- Email: [email protected]
A revised interpretation of the TRH test results in female depressed patients
215
Journal of Affective Disorders, 16 (1989) 215-221 Elsevier
JAD 0613B
A revised interpretation of the TRH test results in female depressed patients Part II: Prolactin responses. Relationships with sex hormones, corticosteroid state, age, ,monoamines and amino acid levels Michael Maes *, Ma rits Vandewoude 2, Leo Maes 3, Chris Schotte * and Paul Cosyns * 1 Departments
of {
2 Internal Medicine, Unive[siiy Hospital of Anhverpen, Wilrijkstraar IO, 2510 Edegem, Belgium L.abot’atory for Analytical Chemistry, CTL, Voskeslaan 270, Ghent, Belgium
*\
(Received 27 April 1988) (Revision received 4 August 1988) (Accepted 7 September 1988)
Prolactin (PRL) levels were recorded in baseline conditions and 20 and 60 min after thyrotropin releasing hormone (TRH) administration (200 pg i.v.) in 60 depressed females categorized according to DSM-III. Peak PRL responses were significantly (r = 0.727, P -z0.001) correlated with their baseline levels. Consequently, the PRL responses to TRH were largely predicted by baseline PRL levels. It was suggested that the PRL responses to TRH consisted of two parts. The first component was a relative exaggeration of basal PRL, reflecting the basal activity of the hormone. The second component was the residual response. This part was estimated by partialling out the relative effects of basal PRL on peak PRL responses by means of regression analysis. Basal PRL and residual PRL responses were uninformative for major depression. Post-menopausal females showed significantly reduced basal PRL levels. There was a significant negative correlation between basal PRL and follicle stimulating hormone levels, age and post-dexamethasone cortisol values. The residual PRL responses were negatively correlated with free triiodothyronine levels and positively with serotonergic variables, i.e., Shydroxyindoleacetic acid in 24-h urine and the ratio L-tryptophan/competing amino acids.
Key words: Depression; TRH test; Prolactin; Hypothalamic-pituitary-gonadal
Address for correspondence: Dr. M. Maes, Department of Psychiatry, University Hospital of Antwerpen, Wihijkstraat 10, 2510 Edegem, Belgium. 0165-0327/89/$03.50
axis
Introduction
In normal subjects the prolactin (PRL) release is largely stimulated by thyrotropin releasing
0 1989 Elsevier Science Publishers B.V. (Biomedical Division)
216
hormone (TRH) administration (200-400 pg i.v.) (Bowers et al., 1971; Jacobs et al., 1971; Rabello et al., 1974). The PRL responses to TRH in major depressive patients have been reported to be reduced as compared to normal controls (Ehrensing et al., 1974; GrCgoire et al., 1977; Linnoila et al., 1979; Linkowski et al., 1980; Winokur et al., 1983; Witschy et al., 1984; Garbutt et al., 1986). Garbutt et al. (1986) suggested that the reductions in PRL responses observed in major depressives could be secondary to lower baseline PRL levels. Some authors found significantly reduced basal PRL levels in major depressives (Linkowski et al., 1980; Garbutt et al., 1986). Nevertheless, the effect of baseline PRL levels on the interpretation of PRL responses has not yet been examined. Furthermore, negative findings with respect to PRL responses to TRH (Maeda et al., 1975; Brambilla et al., 1978; Naeye et al., 1978; Coppen et al., 1980; Langer et al., 1980; Kirkegaard et al., 1981; Targum et al., 1982; Zis et al., 1986) and to baseline PRL levels (Maeda et al., 1975; Witschy et al., 1984) have been reported. PRL data in previous work have rarely been presented in relation to parameters putatively influencing those data, e.g., age, sex hormonal (Rutlin et al., 1977; Agren and Wide, 1986), thyroid (Carlson et al., 1978), corticosteroid (Copinschi et al., 1975) and nutritional state. In addition measures of PRL inhibiting factors (PIF), e.g., dopamine (Delitala et al., 1987), and of PRL releasing factors (PRF), e.g., serotonin (Delitala et al., 1987), have not usually been presented in the same paper. An association between blunted thyrotropin secreting hormone (TSH) levels and reduced PRL responses to TRH was found in depressed patients (Asnis et al., 1981; Kirkegaard et al., 1981; Targum et al., 1982; Witschy. et al., 1984). Subsequently it was concluded that a common factor may influence the PRL and the TSH responsivity of TRH. The present study was undertaken to investigate (1) the effects of basal PRL levels on the interpretation of the PRL responses to TRH, (2) basal PRL and PRL responses in major versus minor depressive females and (3) the relationships between PRL data and age, sex hormonal, thyroid, corticosteroid and nutritional state, dopamine, serotonergic variables and TSH data.
Methods Patients, methods and statistics are described in Part I of this study (Maes et al., 1989). The same patients were investigated for the determination of PRL before and 20 and 60 min after TRH administration. Prolactin was determined using a RIA liquid phase (Prolactin RIA kit, Amersham); inter-assay coefficients of variation (CV) were 5.9% for the low (z = 63.3 ~IU/ml, n = 20), 3.0% for the medium @ = 452 yIU/ml, n = 20) and 2.6% for the high (Iz = 1121 @U/ml, n = 20) ranges. Amino acid determinations were carried out on the same fasting serum samples as taken for basal TSH and PRL. Amino acid determinations were done by a method proposed by Tumell and Cooper (1982) which was slightly modified. Determinations were done by liquid chromatography, precolumn derivatization with orthopdialdehyde and fluorometric detection at h,,, = 330 nm, X,, = 418 (Schoeffel FS970 Fluorimeter). The interassay coefficients of variation (CV) were 8.6% for L-tryptophan (K = 76.6, n = 5) 8.6% for valine (E = 209.4, n = 5) 8.4% for tyrosine (X = 65.6, n = 5), 9.5% for phenylalanine (X = 81.9, n = 5), 10.2% for isoleucine (X = 73.4, n = 5) and 7.9% for leucine (z = 159.6, n = 5) (all in 10e6 mol/l). The ratio between L-tryptophan and the sum of the five other competing amino acids was calculated (LTRP/CAA). 5-Hydroxyindoleacetic acid (5-H&4) was determined in the same sample of 24-h urine as used for the determinations of noradrenaline and dopamine (Maes et al., 1989). The CV values were 11.3% for the low (j? = 1.4, n = 5) and 12.6% for the medium (z = 3.3, n = 5) ranges. Results Baseline PRL, PRL 20 and 60 min after TRH were log-normally distributed. Therefore, the PRL data were processed in their logarithmic transformation. The other variables were assessed in their appropriate transformations as explained in Maes et al. (1989). The results of the measurements of the PRL variables in the several DSM-III subgroups are listed in Table 1. TRH administration yielded a significant increase in PRL after both 20 and 60 min. The PRL responses 20 min after TRH were significantly greater than after 60 min, both
217 TABLE
1
MEASUREMENTS
OF PRL AND THE RESPONSES
DSM-III category Minor depression (3C0.40,309.00) Major depression without melancholia (296.22,296.32) Major depression with melancholia and/or with psychotic features (296.23,296.24, 296.33, 296.34)
OF PRL TO TRH
IN 60 DEPRESSED PRL,
FEMALES PRL,
Ratio
Index
n
PRL,
PRL20
md
19
371.5 (f412.4)
1396.5
( f 779.2) a 641.1 (k439.2)
b 4.999 (f 2.006)
+0.023
(kO.325)
MD-M
26
204.4 (+ 123.3)
1038.7
(+ 898.9) a 443.4 (zt417.5)
b 5.535 (f 3.214)
-0.079
(kO.582)
MD + M
15
250.8 (+ 167.3)
1373.5
( f 1060.9) a 575.1 ( f 425.4) b 6.015 ( f 3.224)
+ 0.108 ( f 0.464)
PRL,, basal prolactin (PRL) levels; PRL2a resp. PRL,, PRL levels 20 resp. 60 min after administration of TRH; Ratio, ratio PRL,,/PRL,; PRL,, residual PRL response. All results are listed as mean ( f 1 SD), the PRL data are given in pIU/ml. a Significantly different from PRL, and PRL,,. b Significantly different from PRLa, all repeated ANOVA on logarithmic transformations (F = 70.13, df = 2/122, P < 10e4).
