Multiple cohorts of pregnant women at different stages of pregnancy were included in this study. The cohorts of patients are summarized in Supplementary Figure 1, Supplementary Tables 1–5. For each cohort, PE was defined as the new onset of hypertension (Systolic BP ≥ 140mmHg and diastolic BP ≥ 90mmHg at two occasions at least 4 hours apart following 20th week of gestation or a systolic BP ≥ 160mmHg and/or a diastolic BP ≥ 110mm Hg within short interval) and proteinuria (≥300 mg in a 24 hours urine collection, or a protein/creatinine ratio ≥ 0.3 or, a dipstick reading of ≥ 1+ after 20th week of gestation), or in the absence of proteinuria, new onset of thrombocytopenia (Platelet count < 1, 00, 000/µL), or renal insufficiency or impaired liver function or pulmonary edema, or cerebral, or visual symptoms and no other disease and therapy for it (4). For the healthy control (CTRL) group, the inclusion criteria were: no hypertension, no proteinuria, and no other comorbidity or therapy. The matched controls were selected based on age and number of pregnancies. Exclusion criteria for PE group included: chronic hypertension, renal diseases, collagen vascular diseases, cancer, thrombosis, multifetal gestation, diabetes, autoimmune disease, smokers, vesicular mole, infections (bacterial, fungal, viral), and in vitro fertilization. Exclusion criteria for the CTRL group comprised hypertension observed after 3 months postpartum, with any kind of ailment or disease, smokers, infections (bacterial, fungal, viral), multifetal gestation, vesicular mole and in vitro fertilization. This study did not include pregnant patients with HELLP (hemolysis, high liver enzymes, and low platelet count) syndrome, nor eclampsia. The cohort of pregnant women were selected from Maharaja Jitendra Narayan Medical College and Hospital (MJNMCH; Cooch Behar, West Bengal, India). Fetal villi samples (1 cm wide pieces) of term end placenta (3^rd trimester) were collected from PE as well as healthy mothers (CTRL) within an hour following the delivery. For histological profiling, tissues were formalin fixed. For mRNA studies, tissues were taken in RNA Liv solution (Hi Media) that helps in stabilizing the RNA in the tissue. For western blot (WB), tissues were snap frozen and stored at -80⁰ C for further experimentation.

Circulating exosomes were isolated at the National Institute for Research in Reproductive and Child Health, Mumbai, India, collected from the serum samples of early onset PE mothers (PE developed at <34 weeks of gestation) of Nowrosjee Wadia Maternity Hospital (NWMH), Mumbai, India.

Serum samples of PE patients (obtained at the time of PE diagnosis) and matched healthy pregnant women (CTRL, matched for mother and gestational age) were from Institute for Maternal and Child Health, IRCCS “Burlo Garofolo” (Trieste, Italy). Placental histological specimens were collected from both PE and healthy (CTRL) mothers (36).

STBMs, prepared via dual-placental perfusion of PE and healthy placentae, were obtained from the Nuffield Department of Obstetrics and Gynaecology, John Radcliffe Hospital, University of Oxford (United Kingdom), as previously described (33).

The Institutional Ethical Committee (IEC/CBPBU/ZOO/2020/001, MJNMC/IEC/77/2024, IECICMR-NIRRCH 320/2017 and IECNWMH/AP/2017/014) approved the study as per Indian Council of Medical Research guidelines (ICMR). Informed written consent forms were obtained from each patient included in the study prior to the collection of the samples. All the records identifying the patients were kept confidential.

Total RNA was isolated from placental tissues of PE and CTRL women with RNeasy Mini Kit (Qiagen, #74104) using silica membrane spin columns, according to manufacturer’s protocol. The purified RNA was quantified using NanoDrop^® spectrometer (Thermo scientific) at 260 nm. 1 ng of total RNA was used to synthesise first-strand cDNA using the Verso cDNA synthesis kit (Thermo Scientific, #AB-1453/A). Real-time quantitative PCR (RT-qPCR) was performed using Bio-Rad CFX Maestro (CFX ConnectTM Real-Time System) and SsoAdvanced Universal SYBR^® Green Supermix (Bio-Rad, #1725270). The reaction involved 40 cycles of denaturation (30 sec at 95 °C) and annealing (30 sec at 52 °C to 60 °C; annealing temperature of the primers). The mRNA expression levels of TNF-α, TGF-β1, VEGF, properdin, C3 and C5 were measured following normalizing with the expression of a housekeeping gene (GAPDH). The primer details are listed in Supplementary Table 6.

