D-Cycloserine Improves Difficult Discriminations in a Pattern Separation Task in Alzheimer’s Disease Patients with Dementia

Abstract

Recent fMRI studies in human identified that pattern separation ability is associated with increased activity in the hippocampal dentate gyrus (DG), whereas no such DG changes are seen during pattern completion. Disruption to neurogenesis in the DG has been associated with Alzheimer’s disease (AD). In a post-hoc analysis of two large unsuccessful AD clinical trials, we examined the effect of D-cycloserine (DCS) on a specific object pattern separation measure, a component of the picture recognition task from the Cognitive Drug Research (CDR) system. This task yields a measure of pattern separation and a measure of pattern completion. Study data were available for 756 AD patients with dementia, randomized to several doses of DCS. Data were available at week 2, 6, 14, and 26 for 732, 707, 653, and 559 patients, respectively. None of the DCS doses had a statistically significant benefit over placebo on pattern completion. However, the DCS 15 mg BID dose significantly increased accuracy over placebo on the pattern separation measure by 5.1%. Further, the magnitude of the benefit of DCS 15 mg BID over placebo was almost doubled relative to the whole study population in a subset of patients whose pattern separation scores were≥2 standard deviations poorer than the CDR norm of age-matched healthy individuals at baseline. These post-hoc analyses suggest a potential value of the pattern separation task for evaluating compounds promoting neurogenesis. Further, the use of a restrictive pattern separation eligibility criterion might facilitate signal detection.

Keywords

Alzheimer’s disease dementia D-cycloserine neurogenesis pattern separation

INTRODUCTION

Emerging evidence has suggested that decreased adult hippocampal neurogenesis, and in particular decreased dentate gyrus (DG) activity may represent an early critical event in the course of Alzheimer’s disease (AD) [1, 2]. In mice models of AD, hippocampal adult neurogenesis is reduced and associated with amyloid-β pathology [3, 4].

Changes in the entorhinal cortex have been identified as a very early feature in AD, and this structure is intimately connected to the DG via the performant pathway, using excitatory amino acids as neurotransmitters, principally L-glutamate in which a decrease has been reported in AD [5.] Activation of the N-Methyl-D-Aspartate (NMDA) receptor, thought to play a key role in memory, requires binding of the excitatory amino acid glutamate, its natural ligand, and glycine, its co-agonist. The NMDA receptor has been linked to neurogenesis [6] and pattern separation modulation, in particular the NR1 subunit of the receptor, the location of the glycine binding site [7].

D-cycloserine (DCS) is a partial agonist at the glycine binding site of the NMDA receptor, and has been shown to promote neuroplasticity in animal models [8, 9] and in human brain [10]. DCS has been evaluated as a treatment for patients with probable AD, but was essentially abandoned in the past twenty years [11 –15] principally due to contradictory or disappointing findings from several clinical studies. For instance, two large multi-center and multi-national studies were completed in 1993 in patients meeting diagnostic criteria for probable AD and with Mini-Mental State Examination (MMSE) scores of 12–24 inclusive that failed to show a significant benefit over placebo using the Cognitive Drug Research (CDR) system, a computerized cognitive battery widely used in clinical trials, across a range of DCS doses [16].

Recently, it has been shown that DG activity may be more specifically assessed using cognitive tasks that require difficult pattern separations [17, 18]. Functional magnetic resonance imaging (fMRI) studies have demonstrated increased DG activity when difficult discriminations are being performed (i.e., the rejection of closely similar pictures), whereas no increased DG activity is found when the task is merely the identification of the original pictures (pattern completion) [19 –21]. One of the CDR System tests, picture recognition, involves the presentation of pictures that must later be recognized as they initially appeared (pattern completion, measuring familiarity with the stimulus) or discriminated from closely similar pictures (pattern separation). This latter discriminating component of the picture recognition task fulfils the requirements for a pattern separation test. Using this pattern separation component of the CDR system, studies have previously identified significant deficits in patients with AD [22] and mild cognitive impairment [23]. A separate analysis of difficult discriminations using this specific CDR pattern separation task was not performed in the original data analysis of the companion AD studies described above [15]. Therefore, we assessed whether DCS had a selective effect on the specific pattern separation task of the CDR in a post-hoc analysis of the study data. In addition, we examined the potential of a more restrictive eligibility criterion using this CDR pattern separation task relative to normative data to enhance signal detection.

