Molecular perspectives in differentiated thyroid cancer

Borealin/CDCA8 deficiency alters thyroid development and results in papillary tumor-like structures

Background

BOREALIN/CDCA8 mutations are associated with congenital hypothyroidism and thyroid dysgenesis. Borealin is involved in mitosis as part of the Chromosomal Passenger Complex. Although BOREALIN mutations decrease thyrocyte adhesion and migration, little is known about the specific role of Borealin in the thyroid.

Methods

We characterized thyroid development and function in Borealin-deficient (Borealin +/-) mice using histology, transcriptomic analysis, and quantitative PCR.

Results

Thyroid development was impaired with a hyperplastic anlage on embryonic day E9.5 followed by thyroid hypoplasia from E11.5 onward. Adult Borealin +/- mice exhibited euthyroid goiter and defect in thyroid hormone synthesis. Borealin +/- aged mice had disorganized follicles and papillary-like structures in thyroids due to ERK pathway activation and a strong increase of Braf-like genes described by The Cancer Genome Atlas (TCGA) network of papillary thyroid carcinoma. Moreover, Borealin +/- thyroids exhibited structural and transcriptomic similarities with papillary thyroid carcinoma tissue from a human patient harboring a BOREALIN mutation, suggesting a role in thyroid tumor susceptibility.

Conclusion

These findings demonstrate Borealin involvement in critical steps of thyroid structural development and function throughout life. They support a role for Borealin in thyroid dysgenesis with congenital hypothyroidism. Close monitoring for thyroid cancer seems warranted in patients carrying BOREALIN mutations.

Prediction of novel target genes and pathways involved in tall cell variant papillary thyroid carcinoma

BACKGROUND

Tall cell variant papillary thyroid carcinoma (TCPTC) is reportedly associated with aggressive clinicopathological parameters and poor outcomes; however, the molecular mechanisms underlying TCPTC remain poorly understood.

METHODS

The gene mutation types and mRNA expression profiles of patients with TCPTC were obtained from The Cancer Genome Atlas (TCGA) database. Differentially expressed genes (DEGs) were identified. Pathways in the interaction network and the diagnostic approaches of candidate markers for TCPTC were investigated.

RESULTS

BRAF mutation was particularly prevalent in TCPTC with a mutation frequency of 78%. TCPTC was associated with a patient age >45 years, tumor multifocality, extrathyroidal extension, a higher T stage, advanced AJCC TNM stages, BRAF V600E mutation, and poor disease-free survival. We identified 4138 TCPTC-related DEGs and 301 TCPTC-specific DEGs. Intriguingly, the gene expression pattern revealed that the dysregulated levels of both putative oncogenes and tumor suppressors in TCPTC were higher than those in classical/conventional variant PTC (cPTC). Functional enrichment analyses revealed that these DEGs were involved in several cancer-related pathways. A protein-protein interaction (PPI) network was constructed from the 301 TCPTC-specific DEGs, and 3 subnetworks, and 8 hub genes were verified. Receiver operating characteristic (ROC) analyses revealed that 6 hub genes, including COL5A1, COL1A1, COL10A1, COL11A1, CCL20, and CXCL5, could be used not only for the differential diagnosis of PTC from normal samples, but also for the differential diagnosis of TCPTC from cPTC samples.

CONCLUSIONS

Our study might provide further insights into the investigations of the tumorigenesis mechanism of TCPTC and assists in the discovery of novel candidate diagnostic markers for TCPTC.

Mut2Vec: distributed representation of cancerous mutations

BACKGROUND

Embedding techniques for converting high-dimensional sparse data into low-dimensional distributed representations have been gaining popularity in various fields of research. In deep learning models, embedding is commonly used and proven to be more effective than naive binary representation. However, yet no attempt has been made to embed highly sparse mutation profiles into densely distributed representations. Since binary representation does not capture biological context, its use is limited in many applications such as discovering novel driver mutations. Additionally, training distributed representations of mutations is challenging due to a relatively small amount of available biological data compared with the large amount of text corpus data in text mining fields.

METHODS

We introduce Mut2Vec, a novel computational pipeline that can be used to create a distributed representation of cancerous mutations. Mut2Vec is trained on cancer profiles using Skip-Gram since cancer can be characterized by a series of co-occurring mutations. We also augmented our pipeline with existing information in the biomedical literature and protein-protein interaction networks to compensate for the data insufficiency.

RESULTS

To evaluate our models, we conducted two experiments that involved the following tasks: a) visualizing driver and passenger mutations, b) identifying novel driver mutations using a clustering method. Our visualization showed a clear distinction between passenger mutations and driver mutations. We also found driver mutation candidates and proved that these were true driver mutations based on our literature survey. The pre-trained mutation vectors and the candidate driver mutations are publicly available at http://infos.korea.ac.kr/mut2vec .

CONCLUSIONS

We introduce Mut2Vec that can be utilized to generate distributed representations of mutations and experimentally validate the efficacy of the generated mutation representations. Mut2Vec can be used in various deep learning applications such as cancer classification and drug sensitivity prediction.