Regulation of Oncogenic Targets by miR-99a-3p (Passenger Strand of miR-99a-Duplex) in Head and Neck Squamous Cell Carcinoma.

To identify novel oncogenic targets in head and neck squamous cell carcinoma (HNSCC), we have analyzed antitumor microRNAs (miRNAs) and their controlled molecular networks in HNSCC cells. Based on our miRNA signature in HNSCC, both strands of the miR-99a-duplex (miR-99a-5p: the guide strand, and miR-99a-3p: the passenger strand) are downregulated in cancer tissues. Moreover, low expression of miR-99a-5p and miR-99a-3p significantly predicts poor prognosis in HNSCC, and these miRNAs regulate cancer cell migration and invasion. We previously showed that passenger strands of miRNAs have antitumor functions. Here, we screened miR-99a-3p-controlled oncogenes involved in HNSCC pathogenesis. Thirty-two genes were identified as miR-99a-3p-regulated genes, and 10 genes (STAMBP, TIMP4, TMEM14C, CANX, SUV420H1, HSP90B1, PDIA3, MTHFD2, BCAT1, and SLC22A15) significantly predicted 5-year overall survival. Notably, among these genes, STAMBP, TIMP4, TMEM14C, CANX, and SUV420H1 were independent prognostic markers of HNSCC by multivariate analyses. We further investigated the oncogenic function of STAMBP in HNSCC cells using knockdown assays. Our data demonstrated that the aggressiveness of phenotypes in HNSCC cells was attenuated by siSTAMBP transfection. Moreover, aberrant STAMBP expression was detected in HNSCC clinical specimens by immunohistochemistry. This strategy may contribute to the clarification of the molecular pathogenesis of this disease.

1. Introduction

Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer, with approximately 650,000 new cases diagnosed annually and 400,000 HNSCC-related deaths worldwide each year [1]. Tobacco and alcohol drinking habits are the major risk factors of HNSCC carcinogenesis [2]. In the past decade, our understanding of the role of human papillomavirus in the development of oropharyngeal squamous cell carcinoma has significantly changed the treatment strategy of this disease [3,4]. HNSCC is typically diagnosed when already at an advanced stage. Despite advancements in surgery, radiation therapy, and chemotherapy, patients with advanced HNSCC have a poor prognosis [1,4] owing to recurrence, metastasis, and treatment resistance [5]. The median overall survival time for patients with recurrence and metastasis is 10–13 months in the setting of first-line chemotherapy and 6 months in the second-line setting [6]. Recently, epidermal growth factor receptor inhibitors and immune checkpoint inhibitors have emerged as therapeutic approaches in HNSCC treatment [7,8]. However, these treatments do not yield satisfactory results.

MicroRNAs (miRNAs) exist widely in eukaryotes, and more than 2500 types of mature miRNAs have been discovered in humans [9,10]. miRNAs are transcribed from the human genome and then processed into mature miRNAs of approximately 18–22 bases [9,10]. miRNAs are classified as noncoding RNAs and function to suppress the translation of mRNAs by binding to the complementary sequence in the 3′-untranslated region (3′-UTR) of the targeted mRNA [9,10]. Notably, one miRNA targets multiple mRNAs, and there are multiple miRNA binding sites in the UTRs of one mRNA [9,10]. Therefore, changes in the expression of miRNAs are involved in various diseases, including human cancers, suggesting that miRNAs play important roles in disease development [11,12,13,14,15].

We have been studying antitumor miRNAs and their oncogenic networks in HNSCC cells based on HNSCC miRNA signatures [16,17,18,19,20]. Our previous studies have shown that the antitumor miR-29 family directly controls laminin-332 and integrins (ITGA6, ITGB4, and ITGB1), and miR-199 family targets ITGA3 in HNSCC cells [21,22,23]. Moreover, the antitumor miR-26 family, miR-29 family, and miR-218 inhibit cancer cell migration and invasion in HNSCC cells, and these miRNAs coordinately regulate lysyl oxidase like 2 [24]. These antitumor miRNAs (the miR-26 family, the miR-29 family, miR-218, and the miR-199 family) target proteins involved in the epithelial-mesenchymal transition, indicating their pivotal roles in metastasis in cancer cells.

In this study, we focused on miR-99a-5p (the guide strand of the miR-99a-duplex) and miR-99a-3p (the passenger strand) based on our HNSCC miRNA signature determined by RNA sequencing [20]. Previous studies have shown that downregulation of miR-99a-5p occurs in various cancers and that the expression of this miRNA attenuates malignant phenotypes in cancer cells, suggesting that miR-99a-5p acts as an antitumor miRNA [25,26]. However, few reports have described the roles of the passenger strand miR-99a-3p in HNSCC, and oncogenic networks controlled by miR-99a-3p are still unknown. In the general concept of miRNA biogenesis, passenger strands of miRNAs are degraded in the cytosol and have no function [9,10]. However, our previous studies showed that some passenger strands of miRNAs, e.g., miR-145-3p, miR-150-3p, and miR-199a/b-3p were downregulated in the signature and acted as antitumor miRNAs in malignant cells. Importantly, several targets regulated by these passenger strands of miRNAs acted as oncogenes, and their aberrant expressions were closely associated with the poor prognosis of the patients [23,27,28,29,30]. Therefore, the analysis of passenger strands of miRNAs is useful for understanding the molecular pathogenesis of HNSCC.