in the study group as a whole and in every individual patient. There were no significant differences in baseline PRL, PRL 20 and 60 min after TRH nor in the ratio peak/basal PRL between the DSM-III categories. Table 2 gives the intercorrelation matrix between basal PRL, the PRL responses to TRH, the ratio peak/basal PRL, free thyroid hormones (FT,, FT,) and age. Baseline PRL was significantly correlated with the PRL responses to TRH. Fifty-two percent of the variance in the PRL levels after TRH administration was explained by the regression on basal PRL. Consequently, PRL responses to TRH were largely
TABLE
2
INTERCORRELATION
MATRIX
pRLB PRL20
PRL60 Ratio PRL, n; FT,
Age
predicted by baseline levels. The PRL responses to TRH thus were built up by two components. The first part was the predictable component, i.e., a relative exaggeration of basal PRL levels. This part was estimated using the equation: Y = 0.47 + 0.86 X ln(basa1 PRL levels). The second part was the orthogonal component residue Y (res Y), computed by using the formula res Y = In Y - Y, in which 1nY represents the natural logarithm of the actual PRL response (criterion variable) and Y the estimated value of the criterion variable determined by the regression analysis. Consequently, the two components of clinical relevance which
0.727 0.698 -0.400 0.000 0.176 - 0.004 -0.387
*** *** **
**
BETWEEN
PRL VARIABLES,
FT,, FT, AND AGE IN 60 DEPRESSED
p=20
PWO
Ratio
PRL,
FT,
0.946 * * * 0.391 * * 0.690 *** - 0.067 - 0.076 - 0.203
0.356 ** 0.645 *** -0.042 - 0.063 -0.146
0.902 *** - 0.330 * * - 0.096 0.260 *
-0.295 * - 0.091 0.206
0.603 * * * -0.156
PRL,, basal PRL; PRL, resp. PRL,, PRL levels 20 resp. 60 min after TRH administration; residual PRL responses; FT,, free triiodothyronine; FT4, free thyroxine. Listed are Pearson’s correlation coefficients. PRL,, PRL2,, PRLsc,, FT, and age are assessed * P -e 0.05, * * P < 0.01, * * * P < 0.001.
Ratio,
FEMALES FT,
- 0.032
ratio PRL,/PRLa;
in their logarithmic
PRL,,
transformation. ..
218
constitute the PRL responses are (1) the predictable component or the basal PRL levels (which are a more ready measure of this component) and (2) the residual PRL response (res Y). The residual PRL responses to TRH (ANCOVA with the actual PRL responses as dependent and basal PRL as independent variables) were not significantly different among the DSM-III groups. These residuals are listed in Table 1. Neither basal PRL nor the residual PRL responses showed significant correlations with the Hamilton Depression Rating Scale score. Basal PRL levels were significantly and negatively correlated with the post-dexamethasone cortisol values (T = - 0.352, P < 0.05) and with age (Table 2). Post-menopausal females showed significantly reduced basal PRL levels as compared to pre-menopausal females (F = 11.94, df = l/59, P < 0.001). Basal PRL was significantly negatively correlated with FSH (rs = - 0.267, P < 0.05). Age and FSH together explained 16.2% of the variance in basal PRL levels (F = 5.53, df = 2/59, P = 0.007). The residual PRL responses were significantly correlated with FT, (Table 2), the ratio L-TRP/CAA (r = 0.336, P < 0.05) and the adjusted 5-HIAA excretion (after controlling by means of multiple regression for 24-h urinary output and creatinine flow). We found no significant correlations either between basal PRL and TSH levels or between residual PRL and TSH responses. The PRL data were not correlated with dopamine contents in 24-h urine and the nutritional parameters, or with the drug state (use of benzodiazepines the day prior to blood sampling or the length of the drug-free period before admission into hospital). Consequently, major effects of the drug state on our results could be disregarded. Discussion Interrelations sponses
between
basal
PRL
and PRL
re-
The results of our study showed that the PRL responses to TRH administration were directly related to the basal PRL levels. Up to 52.9% of the variance in the PRL responses was explained by the regression on basal PRL. Consequently, the PRL responses were largely predicted by the basal
PRL levels. We have shown that the PRL responses to TRH consisted of two components. The first component was a relative exaggeration of the basal PRL levels. The second component was the residual PRL response. This part was estimated by partialling out the relative effects of basal PRL on PRL responses to TRH by means of regression analysis. In clinical practice two factors seems to be relevant for the interpretation of the PRL data: (1) the basal PRL levels reflecting the baseline activity setpoint of the hormone, and (2) the residual PRL values, reflecting the latent capacity of the lactotrope cells in the pituitary to respond to overwhelming amounts of exogenous TRH. It is of interest that the peak/basal PRL ratio, used by several authors as an index for the responsivity of PRL to TRH, shared 81.4% of the variance with the residual PRL responses. Consequently, this ratio offers a relatively good and readily conceptualized measure of the PRL responsivity to exogenous TRH. Baseline PRL levels
The data reported in our paper indicate that the basal PRL levels were no different between major and minor depressives and that they were not related to the severity of illness. These findings agree with those of Maeda et al. (1975) and Witschy et al. (1984). Basal PRL levels were significantly reduced in post-menopausal women and accordingly a significant negative correlation was found between basal PRL and the levels of follicle stimulating hormone (FSH) and age. A reciprocal relationship between PRL and the hypothalamic-pituitarygonadal (HPG) axis could be explained by two factors: (1) a potent PIF was found to be coded within the prohormone for gonadotropin releasing hormone (GnRH) (Nikolics et al., 1985); (2) PRL passes the blood-brain barrier with ease (Belchetz et al., 1982) and PRL inhibits GnRH secretion (Martin and Reichlin, 1987). We have found that basal PRL levels were significantly and negatively correlated with postdexamethasone cortisol levels. Administration of exogenous steroids can affect the PRL responses to various stimuli including TRH (Copinschi et al., 1975; Sowers et al., 1977; Rossier et al., 1980). A plausible explanation for the observed correla-
219
tion is that corticosteroid overdrive, for which a disturbed dexamethasone suppression test is an index (Yerevanian et al., 1983) inhibits PRL secretion (Rothschild et al., 1984) through increased central dopamine activity (Langlais et al., 1984). Although dopamine is a potent PIF (Martin and Reichlin, 1987) we did not find the expected negative correlation with basal PRL. Also other authors (Neill, 1980) have shown that an exact inverse relationship between dopamine and PRL levels was lacking.
