213
Journal of Affeciive Disorders, 21 (1991) 213-218 0 1991 Elsevier Science Publishers B.V. 0165-0327/91/$03.50 ADONIS 0165032791000762
JAD 00791
Blunted growth hormone and prolactin responses to L-tryptophan in depression; a state-dependent abnormality A.K. Upadhyaya,
I. Pennell,
P.J. Cowen ’ and J.F.W. Deakin
Department of Psychiatry, Manchester Royal Infirmary Manchester MI3 9 WL and ’ University Department of Psychiatry and MRC Clinical Pharmacology Unit, Littlemore Hospital, Oxford, U, K. (Received 25 June 1990) (Revision received 11 December 1990) (Accepted 20 December 1990)
Summary We have investigated whether attenuated growth hormone (GH) and prolactin (PRL) responses to L-tryptophan in depression return to normal with clinical recovery. Ten patients who had received intravenous infusions of L-tryptophan (100 mg/kg) when depressed were retested at least 3 months after full recovery and cessation of treatment. In recovered depressives growth hormone responses showed considerable recovery, in all but three cases to within a few units of their healthy age- and sex-matched controls. Prolactin responses increased with clinical recovery in all six male subjects. Results in females were inconclusive because of the effect of weight loss on prolactin responses. The results suggest that GH and PRL responses to tryptophan are state-dependent abnormalities rather than indicators of predisposition to depression. This allows the possibility that impaired functioning in systems with a SHTlA or 5HTlD receptor link may be part of the causal chain in depression.
Key words: Growth hormone; Prolactin;
Response to L-tryptophan;
Introduction Intravenous infusion of the serotonin (5-hydroxytryptamine, 5HT) precursor t_-tryptophan (LTP) induces a dose-dependent increase in circulating prolactin (PRL) and growth hormone (GH) concentrations in healthy volunteers (CharAddress for correspondence: Prof. J.F.W. Deakin, Department of Psychiatry, Manchester Royal Infirmary, Manchester Ml3 9WL, U.K.
Depression
ney et al., 1982). These responses are consistently attenuated in patients with depression (Heninger et al., 1984; Cowen and Charig, 1987; Deakin and Pennell, 1986; Deakin et al., 1990; Meltzer and Nash, 1988). There is evidence that PRL responses to LTP infusion involve 5HT systems. In normal subjects PRL responses to LTP infusion were enhanced by pretreatment with the 5HT reuptake blocker clomipramine and abolished by pretreatment with the mixed 5HT1/2 antagonist metergoline (An-
214
demon and Cowen, 1986; McCance et al., 1987). Since PRL responses to LTP are not blocked by the selective 5HT2-1C antagonist ritanserin (Charig et al., 1987) these responses may be mediated by 5HTlA or SHTlD receptors. However, attempts to block the PRL response using the SHTlA/lB antagonist properties of the p-adrenoceptor antagonists propranolol and pindolol have been equivocal (Upadhyaya et al., 1990; Smith and Cowen, in preparation). While the GH response to LTP was enhanced by clomipramine it was unaffected by metergoline (Anderson and Cowen, 1986; McCance et al., 1987). A clear antagonism was seen with pindolol (Smith and Cowen, in preparation) but not with propranolol (Upadhyaya et al., 1990). Clearly, the 5HT receptor mechanisms involved in the GH and PRL response to LTP remain to be elucidated but there is evidence compatible with the involvement of 5HT systems. Attenuated PRL and GH responses in depressives suggest that depression involves reduced 5HT function. This is consistent with reports of reduced CSF SHIAA concentration in depression and other evidence for 5HT deficiency theories of depression (reviewed in Deakin, 1989; Cowen, 1988). It has been suggested that impaired 5HT function is a trait which increases vulnerability to TABLE
depression and other maladaptive behaviours (Asberg et al., 1976; Willner, 1989; Linnoila et al., 1983). This would predict that attenuated PRL and GH responses in depressives should persist after recovery. In contrast, there is evidence that impaired 5HT function could cause the state of depression. Neuropharmacological studies in animals suggest that antidepressants may work by reversing impaired 5HTl neurotransmission (De Montigny and Aghajanian, 1984). In normal humans and in depressives, some antidepressants enhance PRL and GH responses to LTP (Chamey et al., 1985; Price et al., 1989; Cowen et al., 1986, 1990). According to this hypothesis, attenuated hormonal responses to LTP may be associated with the state of depression and should therefore normalise with recovery. To understand the role of 5HT in the causation of depression, it is clearly important to determine whether attenuated PRL and GH responses to LTP recover with clinical recovery. Material and methods Ten patients with major depressive disorder according to the criteria of DSM-III were studied in two centres: University Department of Psychiatry, Withington Hospital, Manchester and the
1
PATIENT
CHARACTERISTICS
Patient
Weight
age
loss
AND
PRL AND
HAMD Depressed
Recovered
GH RESPONSES
Peak-base
PRL (mu/l)
Depressed
Recovered
TO INTRAVENOUS
TRYPTOPHAN
INFUSION
Peak - base GH (mu/l) Controls
Depressed
Recovered
Controls
Au controls
Females 38
Yes
_
8
640
576
370
43
Yes
27
0
833
143
157
25
No
-
_
274
137
920
0”
33.2
35
25
48
No
19
0
207
1824
139
0
12
29.2
54
29
Yes
24
7
292
564
277
11
36
30
29
32
No
18
6
247
365
870
0
38
No
18
5
589
1081
562
41
No
24
4
30
133
41
No
30
1
60
436
43
No
27
0
86 (261)
Oa 27
3
34.6
38
14
4.8
41
Males
(39.5) Individual
subject data; median
a Raised basal GH.
values in parentheses.
