Abstract
Preeclampsia is a multisystem hypertensive disorder of pregnancy that remains one of the leading causes of maternal and perinatal morbidity and mortality worldwide. The widespread maternal endothelial dysfunction that underlies preeclampsia is thought to arise from excessive placental production of various factors combined with enhanced oxidative stress. While previous studies have reported elevated activin A in women diagnosed with preeclampsia, whether activin A can cause vascular dysfunction has not yet been thoroughly investigated. Here, we demonstrated that different subtypes of activin A receptors were localised to the endothelial and smooth muscle cells of mouse and human aortae. Then, the aorta of healthy female C57Bl6J mice (n = 8) were incubated for 24 h in various concentrations of recombinant activin A to mimic early pregnancy (5 ng/mL), late pregnancy (20 ng/mL) and preeclampsia (50 ng/mL). Vascular reactivity as assessed by wire myography revealed that only the preeclamptic level of activin A impaired agonist-mediated endothelium-dependent relaxation by reducing the vasodilator prostanoid contribution to relaxation. However, agonist-mediated endothelium-independent mechanisms were unaffected. Further investigations carried out on human aortic endothelial cells suggested that the impairment of aorta relaxation could also be driven by increased endothelial cell permeability, and decreased cell viability, adherence and proliferation. This is the first direct evidence to show that activin A can induce endothelial dysfunction in whole blood vessels, suggesting that at high circulating levels it may contribute to the widespread endothelial dysfunction in women with preeclampsia.
Introduction
Preeclampsia is a complex and multisystem hypertensive disorder that can manifest during the second half of pregnancy or immediate postpartum period. It is characterised by new-onset hypertension (≥140 mmHg systolic and/or ≥90 mmHg diastolic) combined with either proteinuria, other maternal end-organ dysfunction (i.e. renal insufficiency) or uteroplacental dysfunction (i.e. fetal growth restriction) (Lowe et al. 2015, Mol et al. 2016). Although clinical symptoms are ordinarily resolved by 12 weeks after birth, mounting emerging evidence has now established a relationship between a history of preeclampsia and future adverse cardiovascular events, such as heart attack, heart failure or stroke (Wu et al. 2017, Benschop et al. 2019).
Preeclampsia is predominantly thought of as a disorder of the vasculature; there is often shallow or deficient trophoblast (placental) cell invasion of the uterine spiral arteries (Cartwright et al. 2010). Consequently, remodelling of the spiral arteries that supply the fetoplacental unit is defective. As such, these critical arteries remain responsive to circulating vasoconstrictors, resulting in periods of intermittent blood flow, often pulsatile and high pressure, which can cause injury at the placental fetal interface and can result in ischemic–reperfusion injury in the developing placenta (Staff 2019). In response to these events, the placenta increases the production of reactive oxygen species and various factors including soluble fms-like kinase-1 (sFLT1) and soluble endoglin (sENG) (Maynard et al. 2003, Venkatesha et al. 2006). Extant literature has drawn a relationship between preeclampsia and abnormal levels of sFLT1 and sENG, which enter the maternal vasculature and can induce endothelial dysfunction (Goulopoulou & Davidge 2015). For years, research has focused on how sFLT1 and sENG contribute to the vascular dysfunction of preeclampsia. However, numerous studies have also reported aberrant activin A levels with significantly elevated in women with preeclampsia (D’Antona et al. 2000, Bersinger et al. 2003, Diesch et al. 2006, Florio et al. 2006, Shahul et al. 2018).
In contrast to a healthy pregnancy, preeclampsia is characterised by widespread reduced endothelium-dependent mechanisms of vessel relaxation, and/or heightened mechanisms of vasoconstriction (Goulopoulou & Davidge 2015). This significant increase in vascular reactivity is considered a response to reduced endothelium-dependent mechanisms of vessel relaxation, and/or heightened mechanisms of vascular smooth muscle contraction (Goulopoulou & Davidge 2015). As the maternal vasculature becomes abnormally sensitive to circulating vasoconstrictors, there is a concomitant loss of normal vessel dilation. This culminates in marked vasospasm, with reduced placental perfusion and reduced oxygenation of the placenta (Karumanchi 2016). Now, evidence suggests that activin A may also play an important role in the development of vascular dysfunction (Hobson et al. 2016).
Activin A is a two-subunit protein belonging to the transforming growth factor β (TGF-β) family of proteins (Bloise et al. 2019). Activin A itself drives cellular proliferation, differentiation, and apoptosis in many organ systems, while also being a key regulator of inflammation (Reddy et al. 2009, Wijayarathna & de Kretser 2016). Like other members of the TGF-β family, activin A plays an essential role in the early stages of embryo implantation, extravillous trophoblast invasion and the production of placental hormones (Stoikos et al. 2010, Wijayarathna & de Kretser 2016). Activin A is normally bound in a complex with follistatin, a protein that neutralises its activity (Wang et al. 2000). In contrast, unbound and free activin A is biologically active. Activin A has been shown to induce cellular changes through four cell surface receptors; type I receptors: activin receptor-like kinases, ALK2 and ALK4 and type II receptors: activin type II and type IIB (ACTRII and ACTRIIB) (Harrison et al. 2006). The formation of complexes made of type I and type II receptor combinations are necessary for activins to transmit their intracellular signals into target tissues (Bloise et al. 2019).
Previous work in human umbilical vein endothelial cells (HUVECs) identified that when treated with activin A, HUVECs upregulate markers of endothelial cell damage and stress (Hobson et al. 2016). In animal models, pregnant mice infused with recombinant human activin A resulted in the development of a preeclamptic phenotype including hypertension and proteinuria (Lim et al. 2015). It is thought that excessive release of activin A reflects an adaptive response by the placenta to try and promote increased trophoblastic invasion (Smith et al. 2002, Muttukrishna et al. 2006, Park et al. 2015). This pathological increase in activin A, in turn, likely enhances the maternal oxidative stress response, contributing to the widespread maternal endothelial vascular dysfunction of preeclampsia (Lim et al. 2015). To the best of our knowledge, the effect of activin A on whole blood vessels has not yet been explored. Therefore, this study investigated whether activin A could induce vascular dysfunction in the aorta of female mice and in human aortic endothelial and smooth muscle cell lines.
Materials and methods
Ethics
All animal experiments were approved by Monash Medical Centre Animal Ethics Committee (MMCB2019/11) and conducted in accordance with the Australian Code of Practice and the National Health and Medical Research Council. Mice were maintained on a 12 h light:12 h darkness cycle at 20°C, with standard food pellets (Barastock, VIC, Australia) and water available ad libitum.
Aorta isolation
Mouse aortae were dissected from a female adult mouse from strain C57BL6J between the ages of 12–24 weeks. Mice were euthanized with isoflurane and cervical dislocation. Then, the entire aorta was isolated and immediately placed into ice-cold Krebs physiological saline solution (mmol/L: NaCl 120, KCl 5, MgSO4 1.2, KH2PO4 1, NaHCO3 25, d-Glucose 11.1, CaCl2 2.5, bubbled with carbogen (95% O2 and 5% CO2)). The abdominal aorta was cleaned of muscle, fat and connective tissue under the guidance of a dissecting microscope (Nikon SMZ445/460). The abdominal aorta was fixed in 10% neutral buffered formalin for at least 48 h before processing and paraffin wax embedding for subsequent histological and immunohistochemical analysis, or cut with fine scissors into 2 mm lengths for wire myography.
