Prenatal Melatonin Modulates Cardiovascular Function and Oxidative Stress in Guinea Pig Neonates Under Normoxic and Hypoxic Gestation

Abstract

Introduction: Gestational hypoxia (GH) increases the risk of cardiovascular diseases by inducing oxidative stress and vascular dysfunction. This study investigates whether prenatal melatonin can mitigate these effects in guinea pigs. Methods: Pregnant guinea pigs were exposed to normoxia or hypoxia and treated with melatonin (1 mg/kg/day). Echocardiography, vascular reactivity, and molecular assays were used to assess cardiovascular structure, function, and redox balance in neonates. Results: GH reduced neonatal birth weight and altered left ventricular (LV) development, resulting in increased LV systolic function and aortic blood flow velocity. Melatonin treatment reversed these effects, restoring endothelial-dependent vasodilation and decreasing oxidative stress in the LV and thoracic aorta. Catalase antioxidant enzyme activity was elevated in melatonin-treated hypoxic neonates. Unexpectedly, melatonin treatment altered cardiac structure in normoxic pregnancies, increasing LV length and decreasing LV myocardial nuclei density. Conclusions: Prenatal melatonin partially modulates GH-induced endothelial dysfunction and oxidative stress, offering potential therapeutic value. However, its effects under normoxic conditions deserve caution, emphasizing the need for targeted use only in pregnancies with evident hypoxic and oxidative stress conditions.

1. Introduction

Cardiovascular diseases (CVDs) affect the function and structure of the heart and blood vessels and are currently the leading cause of global mortality [1]. Importantly, the incidence of CVDs has remained stable in middle- and low-income countries, contributing to a quarter of worldwide deaths in 2021 [2,3]. The main risk factors for CVDs are a sedentary lifestyle, obesity, and hypertension, and the latter has a prevalence of up to 65% in CVD patients [4]. While these conditions usually occur during adulthood, adverse prenatal conditions have been associated with an increased risk of cardiovascular diseases, sparking increasing interest in understanding how prenatal factors contribute to long-term cardiovascular risk [5].

Gestational hypoxia (GH) is defined as an inadequate supply of oxygen during fetal development [6]. This condition can manifest in several pregnancy complications, such as pre-eclampsia, placental insufficiency, and gestation at high altitudes, where the affected offspring has shown increased susceptibility to CVDs in adulthood [7]. During GH, a fetal hemodynamic response allows its survival [6]. The fall in partial pressure of oxygen (PO₂) induces a mechanism known as the fetal brain-sparing response, characterized by norepinephrine-induced and catecholamine-maintained blood flow redistribution to essential circulation, such as cerebral or adrenal vasculature, allowing fetal survival [8]. However, sustained catecholamine release in peripheral territories reduces metabolic growth, shaping fetal growth restriction (FGR; also termed intrauterine growth restriction), defined as birth weight below the 10th percentile, which has deleterious cardiovascular effects during adulthood [8,9]. In addition to FGR, other significant effects of GH are cardiac remodeling and endothelial dysfunction. For instance, GH induced cardiac hypoplasia and ventricular dilation in mice [10], while it reduced left ventricle (LV) diastolic function in full-term guinea pig fetuses [11]. In addition, rat and chicken embryo models of GH have shown reduced endothelium-dependent relaxation in femoral arteries [12,13]. Similarly, in fetal lambs exposed to GH, an increase in femoral vascular resistance and arterial blood pressure was observed [14,15]. Furthermore, in an FGR guinea pig model, a reduced endothelial-dependent vasodilation was observed in the thoracic aorta and femoral arteries in fetuses, associated with a reduced heart weight [16,17]. Importantly, these deleterious effects may persist in adulthood. Endothelial dysfunction in aortas from fetal rats exposed to GH persisted in 8- and 16-week-olds, and was also observed in the thoracic aorta and femoral arteries from 8- and 16-month-old FGR and GH gestated one-year-old guinea pigs [17,18,19,20]. These findings highlight that GH has a significant impact on the left heart circuit, which is critical for systemic blood circulation during adulthood [21].

These GH-induced mechanisms have been associated with oxidative stress (OS), where the production of reactive oxygen species (ROS) surpasses the cellular antioxidant response [7]. ROS are pro-oxidative molecules produced as principal products or byproducts by enzymatic machinery, such as mitochondrial electron transport chain complexes, or nicotinamide adenine dinucleotide phosphate (NADPH) oxidases (NOX) and uncoupled endothelial nitric oxide (NO) synthase (eNOS) enzymes [22]. These molecules are classified as one-electron oxidants, such as free radical superoxide (O₂⁻) and hydroxyl radical (HO⁻), and two-electron oxidants, such as nonradical hydrogen peroxide (H₂O₂) [22]. Furthermore, nitrosative species, such as peroxynitrite (ONOO⁻), are derived from ROS and NO reactions [22]. This is relevant at the cardiovascular level, as it not only induces OS but also decreases the bioavailability of NO, thereby reducing endothelium-dependent vasodilation [23]. On the other hand, the antioxidant machinery is composed of catalase (CAT), glutathione peroxidase (GPX), hemeoxygenase (HMOX/HO), and superoxide dismutase (SOD) isoforms, which are induced transcriptionally mainly by nuclear factor erythroid 2-related factor 2 (NRF2) [24]. At basal levels, ROS play a physiological role; however, high levels of ROS lead to DNA damage, protein oxidation, and lipid oxidation, among other pathophysiological disruptions [25]. The importance of OS on GH-induced effects has been previously demonstrated in several studies. For instance, GH induced an increase in nitrosative stress, lipid peroxidation, and ROS production in the hearts of fetal guinea pigs and rats [26,27,28]. In addition, long-term GH-induced aortic endothelial dysfunction was also associated with increased O₂⁻ and H₂O₂ production [18,20]. Significantly, aortic endothelial dysfunction was associated with an increase in nitrosative stress, an effect that was abolished by antioxidant treatment with vitamin C in rats exposed to GH [29]. Similarly, aortic endothelial dysfunction was reversed by antioxidant treatment with N-acetylcysteine in FGR guinea pig fetuses [16]. Therefore, these findings suggest that antioxidant treatments during pregnancy could represent a promising therapeutic strategy to mitigate the cardiovascular effects induced by GH-associated oxidative stress. However, these treatments must be safe for both the mother and the developing fetus.

In this context, melatonin, a neurohormone synthesized at the pineal gland, which has antioxidant properties and a wide therapeutic window, has been proposed as a solid therapeutic strategy for CVDs [30]. This lipophilic hormone has membrane receptors, type 1 (MT1) and type 2 (MT2), and it exerts antioxidant effects through its cytosolic receptor (MT3), which is a quinone oxidoreductase [31]. Additionally, it induces antioxidant enzymes via its orphan nuclear receptors (ROR/RZR) or through the nuclear translocation of NRF2 [31,32,33]. Furthermore, through direct interaction with ROS, melatonin acts as a radical scavenger, and unlike most small-molecule antioxidants (e.g., ascorbic acid, α-tocopherol, and lipoic acid), it does not undergo redox cycling, acting as a terminal scavenger and preventing it from promoting oxidation, although this mechanism has been debated in biological systems [34]. On the other hand, melatonin can reach the fetus through the placenta, where it modulates fetal development, and a redox improvement with its exogenous administration has been recognized in models of pregnancy complications [35]. Therefore, in the present study, we propose to evaluate the effects of prenatal melatonin on the central cardiovascular health of neonatal guinea pigs exposed to gestational hypoxia.

2. Methods

All procedures were evaluated and authorized by the Institutional Animal Care and Use Committee (CICUA) of Universidad de Chile (certificate 20354-MED-UCH). Furthermore, the study adhered to the ARRIVE guidelines and was conducted in accordance with the U.K. Animals (Scientific Procedures) Act, the E.U. Directive on animal experimentation, and the NIH Guide for the Care and Use of Laboratory Animals.

2.1. Animal Model

Pirbright White guinea pigs (Cavia porcellus) were used as the animal model. Their translational value for DOHaD hypothesis studies has been reviewed elsewhere, which highlights their resemblance with human physiology regarding the placentation process, hormonal levels, small litter size, and perinatal cardiac maturation, in comparison to other murine models [36,37].

2.2. Experimental Design

All animal care and handling protocols described below were performed by a veterinarian with expertise in this area. Pregnancy was confirmed between gestational days (GD) 20 and 25 through visualization of the gestational sac. At GD30, 24 pregnant guinea pigs were equally separated and subjected either to normobaric normoxia (570 m; 760 Torr) or to hypobaric hypoxia (3800 m; 470 Torr) during the rest of gestation. Parallelly, both normoxic (N) and hypoxic (H) groups were further divided into two groups of equal size: vehicle (water) or melatonin (1 mg/kg/day; 0.01% ethanol (EtOH) in water; M5250, Sigma-Aldrich, St. Louis, MO, USA). Melatonin was delivered via drinking water, ensuring a steady increase in its levels throughout the day. Therefore, we designed four experimental groups: normoxic (N) and hypoxic (H) vehicle-treated animals, and normoxic (NM) and hypoxic (HM) melatonin-treated animals (Figure A1). Both oxygen and treatment conditions were maintained throughout the remainder of gestation until delivery.

