Serotonin Transporter-dependent Histone Serotonylation in Placenta Contributes to the Neurodevelopmental Transcriptome

Abstract

Brain development requires appropriate regulation of serotonin (5-HT) signaling from distinct tissue sources across embryogenesis. At the maternal-fetal interface, the placenta is thought to be an important contributor of offspring brain 5-HT and is critical to overall fetal health. Yet, how placental 5-HT is acquired, and the mechanisms through which 5-HT influences placental functions, are not well understood. Recently, our group identified a novel epigenetic role for 5-HT, in which 5-HT can be added to histone proteins to regulate transcription, a process called H3 serotonylation. Here, we show that H3 serotonylation undergoes dynamic regulation during placental development, corresponding to gene expression changes that are known to influence key metabolic processes. Using transgenic mice, we demonstrate that placental H3 serotonylation is dependent on 5-HT uptake by the serotonin transporter (SERT/SLC6A4). SERT deletion robustly reduces enrichment of H3 serotonylation across the placental genome, and disrupts neurodevelopmental gene networks in early embryonic brain tissues. Thus, these findings suggest a novel role for H3 serotonylation in coordinating placental transcription at the intersection of maternal physiology and offspring brain development.

Keywords: 5 total: epigenetics; H3 serotonylation; development; placenta; serotonin transporter.

Figure 1.. H3 serotonylation is associated with developmental gene networks in male and female placenta

(A) Schematic depicting brain 5-HT levels and tissue of origin, adapted from Suri et al. (B) Placental 5-HT levels decrease from E9.5 to E17.5 (unpaired Student’s t-test, t(13) = 3.209, **p = 0.0068), with male and female placental samples clustering together, as noted by circle colors (N=7-8 samples/age). (C) Western blot analysis of H3K4me3Q5ser in male and female placenta tissues at E9.5, E12.5, E14.5 and E17.5 showed a main effect of embryonic age (two-way ANOVA, age F(3,47) = 6.622, p = 0.0008) with no significant effect of sex (F(1,47) = 3.586, p = 0.0644), where histone serotonylation decreased over development (Sidak’s post-hoc test, E9.5 vs E14.5 (*adjusted p = 0.0102); E9.5 vs E17.5 (**adjusted p = 0.0056); E12.5 vs E14.5 (adjusted p = 0.057), E12.5 vs E17.5 (adjusted p = 0.0356), N = 6-8/group). (B, C): Data are normalized to the male E9.5 values and shown as mean ± SEM. (D) Averaged proportion of peaks using annotations from all developmental male and female placentas showed about 68.1% of sites found following H3K4me3Q5ser ChIP-sequencing were located in promoter regions (N = 4 samples/age/sex). (E) There was a ~tenfold greater number of significantly differential peaks comparing E9.5 vs E17.5 in both males and females, compared to sex difference contrasts within embryonic age (p < 0.05, log₂(fold change) > 0.1). (F, G) Heatmaps (F) and profiles (G) of differential peaks from E9.5 vs E17.5 comparisons, separated by directionality and centered on genomic regions to show the majority of altered peaks decrease across placental development. (H) Venn diagram depicting the degree of overlap between male and female E9.5 vs E17.5 comparisons using uniquely annotated peaks, indicating developmental changes are largely conserved between sex. (I) Odds ratio analysis of differential H3K4me3Q5ser peaks (from 1E above) and differentially expressed genes (adjusted p < 0.05; N = 4 samples/age/sex) from E9.5 vs E17.5 comparisons show significant association between altered histone serotonylation regulation and gene expression changes. Insert numbers indicate respective p values for each association (N.S., p > 0.05). (J) Representative genome browser tracks of Hoxa13 and Cxcl1 loci for H3K4me3Q5ser (vs respective DNA input) in E9.5, E12.5 and E17.5 male and female placentas (Hoxa13: **** p < 0.0001 relative to E9.5 within sex; Cxcl1: ***p < 0.001, **p < 0.01 relative to E9.5 within sex; *p < 0.05 denotes significant changes in E12.5 vs E17.5 males and E9.5 vs E12.5 females) Each track represents merged signal for 4 samples. (K) Selected Reactome and GO Biological Process pathways for differential peaks displaying significant associations with gene expression between E9.5 vs E17.5 (from 1I above) for male placenta tissues (FDR < 0.05).

