Abstract

   The aberrant expression of RNA-binding proteins (RBPs) plays important
   roles in the occurrence and progression of cancer. MBNL2 is a member of
   the RNA binding protein MBNL family that is widely expressed in
   mammalian cells. We report here that MBNL2 is downregulated in breast,
   lung and liver cancer tissues, the promoter methylation levels of MBNL2
   are higher in cancer tissues than normal tissues. The enrichment
   analysis of MBNL2 correlated genes indicates the potential function of
   MBNL2 on cancer progression. MBNL2 regulates cancer cell migration and
   invasion by modulating PI3K/AKT-mediated epithelial-mesenchymal
   transition. PI3K/AKT inhibitor overcomes the promotive effect of
   shMBNL2 on metastasis. The expression of MBNL2 is directly targeted by
   miR-182. miR-182 is upregulated in breast, lung and liver cancers and
   has good potential for cancer diagnosis. miR-182 promotes cancer cell
   migration and invasion by inhibiting the expression of MBNL2.
   Re-introduction of exogenous MBNL2 reverses the promotive effect of
   miR-182 on metastasis. Collectively, these findings suggest that MBNL2
   plays a tumor suppressive function through miR-182-MBNL2-AKT-EMT
   signaling pathways.

   Keywords: MBNL2, RNA-binding protein, miR-182, PI3K/AKT, Cancer,
   Metastasis

Introduction

   RNA-binding proteins (RBPs) plays central roles in post-transcriptional
   regulation. RBPs can interact with various RNAs (mRNAs, ncRNAs, tRNAs,
   etc.) to control all steps of RNA metabolism, such as microRNA
   processing, alternative splicing and polyadenylation, subcellular
   localization, stability, and translation of RNAs [41]^1^-[42]^3.
   Dysregulated RBPs are associated with various human diseases, including
   cancer. Several RBPs are predominantly up- or down-regulated in human
   cancers and show distinct molecular characteristics and prognostic
   effects [43]^4^,[44]^5. Aberrant expression of RBPs have been
   implicated in controlling the expression of RNAs related to cancer cell
   proliferation, apoptosis, angiogenesis, senescence, and
   epithelial-mesenchymal transition (EMT), ultimately contributing to
   oncogenesis and cancer development [45]^6^-[46]^9. Recent studies
   demonstrate that RBP gene-related signature has good predictive value
   and may have good prospects for improving clinical treatment and
   patient prognosis [47]^10^-[48]^12. Given their crucial roles in
   cancer, RBPs have emerged as promising targets of cancer therapeutics.

   Muscleblind-like (MBNL) proteins are conserved multifunctional
   RNA-binding proteins which regulate tissue specific alternative
   splicing, mRNA stability and trafficking, alternative polyadenylation,
   as well as microRNA biogenesis [49]^13^-[50]^17. The MBNL family
   consists of three closely related paralogs in mammals, MBNL1, MBNL2,
   and MBNL3. MBNL1 and MBNL2 are ubiquitously expressed, whereas MBNL3 is
   predominantly in the placenta [51]^18^,[52]^19. MBNL family plays
   important roles in the differentiation of embryonic stem cells, as well
   as in neuronal differentiation. MBNL deficient mice develop myotonia
   and skeletal myopathy, key features of human myotonic dystrophy (DM)
   [53]^20^-[54]^23. Loss of function of MBNL1/2 also causes fundamental
   defects in cortical neuron distribution and morphological changes in
   dendritic spines and postsynaptic densities [55]^24.

