All experimental protocols were performed in accordance with the recommendations of the Brazilian Guidelines for the Care and Use of animals for Scientific and Teaching Purposes and the Guidelines for the Practice of Euthanasia (CONCEA-MCT, 2013), having been approved by the Ethics Committee on Animal Use of the University of São Paulo, Ribeirão Preto, Brazil (Protocol n° 206/2016).

Four-week-old male C57Bl/6J mice were obtained and maintained in the Animal Facility of the University of São Paulo, Ribeirão Preto, Brazil on 12-h light/dark cycles under controlled temperature (22 ± 1°C) and humidity (50–60%) with ad libitum access to food and water. Mice were anesthetized (ketamine and xylazine at doses of 75 and 15 mg/kg, respectively, intraperitoneally), the efferent duct of each testicle was ligated, and the testicles were removed. Sham mice were submitted to the same surgical procedures without efferent duct ligation or testicles removal. After a 1-week acclimatization period, mice were divided into two groups: (1) mice maintained on a control diet [CD (protein 22%, carbohydrate 70%, and fat 8% of energy, PragSolucoes)], and (2) mice receiving a high-fat diet [HFD (protein 10%, carbohydrate 25%, and fat 65% of energy, PragSolucoes)]. After 18 weeks on the CD or HFD, mice were euthanized by carbon dioxide (CO₂) inhalation without removing mice from their home cage. Death was confirmed by cardiac arrest. Studies followed the ARRIVE guidelines for reporting experiments on experimental animals.

Glucose levels were determined using a glucose analyzer (Accu-Check, Roche Diagnostics). Total cholesterol and triglyceride concentrations were determined in serum samples from mice fasted for 12 h (da Costa et al., 2018), by enzymatic colorimetric method (Doles^®). Plasma insulin concentration (ng/mL) was determined by radioimmunoassay (Insulin Kit^®).

Plasma testosterone levels were determined in sham, castrated and testosterone-treated mice using the IMMULITE 1000 Immunoassay System (Enzo Life Sciences^®, New York, NY, United States). After 18 weeks on the CD or HFD, testosterone-treated mice received intramuscular injections of testosterone-propionate [dissolved in peanut oil, 5 mg/kg body weight per day (Sigma Aldrich, San Luis, MO, United States)], for 15 days.

Thoracic aorta was rapidly removed, transferred to an ice-cold (4°C) Krebs Henseleit-modified solution [(in M) 130 NaCl, 14.9 NaHCO₃, 4.7 KCl, 1.18 KH₂PO₄, 1.17 MgSO₄⋅7H₂O, 5.5 glucose, 1.56 CaCl₂⋅2H₂O and 0.026 EDTA] gassed with 5% CO₂ and 95% O₂ to maintain a pH of 7.4, and dissected into 2 mm rings. Aortic rings were mounted in a wire myograph to measure isometric tension, as previously described (da Costa et al., 2017). Vessels were allowed to equilibrate for about 30 min in Krebs Henseleit solution. After the stabilization period, endothelial function was assessed by testing the relaxant effect of acetylcholine (ACh, 10^–6 M) on vessels contracted with phenylephrine (PE, 10^–7 M). Aortic rings exhibiting a vasodilator response to ACh greater than 70–80% were considered endothelium-intact vessels. Cumulative concentration-response curves to ACh (10^–10 to 10^–4 M) were performed in aortic rings from the various experimental groups.

Relaxation responses to ACh were also determined after incubation with either the selective superoxide anion (O₂^–) scavenger tiron (10^–4 M, 30 min) or the Nrf2 activator bardoxolone (10^–6 M, 3 h). Each vascular preparation was tested with a single agent.

Endothelial cells were commercially purchased. An immortalized line of human umbilical vein endothelial cells, EA.hy926, (ATCC^® CRL-2922™) was used. Cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS) in a CO₂ incubator, at 37°C, 5% CO₂. After confluence, cells were maintained in the presence of 2% FBS (Rodrigues et al., 2021). Cells were stimulated with testosterone for various times (10^–7 M, 15 min to 6 h). To determined testosterone effects on NADPH oxidase activity, cells were preincubated with a NOX1 inhibitor ML171 (10^–4 M, 1 h). To investigate genomic and non-genomic effects of testosterone, cells were incubated with the androgen receptor antagonist flutamide (10^–5 M, 1 h) and the gene transcription inhibitor actinomycin D (10^–5 M, 1 h), respectively, previous to exposure to testosterone.

