Abstract
BACKGROUND: Colorectal carcinoma exhibits high heterogeneity, comprising subtypes that show poor efficacy of T cell-based immunotherapies, such as programmed cell death protein 1 (PD-1) checkpoint inhibitors. Although natural killer (NK) cells are considered a promising approach for cancer immunotherapy, it remains unclear what molecular mechanisms drive NK cell activation or suppression within the tumor microenvironment. Moreover, limitations in human tumor models that reflect the diversity of individual patient tumors hinder the ability to effectively select patients who would benefit most from NK cell-based therapies. METHODS: Here, we established a co-culture platform of genetically diverse colorectal cancer (CRC) patient-derived organoids (PDOs) with primary allogeneic NK cells. We performed bulk RNA sequencing analysis of sorted NK cells after exposure to PDOs and aligned gene expression signatures derived from our findings with publicly available single-cell RNA sequencing data of NK cells from peripheral blood and CRC tissues of patients. Moreover, we evaluated identified pathways using flow cytometry and IncuCyte live-cell imaging analysis to quantify phenotypic alterations and NK cell-mediated killing of PDOs over time, respectively. Ultimately, we tested CRISPR-Cas9-edited NK cells and PDOs, small molecule compounds, and clinically relevant monoclonal antibodies (mAbs) to increase NK cell potency. RESULTS: On co-culture, NK cells acquired common transcriptional signatures related to hypoxia and transforming growth factor-beta (TGF-β), similar to NK cells infiltrating CRC tissues of patients. In addition, we observed patient-specific differential PDO susceptibilities to NK cell-mediated lysis. Major histocompatibility complex class I deficiency and natural killer group 2, member D (NKG2D)-ligand expression on PDOs facilitated NK cell-mediated cytotoxicity, and induced phenotypic NK cell diversification related to activation and the acquisition of inflammation and tissue-residency-related transcriptional signatures. Genetic or pharmaceutical targeting of hypoxia-inducible factors HIF1A/EPAS1 or TGF-βR1, or the addition of anti-CEACAM1 mAbs, enhanced NK cell-mediated PDO killing or activation, respectively. CONCLUSIONS: The NK cell/PDO co-culture platform allows the identification of both common and patient-specific impacts of the tumor microenvironment on NK cell function and can aid the development of patient-tailored immunotherapies. The majority of CRC (CMS2/CMS3) PDOs from our cohort were susceptible to NK cell-mediated killing and induced NK cell activation, highlighting the potential of NK cells for CRC immunotherapies.