3.6
7.1
30
3.1
12.5
10.5
37
323
5
35.6
38
36
90
6
25
35.9
38
106
202
0
14.6
18.7
40
(400)
(300)
(01.5)
(14.3)
(29.6)
(37.5)
215
University Department of Psychiatry and MRC Clinical Pharmacology Unit, Littlemore Hospital, Oxford. LTP infusions were administered to the patients when depressed and after clinical recovery, the interval between two tests varying from 6 to 40 months. They were free from any physical illness and had not taken psychotropic drugs for at least 3 months before the tests except for small doses of benzodiazepines (equivalent to less than 15 mg of diazepam/day). All patients had been free of depressive symptoms for at least 3 months. Severity of depression was rated on the 17-item Hamilton depression rating scale on both occasions. Signifi-
Prolactin
(mUA)
cant weight loss was defined as more than 3 kg, a Hamilton weight loss rating of 2. None of the controls or the recovered depressives had significant weight loss. The clinical characteristics of the patients and the details of the age/sex-matched controls are shown in Table 1. After an overnight fast, an intravenous cannula was inserted into a forearm vein. After 30-60 min a blood sample was taken for estimation of the basal serum GH and PRL. LTP (10 g/l in 0.72% saline, 0.05% sodium bisulphite vehicle) was infused at a dose of 100 mg/kg over 30 min. Blood samples were obtained for estimation of PRL and GH at 30, 40, 50, 60, 70 and 90 min after the
Growth hormone (mUA)
(1824) 1200
1000
800
600
Fig. 1. Prolactin and growth hormone responses to L-tryptophan infusion in patients while depressed and after recovery with matched controls. Solid lines indicate individual patients and dotted lines the individuals’ matched controls.
compared
216
beginning of the infusion. The subjects remained supine throughout the test and they were not allowed to sleep. Serum GH and PRL were determined in coded samples, by radioimmunoassay as described previously (Cowen and Charig, 1987; Deakin et al., 1990). Hormonal responses were measured as peak increase over baseline. Individual results are presented without statistical tests of significance since the changes are clear as are the limitations of small sample size, effects of weight loss and gender. Results
Depressed patients clearly had lower GH responses to LTP than controls. Several depressives showed absent responses but in two cases this may have been due to their raised pre-infusion GH concentration (16 and 23 mu/l). Two of the three patients with weight loss had the greatest GH responses to LTP (Table 1). GH responses showed unequivocal increases with recovery from depression in all but one patient with an extreme GH response during the depressed phase associated with severe weight loss. Eight of the 10 recovered patients had responses in the range of the matched controls and close to their paired control. PRL responses to LTP were greater in females than males and in those with significant weight loss (Goodwin et al., 1987; Deakin et al., 1990). In the present study, only two depressed females did not have weight loss and one showed a marked increase in the PRL response with recovery (2071824 mu/l) while the other did not (274-137 mu/l). In four of the six depressed males, PRL responses were reduced compared to their matched controls. In every male patient, PRL responses increased with recovery (Fig. 1) producing a median PRL response slightly greater than the matched control median (Table 1). Discussion
It is clear that attenuated GH responses to LTP in depression normalise with recovery and so reflect the depressed state rather than a trait underlying vulnerability to depression.
