Cardiac TRPV1 afferent signaling promotes arrhythmogenic ventricular remodeling after myocardial infarction

Epicardial RTX application induces degeneration of TRPV1 afferent fibers.

To target cardiac-specific TRPV1-expressing afferent fibers (Figure 1A), we utilized a percutaneous approach to transiently apply RTX (12.5 μg/mL; Supplemental Methods) to the epicardium in MI subjects. We previously established the feasibility of this approach in normal porcine (12). Four weeks later, depletion of calcitonin gene–related peptide (CGRP), a marker of afferent fibers that colocalize with TRPV1 (13), was confirmed by epicardial IHC (Figure 1, B and C). Adrenergic fibers labeled by tyrosine hydroxylase (TH) immunoreactivity were not affected. Depletion of TRPV1 afferent fibers was also demonstrated functionally using TRPV1 agonists capsaicin (Figure 1, D and E) and bradykinin (Figure 1F). These activators normally result in robust adrenergic-induced hemodynamic responses but were abolished following RTX treatment. Additionally, immunohistochemical studies of dorsal root ganglia (DRG) from MI + RTX compared with control and MI subjects demonstrated regions of axon loss but no differences in DRG soma size (Supplemental Figure 1, A–C; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.124477DS1).

Given the small diameter of C- and Aδ-axons that characterize TRPV1 afferents (14), selective loss of TRPV1 neurons would be expected to bias the distribution of DRG soma size toward larger neurons. The absence of this finding, in the presence of loss of axons, indicates that epicardial RTX application caused axonal degeneration of TRPV1 afferents and not significant elimination of sensory afferent soma in the DRG. Both chemical (TRPV1-expressing) and multimodal (mechanical and chemical) cardiac afferents were likely impacted by RTX, given the expression of TRPV1 in these fibers.

Cardiac TRPV1 afferent depletion reduces VAs after MI.

To determine whether depletion of cardiac TRPV1 afferents reduces VAs after MI, we performed programmed extrastimulation (PES) from a right ventricular endocardial catheter 4 weeks after anteroseptal infarct in MI and MI + RTX subjects. Ventricular tachycardia (VT) or ventricular fibrillation (VF) was induced in 12 of 13 MI subjects (92%) but only in 1 of 7 MI + RTX subjects (14%) (P = 0.005; Figure 2, A and B). The ease of inducibility (i.e., degree of aggressiveness of the PES protocol) was quantified as the arrhythmogenicity index (AI), which was higher when fewer ventricular extrastimuli were required for induction. MI subjects compared with MI + RTX showed greater AI (Figure 2, C–E).

Arrhythmogenesis was also assessed by bolus i.v. administration of cesium chloride (CsCl), a nonspecific potassium channel antagonist that transiently elicits afterdepolarization-induced triggered activity (15) (Supplemental Methods). Premature ventricular contractions (PVCs), as well as the number and length of nonsustained VT (NSVT) episodes, were reduced in MI + RTX (Figure 2, F–I) compared with MI.

In summary, these findings show that VAs, assessed by independent provocations, are significantly reduced by depletion of cardiac TRPV1 afferents.

TRPV1 afferent depletion does not mediate its antiarrhythmic effects by altering cardiomyocyte electrophysiology.

Next, to determine the mechanisms by which cardiac TRPV1 afferent depletion reduced VAs after MI, we examined the expression of ion channels causative of VAs (16) in the scar BZ, where postinfarct VAs typically originate (17). Expression of SCN5A (I_NA), KCND2 (I_to), HERG (I_kr), KCNQ1 (I_ks), and KCNJ4 (I_k1) was not different (Supplemental Figure 2A) among control, MI, and MI + RTX subjects. Additionally, since β-adrenergic signaling modulates ion channel function, we examined and found no differences in β₁AR mRNA expression, protein levels, or functional responses to sympathetic stimulation (Supplemental Figure 2B) elicited by right stellate ganglion stimulation (RSGS).

Although expression of ion channels was not different, alterations in ion channel function by cardiac TRPV1 afferent depletion may underlie the protection from VAs. To examine this possibility, we recorded action potentials (APs) from scar BZ cardiomyocytes isolated from MI and MI + RTX subjects. We confirmed that isolated cells from MI and MI + RTX subjects exhibited no morphological differences, since RTX was applied to the epicardium directly (Supplemental Figure 3, A and B). The APs of BZ cardiomyocytes from MI and MI + RTX subjects showed no differences in resting membrane potential, AP amplitude, upstroke velocity (dV/dt_max), and duration (AP duration-90 [APD₉₀] and APD₅₀) (Figure 3, A–F).