MATERIALS AND METHODS

Data for this post-hoc analysis were obtained from two randomized, double-blind, placebo controlled, 6-month, multi-center, and multi-national studies of DCS versus placebo conducted in AD patients with dementia [16 , 25]. The trials were run between 1991 and 1993. The data from these two companion studies were combined because they used the same inclusion and exclusion criteria, overall design, and outcome measures. Protocol EC6-93-06-025 enrolled 384 patients and was conducted at 34 centers in six European countries and Israel [24]. Protocol NC6-90-02-009 enrolled 410 patients and was conducted at 29 centers in the USA and Canada [25]. Following training on the use of the CDR System at the screen visit, patients were assessed at the baseline visit and weeks 2, 6, 14, and 26. A total of 794 patients met the NINCDS-ADRDA revised (DSM-III-R) criteria (21) for probable AD [26] supported by CAT scan or MRI. Eligible patients had a MMSE score between 12 and 24 inclusive, a Global Deterioration Scale score between 3 and 5, a Hamilton depression scale score < 18, a Hachinski ischemic scale score < 4, evidence of progressive worsening of memory and a deficit in one more areas of cognition (orientation, language, praxis, attention, visual perception, problem solving, social function). Eligible subjects were randomized to DCS 1, 5, 15, 50, and 100 mg BID or placebo in the two studies. The primary outcome measure was the CDR System, a set of computerized cognitive tests assessing attention, information processing, executive function, working memory and verbal and non-verbal episodic memory. One of the CDR System tests, picture recognition, involves the presentation of pictures that must later be recognized (pattern completion) or discriminated from closely similar pictures (pattern separation). In the original study analysis, a signal detection theory index of stimulus sensitivity was calculated which combined the measures of the ability to correctly recognize the original pictures with the ability to correctly reject the closely similar pictures. This post-hoc analysis separately investigated the effects of DCS on the pattern completion and pattern separation measures of the task. The analysis was also repeated in a subgroup of the Intent To Treat (ITT) set who showed a pattern separation score≥2 standard deviations (SDs) poorer than the norm from the CDR System normative database for age-matched healthy individuals [27] at the baseline visit.

Statistical analysis

The pattern completion and pattern separation scores were analyzed separately. The baseline scores were subtracted from each subsequent study visit to produce change from baseline scores. The change from baseline scores were analyzed using mixed model repeated measures analyses of variance. Terms were fitted to the model for treatment group, study visit and the interaction between them. The study visits were fitted as repeated measures. The analysis was conducted using SAS version 9.4 (SAS Institute, Cary, NC, USA).

RESULTS

There were 794 subjects (451 females) with AD dementia enrolled in the two companion studies. The mean age was 72.7 years (range 48–91) and mean MMSE 19.6 (range 12–24). Data from 756 subjects were available for this post-hoc analysis.

The pattern completion measures declined progressively over the course of the study in all groups except the DCS 1 mg BID dose. The decline was statistically significant in the placebo group (p = 0.009), DCS 5 mg (p < 0.0001), DCS 15 mg (p = 0.0247), DCS 50 mg (p = 0.0152), and DCS 100 mg BID (p = 0.0001) groups. None of the DCS doses, including the 1 mg dose had a statistically significant beneficial effect over placebo on the pattern completion stimuli in this patient population at any individual study visit including the endpoint (Fig. 1a).

