Peripheral chemoreceptors sustain central chemoreflex potentiation and cardiorespiratory abnormalities in high-output heart failure

Abstract

Central chemoreflex activation worsens cardiorespiratory dysfunction in high-output heart failure (HO-HF). Recently, interdependence between both peripheral and central chemoreceptors has been linked to alterations in cardiorespiratory regulation. Whether central chemoreflex potentiation in HO-HF requires sensory inputs from peripheral chemoreceptors remains completely unknown. Accordingly, we hypothesized that peripheral–central chemoreceptor interaction promotes cardiorespiratory dysfunction in non-ischemic HO-HF. We used male Sprague–Dawley rats to investigate the role of carotid bodies (CBs), the main peripheral chemoreceptors, on autonomic, respiratory, and cardiac function alterations during the progression of HO-HF. CB denervation (CBD) was used to eliminate CB inputs in HO-HF rats. The effect of CBD on HO-HF related cardiac, autonomic, and ventilatory function was measured using echocardiography, pressure–volume loop analysis, electrocardiography, plethysmography, and telemetry. HO-HF rats exhibited enhanced central chemoreflex drive, irregular breathing, autonomic imbalance, cardiac electrophysiological abnormalities, cardiac diastolic dysfunction, and cardiac hypertrophy. Remarkably, CBD completely normalized central chemoreflex function in HO-HF rats, restored ventilatory stability, reduced apnea–hypopnea incidence, improved heart rate variability, shortened QRS and PR intervals, attenuated collagen deposition, and ameliorated diastolic dysfunction. Additionally, CBD also corrected respiratory–cardiovascular coupling abnormalities in HO-HF rats. These findings demonstrate that an intact and functional CB is necessary for the development of cardiorespiratory disturbances in non-ischemic HO-HF. Targeting CB–central chemoreceptor interdependence may represent a novel therapeutic approach for non-ischemic HO-HF.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00395-025-01154-5.

Introduction

Heart failure (HF) is a life-threatening disease characterized by cardiorespiratory decline; it affects more than 64 million adults worldwide, and it’s especially prevalent in the aging population [17, 44]. Non-ischemic HF, particularly, is increasing in prevalence, as it remains understudied. The treatment of non-ischemic HF is challenging, as its pathophysiological mechanisms are poorly understood. The standard treatment for HF is to block the adrenergic and renin–angiotensin systems, but this approach is often ineffective in the long-term [47, 53, 55]. An alternative approach is to intervene in the malfunction of visceral reflexes [14–16, 21]. In clinical studies of HF patients, dysregulated chemoreception played a crucial role in survival: when both peripheral and central chemoreceptors become overactive, survival decreases [14–16]. This suggests that reducing this overactivation may be an effective treatment to reduce cardiorespiratory disturbances in patients with HF.

The central and peripheral chemoreflex responses are controlled, respectively, by the carotid bodies (CBs) and the retrotrapezoid nucleus (RTN). CBs and RTN are chemoreceptors that sense the arterial pressure of O₂ (P_aO₂) and/or CO₂ [18, 22]. CBs communicate with the RTN through a glutamatergic anatomical-functional connection mediated by the caudal nucleus of the solitary tract (cNTS) [12, 49]. In patients, both acute and long-term removal of CBs depresses the central chemoreflex drive [5, 13, 28, 34]; this strongly suggests that both CBs and RTN work cooperatively to modulate the ventilatory response through interdependent feedback [18] and that central chemosensory activity could be modulated by peripheral CB chemoreceptors.

It is unknown whether CB–RTN interdependent feedback plays a role in high-output non-ischemic HF pathophysiology. We hypothesize that communication between CBs and RTN promotes the development of cardiorespiratory distress during the progression of non-ischemic HF, and that denervation of CBs could treat the cardiovascular and respiratory abnormalities present in non-ischemic HF. Accordingly, we study the effects of CB denervation (CBD) on autonomic, respiratory, and cardiac function alterations during the progression of high-output non-ischemic HF in rats, a model that recapitulates key aspects of the pathophysiology observed in patients with HF with preserved ejection fraction (HFpEF).

Materials and methods

Animal care and housing

Male Sprague Dawley rats (P90) were housed at 25 °C and a light/dark cycle of 12 h with ad libitum access to water and food. All experiments were approved by the Comité de Bioética from the Pontificia Universidad Católica de Chile (Protocol #170710022) and by the University of Kansas Medical Center IACUC (Protocol #24-02-378). All experiments were designed and performed according to the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. Rats were randomly assigned to 3 experimental conditions: i) Sham (control), ii) high-output HF (HO-HF), and iii) high-output HF + CB denervation (HO-HF_CBD). The experimental timeline is shown in Supplemental Fig. 1.

