SPP1 integrates trophoblast-endothelial crosstalk via the αVβ3–PI3K/Akt–MMP-9 axis to govern vascular remodeling in preeclampsia

Abstract

Background

Preeclampsia (PE) is a major hypertensive disorder of pregnancy rooted in defective placental vascular remodeling. Osteopontin (SPP1) is implicated in placental development, but its role in orchestrating the critical paracrine dialogue between trophoblasts and endothelial cells remains undefined.

Methods

Placental SPP1 expression and microvessel density (MVD) were analyzed in PE and control pregnancies. An in vitro hypoxia-reoxygenation (H/R) model with SPP1 manipulation in HTR-8/SVneo trophoblasts was used to assess cell functions. Trophoblast-endothelial communication was evaluated via a Transwell co-culture system with HUVECs. Key signaling (integrin αVβ3–PI3K/Akt/FAK) and the MMP-9/TIMP-1 balance were examined. Serum SPP1, sFlt-1, and PlGF were measured. In vivo, we established an N-nitro-L-arginine methyl ester (L-NAME)-induced PE rat model, and evaluated the therapeutic effect of recombinant SPP1 supplementation on PE-like phenotypes, placental vascular remodeling and the underlying signaling changes.

Results

SPP1 was downregulated in PE placentas and positively correlated with MVD. H/R reduced SPP1 secretion, impairing trophoblast invasion/proliferation and suppressing endothelial tube formation via paracrine signaling. Mechanistically, trophoblast-derived SPP1 promoted angiogenesis by activating endothelial integrin αVβ3–PI3K/Akt/FAK signaling and restoring the MMP-9/TIMP-1 balance. Downregulated SPP1 disrupts EVT–endothelial paracrine signaling required for spiral artery‑like remodeling. In vivo, SPP1 supplementation ameliorated PE phenotypes, restored placental vascular density, and reversed signaling defects in a rat model. Clinically, serum SPP1 levels were reduced in PE and inversely correlated with the sFlt-1/PlGF ratio and disease severity, highlighting its predictive value.

Conclusion

SPP1 is a central paracrine mediator of trophoblast-endothelial crosstalk, and its hypoxia-induced reduced expression drives vascular dysfunction in PE via the integrin αVβ3–PI3K/Akt–MMP-9/TIMP-1 axis. Our findings nominate SPP1 as both a promising biomarker and a potential therapeutic target for PE.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11033-026-11867-y.

Introduction

Preeclampsia (PE) is a pregnancy-specific multisystem metabolic and vascular disorder centered on placental dysfunction, and represents a leading cause of increased maternal and perinatal morbidity and mortality [1]. It is widely accepted that inadequate trophoblast invasion and impaired spiral artery remodeling are key factors in the pathogenesis of PE. These defects lead to placental hypoxia and maldevelopment, triggering the release of anti-angiogenic factors, which in turn induce widespread endothelial dysfunction and a systemic inflammatory response, ultimately manifesting as multisystem clinical disease [2, 3].

Secreted phosphoprotein 1 (SPP1), is a pleiotropic extracellular matrix protein [4] that is expressed in the placenta under both normal and preeclamptic conditions. It critically supports trophoblast adhesion and motility, thereby contributing to placental development and function [5]. Previous studies have established a critical role for SPP1 in modulating trophoblast invasion and adhesion through interactions with integrins [6]. Notably, SPP1 also exerts potent pro-angiogenic effects: it engages with endothelial integrin receptors αVβ3 activating downstream signaling cascades that enhance tube formation, migration, and proliferative capacity [7]. Furthermore, SPP1 modulates matrix metalloproteinase activity, facilitating extracellular matrix degradation and remodeling to promote angiogenic expansion [8]. Given its dual role in modulating trophoblast behavior and endothelial function, along with its recognized capacity for integrin-mediated paracrine signaling in tumor angiogenesis—this process shares core mechanisms with placental angiogenesis including endothelial activation and ECM remodeling [7]—we hypothesize that SPP1 and its cognate integrin receptors form a critical signaling axis at the trophoblast–endothelial interface. Nevertheless, the spatiotemporal dynamics of SPP1 and integrin αVβ3 expression within the placental bed across normal and PE pregnancies, the precise mechanism of their interaction, and their collective impact on vascular remodeling processes remain poorly defined. Similarly, the pathophysiological contribution of downregulated SPP1 to aberrant spiral artery transformation in PE is not yet fully elucidated.

We hypothesized that SPP1, as a critical placenta-derived signaling molecule, mediates paracrine crosstalk between trophoblasts and endothelial cells. Of note, SPP1 is constitutively expressed in various systemic tissues including vascular endothelial cells under physiological conditions, but in the human placenta, SPP1 is predominantly localized to trophoblasts according to existing evidence [9]. Therefore, we proposed that the downregulation of trophoblast-derived SPP1 is a pivotal event in the aberrant vascular remodeling of PE, thereby linking placental ischemia to systemic vascular dysfunction. We aim to delineate the molecular mechanisms through which SPP1 regulates intercellular communication, endothelial homeostasis, and angiogenic programming, thereby uncovering its fundamental role and mechanistic underpinnings in PE-associated failure of placental vascular remodeling.

Of note, SPP1 is constitutively expressed in various systemic tissues including vascular endothelial cells under physiological conditions, but in the human placenta, SPP1 is predominantly localized to trophoblasts rather than vascular endothelial cells according to existing evidence.

Therefore, we proposed that the downregulation of trophoblast-derived SPP1 is a pivotal event in the aberrant vascular remodeling of PE, thereby linking placental ischemia to systemic vascular dysfunction.

Materials and methods

Clinical sample collection

Sample source:

The inclusion criteria for the experimental group included that the age of the participants were 25 to 40 years old, with clinical features meeting the diagnostic criteria for sPE according to the International Society of Hypertension Global Hypertension Practice Guidelines [10]. The clinical characteristics of the patients are shown in Table 1.

Table 1.

Clinical information on patients with PE and healthy controls

			Control (n = 30)	PE (n = 30)	t	p
Maternal Status	Basic Information	Age (years)	32.933 ± 2.8403	30.868 ± 2.9928	1.714	0.09
		Gestational weeks (weeks)	37(34, 38)	33(30.75, 34)	5.68	< 0.001
		Gravidity (times)	2(1, 3)	1(1,3)	2.227	0.029
		Parity (times)	1(0, 1)	0(0,1)	1.771	0.08
		BMI(kg/m²)	29.1376 ± 4.1393	30.4616 ± 3.9874	-1.476	0.541
	Vascular Parameters	Systolic BP(mmHg)	117.111 ± 11.9187	148.447 ± 13.8578	-11.077	< 0.001
		Diastolic BP(mmHg)	75.733 ± 10.1632	127.395 ± 15.57223	-2.222	0.029
	Cardiac Function	CK-MB(U/L)	12.7391 ± 4.18023	21.0393 ± 7.81311	-4.564	< 0.001
		CK(U/L)	40.926 ± 19.2519	66(47, 125)	-1.778	0.082
	Liver Function	AST(U/L)	18.71 ± 7.910	24.3 ± 11.332	-2.57	0.012
		ALT(U/L)	12.8 ± 11.964	20.57 ± 28.562	-1.623	0.109
	Renal Function	β2-microglobulin(mg/L)	2.39 ± 2.478	2.61 ± 0.613	-0.514	0.609
		Creatinine (Cr)(µmol/L)	49.75 ± 11.163	59.57 ± 13.270	-3.574	0.001
		Uric Acid (UA)(µmol/L)	316.85 ± 89.909	433.59 ± 106.488	-0.514	< 0.001
		Urine protein	2+(0+, 3+)	3+(2+, 3+)	-0.401	0.689
		24 h Urine protein(g/24 h)	0.24(0.1125, 1.00)	0.81(0, 4.86)	0.216	0.829
	Hematology	Platelet (PLT)(×109/L)	202.711 ± 67.2656	186.694 ± 74.6852	1.647	0.104
Neonatal Status	Basic Information	Apgar-1 min	9(8, 9)	8(7, 8)	3.874	< 0.001
		Apgar-5 min	10(9, 10)	9(8, 9)	4.419	< 0.001
		Apgar-10 min	10(9, 10)	9(8, 9)	3.980	< 0.001
		Birth weight (kg)	2927.78 ± 664.365	1693.16 ± 663.260	8.441	< 0.001

Inclusion criteria:

The inclusion criteria were based on the diagnostic criteria for PE from the Global Practice Guidelines for Hypertension [10].