Immunohistochemistry
Mouse aortae were cleaned of excess tissue and then processed and embedded in paraffin by the Monash University Histology Platform. Paraffin-embedded sections were cut into 5 µm thick sections and left to dry overnight on SuperFrost™ adhesive slides (ThermoFisher). Slides were baked for 20 min at 60°C before dewaxing and rehydrating. For cells, primary human aortic endothelial cells (HAoECs) and primary human aortic smooth muscle cells (HASMCs) were cultured in eight well chamber polystyrene slides (Falcon) and fixed with ice-cold 70% ethanol for 10 min and allowed to air dry. Cells were rehydrated in TBS for 5 minutes. For all tissues, endogenous peroxidase activity was quenched at room temperature with 3% hydrogen peroxide in absolute methanol for 15 min followed by blocking performed with serum-free Protein Block agent (Agilent Technologies) for 30 min. Primary antibodies were applied overnight at 4°C. Anti-activin receptor type IA-ab60157 (aorta: 1/50, HAoECs: 1/100, HASMCs: 1/100), receptor type IB-ab227234 (aorta: 1/50, HAoECs: 1/100, HASMCs: 1/100), receptor type IIA-ab216960 (aorta: 1/1000, HAoECs: 1/50, HASMCs: 1/50) and receptor type IIB-ab128544 (aorta 1/100, HAoECs: 1/50, HASMCs: 1/50) (Abcam). The negative control was generated by replacing the primary antibody with rabbit IgG (Sigma-Aldrich). Slides were washed in TBS 0.025% Triton X and the EnVision+ System- HRP Labelled Polymer Anti-Rabbit secondary antibody (Agilent Technologies) was applied for 30 min at room temperature. Slides were washed in TBS 0.025% Triton X and staining visualised using Liquid DAB+ Substrate Chromatin System (Agilent Technologies) for 2–3 min. All slides were scanned by the Aperio Scanscope AT Turbo and images were prepared using ImageJ. Counterstaining to identify endothelial and smooth muscle cells of the mouse aorta was not performed due to the localisation of the receptor in individual HAoECs and HASMCs.
Vascular studies
Vascular dysfunction ex vivo
Each aortic segment was incubated for 24 h at 37°C in room air with 5% CO2 in a 24-well plate, containing 0.5 mL of media (Dulbecco’s Modified Eagle Medium/Nutrient Mixture F-12; ThermoFisher) and recombinant human/mouse/rat activin A (R&D Systems). First, dose optimisation experiments were undertaken. Arteries were incubated with either control (4 nM hydrochloric acid, HCl), or with increasing concentrations of activin A (5, 20 or 50 ng/mL). These concentrations represent levels in early pregnancy (5 ng/mL), levels in late pregnancy (20 ng/mL), women with preeclampsia (50 ng/mL) and in later investigations, supraphysiological (100 ng/mL) concentrations (Petraglia et al. 1995, Fowler et al. 1998).
After incubation, each segment was threaded with two 40 μm stainless steel wires and mounted on a four-channel wire myograph (model 610M; Danish Myo Technology, Aarhus, Denmark) before normalisation. Vascular reactivity was measured in real-time using LabChart software (ADInstruments, NSW, Australia).
Assessment of vascular reactivity
A hallmark of preeclampsia is vascular dysfunction, which displays decreased relaxation and/or increased contraction of the maternal vasculature. Wire myography allows for the interrogation of vascular reactivity to vasoactive agents. After mounting the aorta segments onto the wire myograph, each segment was submerged in 5mL of Krebs-buffer maintained at 37°C and bubbled with carbogen (95% O2, 5% CO2). Once the aorta was left for ~10 min to stabilise, the tension of each segment was increased and maintained at 5 mN for ~30 min before experimentation (Marshall et al. 2020).
Initially, each aorta segment is assessed to ensure both the smooth muscle and endothelial cells are functional. To test the vascular smooth muscle cells, each segment was constricted using the thromboxane A2 mimetic U46619 (0.5 × 10−6 M). This dose of U46619 caused maximum vessel constriction (Emax). Then, the endothelium-dependent dilator acetylcholine (ACh, 10−5 M) was added at 50–70% of the vessel’s maximum constriction to induce relaxation and to test whether the endothelium was functional. Once vessel segments were confirmed to be functional, arteries were rested for 15–20 min before undertaking constriction or relaxation curves (Marshall et al. 2018, Langston-Cox et al. 2020).
To assess endothelium-dependent and endothelium-independent vasodilation, the aorta was pre-constricted to 50–70% of their maximum using U46619 (~0.2 × 10−8 M). To assess vessel relaxation, relaxation curves to ACh (10−9 to 10−5 M), and the endothelium-independent agent, sodium nitroprusside (SNP, 10−10 to 10−5 M) were performed. Relaxation is expressed as a percentage of pre-constriction to U46619. Smooth muscle reactivity to different vasoconstrictors was also evaluated. Contraction curves to the endothelium-dependent vasoconstrictors endothelin-1 (ET-1, 10−10 to 10−5 M) and U46619 (10−11 to 10−6 M) were also undertaken.
To further explore activin A-induced vascular dysfunction, responses to ACh were performed as described above after a 20-min incubation with one or more pharmacological blocking agents. The cyclooxygenase inhibitor indomethacin (Indo; 0.5 × 10−6 M) prevents the aorta from producing prostanoids (i.e. PGI2). A combination of Indo and the nitric oxide synthase (NOS) inhibitor Nω-nitro-l-arginine methyl ester (l-NAME; 0.5 × 10−6 M) prevents the aorta from producing both prostanoids and NO. The remaining relaxation after blocking both NO and prostanoids with their respective blocking agents is attributed to endothelium-derived hyperpolarisation (EDH). The contribution of intermediate- and small-conductance calcium-activated potassium channels to EDH-mediated relaxation was assessed by pre-incubation with TRAM-34 (0.5 × 10−3 M) and apamin (0.5 × 10−6 M), respectively, in the presence of l-NAME + Indo (Marshall et al. 2017, 2018).
Vascular reagents
All vascular drugs were purchased from Sigma-Aldrich. Drugs were dissolved in distilled water, with the exception of indomethacin (0.1 mol/L sodium carbonate) and TRAM-34 (DMSO), with subsequent dilutions in distilled water.
Human aortic endothelial cells and human aortic smooth muscle cells cell culture
HAoECs and HASMCs were purchased from American Type Culture Collection (ATCC PCS-100-011 and PCS-100-012; Manassas, VA). HAoECs (from passages 3 to 7) were cultured in Endothelial Growth Medium-2 BulletKit (Lonza, USA, Cat. No. CC-3162) or HASMCs (from passages 3 to 7) were cultured in Smooth Muscle Growth Medium-2 BulletKit (Lonza, USA, Cat. No. CC-3182; ascorbic acid was discarded). All cells were cultured at 37°C with 5% CO2. Each passage represented one nnumber. Once cells reached ~80% confluence, the cells were sub-cultured or used for experimentation. Briefly, media was aspirated, cells washed twice with 10 mL of PBS and then replaced with trypsin (0.25%) for 5 min at 37°C. Cells were dislodged and an equal volume to the trypsin of Hank’s Balanced Salt Solution (HBSS) was added to neutralise trypsin. The cell suspension was then transferred into a 50-mL falcon tube centrifuged at 300× g for 5 min at room temperature. The supernatant was aspirated and the cell pellet was carefully resuspended in media. Cell counts were performed using an automated haemocytometer (Countess II Automated Cell Counter, Invitrogen, ThermoFisher). Cells were then diluted to the required concentration for subsequent experimentation.
