Yorkshire pigs (n = 23, 11 males and 12 females), weighing 45 ± 1 kg, were included in the study. Animal preparations were conducted as previously reported (Howard-Quijano et al., 2017). Animals were sedated with Telazol (4 mg/kg, intramuscular) and Xylazine (2 mg/kg, intramuscular), underwent tracheal intubation, and were mechanically ventilated and general anesthesia was maintained with inhaled isoflurane (1 to 3%) during surgical preparation. A Prucka CardioLab recording system (GE Healthcare, Fairfield, CT, USA) was used for Heart rate (HR) and surface electrocardiogram (ECG) monitoring. The carotid and femoral arteries were catheterized for blood pressure monitoring and Millar catheter insertion (SPR-350). In addition, jugular and femoral veins were cannulated for saline infusion (10 ml/kg/hr) and drug administration. Arterial blood gas was checked every hour with adjustment of ventilation if needed to maintain acid-base equilibrium. Body temperature was maintained by external warmers. Dorsal spinal laminectomy was done in prone position to expose the spinal cord and left dorsal root of T2, median sternotomy was then performed supine to expose the heart, and for terminal experiment animals were in left lateral position for DRGS and cardiac ischemia induction. After the completion of surgical preparation, general anesthesia was transitioned to intravenous alpha-chloralose (50 mg/kg initial bolus followed by a 20 mg/kg/hr continuous infusion) for less impact on cardiac sympathetic excitability as we previously described (Dale et al., 2020; Omura et al., 2021). Depth of anesthesia was assessed throughout the experiments by monitoring corneal reflexes, jaw tone, and hemodynamic indices. At experiment end, animals were euthanized by inducing ventricular fibrillation via injection of potassium chloride under deep anesthesia.

Myocardial ischemia was performed as previously described (Howard-Quijano et al., 2017). A prolene suture was placed around the left anterior descending coronary artery (LAD) below the second diagonal branch of the LAD. The suture was ligated with a short polyethylene tubing segment to create myocardial ischemia for 1 h. Ischemia was confirmed by the presence of ST-segment elevations.

A 4-contact Axium™ DRG lead (St. Jude Medical, Minneapolis, MN, USA) was placed in the epidural space of the left DRG along the root of the second thoracic spinal level (T2), without removing the bone that covers the DRG. The stimulating electrode was enclosed in the body of intervertebral canal and ligaments which creates stabilization for the DRG lead (Figure 1A), similar to the clinical settings (van Velsen et al., 2018; Esposito et al., 2019). This stimulating lead was placed only in the DRGS group, with no placement in the CONTROL group. The T2 level was selected for DRGS as multiple cardiac afferent nerve fibers travel to the thoracic spinal cords through the C6-T6 DRGs bilaterally, with the largest numbers to the T2–T4 (Hopkins and Armour, 1989), and previous work has shown that cardiac afferents, as well as efferent output, are distributed symmetrically between the left and right sides (Yamakawa et al., 2016; Huang et al., 2017; Akgul Caglar et al., 2021) and thoracic DRGS at one level still gave cardioprotection (Kuwabara et al., 2021).

An A-M systems stimulator (Model 2100, A-M Systems, Sequim, WA, USA) was used to stimulate DRG. Motor threshold (MT) was assessed by increasing stimulus current intensity until muscle contractions were observed in the left shoulder, using a frequency of 2 Hz and pulse width of 0.4 ms. The average MT was 0.24 ± 0.03 mA in this study. Left T2 DRG was stimulated at frequency of 1 kHz, 0.03 ms pulse width, and current amplitude of 90% of MT.

The timeline of the experimental protocol is shown in Figure 1A. Twenty-three animals were randomized to two groups; LAD ischemia (CONTROL, n = 13) and LAD + dorsal root ganglion stimulation (DRGS, n = 10). After finishing the surgical preparation, the animal was positioned in the left lateral decubitus and a DRGS catheter was placed. Animals in both the CONTROL and DRGS groups were subjected to myocardial ischemia, but only the DRGS group received catheter-based neuromodulation.

