Journal of Affective Disorders 93 (2006) 239–243 www.elsevier.com/locate/jad

Brief report Controlled trial of D-cycloserine adjuvant therapy for treatment-resistant major depressive disorder ⁎ Uriel Heresco-Levy a, , Daniel C. Javitt b,c, Yovgenia Gelfin a, Elena Gorelik d,e, Marina Bar d,e, Monica Blanaru d,e, Ilana Kremer d,e

a Ezrath Nashim-Herzog Memorial Hospital and Department of Psychiatry, Hadassah Medical School, Hebrew University, Jerusalem, Israel b Nathan S. Kline Institute for Psychiatric Research, Orangeburg, NY, USA c Department of Psychiatry, New York University, NY, USA d Psychiatry Department, Haemek Medical Center, Afula, Israel e Faculty of Medicine, Technion, Haifa, Israel Received 12 July 2005; received in revised form 8 March 2006; accepted 10 March 2006 Available online 4 May 2006

Abstract

Background: Compounds that reduce N-methyl-D-aspartate receptor (NMDAR) function, including NMDAR antagonists and partial agonists at the NMDAR-associated (GLY) site, may act as antidepressants. The drug D-cycloserine (DCS) acts as a partial agonist at the NMDAR-GLY site. Preclinical and clinical data suggest that at dosages ≥100 mg/day DCS acts as a functional NMDAR antagonist and may have antidepressant effects. Methods: Twenty-two treatment resistant major patients participated in a double-blind, placebo-controlled 6-week crossover trial with 250 mg/day DCS added to their ongoing antidepressant medications. Results: DCS treatment was well tolerated and resulted in symptom reductions. However, biweekly-performed clinical assessments, including the Hamilton Depression Rating Scale, Hamilton Rating Scale for Anxiety and Zung Self-Rating Depression Scale did not reveal statistically significant therapeutic advantages of DCS vs. placebo adjuvant treatment. Limitations: Small sample, uneven treatment resistance criteria across subjects. The exposure to DCS (dose/length of treatment) may not have been sufficient. Conclusions: This exploratory study represents the first attempt to assess the effects of a NMDAR-GLY site partial agonist in depression treatment. The findings and limitations of this study should be taken into account in the planning of future clinical trials with NMDAR modulators in depression. © 2006 Elsevier B.V. All rights reserved.

Keywords: NMDAR; Depression treatment; D-cycloserine

1. Introduction

Despite the improvement in pharmacological treat- ⁎ Corresponding author. Psychiatry Division, Ezrath Nashim- Herzog Memorial Hospital, P.O. Box 3900, Jerusalem 91035, Israel. ment options, depression remains a chronic and/or Tel.: +972 2 5316 906; fax: +972 2 6536 075. recurrent condition that is often characterized by persis- E-mail address: [email protected] (U. Heresco-Levy). tence of symptoms. These caveats of presently available