High-output heart failure model

A high-output heart failure (HO-HF) model in rats was used in this study. Importantly, HO-HF is characterized by preserved or elevated cardiac output, cardiac remodeling, increased filling pressures, and neurohumoral activation, a constellation of hemodynamic features that differ from pressure-overload or ischemic HF models and more closely parallels non-ischemic HF and preserved ejection fraction HF [2, 9, 10, 27, 45, 46, 50, 52]. Briefly, under isoflurane anesthesia (5% for induction and 2% for maintenance, balanced with O₂), we performed a laparotomy to clean and isolate the abdominal aorta and inferior cava vein. The vessels were clamped and connected by an arterio-venous shunt (AV-shunt) using a 19G needle. Then, the aperture was closed using tissue glue (Hystoclear, Bauman). Laparotomy was finally sutured in layers with sterile absorbable silk 4–0. Sham animals underwent a similar procedure, with the exception that an AV-shunt was not performed.

Echocardiography

At 4 and 8 weeks after induction of heart failure, we performed echocardiography recordings of the left ventricle (LV). Transthoracic M-mode echocardiography (Mindray Z6 Vet) was made from the left parasternal short-axis in isoflurane-anesthetized rats (5% for induction and 2% for maintenance, balanced with O₂). The LV end-diastolic and systolic diameter (EDD and ESD, respectively) were averaged from 3 consecutive cardiac cycles and used for mathematical estimation of end-systolic and end-diastolic volumes (EDV and ESV, respectively) derived from Teicholz method according to the equation: $LVESV=\frac{7\times {\mathit{ESD}}^3}{2.4+ESD}$ and $LVEDV=\frac{7\times {\mathit{EDD}}^3}{2.4+EDD}$. Then, ejection fraction (EF), cardiac output (CO), stroke volume (SV), heart rate (HR) and fractional shortening (FS) were calculated. To consider a rat to be under HO-HF, ESV, EDV, and SV values must be 2.5 standard deviation over Sham average [2, 9, 10, 27, 45, 46, 50, 52], no load-dependent parameters were considered to allocate animals into the HO-HF group.

Carotid body chemosensory recordings

Sham and HO-HF rats were anesthetized with sodium pentobarbitone (60 mg/kg, i.p.) and placed in a supine position. A neck midline incision was made to visualize the trachea, and a plastic tube was inserted. Then, the carotid bifurcation was exposed, and the carotid sinus nerves were separated from surrounding tissue and severed (bilaterally) at their origin within the glossopharyngeal nerve. One nerve was placed in paired Pt/Ir electrodes connected to an AC preamplifier (P55, Natus Neurology, USA) and covered with warm mineral oil [35, 41]. The electroneurogram was band-pass filtered (10 Hz–1 kHz) and amplified. The signal was digitally acquired at 2 kHz (PowerLab 8SP, ADInstruments, USA) and stored as binary files. A spike amplitude discriminator window was used to assess the chemosensory discharge frequency in 1 s bins (ƒx) [35, 41]. Animals breathed spontaneously throughout the experiment and were submitted to acute changes in oxygen inspiratory fraction (F_iO₂: 0–100%).

Carotid body denervation

Animals were anesthetized with isoflurane (5% for induction and 2% for maintenance, balanced with O₂), and a midline neck incision was performed to expose the carotid bifurcation [15, 34]. Both carotid sinus nerves (CSN) were isolated, and a drop of 2% lidocaine was applied before bilateral transection. After denervation, the skin was sutured with absorbable 4–0 silk, and rats received meloxicam (2 mg/kg s.c.) and enrofloxacin (5 mg/kg s.c.) for 3 days. Then, CB denervated animals underwent volume-overload HO-HF surgery. CB denervation (CBD) was not performed on healthy animals in this study, as the effects of CBD on cardiorespiratory physiology in healthy animals have already been extensively studied [8, 34, 42, 48].

Ventilatory and chemoreflex function

Respiratory function was recorded by whole body plethysmography (Emka, France) in non-anesthetized unrestrained rats. Ventilatory parameters were tidal volume (V_T), respiratory frequency (R_F), and minute ventilation (V_T * R_F = V_E) [9, 10, 45, 46, 50–52]. Resting breathing was recorded for 2 h, while the plethysmograph chamber was flushed with room air at 2.5 L/min. We quantified breathing disorders (apnea, hypopnea, sigh, and post-sigh apneas (PSA)) and their duration from the last hour of the recording, as previously described [8, 9, 44–46]. Breathing irregularity was also studied by calculation of the irregularity score (IS) (TTOT_n – TTOT_n-1/TTOT_n), coefficient of V_T variation, and breath-to-breath (B-B) interval dispersion in the long (SD1) and short (SD2) axes of the Poincare plot [9, 10, 50–52]. Peripheral and central chemoreflex functions were assessed by calculating the hypoxic and hypercapnic ventilatory responses (HVR and HCVR, respectively), corresponding to the slope between baseline and maximum ventilatory response following a hypoxic (F_IO₂ 10%) or hypercapnic (F_ICO₂ 7%) challenge [9, 10, 45, 46, 50, 52].