A 56-electrode mesh sock electrode system was placed around the heart and unipolar electrograms (0.05–500 Hz) were measured using a Prucka CardioLab electrophysiology mapping system (GE Healthcare, Fairfield, CT, USA) (Figure 1B) (Kuwabara et al., 2021). All physiological measures were recorded at- (1) baseline (BL), (2) during 30-min of DRGS and for 30-min in the control (No-DRGS group), and (3) during 1-h of acute myocardial ischemia (or until releasing the ligation). We assessed activation recovery interval (ARI), which has been shown as a surrogate of local action potential duration (Millar et al., 1985; Vaseghi et al., 2012). ARIs were calculated with customized software (iScalDyn, University of Utah, Salt Lake City, UT, USA) (Figure 1C), as previously described (Howard-Quijano et al., 2017). A 2-dimensional polar map was created with ARI data from the sock electrodes by using publicly available software (Map3d; Scientific Computing and Imaging Institute, University of Utah, Salt Lake City, UT, USA). Sympathetic stimulation is associated with shortened ARI duration and increased dispersion of repolarization (DOR) (Yamakawa et al., 2016). Raw ARIs analyzed by the mapping system are affected by HR. In this study, ARIs were corrected by HR using Bazett’s formula (Dahlberg et al., 2021). Ischemic regions were defined as the electrodes where ST elevations were seen. To ensure accuracy of ARI measurement, each electrogram with ST segment changes was both measured by semiautomated accepted software and then checked by hand following the guidelines described by Haws and Lux for ARI measurement in ischemia (Haws and Lux, 1990). Global epicardial DOR was analyzed using the variance of all mean ARIs. Epicardial dispersion is associated with a heterogeneity of repolarization time and increased risk for ventricular arrhythmias (Vaseghi et al., 2012).

To compare the electrophysiological changes during ischemia, the time point before the ischemia induction (30 min of No-DRGS or DRGS) was compared.

Tp-Te interval was analyzed from the clearest limb lead in each animal. Tp-Te interval is an important marker for arrhythmogenicity in acute myocardial ischemia (Ozbek and Sokmen, 2019) and is also an independent risk predictor for sudden cardiac death (Yagishita et al., 2015). Tp-Te interval was assessed in limb leads with the clearest T wave at 200 mm/s paper speed as the average of five consecutive beats. Tp-Te interval was measured from the maximal T wave to the end of the T wave. The peak of the T wave was visually determined, and the end of the T wave was defined as the intersection of the tangent to the slope of the T wave and the isoelectric line (Figure 3A). Tp-Te interval was compared right before ischemia and at 30 min after ischemia onset to evaluate the arrhythmogenesis between the groups.

Ventricular arrhythmias, which include premature ventricular contractions (PVCs), ventricular tachycardia (VT), and ventricular fibrillation (VF) episodes, were counted using Prucka CardioLab system. VT was classified as three or more consecutive PVCs and terminating spontaneously in less than 30 s (Figure 3C; European Heart Rhythm Association et al., 2006). A ventricular arrhythmia score (VAS) was calculated for each animal during myocardial ischemia according to the criteria previously reported (Curtis and Walker, 1988; Howard-Quijano et al., 2021). The scoring system assigns a numeric value based upon the severity of arrhythmia with larger scores demonstrating greater physiological severity. In brief, the VAS was characterized as: “0: No PVCs, VT, or VF,” “1: PVCs,” “2: 1–5 episodes of VT,” “3: > 5 episodes of VT or 1 episode of VF,” “4: 2–5 episodes of VF,” “5: > 5 episodes of VF.”

Hemodynamic measures were recorded using a five French Millar Mikro-Tip pressure transducer catheter (SPR-350) inserted into the left ventricle via the left carotid artery with continuous data recorded in an MPVS Ultra Pressure-Volume Loop System (Millar Instruments, Houston, TX, USA) throughout the experiment. Left ventricular systolic function was evaluated by the maximum rate of pressure change (dP/dt max). ECG was continuously recorded on Prucka CardioLab system (GE Healthcare, Fairfield, CT, USA).