Our functional assays indicated that ectopic expression of both strands of the miR-99a-duplex significantly attenuated malignant phenotypes in HNSCC cells. We further analyzed miR-99a-3p-regulated oncogenic genes involved in HNSCC molecular pathogenesis. In total, 32 genes were identified as miR-99a-3p-controlled genes, and 10 genes (STAMBP, TIMP4, TMEM14C, CANX, SUV420H1, HSP90B1, PDIA3, MTHFD2, BCAT1, and SLC22A15) significantly predicted 5-year overall survival in patients with HNSCC. Moreover, our findings revealed that aberrant expression of STAMBP enhanced cancer cell aggressiveness in HNSCC.

2. Materials and Methods

2.1. Clinical Human HNSCC Specimens and HNSCC Cell Lines

Twenty-two clinical specimens were obtained from patients with HNSCC following surgical tumor resection at Chiba University Hospital (2008–2013, Chiba, Japan). The patients’ clinical characteristics are shown in Table 1. Written informed consent was obtained from all patients before the use of their specimens. This study was approved by the Bioethics Committee of Chiba University (approval number: 811(690)). Normal tissue was collected from the most distant cancerous part of the same specimen. A total of 22 pairs of HNSCC tissues and adjacent normal (noncancerous) tissues were obtained in this study.

Table 1

Clinical features of 22 HNSCC patients.

No.	Age	Sex	Location	T	N	M	Stage	Differentiation
1	66	M	hypopharynx	2	2c	0	IVa	moderate
2	66	M	hypopharynx	4a	2c	0	IVa	well
3	66	M	hypopharynx	4b	2c	0	IVb	moderate
4	76	M	hypopharynx	4a	1	0	IVa	well
5	74	M	hypopharynx	4a	2c	0	IVa	poor
6	45	M	hypopharynx	4a	2c	0	IVa	moderate
7	75	M	hypopharynx	4a	2c	0	IVa	well
8	58	M	hypopharynx	4a	0	0	IVa	well
9	69	M	larynx	3	0	0	III	well
10	70	M	larynx	4a	1	0	IVa	well-moderate
11	84	M	larynx	4a	0	0	IVa	moderate
12	50	M	larynx	4a	2b	0	IVa	moderate
13	82	M	larynx	4a	0	0	IVa	moderate
14	85	M	larynx	3	2b	0	IVa	moderate
15	66	M	tongue	2	0	0	II	moderate
16	73	M	tongue	3	1	0	III	poor
17	74	M	tongue	1	0	0	I	well
18	72	M	tongue	4a	2b	0	IVa	moderate
19	83	M	oral floor	2	1	0	III	well
20	68	F	oral floor	4a	1	0	IVa	well
21	77	M	oral floor	2	2b	0	IVa	moderate
22	69	M	oropharynx	1	0	0	I	well

T: Primary tumor stage, N: Regional lymph nodes stage, M: Distant metastasis stage. All according to the UICC (The Union for International Cancer Control) classification.

Two HNSCC cell lines, FaDu (American Type Culture Collection, Manassas, VA, UAS) and SAS cells (RIKEN Cell Bank, Tsukuba, Ibaraki, Japan), were used in this study.

2.2. RNA Extraction and Quantitative Real-Time Reverse Transcription Polymerase Chain Reaction (qRT-PCR)

RNA was extracted from clinical specimens and cell lines as previously described [20,23,30,31,32]. miRNA expression levels were evaluated using qRT-PCR as described previously [20,23,30,31,32]. The TaqMan probes and primers used in this study are listed in Table S2.

2.3. Transfection of miRNAs, siRNAs, and Plasmid Vectors into HNSCC Cells

The procedures for transfection of miRNAs, siRNAs, and plasmid vectors into HNSCC cells were described previously [20,23,30,31,32]. The reagents used in this study are listed in Table S2.

2.4. Functional Assays in HNSCC Cells (Cell Proliferation, Migration and Invasion Assays)

The procedures for functional assays in cancer cells (proliferation, migration, and invasion) are described in our previous studies [20,23,30,31,32]. Cells were transfected with 10nM miRNAs or siRNAs. Cell proliferation was evaluated with XTT assays. Migration assays were performed with uncoated transwell polycarbonate membrane filters, invasion assays with modified Boyden chambers.

2.5. Measurement of miR-99a-3p Incorporated into the RISC

Immunoprecipitation using anti-Ago2 antibodies was performed to determine whether miR-99a-3p was incorporated into the RISC. FaDu and SAS were transfected with 10nM miRNAs for 48 h and the collected cells went through immunoprecipitation using human anti-Ago2 antibodies (microRNA Isolation Kit, Human Ago2; Wako, Osaka, Japan) according to the manufacture’s protocol. Obtained miRNAs proceeded to qRT-PCR. For normalization of the results, miR-26a was measured, whose expression was not affected by miR-99a-5p/3p transfection. The procedure for immunoprecipitation was described in previous studies [23,30,31,32]. The reagents used in this study are listed in Table S2.

