INTERIM REPORT FROM A PHASE 2 MULTICENTER STUDY OF TAZEMETOSTAT, AN EZH2 INHIBITOR: CLINICAL ACTIVITY AND FAVORABLE SAFETY IN PATIENTS WITH RELAPSED OR REFRACTORY B-­CELL NON-­HODGKIN LYMPHOMA

1

Franck Morschhauser, Gilles Salles, Pamela McKay, Hervé Tilly, Anna Schmitt,John Gerecitano, Peter Johnson, Steven Le Gouill, Michael J. Dickinson,Christophe Fruchart, Thierry Lamy, Aristeidis Chaidos, Wojciech Jurczak, Stephen Opat,John Radford, Pier Luigi Zinzani, Sarit Assouline, Guillaume Cartron, Alicia Clawson,Natasha Picazio, Scott Ribich, Stephen J. Blakemore, John Larus, Harry Miao,Mark Woodruff, Peter T. Ho, Vincent Ribrag

International Conference on Malignant Lymphoma (ICML)June 14-­17, 2017

TAZEMETOSTAT FOR THE TREATMENT OF B-­CELL NHL

• EZH2 is an epigenetic regulator of gene expression and plays a critical role in multiple forms of cancer

‒ Activating mutations of EZH2 can act as an oncogenic driver for cancers, especially in FL and GCB-­DLBCL, present in ~20% of patients

• Tazemetostat

‒ First-­in-­class, potent and selective oral inhibitor of mutated and wild-­type EZH2

‒ Preclinical activity in DLBCL cells lines, with greater activity in EZH2 mutant models

‒ Monotherapy activity and favorable safety in phase 1 studies in patients with relapsed or refractory (R/R) NHL, as well as certain genetically defined solid tumors

2

K27me3K27me3

K27me3K27me3

Transcriptional Repression

Compacted Chromatin

EZH2

PRC2

Y646F/N/H/S/CA682GA692V

Tazemetostat

TAZEMETOSTAT PHASE 2 NHL STUDY DESIGN

• Global, multi-­center, open-­label study in 6 cohorts of patients with R/R DLBCL or FL‒ Patients prospectively stratified by EZH2 mutational status and cell of origin‒ ≥2 prior therapies

• Primary endpoint: overall response rate 1 (ORR)‒ Secondary efficacy endpoints: progression-­free survival (PFS) and duration of response

• Study initiated with 5 monotherapy cohorts (n=270)‒ 3 cohorts of DLBCL (n=60 each) and 2 cohorts of FL (n=45 each)‒ Cohort of tazemetostat + prednisolone (n=70) in DLBCL added in 2017

31 Objective response assessed by IWG-­NHL criteria (Cheson 2007)Restaging every 8 weeks for 6 cycles, then every 12 weeks thereafter

PRE-­SCREENING

COHORT ALLOCATION

ELIGIBILTY,

ENROLLMENT

EOT FOLLOW-­UP

DLBCL, GCBEZH2 Mt (N=60)

FL, EZH2 WT (N=45)

FL, EZH2 Mt (N=45)

DLBCL, NON-­GCB (N=60)

DLBCL, GCBEZH2 WT (N=60)Archival Tissue

Central Lab COO, EZH2

Tazemetostat, 800 mg BID until PD or withdrawal

ORR, PFS, DOR, safety, PK

OS

PROTOCOL SPECIFIED MOLECULAR CHARACTERIZATION OF B-­CELL NHL PATIENTS

4

• Prospective testing: required for cohort allocation‒ FL & DLBCL – cobas® EZH2 Mutation Test (Roche Molecular Systems, in development) • Allele specific PCR test for EZH2 hot spot mutations

‒ DLBCL only – Hans IHC for determination of cell of origin (COO)

• Retrospective testing:‒ NGS on archival and circulating tumor DNA collected at screening for common NHL mutations in a panel of 62 genes

‒ Lymph2Cx nanoString assay for more accurate COO determination• Published discordance rate of ~20% between nanoString vs. Hans

PHASE 2 NHL STUDY PROGRESS

• Data cut-­off: June 1, 2017

• 218 patients enrolled into monotherapy cohorts‒ 81% of total of 270‒ Closed to accrual:• FL wild-­type EZH2 – 54 pts, 54 evaluable for efficacy• DLBCL wild-­type EZH2 (GCB EZH2 WT + non-­GCB) – 120 pts, 119 evaluable for efficacy

