Mitochondrial-targeted MitoQ attenuates PFOS-induced vasoconstriction and cAMP suppression in human placental arteries

Abstract

Perfluorooctanesulfonic acid (PFOS), a persistent environmental pollutant, is elevated in the plasma of preeclamptic women and may contribute to gestational hypertension. However, its direct effects on placental vascular function remain unclear. This study investigated PFOS’ impact on vascular reactivity in human placental chorionic plate arteries (CPAs) and evaluated the therapeutic potential of the mitochondrial-targeted antioxidant mitoquinone (MitoQ). CPAs from normotensive pregnancies were treated with PFOS (10 μM) or vehicle for 24 h. Wire myography assessed contractile responses to KCl and endothelin-1 (ET-1), as well as cyclic adenosine monophosphate (cAMP)-mediated (isoproterenol, forskolin) and cyclic guanosine monophosphate (cGMP)-mediated (sodium nitroprusside, SNP) vasodilation. Adenosine triphosphate (ATP), cAMP, and cGMP levels were quantified, and MitoQ (100 nM) co-treatment was tested for rescue effects. PFOS-exposed CPAs exhibited enhanced maximal contraction to KCl (9.73 ± 0.96 vs. 5.60 ± 0.41 mN) and ET-1 (9.84 ± 1.05 vs. 5.77 ± 0.49 mN). cAMP-dependent relaxation was impaired (isoproterenol: 31.16 ± 4.55% vs. 65.27 ± 6.08%; forskolin: 65.59 ± 2.72% vs. 87.55 ± 1.89%), while SNP-induced cGMP-mediated relaxation remained unaffected. PFOS reduced ATP by 58% (30.43 ± 2.89 vs. 73.71 ± 9.15 μmol/mg) and cAMP by 57% (46.15 ± 9.54 vs. 109.1 ± 8.88 nmol/mg) but did not affect cGMP levels. MitoQ restored ATP/cAMP levels and normalized vascular function, reversing PFOS-induced hypercontractility and cAMP pathway suppression. These findings demonstrate that PFOS directly impairs placental vascular homeostasis via mitochondrial dysfunction and cAMP signaling disruption, implicating it as a mediator of gestational hypertension. Mitochondrial-targeted interventions like MitoQ may mitigate PFOS-associated vascular dysfunction, highlighting therapeutic avenues for high-exposure pregnancies.

PFOS exposure induces hypercontractility and impaired cAMP-mediated vasodilation via ATP/cAMP depletion—a mechanism rescued by the mitochondrial antioxidant MitoQ, directly linking PFAS toxicity to hypertensive disorders of pregnancy.

Graphical Abstract

Graphical Abstract.

Introduction

The human placenta represents one of nature’s most sophisticated vascular networks, developing from simple vasculogenesis during the third week post-conception into a highly specialized organ capable of sustaining fetal growth throughout gestation [1]. This transient yet vital organ orchestrates the complex exchange of nutrients, gases, and waste between maternal and fetal circulations through an intricate vascular architecture characterized by low resistance and high compliance. The placental vascular system undergoes extensive and continual remodeling throughout pregnancy, with gradual increases in capillary number, volume, and surface area within placental villi during the first and early second trimesters [1]. This adaptive remodeling ensures optimal blood flow dynamics essential for supporting the exponentially increasing metabolic demands of the developing fetus.

Central to placental perfusion are the chorionic plate arteries (CPAs), which form the terminal resistance vessels of the fetoplacental circulation [2, 3]. Unlike other vascular beds, placental vessels lack autonomic innervation, relying entirely on locally derived and circulating vasoactive molecules—including endothelin-1 (ET-1), prostaglandins, and nitric oxide (NO)—to regulate vascular tone [4, 5]. This unique regulatory mechanism makes placental vessels particularly vulnerable to disruptions in molecular signaling pathways that govern smooth muscle contractility and endothelial function.

Adequate fetoplacental perfusion is critical for successful pregnancy outcomes, as evidenced by the strong association between fetal growth restriction and compromised placental blood flow [6]. Disruptions in placental vascular homeostasis are central to hypertensive disorders of pregnancy (HDP), particularly preeclampsia (PE)—a condition marked by new-onset hypertension and organ damage after 20 weeks of gestation [7]. PE affects 2%–8% of pregnancies globally and represents a leading cause of maternal and perinatal morbidity and mortality. Beyond immediate pregnancy complications, PE predisposes offspring to lifelong cardiovascular risks, establishing a transgenerational cycle of vascular dysfunction [8].

The pathophysiology of PE involves complex interactions among inadequate placental perfusion, endothelial dysfunction, and systemic inflammatory responses that culminate in the characteristic maternal syndrome. Although PE is a major cause of maternal and fetal morbidity and mortality, the mechanisms of this disorder have not been clearly identified. Studies in animal models of hypertension during pregnancy have proposed that localized reduction in placental perfusion and resulting placental hypoxia lead to the release of vasoactive factors into the maternal circulation. These circulating factors then cause systemic vascular changes, increased vascular resistance, and hypertension [9–11]. Although several vasoactive factors have been proposed, a key role for environmental exposures in the pathogenesis of hypertension during pregnancy has been hypothesized [12–14].