Melatonin dosage was determined through a preliminary exploratory assessment comparing 1 mg/kg and 10 mg/kg doses. While control animals displayed an endogenous daily variation of approximately 1–5 fold, the 1 mg/kg dose achieved a physiologically relevant circadian amplitude (approx. 4–50 fold increase over controls). By contrast, the 10 mg/kg dose markedly amplified daily variation beyond endogenous limits (7–90 fold) and was therefore excluded to avoid supraphysiological or non-physiological effects.

To ensure dose standardization and prevent confounding variables associated with hypoxia-induced polydipsia, water intake was monitored daily. Melatonin concentration was standardized based on a mean body weight of 700 g and a conservative minimum intake threshold of 150 mL/day to satisfy the 1 mg/kg requirement. This concentration remained constant throughout the treatment, and compliance was verified by daily weighing of the bottles to ensure all animals met or exceeded the calculated volume for full dosage. Water was provided ad libitum in light-protected bottles, with volumes adjusted to always exceed the maximum expected daily intake for each group. The solution was renewed every 24 h to maintain compound stability. This chronic administration method was prioritized to achieve sustained bioavailability. Throughout the experimental design, all animals were housed under standard conditions (35–40% humidity, 20–21 °C, and a 12:12 h light–dark cycle) and fed with standard chow (LabDiet 5025, LabDiet, St. Louis, MO, USA, 40 g/day).

2.3. In Vivo Approaches

2.3.1. Monitoring of Maternal Body Weight

During gestation, body weight (BW) was recorded daily in the pregnant dams, expressed as an absolute value, and normalized by BW gain: $\left(\frac{Final BW \left(g\right) \times 100}{Initial BW \left(g\right)}\right)-100\%$.

2.3.2. Ultrasound Assessment

Ultrasound echocardiography and vascular doppler assessments in guinea pig fetuses were evaluated weekly from GD45 until GD70 (near term). Unsedated pregnant dams were placed on a pillow in dorsal decubitus and immobilized gently by a highly experienced technician. Fetal cardiovascular structure and function were determined in all of the fetuses (2–3 per litter) using a portable ultrasonographer (Z6 VET, Mindray, Shenzhen, China) with an L14-6P transducer (Z60 L14-6P, Mindray).

Total cardiac length and width, as RV and LV structural parameters, were measured using B-mode imaging. In addition, LV functional parameters, such as end-diastolic thickness (EDT), end-systolic thickness (EST), end-diastolic diameter (EDD), end-systolic diameter (ESD), shortening fraction (SF), mitral E/A index, Tei index, and mitral annular plane systolic excursion (MAPSE), were assessed using M-mode imaging. Additionally, vascular structural and functional parameters, such as the diameter of the left ventricular outflow tract (LVOT), were assessed using B-mode imaging. Aortic maximal (Vmax) and mean (Vmean) blood flow velocities, as well as acceleration time, were determined using pulsed-wave Doppler mode. All of the mentioned variables were averaged from three consecutive measurements performed during each ultrasound examination.

2.4. Ex Vivo Approaches

2.4.1. Euthanasia and Cardiovascular Tissue Collection

At birth, the newborns were returned to normobaric normoxia conditions. The total number of pups was 13 (6 dams), 17 (6 dams), 19 (6 dams), and 15 (6 dams) for the N, H, NM, and HM groups, respectively. One animal from each litter was then selected, weighed, and assigned to its corresponding experimental group. Each group consisted of 6 neonates, equally divided between females and males. The remaining animals from each litter were returned to their mother for future studies.

The one-day old newborns were euthanized with an overdose of sodium pentobarbital (1 mg/kg; Sigma-Aldrich; P3761) via intraperitoneal injection. Subsequently, the heart was removed, weighed, and cut transversely, and the upper section was fixed in 4% formaldehyde (Sigma-Aldrich; F8775) to perform histological assays, while the lower section was separated into the RV and LV and subsequently frozen at −80 °C to perform molecular biology assays. In addition, the descending thoracic aorta was separated from the aortic isthmus and sectioned into ~6 mm rings, which were fixed in 4% formaldehyde for histological assays or maintained in continuously oxygenated Krebs for vascular reactivity assays.

2.4.2. Vascular Reactivity Assessment

Aortic sections were cut into 2 mm rings, mounted on a wire myograph (Multi-Myograph System 620 M, Danish Myo Technology, Hinnerup, Denmark), and maintained in Krebs buffer under a constant supply of a gas mixture composed of 95% O₂ and 5% CO₂ throughout the protocol. The arterial rings were stretched radially and their tension standardized using the Mulvany–Halpern protocol, achieving a passive tension equivalent to a physiological neonatal arterial systolic pressure (~60 mmHg), based on the mean arterial pressure range of 50–55 mmHg reported in neonatal guinea pigs [38,39].

Subsequently, the maximum contractile capacity was evaluated through a modified Krebs buffer curve with increasing concentrations of potassium chloride (KCl; 4.75–125 mM). In addition, the contractile response to phenylephrine (PE; P6126, Sigma-Aldrich), as well as the endothelium-dependent vasodilatory response to methacholine (MetCh; A2251, Sigma-Aldrich) and the endothelium-independent response to sodium nitroprusside (SNP; 71778, Sigma-Aldrich), previously pre-contracted with PE (10⁻⁵ M), was evaluated through cumulative concentration–response curves (CCRCs; 10⁻¹⁰–10⁻⁴ M). To assess the contribution of NO to endothelium-dependent vasodilation, before a MetCh CCRC, the arteries were preincubated for 10 min with L-NG-Nitro arginine methyl ester (L-NAME; 10⁻⁵ M; N5751, Sigma-Aldrich), an inhibitor of NO synthase isoforms. Each CCRC assay was evaluated by measuring the response two minutes after each concentration increase, with at least 30 min of resting in Krebs buffer between tests to ensure vascular recovery.

2.5. In Vitro Approaches

2.5.1. Neonatal Cardiovascular Structure

Cardiac and aortic sections were fixed in 4% formaldehyde for 24 h and then washed in 1× phosphate-buffered saline (PBS) and maintained in 50% EtOH in distilled water. These sections were subsequently dehydrated, embedded in paraffin, sectioned, and mounted on 5 μm thick slides. These slides were incubated overnight at 37 °C, then deparaffinized and hydrated. Afterward, the slides were stained with hematoxylin-eosin (H&E) and Van Gieson, as previously reported [40,41].

2.5.2. Immunohistochemistry Assays

Sections of 5 μm thickness, previously deparaffinized and hydrated, were subjected to antigen retrieval in citrate buffer (10 mM sodium citrate and 0.05% Tween 20 in distilled water, pH 6.0) by heating at 100 °C for 40 min, followed by cooling for 30 min in cold citrate buffer. Sections were incubated for 15 min with an endogenous peroxidase blocker. Cardiac sections were incubated overnight at 4 °C with a specific primary antibody against the antioxidant-related transcription factor, NRF2 (1:400; ab31163, Abcam, Cambridge, UK), and both cardiac and aortic sections were incubated with specific primary antibodies against the OS markers 3-nitrotyrosine (3-NT; 1:500; A-21285, Invitrogen, Carlsbad, CA, USA) and 4-hydroxy-2-nonenal (4-HNE; 1:400; ab46545, Abcam). Subsequently, the presence of NRF2 and OS markers was detected using a commercial IHC kit (Mouse/Rabbit PolyDetector Plus DAB HRP Brown Kit, BSB0261, BioSB, Goleta, CA, USA) following the manufacturer’s protocol. Additionally, after DAB revelation, the slides were counterstained with hematoxylin (26754, Electron Microscopy Sciences, Hatfield, PA, USA) for a 1 min incubation, blued with 1% ammonia for 10 s, dehydrated (in reverse of the hydration steps mentioned above), and mounted with a synthetic resin solution [20].

Levels of NRF2 and OS markers were determined in the myocardium of the free walls of the ventricles (≥3 fields per ventricle) and in the intima-media of the aorta (≥4 fields per aorta) by quantifying the pixels colored in the brown range (DAB precipitate, which indicates the presence of the protein/marker of interest) relative to the total tissue area evaluated in the sample, using 40X microphotographs (BX41, Olympus; Tokyo, Japan) and Adobe Photoshop CS6 (Adobe Inc., San Jose, CA, USA), as previously described [20]. In addition, NRF2 nuclear localization at the LV myocardium (≥3 fields per ventricle) was assessed using the Fiji package of ImageJ 1.54 g software by quantifying DAB-positive nuclei to total nuclei [42].

2.5.3. Transcriptional Expression Evaluations

Antioxidant enzyme mRNA levels were evaluated from ~20 mg of frozen RV and LV tissue. Total RNA was extracted using an acid-guanidinium-phenol-based reagent (TRIzol, 15596026, Invitrogen) following the manufacturer’s protocol. RNA concentration was determined using a spectrophotometer (Genesys 10uv, Thermo Spectronic, Waltham, MA, USA) by measuring the absorbance at 260 nm, while purity was assessed through the 260/280 nm absorbance ratio, with acceptable values ranging from 1.8 to 2.0. Additionally, RNA integrity was evaluated using a fluorometer (Qubit 4, Q33238, Invitrogen) and a commercial kit (Qubit RNA IQ Assay Kit, Q33221, Invitrogen). All samples displayed RNA integrity scores of 8/10 or higher. Subsequently, 2 µg of total RNA was used as a template for cDNA synthesis using a reverse transcriptase enzyme (M-MLV Reverse Transcriptase, M1701, Promega, Madison, WI, USA), following the manufacturer’s protocol [43].

