GRP78 Is Expressed on the Cell Surface of Pediatric AML Blasts and Can be Targeted with GRP78-CAR T Cells without Toxicity to Hematopoietic Progenitor Cells

Finding the ideal immunotherapy target for AML has proven challenging and is limited by overlapping expression of antigens on hematopoietic progenitor cells (HPCs) and AML blasts. Intracellular Glucose-regulated-protein 78 (GRP78) is a key UPR regulator, which normally resides in the endoplasmic reticulum (ER). GRP78 is overexpressed and translocated to the cell surface in a broad range of solid tumors and hematological malignancies in response to elevated ER stress, making it an attractive target for immune-based therapies with T cells expressing chimeric antigen receptors (CARs). The goal of this project was to determine the expression of GRP78 on pediatric AML samples, generate GRP78-CAR T cells, and evaluate their effector function against AML blasts in vitro and in vivo.

To demonstrate overexpression of GRP78 in AML, we performed gene expression analysis by RNAseq on a cohort of cord blood CD34+ cell samples (N=5) and 74 primary AML samples. Primary AML samples included RUNX1-RUNX1T1 (N=7), CBFB-MYH11(N=17), KMT2A rearrangement (N=28) and NUP98 (N=22). Analysis showed increased GRP78 expression in AML samples, especially in KMT2A- and NUP98-rearranged AML. To demonstrate surface expression of GRP78, we performed flow cytometry of AML (Kg1a, MOLLM13, THP-1, MV4-11) cell lines as well as 11 primary AML samples and 5 PDX samples; non transduced (NT) T cells served as control. All AML samples, including cell lines, primary AML blasts, and PDX samples, showed increased expression of GRP78 on their cell surface in comparison to NT T cells

We then designed a retroviral vector encoding a GRP78-CAR using a GRP78-specific peptide as an antigen recognition domain, and generated GRP78-CAR T cells by retroviral transduction of primary human T cells. Median transduction efficiency was 82% (± 5-8%, N=6), and immunophenotypic analysis showed a predominance of naïve and terminal effector memory subsets on day 7 after transduction (N=5). To determine the antigen specificity of GRP78-CAR T cells, we performed coculture assays in vitro with cell surface GRP78⁺ (AML cell lines: MOLM13, MV-4-11, and THP-1 and 3 AML PDX samples) or cell surface GRP78^- (NT T cells) targets. T cells expressing CARs specific for HER2-, CD19-, or a non-functional GRP78 (DGRP78)-CAR served as negative controls. GRP78-CAR T cells secreted significant amounts of IFNg and IL-2 only in the presence of GRP78⁺ target cells (N=3, p<0.005); while control CAR T cells did not. GRP78-CAR T cells only killed GRP78⁺ target cells in standard cytotoxicity assays confirming specificity. To test the effects of GRP78-CAR T cells on normal bone marrow derived HPCs, we performed standard colony forming unit (CFU) assays post exposure to GRP78-CAR or NT T cells (effector to target (E:T) ratio 1:1 and 5:1) and determined the number of BFU-E, CFU-E, CFU-GM, and CFU-GEMM. No significant differences between GRP78-CAR and NT T cells were observed except for CFU-Es at an E:T ratio of 5:1 that was not confirmed for BFU-Es. Finally, we evaluated the antitumor activity of GRP78-CAR T cells in an in vivo xenograft AML model (MOLM13). Tumor growth was monitored by serial bioluminescence imaging. A single intravenous dose of GRP78-CAR T cells induced tumor regression, which resulted in a significant (p<0.001) survival advantage in comparison to mice that had received control CAR T cells.

In conclusion, GRP78 is expressed on the cell surface of AML. GRP78-CAR T cells have potent anti-AML activity in vitro and in vivo and do not target normal HPCs. Thus, our cell therapy approach warrants further active exploration and has the potential to improve outcomes for patients with AML.