Environmental exposures constitute one of several hypotheses concerning the pathogenesis of hypertension during pregnancy. This “environmental hypothesis” has gained substantial support through studies demonstrating elevated plasma levels of per- and polyfluoroalkyl substances (PFAS) in affected women [15, 16]. Several other studies have shown that plasma concentrations of perfluorooctane sulfonic acid (PFOS), the predominant PFAS compound, are elevated approximately 2- to 3-fold in women with PE compared to normotensive controls [16, 17]. Experimental studies have also shown that elevated PFOS levels in pregnant rats significantly increase systemic vascular resistance and arterial pressure [18, 19]. Furthermore, endothelium-dependent vascular relaxation is reduced in the uterine arteries of pregnant rats exposed to PFOS in drinking water [18, 19]. However, a fundamental question is whether the hemodynamic effects of elevated PFOS during pregnancy reflect the direct impact of PFOS on the mechanisms of vascular contraction/relaxation. Although studies in PFOS-exposed pregnant rats have suggested a possible association between endothelial cell dysfunction and hypertension [18, 19], it is not clear whether PFOS causes the reduction in endothelium-dependent vascular relaxation or whether it is merely a consequence of the hypertension developed during chronic PFOS exposure. Also, PFAS may activate other vasoactive factors in vivo [16], raising the possibility that the chronic vascular effects of PFOS may not be caused directly by PFOS but rather by other vasoactive factors. Furthermore, the inherent limitations of rodent models—which cannot fully replicate human gestational hemodynamics or placental vascular biology—necessitate investigation using human placental tissues. Importantly, human placental vasculature—with its unique trophoblast-endothelial interactions and gestational signaling pathways—may respond differently to PFOS than rodent arteries. This makes it important to investigate the direct effects of PFOS on the mechanisms of vascular contraction/relaxation in human CPAs, a clinically relevant vascular bed governing placental perfusion.

Mitochondrial dysfunction represents a central mechanism linking environmental exposures to placental vascular pathology in PE. Placental tissues from PE pregnancies consistently demonstrate impaired oxidative phosphorylation, increased reactive oxygen species production, and compromised adenosine triphosphate (ATP) synthesis. These bioenergetic deficits have profound implications for vascular function, as ATP serves as the essential substrate for cyclic adenosine monophosphate (cAMP) synthesis via adenylyl cyclase activation [20–22]. Recent investigations by Alissa et al. demonstrated that PFOS directly disrupts mitochondrial function in placental trophoblasts, inducing oxidative stress, membrane depolarization, and severe ATP depletion [20]. The connection between mitochondrial dysfunction and vascular regulation extends beyond simple energy deficits. Reduced ATP availability compromises cAMP production, which subsequently impairs protein kinase A (PKA)-mediated vasodilation and calcium desensitization mechanisms vital for maintaining appropriate vascular tone [23, 24]. The cAMP-mediated relaxation pathway, governed by β-adrenergic receptors and adenylyl cyclase, operates independently of the NO/cyclic guanosine monophosphate (cGMP) pathway, yet both systems prove indispensable for vascular homeostasis [25]. While PFOS broadly impairs endothelial function, its selective disruption of smooth muscle signaling pathways—particularly the preferential targeting of cAMP and cGMP systems—remains incompletely characterized in human placental vessels.

The present study addresses these critical knowledge gaps by testing the hypothesis that PFOS directly enhances vascular contraction while impairing relaxation responses in human CPAs, and that mitoquinone (MitoQ)—a mitochondria-targeted antioxidant—can counteract PFOS-induced vascular dysfunction. Using ex vivo CPAs isolated from normotensive pregnancies, we pursued four specific objectives: (1) Determine whether PFOS potentiates vascular contraction to depolarizing stimuli (KCl) and ET-1; (2) Assess PFOS effects on both cAMP- and cGMP-mediated vasodilation pathways; (3) Establish mechanistic links between functional vascular changes and PFOS-induced alterations in cellular ATP, cAMP, and cGMP levels; and (4) Evaluate MitoQ’s therapeutic potential for rescuing PFOS-induced placental vascular dysfunction.

Materials and methods

Ethical approval and tissue collection

This study received ethical approval from the Institutional Review Board of the University of Wisconsin, Madison (Protocol #2004–006), in compliance with the Declaration of Helsinki and institutional guidelines for human tissue research. Written informed consent was obtained from all participants enrolled in the UnityPoint Health-Meriter Hospital (Madison, Wisconsin) prior to placental collection. Placentas (n = 20) were collected within 30 min of uncomplicated Caesarean deliveries (37–42 weeks gestation) from normotensive pregnancies without maternal comorbidities (e.g. hypertension, diabetes) or fetal growth restriction. Demographic and clinical characteristics are summarized in Table 1. All patients are Caucasians with no current or past history of other major complications. To ensure tissue integrity, biopsies were obtained from non-calcified, well-perfused regions of the chorionic plate, as identified by visual inspection and palpation.

Table 1.

