#4416

EZH2 Inhibitor EPZ-6438 Synergizes With Anti-Lymphoma Therapies in Preclinical Models

L. Danielle Johnston1*, Sarah K. Knutson1*, Natalie Warholic1, Christine R. Klaus1, Tim Wigle1, Dorothy Iwanowicz1, Elizabeth A. Admirand1, Bruce A. Littlefield2, Margaret Porter Scott1, Jesse J. Smith1, Mikel Moyer1, Robert A. Copeland1, Roy M. Pollock1, Kevin Kuntz1, Heike Keilhack1†, Alejandra Raimondi1† 1Epizyme Inc., 400 Technology Square, 4th floor, Cambridge MA 02139, USA 2Eisai Inc., 4 Corporate Drive, Andover MA 01810, USA *contributed equally, † corresponding authors Abstract Results

Preclinical data have suggested that small molecule inhibitors for the histone methyltransferase Combination Benefit with CHOP Components and EPZ-6438 in Mutant Glucocorticoid Agonists Enhance Potency of EPZ-6438 in EZH2 Mutant Combination Benefit of EPZ-6438 with Glucocorticoid Agonists in EZH2 EZH2 may represent potential new treatment modalities for Non-Hodgkin lymphomas (NHL) EZH2 Germinal Center B-cell Lymphoma Cell Lines Lymphoma Lines WT Germinal Center But Not Activated B-Cell Lymphoma Lines WSU-DLCL2 (Y646F) expressing EZH2 change of function mutations that merit further investigation in a preclinical or WSU-DLCL2 (Y646F) A EPZ-6438 + Prednisolone B EPZ-6438 + Dexamethasone clinical setting, as appropriate. Our group has previously reported that selective inhibition of EPZ-6438 + (50:1) EPZ-6438 + (400:1) EPZ-6438 + Mafosfamide (1:20) A EPZ-6438 + Prednisolone B EPZ-6438 + Dexamethasone 5.0 5.0 5.0 Fa CI Fa CI 0.36 3.05 4.0 4.0 EZH2 results in specific killing of lymphoma cells bearing EZH2 mutations in vitro and in vivo, with 0.53 0.63 Fa CI 4.0 0.57 0.93 0.68 0.44 0.46 1.87 0.70 0.85 3.0 3.0 0.58 1.20 3.0 1 0.82 0.34 0.82 0.93 minimal effects on non-mutant lymphoma cells [Knutson et al. Nature Chemical Biology 2012 ; 0.89 0.39 0.70 0.91 GCB: 2.0 0.80 1.05 0.95 0.48 2.0 2.0 0.97 1.06 CI +/-CI 1.96 s.d. CI +/-CI 1.96 s.d. 2 +/-CI 1.96 s.d. Keilhack et al. Blood (ASH Annual Meeting Abstracts) 2012, 120, Abstract 3712 ]. Since epigenetic 1.0 1.0 1.0 DOHH2 0 0 0 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 changes have been suggested to be involved in resistance of cancer cells to many anticancer Fractional Effect Fractional Effect Fractional Effect agents, we studied EPZ-6438 (E7438), our clinical stage EZH2 inhibitor, in combination with EPZ-6438 + Doxorubicin (10:3) EPZ-6438 + Vincristine (800:1) EPZ-6438 + Mafosfamide (4:25) 5.0 Fa CI 5.0 5.0 Fa CI Fa CI 0.49 0.71 SU-DHL-10 (Y646F) standard of care agents for NHL, second line therapies, or targeted therapies that are being 4.0 0.11 1.28 0.20 0.77 0.60 0.65 4.0 4.0 C EPZ-6438 + Prednisolone D EPZ-6438 + Dexamethasone 0.30 1.01 0.33 0.76 0.64 0.95 0.35 1.71 0.39 1.19 3.0 0.73 0.95 3.0 3.0 C EPZ-6438 + Prednisolone D EPZ-6438 + Dexamethasone 0.65 1.35 0.62 1.00 explored in this indication. With continuous exposure to EPZ-6438, cell-based assays of two 0.81 1.05 2.0 2.0 0.93 0.64 2.0 0.82 0.83 CI +/-CI 1.96 s.d. CI +/-CI 1.96 s.d. CI +/-CI 1.96 s.d. different EZH2 mutant cell lines demonstrated combination benefits with all components of the 1.0 1.0 1.0