2.6. Identification of miR-99a-3p and miR-99a-5p Targets in HNSCC Cells

The strategy for identification of miRNA targets in this study is summarized in Figure S5. Two expression profiles (i.e., miR-99a-5p-transfected FaDu cells [GEO accession number: GSE123318], miR-99a-3p-transfected FaDu cells [accession number: GSE123318]) were used in this screening. The TargetScanHuman database (http://www.targetscan.org/vert_72/) was used to predict miRNA binding sites.

2.7. Plasmid Construction and Dual-Luciferase Reporter Assays

Plasmid vectors, including vectors containing the wild-type sequences of miR-99a-3p binding sites in the 3′-UTR of STAMBP or the deletion sequences of miR-99a-3p binding sites in the 3′-UTR of STAMBP, were prepared. The inserted sequences are shown in Figure S7. The procedures for transfection and dual luciferase reporter assays were described in our previous studies [20,23,30,31,32]. The reagents used in this study are listed in Table S2.

2.8. Clinical Data Analyses of miRNAs and Target Genes in HNSCC Specimens

TCGA (https://tcga-data.nci.nih.gov/tcga/) was applied to investigate the clinical significance of miRNAs and their target genes. Gene expression and clinical data were obtained from cBioPortal (http://www.cbioportal.org/) and OncoLnc (http://www.oncolnc.org/) (data downloaded on 1 August 2019).

2.9. Western Blotting and Immunohistochemistry

The procedures for Western blotting and immunohistochemistry were described in our previous studies [20,23,30,31,32]. The antibodies used in this study are listed in Table S2.

2.10. Statistical Analyses

Mann–Whitney U tests were applied for comparisons between two groups. For multiple groups, one-way analysis of variance and Tukey tests for post-hoc analysis were applied. These analyses were performed with JMP Pro 14 (SAS Institute Inc., Cary, NC, USA).

3. Results

3.1. Downregulation and Clinical Significance of miR-99a-5p and miR-99a-3p in HNSCC Clinical Specimens

The clinical features of HNSCC specimens are listed in Table 1. Expression levels of miR-99a-5p and miR-99a-3p were significantly low in cancer tissues compared with those in normal tissues from the same patients (p < 0.0001 and p < 0.0001, respectively; Figure 1A and Figure S1). The expression levels of these miRNAs in two HNSCC cell lines (FaDu and SAS cells) were also very low compared with those in normal tissues (Figure 1A and Figure S1). A positive correlation was detected between miR-99a-5p and miR-99a-3p expression levels by Spearman’s rank analysis (R = 0.716, p < 0.0001; Figure 1B).

Figure 1

Expression and clinical significance of miR-99a-5p and miR-99a-3p in HNSCC clinical specimens. (A) Expression of miR-99a-5p and miR-99a-3p was significantly reduced in HNSCC clinical specimens and cell lines (FaDu and SAS cells). Data were normalized to the expression of RNU48. (B) Spearman’s rank tests showed positive correlations between expression levels of miR-99a-5p and miR-99a-3p in clinical specimens. (C) Kaplan-Meier survival curve analyses of patients with HNSCC using data from The Cancer Genome Atlas (TCGA) database. Patients were divided into two groups according to miRNA expression, high group and low group (according to median expression). The red line shows the high expression group, and the blue line shows the low expression group.

Cohort analysis using data from The Cancer Genome Atlas (TCGA) database revealed that low expression of miR-99a-5p and miR-99a-3p was associated with poorer survival in patients with HNSCC (p = 0.0008 and p = 0.0012, respectively; Figure 1C). We confirmed positive correlation of these miRNAs expression by using TCGA database sets (Figure S2).

3.2. Ectopic Expression of miR-99a-5p and miR-99a-3p on Cell Proliferation, Migration and Invasion in HNSCC Cells

To investigate the anti-tumor functions of miR-99a-5p and miR-99a-3p in HNSCC cells, we assessed changes in cell proliferation, migration, and invasion after transfection of these miRNAs into FaDu and SAS cells. Notably, ectopic expression of miR-99a-5p significantly decreased cell proliferation (Figure 2A). However, cell proliferation was not affected by miR-99a-3p transfection. Additionally, the migration and invasion of FaDu and SAS cells were significantly suppressed by miR-99a-5p and miR-99a-3p transfection (Figure 2B,C). Photomicrographs are presented in Figure S3.

Figure 2

Functional assays of cell proliferation, migration, and invasion following ectopic expression of miR-99a-5p and miR-99a-3p in HNSCC cell lines (FaDu and SAS cells). (A) Cell proliferation was assessed using XTT assays. Data were collected 72 h after miRNA transfection (* p < 0.0001). (B) Cell migration was assessed with membrane culture system. Data were collected 48 h after seeding the cells into the chambers (* p < 0.0001). (C) Cell invasion was determined 48 h after seeding miRNA-transfected cells into chambers using Matrigel invasion assays (* p < 0.0001).

3.3. Incorporation of miR-99a-5p and miR-99a-3p into the RNA-Induced Silencing Complex (RISC) in HNSCC Cells

Ago2 is an essential component of the RISC. Therefore, to verify that miR-99a-5p and miR-99a-3p had functions in HNSCC cells, immunoprecipitation assays were carried out using anti-Ago2 antibodies. After transfection of both miRNAs to SAS cells, the amounts of miR-99a-5p and miR-99a-3p were significantly increased relative to that in control (untransfected) cells (Figure S4). These data showed that miR-99a-5p (the guide strand) and miR-99a-3p (the passenger strand) were both incorporated into the RISC in HNSCC cells.

