Two miRNA gene expression profiles by microarray (GSE37053 and GSE24371) from MM patients were acquired from the GEO database (https://www.ncbi.nlm.nih.gov/geo/)9, 10, 11. The dataset GSE37053 was generated in the GPL8227 platform (Agilent-019118 Human miRNA Microarray 2.0 G4470B) by performing miRNA profiling of CD138 plasma cells obtained from 39 MM patients and 18 disease controls. GSE24371 was produced in the GPL10385 platform (microRNA spotted oligo array version 17.0) by profiling the miRNAs of CD138 plasma cells purified from 33 MM patients and 9 disease controls. All samples in these two datasets are categorized as either MM or control.

Analysis of differentially expressed miRNAs (DEmiRs) was performed using the GEO2R platform (https://www.ncbi.nlm.nih.gov/geo/geo2r /). The GEO2R platform analyzed and presented the DEmiRs between MM patients and disease controls. In the GSE37053 and GSE24371 datasets, miRNAs with a |log₂fold change (FC)|≥ 0.4 were identified as significant DEmiRs using limma packages from the Bioconductor project12. Venn diagrams of the DEmiRs of GSE37053 and GSE24371 datasets were drawn using an online tool, Venny 2.1.0 (https://bioin fogp.cnb.csic.es/tools /venny /). Common DEmiRs from these two datasets were selected for further analysis.

Target genes of selected common DEmiRs were predicted and identified using the bioinformatics prediction software miRNA Data Integration Portal (mirDIP; (http://ophid.utoronto.ca/mirDIP)13. This software can obtain the data from 26 different databases for miRNAs to generate robust data on miRNA target genes.

Gene ontology (GO) comprises three classes of data: biological processes, molecular functions, and cellular components. GO and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment for probing the function of miRNA target genes were analyzed using the Database for Annotation Visualization and Integrated Discovery (DAVID 6.8, https://david.ncifc rf.gov/tools.jsp) and R software14. Terms with P < 0.05 were considered significantly enriched.

To examine the connections among the target genes, a protein–protein interaction (PPI) network was constructed using the Search Tool for the Retrieval of Interacting Genes database (STRING version 11.5, http://stringdb.org)15. Cytoscape (version 3.7.2) software was used for the visualization of the PPI network16.

Interactions between circRNA and the selected common DEmiR were determined using the web tool Circular RNA Interactome (CircInteractome, https://circinteractome.nia.nih.gov/)17. This computational platform is able to predict and map the interaction sites between circRNA and miRNA.

The human MM cell line RPMI 8226 was procured from the National Centre for Cell Science, India. The BTZ-resistant RPMI 8226/PS100 cell line, which is resistant to 100 nM BTZ and was generated by exposing RPMI 8226 cells to stepwise increasing concentrations of BTZ with a starting concentration of 0.2 nM for approximately 15 months, was donated by Prof. Jacqueline Cloos18. DG75, a non-myeloma B-lymphoblast cell line, was used as a control. All cells were cultured in RPMI 1640 (Gibco) media supplemented with 10% (v/v) heat-inactivated fetal bovine serum (Gibco). BTZ-resistant RPMI 8226/PS100 cells were cultured and treated with 100 nM BTZ (Natco Pharma). Cell lines were maintained at 37°C with 5% CO₂.

Total RNA was isolated using RNA-XPress reagent (HiMedia) and the concentration and quality were assessed using a NanoDrop One (Thermo Fisher Scientific). A total of 1.5 μg of total RNA was used for cDNA synthesis using the iScript cDNA Synthesis Kit (Bio-Rad) according to the manufacturer’s protocol (5 min at 25°C, 20 min at 46°C, and 1 min at 95°C). Quantitative RT-PCR was performed using SYBR Green reagent (Bio-Rad) in a CFX-Connect real time system (Bio-Rad). HPRT1 was used as the endogenous control to normalize the expression of genes using the 2^{- ΔΔCt} method. All the samples were run in technical triplicate for 40 cycles of 5 min at 95°C, 10 sec at 95°C, 20 sec at 52°C, and 30 sec at 72°C. The primer sequences used for quantitative RT-PCR analysis in this study were as follows:

Stem-loop primer for hsa-mir-151a-3p:

Cells were harvested and lysed with ice-cold radioimmunoprecipitation assay buffer [20 mM Tris-HCl (pH 8.0), 137 mM NaCl, 10% glycerol, 1% Nonidet P-40, 2 mM EDTA, 200 mM Na3VO4, 100 mM phenyl-methyl-sulfonyl fluoride, protease inhibitor cocktail, and phosphatase inhibitor]. Protein concentration was measured using Bradford reagent (Bio-Rad). Equal amounts of protein were loaded and run on sodium dodecyl sulfate–polyacrylamide gel electrophoresis and electro-transferred onto a nitrocellulose membrane (G-Biosciences). The membrane was blocked with 5% non-fat dry milk in TBS containing 0.1% Tween 20 (TBST). Primary antibodies against p53 (Santa Cruz) and β-actin (Cell Signaling Technology) were incubated with membranes at 4°C overnight. After incubation with primary antibody, membranes were washed with TBST and probed with secondary antibodies (anti-mouse and anti-rabbit conjugated to horseradish peroxidase; Cell Signaling Technology). The blots were then developed using enhanced chemiluminescent substrate (Thermo Fisher Scientific).

