Endocrine Abstracts

Objectives: Hürthle cell (oncocytic) lesions can be metaplastic or neoplastic events. The neoplastic entities -formerly, oncocytic variant of follicular tumours; Hürthle cell neoplasm (HCN) - include HC adenomas and HC carcinomas and have been recognized as a separate class of tumours by the WHO. The remaining of the thyroid tumours demonstrating “mitochondrion-rich cells” and oncocytic morphology are referred under the umbrella term of “oncocytic variant”. However, this rigid separation between the two groups of tumors, characterized by the presence of a mitochondrial-rich cytoplasm, is not clarified, and is not universally accepted. This brings the question if the “oncocytic cells” are different from “mitochondrion-rich cells”, or if they represent a sort of continuum? Thus, we conducted a comparative transcriptomics, proteomic and CNV analysis in the group of HCNs vs mitochondrial-rich non-Hürthle thyroid tumours (other thyroid tumours demonstrating oncocytic morphology).

Methods: Eighteen thyroid tissue samples obtained from the biobank of the pathology department. The cohort of 12 HCN (7 HCC, 5 HCA) and 6 non-HCNs (2 OV-PTC, 3 WT-UMP, 1 NIFTP with oncocytic morphology). We have compared the groups HCNs vs non-HCNs. RNA was sequenced using Ion AmpliSeq Transcriptome Humane Gene Expression Kit and analysed using Transcriptome Analysis Console Software. For proteomic analysis, liquid Chromatography Analysis/Mass Spectrometry were done by using the software Proteome Discoverer 2.4.0.305. Shallow whole-genome sequencing of DNA samples was performed using Illumina platform. BWA was used to map sequencing reads to human reference, Them QDNAseq was used to access the CNV among the genome regions. Enrichment analyses was performed for RNA, proteomic and CNV, with package enrichGO and with annotation of the org.Hs.eg.db packge. For all plots, R and the package “ggplot2” were used. Statistical analyses were performed also using R, in particular the nonparametric Wilcoxon rank-sum test.

Results: The mitochondrial translation system related pathways for RNA sequencing, extracellular matrix related, metabolic process and Immune response related pathways, for proteomics and plasticity related pathways for CNV analysis were found to be significantly increased in the HCN. HCN display, a higher number of gains (chromosomes 2, 4 and 12, P<0.0001−0.004) and lower number of losses (chromosome 3, P<0.001−0.004) when compared with the non-Hürthle group.

Conclusions: Our results revealed some significant differences between the two groups in the omics platform and in the CNV analyses. Our results suggest that HCNs vs mitochondrial-rich non-Hürthle thyroid tumours may be just "different phases" of a spectrum that ends with the Hürthle cell phenotype.