To perform each qPCR assay, 15 ng of newly synthesized cDNA was used with a commercial DNA polymerase mix (KiCqStart SYBR Green qPCR ReadyMix, KCQS01, Sigma-Aldrich), following the manufacturer’s instructions. Specific primers for different gene transcripts related to the antioxidant response were used in each assay (Table A1). Reactions were carried out in a real-time thermal cycler (AriaMX, G8830A, Agilent, Santa Clara, CA, USA) using the following thermal cycling program: an initial denaturation step at 95 °C for 10 min, followed by 40 cycles of 15 s at 95 °C, 20 s at 55–60 °C, and 20 s at 72 °C, ending with a standard melting curve analysis. To determine the transcriptional levels of each transcript evaluated, these were quantified using the 2^−ΔΔCt method, relative to the reference gene transcript β-actin (Bact) [44].

2.5.4. Immunoblot Assays

To evaluate antioxidant protein expression levels, ~20 mg of LV tissue was resuspended in RIPA buffer (20 mM Tris, 150 mM NaCl, 1 mM EDTA, 0.8% Triton X-100, and 0.8% deoxycholate) supplemented with a commercial protease inhibitor cocktail (Halt Protease Inhibitor Cocktail, 78430, Thermo Fisher, Waltham, MA, USA). The samples were homogenized and centrifuged at 10,000× g for 20 min at 4 °C, and the supernatant was recovered. Protein concentration in the supernatants was determined using the Bradford method (Protein Assay Dye Reagent Concentrate, 5000006, Bio-Rad, Hercules, CA, USA) with a microplate reader (Sunrise Absorbance Reader, Tecan, Grödig, Austria). Absorbance values were interpolated on a standard curve prepared with bovine serum albumin (BSA; Pierce BSA Standard, 2 mg/mL; 23210, Thermo Fisher), following the manufacturer’s instructions. Protein concentrations were normalized to 2 μg/μL using denaturing buffer (62.5 mM Tris, pH 6.8, 0.01% bromophenol blue, 1.43 mM 2-mercaptoethanol, 2% SDS, and 0.1% glycerol).

For immunoblot assays, 20 μg of denatured protein (95 °C, 10 min) from each sample was loaded onto 10–14% bis-acrylamide gels (40% Acrylamide/Bis Solution, 37.5:1, 1610148, Bio-Rad), separated by electrophoresis, and transferred onto nitrocellulose membranes (1215471, GVS Filter Technology, Sanford, ME, USA). Given that a total of 24 samples were analyzed, they were loaded in batches of 12 samples at a time, with one batch consisting of vehicle-treated samples and the other composed of melatonin-treated samples. To enable comparison between two immunoblots, a loading control (LC) sample was included, which remained consistent across all immunoassays.

Equivalent total protein loading was confirmed by staining membranes with 0.1% Ponceau S in 5% acetic acid (Ponceau S, 114275, Sigma-Aldrich). The stain was removed by washing with 1X PBS before blocking with 5% BSA (Probumin, 81003, Sigma-Aldrich) for one hour at room temperature. Subsequently, the membranes were incubated overnight at 4 °C with specific primary antibodies against antioxidant proteins and the reference protein α/β-tubulin (TUBA) (Table A2). Next, membranes were washed with 0.05% Tween 20 in 1X PBS and incubated with anti-mouse (1:2000, 115035062, Jackson ImmunoResearch, West Grove, PA, USA) or anti-rabbit (1:2000, 111035045, Jackson ImmunoResearch) secondary antibodies at 4 °C for 1 h. Immunostaining was performed using chemiluminescent reagents (SuperSignal West Pico Plus; 34080; Thermo Scientific and SuperSignal West Femto, 34095; Thermo Scientific). Chemiluminescent images were captured using an imaging system (Alliance Q9 Advanced, Uvitec, Cambridge, UK), and protein levels were quantified via densitometry (Image Studio Lite, LI-COR, Lincoln, NE, USA), with normalization to the TUBA protein signal for each sample [20].

2.5.5. Antioxidant Enzyme Activity

Cardiac antioxidant enzyme activity was determined from ~20 mg of RV and LV frozen tissue, which was resuspended and homogenized in phosphate buffer (50 mM KH₂PO₄, 1 mM EDTA, and pH 7.0). The tissue was then centrifuged at 10,000 RCF for 15 min at 4 °C, and the supernatant was collected. The total activities of CAT (Catalase Assay Kit, 707002, Cayman Chemical, Ann Arbor, MI, USA), GPX (Glutathione Peroxidase Assay Kit, 703102, Cayman Chemical) and SOD (Superoxide Dismutase Assay Kit, 706002, Cayman Chemical) were evaluated, following the manufacturer’s instructions. The absolute enzyme activity was normalized to the total protein concentration in the samples, which was quantified using the Bradford method, as previously described [45,46].

2.6. Statistical Analyses

All data analyses were performed using statistical software (GraphPad Prism 8.02). In general, data were checked for normal distribution using both Shapiro–Wilk and Kolmogorov–Smirnov tests. Additionally, potential outliers were identified using the Grubbs test and excluded from subsequent comparisons. When appropriate, all data were expressed as mean ± SEM.

2.6.1. Sample Size

The primary outcome was a reduction in oxidative stress, reflected by a decrease in 3-NT levels in the left ventricle. The sample size was determined a priori using statistical power analysis software (G*Power 3.1.9.7) for a two-way ANOVA (Gestational Environment × Treatment), with α = 0.05, power = 0.80, a large effect size (f = 0.60), and a numerator df of 1 (4 groups). The calculation yielded n = 6 animals per group (total n = 24). In addition, a post hoc power calculation using the same parameters (f = 0.60, α = 0.05, and n = 6 per group) yielded an achieved power = 0.816, supporting the adequacy of the group size.

2.6.2. In Vivo Data—Ultrasound Measurements

For the analysis of ultrasound data over time, extra sum-of-squares F-tests were used. Initially, a comparison of nonlinear model fits (first-order linear vs. second-order polynomial) was conducted to determine the best fit for each group. If the four groups fit different models, it was inferred that they exhibited distinct developmental patterns of the variables. Conversely, if the four groups fit the same model, an additional nonlinear curve-fitting analysis was performed to evaluate whether a single shared curve equation could describe the datasets. If one of the groups did not fit the same curve equation, further curve-fitting analyses were conducted to determine whether the differences were attributable to the slope or the intercept. These comparisons used the slopes and intercepts of the normoxic and hypoxic vehicle-treated groups as references.

Additionally, ultrasound values were segmented into two groups: the first (G45–50) and last (G65–70) weeks of echocardiographic measurement to facilitate appropriate comparisons of group differences during the early and late stages of fetal development.

2.6.3. Ex Vivo Data—Wire Myography

The potassium chloride contraction curves (tension, mN) were fitted to a sigmoidal curve with a Boltzmann model, from which the maximum contraction values (KCl Emax) and sensitivity (V50) to the different potassium concentrations were determined.

On the other hand, both the contraction (induced by PE) and the relaxation (induced by MetCh or SNP) CCRCs were adjusted to a sigmoidal dose–response model, obtaining the values of maximum response (Emax) and pharmacological potency (i.e., sensitivity, −LogEC50, pD₂) of each group [20]. In addition, the relationship between SNP and MetCh Emax values was evaluated using Spearman’s rank correlation coefficient on paired data. Finally, NO-dependent and -independent contributions to MetCh-induced vasodilatation were calculated with MetCh area under the curve (AUC) values: $NO dependent = Control CCRC AUC \left(total\right)- LNAME CCRC AUC \left(NO independent\right)$.

2.6.4. In Vitro Data—General Analyses

To determine whether there were differences in the remaining parameters between groups, two- and three-way ANOVAs were performed to evaluate main effects (gestational age, hypoxia and melatonin treatment) and interactions. When ANOVA indicated significant effects or interactions (p ≤ 0.05), post hoc pairwise comparisons were performed using the False Discovery Rate (FDR) with Benjamini, Krieger, and Yekutieli corrections to identify significant differences (q ≤ 0.05) [47,48].

3. Results

3.1. Gestational Hypoxia Alters Neonatal Body Weight

The body weight of pregnant guinea pig dams was assessed during gestation, and we observed similar maternal net weights and weight gain (Figure A2A,B) between groups. Furthermore, GH reduced the birth weight of guinea pigs exclusively in the vehicle-treated group, compared to its normoxic counterpart (Figure A2C). In addition, the sex distribution per litter, expressed as the percentage of females, was as follows: N (36.12 ± 11.72%), H (50.00 ± 14.28%), NM (45.83 ± 20.83%), and HM (50.00 ± 18.76%), without significant differences between groups.

