Premature ventricular contractions activate vagal afferents and alter autonomic tone: implications for premature ventricular contraction-induced cardiomyopathy

Abstract

Mechanisms behind development of premature ventricular contraction (PVC)-induced cardiomyopathy remain unclear. PVCs may adversely modulate the autonomic nervous system to promote development of heart failure. Afferent neurons in the inferior vagal (nodose) ganglia transduce cardiac activity and modulate parasympathetic output. Effects of PVCs on cardiac parasympathetic efferent and vagal afferent neurotransmission are unknown. The purpose of this study was to evaluate effects of PVCs on vagal afferent neurotransmission and compare these effects with a known powerful autonomic modulator, myocardial ischemia. In 16 pigs, effects of variably coupled PVCs on heart rate variability (HRV) and vagal afferent neurotransmission were evaluated. Direct nodose neuronal recordings were obtained in vivo, and cardiac-related afferent neurons were identified based on their response to cardiovascular interventions, including ventricular chemical and mechanical stimuli, left anterior descending (LAD) coronary artery occlusion, and variably coupled PVCs. On HRV analysis before versus after PVCs, parasympathetic tone decreased (normalized high frequency: 83.6 ± 2.8 to 72.5 ± 5.3; P < 0.05). PVCs had a powerful impact on activity of cardiac-related afferent neurons, altering activity of 51% of neurons versus 31% for LAD occlusion (P < 0.05 vs. LAD occlusion and all other cardiac interventions). Both chemosensitive and mechanosensitive neurons were activated by PVCs, and their activity remained elevated even after cessation of PVCs. Cardiac afferent neural responses to PVCs were greater than any other intervention, including ischemia of similar duration. These data suggest that even brief periods of PVCs powerfully modulate vagal afferent neurotransmission, reflexly decreasing parasympathetic efferent tone.

NEW & NOTEWORTHY Premature ventricular contractions (PVCs) are common in many patients and, at an increased burden, are known to cause heart failure. This study determined that PVCs powerfully modulate cardiac vagal afferent neurotransmission (exerting even greater effects than ventricular ischemia) and reduce parasympathetic efferent outflow to the heart. PVCs activated both mechano- and chemosensory neurons in the nodose ganglia. These peripheral neurons demonstrated adaptation in response to PVCs. This study provides additional data on the potential role of the autonomic nervous system in PVC-induced cardiomyopathy.

INTRODUCTION

Frequent premature ventricular contractions (PVCs) are known to cause heart failure (1, 13, 20, 47), and increased burden of PVCs is associated with increased risk of cardiomyopathy (4, 25, 29, 43). Precise mechanisms behind development of PVC-induced cardiomyopathy are unclear, but several studies have suggested changes in ventricular dynamics and dyssynchrony (31, 43), calcium handling, and oxygen consumption (27) as potential mechanisms.

The autonomic nervous system modulates every aspect of cardiac function (21, 38). Sympathetic activation and parasympathetic dysfunction often work in concert to increase risk of heart failure and sudden death (19, 34, 38, 46). It is possible that PVCs cause imbalances in the autonomic nervous system that contribute to cardiac dysfunction. PVCs can elevate sympathetic tone and have been reported to alter stellate ganglion neural activity (10, 26) and increase muscle sympathetic nerve activity, coronary sinus norepinephrine levels (15, 28, 41, 50), and the low-frequency component of heart rate variability (HRV) (3), all indices of elevated sympathetic tone. It has also been shown that PVCs can alter the activity of intrinsic cardiac ganglion neurons in vivo (16). However, the effect of PVCs on cardiac parasympathetic neurotransmission is unclear. Over 80% of the vagal trunk consists of sensory afferent fibers, for which neurons reside in the inferior vagal (nodose) ganglia. These neurons transduce visceral activity, including beat-to-beat variability of cardiac function, to alter subsequent parasympathetic efferent tone (21). Reflex withdrawal of parasympathetic tone would eliminate critical peripheral restraining influences on cardiac adrenergic function.

In this study, we hypothesized that PVCs adversely influence autonomic control by altering the activity of cardiac sensory nodose neurons, subsequently decreasing parasympathetic efferent neurotransmission to the heart. We also aimed to compare the effect of PVCs with ventricular ischemia, a known modulator of autonomic tone. Finally, we aimed to assess whether PVCs purely activate cardiac mechanosensitive neurons (as might be expected) or also alter activity of chemosensitive neurons, amplifying autonomic dysfunction.

