AU2007291235A1 - Bifeprunox doses for treating schizophrenia - Google Patents

Description

WO 2008/025781 PCT/EP2007/058958 BIFEPRUNOX DOSES FOR TREATING SCHIZOPHRENIA [001] The present invention relates to a daily dose of bifeprunox for the treatment of a patient with schizophrenia, the treatment with bifeprunox of patients with stable schizophrenia and of patients with actute exacerbations of schizophrenia and to a pharmaceutical composition comprising a dose of at least one bifeprunox compound. [002] Schizophrenia is a lifelong disabling psychiatric disorder characterized by severe and variable symptoms, including positive and negative symptoms, cognitive deficits and depression. The course of the illness can be divided into 4 major phases: premorbid, acute, stable/maintenance and late course. The premorbid phase refers to symptoms, which occur before the onset of positive symptoms. During the acute phase patients experience overt positive symptoms such as delusions and hallucinations. The stable/maintenance phase may be divided into two subphases. The first 5 to 10 years of illness are often characterized by multiple exacerbations of positive symptoms, with more stable periods interspersed between acute episodes. This subphase is followed by a plateau phase which is characterized by a stabilization of symptoms and a reduced number of exacerbations. Key treatment goals during the maintenance phase are to facilitate the patient's return to the community and establish a long-term maintenance plan. In the late course phase of the illness, positive symptoms tend to diminish with age and many patients with long-term impairments regain some degree of social and occupational competence, however, the effects of years of dysfunction are rarely overcome. [003] In the maintenance and late course phases of schizophrenia, the majority of patients continue to face a number of problems, such as a need for further symptomatic improvement and long-term control of psychotic symptoms, including prevention of relapses; maintenance of cognitive function; prevention of weight increase, hyperglycemia, and dyslipidemia; and improvement in quality of life. In addition, positive symptoms may become more resistant to treatment with each succeeding episode. Consistent with this notion, 85%-90% of patients with schizophrenia experience clinical deterioration. Further, at least half of patients given antipsychotic agents do not comply with the treatment regimen prescribed and are thus at risk for relapse. Therefore, despite intense effort aimed at improving treatment, a large proportion of patients with schizophrenia are severely disabled; they relapse often and may require hospitalization.

WO 2008/025781 PCT/EP2007/058958 2 [004] Compounds currently used to treat schizophrenia have been associated with several undesirable side effects. These side effects include weight gain, hyperprolactinemia, elevated triglyceride levels, the metabolic syndrome (markers: diabetes, hyperlipidemia, hypertension, and obesity), prolonged QTc intervals, glucose 5 abnormalities, and the exhibition of extrapyramidal symptoms. For example, prolonged QTc intervals, i.e., the corrected QT interval in an electrocardiogram, can lead to problematic heart rhythms, or heart arrhythmias. Similarly, the weight gain observed with conventional atypical antipsychotics, such as risperidone and olanzapine, has been associated with an increased risk of cardiovascular disease and diabetes mellitus. 10 [005] In addition, schizophrenia treatment using medication may be over extended periods of time. As such, these undesirable side effects affect patients on a daily basis as well as contributing to their long-term health. These side effects may also lead to noncompliance with a patient's treatment regimen. Even when treatment is for a limited and/or short duration, side effects affect a patient's willingness to comply with 15 treatment regimens. [006] Therefore, a need exists for methods for avoiding and/or reducing these undesirable side effects as well as maintaining, reducing, and /or improving baseline conditions of a patient undergoing schizophrenia treatment. [007] The present inventors have discovered that treating schizophrenia 20 patients with (a pharmaceutical composition) comprising a dose of at least one bifeprunox compound, particularly a daily dose of 20-30 mg, makes it possible to reduce and/or avoid one or more of these side effects and one or more symptoms of schizophrenia. In an embodiment of the invention, the dose is administered once-daily. Embodiments of specific doses are a 20 mg dose and 30 mg dose. Particularly preferred 25 is a daily dose of 20 mg. [008] Favourable effects of this treatment include, but are not limited to, reduction of the Positive and Negative Syndrome Scale (PANSS) total score in a patient, maintenance of body weight, maintenance and/or improvement of triglycerides levels and/or total cholesterol levels, maintenance of clinical stability of schizophrenia in 30 particular in patients with chronic, stable schizophrenia (treatment effects are e.g. an increase of the time to deterioration), improvement of one or more psychotic symptoms or maintenance and or/reduction of extrapyramidal signs and symptoms (EPS) profile similar to baseline measurements before administration. Other favourable effects are a WO 2008/025781 PCT/EP2007/058958 3 reduction of the incidence of hyperglycemia and/or one or more diabetes-related adverse events. [009] In an embodiment of the invention bifeprunoxis used for the long-term treatment of a patient with schizophrenia, in particular in a daily dose of 20-30 mg. The 5 term "long term treatment" refers for example to at least 3 months or at least 6 months treatment. [010] The embodiments disclosed herein, while providing some general overview of various embodiments of the present disclosure, are not intended to limit the scope of the present disclosure in any manner. 10 [011] Bifeprunox compounds are described in U.S. Patent No. 6,225,312 and U.S. Patent No. 7,030,241, the contents of which are incorporated herein by reference. The hydrochloric acid salt of this compound (7-[4-([1,1'-biphenyl]-3-ylmethyl)-1 piperazinyl]-2(3H)-benzoxazolone (bifeprunox) is described and claimed in W097/36893 and the monomethanesulfonate salt is described and claimed in W002/066449. In the 15 second of these patent publications, the direct formation of the monomethanesulfonate salt by the reaction between the reactive mesylate ester of N,N,N-bis(2-ethanol)-m phenylbenzyl amine and 7-amino-2(3H)-benzoxazolone is disclosed. A stable polymorph of bifeprunox monomethanesulfonate salt is disclosed and claimed in WO 2005/016898. Also included in the term "bifeprunox compounds" are bifeprunox N-oxides. A bifeprunox 20 N-oxide is disclosed and claimed in WO 2007/023141. [012] Bifeprunox compounds are indicated for the treatment of CNS (central nervous system) disorders, including schizophrenia, other psychotic disorders (in particular psychosis) and Parkinson's disease. In the framework of the present invention, dosage strength (or dose) is expressed in an amount equivalent to a 25 bifeprunox base. As used herein, the term "bifeprunox base" refers to the compound 7 [4-([1,1 '-biphenyl]-3-ylmethyl)-1 -piperazinyl]-2(3H)-benzoxazolone (INN bifeprunox) having the following formula: WO 2008/025781 PCT/EP2007/058958 4 H -~N 0 CN N [013] As used herein, the term "bifeprunox compound(s)" refers to the active compound 7-[4-([1,1'-biphenyl]-3-ylmethyl)-1-piperazinyl]-2(3H)-benzoxa-zolone, its N-oxide and pharmaceutically acceptable salts, solvates and hydrates thereof. When 5 the N-oxide is used as the bifeprunox compound, the amount in milligrams is the same amount as the amount the person skilled in the art would select for the bifeprunox compound without the oxide. In addition, pharmaceutically acceptable salts of bifeprunox or its N-oxide may be obtained using standard procedures well known in the art, for example, by mixing a compound of the present invention with a suitable acid, for 10 instance an inorganic acid or an organic acid. [014] The present disclosure is directed to the use of bifeprunox in the treatment of schizophrenia for maintaining, reducing and/or improving conditions associated with the treatment of schizophrenia by administering to a patient in need thereof a pharmaceutical composition comprising a dose of at least one bifeprunox 15 compound. For example, the pharmaceutical composition comprises at least one bifeprunox compound in an amount ranging from 5 mg to 40 mg, such as, from 10 mg to 40 mg or further for example, from 20 mg to 30 mg of a bifeprunox compound. In an embodiment, bifeprunox is used in the treatment of a patient with schizophrenia having weight problems or susceptible to weight problems. 20 [015] An embodiment of the invention relates to a daily dose of bifeprunox for the treatment of a patient with schizophrenia, wherein the dose is 20-30 mg of at least one bifeprunox compound. Particularly, said dose is useful for maintaining clinical stability in a patient with stable schizophrenia, and more particularly chronic, stable schizophrenia. Embodiments thereof are a dose of 20 mg and a dose of 30 mg, 25 respectively. [016] In a further embodiment the present invention relates to a daily dose of bifeprunox for the treatment of a patient with acutely exacerbated schizophrenia wherein WO 2008/025781 PCT/EP2007/058958 5 the dose is 20-30 mg of at least one bifeprunox compound. Embodiments thereof are a dose of 20 mg and a dose of 30 mg, respectively, used in said treatment with favourable side effects. [017] An embodiment of the present invention relates to bifeprunox for use in the 5 treatment of patients (in particular with schizophrenia) with psychoses and mood disorders wherein bifeprunox (i.e. at least one bifeprunox compound) is administered in combination with the mood-stabilizing drug lithium, and a kit for said use. [018] A further embodiment of the present invention relates to bifeprunox for use in the treatment of patients with a CNS disorder (in particular with schizophrenia) 10 wherein at least one bifeprunox compound is administered in combination with an antidepressant (in particular an SSRI, specifically paroxetine), and a kit for said use.. [019] Further embodiments of the invention relate to (methods for) co administration of bifeprunox with a CYP2C9 inhibitor (for example fluconazole), with a CYP3A4 inhibitor (for example ketoconazole and carbamazepine), with a CYP2D6 15 inhibitor (for example paroxetine) and with a H2-antagonist (for example famotidine), respectively, and kits for said treatment. [020] In an embodiment of the present disclosure, the at least one bifeprunox compound comprises bifeprunox mesylate. Preferably, the at least one bifeprunox compound is bifeprunox mesylate. The bifeprunox mesylate may be chosen from the a, 20 y, or 6 crystalline polymorphic forms, and mixtures thereof. For example, the at least one bifeprunox compound comprises at least one polymorphic form chosen from the a and y polymorphic forms. [021] The crystalline polymorphic form of a bifeprunox mesylate according to the present disclosure is defined by at least the physicochemical parameters as disclosed in 25 WO 2005/016898 [022] In another embodiment, the present disclosure provides bifeprunox mesylate in which at least about 50 percent by weight (wt. %), at least about 60 wt. %, at least about 70 wt. %, at least about 80 wt. %, at least about 90 wt. %, or at least about 95 wt. % of the bifeprunox mesylate is in the polymorphic a form. In another 30 embodiment, the pharmaceutical composition is substantially devoid of any y or 5 polymorphic forms of bifeprunox mesylate. In another embodiment, the bifeprunox mesylate provided by the present disclosure comprises less than 10 wt. %, less than 5 wt. %, or less than 2.5 wt. % of the y or 5 polymorphic forms of bifeprunox mesylate. In WO 2008/025781 PCT/EP2007/058958 6 another embodiment, at least about 99 wt. % of bifeprunox mesylate is in the polymorphic cx form. [023] The preparation of polymorphic form ax can be carried out according to the procedures decribed in WO 2005/016898. 5 [024] The at least one bifeprunox compound according to the present disclosure can be formulated into dosage forms in which the active substance is present in the solid form by methods known in the art. Examples of said dosage forms are (optionally coated) tablets, capsules, granular aerosols, suppositories and suspensions. Such dosage forms can be prepared by mixing the at least one 10 bifeprunox compound with inert pharmaceutically acceptable excipients and carriers. [025] Pharmaceutical compositions of the present disclosure can comprise at least one pharmaceutical excipient. Non-limiting examples of suitable excipients include suspending agents (for example, gums, xanthans, cellulosics and sugars), humectants (for example, sorbitol), solubilizers (for example, ethanol, water, PEG and propylene 15 glycol), surfactants (for example, sodium lauryl sulfate, Spans, Tweens, and cetyl pyridine), preservatives, antioxidants (for example, parabens, and vitamins E and C), anti-caking agents, coating agents, chelating agents (for example, EDTA), stabilizers, antimicrobial agents, antifungal or antibacterial agents (for example, parabens, chlorobutanol, phenol, sorbic acid), isotonic agents (for example, sugar, sodium 20 chloride), thickening agents (for example, methyl cellulose), flavoring agents (for example, chocolate, thalmantin, aspartame, root beer or watermelon or other flavorings stable at pH 7 to 9), anti-foaming agents (e.g., simethicone, Mylicon*), disintegrants, flow aids, lubricants, adjuvants, colorants, diluents, moistening agents, preservatives, carriers, binders (for example, hydroxypropylmethylcellulose, polyvinyl pyrilodone, other 25 cellulosic materials and starch), diluents (for example, lactose and other sugars, starch, dicalcium phosphate and cellulosic materials), disintegrating agents (for example, starch polymers and cellulosic materials), glidants and water insoluble or water soluble lubricants or lubricating agents. [026] One illustrative dosage form comprises, apart from the milled and sieved 30 dose of the active substance (bifeprunox as described herein), lactose monohydrate, microcrystalline cellulose, sodium starch glycolate (for example, type A), sodium stearyl fumarate and optionally colloidal anhydrous silica. In one embodiment, lactose is present in an amount of about 20% to about 90% by weight, about 70% to about 90% by weight, or about 75% to about 85% by weight, based on the total weight of the tablet WO 2008/025781 PCT/EP2007/058958 7 core. Microcrystalline cellulose is present in an amount of about 5% to about 90% by weight, about 10% to about 15% by weight, or about 11% to about 12% by weight, based on the total weight of the tablet core. Sodium starch glycolate (e.g. type A) is present in an amount of about 0.1% to about 2.5 % by weight, about 0.3% to about 0.7 5 % by weight, or about 0.5% by weight, based on the total weight of the tablet core. Sodium stearyl fumarate is present in an amount of about 0.1% to about 1.5% by weight, about 0.6% to about 1.3% by weight, or about 1.0% by weight, based on the total weight of the tablet core. Colloidal anhydrous silica is optionally added to the formulation in order to improve the flow properties of the powder. If desired, colloidal anhydrous silica 10 is typically present in an amount of about 0.05% to about 0.5% by weight or about 0.4% by weight, based on the total weight of the tablet core. The amount of optional coating is about 2.0% to about 5.0% by weight, about 3.0% to about 4.0% by weight, or about 3.5% by weight, based on the total weight of the tablet core. [027] In at least one embodiment, the pharmaceutical compositions 15 comprising the at least one bifeprunox compound according to the present disclosure can be administered to a subject, for example a human subject, in need thereof. [028] The present disclosure is also directed to, but not limited to, reducing a PANSS total score in a patient, maintaining body weight, maintaining and/or improving triglycerides levels and/or total cholesterol levels, maintaining clinical stability of 20 schizophrenia, improving one or more psychotic symptoms or maintaining an EPS profile similar to baseline measurements before administration. The present disclosure is also directed to methods for reducing the incidence of hyperglycemia and/or diabetes-related adverse events. These methods are exemplified in the following clinical examples provided below. 25 [029] U.S. Patent Application Nos. 10/920,361, 10/920,386, and 11/354,652 are hereby incorporated herein by reference in their entireties. All data herein are understood to be approximate and subject to normal measurement error depending, for example, on the apparatus used and other parameters influencing peak positions and peak intensities. Unless specifically defined otherwise or unless the context demands 30 otherwise, the word "about" as used herein generally means ± 5% of the recited value. BRIEF DESCRIPTION OF THE DRAWINGS [030] Fig.1 shows Panel 34 PANSS Positive Score (FAS, LOCF). [031] Fig.2 shows Panel 35 PANSS Negative Score (FAS, LOCF).

WO 2008/025781 PCT/EP2007/058958 8 [032] Fig.3 shows Panel 36 PANSS General Psychopathology Score (FAS, LOCF). [033] Fig.4 shows Panel 42 Proportion of Patients with at Least a 25% Reduction of PANSS Total Score (FAS, LOCF). 5 [034] Fig. 5 shows Panel 43 Proportion of Patients with a CGI-I Score of 2 or Less (FAS, LOCF). EXAMPLES [035] The following examples are only intended to further illustrate the present 10 disclosure, in more detail, and therefore these examples are not deemed to restrict the scope of the present disclosure in any way. [036] EXAMPLE 1. EFFICACY OF BIFEPRUNOX IN THE TREATMENT OF SCHIZOPHRENIA 15 [037] A six-week, randomized, double-blind, placebo-controlled and risperidone referenced study was used to assess the efficacy and safety of fixed doses of bifeprunox in the treatment of schizophrenia. A total of 599 subjects were randomized. [038] Treatment started with a single-blind placebo lead-in period of at least 20 three days, followed by titration from 0.25 mg up to 30 mg/day or 40 mg/day of bifeprunox for bifeprunox-treated subjects. Risperidone-treated subjects were titrated from 2 mg to 6 mg daily over a three-day period and maintained at 6 mg/day for the remainder of the treatment period. [039] Rating scale assessments were performed weekly, except for week 5, to 25 measure the change from baseline to endpoint of the PANSS total score. Other assessments included: PANSS positive symptom subscale score, PANSS negative symptom subscale score, PANSS general pschopathology subscale score, BPRS total score, BPRS psychosis score, the CGI-S score, the CGI-I score, responder rates based on the PANSS total score, and the Calgary Depression Scale for Schizophrenia (CDSS). 30 [040] Safety and tolerability measures included physical examinations, weight, waist circumference, vital signs, 12-lead electrocardiogram (ECG), clinical laboratory assessments (hematology, biochemistry, urinalysis, special laboratory assessments for prolactin, IGF-1, IGFBP-3, thyroid function), need for anticholinergic treatment during WO 2008/025781 PCT/EP2007/058958 9 double-blind treatment period, concomitant medication use, adverse event monitoring and assessments of normal movement. [041] The 20 mg bifeprunox treatment group showed a statistically significant difference from placebo for the primary endpoint of change from baseline to endpoint in 5 PANSS total score. The mean change (standard deviation) from baseline to endpoint in PANSS total score was -13.5 (20.1) for the 30 mg bifeprunox group, -10.3 (20.5) for the 40 mg bifeprunox group, -7.7 (19.2) for the placebo group, and -19.7 (19.3) for the risperidone group. [042] The 30 mg bifeprunox group showed notable differences from placebo 10 for CGI-S score, PANSS negative symptom subscale score, and PANSS positive symptom subscale score. Notable differences were also observed between the 30 mg bifeprunox group and placebo for the change from baseline to endpoint in PANSS general psychopathology subscale score, BPRS total score, BPRS psychosis cluster score, PANSS responder rate, and CGI-I responder rate. In the study, a PANSS 15 responder refers to a subject whose PANSS total score decreased by 20% or more from baseline to endpoint. A CGI-I responder refers to a subject who was categorized as "very much improved" or "much improved" in the CGI Global Improvement scale at endpoint. [043] The 40 mg bifeprunox group showed a notable difference from placebo 20 for the change in PANSS positive symptom subscale score and BPRS psychosis cluster score. [044] Decreases in body weight were seen in bifeprunox treatment groups in contrast to an increase in the placebo and risperidone treatment groups. [045] The bifeprunox groups had a lower incidence of N to H shifts in 25 triglycerides, VLDL, and LDL, and a higher incidence of N to L shifts in total cholesterol and VLDL compared with the placebo and risperidone groups. EXAMPLE 2. CLINICAL STUDIES DIRECTED TO BIFEPRUNOX IN THE TREATMENT OF SCHIZOPHRENIA 30 Example 2a - CLINICAL STUDY ONE [046] Objectives: This clinical study's primary objective was to investigate whether six weeks of treatment with 5 mg, 10 mg, or 20 mg bifeprunox is superior to treatment with placebo in adult subjects with schizophrenia, using the change from Baseline to Endpoint of the Positive and Negative Symptom Scale (PANSS) total score WO 2008/025781 PCT/EP2007/058958 10 as the primary outcome measure. The secondary objectives were to assess the efficacy of bifeprunox in treating schizophrenia using the positive symptom sub-scale score of the PANSS, the negative symptom sub-scale score of the PANSS, the general psychopathology sub-scale of the PANSS, the Brief Psychiatric Rating Scale (BPRS) 5 total score derived from the PANSS, the BPRS psychosis derived from the PANSS, the Clinical Global Impressions Severity of Illness score (CGI-S), the Clinical Global Impressions Improvement score (CGI-1), and responder rate based on the PANSS total score and CGI-I responder rate. It was also an objective of this study to To assess the safety and tolerability of bifeprunox using physical examination, weight, vital signs 10 (including pulse rate and systolic/diastolic blood pressure (BP) -both lying after five minutes and standing after two minutes, and oral temperature), 12-lead electrocardiogram (ECG), safety laboratory assessments including hematology, biochemistry, and urinalysis, need for anticholinergic treatment during double-blind treatment period, concomitant medication use, adverse event monitoring, and 15 assessments of abnormal movement including the Simpson-Angus Scale (SAS), Barnes Akathisia Scale (BAS), and Abnormal Involuntary Movement Scale (AIMS). [047] Methodology: This was a randomized, double-blind, fixed-dose, placebo-controlled, risperidone-referenced, parallel group, multi-center study in adult subjects with schizophrenia. There were five treatment groups in this study. The 20 treatment groups were as follows: bifeprunox 5 mg, bifeprunox 10 mg, bifeprunox 20 mg, risperidone 6 mg and placebo. Study medication was administered once daily. After Baseline measurements were performed, the titration phase was begun. Bifeprunox treated subjects were titrated up to 5 mg, 10 mg or 20 mg according to a standardized titration schedule (Day 1: 0.125 mg, Day 2: 0.25 mg, Day 3: 0.5 mg, Day 4: 1.0 mg, Day 25 5: 2.0 mg, Day 6: 5.0 mg, Day 7: 10.0 mg, Day 8: 20 mg). When the assigned dose was reached, subjects were maintained at that dose for the remainder of the six-week treatment period. Risperidone treated subjects were titrated to 6 mg over three days (Day 1: 2 mg, Day 2: 4 mg, Day 3: 6 mg) using a once-daily regimen. [048] Number of Subjects (Planned, Screened, Randomized and 30 Analyzed): A total of 575 patients with Schizophrenia were planned for inclusion in the study. A total of 836 subjects were screened at 40 centers and a total of 589 subjects (5 mg bifeprunox: 115 subjects; 10 mg bifeprunox: 120 subjects; 20 mg bifeprunox: 115 subjects; placebo: 119 subjects; risperidone: 120 subjects) were randomized at 37 centers.

WO 2008/025781 PCT/EP2007/058958 11 [049] Test Product, Dose and Mode of Administration: Bifeprunox tablets, total daily dose 0.125 mg to 20 mg administered orally using a once daily dosing regimen. [050] Reference Therapy, Dose and Mode of Administration: Placebo and 5 risperidone, 2 mg to 6 mg administered orally using a once daily dosing regimen. [051] Efficacy Results: The 20 mg bifeprunox dose was effective in reducing both positive and negative symptoms of schizophrenia and lessening overall psychopathology compared to placebo as shown by the statistically significant comparisons from the analysis of PANSS total and subscale scores. Subjects in the 20 10 mg bifeprunox group had a 5.8 point greater improvement from Baseline in PANSS total score compared with placebo subjects. The lower doses of bifeprunox were not effective. The 10 mg dose of bifeprunox did not show statistically greater improvement for any efficacy endpoint. The 5 mg dose showed greater improvement on some secondary efficacy measures compared to placebo, but did not show superiority to 15 placebo for the primary efficacy endpoint (PANSS total score). Risperidone 6 mg was used as an active reference in this study and showed clear separation from the placebo group. In general, the magnitude of the improvements seen in the 20 mg bifeprunox group was smaller than those seen in the risperidone group for most efficacy endpoints. [052] Safety Results: The percentage of subjects with at least one 20 treatment-emergent adverse event (TEAE) was similar across treatment groups: 89% (102 subjects) in the 5 mg bifeprunox group, 87% (104 subjects) in the 10 mg bifeprunox group, 83% (95 subjects) in the 20 mg bifeprunox group, 85% (101 subjects) in the placebo group, and 89% (107 subjects) in the risperidone group. Overall, out of 349 subjects treated with bifeprunox, the most frequently reported TEAEs were headache, 25 dyspepsia, insomnia, nausea, vomiting NOS, constipation, and agitation. TEAEs with a higher incidence (25% difference) in the 20 mg bifeprunox group (N = 114 subjects) compared with the placebo group (N = 119 subjects) included constipation, dyspepsia, and vomiting NOS. Related TEAEs with a higher incidence (25% difference) in the 20 mg bifeprunox group compared with placebo were constipation, vomiting NOS, and 30 headache NOS. The percentage of subjects with at least one severe TEAE was lowest in the 20 mg bifeprunox group (11 subjects, 10%) followed by the placebo group (15 subjects, 13%). For the remaining groups, the percentage of subjects with at least one severe TEAE ranged from 16% to 18%. The incidence of TEAEs considered to be severe was similar (<5% difference) between the 20 mg bifeprunox group and the WO 2008/025781 PCT/EP2007/058958 12 placebo group for all TEAEs. There were no dose-related trends in the bifeprunox groups in overall incidence, incidence of related, or incidence of severe TEAEs observed for any event. [053] The total number of subjects with at least one SAE was higher in the 5 active treatment groups(bifeprunox groups: 12-15%, risperidone group: 16%) compared with the placebo group (9%). The most commonly reported SAEs (25% in any treatment group) were aggravated psychosis and aggravated schizophrenia NOS. There was a slightly higher incidence of aggravated schizophrenia NOS in the 20 mg bifeprunox group (8 subjects, 7%) compared with placebo and risperidone (5 subjects each, 4%). 10 Suicide attempt was reported for one subject in the 10 mg bifeprunox group (<1%), two subjects in the 20 mg bifeprunox group (2%), no subjects in the placebo group, and no subjects in the risperidone group. The serious adverse event of suicidal ideation was reported for one subject each (<1%) in the 10 mg bifeprunox and risperidone groups (one additional subject in the 20 mg bifeprunox group experienced a non-serious 15 adverse event of suicidal ideation). No trend of dose-related increase in incidence of any SAE was observed for the bifeprunox groups. Bifeprunox was safe and well tolerated at all dose levels. The total number of subjects with at least one AE that led to discontinuation was similar among treatment groups (5 mg bifeprunox: 13 subjects, 11%; 10 mg bifeprunox: 17 subjects, 14%; 20 mg bifeprunox: 11 subjects, 10%; placebo: 20 15 subjects, 13%; risperidone: 17 subjects, 14%). The most common (reported by >2% of subjects in any treatment group) AEs that led to discontinuation were agitation, aggravated psychosis, and aggravated schizophrenia NOS. There were no treatment group differences in incidence of AEs leading to discontinuation of study medication between the 20 mg bifeprunox group and the placebo group. There was no trend of 25 dose-related increase in incidence of AEs leading to discontinuation in the bifeprunox groups. Evaluation of laboratory, vital sign, ECG and physical exam findings did not raise any unexpected safety concerns. Bifeprunox subjects exhibited a decrease in prolactin compared with the placebo group. Mild weight loss was observed in the bifeprunox groups but not in the placebo or risperidone groups. 30 [054] There were no notable differences between treatment groups in changes from Baseline to Endpoint in BAS, SAS, or AIMS scores. Use of anticholinergic medication for patients treated with bifeprunox was similar to that of patients on placebo and less than in patients on risperidone.

WO 2008/025781 PCT/EP2007/058958 13 [055] Conclusion: On conclusion of this study is that a 20 mg dose of bifeprunox given once daily for six weeks was effective in reducing both positive and negative symptoms of schizophrenia. Overall, all doses of bifeprunox were safe and well tolerated by schizophrenic subjects. No dose-response relationship of safety/tolerability 5 was seen. Example 2b - CLINICAL STUDY TWO [056] Primary Objectives: To investigate whether six weeks of treatment with fixed doses of bifeprunox (30 mg/day or 40 mg/day) can demonstrate superior efficacy 10 compared with placebo in adult subjects with schizophrenia, using the change from Baseline to Endpoint of the Positive and Negative Symptom Scale (PANSS) total score as the primary outcome. [057] Secondary Objectives: To assess the efficacy of bifeprunox in treating schizophrenia using the positive symptom sub-scale score of the PANSS, the negative 15 symptom sub-scale score of the PANSS, the general psychopathology sub-scale of the PANSS, Brief Psychiatric Rating Scale (BPRS) total score derived from the PANSS, BPRS psychosis score derived from the PANSS, the Clinical Global Impression Severity of Illness score (CGI-S), the Clinical Global Impression Improvement score (CGI-1), responder rate based on the PANSS total score, Calgary Depression Scale for 20 Schizophrenia (CDSS), and subject satisfaction rating. Pharmacokinetic (PK) data of bifeprunox in schizophrenic subjects were also assessed and are presented in a separate report (combined with data from study S1543003). To assess the safety and tolerability of bifeprunox using physical examination, weight, waist circumference, vital signs (pulse rate and systolic/diastolic blood pressure [BP] -both lying after five minutes 25 and standing after two minutes, and oral temperature), 12-lead electrocardiogram (ECG), clinical laboratory assessments (hematology, biochemistry, urinalysis, special laboratory assessments for prolactin, IGF-1, IGFBP-3, and thyroid function), need for anticholinergic treatment during double-blind treatment period, concomitant medication use, adverse event (AE) monitoring, and assessments of abnormal movement including 30 the Simpson-Angus Scale (SAS), Barnes Akathisia Scale (BAS), and Abnormal Involuntary Movement Scale (AIMS). [058] Methodology: This was a Phase Ill, six-week randomized, double blind, placebo-controlled, risperidone-referenced, parallel-group, multi-center study of the efficacy, tolerability, and safety of bifeprunox in adult subjects with schizophrenia.