could best be interpreted using two factors: firstly basal PRL levels and secondly the residual PRL levels or the peak/basal PRL ratio. (2) Basal PRL and PRL responses to TRH were uninformative for major depression. (3) Basal PRL levels were linked to the HPG axis function, age and postdexamethasone cortisol values. The residual PRL responses were related to the thyroid state and to measures of serotonergic activity.
Residual PRL. responses
We are indebted to Professor Dr. Blockx P., M.D., Ph.D. (nuclear medicine), Professor W. De Potter, Ph.D. (neurochemistry) and to Professor Dr. S. Scharpe, M.D., Ph.D. (pharmacology) who provided laboratory support. The authors express their appreciation to Mrs. C. Maes-Steyaert for secretarial support.
We found that the responsivity of PRL to TRH administration related neither to the diagnostic DSM-III groups nor to the severity of illness. Our findings are thus in agreement with the studies mentioned in our introduction, reporting negative findings on PRL responses to TRH. We found that the residual PRL responses to TRH were significantly and negatively correlated with free triiodothyronine levels. This agrees with the findings that thyroid hormones can inhibit PRL release (Carlson et al., 1978; Garbutt et al., 1986). We have found significant and positive relationships between the PRL responses and serotonergic variables, i.e., the ratio L-tryptophan/competing amino acids (L-TRP/ CAA) and 5hydroxyindoleacetic acid (5-HIAA). The LTRP/CAA ratio is an index of the availability of L-TRP to the brain (Curzon and Sama, 1984; Femstrom, 1984). The brain serotonin synthesis largely depends on the availability of L-TRP (Moir and Eccleston, 1968). 5-HIAA is the major metabolite of serotonin. Our results accord with the established knowledge that plasma PRL responses to TRH are in part mediated through serotonergic mechanisms (Coppen et al., 1980). We could not find a significant relationship between the TSH and PRL responses to TRH administration. This agrees with the results of Loosen et al. (1983) and Zis et al. (1986). Other authors (Frey and Hang, 1977; Geras et al., 1982) also concluded that the mechanisms of TSH release after TRH were different from TRH-induced PRL release. Conclusions
(1) In clinical practice PRL responses to TRH
Acknowiedgements
References .&en, H. and Wide, L. (1986) TRH tests with analyses of TSH, prolactin and GH responses in subtypes of patients with major depressive disorders. Acta Psychiatr. Stand. 73, 549-558. Asnis, G.M., Sachar, E.J., Halbreich, U., Swami Nathan, R., Ostrow, L., Solomon, M. and HaIpem, F.S. (1981) Endocrine responses to thyrotropin-releasing hormone in major depressive disorders. Psychiatr. Res. 5, 205-215. Belchetz, P.E., Ridley, R.M. and Baker, H.F. (1982) Studies on the accessibility of prolactin and growth hormone to brain: effect of agonists on hormone levels in serial, simultaneous plasma and cerebrospinal fluid samples in the rhesus monkey. Brain Res. 239, 310-314. Bowers, C.Y., Friesen, H.G., Hwang, P., Guyda, H.J. and Folkers, K. (1971) Prolactin and thyrotropin release in man by synthetic pyroglutamyl-histidyl-prolinamide. B&hem. Biophys. Res. Commun. 45, 1033-1041. Brambilla, F., Smeraldi, E., Sacchetti, E., Negri, F., Cocchi, D. and Miiller, E.E. (1978) Deranged anterior pituitary responsiveness to hypothalamic hormones in depressed patients. Arch. Gen. Psychiatry 35, 1231-1238. Carlson, H.E., Sawin, C.I.,, Krugman, L.G., Meyer, N.C. and Hershman, J.M. (1978) Effect of thyroid hormones on the prolactin response to thyrotropin-releasing hormone in normal persons and euthyroid goitrous patients. J. Clin. Endocrinol. Metab. 47, 275-279. Copinschi, G., L’Hermite, N., Leclercq, R., Goldstein, J., Vanhaelst, L., Virasoro, E. and Rofyn, C. (1975) Effects of glucocorticoids on pituitary hormonal response to hypoglycemia. Inhibition of prolactin release. J. Clin. Endocrinol. Metab. 40, 442-449. Coppen, A., Rama Rao, V.A., Bishop, M., Abou-Saleh, M.T.
