Long Non-Coding RNA Deregulation in Tongue Squamous Cell Carcinoma

Background. The deregulated tumorigenic long non-coding RNA (lncRNA) has been reported in several malignancies. However, there is still no comprehensive study on tongue squamous cell carcinoma (SCC). Methods. Functional reannotation for the human lncRNA was carried out by ncFANs. Real-time quantitative PCR was used to validate the identified lncRNAs. Results. Using the functional annotation algorithm from ncFANs, 8 differentially expressed lncRNAs were identified. Lnc-PPP2R4-5, lnc-SPRR2D-1, lnc-MAN1A2-1, lnc-FAM46A-1, lnc-MBL2-4:1, and lnc-MBL2-4:3 were upregulated in the microdissected tongue SCC tissues. In comparison, lnc-AL355149.1-1 and lnc-STXBP5-1 showed significant downregulation. High level of lnc-MBL2-4:3 was significantly associated with the node positive tongue SCC patients. Further, patients with advanced T-stage demonstrated a further reduction of lnc-AL355149.1-1 in the tumor tissues. Treatment of tongue SCC cells with 5-fluorouracil and paclitaxel can reserve the expression patterns observed in the tongue SCC tissues. Further, changes of lnc-MBL2-4:3 and lnc-AL355149.1-1 expression levels were noticed in the cisplatin-resistant tongue SCC cells. Conclusions. Our results demonstrated that functional reannotation allows us to identify novel lncRNAs using the existing gene expression array dataset. The association of lncRNA with the T-stage and nodal status of tongue SCC patients suggested that lncRNA deregulation was involved in the pathogenesis of tongue SCC.


Introduction
Tongue squamous cell carcinoma (SCC) is a common epithelial cancer identified in the oral cavity. In USA, tongue is the most common site of oropharyngeal cancer (in comparison with mouth, pharynx, and others in oral cavity) and the major histological form is squamous cell carcinoma [1]. Major causative factors included tobacco consumption and alcohol abuse. Viral infection including human papilloma virus and Epstein-Barr virus infection might also play a part in the development of tongue SCC [2]. Tongue SCC is an aggressive tumor with rapid growth rate and high chance of regional and distant metastasis. Tumor dimension and the existence of extracapsular spread (ECS) are predictors of survival [3]. In addition, regional spreading to the cervical lymph node and distant metastasis of tongue SCC are indicators of poor prognosis [4,5].
Different from the mRNA transcript, the codon on the long non-coding RNA (lncRNA) do not code for any peptides or proteins. lncRNA refers to RNA molecules with size over 200 b.p. long and without protein coding functions. To date, the precise control mechanisms of lncRNA are not completely understood. LncRNA could be transcribed by RNA polymerases I, II, and III [6]. LncRNA functions as epigenetic regulators in the somatic cells and is directly involved in cell cycle regulation and cell differentiation [7]. Although the precise regulatory mechanism of lncRNA is not fully understood, evidence suggested that expression of certain lncRNA is being modulated by external stimulus, such as cellular irradiation and chemotherapeutic agents treatment [7,8]. Further, lncRNA is alleged to be involved in the development of resistant phenotype and deters the efficacy of cancer treatment [9].
To the best of our knowledge, comprehensive lncRNA study aiming at identifying novel lncRNA signature has not yet been carried out in the tongue squamous cell carcinoma. Hence, in the present study, we aimed at identifying the candidate lncRNA associated with tongue SCC using the  OTTHUMG00000020778  lnc-PPP2R4-5  tggattttcatgcctgctg  ggctgcattaccagaaaggt  3  OTTHUMG00000012449  lnc-SPRR2D-1  gcctctcctgcaagtgtga  tcctcatttatgacattttcagtctc  5  OTTHUMG00000012147  lnc-MAN1A2-1  gagaccgaggaatcttgctg  ctcagtgggctcagtaatgct  21  OTTHUMG00000015099  lnc-FAM46A-1  aggggtctcttgtccttggt  atcctcttattggcacactgc  26  OTTHUMT00000048111  lnc-MBL2-4:1  gcagccctggagagtttatct  cagcataatatggatgtttgaagg  67  OTTHUMT00000048112  lnc-MBL2-4:3  gagccagcaaaggagactga  cccagaaggggctcttactc  36  OTTHUMG00000002490  lnc-AL355149.1-1  gaaaactaggcgtctgggaac  caaacaatgggagcaagtcc  25  OTTHUMG00000015764 lnc-STXBP5-1 gctatgggaattatttttcctgtg ggtaagccagttttcccttttt 16 * Probe number in the universal probe library. microarray dataset available in the public microarray data repository. Real-time quantitative PCR was then used to confirm the expression and validate the results in the primary tongue SCC tissues and the paired normal epithelia. We also correlated the expression patterns with the clinical characteristics of tongue SCC patients in order to reveal any potential clinical use of the lncRNA.

Functional Reannotation.
The probes on the HG-U133 Plus 2.0 array were reannotated for human lncRNA using non-coding RNA function annotation server (ncFANs) as described [10]. Differential expressed lncRNAs were selected using Student's t-test. A value below 0.05 was considered as differentially expressed lncRNAs. The value of differential expression was adjusted with the Benjamini and Hochberg correction for multiple comparisons.

