BIOL PSYCHIATRY 1992;3 ! :975-983
975
Effect of Clomipramine and Lithium on Fenfluramine-Induced Hormone Release in Major Depression Baruch Shapira, Mohammed J. Yagmur, Cornelius Gropp~ Michael Newman, and Bernard Lerer
Prolactin (PRL) and cortisol responses to oral administration of d-I fenfluramine hydrochloride (60 mg) and placebo were examined in patients with endogenous major depressive disorder on three separate occasions: prior to treatment with clomipramine (CMi), after 4 weeks of CMl administration (175-250) rag/day), and 3 weeks after addition of lithium (Li) carbonate (serum level 0.5-0.9 retool) to the treatment regimen. CM! significantly increased baseline PRL levels which were further elevated following Li supplementation. PRL response to fenfluramine (minus elevated baseline PRL levels) but not to placebo, was significantly increased by CMI administration, reflected over the 6-hr time course examined and in peak minus baseline values. Following addition of Li, The degree of enhancement was diminished although the peak minus baseline value remained significant relative to the pretreatment response. Ccrtisol levels were not increased by fenfluramine and were not altered by CM! or CM! + Li administration. The effect of CM! extends previous observations regarding the action of antidepressant treatment on serotenergically mediated hormone release. Methodological considerations relevant to the effect of CMi + Li are discussed.
Introduction Although studies in laboratory animals have yielded a substantial body of data regarding the effect of antidepressant treatments on serotonergic neurotransmission, a precise understanding of their mechanism of action has yet to emerge (Heninger and Charney 1987). It is not clear whether signal transduction via serotonergic neuronal pathways is enhanced following chronic antidepressant treaunent, as suggested by electrophysiological studies (De Montigny and Blier 1984) or is ~duced, as suggested by radioligand binding studies of rat cortical 5-HT2 receptors and their associated phosphatidyl inositol second messenger ~y,,,,-.,,, ,,~:~.,,~ and Lerer 1989). lVloreover, elec_.~onvulsive shock (ECS) has been shown to increase serotonin (5-HT2) receptor number and the results of electrophysio,,~,,.~...~
ql~T ....
From the Depression Treatment Unit, Herzog Hospitzd, Jt:rasalem, Israel (BS, CG); the Depamnent of Psychiatry, Hebrew University-Hedassah Medical School, Jerusalem, Israel (BS, MN, BL); and the Arieh Jmos-Kfar Shaul Mental Health Center, Jerusalem,, Israel (MJY). Address reprint reque:as to Bamch Shapiro, M.D., Depression Treatment Unit, Herzog Hospital, P.O. Box 140, Jerusalem 91001, Israel. Received July 26, 1991; revised December 21, 1991. © 1992 Society of Biological Psychiatry
0006-3223/92/$05.00
~6
BIOL PSYCHIATRY 1992~31:975-983
B. Shapira et al
logical studies with ECS parallel those obtained with antidepressants (Lerer 1987). Considerable interest has therefore been focused on the effect of antidepressant treatments on 5-HT neurotr~smission in human subjects in order to resolve the questions arising from preclinical studies and to seek a common mechanism of action (Price et al 1990). Neur~ndocrine challenge tests have been the most frequently used strategy for evaluating cenwal 5-HT function in depressed patients. Plasma prolactin (PRL) release after intravenous loading with L-tryptophan has been evaluated in a number of studies which have shown blunted responses in depressed patients and enhancement following tricyclic (TCA) and other antidepressant drug treatment (e.g., Heninger et al 1984; Cowen and Charig 1987; Chamey et al 1984; Price et al 1989a). Studies with orally administered d1 fenfluramine hydrochloride have also shown blunted PRL release, particularly in depressed patients with endogenous features (Siever et al 1984; Coccaro et al 1989; Mitchell and Smythe 1990). Shapira et al (1989) have demonstrated enhancement of the PRL response to fenfluramine in depressed ~atients following treatment with imipramine and also after electroconvulsive therapy (E~l") (Shapira et al 1992). The primary action of fenfluramine is to promote serotonin release from presynaptic neurons. The evidence supporting ser¢~tonergic mediation of PRL release by this agent in animals and humans has been reviewed elsewhere (Coccaro et el 1989). There is also evidence that fenfluramine stimulates c¢~tisol release (Lewis and Sherman 1984), although findings in this regard are less consistent (Lerer et al 1988; Mitchell and Smythe 1990), and a report of blunted responsiveness in depression (Weizman et al 1988) has not been replicated (Mitchell and Smythe 1990). The purpose of this study was to evaluate the effect of antidepressant treatment with clomipramine (CMI), an agent with a vreferential but not exclusive action on serotoninergic neurotransmission (Hall and Ogren 1981), on fenfluramine-induced hormone release. A further objective was to examine the effect of concurrently administered lithium carbonate (Li) on the response. There is considerable clinical interest in the effect of Li to augment the therapeutic action of TCA and other antidepressants in patients whose depressive symptoms are refractory to these agents (Schou 1990). Li supplementation has been proposed to act by enhancing 5-HT neurotransmission (De Montigny et al 1983). It was therefore of interest to examine the effect of Li supplementation on serotonergically mediated hormone release in depressed patients treated with CMI. Methods Nine physically healthy patients (four men, 5 women) hospitalized in the Herzog HospitalEzrath Nashim Depression Treatment Unit gave written informed consent to participate in the study. All fulfilled Research Diagnostic Criteria (Spitzer et al 1978) for major depressive disor !.,-r,endogenous subtype (eight were unipolar, one bipolar) and were free of all medication other than chloral hydrate for nocturnal sedatim~, for at least 2 weeks before the study. Their mean (_+SEM) age was 48.8 _+ 2.79 years. Their Hamilton Depression Rating Scale (HDRS-21 item) score (Guy 1976) was 24.8 _ 3.41. All the patients were treated with CMI (t.i.d. dosage schedule) which was increased over a period of 7-10 days to a minimum dose of 175 mg/day (mean 203.1 _+ 11.04, range 175-250 mg/day) and then continued at thi,~; dosage for a further 3 weeks. Four patients manifested little or no clinical change (HDRS |8 _ 2.64) whereas five patients were remitted or had significantly improved (HDRS 6.4 _+ 1.2). All then commenced treatment with Li (b.i.d. schedule) which was increased to maintain a serum level of
