C2 spinal cord stimulation induces dynorphin release from rat T4 spinal cord: potential modulation of myocardial ischemia-sensitive neurons

Abstract

During myocardial ischemia, the cranial cervical spinal cord (C1–C2) modulates the central processing of the cardiac nociceptive signal. This study was done to determine 1) whether C2 SCS-induced release of an analgesic neuropeptide in the dorsal horn of the thoracic (T4) spinal cord; 2) if one of the sources of this analgesic peptide was cervical propriospinal neurons, and 3) if chemical inactivation of C2 neurons altered local T4 substance P (SP) release during concurrent C2 SCS and cardiac ischemia. Ischemia was induced by intermittent occlusion of the left anterior descending coronary artery (CoAO) in urethane-anesthetized Sprague-Dawley rats. Release of dynorphin A (1-13), (DYN) and SP was determined using antibody-coated microprobes inserted into T4. SCS alone induced DYN release from laminae I–V in T4, and this release was maintained during CoAO. C2 injection of the excitotoxin, ibotenic acid, prior to SCS, inhibited T4 DYN release during SCS and ischemia; it also reversed the inhibition of SP release from T4 dorsal laminae during C2 SCS and CoAO. Injection of the κ-opioid antagonist, nor-binaltorphimine, into T4 also allowed an increased SP release during SCS and CoAO. CoAO increased the number of Fos-positive neurons in T4 dorsal horns but not in the intermediolateral columns (IML), while SCS (either alone or during CoAO) minimized this dorsal horn response to CoAO alone, while inducing T4 IML neuronal recruitment. These results suggest that activation of cervical propriospinal pathways induces DYN release in the thoracic spinal cord, thereby modulating nociceptive signals from the ischemic heart.

myocardial ischemia activates multimodal sensory afferent neurites that respond to changes in the chemical and mechanical milieu of the heart (9, 13, 28). The high cervical spinal cord (at the C1–C2 level) plays an important role in modulating the transmission of the ischemic cardiac afferent sensory signal within the thoracic spinal cord via descending inhibitory propriospinal pathways (9, 13, 14). High cervical spinal cord stimulation (SCS) has been shown to be effective in treating refractory angina with no detrimental effects on cardiac function (14, 23, 26). Our laboratory has investigated the neurochemical mediation of the cardiac ischemic signal and the effects of electroneuromodulation on the release of some of the neuropeptides that may be involved with the transmission of this signal (10, 19–21). Recent evidence indicates that substance P (SP) is released from dorsal horn laminae in the thoracic spinal cord in response to cardiac ischemia (10, 21), induced by brief coronary artery occlusion and that this peptide serves as one of the mediators of the cardiac nociceptive signal. We also found that preemptive C2 SCS reduces SP release (10) and SP mRNA levels in the T4 spinal cord (11), and it decreases neuronal activation, assessed using c-Fos, in the T4 spinal cord during transient cardiac ischemia (10).

Previous studies showed that direct electrical stimulation of the upper cervical spinal cord (C1–C2) decreased the electrical activity of spinothalamic tract neurons activated by cardiac ischemia-sensitive afferent neurites, regardless of the type of stimulus, exciting the cardiac ischemia-sensitive neurites (13, 25, 27). Other forms of electroneuromodulation (i.e., vagal afferent stimulation) have been shown to excite cranial cervical propriospinal neurons (8, 13, 32), inhibit neurons in the upper thoracic spinal segments (1, 14, 20), and suppress coronary artery occlusion-induced release of SP from the T4 spinal cord (19), while causing the release of dynorphin (19). These observations led to the hypothesis that electrical neuromodulation at the C2 level, whether initiated by vagal afferent stimulation or direct spinal cord stimulation, attenuates the nociceptive signal (i.e., SP release) at the thoracic level, in part, via the release of an analgesic neuropeptide (such as dynorphin) at the thoracic level. Such an inhibitory peptide could be released either directly from the activated descending propriospinal pathways projecting onto the sensory fiber endings in the superficial thoracic spinal cord or from descending signals projecting from supraspinal sites onto thoracic spinal cord neurons.

