Association between DNA methylation profile and malignancy in follicular-patterned thyroid neoplasms

Classification of Thyroid Tumors Based on DNA Methylation Patterns

Background: Alterations in DNA methylation are stable epigenetic events that can serve as clinical biomarkers. The aim of this study was to analyze methylation patterns among various follicular cell-derived thyroid neoplasms to identify disease subtypes and help understand and classify thyroid tumors. Methods: We employed an unsupervised machine learning method for class discovery to search for distinct methylation patterns among various thyroid neoplasms. Our algorithm was not provided with any clinical or pathological information, relying exclusively on DNA methylation data to classify samples. We analyzed 810 thyroid samples (n = 256 for discovery and n = 554 for validation), including benign and malignant tumors, as well as normal thyroid tissue. Results: Our unsupervised algorithm identified that samples could be classified into three subtypes based solely on their methylation profile. These methylation subtypes were strongly associated with histological diagnosis (p < 0.001) and were therefore named normal-like, follicular-like, and papillary thyroid carcinoma (PTC)-like. Follicular adenomas, follicular carcinomas, oncocytic adenomas, and oncocytic carcinomas clustered together forming the follicular-like methylation subtype. Conversely, classic papillary thyroid carcinomas (cPTC) and tall cell PTC clustered together forming the PTC-like subtype. These methylation subtypes were also strongly associated with genomic drivers: 98.7% BRAFV600E-driven cancers were PTC like, whereas 96.0% RAS-driven cancers had a follicular-like methylation pattern. Interestingly, unlike other diagnoses, follicular variant PTC (FVPTC) samples were split into two methylation clusters (follicular like and PTC like), indicating a heterogeneous group likely to be formed by two distinct diseases. FVPTC samples with a follicular-like methylation pattern were enriched for RAS mutations (36.4% vs. 8.0%; p < 0.001), whereas FVPTC- with PTC-like methylation patterns were enriched for BRAFV600E mutations (52.0% vs. 0%, Fisher exact p = 0.004) and RET fusions (16.0% vs. 0%, Fisher exact p = 0.003). Conclusions: Our data provide novel insights into the epigenetic alterations of thyroid tumors. Since our classification method relies on a fully unsupervised machine learning approach for subtype discovery, our results offer a robust background to support the classification of thyroid neoplasms based on methylation patterns.

Association of DNA Promoter Methylation and BRAF Mutation in Thyroid Cancer

The BRAF V600E mutation and DNA promoter methylation play important roles in the pathogenesis of thyroid cancer (TC). However, the association of these genetic and epigenetic alterations is not clear. In this study, using paired tumor and surrounding normal tissue from the same patients, on a genome-wide scale we tried to identify (a) any association between BRAF mutation and DNA promoter methylation, and (b) if the molecular findings may provide a basis for therapeutic intervention. We included 40 patients with TC (female = 28, male = 12) without distant metastasis. BRAF mutation was present in 18 cases. We identified groups of differentially methylated loci (DML) that are found in (a) both BRAF mutant and wild type, (b) only in BRAF mutant tumors, and (c) only in BRAF wild type. BRAF mutation-specific promoter loci were more frequently hypomethylated, whereas BRAF wild-type-specific loci were more frequently hypermethylated. Common DML were enriched in cancer-related pathways, including the mismatch repair pathway and Wnt-signaling pathway. Wild-type-specific DML were enriched in RAS signaling. Methylation status of checkpoint signaling genes, as well as the T-cell inflamed genes, indicated an opportunity for the potential use of PDL1 inhibitors in BRAF mutant TC. Our study shows an association between BRAF mutation and methylation in TC that may have biological significance.

DNA Methylation Haplotype Block Markers Efficiently Discriminate Follicular Thyroid Carcinoma from Follicular Adenoma

CONTEXT

Follicular thyroid carcinoma (FTC) is the second most common type of thyroid carcinoma and must be pathologically distinguished from benign follicular adenoma (FA). Additionally, the clinical assessment of thyroid tumors with uncertain malignant potential (TT-UMP) demands effective indicators.

OBJECTIVE

We aimed to identify discriminating DNA methylation markers between FA and FTC.

