Psychoneurocndocrinology,Vol. 15, No. 4, pp. 253-259,
1990
0306-4.530/90 $3.00 + 0.0(; P~gmnon Press plc
Printed in Great Britain
GROWTH HORMONE (GH) RESPONSE TO CLONIDINE A N D GROWTH HORMONE RELEASING FACTOR (GRF) IN NORMAL CONTROLS MANUEL E. TANCER, MURRAY B. STEIN a n d THOMAS W. UHDE Section on Anxiety and Affective Disorders, Biological Psychiatry Branch, National Institute of Mental Health, Bethesda, Maryland, U.S.A. (Received 4 August 1989; infinalform 17 May 1990)
SUMMARY To investigate the relationship between the plasma growth hormone (GH) response to provocative challenge with the hypothalamic peptide growth hormone-releasing factor (GRF) and the o~z-adrenergic agonist clonidine, we administered GRF (1 I.tg/kg), clonidine (2 p.g/kg), and placebo to 21 healthy normal controls (13 men and eight women). Both elonidine and GRF caused significant increases in plasma GH levels over baseline. The peak GH-responses to GRF and elonidine were similar (GRF = 8.7+ 6.7 ng/ml; elonidine = 6.5+5.9 ng/ml; W'tleoxon test: s = 361, z ffi-1.31, p = NS). The GH responses to GRF and elonidine were significantly correlated (rs = 0.62, n = 20, p = 0.004). Unexpectedly, we found that five of the 21 (26%) normal controls had no GH secretory response to either GRF or elonidine. There was a modest gender effect with elonidine (men > women; p < 0.06) and a negative correlation between GH secretion and age with both GRF and elonidine. Neither GRF nor elonidine had an effect on eortisol levels (DRUG × TIME interaction: F(8,152)•0.60, p = NS). These findings are consistent with animal studies suggesting that the GH response to clonidine is mediated by GRF. The age and gender effects underscore the importance of careful matching for these factors in studies measuring the GH secretory response.
INTRODUCTION THE GROWTH HORMONE (GH) response to neuroendocrine challenge paradigms has been widely used as a trait marker for psychiatric disorders. "Blunted" or subnormal GH responses to the a2adrenergic agonist clonidine has been consistently reported in patients with major depression (Matussek et al., 1980; Chamey et al., 1982; Siever et al., 1984) and in four o f five reports to date in patients with panic disorder (Uhde et al., 1986; Chamey & Heninger, 1986; Nutt, 1989; Curtis et al., 1989; but see Schittecatte et al., 1988). Furthermore, normal GH responses to clonidine have been reported in "reactive" depression (Matussek et al., 1980), anorexia nervosa (Brambilla et al., 1987), an social phobia (Tancer & Uhde, 1989). The subnormal G H responses to clonidine are generally interpreted as indirect evidence of post-synaptic adrenergic receptor down-regulation (Uhde, 1988). The identification and synthesis o f growth h o r m o n e releasing factor (GRF) by Guillemin et al. (1982) and Rivier et al. (1982) provided a specific tool to study the dynamic regulation of GH-release. Studies to date with G R F in psychiatric patients have yielded mixed results, Address correspondence and reprint requests to: Dr. Thomas W. Uhde, Chief, Section on Anxiety and Affeetive Disorders, Biological Psychiatry Branch, National Institute of Mental Health, 9000 Rockville Pike, Building 10, Room 3S-239, Bethesda MD 20892, USA. 253
254
M.E. TANCERet al.
