Cell Adhesion of ALL to Stromal Cells May Mediate CAR T-Cell Resistance: A Novel Escape Mechanism for Immunotherapy

Abstract

Backgroud: Despite advances in therapy and improved survival, relapsed and refractory B-cell precursor acute lymphoblastic leukemia (r/r BCP-ALL) in pediatric and adult patients still remains a problem. Chimeric antigen receptor T cells against CD19 (CD19 CAR T) show promising results in patients with r/r BCP-ALL. However, relapse of the disease still occurs with appreciable frequency even with this novel therapy. As a significant number of relapses post-CAR T lack surface CD19 expression, further CD19-directed therapy is not an option for these cases. Hypothesis: Sometimes despite CAR T engraftment and establishment of B-cell aplasia, relapse still occurs. We hypothesized that, similarly to cell adhesion mediated chemotherapeutic drug resistance (CAM-DR), cell adhesion mediated CAR T-cell resistance (CAM-CART-R) can contribute to relapse of ALL. Results: To test our hypothesis, primary ALL cells were treated with CD19 CAR T cells either with murine calvaria-derived bone marrow stromal cells, OP9, or cultured only with media in short term cultures. We observed B-ALL cells treated with CD19 CAR T on OP9 has 10-20% higher viability compared to B-ALL and CD19 CAR T co-culture in medium alone, supporting the notion of CAM-CART-R. We also determined that soluble factors in OP9 primed medium may contribute to CAM-CART-R. However, the direct stromal contact mediated significant protection again CAR T induced apoptosis of B-ALL cells. To determine the molecular mechanisms underlying the survival promoting effects of stromal cells on CD19-, these cells were starved in serum-free media for 4hours and then treated with PI3Kδ inhibitor CAL-101 or DMSO and co-cultured with OP9 cells for 1 hour. We found that p-Akt is upregulated by stromal contact in CD19-negative B-ALL cells post-CAR T therapy and that PI3Kδ inhibition using can downregulate p-Akt in CD19-negative B-ALL patients. Critically, we investigated whether CD19 CAR T cells were functional under these conditions. For this purpose, we determined if stromal contact of ALL cells or stromal contact of CAR T cells changes the intracellular cytokine milieu of CD19 CAR T cells and found that intracellular IL-6, TNF- α and IFN-γ were reduced upon stromal contact supporting our hypothesis of a role of stromal cells in CAM-CART-R. We also determined that immune checkpoints molecules on T cells are unaffected by OP9 cells. Despite the reduction of cytokine level in T cells upon co-culture with B-ALL cells on OP9, PD-1, TIM-3 and LAG3 expression on CD19 CAR T cells after 2 days of co-culture was not altered as determined by flow cytometry. Resistance of ALL cells to CD19 CART cells was not mediated through checkpoint inhibition, since the PD-1/PD-L1 inhibitor Nivolumab failed to enhance ALL killing. Phenotypic profiling of thirteen cases of primary ALL relapse post-CD19 CAR T cell therapy showed high expression of adhesion molecules including integrin α4. Phenotypic analysis also revealed high expression of integrins is retained in primary ALL cells after CD19 knockout in one case. To explore possible solutions to overcome CAM-CART-R, we examined a strategy of blocking specifically integrin α4. We have previously shown that blocking integrin α4 can de-adhere CD19-negative B-ALL relapse post-CAR T cell therapy from their respective counter-ligands in vitro and that these cells can benefit from integrin blocking therapy in vivo. We have now confirmed this in NSG mice injected with CD19-negative B-ALL cells from a patient with post-CAR T cell relapse. Mice were treated intraperitoneally (n=6/group) with total immunoglobulin (Ig) control or humanized anti-human integrin α4 antibody Natalizumab (NZM). As a result, Natalizumab monotherapy significantly prolonged survival of leukemic mice compared to control Ig group (66 days (Ig) vs 85 days (NZM) p<0.005). Further combination treatments with chemotherapy are in progress. Conclusion: In summary, our data indicate that similarly to CAM-DR, CAM-CART-R can occur resulting in relapse of ALL. Targeting adhesion molecules may be a new approach to treat or prevent relapse following CD19 CAR T cell therapy for . Disclosures Ahmed: CellMedica: Other: Royalties; Celgene: Other: Royalties; Adaptimmune: Membership on an entity's Board of Directors or advisory committees. Babak:Simurx. Inc: Other: Founder . Pulsipher:Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Jazz: Other: Education for employees; Adaptive: Membership on an entity's Board of Directors or advisory committees, Research Funding; CSL Behring: Membership on an entity's Board of Directors or advisory committees; Amgen: Other: Lecture; Bellicum: Consultancy; Miltenyi: Research Funding; Medac: Honoraria. Wayne:AbbVie: Consultancy; Kite, a Gilead Company: Consultancy, Research Funding; Servier: Consultancy; Spectrum Pharmaceuticals: Consultancy, Research Funding. Abdel-Azim:Adaptive: Research Funding.