To determine whether calcium handling, an accepted mechanism for arrhythmogenesis when dysfunctional (18), was involved in the reduced arrhythmogenic potential observed in MI + RTX animals, we investigated Ca²⁺ transients in cardiomyocytes loaded with the Ca²⁺-sensitive indicator Fluo-4. We found no significant differences in the kinetics of calcium release (time to peak) and reuptake (half maximal recovery [τ₅₀]), as shown in Figure 3, G–I. Similarly, there were no differences in expression of the L-type calcium channel or other calcium handling proteins, as determined by RNA sequencing (RNA-seq) (Supplemental Figure 4A) and quantitative PCR (qPCR) (Supplemental Figure 4B).

These data indicate that modulation of cardiomyocyte AP and calcium handling do not underlie the antiarrhythmic protection conferred by cardiac TRPV1 afferent depletion following MI. It is possible that TRPV1 afferent depletion is associated with changes in the functional regulation of ion channels or calcium handling proteins; however, if these changes exist, they do not appear to translate to the cardiomyocyte AP or Ca²⁺ transient properties.

BZ electrical heterogeneity, fibrosis, and altered connexin distribution are attenuated by TRPV1 afferent depletion.

The critical dependence of VAs on structural abnormalities that permit slowed and heterogeneous electrical propagation in the BZ of infarcted hearts is well established (17). To determine whether chronic TRPV1 afferent depletion exerts its protective effects by ameliorating cardiac structural abnormalities following MI, we quantified fibrosis, a determinant of electrical propagation, and performed high-resolution cardiac electrical mapping of the BZ in MI and MI + RTX subjects.

Trichrome staining of BZ myocardium demonstrated a substantial reduction in transmural fibrosis and wall thinning in MI + RTX versus MI subjects (Figure 4, A and B) in the scar BZ but not in remote myocardium. Next, we turned to RNA-seq of myocardium from the scar BZ to profile the expression of key proteins known to underlie collagen deposition following MI. As shown in Figure 4C, several genes involved in the regulation of collagen deposition after MI were differentially expressed in MI + RTX myocardium compared with MI, suggesting that several pathways key to infarct-related remodeling are impacted by cardiac TRPV1 afferent depletion.

To examine whether the reduction in fibrosis was of functional significance, we performed electrical mapping of the scar BZ as previously published (19). These studies demonstrated that electrical activation in scar BZ appeared slower and more heterogeneous in MI subjects, compared with MI + RTX. Specifically, mean activation time, activation dispersion, activation delay (i.e., duration of propagation from the earliest to latest region of the mapped area), and maximal activation time (i.e., longest activation time recorded in the mapped region) were improved by TRPV1 afferent depletion in MI + RTX subjects (Figure 4, D and E). Interestingly, RTX treatment had no significant impact on repolarization properties of the BZ tissue. Epicardial activation recovery intervals (ARIs, a surrogate for APD) from the scar BZ (Supplemental Figure 5, A and B) were not different between the 2 groups (consistent with AP data shown earlier in Figure 3, A–F), nor were ARI dispersion and ARI restitution properties of MI and MI + RTX subjects (Supplemental Figure 5, C and D).

Connexin 43 (Cx43) distribution in the scar BZ is known to be reduced and lateralized following MI, and it is associated with VAs (20). We examined whether TRPV1 depletion also attenuated Cx43 remodeling in its reduction of postinfarct VAs. We observed restored Cx43 distribution in MI + RTX subjects in the scar BZ, while no significant differences were identified in myocardium remote to the infarcted region (Supplemental Figure 6).

In summary, these data suggest that cardiac TRPV1 afferent depletion exerts its antiarrhythmic effects by attenuating scar BZ fibrosis, Cx43 distribution, and activation heterogeneity at the tissue level — not by altering cellular AP or tissue repolarization properties.

Depletion of cardiac TRPV1 afferents preserves global cardiac function after MI.

We sought to determine whether the structural changes mitigated by TRPV1 afferent depletion at the time of MI improved cardiac function as determined by echocardiography. MI resulted in reduced left ventricular ejection fraction (LVEF) and increased LV end-systolic volume (LVESV) and internal diameter in diastole (LVIDd), as well as LV wall thinning; however, treatment with RTX at the time of MI attenuated these changes (Figure 5, A–E).

Chronic cardiac TRPV1 afferent signaling promotes cardiac hyperdynamic adrenergic dysfunction.

