Impact of Oncogenic Targets by Tumor-Suppressive miR-139-5p and miR-139-3p Regulation in Head and Neck Squamous Cell Carcinoma

We newly generated an RNA-sequencing-based microRNA (miRNA) expression signature of head and neck squamous cell carcinoma (HNSCC). Analysis of the signature revealed that both strands of some miRNAs, including miR-139-5p (the guide strand) and miR-139-3p (the passenger strand) of miR-139, were downregulated in HNSCC tissues. Analysis of The Cancer Genome Atlas confirmed the low expression levels of miR-139 in HNSCC. Ectopic expression of these miRNAs attenuated the characteristics of cancer cell aggressiveness (e.g., cell proliferation, migration, and invasion). Our in silico analyses revealed a total of 28 putative targets regulated by pre-miR-139 (miR-139-5p and miR-139-3p) in HNSCC cells. Of these, the GNA12 (guanine nucleotide-binding protein subunit alpha-12) and OLR1 (oxidized low-density lipoprotein receptor 1) expression levels were identified as independent factors that predicted patient survival according to multivariate Cox regression analyses (p = 0.0018 and p = 0.0104, respectively). Direct regulation of GNA12 and OLR1 by miR-139-3p in HNSCC cells was confirmed through luciferase reporter assays. Moreover, overexpression of GNA12 and OLR1 was detected in clinical specimens of HNSCC through immunostaining. The involvement of miR-139-3p (the passenger strand) in the oncogenesis of HNSCC is a new concept in cancer biology. Our miRNA-based strategy will increase knowledge on the molecular pathogenesis of HNSCC.


Introduction
Head and neck squamous cell carcinoma (HNSCC) is a malignant neoplasm that originates in the epithelium of the oral cavity, larynx, and pharynx, and it accounts for approximately 90% of head and neck cancers [1,2]. HNSCC is the sixth most common cancer worldwide, and over 10% of the patients are at an advanced stage at the time of initial diagnosis [3]. In addition, approximately half of the patients will experience recurrence after treatment [4]. Effective treatments for patients with metastasis or recurrence are limited, and the prognosis of patients with treatment failure is extremely poor [5]. The elucidation of the molecular pathogenesis of the malignant transformation of HNSCC is indispensable for the development of new therapeutic strategies. As a result of the human genome project, a vast number of non-protein-coding RNA molecules (ncRNAs) have been transcribed and have been found to have actual functions in human cells [6,7]. MicroRNAs (miRNAs), a type of ncRNA, are single-stranded RNA molecules consisting of only 19−23 bases [8,9]. A specific characteristic of miRNAs is that a single miRNA can control a large number of RNA transcripts (protein-coding RNAs and ncRNAs) in a sequence-dependent manner [9]. Over the last decade, numerous studies have reported aberrantly expressed (upregulated or downregulated in cancer tissues) miRNAs in a wide range of cancers [10]. Aberrantly expressed miRNAs trigger the disruption of intracellular RNA networks and promote cancer cell initiation, metastasis, and drug resistance [11]. Therefore, up-or downregulated miRNAs in cancer cells are closely involved in human oncogenesis [7].
It is important to determine target miRNAs with altered expression in cancer tissues. Based on this concept, we determined miRNA expression signatures by using RNA sequencing in several cancers [12,13]. Our recent studies showed that some passenger strands (e.g., miR-145-3p, miR-99a-3p, and miR-199-3p) derived from miRNA duplexes act as tumor-suppressive miRNAs by targeting several oncogenes [14][15][16]. According to the previous theory of miRNA biogenesis, miRNA guide strands are incorporated into the RISC to regulate their target genes. In contrast, the passenger strands of miRNAs are degraded in the cytoplasm and have no function [17]. Recent large-scale bioinformatics analyses showed that both strands of several miRNAs (e.g., miR-30a, miR-143, and miR-145) function jointly as tumor-suppressive miRNAs in cancer cells [18][19][20]. The involvement of miRNA passenger strands in human oncogenesis is a new concept in cancer biology.
In this study, we newly generate a miRNA expression signature in clinical specimens of HNSCC by using RNA sequencing. Analysis of the signature is used to reveal the downregulation of miR-139-5p (the guide strand) and miR-139-3p (the passenger strand) in cancer tissues, and their downregulation is confirmed by an analysis of The Cancer Genome Atlas (TCGA). We demonstrate the tumor-suppressive functions of these two miRNAs and identify their target oncogenes in HNSCC. This is the first report of the involvement of the passenger strand miR-139-3p and its target genes in HNSCC. Our miRNA-based strategy will increase knowledge on the molecular pathogenesis of HNSCC.