The PRL results are complicated by the effects of recent weight loss in females (Goodwin et al., 1987; Cowen and Charig, 1987; Deakin et al., 1990) which may account for the exaggerated responses in two of the four depressed females. Only one of the two females without weight loss showed an increased PRL response with recovery. However, recovery of PRL responses clearly occurred with clinical recovery in the male depressives. It therefore seems probable that attenuated PRL and GH responses to LTP reflect the depressive state since they recover with clinical recovery, Chamey et al. (1982) reported that 4 weeks’ treatment with amitriptyline or imipramine increased PRL responses to LTP infusions in depressives, though the increase did not correlate with clinical improvement. Similar findings have been reported with tranylcypromine and fluvoxamine treatment (Price et al., 1985, 1989). Interpretation of this increase in PRL response is confounded by the simultaneous effect of drug treatment and change in clinical state. In the present study, however, the patients had been well and free of medication for many weeks. Siever et al. (1986) reported that attenuated PRL responses to the 5HT-releasing agent fenfluramine were as frequent in a group of recovered depressives as in a separate group of patients who were depressed at the time of testing. This result does not seem decisive evidence for a trait marker since the same patients were not followed from depression to recovery. 5HT deficiency theories of depression predict the observed blunting of hormonal responses to LTP and the hypothesis survives this test. However, GH and PRL secretion are influenced by a number of interacting neurotransmitters in addition to SHT, abnormalities of which could result in attenuation of responses to LTP. Some altemative explanations have been tested but the results are not conclusive. Reduced GH responses could indicate reduced responsiveness to GHRH. However, attenuation of GH response to LTP is reproducible and marked whereas normal as well as blunted GH responses to GHRH have been reported (Eriksson et al., 1988; Lesch et al., 1987; Thomas et al., 1989; Krishnan et al., 1988). Similarly, studies of GH responses to apomorphine have not con-
217
sistently revealed evidence of impaired dopamine receptor function in depression (Corn et al., 1984). Perhaps the most reproducible neuroendocrine abnormality in depression is the blunted GH response to clonidine challenge (Checkley et al., 1984). This may indicate that diminished (Ye function occurs in depression. However, there is evidence in a small number of patients that this abnormality does not normalise with recovery (Mitchell et al., 1988). In contrast, the GH response to LTP appears to recover in the present study and this suggests different mechanisms are involved in the attenuation of GH responses to clonidine and LTP. Pituitary secretion of PRL in response to metoclopramide and TRH is reported to be normal in depressives (Meites and Sonntag, 1981). The PRL response to morphine was reduced in depressives in one study (Extein et al., 1980) but this could be due to involvement of SHT in morphine’s actions. It can be seen that where the possible non-SHT mechanisms for impaired hormonal responses to LTP have been tested, they have not proved convincing. Impaired 5HT function in depression remains a viable explanation of attenuated hormonal responses to LTP in depression. Normalisation of hormonal responses to LTP with clinical recovery from depression raises the possibility that impaired 5HT function could be closely related to causation of depression. The ability of antidepressants to enhance 5HT function, as suggested for example by electrophysiological studies of 5HT synapses (De Montigny and Aghajanian, 1984) might then cause recovery from depression by reversing the impairment of 5HT function. This is corroborated by the finding that reducing tryptophan availability by a dietary manipulation exacerbated depressed mood in patients being treated with antidepressants (Delgado et al., 1989). The present findings suggest that impaired 5HT function may be a state-dependent abnormality rather than a long-term vulnerability marker and this begs the question of its cause. Deakin et al. (1990) pointed out the evidence that hypercortisolaemia may reduce 5HTl receptor function (Kennett et al., 1985; Bagdy et al., 1989) and this includes the PRL response to 5HT drug challenge
in rats (Bagdy et al., 1989) and possibly in humans (Deakin et al., 1990). Adverse life events, lack of social support and other psychosocial factors could be associated with depression because they cause hypersecretion of cortisol which impairs 5HTl receptor function. References Anderson, I.M. and Cowen, P.J. (1986) Clomipramine enhances prolactin and growth hormone responses to L-tryptophan. Psychopharmacology 89, 131-133. Asberg, M., Cowen, P.J., Traskman, L., Bertilsson, L. and Ringerger, V. (1976) ‘Serotonin depression’, a biochemical subgroup within the affective disorders. Science 191, 478480. Bagdy, G., Calogero, A.E.. Aulakh, C.S., Szemeredi, K. and Murphy, D.L. (1989) Long-term cortisol treatment impairs behavioural and nemoendocrine responses to 5HTl agonists in the rat. Neuroendocrinology 50, 241-247. Charig, E.M., Anderson, I.M., Robinson, J.M., Nutt, D.J. and Cowen, P.J. (1987) L-Tryptophan and prolactin release: evidence for interaction between 5-HTl and 5-HT2 receptors. Hum. Psychopharmacol. 