Cardiac function and respiratory-triggered modulation of cardiac cycle

We anesthetized animals with alpha-chloralose and urethane (40 and 800 mg/kg i.p., respectively), then recorded close chest pressure–volume loops (PV-loop) from the LV using a conductance catheter (SPR-869, Millar) connected to an analog–digital recording system PowerLab 16/35 (ADInstruments) and LabChart software v7.2.1 (ADInstruments). The catheter was slowly introduced to the LV through the common carotid artery and left in place for 30 min for signal stabilization. After catheter positioning, a laparotomy was performed to localize and isolate the inferior cava vein. Preload-dependent and independent cardiac function-related parameters were calculated as previously described [2, 10, 50, 52]. Ventilation-dependent modulation of cardiac function was assessed using single-beat PV loop analysis [10, 25, 27, 45]. For this, data was binned to 5 s intervals and LV pressures and volumes (EDP and EDV) were analyzed in relation to the inspiratory and expiratory phases of the breathing cycle [25, 45].

Electrocardiogram and heart rate variability

The second lead electrocardiogram (ECG) was recorded using LabChart v7.2.1 software in anesthetized rats. After signal stabilization (30 min), ECG was recorded for 20–30 min. Alterations in cardiac electrophysiology were determined by analyzing P, Q, R, and T wave amplitude, and P-R segment, P wave, QRS complex, and QT interval duration [27, 40]. Heart rate variability (HRV) was analyzed in 10 min artifact-free lead II ECG recordings using the HRV plug-in from LabChart v7.2.1 software. Briefly, estimation of power spectral density (PSD) was performed using the Fourier Fast Transformation algorithm after Hann windowing with 1/2 overlapping. Cut-off frequencies were defined as follows: low frequency (LF) 0.04–0.6; high frequency (HF) 0.6–2.5 Hz, as previously described [2, 9, 10, 50, 51].

Radiotelemetry implant and cardio-respiratory coupling

Under isoflurane anesthesia (5% for induction and 2% for maintenance, balanced with O₂), we performed a leg skin incision to isolate the femoral artery. Briefly, the transducer tip (HD-S10, DSI) was introduced in the femoral artery and fixed with non-absorbable suture 6–0, while the transmitter was positioned subcutaneously. The arterial blood pressure from unrestrained and non-anesthetized animals was recorded using an indwelling radio-telemetry device (Data Science International) at a sample rate of 1000 Hz [2, 9, 10, 45, 51]. Once the animal recovered from telemetry implantation surgery, they were transferred to the plethysmograph system to perform simultaneous hemodynamic and respiratory recordings. Respiratory-cardiovascular coupling was examined by evaluating coherence between low frequency oscillations in respiration (i.e., oscillatory breathing) and the low frequency domain of HR (i.e., sympathetic predominance) as previously described [9, 10, 30, 51]. Briefly, 3 min artifact-free recordings underwent auto- and cross-spectral analysis employing Welch’s overlapped segment averaging method in a custom-made Matlab routine. Signal derived from respiratory slow oscillations served as the input, while the HRV low frequency served as the output for coherence analysis. The mean square coherence magnitude was evaluated within a 0.015 Hz range centered on the frequency of the maximum V_T spectral peak in the very low frequency (vLF) domain (vLF: 0.01–0.25 Hz) to capture breathing oscillations [9, 10, 30, 51]. Coherence values exceeding 0.5 indicated positive interaction between these signals. Additionally, the phase was calculated at the respiratory spectral peak.

Picrosirius red protocol

At the end of the protocol, rats were fixed with paraformaldehyde 4% by intracardiac perfusion. The hearts were included in paraffin, and 10 µm transversal sections were obtained using a microtome (Leica). Heart sections were deparaffinized, rehydrated, and incubated in concentrated picric acid at 60 °C for 30 min, then incubated with picrosirius red (ScyTek Laboratories, SRS250) for 30 min at room temperature. Finally, the samples were mounted with Entellan (Merck, HX1260033). Optical images of the stained sections were captured at 100 × magnification using a light microscope (Olympus CX31). Collagen deposition in myocardial tissue was quantified using Fiji Software [27].

Retrotrapezoid nuclei microdialysis

We anesthetized animals using a mixture of ketamine/xylazine (100 and 10 mg/kg, respectively), placed the animal on a stereotaxic holder (model 900; David Kopf 161 Instruments, USA), and exposed the upper side of the skull to localize the RTN with respect to bregma (AP: − 11.64, ML: ± 2, DV: 8.7 mm) [9]. The microdialysis probe (shaft length of 7 mm with a 2 mm long cuprophane) was activated at 2 µL/min flow using sterile cerebrospinal fluid (aCSF). The membrane was inserted directly into the animal while it was secured in the stereotaxic frame and slowly crossed through the parenchyma at a speed not exceeding 100 µm/s. Then, we allowed the probe to equilibrate for at least 1 h before starting extracellular fluid collections to allow recovery from potential tissue damage by its insertion. Samples were collected for 20 min at 1 µL/min during resting breathing in normoxic air. Body temperature was maintained at 35–37 °C.

