Background Natural killer (NK) cells are emerging as a promising therapeutic option in cancer. To better understand how cancer cells evade NK cells, we studied interacting NK and blood cancer cells using single-cell and genome-scale functional genomics screens.

Methods We performed multiplexed single-cell RNA-seq (scRNA-seq) on co-cultures of NK cells and 26 cell lines representing diverse blood cancers. Using screens of pooled DNA-barcoded cell lines (PRISM), we quantified the sensitivity of over 60 blood cancer cell lines and integrated the results with CCLE multi-omics to uncover molecular correlates of tumor cell susceptibility to NK cells. We performed 12 genome-scale CRISPR loss-of-function and gain-of-function screens of cancer-cell intrinsic NK cell resistance mechanisms across 7 blood cancer cell lines. Finally, we investigated the mechanisms-of-action of 65 genome-scale screen hits using CRISPR screens with scRNA-seq readout in both tumor and NK cells.

Results At single-cell resolution, interaction of NK and cancer cells induced distinct activation states in both cell types depending on the cancer cell lineage and molecular phenotype. NK cells transitioned either into an activated state characterized by 4-1BB, GITR, TIM-3, and TIGIT or a state marked by type I interferon signature. Tumor cells responded to NK cell attack by activating interferon gamma (IFNy) signaling, inducing MHC class I. The activation states correlated with sensitivity to NK cells, ranging from more sensitive myeloid to more resistant B-lymphoid cancers. Molecular correlates of increased sensitivity included expression of activating receptor ligands NCR3LG1, PVR, and ULBP1 and mutations in the NF-kB regulator TRAF3. CRISPR screens uncovered cancer cell-intrinsic genes driving sensitivity and resistance, including antigen presentation and death receptor signaling mediators and adhesion molecules. The screens identified new blood cancer-specific NK cell inhibitory regulators (SELPLG, SPN, and MYB ) and genes previously underappreciated in NK cell evasion, including protein fucosylation and transcriptional regulators (e.g. GFI1B). CRISPR screens with scRNA-seq readout identified MHC-I, IFNy, and NF-κB regulation as underlying mechanisms. Cancer cell knockout of positive regulators of NK cell response (CD58, NCR3LG1) induced an inactive NK cell state, providing experimental evidence how cancer cell-intrinsic genetic alterations can shape the molecular profile of attacking immune cells to promote immune evasion.

Conclusions By integrating diverse functional genomics screens and patient genomic profiles, we provide a comprehensive landscape of potential biomarkers and functionally validated genetic mechanisms which influence how NK cells recognize and kill malignant cells. The results offer a roadmap to facilitate development of NK-cell based immunotherapy for blood cancers and beyond.

Ethics Approval The study was approved by the Helsinki University Hospital ethics committee, (permit number 303/13/03/01/2011), and abided by the principles of the Declaration of Helsinki.