Control group was selected by randomly assigning 30 healthy pregnant women with singleton mid-to-late pregnancies who had no pregnancy complications during the same hospitalization period as the case group.

Exclusion criteria:

Comorbidities such as diabetes, chronic hypertension, heart disease, severe liver/kidney disorders, other endocrine diseases, hematologic disorders, recent infections, or substance abuse.

Control group exclusion criteria included: infection history.

Specimen Processing:

Three to four pieces of placental tissues (each measuring approximately 1 cm×1 cm×1.5 cm) were rapidly collected from the maternal-facing basal plate adjacent to the umbilical cord insertion, avoiding areas of necrosis, calcification, and hemorrhage. After washing with saline solution, one of these tissue blocks was fixed in 4% formalin for Immunofluorescence(IF) assay, and the others were snap-frozen in liquid nitrogen and stored at -80 °C in a refrigerator overnight for later RNA extraction. Maternal venous blood samples were collected within 24 h before delivery. Samples were centrifuged immediately, and the supernatant was separated and stored at -80 °C until analysis.

Histological analysis

Immunofluorescence and image analysis:

Fresh placenta tissues were fixed in 4% paraformaldehyde/PBS for 48 h, dehydrated, paraffin-embedded, and sectioned at 5 μm. Post-deparaffinization/rehydration, heat-induced antigen retrieval was performed, followed by 1 h blocking with 5% BSA. Sections were incubated overnight at 4℃ with 5% BSA-diluted primary antibodies, PBS-washed, and incubated 1 h in the dark with species-matched fluorophore-conjugated secondary antibodies (5% BSA-diluted). After additional PBS washes, slides were DAPI-counterstained for 5–10 min, rewashed, and coverslipped with anti-fade medium. Images were captured via fluorescence microscopy. For quantitative analysis, five random non-overlapping fields per section were captured at 200× magnification. The mean fluorescence intensity (MFI) or integrated optical density (IOD)/area was quantified using ImageJ software after background subtraction. Primary antibodies used for immunofluorescence included anti-SPP1 (1:200, Abcam), anti-CD31 (1:200, Abcam), anti-HIF-1α (1:200, Cell Signaling Technology), anti-MMP-9 (1:200, Abcam), and anti-TIMP-1 (1:200, Abcam). Secondary antibodies included Alexa Fluor 488- and 594-conjugated goat anti-rabbit/mouse IgG (1:500, Invitrogen). Nuclei were counterstained with DAPI (Sigma-Aldrich). Images were captured using a fluorescence microscope (Olympus, Japan).

Microvessel density:

CD31-immunofluorescent sections were used to assess MVD (reflecting vessel density). High-MVD regions were first identified at 100× magnification, with vessel counting performed at 200×. Counts from 5 distinct fields were averaged to determine mean MVD.

Cell culture and treatment

Cell lines:

Human extravillous trophoblasts (HTR-8/SVneo; Ek-Bioscience, Shanghai) were cultured in RPMI-1640 (Gibco) supplemented with 10% foetal bovine serum (FBS; Viva Cell Biosciences) and 1% penicillin-streptomycin-glutamine (Solarbio) at 37℃/5% CO₂. Human umbilical vein endothelial cells (HUVEC; gifted by the Central Laboratory of Hebei Medical University Second Hospital) were cultured in DMEM (Gibco) with 5% FBS (Viva Cell Biosciences) and 1% penicillin-streptomycin-glutamine (Solarbio) under the same conditions.

Hypoxia-reoxygenation model:

Logarithmic phase HTR8/SVneo cells were seeded in 6-well plates and incubated 12 h in hypoxic environments (94% N2, 5% CO2, 1% O2) to simulate the ischaemic–hypoxic state in early pregnancy in vitro, and then reoxygenation (21% O2) for less than 72 h, a model that has been demonstrated to effectively replicate key aspects of PE pathophysiology [11].

Genetic manipulation and Cell transfection:

The SPP1 knockdown, overexpression as well as control plasmids were constructed by Mailgene (Beijing, China). HTR8/SVneo cells were transfected with equal amounts of plasmid and NC using Lipofectamine 3000 (Invitrogen, USA). The transfected cells were firstly incubated at 37℃, 5% CO2 for 6 h then replaced with fresh FBS medium.

Transwell Co-culture System:

Pretreated HTR-8/SVneo cells were seeded in Transwell inserts (0.4 μm pore), and HUVEC in lower chambers, co-cultured for 24 h under normoxia. They were cultured in 45%DMEM and 45% RPMI 1640 with 10% FBS and 1% penicillin/streptomycin.

Functional assays

Trophoblast function

Cell invasion:

HTR-8/SVneo cells post-treatment were seeded in Matrigel-precoated Transwell inserts (5 × 10⁴/insert). After attachment, inserts were placed in 24-well plates with medium. Following 24 h incubation, non-invading cells were swabbed off; invading cells on the lower membrane were methanol-fixed, crystal violet-stained, and imaged. Invading cells were counted in 5 random fields/insert via ImageJ.

CCK-8 assay:

HTR-8/SVneo post-co-culture were seeded in 96-well plates (5 × 10³/well), incubated with 10 µL/well CCK-8 reagent (absin, abs50003) for 2 h at 37 °C, and OD₄₅₀ measured by microplate reader (TECAN).

Endothelial cell function

Tube formation assay:

HUVEC post-co-culture were seeded on Matrigel (Corning, 3413), incubated for 6 h at 37 °C, imaged by inverted microscope (Carl Zeiss, Germany), and total tube length/branch points quantified using Image J.

Molecular biology assays

RNA isolation and qRT-PCR analysis

Total RNA from primary trophoblast tissue and cultured cells was isolated with TRIzol (Takara). 1 µg RNA was reverse transcribed into 20 µl cDNA using a qRT-PCR Kit (Takara), and mRNA levels were measured via LightCycler1480 (Roche) with PCR conditions: 95˚C for 30s (1 cycle); 95˚C for 5s, 60˚C for 20s (40 cycles); 95˚C for 5s; 60˚C for 60s; 95˚C (1 cycle). Relative expression was calculated using the 2-ΔΔCt method, with primers listed in Table 2.

Table 2.

Primers used for qRT-PCR

	SPP1	HIF-1α	GAPDH
Forward	5’- CAGTGATTTGCT TTTGCCTGAG − 3’	5’- TATGAGCCAGAAGAACTTTTAGGC − 3’	5’- GGAGCGAGATCCCTCCAAAAT − 3’
Reverse	5’- GGTGAGAATCATCAGTGTCCTGG − 3’	5’- CACCTCTTTTGGCAAGCATCCTG − 3’	5’- GGCTGTTGTCATACTTCTCATGG − 3’

Western blot:

Total proteins were extracted with RIPA buffer (Solarbio), quantified, separated on 10% SDS-PAGE gels, and transferred to PVDF membranes (Millipore). Membranes were blocked with rapid blocking buffer for 30 min, incubated overnight at 4 °C with primary antibodies, then 1 h with Dylight800-conjugated secondary antibody (1:5000; A23920, Abbine). Protein signals were visualized via the Odyssey SLX system (LLX-1215). Membranes were incubated overnight at 4 °C with the following primary antibodies: anti-SPP1 (1:1000, Abcam), anti-p-PI3K (1:1000, Cell Signaling Technology), anti-PI3K (1:1000, CST), anti-p-Akt (1:1000, CST), anti-Akt (1:1000, CST), anti-p-FAK (1:1000, CST), anti-FAK (1:1000, CST), anti-MMP-2 (1:1000, Abcam), anti-MMP-9 (1:1000, Abcam), anti-TIMP-1 (1:1000, Abcam), and anti-GAPDH (1:5000, Proteintech). Secondary antibodies included HRP-conjugated goat IgG (1:5000, MultiSciences). Protein bands were visualized using an ECL chemiluminescence detection system (Millipore).