3.4. Screening of Molecular Targets Regulated by miR-99a-5p and miR-99a-3p in HNSCC Cells

To identify the genes controlled by miR-99a-5p and miR-99a-3p in HNSCC cells, we used gene expression data obtained by RNA microarray analysis of miR-99a-5p- or miR-99a-3p-transfected FaDu cells and data from the TargetScanHuman database (release 7.2), which provided annotated putative targets for each miRNA. Our strategy searching for miR-99a-5p and miR-99a-3p target genes is shown in Figure S5.

Using this strategy, only genes from miR-99a-3p-transfected FaDu cells endured the selection process. For miR-99a-3p, 114 genes were identified as putative target genes in HNSCC cells (Table 2). Eighteen genes were identified as miR-99a-5p-controlled genes, none of which showed correlations with prognosis in TCGA database (Table S1).

Table 2

Candidate target genes regulated by miR-99a-3p.

Entrez GeneID	Gene Symbol	Gene Name	Total Sites	FaDu miR-99a-3pTransfectant FC (log2)	TCGAOncoLnc5-Year OS p-Value
10617	STAMBP	STAM binding protein	1	−1.0548	0.0032
7079	TIMP4	TIMP metallopeptidase inhibitor 4	1	−2.2976	0.0003
51522	TMEM14C	transmembrane protein 14C	1	−3.0740	0.0112
51111	SUV420H1 (KMT5B)	suppressor of variegation 4-20 homolog 1	1	−1.0238	0.0432
821	CANX	calnexin	1	−2.3590	0.0436
10797	MTHFD2	methylenetetrahydrofolate dehydrogenase (NADP+ dependent) 2,methenyltetrahydrofolate cyclohydrolase	1	−1.6298	0.0124
2923	PDIA3	protein disulfide isomerase family A, member 3	1	−1.3332	0.0162
586	BCAT1	branched chain amino-acid transaminase 1, cytosolic	3	−2.0236	0.0296
7184	HSP90B1	heat shock protein 90kDa beta (Grp94), member 1	1	−2.3549	0.0305
55356	SLC22A15	solute carrier family 22, member 15	1	−1.9007	0.0412
23786	BCL2L13	BCL2-like 13 (apoptosis facilitator)	1	−1.2622	0.0604
29967	LRP12	low density lipoprotein receptor-related protein 12	1	−1.0258	0.0897
112752	IFT43	intraflagellar transport 43	1	−1.4595	0.0922
23516	SLC39A14	solute carrier family 39 (zinc transporter), member 14	1	−1.6435	0.1124
55255	WDR41	WD repeat domain 41	1	−1.2410	0.1145
56886	UGGT1	UDP-glucose glycoprotein glucosyltransferase 1	1	−1.0232	0.1228
6137	RPL13	ribosomal protein L13	1	−1.6427	0.1408
114971	PTPMT1	protein tyrosine phosphatase, mitochondrial 1	1	−1.2094	0.1545
27	ABL2	ABL proto-oncogene 2, non-receptor tyrosine kinase	1	−1.8956	0.1549
114818	KLHL29	kelch-like family member 29	1	−1.4701	0.1551
2512	FTL	ferritin, light polypeptide	1	−1.2722	0.1809
84803	AGPAT9 (GPAT3)	1-acylglycerol-3-phosphate O-acyltransferase 9	1	−1.8971	0.2008
23271	CAMSAP2	calmodulin regulated spectrin-associated protein family, member 2	1	−1.0735	0.3277
122953	JDP2	Jun dimerization protein 2	1	−2.2594	0.3311
219902	TMEM136	transmembrane protein 136	1	−1.4956	0.3455
440026	TMEM41B	transmembrane protein 41B	1	−2.3021	0.3843
54629	FAM63B	family with sequence similarity 63, member B	1	−1.2972	0.3940
182	JAG1	jagged 1	1	−1.0082	0.3971
2121	EVC	Ellis van Creveld syndrome	1	−1.7233	0.4006
490	ATP2B1	ATPase, Ca++ transporting, plasma membrane 1	1	−2.0436	0.5557
9208	LRRFIP1	leucine rich repeat (in FLII) interacting protein 1	1	−1.1882	0.6483
50848	F11R	F11 receptor	1	−1.0283	0.7312
79152	FA2H	fatty acid 2-hydroxylase	1	−1.7458	0.0877
23049	SMG1	SMG1 phosphatidylinositol 3-kinase-related kinase	1	−1.1426	0.1267
5337	PLD1	phospholipase D1, phosphatidylcholine-specific	1	−1.0968	0.1538
5935	RBM3	RNA binding motif (RNP1, RRM) protein 3	1	−1.3755	0.1554
135398	C6orf141	chromosome 6 open reading frame 141	1	−1.4022	0.1594
5251	PHEX	phosphate regulating endopeptidase homolog, X-linked	1	−1.0845	0.1616
201229	LYRM9	LYR motif containing 9	1	−1.8318	0.1709
6095	RORA	RAR-related orphan receptor A	1	−1.6688	0.1809
85439	STON2	stonin 2	2	−1.0335	0.2107
114781	BTBD9	BTB (POZ) domain containing 9	1	−1.4022	0.2145
144348	ZNF664	zinc finger protein 664	1	−1.1221	0.2262
27125	AFF4	AF4/FMR2 family, member 4	1	−1.3648	0.2620
152007	GLIPR2	GLI pathogenesis-related 2	1	−1.8377	0.2882
688	KLF5	Kruppel-like factor 5 (intestinal)	1	−1.0461	0.3321
27250	PDCD4	programmed cell death 4 (neoplastic transformation inhibitor)	1	−1.3298	0.3380
440295	GOLGA6L9	golgin A6 family-like 9	2	−1.5553	0.3385
55175	KLHL11	kelch-like family member 11	1	−1.1230	0.3808
85015	USP45	ubiquitin specific peptidase 45	1	−1.0026	0.3813