Data analysis

Data is presented as mean ± SEM in main text and box-plot min to max in Figures. Two-way or one-way ANOVA parametric test followed by Holm-Sidak post hoc was used to assess statistical difference using Graph Pad Prism 8.0 software. P < 0.05 was considered significant.

Results

Cardiac left ventricle dimensions and baseline hemodynamic in HO-HF rats

Echocardiography was used to determine the effects of CBD on LV dimension and hemodynamic function in rats compared to Sham. Both HO-HF and HO-HF_CBD had increased LV chamber dimensions (Table 1), resulting in higher volumes during systole and diastole (ESV and EDV, respectively). HO-HF increased stroke volume (SV) and cardiac output (CO) compared to Sham but did not alter heart rate (HR). HO-HF also increased cardiac index and hypertrophy (heart weight to body weight) in both HO-HF and HO-HF_CBD animals compared to Sham. No significant changes in body weight were found between experimental conditions.

Table 1.

Baseline hemodynamic and ventilatory parameters

	Sham (n = 10)	HO-HF (n = 10)	HO-HFCBD (n = 12)
LVEDD (mm)	6.1 ± 0.22	9.11 ± 0.22*	9.26 ± 0.23₸
LVESD (mm)	3.2 ± 0.36	5.65 ± 0.21*	5.67 ± 0.23₸
HR (bpm)	331.4 ± 9.78	346.5 ± 14.11	336.2 ± 6.67
LVEDV (µL)	191.0 ± 16.23	463.3 ± 25.65*	481.70 ± 25.69₸
LVESV (µL)	49.0 ± 12.13	158.0 ± 14.18*	162.20 ± 13.56₸
SV (µL)	142.0 ± 9.04	305.3 ± 20.76*	319.40 ± 19.63
EF (%)	76.4 ± 4.55	65.8 ± 2.50	66.46 ± 2.31
FS (%)	48.2 ± 4.48	37.97 ± 1.88	38.87 ± 1.88
CO (mL/min)	47.1 ± 3.26	104.8 ± 6.89*	107.60 ± 6.87₸
CI (mL/min/cm2)	0.07 ± 0.01	0.16 ± 0.01*	0.17 ± 0.01₸
HW/BW (mg/g)	2.8 ± 0.08	4.0 ± 0.34*	3.9 ± 0.26₸
BW (g)	500.4 ± 4.73	491.6 ± 16.68	483.10 ± 9.56
RF (bpm)	87.9 ± 2.2	88.8 ± 2.9	89.3 ± 3.2
VT (ml/100)	0.24 ± 0.01	0.26 ± 0.01	0.25 ± 0.01
VE (ml/min/100 g)	21.4 ± 1.3	22.7 ± 0.8	21.8 ± 1.0

Data is presented as mean ± SEM

LVEDD left ventricle end-diastolic diameter, LVESD left ventricle end-systolic diameter, HR heart rate, LVEDV left ventricle end-diastolic volume, LVESV left ventricle end-systolic volume, SV stroke volume, EF ejection fraction, FS fractional shortening, CO cardiac output, CI cardiac index, BW body weight, R_F respiratory frequency, V_T tidal volume, V_E minute volume

*P < 0.05 Sham vs. HO-HF

^₸P < 0.05 Sham vs HO-HF_CBD

CBD prevented central chemoreflex potentiation in HO-HF rats

Dysfunctional chemoreflex is associated with severe pathophysiology in patients and animals with HF [14–16, 50]. We found that HO-HF rats showed potentiation of the central chemoreflex, but not the peripheral chemoreflex. Indeed, for the hypercapnic ventilatory responses (HCVR), both the magnitude (∆V_E: 28.8 ± 3.4 vs. 40.4 ± 3.2 ml/min/100 g, Sham vs. HO-HF) and sensitivity (4.1 ± 0.5 vs. 5.8 ± 0.4 ∆V_E/%F_ICO₂, Sham vs. HO-HF) were significantly augmented in HF compared to Sham (Fig. 1A, E–G). In contrast, HO-HF did not change the hypoxic ventilatory responses compared to Sham (Fig. 1A–D; Supplementary Table 1). We confirmed this result by performing direct afferent recordings of the CB and found no differences in the chemosensory response to hypoxia between HO-HF and Sham rats (Supplemental Fig. 2). Remarkably, HO-HF_CBD rats had significantly reduced potentiation of central chemoreflex drive compared to both HO-HF and Sham rats, suggesting that it requires functional CB afferent input (Fig. 1A, C, and D). Indeed, CBD normalized both the magnitude (∆V_E: 28.8 ± 3.4 vs. 28.6 ± 2.2 ml/min/100 g, Sham vs. HO-HF_CBD) and sensitivity (4.1 ± 0.5 vs. 4.1 ± 0.3 ∆V_E/%F_ICO₂, Sham vs. HO-HF_CBD) of the hypercapnic ventilatory responses, as HO-HF_CBD rats were not significantly different from Sham rats with respect to these phenotypes. No differences in resting ventilation parameters were found between experimental groups (Table 1).