ELISA:

SPP1 (supernatants/serum; R&D Systems, DY1433) and PlGF, sFlt-1 (Cloud-Clone Corp, Wuhan) were quantified via ELISA per manufacturers’ instructions. The sFlt-1/PlGF ratio was calculated as sFlt-1/PlGF concentration. Absorbance was measured at 450 nm with a microplate reader.

Animal model establishment and grouping

Animal selection:

Specific pathogen-free (SPF) grade Sprague-Dawley (SD) rats (female, 8–10 weeks old, 200–220 g; male, 10–12 weeks old, 250–280 g) were purchased from Beijing Sbeifu Biotechnology Co., Ltd. Animals were housed in a temperature-controlled environment (22 ± 2 °C) with 12 h light/dark cycle, free access to food and water, and acclimatized for 1 week before experimentation. All procedures were approved by the Animal Ethics Committee of The Second Hospital of Hebei Medical University and conducted in accordance with the Guide for the Care and Use of Laboratory Animals.

Preparation of gestational rat model of preeclampsia:

The rats were housed in a cage at a ratio of female to male rats = 2:1. The next day, vaginal secretions were collected and smears were observed, revealing a large number of epithelial cells arranged in a “leaved” pattern and a significant number of sperm, which was designated as day 0.5 of gestation. The pregnant rats were randomly divided into two groups: the control group (NC) and the PE group. The PE group (n = 6) received intraperitoneal injection of N-nitro-L-arginine methylesterhydrochloride (L-NAME) 200 µL (75 mg/kg/day) from day 8.5 to day 18.5 for a total of 10 days. The control group (n = 6) received intraperitoneal injection of 0.9% normal saline 200 µL/d from day 8 to day 18, at 8–9:00 AM daily for a total of 10 days. During this period, the rats were weighed every 2 days at the same time to record weight gain. Tail blood pressure was measured using a non-invasive sphygmomanometer on days 8.5, 10.5, 12.5, 14.5, 16.5, and 18.5, and qualitative urine protein tests were performed. The model was considered successful if the systolic blood pressure of the tail artery was 30mmHg higher than the baseline blood pressure or reached 140mmHg after drug administration.

Preparation and Administration of Recombinant SPP1:

Recombinant SPP1 protein was reconstituted in 10 mM phosphate buffer (pH 7.4) to a stock concentration of 0.1–1.0 mg/mL. Recombinant SPP1 was administered by intraperitoneal (i.p.) injection at a therapeutic dose of 1 mg/kg from gestational day 12.5 to 18.5, once daily. To avoid interference, recombinant SPP1 and L-NAME were injected with a 12-hour interval each day. The dose was selected based on our preliminary experiments (Fig. S4).

Pregnancy Outcomesss:

At 20.5 days of gestation, rats were administered intraperitoneal injection of 5% pentobarbital sodium (150 mg/kg) for anesthesia. Fetal rats and placentas were then retrieved, with records kept for fetal rat count, fetal rat weight, placental count, placental weight.

Statistical analysis

Statistical analyses were performed with GraphPad Prism 8, Excel, and ImageJ. Data are presented as mean ± SD. Normality was tested via Anderson-Darling or Shapiro-Wilk test. Two-group comparisons used two-tailed Student’s t test or Mann-Whitney U test; multigroup analyses employed one-way ANOVA with Tukey’s test, Kruskal-Wallis test, two-way ANOVA with Sidak’s test, or Scheirer-Ray-Hare test. Pearson’s correlation analyzed associations between serum SPP1 and clinical parameters. Significance was set at P < 0.05, denoted as *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, and ns (not significant).

Results

SPP1 expression is downregulated in preeclamptic placentas and correlates with hypoxic stress and impaired vascularization

To investigate the role of SPP1 in PE, we first examined its expression and localization in placental tissues from normal control (n = 30) and PE (n = 30) patients. IF staining revealed that SPP1 was primarily localized to extravillous trophoblast (EVTs) and perivascular regions in control placentas, and its expression was markedly reduced in PE placentas (Fig. 1A). Western blot and qRT-PCR confirmed that SPP1 expression was significantly downregulated in PE patients at both the protein (****P < 0.0001, Fig. 1B-C) and mRNA (P < 0.01, Fig. 1D) levels, compared with the control group.

Fig. 1.

SPP1 expression is downregulated in preeclamptic placentas and correlates with microvessel density and HIF-1α expression. A Representative immunofluorescence (IF) images showing the localization of SPP1 (green) in extravillous trophoblasts (EVTs) and perivascular regions of placental tissues from control (n = 30) and preeclampsia (PE, n = 30) groups. DAPI (blue) was used for nuclear staining. Scale bars: 20× magnification (Merge 20× panels), 40× magnification (Merge 40× panels, zoomed from red boxes). B Western blot analysis of SPP1 protein expression in control and PE placental tissues, with GAPDH as the internal reference. C-D Quantitative analysis of SPP1 protein (C) and mRNA (D) expression levels in control and PE groups, normalized to GAPDH. ****P < 0.0001 (C), P < 0.01 (D) vs. control group, n = 30 per group. E Representative double IF images showing co-localization of CD31 (green, vascular endothelial marker) and SPP1 (red) in control and PE placentas. DAPI (blue) was used for nuclear staining. Scale bars: 20× magnification (Merge 20× panels), 40× magnification (Merge 40× panels, zoomed from red boxes). F Quantitative analysis of CD31 relative fluorescence intensity in control and PE placentas. ***P < 0.001 vs. control group, n = 30 per group. G Representative double IF images of CD31 (green) and SPP1 (red) for MVD assessment in control and PE groups. DAPI (blue) was used for nuclear staining. Scale bars: 20× and 40× magnification as indicated. H Quantitative analysis of MVD in control and PE placentas. ****P < 0.0001 vs. control group, n = 30 per group. I Pearson correlation analysis showing a strong positive correlation between SPP1 expression levels and MVD in placental tissues (R = 0.99, P < 0.001). To eliminate gestational age bias (PE group: median 33 weeks; control group: median 37 weeks, P < 0.001), each group was stratified into 3 gestational age-matched subgroups, and 6 subgroup mean values were used for correlation analysis. J Representative double IF images showing co-localization of hypoxia-inducible factor-1α (HIF-1α, green) and SPP1 (red) in control and PE placentas. DAPI (blue) was used for nuclear staining. Scale bars: 20× and 40× magnification as indicated. K Quantitative analysis of HIF-1α relative fluorescence intensity in control and PE placentas. P < 0.01 vs. control group, n = 30 per group. L Pearson correlation analysis revealing a significant negative correlation between SPP1 expression and HIF-1α levels in placental tissues (R=-0.94, P = 0.0057). Consistent with the above strategy, 6 gestational age-matched subgroup mean values were used for analysis to eliminate gestational age confounding. Data are presented as mean ± SD. **P < 0.01, ***P < 0.001, **P < 0.0001 vs. control group

To assess the association between SPP1 and placental vascularization, we performed double IF staining for CD31 (a specific vascular endothelial marker) and SPP1. The results showed that SPP1 co-localized with CD31-positive vascular structures in control placentas, while this co-localization was significantly diminished in PE placentas (Fig. 1E, G). Quantitative analysis demonstrated a significant reduction in CD31 relative fluorescence intensity (***P < 0.001, Fig. 1F) and MVD (****P < 0.0001, Fig. 1H) in PE placentas compared with controls. Notably, there was a significant difference in gestational age between the preterm PE group (median 33 weeks) and term control group (median 37 weeks, P < 0.001). As gestational age is a critical biological confounder that directly modulates both placental SPP1 expression and MVD, we stratified each group into 3 gestational age-matched subgroups, generating 6 data points for correlation analysis to eliminate gestational age bias. Pearson correlation analysis indicated a strong positive correlation between SPP1 expression levels and MVD in placental tissues (R = 0.99, P < 0.001, Fig. 1I).