   Recent studies indicate that MBNL family proteins also participate in
   the occurrence and progression of cancer. MBNL1-mediated alternative
   mRNA splicing patterns contribute to the pathogenesis of MLL-rearranged
   leukemias [56]^25. MBNL1 inhibits glioma stem cell self-renewal and
   tumorigenic potential, hypoxic response within the tumor inhibits MBNL1
   activity, promoting stem-like phenotypes and tumor growth [57]^26.
   MBNL1 is downregulated in several common cancers, low MBNL1 is
   correlated with poor overall survival and increased relapse and distant
   metastasis [58]^27. MBNL1 also inhibits tumor cell metastasis through
   regulating the stability of Snail, TACC1 and DBNL [59]^28^,[60]^29. The
   role of MBNL2 on cancer is contradictory. MBNL2 is upregulated in
   metastatic renal carcinoma tissues and drives the abnormal
   cancer-associated transcripts, suggesting an oncogenic function of
   MBNL2 [61]^30. Chronic hypoxia specifically induces MBNL2, MBNL2 drives
   cancer cell proliferation and migration by regulating the expression
   and alternative splicing of hypoxia-induced genes, such as VEGFA
   [62]^31. In contrast, some studies indicate that MBNL2 possesses
   antitumor activity. MBNL2 expression level is negatively correlated
   with tumor size and stage in hepatocellular carcinoma, overexpression
   of MBNL2 inhibits liver cancer cell growth and invasion both in vitro
   and in vivo [63]^32. MBNL2 increases the anticancer ability of a
   natural compound Neobractatin [64]^33. MBNL2 also regulates tumor cell
   proliferation and DNA damage response by stabilizing p21 [65]^34.

   In this report, we further explore the molecular mechanism underlying
   the tumor suppressive function of MBNL2 in a panel of human cancers. We
   show that MBNL2 is downregulated in breast, liver and lung cancers,
   overexpression of MBNL2 inhibits cancer cell metastasis by modulating
   PI3K/AKT mediated EMT process. High levels of MBNL2 promoter
   methylation and miR-182 may contribute to decreased MBNL2 expression in
   cancer tissues. miR-182 directly targets the expression of MBNL2,
   exogenous MBNL2 can overcome the promotive effect of miR-182 on cancer
   cell metastasis. Thus, we propose that MBNL2 plays tumor suppressive
   function by modulating miR-182-MBNL2-AKT-EMT signaling pathway.

Materials and Methods

Cell culture and transfection

   MDA-MB-231, A549, MCF7, HEK 293T and HepG2 cells were obtained from
   American Type Culture Collection (Manassas, VA, USA). Cells were grown
   in Dulbecco's modified Eagle medium (Thermo Fisher Scientific, Waltham,
   MA, USA), supplemented with 10% fetal bovine serum (Biowest, France)
   and 1% Antibiotic-Antimycotic (Thermo Fisher Scientific, Waltham, MA,
   USA). To generate stable MBNL2 knockout cells, shMBNL2 lentiviral
   constructs were purchased from Genechem and infected MDA-MB-231, A549
   and HepG2 cells. To generate stable MBNL2-overexpressed cell lines,
   pVSVG (Addgene #8454), psPAX2 (Addgene #12260), PCDH-EF1and
   MBNL2-PCDH-EF1 (Unibio) plasmids were constructed, and HEK 293T cell
   used to generate high titer lentivirus. The lentivirus-infected cell
   lines and plasmid were cultured under the selection pressure of 2 μg/mL
   puromycin (Thermo Fisher Scientific, Waltham, MA, USA). Lipofectamine
   2000 (Invitrogen Life Technologies, Waltham, MA, USA) was used to
   transfect miRNA, miRNA inhibitor or siRNA.

Bioinformatics Analysis

   The expression signatures of MBNL2 were investigated in BRCA, LIHC,
   LUAD, LUSC samples from the TCGA database ([66]https://www.cancer.gov)
   and GEO database (lung cancer chip [67]GSE19804 and breast cancer chip
   [68]GSE10797) ([69]https://www.ncbi.nlm.nih.gov/geo/). The promoter
   methylation level of MBNL2 in BRCA, LIHC, LUAD, LUSC were investigated
   by UALCAN database ([70]http://ualcan.path.uab.edu/). The correlation
   between survival and MBNL2 expression was analyzed by Kaplan-Meier
   Plotter database ([71]https://kmplot.com/analysis/).

   For enrichment analysis of MBNL2 related signaling pathways, MBNL2 was
   used as the target gene to search for similar genes in GEPIA, according
   to the pearman correlation coefficient (PCC), the top 1000 genes were
   selected in BRCA, LIHC and LUAD and download the gene list, the three
   gene lists were put into Metascape for pathway enrichment analysis,
   Metascape depicts top enriched clusters and their enrichment patterns
   as a clustered heatmap.