METHODS

Sixteen Yorkshire pigs (59.2 ± 3.2 kg) were used in this study. Animal experiments were performed in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and approved by the University of California, Los Angeles Chancellor’s Animal Research Committee.

Animal Preparation

Animals were sedated (4–8 mg/kg im tiletamine-zolazepam) and intubated. Anesthesia for the surgical portion of the experiment was provided using isoflurane (1–5% inhaled). Following completion of surgical procedures and for neural recordings, anesthesia was switched to α-chloralose (50 mg/kg initial bolus, thereafter 20–35 mg·kg⁻¹·h⁻¹ iv; Sigma-Aldrich). The level of anesthesia was adjusted during the procedure by examining the corneal reflex, jaw tone, and hemodynamic indices. End-tidal CO₂ and oxygen saturation via pulse oximetry were monitored throughout the experiment. A water heating pad (T/PUMP; Gaymar Industries, Orchard Park, NY) was used to maintain temperature. Twelve-lead electrocardiography (ECG) was recorded via CardioLab System (GE Healthcare). Because of sternotomy, anterior precordial leads were placed on left dorsal aspect of the animal. Left and right femoral arteries and veins were cannulated to measure blood pressure and administer saline and drugs, respectively. The carotid artery was cannulated to obtain access to the left ventricle (LV). ECG and arterial blood pressure were digitized (Power1401; Cambridge Electronic Design, Cambridge, England), stored, and analyzed offline by the Spike2 program (Cambridge Electronic Design). Arterial blood gases were assessed every hour, and the tidal volume or respiratory rate was adjusted or sodium bicarbonate was given as needed to maintain a normal pH. Sternotomy was performed after a fentanyl bolus (20–30 µg/kg). An overdose of pentobarbital sodium (100 mg/kg iv; Med-Pharmex) followed by saturated KCl (1–2 mg/kg iv; Sigma-Aldrich) was used for euthanasia.

Histological Confirmation of the Nodose Ganglion

Nodose ganglia were removed after euthanasia and fixed in 4% paraformaldehyde for 24 h at 4°C and then embedded in paraffin. Tissue was sectioned (5 μm) and rehydrated in two toluene washes followed by three ethanol washes and water. Slides were blocked for 1 h in 3% BSA-TBS-0.2% Triton X-100 with 5% donkey serum and incubated overnight at 4°C with rabbit anti-PGP9.5 (1:200; ab108986; Abcam) and mouse anti-S100 (1:200; ab4066; Abcam) at 4°C followed by 2-h incubation at room temperature with Alexa Fluor 555 donkey anti-rabbit IgG and Alexa Fluor 488 donkey anti-mouse IgG (Invitrogen), respectively. Sections were imaged using a Zeiss LSM 880 with Airyscan (Zeiss).

HRV Analysis

HRV was analyzed in 10 animals that had suitable/clean ECG recordings pre- and post-PVCs using the Lomb periodogram nonparametric method for spectral analysis (LabChart software; ADInstruments, Colorado Springs, CO). HRV during the 1 min before induction of PVCs was compared with 1 min after cessation of PVCs. Given the short duration of the intervention and period of time (1 min) pre- versus post-PVC evaluated, only the high-frequency component of HRV was analyzed and used to assess parasympathetic function (0.15–0.4 Hz) (36). For HRV time-domain analysis, the following parameters were calculated: standard deviation of RR interval, coefficient of variance of RR intervals, standard deviation of successive RR interval differences between adjacent RR intervals, root-mean-square successive difference, and number of pairs of adjacent RR intervals differing by >50 ms to all RR intervals.

Premature Ventricular Contractions

PVCs were induced for 1 min using intermittent endocardial right ventricular (RV) pacing via a quadripolar catheter (Abbott, St. Paul, MN) placed from the femoral vein and attached to a cardiac stimulator (EPS320; Micropace, Canterbury, Australia). PVCs were generated on average every 5 beats (range 4–7) with random/variable coupling intervals (300–600 ms).

Afferent Nodose Neuronal Activity Recordings

To evaluate cardiac afferent neurotransmission, activity of individual nodose ganglia afferent neurons was recorded at baseline and during cardiac stimuli, including PVCs. Right and left nodose ganglia were exposed via a lateral neck cut down (Fig. 1A and Supplemental Fig. S1; all supplemental material is available at https://doi.org/10.6084/m9.figshare.8139512.v1).

Fig. 1.

