All eight patients with MM were admitted to the Affiliated Hospital of Nantong University from September 2016 to December 2019, and the diagnostic criteria and treatment efficacy were referenced to the International Myeloma Working Group. All the cases included in this study were primary and untreated MM patients, with a mean age of 66 years (50–70 years). In the control group, 5 cases with a mean age of 62 years (50–70 years) were selected from patients with benign diseases (BD), including iron deficiency anemia and thrombocytopenia. The collection and use of clinical specimens were approved by the ethics committee of Affiliated Hospital of Nantong University with written informed consent. BM samples from newly diagnosed MM and BD patients were harvested from their iliac bones. Total 3–4 ml of BM was collected and placed into EDTA anticoagulant tubes. All procedures were carried out under sterile conditions.

BM mononuclear cells were isolated using human BM mononuclear lymphocytes separating medium (Biological Products Technology; Tianjin Haoyang; China). In brief, BM was added slowly into the separating medium and then centrifuged at 1,000 × g for 20 min at 20°C. Slow acceleration and deceleration settings were used. Finally, primary BMSCs were cultured and selected in plastic flasks by adherence method and used at the third to fourth passage. The MM cell lines RPMI 8226 and U266 used in in vitro experiments were purchased from the Cell Bank of the Chinese Academy of Sciences (Shanghai, China).

Flow cytometry was used to identify phenotypic surface markers of BMSCs with antibodies, including PE-conjugated anti-CD29, anti-CD73, anti-CD105, anti-CD11b, anti-CD34 and anti-CD45 antibodies (Cyagen, China). Appropriate isotype controls were used. Data analysis was performed on FlowJo 10.6.2 (FlowJo, LLC).

The osteogenic differentiation of BMSCs was performed using the osteogenic differentiation medium kit (HUXMA-90021, Cyagen, China). The osteogenic differentiation complete medium was prepared according to the kit instructions and culture medium was changed every 3 days. Finally, the osteogenic differentiation was performed and analyzed by alizarin red staining after the induction of differentiation.

Lipogenic differentiation of BMSCs was performed using the lipogenic differentiation medium kit (HUXMA-90031, Cyagen, China). Lipogenesis-induced differentiation A and B solutions were prepared according to the kit instructions and induced until large and round lipid droplets appeared. Then, Oil Red O was used to detect lipid droplets.

Total RNA was extracted from BMSCs isolated from MM and BD patients. RNA was quantified using a NanoDrop ND-2000 (Thermo Fisher Scientific, Waltham, MA, United States), and RNA integrity was assessed using Agilent Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, United States). After RNA qualification, sample labeling, microarray hybridization, and washing were performed according to the standard process. First, total RNA was transcribed to double-stranded cDNA and then Cyanine-3-CTP(Cy3)-labeled cRNA was synthesized; the labeled cRNA was hybridized to the microarray [miRNA: Agilent Human miRNA Microarray Kit, release 21.0,8 × 60K (DesignID:070,156)]. The labeled cRNA was hybridized to the chip, and the original image was scanned by Agilent Scanner G2505C (Agilent Technologies) after elution. After washing, the arrays were scanned using a G2505C scanner (Agilent Technologies).

Microarray analysis was performed by OE Biotech (Shanghai, China). Feature Extraction software (version10.7.1.1, Agilent Technologies) was used to analyze array images to get raw data. GeneSpring (version 14.8, Agilent Technologies) was employed to finish the basic analysis with the raw data. At first, the raw data were normalized with the quantile algorithm. The probes that at least one out of two conditions have flags in ‘Detected’ were chosen for further data analysis. Differentially expressed genes were then identified through fold change as well as p value calculated with t-test. The threshold set for up-and down-regulated genes was a fold change ≥1.5 and a p-value ≤ 0.05. Afterward, GO and KEGG analyses were applied to determine the roles of these differentially expressed mRNAs. Finally, Hierarchical Clustering was performed to display the differentially expressed genes’ expression pattern among samples.

