Efficacy and Safety of 30 Mg/D and 45 Mg/D Nemifitide Compared To

ARTICLE International Journal of Neuropsychopharmacology (2006), 9, 517–528. Copyright f 2005 CINP doi:10.1017/S1461145705006139 Eﬃcacy and safety of 30 mg/d and 45 mg/d nemiﬁtide compared to placebo in major depressive disorder

Stuart A. Montgomery1, John P. Feighner2, Lev Sverdlov2, Ram K. Shrivastava3, Lynn A. Cunningham4, Ari Kiev5, Joseph Hlavka2 and George Tonelli Jr.2 1 Imperial College School of Medicine, London, UK 2 Tetragenex Pharmaceuticals, Inc., Park Ridge, NJ, USA 3 East Side Comprehensive Medical Service, New York, NY, USA 4 Vine Street Clinical Research, Springﬁeld, IL, USA Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 5 Social Psychiatry Research Institute, New York, NY, USA

Abstract

Nemifitide is a novel pentapeptide antidepressant, which appears to be effective in the treatment of major depressive disorder (MDD). In the present study 81 patients with MDD, DSM-IV criteria were random- ized following a 1-wk screening period to receive 30 mg/d nemifitide, 45 mg/d nemifitide or placebo in a 6-wk double-blind, multicentre, outpatient efficacy study. Nemifitide or placebo was delivered by subcutaneous injection for 2 wk daily for 5 days (Monday to Friday) in the first 2 wk and patients were followed up for a further 4 wk. The primary efficacy measure was the change from baseline on the Montgomery–Asberg Depression Rating Scale. Secondary measures included the 17-item Hamilton Psychiatric Rating Scale for Depression (HAMD), the CGI severity and improvement scale and the Carroll Self-Rating Scale for Depression. This proof-of-principle study demonstrated a statistically significant superiority of the 45-mg/d dose vs. placebo at the time-point of peak effect (1 wk after the end of treatment). There appeared to be a greater effect with the 45 mg/d nemifitide dose than with 30 mg/d. An additional exploratory analysis by stratification of all patients by severity above and below or equal to the median baseline HAMD score of 22 showed a higher percentage of responders for both doses of nemifitide with statistical separation from placebo for patients with baseline HAMD score of >22 (above the median). There was no significant difference among treatment groups for patients with baseline HAMD score of f22. Nemifitide showed a good tolerability and safety profile. There were no dropouts due to adverse events, and the incidence of side-effects with nemifitide was comparable with that of placebo. Received 20 March 2005; Reviewed 20 May 2005; Revised 29 June 2005; Accepted 29 July 2005; First published online 19 December 2005

Key words: Antidepressant, clinical trial, eﬃcacy, nemiﬁtide, safety.

Introduction the last three decades, antidepressant drugs have substantially changed the therapy of depression and Major depression is a serious psychiatric illness with a a variety of pharmacological agents have proved very high social and economic burden (APA, 1994; to be effective in the treatment of major depression Greenberg et al., 1993). Depression affects people of (Frazer, 1997; Montgomery, 1999). The difference in both sexes (with prevalence in women from 1.7 to 2.7 the mechanism of action between current and future times greater than in men) and a large age range, from antidepressants is under scientific investigation early adolescence to lifelong risk (Kessler et al., 1993; (Richelson, 2003; Stahl and Grady, 2003). Numerous Sambunaris et al., 1997; Weissman et al., 1996). During placebo-controlled, double-blind studies attest to the fact that antidepressant drugs are effective, but may Address for correspondence : Dr Lev Sverdlov, Tetragenex require up to 4–6 wk before meaningful or optimal Pharmaceuticals, Inc., 1 Maynard Drive, Suite 205, Park Ridge, NJ clinical effect is achieved. Side-effects of the newer 07656, USA. Tel. : 201-505-1300 Fax : 201-505-1501 antidepressants are generally milder than older E-mail : [email protected] tricyclic antidepressants (TCAs) and monoamine 518 S. A. Montgomery et al. oxidase inhibitors (MAOIs), but many patients a stable reproducible clinical effect of re-treatment with discontinue treatment before full therapeutic effect is nemifitide for most patients who relapsed. Mean achieved (Bull et al., 2002; Eaddy et al., 2003; Masand, duration between re-treatments of responders in the 2004). An antidepressant providing rapid sympto- extension study was 3.3 months, which was similar to matic relief with minimal side-effects would be the duration recorded in responders in the initial study. valuable not only in terms of immediate benefit to the This paper presents the results of a new proof- patients, but also in overall management (Barranco of-principle study with nemifitide comparing doses of et al., 1979; Rothshild and Duval, 2003; Stahl, 2003). 30 mg/d and 45 mg/d vs. placebo for the treatment of Nemifitide (4-fluoro-L-phenylalanyl-trans-4-hydroxy- DSM-IV major depression. L-prolyl-L-arginylglycyl-L-tryptophanamide ditrifluoro- acetate; Hlavka et al., 1997) is the first drug developed Method from a new class of small-chain peptide anti-

