Cardioprotection via the skin: nociceptor-induced conditioning against cardiac MI in the NIC of time

Abstract

Timely reperfusion is still the most effective approach to limit infarct size in humans. Yet, despite advances in care and reduction in door-to-balloon times, nearly 25% of patients develop heart failure postmyocardial infarction, with its attendant morbidity and mortality. We previously showed that cardioprotection results from a skin incision through the umbilicus in a murine model of myocardial infarction. In the present study, we show that an electrical stimulus or topical capsaicin applied to the skin in the same region induces significantly reduced infarct size in a murine model. We define this class of phenomena as nociceptor-induced conditioning (NIC) based on the peripheral nerve mechanism of initiation. We show that NIC is effective both as a preconditioning and postconditioning remote stimulus, reducing infarct size by 86% and 80%, respectively. NIC is induced via activation of skin C-fiber nerves. Interestingly, the skin region that activates NIC is limited to the anterior of the T9−T10 vertebral region of the abdomen. Cardioprotection after NIC requires the integrity of the spinal cord from the region of stimulation to the thoracic vertebral region of the origin of the cardiac nerves but does not require that the cord be intact in the cervical region. Thus, we show that NIC is a reflex and not a central nervous system-mediated effect. The mechanism involves bradykinin 2 receptor activity and activation of PKC, specifically, PKC-α. The similarity of the neuroanatomy and conservation of the effectors of cardioprotection supports that NIC may be translatable to humans as a nontraumatic and practical adjunct therapy against ischemic disease.

NEW & NOTEWORTHY This study shows that an electrical stimulus to skin sensory nerves elicits a very powerful cardioprotection against myocardial infarction. This stimulus works by a neurogenic mechanism similar to that previously elucidated for remote cardioprotection of trauma. Nociceptor-induced conditioning is equally potent when applied before ischemia or at reperfusion and has great potential clinically.

INTRODUCTION

Acute myocardial infarction (MI) is a leading cause of death worldwide, and every year ~720,000 Americans suffer new or recurrent coronary attacks (5). The incidence of MI is rising worldwide (47a), and despite improvements in percutaneous coronary intervention (PCI) and bypass surgery, a substantial number of people have heart failure after damage to the heart post-MI. Thirty-two years ago, the phenomenon of ischemic preconditioning (IPC) was discovered, demonstrating that the mammalian heart has endogenous cardioprotective mechanisms (43). In 1993, Przyklenk et al. (45) demonstrated the existence of remote preconditioning, a cardioprotective phenomenon in which the protective ischemic stimulus is distant from the tissue protected against a subsequent injurious insult, and this was followed up by a series of important confirmatory papers extending the concept (7, 22, 57). In a series of papers, Vinten-Johansen and colleagues demonstrated the concept of ischemic postconditioning, that a stimulus given at the time of coronary reperfusion can elicit cardioprotection (61), and showed that postconditioning can be elicited remotely (RIPostC) (31). Importantly, RIPostC provides the possibility that cardioprotection may be initiated at a clinically relevant time, from an organ other than the heart, thereby avoiding damage to the heart. Since this discovery, IPC and RIPostC have been demonstrated to occur in mice, rats, dogs, sheep, pigs, and humans (26, 37, 38), and the mechanisms have been studied in detail (1, 39, 51, 56), with many attempts to translate via clinical trials (6, 8, 12, 13, 25, 26, 28, 29, 32, 33, 36, 37, 55, 60). Although clinical trials have shown some limited benefits, nothing has yet proven reliable and effective in large clinical trials. An easy to implement and efficacious cardiac conditioning therapy with no or low adverse events remains the goal in the field.

We have previously identified a novel and powerful cardioprotective phenomena that occurs subsequent to abdominal skin incision, called remote cardioprotection of trauma (RPCT) (30, 46). The discovery that skin incision of a specific spot around the umbilicus initiates RPCT against MI was accidental, as previously described (46). We showed that RPCT is TNF-α independent (unlike IPC) (46) and involves sensory nerves in the skin, and we traced a pathway from the skin to the cardiac innervation (30). Remote stimulation of the cardiac nerves activates PKC in the myocardium, and this activates ATP-sensitive K⁺ (K_ATP) channels, leading to cardioprotection via a mechanism much like that of early phase IPC (30). However, surgical incision would not be practical in the clinical setting.

In the present study, we hypothesized that direct electrical stimulation (ES) of the skin activates C-fiber sensory nerves to produce a remote conditioning effect similar to RPCT and that the mechanism involves a neurogenic signal that propagates from the skin to the cardiac nerves, releasing neuropeptides that signal myocardial cells to activate PKC-dependent cardioprotection. We further showed that ES may be used at reperfusion to initiate cardioprotection. We found that ES in a very specific region (anterior abdominal T9−T10 dermatome) initiates this cardioprotection, and we term this phenomenon nociceptor-induced conditioning (NIC). We show that NIC provides protection against infarction as both a preconditioning and postconditioning stimulus. The mechanism involves skin C-fiber neurons and requires spinal integrity between the T9−T10 dermatome and the origin of the cardiac innervation (T1−T5) but does not require signal from the central nervous system (CNS). We ruled out indirect cardioprotective effects of vascular dilatation. NIC is noninvasive, does not cause tissue injury, is highly clinically feasible, and therefore is important to develop translationally and test as a therapeutic strategy.

METHODS

Animals.

Mice were maintained according to Institutional Guidelines (Loyola University, Chicago, IL) and the National Institutes of Health Guide for Care and Use of Laboratory Animals (Revised 2010) and the Position of the American Heart Association on Research Animal Use (1985). Wild-type (WT) mice were B6129SF2/J F₂. Bradykinin 2 receptor knockout mice (BK2R-KO; JAX strain 101045) and transient receptor potential vanilloid 1 knockout mice (TRPV1-KO; JAX strain 003770) (62) were on the same background (C57, JAX strain 000664). For these knockouts, mice of the C57 background were used for controls; sibling controls could not be used as knockouts were maintained in the homozygous state. All experimental groups consisted of male and female mice in equal numbers; post hoc analyses revealed no differences between sexes in end point measurements in these experiments.