   To investigate potential microRNAs target MBNL2, we identified 78
   potential miRNAs by combining four well-recognized miRNA databases,
   mirDIP ([72]http://ophid.utoronto.ca/mirDIP/), TargetScan
   ([73]http://www.targetscan.org/vert_72/), miRDB ([74]http://mirdb.org)
   and miRWalk ([75]http://zmf.umm.uni- heidelberg.de/apps/zmf/mirwalk2/).
   The BRCA-miRNA data were downloaded from TCGA and divided into high and
   low expression groups by R package limma, we identified 70 highly
   expression miRNAs and 82 lowly expression miRNAs. Then, 78 microRNAs
   identified from four microRNA databases were compared with the high and
   low expression miRNAs in TCGA BRCA data, the results showed that there
   were 9 up-regulated miRNAs and 5 down regulated miRNAs. Finally, the
   expression signatures of these miRNAs were investigated in BRCA, LIHC,
   LUAD, LUSC samples from the TCGA miRNA database.

Transcriptome analysis

   Transcriptome analysis was performed by Gene Denovo Biotechnology Co.
   (Guangzhou, China). Briefly, total RNAs in control or MBNL2 knockout
   cells were extracted using Trizol reagent. Eukaryotic mRNA was enriched
   by Oligo (dT) beads and reversed transcribed into cDNA with random
   primers. The cDNA fragments were purified, end-repaired, poly A added,
   and ligated to Illumina sequencing adapters. The ligation products were
   sequenced using Illumina HiSeq^TM 2500. Differently expressed genes
   (DEGs) were identified with a fold change ≥ 2 and a false discovery
   rate (FDR) < 0.05 in a comparison as significant DEGs, DEGs were then
   subjected to enrichment analysis of GO functions and KEGG pathways. The
   Q value is an adjusted p value after multi-hypothesis test correction,
   taking FDR ≤ 0.05 as a threshold. The Richfactor refers to the ratio of
   the number of the differentially expressed to the total number of
   annotated genes, the larger the Richfactor value, the greater the
   degree of enrichment.

Immunoblotting

   Samples were collected and lysed in protein lysis buffer (5 mL 1 M
   Tris-HCl (pH 8.0), 26 g urea, 1 mL Triton X-100, 2 pieces of protease
   inhibitor cocktail, extended to the total volume of 100 mL with ddH2O).
   Protein samples were separated by SDS-PAGE and then transferred to
   nitrocellulose membrane (PALL, NY, USA). The membranes were incubated
   in primary antibodies (anti-MBNL2, ab171551, 1:750; anti-MBNL2, CST
   29733S, 1:1000; anti-GAPDH, Proteintech 10442-1-AP, 1:10000; anti-ZEB1,
   ab203829, 1:500; anti-Vimentin, abD21H3, 1:500; anti-E-cadherin, CST
   24E10, 1:1000; 1:500; anti-AKT, CST C67E7, 1:1000; anti-p-AKT(ser473),
   CST D9E, 1:1000; anti- PI3Kinase P110β, CST 3011, 1:1000; anti- PTEN,
   CST 138G6, 1:1000; anti- P-GSK3β, CST D85E12, 1:1000), followed
   incubation with secondary antibodies (1:5000, 52200336, SeraCare,
   Milford, MA, USA), Membrane was finally supplemented with the Pierce
   ECL Western Blotting Substrate (Thermo Fisher Scientific) and imaged
   with an autoradiography film (FUJIFILM, Minato, Japan).

Migration and invasion assays

   Cells were suspended in serum-free medium and seeded into the upper
   chamber of transwell insert membranes of an 8 μm pore size (Corning) or
   coated with Matrigel (Corning) in a 24-well plate (3-10×10^4 cells per
   well). Approximately 10% fetal bovine serum (FBS) was used in the
   bottom chamber as the chemoattractant. The cells in the upper chamber
   were removed 24-72 h later using a cotton swab and washed one to two
   times with PBS and fixed by methanol and then stained with 0.5% crystal
   violet. The number of successfully translocated cells was counted under
   microscope.