0165-0327/$ - see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.jad.2006.03.004 240 U. Heresco-Levy et al. / Journal of Affective Disorders 93 (2006) 239–243 antidepressant drugs highlight the need for innovative Table 1 Demographic and clinical characteristics of twenty-two major treatment strategies. N-methyl-D-aspartate glutamate depressive disorder patients receiving adjuvant treatment with receptor (NMDAR) antagonists and partial agonists at 250 mg/day D-cycloserine the NMDAR-associated glycine (GLY) site may repre- sent a novel type of antidepressant medications. Glu- Age (y) 56.9±12.4 tamate antagonists acting at the NMDAR protect Male/female (n) 12/10 vulnerable neurons against a variety of insults, including Marital status Single=1, Married=16, Divorced=2, Widowed=3 stress and glucocorticoid-induced damage and increase Additional DSM-IV diagnoses brain-derived neurotrophic factor (BDNF) synthesis and Dependent Personality Disorder 3 neurogenesis in the dentate gyrus. These effects may Years of education 9.8±4.0 serve to enhance and maintain normal synaptic connec- Age at onset of first episode (y) 46.5±17.4 tivity, thereby allowing the chemical signal to reinstate Total number of episodes 3.0±2.4 Total number of hospitalizations 2.8±1.5 normal affective functioning (Manji et al., 2003; Zarate Number of suicide attempts 1.3±0.5 et al., 2002). This type of compounds also exhibit Duration of present episode (wks) 19.8±13.0 antidepressant-like effects in a variety of animal models Baseline HAMD score 27.2±5.2 (Krystal et al., 2002). Trullas and Skolnick (1990) exa- Baseline HAMA score 26.1±6.9 mined functional NMDAR antagonists and found that a Baseline Zung score 61.9±6.9 competitive NMDAR antagonist (2-amino-7-phospho- Values are mean±SD for continuous variables and number of subjects noheptanoic acid [AP-7]), a noncompetitive NMDAR for categorical values. HAMD=Hamilton Depression Rating Scale; HAMA=Hamilton Rating Scale for Anxiety; Zung=Zung Self-Rating antagonist (dizolcipine [MK-801]), and a partial agonist Depression Scale. at the NMDAR-GLY site (1-aminocyclo-propanecar- boxylic acid [ACPC]) mimic the effects of clinically effective antidepressants. Furthermore, down-regulation used DCS, 500 mg/day, with 30 patients of NMDAR-GLY site appears to be a common feature of suffering from various, mainly neurotic, mental dis- current antidepressant medications. Tricyclics, turbances. An overall improvement of mental condition reuptake inhibitors, electroconvulsive stimuli and bupro- in 47% of patients was reported, DCS effects being most pion reduce the potency of GLY to inhibit [3H]-5,7- pronounced against insomnia, muscular tension, an- dichlorkynurenic acid (5,7-DCKA) binding to the GLY orexia and depression. These observations support the site, a change consistent with down-regulation of the concept that DCS may have antidepressive effects when NMDAR function (Krystal et al., 2002). These fin- acting as an NMDAR antagonist. In this pilot study, we dings suggest that (1) adaptive NMDAR changes could investigate the therapeutic potential of augmentation be a feature common to treatment with structurally un- treatment with DCS in treatment resistant major related antidepressants, and (2) modulation of NMDAR- depression. GLY sites activity may be therapeutically beneficial in depression. 2. Methods D-cycloserine (DCS) is a broad spectrum antibiotic (Seromycin) used for over 30 years in tuberculosis and Patients evaluated for the study meet DSM-IV diag- urinary tract infections. In addition, it binds with high nostic criteria for major depressive disorder established affinity to NMDAR having, similarly to ACPC, partial on the basis of semi-structured psychiatric interviews, agonist characteristics at the NMDAR-associated GLY review of all available medical records and confirma- site (Hood et al., 1989; Emmett et al., 1991). Preclinical tion by at least two board-certified psychiatrists. Pa- data support a dose-dependent biphasic DCS effect: at tients could enter the study if they had at least 4 weeks low doses DCS behaves as an agonist whereas at higher of treatment with an antidepressant drug at clinically doses it behaves as an antagonist of NMDAR-mediated adequate doses and a score of at least 18 on the 21-item neurotransmission. It is hypothesized that GLY site Hamilton Depression Rating Scale (HAMD (Hamilton, antagonistic effects of DCS may be obtained at doses in 1960)), with at most mild improvement from onset of excess of 100 mg/day (Millan, 2002). During the 1950s antidepressant treatment. Patients who underwent ECT tuberculosis patients receiving 500–1000 mg/day DCS treatment during the 3 months preceding the study, regimens reported significant psychic improvement change in psychotropic medications doses during the 4 (e.g., increase of appetite and a sense of well-being weeks preceding the study, had a concurrent unstable bordering on euphoria) attributable to the action of the medical or neurological illness, or a history of drug/ drug (Crane, 1959; Epstein et al., 1955). Crane (1961) abuse were excluded. Medical and neurological U. Heresco-Levy et al. / Journal of Affective Disorders 93 (2006) 239–243 241 examinations, clinical laboratory tests and EKG were Clinical assessments were obtained biweekly throug- performed before trial inclusion. The research protocol hout the study using the HAMD, the Hamilton Rating was approved by the appropriate review boards and Scale for Anxiety (HAMA) (Hamilton, 1959), the Zung written informed consent was obtained from subjects Self-Rating Depression Scale (Zung, 1965), the Positive following complete description of the study orally and and Negative Syndrome Scale (PANSS) (Kay et al., in writing. A total of 22 patients entered the study 1987) and the UKU Side Effects Rating Scale (Table 1). All patients have had at least one previous (Lingjaerde et al., 1987), administered by a single unsuccessful pharmacological intervention during the trained research psychiatrist. According to the research current depression episode and the majority of patients protocol, patients requiring medication changes during were receiving, at study entry, multiple and/or the study, as evidenced by appearance of side effects combination treatments to which they did not have a and/or a HAMD total score increase of ≥20% had to be clinically significant response (Table 2). withdrawn from the study. The study had a total length of 16 wk using a random Statistical analyses were performed using the SPSS assignment, double-blind, placebo-controlled, crossover for Windows (SPSS Inc., Chicago IL, USA) computer design. After a 2 week (wks-2 to 0) baseline assessment program. All cited p values are two-tailed, with a period, subjects were randomly allocated, without significance level of 0.05. Prior to statistical analysis all blocking, stratification or other restrictions, to receive variables were tested for normality using the one-sample under double-blind conditions, 250 mg/day DCS or Kolmogorov–Smirnov test (SPSS). No significant placebo for 6 weeks (wks 0–6). Patients then underwent deviations from normality were observed (all p>0.35). a 2-wk experimental treatment washout, followed by 6 wk of crossover treatment (wks 8–14). DCS and 3. Results placebo were administered orally, prior to breakfast, in identically looking capsules prepared by the hospital Of the 22 patients enrolled in the study, 9 were pharmacy. Clinical and research staff, patients and their randomized to receive DCS during the first treatment families were unaware of and could not determine the phase; 13 received placebo. For all subjects symptoms study drug assignment by appearance or otherwise. were stable for at least 2 weeks before initiation of

