Identification and validation a TGF-β-associated long non-coding RNA of head and neck squamous cell carcinoma by bioinformatics method

Abstract

Background: The role of transforming growth factorβ (TGF-β)-induced tumor progression in advanced malignancy is well established, but the involvement of long non-coding RNAs (lncRNAs) in TGF-β signaling remains unclear. This study aimed to identify TGF-β-associated lncRNAs in head and neck squamous cell carcinoma (HNSCC).

Methods: Expression profiling of lncRNAs was obtained using Gene Expression Omnibus and The Cancer Genome Atlas. Real-time quantitative PCR was used to analyze the expression of EPB41L4A-AS2 in HNSCC cell line. We used bioinformatics resources (DAvID) to conduct Gene Ontology biological processes and KEGG pathways at the significant level. Wound healing assay, cell migration and invasion assays, were used to examine the effects of EPB41L4A-AS2 on tumor cell metastasis in vivo. Protein levels of EPB41L4A-AS2 targets were determined by western blot.

Results: A novel TGF-β-associated lncRNA, EPB41L4A-AS2, was found downregulated by TGF-β and associated with invasion and metastasis. The relationship of EPB41L4A-AS2 with the clinicopathological features and prognosis of HNSCC patients was evaluated. Bioinformatic analyses revealed that EPB41L4A-AS2 may be involved in processes associated with the tumor-associated signaling pathway, especially the TGF-β signaling pathway. Furthermore, a TGF-β-induced epithelial-to-mesenchymal transition (EMT) model was established. Low EPB41L4A-AS2 expression was determined, and overexpression of this gene inhibited cell migration and invasion in the EMT model. Moreover, EPB41L4A-AS2 suppressed TGFBR1 expression.

Conclusions: EPB41L4A-AS2 might serve as a negative regulator of TGF-β signaling and as an effective prognostic biomarker and important target in anti-metastasis therapies of HNSCC patients.

Keywords: EMT; EPB41L4A-AS2; Head and neck squamous cell carcinoma; TGF-β.

Fig. 1

The flow chart for bioinformatics analysis

Fig. 2

TGFβ upregulation in HNSCC and association with worse patient survival. a, b TGFβ is upregulated in HNSCC. c ROC analysis was performed to determine the cutoff value of TGFβ that could differentiate between the HNSCC patients with low and high TGFβ levels (p < 0.001; area under the ROC curve, 0.939; cutoff value, 10.775). d Based on the expression levels of TGFβ Kaplan–Meier survival analysis, patients with high TGFβ expression had significantly worse overall survival than those with low TGFβ expression. The log-rank test was used to calculate p value. *** p < 0.001

Fig. 3

Three dysregulated TGF-β-associated lncRNAs were found in HNSCC. a, b A negative correlation was observed between EPB41L4A-AS2 as well as LINC00515 and TGF-β. c A positive correlation was observed between MIR4435-2HG and TGF-β

Fig. 4

Significant correlation of EPB41L4A-AS2 with clinicopathological features and patient survival in HNSCC. a–c A negative relationship between EPB41L4A-AS2 expression and N classification, T classification, and tumor stage was found. d Low EPB41L4A-AS2 expression significantly correlated with patients with perineural invasion. e, f High EPB41L4A-AS2 expression was significantly positively associated with OS and RFS in HNSCC

Fig. 5

EPB41L4A-AS2 is downregulated in the CNE1-EMT model. a TGF-β-induced cell scattering and morphologic changes in CNE-1 cells. b Western blot analysis to detect the expression of Snail, vimentin, E-cadherin, N-cadherin and β-actin. c The relative protein expression levels were represented as columns. d The expression of EPB41L4A-AS2 was downregulated in TGF-β-induced EMT model. e EPB41L4A-AS2 expression level in the 3-day TGF-β treatment was downregulated. * p < 0.05 and ** p < 0.01

Fig. 6

EPB41L4A-AS2 inhibits cell migration and invasion in the CNE1-EMT model. a Quantification of EPB41L4A-AS2 mRNA expression by qRT-PCR. b Effect on migration by wound healing assays. c Effect on migration by transwell assays. d Effect on invasion by transwell assays. e Western blot analysis to detect the expression of E-cadherin, vimentin, N-cadherin and β-actin. f The relative protein expression levels were represented as columns. * p < 0.05 and *** p < 0.001

Fig. 7

EPB41L4A-AS2 inhibits the TGFBR1/SMAD2/SMAD3 axis. a Western blot analysis to detect the expression of TGFBR1, Smad2, p-Smad2, Smad3, p-Smad3, Snail and β-actin. b The relative protein expression levels were represented as columns. c–e A negative correlation was found between EPB41L4A-AS2 and TGFBR1 or Snail1/2. * p < 0.05 and ** p < 0.01

Fig. 8

Analyses of the potential molecular mechanisms of EPB41L4A-AS2 by a bioinformatics-based method. a, b ceRNA and lncRNA-mRNA pathway network was established. c Regulatory network of EPB41L4A was constructed