PARP Inhibitors: Usurping DNA repair to target cancer Lee Schwartzberg MD, FACP Chief Medical Officer The West Clinic

Question 1 DNA repair mechanisms are important in Cancer cells only Both cancer and normal eukaryotic cells Predominantly in rapidly growing cells like bone marrow precursors Predominantly cancer cells with BRCA mutations

Question 2 PARP inhibitors have demonstrated activity in: BRCA 1 mutation carrier breast cancer BRCA 2 mutation carrier breast cancer Triple negative breast cancer 1 and 3 only 1 and 2 only All of the above

All cells are under constant risk of DNA damage Ultraviolet light Ionizing radiation Man-made and natural chemicals Reactive oxygen species most are generated “endogenously” 10,000 Single Strand Breaks/ cell/day ~100,000,000,000,000,000 DNA lesions in a human body every day1-3 Mammalian cells are constantly at risk of damage to their DNA from multiple causes1, including ultraviolet light, ionizing radiation, reactive oxygen species, several endogenous and synthetic compounds, and during normal DNA replication. Approximately 1x104 DNA lesions (single-strand breaks [SSBs]) are generated in a metabolically active mammalian cell each day.1,2 There are approximately 1x1013 cells in the human body, so this corresponds to a total of around 1x1017 DNA lesions per person, per day.1-3 References 1. Jackson SP. Biochem Soc Trans 2001; 29: 655-661. 2. Lindahl T. Nature 1993; 362: 709-715. 3. Jackson SP, Bishop CL. Drug Discovery World 2003; Fall: 41-45. 1. Jackson SP. Biochem Soc Trans 2001;29:655-661 2. Lindahl T. Nature 1993;362:709-715 3. Jackson SP, Bishop CL. Drug Discovery World 2003;(Fall):41-45

Cellular Response To DNA Damage

Cancer cells are highly susceptible to DNA repair inhibition Undergo deregulated proliferation Less time for DNA repair than in normal cells Grow under stress, which causes ongoing DNA damage Have DNA repair defects P53, BRCA1, BRCA 2, ATM, Fanconi’s Anemia Allow growth despite ongoing genome instability Are reliant on the DNA repair pathways they still retain Eukaryotic cells have evolved mechanisms to monitor the integrity of their genome and repair damaged DNA before the cell cycle progresses. The mechanisms that monitor for damaged DNA are intimately linked with cell-cycle events and the process known as “checkpoint” control. Loss of checkpoint function predisposes a cell to acquire selective oncogenic mutations and, therefore, it is an important prognostic indicator. People with genetic instability disorders that cause defects in checkpoint function have an increased incidence of many cancers.1 Cancer cells undergo deregulated proliferation, losing cell-cycle control and resting in the G0 phase of the cell cycle for little or no time. Thus, there is less opportunity for DNA repair to occur than in normal cells. Growth under stress increases the risk of DNA damage in cancer cells. In addition, the genomic instability associated with cancer cell formation (“mutator phenotype”) results in an increased DNA mutation rate.2 Many DNA repair pathways are lost or deficient in cancer cells. In particular, breast or ovarian tumor cells with hereditary breast cancer associated gene (BRCA) 1 or BRCA2 mutations are HR deficient and so they lack the ability to efficiently repair DSBs.3 Cancer cells continue to grow despite ongoing genomic instability and so the repair pathways they still retain become integral to their survival. Interest in targeting these remaining DNA repair pathways in cancer cells as a selective anti-cancer therapy is developing. References 1. Khanna K. J Natl Cancer Inst 2000; 92: 795-802. 2. Bielas JH, Loeb LA. Environ Mol Mutagen 2005; 45: 206-213. 3. Lomonosov M et al. Genes Dev 2003; 17: 3017-3022.

DNA Excision Repair Mechanisms

Poly(ADP-Ribose) Polymerase (PARP) A key role in the repair of DNA single-strand breaks Through the base excision repair pathway (BER) Binds directly to sites of DNA damage Once activated, it uses NAD as a substrate, and generates large, branched chains of poly (ADP-ribose) polymers on multiple target proteins Recruits other DNA repair enzymes PAR Lig3 XRCC1 PNK Polß 8 8