220 and Wood, K. (1980) Neuroendocrine studies in affective disorders: part 1: plasma prolactin response to thyrotropin-releasing hormone in affective disorders: effect of ECT. J. Affect. Disord. 2, 311-315. Cutzon, G. and Sama, G.S. (1984) Tryptophan transport to the brain: newer findings and older ones reconsidered. In: H.G. Schlossberger, W. Kochen, B. Linzen and H. Steinhart (Eds.), Progress in Tryptophan and Serotonin Research. Walter de Gruyter, New York, NY, pp. 145-160. Delitala, G., Tomasi, P. and Virdis, R. (1987) Prolactin, growth hormone and thyrotropin-thyroid hormone secretion during stress states in man. In: A. Grossman (Ed.), Clinical Endocrinology and Metabolism, Vol. 1, Neuroendocrinology of Stress. Ball&e Tindal, London, pp. 391-414. Ehrensing, R.H., Kastin, A.J., Schalch, D.S., Friesen, H.G., Vargas, J.R. and Schally, A. (1974) Affective state and thyrotropin and prolactin responses after repeated injections of thyrotropin-releasing hormone in depressed patients. Am. J. Psychiatry 131, 714-718. Fernstrom, J.D. (1984) Tryptophan availability and serotonin synthesis in rat brain: effects of experimental diabetes. In: H.G. Schlossberger, W. Kochen, B. Linzen and H. Steinhart (Eds.), Progress in Tryptophan and Serotonin Research. Walter de Gruyter, New York, NY, pp. 161-172. Frey, M.M. and Hang, E. (1977) Effect of prolonged oral administration of TRH on plasma levels of thyrotrophin and prolactin in normal individuals and in patients with primary hypothyroidism. Acta Endocrinol. 85, 744-752. Garbutt, J.C., Loosen, P.T., Blacharsh, K. and Prange, A.J. (1986) The prolactin response to thyrotropin-releasing hormone in depressed patients and normal subjects. Psychoneuroendocrinology 11, 213-219. Geras, E., Rebecchi, M.J. and Gershengom, M.C. (1982) Evidence that stimulation of thyrotropin and prolactin secretion by thyrotropin-releasing hormone occur via different calcium mediated mechanisms: studies with verapamil. Endocrinology 110, 901-906. Gregoire, F., Brauman, H., De Buch, R. and Corvilain, J. (1977) Hormone release in depressed patients before and after recovery. Psychoneuroendocrinology 2, 303-312. Jacobs, L.S., Snyder, P.J., Wilber, J.F., Utiger, R.D. and Daughaday, W.H. (1971) Increased serum prolactin after administration of synthetic thyrotropin releasing hormone (TRH) in man. J. Clin. Endocrinol. Metab. 33, 996-998. Kirkegaard, C., Eskildsen, P.C. and Bjerum, N. (1981) Parallel changes of the responses of thyrotropin, growth hormone and prolactin to thyrotropin-releasing hormone in endogenous depression. Psychoneurcendocrinology 6, 253-259. Langer, G., Schoenbeck, G., Koinig, G., Reiter, H., Schuessler, M., Aschauer, H. and Lesch, 0. (1980) Evidence for neuroendocrine involvement in the therapeutic effects of antidepressant drugs. In: F. Brambilla, 0. Racagni and D. De Wied (I%.), Progress in Psychoneuroendocrinology. Elsevier, New York, NY, pp. 197-208. Langlais, P.J., Rothschild, A.J., Schatzberg, A.F., Cole, J.O. and Bird, E.D. (1984) Dexamethasone elevates dopamine in human plasma and rat brain. Psychopharmacol. Bull. 20, 365-370. Linkowski, P., Brauman, H. and Mendlewicz, J. (1980) Pro-
lactin secretion in women with unipolar and bipolar depression. Psychiatr. Res. 3, 265-271. Linnoila, M., Lamberg, B.A., Rosberg, G., Karonen, S.-L. and Welin, M.S. (1979) Thyroid hormones and TSH, prolactin and LH responses to repeated TRH and LHR injections in depressed patients. Acta Psychiatr. Stand. 59, 536-544. Loosen, P.T., Kistler, K. and Prange, A.J. (1983) Use of TSH response to TRH as an independent variable. Am. J. Psychiatry 140, 700-703. Maeda, K., Kato, Y., Ohgo, S., Chihara, K., Yoshimoto, Y., Yamaguchi, N., Kuromaru, S. and Imura, H. (1975) Growth hormone and prolactin release after injection of thyrotropin-releasing hormone in patients with depression. J. Clin. Endocrinol. Metab. 40, 501-505. Maes, M., Vandewoude, M., Maes, L., Schotte, C. and Cosyns, P. (1989) A revised interpretation of the TRH test results in female depressed patients. Part I: TSH responses. Effects of severity of illness, thyroid hormones, monoamines, age, sex hormonal, corticosteroid and nutritional state. J. Affect. Disord. 16, 203-213. Martin, J.B. and Reichlin, S. (1987) Clinical Neuroendocrinology, 2nd edn. F.A. Davis Company, Philadelphia, PA. Moir, A.T.B. and Eccleston, D. (1968) The effect of precursor loading in the cerebral metabolism of 5-hydroxy-indoles. J. Neurochem. 15, 1093-1108. Naeye, R., Goldstein, J., Zegers de Beyl, D., Linkowski, P., Mendlewicz, J., Copinschi, G., Badawi, M. and Leclercq, R. (1978) Thyrotrophin, prolactin and growth hormone responses to TRH in barbiturate coma and in depression. Clin. Endocrinol. 9, 49-58. Neill, J.D. (1980) Neuroendocrine regulation of prolactine secretion. In: L. Martini and W.F. Ganong (Eds.), Frontiers in Neuroendocrinology, Vol. 6, Raven Press, New York, NY, pp. 129-155. Nikolics, K., Mason, A.J., Szonyi, E., Ramachandran, J. and Seeburg, P.H. (1985) A prolactin-inhibiting factor within the precursor for human gonadotropin-releasing hormone. Nature 316, 511-517. Rabello, M.M., Snyder, P.J. and Utiger, R.D. (1974) Effects on the pituitary-thyroid axis and prolactin secretion of single and repetitive oral doses of thyrotropin-releasing hormone (TRH). J. Clin. Endocrinol. Metab. 39, 571-578. Rossier, J., French, E., Rivier, C., Shibasaki, T., Guillemin, R. and Bloom, F.E. (1980) Stress-induced release of prolactin: blockade by dexamethasone and naloxone may indicate /I-endorphin mediation. Proc. Natl. Acad. Sci. U.S.A. 77, 666-669. Rothschild, A.J., Langlais, P.J., Schatzberg, A.F., Walsh, F.Y., Cole, J. and Bird, E.D. (1984) Dexamethasone increases plasma free dopamin in man. J. Psychiatr. Res. 18,217-223. Rutlin, E., Hang, E. and Torjesen, P.A. (1977) Serum thyrotropin, prolactin and growth hormone response to TRH during oestrogen treatment. Acta Endocrinol. 84, 23-35. Sowers, J.R., Carlson, H.E., Brautbar, N. and Her&man, J.M. (1977) Effect of dexamethasone on prolactin and TSH response to TRH and metoclopramide in man. J. Clin. Endocrinol. Metab. 44, 237-241. Targum, S.D., Sullivan, A.C. and Bymes, S.M. (1982) Com-
221 pensatory pituitary-thyroid mechanisms in major depressive disorder. Psychiatr. Res. 6, 85-96. Tumell, D. and Cooper, J. (1982) Rapid assay for amino acids in serum or urine by precohmm derivatisation and reversed-phase liquid chromatography. Clin. Chem. 28, 527-531. Winokur, A., Amsterdam, J., Oler, J., Mendels, J., Snyder, P.J., Caroff, S.N. and Brunswick, D.J. (1983) Multiple hormonal responses to protirelin (TRH) in depressed patients. Arch. Gen. Psychiatry 40, 525-531. Witschy, J.K., S&lesser, M.A., Fulton, C.L., Orsulak, P.J.,
Giles, D.E., Fairchild, C., Crowley, G. and Rush, J. (1984) TRH-induced prolactin release is blunted in females with endogenous unipolar major depression. Psychiatr. Res. 12, 321-331. Yerevanian, B.I., Woolf, P.D. and Iker, H.P. (1983) Plasma ACTH levels in depression before and after recovery: relationship to the dexamethasone suppression test. Psychiatr. Res. 10, 175-181. Zis, A.P., Albala, A.A., Haskett, R.F., Carroll, B.J. and Lohr, N.E. (1986) Prolactin response to TRH in depression. Psychiatr. Res. 20, 77-82.