Nodose ganglia in vivo neuronal recordings. A: anatomic locations of the nodose ganglion and carotid artery are shown. Protein gene product (PGP) 9.5 staining of the nodose ganglion was used to confirm location of the ganglion at the end of the experiments. B: customized 16-channel linear microelectrode array was used for individual neuronal recordings. C: representative sorted neuronal action potentials from the nodose ganglion of 1 animal. Representative action potentials from individual neurons are illustrated in the boxes. D: representative activities generated by 6 nodose neuronal soma in response to premature ventricular contractions (PVCs) in 1 animal. Increase in the activity of these nodose neurons (P < 0.05) can be observed during PVCs in this animal. CH, electrode channel; LVP, left ventricular pressure (mmHg).

Vagus nerve branches were kept intact during recordings. As no differences in the transduction capabilities of left versus right nodose ganglion have been previously observed (44), left (n = 3), right (n = 4), or bilateral (n = 9) nodose ganglion soma activity was recorded in vivo using custom-made 16-channel linear microelectrode arrays (LMA; Microprobes for Life Science, Gaithersburg, MD; Fig. 1B). The LMA (25-µm-diameter platinum/iridium electrodes, 16 electrodes/channels, 250-µm interelectrode distance) was connected to a 16-channel preamplifier (NeuroNexus, Ann Arbor, MI). Neural signals were sampled at 20 kHz (filtered at 300–10,000 Hz), amplified, and digitized (SmartBox acquisition system; NeuroNexus). If cardiac-related neuronal activity could not be identified, LMA position was adjusted slightly and neural activity was re-evaluated. At the end of each experiment, nodose ganglia were removed and histology was performed for confirmation of the location of the nodose as described above (Fig. 1A). Before removal of ganglia, ipsilateral vagus nerve stimulation was performed (1 Hz, 1 ms, 6 mA) at the end of the experiment to confirm backfiring (antidromic activation) of nodose ganglion neurons (Supplemental Fig. S2).

Cardiovascular Interventions

Interventions were performed in all animals in a random order. PVCs were induced as described above. To further characterize the nodose ganglion neurons identified, other cardiac interventions were performed. These included 1) anterior LV and RV mechanical stimulation and 2) chemical stimulation, 3) rapid RV pacing, 4) inferior vena cava (IVC) and then 5) aortic artery occlusion and 6) 1 min of left anterior descending (LAD) ischemia. Depending on the intervention, 10–30 min was allowed between stressors for hemodynamic indices to return to baseline.

Ventricular epicardial afferent inputs.

To assess the response of nodose ganglion afferent neurons to epicardial mechanical stimuli, gentle pressure (~10 g) was applied to the RV and LV anterior walls for 15 s via a saline-soaked-cotton-tipped applicator. Thereafter, chemical transduction of individual neurons was tested by 1-min application of a gauze soaked in adenosine (100 µM), bradykinin (10 µM), capsaicin (1 µM), or veratridine (10 µM) to the anterior RV and LV epicardium (same region that mechanical stimuli had been applied). Warmed saline was used to wash off the chemicals after each application.

Rapid ventricular pacing.

Rapid RV pacing was performed for 1 min at 15% above the baseline heart rate using the same pacing catheter used for PVCs. Neuronal responses to ventricular pacing were evaluated.

Great vessel occlusions.

Two segments of umbilical tape were placed around the IVC and descending aorta. IVC and then aortic occlusions were performed separately for 30 s to evaluate and compare neuronal responses to changes in preload and afterload, respectively.

Regional ventricular ischemia.

A silk suture was placed around the LAD coronary artery after its first diagonal branch. The LAD was occluded for 1 min to evaluate neural responses to coronary arterial occlusion. Ischemia was confirmed by ST elevation or T-wave inversion on ECG and decrease in LV systolic pressure (LVSP).

Neural Signal Processing and Analysis

Detailed spike identification of the activity generated by individual neurons was performed by assessing all neural channels. Simultaneous and similar waveforms present on all channels were identified as artifacts and removed. Neural waveforms with a signal-to-noise ratio ≥3 were identified (Fig. 1C). Neuronal activity classification was performed using principal component, cluster on measurements, and k-means clustering analysis using Spike2 (5, 16, 33). Time series of individual neuronal activity for the entire experiment was transferred to MATLAB (MathWorks, Natick, MA) for post hoc neural analysis.

If activity of a neuron changed significantly during at least one cardiovascular stressor (1-min baseline vs. during the intervention), then the neuron was considered to be a cardiac-related neuron (Fig. 1D). Neurons that did not respond to any cardiovascular interventions, and may be sensing other visceral organs, were excluded from the analysis. Chemosensitive cardiac neurons were defined as those that responded to at least one ventricular chemical stimulus and not to the epicardial mechanical stimuli. Mechanical neurons were defined as those that responded only to mechanical stimuli. Multimodal neurons were identified as those that responded to at least one mechanical and at least one chemical stimulus.

