Plasmas of this ancillary study were obtained from a prospective cohort study conducted at the department of Gynaecology and the Haematology Outpatients Department, Nîmes University Hospital (France). Participants were included from June 2015 to December 2018 (“Grosspath” study. clinicaltrials.gov identifier: NCT 01736826). The sample-size estimation was based on the primary objective to compare plasma nucleosomes and cfDNA levels between women with normal pregnancies and women developing placenta-mediated pregnancy complications. Participants were recruited as previously described (Bouvier et al., 2020). Briefly, in the “GrossPath” study, we recruited 30 healthy, non-pregnant volunteers (Group HV), 50 pregnant women with normal pregnancies (Group NP), and 14 pregnant women who had developed preeclampsia (Group PE).

The healthy, non-pregnant women were recruited after a general call to all hospital staff. Ages of controls were matched with the ages of pregnant women. Exclusion criteria were: pregnancy and postpartum period (up to 3 months after childbirth), any history of thrombotic events, chronic disease, chronic or recent infection and cancer.

The 50 pregnant women who finally experienced normal pregnancies were recruited during their initial consultation at the department of gynaecology and obstetrics in their 12th week of gestation. These women gave a single blood sample at the time of delivery. Exclusion criteria were: any previous placenta-mediated pregnancy complications, any history of thrombotic events, chronic disease, chronic or recent infection and cancer. Any patients from Group NP who subsequently developed pregnancy complications during follow-up were then transferred to the abnormal pregnancy group and analysed in that group.

A total of 14 pregnant women who had developed PE were included at the onset of the disease. Among these 14 patients, one woman from Group NP was transferred following the occurrence of preeclampsia in the seventh month of pregnancy. Blood samples were obtained in order to compare biological markers between Group NP and Group PE just before delivery to avoid a bias caused by labour and delivery between the two groups. PE was diagnosed according to the international criteria (Brown et al., 2018; ACOG, 2019). If 24 h-urine collection was not available, the protein-creatinine ratio was measured from a random urine sample (Price et al., 2005).

All plasmas available from preeclamptic women (n = 13/14) were tested for this study. One plasma from the GrossPath study could not be tested due to insufficient quantity. In parallel, 13 plasmas from normal pregnancies were tested. The 13 women with normal pregnancies were selected via the propensity score method with maternal and gestational ages as matching variables (+/−2 years old and+/−1 week of gestation respectively).

The main and ancillary studies (clinicaltrials.gov identifier: NCT 01736826 and NCT 05470712, respectively) were approved by the Institutional Review Board (IRB) and ethics committee (IRB no. 22.07.01) at Nîmes University Hospital and by the local Committee for the Protection of Persons undergoing Biomedical Research (CPP Sud Méditerranée III). This clinical investigation was performed in accordance with the Helsinki declaration of 1975 as revised in 1996. All participants had given written informed consent to participate in the trial (NCT 01736826) and to constitute a biobank to be used later for the same theme. All pregnancies were singletons.

The patients’ clinical data—age, smoking habits, blood pressure, information about current and previous pregnancies, delivery, and perinatal outcomes—were all collected.

PMN cells from three healthy consenting volunteers were immediately purified after blood collection using density gradient centrifugation according to the manufacturer’s instructions (Polymorphoprep™, Proteogenix). A complete blood cell count was made in parallel. At the end of purification, 4.5 *10⁵ purified PMNs were seeded on 0.001% poly-L-lysine coated coverslips in a 24-well plate using Roswell Park Memorial Institute medium (RPMI) with 50% phosphate buffer saline (PBS) Ca²⁺, Mg²⁺ without fetal bovine serum (FBS).

For each experiment, several conditions were tested. Purified neutrophils were non-stimulated or stimulated with either plasma from non-pregnant women (negative controls), phorbol-12-myristate-13-acetate at 25 nM, (PMA, Merck Biodevelopment, France, chemical NETs inductor, positive controls (Urban et al., 2009; O’Donoghue et al., 2013; Douda et al., 2015) or plasma from participants (1:1 dilution in RPMI with 50% PBS, Ca²⁺, Mg²⁺ without FBS): normal pregnancy (Group NP) or women having developed PE (Group PE) for 4 h at 37°C as described by Hu et al. (2018).