Briefly, 1 mg of washed dynabeads (Invitrogen; #14311D) were incubated with 30 µg of anti-placental alkaline phosphatase (anti-PLAP) antibody (Invitrogen, # MA1-19354) on an end-to-end rotator at 37°C overnight. Next day, the supernatant was discarded by placing the tubes on a magnetic stand and the antibody coupled beads were washed with HB, LB and SB buffers provided in the kit. Serum-derived exosomes (100 µl) from early onset PE (EOPE, n=4) and CTRL (n=4) groups were incubated with anti-human Fc receptor blocking reagent and anti-PLAP antibody bound beads at 4°C overnight. The bound PLAP⁺ exosomes were eluted using 180 µl of elution buffer (0.1M glycine-HCl, pH 2.8) and neutralized with 1M Tris-buffer, pH 8.0. The flowthrough containing PLAP-exosomes were then used for isolating HLA-G⁺ exosomes after incubation with Anti-HLA-G coated dynabeads at 4°C overnight. Finally, PLAP⁺ and HLA-G⁺ exosomes were pooled to obtain circulating placental exosomes. For the extraction of proteins, 250 μl of placental exosomal suspension from PE cases and healthy controls were lysed in 50 μl RIPA buffer (Sigma-Aldrich; #R0278). Lysates were centrifuged at 16000×g for 20 min at 4°C following 60 min incubation on ice with intermittent vortexing. The protein-containing supernatant was collected in a fresh sterile microcentrifuge tube and used for WB. Serum-derived from 3^rd trimester healthy pregnant women was used as a positive control.

Radioimmunoprecipitation assay lysis buffer (RIPA) containing a protease inhibitor cocktail (HiMedia; #ML051) was used to homogenise placental tissues (~40 mg). Homogenised lysates were centrifuged at 15000 rpm for 20 min at 4 °C. The protein was quantified using Folin-Lowry’s method. 40 µg of protein was loaded in each well and was resolved under reducing conditions via 8% v/v SDS-PAGE. The proteins were transferred onto a nitrocellulose membrane (Bio-Rad). The membrane was then blocked for 45 min at room temperature using 2% w/v bovine serum albumin (BSA; Sigma Aldrich; #A2153) in TBS containing 0.1% v/v Tween-20 (SRL; #65296) (TBST) for properdin, while 3% non-fat dry milk powder (SRL; #28582) in TBST was used in the case of C3, C5 and β-actin. The membrane was probed with mouse anti-human properdin monoclonal antibody (#HYB 3904, 0.9mg/ml in PBS, 1:2000 dilution in 1% BSA in TBST), rabbit anti-C3 polyclonal antibody (ABclonal A16781, 1:2000 dilution in TBST) and rabbit anti-C5/C5a polyclonal antibody (ABclonal A8104, 1:700 dilution in TBST), respectively, at 4 °C overnight. Rabbit polyclonal antibody against human β-actin (Abcam, ab-16039, 1:2000 dilution in TBS) was used as loading control. Membranes were washed in TBST and probed with HRP-conjugated goat anti-mouse IgG (H+L) (ABclonal AS003, 1:4000 dilution in 1% BSA in TBST) for properdin, and HRP-conjugated goat anti-rabbit IgG (H+L) (ABclonal AS014, 1:10000 dilution in TBST for C3, and 1:2000 dilution in TBST for C5) for 2h at room temperature followed by washing. Finally, immobilon western blotting substrate (Milipore; #WBLUF0100) was used to detect the protein bands. The WB experiments were repeated three times and representative images were selected for presentation. Intensity was measured using ChemiDocTM MP Imaging System (Bio-Rad) and densitometric analysis was carried out using ImageJ software.

