8:50-9:10 The preeclampsia gene STOX1 locally affects genome methylation in trophoblast cells

STOX1 variants were posited as major determinants of risks of familial forms of preeclampsia in 2005 [1]. Since then, it has been demonstrated that overexpression of its complete isoform, STOX1A, produces, in trophoblast cell lines such as JEG-3 and BeWo, a gene expression profiles having a lot in common with the one of the preeclamptic placenta [2, 3]. Interestingly the utero placental expression of this isoform results in a preeclamptic syndrome in mice [4], with a Fetal Growth Restriction phenotype in the litter [5], with a major alteration of the oxidative/nitrosative stress balance [6]. Recently, we identified the DNA binding site of STOX1, STRE1 [3]. EPIC arrays interrogating 816757 CpGs were used to pinpoint possible methylation alterations due to STOX1 overexpression in JEG-3 cells (two overexpressing cell lines versus three controls cell lines). The most differentially methylated genes were identified, as well as the abnormally methylated regions (DMR) nearby. Then a bioinformatic search was performed using the MEME suite to search for a possible STRE1 site in the vicinity. To better characterize the mechanistical impact of STOX1 in terms of methylation, an approach of invalidation of the TFBS STRE1 was initiated using CrispR-cas9 genome editing, with a protocol derived from the ouabain selection method [7]. Locus-specific methylation analysis was performed by a HpaII-qPCR procedure and by cloning sequencing of bisulphite treated DNA. 17494 CpG had a p value below 0.05, 15059 (86,1%) were with a decreased methylation while 2435 (13.9%) presented with an increased methylation. Since we did not have the power of defining individual significant CpG after correction for multiple testing, we focused on DMRs that are sustained by a series of consecutive modified CpG. Concentrating on coding genes, we found 1992 which were different from more than one CpG, thus considered as harboring a DMR. Among the up-methylated genes, the highest modification was discovered in FAD-Dependent Oxidoreductase domain-containing protein 2 (FOXRED2). In its promoter there is a CpG island covering about 2000 base pairs of DNA. Inside this sequence a specific region of ~450 bp has its methylation increased by ~60-70% following STOX1A overexpression. In this specific case, microarray analysis revealed a down regulation ~30 fold of the gene, largely confirmed by qRT-PCR. Having found a STRE1 binding site ~400 bp upstream of the DMR, we surmised that STOX1 was acting through this sequence. Therefore, we conducted a CrispR-Cas9 experiment to invalidate this sequence. We obtained 58 clones, 10 of which were chosen after bulk sequencing for having a STRE1 importantly modified. We showed that in these clones the methylation was back to the level of cells that do not express STOX1 and studied the expression alterations in the cell lines. We show for the first time how a transcription factor is able to act specifically locally to modify epigenetic marks (i.e. DNA methylation). A lot remains to be done to find STOX1 partners that are operative in this specific modification. The function of FOXRED2 in the context of preeclampsia or more generally oxidative stress management, remains to be elucidated. van Dijk, M., et al., Nat Genet, 2005. 37(5): p. 514-9. Rigourd, V., et al., PLoS One, 2008. 3(12): p. e3905. Ducat, A., et al., iScience, 2020. 23(5): p. 101086. Doridot, L., et al., Hypertension, 2013. 61(3): p. 662-8. Collinot, H., et al., J Hypertens, 2018. 36(6): p. 1399-1406. Doridot, L., et al., Antioxid Redox Signal, 2014. 21(6): p. 819-34. Agudelo, D., et al., Nat Methods, 2017. 14(6): p. 615-620.