Lessons learnt in recent trials in negative symptoms D. Umbricht, - - PowerPoint PPT Presentation

ISCTM ~ ECNP Joint Conference ▪ 1 September 2017 ▪ Paris ▪ France

Lessons learnt in recent trials in negative symptoms

• D. Umbricht, N. Schooler, D. Fraguas,
• A. Khan, A. Kott, D. Daniel, C. Arango

ISCTM Paris, September 1st, 2017

1

Disclosures

• I am an employee of F

. Hoffmann – La Roche

• I hold stocks of F

. Hoffmann – La Roche, Novartis, and Basilea

• The views and opinions expressed in this

presentation only present the personal views

• f Dr. Umbricht and not those of F

. Hoffmann – La Roche, Ltd.

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

2

Outline

• Results of a survey among experienced trialists
• Some lessons learnt from the bitopertin phase 3

trials

• Results of a meta-analysis of recent negative

symptom trials

• M onotherapy or adjunctive treatment?

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

3

Definitions*

• Prominant or dominant negative symptoms

– High negative symptoms but also a “substantial burden

from psychotic symptoms including hallucinations and delusions”

Example: A score of ≥4 on at least 3, or ≥5 on at least 2, of the 7

negative subscale items of the Positive and Negative Syndrome Scale (P ANSS) (Stauffer et al 2012).

• Predominant negative symptoms

– High negative symptoms, but mild and stable positive

symptoms (and low EPS and depression)

Example: A score of ≥4 on at least 3, or ≥5 on at least 2, of the 7

negative subscale items of the Positive and Negative S yndrome Scale plus a P ANSS positive score of <19, a Barnes Akathisia score of <2, a Simpson–Angus score of <4, and a Calgary Depressive Scale score of <9 (Stauffer et al 2012)

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

4

* M arder et al, 2013

Results of a survey among experienced trialists

• 12 colleagues interviewed
• 8 from industry, 4 from academia
• Responsible for 1-4 trials, average 2, in total

23 trials (9 academia, 14 industry)

• M ethods

– First a questionnaire* was sent out – Followed up by personal interviews* *

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

5

* Developed by Nina Schooler, Celso Arango and Daniel Umbricht * * Conducted by Nina Schooler and Daniel Umbricht

Key ‘hot’ topics

• Patient population/ Inclusion criteria
• Scales/Assessments
• Role of informant
• Role of a psychosocial «platform» in a trial

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

6

Patient population/ Inclusion criteria

• Predominant versus dominant negative symptoms

– General agreement that inclusion criteria may have been

‘overengineered’ with too much a focus on keeping positive symptoms low excluding a large number of subjects from studies

– General agreement that substantial positive symptoms should be

allowed as long as they are stable and not ‘disruptive’

– Possible solution: Stratify predominant/ dominant neg sx patients

• Severity of negative symptoms

– Concern that including only patients with relatively high negative

symptoms selects more ‘treatment resistant’ and least engaged patients

– Solution: Include patients with less severe negative symptoms, stratify

by severity

• Duration of illness

– Focus on patients earlier in their illness

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

7

Effects of neg sx inclusion criteria

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

8

0.25 0.31 0.35 0.33 0.32 0.40 0.23 0.36

Correlation between changes from BL in positive and negative subscales

• M ore restrictive baseline symptom severity thresholds yielded a considerably smaller sample size and higher

negative and lower positive symptoms at baseline.

• Unadjusted negative symptom change greater with more restrictive criteria;
• When adjusted for baseline severity the magnitude of change comparable across subsets.
• The amount of variance in negative symptom change attributed to positive symptom change also comparable

across subsets.

Dunayevich et al, European Neuropsychopharmacology(2014) 24, 1615–1621

Change in neg sx (NSFS*) tends to be greater when positive symptoms are “relatively” lower - independent of negative symptom level**

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

9

Pos Sx > Neg Sx Neg Sx > Pos Sx

NSFS* at baseline

N= 861 N= 62 N= 203 N= 174 N= 107 N= 376

* NSFS= P ANSS Negative S ymptom Factor Score (M arder factor; item scoring 0-6)) * * Data from phase 3 bitopertin suboptimally controlled symptoms studies

Change in neg sx (NSFS*) tends to be greater when positive symptoms are lower - independent of negative symptom level**

