|Home | About | Journals | Submit | Contact Us | Français|
Although remote ischemic stimuli have been shown to elicit cardioprotection against ischemia/reperfusion injury, there is little known about the effects of nonischemic stimuli. We previously described a remote cardioprotective effect of nonischemic surgical trauma (abdominal incision) called remote preconditioning of trauma (RPCT). In the present study, we elucidate mechanisms underlying this phenomenon.
We used a murine model of myocardial infarction to evaluate ischemia/reperfusion injury, and either abdominal surgical incision, or application of topical capsaicin, to elicit cardioprotection. We show that the cardioprotective effect of RPCT is initiated by skin nociception, and requires neurogenic signaling involving spinal nerves and activation of cardiac sensory and sympathetic nerves. Our results demonstrate bradykinin-dependent activation and repression, respectively, of PKCε and PKCδ in myocardium after RPCT, and we show involvement of the KATP channels in cardioprotection. Finally, we show that topical application of capsaicin, which selectively activates C sensory fibers in the skin, mimics the cardioprotective effect of RPCT against myocardial infarction.
Nontraumatic nociceptive preconditioning represents a novel therapeutic strategy for cardioprotection with great potential clinical utility.
Cardiac ischemia/reperfusion (I/R) injury contributes significantly to morbidity and mortality throughout the world.1 Over the past 2 decades, various strategies for protecting the heart against myocardial infarction (MI) and I/R dysfunction have been developed.2–4 Ischemic preconditioning (IPC), especially, has been extensively studied in multiple species and in the clinical setting.5–9 Importantly, remote IPC that results from brief episodes of ischemia occurring at a distant organ site has been shown to be cardioprotective.4,10,11 Previously, we observed that infarct size after in vivo I/R was altered by nonischemic surgical trauma. Vascular surgery performed to catheterize the carotid artery increased infarct size, whereas transverse abdominal incision resulted in a significantly decreased infarct size.12 To describe this nonischemic preconditioning phenomenon, we coined the term “remote preconditioning of trauma” (RPCT).
It has been demonstrated that bradykinin (BK), adenosine, opioids, and norepinephrine (NE) all have roles in remote IPC.10,13–16 BK is one of several oligopeptides called kinins that are produced by sympathetic nerve endings (ie, synaptosomes), myocytes, and endothelial cells in the heart.17–19 The actions of BK are mediated by two major receptor subtypes, BK receptors 1 and 2 (BK1R and BK2R).20,21 BK1R are inducible by inflammatory stimulation or tissue injury, and BK1R seem to play an injurious role in myocardial I/R.22 BK2R is constitutively expressed and mediates most of the physiological actions of kinins. Several studies demonstrate that activation of BK2R is involved in both IPC and in remote IPC.10,23 Previous studies demonstrate that endogenous BK activates sympathetic cardiac afferents during I/R and that this reduces cardiac dysfunction and MI.10,24–26 Calcitonin gene-related peptide (CGRP) and substance P are released along with BK from sensory nerves and can act on cardiac sympathetic nerves to provoke NE release.17,24 There is evidence that BK-induced protection requires protein kinase C (PKC) activation.27,28 In particular, PKC-δ has been shown to be a critical mediator of postischemic cardiomyocyte necrosis and contractile dysfunction after I/R, and PKCε is a mediator of cardioprotection.27,29–31
There is currently nothing known about the molecular initiators that instigate cardioprotection after RPCT. The fact that cardioprotection against MI mediated by RPCT does not require TNF-α 12,13 suggests that the mechanism is not the same as IPC.32 To delineate the mechanism of RPCT, we performed pharmacological, genetic, biochemical, and physiological analyses in this study. Our results show that an abdominal incision elicits cardioprotection against MI via stimulation of peripheral nociception. Nociception triggers neurogenic signaling via spinal nerves, which activates the sympathetic nervous system in the heart and elicits activation of PKCε and inhibition of PKCδ in a BK2R-dependent manner. Activation of the mitochondrial KATP (mitoKATP) channels is required for cardioprotection. Direct activation of C sensory fibers in skin using topical capsaicin mimics the cardioprotective effect of RPCT, supporting that peripheral nociceptor stimulation has great clinical potential.