Development of Cisplatin-Resistant HN21B Cell Line.
The cisplatin-resistant HN21B cell line was developed by chronic cisplatin treatment. HN21B cells were exposed to cisplatin for 3 days, followed by growth recovery in drugfree medium. The concentration of cisplatin was increased in the subsequent cycle and the procedure was repeated until resistance was achieved in HN21B cells.

Statistical
Analysis. Statistical analysis was performed using SPSS V16.0 (SPSS, Chicago, IL). The statistical difference between tongue SCC and paired normal epithelia was examined using Wilcoxon signed-rank test. All the tests were two-sided. value <0.05 was considered as statistically significant.

Discussion
Expression alterations of lncRNA have been reported in several human tumors. In the past, lncRNA was regarded as functionless RNA fragment as it carries no protein coding information. With the characterization of functional noncoding RNA such as microRNA and PIWI-interacting RNAs (piRNAs) in the recent years, considerable attention has been dedicated to identify the active players during the progression and development of human malignancies. Although the deregulated lncRNA patterns have been identified in several cancers, such as prostate and liver cancers, little is known in oral tongue carcinoma. Using serial analysis of gene expression (SAGE), lncRNA expression has been demonstrated in the oral cavity and premalignant oral lesions located on tongue, gingiva, and buccal mucosa [11]. Recently, Fang et al. evaluated the expression patterns of lncRNA UCA1 (urothelial cancer-associated 1) in tongue squamous cell carcinoma and revealed that high UCA1 expression was linked to the migratory ability of the epithelial cancer cells and regional lymph node metastasis [12]. Another example is lncRNA MEG3 (maternally expressed gene 3). MEG3 is a lncRNA with the ability to regulate DNA methyltransferase 3B and was found to be downregulated in the tongue SCC tissues [13]. Low MEG3 level was an independent prognostic indicator and was associated with poor survival of tongue SCC patients [13]. The differentially expressed lncRNA could also be detected in the saliva and was suggested to be a useful noninvasive biomarker for oral cancer detection [14].
In this study, we used the existing microarray data to explore the differentially expressed lncRNA patterns in the tongue SCC tissues. The probesets on the microarray chips were originally designed to detect the protein-coding genes. Usually, multiple probes are assigned to individual genes to cover the entire length of the transcript in order to ensure the accuracy of measurement. With the advance in our understanding of lncRNA and their sequence, it was recognized that particular probesets on the microarray chips match with the lncRNA sequence. By reannotation of the microarray for lncRNA, the expression information of the microarray dataset could be revealed for subsequent inspection [15]. To ascertain that the identified lncRNA is deregulated in the tongue SCC, we carried out real-time quantitative PCR to validate the expression discrepancy between the epithelial cancers and normal epithelia of the same patients. One of the limitations of this study is the sample size of paired tumor and normal tissue is not big enough. Of the 8 lncRNAs, 4/8 (50%) were found to be deregulated in our tongue SCC cohort. To explore the potential role of the lncRNAs in the pathogenesis of tongue SCC, we correlated the expression levels with the clinicopathological parameters of tongue SCC patients. The association of lncRNA with T-stages and regional nodal status suggested that the lncRNA deregulation is potentially implicated in the rapid growth and high migratory property of the tongue SCC.
LncRNA could contribute to the cancer progression by controlling the process of proliferation and migration [16,17]. In view of the association with tumor stages (OTTHUMG00000002490) and regional nodal status (OTTHUMT00000048112) of oral tongue patients, we suggested that lncRNA expression could possibly link to the progression of tongue squamous cell carcinoma. To explore the association, we treated the tongue cancer cell line with chemotherapeutic drugs which target highly proliferating cell and cell with high migratory potency [18,19]. Both OTTHUMG00000002490 and OTTHUMT0000004811 demonstrated significant changes and the expression changes were responsive to the treatment dosage of 5-Fu and paclitaxel, revealing that the 2 lncRNAs can modulate the response of tongue cancer cells to chemotherapeutic agents. In addition, targeting lncRNA is suggested to be a feasible approach to overcome the resistance against chemotherapeutic drugs due to the observation that cancer cells expressing particular lncRNA are less responsive to the cytotoxic drugs [20,21]. For example, in non-small cell lung cancer, it has been reported that cancer cells with high lncRNA AK126698 expression are associated with the cisplatin resistant phenotype [22]. As cisplatin-resistance is a challenge to the treatment efficacy of tongue squamous cell carcinoma, we developed a cisplatin-resistant model to examine the potential implication of the 2 lncRNAs to the development of resistant phenotype in oral tongue cancers. The significant changes in OTTHUMG00000002490 and OTTHUMT0000004811 observed in the resistant cell line indicated the potential association with the development of cisplatin resistance in tongue SCC.

Conclusions
In conclusion, our results indicated that oncogenic/tumorsuppressing lncRNA could potentially be identified through computational exploration of the existing microarray data. The associations of lncRNA with the clinical features of tongue SCC patients substantiate the claim that lncRNA deregulation is linked to the biology activity of tongue SCC. Whether the deregulated lnRNA could possibly be used as diagnostic/prognostic indicators has yet to be elucidated. Further studies are warranted to decipher the functional roles of lncRNA in the pathogenesis of tongue SCC.