Endocrine Abstracts

JOINT2387

Background: Glucocorticoids (GC) modulate the epigenetic machinery, thereby influencing gene expression. In cases of Cushing’s disease (CD) in remission, the persistence of some comorbidities, such as hypertension, obesity, and diabetes, may be related to epigenetic changes. Our objective was to evaluate a) the reversibility of GC-induced changes in DNA methylation (DNAm) in patients with CD after at least one year of disease remission and b) the impact of DNAm on the persistence of comorbidities.

Methods: Forty-six patients diagnosed with CD and 59 BMI and gender-matched control subjects were included. Blood samples were collected during the active phase of CD (active_CD) (n=40) and at least one year of remission (remission_CD)(n=30, mean 49.76±20.02 months), including 24 paired samples. Methylation levels of leukocyte-derived DNA were assessed using the Illumina Infinium Human Methylation 850K array. A difference in beta values of 0.05 with a Benjamini-Hochberg adjusted p-value of<0.05 was used to define differentially methylated probes/CpGs (DMPs).

Results: A total of 4264 DMPs were detected between control and active_CD (hypomethylated n=2195 and hypermethylated n=1349), 184 in control vs remission_CD (hypomethylated n=150 and hypermethylated n=34), and 2384 in active_CD vs remission_CD (hypomethylated n=453 and hypermethylated n=1931). Unsupervised hierarchical clustering revealed 3 main clusters; Cluster 1 comprised almost entirely of controls, the adjacent Cluster 2 included predominantly remission_CD, whereas most active_CD clustered in the distal Cluster 3. Analysis of DMPs associated with gene promoter regions yielded similar results. These results suggest that differences in methylation persist between the control and remission groups despite disease remission. Gene set enrichment analysis (GSEA) using DMPs revealed enrichment of cellular metabolic processes by comparing active CD or remission with controls, whereas immune processes were primarily enriched by comparing active CD and remission. A sub-analysis of 24 patients with paired samples from active disease and remission showed similar results. No association was observed between DMPs and comorbidities, such as diabetes, hypertension, obesity, and dyslipidemia. Using a random forest model, we could classify control, active, and remission samples in our cohort with an overall error rate of 14% (%; controls=0, active_CD=23.5, remission_CD=22).

Conclusion: Cortisol excess in active CD is associated with a spectrum of methylation changes in leukocyte-derived DNA, which are not completely reversible after long-term remission. This indicates that cortisol has sustained effects on the epigenetic machinery. Future studies should explore prolonged epigenetic changes in other cell types in CD and examine their relation to persistent morbidities.