Mice were maintained in accordance with institutional guidelines, the Guide for the Care and Use of Laboratory Animals (NIH, revised 1985), and the Position of the American Heart Association on Research Animal Use (1984). Wild-type (B6129SF2/J F2) and BK2 receptor knockout mice (B6129SF2/J F2, strain 101045) were obtained from the Jackson Laboratories (Bar Harbor, Me). All groups of mice consisted of males and females distributed equally among groups; post hoc analyses confirmed previous results that there were no gender-related differences in these studies.
Mice were subjected to surgical protocols as delineated in Figure 1. A minimally traumatic mouse model was used for in vivo studies of I/R injury and RPCT as described previously.12,33–35 All mice were continuously monitored by electrocardiography, and mice without evidence of ischemia and timely reperfusion were excluded from the studies (1%); survival in this study was 96%. Coronary occlusion was for 45 minutes. In experiments in which infarct size was the end point, infarct size was determined at 4 or 24 hours after reperfusion, as previously described,6,12 and is presented as area of the infarct normalized to the area of the region at risk. For all studies, the region-at-risk was not significantly different between groups (Table 1, supplement). The 4-hour time point was used only in the spinal transection and related control studies to prevent the mice from regaining consciousness in that study, for ethical reasons. Abdominal incision was used as the nonischemic stimulus for RPCT as previously described.12 The incision was through the skin, subcutaneous, fat, muscle, and peritoneum, and was 2 cm in length through the abdominal midline; we refer to this as the RPCT stimulus. Afterward, the incision was sutured immediately using 7-0 polypropylene sutures. For skin incision, the incision was made anatomically in the same location but care was taken to cut the skin only. Sham control groups were used for the abdominal and skin incisions. In these groups, mice were subjected to intubation and anesthesia as described, the skin was sterilized (Betadine scrub), and the abdomen was shaved, just as in the experimental groups; the only difference was that the incision was not made.
Nociception-induced preconditioning was elicited by application of topical capsaicin (≈150 μL of 0.1%; Chattem, Inc) along the same line as surgical incision, using a mask that exposes a strip of skin 2×1 cm through the umbilicus. Sham control mice for this study were treated exactly the same, except with a control gel base without capsaicin.
To determine the role of the central nervous system versus spinal ganglia, we transected the spinal cord at different levels 15 minutes before the abdominal incision. Mice were anesthetized and bone markers were identified at the thoracic vertebra C7 and T7 levels. The spinal cords of C7 and T7 were transected by a sharp knife with limited bleeding under a microscope, followed by immediate abdominal incision, I/R 15 minutes later, and infarct size determination 4 hours later. Sham controls were subjected to the same surgical manipulations and procedures (ie, anesthesia, shaving, opening the skin and muscle, exposing the spinal column), but without actually cutting the cord.
To track the connection between nerve cells from the sensory fibers of the skin to the spine, a fluorescent dye, 1,1′-dioactadecyl-3,3,3′,3′-tetramethylindocarbodyanine perchlorate (Dil),36 was injected subcutaneously at the abdominal incision level (thoracic vertebra T9–T10 of spine). One week after the injection, mice were perfused with 4% formaldehyde. The fixed spinal cord and dorsal root ganglia at thoracic vertebra T9–T10 and T1–T5 levels were dissected. The sectioned spinal cord (40 μm) and whole-mount dorsal root ganglion were placed under a confocal microscope for direct visualization and image capture.