Table 2 Ongoing psychotropic medications of twenty-two major depressive disorder patients receiving adjuvant treatment with 250 mg/day D-cycloserine Patient Drug 1 Drug 2 Drug 3 number Brand name Dose Length of Brand name Dose Length of Brand name Dose Length of (mg/day) treatment (wks) (mg/day) treatment (wks) (mg/day) treatment (wks) 1 20 18 Clonazepam 1.5 4 2 Mianserin 60 16 10 16 3 150 6 Clonazepam 2 6 4 Imipramine 150 8 Clonazepam 1 26 5 Mianserin 60 12 Clonazepam 4 8 6 Fluoxetine 40 20 50 8 Clonazepam 2 12 7 300 32 Sulpiride 150 22 20 32 8 Mianserin 60 34 Clonazepam 1.5 10 Sulpiride 100 6 9 Mianserin 45 24 0 Mianserin 60 36 50 8 1 12 11 187.5 15 Sulpiride 150 10 Clonazepam 1.5 15 12 150 12 Sulpiride 150 10 Clonazepam 0.5 12 13 Clomipramine 225 40 Clonazepam 6 40 14 Clomipramine 150 16 Chlorpromazine 150 12 Clonazepam 15 12 15 Fluoxetine 20 12 Clonazepam 0.5 6 16 Mianserin 45 30 Clonazepam 0.75 30 17 20 6 Clonazepam 1.5 6 18 Venlafaxine 225 16 Oxazepam 20 6 19 30 14 4 8 20 Paroxetine 40 8 Lorazepam 4 10 21 Paroxetine 40 8 22 Clomipramine 150 16 242 U. Heresco-Levy et al. / Journal of Affective Disorders 93 (2006) 239–243

Table 3 HAMD, HAMA, PANSS and Zung Self-Rating Depression Scale scores during adjuvant treatment with 250 mg/day D-cycloserine or placebo as a function of treatment week Scale Treatment Week within treatment phase LOCF F1 p Assignment Week 0 Week 2 Week 4 Week 6