Evans Blue and triphenyl tetrazolium chloride (TTC) were used for the heart staining as previously shown (Howard-Quijano et al., 2021). At the very end of the experiment after the animal was sacrificed, the LAD ligation was tied again and a cross clamp was placed on the aorta just above the base of the heart with the animal in the supine position. We made sure that the aorta was completely sealed in order to not leak the dye. Evans Blue dye was injected via the needle punctured right below the cross clamp with the heart being hand-massaged, then the heart was excised and sliced. The heart slices were then socked with TTC solution for 30 min. Area at risk (AAR) was defined as the area where it was not stained by Evans Blue, and it was measured as the ischemic insult in each animal. Infarct area was defined as the area where it turned to white color.

After myocardial ischemia and reperfusion (Figure 1A), T2 thoracic dorsal root ganglions and spinal cord were harvested for further analyses. Tissue preparation was performed as previously described (Howard-Quijano et al., 2021). Sections were processed in 4% paraformaldehyde (PFA, Thermo Fisher Scientific, Rockford, IL, USA) at 4°C for 2 days, followed by a 30% buffered (PBS) sucrose solution containing ∼ 0.01% sodium azide until tissues sank (11–14 days). The tissues were embedded with O.C.T. compound (Fisher Scientific, Fair Lawn, NJ, USA) and stored at −80°C until sectioning.

All data are reported as mean and standard error (SE) for continuous data with normal distribution. Statistical normality was confirmed with the Shapiro-Wilk test. Differences in cardiac electrophysiologic measures between the CONTROL and DRGS groups were compared using unpaired t-test or Mann-Whitney tests per data normality. Two-way repeated-measures ANOVA with post hoc Holm-Sidak’s test was used to compare hemodynamic and electrophysiologic measures at different time points within the group and between groups. An unpaired t-test was performed to compare episodes of PVCs and area at risk, and a Mann-Whitney test was performed to compare arrhythmia score, episodes of VT or VF, and infarct size between the groups. Immunoreactive cell counts were compared using an unpaired t-test between the groups. Statistical analysis was done using Prism software (version 8, GraphPad Software Inc., San Diego, CA, USA). A two-tailed p value less than or equal to 0.05 (p ≤ 0.05) was considered to be statistically significant. The sample size was based on previous experience with a similar experimental model and design (Howard-Quijano et al., 2017, 2021).

Thirty min DRGS alone did not change cardiac electrophysiological measurements. ARI and DOR also had no significant differences, as compared to baseline, within each group and between the groups (Table 1).

Two animals in the CONTROL group and one animal in the DRGS group had VF before 30 min of ischemia, therefore the data at 30 min of ischemia was analyzed and reported at the time right before VF happened for those animals.

The ischemic insult as measured by area at risk was greater in the DRGS group than in the CONTROL group (CONTROL: 24 ± 3.4%/left ventricle, DRGS: 31 ± 2.2%/left ventricle, p = 0.017), and the infarct size within the area at rick was comparable between the groups (CONTROL: 13 ± 3.5%, DRGS: 27 ± 7.0%, p = 0.0729).

In each group, as expected in ischemic model, ARI decreased, and DOR increased after ischemia (Table 1). There was no significant difference in the number of ST elevated electrodes during ischemia (Figure 2A). However, as shown in Figure 2B, the ischemic region had lower ARI in the CONTROL group, and when comparing the difference before and after ischemia, there was a significant reduction in the magnitude of global ARI shortening in the DRGS group (CONTROL, −111 ± 10.8 m vs. DRGS, −78.5 ± 6.0 ms, p = 0.0259, Figure 2C). Cardiac electrophysiologic changes were also examined regionally in the ischemic vs. non-ischemic region of the myocardium. In the ischemic zone, the change in ARI was significantly reduced by DRGS (CONTROL, −201 ± 9.8 ms, DRGS, −170 ± 9.4 ms, p = 0.0373, Figure 2D). On the other hand, there was no significant difference in the change in ARI in the non-ischemic region (CONTROL, −38 ± 11 ms. DRGS, −27 ± 8.6 ms. p = 0.503, Figure 2E). In addition, change in DOR during ischemia was also reduced by DRGS (CONTROL, 9546 ± 763 ms², DRGS, 6491 ± 636 ms², p = 0.0076, Figure 2F).