with " b l u n t e d " and " e x a g g e r a t e d " G H responses reported (Krishnan et al., 1988; Lesch e t a / , , 1988; Rapaport et al., 1989; Brambilla et al., 1989). The relationship between the G H response to clonidine and to G R F is not entirely understood. There is s o m e e v i d e n c e in rats that the G H response to clonidine is, in fact, m e d i a t e d by GRF (Miki et al., 1984; Celia et al., 1987). A positive correlation between the G H responses to clonidine and G R F across individuals would be consistent with such a mechanism. Such a positive correlation has been reported b y Brambilla et al. (1989) in a c o m b i n e d sample o f women controls and patients with anorexia nervosa and by Lesch et al. (1988) in depressed patients and controls (14 w o m e n and six men). In order to further explore the relationship between the G H responses to clonidine and to GRF, we administered clonidine (2 ~tg/kg), G R F (1 ~tg/kg), and placebo to 21 healthy volunteers in a randomized, double-blind fashion. Because o f discrepancies in the literature regarding the effect of clonidine on cortisol levels (Lal et al., 1975; C h a m e y et al., 1982; Stein & Uhde, 1988; Tancer & Uhde, 1989) and a single report o f the effects o f G R F on cortisol levels ( T h o m e r et al., 1983), we also report the p l a s m a cortisol responses to GRF, clonidine, and placebo in the subjects. SUBJECTS A N D METHODS Subjects The subjects were 21 normal voluntee.rs [13 men and eight women; mean age 28.2+5.9 years (range 20-44 yr)], who had no current or past history of psychopathology as determined by a structured interview modified from the Schedule for Affective Disorders and Schizophrenia (SADS,LA) (Mannuzza et al., 1985). All had negative family histories of depressioni drug or alcohol abuse, anxiety disorders, and schizophrenia within :first degree relatives. The participants were in good physical health as d e t ~ by history, physical e x ~ t i o u ~ electrocardiogram, and laboratory studies (SMAC-20, CBC, urinalysis, and thyroid indices) and were medication-free for at least one month prior to participation: The mean weight of the subjects was 70.2+ 12.4 kg (range 48:100 kg); their mean Quetelet index was 24.24"3.9 kg/m 2 (range 17.7-33.0). The women, who were all pre- menopausal, were studied during the first i0 days of their menstrual cycles. All subjects gave oral and written informed consent prior to entering the study. Methods All subjects were on a low monoamine, low caffeine (< 100 mg/day) diet for at least 72 hr prior to each of the three study days. Following an overnight fast, participants came to the outpatient clinic at 0900h. At that time~ with the subjects in a recumbent position, an intravenous catheter was inserted in the non-dominant antecubital vein and was kept patent with a slow infusion of normal saline. One hour following the intravenous catheter insertion, subjects received, in a double,blind, randomized fashion, clunidine (2 ~tg/kg), GRF (1 gg/kg) (Bacbem GRF 1-44 amide), or placebo over 1 rain. Each control received all three infusions, with at least 48 hr ~tween study days. Blood samples were obtained via a three-way stopcock at -15, 0, +15, +30, +45, and +60 min. GH and C o , sol were measured by radioimmunoassay (Hazelton Laboratory, Vienna, Virginia). The GH assay has a sensitivity ol 0.5 ng/ml and inter- and intra- assay variabilities of 7.3% and 4.9%, respectively. The cortisol assay has inter- and intra- assay variabilities of 12.3% and 5.5%, respectively. Statistical Methods As the G H responses to clonidine and GRF were not normally