STOX1 deficiency is associated with renin-mediated gestational hypertension and placental defects

Abstract

The pathogenesis of preeclampsia and other hypertensive disorders of pregnancy remains poorly defined despite the substantial burden of maternal and neonatal morbidity associated with these conditions. In particular, the role of genetic variants as determinants of disease susceptibility is understudied. Storkhead-box protein 1 (STOX1) was first identified as a preeclampsia risk gene through family-based genetic linkage studies in which loss-of-function variants were proposed to underlie increased preeclampsia susceptibility. We generated a genetic Stox1 loss-of-function mouse model (Stox1 KO) to evaluate whether STOX1 regulates blood pressure in pregnancy. Pregnant Stox1-KO mice developed gestational hypertension evidenced by a significant increase in blood pressure compared with WT by E17.5. While severe renal, placental, or fetal growth abnormalities were not observed, the Stox1-KO phenotype was associated with placental vascular and extracellular matrix abnormalities. Mechanistically, we found that gestational hypertension in Stox1-KO mice resulted from activation of the uteroplacental renin-angiotensin system. This mechanism was supported by showing that treatment of pregnant Stox1-KO mice with an angiotensin II receptor blocker rescued the phenotype. Our study demonstrates the utility of genetic mouse models for uncovering links between genetic variants and effector pathways implicated in the pathogenesis of hypertensive disorders of pregnancy.

Keywords: Cardiology; Hypertension; Mouse models; Obstetrics/gynecology; Reproductive Biology.

Figure 1. Generation of Stox1-KO mouse and placental expression of Stox1.

(A) Targeting strategy used to generate Stox1-KO mice containing an approximately 1 kb deletion of exon 3. (B and C) Total number of embryos (B) and (C) percentage of resorbed embryos per litter (n = 13 litters). (D) Stox1 mRNA expression in various tissues by reverse transcription PCR (RT-PCR), normalized to β-actin. (E and F) In situ hybridization (E) and immunohistochemistry (F) showing normal Stox1 expression in WT placenta and lack of expression in KO. Scale bars: 50 μm. D, decidua; JZ, junctional zone; L, labyrinth. (G) Biallelic expression of Stox1 mRNA in the placenta, evidenced by expression from either maternal or paternal allele in heterozygous placentas (RT-PCR). Stox1 genotypes of female and male mice shown. F, female; M, male. Results are shown as mean ± SEM. Two-tailed, unpaired t test.

Figure 2. Loss of STOX1 results in gestational hypertension.

(A) Systolic blood pressure at different time points for WT and Stox1-KO cohorts. A subset of Stox1-KO, pregnant mice were treated daily with losartan (KO + Los) starting at E14.5 until sacrifice. NP, nonpregnant; SBP, systolic blood pressure; PP, postpartum (day 10). Details including mouse numbers and all comparisons are in Supplemental Table 1. (B) Urine albumin/creatinine ratio for NP (WT, n = 8; KO, n = 9) and pregnant mice at E17.5 (WT, n = 10; KO, n = 14; KO + Los, n = 7). (C) Representative images of H&E-stained maternal kidney sections from the indicated groups showing normal tubules and glomeruli. Scale bar: 50 μm. (D) Transmission electron micrographs of glomeruli. Arrowheads point to swollen podocyte foot processes; arrow points to exuberant endothelial cells with processes extending into the lumen in Stox1 KO. Scale bar: 100 nm. (E and F) Serum PlGF (E) and sFlt-1 (F) levels measured by ELISA in nonpregnant WT (n = 8) and Stox1 KO (n = 10) and pregnant WT and KO at E17.5 (n = 10). (G and H) Similar weights of embryos (G) and placentas (H) from WT (n = 41 embryos/placentas from 7 litters) and KO (n = 47 embryos/placentas from 8 litters) collected at E17.5. (I) Similar embryo/placenta ratio in WT and KO. Data are mean ± SEM. One-way ANOVA with Bonferroni’s correction (A, B, E, and F); 2-tailed, unpaired t test (G–I). Adjusted *P ≤ 0.05, **P ≤ 0.01, ***P ≤ 0.001, ****P ≤ 0.0001.

