Epigenetic silencing of miR-340-5p in multiple myeloma: mechanisms and prognostic impact

Background miR-340-5p, localized to 5q35, is a tumor suppressor miRNA implicated in multiple cancers. As a CpG island is present at the putative promoter region of its host gene, RNF130, we hypothesized that the intronic miR-340-5p is a tumor suppressor miRNA epigenetically silenced by promoter DNA methylation of its host gene in multiple myeloma. Results By pyrosequencing-confirmed methylation-specific PCR, RNF130/miR-340 was methylated in 8/15 (53.3%) myeloma cell lines but not normal plasma cells. Methylation correlated inversely with the expression of both miR-340-5p and RNF130. In completely methylated WL-2 and RPMI-8226R cells, 5-AzadC treatment led to demethylation and re-expression of miR-340-5p. In primary samples, RNF130/miR-340 methylation was detected in 4 (22.2%) monoclonal gammopathy of undetermined significance, 15 (23.8%) diagnostic myeloma, and 7 (23.3%) relapsed myeloma. RNF130/miR-340 methylation at diagnosis was associated with inferior overall survival (median 27 vs. 68 months; P = 3.944E−5). In WL-2 cells, overexpression of miR-340-5p resulted in reduced cellular proliferation [MTS, P = 0.002; verified in KMS-12-PE (P = 0.002) and RPMI-8226R (P = 2.623E−05) cells], increased cell death (trypan blue, P = 0.005), and enhanced apoptosis by annexin V-PI staining. Moreover, by qRT-PCR, overexpression of miR-340-5p led to repression of both known targets (CCND1 and NRAS) and bioinformatically predicted targets in MAPK signaling (MEKK1, MEKK2, and MEKKK3) and apoptosis (MDM4 and XIAP), hence downregulation of phospho-ERK1/2 and XIAP by Western blot. Furthermore, by qRT-PCR, in CD138-sorted primary samples (n = 37), miR-340-5p and XIAP were inversely correlated (P = 0.002). By luciferase assay, XIAP was confirmed as a direct target of miR-340-5p via targeting of the distal but not proximal seed region binding site. Conclusions Collectively, tumor-specific methylation-mediated silencing of miR-340-5p is likely an early event in myelomagenesis with adverse survival impact, via targeting multiple oncogenes in MAPK signaling and apoptosis, thereby a tumor suppressive miRNA in myeloma. Electronic supplementary material The online version of this article (10.1186/s13148-019-0669-2) contains supplementary material, which is available to authorized users.