Maternal and neonatal characteristics of the study cohort

Characteristic	Value
Maternal age (years)	30.2 ± 4.8
Pre-pregnancy BMI (kg/m2)	24.8 ± 3.2
Parity	Nulliparous: 55%; Multiparous: 45%
Gestational age (weeks)	38.5 ± 1.1
Smoking (%)	0
Fetal sex (M/F)	11/9
Neonatal birth weight (g)	3330 ± 290

Chorionic plate artery isolation and preparation

CPAs were traced from the umbilical cord insertion point and third-generation resistance-sized vessels (200–400 μm internal diameter) were excised [3, 26]. Arteries were immediately immersed in ice-cold Krebs physiological salt solution (KPS; composition in mM: NaCl 119, KCl 4.7, CaCl₂ 2.5, MgSO₄ 1.17, NaHCO₃ 25, KH₂PO₄ 1.18, EDTA 0.026, d-glucose 5.5, and pH 7.4). Connective tissue was removed under a stereomicroscope, and arteries were sectioned into 1 mm rings. Rings were incubated in phenol red-free DMEM supplemented with 10% charcoal-stripped fetal bovine serum and antibiotics (100 U/mL penicillin, 100 μg/mL streptomycin) at 37°C in 5% CO₂ [27]. To assess PFOS effects, CPA rings were randomly assigned to be treated with vehicle (0.1% DMSO) or PFOS (10 μM) for 24 h. This concentration reflects bioaccumulation in placental tissue, which exceeds serum levels (0.6–4.8 μM) due to PFOS’ pharmacokinetic properties [28–30] and the fact that more than 99% of PFOS is protein-bound, yielding 0.1–0.2 μM free PFOS [31]. For MitoQ co-treatment, arteries were simultaneously exposed to 100 nM MitoQ, a concentration previously validated for mitochondrial targeting without cytotoxicity [32, 33].

Wire myography and normalization

After treatment, CPA rings were mounted on a wire myograph (Danish Myo Technology, Aarhus, Denmark) for isometric tension recording. Vessels were normalized to an internal diameter corresponding to 0.9 × IC₁₀₀ (where IC₁₀₀ is the circumference at 100 mmHg transmural pressure), following Mulvany method [34, 35]. Briefly, each CPA ring was stretched in 0.5-mN increments until the maximal active force to 80 mM KCl was achieved. The vessel circumference at 100 mmHg (IC₁₀₀) was calculated via the Laplace-based algorithm, and rings were set to 0.9 × IC₁₀₀, equilibrating for 30 min before testing. From each donor, an average of six rings (range 4–8) were included in analyses, with data averaged to yield one n per patient. Arteries were equilibrated in oxygenated KPS (95% O₂/5% CO₂) for 30 min before experimentation [36, 37].

Assessment of contractile responses

Vessel viability was confirmed via three successive exposures to 80 mM KCl, with rings generating <1 mN force excluded [3, 26]. Cumulative concentration-response curves to KCl (0–80 mM) and ET-1 (10⁻¹⁰–10⁻⁷ M) were constructed.

Assessment of vascular relaxation responses

Steady-state, submaximal pre-constrictions with EC₈₀ ET-1 were obtained and cAMP-mediated relaxation was assessed using cumulative additions of isoproterenol (10⁻¹⁰–10⁻⁵ M) or forskolin (10⁻¹⁰–10⁻⁵ M). cGMP-mediated relaxation was determined using sodium nitroprusside (SNP) (10⁻⁹–10⁻⁵ M).

Biochemical assays

ATP measurement

Snap-frozen CPAs (20 mg) were homogenized in ice-cold phosphate-buffered saline (PBS, pH 7.4; 200 μL) using a homogenizer (Fisherbrand 150, ThermoFisher Scientific, Newark, DE) on ice to ensure complete tissue lysis while maintaining ATP stability and minimizing degradation by endogenous ATPases. Following homogenization, samples were centrifuged at 12,000 × g for 5 min at 4°C to pellet cellular debris, unbroken cells, and insoluble material. The resulting clarified supernatant containing soluble cytoplasmic ATP was collected. ATP quantification was performed according to the manufacturer’s instructions using the EnzyLight ATP Kit (cat# EATP-100; BioAssay Systems) measured using a SpectraMax i3× multi-mode plate reader (Molecular Devices, San Jose, CA). Total protein content in the supernatant was determined using the bicinchoninic acid (BCA) assay. ATP levels were normalized to protein content (BCA assay; Pierce #23225, ThermoFisher Scientific, Newark, DE) and expressed as μmol/mg protein.

Cyclic adenosine monophosphate and cyclic guanosine monophosphate quantification

Snap-frozen CPAs (50 mg wet weight) were homogenized in 500 μL ice-cold 0.1 M hydrochloric acid (HCl; 10% w/v) using a handheld rotor-stator homogenizer (Fisherbrand 150, ThermoFisher Scientific, Newark, DE) equipped with a stainless-steel saw-tooth probe. Homogenization was performed in 30-s pulses on ice to inhibit endogenous phosphodiesterases (PDEs), stabilize cyclic nucleotides, and ensure complete lysis. The samples were vortexed vigorously (10 sec) and centrifuged at 800 × g for 10 min at 4°C. The resulting supernatant was collected for immediate analysis. cAMP and cGMP were quantified using acetylated competitive ELISAs (cAMP Complete ELISA kit #ADI-900-163; cGMP Complete ELISA kit #ADI-900-164; Enzo Life Sciences, Farmingdale, NY) according to manufacturer protocols. Briefly, samples and standards were acetylated (acetic anhydride/triethylamine) to enhance sensitivity prior to incubation in antibody-coated microplates. After sequential incubation with cyclic nucleotide conjugates and anti-rabbit IgG-HRP, tetramethylbenzidine substrate was added. Reactions were stopped with sulfuric acid, and absorbance was measured at 450 nm (reference: 590 nm) using a SpectraMax i3x multi-mode plate reader (Molecular Devices, San Jose, CA). Total protein was determined from a parallel homogenate prepared in ice-cold RIPA buffer using the BCA assay. Cyclic nucleotide concentrations (nmol) were normalized to total protein (mg) and expressed as nmol/mg protein.