The objectives of this study were threefold: to determine 1) whether high cervical spinal cord stimulation induces the release of dynorphin from dorsal horn sites in the thoracic spinal cord, 2) whether the source of the dynorphin is cervical propriospinal neurons, and 3) whether chemically inactivating C2 neurons, one possible source of this dynorphin, alters the pattern of SP release from sites in the thoracic dorsal horn during myocardial ischemia in the presence of C2 SCS. Secondarily, c-Fos histochemistry of the T4 spinal segment and caudal brain stem was used to determine whether the pattern of neuronal activation (including preganglionic sites of the intermediolateral cell columns) was altered by the disruption of the pathways originating from the cervical spinal cord during SCS and cardiac ischemia.

METHODS

Surgical preparation of animals.

Male Sprague-Dawley rats (n = 38, 277.6 ± 4.1 g body wt; Harlan, Indianapolis, IN) were used in these studies. All procedures and experimental protocols were reviewed and approved by the East Tennessee State University Institutional Committee on Animal Care and Use and conformed to the Animal Welfare Act, according to the Public Health Policy on Humane Care and Use of Laboratory Animals and adhere to the APS's Guiding Principles in the Care and Use of Animals. Rats were anesthetized with urethane (1.5 g/kg body wt) injected intraperitoneally. Supplemental injections (30 mg/kg) of urethane were made every 90 min (or more frequently, if needed) throughout the entire experiment through the left femoral vein to maintain surgical-level anesthesia. Rats were surgically prepared as described before (19–21). Cardiovascular data [pressure pulse, mean arterial pressure (MAP; mmHg), the ECG, and heart rate (HR; beats/min, bpm)] were recorded and analyzed via PowerLab (ver. 4.24) and are presented as means ± SE. Data are reported for baseline, resting levels, steady-state levels during an experimental intervention, and recovery levels. Only one experimental intervention was performed on a group of animals (Fig. 1; sham, coronary occlusion, cervical spinal cord stimulation with or without coronary artery occlusion, spinal cord stimulation with or without ibotenic acid, etc.). For hemodynamic data, significance was determined from Student's t-test for paired data within a specific experimental group and using one-way repeated-measures ANOVA between groups with follow-up comparisons using the Tukey test. P ≤ 0.05 was taken as the minimum level of significance.

Fig. 1.

Protocols for the in situ experiments, in which rats in each respective group were exposed to spinal cord stimulation (SCS: gray boxes), with or without transient, intermittent occlusion (solid boxes) of the left anterior descending coronary artery (LAD CoAO). Solid dashed lines indicate placement of either DYN A (1-13)- or SP-antibody-coated microprobes (Prb) for 10-min periods. Baseline, rest probes (Prb 1–3) were inserted for 10 min each prior to experimental intervention. For SCS alone, electrical stimulation (50 Hz, 200 μs, 90% muscle threshold) was delivered to the left-sided dorsal aspect of the C2 spinal cord for a continuous 45 min. For CoAO, the LAD was occluded for 90 s with 60 s occlusion off, repeated for 10 min. At the end of 10-min period, a new microprobe was inserted into the left T4 spinal cord and the intermittent CoAO cycle repeated two more times (i.e., Prb 4–6). For SCS+CoAO, stimulation of the C2 spinal cord was initiated 15 min prior to beginning the occlusion protocol (total time for SCS with CoAO, 45 min). For protocols involving ibotenic acid (IBO) injections, the IBO was injected bilaterally into the C2 spinal cord 60 min before initiating either the SCS or SCS+CoAO protocols. For protocols involving nor-binaltorphimine (BNT), the BNT was injected bilaterally into the T4 spinal cord 10 min before placing Prb 1.

Laminectomy and microprobe placement.