Murine models and surgical procedures.

Mice were subjected to surgical procedures as previously described in detail (30, 46). Briefly, we used a murine model of ischemia-reperfusion (I/R) to determine whether direct ES of skin sensory nerves can elicit cardioprotection against MI. The coronary artery occlusion site was exposed through a very small left thoracotomy, and a loop occluder was applied to provide I/R in the open chest state. All mice were continuously monitored via ECG, and those without evidence of ischemia or of reperfusion were disqualified from the study (2%). Survival was 95% overall. We attached six leads from a KWD-808-I multipurpose health device (Yingdi) spread along a line on the anterior abdomen located in the T9−T10 dermatome or at other experimental locations. With the use of finger and facial sterile acupuncture needles (0.18 × 7 mm, Millenia Acupuncture, Alhambra, CA), mice were subjected to a transcutaneous ES (10-V 100-Hz pulse width 400 μs) for 20 min, either before ischemia (preconditioning) or at reperfusion (postconditioning), using a 45-min coronary occlusion. Infarct size was measured as previously described (30). In these experiments, there were no significant differences between the size of the region at risk of experimental versus control groups, as previously reported for the model (Table 1) (30). Notably, this stimulation involved needles inserted and looped through the skin and not penetrating underlying tissues. This is unlike electroacupuncture, which involves inserting the needles deeply into muscle and other tissues at predefined acupoints (mapped based on the acupuncture medicinal theory of meridians and observation of sensation) to stimulate (through mechanical or electrical means) that acupoint (not specifically a nerve). The ES stimulus used to initiate NIC is not traditional acupuncture but direct ES of the skin and associated sensory nerves. Our rationale for this approach is based on our previous discovery that sensory nerve blockade prevents RPCT (30). After stimulation, needles were carefully unhooked from the electrical wires and were removed at the end of the procedure before the mice were recovered. Infarct size was measured after 24-h reperfusion except for mice having undergone spinal transection, which received 4-h reperfusion for humane reasons (see below).

Table 1.

Values for infarct/LV, risk/LV, and infarct/risk of the experimental groups in this study

Group	Procedure	Infarct/LV	Risk/LV	Infarct/Risk	P Values Versus Control for Group (Risk/LV)	P Values Versus Control for Group (MI/Risk)	Figure
1	ES pre-CON	5.8 ± 1.6	58.7 ± 3.2	8.7 ± 5.5	P > 0.05	P ≤ 0.001	1C
2	Sham	36.3 ± 5.3	55.4 ± 5.6	60.2 ± 4.2	Control for 1	Control for 1	1C
3	ES post-CON	5.5 ± 2.1	57.1 ± 4.4	10.6 ± 7.2	P > 0.05	P ≤ 0.001	1C
4	Sham	33.1 ± 5.8	52.6 ± 3.6	52.6 ± 2.8	Control for 3	Control for 3	1C
5	ES T9-T10	7.4 ± 1.6	54.7 ± 3.2	11.2 ± 2.9	P > 0.05	P ≤ 0.001	2A
6	Control T9-T10	36.3 ± 5.3	55.4 ± 5.6	57.3 ± 4.4	Control for 6–9	Control for 5 and 7	2A
7	Forepaw	36.2 ± 5.1	59.0 ± 3.6	53.4 ± 4.2	Control for 6–9	P > 0.05	2A
8	ES T7-T8	22.5 ± 9.3	56.97 ± 3.2	25.9 ± 12.9	Control for 6–9	P ≤ 0.001	2A
9	Sham T7-T8	37.6 ± 2.9	59.3 ± 2.98	53.7 ± 3.9	Control for 6–9	Control for 8	2A
10	Caps T9-T10	7.2 ± 3.6	54.9 ± 5.3	11.5 ± 3.5	P > 0.05	P ≤ 0.001	2B
1	Sham Caps	33.9 ± 5.7	59.75 ± 2.3	46.7 ± 2.6	Control for 11–17	Control for 10	2B
12	Forepaw Caps	38.7 ± 5.9	55.7 ± 6.1	57.2 ± 6.3	Control for 11–17	P > 0.05	2B
13	Caps middle 1/3	32.6 ± 4.7	60.1 ± 4.2	45.1 ± 3.6	Control for 11–17	P > 0.05	2B
14	Caps left 1/3	40.1 ± 4.9	61.7 ± 6.2	53.2 ± 5.1	Control for 11–17	P > 0.05	2B
15	Caps right 1/3	41.98 ± 1.4	59.43 ± 1.5	55.53 ± 1.8	Control for 11–17	P > 0.05	2B
16	Caps T7-T8	38.6 ± 6.5	60.2 ± 1.18	52.9 ± 6.4	Control for 11–17	P > 0.05	2B
17	Sham Caps	38.5 ± 3.9	57.8 ± 2.95	51.3 ± 2.3	Control for 11–17	Control for 12–16	2B
18	ES + lidocaine	37.1 ± 4.9	51.7 ± 4.2	50.8 ± 8.2	P > 0.05	P > 0.05	3A
19	Sham	43.1 ± 4.9	55.7 ± 4.2	57.2 ± 4.1	Control for 18	Control for 18	3A
20	ES + vehicle	4.2 ± 2.6	56.8 ± 5.4	6.5 ± 1.5	Control for 18	P ≤ 0.001	3A
21	ES + sham	5.8 ± 1.6	58.7 ± 3.2	8.7 ± 5.5	Control for 18	P ≤ 0.001	3A
22	Sham no drug	36.3 ± 5.3	55.4 ± 5.6	60.2 ± 4.2	Control for 18	Control for 20 and 21	3A
23	NIC + TRPV1-KO	33.4 ± 7.8	59.5 ± 4.5	46.0 ± 9.6	P > 0.05	P ≥ 0.05	3B
24	Control TRPV1-KO	36.6 ± 5.1	61.8 ± 2.5	57.9 ± 6.3	Control for 23	Control for 23	3B
25	NIC + WT	5.3 ± 3.6	55.9 ± 4.7	9.3 ± 2.2	Control for 23	P ≤ 0.001	3B
26	Control WT	36.3 ± 5.3	55.4 ± 5.6	60.2 ± 4.2	Control for 23	Control for 25	3B
27	ES + C7	6.8 ± 4.6	54.4 ± 3.2	10.2 ± 3.5	P > 0.05	P > 0.05	3C
28	Control ES	8.8 ± 2.5	60.4 ± 3.2	12.4 ± 3.8	Control for 27	Control for 28	3C
29	ES + T7	35.3 ± 3.3	56.4 ± 5.1	54.2 ± 5.1	Control for 27	P ≤ 0.001	3C
30	Sham	6.2 ± 3.6	58.4 ± 4.2	11.2 ± 5.4	Control for 27	Control for 30	3C
31	ES + BK2R-KO	37.3 ± 2.9	53.2 ± 3.5	54.8 ± 4.4	P > 0.05	P > 0.05	4A
32	Control BK2R-KO	34.3 ± 8.3	60.4 ± 7.6	56.2 ± 7.2	Control for 31	Control for 32	4A
33	ES + WT	36.37 ± 5.5	59.4 ± 6.6	60.2 ± 4.2	Control for 31	P ≤ 0.001	4A
34	Control WT	8.0 ± 6.5	63.4 ± 6.2	13.4 ± 5.5	Control for 31	Control for 33	4A
35	ES + chelerythrine	36.7 ± 5.6	58.2 ± 3.6	53.6 ± 4.2	P > 0.05	P ≤ 0.001	4B
36	ES + vehicle	4.2 ± 3.1	55.1 ± 8.2	6.3 ± 2.1	Control for 35	Control for 35	4B
37	Vehicle no ES	35.4 ± 4.2	63.2 ± 4.1	56.6 ± 7.9	Control for 35	P > 0.05	4B
38	Control no ES	38.4 ± 5.4	60.8 ± 5.7	57.3 ± 3.5	Control for 35	Control for 37	4B