Base Excision Repair

Inhibiting PARP-1 Increases Double-Strand DNA Damage DNA single strand break (SSB) damage XRCC1 LigIII PNK 1 pol β PARP Inhibition of PARP-1 prevents recruitment of DNA repair enzymes leads to failure of SSB repair -accumulation of SSBs Inhibition of PARP-1 activity prevents the recruitment of DNA repair enzymes and leads to failure of SSB repair and accumulation of SSBs. During the S-phase of the cell cycle, the replication fork is arrested at the site of an SSB, which then degenerates into a DSB. In normal cells, this triggers activation of the HR pathway to repair the DSB.1 Abbreviation on slide: PNK 1, polynucleotide kinase 1 Reference 1. Helleday T et al. Cell Cycle 2005; 4: 1176-1178. During S-phase, replication fork is arrested at site of SSB Degeneration into Double strand breaks

BRCA1 And 2 Are Required for Efficient Repair of Double Stranded DNA Breaks DNA DSB ATM/R gH2AX BRCA1 Rad50 MRE11 NBS1 Non-homologous end-joining Homologous recombination Ku 70/80 BRCA2 Rad 51 Cell survival Cancer cell death DNA-PKcs RPA Rad 52/4 ERCC1 XRCC3 XRCC4 Ligase IV Predominant in G1 Error-prone Gross Genomic instability Major pathway for repair Error-free Cells with BRCA mutations are deficient in homologous recombination and lack the ability to efficiently repair DSBs.

The Concept of Synthetic Lethality (BRCA) (PARP) Ashworth, A. J Clin Oncol; 26:3785-3790 2008

BRCA1 and BRCA2 -/- cells are very sensitive to PARP inhibition Increased levels of chromosomal aberrations in PARP inhibitor treated BRCA2 -/- cells BRCA2 +/- BRCA2 -/- Wild type Log surviving fraction - 4 3 2 1 PARP inhibitor concentration (M) 10-9 10-8 10-7 10-6 10-5 10-4 Wild type Control + PARP inhibitor In vitro and in vivo studies have investigated the potential therapeutic effect of PARP inhibition on cells that have deficient DNA repair systems.1,2 One of these studies by Farmer et al used mouse embryonic stem (ES) cells lacking BRCA1 or BRCA2.1 BRCA1 and BRCA2 are essential for efficient DNA DSB repair by HR3 and mutations in these genes predispose to breast, ovarian, and other cancers.4 In HR-deficient cells, the activity of PARP-1 reduces the DNA damage burden. Farmer et al demonstrated that BRCA2 (and BRCA1)-deficient mouse ES cells were extremely sensitive to PARP inhibition.1 As a result of HR deficiency in these cells, the addition of PARP inhibitors led to chromosomal instability, cell-cycle arrest, and subsequent cell death.1 The effects were rapid and irreversible. Furthermore, to investigate the possibility that PARP inhibition may be used to treat BRCA2-deficient tumors, Bryant et al2 conducted a xenograph study in nude mice. The results indicated that BRCA2-deficient tumors were susceptible to PARP inhibition alone.2 References 1. Farmer H et al. Nature 2005; 434: 917-921. 2. Bryant HE et al. Nature 2005; 434: 913-917. 3. Tutt A, Ashworth A. Trends Mol Med 2002; 8: 571-576. 4. Wooster R, Weber BL. N Engl J Med 2003; 348: 2339-2347. BRCA2 -/- Control + PARP inhibitor Farmer H et al. Nature 2005;434:917-920 Personal communication, Alan Ashworth

PARP Inhibitors in Clinical Development Differing chemical structures Differing toxicity Differing schedules and routes of administration

Chemotherapeutic Agents: Double Strand DNA Breaks Alkylators DNA interstrand cross-links  double strand (DS) DNA breaks Cyclophosphamide Platinums Forms adducts with DNA Cisplatin Carboplatin Oxaliplatin Topoisomerase I poisons Arrest of DNA replication forks Etoposide Irinotecan Topotecan Mitoxantrone Topoisomerase II poisons DNA interstrand cross-linking, generation of O2 free radicals Doxorubicin Epirubicin Bleomycin Directly damages DNA  DS DNA breaks Kennedy R et al. JNCI 2004; 96:1659-1668

PARP Inhibitors in BRCA 1/2 Mutated Tumors

Phase I Trial of Olaparib in Patients with Solid Tumors Escalation and expansion phase, n = 60 Recommended phase II dose: 400 mg PO BID Toxicities Nausea (32%), fatigue (30%), vomiting (20%), taste alteration (13%), anorexia (12%), anemia (5%) Clinical activity = 12/19 patients with BRCA mutations Tumor BRCA No. of pts Response Breast 2 1 CR, 1 SD Ovarian 1 or 2 8 8 PRs Fallopian tube 1 PR Prostate Fong PC et al. N Engl J Med 2009; 361:123-134