WO 2008/025781 PCT/EP2007/058958 14 Subjects who completed this study had the option to continue in the long-term. The treatment groups were: bifeprunox 30 mg/day, bifeprunox 40 mg/day, risperidone 6 mg/day, and placebo. After completing a Single-Blind placebo lead-in period of at least three days, bifeprunox-treated subjects were titrated from 0.25 mg up to 30 mg/day or 40 5 mg/day according to a standardized titration scheme over an eight-day period. When the assigned dose was reached, subjects were maintained at that dose for the remainder of the six week treatment period. Risperidone-treated subjects were titrated from 2 mg to 6 mg daily over a three-day period and were subsequently maintained at 6 mg/day for the remainder of the treatment period. Rating scale assessments for efficacy and abnormal 10 movement disorders were done weekly except for Week 5. Safety assessments were performed at Screening, during treatment and at the end of the study. Subject satisfaction with study medication was assessed at Week 6. Samples of blood were obtained at Weeks 2, 4, and 6 for determination of bifeprunox in plasma. Blood samples were also obtained at Screening/Baseline, Week 3, and Week 6 for clinical laboratory 15 assessments. [059] Number of Subjects (Planned, Consented, Randomized and Analyzed): A total of 576 subjects with schizophrenia were planned for inclusion in the study. Of 783 screened subjects, a total of 599 subjects were randomized (30 mg bifeprunox: 148 subjects, 40 mg bifeprunox: 148 subjects; placebo: 149 subjects; 20 risperidone: 154 subjects). [060] Diagnosis and Main Criteria for Inclusion: Male or female subjects 18-75 years of age with schizophrenia (per DSM-IV-TR criteria). Subjects had to have a total score on the PANSS between 70 and 120; at least two of four PANSS items (conceptual disorganization, hallucinatory behavior, suspiciousness, unusual thought 25 content) had to have a score 4; and the score on the CGI-S had to be at least 4. [061] Reference Therapy, Dose and Mode of Administration: Placebo and risperidone, 6 mg administered orally once daily. [062] Efficacy Results: The 30 mg bifeprunox treatment group showed a statistically significant difference from placebo for the primary endpoint of change from 30 Baseline to Endpoint in PANSS Total Score (LOCF) based on the Hochberg adjusted p value (adjusted p=0.020). The 40 mg bifeprunox treatment group was not significantly different from the placebo group (adjusted p=0.156) for the primary efficacy endpoint. The mean change (SD) from Baseline to Endpoint in PANSS total score was -13.5 (20.1) for the 30 mg bifeprunox group, -10.3 (20.5) for the 40 mg bifeprunox group, -7.7 (19.2) WO 2008/025781 PCT/EP2007/058958 15 for the placebo group, and -19.7 (19.3) for the risperidone group. The treatment effect values (for mean change from Baseline at Endpoint [LOCF]) corresponding to the difference between bifeprunox and placebo were: -5.9 for the 30 mg bifeprunox group and -3.2 for the 40 mg bifeprunox group. No statistically significant difference based on 5 the planned step-down procedure was seen in change from Baseline to Endpoint in CGI S for the 30 mg bifeprunox treatment group compared to placebo. Consequently, the differences between placebo and 30 mg bifeprunox groups for change from Baseline to Endpoint in PANSS Negative Symptom and PANSS Positive Symptom subscale scores were not evaluated using the stepdown procedure for statistical significance compared 10 to placebo. The 30 mg bifeprunox group showed notable differences from placebo for CGI-S score (nominal p=0.028), PANSS Negative Symptom subscale score (nominal p=0.027), and PANSS Positive Symptom subscale scores (nominal p=0.010). [063] Notable differences were observed between the 30 mg bifeprunox group and placebo for the change from Baseline to Endpoint in PANSS general 15 psychopathology subscale score (p=0.025), BPRS total score (p=0.019), BPRS psychosis cluster score (p=0.002), PANSS (30%) responder rate (p=0.019), and CGI-I responder rate (p=0.039). No notable differences at Endpoint were observed between the 30 mg bifeprunox group and placebo for subject satisfaction (p=0.051), CGI Improvement score, or CDSS score. For the 40 mg bifeprunox group, no statistically 20 significant difference from placebo was seen for the primary efficacy parameter. No notable differences from placebo were noted for the 40 mg bifeprunox group for any of the secondary efficacy parameters with the exception of PANSS Positive Symptom Subscale score (p=0.020) and BPRS Psychosis cluster score (p=0.031). [064] Safety Results: The percentage of subjects with at least one TEAE 25 was comparable in the bifeprunox dose groups (74%-76%) and was higher than the placebo group (64%) but slightly lower than in the risperidone group (78%). Treatment emergent AEs with a higher (25% difference) incidence in the bifeprunox groups compared with the placebo group included nausea, vomiting, constipation, dyspepsia, diarrhoea, and dizziness. No difference between bifeprunox treatment groups was 30 generally observed in the incidence of individual TEAEs in the bifeprunox groups. Of TEAEs occurring in at least 5% of subjects in any treatment group, those having a higher (22% difference) incidence in the 40 mg bifeprunox group compared with the 30 mg bifeprunox group included nausea, vomiting, toothache, anorexia, akathisia, dizziness, headache, and insomnia. In contrast, dry mouth, salivary hypersecretion, decreased WO 2008/025781 PCT/EP2007/058958 16 appetite, sedation, somnolence, anxiety, and vaginitis occurred with a higher (22% difference) incidence in the 30 mg bifeprunox group compared with the 40 mg bifeprunox group. The incidence of severe TEAEs was generally low (51% incidence) with the exception of TEAEs of psychotic disorder (56%) and schizophrenia (3% each in the 5 bifeprunox and placebo groups and 2% in the risperidone group). The incidence of severe TEAEs was generally comparable across treatment groups and there was no difference between bifeprunox treatment groups in the incidence of severe TEAEs for any event. Special interest TEAEs were defined prior to database lock and included events related to suicide, suicide attempt, sexual dysfunction, syncope, vasovagal 10 attack, and postural hypotension. A total of 76 subjects (13%) overall reported at least one TEAE of special interest during the study. The percentage of subjects with at least one special interest TEAE was higher in the bifeprunox and risperidone treatment groups (12%-16%) compared with the placebo group (6%). The majority of TEAEs of special interest occurred in 51% of subjects in any treatment group. Dizziness was the 15 most commonly reported and occurred at a slightly higher incidence in the bifeprunox treatment groups compared with the other two groups. [065] A total of 60 subjects (10%) overall experienced 81 SAEs (including deaths). Most SAEs occurred in 51% of subjects. The exceptions were psychotic disorder (55%) and schizophrenia (3% in each of the four treatment groups). The 20 incidence of SAEs was comparable among bifeprunox groups and placebo. There were no notable differences between treatment groups on measures of abnormal movement (BAS, SAS, or AIMS scores). Use of anticholinergic medication for subjects treated with bifeprunox was similar to that of subjects treated with placebo. Overall, there were no clinically meaningful changes in clinical laboratory parameters, physical examination 25 findings, or ECG readings with bifeprunox. The bifeprunox groups had a lower incidence of N to H shifts in triglycerides, VLDL, and LDL and a higher incidence of N to L shifts in total cholesterol and VLDL compared with the placebo and risperidone groups. The incidence of markedly abnormal total cholesterol values was similar across treatment groups (1% to 2%). Markedly abnormal triglyceride values were reported by slightly 30 fewer subjects in the 40 mg bifeprunox group compared with other groups. There was a slightly greater incidence of N to H shifts in tri-iodine thyronine in the 30 mg bifeprunox and risperidone groups (6% each) compared with the placebo (4%) and 40 mg bifeprunox (5%) groups and slightly greater incidence of N to L shifts in TSH in the 30 mg bifeprunox (4%) and the 40 mg bifeprunox (5%) compared with the placebo (2%) WO 2008/025781 PCT/EP2007/058958 17 and risperidone (<1%) groups. Small comparable decreases in body weight were seen in bifeprunox treatment groups in contrast to an increase in the placebo and risperidone treatment groups. The incidence of markedly abnormal decreases in body weight was slightly higher in both the bifeprunox treatment groups relative to the placebo and 5 risperidone groups, while the incidence of markedly abnormal increases in body weight was higher in the risperidone treatment group relative to the bifeprunox and placebo groups. [066] Conclusion: One conclusion of this study is that a 30 mg dose of bifeprunox given once daily for six weeks was effective in reducing the symptoms of 10 schizophrenia as shown by the statistically significant comparisons from the analysis of PANSS total scores. Subjects in the 30 mg bifeprunox group had a 5.9 point greater improvement from Baseline to Endpoint in PANSS total score compared with placebo subjects based on LOCF data. The statistically significant differences noted between the active control, risperidone and placebo treatments for the primary and secondary 15 efficacy parameters demonstrate that this is a valid study. [067] The 30 mg bifeprunox treatment group showed a notable difference from placebo (based on nominal p-values) for the three key secondary efficacy endpoints, change from Baseline to Endpoint in CGI-S, PANSS Negative, and Positive Symptom subscale scores; statistical significance was not achieved based on the step 20 down procedure for these three key secondary efficacy endpoints. Notable differences were observed between the 30 mg bifeprunox group and placebo at Endpoint for most of the other secondary efficacy parameters (change from Baseline to Endpoint in PANSS general psychopathology subscale score, BPRS total score, and BPRS psychosis cluster score). Notable differences between the 30 mg bifeprunox dose and placebo 25 were also demonstrated at Endpoint for PANSS and CGI-I Responder rates. The 40 mg bifeprunox treatment group did not show statistically significant differences from placebo for the primary efficacy parameter. There were no notable differences between the 40 mg bifeprunox and placebo groups for any of the secondary efficacy parameters with the exception of PANSS Positive Symptom Subscale and BPRS cluster score. Overall, 30 30 mg and 40 mg doses of bifeprunox were safe and well tolerated by schizophrenic subjects.

WO 2008/025781 PCT/EP2007/058958 18 [068] Example 2c: CLINICAL STUDY THREE Primary Objective: To investigate whether six weeks of treatment with fixed doses of bifeprunox (20 mg/day or 30 mg/day) can demonstrate superior efficacy compared with placebo in adult subjects with schizophrenia, using the change from Baseline to 5 Endpoint of the Positive and Negative Symptom Scale (PANSS) total score as the primary outcome. Secondary Objectives: To assess the efficacy of bifeprunox in treating schizophrenia using the Positive Symptom sub-scale score of the PANSS, the Negative Symptom sub scale score of the PANSS, the General Psychopathology sub-scale of the PANSS, Brief 10 Psychiatric Rating Scale (BPRS) total score derived from the PANSS, BPRS psychosis score derived from the PANSS, the Clinical Global Impression Severity of Illness score (CGI-S), the Clinical Global Impression Improvement score (CGI-1), CGI-I responder rate, PANSS responder rate based on the PANSS total score, Calgary Depression Scale for Schizophrenia (CDSS), and Subject Satisfaction rating. To assess the safety and 15 tolerability of bifeprunox using physical examination, weight, waist circumference, vital signs (pulse rate and systolic/diastolic blood pressure [BP], and oral temperature), 12 lead electrocardiogram (ECG), clinical laboratory assessments (hematology, biochemistry including fasting insulin level, fasting glucose, fasting lipid profile, urinalysis, special laboratory assessments for prolactin, IGF-1, IGFBP-3, and thyroid 20 function), need for anticholinergic treatment during double-blind treatment period, concomitant medication use, adverse event (AE) monitoring, and assessments of abnormal movement including the Simpson-Angus Scale (SAS), Barnes Akathisia Scale (BAS), and Abnormal Involuntary Movement Scale (AIMS). Methodology: This was a Phase Ill, six week randomized, double-blind, placebo 25 controlled, olanzapine-referenced, parallel group, multi-center study of the efficacy, tolerability, and safety of bifeprunox in adult subjects with schizophrenia. There were four treatment arms in this study, approximately 144 subjects per arm. The treatment groups were: bifeprunox 20 mg/day, bifeprunox 30 mg/day, olanzapine 15 mg/day, and placebo. After completing the Single-Blind Placebo Lead-in Period of at least three days, 30 bifeprunox-treated subjects were titrated from 0.25 mg up to 20 mg/day or 30 mg/day according to a standardized titration scheme over a seven- or eight-day period, respectively. When the assigned dose was reached, subjects were maintained at that dose for the remainder of the six week treatment period. Olanzapine-treated subjects began dosing at 10 mg/day for the initial seven day period and, were subsequently WO 2008/025781 PCT/EP2007/058958 19 maintained at 15 mg/day for the remainder of the treatment period. Subjects were hospitalized (if not already inpatients) after eligibility for study entry was verified, starting from the Screening Visit until at least 10 days after Baseline. Subjects could be hospitalized longer than 10 days if considered medically necessary by the Investigator. 5 Rating scale assessments for efficacy and abnormal movement disorders were done weekly except for Week 5. Safety assessments were performed at Screening, during treatment and at the end of the study. Subject satisfaction with study medication was assessed at Week 6. Samples of whole blood were obtained at Weeks 2, 4, and 6 for determination of bifeprunox in plasma. 10 Number of Subjects (Planned, Consented, Randomized and Analyzed): A total of 576 subjects with schizophrenia were planned for inclusion in the study. A total of 814 subjects were screened at 32 centers and a total of 604 subjects (20 mg bifeprunox: 154 subjects, 30 mg bifeprunox: 150 subjects; placebo: 150 subjects; olanzapine: 150 subjects) were randomized at 32 centers. 15 Diagnosis and Main Criteria for Inclusion: Male or female subjects 18-75 years of age with schizophrenia (per DSM-IV-TR criteria). Subjects must have had a total score on the PANSS between 70 and 120; at least two of four PANSS items (conceptual disorganization, hallucinatory behavior, suspiciousness, unusual thought content) must have had a score 4; the score on the CGI-S must have been at least 4. 20 Test Product, Dose and Mode of Administration: Bifeprunox tablets, total daily dose 20 mg or 30 mg (one 20 mg tablet and one 10 mg tablet) administered orally using a once daily dosing regimen. Reference Therapy, Dose and Mode of Administration: Placebo and olanzapine, 5 mg and 15 mg administered orally using a once daily dosing regimen. 25 Efficacy Results: Both the 20 mg and 30 mg treatment groups showed improvements over Baseline at Endpoint but did not demonstrate efficacy as compared to the placebo group with respect to the primary and key secondary efficacy parameters. However, the 20 mg bifeprunox group showed notable improvement over the placebo group for the change from Baseline to Endpoint in a secondary efficacy parameter, CGI improvement 30 score (nominal p = 0.027). Additionally in one other secondary efficacy parameter, the difference between the 20 mg bifeprunox group and the placebo group approached being notable (p = 0.061) for PANSS-20%-responder rate. However, these occurrences of notable and nearly notable differences among the secondary efficacy parameters do not exceed what is expected to happen by chance, i.e., in 5% of the treatment WO 2008/025781 PCT/EP2007/058958 20 comparisons. In all other secondary and key secondary parameters, neither bifeprunox treatment group showed notable improvement over the placebo group for any efficacy endpoint. Olanzapine at a dose of 15 mg was used as an active reference in this study. The difference of the PANSS total score between olanzapine and placebo was analyzed 5 in accordance with sensitivity analyses. These results showed that olanzapine was notably different from placebo (p < 0.001). In general, the magnitude of the improvements seen in the bifeprunox dose groups was higher than those seen in the placebo group, but lower than those seen in the olanzapine group for most efficacy endpoints. 10 Safety Results: The percentage of subjects with at least one treatment-emergent adverse event (TEAE) was highest in the 20 mg bifeprunox group (126 subjects, 82%) followed by the 30 mg bifeprunox (115 subjects, 77%), olanzapine (110 subjects, 73%), and placebo (107 subjects, 72%) groups. Overall, out of 304 subjects treated with bifeprunox, the most frequently reported TEAEs were headache, nausea, vomiting, 15 dyspepsia, and insomnia. TEAEs with a higher incidence (25% difference) in the bifeprunox groups compared with the placebo group included nausea, vomiting, and constipation. In general, no clear dose-related increase in the incidence of individual TEAEs was observed in the bifeprunox groups. TEAEs with a slightly higher (22% difference) incidence in the 30 mg bifeprunox group compared with the 20 mg bifeprunox 20 group included fatigue, dizziness, and sedation. The percentage of subjects with at least one severe TEAE was comparable in the bifeprunox treatment groups and the placebo group (9% to 12%) and slightly less in the olanzapine treatment group (6%). There was no clear indication of a dose-related increase in the incidence of severe TEAEs in the bifeprunox groups for any event. Special interest TEAEs were defined prior to database 25 lock and included the following events: suicide, suicide attempt, events related to sexual dysfunction, syncope, vasovagal attack, and postural hypotension. A total of 58 subjects (10%) overall reported at least one TEAE of special interest during the study. The total number of subjects with at least one special interest TEAE was higher in the bifeprunox treatment groups (30 mg bifeprunox: 20 subjects, 13%; 20 mg bifeprunox: 17 subjects, 30 11%) compared with the placebo (11 subjects, 7%) or olanzapine (10 subjects, 7%) groups. The majority of TEAEs of special interest occurred in 51% of subjects in any treatment group. The most commonly reported TEAE of special interest was dizziness which had a slightly higher incidence in the bifeprunox treatment groups compared with the other two groups (30 mg bifeprunox: 15 subjects, 10%; 20 mg bifeprunox: 13 WO 2008/025781 PCT/EP2007/058958 21 subjects, 8%; placebo: nine subjects, 6%; olanzapine: eight subjects, 5%). Other special interest TEAEs occurring in at least two subjects within a treatment group included syncope vasovagal (30 mg bifeprunox: two subjects) and orthostatic hypotension (30 mg bifeprunox: three subjects). 5 The percentage of subjects with at least one SAE was least in the olanzapine treatment group (six subjects, 4%) followed by the bifeprunox groups (20 mg: 15 subjects, 10%; 30 mg: 12 subjects, 8%), with the highest incidence of SAEs noted in the placebo group (20 subjects, 13%). The most commonly reported SAEs were psychotic disorder (4% of subjects overall) and schizophrenia (2% overall). The incidence of these SAEs was 10 similar or lower in the bifeprunox groups compared with the placebo group. Two subjects in the 30 mg bifeprunox group had a SAE of syncope vasovagal compared with no subjects in the other treatment groups. Apart from this possible exception, there were no other notable indications of a doserelated increase in the incidence of any other SAE observed for the bifeprunox groups. The percentage of subjects who discontinued study 15 medication due to an AE was greatest in the placebo group (11%) followed by the bifeprunox groups (8% each) and the olanzapine treatment group (6%). The percentage of subjects with at least one AE that led to study termination was greatest in the placebo group (12%) with comparable percentages of subjects having at least one AE leading to study termination in the 20 mg bifeprunox (8%), 30 mg bifeprunox (7%), and olanzapine 20 (6%) treatment groups. The most common (reported by >2% of subjects in any treatment group) AEs that led to discontinuation of study medication were psychotic disorder and schizophrenia. There was no clear trend of a dose-related increase in the incidence of AEs leading to discontinuation of study medication in the bifeprunox groups. Evaluation of laboratory, vital sign and ECG findings did not raise any unexpected safety concerns. 25 Subjects in the bifeprunox group exhibited a decrease in prolactin compared with subjects in the placebo and olanzapine groups. The incidence of markedly abnormal decreases in body weight was comparable among bifeprunox treatment and placebo groups (5% to 6%) and was lower in the olanzapine treatment group (<1%). The incidence of markedly abnormal increases in body weight was comparable in the 30 bifeprunox and placebo groups (1% to 3%) and much higher in the olanzapine (19%) group. There were no notable differences between the treatment groups in changes from Baseline to Endpoint in BAS, SAS, or AIMS scores. Use of anticholinergic medication for subjects treated with bifeprunox was similar to that of subjects on placebo.

WO 2008/025781 PCT/EP2007/058958 22 Discussion and conclusion: This was a six-week randomized, double-blind, fixed dose, placebo-controlled, parallel group, multi-center study of the efficacy, tolerability and safety of bifeprunox with olanzapine as an active reference in the treatment of 604 subjects with schizophrenia. The study was conducted at 32 centers in the United States 5 (26), Colombia (3), and India (5). This was a valid study evidenced by the results demonstrating statistically significant differences between the active control olanzapine, and placebo treatment for the primary efficacy parameter. 10 Both the 20 mg and 30 mg treatment groups showed improvements over Baseline at Endpoint but did not demonstrate efficacy as compared to the placebo group with respect to the primary and key secondary efficacy parameters. However, the 20 mg bifeprunox group showed notable improvement over the placebo group for the change from Baseline to Endpoint in a secondary efficacy parameter, CGI improvement score 15 (nominal p = 0.027). Additionally in one other secondary efficacy parameter, the difference between the 20 mg bifeprunox group and the placebo group approached being notable (p = 0.061) for PANSS-20%-responder rate. However, these occurrences of notable and nearly notable differences among the secondary efficacy parameters do not exceed what is expected to happen by chance, i.e., in 5% of the treatment 20 comparisons. In all other secondary and key secondary parameters, neither bifeprunox treatment group showed notable improvement over the placebo group for any efficacy endpoint. Olanzapine at a dose of 15 mg was used as an active reference in this study. In general, the magnitude of the improvements seen in the bifeprunox dose groups was higher than 25 those seen in the placebo group, but lower than those seen in the olanzapine group for most efficacy endpoints. In contrast, in the large study in Clinical Study One, the 20 mg bifeprunox dose was effective in reducing both positive and negative symptoms of schizophrenia and lessening overall psychopathology compared to placebo as shown by statistically 30 significant comparisons from the analysis of PANSS total and subscale scores. In the present study, using the observed values analysis of change from Baseline to Week 6 in the PANSS total score, results were more similar between the 20 mg bifeprunox (-24.42 [15.6]), 30 mg bifeprunox (-24.56 [17.02]), and olanzapine (-29.11 [16.88]) groups, however the placebo group also demonstrated similar results (-22.29 WO 2008/025781 PCT/EP2007/058958 23 [19.14]). This indicates that subjects who stayed in the study for six weeks responded well to their treatment regimen. Bifeprunox was well tolerated at both dose levels. The rate of withdrawal due to adverse events was lower in the bifeprunox groups compared with the placebo group. Adverse 5 events appearing more frequently in bifeprunox treated subjects than in placebo subjects were mainly gastrointestinal in nature and mild to moderate in severity. Only two subjects (in the 20 mg bifeprunox group) discontinued study medication due to gastrointestinal AEs (one subject discontinued due to nausea, one subject discontinued due to nausea and vomiting). The percentage of subjects with at least one SAE was 10 lower in the bifeprunox groups (7% to 8%) compared with the placebo group (13%). The most commonly reported SAEs were psychotic disorder (4% overall) and schizophrenia (1% overall). The incidence of these SAEs was similar or lower in the bifeprunox groups compared with the placebo group. Two subjects in the 30 mg bifeprunox group had a SAE of vasovagal syncope compared with no subjects in the other treatment groups. 15 Apart from this possible exception, there were no other notable indications of a dose related increase in the incidence of any other SAE for the bifeprunox groups. Evaluation of laboratory, vital sign, and ECG findings did not raise any unexpected safety concerns. Bifeprunox was associated with a decrease in prolactin consistent with what has been seen in previous studies and expected on the basis of the partial 20 dopamine agonistic profile of the drug, and was not associated with any AEs. In contrast, increases in prolactin were observed in the placebo and olanzapine groups. Small decreases in mean body weight were noted in the bifeprunox groups while an increase in weight was noted for the olanzapine group. The incidence of markedly abnormal decreases in body weight was comparable among bifeprunox treatment and 25 placebo groups (5 % to 6%) and was lower in the olanzapine treatment group ( 1%). The incidence of markedly abnormal increases in body weight was comparable in the bifeprunox and placebo groups (1% to 3%) and much higher in the olanzapine (19%) group. These findings are consistent with what has been observed in other bifeprunox studies and are noteworthy given that weight gain has been problematic for atypical 30 antipsychotics and is associated with an increased risk of cardiovascular disease and diabetes mellitus. There were no notable differences between treatment groups on measures of abnormal movement (BAS, SAS, or AIMS scores). Use of anticholinergic medication for subjects treated with bifeprunox was similar to that of subjects on placebo. The incidence of WO 2008/025781 PCT/EP2007/058958 24 extrapyramidal disorder was low in all groups (bifeprunox groups:<1 % to 3%; olanzapine: 1%; placebo: 3%). The 20 mg and 30 mg doses of bifeprunox did not show statistically significantly greater improvement compared with placebo for most efficacy endpoints. Better CGI improvement scores were seen in the 20 mg (but not 30 mg) 5 bifeprunox group compared with the placebo group. Nausea, vomiting, and constipation were the more notable AEs relative to placebo. Overall, 20 mg and 30 mg doses of bifeprunox were safe and well tolerated by subjects with schizophrenia. 10 Example 2d - CLINICAL STUDY FOUR [069] Primary Objective: to investigate whether 6 months of bifeprunox treatment is superior to treatment with placebo in patients with chronic schizophrenia, using the time to deterioration from randomization as the primary outcome measure. [070] Secondary Objective: to investigate whether the acute effect after 6 15 weeks of bifeprunox treatment is superior to treatment with placebo, using the change from baseline in Positive and Negative Syndrome Scale (PANSS) total score as the outcome measure. [071] Other Secondary Objective: To evaluate the long-term safety and tolerability of bifeprunox versus placebo. 20 [072] Methodology: This study was a multinational, multicentre, randomised, double-blind, parallel-group, placebo-controlled, fixed-dose study. The study consisted of a 3- to 6-day antipsychotic-free run-in period, after which patients were randomised to 6 months of double-blind treatment with fixed doses of bifeprunox (20mg/day (BX20) or 30mg/day (BX30)) or placebo (PBO). Patients allocated to the BX groups were up 25 titrated from 0.25mg/day over 7 days (BX20) or 8 days (BX30), and then continued on these doses for the remainder of the study. Efficacy assessments were made at baseline (except CGI-1) and at Weeks 1, 2, 4, 6, and 9, and at Months 3, 4, 5, and 6. Safety assessments were performed at screening, during treatment, and at the end of the study. At predetermined time points, blood samples were obtained for drug 30 concentration analysis of BX and its major metabolites (3'- and 4'-sulfate conjugates of BX), and pharmaco-economic assessments were performed. [073] Number of Patients Planned and Analyzed: There were a total of 495 patients that were planned for enrolment: 165 in each treatment group.

WO 2008/025781 PCT/EP2007/058958 25 [074] Diagnosis and Main Inclusion Criteria: Patients with a primary diagnosis of schizophrenia, according to DSM-IV-TR criteria, for more than 2 years, who: had a PANSS total score 260 and a CGI-S score 4 (moderately ill) at screening and baseline; had PANSS items P7 (hostility) and G8 (uncooperativeness) scores 54 5 (moderate) at screening and baseline; were between 18 and 65 years of age (extremes included); were inpatients, partially hospitalized, or outpatients followed up in a day care program within 90 days; and prior to screening had no modification of antipsychotic medication within 1 month prior to screening. [075] Investigational Product, Dose and Mode of Administration, Batch 10 Number: Bifeprunox - up-titration over 7 (20 mg) or 8 days (30 mg) to 20 or 30 mg once daily; encapsulated tablets, orally. [076] Reference Therapy, Dose and Mode of Administration: Placebo encapsulated tablets, orally. [077] Efficacy Results: The primary efficacy variable was the time to 15 deterioration and the analysis was based on the FAS. The primary efficacy analysis rejected the hypothesis of equal time to deterioration of schizophrenia in the three treatment groups (Cox model, p = 0.008). Subsequent pairwise comparisons of each of the BX groups and the PBO group showed that patients in the BX groups had a statistically significantly longer time to deterioration of schizophrenia than patients in the 20 PBO group (BX20: p = 0.008 and BX30: p = 0.006). The proportion of patients who deteriorated was 59% in the PBO group, 41% in the BX20 group, and 38% in the BX30 group. The Cox proportional hazards model gave an estimated hazard ratio of 0.66 (BX20) and 0.65 (BX30) relative to PBO; that is, the risk of deterioration was approximately 1.5 times higher for patients in the PBO group than for patients in the 25 BX20 or BX30 groups. Bifeprunox was also statistically significantly superior to PBO in the analysis of time to deterioration based on the PPS. Since most of the patients in the PPS participated for most of the study, the results were very close to the results of the primary analysis, both for the estimated hazard ratios and the p-values obtained. This illustrates the robustness of the conclusion of the primary efficacy analysis. The 30 secondary efficacy variable was the PANSS total score at Week 6. The adjusted mean change from baseline to Week 6 in PANSS total score (FAS, LOCF) for each BX group (BX20: -4.0; BX30: -2.7) was statistically significantly greater than that for the PBO group (1.1) (BX20: p = 0.002; BX30: p = 0.017) according to Hochberg's Step-up Method.