‒ Open to accrual: • FL mutated EZH2 – 19 pts, 13 evaluable for efficacy• DLBCL mutated EZH2 – 22 pts, 17 evaluable for efficacy

• Evaluable population‒ Safety: 210 pts‒ Efficacy: 203 pts

5

Characteristic Follicular Lymphoma DLBCL

EZH2 Status Mutant Wild-­type Mutant Wild-­typen 13 54 17 120

Age, median years 62 61 61 69

Males 46% 63% 53% 58%

ECOG PS, median (range) 0 (0 -­ 2) 0 (0 -­ 2) 1 (0 -­ 2) 1 (0 -­ 2)

Prior lines of therapy, n (%) 1 1 ( 8%) 0 0 3 ( 3%)

2 2 (15%) 11 (20%) 4 (24%) 40 (33%)

3 3 (23%) 9 (17%) 7 (41%) 28 (23%)

4 1 ( 8%) 14 (26%) 3 (18%) 18 (15%)

≥ 5 6 (46%) 20 (37%) 3 (18%) 31 (26%)

median 4 4 3 3

Refractory to last regimen, n (%) 7 (54%) 26 (48%) 14 (82%) 75 (63%)

Prior HSCT 23% 41% 41% 24%

Median time from initial diagnosis years 7.4 4.9 1.0 2.0

Median time from last prior therapy weeks 13.0 41.3 8.6 11.6

Data as of 6/1/2017Refractory to last regimen defined as SD or PD as best response to most recent prior therapy

PHASE 2 NHL DEMOGRAPHICS & DISEASE CHARACTERISTICS

6

Characteristic Follicular Lymphoma DLBCL

EZH2 Status Mutant Wild-­type Mutant Wild-­typen 13 54 17 120

Age, median years 62 61 61 69

Males 46% 63% 53% 58%

ECOG PS, median (range) 0 (0 -­ 2) 0 (0 -­ 2) 1 (0 -­ 2) 1 (0 -­ 2)

Prior lines of therapy, n (%) 1 1 ( 8%) 0 0 3 ( 3%)

2 2 (15%) 11 (20%) 4 (24%) 40 (33%)

3 3 (23%) 9 (17%) 7 (41%) 28 (23%)

4 1 ( 8%) 14 (26%) 3 (18%) 18 (15%)

≥ 5 6 (46%) 20 (37%) 3 (18%) 31 (26%)

median 4 4 3 3

Refractory to last regimen, n (%) 7 (54%) 26 (48%) 14 (82%) 75 (63%)

Prior HSCT 23% 41% 41% 24%

Median time from initial diagnosis years 7.4 4.9 1.0 2.0

Median time from last prior therapy weeks 13.0 41.3 8.6 11.6

PHASE 2 NHL DEMOGRAPHICS & DISEASE CHARACTERISTICS

Data as of 6/1/2017Refractory to last regimen defined as SD or PD as best response to most recent prior therapy

7

Characteristic Follicular Lymphoma DLBCL

EZH2 Status Mutant Wild-­type Mutant Wild-­typen 13 54 17 120

Age, median years 62 61 61 69

Males 46% 63% 53% 58%

ECOG PS, median (range) 0 (0 -­ 2) 0 (0 -­ 2) 1 (0 -­ 2) 1 (0 -­ 2)

Prior lines of therapy, n (%) 1 1 ( 8%) 0 0 3 ( 3%)

2 2 (15%) 11 (20%) 4 (24%) 40 (33%)

3 3 (23%) 9 (17%) 7 (41%) 28 (23%)

4 1 ( 8%) 14 (26%) 3 (18%) 18 (15%)

≥ 5 6 (46%) 20 (37%) 3 (18%) 31 (26%)

median 4 4 3 3

Refractory to last regimen, n (%) 7 (54%) 26 (48%) 14 (82%) 75 (63%)

Prior HSCT 23% 41% 41% 24%

Median time from initial diagnosis years 7.4 4.9 1.0 2.0

Median time from last prior therapy weeks 13.0 41.3 8.6 11.6

PHASE 2 NHL DEMOGRAPHICS & DISEASE CHARACTERISTICS

Data as of 6/1/2017Refractory to last regimen defined as SD or PD as best response to most recent prior therapy