PMNs were then fixed on coverslip using 4% paraformaldehyde (Santa Cruz Biotechnology, INC) for 10 minutes and stained after three washes with PBS. After permeabilization with PBS/Tween-20 0.2% for 10 minutes, non-specific antigenic PMN sites were blocked with PBS/Tween 0.1% with bovine serum albumin (BSA) 1% without EDTA for 30 minutes. After three additional washes, cells were incubated with antibodies: Fluorescein isothiocyanate (FITC) anti-myeloperoxydase antibody (mouse anti-human clone 5B8, BD Biosciences™, United States, 1:10 dilution); uncoupled anti-citrullinated histone H3 antibody (Rabbit polyclonal to Histone H3 (citrulline R2 + R8 +R17), Abcam™, Canada, 1/1,000 dilution) overnight at 4°C in the dark. Cells were then washed with PBS and counterstained with Alexa Fluor-568 (Donkey anti-rabbit, Abcam™, Canada, 1/2000 dilution) for 30 minutes at room temperature in the dark and, afterwards, with Fluoroshield mounting medium (Abcam™, Canada) containing blue-fluorescent Di Aminido Phenyl lndol (DAPI). Immunofluorescence microscopy was performed on an Olympus BX 60 microscope (CytoVision® 7.5. software, Leica Biosystems). For each condition, six fields evenly distributed on the coverslip were analysed. NETs were identified as elongated extracellular DAPI-stained DNA fibers associated with myeloperoxydase using FIJI software as described by Brinkmann et al. and Arpinati et al. (Brinkmann et al., 2004; Arpinati et al., 2020).

NETs from vital NETosis were characterized by a positive DAPI staining and citrullinated histone H3 (CitH3): MPO+, CitH3+ (Erpenbeck et al., 2016; Arpinati et al., 2020). NETs from non-vital NETosis were deducted from the difference between total NETs (DAPI+) and NETs from vital NETosis (MPO+, CitH3+). Image analysis was performed blindly, without knowledge of the cell-processing conditions.

The same experiments as described previously were performed using diphenyleiodonium (DPI), a ROS pathway inhibitor before stimulation, at 10 μg/L, 1 h, at 37°C and 5% CO_2, to inhibit suicidal NETOsis (O’Donnell et al., 1993; Douda et al., 2015).

As vital NETosis mainly passes through the TLR2 and four pathway (Clark et al., 2007; Semeraro et al., 2011; Erpenbeck et al., 2016; Muñoz-Caro et al., 2021), we also performed the same experiments with anti-TLR2 and anti-TLR4 blocking antibodies (ANTI-HU CD284 HTA125 and ANTI-HU CD282 TL2.1 FG, Life Technologies SAS) at 5 μg/mL for 1 h at 37°C, 5% CO2.

The inhibitors were added to the culture medium in the wells with purified neutrophils before stimulating with plasma.

For each experiment using DPI or blocking antibodies, a non-stimulated well was prepared with dimethylsulfoxyde (DMSO), the solvent present in inhibition reagents, as negative controls.

Statistics were analysed with the GraphPad statistical software program (version 6.07. San Diego, United States) using a non-parametric Wilcoxon test for paired data and the non-parametric Mann-Whitney test for unpaired data. When multiple comparisons were performed, a Bonferroni correction was applied. Quantitative data were expressed as medians and interquartile ranges. Qualitative data were expressed as absolute number and frequency (%). A p-value <0.05 was considered statistically significant.

Twenty-six participants were analyzed in the study: 13 pregnant women with PE in Group PE, 13 pregnant women without complications in Group NP with maternal and gestational ages as matching variables (+/−2 years old and+/−1 week of gestation respectively) using the propensity score method.

The flowchart is detailed in Figure 1. The characteristics of the population are summarized in Table 1.

We found a significant increase in the number of total NETs in women with PE compared to women with normal pregnancies p = 0.03 (Figure 2A).