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

10

Pos Sx <14 Pos Sx = 14-18 Pos Sx > 18

NSFS* at baseline

N= 101 N= 361 N= 53 N= 159 N= 34 N= 149

* NSFS= P ANSS Negative S ymptom Factor Score (M arder factor; item scoreing 0-6) * * Data from phase 3 bitopertin suboptimally controlled symptoms studies

N= 471 N= 165 N= 300

Change in positive sx (PNSFS*) tends to be greater when positive symptoms are higher - independent of negative symptom level**

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

11

Pos Sx <14 Pos Sx = 14-18 Pos Sx > 18

NSFS at baseline

N= 101 N= 361 N= 53 N= 159 N= 34 N= 149

* PSFS= P ANSS Positive S ymptom Factor Score (M arder factor; item scoring 0-6) * * Data from phase 3 bitopertin suboptimally controlled symptoms studies

N= 165 N= 300

IRT analysis, Bitopertin P3 negative sx studies:

Key ‘avolition’ items of PANSS NSFS perform best around or below mean

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

12

• Analysis supports the view that patients with less severe neg sx should be enrolled

N=1878

Analysis performed by A. Khan, NeurocogTrials

N2 Emotional Withdrawal

• N4. Passive/ apathetic social withdrawal
• N1. Blunted affect
• G16. Active social avoidance
• G7. M otor retardation
• N6. Lack of spontaneity and flow of conversation
• N3. Poor rapport

Item information

M ean NSFS at baseline

IRT analysis, Bitopertin P3 neg sx studies:

Key ‘avolition’ items of PANSS NSFS perform best around or below mean

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

13

• Analysis supports the view that patients with less severe neg sx should be enrolled

200 200 200 200

Analysis performed by A. Khan, NeurocogTrials

N2 Emotional Withdrawal

• N4. Passive/ apathetic social withdrawal
• N1. Blunted affect
• G16. Active social avoidance
• G7. M otor retardation
• N6. Lack of spontaneity and flow of conversation
• N3. Poor rapport

N=1878

Item information

NSFS distribution at baseline

IRT analysis of NSA, Bitopertin P3 neg sx studies

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

14

75

N=1783

Analysis performed by A. Khan, NeurocogTrials

NSA Total Score Distribution at baseline

Item information

Individual items of NSA-16 perform the best around or below mean (IRT analysis, Bitopertin P3 neg sx studies)

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

15

Analysis performed by A. Khan, NeurocogTrials

Prolonged time to respond Restricted speech quantity Impoverished speech content Inarticulate speech Emotion: Reduced range Affect: Reduced modulation of intensity Affect: Reduced display on demand Reduced social drive

NSA Total Score Distribution at baseline

N=1783

Item information

Individual items of NSA-16 perform the best around or below mean (IRT analysis, Bitopertin P3 neg sx studies)

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

16

Analysis performed by A. Khan, NeurocogTrials

Poor rapport with interviewer Sexual interest Poor grooming and hygiene Reduced sense of purpose Reduced interests Reduced daily activity Reduced expressive gestures Slowed movements

Additional topics

Role of Informant

–

Deemed unreliable as informant, more important for compliance

–

Insistence on informant may limit patients to lower functioning patients, potentially excluding patients who live independently and may respond best

Scales/ Assessments

–

Include scales that measure avolition and expressive deficits separately (CAINS, BNSS), keep P ANSS for legacy reason

–

M ost colleagues in favor of either centralized ratings, video taping/ independent assessment

• r ‘Blended’ approaches with site rater responsible for enrollment, CR or videotaped

interviews used for outcome

–

Some scepticism that CR could not capture all nuances of negative symptoms also expressed

Role of a psychosocial «platform» in a trial

–

Biggest difference between academic and industry

• Colleagues from academia in favor of a psychosocial platform, also to increase number
• f visits to provide a «low level» psychosocial platform
• Colleagues from industry were less enthusiastic, favored ‘clean’ studies with fewer visits

Biomarkers

–

If biomarkers were considered, effort-choice tasks recommended to characterize patients

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

17

Drivers of placebo response in negative symptoms trials

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

18

Larger Placebo Response in Antipsychotic Trials Obscures True Treatment Effect

19 Factors associated with placebo response

• Year of study
• Number of sites
• Percentage of university or VA settings

Line A = M ean drug response Line B = Identity line i.e. size of drug response (vertical axis) equal to the placebo response (horizontal axis); vertical deviation from this identity line shows the drug-placebo difference. ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