Since enhanced cardiac adrenergic tone results from cardiac TRPV1 afferent activation and is a powerful modulator of cardiac electrical function, we sought to examine whether sympathetic neuronal dysfunction (known to occur following MI) was the link between TRPV1 afferent activation and adverse BZ remodeling. This question would not be answered by βAR or calcium channel blockade because, independent of arrhythmia mechanism, βAR and calcium channel blockers are established antiarrhythmic agents. We turned to extracellular neuronal recordings of sympathetic neurons located in the cardiac pole of the stellate ganglion, the major nexus point for efferent adrenergic outflow to the heart. We performed such recordings of stellate ganglion neurons (SGNs) in control, MI, and MI + RTX animals. Filtered recordings were subjected to artifact removal and spike sorting using principal component analysis (PCA) and cluster on measurement analysis (Figure 6A). This identified the activity of individual neurons, which are designated as unique colors. A few examples are shown in Figure 6A. These approaches are detailed in the Supplemental Methods.

Basal SGN activity was greater in MI compared with control subjects (Figure 6B). Additionally, the proportion of neurons exhibiting spontaneous activity was greater in MI compared with control subjects (Figure 6C), indicating higher sympathetic tone after MI. Both indices of enhanced adrenergic tone were attenuated by cardiac TRPV1 afferent depletion (Figure 6, B and C). The responses of SGNs to cardiovascular stimuli were used to characterize reflex processing by SGNs (Figure 6D). SGN responses to cardiac mechanical stress (rapid ventricular pacing; RVP), afterload increase (aortic occlusion; Ao Occ), and preload reduction (inferior vena cava occlusion; IVC Occ) caused paradoxical changes in SGNs from MI compared with control subjects. TRPV1 afferent depletion in MI subjects (MI + RTX) largely restored these response patterns to those seen in control subjects. Conditional probability analysis (21, 22) was used to assess the degree of coordinated network function among cardiac-related SGNs to RVP, IVC Occ, and Ao Occ. It revealed a diminution in network processing capabilities of MI subjects compared with controls, and this was similarly normalized by cardiac TRPV1 afferent depletion (Figure 6E). These data indicate enhanced but uncoordinated reflex function of SGNs following MI; however, these adverse neural network processes were prevented by TRPV1 afferent depletion.

Study limitations.

TRPV1 channels have been reported in cardiomyocytes, although the intracellular location remains controversial (27, 28). It is unlikely that myocardial TRPV1 modulation influenced our findings because (a) experiments were performed 4 weeks after RTX was applied and (b) the expected effect of RTX activation of the TRPV1 channel is an irreversible calcium influx, which would result in cell death. This did not occur, as verified by histology, LV function (assessed by echocardiography), and ex vivo high-resolution gadolinium–enhanced cardiac MRI (1 mm voxel resolution) in normal hearts. Toxic irreversible activation of TRPV1 in cardiomyocytes would also worsen LV function after MI, instead of improving it, as shown in the present study (Supplemental Figure 7).

Delivery of RTX required percutaneous epicardial access, which may limit clinical translation of this approach. However, percutaneous epicardial access is increasingly used for ablation and left atrial appendage closure. Genetic ablation of TRPV1 may further clarify our findings; however, such models do not exist in large animals. Although porcine subjects more closely recapitulate human pathophysiology than murine models, some caution should be exercised in extrapolating these findings to humans, given some resistance to βAR blockers in preventing VT/VF during acute ischemia in 1 porcine study (29).

Conclusions.

These data in composite demonstrate that epicardial RTX application in the early ischemic period is a feasible method to deplete TRPV1 sensory afferents. These afferent fibers chronically drive sympathoexcitation following MI, promoting adverse myocardial remodeling, electrical heterogeneity, and VT/VF risk (Supplemental Figure 7). Cardiac-restricted targeting of chronic CSAR ameliorates these pathologic processes and renders infarcted hearts more stable and resistant to VAs. This implicates TRPV1 afferents as key targets in reducing post-MI cardiac dysfunction and arrhythmias, to which novel minimally or noninvasive approaches can be developed.

Animal models and protocols.

Yorkshire pigs (n = 39) with normal hearts (control; n = 11; 50.89 ± 1.66 kg), and chronic anterior MI (n = 16; 52.04 ± 1.91 kg) were compared with RTX-treated pigs with chronic anterior MI (MI + RTX; n = 12; 50.62 ± 3.87 kg) (S&S Farms). Each subject was sedated with terconazole (4–8 mg/kg, intramuscular), intubated, and placed under general anesthesia. Inhaled isoflurane (2%) with supplemental oxygen (2 L/minute) during the experimental study. Surface electrocardiogram (ECG) and arterial blood pressure (ABP) were continuously monitored.

