P-414: Single-Cell Profiling Reveals Aggressive Tumor Subcluster and Compromised Immune State in Multiple Myeloma

Introduction: Despite advancements in managing multiple myeloma (MM) patients, a subset of individuals still undergoes rapid disease progression and early death (ED) within two years of diagnosis. To gain insights into the disease and identify treatment targets, we examined the molecular characteristics of malignant and non-malignant cells in the bone marrow of MM patients with ED.

Methods: We used single-cell RNA (scRNA) sequencing to analyze bone marrow aspirates from four MM patients with ED, eight without ED (nED), and seven healthy donors. Additionally, we incorporated CHIP-seq data to investigate epigenetic modifications. We constructed a comprehensive differentiation trajectory from stem cells to plasma cells and identified mutation hotspots using whole-exome sequencing. Non-malignant cell variations within the myeloma microenvironment were explored using pan-cancer references. The scRNA findings were validated through in vitro and in vivo experiments.

Results: We classified Ig light chain-restricted tumor cells into ten subclusters and found that subcluster 4 had a higher presence in ED patients. Other subclusters grouped with the classical MM cells phenotype (CD38+CD19-CD56+/-CD24-). Subcluster 4 displayed a plasmablast phenotype (CD38+CD19+CD56+CD24+) with increased proliferation and drug resistance. Chromosomal aberrations and mutation hotspots were more prevalent in this subcluster. The genes specific to subcluster 4 were enriched in H3K27ac-mediated epigenetic regulation. We constructed a gene score, with LILRB4 showing high specificity for subcluster 4, effectively identifying ED patients. Combining this gene score with clinical factors improved patient stratification. In vitro and in vivo experiments confirmed the enhanced properties of LILRB4+ cells.

The evolution trajectory of CD8+ T cells was divided into cell fate 1 (exhausted) and cell fate 2 (activated). Exhausted T cells were more abundant in ED patients and showed metabolic inhibition, reduced cell cycle activity, and upregulated apoptosis pathways. Memory T cell differentiation was impaired in the ED group, with increased expression of immunosuppressive genes like PIM3. In vitro experiments demonstrated that inhibiting PIM promoted memory T cell differentiation.

Cell-cell interaction analysis revealed a significant increase in MM cell-CD14+ monocyte interactions in the ED group, particularly through LILRB4-interacting molecules. Co-culture experiments revealed that LILRB4+ MM cells promoted the differentiation of CD14+ monocytes into myeloid-derived suppressor cells.

Conclusions: Our study provides insights into aggressive tumor subclusters and compromised immune states in MM patients with ED. We identified a unique population of LILRB4+ tumor cells and developed a gene score to identify ED patients. We also confirmed T cell dysfunction and highlighted the role of LILRB4+ tumor cells in immune evasion. Targeting LILRB4 could be a promising therapeutic strategy for MM treatment.