FDA WARNING LETTERS Protocol-specified PK sampling PK sampling Actual time of time according schedule (post PK sampling to protocol infusion

(2400 hours) (2400 hours) initiation) Subject 005 Subject 005 Day Subject 005 Day Day 1 Study 1 Study Drug 1 Study Drug Drug Infusion Infusion Time: Infusion Time: Time: 12451245-1316 1245-1316 1316 Just prior to dose Just prior to dose 1220 15 minutes (post 1300 1335 infusion initiation) 30 minutes (post 1315 1347 infusion initiation) 45 minutes (post 1330 1403 infusion initiation) 1 hour (post 1345 1418 infusion initiation) 2 hours (post 1445 1518 infusion initiation)

In your September 14, 2009, written response, you indicated that your site misunderstood the timing of the PK samples. You noted that your site interpreted the protocol-required PK sampling to begin at the end of the drug infusion time rather than at the start of the drug infusion time. You stated that your site will pay particular attention to the protocol requirement for PK sampling in the future. We acknowledge your response. However, we are concerned that the response is not adequate to prevent future recurrence of the violation noted above. In particular, we are concerned that you did not properly understand the protocol specified PK sampling schedule from the start of the study, and that you did not ensure that the PK samples were collected according to the protocol throughout the study. Your failure to collect PK samples as specified in the protocol significantly undermines the reliability and integrity of the data captured at your site. Regarding Protocols (b)(4) and (b)(4) 2. You failed to maintain adequate and accurate case histories that record all observations and other data pertinent to the investigation [21 CFR 312.62(b)]. The inspection revealed that there were numerous adverse events recorded in progress

FDA WARNING LETTERS

notes of subjects' records that had not been reported on the case report forms (CRFs), and there was no explanation regarding this discrepancy. The following are examples of adverse events that had not been reported on the electronic CRFs at the time of the inspection. Note that since the inspection occurred in August 2009, the examples below represent delays in transcribing adverse events from progress notes to electronic CRFs that range from at least 8 months to 15 months. a. Subject (b)(4) 071 003: Progress notes documented weakness of upper and lower extremities on May 7, 2008. b. Subject (b)(4) 071 005: Progress notes documented nausea and vomiting on July 2, 2008. c. Subject (b)(4) 071 008: Progress notes documented alopecia on August 14, 2008, as well as dark nails on October 28, 2008. d. Subject (b)(4) 071 012: Progress notes documented difficulty swallowing food on November 3, 2008, as well as nausea and vomiting on December 31, 2008. e. Subject (b)(4) 071 001: Progress notes documented decreased breath sounds with wheezing on September 10, 2008, as well as nausea and vomiting on December 8, 2008. f. Subject (b)(4) 071 004: Progress notes documented difficulty reading on January 5, 2009. g. Subject (b)(4) 071 005: Physical examination notes documented a small lump on the left posterior scalp on January 12, 2009; progress notes documented alopecia on January 28, 2009. h. Subject (b)(4) 071 006: Progress notes documented difficulty sleeping on December 10, 2008, due to left thigh pain. In your September 14, 2009 written response, you stated that your site staff enters all adverse events and concomitant medications in subjects' charts at the time of each visit. You indicated that at the start of the study, your site staff members were entering data (adverse events and concomitant medications) in CRFs prior to the