Fig. 1.

CBD prevents central chemoreflex potentiation in HF rats. A Representative recording for ventilatory flow (left Y axis) and respiratory frequency (right Y axis) from one animal per group. B Summary data showing changes in magnitude and C gain of the hypoxic ventilatory response (HVR). D Change in minute ventilation (V_E) pre-, during (gray square), and post-hypoxic (F_IO₂ 10%) stimulation. E Summary data showing changes in magnitude and F gain of the hypercapnic ventilatory response (HCVR). G Change in minute ventilation (V_E) pre-, during (gray square), and post-hypoxic (F_ICO₂ 7%) stimulation. One-way ANOVA was performed to compare groups followed by Holm Sidak post hoc. Sham, n = 10; HO-HF, n = 10; HO-HF_CBD, n = 12

CBD improves breathing disturbances in HO-HF rats

Patients and animals with HF experience breathing irregularity. We found that HO-HF rats displayed marked alterations in the resting breathing patterns compared to Sham rats, as evidenced by significant increases in the breath-to-breath interval variability, tidal volume variations, and breathing irregularity scores (Fig. 2A–G). We next examined how CBD affected the development of irregular breathing (Fig. 2A–G). We found that, compared to HO-HF, HO-HF_CBD rats had improved breath-to-breath interval variability (SD1: 70.5 ± 3.2 vs. 46.6 ± 3.1 ms; SD2: 102.2 ± 8.1 vs. 76.8 ± 4.2 ms, HO-HF vs. HO-HF_CBD), reduced V_T amplitude oscillation (25.5 ± 0.8% vs. 18.7 ± 0.7%, HO-HF vs. HO-HF_CBD), and improved breathing irregularity score (IS: 13.4 ± 0.6% vs. 11.0 ± 0.5%, HO-HF vs. HO-HF_CBD). More remarkably, CBD reduced by 65% the incidence of apneas/hypopneas in HO-HF rats (Fig. 2H). There were no differences in the number of sighs or post-sigh apneas between the three groups (Supplemental Table 2). Since the interplay between CB-RTN may be glutamatergic in nature, we next measured the baseline concentration of glutamate in the RTN. We found that HO-HF rats exhibited evidence of glutamate spillover in the RTN: compared to Sham rats, HO-HF rats had a twofold increase in extracellular glutamate (4.3 ± 1.01 vs. 8.1 ± 0.92 pmol/µl, Sham vs. HO-HF). In HO-HF rats, CBD prevented RTN glutamate spillover (8.1 ± 0.92 vs. 2.6 ± 0.82 pmol/µl, HO-HF vs. HO-HF_CBD) (Supplemental Fig. 3).

Fig. 2.

CBD reduces disordered breathing in HF rats. A Representative ventilatory flow and respiratory frequency (R_F) recording, B histogram of tidal volume (V_T) amplitude distribution, and C 3D Poincaré plots from one animal per group. Arrowheads in A showed the incidence of breathing disturbances on the resting breathing pattern. Summary data showing D short axis (SD1), E long axis (SD2) breath-to-breath interval variability, F irregularity score (IS), G coefficient of V_T variation, and H apnea–hypopnea index (AHI). One-way ANOVA was performed to compare groups followed by Holm–Sidak post hoc. Sham n = 10, HO-HF n = 10, HF_CBD n = 12

CBD prevents autonomic and cardiac electrical alterations in HO-HF rats

Next, we studied the effect of HO-HF and CBD on cardiac autonomic control by measuring heart rate variability (HRV). We found that HO-HF rats displayed evidence of autonomic imbalance with sympathetic predominance. The low frequency (LF_HRV) component of the HRV was significantly higher in HO-HF rats compared to Sham rats (LF_HRV: 26.1 ± 5.5 vs. 49.1 ± 3.7 nu, Sham vs. HO-HF) (Fig. 3A and D). Parasympathetic withdrawal was also found in HO-HF rats (Fig. 3E). In contrast, CBD normalized cardiac autonomic control in HO-HF rats (Fig. 3A, D–F): in HO-HF_CBD rats, LF/HF_HRV was significantly lower than in HF rats, and not significantly different than in Sham rats (LF/HF_HRV: 0.47 ± 0.15 vs. 0.44 ± 0.07 nu, Sham vs. HO-HF_CBD). Since cardiac autonomic imbalance is highly linked to alterations in cardiac electrical conduction, we next evaluated the effects of CBD on cardiac electrophysiology in HO-HF. Compared to Sham rats, HO-HF rats had significant alterations in the duration of the QRS complex (0.0195 ± 0.001 vs. 0.0225 ± 0.001 s, Sham vs. HO-HF) as well as in the PR interval (0.0478 ± 0.002 vs. 0.0604 ± 0.003 s, Sham vs. HO-HF). CBD prevented the development of prolonged QRS and PR intervals in HO-HF rats (Fig. 3B, C and G–H). No significant alterations in the amplitude of ECG waves were found between groups (Supplemental Table 3).

Fig. 3.