LV, left ventricle; MI, myocardial infarction; WT, wild type; ES, electrical stimulation; BK2R-KO, bradykinin 2 receptor knockout mice; TRPV-KO, transient receptor potential vanilloid knockout mice; CON, conditioning; Caps, capsaicin.

Spinal transections.

The spine was transected separately at two levels to determine the role of the CNS and the integrity of the spinal nerves between T7 and C7 in the development of cardioprotection after NIC. This was performed by exposing the spinal column and cutting with a sharp surgical blade under a microscope (limited bleeding and coagulation of bleeders). Sham control mice were subjected to the identical surgical approach, but the cord was not cut. After 15 min, the wounds were closed and the mouse was moved to a dorsally recumbent position for the cardiac surgery. Mice were not allowed to regain consciousness. This was for humane reasons of avoiding discomfort to the animals and based on observation of many in the field that infarct size is the same at 4 and 24 h post-I/R per Institutional Animial Care and Use Committee approval (53).

Pharmacological agents.

Pharmacological agents were used to block sensory nerve activity and to induce NIC by chemical means. To block skin sensory nerves, lidocaine (100 μl of 1% in saline) was administered subcutanously. Five minutes before skin stimulation, control was saline. To chemically induce NIC, topical capsaicin (0.1% cream, Chatten, 150 μl/25 g per mouse) was administered topically to a 1.0 × 2.0-cm area located lateral to the umbilicus, using gloves and a mask to avoid exposure to personnel. Nitroglycerin was given intravenously (0.15 mg/kg) for assessment of the echocardiography contrast methodology in measuring vascular dilation.

Echocardiography and contrast experiments.

A Vevo 2100 imaging system (Visualsonics, Toronto, ON, Canada) was used to obtain the images using nonlinear contrast mode with a MS-250 probe (21-MHz centerline frequency); using first the B-mode, a clear parasternal long-axis (PSLAX) view was obtained. Subsequently, the image was changed to nonlinear contrast mode at 10% intensity (peak to peak pressure of 4.7 MPa), and a baseline PSLAX view was captured and cine stored. Twenty microliters of contrast (echogenic liposomes, BRACCO Research, Geneva, Switzerland) was injected, and images were captured every minute for 20 min on contrast mode. After acupuncture needles were placed, contrast agent was injected, and images were captured every minute for 20 min as described above. Subsequently, the electric current was turned on to initiate ES, another 20 μl of contrast injected, and images were captured every minute for 20 min (3 sets of data for each mouse). To use this methodology for measurement of vascular flow, we used eight mice. These mice underwent contrast agent injection. Images were obtained as described above but were obtained every 10 s for 3 min and then every minute for another 2 min. Then, in a second run, mice were given 90 μl iv nitroglycerin (0.15 mg/kg) simultaneous with contrast injection and subjected to echocardiography again with images obtained every 10 s for 3 min and then every minute for another 2 min. Mice were euthanized at the end of the procedure. All saved cine images were transferred to a separate computer station for postprocessing.