FDA WARNING LETTERS

sponsor's monitoring visits. However, you indicated that your site staff stopped entering data in CRFs in advance of the monitoring visits so that they could enter data with the monitors present, to avoid additional system queries. As a corrective measure, you noted that your site staff members have implemented a policy requiring that all CRFs be completed in a timely manner, defined as 30 days following a study visit, regardless of scheduled monitoring visits. We acknowledge your response. However, the lack of timely adverse event information in the electronic CRFs may have jeopardized subject safety as well as the reliability and integrity of the data captured on the CRFs at your site. We find your corrective measures to prevent future recurrence to be acceptable, if implemented as proposed. This letter is not intended to be an all-inclusive list of deficiencies with your clinical study of an investigational drug. It is your responsibility to ensure adherence to each requirement of the law and relevant FDA regulations. You should address these deficiencies and establish procedures to ensure that any ongoing or future studies will be in compliance with FDA regulations. Within fifteen (15) working days of your receipt of this letter, you should notify this office in writing of the actions you have taken to prevent similar violations in the future. Failure to adequately and promptly explain the violations noted above may result in regulatory action without further notice. 1. You failed to maintain adequate and accurate case histories that record all observations and other data pertinent to the investigation on each individual administered the investigational drug or employed as a control in the investigation [21 CFR 312.62(b)]. Discrepancies in records were found in all six subjects enrolled in Protocol (b)(4), raising significant questions about the reliability of data at your site. a. For Subject 34002, the following deficiencies were noted: i. On January 20, 2005, this subject was enrolled in the study and was administered the study drug. You submitted a protocol deviation/violation report to the Institutional Review Board (IRB) on May 2, 2005 stating that this subject was enrolled without having signed an Informed Consent Form (ICF). The subjects records contained two copies of the ICF. One was signed on May 10, 2005, and another copy was an altered version which reflected a signature date of January 10, 2005. You stated in your affidavit obtained by the field investigator during the FDA inspection that you believe that the ICF

FDA WARNING LETTERS with a signature date of January 10, 2005 is an altered document. You stated that you did not alter the document and that you do not know who altered it.

ii. The electronic Case Report Form (eCRF) for the Baseline Visit states that the subject signed the ICF on January 20, 2005. However, according to the Baseline Visit worksheet, the ICF was signed on September 18, 2003. The subjects records did not contain an ICF dated September 18, 2003 or January 20, 2005. In addition, as stated above under 1(a)(i), the subjects records contained two copies of the ICF, one copy was signed on May 10, 2005 and the other was altered to indicate it was signed on January 10, 2005. Thus, it appears that the subject signed the ICF on May 10, 2005, which was not accurately recorded in the Baseline Visit eCFR or worksheet. iii. The eCRF for the 12-Month Visit states that pre-bronchodilator spirometry was conducted on January 17, 2005 and that post-bronchodilator spirometry was conducted on January 17, 2006. However, the 12-Month Visit worksheet states that both of these spirometries were conducted on July 17, 2006. We note that your office manager stated during the inspection that this worksheet was mistakenly documented as the 24-month visit, but was in fact the 12-Month Visit worksheet. iv. Pages 1 through 5 were missing from the ICFs dated May 10, 2005 and January 10, 2005. b. For Subject 60060, the following deficiency was noted: The February 9, 2006 30-Month Visit worksheet contains a check mark in the box labeled Yes for the question, Has the patient had any of the following during the past 6 months? These four choices follow the question: non-serious malignancy, serious malignancy, serious asthma exacerbation, and other serious adverse event. The same page of the worksheet indicates that if the answer to this question is yes, an adverse event (AE) worksheet packet is to be completed for each AE. However, no completed AE worksheet packet was in the subjects records. c. For Subject 34004, the following deficiencies were noted: i. The AE section of the 6-Month Visit worksheet contained conflicting information about whether the subject suffered an AE or not. On one page, it states that the patient did not have a non-serious malignancy, serious malignancy, serious asthma exacerbation, or other serious adverse event. However, on another page, in response to the question What type of adverse event was this? the non-serious malignancy box is checked. ii. Pages 1 through 8 were missing from the ICF.

FDA WARNING LETTERS d. For Subject 60667, the 12-Month Visit worksheet did not document the subjects response regarding whether an adverse event had been experienced in the previous six months. e. For Subject 62694, the following deficiencies were noted: i. The eCRF for the 12-Month Visit states that spirometry was conducted on June 26, 2007. However, the 12-Month Visit worksheet states that the spirometry was conducted on July 26, 2007. ii. Page 9 was missing from the ICF.