Due to impaired spiral artery remodeling, PE placentas exhibit inadequate perfusion and localized hypoxia. Consistent with this, double IF staining showed that hypoxia-inducible factor-1α (HIF-1α) expression was markedly elevated in PE placentas and co-localized with SPP1-positive areas (Fig. 1J). Quantitative analysis verified the significant upregulation of HIF-1α relative fluorescence intensity in the PE group (P < 0.01, Fig. 1K). Following the same gestational age stratification strategy to eliminate confounding, Pearson correlation analysis revealed a significant negative correlation between SPP1 expression and HIF-1α levels in placental tissues (R=-0.94, P = 0.0057, Fig. 1L).

Hypoxia suppresses SPP1 expression and secretion, impairing trophoblast invasion and proliferation

To simulate the hypoxic pathological state of the preeclamptic placenta, HTR-8/SVneo trophoblasts were subjected to H/R treatment. H/R treatment significantly suppressed SPP1 protein expression in HTR-8/SVneo cells (P < 0.01, Fig. 2A-B), mirroring its downregulation in clinical PE placentas. Subsequently, we performed genetic manipulation of SPP1 in HTR-8/SVneo cells: short hairpin RNA-mediated SPP1 knockdown (sh-SPP1) with a negative control (sh-NC), and SPP1 overexpression (SPP1-OE) with a negative control (SPP1-NC, hereafter referred to as OE-NC). We then evaluated the effects of altered SPP1 expression on trophoblast biological behaviors under normoxic and H/R conditions. In Transwell invasion assays, compared with normoxia, H/R treatment significantly reduced the invasiveness of HTR-8/SVneo cells (P < 0.05, Fig. 2C-D). Under H/R conditions (mimicking the PE pathological microenvironment), sh-SPP1 significantly decreased the number of invaded cells relative to sh-NC (P < 0.01, Fig. 2E), whereas SPP1-OE significantly enhanced invasiveness compared with OE-NC (P < 0.01, Fig. 2F). Consistently, cell proliferation was significantly impaired in the H/R group compared with the normoxia group (P < 0.0001 at 48 h, Fig. 2G). Under H/R conditions, sh-SPP1 further suppressed cell proliferation relative to sh-NC at all time points (P < 0.05, P < 0.001, Fig. 2H), while SPP1-OE notably increased the proliferation rate compared with OE-NC (P < 0.05 at 36 h, Fig. 2I).

Fig. 2.

Hypoxia-reoxygenation (H/R) suppresses SPP1 expression and secretion, and SPP1 modulates trophoblast invasion and proliferation. A Western blot analysis showing SPP1 protein expression in HTR-8/SVneo cells under normoxia (N) and H/R conditions. GAPDH (36 kDa) was used as the loading control. B Quantitative densitometry of SPP1 protein levels, normalized to GAPDH, demonstrating significant downregulation of SPP1 in H/R-treated cells (P < 0.01 vs. normoxia, n = 3 per group). C Representative images of Transwell invasion assays in HTR-8/SVneo cells under normoxia, H/R, H/R + sh-SPP1 (sh-SPP1), H/R + sh-NC (NC-sh), H/R+SPP1-OE (SPP1-OE), and H/R + OE-NC (SPP1-NC) conditions. Scale bar: 100 μm. D Quantitative analysis of Transwell invasion assays, showing significantly reduced invasive capacity in H/R-treated cells compared with normoxia (P < 0.05 vs. normoxia, n = 3 per group). E Quantitative analysis of invasion in sh-SPP1 vs. sh-NC groups under H/R conditions, showing significantly decreased invasiveness in sh-SPP1 cells (P < 0.01 vs. sh-NC, n = 3 per group). F Quantitative analysis of invasion in SPP1-OE vs. OE-NC groups under H/R conditions, showing significantly enhanced invasiveness in SPP1-OE cells (P < 0.01 vs. OE-NC, n = 3 per group). G CCK-8 assay results showing the proliferative capacity of HTR-8/SVneo cells under normoxia and H/R conditions at 12, 24, 36, 48, and 60 h, with significantly impaired proliferation in the H/R group (**P < 0.0001 vs. normoxia at 48 h, n = 6 per group). H CCK-8 assay results showing proliferation in sh-SPP1 vs. sh-NC groups under H/R conditions, with significantly suppressed proliferation in sh-SPP1 cells at all time points (*P < 0.05, P < 0.001 vs. sh-NC, n = 6 per group*). I CCK-8 assay results showing proliferation in SPP1-OE vs. OE-NC groups under H/R conditions, with significantly increased proliferation in SPP1-OE cells at 36 h (*P < 0.05 vs. OE-NC, n = 6 per group).Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Trophoblast-derived SPP1 mediates paracrine promotion of endothelial angiogenesis via integrin αVβ3

To investigate the paracrine effect of trophoblast-derived SPP1 on endothelial angiogenesis, we performed tube formation assays in HUVECs co-cultured with HTR-8/SVneo trophoblasts under normoxia and H/R conditions, with genetic manipulation of SPP1 (knockdown/overexpression). Unexpectedly, H/R treatment significantly increased the tube-forming capacity of HUVECs compared to normoxia, as demonstrated by elevated mesh points (*P < 0.001) and branching length (****P < 0.0001, Fig. 3A). Consistent with our hypothesis that SPP1 exerts a pro-angiogenic role, SPP1 knockdown (sh-SPP1, with negative control NC-sh) significantly impaired angiogenesis in HUVECs co-cultured under H/R conditions, with reduced mesh points (***P < 0.001) and branching length (****P < 0.0001) relative to NC-sh. Conversely, SPP1 overexpression (SPP1-OE, with negative control SPP1-NC) further enhanced angiogenic capacity, with significant increases in both mesh points (*P < 0.05) and branching length (*P < 0.05) compared to SPP1-NC (Fig. 3A). These results indicate that SPP1 retains a pro-angiogenic function even under H/R conditions, while the net pro-angiogenic effect of H/R suggests that additional hypoxia-driven pathways (e.g., VEGF signaling) dominate over SPP1 reduction in this acute setting. Enzyme-linked immunosorbent assay (ELISA) revealed that the level of soluble SPP1 secreted by HTR-8/SVneo cells was significantly reduced under H/R conditions compared to normoxia (P < 0.01, Fig. 3H). Further knockdown of SPP1 led to an additional decrease in SPP1 concentration in the supernatant (****P < 0.0001 vs. NC-sh), while SPP1 overexpression markedly increased its secretion (****P < 0.0001 vs. SPP1-NC, Fig. 3H). We next examined the expression of integrin αVβ3, a well-characterized receptor for SPP1, in HUVECs. Western blot analysis confirmed that integrin αVβ3 protein expression remained consistent across all experimental conditions (normoxia, H/R, SPP1 knockdown/overexpression), with no significant differences observed (all P > 0.05, ns, Fig. 3I-J). To functionally validate the role of integrin αVβ3 in SPP1-mediated angiogenesis, we treated HUVECs with a specific integrin αVβ3 inhibitor during co-culture with SPP1-overexpressing HTR-8/SVneo cells. Blocking integrin αVβ3 markedly attenuated the pro-angiogenic effect of SPP1 overexpression, as evidenced by significant reductions in both mesh points (P < 0.01) and branching length (P < 0.05) in HUVECs (Fig. 3K-M).