Figure 3. Hypoxia and renin upregulation in Stox1-KO placentas.

(A) Representative low-magnification images of H&E-stained WT and KO placenta at E17.5. Scale bar: 200 μm. (B) High-magnification images of the junctional zone and labyrinth. Scale bar: 50 μm. (C) Hypoxyprobe immunohistochemistry of the placenta from WT, KO, and no probe control shows increased hypoxia in the junctional zone of the KO. Scale bar: 100 μm. (D) CD31 immunofluorescence identifies blood vessels in the decidua in WT, KO, and KO treated with losartan (KO + Los) groups at E17.5. Scale bar: 50 μm. (E) Increased junctional zone laminin deposition in KO compared with WT and KO + Los. Scale bar: 50 μm. (F) Quantification of CD31-positive microvessels in D in WT, KO, and KO + losartan placentas (n = 6 placentas from 3–4 litters, 3 high-power fields per placenta). (G) Renin expression in the placenta by Western blot. (H) Quantification of active renin 38 kDa peptide in replicate blots (n = 3). (I) Immunohistochemistry for placental renin. Scale bar: 50 μm. D, decidua; JZ, junctional zone; L, labyrinth. Solid line marks border between the decidua and junctional zone; dotted line marks border between junctional zone and labyrinth. Data are shown as mean ± SEM. One-way ANOVA with Bonferroni’s correction (F); 2-tailed, unpaired t test (H). *P ≤ 0.05, **adjusted P ≤ 0.01.

Figure 4. Analysis of scRNA-Seq data from E9.5 mouse placenta shows coexpression of Stox1 with endothelial cell and SpA-TGC markers.

(A) Diagram showing 6 cell clusters defined in data set from Nelson et al. (38). (B) Overview of the number of variable features producing 6 cell populations obtained from the analyzed samples. (C) Scatter plot showing Stox1 expression in profiled single cells. Population of Stox1⁺ cells analyzed in D and E shown in red. (D) Selected cell population markers expressed in Stox1⁺ cells. Asterisks mark Kdr and Cd34, which were also markers of the novel Prdm1⁺ TGC population. (E) Expression of renin-angiotensin pathway members in Stox1⁺ cells. DEGs, differentially expressed genes; DS, decidual stroma; EC, endothelial cell; NK, natural killer; SpA-TGCs, spiral artery trophoblast giant cell; TGC, trophoblast giant cell. Data are normalized expression values: per cell (C), median ± IQR (D), or mean (E).

Figure 5. STOX1 regulates renin expression in human cytotrophoblasts.

(A) siRNA knockdown of STOX1 in HTR-8/SVneo cells results in increased renin transcription (RT-PCR). (B) Increased active renin peptide (38 kDa) with STOX1 knockdown by Western blot. (C) Quantification of replicate Western blots, normalized to β-actin (n = 7; 1 outlier identified by ROUT method excluded from Stox1 siRNA group). (D) Dual-luciferase assay of HTR-8/SVneo cells transfected with empty luciferase vector control or human REN promoter-luciferase construct. Baseline renin promoter-luciferase activity detected in control siRNA condition. STOX1 siRNA knockdown resulted in decreased renin promoter activity, suggesting negative feedback regulation by renin. Relative luciferase expression normalized to vector control with Ctrl siRNA (n = 3). (E) Human renin 3′-UTR-luciferase reporter showing increased reporter activity with Stox1 knockdown suggesting that STOX1 repression of renin involves the 3′-UTR (n = 3). Ctrl, control. Data are shown as mean ± SEM. Two-tailed, unpaired t test. *P ≤ 0.05, ***P ≤ 0.001.

Figure 6. STOX1 is conserved and Y153H is the most common variant in population data.

(A) Comparison of STOX1 isoforms in human and mouse with sequence alignment of WH DNA-binding domain show high sequence identity (shaded blue). Location of Y153H variant marked. NLS, nuclear localization signal; WH, winged helix. (B) Frequency and number of homozygotes for the 10 most common missense variants in STOX1 and (C) population frequencies of the Y153H variant from gnomAD.