All values are represented as mean ± SEM, and p-values < 0.05 were considered statistically significant. The significance was calculated using a standard Student’s t-test, with * p < 0.05, ** p < 0.01, and *** p < 0.001. The gene expression levels were analyzed with an unpaired two-tailed t-test. All the data were analyzed using GraphPad Prism software (version 5.01).

Table 1

Regulation	miRNA Symbol
Up-regulated	hsa-miR-222, hsa-miR-221, hsa-miR-188-5p, hsa-miR-96, hsa-miR-29c, hsa-miR-148a, hsa-miR-874, hsa-miR-34c-3p, hsa-miR-151-3p
Down-regulated	hsa-miR-142-3p, hsa-miR-26b, hsa-miR-103, hsa-miR-185, hsa-miR-25, hsa-miR-33a, hsa-miR-107

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

The microarray-based gene expression profiles GSE37053 and GSE24371 were analyzed to identify miRNAs that were differentially expressed between MM patients and disease controls. Data normalization of these two datasets was performed to compare the gene expression between MM and control (Figure 1 a,b). miRNAs with |log₂FC| ≥ 0.4 and adjusted p < 0.05 were considered DEmiRs. The common DEmiRs from these two datasets consisted of 16 miRNAs. Among these common miRNAs, 9 miRNAs were upregulated and 7 miRNAs were downregulated (Figure 1 c,d; Table 1).

Among 16 DEmiRs, one miRNA, hsa-mir-151a-3p, whose role has not yet been examined in MM to our knowledge, was selected for further investigation. hsa-mir-151a-3p was found to be upregulated significantly in both the datasets of MM patients, GSE37053 and GSE24371. Its expression was assessed in the RPMI 8226 MM cell line and BTZ-resistant RPMI 8226/PS100 cells, and it was found to be significantly upregulated in both MM cells and BTZ-resistant MM cells (Figure 2). Therefore, hsa-mir-151a-3p may have a role in the pathogenesis of MM as well as in drug resistance mechanisms.

Analysis with the mirDIP database detected several genes that could be targeted by hsa-mir-151a-3p with very high confidence. A total of 159 genes showing interactions with the miRNA with very high confidence (top 1%) were identified and selected for further analysis (Supplementary File 1).

To explore the biological function of target genes of hsa-mir-151a-3p, GO and KEGG pathway enrichment analysis were performed. The top GO terms were determined based on the counts of the genes from the DAVID database (version 6.8). GO analysis revealed that the target genes were enriched in the biological processes positive regulation of angiogenesis, regulation of cell cycle, intracellular signal transduction, apoptotic process, regulation of transcription, DNA-templated, and regulation of transcription from RNA polymerase II promoter (Figure 3 a); the cellular components extracellular exosome, perinuclear region of cytoplasm, cytoplasm, nucleoplasm, and nucleus (Figure 3 b); and the molecular functions p53 binding, ubiquitin protein ligase binding, ATPase activity, protein serine/threonine/tyrosine kinase activity, chromatin binding, and mRNA binding (Figure 3 c). KEGG analysis showed that the most significantly enriched pathways were the inflammation mediated by chemokine and cytokine signaling pathway, the p53 pathway, the PI3 kinase pathway, the PDGF signaling pathway, the Wnt signaling pathway, angiogenesis, the insulin/IGF pathway-protein kinase B signaling cascade, the MAPK cascade, hypoxia response via HIF activation, the apoptosis signaling pathway, the interleukin signaling pathway, the integrin signaling pathway, and the Ras pathway (Figure 3 d).

The PPI network of the 159 target genes of hsa-mir-151a-3p revealed various clusters of target genes (Figure 4). One prominent cluster was found around p53, which interacted with AKT3, PTEN, HIF1A, FOXP1, SMAD2, MAF, AGO2, AGO3, and other gene products. Therefore, it could be presumed that p53 is involved in the process of pathogenesis in MM.

Analysis using the CircInteractome database identified several circRNAs that could potentially interact with hsa-mir-151a-3p. One such circRNA was hsa_circ_0000073, which could interact with hsa-mir-151a-3p (Figure 5 a). The expression of this circRNA was assessed in RPMI 8226 MM cells and BTZ-resistant RPMI 8226/PS100 MM cells. hsa_circ_0000073 was found to be significantly downregulated in BTZ-resistant MM cells (Figure 5 b). Thus, it could be speculated that the downregulation of this circRNA may be one of the causes of the upregulation of the miRNA hsa-mir-151a-3p in BTZ-resistant MM and may contribute to drug resistance in MM.

As p53 was found to be targeted by the miRNA hsa-mir-151a-3p by mirDIP software and forms a prominent cluster in the PPI network of target genes of hsa-mir-151a-3p, the expression of p53 was further examined in RPMI 8226 MM cells as well as BTZ-resistant MM cells at both the transcript and protein levels. p53 was found to be significantly downregulated in both MM cells and BTZ-resistant MM cells at both the mRNA and protein levels (Figure 6 b,c,d). Thus, it could be assumed that downregulation of p53 might impart drug resistance in MM.