MTS cell viability assay
Cell viability, as measured by cell metabolism, was assessed using a CellTiter 96 AQueous One Solution Cell Proliferation Assay (MTS) kit (Promega). Briefly, HAoECs or HASMCs were plated at a seeding density of 50,000 cells/mL (5000 per well) in a 96-well plate. Control cells were incubated with HCl. Treated cells were incubated with increasing concentrations of activin A (5, 20 or 50 ng/mL) at 37°C for 24 h in room air with 5% CO2. After 24-h incubation, the CellTiter 96 AQueous One Solution Reagent (Cat. No. G3580, Promega) was defrosted and protected from light using foil. A multichannel pipette was used to add 20 μL of the reagent to each well. The plate was gently tapped to mix the reagent with the cells. The plate was placed back into the incubator for 4 h. Absorbance at 490 nm was measured using a plate reader (SpectraMax i3, Molecular Devices).
Angiogenesis assay
To assess whether activin A affects angiogenesis of HAoECs, μ-Plate angiogenesis 96-well plates were utilised (ibidi, Germany, Cat. No. 89646) following the manufacturer’s instructions. Briefly, wells were coated with 10 μL of Matrigel and incubated for 1 h at 37°C. Cells were added at 10,000 per well with control cells incubated with HCl and other cells incubated with increasing concentrations of activin A (5, 20, 50 or 100 ng/mL) at 37°C for 24 h in room air with 5% CO2. Images were taken after 8 and 24 h of incubation. Images were analysed using ImageJ ‘Angiogenesis Analyzer’ software. Various parameters were analysed however only the total number of branches, total branch length, total number of segments and total segment length are presented.
Permeability
To assess whether activin A induces permeability of an endothelial cell monolayer, a FITC-dextran based permeability assay was used. HAoECs were cultured onto gelatinised transwells (Costar® 6.5 mm Transwell®, 0.4 µm Pore Polyester Membrane Inserts, Stemcell Technologies, Australia, Cat. No. 38024) at a density of 60,000 cells per insert in 24-well plates with each well containing 600 µL of media at 37°C for 24 h to form a monolayer. Then, media was replaced and cells treated with HCl or activin A (5, 20, 50 or 100 ng/mL) for 24 or 48 h. After incubation, transwell media was replaced with a working solution of 1 mg/mL of FITC-dextran in HBSS and 20% FBS and placed in a new 24-well plate containing 600 µL of HBSS and 20% FBS in each well. The plate was protected from light and incubated at 37°C for 1 h in room air with 5% CO2. Media was removed from the 24-well plate and placed in a black-walled 96-well plate before fluorescence readings were measured at 485 nm excitation and 535 nm emission.
Cell adhesion and proliferation
Cell adhesion and proliferation were assessed using the xCELLigence Real-Time Cell Analyser system (ACEA Biosciences Inc, San Diego, CA, USA), which continuously monitors cellular behaviour by measuring changes in electrical impedance in the cellular monolayer (reported as cell index). Briefly, cells were plated at a density of 40,000 cells per well in a 16-well E-Plate PET in 200 µL volume with HCl or activin A (5, 20, 50 or 100 ng/mL). Adhesion was measured in the first 5 h, while proliferation was measured over the following 24 h.
Statistical analysis
Results are presented as mean ±s.e.m., with n representing the number of animals or a new passage number. Statistical analysis was performed using one-way ANOVA followed by multiple comparisons using Tukey post hoc test (GraphPad Prism 8.0). Concentration–response curves used non-linear regression analysis was used to calculate the sensitivity of each agonist (pEC50). Maximum relaxation (Rmax) to ACh and SNP were measured as a percentage of the pre-constriction to U46619. Maximum contraction (Emax) to U46619 and ET-1 were measured as a percentage of the level of maximum pre-constriction to U46619. Data were tested for normality using Shapiro–Wilk test. Group pEC50, Rmax, Emax and area under the curve (AUC) were compared using one-way ANOVA with post hoc analysis using Tukey’s multiple comparisons test, unpaired Student’s t-test or Kruskal–Wallis test where appropriate. The relative contribution of NO, prostacyclin (PGI2) and EDH was determined through AUC analysis of ACh-mediated relaxation using a two-way repeated-measures ANOVA for multiple comparisons, followed by Sidak’s post hoc test. P < 0.05 was considered statistically significant.
Results
Localisation of the activin A receptors within mouse aorta and human aortic cells
To determine if activin A had the potential to influence vascular function in the aorta of female mice, immunohistochemistry was used to localise the protein of four different activin A receptors in both the aorta of female mice (Fig. 1) and in human aortic endothelial and smooth muscle cells (Fig. 2). All four receptors investigated were localised in both endothelial and vascular smooth muscle cells of the mouse aorta and human aortic cells, with no localisation in the negative control.
Localization of four different activin A receptors by immunohistochemistry in endothelial cells (ECs) and smooth muscle cells (SMCs) in the aorta of female mice. (A) Type IA receptor (ACTRIA), (B) type IB (ACTRIA), (C) type IIA (ACTRIIA) and (D) type IIB (ACTRIIB). Scale bars = 100 μm.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Localization of four different activin A receptors by immunohistochemistry in endothelial cells (ECs) and smooth muscle cells (SMCs) in the aorta of female mice. (A) Type IA receptor (ACTRIA), (B) type IB (ACTRIA), (C) type IIA (ACTRIIA) and (D) type IIB (ACTRIIB). Scale bars = 100 μm.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Localization of four different activin A receptors by immunohistochemistry in endothelial cells (ECs) and smooth muscle cells (SMCs) in the aorta of female mice. (A) Type IA receptor (ACTRIA), (B) type IB (ACTRIA), (C) type IIA (ACTRIIA) and (D) type IIB (ACTRIIB). Scale bars = 100 μm.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Localization of four different activin A receptors by immunohistochemistry in human aortic endothelial cells: (A) type IA receptor (ACTRIA), (B) type IB (ACTRIA), (C) type IIA (ACTRIIA) and (D) type IIB (ACTRIIB) and in human aortic smooth muscle cells: (E) ACTRIA, (F) ACTRIA, (G) ACTRIIA and (H) ACTRIIB. Scale bars = 50 μm.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Localization of four different activin A receptors by immunohistochemistry in human aortic endothelial cells: (A) type IA receptor (ACTRIA), (B) type IB (ACTRIA), (C) type IIA (ACTRIIA) and (D) type IIB (ACTRIIB) and in human aortic smooth muscle cells: (E) ACTRIA, (F) ACTRIA, (G) ACTRIIA and (H) ACTRIIB. Scale bars = 50 μm.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Localization of four different activin A receptors by immunohistochemistry in human aortic endothelial cells: (A) type IA receptor (ACTRIA), (B) type IB (ACTRIA), (C) type IIA (ACTRIIA) and (D) type IIB (ACTRIIB) and in human aortic smooth muscle cells: (E) ACTRIA, (F) ACTRIA, (G) ACTRIIA and (H) ACTRIIB. Scale bars = 50 μm.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Activin A reduces endothelium-dependent relaxation in the mouse aorta
To assess whether activin A would affect vascular function, the aorta of female mice was incubated in control, or increasing concentrations of activin A. Lower doses of activin A (1–20 ng/mL) did not affect ACh-mediated relaxation. However, 50 ng/mL of activin A significantly (P = 0.028) reduced ACh-mediated relaxation in the aorta (Fig. 3A), as analysed by sensitivity (pEC50, Fig. 3B and Table 1). Activin A did not affect SNP-mediated relaxation (Fig. 3C, D and Table 1), nor ET-1-mediated contraction (Fig. 3E, F and Table 1) nor U46619-mediated contraction (Fig. 3G, H and Table 1).