Background
Multiple myeloma is the second most common blood cancer, accounting for approximately 10% of all hematologic malignancies [1]. Patients with myeloma are often preceded by an asymptomatic pre-malignant stage of monoclonal gammopathy of undetermined significance (MGUS), which progresses into symptomatic myeloma at a rate of 1% per year [2]. Genetically, about half of patients with myeloma carries a non-hyperdiploid karyotype, which is associated with recurrent translocations involving immunoglobulin gene located at 14q32, whereas the other half carries a hyperdiploid (HPRD) karyotype, characterized by trisomies of odd number chromosomes [3]. Apart from chromosomal copy number alterations, recurrent genetic mutations have been identified, in particular, amongst genes involved in RAS (NRAS, KRAS, and BRAF), NF-kB (TRAF3, CYLD, and LTB), and TP53 (TP53, ATM, and ATR) signaling pathways [4][5][6].
MicroRNAs (miRNAs) are endogenous, single-stranded, small non-coding RNAs of~22 nt in length [7,8]. Functionally, a miRNA will target and suppress the expression of their target protein-coding genes by complementary binding of the seed region, i.e., the second to seventh nucleotides, to the seed region binding site (SRBS) in the 3′-UTR of the targeted mRNAs of protein-coding genes, leading to translational blockade or mRNA degradation [9]. Deregulated miRNA expression has been found in cancers including hematologic malignancies [10,11], in which oncogenic miRNAs targeting tumor suppressor genes are upregulated, whereas tumor suppressive miR-NAs targeting oncogenes are downregulated [8,11]. For instance, overexpression of miR-21, of which PTEN was a direct target, led to the activation of PI3K/AKT signaling pathway, and hence, miR-21 is an oncomiR promoting cellular proliferation [12]. On the other hand, miR-30c, targeting BCL9, has been shown downregulated in myeloma. BCL9 is an essential effector component for transcription of oncogenic Wnt target genes [13]. Moreover, restoration of miR-30c led to inhibition of cell proliferation, invasion, and migration in addition to enhancing apoptosis of myeloma cells, hence a tumor suppressor miRNA [14].
DNA methylation refers to the addition of a methyl (-CH 3 ) group to carbon five position of the cytosine ring in a CpG dinucleotide [15]. CpG dinucleotides may cluster as a CpG island, which is defined as any genomic region of > 200 bp with a high GC content of > 50% and a high ratio of observed/expected CpG > 0.60 [16,17]. In the mammalian genome, promoter-associated CpG islands are localized to or in close proximity to the promoter region of more than half of the human genes [18] and involved in the regulation of gene expression by DNA methylation [19]. In normal cells, the majority of promoter-associated CpG islands are unmethylated, associated with a euchromatin configuration, and hence transcriptionally ready or active for gene expression [20]. Conversely, CpG islands/sites in the intergenic regions, imprinted regions, and X-chromosome are hypermethylated, leading to repression of repetitive elements, such as SINE and LINE elements, imprinted genes, and X-linked genes respectively [19]. In contrast to normal cells, cancer cells are characterized by global DNA hypomethylation and locus-specific hypermethylation of promoterassociated CpG islands of tumor suppressor genes or miR-NAs, resulting in downregulation, and hence loss of tumor suppressor functions [8,21,22]. For instance, in myeloma, tumor suppressive miR-34b/c [23], miR-203 [24], and miR-129-2 [25] have been shown to be silenced by promoter DNA methylation. Moreover, epigenetic silencing of miR-137 has been found to correlate with shorter progression-free survival in myeloma [26]. miR-340-5p, localized to 5q35, is embedded in the second intron of its host gene, RNF130, and has been shown to be a tumor suppressor downregulated in several cancers, such as breast cancer [27], ovarian cancer [28], and hepatocellular carcinoma [29]. In hepatocellular carcinoma, overexpression of miR-340-5p by oligo transfection resulted in inhibition of cell proliferation, migration, and invasion in vitro and suppression of tumor growth in vivo by directly targeting Janus kinase 1 [29]. As a CpG island is present at the promoter region of its host gene, RNF130, we hypothesized that miR-340-5p is an intronic tumor suppressor miRNA epigenetically silenced by RNF130/miR-340 promoter DNA hypermethylation in multiple myeloma (Additional file 3: Figure S3).