For WB using circulating placental exosomes, approximately 50 µg of protein extracted from serum and circulating placental exosomes of PE and CTRL pregnant women were separated on 12% SDS-PAGE gradient gel at 100V (90 min). The gel was transferred to polyvinylidene difluoride (PVDF) membrane (Pall Life Sciences). After blocking the membrane with 3% BSA in PBS-T (PBS containing 0.1% Tween-20), the membrane was incubated overnight with mouse anti-human properdin primary antibody (1:1000; #HYB 3904). Next day, the membrane was washed with PBS-T and PBS for 5 min. This followed incubation with goat anti-mouse HRP-conjugate (1:5000; Invitrogen; #31430). After washing with PBS-T and PBS of 5 min each, the blot was developed. The immunoblot experiments were repeated three times and representative images were selected. The relative protein levels of properdin in individual EOPE and control placental exosome lysates were estimated by densitometric analysis using ImageJ software. CD9 (1:500; Abcam; #ab58989) was used as a loading control for the densitometric analysis, which is a standard marker of exosomes.

Placental STBMs, isolated from CTRL or PE patients, were prepared using a dual-placental perfusion system, as previously described (33, 48). 20 μg of STBMs, diluted in 2x Laemmli buffer, were subjected to a 10% v/v SDS-PAGE under reducing conditions. Proteins were transferred to a nitrocellulose membrane using the wet system of Mini Blot Module (Thermo Fisher Scientific). After blocking for 1h with 5% skimmed milk in TBST, the membranes were incubated overnight with primary antibodies at 4 °C: anti-human properdin (1:1000; Millipore; #MABF2127), and anti-β-actin (1:1000 dilution; Santa Cruz; #sc-8432). The following day, the membrane was incubated with anti-mouse LI-COR IRDye 800CW and anti-mouse LI-COR IRDye 680CW (1:10, 000 dilution; LICOR Biosciences, Lincoln, NE, USA) for 1h at room temperature. The fluorescence intensity was acquired in the Odyssey^® CLx near-infrared scanner (LI‐COR Biosciences, Lincoln, NE, USA) and normalized for anti-β-actin. Image acquisition, processing and data analysis were performed via Image Studio Ver 5.2 (LICOR Biosciences).

PE and CTRL placental cotyledons were sampled in the marginal zone bordering the yolk sac, maximum 4 h after birth, and the sample was washed with saline to remove blood before being fixed in 10% v/v buffered formalin, embedded in paraffin, and stored at 4 °C. Tissue sections of 3-4 µm were deparaffinized with xylene and rehydrated in decreasing concentrations of ethanol (100%, 95%, 70%) and dH₂O. Antigen unmasking was performed using citrate buffer (pH 6) for 20 min at 98 °C in a thermostatic bath. After neutralization of the endogenous peroxidase activity with 3% v/v H₂O₂ for 5 min, the sections were incubated in PBS + 2% w/v BSA for 30 min to block non-specific binding. Next, mouse anti-human properdin monoclonal antibody (1:50; Santa Cruz) was incubated at 4 °C. Staining was revealed via anti-mouse horseradish peroxidase (HRP)-conjugate (1:500 dilution, Sigma), incubated for 30 min at room temperature, and the substrate, 3-amino-9-ethylcarbazole (AEC, Vector Laboratories). Sections were counterstained with Mayer Haematoxylin (DiaPath, Italy) and examined under a Leica DM 2000 optical microscope. Kidney tissue was stained as a positive control. Images were collected using a Leica DFC 7000 T digital camera (Leica Microsystems, Wetzlar, Germany). To quantify the staining, H score was calculated via ImageJ.

