(P 109) INTENSE PRECLINICAL ACTIVITY BY B7H3-SPECIFIC CAR.CIK LYMPHOCYTES AGAINST EPITHELIOID SARCOMA

Objective: Epithelioid sarcoma is an ultra-rare mesenchymal tumor characterized by either genetic or epigenetic loss of INI1/SMARCB1 function. This event leads to constitutive EZH2 activation, promoting tumorigenesis epigenetically. In advanced stages, epithelioid sarcoma is characterized by dismal prognosis and limited therapeutic options. The advent of EZH2 inhibitors has revived hope for advanced epithelioid sarcoma; however, as seen with the majority of targeted treatments, resistance typically ensues. Hence, epithelioid sarcoma represents a relevant unmet medical need; reliable preclinical studies for the development of innovative, effective treatments are awaited. Purpose of this study is to explore and define a preclinical model for a novel CAR-based cellular immunotherapy against epithelioid sarcoma, including the possible targeting of sarcoma cell subsets with stemness features that are emerging as relevant components for chemoresistance and disease relapse in several solid tumors. The strategy is focused on Cytokine-induced Killer lymphocytes (CIK), ex vivo expanded T-NK lymphocytes endowed with intrinsic HLA-independent anti-tumor activity (NKG2D mediated). CIK will be redirected with a chimeric antigen receptor (CAR) against B7 homolog 3 protein (B7-H3), an emerging target expressed by multiple solid tumors and associated with immunosuppressive functions and stemness.

Methods: The experimental setting is based on patient-derived CAR.CIK and 2 proximal epithelioid sarcoma (EpS) cell lines (ES1, ES2) with confirmed loss of INI1/SMARCB1, derived from two patients affected by advanced EpS. CAR.CIK were generated by engineering circulating PBMC precursors with a retroviral vector encoding for a 2nd generation anti-B7-H3 CAR with CD28 co-stimulatory domain and ex vivo expanded (3 weeks) according to a standard protocol. Tumor killing activity by CAR.CIK was assessed by coculture in vitro (48h) and analyzed by flow cytometry. To visualize and track the fate of putative sarcoma cancer stem cells (sCSCs), EpS cells were engineered with a lentiviral CSC-detector wherein the promoter of stem-gene Oct4 controls the expression of eGFP, previously validated in sarcomas (Sangiolo et al, Cancer Research 2014; Oncoimmunology 2018).

Results: We effectively ex vivo generated CAR.CIK from patients (n=3). The membrane expression of anti-B7H3 CAR ranged from 50 to 80%. The main phenotype of mature CAR.CIK was:CD3+CD56+=30% CD3+CD8+: 70%, NKG2D= 60%, PD-1=15%), comparable to paired unmodified control (NTD.CIK).
The CAR-target B7-H3 was confirmed to be intensely expressed by our ES cell lines (85-90%), along with NKG2D ligands (MIC A/B >80%), while a modest expression of both PD-L1 (10-20%) and PD-L2 (10-15%) was detected. The tumor killing activity in vitro of anti-B7-H3 CAR.CIK against ES was intense and significantly superior as compared with control NTD-CIK, especially at very low effector/target (E/T) ratios, ideally more representative of hypothetical clinical scenarios (95% vs 68% E/T=10:1; 70% vs 20% E/T= 1:1, 60% vs 2% E/T=1:4, p< 0.005 n=4). Within ES cells, we could visualize a subset of putative GFP+ sarcoma CSCs (8-10% n=2) that were equally killed by B7H3 CAR.CIK.

Conclusion: We report first proof of concept of a CAR-based cellular immunotherapy against epithelioid sarcomas, including a subset of putative CSCs, underscoring B7-H3 as a relevant target in this challenging, ultra-rare soft tissue sarcoma histotype.