Maternal preconception stress produces sex-specific effects at the maternal:fetal interface to impact offspring development and phenotypic outcomes†

Abstract

Entering pregnancy with a history of adversity, including adverse childhood experiences and racial discrimination stress, is a predictor of negative maternal and fetal health outcomes. Little is known about the biological mechanisms by which preconception adverse experiences are stored and impact future offspring health outcomes. In our maternal preconception stress (MPS) model, female mice underwent chronic stress from postnatal days 28-70 and were mated 2 weeks post-stress. Maternal preconception stress dams blunted the pregnancy-induced shift in the circulating extracellular vesicle proteome and reduced glucose tolerance at mid-gestation, suggesting a shift in pregnancy adaptation. To investigate MPS effects at the maternal:fetal interface, we probed the mid-gestation placental, uterine, and fetal brain tissue transcriptome. Male and female placentas differentially regulated expression of genes involved in growth and metabolic signaling in response to gestation in an MPS dam. We also report novel offspring sex- and MPS-specific responses in the uterine tissue apposing these placentas. In the fetal compartment, MPS female offspring reduced expression of neurodevelopmental genes. Using a ribosome-tagging transgenic approach we detected a dramatic increase in genes involved in chromatin regulation in a PVN-enriched neuronal population in females at PN21. While MPS had an additive effect on high-fat-diet (HFD)-induced weight gain in male offspring, both MPS and HFD were necessary to induce significant weight gain in female offspring. These data highlight the preconception period as a determinant of maternal health in pregnancy and provides novel insights into mechanisms by which maternal stress history impacts offspring developmental programming.

Keywords: PVN; developmental programming; extracellular vesicles; placenta; preconception stress; sex-specific; uterus.

Graphical Abstract

Figure 1

Maternal preconception stress dampens pregnancy-induced shift in circulating EV protein cargo and glucose processing. (A) Schematic of experimental timeline for control, stress, non-pregnant, and pregnant groups. (B) Zetaview particle analysis identified a significant increase in concentration during pregnancy, plotted as average values across replicates (N = 4–8/group) and area under the curve of size distribution plot confirming an overall increase in circulating EV concentration during pregnancy (F_1,19 = 6.637, * p < 0.05, N = 4–8/group). (C) Heatmap of total identified EV-isolated proteins by mass spectrometry 4 weeks post stress (non-pregnant) or at embryonic day 12.5 (pregnant), compared to controls with hierarchical clustering of samples (light green, control non-pregnant; dark green, control pregnant; orange, stress non-pregnant; red, stress pregnant). Z-scores plotted across individual rows for each protein. (D) MPS imparted a lasting change in the circulating EV proteome of non-pregnant females. Green indicates increased and red indicates decreased abundance in stress relative to control (N = 6–8/group). (E) Effect of pregnancy on the EV proteome presents a distinct profile in MPS dams relative to controls. Green indicates increased and red indicates decreased abundance in pregnant relative to non-pregnant (N = 4–5/group). (F) Clearance of a glucose bolus was unaffected by preconception stress in non-pregnant females (One-way rmANOVA, time × MPS (F_4,68 = 0.0768, NS p = 0.989), time (F_2.348,39.91 = 172.5, p < 0.0001), MPS (F_1,17 = 0.345, NS p = 0.565); N = 8–11). (G) Preconception stressed females impaired glucose clearance at E12.5 (time × MPS (F_4,48 = 3.882, p < 0.05), time (F_1.84,22.12 = 255.2, p < 0.0001), MPS (F_1,12 = 2.065, NS p = 0.176); N = 6–8). Error bars represent SEM.

Figure 2

Stress prior to conception has a lasting offspring sex-specific effect on the placental transcriptome. (A) Experimental timeline detailing generation of control and MPS groups prior to collection of individual maternal:placental:fetal units for RNA sequencing. (B) Representative bubble plot of gene set enrichment analysis (GSEA) illustrates enrichment of genes involved in immune responses in female MPS placentas relative to controls at embryonic day 12.5 (N = 4–5/maternal condition, NES > |1.6|, and FDR < 0.05). (C) Genes involved in lipid and steroid metabolism pathways are reduced MPS male placentas relative to controls at embryonic day 12.5 (N = 4–5/maternal condition, NES > |1.6|, and FDR < 0.05). (D) Male and female placentas have opposing transcriptional responses to MPS in genes involved in lipid transport and IGF signaling (N = 4–5/treatment, NES > |1.6|, and FDR < 0.05).