Fig.1

Changes from baseline on Cognitive Drug Research (CDR) System Pattern Separation and Completion measures between placebo and D-cycloserine treated patients in the whole study population (ascending scores reflect improvements).

As shown in Figure 1b, the pattern separation measures progressively declined in all groups except the DCS 15 mg BID group. The decline on the pattern separation measure was statistically significant in the placebo group (p = 0.0312), DCS 1 mg (p < 0.0001), and DCS 50 mg (p = 0.0112) groups. In contrast, the DCS 15 mg BID group improved significantly over time (p = 0.003).

DCS 15 mg BID was significantly better than placebo on the pattern separation task at week 14 (p = 0.0192) and week 26 (p = 0.0066). Over all time points, DCS 15 mg BID was significantly better than placebo on the pattern separation task by 5.1% (s.e.m. 1.4; p = 0.0003). The improvement with DCS 15 mg BID over all time points had an effect size of 0.24, and at week 26 the effect size was 0.46 (Fig. 2a).

Fig.2

Change from baseline scores over the study period, presented as differences to placebo, for the Cognitive Drug Research (CDR) System Pattern Separation and Completion measures. Plot (a) is for the whole study population, and (b) for subjects > 2 SD from the CDR System normative database at baseline (ascending scores reflect improvements over placebo).

Fig.3

Changes from baseline on Cognitive Drug Research (CDR) System Pattern completion (a) and Pattern Separation (b) measures between placebo and D-cycloserine treated patients for subjects > 2 SD from the CDR System normative database at baseline (ascending scores reflect improvements).

We identified 363 patients whose pattern separation scores at baseline were≥2 SDs poorer than the age-matched norm on the CDR System normative database. In this sub-set of patients, we found an enhanced signal of 9.9% (s.e.m. 2.8; p = 0.0005) for the DCS 15 mg BID dose over placebo over the study period that was greater than the whole study population of 5.1% noted above (Fig. 2b). Over the entire study, the improvement with DCS 15 mg BID for this restricted population had an effect size of 0.46, and 0.82 at week 26 specifically.

DISCUSSION

Most of the patients in these companion AD studies showed a progressive decline on both the pattern completion and pattern separation tasks of the CDR suggesting a generalized hippocampal connectivity defect since pattern separation and pattern completion have been linked to hippocampal function [28,29, 28,29] in line with AD pathology. In contrast to all other patients in these studies, patients treated with DCS 15 mg BID revealed a significantly improved performance on the pattern separation task that was statistically significantly better than placebo at 14 weeks and 26 weeks. To our knowledge, this is the first report of a beneficial drug effect on a pattern separation task in AD patients.

The observed dose response at week 14 and 26 appeared to follow a relatively bell-shaped pattern across the different DCS doses with the optimal dose at 15 mg. BID. Such a pattern could be expected from a partial agonist. The bell-shaped pattern is reminiscent of the acute effect of DCS on visual memory reported in Rhesus monkeys [30]; scopolamine-induced non-spatial reference memory deficit reported in rats [31] and reports in healthy volunteers who were subjected to a scopolamine infusion and cognitively assessed using the CDR system [32]. However, and for the latter, one might expect the mechanisms of action linked to NMDA receptor activation and subsequent cognitive improvement to be different after acute and chronic regimen, since generation of new-born hippocampal neurons and full network integration is not expected after acute administration.

The failure of DCS to demonstrate a significant benefit over placebo in the overall study population of this clinical trial is in line with some preclinical and clinical findings [12 , 15] but the differential benefit of DCS 15 mg BID on pattern separation is consistent with other reports showing selective cognitive benefit [11,14, 11,14]. The dose response differences could be explained by the reported bi-phasic dose response of DCS on NMDA receptor function with lower doses having an agonistic effect and higher dose an antagonistic effect [33]. Moreover, such a bell-shaped pattern is reminiscent of the effect of the non-competitive NMDA receptor antagonist memantine [34 , 36].