RNA isolation and quantitative RT-PCR

   Total RNA was extracted using the RNAiso Plus Reagent (TaKaRa, Mountain
   View, CA, USA), from culture cells according to the manufacturer's
   instruction. The cDNA was synthesized from total RNA using ReverTra Ace
   qPCR RT Master Mix (TOYOBO, Osaka, Japan) or TaqMan™ MicroRNA Reverse
   Transcription Kit (Life Technologie). qPCR was performed with SYBR
   Green Realtime PCR Master Mix (TOYOBO, Osaka, Japan) and primer pairs
   on a qTOWER 2.2 (Jena, Germany). The level of mature miR-182 was
   normalized relative to U6 endogenous control and MBNL2 expression was
   normalized relative to GAPDH (endogenous control) by the 2^-∆∆CT
   method. The primer of miR-182 and U6 purchased from Thermo Fisher.
   Primer sequences are listed as follows: MBNL2, forward primer
   5′-TCAAAGAGGAACATGCTCACG-3′, reverse primer
   5′-AACGGCCCTTTAGGGAATCAA-3′; GAPDH, forward primer
   5′-AAGGCTGTGGGCAAGG-3′, reverse primer 5′-TGGAGG AGTGGGTGTCG-3′.

Luciferase reporter assay

   The pGL3 luciferase reporters containing wild-type MBNL2 (WT) or
   mutant-type (MUT) miR-182 binding sites were constructed by YouBio
   (Changsha, China). For luciferase assay, miR-182 mimic plus WT or MUT
   3'UTR of MBNL2 were transfected into MDA-MB-231 cells by lipofectamine.
   After 48-72 hours, the cells were collected and firefly and renilla
   luciferase activities were measured with Promega Dual-Luciferase®
   Reporter Assay System (Promega Corporation, Madison, WI, USA).

Immunohistochemistry

   The breast cancer, liver cancer and lung adenocarcinoma tissue
   microarray purchased from SHANGHAI OUTDO BIOTECH CO.LTD, the array
   numbers are HBre-Duc060CS-04, HLivH060CS01 and HLugA060PG02
   respectively. The primary antibodies for IHC staining are mouse
   monoclonal Ab against human MBNL2 (sc-136167, Santa Cruz Biotechnology,
   CA, USA). The paraffin-embedded slide was deparaffinized, rehydrated,
   and block in sheep serum for 30 min, followed by incubation with
   anti-MBNL2 overnight at 4 °C. The slide was mounted with D.P.X. for
   histology analysis. A semi-quantitative method based on the staining
   intensity and the area of positive cells was used. The staining scores
   were calculated by multiplying the above two parameters. The MBNL2
   staining results were assessed according to the staining scores.

Statistical analysis

   Student's t test, the Mann-Whitney U test were used to analyze
   differences among different groups. All the results were presented as
   the mean ± SD and P values < 0.05 were considered statistically
   significant.

Results

MBNL2 is downregulated in some human cancers

   To study the function of MBNL2 in human cancers, we analyzed TCGA
   RNA-seq data sets. A significant downregulation of MBNL2 was observed
   in several cancer types, including breast cancer, liver cancer, lung
   adenocarcinoma, and lung squamous cell carcinoma (Figure [76]1A). The
   microarray data from GEO also showed downregulation of MBNL2 in breast
   and lung cancer tissues (Figure [77]1B). We detected MBNL2 protein
   expression by immunohistochemistry in commercial tissue microarrays
   which containing cancer tissues of patients with primary operable
   breast cancer, lung cancer, or liver cancer and adjacent normal tissues
   (Table [78]1). MBNL2 protein was absent or weakly detected in cancer
   tissues, whereas adjacent normal tissues expressed higher levels of
   MBNL2 (Figure [79]1C).

Figure 1.

