Results from the Completed Dose-Escalation Phase I SHRINK

Poster 331 Results from the completed dose-escalation phase I SHRINK study evaluating the autologous NKG2D-based CAR T-cell therapy CYAD-01 in metastatic colorectal cancer patients

Leila Shaza 1, Alain Hendlisz 1, Ahmad Awada 1, Jean‐Luc Canon 2, Javier Carrasco 2, Eric Van Cutsem 3, Jeroen Dekervel 3, Erik Alcantar‐Orozco 4, Florence Renard 4, Emilie Cerf 4, Caroline Lonez 4, Anne Flament 4, Jean‐Pascal H. Machiels 5, Marc Van Den Eynde 5 1. Institut Jules Bordet, Université Libre de Bruxelles, Brussels, Belgium; 2. Grand Hôpital de Charleroi (GHdC), Charleroi, Belgium; 3. University Hospital Leuven (UZ Leuven), Leuven, Belgium; 4. Celyad, Mont‐Saint‐Guibert, Belgium; 5. Cliniques Universitaires Saint‐Luc, Université Catholique de Louvain, Brussels, Belgium BACKGROUND TABLES & FIGURES MAIN RESULTS

● Colorectal cancer (CRC) is the third most common type of cancer among both men and Table 1: Patient, tumor and treatment characteristics ● In total 9mCRCpatientshave been enrolled in this Phase I study (3 at DL1, 3 at TOTAL DL2 and 3 at DL3) (Table 1). women worldwide and is the fourth in terms of mortality. According to National Institutes Patient, tumor and treatment characteristics Figure 1: Change in tumor burden from baseline N=9 of Health data, only 40% of CRC are diagnosed with localized-stage disease, when the 5- ● At the study snapshot (Oct 17, 2019), only one patient experienced a Grade 3 Mean age and range (year) 56 (28-73) Neoadjuvant 1st-line mCRC a Refractory mCRC year survival rate is close to 90%. At the advanced stage and/or when the cancer has Male / Female 4 (44.4%) / 5 (55.6%) related adverse event (AE) and no patient experienced Grade ≥ 4 related AEs invaded distant organs, 5-year survival rate drops till 13%. ECOG Grade 0 / 1 6 (66.7%) / 3 (33.3%) 90 (Table 2 and 3). There was no report of dose-limiting toxicity (Table 2). Sites of the target lesions 80 PD 1x108 cells/inf. ● Neoadjuvant chemotherapy is often used before resection in CRC patients with liver Liver a 7 (77.8%) a Post-baseline tumor evaluation before ● Patients in the neoadjuvant setting received one cycle of 3 infusions of 70 3x108 cells/inf. Lymph node 3 (33.3%) surgery [neoadjuvant setting] metastases considered as resectable with the aim to treat micro-metastases and reduce 60 CYAD-01inassociationwithFOLFOXasper protocol. Four patients out of the Lung 3 (33.3%) 1x109 cells/inf. the tumor volume for surgical resection. 50 PR = partial response; Rectum 1 (11.1%) five refractory mCRC setting who were not in PD at first tumor assessment 40 SD = stable disease; ● For refractory metastatic CRC (mCRC) patients, rechallenge with the same chemotherapy Skin 1 (11.1%) SD PD PD = progressive disease; received a second cycle of treatment with CYAD-01 (Figure 2). 30 according to modified RECIST criteria 1.1. agents can provide a continuum of treatment although it may lack of efficacy in patients Number of target lesions st 1 1 (11.1%) 20 mCRC = metastatic colorectal cancer ● In the neoadjuvant 1 line mCRC population (n=4), 1 partial response (PR) and 2 who have already progressed after a similar regimen. 2 to 3 6 (66.7%) 10 patients with stable disease (SD >3 months) were observed (Table 4 and 2). SD PR SD SD SD SD ● Chimeric antigen receptor T-cell (CAR T-cell) therapy has delivered major clinical results ≥ 4 2 (22.2%) 0 Prior adjuvant treatment 2 (22.2%) -10 ● In the refractory mCRC population (n=5), 4 patients presented SD >3 months