Agid et al, Am J Psychiatry 2013; 170:1335–1344

Placebo Response >40% Obscures Treatment Effect in M DD

Meta-analysis of adjunctive treatment in MDD

Placebo (+SoC) response >40% showed a trend to lower risk ratio of response to the adjunctive drug vs. placebo (Iovieno & Papaksotas 2012)

20

Response Rate (RR)

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

Response* Rates (ITT population) in Bitopertin Phase 2 and Phase 3 Trials in Negative Symptoms of Schizophrenia

21

61% 58% 56% 44% 61% 58% 60% 58%

30% 35% 40% 45% 50% 55% 60% 65%

Response Rate Placebo Bitopertin 10mg

Phase 2 Trial

• ------------------------- Phase 3 Trials ---------------------------

* Response defined as ≥ 20% improvement on the NSFS

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

Categorization of patients

Responding to any intervention

• Active +
• Placebo +
• Not informative

Responding to active, but not placebo

• Active +
• Placebo ±
• Informative

Not responding to any intervention

• Active -
• Placebo -
• Not informative

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

22

30% >-7.25 25% >-8 20% >-9.3 15% >-9 10% >-10.5 5% >-14

≤-3.75 ≤-3.3 ≤-3 ≤-2.3 ≤-2 ≤-1.2

Adapted from Borsook, Becerra and Fava, Transl Psychiatry. 2013

Low Placebo response High Placebo response Identify sites that show high or low average placebo response and remove all data including data from patients on active treatment, assuming that average placebo response is indexing category of patients recruited at that site → composition of patients contains more patients in the green group with filter narrowing

Effect of Sites with “Normal” PBO Response: NS, 10mg, M M RM Week 24, Bitopertin Phase 3 neg sx studies

Treatment<<Favors>> Placebo

30% >-7.25 25% >-8 20% >-9.3 15% >-9 10% >-10.5 5% >-14

≤-3.75 ≤-3.3 ≤-3 ≤-2.3 ≤-2 ≤-1.2

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France 23

Results in “Normal” Pbo response sites better than the overall population in all NS studies

Correlation between Number of Study Sites and Placebo Response* Rates (ITT population) in Bitopertin Phase 2 and Phase 3 Trials in Negative Symptoms of Schizophrenia

24

61% 58% 56% 44%

R² = 0.92

30% 35% 40% 45% 50% 55% 60% 65% 40 50 60 70 80 90 100 110

Placebo Response

Number of Sites

Placebo Linear (Placebo)

Phase 2 Trial Phase 3 Trials

* Response defined as ≥ 20% improvement on the NSFS

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

Correlation between Number of Study Sites and Placebo Response* Rates (ITT population) in Bitopertin Phase 2 and Phase 3 Trials in Negative Symptoms of Schizophrenia

25

61% 58% 56% 44%

R² = 0.92

30% 35% 40% 45% 50% 55% 60% 65% 40 50 60 70 80 90 100 110

Placebo Response

Number of Sites

Placebo Bitopertin 10mg Linear (Placebo) Linear (Bitopertin 10mg)

Phase 2 Trial Phase 3 Trials

* Response defined as ≥ 20% improvement on the NSFS

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

Investigating predictors of placebo response in phase 3 bitopertin neg sx trials

• Erratic ratings
• Change in the first four weeks

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

26

<.0001 <.0001 <.0001 <.0001 <.0001 <.0001

• 12
• 11
• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=199) Not affected subjects (N=149) Affected subjects (N=16)

Protocol WN25308

Erratic ratings of NSFS in placebo treated patients– change of at least 20% in opposite directions across 2 consecutive visits associated with greater placebo response at patient level

ES = 0.09

Analysis performed by A. Kott and X. Wang, Bracket

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

27

0.4974 0.0075 0.0035 <.0001 <.0001 0.0174

• 12
• 11
• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=202) Not affected subjects (N=162) Affected subjects (N=17)

Protocol WN25309

Erratic ratings of NSFS in placebo treated patients– change of at least 20% in opposite directions across 2 consecutive visits associated with greater placebo response at patient level

ES = 0.04

Analysis performed by A. Kott and X. Wang, Bracket

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

28

0.0123 0.0009 0.0023 0.0003 0.0007 0.0042

• 12
• 11
• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=197) Not affected subjects (N=142) Affected subjects (N=20)

Protocol NN25310

Erratic ratings of NSFS in placebo treated patients– change of at least 20% in opposite directions across 2 consecutive visits associated with greater placebo response at patient level