Effect of Drugs on Fenfluramine Hormone Release
BIOL PSYCHIATRY 1992;3h975-983
9"]7
0.63 4. 0.04 mmol (range 0.5-0.9 mmol). Clinical justification for the addition of Li was either as a potential augmenter in the nonimproved patients or as a p~z]ude to maintenance therapy in those who had remitted. Two of the patients who had not improved on CMI alone responded to Li supplementation (HDRS change >50%). Fenfluramine challenge tests were performed in the drug-free state prior to conLrnencing CMI treatment and after 4 weeks of CMI. Three weeks after Li was added to the treatment regimen a further test was performed. Fenfluramine (60 mg orally) and placebo challenges were administered after a 12-hr fast (including no medication) under random assignment, double-blind conditions as previously described (Lerer et al 1988; Shap/ra et al 1989) and were separated by at least 48 hr. At 8:00 AM on the challenge day, an intravenous catheter was inserted in a forearm vein and was kept patent by a heparin lock. Thirty min later, the first baseline blood sample was drawn, followed after 15 min by the second baseline sample. The challenge medication was then administered. Blood sampling continued hourly for the next 6 hr. Uniform, nontryptophan-contalning snacks were permitted 1 and 4 hr following administration of the medication (after blood sampling), and locomotion was restricted as much as possible. Blood samples were. drawn into EDTAtreated tubes which were kept in an icewater bath until completion of the procedure. Plasma was extracted from the samples and stored at -80°C until assayed. All samples from a particular subject were assayed on a single day. Prolactin and cortisol levels were determined using commercially available radioimmunoassay kits as previously described (Lerer et al 1988; Shapira et al 1992). For prolactin, interassay variation was 5.6% and intraassay variation 6.2%; for cortisol 5.2% and 3.5%, respectively. Serum Li levels were determined by flame photometry. Data were analyzed by analysis of variance (ANOVA) with repeated measures and paired t-tests (two-tailed). Correlations were computed by the Pearson correlation coefficient. Values are presented as mean -+ SEM. For basehne values, means of the - 15 and 0 min samples were used as these did not differ significantly. One subject (a nonresponder to CMI and CMI + Li) was omitted from the analysis of the hormone data because her pretreatment baseline PRL levels were more than 3 SDs higher than the group mean. Results
Figure 1 shows the effect of CMI treatment (4 weeks) and CMI + Li (3 weeks) on the PRL response to fenfluramine (upper panel) and placebo challenge (lower panel). Results are presented as delta values (minus baseline) for each hr as baseline prolactin levels increased significantly during CMI and CMI 4- Li treatment (pretreatment 6.01 _+ 0.61; post-CMI 9.77 _+ 0.81 ng/ml, paired t - 4.18, p - 0.0008; post-Li + CMI 13.42 4. 1.69 ng/ml, t - 4.61, p - 0.0004) CMI treatment significantly increased PRL release after fenfluramine challenge ( F - 10.3; df 1,88; p = 0.01). Maximal enhancement by CMI treatment of PRL release was observed at the 4- and 5-hr time points. There was no evident trend for PRL release following fenfluramine to be more enhanced in the patients who responded to CMI treatment compared with those who did not. "ihere was no effect of CMI or CMI + Li on PRL levels following placebo challenge. Addition of Li resulted in slightly lower PRL levels in response to fenfluramine and a loss of significance versus pretreatment values (when comparing the full time course of the hormone responses). Figure 2 shows peak minus baseline PRL response to fenfluramine and placebo. On this measure, the response to fenfluramine was greater than
978
B. Shapira et al
BIOL PSYCHIATRY 1992 ;31:975-983
FENFLURAMINE CHALLENGE
12
10 8-
(*) o PRETREATMENT B CLOMIPRAMINE
~
~"
/)lT
CLOMIP. + LITHIUM
/
",..T T
........
T T ..........v"" .~.,, /...... ~...
Z
......... /
n
0 -2 1
E ¢m i¢
3
4
5
6
PLACEBO CHALLENGE i
12
10-
2
0
PRETREATMENT
B
CLOMIPRAMINE
v
CLOMIP. + LITHIUM
Z m
I-
¢j
_J 0 (1.
~ ~ ,,"~ r~''''
'"°.~T
-2 1
2
3
TIME
4
5
6
(hours)
Figure 1. Effect of CMI and CMI + Li on PRL response to fenfluramine and placebo challenge in patients with endogenous major depressive disorder. Bars show SEM. PRL levels (minus baseline) are significantly enhanced by CM! (F = 10.3; df 1,88; p - 0.01) but not CMI + Li. * p = 0.01, (*) p - 0.09 (versus corresponding time point pretreatment).
pretreatment (F -- 3.58; df 2, 7; p = 0.05) in both the CMI (p -- 0.04) and CMI + Li (p - 0.03) conditions but did not show further enhancement by Li compared with CMI alone. There was no trend for CMI responders and CMI nonresponders to differ in this regard. In view of a possible effect of baseline prolactin levels on the prolactin response to fenfluramine, correlations were examined under all three treatment conditions. In the pretreatment, drug-free state (r = - 0 . 1 0 ; p - 0.8) and during CMI administration
Effect of Drugs on Fenfluramine Hormone Release
BIOL PSYCHIATRY 1992;31:975-983
979
E gz
I PRETREATMENT ~ CLOMIPRAMINE
lS-
z m
IO
~
12
'A"
CLOMIP.+ LITHIUM
..I
O er ttl Z __1 ILl O~ Ill m
C~
"r
.
Figure 2. Peak minus baseline values for PRL response to fenfluramine and placebo challenge--pretreatment, during CMI administration and after addition of Li. Bars show SEM. * p - 0.04 (for CMI), p 0.03 (for CMI + Li) versus pretreatment.
