GBR 1302: EFFECT OF CD3-HER2, A BISPECIFIC T CELL ENGAGER ANTIBODY, IN TRASTUZUMAB-RESISTANT CANCERS JONATHAN BACK 1 ■ MARTIN WERMKE 2 ■ JULIE MACOIN 1 ■ AMELIE CROSET 1 ■ JOHN KAUH 3 ■ VENKATESHWAR REDDY 3 1 GLENMARK PHARMACEUTICALS SA, LA CHAUX-DE-FONDS, SWITZERLAND ■ 2 UNIVERSITY HOSPITAL CARL-GUSTAV-CARUS, DRESDEN, GERMANY 3 GLENMARK PHARMACEUTICALS INC, PARAMUS, NEW JERSEY, USA INTRODUCTION ◾ Although current therapies targeting HER2-overexpressing cancers have proven beneficial, including trastuzumab (Herceptin®) and ado-trastuzumab emtansine (T-DM1; Kadcyla®), patients often progress on therapy due to primary and/or acquired resistance mechanisms 1,2 ◾ Immune checkpoint inhibitors have demonstrated the potential of mobilizing T cells to elicit anti-tumor responses, but these tumor-specific immune responses are highly context-dependent 3 ◾ GBR 1302 is a novel HER2xCD3 bispecific antibody engineered (using the Glenmark Bispecific Engagement by Antibodies based on the T cell receptor [BEAT®] platform) as an alternative way to leverage T cell potency against tumor cells independent of existing tumor immune response (Figure 1 ) • GBR 1302 directs T cells to HER2-expressing tumor cells by simultaneously engaging the CD3 molecule on T lymphocytes and the HER2 antigen on tumor cells (HER2 2+ or 3+ overexpression), thereby killing the bound target cells through redirected lysis Figure 1. GBR 1302 BEAT® Design and Redirected Lysis Fab arm with anti-CD3ε specificity Single-chain variable fragment arm with anti-HER2 specificity Redirected Lysis CD3+ T Cell Tumor Cell a Based on trastuzumab. Fab, fragment antigen binding. METHODS/RESULTS Therapeutic Window of GBR 1302 ◾ GBR 1302 triggers potent killing of HER2 positive (IHC3+) and HER2 equivocal (IHC2+) cancer cells, including the trastuzumab-resistant JIMT-1 cell line, while maintaining an acceptable therapeutic window (up to 1000-fold greater) on cells expressing normal levels of HER2 (Figure 2) • Interdonor variability of the therapeutic window was small • Potency on IHC3+ and IHC2+ overlapped Figure 2. Therapeutic Window of GBR 1302 EC 50 (ng/mL) 10 100 1000 1 0.1 0.01 0.001 Cardiomyocytes MDA-MB-231 HT-1080 MDA-MB-175 JIMT-1* T-47D ZR-75-1 MDA-MB-453 SNU-216 BT-474 SK-Br-3 NCI-N87 Interdoner Variability (n=8 to 27) Not Characterized IHC0 IHC1+ IHC2+ IHC3+ HercepTest® Rank Cell lines were characterized using the HercepTest®. Cardiomyocytes, T-47D cells, and ZR-75-1 cells were not characterized by HercepTest. HercepTest rank for primary cardiomyocytes was set arbitrarily as zero. Ratio of effector (healthy donor PBMC):target cells, 10:1. T-47D and ZR-75-1 cells had HER2 levels by flow cytometry (specific antibody binding capacity) corresponding to IHC1/2+ and IHC2+, respectively. *JIMT-1 cells are IHC2+ but due to masking of HER2 epitope are equivalent to IHC1+ for binding to trastuzumab. IHC, immunohistochemistry; PBMC, peripheral blood mononuclear cell. Cytotoxic Potential of GBR 1302 Versus Standard-of-Care Therapies ◾ GBR 1302 displayed consistently superior killing of HER2-overexpressing target cells versus trastuzumab or T-DM1 independent of cell cycle, proliferation, and antibody-dependent cellular cytotoxicity ◾ In a trastuzumab-resistant model, GBR 1302 demonstrated potent tumor growth inhibition (Figure 3) Figure 3. Cytotoxic Potential of GBR 1302: In Vivo Tumor Model JIMT-1 + PBMC Tumor Volume (mm 3 ) Breast Carcinoma Days 1600 1400 1200 1000 800 600 400 200 0 0 10 20 30 40 50 60 Trastuzumab-Resistant Model (Subcutaneous) GBR 1302 JIMT-1 JIMT-1 + hPBMC JIMT-1 + hPBMC + GBR 1302 Human trastuzumab-resistant breast carcinoma cells (JIMT-1; IHC2+ but 1+ for binding to trastuzumab) were grafted subcutaneously with human PBMCs into immunodeficient mice. Mice were treated intravenously with GBR 1302 (0.05 mg/kg) three times per week beginning on Day 0 for 6 total doses (black arrows). n=20 mice with PBMCs from 3 donors were used per condition. Ratio of effector (healthy donor PBMC):JIMT-1 cells, 1:1. IHC, immunohistochemistry; PBMC, peripheral blood mononuclear cell. Response Rates by HER2 Status in Metastatic Breast and Gastric Cancer ◾ Metastatic breast cancer patients with IHC2+ and 3+ responded favorably to single-agent GBR 1302 and to combination with anti-PD1, compared