depressants. Recent clinical results (Feighner et al., Study design Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 2000, 2001a,b, 2003; Feighner and Sverdlov, 2002) This study was a double-blind, placebo-controlled, suggest that nemifitide may have significant advan- parallel-group comparison in depressed outpatients tages over current therapies including a rapid onset of at three centres in the USA: Vine Street Clinical response (3–5 doses) with minimal side-effects and Research, Springfield, IL; Social Psychiatry Research that it represents a potential new treatment paradigm Institute, New York, NY; and East Side Comprehensive for major depression. Nemifitide was developed Medical Service, New York, NY. The study design is by substantial chemical modification of a previously presented in Figure 1. Following a 1-wk assessment known tripeptide MIF-1 that demonstrated anti- period (baseline) patients were randomly assigned depressant activity in animal models (Pignatiello et al., with a ratio of 1:1:1 to one of three treatment 1989) and in clinical studies with few apparent side- groups: 30 mg/d nemifitide, 45 mg/d nemifitide effects (Ehrensing and Kastin, 1978, 1994). or placebo. The doses were selected on the basis In a series of studies to elucidate the mechanism of of pharmacokinetic data indicating that the majority of action of nemifitide (Feighner, 2003) it has been shown patients receiving nemifitide at a dose of o30 mg/d that, following subcutaneous administration in rats, should achieve the minimum projected therapeutic nemifitide quickly crossed the blood–brain barrier and concentration (Feighner, 2003) and were approximately is found at nanomolar concentration in the amygdala, twice those used in early clinical studies (Feighner et al., frontal cortex and hippocampus, sites that express 2000, 2001a). A fixed dose of nemifitide or identical high concentrations of receptors involved with lactose placebo was administered subcutaneously for the pathophysiology of depression. In-vitro studies 10 d (two 5-d treatment cycles from Monday to Friday, indicate that both nemifitide and its active metabolite separated by two non-treatment days, Saturday (M1) bind at micromolar concentrations to several and Sunday). After the 10-d treatment patients were receptors including the 5-HT2 , 5-HT2 , melanocortin A C followed up for 4 wk. Evaluation of safety and efficacy MC4, MC5 and bombesin. In-vivo studies, for example was carried out after 3, 7, 10 d treatment and weekly serotonin releaser d-fenfluramine challenge, indicate for the following 4 wk. All potential participants in the that the serotonergic pathway is involved in the mode study provided written informed consent. The protocol of action of nemifitide by blocking the hyperthermic and study centres were approved by a centralized effect of the 5-HT2 agonist DOI indicates 5-HT2 A A institutional review board. antagonist properties. Early pilot clinical in-patient (Feighner et al., 2000) Patients and outpatient (Feighner et al., 2001a) studies reported statistically significant efficacy compared with placebo Male or female patients aged at least 18 yr who ful- for low doses of nemifitide (11–20 mg/d) in patients filled DSM-IV criteria for major unipolar depressive with plasma drug concentration above a minimum disorder with a severity of depression score of o25 projected therapeutic concentration (measured for on the Montgomery and Asberg, Depression Rating

Cmax or at 1 h after dose). A potential for a rapid onset Scale (MADRS; Montgomery and Asberg, 1979) and a of action was observed with a significant difference Clinical Global Impression of severity (CGI; Guy, 1976) from placebo on key efficacy parameters as early as score of o4 at screening and baseline were eligible to day 5, and the time to peak effect between 1 and 2 wk participate in the study. Female patients of childbearing following treatment. A long-term open-label out- potential were included if they had a negative serum patient extension study (Feighner et al., 2003) reported pregnancy test at screening and were using an Efficacy and safety of nemifitide 519

M - F M T W T F M T W T F M - F M - F M - F M - F M

Placebo Placebo

Nemifitide Nemifitide (30 mg) (30 mg)

Nemifitide Nemifitide (45 mg) (45 mg) Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 5 d of 5 d of Treatment Treatment

Screen

Day -7 Baseline/BeginBaseline/Begin BeginBegin Effect ofof EffectEffect of of Effect ofof to -3 Termination SeriesSeries 1 1 - - SeriesSeries 2 2 - - treatment Treatmenttreatment treatment Termination TreatmentTreatment TreatmentTreatment Day 15 Day 22 Day 29 Day 43 Day 1 Day 8 ±3 d ±3 d ±3 d

Figure 1. Study design (three centres, n=81). acceptable method of birth control for at least 3 baseline visit, and frequency and intensity of the months before screening. Patients with renal, hepatic, therapy is not expected to change. Current treatment endocrine, gastrointestinal or nervous system dis- with L-dopa or other dopaminergic agents was order, history or presence of significant cardiovascular excluded. disease (including significant electrocardiogram abnormalities) or other conditions that could interfere Efficacy measurements with the safety and conduct of this study were The primary efficacy measure was the MADRS excluded. Patients with a history of substance abuse total score change from baseline. Secondary efficacy within a year of the study were excluded. Other measures included the 17-item Hamilton Rating Scale exclusions were significant physical or CNS disease, (HAMD; Hamilton, 1960), CGI [measured for severity history of bipolar I or II disorder, eating disorder, of illness (CGI-S), clinician’s assessment of global obsessive–compulsive disorder, panic disorder, or improvement (CGI-I), and patient’s global impres- seasonal affective disorder, patients who, in the sion] and the 52-item Carroll Self-Rating Scale for judgement of the investigator, represented a signifi- Depression (CSRSD; Carroll et al., 1981). Total cant risk for suicide or had been refractory to anti- scores, absolute and percentage change from baseline depressant medical treatment, borderline, histrionic, responder rates were also analysed. Responders were schizotypal, or anti-social personality disorders. Use defined as a decrease in MADRS score of at least 50% of antipsychotic or investigational drugs within the of baseline value and on the CGI-I as a score of 1 or last 30 d, ECT within 3 months prior to screening, 1 2 (much or very much improved). Remission was fluoxetine hydrochloride (Prozac ) within 4 wk of defined as a score of <8 for the 17-item HAMD (Frank screening, MAOIs or St. John’s Wort within 14 d prior et al., 1991) and a score of <10 for the MADRS (Hawley to screening, or other antidepressants within 7 d prior et al., 2002; Zimmerman and Posternak, 2004). to screening were excluded. Regular use of benzodia- zepines (4 or more per week) in the last 3 months was Safety measurements excluded and all benzodiazepine use had to be dis- continued before entry to the study for a period Safety was assessed by monitoring the frequency and equivalent to 3–5 half-lives of the respective benzo- severity of adverse events (AEs), evaluating the injec- diazepine. Depression-oriented psychotherapeutic tion sites for primary local irritation (erythema and treatment (e.g. cognitive behavioural therapy or inter- oedema), concomitant medication, routine laboratory personal therapy) was allowed during the study if testing (clinical chemistry, haematology and urinal- it has been ongoing for at least 12 wk prior to the ysis), vital signs, general physical examinations, and 520 S. A. Montgomery et al.