Methylation and expression of RNF130/miR-340 in primary bone marrow samples of MGUS, diagnostic myeloma, and relapsed myeloma By MSP, methylation of RNF130/miR-340 was detected in 4 (22.2%) MGUS, 15 (23.8%) diagnostic myeloma, and 7 (23.3%) relapsed myeloma primary bone marrow samples (Fig. 3a). Methylation frequency of RNF130/miR-340 was not significantly different among those consecutive clinical stages of myeloma (MGUS vs. diagnostic myeloma: P = 1.000; diagnostic myeloma vs. relapsed myeloma: P = 1.000). However, in contrast to the absence of methylation in normal, the appearance of methylation in MGUS and a comparable frequency in consecutive stages from MGUS to diagnostic myeloma and to relapsed myeloma indicated it might be an early event in the pathogenesis of myeloma. Moreover, among 26 diagnostic samples with paired CD138-sorted RNA samples, methylation of RNF130/miR-340 had a trend of associating with lower expression of miR-340-5p ( Fig. 3b; P = 0.223).
Of the known targets of miR-340-5p, overexpression of miR-340-5p resulted in significant downregulation of CCND1 and NRAS and all of the predicted target genes, including MEKK1, MEKK2, MEKKK3, MDM4, and XIAP (Fig. 5a). Moreover, by Western blot, protein expression level of XIAP, which is an inhibitor of apoptosis, and p-ERK1/2, which is a key effector downstream to NRAS, MEKK1, and MEKKK3 in MAPK signaling, were decreased by 27% and 60% respectively upon overexpression of miR-340-5p (Fig. 5b). Furthermore, in CD138-sorted primary samples (n = 37), a higher expression level of miR-340-5p was significantly correlated with a lower level of XIAP by qRT-PCR (Fig. 5c, R 2 = 0.2466, P = 0.002). These data suggested that miR-340-5p might exert its tumor suppressive function by regulating multiple oncogenic target genes involved in MAPK signaling and apoptotic pathways.
Identification of XIAP as a direct target of miR-340-5p As overexpression of miR-340-5p had induced a significant increase of apoptosis in myeloma cells (Fig. 4d), luciferase reporter assay was employed to verify if XIAP, an inhibitor of apoptosis, is a direct target of miR-340-5p. By bioinformatics, two conserved miR-340-5p SRBSs were identified in the 3′-UTR of XIAP (Additional file 2: Figure S2A and Fig. 6a). DNA fragments containing either wild-type or mutant SRBS were generated and cloned into a dual firefly/ Renilla luciferase reporter vector (Additional file 2: Figure  S2B and Fig. 6b). For each SRBS, luciferase vector containing the wild-type or mutant 3′-UTR of XIAP was co-transfected with miR-340-5p mimics or scramble control into Hela cells for luciferase assay at 48 h. Upon co-transfection of the wild-type SRBS 1, which was predicted to form 7mer-A1 binding with miR-340-5p, overexpression with miR-340-5p mimics resulted in comparable luciferase signal as compared with scramble control (Additional file 2: Figure S2C). Moreover, co-transfection of The numbers were assigned for illustration purpose, and hence, the identical Arabic numerals in different disease stages are not serial samples from the same patient. b In 26 patients with both CD138-sorted DNA and RNA, there was a trend that patients with methylation of RNF130/miR-340 (n = 5) had a lower expression and hence larger ΔCt of mature miR-340-5p than patients without methylation (n = 21). c Kaplan-Meier analysis of OS in patients with and without methylation of RNF130/miR-340 mutant SRBS 1 with miR-340-5p mimics had a similar luciferase activity as compared with scramble control (Additional file 2: Figure S2C). In contrast, upon co-transfection with the wild-type SRBS 2, which was predicted to form 8mer binding with miR-340-5p, overexpression with miR-340-5p mimics (Fig. 6c) led to a reduction of luciferase activity by 34.9%, as compared with scramble control ( Fig. 6d; P = 9.981E−05). Moreover, upon co-transfection of the mutant SRBS 2 with miR-340-5p, the luciferase activity was restored to a comparable level as compared with scramble control ( Fig. 6d; P = 0.175). Thus, these data suggested that XIAP is a direct target of miR-340-5p through binding at SRBS 2 in the 3′-UTR.