in B-cell malignancies. However, this immunotherapy has yet to deliver significant objective % change baseline from Number of prior metastatic treatment lines in sum of longest diameter -20 (Table 4 and Figure 2) with 3 out of 4 SD patients showing some evidence of a responses in patients with advanced solid tumors. 0 4 (44.4%) -30 tumor burden decrease (Figure 1). 1 2 (22.2%) ● CYAD-01 is a multi-complex, second-generation NKG2D CAR T-cells comprising the -40 2 to 3 0 (0%) -50 ● CYAD-01 cells are detected in the peripheral blood at least up to 40 days after ≥4 3 (33.3%) st human natural killer group 2D (NKG2D) receptor fused to the intracellular domain of CD3ζ. -60 1 infusion (Figure 3). Detailed discussion of cell kinetics is presented in poster Mean (range) 3 (1-6) Evaluable mCRC patients (N=9) ● NKG2D targets 8 stress ligands found at high frequency across a range of solid tumors Number of prior metastatic lines containing oxaliplatin P147. including metastatic CRC. Interestingly, other non-malignant stromal cells within the 0 6 (66.7%) a tumor also express NKG2D ligands suggesting that the mechanism of action of CYAD-01 1 1 (11.1%) 2 2 (22.2%) Figure 2: Time to response and duration of treatment likely involves targeting of malignant and non-malignant cells within the tumor CYAD-01 FOLFOX Number of prior metastatic lines containing molecular-targeting agents CONCLUSIONS microenvironment which, in pre-clinical models, resulted in a broader anti-tumor response. None 6 (66.7%) a Best overall response DL-1: 1x108 cells/inf. EGFR targeted therapy 1 (11.1%) 8 ● CYAD-01 has been tested as a monotherapy without pre-conditioning chemotherapy and Disease progression DL-2: 3x10 cells/inf. Bevacizumab 2 (22.2%) 9 showed initial clinical responses including transient complete remissions in Surgery DL-3: 1x10 cells/inf. ● The SHRINK study evaluates the autologous CYAD-01 with prior FOLFOX a Prior metastatic agents 4 patients were enrolled in the neoadjuvant setting as 1st-line mCRC with resectable a in advanced metastatic colorectal cancer patients. relapsed/refractory acute myeloid leukemia patients (THINK study – NCT03018405) [1]. liver-dominant metastases.