ECG parameters. All new or continuing complications of analysis because, as a secondary eﬃcacy variable, it a patient’s medical condition (observed by site personnel was less likely to be aﬀected by the inclusion criteria or volunteered by patient, spontaneously, or in response and could have a greater range of values for baseline. to query) that were not present at screening were Statistical analysis of safety data was performed for recorded as AEs at each visit during the study. Any the ITT population. Treatment group comparisons medical condition that remained unchanged or were made with descriptive statistics (vital signs, improved after screening was not recorded as an AE at clinical chemistry and haematology, some items for subsequent visits. However, if there was a worsening of a urinalysis) or on the basis of the incidence rates (AEs, patient’s medical condition that was present at screening, local irritation, concomitant medication, urinalysis, then it was also considered as a new AE and reported. general physical examinations, abnormalities for ECG parameters). Change for clinical laboratory evalu- ations were examined with descriptive statistics and Statistical analysis

shift tables. Where appropriate, Fisher’s exact test was Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 All clinical data were managed according to good used to analyse the differences in the incidence rates 1 clinical practice and analysis was performed using SAS among three treatment groups. version 8 statistical analysis system software (SAS Institute Inc., Cary, NC, USA). All hypotheses, unless Results otherwise noted, assumed a significant level of 0.05 for two-tailed tests. Because of the small sample size where Of the 81 patients enrolled in the study (ITT) and a significant difference was not observed exploratory randomly assigned to treatment (28 in the 30-mg/d analysis for trends and non-parametric statistics were nemifitide group, 27 in the 45-mg/d nemifitide group used, a significant level of 0.10 was also considered. and 26 in the placebo group), 78 (96.3%) completed the Demographic and other baseline continuous vari- treatment phase and did not have any protocol viol- ables were analysed using ANOVA with the treatment ations (PP). Three patients (3.7%) did not complete group as a factor. The x2 test or non-parametric dosing (withdrew consent), receiving only 1–7 of Fisher’s exact test were used for categorical variables. 10 planned treatment doses (two patients from the The primary analysis of efficacy data was per- 45-mg/d treatment group and one subject from formed for the intention-to-treat (ITT) population, the placebo group). Forty-seven patients of the 81 which included all patients who received at least enrolled (58.0%) completed the last visit (day 43). one dose of research drug and at least one post- Discontinuation was primarily due to lack of per- randomization assessment. Actual score, change from ceived efficacy and there were no discontinuations for baseline, per cent change from baseline were sum- safety reasons. Table 1 presents the summary of marized for each treatment group by time-point with demographics and baseline characteristics for the ITT descriptive statistics (e.g. n, mean, standard deviation, population. There was a relatively even distribution standard error of mean, median, minimum, maximum of demographic characteristics (age, sex, weight and values and range). Treatment group comparison was medical history) in the three treatment groups. The made with ANOVA model with centre and treatment overall mean and standard error of mean (S.E.M.) groups as factors. When it was appropriate, t test to of scores at screening and baseline at day 1 before compare two treatment groups was used. For the dosing were very similar for both the MADRS (screen responder analysis the Cochran/Mantel–Haenszel 31.3¡0.39, baseline 31.9¡0.43) and HAMD (screen (CMH) x2 test adjusted for centre was used. The sec- 22.4¡0.37, baseline 22.8¡0.41). There were no sig- ondary analysis of efficacy data included all patients nificant differences among the three treatment groups who received at least eight doses and did not have any in the baseline values of any of the severity measures major protocol violations (per protocol population, PP). (ANOVA model: p=0.875 for MADRS, p=0.987 for For both primary and secondary efficacy analysis, the HAMD and p=0.979 for CSRSD). last observation carried forward (LOCF) approach start- The peak effect on the MADRS, the primary efficacy ing at day 11 until day 43 was used for missing data. scale, and on the secondary efficacy scales was A subgroup analysis to examine the effect of base- detected on day 22, 1 wk after the end of treatment. line severity of symptoms for efficacy parameters Table 2 shows the mean scores on the efficacy scales at (Indrayan and Sarmukaddam, 2001; Khan et al., 2002; day 22 and the end of the study (day 43). There was Olfson et al., 1998) was carried out in patients with a numerical advantage for 45-mg/d nemifitide vs. baseline severity above the median and below or equal placebo for the majority of time-points measured as to the median. The HAMD was selected for subgroup the change from baseline on the MADRS, HAMD, and Efficacy and safety of nemifitide 521

