Bone marrow‐derived mesenchymal stem cells inhibit NK cell function via Tim‐3/galectin‐9 in multiple myeloma patients

Dear Editor, Our research clarifies that the characteristics of NK cell immunodepleting in the bone marrow micro-environment of newly diagnosed multiple myeloma (NDMM) patients. Mechanically, bone marrow mesenchymal stem cells (BMSCs) from multiple myeloma (MM) patients can promote NK cell exhaustion via Tim-3/galectin-9 through direct or indirect cell-to-cell communication. MM is known as haematological neoplasia, in which aberrant plasma cells develop clonally in the bone marrow, leading to anaemia, renal damage, osteolytic destruction and abnormal immune function.1 NK cell is a lymphocyte that kills tumour cells without prior sensitization. The exhaustion of NK cells plays a vital role in MM bone marrow inhibiting the immune environment.2 T-cell immunoglobulin mucin 3 (Tim-3), which can bind galectin-9 on myeloma cells, is known to be a negative regulatory molecule of NK cells.3 BMSC is crucial for modulating functions of immune cells. Studies have shown that MM-derived BMSCs can interact with immune cells in the bone marrow micro-environment through direct or indirect cell-to-cell communication, regulating immune cell function and thus influencing the onset and development of MM.4, 5 However, there is little known about Tim-3 ligands on BMSCs and the ways interact with NK cells. First, we used clinical samples to evaluate the NK cell status in MM (Supplemental Table S1). In terms of quantity, the ratios of total NK cells (CD3−CD56+) and CD56dimNK cells (CD16+CD56dim) were significantly higher in NDMM and complete remission multiple myeloma (CR) than in Healthy donors (HDs), but the ratios of CD56brightNK cells (CD16−CD56bright) were not significantly different between three groups (Supplemental Figure S1A and Supplemental Table S2). In terms of functionality, expression of NK cell functional molecules (CD107a, NKG2D, INF-γ, and perforin) was lower in NDMM than in HDs (Supplemental Figure S1B-E and Supplemental Table S3). In short, we observed that NK cells of MM have an increased quantity but a decreased function. So, why did NK cells functionally exhaust in MM? Single-cell RNA sequencing datasets GSE188632 (end-stage MM) and GSE166902 (healthy donor, HD) were reanalysed.6 It showed that HAVCR2 gene (Tim-3) expression of NK cells was higher in MM (Figure 1A–C). GSE27838 (NK cell RNA sequencing data from eight HDs and MM) was also reanalysed.7 HAVCR2 gene expression was also higher in MM (Figure 1D), but other immune checkpoints were not significantly different (Supplemental Figure S2). Meanwhile, GSE113736 (BMSCs RNA sequencing data) was also reanalysed to verify Tim-3 ligands expression.8 LGALS9 (galectin-9) expression was higher in MM than in HDs, while other Tim-3 ligands (HMGB1, CEACAM1, and PtdSer) had no significant differences (Figure 1E). Using the MMRF-COMMPASS database, we demonstrated that survival was shorter in the high Tim-3 group than in the low Tim-3 group (Figure 1F). These suggested that Tim-3 can influent the early survival of MM and Tim-3 may regulate NK cell function by interaction with GSE113736 in BMSCs. Next, we used clinical samples to verify the significant high Tim-3 expression in MM. Tim-3 expression of total NK and CD56dimNK cells was significantly greater in NDMM than in CR and HDs. Tim-3 expression of CD56brightNK cells was also greater in NDMM than in HDs (Supplemental Figure S3A and Supplemental Table S4). As for functions of Tim-3-positive NK cells (CD3−CD56+Tim-3+), the expression of CD107a, NKG2D and INF-γ on Tim-3-positive NK cells considerably decreased in NDMM than HDs (Supplemental Figure S3B and Supplemental Table S5). Then, functions of Tim-3-positive NK and Tim-3-negative NK cells in each MM patient were compared. Expression of NK cell functional molecules on Tim-3-positive NK cells significantly decreased than those on Tim-3-negative NK cells (Supplemental Figure S3C). Tim-3 expression and NK cell functional molecules expression were all negatively correlated (Supplemental Figure S3D). These results confirmed that Tim-3 was a negative regulatory molecule for NK cells in MM. Next, Tim-3 ligands on BMSCs and human MM cell lines (U266, RPMI-8226) were detected