0 0 0 ABC: 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 CHOP regime, second line therapies, but also with several targeted therapies (for Fractional Effect 0 0.2 Fractional0.4 Effect0.6 0.8 1.0 Fractional Effect Toledo CI < 1 synergy instance other epigenetic drugs, PI3K pathway or other inhibitors). These effects were not CI = 1 additive observed in an EZH2 wild type lymphoma cell line of the activated B cell type. Strong SU-DHL-10 (Y646F) CI > 1 antagonism combination benefit with CHOP was also observed in two different EZH2 mutant xenograft EPZ-6438 and doxorubicin act synergistically in the WSU-DLCL2 cells and produce an additive Combination benefit was observed in DOHH2 EZH2 wild type GCB cells upon treatment Potency of EPZ-6438 is dramatically increased when combined with glucocorticoid models. For instance, in the SUDHL6 Y646N xenograft model neither EPZ-6438 nor CHOP effect in SU-DHL-10 cells. Combination benefit is observed with mafosfamide and vincristine in with EPZ-6438 and prednisolone (A) or dexamethasone (B), according to pre-treatment both EZH2 Y646 mutant cell lines. In WSU-DLCL2 doses ranged from 0.16-20nM for agonists. The addition of prednisolone (A, C) or dexamethasone (B, D) in 2 EZH2 model A (see Methods section). In contrast, no combination benefit was observed in chemotherapy alone induced a significant antitumor effect, yet their combination produced doxorubicin, 0.04-5nM for vincristine, 0.156-10µM for mafosfamide, and 15-1000nM for EPZ- Y646F mutant DLBCL lines according to pre-treatment model A (see Methods section) Toledo cells, an EZH2 wild type ABC lymphoma line. Doses ranged from 15nM-1000nM produces a dose dependent shift in the IC durable tumor regressions even after cessation of dosing. Importantly, this effect was preserved 6438. In SU-DHL-10 cells doses ranged from 0.5-60nM for doxorubicin, 0.016-2nM for 50 of EPZ-6438. Doses ranged from 15nM- for prednisolone and from 1.5nM-100nM for dexamethasone in both cells lines. EPZ- vincristine, 0.156-10µM for mafosfamide, and 1.56-100nM for EPZ-6438. Cells were treated 1000nM for prednisolone and 1.5nM-100nM for dexamethasone in both cell lines. 6438 ranged from 0.16-10µM in DOHH2 cells and 15.6-1000nM in Toledo cells. when doxorubicin was omitted from the CHOP chemotherapy regime in a third study with according to pre-treatment model A, and data analyzed with the Calcusyn software (see Doses of EPZ-6438 ranged from 15-1000nM in WSU-DLCL2 cells and 1.5-100nM in SU- another EZH2 mutant xenograft model. Subsequently we showed that glucocorticoid receptor Methods section). DHL-10 cells. EPZ-6438-CHOP Combinations Show Enhanced Anti-tumor Activity agonism may be a key mechanism of the combination benefit observed with CHOP, as the Compared to Single Agents in Several EZH2 Mutant Lymphoma EPZ-6438/Glucocorticoid Agonist Combination Overcomes EZH2i Very Strong Synergy Observed in EZH2 Mutant Lymphoma Cell Line Xenograft Models antiproliferative effect of EPZ-6438 was enhanced by either prednisolone or dexamethasone Insensitivity in Cell Lines Resistant to EZH2i with Combination of EPZ-6438 and Other Targeted Therapies A EPZ-6438 WSU-DLCL2 alone, in several EZH2 mutant lymphoma cell lines (in vitro). Taken together these data suggest

that the single agent activity of EPZ-6438 in EZH2 mutant NHL may be further enhanced and 4 Day EPZ-6438 IC50 (µM) 7 Day EPZ-6438 IC50 (µM) EPZ-6438 Synergizes with EPZ-6438 Synergizes with Cell Line BCL2 inhibitor-Navitoclax mTOR inhibitor-Everolimus EPZ-6438 expanded through rational combination strategies. EPZ-6438 EPZ-6438 4d 6438 Pre/ 4d Pred Pre/ EPZ-6438 C Alone 4d Co-treatment 3d Co-treat 3d Co-treat 7d Co-treatment CHOP CHOP WSU-DLCL2: EPZ-6438 + Navitoclax (1:5) WSU-DLCL2: EPZ-6438 + Everolimus (400:1) WSU EPZ-6438 0.53 0.020 0.011 >1 0.014 5.0 5.0 B (Y646-Sens) Methods Fa CI Fa CI 4.0 4.0 SU-DHL-10 0.22 1.44 0.43 0.61 COP COP 0.34 0.76 0.57 0.35 0.64 0.0092 0.0027 0.52, >1 0.020 3.0 SUDHL10 Design of in vitro Combination Assays (Y646-Sens) 3.0 0.43 0.68 0.74 0.12 0.57 0.43 0.81 0.10 EPZ-6438 RL 2.0 0.73 0.20 2.0 0.88 0.07