Fig. 3.

Trophoblast-derived SPP1 promotes endothelial angiogenesis via integrin αVβ3 in a paracrine manner. A Representative images of tube formation assays in human umbilical vein endothelial cells (HUVECs) co-cultured with HTR-8/SVneo trophoblasts under normoxia, hypoxia-reoxygenation (H/R), H/R + NC-sh (negative control for SPP1 knockdown), H/R+SPP1-sh (SPP1 knockdown), H/R+SPP1-NC (negative control for SPP1 overexpression), and H/R+SPP1-OE (SPP1 overexpression) conditions. Scale bar: 200 μm. B-C Quantitative analysis of tube formation assays: B Mesh points in normoxia vs. H/R groups (***P < 0.001 vs. normoxia, n = 3 per group); C Branching length in normoxia vs. H/R groups (****P < 0.0001 vs. normoxia, n = 3 per group). D-E Quantitative analysis of tube formation in SPP1-sh vs. NC-sh groups under H/R conditions: D Mesh points (***P < 0.001 vs. NC-sh, n = 3 per group); E Branching length (****P < 0.0001 vs. NC-sh, n = 3 per group). F-G Quantitative analysis of tube formation in SPP1-OE vs. SPP1-NC groups under H/R conditions: F Mesh points (*P < 0.05 vs. SPP1-NC, n = 3 per group); G Branching length (*P < 0.05 vs. SPP1-NC, n = 3 per group). H ELISA results showing the concentration of soluble SPP1 in the supernatant of HTR-8/SVneo cells under different conditions (normoxia, H/R, NC-sh, SPP1-sh, SPP1-NC, SPP1-OE). **P < 0.01 (normoxia vs. H/R), ***P < 0.0001 (NC-sh vs. SPP1-sh, SPP1-NC vs. SPP1-OE), n = 3 per group. I Western blot analysis showing integrin αVβ3 protein expression in HUVECs under different experimental conditions (normoxia, H/R, NC-sh, SPP1-sh, SPP1-NC, SPP1-OE). GAPDH (36 kDa) was used as the loading control. J Quantitative densitometry of integrin αVβ3 protein levels, normalized to GAPDH, demonstrating no significant differences across all groups (all *P > 0.05, ns, n = 3 per group). K Representative images of tube formation assays showing the effect of integrin αVβ3 inhibition on SPP1-mediated angiogenesis. HUVECs were co-cultured with SPP1-overexpressing HTR-8/SVneo cells (OE-SPP1) in the presence or absence of an integrin αVβ3 inhibitor (SPP1-sh + αvβ3), with OE-NC as the negative control. Scale bar: 200 μm. L-M Quantitative analysis of tube formation in the inhibitor experiment: L Mesh points (****P < 0.0001 vs. OE-NC, **P < 0.01 vs. OE-SPP1, n = 3 per group); M Branching length (**P < 0.01 vs. OE-NC, *P < 0.05 vs. OE-SPP1, n = 3 per group).Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns: not significant vs. corresponding control group

SPP1 activates the PI3K/Akt/FAK pathway and restores the MMP-9/TIMP-1 balance to facilitate angiogenesis

To elucidate the molecular mechanisms underlying SPP1’s pro-angiogenic effects, we first examined the activation status of the phosphatidylinositol 3-kinase/protein kinase B/focal adhesion kinase (PI3K/Akt/FAK) signaling pathway in HUVECs. Western blot analysis revealed that compared with the normoxia group, hypoxia-reoxygenation (H/R) treatment significantly reduced the phosphorylation levels of Akt, PI3K, and FAK (all P < 0.01, Fig. 4C, E, G), while the total protein levels of Akt, PI3K, and FAK remained unchanged (all *P > 0.05, ns, Fig. 4B, D, F). Consistent with the central role of SPP1 in this pathway, SPP1 knockdown (SPP1-sh, with negative control NC-sh) further exacerbated the H/R-induced suppression of Akt, PI3K, and FAK phosphorylation (P < 0.01, P < 0.001, Fig. 4C, E, G). In contrast, SPP1 overexpression (SPP1-OE, with negative control SPP1-NC) significantly reversed the H/R-induced dephosphorylation of these signaling molecules, restoring the activity of the PI3K/Akt/FAK axis (P < 0.05, Fig. 4C, E, G). Furthermore, SPP1-modulated signaling was closely associated with altered expression of vascular endothelial growth factor receptor 2 A (VEGFR2A) and angiopoietin-1 (Ang-1), as well as the anti-angiogenic factor soluble fms-like tyrosine kinase-1 (sFlt-1). Under H/R conditions, SPP1-sh significantly downregulated the expression of pro-angiogenic factors VEGFR2A and Ang-1 (P < 0.05, P < 0.01, Fig. 4I, J), while upregulating the expression of the anti-angiogenic factor sFlt-1 (P < 0.05, Fig. 4K). Conversely, SPP1-OE significantly upregulated VEGFR2A and Ang-1 expression (P < 0.01, P < 0.05, Fig. 4I, J) and downregulated sFlt-1 expression (P < 0.05, Fig. 4K) compared with SPP1-NC. Consistent with the functional changes in angiogenesis, SPP1 modulation also altered the expression of matrix metalloproteinases (MMPs) and their tissue inhibitors (TIMPs). H/R treatment significantly reduced the expression of matrix metalloproteinase 2 (MMP2) and matrix metalloproteinase 9 (MMP9), while increasing the expression of TIMP metallopeptidase inhibitor 1 (TIMP-1) (P < 0.05, P < 0.01, Fig. 4M, N, O). SPP1-sh further exacerbated the H/R-induced reduction in MMP2/MMP9 and increase in TIMP-1 (P < 0.05, P < 0.01, Fig. 4M, N, O), whereas SPP1-OE reversed these trends, significantly restoring MMP2/MMP9 levels and reducing TIMP-1 expression (P < 0.05, P < 0.001, Fig. 4M, N, O).

Fig. 4.

SPP1 activates the PI3K/Akt/FAK pathway and regulates the MMP-9/TIMP-1 balance in HUVECs. A Western blot analysis showing the protein expression levels of total and phosphorylated PI3K/Akt/FAK signaling components in HUVECs under normoxia, H/R, H/R + NC-sh (negative control for SPP1 knockdown), H/R+SPP1-sh (SPP1 knockdown), H/R+SPP1-NC (negative control for SPP1 overexpression), and H/R+SPP1-OE (SPP1 overexpression) conditions. GAPDH (36 kDa) was used as the loading control. B-D Quantitative densitometry of (B) p-Akt/Akt, (C) p-PI3K/PI3K, and (D) FAK protein levels. **P < 0.05, **P < 0.01, *P < 0.001 vs. NC-sh (for sh-SPP1) or SPP1-NC (for SPP1-OE); ns: not significant, n = 3 per group. E-G Quantitative densitometry of (E) p-FAK/FAK protein levels. **P < 0.05, P < 0.01 vs. NC-sh (for sh-SPP1) or SPP1-NC (for SPP1-OE); ns: not significant, n = 3 per group. H Western blot analysis showing the expression of vascular endothelial growth factor receptor 2 A (VEGFR2A), angiopoietin-1 (Ang-1), soluble fms-like tyrosine kinase-1 (sFlt-1), and GAPDH (36 kDa) in HUVECs under different experimental conditions. I-K Quantitative densitometry of (I) VEGFR2A, (J) Ang-1, and (K) sFlt-1 protein levels. **P < 0.05, P < 0.01 vs. NC-sh (for sh-SPP1) or SPP1-NC (for SPP1-OE); ns: not significant, n = 3 per group. L Western blot analysis showing the expression of matrix metalloproteinase 2 (MMP2), matrix metalloproteinase 9 (MMP9), TIMP metallopeptidase inhibitor 1 (TIMP-1), and GAPDH (36 kDa) in HUVECs under different experimental conditions. M-O Quantitative densitometry of (M) MMP2, (N) MMP9, and (O) TIMP-1 protein levels. **P < 0.05, **P < 0.01, *P < 0.001 vs. NC-sh (for sh-SPP1) or SPP1-NC (for SPP1-OE); ns: not significant, n = 3 per group.Data are presented as mean ± SD