No significant differences were found in HR, SBP, LVESP, and dP/dt max, within the group and between the groups before ischemia (Table 2). During ischemia, there were no significant differences in hemodynamic measurements between the groups. Within-group, SBP and LVESP decreased more in the CONTROL group than the DRGS group after ischemia, while the other parameters had no significant changes (Table 2).

In the superficial dorsal horn of the T2 spinal cord, myocardial ischemia-induced neural cell activation, represented by percent of cFos⁺/NeuN⁺ double-labeled cells (41 ± 13%), was significantly reduced by DRGS (7.8 ± 3.4%, p = 0.0476, Figure 4A, representative expressions shown in Figure 4B). In contrast, in the deep dorsal horn of the spinal cord (laminae lll - V, Vll, X), there was no significant difference in myocardial ischemia-induced neural cell activation between the groups (CONTROL: 46 ± 12%, DRGS: 65 ± 16%, p = 0.36).

Myocardial ischemia induced neuronal apoptosis represented by percent of CC3⁺/NeuN⁺ double-labeled cells (34 ± 15%, 466 of 1,500 cells) in spinal cord dorsal horn, which was significantly reduced by DRGS (22 ± 10%, 255 of 1,641 cells, p = 0.0084, Figure 5A, representative expressions shown in Figure 5B).

We found that DRGS attenuated cardiac ischemia induced sympathoexcitation and ARI shortening. There is a close correlation between action potential duration (APD) and ARI (Haws and Lux, 1990) and activation of myocardial sympathetic afferents during ischemia initiates activation of the sympathetic preganglionic neurons in the IML region in the thoracic spinal cord, which lead to increased sympathetic autonomic nervous system output and life-threatening arrhythmias (Fukuda et al., 2015; Franciosi et al., 2017).

Following myocardial ischemia, conduction blocks occur in the ischemic zone, which cause APD and T-wave alternans, leading to reentrant arrhythmias (Bernus et al., 2005; Downar et al., 2021). Also, after myocardial infarction, there is regional remodeling of the intrinsic cardiac nervous system, thus altering afferent neural signals and neural processing. These changes most frequently occur in the ischemic zones (border zones) and dynamically amplify the risk for ventricular arrhythmias (Vaseghi et al., 2012; Rajendran et al., 2016).

Cardiac primary afferent sympathetic neurons, with cell bodies located in DRG, are an essential component of this pathway for transmission of cardiac signals to the central nervous system during myocardial ischemia (Fukuda et al., 2015). Neuromodulation through bioelectronic DRGS can directly modulate dorsal horn neurons to decrease spinal neuronal sympathetic output to the heart (Sverrisdottir et al., 2020; Kuwabara et al., 2021). DRGS has also been previously shown to augment T-junction filtering through a reduction in action potentials from peripheral systems (Kent et al., 2018). Our results demonstrate that DRGS reduced ischemia induced cardiac sympathoexcitation likely through a reduction in the afferent neural transmission to the spinal cord.