2. You failed to obtain informed consent in accordance with the provisions of 21 CFR part 50 [21 CFR 312.60]. a. Subject 34002 was enrolled in the study before you obtained the subjects legally effective informed consent, in violation of 21 CFR 50.20. As explained above under 1(a)(i), on January 20, 2005, Subject 34002 was enrolled in the study and was administered the study drug. You submitted a protocol deviation/violation report to the IRB on May 2, 2005 stating that this subject was enrolled without having signed an informed consent form. The subjects records contained two copies of the informed consent form. One was signed on May 10, 2005, and another copy was an altered version which reflected a signature date of January 10, 2005. You stated in your affidavit obtained by the field investigator during the FDA inspection that you believe that the ICF with a signature date of January 10, 2005 is an altered document. You stated that you did not alter the document and that you do not know who altered it. Thus, available documentation indicates that you did not obtain the subjects informed consent until May 10, 2005, months after the subject was enrolled in the study. b. For three subjects, you failed to document informed consent by the use of a written consent form approved by the IRB, as required by 21 CFR 50.27. Specifically, ICFs, signed by the following subjects on the indicated dates, were sample ICFs, which were not approved by the IRB: a. Subject 34002, signed on February 7, 2007 b. Subject 34004, signed on May 29, 2006 c. Subject 60060, signed on October 12, 2006 3. You failed to promptly report to the IRB all changes in the research activity [21 CFR 312.66]. The following Protocol Amendments were not submitted to the IRB for review and approval:

FDA WARNING LETTERS a. Amendment A3 dated October 15, 2007, which contained additional information regarding expert oncology review of all malignancy adverse events. b. Amendment A4 dated April 14, 2008, which increased the frequency of physical examination, review of systems, review of laboratory results, and review of radiologic findings from only at screening and baseline to every 6 months. 4. You failed to ensure that the investigation was conducted according to the signed investigator statement and the investigational plan [21 CFR 312.60].

a. The protocol specified that clinical examinations were to be performed by the investigator (you), and you signed the Form FDA 1572, Investigator Statement, indicating that you were the investigator and that there were no sub-investigators. However, the clinical examinations for the following visits were performed by (b) (4), M.D: i. 12-Month Visit for Subject 62694 ii. 18-Month Visit for Subject 34004 iii. 30-Month Visit for Subject 34004 iv. 18-Month Visit for Subject 60667 v. 12-Month Visit for Subject 60060 b. The protocol specified that subjects who miss any visit were to be contacted by telephone. There is no documentation that you attempted to contact Subject 60667 after the subject missed the 6-Month visit. c. The protocol specified that spirometry was to be performed at the baseline visit. You did not perform spirometry at the baseline visit for the following subjects: i. Subject 34002: Baseline visit on January 20, 2005; spirometry on August 23, 2003 ii. Subject 34004: Baseline visit on May 12, 2005; spirometry on July 21, 2005 iii. Subject 60060: Baseline visit on August 8, 2005; spirometry on June 24, 2004 5. You failed to maintain adequate records of the disposition of the drug, including dates, quantity, and use by subjects [21 CFR 312.62(a)]. You did not maintain records of study drug disposition for Subjects 62694 and 34004.

FDA WARNING LETTERS

6. You failed to retain records required to be maintained under 21 CFR part 312 until 2 years after the investigation was discontinued and FDA was notified [21 CFR 312.62(c)]. By a letter dated June 13, 2008, the sponsor notified you that your participation in Protocol (b)(4) was terminated, thereby discontinuing your investigation. The sponsor also notified FDA, by a letter dated June 13, 2008, that your participation in Protocol (b)(4) was terminated. . The protocol eligibility criteria state that a subject must have a primary diagnosis of Bipolar I Disorder as defined by the DSM-IV criteria and confirmed by the [Kiddie-Schedule for Affective Disorders and Schizophrenia (K-SADS)]. Subject 1001 was screened and randomized to the study without a primary diagnosis of Bipolar I disorder confirmed by a fully-completed K-SADS assessment. In particular, the K-SADS assessment for Subject 1001 did not contain the indicated Depressive Disorders and the Mania supplements. ii. The protocol eligibility criteria state that a subject must have a Young Mania Rating Scale (YMRS) score greater than or equal to 17 at the screening and baseline visits. Subjects 1009 and 1010 were each randomized to the study with a YMRS score of 8 at baseline. b. The protocol states that the first dose of the study medication is taken at the baseline visit after completion of baseline visit procedures to confirm subject eligibility and to perform post-dosing pharmacokinetic (PK) sampling. Subject 1010 was dosed on December 8, 2006, three days prior to the baseline visit on December 11, 2006 and as such did not have eligibility confirmed and did not have PK sampling performed. c. The protocol requires that efficacy be assessed using the YMRS, the Childrens Global Assessment Scale (CGAS), and Clinical Global Impression (CGI) scale at each visit. The CGI assessment was not performed at the Week 1 study visit for Subject 1010. d. You failed to follow protocols related to the enrollment of wards of the state. The Institutional Review Board (IRB) Initial Review Submission Form Box 28 in this study addressed vulnerable subject categories. You failed to indicate that you intended to enroll Subjects 1001 and 1007, both wards of the state and thus vulnerable subjects, to the IRB on the Initial Review Submission Form. In turn, the IRB was prevented from assessing whether it was appropriate to enroll wards of the State in this particular clinical investigation and was prevented from requiring that an advocate be appointed for each child who is a ward pursuant to 21 CFR 50.56.