The spinal cord from C1–C3 and from T2–T4 was exposed on all of the rats by removing the appropriate vertebral processes as described previously (10).

Each of the in vivo probes was positioned at the rostro-caudal midpoint of the T4 spinal segment with the aid of a digital micropositioner (Stoelting) and a stereotaxic surgical microscope with color video display. To ensure repeatability of their placement, each probe was visually positioned at the midline and surface of the spinal cord, moved 0.5 mm left of the midline, and inserted to a depth of 2 mm. Each probe remained in situ for 10 min. New probes were used for each 10 min pre- or experimental procedure, and these were designated as “rest” (Fig. 1, probes 1–3), spinal cord stimulation (SCS), or coronary occlusion with spinal cord stimulation (SCS+ CoAO) (Fig. 1, probes 4–6). At the completion of the 10 min in situ time, probes were withdrawn from the spinal cord, washed briefly in ice-cold PBS and incubated with radiolabeled DYN or SP (0.01 μCi/5 μl PBS, 7.4) for 24 h at 4°C and then processed as described before (10, 12, 19, 21). Probe placement was verified, as described previously (10, 19, 21), by visualizing the deposit of 20 nl Pontamine blue dye in a coronal section of T4 that was injected through the final microprobe.

Measurement of immunoreactive DYN or SP using immobilized antibody microprobe technique.

The release of endogenous immunoreactive-DYN or -SP from sites in the thoracic spinal cord was measured using the antibody-coated microprobe technique, as previously described (10, 12, 19, 21). In all experiments where probes were inserted into the thoracic spinal cord, a set of control probes (designated as in vitro probes) was identically and simultaneously prepared as the in vivo probes. The in vitro probes were used to determine the sensitivity of the binding of radiolabeled ligand [¹²⁵I-Tyr⁸ SP, or ¹²⁵I-Dyn A (1–13) Phoenix Pharmaceuticals] (see 18, 19, 21) and to confirm the uniformity of binding of the silane and antibody along the probe shaft.

Image analysis of the microprobes.

Autoradiographic images of the microprobes were analyzed for patterns of inhibition of binding of the radiolabeled ligand along the length of the probe. Such inhibition is indicative of where unlabeled (e.g., endogenously released) DYN or SP bind to their antibodies on the respective probes during the in vivo exposure time. Since the position of the probe tips was marked by a deposit of blue dye in the spinal cord, sites of release of the peptide could be determined from differences in the optical densities of the probe images on the autoradiographic film. The analysis was carried out based on initial methods described by Hendry et al. (18) and modified by this laboratory (10, 19, 21). A computerized image analysis system (MCID, Imaging Research, Canada) was used to linearly integrate a 4-mm length of the probe images, starting from the probe tip. Background grayness, due to the exposed X-ray film alone, was subtracted from each pixel of the probe image.

In the diagrams presented, the mean optical density of the probe image is converted to a gray scale in arbitrary units of 0-1026 (with 1026 being the darkest gray level). Each probe image was analyzed for 4 mm: the first 2 mm, starting at the tip, corresponded to the segment of the probe inserted into the spinal cord (designated 2 to 0); while the next 2 mm corresponded to the part of the probe that remained outside the spinal cord (designated as “0 to –2 mm”). The 2 mm outside the spinal cord served as an internal control area along each probe and for between-group controls via comparison to similar segments on the in vivo and in vitro probes. Typical pseudocolor images of microprobes were presented previously (19, 21), together with a gray level color-equivalent scale (36). The data presented in the image analysis figures are given as the mean gray levels ± SE for each specified group of probes. Differences in the patterns of binding of radiolabeled DYN or SP along the probes during various experimental interventions were determined by Student's t-test. The calculated T value, where P = 0.05 (the minimum level of significance taken), is plotted along the lower portion of the image analysis graphs (just above the abscissa). The T-value for each pixel along the analyzed image was calculated and plotted in relation to the T-value of P = 0.05. Any points along the length of the probes that were different from each other appear above the T-value line. Because the resolution of detecting a meaningful difference in the binding of radiolabeled peptide is 100 μm (12), biological significance was defined only when the difference between two groups (i.e., the T-value) was maintained above the P = 0.05 line for a linear distance of at least 100 μm. This technique determined whether DYN or SP was released, what specific sites in the spinal cord released DYN or SP, and whether an experimental intervention altered the spatial pattern of DYN or SP release within the T4 segment of the spinal cord. This technique is not used to determine differences in specific amounts of peptides released, and this should not be inferred from the data presented (e.g., the height of the T-value above the significance line).