Thiobarbituric acid reactive substances (TBARS) concentration was determined colorimetrically (at 540 nm) in aortas homogenates with a commercially available kit (#10009055, Cayman Chemical, United States), following the manufacturer’s instructions. TBARS concentration was determined using a standard curve of malondialdehyde and the results are expressed as μM/mg protein.

Plasma nitrate concentrations were determined by subtracting the nitrite values from the NOx values (nitrite + nitrate). For the determination of NOx, the Griess reaction was used. Plasma samples were incubated with nitrate reductase, at 37°C for 12 h, in a light-free environment. The Griess reagent was added and reading was performed by spectrophotometry at 540 nm.

Thioredoxin reductase (TrxR) activity was determined in aortas using a commercially available kit (#CS0170, Sigma-Aldrich, San Luis, MO, United States). Enzymatic activity was assessed by determining the difference between the time-dependent increase in absorbance at 412 nm in the presence of the TrxR activity inhibitor from total activity. One activity unit equaled 1 mM 5′-thionitrobenzoic acid formed/mg protein. Thoracic aortae were homogenized in 300 μL of Krebs Henseleit and centrifuged at 18,000 × g for 15 min at 4°C. The supernatants were analyzed for superoxide dismutase (SOD) or catalase (CAT) activities. SOD activity was evaluated using a commercially available kit (#19160, Sigma-Aldrich, San Luis, MO, United States). The results were normalized for protein concentration, and SOD activity is expressed as % inhibition rate/mg protein. CAT activity was assayed by H₂O₂ consumption and measured in a spectrophotometer at 240 nm. One CAT unit (U) was defined as the amount of enzyme required to decompose 1 μM of H₂O₂/min/mg protein.

To determine nuclear accumulation of Nrf2, the nuclear fraction from cell or tissue lysates were separated using the Active Motif nuclear extract kit (Active Motif, Carlsbad, CA, United States) following the manufacturer’s protocol. Cells or tissue were resuspended in 1X hypotonic buffer and centrifuged for 30 s at 12,800 × g at 4°C. Nuclear pellets were resuspended in lysis buffer provided by the manufacturer. The suspension was incubated for 30 min on ice on a rocking platform set at 5 × g and then centrifuged for 10 min at 12,800 × g. The supernatant was transferred to a pre-chilled microcentrifuge tube. TransAM™ Nrf2 ELISA kit (Active Motif) was used to measure nuclear accumulation of Nrf2 at a wavelength of 450 nm.

Total RNA was isolated from aortae using Trizol^® (Invitrogen, Carlsbad, CA, United States). RNA was treated with DNAse I (1 U/μL, Promega) and used for first-strand cDNA synthesis, accordingly to the manufacturer instructions. mRNA levels were quantified in triplicate by qPCR StepOnePlus™ Life Technologies. Specific primers (TaqMan™) for RT-qPCR were as follows: mouse SOD-1 [Mm01344233_g1], Catalase [Mm00437992_m1], Heme oxygenase-1 [Mm00516005_m1], Peroxiredoxin-1 [Mm01621996_s1] and β-actin [Mm00607939_s1], purchased from Life Technologies. PCR cycling conditions included 10 min at 95°C, followed by 40 cycles at 95°C for 15 s, 60°C for 1 min, and 72°C for 60 s. Dissociation curve analysis confirmed that signals corresponded to unique amplicons. Specific mRNA expression levels were normalized relatively to β-actin mRNA levels using the comparative ΔΔCt method.