ES = 0.04

Analysis performed by A. Kott and X. Wang, Bracket

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

29

0.0001 <.0001 <.0001 <.0001 <.0001 0.0001

• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24

Erratic ratings of NSFS in placebo treated patients – change of at least 20% in opposite directions across 2 consecutive visits associated with greater placebo response at site level*

Study week

Study Placebo response (N=199) Not affected subjects (N=161) @ 73 sites Affected subjects (N=38; 19%) @ 19 sites

Protocol WN25308

ES = 0.11

* sites with at least one patient with erratic ratings in any treatment arm

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

30

Analysis performed by A. Kott and X. Wang, Bracket

0.1286 0.0042 0.0031 <.0001 <.0001 <.0001

• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=202) Not affected subjects (N=154) @ 72 sites Affected subjects (N=48; 23%) @ 24 sites

Protocol WN25309

Erratic ratings of NSFS in placebo treated patients – change of at least 20% in opposite directions across 2 consecutive visits associated with greater placebo response at site level*

ES = 0.14

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

31

* sites with at least one patient with erratic ratings in any treatment arm

Analysis performed by A. Kott and X. Wang, Bracket

0.0323 0.001 0.0006 0.0013 0.0082 0.0894

• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=197) Not affected subjects (N=152) @ 79 sites Affected subjects (N=45; 27%) @ 26 sites

Protocol NN25310

Erratic ratings of NSFS in placebo treated patients – change of at least 20% in opposite directions across 2 consecutive visits associated with greater placebo response at site level*

ES = 0.04

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

32

* sites with at least one patient with erratic ratings in any treatment arm

Analysis performed by A. Kott and X. Wang, Bracket

Placebo response – Improvement in NSFS above 95th percentile at week 4 at patient level

<.0001 <.0001 <.0001 <.0001 <.0001 <.0001

• 12
• 11
• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=199) Not affected subjects (N=187) Affected subjects (N=12)

Protocol WN25308

ES = 0.06

Analysis performed by A. Kott and X. Wang, Bracket

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

33

Placebo response – Improvement in NSFS above 95th percentile at week 4 at patient level

<.0001 <.0001 0.0001 0.0002 0.0004 0.0001

• 12
• 11
• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=202) Not affected subjects (N=195) Affected subjects (N=7)

Protocol WN25309

ES = 0.09

Analysis performed by A. Kott and X. Wang, Bracket

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

34

<.0001 <.0001 <.0001 <.0001 <.0001 <.0001

• 18
• 16
• 14
• 12
• 10
• 8
• 6
• 4
• 2

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=197) Not affected subjects (N=186) Affected subjects (N=11)

Protocol NN25310

Placebo response – Improvement in NSFS above 95th percentile at week 4 at patient level

ES = 0.11

Analysis performed by A. Kott and X. Wang, Bracket

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

35

<.0001 <.0001 <.0001 <.0001 <.0001 0.0156

• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=199) Not affected subjects (N=169) @ 75 sites Affected subjects (N=30; 15%) @ 17 sites

Protocol WN25308

Placebo response – Improvement in NSFS above 90th percentile at week 4 at site level

ES = 0.07

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

36

* sites with at least one patient with improvement above 90th percentile in any treatment arm

Analysis performed by A. Kott and X. Wang, Bracket

0.0073 <.0001 <.0001 <.0001 <.0001 <.0001

• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20

Placebo response – Improvement in NSFS above 90th percentile at week 4 at site level

24 Study week

Study Placebo response (N=202) Not affected subjects (N=174) @ 77 sites Affected subjects (N=28;13%) @ 19 sites

Protocol WN25309

ES = 0.09

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

37

* sites with at least one patient with improvement above 90th percentile in any treatment arm

Analysis performed by A. Kott and X. Wang, Bracket

<.0001 <.0001 <.0001 <.0001 <.0001 <.0001

• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=197) Not affected subjects (N=165) @ 91 sites Affected subjects (N=32 ;16%) @ 14 sites

Protocol NN25310

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

Placebo response – Improvement in NSFS above 90th percentile at week 4 at site level

ES = 0.20

38

* sites with at least one patient with improvement above 90th percentile in any treatment arm

Analysis performed by A. Kott and X. Wang, Bracket

Placebo response – Improvement in NSFS above 90th percentile at week 4 at site level

<.0001 <.0001 <.0001 <.0001 <.0001 <.0001

• 10
• 9
• 8
• 7
• 6
• 5
• 4
• 3
• 2
• 1

Change in Marder Negative Factor Score LSMean +/- SE 4 8 12 16 20 24 Study week

Study Placebo response (N=197) Not affected subjects (N=165) @ 91 sites Affected subjects (N=32) @ 14 sites

Protocol NN25310

ES = 0.20

ES of change in placebo response 0.04 to 0.2 potentially jeopardizing signal

• detection. These observations raise the issues of patient selection and

assesssment quality.