Ibotenic acid microinjections.

For those rats (n = 14) undergoing bilateral ibotenic acid (IBO) injections, the dura above the dorsal aspect of the C2 spinal cord was removed during the basic surgical preparation of the animal. Following the completion of the 60 min postsurgery recovery time, a glass microprobe attached to a Nanoject II (Drummond Instruments) was positioned 0.8 mm lateral to midline and inserted 1.0 mm into the C2 spinal segment. After 60 s, 200 nl of 5 μg/μl IBO was injected into either the right or left side of the spinal segment. After 60 s, the microprobe was removed and positioned on the opposite side of the spinal cord, and the injection was repeated. The order of which side of the spinal cord received the first injection was randomized. Pledgets soaked in the same 5 μg/μl solution of IBO were placed across the dorsal surface of the C2 spinal cord, and a 60-min postinjection time elapsed before beginning the experimental protocol. This injection sequence and time protocol are similar to and was based on an earlier study involving microinjections of IBO in the central nervous system (CNS) (34). As a control, 3 rats underwent bilateral 200-nl injections of normal buffered saline instead of the IBO.

Nor-binaltorphimine microinjections.

Another group (n = 6) of rats received the κ-opioid receptor antagonist, nor-binaltorphimine (BNT). The injection procedure was identical to the one described for IBO, except the microinjector was positioned 0.5 mm lateral to the midline and inserted 1.0 mm into the T4 spinal cord. After 60 s, 100 nl of 0.75 μg/μl BNT (made fresh immediately before injections) was injected bilaterally into the T4 spinal segment. Pledgets soaked in the same BNT solution were placed across the dorsal aspect of T4 following the injections and 10 min later, the pledget was discarded and the first SP antibody-coated microprobe was inserted for the initial 10 min rest period (see Fig. 1). Following the removal of rest probe 3, a fresh pledget was placed on T4 surface, while the initial 15 min of preemptive SCS was applied. This pledget was removed once probe 4 was inserted in T4 and CoAO initiated.

Coronary artery occlusion.

A left thoracotomy was performed between the 4th and 5th ribs, while the animals were on ventilation. The pericardium was opened, and a segment of saline-soaked 5-0 suture was looped around the left anterior descending coronary artery, near its branch point from the left coronary artery, as previously described (10, 19, 21). A DYN- or SP-antibody-coated microprobe was positioned in the T4 spinal level for 10 min (rest, Prb1 probe) followed by a second (Prb2), and then third rest probe (Prb3) inserted for subsequent 10-min periods prior to the beginning of the occlusion sequence (Fig. 1). For each of the following probes (Prb4 to Prb6), coronary artery occlusion was applied sequentially for 90 s with a 60-s rest interval over a 10-min period (i.e., 4 occlusions were applied over the 10-min period). This constituted the coronary occlusion intervention, and a single microprobe remained in the spinal cord during this 10-min period. The probe was then replaced, and the coronary occlusion cycle repeated a 2nd (probe 5) and 3rd (probe 6) time for a total duration of 30 min (Fig. 1). The average replacement time between probes was ∼120 s. At the completion of the experiment, the heart was removed, and the placement of the snare suture was verified visually.

Spinal cord stimulation.