Table 1. Demographics, intention-to-treat population

Treatment groups

Placebo 30 mg/d 45 mg /d Overall Parameters Statistics (n=26) (n=28) (n=27) (n=81) p value

Sex Male n (%) 12 (46.2%) 12 (42.9%) 10 (37.0%) 34 (42.0%) 0.792 Female n (%) 14 (53.8%) 16 (57.1%) 17 (63.0%) 47 (58.0%) Race Caucasian n (%) 21 (80.8%) 18 (64.3%) 19 (70.4.0%) 58 (71.6%) 0.744 Hispanic n (%) 2 (7.7%) 3 (10.7%) 2 (7.4%) 7 (8.7%) African-American n (%) 1 (3.8%) 5 (17.9%) 3 (11.1%) 9 (11.1%) Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 Asian n (%) 0 (0.0%) 0 (0.0%) 1 (3.7%) 1 (1.2%) Other n (%) 2 (7.7%) 2 (7.1%) 2 (7.4%) 6 (7.4%)

Age Mean¡S.E.M. 40.5¡2.35 41.2¡2.22 37.4¡2.26 39.7¡1.31 0.444 Median 41.0 40.5 37.0 39.0 Min 19 22 22 19 Max 61 61 65 65 Age by sex Male Mean¡S.E.M. 46.0¡3.62 40.2¡3.88 38.1¡4.63 41.6¡2.32 0.364 Median 48.0 40.5 35.5 40.5 Min 19 22 23 19 Max 61 61 65 65 Female Mean¡S.E.M. 35.9¡2.56 42.0¡2.69 36.9¡2.46 38.3¡1.51 0.212 Median 35.0 40.5 37.0 37.0 Min 19 24 22 19 Max 51 60 55 60

S.E.M., Standard error of mean. Continuous variables were analysed using ANOVA model with treatment group as a factor. For categorical variables, comparability of treatment groups was assessed using the x2 test.

CGI-I but this did not reach statistical significance between nemifitide groups and placebo. In the analy- (Figure 2). On the HAMD and CGI there was a trend sis of remission, defined as a MADRS score of f9, the for an advantage compared with placebo for the 45-mg remission rate on day 43 (end of observation) was dose (p<0.10) on day 22. On the CSRSD (not shown in 33.3% in the 45-mg/d nemifitide group, compared Figure 2) there was a significant advantage for 45 mg with 17.9% in the 30-mg/d nemifitide group, and nemifitide (p<0.05) on days 22, 29 and 43 with best 20.0% in the placebo group. Remission rate, based on effect at day 22. The separation at days 29 and 43 was a HAMD score of f7, was comparable with the less than the peak effect at day 22. remission rate for a MADRS score of f9. An advantage for the 30-mg/d nemifitide dose In the patients treated with 30 mg/d nemifitide compared with placebo was observed on the HAMD there were 50% responders compared with 40% in the and CGI-I at day 22, but not on the MADRS (Figure 2). placebo group and the remission rate was 17.9% In the responder analysis, defined as at least 50% compared with 20% in the placebo group. reduction in scale score, on the HAMD at day 22, 52% Figure 3 shows the results of stratification analysis of patients responded on the 45-mg dose vs. 39.1% on for patients with baseline HAMD score above the placebo. Based on the responder definition (CGI-I f2), overall baseline median (HAMD >22) and below or 60% of patients responded in the 45-mg/d nemifitide equal to the baseline median (HAMD f22). Patients group, compared with 40% in the placebo group, 50% with more severe depression scores at baseline (Figure in the 30-mg/d nemifitide group, and 40% in the 3a) showed a greater separation from placebo in both placebo group. There was no significant difference in doses of nemifitide than the whole population treated percentage of responders measured on the MADRS with nemifitide. At the peak effect (day 22) there was 522 S. A. Montgomery et al.

Table 2. Summary of psychometric scores by treatment groups intention-to-treat population, LOCF from day 11 to day 43

Treatment group/ Baseline (day 1) Day 22 End-point (day 43) psychometric scores Mean (¡S.E.M.) Mean (¡S.E.M.) Mean (¡S.E.M.)

Placebo (n=26) MADRS 31.8 (0.74) 21.3 (2.32) 21.2 (2.46) Responders – 36.0% 36.0% HAMD 22.8 (0.80) 15.9 (1.85) 15.9 (1.93) Responders – 39.1% 39.1% CSRSD 24.8 (1.41) 19.6 (2.52) 19.8 (2.53) Responders – 26.1% 26.1% CGI-I – 3.0 (0.25) 3.0 (0.30)

Responders – 40.0% 44.0% Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 Nemiﬁtide 30 mg/d (n=28) MADRS 32.1 (0.83) 21.1 (2.26) 22.1 (2.31) Responders – 42.9% 25.0% HAMD 22.9 (0.68) 13.6 (1.58) 15.3 (1.66) Responders – 46.2% 30.8% CSRSD 24.7 (1.11) 17.3 (1.94) 18.5 (1.98) Responders – 36.0% 28.0% CGI-I – 2.8 (0.25) 2.9 (0.24) Responders – 50.0% 35.7% Nemiﬁtide 45 mg/d (n=27) MADRS 31.6 (0.68) 16.5 (2.17) 18.4 (2.51) Responders – 41.7% 37.5% HAMD 22.8 (0.66) 11.8 (1.64) 14.1 (1.65) Responders – 52.0% 40.0% CSRSD 25.1 (1.11) 13.6 (1.69) 15.6 (1.87) Responders – 40.0% 40.0% CGI-I – 2.4 (0.26) 2.8 (0.25) Responders – 60.0% 40.0%

