ECE2020 Audio ePoster Presentations Pituitary and Neuroendocrinology (217 abstracts)
Hypercortisolism-related molecular signature: Results from whole blood methylome analysis
Roberta Armignacco 1 , Amandine Septier 1 , Anne Jouinot 1,2 , Mario Neou 1 , Cassandra Gaspar 3 , Karine Perlemoine 1 , Lucas Bouys 1 , Leah Braun 4 , Anna Riester 4 , Maria-Christina Zennaro 5,6 , Martin Reincke 4 , Jérôme Bertherat 1,7 , Felix Beuschlein 4,8 & Guillaume Assié 1,7
1Université de Paris, Institut Cochin, INSERM, CNRS, Paris, France; 2Medical Oncology, Assistance Publique Hôpitaux de Paris, Hôpital Cochin, Paris, France; 3Sorbonne Université, Inserm, UMS Pass, Plateforme Post-génomique de la Pitié-Salpêtrière, P3S, Paris, France; 4Department for Endocrinology, Medizinische Klinik und Poliklinik IV, Ludwig-Maximilians-University, Munich, Germany; 5Université de Paris, PARCC, INSERM, Paris, France; 6Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Service de Génétique, Paris, France; 7Department of Endocrinology, Center for Rare Adrenal Diseases, Assistance Publique-Hôpitaux de Paris, Hôpital Cochin, Paris, France; 8Klinik für Endokrinologie, Diabetologie und Klinische Ernährung, Universitätsspital Zürich, Zurich, Switzerland
The effective treatment and optimal prognosis of hypercortisolism (Cushing’s syndrome – CS) depend on accurate and early diagnosis. However, hormonal assays can be complex, requiring multiple tests, and not predictive for any related complications, neither for their duration and severity. Identifying novel, specific and easily measurable biomarkers may improve CS diagnosis as well as the evaluation of complications. Since stress-associated epigenetic markers can be measured at blood level, we analyzed the whole blood methylome of patients, before and after hypercortisolism cure, in order to identify a specific methylation signature of cortisol excess. We collected paired blood samples of 47 patients with confirmed hypercortisolism, obtained before (Pre) and several months after (Post) treatment, and we extracted leucocyte DNA. Methylome data were generated by the Infinium MethylationEPICBeadChip (850K probes; Illumina), and pre-processed and pre-analyzed by adapting different packages (minfi, ChAMP) specifically developed for methylation array analysis for the R software. The entire set of probes was analyzed by both unsupervised and supervised approaches in a training sub-cohort of 48 paired samples (Pre/Post treatment) from 24 patients. The results were then tested on the rest of samples, used as validation cohort. Unsupervised clustering, based on the most variable features, showed a distribution of samples in pairs, each corresponding to an individual, thus expectedly accounting for the highest source of variability among samples. Interestingly, all sample pairs showed a common group of CpGs differentially methylated in the Pre compared to the Post condition, thus indicating the presence of a specific cortisol-related methylation signature. Consistently, the projection of the two most representative components of variability (PCA analysis) allowed to well discriminate Pre samples from Post samples. A supervised pairwise comparison of the samples, performed taking into account some meaningful covariates (particularly blood cell composition), allowed to identify 12 significant CpGs (adjusted P-value < 0.05) perfectly discriminating the Pre from the Post samples. Hierarchical clustering performed on the validation cohort using the same 12 selected CpGs allowed to classify samples according to their status of hypercortisolism with an accuracy of 0.81. The prediction power of the model raised to 0.97 (where 1 corresponds to perfect prediction) when we calculated the C statistic after having performed logistic regression on the validation cohort. Our results show that a specific hypercortisolism-related epigenetic signature exists and that it is measurable at the whole blood level. This approach promises to be powerful to identify specific biomarkers of cortisol excess and of its related complications.
Article tools
My recent searches
My recently viewed abstracts
Authors
Endocrine Abstracts
ISSN 1470-3947 (print) | ISSN 1479-6848 (online)
© Bioscientifica 2024 |
Privacy policy |
Cookie settings
BiosciAbstracts
Bioscientifica Abstracts is the gateway to a series of products that provide a permanent, citable record of abstracts for biomedical and life science conferences.
Find out more