Discussion
There are a number of interesting observations in this study.
Firstly, methylation of miR-340-5p in myeloma cell lines and primary myeloma cells was tumor-specific as evidenced by the absence of methylation in normal controls, similar to the tumor-specific methylation of other tumor suppressive miRNAs, such as miR-124 [34,35], miR-203 [24], and miR-34 family miRNAs [23,36] in myeloma. This contrasted with the tissue-specific methylation of miRNAs [37,38], in which miRNA methylation was detected in both tumor cells and the corresponding normal counterparts, and hence likely unimportant in carcinogenesis.
Secondly, in primary samples, methylation of RNF130/ miR-340 appeared as early as MGUS, at a frequency comparable to that of myeloma at diagnosis and relapsed myeloma. Therefore, it is likely that methylation of RNF130/miR-340 is an early event in the pathogenesis of myeloma, similar to methylation of miR-203 [24] and Fig. 4 Restoration of miR-340-5p in HMCLs. a By qRT-PCR, miR-340-5p was shown to be successfully overexpressed in KMS-12-PE, WL-2, and RPMI-8226R cells. By MTS assay, overexpression of miR-340-5p significantly inhibited cellular proliferation in all three cell lines. b By trypan blue staining, overexpression of miR-340-5p increased the number of dead cells by 3.73% in WL-2 cells. Error bars represent the standard deviation from three independent transfections. c By annexin V/PI analysis, overexpression of miR-340-5p increased cellular apoptosis by 9.60% in WL-2 cells miR-342 [39]. By contrast, miR-129-2 methylation was implicated in the progression from MGUS to symptomatic myeloma [25], and miR-34b/c methylation at relapse or progression of myeloma [23]. Moreover, methylation of RNF130/miR-340 correlated with shorter OS in newly diagnosed myeloma, similar to CDKN2A [40,41] and DAPK1 [42] methylation. As this cohort of myeloma patients was uniformly treated [43] with bortezomib-based induction regimen, followed by ASCT, and then thalidomide maintenance, methylation of RNF130/miR-340 was a potential novel prognostic marker for myeloma and hence warrants a prospective study in larger cohorts of myeloma samples. Similarly, dysregulated expression of a number of miRNAs, such as miR-135a/b, miR-200a/b, and miR-596, has been shown to carry prognostic impact in myeloma [44]; hence, the prognostic significance of miR-340 expression warrants further investigation.
Thirdly, we showed that the expression of miR-340-5p was regulated by promoter hypermethylation of the host gene. First, we showed that miR-340-5p expression correlated with that of RNF130, consistent with the data showing that intronic miRNAs are transcribed in parallel with their host genes, hence are co-regulated [45]. Similarly, when miRNA expression were correlated with gene expression in myeloma cell lines [46], co-expression of 32 pairs of intronic miRNAs and their host genes were found, including miR-340/RNF130. Moreover, we showed low expression of both miR-340-5p and RNF130 in methylated myeloma lines and high expression in unmethylated lines. Furthermore, upon 5-AzadC demethylation, RNF130 promoter demethylation was associated with re-expression of miR-340-5p. Therefore, our data provided additional evidence that intronic miRNAs are silenced by promoter DNA methylation of their host genes [39,47].  5 Effect of miR-340-5p in the regulation of target genes and signaling pathways. a Upon overexpression of miR-340-5p, relative expression levels of myeloma-related known target genes of miR-340-5p, including CCND1, MDM2, NFKB1, and NRAS, and bioinformatically predicted oncogenic target genes involved in MAPK (MEKK1, MEKK2, and MEKKK3), apoptosis (XIAP), and TP53 (MDM4) signaling pathways were shown. Error bars represent the standard deviation from three independent qRT-PCR. b Western blot analysis of phospho-ERK1/2, total-ERK1/2 (downstream to MAPK signaling), and XIAP upon transfection with miR-340-5p mimics or scramble control were shown. β-Actin was set as the endogenous control and normalizer for densitometric analysis of protein levels. Relative normalized protein levels are shown above the corresponding band. c In CD138-sorted primary samples, using ΔCt, the expression of miR-340-5p was plotted against XIAP, and a significant inverse correlation was demonstrated Fourthly, our data demonstrated that miR-340-5p is a tumor suppressor in myeloma. This is supported by the downregulation of mRNA of both known targets (NRAS and CCND1) and bioinformatically predicted targets (MEKK1, MEKK2, MEKKK3, and XIAP) upon miR-340-5p overexpression. Moreover, by Western blot, the protein level of p-ERK1/2, a key effector of MAPK signaling downstream to NRAS, MEKK1, and MEKKK3, was found repressed upon overexpression of miR-340-5p. Therefore, methylation-mediated silencing of miR-340-5p may account for the constitutive activation of MAPK signaling in myeloma pathogenesis [48,49]. Moreover, overexpression of miR-340-5p led to downregulation of CCND1, which is responsible for the transition from G1 to S phase of the cell cycle. Interestingly, upregulation of D-type cyclins is a unified theme in the pathogenesis of myeloma [50]. As CCND1 that encodes for cyclin D1 has been proven to be a direct target of miR-340-5p by luciferase assay [32], our data of methylation-mediated silencing of miR-340-5p may account for the overexpression of CCND1 in myeloma [50]. On the other hand, we showed that XIAP, an inhibitor of apoptosis and an oncogene overexpressed in multiple cancers, is inversely correlated with and a direct target of miR-340-5p. These are evidenced by, firstly, the lower expression of XIAP correlated with higher expression of miR-340-5p in CD138-sorted primary samples and, secondly, by the suppression of luciferase activity upon co-transfection with miR-340-5p and the WT XIAP 3′ UTR, which was restored by co-transfection with miR-340-5p and the mutant XIAP 3′ UTR construct. Interestingly, the suppression of XIAP by miR-340-5p was mediated by the distal (8mer) but not the proximal (7mer-A1) SRBS, hence consistent with the higher predictive value of SRBS by 8mer than 7mer-A1 [7]. Indeed, high XIAP expression has been demonstrated in primary myeloma plasma cells and cell lines and yielded larger tumors than myeloma cells with XIAP knock-down in a mouse xenograft model [51]. Therefore, the tumor suppressor activity of miR-340-5p was mediated by targeting, and hence downregulation, of XIAP in myeloma. Collectively, these results indicated that miR-340-5p is a tumor suppressor miRNA in myeloma by inhibition of MAPK signaling, cell proliferation, and induction of apoptosis. Conversely, methylation-mediated silencing of miR-340-5p enhanced myeloma plasma cell proliferation via upregulation of CCND1 and MAPK signaling (NRAS/ MEKK1/MEKKK3/p-ERK) in addition to enhancing myeloma cell survival by targeting XIAP. Similarly, miR-340 has been shown to carry tumor suppressor function by inhibiting myeloma cell-induced angiogenesis upon co-culture of myeloma cells with exosomes enriched in miR-340 [52].