   [80]Figure 1
   [81]Open in a new tab

   Analysis the expression of MBNL2 in breast, lung and liver cancers. (A)
   Scatter diagram derived from gene expression data in TCGA comparing the
   expression of MBNL2 in tumor tissue and normal tissue in BRCA, LIHC,
   LUAD and LUSC, ****p<0.0001. (B) Scatter diagram derived from gene
   expression data in GEO datasets ([82]GSE10797 and [83]GSE19804)
   comparing the expression of MBNL2 in tumor tissue and normal tissue in
   breast and lung cancers, *p<0.05, **p<0.01. (C) Representative images
   of IHC staining in human breast, lung and liver cancer tissue with
   MBNL2 antibody. (D) Promoter DNA methylation level of MBNL2 in BRCA,
   LUAD, LUSC and LIHC, data from UALCAN, *p<0.05, ** p<0.01,
   ****p<0.0001.

Table 1.

   Basic clinicopathological factors of patients included

   [84]graphic file with name jcav12p6715g001.jpg
   [85]Open in a new tab

   Aberrant promoter methylation is considered a hallmark of cancer
   involved in silencing of tumor suppressor genes and activation of
   oncogenes. We analyzed MBNL2 promoter methylation levels in different
   cancer types. Methylation indices for MBNL2 in BRCA, LIHC, LUSC and
   LUAD patients were found to be significantly higher as compared to that
   in normal individuals (Figure [86]1D). These findings indicate
   increased methylation of MBNL2 promoter may contribute to reduced gene
   expression in cancer.

MBNL2 regulates cancer cell migration and invasion

   To explore the biological function of MBNL2 in cancer, we identified
   the top 1000 MBNL2 correlated genes from TCGA breast, lung and liver
   cancer data sets and the gene lists were submitted to Metascape to
   perform gene ontology enrichment analysis. According to GO term
   analysis, genes correlated with MBNL2 were significantly enriched for
   biological processes associated with membrane trafficking, TGF-beta
   signaling, VEGFR and EGF/EGFR signaling, which are all related to
   cancer cell invasion and metastasis (Figure [87]2A, Supp Figure [88]1).
   We overexpressed MBNL2 in cancer cells and performed transcriptome and
   gene expression analysis. GO and KEGG pathway enrichment analysis
   revealed that ECM-receptor interaction, focal adhesion and PI3K-AKT
   signaling are among the top five enriched pathways, these pathways are
   closely associated with cancer metastasis (Figure [89]2B). Therefore,
   we hypothesized that MBNL2 might function as a tumor suppressor in
   breast, liver and lung cancers. To this end, we knocked down MBNL2
   expression by using shRNA, or overexpressed His-MBNL2 in MDA-MB-231
   breast cancer cells, A549 non-small-cell lung cancer cells, and HepG2
   liver cancer cells, the knock-down or knock-in efficiency were verified
   by qRT-PCR and western blot (Figure [90]2C). Transwell migration assay
   and matrigel invasion assay revealed that overexpression of MBNL2
   dramatically reduced the number of migrated and invaded cells, whereas
   MBNL2 shRNA showed the opposite effects (Figure [91]2D and E, supp
   Figure [92]2A and B).

Figure 2.

   [93]Figure 2
   [94]Open in a new tab

   MBNL2 regulates cancer cell migration and invasion. (A) Metascape
   depicts top enriched clusters and their enrichment patterns across
   multiple gene lists (MBNL2 correlated genes in BRCA, LUAD and LIHC,
   data from GEPIA) as a clustered heatmap. (B) KEGG pathway enrichment
   analysis of RNA-seq data from cells transfected with empty vector or
   His-MBNL2. (C) MBNL2 mRNA and protein expression in MDA-MB-231 and A549
   cell lines measured by qRT-PCR and Western blotting. (D) Migration
   assays and (E) Invasion assays showing that overexpression of MBNL2
   inhibited the ability of migration and invasion in MDA-MB-231 and A549
   cells, knockdown of MBNL2 increased the number of migrated and invaded
   cells. ***p<0.001.

MBNL2 regulates EMT process through PI3K/AKT signaling

   Epithelial-mesenchymal transition (EMT), an evolutionarily conserved
   developmental process, is well recognized as a key step in cancer
   invasion and metastasis. Therefore, we investigated whether MBNL2
   regulates metastasis through EMT in MDA-MB-231, A549 and HepG2 cells.
   Western blotting revealed that knockdown of MBNL2 downregulated the
   expression of epithelial marker E-cadherin and upregulated the
   expression of mesenchymal markers Vimentin and ZEB1, however,
   overexpression of MBNL2 showed the opposite effects (Figure [95]3A,
   supp Figure [96]2C).

