Integrative genomic, single-cell, and functional profiling uncovers immune consequences of the CD48–CD244 axis in NK cells in multiple myeloma

Abstract Background: Natural killer (NK) cells play a key role in immune surveillance of multiple myeloma (MM), but their function becomes progressively impaired as the disease advances. CD48, one of the most abundant NK-cell ligands on MM cells, engages the receptor CD244 (2B4), transmitting either activating or inhibitory signals depending on downstream adaptors such as SAP or inhibitory phosphatases like SHIP-1. However, the net effect of CD48–CD244 interaction (whether it promotes NK-cell activation or dysfunction) remains unclear. Methods: We combined bulk and single-cell RNA-seq (scRNA-seq), ATAC-seq, machine learning, genome-wide CRISPRa/KO screens, in vitro functional assays, and in vivo murine modeling to investigate CD48's regulation and role in NK-cell immunity. Results: In CoMMpass RNA-seq data (n=774), CD48 expression was highest in high-risk MM subsets (t(4;14), gain(1q21), p=<0.0001) and associated with inferior survival (p=0.0044). Unsupervised clustering of NK-ligand expression in tumor cells (including HLA-A/B/C/E, PCNA, MICA/B, PVR, ULBP1–3, NECTIN2, in addition to CD48) stratified patients into groups with distinct cytogenetics and outcomes, with broader NK-ligand patterns, such as Cluster B characterized by High expression of HLA-A/B/C/E and MICA/B, linked to improved survival. To uncover CD48 regulation, we analyzed ATAC-seq and ENCODE ChIP-seq data and trained an XGBoost model to predict CD48 expression from transcription factor activity. IRF4 and TFAP2A emerged as top positive regulators, while MEF2B acted as a negative regulator (R² = 0.40). Genome-wide CRISPRa/KO screens in MM cell lines identified 52 candidate regulators of surface CD48 expression, which clustered into functional categories including steroid receptor signaling, ER trafficking, and GPI-anchor biosynthesis. CRISPR-mediated validation confirmed that ESRRA represses and TFAP2A promotes CD48 surface expression. To test functional consequences, we engineered Vk*MYC MM cells to overexpress or knock out CD48 and co-cultured them with murine KIL C.2 NK cells. CD48 overexpression enhanced NK cytotoxicity and cytokine secretion in vitro, while CD48KO impaired activation. SHIP1 deletion in NK cells failed to restore cytotoxicity, suggesting deeper exhaustion mechanisms. Human scRNA-seq analysis (GSE223060, n=54 MM, 8 HD) revealed CD48–CD244 signaling from plasma cells and monocytes to NK cells, inferred using CellChat, was dominant in MM but not HD. MM-NK cells exhibited transcriptional exhaustion: upregulated SHIP1, SAP, and HLA-inhibitory receptors (e.g., NKG2A), alongside downregulation of AP-1 family members (FOS, JUN, JUNB). An CMV “adaptive-like” CD56-dim NK subset in MM co-expressed FGFBP2, KLRC2/3, and downregulated KLRB1/KLRF1, consistent with chronic stimulation. An independent scRNA-seq cohort (n>200) confirmed AP-1 family members suppression and elevated CD48 expression in MM over MGUS, SMM, or controls. In vivo, in the immunocompetent Vk*MYC model, CD48-overexpressing tumors showed delayed growth and extended survival vs. WT, while NK-cell depletion accelerated disease. NK cells isolated ex vivo from tumor-bearing mice showed functional exhaustion with reduced cytokine secretion and degranulation, even after PMA/ionomycin stimulation. Finally, murine scRNA-seq of Vk*MYC bone marrow samples across disease stages (ND, MM, AMM) revealed stepwise NK dysfunction: decreasing cytotoxicity, rising exhaustion and inflammatory module scores, and upregulation of HLA-inhibitory receptors, paralleling human MM data. Conclusion: CD48 exerts dual effects in MM, potentiating acute NK activation while driving chronic exhaustion via sustained CD244 engagement. Our multi-platform analysis uncovers transcriptional and post-translational CD48 regulators, validates immunocompetent Vk*MYC as a preclinical model of NK dysfunction, and identifies CD48–CD244 signaling as a key immunosuppressive axis. These insights suggest that modulating CD48 or its downstream effectors could restore NK surveillance and inform new immunotherapies for MM.