Subsequent to the post-hoc analysis of the whole study population, we identified a sub-set of patients whose CDR pattern separation scores at baseline were≥2 SDs poorer than the age-matched healthy norm obtained from the CDR normative database. This sub-set of patients with greater pattern separation deficits at baseline yielded an enhanced signal favoring DCS 15 mg BID over placebo relative to the whole study population (Fig. 2).

This finding suggests that the addition of a pattern separation criterion to the study inclusion criteria for patient selection might enhance signal detection in AD trials. A similar enhanced effect has been reported in patients with Parkinson’s disease dementia who were assessed with the CDR system [37] where a subset of patients were selected based upon baseline scores that were≥2 SDs poorer than the norm from the CDR System normative database on four cognitive measures. According to the Yerkes-Dodson principle, a broader range of impairment differences for any given cognitive domain between patients increases the variability of drug response since patients with greater impairment may require higher doses than patients with lesser impairment [38, 39].

Conclusions

In this post-hoc analysis, DCS 15 mg BID revealed a statistically significant benefit over placebo on a specific CDR System pattern separation measure in AD patients with dementia, and in contrast to all other DCS doses. Pattern separation is a process that has been linked to neurogenesis and DG activity. The beneficial data shown for the DSC 15 mg BID dose are consistent with the dose-response pattern previously reported in other studies.

Our findings suggest a potential value of using a pattern separation task to evaluate compounds that might enhance neurogenesis. Further, a subsequent analysis using a subset of patients (n = 363) with a baseline pattern separation score that was≥2 SD poorer than healthy controls from the CDR System normative database yielded an enhanced signal for DCS 15 mg BID relative to the whole study population. This finding suggests that a more restrictive eligibility criterion that uses the CDR System pattern separation task might facilitate signal detection in AD clinical trials. Moreover, these findings identify the value of more precise dose ranging prior to embarking in large and long clinical trials. This post-hoc analysis represents only one small study and clearly requires further exploration.

Footnotes

ACKNOWLEDGMENTS

Both Drs. Targum and Wesnes received grants from the sponsor, GD Searle (Skokie, Illinois) at the time of this study (1991-1993) to participate in the conduct and/or analysis of the study. However, the sponsor was not involved in any way in the post-hoc analyses or interpretation of the data presented in this report, not involved in the manuscript review, and not involved in the decision to submit the manuscript for publication. Partial support for the post-hoc analysis came from Bracket LLC.

Authors’ disclosures available online ().

References

Hollands

, Bartolotti

, Lazarov

(2016) Alzheimer’s disease and hippocampal adult neurogenesis; exploring shared mechanisms. Front Neurosci 10, 178.

Rodríguez

, Verkhratsky

(2011) Neurogenesis in Alzheimer’s disease. J Anat 219, 78–89.

Adeosun

, Hou

, Zheng

, Stockmeier

, Ou

, Paul

, Mosley

, Weisgraber

, Wang

(2014) Cognitive deficits and disruption of neurogenesis in a mouse model of apolipoprotein E4 domain interaction. J Biol Chem 289, 2946–2959.

Rodríguez

, Jones

, Tabuchi

, Allan

, Knight

, LaFerla

, Oddo

, Verkhratsky

(2008) Impaired adult neurogenesis in the dentate gyrus of a triple transgenic mouse model of Alzheimer’s disease. PLoS One 3, e2935.

Hyman

, Van Hoesen

, Damasio

(1987) Alzheimer’s disease: Glutamate depletion in the hippocampal perforant pathway zone. Ann Neurol 22, 37–40.

Joo

, Kim

, Lee

, Park

, Kim

, Yun

, Lee

, Rhim

, Son

(2007) Activation of NMDA receptors increases proliferation and differentiation of hippocampal neural progenitor cells. J Cell Sci 120, 1358–1370.

