Adult male Wistar Kyoto rats (n = 12, 276±5 g) were randomly divided into two groups: the Vehicle group (Veh, n = 6) and the OXY group (n = 6). The Veh group received NaCl 0.9% solution (200 µL, administered intraperitoneally). In contrast, the OXY group received OXY (1 mL, 5 U.I. in NaCl 0.9%, Laboratorio Sanderson, Ñuñoa, Chile) (0.433 µg/kg, 200 µL, administered intraperitoneally) once daily, for 14 days. Both animal groups increased their body weight after 14 days, which was not associated with OXY administration (Veh: 281 ± 8 vs. 298 ± 8 g, pre vs. post, p = 0.003) (OXY: 264 ± 8 vs. 278 ± 12 g, pre vs. post, p = 0.006). The animals were sourced from the animal facility at the Universidad de Antofagasta and were housed under controlled temperature and humidity conditions with a standard 12-hour light/dark cycle. The rats had unrestricted access to water and a standard diet (Prolab RMH3000; LabDiet, USA).

Animal welfare guidelines for this study were established by the American Physiological Society. The protocols were approved by the Ethics Committee on Scientific Research at the Universidad de Antofagasta (CEIC-UA 438/2022). At the end of the experiments, all animals were humanely euthanized with an overdose of sodium pentobarbital (100 mg/kg, i.p.).

The experimental design is illustrated in Fig. 1. Before (pre-, 2 days prior to beginning treatments) and after (post-, at days 13 and 14) the Veh and OXY administration period, the exercise performance was assessed using an incremental exercise test under both NN and HH conditions (Fig. 5A). Afterward, the animals were subjected to physiological monitoring (heart rate variability (HRV), baroreflex, and hypobaric hypoxic ventilatory response (HHVR)) at the end of the experiment, similar to those previously described [4, 25]. At the endpoint (14 days after Veh or OXY administration), the rats were anesthetized using 40 mg/kg and 1 g/kg α-chloralose and urethane (Sigma-Aldrich), respectively. Once deeply anesthetized, they were positioned supine, maintaining their body temperature at 38.0 ± 0.5 °C through a controlled warming mat (Kent Scientific model RT-0515). A flexible tube was inserted into the trachea to monitor airflow and was connected to a pneumotachograph to study the HHVR. Subsequently, a catheter was placed in the jugular vein (PE50 polyethylene tubing containing a saline solution) for drug administration, and another catheter (PE-10 connected to PE-50 tubing filled with a heparin saline solution, concentration: 5 I.U.) was placed in the left femoral artery to measure blood pressure (BP) (PowerLab/4SP, ADInstruments, Castle Hill, NSW, Australia).

Heart rate variability (HRV) indirectly indicates the autonomic balance in the heart [10, 25, 26]. The R-R time series was calculated from dP/dt from a 10-minute BP recording. By applying an autoregressive algorithm following Hann windowing with a 50% overlap, the power spectral density of the HRV was derived. Low-frequency HRV (LFHRV), range: 0.04–0.6 Hz; and high-frequency HRV (HFHRV), range: 0.6–2.4 Hz [25]. In addition, the LFHRV-to-HFHRV ratio (LF/HFHRV) was employed as a global autonomic index. LFHRV and HFHRV are reported in normalized units (n.u.). We used spectral non-stationary analysis with a 2-second resolution to evaluate the short-term variability. This analysis was performed with Kubios HRV Premium Software V 3.1 (Kubios, Finland).

