NT5E DNA methylation in papillary thyroid cancer: Novel opportunities for precision oncology

• Published in: Molecular and cellular endocrinology 2023-06, Vol.570, p.111915-111915, Article 111915
• Main Authors: Bertoni, Ana Paula Santin, Valandro, Cleiton Felipe, Brasil, Rafael Ávila, Zeiser, Felipe André, Wink, Márcia Rosângela, Furlanetto, Tania Weber, da Costa, Cristiano André
• Format: Article
• Language: English
• Subjects: 5'-Nucleotidase - genetics; Carcinoma, Papillary - genetics; Carcinoma, Papillary - pathology; DNA methylation; DNA Methylation - genetics; GPI-Linked Proteins - genetics; Humans; Machine learning; NT5E/CD73; Precision Medicine; RNA, Messenger - genetics; RNA, Messenger - metabolism; Thyroid cancer; Thyroid Cancer, Papillary - genetics; Thyroid Neoplasms - genetics; Thyroid Neoplasms - pathology
• ISSN: 0303-7207; 1872-8057
• DOI: 10.1016/j.mce.2023.111915

The ectoenzyme CD73, encoded by the NT5E gene, has emerged as a potential prognostic and therapeutic marker for papillary thyroid carcinoma (PTC), which has increased in incidence in recent decades. Here, from The Cancer Genome Atlas Thyroid Cancer (TCGA-THCA) database, we extracted and combined clinical features, levels of NT5E mRNA, and DNA methylation of PTC samples and performed multivariate and random forest analyses to evaluate the prognostic relevance and the potential of discriminating between adjacent non-malignant and thyroid tumor samples. As a result, we revealed that lower levels of methylation at the cg23172664 site were independently associated with BRAF-like phenotype (p = 0.002), age over 55 years (p = 0.012), presence of capsule invasion (p = 0.007) and presence of positive lymph node metastasis (LNM) (p = 0.04). The methylation levels of cg27297263 and cg23172664 sites showed significant and inversely correlations with levels of NT5E mRNA expression (r = −0.528 and r = −0.660, respectively), and their combination was able to discriminate between adjacent non-malignant and tumor samples with a precision of 96%–97% and 84%–85%, respectively. These data suggest that combining cg23172664 and cg27297263 sites may bring new insights to reveal new subsets of patients with papillary thyroid carcinoma. [Display omitted] •PTC with high NT5E expression correlates with hypomethylation.•Low methylation levels at cg23172664 site predicts poor PTC outcome.•This study combines NT5E mRNA and methylation levels in machine learning approach.•Machine learning approach effectively discriminates non-malignant than PTC samples.