1, 93-97. Charney, D.S., Heninger, G.R., Renhard, J.F., Stemberg, D.E. and Hafstead, K.M. (1982) Effect of intravenous L-tryptophan on prolactin and growth hormone and mood in healthy subjects. Psychopharmacology 77, 217-222. Checkley, S.A., Glass, I.B.. Thompson, C.. Corn, T. and Robinson, P. (1984) The GH response to clonidine in endogenous as compared with reactive depression, Psychol. Med. 14, 7733777. Corn, T.H., Hale, AS., Thompson, C., Bridges, P.K. and Checkley, S.A. (1984) A comparison of the growth hormone responses to clonidine and apomorphine in the same patients with endogenous depression. Br. J. Psychiatry 144, 636-639. Cowen, P.J. (1988) Recent views on the role of 5.hydroxytryptamine in depression. Curr. Opinion Psychiatry 1, 5659. Cowen, P.J. and Charig, E.M. (1987) Neuroendocrine responses to tryptophan in major depression. Arch. Gen. Psychiatry 44, 958-966. Cowen, P.J., Geaney, D.P., Schachter, M., Green, A.R. and Elliott, J.M. (1986) Desipramine treatment in normal subjects: effects on neuroendocrine responses to tryptophan and platelet serotonin (5-HT)-related receptors. Arch. Gen. Psychiatry 43, 61-67. Cowen, P.J., McCance, S.L., Gelder, M.G. and Grahame-Smith, D.G. (1990) The effect of amitriptyline on endocrine responses to intravenous t.-tryptophan. Psychiatr. Res. 31, 201-208. De Montigy, C. and Aghajanian, G.K. (1984) Tricyclic antidepressants: long-term treatment increases responsivity of rat forebrain neurones to serotonin. Science 202, 1303-1306. Deakin, J.F.W. (1989) Role of 5HT receptor subtypes in depression. In: T. Archer, P. Bevan and A. Cools (Eds.),
218 Behavioural Pharmacology of 5HT. Lawrence Erlbaum, New York, NY, pp. 179-204. Deakin, J.F.W. and Pennell, I. (1986) 5HT receptor subtypes in depression. Psychopharmacology 89, S24. Deakin, J.F.W., Pennell, I., Upadhyaya, A.J. and Lofthouse, R. (1990) A neuroendocrine study of 5HT function in depression: evidence for biological mechanisms of endogenous and psychosocial causation. Psychopharmacology 101, 8592. Delgado, P.L., Charney, D.S., Price, L.H., Landis, H. and Heninger, G.R. (1989) Neuroendocrine and behavioural effects of dietary tryptophan restriction in healthy subjects. Life Sci. (in press). Eriksson, E., Balldin, I., Lindstedt, G. and Mod&h, K. (1988) Growth hormone response to the alpha-adrenoceptor agonist guanfacine and to growth hormone releasing hormone in depressed patients and controls. Psychiatr. Res. 26, 59-67. Extein, I., Pottash, A.L.C., Gold, M.J., Sweeney, D.R.. Martin, D.M. and Goodwin, F.K. (1980) Deficient prolactin response to morphine in depressed patients. Am. J. Psychiatry 137, 845-846. Goodwin, G.M., Fairburn, CC. and Cowen, P.J. (1987) The effects of dieting and weight loss on neuroendocrine responses to L-tryptophan, clonidine and apomorphine in volunteers: important implications for neuroendocrine investigation in depression. Arch. Gen. Psychiatry 44, 952955. Heninger. G.R., Charney, D.S. and Sternberger, D.E. (1984) Serotonergic function in depression. Arch. Gen. Psychiatry 41, 398-402. Kennett, G.A., Dickinson, S.L. and Cruzon, G. (1985) Central serotonergic responses and behavioural adaptation to repeated immobilisation: the effect of the corticosterone synthesis inhibitor metyrapone. Eur. J. Pharmacol. 119, 1433274. Krishnan, K.R.R., Manepalli, A.N., Ritchie, J.C., Rayasam. K., Melville, M.L., Daughtry, G., Throner, M.O., Rivier, J.E., Vale, W.W., Nemeroff, C.B. and Carroll, B.J. (1988) Growth hormone-releasing factor stimulation test in depression Am. J. Psychiatry 145, 90-92. Lesch. K.P., Laux, G., Erb, A., Pfuller, H. and Beckmann, H. (1987) Growth hormone (GH) responses to GH releasing hormone in depression: correlation with GH release following clonidine. Psychiatr. Res. 25, 301-310.
Linnoila, M., Virkkunen, M., Scheinin, M., Muutila, A., Rimon. R. and Goodwin, F.K. (1983) Low cerebrospinal fluid 5-hydroxyindolacetic acid concentration differentiates impulsive from nonimpulsive violent behaviour. Life Sci. 33, 2609-2614. McCance, S.L., Cowen, P.J., Wailer, H. and Grahame-Smith, D.G. (1987) The effect of metergoline on endocrine responses to t_-tryptophan. J. Psychopharmacol. 2, 90-94. Meites, J. and Sonntag, W.E. (1981) Hypothalamic, hypophysiotropic hormones and neurotransmitter regulation: current views. Annu. Rev. Pharmacol. Toxicol. 21, 295-322. Meltzer, H.Y. and Nash, G.F. (1988) Serotonin and mood: nemoendocrine aspects. In: D. Gantan and D. Pfaff (Eds.), Current Topics in Neuroendccrinology, Vol. 8. Springer Verlag, Berlin, pp. 183-210. Mitchell, P.B., Beam, J.H.A., Corn, T.H. and Checkley, S.A. (1988) Growth hormone response to clonidine after recovery in patients with endogenous depression. Br. J. Psychiatry 152, 34-38. Price, L.H., Chamey. D.S. and Heninger, G.R. (1985) Effects of tranylcypromine treatment on neuroendocrine behavioural and autonomic responses to tryptophan in depressed patients. Life Sci. 37, 8099818. Price, L.H., Charney, D.S.. Delgado, P.L., Anderson, G.M. and Heninger, G.R. (1989) Effects of desipramine and fluvoxamine treatment on the prolactin response to t.-tryptophan. Arch. Gen. Psychiatry 46, 6255631. Siever, L.J., Coccaro, E.F., Benjamin, E., Rubinstem, K. and Davis, K.L. (1986) Adrenergic and serotonergic receptor responsiveness in depression in antidepressants and receptor function. In: Ciba Foundation Symposium 123. Wiley, Chichester, pp. 148-163. Thomas, R.. Beer, R., Harris, B.. John, R. and Scanlon, M. (1989) GH responses to growth hormone releasing factor in depression. J. Affect. Disord. 16, 1333137. Upadhyaya, A.K., Deakin, J.F.W. and Pennell, I. (1990) Hormonal response to L-tryptophan infusion: effect of propranolol. Psychoneuroendocrinology 15, 3099312. Wilier, P. (1989) Towards a theory of serotonergic dysfunction in depression. In: P. Bevan. A.R. Cooles and T. Archer (Eds.), Behavioural Pharmacology of 5-HT. Laurence Erlbaum, New York, NY, pp. 157-179.
