RESUMO
BACKGROUND: Familial hypercholesterolemia (FH) is a monogenic disorder characterized by elevated low-density lipoprotein cholesterol (LDL-C). A FH causing genetic variant in LDLR, APOB, or PCSK9 is not identified in 12-60% of clinical FH patients (FH mutation-negative patients). We aimed to assess whether altered DNA methylation might be associated with FH in this latter group. METHODS: In this study we included 78 FH mutation-negative patients and 58 FH mutation-positive patients with a pathogenic LDLR variant. All patients were male, not using lipid lowering therapies and had LDL-C levels >6 mmol/L and triglyceride levels <3.5 mmol/L. DNA methylation was measured with the Infinium Methylation EPIC 850 K beadchip assay. Multiple linear regression analyses were used to explore DNA methylation differences between the two groups in genes related to lipid metabolism. A gradient boosting machine learning model was applied to investigate accumulated genome-wide differences between the two groups. FINDINGS: Candidate gene analysis revealed one significantly hypomethylated CpG site in CPT1A (cg00574958) in FH mutation-negative patients, while no differences in methylation in other lipid genes were observed. The machine learning model did distinguish the two groups with a mean Area Under the Curve (AUC)±SD of 0.80±0.17 and provided two CpG sites (cg26426080 and cg11478607) in genes with a possible link to lipid metabolism (PRDM16 and GSTT1). INTERPRETATION: FH mutation-negative patients are characterized by accumulated genome wide DNA methylation differences, but not by major DNA methylation alterations in known lipid genes compared to FH mutation-positive patients. FUNDING: ZonMW grant (VIDI no. 016.156.445).