Total RNA was extracted with Trizol reagent (Invitrogen, CA, United States), and the cDNA was synthesized using a reverse transcription kit (Ribobio, China or Vazyme, Nanjing, China) according to the manufacturer’s instructions. miRNA-related primers were produced by Ribobio Company. The sequences of β-Actin primers were as follows: forward, 5′-CAC​GAA​ACT​ACC​TTC​AAC​TCC-3’; reverse, 5′-CAT​ACT​CCT​GCT​TGC​TGA​TC-3′. The sequences of TIMP2 primers were as follows: forward, 5′-GCT​GCG​AGT​GCA​AGA​TCA​C-3′; reverse, 5′-TGG​TGC​CCG​TTG​ATG​TTC​TTC-3′. We used quantitative PCR to detect the expression of miR-483-5p (miDETECT A TrackTM, Ribobio, China) and TIMP2 (Vazyme, Nanjing, China). U6 snRNA or β-Actin were used as internal controls. All samples were run in triplicate, and all reactions were repeated three times independently to ensure reproducibility.

Exosomes were isolated using the ultracentrifugation method. In brief, cells were grown in α-MEM medium supplemented with 10% exosome-depleted FBS. After 48 h, the culture medium was collected and subjected to sequential centrifugation at 300 × g for 20 min, 2,000 × g for 20 min, and 10,000 × g for 30 min at 4°C to remove residual cells, debris and organelles. Then, ultracentrifugation at 100,000 × g for 2 h was used to collect exosomes.

miR-483-5p inhibitor (5′-CCC​UCC​ACC​AUG​CAA​GGG​AUG-3′), inhibitor negative control (5′-CUC​CCU​UCU​UUC​CUC​CCG​UCU​U-3′), miR-483-5p mimics (sense, 5′-AAG​ACG​GGA​GGA​AAG​AGG​GAG-3’; antisense, 5′-CCC​UUC​UUU​CCU​CCC​GUC​UUU​U-3′) and mimic negative control (sense, 5′-UUC​UCC​GAA​CGU​GUC​ACG​UTT-3’; antisense, 5′-ACG​UGA​CAC​GUU​CGG​AGA​ATT-3′) (GenePharma, Shanghai, China) were transfected into the cells by using lipofectamine 2000 (Invitrogen, CA, United States) in serum-free medium. After 6 h of incubation, the medium was changed to complete medium.

Cell apoptosis was detected by flow cytometry using Annexin V-Alexa Fluor 647/PI apoptosis detection kit (FcMACS, Nanjing, China). The transfected cells were washed twice with pre-cooled PBS at 4°C and resuspended with the binding buffer and 100 μL of cell suspension (1×10⁶/ml) was incubated with 5 μL of Annexin V-Alexa Fluor 647 and 10 μL of 20 μg/ml of propidium iodide solution at room temperature for 15 min. Finally, 400 μL of PBS was added to the reaction tube and the cells were analyzed by flow cytometry (Beckman Coulter, United States).

Targetscan (http://www.targetscan.org/vert_80/), ENCORI (https://starbase.sysu.edu.cn/), miRDB (http://mirdb.org/) and miRwalk (http://mirwalk.umm.uni-heidelberg.de/) were used to predict the potential target mRNAs of miR-483-5p. RPMI 8226 and U266 cells were co-transfected with miR-483-5p mimics and the luciferase reporter vector containing wild type (WT) 3′-UTR of TIMP2 (GenePharma, Shanghai, China) by lipofectamine 2000. The medium was changed after 6 h of transfection and cells were lysed after 36 h of incubation. At last, luciferase activity was detected by the Reporter Assay Program Dual-Luciferase (Promega).

All data were expressed as means ± SD. All statistical analysis was performed using IBM SPSS 21.0 (SPSS, Chicago, IL, United States) or Graph Pad Prism 7.0 (Graph Pad Software, Inc. San Diego, CA, United States). All reported p-values are two-tailed, and p < 0.05 was considered statistically significant.

BMSCs were isolated by density gradient centrifugation and adherence method. Then we identified BMSCs from BD and MM patients by cell morphology, surface markers and differentiation ability.

Microscopic observations showed that the cell morphology of BD-MSCs and MM-MSCs were similar, both showing long fiber-like spindle shapes (Figure 1A). Flow cytometry results showed that BMSCs positively expressed CD29, CD73, CD105 (>97%), while negatively expressed CD45, CD34, and CD11b (<0.3%) (Figures 1B,C). All results showed that these cells contain a characteristic set of surface markers of BMSCs (Mushahary et al., 2018).

