A placenta-on-a-chip model to determine the regulation of FKBPL and galectin-3 in preeclampsia

Abstract

Preeclampsia is a pregnancy-specific cardiovascular disorder, involving significant maternal endothelial dysfunction. Although inappropriate placentation due to aberrant angiogenesis, inflammation and shallow trophoblast invasion are the root causes of preeclampsia, pathogenic mechanisms are poorly understood, particularly in early pregnancy. Here, we first confirm the abnormal expression of important vascular and inflammatory proteins, FK506-binding protein-like (FKBPL) and galectin-3 (Gal-3), in human plasma and placental tissues from women with preeclampsia and normotensive controls. We then employ a three-dimensional microfluidic placental model incorporating human umbilical vein endothelial cells (HUVECs) and a first trimester trophoblast cell line (ACH-3P) to investigate FKBPL and Gal-3 signaling in inflammatory conditions. In human samples, both circulating (n = 17 controls; n = 30 preeclampsia) and placental (n ≥ 6) FKBPL and Gal-3 levels were increased in preeclampsia compared to controls (plasma: FKBPL, p < 0.0001; Gal-3, p < 0.01; placenta: FKBPL, p < 0.05; Gal-3, p < 0.01), indicative of vascular dysfunction in preeclampsia. In our placenta-on-a-chip model, we show that endothelial cells are critical for trophoblast-mediated migration and that trophoblasts effectively remodel endothelial vascular networks. Inflammatory cytokine tumour necrosis factor-α (10 ng/mL) modulates both FKBPL and Gal-3 signaling in conjunction with trophoblast migration and impairs vascular network formation (p < 0.005). Our placenta-on-a-chip recapitulates aspects of inappropriate placental development and vascular dysfunction in preeclampsia.

Keywords: FKBPL; Galectin-3; Microfluidics; Placental development; Preeclampsia; Vascular remodeling.

Fig. 1

FKBPL and Gal-3 are increased in the placentae and plasma of women with preeclampsia. Protein lysates were generated from placental tissue collected from women with preeclampsia or normotensive controls. a FKBPL expression was determined by Western Blotting and normalized to GAPDH, the loading control. b Gal-3 levels from placental lysates were evaluated by enzyme-linked immunosorbent assay (ELISA). Data were plotted as mean ± SD; n ≥ 6. c, d Plasma FKBPL and Gal-3 levels from women with preeclampsia vs normotensive controls were assessed by ELISA. Data plotted as mean ± SD; n ≥ 17; unpaired student’s t test, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 2

TNF-α treatment of trophoblasts and endothelial cells in 2D monocultures alters FKBPL and Gal-3 protein expression. a–c Western Blotting results of ACH-3Ps protein lysate expression of FKBPL and Gal-3. ACH-3Ps exposed to tumor necrosis factor alpha (TNF-α, 10 ng/mL) for 24 or 72 h. Control, untreated. GAPDH, loading control. d–f Western Blotting of HUVECs protein lysate showing expression of FKBPL and Gal-3. HUVECs exposed to TNF-α (10 ng/mL) for 24 or 72 h. Control, untreated GAPDH, loading control. Data passed Shapiro–Wilk normality test and were analyzed by one-way analysis of variance (ANOVA) with Tukey post-hoc test; n = 3; *p < 0.05, ***p < 0.001

Fig. 3

Endothelial cell presence and TNF-α modify trophoblast migration, FKBPL and Gal-3 expression in a microfluidic chip. In the co-culture set of chips, HUVECs were embedded within the center matrix channel and ACH-3Ps were added to the side channel. a Representative immunofluorescence (IF) images of ACH-3Ps and HUVECs with high expression for FKBPL and CD31, respectively. Nuclei of cells were visualized using DAPI. b Representative IF images of ACH-3Ps invasion across the device (left to right) in the absence or presence of HUVECs and in normal or inflammatory conditions. Cells were IF stained for cytokeratin 7, a marker of trophoblasts, and DAPI. e The number of migrating trophoblast cells from the left side channel were analyzed using ImageJ. c, d ACH-3Ps monoculture chips were also fixed and IF stained for FKBPL, Galectin-3 and DAPI. Chips were treated with TNF-α (10 ng/mL) for 24 or 72 h, with untreated cells as a control. The fold change of f FKBPL expression in ACH-3Ps without HUVECs and h with HUVECs. The fold change of g Gal-3 expression in ACH-3Ps without HUVECs and i with HUVECs Gal-3. Scalebars represent 100 µm. Data plotted as mean fold change ± SEM, ordinary one-way ANOVA or two-way ANOVA with Tukey post hoc test, n = 3, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 4

The presence of ACH-3Ps cells and inflammatory conditions impacts vasculare network formation and FKBPL, Gal-3 and CD31 expression in endothelial cells cultured in a microfluidic device. a In endothelial cells monoculture microfluidic setting, HUVECs were combined with collagen matrix (2.5 mg/mL) and added to the central channel of the microfluidic chips. In the co-culture set of chips, HUVECs were embedded within the central matrix channel and ACH-3Ps were added to the side channel. Chips were treated with TNF-α (10 ng/mL) for 24 or 72 h, with untreated cells as a control. Following 72 h of culture, chips were probed for immunofluorescent imaging of FKBPL, CD31 and Gal-3. a Representative images of cells stained for DAPI, FKBPL and CD31. b Fold change of FKBPL, c Gal-3 and d CD31 expression in HUVECs without ACH-3Ps. e Fold change of FKBPL, f Gal-3 and g CD31expression in HUVECs with ACH-3Ps. Data presented as mean ± SEM, scalebar represents 100 µm. Unpaired student’s t test and ordinary one-way ANOVA with Tukey post hoc test for normally distributed data and Mann–Whitney or Kruskal–Wallis post hoc test for non-normally distributed data; n = 3; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 5

Quantification of vascular network formation. a Representative immunofluorescent images of HUVECs in microfluidic devices under different TNF-α conditions, that were analyzed using the Angiogenesis Analyzer ImageJ macro and their corresponding map outputs. Scalebar represents 100 µm. The b number of master segments, c number of master junctions and d total isolated branches length of HUVECs with and without ACH-3Ps in the system. Data presented as mean ± SEM. Ordinary two-way ANOVA with Tukey post hoc test; n = 3; **p < 0.01, ***p < 0.001, ****p < 0.0001