Figure 3.

   [97]Figure 3
   [98]Open in a new tab

   MBNL2 regulates EMT through PI3K/AKT signaling. (A) Western blot
   analysis of EMT related markers (E-cadherin, Vimentin and ZEB1) in
   MDA-MB-231 and A549 cells overexpressing or silencing MBNL2. (B)
   Oncomine analysis showing the expression of MBNL2 was significantly
   associated with AKT1 and AKT2. (C) Western blot analysis of PI3K/AKT
   related marker (AKT, p-AKT, PI3K, PTEN and p-GSK3β) in MDA-MB-231 and
   A549 cells overexpressing or silencing MBNL2. (D) AKT inhibitor
   LY294002 inhibits shMBNL2-induced cell invasion in MDA-MB-231 cells.

   To further study the molecular mechanism of MBNL2 on EMT, we performed
   gene coexpression analysis using the Oncomine breast cancer databases.
   The results showed that both AKT1 and AKT2 are among the top 20 genes
   that are highly coexpressed with MBNL2 (Figure [99]3B). Because the
   role of PI3K/AKT pathway in EMT is well recognized, thus we aimed to
   study whether MBNL2 regulates EMT through PI3K/AKT. Immunoblotting
   revealed that knockdown of MBNL2 increased the level of phosphorylated
   AKT S473 without changing the protein expression of PI3K, AKT and PTEN,
   whereas exogenous MBNL2 inhibited the phosphorylation of AKT S473 in
   both MDA-MB-231 and A549 cells (Figure [100]3C). Treatment with PI3K
   inhibitor LY294002 in MBNL2 knockdown cells reduced the number of
   invaded cells (Figure [101]3D). Collectively, these data demonstrated
   that silencing MBNL2 promoted EMT process by activating PI3K/AKT
   signaling.

MiR-182 is upregulated in some human cancers

   We were keen to understand the upstream signals responsible for the
   decreased levels of MBNL2 in human cancers. MicroRNAs generally bind to
   the 3'UTR of their target mRNAs and suppress protein production through
   post-transcriptional mechanisms, we therefore explored the miRNAs
   targeting MBNL2 using four miRNA target prediction databases, miRDB,
   mirDIP, miRWalk, and TargetScan. A total of 2010 miRNAs were predicted,
   among which 77 miRNAs were common to all four algorithms (Figure
   [102]4A). Because MBNL2 is downregulated in breast cancer tissues, we
   further identified MBNL2-targetting miRNAs which were upregulated in
   breast cancer. Analysis of TCGA breast cancer miRNA-seq data revealed
   82 upregulated and 70 downregulated miRNAs in breast cancer. The number
   of MBNL2-targeting miRNAs that were up or down regulated in breast
   cancer is represented in Venn diagrams. Finally, we obtained 9
   MBNL2-targeting miRNAs that were upregulated in breast cancer,
   including miR-182, miR-429, miR-97, miR-33a, miR-183, miR-200c,
   miR-493, miR-200b and miR-32 (Figure [103]4B). We analyzed the
   expression levels of these upregulated miRNAs in other human cancer
   types, the results showed that only miR-182 were upregulated in breast,
   liver and lung cancers (Figure [104]4C). In receiver operating
   characteristic (ROC) curve analysis, the area under the curve (AUC) is
   the most widely used index of diagnostic accuracy for assessing the
   sensitivity and specificity of a biomarker. We therefore performed ROC
   curve analysis to evaluate the application of miR-182 in clinical
   diagnosis. The results indicated that miR-182 had a good performance in
   the diagnosis of breast, lung and liver cancers (Figure [105]4D).

Figure 4.