Details of treatment with pharmacological agents, dosage, route, suppliers, and vehicle controls are provided in the supplement, because of space limitation. Briefly, the agents used were the BK2R-antagonist Hoe140 (50 μg/kg, intravenous),10,23,25 the ganglionic blocker hexamethonium (20 mg/kg, intravenous),10,37 KATP channel inhibitors 5HD (100 μmol/L/kg, intravenous),29,38 glibenclamide (0.3 mg/kg, intravenous),27 and CGRP antagonist CGRP5-37 (3 nmol/kg),17,39 and propanolol (2 mg/kg).16,40 Lidocaine (100 μL of 1%, in saline) was administrated by subcutaneous injection at the abdominal level 5 minutes before skin incision. All injectables, except glibenclamide, were dissolved in physiological saline and vehicle controls were saline. Glibenclamide and vehicle controls were 10% DMSO in saline (supplement). Capsaicin (0.1%; Chattem, Inc) was applied topically (150 μL over a 1×2 cm area; supplement). In all studies, vehicle control experiments were performed with the specific vehicle required for each agent referenced.
Group size was determined by Power analysis, as described.12,42 For parameters that require quantification and evaluation for statistical significance, results were expressed as mean±SEM. Statistical significance (P values) was determined using the Student t test (2-tailed distribution and 2-sample unequal variance) with the Bonferroni correction. For multiple group comparisons, 1-way analysis of variance followed by Fisher post hoc test was used. P≤0.05 was considered statistically significant.
We observed a 5- to 6-fold reduction in infarct size after RPCT (Figure 2A; 55.3±3.4% sham, versus 10.2±6.3% RPCT; P≤0.05; n=11). We assessed the extent of apoptosis in both sham and RPCT hearts after I/R, using in situ end labeling of DNA fragmentation (TUNEL staining) and an ELISA-based nucleosome assay. TUNEL results revealed, relative to shams, a significantly decreased proportion of TUNEL-positive nuclei in the myocardium of mice subjected to RPCT (Figure 2B, 2C; 33.6±2.5% versus 4.2±0.01%; P≤0.05; n=4). Results of the quantitative nucleosome assay showed that levels of mononucleosomes and oligonucleosomes were significantly decreased, relative to sham, after RPCT (Figure 2D).
To determine the possible role of a neurogenic pathway, we first addressed whether sympathetic ganglionic transmission is required for cardioprotection against MI by RPCT. Results show that administration of the sympathetic ganglionic blocker hexamethonium (20 mg/kg) before the RPCT stimulus abrogated the protective effect of RPCT against MI (Figure 3A; 9.5±1.2 RPCT versus 55.9±1.9 RPCT hexamethonium; P≤0.05; n=7).
Next, we transected the spinal cord at 2 different levels, vertebral levels C7 and T7, before RPCT (Figure 3B). Transection of the spinal cord at C7 had no effect on RPCT (Figure 3B white bars; 12.1±3.1%, sham versus 10.3±2.2%; C7 transection; P≤0.05; n=4). However, transection at T7 abolished the cardioprotective effects of RPCT against MI (Figure 3B black bars; 13.4±2.5%, sham, versus 53.4±2.8%; T7 transection; P≤0.05; n=6).
We then tested whether the abdominal incision, if limited to the skin layer (location of sensory nerves), was still capable of triggering cardioprotection. The results (Figure 3C) show that incision of skin alone is sufficient to elicit cardioprotection against MI (7.3±3.2% wall incision versus 10.4±2.5% skin incision; P>0.05, n=6). Next, we pretreated the abdominal incision site with lidocaine (1%) 15 minutes before abdominal incision. This treatment completely abrogated the cardioprotective effect of RPCT against MI (Figure 3D; 54.0±3.9 lidocaine RPCT versus 7.2±2.5 RPCT vehicle; P≤0.05; n=7) supporting a critical role for pain sensation (nociception) in initiating RPCT.