Data postprocessing used Vevo 2100 software (version 1.1.X, Visualsonics). Two areas of interest with same size (0.219 mm²) were traced one in the cavity and the other one in the anterior wall of the left ventricle (LV) from the PSLAX view for every cine stored image. The region graph calculation was selected from the software for the areas of interest, and graph analysis was displayed showing the contrast enhancement in both areas expressed in linear arbitrary units (LAU). The LAU is a scalar quantity which is not standardized to any physical property; these values can be compared with one another in a relative fashion since the properties of linearity are still satisfied. The values were exported to an Excel spread sheet, where all LAU values of the contrast were averaged for each time point (~450 values for each time point). The averaged value at each time point, including baseline, was obtained for the cavity and anterior wall. Each time point, for both the cavity and anterior wall, was compared with the individual average baseline to standardize the value. The anterior LAU was compared with the cavity LAU, and a relative myocardial flow value was obtained. The nitroglycerin validation did show a nitroglycerin-dependent significant increase in myocardial flow beginning at 50 s (0.25 ± 0.03 vs. 0.17 ± 0.01, P ≤ 0.01) until 120 s (0.25 ± 0.03 vs. 0.17 ± 0.01, P ≤ 0.01).

Western blots.

Mice were euthanized 15 min after the ES stimulus, the heart was removed and rinsed in sterile cold saline, and segments were flash frozen in liquid nitrogen. Frozen tissue was powdered at liquid nitrogen temperatures, homogenized, and centrifuged as previously described (30) to generate cytosolic and membrane fractions for these assays. Proteins (50 mg/lane) were electrophoresed on 10% SDS-PAGE gels and transferred to membranes, which were then subjected to immunoblot analyses using antibodies to PKC-α, PKC-ε, PKC-δ, and PKC-ζ as previously described (35). Densitometry was performed using a Fluorchem 888p imager. Since normalization of protein load with protein stains is not subject to variations due to experimental treatments and is comparable or better than antibody determination of β-actin (47, 48), we used Ponceau S staining to normalize for transfer efficiency and loading equality of Western blots. This was accomplished by staining the membrane with 0.1% Ponceau in 5% acetic acid and then digitally scanning and densitometrically analyzing the stain in each lane. Estimates of protein loading were made from these data and used to normalize signal intensity of the PKC isoform-specific signals.

Histological, DNA fragmentation, and TUNEL analysis.

Hearts were removed from animals 24 h after stimulation, perfused with sterile chilled PBS, fixed by pressure perfusion, and left overnight in 10% formalin in PBS. Hematoxylin and eosin staining was performed in slices from hearts exposed to NIC or sham stimulus, and hearts removed from naïve animals were also analyzed (normal myocardium). Hearts were postfixed, paraffin embedded, sectioned, and put onto slides as previously described (30). The DeadEnd Fluorometric TUNEL kit (Promega, Madison, WI) was used to assess DNA fragmentation in situ and DAPI staining was used to delineate nuclei. Immunostaining with an antisarcomeric actin antibody (Sigma-Aldrich, St. Louis, MO) was done as previously described to delineate cardiomyocytes (30). TUNEL-positive nuclei (green) were counted from 10 randomly selected fields (n = 5–6, 350–450 nuclei/field), and data were expressed as the percentage of total nuclei (apoptotic index). For a separate analysis of DNA fragmentation, associated with apoptosis, separate groups of mice (n = 4) were used for an ELISA-based assay (cell death detection ELISA, Roche Applied Science, Indianapolis, IN) that determines the relative ratio of histone associated DNA fragments (mono- and oligonucleosomes) in samples (increased fragmentation is associated with apoptosis). Results were normalized to the internal control (with the kit), and data are expressed as percentages of the positive control.

Statistics.

For all experiments, group size was determined by Power analysis as previously described (30, 46) using previous or pilot data. Quantitative end point parameters are expressed as means ± SE. For all data except echocardiographic data, significance was determined using a Student t-test to compare the experimental group with the control group after tests were passed for normal distribution and variance. Variance was not different between groups in all tests and distributions were normally distributed; thus, nonparametric testing was not used. For multiple group comparisons, one-way ANOVA followed by a Fisher post hoc test was used. P ≤ 0.05 was considered statistically significant. For the echocardiographic flow data, linear regression to a cubic function was computed using Matlab. Data were plotted and revealed the dynamic response for each group (ES, needle, and baseline). We computed P values for the upward trend associated with each group by computing the derivative of each cubic fit on a fixed interval.

RESULTS

The results of pre- and postconditioning experiments (Fig. 1, A and B) demonstrated that ES at the abdominal site, previously shown to elicit cardioprotection consequent to surgical incision (30, 46), elicits cardioprotection that reduces infarct size by 85%, relative to sham control (placement of needles but no current). Importantly, since eliciting cardioprotection at reperfusion has high clinical applicability, we showed that the same ES stimulus applied at the beginning of reperfusion is as protective as the preconditioning stimulus (Fig. 1C). The extent of this cardioprotection in terms of the reduction in infarct size reported herein was similar to that previously shown to be elicited by skin incision (30). Histological examination showed that there was decreased infiltration after I/R preceded by ES relative to sham treatment (Fig. 1C). We further showed that ES-induced NIC reduces apoptosis, as evidenced by TUNEL staining and an independent experiment using a DNA fragmentation assay (Fig. 1, E–G).

Fig. 1.

Electrical stimulation (ES) stimulation elicits remote pre- and postconditioning. A: the in vivo mouse model for ischemia-reperfusion (I/R) challenge. ECG monitoring was used to determine I/R (ST segment elevation and normalization, respectively). Ischemia was also confirmed by visual observation (middle: cyanosis) (30, 46). B: representative infarct size after sham (left: control) or 20-min ES [nociceptor-induced conditioning (NIC)] performed at T9−T10 of the anterior abdomen 20 min before I/R. C: graphic representation of protection elicited by NIC as a preconditioning stimulus (left pair of bars) and as a postconditioning stimulus (right pair of bars). The results show that the stimuli (orange) elicited an 80–85% reduction in infarct size compared with sham values (8.7 ± 5.5% vs. 60.2 ± 4.2% for preconditioning, 10.6 ± 7.2% vs. 52.6 ± 2.8% for postconditioning, values are means ± SE, n = 6, *P ≤ 0.001). D and E: hematoxylin and eosin staining (D) and sarcomeric actin staining (E) of sections revealed that most, but not all, apoptotic nuclei were cardiomyocytes. F and G: both TUNEL (F; green fluorescence) and DNA fragmentation (G) were significantly reduced after ES-induced NIC compared with sham in separate groups of mice (n = 4 for fragmentation, P ≤ 0.05; n = 5–6 for TUNEL, P ≤ 0.05).