Generation of a miRNA Expression Signature in HNSCC
A total of six cDNA libraries were prepared from clinical specimens of HNSCC (three pairs of normal and cancer tissues) and were analyzed by using RNA sequencing. The clinical features of the HNSCC patients are summarized in Table S1. We performed RNA sequencing after a trimming procedure, and 955,314-1,927,436 reads were successfully mapped to miRNA loci within the human genome (Table S2). A total of 112 miRNAs were significantly downregulated (log 2 fold change < −2.0 and p-value < 0.05) in the HNSCC tissues ( Figure 1A and Table S3). Notably, our signature showed that both strands of nine pre-miRNAs (miR-1, miR-30a, miR-483, miR-99a, miR-139, miR-190a, miR-199b, miR-204, and miR-378a) were downregulated in the HNSCC tissues (Table S3). We focused on miR-139 out of these nine pre-miRNAs. To confirm the accuracy of our miRNA signature, we evaluated the expression levels of miR-139-5p and miR-139-3p in the HNSCC tissues by using the TCGA-HNSC datasets. The expression levels of miR-139-5p (p < 0.001) and miR-139-3p (p < 0.001) were significantly lower in the HNSCC tissues ( Figure 1B).

Effects of Ectopic Expression of miR-139-5p and miR-139-3p on HNSCC Cell Proliferation, Migration, and Invasion
We examined the expression levels of miR-139-5p and miR-139-3p and performed functional assays of these miRNAs in two HNSCC cell lines, Sa3 and HSC-3. In these cell lines, the expression levels of miR-139-5p and miR-139-3p were lower than those in normal epithelial tissues ( Figure 2). Normal oral tissues were derived from HNSCC patients, and the details are summarized in Table S1.

Effects of Ectopic Expression of miR-139-5p and miR-139-3p on HNSCC Cell Proliferation, Migration, and Invasion
We examined the expression levels of miR-139-5p and miR-139-3p and performed functional assays of these miRNAs in two HNSCC cell lines, Sa3 and HSC-3. In these cell lines, the expression levels of miR-139-5p and miR-139-3p were lower than those in normal epithelial tissues ( Figure 2). Normal oral tissues were derived from HNSCC patients, and the details are summarized in Table S1.

Identification of Putative Oncogenic Targets through miR-139-5p and miR-139-3p Regulation in HNSCC Cells
To identify the putative oncogenic targets of miR-139-5p or miR-139-3p in HNSCC cells, we used the TargetScanHuman database (release 7.2) and gene expression data (GSE172120: genes upregulated in clinical specimens of HNSCC). Our strategy for identifying the miR-139-5p and miR-139-3p target genes is shown in Figure 4.
The analysis of the TargetScan database showed that 3486 and 3145 genes had putative miR-139-5p and miR-139-3p binding sites, respectively, within their sequences ( Figure 4). As the tumor suppressor miRNA targets were assumed to be upregulated in cancer tissue, we also performed a genome-wide gene expression analysis in the same clinical specimens used to generate the miRNA signature. The results revealed that 97 genes were significantly upregulated (log 2 fold change > 1.0 and p-value < 0.05) in HNSCC tissues (Table S4).
The TargetScan and gene expression datasets were merged to determine the oncogenic targets of miR-139-5p or miR-139-3p in HNSCC cells. A total of 28 genes were upregulated in HNSCC tissues, and these genes had a putative miR-139-5p or miR-139-3p binding site in their 3 UTR (Table 1).