S.E.M., Standard error of mean. Peak effect (day 22) is 10 d post-dosing. End of study (day 43) is 31 d post-dosing. Responders in Montgomery–Asberg Depression Rating Scale (MADRS), Hamilton Depression Rating Scale (HAMD) or Carroll Self-Rating Scale for Depression (CSRSD) are the patients with o50% reduction (decrease) from baseline in corresponding scale. Responders for Clinical Global Impression – Improvement (CGI-I) are the patients with score of CGI-I=1 (‘very much improved’) or CGI-I=2 (‘much improved’). a significant difference in the HAMD change from no drug-related serious AEs. A list of reported baseline between patients treated with 45 mg/d incidences occurring in more than one patient is nemifitide vs. placebo (p<0.05) and a trend for those also shown in Table 3. The most frequent side-effects treated with 30 mg/d nemifitide vs. placebo. In (mild or moderate) were headache, dry mouth, contrast, there was no significant difference among nausea, dizziness, injection site reactions, constipation treatment groups for patients with a baseline HAMD and fatigue. There was no statistically or clinically score of f22 (Figure 3b). significant difference in AEs among the three treat- Nemifitide was well tolerated with no evidence ment groups with no indication of any dose effect in of significant AEs. Withdrawal from the study during tolerability with nemifitide. the treatment or follow-up periods was not due to perceived lack of tolerability. The observed AEs were Discussion divided into ‘Non-drug-related’ (NDRAE) and ‘Drug- related’ (DRAE), according to the evaluation made by This placebo-controlled study found significant the blinded clinical investigators (Table 3). There were efficacy 1 wk after discontinuation of treatment for Efficacy and safety of nemifitide 523

0 Nemifitide 30 mg, enrolled (n=28) Nemifitide 45 mg, enrolled (n=27) –3 Placebo, enrolled (n=26)

–6

–9

–12

–15

MADRS, change from baseline –3 1 4 8 11 15 22 29 43

0 Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 –3

–6

–9

–12

HAMD, change from baseline Nemifitide, 45 mg vs. placebo: # –15 –3 1 8 15 22 29 43

3.5

3.0

2.5

CGI – Improvement Nemifitide, 45 mg vs. placebo: # 2.0 –3 1 4 8 11 15 22 29 43 Evaluation day

Figure 2. MADRS (reduction in mean change from baseline), HAMD (reduction in mean change from baseline) and CGI – Improvement (# means pf0.10). nemifitide in a dose of 45 mg/d measured on one of the This placebo-controlled study showed that nemifi- scales used and trends on two others. Bearing in mind tide given at a dose of 45 mg/d was effective com- the small size of the study the results indicate sufficient pared to placebo in the treatment of major depressive separation from placebo to suggest that nemifitide is disorder (MDD) on some, though not all, efficacy an effective antidepressant. For 45 mg nemifitide the measures (p<0.10) 1 wk after discontinuation of difference from placebo at 3 wk was 4.6 on the MADRS treatment. The maximum numerical difference from and 4.1 on the HAMD, which is higher than is nor- placebo in change from baseline for MADRS, HAMD mally seen in pivotal placebo-controlled studies of new and CSRSD at peak effect was more than 4 points for antidepressants and is regarded as a clinically relevant the 45-mg dose that indicates a clinically significant difference. Despite the small numbers a significant difference. difference between 45 mg nemifitide and placebo was The lower dose of 30 mg/d nemifitide was observed on the CSRSD which is based on the HAMD. also numerically better than placebo but the effect A trend was observed on the HAMD and the CGI-I appeared to be smaller, suggesting that there may be a scale but not on the MADRS. The effect was generally dose-response relationship. The study design was lower and less consistent with the 30-mg dose. interesting in that it focused on response, not only 524 S. A. Montgomery et al.

(a) Subjects with HAMD baseline >22 0 Nemifitide 30 mg, enrolled (n = 13) Nemifitide 45 mg, enrolled (n = 13) –3 Placebo, enrolled (n = 14)

–6

–9

–12 Nemifitide, 30 mg vs. placebo: #

HAMD, change from baseline Nemifitide, 45 mg vs. placebo: –15 * –3 1 8 15 22 29 43

(b) Subjects with HAMD baseline р22 Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 0 Nemifitide 30 mg, enrolled (n = 15) Nemifitide 45 mg, enrolled (n = 14) –3 Placebo, enrolled (n = 12)