FDA WARNING LETTERS

Enrollment of subjects who do not meet eligibility criteria, and not performing studyrelated procedures jeopardize subject safety and welfare and compromise interpretation and validity of the investigational endpoints. 2. You failed to maintain adequate and accurate case histories that record all observations and other data pertinent to the investigation on each individual administered the investigational drug or employed as a control in the investigation [21 CFR 312.62(b)]. During the inspection, the study records (including study templates and assessment tools) were observed to be unbound and appeared to lack a system of controls, with entries posted to the wrong visit form and multiple versions completed for the same visit. Subject records contained missing pages, numerous unexplained corrections and conflicting information. Few progress notes were observed Violations included, but were not limited to, the following: a. The protocol eligibility criteria state that a subject must have a YMRS score greater than or equal to 17 at the screening and baseline visits. Study records for Subject 1006 contained two sets of original records with different scoring for the YMRS assessments reportedly obtained during the October 31, 2006 study visit. One of the copies of the YMRS for Subject 1006 dated October 31, 2006 has a score of 11 and the other copy has a score of 23. You provided no explanation for the discrepancies in these records. b. The K-SADS-Present and Lifetime Version is a multi-part diagnostic interview instrument that consists of a screening interview and five additional diagnostic supplements. The additional diagnostic supplements may be indicated based upon scores obtained as determined by the answers given by the subject during the screening interview. The KSADS-PL for Subjects 1001, 1004 and 1006 were missing every other page and/or the required supplements. c. The records for Subject 1004 contained numerous errors in subject number, protocol number, and date or identity of study visit. For example: i. Numerous records reflect that subject number 1002 was used instead of the correct subject number 1004. ii. Forms marked Screening Visit, Week 1 Visit, and Week 4 Visit all are dated October 31, 2006. iii. There are two forms marked Week 1 Visit documenting the Childrens Depression Rating Scale Revised (CDRS-R) evaluation with two different dates, October 31, 2006 and November 8, 2006. iv. You utilized a form from another study entitled Study to document testing completed on this subject for this study.

FDA WARNING LETTERS Failing to maintain adequate and accurate case histories compromises the interpretation of and the validity of the clinical investigational endpoints. 3. You failed to obtain informed consent of each subject in accordance with the provisions of 21 CFR Part 50 [21 CFR 312.60].

Except as provided in 21 CFR 50.23 and 50.24, no investigator may involve a human being as a subject in research unless the investigator has obtained the legally-effective informed consent of the subject or the subject's legally authorized representative [21 CFR 50.20]. As an investigator, it is your responsibility to obtain informed consent in accordance with 21 CFR Part 50. Except as provided in 21 CFR 56.109(c), informed consent shall be documented by the use of a written consent form approved by the IRB and signed and dated by the subject or the subject's legally authorized representative at the time of consent. A copy shall be given to the person signing the form [21 CFR 50.27(a)]. You failed to obtain legally-effective informed consent from Subject 1001 to whom you prescribed the investigational new drug, (b)(4). Specifically, the informed consent form for Subject 1001 in Protocol (b)(4) was signed only by a parent and not by a representative of the Mississippi State Department of Human Services. At the time of the clinical investigation, the child was in the legal custody of the Mississippi State Department of Human Services and thus only the Mississippi State Department of Human Services could serve as the childs legally-authorized representative and grant permission for the child to participate in the clinical investigation. Failing to obtain adequate informed consent jeopardizes the safety and welfare of enrolled subjects by denying them an opportunity to assess the risks and benefits of their participation in the clinical investigation. 4. You failed to promptly report to the IRB all changes in the research activity and you made changes in the research without IRB approval [21 CFR 312.66]. FDA regulations require that the clinical investigator shall assure that he will promptly report to the IRB all changes in the research activity and all unanticipated problems involving risk to human subjects or others, and that he or she will not make any changes in the research with IRB approval, except where necessary to eliminate apparent immediate hazards to human subjects [21 CFR 312.66]. You violated this requirement by administering the investigational new drug (b)(4), to Subject 1010 without obtaining IRB approval of the modified informed consent document. Specifically, the IRB-approved version of the informed consent document for Protocol (b)(4) was altered by hand to state that the subject would not receive payment for participation in accordance with the schedule listed on the form unless