Figure 2.. Placental 5-HT is dependent on SERT expression

(A) Schematic depicting potential modes of placental 5-HT acquisition examined in this study. (B, C) Normalized counts indicating Sert (Slc6a4) and Oct3 (Slc22a3) are expressed in both male and female placental tissues at E12.5, with no differences by sex (unpaired Student’s t-test; Slc6a4: p = 0.3677; Slc22a3: p = 0.5973). (D) The Tph1 gene is not expressed in E12.5 placental tissues. N=4 samples/sex. Data are median ± interquartile range. (E) Assessment of 5-HT levels in E12.5 placental tissues shows significant reductions (one-way ANOVA, F(3,8) = 4.001, p = 0.0004) in Sert KO (Dunnett’s multiple comparisons test; ***adjusted p = 0.0003), Tph1 KO (**adjusted p = 0.0015), and Oct3 KO (*adjusted p = 0.04) tissues. N=3-5/group. (F) Western blot analysis of placental tissues at E12.5, showing reduced H3K4me3Q5ser in Sert KO, Tph1 KO and Oct3 KO tissues (one-way ANOVA, F(3,10) = 15.37, *p = 0.05). Peptide competition assays using H3_1-10 peptides show selective signal of the serotonyl-PTM epitope is predominantly observed in WT placenta. N = 3/group. Data are mean ± SEM.

Figure 3.. SERT deletion alters placental H3 serotonylation patterning

(A) Relative to WT, the greatest number of significantly decreased H3K4me3Q5ser peaks was observed in Sert KO placentas, followed by Tph1 KO and Oct3 KO (left; p < 0.05, log₂(fold change) > 0.1), where the overall pattern of differential sites diverged from those of H3K4me3 alone (right; N = 3 samples/group). (B) Scatter plots of differential H3K4me3Q5ser (left) and H3K4me3 (right) peaks in Sert KO placentas relative to WT, showing the majority of affected sites are downregulated. (C) Odds ratio analysis examining overlap of significantly reduced H3K4me3Q5ser and H3K4me3 peaks (relative to WT, from 3A) with differential H3K4me3Q5ser sites between E9.5 and E17.5 (from 1E), with bubble size representing number of overlapping loci, indicating SERT deletion has greatest impact on developmentally-regulated sites. Insert numbers denote respective p values for each association (NS, p > 0.05), (D, E) Heatmaps (D) and profiles (E) of differential H3K4me3Q5ser loci between E9.5 and E17.5 that are significantly downregulated in Sert KO placentas, separated by directional changes across development and centered on genomic features. (F) Representative genome browser tracks of Hoxa13 and Cxcl1 loci for H3K4me3Q5ser and H3K4me3 (vs respective DNA input) in WT, Sert KO, Tph1 KO and Oct3 KO placentas (Hoxa13: *p < 0.05 relative to WT; Cxcl1: **p < 0.01, *p < 0.05 relative to WT for each histone modification). Each track represents merged signal for 3 samples. (G) Selected Reactome and GO Biological Process pathways for differential loci (vs WT) overlapping with developmentally regulated H3K4me3Q5ser sites (from 1H). Note: there were no significant pathways enriched for overlapping differential peaks from WT vs Oct3 KO comparisons (FDR < 0.05).

Figure 4.. Offspring neurodevelopmental gene expression changes are associated with placental disruptions

(A) Schematics of study design for investigating E12.5 offspring brain changes. (B) Normalized counts showing gene expression for Tph1, Slc6a4, and Slc22a3 are low compared to that for the transporter PMAT (Slc29a4) in embryonic brain. (C) There is no change in 5-HT levels in E12.5 brains when comparing WT vs KO tissues (one-way ANOVA, F(3,14) = 0.027, p = 0.9938). N=4-5 samples/group. (D) There also are no differences in H3K4me3Q5ser in brain tissues (one-way ANOVA, F(3,16) = 0.5861, p = 0.6328). N=5 samples/group. Data are mean ± SEM. (E) Number of differentially expressed genes from bulk RNA-sequencing comparing WT vs. Sert KO, WT vs. Tph1 KO, WT vs. Oct3 KO brain tissues at E12.5 (adjusted p < 0.05). (F) Hierarchical clustering of all differentially expressed genes relative to WT (adjusted p < 0.05). Expression values are averaged within genotype (N=5-6 samples/group). (G) Selected Reactome and GO Biological Process pathways enriched from differentially expressed genes comparing WT vs KO brain tissues at E12.5 (FDR < 0.05).