–6

–9

–12

HAMD, change from baseline –15 –3 1815 22 29 43 Evaluation day

Figure 3. HAMD (reduction in mean change from baseline) for patients with baseline HAMD score of >22 and Baseline HAMD f22 (* means pf0.05; # means pf0.10). during treatment, but also in the period following placebo (p<0.05) measured on CSRSD from day discontinuation of nemifitide. The peak period of 22 (peak effect) until day 43 (end of observation response was at 22 d, 1 wk after discontinuation of the (Figure 4a). In the patients treated with 30 mg/d treatment. This finding may suggest a new paradigm nemifitide the separation from placebo at day 22 for the treatment of MDD. reached a trend with a level of significance of p<0.10 There is an extensive literature of on the effect of with a continued clinical numerical advantage vs. severity of depression on response with response less placebo at days 29 and 43. There was no real difference readily measured seen in patients with milder levels of among treatment groups for patients measured on the depression. In our study we carried out a secondary CSRSD with a baseline HAMD score of f22 (Figure subanalysis to investigate whether response to nemi- 4b). In addition to Figure 3 for HAMD, Figure 4 for fitide was seen more clearly in patients with more CSRSD replicated the same observation about the severe depression, defined as those with baseline difference in response between subjects treated with HAMD scores above the median, which was 22. nemifitide vs. subjects treated with placebo. As We found that patients with a HAMD score of >22 mentioned previously, after separation of all subjects at baseline showed a higher percentage of responders by HAMD baseline, both doses of nemifitide demon- for both doses of nemifitide, with statistical separation strated a significant difference in response for CSRSD from placebo. This finding was supported by the in favour of subjects with a HAMD baseline score of results of an exploratory analysis of CSRSD (Figure 4), >22 vs. subjects with a HAMD baseline score of f22. which had the greatest range of baseline values. On the other hand, for CSRSD the picture for the According to Figure 4a, patients measured on the placebo group is completely different from both doses CSRSD with a baseline HAMD score of >22 treated of nemifitide. Figure 4 shows that the placebo group with 45 mg/d nemifitide demonstrated a stable had a better response in favour of subjects with a response to treatment and statistical separation from HAMD baseline score of f22 (with a major effect from Efficacy and safety of nemifitide 525

Table 3. Non-drug-related, drug-related* and serious adverse events (AEs), incidences of AEs

Treatment groups

Placebo 30 mg/d 45 mg/d Overall Parameters (n=28) (n=27) (n=26) (n=81)

Number of patients with AE 14 (50.0%) 19 (70.4%) 17 (65.4%) 50 (61.7%) Number of patients with 0 (0.0%) 1 (3.7%) 0 (0.0%) 1 (1.2%) serious AE Number of patients with 0 (0.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) drug-related serious AE Number of incidences with AE 28 (24.1%) 58 (50.0%) 30 (25.9%) 116 (100%)

Number of incidences with 10 (21.7%) 23 (50.0%) 13 (28.3%) 46 (100%) Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 non-drug-related AE Number of incidences with 18 (25.7%) 35 (50.0%) 17 (24.3%) 70 (100%) drug-related AE Incidences of AE Headache 2 5 7 14 Mild 2 2 4 8 Moderate 0 3 3 6 Dry mouth 3 2 1 6 Mild 3 2 1 6 Moderate 0 0 0 0 Nausea 2 3 0 5 Mild 2 2 0 4 Moderate 0 1 0 1 Dizziness 0 3 1 4 Mild 0 2 1 3 Moderate 0 1 0 1 Pain at injection site 1 1 2 4 Mild 1 1 2 4 Moderate 0 0 0 0 Constipation 1 2 0 3 Mild 0 2 0 2 Moderate 1 0 0 1 Fatigue 2 1 0 3 Mild 1 1 0 2 Moderate 1 0 0 1

* Drug-related AE was considered either of ‘uncertain’ or ‘probable’ relationship to the study drug by the clinical investigator for double-blind data.

a couple of subjects) vs. subjects with a HAMD baseline (10 doses in 2 wk) at the end of observation (4 wk post- score of >22. Therefore, for both psychometric scores dosing) efficacy results are different from peak effect. there was a very strong tendency to get a better response The study has not directly addressed the issue as to from subjects with a greater level of depression for both how often the treatment should be given. The peak doses of nemifitide. There was no real indication about effect, in the present study at 3 wk, suggests, however, the same tendency for the placebo group. that the second series of treatment may be given 1 The results of this study suggest that patients re- month later. This suggests that additional attention spond rapidly to treatment with 45 mg/d nemifitide. needs to be paid to the effect of a longer series of This is very important when time to response may be injections in the treatment period with nemifitide and crucial, as well as in terms of treatment cost. This a different clinical strategy, for example, treatment study is also consistent with previous studies, which every other day instead of every day. The preclinical have found that after a short period of treatment data support these approaches. 526 S. A. Montgomery et al.

(a) Subjects with HAMD baseline >22 0

Nemifitide 30 mg, enrolled (n = 13) –3 Nemifitide 45 mg, enolled (n = 13) Placebo, enrolled (n = 14) –6

–9 # Nemifitide, 30 mg vs. placebo: –12

Nemifitide, 45 mg vs. placebo: * –15 * *

CSRSD, change from baseline –3 1 8 15 22 29 43

–6

–9

–12

CSRSD, change from baseline –15 –3 1815 22 29 43 Evaluation day

Figure 4. CSRSD (reduction in mean change from baseline) for patients with baseline HAMD score of >22 and baseline HAMD score of f22 (* means pf0.05; # means pf0.10).