3.2. Gestational Hypoxia and Melatonin Alter Fetal Cardiac Structure

During fetal life, both gestational hypoxia and prenatal melatonin treatment were found to alter cardiac development. Comparison of the echocardiographic parameters between the first and last weeks of evaluation revealed that total cardiac volume, as well as RV and LV length, increased progressively with gestation. This increase was significantly magnified by melatonin treatment in both groups, an effect that was detectable from the initial week exclusively in the normoxic group (Table 1). In the melatonin-treated group, gestational hypoxia reduced total and LV length in the initial week and decreased RV length across both the initial and final weeks of evaluation (Table 1). Regarding cardiac width parameters (Table 1), total cardiac width, along with RV and LV width, increased similarly with gestation across all experimental groups, despite an initial melatonin-induced increase in total cardiac width. Gestational hypoxia increased LV width compared to normoxic controls in both the vehicle and melatonin groups. Notably, within the hypoxic cohort, melatonin treatment induced a specific reduction in RV width (at both initial and final weeks) and a reduction in LV width in the final week. The remaining parameters remained similar between groups.

Table 1.

Melatonin affects fetal cardiac structure during both normoxic and hypoxic gestations. Values of total, right ventricle (RV) and left ventricle (LV) length and width, and LV end-diastolic thickness (EDT), end-systolic thickness (EST), end-diastolic diameter (EDD), and end-systolic diameter (ESD) during initial (GD45–50) and final (GD65–70) ultrasound examinations of fetal guinea pigs exposed to gestational normoxia (red sections) or hypoxia (blue sections) and treated with vehicle or melatonin. Values are expressed as mean ± SEM. n = 13(6)–19(6) pups(dams) per group. Data were analyzed using three-way ANOVA with FDR post hoc test. Significant differences (q ≤ 0.05): ^& vs. initial-counterpart, * vs. normoxic-counterpart, ^# vs. vehicle-counterpart.

		Vehicle		Melatonin
Parameter (mm)	Ultrasound Examination	Normoxia	Hypoxia	Normoxia	Hypoxia
Total Length	Initial	7.129 ± 0.296	7.633 ± 0.183	10.28 ± 0.250 #	8.771 ± 0.354 #*
	Final	13.01 ± 0.743 &	13.55 ± 0.420 &	15.51 ± 0.550 &#	15.08 ± 0.444 &#
RV Length	Initial	3.500 ± 0.192	3.780 ± 0.142	5.066 ± 0.142 #	4.006 ± 0.213 *
	Final	6.344 ± 0.449 &	6.200 ± 0.291 &	7.576 ± 0.266 &#	6.952 ± 0.274 &#*
LV Length	Initial	3.764 ± 0.135	3.845 ± 0.139	5.329 ± 0.182 #	4.553 ± 0.297 *
	Final	6.175 ± 0.340 &	6.679 ± 0.357 &	8.255 ± 0.295 &#	7.931 ± 0.305 &#
Total Width	Initial	5.871 ± 0.283	6.327 ± 0.163	7.703 ± 0.220 #	7.082 ± 0.247
	Final	10.20 ± 0.436 &	10.99 ± 0.250 &	10.91 ± 0.293 &	10.96 ± 0.292 &
RV Width	Initial	1.907 ± 0.097	2.408 ± 0.107	2.042 ± 0.100	1.903 ± 0.131 #
	Final	3.569 ± 0.193 &	3.922 ± 0.219 &	3.157 ± 0.177 &	3.270 ± 0.135 &#
LV Width	Initial	2.143 ± 0.132	2.487 ± 0.126	2.063 ± 0.126	2.050 ± 0.148
	Final	3.244 ± 0.156 &	4.232 ± 0.265 &*	3.045 ± 0.206 &	3.575 ± 0.186 &#*
LV EDT	Initial	0.859 ± 0.059	0.714 ± 0.036	0.776 ± 0.052	0.756 ± 0.027
	Final	1.114 ± 0.053 &	1.048 ± 0.049 &	1.048 ± 0.056 &	1.033 ± 0.044 &
LV EST	Initial	1.086 ± 0.061	0.971 ± 0.035	0.828 ± 0.057 #	0.866 ± 0.030
	Final	1.623 ± 0.085 &	1.629 ± 0.065 &	1.202 ± 0.055 &#	1.203 ± 0.051 &#
LV EDD	Initial	2.062 ± 0.120	2.033 ± 0.125	2.887 ± 0.142	2.692 ± 0.179
	Final	3.310 ± 0.307 &	3.490 ± 0.154 &	3.930 ± 0.195 &	3.897 ± 0.152 &
LV ESD	Initial	1.179 ± 0.144	1.079 ± 0.100	2.221 ± 0.194 #	1.943 ± 0.128 #
	Final	1.846 ± 0.194 &	1.960 ± 0.111 &	3.168 ± 0.195 &#	2.845 ± 0.127 &#

Furthermore, the development of LV end-diastolic and end-systolic thicknesses (EDT and EST, respectively) and end-diastolic and end-systolic diameters (EDD and ESD, respectively) during fetal life was assessed. It was observed that LV EDT increased with gestation similarly across all groups (Table 1). By contrast, while LV EST also increased with gestation in all groups, melatonin treatment attenuated these values compared to their vehicle counterparts. Finally, both LV EDD and LV ESD increased with gestation across all groups. However, melatonin treatment increased LV ESD values from the initial week through the final week of gestation compared to their vehicle counterparts (Table 1). The remaining parameters remained similar between groups.

3.3. Melatonin Modulates Fetal Cardiovascular Function over Time

Fetal cardiovascular function was evaluated using left ventricular function analysis. For this, the functional indicators of diastole (mitral E/A index), systole (MAPSE), myocardial performance (Tei index), and shortening fraction were determined. While we observed similar values in the mitral E/A index and Tei index during gestation (Figure 1A,C), GH accelerated MAPSE increase during gestation, reflected by an increased slope, compared to the normoxic group (Figure 1B). Importantly, melatonin treatment induced an alteration in the shortening fraction exclusively in the hypoxic group, as reflected by an increased slope and a reduced intercept during gestation (Figure 1D). Comparing the parameters during gestation, it was observed that MAPSE increased with gestation across all groups, yet both GH and melatonin treatment increased this parameter in the final week compared to vehicle-normoxic controls (Table 2). In addition, the shortening fraction was reduced by melatonin treatment in both groups from the initial week to the final week (Table 2). The remaining parameters remained similar between groups.

Figure 1.

Melatonin affects central cardiovascular function during both normoxic and hypoxic gestations. Mitral E/A index (A), MAPSE (B), left ventricle (LV) Tei index (C), LV shortening fraction (D), and aortic mean (E) and maximal (F) blood flow velocity of guinea pig fetuses exposed to normoxia (N; red) or hypoxia (H; blue) and treated with vehicle (dots) or melatonin (squares) during gestation. Values are expressed as mean ± SEM. n = 13(6)–19(6) pups(dams) per group. Data were analyzed using an F-test. Significant differences (p ≤ 0.05): * vs. normoxic-counterpart, ^# vs. vehicle-counterpart, and b0: intercept-associated change, b1: slope-associated change.

Table 2.

Melatonin affects central cardiovascular function during both normoxic and hypoxic gestations. Values of mitral E/A index, MAPSE, left ventricle (LV) Tei index, shortening fraction, aortic acceleration time, blood flow mean (Vmean) and maximal (Vmax) velocity and LV outflow tract diameter (LVOT) during initial (GD45–50) and final (GD65–70) ultrasound examinations of fetal guinea pigs exposed to gestational normoxia (red sections) or hypoxia (blue sections) and treated with vehicle or melatonin. Values are expressed as mean ± SEM. n = 13(6)–19(6) pups(dams) per group. Data were analyzed using three-way ANOVA with FDR post hoc test. Significant differences (q ≤ 0.05): ^& vs. initial-counterpart, * vs. normoxic-counterpart, ^# vs. vehicle-counterpart.