FDA WARNING LETTERS treated as an outpatient, and this altered form was signed by the parent of Subject 1010.

10

Failure to promptly report changes in the research activity to the IRB and making changes in the research without IRB approval compromises the safety and welfare of subjects enrolled in the clinical investigation. a. Section 3 of Protocols (b)(4) specified that for subjects with a body weight less than 45 kg, the target dose of study drug was 60-80 mg/day; and for subjects with a body weight greater than or equal to 45 kg, the target dose of study drug was 120-160 mg/day. These protocols also specified that the dose level of 160 mg/day was not to be achieved before Day 8 of treatment. These study drug administration limits were not always followed. The study drug and matching placebo were supplied as 20-, 40-, 60-, and 80-mg capsules, with the appearance and size slightly different among the dosage strengths. All medications were packaged in AM (morning dosing) and PM (evening dosing) blister cards that supplied medication for seven days plus three extra days. Each blister card was marked with four different-colored columns labeled with the letters A, B, C, and D to distinguish the different potencies. Each column contained one potency (20-, 40-, 60-, and 80-mg), such that four capsule strengths were available for each dose. Per protocol, the nursing staff and subjects' parent(s) or guardian(s) were to select one capsule for the morning administration and one capsule for the afternoon administration, and leave the remaining capsules in the blister pack. Throughout the study and when dosing modifications were necessary, subjects were to receive clear specifications regarding which capsule was to be taken from the morning and afternoon blister cards. When changes to the original dosage plan were necessary, unscheduled visits were recommended so that subjects could be observed and so that changes to the dosage plan could be fully explained to the subjects and to his or her parent(s) or guardian(s). Regarding Protocol (b)(4): Six of the seven randomized subjects received dosages in excess of protocol-specified limits: i. Subject 1001, age 13 at the time of enrollment, was randomized on May 9, 2006, and discontinued on June 5, 2006, upon completion of the study. Subject 1001, with a documented weight of 46.8 kg, was overdosed on study medication for 20 consecutive days while participating in study (b)(4). Specifically, on May 16 and 17, 2006 (Treatment Days 8 and 9), this subject received 180 mg/day; on May 18 and 19, 2006 (Treatment Days 10 and 11), this subject received 240 mg/day; on May 20 and 21, 2006 (Treatment Days 12 and 13), this subject received 320 mg/day; on May 22, 2006 (Treatment Day 14), this subject received 400 mg/day; and from May 23 through June 4, 2006 (Treatment Days 15 through 27), this subject received 180 mg/day. This subject experienced sedation and dizziness during the study.