Our study has a number of methodological In the three previous placebo-controlled studies, limitations. First, the sample size (n=81) is small. including the dose of 80 mg/d nemifitide, there was However, small pilot studies are sometimes used to no difference in the incidence of headaches between test proof of principle where a trend compared to placebo and nemifitide patients. The data obtained placebo in small numbers is taken to indicate probable from this study confirm our preliminary finding that efficacy. Second, the period of follow-up was relatively nemifitide is a well-tolerated and active treatment short at 4 wk following discontinuation. A period of for patients with major depression, particularly for observation of 6–12 months would help investigate patients with high severity of symptoms at baseline. any potential differences in the duration of drug effect vs. placebo. Third, this study was not designed to Conclusions suggest how nemifitide might be used in an ordinary clinical setting. The main clinical objective of this pilot, This proof-of-principle study demonstrated that proof-of-principle study was to assist in the design of nemifitide, a novel pentapeptide antidepressant, is a future studies to clarify acceptable or maybe optimum promising new drug for the treatment of major dose and dosing schedules for the large pivotal trial depression. The 45-mg/d dose showed greater effect and ultimately for clinical use. However, the data from compared with placebo than 30 mg/d, indicating a our study provide preliminary evidence of the safety possible dose–response curve. Patients with more and the efficacy of nemifitide in the treatment of major severe symptoms of depression (baseline HAMD depression. score >22) showed a better response than patients In this study nemifitide was well tolerated; there with less severe symptoms (baseline HAMD score was a numerical increase of mild/moderate headache f22) in both doses of nemifitide, with better separ- in nemifitide-treated patients compared with placebo ation from placebo. At the doses of 30 mg/d and but the difference was not statistically significant. 45 mg/d, nemifitide showed a good tolerability and Efficacy and safety of nemifitide 527 safety profile. No dropouts occurred because of AEs. Bull SA, Hunkeler EM, Lee JY, Rowland CR, Williamson The percentage of patients with AEs in the nemifitide TE, Schwab JR, Hurt SW (2002). Discontinuing or treatment groups was comparable with the placebo switching selective serotonin-reuptake inhibitors. Annals of group. Pharmacotherapy 36, 578–584. Carroll BJ, Feinberg M, Smouse PE, Rawson SG, Greden JF (1981). The Carroll rating scale for depression. I. Acknowledgements Development, reliability and validation. British Journal of Psychiatry 138, 194–200. The study was supported by Innapharma/Tetragenex Eaddy M, Bramley T, Regan T (2003). Time to antidepressant Pharmaceuticals, Inc. and all study-related materials discontinuation: a comparison of controlled-release were supplied to each clinical study centre in order to paroxetine and immediate-release selective serotonin- conduct the study. All the clinical research sites were reuptake inhibitors. Managed Care Interface 16, 22–27. on a contractual basis and funded directly from Ehrensing RH, Kastin AJ (1978). Dose-related biphasic effect of proyl-leucyl-glycinamide (MIF-1) in depression. Innapharma/Tetragenex Pharmaceuticals, Inc. per Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 American Journal of Psychiatry 135, 562–566. standardized contractual agreement for the conduct of Ehrensing RH, Kastin AJ, Wurzlow GF, Michell GF, clinical research. Mebane AH (1994). Improvements in major depression The authors acknowledge the scientific team of after low subcutaneous doses of MIF-1. Journal of Affective Tetragenex Pharmaceuticals, Inc. (Dr Gabriela Disorders 31, 227–233. Nicolau, Dr Jeffrey Freed and Dr Ken Cartwright) for Feighner JP (2003). Clinical and preclinical overview of their support of this study. The authors thank Mrs nemifitide (INN 00835), a novel pentapeptide Suzanne Connolly for her assistance in preparing this antidepressant. Depression: emerging research and manuscript for publication. The authors also thank the treatment approaches. Euroconferences, Istitute Pasteur, two reviewers and the field editor for their very good Paris, 16–17 January. scientific, clinical and practical comments, which Feighner JP, Ehrensing RH, Kastin AJ, Leonard BE, Sverdlov L, Nicolau S, Patel A, Hlavka J, Abajian H, helped to improve the manuscript. Noble JF (2000). A double-blind, placebo-controlled, efficacy, safety, and pharmacokinetic study of INN 00835, a Statement of Interest novel antidepressant peptide, in the treatment of major depression. Journal of Affective Disorders 61, 119–126. All the patents that have been issued both in Feighner JP, Ehrensing RH, Kastin AJ, Patel A, Sverdlov L, the USA and throughout the world are the property Hlavka J, Abajian HB, Noble JF, Nicolau G (2001a). of Innapharma/Tetragenex Pharmaceuticals, Inc. Double-blind, placebo-controlled study of INN 00835 Relationships to Innapharma/Tetragenex Pharma- (nemifitide) in the treatment of outpatients with major depression. International Clinical Psychopharmacology 16, ceuticals, Inc.: Professor Stuart A. Montgomery is the 345–352. Chairman of the Scientific Advisory Board. Dr John Feighner JP, Nicolau G, Abajian HB, Marricco NC, P. Feighner is the President of Tetragenex Pharma- Morrison J, Sverdlov L, Hlavka J, Tonelli Jr. G (2001b). ceuticals, Inc. Lev Sverdlov, Ph.D. is the Head Clinical pharmacokinetic studies with INN 00835 of Biostatistics of Tetragenex Pharmaceuticals, Inc. (nemifitide), a novel pentapeptide antidepressant. Joseph Hlavka, Ph.D. is the Head of CMC, the Biopharmaceutics & Drug Disposition 23, 33–39. Medicinal Chemist of Tetragenex Pharmaceuticals, Feighner JP, Sverdlov L (2002). The use of discriminant Inc., and the inventor of nemifitide. George Tonelli Jr. analysis to separate a study population by treatment is the Director of Regulatory Affairs of Tetragenex subgroups in a clinical trial with a new pentapeptide Pharmaceuticals, Inc. Dr Ram Shrivastava, Dr Ari antidepressant. Journal of Applied Research 2, 50–57. Kiev and Dr Lynn Cunningham are the Primary Feighner JP, Sverdlov L, Nicolau G, Abajian HB, Hlavka J, Freed JS, Tonelli G (2003). Clinical effectiveness of Investigators for the study. nemifitide, a novel pentapeptide antidepressant, in depressed outpatients: comparison of follow-up References re-treatment with initial treatment. International Journal of Neuropsychopharmacology 6, 207–213. APA (1994). Diagnostic and Statistical Manual of Mental Frank E, Prien RF, Jarrett RB, Keller MB, Kupfer DJ, Lavori Disorder (4th edn). Washington, DC: American Psychiatric PW, Rush AJ, Weissman MM (1991). Conceptualization Association. and rationale for consensus definitions of terms in major Barranco SF, Thrash ML, Hackett E, Frey J, Ward J, depressive disorder. Archives of General Psychiatry 48, Norris E (1979). Early onset of response to doxepin 851–855. treatment. Journal of Clinical Psychiatry 40, Frazer A (1997). Antidepressants. Journal of Clinical Psychiatry 265–268. 58 (Suppl. 6), 9–25. 528 S. A. Montgomery et al.