Journal of Affective Disorders, 16 (1989) 215-221 Elsevier
JAD 0613B
A revised interpretation of the TRH test results in female depressed patients Part II: Prolactin responses. Relationships with sex hormones, corticosteroid state, age, ,monoamines and amino acid levels Michael Maes *, Ma rits Vandewoude 2, Leo Maes 3, Chris Schotte * and Paul Cosyns * 1 Departments
of {
2 Internal Medicine, Unive[siiy Hospital of Anhverpen, Wilrijkstraar IO, 2510 Edegem, Belgium L.abot’atory for Analytical Chemistry, CTL, Voskeslaan 270, Ghent, Belgium
*\
(Received 27 April 1988) (Revision received 4 August 1988) (Accepted 7 September 1988)
Prolactin (PRL) levels were recorded in baseline conditions and 20 and 60 min after thyrotropin releasing hormone (TRH) administration (200 pg i.v.) in 60 depressed females categorized according to DSM-III. Peak PRL responses were significantly (r = 0.727, P -z0.001) correlated with their baseline levels. Consequently, the PRL responses to TRH were largely predicted by baseline PRL levels. It was suggested that the PRL responses to TRH consisted of two parts. The first component was a relative exaggeration of basal PRL, reflecting the basal activity of the hormone. The second component was the residual response. This part was estimated by partialling out the relative effects of basal PRL on peak PRL responses by means of regression analysis. Basal PRL and residual PRL responses were uninformative for major depression. Post-menopausal females showed significantly reduced basal PRL levels. There was a significant negative correlation between basal PRL and follicle stimulating hormone levels, age and post-dexamethasone cortisol values. The residual PRL responses were negatively correlated with free triiodothyronine levels and positively with serotonergic variables, i.e., Shydroxyindoleacetic acid in 24-h urine and the ratio L-tryptophan/competing amino acids.
Key words: Depression; TRH test; Prolactin; Hypothalamic-pituitary-gonadal
Address for correspondence: Dr. M. Maes, Department of Psychiatry, University Hospital of Antwerpen, Wihijkstraat 10, 2510 Edegem, Belgium. 0165-0327/89/$03.50
axis
Introduction
In normal subjects the prolactin (PRL) release is largely stimulated by thyrotropin releasing
0 1989 Elsevier Science Publishers B.V. (Biomedical Division)
216
hormone (TRH) administration (200-400 pg i.v.) (Bowers et al., 1971; Jacobs et al., 1971; Rabello et al., 1974). The PRL responses to TRH in major depressive patients have been reported to be reduced as compared to normal controls (Ehrensing et al., 1974; GrCgoire et al., 1977; Linnoila et al., 1979; Linkowski et al., 1980; Winokur et al., 1983; Witschy et al., 1984; Garbutt et al., 1986). Garbutt et al. (1986) suggested that the reductions in PRL responses observed in major depressives could be secondary to lower baseline PRL levels. Some authors found significantly reduced basal PRL levels in major depressives (Linkowski et al., 1980; Garbutt et al., 1986). Nevertheless, the effect of baseline PRL levels on the interpretation of PRL responses has not yet been examined. Furthermore, negative findings with respect to PRL responses to TRH (Maeda et al., 1975; Brambilla et al., 1978; Naeye et al., 1978; Coppen et al., 1980; Langer et al., 1980; Kirkegaard et al., 1981; Targum et al., 1982; Zis et al., 1986) and to baseline PRL levels (Maeda et al., 1975; Witschy et al., 1984) have been reported. PRL data in previous work have rarely been presented in relation to parameters putatively influencing those data, e.g., age, sex hormonal (Rutlin et al., 1977; Agren and Wide, 1986), thyroid (Carlson et al., 1978), corticosteroid (Copinschi et al., 1975) and nutritional state. In addition measures of PRL inhibiting factors (PIF), e.g., dopamine (Delitala et al., 1987), and of PRL releasing factors (PRF), e.g., serotonin (Delitala et al., 1987), have not usually been presented in the same paper. An association between blunted thyrotropin secreting hormone (TSH) levels and reduced PRL responses to TRH was found in depressed patients (Asnis et al., 1981; Kirkegaard et al., 1981; Targum et al., 1982; Witschy. et al., 1984). Subsequently it was concluded that a common factor may influence the PRL and the TSH responsivity of TRH. The present study was undertaken to investigate (1) the effects of basal PRL levels on the interpretation of the PRL responses to TRH, (2) basal PRL and PRL responses in major versus minor depressive females and (3) the relationships between PRL data and age, sex hormonal, thyroid, corticosteroid and nutritional state, dopamine, serotonergic variables and TSH data.
Methods Patients, methods and statistics are described in Part I of this study (Maes et al., 1989). The same patients were investigated for the determination of PRL before and 20 and 60 min after TRH administration. Prolactin was determined using a RIA liquid phase (Prolactin RIA kit, Amersham); inter-assay coefficients of variation (CV) were 5.9% for the low (z = 63.3 ~IU/ml, n = 20), 3.0% for the medium @ = 452 yIU/ml, n = 20) and 2.6% for the high (Iz = 1121 @U/ml, n = 20) ranges. Amino acid determinations were carried out on the same fasting serum samples as taken for basal TSH and PRL. Amino acid determinations were done by a method proposed by Tumell and Cooper (1982) which was slightly modified. Determinations were done by liquid chromatography, precolumn derivatization with orthopdialdehyde and fluorometric detection at h,,, = 330 nm, X,, = 418 (Schoeffel FS970 Fluorimeter). The interassay coefficients of variation (CV) were 8.6% for L-tryptophan (K = 76.6, n = 5) 8.6% for valine (E = 209.4, n = 5) 8.4% for tyrosine (X = 65.6, n = 5), 9.5% for phenylalanine (X = 81.9, n = 5), 10.2% for isoleucine (X = 73.4, n = 5) and 7.9% for leucine (z = 159.6, n = 5) (all in 10e6 mol/l). The ratio between L-tryptophan and the sum of the five other competing amino acids was calculated (LTRP/CAA). 5-Hydroxyindoleacetic acid (5-H&4) was determined in the same sample of 24-h urine as used for the determinations of noradrenaline and dopamine (Maes et al., 1989). The CV values were 11.3% for the low (j? = 1.4, n = 5) and 12.6% for the medium (z = 3.3, n = 5) ranges. Results Baseline PRL, PRL 20 and 60 min after TRH were log-normally distributed. Therefore, the PRL data were processed in their logarithmic transformation. The other variables were assessed in their appropriate transformations as explained in Maes et al. (1989). The results of the measurements of the PRL variables in the several DSM-III subgroups are listed in Table 1. TRH administration yielded a significant increase in PRL after both 20 and 60 min. The PRL responses 20 min after TRH were significantly greater than after 60 min, both
217 TABLE
1
MEASUREMENTS
OF PRL AND THE RESPONSES
DSM-III category Minor depression (3C0.40,309.00) Major depression without melancholia (296.22,296.32) Major depression with melancholia and/or with psychotic features (296.23,296.24, 296.33, 296.34)
OF PRL TO TRH
IN 60 DEPRESSED PRL,
FEMALES PRL,
Ratio
Index
n
PRL,
PRL20
md
19
371.5 (f412.4)
1396.5
( f 779.2) a 641.1 (k439.2)
b 4.999 (f 2.006)
+0.023
(kO.325)
MD-M
26
204.4 (+ 123.3)
1038.7
(+ 898.9) a 443.4 (zt417.5)
b 5.535 (f 3.214)
-0.079
(kO.582)
MD + M
15
250.8 (+ 167.3)
1373.5
( f 1060.9) a 575.1 ( f 425.4) b 6.015 ( f 3.224)
+ 0.108 ( f 0.464)
PRL,, basal prolactin (PRL) levels; PRL2a resp. PRL,, PRL levels 20 resp. 60 min after administration of TRH; Ratio, ratio PRL,,/PRL,; PRL,, residual PRL response. All results are listed as mean ( f 1 SD), the PRL data are given in pIU/ml. a Significantly different from PRL, and PRL,,. b Significantly different from PRLa, all repeated ANOVA on logarithmic transformations (F = 70.13, df = 2/122, P < 10e4).