27109	ATP5S	ATP synthase, H+ transporting, mitochondrial Fo complex, subunit s (factor B)	1	−1.0441	0.3939
10802	SEC24A	SEC24 family member A	1	−1.1434	0.4081
2639	GCDH	glutaryl-CoA dehydrogenase	1	−1.2904	0.4110
843	CASP10	caspase 10, apoptosis-related cysteine peptidase	1	−1.1677	0.4282
8774	NAPG	N-ethylmaleimide-sensitive factor attachment protein, gamma	1	−1.1850	0.4452
54462	CCSER2	coiled-coil serine-rich protein 2	1	−1.3840	0.4566
9848	MFAP3L	microfibrillar-associated protein 3-like	1	−1.1114	0.4710
64764	CREB3L2	cAMP responsive element binding protein 3-like 2	1	−2.2962	0.4754
334	APLP2	amyloid beta (A4) precursor-like protein 2	1	−1.3719	0.5068
5163	PDK1	pyruvate dehydrogenase kinase, isozyme 1	1	−1.1852	0.5112
10124	ARL4A	ADP-ribosylation factor-like 4A	1	−1.4088	0.5161
145781	GCOM1	GRINL1A complex locus 1	1	−1.0388	0.5289
3987	LIMS1	LIM and senescent cell antigen-like domains 1	1	−1.0346	0.5454
57498	KIDINS220	kinase D-interacting substrate, 220kDa	1	−1.4304	0.5664
285636	C5orf51	chromosome 5 open reading frame 51	1	−1.0631	0.5666
9761	MLEC	malectin	1	−1.6201	0.5725
54477	PLEKHA5	pleckstrin homology domain containing, family A member 5	1	−1.3145	0.5846
10221	TRIB1	tribbles pseudokinase 1	1	−1.7224	0.5958
54014	BRWD1	bromodomain and WD repeat domain containing 1	1	−1.0694	0.6033
390	RND3	Rho family GTPase 3	1	−1.0681	0.6139
55823	VPS11	vacuolar protein sorting 11 homolog (S. cerevisiae)	1	−1.0569	0.6269
8444	DYRK3	dual-specificity tyrosine-(Y)-phosphorylation regulated kinase 3	1	−1.9227	0.6459
1978	EIF4EBP1	eukaryotic translation initiation factor 4E binding protein 1	1	−1.2832	0.6529
8874	ARHGEF7	Rho guanine nucleotide exchange factor (GEF) 7	1	−1.2069	0.6717
309	ANXA6	annexin A6	1	−1.9168	0.6821
5784	PTPN14	protein tyrosine phosphatase, non-receptor type 14	2	−1.2144	0.6888
100534599	ISY1-RAB43	ISY1-RAB43 readthrough	1	−2.4556	0.6928
54431	DNAJC10	DnaJ (Hsp40) homolog, subfamily C, member 10	2	−1.6506	0.7051
63874	ABHD4	abhydrolase domain containing 4	1	−1.6850	0.7071
196	AHR	aryl hydrocarbon receptor	1	−1.1288	0.7219
63897	HEATR6	HEAT repeat containing 6	1	−1.0711	0.7291
10961	ERP29	endoplasmic reticulum protein 29	1	−1.0578	0.7355
126626	GABPB2	GA binding protein transcription factor, beta subunit 2	2	−1.1151	0.7488
79794	C12orf49	chromosome 12 open reading frame 49	1	−1.7522	0.7864
5965	RECQL	RecQ helicase-like	3	−1.2203	0.7885
64651	CSRNP1	cysteine-serine-rich nuclear protein 1	1	−2.1079	0.8036
81558	FAM117A	family with sequence similarity 117, member A	1	−2.0458	0.8054
7706	TRIM25	tripartite motif containing 25	2	−1.2522	0.8360
55339	WDR33	WD repeat domain 33	1	−1.5940	0.8369
10097	ACTR2	ARP2 actin-related protein 2 homolog (yeast)	1	−1.0537	0.8586
23348	DOCK9	dedicator of cytokinesis 9	1	−1.2887	0.8621
10079	ATP9A	ATPase, class II, type 9A	1	−2.2117	0.8625
9497	SLC4A7	solute carrier family 4, sodium bicarbonate cotransporter, member 7	1	−1.1702	0.9121
54832	VPS13C	vacuolar protein sorting 13 homolog C (S. cerevisiae)	2	−1.1051	0.9176
23433	RHOQ	ras homolog family member Q	1	−1.6436	0.9319
55727	BTBD7	BTB (POZ) domain containing 7	1	−1.2974	0.9480
11260	XPOT	exportin, tRNA	1	−1.4398	0.9544
1362	CPD	carboxypeptidase D	2	−1.3036	0.9645
151887	CCDC80	coiled-coil domain containing 80	2	−2.1921	0.9667
116496	FAM129A	family with sequence similarity 129, member A	1	−1.8099	0.9903
9709	HERPUD1	homocysteine-inducible, endoplasmic reticulum stress-inducible, ubiquitin-like domain member 1	1	−3.3737	0.0854*
284723	SLC25A34	solute carrier family 25, member 34	1	−1.5051	0.0757*
83641	FAM107B	family with sequence similarity 107, member B	1	−2.2173	0.0749*
60412	EXOC4	exocyst complex component 4	1	−1.0957	0.0691*
6700	SPRR2A	small proline-rich protein 2A	1	−2.1995	0.0568*
10365	KLF2	Kruppel-like factor 2	1	−1.3378	0.0386*
3572	IL6ST	interleukin 6 signal transducer	1	−1.3786	0.0330*
10551	AGR2	anterior gradient 2	1	−3.7365	0.0134*
9663	LPIN2	lipin 2	1	−2.6332	0.0033*
155435	RBM33	RNA binding motif protein 33	1	−1.3082	0.0029*
54855	FAM46C	family with sequence similarity 46, member C	2	−1.0118	0.0028*
728661	SLC35E2B	solute carrier family 35, member E2B	1	−1.0811	0.0004*
23591	FAM215A	family with sequence similarity 215, member A (non-protein coding)	1	−1.4327	N/A
643707	GOLGA6L4	golgin A6 family-like 4	2	−1.3773	N/A

