A Longitudinal Study of Serotonergic Function in Depression Robert N. Golden, M.D., Amy Durr Heine, R.N., M.S.N., R. David Ekstrom, M.A., M.P.H., Joseph M. Bebchuk, M.D., F.R.C.P.C., Martha E. Leatherman, M.D., and James C. Garbutt, M.D. Several reports have described abnormal neuroendocrine compared with matched healthy control subjects, at each responses to serotonergic challenge tests in depression, but time point over the longitudinal course of their illness and few have studied depressed patients followed longitudinally. recovery. Challenge test results in depressed patients did In order to determine whether blunted prolactin responses not change after response to acute desipramine therapy, at to clomipramine challenge is a “state” vs. “trait” marker in the conclusion of the continuation phase of treatment, or depression, we applied this challenge paradigm to 20 while in a medication-free state of remission. Blunted patients with Major Depression prior to treatment and at prolactin response to clomipramine challenge persists in three additional time points following response to depressed patients after recovery from acute illness, and desipramine: at the completion of acute treatment; at the may reflect an underlying biological vulnerability. end of the continuation phase of treatment; and after a [Neuropsychopharmacology 26:653–659, 2002] minimum of three weeks “washout” following the © 2002 American College of Neuropsychopharmacology. discontinuation of treatment. The prolactin response to Published by Elsevier Science Inc. clomipramine challenge was blunted in depressed patients KEY WORDS: Serotonin; Depression; Clomipramine; centrations of serotonin (5-HT), its metabolites, and Desipramine; Challenge test; Prolactin receptors; platelet 5-HT receptors; cerebrospinal fluid concentrations of 5-hydroxyindoleacetic acid; and neu- Several lines of evidence support the hypothesis that roendocrine responses to 5-HT challenge tests (Maes dysregulation in central serotonergic systems plays a and Meltzer 1995). The induction of clinical relapse in role in the pathogenesis of major depression (Maes and following acute tryptophan depletion in remitted de- Meltzer 1995). Studies have reported abnormalities in pressed patients who had received treatment with sero- various indirect indices of central serotonergic function tonin selective reuptake inhibitors (SSRIs) (Delgado et in depressed patients, including post-mortem brain con- al. 1990) suggests that 5-HT may also be involved in the mechanism of action of at least some successful antide- From the Department of Psychiatry, University of North Carolina pressant treatments, and that the correction of an un- School of Medicine, Chapel Hill, North Carolina (RNG, ADH, RDE, derlying 5-HT dysfunction is linked to the initiation JMB, MEL, JCG), and the Clinical Research Unit, Dorothea Dix Hos- pital, Raleigh, North Carolina (JCG). and maintenance of clinical remission. Address correspondence to: Robert N. Golden, M.D., Professor While there is a substantial body of data linking per- and Chair, Department of Psychiatry, University of North Carolina turbations in these measures of 5-HT function to de- School of Medicine, Campus Box # 7160, Chapel Hill, NC 27599- 7160, Tel.: (919) 966-4738, fax: (919) 966-7659, E-mail: rgolden@ pression, it remains unclear whether such findings re- css.unc.edu flect state characteristics of depressive illness, vs. trait Received May 9, 2001; revised September 20, 2001; accepted Octo- characteristics which are indicative of an ongoing, pre- ber 29, 2001. Online publication: 11/2/01 at www.acnp.org/citations/ sumably genetic vulnerability. Most studies compare Npp110201198. patients who are currently depressed to matched NEUROPSYCHOPHARMACOLOGY 2002–VOL. 26, NO. 5 © 2002 American College of Neuropsychopharmacology Published by Elsevier Science Inc. 0893-133X/02/$–see front matter 655 Avenue of the Americas, New York, NY 10010 PII S0893-133X(01)00406-7 654 R.N. Golden et al. NEUROPSYCHOPHARMACOLOGY 2002–VOL. 26, NO. 5 healthy control subjects who have no personal or family direct 5-HT pharmacological effects from the treating history of mood disorders. Differences between these agent. groups of subjects could reflect either state or trait phe- nomena associated with depression. The most useful approach in distinguishing between these would in- METHODS volve longitudinal studies of depressed patients over Subjects the course of treatment and recovery from their illness. Several treatment studies have applied neuroendocrine This study was approved by the University of North challenge tests to depressed patients before and after Carolina School of Medicine Committee for the Protec- antidepressant treatment, and to date, the results have tion of the