in the study group as a whole and in every individual patient. There were no significant differences in baseline PRL, PRL 20 and 60 min after TRH nor in the ratio peak/basal PRL between the DSM-III categories. Table 2 gives the intercorrelation matrix between basal PRL, the PRL responses to TRH, the ratio peak/basal PRL, free thyroid hormones (FT,, FT,) and age. Baseline PRL was significantly correlated with the PRL responses to TRH. Fifty-two percent of the variance in the PRL levels after TRH administration was explained by the regression on basal PRL. Consequently, PRL responses to TRH were largely
TABLE
2
INTERCORRELATION
MATRIX
pRLB PRL20
PRL60 Ratio PRL, n; FT,
Age
predicted by baseline levels. The PRL responses to TRH thus were built up by two components. The first part was the predictable component, i.e., a relative exaggeration of basal PRL levels. This part was estimated using the equation: Y = 0.47 + 0.86 X ln(basa1 PRL levels). The second part was the orthogonal component residue Y (res Y), computed by using the formula res Y = In Y - Y, in which 1nY represents the natural logarithm of the actual PRL response (criterion variable) and Y the estimated value of the criterion variable determined by the regression analysis. Consequently, the two components of clinical relevance which
0.727 0.698 -0.400 0.000 0.176 - 0.004 -0.387
*** *** **
**
BETWEEN
PRL VARIABLES,
FT,, FT, AND AGE IN 60 DEPRESSED
p=20
PWO
Ratio
PRL,
FT,
0.946 * * * 0.391 * * 0.690 *** - 0.067 - 0.076 - 0.203
0.356 ** 0.645 *** -0.042 - 0.063 -0.146
0.902 *** - 0.330 * * - 0.096 0.260 *
-0.295 * - 0.091 0.206
0.603 * * * -0.156
PRL,, basal PRL; PRL, resp. PRL,, PRL levels 20 resp. 60 min after TRH administration; residual PRL responses; FT,, free triiodothyronine; FT4, free thyroxine. Listed are Pearson’s correlation coefficients. PRL,, PRL2,, PRLsc,, FT, and age are assessed * P -e 0.05, * * P < 0.01, * * * P < 0.001.
Ratio,
FEMALES FT,
- 0.032
ratio PRL,/PRLa;
in their logarithmic
PRL,,
transformation. ..
218
constitute the PRL responses are (1) the predictable component or the basal PRL levels (which are a more ready measure of this component) and (2) the residual PRL response (res Y). The residual PRL responses to TRH (ANCOVA with the actual PRL responses as dependent and basal PRL as independent variables) were not significantly different among the DSM-III groups. These residuals are listed in Table 1. Neither basal PRL nor the residual PRL responses showed significant correlations with the Hamilton Depression Rating Scale score. Basal PRL levels were significantly and negatively correlated with the post-dexamethasone cortisol values (T = - 0.352, P < 0.05) and with age (Table 2). Post-menopausal females showed significantly reduced basal PRL levels as compared to pre-menopausal females (F = 11.94, df = l/59, P < 0.001). Basal PRL was significantly negatively correlated with FSH (rs = - 0.267, P < 0.05). Age and FSH together explained 16.2% of the variance in basal PRL levels (F = 5.53, df = 2/59, P = 0.007). The residual PRL responses were significantly correlated with FT, (Table 2), the ratio L-TRP/CAA (r = 0.336, P < 0.05) and the adjusted 5-HIAA excretion (after controlling by means of multiple regression for 24-h urinary output and creatinine flow). We found no significant correlations either between basal PRL and TSH levels or between residual PRL and TSH responses. The PRL data were not correlated with dopamine contents in 24-h urine and the nutritional parameters, or with the drug state (use of benzodiazepines the day prior to blood sampling or the length of the drug-free period before admission into hospital). Consequently, major effects of the drug state on our results could be disregarded. Discussion Interrelations sponses
between
basal
PRL
and PRL
re-
The results of our study showed that the PRL responses to TRH administration were directly related to the basal PRL levels. Up to 52.9% of the variance in the PRL responses was explained by the regression on basal PRL. Consequently, the PRL responses were largely predicted by the basal
PRL levels. We have shown that the PRL responses to TRH consisted of two components. The first component was a relative exaggeration of the basal PRL levels. The second component was the residual PRL response. This part was estimated by partialling out the relative effects of basal PRL on PRL responses to TRH by means of regression analysis. In clinical practice two factors seems to be relevant for the interpretation of the PRL data: (1) the basal PRL levels reflecting the baseline activity setpoint of the hormone, and (2) the residual PRL values, reflecting the latent capacity of the lactotrope cells in the pituitary to respond to overwhelming amounts of exogenous TRH. It is of interest that the peak/basal PRL ratio, used by several authors as an index for the responsivity of PRL to TRH, shared 81.4% of the variance with the residual PRL responses. Consequently, this ratio offers a relatively good and readily conceptualized measure of the PRL responsivity to exogenous TRH. Baseline PRL levels
The data reported in our paper indicate that the basal PRL levels were no different between major and minor depressives and that they were not related to the severity of illness. These findings agree with those of Maeda et al. (1975) and Witschy et al. (1984). Basal PRL levels were significantly reduced in post-menopausal women and accordingly a significant negative correlation was found between basal PRL and the levels of follicle stimulating hormone (FSH) and age. A reciprocal relationship between PRL and the hypothalamic-pituitarygonadal (HPG) axis could be explained by two factors: (1) a potent PIF was found to be coded within the prohormone for gonadotropin releasing hormone (GnRH) (Nikolics et al., 1985); (2) PRL passes the blood-brain barrier with ease (Belchetz et al., 1982) and PRL inhibits GnRH secretion (Martin and Reichlin, 1987). We have found that basal PRL levels were significantly and negatively correlated with postdexamethasone cortisol levels. Administration of exogenous steroids can affect the PRL responses to various stimuli including TRH (Copinschi et al., 1975; Sowers et al., 1977; Rossier et al., 1980). A plausible explanation for the observed correla-
219
tion is that corticosteroid overdrive, for which a disturbed dexamethasone suppression test is an index (Yerevanian et al., 1983) inhibits PRL secretion (Rothschild et al., 1984) through increased central dopamine activity (Langlais et al., 1984). Although dopamine is a potent PIF (Martin and Reichlin, 1987) we did not find the expected negative correlation with basal PRL. Also other authors (Neill, 1980) have shown that an exact inverse relationship between dopamine and PRL levels was lacking.
