Presenting author disclosures O.H.: Roche consultancy; BMS - PowerPoint PPT Presentation

Longitudinal effect of luspatercept treatment on iron overload and iron chelation therapy in adult patients with β -thalassemia in the BELIEVE trial Olivier Hermine, 1,2 Maria Domenica Cappellini, 3 Ali T . Taher, 4 Thomas D. Coates, 5,6 Vip Viprakasit, 7 Ersi Voskaridou, 8 Ashutosh Lal, 9 Hong Keng Liew, 10 S ilverio Perrotta, 11 Abderrahim Khelif, 12 Antonis Kattamis, 13 hetty, 14 George Zhang, 15 Yu (Olivia) Tian, 15 Dimana Miteva, 14 Tatiana Zinger, 14 Derek Tang, 15 Jeevan K. S . Backstrom, 16 John B. Porter 17 Jay T 1 Imagine Inst it ut e, INS ERM U1163, Universit y of Paris, Paris, France; 2 Depart ment of Hemat ology, Necker Hospit al, Assist ance Publique-Hôpit aux de Paris, Paris, France; 3 Fondazione IRCCS Ca’ Granda Policlinico Hospit al, Universit y of Milan, Milan, It aly; 4 Depart ment of Int ernal Medicine, American Universit y of Beirut Medical Cent er, Beirut , Lebanon; 5 Children's Cent er for Cancer and Blood Diseases, Children's Hospit al Los Angeles, Los Angeles, CA; 6 US chool of Medicine, Los Angeles, CA; 7 S iriraj Hospit al, Mahidol University, Bangkok, Thailand; 8 Thalassemia and S C Keck S ickle Cell Cent er of Laiko General Hospit al, At hens, Greece; 9 Universit y of California S an Francisco, Benioff Children’s Hospit al, Oakland, CA; 10 Hospit al S ult anah Bahiyah, Alor S et ar, Malaysia; 11 Universit à della Campania, Luigi Vanvit elli, Casert a, It aly; 12 Farhat Hached Teaching Hospit al, S ousse Universit y, Tunisia; 13 First Depart ment of Pediatrics, Nat ional and Kapodist rian University of At hens, At hens, Greece; 14 Celgene Int ernat ional, a Brist ol-Myers S quibb Company, Boudry, S witzerland; 15 Brist ol Myers S quibb, Princet on, NJ; 16 Acceleron Pharma, Cambridge, MA; 17 Universit y College London, Universit y College London Hospit als, London, UK Present at ion 1697

Presenting author disclosures O.H.: Roche – consultancy; BMS – consultancy, research funding; AB S cience – consultancy, current equity holder in publicly-traded company, honoraria, patents, royalties, research funding; Alexion, Novartis - research funding. 2

Introduction Figure 1. BELIEVE study trial design • RBC transfusions are the main supportive treatment Patients with β -t halassemia a ≥ 18 years of age, requiring regular RBC for chronic anemia due to β -thalassemia 1 t ransfusions (defined as: 6– 20 RBC unit s in t he 24 weeks prior t o randomization wit h no > 35-day t ransfusion-free period during t hat t ime) – Transfusion-dependent pat ient s require ICT t o prevent (N = 336) iron overload from RBC t ransfusions and associat ed Randomized 2:1 Double-blind period (48 weeks) complicat ions 2 • Thus, there is a clinical need to reduce transfusions Luspatercept b Placebo b 1 mg/ kg s.c. every 21 days and iron burden in patients with anemia due to s.c. every 21 days + BS C β -thalassemia + BS C (n = 224) (n = 112) • Luspatercept, a first-in-class erythroid maturation agent, is approved by the US Food and Drug S t udy Crossover from placebo to Administration for the treatment of anemia in adult unblinding luspatercept permitted (n = 92) patients with β -thalassemia requiring regular RBC transfusions 3– 6 Open-label Current study (up t o 5 years) status c • This analysis assessed the effect of long-term luspatercept use on iron loading and ICT use in the Post -t reat ment phase 3 BELIEVE trial (NCT02604433) 7 follow-up (3 years) a β - thalassemia or Hb E / β - thalassemia (compound β - thalassemia with mutation and/or multiplication of α -globin genes was allowed). b Patients could receive RBC transfusions to maintain their baseline Hb level and ICT. c The trial is fully enrolled and patients continue to receive treatment or follow-up. BSC, best supportive care; Hb, hemoglobin; ICT, iron chelation therapy; RBC, red blood cell; s.c., subcutaneously. 1. Taher AT, et al. Lancet 2018;391:155– 167. 2. Taher AT, Cappellini MD. Blood 2018;132:1781– 1791. 3. Suragani RNVS, et al. Blood 2014;123:3864– 3872. 4. Attie KM, et al. Am J Hemat ol 2014;89:766– 770. 5. Piga A, et al. Blood 2019;133:1279– 1289. 6. Reblozyl (luspatercept-aamt) [package insert]. Summit, NJ: Celgene Corporation; April 2020. 7. Cappellini MD, et al. 3 N Engl J Med 2020;382:1219– 1231.