We next performed a series of experiments to determine the specificity of the anterior abdominal skin field shown to induce NIC after ES. Previously, we showed that incisions in the neck (nonischemic carotid vascular surgery) increases, whereas the anterior abdominal site reduces, infarct size (46). We show here that ES to the anterior abdomen in the T7−T8 dermatome gave a reduced degree of cardioprotection relative to the T9−T10 location (Fig. 2A). We repeated this experiment using topical capsaicin (0.1%) as a stimulus as previously reported (30), and the results were consistent (Fig. 2B). We obtained strong cardioprotection at the T9−T10 level but no protection at the T7−T8 level. In both experiments, we were unable to elicit cardioprotection by electroacupuncture stimulation of the forepaw skin over the PC6 acupuncture sight, a site predicted by acupuncturists and shown to elicit cardioprotection by traditional needling in rats. We also applied capsaicin only to subregions of the T9−T10 site. These stimuli did not initiate cardioprotection (Fig. 2B).

Fig. 2.

Effect of the stimulus location on the cardioprotective effect of nociceptor-induced conditioning (NIC). A: electrical stimulation (ES) positioned at the T9−T10 versus T7−T8 dermatome of the anterior abdomen. The protection against myocardial infarction using the T9−T10 position was strong, as previously reported. The protection after stimulation at T7−T8 was less; although infarct size was significantly lower than control (no current), it was also significantly larger than the T9−T10 group (n = 5–6, *P ≤ 0.05 vs. control and #P ≤ 0.05 vs. T9−T10). B: similar location experiment using topical capsaicin as a stimulus. The results demonstrated that the T9−T10 location produced significant cardioprotection (n = 5–6, *P ≤ 0.05 vs. control). The capsaicin stimulus did not produce any protection at the T7−T8 location. Additional groups (n = 3–5) were added to look at the required size of the stimulus field, divided into thirds. None of these produced significant cardioprotection against myocardial infarction (P > 0.05).

To identify the mechanism underlying initiation of ES-induced NIC, we pretreated the stimulus field with lidocaine (1% lidocaine, 25 μl × 4, sc) 5 min before placement of needles and application of current to the abdominal skin. After 20 min of ES, mice were subjected to 45 min of coronary artery occlusion, and infarct size (relative to the risk region) was assessed 24 h later. The results (Fig. 3) showed that the cardioprotective effect of ES is abrogated by local skin peripheral nerve blockade. We hypothesized that activation of C-fibers in the abdominal skin is the common initiator of NIC due to these stimuli after ES. As the TRPV1 receptor is the major pain receptor in C-fibers in the skin sensory nerves (30), we treated control and TRPV1^−/− mice (n = 6) with ES at the anterior abdominal T9−T10 level (20-min stimulus before I/R, 45-min coronary occlusion) and determined infarct size 24 h later. Protection consequent to NIC was abolished in TRPV1-KO mice (Fig. 3B), similar to previous results with topical capsaicin treatment (46),

Fig. 3.

The cardioprotective effect of electrical stimulation (ES) is initiated by a neurogenic signal from the abdominal skin to the heart via the spine. A: the cardioprotective effect of ES was blocked by local subcutaneous lidocaine (50.8 ± 8.24% vs. 6.5 ± 1.5%, n = 6, *P ≤ 0.001). B: infarct size was not significantly reduced by ES in transient receptor potential vanilloid 1 knockout (TRPV1-KO) mice (57.9% ± 6.3 in TRPV1-KO mice with no ES vs. 46.03% ± 9.6 in TRPV1-KO mice with ES, n = 6, *P > 0.05). C: spinal transection at the T7 level completely blocked ES-induced cardioprotection against myocardial infarction relative to no transection (control), whereas transection at the C7 level left protection intact. n = 6. *P ≤ 0.05 vs. control (spine not cut, see methods).

To determine the involvement of spinal nerves and dorsal root ganglia (DRG), we performed spinal transection at two levels. We severed the spinal cord at the location of the seventh cervical or seventh thoracic vertebra (C7 vs. T7). Sham control surgeries were performed in exactly the same way with the same timing except that the cord was not cut (animals in the sham group underwent laminectomy without spinal transection). Immediately after the spinal surgery, ES was administered, and, 20 min later, mice were subjected to 45-min I/R followed by 4-h reperfusion under anesthesia. Mice were not allowed to regain consciousness (for humane reasons of avoiding discomfort to the animals and based on observation of many in the field that infarct size is the same at 4 and 24 h post-I/R (37). The results (Fig. 3C) were that T7 transection abrogates, whereas C7 transection has no effect, on cardioprotection.

To investigate the role of neuroendocrine factors as mediators of NIC in the myocardium, we used BK2R-KO mice and found that genetic ablation of the receptor completely abolished NIC-induced cardioprotection (Fig. 4A). These results support aspects of the proposed action of the NIC signal on the nerve cells in the heart.

Fig. 4.