Statistical Analysis

Data are presented as means ± SE. Neuronal activity was compared in different time windows (1 min before vs. during intervention) by calculating the average neuronal activity. Significant differences in activity between time windows were compared based on the Skellam distribution (40). This test has been validated for neuronal activity of peripheral ganglia (5, 16, 32, 33) and the central nervous system neurons (40). A χ²-test was used to compare neuronal responses with different stressors. Wilcoxon signed‐rank test was used to compare paired data. One-way ANOVA was used to evaluate differences among groups. Adjustment for multiple comparisons was performed using Tukey multiple comparisons test. A P value ≤ 0.05 was considered statistically significant. Statistical analyses were performed using Prism (GraphPad Software, La Jolla, CA).

RESULTS

Hemodynamic Responses

Hemodynamic parameters at baseline and during interventions are shown in Table 1 and Supplemental Table S1.

Table 1.

Hemodynamic response to applied cardiac stressors (n = 14)

Stimulus	ΔHR, %	ΔLVSP, %	ΔdP/dtmax, %	ΔdP/dtmin, %
EMS	−0.3 ± 0.5	−2.5 ± 0.6	−1.4 ± 1.2	3.7 ± 1.3
ADENOS	4.8 ± 4.7	1.2 ± 1.6	−3.1 ± 2.3	−2.1 ± 1.8
BRADY	4.00 ± 3.4	−1.3 ± 3.0	6.2 ± 4.3	5.8 ± 4.2
CAPS	6.5 ± 6.7	−3.4 ± 4.0	65.2 ± 77.1	1.1 ± 4.1
VERAT	0.6 ± 2.00	−2.8 ± 3.1	−11.6 ± 4.4	6.2 ± 4.2
PVC	3.9 ± 1.7*	−3.8 ± 1.4†	3.7 ± 6.1	0.5 ± 4.5
RVP	13.7 ± 2.8†	−19.5 ± 5.7†	−17.6 ± 5.0†	26.7 ± 5.4†
IVC	3.8 ± 5.8	−52.5 ± 4.7†	−39.1 ± 7.3†	53.7 ± 7.2†
AO	−11.4 ± 3.0†	62.1 ± 7.5†	25.3 ± 13.1*	−50.2 ± 30.0†
LAD	1.3 ± 0.3†	−9.2 ± 1.0†	−11.2 ± 1.7†	19.5 ± 2.1†

Values are means ± SE for percentage change from baseline in heart rate (HR), left ventricular systolic (LVSP), as well as the maximum and minimum 1st derivatives of LV pressure (dP/dt). ADENOS, epicardial adenosine application; BRADY, epicardial bradykinin application; CAPS, epicardial capsaicin application; AO, aortic occlusion; EMS, epicardial mechanical stimulation; IVC, inferior vena cava occlusion; LAD, left anterior descending coronary artery occlusion; PVC, premature ventricular contractions; RVP, right ventricular pacing; VERAT, epicardial veratridine application.

^*P < 0.05 vs. baseline.

^†P < 0.01 vs. baseline. Values in boldface represent statistically significant changes from baseline.

Of note, during the PVC period, excluding the PVC beats, heart rate increased by 3.5 ± 1.6 beats/min (3.9 ± 1.7%; P < 0.05), whereas mean LVSP decreased by 4.1 ± 1.5 mmHg (3.8 ± 1.4%; P < 0.01). LVSP and dP/dt_max decreased in response to IVC occlusion by 58.5 ± 6.6 mmHg (52.5 ± 4.7%; P < 0.01) and 651.4 ± 153.4 mmHg/s (39.1 ± 7.3%; P < 0.01). Aortic occlusion increased both LVSP and dP/dt_max by 60.0 ± 6.7 mmHg (62.1 ± 7.5%; P < 0.01) and 182.3 ± 74.7 mmHg/s (25.3 ± 13.1%; P < 0.05), respectively. LAD occlusion also changed hemodynamic indices significantly by increasing dP/dt_min by 370.6 ± 62.6 mmHg/s (19.5 ± 2.1%; P < 0.01) and decreasing LVSP by 9.7 ± 1.3 mmHg (9.2 ± 1.0%; P < 0.01) and dP/dt_max by 159.5 ± 25.8 mmHg/s (11.2 ± 1.7%; P < 0.01).

