Senescent Syncytiotrophoblast Secretion During Early Onset Preeclampsia

Abstract

Background: Preeclampsia is a severe hypertensive disorder in pregnancy that causes preterm delivery, maternal and fetal morbidity, mortality, and life-long sequelae. Understanding the pathogenesis of preeclampsia is a critical first step toward protecting mother and child from this syndrome and increased risk of cardiovascular disease later in life. However, effective early predictive tests and therapies for preeclampsia are scarce.

Methods: To identify novel markers and signaling pathways for early onset preeclampsia, we profiled human maternal-fetal interface units (fetal villi and maternal decidua) from early onset preeclampsia and healthy controls using single-nucleus RNA sequencing combined with spatial transcriptomics. The placental syncytiotrophoblast is in direct contact with maternal blood and forms the barrier between fetal and maternal circulation.

Results: We identified different transcriptomic states of the endocrine syncytiotrophoblast nuclei with patterns of dysregulation associated with a senescence-associated secretory phenotype and a spatial dysregulation of senescence in the placental trophoblast layer. Elevated senescence markers were validated in placental tissues of clinical multicenter cohorts. Importantly, several secreted senescence-associated secretory phenotype factors were elevated in maternal blood already in the first trimester. We verified the secreted senescence markers, PAI-1 (plasminogen activator inhibitor 1) and activin A, as identified in our single-nucleus RNA sequencing model as predictive markers before clinical preeclampsia diagnosis.

Conclusions: This indicates that increased syncytiotrophoblast senescence appears weeks before clinical manifestation of early onset preeclampsia, suggesting that the dysregulated preeclamptic placenta starts with higher cell maturation resulting in premature and increased senescence-associated secretory phenotype release. These senescence-associated secretory phenotype markers may serve as an additional early diagnostic tool for this syndrome.

Keywords: placenta; plasminogen activator inhibitor 1; pre-eclampsia; pregnancy; pregnancy trimesters.

Figure 1.

Multiomics study design using multiple validation levels with patient cohorts. Schematic illustration of the experimental design across different stages of gestation, early pregnancy (e.ctrl), late healthy term pregnancy (trm.ctrl), and early onset preeclampsia (eoPE). Placental tissue was surgically sampled, and villi and decidua were collected separately for snRNA-seq. Tissues from e.ctrl correspond to 5 to 10 weeks of gestation, eoPE tissue was sampled at <34 weeks of gestation (range, 27–33 weeks of gestation), and trm.ctrl samples at >38 weeks (range, 38–40 weeks of gestation). The gestational age difference between healthy term controls (trm.ctrl) and diseased eoPE was corrected by using additional single-cell RNA sequencing (scRNA-seq) data from preterm (ptrm)-delivered nonhypertensive obstetric pathologies. All key findings were validated in patient cohorts using the methods outlined in the lower figures. FGR indicates fetal growth restriction; ISS, in situ sequencing; PE, preeclampsia; RT-qPCR, quantitative real-time PCR; and snRNA-seq, single-nucleus RNA sequencing.

Figure 2.

Single-nucleus–resolved transcriptomic landscape of the healthy and early onset preeclamptic maternal-fetal interface. A and B, Uniform manifold approximation and projection (UMAP) visualizing (A) maternal (decidual, d) and (B) fetal (villous, v) cell types and states obtained from the single-nucleus RNA sequencing (snRNA-seq) data, with integrated samples from early pregnancy (e.ctrl), late healthy pregnancy (trm.ctrl), and early onset preeclampsia (eoPE). Each dot represents a nucleus; color indicates cell type or state. C, Abbreviations for cell types and states presented in this study. D and E, Distribution of cell type composition at different gestational time points for (D) decidua and (E) villi (% of total tissue nuclei). Numbers under the bars indicate the total high-quality nuclei and their corresponding biological replicate sample size. F, Cell composition comparison by trophoblast, matrisome, and immune compartments between term.ctrl. and eoPE. Two-tailed Welch t test.

Figure 3.

Early onset preeclampsia dysregulates the decidua and its interaction with fetal cells. A, Significantly dysregulated gene expression profiles between term control (trm.ctrl) and early onset preeclampsia (eoPE) decidual cell types (log2-fold change [log2FC] >±0.25; adjusted P<0.05). Log2FC between conditions for each dysregulated gene (dot) visualized. The decidual epithelial cells, monocytes, proliferating natural killer cells, plasma cells, granulocytes, progenitor lymphatic endothelial cells, extravillous trophoblast, and departed syncytiotrophoblast were excluded due to large composition changes between trm.ctrl and eoPE. B, GDF15- protein in decidual deported syncytiotrophoblast (dDSTB) localized by immunohistochemistry in decidual tissue. dDSTB is identified among maternal erythrocytes in maternal vessels as an HLA-G negative (human leukocyte antigen G) β-hCG positive (human chorionic gonadotropin) GDF15 positive (growth differentiation factor 15) fragment (shown in serial sections’ staining as indicated). Arrows indicate the maternal vessel border (n=3, representative area shown). Scale bars, 100 µm (overview) and 20 µm (detailed). C, Interaction between pathologically altered dDSTB-secreted ligands with decidual vascular endothelial cell (dVEC) and smooth muscle cell (dSMC) receptors at the maternal-fetal interface. Only altered receptor-ligand interaction pairs during eoPE are shown; arrows point toward receptors (Wilcoxon rank-sum test, P<0.05, identified by multiple tools and databases). Dot colors encode cell types/states, and squares illustrate the average expression of receptors and ligands of interest. Senescence-associated secretory phenotypes (SASPs) are marked. D, Overlap of differentially expressed genes between decidual immune and matrisome cell-type groups. E, Pathway enrichment analysis of commonly dysregulated genes in decidual immune and matrisome compartments (n=83) calculated using the Metascape tool. In A, C, and D, n=11 villi (6 late controls and 5 eoPEs) and n=9 decidua (4 late controls and 5 eoPEs).