Vascular histology of chorionic plate arteries

CPAs treated with vehicle, PFOS, or MitoQ were fixed by overnight immersion in 10% (vol/vol) neutral buffered formalin (ThermoFisher Scientific, Newark, DE). Tissues were dehydrated, embedded in paraffin wax, cut into 6 μm consecutive cross-sections, and mounted on glass slides. After deparaffinization in xylene, sections were rehydrated and stained with hematoxylin and eosin using standard protocols. For each group, four samples were analyzed, with two sections per sample evaluated. Images were acquired using a Keyence BZ-X700 microscope (Keyence Corporation, Itasca, IL), and vessel characteristics were assessed.

Western blotting using proteins from chorionic plate arteries

CPAs from each group were homogenized in ice-cold RIPA buffer (Cell Signaling Technology) supplemented with Halt protease and phosphatase inhibitor cocktail (Thermo Fisher Scientific). The homogenates were centrifuged at 14 000 g for 10 min at 4°C to collect the lysates. Protein concentration was determined using a BCA assay kit (Pierce, Thermo Fisher Scientific). Lysates (30 μg protein) were mixed with NuPAGE LDS sample buffer and reducing agent (Invitrogen, Thermo Fisher Scientific), then separated by electrophoresis using 4%–12% gradient NuPAGE Bis-Tris gels (Invitrogen) at 120 V for 3 h. Proteins were transferred to Immobilon transfer membranes (Merck Millipore) at 20 V for 60 min. Membranes were blocked with 5% nonfat milk for 1 h at room temperature and incubated overnight at 4°C with primary antibodies: anti-PKA (PKA C-Alpha, Cell Signaling Technology, #5842 T) and anti-GAPDH (Cell Signaling Technology, #4967).

After washing with TBST (three times, 10 min each), membranes were incubated with HRP-conjugated secondary antibody (Goat anti-Rabbit, Invitrogen #A27036) for 1 h at room temperature. Immunoblots were developed using the Pierce ECL detection kit (#32106, Thermo Scientific). Blots were imaged using the LI-COR Odyssey XF system (LI-COR Biosciences). Band intensities were quantified with ImageJ software and expressed as fold change relative to GAPDH and the control group.

Statistical analysis

The contractile force was expressed as active stress (mN). Relaxation responses to isoproterenol, forskolin and SNP were calculated as percent relaxation of the ET-1-induced contraction. Data from multiple rings per donor were averaged, with n representing the number of patients. Concentration-response curves were fitted to a four-parameter sigmoid model (GraphPad Prism 9.0). Maximal response (E_max) and pD₂ (−log[EC₅₀]) were derived via nonlinear regression. All datasets were tested for normality using the Shapiro–Wilk test, and all met the assumption of normality before applying unpaired t-tests or one-way ANOVA with Tukey’s post hoc comparisons. Differences were deemed statistically significant if P < 0.05.

Results

Depolarization-induced contraction

KCl caused concentration-dependent increases in active stress in CPAs, reaching a maximum of 5.60 ± 0.41 mN in vehicle-treated arterial rings. PFOS-treated arteries exhibited significantly enhanced depolarization responses, reaching a maximum stress of 9.73 ± 0.96 mN (P = 0.0014 vs. vehicle; Figure 1A and Table 2). Subgroup analysis of KCl-induced contractions revealed no significant sex differences in PFOS effects between male and female arteries compared to their respective controls (supplementary Figure S1).

Figure 1.

PFOS enhances depolarization- and endothelin-1-Induced contraction in human placental chorionic plate arteries (CPAs). CPA rings were incubated with vehicle (0.1% DMSO) or PFOS (10 μM) for 24 h. Cumulative concentration-response curves to (A) KCl (0–80 mM) and (B) endothelin-1 (ET-1; 10⁻¹⁰–10⁻⁷ M) were generated. Contractions are expressed as active stress (mN). Maximal response (E_max) values are shown in right panels. Data represent mean ± standard error of the mean (SEM) from eight patients. ^*P < 0.05 vs. vehicle.

Table 2.

Maximal active stress, pD₂, and maximal relaxation in human placental chorionic plate arteries (CPAs) treated with vehicle and PFOS (10 μM) for 24 h

		Vehicle	PFOS (10 μM) (P values)
KCl (Emax, mN)		5.60 ± 0.41	9.73 ± 0.96*  (0.0014)
ET-1 (Emax, mN)		5.75 ± 0.49	9.84 ± 1.05*  (0.0033)
IPT	(pD 2 )	−5.97 ± 0.08	−3.94 ± 0.19*  (<0.0001)
	(E max , %)	65.27 ± 6.08	31.16 ± 4.55*  (<0.0001)
FSK	(pD 2 )	−6.29 ± 0.04	−5.64 ± 0.05*  (<0.0001)
	(E max , %)	87.55 ± 1.89	65.59 ± 2.72*  (<0.0001)
SNP	(pD 2 )	−7.45 ± 0.05	−7.41 ± 0.06 (0.5869)
	(E max , %)	92.08 ± 1.63	90.22 ± 2.19 (>0.9999)

Values are expressed as mean ± SEM of eight arterial rings from eight patients in each group. pD₂ is presented as –log M and E_max is presented as maximal stress for contraction or percent relaxation. ET-1 – Endothelin-1; IPT – Isoproterenol; FSK - forskolin and SNP - sodium nitroprusside. ^*P < 0.05 vs. vehicle.