Following these results showing the formation of NETs with plasma from pregnant women with normal or preeclamptic pregnancies, we evaluated the different types of NETosis mechanisms implicated in these situations.

We first compared citrullinated NETs, corresponding to vital NETosis, between normal and preeclamptic pregnancies. We did not find any significant difference between the two groups (p = 0.96) (Figure 2B).

Then, interestingly, our results showed a significant difference between the number of uncitrullinated NETs, corresponding to non-vital NETosis, between women with normal pregnancies compared to women who developed preeclampsia, p = 0.019 (Figure 2C).

We therefore detected the presence of two different types of NETosis in the two situations studied (pictures 2D and 2E). We also observed that the greater part of NETosis in preeclampsia was represented by non-vital NETosis.

Suicidal NETosis described in the literature involves the ROS pathway. This NETosis will be called “ROS-dependent suicidal NETosis” for the rest of our study. To characterize this NETosis in our two groups, we first validated the model with a known activator of suicidal NETosis, PMA, in presence or not of a specific ROS inhibitor, DPI.

NETs from PMA induction were uncitrullinated NETs (p = 0.25) (Figure 3A) and this NETosis was completely inhibited by DPI (p = 0.0039) (Figure 3B). NETs from ROS-dependent suicidal NETosis may therefore be characterized by uncitrullinated NETs inhibited by DPI.

The inhibition by DPI of plasma-stimulated neutrophils from the women in our two groups resulted in a significant decrease in total NETs in the NP group (p = 0.006) and in the PE group (p = 0.001). Moreover, the significant difference in total NETs initially observed between normal and preeclamptic pregnancies (Figure 2A), was maintained in presence of DPI, which meant that total NETs were higher in Group PE than in Group NP (Figure 4A).

The same observation was made with uncitrullinated NETs. There is a significant decrease in the percentage of uncitrullinated NETs before and after inhibition in the NP and PE groups (p = 0.0012 and p = 0.0005 respectively). Again, the significant difference in the percentage of uncitrullinated NETs initially observed (Figure 2C) was maintained in DPI condition (Figure 4B).

In contrast, this inhibition did not result in a significant decrease in citrullinated NETs, either in Group NP (median without inhibition: 15.56% [12.52; 19.63], median with inhibition: 7.21% [1; 19.30], p = 0.15) or in Group PE (median without inhibition: 13.44% [9.63; 23.40], median with inhibition: 16.01% [11.68; 20.93], p = 0.34) confirming the link between ROS-dependent suicidal NETosis and uncitrullinated NETs (Figure 4C).

Thus, we confirmed the presence of ROS-dependent suicidal NETosis in normal pregnancy and in preeclampsia.

To confirm the presence of vital NETosis in normal pregnancy and in pregnancy complicated by preeclampsia, we performed experiments using anti-TLR2 and anti-TLR4 blocking antibodies. Inhibition by anti-TLR2 and anti-TLR4 blocking antibodies led to a significant decrease in the number of total NETs in the PE group (p = 0.0046 respectively) (Figure 5A). The NP group showed the same trend but with no significance, due to the heterogeneity of the values in presence of TLR antibodies (p = 0.043).

It also led to a significant decrease in citrullinated NETs, markers of vital NETosis, in the NP and PE groups (p = 0.021 and p = 0.0005 respectively) (Figure 5B).

However, this inhibition did not lead to a significant decrease in the number of uncitrullinated NETs either in the NP group (median before inhibition: 12.80% [9.05; 18.23], median after inhibition: 8.13% [6.55; 18.43], p = 0.74), or in the PE group (median before inhibition: 20.26% [14.87; 21.47], median after inhibition: 9.73% [4.79; 22.69], p = 0.11) (Figure 5C).

There was also no significant decrease after inhibition by anti-TLR blocking antibodies in total NET rates between NP and PE groups, p = 0.61 (Figure 5A), in citrullinated NET rates between the two groups, p = 0.31 (Figure 5B) or in uncitrullinated NETs p = 0.72 (Figure 5C).

These observations confirm the same induction of vital NETosis in normal and preeclamptic pregnancies as previously described.