WO 2008/025781 PCT/EP2007/058958 26 [078] For the development over time in PANSS total scores, PANSS positive and general psychopathology subscale scores, BPRS total scores, and BPRS psychosis cluster scores (all FAS, LOCF), the same general pattern was seen: initially (until Weeks 6 or 9), the mean scores decreased and subsequently they stabilized in both BX groups, 5 whereas in the PBO group, the mean scores decreased until Weeks 2 or 4, after which the scores increased steadily. For each of these variables, pairwise comparisons of each of the BX groups with the PBO group for the FAS using the LOCF (ANCOVA) showed that treatment with BX (either dose) was generally statistically significantly superior to treatment with PBO from Week 6 or 9 onwards. 10 [079] The mean PANSS negative subscale scores (FAS, LOCF) decreased in both BX groups until Week 6, after which the scores stabilised. In the PBO group, the scores decreased until Week 4, after which the scores tended to increase. In the per visit LOCF analysis (FAS, ANCOVA) of the mean PANSS negative subscale scores, treatment with both BX doses was statistically significantly superior to that with PBO at 15 all time points, except at Weeks 2 and 4 for the BX30 group. [080] For all treatment groups, the mean CDSS total scores (FAS, LOCF) decreased initially (baseline to Week 2), after which the scores either remained stable (both BX groups) or increased (PBO group) over time. In the per-visit LOCF analysis (FAS, ANCOVA) of the mean CDSS total scores, there were no statistically significant 20 differences between either of the BX groups and the PBO group at any time point. This result should be seen in the context that the baseline CDSS scores were low, reflecting a low level of depressive symptomatology. [081] The mean CGI-S scores (FAS, LOCF) decreased in both BX groups until Week 6, after which the scores stabilised. In the PBO group, the mean CGI-S 25 scores decreased until Week 4, after which the scores tended to increase. In the per visit LOCF analysis (FAS, ANCOVA) of the mean CGI-S scores, the BX groups were statistically significantly superior to the PBO group at Months 3, 4, 5, and 6 for BX20 and Months 5 and 6 for BX30. [082] The mean CGI-I scores (FAS, LOCF) decreased in both BX groups until 30 Week 6, after which the scores stabilized. In the PBO group, the scores decreased minimally until Week 2, after which the scores increased steadily. In the per-visit LOCF analysis (FAS, ANCOVA) of the mean CGI-I scores, both BX doses were statistically significantly superior to PBO from Week 6 onwards.

WO 2008/025781 PCT/EP2007/058958 27 [083] For the small number of patients who fulfilled the criteria for predominantly negative symptoms (PBO: 25 patients; BX20: 32 patients; BX30: 36 patients), the adjusted mean change from baseline in PANSS total scores (all groups) and in PANSS positive subscale scores (PBO and BX20) generally followed the same 5 patterns and were within the same range as those for the overall population. In contrast, the adjusted mean change from baseline in PANSS positive subscale score for patients treated with BX30 followed the same pattern over time as for the overall population but the change was larger. The adjusted mean change from baseline in PANSS negative subscale scores (all groups) followed the same pattern over time as the overall 10 population; however, the mean change was generally twice as large in this population relative to the overall population. In the per-visit LOCF analysis (ANCOVA) of PANSS total scores, PANSS positive subscale scores, and PANSS negative subscale scores, BX20 was not statistically significantly different from PBO at any time points. BX30 was statistically significantly superior to PBO in PANSS total scores from Week 6 onwards 15 and in PANSS positive subscale scores from Week 2, onwards whereas BX30 was not statistically significantly different from PBO from Week 2 onwards in PANSS negative subscale scores. [084] For the small number of patients who fulfilled the criteria for predominantly depressive symptoms (PBO: 18 patients; BX20: 15 patients; BX30: 22 20 patients), the PANSS total scores, the PANSS positive subscale scores, the PANSS negative subscale scores, and the CDSS total scores followed generally similar overall patterns as those for the overall population. In the per-visit LOCF analysis (ANCOVA) of the PANSS total scores, PANSS positive subscale scores, PANSS negative subscale scores, and CDSS total scores, there were no statistically significant differences 25 between either of the BX groups and the PBO group at any time point in this sub population, probably due to the low number of patients and the low level of depressive symptoms present at baseline. [085] The proportion of patients with at least a 25% reduction in PANSS total score (FAS, LOCF) in the BX20 group was statistically significantly larger than that in the 30 PBO group from Week 9 onwards, whereas the proportion of responders in the BX30 group was statistically significantly larger than that in the PBO group at Month 5 only. [086] A small proportion (0% to 8%) of patients (irrespective of treatment) had a reduction 35%, 45%, or 55% in PANSS total score. There was no trend between or WO 2008/025781 PCT/EP2007/058958 28 within treatment groups with respect to the distribution in the proportion of patients with a specific reduction. [087] The proportion of patients with a CGI-I score 52 (FAS, LOCF) at Week 6 and Month 6 was larger in the BX groups (Week 6: BX20 17%; BX30 19%; Month 6: 5 BX20 22%; BX30 20%) than in the PBO group (Week 6: 11%; Month 6: 9%). The differences between the BX groups and the PBO group were statistically significant (p<0.05) from Week 9 onwards (BX20) and from Week 9 onwards except Month 3 (BX30). [088] The mean total cholesterol and mean LDL calculated decreased from 10 baseline to Month 6 in all groups (except non-fasting PBO patients) irrespective of treatment and fasting/non-fasting condition. The mean VLDL calculated and triglycerides decreased from baseline to Month 6 in all groups (except non-fasting BX30 patients) irrespective of treatment and fasting/non-fasting condition. The mean HDL increased from baseline to Month 6 in all groups irrespective of treatment and fasting/non-fasting 15 condition. The adjusted mean HDL cholesterol values increased from baseline to Month 6 in all three treatment groups irrespective of fasting/nonfasting condition (PBO: 0.04/0.06 (fasting/non-fasting); BX20: 0.07/0.08; BX30: 0.07/0.08mmol/L). There were no statistically significant differences between either of the BX groups and the PBO group. 20 [089] The adjusted mean triglycerides values decreased from baseline to Month 6 in all three treatment groups irrespective of fasting/non-fasting condition (PBO: 0.06/-0.22 (fasting/non-fasting); BX20: -0.16/-0.21; BX30: -0.37/-0.03mmol/L). There were no statistically significant differences between either of the BX groups (except BX30 (fasting)) and the PBO group. 25 [090] The adjusted mean fasting glucose values increased from baseline to Month 6 in all three treatment groups (PBO: 0.10; BX20: 0.13; BX30: 0.09mmol/L) and there were no statistically significant differences between either of the BX groups and the PBO group. [091] The adjusted mean weight change from baseline to Month 6 (APTS, 30 OC, ANCOVA) was -0.8kg in the PBO group, -0.3kg in the BX20 group, and -0.5kg in the BX30 group. The adjusted mean weight decreases in the BX20 and the BX30 group were not statistically significantly different from that in the PBO group. [092] In all treatment groups, patients lost weight irrespective of whether they had nausea and/or vomiting, although those patients who also had nausea and/or WO 2008/025781 PCT/EP2007/058958 29 vomiting had a greater weight decrease (PBO: -0.6 versus -1.9kg; BX20: -1.0 versus 1.9kg; BX30: -1.1 versus -2.3kg). [093] There was no treatment effect of BX (either dose) on the patients' status of metabolic syndrome during the study. Approximately 75% (range: 70% to 80%) of the 5 patients did not have metabolic syndrome at baseline or at the end of the study. There were no statistically significant treatment differences in the patients' status of metabolic syndrome. [094] There were no clinically relevant changes within or differences between treatment groups in clinical laboratory values, vital signs, metabolic syndrome, or ECG 10 parameters. [095] Conclusions: One conclusion of this study is that both doses of bifeprunox (20 mg/day and 30 mg/day) prevented deterioration of schizophrenia statistically significantly better than placebo. For patients with chronic, stable schizophrenia, the baseline condition was significantly better maintained with BX (both 15 doses) than with PBO either after 6 weeks of treatment or after long-term treatment. A comparison of the safety profiles of bifeprunox and placebo showed a higher incidence of adverse events related to the gastrointestinal system and dizziness in the bifeprunox groups relative to the placebo group. The incidence of nausea and vomiting leading to withdrawal from the study and abnormal movements was higher in the BX30 group than 20 in the BX20 group. A favorable metabolic profile (based on weight changes, blood lipids and the presence/absence of metabolic syndrome) was seen for bifeprunox. RESULTS RELATED TO CLINICAL STUDIES ONE - FOUR PANSSTOTALSCORE 25 [096] CLINICAL STUDY ONE: The mean change (S.D.) from Baseline to Endpoint in PANSS total score was -9.7 (17.5) for the bifeprunox 5 mg group, -5.0 (18.3) for the bifeprunox 10 mg group, -11.3 (17.0) for the bifeprunox 20 mg group , -5.3 (16.3) for the placebo group, and -15.7 (14.9) for the risperidone group. The treatment effect values corresponding to the difference between bifeprunox and placebo mean change 30 from Baseline at Endpoint (LOCF) were -4.1, 0.6, and -5.8 for the bifeprunox 5 mg, 10 mg, and 20 mg groups respectively. From the pairwise comparisons, a statistically significantly greater decrease at Endpoint (LOCF) was seen for the 20 mg bifeprunox group versus placebo (p-value adjusted for multiple comparisons, p=0.031). No WO 2008/025781 PCT/EP2007/058958 30 significant treatment group differences were seen for the 5 mg bifeprunox and 10 mg bifeprunox treatment groups compared to placebo, respectively. Change front Baseline in PANSS Total Score Last Obiervanon Carried Fonvard at Each Visit lntent-ft-Treat Popilation Treat-t Group Bifepranox Bifrprunx Bifeprunos Risperidine StPistim 10 mg 20 mug Plaebo 6 n3 Total Nmdaier of Subjects in teITT N 0 I1S 11 114 116 Biveline n 110 118 i1 114 1i6 Men~ ~ ~ ~~~( 1 .3 9 ( ) 9 .5(16 92 9 (12. 2', 92.1 12-3) 09 6 Medin 9SD. 93~ljjjj 929J 91 82 9 C 6 a- Max 73-2I 67-1'30 -12 5k 0-120 G136 Chsage jfimelin W I 110 S 11i 114 116 Mlean (S.D.) -3.9 (1.) -2. 9 1.? 63(11 -4. 10.1 ) -. (1 J Median 4 -1 -5 Mm- -50 - -- 24 1-1 P-Qlue (ad4 i.D LDO G-3l42 W'eeL 2 L 10 118 11 114 116 Mean (SD} -. 13 42(4 ) - 14. -5.1 (13 6) - 3. (13 9)1 -f n+ (SD, 7. (3$ 14. Media 5 Mm - Max -62 -76 47 - 37 -95 - 0 -44 - 33 -69 -4 P-Vahlc (a'dj) 0.696 LO000 0.11 Week 3 10 1S 11 114 116 Me 5.D.) -7.(16).1 -4 6 ' 11.2 14.5) -6,0 1535) -'5-3 (14-1) ___________Media -7- 11- I _M_ - Max -65 - 76 -54 -41 -S9 - 33 -46- -64- 4 _____________P-Vluecv (a f l 1.0 0 47 Week 4 is t1 S1 1M14 1 Mesn (SD.) -S8.7 (1 4.1 (17 2) 11 2 ( 9 -6.5 (16f0) -4A (14.5) Median -1 -6-3 Min - Max -55 - 7- -5; -41 -6 - 33 -41. - -66- 14 P-Vane adj) ' 1.625 1.000 0 097 EndpcL n1 11: 114 116 Mea (3D-. -9 (1.5 4.0 (1 I -11 ' 170 -5 3163) -157(14.9) AL 1 . Median -8 4 -11 -6 -14 mmn -Miax -66 - 76 -57- 41 -6103 - 33 -2-3 -57 - 1I Tn Effect -4.1 S.6 -53 TEtect CIu 9.2, 0.9) (-44,15)(11 I1 -GA) P- hle 'adj)i 0x2 .031* NoIt: * S-ttcally signiicant after adj'vt r Multiple compariw ymg the Sep Down Dumett' P.rc -edure wiL an overall Type I Errr f 4 P-vaThmes that !ar placebo- aie 7idjusted t. Lu00, Cafideuce tev9 we'e A derved sin accos dance wh the Step Downa Dmea px e .J . Note: Paitwesie ecmpatxiscns versus placebo are fased ox au ANCOVA e< with tent (excluding sispetidae) andpoledenter a.-, fixed fitors 91nd baseline PANISS Ital seem as c . Not': No PANSS &CCrs were record at Week I Ca Wek 2 for Subjezt 11259 in the bifepnmesacz' 2§a'L grf.

WO 2008/025781 PCT/EP2007/058958 31 [097] CLINICAL STUDY TWO: The mean change (SD) from Baseline to Endpoint in PANSS total score was -13.5 (20.1) for the 30 mg bifeprunox group, -10.3 (20.5) for the 40 mg bifeprunox group, -7.7 (19.2) for the placebo group, and -19.7 (19.3) for the risperidone group (as indicated in the below Table). The treatment effect values 5 (for mean change from Baseline at Endpoint [LOCF]) corresponding to the difference between bifeprunox and placebo were: -5.9 for the 30 mg bifeprunox group and -3.2 for the 40 mg bifeprunox group. A statistically significant treatment group difference was seen for the 30 mg bifeprunox treatment group compared to placebo at Endpoint based on the Hochberg adjusted p-value (p=0.020). Differences between 30 mg bifeprunox and 10 placebo treatment groups were also noted at Week 2 through Week 4 from the nominal p-values (p:0.004 at each of Week 2 through Week 4). In addition, the treatment effect increased consistently over the 6 weeks of study (range of effect: -5.6 at week 2 and -6.3 at week 4; Table 3.0.1). No statistically significant difference between 40 mg bifeprunox group and placebo group was evident at Endpoint based on the Hochberg adjusted p 15 value.

WO 2008/025781 PCT/EP2007/058958 32 Change from Baseline, in PANS Total S'core Last Observation Can' iied Fvward at Each Viit Intet-to-Treat Population -'Caaient BfnBif ifepmnox Pdsperide Ttal Nmans of Subjects t ITT NPouat 4a (3 ' 9 -'.1. 1 11 22i Change fcrm Baselme \ieany\D>3M 6 1 n ax 6-0.5 5'i '4e, a ,4 ~ ~ 0 1.2 5!L Vs5 1. 4t S WVeek 14 nn M-etn (SD Medim 10 1 In -VAj U? 09.2 t Media !-Ve 4t 44 Wee'k 461"1 Pj 1"S Median 1,n 3tt t 4 ' 7s 71 -4 4 . S. 4 6s Men (SD)v, 97CI -I1 441 S~ Mreia P-:::I te Raw) I"1 CI . 1 6 P-vnlue s f'i, P pmi- Zk1t m te C.""S Irfie acchngm * te Hoc t-r wTF ~ i N~r T~~31?±±~X ~r3 elll b 'v-4 do ii ANCOl\A rnoca'wi Ue'irniw (n~udg me~~m) a®w iLee n~ fi xedf-ar B'rEtne P&K\' -mvi wue m *t cL-7;"Mediae WeMk 3 A Eainr' t ' tre"" -lb aECdpdrin Ni Hch~e~ phocedlre, ~tMpm7T~n ~\ ~ 'el: 4 e 2y6Sl pri"eFupnrs cC'h7 krt E Affec &$urC': - '%'k Tr-I he 3 Ra 1 [098] CLINICAL STUDY THREE: The mean change (SID) from Baseline to Endpoint in PANSS total score was -13.8 (19.9) for the 20 mg bifeprunox group, -13.1 (20.2) for the 30 mg bifeprunox group, -10.7 (19.4) for the placebo group, and -22.0 5 (18.2) for the olanzapine group (Table 15). The treatment effect values (for mean change from Baseline at Endpoint [LOOF]) corresponding to the difference between bifeprunox and placebo were: -3.5 for the 20 mg bifeprunox group and, -2.2 for the 30 mg bifeprunox group. No statistically significant treatment group differences were seen for the 20 mg or 30 mg bifeprunox treatment groups compared to the placebo group WO 2008/025781 PCT/EP2007/058958 33 based on the Hochberg adjusted p-values. Similarly no notable differences were observed between the placebo group and either of the two bifeprunox dose groups at any other time point during the study (Week 1 through Week 4). When data were analyzed only for those subjects having the same rater for each visit, differences in 5 mean change (SD) from Baseline to Endpoint in PANSS total score (LOCF) were similar to those observed in the primary analysis. The difference of the PANSS total score between olanzapine and placebo was analyzed in accordance with sensitivity analyses. These results showed that olanzapine was notably different from placebo (p < 0.001). Change from Baseline in PANSS Total Score Last Observation Carried Forwardat Each Visit Intent-to-Treat Population BiRfpnox Bifqrnox Olanzapin&* StatIne 2 m 3 m Platele 1I mg Ton INumber of Subjeds irh the ITT Population N 144 14J Ba shne 1 149 445 14J M e: SDn 9 9(11. 0) 156 (10 6) 043 .3 96 17 Median 94 96 94 9 Min Max 71-119 7i. 121 7.i1 0 2 Week I 14P 1444 Me. (S D- 6" (1.7) -6.1 (123) -6.4 (1 1-() -93(1 ) Mda-5-6 -1,1 -6 P-ale 0.618 0.790 149 14,J Me (D -0 (48) -,( O-04 13.7) - 15.1(13: 6) MeanD -W0 -9 - -1 Min Ma -7. 4 -Q7 d5- )2S-6 P-vale 0.043 Week 3 1: 149 "4 15 146 ed i an 131 Miv Max -66, 4P -59, 05 4, 36 -69 P-ahu 0.594 0.6555 Week 4 1 149 4-5 146 Enfear SDF VmDi h J ' Me-SI-s 3,) -2R (189.6) -1".3 (1A 2. 16. 5) Mi Max-7, 49Q -62, 65 -62, 36 -69. 19 Week 6, 1: 149 "4 15 140 E ndpo n LOC:F M."em ( nSD - 9 8 (1.9 -31( .) -17(19A) -20(8 Medi -14 --1 Min' MaRIx -d8. 53 -7.5 -68,' 36 -79. 1 T37 Effectl -7 8. 0.9 -6.6. -22 T Effect -8.4,1 712 P-vale (d41 0 .31 Note: T, eahtm emparismsvrsspcb were based o an A-NC OVA medel vah nreatmerm excbding ohainzs) ma otoed centa a fxed factors and Bsehne PANSS total secse as. a covaria t Note: Sigmicance at Week ) iEndpoi for nuaplme copm as ea aed co3rdmg to zth Hochbeg procedure. Noe -ius and CS pressed fo Week I tLroughi Week 4, ad t pvlen W-eek 6n, ae for 4eic dec rt purposes Only *OLanzapme wasm exchided &Lrom ttselcomparisons of tzeatmnt groups,, as data from thi treat tmenm gs up Was to be considered n suppoxive anahse onslya Tl= Treateq , Adi = Adtred. Data Sone: Table 3 .. 0, Table .I WO 2008/025781 PCT/EP2007/058958 34 [099] CLINICAL STUDY FOUR: The adjusted mean change from baseline to Week 6 in PANSS total score for each BX group (BX20: -4.0 and BX30: -2.7) was statistically significantly greater than that for the PBO (1.1) group (BX20: p = 0.002; BX30: p = 0.017). 5 Panel 26 Adjusted Mean Change from Baseline in PANSS Total Score Week 6 (FAS. LOCF. ANCOVA) Least Squares DIfference to PRO Est I otes Treatment Days n Mean SE mean SE 95% CI 95% C1 p-value Group Lower Upper PB) 42 166 1 06 1,17 BX20 42 158 -4. 01 1.20 -5.07 1 5&i -8. 9 -195 4, 002 BX3o 42 172 -2.65 1.13 -3.71 1,55 -6.76 -O i6 0,)17 1-214 ET PTS11 f 9JUN2006:1 23: 26 13/313 - TFL/SAD Euild: Numbers - F !naI 10 PANSS Positive [0100] CLINICAL STUDY ONE: The Table below presents mean PANSS positive subscale scores at Baseline and the mean change from Baseline by visit using LOCF for the ITT population. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were -1.1, 0.7, -1.5 for the bifeprunox 5 mg, 10 mg, and 20 mg groups 15 respectively. A statistically significantly greater decrease (unadjusted p=0.037) at Endpoint (LOCF) was seen in PANSS positive subscale score for the comparison of the treatment effect estimates for the bifeprunox 20 mg group and placebo.

WO 2008/025781 PCT/EP2007/058958 35 Change from Baseline in PANSS Positive Subscale Score Last Observation Carried Forward at Each Visit Lntent-fo-Trean Population Treatment Group Bifepruno Bifeprnoi Bifrpruns Risp-eriine Smatske 5 mg 10 Ing 20 ng Placebo 6 ng T tal Nunbe of Snbjetainthe ITT N 110 11 Ill 1:14 116 Population 110 111 114 11 24.5 ( ) 24.6 (4.) 249 (3.4) 244 (4.1) 24O (3,S 4 25 24 Min-NTx 1 1 5 835 -3 1 hA-39 6-3 Chns:e Snm Base-m Week I 11 1S 110 114 116 D -. 1(3-) -0.6(3.9) -17 (3.5) -) - 1 ) Mfedim -1 -1--12 Mm'ln-MSax -10 - 14 -2- - -61 -- 3 7 Mv'.~~t) Ii4 _ It '~ P-rahlue 0.B 0.238 .2 Week 21 11811 114 M6 Mean (S.D) -2.1 (4,,) 2(4 -1- (. -2-1 (4A) -43 4,2) Median- -1 -3-2 MinE - Ma-- - 1.1 -1 - 11 -14 - I -17--6 P-value 'S 0.06a ,38 Week3 U 1 1 11 114 11 Is Mean (S.D.) 2. 6 (4) -1. J 1(5 -3-3 (4.4) -23 (49) -0 (4 6) Medi- 1 -5 Min - x -1 14 -14 -106 -13 - 11 -16 - 1 -16 P-value .1 .3 0 11 Week 4 10 11 114 1 6 Mean" {.D.) 1 3 4.8) -2. 3 5. -494.6 -4 1- -4 Mlm - Max -15 14 -3-16 -17 - I11 -13 - 10 -7 P-1 0.32- 0.G5I 00t4 EnKdptm LOCF n 11 114 1'6 Mean (S.D.) 2. (. -5.3 (4 8) - Media -3 1 -5 Mm -hMn -21 14 -18 -16 -1S - 1 -13 -14 TrEffet .i .7 -1. TrifctC1 -. '3 (-'.21 (2 ,0 . aue 0.111 .339 0,3 * Simificat at the 0 e,50 eret Note: Paiwise compianens versP plac are bawed on nM ANCOVA ide wit teame (erchding rmeri de) nd poled Entel as fixed facbars 2d baseline PAN SS positive bsucale scte a- a caare 'Note: No ?ANSS .ecres were recorded at WeI or Week 2 fcr &bjct 11259 in the bifepaox 2 0mg groni. Data Source: Table 3.10 [0101] For the observed values analysis of change from Baseline, similar trends to the PANSS total score were seen in the change from Baseline in PANSS 5 positive subscale score in the bifeprunox treatment groups over time. The pairwise WO 2008/025781 PCT/EP2007/058958 36 comparisons between the bifeprunox and placebo groups were not statistically significant. 5 [0102] CLINICAL STUDY TWO: The Table below presents PANSS Positive Symptom subscale scores at Baseline and the change from Baseline by visit using LOCF for the ITT population. The mean change (SD) from Baseline to Endpoint in PANSS Positive Symptom subscale scores was -4.5 (6.6) for the 30 mg bifeprunox group, -4.2 (6.9) for the 40 mg bifeprunox group, -2.5 (6.0) for the placebo group, and 10 7.2 (6.6) for the risperidone group. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were: -1.9 for the 30 mg bifeprunox group, and -1.7 for the 40 mg bifeprunox group. SEP1 did not yield a significant result, and therefore SEP2 and SEP3 were not evaluated (see Section 7.4.1.1). However, a notable difference between the bifeprunox 30 mg treatment group and placebo was observed at Endpoint (nominal 15 p=0.01). Notable differences between 30 mg bifeprunox and placebo treatment groups were also observed at Week 2 through Week 4 (p:0.006). [0103] Similarly, notable differences between the bifeprunox 40 mg treatment group and placebo were observed at Endpoint (p=0.020). Differences relative to placebo were also observed at Week 1 through Week 3 (p: 0.013).

WO 2008/025781 PCT/EP2007/058958 37 Cinge from Baseline in PANSS Positive Symptom Subscale Score Last Obsenation Caried Fonard at Each Visit Intent-to-Treat Population arep net Group Staatc i.rmox BipansRiprdone 30 Mg 40 g Placeba 6 111 Total Niun:J? 3f Sujct_ th ITT Populationg a IN 40 141 144 51 140 14 1 144 151 25,lean 25 2 25 25 ]ifin. Ma 1.413 17371.3 Chaae mmBasehne Week 140 143 149 ,lean -2a- -3 .

x6 6 -14. 13 P-va0u 0 19 2 13is a 4040 144 151 !/eaSD) -41 (5IS.3) -3 1 57. -I-19 (1. -53 (4 7) Mein -3B -l- -5 Mi, mx -18, I1 - 2 14 3 - 6 - u 0 1 4.M06 Week 3 140 14C 144 151 M ,esa (SD) -4,5(16 -4. (13) - 16 -6.3(3 ta-ledian -4 -2 n. -19. 1 ?-value ?001 0. r Wedepk 4 n1 140 14 144 151 Mvie- - 4A (- . (.) - ( - . (1) Medin -'4- -2 Mi 13 22 13 22, 6 F-value 0 00Z 6 o. 76 W"-Eak &6Eadpint n1 1.40 141 144 151 Mve- -SD) -4,.5 (6) 42(,6- 6) 7266 Media -4-4 M n mx -19,111 -21, 13 -3, 13 -22,. 18 Traffect -1,9 Dr95% CI -3 4: -h5k-3.1 -> 3 9:7.5% C1 -3,6 -A .4 & No-E:Treet comar""isona VerSus pla.ebOwrehdon ACA model wi th tmet(xldn rd7 ee'nez and c ente r as fix7.-ed feton Und Esshe PALNSS poiie , npics sub-se cre aa Nate: p-ale re=ned far Week 1 through- W,'eek 'I 'me fr daescriptiv prosony Daa Soutrce: Tbe .Q ,abe3.1.0. [0104] CLINICAL STUDY THREE: The Table below presents PANSS Positive Symptom subscale scores at Baseline and the change from Baseline by visit using 5 LOCF for the ITT population. The mean change (SD) from Baseline to Endpoint in PANSS Positive Symptom subscale score was -4.3 (6.1) for the 20 mg bifeprunox group, -4.2 (6.8) for the 30 mg bifeprunox group, -3.5 (6.2) for the placebo group, and -7.0 (5.8) for the olanzapine group. The treatment effect values (bifeprunox - placebo) at Endpoint WO 2008/025781 PCT/EP2007/058958 38 LOCF were: -1.1 for the 20 mg bifeprunox group and, -0.7 for the 30 mg bifeprunox group. No differences with nominal p-values (5 0.05) were observed between the placebo group and either of the bifeprunox dose groups at Week 6/Endpoint or at any other timepoint during the study (Week 1 through Week 4). Change from Baseline in PANSS Posidive Symptom Subscale Score Last Observation Carried Forward at Each Xsit Intent-to-Treat Population Treat tment Ca Bifeprmex BifeprLaox OLuainapine \ ai0c 20 mg 30 m1 Pame 15 mg Totfl Nmbe Sbjct n he ITT Popution N 149 14 145 146 RateEhne n 149 148 145 146 MeanL (kD) 24. (4 25. (4.1) 254 (39 "5 36 Mean .4 25325 2 Mist M ;'ax 161, 32 N6_ 3 17 ,35 17, S5 Change fiom Baeline Week n 1 144 14 5 Mean (SD- - 1- Miedian -2 - 2 Mi loM x -13. 1 1. 3-0 1 P-4Ade iA338 Week 2 a 17 145 146 Mean(S) -3.0 .- ( -3.5 4.6) -4.( 4 0) Median -3 -5 M 'L. M x -1 13 -11 1.8-6 Pe0 54 4 149 18 145 146 MIean (SD), -3 .2) -3.6 (&6.-) -4.1 1C.5) -.0(9 Median- -4 -4 -6 Min, Max -19,13 -9 -21 2 -20, 6 P-rahke O.493 241 Week 4 3 149 148 145 146 Mean (SD -v4 2 - - (4 M%'edinn -4 - 3 Mist M';-"ax -1 I1 22 5-5 9 -0 P-Talue 08 92 WVeek, EsdpTint n 149 '49 145 146 Mean(SDJ -4.3- 6.1) -4.2(2 "-) -5&) -O58 Median -4 -37 NT ax -19 14 -2 Tri Effecl- -1 1 - ).7 cezfr ' trior: vi 'BnehLn PAO" Spoq>t ve rlp ern ni. .e o me 5 -grfizL~he -, z4 '"'S'J:r zmpri'c s WXV' evm ' w x c tz ih' -:* b,.te'-p,,L ')YZ Ng, P i 'cl C; 4uT"'I Eek -2 e5 c -214 Ea0dukrab Lx ;prnzo& 'a-I ticc: Tcb4e 3fD.',T ES .0 5 WO 2008/025781 PCT/EP2007/058958 39 [0105] CLINICAL STUDY FOUR: The mean PANSS positive subscale scores (FAS, LOCF) decreased over time in both BX groups until Week 9, after which the scores increased minimally. In the PBO group, the mean PANSS total scores decreased until Week 2, after which the scores increased steadily (Fig. 1; Panel 34). In the per-visit 5 LOCF analysis (FAS, ANCOVA) of the PANSS positive subscale scores, treatment with BX20 or with BX30 was statistically significantly superior to that with PBO from Week 6 onwards. PANSS Negative 10 CLINICAL STUDY ONE: The Table below presents mean PANSS negative subscale scores at Baseline and the mean change from Baseline by visit using LOCF for the ITT population. The bifeprunox 5 mg, 20 mg, and risperidone 6 mg groups showed the greatest improvement over time. Estimates of the treatment effect (bifeprunox placebo) at Endpoint LOCF were -1.0, -0.3, -1.4 for the Bifeprunox 5 mg, 10 mg, and 20 15 mg groups respectively. A statistically significantly greater decrease in PANSS negative subscale score was seen at Week 3 (unadjusted p=0.013) and at Endpoint (LOCF) (unadjusted p=0.026) for the bifeprunox 20 mg group versus placebo.