distributed,they were log-transformed. Both:the actual and log-transformed GH values are presented, Analysis of variance (ANOVA) with repeated measures was used to analyze the log-transfonne,d OH responses and the cortisol responses across TIME (hasoline vs. posy infusion levels) and DRUG (clonidine~ GRF, or placebo) conditions. When the ANOVA indicated significant main effects of TIME or DRUG conditions, Wilcoxon paired tests were used to compare variables. The -15 and 0 time points were averaged and reported as baseline levels. The area under the G H response cmwe from 0 to 60 rain (AUC) was calculatedby the trapezoidal approximation. It should be noted that A U C responses correlated highly with peak responses (rs= 0.96, p0.10 are reported as not significant (NS). Means ate reported + one standard deviation (SD). RESULTS A s expected, clonidine and GRF, but not placebo, administration resulted in significant rises in plasma G H levels c o m p a r e d to baseline (Table I, Fig. 1). The ANOVA revealed significant main effects for D R U G (clonidine vs. G R F vs. placebo) (F = 44.5, d f = 2 , 3 8 , p < 0 . 0 0 0 1 ) and for TIME (baseline vs. subsequent time points) (F =60.5, d f = 4,76, p < 0 . 0 0 0 1 ) , and a significant D R U G × T I M E interaction (F = 2 4 . 3 , d f = 8 , 1 5 2 , p < 0 . 0 0 0 1 ) . Post-hoc testing r e v e a l e d significant differences in the time course of the G H response to clonidine and GRF, with a greater G H response following G R F at +15 min (Wilcoxon test: s = 2 6 6 . 5 , z = -4.02, p < 0 . 0 0 1 ) and a trend for a greater GH-response to G R F at +30 min (Wilcoxon text: s = 345.0, z = - 1 . 9 4 , p < 0 . 0 6 ) . The G H responses
TABLE I. GROWTH HORMONE RESPONSES (NG/ML) TO PROVOCATIVE CHALLENGE Baseline
+15 min
+30 m i n
+45 m i n
+60 rain
GRF (1 ~ g / k g )
0.7+ 0.5 (-0.5+ 0.5)
5.4+ 4.7 (1.2+ 1.1)
7.7+ 6.5 (1.7+ 0.9)
8.4+ 6.1 (1.8+ 0.8)
7.3+ 5.8 (1.7+ 0.9)
CLONIDINE (2 ~ g / k g )
0.5+0.1 (-0.7 + 0.2)
0.9+0.9 (-0.4 + 0.7)
4.8+5.3 (0.9 + 1.3)
6.3+5.7 (1.3 + 1.2)
5.3+5.1 (1.1 + 1.2)
PLACEBO
0.6+ 0.7 (-0.7 + 0.5)
0.9+ 1.5 (-0.6 + 0.8)
1.1 + 2.2 (-0.5 + 0.9)
1.2+ 2.6 (-0.5 + 0.9)
1.0+ 1.7 (-0.5 + 0.8)
Log-transformed values are in parentheses.
1000
-
LU
Z
800 -
.
-
-
rs = 0.62 p< 0.004 n = 20
6oo )
e-~
Z 0 '"
rr" i T (.9
~ ~
'
400
200
DA w
0
v
v
O
I
200
OI
400
I
I
I
I
600
800
1000
1200
GH-RESPONSE TO GRF (AUC-ng-min/ml) FIG. 1: Growth hormone response to intravenous administration of clonldlne (2 pg/kg), GRF (1 ~g/kg), and placebo.
256
M.E. TANCERet al.
to GRF and clonidine did not differ at either +45 min (Wilcoxon test: s=372.5, z= -1.23,p= NS) or + 60 min (Wilcoxon test: s = 369.5, z = -1.30, p = NS). Although the peak GH responses to GRF and clonidine did not differ [GRF =8.7+ 6.7 ng/ml (range 0.4-26.9 ng/ml); clonidine =6.5+_5.9 ng/ml (range 0.0-23.7 ng/ml); Wilcoxon test: s= 361.00, z= -1.31, p = NS], the GH response to GRF, as measured by GHAuo was significantly greater than the GH response to clonidine [GRF = 40.8 + 291.9 n g / m i n / m l (range 1 3 . 5 - 1 1 7 4 . 5 n g / m i n / m l ) ; c l o n i d i n e = 194.4 _+ 203.9 n g / m i n / m l (range 0 - 8 5 1 . 3 ng/min/ml); Wilcoxon test: s = 483.0, z= 1.96, p reflecting the more rapid GH-response to GRF (Fig. 1). One of the subjects had an elevated baseline GH level (> 3 ng/ml) on the clonidine study day and was excluded from the data analysis. One control had a robust 10.8 ng/ml GH response to placebo, a 9.1 n g / m l G H response to clonidine, and an even larger GH response to GRF (22.4 ng/ml). Five of the 21 subjects (24%) had no GH response to either provocative challenge, and three other subjects had a significant (> 5 ng/ml) response to one of the agents but not to the other. The female controls showed a trend towards having lower GH responses to clonidine compared to the men [ w o m e n = 7 8 . 3 + 9 5 . 7 n g / m i n / m l (range 1 . 5 - 2 5 5 . 4 n g / m i n / m l ) ; men=256.9_+ 221.8 ng/min/ml (range 0-851.3 ng/min/ml); Wilcoxon test: s= 49.00, z= 1.90, p
