (1975) These different GABAergic actions might be exerted through Hill, D. R., N. G. Bowery: 3H-baclofen and 3H-GABA bind to bicuculdifferent mechanisms, which might be alternatively sensitive line-insensitive GABAB sites in rat brain. Nature 290: 149—152 to baclofen or sodium valproate stimulation. At present, it is (1981) uncertain whether sodium valproate or baclofen selectively Jefferson, L. J., J. R. Neely: Intermediary metabolism. In: Diabetes Mellitus and obesity; Brodoff B. N. and Bleicher S. J., Eds., Willimodify the release of GHRH or somatostatin, or whether both ams and Wilkins (1982), pp. 3 -24 drugs are active on both neuroregulatory hormones. Kelly, P., K. Moore: Dopamine concentrations in the rat brain following injections into substantia nigra of baclofen, y-amino-butyric In conclusion, the present study shows reducacid, Y-hydroxybutyric acid, apomorphine and amphetamine. tion of baclofen-, but not sodium valproate-stimulated G H Neuropharmacology 17:169-174(1978) secretion in type I diabetics. These findings may be explained Koenig, R. J., D. C. Arajo, A. Cerami: Increased hemoglobin Ai c in by alterations of a GABAergic pathway mediated by B recepdiabetic mice. Diabetes 25:1 — 5(1976) Koenig, R. J, A. Cerami: Synthesis of hemoglobin Aic in normal and tors or by modifications of other baclofen-sensitive neudiabetic mice: potential model of basement membrane thickening. rotransmissions. Proc.Natl. Acad. Sci. 72:3687-3791 (1975) Martin, P., J. Massol, P. Pichat, A. J. Puech: Decreased central GABA Acknowledgements B receptor binding sites in diabetic rats. Neuropsychobiology 19: 146-148(1988) This study was in part supported by a Ministero PubMc Cann, D., L. Kirkish: Evaluation of free testosterone in serum. J. blica Istruzione grant and by Bayer Diagnostics S. P. A. Cavenago BriClin. Immunol. 8:234-236(1985) anza, Italy. Monteleone, P., M. Maj, M. Iovino, L. Steardo: Evidence for a sex difference in the basal growth hormone response to GABAergic stimulation in humans. Acta Endocrinol. 119:353—357 (1988) References Mooradian, A. D.: Diabetic complications of the central nervous system. Endocr. Rev. 9:346-356 (1988) Anden, N. E., H. Wachtel: Biochemical effects of baclofen (parachlorMutter, E. E.: Neural control of somatotropic function. Physiol. Rev. phenyl-GABA) on the dopamine and noradrenaline in the rat 67:962-1053(1987) brain. Acta Pharmacol. Toxicol. 40:310- 320 (1977) Bolli, G., P. Compagnucci, M. G. Cartechini, P. De Feo, F. Santeusanio, Murakami, Y., Y. Kato, Y. Kabayama, K. Tojo, T. Inoue, H. Imura: Involvement of growth-releasing-factor in growth hormone secreA. Puxeddu, P. Brunetti: Urinary excretion and plasma levels of tion induced by gamma-aminobutyric acid in conscious rats. Ennorepinephrine and epinephrine during diabetic ketoacidosis. docrinology 117:787-789(1985) ActaDiabetol.Lat. 16:157-167(1979) Naik, S. R., A. Guidotti, E. Costa: Central GABA receptor agonists: a Bowery, N. G.: Baclofen: 10 years on. TIPS 3:400-403 (1982) comparison of muscimol and baclofen. Neuropharmacology 15: Bowery, N. G., D. R. Hill, A. L. Hudson, A. Doble, D. N. Middlemiss, J. 479-484(1976) Shaw, M. J. Turnbull: H(—) Baclofen decreases neurotransmitter Nayeemunnisa, J. S. Nagaraj: Neurochemical correlates of alloxan release in the mammalian CNS by an action at a novel GABA rediabetes: gamma aminobutyric acid of amphibian brain. Exceptor. Nature 283:92-94 (1980) perientia33:1186-1187(1977) Christensen, N. J.: Plasma norepinephrine and epinephrine in unOkada, Y, C. Nitsch-Hassler, J. S. Kim, I. J. Bak, R. Hassler: Role of treated diabetics during fasting and after insulin administration. Y-aminobutyricacid (GABA) in the extrapyramidal motor system. Diabetes23:1-8(1974) I. Regional distribution of GABA in rabbit, rat, guinea pig, and Dutar, P., R. A. Nicoll: A physiological role of GABAB receptors in the baboon CNS. Expl. Brain Res. 13:514-518(1971) central nervous system. Nature 332:156-158 (1988) Elias, A. N., L. J. Valenta, A. V. Szekeres, M. K. Grossman: Regulatory Parker, C. R. Jr., J. O. Ellegood, V. B. Mahesh: Comparative evaluation of antisera for 5a-dihydrotestosterone radioimmunoassay. J. role of gamma-aminobutyric acid in pituitary hormone secretion. Steroid Biochem. 6:9-13(1974) Psychoneuroendocrinology 7:15—30 (1982) Racagni, G., J. A. Apud, D. Cocchi, V. Locatelli, E. E. Mutter: GABAerFiok, J, Z. Acs, G. B. Makara, S. L. Erdo: Site of y-aminobutyric acid gic control of anterior pituitary hormone secretion. Life Sci. 31: (GABA)-mediated inhibition of growth hormone secretion in the 823-838(1982) rat. Neuroendocrinology 39:510-516(1984)
Downloaded by: Universite Laval. Copyrighted material.