WO 2008/025781 PCT/EP2007/058958 40 Change from Baseline in FANSS Negatve Subscale Score Last Observation Carried Fnrward at Each Visit Iitent-tn-Treat Population Treatment Grou Bifeprunnr Bifeprun cr Bifeprancr Ri peiri)done Statistic m 14 ng 20 Ing Pacebo - g Total 1 1 unber of Sujec in the ITT N 10 113 11 114 Baselie ti i1 14 Mean (S.D.) 2 A (4 ) 23.5 (4.5> 2.6 (4.8) 23.1 (4.9) 21.9 4.4) Min-Max I3 4 2-37 14-4 2 Change fZ(i Ba7Em-ne Wed I U I 10 is I1I1U 114 Mean (S.D.) - ( -1. 13A) - (3 ) -2.1 f3.5t Median -1- 1 -1 -1 Mn13 -I - 6 - -8 -11 -s -16 -5 P-nte1s5 Wek2 10 18 II 114 116 Mean (SD. -1.8s (44) - 4 -20 (4.2i -0 43 B -3 .i(4-2 -Median -2 1 -- 3 Mn-Max. -21 -1 -5- -1 1 -13 5 2 P-.1e "1 09 C - OG54 Week3 0 118 .11 114 Mean (S.D.) -1. s (4. 7j -1 7( ) -. 4 - L-2(4 6) 3: 42 Median - Min-Max -21- 1 15 -1S-11 -13-16 -2'-6 P-value D.290 1-526 cj13* Week 10 1 11 114 6 Nless.'S.fl1 1 -' -3 { Mean RD.) -2 B(4.9i -5 (4.-5) -25 4 -L; (4 5) -3 -2 ? 3 Median -1 Mat - Max -19 - 13 - 10 -1 - 14 6 -9-6 P-am0. 248 C.9017 j.05 Endpint LOCE U 1 1 1 11U1 Meca (SD.) 2.2 (4 S) - - (4 -2.6 (53) - 3 (4, 6) -3.6 (46) Mediian - 1- 1 Mi -Max -- 1; - -1 -16-14 -6 - -21-7 e -i.- -1.4 T'if__CI ( 3) (L Y' (-2.6.-0. al0.118 i.6C 6 6* * Significanta. athe i0$50 'ee1 Ncte: Pauw-e conprnsos v p are baed cm eae Ax1 a ACOVA mx odel wi4h tremment excluded g inperidcne) and pgCdled center as Eixed Pter a a'. ANSS ney'e subscn e sc t a covanWSe Note: No PAN$S acOre - ere recorded at Wceek 1 I Week Sor Subje t 11259 n the bifepnwcx 2 g sp. Data S3 rje: Table 3.2.0 For the observed values analysis of change from Baseline in PANSS negative subscale score, decreases in the scores over time indicated improvement in each of the 5 bifeprunox treatment groups. A statistically significant treatment effect of bifeprunox was observed at Week 2 (p=0.032) for the bifeprunox 5 mg group compared to placebo. No other significant differences were noted. CLINICAL STUDY TWO: The Table below presents PANSS Negative Symptom subscale scores at Baseline and the change from Baseline by visit using LOCF for the 10 ITT population. The mean change (SD) from Baseline to Endpoint in PANSS Negative Symptom subscale scores was -3.1 (5.6) for the 30 mg bifeprunox group, -2.2 (5.4) for the 40 mg bifeprunox group, -1.8 (5.6) for the placebo group, and -3.8 (5.5) for the WO 2008/025781 PCT/EP2007/058958 41 risperidone group. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were: -1.4 for the 30 mg bifeprunox group and -0.9 for the 40 mg bifeprunox group. SEP1 did not yield a significant result, and therefore SEP2 and SEP3 were not evaluated. 5 The 30 mg bifeprunox treatment group showed a notable difference from placebo (nominal p=0.027) at Endpoint. Notable differences between 30 mg bifeprunox and placebo treatment groups were also noted at Week 2 through Week 4 (p:0.021). No differences were observed between placebo and the 40 mg bifeprunox dose group at any time point during the study (Week 1 through Week 6).

WO 2008/025781 PCT/EP2007/058958 42 Change from Basdine in PANSS Negative Synptoin Su bscal Score Last Observation Caiied Forward at Each Visit hnteknt-to-Treat Population Treatment Gticup 5itcBifepnmax ~ Bifepnx Rsperide'ne 0mg 4 g Plcebo 6 mg ±e ITT Poulation 140 141 144 151 Baehne n 140 141 144 1 D ((4A 2 ('5.0) 22(0% 1i Mx 19 3 Chamge ftom Basehe Week 1 n 140 139 1431 MeanD -1 -6 (37 - ( -L edan M'Ein Mu: -19 10 -16, 7 -15 1I -1 1 1 P-value 0 35S9 0,851 Wek2n140 40 144 151 Men(D) -. 6 (A) -L14,( l5 (4-5) -2. 9 (43-) Median 1 M4in, Msx -19, 9 -4, 9 -18 12 -is, 9 W-al e I- 144 Wee 3 ni 14014 1 44 Men (SD -3. ( -2() -1 , 3 -33 (4.9) M '-3 Min, Max% -19, 9 -2 9 -1S, 14 -2,9 P-al0 02 1 .18 ek~ 4 n 140 14 144 1 Median -3 --- 1 Mn x-24, 9 -24 9 -19, 14 -2,. P- he 0.19 .2 Week 6'Bndpoint n 4 4 14415 M'ea! (S)-31() -. 2(4 -LS (5 6) -3(5 mda -3- -1 -4 Miu, \Px -4 -4, 9 ~ -114 -26.9 9h% CI -26 0. '-t, . 9-.

C-

Note: Trennmes2 compwm3.: :rm p 1&tcebz were (3d an ANCOXA model wiLh tre1aetment fxcludin riperidne) and pooled heater a- fsied factNE ad BIseline PANT5 negative . mpom s-scae tXmo as a Note: p-aies pressed ftr Week 1 enough XeTek 4 aue fbr descrpMt puarpoe& mnly. Dat.a Source: Table 10. Table 10. CLINICAL STUDY THREE: The Table below presents PANSS Negative Symptom subscale scores at Baseline and the change from Baseline by visit using LOCF for the ITT population. The mean change (SD) from Baseline to Endpoint in 5 PANSS Negative Symptom subscale score was -3.3 (5.0) for the 20 mg bifeprunox group, -3.1 (5.2) for the 30 mg bifeprunox group, -2.4 (5.1) for the placebo group, and 4.6 (5.2) for the olanzapine group. The treatment effect values (bifeprunox - placebo) at WO 2008/025781 PCT/EP2007/058958 43 Endpoint LOCF were: -0.9 for the 20 mg bifeprunox group and -0.6 for the 30 mg bifeprunox group. No differences with nominal p-values (5 0.05) were observed between the placebo group and either of the bifeprunox dose groups at Week 6/Endpoint or at any other time point during the study (Week 1 through Week 4). Change from Baseline in PANSS Negative Symptom Subscale Score Last Observation Caried Forward at Each Visit Intent-to-Treat Population Treatsen: Gm Bifepnmox Bifepnmaox Oanapine 20mgw0mg Pncebo 15m Tctml Nlumbron Sbet ns the I11 Populain N 149 148 14 14 Bnehne U 149 148 145 146 Mean (SD) 230(.) 2. 4.) 2k9 (4? 215 (4.) Median232234 Mi. ax14. 35 14. 34 1 13 39 Chnge from Bnele Week I U 149 146 144 145 Me (D) 1-5 (3 ) -1.3 (4.0) -. 2 (3.2) Median -1 -1 1 -I M I M-1x -11, 9 -14, 27 - S -5 P-value, 04612 0. 9 26 WeA 49 147 144 ,k - Mean SD -2. 36 -1.9 (4.3) -2.2 (4A4) -3.10 (18I) M-'edian- -1-2 M I -2x 12 -14, 23 -19. 13: -li. P-value 0.2S1 0 41 WeA 3 49 148 14 14 Mean (SD) - 3.-2 (,4A4 -2,5 ("4.8 -'2A4 -4.9) - 4. 0 (4A3) Medi- -2-23 Mn Max -1 9 18 2.3 - 19, is -2a, P-valdue FL.075 0.,79 7 Week 4 U 8 145 146 Me"3n (D) - 3(.)-.6(0 1 4) -4 D (4.) Median - -4 Mi.Max -1 I -17. -19. 1S -24. S P-value 0s058063 Week 6Tndpxint n 149 14 845 14 6 Mean (SD) -3 (00) - -4 (s1) -4.6 (5-2) Medi am 3- 2 Mi.n -1 1 -i3, 3 -9 18 Tr Effie -0 -0.6 Tr Effcs 1.6, D.5 95% CI P-7value .9 0. 317 Nite: Treaqmnt comparisva4 versus placebo were b Uaendj m an ANCeVA vmodl th :'rment (excludmg lanzapin e) and pac ed center at fixd fcrton .and Baeine PANSS negative S-ampaton Su*bZa e a *-a cvr Note: miifiscan a; Week; 6/dc p c pass&cswt cnu-Zed aordin to ts Hobber good®. Nase: PsVah a.d CL psud V.eek ]r- Week A nd WkEadpoir fa dsap:e.po on'y Trt = Teuet Data Souce: Tabe 3.L.0. Table 3.2.0. 5 CLINICAL STUDY FOUR: The mean PANSS negative subscale scores (FAS, LOCF) decreased in both BX groups until Week 6, after which the scores stabilized WO 2008/025781 PCT/EP2007/058958 44 (Fig.2; Panel 35). In the PBO group, the scores decreased until Week 4, after which the scores tended to increase (Panel 35). In the per-visit LOCF analysis (FAS, ANCOVA) of the mean PANSS negative subscale scores, treatment with BX20 or with BX30 was statistically significantly superior 5 to that with PBO at all time points, except at Weeks 2 and 4 for BX30. PANSS DATA FOR CLINICAL STUDY FOUR The mean PANSS total scores (FAS, LOCF) decreased over time in both BX groups until Week 9, after which the scores stabilized (Panel 28). In the PBO group, the 10 mean PANSS total scores decreased until Week 2, after which the scores increased steadily (Panel 28). The LOCF analysis of the PANSS total scores (Panel 28) shows that when the last observation was carried forward, the PANSS total scores in the BX groups remained stable. The changes in both BX groups were statistically significantly superior to that in 15 the PBO group from Week 6 onwards. In addition, BX20 was statistically significantly superior to PBO Week 4 onwards. The OC analysis of the PANSS total scores shows that patients who continued in the study improved over time (Panel 29). The changes in both BX groups were statistically significantly superior to that in the PBO group at a number of time points including Week 6, but excluding Month 6. 20 The differences between the LOCF and POCF analyses indicate that a number of patients who withdrew had a deterioration in their PANSS total score and withdrew at the first visit where a deterioration was seen. This is in line with the study design that required patients to be withdrawn from the study if they were at least minimally worse. For the other efficacy variables, the trends in the development over time in 25 efficacy scores were similar, except for CDSS, where the baseline levels indicate that the patients included in the study were not depressed.

WO 2008/025781 PCT/EP2007/058958 45 Panel 27 Efficacy Scale Scores - Baseline and Adjusted Mean Change from Baseline (FAS, LOC) PBO BX2 :BX30 (i = 166) (n = 168) (i = 172) PANSS tota. Basene 5.S (L.3) 87 2 ( ) S6 9 (0.4) Wk 6 1 1 (. 2) -4.0 (1. 2) * 2.6n- (1 ) * Mpot t6 .1, (. A) -0.8 (1,5,,,* -0.A (14)* PANSS pMAtIr subscale Baselne IRO (4.3) iSA (3.9) 2 2) Wek Y05 (0.4 -0P (CA) * -C.S (GA)k MenthS 6 216 (.5 C (0.) *O .3 (0. PANSS negadvie subtee Baele 247 43) 25.4 (Aj 211 (42) L'b.(3 -2.0( 10.3) vahlec are menu (SD); Week 6 and Monh es are LS ean( d eslcr (SE) fisn E 1he ANCO'vA model wh cenre and tretiei m factor snd the baseline vahe ascavnate. * stansticy sginficanty ddfere tomin PBO aordin t Hci:hbexgtep-up Mehd pr .025 for the indiial BX poup or px.5 for both BX yrcps WO 2008/025781 PCT/EP2007/058958 46 Panel 27 Efficacy Scale Scores - Baseline aud Adjusted Mean Change fmm Baseline (FAS, LOCF) PBO BX20 BX30 (n = 166) (n = 15) (n = 172) Month 6 0. O) - 61(0.4) * -L 7(bA) PANSS general psychopathology Basekne 422 (62 43.2 (6 42.6(5.) Wek 6 Lo (o.) -1L (a') * - & ( 6) Month 6 42 (7 O. (0?) * as (.) BPRS totl Bas e 46.85 w(6) 47. (56) 5Wek 6 (21 (03) * ~13(7) MOnth 6 4 5 (ig 01 (a ) * .5 (.)* BPRS pycho-sis cluster Baseme 12.4 (2-9) 2.1 (2T) 1.2 (21) Wek6 o' 1 (0o.) -0.8 (0-2) * 07 (02) Month 6 L1(1.3) -A.1(0,3) *.0.2 (w.)* CDSS totaI Baseline 22(29 12(27) 25 (35) Week 6 -0 (up -02(.2) -01 ( Month 6 uA (O2) 0(01) 01 (.) CGI-I Week6 4.1 (0.) 3. (0.1 3.7 (0.1) * Mtnth .a 4.6 (1) 4.0(.) * 4 (a t) CGI-S Baselhne 4A4 (0.6) 4.5 (fY)6 4.'(0 Wek t0.( -. 01) -1 (a 1) Month6 0 L (01) -1 (01)* .0(Q.1 * B2sehne valec are Imn (SD); Week 6 and Monrh 6 vams are LS mear error (SE)j fm lse ANCOVA icdel iw centre and treatment factors and the baseline valu a cavariae. * statitcaly significantlydfeent from PEO accdming to H clber' Mep-up Meth; p<0.0 for the Individual BX g" orp fo both BX goWps a Adjusted mean' vanes instre of adutean ch~age frcm baseine WO 2008/025781 PCT/EP2007/058958 47 Panel 31 Adjusted Mean Change from Baseline in PANSS Total Score (FAS, LOCF, ANCOVA) Least:Squras DifTerence to PBO Est I mtes Treatment Days n Mean SE Meaf SE L5% EE ap-va Ius Group poe "plur P20 7 106 -0Q59 0.49 14 16 - 65 0;,!,0 29 166 0.07 1. O 42 166 1.(6 1.17 63 106 2 25 1.29 90 166 3.6 1.3 120 166 5.30 1 150 16a 6.48 1.41 180 16j 7,1 1.42 BX20 7 15 -1 !8: 0.51 -130 0.67 -262 0 02 0, 053 14 1S -3.03 0.22 -2.38 1.09 -4,51 -0.24 0.029 22 15S -3.27 1.03 -33.4 1.36 -6. 1 -0.67 0.014 42 158 -4,01 1.2 -5, 07 1S9 -0 19 -1 95 0. 002 63 1.8 -3.22 1_31 -5.47 1.74 -9.89 -2.05 00:2 si 168 -2.1 1_39 -5.95 1 e4 -9.47 -2.23 0. 002 120 1S8 -1.2 1.41 -6.K 1.87 -1 . 18 -2.82 0. oI 1%0 159 -0 5o 1.44 -7.04 1.91 -10.70 -3.29 . 0.0 190 158 -0 21 146 -7.91 1.93 -11 .70 -4.12 .000 BX3 7 172 -0 73 0.42 -0.14 0.66 -1 .4 1.15 h .35 14 172 -1 45 0.77 -0.79 1.C6 -2.9 1. 2q .455 29 172 -2 13 0.97 - 2.20 1.33 -4.92 0. 41 .. 09e 42 172 - 2.6 11 -3.71 155 - 76 -0-66 0 17 63 172 -2.44 12 4.6A 1.70 -9.03 -1.34 C.0 0 6 90 172 -1.47 1.32 -5.14 1. 80 -9.6A -1.60 C.5 120 172 -0 84 1. 34 -6 14 1 3 -9 73 -2z54 0 c0 150 172 .31 1. - 1 27 -10 46 -313 0c 000 120 172 -0.44 13 -7.54 88; -11 24 -3 -P4 0 00 10214 ET PTS11 W9JUN206:15:23:07 2613/313 - TFL/SAD Bui icId Numbers - Finai WO 2008/025781 PCT/EP2007/058958 48 Panel 32 Adjusted Mean Change from Baseline in PANSS Total Score (FASt OC, ANCOVA) Least. Sqtmrs DIftoren to PRO Estimates Treatment 95ner E oe UIpper Group Days n Mpan SE Mean SE r p-value Pal 7 166 -0.59 0 4 14 1493 -2. 0 .77 28 136 3.52 1. 96 42 T12 -6.26 1.11 63 92 -S 3.3 1 S5 90 78 -9I26 160 120 62 -10.36 1: 2 150 52 -14.39 2.C2 18 44 -14.49 1 L:87 BX20 7 15 -129 t51 - 1 .30 0.67 -2 62 0.02 0,053 14 138 -4.77 f.8f -2.39 102 -4.41 -0. 3 0.020 29 121 - 7.61 103 -4.09 1.26 -6.57 -1. 62 0.001 42 104 -12.40 1 17 -6.14 1.49 -9.07 -3.21 0.000 63 96 -12.25 1:. 7 -45 17 -203 -1 00 0. 12 90 21 -14.17 1.65 - 492 212 -4.09 -. 75 0. 021 120 69 -15.91 1 .4 -5A5 2A2 -10.23 -O 66 .026 150 63 -16 3:6 2.C1 -197 267 - 7. 26 3.31 . 462 120 59 -18.A 9 1.21 -3.60 2 46 -R.47 1.27 C.146 BX30 7 172 -0;73 t4a -0. 14 0.66 -1 49 1 0, 35 14 137 -4 43 0f 0 -2.05 1 .2 -4 06 -0.04 c 046 29 120 -7.14 1C1 -3.63 1.26 -6.10 -1-1 0 4 42 106 -11.23 1 12 -4.97 1.48 -7.&8 -2. O 0.001 63 98 -12.45 1.33 -4.12 1.77 -760 -0. 64 0.021 9 88 -12.71 1 53 -3.46 2.06 -7,53 0.61 0.096 120 73 -14.5 6 1.72 -4 19 2 35 -2.83 0.44 o.076 1SO 64 -15.93 1.93 -1.54 2 61 -6.70 3.62 0.5IS6 190 6-7 -1&.5 1. - 4. 09 2.45 -9.s 0.76 0.097 10214 ET PTSIO O9JUN:2006.:15: 22:47 2613/313 - TFL/SAD ul Id Numbers - F:i nal WO 2008/025781 PCT/EP2007/058958 49 Panel 33 Adjusted Mean Change from Baseline in PANSS Total Score (FAS., POCF, AINCOVA) Least SqtMra5,3 DItfrene to PO Treatment Days n Mean SE Mean SE p-% upper Group L-vr Upper P2C 7 154 -1 47 0 39 14 154 -2.97 0.59 29 -154 -4-11 D74 42 154 -5.07 O.84 C3 54 -5.2D 0.9J 90 154 -5.24 0.93 120 154 -S.9 0.913 150 154 -5 .20 0.9 191 154 -5.41 1.01 BX20 7 142 -2 1 D 41 -133 C. 53 -2A7 -0. 30 C.011 14 142 -5.4S :.62 -2.61 0.90 -4.19 -1.04 .0001 29 142 -774 r 7 -3.63 1 .00 -5. 6 -1.66 0.000 42 142 -9.40 0,p9 -4.32 1.13 -6.55 -2.ID 0.000 63 142 -1Q.W3 D-95 -4.83 1.22 -7.23 -2.43 0. 000 90 142 -10. 43 9q -5. 19 1.26 -7.67 -2.71 0. t 120 142 -10.30 1 C1 -5 21 1 D -7 77 -2 -S 0)00 150 142 -1 0.40 1 03 -5.20 1,33 -7.82 -2.59 C 0 19) 142 -10.4 1.0; -5.43 1?6 -R.11 -2;76 0 &G BXbC 7 148 -2.23 G.3R -).71 0.52 -178 0.26 0.146 14 14 -4.5 0 S :.6 -1.7 0.79 -3 34 -0.22 0.025 29 14 -6 D 75 -2. 74 0.99 -4.69 -0. 79 C.00 42 14 -8. 12 :.5 -3, '5 1.12 -5.26 -0. A L007 63 14e -8-47 0 91 -3.77 1 21 -6 14 -1. 9 0. 02 90 148 -9.25 0 94 -.4.00 1 .25 -3.46 -1 .55 0.01X 120 148 -9. A D 9a -4.59 1.29 -7.12 -2.C5 D.000 150 14S -9.55 I . -4.35 1.32 -6.94 -1.76 0.001 ID) 14 -9.92 1.02 -452 1.35 -716 -1 A7 001 10214 ET PTS12 09AI2006:15:23:42 2613/313 - TFL/SAD &iPild Nuabcrs - Final General Psychopathology Score 5 CLINICAL STUDY ONE: General psychopathology scores decreased from Baseline to Week 6 for each of the bifeprunox treatment groups as shown in theTable below. The estimate of the treatment effect of bifeprunox using LOCF was -2.2, 0.4, and -2.8 for the bifeprunox 5 mg, 10 mg, and 20 mg groups respectively. The bifeprunox 20 mg group had a statistically significantly greater decrease than placebo at Week 2 10 (unadjusted p=0.029), Week 3 (unadjusted p=0.032) and Endpoint (LOCF) (unadjusted p=0.016).

WO 2008/025781 PCT/EP2007/058958 50 Change from Baseline ni PANSS General Psychopatholkgy Subscale Score Last Observation Caried Forvard at Each Visit Intent.-to-Treast Populatiin T1 earment Group Bifeprnnox Bifeprunmox Bifrprnnox Rispnpridne Stadak n mg 10 mg 20 mg Placebo 6 mg Total Nerte of Subjects in the ITT N 1 111 1i4 I 16 Baeline N I! 1 I l11411 ______Me___ (S.D.)3~ 44.1 (6t 55( ) 4 (6.9) 44.7 (63)Q 44.C (6.5) Median 44 46 4 6 45 4 Mw- Max ;1-62 2 6 -1-6 31-6(f 30- 5 Chaage Mam Baseine Wk1 N I 1 116 Mlem 3.D. -47(3 09(.) -. 6 ) -. 55 3 (5 3) Medan -2 9 -2 -4 -3I Week 2 n110 1181 1 1 Mea D.) -. ( S ) - 7 -4.4 7.5) - .73 -. (7.1) Median Mmn-Ma -9-49 -- 23 -7- -2 -6 -30t- 7 - . 4: Week3 0 181 1'46 M (,-,( D.) -3-3 (88) -, L7 -90 -5-2 (7.) - .5 CS& 1) -6-9(73 Median - 3 - -6 -6 Min - Max -29 -- 49 -28 - 23 -9-16 -3-22 -27 - P-value 0-505 Cs-36,'4 C'-032* Week 4 10 1 1 1 M (.D.) -3 (9. -A(S.) .5.2(" -2..3(S. .6.3(7.

Median -4 1 -6 -3 -5 i - Max 1 49 - - -27 1 P-,alue 0). 543 0.13 IS.M62 EndoILC 0 11 14 Min - Max -31 -49 - - 32 - 16 6 Ti Efect -2.2 -. P-valn 1.5 -380 .06 * Significant r th 00 eve NdTe: Pairvse rcmpasncs p bL are based oi an mANCOVA modet wit ieimai 4excuing cspejidenel and pco7d ,aes as fixed facts and baiseme PANS, gmera pfe hopakhology ubrcale - 'wse e a vmnat Nete No PANSS semes were recoded at Week 1 at Week 2 for &ject 11259 in the bifepnmax 20mg grop Data Source: Tabl 3.3.'. For the observed values analysis of change from Baseline in PANSS general psychopathology subscale score, decreases in the scores over time indicated 5 improvement in each of the bifeprunox treatment groups. A statistically significant treatment effect of bifeprunox versus placebo was observed at Week 2 (p=0.038.) for the bifeprunox 20 mg group versus placebo and at Week 6 (p=0.017) for the bifeprunox 5 mg group. No other significant differences were noted.

WO 2008/025781 PCT/EP2007/058958 51 CLINICAL STUDY TWO: Mean PANSS general psychopathology subscale scores at Baseline and change from Baseline in PANSS general psychopathology subscale at each post-Baseline visit for the ITT population (LOCF) are presented in the Table below. The mean change (SD) from Baseline to Endpoint in PANSS general 5 psychopathology subscale score was -6.0 (10.2) for the 30 mg bifeprunox group, -3.8 (10.1) for the 40 mg bifeprunox group, -3.5 (9.7) for the placebo group, and -8.6 (9.8) for the risperidone group. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were: -2.5 for the 30 mg bifeprunox group and -0.7 for the 40 mg bifeprunox group. Notable differences were observed between the 30 mg bifeprunox group and 10 placebo from Week 2 through Endpoint (p 0.025). There were no notable differences between the 40 mg bifeprunox group and placebo at any timepoint during the study (Week 1 through Endpoint).