   [106]Figure 4
   [107]Open in a new tab

   Identification of miRNAs targeting MBNL2 in cancer. (A) 78 miRNAs were
   predicted to target MBNL2 by four miRNA databases (Targetscan, miRDB,
   miRWalk and mirDIP). (B) Overlapping of 77 MBNL2-targeting miRNAs and
   TCGA-BRCA differentially expressed miRNAs, 9 MBNL2-targeting miRNAs
   were up-regulated and 5 MBNL2-targeting miRNAs were downregulated in
   breast cancer. (C) Scatter diagram derived from gene expression data in
   TCGA comparing the expression of miR-182 in tumor tissues and normal
   tissues in BRCA, LIHC, LUAD and LUSC, **** p<0.0001. (D) ROC curve
   analysis. The AUC was analyzed to evaluate the potential of miR-182 in
   cancer diagnosis. The x-axis indicated specificity (%) and y-axis
   indicated sensitivity (%).

MiR-182 targets the expression of MBNL2

   To validate the effect of miR-182 on MBNL2 expression, we performed
   quantitative RT-PCR analysis. Overexpression of miR-182 reduced the
   transcript abundance of MBNL2 in both MDA-MB-231 and MCF-7 cells,
   knockdown of endogenous miR-182 with inhibitor (Int-miR-182) increased
   the amounts of MBNL2 mRNA in both cells (Figure [108]5A, C).
   Immunoblotting results showed that miR-182 suppressed protein
   expression of MBNL2, whereas miR-182 inhibitor moderately increased
   MBNL2 expression in both MDA-MB-231 and MCF-7 cells (Figure [109]5B,
   D). Then, we performed luciferase assay to determine the regulatory
   effect of miR-182 on MBNL2 gene expression. miR-182 dramatically
   reduced the luciferase activity in MCF-7 cells transfected with
   luciferase reporter containing wild-type, but not mutated 3'UTR of
   MBNL2 (Figure [110]5E).

Figure 5.

   [111]Figure 5
   [112]Open in a new tab

   miR-182 directly targets the expression of MBNL2. (A) Relative mRNA
   levels of miR-182 and MBNL2 in MDA-MB-231 and (C) MCF-7 cells
   overexpressing NC, miR-182 mimics, NC inhibitor (Int-NC) and miR-182
   inhibitor (Int-miR-182). (B) The protein levels of MBNL2 were analyzed
   in in MDA-MB-231 and (D) MCF-7 cells transfected with indicated miRNAs
   and inhibitors. (E) Schematic illustration of the predicted miR-182
   binding site in MBNL2 3'UTR. Luciferase reporters containing the wile
   type (WT) or the mutated (Mut) MBNL2 3'UTR were co-transfected with
   either NC or miR-182 into MDA-MB-231 cells, dual luciferase assay was
   performed and normalized to NC. **p<0.01.

MiR-182 controls tumor cell metastasis through MBNL2

   We performed transwell migration assay and matrigel invasion assay to
   determine the effects of miR-182 on cancer cell metastasis. The results
   showed that overexpression of miR-182 mimics led to increased number of
   migrated and invaded cells, inhibition of endogenous miR-182 with miRNA
   inhibitor markedly attenuated migration and invasion in breast cancer
   cells (Figure [113]6A and B). Immunoblotting revealed that miRNA-182
   mimics downregulated the expression of epithelial marker E-cadherin and
   upregulated the expression of mesenchymal markers Vimentin, whereas
   miR-182 inhibitor showed the opposite effects. Moreover, the level of
   phosphorylated AKT S473 was increased in cells overexpressing miR-182
   mimics, indicating that miR-182 may regulate EMT process through AKT
   signaling. We also observed whether re-introduction of exogenous MBNL2
   could provide resistance to miR-182 mediated cell metastasis. Invasion
   assay demonstrated that overexpression of MBNL2 was sufficient to
   rescue the promotive effect of miR-182 on cancer cell metastasis
   (Figure [114]6D).

Figure 6.

   [115]Figure 6
   [116]Open in a new tab

   miR-182 regulates cancer cell migration and invasion through MBNL2. (A)
   Cell migration and (B) invasion assay showing the effect of miR-182 or
   miR-182 inhibitor on cancer metastasis. (C) Western blot analysis of
   EMT related markers (E-cadherin and Vimentin), PI3K, pAKT S473, and
   PTEN in MCF-7 cells transfected with either miR-182 mimics or
   inhibitor. (D) Invasion assays showing that overexpression of MBNL2
   could rescue the promotive effect of miR-182 on cancer cell invasion.