		Vehicle		Melatonin
Parameter	Ultrasound	Normoxia	Hypoxia	Normoxia	Hypoxia
Mitral E/A Index	Initial	0.517 ± 0.023	0.568 ± 0.021	0.652 ± 0.045	0.622 ± 0.025
	Final	0.609 ± 0.028	0.643 ± 0.032	0.719 ± 0.093	0.718 ± 0.040
MAPSE (mm)	Initial	1.146 ± 0.067	1.348 ± 0.112	1.325 ± 0.122	1.380 ± 0.102
	Final	1.561 ± 0.129 &	1.979 ± 0.078 &*	2.267 ± 0.328 &#	1.960 ± 0.129 &
LV Tei Index	Initial	0.549 ± 0.027	0.566 ± 0.031	0.508 ± 0.052	0.572 ± 0.054
	Final	0.601 ± 0.019	0.603 ± 0.020	0.573 ± 0.043	0.533 ± 0.030
Shortening Fraction (%)	Initial	45.34 ± 4.004	49.57 ± 3.561	25.77 ± 3.282 #	27.06 ± 2.418 #
	Final	42.64 ± 4.198	43.39 ± 2.514	18.77 ± 3.203 #	25.32 ± 1.283 #
Acceleration Time (ms)	Initial	27.72 ± 2.601	24.58 ± 2.099	30.45 ± 1.248	29.10 ± 1.096
	Final	20.76 ± 3.352	26.06 ± 2.831	35.92 ± 4.626 #	32.029 ± 2.70
Aortic Vmean (cm/s)	Initial	20.65 ± 2.142	22.75 ± 1.433	16.11 ± 0.921	16.68 ± 0.524 #
	Final	22.18 ± 2.220	27.19 ± 1.804 *	17.61 ± 1.020	17.47 ± 0.955 #
Aortic Vmax (cm/s)	Initial	33.08 ± 3.189	31.98 ± 2.319	26.19 ± 1.479	24.53 ± 0.814 #
	Final	37.11 ± 3.063	40.33 ± 2.554 &	27.36 ± 1.786 #	27.78 ± 1.569 #
LVOT (mm)	Initial	1.157 ± 0.035	1.263 ± 0.040 #	1.750 ± 0.073	1.417 ± 0.059 *
	Final	2.063 ± 0.076 &	1.981 ± 0.049 &	2.300 ± 0.138 &#	2.254 ± 0.073 &#

On the other hand, central vascular function was assessed by measuring aortic mean (Vmean) and maximal (Vmax) blood flow velocities. Importantly, GH was found to accelerate the progressive increase in both aortic Vmax and Vmean during gestation, as reflected by an increased slope, an effect that was abolished by melatonin treatment (Figure 1E,F). In addition, melatonin treatment also reduced the aortic Vmax developmental curve intercept compared to the hypoxic group (Figure 1F). Comparing the parameters during the initial and final evaluation weeks revealed that melatonin treatment increased the acceleration time exclusively in the normoxic group in the final week (Table 2). Furthermore, although GH increased the aortic Vmean in the final week, melatonin treatment reduced this effect, with the difference becoming evident as early as the initial week (Table 2). A gestational age-dependent increase in the aortic Vmax was observed exclusively in the hypoxic group. Melatonin treatment reduced this increase throughout the initial to final examination weeks in the hypoxic group, whereas it induced a similar reduction in the normoxic group only in the final week (Table 2). LVOT diameter increased with gestation across all groups, and this gestational increase was exacerbated by melatonin treatment (Table 2). Interestingly, GH increased the LVOT diameter during the initial week, an effect observed exclusively in the melatonin-treated group. The remaining parameters remained similar between groups.

3.4. Melatonin Alters Neonatal Cardiac Structure

Cardiac weight (N: 0.614 ± 0.065; H: 0.643 ± 0.063; NM: 0.610 ± 0.050; HM: 0.578 ± 0.046% to total BW) was similar between groups. In addition, the cardiac area, as well as the RV and LV luminal areas, and the thicknesses of the septum, RV, and LV, remained similar between groups (Figure 2A–G). However, melatonin treatment decreased the LV myocardial nuclei density exclusively in the normoxic group (Figure 2H), whereas a sustained cardiomyocyte cross-sectional area was observed between groups (Figure 2I).

Figure 2.

Melatonin alters cardiac structure during normoxic gestation. Representative photographs (A) and values of cardiac area (B), right ventricle (RV; (C)) and left ventricle (LV; (D)) luminal area, septum (E), RV (F) and LV (G) thickness of hearts, and nuclei density (H) and cardiomyocyte cross-section area (I) in the left ventricle from newborn guinea pigs exposed to gestational normoxia (N; red) or hypoxia (H; blue) and treated with vehicle or melatonin during gestation. Values are expressed as mean ± SEM. n = 5–6 independent samples (one per pup) per group. Data were analyzed using two-way ANOVA with FDR post hoc test. Significant differences (q ≤ 0.05): ^# vs. vehicle-counterpart. Scale bar: 100 µm.

3.5. Melatonin Compensates Gestational Hypoxia-Induced Neonatal Cardiac Oxidative Stress

GH increased the levels of the nitrosative stress marker 3-NT, melatonin treatment exhibited a trend toward (p = 0.075) decreasing this effect (Figure 3A,B). On the other hand, 4-HNE levels remained similar between groups (Figure 3C). While total NRF2 levels remained sustained between groups (N: 0.344 ± 0.019; H: 0.335 ± 0.023; NM: 0.306 ± 0.035; HM: 0.321 ± 0.028 pixels/µm²), melatonin treatment increased the levels of NRF2-positive nuclei exclusively in the normoxic animals (Figure 3D). Interestingly, melatonin treatment increased CAT activity exclusively in the hypoxic group, which exhibited higher antioxidant activity than its normoxic and vehicle-treated counterparts (Figure 3E). On the other hand, melatonin treatment increased GPX activity exclusively in the normoxic group (Figure 3F). By contrast, SOD activity remained constant in different experimental groups (Figure 3G).

Figure 3.

Melatonin alters cardiac oxidative balance during both normoxic and hypoxic gestations. Representative photographs (A) and levels of 3-nitrotyrosine (3-NT; (B)), 4-hydroxinonenal (4-HNE; (C)), NRF2 positive/total nuclei (D), and catalase (CAT; (E)), glutathione peroxidase (GPX; (F)), and superoxide dismutase (SOD; (G)) activity in left ventricle from newborn guinea pigs exposed to gestational normoxia (N; red) or hypoxia (H; blue) and treated with vehicle or melatonin during gestation. Values are expressed as mean ± SEM. n = 5–6 independent samples (one per pup) per group. Data were analyzed using two-way ANOVA with FDR post hoc test. Significant differences (q ≤ 0.05): * vs. normoxic-counterpart, ^# vs. vehicle-counterpart. Scale bar: 100 µm.

Furthermore, melatonin treatment increased Gpx1 levels exclusively in the normoxic group (Table 3). In addition, Hmox1 transcriptional levels were increased by melatonin treatment, and this effect was reversed by gestational hypoxia, resulting in decreased levels compared to their normoxic counterparts (Table 3). The remaining mRNA levels evaluated remain similar between groups.

Table 3.

Melatonin alters transcriptional levels of antioxidant enzymes during both normoxic and hypoxic gestations. Transcriptional levels (expressed as fold-change to vehicle-treated normoxic group) of catalase (Cat), glutathione peroxidase isoform-1 (Gpx1), -2 (Gpx2), -3 (Gpx3) and -4 (Gpx4), superoxide dismutase isoform-1 (Sod1) and -2 (Sod2) and heme oxygenase-1 (Hmox1) at the left ventricle of newborn guinea pigs exposed to gestational normoxia (red sections) or hypoxia (blue sections) and treated with vehicle or melatonin. Values are expressed as mean ± SEM. n = 5–6 independent samples (one per pup) per group. Data were analyzed using two-way ANOVA with FDR post hoc test. Significant differences (q ≤ 0.05): * vs. normoxic-counterpart, ^# vs. vehicle-counterpart.

	Vehicle		Melatonin
mRNA	Normoxia	Hypoxia	Normoxia	Hypoxia
Cat	1.000 ± 0.444	1.198 ± 0.382	2.527 ± 1.044	1.514 ± 0.168
Gpx1	1.000 ± 0.277	3.654 ± 1.395	12.086 ± 5.322 #	4.275 ± 0.558
Gpx2	1.000 ± 0.184	2.439 ± 0.782	3.188 ± 1.003	4.576 ± 2.68
Gpx3	1.000 ± 0.372	2.870 ± 1.297	11.993 ± 5.904	1.513 ± 0.403
Gpx4	1.000 ± 0.376	1.197 ± 0.415	6.041 ± 2.707	1.702 ± 0.593
Sod1	1.000 ± 0.240	1.960 ± 1.001	6.895 ± 3.457	1.593 ± 0.477
Sod2	1.000 ± 0.393	3.106 ± 0.908	5.134 ± 2.350	3.831 ± 1.383
Hmox1	1.000 ± 0.182	2.807 ± 0.783	5.512 ± 2.084 #	1.819 ± 0.635 *

Interestingly, at the protein level, melatonin treatment reduced CAT levels exclusively in the hypoxic group, compared to its vehicle-treated counterpart (Figure A3A,B). In addition, melatonin treatment reduced GPX1/2 and SOD1 protein levels under both gestational conditions, compared to its vehicle-treated counterpart (Figure A3C,D). Importantly, GH reduced SOD1 levels in vehicle-treated animals (Figure A3D). Furthermore, melatonin treatment increased SOD2 levels exclusively in the normoxic group (Figure A3E). Finally, GH reduced HO-1 protein levels exclusively in the melatonin-treated group (Figure A3F).

3.6. Melatonin Compensates Gestational Hypoxia-Induced Neonatal Vascular Dysfunction

When evaluating the neonatal aortic structure, it was observed that the parameters studied, including luminal and total diameter and intima-media and adventitia thickness and occupation, remained similar between groups (Figure 4A–G). On the other hand, while hypoxia did not affect OS markers in the aorta, melatonin treatment decreased 3-NT levels exclusively in the normoxic group (Figure 4H). In addition, 4-HNE levels were reduced by melatonin treatment under both conditions (Figure 4I). Furthermore, 4-HNE levels remained higher in the melatonin-treated hypoxic group compared to its normoxic counterpart (Figure 4I).

Figure 4.