The BR sensitivity was assessed using serial doses (50 µL) of sodium nitroprusside (SNP, 0.4, 0.8, 1.6, 3.2, 6.4, 12.8, and 25.6 µg/kg; Sigma-Aldrich, United States) and phenylephrine (PHE, 0.2, 0.4, 0.8, 1.6, 3.2, 6.4, and 12.8 µg/kg; Sigma-Aldrich, United States). These drugs were administered to facilitate a decrease (SNP) or increase (PHE) in BP, with subsequent HR changes. SNP and PHE injections were administered in a dose-dependent manner, from lowest to highest. The cardiac BR function was analyzed using logistic regression across the entire pressure range [4, 27,28,29]. The data were adjusted to the equation: FC = A / [1 + exp (B (SBP-C))] + D, where A is the range of HR, B is the slope coefficient, C is the pressure at the midpoint of the range (BP50), and D is the minimum HR. The maximum slope (maximum gain) was determined by the first derivative of the baroreflex curve and calculated using the equation gain = A1 × A2 × [1/4], where A1 is the range and A2 is the average slope. Further HR at the midpoint of the range (HR50) was calculated. All analyses were performed using the GraphPad Prism software (version 10.0; La Jolla, CA, USA) and Excel.

The BP assessment was performed in the left femoral artery. The BP signal was recorded using an analog-digital system (PowerLab/4SP, AD Instruments, Castle Hill, NSW, Australia) and analyzed using LabChart 7.0 software (AD Instruments, Castle Hill, NSW, Australia). From the systolic blood pressure (SBP) and diastolic BP (DBP) raw signals, the pulse pressure (PP = SAP-DAP) and mean arterial blood pressure (MABP = 1/3 of SAP + 2/3 of DAP) were calculated. Heart rate (HR) was calculated from the first derivative of BP (dP/dt) [25, 26].

Initially, the animals were maintained under normobaric normoxia (NN: O2 partial pressure [PO2], 155 mmHg), and the baseline breathing was assessed during 10 min recording. Subsequently, the rats were subjected to HH (PO2: 100 mmHg; speed decay: 1.57 mmHg/min) environmental conditions in a hypobaric hypoxic chamber (Genetic Core, Peru) simulating 3,500 m (speed ascend: 100 m/min). The breathing at HH was assessed during a 10-minute recording. To determine HHVR, the animals were anesthetized inside the hypobaric hypoxic chamber. After deep anesthesia, a catheter was placed in the trachea and connected to a spirometer pod (PowerLab/4SP, ADInstruments, Castle Hill, NSW, Australia), which was previously calibrated. From ventilatory flow, the minute ventilation (tidal volume [V̇T] ⋅ respiratory frequency [Rf] = V̇E) was calculated [30]. HHVR was expressed as a percentage difference between hypoxia and normoxia (% of normoxia). The breathing signal was analyzed using LabChart 7.0 software (AD Instruments, Castle Hill, NSW, Australia).

Before and after Veh or OXY administration, a progressive exercise test was used to assess the exercise capacity of all animals, as previously described [31]. Before the exercise test, the animals were familiarized with the treadmill (LE 8700 TS model, Panlab Harvard Apparatus, Spain) for 3 min, and after that, started to run at 16 cm/s without inclination. This acclimatization period was for one week. After the familiarization period, the rats began running at 41.7 cm/s with a 2º incline; the treadmill speed was increased by 5 cm/s every minute with a constant inclination until exhaustion. A mild electrical grid (0.5 mA) was used to motivate exercise at the back of the treadmill. The test concluded when the rats displayed signs of exhaustion, defined as remaining on the electrical grid for 10 s without being able to continue or keep up with the treadmill’s pace (Fig. 5). The work done until exhaustion was expressed in Joules (J).

Data are expressed as mean ± standard error of the mean (S.E.M.) in the results section, and median ± min-max in the figures, except for exercise performance, which is shown as mean ± standard deviation (S.D.). The normality of the data was assessed using the Shapiro-Wilk test, and Levene’s test determined the homoscedasticity of the variance. Differences between groups were evaluated using repeated measures ANOVA (2 × 2) followed by Holm-Sidak posthoc. The cardiorespiratory responses to HH were analyzed with a non-parametric Mann-Whitney test. In addition, the statistical power (1-β) (SP) was calculated for every significant comparison (G*Power, Germany). Differences were considered statistically significant at P < 0.05. All analyses were performed using the GraphPad Prism software (version 10.0, La Jolla, CA, USA).