Tissue-Resident Adaptive CD56bright NK Cell Responses in Cancer

ABSTRACT Natural killer (NK) cells are key mediators of anti-tumor immune responses, acting to directly kill cancer cells (CD56dim subtype) and recruit pro-inflammatory immune populations (CD56bright subtype). Recent discoveries, however, have revealed other NK cell subtypes with properties of tissue-resident and adaptive, memory-like

phenotypes. Tissue-resident NK cells have been shown to have elevated production of chemokines, such as CCL3 and XCL1 that, in turn, recruit conventional type 1 dendritic cells (cDC1) and CD8+ T cells to tumors. Adaptive NK cells have been shown to persist for months longer than conventional NK cells and display

enhanced tumor cell killing. Given the diversity of the NK cell repertoire, it is crucial to identify which NK cell subtypes exist in tumors and are key to anti-tumor responses. To address this, our group has generated a pan- cancer atlas of NK cells using single-cell RNA-sequencing data from 326 patients across 17 cancer types. In

doing so, we identified prominent populations of CD56bright NK cells that display gene signatures suggestive of both tissue-resident and adaptive NK phenotypes (bright-taNKs). We found bright-taNKs to display higher expression of chemokines and cytokines, such as CCL3, XCL1, and IFNG, suggesting a functional role in

immune cell recruitment to tumors. Using a gene signature to estimate the relative abundance of bright-taNKs in bulk RNA-seq data, we found that bright-taNKs are associated with significantly longer disease-free survival in cancer patients. Thus, bright-taNKs may be critical for anti-tumor immunity and represent attractive

phenotypes to mimic for the next generation of engineered NK cell therapies. As little is currently known about bright-taNKs, this proposal aims to advance our understanding of the functions of bright-taNKs in tumors and identify mechanisms promoting their distinct phenotypes. In Aim 1, we will use orthogonal methods to determine

whether bright-taNK populations are associated with increased recruitment of anti-tumor immune populations, such as cDC1s and CD8+ T cells. We will develop a bioinformatics tool that can be paired with well-established tools to estimate fractions of different NK cell subtypes and other immune cell types in bulk RNA-seq data.

Additionally, we will apply highly-multiplexed spatial imaging tools to interrogate the spatial distribution of bright- taNKs, and other NK cell populations, in tumors and their associations with other immune cell types. In Aim 2, we will use gene delivery platforms to screen for transcription factors that promote features of tissue-resident

and adaptive phenotypes in CD56bright NK cells. We will then transition the most promising gene targets to our clinical-grade induced pluripotent stem cell to NK cell (iNK) platform to holistically evaluate the effects on anti- tumor functions. Altogether, this proposal aims to advance our understanding of bright-taNKs, and other NK cell

subtypes, in tumors. Ultimately, drivers of favorable NK cell properties can be readily integrated into our group’s bench-to-bedside pipeline for NK cell therapies to enhance their efficacy and durability.