For groups with spinal cord stimulation (SCS), with and without coronary occlusion (Fig. 1), following completion of all thoracic surgical procedures, a spring-loaded unipolar ball electrode was placed epidurally on the dorsal surface of the left C2 spinal cord segment. The electrode was covered with mineral oil to prevent current spread. Stimulation was applied using a Grass constant current stimulator (model S48) with an isolation unit (Grass Instruments, model SIU 7) using 50 Hz, 0.2-ms duration and voltage at 90% of motor threshold. Preemptive SCS was applied continuously for 45 min (stimulation alone animals, n = 14, groups 1 and 3 Fig. 1), and in the group of rats also undergoing CoAO (n = 21, groups 2, 4, and 5, Fig. 1), with SCS starting 15 min before beginning the intermittent CoAO, and sustained for a total duration of 45 min.

Immunohistochemistry.

Ninety minutes after completion of the experimental period, animals were given a large dose of urethane (300 mg/0.5 ml) and then perfused transcardially as described before (20).

Sections of T4, C2 and caudal brain stem were processed for the presence of c-Fos like-immunoreactivity (Fos-LI), as previously described (10, 20). The localization of Fos-LI was evaluated by bright-field microscopy using an Olympus BH2 microscope. Sites in the CNS were identified using the atlas of Paxinos and Watson (30). Quantification of the Fos-positive cells was done by counting the number of dark brown nuclei as described previously (10, 20). Data were then grouped and averaged ± SE for each experimental category. Differences in the number of Fos-positive neurons were determined using one-way ANOVA (SigmaStat) and post hoc comparisons between groups were done using the Tukey test. A significance of P < 0.05 was used.

RESULTS

DYN release from T4 spinal cord in response to SCS: the effect of IBO.

Preemptive, sustained SCS (n = 8 rats) applied at the C2 spinal level did not alter MAP or HR from resting levels (89 ± 4 mmHg and 378 ± 5 bpm, respectively). While transient, intermittent CoAO (n = 4 rats) caused a decrease in MAP and HR from resting levels, this change was not significant (P = 0.561, F = 0.617, and P = 0.151, F = 2.347, respectively). Bilateral injection of either 200 nl of PBS or IBO (n = 6 rats) in the C2 spinal cord, prior to SCS also did not affect blood pressure (BP) or HR, nor did bilateral injection of nor-binaltorphimine (BNT) (n = 6 rats). Verification that the microinjections stayed in the C2 spinal segment was done by injecting 200 nl Pontamine blue dye into the C2. It showed that this volume of fluid spread ∼2.5 ± 0.07 mm rostrocaudally (with a concentrated spread of ∼360 μm) by 1 mm². This indicated that the IBO remained in the C2 spinal segment, which is ∼3.2 mm in length (17).

Volume injection alone into the C2 spinal cord did not alter the pattern of DYN release from the thoracic spinal cord. There were no differences in the average gray levels of rest microprobes after 200 nl PBS was injected into C2 compared with the gray levels of the rest probes with no C2 microinjection or after IBO injections (graph not shown).

The average gray levels of the microprobes inserted into the T4 spinal cord during SCS were lower than the gray levels of the rest probes from 0 mm (corresponding to the surface of the spinal cord) to 1 mm in the spinal cord (Fig. 2A). This indicated that SCS caused an increase in the release of immunoreactive DYN (irDYN) above basal, resting levels, from laminae I–V in the T4 thoracic spinal cord. This pattern was reversed following injection of IBO into C2 (see Fig. 2C). The gray levels of the rest microprobes were significantly lower than the levels of the SCS probes at the superficial sites near the surface of the spinal cord, from 0 to 0.2 mm (see Fig. 2C) and from 1 to 2 mm in the spinal cord. The interposed laminae approached but did not achieve significance. At the very least, this indicated that C2 applied IBO inhibited the C2 SCS evoked release of DYN from the T4 superficial laminae I–II and from laminae V–VII.

Fig. 2.