It is known that sensory fibers from the abdominal skin project to the spinal cord at the vertebral T9–T10 levels, whereas the sensory nerves innervating the heart project to the vertebral T1-5 level43,44 (Figure 3A). Therefore, we investigated the possibility that sensory nerves originating in the skin or muscle underneath the skin at the incision site may connect to a higher level in the spinal cord. We found that dye injection (Dil) in skin at the abdominal midline (T9-10 vertebral level) labeled spinal neurons of the dorsal horns at both the vertebral T1–T5 level (Figure 3E), as well as at the T9-10 level.
To elucidate whether RPCT requires BK/BK2R signaling, we treated mice with the BK2R-selective antagonist Hoe140 (50 μg/kg) 15 minutes before abdominal incision. Inhibition of BK2R, relative to vehicle-treated RPCT controls, abolished the protective effect of RPCT on infarct size (Figure 4A; 10.4±2.5% RPCT vehicle versus 52.4±2.4% RPCT Hoe140; P≤0.05; n=7). Furthermore, RPCT did not have a cardioprotective effect against myocardial infarction in BK2R knockout mice (Figure 4B; 55.3±3.4% BK knockout sham versus 54.8±2.4% BK knockout RPCT; P>0.05; n=7).
To assess the role of β-adrenergic signaling after RPCT, we treated groups of mice with propanolol (2 mg/kg) or vehicle and subjected them to RPCT, followed by 45 minutes of coronary occlusion 15 minutes later (Figure 4C). Analysis of infarct size assessed 24 hours later demonstrated complete blockade of the cardioprotection afforded by RPCT against MI (54.4±4.7 propanolol versus 10.4±5.7 vehicle; P≤0.05; n=7). To determine the involvement of sensory nerve transmission, we assessed the requirement for CGRP using the antagonist CGRP5-37 (3 nmol/kg; Figure 4D). Blockade of CGRP prevented cardioprotection against MI after RPCT (53.2±3.8 CGRP5-37 versus 7.2±5.2 vehicle; P≤0.05; n=7).
Groups of mice were treated with chelerythrine (5 mg/kg, intravenous) or vehicle 15 minutes before RPCT (Figure 4E) and infarct size assessed (45-minute ischemia, 24-hour reperfusion). The results demonstrate that blockade of PKC abrogates the cardioprotection afforded by RPCT (58.3±4.5 chelerythrine versus 9.5±1.2 vehicle; P≤0.05; n=7).
We next determined the alterations of PKC activation in myocardium after RPCT, measured by the ratio of the membrane-associated to cytoplasm-associated fraction of PKC. Quantitative immunoblotting demonstrated that PKCε is activated, whereas PKCδ activity is repressed 15 minutes after the RPCT stimulus (Figure 5A–D). There was no effect on PKCα activity. Importantly, the results of similar experiments using BK2R knockout hearts demonstrated that ablation of BK2R prevented the RPCT-induced effects on PKCε and PKCδ (Figure 5E–H).
Administration of 5HD (100 μg/kg, intravenous)27,38 completely eliminated the protective effect of RPCT against MI (Figures 6, ,8;8; 1±1.5 RPCT saline versus 51.3±1.6 RPCT 5HD; P≤0.05; n=8), whereas cardioprotection was partially abrogated by glibenclamide (300 μg/kg, intravenous; 8.1±1.5 RPCT saline versus 38.9±2.8 RPCT glibenclamide; P≤0.05; n=7).
To determine whether direct chemical stimulation of sensory C-fibers in the skin elicits cardioprotection similar to that of RPCT, we applied capsaicin topically (0.1%) to the abdominal midline, along the same line used for the abdominal incision. The mice were subjected to a 45-minute coronary occlusion/reperfusion, and infarct size was measured 24 hours later (Figure 7). Infarct size was significantly reduced by topical capsaicin (51.2±1.46 sham versus 7.48±1.99 capsaicin; P≤0.05; n=6).