ET-1-induced contraction

ET-1 elicited concentration-dependent increases in active stress in CPAs, with PFOS exposure shifting the concentration-response curve leftward (Figure 1B). Maximal ET-1-induced stress increased from 5.75 ± 0.49 mN in controls to 9.84 ± 1.05 mN in PFOS-treated rings (P = 0.0033;  Figure 1B,  Table 2). No significant sex differences were observed in ET-1 contractile responses in PFOS-treated male and female CPAs relative to controls (supplementary Figure S1).

Cyclic adenosine monophosphate-mediated relaxation

β-Adrenergic signaling

Isoproterenol-induced relaxation (Emax: 65.27 ± 6.08% in controls) was attenuated by PFOS (Emax: 31.16 ± 4.55%; P < 0.0001; Figure 2A and Table 2). Because the magnitude of ET-1 contraction was different between PFOS-treated and vehicle controls, the ET-1 concentration was adjusted to produce a submaximal contraction that was similar in magnitude to that in vehicle controls.

Figure 2.

PFOS selectively impairs cAMP- but not cGMP-dependent vasodilation in CPAs. CPAs were treated with vehicle or PFOS (10 μM) for 24 h. Submaximal preconstriction with ET-1 (EC₈₀) was followed by cumulative additions of (A) isoproterenol (β-adrenergic agonist), (B) forskolin (adenylyl cyclase activator), or (C) sodium nitroprusside (SNP; NO donor) and relaxation (% of ET-1-induced contraction) was assessed. Bar graphs of the inset show summarized pD₂ values. Data represent mean ± standard error of the mean (SEM) from eight patients. ^*P < 0.05 vs. vehicle.

Adenylyl cyclase activation

Forskolin-mediated relaxation decreased from 87.55 ± 1.89% (vehicle) to 65.59 ± 2.72% following PFOS exposure (P < 0.0001; Figure 2B and Table 2).

Cyclic guanosine monophosphate-mediated relaxation

SNP produced comparable relaxation in PFOS-treated (Emax: 90.22 ± 2.19%; pD₂: −7.41 ± 0.06) and control arteries (Emax: 92.07 ± 1.63%; pD₂: −7.45 ± 0.05; P > 0.5; Figure 2C and Table 2).

Bioenergetic and second messenger levels

PFOS reduced ATP by 58.35% (30.43 ± 2.89 vs. 73.71 ± 9.15 μmol/mg protein; P = 0.0005) and cAMP by 57.34% (46.15 ± 9.54 vs. 109.1 ± 8.88 nmol/mg protein; P = 0.0003). cGMP levels remained unchanged (Figure 3).

Figure 3.

PFOS depletes ATP and cAMP levels in CPAs without affecting cGMP. CPAs were exposed to vehicle or PFOS (10 μM) for 24 h. (A) ATP (μmol/mg protein), (B) cAMP (nmol/mg protein), and (C) cGMP (nmol/mg protein) levels were quantified using EnzyLight ATP (EATP-100; BioAssay Systems) and ELISA kits (ADI-900-163A cAMP, ADI-900-164 cGMP; Enzo Life Sciences). Data represent mean ± SEM (n = 8). ^*P < 0.05 vs. vehicle.

Mitochondrial antioxidant rescue by mitoquinone

Contraction

Co-treatment with MitoQ normalized PFOS-enhanced KCl (Emax: 6.76 ± 0.69 mN vs. 10.25 ± 0.81 mN; P = 0.0038) and ET-1 responses (Emax: 6.46 ± 0.78 mN vs. 10.06 ± 0.92 mN; P = 0.015; Figure 4 and Table 3).

Figure 4.

Mitoquinone (MitoQ) attenuates PFOS-induced hypercontractility in CPAs. CPAs were co-treated with PFOS (10 μM) and MitoQ (100 nM) for 24 h. Cumulative responses to (A) KCl and (B) ET-1 were measured. Emax values are shown in right panels. Data represent mean ± SEM (n = 8). Different lowercase letters denote significant differences between groups (P < 0.05).

Table 3.

Maximal active stress, pD₂, and maximal relaxation in human placental chorionic plate arteries (CPAs) treated with vehicle and PFOS (10 μM) in the presence or absence of mitoquinone (MitoQ) for 24 h

		Vehicle	PFOS (10 μM)	MitoQ (100 nM) + PFOS (10 μM)	MitoQ (100 nM)
KCl (Emax, mN)		6.17 ± 0.22a	10.25 ± 0.81b (0.0007)	6.76 ± 0.69a (0.9141)	5.55 ± 0.63a (0.9080)
ET-1 (Emax, mN)		5.91 ± 0.78a	10.06 ± 0.92b (0.0042)	6.46 ± 0.78a (0.9572)	5.35 ± 0.61a (0.9585)
IPT	(pD 2 )	−5.81 ± 0.11a	−4.04 ± 0.19b (<0.0001)	−5.21 ± 0.09c (0.0138)	−5.36 ± 0.08a,c (0.0842)
	(E max , %)	59.54 ± 7.60a	33.42 ± 5.29b (<0.0001)	53.66 ± 4.59a (0.5560)	52.40 ± 3.12a (0.3844)
FSK	(pD 2 )	−6.37 ± 0.04a	−5.64 ± 0.04b (<0.0001)	−6.34 ± 0.05a (0.9977)	−6.64 ± 0.12a (0.1565)
	(E max , %)	89.16 ± 1.78a	64.22 ± 2.48b (<0.0001)	85.24 ± 2.58a (0.8206)	86.31 ± 3.81a (0.9589)
SNP	(pD 2 )	−7.47 ± 0.05	−7.45 ± 0.08 (0.9969)	−7.48 ± 0.09 (0.9933)	−7.51 ± 0.05 (0.9813)
	(E max , %)	91.88 ± 1.93	90.22 ± 2.18 (>0.9999)	88.21 ± 4.48 (>0.9999)	90.08 ± 3.24 (>0.9999)