Results Table. Patient baseline characteristics Luspatercept Placebo Characteristic (N = 224) (N = 112) Age, median (range), years 30 (18– 66) 30 (18– 59) Female, n (% ) 132 (58.9) 63 (56.3) Hb (24 weeks), a median (range), g/dL 9.31 (4.5– 11.4) 9.15 (5.8– 11.7) RBC transfusion burden, median (range), units/12 weeks 6.1 (3– 14) 6.3 (3– 12) RBC transfusion burden, median (range), units/24 weeks 14 (6– 24) 15 (6– 26) Splenectomy, n (% ) 129 (57.6) 65 (58.0) SF, mean (SD), μg/L 2,097 (1,757) 1,845 (1,669) LIC, mean (SD), mg/g dw 12.0 (14.8) 10.1 (11.5) 154 (68.8) 75 (67.0) > 3 mg/ g dw, n (% ) Myocardial iron by T2* MRI, mean (SD), ms 33.5 (16.2) 34.8 (10.7) ICT use b,c , n (% ) Deferasirox 139 (62.3) 63 (57.8) Deferiprone 92 (41.3) 40 (36.7) Deferoxamine mesylat e / deferoxamine 83 (37.2) 39 (35.8) a Defined as the mean of all documented pre-transfusion Hb values during the 24 weeks prior to first dose for each patient. b Defined as started before the start of study treatment and either ended before the start of the study treatment or continued after study treatment. c Analyzed using the safety population (luspatercept n = 223, placebo n = 109). 4 dw, dry weight; LIC, liver iron concentration; SD, standard deviation; SF, serum ferritin; T2* MRI, T2-weighted magnetic resonance imaging.

Results (cont.) Figure 2. Decrease in SF categories baseline ≥ 1,000 μ g/ L to post-baseline < 1,000 μ g/ L (A) and baseline ≥ 2,500 μ g/ L to post-baseline < 2,500 μ g/ L (B) A B Baseline SF ≥ 1,000 μ g/L to post-baseline < 1,000 μ g/L Baseline SF ≥ 2,500 μ g/L to post- baseline < 2,500 μg/L 100 Placebo a Luspat ercept 90 80 +57.6% +40.5% 70 ) Patients (% +24.3% +24.6% 60 +54.8% +39.3% 46.4 (26/ 56) 50 P = 0.017 33.3 40 P = 0.227 P = 0.023 29.5 (7/ 21) P = 0.527 (38/ 129) 23.7 23.7 30 19.0 (32/ 135) (14/ 59) 17.0 (12/ 63) 12.3 (24/ 141) 20 8.8 8.3 7.7 (8/ 65) 5.0 (5/ 57) (2/ 24) (2/ 26) 10 (3/ 60) 0 Weeks Weeks Weeks Weeks Weeks Weeks Weeks Weeks 1– 24 25– 48 49– 72 73– 96 1– 24 25– 48 49– 72 73– 96 Data cutoff: July 1, 2019. P values are estimated from Cochran-Mantel-Haenszel test. SF levels ≥ 1,000 μg/L and ≥ 2,500 μg / L indicate iron overload and increased risk of cardiac-related mortality, respectively. 5 a Placebo patients evaluated up to Week 48.

Results (cont.) Figure 3. Decrease in LIC (A) and myocardial iron T2* categories (B) A B LIC Myocardial iron T2* Baseline > 3 mg/g dw to post- baseline ≤ 3 mg/g dw Baseline ≤ 20 ms to post -baseline > 20 ms Placebo a Luspat ercept 14.3 16 30 25.0 (15/ 105) P = 0.417 P = 0.357 (6/ 24) 14 25 20.0 12 9.7 (6/ 30) P = 0.486 20 (13/ 134) ) ) Patients (% Patients (% 10 6.6 8 15 5.9 (4/ 61) (4/ 68) 9.1 4.2 6 (1/ 11) 10 (5/ 120) 4 5 2 0 0 Week 24 Week 48 Week 96 Week 48 Week 96 Data cutoff: July 1, 2019. Haenszel test. Patients with LIC ≥ 3 mg/g dw are considered to have iron overload. Myo cardial iron T2* < 20 ms indicates increased cardiac risk. P values are estimated from Cochran– Mantel– 6 a Placebo patients evaluated up to Week 48.

Results (cont.) Figure 4. Luspatercept-treated patients receiving ≥ 1 ICT over time a • The mean (S D) daily dose of deferasirox in all Luspat ercept Luspat ercept Luspat ercept responders b (overall) non-responders luspatercept-treated patients at baseline was 100 1,332.37 mg (1,087.0) • During Weeks 85– 96, the mean change (S D) in 90 daily dose of deferasirox from baseline in the ) Patients (% luspatercept arm overall (n = 118), responders (n = 32), and non-responders (n = 86) was 80 −189.8 mg (768.5), −201.1 mg (740.2), and −185.6 mg (783.0), respectively 70 • The mean change (S D) in the luspatercept arm overall at Week 144 (n = 25) was −810.1 mg 60 (1,614.8) Weeks Overall 219 217 207 200 195 186 169 161 148 142 133 92 28 Responders 46 46 46 45 44 42 41 42 41 38 37 27 8 Non-responders 173 171 161 155 151 144 128 119 107 104 96 65 20 Data cutoff: July 1, 2019. Only a small number of patients could be evaluated at the later time points due to poor tracking and ICT adherence. a Includes only patients initially randomized to receive luspatercept. b Responders are defined as patients achieving ≥ 33% reduction in transfusion burden during Weeks 13 – 24, with a reduction of ≥ 2 RBC units, versus baseline. 7

S ummary • A significant number of luspatercept- treated patients with baseline SF ≥ 1,000 μg/L shifted to SF < 1,000 μg/L during the first 48 weeks • The proportion of patients receiving ≥ 1 ICT gradually declined in both luspatercept responders and non-responders over time up to Week 144 • Both luspatercept responders and non-responders experienced decreases in mean daily dose of deferasirox over time 8