>1 0.0096 <<0.004 0.38 <0.004 0.86 0.07 +/-CI 1.96 s.d. 0.89 0.11 CI +/-CI 1.96 s.d. Pre-treatment Model (Y646-Res) 1.0 1.0 CHOP SUDHL6 SU-DHL-4 0 0 >1 >1, 0.20, >1 0.035 >1 0.51 0 0.2 0.4 0.6 0.8 1.0 Day 0 Day 4 Day 7 (Y646-Res) 0 0.2 Fractional0.4 Effect0.6 0.8 1.0 Fractional Effect EPZ-6438 DOHH2 >1 0.20 1.9 >1 0.34 CI < 1 synergy Pre-treat with Co-treatment with dose matrix of EPZ- ATP (WT) CI = 1 additive COP COP 7 Doses EPZ-6438 Normalize cell number 6438 and Compound of Interest A OCI-Ly19 CI > 1 antagonism Re-dose >1 0.19 0.0055 >1 0.026 (WT) Pre-treat with Normalize cell number Co-treatment with 7 Doses EPZ-6438 B 1 Dose Compound of Interest Re-dose and 1 Dose Compound of Interest Very strong synergy is observed when EPZ-6438 is combined with the Bcl-2 inhibitor CHOP navitoclax, as well as with mTOR inhibitor everolimus. Dose ranges for Navitoclax Pre-treat with Normalize cell number Co-treatment with 7 Doses EPZ-6438 Overall, a combination of prednisolone and EPZ-6438 leads to greater sensitivity in A. WSU-DLCL2 28 day xenograft study, treated with single agent, CHOP, or combination. are 0.16-10µM, 0.04-5nM for everolimus, and 31-2000nM for EPZ-6438. These data C 7 Doses EPZ-6438 Re-dose and 1 Dose Compound of Interest all GCB cell lines tested, not just EZH2i sensitive cell lines. Sequence of drug B. SU-DHL-6 xenograft study, 28 day dosing with single agent, CHO P, or combination, and tumor were generated in the pre-treatment model A and data analyzed with Calcusyn growth delay up to 60 days. Top graph is mean tumor volume. Bottom graph is survival curve addition is crucial. For all lines except RL, preincubation with prednisolone, Co-treatment Model: 4 or 7 day software (see Methods section). (animals euthanized if their tumor volume exceeded 2000mm3). followed by EPZ-6438, is not effective. Day 0 x 7 doses EPZ - 6438 C. SU-DHL-10 xenograft study, 28 days dosing with single agent, COP (CHOP without doxorubicin), x 1 constant dose of compound of interest or combination, and tumor growth delay up to the 60 days. Top graph is mean tumor volume. Day 4 Day 7 Summary Table of Combinations with EPZ-6438 Bottom graph is survival curve (animals euthanized if their tumor volume exceeded 2000mm3).

WSU-DLCL2 SU-DHL-10 Toledo DOHH2 Data Analysis-Chou Talalay: Pre-treatment Model A (EZH2 mutant GCB) (EZH2 mutant GCB) (WT EZH2 ABC) (WT EZH2 GCB) Compound A 7x potency 2x potency 2x potency Prednisolone no effect Conclusions enhancement enhancement enhancement Fa-CI Plot Doxorubicin synergy additive no effect no effect 5.0 Standard of • EPZ-6438 and glucocorticoid receptor agonists (GRag) cooperate to dramatically enhance antiproliferative activity in NHL of the Mafosfamide additive additive no effect no effect Care 4.0 Fa CI DLBCL Vincristine additive additive no effect no effect germinal center B subtype 0.53 0.63 3.0 CI < 1 synergy 0.68 0.44 Cisplatin synergy additive no effect no effect • The in vitro cell killing activity of EPZ-6438 is enhanced by GRag in EZH2 mutant bearing cells. 2.0 CI = 1 additive Compound B Compound 0.82 0.34 AraC synergy additive no effect no effect