Concentration–response curves (A) acetylcholine (ACh), (B) sodium nitroprusside (SNP), (C) endothelin-1 (ET-1) and (D) U46619 in the aorta artery of female mice exposed to activin A. This demonstrates that activin A reduces vascular relaxation to ACh but does not enhance vascular contraction in the mouse aorta. Data are shown as mean ± s.e.m. (n = 8). *P < 0.05 sensitivity (pEC50) compared to control.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Concentration–response curves (A) acetylcholine (ACh), (B) sodium nitroprusside (SNP), (C) endothelin-1 (ET-1) and (D) U46619 in the aorta artery of female mice exposed to activin A. This demonstrates that activin A reduces vascular relaxation to ACh but does not enhance vascular contraction in the mouse aorta. Data are shown as mean ± s.e.m. (n = 8). *P < 0.05 sensitivity (pEC50) compared to control.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Concentration–response curves (A) acetylcholine (ACh), (B) sodium nitroprusside (SNP), (C) endothelin-1 (ET-1) and (D) U46619 in the aorta artery of female mice exposed to activin A. This demonstrates that activin A reduces vascular relaxation to ACh but does not enhance vascular contraction in the mouse aorta. Data are shown as mean ± s.e.m. (n = 8). *P < 0.05 sensitivity (pEC50) compared to control.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Reactivity of the aorta after incubation in different concentrations of activin A. Values are expressed as mean ± s.e.m. Statistical analysis is one-way ANOVA.
| Treatment | pEC50† | Rmax (%) | Emax (%) | n |
|---|---|---|---|---|
| ACh | ||||
| Control | 6.68 ± 0.09 | 78.78 ± 2.26 | 9 | |
| 5 ng/mL activin A | 6.73 ± 0.07 | 83.46 ± 3.92 | 4 | |
| 20 ng/mL activin A | 6.68 ± 0.15 | 81.15 ± 3.14 | 4 | |
| 50 ng/mL activin A | 6.23 ± 0.16* | 62.71 ± 9.63 | 8 | |
| SNP | ||||
| Control | 7.71 ± 0.15 | 91.43 ± 2.93 | 9 | |
| 5 ng/mL activin A | 7.71 ± 0.26 | 85.37 ± 12.69 | 4 | |
| 20 ng/mL activin A | 7.35 ± 0.26 | 91.18 ± 5.88 | 4 | |
| 50 ng/mL activin A | 7.76 ± 0.20 | 97.36 ± 3.16 | 8 | |
| ET-1 | ||||
| Control | 8.88 ± 0.28 | 75.93 ± 5.93 | 8 | |
| 50 ng/mL activin A | 8.72 ± 0.26 | 69.96 ± 4.70 | 8 | |
| U46619 | ||||
| Control | 8.29 ± 0.19 | 96.41 ± 7.72 | 8 | |
| 50 ng/mL activin A | 8.16 ± 0.15 | 100.80 ± 6.07 | 8 | |
†Values are Log M dose; *Significantly (P < 0.05) different compared to control.
Emax, maximum contraction as a % of KPSS; pEC50, sensitivity, Rmax, maximum relaxation as % of pre-constriction.
Activin A reduces agonist-stimulated vasodilatory prostanoid pathway
As activin A at 50 ng/mL reduced ACh-mediated relaxation, pharmacological blocking agents were then used to decipher which relaxation pathways had been disrupted. In control treated aorta, prostanoid inhibition significantly (RmaxP = 0.006; AUC P < 0.0001) reduced ACh-mediated relaxation (Fig. 4A, C and Table 2), but did not significantly affect ACh-mediated relaxation in activin A treated aorta (Fig. 4B, D and Table 2). Prostanoid and NOS inhibition significantly (pEC50, Rmax& AUC P < 0.0001) reduced ACh-mediated relaxation in control (Fig. 4A, C and Table 2) and activin A treated aorta (Rmax& AUC P < 0.0001; Fig. 4B, D and Table 2). The relaxation remaining in the presence of Indo+l-NAME is attributed to EDH. EDH relaxation was all but abolished with the addition of IKCa and SKCa channel inhibition in both control and activin A treated aorta (Fig. 4A, B, C, D and Table 2).
Concentration–response curves for acetylcholine (ACh) in the (●) absence (control) or presence of (□) Indo, (▲)Indo+l-NAME, (♦)Indo+l-NAME+TRAM34+Apamin in the aorta of female mice exposed to (A) control or (B) activin A (50 ng/mL) (n = 10). (C) Area under the curve (AUC) after treatment in (C) control or (D) activin A (50 ng/mL) to (E) assess the relative contribution of nitric oxide (NO), prostanoids (PG) and endothelium-derived hyperpolarization (EDH) in acetylcholine-mediated relaxation (n = 10). Control (filled bars) and activin A (grey bars). Activin A treatment resulted in a reduction of PG contribution to ACh-mediated relaxation. Data are shown as mean ± s.e.m. *P < 0.05.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Concentration–response curves for acetylcholine (ACh) in the (●) absence (control) or presence of (□) Indo, (▲)Indo+l-NAME, (♦)Indo+l-NAME+TRAM34+Apamin in the aorta of female mice exposed to (A) control or (B) activin A (50 ng/mL) (n = 10). (C) Area under the curve (AUC) after treatment in (C) control or (D) activin A (50 ng/mL) to (E) assess the relative contribution of nitric oxide (NO), prostanoids (PG) and endothelium-derived hyperpolarization (EDH) in acetylcholine-mediated relaxation (n = 10). Control (filled bars) and activin A (grey bars). Activin A treatment resulted in a reduction of PG contribution to ACh-mediated relaxation. Data are shown as mean ± s.e.m. *P < 0.05.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Concentration–response curves for acetylcholine (ACh) in the (●) absence (control) or presence of (□) Indo, (▲)Indo+l-NAME, (♦)Indo+l-NAME+TRAM34+Apamin in the aorta of female mice exposed to (A) control or (B) activin A (50 ng/mL) (n = 10). (C) Area under the curve (AUC) after treatment in (C) control or (D) activin A (50 ng/mL) to (E) assess the relative contribution of nitric oxide (NO), prostanoids (PG) and endothelium-derived hyperpolarization (EDH) in acetylcholine-mediated relaxation (n = 10). Control (filled bars) and activin A (grey bars). Activin A treatment resulted in a reduction of PG contribution to ACh-mediated relaxation. Data are shown as mean ± s.e.m. *P < 0.05.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Reactivity of the aorta after pharmacological blockade. Values are expressed as mean ± s.e.m. Statistical analysis is one-way ANOVA.