FX Refractory mCRC FiRi Bev FiRiX Pmab Cetux ● This combination has been found to have a favorable safety profile at all 1. Sallman DA et al. (2018), Haematologica 103: e242-e246. Others Table 2: Incidence of adverse events (AEs) Y b Y b ---YYb SD SD doses evaluated with no dose-limiting toxicities reported. TOTAL -Y-----SD SD N=9 (39 infusions) ● No cumulative toxicities were noted upon multiple administrations of All AEs All treatment-related AEs Y b Y---YYSD CYAD-01 with concurrent FOLFOX treatment. SHRINK STUDY Any Grade 164 (100%) 16 (100%) Grade ≥ 2 60 (36.6%) 2 (12.5%) YYb --Y-Y b SD ● Based upon this safety profile, a dose of 1x109 CYAD-01 treatment is Grade ≥ 3 11 (6.7%) 1 (6.3%) -Y----- ● The Phase I SHRINK study (NCT03310008) evaluates the autologous CAR T-cell product Grade 4 0 (0%) 0 (0%) PD considered safe for further studies combining with FOLFOX chemotherapy CYAD-01, administered concurrently with FOLFOX (Folinic acid (leucovorin), Fluorouracil Dose limiting toxicity 0 (0%) 0 (0%) Serious adverse events 1 (0.6%) 0 (0%) Neoadjuvant 1st-line mCRC ● We await further information from tumor analysis collected from patients (5-FU) and Oxaliplatin) chemotherapy. AEs leading to treatment ------PR undergoing surgery in the neo-adjuvant group. ● The objective of this study is to evaluate if the FOLFOX chemotherapy can ensure the discontinuation 0 (0%) 0 (0%) PR engraftment of the CYAD-01 CAR T-cells, without impacting the CYAD-01 safety profile, ------SD SD ● Given the small sample size of the populations in this dose-escalation and potentially improve clinical outcome. Table 3: Patient incidence of treatment-related AEs ------SD SD study, it is not possible to dissect the contribution of CYAD-01 and TOTAL ------PD FOLFOX to clinical responses. However, in the refractory setting, achieving ● The patient population includes two different mCRC patient populations: Treatment related AEs N=9 (39 infusions) All Grade ≥ 3 SD in patients who previously received oxaliplatin-based therapies st 0w 4w 8w 12w 16w 20w 24w 28w o Resectable liver dominant mCRC with FOLFOX chemotherapy as 1 line metastatic Total pts with at least 1 related AE (%) 6 (66.7%) 1 (11.1%) FOLFOX (3/4 patients) and FOLFIRI (4/4 patients) provides treatment (ie, neoadjuvant population), and Pyrexia 2 (22.2%) 0 (0) Anemia 1(11.1%) 1(11.1%) encouragement for the therapeutic combination of CYAD-01 and Cytokine release syndrome 2 (22.2%) 0 (0) a Other metastatic treatment lines: Irinotecan drug-eluting embolization beads, trifluridine/tipiracil, aflibercept, undisclosed Phase I agent FOLFOX. o Non-resectable mCRC with prior chemotherapy lines for mCRC including oxaliplatin b Infusion site reaction 1 (11.1%) 0 (0) ≥ 2 metastatic lines FX = FOLFOX; FiRi = FOLFIRI; FiRiX = FOLFIRINOX; Cetux = Cetuximab; Pmab = Panitumumab; Bev = Bevacizumab and irinotecan-based chemotherapy (ie, rechallenge population or refractory mCRC). Atrial tachycardia 1 (11.1%) 0 (0) PR = Partial response; SD = Stable disease; PD = Progressive disease ● The results from this study will be compared to its sister study evaluating Fatigue 1 (11.1%) 0 (0) ● The study design: International normalized ratio the allogeneic analog of CYAD-01 in mCRC that combine to support the increased 1 (11.1%) 0 (0) development of CAR T-cell therapy in solid tumors (poster 330). o Dose escalation with a Fibonacci 3+3 design with three dose levels (DL) of the NKG2D CAR T-cells: 1x108,3x108 and 1x109 cells per infusion. Figure 3: CYAD-01 kinetics in the peripheral blood o Three intravenous CAR T-cell infusions at a 2-week interval administered at Day 3 of 1x108 cells/infusion 3x108 cells/infusion 1x109 cells/infusion three consecutive FOLFOX chemotherapy cycles. Table 4: Best tumor responses (3 patients) (3 patients) (3 patients) ACKNOWLEDGEMENTS & DISCLOSURES TOTAL TOTAL o Potential new cycle of treatment (3x 3 infusions) in case of no PD at first scheduled Clinical outcome Neoadjuvant Refractory mCRC tumor assessment . N = 4 N=5 ● Celyad thanks patients & their families, physicians, and study teams at all o Primary endpoint of this study is the occurrence of dose-limiting toxicity (DLT) during Best overall responses – n (%) participating centers. Partial response (PR) 1 (25%) the treatment phase. Key secondary endpoints include cell kinetics, objective Stable disease (SD) 2 (50%) 4 (80%) ● This study was funded and sponsored by Celyad SA (ClinicalTrials.gov identifier: responses and duration of responses. Progressive disease (PD) 1 (25%) 1 (20%) NCT03310008). Objective response rate a - % 25% ● In this context, FOLFOX should help to: Disease control rate b - % 75% 80% ● LS, AH, AA, JLC, JC, EVC, JD, JPHM and MVDE are Investigators on the SHRINK trial. EAO, FR, EC, CL and AF are employed by Celyad SA. o Control of tumor progression during the CYAD-01 manufacturing period (3 weeks), a Evaluation of clinical response is based on modified RECIST 1.1 criteria o Facilitate CAR T-cell engraftment by inducing a transient lymphodepletion, and b PR or SD ● This poster is published for information only. o Further increase the NKG2D ligand expression on tumor tissues. The views expressed are those of the authors and not necessarily those of the organizations named herein.