A role for Notch signaling in trophoblast endovascular invasion and in the pathogenesis of pre-eclampsia

Abstract

Placental trophoblasts (TBs) invade and remodel uterine vessels with an arterial bias. This process, which involves vascular mimicry, re-routes maternal blood to the placenta, but fails in pre-eclampsia. We investigated Notch family members in both contexts, as they play important roles in arterial differentiation/function. Immunoanalyses of tissue sections showed step-wise modulation of Notch receptors/ligands during human TB invasion. Inhibition of Notch signaling reduced invasion of cultured human TBs and expression of the arterial marker EFNB2. In mouse placentas, Notch activity was highest in endovascular TBs. Conditional deletion of Notch2, the only receptor upregulated during mouse TB invasion, reduced arterial invasion, the size of maternal blood canals by 30-40% and placental perfusion by 23%. By E11.5, there was litter-wide lethality in proportion to the number of mutant offspring. In pre-eclampsia, expression of the Notch ligand JAG1 was absent in perivascular and endovascular TBs. We conclude that Notch signaling is crucial for TB vascular invasion.

Fig. 1.

CTBs modulated expression of Notch receptors/ligands as they invaded the uterine wall and vasculature. (A) CTBs of anchoring villi (av) differentiate as they invade the maternal uterus. Undifferentiated CTB progenitors (yellow) surround the mesenchymal cores of anchoring villi. These progenitors differentiate as they move into the proximal and distal regions of the cell columns (cc; orange and purple). Invasive CTBs (red) migrate through the uterus. Endovascular CTBs (blue) disrupt the smooth muscle layer of maternal blood vessels (bv), where they also replace ECs. The depth of CTB invasion in normal pregnancy and in pre-eclampsia is indicated by arrowheads. (B-I) Double indirect immunofluorescence was performed on tissue sections of 2nd trimester basal plate biopsies using antibodies that specifically reacted with KRT7 (red), which is expressed exclusively by trophoblasts, and with Notch family members (green). Nuclei were labeled with DAPI (blue). (B) CTB progenitors (upper inset) upregulated NOTCH2 expression (lower inset) as they invaded the uterine wall and blood vessels. (C,D) CTBs expressed NOTCH3 at all stages of differentiation/invasion (insets). (E) CTB progenitors and cell columns (upper inset) expressed NOTCH4, which was downregulated in proximity to maternal blood vessels (lower inset). (F) DLL1 was expressed by maternal cells in the uterus that associated with CTBs (inset). (G) DLL4 immunoreactivity, which was absent in progenitors (upper inset), increased as CTBs entered the cell columns (lower inset) and declined with deeper invasion. (H,I) CTBs expressed JAG1 only in proximity to maternal spiral arterioles. Scale bars: 100 μm. The direction of invasion is indicated by arrows.

Fig. 2.

Notch inhibition reduced CTB invasion and EFNB2 expression in vitro. (A) Expression of Notch family member RNA in CTBs was quantified as they differentiated in culture over a 16-hour period. NOTCH2 and DLL4 expression increased, NOTCH3 expression remained constant, and NOTCH4 expression decreased. (B,C) The level of CTB apoptosis, which was assessed using the TUNEL method, was negligible in experimental cultures that contained the Notch inhibitor L-685,458 (10 μM) (B) and in control conditions (C). TUNEL-positive apoptotic cells were visualized with FITC (green) and cell nuclei were stained with DAPI (blue). Scale bars: 100 μm. (D) Notch inhibition reduced CTB invasion through Matrigel by 40% (n=10). (E) CTB EFNB2, EPHB4 and MMP9 RNA expression was measured by Taqman at 0 hours and after the cells had fully differentiated in culture (36 hours) with or without the Notch inhibitor. Compared with controls, CTBs cultured with L-685,458 expressed lower levels of EFNB2. GAPDH served as a loading control. Data are represented as mean+s.d.

Fig. 3.

Notch activity was highest in invasive GlyTCs and TGCs that associated with uterine spiral arterioles. (A-C) Tissue sections from the placentas of Tnr mice enabled assessment of Notch activity (green) across gestation. Nuclei were stained with DAPI (blue). (A) At E8.5 (n=5), Notch activity was observed in the ectoplacental cone (epc). (B,C) At E14.5 (n=9), Notch activity was highest in invasive trophoblast cells surrounding maternal blood vessels (bv). (D-F) Double (D) or single (E,F) in situ hybridization was performed on sections of E12.5 Tnr placentas (n=13) to detect Notch activity (purple) and expression of the mouse trophoblast lineage-specific markers Prl2c2 (TGCs; brown) or Pcdh12 (GlyTCs; purple). Nuclei were stained with nuclear Fast Red. Arrowheads indicate double-positive cells. Approximately half of spiral artery-associated (D) TGCs and (E) GlyTCs had (F) Notch activity. Scale bars: 100 μm. ca, canal; cp, chorionic plate; dec, decidua; lab, labyrinth; sp, spongiotrophoblast. (G) Mouse trophoblast stem cells regulated the expression of Notch family members as they differentiated in vitro. Notch RNA expression levels were normalized to Rn18s. Notch2, Jag1 and Dll4 were upregulated with trophoblast differentiation; levels of all other family members declined. Data are represented as mean±s.d.