Values are expressed as mean ± SEM of eight arterial rings from eight patients in each group. pD₂ is presented as –log M and E_max is presented as maximal stress for contraction or percent relaxation. ET-1 – Endothelin-1; IPT – Isoproterenol; FSK - forskolin and SNP - sodium nitroprusside. Different lowercase letters denote significant differences (P ≤ 0.05) compared to the other groups.

Relaxation

MitoQ restored isoproterenol efficacy in PFOS-exposed rings (Emax: 53.66 ± 4.59% vs. 33.42 ± 5.29%; P < 0.0001) and forskolin responses (Emax: 85.24 ± 2.58% vs. 64.22 ± 2.48%; P < 0.0001; Figure 5A and B and Table 3). SNP responses were unaffected (Figure 5C and Table 3).

Figure 5.

MitoQ rescues cAMP-mediated vasodilation in PFOS-exposed CPAs. CPAs were co-treated with PFOS (10 μM) and MitoQ (100 nM) for 24 h. Relaxation responses to (A) isoproterenol, (B) forskolin, and (C) SNP were assessed. Bar graphs of the inset show summarized pD₂ values. Data represent mean ± SEM (n = 8). Different lowercase letters denote intergroup differences (P < 0.05).

Biochemical recovery

MitoQ reversed PFOS-induced ATP (68.70 ± 8.78 vs. 32.05 ± 6.70 μmol/mg protein; P = 0.0216) and cAMP deficits (145.90 ± 12.06 vs. 58.69 ± 12.58 nmol/mg protein; P = 0.0036; Figure 6).

Figure 6.

MitoQ restores PFOS-induced ATP and cAMP deficits in CPAs. CPAs were co-treated with treated with PFOS (10 μM) and MitoQ (100 nM) for 24 h. (A) ATP, (B) cAMP, and (C) cGMP levels were quantified. Data represent mean ± SEM (n = 8). Different lowercase letters denote intergroup differences (P < 0.05).

Morphological and protein expression analysis

Hemotoxylin and eosin (H&E) staining showed no significant differences in vessel wall thickness between groups (Figure 7). Immunohistochemistry for α-smooth muscle actin (α-SMA) revealed comparable smooth muscle density across treatments (supplementary Figure S2). Western blotting indicated unchanged total PKA expression with PFOS and/or MitoQ (Figure 8).

Figure 7.

Vascular histology of CPAs using hematoxylin and eosin staining. Vehicle, PFOS (10 μM), MitoQ (100 nM), and MitoQ (100 nM) + PFOS (10 μM) treated CPAs at 4× and 10× magnification following H&E staining. The bar graph shows the percentage area of H&E staining in vascular tissues (mean ± SEM, n = 4). No detectable changes were observed in vasculature between groups.

Figure 8.

Protein expression levels of protein kinase A in CPAs. Protein expression in vehicle, PFOS (10 μM), MitoQ (100 nM), and MitoQ (100 nM) + PFOS (10 μM) treated CPA samples were analyzed by western blot. Representative blots are shown (top), with densitometric quantification of PKA normalized to GAPDH (bottom, mean ± SEM, n = 4; one-way ANOVA and Dunnett post hoc test).

Discussion

The present study demonstrates that PFOS directly induces vascular dysfunction in human CPAs through mitochondrial bioenergetic failure and selective impairment of cAMP-mediated vasodilation. Key findings include exaggerated contractility to ET-1 and KCl, impaired cAMP-dependent relaxation via β-adrenergic and adenylate cyclase pathways, while preserving cGMP-dependent relaxation. These effects correlate with significant reductions in ATP and cAMP levels. Importantly, mitochondria-targeted antioxidant MitoQ restored ATP/cAMP levels and normalized vascular reactivity. These results bridge epidemiological associations between PFAS exposure and HDP with molecular mechanisms, offering novel insights into PFOS-specific placental pathophysiology.

The observed PFOS-induced enhancement of both KCl- and ET-1-mediated contractions in human CPAs stands in stark contrast to prior rodent studies, which reported no significant effect of PFOS on KCl responses in uterine arteries [18, 19]. This interspecies divergence likely arises from critical differences in vascular biology and exposure paradigms. In rodent models, chronic in vivo PFOS exposure (e.g. 3–5 weeks via drinking water) may activate compensatory mechanisms—such as upregulation of antioxidant defenses or adaptive changes in receptor density—that mask acute vascular effects [18, 19, 38–40]. Conversely, our human CPA model isolates direct PFOS actions over 24 h, circumventing systemic adaptations and revealing PFOS’ immediate mitochondrial effects.