Clinical Significance of the Putative Target Genes of miR-139-5p and miR-139-3p Determined by the TCGA Analysis
To determine clinical usefulness, a clinicopathological analysis of these putative target genes was performed with the TCGA-HNSC data. Five of the genes (PLEC, YWHAQ, GNA12, OLR1, and ACOT9) were upregulated in HNSCC ( Figure 5A), and a significantly worse prognosis was predicted in patients with HNSCC ( Figure 5B). Of those genes, GNA12 and OLR1 were identified as independent prognostic factors through a multivariate Cox regression analysis ( Figure 5C).

Direct Regulation of GNA12 and OLR1 by miR-139-3p in HNSCC Cells
To investigate whether the expression of GNA12 and OLR1 is regulated directly by pre-miR-139 in HNSCC cells, we performed real-time PCR, Western blotting, and dual luciferase reporter assays. In HSC-3 cells transfected with miR-139-3p, both the mRNA and protein levels of GNA12 were markedly reduced ( Figure 6A,B).

Regulation in HNSCC Cells
To identify the putative oncogenic targets of miR-139-5p or miR-139-3p in HNSC cells, we used the TargetScanHuman database (release 7.2) and gene expression da (GSE172120: genes upregulated in clinical specimens of HNSCC). Our strategy for iden fying the miR-139-5p and miR-139-3p target genes is shown in Figure 4.  The dual luciferase reporter assays showed that miR-139-3p binds directly to GNA12. Luciferase activity was significantly reduced after transfection of a vector harboring the wild-type GNA12 3 -UTR sequence (containing the miR-139-3p binding site). In contrast, luciferase activity was not reduced after transfection of a vector harboring a GNA12-deleted sequence (lacking the miR-139-3p binding site) ( Figure 6C). These data suggest that miR-139-3p binds directly to the 3 -UTR of GNA12 to control its expression.
The TargetScan database shows that the binding sites of both miR-139-5p and miR-139-3p are located within the 3 UTR of the OLR1 gene. We investigated the involvement of these miRNAs in the expression control of OLR1 in HNSCC cells. As a result, miR-139-5p was not involved in OLR1 expression ( Figure S3). Both the mRNA and protein levels of OLR1 were significantly downregulated after miR-139-3p transfection in HSC-3 cells ( Figure 6D,E). Luciferase activity was significantly reduced after transfection with a vector harboring the wild-type sequence-but not with that harboring a deleted sequence-of the OLR1 3 -UTR ( Figure 6F). These data suggest that miR-139-3p directly controls OLR1 expression by binding to its 3 -UTR in HNSCC cells.

Overexpression of GNA12 and OLR1 in Clinical Specimens of HNSCC
The expression of the GNA12 and OLR1 proteins was evaluated by immunostaining in clinical specimens of HNSCC. High expression of GNA12 was detected in HNSCC lesions ( Figure 7A-C). In contrast, normal epithelial tissues showed weak immunostaining ( Figure 7D). High expression of OLR1 was detected in HNSCC lesions, in contrast to the weak staining in the normal epithelium ( Figure 7E-H). The clinical features of three of the HNSCC cases used for immunohistochemical staining are summarized in Table S5.

GNA12-and OLR1-Mediated Pathways in HNSCC Cells
Finally, we investigated the genes that were differentially expressed between the high and low GNA12 or OLR1 expression groups in an HNSCC cohort from TCGA using a gene set enrichment analysis. As a result, 19 and 14 gene sets were significantly enriched (FDR q-value < 0.05) in the high GNA12 and OLR1 expression groups, respectively ( Table 2). The most enriched gene set in both the GNA12 and OLR1 expression groups was that of the "epithelial-mesenchymal transition" (Table 2 and Figure S4). These results indicate that GNA12 and OLR1 play roles as oncogenes in HNSCC via the epithelial-mesenchymal transition pathway. We validated the expression levels of these 28 target genes by using RNA sequencin data obtained from clinical specimens of HNSCC in TCGA . The expression of 24 gen  (BCL11B, DOCK5, PLEC, SRPK2, TNFSF10, ELAVL2, YWHAQ, GNA12, HIST2H2B  SLC43A2, TMEM184B, STRIP2, PDK1, OLR1, NAV1, MKI67, SLC25A37, OSBPL3, HES  ABCA1, ACOT9, NKD1, APP, and SLC16A3) was upregulated in the HNSCC tissues (n 518) compared with normal tissues (n = 44) ( Figure S2A,B).