Fig. 4.

Conditional deletion of Notch2 in invasive trophoblasts reduced TGC and GlyTC invasion of spiral arterioles. (A-H) In situ hybridization was performed on tissue sections of E12.5 Notch2^floxl+;Tpbpa-Cre and Notch2^floxlflox;Tpbpa-Cre placentas for the TGC marker Prl2c2 (A,B, magnified in E,F, respectively) and the GlyTC marker Pcdh12 (C,D, magnified in G,H, respectively). Hybridization was visualized by NBT/BCIP staining (purple). Nuclei were stained with nuclear Fast Red. The decidua (dec), junctional zone (jz) and labyrinth (lab) are shown. In Notch2^floxlflox;Tpbpa-Cre mice, very few TGCs and GlyTCs invaded spiral arterioles (arrowheads). Scale bars: 400 μm in A-D; 100 μm in E-H.

Fig. 5.

The trophoblast-lined vascular canals that supply blood to the placenta were smaller in Notch2^floxlflox;Tpbpa-Cre mice. (A-H) Vascular corrosion casts of the maternal blood spaces of the placenta were prepared at E10.5 and E14.5. In placentas of Notch2 floxl+;Tpbpa-Cre and Notch2^floxlflox;Tpbpa-Cre offspring, a complex array of maternal spiral arterioles converged into several large canals before branching into the labyrinthian sinusoids. At E10.5, the canals in Notch2^floxl+;Tpbpa-Cre placentas (A; enlarged in C) (n=6) were larger than those in Notch2^floxlflox;Tpbpa-Cre placentas (B; enlarged in D) (n=7). We noted similar differences between Notch2^floxl+;Tpbpa-Cre (E; enlarged in G) (n=8) and Notch2^floxlflox;Tpbpa-Cre animals (F; enlarged in H) (n=4) at E14.5. Scale bars: 1 mm in A,B,E,F; 500 μm in C,D,G,H. (I) The vessel diameter was used to calculate the cross-sectional areas of the canals. At E10.5 and E14.5, vessels in Notch2^floxlflox;Tpbpa-Cre placentas were 40% and 34% smaller, respectively, when compared with Notch2^floxl+;Tpbpa-Cre animals. Group means are represented as solid bars.

Fig. 6.

Placental perfusion was compromised in Notch2^floxlflox;Tpbpa-Cre animals. (A-D) To approximate placental perfusion, animals were injected i.v. with fluorescein-conjugated dextran and tissue sections of the labyrinth (lab) were analyzed for the short-term (3-minute) accumulation of this tracer. The FITC signal, which appeared to be stronger in the labyrinth (outlined) of Notch2^floxl+;Tpbpa-Cre placentas (A; enlarged in B) (n= 15) when compared with Notch2^floxlflox;Tpbpa-Cre littermates (C; enlarged in D) (n=8), was quantified by using ImageJ (E). Placental perfusion was reduced by ~23% in Notch2^floxlflox;Tpbpa-Cre placentas. Group means are represented as solid bars. Scale bars: 1 mm in A,C; 200 μm in B,D. dec, decidua.

Fig. 7.

In PE, endovascular and perivascular CTBs failed to express JAG1. (A-F) JAG1 in situ hybridization (purple) and KRT7 immunolocalization (brown) were performed on tissue sections of basal plate biopsies from PTL and PE cases (SPE and HELLP). (A-C) In PTL, endovascular and perivascular CTBs (arrowheads) expressed JAG1. (D-F) In PE, fewer vessels were modified and many of the associated CTBs lacked JAG1 expression. (G,H) Double indirect immunolocalization of JAG1 and KRT7 confirmed these results at the protein level. Nuclei were stained with DAPI. (I) Tissue sections from six PTL, ten SPE and three HELLP cases were scored for the percentage of modified maternal vessels that contained JAG1-positive CTBs. The mean decreased to 53% in SPE and to 26% in HELLP. Individual results are reported in Table S5 in the supplementary material. Data are represented as mean+s.d. Scale bars: 100 μm. bv, blood vessel.