Z m
PLACEBO
FENFLURAM!NE
ILl
L
(r ffi 0.50; p ffi 0.2), no relationship was observed. During CMI + Li administration, however, there was a significant inverse relationship (r ffi -0.84; p = 0.008). Unlike PRL, baseline cortisol levels were not significantly altered by CMI treatment. There was no relationship between baseline cortisol and baseline PRL levels nor between baseline cortisol and peak or peak minus baseline PRL levels in the drug-free state or during either of the treatment conditions. As shown in Figure 3, fenfluramine challenge did not alter cortisol levels relative to the effect of placebo. There was no significant effect of treatment with CMI or CMI + Li on cortisol response to the challenge agent.
Discussion The present timings extend our earlier report (Shapira et al 1989) which showeci that 3 weeks of treatment with imipramine significantly enhanced the PRL response to fenfluramine but not placebo challenge. The results complement a number of other studies which used the PRL response to L-tryptophan challenge as a measure of serotonergic responsitivity and also demonstrated enhancement after a variety of antidepressant drugs in normal subjects (Cowen et al 1986) as well as in depressed patients (Charney et al 1984; Price et al 198%). We have also shown that treatment of depressed patients with ECT results in a significant enhancement of fenfluramine-induced PRL release (Shapira et ai 1992). The absence in our stud), of a Li effect is inconsistent with the findings of one group but is partially supported by those of another (although both these groups used a different serotonergic challenge and the results are not directly comparable). Glue et al (1986) found that both short- (4 days) and long-term (20 days) administration of Li to normal volunteers significantly enhanced the PRL response to L-tryptophan challenge. In depressed patients resistant to treatment with TCA, addition of Li for 20 days also resulted
980
B. Shapira et al
BIOL PSYCHIATRY
1992;31:975-983
FENFLURAMINE CHALLENGE 12 o PRETREATMENT 10
m CLOMIPRAMINE CLOMIP. + LITHIUM
E w
I=
T
.J O 03 i I--er O O
.
~
,,'"
..."
; ........ Z & / / / /J
1J,
-2
I
2
3
4
5
6
PLACEBO CHALLENGE
12
o PRETREATMENT 10-
g CLOMIPRAMINE .
O) ¢:
v CLOMIP, + LITHIUM
.
.,J 0 03 m I-, tr 0 (,,1
,
''''''''''',
I, b
.
-2
f 1
2
3
TIME
4
5
6
(hours)
Figure 3. Effect of CMI and CMI + Li on cortisol levels (minus baseline) following fenfluramine and placebo administration. Bars show SEM. No differences statistically significant at p < 0.1.
in enhancement of the response compared with the TCA-only condition (Cowen et al 1989). In contrast, Price et al (1986) found that although short-term Li administration increased the PRL response to L-tryptophan in the patients they examined, longer-term Li administration did not. Various factors that may have influenced our results should be noted. One consideration is the small number of subjects, although the absence of any trend towards enhancement by Li renders this possibility less likely. A second possibility is that this result may be peculiar to CMI which, unlike other TCA, was found to increase baseline PRL levels.
Effect of Drugs on Fenfluramine Hormone Release
elOLPS¥Ct;~Y 1992;31:975-983
981
This effect of CMI has been previously reported (Zohar et al 1988). Baseline PRL levels increased further after addition of Li to the treatment regimen. This effe~rt could have been d,~e to a continued and cumulative action of CMI, althoagh in the absence of a group of subjects treated for 7 weeks with Cg~ alone, this possibility is speculative. However, Li has not been shown to raise baseline prolactin levds in studies of this type (e.g., Glue et al 1986; Price et al 1989b). Whatever mechanism might have been operative in raising baseline prolactin levels, it is noteworthy that during the fenfluraraine, challenge conducted under CMI + Li adr.ainistration, a significant inverse correlation between baseline prolactin level and peak minus baseline prolactin response to fenfluramine was observed. This might well have masked an effect of Li to enhance fenflur,~mine-induced prolactin release. Finally, the fact that the subject group included patients who had responded to CMI when Li was added, as well as those who were CMI-refractory, must also be considered. In this regard, the present sample differs from the TCA-refractory sample studied by Cowen et al (1989~ in groom PRL response to tryptophan was enhanced
by t,i. Cortisol rrsponses to fenfluramine were not significantly different from those after placebo and were not altered by CMI or CMI plus Li administration. This is consistent with our own previous observations (Lerer et al 1988) and those of other authors (Mitchell and Smythe 1990).-egarding fenfluramine effects on cortisol release, (It is, however, possible that a stimulatory effect of fenfluramine on cortisol release might be better demonstrated in the early evening when baseline cortisol levels are lower and more stable than in the morning.) We also found no relationship between baseline cortisol and peak or peak minus baseline PRL responses to fenflur~Tdne. Such a relationship has been reported by Mitchell and Smythe (1990) who found that blunted PRL responses to fenfluramine in depressed patients were dependent on baseline cortisol levels. However, the subjects in the present study were all endogenously depressed and the relationship between baseline cortisol levels and PRL response to fenfluramine was not found by Mitchell et al (1990) to be significant in this subgroup of patients. In conclusion, the present report provides additional evidence that serotonergically mediated PRL release is enhanced by TCA treatment. Together with studies employing L-tryptophan challenge, it suggests that antidepressant drugs as well as ECT enhance central 5-HT function in depressed patients. This effect was present in patients who responded to treatment as well as those who did not, and must thus be considered a necessary rather than a sufficient concomitant of antidepressant response, as has been previously noted (Price et al 1989a). The precise receptor mechanisms involved in the enhancement by antidepressant treatment of serotonergically mediated hormone release as well as the role of other neurotransmitter systems in the antidepressant response, remains to be elucidated. The failure of Li to further augment fenfluramine-induced PRL release requires reevaluation, taking into account the methodological considerations noted. Supported in part by NIMH Grant #MH-43873.