Journal of Affeciive Disorders, 21 (1991) 213-218 0 1991 Elsevier Science Publishers B.V. 0165-0327/91/$03.50 ADONIS 0165032791000762
JAD 00791
Blunted growth hormone and prolactin responses to L-tryptophan in depression; a state-dependent abnormality A.K. Upadhyaya,
I. Pennell,
P.J. Cowen ’ and J.F.W. Deakin
Department of Psychiatry, Manchester Royal Infirmary Manchester MI3 9 WL and ’ University Department of Psychiatry and MRC Clinical Pharmacology Unit, Littlemore Hospital, Oxford, U, K. (Received 25 June 1990) (Revision received 11 December 1990) (Accepted 20 December 1990)
Summary We have investigated whether attenuated growth hormone (GH) and prolactin (PRL) responses to L-tryptophan in depression return to normal with clinical recovery. Ten patients who had received intravenous infusions of L-tryptophan (100 mg/kg) when depressed were retested at least 3 months after full recovery and cessation of treatment. In recovered depressives growth hormone responses showed considerable recovery, in all but three cases to within a few units of their healthy age- and sex-matched controls. Prolactin responses increased with clinical recovery in all six male subjects. Results in females were inconclusive because of the effect of weight loss on prolactin responses. The results suggest that GH and PRL responses to tryptophan are state-dependent abnormalities rather than indicators of predisposition to depression. This allows the possibility that impaired functioning in systems with a SHTlA or 5HTlD receptor link may be part of the causal chain in depression.
Key words: Growth hormone; Prolactin;
Response to L-tryptophan;
Introduction Intravenous infusion of the serotonin (5-hydroxytryptamine, 5HT) precursor t_-tryptophan (LTP) induces a dose-dependent increase in circulating prolactin (PRL) and growth hormone (GH) concentrations in healthy volunteers (CharAddress for correspondence: Prof. J.F.W. Deakin, Department of Psychiatry, Manchester Royal Infirmary, Manchester Ml3 9WL, U.K.
Depression
ney et al., 1982). These responses are consistently attenuated in patients with depression (Heninger et al., 1984; Cowen and Charig, 1987; Deakin and Pennell, 1986; Deakin et al., 1990; Meltzer and Nash, 1988). There is evidence that PRL responses to LTP infusion involve 5HT systems. In normal subjects PRL responses to LTP infusion were enhanced by pretreatment with the 5HT reuptake blocker clomipramine and abolished by pretreatment with the mixed 5HT1/2 antagonist metergoline (An-
214
demon and Cowen, 1986; McCance et al., 1987). Since PRL responses to LTP are not blocked by the selective 5HT2-1C antagonist ritanserin (Charig et al., 1987) these responses may be mediated by 5HTlA or SHTlD receptors. However, attempts to block the PRL response using the SHTlA/lB antagonist properties of the p-adrenoceptor antagonists propranolol and pindolol have been equivocal (Upadhyaya et al., 1990; Smith and Cowen, in preparation). While the GH response to LTP was enhanced by clomipramine it was unaffected by metergoline (Anderson and Cowen, 1986; McCance et al., 1987). A clear antagonism was seen with pindolol (Smith and Cowen, in preparation) but not with propranolol (Upadhyaya et al., 1990). Clearly, the 5HT receptor mechanisms involved in the GH and PRL response to LTP remain to be elucidated but there is evidence compatible with the involvement of 5HT systems. Attenuated PRL and GH responses in depressives suggest that depression involves reduced 5HT function. This is consistent with reports of reduced CSF SHIAA concentration in depression and other evidence for 5HT deficiency theories of depression (reviewed in Deakin, 1989; Cowen, 1988). It has been suggested that impaired 5HT function is a trait which increases vulnerability to TABLE
depression and other maladaptive behaviours (Asberg et al., 1976; Willner, 1989; Linnoila et al., 1983). This would predict that attenuated PRL and GH responses in depressives should persist after recovery. In contrast, there is evidence that impaired 5HT function could cause the state of depression. Neuropharmacological studies in animals suggest that antidepressants may work by reversing impaired 5HTl neurotransmission (De Montigny and Aghajanian, 1984). In normal humans and in depressives, some antidepressants enhance PRL and GH responses to LTP (Chamey et al., 1985; Price et al., 1989; Cowen et al., 1986, 1990). According to this hypothesis, attenuated hormonal responses to LTP may be associated with the state of depression and should therefore normalise with recovery. To understand the role of 5HT in the causation of depression, it is clearly important to determine whether attenuated PRL and GH responses to LTP recover with clinical recovery. Material and methods Ten patients with major depressive disorder according to the criteria of DSM-III were studied in two centres: University Department of Psychiatry, Withington Hospital, Manchester and the
1
PATIENT
CHARACTERISTICS
Patient
Weight
age
loss
AND
PRL AND
HAMD Depressed
Recovered
GH RESPONSES
Peak-base
PRL (mu/l)
Depressed
Recovered
TO INTRAVENOUS
TRYPTOPHAN
INFUSION
Peak - base GH (mu/l) Controls
Depressed
Recovered
Controls
Au controls
Females 38
Yes
_
8
640
576
370
43
Yes
27
0
833
143
157
25
No
-
_
274
137
920
0”
33.2
35
25
48
No
19
0
207
1824
139
0
12
29.2
54
29
Yes
24
7
292
564
277
11
36
30
29
32
No
18
6
247
365
870
0
38
No
18
5
589
1081
562
41
No
24
4
30
133
41
No
30
1
60
436
43
No
27
0
86 (261)
Oa 27
3
34.6
38
14
4.8
41
Males
(39.5) Individual
subject data; median
a Raised basal GH.
values in parentheses.