Melatonin alters oxidative status in neonatal aorta. Representative photographs (A) and values of luminal (B) and total (C) diameter, intima-media (D) and adventitia (E) thickness, intima-media (F) and adventitia (G) occupation, as levels of 3-nitrotyrosine (3-NT; (H)), 4-hydroxinonenal (4-HNE; (I)) in thoracic aortas from newborn guinea pigs exposed to gestational normoxia (N; red) or hypoxia (H; blue) and treated with vehicle or melatonin during gestation. Values are expressed as mean ± SEM. n = 4–6 independent samples (one per pup) per group. Data were analyzed using two-way ANOVA with FDR post hoc test. Significant differences (q ≤ 0.05): * vs. normoxic-counterpart, ^# vs. vehicle-counterpart. Scale bar: 1000 µm for structure and 100 µm for oxidative stress markers microphotographs.

When evaluating the vasoreactivity of the aortas, it was observed that the KCl-induced maximal contractile capacity of the aortas was increased by GH, an effect that was sustained in the melatonin group (Figure 5A–C). This was accompanied by a reduction in melatonin sensitivity to KCl exclusively in the hypoxic group, compared to its vehicle counterpart (Figure 5C). Moreover, phenylephrine Emax was increased exclusively in the normoxic group treated with melatonin (N: 139.12 ± 10.404; H: 109.99 ± 7.728; NM: 85.73 ± 11.752; HM: 83.11 ± 6.269% constriction to 64.86 mM KCl). Furthermore, sensitivity to PE was reduced by GH and melatonin treatment, effects observed exclusively in comparison to the normoxic-untreated group (N: 5.99 ± 0.147; H: 5.56 ± 0.061; NM: 5.05 ± 0.097; HM: 5.26 ± 0.141).

Figure 5.

Melatonin compensates gestational hypoxia-induced endothelial dysfunction. Concentration–response curves, maximal response (Emax), and sensitivity (V50/pD₂) to potassium chloride (KCl; (A–C)), sodium nitroprusside (SNP; (D–F)) and metacholine (MetCh; (G–I)), scatter plot of Emax values to MetCh and SNP (J), and values of NO contribution to vasodilation (K) in thoracic aortas from newborn guinea pigs exposed to gestational normoxia (N; red) or hypoxia (H; blue) and treated with vehicle (solid lines) or melatonin (dashed lines) during gestation. Values are expressed as mean ± SEM. n = 4–6 independent samples (one per pup) per group. Data were analyzed using two-way ANOVA with FDR post hoc test. Significant differences (q ≤ 0.05): * vs. normoxic-counterpart, ^# vs. vehicle-counterpart.

On the other hand, melatonin treatment reduced the maximal response to the endothelium-independent vasodilator SNP exclusively in the hypoxic group, an effect that was reversed by melatonin treatment (Figure 5D,E). In addition, melatonin treatment increased the sensitivity to SNP exclusively in the hypoxic counterpart (Figure 5F). Importantly, GH decreased the maximal response to the endothelium-dependent vasodilator MetCh, an effect that was fully reversed by melatonin treatment (Figure 5G,H). Additionally, melatonin treatment also increased MetCh Emax in normoxic animals compared to their vehicle counterparts (Figure 5H). Furthermore, sensitivity to MetCh remained constant among the different groups (Figure 5I). Interestingly, Spearman’s rank correlation analysis revealed a significant positive correlation between SNP Emax (independent variable) and MetCh Emax, indicating a moderate association between endothelium-independent and endothelium-dependent vasodilatory responses (Figure 5J). Notably, offspring treated with melatonin exhibited the highest responses to both SNP and MetCh, clustering at the upper end of the regression line (Figure 5J). These effects were exclusively associated with a melatonin-induced increase in the NO-independent contribution in both the normoxic and hypoxic groups (Figure 5K).

4. Discussion

The present study investigated the impact of antenatal melatonin treatment on the perinatal cardiovascular health of guinea pigs exposed to gestational hypoxia, a condition characterized by oxidative stress and an increased risk of cardiovascular diseases in adulthood [5,7]. The results showed that melatonin treatment during gestation was able to reverse some of the deleterious effects of GH at multiple levels; however, it also exhibited some effects under non-pathological normoxic conditions.

4.1. Prenatal Melatonin Effects in the Gestational Hypoxia Context

A consistent finding is that GH reduces birth weight, as observed in our study and in multiple reports from animal models and high-altitude human populations [5]. Importantly, GH has been associated with alterations in maternal metabolism, including increased insulin sensitivity, reduced maternal adipose tissue, and decreased gluconeogenic activity, which together limit the energy reserves required for fetal development [49]. Moreover, this hypoxic environment not only limits fetal development by a metabolic alteration, but also alters both maternal and fetal hemodynamics [11]. Regarding maternal effects, GH induces a lower proliferative capacity, an alteration of myogenic tone, and insufficient blood flow in uterine arteries [50,51,52]. On the other hand, the fetal hemodynamic response maintains perfusion in essential vascular beds such as the brain, associated with an increase in peripheral vascular resistance and a lower metabolic capacity [8]. Interestingly, several studies have linked oxidative stress with a reduction in birth weight as another pregnancy complication [53]. Thus, the absence of this effect in the melatonin-treated group may suggest its ability to attenuate oxidative stress and hemodynamic alterations induced by hypoxia.

During pregnancy, GH affects the fetal pattern of LV width development, increasing this parameter in late gestation. Interestingly, these changes were associated with similar development of LV EDT and EDD, as well as the mitral E/A index, with similar values at the end of gestation, suggesting that LV diastolic function was preserved during hypoxia in our fetal guinea pigs. These findings differ from those observed in other models of gestational hypoxia. In full-term fetal guinea pigs, GH increases the LV mitral E/A index [11]. Furthermore, GH exhibits an increased LV EDD, associated with reduced contractility of LV muscle bundles in near-term chicken embryos [54]. These differences could be due to the contrasting hypoxic conditions to which the animals were exposed, where an oxygen availability of 10.5% (~5500 m) was used for the guinea pig model, and an absent contribution of the maternal response was found in the chicken model, increasing the severity of the hypoxic episode [11,54]. Moreover, an increased LV mitral E/A index has been observed in ≥34 weeks SGA fetuses and pre-eclamptic pregnancies, compared to non-complicated pregnancies, suggesting a reduced LV diastolic function [55,56]. By contrast, FGR neonates exhibited a reduced LV EDD compared to the control group, implying a lower diastolic capacity [57]. These findings suggest that multiple factors, including environmental conditions and structural or functional limitations, may have contributed to the preservation of LV diastolic function in our guinea pig model.

On the other hand, GH caused several changes at the LV systolic level, affecting the development of MAPSE, which, at the end of gestation, had higher values compared to the normoxic group. These adaptation mechanisms are associated with an increase in peripheral vascular resistance, a phenomenon observed in several GH animal models [12,58,59]. Therefore, this hemodynamic alteration increases LV afterload and consequently induces a rise in LV contractile capacity, maintaining cardiac output [5]. This is reflected by an increase in LV width, a preserved LV ESD and EST, a more positive slope in MAPSE development, and an increased MAPSE value at the end of gestation, all of which are indicative of enhanced systolic function. This suggests a disruption of cardiac functional homeostasis, as the RV, which normally carries the highest fetal workload, now requires an augmented LV contribution to compensate for the increased afterload induced by GH [60,61]. While these outcomes may be beneficial for the developing fetus under gestational stress, they may induce long-term maladaptive mechanisms, such as ventricular hypertrophy and subsequent systolic dysfunction after birth [62]. Therefore, these results suggest that the observed increase in systolic function during gestation may reflect adaptive mechanisms that counteract peripheral vascular resistance, as reported in multiple animal models under conditions of gestational hypoxia [12,58,59].

Importantly, melatonin treatment modulated GH-induced alterations at multiple levels. It markedly affected cardiac structure, increasing total heart length in both ventricles during gestation, with this effect persisting in the LV at term. Notably, ventricular elongation decreases longitudinal stress and increases circumferential stress, thereby influencing contractile performance [63,64,65]. In line with this, fetal guinea pigs treated with melatonin exhibited lower LV EST and higher ESD at the end of gestation and were further characterized by lower LV SF, suggesting a reduction in contractile capacity. Importantly, this cardiomegaly may originate from constant volume overload [65,66]. This hemodynamic condition may result from a high-output circulatory state, potentially driven by low systemic vascular resistance, as seen in cases of excessive NO release [67,68]. Melatonin treatment increases blood flow and reduces vascular resistance through various mechanisms, including increased endothelium-dependent vasodilation, antioxidant defense, and cholinergic activity, and decreased adrenergic tone [30,69,70]. These findings suggest that melatonin treatment reduces vascular resistance, potentially leading to volume overload and contributing to the observed cardiac changes. Importantly, cardiac elongation is a recognized phenotype in FGR, and unlike hypertrophic remodeling, which emerges early in gestation, the elongated phenotype is generally associated with better outcomes [64,65]. These findings suggest that melatonin’s potential modulation of vascular dynamics during gestation could have significant implications for fetal heart development. Further research is needed to determine whether these changes contribute to maladaptive responses under hypoxic and non-pathological conditions.