A: the cardioprotective effect of nociceptor-induced conditioning (NIC) was abolished in bradykinin 2 receptor knockout (BK2R-KO) mice (54.8 ± 4.4% vs. 13.4 ± 5.5%, n = 6, *P ≤ 0.001). B: the cardioprotective effect of electrical stimulation (ES) was blocked by administration of chelerythrine (53.6 ± 4.2% vs. 6.3 ± 2.1%, n = 6, *P ≤ 0.001). C: representative Western blots of PKC-α, PKC-ε, PKC-δ, and PKC-ζ in the cytosol (Cy) and membrane (M) fractions of the heart subjected to either ES or sham (no ES). D–G: quantitative results of the Western blots. The results showed an increase in the translocation of PKCα (P ≤ 0.05, n = 6) in NIC after ES, whereas the ratios of PKC-ε, PKC-δ, and PKC-ζ were not significantly changed after ES (n = 6, P > 0.05).

An important downstream mediator of cardioprotection in many scenarios is PKC. Therefore, we treated mice with a relatively selective general PKC inhibitor (chelerythrine, 5 mg/kg iv) before ES stimulation and assessed the effect on infarct size. The results demonstrate that the protective effect of NIC was significantly blocked by PKC inhibition (Fig. 4B). Using Western blot analysis, we determined that PKC-α in the myocardium is the isoform that is significantly (n = 6, P ≤ 0.05) activated after remote peripheral NIC stimulation, whereas PKC-ε, PKC-δ, and PKC-ζ were not activated (P > 0.05; Fig. 4, C–G).

Since our results suggested a role for bradykinin, known to be cardioprotective but also to cause vascular dilation, we investigated whether coronary flow is altered after NIC stimulated by peripheral ES, since altered coronary flow could lead to enhanced salvage. To do this, we used echogenic liposomes (BRACCO Research) to evaluate, in real time, the vascular volume of the LV anterior and posterior walls before, during, and after stimulation. The results (Fig. 5) demonstrated a significant increase in flow, as signified by increased echogenic material, in both LV walls during the time interval from 11 min (0.98 ± 0.06 and 0.94 ± 0.05 vs. 0.7 ± 0.11, P ≤ 0.05) to 17 min (1.25 ± 0.14 and 1.04 ± 0.04 vs. 0.77 ± 0.10, P ≤ 0.05) poststimulus. Interestingly, this increased flow occurred after ES (green) but also with needle placement with no current (red) relative to baseline (blue). There was no significant difference between needle placement and ES at any time point (P > 0.05). This method was validated by assessing the change in flow after giving nitroglycerin (Fig. 5, inset; see methods), showing a significant transient increase in flow attributable to vasodilation between 50 s and 2 min after nitroglycerin treatment, consistent with its known action and half-life.

Fig. 5.

Effect of bradykinin on vascular flow in the heart after electrical stimulation (ES). Echogenic liposomes from BRACOO Research were injected into mice and flow in two segments of the heart monitored by echocardiography over 20 min. Since the amount of contrast material measured in real time is proportionate to blood flow, we can detect changes in blood flow over time by taking baseline measurements (blue) and comparing these with measurements taken after manipulation [ES (green) or sham ES (red)]. Statistical analyses were used (see Statistics and text) to determine whether separations between treatments are statistically significant (yellow box, n = 8, P ≤ 0.05). The results support that flow was significantly increased between 11 and 17 min posttreatment after both sham and ES treatment relative to baseline. Since needle placement alone does not reduce infarct size, the cardioprotection consequent to NIC cannot be related to this vasodilatory effect. LAU, linear arbitrary units.

DISCUSSION

This study demonstrates that skin nociceptor stimulation by ES, designated as NIC herein, induces a powerful cardioprotection against MI, at least in part, by reducing cardiomyocyte apoptosis. This is initiated by skin C-fibers and requires a specific neurological pathway and activation of effectors in the myocardium. We show that NIC can be initiated by ES to skin and by topical capsaicin at specific locations on the anterior abdomen near the T9−T10 dermatome. The topical capsaicin at the doses used is not expected to lead to detectable serum levels of capsaicin; the effects are limited to the skin.

ES-induced cardioprotection was abrogated with TRPV1-KO, and this is interpreted as showing the involvement of TRPV1 in peripheral C-fibers. TRPV1 is the receptor for capsaicin, and capsaicin-specific effects are dependent on this receptor (30). This suggests that capsaicin and ES-induced cardioprotection share involvement of TRPV1.

Lidocaine is known to activate sensory nerves through multiple channels and can potentiate TRPV1 and TRPA channel activity. However, at the dosage we used, lidocaine depletes the nerves of neuropeptides and is an effective and relatively selective blocker of C-fiber nociceptors, as previously shown (30). Apparently, at this concentration, the transient activation of the nerves by lidocaine itself is not sufficient to activate cardioprotection but does prevent activation of C-fiber nociceptor-containing cells by ES and blocks cardioprotection by NIC. This result supports that ES-induced cardioprotection requires activation of skin sensory nerves. This is consistent with our previously published results demonstrating that cardioprotection can be initiated by noxious stimuli (skin incision or topical capsaicin) administered to the same field used in these studies (30, 46).

We show that the skin sensory nerves that initiate NIC must be in the region of the T9−T10 dermatome, lateral to and on both sides of the umbilicus (Fig. 2). We conclude that there exists a skin sensory neural field, located on the anterior portion of the T9−T10 dermatome, that has a special relationship to the cardiac innervation. The nature and developmental genesis of this connection will require further investigation.