WO 2008/025781 PCT/EP2007/058958 52 Change from Baseline in PANSS General Psychopathology Subscale Score Last Observation Carried Fonvard at Each Visit Intent-to-Treat Populition Tea~tmnt Group Bfermn"ax Biferpnmx Rispeidone Statistic 3 m1 40 Mg pLacebo . ng Tota Number of S jects iu the ITTPIputio- N 140 144 Basehne E 140 14- 144 1 Meaa (SD) 459 (5 7 44.8 64 45_5 (6.) 44.9 (6.2) Mledian11 46 44 464 Nm, Max 29, 67 , 31, 60 Week 1140 139 143 149 Measa (SD) 47 -2.,64) -3.9 (5.6) Median-2 -2 -4 -IMa -23 2I -2 1..1 1 ' -19, 14 P-Valu 0. 51 0.501U Wek 2 Y 140 140 144 11 Mesn (SD) -4,9 (&S 1D: -3._4 (.) -. S ) -. 7 Me~dian -4 -3-6 Mi -Max :'23, 3< -23, 2-2 25 -21 , 13 z4f Wek 3 140 1 1 Mes ~ _ (SD -58(3 38(2 33 (9.A) -7.6 (7,7) MedinD -- 3 -, Min M3ax -I2 -24, 13 P-mnke C.15 .7 Week 4 n 14' 40 14415 Mean(SD) -N5 3 -3.4(.3 - 4 -'(5 M--2 -4 -9 Nm, Max -,-2 -3, 2 3 13 P-L) 0 07 ** a.632 Week 62Endpta L 40 14 Mem:' ) K (I2 -3.1. ) -3.5(907 -&.6(3 MedIan -- -4 TrdE-ect -25-0 5%C7 -4., -. 3 -2..9. L5 P-valu*'e 0,10215 * r". 545 * Notaby dfferent ficem placebo, p'-:50; Notaby differ5et fivmeplc p'u Nate: Trea:nes. comparns ver-us p we're bas tn an ANCOVA node uh teatment ( exPlndmlg Rpeidcoe) and pooled center as fied factors and Base 'sm e PANSS general p.y1patholoy sub cale score i:a cvanate. 3t = TfaientL Data Source: LTab 3. CLINICAL STUDY THREE: Mean PANSS General Psychopathology subscale scores at Baseline and change from Baseline in PANSS General Psychopathology subscale scores at each post-Baseline visit for the ITT population (LOCF) are presented 5 in the table below. The mean change (SD) from Baseline to Endpoint in PANSS General Psychopathology subscale score was -6.1 (10.7) for the 20 mg bifeprunox group, -5.8 (10.3) for the 30 mg bifeprunox group, -4.8 (10.0) for the placebo group, and -10.5 (9.4) for the olanzapine group. The treatment effect values (bifeprunox - placebo) WO 2008/025781 PCT/EP2007/058958 53 at Endpoint LOCF were: -1.5 for the 20 mg bifeprunox group and -0.9 for the 30 mg bifeprunox group. No notable differences were observed between the placebo group and either of the two bifeprunox dose groups at Endpoint or at any other time point during the study (Week 1 through Week 4). Change from Baseline in PANSS General Psychopathology Subscale Score Last Observation Carried Forvard at Each Visit Intent-to-Treat Population T reament Group Bifeprunox Bifeprunox Oanapine Statistic 20 mg 30 Ing Placdbo 15 mg Total Nuiber of Subjects in the IT Populanion N 149 148 145 146 BaseLne n 149 148 145 146 Mtan (S) 4 6.-3 6. ) 47. (. 465 (7.1) 47.6 (6 9) Meca47 48 46 48 Min- Max ',L 61 30, 68 27- 64 3L 65 Change kfom Baseline W ik I 146 144 145 Mean (SD -24 (6.7) -2.8 ( (18) -4.5 (6- 2) Mkchma-n3 2 Mhin Mx -2 2 -35 -34 5 P-value 0.405 G.607 eek2 n49 147 145 146 Me (SD) -4. 4 - 2) -4.7 (7.6) -78 Nedlian4 -5 -4 Max Ma_321 '28 -2.32 -28, 11 -30- V2 P-value 0.961 a5 49 148 145 146 Mean (SD) 6 (9 ) -0(9.3) -54 (8t -9.3() Median -6 -4 MmL. Max -37- 28 -30. 3 -31, 15 -31, 12 P-vahu 1). 78654 Wek 4 n i9 148 145 146 Mean (SD) -5 (. 4 (9.5) -2 (9 -9.6 (8.6) Mechn --- 4 -9 Min- MaX -928 -30, 32 -4. 15 -38, 12 P-vale 0.47 1 0.981 Week 6/Endncln n 149 142 145 146 Mean (SD) -6 1 (1 7) -58 (510.3 -4.S (l0) -10.5 (9.4, Media -7 -6 -3 -9 MeN. Max -49, 28 -30tj3 -33, 21 -39 12 Trt Effee -L Tr Effet -3. 7 -3.1. 14 95%t Cl P-va .18S0 188446 Note: Treament compaisons veLu. pLcwere bas d n ANCO'VA model wit treatment (excluding oLnzapine) and Pooled center as fixed facrs and Baseline PAN'S generm psychopathokogy ubscale score as a coxanate. Trt = Treatmennit Data Souce: Table 3.30 5 WO 2008/025781 PCT/EP2007/058958 54 CLINICAL STUDY FOUR: The mean PANSS general psychopathology subscale scores generally followed the same pattern as the PANSS total scores. The mean PANSS general psychopathology subscale scores (FAS, LOCF) decreased in both BX groups until Week 6, after which the scores stabilized (Fig.3; Panel 36). In the 5 PBO group, the mean PANSS general psychopathology subscale scores decreased until Week 2, after which the scores increased steadily (Panel 36). In the per-visit LOCF analysis (FAS, ANCOVA) of the PANSS general psychopathology subscale scores, both BX20 and BX30 were statistically significantly superior to PBO from Week 9 onwards. In addition, BX20 was statistically significantly 10 superior to PBO at Week 6. BPRS Total Score CLINICAL STUDY ONE: As shown in the Table below, the bifeprunox 20 mg group showed greater change in BPRS total score than the placebo group. Statistically significant treatment group differences were seen for bifeprunox 20 mg versus placebo 15 at every time point beginning at Week 2 (unadjusted p=0.042); Week 3 (unadjusted p=0.020); Week 4 (unadjusted p=0.024); Endpoint (LOCF) (unadjusted p=0.012). Change from Baseline in BPRS Total Scowe Last Observation Carried Forward at Ead Visit hitent-to-Treat Population Treatsment Group Bifepranox Bifeprunex Bifeprunx Risperidone Smg I LMg 20 Min Placebo 6 mg Totai Numer zf Subject he ITT N 10 Ie 11 114 116 Bafln 113 1, SD) 5 (.3 6 5U 's 542( 1' 5_2.(.2 ___________Median '5 54 Mm__ _ __ _ Lr 4r4 41-67 3' 9-4 41 40 -4 36 CUan'ge ft1;1n Bnase v" 'eel- I1n 1 " a ' _______ ___ \ei (S. Dy . 5)~L -2C(D -4. 2-47 3 -4 (.3 WO 2008/025781 PCT/EP2007/058958 55 MIedian _3 3-3 MiEn - a-23 44 2 - 17i P-vahse( (.999 . Week 2 W 11 1 14 16 Mes (S.D.) -4.2 3 ) -6.1 ( ) (8) - ( Median -4 37 Mn- Ma -29 -44 -29 - 24 6 - 6 -2- 2 6 - 10 P-value 0.649 0A93 (. 4* WeEk 3 n 0 1 1 14 16 Mea~n (.D.) -5. 2 (9 7 -3 1. 9) -,7.5 (91 . 9.7) -9. ( O 4) Min - Mlax -32 - 44 -4 -2 -7 - 19 -3 - 2 -2-i P-vahse 0. 50 2 0.67.2* WeV'ek 4 t 10 116 Mean (S.D 1 - 9) -45.6 dian - 4 -7 .Min wa - 44 -34 - 19 P ale0.5 027 (.024* E£iscanLOCF n n1 1 11 11 116 Men SD 6. 2 (e.a 5) -3.4 (1 L5) -7: f) .1) -;.8 (0.3 -9.2 (1 9 Median - 2-S- -9 M-Max 4 -44 -3449 -31- 10 Tr ifet 5 Tr( Effet C. ) -6.2, -S) P-vthie S.8f '.66 A..12 Significar at the .'0 leve. Note: Pa7itwEa campflisonS versus. pa ne based cm anANCOVA o' m e' with W-ratrieit (exchdng nsei done) aml pcued center n fixed f rs d baEline BPRS totl sore as a caiate. Note: No PANSS scores were record at Wete 1 or Week 2 Subjet 11 9 in the bifiprusaox 2a s up. Data %cnrze: Table 3.4.') Similar trends observed in the LOCF analysis were noted in the change from Baseline in BPRS total score analysis using observed values. The comparison of bifeprunox treatment effect to placebo was statistically significant at Week 6 (p=0.024) 5 for the 5 mg group only. No other significant comparisons were noted. CLINICAL STUDY TWO: The table below presents mean BPRS total scores (derived from the PANSS) at Baseline and change from Baseline in BPRS total scores at each post-Baseline visit for the ITT population (LOCF). The mean change (SD) from Baseline to Endpoint in BPRS total score was -8.1 (12.3) for the 30 mg bifeprunox 10 group, -6.5 (11.7) for the 40 mg bifeprunox group, -4.9 (11.5) for the placebo group, and -12.2 (11.7) for the risperidone group. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were: -3.2 for the 30 mg bifeprunox group and -1.9 for the 40 mg bifeprunox group. Notable differences were observed between the 30 mg bifeprunox group and placebo from Week 2 through Endpoint (p 0.019). No notable differences 15 were seen between the 40 mg bifeprunox group and placebo at any time point during the study (Week 1 through Endpoint).

WO 2008/025781 PCT/EP2007/058958 56 Change from Baselne in BPRS Total Score Last Observation Canied Fonard at Each Visit Intent-to-Treai Population Tietxm1-2uiCcr Bifepnnox BifCpm!ex e Stthizi 0m 0m imo6m Timai Nimnber of SubjecsC in se ITT Pdatim N 140 14 144 151 BaseLmZ 140 141 144 11 Men ('D' 54. ( ) 541 (6. 9) 5A(.) 5 71 Camae nasa Bnselir M emin (SD) -3.8 70) -. 71 350.) -. 63 Min. Maix -2 6, 4S -27 23 -33. 2 2 -31 Week 2 4140 14 144 151 Mean (SD) C7. ,'9.5)' -5 (93 39 (9.7) -. & in- -4 -9 N-iin. M'.ax 28, 21 -36 18 -31 32 25 P-vhie 0:004* 0.74 1,40 140 144 Meia -6 - 6 -- 1 1 Min. M ax -36, '21 -7 19 -4L 1.-5.1 Men SD- - 2 ('5) -7 j(I,', - 4.6 . L4,) -1 10. ) 7-13 Mia, M"'AX -721 - 9 -4- 32 -35. 1 Week 6 0 141 144 151 M Sa SD) -1 -2.3) -6. 1 -4 9 (1.5) - 12.(1 M edin - - 1 Min. M'ax -41, 1 -39;. 2.2 -42'1 3 2-3. n., Bfect -13-L 9%C1 58 & -. 0 19 *Ntabl y Liffern i mE oE1 , placeo, p, 05 ** *: Notbl d.en &on plate. p 01A. NLote: Treatmnt Cmaurism'ons veCLheb ee ae -ACV moadel w:vith tset(x dn risperdae, mnd es t ad Eed factOrs and Baseh',e BPR totl 5.Core ',S R COVan!.Ate. Data 4erm Tab1e341 CLINICAL STUDY THREE: The Table below presents mean BPRS total scores (derived from the PANSS) at Baseline and change from Baseline in BPRS total scores at each post-Baseline visit for the ITT population (LOCF). The mean change (SD) from 5 Baseline to Endpoint in BPRS total score was -8.5 (11.9) for the 20 mg bifeprunox group, -7.9 (12.0) for the 30 mg bifeprunox group, -6.7 (11.7) for the placebo group, and -13.2 (10.7) for the olanzapine group. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were: -2.1 for the 20 mg bifeprunox group and -1. 1 for the 30 mg bifeprunox group. No notable differences were observed between the placebo group and WO 2008/025781 PCT/EP2007/058958 57 either of the two bifeprunox dose groups at Endpoint or at any other time point during the study (Week 1 through Week 4). Change from Baseline in BPRS Total Score Last Observation Carried Forward at Each Visit Intent-to-Treat Population Treannent Group Biferrunrx Bifepmunox Ohlnzapme Ste 20 mg 30 mg Placebo 15 mg ToNtaInaber of Subjects i the ITT Ppuli on N 149 148 145 146 Baslne n 149 148 145 146 Man (SD544(6) 54 6.7 , 55.7 55.6 (6.9) 55 56 5> r6 Min Max 37(57 36, 74 40, 73 Change frni BtLine Week 1 n 149 146 144 145 Mean (SD -31 (7.3 -3.6 (7.3) -4.(7. -5.6 (6.7) Ma3n -3-4 -4 Min. Max -29. 2 2 6 -42 12 -37 P-value. 1) 035 0.f14 7 7 Week2 n 149 147 145 146 Mean (SD) -6.

0 (9 1) -3 9 3 -65(2 -9 (2) Mthm]n -7 -6 -5 -8' MiSl Max -37- 28 -33 35 -3 16 -37. 10 P-value 1 0.219 Week3 149 148 145 145 a (SD) -5 (103) -6.7 (107) -7.5 (10.4) -114 (9.3 Maha~n -8 -i Mi Max -41, 2 -35, 522 -39, 10 P-ahe0.896 7 0.446 Week 4 n 149 14S 145 146 Mean (SD) -8.1 (11-3, - 103) (1M) -120 (9.) Mhn'x -9 -S -6 -11 Mil. Max -42.- 28 -36. 35 -35, 22 -38. 10 P-value 0.3160.7 Week 6/Endposn n1 149 14S8 145 146 Mean (SD) -5 19' - 1 -6.7 (11.7) -112 (107) Mahasn -8- -5 1 Mil Max -42 30 -43 35 -40, 22 -42. 10 Tit Effect -- 1 -1.1 Tit Effect -4.7, 0.5 -3.7. 1 5 95% CI P-value .115 10.411 Note Trentment compads versus placebo -wre basd on an ANCOVA model wih treatnitait (excluinfig oktazapine) and poied centr as used faacor and Baseline:BPRS iotal scnre as a covariate. Trt = Treatmen.m Data Source: Table 34 3. BPRS Psychosis Cluster Score 5 CLINICAL STUDY ONE: The Table below presents mean BPRS psychosis cluster score at Baseline and the mean change from Baseline by visit using LOCF for the ITT population. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were -0.9, 0.4, -1.1 for the Bifeprunox 5 mg, 10 mg, and 20 mg groups respectively. A statistically significantly greater decrease in the difference between bifeprunox and WO 2008/025781 PCT/EP2007/058958 58 placebo mean change from Baseline (unadjusted p=0.044) at Endpoint (LOCF) was seen in BPRS psychosis cluster score for the bifeprunox 20 mg group versus placebo. Change from Baseline in BPRS Psychads Clusfte Scre Last Observation Carried Forward at Each Visit hItent-to-Treat Population Treatment Grrn Bifeprunox Bifeprunox Bifeprnnox Riperidone 5 mMg10 mg Ig Placebo 6 mg TC-1al Numbt .er o-f Sujett ete ITT N 10 118 1111 P4 116 Mearn (:S D. 80 (1.i6 1 A (2.7') 1O A (2.5) A(2 a 9) 1.0 (2.6 Mediau 18 88 17 MnMx -2613 3-7 42 -12-7 126 Change fwm'iBaseie \We 1a !010 10 1111 MeLn(D.2. - -0.7 ( ) -L5( -2. 2 (.7) Median - 1 -1 1 Min- Mr -9-13 -10-3 -9-4 -1 - 4 P-value 0.73 C19 0110 \Wek3 ,0 1,& 0 1141 Mean( OD.) -2 (3. - ( -2.3( -. 6(3.4) - (2) Medi2 - - - 4 M- M x -1 -34- B-12- -1-41 P- ue .76- 0204 0.1 Week3 1 1 1 i 14 116 I' 141 MiesL D.) -2(3O) -t t (4.2) -2.7 (3) -L7 (3-9) -4.1 ( 4) Mein -2 -1 -- -4 Ma Max - -14 - -11 -9 P-vllue 0.449 C0,151 0.0589 Week 4 a I1s 1'8 1114 t 16 Men S D.) -2. 4 (3 R) -' 4 ) -. (l L8(.9 3 33 Mk< Ma -4 1 -I- -10- - -43 P-Value 0.2 . H.0 Enaoit LOCF i 1 181111 1 MeS an (S D. -2.5 ( ) -1. 2 (4.) -2.7 (19) -L 6 (41): -4., I (15) Median -3C, -3 -e'-4 M ua - a 14 -1IS -11 2 -11 -9 -11- -e 1 Tnf Eec -0 9 CA L Tn ,E! c vC1 (-2. 1. . 2.i (-.,(5 -2.2., -0.t) * Si 2nificant at th e 'D C50 evet Note: Fai.ise compaSsos ve w) S.Cebo are ba'ed .n ANCOVA mode 'ith is9eaunetxchidius spide d cener as e ft ad b E p da re a aCmiate Note No PANSS s-es -wete recorded at Week 1 r 'ek 2 for Subjeci 11259 ri t.htebpsx 2rg rup. Data Snitoce: Table 3.5.!. For the observed values analysis of change from Baseline in BPRS psychosis 5 cluster score, larger decreases in the scores compared to placebo were seen for each of the three bifeprunox treatment groups beginning at Week 4. No statistically significant differences were noted. CLINICAL STUDY TWO: The Table below presents mean BPRS psychosis cluster score at Baseline and the mean change from Baseline by visit using LOCF for 10 the ITT population. The mean change (SD) from Baseline to Endpoint in BPRS WO 2008/025781 PCT/EP2007/058958 59 psychosis cluster score was -3.2 (4.2) for the 30 mg bifeprunox group, -2.6 (4.4) for the 40 mg bifeprunox group, -1.8 (3.5) for the placebo group, and -4.9 (4.0) for the risperidone group. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were: -1.4 for the 30 mg bifeprunox group and -1.0 for the 40 mg bifeprunox group. 5 Notable differences were observed between the 30 mg bifeprunox group and placebo from Week 2 through Endpoint (p 0.002). Notable differences were seen between the 40 mg bifeprunox group and placebo at Endpoint (p=0.031) and at Week 2 and Week 3 (p 0.036). Change from Baseline in BPRlS Psychosis Chster Score Last Observation CarriId Forvard at Each Vit Intent-to-Treat Populion Tretmtw Goe 3ifex Bifeprumox Statistic 3 0 40 Picabe 6 ng te ITT PoN 141 14 11 M-an (.SD) 26A (25 1 5 6.3 (2-4) 16.5 s 4) Medin 616 6 Min Max 2-24 .4 14 C. 23 Chage nYon Basebne: Week 1 a14131419 M-an ('SD) -3(13) -1.? 16) -. (.) -. 24 Mi, ax -9f)-1, 6 -3, 7 -R. 4 F-vak-!e .1 5 . 9 WeEk 2 MeL (ID) -2. 61) -. (35)- \1) -. 41 Me2 2 2-1 Min, mx -1, 6 -12, 5 - .1)-1 -4 " i'ud"1 Pdne 40 01 ** 0. 03 1* Wed: 3 4441 M an (SD) -. (i)-. 40 13(-) -43(A Meia - - 4 -. -4 -4 : , x -3n 6-14 7 - 7 1 4 WEAk 4 1441 Medi1 -3-13- -5, N: is, M x"I -13, 6 -14, 7 -3 1 1144 -hE~ l .0x ** 0.0711 Weak 6/Endpman n 4 4 4 5 Mem(,D) -3-2'42) - 44 ( ) -4 4) Min, ax -13-,15 -5 147-5 T kte -".4 -. T tfB~ect 9P% C1 -2C . I.t - -0.1 P-Value 0.02** z2 *Notabv4iy ~ent frcn p ceba, p0.5 ** Nvaly dtiferen ' mnf parce p< t.1. Note: Trnnen emnpasrs vemn pacee- t wer ae en an ANCVA mc'el wi te atmntAudn rispendone) rd pno ed center as fed tAcTors md Bas eline BPRS ps c te rc a ' r Dr4a Scarce- Toht 3.5- WO 2008/025781 PCT/EP2007/058958 60 CLINICAL STUDY THREE: The Table below presents mean BPRS psychosis cluster score at Baseline and the mean change from Baseline by visit using LOCF for the ITT population. The mean change (SD) from Baseline to Endpoint in BPRS psychosis cluster score was -3.1 (3.9) for the 20 mg 5 bifeprunox group, -3.2 (4.4) for the 30 mg bifeprunox group, -2.6 (4.0) for the placebo group, and -4.9 (4.0) for the olanzapine group. The treatment effect values (bifeprunox placebo) at Endpoint LOCF were: -0.6 for the 20 mg bifeprunox group and -0.5 for the 30 mg bifeprunox group. No notable differences were observed between the placebo group and either of the two bifeprunox dose groups at Endpoint or at any other time point 10 during the study (Week 1 through Week 4).

WO 2008/025781 PCT/EP2007/058958 61 Change from Baseline in BPRS Psychosis Cluster Score Last Observation Carried Forward at Each Visit Intent-to-Treat Population Treatment Group Bfepnnmox Bifpunmox Olanzapme Statistic 20 mg 30 Im Placebo 15 mg Total Nunber of Subjects in the ITT Poplation N 149 14 145 146 Bas~ehne ai 149 148 14f 1465 Me an (SD) 15.1 C) 168 (.3) 16.6c2.4) 16 A2.3) Median 16 6 16 17 Nn.Mx 1, 12. 23 12 23 11 23 Change from Basehn Week 1 149 146 144 145 Mean (SD) -14t.) -1. (2.5) 61 9 (2.3) eda -1 -2 -1 M )Mx -0 4 6 P-whlse 0.871 0.8~24 Wek 2 a 149 147 145 146 Mean (SD) -1 ) 2. 3) 3 -,.4 (1 16 Median -2 MmL, Mx -12, 4 -11s -11, 1 P-value 0.6180.0 Wik 3 149 145 146 Mean (SD) 6 (3. 4) -3-4.0 (2) IMdia -s-3 -3 -4 Min, Max - 14, 5 -12, 8 -14, 7 -1 P-vaith 0.697 0.460 Wck 4 a 1 4 9 14? 145 146 Me an (SD) - -3.0 (4.0) -2. 39) -4.4 (3.5) M, -1Mx -1 -16, -133 P-alue 0.5 fl873 Wek 6Endpoin I 19 148 145 146 Mean (SD) -;-1 (39) 2(44 -2.6 (4.0) -4. 9 (40) Median -3 - -2 Min- Max -14,9 -16 S -16- 7 -1 Tr ffes -0.6 -0.5 Tra Pfect -1.5, 0.3 -1A.0.4 95% C1I Pl .165 -0.245 N Treatment comllpansons: versus placebo based on an ANCOVA imodel with treatment (excludig olnzarspme) and pooled center as fixed factors and Basehne psychosis chster score as a cvrae Trt = Trenment. Data Source: Table 310 CGI-S CLINICAL STUDY ONE: The Table below presents mean CGI Severity of Illness scores at Baseline and the mean change from Baseline by visit using LOCF for 5 the ITT population. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were -0.31, 0.12, -0.19 for the Bifeprunox 5 mg, 10 mg, and 20 mg groups respectively. Pairwise treatment group comparisons showed statistically significantly greater decreases in the bifeprunox 20 mg versus placebo at Week 2 (p=0.008), Week 3 (p=0.020) and Week 4 (p=0.032), but not at Endpoint (LOCF). The treatment group WO 2008/025781 PCT/EP2007/058958 62 comparison for bifeprunox 5 mg vs placebo was significant at Week 3 (p=0.049), Week 4 (p=0.033), and Endpoint (LOCF) (p=0.013). Change from Baseline in CGI Severity of Iflness Score Lat Observation Carried Forward at Each Visit Intent4fo-Tr-eat Population T reatment Gron Biieprun.x Bifeprunox Biftprimui Ris-peridqke Static Ig 10 mg 0 mg Phebo 6 Ig Total Nysaier cf Subj t es in !Le ITT N I IS 111 14 116 Baseline a 110 11s M 4 116 Mea( fS.D-j 4.57 63 4 64(06) 4.t rt66 4.54 a64) 460(Q.68) Median 5. 5. 5 04 ( 4.5 Min -Max 4.C. - 7.0 4. - 6.0 4. 0 - 5.0 4. C - 6 o 4. 0 - 710 Chage: fram Basenne Wek 114 115 Mea (S.D.) -0.IS(0 8s -013(&75) -0.31 (0.74) - 6 A - 0' tean 0. 0.C 0. 0 Ct 0' Mn - Max -2s -- I 3 - -4. 0 - L. - L 0 3 0 - 2.3 P 754 054 fl 172 Week 1 7 t11 14 1 5 M,-ean(kS-D.) -0.3720.83) -0.23( 2) -G. 50 (CR5) -).."(S0t 2) -0.60Q 3 Me .M 0.' 0 .Q P S.Q Ma Ma- 2. -4.0 -2.0 -4.0-10 -0 2. -10 2. WeIk'S r]11 0 '114 115 MAn 6.D.} -04 (0 9 2 -2 C. I -0.(0.90 -t1.24 (0 8M) - 10f0 0's I Median 0.0 0.'0 .. 0.0 Mm, - IMaa:I -1.0 - -0 -4. 0 - 1.0 -3.0 - L.e -4. 0 - 2 C, -2. - L-0 P-value 0.04*' l623 0. 0 * Week 11 117 -'4 11 MLeN fS.D. -. 52 (5 1) A I 099 -0 57 (f89l -J 25 '(&,4 -034(( .85) Mfedia.C . 1 M0 '. in - Masx -. -12 -_4 . -720 -. - L- 2 P-vahlm ft.33 0.2,76 (10 WEdkot LCI 17 11 4 15 Men(SD.) - .5( 0 -018L4 -A 521 (C0.96)C -(.2 C( 7) -. 6f,.93) Mded-ian_ _ 0 . t 0 0 Mm - 4. -2-"0 -4 0 - . -30 - 1. - 0 Ti E5mo -Q31 0.12E-9 Ti -t t C1 I -0-56- -3 07) (0. 0 6) (-c44, A 4. ) Pi e . -> D13* 0 31 0. 1 Sigific ant at Ihe 0.050e Nte: Painvise c-mpsos ar based r e a ANCOVA mc,3 wicrandent iexcutlts neridone) and p.>& center a' fixed" sn andt basehne CG- Severity of Riness seCe as eari iea.e. D-ata Srsre: Tablt 3 & C 5 CLINICAL STUDY TWO: The Table below presents mean CGI-S scores at Baseline and the mean change from Baseline by visit using LOCF for the ITT population (LOCF). The mean change (SD) from Baseline to Endpoint in CGI severity of illness score was -0.69 (1.19) for the 30 mg bifeprunox group, -0.54 (1.12) for the 40 mg bifeprunox group, -0.37 (1.07) for the placebo group, and -1.06 (1.20) for the risperidone 10 group. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were: - WO 2008/025781 PCT/EP2007/058958 63 0.28 for the 30 mg bifeprunox group and -0.18 for the 40 mg bifeprunox group. No statistically significant difference in change from Baseline to Endpoint in CGI-S was seen for the 30 mg bifeprunox treatment group compared to placebo based on the step-down procedure (adjusted p=0.056). The 30 mg bifeprunox group showed a notable difference 5 from placebo at endpoint (nominal p=0.028). Notable differences between 30 mg bifeprunox and placebo treatment groups were also noted at Week 2 through Week 4 (p 0.015). No differences were observed between placebo and the 40 mg bifeprunox dose group at any time point during the study (Week 1 through Week 4). Change fran Baseline in CGI Seveiitv of Ilfness Score Last Observation Carried Forward at Each Visit Intent-to-Treat Population Tarment Grw' IBife;;mx B£pmn:xpeio Sta-S1 30 mI 40 mg Placebo 6 mg iTtal Nunber of Si N1 144 151 m the T Pepriamic'i Baselme N 140 11 144 (SD) 4 456t .. 476? Mire ax 4AKGD, 6.0 4I 4. 4, 6.0 Change fbvn Baselmeh W~e-ek1 N 139 3 14414 MenSD 0 33 ( ?) -2 0 .) -2 29 0. .66)- -,;.39 Q3 Medium 0. 0. Min.M x -3.0 ., . 4., 2.0 P-ahe0.171 f0 249 Week. 2 N 140 140 11445 MedIafR 0 Q. Q 0 M am M.. -IA. 10 N~e 13 "m"SD - 2.(08 -3.54(.02 -&4 LO) -94L ) t liAe 7'V 4 Meed4 10 1-t-4 N- an 4SD C0,4 Mu Mx -41 2. 3A20 -41". 2.0 -4J 2.0 p-Value 0.01 1197 440 14 44 Mean fS) 6() -R 11.18) -0,341.0) -. 2 9 M in. MAax -4.A 2., -3D. 1.0 -A"2.2-A . Week 6/Endpomi. N 140 41415 Ft'l S± S 1 ~ 6 ' %0 53. -1-i Q7 9 7.5 C- -0.10-6, O-i he (Raw) .1 Pake(Ad) 056 Note: Treatme cpo u p1ra n ANCOA& model "r tnament (excl d rispenicme) amd center a fixed fieton ad Bashne Cid n oft±x i o>)e n, a cOnrate. Note: Adjued p-vwie was based an the HOchhezg Jxweds&e for amide comio Nte.:? Psal3eL pesete foI Wek 1 thr Week 4 re & descripti piues only. Tt= Trk Aj Adi d D ucT T-abe 05..0 WO 2008/025781 PCT/EP2007/058958 64 CLINICAL STUDY THREE: The Table below presents mean CGI Severity of Illness scores at Baseline and the mean change from Baseline by visit using LOCF for the ITT population. The mean change (SD) from Baseline to Endpoint in CGI severity of illness score was -0.63 (0.98) for the 20 mg bifeprunox group, -0.62 (1.03) for the 30 mg 5 bifeprunox group, -0.49 (1.06) for the placebo group, and -1.03 (1.00) for the olanzapine group. The treatment effect values (bifeprunox - placebo) at Endpoint LOCF were: -0.13 for the 20 mg bifeprunox group and -0.09 for the 30 mg bifeprunox group. No differences were notable with nominal p-values (50.05) between the placebo group and either of the bifeprunox dose groups at Week 6/Endpoint or at any other time point during the study 10 (Week 1 through Week 4).

WO 2008/025781 PCT/EP2007/058958 65 Change from Baseline in CGI Severity of Illness Score Last Observation Carried Fonvard at Each Visit Intent-to-Treat Population Trneai Group Bifepfmno Bitfpnmox Olmmapine Statitic 20 rag 30' mg Placeb I1 mg Toua Number of Subjects in the ITT Population N 149 14- 145 146 Baseline 1449 145 146 Mean (SD 4 75 (0.69) 4.80 (0 69) 470 (0.64) 4.86 0.71) Me&an 5. 0 5.)1 Min. \Mx 4., 6.0 31 6.0 4-0- 60 4-0, 7.0 Change kromi aemne Week 1 149 147 144 146 Mean (SD' -0.23 (v.63) -0.26 (0 60) -027 (061) -0.40 .65) M an0.0 0 D0.0 0.0 Mill ax -2.0,2.0 -, .0 -2 0 .3AL P-value u 50 Q54 Wk 149 148 14 146 Mean (SD) -. 48 (0.81) -0.41 (0.76) -G.46 (0 83) -0 - (0%6) Median RO U. D 1 -L.0 Mu; Ma2x -0 2.0 -. L -3.0. L-3 ) P--ahue 0995 0.2 -28 Week 3 1 149 148 145 146 Mean (SD) -0 9 (89) -0.50 (Q90) -C-54 (0 89) -90 (0.7) 'Mn Mx -1, 2. - 34 2 -3240 1.0 Week 4 K 149 148 i 1456 Mean (SD) -065 (9 -0. 57 (R S9) -Q54 (0. -09 (0 9) Medin -L -1.0 -L Mi -Max -1a 2.0 -3.0 2 0 -3., 2.0 - 1.0, LO P-v;ahse 0349 10 999 Wek /Edoitf 49 148 145 146 Men.D a3(98) -'o62 (L O3) -0.49 (16) -LG3 (1. G00) Median -L O -1.0 D0.0 -tLo Mm, Mdax -4-CA 2.0 -4.0, 2. 0 -4.Q 20 -40, 1.0 TI Effec -) 13 -009 Trt Effen -0.35 0.9 -03L n 13 95% CI P. 'e .62 ,D-4 4 Nte: Teaet comparson vrsus p w e sed on an ANCOVA model with reatment (excludig olainapinv) and pooled center as fixtd factors aid Baselne CGI severiy of illness scre as a coariate. Note: Sigufiate at ek 5Landcimt for nudt rmnisom was eRvauated accordig to e Hotherg preceduwr Note: P-vale a'd C preiemed f Week I through Week 4, and Week &Endpomt ae fonie 'cripe puposaes on. Trt= Treatment. Data Source: Tab-e 3.0-, Table 3.6. CGI-I CLINICAL STUDY ONE: Improvement was noted in the CGI improvement 5 scores for the 20 mg bifeprunox treatment groups as shown in Table 24. The treatment group difference was statistically significant at Week 1 (p=0.040) and Week 2 (p=0.016) for the bifeprunox 20 mg group versus placebo in the LOCF analysis, but not at Endpoint. No other significant differences were noted.