* poor prognosis in patients with low gene expression.

3.5. Clinical Significance of miR-99a-3p Targets in HNSCC Pathogenesis

By using TCGA database, we narrowed down the list of 114 genes according to correlations with 5-year overall survival rates. Among the genes, high expression of 10 genes (STAMBP, TIMP4, TMEM14C, CANX, SUV420H1, HSP90B1, PDIA3, MTHFD2, BCAT1, and SLC22A15) was associated with poor prognosis (5-year overall survival rate: p < 0.05) in patients with HNSCC (Table 2 and Figure 3). Furthermore, multivariate analysis elucidated that expression levels of five genes (STAMBP, TIMP4, TMEM14C, SUV420H1, and CANX) were independent prognostic factors for 5-year overall survival in these patients (Figure 4).

Figure 3

Clinical significance of miR-99a-3p target genes in TCGA database. Among putative targets of miR-99a-3p in HNSCC cells, high expression of 10 genes (STAMBP, TIMP4, TMEM14C, CANX, SUV420H1, HSP90B1, PDIA3, MTHFD2, BCAT1, and SLC22A15) was significantly associated with poor prognosis in patients with HNSCC. Kaplan-Meier curves of 5-year overall survival for each gene are shown.

Figure 4

Forest plot of multivariate analysis of five genes (STAMBP, TIMP4, TMEM14C, SUV420H1, and CANX), which were independent prognostic factors for overall survival after adjustment for patient age, disease, stage, and pathological grade.

3.6. Direct Regulation of STAMBP by miR-99a-3p in HNSCC Cells

In cells transfected with miR-99a-3p, the levels of STAMBP mRNA and STAMBP protein were significantly lower than in mock- or miR-control-transfected cells (Figure 5A,B). The whole pictures of Western blotting are shown in Figure S6. In addition, we investigated whether the other four genes (TIMP4, TMEM14C, SUV420H1, and CANX) were controlled by miR-99a-3p in HNSCC cells at the RNA levels. Consistent with the estimation of TargetScanHuman database, the expression levels of the four genes were also downregulated by miR-99a-3p transfection in HNSCC cells (Figure S8).

Figure 5

Expression of STAMBP/STAMBP was directly regulated by miR-99a-3p in HNSCC cells. (A) Expression of STAMBP mRNA was significantly reduced by miR-99a-3p transfection into FaDu and SAS cells (72 h after transfection; * p < 0.0001, N.S.: Not significant). Expression of GAPDH was used as an internal control. (B) Expression of STAMBP protein was reduced by miR-99a-3p transfection into HNSCC cells (72 h after transfection). Expression of GAPDH was used as an internal control. (C) TargetScanHuman database analyses predicted one putative miR-99a-3p binding site in the 3′-UTR of STAMBP. (D) Dual luciferase reporter assays showed that luminescence activities were reduced by cotransfection with wild-type (miR-99a-3p binding site) vectors and miR-99a-3p in FaDu and SAS cells. Normalized data were calculated as Renilla/firefly luciferase activity ratios (N.S.: Not significant).