3.6
7.1
30
3.1
12.5
10.5
37
323
5
35.6
38
36
90
6
25
35.9
38
106
202
0
14.6
18.7
40
(400)
(300)
(01.5)
(14.3)
(29.6)
(37.5)
215
University Department of Psychiatry and MRC Clinical Pharmacology Unit, Littlemore Hospital, Oxford. LTP infusions were administered to the patients when depressed and after clinical recovery, the interval between two tests varying from 6 to 40 months. They were free from any physical illness and had not taken psychotropic drugs for at least 3 months before the tests except for small doses of benzodiazepines (equivalent to less than 15 mg of diazepam/day). All patients had been free of depressive symptoms for at least 3 months. Severity of depression was rated on the 17-item Hamilton depression rating scale on both occasions. Signifi-
Prolactin
(mUA)
cant weight loss was defined as more than 3 kg, a Hamilton weight loss rating of 2. None of the controls or the recovered depressives had significant weight loss. The clinical characteristics of the patients and the details of the age/sex-matched controls are shown in Table 1. After an overnight fast, an intravenous cannula was inserted into a forearm vein. After 30-60 min a blood sample was taken for estimation of the basal serum GH and PRL. LTP (10 g/l in 0.72% saline, 0.05% sodium bisulphite vehicle) was infused at a dose of 100 mg/kg over 30 min. Blood samples were obtained for estimation of PRL and GH at 30, 40, 50, 60, 70 and 90 min after the
Growth hormone (mUA)
(1824) 1200
1000
800
600
Fig. 1. Prolactin and growth hormone responses to L-tryptophan infusion in patients while depressed and after recovery with matched controls. Solid lines indicate individual patients and dotted lines the individuals’ matched controls.
compared
216
beginning of the infusion. The subjects remained supine throughout the test and they were not allowed to sleep. Serum GH and PRL were determined in coded samples, by radioimmunoassay as described previously (Cowen and Charig, 1987; Deakin et al., 1990). Hormonal responses were measured as peak increase over baseline. Individual results are presented without statistical tests of significance since the changes are clear as are the limitations of small sample size, effects of weight loss and gender. Results
Depressed patients clearly had lower GH responses to LTP than controls. Several depressives showed absent responses but in two cases this may have been due to their raised pre-infusion GH concentration (16 and 23 mu/l). Two of the three patients with weight loss had the greatest GH responses to LTP (Table 1). GH responses showed unequivocal increases with recovery from depression in all but one patient with an extreme GH response during the depressed phase associated with severe weight loss. Eight of the 10 recovered patients had responses in the range of the matched controls and close to their paired control. PRL responses to LTP were greater in females than males and in those with significant weight loss (Goodwin et al., 1987; Deakin et al., 1990). In the present study, only two depressed females did not have weight loss and one showed a marked increase in the PRL response with recovery (2071824 mu/l) while the other did not (274-137 mu/l). In four of the six depressed males, PRL responses were reduced compared to their matched controls. In every male patient, PRL responses increased with recovery (Fig. 1) producing a median PRL response slightly greater than the matched control median (Table 1). Discussion
It is clear that attenuated GH responses to LTP in depression normalise with recovery and so reflect the depressed state rather than a trait underlying vulnerability to depression.