This interspecies divergence likely arises from critical differences in vascular biology and exposure paradigms [18, 19]. Furthermore, fundamental distinctions between vascular beds contribute to this divergence: uterine arteries (maternal origin) and CPAs (fetal origin) exhibit differential responses to vasoactive stimuli due to developmental and physiological heterogeneity. These differences arise from both developmental programming and adaptive plasticity of vascular smooth muscle cells (VSMCs). VSMCs dynamically remodel contractile properties in response to local biomechanical stress, aging, and environmental cues [41–43]. When considered alongside differences in developmental origin, our results suggest that fetal-origin vessels (CPAs) may be particularly vulnerable to ATP depletion, whereas maternal-origin vessels (uterine arteries) may be more susceptible to transcriptional disruption. Specifically, CPAs lack autonomic innervation [43], rendering them exquisitely dependent on local ATP/cAMP signaling. This distinction explains why fetal-origin vessels may be more vulnerable to PFOS-induced bioenergetic failure, whereas maternal-origin vessels show greater susceptibility to transcriptional disruption. These findings align with clinical observations of altered fetoplacental hemodynamics in PFOS-exposed pregnancies. Recent epidemiological work linked PFAS exposure (including PFOS) to a reduced cerebroplacental ratio—a key indicator of compromised fetoplacental circulation [44]. Though umbilical artery pulsatility index changes did not reach significance, this trend provides a clinical correlate to our ex vivo observations of hypercontractility and suppressed cAMP-mediated vasodilation. Mechanistically, the concurrent potentiation of depolarization- and receptor-mediated contractions suggest PFOS targets shared downstream pathways (e.g. ATP depletion) rather than agonist-specific receptors. These results mirror clinical reports of enhanced ET-1 vasoconstriction in preeclamptic chorionic arteries [43, 45–48], underscoring PFOS’ role in disrupting placental perfusion.

Although our ex vivo CPA model isolates direct vascular effects of PFOS, in vivo exposure also provokes systemic inflammatory and hormonal dysregulation that likely amplifies vascular dysfunction. PFOS elevates circulating IL-6 and TNF-α in pregnant women and animal models, cytokines known to impair endothelial NO bioavailability, increase oxidative stress, and upregulate adhesion molecules. Additionally, PFOS suppresses estradiol synthesis, removing estrogen’s vasoprotective effects such as eNOS upregulation and antioxidant defense. We propose future studies to integrate these systemic contributors by assessing fetoplacental hemodynamics in PFOS-exposed pregnant animals while measuring serum IL-6, TNF-α, and estradiol; co-incubating CPAs with PFOS plus recombinant cytokines; and supplementing with physiologic estradiol levels to dissect cooperative effects on vascular reactivity.

We also confirmed that fetal sex did not bias our findings: CPAs were derived from 20 placentas (11 male, nine female) randomized across all treatment groups, and subgroup analyses of KCl- and ET-1–induced contractions revealed no significant sex differences (supplementary Figure S1). This balance ensures generalizability across sexes and aligns with previous reports of minimal sex-based variability in normotensive placental vascular function.

To elucidate mechanisms underlying PFOS-induced hypercontractility, we interrogated key vasodilatory pathways. PFOS exposure impaired isoproterenol-induced relaxation, a β-adrenergic receptor-dependent process, but also attenuated responses to forskolin—a direct adenylyl cyclase activator. This dual inhibition rules out receptor-level dysfunction and instead implicates downstream cAMP synthesis or signaling. Crucially, SNP-mediated cGMP-dependent relaxation remained intact in PFOS-treated arteries, demonstrating pathway specificity. Direct biochemical analysis confirmed this specificity: PFOS reduced cAMP by 57% while leaving cGMP unchanged. These findings align with clinical studies of preeclamptic placentas, where cGMP signaling persists despite severe vascular dysfunction [49], suggesting cGMP suppression represents a conserved mechanism in gestational hypertension.

Mechanistically, PFOS impairs cAMP-driven vasodilation by depleting mitochondrial ATP—the essential substrate for adenylyl cyclase activity—while leaving cGMP pathways intact because soluble guanylyl cyclase utilizes GTP. This pathway-specific vulnerability underscores the central role of mitochondrial bioenergetics in supporting energy-intensive signaling cascades such as cAMP-PKA–mediated relaxation, and explains why cGMP-mediated responses to SNP were preserved. Clinically, this distinction suggests that NO donors or sGC stimulators could bypass PFOS-induced ATP deficits and restore vasodilation in affected placental beds.

The observed PFOS-induced suppression of vascular cAMP levels likely stems from mitochondrial dysfunction, a mechanism supported by converging lines of evidence. PFOS disrupts mitochondrial respiratory capacity in placental trophoblasts, impairing Complex II/III activity and reducing ATP production by >50% [20, 50]. Critically, our previous work demonstrated that PFOS reduces mitochondrial biogenesis in human trophoblasts [20, 50], directly decreasing mitochondrial content. Given the established relationship between mitochondrial abundance and ATP-generating capacity [51], this loss of mitochondrial mass likely contributes to the bioenergetic dysfunction observed in CPAs. Our findings in CPAs mirror this bioenergetic failure, with PFOS-exposed arteries showing a 58% reduction in ATP—a substrate essential for cAMP synthesis via adenylate cyclase [22, 52]. The selective impairment of cAMP signaling (but preserved cGMP responses) aligns mechanistically, as cAMP synthesis is ATP-dependent [21, 52], while cGMP pathways utilize GTP. This ATP-cAMP axis disruption explains impaired vasodilation and MitoQ’s efficacy: by scavenging mitochondrial ROS, MitoQ restores electron transport chain function, rescuing ATP production and cAMP synthesis. While ATP depletion drives cAMP deficits, we cannot exclude ancillary mechanisms (e.g. direct adenylate cyclase inhibition). Nevertheless, MitoQ’s rescue of cAMP and vasorelaxation implicates mitochondrial bioenergetics as the dominant pathway.