Heart Rate Variability

HRV was analyzed before and after induction of PVCs. The heart rate in the minute before initiation of PVCs was 96.3 ± 4.6 beats/min. The heart rate was 97.4 ± 4.7 beats/min in the minute after cessation of PVCs. High-frequency component analysis of HRV showed that PVCs significantly decreased indices of parasympathetic tone [relative power of high frequency (HF): 78.6 ± 2.7 to 57.9 ± 6.8%, P < 0.05; normalized power of HF: 83.6 ± 2.8 to 72.5 ± 5.3, P < 0.05], suggesting parasympathetic withdrawal (Table 2). All time-domain parameters pre- versus post-PVCs showed a consistent trend for decrease in variability, consistent with a decrease in parasympathetic tone (Table 2), but did not reach statistical significance. This was likely due to the limited duration of the intervention as well as the short period of time that could be used for analysis pre- and postintervention.

Table 2.

Heart rate variability analysis

	HF, %	nHF	SDRR, ms	CVRR	SDSD, ms	RMSSD, ms	pRR50, %
Baseline	78.6 ± 2.7	83.6 ± 2.8	31.6 ± 16.8	0.05 ± 0.03	43.75 ± 24.0	43.5 ± 23.9	0.03 ± 0.02
Post-PVC	57.9 ± 6.8	72.5 ± 5.3	14.9 ± 8.2	0.02 ± 0.01	23.96 ± 13.0	24.3 ± 13.2	0.03 ± 0.02
P value	<0.05	<0.05	0.37	0.33	0.49	0.51	0.97

Values are means ± SE. Wilcoxon signed‐rank test was used to compare baseline vs. post-PVC periods; n = 10 animals. CVRR, coefficient of variance of RR intervals; HF, high-frequency component of heart rate variability; nHF, normalized HF; pRR50, proportion of the number of pairs of successive RR intervals that differ by >50 ms divided by all of the RR intervals given as a percentage; PVC, premature ventricular contraction; RMSSD, root mean square of the successive RR differences; SDRR, standard deviation of RR interval; SDSD, standard deviation of successive RR interval differences between adjacent RR intervals.

Neuronal Responses to Cardiac Stressors

Given evidence for withdrawal of parasympathetic efferent tone during PVCs, we evaluated whether PVCs affect vagal afferent neurotransmission, which can reflexly decrease central parasympathetic drive. Of 16 animals, cardiac-related neurons were meticulously identified in the nodose ganglia of 9 animals, as the nodose ganglia contain neurons from all visceral organs (26). Location of nodose ganglion was confirmed histologically in all 16 animals (Fig. 1A). Of the 212 nodose afferent neurons (13 ± 4 per animal) recorded, 89 cardiac-related neurons were identified based on their significant responses (P < 0.05) to cardiac stressors (Fig. 2A). Basal activity of cardiac afferent neurons was relatively low (0.19 ± 0.05 Hz). Of these 89 neurons, the greatest number (51%) responded to PVCs, whereas only 31% responded to LAD occlusion of similar duration (P < 0.05; Fig. 2B). In addition, 19 and 18% of these sensory neurons responded to IVC and aortic occlusions, respectively. Changes in percentage of neurons affected by other cardiovascular interventions, including RV pacing, are shown in Fig. 2B.

Fig. 2.

Response of neurons to cardiovascular stimuli. A: 89 cardiac afferent neurons (each row is an individual neuron) were identified based on their response to cardiovascular interventions, including PVCs. B: percentage of cardiac neurons that responded to each stressor is shown. PVCs engaged the highest number of afferent neurons. *P < 0.05 vs. other interventions (ANOVA). C: average change in firing rate of all nodose cardiac afferent neurons (n = 89). PVCs, left anterior descending (LAD) occlusion, and IVC occlusion caused the greatest change in firing rates of neurons. *P < 0.01 (ANOVA).

The absolute value of the change in firing rate from baseline for each intervention is shown in Fig. 2C. PVCs, LAD, and IVC occlusions caused the greatest change in the firing rates (PVCs: 0.29 ± 0.06 Hz; LAD: 0.24 ± 0.06 Hz; IVC occlusion: 0.19 ± 0.06 Hz; Fig. 2C). Cardiac sensory neuronal firing activity in response to PVCs and LAD was significantly higher (P < 0.01) than epicardial mechanical stimuli, aortic occlusion, RV pacing, and chemical stimulation.

