**2. Results**

#### *2.1. E*ff*ects of Hydroxychloroquine on Placental Secretion*

Hypoxia significantly increased the secretion of sFlt-1 (Figure 1a, *p* = 0.02), sEng (Figure 1b, *p* = 0.02), and TNF-α (Figure 1c, *p* = 0.02) from explant cultures after 24 h incubation. In the presence of X-XO (xanthine/xanthine oxidase system), explants cultured for 48 h significantly increased secretions of 8-isoprostane (Figure 2a, *p* = 0.03) and activin A (Figure 2b, *p* = 0.01) compared to controls. Co-incubation with 1 μg/mL hydroxychloroquine did not alter either the hypoxia-induced secretion of sFlt-1 (Figure 1a), sEng (Figure 1b), or TNF-α (Figure 1c), or the X-XO-induced increase in 8-isoprostane (Figure 2a) and activin A (Figure 2b).

#### *2.2. E*ff*ect of Hydroxychloroquine on HUVEC Viability*

Previously we have demonstrated that, compared to untreated controls, there was no effect of hydroxychloroquine on human umbilical vein endothelial cell (HUVEC) viability across a dose range of 0.1, 1, and 10 μg/mL over 120 h in culture [25]. However, treatment of cells with

100 μg/mL hydroxychloroquine significantly reduced cell viability at 24 h (*p* < 0.001) [25]. Dosing of hydroxychloroquine for all subsequent experiments were based on these results.

**Figure 1.** Release of (**a**) soluble fms-like tyrosine kinase-1 (sFlt-1), (**b**) soluble endoglin (sEng), and (**c**) tumour necrosis factor-α (TNF-α) by placental explants of human term normal pregnancy placentae after 24 h incubation at 5% oxygen concentration (normoxia) versus 1% oxygen (hypoxia). The explants were incubated in the hypoxic environment in the absence or presence of 1 μg/mL hydroxychloroquine. Data are mean ± standard error of the mean (SEM) from 10 independent biological replicates. \* denotes *p* < 0.05. NT: non treated, HCQ: hydroxychloroquine.

**Figure 2.** Release of (**a**) 8-isoprostane and (**b**) activin A by placental explants of human term normal pregnancy placentae after 48 h incubation at 20% oxygen concentration with 5% CO2. The explants were incubated in media containing xanthine (2.3 mM) + xanthine oxidase (15 mU/mL) in the absence or presence of 1 μg/mL hydroxychloroquine. Data are mean ± SEM from 10 independent biological replicates. \* denotes *p* < 0.05. X/XO: xanthine/xanthine oxidase, HCQ: hydroxychloroquine.

#### *2.3. E*ff*ects of Hydroxychloroquine on Endothelial Function In Vitro*

HUVECs were treated in the absence or presence of (i) TNF-α (100 ng/mL), (ii) sera from normal pregnancies (20%), or (iii) sera from preeclamptic women (20%) in the presence or absence of hydroxychloroquine (1 μg/mL) to assess endothelial dysfunction (Figure 3). Compared to controls, incubation of HUVECs with TNF-α (Figure 3a,c) or sera from preeclamptic women (Figure 3b,d) significantly increased both NADPH oxidase 2 (NOX2) mRNA expression (*p* < 0.001 and *p* = 0.01, respectively) and 8-isoprostane secretion (*p* = 0.02 and *p* = 0.04, respectively). Co-treatment of HUVECs with TNF-α and hydroxychloroquine significantly reduced NOX2 mRNA expression (Figure 3a, *p* = 0.03) and secretion of 8-isoprostane (Figure 3c, *p* = 0.04). Co-treatment of HUVECs with serum from preeclamptic women and hydroxychloroquine did not significantly alter the expression of NOX2

mRNA or 8-isoprostane. However, 100 μM apocynin, a NOX inhibitor, significantly reduced the NOX2 mRNA expression and 8-isoprostane release induced by serum from preeclamptic women (Figure 3b,d, respectively, *p* < 0.01 for both).