FDA WARNING LETTERS

11

ii. Subject 1003, age 15 at the time of enrollment, was randomized on June 1, 2006, and discontinued on June 27, 2006, upon completion of the study. Subject 1003, with a documented weight of 65 kg, was overdosed on study medication for 21 consecutive days while participating in study (b)(4) Specifically, on June 7, 2006 (Treatment Day 7), this subject received 180 mg/day, in excess of the protocol-specified maximum dose of 160 mg/day prior to Day 8 of treatment. On June 8 and 9, 2006 (Treatment Days 8 and 9), this subject received 240 mg/day. From June 10 through 26, 2006 (Treatment Days 10 through 26), this subject received 400 mg/day. On June 27, 2006 (Treatment Day 27), this subject received 280 mg/day. iii. Subject 1004, age 16 at the time of enrollment, was randomized on June 1, 2006, and discontinued on June 14, 2006, prior to study completion, due to lack of efficacy and sedation. Subject 1004 received doses in excess of the maximum target dose (160 mg/day) for 3 consecutive days while participating in Study (b) (4). Specifically, on June 9, 2006 (Treatment Day 9), this subject received 320 mg/day. On June 10 and 11, 2006 (Treatment Days 10 and 11), this subject received 240 mg/day. This subject experienced sedation during the study. iv. Subject 1005, age 15 at the time of enrollment, was randomized on June 14, 2006, and discontinued on July 18, 2006, prior to study completion, due to noncompliance. Subject 1005 received doses in excess of the maximum target dose (160 mg/day) for 16 consecutive days while participating in study (b)(4). Specifically, on June 21 and 22, 2006 (Treatment Days 8 and 9), this subject received 180 mg/day. From June 23 through 27, 2006 (Treatment Days 10 through 14), this subject received 240 mg/day. From June 28 through July 3, and July 5 through July 6, 2006 (Treatment Days 15 through 20 and Days 22 through 23), this subject received 400 mg/day. On July 4, 2006 (Treatment Day 21), this subject received 200 mg/day. v. Subject 1006, age 13 at the time of enrollment, was randomized on June 26, 2006, and discontinued on July 25, 2006, upon study completion. Subject 1006, with a documented weight less than 45 kg, was overdosed on study medication for a total of 7 consecutive days while participating in study (b)(4). Specifically, from July 11 through 17, 2006 (Treatment Days 16 through 22), this subject received 400 mg/day. vi. Subject 1007, age 10 at the time of enrollment, was randomized on June 27, 2006, and discontinued on July 25, 2006, upon study completion. Subject 1007, with a documented weight of 39.5 kg, was overdosed on study medication for 13 consecutive days while participating in study (b)(4) Specifically, from July 12 through 24, 2006 (Treatment Days 16 through 28), this subject received 120 mg/day. Regarding Protocol (b)(4): The only subject randomized in this study received dosages in excess of protocol-specified limits:

FDA WARNING LETTERS vii. Subject 1001, an adolescent, was randomized on May 1, 2006, and discontinued on May 23, 2006, prior to study completion, due to worsening auditory hallucinations that apparently caused the subject to lacerate her wrists. Subject 1001, with a documented weight of 66.5 kg, was overdosed on study medication for 17 consecutive days while participating in study (b)(4). Specifically, on May 5, 2006 (Treatment Day 5), this subject received 320 mg/day. From May 6 through 21, 2006 (Treatment Days 6 through 21), this subject received 400 mg/day.

12

In your March 13, 2009 response, you stated that you were first notified on July 26, 2006 by (b)(4) data management personnel that there were dosing errors in the conduct of the studies, stemming from your mistaken assumption that all capsules contained in the study drug blister pack were of the same 20 mg strength. Immediately upon becoming aware of this issue, you stated that you telephoned each subject's parent or guardian to inform him or her of this error, and inquired about any possible safety concerns or issues, of which there were none. You stated that during these calls, you instructed the parents and guardians to stop dispensing the currently prescribed dosage and begin the corrected dosage amount. A follow-up letter was also sent to the parents and guardians on August 4, 2006. You noted that the (b)(4) study monitor conducted a targeted review with you and your research staff on the correct dosing and dispensing of study medication on July 26, 2006. Additionally, on July 31, 2006, your research staff attended the (b)(4)-hosted Question and Answer (b) (4) Recruitment meeting that included training on the dosing and dispensing for all (b)(4) research sites conducting the above-referenced studies in which you were involved. As part of your corrective action plan, you made formal revisions to your Standard Operating Procedures (SOPs). Specifically, you indicated that the Dosing and Dispensing of Investigational Product SOP was revised to add a more detailed inpatient and outpatient dosing process and to further specify how the clinical investigator, or designee, should document dosing instructions in both milligrams and tablets/capsules, as well as clearly document original dosing schematics and any titration throughout the study. You also noted that the Study Launch Procedures SOPs were revised to: (1) emphasize the materials reviewed and received at the investigator meetings, (2) include the review of the applicable Institutional Review Board (IRB) manuals for investigators, and (3) include internal documentation of the dosing and procedures to be performed on a mock subject from screening through study completion. Additionally, you stated that the Quality Assurance/Quality Control (QA/QC) Program SOP was also revised to ensure the oversight of two subjects, or 10% of the population randomized, with ongoing review of dosing and dispensing, protocol compliance, prior/concomitant medications, and prohibited medications. We acknowledge your response. However, we are concerned that the response is not adequate to prevent future recurrence of the violation noted above. In particular, we are concerned that you did not properly understand the study drug packaging from the

