250 Manipulating T cell epigenetic programs to improve persistence against group 3 pediatric medulloblastoma

Background

Safe, effective therapies are desperately needed to improve survival and long-term quality of life for patients with medulloblastoma (MB), a malignant pediatric brain cancer. Chimeric antigen receptor (CAR) T cells are a promising therapeutic option as they specifically target T-cell cytotoxic potential to surface antigens only presented by tumor cells. All subgroups of MB highly express the surface protein B7 homolog 3 (B7-H3) whereas healthy brain tissue does not, making it an ideal target for CAR T cell therapy. However, a major roadblock to long-term tumor control is the inability of CAR T cells to persist, suggesting that genetic modifications to improve persistence will benefit overall therapeutic efficacy. The goal of this study was to compare head-to-head the impact of knocking out DNMT3A and TET2 negative regulators of CAR T cell proliferation and persistence against group 3 MB (G3MB), the most malignant subtype.

Methods

We generated second generation B7-H3 CAR T cells with CD28ζ signaling domain and performed CRISPR/Cas9 knockout (KO) of DNMT3A, TET2, and AAVS1 as a control. We compared resulting CAR T cell persistence via repeat stimulation assays and tumor killing capacity via MTS cytotoxicity assays in vitro. We further compared genetically modified CAR T cells in NSG mice in vivo against the HDMB03 G3MB cell line.

Results

Our data shows that KO of DNMT3A in B7-H3 CAR T cells consistently improves persistence against G3MB cell lines in vitro. KO of TET2, however, shows inconsistent improvement of persistence that is dependent on the healthy donor used to generate the CAR T cells. In 2 out of 4 donors tested, TET2 KO CAR T cells persisted longer than AAVS KO control CAR T cells and DNMT3A KO CAR T cells. In the other 2 donors, TET2 KO did not improve CAR T cell persistence. In donors where TET2 did produce a persistence benefit, TET2 KO CAR T cells failed to elicit strong tumor killing after multiple repeat stimulations compared to DNMT3A KO.

Conclusions

Our study demonstrates that genetic KO of negative regulators of CAR T cell persistence can improve their anti-tumor function against G3MB. We further conclude that for the CD28ζ B7-H3 CAR, DNMT3A KO is superior for achieving improved persistence over TET2 KO. We are currently validating this result in other G3MB cell lines with differential B7-H3 antigen density to determine how antigen quantity impacts the performance of DNMT3A and TET2 KO CAR T cells.