WO 2008/025781 PCT/EP2007/058958 66 Actual Vahes of CGImprovemnt Score Last Observanian Carried Fnvard at Each Visit Inatent-to-Treat Population Tre.atment Gi oup Bifqprutnx Bifeprunn Biffuprunlex Risperidone Statitic 5 mg 20 Mg Placebo 6 mg T Otal Nmsber Mf Subjects m tLe ITT N 1 118 11 14 116 Achn& Valne We 1 f1 -1 4 115 Mean (5 D 3.76 (.) 8 1 3. 54 (.) . (0.93) 343 (0 97) Mi-Max I.C. ILC-6. 10 -6.0 L -6. Week2 a 1 0 7I 115 MEaz (S.D.) 3.46( G; 30 1 2 3.30(L 34 '.65L18 3S 4 05) Median 3.0 4. 2 0 4.0 3 Min-MJ\ I -6- Lt, -5 1 -61 . - 1 L 6.! F-vale 1- 4 14 0216" W e " l 11 4 115 Mear(ED. A 2L7) 3119(L. 2) 1.1(.9 3. 5(L ) 297(105 Medn 4. 3. I j . .0 0.279 (14 0. S Wek4 v U 17 e1 14 t 2 5 UV~k 4i U4" Mean (SD.)1 .7 .5 3. 636 3 5(L2 9) _K5(39) 3.5L II Median C340 . 3 . 4.C Min -M 1.' -7. LO -. 0 -6.0 1. -60 P -wh v -.I I51 OMl k0.220 7Edpoin; LOCF n 1 1 4 11 Mea: 4- 3. :. Min -M LN -6,0 L -7.0 I.C!-6.0 .LD-6. L -6.(l TnEffec -0.28 .28 -0.21 Tn Effe I (-. 62, 1.t6 (0..6, 62) (-05.5 0.14) P-ahme 0.110 109 0235 SSignifican t Le '0 050 le . Note: Pairwise compaiseans ve&s pare ba:ed ox an ACOVA mode: with treaUnent (excludmg ijpenidaze) Zmdpoole zater as fixed fatcrs cd basehae CGI Ser.venty n are a a carie. Data Soarc: Table 3 .7.. CLINICAL STUDY TWO: The percentages of subjects who reported much or very much improvement from Baseline to Endpoint (LOCF) were: 36% in the 30 mg 5 bifeprunox group, 28% in the 40 mg bifeprunox group, 25% in the placebo group, and 52% in the risperidone group (Table below). At Endpoint, no notable differences were observed between placebo and either of the two bifeprunox dose groups; however, notable differences between the 30 mg bifeprunox group and placebo were seen at Week 2 through Week 4 (p 0.024). No notable differences were seen between the 40 10 mg bifeprunox group and placebo at any time point during the study (Week 1 through Endpoint). Data for Week 2 through Endpoint are summarized in the Table below.

WO 2008/025781 PCT/EP2007/058958 67 CGIn Impromenent Scare Last Observation Caried Fonward at Each Visit Intient-to-Treat Population Treatmera Growe 3 feprecx Bifiprumox Pdspertdne 3ng 40mg Placebo 5mg Total Nunbr Sjin nhe 1T P ul~ 'atn nM, 140 141 144 151 Wek14tn4 140 143; 151 Nor A-'seed, n 0 Ver uch~ Imon I % 4' (3% 5 4) 6(4) 8(% Mimlly Impyroved ' (%463 46 2% No Change( 34 (24 31 -(22 41 (% MmimallylkWoe., (%n1(0 0(% 16(1) 8(% Much n 4 (3%- 1 (11%) 1 %) P ~ ~ ~ t - -l 14 140 143 151 Ni3 42 (%) 445(% No Change 6037 26%) 32 (3%) 34 (4) 19 (1%)I Mmal Waq "") .j7% 9 (6%).'\ 16 %) 7() Very Muc WoeW n(() % 1<% N§ 7Cna.rQ P--aine 012 251, t Weel 4. a 4 140 143 151 Not "s7eon 1%) 0 7 0 MudvCh Inmi n CO 37U(26%) - 28 (20) 7 19*.. 58 (38% Mn1im..ally Impro n 26' 49 3 No Chane n )4 "24%1 6) 2 1 W Ek JEdon~ 140 14 1 143 15 1 Nt Asesed a 0 11 vey M uch Imw'ed : 12 r (9% 13 (9%) 6 "4%) 24 (16%) Much~~~~~~ In7e~ (.QR 39(8) 2 9) 2% 54(6% M>jimal Imroed t0 21)' 3 (2%) 40 (2% 4 2% Noc Change- U % 34 (24%).* 2 8 (2%) 31(2% 2G% Lir'ah Wce, 'e % 16 (11%> 15 0 1 6 vryMich Worse. W%) 1 ( 1 1% 1 (-5%) P-Value Q.C79 OA$2 No&ny di ent fiom placebo, ' * Nothly different t.:om pla pix o 0 Note: Piues were h~eed co a CMH ~t s trai~d by pocded enter. DataC C Sou ce .7. CLINICAL STUDY THREE: The Table below presents frequencies for the seven categories of CGI Improvement ratings by visit using LOCF for the ITT population. Statistical differences among treatment groups regarding the mean CGI Improvement 5 score were evaluated using a CMH test stratified by pooled center applied to the frequencies of the seven CGI Improvement categories with modified ridit scores. The percentages of subjects who reported much or very much improvement combined, (derived by summing the corresponding individual percentages in Table 23) from WO 2008/025781 PCT/EP2007/058958 68 Baseline to Endpoint were: 38% in the 20 mg bifeprunox group, 34% in the 30 mg bifeprunox group, 32% in the placebo group, and 46% in the olanzapine group. Comparison of the 20 mg bifeprunox group versus placebo group at Endpoint was found to be notable (p=0.027). The treatment comparison of the 30 mg bifeprunox group 5 versus placebo group was not notable (p=O.162). No notable differences were observed between the placebo group and either of the two bifeprunox dose groups at any other time point during the study (Week 1 through Week 4). CGI Irmprivement Score Last Observation Carried Forward at Each Visit Intent-to-Treat Population Treatment Group Bifeprnx Bifeprunox Olanzapine 20 mg 30 m Total Number of Subjects m the ITT Popuiation 149 148 145 146 Week 2. n1 149 148 145 146 Not asese ny (%o 0 0 Veiy ImIcd 111)od n %) ( 1 % 4 (3%) 6 (4%) M uch hulproved G2 (2%V 2 2% 340(3)3 2% Minimlfly i n 67 (45%) 551 (7%) 49 (4%) 6(% NO chasig, a(% 218 (19%,) 33(2% 35 ('24%.) '26 ( 1%) Mimially worse, n (%) 13 (9%) 11 7% 1 (2 9 (6% MuchI& worse, n (%) 6(4%111) 1 (10%), D 5 (3%)1-<% Very muc0h wr 0 1 (% P-ale0 334 0.661 Week 4 149 A4 14i 146 Not 0 0 0, Very muc1h improv 0 n ) ( 3 10 r7% 1. ) Muchd Imroe 47 (3 2%") 4.5 (30% 3 S'(2 6% 53 (3 6% ') Mimmaiyi impme n(% 44 (30%) 43 ( 28 (%) 3(9%) Nd chage, nU (%) 2U(7) 2 (7)2 20) 2 1% Mininually worse, In %)16 (11% 12 (SO 2(2% 0(% NM worse, L (%) 9 ( 6(1% 86% 5 % Ve~ rmuh wore 0 ) 0 0 P-value 0. 0 57 036 Week 6/Endpoint 149 148 149 146 Not assess.ed' n (% Oj Ver much impove n %7 (5% 1 () 9(6%) I (12 Much h1proved 49 (3%) 40 (27%) 33 4 Minml im r 42 (28 % 44 (Y%) 2 7 (19%) 43 (29%) No chmig(e, n (% 22 (15% 23 (16%) 31 (21% 21(1% Minimally woarse. n % 19 (1% 12 f78%) 3 1 1%) 1;0 (7%) wj 7 ( 7 ) 17 (11%) 1 ( 5 (3%) IVey much Wors 0 2(% P-vahm0.027* 0..162 * N rabl diftferent &i-m pla cebon ,O 0150.. Note: P-values we be on CMH-I test stfanfied by poled center Data Soubrce 33.0. PANSS Responder Rates 10 CLINICAL STUDY ONE: The PANSS responder rates were higher for each of the bifeprunox treatment groups compared to the placebo group. The PANSS responder rates were 28%, 24%, and 34% for the three dose groups respectively using the 20% WO 2008/025781 PCT/EP2007/058958 69 definition from the study protocol. Responder rates for PANSS using all four definitions are presented in the Table below. FANS$ Responder Rates At Endpoint Last (smvanan Carried Forward Intent-to-Treat Populatimn TreatmenT Grru Bifeprun r Bifeprn o Bifepriun ox Risperidone Stadsde l m lu g 2 g Placbo 6 Ing TotalNumera cf ubjecinthe ITT N 1 18 111 1 16 tDetfintu ') 109 L,4 -3, 12~ 0 14, 2 034 9 ADefr i=o 2 ( I C IS(l PA ESRegnd apo t (L 40C") *inince9nt at the 0.5 <eve ** Sigifcamn he 0.010 level Ncte: A PANSS respn dens d-;efne d -,s a object whse PA-Nsy tota Sexe deaeawed ry the specified p-ercentge A significant difference in the PANSS responder rate using LOCF was seen 5 using the 25% definition for bifeprunox 5 mg versus placebo, and for the 25%, 30%, and 35% definitions for the bifeprunox 20 mg groups versus placebo. CLINICAL STUDY TWO: A PANSS responder was defined as a subject whose PANSS total score decreased by 20% or more from Baseline to Endpoint based on LOCF data. As exploratory analyses, responder rates were also analyzed using a 30%, 10 35%, 40%, and 50% definition of responder. Summaries of the PANSS responder rates at Endpoint (LOCF) are presented in the Table 25 below. The PANSS responder rates were slightly higher for the two bifeprunox treatment groups compared to the placebo group. The PANSS 20% responder rates were: 36% in the 30 mg bifeprunox group, 30% in the 40 mg bifeprunox group, 26% in the placebo 15 group, and 54% in the risperidone group. However, the difference between 30 mg bifeprunox group and placebo was not notable (p=0.052). The PANSS 30% responder rates were: 24% in the 30 mg bifeprunox group, 22% in the 40 mg bifeprunox group, 14% in the placebo group, and 31% in the risperidone group. The difference between the 30 mg bifeprunox group and placebo was notable (p=0.01 9). Notable differences 20 between 30 mg bifeprunox group and placebo were also seen for 35% and 50% responder rates WO 2008/025781 PCT/EP2007/058958 70 PANSS Responder Rates at Endpoint Last Observation Carried Forward Inte-o-Treat Population Treanted- Group Bikpyrmo Bifeprnex Rispe '-r id L3e 30 mg 40 mg Paeb a TZtal Nium er of SunbjiectsN m 1, 1TT 140 "4. 144 11 20% Deaimfio Re ed at W eek 6 YeSi()3 (3% 42 (30%) 32 (26%) 82 (54%) P- h 4.1 (HI0..4 3%Denniton Responder e- 34 (2 4 % 1 (2% 2 % 47 * dkent from pa p9AP,5.. N p- ve wee bed ema the CMIH t strified by poold cente. Note: A AS epae a efmedm aL" Uujc hsePNSsor erae y20 2% defation:, r 3%(34)% denstcn. D &,s Seurm: TaMe 3.9.0. CLINICAL STUDY THREE: A PANSS responder was defined as a subject 5 whose PANSS total score decreased by 20% or more from Baseline to Endpoint. As exploratory analyses, responder rates were also analyzed using a 30%, 35%, 40%, and 50% definition of responder. Summaries of the PANSS responder rates at Endpoint (LOCF) are presented in the Table below. The 20% PANSS responder rates were slightly higher for the two bifeprunox treatment groups compared to the placebo group: 10 42% of the subjects in the 20 mg bifeprunox group, 39% of the subjects in the 30 mg bifeprunox group, 32% of the subjects in the placebo group, and 54% of the subjects in the olanzapine group improved by 20% or more from Baseline to Endpoint in PANSS total score. The difference between the 20 mg bifeprunox group and the placebo group approached being notable (p=0.061). However, when more stringent criteria (30% 15 50%) for definition of responder were used, differences between the 20 mg or 30 mg bifeprunox groups and the placebo group were not notable (all p>0. 118).

WO 2008/025781 PCT/EP2007/058958 71 PANSS Responder Rates at Endpoint Last Observation Carried Forvaird Intent-to-Treat Populatio Treatint Group Blfpnm ox Bf1eprnox Olanzapmie 20 mng 30 rug Placebo 15 mug Total Nuber of Subjects i the 149 148 145 146 ITT Population 20%6 defmition s rate Week 6/Endpoint Yes 62 (42%N) 57 (39%) 46 (32%) 79 (54%) No 87 (8%) 91 (61% ) S9 68%) 67 (46%) P-value 46. 1S5 Note: P vlewee basedc o CM test trifte by- pold nt Note: A PANSS responder wasde s a sujetwhs PANSS scoredcrae by 2:2% frm Baselme to Endpoint Dara Sorcr Table 3.9.0. CLINICAL STUDY FOUR: The proportion of patients with a 25%, 35%, 45%, or 55% reduction in PANSS total score relative to baseline at each visit (PANSS responders) is shown by LOCF (Fig. 4; Panel 42). 5 CGI Responder Rates CLINICAL STUDY ONE: A CGI responder is defined as a subject who is categorized as "very much improved" or "much improved" on the CGI Improvement scale. The CGI responder rates for the bifeprunox 5 mg and 20 mg groups were higher than the responder rates for the placebo group. No statistically significant differences 10 were seen for any of the three bifeprunox dose groups when compared to placebo. CGI Reiponder Rate At Endpoint Last Obsevatiou Cardied Forward Intent-to-Ireat Populatioan Treatment Groupa Biepune0 Bifepnx Eifrprianox Rispridone Stthfie lug Z10 mug Placebo C, Ing TotalNnber of Subiets nhe ITT N 110 1 11 114 116 Population CG-iRepnr at n (%) 6 (33) 2I (S)28(5 25 (2) 41 (35) P-v alue 00 5 0,?21 0. 513 the CID12mrvmnsae N e:P-7values- 31:be d on a Ccr -M ae- enzlChi1-Aq Uae test 'Aatined by ooe cent,-. Data Souce: Table 3.RA CLINICAL STUDY TWO: A CGI responder is defined as a subject who is categorized as "very much improved" or "much improved" on the CGI Improvement scale. Summaries of CGI-I responder rates at Endpoint (LOCF) are presented in the 15 Table below. The CGI-I responder rates were: 36% in the 30 mg bifeprunox group, 28% WO 2008/025781 PCT/EP2007/058958 72 in the 40 mg bifeprunox group, 25% in the placebo group, and 52% in the risperidone group. A notable difference between the 30 mg bifeprunox group and placebo was seen (p=0.039). CGI Responder Rate at Endpoint Last Observation Carried Forward lntent-to-:Treat Populafion Trament rp Bifepm Bifepn UMX Rispecidone 40nmg Placebo 6cv Ttal Nu in f e iM .he ITT 140 141 44 PeOu1-laenIn OG Reponder At Endpcn a 140 141 143 11 Ye n(".5 (3% 4C C8%A 36 (2i&%) 71 (52%) N Q( (%) 17 5%) 73 48%) ,u-eb *NotAbly difkrent fom placezbo p<7Q05&. Note: p3-vlhe1 ere bed on a CMH tesz ftafied bypc--oed enter. N G rderz defined a; a subject whao wa c'atecgnzed as "wv y rachn or rmuch improved" in the CGIi mpvemente D-a 0urce: Table 3 10 5 CLINICAL STUDY THREE: A CGI responder is defined as a subject who is categorized as "very much improved" or "much improved" on the CGI Improvement scale. Summaries of CGI-I responder rates at Endpoint (LOCF) are presented in Table 3.10.0 (LOCF) and Table 26. The CGI-I responder rates were: 38% in the 20 mg bifeprunox group, 34% in the 30 mg bifeprunox group, 32% in the placebo group, and 10 46% in the olanzapine group. No notable differences were observed between the placebo group and either of the two bifeprunox dose groups. CGI Responder Rate at Endpoint Last Observation Carried Forward Intent-to-Treat Population Tretmhlent Group Bifeprunax Bifeprunox Of anzapine 20 mg 30 rug Placebo 15 mg total Number of Subiects in the 149 148 14 146 ITT Population Week &Endpoint, n 149 148 145 146 Yes 56 50 (34%) 46 (32%) 67 (46%) No 9(6% 9 (6S%) 79 (54%) P-vah 0.264 0.674 Note: P vahe. were based en a CMH test stratifid by pooIled center. N-ote: A CGI responder was defimtdas a subjecr w was categorizedas 'very much improved or munch impvrcred in the CCI impovement scAae Data Sourcet: Table 3100.

CDSS

WO 2008/025781 PCT/EP2007/058958 73 CLINICAL STUDY TWO: Calgary Depression Scale for Schizophrenia scores are presented in the Table below. This Table presents mean CDSS scores at Baseline and the mean change from Baseline by visit using LOCF for the ITT population. The mean change (SD) from Baseline to Endpoint in CDSS score (LOCF) was -0.66 (3.64) 5 for the 30 mg bifeprunox group, 0.01 (3.66) for the 40 mg bifeprunox group, -0.39 (3.45) for the placebo group, and -0.89 (3.42) for the risperidone group. The treatment effect values (bifeprunox, placebo) at Endpoint LOCF were: -0.38 for the 30 mg bifeprunox group and 0.29 for the 40 mg bifeprunox group. No notable differences were observed between placebo and either of the two bifeprunox dose groups at Endpoint or at any 10 other timepoint during the study (Week 1 through Week 4). Change from Baseline in CDSS Total Score La~d Obser-vation Carried Fonvard at Each Viut In1tent-to-Treat Popultiion, Treatmas Group Bifepam BeymaRispende Stada:6 20m 0 g Placebo '6 1.;2 me .T Ppai 4 4 4 5 Basehn'e N 3 4 44 5 Mesa (.SD) 3 ( 0 3.52 (4.04), 30 (34) 332 M 2 3 0 MMa, 190 0 P .3Lf t'.i~3 We a2 39 i43 4 -MD) 41 ) 42 Median 0,0 0.0 .0. nMax - 110 - 12 C. - 12.0 3 '14-2 P-value 5224 WL ek 4 139 139 1435 alt Me. 0 10 0 P-valu3e: 0 1 0169Q 3 314 Me D) -05 0 -. 0.0' 0D ":-I. M i M -0.. 4. Week. 4n 13 t39 143 05 MutMa -1 0."- 10 44, 40 10,13 0 1. -a0.413 3 Mesd x f hct' InS" 66 (3CD 0 '- 1 66) 439 -, 5 -0 A2 '771 C=41,R5 O4 O P-vaue 31 0a WO 2008/025781 PCT/EP2007/058958 74 CLINICAL STUDY THREE: The Table below presents mean CDSS scores at Baseline and the mean change from Baseline by visit using LOCF for the ITT population. The mean change (SD) from Baseline to Endpoint in CDSS score was -1.30 (3.85) for the 20 mg bifeprunox group, -0.79 (3.02) for the 30 mg bifeprunox group, -0.59 (4.20) for 5 the placebo group, and -1.47 (3.95) for the olanzapine group. The treatment effect values (bifeprunox, placebo) at Endpoint LOCF were: -0.54 for the 20 mg bifeprunox group and -0.34 for the 20 mg bifeprunox group. No notable differences were observed between the placebo group and either of the two bifeprunox dose groups at Endpoint or at any other time point during the study (Week 1 through Week 4). Clunge from Baseline in CDSS Total Score Last Observation Carried Forward aT Each Visit hitent-to-Treat Population rea&-menfl Group BifeprnrOx Bifepnmnox Okuzapin Statistic 20 mg 30 mg Pl3cebo 1 mg Tetavl Ntunber of Subject 149 48 145 146 in the ITT Popuatkon N Baseline 1i 14 14- 145 146 Mean (SD) 4.34 (4.12 3i3 (;67 4.01 (N3 88) 399 (3 64) Median 4- - 0 3.0 30 Mna Max 0.0, 16.0 Df- 15. 0. 1 H 0., 14t. 0 Change from Baseine Week 149 146 144 146 Mean (SD) -086 C27) -0.77 (2- 63) -064 C2 94) -1.01 9291 Median 0.0 0) 0.0 -1.C1 Min. Max -11. .0S -10.0, 7. -14.0, S.0 -12.0. 7. Pvle0.6l3 3 i). 75 W 2 1 n1 147 145 146 Metan (SD) 19(30) -Q6 9(2.97) -Q63 373) -. 3 343.33) Median -I .0 s-0 -1.0 Mn, Max -11' 9 10 10 -16.0, . -2., 12 0 P-valt 0.215 0. 561 Wek 3 n 149 147 145 146 Mean (SD) -1.41 (3 ,7) -4A.-6 I (3.1) -1 12 (3.83) -1.9 (3.43) Median -1.0 0- -1.0 Min. Max -11. . 9. - 130 -18.0, 1i1 - 1 9.0. P-value 0.30 0.2 Week 4 a 149 147 145 146 Mean (SD) -I 3 5 ( 66) -.6 5 (107) -. 6(48 -1.4913.99) Median 0.0 -1 0 MN, Max -120, 9.0 -11.. 13 0 -18.0, 11 0 -11.0, 9.0 e03S 0U81 We 61ndp9 n 4 147 145 146 Meaniji (SD) 30 ( 0. ( J) -0.59(4.2 0 -1.47 ( 95) Median 0 - 0.0 -1.0 Min. Max -11 G. 9 k -'1 0, 13.0 -18.0, 11-0 -11.0. 9.0 Tt Effect -0.54 -1.34 Tu Effect C1 -1 ,Q 19 - .1- 0.3 P-vale 0.145 0 3 -4 Note: Treament compansons ndi least square mnean were based on an ANCOVA model vth trnnent (?xcling olnzapin) and poled center as fixed factors m11d Basline CDSS total score 3s a covariatt Trt = Treatmes 10 Dna So"rce Tak 2-K CLINICAL STUDY FOUR: The proportion of patients with a CGI-I score 52 at each visit (CGI-I responders; FAS, LOCF) is show below in the Table (Fig. 5; Panel 43).

WO 2008/025781 PCT/EP2007/058958 75 EPS EPS were assessed using treatment -emergent adverse events such as akathisia, dyskinesia, parkinsoniam, etc. and/or formal rating scales such as SAS, BAS, 5 and/or AIMS. Simpson-Angus Scale16 (SAS) The SAS is used to measure Parkinsonian-type symptoms in patients exposed to antipsychotics. The scale consists of 10 items, each rated on a 5-point scale ranging from O(complete absence of the condition) to 4 (presence of the condition in extreme 10 form). The SAStotal score is defined as the sum of all item scores, and the range is 0 to 40. A SAS total score of up to 3 is considered normal. Barnes Akathisia Scalel7 (BAS) The BAS is used to rate observable, restless movements of drug-induced akathisia as well as the subjective awareness of restlessness and any distress 15 associated with the akathisia. The scale consists of 3 items that is rated from 0 (no evidence of akathisia) to 3 (severe akathisia). The BAS total score is defined as the sum of the these three BAS item scores and ranges from is 0 to 9. In addition, a global clinical assessment of akathisia is rated from 0 (no evidence of akathisia) to 5 (severe akathisia). 20 Abnormal Involuntary Movement Scale 18 (AIMS) The AIMS, designed to record the occurrence of dyskinetic movements, consists of 12 items. Items 1 to 7 measure specific involuntary movements on a scale from 0 (none) to 4 (severe). Items 8 to 10 measure global assessment of abnormal movement on a scale from 0 (no awareness) to 4 (aware, severe distress). Items 11 and 12 are 25 questions regarding the dental condition of the patient, with yes/no answers. The total score is calculated by summing AIMS items 1 through 10 and ranges from 0 to 40; the non-global total score is calculated by summing items 1 through 7.