| pEC50 | Rmax (%) | AUC | n | |
|---|---|---|---|---|
| Control | ||||
| ACh- control | 7.17 ± 0.09 | 86.77 ± 2.88 | 179 ± 9.89 | 11 |
| +Indo | 6.79 ± 0.07 | 70.43 ± 3.99* | 121.2 ± 10.3* | 10 |
| +Indo+l-NAME | 5.88 ± 0.27* | 16.19 ± 4.07* | 10.66 ± 4.33* | 10 |
| +Indo+l-NAME+ TRAM-34+Apamin | ND | 0 ± 0* | 0 ± 0* | 4 |
| Activin A | ||||
| ACh- control | 6.59 ± 0.14 | 70.54 ± 6.05 | 109.30 ± 15.56 | 11 |
| +Indo | 6.61 ± 0.15 | 67.73 ± 5.68 | 103.00 ± 1.21 | 10 |
| +Indo+l-NAME | 6.49 ± 0.33 | 6.52 ± 4.24* | 1.82 ± 1.21* | 10 |
| +Indo+l-NAME+ TRAM-34+Apamin | ND | 0 ± 0* | 0 ± 0* | 4 |
*Statistically significant (P < 0.05) different compared to control.
AUC, area under the curve; Emax, maximum contraction as a % of KPSS; pEC50, sensitivity; ND, not determined; Rmax, maximum relaxation as % of pre-construction.
Overall, AUC for relaxation was significantly (P = 0.012) smaller in activin A-treated aorta, demonstrating endothelial dysfunction (Fig. 4E). The AUC analysis further revealed that the contribution of prostanoids to ACh-mediated relaxation was significantly (P= 0.036) reduced in activin A-treated aorta. However, the contribution of NO and EDH was not different based on treatment (Fig. 4E).
Activin A reduces cell viability of human aortic endothelial cells
To further explore the effect of activin A, human aortic cell lines were utilised to assess effects in human models. Cell viability was assessed in human aortic endothelial and smooth muscle cells treated with various concentrations of activin A after 24 and 48 h by measuring the cell metabolism (Fig. 5). Activin A was able to significantly reduce cell viability of endothelial cells at a concentration of 5 ng/mL (P = 0.006), 20 ng/mL (P = 0.002) and 50 ng/mL (P = 0.0004) after 24 (Fig. 5A) and 48 h (P < 0.0001, Fig. 5B). However, activin A did not affect smooth muscle cell viability after 24 (Fig. 5C) or 48 h (Fig. 5D).
Cell viability of human aortic endothelial cells (HAoECs) after (A) 24 h or (B) 48 h treatment in control or activin A (5, 20 or 50 ng/mL) (n = 6–7). Cell viability of human aortic smooth muscle cells (HASMCs) after (C) 24 h or (D) 48 h treatment in control or activin A (5, 20 or 50 ng/mL) (n = 8). Data are shown as mean ± s.e.m. *P < 0.05 relative to control.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Cell viability of human aortic endothelial cells (HAoECs) after (A) 24 h or (B) 48 h treatment in control or activin A (5, 20 or 50 ng/mL) (n = 6–7). Cell viability of human aortic smooth muscle cells (HASMCs) after (C) 24 h or (D) 48 h treatment in control or activin A (5, 20 or 50 ng/mL) (n = 8). Data are shown as mean ± s.e.m. *P < 0.05 relative to control.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Cell viability of human aortic endothelial cells (HAoECs) after (A) 24 h or (B) 48 h treatment in control or activin A (5, 20 or 50 ng/mL) (n = 6–7). Cell viability of human aortic smooth muscle cells (HASMCs) after (C) 24 h or (D) 48 h treatment in control or activin A (5, 20 or 50 ng/mL) (n = 8). Data are shown as mean ± s.e.m. *P < 0.05 relative to control.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Activin A increases endothelial cell permeability while decreasing cell adherence and proliferation
To investigate effects on angiogenesis, tube formation assays were utilised to assess whether activin A injures human aortic endothelial cells (Fig. 6). Activin A did not alter various angiogenesis parameters after 8 or 24 h of treatment, including the total number of branches (Fig. 6A and B), total branch length (Fig. 6C and D), total number of segments (Fig. 6E and F) and total segment length (Fig. 6G and H). In addition, activin A did not affect the total number of junctions, extremities, master joints or nodes (data not shown).
Angiogenesis tube-formation assay of human aortic endothelial cells (HAoECs) after treatment with control or activin A (5, 20, 50 or 100 ng/mL) for 8 h (left) or 24 h (right). Assessment included (A and B) total number of branches, (C and D) total branch length, (E and F) total number of segments and (G and H) total segment length (n = 8). Data are shown as mean ± s.e.m.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Angiogenesis tube-formation assay of human aortic endothelial cells (HAoECs) after treatment with control or activin A (5, 20, 50 or 100 ng/mL) for 8 h (left) or 24 h (right). Assessment included (A and B) total number of branches, (C and D) total branch length, (E and F) total number of segments and (G and H) total segment length (n = 8). Data are shown as mean ± s.e.m.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Angiogenesis tube-formation assay of human aortic endothelial cells (HAoECs) after treatment with control or activin A (5, 20, 50 or 100 ng/mL) for 8 h (left) or 24 h (right). Assessment included (A and B) total number of branches, (C and D) total branch length, (E and F) total number of segments and (G and H) total segment length (n = 8). Data are shown as mean ± s.e.m.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Activin A at 100 ng/mL significantly increased cell permeability after 24 h (P = 0.014, Fig. 7A) and 48 h (P = 0.026, Fig. 7B) of treatment. Next, cellular adherence (Fig. 7C) and proliferation (Fig. 7D) were accessed using the xCELLigence real-time system. Overall, average cell index was significantly (P = 0.031) reduced during cell adhesion after treatment with 100 ng/mL of activin A (Fig. 7E). In addition, cell proliferation was also significantly reduced by activin A treatment of 50 ng/mL (P = 0.029) and 100 ng/mL (P = 0.011, Fig. 7E).
Permeability assay of human aortic endothelial cells (HAoECs) after treatment with control or activin A (5, 20, 50 or 100 ng/mL) for (A) 24 h or (B) 48 h. Cell index of HAoECs after treatment with control or activin A (5, 20, 50 or 100 ng/mL) using the xCELLigence system to monitor (C) cell adhesion and (D) cell proliferation. Average cell index from (E) 0–5 h and (F) 5–24 h of culture. (n = 7–8). Data are shown as mean ± s.e.m. *P < 0.05.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Permeability assay of human aortic endothelial cells (HAoECs) after treatment with control or activin A (5, 20, 50 or 100 ng/mL) for (A) 24 h or (B) 48 h. Cell index of HAoECs after treatment with control or activin A (5, 20, 50 or 100 ng/mL) using the xCELLigence system to monitor (C) cell adhesion and (D) cell proliferation. Average cell index from (E) 0–5 h and (F) 5–24 h of culture. (n = 7–8). Data are shown as mean ± s.e.m. *P < 0.05.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Permeability assay of human aortic endothelial cells (HAoECs) after treatment with control or activin A (5, 20, 50 or 100 ng/mL) for (A) 24 h or (B) 48 h. Cell index of HAoECs after treatment with control or activin A (5, 20, 50 or 100 ng/mL) using the xCELLigence system to monitor (C) cell adhesion and (D) cell proliferation. Average cell index from (E) 0–5 h and (F) 5–24 h of culture. (n = 7–8). Data are shown as mean ± s.e.m. *P < 0.05.
