Male Wistar rats (250–270 g) were obtained from the Animal Care Facility of Federal University of Sergipe and maintained under a controlled 12-h light/12-h dark cycle at room temperature (23 ± 2°C). All experimental procedures were previously approved by the Ethics Committee for Animal Research of the Federal University of Sergipe, Brazil (Protocol #36/10).

Animals were randomly divided into four experimental groups and the treatments were performed for 8 days: CTR, received saline (i.p. daily); ISO, received isoproterenol (4.5 mg/kg/day, i.p.); GBE, received G. biloba extract (100 mg/kg/day, v.o.) plus saline (i.p.); ISO + GBE, received isoproterenol (4.5 mg/kg/day, i.p.) plus G. biloba extract (100 mg/kg/day, v.o.). Isoproterenol was used as a non-specific β-adrenergic receptor (β-AR) agonist to induce cardiac hypertrophy, as previously described (Gavioli et al., 2014). The GBE used in this study was a standardized extract obtained from leaves, in which containing 25.5% Ginkgo flavonoids, 24.14% quercetin, 0.86% kaempferol, 0.55% isorhamnetin, and 6% terpenoids (ginkgolides and bilobalide). The GBE was provided by Fármacos (Sergipe) from Ningbo Traditional Chinese Pharmaceutical (China).

The animals were anesthetized with thiopental sodium (50 mg/kg, i.p.) and a polyethylene catheter was implanted into the femoral artery. The catheter was tunneled into the back of the rats and exteriorized in the nape. After 24 h, the catheter was connected to a pressure transducer (FE221, Bridge Amp, ADInstruments, Bella Vista, NSW, Australia) coupled to a pre-amplifier (Powerlab 8/35, ADInstruments). Blood pressure and heart rate (HR) were recorded for 30 min and processed using a dedicated software (LabChart 7 Pro, ADInstruments).

Cardiac autonomic balance was evaluated by frequency domain (Macedo et al., 2016). To perform this analysis, the CardioSeries v2.4 software was used¹. First, the beat-by-beat series obtained from pulsate arterial pressure recordings and HR were converted to data points every 100 ms using cubic spline interpolation (10 Hz). The interpolated series were divided into half-overlapping sequential sets of 512 data points (51.2 s). Before calculation of the spectral power density, the segments were visually inspected and the non-stationary data were not taken into consideration. The spectrum was calculated using the Fast Fourier Transformation algorithm and Hanning window was used to attenuate side effects. The spectrum is composed by bands of low frequency (LF; 0.2–0.75 Hz) and high frequency (HF; 0.75–3 Hz), being the results expressed as normalized units (nu), by calculating the percentage of the LF and HF variability with respect to the total power after subtracting the power of the very LF component (VLF frequencies < 0.20 Hz), namely LF/HF ratio. HF indicates the cardiac parasympathetic activity, while LF is an index of cardiac sympathetic activity and LF/HF ratio represents the sympatho-vagal balance to the heart.

The baroreflex sensitivity (BRS) was measured in the time domain by the sequence method (Bertinieri et al., 1985). Sequences of at least four heart beats with increased SAP followed by pulse interval lengthening or subsequent decrease of SAP with pulse interval shortening with correlation greater than 0.85 were identified as baroreflex sequence. The slope of the linear regression between SAP and pulse interval was considered as a measure of BRS (mmHg/s).

The rats were anesthetized and kept in the supine position with spontaneous breathing. For surface ECGs recording, three stainless steel electrodes were subcutaneously implanted and ECG signals were amplified (HP7754A, HP7754B, Hewlett-Packard, Chicago, IL, USA), digitalized (DI-710, Windaq Pro, Dataq, Akron, OH, USA) and stored in a computer for off-line processing. In all of the experimental groups, HR, corrected QT interval (QTc), QRS complex duration, and intrinsicoid deflection (ID) were measured in 10 consecutive beats. QT interval was corrected by HR using Bazett’s equation.

After 15 min of heparin administration (1,000 I.U., i.p.), the heart was quickly removed and carefully mounted in an aortic perfusion system (Langendorff technique) on a constant flow (8 mL/min) (Milan Peristaltic Pump, Paraná, Brazil). Then, the heart was perfused with Krebs solution (in mM: 118.0 NaCl, 4.7 KCl, 1.2 MgSO₄, 25.0 NaHCO₃, 1.8 CaCl₂, 11.1 glucose, 1.2 KH₂PO₄; pH was adjusted to 7.4) that had been previously filtered through a cellulose acetate membrane (0.45 μm), oxygenated (95% O₂ + 5% CO₂) and kept at 37 ± 0.1°C (Haake F3, Berlin, Germany). The left intraventricular pressure was measured using a water-filled balloon introduced into the cavity of the left ventricle. This device was coupled to a pressure transducer (HP 1290A, Hewlett-Packard, Chicago, IL, USA). Signals were amplified (HP7754A, HP7754B), digitized (DI-710, Windaq Pro, Dataq, Akron, OH, USA) and stored in a computer. The system was calibrated using a mercury column. Coronary pressure was measured at the tip of the aortic cannula and monitored in a water column.