Image analysis graphs of DYN A-(1-13) antibody-coated microprobes. Comparison of the binding patterns of radiolabeled DYN A-(1-13) to microprobes and the resulting differences in optical density [converted to gray levels in arbitrary units (a.u.)] of autoradiographic probe images. Gray levels are given as the means (solid colored lines) ± SE (dotted black traces on either side of main colored trace) and are plotted against the length of probe inserted into the spinal cord (2 to 0 mm) together with the 2-mm length that remained outside the spinal cord (0 to –2 mm). Probes were inserted 0.5 mm lateral to the midline on the left side of the T4 spinal segment. Plots of the T value (solid blue line; P = 0.05, minimum level of significance) are superimposed just above the x-axis in each panel. Areas along the length of the probes where the difference in the mean gray levels were ≥P = 0.05 for a continuous distance of at least 100 μm indicate significant differences between the corresponding two sets of probes (experimental conditions). A: comparison of rest probes inserted into the left T4 spinal cord during basal (rest) conditions (green trace) to those probes inserted during left-sided C2 spinal cord stimulation (SCS) (red trace). B: comparison of probes inserted into the T4 spinal cord during control rest conditions without IBO injections (black trace) to those rest probes inserted into the spinal cord after IBO was injected into the C2 spinal cord (red trace). C: comparison of rest (basal) probes inserted into the T4 spinal segment following bilateral injections of IBO into the C2 spinal cord (blue trace) to those probes inserted into the T4 spinal cord during C2 SCS following IBO injections (red trace). D: comparison of probes inserted into the T4 spinal cord during control C2 SCS without IBO injections (green trace) to those during C2 SCS after IBO was injected into the C2 spinal cord (red trace). n, number of probes in each group.

The inhibitory effect of C2 applied IBO on DYN release is further illustrated when comparing the average gray levels of the resting microprobes inserted into T4 during control (rest) conditions, when no IBO was injected, to those inserted during rest following IBO (see Fig. 2B). There was no difference between these two groups of probes along the 2 mm of probe residing in the thoracic spinal cord. In contrast, the gray levels of the SCS microprobes following C2 IBO injection were consistently higher than those of the microprobes during C2 SCS in the control condition (see Fig. 2D), indicating the IBO applied at the C2 region inhibited the C2 SCS-evoked release of DYN from T4 laminae I–IV and from VI–VII. Preemptive C2 SCS alone did not alter the number of Fos-positive neurons compared with the sham control in the dorsal horn laminae of T4 (Fig. 3A) or in the nucleus tractus solitarii (NTS) in the caudal brain stem (Fig. 3B), but it did recruit additional neurons (i.e., higher number of Fos-positive neurons) within the intermediolateral column (IML) in the T4 spinal cord (Fig. 3C). Injection of IBO in the C2 spinal cord resulted in an increased number of Fos-positive neurons during SCS in laminae I–V in T4 compared with SCS alone (Fig. 3A). This treatment did not affect neuronal activation in the IML (Fig. 3C) or within the NTS (Fig. 3B).

Fig. 3.

A: quantification of c-Fos-positive neurons in T4 spinal segment. Fos-positive cells were counted under bright-field microscopy within laminae I–V (0.3-mm² area) for the left side of spinal segments, averaged and expressed as a function of the cross-sectional area measured (means ± SE). B: quantification of c-Fos-positive neurons in the left nucleus tractus solitarius (NTS) (in a 0.6-mm² area) at brain stem level bregma −13.8 mm. C: quantification of c-Fos-positive neurons in left side T4 intermediolateral cell column (IML) (in 0.23-mm² area). Bars indicate number of Fos-positive neurons in each of these central nervous system regions in sham time controls, during SCS ± bilateral injections of IBO into the C2 spinal region, in response to transient LAD coronary artery occlusion (CoAO) and during combined SCS and CoAO (±IBO). *P < 0.05 vs. sham; +P < 0.05 vs. CoAO; #P < 0.05 for SCS+CoAO vs. SCS+CoAO+IBO.