WO 2008/025781 PCT/EP2007/058958 76 CLINICAL STUDY ONE: Treatment GI"up System Crgaia ilase! Preferred Term Stat istic Bitfeprunox ifepruncer 5ifeprnax Risperidcas 5 mo 1C' Mg 20 mg Placebo Z ma overall 1FJSCUISELETAL L-D O MNN4EI'ZE TISSUE D2IDKR 'cnt'd' "uscle Spasms n ri) a 2 ( 241 0 2 C 23 4 C M-ncle Stiffness n C%) 1 (<1% 1 (<1 2 Muses Tightness n ... I (IlI ' 15 uscle Twitcain3 n . 2 C 3M% 1 (<1%' I ( 1 4 C<. Nusacnic'eleIetal Discocfcrt n %C 5 1 (<1 1 Cc musculoekeletal Stiffness n Ci: ' 5 1 i 1 1' myalgia n C 0 .1%) 1 ( I% i 'A11 2 2% 5 Nech Pain n C 1 (<1% 3 ( 3C 2 12% 0 2 2%) a Neck Stiffness n 1 l Pain in Jaw 1% ( 1%) (I! 1 <4 Pain in Limb n <l' 2 ( 2C 3 3%i 4 ( 3%) 2 2%) 12 Sensation of Reavinee n (4:u Q <. 1 NE-PLAB1MS 2EN22f, MALIGEMNT AND tUNSFPECTFIED CINCL CYTS AND C c) 2 2( POLYPS, 2enigo Skin Neoplsmn No. n1%) ' 1 skin Papilloma <1% c IIER'GCS TYSTENi DTISRDERS n 'C. I% D '. (n53%) is. ( -- 52 ;4%m 5i 2 75 14 Akathisia n CC I (<1%C 2 1 2CC 3 C 3CC 0 5 C 5%) 12 C balance Impaired No n C 2 ' 1 C1% C 5 N'te: Percentages for gender-specific adverse evewns are based on the number of subjects in the: appropriate gender. Praoentagas for all other adverse events are based oii the total number of subjects in the Safety population. Note: Sach sutlect is counted at most on(c6 within a esstem organ class and preferred term. Adverse events were code. to system organ class and preferred term using the NeICal dictionary, Verslon 5.0 Iiete- Treat ment:-emergent adverse events are defined as any adverse e.ent that occurred following inititon of stuy medication or any pre-existing medical condition that worsened after randomizatfon. Treatment-emergent adverse events include all adverse events re-ported through a subject's study disccntinatio-n visit and all SAss reported within 20 days after permanent disccotinstion of study medication. Soulrce- PP'TC';L \'JSSTP-NAS'1\VMEDATk\ScRKo0a50tA\MLTSISiTABLELIBIlNAL\kECEHALLS'FETY.SAS, I0TDTILES CUS) I un 2APn04 122 Overt r1Ecidene of Treatment-E nergent Adverse Events, Safety Population Treatment Group System Organ Class/ Preferre-d Term Statietic Bi5&prunIIx Bifeprunox Bifepnnov Rlsperidee 5 mg 1' wg 20 mg. Placab n mg overall NERvos. PYSTEN Dtst1ThmiRS i cont'dC Distuarba:ice in ALtenticui i s 1 .1 1'<ll 2 C'1% DI3U1ele n 't 1z [21%' 12 '10f a k 7t) 12 '10%) S '4 (:34 DrooLrj n S Dr ol iignJ-. tcle) 2 - 4 3 Itl Lycartbria n !D 1 1%) C 2 2t1 3 '1% ryskilmnesia n d 1 ' 1 gi1 r 0 ! 2 '<14 ts t'onia n Qs I l ii 1 :<1C1 S 4%) 7 (1 Extrapyramidal Eisorder n0 ' 2C Gait lanormal Na le) 2 ( 2i)s 2 (l Gait Festinating n ( C 1l'%) 1(1 Grand Mal Cntir 1 ilc 0 1 (1t Headache I-o n 4) 34 (30%> as (3t i 2r (C2'l 2T 21t) 23 '2a9'% 1-53 2 eyperact ivity -yndrome Aggravated n :t C I klt) 1 '<A nypcaesthesia I 1 lt: 1 'Al) 2 Lees cf PropriocEptic n 1 (,1*) C C 1 '1% Memory Impairment n 0 3 (3%) C MIgraine Agravated n 10 l /<I I i% Nigrasine NOS n l (,I1. COulogyric 'rints n 1 1'l 1 1 2 cl Paraesthesia n 4 v l Parkinonian Rest Tremor n ( 1%( 0 Parkioniam n I i1 0 1 .1 Note:- Percentages for gendr-pecific adverse events are based c the number of sub)ects In the prprcriate gender Percentages for all other adverse events are based cn the tc'tail ntmder of subjects is the: safety population. [1te Each> subject is mccmntEd at most ni2-e within a system organ class and preferred ter. Adve:rse evant were caded to nyate organ class and preferred term using the NedDLeA dictionary. Version 5.. Note: Treatment-emergent adverse events are defined as any adverse event that occurred followIng initiation of study medIcation or any pre-existing medical c condition that worsened after randomirat-Ion. Treatment-emergent adverse events inside all adverse events reported through a Lject's study diecontinuatina iait antd all Ses reported within I. days after permanent dis-ontnuation of study medicatiat Scarce: PRT00iL U'R'YTP-MS\OWbSDATA\SAStO000591\CM LTIS\TILLT\TL\AERALL SAPETY.SAC., -rIDTIL1Ef ) 'Rmun 29LFR04 12:26 WO 2008/025781 PCT/EP2007/058958 77 Overall Incidence of TreatAment-Einergent Adverse Events: Safety Populatin Treatment iroup S,'stem crgan Class. Preferred Term Stati stic 2iEeprunnx .tfeprunccc Sifeprnox Risperidcna 5 mg 10 mg 20 ltg Placeh 'S mg Kverall JPSVK#US 2YSTE1 DISOPE2RS icCnt d Sedation n r? 3 5? 1C4 f a 10C (t*) 10 ( at is 13h 51 I scamnclence n % 2 ( 2' I - ' E%) E 4%) S 4) 22 Synccpe I (,1%) 3 3v I 1 1 Tardive Eyskinesia n V0 1 (% 1 1 2 211 Tremor n% 1 ,1%) 6 % J TrIamue n 4 1 (,1%) F '1 1 (21 PSYCRTATRTC DTsaP~EPS n 47 (41% 30 r32% 46 :4%. 44 '7%) 47 2 Acute psychrsis u, 1 (,1%) 1 AdiuLtmEIt Dizorder Ncis n 0 l1 Affect Lail1% 1 I21%) 2 J Agrsso .J.1(<%) .' 12ct et (41 Agit atlic n 13 (11% 11 ( %) 13 tilll l . S1S I St) S7 (1 Agitat ioin ALravated n4t 5 ( 4C I :,lt! 2%) 0 12 Al~chli Haagever n t DI.1 44 y41 Anbedonia n (C 0 0 l.1 vl Aaxiety n (r 1 (1 a akxI.Av Ig-Tu~ e 14 AnSit Agavated n t 2 3%) 3 ; 3% 2 n21 2 2k) 4 . 1)40 Bruxism n 1 %rl Qatatlia (1%) 0 1 (<j cafusi'on ( ,I) 1 l 2 Dot : Percitages for gender-specific adverse events are based on the number of subject in the appropiat gendr . Percentages Eor all other adverse events are based on the total nLniber of aub cts in the Safety population. Mite: 1ach subject t3 :counted at most once within a system organ class and preferred term. Adverse events wer coded to system -rgan class and preferred termI uing the M-ER dictionary, Version 5.0. Nor&: Treatment -emergent adverse events are def ined as any adverse e.ent that occurred following i nitia-tion of study medicatiton or an pre-existing medical cndit ion that worsened after rand-mizat i.n. Treatment-emergent adverse events inlude all adverse events reported through a subject's study disccntination visit and all O-IEe reported within 20 days after permanent discontination of Study medlcatimon ourc':-e: PRCT'~-OL \KUSRTP-taSQ.lkMSDATAs \S\RKh00es10\kAILTSIG\TEELE\lThLiAEOVEJ.RLL 2SETY.22S. QT.IIJDTILES *r29 Run 2-9&PF4 12:2s WO 2008/025781 PCT/EP2007/058958 78 CLINICAL STUDY TWO: OCern Incidence of Treatment-Emergeut Advere Events: Safety Populadon Prctcc:I l £14390'J1 Selvav Phaimrmaeutica Fina1 9tudy, Rep.rt Treatmtnt 3rup Cystern Orca cl / feprnr EprmCr isperir1nme Preferred Term taticit ic 3 mg 40 mg Pla:ebo mvera tEPV';cr CYSTEM [-fERRtE:n; 2 G r. 3 C E: i 4: f - rr 2 cl 2) E29 -YATHICI A nFt E( 10 ( 71! Ei2I,) 16I 10&) 2G BAIIW'E DISORDER n05 l) DC 1 BRAr&KIllEsIA n0 02 1 LR'Of(EATRETSI& 0 0 *G~REEL EIGTETTn1 1. l D1' TUFEkN A I' TTELITLMI n: 1) DIZ:IE-S n 1 2 '0 14S) 6IN 4, 9 i GU 46 IARTMFIA 0 0 3 a 2ki 'I DY3EUSIA n1 0 0 1 E{T2.PYP4MIDL DISDRDER n ~ II4 S E) 21 14l ill aEADACH2 n (i 12 (2.) ZE 17 17 21 i4) 23 15') 2 q - DE>nots gender-speefit advere events. Percent for geder-pecif adve2re events are dn the number of ubjct, in the aprrpri~ete gender. Perontag e fr a1 other _ avene everimtn are 1eon the total mIner f ubjectS :in Gafety Pplat ien Note Eazh SLLbject iocounte:e at mos n within a yotem orgAn claor and praferre:S trm. Adverse event we-r C-ded tc- svota. crgan and preEered te.e us'ig the Me1DK dizticar-.. Versi-n . Nt,.: Trt.nri-eerent averze event are define a any HAve:rxe event that =courr-d fllowing iiit ie~itin ,f otucly me'iation Cr any prt-existing medial cn'ition that vwrmened in severty aftr randomizstion. 'Treatrnt -ergemt A'erse -nt, in-,ld, all ad'erse e-vente reirnt~ed through a subject 's stuiy dicontinuation vioit -d nll AEo reported within 0 dayo aEter tudy 1kt-:: Advarnent i lude nw ,rcened phyi.via1 xaminatisn abnrm.alitoes. D. t,, f-re Tablt 4.1 0 Overall Incidence of Treatment-Emergent Adverse Events: Safety PopuLi:oni Prote'el RL14;001 Solvoy Pharrmoeu-tials pinl Stu& Report Treatment Group Gvytem Qrgn Claso/ Cifpruns Di.prnr- Rimperirdce Praferred Term Sttistic 30 mg 40 mg PlacebC 6 g Ove 1RPV7US SYSTEM rISOPIERS (cont' d) 117P09ET4ET n ). 0 0 0 1 1-) 1Q MEVRY ~[MPLPfMET n 1t 1 'sc) 2 FIRAINE n ( 1 I'll 0 0 C t PARIE--VEEj nm 21 14 pSYl4H. TOR J4iPER2CTlVIT n 0 S-ETTKN n )E 4 3 ) E I -) 24i C~qElRDACHEnM) <% TA-RDTvE nYSrrrESTA n )1 . 2n.) TEEML n ( 9 ( M 10 (7' 7 ? V 7 SU 5 33 A Eeote gender- opec if ic adverse event. ot: per-ntog-s for n p i advero event" are ad on the nuber of object. the appr-priat gedr Pern-tgez f-r a] .thr adver:e event re 1-- the tet al numner rf :!bjecto iar-fety Populati. Eh -ubject i counted at, mo once within yitemrgn c3 - nd preferrei term.. Ayereeven t r Aed t yter -3g C1l... :-d pr-Efere t-.r u-ng the MedDFA dizti:ma.. ce'o~ .1. Nt :: Tr et-Eergent aere evt:I re def.ined a ay I :re e-t that occurred fllC1eing initiaticn CIE sty Imedicaticor ny pre-ei -tng medical condflit'on that opened in -everity afterr Treatmt-eegnt dvem ente inlude al verse even~tr rei:'orted through a subject'. study discontinu.ion visit and ll SAEz reported within 1 days after study 2~di1 nLinuati~on. l1t.: Adv-re events iclude n -w:or roene-d phyni-al e-aImintin bnomalitiez, Dtate Smrce Table 4L 5 WO 2008/025781 PCT/EP2007/058958 79 CLINICAL STUDY THREE: Overall hiddence: of Treatment-Emergent Adve Events: Safety Popubtion Tr-atn-nt Group Eifepnrut: PifEpuncz Ol aspine SymtLem Or9an Clac/ ') ,rg 26 rng Placebo 1E mg C'serall MtscULcELETL TF'S n1 ) 1 )< " G Y A [ 0 ST I nVI C -N f r,)r I G l D r O L 2 f d i , - t) 0 6 . l FT1AL12IA z 1 HECK PAI ni 1% 4 PAIN .lW EXTFEITf n: 4 it0 3 FAIlV IN JAN, t dl1 ' l1 tfFOPLSMs BEIrt. roALIGAUT V J.- UtSPECIFIED IDiIQL CYfSTS 211]E F'3IYPE n 0 1 v I:!) , 0 1(:1 CYST n C0 i -:1 2 0 l ( j NERVOUS EYISTEFI IIDEsaP n) soS vr ( yi-') SO 40f) C( 27ki 22 1-R) AETRII A n S E ) I 4'' 4 s i I& 2'' BALAECF DISORDER n I l 1 g!PJBl EEATION n. ). 2 1 2 .l + D:ndts g-nder-pifiz advreevents. ot,, Pewaentages for gndtr-opec'fic adverse events ar-e e n the nuber of ubjct. in the apprrte gender. Percentages fo all other schrerne event are b:end1 n the totdl fnmer esf subiects in the Safetv population. otea Each r ubjact is counted at n ot core withiri -'ytem crgan 1-o n preferrd tr. 53e:r-events were -- r.d!e tov yst. org".n erred tm using tle MecEA dicticnary Verien g.1 Hot,- Treatment-emnerget adverse events ar defined ny a&serze -evet that c-urrd fo1lwing initiAtion of stucv medicationr ari Fre-eiting medical ccndit.!on that worened in severity aEtr randomizatin. Tratment-enerent adverse vent. in1ude all dr events reprrted- I thrugh a subj::.-s tuIy diiccatinuatiin vi3it -n all SAEo rep-rted vithin ~ day. after o-tudy diaDntillatic . Dat&1:urc,, Tabl 4.1.. Overall iciden~e of Trearmwnt-Emergen Adwrse Eveni: Sarety Population Treatant Group Efeprun- l feprur:.xOlanpin Syotem Crgnn Claso./ Wj g 0 ur. Fla-eb: 1E Trg overall PreferreA Term Stt iotic (.D-1G4 ifl-150) iH-t9 (M-1E: E-GO.) NlRST1US SYSTEM D1h2RlEES (ccnt 'd) MINVULSTON1 -n (2- 1 , 1-'. IS) U C. 0 2 -I.Nl) DISTURBANGE IN ATTEUTIN ( 0 1 <lk) 0 1 <1 ) DIZERTHES n12 e E )) IC ,D 7j .DYiST.PTRPIA ni0 1 TI DYSKITEGIA ) 1\ DY-TCNt1A 2 ( C. 2 ( 1I: C 4 ( lo) EXTRAFYPhd ML DIgDPJER n 2i E jf 1 1) 4 1) 12 PACIAL PALSY nt 0 1 1l 1 HZADACH E n( 3c0 i 19) 29 l ( 9% 32 21'l,) 12e 121,i 109 I 19\ RYAETRE lA As 1 6: 1 1 l)(A RYPOKIIf n 1 0 0 1 I I 1 i IMMTAIL lMP.1lRMENT n(19 0 1 ( 1) 0 1 i 1') e .Dentas 3rnder-opeczif iz adlvere evento.. Eote: Perentage for gnder-pecffic asers'e event re - n the number of -ubject 31n the apprcpriat gnder Perceitage, f-r all ath-r a~dvrse e'ento are bxd1rn the total nunLer nf subjcto in the Safety populat ion. Iote Eoh ubject is counted at .ot .n-, within , trm r clas and preferred t trm. tAr ,e events wra :':ed t, syotem orjn l, and peEerred t:e:r i thE MedDPA di.tinary. er4 in -. 1. tbte, Tratmnt- emergent adver-e events are defined ,a any ad e event that ccutrrd foll-ing inibititi of t.d rmediention r any pre-istng3 med-l ccandition tlht -r.ened in everty after r Enndamiratin Treatment-emergent adver - events in-lud 11 erLa evenLs rep~zrted thr-ugh a -ubjet:'o tudy discontinuation vioit and ell LAo reported within 2.0 Jay after study Da.taSource Table 4,1.6 WO 2008/025781 PCT/EP2007/058958 80 Ovr'al hiddence of Tieatinmied inrgen Advent- Events:. Safety Population Treatment Group Eiftpac Eifepruo 01 snepine tyatemi Organ. Claaf 2 r 20 a Placebo it mra Overall Preferred Term ,tat ietic (1-t4) (t11-1t01 iR-19 )011-1t01 )01i- i 1lS!V01T] SYSTEM ISOPMDERG cost' c) TIEUJOSIICAL SYlPMT.W nsi 0 1 < 312) 0 0 1 < I:1 PARAE 'IRGIjAn(i (1) ( l) 1(1) 4(<1 RESTLESS LEG2 SBRCF Et 10 7 C 4 12 01 i 3 C 64 IRn0 FfEAAIC-1E GORULEZE n;t; CE 3 2) 11 8 ~ W) l 9"t) 29 ItG SPEECH DISERn( 0 0 0 : SYrIDPE VAEG i'l n S TAPDIVE DYSKITMSIA n () 1(ek 1 TEM. 1) 4 3-) 1 ":) 4 1 3%) I5 I 2E 5 D-note gender-apeoiE:5n adverse eventa lAste2 Preretgeo far geder-opecific aodve:rse events arrc hard c-n the number of subjects i-n the apgropriate perier - Percentagee fEsr all other odrerse events are based an the total number of subject in the tofety population. Nste Each subject is counted at noot onc-e within e sy-tem organ r-s ond preferred term. Averce event eme oleI t. nc-te crg-an 1ons and pre erred tern ui5.g the MedDPRA dictinary., Verian -.. 1. Note Tr-eatent-emergent adverse events arec defined a any a&erse event that occurred folldings initiation of tudy mediat ion -or a-y pre-existini si::il condtin that wamened in ceve:rity after raadomiration. Treotment-rrcqe:nt adverar event. include oll adverne evennta reported though a Subject.'o nto::y d:imconti-ristion viait and all SAs reported witin 10 day & after rtudy di 0.0 on tusaan Drteurce:. 'Table 4.,0 CLINICAL STUDY FOUR: The proportion of patients with TEAEs related to EPS was lower in the PBO group (4%) than in either of the BX groups (BX20: 10%; BX30: 15%). TEAEs related to 5 EPS for which there were 3 patients in either BX group relative to the PBO group comprised: BX20 - akathisia; BX30 - dyskinesia, akathisia, extrapyramidal disorder. In total, 15 patients had TEAEs related to EPS that led to withdrawal (2 in the PBO group; 4 in the BX20 group; 9 in the BX30 group).

WO 2008/025781 PCT/EP2007/058958 81 Panel 48 All EPS TEAEs by SOC and Preferred Term (APTS) Pao BX20 BX3O System Organ Class I Pref'erred Term Pat ients Treated 166 159 172 Patients with EPS TEAEs 7 (4.2) 16 (1 o1) 25 (14.5) NERVOU3S SYSTEM DISORDERS 7 (4.2) 16 (1 .1) 25 (14.5) Dy skin 1 ( J) 2 (.3) 11 (b.4) Akathlit 2( 2) 11 (6.a9) 8 (4.7) Extrapyranidal disorder 1 6) 2.9) Athetos i s1 ChorF- the t i s1 . 3 (1 i) 1 . Dystorn 1 (0 i MUSCULOSKELETAL AND 1 ( ) CONNECTIVE TISSUE D ISORDERS uscle rigidity 1( ) (.) S} Refers to a G ,ender Specific AE Coding is done in MadDRA version R 1: 10214 ST AE04 EPS 14:-07 a :14 737 200/313 TFL/SAD Bu l Id Nulm:ber - F ina SAS, BAS & AIMS CLINICAL STUDY ONE: Simpson-Angus Scale global scores were evaluated as normal or abnormal at Baseline and Endpoint. The percentages of subjects with 5 normal and abnormal scores at each time point. At Baseline, the percentages of subjects with abnormal SAS scores ranged from 8% to 12%. At Endpoint, the percentages of subjects with abnormal SAS scores ranged from 3% to 10%. The percentage of subjects with abnormal SAS scores decreased between Baseline and Endpoint in all treatment groups. There were no statistically significant differences among the treatment groups in 10 the percentage of subjects shifting between categories from Baseline to Endpoint (p=0.800). The BAS score consists of an objective score, a score awareness of restlessness, a score of distress related to restlessness, a 3-item total score, and a global clinical assessment score. There were no statistically significant differences 15 among the treatment groups at Baseline or Endpoint for the objective score, awareness of restlessness score, distress related to restlessness score, the 3-item total score, or the global clinical assessment score. There were no statistically significant differences among the treatment groups in the percentage of subjects shifting between categories from Baseline to Endpoint in BAS 20 total scores or BAS Global Clinical Assessment scores. There were no statistically significant differences among the treatment groups at Baseline or Endpoint. There was WO 2008/025781 PCT/EP2007/058958 82 no statistically significant difference among the treatment groups in the values for change from Baseline to Endpoint. CLINICAL STUDY TWO: Simpson-Angus Scale global scores were evaluated as normal or abnormal at Baseline and Endpoint. At Baseline, the percentages of 5 subjects with abnormal SAS scores were: 14% in the 30 mg bifeprunox group, 11% in the 40 mg bifeprunox group, 7% in the placebo group, and 6% in the risperidone group. The percentage of subjects with abnormal SAS scores between Baseline and Endpoint decreased in the bifeprunox groups, were unchanged in the placebo group, and increased in the risperidone group. At Endpoint, the percentages of subjects with 10 abnormal SAS scores ranged from 7% in the 40 mg bifeprunox and placebo groups to 14% in the risperidone group. Shifts from normal at Baseline to abnormal at Endpoint were the highest in the risperidone group (19 subjects, 12%) compared to the other three treatment groups (30 mg bifeprunox: one subject, <1%; 40 mg bifeprunox: five subjects, 4%; placebo (six 15 subjects, 4%). A statistically significant difference (p<0.001) among the treatment groups was observed. The BAS score consists of an objective score, a subjective awareness of restlessness score, a score of distress related to restlessness, a 3-item total score, and a global clinical assessment score. There were no statistically significant differences 20 across the treatment groups at Endpoint for the objective score (p=0.093), awareness of restlessness score (p=0.368), distress related to restlessness score (p=0.779), the 3 item total BAS score (p=0.433), or the global clinical assessment of akathisia score (p=0.541). There were also no statistically significant differences across treatment groups at Baseline for the above measures (p 0.168). 25 There were no statistically significant differences across the treatment groups in the percentage of subjects shifting between categories from Baseline to Endpoint in BAS total scores (p = 0.482) or BAS Global Clinical Assessment scores (p=0.91 1). There were no statistically significant differences across the treatment groups at Baseline (mean range = 1.0 to 1.2; p=0.923) or Endpoint (mean range = 0.9 to 1.6; 30 p=0.138). There was no statistically significant difference across the treatment groups in the values for change from Baseline to Endpoint (mean range=-0.3 to 0.4; p=0.138). CLINICAL STUDY THREE: Simpson-Angus Scale global scores were evaluated as normal or abnormal at Baseline and Endpoint. At Baseline, the percentages of subjects with abnormal SAS scores ranged from 6% to 10%. At Endpoint, the WO 2008/025781 PCT/EP2007/058958 83 percentages of subjects with abnormal SAS scores ranged from 3% to 8%. The percentage of subjects with abnormal SAS scores decreased between Baseline and Endpoint in all treatment groups. There were no statistically significant differences among the treatment groups in the percentage of subjects shifting 5 between categories from Baseline to Endpoint (p = 0.463). The BAS score consists of an objective score, a subjective awareness of restlessness score, a score of distress related to restlessness, a 3-item total score, and a global clinical assessment score. There were no statistically significant differences among the treatment groups at Endpoint for the objective score (p = 0.421), awareness of restlessness score (p = 10 0.584), distress related to restlessness score (p = 0.254), the 3-item total score (p = 0.865), or the global clinical assessment score (p = 0.518). There was a statistically significant difference among treatment groups at Baseline for the subjective distress related to restlessness score (p = 0.029), but not for any other scores. There were no statistically significant differences among the treatment groups in 15 the percentage of subjects shifting between categories from Baseline to Endpoint in BAS total scores (p = 0.808) or BAS Global Clinical Assessment scores (p = 0.525). There were no statistically significant differences among the treatment groups at Baseline (p = 0.362) or Endpoint (p = 0.187). There was no statistically significant difference among the treatment groups in the values for change from Baseline to 20 Endpoint (p = 0.187). CLINICAL STUDY FOUR: SAS At baseline, the mean SAS total scores in all three treatment groups ranged from 25 2.17 to 2.33. There were minor fluctuations in the mean SAS total scores in all three treatment groups during the study, and at Month 6, the mean SAS total scores ranged from 0.57 to 1.16 (APTS, OC); thus, the mean total scores at baseline and at Month 6 were within normal range. The adjusted mean maximal changes from baseline to Month 6 in SAS 30 total scores were <1 in each treatment group (APTS, OC, ANCOVA). There were no clinically relevant differences between any of the treatment groups in SAS total scores. At baseline, the majority of patients (PBO: 74%; BX20: 71%; BX30: 72%) were normal with respect to SAS status and the proportion of normal patients had increased at Month 6 (PBO: 85%; BX20: 89%; BX30: 78%). There were no statistically significant WO 2008/025781 PCT/EP2007/058958 84 differences between any of the treatment groups in the shifts in SAS status (categories: no change; abnormal to normal; normal to abnormal). BAS At baseline, the mean BAS total scores in all three treatment groups ranged and 5 from 0.44 to 0.46. There were minor fluctuations in the mean BAS total score in all three treatment groups during the study, and at Month 6, the mean BAS total scores ranged from 0.05 to 0.17 (APTS, OC). The adjusted mean maximal changes from baseline to Month 6 in BAS total scores were <0.6 in each treatment group (APTS, OC, ANCOVA). There were no clinically relevant differences between any of the treatment groups in 10 BAS total scores. For the BAS global assessment scores, the trends were similar: there were minor fluctuations in the mean BAS global assessment scores in all three treatment groups during the study; the mean scores ranged from 0.22 to 0.26 at baseline, and from 0.07 to 0.14 at Month 6 (APTS, OC). There were no clinically relevant differences between any 15 of the treatment groups in BAS global assessment scores at Month 6 (APTS, OC). There were no statistically significant differences between any of the treatment groups in the distribution of BAS items 1, 2, or 3. AIMS The mean AIMS total scores in all three treatment groups were low and ranged 20 from 1.04 to 1.35 at baseline. There was for all treatment groups an initial increase in the adjusted mean scores (largest in the BX30 group), after which the mean scores decreased over time; at Month 6, the scores had returned to baseline level in all three treatment groups (PBO: 0.11; BX20: 1.08; BX30: 0.86 (APTS, OC)). The adjusted mean changes in AIMS total scores in the BX20 (all time points) and BX30 (Months 5 and 6) 25 groups were not statistically significantly different from those in the PBO group, whereas the adjusted mean change for BX30 (from Week 1 to Month 4) group was. The adjusted mean maximal changes from baseline to Month 6 in AIMS total score were small (PBO: 0.39; BX20: 1.21; BX30: 2.45, APTS, OC, ANCOVA). None of the differences were considered clinically significant.

WO 2008/025781 PCT/EP2007/058958 85 Panel 49 Movement Rahng Scales Scores (APTS) PBO" BX201 BX30 OC LOCF OC LOCF OC LOCT SAS BasA7ne 2 17 (3.541 2.

1 7(35) 227 (3) 2 ) 2.33 (3.53) 2.33; (3.53 Week 6 L42 (2.44) 1. 54 (2 79) 1.31 (2.69) 1.43(27 2.06 (3.41) 2.35 (3.IS7) Month 6 0.57 1.6) 1 33 2 6S) 0.97 (2 20) 1 22 (2.47 L 16 (278) 2.09 3. 64) BAS Basehne .46 (1.13) 3.46 (1 13) 0.44 ( 10) 0.44 (1 I0) 0.45 (1.11) .

4 5 (L11) Week 6 0 (0. 6) 0.34 (0 94) 0.36 ( 99) U.49 (1 21) 3 (L01) .66 ( 55) Month 6 u.5 (1) 0.34 (0.99) 0.17 (0 42) 0.4 1 5 16 (0.65) 0.69 ( ALMS Baseline 1 10 (2.82 1 10 (2 82) 1 4 (,6) 1 04 26) 35 (29) 1335 (279) Week6 092 (70 3) 1. 05'( 12) 1 16 (2.55)l 1 8 9 (4.5) 2.83 (680 ( Month 6 0 11 (0.44) 8 ( 0 (2.3) 1 55 ( .92) _ 6 (18 2.77 (6. 09) Valuesi. are meansD a PBO - OC: basee n = 166, W\eek 6 n= 1 1 Month 6 n = 44: LOCF: n = 16 b BX20 - OC: basefne: n = 159, Week 6: n = 104.. Mnth 6: n = 59; LOCF: n= 159 c X3 - OC: bas-ene: n = 172, Week 6: n = I0 6, Mth 6: n = OCF: n 17 Body Weight 5 CLINICAL STUDY ONE: Small decreases in weight were observed in the bifeprunox treatment groups, but not in the placebo treatment group. At Endpoint, the mean weight loss for subjects in the bifeprunox treatment groups was approximately 1 lb (5 mg bifeprunox, -1.0 lb; 10 mg bifeprunox, -1.3 lb; 20 mg bifeprunox, -0.6 lbs). The placebo treatment group had a 10 mean weight gain of 1.9 lbs. Statistical testing of the change from Baseline in weight was performed using similar methodology to that used for the secondary efficacy parameters (an ANCOVA model with factors for treatment and weight at Baseline). Pairwise comparisons of the bifeprunox treatment groups versus placebo were all statistically significant (bifeprunox 5 15 mg [p=0.025], bifeprunox 10 mg [p=0.009] and bifeprunox 20 mg [p=0.031] for the difference between bifeprunox and placebo mean change from Baseline in weight at Endpoint. The mean weight change of subjects in the risperidone treatment group was a 4.8 lb increase.

WO 2008/025781 PCT/EP2007/058958 86 Increases in weight: The percentages of subjects whose weight increased by more than 7% in the bifeprunox treatment groups (2 to 4%) were less than or similar to the percentage observed in the placebo treatment group (5%). In the risperidone treatment group, 13% of subjects increased their weight by 7%. 5 Decreases in weight: The percentage of subjects whose weight decreased by more than 7% was higher in the bifeprunox treatment groups (5 mg, 6%, 10 mg, 7%; 20 mg, 8%) than in the placebo treatment group (3%). No subjects in the risperidone treatment group decreased their weight by 27%. 10 CLINICAL STUDY TWO: Small decreases in mean body weight were observed in the bifeprunox treatment groups and the placebo group at Endpoint, while an increase was noted for the risperidone treatment group (Table below). At Endpoint, the mean weight changes were: -2.2 lbs in 30 mg bifeprunox group; -1.9 lbs in 40 mg bifeprunox group; 0.5 lbs in placebo 15 group; and 3.2 lbs in risperidone group. In the bifeprunox treatment groups, mean changes at Week 6 were higher than those at Endpoint. Chang es in Body Weight at XVek 6 and Endpoint, Safety PopuLatin Tesmet Gamp Bifsp5ex ifpsRipda ai f45 N=17 N349 N=154 145 1 49 154 Medsk16 11515U Ma.ax S4., 360 W4 37S 7S 3 'K. 4'4 Chaee om' B eleY at Week 6S, 6 6 4 4 8 -1. -4~ %3i Min Max -3.8-P15-11 1.3 Chag fus einat i n 134 4 1 edIt e n 10V -1 0 Data Swnw Tale 9.4 WO 2008/025781 PCT/EP2007/058958 87 CLINICAL STUDY THREE: Small decreases in mean body weight were observed in the bifeprunox treatment groups and the placebo group, while an increase was noted for the olanzapine treatment group. At Endpoint, the mean weight changes were: -2.3 lbs in 20 mg bifeprunox group; -1.1 lbs in 30 mg bifeprunox group; -1.3 lbs in 5 placebo group; and 5.2 lbs in olanzapine group. Mean changes at Week 6 were comparable to those at Endpoint ( Table below). The incidence of markedly abnormal ( 7%) decrease in body weight was comparable among bifeprunox treatment and placebo groups (20 mg bifeprunox: 6%; 30 mg bifeprunox: 5%; placebo: 5%) and was lower in the olanzapine treatment group (< 10 1%). The converse was seen in the incidence of markedly abnormal ( 7%) increase in body weight with 19% in the olanzapine treatment group compared to 5% in the 30 mg bifeprunox, 1% in the 20 mg bifeprunox, and 5% in the placebo group who experienced this. Changes in Body Wght a Xeek 6 and Endpoint Safety Population 3U mg~ Pla&cebo In' 7c.al NumbeP of Subji '. N"=1 54 N=150C 1 4 9 N15 0 MemSD 10.6617) 179 1t"6 2.4) 17 430553)1 75 527 Clunae i e Baelin Wek 6- n 746-"79 al&i Men 1 73.97 AMbdin -1 -20 -Na Man L -34, 2:6 -13, 39-5, 7 3 Charenm Bse iuak Endpames 1 4 1 N '43 14 Baselim~ Mea 1 279.77521. 3.rSD) - ,,1 -1.1 (17. - 3 1 12 (6 9' C. 4 -4, 26 5, 4 1 3 'e: E w d d4la ce qhlAdgJem taint n e7m Data Sm-,7ze TalAe 9.4.'-- WO 2008/025781 PCT/EP2007/058958 88 CLINICAL STUDY FOUR: Weight and BMI and changes therein relative to baseline are summarized in the Tables below. In all groups, the adjusted mean weight and BMI decreased from baseline to Month 6. The adjusted mean weight change from baseline to Month 6 5 (APTS, OC, ANCOVA) was -0.8kg in the PBO group, -0.3kg in the BX20 group, and 0.5kg in the BX30 group. The adjusted mean weight decreases in the BX20 and the BX30 group were not statistically significantly different from that in the PBO group. Weight decreased was reported as a TEAE for patients in all treatment groups (PBO: 5%; BX20: 8%; BX30: 8%) (Panel 45). Weight increased was reported as a 10 TEAE for patients in the PBO group (1.8%) and in the BX30 group (1.2%). The adjusted mean weight change from baseline to Month 6 (APTS, LOCF, ANCOVA) in patients with/without nausea and/or vomiting showed that patients in all groups lost weight irrespective of whether they had nausea and/or vomiting, although those patients who also had nausea and/or vomiting had a greater weight decrease 15 (PBO: -0.6 versus -1.9kg; BX20: -1.0 versus -1.9kg; BX30: -1.1 versus -2.3kg.