Circulating SPP1 levels are reduced in preeclampsia and correlate with disease severity and the sFlt-1/PlGF ratio

We further analyzed maternal plasma samples collected in the third trimester using ELISA. The results revealed that circulating SPP1 levels were significantly lower in patients with PE compared to normotensive controls (****P < 0.0001, Fig. 5A). Subsequent correlation analyses demonstrated that maternal plasma SPP1 levels were strongly negatively correlated with both systolic blood pressure (r=-0.5079, P < 0.0001, Fig. 5B) and diastolic blood pressure (r=-0.5390, P < 0.0001, Fig. 5C) in the overall cohort. Additionally, SPP1 levels showed a significant positive correlation with neonatal birth weight (r = 0.6395, P < 0.0001, Fig. 5D). We also measured the soluble Fms-like tyrosine kinase-1 to placental growth factor ratio (sFlt-1/PlGF ratio), a well-established clinical biomarker for PE severity. As expected, the sFlt-1/PlGF ratio was significantly elevated in PE patients. Notably, maternal plasma SPP1 levels were inversely correlated with the sFlt-1/PlGF ratio (r=-0.6069, P < 0.0001, Fig. 5E).

Fig. 5.

Circulating SPP1 levels are reduced in PE patients and correlate with disease severity and the sFlt-1/PlGF ratio. A ELISA results showing maternal plasma SPP1 levels in normotensive control (n = 30) and preeclampsia (PE, n = 30) groups. ****P < 0.0001 vs. control group. B Pearson correlation analysis showing a significant negative correlation between maternal plasma SPP1 levels and systolic blood pressure in the overall cohort (r=-0.5079, P < 0.0001, n = 60). C Pearson correlation analysis showing a significant negative correlation between maternal plasma SPP1 levels and diastolic blood pressure in the overall cohort (r=-0.5390, P < 0.0001, n = 60). D Pearson correlation analysis showing a significant positive correlation between maternal plasma SPP1 levels and neonatal birth weight in the overall cohort (r = 0.6395, P < 0.0001, n = 60). E Pearson correlation analysis showing a significant negative correlation between maternal plasma SPP1 levels and the soluble Fms-like tyrosine kinase-1/placental growth factor (sFlt-1/PlGF) ratio in the overall cohort (r=-0.6069, P < 0.0001, n = 60). Data are presented as mean ± SD. ****P < 0.0001 vs. control group

To further explore whether the observed correlations were driven by group-specific effects, we performed stratified correlation analyses in the control and PE groups separately. No significant correlations between circulating SPP1 levels and clinical parameters were observed in the control group, while a significant positive correlation between SPP1 and neonatal birth weight was detected specifically in the PE group (R = 0.47, p = 0.0022, Supplementary Fig. 3). These results indicate that the overall correlations in the combined cohort are primarily driven by the PE group, suggesting a specific role of SPP1 in the pathophysiology of preeclampsia and its associated fetal outcomes.

Effects of SPP1 supplementation on vascular remodeling and related molecules in a rat model of preeclampsia

To further verify the role of SPP1 in PE-related vascular remodeling, we established a rat model of PE and performed SPP1 supplementation experiments. Compared with the Control group, PE model rats exhibited significantly elevated systolic blood pressure and urinary protein levels from mid-gestation onward (all ****P < 0.0001 vs. Control, Fig. 6A-B). Concurrently, PE rats showed significantly decreased fetal weight (****P < 0.0001 vs. Control, Fig. 6C) and placental weight (P < 0.01 vs. Control, Fig. 6D). Notably, supplementation with recombinant SPP1 partially ameliorated these maternal and fetal phenotypes: systolic blood pressure and urinary protein were significantly reduced compared with the PE group (all *P < 0.05 vs. PE), while fetal weight and placental weight were significantly restored toward normal levels (*P < 0.05, **P < 0.01 vs. PE, Fig. 6A-D). It should be noted that while SPP1 supplementation improved these parameters, they remained significantly different from the Control group, indicating partial rather than complete reversal of the PE phenotype.

Fig. 6.

SPP1 supplementation ameliorates PE phenotypes and restores placental vascular density in a rat model of PE. A Systolic blood pressure of rats in Control, PE, and PE+SPP1 groups measured on days 8.5, 10.5, 12.5, 14.5, 16.5, and 18.5 of gestation. ****P < 0.0001 vs. Control group, n = 8 per group. B Urinary protein levels of rats in each group measured at the same gestational time points. ****P < 0.0001 vs. Control group, n = 8 per group. C Fetal weight in each group at gestational day 18.5. ****P < 0.0001 vs. Control, **P < 0.01 vs. PE, n = 8 per group. D Placental weight in each group at gestational day 18.5. **P < 0.01 vs. Control, *P < 0.05 vs. PE, n = 8 per group. E Representative immunofluorescence images showing SPP1 (red) and CD31 (green) expression in placental tissues of rats in each group. DAPI (blue) was used for nuclear staining. Scale bar: 20× magnification. F Quantitative analysis of relative fluorescence intensity of CD31 (a marker of MVD) in rat placental tissues. ****P < 0.0001 vs. Control, **P < 0.01 vs. PE, ns: not significant (PE+SPP1 vs. Control), n = 8 per group. G Quantitative analysis of relative fluorescence intensity of SPP1 in rat placental tissues. ****P < 0.0001 vs. Control, ***P < 0.001 vs. PE, ns: not significant (PE+SPP1 vs. Control), n = 8 per group. Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns: not significant

Immunofluorescence staining of rat placental tissues confirmed that, compared with the control group, the expression of SPP1 in PE rat placentas was significantly decreased (****P < 0.0001 vs. Control, Fig. 6E, G), and MVD, labeled by cluster of differentiation 31 (CD31)) was also significantly reduced (****P < 0.0001 vs. Control, Fig. 6E, F). After SPP1 supplementation, SPP1 expression in PE rat placentas was significantly upregulated (***P < 0.001 vs. PE, Fig. 6G), and CD31-labeled MVD was significantly increased (**P < 0.01 vs. PE, Fig. 6F), with both parameters approaching control levels.

Effects of SPP1 on angiogenic factor expression and vascular homeostasis in a rat model of preeclampsia

Western blot results showed that, compared with the Control group, the expression of SPP1 and the phosphorylation levels of phosphatidylinositol 3-kinase (PI3K), protein kinase B (Akt), and focal adhesion kinase (FAK) in placental tissues of preeclampsia (PE) model rats were significantly decreased (SPP1: P < 0.01 vs. Control; p-Akt: P < 0.01 vs. Control; p-FAK: P < 0.05 vs. Control, Fig. 7A-B, D, H). After SPP1 supplementation, SPP1 expression and the phosphorylation levels of PI3K, Akt, and FAK were significantly increased and partially restored toward normal levels (SPP1: *P < 0.05 vs. PE; p-Akt: *P < 0.05 vs. PE; p-FAK: *P < 0.01 vs. PE**, Fig. 7A-B, D, H), indicating significant improvement in the activation of the PI3K/Akt/FAK signaling pathway. Total protein levels of PI3K, Akt, and FAK remained unchanged across all groups (all *P > 0.05, ns, Fig. 7E, F, G). Meanwhile, compared with the Control group, the expression of matrix metalloproteinase 9 (MMP-9) in placental tissues of PE rats was significantly downregulated (P < 0.05 vs. Control), while the expression of tissue inhibitor of metalloproteinases 1 (TIMP-1) was significantly upregulated (***P < 0.0001 vs. Control), resulting in an imbalance of the MMP-9/TIMP-1 ratio. SPP1 supplementation partially reversed this imbalance: MMP-9 expression was markedly increased (P < 0.01 vs. PE), and TIMP-1 expression was significantly decreased (P < 0.01 vs. PE), with the MMP-9/TIMP-1 ratio partially restored to a level close to that of the Control group (Fig. 7I-L). The expression of MMP-2 was also significantly reduced in PE rats and partially restored by SPP1 supplementation (*P < 0.05 vs. Control, *P < 0.01 vs. PE, Fig. 7I-J).