Clinical Significance of the Putative Target Genes of miR-139-5p and miR-139-3p Determined by the TCGA Analysis
To determine clinical usefulness, a clinicopathological analysis of these putative ta get genes was performed with the TCGA-HNSC data. Five of the genes (PLEC, YWHAQ GNA12, OLR1, and ACOT9) were upregulated in HNSCC (Figure 5A), and a significant worse prognosis was predicted in patients with HNSCC ( Figure 5B). Of those gene GNA12 and OLR1 were identified as independent prognostic factors through a multiva iate Cox regression analysis ( Figure 5C).   Low expression of all five genes was significantly predictive of a worse prognosis in patients with HNSCC. The patients were divided into two groups-the high-and low-expression groups-according to the median miRNA expression level. The red and blue lines represent the high-and low-expression groups, respectively. (C) Forest plot presenting the results of a multivariate Cox regression analysis of the prognostic value of two putative target genes (GNA12 and OLR1) identified in an HNSCC dataset from TCGA. The expression levels of GNA12 and OLR1 were determined to be independent prognostic factors in terms of the 5-year overall survival rate after adjustments for tumor stage, age, and pathological grade (p < 0.05).

Direct Regulation of GNA12 and OLR1 by miR-139-3p in HNSCC Cells
To investigate whether the expression of GNA12 and OLR1 is regulated directly by pre-miR-139 in HNSCC cells, we performed real-time PCR, Western blotting, and dual luciferase reporter assays. In HSC-3 cells transfected with miR-139-3p, both the mRNA and protein levels of GNA12 were markedly reduced ( Figure 6A,B).

Overexpression of GNA12 and OLR1 in Clinical Specimens of HNSCC
The expression of the GNA12 and OLR1 proteins was evaluated by immunostaining in clinical specimens of HNSCC. High expression of GNA12 was detected in HNSCC lesions ( Figure 7A-C). In contrast, normal epithelial tissues showed weak immunostaining ( Figure 7D). High expression of OLR1 was detected in HNSCC lesions, in contrast to the weak staining in the normal epithelium ( Figure 7E-H). The clinical features of three of the HNSCC cases used for immunohistochemical staining are summarized in Table S5.