Figure 4.

Syncytiotrophoblast populations are the most affected placental cell population in early onset preeclampsia. A, Significantly dysregulated gene expression profiles between term control (trm.ctrl) and early onset preeclampsia (eoPE) villous cell types (log2-fold change [log2FC] >±0.25; adjusted P<0.05). Log2FC between conditions for each dysregulated gene (dot) visualized. B, Overlap of differentially expressed genes (DEGs) between villous cell-type groups. C, Convergence of DEGs in the nucleus states from the syncytiotrophoblast (STB) where each dot represents a single gene (shared upregulated: red; shared downregulated: blue; and differential gene expression: yellow). D, Heatmap of the mean log2FC values of overlapping genes of concordant dysregulation directionality (n=28) between nuclear STB (vSTB) states as visualized in C. E, Overlap of senescence-associated secretory phenotype (SASP) genes derived from the human SASP atlas (Buch Institute) with all and shared genes dysregulated in vSTB nuclear states (min. log2FC 0.4, genes expressed in ≥30% of nuclei and adjusted P<0.05). F, Network of statistically significant hub genes of vSTB DEGs. Node color represents log2FC from the DEG between trm.ctrl and eoPE. Overall, n=12 placentas (6 term controls and 6 eoPEs).

Figure 5.

Senescence profile and spatial dysregulation of the maternal-fetal interface in early onset preeclampsia. A, Heatmap illustrating log2-fold changes (log2FC) of senescence-associated secretory phenotype (SASP)–related genes dysregulated in ≥2 nuclear villous syncytiotrophoblast (vSTB) states in early onset preeclampsia (eoPE; min. log2FC 0.4, expressed in ≥30% of nuclei and adjusted P<0.05 per gene). B, Upregulated mRNA expression (quantitative real-time PCR [RT-qPCR]) of SASP markers GDF15 (growth differentiation factor 15), INHBA, and SERPINE1 in villous tissue of eoPE in a clinical cohort compared with healthy matched preterm controls (ptrm: n=18; eoPE: n=61; sig.level **<0.01, ***<0.001, and ****<0.0001; and unpaired 2-tailed Welch t test). C, Representative immunofluorescence staining of GDF15 in term controls (trm.ctrl) and eoPE (n=3 per group; scale bar, 50 µm). D, Representative immunohistochemical staining of PAI-1 (plasminogen activator inhibitor 1; encoded by the SERPINE1 gene) in trm.ctrl vs eoPE placentas suggests overall increased expression in the vSTB layer (n=4 per group; scale bar, 40 µm). E, Left, Targeted high-resolution spatial transcriptomic data acquired through spatially resolved in situ sequencing (ISS), with spatial context of mRNA molecules implicating highly structured tissue in an unsupervised 2-dimensional embedding (n=2 total, representative area shown; scale bar, 60 µm). Right, Senescence-associated INHBA expression is spatially variable in ISS data and is closer to vascular transcripts (black arrows) in eoPE compared with controls. F, mRNA expression (RT-qPCR) of SASP markers, GDF15, INHBA, and SERPINE1, is not related to fetal growth restriction (FGR) in eoPE (eoPE without FGR: n=16; eoPE with FGR: n=61; sig.level **<0.01, ***<0.001, and ****<0.0001; and unpaired 2-tailed Welch t test).

Figure 6.

Senescence factors derived from the maternal-fetal interface are detectable in maternal blood and serve as good predictors for preeclampsia (PE). A, Dysregulated and secreted ligands from the villous syncytiotrophoblast (vSTB) in early onset PE (eoPE) act on highly expressed receptors in decidual vascular endothelial cells (vVECs) and smooth muscle cells (dSMCs). This highlights ligand pressure, that is, increased ligand expression with unaltered receptor expression at the maternal-fetal interface. Arrows point toward receptors (Wilcoxon rank-sum test, P<0.05, identified by multiple tools and databases). Dot colors encode cell types/states, ligand squares represent upregulated secreted log2-fold change (log2FC) expression per nuclear state, and receptor squares encode average expression. Senescence-associated secretory phenotypes (SASPs) are marked. B, Schematic illustrating the maternal blood sampling strategy at 36 weeks of pregnancy before diagnosis of PE. Samples are part of the large-scale Australian Triple B Study: Bumps, Babies and Beyond (BUMPS) cohort (n=222). C, Circulating GDF15 (growth differentiation factor 15) and PAI-1 (plasminogen activator inhibitor 1; encoded by SERPINE1) are characteristic for the preeclamptic syndrome and are used as predictive markers for PE (n=21; healthy term controls: n=201). D, Schematic illustrating published maternal first-trimester serum sampling strategy of prospectively recruited women healthy term, and eoPE pregnancies were diagnosed in later gestation (n=972). E, Circulating activin A, synthesized by inhibin βα dimers (encoded by INHBA), predicts eoPE in a conditional logistic regression model and shows that the currently used Fetal Maternal Foundation (FMF) algorithm underestimates eoPE risk in women with high first-trimester activin A (eoPE: n=27; healthy term controls: n=47, matched for body mass index, gestational age, and maternal age; subset of cohort is shown in D). AUC indicates area under the curve.