CBD prevents alterations in cardiac autonomic control and electrical conductivity in HF rats. A Representative heart rate variability (HRV) power spectral density (PSD) plots, B Electrocardiogram (ECG) waterfall plot and C averages traces from one animal per condition. Summary data showing D Low Frequency band, E High Frequency (HF) band, F LF/HF ratio G QRS complex duration, and H PR interval duration. One-way ANOVA was performed to compare groups followed by Holm Sidak post hoc. ECG: Sham, n = 6; HO-HF, n = 7; HO-HF_CBD, n = 5. HRV: Sham, n = 9; HO-HF, n = 10; HO-HF_CBD, n = 7

CBD ameliorates cardiac dysfunction and remodeling in HO-HF rats

HF is characterized by marked impairment of cardiac function [2, 33, 50]. HO-HF rats displayed both diastolic and systolic impairment, as evidenced by deterioration in the end-diastolic and systolic pressure volume relationships (EDPVR and ESPVR, respectively). Compared to Sham rats, HO-HF rats had a ~ fourfold increase in EDPVR and a ~ threefold reduction in ESPVR (Fig. 4A, C, D). This was accompanied by overt signs of cardiac fibrosis (Fig. 4B) and increased LV end-diastolic pressures (EDP) (Fig. 4E), but no significant changes in LV end-systolic pressures (Fig. 4F). Next, we found that CBD improved cardiac function in HO-HF rats. In HO-HF_CBD rats, deterioration in EDPVR was completely prevented (EDPVR: 0.0095 ± 0.001 vs. 0.003 ± 0.001 mmHg/µL, HO-HF vs. HO-HF_CBD) (Fig. 4A and C), and the increase in EDP was significantly reduced (Fig. 4E). Collagen deposition in the LV increased in HO-HF rats compared to Sham rats; this increase was abrogated in HO-HF_CBD rats (Fig. 4B, G). CBD modestly improved cardiac systolic function in HO-HF animals (Fig. 4D). Lastly, CBD also prevented alterations in the filling properties of the heart during HO-HF (Table 2).

Fig. 4.

CBD improves cardiac diastolic function and reduces intramuscular collagen deposition in HF rats. A Representative pressure–volume (PV) loops obtained from one animal per condition. B Representative histological LV sections stained with picrosirius red. Summary data showing C LV end-diastolic pressure volume relationship (EDPVR), D LV end-systolic pressure volume relationship (ESPVR), E LV end-diastolic pressure (EDP), F LV end-systolic pressure (ESP), and G LV collagen content. One-way ANOVA was performed to compare groups followed by Holm–Sidak post hoc. PV-loop: Sham, n = 7; HO-HF, n = 10; HO-HF_CBD, n = 7. PSR: Sham, n = 10; HO-HF, n = 10; HF_CBD, n = 8

Table 2.

PV-loop LV cardiac function parameters

	Sham (n = 7)	HO-HF (n = 8)	HO-HFCBD (n = 6)
CO (mL/min)	64.11 ± 6.18	116.23 ± 12.7*	106.81 ± 5.71₸
SV (µL)	170.30 ± 15.22	330.40 ± 22.03*	282.10 ± 16.05₸
EDV (µL)	252.4 ± 17.47	464.6 ± 27.23*	423.1 ± 20.98₸
ESV (µL)	82.15 ± 10.53	134.3 ± 14.79*	153.3 ± 8.06₸
ESP (mmHg)	110.7 ± 7.09	113.7 ± 4.24	115.5 ± 3.19
EDP (mmHg)	3.30 ± 0.27	5.91 ± 0.78	4.03 ± 0.28
HR (bpm)	377.9 ± 21.16	348.6 ± 25.21	383.5 ± 10.61
EF (%)	74.56 ± 4.40	77.77 ± 4.31	70.39 ± 2.38
Ea (mmHg/µL)	0.54 ± 0.04	0.28 ± 0.01*	0.33 ± 0.02₸
dP/dt max (mmHg/s)	9410 ± 1496	8822 ± 814	10,594 ± 1190
dP/dt min (mmHg/s)	− 6268 ± 788	− 5888 ± 639	− 6249 ± 789
dV/dt max (µL/s)	15,752 ± 2721	21,360 ± 3749*+	15,250 ± 2655
dV/dt min (µL/s)	− 12,997 ± 1893	− 17,937 ± 1614*+	− 12,825 ± 879
Tau (s)	9.33 ± 1.2	8.65 ± 0.62	10.43 ± 0.74

⁺P < 0.05 HO-HF vs HO-HF_CBD

Data is presented as mean ± SEM

CO cardiac output, SV stroke volume, EDV end diastolic volume, ESV end systolic volume, EDP end-diastolic pressure, HR heart rate, EF ejection fraction, Ea elastance, dP/dt max maximal rate of rise of ventricular pressure, dP/dt min minimal rate of left ventricular pressure decay, dV/dt max filling velocity, dV/dt min empty rate velocity, Tau the time constant of relaxation. One-Way ANOVA test was performed