Citation: Reproduction 163, 3; 10.1530/REP-21-0368
Discussion
The present study demonstrates that activin A is capable of inducing endothelial dysfunction in the aorta of mice and in human aortic endothelial cells. Activin A was able to reduce agonist-mediated relaxation in the aorta by reducing prostanoid production. Activin A reduced endothelial cell viability as measured by cell metabolism, decreased cell adherence and proliferation, while increasing permeability. Activin A had no effect on the smooth muscle of mouse aorta, nor did it affect cell viability of human aortic smooth muscle cells, overall suggesting activin A specifically targets the endothelial cells of the vasculature. Therefore, high levels of activin A may also contribute to the vascular dysfunction reported in women with preeclampsia.
The concentrations of activin A used in our study were selected as they represent local circulating levels in the maternal circulation in a healthy pregnancy (5 ng/mL), mildly elevated levels (20 ng/mL), concentrations reported in the third trimester of women diagnosed with early-onset preeclampsia (50 ng/mL) (Lim et al. 2015) or a supraphysiological concentration (100 ng/mL). Excitingly, this study is the first to demonstrate that physiological relevant levels of activin A reported in women with early-onset preeclampsia (50 ng/mL) are able to induce vascular dysfunction in both the mouse aorta and human aortic endothelial cells.
Previous literature has often studied the effect of activin A on transient cells that are a part of the maternal vasculature for a limited time, such as HUVECs. Therefore, this study aimed to assess the effects of activin A on human vascular cells and whole blood vessels that model the maternal vasculature. After establishing that various activin A receptors were present in the aorta of mice and human aortic cells, the effects of activin A on the vasculature were assessed. It was found that circulating concentrations of activin A in the third trimester of women with preeclampsia (50 ng/mL) were capable of significantly reducing endothelial-dependent relaxation in the aorta of female mice. This effect is characteristic of the blood vessels of women with preeclampsia (Goulopoulou & Davidge 2015). Importantly, this study also found that there was no effect on endothelial-independent vascular dysfunction, meaning the smooth muscle cells were unaffected. These results are in line with previous research demonstrating that activin A can induce dysfunction of endothelial cells (Gurusinghe et al. 2014, Lim et al. 2015, Yong et al. 2015, Hobson et al. 2016).
Importantly, we demonstrate the endothelial dysfunction observed in the mouse aorta was due to a reduction in vasodilatory prostanoids expected to be prostacyclin. Prostacyclin is a potent vasodilator produced by blood vessels (Smyth et al. 2009). Prostacyclin concentrations have been reported to be significantly reduced in the urine and plasma of women with preeclampsia (Downing et al. 1980, Remuzzi et al. 1980, Barden et al. 1994). Throughout pregnancy this decrease in prostacyclin is associated with an increase in the vasoconstrictor thromboxane alpha-2 (TxA2), implying that women who go on to develop preeclampsia, favour vasoconstriction in early pregnancy (Chavarría et al. 2003). The reduction in prostacyclin has been hypothesised to be due to the increased oxidative stress in preeclampsia (Walsh 2004). In oncology research, activin A has been found to independently upregulate thromboxane synthase, the TxA2 precursor (Yamashita et al. 1991). In addition, activin A has been demonstrated to stimulate TxA2 in rat bone marrow macrophages (Nüsing & Barsig 1999). It is therefore possible that high levels of circulating activin A contributes to the reduced prostacyclin reported in preeclamptic women, theoretically causing an imbalance between vasodilators and vasoconstrictors, inducing a vasoconstrictive phenotype.
In support, our group has previously demonstrated that activin A induces endothelial dysfunction in HUVECs by increasing expression of vascular cell adhesion molecule one (VCAM-1) and intercellular adhesion molecule one (ICAM1), and by increasing ROS and permeability (Lim et al. 2015, Hobson et al. 2016). In addition, others have demonstrated that activin A inhibits endothelial cell growth and proliferation in HUVECs (McCarthy & Bicknell 1993, Breit et al. 2000, Kaneda et al. 2011) and in gastric cancer specimens (Kaneda et al. 2011). Numerous studies have also reported that activin A is also capable of suppressing angiogenesis in HUVECs (Kaneda et al. 2011, Gurusinghe et al. 2014). However, in contrast, we did not observe any effect of activin A on angiogenesis in HAoECs. We believe this is due to the nature of aortic endothelial cells, which have been shown to exhibit stronger angiogenic potential than HUVECs (Seo et al. 2016). Potentially this affords them some protection implying a higher concentration and/or duration of activin A treatment may be required to affect angiogenesis in these cells. Or, activin A does not affect angiogenesis specifically in endothelial cells of the human aorta.
This study also showed that activin A was unable to significantly decrease aortic smooth muscle cell viability at any dose. Looking to the limited studies on activin A and smooth muscle, activin A has been reported to be a weak mitogen for rat aortic smooth muscle cells (Pawlowski et al. 1997), while also having no effect on rat aortic smooth muscle growth (Nakaoka et al. 1997). These studies support the concept that activin A may not directly induce dysfunction of smooth muscle cells. Moreover, as this study was undertaken in the aorta, it might not be that vascular smooth muscle in general is unaffected during preeclampsia, but rather the aorta is not specifically affected. The aorta is a conduit vessel and is not a part of the systemic vasculature, thus it does not affect blood pressure. Preeclampsia is a disorder of the systemic vasculature (Powe et al. 2011), with systemic vessels being those that regulate blood pressure control. Given that the key clinical feature of preeclampsia is hypertension, it is still possible that the arterial beds involved in blood pressure maintenance and flow are more affected. Future studies are now planned to study the effects of activin A on the systemic vasculature during pregnancy in rodents.
In conclusion, this study provides key further evidence to support the concept that activin A may also contribute to the widespread vascular dysfunction reported in women diagnosed with preeclampsia. Vascular dysfunction caused by activin A is also likely to localised to endothelial cells and may be contributing to this dysfunction via reducing the availability of the vasodilator prostanoid prostacyclin, decreasing endothelial cell adhesion and proliferation, while increasing permeability. Though many questions remain about the effects of high concentrations of activin A in pregnancy, it is possible that activin A may be another antiangiogenic factor involved in the development and progression of preeclampsia.
Declaration of interest
Natalie Hannan is on the editorial board of Reproduction. Natalie Hannan was not involved in the review or editorial process for this paper, on which she is listed as an author. The other authors have nothing to disclose.
Funding
This work was supported by the National Health and Medical Research Council Programme Grant (1113902).
Author contribution statement
S M and E W conceived the experiments. C B, Y Y and S M performed the experiments and analysed the data. C B, Y Y, N H, S M and E W wrote and reviewed the manuscript.
Acknowledgements
The authors acknowledge use of the facilities and technical assistance of Monash Histology Platform, Department of Anatomy and Developmental Biology, Monash University. The authors also acknowledge Monash Micro Imaging, Monash University, for the provision of instrumentation, training and technical support.