References Charney DS, Heninger GR, Steinberg DE (1984): Serotonin function and mechanism of action of antidepressant treatment: Effects of aitriptyline and desipramine. Arch Gen Psychiatry 41:359365. Coccaro EF, Siever LJ, Klar HM, et al (1989): Serotonergic studies in patients with affective and
982
BIOL PSYCHIATRY 1992;31:975--983
B. Shapira et al
personality disorders: Correlates with suicidal and impulsive aggressive behaviour. Arch Gen Psychiatry 46:587-599. Cowen PJ, Charig EM (1987): Neuroendocrine responses to intravenous tryptophan in major depression. Arch Gen Psychiatry 44:958-966. Cowen PJ, Geaney DP, Schachter M, et al (1986): Desipramine treatment in normal subjects: Effects on neur~endocrine responses to tryptophan and on platelet serotonin (5-HT)-related receptors. Arch Gen Psychiatry 43:61-67. Cowen PJ, McCance SL, Cohen PR, Julier DL (1989): Lithium increases 5-HT mediated neuroendocrine responses in tficyclic resistant depressions. Psychopharmacology 99:230-232. De Montigny C, Blier P (1984): Effects of antidepressant treatments on 5-HT neurotransmission: Electrophysiological and clinical studies. In Usden E (ed), Advances in Biochemical Psychopharmacology, vol. 39: Frontiers in Biochemical and Pharmacological Research in Depression. New York: Raven Press. De Montigny C, Cournoyer G, Morisette R, Langlois R, Caille G (1983): Lithium carbonate addition in tficyclic antidepressant-resistant unipolar depression: Correlations with the neurobiologic actions of tricyclic antidepressant drugs and lithium ion on the serotonin system. Arch Gen Psychiatry 40:1327-1334. Glue PW, Cowen PJ, Nutt DJ, Kolakowska T, Grahame-Smith DG (1986): The effect of lithium on 5-HT-mediated neuroendocrine responses and platelet 5-HT receptors. Psychopharmacology 90:498-502. Guy W (1976): ECDEU Assessment Manual for Psychopharmacology. US Dept. of Health, Educati(~n and Welfare, National Institute of Mental Health, Rockville, MD. Hall H, Ogren SO (1981): Effects of antidepressant drugs on different receptors in the brain. Fur J Pharmacol 70:393-407. Heninger GR, Chamey OJ (1987): Mechanism of action of antidepressant treatments: Implications for the etiology and treatment of affective disorders. In Meltzer HY (ed), Psychopharmacology: The Third Generation of Progress. New York: Raven Press. Heninger GR, Charney DS, Steinberg DE (1984): Serotonergic function in depression: Prolactin response to intravenous tryptophan in depressed patients anti healthy subjects. Arch Gen Psychiatry 41:398-402. Lerer B (1987): Neurochemical and other neurobiological consequences of ECT: Implications for t.hq~,pathogenesis and treatment of affective disorders. In Meitzer HY (ed), Psychopharmacology: Tke Tl:ird Generation of Progress. New York: Raven Press. Lerer B Ran A, Blacker M, et al (1988): Neuroendocrine responses in chronic schizophrenia: Evidence for serotonergic dysfunction. Schizophr Res 1:405-410. Lewis DA, Sherman BM (1984): Serotonergic stimulation of adrenocorticotropin secretion in man. J Cl.in Endocrinol Metab 58:458-462. Mitchel~ P, Smythe G (1990): Hormonal responses to fenfluramine in depressed and control subjects. J Aff.:ct Disord !9:43-56. Mitchell P, Smy:itc G, Parker G, et al (1990): Hormonal responses to fenfluramine in depressive subtyI~es. Br J Psychiatry 157:551-557. Newma~ M, Lerer B (1989): Modulation of second messenger function in rat brain by in ~'ivo alterwtion of receptor sensitivity: Relevance to the mechanism of action of electroconvulsive • ~rapy and antidepressants. Prog Neuropsy~hopharmacol Biol Psychiatry 13:1-30. Price LH, Chamey DS, Delgado PL, et al (1989a~: Effects of disipramine and fluvoxamine treatment of the prolactin response to tryptophan: Serotonergic function and the mechanism of antidepressant action. Arch Gen Psychiatry 46:625-631. Price LH, Charney DS, Delgado PL, Heninger GR (1989b): Lithium treatment and serotoninergic function. Arch Gen Psychiatry 46:13-19. Price LH, Charney DS, Delgado PL, et al (1990): Clinical data on the role of serotonin in the mechanism of action of antidepressant drugs. Y Clin Psychiatry 51 (Suppl 4):44-50.
Effect of Drugs on Penfluramine Hormone Release
BIOLPSYCHIATRY 1992;31:975-983
983
Schou M (1990): Lithium and treatment-resistant depressions: A review. Lithium 1:3-8. Shapira B, Reiss A, Kaiser N, Kindler S, Lever B (1989): Effect of imipramine treatment on the prolactin response to fenfluramine and placebo challenge in depressed patients. J Affect Disord 16:1-4. Shapira B, Lerer B, Kindler S, et al (1992): Enhanced serotonergic r~sponsivity following electroconvulsive therapy in patients with major depression. Br J Psychiatry (in press). Siever LJ, Mutphy DL, Slater S, de la Vega E, Lipper S (1984): Plasma prolactin changes following fenfluramine in depressed patients compared to controls: An evaluation of central serotonergic responsivity in depression. Life Sci 34:1029-1039. Spitzer RL, Endicott J, Robins E (1978): Research Diagnostic Criteria: Rationale and reliability. Arch Gen Psychiatry 35773-782. Weizman A, Mark M, Gil-Ad I, Tyano S, Laron 2 (1988): Plasma cortisol, prolactin, growth hormone and immunoreactive @-endorphin response to fenfluramine challenge in depressed patients. Clin Neurophamacol 11:250-256. Zohar J, Insel TR, Zohar-Kadouch RC, Hill JL, Murphy DL (1988): Serotonergic responsivity in obsessive compulsive disorder. Arch Gen Psychiatvy 45:167-172.