with trastuzumab ( Figure 4 ) • There was an overall predictive response rate of 20% to single-agent GBR 1302 and 36% to combination therapy of GBR 1302 with anti-PD1 in metastatic breast and gastric cancer patients Figure 4. Primary Human Ex Vivo (Canscript) Studies to Identify Responders Rx2 80% 20% Rx3 88% 12% Rx4 64% 36% Number of Tumors 100 80 60 40 20 0 Responder Non-Responder Predictive responders (treatment outcome) are based on M-score that takes into account multiple input parameters for the given tumor specimen. Positive prediction of response: M-score >15; negative prediction of response: M-score <25. Rx2, GBR 1302; Rx3, trastuzumab; Rx4, GBR 1302 + anti-PD1. HER2 Status Number of Samples Number of Responders (% Responders) GBR 1302 Trastuzumab GBR 1302 + anti-PD1 Negative 4 0 0 1 (25%) 1+ 6 0 0 1 (16.7%) 2+ 8 3 (37.5%) 1 (12.5%) 5 (62.5%) 3+ 7 2 (28.5%) 1 (14.3%) 3 (42.8%) Total 25 5 (20%) 2 (8%) 10 (40%) HER2 Status Number of Samples Number of Responders (% Responders) GBR 1302 Trastuzumab GBR 1302 + anti-PD1 Negative 14 2 (14.3%) 2 (14.3%) 5 (35.7%) 1+ 4 1 (25%) 1 (25%) 1 (25%) 2+ 3 0 0 0 3+ 4 2 (50%) 1 (25%) 2 (50%) Total 25 5 (20%) 4 (16%) 8 (32%) Immune Cell Signatures of Tumors ◾ Immune signatures of breast cancers differed from gastric cancers (Figure 5) • In some instances, HER2 status tumors clustered together or close to each other • Diverse relative abundance of immune cells was observed at a subset level between samples • Relative to control mAb (Rx1), these signatures increased under the pressure of GBR 1302 (Rx2) and GBR 1302 + anti-PD1 (Rx4) compared with trastuzumab (Rx3) • The overall immune signature of the tumors changed from cold (blue) to hot (red) when treated with GBR 1302 Figure 5. Immune Cell Signatures of Metastatic Breast Cancer and Gastric Cancer Hot Cold Hot Cold Rx1: Control mAB Rx2: GBR 1302 Rx3: Trastuzumab Rx4: GBR 1302 + anti-PD1 TIS and GAJ are two published RNA-based signatures with relevance to clinical response to anti-PD1 in melanoma and head and neck squamous cell carcinoma. CaBr, breast cancer; DC, dendritic cell; GS, gastric cancer; mAb, monoclonal antibody. Difference in Immune Signatures with GBR 1302 ◾ CXCL9, 10, and 11 genes were regulated by IFNγ and DE genes in GBR 1302 predictive responders (Figure 6) ◾ Expression of potential immune resistance mechanism genes (eg, LAG3, IDO, CTLA4) was observed Figure 6. GBR 1302 Versus Control Immune Signatures mAb, monoclonal antibody; Rx1, control mAB; Rx2, GBR 1302; Rx3, trastuzumab; Rx4, GBR 1302 + anti-PD1; Trt, treatment. CONCLUSIONS ◾ GBR 1302 demonstrates potent killing of HER2 positive (IHC3+) and HER2 equivocal (IHC2+) cancer cells ◾ GBR 1302 displays superior cytotoxic potential versus standard-of- care therapies in multiple in vitro assays and in vivo tumor models, including in trastuzumab-resistant cells ◾ Translational studies identify predictive responders to GBR 1302 and to the combination of GBR 1302 with anti-PD1 • Expansion of effector and memory T cells was observed in predictive responders in Canscript studies • Pro-inflammatory, IFNγ responsive genes are upregulated in predictive responders ◾ GBR 1302 is currently in a phase 1 dose escalation clinical trial in HER2 positive and equivocal cancers (NCT02829372) • Preliminary data from peripheral blood biomarkers indicate that GBR 1302 triggers relevant T cell activation and cytokine production 4 REFERENCES 1. Connell CM and Doherty GJ. ESMO Open. 2017;2:e000279. 2. Yu S, Liu Q, Han X, et al. Exp Hematol Oncol . 2017;6:31. 3. Li X, Shao C, Shi Y, Han W. J Hematol Oncol. 2018;11(1):31. 4. Wermke M, Alt J, Kauh J, et al. Poster presented at: ASCO-SITC Clinical Immuno-Oncology Symposium; January 25-27, 2018; San Francisco, CA. DISCLOSURES This study was funded by Glenmark Pharmaceuticals SA. Medical writing and editorial assistance were provided by Prescott Medical Communications Group, Chicago, IL. Copies of this poster obtained through Quick Response (QR) Code are for personal use only and may not be reproduced without permission from ASCO® and the author of this poster. Corresponding author: Venkat Reddy, Venkat.Reddy@glenmarkpharma.com C POSTER PRESENTED AT THE 54TH ANNUAL MEETING OF THE AMERICAN SOCIETY OF CLINICAL ONCOLOGY | JUNE 1-5, 2018 | CHICAGO, IL 12053