There is recent evidence supporting the cardioprotective effects of remote ischemic stimuli and the mechanisms by which these produce cardioprotection.4,10,11,13 However, there is nothing known concerning the effect of remote nonischemic stimuli, including surgical injury, on cardioprotection or myocardial I/R injury. We published the first report12 to our knowledge describing the effects of remote nonischemic surgical stimuli on cardiac I/R injury and showed that, depending on the site, surgical incisions can be either cardioprotective or injurious. We showed that an abdominal surgical incision elicits a preconditioning effect with early and late phases12 of protection against MI. We named this phenomenon RPCT. We determined that RPCT reduces infarction, in part by reducing apoptosis (Figure 2), and that protection extends to ventricular dysfunction after I/R (see supplement). We also have evidence that there is a postconditioning phenomenon related to nociceptive stimulation (supplement). Herein, we delineate major aspects of the mechanism underlying RPCT.
Although some evidence supports that a neurogenic pathway is involved in the cardioprotection of remote IPC,10,45 other evidence implicates diffusible humoral factors.32,37 We report that administration of hexamethonium, a ganglionic blocker that inhibits impulse transmission from the preganglionic neurons to the postganglionic neurons of both the sympathetic and parasympathetic systems, abrogates the protection of RPCT against MI (Figure 3A), supporting a neurogenic mechanism. Our results with spinal transection (Figure 3B) also support a neurogenic mechanism and rule out an essential diffusible humoral factor as the cause of cardioprotection after RPCT. We also demonstrate that a shallow skin incision is sufficient to initiate RPCT (Figure 3C). Our result (Figure 3D) that lidocaine completely blocks RPCT supports that peripheral nociception via skin sensory fibers is required for RPCT. These peripheral nerves are essentially the axons of the dorsal root ganglion neurons. We propose that after nociceptive stimulation, peripheral nerve depolarization leads to a dorsal root reflex at the T9-10 vertebral level of the spinal cord, the level of the abdominal incision. Our results with spinal transection (Figure 3B) demonstrate that RPCT after stimulation of sensory nerves at the T9-10 vertebral level requires an intact spine to the T7 level. That spinal integrity above C7 is not required demonstrates that the central nervous system is not involved in RPCT. The dorsal root reflex at T9-10 likely activates spinal nerves at higher levels, leading to activation of the cardiac nerves, which mediate the cardioprotection. This is consistent with evidence that a dorsal root reflex can activate the dorsal horn neurons at higher levels of the spinal cord.45 Further, it is known that action potentials can travel antidromically along the dorsal root and activate the sensory fibers innervating the heart,46,47 that sensory fibers in the heart activate the cardiac sympathetic nervous system,48 and that cardioprotection is elicited by activation of adrenergic neurons after spinal cord stimulation.45 To support the dorsal root reflex hypothesis, we performed retrograde nerve tracing studies with Dil to determine if the sensory nerves along the abdominal midline connect to the dorsal root at T9-10 and T1-5. We did see fluorescence at both levels (Figure 3E). These results are consistent with 2 interpretation: (1) that a dorsal root reflex activates sensory neurons innervating the heart and stimulates the cardiac sympathetic system as suggested and as supported by other studies;10,45–47 or (2) that a dorsal root reflex may directly activate preganglionic neurons in the lateral horn of the spine that activate the sympathetic nervous system. Future studies will test functionally the involvement of the dorsal root reflex in RPCT.
To summarize the neurogenic aspect of our findings, peripheral sensory nerve stimulation leads, via spinal nerves, to activation of the cardiac sensory nerves, which, in part through CGRP release, stimulate cardiac sympathetic nerves. Release of NE and BK from the sympathetic nerves underlies cardioprotective response in the myocardium after RPCT.