8

TAZEMETOSTAT DEMONSTRATED FAVORABLE SAFETY PROFILETreatment-­EmergentAdverse Event

Patients (n=210) with:

All TEAEs Treatment-­Related TEAEs

All Grades Grade ≥3 All Grades Grade ≥3Nausea 42 (20%) 1 (<1%) 29 (14%) 0

Thrombocytopenia 39 (19%) 19 ( 9%) 28 (13%) 12 ( 6%)

Anaemia 33 (16%) 16 ( 8%) 21 (10%) 9 ( 4%)

Cough 30 (14%) 1 (<1%) 4 ( 2%) 1 (<1%)

Fatigue 26 (12%) 5 ( 2%) 15 ( 7%) 2 ( 1%)

Diarrhoea 24 (11%) 1 (<1%) 17 ( 8%) 1 (<1%)

Asthenia 22 (10%) 3 ( 1%) 16 ( 8%) 1 (<1%)

Neutropenia 21 (10%) 15 ( 7%) 19 ( 9%) 13 ( 6%)

Pyrexia 21 (10%) 1 (<1%) 2 ( 1%) 0

Vomiting 21 (10%) 2 ( 1%) 7 ( 3%) 1 (<1%)

Bronchitis 14 ( 7%) 0 2 ( 1%) 0

Constipation 13 ( 6%) 1 (<1%) 4 ( 2%) 1 (<1%)

Decreased appetite 13 ( 6%) 0 6 ( 3%) 0

Upper respiratory tract infection 13 ( 6%) 0 1 (<1%) 0

Abdominal pain 12 ( 6%) 3 ( 1%) 4 ( 2%) 0

Headache 12 ( 6%) 0 4 ( 2%) 0

Urinary tract infection 12 ( 6%) 0 4 ( 2%) 0

Back pain 11 ( 5%) 2 ( 1%) 1 (<1%) 0

Oedema peripheral 11 ( 5%) 2 ( 1%) 1 (<1%) 0

Dysgeusia 10 ( 5%) 0 7 ( 3%) 0

Rhinitis 10 ( 5%) 0 1 (<1%) 0Data as of 6/1/20171 One (1) TEAEs of Febrile Neutropenia Adverse events reported in ≥5% of patients

9

ADVERSE EVENTS LED TO LOW RATE OFDOSE REDUCTIONS AND DISCONTINUATIONS

Patients(n=210)

Treatment-­Emergent Adverse Events (TEAEs) *

Treatment-­Related TEAEs

Adverse Event (any) 190 (90%) 123 (59%)

Grade ≥ 3 91 (43%) 38 (18%)

Serious AE 81 (39%) 20 (10%)

AE Leading to Dose Interruption 50 (24%) 31 (15%)

AE Leading to Dose Reduction 8 ( 4%) 7 ( 3%)

AE Leading to Drug Discontinuationor Study Withdrawal 26 (12%) 5 ( 2%)

Data as of 6/1/2017* TEAEs are adverse events that first appear during treatment, which were absent before or which worsen relative to the pre-­treatment

10

TAZEMETOSTAT DEMONSTRATED HIGHER RESPONSE RATES IN EZH2 MUTATED NHL

Best ResponseFL

EZH2 MT(n=13)

FLEZH2 WT(n=54)

DLBCLEZH2 MT(n=17)

DLBCLEZH2 WT(n=119)

Objective Response Rate (CR + PR) 12 (92%) 14 (26%) 5 (29%) 18 (15%)

Complete Response (CR) 1 (8%) 3 (6%) 0 10 (8%)

Partial Response (PR) 11 (85%) 11 (20%) 5 (29%) 8 (7%)

Stable Disease 1 (8%) 23 (43%) 6 (35%) 22 (18%)

SD study drug ongoing 1 (8%) 12 (22%) 1 (6%) 4 (3%)

Progressive Disease 0 13 (24%) 6 (35%) 60 (50%)

No Data, Unknown (UNK) 0 4 (7%) 0 19 (16%)

Time to first Response (wks)median (range)

11.9(6.9 – 35.9)

15.2(8.1 -­ 32.1)