McHugh

, Jones

, Quinn

, Balthasar

, Coppari

, Elmquist

, Lowell

, Fanselow

, Wilson

, Tonegawa

(2007) Dentate gyrus NMDA receptors mediate rapid pattern separation in the hippocampal network. Science 317, 94–99.

Ren

, Li

, Zhang

, Li

, Wang

, Ma

(2013) The effects of intra-hippocampal microinfusion of D-cycloserine on fear extinction, and the expression of NMDA receptor subunit NR2B and neurogenesis in the hippocampus in rats. Prog Neuropsychopharmacol Biol Psychiatry 44, 257–264.

Sta Maria

, Reger

, Cai

, Baquing

MAT

, Buen

, Ponnaluri

, Hovda

, Harris

, Giza

(2017) D-cycloserine restores experience-dependent neuroplasticity after traumatic brain injury in the developing rat brain. J Neurotrauma 34, 1692–1702.

10.

Forsyth

, Bachman

, Mathalon

, Roach

, Asarnow

(2015) Augmenting NMDA receptor signaling boosts experience-dependent neuroplasticity in the adult human brain. Proc Natl Acad Sci U S A 112, 15331–15336.

11.

Tsai

, Falk

, Gunther

, Coyle

(1999) Improved cognition in Alzheimer’s disease with short-term D-cycloserine treatment. Am J Psychiatry 156, 467–469.

12.

Fakouhi

, Jhee

, Sramek

, Benes

, Schwartz

, Hantsburger

, Herting

, Swabb

, Cutler

(1995) Evaluation of cycloserine in the treatment of Alzheimer’s disease. J Geriat Psychiatry Neurol 8, 226–230.

13.

Randolph

, Roberts

, Tierney

, Bravi

, Mouradian

, Chase

(1994) D-cycloserine treatment of Alzheimer disease. Alzheimer Dis Assoc Disord 8, 198–205.

14.

Schwartz

, Hashtroudi

, Herting

, Schwartz

, Deutsch

(1996) d-Cycloserine enhances implicit memory in Alzheimer patients. Neurology 46, 420–424.

15.

Schade

, Paulus

(2016) D-cycloserine in neuropsychiatric diseases: A systematic review. Int J Neuropsychopharmacol 19, pyv102.

16.

Jones

, Laake

, Øksengård

(2002) D-cycloserine for Alzheimer’s disease. Cochrane Database Syst Rev 2, CD003153.

17.

Leutgeb

, Leutgeb

, Moser

(2007) Pattern separation in the dentate gyrus and CA3 of the hippocampus. Science 315, 961–966.

18.

Schmidt

, Marrone

, Markus

(2012) Disambiguating the similar: The dentate gyrus and pattern separation. Behav Brain Res 226, 56–65.

19.

Berron

, Schütze

, Maass

, Cardenas-Blanco

, Kuijf

, Kumaran

, Düzel

(2016) Strong evidence for pattern separation in human dentate gyrus. J Neurosci 36, 7569–7579.

20.

Kirwan

, Stark

(2007) Overcoming interference: An fMRI investigation of pattern separation in the medial temporal lobe. Learn Mem 14, 625–633.

21.

Yassa

, Stark

(2011) Pattern separation in the hippocampus. Trends Neurosci 34, 515–525.

22.

Wesnes

, Annas

, Basun

, Edgar

, Blennow

(2014) Performance on a pattern separation task by Alzheimer’s patients shows possible links between disrupted dentate gyrus activity and apolipoprotein E ɛ4 status and cerebrospinal fluid amyloid-β42 levels. Alzheimers Res Ther 6, 20.

23.

Wesnes

(2010) Visual object pattern separation: A paradigm for studying the role of the dentate gyrus in memory disorders. Alzheimers Dement 6, Supplement e45.

24.

