Differential activation of placental unfolded protein response pathways implies heterogeneity in causation of early- and late-onset pre-eclampsia

Abstract

Based on gestational age at diagnosis and/or delivery, pre-eclampsia (PE) is commonly divided into early-onset (<34 weeks) and late-onset (≥34 weeks) forms. Recently, the distinction between 'placental' and 'maternal' causation has been proposed, with 'placental' cases being more frequently associated with early-onset and intrauterine growth restriction. To test whether molecular placental pathology varies according to clinical presentation, we investigated stress-signalling pathways, including unfolded protein response (UPR) pathways, MAPK stress pathways, heat-shock proteins and AMPKα in placentae delivered by caesarean section for clinical indications at different gestational ages. Controls included second-trimester, pre-term and normal-term placentae. BeWo cells were used to investigate how these pathways react to different severities of hypoxia-reoxygenation (H/R) and pro-inflammatory cytokines. Activation of placental UPR and stress-response pathways, including P-IRE1α, ATF6, XBP-1, GRP78 and GRP94, P-p38/p38 and HSP70, was higher in early-onset PE than in both late-onset PE and normotensive controls (NTCs), with a clear inflection around 34 weeks. Placentae from ≥ 34 weeks PE and NTC were indistinguishable. Levels of UPR signalling were similar between second-trimester and term controls, but were significantly higher in pre-term 'controls' delivered vaginally for chorioamnionitis and other conditions. Severe H/R (1/20% O2 ) induced equivalent activation of UPR pathways, including P-eIF2α, ATF6, P-IRE1α, GRP78 and GRP94, in BeWo cells. By contrast, the pro-inflammatory cytokines TNFα and IL-1β induced only mild activation of P-eIF2α and GRP78. AKT, a central regulator of cell proliferation, was reduced in the < 34 weeks PE placentae and severe H/R-treated cells, but not in other conditions. These findings provide the first molecular evidence that placental stress may contribute to the pathophysiology of early-onset pre-eclampsia, whereas that is unlikely to be the case in the late-onset form of the syndrome.

Keywords: cell stress responses; growth restriction; placenta; pre-eclampsia; pregnancy; unfolded protein response.

Figure 1

Continued. High activation of UPR pathways in early-PE (<34 weeks) but not late-PE (≥34 weeks) placentae and a strong negative correlation between activation and gestational age of PE placentae. (A) Comparison among early-PEs, late-PEs and NTCs; the intensity of bands was quantified and normalized with β-actin: data are presented as medians; a non-parametric Kruskal–Wallis test with Dunn's multiple comparison test was used, with p ≤ 0.05 considered significant; a, b, significant differences between < 34 weeks PEs and term controls or ≥ 34 weeks PEs, respectively; term con, term controls. (B) Electron micrographs show dilation of ER cisternae in the early-PE placenta (29 weeks) but not in the late-PE (38 weeks) and NTC (39 weeks) placentae: arrows, normal ER cisternae morphology; arrowheads, dilated ER cisternae; magnification = ×1900; scale bar = 2 µm. (C) Pearson correlation analysis was performed on the arbitrary value of quantified data plotted against gestational age, with p ≤ 0.05 considered significant; power regression lines were fitted to display the relationship; diamonds, PE samples; squares, NTC samples; ****p < 0.0001; ***p < 0.001

Figure 2

A negative correlation between activated p38 kinase, AMPKα, HSP70 and gestational age of PE placentae. (A) Pearson correlation analysis was performed on the arbitrary value of quantified data plotted against gestational age, with p ≤ 0.05 considered significant; power regression lines were fitted to display the relationship: diamonds, PE samples; squares, NTC samples; ***p < 0.001; **p < 0.01; *p < 0.05. (B) Comparison among early-PEs, late-PEs and NTCs; the intensity of bands was quantified and normalized with β-actin: data are presented as medians; a non-parametric Kruskal–Wallis test with Dunn's multiple comparison test was used, with p ≤ 0.05 considered significant; a, b, significant differences between < 34 weeks PE and NTC or ≥ 34 weeks PE, respectively

Figure 3

Placentae from early-PE cluster separately from the majority of late-PEs and NTCs, based on the levels of stress-response proteins. Cluster analysis of western blot data was performed using Ward's agglomerative method: d, days

Figure 4

No change of UPR activation in second-trimester placentae compared to normotensive term placentae. (A) Tissue lysates from seven second-trimester (ST) and seven NTC placentae were resolved in SDS–PAGE before probing with antibodies against ER stress markers. (B) The intensity of bands was quantified and normalized with β-actin; phosphorylation status is presented as the ratio between phosphorylated and total protein: data are medians for seven placentae/group; no significant changes were observed in any UPR pathway

Figure 5

Severe but not mild hypoxia–reoxygenation activates the UPR and reduces proliferation of BeWo cells. (A) Cells were incubated under either 6 h repetitive cycles of 1% and 20% O₂ (severe H/R) or 5% and 20% O₂ (mild H/R) for 24 h. Cell lysates were collected and resolved in SDS–PAGE before probing with primary antibodies specific for UPR markers; β-actin was used as the loading control. (B) Densitometry of bands expressed relative to normal controls; phosphorylation status is presented as the ratio between phosphorylated and total protein: data are mean ± SEM for four independent experiments; a, significance at p < 0.05 compared to normoxic controls (20 n). (C) Cells used for cell proliferation assay were subjected to 48 h H/R challenge and the numbers compared to the 20% O₂ normoxic controls (100%): data are mean ± SEM from four independent experiments; *p < 0.05

Figure 6

Pro-inflammatory cytokines, TNFα and IL-1β, have a subtle effect on UPR pathways but do not affect cell proliferation. (A) Cells were treated with TNFα (50 ng/ml), IL-1β (50 ng/ml) or both for 24 h. Cell lysates were subjected to western blotting analysis with antibodies specific for UPR markers; β-actin was used as the loading control. (B) Densitometry of bands expressed relative to normal controls; phosphorylation status is presented as the ratio between phosphorylated and total protein: data are mean ± SEM for five independent experiments; a, significance at p < 0.05 compared to non-treatment controls (Control). (C) Cells used for the cell proliferation assay were subjected to 48 h incubation and the numbers compared to the non-treatment controls (100%): data are mean ± SEM from four independent experiments

Figure 7

Down-regulation of AKT in the < 34 weeks PE placentae and severe H/R-treated cells, but not in the ≥ 34 weeks PE placentae, second-trimester, mild H/R-treated and TNFα + IL-1β-treated cells. (A) Comparison among early-PEs, late-PEs and NTCs: data are presented as medians; a non-parametric Kruskal–Wallis test with Dunn's multiple comparison test was used, with p ≤ 0.05 considered significant; a, b, significant differences between < 34 weeks PE and term control or ≥ 34 weeks PE, respectively. (B) Correlation analysis on the arbitrary value of quantified data was plotted against gestational age; a Pearson correlation was performed, with p ≤ 0.05 being considered significant; power regression lines were fitted to display the relationship: diamonds, PE samples; squares, NTC samples; ****p < 0.0001. (C) AKT protein level was analysed in the second-trimester placentae and normotensive term controls by western blot. (D) H/R or TNFα and IL-1β treatment was performed and western blotting was used to analyse AKT protein level; β-actin was used as the loading control: data are mean ± SEM for four or five independent experiments; p = 0.068 compared to normoxic control

Figure 8

Schematic diagram illustrating how our findings fit into a model integrating the pathophysiology (degree of spiral artery conversion), different potential aetiologies (placental pathology and maternal susceptibility) and clinical outcomes of pre-eclampsia at different gestational ages