References
Barden A, Beilin LJ, Ritchie J, Walters BN & Michael CA 1994 Plasma and urinary endothelin 1, prostacyclin metabolites and platelet consumption in pre-eclampsia and essential hypertensive pregnancy. Blood Pressure 3 38–46. (https://doi.org/10.3109/08037059409101520)
Benschop L, Duvekot JJ & Roeters van Lennep JE 2019 Future risk of cardiovascular disease risk factors and events in women after a hypertensive disorder of pregnancy. Heart 105 1273–1278. (https://doi.org/10.1136/heartjnl-2018-313453)
Bersinger NA, Smarason AK, Muttukrishna S, Groome NP & Redman CW 2003 Women with preeclampsia have increased serum levels of pregnancy-associated plasma protein A (PAPP-A), inhibin A, activin A and soluble E-selectin. Hypertension in Pregnancy 22 45–55. (https://doi.org/10.1081/PRG-120016794)
Bloise E, Ciarmela P, Dela Cruz C, Luisi S, Petraglia F & Reis FM 2019 Activin A in mammalian physiology. Physiological Reviews 99 739–780. (https://doi.org/10.1152/physrev.00002.2018)
Breit S, Ashman K, Wilting J, Rössler J, Hatzi E, Fotsis T & Schweigerer L 2000 The N-myc oncogene in human neuroblastoma cells: down-regulation of an angiogenesis inhibitor identified as activin A. Cancer Research 60 4596–4601.
Cartwright JE, Fraser R, Leslie K, Wallace AE & James JL 2010 Remodelling at the maternal-fetal interface: relevance to human pregnancy disorders. Reproduction 140 803–813. (https://doi.org/10.1530/REP-10-0294)
Chavarría ME, Lara-González L, González-Gleason A, García-Paleta Y, Vital-Reyes VS & Reyes A 2003 Prostacyclin/thromboxane early changes in pregnancies that are complicated by preeclampsia. American Journal of Obstetrics and Gynecology 188 986–992. (https://doi.org/10.1067/mob.2003.203)
D’Antona D, Reis FM, Benedetto C, Evans LW, Groome NP, de Kretser DM, Wallace EM & Petraglia F 2000 Increased maternal serum activin A but not follistatin levels in pregnant women with hypertensive disorders. Journal of Endocrinology 165 157–162. (https://doi.org/10.1677/joe.0.1650157)
Diesch CH, Holzgreve W, Hahn S & Zhong XY 2006 Comparison of activin A and cell-free fetal DNA levels in maternal plasma from patients at high risk for preeclampsia. Prenatal Diagnosis 26 1267–1270. (https://doi.org/10.1002/pd.1606)
Downing I, Shepherd GL & Lewis PJ 1980 Reduced prostacyclin production in pre-eclampsia. Lancet 2 1374. (https://doi.org/10.1016/s0140-6736(8092443-5)
Florio P, Reis FM, Severi FM, Luisi S, Imperatore A, Palumbo MA, Bagnoli F, Gioffre W & Petraglia F 2006 Umbilical cord serum activin A levels are increased in pre-eclampsia with impaired blood flow in the uteroplacental and fetal circulation. Placenta 27 432–437. (https://doi.org/10.1016/j.placenta.2005.04.008)
Fowler PA, Evans LW, Groome NP, Templeton A & Knight PG 1998 A longitudinal study of maternal serum inhibin-A, inhibin-B, activin-A, activin-AB, pro-alphaC and follistatin during pregnancy. Human Reproduction 13 3530–3536. (https://doi.org/10.1093/humrep/13.12.3530)
Goulopoulou S & Davidge ST 2015 Molecular mechanisms of maternal vascular dysfunction in preeclampsia. Trends in Molecular Medicine 21 88–97. (https://doi.org/10.1016/j.molmed.2014.11.009)
Gurusinghe S, Wallace EM & Lim R 2014 The relationship between activin A and anti-angiogenic factors in the development of pre-eclampsia. Pregnancy Hypertension 4 3–6. (https://doi.org/10.1016/j.preghy.2013.09.002)
Harrison CA, Chan KL & Robertson DM 2006 Activin-A binds follistatin and type II receptors through overlapping binding sites: generation of mutants with isolated binding activities. Endocrinology 147 2744–2753. (https://doi.org/10.1210/en.2006-0131)
Hobson SR, Acharya R, Lim R, Chan ST, Mockler J & Wallace EM 2016 Role of activin A in the pathogenesis of endothelial cell dysfunction in preeclampsia. Pregnancy Hypertension 6 130–133. (https://doi.org/10.1016/j.preghy.2016.03.001)
Kaneda H, Arao T, Matsumoto K, De Velasco MA, Tamura D, Aomatsu K, Kudo K, Sakai K, Nagai T & Fujita Y et al.2011 Activin A inhibits vascular endothelial cell growth and suppresses tumour angiogenesis in gastric cancer. British Journal of Cancer 105 1210–1217. (https://doi.org/10.1038/bjc.2011.348)
Karumanchi SA 2016 Angiogenic factors in preeclampsia: from diagnosis to therapy. Hypertension 67 1072–1079. (https://doi.org/10.1161/HYPERTENSIONAHA.116.06421)
Langston-Cox A, Leo CH, Tare M, Wallace EM & Marshall SA 2020 Sulforaphane improves vascular reactivity in mouse and human arteries after ‘preeclamptic-like’ injury. Placenta 101 242–250. (https://doi.org/10.1016/j.placenta.2020.09.001)
Lim R, Acharya R, Delpachitra P, Hobson S, Sobey CG, Drummond GR & Wallace EM 2015 Activin and NADPH-oxidase in preeclampsia: insights from in vitro and murine studies. American Journal of Obstetrics and Gynecology 212 86.e1–86. e 12. (https://doi.org/10.1016/j.ajog.2014.07.021)
Lowe SA, Bowyer L, Lust K, McMahon LP, Morton M, North RA, Paech M & Said JM 2015 SOMANZ guidelines for the management of hypertensive disorders of pregnancy 2014. Australian and New Zealand Journal of Obstetrics and Gynaecology 55 e1–29. (https://doi.org/10.1111/ajo.12399)
Marshall SA, O’Sullivan K, Ng HH, Bathgate RAD, Parry LJ, Hossain MA & Leo CH 2017 B7-33 replicates the vasoprotective functions of human relaxin-2 (serelaxin). European Journal of Pharmacology 807 190–197. (https://doi.org/10.1016/j.ejphar.2017.05.005)
Marshall SA, Senadheera SN, Jelinic M, O’Sullivan K, Parry LJ & Tare M 2018 Relaxin deficiency leads to uterine artery dysfunction during pregnancy in mice. Frontiers in Physiology 9 255. (https://doi.org/10.3389/fphys.2018.00255)
Marshall SA, Qin CX, Jelinic M, O’Sullivan K, Deo M, Walsh J, Li M, Parry LJ, Ritchie RH & Leo CH 2020 The novel small-molecule annexin-A1 mimetic, compound 17b, elicits vasoprotective actions in streptozotocin-induced diabetic mice. International Journal of Molecular Sciences 21 E1384. (https://doi.org/10.3390/ijms21041384)
Maynard SE, Min JY, Merchan J, Lim KH, Li J, Mondal S, Libermann TA, Morgan JP, Sellke FW & Stillman IE et al.2003 Excess placental soluble fms-like tyrosine kinase 1 (sFlt1) may contribute to endothelial dysfunction, hypertension, and proteinuria in preeclampsia. Journal of Clinical Investigation 111 649–658. (https://doi.org/10.1172/JCI17189)