In addition to reducing cardiac sympathoexcitation, this study also demonstrates that DRGS reduced ventricular arrhythmogenesis and resultant ventricular arrhythmias. Myocardial ischemia is accompanied by an increase in the dispersion of ventricular repolarization, primarily caused by disparate shortening of the action potential duration in areas most affected by ischemia (Behrens et al., 1997). Global DOR and Tp-Te interval are closely correlated and they have a strong link with arrhythmogenicity (Vaseghi et al., 2012, 2013), which is consistent with our data showing the reduced magnitude of change with DRGS. Previous results demonstrating a reduction in myocardial excitability created by extraventricular stimulus during DRGS is consistent with our current results (Kuwabara et al., 2021). Importantly, the VAS and episodes of VF during ischemia were significantly reduced by DRGS, which reinforces our electrophysiological data and the effects of DRGS on myocardial sympathoexcitation during ischemia.

The mechanism underlying reduction in ischemia-induced cardiac sympathoexcitation and arrhythmogenesis with DRGS in this study is postulated to be through modulation of myocardial afferent neural input into the dorsal horn of the spinal cord. This mechanism is supported by our data demonstrating a reduction in cFos + neuronal cell activation specifically in the superficial dorsal horn of the spinal cord. We and others have previously shown that cFos + neurons in the spinal cord are an accurate marker of neuronal activation during cardiac ischemia and neuromodulation with SCS (Howard-Quijano et al., 2021). Myocardial ischemia-sensitive ventricular sensory afferent neurons are capable of signaling to central neurons (Fukuda et al., 2015) and coronary artery occlusion has been shown to increase neuronal activity in the left T2 dorsal root ganglia with nerve endings projecting to the superficial dorsal horn of the spinal cord (Armour, 1999). Thus, the reduction in cFos + neurons in the superficial dorsal horn during ischemia with DRGS likely represents a reduction in ischemic afferent neural transmission to the spinal cord, which then leads to reduced efferent output from the thoracic IML neurons to the heart.

Myocardial ischemia activates the afferent neurons in the heart, and this excitatory response will trigger the spinal cord neural network and influence the intermediolateral (IML) neurons to cause the sympathoexcitation through the IML-sympathetic chain pathway (Dale et al., 2020; Omura et al., 2021). One of the potential mechanisms of action through which DRGS works is by blocking of the afferent signaling and preventing excitatory afferent signals from entering the spinal neural network during the ischemia, and thereby the sympathoexcitation is reduced or blunted. Cardiac neuraxis consists of three main levels of control, including intrinsic cardiac ganglia, intrathoracic extracardiac ganglia, and the central nervous system (Armour, 2004; Shivkumar et al., 2016). Brainstem, hypothalamus and the telencephalon areas of the brain are involved in the supraspinal neural processing of myocardial ischemia (Foreman, 1999; Janig, 2014). The sympathetic efferent pathway includes the neural hierarchy of supraspinal, cervical and thoracic spinal cord, sympathetic chain, and the intrinsic cardiac neural systems (Armour, 2004; Shivkumar et al., 2016). Spinal cord, intrathoracic extracardiac ganglia, and intrinsic cardiac nervous system have internal feedback loops which make them capable of receiving afferent information and providing the efferent outflow without communicating with the higher neural centers (Armour, 2004; Shivkumar et al., 2016). During the myocardial ischemia, the afferent neurons are activated (Salavatian et al., 2019) which will directly (through the internal reflex loop) and indirectly (from supraspinal to spinal cord to sympathetic chain and eventually to the intrinsic cardiac nervous system) trigger the sympathoexcitatory response. DRGS modulates the afferent pathway at the level of the spinal cord, and therefore, it has the potential to mitigate the sympathoexcitation that is modulated through the spinal cord.

Additionally, we also found a reduction in neuronal apoptosis in the DRG with DRGS, as demonstrated by a downregulation of neuronal CC3 expression. Previous reports have demonstrated that myocardial ischemia can lead to induction of inflammatory gene expression and/or neuronal apoptosis in different components of the cardio-spinal reflex pathway, including spinal cord, stellate ganglion, and DRG (Gao et al., 2017; Wang et al., 2018). Thus, these evidence and current results suggest that DRGS’s anti-arrhythmic effects, in part, may be due to its ability to mitigate cardiac ischemia-induced loss of neuron cell viability. However, further studies are needed to determine the precise mechanisms through which DRGS impacts DRG cells.