Rights of Human Subjects. Oral and written been inconclusive. Some reports describe enhancement informed consent was obtained from each subject be- of the characteristically blunted prolactin response to fore enrollment. All potential subjects received a com- 5-HT challenge in depressed patients following treat- prehensive medical evaluation (which included a phys- ment, but this finding is not consistent (O’Keane et al. ical examination, medical history with review of 1992; Shapira et al. 1989, 1992; Leatherman et al. 1993; systems, and clinical laboratory testing for hematologic, Kasper et al. 1990). Others have compared 5-HT neu- hepatic, and endocrine disease), and only those who roendocrine challenge profiles in depressed patients were free of any medical condition that could jeopar- who are receiving antidepressant pharmacotherapy, or dize their safety or complicate the interpretation of the who are in a current state of remission, to those of data were eligible to participate. Female subjects were healthy volunteers or untreated patients (Flory et al. given a pregnancy test immediately prior to the study 1998; Meltzer et al. 1997). For a variety of practical and to insure that those with positive results would be ex- pragmatic reasons, it is difficult to reexamine in a longi- cluded. All subjects received modest financial compen- tudinal patient cohort 5-HT indices over the course of sation for their participation, as approved by our IRB. the acute and continuation phases of their illness, in- Depressed patients were recruited from the clinical cluding a medication-free state of remission. However, services at the Department of Psychiatry at the Univer- this approach offers the greatest opportunity to dissect sity of North Carolina Hospitals and Dorothea Dix Hos- the trait vs. state nature of 5-HT dysregulation in de- pital and from newspaper advertisements. Diagnoses pression. were assigned by a board certified psychiatrist (RNG), The clomipramine (CMI) challenge test meets the based on all available clinical data, including the infor- van Praag et al. (1987) criteria for a valid 5-HT probe: mation generated by a semi-structured diagnostic inter- (1) a dose response relationship exists for the neuroen- view (SCID) performed by a member of the research docrine response to CMI challenge (Golden et al. 1989); team. All depressed patients met DSM-III-R (American (2) CMI challenge does not appear to affect catechola- Psychiatric Association 1987) criteria for Major Depres- minergic systems (Golden et al. 1989); and (3) the neu- sion, and achieved a score of 17 or greater on the roendocrine response to CMI challenge can be blocked Hamilton Depression Rating Scale (HDRS) (Hamilton with a 5-HT functional antagonist (Laakmann et al. 1960), 21 item version, at baseline. 1983). Several studies have described abnormal neu- Healthy control subjects were recruited via printed roendocrine responses to CMI challenge in depressed advertisements distributed throughout the University patients, in particular, blunted prolactin responses of North of Carolina campus and placed in the local (Golden et al. 1990a,b, 1992; Anderson et al. 1992). The and student newspapers. All potential control subjects blunted prolactin response to CMI challenge in underwent careful medical and psychiatric screening, depression does not appear to be attributable to lac- including the administration of a semi-structured diag- totroph dysfunction, since depressed patients with nostic interview (SADS-L) (Spitzer and Endicott 1978). blunted prolactin responses to CMI mount robust pro- Only subjects with no personal history of psychiatric ill- lactin responses to other stimuli, i.e. TRH challenge ness and a negative family history for mood and anxi- (Anderson et al. 1992; Golden et al. 1990a,b). We applied ety disorders and alcoholism were eligible. Control sub- the CMI challenge test in a longitudinal study of de- jects were free of exposure to psychoactive medications pressed patients, comparing them to healthy volunteers and alcohol for a minimum of three weeks prior to at baseline, and then following their CMI challenge test study. profile over the course of treatment and recovery from their acute episode of illness. We chose as our treatment Procedures agent desipramine, a secondary amine tricyclic antide- pressant which is a norepinephrine reuptake inhibitor All depressed patients and healthy control subjects un- with minimal direct pharmacological effects on 5-HT derwent standardized CMI challenge testing, as pre- neurotransmission (Hyttel 1982), in order to avoid con- viously described (Golden et al. 1989), at baseline. founding our study of 5-HT function in depression with Subjects were placed on a standard low monoamine, NEUROPSYCHOPHARMACOLOGY