8.3(4.6 – 48.1)

8.5(5.3 – 24.7)

Data as of 06/1/2017Ongoing patients with Best Response of 'No Data, Unknown' are not included in this table

11

TUMOR REDUCTION IN FOLLICULAR LYMPHOMA

75% of patients experienced reduction of tumor burden

Data as of 6/1/201712

DURATION OF TUMOR RESPONSE IN FOLLICULAR LYMPHOMA

Data as of 6/1/2017

48% of patients remain on study

0 5 1510Months Since Treatment Initiation

13

TUMOR RESPONSE IN FL WITH MUTATED EZH2cobas® Test Y646F;; NGS Y646F;; ctDNA Y646F

14

Baseline

Week 24

68 y.o female

1999

Tazemetostat: week 36+

2017

ChlorambucilPrednisone R-­CHOP

2007 2012 20141998

PI3K/mTOR inhibitor

CR CR SD2016

PR at week16

DURATION OF TUMOR RESPONSE IN DLBCL

Data as of 6/1/2017

12% of patients remain on study

0 5 2010 15Months Since Treatment Initiation 15

TUMOR RESPONSE IN DLBCL WITH MUTATED EZH2cobas® Test Y646X;; NGS Y646H;; ctDNA Y646H

16

0 2 0 4 0 6 00

2 0 0 0

4 0 0 0

6 0 0 0

8 0 0 0

1 0 0 0 0

T im e p o s t t r e a t m e n t , w k

Lesio

n size, SPD

2003

Tazemetostat: week 60+

2016

R-­CHOPCarmustineEtoposideCytarabineMelphalan

RituximabEpratuzumab

1999 2005

hA20 Rituximab

2007 2012

LenalidomideObinutuzumab Bendamustine

2014

Baseline Week 24

CR PR PR PR PD SD2017

PR at week 48

61 y.o. male

RESPONSE AFTER INITIAL DISEASE PROGRESSION IN DLBCL NON-­GCB (PMBCL)

17

Screening Week 8 Week 16

0 8 1 6 2 40

5 0 0 0

1 0 0 0 0

1 5 0 0 0

T im e p o s t t re a tm e n t , w k

Les

ion size, SPD

36 y.o. male

• NGS analysis performed on archive tumor and circulating tumor DNA (ctDNA) for a subset (n = 92) of patients‒ Custom 62 gene panel includes common NHL somatic mutations‒ Responder (CR+PR) vs. Non-­Responder analyses‒ Details presented in Poster #154 (Blakemore et al.)

• Results: ‒ Positive and negative predictors for tazemetostat response (PR/CR) identified• Positive predictors = EZH2 & MYD88 activating mutations• Negative predictors = MYC, TP53 and HIST1H1E• EZH2 and MYD88 mutually exclusive in this cohort i.e. potential for independent mechanism of sensitivity to tazemetostat in these patients

‒ Detection of EZH2 mutations in ctDNA indicates potential for future potential use of plasma for patient identification

MOLECULAR PROFILING IDENTIFIES POTENTIAL TAZEMETOSTAT RESPONSE PREDICTORS

18

364

421

020406080100

MYD88 MT MYD88 WT

Patients, %

Non-­Responder Responder

3

637

18

020406080100

EZH2 MT EZH2 WT

Patients, %

Non-­Responder Responder

Data as of 6/1/2017

SUMMARY

19

• Tazemetostat shows efficacy in heavily pretreated relapsed/refractory FL and DLBCL‒ Follicular Lymphoma

• 92% ORR in patients with mutated EZH2• 26% ORR in patients with wild-­type EZH2 with 22% of patients on-­study in SD• 48% overall with treatment ongoing

‒ DLBCL• 29% ORR in patients with mutated EZH2• 15% ORR with 8% CRs in WT patients• Durable responses in both WT and mutants

• Late responses and conversion to CR in both subtypes

• Molecular profiling may help predict response

• Tazemetostat is safe‒ Low incidence of treatment-­related grade ≥3, mainly thrombocytopenia (6%) and neutropenia (6%)

• Enrollment of EZH2 mutated DLBCL and FL continues

20

ACKNOWLEDGEMENTS

We wish to thank all of the physicians, nurses, study staff, scientists, and

most of all, the patients and families who contributed to this study