Identifying the Transcriptional Regulatory Network Associated With Extrathyroidal Extension in Papillary Thyroid Carcinoma by Comprehensive Bioinformatics Analysis

Abstract

Extrathyroidal extension (ETE) affects papillary thyroid cancer (PTC) prognosis. The objective of this study was to identify biomarkers for ETE and explore the mechanisms controlling its development in PTC. We performed a comprehensive bioinformatics analysis using several datasets. Differential expression analysis and weighted gene coexpression network analysis (WGCNA) on 58 paired PTC samples from The Cancer Genome Atlas (TCGA) were used to detect ETE-related mRNA and long noncoding (lnc) RNA modules and construct an lncRNA/mRNA network. An independent TCGA dataset containing 438 samples was utilized to validate and characterize the WGCNA results. Functional annotation was used to identify the biological functions and related pathways of ETE modules. Two independent RNA sequencing datasets were combined to crossvalidate relationships between lncRNAs and mRNAs by Pearson correlation analysis. Transcription factors (TFs) for affected genes were predicted using the binding motif data from Ensembl Biomart to construct a TF/lncRNA/mRNA network. Other two independent datasets were used to crossvalidate TF-mRNA associations. Finally, receiver operating characteristic, survival analyses, and Cox proportional hazard regression model were performed to explore the significance of hub genes in ETE diagnosis and PTC prognosis. Three mRNA modules and two lncRNA modules were significantly associated with ETE. Enrichment analysis showed extracellular matrix changes was closely related to the development of ETE. A TF/lncRNA/mRNA regulatory network was constructed containing 33 validated hub genes, 64 lncRNAs, and 64 TFs, all differentially expressed between ETE and non-ETE samples. Unc-5 family C-terminal like [area under the curve (AUC): 0.711], sushi repeat containing protein X-linked 2 (AUC: 0.706), lysyl oxidase (AUC: 0.704), collagen type I alpha 1 chain (AUC: 0.704), and collagen type X alpha 1 chain (AUC: 0.704) were the most highly significant hub genes for ETE diagnosis. The Cox proportional hazard regression model constructed with hub genes showed significant survival differences between low-and high-risk groups (p = 0.00025) and performed good prediction for PTC prognosis(AUC = 0.794;

Long Non-Coding RNAs as Determinants of Thyroid Cancer Phenotypes: Investigating Differential Gene Expression Patterns and Novel Biomarker Discovery

“…Of note, authors reported no significant change in expression of TINCR in PTC vs. normal tissue, contradicting previous reporting of its upregulation in the same disease [85]. This displays the complexity of lncRNA expression, as well as the varying consequences of their dysregulated expression patterns [86].…”

Hub gene identification and prognostic model construction for isocitrate dehydrogenase mutation in glioma

“…Then, the modules whose eigengenes showed correlation above 0.75 (cut line for merging of modules = 0.25) were merged (Y. [5] ), and ten modules were finally obtained. The eigengene adjacency heatmap was shown in Fig.…”

Comprehensive Analysis of Prognostic Alternative Splicing Signature Reveals Recurrence Predictor for Papillary Thyroid Cancer

“…By using TCGA data, reported studies involved to PTC prognostic have shown that long noncoding RNA (lncRNAs), microRNA, and methylation data can work as prognostic factors. Chen et al, using binding motif data from Ensembl Biomart, predicted transcription factors (TFs) for affected genes to construct a TF/lncRNA/mRNA network, which predicted PTC prognosis with an AUC of 0.794 (28). Wang et al established an N6-methyladenosine (m6A) RNA methylation-related risk signature of disease-free survival for a total PTC cohort with an AUC of 0.817.…”