Mechanosensitive Versus Chemosensitive Neural Responses

To further evaluate the type of neurons activated by PVCs, we focused on sensory neurons receiving input from the regions of the epicardial RV and LV anterior wall, where chemical interventions could be directly applied (Fig. 3A). Basal firing rate of different types of cardiac afferent neurons (mechanical vs. chemical vs. multimodal) was not significantly different. PVCs not only activated mechanosensitive neurons (42%), but also altered activity of chemosensitive (25%) and multimodal neurons (33%) innervating the RV and LV anterior epicardium (Fig. 3, B and C). No difference in basal activity or change in firing rate of mechanical sensory nodose neurons innervating the RV compared with the LV anterior wall was noted. Also, no significant difference between the number of neurons that responded to the various chemical stimuli was observed.

Fig. 3.

Type of afferent neurons modulated by premature ventricular contractions (PVCs). A: location of right ventricular (RV) and left ventricular (LV) application of mechanical and chemical stimuli is shown by the shaded area (white square). Mechanical and chemical stimuli were applied to the same receptive field. B: PVCs activated both mechanosensitive and chemosensitive neurons. C: percentages of mechanical, chemical, and multimodal neurons that responded to PVCs are shown. ADE, adenosine; BRA, bradykinin; C, chemical neurons; CAP, capsaicin; CHEM, chemical stimulation; EMS, epicardial mechanical stimulation; LAD, left anterior descending; M, mechanical neurons; MM, multimodal neurons; VER, veratridine.

Temporal Profile of Nodose Sensory Neuronal Activity in Response to PVCs

Analysis of the activity of nodose neurons that responded to PVCs demonstrated that after cessation of PVCs, their activity remained elevated and did not return to baseline for ≥1 min postcessation of PVCs (Fig. 4A), suggesting existence of neuronal memory. The first five PVCs caused the highest impact on cardiac neuronal activity (Fig. 4B). This activity then somewhat diminished (Fig. 4B), indicating receptor-mediated adaptation, given somewhat of a decrease in response to a constant stimulus (49).

Fig. 4.

Response of cardiac afferent neurons activated with premature ventricular contractions (PVCs). A: response of neurons that were activated during PVCs shows elevated firing rates even after cessation of PVCs (response evaluated over the minute following discontinuation of PVCs). B: response of neurons as a function of the PVC number during the 1-min period is shown. C: representative example of a PVC coupling interval and underlying sinus cycle length. Ratio of the coupling interval to the sinus cycle length was used to assess effect of early vs. late-coupled PVCs on neuronal activation. Representative sinus cycle length (569 ms, as shown) and PVC coupling interval (395 ms, as shown) are shown. D: response of neurons to PVCs as a function of the coupling interval to sinus rhythm ratio. Late-coupled PVCs (ratio of 80–90%) caused the greatest neuronal activity. *P < 0.05 (ANOVA). BL, baseline; CI, coupling interval; CL, cycle length; LVP, left ventricular pressure.

Effects of PVC Coupling Interval on Afferent Activity

To assess the effect of the PVC coupling intervals on neuronal activity, the ratio of the PVC coupling interval to the sinus cycle length was used to define early versus late coupled PVCs, as different sinus rates were observed in different animals (Fig. 4C). Increasing afferent neuronal activity was observed as the PVC coupling interval increased (Fig. 4D). Cardiac afferent neurons showed the highest activity during PVCs with the longest coupling interval (ratio of PVC coupling interval to sinus cycle length of 80–90%), suggesting that late-coupled PVCs exert the greatest effect on vagal afferent neural transduction.

DISCUSSION

To our knowledge, this is the first study to evaluate effects of PVCs on cardiac vagal afferent neurotransmission using direct recordings of cardiac nodose sensory neurons in vivo. The major findings of this study are 1) PVCs are powerful modulators of cardiac vagal afferent neurotransmission, exerting even greater effects than ventricular ischemia of similar duration, significantly increasing sensory neural inputs to the brain stem, 2) PVCs activate mechanosensory as well as chemosensory neurons in the inferior vagal ganglia, and 3) nodose ganglion sensory neurons are capable of state-dependent adaptations and display memory in response to cardiac stimuli. Because of this capacity, any excitation of this population of cardiac afferent neurons persists following the initial insult, further exacerbating the pathology of cardiac arrhythmias and cardiomyopathy.

PVC Induced Parasympathetic Efferent Withdrawal

Previous studies of patients with PVCs using HRV analysis have shown existence of sympathetic activation based on an increase in the low-frequency component (3). PVCs have also been shown to increase muscle sympathetic nerve activity and activate neurons in the stellate ganglia (10, 15, 26, 50). However, the effects of PVCs on the parasympathetic nervous system remained unclear. HRV data have shown controversial results, with some studies demonstrating a decrease and others showing no clear changes in HF component of HRV in the presence of frequent PVCs (3, 11, 18, 39).