McCarthy SA & Bicknell R 1993 Inhibition of vascular endothelial cell growth by activin-A. Journal of Biological Chemistry 268 23066–23071. (https://doi.org/10.1016/S0021-9258(1949426-X)
Mol BWJ, Roberts CT, Thangaratinam S, Magee LA, de Groot CJM & Hofmeyr GJ 2016 Pre-eclampsia. Lancet 387 999–1011. (https://doi.org/10.1016/S0140-6736(1500070-7)
Muttukrishna S, Hyett J, Paine M, Moodley J, Groome N & Rodeck C 2006 Uterine vein and maternal urinary levels of activin A and inhibin A in pre-eclampsia patients. Clinical Endocrinology 64 469–473. (https://doi.org/10.1111/j.1365-2265.2006.02476.x)
Nakaoka T, Gonda K, Ogita T, Otawara-Hamamoto Y, Okabe F, Kira Y, Harii K, Miyazono K, Takuwa Y & Fujita T 1997 Inhibition of rat vascular smooth muscle proliferation in vitro and in vivo by bone morphogenetic protein-2. Journal of Clinical Investigation 100 2824–2832. (https://doi.org/10.1172/JCI119830)
Nüsing RM & Barsig J 1999 Induction of prostanoid, nitric oxide, and cytokine formation in rat bone marrow derived macrophages by activin A. British Journal of Pharmacology 127 919–926. (https://doi.org/10.1038/sj.bjp.0702626)
Park HJ, Shim SS & Cha DH 2015 Combined screening for early detection of pre-eclampsia. International Journal of Molecular Sciences 16 17952–17974. (https://doi.org/10.3390/ijms160817952)
Pawlowski JE, Taylor DS, Valentine M, Hail ME, Ferrer P, Kowala MC & Molloy CJ 1997 Stimulation of activin A expression in rat aortic smooth muscle cells by thrombin and angiotensin II correlates with neointimal formation in vivo. Journal of Clinical Investigation 100 639–648. (https://doi.org/10.1172/JCI119575)
Petraglia F, De Vita D, Gallinelli A, Aguzzoli L, Genazzani AR, Romero R & Woodruff TK 1995 Abnormal concentration of maternal serum activin-A in gestational diseases. Journal of Clinical Endocrinology and Metabolism 80 558–561. (https://doi.org/10.1210/jcem.80.2.7852520)
Powe CE, Levine RJ & Karumanchi SA 2011 Preeclampsia, a disease of the maternal endothelium: the role of antiangiogenic factors and implications for later cardiovascular disease. Circulation 123 2856–2869. (https://doi.org/10.1161/CIRCULATIONAHA.109.853127)
Reddy A, Suri S, Sargent IL, Redman CW & Muttukrishna S 2009 Maternal circulating levels of activin A, inhibin A, sFlt-1 and endoglin at parturition in normal pregnancy and pre-eclampsia. PLoS ONE 4 e4453. (https://doi.org/10.1371/journal.pone.0004453)
Remuzzi G, Marchesi D, Zoja C, Muratore D, Mecca G, Misiani R, Rossi E, Barbato M, Capetta P & Donati MB et al.1980 Reduced umbilical and placental vascular prostacyclin in severe pre-eclampsia. Prostaglandins 20 105–110. (https://doi.org/10.1016/0090-6980(8090010-6)
Seo HR, Jeong HE, Joo HJ, Choi SC, Park CY, Kim JH, Choi JH, Cui LH, Hong SJ & Chung S et al.2016 Intrinsic FGF2 and FGF5 promotes angiogenesis of human aortic endothelial cells in 3D microfluidic angiogenesis system. Scientific Reports 6 28832. (https://doi.org/10.1038/srep28832)
Shahul S, Ramadan H, Nizamuddin J, Mueller A, Patel V, Dreixler J, Tung A, Lang RM, Weinert L & Nasim R et al.2018 Activin A and late postpartum cardiac dysfunction among women with hypertensive disorders of pregnancy. Hypertension 72 188–193. (https://doi.org/10.1161/HYPERTENSIONAHA.118.10888)
Smith GC, Stenhouse EJ, Crossley JA, Aitken DA, Cameron AD & Connor JM 2002 Early pregnancy levels of pregnancy-associated plasma protein A and the risk of intrauterine growth restriction, premature birth, preeclampsia, and stillbirth. Journal of Clinical Endocrinology and Metabolism 87 1762–1767. (https://doi.org/10.1210/jcem.87.4.8430)
Smyth EM, Grosser T, Wang M, Yu Y & FitzGerald GA 2009 Prostanoids in health and disease. Journal of Lipid Research 50 (Supplement) S423–S428. (https://doi.org/10.1194/jlr.R800094-JLR200)
Staff AC 2019 The two-stage placental model of preeclampsia: an update. Journal of Reproductive Immunology 134–135 1–10. (https://doi.org/10.1016/j.jri.2019.07.004)
Stoikos CJ, Salamonsen LA, Hannan NJ, O’Connor AE, Rombauts L & Dimitriadis E 2010 Activin A regulates trophoblast cell adhesive properties: implications for implantation failure in women with endometriosis-associated infertility. Human Reproduction 25 1767–1774. (https://doi.org/10.1093/humrep/deq097)
Venkatesha S, Toporsian M, Lam C, Hanai J-i, Mammoto T, Kim YM, Bdolah Y, Lim KH, Yuan HT & Libermann TA et al.2006 Soluble endoglin contributes to the pathogenesis of preeclampsia. Nature Medicine 12 642–649. (https://doi.org/10.1038/nm1429)
Walsh SW 2004 Eicosanoids in preeclampsia. Prostaglandins, Leukotrienes, and Essential Fatty Acids 70 223–232. (https://doi.org/10.1016/j.plefa.2003.04.010)
Wang Q, Keutmann HT, Schneyer AL & Sluss PM 2000 Analysis of human follistatin structure: identification of two discontinuous N-terminal sequences coding for activin A binding and structural consequences of activin binding to native proteins. Endocrinology 141 3183–3193. (https://doi.org/10.1210/endo.141.9.7675)
Wijayarathna R & de Kretser DM 2016 Activins in reproductive biology and beyond. Human Reproduction Update 22 342–357. (https://doi.org/10.1093/humupd/dmv058)
Wu P, Haththotuwa R, Kwok CS, Babu A, Kotronias RA, Rushton C, Zaman A, Fryer AA, Kadam U & Chew-Graham CA et al.2017 Preeclampsia and future cardiovascular health: a systematic review and meta-analysis. Circulation: Cardiovascular Quality and Outcomes 10 e003497. (https://doi.org/10.1161/CIRCOUTCOMES.116.003497)
Yamashita T, Shimizu T & Ogata E 1991 Activin A/erythroid differentiation factor induces thromboxane A2 synthetic activity in murine erythroleukemia cells. Journal of Biological Chemistry 266 3888–3892. (https://doi.org/10.1016/S0021-9258(1967877-4)
Yong HE, Murthi P, Wong MH, Kalionis B, Cartwright JE, Brennecke SP & Keogh RJ 2015 Effects of normal and high circulating concentrations of activin A on vascular endothelial cell functions and vasoactive factor production. Pregnancy Hypertension 5 346–353. (https://doi.org/10.1016/j.preghy.2015.09.006)