From the skin sensory nerves, in theory, there are several pathways that a signal may travel to the heart, via 1) humoral signals; 2) neural pathway involving the spinal column, DRG, and CNS; 3) neural pathway involving the spinal column and DRG (not CNS); and 4) neural pathway involving vagal innervation. The results of spinal transection show that the CNS is not involved, while the fact that T7 transection abrogates NIC supports that the spinal column/DRG must be intact between the initiating skin region and cardiac spinal innervation (Fig. 3). Vagal nerve stimulation has been shown to be cardioprotective (17, 35). In our study, the vagal nerves would be intact after T7 transection, so they are ruled out as mediating the signal as well. Humoral factors could presumably be secreted in the skin, nerves, or heart after ES. If cardioprotection were mediated by these factors alone, then T7 spinal transection would not prevent cardioprotection unless, as has been supported (41), intact innervation to the organ was still required. However, in the instance that a humoral factor were secreted from skin or associated nerves and integrity with the nervous system were required for secretion of such a factor, C7 transection would not have been protective. Thus, a humoral factor secreted from the skin or lower spinal innervation is unlikely to be critical for NIC. We cannot rule out that an autocrine humoral factor is secreted by the heart or its innervation. However, we can definitively conclude that the integrity of the innervation from the skin to the cardiac nerves via the spinal DRG is absolutely required for NIC. Our interpretation is that after stimulation, a neural signal from the skin sensory nerves at T9−T10 travels to the T9−T10 DRG, up the spinal cord to the cardiac associated DRG (T1−T5) (30), and thence to the heart. It has been previously shown that a “dorsal root reflex” can be caused by high intensity C-fiber activation (21, 44). The dorsal root reflex is a volley of action potentials that can be recorded emerging antidromically from the spinal cord in the dorsal roots after ES of an adjacent dorsal root or sensory nerve (3, 9). Although much work showing a dorsal root reflex has been performed in isolation with chilled preparations, two groups have demonstrated in vivo spread of the dorsal root potentials both along and across the cord (11a, 50). Our proposed mechanism is biologically supported by this body of work. Thus, the neurological mechanism of cardiac conditioning by NIC appears to be a reflex between a specific peripheral nerve field and the heart that does not involve the CNS.

The next question we posed relates to how the neural signal reaching the heart acts within the myocardium. Based on previously published studies (11, 54), we hypothesized that a signal from the DRG at the level of the cardiac innervation (T1−T5) spreads antidromically to the heart along the sensory nerves and activates the cardiac sympathetic nerves, as previously supported for spinal cord stimulation (1, 19, 54). There is precedent for involvement of the cardiac innervation in cardioprotection and sensory nerve blockade in the heart can block cardioprotection (18, 34). In the heart, this would involve release of calcitonin gene-related peptide from sensory nerves, which may then act on the sympathetic nerves as previously shown by us and others (15, 16, 30, 54, 59). In this study, we proposed that such stimulation of the sympathetic nerves would then trigger release of bradykinin, which can act directly on cardiomyocytes, through the bradykinin receptor 2 (BK2R) (52). Bradykinin can stimulate nerve cells to release catecholamines that would act on the β-adrenergic receptor of cardiomyocytes (norepinephrine). We have previously shown that calcitonin gene-related peptide activity is required for capsaicin-induced RPCT (30). An important downstream mediator of cardioprotection in many scenarios is PKC, which is downstream of both β-adrenergic receptor and BK2R (20, 42, 49). Collectively, our results (Fig. 3) support that PKC signaling is required for cardioprotection afforded by NIC and that PKC-α in the heart is activated by the ES stimulus. PKC-α has previously been shown to be cardioprotective and appears to convey its effect through modulation of the mitochondrial permeability transition pore channel. PKC-α has been implicated in sildenafil-induced cardioprotection (14) and is downstream of mitochondrial K_ATP channel activation after IPC in the human myocardium (27). This agrees with our previous data that NIC elicited by topical capsaicin requires activation of K_ATP channels (30). We posit that BK2R activation of PKC-α and subsequent intracellular signaling mediate the cardioprotective effects of NIC against MI in vivo.

Since our results demonstrated a role for bradykinin, known to be cardioprotective but also to cause vascular dilation, we investigated whether coronary flow is altered after NIC stimulated by peripheral ES and whether such an effect might underlie cardioprotection. Our results with echogenic liposomes (Fig. 5, see methods) demonstrate a significant increase in flow, as signified by increased echogenic material, in both LV walls beginning 11 min after stimulation. The increase appears to last up to 17 min post-ES. Interestingly, this increased flow occurs after ES but also with needle placement with no current (needle only) but not at baseline. This suggests that there may be a neurally activated pathway that affects coronary dilation and that needle placement can activate this. However, needle placement alone does not affect infarct size (Fig. 1), leading us to conclude that coronary flow changes do not underlie NIC-induced cardioprotection. This supports that NIC-induced cardioprotection is due to signaling within cardiac component cells and not to vascular effects. Nevertheless, the effects of needling on coronary flow will be of interest and could have clinical application if effective in humans.

Our overall results are consistent with the following mechanism (Fig. 6): the peripheral stimulus activates skin sensory nerve C-fibers that are localized so as to trigger a reflex signal to the heart. This travels via the spinal cord and DRG (30) to the DRG that connects to the cardiac sensory nerves (T1−T5). The signal travels from these cardiac-associated DRGs to cardiac sensory nerves via an antidromic signal and the sensory nerves of the heart activate the cardiac sympathetic nerves transiently. The interneuronal signaling between sensory and adrenergic nerve endings in the heart acts to releases bradykinin and norepinephrine, which activates cardioprotection via PKC-α activation in the myocardium. Bradykinin release may also initiate coronary dilation, although further studies need to be done to confirm this. The results suggest that PKC-α intracellular signaling plays an important role in the cardioprotection of NIC against MI. These findings provide new insights regarding both neural and molecular mechanisms involved in NIC.

Fig. 6.

Schematic outline of how electrical stimulation (ES) activates the cardiac nerves via the spinal cord/dorsal root ganglia (DRG), presumably via a dorsal root reflex. Within the cardiac sympathetic nerve endings, vesicles containing neurotransmitters including bradykinin (BK) sit “docked” and ready at the synaptic membrane. ES triggers a nerve impulse that leads to the release of the neurotransmitters BK and norepinephrine (NE). BK and NE then bind to their cognate receptors, activating PKC-α and mediating the cardioprotection against ischemia/reperfusion injury that defines nociceptor-induced conditioning (NIC). β-AR, β-adrenergic receptor.

