Rationale: Peri-conceptional folate status is essential for offsprings' brain structure and function. We hypothesize that maternal genetic variations in folate metabolism during pregnancy may contribute to the inter-individual variability of adverse cognitive outcome seen in offspring of women with epilepsy (WWE). Methods: WWE in this cohort were part of the prospecitve Maternal Outcomes and Neurodevelopmental Effects of Antiepileptic Drugs (MONEAD) study. Their children were assessed at two years old via the Bayley Scales of Infant and Toddler Development-3 (BSID-III) yielding 5 Scale scores (Language, Motor, Cognitive, Social-Emotional, Adaptive Behavior). Other clinical features assessed included WWE’s IQ, WWE’s education, maximum observed ASM blood levels and maximum % defined daily dose in the third trimester.Samples underwent whole exome sequencing (WES) with collaboration of Epi25 at the Broad Institute of Harvard on the Illumina HiSeq X platform. To guarantee high quality of the sequencing data, we performed several quality control (QC) procedures: 1) Variants were excluded if the p value < 1e−6 in Hardy-Weinberg equilibrium test; 2) Call rate < 90%, occurring as multi-allelic SNVs or appearing in other than ”PASS” from the variant quality score recalibration in GATK metric; 3) Samples excluded if their distribution of singleton calls, heterozygous calls or inbreeding coefficient resided more than three standard deviation from the mean, or had >1% missing calls; 4) Identify-by-state was computed and in each pair of samples that were similar, only one of them were kept for further analysis; 5) Population admixture analysis was performed using ADMIXTURE with reference of 1000 Genomes dataset (Phase 3) and samples with < 70% European ancestry were excluded.
Different folate pathway gene panels including folate synthesis pathway genes (26 genes), folate one carbon pathway genes (53 genes), and folate metabolism pathway genes (69 genes), were extracted from genecard.com. Rare variants in genes of interest were identified with occurrence < 1% in the European populations in gnomAD v2. Results: In these 118 WES samples of WWE, 111 passed QC and 72 were deemed as European ancestry. Among these 72 European ancestry WWE, 56 children completed BSID-III at two years old and were used for downstream analysis. Total rare variant counts in folate biosynthesis or folate one carbone pathway were not significantly associated with any of the 5 BSID-III scores, but total rare variant counts in folate metabolism gene panels were significantly associated with language (r2 = -0.317, p=0.017) and cognitive (r2 = -0.274, p=0.041) domains of the BSID-III. In addition, all clinical characteristics were also assessed and variables with statistical significance were used for further regression analysis. In adjusted analysis including WWE’s IQ and education background, rare genetic variant counts in the folate metabolism pathway remained as an independent factor to predict the poorer language (-2.9; 95% CIs= -7.9, -1.5, p=0.005) and cognitive (-2.3; 95% CIs= -5.0, -0.32, p=0.027) scores in BSID-III. Conclusions: Increased burden of rare genetic variations in folate metabolism in European ancestry WWE is associated with poorer neuropsychological outcomes of their children at two years old for language and cognitive domains. The findings in this cohort should be replicated at older ages, where neuropsychological testing is more sensitive. In addition, future studies using the children’s genes and replication in other cohorts are needed. Funding: Please list any funding that was received in support of this abstract.: NIH, NINDS and NICHD #U01-NS038455