**Figure 3.** NADPH oxidase 2 (NOX2) RNA expression of human umbilical vein endothelial cells (HUVECs) treated with 100 ng/mL TNF-α (**a**) and 20% preeclampsia (PE) sera (**b**). Release of 8-isoprostane by HUVECs treated with 100 ng/mL recombinant TNF-α (**c**) and 20% preeclampsia sera (**d**). Data are mean ± SEM from eight independent biological replicates. \* denotes *p* < 0.05; \*\*\*\*p<0.001.

Compared to controls, incubation of HUVECs with TNF-α (Figure 4a) or 20% sera from preeclamptic women (Figure 4b) increased immunoreactivity for NOX2 protein. Once again, co-treatment of HUVECs with TNF-α and either apocynin or hydroxychloroquine reduced immunoreactive NOX2 protein expression (Figure 4a). Similarly, co-treatment of HUVECs with sera from preeclamptic women and either apocynin or hydroxychloroquine also showed reduced immunoreactive NOX2 protein expression (Figure 4b).

**Figure 4.** Western blot representative for NOX2 protein expression of HUVECs untreated (cont) or treated with 100 ng/mL TNF-α (**a**) or 20% preeclampsia (PE) sera (**b**) with or without apocynin (apo, 100 μM) or hydroxychloroquine (HCQ, 1 μg/mL). β-actin was used as a loading control.

#### *2.4. E*ff*ect of Hydroxychloroquine on Vascular Permeability*

Both TNF-α (Figure 5a) and sera from preeclamptic women (Figure 5b) significantly increased HUVEC monolayer permeability compared to controls (*p*=0.02 and *p*=0.005, respectively). These effects were mitigated by co-treatment with hydroxychloroquine (*p* = 0.04 and *p* = 0.007, respectively). Hydroxychloroquine prevented the significant loss of zonula occludens 1 (ZO-1) induced by both TNF-α (Figure 5c, *p* = 0.003) and sera from preeclamptic women (Figure 5d, *p* = 0.02).

**Figure 5.** HUVECs' permeability when treated with 100 ng/mL recombinant TNF-α (**a**) and 20% preeclampsia sera (**b**) (*n* = 9). Mean zonula occludens 1 (ZO-1) fluorescence when treated with 100 ng/mL recombinant TNF-α (**c**) and 20% preeclampsia sera (**d**) (*n* = 6). Data are means ± SEM from nine and six independent biological replicates, respectively. \* denotes *p* < 0.05, and \*\* denotes *p* < 0.005.

#### *2.5. E*ff*ect of Hydroxychloroquine on Zonula Occludens 1 (ZO-1) Immunohistochemistry*

Figure 6 contains representative images of ZO-1 immunostaining. There was normal ZO-1 immunostaining in untreated or HUVECs treated with sera from normal pregnancies (Figure 6A,D), with the loss of immunostaining in cells treated with either TNF-α (Figure 6B) or preeclampsia sera (Figure 6E). Hydroxychloroquine rescued the loss of ZO-1 induced by both TNF-α (Figure 6C) and preeclampsia sera (Figure 6F).

**Figure 6.** Immunofluorescent staining of ZO-1 on HUVECs treated with 100 ng/mL recombinant TNF-α or 20% preeclampsia sera for 16–22 h. Representative images from one of six experiments are shown. (**A)** Control untreated HUVECs, (**B**) TNF-α 100 ng/mL, (**C**) TNF-α 100 ng/mL with hydroxychloroquine 1 μg/mL, (**D**) control HUVECs treated with 20% normal pregnancy sera, (**E**) 20% preeclampsia sera, and (**F**) preeclampsia sera with hydroxychloroquine 1 μg/mL. Arrows show the ZO-1 staining on the endothelial cell border.