*P < 0.05 Sham vs. HO-HF

^₸P < 0.05 Sham vs. HO-HF_CBD

CBD normalizes respiratory-cardiovascular pathological coupling in HO-HF rats

Dysregulation in the respiratory-cardiovascular coupling contributes to the progression of HF pathophysiology [8, 30, 39]. We studied respiratory-cardiovascular coupling using analysis of coherence between oscillatory breathing and cardiac sympathetic-like activity. We found that HO-HF rats displayed evidence of respiratory-sympathetic coupling, as shown by significant coherence between low-frequency oscillation in breathing and heart rate (Fig. 5A and D). Notably, CBD prevented the pathological entrainment of respiration and sympathetic activity in HO-HF rats: HO-HF_CBD rats showed no significant coherence (0.56 ± 0.02 vs. 0.29 ± 0.04, HO-HF vs. HO-HF_CBD). Importantly, respiratory-cardiovascular coupling in HO-HF rats translated into significant impairments of breathing modulation of cardiac function. HO-HF rats presented active modulation of cardiac diastolic function during expiratory phases (of breathing), which is absent on healthy Sham rats (Fig. 5F). Accordingly, HO-HF rats displayed a ~ tenfold increase in LV developed diastolic pressures between expiration and inspiration compared to Sham rats (Fig. 5G), and these effects were abrogated in HO-HF_CBD rats (Fig. 5E-G).

Fig. 5.

CBD abolishes pathological respiratory-sympathetic coupling in HF rats. A Representative blood pressure (top), tidal volume (V_T, middle), and wavelet based-coherence (bottom) analysis obtained from one animal per group. B Representative simultaneous recordings of LV pressures (top) and ventilatory flows (bottom), and C single-beat PV-loop analysis during the expiratory and inspiratory phases of the breathing cycle obtained in one rat per group. Summary data for D very Low Frequency (vLF) Coherence, E single-beat analysis of EDPVR (V₀), F intraventricular pressure at normalized end-diastolic volume (nEDV) for each phase of respiration, and G percent changes in ∆LVpressure at Exp-Insp for each group. Paired t-test (E, F) and one-way ANOVA followed by Holm Sidak post hoc (D, G). PV-loop vent-dependent: Sham, n = 7; HO-HF, n = 10; HO-HF_CBD, n = 7. Cardio-respiratory coupling: Sham, n = 4; HO-HF, n = 6; HO-HF_CBD, n = 4

Discussion

Chemoreflex is the primary mechanism that controls ventilatory response to fluctuations in arterial gases. Chemoreceptors, whether peripheral or central, have a major influence over neural circulatory control in scenarios marked by substantial variations in arterial O₂ and/or CO₂ levels, inducing a range of reflex cardiovascular and respiratory responses [14–16, 37]. Abnormal central and/or peripheral chemoreflex function is closely associated with poor cardio-respiratory outcomes in several conditions, including HF, and reduces quality of life and overall survival [14–16, 19, 37]. Central chemoreceptors in the RTN (a major site for central chemoreception) have gained attention in several disease conditions [14–16, 50]: in HF, central chemoreceptors contribute to the continuous deterioration of cardiac function through the development/maintenance of autonomic imbalance and disordered breathing [9, 50]. When the RTN is ablated in HF, it restores normal chemoreflex drive and breathing, suggesting that RTN neurons are necessary for chemoreflex potentiation in the setting of HF. However, the pathophysiological mechanisms of RTN dysfunction in HF remain unknown. Interestingly, several studies have proposed that central (i.e., RTN) and peripheral (i.e., CB) chemoreceptors are interdependent and function together to orchestrate the cardio-respiratory chemoreflex response in a supra-additive, additive, or hypo-additive manner, dependent on the physiological context [3–5, 49]. It is unknown whether chemoreceptor interdependence plays a role in HF pathophysiology. Here, we hypothesized that CB chemosensory inputs may ‘tune’ RTN chemoreceptor function in HF by adjusting central chemoreflex drive; therefore, we reduced peripheral chemoreflex drive through CBD, and investigated whether it modulated central chemoreflex drive in HO-HF. We showed, for the first time, that CB tonicity is necessary to adjust central chemoreflex drive in HO-HF through a mechanism partially linked to glutamate spillover in the RTN. Remarkably, eliminating CB afferent drive abolished the potentiation of central chemoreflex, reduced the incidence of disordered breathing and sympathoexcitation, prevented the development of pathological respiratory-cardiovascular coupling, and reduced cardiac dysfunction.