Aortas were frozen in liquid nitrogen and homogenized in a lysis buffer [50 mM Tris/HCl, 150 mM NaCl, 1% Non-idet P40, 1 mM ethylenediaminetetraacetic acid (EDTA), 1 μg/mL leupeptin, 1 μg/mL pepstatin, 1 μg/mL aprotinin, 1 mM sodium orthovanadate, 1 mM phenylmethylsulphonyl fluoride (PMSF), and 1 mM sodium fluoride]. The tissue extracts were centrifuged, and total protein content was quantified using the Bradford method (Bradford, 1976). Proteins (40 μg) were separated by electrophoresis on 10% polyacrylamide gel, and transferred on to nitrocellulose membranes. Non-specific binding sites were blocked with 5% bovine serum albumin (BSA) in Tris-buffered saline (TBS) containing 0.1% Tween 20 (for 1 h at 24°C). Membranes were incubated with antibodies (at the indicated dilutions) overnight at 4°C. Antibodies were used as follows: anti-keap1 (1:1,000 dilution; Abcam), anti-NOX1 (1:1,000 dilution; Cell Signaling), anti-NOX4 (1:1,000 dilution; Cell Signaling), anti-βactin (1:1,0,000 dilution; Sigma), and anti-GAPDH (1:10,000 dilution; Sigma). After incubation with secondary antibodies, signals were obtained by chemiluminescence, visualized by autoradiography and quantified densitometrically.

The individual concentration–response curves were fitted into a curve by non-linear regression analysis. pD₂ (defined as the negative logarithm of the EC₅₀ values) and maximal response (Emax) were compared by Two-way ANOVA with Bonferroni post-test. The results of the molecular experiments were analyzed by Mann Whitney test or Two-way ANOVA, followed by the Bonferroni post-test. Data were assessed for normality with Shapiro-Wilk test. The Prism software, version 9.0 (GraphPad Software Inc., San Diego, CA, United States) was used to analyze these parameters as well as to fit the sigmoidal curves. Data are presented as mean ± SEM. N represents the number of mice used. p values less than 0.05 were considered significant.

Table 1 details the characteristics of the animal groups. At the beginning of the experimental protocol, sham and castrated mice displayed similar body weight. Both groups gained weight during the protocol development. However, the weight gain was higher in mice fed the HFD as compared to the CD-fed group. Castrated and Sham mice on the HFD showed similar body weight gain. Similarly, testosterone treatment did not modify weight gain in HFD- or CD-fed mice. Castration effectively reduced endogenous androgen levels, as evidenced by the undetectable levels of testosterone in castrated animals. Castrated mice fed the CD and treated with testosterone propionate (5 mg/kg body weight) exhibited supraphysiological levels of testosterone, whereas castrated HFD-fed mice treated with the same dose of testosterone propionate (5 mg/kg body weight) exhibited testosterone levels similar to those observed in the sham group.

No differences in serum glucose were observed between sham and castrated mice fed the CD. However, HFD increased serum glucose in sham mice. Insulin and total cholesterol plasma levels were also increased in HFD-fed sham mice compared to CD-fed sham mice. Castration did not change increased serum glucose, insulin or total cholesterol levels induced by the HFD, but decreased triglyceride plasma levels in mice fed both the CD and HFD (Table 1). These results indicate that metabolic changes evoked by the HFD in young mice are not affected by either castration or supraphysiological levels of circulating testosterone.

We then assessed the involvement of testosterone on vascular dysfunction in HFD-fed mice considering its independent association with cardiovascular complications. Castrated and sham-operated mice were fed the HFD and vascular function was evaluated. As shown in Figure 1A, no differences in SNP-induced vascular relaxation were observed between CD sham and HFD sham mice. In contrast, HFD reduced ACh-induced vascular relaxation in sham mice in comparison with CD-fed sham mice (Figure 1B). In mice fed the CD, there were no differences in ACh-induced vasodilatation between the sham and castrated groups. However, castration prevented HFD-induced vascular dysfunction (Figure 1C). Treatment of castrated mice maintained on the CD with testosterone propionate decreased ACh-induced vasodilation. The effects of castration on ACh vasodilation in HFD-fed mice, i.e., increased ACh vasodilatation, were lost upon testosterone propionate treatment (Figure 1D). Table 2 shows the maximum response and pD₂ values.