FDA WARNING LETTERS start of the study and you did not ensure that the appropriate dose was administered throughout the study. We are aware that the sponsor subsequently changed the packaging for the investigational product following observations that clinical investigators were not properly dosing subjects. Notwithstanding the packaging concerns, as clinical investigator, you retain responsibility for ensuring that the protocol is followed, including ensuring that the proper dose is administered to subjects.

13

b. Section 5.3.3 of Protocols (b)(4) and (b)(4) specified that you were to have developed a titration plan to increase the subject's total daily dose of study medication from a 20 mg starting dose (at the baseline visit) to a target dose achieved over 2 weeks, in general. In developing the titration plan, you were to have considered the subjects psychiatric history, current psychiatric status, and weight. The Protocol also permitted you to modify the plan at any time based on a subjects clinical response and toleration. You or a staff member were to have explained the titration plans to subjects, and to have clearly specified which capsules the subjects were to have taken from each AM and PM column for the first seven days of the titration plan. Furthermore, Section 5.3.4 of Protocols (b)(4) and (b)(4) specified that you were to have assessed subjects' compliance with study medication administration at Week 1 and at each subsequent visit. Additionally, you were to have counted medication returned by subjects to reconcile medication dispensed with subjects' reported usage. There is no evidence that you developed dosing titration plans for the subjects enrolled in these studies, or that you assessed subjects' compliance and performed the requisite drug reconciliation as specified in the protocols. In your March 13, 2009 response, you acknowledged your failure to fully document specific dosing instructions and titration plans. You stated that you recognized that this specific documentation was necessary in order to ensure accurate dosing. As a corrective measure, you stated that your Dosing and Dispensing of Investigational Product SOP will be revised to include more specific information on how, why, and when dosing changes are documented. We acknowledge your assurance that corrective actions will be taken. However, we note that your response did not contain a detailed outline of procedures or processes that would be implemented to prevent future recurrence of the violation noted above. Specifically, we are concerned that your lack of adherence to the protocol in developing dosing titration plans, assessing subjects' compliance, and performing drug reconciliation led to the significant overdoses and resultant adverse events experienced by the pediatric subjects enrolled at your site as described above. As you noted in your response, you were first notified on July 26, 2006 by (b)(4) data management personnel that there were dosing errors in your conduct of these pediatric studies. It is likely that your lack of adherence to the protocols led to your failure to recognize the overdosing at your site, until it was discovered and brought to your attention by the sponsor's data management unit. While there may have been concerns with the packaging of the study drug, your failure to conduct the requisite safety measures contributed to the unnecessary

FDA WARNING LETTERS exposure of pediatric subjects to significant overdoses, which jeopardized the subjects' rights, safety, and welfare.