Discussion
Due to the high rate of recurrence and metastasis after the initial treatment, the prognosis of patients with HNSCC remains poor [4,21]. Recently developed molecule-targeted drugs and immune checkpoint inhibitors have not significantly improved the prognosis of patients [5,22,23]. Therefore, novel treatment options for patients with advanced HN-SCC are urgently needed. The elucidation of the molecular mechanisms underlying the oncogenesis of HNSCC is indispensable for finding therapeutic target molecules.
We have continued to use miRNA-based approaches in order to identify potential prognostic markers and therapeutic targets for HNSCC [24]. Our previous studies showed that certain integrins are controlled directly by tumor-suppressive miRNAs (e.g., miR-29-family, miR-150, and miR-199a) and that the activation of signaling mediated by ITGA3/ITGB1 and ITGA6 enhances the aggressiveness of HNSCC [12,16,25]. In addition, the expression of ITGA3 predicted poorer survival of patients with HNSCC [16].
In this study, we determined the miRNA expression signature of HNSCC by using clinical specimens. The analysis of the signature showed that a total of 112 miRNAs were significantly downregulated (log 2 fold change < −2.0 and p-value < 0.05) in cancer tissues. Interestingly, among these downregulated miRNAs, 31 were annotated as passenger strands of miRNAs in the miRBase database. Functional analysis of these miRNAs and their regulated molecular networks will help elucidate the molecular pathogenesis of HNSCC.
Our previous studies revealed that miR-139-3p (the passenger strand of pre-miR-139) acts as a tumor-suppressive miRNA in bladder cancer [33] and renal cell carcinoma [34]. A single miRNA can control a vast number of RNAs in cells, and the genes controlled by a miRNA vary from cell to cell. In this study, we revealed that both strands of pre-miR-139 act as tumor-suppressive miRNAs in HNSCC cells. Previously, a vast number of studies reported that miR-139-5p plays a tumor-suppressive role in multiple cancers [35]. In oral squamous cell carcinoma (OSCC), miR-139-5p inhibited tumorigenesis by targeting HOXA9 and CXCR4 [36,37]. Recent studies showed that long non-coding RNAs (lncRNAs) are closely associated with human oncogenesis [11]. Several studies showed that LINC00152 acts as a cancer-promoting and metastatic lncRNA in multiple cancers [38]. A recent study indicated that the lncRNA LINC00152 adsorbs miR-139-5p and behaves as an oncogene in OSCC [39].
On the other hand, because it is a passenger strand, few studies have evaluated the functional significance of miR-139-3p in HNSCC cells. A previous study showed that the expression of miR-139-3p was suppressed by human papillomavirus 16, which, in turn, promoted oncogenesis in head, neck, and cervical cancers [40]. Another study showed that miR-139-3p is significantly downregulated in laryngeal squamous cell carcinoma and that its expression regulates several cancer-related genes, (e.g., RAB5A, ITGB1, FAK, PXN, VEGF, and MMP9) [41]. In breast cancer, expression of miR-139-3p was downregulated in cancer tissues and cell lines, and its expression was associated with a poor prognosis in breast cancer patients [42]. Overexpression of miR-139-3p inhibited malignant cancer cell phenotypes by targeting RAB1A [42]. Judging from previous studies and our studies, both the guide and passenger strands of miR-139 have tumor-suppressive functions, and the search for molecular pathways controlled by each strand will contribute to the elucidation of the molecular mechanisms of HNSCC.
Our next focus was to identify the oncogenic targets regulated by each strand of miR-139 (miR-139-5p and miR-139-3p) in HNSCC cells. Our in silico analysis identified five genes (PLEC, YWHAQ, GNA12, OLR1, and ACOT9) that were closely associated with a poor prognosis of patients with HNSCC. Importantly, two of those genes (GNA12 and OLR1) were independent prognostic markers in patients with HNSCC.
GNA12 (Ga12) is the alpha subunit of G proteins (heterotrimeric guanine nucleotidebinding protein), which are classified into four subfamilies: Gs, Gi, Gq, and G12 [43]. Ga12 and RhoGEF (Rho guanine nucleotide exchange factors) activate Rho-mediated signaling, which, in turn, enhances cancer cell progression, invasion, and metastasis [44]. High expression of GNA12 was detected in OSCC patients and contributed to cancer cell migration and invasion and lymph node metastasis [45,46].
A previous study showed that OLR1 is a lectin-like scavenger receptor. This receptor binds to several ligands, such as oxidized low-density lipoprotein, polyanionic chemicals, and anionic phospholipids [47]. Aberrant expression of OLR1 has been reported in multiple cancers; its expression induces the release of inflammatory cytokines and is closely related to cancer development and metastasis [48]. Detailed functional analyses of these genes will provide important information for elucidating the molecular pathogenesis of HNSCC.
In conclusion, our new RNA-sequencing-based miRNA signature revealed aberrantly expressed miRNAs in HNSCC tissues, including certain passenger strands of miRNAs. Both strands of pre-miR-139 (miR-139-5p and miR-139-3p) acted as tumor-suppressive miRNAs in HNSCC cells. A total of 28 genes were identified as targets of these miRNAs. Among these targets, GNA12 and OLR1 were demonstrated to be independent prognostic markers in patients with HNSCC. Identification of novel tumor-suppressive miRNAs and their regulated oncogenic targets may improve our knowledge of the molecular mechanisms of the oncogenesis of HNSCC.

Clinical HNSCC Specimens, Normal Epithelial Specimens, and HNSCC Cell Lines
Total RNA was extracted from three paired HNSCC and three normal oral epithelial tissue specimens for RNA sequencing and microarray analysis. The clinical features of HNSCC patients are summarized in Table S1. This study was approved by the Bioethics Committee of Chiba University (approval number: 28-65, 10 February 2015). Informed consent was obtained from all patients. The study's methodologies conformed to the standards set by the Declaration of Helsinki.
The HNSCC cell lines used in this study were obtained from the RIKEN BioResource Center (Tsukuba, Ibaraki, Japan) and are summarized in Table S6.