We next determined whether testosterone contributes to increased vascular ROS generation in mice on the HFD. As expected and already reported, HFD decreased plasma NOx (Figure 2A) and increased aortic TBARS levels (Figure 2B). Basal levels of ROS, determined by DHE fluorescence intensity (Figure 2C), were significantly increased in arteries from HFD-fed sham mice. In addition, aortas from HFD sham mice exhibited increased generation of O₂^– (Figure 2D) and H₂O₂ (Figure 2E), determined by lucigenin chemiluminescence and amplex red, respectively. Castration prevented increased vascular ROS induced by HFD. No differences were observed in ROS generation between sham and castrated mice on the CD. Similarly to castration, tiron improved ACh-induced vascular relaxation in HFD sham mice (Figure 2F). No additional effects were observed with the combination of tiron and castration, i.e., tiron did not further improve ACh vasodilation in HFD castrated mice (Figure 2G), indicating that both tiron and castration interfere with similar pathways. Table 3 shows the maximum response and pD₂ values.

Nuclear factor E2–related factor 2 signaling is a major regulator of cellular antioxidant systems. HFD increased vascular Keap-1 protein expression (Figure 3A). In addition, HFD decreased nuclear accumulation of Nrf2 (Figure 3B), which was associated with a decrease in mRNA expression of Nrf2-regulated enzymes, such as SOD-1 (Figure 3C), catalase (Figure 3D), heme oxygenase-1 (Figure 3E), peroxiredoxin (Figure 3F) as well as with decreased activity of the enzymes thioredoxin (Figure 3G), SOD (Figure 3H) and catalase (Figure 3I). These events were also prevented in castrated HFD-fed mice, indicating that Nrf2-regulated redox balance in the vasculature of obese mice is modulated by testosterone. No differences were observed in Nrf2 antioxidant system between sham and castrated mice on the CD.

Activation of the Nrf2 system by bardoxolone improved ACh-induced vascular relaxation in HFD-fed sham mice (Figure 4A). The effects of bardoxolone and castration on ACh vasodilation were similar in HFD-fed mice (Figure 4B), with no additional effects being observed in arteries from castrated mice incubated with bardoxolone. Bardoxolone did not change vascular reactivity in sham or castrated mice fed the CD. Bardoxolone increased nuclear accumulation of Nrf2 in vessels from HFD-fed sham mice (Figure 4C) and decreased HFD-induced vascular ROS generation (Figure 4D). Notably, no synergism was observed between the effects of bardoxolone and castration. In addition, bardoxolone improved catalase and heme oxygenase-1 mRNA expression (Figure 4E). Table 4 shows the maximum response and pD₂ values.

NADPH oxidase is the main source of free radicals in the vasculature. HFD increased NOX1 expression in aortas, which was prevented in castrated HFD-fed mice (Figure 5A). No difference was observed in the antioxidant system between sham and castrated mice fed the CD. In addition, no difference was observed in NOX4 (Figure 5B) expression among the different experimental groups.

Considering that testosterone modulates endothelial cell function as well as oxidative and inflammatory processes in endothelial cells, we further determined molecular mechanisms involved in testosterone-mediated endothelial dysfunction, by using EA.hy926 cells. Testosterone induced a time-dependent ROS generation in EA.hy926 cells (Figure 6A), with maximum generation occurring at 2 h. In addition, testosterone increased nuclear accumulation of Nrf2 from 30 min, but this increase was reduced after 3 h of stimulation with testosterone (Figure 6B). Incubation of EA.hy926 cells with bardoxolone prevented ROS generation (Figure 6C). Testosterone-induced ROS generation at 3 and 6 h was prevented by ML171, a NOX1 inhibitor (Figure 6D). To investigate the possible mechanisms (genomic and non-genomic) whereby testosterone induces ROS production at 3 and 6 h, ROS generation was determined in the presence of flutamide and actinomycin D. Pre-incubation with flutamide prevented testosterone-induced ROS generation at 3 h, but not at 6 h. On the other hand, incubation with actinomycin D prevented testosterone-induced ROS generation at 6 h, but not at 3 h (Figure 6E). Actinomycin D prevented the decrease in Nrf2 activity induced by testosterone. Testosterone-induced decrease in Nrf2 activity was not observed in the presence of flutamide (Figure 6F).