The PRL results are complicated by the effects of recent weight loss in females (Goodwin et al., 1987; Cowen and Charig, 1987; Deakin et al., 1990) which may account for the exaggerated responses in two of the four depressed females. Only one of the two females without weight loss showed an increased PRL response with recovery. However, recovery of PRL responses clearly occurred with clinical recovery in the male depressives. It therefore seems probable that attenuated PRL and GH responses to LTP reflect the depressive state since they recover with clinical recovery, Chamey et al. (1982) reported that 4 weeks’ treatment with amitriptyline or imipramine increased PRL responses to LTP infusions in depressives, though the increase did not correlate with clinical improvement. Similar findings have been reported with tranylcypromine and fluvoxamine treatment (Price et al., 1985, 1989). Interpretation of this increase in PRL response is confounded by the simultaneous effect of drug treatment and change in clinical state. In the present study, however, the patients had been well and free of medication for many weeks. Siever et al. (1986) reported that attenuated PRL responses to the 5HT-releasing agent fenfluramine were as frequent in a group of recovered depressives as in a separate group of patients who were depressed at the time of testing. This result does not seem decisive evidence for a trait marker since the same patients were not followed from depression to recovery. 5HT deficiency theories of depression predict the observed blunting of hormonal responses to LTP and the hypothesis survives this test. However, GH and PRL secretion are influenced by a number of interacting neurotransmitters in addition to SHT, abnormalities of which could result in attenuation of responses to LTP. Some altemative explanations have been tested but the results are not conclusive. Reduced GH responses could indicate reduced responsiveness to GHRH. However, attenuation of GH response to LTP is reproducible and marked whereas normal as well as blunted GH responses to GHRH have been reported (Eriksson et al., 1988; Lesch et al., 1987; Thomas et al., 1989; Krishnan et al., 1988). Similarly, studies of GH responses to apomorphine have not con-
217
sistently revealed evidence of impaired dopamine receptor function in depression (Corn et al., 1984). Perhaps the most reproducible neuroendocrine abnormality in depression is the blunted GH response to clonidine challenge (Checkley et al., 1984). This may indicate that diminished (Ye function occurs in depression. However, there is evidence in a small number of patients that this abnormality does not normalise with recovery (Mitchell et al., 1988). In contrast, the GH response to LTP appears to recover in the present study and this suggests different mechanisms are involved in the attenuation of GH responses to clonidine and LTP. Pituitary secretion of PRL in response to metoclopramide and TRH is reported to be normal in depressives (Meites and Sonntag, 1981). The PRL response to morphine was reduced in depressives in one study (Extein et al., 1980) but this could be due to involvement of SHT in morphine’s actions. It can be seen that where the possible non-SHT mechanisms for impaired hormonal responses to LTP have been tested, they have not proved convincing. Impaired 5HT function in depression remains a viable explanation of attenuated hormonal responses to LTP in depression. Normalisation of hormonal responses to LTP with clinical recovery from depression raises the possibility that impaired 5HT function could be closely related to causation of depression. The ability of antidepressants to enhance 5HT function, as suggested for example by electrophysiological studies of 5HT synapses (De Montigny and Aghajanian, 1984) might then cause recovery from depression by reversing the impairment of 5HT function. This is corroborated by the finding that reducing tryptophan availability by a dietary manipulation exacerbated depressed mood in patients being treated with antidepressants (Delgado et al., 1989). The present findings suggest that impaired 5HT function may be a state-dependent abnormality rather than a long-term vulnerability marker and this begs the question of its cause. Deakin et al. (1990) pointed out the evidence that hypercortisolaemia may reduce 5HTl receptor function (Kennett et al., 1985; Bagdy et al., 1989) and this includes the PRL response to 5HT drug challenge
in rats (Bagdy et al., 1989) and possibly in humans (Deakin et al., 1990). Adverse life events, lack of social support and other psychosocial factors could be associated with depression because they cause hypersecretion of cortisol which impairs 5HTl receptor function. References Anderson, I.M. and Cowen, P.J. (1986) Clomipramine enhances prolactin and growth hormone responses to L-tryptophan. Psychopharmacology 89, 131-133. Asberg, M., Cowen, P.J., Traskman, L., Bertilsson, L. and Ringerger, V. (1976) ‘Serotonin depression’, a biochemical subgroup within the affective disorders. Science 191, 478480. Bagdy, G., Calogero, A.E.. Aulakh, C.S., Szemeredi, K. and Murphy, D.L. (1989) Long-term cortisol treatment impairs behavioural and nemoendocrine responses to 5HTl agonists in the rat. Neuroendocrinology 50, 241-247. Charig, E.M., Anderson, I.M., Robinson, J.M., Nutt, D.J. and Cowen, P.J. (1987) L-Tryptophan and prolactin release: evidence for interaction between 5-HTl and 5-HT2 receptors. Hum. Psychopharmacol. 