Sympathoexcitation is a distinctive feature of HF [2, 30]. We hypothesize two possible mechanisms by which CBD may contribute to autonomic modulation in HF. First, through a direct effect on RTN chemoreception, by regulating glutamate content within the RTN. We found significant increases in extracellular glutamate in the RTN of HO-HF rats, which were prevented by CBD. We proposed that CB-RTN crosstalk encompasses, at least in part, fine-tuning of extracellular glutamate levels. Therefore, the effects of CBD on autonomic regulation may be associated with the normalization of the central chemoreflex drive in HO-HF. Second, by indirect regulation of pre-sympathetic motoneurons at the rostral ventrolateral medulla (RVLM), through second-order synaptic projections from the caudal NTS (the main site for CB-mediated afferent integration) [12, 24, 25]. Further work is necessary to investigate whether this effect may be linked to neural plasticity and/or alterations in neurochemical balance within the RVLM of HF animals. However, we did show that CBD prevented the enhanced cardiac sympathetic drive at rest during eupneic conditions; we speculate that CB tonicity may also be relevant to adjusting excitability thresholds on RVLM neurons in the pathology of HF.

We found that performing CBD before the induction of HO-HF prevented alterations in cardiac electrophysiology and cardiac function. The latter effect, particularly, could have therapeutic implications, as CB modulation has been proposed as a treatment for co-morbidities in pathological scenarios [15, 16, 30, 38]. Here, we found that HO-HF elevated QRS and PR interval duration, and that CBD prevented these electrophysiological changes. Importantly, prolonged QRS and PR intervals have major implications for cardiac arrhythmogenesis, as decreases in QRS complex duration are associated with a lower risk of arrhythmia incidence [7, 23, 27]. It is important to emphasize that the QRS and PR interval prolongations observed in mice with HO-HF are minimal and far smaller than those typically seen in human HF. Therefore, they cannot be extrapolated to clinically significant conduction disturbances such as left bundle branch block, that characterize a subset of HF patients and have major prognostic implications. Nevertheless, we found that CBD in HO-HF rats reduced extracellular collagen deposition, a structural substrate known to promote adverse remodeling-associated arrhythmias [27]. Thus, while the electrophysiological changes in this model are modest, the observed reduction in fibrotic remodeling may still hold mechanistic relevance for arrhythmic vulnerability. Of note, we did not test whether the beneficial effects of CBD were exclusively related to a decrease in adrenergic signaling to the heart [13, 29, 30]; this hypothesis requires further investigation.

Cardiorespiratory coupling (CRC) implies a mutual interplay between autonomic and respiratory control systems [11]. The emergence of CRC accentuates increased sympathetic tone and breathing instability in cardiovascular diseases [30, 50]. Our results confirm the presence of pathological CRC in HO-HF and additionally demonstrate how CB tonicity contributes to CRC during HO-HF. Interestingly, oscillatory breathing appears to trigger sympathetic entrainment in HO-HF, and CBD is sufficient to completely abolish CRC. Furthermore, pathological respiratory modulation of cardiac diastolic function was also evident in HO-HF animals, which showed expiration-induced further diastolic dysfunction. CBD prevented expiratory modulation of cardiac diastolic function in HO-HF rats.

Although high-output heart failure models provide valuable mechanistic insights, they have inherent limitations regarding clinical translation. The etiological factors leading to high-output states, such as anemia, hyperthyroidism, and arteriovenous shunting, are uncommon in contemporary HF populations, which are more frequently characterized by metabolic dysfunction, obesity, renal disease, or hypertension [6, 20, 31]. High-output states also evoke distinct compensatory adaptations, including peripheral vasodilation and plasma volume expansion, which can modulate ventricular stiffness, diastolic function, and ventricular–arterial coupling [1, 26, 43, 54]. In addition, comorbidities in human HF are frequently linked to chronic low-grade inflammation [32, 36], which may not be fully captured by high-output heart failure models. Thus, while translational extrapolation must be cautious, it is worth noting that the high-output model of HF importantly recapitulates several clinically relevant cardiorespiratory disturbances observed across the HF spectrum, including chemoreflex sensitization, ventilatory instability, and sympathetic activation.

Our findings demonstrate, for the first time, that an intact and functional CB is necessary for the development of high-output non-ischemic HF-associated cardiorespiratory disturbances. Therefore, targeting CB–central chemoreceptor interdependence may represent a novel therapeutic approach to control cardiovascular and respiratory dysfunction in non-ischemic HF.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

K.V.P. collected, analyzed, and interpreted data, and contributed to the preparation of the manuscript. E.D.J., K.S., S.V., F.O., I.B., and C.T. performed data analysis and contributed to the preparation of the manuscript. K.P., F.O., and R.D.R. performed data interpretation and contributed to the preparation of the manuscript. R.D.R. contributed to the concept of the project and experimental design. All data collecting was undertaken in the laboratories of R.D.R. All authors approved the final version of the manuscript.

Funding

Research reported in this publication was supported by Fondo de Desarrollo Científico y Tecnológico FONDECYT 1180172, 1220950 and by the National Heart, Lung, And Blood Institute of the National Institutes of Health under the Award number R01HL176779. The content is solely the responsibility of the authors and does not necessarily represent the official view of the National Institutes of Health.

Data availability

All data generated in this study are available in the manuscript and Supplementary Information. Additional data are available from the corresponding author upon reasonable request.

Declarations

Conflict of interest

The authors have no conflicts to disclose.