Phase II Trial of Olaparib in BRCA-deficient Metastatic Breast Cancer Eligibility Confirmed BRCA1 or 2 mutation Stage IIIB/C or IV BC after progression ≥ 1 prior chemotherapy for advanced disease (Non-randomized sequential cohorts) Cohort 1 Olaparib 400 mg po bid (MTD) 28-day cycles Cohort 2* Olaparib 100 mg po bid (maximal PARP inhibition) 28-day cycles Primary Endpoint: Response rate * Following an interim review, patients in the 100 mg bid cohort were permitted to crossover to receive 400 mg bid Tutt A et al. J Clin Oncol 2009;27(18S):803s (abstr CRA501)

Olaparib in BRCA-deficient Metastatic Breast Cancer: Select Toxicities Olaparib 400 mg BID (n = 27) Olaparib 100 mg BID Grade 1/2 Grade 3 Fatigue 15 (56) 4 (15) 2 (7) Nausea 11 (41) 5 (19) Vomiting 7 (26) 3 (11) 6 (22) Headache 10 (37) 1 (4) Constipation 8 (30) Tutt A et al. J Clin Oncol 2009;27(18S):803s (abstr CRA501)

Olaparib in BRCA-deficient Metastatic Breast Cancer: Results Best percent change from baseline in target lesions by genotype Median 3 prior lines of therapy ITT cohort 400 mg BID N = 27 100 mg BID ORR 11 (41%) 6 (22%) CR 1 (4%) PR 10 (37%) Median PFS 5.7 mo (4.6-7.4) 3.8 mo (1.9 – 5.6) Tutt A et al. J Clin Oncol 2009;27(18S):803s (abstr CRA501)

PARPi Monotherapy in BRCA Mutated tumors Drug Phase Dose Tumor N CBR (%) RR (%) MDR (MOS) PFS (MOS) Olapirib 1 Varies Ovarian 50 46 40 6.5 NR 2 400 mg BID 33 35 9.6 100 mg BID 24 13 9.0 400 mg Breast 27 41 5.7 22 3.8 MK-4827 19 45 4

Prior response to platinum may predict response to olaparib in BRCA mutated Ovarian Cancer Gelmon K, et al J Clin Onc 2010

PARP Inhibitors beyond BRCA mutation carriers

Triple Negative Breast Cancer (TNBC) ‘Triple negative’: ER-negative, PR-negative, HER2-negative Depending on thresholds used to define ER and PR positivity and methods for HER2 testing TNBC accounts for 10–17% of all breast carcinomas Significantly more aggressive than other molecular subtype tumors Higher relapse rate than other subtypes No specific targeted therapy Reis-Filho JS, et al. Histopathology 2008;52:108-118.

Triple Negative/Basal-Like1,2,3 TNBC Shares Clinical and Pathologic Features with BRCA-1-Related Breast Cancers (“BRCAness”) Characteristics Hereditary BRCA1 Triple Negative/Basal-Like1,2,3 ER/PR/HER2 status Negative TP53 status Mutant BRCA1 status Mutational inactivation* Diminished expression* Gene-expression pattern Basal-like Tumor histology Poorly differentiated (high grade) Chemosensitivity to DNA-damaging agents Highly sensitive *BRCA1 dysfunction due to germline mutations, promoter methylation, or overexpression of HMG or ID44 1Perou et al. Nature. 2000; 406:747-752 2Cleator et al.Lancet Oncol 2007;8:235-44 3Sorlie et al. Proc Natl Acad Sci U S A 2001;98:10869-74 4 Miyoshi et al. Int J Clin Oncol 2008;13:395-400

Targeting DNA Repair Pathway in TNBC Clustering analyses of microarray RNA expression have shown that familial BRCA-1 tumors strongly segregate with basal-like/ triple-negative tumors Suggests that sporadic TNBC may have acquired defects in BRCA1-related functions in DNA repair Basal-like = BRCA1+ = BRCA2+ Sorlie T et al. PNAS 2003;100:8418-8423

Predictors of Response to Cisplatin in TNBC Silver, D. P. et al. J Clin Oncol; 28:1145-1153 2010