Conclusions
In conclusion, in myeloma, methylation-mediated silencing of miR-340-5p is tumor-specific, reversible, associated with inferior OS, and likely an early event in Fig. 6 Identification of XIAP as a direct target of miR-340-5p. a Location, sequence, and predicted binding of miR-340-5p seed region to seed region binding site 2 (SRBS 2) on 3′-UTR of XIAP mRNA, and the design of mutant sequence. b Direct sequencing confirmed the sequences of both wild-type (5′-CTTTATAA-3′) and mutant (5′-CGGGGGAA-3′) in the corresponding luciferase reporter vector. c By qRT-PCR, miR-340-5p was overexpressed when co-transfection with luciferase reporter vector containing wild-type or mutant SRBS 2. d Normalized luciferase activity of luciferase reporter vector carrying wild-type or mutant SRBS 2, in the presence of miR-340-5p mimics or scramble control. Error bars represent the standard deviation from three independent transfections myelomagenesis. Moreover, miR-340-5p potentially exerts its tumor suppressive function via regulation of MAPK signaling and apoptotic pathways.

Patient information
Bone marrow samples were obtained from patients with MGUS (n = 18), newly diagnosed myeloma (n = 63), and myeloma relapse from complete remission (n = 30). Diagnosis of myeloma was based on standard criteria of the International Myeloma Working Group (IMWG) [53]. Complete staging work-up consisted of bone marrow examination, skeletal survey, serum and urine protein electrophoresis, and serum immunoglobulin levels.
Of the 63 patients with newly diagnosed myeloma, there were 28 females and 35 males, with a median age of 58  years. Apart from 1 patient lacking International Staging System (ISS) data [54], there were 20 stage I, 16 stage II, and 26 stage III cases. There were 8 IgA, 32 IgG, 5 IgD, 14 light chain, and 4 non-secretary myelomas. According to the criteria of the European Group for Blood and Marrow Transplantation Myeloma Registry [55], "relapse" was defined as the reappearance of the same paraprotein detected by serum/urine protein electrophoresis, appearance of new bone lesion or extramedullary plasmacytoma, or unexplained hypercalcemia after prior complete remission. The study has been approved by the Institutional Review Board of Queen Mary Hospital (UW 05-269 T/932), and written informed consent was obtained from patients for the publication of this article and any accompanying data or images.  [43]. Treatment response was defined in accordance to the IMWG criteria [56].  [57] and MMKKF (unpublished) were established from the myelomatous pleural effusion of myeloma patients. Cell lines were cultured in RPMI-1640 medium (IMDM for LP-1, DMEM + IMDM for MMLAL), supplemented with 10% or 20% fetal bovine serum, 50 U/mL of penicillin, and 50 μg/mL streptomycin, in a humidified atmosphere of 5% CO 2 at 37°C. All culture reagents were purchased from Invitrogen (Carlsbad, CA, USA). Cell lines obtained from DSMZ and ATCC have been authenticated using short tandem repeat DNA profiling analysis by them. Cells were used for biology assay within 10 passages from thawing.