The PFOS-induced reduction in vascular cAMP levels not only attenuates cAMP-mediated relaxation but can also enhance contractile mechanisms. Reduced cAMP disinhibits myosin light chain kinase, enhancing calcium-calmodulin activation, and myosin phosphorylation [53]. Concurrently, cAMP loss activates RhoA/Rho-kinase signaling, suppressing myosin phosphatase and sustaining contraction via calcium-independent mechanisms [54, 55]. These findings position PFOS as a dual disruptor of vascular homeostasis, simultaneously attenuating vasodilatory capacity while hyperactivating contractile machinery through convergent calcium-dependent and -independent pathways.

Beyond mitochondrial dysfunction, PFOS may also exert non-bioenergetic toxicity by directly inhibiting adenylyl cyclase or enhancing PDE–mediated cAMP hydrolysis, as documented in astrocyte and pancreatic β-cell models. Although our analysis found no change in total PDE protein levels, this does not rule out alterations in their enzymatic activity or regulation. Therefore, future work should aim to delineate these potential mechanisms by measuring AC and PDE enzymatic activity directly in CPA homogenates and employing pharmacological dissection with specific AC inhibitors and PDE inhibitors. Furthermore, phosphoproteomic profiling could identify PFOS-induced phosphorylation changes in AC isoforms and PDE family members, which may regulate their function. Finally, to fully characterize the downstream impact, these approaches should be coupled with direct PKA activity assays to definitively rule out secondary effects on PKA signaling and to map specific alterations in the phosphorylation of its vascular smooth muscle targets.

However, this study has limitations. First, the ex vivo model isolates PFOS’ direct vascular effects but excludes systemic contributors to HDP, such as immune cells and hormones—processes that may synergize with vascular dysfunction in vivo (e.g. PFOS upregulates placental cytokines [56, 57] and suppresses estradiol synthesis [58]. Second, CPAs from normotensive pregnancies may not fully replicate preeclamptic pathophysiology. Third, experimental design constraints warrant consideration: (1) Human exposures involve complex PFAS mixtures (not PFOS alone), with epidemiological evidence showing additive/synergistic effects from PFOS, PFOA, and other congeners [59, 60]. (2) Our use of a single 10 μM PFOS dose reflects placental bioaccumulation kinetics (>99% protein binding yields ~0.1–0.2 μM free PFOS [31]) but lacks dose-response assessment. (3) PFOS levels in participants were not measured, precluding exposure-response correlations. (4) Technical limitations precluded endothelial-dependent testing: isolated CPAs failed to relax to classical agonists (e.g. acetylcholine, bradykinin) as reported previously [23, 24], necessitating focus on smooth muscle cAMP/cGMP pathways. Finally, therapeutic translation requires caution: MitoQ’s safety in pregnancy is underexplored, with studies showing gestational-age-dependent outcomes (beneficial in late-gestation sheep [61] but harmful in early-gestation rats [62]). Our findings await in vivo validation. To address these gaps, future work should: (1) Test PFAS mixtures at serum-equivalent ratios using mixture-response modeling. (2) Map dose-temporal dynamics of PFOS (1–5 μM; chronic/organ-on-chip models). (3) Correlate maternal PFOS levels with vascular dysfunction. (4) Validate rescue in PFOS-exposed pregnant rats ± MitoQ by measuring maternal blood pressure, uterine artery function, and placental/fetal growth metrics. (5) Define MitoQ’s safety window via reproductive-toxicity studies and mechanistic profiling (e.g. mitophagy [63], biogenesis [64]). (6) Quantify mitochondrial ROS, ΔΨm, and PKA activity in CPAs to establish causality between mitochondrial dysfunction and vascular impairment.

In conclusion, this study establishes mitochondrial bioenergetic failure as a central mechanism of PFOS-induced vascular dysfunction in human CPAs. PFOS-induced inhibition of ATP/cAMP production and vascular relaxation and enhancement of vascular contraction in CPAs supports a direct role for PFAS as one possible mediator of the increased vascular resistance associated with HDP. Mitigating mitochondrial dysfunction—via antioxidants like MitoQ or dietary interventions—could attenuate PFOS-associated HDP risk, particularly in high-exposure populations. Regulatory agencies should consider placental vascular toxicity when evaluating PFAS safety thresholds, as current guidelines may inadequately protect this critical developmental window.

Author contributions

PY and JM participated in concept development and study design, performed experimental work, data analysis and drafted the manuscript; RS and AH performed experimental work; SK conceived the conception and study design, obtained funding, provided important intellectual support, and edited the manuscript. All authors contributed to the article and approved the submitted version.

Conflict of interest: The authors have declared that no conflict of interest exists.

Data availability

The authors confirm that the data supporting the findings of this study are available within the article.