Discussion

   Except the well-studied function in DM, recent reports suggested that
   MBNL2 also participates in tumorigenesis and tumor progression.
   However, the role of MBNL2 on cancer is controversial. Previously, we
   report that MBNL2 post-transcriptionally regulates p21 mRNA stability
   independently of p53, depletion of MBNL2 promotes colon cell
   proliferation [117]^33. In this study, we further explored the
   molecular mechanism underlying the tumor suppressive function of MBNL2
   in a panel of human cancer cells. We observed that MBNL2 is down
   regulated in breast, lung, and liver cancer. MBNL2 modulates multiple
   pathways involving tumorigenesis and tumor progression. Depletion of
   MBNL2 promotes whereas exogenous MBNL2 inhibits migration and invasion
   in breast, lung and liver cancer cells. We identified that MBNL2
   modulates the activation of AKT without changing the protein expression
   of PI3K, AKT and PTEN. Treatment of PI3K/AKT inhibitor could overcome
   the promotive effect of shMBNL2 on cancer cell invasion. Together,
   these results further confirm the tumor suppressive function of MBNL2
   in some types of human cancer.

   MiRNAs control gene expression through a post-transcriptional
   mechanism. Human miRNAs are frequently located at fragile sites and
   chromosomal regions affected in cancer [118]^35^-[119]^36. miR-182 is
   one of the most frequently studied cancer-related miRNAs that is
   dysregulated in various human cancers, miR-182 may function as an
   oncogenic or tumor suppressor miRNA in different cancer types. Some
   studies report that miR-182 promotes cancer cell proliferation and
   invasion in hepatocellular carcinoma, prostate cancer, breast cancer
   and glioma [120]^37^-[121]^40. However, some studies show that miR-182
   may function as tumor suppressor, miR-182 is frequently downregulated
   in gastric cancer and metastatic NSCLC, overexpression of miR-182
   inhibits cell growth, EMT and metastasis [122]^41^-[123]^42. miR-182
   sensitizes human ovarian cancer cells to cisplatin through direct
   targeting CDK6 [124]^43. miR-182 also inhibits colon cancer
   tumorigenesis, angiogenesis and lymphangiogenesis by targeting VEGF-C
   [125]^44. These conflicting results suggest the role of miR-182 is
   complex in cancer. To identify the upstream signals responsible for
   downregulation of MBNL2 in cancer, we explored the MBNL2-targeting
   miRNAs upregulated in breast cancer tissues using TCGA miRNA-seq data.
   We show miR-182 is upregulated in breast, liver and lung cancer
   tissues. ROC curve analysis indicates miR-182 may serve as a valuable
   biomarker for diagnosis. miR-182 plays oncogenic function in breast
   cancer, lung and liver cancers. Overexpression of miR-182 promotes
   invasion through MBNL2, exogenous MBNL2 can rescue the inhibitory
   effect of miR-182 on metastasis. Since miR-182 directly targets the
   expression of MBNL2, the upregulation of miR-182 may contribute to the
   downregulation of MBNL2 in these cancer types.

   The aberrant expression of RBPs have been implicated in various types
   of human cancers. DNA methylation has been widely described as the main
   epigenetic mechanism that plays a significant role in gene expression,
   X-chromosome inactivation, genetic imprinting, differentiation and
   development [126]^45^-[127]^48. Hypermethylation occurring in the
   context of the CpG at a promoter region prevents the binding of
   transcription factors, thereby inhibiting gene transcription [128]^49.
   We show here the methylation status of MBNL2 promoter is significantly
   higher in tumor tissues comparing to normal tissues. Hypermethylation
   of MBNL2 is associated with downregulation of MBNL2 in tumor tissues.
   When taken together, we hypothesized that both DNA methylation and
   miR-182 may contribute to MBNL2 downregulation in cancer. MBNL2 plays a
   tumor suppressive function through miR-182-MNL2-AKT-EMT signaling
   pathway.

Supplementary Material

   Supplementary figures.
   [129]Click here for additional data file.^ (613.8KB, pdf)

Acknowledgments