pensatory mechanisms. On the other hand, it is possible that the G H releasing effect of sodium valproate is mediated only by GABA A receptors and thus that it is not affected by GABA B receptor changes. This hypothesis requires a complete separation between GABA A and GABA B pathways and is supported by the observation that GABA A and GABA B sites have different distributions in the central nervous system (Dutar and Nicoll 1988). G H secretion from the pituitary is under the dual regulation of the inhibitory peptide somatostatin and the stimulatory hormone GHRH (Tuomisto and Mannisto 1985; Mutter 1987). Therefore, at least two hypothalamic neuroanatomic pathways may be supposed in the control of G H secretion.
Horm. metab. Res. 23 (1991)
V. Coiro, L. Capretti, L. Bianconi, A. Castelli, L. Cerri, G. Roberti, A. Marcato et al.
Robertson, R. P., Halter, J. B., D. J. Porte Jr.: A role for alpha-adrenergic receptors in abnormal insulin secretion in diabetes mellitus. J. Clin. Invest. 57:791-795(1976) Sawynok, J., F. S. La Bella: GAB A and baclofen potentiate the K + evoked release of methionine-enkephalin from rat striatal slices. Eur. J.Pharmacol. 70:103-110(1981) Schalch, D. S., M. L. Parker: A sensitive double antibody immunoassay for human growth hormone in plasma. Nature 203:1141 —1145 (1964) Speroni, G., G. P. Ceda, L. Capretti, G. Volenti: Clonidine and GH secretion in insulin dependent diabetes (ISS). Horm. Met. Res. 15: 46-47(1983) Stryker, T. D., T. Conlin, S. Reichlin: Influence of a benzodiazepine, midazolam, and gamma-aminobutyric acid (GABA) on basal somatostatin secretion from cerebral and diencephalic neurons in dispersed cell culture. Brain Res. 362:339-343 (1986) Takahara, J., S. Yunoki, H. Hosogi, W. Yakushiji, J. Kageyama, T. Ofuji: Concomitant increase in serum growth hormone and hypothalamic somatostatin in rats after injection of gamma-
aminobutyric acid, aminooxyacetic acid or gamma-hydroxybutyricacid. Endocrinology 106:343-347 (1980) Tappaz, M. L.: GABA and anterior pituitary function: anatomical data. Psychoneuroendocrinology 9: 85—95 (1984) Tuomisto, J., P. Mannisto: Neurotransmitter regulation of anterior pituitary hormones. Pharmacol. Rev. 37:249-332(1985) Vijayan, E., S. M. Mc Cann: Effects of intraventricular injection of gamma-aminobutyric acid (GABA) on plasma growth hormone and thyrotropin in conscious ovariectomized rats. Endocrinology 103:1888-1893(1978) Requests for reprints should be addressed to: Vittorio Coiro, M. D. Cattedra di Clinica Medica Generale e Terapia Medica Universita di Parma Via Gramsci l4 1-43100 Parma (Italy)
Downloaded by: Universite Laval. Copyrighted material.
604