Term placental tissues from PE and CTRL groups were fixed in 10% v/v buffered formalin, paraffin-embedded and stored at 4 °C. Tissue sections of 3-4 µm were deparaffinized with xylene for 30 mins twice and rehydrated with decreasing concentrations of ethanol (100%, 95%, 70%, 50% and 30%) for 10 min each. The antigen unmasking was performed using citrate buffer, pH 6 for 2 min at 98°C in a microwave. Sections were then washed five times in PBS buffer, pH 7.4 and incubated with PBS + 10% w/v BSA at 4 °C for 1.5 h to block non-specific binding. Next, mouse anti-human properdin monoclonal antibody (HYB 3904; 1:100) and rabbit polyclonal antibody against human cleaved caspase 3 (Cell Signaling Technology; 1:400 dilution) in 2% BSA in PBS were added to the sections, and incubated overnight at 4°C. Next day, after washing with PBS five times, sections were incubated overnight with goat anti-mouse Alexa FluorTM 488 (Thermo Fisher Scientific; 1:500 dilution) and chicken anti-rabbit Alexa FluorTM 647 (Thermo Fisher Scientific; 1:500 dilution) conjugates, respectively, at 4°C for 2 h, followed by 4’, 6-diamidino-2-phenylindole (DAPI) staining for 10 min in dark at room temperature. Following two 5-min washes with PBS, the sections were mounted with Dibutylphthalate polystyrene xylene (DPX) (Merck). Images were captured using fluorescence microscope (ZEISS Scope.A1).

PE placental tissues were fixed in 2.5% v/v glutaraldehyde (TAAB Laboratories, UK, #G002) and 2% v/v paraformaldehyde (Thermo Fisher) in 0.1 M phosphate buffer (pH 7.3) for 2 h at 4 °C. After washing in PBS, sections were fixed in 1% osmium tetroxide (TAAB Laboratories, UK, #O001) for 1 h and dehydrated in acetone (30%-50%-70%-80%-90%-95% twice for 10 min at 4°C). The samples were embedded in araldite CY212 (TAAB Laboratories, UK) and then polymerized at 60 °C for 72 h. Ultra-thin sections (60–70 nm) were stained with aqueous uranyl acetate and alkaline lead citrate (LADD Research, USA). The sections were observed under a Tecnai G2–20 S-twin transmission electron microscope (Fei, Eindhoven, Netherlands) at a magnification range of 2000–10000X. TEM study was carried out at the All India Institute of Medical Sciences, New Delhi, India.

For measuring the levels of properdin in PE and CTRL sera, human complement properdin ELISA kit (Hycult Biotech; # HK334) was used; instructions provided by the manufacturer were followed for the assay. The plate was read by the PowerWave X Microplate Reader (Bio-Tek Instruments) spectrophotometer at 450 nm.

For statistical analysis, the non-parametric Mann-Whitney U test was used to compare two groups (CTRL and PE) for IHC, WB (placental tissues, exosomes, and STBM analyses), and RT-qPCR. Matched data comparing PE vs healthy CTRL were analysed through paired Wilcoxon signed-rank test. Results were expressed as mean ± standard deviations. All statistical analyses were performed using GraphPad Prism software 9.0 (GraphPad Software Inc., USA). Results were considered statistically significant for all analyses when *p <0.05, **p <0.01.

As little information is available regarding properdin expression in placentae, we initially focused on mRNA expression in PE placentae compared with CTRL tissues via RT-qPCR. To assess the inflammatory condition of the PE placenta, we first examined the mRNA expression levels of three key inflammatory markers (i.e., TGF-β1, TNF-α and VEGF), which are commonly dysregulated in PE. The PE placentae, belonging to our cohort of patients, showed significantly higher levels of mRNA expression of TGF-β1 (6-fold, Figure 1A), TNF-α (3-fold, Figure 1B), and VEGF (3-fold) (Figure 1C), compared to CTRL. The mRNA levels of properdin in the same placentae were nearly 10-fold higher in PE compared to CTRL (Figure 1D). Since properdin is the only known positive regulator of the complement alternative pathway, we also analyzed the mRNA expression of C3 and C5 genes, which were found to be significantly over-expressed in PE placentae compared to the CTRL (Figures 1E, F).

Secondly, we analysed them at the protein level by WB. Our results confirmed what was observed for the properdin, C3, and C5 transcripts (Figure 2). The WB analysis revealed significantly higher levels of these complement proteins in PE placental tissues compared to CTRL (Figures 2A, B). Thus, properdin mRNA and protein levels remain significantly higher in PE placenta compared to CTRL (Figures 1D, 2A, B), along with the complement proteins, C3 and C5 (Figures 1E, F, 2A, B).