Next, we performed dual-luciferase reporter assays to determine whether STAMBP was directly regulated by miR-99a-3p. We used vectors encoding the partial wild-type sequences of the 3′-UTR of STAMBP, including the predicted miR-99a-3p target site deletion vector lacking the miR-99a-3p binding site (Figure 5C and Figure S7). We found that luciferase activity was significantly decreased by cotransfection with miR-99a-3p and the vector carrying the wild-type 3′-UTR of STAMBP, whereas transfection with the deletion vector blocked the decrease in luminescence in FaDu and SAS cells (Figure 5D). These data demonstrated that miR-99a-3p directly bound to the 3′-UTR of STAMBP.

3.7. Effects of STAMBP Knockdown on Cell Proliferation, Migration, and Invasion in HNSCC Cells

To investigate the oncogenic functions of STAMBP in HNSCC cells, knockdown assays were conducted using small interfering RNAs (siRNAs). Both mRNA and protein expression levels were successfully suppressed by siSTAMBP-1 and siSTAMBP-2 transfection into FaDu and SAS cells (Figure 6A,B). The whole pictures of Western blotting are shown in Figure S6.

Figure 6

Effects of STAMBP knockdown on cell proliferation, migration, and invasion in HNSCC cells. (A) Expression of STAMBP mRNA was significantly reduced by siRNA transfection into HNSCC cells (* p < 0.0001). Expression of GAPDH was used as an internal control. (B) Expression of STAMBP protein was markedly reduced by siRNA transfection into HNSCC cells. Expression of GAPDH was used as an internal control. (C) Cell proliferation was assessed using XTT assays. Data were collected 72 h after miRNA transfection (* p < 0.0001). (D) Cell migration was assessed with a membrane culture system. Data were collected 48 h after seeding the cells into the chambers (* p < 0.0001). (E) Cell invasion was determined 48 h after seeding miRNA-transfected cells into chambers using Matrigel invasion assays (* p < 0.0001).

In functional assays, cell proliferation was not suppressed by siSTAMBP transfection into FaDu cells. Besides, in SAS cells, cell proliferation was significantly suppressed by siSTAMBP transfection (Figure 6C). Cell migration and invasive abilities were significantly blocked by knockdown of STAMBP (siSTAMBP-1 and siSTAMBP-2) in FaDu and SAS cells (Figure 6D,E). The photomicrographs are shown in Figure S3. Regarding the cell proliferation assay, the results differed between FaDu cells and SAS cells. To explain this phenomenon, a detailed analysis of genes involved in cell cycle and cell division for two cell lines will be necessary.

3.8. Overexpression of STAMBP in HNSCC Clinical Specimens

Expression of STAMBP protein was evaluated using HNSCC clinical specimens. Overexpression of STAMBP was detected in cancer lesions in HNSCC clinical specimens (Figure 7A–H). In contrast to cancer lesions, expression of STAMBP was extremely weak in normal mucosa (Figure 7J). Information on clinical specimens used for immunostaining is shown in Table 3.

Figure 7

Overexpression of STAMBP in HNSCC clinical specimens. (A–I) Expression of STAMBP was investigated by immunohistochemical staining of HNSCC clinical specimens. Overexpression of STAMBP was detected in the nuclei and/or cytoplasm of cancer cells. (J) Extremely weak expression of STAMBP in normal mucosa of larynx and pharynx.

Table 3

Clinical features of 9 HNSCC cases used for immunohistochemical staining.

	Age	Sex	Location	T	N	M	Stage	Differentiation
A	80	M	larynx	3	2c	0	IVa	moderate
B	73	M	larynx	3	0	0	III	poor
C	77	M	oral	2	2b	0	Iva	moderate
D	42	F	oral	4a	0	0	IVa	poor
E	51	M	oral	2	0	0	II	well
F	52	F	oral	4a	2c	1	Ivc	well
G	72	M	hypopharynx	2	0	0	II	moderate
H	64	M	hypopharynx	2	2b	0	IVa	well
I	70	M	hypopharynx	2	2b	0	Iva	well

To confirm our immunostaining results, we analyzed gene expression data of GEO database (accession number: GSE6631). Analysis of gene expression data showed that expression of STAMBP was significantly upregulated in HNSCC clinical specimens (Figure S9).

4. Discussion

Owing to the high rate of recurrence and metastasis in HNSCC, HNSCC is still a deadly cancer, with an average 50% overall 5-year survival rate [1,2,3,4,5,6]. In order to improve treatment outcomes in patients with HNSCC, it is essential to develop treatments for cases with recurrence and metastasis. Advanced genomic approaches are effective for elucidating the molecular pathogenesis of HNSCC, leading to the identification of molecular targets for treatment.

As part of the unique biological nature of miRNAs, a single miRNA can control (directly or indirectly) many RNA transcripts in each cell. Therefore, the aberrant expression of miRNA influences multiple pathways, including cell proliferation, migration, invasion, and apoptosis. Aberrantly expressed miRNAs disrupt RNA expression networks, resulting in cancer cell initiation, development, metastasis, and drug resistance [11,12,13,14,15]. Accordingly, we have sequentially identified antitumor miRNAs and their controlled molecular targets and pathways in HNSCC cells based on miRNA signatures [16,17,18,19,20]. We recently created an HNSCC miRNA expression signature by RNA sequencing [20]. Notably, our signatures revealed that some miRNA passenger strands, e.g., miR-143-5p, miR-145-3p, miR-150-3p, miR-199a-3p, and miR-199b-3p were downregulated in HNSCC tissues and that their expression status was closely involved in HNSCC molecular pathogenesis [20,23,27]. More recently, our group revealed that passenger strands of miRNAs exert antitumor roles by targeting several oncogenes in prostate cancer, renal cell carcinoma, esophageal squamous cell carcinoma, and lung cancer [28,29,30,31,32,33,34]. The participation of passenger strands of miRNAs in carcinogenesis is a new concept in cancer research.

