Identification of Key Functional Gene Signatures Indicative of Dedifferentiation in Papillary Thyroid Cancer

Abstract

Background: Differentiated thyroid cancer (DTC) is the most common type of thyroid cancer. Many of them can relapse to dedifferentiated thyroid cancer (DDTC) and exhibit different gene expression profiles. The underlying mechanism of dedifferentiation and the involved genes or pathways remained to be investigated.Methods: A discovery cohort obtained from patients who received surgical resection in the Fudan University Shanghai Cancer Center (FUSCC) and two validation cohorts derived from Gene Expression Omnibus (GEO) database were used to screen out differentially expressed genes in the dedifferentiation process. Weighted gene co-expression network analysis (WGCNA) was constructed to identify modules highly related to differentiation. Gene Set Enrichment Analysis (GSEA) was used to identify pathways related to differentiation, and all differentially expressed genes were grouped by function based on the GSEA and literature reviewing data. Least absolute shrinkage and selection operator (LASSO) regression analysis was used to control the number of variables in each group. Next, we used logistic regression to build a gene signature in each group to indicate differentiation status, and we computed receiver operating characteristic (ROC) curve to evaluate the indicative performance of each signature.Results: A total of 307 upregulated and 313 downregulated genes in poorly differentiated thyroid cancer (PDTC) compared with papillary thyroid cancer (PTC) and normal thyroid (NT) were screened out in FUSCC cohort and validated in two GEO cohorts. WGCNA of 620 differential genes yielded the seven core genes with the highest correlation with thyroid differentiation score (TDS). Furthermore, 395 genes significantly correlated with TDS in univariate logistic regression analysis were divided into 11 groups. The areas under the ROC curve (AUCs) of the gene signature of group transcription and epigenetic modification, signal and substance transport, extracellular matrix (ECM), and metabolism in the training set [The Cancer Genome Atlas (TCGA) cohort] and validation set (combined GEO cohort) were both >0.75. The gene signature based on group transcription and epigenetic modification, cilia formation and movement, and proliferation can reflect the patient's disease recurrence state.Conclusion: The dedifferentiation of DTC is affected by a variety of mechanisms including many genes. The gene signature of group transcription and epigenetic modification, signal and substance transport, ECM, and metabolism can be used as biomarkers for DDTC.

Homeobox and Polycomb target gene methylation in human solid tumors

“…In thyroid cancer, TGIF1 is upregulated in the papillary subtype, which aligns with our observation of hypomethylation in a DMR in the gene 75 . We show that CERS4 is differentially methylated in thyroid cancer, with evidence from the literature that it is also differentially expressed 76 . Finally, PKNOX1 has been identified as a key component in angiogenesis in thyroid carcinomas 77 .…”

Insight of novel biomarkers for papillary thyroid carcinoma through multiomics

“…Notably, these pathways were predominantly related to the extracellular matrix (ECM), including extracellular matrix organization, collagen-containing extracellular matrix, and extracellular matrix structural constituent ( Figure 2D ). Previous studies have highlighted the dysregulation of ECM in PTC, which is closely associated with tumor activities such as migration and invasion ( 21 ). In addition, the KEGG analysis revealed seven significant pathways, including protein digestion and absorption, cell adhesion molecules, ECM-receptor interaction, malaria, AGE-RAGE signaling pathway in diabetic complications, focal adhesion, and axon guidance.…”

Aberrant expression of PROS1 correlates with human papillary thyroid cancer progression

“…S1. Additionally, P-value < 0.05 and |log2 (fold change) > 1 were considered as the threshold for identifying DEGs (Xu et al, 2021;Yan et al, 2019b;Zhang et al, 2019). For visualization, the Venn diagrams of the five datasets were drawn with the online tool: Venn diagrams (http://bioinformatics.psb.ugent.be/webtools/Venn/), and volcano plots of DEGs were drawn by ggplot2 package in R.…”