Our study has demonstrated the cardiac protective effects of DRGS, however, the clinical translation is still premature and requires further investigations. We were able to show the effectiveness of DRGS in this study, where DRGS could modulate autonomic imbalances and reduce arrhythmogenesis in the ischemic myocardium in a healthy porcine model. Porcine models are very effective for cardiovascular research because they have similar physiology, heart size, immune system, anatomy, and coronary circulation also resembles human (Spannbauer et al., 2019).

While there are multiple modalities of neuromodulation available (Koetsier et al., 2020), we used 1 kHz high-frequency DRGS in this study based on its clinical efficacy and reduced side effect profile compared to SCS. High-frequency SCS is thought to be a more promising strategy as it has less paresthesias in its clinical application for pain therapies (North et al., 2016) and a recent clinical study showed high-frequency DRGS also worked well to relieve pain without paresthesia (Billet et al., 2017). Prior studies shows that there is no significant difference between low and high-frequency DRGS (Koetsier et al., 2020). It has also been shown that even the conventional DRGS causes less paresthesia compared to the SCS during the stimulation (Deer et al., 2019) and it also improves over time (Mekhail et al., 2020). This indicates that the use of high-frequency stimulation in the kilohertz range, which is usually associated with a higher power consumption, only for the purpose of avoiding paresthesia would not be necessary. In our previous report, we demonstrated that both 20 Hz and 1 kHz DRGS reduced the arrhythmogenicity induced by extra-ventricular stimuli with S1/S2 pacing with no superiority (Kuwabara et al., 2021). Additionally, there are currently no optimal stimulation parameters described for DRGS, especially during myocardial ischemia and more studies are needed to distinguish the therapeutic differences between conventional and kilohertz DRGS in more detail.

Interestingly, we were able to demonstrate efficacy of DRGS while stimulating DRG at a single thoracic level (T2) on the left side supporting results from previous reports that suggest both left or right cardio-spinal neural reflex pathways from the heart contribute to transmission of sympathetic afferent inputs to upper thoracic dorsal spinal columns. Recent evidence has shown that cardiac sensory soma is distributed symmetrically between the left and right sides of DRG (Akgul Caglar et al., 2021). Further, in porcine model, no change in the magnitude of ARI shortening were seen between the left or right spinal dorsal root transections from T1 to T4 spinal levels following sympathetic activation by stellate ganglion stimulation (Yamakawa et al., 2016). While we did not evaluate if there would be an incremental benefit of multiple spinal level or bilateral DRGS, the simplicity of our approach lends to better clinical translation.

We showed the effectiveness of DRGS in a myocardial ischemia porcine model, but this study has some limitations. Anesthesia with isoflurane can suppress nerve activity and change cardiac electrophysiology. However, we discontinued isoflurane after finishing surgical procedures and used alpha-chloralose to minimize the autonomic effects. The protocol remained similar in both the groups, thus the results seen in this study should be least impacted. Also in this study, DRGS was initiated prior to ischemia and we showed DRGS decreased cardiac sympathoexcitation during ischemia, which we think would be considered as the potential strategy in a clinical situation where a second acute myocardial ischemia and/or recurrent ventricular arrhythmia happen. While we performed this study in a healthy porcine model with acute ischemia, there is a possibility that the result may not be directly applicable in the setting of chronic myocardial ischemia/infarction. In this study, the DRGS lead was not placed in the CONTROL group, but the surgical preparation including the laminectomy was performed similarly in both groups. We do not expect that the placement of the DRGS lead in the CONTROL group would significantly change the results of the study as the DRGS lead placement alone in the pain studies could not suppress the activity of sensory neurons to provide the therapeutic effect (Yu et al., 2020). Finally, our molecular findings provide important insight into how DRGS may be reducing ischemia-induced cardiac arrhythmias, further studies are needed to elucidate the specific molecular pathways through DRGS could work in the DRG or spinal cord.