Altered DNA methylation profiles in blood from patients with sporadic Creutzfeldt-Jakob disease.
Luke Child DabinFernando GuntoroTracy CampbellTony BélicardAdam R SmithRebecca G SmithRachel RaybouldJonathan M SchottKatie LunnonPeter SarkiesJohn CollingeSimon MeadEmmanuelle ViréPublished in: Acta neuropathologica (2020)
Prion diseases are fatal and transmissible neurodegenerative disorders caused by the misfolding and aggregation of prion protein. Although recent studies have implicated epigenetic variation in common neurodegenerative disorders, no study has yet explored their role in human prion diseases. Here we profiled genome-wide blood DNA methylation in the most common human prion disease, sporadic Creutzfeldt-Jakob disease (sCJD). Our case-control study (n = 219), when accounting for differences in cell type composition between individuals, identified 38 probes at genome-wide significance (p < 1.24 × 10-7). Nine of these sites were taken forward in a replication study, performed in an independent case-control (n = 186) cohort using pyrosequencing. Sites in or close to FKBP5, AIM2 (2 probes), UHRF1, KCNAB2 successfully replicated. The blood-based DNA methylation signal was tissue- and disease-specific, in that the replicated probe signals were unchanged in case-control studies using sCJD frontal-cortex (n = 84), blood samples from patients with Alzheimer's disease, and from inherited and acquired prion diseases. Machine learning algorithms using blood DNA methylation array profiles accurately distinguished sCJD patients and controls. Finally, we identified sites whose methylation levels associated with prolonged survival in sCJD patients. Altogether, this study has identified a peripheral DNA methylation signature of sCJD with a variety of potential biomarker applications.
Keyphrases
- dna methylation
- genome wide
- case control
- gene expression
- machine learning
- end stage renal disease
- endothelial cells
- copy number
- ejection fraction
- newly diagnosed
- small molecule
- prognostic factors
- chronic kidney disease
- high throughput
- artificial intelligence
- deep learning
- mass spectrometry
- single molecule
- induced pluripotent stem cells
- patient reported outcomes
- big data
- chemotherapy induced
- binding protein