975
Effect of Clomipramine and Lithium on Fenfluramine-Induced Hormone Release in Major Depression Baruch Shapira, Mohammed J. Yagmur, Cornelius Gropp~ Michael Newman, and Bernard Lerer
Prolactin (PRL) and cortisol responses to oral administration of d-I fenfluramine hydrochloride (60 mg) and placebo were examined in patients with endogenous major depressive disorder on three separate occasions: prior to treatment with clomipramine (CMi), after 4 weeks of CMl administration (175-250) rag/day), and 3 weeks after addition of lithium (Li) carbonate (serum level 0.5-0.9 retool) to the treatment regimen. CM! significantly increased baseline PRL levels which were further elevated following Li supplementation. PRL response to fenfluramine (minus elevated baseline PRL levels) but not to placebo, was significantly increased by CMI administration, reflected over the 6-hr time course examined and in peak minus baseline values. Following addition of Li, The degree of enhancement was diminished although the peak minus baseline value remained significant relative to the pretreatment response. Ccrtisol levels were not increased by fenfluramine and were not altered by CM! or CM! + Li administration. The effect of CM! extends previous observations regarding the action of antidepressant treatment on serotenergically mediated hormone release. Methodological considerations relevant to the effect of CMi + Li are discussed.
Introduction Although studies in laboratory animals have yielded a substantial body of data regarding the effect of antidepressant treatments on serotonergic neurotransmission, a precise understanding of their mechanism of action has yet to emerge (Heninger and Charney 1987). It is not clear whether signal transduction via serotonergic neuronal pathways is enhanced following chronic antidepressant treaunent, as suggested by electrophysiological studies (De Montigny and Blier 1984) or is ~duced, as suggested by radioligand binding studies of rat cortical 5-HT2 receptors and their associated phosphatidyl inositol second messenger ~y,,,,-.,,, ,,~:~.,,~ and Lerer 1989). lVloreover, elec_.~onvulsive shock (ECS) has been shown to increase serotonin (5-HT2) receptor number and the results of electrophysio,,~,,.~...~
ql~T ....
From the Depression Treatment Unit, Herzog Hospitzd, Jt:rasalem, Israel (BS, CG); the Depamnent of Psychiatry, Hebrew University-Hedassah Medical School, Jerusalem, Israel (BS, MN, BL); and the Arieh Jmos-Kfar Shaul Mental Health Center, Jerusalem,, Israel (MJY). Address reprint reque:as to Bamch Shapiro, M.D., Depression Treatment Unit, Herzog Hospital, P.O. Box 140, Jerusalem 91001, Israel. Received July 26, 1991; revised December 21, 1991. © 1992 Society of Biological Psychiatry
0006-3223/92/$05.00
~6
BIOL PSYCHIATRY 1992~31:975-983
B. Shapira et al
logical studies with ECS parallel those obtained with antidepressants (Lerer 1987). Considerable interest has therefore been focused on the effect of antidepressant treatments on 5-HT neurotr~smission in human subjects in order to resolve the questions arising from preclinical studies and to seek a common mechanism of action (Price et al 1990). Neur~ndocrine challenge tests have been the most frequently used strategy for evaluating cenwal 5-HT function in depressed patients. Plasma prolactin (PRL) release after intravenous loading with L-tryptophan has been evaluated in a number of studies which have shown blunted responses in depressed patients and enhancement following tricyclic (TCA) and other antidepressant drug treatment (e.g., Heninger et al 1984; Cowen and Charig 1987; Chamey et al 1984; Price et al 1989a). Studies with orally administered d1 fenfluramine hydrochloride have also shown blunted PRL release, particularly in depressed patients with endogenous features (Siever et al 1984; Coccaro et al 1989; Mitchell and Smythe 1990). Shapira et al (1989) have demonstrated enhancement of the PRL response to fenfluramine in depressed ~atients following treatment with imipramine and also after electroconvulsive therapy (E~l") (Shapira et al 1992). The primary action of fenfluramine is to promote serotonin release from presynaptic neurons. The evidence supporting ser¢~tonergic mediation of PRL release by this agent in animals and humans has been reviewed elsewhere (Coccaro et el 1989). There is also evidence that fenfluramine stimulates c¢~tisol release (Lewis and Sherman 1984), although findings in this regard are less consistent (Lerer et al 1988; Mitchell and Smythe 1990), and a report of blunted responsiveness in depression (Weizman et al 1988) has not been replicated (Mitchell and Smythe 1990). The purpose of this study was to evaluate the effect of antidepressant treatment with clomipramine (CMI), an agent with a vreferential but not exclusive action on serotoninergic neurotransmission (Hall and Ogren 1981), on fenfluramine-induced hormone release. A further objective was to examine the effect of concurrently administered lithium carbonate (Li) on the response. There is considerable clinical interest in the effect of Li to augment the therapeutic action of TCA and other antidepressants in patients whose depressive symptoms are refractory to these agents (Schou 1990). Li supplementation has been proposed to act by enhancing 5-HT neurotransmission (De Montigny et al 1983). It was therefore of interest to examine the effect of Li supplementation on serotonergically mediated hormone release in depressed patients treated with CMI. Methods Nine physically healthy patients (four men, 5 women) hospitalized in the Herzog HospitalEzrath Nashim Depression Treatment Unit gave written informed consent to participate in the study. All fulfilled Research Diagnostic Criteria (Spitzer et al 1978) for major depressive disor !.,-r,endogenous subtype (eight were unipolar, one bipolar) and were free of all medication other than chloral hydrate for nocturnal sedatim~, for at least 2 weeks before the study. Their mean (_+SEM) age was 48.8 _+ 2.79 years. Their Hamilton Depression Rating Scale (HDRS-21 item) score (Guy 1976) was 24.8 _ 3.41. All the patients were treated with CMI (t.i.d. dosage schedule) which was increased over a period of 7-10 days to a minimum dose of 175 mg/day (mean 203.1 _+ 11.04, range 175-250 mg/day) and then continued at thi,~; dosage for a further 3 weeks. Four patients manifested little or no clinical change (HDRS |8 _ 2.64) whereas five patients were remitted or had significantly improved (HDRS 6.4 _+ 1.2). All then commenced treatment with Li (b.i.d. schedule) which was increased to maintain a serum level of