At birth, cardiac weight, structure, cardiomyocyte nuclei density, and CSA were similar across all groups in neonatal guinea pigs. These results indicate that there is an adaptive response to GH during fetal life that does not affect the structure of the heart in immediate postnatal life. Similarly, the cardiac structure remained sustained in rabbit fetal hearts exposed to prenatal hypoxia by uteroplacental artery ligation [71]. Moreover, the LV structure remained unaltered in 90-day-old GH-gestated guinea pigs [72]. However, GH generates deleterious effects in cardiac structure over the long term, as reflected by an increased LV luminal area in one-year-old female guinea pigs [20]. These findings suggest that even without evident changes early on, GH may trigger latent defects leading to long-term cardiac dysfunction.

Importantly, the main GH outcome observed is oxidative stress [7]. In our model, GH increased 3-NT levels in the LV, associated with unchanged NRF2 localization and similar transcript and activity levels of antioxidant enzymes. Similarly, other GH models in fetal guinea pigs observed an increase in 3-NT in the hearts of the hypoxic animals, associated with an increase in ONOO⁻ generation by inducible nitric oxide synthase (iNOS), reflecting nitrosative stress, a well-established component of the broader oxidative stress, as a common outcome from GH [26,73]. Moreover, under unstressed conditions, basal levels of NRF2 remain low, mainly due to its interaction with KEAP1, an adaptor protein for the CUL3-RBX1 ubiquitin ligase complex that induces NRF2 ubiquitination and subsequent proteasomal degradation [74]. However, under oxidative stress, the reaction of ROS with KEAP1 sensor cysteines allows the release of NRF2 and its subsequent nuclear translocation, where it induces the expression of antioxidant enzymes genes, such as CAT, GPX, HMOX and SOD isoforms [75]. The increase in oxidative stress in the LV does not appear to align with enhanced NRF2 activity. Interestingly, KEAP1-independent mechanisms, such as glycogen synthase kinase-3β (GSK-3β), which becomes activated under chronic hypoxia, can suppress NRF2/ARE-mediated antioxidant responses in several cell types, including cardiomyocytes [76,77]. These findings suggest a potential target for studying the reduced antioxidant capacity outcome in this gestational condition.

Importantly, melatonin treatment modified the redox balance in the LV of neonatal guinea pigs, increasing CAT activity exclusively in hypoxic-treated animals. This occurred despite unchanged NRF2 nuclear localization and CAT transcription, and was accompanied by reduced CAT protein levels, suggesting enhanced specific activity of the enzyme. Our previous report indicated that melatonin treatment increased CAT activity in the cardiac tissue of newborn lambs exposed to GH [78]. The importance of CAT lies in its capacity to degrade H₂O₂ to water and oxygen, and its catalytic activity depends on several factors, including the formation of inactive intermediates and binding to its structural cofactor, heme [79]. The iron ion from the heme pool is altered by oxidative stress, forming either ferrous (Fe²⁺) or ferric (Fe³⁺) oxidation states, as well as by enzymatic degradation, for example, by Hmox proteins, to produce biliverdin, ferrous iron, and carbon monoxide [80]. Importantly, melatonin treatment reduced Hmox1 transcriptional and protein levels exclusively in the hypoxic group. Thus, melatonin’s scavenging effects, together with the reduction in Hmox1 levels, may increase heme availability and thereby augment CAT catalytic efficiency; however, future studies should determine whether a causal relationship exists between these mechanisms [81]. Furthermore, melatonin treatment increased GPX activity in the normoxic group, but reduced GPX protein levels, despite enhanced NRF2 nuclear localization and increased GPX1 transcription. In the hypoxic group, melatonin treatment decreased GPX protein levels, accompanied by sustained enzyme activity, suggesting improved enzyme efficiency, similar to that observed with CAT. GPX generates water and oxygen from H₂O₂ in a reaction where reduced glutathione (GSH) is oxidized (GSSG) [82]. Melatonin treatment has been observed to increase GSH levels, as a decreased GSH:GSSG ratio, in several experimental systems [83]. Therefore, melatonin boosts antioxidant capacity in the context of GH at the cardiac level, which is important, as the deleterious effects of oxidative stress on the heart have been observed in multiple hypoxia-related studies [5,7,78]. By increasing the activity of these enzymes, melatonin helps balance oxidative stress, potentially mitigating the long-term negative effects of GH on cardiac function.

On the other hand, GH induced notable alterations in central vasculature during both gestation and the postnatal period. In parallel with the increase in LV systolic function, GH accelerated the development of aortic blood-flow dynamics, with Vmean and, to a lesser extent, Vmax being elevated during gestation, and Vmean remaining higher than in normoxic fetuses at term. Aortic velocity has a positive correlation with MAPSE, suggesting that LV systolic function also conditions aortic hemodynamics [84]. Thus, the enhanced fetal systolic function likely increases aortic blood-flow velocity to sustain systemic circulation, a response that during GH occurs in the context of elevated peripheral vascular resistance [62]. Importantly, we observed a similar intima-media thickness in newborn animals; however, we previously reported an increased intima-media thickness in the aortas of one-year-old female guinea pigs gestated in hypoxia [20]. Previous studies have reported increased arterial stiffness and other deleterious outcomes associated with this type of stress in FGR guinea pigs [85]. These findings indicate that GH-induced fetal cardiovascular changes may lead to long-term arterial remodeling and increased vascular stiffness.

Moreover, endothelial dysfunction was observed in the aorta of the GH offspring, an important risk factor in the development of cardiovascular diseases [16,18]. In a rat model of prenatal hypoxia, maximal acetylcholine vasodilation was found to be reduced in fetal aortas [18]. Similarly, fetal guinea pigs exposed to FGR induced by uterine artery occlusion exhibited endothelial dysfunction in the aorta [16]. Moreover, the endothelial dysfunction exhibited in hypoxic animals was accompanied by a lower maximum response to SNP, which could reflect impaired downstream mechanisms, such as alterations in the sGC-cGMP-PKG pathway [86]. Interestingly, hypoxia did not alter the NO-dependent contribution to endothelium-dependent vasodilation, suggesting that the overall impairment reflects combined changes in NO-dependent and NO-independent pathways rather than a single mechanism. Thus, GH activates initially compensatory vascular adaptations that ultimately promote endothelial dysfunction and remodeling, establishing key risk factors for long-term cardiovascular disease [5].

Melatonin treatment abolished GH-induced aortic blood flow velocity during gestation, reduced oxidative stress, and restored endothelial function in newborn guinea pigs. In a late-gestation fetal sheep model, melatonin modulated the fetal cardiometabolic response to acute hypoxia by increasing NO bioavailability, suppressing blood flow redistribution [14]. Importantly, these effects were NO-dependent, as NO blockade reversed them [14]. In addition, in a chick embryo model of hypoxic incubation, melatonin treatment restored endothelial function, reduced oxidative stress, and increased nitric oxide bioavailability [87]. Similarly, other studies have suggested that perinatal melatonin abolishes oxidative stress in various gestational contexts [35,78,88,89,90]. In our study, melatonin treatment reversed GH-induced endothelial dysfunction, associated with an improvement in the response to SNP, as evidenced by both increased maximum response and enhanced sensitivity to this drug. Furthermore, a positive correlation was found between the endothelium-dependent vasodilatory response (MetCh) and the smooth muscle-dependent vasodilatory response (SNP), indicating a dependence on the smooth muscle response. Finally, we also observed that melatonin treatment increased the NO-independent contribution in the thoracic aorta. These results suggest that the improvement in endothelial function is related to changes downstream of eNOS in the redox-sensitive sGC-cGMP-PKG signaling pathway [86,91]. Additionally, a reduction in NO-independent vasodilation points to potential involvement of enzymes that synthesize other gasotransmitters, such as carbon monoxide and hydrogen sulfide, which are expressed in both the vascular endothelium and smooth muscle and require further study [92]. Therefore, our results suggest that melatonin treatment reverses GH-induced endothelial dysfunction, potentially through a reduction in oxidative stress and an increase in NO-independent pathways.