BK is both a hormone and a neurotransmitter, and it is secreted from sympathetic nerves in the heart.45,48 BK/BK2R has been shown to be an important mediator of remote IPC10,23 and is known to trigger NE release from cardiac sympathetic nerves.10,48 Release of substance P and CGRP from afferent nerves can act on sympathetic nerves to stimulate release of NE and BK, and both can act on cardiomyocytes, which possess both βAR and BK2R.17 Both of these G-coupled receptor systems have been shown to activate PKC in cardiomyocytes; particularly, PKCε is known to be an essential mediator of IPC.28 Our results demonstrate that PKCε and PKCδ are activated and repressed, respectively, and that PKCα is unaffected by a RPCT stimulus. PKCε is thought to work, at least in part, through repression of proapoptotic pathways including via activation of KATP channels.27,29 Conversely, PKCδ has been shown to be pro-cell death.30,31,49 Our results demonstrate that both the activation of PKCε and inhibition of PKCδ after RPCT are BK2R-dependent (Figure 5), and that the cardioprotection against MI of RPCT is PKC-dependent, BK2R-dependent (Figure 4), and that the action of the mitoKATP and perhaps the sarcKATP channels (Figure 6) are required for RPCT. These results are consistent with a PKC-mediated cardioprotection elicited by neurogenic stimulation of cardiomyocytes after RPCT.
Our results, interpreted in the light of recent discoveries by others, cited herein, support a proposed mechanism of RPCT (Figure 8). We propose that nociceptive stimulation of sensory nerves in the skin of the abdomen triggers a neurogenic signal (initiator of RPCT) that is transmitted via nerve fibers and causes a dorsal root reflex that activates spinal nerves higher in the spinal cord, ultimately leading to activation of the cardiac sympathetic nervous system. Most likely (and supported by our results with CGRP blockade), the stimulation of cardiac sensory nerves triggers the activation of the cardiac sympathetic system, which involves NE and BK release and activation of βAR and BK2R in myocardium. Finally, the signal results in activation of PKCε and mitoKATP and inhibition of PKCδ, which together mediate cardioprotection against MI (mediators of RPCT; Figure 8).
The cardioprotective effect of RPCT against MI is the most powerful noted to date in the mouse (80% decrease in infarct size). Although PKCε and PCKδ have been previously implicated in cardioprotection, the clear-cut activation of a protective isoform (PKCε) and repression of a presumably injurious isoform (PKCδ) are unique to RPCT thus far. The ability of a nonischemic surgical stimulus to elicit this form of cardioprotection is completely novel. Although the mechanism may in some aspects be similar to cardioprotection afforded by remote IPC and spinal stimulation,10,28,45 the protection that we observe is much more powerful (80% reduction in infarct size compared to 45% reduction), is TNF-α–independent,12 and therefore is mechanistically different from IPC. As we demonstrate (Figure 7), remote nociceptive stimulation can be accomplished by chemical stimulation (capsaicin) of skin sensory nerves. This and the fact that spinal cord stimulation for angina was found to be effective and provided functional benefit46,47,50 support that this study is of high clinical relevance. Capsaicin is FDA-approved, inexpensive, widely available, and used topically to treat pain. Most importantly, topical capsaicin has no known serious adverse effects and could be easily applied in an ambulance or emergency room setting, well in advance of coronary reperfusion. If proven efficacious in humans, this simple therapy has the potential to reduce myocardial injury in the setting of I/R, thereby reducing the extent and consequences of acute MI.
Sources of Funding
This study was supported by NIH grants HL63034 and HL091478 (Jones), HL-087861 (Fan), NS55860 (Zhang), HL076684 and HL62984 (Weintraub), and NS45594 (Zhang).
Presented in part at American Heart Association Scientific Sessions 2008, November 8–12, 2008, New Orleans, La.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/cgi/content/full/120/11_suppl_1/S1/DC1.
Drs Jones, Ren and Weintraub are coinvestors of provisional patent entitled “Methods of Preventing Ischemic Injury Using Peripheral Nociceptive Stimulation.”