Effect of Drugs on Fenfluramine Hormone Release
BIOL PSYCHIATRY 1992;3h975-983
9"]7
0.63 4. 0.04 mmol (range 0.5-0.9 mmol). Clinical justification for the addition of Li was either as a potential augmenter in the nonimproved patients or as a p~z]ude to maintenance therapy in those who had remitted. Two of the patients who had not improved on CMI alone responded to Li supplementation (HDRS change >50%). Fenfluramine challenge tests were performed in the drug-free state prior to conLrnencing CMI treatment and after 4 weeks of CMI. Three weeks after Li was added to the treatment regimen a further test was performed. Fenfluramine (60 mg orally) and placebo challenges were administered after a 12-hr fast (including no medication) under random assignment, double-blind conditions as previously described (Lerer et al 1988; Shap/ra et al 1989) and were separated by at least 48 hr. At 8:00 AM on the challenge day, an intravenous catheter was inserted in a forearm vein and was kept patent by a heparin lock. Thirty min later, the first baseline blood sample was drawn, followed after 15 min by the second baseline sample. The challenge medication was then administered. Blood sampling continued hourly for the next 6 hr. Uniform, nontryptophan-contalning snacks were permitted 1 and 4 hr following administration of the medication (after blood sampling), and locomotion was restricted as much as possible. Blood samples were. drawn into EDTAtreated tubes which were kept in an icewater bath until completion of the procedure. Plasma was extracted from the samples and stored at -80°C until assayed. All samples from a particular subject were assayed on a single day. Prolactin and cortisol levels were determined using commercially available radioimmunoassay kits as previously described (Lerer et al 1988; Shapira et al 1992). For prolactin, interassay variation was 5.6% and intraassay variation 6.2%; for cortisol 5.2% and 3.5%, respectively. Serum Li levels were determined by flame photometry. Data were analyzed by analysis of variance (ANOVA) with repeated measures and paired t-tests (two-tailed). Correlations were computed by the Pearson correlation coefficient. Values are presented as mean -+ SEM. For basehne values, means of the - 15 and 0 min samples were used as these did not differ significantly. One subject (a nonresponder to CMI and CMI + Li) was omitted from the analysis of the hormone data because her pretreatment baseline PRL levels were more than 3 SDs higher than the group mean. Results
Figure 1 shows the effect of CMI treatment (4 weeks) and CMI + Li (3 weeks) on the PRL response to fenfluramine (upper panel) and placebo challenge (lower panel). Results are presented as delta values (minus baseline) for each hr as baseline prolactin levels increased significantly during CMI and CMI 4- Li treatment (pretreatment 6.01 _+ 0.61; post-CMI 9.77 _+ 0.81 ng/ml, paired t - 4.18, p - 0.0008; post-Li + CMI 13.42 4. 1.69 ng/ml, t - 4.61, p - 0.0004) CMI treatment significantly increased PRL release after fenfluramine challenge ( F - 10.3; df 1,88; p = 0.01). Maximal enhancement by CMI treatment of PRL release was observed at the 4- and 5-hr time points. There was no evident trend for PRL release following fenfluramine to be more enhanced in the patients who responded to CMI treatment compared with those who did not. "ihere was no effect of CMI or CMI + Li on PRL levels following placebo challenge. Addition of Li resulted in slightly lower PRL levels in response to fenfluramine and a loss of significance versus pretreatment values (when comparing the full time course of the hormone responses). Figure 2 shows peak minus baseline PRL response to fenfluramine and placebo. On this measure, the response to fenfluramine was greater than
978
B. Shapira et al
BIOL PSYCHIATRY 1992 ;31:975-983
FENFLURAMINE CHALLENGE
12
10 8-
(*) o PRETREATMENT B CLOMIPRAMINE
~
~"
/)lT
CLOMIP. + LITHIUM
/
",..T T
........
T T ..........v"" .~.,, /...... ~...
Z
......... /
n
0 -2 1
E ¢m i¢
3
4
5
6
PLACEBO CHALLENGE i
12
10-
2
0
PRETREATMENT
B
CLOMIPRAMINE
v
CLOMIP. + LITHIUM
Z m
I-
¢j
_J 0 (1.
~ ~ ,,"~ r~''''
'"°.~T
-2 1
2
3
TIME
4
5
6
(hours)
Figure 1. Effect of CMI and CMI + Li on PRL response to fenfluramine and placebo challenge in patients with endogenous major depressive disorder. Bars show SEM. PRL levels (minus baseline) are significantly enhanced by CM! (F = 10.3; df 1,88; p - 0.01) but not CMI + Li. * p = 0.01, (*) p - 0.09 (versus corresponding time point pretreatment).
pretreatment (F -- 3.58; df 2, 7; p = 0.05) in both the CMI (p -- 0.04) and CMI + Li (p - 0.03) conditions but did not show further enhancement by Li compared with CMI alone. There was no trend for CMI responders and CMI nonresponders to differ in this regard. In view of a possible effect of baseline prolactin levels on the prolactin response to fenfluramine, correlations were examined under all three treatment conditions. In the pretreatment, drug-free state (r = - 0 . 1 0 ; p - 0.8) and during CMI administration
Effect of Drugs on Fenfluramine Hormone Release
BIOL PSYCHIATRY 1992;31:975-983
979
E gz
I PRETREATMENT ~ CLOMIPRAMINE
lS-
z m
IO
~
12
'A"
CLOMIP.+ LITHIUM
..I
O er ttl Z __1 ILl O~ Ill m
C~
"r
.
Figure 2. Peak minus baseline values for PRL response to fenfluramine and placebo challenge--pretreatment, during CMI administration and after addition of Li. Bars show SEM. * p - 0.04 (for CMI), p 0.03 (for CMI + Li) versus pretreatment.