could best be interpreted using two factors: firstly basal PRL levels and secondly the residual PRL levels or the peak/basal PRL ratio. (2) Basal PRL and PRL responses to TRH were uninformative for major depression. (3) Basal PRL levels were linked to the HPG axis function, age and postdexamethasone cortisol values. The residual PRL responses were related to the thyroid state and to measures of serotonergic activity.
Residual PRL. responses
We are indebted to Professor Dr. Blockx P., M.D., Ph.D. (nuclear medicine), Professor W. De Potter, Ph.D. (neurochemistry) and to Professor Dr. S. Scharpe, M.D., Ph.D. (pharmacology) who provided laboratory support. The authors express their appreciation to Mrs. C. Maes-Steyaert for secretarial support.
We found that the responsivity of PRL to TRH administration related neither to the diagnostic DSM-III groups nor to the severity of illness. Our findings are thus in agreement with the studies mentioned in our introduction, reporting negative findings on PRL responses to TRH. We found that the residual PRL responses to TRH were significantly and negatively correlated with free triiodothyronine levels. This agrees with the findings that thyroid hormones can inhibit PRL release (Carlson et al., 1978; Garbutt et al., 1986). We have found significant and positive relationships between the PRL responses and serotonergic variables, i.e., the ratio L-tryptophan/competing amino acids (L-TRP/ CAA) and 5hydroxyindoleacetic acid (5-HIAA). The LTRP/CAA ratio is an index of the availability of L-TRP to the brain (Curzon and Sama, 1984; Femstrom, 1984). The brain serotonin synthesis largely depends on the availability of L-TRP (Moir and Eccleston, 1968). 5-HIAA is the major metabolite of serotonin. Our results accord with the established knowledge that plasma PRL responses to TRH are in part mediated through serotonergic mechanisms (Coppen et al., 1980). We could not find a significant relationship between the TSH and PRL responses to TRH administration. This agrees with the results of Loosen et al. (1983) and Zis et al. (1986). Other authors (Frey and Hang, 1977; Geras et al., 1982) also concluded that the mechanisms of TSH release after TRH were different from TRH-induced PRL release. Conclusions
(1) In clinical practice PRL responses to TRH
Acknowiedgements
References .&en, H. and Wide, L. (1986) TRH tests with analyses of TSH, prolactin and GH responses in subtypes of patients with major depressive disorders. Acta Psychiatr. Stand. 73, 549-558. Asnis, G.M., Sachar, E.J., Halbreich, U., Swami Nathan, R., Ostrow, L., Solomon, M. and HaIpem, F.S. (1981) Endocrine responses to thyrotropin-releasing hormone in major depressive disorders. Psychiatr. Res. 5, 205-215. Belchetz, P.E., Ridley, R.M. and Baker, H.F. (1982) Studies on the accessibility of prolactin and growth hormone to brain: effect of agonists on hormone levels in serial, simultaneous plasma and cerebrospinal fluid samples in the rhesus monkey. Brain Res. 239, 310-314. Bowers, C.Y., Friesen, H.G., Hwang, P., Guyda, H.J. and Folkers, K. (1971) Prolactin and thyrotropin release in man by synthetic pyroglutamyl-histidyl-prolinamide. B&hem. Biophys. Res. Commun. 45, 1033-1041. Brambilla, F., Smeraldi, E., Sacchetti, E., Negri, F., Cocchi, D. and Miiller, E.E. (1978) Deranged anterior pituitary responsiveness to hypothalamic hormones in depressed patients. Arch. Gen. Psychiatry 35, 1231-1238. Carlson, H.E., Sawin, C.I.,, Krugman, L.G., Meyer, N.C. and Hershman, J.M. (1978) Effect of thyroid hormones on the prolactin response to thyrotropin-releasing hormone in normal persons and euthyroid goitrous patients. J. Clin. Endocrinol. Metab. 47, 275-279. Copinschi, G., L’Hermite, N., Leclercq, R., Goldstein, J., Vanhaelst, L., Virasoro, E. and Rofyn, C. (1975) Effects of glucocorticoids on pituitary hormonal response to hypoglycemia. Inhibition of prolactin release. J. Clin. Endocrinol. Metab. 40, 442-449. Coppen, A., Rama Rao, V.A., Bishop, M., Abou-Saleh, M.T.