Pauvlik

(1993) Searle clinical research report. A two year, multicenter, randomized, double-blind controlled parallel group study of cycloserine in the treatment of Alzheimer’s disease with six months placebo control. Report No. EC6-93-06-025, July 30, 1993. GD Searle, Skokie, IL.

25.

Pauvlik

(1993) Searle clinical research report. A two-year, multicenter, randomized, double-blind controlled parallel group study of cycloserine in the treatment of Alzheimer’s disease with six months placebo control. Report No. NC6-93-06-009, September 29, 1993. GD Searle, Skokie, IL.

26.

Lopez

, McDade

, Riverol

, Becker

(2016) Evolution of the diagnostic criteria for degenerative and cognitive disorders. Curr Opin Neurol 24, 532–541.

27.

Wesnes

, McNamara

, Annas

(2016) Norms for healthy adults aged 18-87 years for the Cognitive Drug Research System: An automated set of tests of attention, information processing and memory for use in clinical trials. J Psychopharmacol 30, 263–272.

28.

Guzman

, Schlögl

, Frotscher

, Jonas

(2016) Synaptic mechanisms of pattern completion in the hippocampal CA3 network. Science 353, 1117–1123.

29.

Rolls

, Kesner

(2016) Pattern separation and pattern completion in the hippocampal system. Introduction to the Special Issue. Neurobiol Learn Mem 129, 1–3.

30.

Matsuoka

, Aigner

(1996) D-cycloserine, a partial agonist at the glycine site coupled to N-methyl-D-aspartate receptors, improves visual recognition memory in rhesus monkeys. J Pharmacol Exp Ther 278, 891–897.

31.

Andersen

, Lindberg

, Myhrer

(2002) Effects of scopolamine and D-cycloserine on non-spatial reference memory in rats. Behav Brain Res 129, 211–216.

32.

Wesnes

, Jones

, Kirby

(1991) The effects of D-cycloserine, a glycine agonist, in a human model of the cognitive deficits associated with ageing and dementia. Br J Clin Pharmacol 31, 577–578.

33.

Emmett

, Mick

, Cler

, Rao

, Iyengar

, Wood

(1991) Actions of d-cycloserine at the N-methyl-d-aspartate-associated glycine receptor site in vivo. Neuropharmacology 30, 1167–71.

34.

Lench

, Robson

, and Jones

RSG

(2015) Differential effects of D-cycloserine and ACBC at NMDA receptors in the rat entorhinal cortex are related to efficacy at the co-agonist binding site. PLoS One 10, e0133548.

35.

Berridge

, Shumsky

, Andrzejewski

, McGaughy

, Spencer

, Devilbiss

, Waterhouse

(2012) Differential sensitivity to psychostimulants across prefrontal cognitive tasks: Differential involvement of noradrenergic α₁- and α₂-receptors. Biol Psychiatry 71, 467–473.

36.

Canal

, Imbimbo

(1996) Relationship between pharmacodynamic activity and cognitive effects of eptastigmine in patients with Alzheimer’s disease. Eptastigmine Study Group. Clin Pharmacol Ther 60, 218–228.

37.

Goetghebeur

PJD

, Micaletto

, Wesnes

(2018) Computerized assessment of attentional processes and drug efficacy in Parkinson’s disease dementia using the CDR System. Poster presented at: Advances in Alzheimer’s and Parkinson’s therapies – An AAT-AD/PD focus meeting. 15-18 March 2018, Torino, Italy.

38.

Mattay

, Goldberg

, Fera

, Hariri

, Tessitore

, Egan

, Kolachana

, Callicott

, and Weinberger

(2003) Catechol O-methyltransferase val158-met genotype and individual variation in the brain response to amphetamine. Proc Natl Acad Sci U S A 100, 6186–6191.

39.

Farrell

, Tunbridge

, Braeutigam

, Harrison

(2012) COMT Val(158)Met genotype determines the direction of cognitive effects produced by catechol-O-methyltransferase inhibition. Biol Psychiatry 71, 538–544.