Z m
PLACEBO
FENFLURAM!NE
ILl
L
(r ffi 0.50; p ffi 0.2), no relationship was observed. During CMI + Li administration, however, there was a significant inverse relationship (r ffi -0.84; p = 0.008). Unlike PRL, baseline cortisol levels were not significantly altered by CMI treatment. There was no relationship between baseline cortisol and baseline PRL levels nor between baseline cortisol and peak or peak minus baseline PRL levels in the drug-free state or during either of the treatment conditions. As shown in Figure 3, fenfluramine challenge did not alter cortisol levels relative to the effect of placebo. There was no significant effect of treatment with CMI or CMI + Li on cortisol response to the challenge agent.
Discussion The present timings extend our earlier report (Shapira et al 1989) which showeci that 3 weeks of treatment with imipramine significantly enhanced the PRL response to fenfluramine but not placebo challenge. The results complement a number of other studies which used the PRL response to L-tryptophan challenge as a measure of serotonergic responsitivity and also demonstrated enhancement after a variety of antidepressant drugs in normal subjects (Cowen et al 1986) as well as in depressed patients (Charney et al 1984; Price et al 198%). We have also shown that treatment of depressed patients with ECT results in a significant enhancement of fenfluramine-induced PRL release (Shapira et ai 1992). The absence in our stud), of a Li effect is inconsistent with the findings of one group but is partially supported by those of another (although both these groups used a different serotonergic challenge and the results are not directly comparable). Glue et al (1986) found that both short- (4 days) and long-term (20 days) administration of Li to normal volunteers significantly enhanced the PRL response to L-tryptophan challenge. In depressed patients resistant to treatment with TCA, addition of Li for 20 days also resulted
980
B. Shapira et al
BIOL PSYCHIATRY
1992;31:975-983
FENFLURAMINE CHALLENGE 12 o PRETREATMENT 10
m CLOMIPRAMINE CLOMIP. + LITHIUM
E w
I=
T
.J O 03 i I--er O O
.
~
,,'"
..."
; ........ Z & / / / /J
1J,
-2
I
2
3
4
5
6
PLACEBO CHALLENGE
12
o PRETREATMENT 10-
g CLOMIPRAMINE .
O) ¢:
v CLOMIP, + LITHIUM
.
.,J 0 03 m I-, tr 0 (,,1
,
''''''''''',
I, b
.
-2
f 1
2
3
TIME
4
5
6
(hours)
Figure 3. Effect of CMI and CMI + Li on cortisol levels (minus baseline) following fenfluramine and placebo administration. Bars show SEM. No differences statistically significant at p < 0.1.
in enhancement of the response compared with the TCA-only condition (Cowen et al 1989). In contrast, Price et al (1986) found that although short-term Li administration increased the PRL response to L-tryptophan in the patients they examined, longer-term Li administration did not. Various factors that may have influenced our results should be noted. One consideration is the small number of subjects, although the absence of any trend towards enhancement by Li renders this possibility less likely. A second possibility is that this result may be peculiar to CMI which, unlike other TCA, was found to increase baseline PRL levels.
Effect of Drugs on Fenfluramine Hormone Release
elOLPS¥Ct;~Y 1992;31:975-983
981
This effect of CMI has been previously reported (Zohar et al 1988). Baseline PRL levels increased further after addition of Li to the treatment regimen. This effe~rt could have been d,~e to a continued and cumulative action of CMI, althoagh in the absence of a group of subjects treated for 7 weeks with Cg~ alone, this possibility is speculative. However, Li has not been shown to raise baseline prolactin levds in studies of this type (e.g., Glue et al 1986; Price et al 1989b). Whatever mechanism might have been operative in raising baseline prolactin levels, it is noteworthy that during the fenfluraraine, challenge conducted under CMI + Li adr.ainistration, a significant inverse correlation between baseline prolactin level and peak minus baseline prolactin response to fenfluramine was observed. This might well have masked an effect of Li to enhance fenflur,~mine-induced prolactin release. Finally, the fact that the subject group included patients who had responded to CMI when Li was added, as well as those who were CMI-refractory, must also be considered. In this regard, the present sample differs from the TCA-refractory sample studied by Cowen et al (1989~ in groom PRL response to tryptophan was enhanced
by t,i. Cortisol rrsponses to fenfluramine were not significantly different from those after placebo and were not altered by CMI or CMI plus Li administration. This is consistent with our own previous observations (Lerer et al 1988) and those of other authors (Mitchell and Smythe 1990).-egarding fenfluramine effects on cortisol release, (It is, however, possible that a stimulatory effect of fenfluramine on cortisol release might be better demonstrated in the early evening when baseline cortisol levels are lower and more stable than in the morning.) We also found no relationship between baseline cortisol and peak or peak minus baseline PRL responses to fenflur~Tdne. Such a relationship has been reported by Mitchell and Smythe (1990) who found that blunted PRL responses to fenfluramine in depressed patients were dependent on baseline cortisol levels. However, the subjects in the present study were all endogenously depressed and the relationship between baseline cortisol levels and PRL response to fenfluramine was not found by Mitchell et al (1990) to be significant in this subgroup of patients. In conclusion, the present report provides additional evidence that serotonergically mediated PRL release is enhanced by TCA treatment. Together with studies employing L-tryptophan challenge, it suggests that antidepressant drugs as well as ECT enhance central 5-HT function in depressed patients. This effect was present in patients who responded to treatment as well as those who did not, and must thus be considered a necessary rather than a sufficient concomitant of antidepressant response, as has been previously noted (Price et al 1989a). The precise receptor mechanisms involved in the enhancement by antidepressant treatment of serotonergically mediated hormone release as well as the role of other neurotransmitter systems in the antidepressant response, remains to be elucidated. The failure of Li to further augment fenfluramine-induced PRL release requires reevaluation, taking into account the methodological considerations noted. Supported in part by NIMH Grant #MH-43873.