Both clinical and basic science research evidence demonstrates that electrical spinal cord stimulation improves cardiac functional status, relieves symptoms in patients with refractory angina pectoris, and reduces infarct size associate with spinal cord and cardiac sympathetic nerve activation (53, 54). Spinal cord stimulation involves implantation of a stimulator into the spine, which cannot be done on an emergent basis for cardiovascular disease. Furthermore, these stimulators become dislodged at a significant rate. Finally, spinal cord stimulation was only effective as a preconditioning stimulus, and the infarct sparing effect was not as strong as ES-induced cardioprotection, which is equally effective when performed before ischemia (preconditioning) or at reperfusion (postconditioning). Similar protective effects are under study for remote postconditioning by inflation of blood pressure cuffs on the appendages. This approach has both neural and humoral components to its mechanism but seems to require active innervation of the limb for the production of humoral factor, which then acts on the heart (39, 41). This approach has limitations in that some patients cannot undergo the procedure, and clinical efficacy has been elusive. The effect may also be limited by distribution of humoral factor and the fact that it cannot act on ischemic muscle until after the vessel occlusion is resolved. We report that direct sensory nerve stimulation of a particular peripheral nerve field can elicit some of the same effects and can be applied in a noninvasive, nontraumatic manner. By nontraumatic, we mean that the stimulus does not cause tissue injury or significant pain (verified using the same stimulus in electroacupuncture and subjective testing by authors on themselves). As we have shown, the CNS is not involved, and spinal transection results support that the neural signal acts directly on the heart via reflex. If a humoral signal is released, it is likely released by the heart and may act elsewhere in extending protection to other organs; this is actively under study. The cardioprotection induced by NIC is expected to protect the myocardium independent of reperfusion, meaning that reestablishment of coronary flow is not required for the neurosignal to reach ischemic muscle. This may underlie the high potency of NIC against MI in the murine model.

In conclusion, this study demonstrates that transdermal ES at the anterior abdominal T9−T10 level activates C-fiber sensory nerves in the dermis, which elicits a dorsal root reflex that initiates, by direct neurohormonal activity in the myocardium, an extremely strong cardioprotective effect against MI, termed herein NIC. Clinically, if these discoveries are translatable to humans, one could anticipate an electrical device that could be applied to the anterior abdominal region during a heart attack or at reperfusion (postconditioning). It is possible that NIC could be performed at the location of an MI occurrence or in the ambulance and could extend the golden hour, salvaging sufficient myocardium to make the difference between life and death in some patients, and may reduce morbidity in many more. In any case, an effective adjunct therapy to PCI that is noninvasive and easy to deploy could improve outcomes after MI and may have effects on other remote organs as well.

Limitations.

Although we have demonstrated that NIC induces cardioprotection as a preconditioning and postconditioning stimulus, most of the experiments presented in this initial mechanistic study were performed with the preconditioning stimulus since certain procedures would be difficult to perform at the time of reperfusion in the surgical model. Judging by the mechanistic similarities between ischemic pre- and postconditioning, it is highly likely that the proposed mechanism pertains to postconditioning, although this will have to be verified in large animal models. Although these experiments were performed using acupuncture needles, they were deployed only into the skin and are not intended to work as traditional acupuncture stimulus (see methods for details). The general PKC inhibitor used was relatively nonisoform selective, and future studies with isoform-specific inhibitors will have to be performed to map out the downstream signaling cascades involved in cardioprotection; however, this was beyond the scope of the present study. Finally, there are limitations inherent with rodent models. Many cardioprotective therapies that have worked in rodents and even in large animals have not worked well in humans. This is likely due to multiple differences including genetic and physiological homogeneity of the models, which is not a feature of human populations, physiological differences between humans and model species, human medications, and the exact timing and circumstances of MI (including the timing and location and extent of coronary occlusion). However, the basic finding that RPCT can elicit cardioprotection has been confirmed in four species to date (10, 23, 24, 30, 40), and the highly conserved neuroanatomy support that attempts should be made to translate NIC to the clinic.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grant 5-R01-HL-091478 (to W. K. Jones).

DISCLOSURES

W. K. Jones and N. L. Weintraub have an equity interest in CardioCeption, LLC, a startup focused on bringing nociceptor-induced conditioning to clinical trial, and hold >5% equity but to date receive no salaries or funds from this effort.

AUTHOR CONTRIBUTIONS

W.K.J. conceived and designed research; X.R., A.E.R., T.L.L., L.H., F.M., Y.L., M.T., A.E.R., T.L.L., L.H., F.M., Y.L., M.T., M.R., G.-C.F., J.-M.Z., E.K., A.A., S.E.K., M.J., S.D., A.C., J.R., and W.K.J. analyzed data; X.R., A.E.R., T.L.L., L.H., F.M., M.T., M.R., S.D., A.C., J.R., N.L.W., and W.K.J. interpreted results of experiments; X.R., F.M., Y.L., M.R., G.-C.F., J.-M.Z., S.E.K., M.J., S.D., A.C., and W.K.J. prepared figures; X.R., L.H., Y.L., and W.K.J. drafted manuscript; X.R., A.E.R., T.L.L., L.H., M.T., W.R.X., G.-C.F., J.-M.Z., E.K., A.A., S.E.K., S.D., A.C., J.R., N.L.W., and W.K.J. edited and revised manuscript; X.R., T.L.L., L.H., F.M., Y.L., M.T., M.R., W.R.X., G.-C.F., J.-M.Z., E.K., A.A., S.E.K., M.J., S.D., A.C., J.R., N.L.W., and W.K.J. approved final version of manuscript.