RTX administration.

Isoflupredone acetate (0.04 mg/kg) and famotidine (0.5 mg/kg) were given intramuscularly 24 hours prior to RTX administration. The left sternocostal angle was anesthetized with 1% lidocaine before making a 0.5–1.0 cm incision. Percutaneous access to the intrapericardial space was obtained using a Tuohy needle under fluoroscopic observation (14). A small amount of contrast agent was used to ensure the advancement of the needle tip, and intrapericardial access was confirmed by cardiac ectopy and tension before entering the pericardial space. Access was confirmed by a short J-tip guidewire taking a cardiac border–forming configuration fluoroscopically. The incision was progressively dilated with 6- to 8-Fr dilators. A multipurpose sheath was then advanced over the J-tip wire in order to aspirate pericardial fluid and administer 10 mL of RTX (12.5 μg/mL). Twenty minutes later, RTX was aspirated, and methylprednisolone sodium succinate (50 mg) was injected to prevent pericardial adhesions. Buprenorphine (0.02 mg/kg) was administered intramuscularly starting 6 hours after the procedure and then every 24 hours for 2 days. Carprofen (4 mg/kg) and famotidine (20 mg/day) were also orally given every 24 hours for 3 days after procedure. In sham subjects, the protocol with repeated identically, with saline vehicle infused instead.

Isolation of BZ ventricular myocytes.

Following euthanasia, BZ myocardium (adjacent but distinct from dense scar) was rapidly biopsied and 0.8 cm³ samples were placed in storage solution as described by Workman et al. (30). We then performed collagenase/protease isolation of ventricular myocytes as described (31). Myocytes were stored at room temperature in Kraft-Bruhe solution (pH 7.4) until use within 4 hours of isolation.

AP recordings.

For AP recordings, we placed isolated cardiomyocytes in an experimental chamber (0.5 mL) mounted on the stage of a Nikon Diaphot inverted microscope modified for simultaneous patch clamping and cellular epifluorescence recording. A modified Tyrode bath solution (in mmol/L: NaCl 136, KCl 5.4, NaH₂PO₄ 0.33, MgCl₂ 1.0, HEPES 10, CaCl₂ 1.8, glucose 10, pH 7.4) continuously perfused the chamber. Patch electrodes were pulled from borosilicate glass (TW150F-3, World Precision Instruments) on a horizontal puller (P-97, Sutter Instruments). The fire-polished electrodes had a tip resistance of 1–2 MΩ. The internal solution contained (in mmol/L) NaCl 10, KCl 130, HEPES 10, MgATP 5, phosphocreatine 5, cAMP 0.05, and MgCl₂ 1, pH adjusted to 7.2 with KOH. We used an Axopatch 200B patch-clamp amplifier (Molecular Devices) under the control of pCLAMP 10 software (Molecular Devices) and a Digidata 1440A PC interface (Molecular Devices) to induce and record APs in whole-cell current clamp mode.

Measurement of Ca²⁺ transients.

We loaded isolated cardiomyocytes with the Ca²⁺ indicator Fluo-4 by incubating cells in bath solution containing Fluo-4 AM (10 μM) and Pluronic F-127 (0.02%) for 30 minutes, followed by 3 washes in dye-free bath solution for 5 minutes each. The loading and washout times were sufficient for deesterification of the dye. We then recorded Fluo-4 fluorescence signals using a photodiode during simultaneous current clamp with a custom-designed photometric epifluorescence detection system attached to the patch clamp microscope described above. We used a 488 nm LED light source for excitation and a longpass filter to restrict emission to > 510 nm. Fluorescence ratios were normalized to baseline (F/F₀) after background subtraction.

Statistics.

Variables are presented as means ± SEM. The Shapiro-Wilk test was used for assessing distribution. A 2-tailed Student’s t test and 2-way ANOVA test were used for data that were normally distributed, and the Mann-Whitney U test and Kruskal-Wallis test were used for data that were not normally distributed. VT inducibility in MI vs MI + RTX subjects were analyzed using the χ² test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. Analysis was performed using Microsoft Excel and Graphpad Prism.

Study approval.

Studies were performed in accordance with the University of California Institutional Animal Care and Use Committee (IACUC) guideline and the NIH Guide for the Care and Use of Laboratory Animals (National Academies Press, 2011).

Version 1. 12/17/2019

In-Press Preview

Version 2. 02/13/2020

Electronic publication