HAMD D-cycloserine 20.8±9.6 (19) 19.7±10.4 (19) 17.0±8.1 (17) 14.6±7.8 (16) 16.4±9.1 (19) Placebo 24.4±7.6 (20) 22.5±8.4 (19) 20.2±10.5 (18) 17.4±8.4 (16) 21.3±11.0 (20) 0.56 0.51 HAMA D-cycloserine 19.1±9.9 (19) 18.6±10.2 (19) 14.9±8.9 (17) 15.2±8.2 (16) 16.8±9.5 (19) Placebo 23.2±9.3 (20) 20.7±9.0 (19) 19.8±11.7 (18) 15.8±8.9 (16) 19.9±12.1 (20) 0.011 0.92 Zung D-cycloserine 56.1±10.0 (19) 54.6±11.7 (19) 50.5±11.0 (17) 47.6±11.0 (16) 49.7±11.9 (19) Placebo 58.2±11.6 (20) 56.8±9.9 (19) 54.0±10.9 (18) 52.5±8.9 (16) 55.8±10.5 (20) 2.82 0.10 PANSS Positive D-cycloserine 7.5±1.0 (19) 8.0±2.1 (19) 7.6±1.5 (17) 7.6±1.4 (16) 7.7±1.6 (19) Placebo 7.7±1.3 (20) 7.4±0.9 (18) 7.6±1.5 (18) 7.3±0.6 (16) 7.6±1.4 (20) 1.08 0.31 PANSS Negative D-cycloserine 8.7±2.5 (19) 8.9±3.7 (19) 8.9±3.1 (17) 8.4±2.9 (16) 8.5±2.9 (19) Placebo 8.8±2.6 (20) 8.5±2.5 (18) 8.5±2.3 (18) 8.2±2.1 (16) 8.6±2.2 (20) 0.01 0.96 PANSS General D-cycloserine 34.9±7.9 (19) 32.6±11.9 (19) 31.6±7.5 (17) 28.9±7.7 (16) 30.5±8.8 (19) Placebo 37.6±6.0 (20) 36.3±6.8 (18) 35.9±8.8 (18) 33.3±7.0 (16) 35.6±8.6 (20) 2.22 0.15 Values are mean±SD(n). HAMD=Hamilton Depression Rating Scale; HAMA=Hamilton Rating Scale for Anxiety; Zung=Zung Self-Rating Depression Scale; PANSS=Positive and Negative Syndrome Scale. LOCF=Last observations carried forward. 1df=1,36 experimental treatment. Fifteen patients completed the a mean of 2.26±7.63 (11.8%), 4.37±6.77 (21.1%) and entire trial. Two patients were withdrawn while 6.42±9.06 (11.4%) units on the HAMD, HAMA and receiving DCS: patient 5 at study wk 4 due to symptom Zung scales, respectively, under the experimental exacerbation and patient 16 at wk 12 due to medical treatment regimen. However, this outcome did not differ reasons not related to the experimental treatment. Five significantly from the effects registered during the patients were withdrawn while receiving placebo: placebo treatment phase. A number of limitations of patients 15, 18, 7 and 2 at wks 1, 4, 6 and 10 this exploratory investigation may have contributed to respectively, due to symptom exacerbation, and patient this outcome. The treatment resistance criteria among 20 at wk 8 due to medical reasons. Overall, 20 patients subjects were uneven. Ongoing antidepressant treat- entered the placebo treatment phase and 19 the DCS ments and treatment periods were heterogenous and phase. some patients (1, 15, 17) were receiving minimal Data were analyzed using separate ANCOVA for antidepressant doses. Furthermore, only 16 patients each measure, with LOCF end treatment values serving completed the DCS treatment phase and the 6 wk active as primary dependent variable and pretreatment (week treatment period may have been too short to evidence 0) scores serving as covariates. Treatment arm was the full DCS treatment potential. coded as a between-subject variable. No significant The issue of DCS dose range to be explored in between-treatment differences were observed for any depression requires further consideration. Dose corre- of the variables studied, although a trend (p=0.1) spondences between agonistic/antagonistic DCS effects toward improvement was registered for the Zung scale in rodents and humans are poorly understood (Millan, (Table 3). Similar results were obtained if analyses 2002). Moreover, the mood elevating effects registered were performed using data from completers only. with DCS in tuberculosis patients were usually evident Overall, DCS treatment was well tolerated through- at >500 mg/day doses. Thus, the 250 mg/day dose out the study and no significant DCS-induced clinical or employed in this pilot study may not have been laboratory side effects were registered. sufficient. Recently, no side effects were reported when a 500 mg DCS acute regimen was used with 4. Discussion phobic individuals (Ressler et al., 2004). However, the use of “high dose” DCS may carry the potential danger This study represents to