In this study, we revealed that both strands of the miR-99a-duplex (miR-99a-5p and miR-99a-3p) acted as antitumor miRNAs in HNSCC cells. Moreover, we showed that 18 and 30 genes were putative targets of miR-99a-5p and miR-99a-3p regulation, respectively. Many studies have shown that miR-99a-5p acts as an antitumor miRNA in various cancers by targeting many oncogenes [35,36,37,38,39]. In contrast to miR-99a-5p, few papers have analyzed the functional significance of miR-99a-3p in cancer cells. Previously, downregulation of miR-99a-3p was detected in castration-resistant prostate cancer (CRPC), and ectopic expression of miR-99a-3p was found to attenuate cancer cell aggressive phenotypes [40]. Moreover, miR-99a-3p was shown to regulate non-SMC condensin I complex subunit G directly, and its overexpression was detected in CRPC clinical specimens, showing a significant association with shorter disease-free survival and advanced clinical stage [40]. In renal cell carcinoma cells, miR-99a-3p significantly inhibits cell proliferation and colony formation through regulating ribonucleotide reductase regulatory subunit-M2 [41]. More recently, lower expression of miR-99a-3p and its mediated molecular pathways were detected in HNSCC by in silico analysis, TCGA database, and Genotype-Tissue Expression sequencing databases [42]. These studies indicated that the downregulation of miR-99a-3p was closely involved in cancer pathogenesis.

In this study, we aimed to identify oncogenic targets regulated by miR-99a-3p in HNSCC cells. In total, 114 genes were identified as miR-99a-3p-controlled genes, and 10 of these genes (STAMBP, TIMP4, TMEM14C, CANX, SUV420H1, HSP90B1, PDIA3, MTHFD2, BCAT1, and SLC22A15) significantly predicted 5-year overall survival. Notably, among these genes, STAMBP, TIMP4, TMEM14C, CANX, and SUV420H1 were independent prognostic markers of HNSCC, as demonstrated by multivariate analyses. Our preliminary analysis has shown that these genes were controlled by miR-99a-3p in HNSCC cells (Figure S8). Further detailed examinations are necessary in the future. A previous study showed that TIMP4 secretion was regulated by the expression of LOX/SNAI2 axis and contributed to the malignant phenotype of cancers, e.g., thyroid cancer, colon cancer, and breast cancer [43]. Calnexin CANX is an integral protein of the endoplasmic reticulum and acts as a chaperon. In colorectal cancer, overexpression of CANX predicted poor prognosis of the patients, and its knockdown attenuated aggressive phenotypes of cancer cells [44]. Another study showed that serum levels of CANX were significantly higher in patients with lung cancer, and its expression was a useful sero-diagnostic marker of the patients [45]. Overexpression of SUV420H1 (acts as lysine methyl transferase) enhanced oncogenic ERK signaling through ERK phosphorylation and transcription [46]. These genes may be candidate prognostic markers and therapeutic targets in HNSCC. Functional analysis of these genes will reveal new molecular pathologies for HNSCC.

Among these targets, we further investigated the oncogenic roles of STAMBP in HNSCC cells. STAM-binding protein (STAMBP) is a deubiquitinating enzyme that interacts with the SH3 domain of STAM. This protein plays key roles in cell surface receptor-mediated endocytosis and sorting and in cytokine-mediated signaling for MYC induction and cell cycle progression [47,48,49,50,51]. Whole-exome sequencing revealed that the microcephaly-capillary malformation syndrome was related to recessive mutations in STAMBP [52]. In cancer research, almost no functional analysis of STAMBP has been conducted. A recent study showed that STAMBP expression contributes to melanoma cell migration and invasion through the stabilization of SLUG expression [53]. This result was consistent with our current HNSCC data. In this study, overexpression of STAMBP was detected in HNSCC clinical specimens; knockdown assays using siRNAs demonstrated that migration and invasion were significantly reduced in HNSCC cells. Thus, overexpression of STAMBP may promote cancer cell metastasis. Further studies are needed to analyze the molecular networks controlled by STAMBP in various cancers.

5. Conclusions

Based on the miRNA expression signature of HNSCC, we revealed that miR-99a-3p (the passenger strand) acted as an antitumor miRNA in HNSCC cells. In total, 10 genes (STAMBP, TIMP4, TMEM14C, CANX, SUV420H1, HSP90B1, PDIA3, MTHFD2, BCAT1, and SLC22A15) were regulated by miR-99a-3p in HNSCC cells and were closely involved in HNSCC molecular pathogenesis. STAMBP expression was directly controlled by miR-99a-3p, and its overexpression enhanced cancer cell migration and invasion. Our strategy, i.e., identification of antitumor miRNAs and their targets, may be an attractive tool to reveal novel prognostic and therapeutic targets in HNSCC.