Fig. 7.

SPP1 supplementation restores the PI3K/Akt/FAK signaling pathway and corrects the MMP-9/TIMP-1 imbalance in the placenta of PE rats. A Western blot analysis showing the protein expression levels of SPP1, total PI3K, phosphorylated PI3K (p-PI3K), total Akt, phosphorylated Akt (p-Akt), total FAK, and phosphorylated FAK (p-FAK) in placental tissues of rats in Control, PE, and PE+SPP1 groups. GAPDH (36 kDa) was used as the loading control. B Quantitative densitometry of SPP1 protein expression, normalized to GAPDH. *P < 0.01 vs. Control, P < 0.05 vs. PE, n = 6 per group. C Quantitative densitometry of p-PI3K protein expression, normalized to GAPDH. ns: not significant, n = 6 per group. D Quantitative densitometry of p-Akt protein expression, normalized to GAPDH. *P < 0.01 vs. Control, P < 0.05 vs. PE, n = 6 per group. E Quantitative densitometry of total Akt protein expression, normalized to GAPDH. ns: not significant, n = 6 per group. F Quantitative densitometry of total PI3K protein expression, normalized to GAPDH. ns: not significant, n = 6 per group. G Quantitative densitometry of total FAK protein expression, normalized to GAPDH. ns: not significant, n = 6 per group. H Quantitative densitometry of p-FAK protein expression, normalized to GAPDH. *P < 0.05 vs. Control, P < 0.01 vs. PE, n = 6 per group. I Western blot analysis showing the protein expression levels of matrix metalloproteinase 2 (MMP2), matrix metalloproteinase 9 (MMP9), and tissue inhibitor of metalloproteinases 1 (TIMP-1) in placental tissues of rats in each group. GAPDH (36 kDa) was used as the loading control. J Quantitative densitometry of MMP2 protein expression, normalized to GAPDH. *P < 0.05 vs. Control, P < 0.01 vs. PE, n = 6 per group. K Quantitative densitometry of MMP9 protein expression, normalized to GAPDH. *P < 0.05 vs. Control, P < 0.01 vs. PE, n = 6 per group. L Quantitative densitometry of TIMP-1 protein expression, normalized to GAPDH. ****P < 0.0001 vs. Control, P < 0.01 vs. PE, n = 6 per group.Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns: not significant. Statistical comparisons are shown for Control vs. PE, and PE vs. PE+SPP1

Discussion

During normal pregnancy, EVTs invade the uterine myometrial inner layer and remodel spiral arteries into high-flow, low-resistance vessels to ensure placental perfusion, whereas defective EVT invasion-induced impaired spiral artery remodeling causes placental ischemia and systemic endothelial dysfunction, a core pathological feature of PE [12]. A previous study has confirmed SPP1 downregulation in PE placentas, linked its downregulation to vascular remodeling defects [13]and identified it as a potential early-onset PE predictive biomarker via bioinformatic analyses [14], but these findings only establish phenotypic and predictive associations, without elucidating SPP1’s function as a paracrine mediator at the trophoblast-endothelial interface or its underlying molecular regulatory mechanisms. Building on these findings, our study for the first time reveals, via clinical, in vitro and in vivo triple evidence, that SPP1 is a key paracrine mediator orchestrating trophoblast-endothelial crosstalk: it synergistically promotes EVT invasion and endothelial angiogenesis to maintain placental vascular homeostasis by activating the αVβ3–PI3K/Akt pathway and regulating the MMP-9/TIMP-1 balance, and hypoxia-induced SPP1 downregulation is a central molecular event driving abnormal placental vascular remodeling in PE. This work fills the critical gap in understanding SPP1’s molecular mechanism in PE vascular dysfunction, elevates SPP1 from a mere PE-associated and predictive biomarker to a core regulatory molecule of placental vascular remodeling, and provides a novel potential target for PE diagnosis and targeted therapy.

Downregulated SPP1 links placental hypoxia to trophoblast dysfunction

SPP1 is a critical molecule regulating EVT function. In clinical PE placental tissues, SPP1 expression was significantly downregulated and positively correlated with MVD, consistent with the findings of Kramer et al. [15]. Using the H/R model to mimic the hypoxic pathological microenvironment of PE placentas, we found that H/R impaired the invasion and proliferation of HTR-8/SVneo trophoblasts; SPP1 knockdown exacerbated these impairments, while SPP1 overexpression restored trophoblast invasive capacity and proliferative activity. These results confirm that SPP1, a trophoblast-secreted molecule, directly modulates core EVT functions by binding to surface integrin αVβ3 to activate downstream signaling—enhancing EVT invasion into spiral arteries and sustaining trophoblast proliferation to provide sufficient functional cells for vascular remodeling.

SPP1 and the hypoxia marker HIF-1α were negatively correlated and colocalized in PE placentas, supporting hypoxia as a core driver of placental dysfunction in PE [16, 17]. Defective spiral artery remodeling induces placental hypoperfusion and HIF-1α-mediated hypoxic adaptive responses; while HIF-1α activates pro-angiogenic target genes for compensatory angiogenesis, this mechanism fails in PE due to reduced SPP1 expression, and HIF-1α may further inhibit SPP1 expression to exacerbate vascular dysfunction [18, 19]. Notably, the regulatory relationship between HIF-1α and SPP1 is cell context-dependent [20], which underlies the complex placental responses to ischemic stress and may explain divergent findings across experimental models [21]. SPP1 and HIF-1α colocalization in EVTs forms a reciprocal vicious cycle: hypoxia inhibits SPP1 expression via HIF-1α activation, and downregulated SPP1 impairs trophoblast hypoxic adaptability; this in turn aggravates hypoxia by inducing vascular remodeling defects, a regulatory pattern consistent with previous observations [18, 19]. Additionally, hypoxia-regulated non-coding RNAs post-transcriptionally target SPP1 [22], adding another layer of complexity to SPP1 modulation in PE. Collectively, these data identify SPP1 as a key molecular node linking hypoxic stress to trophoblast dysfunction, whose downregulation impairs EVT invasion and directly hinders spiral artery remodeling.

SPP1 integrin signaling maintains endothelial angiogenic competence and ECM homeostasis

Endothelial cells form the structural basis of vascular remodeling, with their angiogenic capacity governing placental microvascular network formation and perfusion efficiency, and SPP1 regulates endothelial cell function by activating the PI3K/Akt/FAK signaling pathway and MMP-9/TIMP-1 balance in a paracrine-dependent manner. As the core endothelial receptor mediating SPP1 biological effects [7, 23], integrin αVβ3 regulates endothelial tube formation for placental vascular network establishment; our study confirmed its stable expression on HUVECs under normoxia, H/R, and co-culture with SPP1-modified trophoblasts, ruling out receptor dysregulation and identifying reduced SPP1 ligand levels as the primary driver of PE-related endothelial defects. Functionally, SPP1 binds integrin αVβ3 to activate the PI3K/Akt/FAK pathway, which enhances endothelial proliferation and migration to promote tube formation and modulates angiogenic factor balance—H/R and SPP1 knockdown reduced pro-angiogenic Ang-1 and increased anti-angiogenic sFlt-1, while SPP1 overexpression reversed this imbalance, consistent with the pathway’s established role in sustaining endothelial survival and vascular integrity during placental development [24].