Methylation-specific polymerase chain reaction (MSP)
Genomic DNA was extracted using QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany), bisulfite-converted using EpiTect Bisulfite Kit (Qiagen), and hence templates for methylated MSP (M-MSP) and unmethylated MSP (U-MSP). Primer sequences and conditions are in Table 1. Detailed procedures of MSP have been previously described [39,47].

Quantitative reverse transcription polymerase chain reaction (qRT-PCR)
Total RNA was isolated using the mirVana™ miRNA Isolation Kit (Ambion). For miR-340-5p, reverse transcription was performed using the TaqMan MicroRNA RT Kit (ABI, Waltham, MA, USA), followed by qRT-PCR using TaqMan assay (ABI), with RNU48 as the endogenous control. The expression of miR-340-5p were analyzed by the ΔCT method. Correlation between methylation and expression of miR-340-5p was calculated by Student's t test. For other genes, reverse transcription was performed using the QuantiTect Reverse Transcription Kit (QIAGEN). RNF130, MEKK1, MEKK2, MEKKK3, XIAP, MDM4, and CCND1 were quantified using TaqMan assays (ABI). MDM2, NFKB1, and NRAS were quantified using SYBR Green Master Mix (ABI). GAPDH was used as the endogenous control. Primers are listed in Table 1. The expression of these genes were calculated by the ΔCT method. Relative expression of gene in response to overexpression of miR-340-5p as compared with scramble control was analyzed by the 2 −ΔΔCT method.

Trypan blue exclusion assay
At day 5 after transfection of miR-340-5p mimics or scramble control, cell death was analyzed by trypan blue (Sigma-Aldrich, St. Louis, MO, USA). Cells in five random microscopic fields were counted for each sample under a microscope. Dead cell (stained in blue) percentage is equal to the average number of dead cells per microscopic field over the average number of total cells per microscopic field. Data represents the mean dead cell percentage derived from three independent transfections with triplicate in each.

MTS assay
The number of viable cells in proliferation was measured by CellTiter 96® AQ ueous One Solution Cell Proliferation Assay Kit (Promega, Madison, WI, USA) following the manufacturers' instructions. The reagent contains a yellowish tetrazolium compound [3-(4,5-dimethylthiazol-2-yl)-5-(3carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt; MTS], which can be bioreduced by live cells into a purple-colored formazan product that is soluble in tissue culture medium and measurable by the colorimetric method. Relative proliferation percentage of miR-340-5p overexpressed cells compared with scramble control was calculated. Data represents the mean relative proliferation percentage derived from three independent transfections with four replicates in each.

Annexin V/propidium iodide (PI) staining assay
Cell apoptosis was tested by flow cytometry using FITC Annexin V Apoptosis Detection Kit I (BD Biosciences,  19.0. The differences between WL-2 cells transfected with miR-340-5p mimics and scramble oligonucleotide control in trypan blue exclusion assay and MTS assay were studied by Student's t test. All P values were two-sided, and P < 0.05 was defined as a significant difference.

Additional files
Additional file 1: Figure S1. The funders had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Availability of data and materials All data generated or analyzed during this study are included in this published article and its supplementary information files.
Authors' contributions ZL carried out the experiments. ZL, KYW, and CSC drafted the manuscript. KYW and CSC conceived and designed the study. GAC advised on the experimental design, and WJC provided the cell lines and additional primary myeloma samples. All authors participated in the analysis and interpretation of data. All authors read and approved the final manuscript.

Ethics approval and consent to participate
The study has been approved by the Institutional Review Board of Queen Mary Hospital (UW 05-269 T/932), and written informed consent has been obtained from patients for the participation of this study.

Consent for publication
We have obtained consents from patients for the publication of this report and any accompanying data or images.