Phase II Study of the PARP inhibitor Iniparib in Combination with Gemcitabine/Carboplatin in Triple Negative Metastatic Breast Cancer Background and Rationale PARP1 Upregulated in majority of triple negative human breast cancers1 Iniparib (BSI-201) Small molecule IV PARP inhibitor Potentiates effects of chemotherapy-induced DNA damage No dose-limiting toxicities in Phase I studies of BSI-201 alone or in combination with chemotherapy Marked and prolonged PARP inhibition in PBMCs O’Shaughnessy J, et al. NEJM 2011

Phase II TNBC Study: Treatment Schema Metastatic TNBC N = 120 RANDOMIZE 1st -3rd line MBC Eligible BSI-201 (5.6 mg/kg, IV, d 1, 4, 8, 11) Gemcitabine (1000 mg/m2, IV, d 1, 8) Carboplatin (AUC 2, IV, d 1, 8) Gemcitabine (1000 mg/m2, IV, d 1, 8) Carboplatin (AUC 2, IV, d 1, 8) 21-Day Cycle RESTAGING Every 2 Cycles * Patients randomized to gem/carbo alone could crossover to receive gem/carbo + BSI-201 at disease progression

Safety – Hematologic Toxicity Phase II Gem Carbo +/- Iniparib BSI-201 + Gem/Carbo (n = 57) Grade 2 Grade 3 Grade 4 Anemia, n (%) 12 (20.3%) 7 (11.9%) (0.0%) 15 (26.3%) 7 (12.3%) Thrombocytopenia, n (%) 6 (10.2%) 4 (7.0%) (10.5%) Neutropenia, n (%) 18 (30.5%) 13 (22.0%) 7 (12.3%) (31.6%) Febrile neutropenia, n (%) 3 (5.1%) 1 (1.7%) RBC treatment*, n (%) 5 (8.5%) 2 (3.4%) (5.3%) (8.8%) (3.5%) G-CSF Use, n (%) (1.8%) *Transfusion and/or EPO use O’Shaughnessy J, et al. NEJM 2011

Safety – Non-Hematologic Toxicity Phase II Gem Carbo +/- Iniparib BSI-201 + Gem/Carbo (n = 57) Grade 2 Grade 3 Grade 4 Nausea, n (%) 10 (16.9%) 2 (3.4%) (0.0%) 7 (12.3%) Vomiting, n (%) 9 (15.3%) 4 (7.0%) 1 (1.8%) Fatigue, n (%) 10 (16.9%) 6 (10.2%) 10 (17.5%) Neuropathy, n (%) Diarrhea, n (%) (1.7%) O’Shaughnessy J, et al. NEJM 2011

Final Results: Phase II: Gem Carbo +/- Iniparib in TNBC O’Shaughnessy J et.al. NEJM 2011

Final Results: Phase II Gem Carbo +/- Iniparib in TNBC O’Shaughnessy J, et.al. NEJM 2011

Phase I: Olaparib + Paclitaxel in 1st and 2nd line MBC BKG: Olaparib single agent activity in BRCA 1/2 mutated MBC Olaparib + paclitaxel, N=19, 70% 1st line, unselected for BRCA mutations 33-40% RR; no CRs Median PFS: 5.2-6.3 months Hematologic toxicity high, requires G-CSF Dose reductions common Unclear whether combination be taken forward

Resistance to PARP Inhibitors: Reversion of BRCA2 mutations Partial function of BRCA2 is restored and cells become competent for homologous recombination repair Selection of resistant clones from a BRCA2 mutated cell line reveals that resistance to PARP inhibitors is associated with reversion of BRCA2 mutation. Investigators go on to show that partial function of BRCA2 is restored and that cells become competent for HR repair. A number of important points from this study: (1) Validates the mechanism of action of PARP inhibitors in BRCA2 deficient patients, (2) PARP inhibition unlikely to be a “magic bullet”, (3) Continued HR deficiency is not required for tumor survival and development. Edwards SL et al. Nature 2008; 451:1111-1115

The Future of PARP inhibitors: Many Unanswered Questions Can we use these agents more broadly? To treat other tumors with specific DNA repair defects, i.e. sporadic loss of BRCA 1/2, tumors with PTEN mutations Challenge is to identify them Timing of PARP inhibitor in relation to cytotoxic agent (before it, with it, how long to continue it?)

Conclusions Targeting DNA repair mechanisms in tumor cells is a rational target PARP is an integral enzyme in DNA repair Multiple PARP inhibitors are available Preliminary results show activity in BRCA mutated cancers (Breast and Ovarian) Preliminary results show activity of iniparib with chemotherapy in TNBC Phase III results forthcoming