References Charney DS, Heninger GR, Steinberg DE (1984): Serotonin function and mechanism of action of antidepressant treatment: Effects of aitriptyline and desipramine. Arch Gen Psychiatry 41:359365. Coccaro EF, Siever LJ, Klar HM, et al (1989): Serotonergic studies in patients with affective and
982
BIOL PSYCHIATRY 1992;31:975--983
B. Shapira et al
personality disorders: Correlates with suicidal and impulsive aggressive behaviour. Arch Gen Psychiatry 46:587-599. Cowen PJ, Charig EM (1987): Neuroendocrine responses to intravenous tryptophan in major depression. Arch Gen Psychiatry 44:958-966. Cowen PJ, Geaney DP, Schachter M, et al (1986): Desipramine treatment in normal subjects: Effects on neur~endocrine responses to tryptophan and on platelet serotonin (5-HT)-related receptors. Arch Gen Psychiatry 43:61-67. Cowen PJ, McCance SL, Cohen PR, Julier DL (1989): Lithium increases 5-HT mediated neuroendocrine responses in tficyclic resistant depressions. Psychopharmacology 99:230-232. De Montigny C, Blier P (1984): Effects of antidepressant treatments on 5-HT neurotransmission: Electrophysiological and clinical studies. In Usden E (ed), Advances in Biochemical Psychopharmacology, vol. 39: Frontiers in Biochemical and Pharmacological Research in Depression. New York: Raven Press. De Montigny C, Cournoyer G, Morisette R, Langlois R, Caille G (1983): Lithium carbonate addition in tficyclic antidepressant-resistant unipolar depression: Correlations with the neurobiologic actions of tricyclic antidepressant drugs and lithium ion on the serotonin system. Arch Gen Psychiatry 40:1327-1334. Glue PW, Cowen PJ, Nutt DJ, Kolakowska T, Grahame-Smith DG (1986): The effect of lithium on 5-HT-mediated neuroendocrine responses and platelet 5-HT receptors. Psychopharmacology 90:498-502. Guy W (1976): ECDEU Assessment Manual for Psychopharmacology. US Dept. of Health, Educati(~n and Welfare, National Institute of Mental Health, Rockville, MD. Hall H, Ogren SO (1981): Effects of antidepressant drugs on different receptors in the brain. Fur J Pharmacol 70:393-407. Heninger GR, Chamey OJ (1987): Mechanism of action of antidepressant treatments: Implications for the etiology and treatment of affective disorders. In Meltzer HY (ed), Psychopharmacology: The Third Generation of Progress. New York: Raven Press. Heninger GR, Charney DS, Steinberg DE (1984): Serotonergic function in depression: Prolactin response to intravenous tryptophan in depressed patients anti healthy subjects. Arch Gen Psychiatry 41:398-402. Lerer B (1987): Neurochemical and other neurobiological consequences of ECT: Implications for t.hq~,pathogenesis and treatment of affective disorders. In Meitzer HY (ed), Psychopharmacology: Tke Tl:ird Generation of Progress. New York: Raven Press. Lerer B Ran A, Blacker M, et al (1988): Neuroendocrine responses in chronic schizophrenia: Evidence for serotonergic dysfunction. Schizophr Res 1:405-410. Lewis DA, Sherman BM (1984): Serotonergic stimulation of adrenocorticotropin secretion in man. J Cl.in Endocrinol Metab 58:458-462. Mitchel~ P, Smythe G (1990): Hormonal responses to fenfluramine in depressed and control subjects. J Aff.:ct Disord !9:43-56. Mitchell P, Smy:itc G, Parker G, et al (1990): Hormonal responses to fenfluramine in depressive subtyI~es. Br J Psychiatry 157:551-557. Newma~ M, Lerer B (1989): Modulation of second messenger function in rat brain by in ~'ivo alterwtion of receptor sensitivity: Relevance to the mechanism of action of electroconvulsive • ~rapy and antidepressants. Prog Neuropsy~hopharmacol Biol Psychiatry 13:1-30. Price LH, Chamey DS, Delgado PL, et al (1989a~: Effects of disipramine and fluvoxamine treatment of the prolactin response to tryptophan: Serotonergic function and the mechanism of antidepressant action. Arch Gen Psychiatry 46:625-631. Price LH, Charney DS, Delgado PL, Heninger GR (1989b): Lithium treatment and serotoninergic function. Arch Gen Psychiatry 46:13-19. Price LH, Charney DS, Delgado PL, et al (1990): Clinical data on the role of serotonin in the mechanism of action of antidepressant drugs. Y Clin Psychiatry 51 (Suppl 4):44-50.
Effect of Drugs on Penfluramine Hormone Release
BIOLPSYCHIATRY 1992;31:975-983
983
Schou M (1990): Lithium and treatment-resistant depressions: A review. Lithium 1:3-8. Shapira B, Reiss A, Kaiser N, Kindler S, Lever B (1989): Effect of imipramine treatment on the prolactin response to fenfluramine and placebo challenge in depressed patients. J Affect Disord 16:1-4. Shapira B, Lerer B, Kindler S, et al (1992): Enhanced serotonergic r~sponsivity following electroconvulsive therapy in patients with major depression. Br J Psychiatry (in press). Siever LJ, Mutphy DL, Slater S, de la Vega E, Lipper S (1984): Plasma prolactin changes following fenfluramine in depressed patients compared to controls: An evaluation of central serotonergic responsivity in depression. Life Sci 34:1029-1039. Spitzer RL, Endicott J, Robins E (1978): Research Diagnostic Criteria: Rationale and reliability. Arch Gen Psychiatry 35773-782. Weizman A, Mark M, Gil-Ad I, Tyano S, Laron 2 (1988): Plasma cortisol, prolactin, growth hormone and immunoreactive @-endorphin response to fenfluramine challenge in depressed patients. Clin Neurophamacol 11:250-256. Zohar J, Insel TR, Zohar-Kadouch RC, Hill JL, Murphy DL (1988): Serotonergic responsivity in obsessive compulsive disorder. Arch Gen Psychiatvy 45:167-172.
