Endocrine Abstracts

Objectives: Somatotropinomas form three subgroups with distinct methylation profiles¹, one matching the sparsely granulated (SG) and the other two matching the densely granulated phenotypes (DG-A and DG-B). Publicly accessible raw methylation data¹ and our RNA microarray data were analysed to identify differentially methylated regions (DMRs) and enriched genes between SG, DG-A and DG-B and normal anterior pituitary gland (NP), genes showing differential methylation of promoters with differential expression.

Methods: DMRs were identified using bumphunter (minimum number of probes/DMRs=7, adjusted P<0.05, false discovery rate (FDR) <0.05) with gene set enrichment analysis (GSEA) using Fisher’s exact test (P<0.05). Interaction hotspots were identified using EpiMod (P<0.05). Differentially expressed genes were identified using Affymetrix data with FDR <0.05 and fold change of ≥2.

Results: 3 distinct methylation profiles were noted (Figure 1). 3 genes (SG vs NP=2, DG-B vs NP=1) showed hyper-methylation of promoter with decreased expression. 21 genes (SG vs NP=1, DG-A vs NP=5, DG-B vs NP=15) showed hypo-methylation of promoter with increased expression. Genes of interest included IER3 (SG vs NP, methylation value 1.12, log-fold change=−5.28108) and TET1 (DG-B vs NP, methylation value =−1.13, log-fold change=3.09). One gene set (BRIDEAU_IMPRINTED_GENES) and 6 interaction hotspots (seeds mainly involving inflammatory mediators, Figure 2) were common across all analyses comparing tumour subgroups to normal anterior pituitary gland.

Conclusions: Pathways involving apoptosis (IER3), demethylation (TET1), inflammation and degradation of extracellular matrix appear central to pituitary tumorigenesis. Further analysis and validation is in progress.

Figure 1 Heatmap of methylation values across tumour subgroups and normal pituitary.

Figure 2 Interaction hotspots with (A) CCL11 and (B) MMP3 as seeds.