As sympathetic efferent activation and parasympathetic efferent dysfunction often act in concert to increase progression of cardiomyopathy, we hypothesized that PVCs may lead to a low parasympathetic tone due to activation of cardiac sensory neurons in the nodose ganglia that provide input to the nucleus tractus solitarius of the medulla. To confirm that parasympathetic tone was indeed decreased, we measured HRV before and immediately after cessation of PVCs to avoid the confounding factor of presence of PVCs on HRV analysis that may have complicated results of other studies. We found that, indeed, the HF component decreased, suggesting with parasympathetic withdrawal. As such, we hypothesized that afferent parasympathetic activation contributes to parasympathetic efferent withdrawal associated with PVCs.

Sensory Transduction of PVCs by the Nodose Ganglia

The nodose ganglia provide sensory innervation from the lungs, gastrointestinal tract, and other visceral organs. Only a small percentage of these neurons are cardiovascular (26, 32). Therefore, it is not unexpected that in some animals, despite recording multiple neurons, we could not identify neurons that responded to cardiovascular interventions.

A previous study suggested that LAD occlusion is, in general, a very potent modulator of cardiac afferent neurons (32). In this study, we were surprised to find that variably coupled PVCs engaged even more cardiac-related neurons than LAD occlusion, IVC occlusion (which decreased cardiac preload by 50–60 mmHg), or aortic occlusion (which increased cardiac afterload by >60 mmHg). Variably coupled PVCs also activated more sensory neurons than rapid RV pacing, consistent with a previous porcine study demonstrating that chronic PVCs are more likely to cause cardiomyopathy than rapid RV pacing at 140 beats/min (47). Our data suggest that even brief periods of PVCs are important stressors and modulators of the parasympathetic nervous system.

Sensory Transduction of Mechanical and Chemical Milieu

A major strength of our study is evaluation and characterization of individual cardiovascular neuronal types rather than measurement of global vagal nerve activity, which reflects the average effects of afferent and efferent neurotransmission. Cardiac afferent neurons in the nodose ganglia have been shown to be mechanosensitive, chemosensitive, or capable of transducing both types of stimuli (multimodal neurons) (2, 44). We hypothesized that PVCs, in altering mechanical stretch and synchrony (47), would likely activate mechanosensitive neurons. We were surprised to find that PVCs also activated chemosensitive neurons. These data suggest that PVCs may cause changes in supply/demand that may lead to release of reactive oxygen species and activate chemosensitive neurons. Arrhythmias, including PVCs, can be associated with release of cellular reactive oxygen species by altering both ion channel (calcium channel) and mitochondrial function (22, 27, 48).

Of note, 33 (37%) of cardiac afferent neurons responded to PVCs without responding to any of the other RV and LV epicardial mechanical or chemical stimuli. This might be due to the fact that 1) the sensory axons of these neurons are in the dorsal/posterior aspect of the heart or on the endocardium, which would not be affected by our local anterior RV and LV applied mechanical or chemical cardiac stimuli or 2) these neurons may sense other molecules/chemicals released during PVCs not tested by our chemical interventions.

Temporal Profile of Neuronal Activity During PVCs

Cardiac sensory nodose neurons were maximally activated immediately after the initiation of PVCs and specifically during the period encompassing the first five PVCs. After the first five PVCs, these afferent neurons remained active but at a lower firing rate. These data suggest that neurons of the nodose ganglia adapt during an ongoing pathology. It is important to note, however, that activity of these neurons remained elevated compared with baseline for ≥1 min postcessation of PVCs. These data are consistent with our HRV data indicating that withdrawal of cardiac parasympathetic efferent tone persisted after arrhythmia termination.

Effects of PVC Coupling Interval on Afferent Activity

It has been shown that PVCs with the longer coupling interval cause more pronounced LV dyssynchrony, and this may be a more important variable than even location of PVCs in causing heart failure (31). In this study, PVCs with the longest coupling interval caused a greater activation of cardiac sensory neurons than shorter coupling interval PVCs. This could potentially be due to the more pronounced LV dyssynchrony observed in previous studies with longer coupled PVCs (31).