1990
0306-4.530/90 $3.00 + 0.0(; P~gmnon Press plc
Printed in Great Britain
GROWTH HORMONE (GH) RESPONSE TO CLONIDINE A N D GROWTH HORMONE RELEASING FACTOR (GRF) IN NORMAL CONTROLS MANUEL E. TANCER, MURRAY B. STEIN a n d THOMAS W. UHDE Section on Anxiety and Affective Disorders, Biological Psychiatry Branch, National Institute of Mental Health, Bethesda, Maryland, U.S.A. (Received 4 August 1989; infinalform 17 May 1990)
SUMMARY To investigate the relationship between the plasma growth hormone (GH) response to provocative challenge with the hypothalamic peptide growth hormone-releasing factor (GRF) and the o~z-adrenergic agonist clonidine, we administered GRF (1 I.tg/kg), clonidine (2 p.g/kg), and placebo to 21 healthy normal controls (13 men and eight women). Both elonidine and GRF caused significant increases in plasma GH levels over baseline. The peak GH-responses to GRF and elonidine were similar (GRF = 8.7+ 6.7 ng/ml; elonidine = 6.5+5.9 ng/ml; W'tleoxon test: s = 361, z ffi-1.31, p = NS). The GH responses to GRF and elonidine were significantly correlated (rs = 0.62, n = 20, p = 0.004). Unexpectedly, we found that five of the 21 (26%) normal controls had no GH secretory response to either GRF or elonidine. There was a modest gender effect with elonidine (men > women; p < 0.06) and a negative correlation between GH secretion and age with both GRF and elonidine. Neither GRF nor elonidine had an effect on eortisol levels (DRUG × TIME interaction: F(8,152)•0.60, p = NS). These findings are consistent with animal studies suggesting that the GH response to clonidine is mediated by GRF. The age and gender effects underscore the importance of careful matching for these factors in studies measuring the GH secretory response.
INTRODUCTION THE GROWTH HORMONE (GH) response to neuroendocrine challenge paradigms has been widely used as a trait marker for psychiatric disorders. "Blunted" or subnormal GH responses to the a2adrenergic agonist clonidine has been consistently reported in patients with major depression (Matussek et al., 1980; Chamey et al., 1982; Siever et al., 1984) and in four o f five reports to date in patients with panic disorder (Uhde et al., 1986; Chamey & Heninger, 1986; Nutt, 1989; Curtis et al., 1989; but see Schittecatte et al., 1988). Furthermore, normal GH responses to clonidine have been reported in "reactive" depression (Matussek et al., 1980), anorexia nervosa (Brambilla et al., 1987), an social phobia (Tancer & Uhde, 1989). The subnormal G H responses to clonidine are generally interpreted as indirect evidence of post-synaptic adrenergic receptor down-regulation (Uhde, 1988). The identification and synthesis o f growth h o r m o n e releasing factor (GRF) by Guillemin et al. (1982) and Rivier et al. (1982) provided a specific tool to study the dynamic regulation of GH-release. Studies to date with G R F in psychiatric patients have yielded mixed results, Address correspondence and reprint requests to: Dr. Thomas W. Uhde, Chief, Section on Anxiety and Affeetive Disorders, Biological Psychiatry Branch, National Institute of Mental Health, 9000 Rockville Pike, Building 10, Room 3S-239, Bethesda MD 20892, USA. 253
254
M.E. TANCERet al.