WO 2008/025781 PCT/EP2007/058958 89 Table 212 Weight BM1 and Waid Measuremnen (APTS) PCs LOW High Assessment Treatment Visit n Ean SD tedian Min Max n () n () BM1I Pe2 Bas ljine 166 25,2 4.e 24,2 17.4 41,5 Wo@k 6 112 25.0 5.U 23.9 17.2 41.1 min3th 3 719 25,1 5 3 23,8 17.7 42.1 Month 6 43 25.3 5.1 24.2 18.1 42.1 Last* 150 25.1 5.0 24.2 16 42.1 9X2O G 1a sine 159 24.7 4.3 24.0 17 3 40. 9 Wek :6 104 24.3 4;2 23. 167 39.3 Month 3 81 23.9 4 .2 23.1 16 3 40.2 MLanth 6 59 239 3.7 231 1159 32 Last* 143 24.3 4.2 23.7 15. 40.2 BX3' gasline 172 24.9 4.5 23.9 14.3 43.6 Wek b6 1(6 24,8 4.3 23,9 13. 42,6 Mamnth 3 8 24.4 4.3 23.4 13.2 41.5 Mo-nth 6 55 24,4 3.3 23,7 17.6 31,2 Last' 148 24.5 4.4 23.7 13.2 41.5 PCS: LOW:N/A, HIGI :N /A WA|IST PBO pasiaS|n 90 4.0 12 ;8 20 62 117 WeekA :6 62 93.2 1:2,4 92. 0 F,2 117 Month 3 47 84.3 12 0 82.2 62 115 Manth 6 31 85.1 12 4 85.0 7 116 Last* 91 84.7 12 2 93.9 62 116 2X2 gasol I ine 84 82,2 12 6 SO 0 59 123 W-ek .6 55 81.6 14 5 79A) U , 123 month 3 54 79-,5 13. 2 76,5 57.9 123 Mnlnth 6 44 79.4 1 2.1 7717 58.5 110 Last* 89 81,6 12.4 79,5 58.5 123 X3I0 Baslc 1in 93 83.7 15.8 91.0 55 1 I Week 6 59 23.9 16.7 20.0 62 160 Minth 3 54 91.6 14.8 79.5 61 154 Ma:nth 6 38 81A1 10,9 781 62,5 112 Last* 94 92.6 15 3 80.0 61 152 PCS: LOW:N/A. HIGH4/A WEI: GHT PBO Basl ine 166 71.3 14 5 70 0 42 117 Wook 6 112 71,0 1 5. 9, 0 43 117 2 (1.8) 0 (0. G) Mrnth 3 78 707 16.6 58 .0 43 114 7 9. 0) 2 2.) month 6 43 71 6 16.2 68,0 0 4 117 4 (q. 3) 3 (7 G) Last, 150 7112 1 5.S A 09 43 117 11 7.3) 4 (2.7) BX20 BasRI ine 159 70.7 13.9 6. 0 45 125 We:k 6 104 69.7 13.8 67.0 46 123 5 4 .8) 0 (0 L) Ma:nth 3 81 68 1 14 1 66 0 45 126 9 (9 9) 0 (0,0) Minth 6 59 68. 2 12 3 67.0 45 99 6 (10.2) 4 (6. 8) Last* 143 697 13 5 68.0 45 126 13 (9 1) 4 (2,8) X3' BaI li no 172 71 2 15 2 700 3a 126 Week 6 106 72.2 15 6 70.U 37 123 5 (4 7): 0 (D0): Month 3 AS 70. 1 14.7 68.0 35 120 10 (1l 4) 1 c. 1 Manth 6 5S 71 4 12. 2 7O.0 49 103 4 '(7 -3) 4 (7-3): Last* 148 7 0.9 14.9 68.5 35 120 14 (9. 5) 4 (2.7) PCS: LOW: De -creamg from basull inc >= 7 %, NIlGH: increase from bascI irn > 7 % Last post-asel ine observation 10214 ST WB21 09SJUN2006:15:5559 261:/313 - TFL/SAD BuLildid tCbr$ - Final WO 2008/025781 PCT/EP2007/058958 90 Table 13 Change from Baehne in Weight BMI nmd Waist Measurement (APTS) PCS LOW High Assossamt Treatment Visit Tl Mean SD tad Ian Min Max n (%) f (%) BMl P2o Wiek 6 112 -0.1 0.7 0,0 -2.9 1 54 Month: 3 78 -0.1 1 I 0.0 -3.3 4 19 Month: 6 43 -0.1 1.4 0,0 -4.6 3 43 Last* 150 -0.1 1.2 0 -4.6 8 55 2X20 Week 6 104 -0.2 0. 7 0 0 -2.3 1. 42 Month 3 S1 -0.2 1 0 0.0 -3:.2 1.63 Month. 6 59 -0,2 1,3 0.0 -4.6 2.29 Last* 143 -0.4 1 0 0.0 -4.6 2.29 2X30 tpok 6 10& -073 11 D 0 -9,3 1. 31 Month. 3 S8 -0,4 1,0 -C.1 -3.7 2.13 Month P 5 L -O 2 11 1 C.0 -3 2.61 Last* 148 -0.4 1.3 .0 -9.3 2.61 PCS: LOW:N/A. H1G :NI/A WA IST PBO Wiek 6 61 0.0 0. 2 0,0 -0.5 1 Month: 3 43 0.6 4.4 0,0 -9 17, 4 Month: 6 27 -0.0 4.9 0,0 -9.2 14 2 Last* S1 -0.1 3.7 0.0 -9.2 17.4 2X20 We ek 6 55 -0.0 0.3 0.0 -2 0 Month 3 44 -1,1 3.9 0.0 -21 3.2 Month: 6 35 -1.1 3.8 0.0 -12 4 4 Last 75 -09 3.3 0.0 -15 4.4 2X30 wiek I 59 -0.0 0 3 0. -2 0 Month: 3 47 -04 2;5 0.0 -10 4.5 Munth: 1 30 -O G 4.1 0.1 -20 5 Lat* 84 -0A 4 2.9 0.0 -20 5 PS: LW: N/A HI GH:N/A WElIGHT PBo Wek 6 112 -0.3 1.9 0.0 -7 5 2 (1. 9) 0 (Ci. Li) Month: 3 78 -0.4 3.3 00 -1:1 11 7 (9.0) 2 (2 6) Month: 6 43 -0.4 3.9 0.0 -14 9 4 (9.3) 3 (7 0) Last* 150 -0.4 3.4 0 0 -14 23 11 (7.3) 4 (2 7) BX210 Week 6 104 -0,6 2 0 0.0 -7 4 5 (4. 8 0 (0 0i Month: 3 91 -0.6 2.7 0.0 - 5 9 (9.9) 8 ( 0 0 Month 6 59 -0.6 3,7 0.0 -13 7 6. (10.2) 4 (6..8 La.t* 143 1 1 2 j . -13 7 13 ( 91' 4 (2 BX3f) Wek 6 10 -0-9 3;7 0.0 -32 4 5 (4.7) 0 (0.0} Month: 3 SB -1 1 30 -0.5 -10 6 10 (11 -4) 1 >1. 1) Month. 6 55 -0,1 3,3 0.0 -1G 9 4 (7.3 4 (7.3) Last* 148 -1-2 3 9 0.0 -32 8 14 9. S) 4 (2.7) PCS: LOWi Decreas;' frm basIi ne >=7 %7 Ni 1 G+: Increase from basaI linn >= 7 % Last post- Pa o ti no abservat ion 10214 ST WB02 09UN206:15:56:05 2618/313 - TFL/AD Build Number3s - Final WO 2008/025781 PCT/EP2007/058958 91 Time to Deterioration CLINICAL STUDY FOUR: The primary efficacy variable was the time to deterioration and the analysis was based on the FAS. The primary efficacy analysis rejected the hypothesis of equal time to 5 deterioration of schizophrenia in the three treatment groups (p = 0.008). The proportion of patients who deteriorated was 59% in the PBO group, 41% in the BX20 group, and 38% in the BX30 group. The Cox proportional hazards model gave an estimated hazard ratio of 0.66 (BX20) and 0.65 (BX30) relative to PBO; that is, the risk of deterioration was approximately 1.5 times higher for patients in the PBO group than for patients in the 10 BX20 or BX30 groups. Subsequent pairwise comparisons of each of the BX groups and the PBO group showed that patients in the BX groups had a statistically significantly longer time to deterioration of schizophrenia than patients in the PBO group (BX20: p = 0.008 and BX30: p = 0.006). The primary efficacy analysis was repeated for the PPS. Since most of 15 the patients in the PPS participated for most of the study, the results were very close to the results of the primary analysis, both for the estimated hazard ratios and the p-values obtained. This illustrates the robustness of the conclusion of the primary efficacy analysis. Panel 25 Log-iauk Test Analysing Time to Deterioration - OveraRl Test (FAS) Hazard Treatment N oa D Rat io N Deter iorated Deter i 0rat~dRai P-va ne PEG 16 98 5 BX20 '15 64 40. 5 0,656 BX30 172 &G 384 0 653 All 496 22: 46-0 .. 60. a 10214 ET RELAP 01 Cox 09JUN2006:15:38:0; 2615/313 - TFL/SAD BuiId :Nribers - Final 20 Lipid Profile CLINICAL STUDY FOUR: The adjusted mean HDL cholesterol values increased from baseline to Month 6 in all three treatment groups irrespective of fasting/non-fasting condition (PBO: 0.04/0.06 (fasting/nonfasting); BX20: 0.07/0.08; BX30: 0.07/0.08mmol/L). There were no statistically significant differences between 25 either of the BX groups and the PBO group. The adjusted mean triglycerides values decreased from baseline to Month 6 in all three treatment groups irrespective of fasting/non-fasting condition (PBO: -0.06/-0.22 (fasting/non-fasting); BX20: -0.16/-0.21; WO 2008/025781 PCT/EP2007/058958 92 BX30: -0.37/-0.03mmol/L). There were no statistically significant differences between either of the BX groups (except BX30 (fasting)) and the PBO group. The adjusted mean fasting glucose values increased from baseline to Month 6 in all three treatment groups (PBO: 0.10; BX20: 0.13; BX30: 0.09mmol/L) and there were 5 no statistically significant differences between either of the BX groups and the PBO group. Panel 50 below summarizes the lipid profile at baseline (mean values) and at Month 6 (mean change from baseline) in fasting and non-fasting patients. The meantotal cholesterol and mean LDL calculated decreased from baseline to Month 6 in 10 all groups (except non-fasting PBO patients) irrespective of treatment and fasting/non fasting condition. The mean VLDL calculated and mean triglycerides decreased from baseline to Month 6 in all groups (except non-fasting BX30 patients) irrespective of treatment and fasting/non-fasting condition. The mean HDL increased from baseline to Month 6 in all groups irrespective of treatment and fasting/nonfasting condition.

WO 2008/025781 PCT/EP2007/058958 93 Panel 50 Lipid Profile Time poinr RBO BX2-0 BX30 3 Co nc, n C ,C Fasjtina Bas-eline 91D 4.96; 91 4.S2 88 4. 87 Choleneroi.lenda ~ ~ ~ - A M4e= chanize" 27 -0.0 31 -l 37 0.2 ( oL) Nnftng Bas ,ehn 73 47 7 4. 69 827 5.0 M, ca-e0 1014 20 -O 10 Fasting Basetnie- V9 27 7 .8 5 29

LDL

h M2ean chanlzez 27 -09 3 -0.2 1 37 .1 Calculated (mmnl/) tNo-f ng. Baselne 2 2 82 76 275 0 06 Mean change" 17 105 19 -035 2 0.2 1 FastiB n 90 071 81 .72 88 071 VLDL " Mencane 7 -12 3 0 7 -1 Calculated Non-fasting Ba etn7e .7 .73 0.69 Nean change 1 -6 19 0 19 2 0.0Z Fasting Baseine 90 1 27 1 188 LIS HDL, direct Nfea Change" 279 387 0.0 (mmolL) Non-fasting Ba7-n3e 7 2 2 6 MaIn change 1 19 0.12 20 0 1 Fatia asine 90 L51:6 81 L6'7 8 6 Triglycerides Mean change' 27 -005 38L -ODS 37 -A (mmol"/L) Nan-fasang Ba-ehvne 73 L.M6 77 1.64 L2 L53 Man ch an'me. 17 7 1 14 0 00 \fe~i~~e 1 9 -0.4! A!H vanau are mean vah-mn. a Mean change from baselne to MN1h 6 b See ano Thle 197 for LDL direct (measwed a ~aseinIe in 2 patients (PBO: 2; BX2: 5; PX: 5) and at Hyperglycemia and Diabetes Mellitus Hyperglycemia, in some cases extreme and associated with ketoacidosis or hyperosmolar coma or death, has been reported in patients treated with atypical 5 antipsychotics. The incidence of hyperglycemia and diabetes-related adverse events (such as hyperglycemia, elevated blood glucose, impaired glucose tolerance, diabetes mellitus, inadequately controlled diabetes) in patients treated with bifeprunox was 0.5% (5/1050) and in placebo-treated patients was 0.6% (3/469) in six-week placebo controlled trials. In a 26-week placebo-controlled trial, no patients reported 10 hyperglycemia or diabetes-related adverse events. Assessment of the relationship between atypical antipsychotic use and glucose abnormalities is complicated by the possibility of an increased background risk of diabetes mellitus in patients with schizophrenia and the increasing incidence of diabetes mellitus in the general WO 2008/025781 PCT/EP2007/058958 94 population. Given these confounders, the relationship between atypical antipsychotic use and hyperglycemia-related adverse events is not completely understood. However, epidemiological studies that did not include bifeprunox suggest an increased risk of treatment-emergent hyperglycemia-related adverse events in patients treated with the 5 atypical antipsychotics included in these studies. Because bifeprunox was not marketed at the time these studies were performed, it is not known if bifeprunox is associated with this increased risk. Precise risk estimates for hyperglycemia-related adverse events in patients treated with atypical antipsychotics are not available. 10 PR, QT, QTc, QRS CLINICAL STUDY ONE: Changes in PR, QT, QTc, QRS intervals and heart rate over time were evaluated by assessing changes in mean values between Baseline and Endpoint. There was very little change in mean values for PR, QTc, or QRS intervals between Baseline and Endpoint in any treatment group. Mean changes for these 15 intervals ranged from approximately -2 msec to 4 msec. There were no trends in mean changes by treatment group. Changes in mean heart rate between Baseline and Endpoint ranged from -1.5 bpm to 0.9 bpm. There were no trends in mean changes by treatment group. 20 CLINICAL STUDY TWO: Changes in PR, QT, QTc, QRS intervals and heart rate over time were evaluated by assessing changes in mean values between Baseline and Endpoint. There was very little change in mean values for PR, QTc, or QRS intervals between Baseline and Endpoint in any treatment group. Ranges for the mean changes across the treatment groups were: PR=-2.5 msec to 0.2 msec; QT= -2.8 msec to 7 25 msec, QTc=0.4 msec to 3.9 msec; QRS=-0.4 msec to 1.4 msec. There were no trends in mean changes by treatment group. Changes in mean heart rate between Baseline and Endpoint across the treatment groups ranged approximately from -1.3 bpm to 1.7 bpm. There were no trends in mean changes by treatment group. CLINICAL STUDY THREE: Changes in PR, QT, QTc, QRS intervals and heart rate 30 over time were evaluated by assessing changes in mean values between Baseline and Endpoint. For these summaries, the Bazett corrected QT interval was presented as QTc. There was very little change in mean values for PR, QTc, or QRS intervals between Baseline and Endpoint in any treatment group. Mean changes for these intervals ranged from approximately 3 msec to 4 msec. There were no trends in mean changes by WO 2008/025781 PCT/EP2007/058958 95 treatment group. Changes in mean heart rate between Baseline and Endpoint ranged from -1 bpm to 1 bpm. There were no trends in mean changes by treatment group. 5 EXAMPLE 3a: PHARMACOKINETICS OF BIFEPRUNOX The objective of this study is to assess the pharmacokinetics (PK) of bifeprunox. The PK of bifeprunox in healthy subjects were investigated based on a pooled analysis of PK parameters from 21 clinical pharmacology studies. The pooled analysis included PK profiles after single and multiple doses to 132 and 399 subjects, respectively, and 10 explored the potential effects of age, gender, body weight, and race. In addition, PK in patients with schizophrenia were investigated using a population PK approach based on samples from 376 patients in phase Il studies and 434 patients in phase Ill studies. Bifeprunox was rapidly absorbed after oral administration (tmax from 1.5 to 2 hours at all dose levels). Bifeprunox multiple-dose PK were dose-proportional in the 20-40 mg/day 15 range. Steady-state mean apparent clearance and apparent volume of distribution were 62.2L/h and 1300L, respectively. Bifeprunox was eliminated with a mean plasma steady-state half-life of 14.4 hours. Administration of a 40 mg dose with a standard high fat meal was associated with a slight delay in tmax (1.5 hours) and a small increase in Cmax (10%) and AUC (29%). Bifeprunox is approximately 99% bound to serum proteins. 20 Bifeprunox is metabolized by CYP2C9, CYP3A4 and to a lesser extent, CYP2D6. Bifeprunox exposure was increased by co-administration with fluconazole (CYP2C9 inhibitor) and to a minor extent ketoconazole (CYP3A4 inhibitor), but not by coadministration with paroxetine (CYP2D6 inhibitor) and famotidine (a H2-antagonist). Bifeprunox exposure was reduced by co-administration of carbamazepine (CYP3A4 25 inducer). Co-administration of the narrow therapeutic index compounds warfarin and lithium (see EXAMPLE 3b) with bifeprunox did not affect the PK of these compounds to any relevant extent. In CYP2C9 slow/intermediate metabolizers, higher plasma levels of bifeprunox were observed than in subjects with normal enzyme activity. After a single oral dose of [14C]-labeled bifeprunox, 13% and 74% of the radioactivity was excreted in 30 the urine and feces, respectively. No clinically significant age-, gender-, body weight- or race-related effects on bifeprunox PK were noted. PK in patients with schizophrenia were similar to that seen in healthy subjects.

WO 2008/025781 PCT/EP2007/058958 96 One conclusion of this study is that bifeprunox is rapidly absorbed after oral administration; the mean elimination half-life is about 14 hours. Multiple-dose PK were dose-proportional in the 20-40 mg range. Bifeprunox has a low interaction potential. 5 EXAMPLE 3b: PHARMACOKINETIC INTERACTION OF LITHIUM AND BIFEPRUNOX IN HEALTHY MALE SUBJECTS PURPOSE: Bifeprunox, a partial agonist for dopamine D2 and 5-HT1A receptors, is being developed for the treatment of schizophrenia. Because bifeprunox may be used in 10 combination with the mood-stabilizing drug lithium for the treatment of patients with psychoses and mood disorders, the effect of multiple doses of bifeprunox on the pharmacokinetic (PK) profile of lithium was evaluated. Lithium has a narrow therapeutic index that can complicate therapy, and serum levels greater than 1.5 mmol/L carry a greater risk of lithium toxicity than do lower levels. METHODS: This was a single center, 15 double-blind, randomized, placebo-controlled, parallel-design study in 48 healthy male subjects. All subjects were to receive open-label lithium (450 mg) twice daily on days 1 through 8 and, provided serum levels were stable on days 5 through 7, again on days 9 through 20. Subjects whose serum levels were stable on days 5 through 7 were included in those randomized to receive, in addition to lithium 450 mg twice daily, either placebo 20 or a rising dose of bifeprunox (0.025-40 mg) once daily on days 9 through 17, and placebo or bifeprunox 40 mg on days 18 through 21. Only a single morning dose of lithium 450 mg was administered on day 21. Lithium, steady-state Cmax, AUC over the dosing interval (0-t), and renal clearance (CL-R) values were compared between the bifeprunox and placebo groups using ANCOVA with baseline values of lithium measured 25 on day 8 as covariate. RESULTS: There were small increases in mean Cmax and AUC(0-t) of lithium in the bifeprunox group, but the ratios of the geometric least square means and the 90% confidence intervals (Cl) of the two treatments for AUC, Cmax, and CL-R were within the predefined range of 0.80 to 1.25. The treatment ratios and 90% Cl for Cmax, AUC(0-t), and CL-R were 1.11 (90% Cl: 1.01-1.20), 1.13 (90% Cl: 1.06-1.21), 30 and 0.94 (90% Cl: 0.87-1.02) respectively. Combined administration of bifeprunox up to 40 mg per day and lithium 450 mg twice daily was well tolerated. CONCLUSION: There was no clinically relevant effect of co-administration of multiple doses of bifeprunox on lithium steady-state pharmacokinetics. Results suggest that no dosage adjustment of lithium would be required during concomitant administration with bifeprunox.

WO 2008/025781 PCT/EP2007/058958 97 EXAMPLE 4: Efficacy and Safety of Bifeprunox in Treatment of Patients with Acutely Exacerbated Schizophrenia Objective: To evaluate the efficacy and safety of bifeprunox in the treatment of acutely 5 ill patients with schizophrenia. Method: A 6-Week randomized, placebo-controlled, risperidone-referenced dose-finding study included 589 randomized patients with an acute exacerbation of schizophrenia (DSM-IV-TR). Patients were randomly assigned to bifeprunox 5 mg (n= 115), bifeprunox 10 mg (n=120), bifeprunox 20 mg (n=1 15), placebo (n=1 19) or risperidone 6 mg 10 (n=120). Treatment with all bifeprunox doses were titrated up to target dose, beginning with a dose of 0.125 mg on day 1, 0.25 mg on day 2, 0.5 mg on day 3, 1 mg on day 4, 2 mg on day 5, and 5 mg on day 6, 10 mg on day 7 or 20 mg on day 8, while treatment with risperidone was titrated over 3 days. The change in the Positive and Negative Symptom Scale (PANSS) total score, from baseline to endpoint, was the primary 15 outcome measure. Secondary efficacy measures included: PANSS positive, PANSS negative, PANSS general psychopathology (GPP) score, PANSS-derived Brief Psychiatric Rating Scale (BPRS) score, Clinical Global Impressions-Severity of Illness (CGI-S), CGI-Improvement (CGI-1) scores, and responder rates. Safety and tolerability evaluations included extrapyramidal symptoms (EPS), weight gain, lipid profile, and 20 serum prolactin. Risperidone was included for assay sensitivity. Results: Reduction in the PANSS total score for bifeprunox was statistically significant (P < 0.05) compared to placebo at Week 3 and Week 6/endpoint for the 20 mg dose. The positive effect of bifeprunox 20 mg was also observed on the secondary efficacy measures PANSS positive, negative, GPP subscales, BPRS, and responder rates. 25 Risperidone 6 mg was statistically significant at endpoint compared to placebo for all efficacy measures. The most common adverse events (incidence >5% and twice for placebo) included: dyspepsia, nausea, vomiting, and constipation. A dose relationship was not evident for any of the most frequent adverse events. Bifeprunox was associated with decreased prolactin levels, and rates of EPS that were comparable 30 to placebo. In addition, patients receiving bifeprunox experienced statistically significant (P < 0.05) weight decrease, and demonstrated statistically significant improvements in non-fasting triglycerides (P < 0.005) and total cholesterol (P < 0.005) compared to placebo.

WO 2008/025781 PCT/EP2007/058958 98 Conclusion: In this study, bifeprunox 20 mg was shown to be effective in the treatment of acute schizophrenia. Bifeprunox may have safety advantages, stemming from a decrease in weight, and improvement in the lipid profile. 5

Claims (45)

1. A daily dose of bifeprunox for the treatment of a patient with a CNS disorder, wherein the dose is 20-30 mg of at least one bifeprunox compound.

2. The dose of claim 1 for the treatment of a patient with schizophrenia. 5

3. The dose of claim 1 or 2 wherein the dose is a once-daily dose.

4. The dose of any one of claims 1 -3 for long-term treatment of a patient with schizophrenia.

5. The dose of any one of claims 1 -4 for maintaining clinical stability in a patient with schizophrenia. 10

6. The dose of claim 5, wherein the dose is 20 mg of at least one bifeprunox compound.

7. The dose of claim 5, wherein the dose is 30 mg of at least one bifeprunox compound.

8. The dose of any one of claims 1 - 3 for the treatment of a patient with 15 acutely exacerbated schizophrenia.

9. The dose of claim 8, wherein the dose is 20 mg of at least one bifeprunox compound.

10. The dose of any one of claims 1 - 9 wherein the bifeprunox compound is bifeprunox mesylate. 20

11. The dose of claim 10, wherein the bifeprunox compound is the alpha polymorph of bifeprunox mesylate.

12. A pharmaceutical composition comprising 30 mg of at least one bifeprunox compound and at least one pharmaceutically acceptable carrier for use in the treatment of a patient with schizophrenia. 25

13. A pharmaceutical composition comprising 20 mg of at least one bifeprunox compound and at least one pharmaceutically acceptable carrier for use in the treatment of a patient with schizophrenia.

14. The pharmaceutical composition of claim 12 or 13 for use in long-term treatment of a patient with schizophrenia. WO 2008/025781 PCT/EP2007/058958 100

15. The pharmaceutical composition of any one of claims 12 - 14 for use for maintaining clinical stability in a patient with schizophrenia.

16. The pharmaceutical composition of claims 12 or 13 for use for the treatment of a patient with acutely exacerbated schizophrenia. 5

17. The pharmaceutical composition of any one of claims 12 -16, wherein the bifeprunox compound is bifeprunox mesylate.

18. The pharmaceutical composition of claim 17, wherein the bifeprunox compound is the alpha polymorph of bifeprunox mesylate.

19. Bifeprunox for use in the long-term treatment of a patient with 10 schizophrenia.

20. Bifeprunox for use according to claim 19 in the treatment of a patient with stable schizophrenia.

21. Bifeprunox for use in the treatment of a patient with stable schizophrenia.

22. Bifeprunox for the use of any one of claims 19 - 21, wherein a daily dose 15 of 20-30 mg of at least one bifeprunox compound is administered to the patient.

23. Bifeprunox for the use of claim 22, wherein a daily dose of 20 mg of at least one bifeprunox compound is administered to the patient.

24. Bifeprunox for the use of claim 22, wherein a daily dose of 30 mg of at least one bifeprunox compound is administered to the patient. 20

25. Bifeprunox for use in the treatment of a patient with acutely exacerbated schizophrenia.

26. Bifeprunox for the use of claim 25, wherein a daily dose of 20 mg of at least one bifeprunox compound is administered to the patient.

27. Bifeprunox for the use of claim 25, wherein a daily dose of 30 mg of at 25 least one bifeprunox compound is administered to the patient.

28. Bifeprunox for the use of any one of claims 19-27 wherein at least one bifeprunox compound is administered in combination with the mood-stabilizing drug lithium. WO 2008/025781 PCT/EP2007/058958 101

29. Bifeprunox for use in the treatment of patients with psychoses and mood disorders wherein at least one bifeprunox compound is administered in combination with the mood-stabilizing drug lithium.

30. Bifeprunox for use of any one of claims 19-27 wherein at least one 5 bifeprunox compound is administered in combination with an antidepressant.

31. Bifeprunox for use of claim 30 wherein the antidepressant is paroxetine.

32. Bifeprunox of any one of claims 19 - 31, wherein the bifeprunox compound is bifeprunox mesylate.

33. Bifeprunox of claim 32, wherein the bifeprunox compound is the alpha 10 polymorph of bifeprunox mesylate.

34. A kit for the treatment of a patient with psychoses and mood disorders comprising a composition comprising at least one bifeprunox compound and a composition comprising the mood-stabilizing drug lithium.

35. A kit for the treatment of a patient with a CNS disorder comprising a 15 composition comprising at least one bifeprunox compound and a composition comprising an antidepressant.

36. The kit of claim 35 wherein the antidepressant is paroxetine.

37. The kit of any one of claims 34 - 36, wherein the bifeprunox compound is bifeprunox mesylate. 20

38. The kit of claim 37, wherein the bifeprunox compound is the alpha polymorph of bifeprunox mesylate.

39. Use of bifeprunox for the preparation of a medicament for the treatment of a patient with schizophrenia, wherein a daily dose of 20-30 mg of at least one bifeprunox compound is administered to the patient and said treatment results in a favorable 25 metabolic profile.

40. Use of claim 40, wherein the treatment with bifeprunox of a patient with schizophrenia avoids and/or reduces side effects selected from weight gain, disorders of WO 2008/025781 PCT/EP2007/058958 102 triglyceride levels and/or total cholesterol levels, the incidence of hyperglycemia and/or one or more diabetes-related adverse events.

41. Use of bifeprunox for the preparation of a medicament for the treatment of a patient with stable schizophrenia, wherein a daily dose of 20-30 mg of at least one 5 bifeprunox compound is administered to the patient.

42. Use of bifeprunox for the preparation of a medicament for the treatment of a patient with acutely exacerbated schizophrenia, wherein a daily dose of 20-30 mg of at least one bifeprunox compound is administered to the patient.

43. The use of any one of claims 39 - 42, wherein the bifeprunox compound 10 is bifeprunox mesylate.

44. The use of claim 43, wherein the bifeprunox compound is the alpha polymorph of bifeprunox mesylate.

45. Bifeprunox for use in the treatment of a patient with schizophrenia having weight problems or susceptible to weight problems. 15