CI +/-CI 1.96 s.d. CI > 1 antagonism • EZH2i enhances the GRag antiproliferative activity in WT EZH2 GCB cell lines. 1.0 0.89 0.39 additive additive no effect no effect 0.95 0.48 0 Epigenetic * additive Not tested Not tested Not tested • The combination of EPZ-6438 and GRag reverses insensitivity in EZH2i resistant mutant cell lines. 0 0.2 Fractional0.4 Effect0.6 0.8 1.0 Drugs Combination Index Equation Azacytadine* additive Not tested Not tested Not tested • In two different EZH2 mutant xenograft models, strong combination benefit was demonstrated with EPZ-6438 and CHOP, and this Everolimus very strong synergy strong synergy no effect no effect effect was preserved in a study in a third EZH2 mutant xenograft model in which doxorubicin was omitted from the chemotherapy 15x potency 4x potency Dexamethasone 5x potency enhancement no effect Other enhancement enhancement Therapies regime. Pre-treatment model A: Lymphoma cells were pre-treated in flasks with 7 doses of EPZ-6438 or DMSO for 4 days. Cell densities were normalized Navitoclax Very strong synergy 2x potency enhancement Not tested No effect and co-treated with EPZ-6438 and compound of interest (3 replicate plates in matrix of two compound combinations in a constant ratio) using the Obatoclax additive additive No effect No effect • These results suggest that: CI < 1 synergy HP D300 digital dispenser. After 3 days of co-treatment, cell viability was measured via ATP content using CellTiter-Glo® No effect = No change in drug IC50 upon addition of EPZ-6438 • Glucocorticoid receptor agonism may play a key role in the amplified anti-tumor activity observed with combinations of EPZ-6438 and * Experiments were performed with EZH2i tool compound instead of EPZ-6438. CI = 1 additive Pre-treatment model B,C: Lymphoma cells were pretreated with 1 dose of compound of interest (B) or 7 doses of EPZ-6438 (C) and DMSO for 4 CI > 1 antagonism CHOP in EZH2 mutant lymphoma xenografts. days. Cell densities were then normalized and co-treated with 1 dose of compound of interest and 7 doses of EPZ-6438 and DMSO for 3 days. Combination benefit with EPZ-6438 is achieved with all drugs tested in EZH2 Viability was determined using Guava ViaCount® Reagent . mutant lymphoma lines. Glucocorticoid receptor agonists demonstrate • EZH2i may have a broader role in GCB lymphoma beyond the subset with an EZH2 mutation. Co-treatment Model: Lymphoma cells were treated with 7 doses of EPZ-6438 and 1 dose of compound of interest for either 4 or 7 days. Viability combination benefit with EZH2 WT and mutant GCB lymphoma lines, but not in a References was determined using Guava ViaCount® Reagent . ABC lymphoma cell line. 1. Knutson et al. A selective inhibitor of EZH2 blocks H3K27 methylation and kills mutant lymphoma cells, Nature Chemical Biology, 2012. 8:890-896. Data Analysis using Chou-Talalay method: Synergy quantification is performed using the Chou-Talalay method3 for drug combination. The 2. Keilhack et al. Preclinical characterization of E7438, a potent, selective inhibitor of protein methyltransferase EZH2 with robust antitumor activity against EZH2 mutated non-Hodgkin Combination Index (CI) equation offers a quantitative definition for additivity (CI=1), synergism (CI < 1), and antagonism (CI > 1). This equation used lymphoma xenografts in mice, Blood (ASH Annual Meeting Abstracts) Nov 2012: 120: 3712. Fa values from a constant ratio of drug combination to determine CI values. The resulting plot (Fa-CI) plot shows the resultant CI values bracketed 3. Chou theoretical basis, experimental design and computerized simulation of synergism and antagonism in drug combination studies, Pharmacol Rev, 2006. 58:621-681. by 95% confidence intervals. These Fa-CI plots are generated using the Calcusyn software. Statistically significant CI values for synergy are for Disclosures: Johnston, Iwanowicz, Admirand, and Moyer: Epizyme: Employment, Equity Ownership, stock options Other. Knutson, Warholic, Klaus, Wigle, Porter Scott, Smith, Pollock, Kuntz, Keilhack, and Raimondi: Epizyme, Inc.: Employment, Equity example those CI value< 1 with the confidence interval lines also below 1. Ownership, Patents & Royalties, stock options Other. Littlefield: Eisai Inc.: Employment. Copeland: Epizyme Inc. : Employment, Equity Ownership, Patents & Royalties, stock options Other; Mersana: Membership on an entity’s Board of Directors or advisory www.epizyme.com committees.