our knowledge the first of psychotomimetic effects similar to those induced by attempt to assess the effects of a NMDAR-GLY site NMDAR noncompetitive antagonists (Heresco-Levy, partial agonist in the treatment of major depression. We 2002). failed to find evidence that adjuvant treatment with Ultimately, a conceptual paradox should be noted. If, 250 mg/day DCS might be beneficial. Patients improved as hypothesized (Skolnick, 1999; Krystal et al., 2002) U. Heresco-Levy et al. / Journal of Affective Disorders 93 (2006) 239–243 243 antidepressant drugs directly or indirectly reduce Hood, W.F., Compton, R.P., Monahan, J.B., 1989. DCS: a for NMDAR function, the addition of DCS may not the N-methyl-D-aspartate coupled glycine receptor has partial agonist characteristics. Neurosci. Lett. 98, 91–95. further down-regulate NMDAR-mediated neurotrans- Kay, S.R., Opler, L.A., Fiszbein, A., 1987. Positive and Negative mission which may be already reduced due to the Syndrome Scale (PANSS) Rating Manual. Social and Behavioral ongoing antidepressant treatment. From this perspec- Sciences Documents, San Rafael, CA. tive, the assessment of DCS as antidepressant mono- Krystal, J.H., Sanacora, G., Blumberg, H., Anand, A., Charney, D.S., therapy, rather than adjuvant treatment, may be highly Marek, G., Epperson, C.N., Goddard, A., Mason, G.F., 2002. Glutamate and GABA systems as targets for novel antidepressant informative. and mood-stabilizing treatments. Mol. Psychiatry 7 (suppl. 1), 571–580. Acknowledgments Lingjaerde, O., Ahlfors, V.G., Beck, P., Dencker, S.J., Elgen, K., 1987. The UKU side effect rating scale. Acta Psychiatr. Scand. 384, – This research was supported by a grant from the 85 95 (suppl.). Manji, H.K., Quiroz, J.A., Sporn, J., Payne, J.L., Denicoff, K., Gray, National Alliance for Research on Schizophrenia and N.A., Zarate Jr., C.A., Charney, D.S., 2003. Enhancing neuronal Depression (NARSAD), NY, USA (U. H-L). plasticity and cellular resilience to develop novel, improved therapeutics for difficult-to-treat depression. Biol. Psychiatry 53, 707–742. References Millan, M.M., 2002. N-methyl-D-aspartate receptor-coupled glycine B receptors in the pathogenesis and treatment of schizophrenia: a Crane, G.E., 1959. Cycloserine as an antidepressant agent. Am. J. critical review. Curr. Drug Targets, CNS Neurol. Disord. 1, Psychiatry 115, 1025–1026. 191–213. Crane, G.E., 1961. The psychotropic effects of cycloserine: a new use Ressler, K.J., Rothbaum, R.O., Tannenbaum, L., Anderson, P., Graap, for an antibiotic. Compr. Psychiatry 2, 51–59. K., Zimand, E., Hodges, L., Davis, M., 2004. Cognitive enhancers Emmett, M., Mick, S., Cler, J., 1991. Actions of D-cycloserine at the as adjuvants to psychotherapy. Arch. Gen. Psychiatry 61, N-methyl-D-aspartate-associated glycine receptor site in vivo. 1136–1143. Neuropharmacology 30, 1167–1171. Skolnick, P., 1999. Antidepressants for the new millenium. Eur. J. Epstein, I.G., Nair, K.G.S., Boy, L.G., 1955. The treatment of human Pharmacol. 375, 31–40. tuberculosis with cycloserine. Prog. Rep. Dis. Chest 29, 241–257. Trullas, R., Skolnick, P., 1990. Functional antagonists at the NMDA Hamilton, M., 1959. The assessment of anxiety states by rating. Br. J. receptor complex exhibit antidepressant actions. Eur. J. Pharmacol. Med. Psychol. 175, 17–22. 185, 1–10. Hamilton, M., 1960. A rating scale for depression. J. Neurol. Zarate, C.A., Quiroz, J., Payne, J., Manji, H.K., 2002. Modulators of Neurosurg. Psychiatry 23, 56–62. glutamatergic system: implications for the development of Heresco-Levy, U., 2002. Schizophrenia and other psychotic disorders: improved therapeutics in mood disorders. Psychopharmacol. transmitter systems. In: D'Haenen, H., den Boer, J.A., Bull. 4, 35–83. Willner, P. (Eds.), Biological Psychiatry. John Wiley and Sons, Zung, W.W.K., 1965. A self-rating depression scale. Arch. Gen. London, UK, pp. 587–600. Psychiatry 12, 63–70.