Endocrine Abstracts

Objectives: Somatotropinomas can be divided into three subgroups based on their distinct DNA methylation profiles¹, one matching sparsely granulated (SG) and the other two matching densely granulated phenotypes (DG-A and DG-B). Sparsely granulated adenomas show fibrous body formation on cytokeratin immunohistochemistry, compared to diffuse staining in densely granulated adenomas. Methylation¹ and gene expression data were analysed to identify (i) differentially methylated regions (DMRs) and genes between SG, DG-A and DG-B and normal pituitary (NP), and (ii) genes showing differential promoter methylation with differential expression.

Methods: We identified.

• DMRs using bumphunter (minimum probes/ DMR=7, adjusted P<0.05),

• differentially expressed genes from Affymetrix data (false discovery rate <0.05, fold change of ≥2),

• gene sets using gene set enrichment analysis (Fisher’s exact test (P<0.05)),

• interaction hotspots using EpiMod (P<0.05).

Results: All tumour subgroups showed predominant hypomethylation compared to normal, with striking profound hypomethylation observed in DG-B. Three genes showed promoter hypermethylation with decreased RNA expression, including IER3 (SG vs.NP), with a known pro-apoptotic role in cervical cancer. Twenty-one genes showed promoter hypomethylation with increased RNA expression, including TET1 (DG-B vs. NP), which is a demethylating agent and may cause the profound hypomethylation in DG-B tumours. ITPR2 and SFRP1 showed promoter hypermethylation with decreased expression in SG vs. DG-B tumours. ITPR2 mediates oncogene-induced senescence and SFRP1 inhibits Wnt signalling, both consistent with tumour suppressive roles. Subnetworks based around inflammatory mediators show predominant promoter hypo-methylation (DG-B vs. DG-A/ DG-B vsSG), suggesting a key role for inflammatory mediators in DG-B tumours.

Conclusions: Pathways involving apoptosis, demethylation and inflammation mediate different tumorigenic pathways in pituitary adenomas with different cytokeratin-staining patterns on immunohistochemistry.

Reference: 1. Capper, D. et al. DNA methylation-based classification of central nervous system tumours. Nature 555, 469–474, doi:10.1038/nature26000 (2018).