220 and Wood, K. (1980) Neuroendocrine studies in affective disorders: part 1: plasma prolactin response to thyrotropin-releasing hormone in affective disorders: effect of ECT. J. Affect. Disord. 2, 311-315. Cutzon, G. and Sama, G.S. (1984) Tryptophan transport to the brain: newer findings and older ones reconsidered. In: H.G. Schlossberger, W. Kochen, B. Linzen and H. Steinhart (Eds.), Progress in Tryptophan and Serotonin Research. Walter de Gruyter, New York, NY, pp. 145-160. Delitala, G., Tomasi, P. and Virdis, R. (1987) Prolactin, growth hormone and thyrotropin-thyroid hormone secretion during stress states in man. In: A. Grossman (Ed.), Clinical Endocrinology and Metabolism, Vol. 1, Neuroendocrinology of Stress. Ball&e Tindal, London, pp. 391-414. Ehrensing, R.H., Kastin, A.J., Schalch, D.S., Friesen, H.G., Vargas, J.R. and Schally, A. (1974) Affective state and thyrotropin and prolactin responses after repeated injections of thyrotropin-releasing hormone in depressed patients. Am. J. Psychiatry 131, 714-718. Fernstrom, J.D. (1984) Tryptophan availability and serotonin synthesis in rat brain: effects of experimental diabetes. In: H.G. Schlossberger, W. Kochen, B. Linzen and H. Steinhart (Eds.), Progress in Tryptophan and Serotonin Research. Walter de Gruyter, New York, NY, pp. 161-172. Frey, M.M. and Hang, E. (1977) Effect of prolonged oral administration of TRH on plasma levels of thyrotrophin and prolactin in normal individuals and in patients with primary hypothyroidism. Acta Endocrinol. 85, 744-752. Garbutt, J.C., Loosen, P.T., Blacharsh, K. and Prange, A.J. (1986) The prolactin response to thyrotropin-releasing hormone in depressed patients and normal subjects. Psychoneuroendocrinology 11, 213-219. Geras, E., Rebecchi, M.J. and Gershengom, M.C. (1982) Evidence that stimulation of thyrotropin and prolactin secretion by thyrotropin-releasing hormone occur via different calcium mediated mechanisms: studies with verapamil. Endocrinology 110, 901-906. Gregoire, F., Brauman, H., De Buch, R. and Corvilain, J. (1977) Hormone release in depressed patients before and after recovery. Psychoneuroendocrinology 2, 303-312. Jacobs, L.S., Snyder, P.J., Wilber, J.F., Utiger, R.D. and Daughaday, W.H. (1971) Increased serum prolactin after administration of synthetic thyrotropin releasing hormone (TRH) in man. J. Clin. Endocrinol. Metab. 33, 996-998. Kirkegaard, C., Eskildsen, P.C. and Bjerum, N. (1981) Parallel changes of the responses of thyrotropin, growth hormone and prolactin to thyrotropin-releasing hormone in endogenous depression. Psychoneurcendocrinology 6, 253-259. Langer, G., Schoenbeck, G., Koinig, G., Reiter, H., Schuessler, M., Aschauer, H. and Lesch, 0. (1980) Evidence for neuroendocrine involvement in the therapeutic effects of antidepressant drugs. In: F. Brambilla, 0. Racagni and D. De Wied (I%.), Progress in Psychoneuroendocrinology. Elsevier, New York, NY, pp. 197-208. Langlais, P.J., Rothschild, A.J., Schatzberg, A.F., Cole, J.O. and Bird, E.D. (1984) Dexamethasone elevates dopamine in human plasma and rat brain. Psychopharmacol. Bull. 20, 365-370. Linkowski, P., Brauman, H. and Mendlewicz, J. (1980) Pro-
lactin secretion in women with unipolar and bipolar depression. Psychiatr. Res. 3, 265-271. Linnoila, M., Lamberg, B.A., Rosberg, G., Karonen, S.-L. and Welin, M.S. (1979) Thyroid hormones and TSH, prolactin and LH responses to repeated TRH and LHR injections in depressed patients. Acta Psychiatr. Stand. 59, 536-544. Loosen, P.T., Kistler, K. and Prange, A.J. (1983) Use of TSH response to TRH as an independent variable. Am. J. Psychiatry 140, 700-703. Maeda, K., Kato, Y., Ohgo, S., Chihara, K., Yoshimoto, Y., Yamaguchi, N., Kuromaru, S. and Imura, H. (1975) Growth hormone and prolactin release after injection of thyrotropin-releasing hormone in patients with depression. J. Clin. Endocrinol. Metab. 40, 501-505. Maes, M., Vandewoude, M., Maes, L., Schotte, C. and Cosyns, P. (1989) A revised interpretation of the TRH test results in female depressed patients. Part I: TSH responses. Effects of severity of illness, thyroid hormones, monoamines, age, sex hormonal, corticosteroid and nutritional state. J. Affect. Disord. 16, 203-213. Martin, J.B. and Reichlin, S. (1987) Clinical Neuroendocrinology, 2nd edn. F.A. Davis Company, Philadelphia, PA. Moir, A.T.B. and Eccleston, D. (1968) The effect of precursor loading in the cerebral metabolism of 5-hydroxy-indoles. J. Neurochem. 15, 1093-1108. Naeye, R., Goldstein, J., Zegers de Beyl, D., Linkowski, P., Mendlewicz, J., Copinschi, G., Badawi, M. and Leclercq, R. (1978) Thyrotrophin, prolactin and growth hormone responses to TRH in barbiturate coma and in depression. Clin. Endocrinol. 9, 49-58. Neill, J.D. (1980) Neuroendocrine regulation of prolactine secretion. In: L. Martini and W.F. Ganong (Eds.), Frontiers in Neuroendocrinology, Vol. 6, Raven Press, New York, NY, pp. 129-155. Nikolics, K., Mason, A.J., Szonyi, E., Ramachandran, J. and Seeburg, P.H. (1985) A prolactin-inhibiting factor within the precursor for human gonadotropin-releasing hormone. Nature 316, 511-517. Rabello, M.M., Snyder, P.J. and Utiger, R.D. (1974) Effects on the pituitary-thyroid axis and prolactin secretion of single and repetitive oral doses of thyrotropin-releasing hormone (TRH). J. Clin. Endocrinol. Metab. 39, 571-578. Rossier, J., French, E., Rivier, C., Shibasaki, T., Guillemin, R. and Bloom, F.E. (1980) Stress-induced release of prolactin: blockade by dexamethasone and naloxone may indicate /I-endorphin mediation. Proc. Natl. Acad. Sci. U.S.A. 77, 666-669. Rothschild, A.J., Langlais, P.J., Schatzberg, A.F., Walsh, F.Y., Cole, J. and Bird, E.D. (1984) Dexamethasone increases plasma free dopamin in man. J. Psychiatr. Res. 18,217-223. Rutlin, E., Hang, E. and Torjesen, P.A. (1977) Serum thyrotropin, prolactin and growth hormone response to TRH during oestrogen treatment. Acta Endocrinol. 84, 23-35. Sowers, J.R., Carlson, H.E., Brautbar, N. and Her&man, J.M. (1977) Effect of dexamethasone on prolactin and TSH response to TRH and metoclopramide in man. J. Clin. Endocrinol. Metab. 44, 237-241. Targum, S.D., Sullivan, A.C. and Bymes, S.M. (1982) Com-
221 pensatory pituitary-thyroid mechanisms in major depressive disorder. Psychiatr. Res. 6, 85-96. Tumell, D. and Cooper, J. (1982) Rapid assay for amino acids in serum or urine by precohmm derivatisation and reversed-phase liquid chromatography. Clin. Chem. 28, 527-531. Winokur, A., Amsterdam, J., Oler, J., Mendels, J., Snyder, P.J., Caroff, S.N. and Brunswick, D.J. (1983) Multiple hormonal responses to protirelin (TRH) in depressed patients. Arch. Gen. Psychiatry 40, 525-531. Witschy, J.K., S&lesser, M.A., Fulton, C.L., Orsulak, P.J.,
Giles, D.E., Fairchild, C., Crowley, G. and Rush, J. (1984) TRH-induced prolactin release is blunted in females with endogenous unipolar major depression. Psychiatr. Res. 12, 321-331. Yerevanian, B.I., Woolf, P.D. and Iker, H.P. (1983) Plasma ACTH levels in depression before and after recovery: relationship to the dexamethasone suppression test. Psychiatr. Res. 10, 175-181. Zis, A.P., Albala, A.A., Haskett, R.F., Carroll, B.J. and Lohr, N.E. (1986) Prolactin response to TRH in depression. Psychiatr. Res. 20, 77-82.