NK Cell States in Multiple Myeloma: Pathways to Novel Immunotherapies

Introduction

The tumor microenvironment (TME) in multiple myeloma (MM) is characterized by immune suppression, which hinders effective anti-tumor responses. Natural Killer (NK) cells, crucial for tumor surveillance, often exhibit dysfunction and exhaustion in the MM TME. Characterizing the TME and the states of NK cells is essential for identifying therapeutic targets to restore NK cell activity. This study provides a comprehensive single-cell atlas of NK cells in MM and healthy bone marrow, emphasizing therapeutic target identification for NK cell-based therapies.

Methods

We constructed an MM single-cell atlas using publicly available single-cell RNA sequencing (scRNA-seq) datasets from GEO, encompassing 91 samples across all disease stages of MM. For comparison, a healthy control dataset was built, including 37 samples. Sequencing data were processed using a standardized pipeline to ensure comparability. Cell type and state annotations were performed using SingleR and a gene expression signature-based scoring algorithm, classifying NK cells into active, tissue-resident (rNK), or exhausted (eNK) states. Differential gene expression (DGE) analysis and gene ontology (GO) enrichment were conducted to profile eNK cells. Cell-cell interaction analysis characterized the interactions within the TME. Transcription Factor (TF) enrichment analysis was performed with pySCENIC on eNK cells for each disease stage to identify key regulatory elements driving the observed NK cell states and potential genes to target for NK reactivation.

Results

The MM atlas consisted of 320,479 cells, while the healthy dataset included 140,229 cells. NK cells were classified into rNK and eNK states. In comparison to the healthy dataset (mean eNK: 19.3%, CI: 15.8,19.3), eNK proportion was elevated in monoclonal gammopathy with uncertain significance (MGUS; 32.6%,p=0.04) and continually increased with disease progression: smoldering MM (SMM; 40.9%,p<0.001), primary MM (PMM; 40.8%,p<0.001) and relapsed/refractory MM (RRMM; 59.2%,p<0.001)). DGE analysis identified 425 differentially expressed genes between rNK and eNK cells in MM, amongst which genes involved in NK cell proliferation were downregulated in eNK cells (e.g. FOS/JUN family members and CD69), and genes participating in immune activation (KLRC2 and CCL5), and NK cell exhaustion, including TIGIT and CD47 were upregulated. GO enrichment analysis revealed non-overlapping biological processes associated with NK cell exhaustion in healthy vs. MM conditions. Cell-cell interaction analysis identified 19 activated immune suppression-related receptors in healthy samples, 19 in MGUS, 17 in SMM, 24 in PMM, and 22 in RRMM. No significant differences were observed in the number of cell types interacting with these receptors. Notably, TIGIT and LAG3 were selectively activated in PMM eNK cells. Activation of eNK-expressed receptors was associated with activation of STAT-, ETF1-, ELF1 transcription factor driven programs.

Conclusions

This study presents a detailed single-cell atlas of NK cells during MM development and progression, highlighting the continuous accumulation of eNK cells as the disease evolves. The distinct gene expression profiles and enriched biological processes in MM eNK cells suggest unique, MM-specific pathways of immune exhaustion and activation with several TME cell types predicted to contribute to NK cell exhaustion. The TF analysis identified key TF-driven programs involved in NK biology and immune regulation, warranting their study as potential therapeutic targets to revert NK cell exhaustion in MM.

Verga:ONK Therapeutics: Research Funding. Szegezdi:ONK Therapeutics: Research Funding.