1, 93-97. Charney, D.S., Heninger, G.R., Renhard, J.F., Stemberg, D.E. and Hafstead, K.M. (1982) Effect of intravenous L-tryptophan on prolactin and growth hormone and mood in healthy subjects. Psychopharmacology 77, 217-222. Checkley, S.A., Glass, I.B.. Thompson, C.. Corn, T. and Robinson, P. (1984) The GH response to clonidine in endogenous as compared with reactive depression, Psychol. Med. 14, 7733777. Corn, T.H., Hale, AS., Thompson, C., Bridges, P.K. and Checkley, S.A. (1984) A comparison of the growth hormone responses to clonidine and apomorphine in the same patients with endogenous depression. Br. J. Psychiatry 144, 636-639. Cowen, P.J. (1988) Recent views on the role of 5.hydroxytryptamine in depression. Curr. Opinion Psychiatry 1, 5659. Cowen, P.J. and Charig, E.M. (1987) Neuroendocrine responses to tryptophan in major depression. Arch. Gen. Psychiatry 44, 958-966. Cowen, P.J., Geaney, D.P., Schachter, M., Green, A.R. and Elliott, J.M. (1986) Desipramine treatment in normal subjects: effects on neuroendocrine responses to tryptophan and platelet serotonin (5-HT)-related receptors. Arch. Gen. Psychiatry 43, 61-67. Cowen, P.J., McCance, S.L., Gelder, M.G. and Grahame-Smith, D.G. (1990) The effect of amitriptyline on endocrine responses to intravenous t.-tryptophan. Psychiatr. Res. 31, 201-208. De Montigy, C. and Aghajanian, G.K. (1984) Tricyclic antidepressants: long-term treatment increases responsivity of rat forebrain neurones to serotonin. Science 202, 1303-1306. Deakin, J.F.W. (1989) Role of 5HT receptor subtypes in depression. In: T. Archer, P. Bevan and A. Cools (Eds.),
218 Behavioural Pharmacology of 5HT. Lawrence Erlbaum, New York, NY, pp. 179-204. Deakin, J.F.W. and Pennell, I. (1986) 5HT receptor subtypes in depression. Psychopharmacology 89, S24. Deakin, J.F.W., Pennell, I., Upadhyaya, A.J. and Lofthouse, R. (1990) A neuroendocrine study of 5HT function in depression: evidence for biological mechanisms of endogenous and psychosocial causation. Psychopharmacology 101, 8592. Delgado, P.L., Charney, D.S., Price, L.H., Landis, H. and Heninger, G.R. (1989) Neuroendocrine and behavioural effects of dietary tryptophan restriction in healthy subjects. Life Sci. (in press). Eriksson, E., Balldin, I., Lindstedt, G. and Mod&h, K. (1988) Growth hormone response to the alpha-adrenoceptor agonist guanfacine and to growth hormone releasing hormone in depressed patients and controls. Psychiatr. Res. 26, 59-67. Extein, I., Pottash, A.L.C., Gold, M.J., Sweeney, D.R.. Martin, D.M. and Goodwin, F.K. (1980) Deficient prolactin response to morphine in depressed patients. Am. J. Psychiatry 137, 845-846. Goodwin, G.M., Fairburn, CC. and Cowen, P.J. (1987) The effects of dieting and weight loss on neuroendocrine responses to L-tryptophan, clonidine and apomorphine in volunteers: important implications for neuroendocrine investigation in depression. Arch. Gen. Psychiatry 44, 952955. Heninger. G.R., Charney, D.S. and Sternberger, D.E. (1984) Serotonergic function in depression. Arch. Gen. Psychiatry 41, 398-402. Kennett, G.A., Dickinson, S.L. and Cruzon, G. (1985) Central serotonergic responses and behavioural adaptation to repeated immobilisation: the effect of the corticosterone synthesis inhibitor metyrapone. Eur. J. Pharmacol. 119, 1433274. Krishnan, K.R.R., Manepalli, A.N., Ritchie, J.C., Rayasam. K., Melville, M.L., Daughtry, G., Throner, M.O., Rivier, J.E., Vale, W.W., Nemeroff, C.B. and Carroll, B.J. (1988) Growth hormone-releasing factor stimulation test in depression Am. J. Psychiatry 145, 90-92. Lesch. K.P., Laux, G., Erb, A., Pfuller, H. and Beckmann, H. (1987) Growth hormone (GH) responses to GH releasing hormone in depression: correlation with GH release following clonidine. Psychiatr. Res. 25, 301-310.
Linnoila, M., Virkkunen, M., Scheinin, M., Muutila, A., Rimon. R. and Goodwin, F.K. (1983) Low cerebrospinal fluid 5-hydroxyindolacetic acid concentration differentiates impulsive from nonimpulsive violent behaviour. Life Sci. 33, 2609-2614. McCance, S.L., Cowen, P.J., Wailer, H. and Grahame-Smith, D.G. (1987) The effect of metergoline on endocrine responses to t_-tryptophan. J. Psychopharmacol. 2, 90-94. Meites, J. and Sonntag, W.E. (1981) Hypothalamic, hypophysiotropic hormones and neurotransmitter regulation: current views. Annu. Rev. Pharmacol. Toxicol. 21, 295-322. Meltzer, H.Y. and Nash, G.F. (1988) Serotonin and mood: nemoendocrine aspects. In: D. Gantan and D. Pfaff (Eds.), Current Topics in Neuroendccrinology, Vol. 8. Springer Verlag, Berlin, pp. 183-210. Mitchell, P.B., Beam, J.H.A., Corn, T.H. and Checkley, S.A. (1988) Growth hormone response to clonidine after recovery in patients with endogenous depression. Br. J. Psychiatry 152, 34-38. Price, L.H., Chamey. D.S. and Heninger, G.R. (1985) Effects of tranylcypromine treatment on neuroendocrine behavioural and autonomic responses to tryptophan in depressed patients. Life Sci. 37, 8099818. Price, L.H., Charney, D.S.. Delgado, P.L., Anderson, G.M. and Heninger, G.R. (1989) Effects of desipramine and fluvoxamine treatment on the prolactin response to t.-tryptophan. Arch. Gen. Psychiatry 46, 6255631. Siever, L.J., Coccaro, E.F., Benjamin, E., Rubinstem, K. and Davis, K.L. (1986) Adrenergic and serotonergic receptor responsiveness in depression in antidepressants and receptor function. In: Ciba Foundation Symposium 123. Wiley, Chichester, pp. 148-163. Thomas, R.. Beer, R., Harris, B.. John, R. and Scanlon, M. (1989) GH responses to growth hormone releasing factor in depression. J. Affect. Disord. 16, 1333137. Upadhyaya, A.K., Deakin, J.F.W. and Pennell, I. (1990) Hormonal response to L-tryptophan infusion: effect of propranolol. Psychoneuroendocrinology 15, 3099312. Wilier, P. (1989) Towards a theory of serotonergic dysfunction in depression. In: P. Bevan. A.R. Cooles and T. Archer (Eds.), Behavioural Pharmacology of 5-HT. Laurence Erlbaum, New York, NY, pp. 157-179.