with " b l u n t e d " and " e x a g g e r a t e d " G H responses reported (Krishnan et al., 1988; Lesch e t a / , , 1988; Rapaport et al., 1989; Brambilla et al., 1989). The relationship between the G H response to clonidine and to G R F is not entirely understood. There is s o m e e v i d e n c e in rats that the G H response to clonidine is, in fact, m e d i a t e d by GRF (Miki et al., 1984; Celia et al., 1987). A positive correlation between the G H responses to clonidine and G R F across individuals would be consistent with such a mechanism. Such a positive correlation has been reported b y Brambilla et al. (1989) in a c o m b i n e d sample o f women controls and patients with anorexia nervosa and by Lesch et al. (1988) in depressed patients and controls (14 w o m e n and six men). In order to further explore the relationship between the G H responses to clonidine and to GRF, we administered clonidine (2 ~tg/kg), G R F (1 ~tg/kg), and placebo to 21 healthy volunteers in a randomized, double-blind fashion. Because o f discrepancies in the literature regarding the effect of clonidine on cortisol levels (Lal et al., 1975; C h a m e y et al., 1982; Stein & Uhde, 1988; Tancer & Uhde, 1989) and a single report o f the effects o f G R F on cortisol levels ( T h o m e r et al., 1983), we also report the p l a s m a cortisol responses to GRF, clonidine, and placebo in the subjects. SUBJECTS A N D METHODS Subjects The subjects were 21 normal voluntee.rs [13 men and eight women; mean age 28.2+5.9 years (range 20-44 yr)], who had no current or past history of psychopathology as determined by a structured interview modified from the Schedule for Affective Disorders and Schizophrenia (SADS,LA) (Mannuzza et al., 1985). All had negative family histories of depressioni drug or alcohol abuse, anxiety disorders, and schizophrenia within :first degree relatives. The participants were in good physical health as d e t ~ by history, physical e x ~ t i o u ~ electrocardiogram, and laboratory studies (SMAC-20, CBC, urinalysis, and thyroid indices) and were medication-free for at least one month prior to participation: The mean weight of the subjects was 70.2+ 12.4 kg (range 48:100 kg); their mean Quetelet index was 24.24"3.9 kg/m 2 (range 17.7-33.0). The women, who were all pre- menopausal, were studied during the first i0 days of their menstrual cycles. All subjects gave oral and written informed consent prior to entering the study. Methods All subjects were on a low monoamine, low caffeine (< 100 mg/day) diet for at least 72 hr prior to each of the three study days. Following an overnight fast, participants came to the outpatient clinic at 0900h. At that time~ with the subjects in a recumbent position, an intravenous catheter was inserted in the non-dominant antecubital vein and was kept patent with a slow infusion of normal saline. One hour following the intravenous catheter insertion, subjects received, in a double,blind, randomized fashion, clunidine (2 ~tg/kg), GRF (1 gg/kg) (Bacbem GRF 1-44 amide), or placebo over 1 rain. Each control received all three infusions, with at least 48 hr ~tween study days. Blood samples were obtained via a three-way stopcock at -15, 0, +15, +30, +45, and +60 min. GH and C o , sol were measured by radioimmunoassay (Hazelton Laboratory, Vienna, Virginia). The GH assay has a sensitivity ol 0.5 ng/ml and inter- and intra- assay variabilities of 7.3% and 4.9%, respectively. The cortisol assay has inter- and intra- assay variabilities of 12.3% and 5.5%, respectively. Statistical Methods As the G H responses to clonidine and GRF were not normally distributed,they were log-transformed. Both:the actual and log-transformed GH values are presented, Analysis of variance (ANOVA) with repeated measures was used to analyze the log-transfonne,d OH responses