Clinical Implications

PVCs are routinely observed in many patients with an estimated prevalence of 1–4% (23). However, several studies have shown that a high burden of PVCs is associated with development of left ventricular dysfunction and heart failure over time (4, 12, 29). The extent of LV dysfunction also correlates with frequency of PVCs. Currently, there are no clear cut-off points that delineate the PVC burden at which cardiomyopathy may develop (4, 6, 17). However, several studies have suggested that a frequency >10%, and especially >24%, is associated with development of cardiomyopathy (4, 17, 37). In line with these studies, we chose to deliver PVCs at an average burden of every 5 beats (~20%). Of note, most patients with PVC-induced heart failure do not have scar on cardiac imaging, suggesting an important role for the autonomic nervous system in development of this disease. Treatment of PVCs, including with medications that reduce burden of PVCs or catheter ablation of PVCs, results in improvement of LV function and heart failure (4, 14, 17, 24, 35, 42). By evaluating both the afferent and efferent limbs of the cardiac parasympathetic nervous system, neural data from this study provide supportive evidence for powerful modulation of vagal neurotransmission by even brief periods of PVCs. Effects of PVCs were greater than LAD ischemia or burst RV pacing of similar duration. As these effects persist following the cessation of PVCs, it appears that this neural transduction involves memory. These data have implications for autonomic imbalances that can be caused by PVCs and can contribute to development of cardiomyopathy. Furthermore, given the important role of the autonomic nervous system in development of PVC-induced cardiomyopathy, it is possible that in patients in whom ablation or medication fail to treat PVCs, autonomic modulation can provide an additional avenue for therapy. Consistent with this, bilateral stellate ganglion blockade and cardiac sympathetic denervation have been reportedly used in case series of patients to treat refractory PVCs and ventricular arrhythmias (9, 45).

Limitations

The basal activity of nodose ganglion neurons measured in this study was relatively low, which may reflect effects of anesthesia. To minimize effects of isoflurane, neural recordings were obtained after anesthesia was switched to α-chloralose. Because inhaled isoflurane can suppress neuronal responses to cardiovascular interventions, the results of this study may represent a conservative estimate of the effects of PVCs on parasympathetic function. Given the acute nature of these studies, the effects of PVCs were evaluated over short durations. It is possible that long-term autonomic effects of PVCs in development of cardiomyopathy may not parallel those seen in this study. However, given the powerful modulatory effect of PVCs on neural activity observed in this study and the persistence of these effects beyond cessation of PVCs, it would be expected that chronic frequent PVCs would continue to modulate afferent neuronal neurotransmission and reflexly decrease cardiac parasympathetic efferent tone. As HRV is noninvasive, it has been frequently used to assess autonomic tone in clinical studies and has been associated with increased risk of death in setting of heart failure (30); however, HRV have can have significant limitations, especially in assessing situations where heart rate changes. A decrease in HRV can be driven by an increase in heart rate (7, 8). In this study, there was a small increase in heart rate from 96 to 97 beats/min when comparing the pre-PVC with the post-PVC period, and it is possible that a small portion of the HRV changes are driven by this mild increase in heart rate. Parameters in this study were statistically significant in the frequency domain but not the time domain. This may be due to the short duration of the intervention and period for pre- and post-PVC time analysis as well as lack of large numbers of animals. Finally, as the primary goal of this study was to compare the effect of PVCs on afferent neurotransmission and parasympathetic efferent outflow compared primarily with LAD ischemia, effects of premature atrial contractions and various locations of PVCs were not assessed. However, evaluation of these comparisons remains an important part of future studies in our laboratory.

Conclusions

PVCs cause vagal afferent activation and decrease parasympathetic efferent tone to the heart. As such, even brief periods of PVCs act as significant modulators of cardiac control, continuing to exert a powerful influence on neurotransmission even after the event and more than ventricular ischemia.

GRANTS

This work was supported by National Institutes of Health Grants DP2-HL-132356 and SPARC OT2OD023848 to M. Vaseghi.

DISCLOSURES

M. Vaseghi and J. L. Ardell have founder shares in Neurcures, Inc. University of California, Los Angeles, has patents developed by M. Vaseghi and J. L. Ardell relating to cardiac neural diagnostics.

AUTHOR CONTRIBUTIONS

S.S., J.L.A., J.A.A., and M.V. conceived and designed research; S.S., N.Y., J.H., S.P., and M.V. performed experiments; S.S., N.L., S.P., and M.V. analyzed data; S.S., J.L.A., J.A.A., and M.V. interpreted results of experiments; S.S. and M.V. prepared figures; S.S. and M.V. drafted manuscript; S.S., N.Y., J.H., N.L., S.P., J.L.A., J.A.A., and M.V. edited, revised, and approved final version of manuscript.