• Erratic ratings not compatible with known rate of change in patients

⇒

Supports the use of centralized or videotaped ratings and/ or performance based and ecologically momentary assessments

• Dramatic improvement post randomization not expected in true

negative symptom patients

⇒ Need for biomarkers and behavioral characterization (e.g. Effort

choice task)

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

39

M eta-regression analysis of placebo response in neg sx trials*

• Eighteen clinical trials (12 academia, 6 industry) conducted in the last 15 years

from seventeen publications, assessing the effect of 13 drugs versus placebo on negative symptom in 998 patients on stable AP background treatment

• Placebo response was significant (p<0.001) and clinically relevant (Cohen’s d: 2.91,

95% CI: 2.05 to 3.77), but there was significant heterogeneity and high risk of publication bias.

• M ultivariable meta-regression analysis found that a higher placebo response was

significantly and independently associated with

–

Higher numbers of arms in the trial (p=0.001)

–

M ore study sites (p<0.001)

–

Industry sponsorship (p=0.001)

• Severity of negative and positive symptoms at baseline were not associated with

placebo response when controlling for other factors

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

40

* Fraguas D, Díaz-Caneja CM , Pina-Camacho L, Umbricht D, Arango C: manuscript in preparation

Adjunctive versus monotherapy?

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

41

Adjunctive versus monotherapy?

Treatment Period 1 3 weeks Wash out 14 days Treatment Period 2 3 weeks Treatment Period 3 3 weeks

§ Proof-of-M echanism (POM ) study, randomized, double-blind, placebo-

controlled, three-way crossover design

§ Six different treatment sequences (n=5) of PDE10 inhibitor RG7203 at 5 mg

QD and 15 mg QD and placebo on top of stable antipsychotic treatment

§ N=24 (completers; 33 recruited) Schizophrenia patients with negative

symptoms (mild/ moderate)

§ At end of each treatment period imaging (monetary incentive delay task) and

behavioral (effort choice task) assessment of reward anticipation and reward valuation

Wash out 14 days

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

42

M ID Task: Increased discrimination between reward and non- reward at low dose in the context of an overall blunted activation in drug conditions

Error areas represent the standard error of mean

Control Condition Anticipation of low reward Anticipation of high reward

43 p=0.039* p=0.053*

* two-sided p-value for paired t- test versus placebo

Figures show time-dependent fitted BOLD response

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

Patients choose high effort high reward option significantly less often during treatment than during placebo

44 ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France M ultiple regression shows that overall activation but not differential activation (reward anticipation versus control) is related to effortful behavior

Effect of risperidone in healthy volunteers (N=21): Risperidone blunts the overall response in the M ID task

p<0.01

• Do antipsychotics curtail potential benefits of adjunctive treatments for negative

symptoms?

• In-hous preclinical data would support that
• “ The preferred design is a double-blind comparison with placebo, especially since no

standard treatment for negative symptoms is recognized… ..(M öller et al, Working group on negative symptoms in schizophrenia, Psychopharmacology, 1994)

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

45

Conclusions

• Patient selection

–

Drop restrictions on level of positive symptoms, as long as they are not ‘disruptive’, and stratification into dominant vs predominant neg symptom patients

–

Consider enrolling patients with moderate negative symptoms (i.e. 14 and up on NSFS [scoring 0-6])

–

Include patients with shorter duration of illness

–

Consider patients who can live independently, that is have more potential for improvment, i.e. consider dropping requirement of informant

• Assessments

–

Consider use of effort choice task to characterize patients

–

Consider centralized or videotaped independent ratings

• Study design/ operational aspects

–

Keep site numbers low; in phase 3 consider separate safety studies

–

Keep number of arms to a minimum

–

Consider a monotherapy trial in patients who have predominant negative symptoms?

• Industry versus academic trials? Result of commercialization of drug development?

How can we involve academia more?

ISCTM ~ ECNP Joint Conference ▪1 September 2017 ▪Paris ▪ France

46