Greenberg PE, Stiglin LE, Finkelstein SN, Berndt ER (1993). Montgomey SA, Asberg M (1979). A new depression scale The economic burden of depression in 1990. Journal of designed to be sensitive to change. British Journal of Clinical Psychiatry 54, 405–418. Psychiatry 134, 382–389. Guy W (1976). Early Clinical Drug Evaluation UNIT (ECDEU) Olfson M, Marcus S, Sackeim HA, Thompson J, Pincus HA Assessment Manual for Psychopharmacology (revised). (1998). Use of ECT for the inpatient treatment of NIMH publication DHEW publ. no. (Adm) 76-338. recurrent major depression. American Journal of Psychiatry Bethesda MD: National Institute of Mental Health, 155, 22–29. 218–222. Pignatiello MF, Olson GA, Kastin AJ, Ehrensing RH, Hamilton M (1960). A rating scale for depression. Journal McLean JH, Loson RD (1989). MIF-1 is active in a chronic of Neurology, Neurosurgery and Psychiatry 23, 56–62. stress animal model of depression. Pharmacology, Hawley CJ, Gale TM, Sivakumaran T (2002). Defining Biochemistry & Behavior 32, 737–742. remission by cut off score on the MADRS: selecting Richelson E (2003). Interactions of antidepressants with the optional value. Journal of Affective Disorders 72, 177–184. neurotransmittertransporters andreceptorsand their clinical Hlavka JJ, Nicolau G, Noble JF, Abajian H (1997). relevance. Journal of Clinical Psychiatry 64 (Suppl. 13), 5–12. Downloaded from https://academic.oup.com/ijnp/article/9/5/517/750864 by guest on 25 September 2021 INN 00835 – antidepressant. In Drug of the Future Rothshild AJ, Duval SE (2003). How long should patients (pp. 1314–1318). Barcelona: Prous Science Publishers. with psychotic depression stay on the antipsychotic Indrayan A, Sarmukaddam SB (2001). Medical Biostatistics. medications? Journal of Clinical Psychiatry 64, 390–396. New York: Marcel Dekker, Inc. Sambunaris A, Hesselink JK, Pinder R, Panagides J, Stahl Kessler RC, McGonagle KA, Swartz M, Blazer DG, SM (1997). Development of new Antidepressants. Journal Nelson CB (1993). Sex and depression in the National of Clinical Psychiatry 58 (Suppl. 6), 40–53. Comorbidity Survey, 1: lifetime prevalence, chronicity and Stahl SM (2003). At long last, long-lasting psychiatric recurrence. Journal of Affect Disorders 29, 85–96. medications: an overview of controlled-release Khan A, Leventhal RM, Khan SR, Brown WA (2002). technologies. Journal of Clinical Psychiatry 64, 355–356. Severity of depression and response to antidepressants Stahl SM, Grady MM (2003). Difference in mechanism of and placebo: an analysis of the Food and Drug action between current and future antidepressants. Administration Database. Journal of Clinical Journal of Clinical Psychiatry 64 (Suppl. 13), 13–17. Psychopharmacology 22, 40–45. Weissman MM, Bland RC, Canino CJ, Faravelli C, Greenwald Masand PS (2004). Depression – treatment and unmet S, Hwu HG, Joyce PR, Karam EG, Lee CK, Lellouch J, et al. medical needs. Supplement to Clinical Psychiatry News, (1996). Cross-national epidemiology of major depression Atypical Antipsychotics in Treatment-Resistant and bipolar disorder. Journal of the American Medical Depression, pp. 6–8. Association 276, 293–299. Montgomery SA (1999). New development in the treatment Zimmerman M, Posternak MA (2004). Defining remission on of depression. Journal of Clinical Psychiatry 60 (Suppl. 14), the Montgomery–Asberg Depression Rating Scale. Journal 10–15. of Clinical Psychiatry 65, 163–168.