The release of DYN and SP during SCS and cardiac ischemia: the effect of IBO.

Cardiac ischemia induced by transient, intermittent CoAO, applied during preemptive continuous C2 SCS, caused an initial decrease in BP, usually within the first 10 s and an augmentation in the amplitude of the QRS complex accompanied by an elevation in the ST segment. ST segments rapidly returned to basal levels after reperfusion. Overall, there were no significant changes from baseline for MAP and HR during CoAO in any of the groups except for the SCS+CoAO+IBO group, when MAP decreased (from 80 ± 5 to 64 ± 7 mmHg, P < 0.05) and HR decreased (from 402 ± 11 to 369 ± 13 bpm, P < 0.05) during the CoAO stress.

The effects of preemptive C2 SCS on irDYN release in the T4 spinal cord during cardiac ischemia are shown in Fig. 4. The gray levels of the SCS+CoAO probes were consistently lower than the preischemic rest probes, indicating that preemptive C2 SCS continued to induce DYN release within the T4 thoracic spinal cord from laminae I–VII during CoAO (Fig. 4A). Moreover, there was no synergistic effect between cardiac ischemia and C2 SCS on T4 DYN release, as seen in Fig. 4B; the gray levels of the microprobes during C2 SCS alone were the same as the gray levels of the microprobes during C2 SCS and CoAO. This finding is consistent with previous results that showed CoAO alone did not cause the release of irDYN from rat thoracic spinal cord (see Fig. 2, Ref. 19).

Fig. 4.

Image analysis graphs of DYN antibody-coated microprobes during cardiac ischemia. A: comparison of probes inserted into the T4 spinal cord during rest (blue trace) to those probes inserted during preemptive, continuous C2 spinal cord stimulation, and coronary artery occlusion, SCS+CoAO (red trace). B: comparison of probes inserted during C2 SCS (green trace) to those inserted during C2 SCS and coronary occlusion (red trace). Details of graphs described in Fig. 2.

Cardiac ischemia increased the number of Fos-positive neurons in the T4 dorsal horns and the NTS compared with resting sham levels (Fig. 3, A and B). The change in the number of Fos-positive neurons by CoAO localized within the IML (sympathetic preganglionic neurons) at the T4 spinal level was not significant (Fig. 3C). Compared with CoAO alone, C2 SCS significantly reduced the number of Fos-positive neurons in the T4 dorsal horns (Fig. 3A) and NTS (Fig. 3B), while increasing the number within the T4 IML region (Fig. 3C). Injection of IBO in the C2 spinal cord did not affect the number of Fos-positive neurons in the T4 dorsal horn or the NTS compared with the number observed during SCS+CoAO without IBO. In contrast, IBO pretreatment to the C2 region resulted in fewer Fos-positive neurons in the T4 IML during SCS+CoAO compared with the control SCS+CoAO alone (Fig. 3C).

Figure 5 shows the effects of IBO on the release of substance P (SP) from the thoracic spinal cord. C2 applied ibotenic acid had no effect on the basal resting levels of SP at T4 (Fig. 5A). However, in rats receiving C2 IBO, the gray levels of the SP antibody-coated probes during C2 SCS+CoAO were lower than those without IBO at 0 to 0.25 mm and 0.6 to 1.6 mm (see Fig. 5B). This indicated a greater release of SP from the superficial laminae I–II and IV–VII during cardiac ischemia following IBO injections into the C2 region.

Fig. 5.

The effect of IBO injection into the C2 spinal cord on SP release from the T4 spinal cord. Image analysis graphs of SP antibody-coated microprobes (A) during rest in control conditions (blue trace) or during rest following IBO injection (red trace) or (B) during preemptive continuous C2 SCS+CoAO in the control condition (blue trace) or following IBO injection (red trace). Details of graphs are described in Fig. 2.

Release of SP during SCS and cardiac ischemia: the effect of BNT.

