Revised International Staging System (R-ISS) stage-dependent analysis uncovers oncogenes and potential immunotherapeutic targets in multiple myeloma

Multiple myeloma (MM), characterized by high intratumour heterogeneity, accounts for ∼10% of all haematologic malignancies. Stratified by the Revised International Staging System (R-ISS), little is known about R-ISS-related plasma cell (PC) heterogeneity, gene expression modules in cytotoxic T/NK cells and immunoregulatory ligands and receptors. Herein, we constructed a single-cell transcriptome atlas of bone marrow in normal and R-ISS-staged MM patients. Focusing on PCs, we identified and validated a subset of GZMA+ cytotoxic PCs. In addition, a malignant PC population with high proliferation capability (proliferating PCs) was associated with unfavourable prognosis and EBV infection in our collected samples. Ribonucleotide Reductase Regulatory Subunit M2 (RRM2), a specific marker of proliferating PCs, was shown to induce MM cell line proliferation and serve as a detrimental marker in MM. Subsequently, three R-ISS-dependent gene modules in cytotoxic CD8+ T and NKT cells were identified and functionally analysed. Finally, cell-cell communication between neutrophils and proliferating PCs with cytotoxic CD8+ T and NKT cells was investigated, which identified intercellular ligand receptors and potential immunotargets such as SIRPA-CD47 and TIGIT-NECTIN3. Collectively, this study provides an R-ISS-related single-cell MM atlas and reveals the clinical significance of two PC clusters, as well as potential immunotargets in MM progression.

Introduction cytotoxic plasma cells in MM remain limited. The existence of cytotoxic plasma cells prompted 120 us to investigate their existence and clinical relevance in MM. Considering that C21 accounts 121 for a rare population (average 2.04%, ranging from 0%-10.00%) in all plasma cells (Fig 2B), 122 we first validated its existence with another MM single-cell dataset. In studies focusing on 123 plasma cell heterogeneity of symptomatic and asymptomatic myeloma (dataset GSE117156)

156
Next, we discovered 75 significantly up-regulated genes in the MM (UGM) dataset of 157 GSE6477 as compared with normal samples (Fig 4A-4B). The top 150 specific genes in C10 158 were compared with 75 UGMs, and 7 genes (CADM1, HIST1H1C, CD48, RRM2, PPIB, LDHB 159 and HINT1) were acquired ( Fig 4C). The expression of 7 UGMs is shown with the R-ISS stage 160 in Fig 4D. We calculated a 7-gene signature score and analyzed the relevance of the score 161 with respect to clinical parameters ( Fig 4E). Next, the prognostic significance of these 7 genes 162 was analyzed. RRM2 and HINT1 showed good performance as unfavorable markers, with 163 HR=2.3 (95% CI=1.4-3.6, p-value<0.000402) and HR=1.9 (95% CI=1.2-2.9, p-value= 164 0.005496), respectively ( Fig 4F). Then, the expression of RRM2 and HINT1 was examined in    Finally, we calculated the differentially expressed genes (DEGs) with fold change >=2 or < 171 =0.5 and adjusted p value <0.05 in C10 by comparing R-ISS stages I and III. All 150 DEGs 172 were acquired. Then, we constructed a functional analysis of these 150 DEGs. As shown in

199
First, re-clustering of T cells generated 21 clusters (Fig 6A), T1 to T21, belonging to CD4+ T cells, CD8+ T cells, NK cells and NKT cells (Fig 6B and 6D). No biased distribution was 201 observed in 11 samples ( Fig 6C). The percentages of T1 to T21 in MM versus healthy controls 202 and MM R-ISS I to III are presented in Fig 6E and Fig 6F, respectively. It is worth noting that 2 203 clusters conform to the hypothesis of decreased percentage along with R-ISS stages: T2 and 204 T10. T2 was marked by high expression of CD8A and no expression of NKG7 and was 205 identified as CD8+ T cells. T10 cells express both CD8A and NKG7 and were defined as NKT 206 cells. We concentrate on T2 and T10 in the following work.

207
To identify R-ISS-dependent gene modules in T2 and T10, the R package MFUZZ [46] was 208 applied. As a result, in T2 CD8+ T cells, 12 gene modules with distinct expression patterns 209 were generated, and module 5 (gradually increased expression with R-ISS stage) and module 210 3 (gradually decreased expression with R-ISS stage) were chosen for subsequent analysis 211 ( Fig 7A-7B). As expected, genes in module 5 were functionally related to antigen processing 212 and presentation, T cell activation and haemopoiesis ( Fig 7D). Surprisingly, genes in module 3 213 were involved in neutrophil activation (Fig 7E), which prompted us to examine neutrophils in 214 MM. Significantly, the proportion of activated neutrophils characterized by CXCR2 expression

269
In addition, we also identified multiple immunotherapeutic targets in MM. CD24 is a highly