Figure 3

Maternal preconception stress induces sex-specific effects on the uterine transcriptome and offspring fetal development. (A) Gene sets involved in innate immunity were significantly enriched in MPS uterine tissue apposing female placentas (N = 4–5/treatment, NES > |1.8|, and FDR < 0.05). (B) Gene sets involved in cellular metabolism and respiration were significantly reduced in MPS uterine tissues apposing male placentas (N = 4–5/treatment, NES > |1.8|, and FDR < 0.05). (C) Rank–rank hypergeometric overlap (RRHO) analysis comparing gene expression between female placentas and their apposing uterine tissues showed opposing directionality of expression in genes relating to lipid metabolism, angiogenesis, and DNA damage repair. (D) RRHO analysis of overlap in gene expression between male placentas and their apposing uterine tissue showed parallel expression patterns in genes related to protein export, O-glycan biosynthesis, mitochondrial translation, and the citric acid cycle. Data are plotted as increasing ratios of gene expression across the y- and x-axes for the placenta and uterus, respectively. Each pixel represents the -log₁₀(nominal p-value) of overlapping genes via the hypergeometric distribution and color coded according to degree of significance. The resulting heatmap consists of four quadrants, where the bottom-left and top-right represent concordant changes in the directionality of gene expression between the two tissues (N = 4–5/treatment/sex, max log₁₀(p-value) = 5). Representative Venn diagrams of gene expression overlap in the concordant quadrants. (E) Comparison of gene sets revealed significant reduction in pathways involved in neurodevelopment in MPS female fetal brains (N = 4–5/treatment, NES > |1.8|, and FDR < 0.05). (F) GSEA revealed an increase in genes involved in neuronal differentiation in MPS male fetal brains (N = 4–5/treatment, NES > |1.8|, and FDR < 0.05).

Figure 4

Maternal preconception stress has enduring effects on the developmental trajectory of a PVN-enriched neuronal population. (A) Schematic of experimental timeline detailing maternal experimental conditions and timepoints for collection and isolation of Sim1CrexRiboTag offspring brains for sequencing. (B) Representative Sim1Cre expression in the juvenile mouse hypothalamus. (C) Heatmap of differentially expressed genes at PN21 in Sim1Cre positive neurons in the PVN of MPS and control female offspring (N = 4/group). Z-scores plotted across individual rows for each gene. (D) Representative heatmap of differentially expressed genes at PN21 in the control vs MPS male PVN-enriched population of Sim1Cre neurons. (E) Gene ontology (GO) terms determined by DAVID analysis representing genes enriched in control and MPS female offspring PVN (fold enrichment > 1.5, FDR < 0.05). (F) Gene set enrichment analysis (GSEA) determined minimal effects of MPS in the female PVN at PN70 (N = 4–5/treatment, NES > |1.8|, and FDR < 0.05). (G) Relative to controls, biological processes involved in cell signaling and neural transmission were reduced in male MPS PVN in adulthood (N = 4–5/treatment, NES > |1.8|, and FDR < 0.05).

Figure 5

Enduring effects of maternal preconception stress on offspring response to a high-fat low-fiber diet. (A) Schematic of experimental timeline detailing maternal and offspring experimental conditions and timing to generate four final groups: Control-Chow, MPS-Chow, Control-HFD, and MPS-HFD. (B) MPS did not affect litter size (t(26) = 0.497, p = 0.624), (C) sex ratio (t(26) = 1.042, p = 0.307), (D) or offspring weight prior to weaning (F_2,56 = 0.177 p = 0.838). (E) Dietary challenge revealed female MPS offspring increased susceptibility to high-fat diet-induced weight gain (two-way rmANOVA, time × MPS × HFD (F_7,273 = 2.284, p = 0.0282), HFD × MPS (F_1,39 = 4.215, p = 0.0468), time × MPS (F_7,273 = 5.047, p < 0.0001), time × HFD (F_7,273 = 7.589, p < 0.0001), time (F_7,273 = 700.9, p < 0.0001), MPS (F_1,39 = 4.112, p = 0.0494), HFD (F_1,39 = 0.002, NS p = 0.9634), N = 9–13/treatment/diet). (F) Female MPS HFD-fed offspring weighed more than all other groups after 6 weeks on HFD (two-way ANOVA, F_3,39 = 5.946, p = 0.0019, Tukey control vs MPS + HFD p = 0.009, MPS vs MPS + HFD p = 0.0342, control + HFD vs MPS + HFD p = 0.0021, N = 9–13/treatment/diet). (G) Female MPS offspring maintained an increased caloric intake of high-fat low-fiber diet for 2 weeks after the first week, relative to control HFD-fed females (W5 t(9) = 3.539 *p = 0.00632, W6 t(9) = 2.277 *p = 0.0488, N = 5–6 cages/treatment). (H) High-fat low-fiber diet altered the developmental trajectory of male offspring body weight regardless of maternal condition (two-way rmANOVA, time × MPS × HFD (F_7,223 = 3.943, p = 0.0004), HFD × MPS (F_1,32 = 0.1917, NS p = 0.6644), time × MPS (F_7,223 = 1.203, NS p = 0.3023), time × HFD (F_7,223 = 21.86, p < 0.0001), time (F_2.003,63.82 = 545.1, p < 0.0001), MPS (F_1,32 = 1.898, NS p = 0.1778), HFD (F_1,32 = 7.100, p = 0.0120), N = 9–10/treatment/diet). (I) Male MPS offspring were heavier than their control and chow-fed counterparts after 6 weeks of dietary challenge (two-way ANOVA, F_3,30 = 9.257, p = 0.0002, Tukey control vs MPS + HFD p = 0.0006, MPS vs MPS + HFD p = 0.0003, control + HFD vs MPS + HFD p = 0.0384, N = 9–10/treatment/diet). (J) Male MPS offspring did not consume significantly more calories than controls. Error bars represent SEM.