Polymorphisms in EPHX2 and CYP2J2 moderate associations between white matter hyperintensities and temporal lobe atrophy

Background White matter hyperintensities (WMH) of presumed vascular origin commonly coincide with neurodegeneration. Previous studies have reported heterogeneous relationships between WMH and atrophy. Cytochrome P450 (CYP) 2J2 (CYP2J2) is involved in the vascular ischemic response and soluble epoxide hydrolase (EPHX2) metabolizes its products, which have been implicated in vascular pathology and WMH. Here we investigate the potential moderation effect of CYP2J2 and EPHX2 single nucleotide polymorphisms (SNPs), individually and interactively, on the association between WMH and atrophy. Methods Patients with vascular cognitive impairment or neurodegenerative diagnoses (n=447) from the Sunnybrook Dementia Study (NCT01800214), Vascular Brain Health (VBH) study, and the Brain‐Eye Amyloid Memory (BEAM) study were genotyped for SNPs in EPHX2 and CYP2J2. SNPs were genotyped by the Illumina Neurochip and were selected a priori based on literature. All participants were confirmed to have European ancestry via multi‐dimensional scaling using the genotype data. Imaging analysis was performed on 3D T1 and T2/PD weighted images for brain segmentation and volumetric acquisition using SABRE, and WMH volumes were acquired on T2/PD/FLAIR images using Lesion Explorer. Linear regression models controlling for age, sex, Mini Mental State Exam, and head size were used to assess moderation effects of the SNPs (WMH×SNP interactions) on associations between WMH and temporal atrophy using SPSS (v.26). Presence of minor alleles were assessed using a dominant genetic model. Results The intron variant EPHX2 rs7816586 G/A polymorphism had a trending moderation effect on the association between periventricular WMH and left temporal atrophy where the minor A‐allele (frequency=11.52%) was protective (F=3.958, p=0.047) in extensive WMH. The promoter variant CYP2J2 rs10889162 C/T had a significant moderation effect on the association between deep WMH and right temporal atrophy where the minor T‐allele (frequency=8.61%) was protective in cases of extensive WMH (F=5.771, p=0.017). No interaction between the EPHX2 and CYP2J2 SNPs was detected. Conclusion We identified potential moderation effects of genetic variation in EPHX2 and CYP2J2 on the association between WMH and temporal lobe atrophy across neurodegenerative diagnoses. The lack of SNP×SNP interactions suggests that these genes may act independently, consistent with WMH×SNP interaction effects that occurred in association with differen