METHODS

DNA methylation patterns were investigated in 33 FTC and 33 FA samples using reduced representation bisulfite sequencing and methylation haplotype block-based analysis. A prediction model was constructed and validated in an independent cohort of 13 FTC and 13 FA samples. Moreover, 36 TT-UMP samples were assessed using this model.

RESULTS

A total of 70 DNA methylation markers, approximately half of which were located within promoters, were identified to be significantly different between the FTC and FA samples. All the Gene Ontology terms enriched among the marker-associated genes were related to "DNA binding," implying that the inactivation of DNA binding played a role in FTC development. A random forest model with an area under the curve of 0.994 was constructed using those markers for discriminating FTC from FA in the validation cohort. When the TT-UMP samples were scored using this model, those with fewer driver mutations also exhibited lower scores.

CONCLUSION

An FTC-predicting model was constructed using DNA methylation markers, which distinguished between FA and FTC tissues with a high degree of accuracy. This model can also be used to help determine the potential of malignancy in TT-UMP.

An Integrative Computational Approach Based on Expression Similarity Signatures to Identify Protein-Protein Interaction Networks in Female-Specific Cancers

Breast, ovarian, and endometrial cancers have a major impact on mortality in women. These tumors share hormone-dependent mechanisms involved in female-specific cancers which support tumor growth in a different manner. Integrated computational approaches may allow us to better detect genomic similarities between these different female-specific cancers, helping us to deliver more sophisticated diagnosis and precise treatments. Recently, several initiatives of The Cancer Genome Atlas (TCGA) have encouraged integrated analyses of multiple cancers rather than individual tumors. These studies revealed common genetic alterations (driver genes) even in clinically distinct entities such as breast, ovarian, and endometrial cancers. In this study, we aimed to identify expression similarity signatures by extracting common genes among TCGA breast (BRCA), ovarian (OV), and uterine corpus endometrial carcinoma (UCEC) cohorts and infer co-regulatory protein-protein interaction networks that might have a relationship with the estrogen signaling pathway. Thus, we carried out an unsupervised principal component analysis (PCA)-based computational approach, using RNA sequencing data of 2,015 female cancer and 148 normal samples, in order to simultaneously capture the data heterogeneity of intertumors. Firstly, we identified tumor-associated genes from gene expression profiles. Secondly, we investigated the signaling pathways and co-regulatory protein-protein interaction networks underlying these three cancers by leveraging the Ingenuity Pathway Analysis software. In detail, we discovered 1,643 expression similarity signatures (638 downregulated and 1,005 upregulated genes, with respect to normal phenotype), denoted as tumor-associated genes. Through functional genomic analyses, we assessed that these genes were involved in the regulation of cell-cycle-dependent mechanisms, including metaphase kinetochore formation and estrogen-dependent S-phase entry. Furthermore, we generated putative co-regulatory protein-protein interaction networks, based on upstream regulators such as the ERBB2/HER2 gene. Moreover, we provided an ad-hoc bioinformatics workflow with a manageable list of intertumor expression similarity signatures for the three female-specific cancers. The expression similarity signatures identified in this study might uncover potential estrogen-dependent molecular mechanisms promoting carcinogenesis.

DNA methylation in thyroid cancer

In recent years, cancer genomics has provided new insights into genetic alterations and signaling pathways involved in thyroid cancer. However, the picture of the molecular landscape is not yet complete. DNA methylation, the most widely studied epigenetic mechanism, is altered in thyroid cancer. Recent technological advances have allowed the identification of novel differentially methylated regions, methylation signatures and potential biomarkers. However, despite recent progress in cataloging methylation alterations in thyroid cancer, many questions remain unanswered. The aim of this review is to comprehensively examine the current knowledge on DNA methylation in thyroid cancer and discuss its potential clinical applications. After providing a general overview of DNA methylation and its dysregulation in cancer, we carefully describe the aberrant methylation changes in thyroid cancer and relate them to methylation patterns, global hypomethylation and gene-specific alterations. We hope this review helps to accelerate the use of the diagnostic, prognostic and therapeutic potential of DNA methylation for the benefit of thyroid cancer patients.