14

c. Section 6.2 of Protocol (b)(4) required serum pregnancy and urine drug screening tests to be performed at baseline. These tests were not performed for the following subjects: i. Subject 1002's serum pregnancy and urine drug-screening tests were not performed at baseline. ii. Subject 1003's serum pregnancy and urine drug-screening tests were not performed at baseline. iii. Subject 1004's urine drug-screening tests were not performed at baseline. iv. Subject 1007's urine drug-screening tests were not performed at baseline. In your March 13, 2009 response, you stated that Subjects 1002, 1003, 1004, and 1007 were inpatients at baseline. Given that the pregnancy and urine drug screen tests were negative for these subjects at the time of screening, you indicated that your research staff did not repeat the tests at baseline as they assumed there was no opportunity for the results to change. You noted that a Visit Checklist was created to ensure that all study-related procedures are being conducted for each visit, and that completion of this checklist is now standard work practice at your site. We acknowledge your response. However, regardless of the results obtained at screening, the serum pregnancy and urine drug tests should also have been performed at baseline as specified in the protocol. The Visit Checklist you have instituted as standard practice at your site to prevent any similar future occurrences is acceptable. 2. You failed to promptly report to the Institutional Review Board (IRB) all changes in the research activity [21 CFR 312.66]. On August 2, 2006, upon confirming study drug dosing errors, the sponsor instructed you to cease subject enrollment at your site. On February 7, 2007, the sponsor notified you of its decision to terminate your investigational site's participation in the (b)(4) pediatric studies, including Protocols (b)(4) due to good clinical practice (GCP) noncompliance related to dosing errors and continued delays in resolution of queries and data-clarification forms. You failed to notify the IRB of these decisions made by the sponsor. Documentation available at your site indicates that notification of the IRB was limited to the submission of protocol deviation reports concerning subjects' overdosing errors. Furthermore, you failed to notify the IRB of the sponsor's decisions in the continuing review reports and study close-out letters. In your March 13, 2009 response, you stated that you assume full responsibility for the lack of notification to the IRB. To ensure continued compliance with IRB guidelines, you noted that formal training concerning IRBs was conducted at your site on March 17, 2007, and the Western Institutional Review Board Guide for Researchers Versions 1.5 and 1.6 was distributed at your site for independent review. We acknowledge your response and find your corrective actions acceptable.

FDA WARNING LETTERS

15

This letter is not intended to be an all-inclusive list of deficiencies with your clinical study of an investigational drug. It is your responsibility to ensure adherence to each requirement of the law and relevant FDA regulations. You should address these deficiencies and establish procedures to ensure that any on-going or future studies will be in compliance with FDA regulations. Within fifteen (15) working days of your receipt of this letter, you should notify this office in writing of the actions you have taken to prevent similar violations in the future. Failure to adequately and promptly explain the violations noted above may result in regulatory action without further notice. Subjects were randomized postoperatively, as opposed to preoperatively, in violation of the protocol. Per Section 4.6.4.1 of the protocol, randomization was to take place following screening on Day 0, a day prior to surgery. In a Note to File dated April 8, 2009, your site staff notified the Institutional Review Board (IRB) that 44 subjects were randomized post-surgery, in violation of the protocol. The following are examples of subjects that were randomized postoperatively: Subjects 001, 002, 003, 004, 005, 006, 010, 011, 030, 031, 040, 058, 061, and 064. In your June 3, 2009 written response, you stated that this violation occurred due to a misunderstanding of the protocol. You noted that you intended to randomize subjects only after verifying that all inclusion and exclusion criteria were met. You indicated that you understood that the protocol required subjects to have had a (b) (4) and as such, you mistakenly randomized subjects postoperatively to verify that they had indeed undergone the protocol-specified procedure. However, on May 11, 2007, your site was forwarded a monitor's email reiterating that subjects were to be randomized before surgery. Nonetheless, you continued to randomize subjects postoperatively until October 2007. Additionally, in your written response, you indicated that you notified the IRB of this protocol deviation. We note that under section 5.2 of the protocol, any protocol deviations were to be submitted to the IRB as soon as possible. However, you did not notify the IRB until April 8, 2009, almost a year following notification of your site's closure of the study on April 24, 2008, and almost two years after you were notified by the monitor that randomization was to occur prior to surgery. In your response, as corrective measures, you promised to randomize subjects as required by the protocol in the future, and to seek further clarification from the sponsor for any future ambiguities in protocol specifications. We acknowledge your response. However, we are concerned that the response is not adequate to prevent future recurrence of the violation noted above because it provides no specific or detailed plans or procedures to prevent future recurrence

FDA WARNING LETTERS

16

of this or similar violations. In this case, not only did you fail to properly understand the protocol randomization procedure from the start of the study, and therefore did not ensure that the randomization procedure was carried out according to the protocol from the outset, but also you did not remedy the problem for more than four months after having received clarification of the protocol. This letter is not intended to be an all-inclusive list of deficiencies with your clinical study of an investigational drug. It is your responsibility to ensure adherence to each requirement of the law and relevant FDA regulations. You should address these deficiencies and establish procedures to ensure that any on-going or future studies will be in compliance with FDA regulations.