4.2. Prenatal Melatonin Effects in the Non-Pathological Context

Most importantly, melatonin treatment induced substantial alterations in cardiovascular structure and redox homeostasis during physiological gestation. It increased cardiac length and modified LV function in utero, and at birth reduced LV myocardial nuclei density while altering aortic structure and contractile capacity. These findings suggest that melatonin exogenous supplementation during gestation may interfere with normal cardiac development [93,94]. In fact, melatonin treatment has been reported to alter cardiac plasticity, increasing both proliferative capacity and fetal cardiac protein markers during embryonic development. This has been shown in neonatal mouse or in vitro models, regulated by melatonin-sensitive pathways such as HIF-1α or miR-143-3p [94,95]. Multiple studies have evaluated and observed the beneficial effects of prenatal melatonin [14,35,87,89,90,96]. However, few studies have pointed out some deleterious effects. One report evaluated the effect of prenatal melatonin in a model of spontaneous hypertension, at a dose of 10 mg/kg/day through drinking water, and the control group exhibited high mortality in the offspring of the treated control group compared to the pathological group, reaching 95% mortality in 6-week-olds [97]. On the other hand, we observed in a sheep GH model that prenatal melatonin, at a dose of 10 mg/kg/day, reduced birth weight, which was associated with increased cortisol levels [98]. These results suggest deleterious effects of prenatal melatonin supplementation under chronic hypoxia, potentially leading to alterations at the cardiac or systemic level, for example, through dysregulation of the central or endogenous circadian rhythms within tissues [90,99]. Importantly, despite this study using a substantially lower dose (10-fold lower than 10 mg/kg/day), we still observed potentially adverse cardiac remodeling under normoxic conditions, underscoring the sensitivity of fetal cardiovascular development to exogenous melatonin and the critical importance of careful dose selection in translational research for both early and long-term outcomes. Finally, melatonin treatment has been reported to induce the nuclear translocation of NRF2, thereby increasing the expression of antioxidant genes [32,33]. In our study, melatonin treatment increased the nuclear localization of NRF2, and this effect was associated with the induction of Gpx1 and Hmox1 transcript levels; however, this was not associated with changes in protein levels in the same direction. Our results highlight a transcriptional–translational decoupling within the NRF2 pathway under normoxic conditions. Because mRNA is an unreliable predictor of protein abundance, this finding suggests that post-transcriptional control mechanisms limit the final expression of antioxidant proteins [100]. Therefore, we hypothesize that even under normoxic conditions, where ROS levels are low and serve a physiological role, melatonin may influence cellular redox balance (e.g., by scavenging ROS) and/or exert its effects through receptor-mediated gene induction [31].

4.3. Limitations

This study aimed to exploratorily characterize the cardiovascular effects of antenatal melatonin treatment during the immediate postnatal period in the context of gestational hypoxia, focusing on the morphological and functional alterations induced by GH and the potential therapeutic effects of melatonin through its reported antioxidant mechanisms [31]. Consequently, one limitation of this work is that we did not investigate tissue-specific molecular pathways in each cardiovascular structure; instead, we assessed their overall status. In addition, another limitation was the inability to evaluate sex dimorphism, despite evidence suggesting that sex-specific differences may emerge as early as the fetal stage [28,72]. Furthermore, evaluating the heart at the immediate postnatal stage is a limitation, as cardiomyocytes are within a critical perinatal window of transition. During this period, which can extend up to two weeks, the heart undergoes profound maturation involving cell cycle exit, binucleation, and expression shifts in contractile proteins and ion channels [101,102]. This period of high developmental plasticity may have masked significant structural or functional alterations that would only manifest once a stable, mature phenotype is reached. All of these aspects should be further explored in future studies.

5. Conclusions

Our findings suggest that prenatal melatonin can mitigate some of the detrimental effects of gestational hypoxia, including endothelial dysfunction, a known risk factor for cardiovascular disease. These cardiovascular benefits appear to be linked to reduced oxidative stress and the activation of antioxidant pathways under conditions of prenatal hypoxia. Importantly, melatonin also exerts physiological effects through mechanisms not exclusively associated with hypoxic pathways, acting as a functional counterweight that compensates for cardiovascular deficits. However, the observation of unexpected effects in non-pathological pregnancies raises concerns, as melatonin exogenous supplementation may interfere with normal cardiovascular development and potentially promote adverse outcomes later in life in normoxic gestations. Given that this study provides exploratory insights into the effects of prenatal melatonin, it highlights the need for further investigation to evaluate the long-term impact of melatonin across different vascular territories.

Appendix A

Figure A1.

Experimental Design. Pregnant guinea pigs were subjected to normoxia (red) or hypoxia (blue) and treated with vehicle or melatonin (1 mg/kg/day) from gestational day (GD) 30. During gestation, the fetal cardiovascular function and structure was monitored. At birth, newborns were euthanized and both aortic and left ventricles were dissected and processed for ex vivo and in vitro assays.

Figure A2.

Maternal and newborn body weight. Maternal body weight (BW) at gestational days (GD) 0, 30, and 70 (A), maternal BW gain during gestation (B), and neonatal birth weight (C) values from normoxia (N; red) or hypoxia (H; blue) guinea pig pregnancies treated with vehicle (dots) or melatonin (M; squares). The dotted line indicates the GD at which both oxygenic and antioxidant treatments were initiated. Values are expressed as mean ± SEM. n = 5–6 (dams/pups) per group. Data were analyzed using an F-test and two-way ANOVA with FDR post hoc test. Significant differences (q ≤ 0.05): * vs. normoxic counterpart.

Figure A3.

Melatonin alters oxidative balance during both normoxic and hypoxic gestations. Representative photographs (A) and protein levels of catalase (CAT; (B)), glutathione peroxidase 1/2 (GPX1/2; (C)), superoxide dismutase-1 (SOD1; (D)), superoxide dismutase-2 (SOD2; (E)) and hemeoxigenase-1 (HO-1; (F)) in left ventricle from newborn guinea pigs exposed to gestational normoxia (N; red) or hypoxia (H; blue) and treated with vehicle or melatonin during gestation. In panel (A), the separated band corresponds to the loading control (LC), which was used to normalize all blots based on densitometry. Values are expressed as mean ± SEM. n = 5–6 independent samples (one per pup) per group. Data were analyzed using two-way ANOVA with FDR post hoc test. Significant differences (q ≤ 0.05): * vs. normoxic-counterpart, ^# vs. vehicle-counterpart.

Table A1.

List of primers.

mRNA Target	Gene Symbol	Forward Primer (5′ ► 3′)	Reverse Primer (5′ ► 3′)	Product Length (bp)	NCBI RefSeq
Catalase	Cat	GACAAAATGCTTCAGGGCCG	ACCTTGGTTGTCAGTCACGC	156	NM_001439566.1
Glutathione Peroxidase 1	Gpx1	TCATTGAGAATGTGGCCTCCC	GGACGTACTTGAGCGAATGC	177	XM_003476448.5
Glutathione Peroxidase 2	Gpx2	ACAGCCGCACCTTTCATACC	GCAAGGCTATTGGGTGAACG	163	XM_005004774.4
Glutathione Peroxidase 3	Gpx3	TGTGAGCGGAACCATCTACG	TCTCCCGGCTCTTGTTTTCC	228	XM_003464498.5
Glutathione Peroxidase 4	Gpx4	CTCCATGCACGAGTTCTCCG	CAGACCACACTCAGCGTACC	169	NM_001256319.1
Superoxide Dismutase 1	Sod1	CCGTTGTGGTAAAGGGACGC	AGTCCTCGATGGATACATTGGC	222	XM_003467248.5
Superoxide Dismutase 2	Sod2	CCTCCCCGATTTACCCTACG	CCGTTGAACTTCAGTGCAGG	195	XM_003466367.5
Heme Oxigenase 1	Hmox1	GCTGGTGATGGCCTCACTGTACC	CGTACCAGAAGGCCATGTCCTGC	149	XM_023561275.2
β-Actin	Actb	ATGGGCCAGAAGGACTCCTACG	TCAGGGGCCACACGCAATTC	158	NM_001172909.1

Table A2.

List of antibodies.

Protein Target	Abbreviation	Dilution Factor	Product Reference	NCBI GeneID
Catalase	CAT	1:5000	#ab1877, Abcam, Cambridge, UK	100135492
Glutathione Peroxidase 1/2	GPX1/2	1:2000	#sc-133160, Santa Cruz Biotechnology, Dallas, USA	100729115; 100714682
Superoxide Dismutase 1	SOD1	1:2000	#sc-17767, Santa Cruz Biotechnology, Dallas, USA	100135622
Superoxide Dismutase 2	SOD2	1:2000	#06-984, Millipore, Burlington, USA	100135623
Heme Oxigenase 1	HMOX1/HO-1	1:2000	#MA-1-112, Invitrogen, Carlsbad, USA	100715748
αβ-Tubulin	TUBA	1:2000	#2148, Cell Signaling, Danvers, USA	100716481; 100718713

Author Contributions

Conceptualization, A.A.P. and E.A.H.; Methodology, A.A.P., T.A.J., and E.A.H.; Validation, A.A.P., T.A.J., and E.A.H.; Formal Analysis, A.A.P., T.A.J., and E.A.H.; Investigation, A.A.P., P.H., J.C., D.C., J.I.-G., T.A.J., J.N.P., M.S., R.J., F.A.B., and E.G.F.; Resources, A.G.-C. and E.A.H.; Data Curation, A.A.P., T.A.J., and E.A.H.; Writing—Original Draft Preparation, A.A.P. and E.A.H.; Writing—Review & Editing, A.A.P. and E.A.H.; Visualization, A.A.P. and E.A.H.; Supervision, E.A.H.; Project Administration, E.A.H.; Funding Acquisition, A.G.-C. and E.A.H. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

The animal study protocol was approved by the Institutional Review Board of theComité Institucional de Cuidado y Uso de Animales (CICUA) of Universidad de Chile (certificate 20354-MED-UCH; Approved 19 December 2023).

Informed Consent Statement

Not applicable.

Data Availability Statement

The raw data and image files supporting this study are openly available in Zenodo at https://doi.org/10.5281/zenodo.17487466.

Conflicts of Interest

The authors declare no conflicts of interest.

Funding Statement

This manuscript was funded by ANID grant No. 21230893, FONDECYT de Inicio grant N° 11200798, and FONDECYT Regular grant N° 1201283 & 1241502.