via flow cytometry. Galectin-9 was highly expressed on BMSCs compared to HMGB1, CEACAM1 and PtdSer. However, no significant differences were seen in MM cells (Figure 2A; Supplemental Figure S4 and Supplemental Table S6). Furthermore, we observed the expression of galectin-9 on BMSCs was positively correlated with M protein and bone marrow plasma cell ratio in MM patients (Figure 2B; Supplemental Figures S7 and S8). We speculated that BMSCs regulate NK cells via Tim-3/ galectin-9 and confirmed it by using in vitro co-culture. The BMSCs/NK co-culture systems were constructed (Figure 2C). Expression of CD107a, NKG2D, INF-γ and perforin on NK cells was significantly decreased after co-culture with BMSCs for 6 days, while these markers were restored in the Tim-3 inhibitor group (Figure 2D and Supplemental Table S7). U266 cells were added to each co-culture group, and the apoptosis of U266 cells was examined after 72 h. The apoptosis ratios of U266 were significantly decreased after co-culture with BMSCs and restored in the Tim-3 inhibitor group (Figure 2E). Moreover, the NK cells functional molecules expression and the galectin-9 expression on BMSCs were negatively correlated (Figure 2F). Those results indicated that BMSCs can negatively regulate NK cell function by Tim-3/galectin-9. Whether BMSCs only regulated NK cells through direct contact? The BMSCs/NK indirect co-culture systems were constructed (Figure 3A). Expression of CD107a, NKG2D, INF-γ, and perforin was significantly decreased after co-culture with BMSCs, and these markers were restored in the Tim-3 inhibitor group and exosome inhibitor group (Figure 3B and Supplemental Table S8). The apoptosis of U266 was significantly decreased after co-culture with BMSCs and restored in the Tim-3 inhibitor group and exosome inhibitor group (Figure 3C). Then, we knocked down the galectin-9 on BMSCs to purified BMSCs-derived exosomes (Figure 3D; Supplemental Figures S5 and S6A; Supplemental Tables S9 and S10), Western-blot showed galectin-9 expression on galectin-9 knockdown BMSCs-derived exosomes was decreased significantly than BMSCs-derived exosomes9, 10 (Supplemental Figure S6B). After that, NK cells were co-cultured with BMSCs, galectin-9 knockdown BMSCs, BMSCs-derived exosomes, and galectin-9 knockdown BMSCs-derived exosomes respectively (Figure 3E). Expression of NK cell functional molecules was significantly restored in the galectin-9 knockdown BMSCs group compared to the normal BMSCs group; furthermore, expressions of NK cell functional molecules were significantly decreased after co-culture with BMSCs-derived exosomes, and these markers were restored in the galectin-9 knockdown BMSCs-derived exosomes group (Figure 3F). The apoptosis of U266 was significantly increased in the galectin-9 knockdown BMSCs-derived exosomes group than co-cultured with BMSCs-derived exosomes (Figure 3G). Finally, we detected the related protein changes in the Tim-3/galectin-9 pathway, Notch1 and EOMES expression decreased when NK co-culture with BMSCs, which indicated that the Tim-3/galectin-9 pathway might exhaust NK cells through down-regulating Notch1/EOMES (Figure 3H). In summary, NK cells are exhausted in the bone marrow micro-environment of MM patients. Mechanically, MM-derived BMSCs can inhibit the immune response of NK cells via Tim-3/galectin-9, both through direct cell-to-cell contact and indirect contact mediated by exosomes, and the application of Tim-3 inhibitors or exosome inhibitors can restore the exhaustion of NK cells (Figure 4). This work was supported by the National Natural Science Foundation of China Youth Project (grant no. 81900131), the Tianjin Municipal Natural Science Foundation (grant no. 18JCQNJC80400), the Tianjin Education Commission Research Project (grant no. 2018KJ043), the Tianjin Education Commission Research Project (grant no. 2018KJ045), the Tianjin Science and Technology Planning Project (grant no. 20YFZCSY00060), Tianjin Municipal Health Commission Youth Project (grant no. TJWJ2021QN001), Medjaden Academy & Research Foundation for Young Scientists (grant no. MJR20221011), Tianjin Key Medical Discipline(Specialty) Construction Project (grant no. TJYXZDXK-028A). The authors declare no conflicts interests. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.