Multi-Omic Profiling of NK Cell Dysfunction and Tumor Immune Escape in Multiple Myeloma Evolution

Abstract

Introduction: Natural killer (NK) cells play a key role in immune surveillance of multiple myeloma (MM), targeting malignant plasma cells. Genome-wide CRISPR-Cas9 screening has identified genes regulating tumor susceptibility or resistance to NK-mediated killing, highlighting mechanisms of immune escape. Thus,this study aims to characterize NK cells dynamics and tumor immune visibility in MM through integrated functional genomic screening, single-cell transcriptomics, and multiparametric flow cytometry across disease stages. Methods: We analyzed a published genome-wide CRISPR-Cas9 screen in NK-tumor co-cultures to define gene signatures linked to NK sensitivity or resistance in MM cell lines. These signatures were applied to single-cell RNA-seq data from 209 patients, profiling 209,175 malignant plasma cells and 48,614 NK cells (Resting, Cytokine-activated, Adaptive, Type I, Activated). Healthy donor (HD) samples served as reference. Additionally, we performed fresh bone marrow sampling from 46 MM patients and conducted multiparametric flow cytometry. FlowCT was used for cytometric analysis and unsupervised clustering. Results: Both NK sensitivity and resistance gene signatures declined with disease progression, indicating reduced tumor immune visibility and impaired NK recognition. NK cells showed marked inter-patient heterogeneity and different functional states. Clustering of integrated tumor and NK profiles revealed seven transcriptional clusters. Cluster 1 (68% MM) showed cytokine-responsive NKs (e.g., IFNG, CD69), while Cluster 4 (mainly HD) retained resting NK features (e.g., NKG2A, GZMK) and higher tumor immunogenicity. Clusters 5–6 (mostly MM) showed low susceptibility and expression of immunoregulatory genes (e.g., TIGIT, CD96). Cluster 7 (relapsed MM) exhibited adaptive NK features (e.g., KLRC2, IL32). MGUS and SMM aligned more closely with HD profiles, indicating preserved immune control in early disease. Flow cytometry confirmed progressive downregulation of TIGIT and NKG2A in NK cells with disease progression. In MM patients with osteolytic lesions, we observed expansion of activated NK subsets (TIM3+ CD57+/− TIGIT− PD1− NKG2A+/−) in bone marrow, suggesting a local immune response to inflammation and bone damage. Conclusions: MM progression involves loss of tumor immunogenicity and remodeling of NK cell states, with emergence of suppressive phenotypes in advanced stages and expansion of activated NKs in osteolytic disease. MGUS and SMM retain more physiological NK-tumor interactions. These findings offer a multiomic framework for understanding NK dysfunction in MM and support the development of NK-based immunotherapies and monitoring strategies across disease stages.