and the cortisol responses across TIME (hasoline vs. posy infusion levels) and DRUG (clonidine~ GRF, or placebo) conditions. When the ANOVA indicated significant main effects of TIME or DRUG conditions, Wilcoxon paired tests were used to compare variables. The -15 and 0 time points were averaged and reported as baseline levels. The area under the G H response cmwe from 0 to 60 rain (AUC) was calculatedby the trapezoidal approximation. It should be noted that A U C responses correlated highly with peak responses (rs= 0.96, p0.10 are reported as not significant (NS). Means ate reported + one standard deviation (SD). RESULTS A s expected, clonidine and GRF, but not placebo, administration resulted in significant rises in plasma G H levels c o m p a r e d to baseline (Table I, Fig. 1). The ANOVA revealed significant main effects for D R U G (clonidine vs. G R F vs. placebo) (F = 44.5, d f = 2 , 3 8 , p < 0 . 0 0 0 1 ) and for TIME (baseline vs. subsequent time points) (F =60.5, d f = 4,76, p < 0 . 0 0 0 1 ) , and a significant D R U G × T I M E interaction (F = 2 4 . 3 , d f = 8 , 1 5 2 , p < 0 . 0 0 0 1 ) . Post-hoc testing r e v e a l e d significant differences in the time course of the G H response to clonidine and GRF, with a greater G H response following G R F at +15 min (Wilcoxon test: s = 2 6 6 . 5 , z = -4.02, p < 0 . 0 0 1 ) and a trend for a greater GH-response to G R F at +30 min (Wilcoxon text: s = 345.0, z = - 1 . 9 4 , p < 0 . 0 6 ) . The G H responses
TABLE I. GROWTH HORMONE RESPONSES (NG/ML) TO PROVOCATIVE CHALLENGE Baseline
+15 min
+30 m i n
+45 m i n
+60 rain
GRF (1 ~ g / k g )
0.7+ 0.5 (-0.5+ 0.5)
5.4+ 4.7 (1.2+ 1.1)
7.7+ 6.5 (1.7+ 0.9)
8.4+ 6.1 (1.8+ 0.8)
7.3+ 5.8 (1.7+ 0.9)
CLONIDINE (2 ~ g / k g )
0.5+0.1 (-0.7 + 0.2)
0.9+0.9 (-0.4 + 0.7)
4.8+5.3 (0.9 + 1.3)
6.3+5.7 (1.3 + 1.2)
5.3+5.1 (1.1 + 1.2)
PLACEBO
0.6+ 0.7 (-0.7 + 0.5)
0.9+ 1.5 (-0.6 + 0.8)
1.1 + 2.2 (-0.5 + 0.9)
1.2+ 2.6 (-0.5 + 0.9)
1.0+ 1.7 (-0.5 + 0.8)
Log-transformed values are in parentheses.
1000
-
LU
Z
800 -
.
-
-
rs = 0.62 p< 0.004 n = 20
6oo )
e-~
Z 0 '"
rr" i T (.9
~ ~
'
400
200
DA w
0
v
v
O
I
200
OI
400
I
I
I
I
600
800
1000
1200
GH-RESPONSE TO GRF (AUC-ng-min/ml) FIG. 1: Growth hormone response to intravenous administration of clonldlne (2 pg/kg), GRF (1 ~g/kg), and placebo.
256
M.E. TANCERet al.
to GRF and clonidine did not differ at either +45 min (Wilcoxon test: s=372.5, z= -1.23,p= NS) or + 60 min (Wilcoxon test: s = 369.5, z = -1.30, p = NS). Although the peak GH responses to GRF and clonidine did not differ [GRF =8.7+ 6.7 ng/ml (range 0.4-26.9 ng/ml); clonidine =6.5+_5.9 ng/ml (range 0.0-23.7 ng/ml); Wilcoxon test: s= 361.00, z= -1.31, p = NS], the GH response to GRF, as measured by GHAuo was significantly greater than the GH response to clonidine [GRF = 40.8 + 291.9 n g / m i n / m l (range 1 3 . 5 - 1 1 7 4 . 5 n g / m i n / m l ) ; c l o n i d i n e = 194.4 _+ 203.9 n g / m i n / m l (range 0 - 8 5 1 . 3 ng/min/ml); Wilcoxon test: s = 483.0, z= 1.96, p reflecting the more rapid GH-response to GRF (Fig. 1). One of the subjects had an elevated baseline GH level (> 3 ng/ml) on the clonidine study day and was excluded from the data analysis. One control had a robust 10.8 ng/ml GH response to placebo, a 9.1 n g / m l G H response to clonidine, and an even larger GH response to GRF (22.4 ng/ml). Five of the 21 subjects (24%) had no GH response to either provocative challenge, and three other subjects had a significant (> 5 ng/ml) response to one of the agents but not to the other. The female controls showed a trend towards having lower GH responses to clonidine compared to the men [ w o m e n = 7 8 . 3 + 9 5 . 7 n g / m i n / m l (range 1 . 5 - 2 5 5 . 4 n g / m i n / m l ) ; men=256.9_+ 221.8 ng/min/ml (range 0-851.3 ng/min/ml); Wilcoxon test: s= 49.00, z= 1.90, p