Vascular remodeling relies on ordered ECM degradation and reorganization, and the MMP‑9/TIMP‑1 ratio serves as a key measure of ECM remodeling capacity [25, 26]. Our study demonstrated that inhibition of the SPP1–PI3K/Akt/FAK axis reduced MMP‑9 activity and increased TIMP‑1 expression, while SPP1 overexpression restored the MMP‑9/TIMP‑1 balance. These findings support a critical role for SPP1 in maintaining endothelial vascular network formation by fine‑tuning ECM remodeling efficiency, a process that is severely impaired in PE [27]. Of note, an apparent discrepancy was observed in our functional assays: hypoxia‑reoxygenation (H/R) promoted endothelial tube formation despite reducing SPP1 expression. This is not contradictory. Hypoxia‑reoxygenation is a potent inducer of multiple pro‑angiogenic factors, including HIF‑1α target genes such as VEGF and ANGPT2, which can collectively enhance angiogenesis even when a single factor like SPP1 is decreased. Importantly, SPP1 knockdown still suppressed angiogenesis under H/R, confirming that SPP1 is necessary for full angiogenic capacity. Thus, the H/R‑induced increase in overall angiogenic activity represents the net outcome of competing pathways, with global pro‑angiogenic signals outweighing the partial loss of SPP1. In line with previous studies [25], MMP/TIMP imbalance is closely associated with defective vascular remodeling in placental ischemia. SPP1‑mediated endothelial regulation is also modulated by hypoxic stress: high HIF‑1α expression in PE placental endothelial cells may suppress SPP1 and weaken its protective effects. SPP1 downregulation further compromises endothelial tube formation and ECM remodeling, ultimately leading to microvascular rarefaction and impaired vascular remodeling—consistent with the local hypoxia–angiogenesis vicious cycle underlying PE pathogenesis [28].

SPP1 as a critical paracrine mediator of trophoblast-endothelial crosstalk

Physiologically, SPP1 exhibits stage-specific expression during pregnancy: it is highly expressed in the first trimester to mediate embryo implantation and early trophoblast invasion for initial placental vascular construction, and its expression is adaptively upregulated in the second and third trimesters to sustain the maturation and homeostasis of the placental vascular network [15, 29]. Clinical serum data from our study further support the physiological and pathological importance of SPP1: serum SPP1 was significantly lower in PE patients, and its levels were correlated with key indicators of PE severity, including maternal blood pressure, neonatal birth weight, and the sFlt-1/PlGF ratio (a gold-standard marker of PE vascular dysfunction) [30, 31].

Beyond its stage-specific expression and clinical relevance, SPP1 plays a pivotal role as a paracrine mediator in trophoblast-endothelial crosstalk—an essential process for placental vascular remodeling. Our Transwell co-culture assays demonstrated that trophoblast-secreted SPP1 regulates endothelial function in a paracrine manner, and impaired paracrine signaling contributes to abnormal vascular remodeling in PE. Mechanistically, H/R-treated trophoblasts secreted less SPP1 and inhibited HUVEC tube formation, whereas exogenous SPP1 or SPP1 overexpression significantly restored angiogenic activity; these findings are consistent with studies showing that hypoxia disrupts trophoblast-endothelial communication in PE [32, 33] and that SPP1 maintains vascular homeostasis across systems [34, 35]. SPP1 acts through the integrin αVβ3–PI3K/Akt/FAK axis to synchronously coordinate EVT invasion, endothelial angiogenesis, and ECM remodeling, and inhibition of αVβ3 abolished the pro-angiogenic effect of SPP1 overexpression. These results extend the findings of Ke et al. [36] and identify integrin αVβ3 as a common receptor for SPP1-mediated crosstalk between trophoblasts and endothelial cells.

Serum SPP1 therefore represents a promising biomarker for impaired trophoblast-endothelial crosstalk and defective vascular remodeling in PE. Furthermore, the ability of exogenous SPP1 to rescue endothelial dysfunction supports its potential as a therapeutic target, and restoring SPP1-mediated paracrine signaling may represent a novel strategy to improve vascular remodeling in PE.

In vivo validation and therapeutic potential of SPP1 in PE

The translational relevance of our in vitro findings is supported by our in vivo experiments using a rat model of PE. The L-NAME-induced PE model successfully recapitulated key clinical features, including hypertension, proteinuria, and fetal growth restriction, paralleled by a significant reduction in placental SPP1 expression and MVD. Critically, supplementation with recombinant SPP1 protein during gestation not only ameliorated these core maternal and fetal phenotypes but also restored placental vascular density. This direct in vivo evidence firmly establishes that downregulated SPP1 is not merely a correlative biomarker but a contributory factor to the PE syndrome, and more importantly, that its restoration can therapeutically reverse key aspects of the disease.

At the molecular level, the PE rat model mirrored our cellular observations: placental PI3K, Akt and FAK phosphorylation was suppressed, with an MMP-9/TIMP-1 imbalance that inhibited extracellular matrix remodeling. Exogenous SPP1 administration rescued these signaling defects, reinstating integrin αVβ3–PI3K/Akt/FAK pathway activity and correcting the MMP-9/TIMP-1 imbalance. These in vivo data confirm that the SPP1-integrin αVβ3 signaling axis identified in vitro is functionally active in the complex tissue microenvironment of the compromised placenta. They bridge cellular mechanisms and whole-organism PE pathophysiology, providing compelling proof-of-concept that targeting the SPP1 pathway has therapeutic potential for mitigating systemic PE manifestations.

Study limitations

The present study only collected serum and placental samples from the third trimester of pregnancy, thus failing to elucidate the dynamic changes of SPP1 expression across the entire gestational period. Future longitudinal studies including first, second and third trimester samples are needed to fully characterize the gestational regulation of SPP1 in normal and PE pregnancies. Nevertheless, our findings firmly confirm the pathological relevance of SPP1 downregulation in late-onset PE, which is well-supported by clinical, in vitro and in vivo evidence.

Conclusions

In summary, our study delineates a novel pathway wherein hypoxia-driven loss of SPP1 impairs integrin αVβ3-mediated activation of the PI3K/Akt/FAK pathway, which in turn reduces MMP-9 activity and increases TIMP-1 expression, leading to dysregulated MMP-9/TIMP-1 balance and ultimately defective vascular remodeling in PE. This mechanistic cascade is supported by evidence spanning clinical samples, in vitro co-culture models, and an in vivo preclinical model, where SPP1 supplementation demonstrated therapeutic efficacy. Our findings position SPP1 as a central mediator of trophoblast-endothelial communication, its serum levels as a promising biomarker, and SPP1 itself as a potential therapeutic target worthy of further investigation. This work advances our understanding of PE pathogenesis and provides a foundational framework for future diagnostic and therapeutic innovations aimed at restoring SPP1-mediated trophoblast-endothelial crosstalk and vascular homeostasis.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Author contributions

All authors contributed to the study conception and design. XYM primarily conducted the study design, experimentation, data analysis, and manuscript writing. HFK additionally contributed to experimentation, data collection. FW, YS and YQC were involved in clinical sample collection. HX supervised the project, acquired funding, and gave final approval of the manuscript. All authors read and approved the final manuscript.

Funding

This research was funded by the “S&T Program of Hebei, grant number 21377707D” and the “Medical Science Research Project of Hebei, grant number 20250045.

Data availability

Data is provided within the manuscript or supplementary information files.

Declarations

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

Approval was obtained from the ethics committee of the second Hospital of Hebei Medical University. The procedures used in this study adhere to the tenets of the Declaration of Helsinki. standards. Informed consent was obtained from the participants(approval number: 2024-R581). This study was approved by the Animal Ethics Committee of the second Hospital of Hebei Medical University (approval number: 2024-AE349).

Consent for publication

Not applicable.