Western blots were performed as previously described (Mota et al., 2015), with some modifications. Thirty to fifty microgram of protein were resolved on SDS–PAGE, transferred to a PVDF membrane, and incubated with the following primary antibodies: anti-eNOS (1:1000, sc-654), anti-nNOS (1:1000, sc-8309), M₂ (1:1000, sc-9107) anti-peNOS^ser1177 (1:1000, sc-12972), anti-pnNOS^ser852 (1:1000, sc-19826), anti-SERCA2 (1:2500, sc-376235), and anti-GAPDH (1:3000, sc-32233) from Santa Cruz Biotechnology or β₁-AR (1:1000, ab3442, Abcam). All of them incubated at 4°C overnight. After incubation with appropriate secondary peroxidase-coupled antibodies for 1 h, immunodetection was carried out using enhanced chemiluminescence (Amersham Biosciences) followed by densitometric analysis with software ImageJ. Protein levels were expressed as a ratio of optical densities. GAPDH was used as a control for any variations in protein loading.

After experimental procedures, rats were anesthetized and euthanized by applying potassium chloride solution (KCl – 10%) into the jugular vein. The hearts were fixed in formalin (10%), embedded in paraffin and cut at 5 μm thickness followed by staining with hematoxylin-eosin. Morphometric analysis was performed using the software ImageJ. The mean nuclear area was extracted from each histological slide.

Rat neonatal cardiomyocytes (3-days old) were cultured as previously described (Rocha-Resende et al., 2012). Briefly, cardiac cells were plated in dishes containing M199 medium supplemented with 100 units/mL penicillin, 100 μg/mL streptomycin, 10% fetal bovine serum, and 2 mM/L L-glutamine. To prevent growth of fibroblasts, medium was supplemented with 20 μg/mL cytosine-D-arabinofuranoside (ARA-c). After 48 h, neonatal cardiomyocytes were exposed to isoproterenol (10 μM) and/or GBE (100 μg/mL). When appropriated, cells were incubated with atropine (AT, 10 μM) or Nω-nitro-L-arginine methyl ester hydrochloride (L-NAME, 10 μM) for 48 h. Afterward, the cells were then used for immunofluorescence.

Neonatal cardiomyocytes were fixed in a 4% paraformaldehyde solution and permeabilized with 0.5% Triton-X100. After blocking, cardiomyocytes plated onto glass coverslips were incubated for 1 h at room temperature with Alexa Fluor 488-conjugated anti-phalloidin (1:100, A12379, Invitrogen). Nuclear staining was obtained by incubating with 4,6-diamidino-2-phenylindole (DAPI, 1:50). Surface area of cardiomyocytes was measured in phalloidin stained cells. Images were acquired with a Zeiss LSM 510 confocal system located at Center of Acquisition and Processing of Images (CAPI–ICB, UFMG). All images were representative of two independent experiments in which multiple cells were evaluated.

All data are expressed as mean ± SEM. Statistical comparisons were performed using GraphPad Prism 5.1 (San Diego, CA, USA). Normality and equality of variance were tested by Shapiro–Wilk and Levene test, respectively. Significant differences between groups were determined with one-way ANOVA followed by the Bonferroni post hoc test. Differences were considered to be statistically significant when p < 0.05.

First, we show that ISO-treated rats (4.5 mg/kg/day for 8 days) developed prominent cardiac hypertrophy, indicated by an increase in heart weight to body weight or tibia length ratios compared with untreated rats. Remarkably, concomitant GBE treatment (100 mg/kg) prevented the cardiac hypertrophy induced by chronic β-AR stimulation (Table 1). In vivo evaluation of cardiovascular function demonstrated a decreased HR in ISO-treated rats, which was not prevented after GBE treatment. Despite a tendency toward lower mean arterial pressure in ISO group, there was no significant change when compared to untreated group (Table 2).

It is well known that chronic β-AR stimulation induces profound alterations in autonomic nervous system (Shivkumar and Ardell, 2016), therefore we next evaluated whether GBE modulates the autonomic balance and BRS of hypertrophic hearts. As expected, chronic β-AR treatment led to a greater power in the LF band (Figure 1A) and smaller power in the HF band (Figure 1B) when compared to control, an indication of sympathovagal imbalance. Notably, GBE treatment fully restored sympathovagal balance in ISO-treated rats (LF/HF ratio, Figure 1C). Moreover, as shown in Figure 1D ISO-treated rats displayed decreased spontaneous BRS, which was rescued by GBE treatment. Altogether, our data indicate that GBE treatment modulates the sympathovagal balance of ISO-treated rats, suggesting the involvement of the cholinergic parasympathetic drive in the restoration of autonomic balance mediated by GBE.