Generation of a miRNA Expression Signature in HNSCC through RNA Sequencing
To construct the miRNA expression signature of HNSCC, six cDNA libraries were sequenced by using HiSeq 2000 (Illumina, San Diego, CA, USA). The small-RNA-sequencing and data-mining procedures have been described in our previous studies [12,[26][27][28].

Ectopic miRNA Expression Assays of Cell Proliferation, Migration, and Invasion
The procedures used for transient transfection assays of miRNAs, siRNAs, and plasmid vectors were described in our previous studies [12,[14][15][16]. The reagents used are listed in Table S7.

Identification of Putative Targets Controlled by miR-139-5p and miR-139-3p in HNSCC Cells
The strategy used for the selection of the miR-139-5p and miR-139-3p target genes in this study is summarized in Figure 4. We selected putative target genes with the miR-139-5p and miR-139-3p binding sites by using TargetScanHuman ver. 7.2 (http://www. targetscan.org/vert_72/, accessed on 10 July 2020) [49]. The expression profiles of the clinical specimens of HNSCC (genes downregulated in HNSCC tissues) were used to screen miRNA target genes. Our expression data were deposited in the GEO database (accession number: GSE172120). Furthermore, we narrowed down the candidate genes by factoring in clinical information from TCGA-HNSC analyses.
Multivariate Cox regression analyses were also performed with the TCGA-HNSC clinical data and survival data according to the expression levels of each gene from OncoLnc to identify factors associated with HNSCC patient survival. In addition to gene expression, the tumor stage, pathological grade, and age were evaluated as potential independent prognostic factors. The multivariate analyses were performed using JMP Pro 15.0.0 (SAS Institute Inc., Cary, NC, USA).

Direct Regulation of Target Genes by miR-139-3p Using Dual Luciferase Reporter Assays
The psiCHECK-2 vector (C8021; Promega, Madison, WI, USA) was inserted into the synthesized DNA with or without the miR-139-3p binding sequences of GNA12 or OLR1. The plasmid vectors were then transfected into cells using Lipofectamine 2000 (Invitrogen) at a final concentration of 50 ng/well.

Western Blotting and Immunohistochemistry
The Western blotting and immunohistochemistry procedures were detailed in our previous study [12]. The antibodies used are listed in Table S7. The clinical features of 3 HNSCC cases used for immunohistochemical staining are summarized in Table S5.

Analysis of Molecular Pathways Related to Target Genes Controlled by miR-139-3p
To analyze the molecular pathways related to the target genes of miR-139-3p, we performed gene set enrichment analysis (GSEA) for GNA12 and OLR1.
We divided the TCGA-HNSC RNA-seq data into a high-expression group and lowexpression group according to the Z-score of GNA12 or OLR1 on cBioportal. A ranked list of genes was generated by comparing the expression levels of each gene between the two groups. We loaded the obtained gene list into gene set enrichment analysis (GSEA) software [51,52] and used the Molecular Signatures Database (MSigDB) hallmark gene set [53].

Statistical Analysis
All data were presented as mean values and standard errors of at least three independent experiments. Statistical analyses were performed using JMP Pro 15 (SAS Institute Inc., Cary, NC, USA). Differences between 2 groups were evaluated with t-tests. Dunnet's test was used for multiple-group comparisons. p-value < 0.05 was considered statistically significant.

Conclusions
In this study, we focused on miR-139-5p and miR-139-3p based on our miRNA signatures. Our functional assays indicated that these miRNAs play an oncogenic role in HNSCC cells. Using an in silico database analysis to identify the gene targets regulated by miR-139-5p and miR-139-3p, we rapidly identified candidate oncogenes in HNSCC. Our HNSCC miRNA signature and miRNA-based analyses will provide important insights into the molecular pathogenesis of HNSCC.  Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
Our expression data were deposited in the GEO database (accession number: GSE172120).