To investigate the distribution of properdin in placental tissues from both PE and normal pregnancies, we assessed the properdin levels in term placental tissues via IHC. Staining for properdin was detected in both CTRL and PE placenta, particularly within the fetal villi endothelium (indicated by the blue arrows) and in the syncytiotrophoblast monolayer (red arrows) (Figures 3A, B, Supplementary Figures 2A, B). However, the intensity of properdin staining was comparatively lower in CTRL vs PE samples (Figure 3C). The PE placentae showed quite a high level of properdin in the syncytial knots (black arrows), which were high in number, and also throughout the syncytiotrophoblast monolayer (red arrows) across almost all samples (Figure 3B; Supplementary Figure 2B).

Since properdin was highly concentrated in the syncytial knots of PE placentae, we decided to carry out TEM study of this region. The PE placenta exhibited intense haematoxylin nuclear staining at the placental knot regions (black arrow) under a bright-field microscope (Figure 3D). A high-resolution image of the syncytial knots in PE placenta stained deep red with anti-properdin antibody (black arrows), which is shown in Figure 3E. The TEM images revealed the presence of several pyknotic nuclei inside the syncytial knot structures in the PE placenta, which seems to be mostly at their late degenerative state (Figure 3F). The small round pyknotic nuclei were present in the syncytium with dense peripheral condensed heterochromatin. The heterochromatin was abundant compared to euchromatin and was surrounded by nuclear membrane, leaving little space in between the membranes.

Pyknotic nuclei can form during apoptosis as well as necrosis, and evidence of both of these processes was observed in the PE placenta (15, 49–51). Thus, to confirm the apoptotic characteristics of the nuclei in PE syncytial knots, we investigated the expression of cleaved caspase 3, an apoptotic protein marker. Although the data obtained were not quantitative, the PE placentae showed strong immunofluorescence with anti-cleaved caspase 3 antibody (red colour) compared to CTRL placentae (Figures 4A, B). The serial sections of PE placenta also exhibited higher immunofluorescence for anti-properdin (green colour) compared to CTRL placental sections (Figures 4C, D). High magnification images showed the presence of both cleaved caspase 3 (Figure 5A) and properdin proteins (Figure 5B) in the syncytial knot regions of the PE placenta, which also revealed high nuclear staining (DAPI).

Having noted high level of properdin mRNA and protein in the PE placental tissue, we sought to estimate the level of properdin protein in the PE serum by ELISA. Strikingly, properdin levels in the PE serum were significantly lower compared to healthy mothers (CTRL) (Figure 6A). Indeed, the properdin levels ranged between 7.08 and 18.87 µg/ml in PE sera (mean 13.14 ± 3.3), and 10.68 and 22.99 µg/ml (mean 16.26 ± 2.9) in healthy CTRL.

Placental extracellular vesicles are increasingly recognized as key mediators in the development of PE, even though the underlying mechanisms are not fully understood. Given that both placental STBMs and circulating placental exosomes were found to be significantly increased in PE mothers in our earlier studies, we decided to examine whether these microvesicles were differentially loaded with properdin in PE and CTRL. TEM images of the PE placenta revealed a substantial shedding of syncytiotrophoblast fragments as extracellular vesicles of varied sizes, along with cell debris and apoptotic nuclei, in the intervillous space. The exosomes (yellow arrowheads), STBMs (green arrowheads), and pyknotic nuclei (red arrowheads) were observed in the PE placenta exposed towards the intervillous space (Figure 6Bi). CTRL placenta showed an almost intact syncytiotrophoblast layer (outer layer facing intervillous space) and cytotrophoblast layer lying underneath (Figure 6Bii). The level of properdin protein in STBMs from PE mothers compared to CTRL showed no significant difference in WB analysis (Figure 6C). However, the properdin protein levels in circulating placental exosomes from PE mothers were significantly lower compared to those in CTRL (Figure 6D).