PF439 EXPLORING THE THERAPEUTIC POTENTIAL OF T‐CELL‐RECEPTOR REDIRECTED CD3+ NATURAL KILLER CELLS FOR ADOPTIVE CELLULAR THERAPY TO ACUTE MYELOID LEUKEMIA

Abstract

Background:Adoptive cellular therapy (ACT) with redirected T cells expressing a chimeric antigen receptor (CAR) or transgenic T‐cell‐receptor (tTCR) has revolutionized cellular immunotherapy to hematological neoplasia, in particular to acute lymphoid leukemia, and also shows great promise as therapy for solid tumors. While CARs can only detect fully cell surface expressed target structures, TCR‐mediated recognition is not limited to surface antigens, but covers processed tumor neoantigens derived from the whole proteome. However, mispairing of transgenic and endogenous TCRs and restriction to patient‐derived, autologous T lymphocytes with variable “fitness” and T cell subsets due to individual health conditions and age of the patient exemplify the current limitations encountered in TCR‐redirected ACT. The natural killer (NK) cell line NK92 elicits lytic activity comparable to T cells, has been approved by the FDA for ACT and shown not to cause graft‐vs‐host disease. Since NK cell redirection is currently limited to CARs, NK92 cells engineered to express a CD3/TCR or CD3/CD8/TCR complex might evolve as an attractive, standardized cellular source for off‐the‐shelf TCR‐based ACT.Aims:Thus, the goal of this study was to explore NK92‐CD3+ and NK92‐CD3+CD8+ variants redirected to acute myeloid leukemia (AML) by expression of different AML‐reactive TCRs for antileukemic immunity in vitro and in vivo.Methods:NK92 cells expressing the γδεζ‐chains of the CD3 complex (NK92CD3+) were kindly provided by Dr. C. Wölfel (III. Dept. of Medicine) and further genetically engineered to express different human CD8 isoforms. Upon retroviral transfer (RT) of 3 different codon‐optimized and murinized TCRs obtained from AML‐specific CD8+CTL clones into NK92‐CD3+cells TCR expression was examined by flow cytometry (FC). NK92CD3+TCR+as well as NK92CD3+CD8+TCR+cells (together referred to as NK92TCR+) were enriched using antibiotic selection and immunomagnetic cell sorting (MACS®). Biological reactivity of redirected NK92TCR+cells to patient‐derived (primary) AML blasts was tested in vitro by IFN‐γ ELISPOT and 51Cr‐release assays. Comparable expression‐profiling of checkpoint molecules on NK92 and NK92TCR+cells such as e.g. CD80/86, PD1/PDL1, NKG2A, TIM3, TIGIT and HLA‐E was performed by FC. A patient‐derived AML‐xenograft (PDX) model in NSG mice has been already established.Results:Following RT and enrichment >80% of NK92‐CD3+cells expressed a given TCR, and >90% of NK92CD3+CD8+TCR+cells could be obtained. Upon coculture with different primary AML samples and corresponding EBV‐LCL blasts as controls NK92CD3+TCR+ elicited TCR dependent IFN–γ release and exhibited cytolytic activity to some but not all AML samples. CD8 coexpression slightly increased reactivity as measured by IFN‐γ release. Interestingly, as TCR‐redirected NK92 antileukemic responses appeared to be inhibited to some but not all AML samples we found high expression of HLA‐E on these AML blasts and of NGK2A on NK92TCR+ cells. Current blocking studies to examine whether a HLA‐E/NKG2A interaction might impair NK92TCR+ mediated AML reactivity are ongoing and will be presented. Moreover, first preliminary in vivo experiments of adoptively transferred NK92TCR+ cells into a NSG‐AML PDX model indicated biological activity but further studies are in progress to evaluate this point in more detail.Summary/Conclusion:These studies demonstrate that the established and FDA approved NK92 cell line can be redirected to elicit TCR‐mediated antitumoral immunity to AML. NK92TCR+ cells might thus represent a promising universal tool for an ‘off‐the‐shelf’ ACT product