To investigate whether GBE modulates expression levels of muscarinic receptor (M₂) and β₁-AR in the heart, we next performed the western-blot technique. As shown in the Figure 2, protein levels of M₂ were upregulated (Figure 2A), while the expression of β₁-AR was downregulated (Figure 2B) in left ventricles from ISO-treated rats when compared to control. On the other hand, GBE treatment alone led to a decreased expression of M₂ (Figure 2A), while the protein levels of β₁-AR was upregulated (Figure 2B). Importantly, GBE treatment was able to prevent alterations of M₂ and β₁-AR induced by ISO suggesting that under this condition appropriated autonomic balance was restored. Once again, this finding reinforces the idea that G. biloba-mediated cardioprotective actions involve the activation of cholinergic activity.

It is well known that augmented local cholinergic activity leads to an increase in NO levels (Rocha-Resende et al., 2012). Therefore, in order to better understand whether the synthesis of NO is involved in cardioprotective actions of GBE, we next evaluated protein expression and activity of the constitutive NO synthesis (eNOS and nNOS) in the heart. As shown in the Figure 3A, ISO-treated rats presented an upregulation of eNOS, while its activity was markedly reduced (Figure 3B). In contrast, GBE treatment of ISO-treated rats fully restored eNOS levels and activity when compared to control. Importantly, neither nNOS expression nor activity was changed by ISO or GBE treatment (Figures 3C,D). Altogether, our data show that GBE treatment restores impaired eNOS.

To confirm whether GBE treatment exhibits cardioprotection through impeding typical hypertrophic remodeling induced by chronic β-AR stimulation, morphometric analysis of nuclear cross-sectional area was evaluated. As shown in the Figure 4, left ventricular histological sections display increased nuclear area in ISO-treated rats, which was restored to the control condition with GBE treatment. GBE treatment alone had no effect on nuclear size. Moreover, ISO group showed areas that were characterized by intense spindle cell proliferation, identified as fibroblasts replacing the cardiac parenchyma (Figure 5). However, GBE treatment of ISO-treated rats exhibited markedly less expressive hypercellular areas, suggesting a minor replacement of cardiac parenchyma and therefore, a considerable decrease in the degree of cardiac remodeling. GBE-treated group demonstrated a typical cardiac parenchyma similar to untreated group, suggesting that the administration of GBE did not promote remarkable changes in the histological architecture of the heart.

Maladaptive cardiac remodeling is commonly associated with myocardial electric remodeling, therefore we next assessed whether GBE treatment prevents such electrical dysfunction through surface ECG recordings. Figure 6A shows typical ECG tracing in four experimental groups. It worthwhile note that depression of ST segment (negative T-wave) induced by chronic β-AR stimulation was abolished in GBE-treated rats. Moreover, rats that received ISO presented ECG changes that are typical of hypertrophic hearts such as, marked enlargement of QRS complex duration (Figure 6B), prolonged QTc interval (Figure 6C), and increased ID (Figure 6D) compared with control. Notably, concomitant treatment with GBE fully abolished all ECG changes found in ISO-treated rats (Figures 6A–C).

Impaired myocardial contractility is a hallmark of pathological ventricular remodeling, therefore we also assessed whether GBE prevents the loss of ventricular contractility through Langendorff-isolated heart Figure 7A. As shown in the Figure 7B, chronic β-AR stimulation markedly decreased the LVDP when compared with control group, whereas GBE treatment significantly ameliorated the contractile dysfunction of hypertrophied hearts. GBE treatment alone did not affect the LVDP. Furthermore, during heart perfusion, the coronary perfusion pressure was simultaneously recorded in order to evaluate the coronary vasomotor tone. Chronic β-AR stimulation led to a higher coronary pressure when compared with control, whereas GBE treatment fully abolished this effect (Figure 7C). Interestingly, animals that only received GBE showed a reduced coronary pressure compared to untreated rats.

Abnormal Ca²⁺ handling is a common feature of impaired cardiac contractility, therefore it was evaluated the sarcoplasmic reticulum Ca²⁺ pump (SERCA2) expression in the heart. SERCA2 protein levels were significantly decreased in ventricles from ISO-treated rats compared to control (Figure 7D). In contrast, GBE treatment of ISO-treated rats fully restored the physiological levels of SERCA2.

Taken together, our data suggest that the cardioprotective action of GBE involves the activation of cholinergic signaling. Therefore, to validate our findings, we next performed experiments on primary cultures of neonatal rat ventricular myocytes, which constitute a reliable in vitro model. Thus, in order to mimic the pathological conditions elicited by sustained β-AR stimulation, neonatal cardiomyocytes were treated with ISO and cellular hypertrophy was evaluated by measurement of myocyte surface area. As shown in the Figure 8, cardiomyocytes treated with ISO showed increased cell surface area, whereas co-treatment with GBE fully prevented this effect. Importantly, antihypertrophic action of GBE was abolished by atropine, a muscarinic receptor antagonist, or L-NAME, an inhibitor of NOS. Moreover, GBE alone had no effect on cellular area. Altogether, these data show that GBE antihypertrophic effect occurs via activation of M₂/NO pathway.