Osteogenesis induction results showed that primary cultured BMSCs formed small lamellar mineralized nodules after 21 days of culture, but the bone nodules from BD-MSCs were darker and larger in appearance compared to those from MM-MSCs (Figure 1D). The results of lipogenesis induction showed that primary cultured BMSCs were easily induced to differentiate into adipogenic direction after 21 days. After Oil Red O staining, a large number of granular red lipid droplets in the cytoplasm were shown. Moreover, MM-MSCs had more lipid droplets (Figure 1E). Therefore, we speculated that the osteogenic capacity of MSCs in MM patients was reduced while their lipogenic capacity was enhanced.

To explore the differences between BMSCs of MM and BD patients, We selected two paired BD-BMSCs and MM-BMSCs for differential gene screening by gene microarray, and the screening criteria were up-regulated or down-regulated fold change value ≥ 1.5 and p-value ≤0.05. Cluster and scatter plots demonstrated the differentially expressed miRNAs among them (Figures 2A,B). The clustered map showed 39 up-regulated miRNAs, including miR-483-5p and 32 down-regulated ones in MM-MSCs compared to BD-MSCs (Figure 2A). qPCR verification showed that miR-483-5p was up-regulated in BMSCs from MM patients compared to BD patients (Figure 2C).

GO enrichment analysis bar graphs showed the entries of differential gene-related biological processes (BP), cellular components (CC), and molecular functions (MF). The different color distributions correspond to BP, CC, and MF. We performed enrichment analyses for all, up-regulated, and down-regulated differential genes for BP, CC, and MF, respectively. The top 10 most significant GO entries in the three major GO categories were shown (Figure 2D).

KEGG enrichment analysis was performed to explore pathways associated with differential genes. This KEGG enrichment analysis was performed separately for all, up-regulated, and down-regulated differential genes (Figure 2E). KEGG enrichment analysis explored the signaling pathways that might be associated with the differential genes (Figure 2E).

Exosomes act as intermediaries for intercellular communication. Exosomes from BMSCs were isolated by ultracentrifugation and characterized by Western blot (Figure 3A), NTA (Figure 3B) and TEM (Figure 3C). Interestingly, MM cells treated with conditioned medium (CM) from MM-MSCs showed an enhanced proliferation ability than those treated with CM from BD-MSCs (Figure 3D). We also observed a similar effect in MM cells treated with exosomes from MM-MSCs (Figure 3E). Furthermore, MM-MSC-Ex treatment significantly increased miR-483-5p expression in MM cells (Figure 3F).

To explore the role and mechanism of miR-483-5p in the malignant progression of MM cells, we transfected miR-483-5p inhibitor into U266 cells, which expressed higher miR-483-5p than RPMI 8226 (Supplementary Figure S1). First, we examined the transfection efficiency of miR-483-5p inhibitor, and we found that miR-483-5p level was significantly decreased after transfection (Figure 4A). In addition, cell proliferation ability of U266 was impaired after transfection (Figure 4B), while cell apoptosis was promoted (Figure 4C). In accordance, Western blot results revealed that miR-483-5p inhibitor resulted in a down-regulated level of c-Myc, Bcl-2, and up-regulated p21 expression in MM cells (Figure 4D), which are markers of cell proliferation and apoptosis. In contrast, the opposite results were observed with transfection of miR-483-5p mimics in RPMI 8226 cell line (Figures 4E–G). In summary, miR-483-5p impacted MM cell proliferation and apoptosis, thereby accelerating the malignant process of MM.

To search for a potential downstream target of miR-483-5p, we used miRNA prediction programs including Targetscan, ENCORI, miRDB and miRwalk as bioinformatics methods. TIMP2 was one intersection among them (Figure 5A). miR-483-5p and TIMP2 contained a perfect 7mer-m8 seeding match (Figure 5B). To detect their relation, we found that TIMP2 mRNA and protein expression was negatively correlated with miR-483-5p, tested by qRT-PCR and Western blot, respectively (Figures 5C,D). To further verify the relationship between miR-483-5p and TIMP2, the TIMP2 luciferase reporter plasmid was co-transfected with miR-483-5p mimics into MM cells. Compared to nective control, co-transfection with pmirGLO-TIMP2 WT and miR-483-5p mimics suppressed the luciferase activity by around 50% (Figure 5E). We speculated that miR-483-5p might promote MM progression by targeting TIMP2. For further analysis, we downloaded the GSE6477 dataset from GEO database, and we found that TIMP2 expression was lower in both newly diagnosed and relapsed MM patients than normal donors, which was consistent with our expectation. These results suggested that miR-483-5p/TIMP2 axis may play a vital role in MM (Figure 6).