The results reported here indicate that SCS induces an increased release of irDYN and suggest that irDYN may, in part, contribute to the inhibition of SP release by preemptive SCS during cardiac ischemia. If irDYN participates in the inhibition of SP from thoracic dorsal horn sites, introducing a specific κ-opioid receptor antagonist, such as nor-binaltorphimine (BNT) should reverse the inhibitory effects of SCS. Injection of BNT had no effect on MAP or HR compared with control levels or during SCS+CoAO. Figure 6 shows the effects of BNT on SP release during SCS and CoAO. The gray levels of the SCS+CoAO probes in rats receiving BNT were lower than the corresponding rest probes in rats with BNT (see Fig. 6A) at 0 to 1.6 mm. This binding pattern is quite similar to that previously reported (see Fig. 5, Ref. 21) in rats undergoing CoAO alone. Thus, within this group of rats, the CoAO afferent signal induced SP release from laminae I–V, even though preemptive C2 SCS was applied. The gray levels of the SP microprobes in BNT-injected rats during SCS+CoAO were significantly lower than the gray levels of probes from control rats (without BNT) during SCS+CoAO at 0.3 to 0.47 mm and 1.0 to 1.6 mm (see Fig. 6B). This binding pattern is quite similar to that described in Fig. 5B and indicates a release of irSP from laminae II–III and IV–VII. Furthermore, there was no difference in the gray levels of SP microprobes from rats receiving BNT at T4 to those receiving IBO at C2, as shown in Fig. 6C in response to SCS and CoAO. This indicates that both IBO in C2 or BNT in T4 permitted the release of SP in response to ischemia even though C2 SCS was applied.

Fig. 6.

The effect of nor-binaltorphimine (BNT) injection into the T4 dorsal spinal cord on SP release from the T4 spinal cord. Image analysis graphs of SP antibody-coated microprobes during preemptive continuous C2 spinal cord stimulation and coronary artery occlusion (SCS+CoAO): (A) following BNT injection compared with rest (black main trace) and stimulation with occlusion (regular trace); (B) in the control condition (without BNT, green trace) and following BNT injection (red trace); and (C) following BNT injection (red trace) or IBO C2 injections (blue trace). Details of graphs are described in Fig. 2.

DISCUSSION

It is well documented that electroneuromodulation represents an effective clinical therapy for patients with intractable chronic angina (25). Its application provides benefit not only in reducing the pain associated with coronary artery disease (25) but also in ameliorating the ischemia-induced damage to the myocardium (35). However, the neurochemical mediators transducing the antinociceptive effects of electroneuromodulation are still not well understood. Our current understanding of the events that regulate not only the perception of the cardiac nociceptive signal but also the autonomic reflex adjustments to the stressed myocardium include a hierarchy of neural pathways, involving intrathoracic and central (propriospinal and supraspinal) circuits. Most likely, there are multiple neurochemicals involved in transmitting and modulating the cardiac nociceptive signal, and the pattern of such release can be impacted by site-specific electrical or chemical stimulation of central and peripheral neurons (10, 19, 21, 25, 27–28). Our studies have focused on two such candidates, the putative nociceptive peptide, substance P, and the putative analgesic peptide, dynorphin A (1–13). The results from the current experiments provide further insight into the potential interaction between these two neuropeptides and their involvement with the processing of the cardiac ischemic afferent signal at the level of the spinal cord.

The important findings of the present study are that 1) cranial cervical spinal cord stimulation induces irDYN (dynorphin A 1–13) release from T4 dorsal horn sites (laminae I–V); 2) application of the excitotoxin, ibotenic acid, at the C2 level, eliminated such C2-SCS induced DYN release at T4; 3) the application of IBO at the C2 level blocked the C2-SCS-induced suppression of SP neuronal release in the T4 spinal segment associated with transient myocardial ischemia; 4) application of the specific κ-opioid receptor antagonist, nor-binaltorphimine, directly in the thoracic dorsal horn, also blocked the C2 SCS-induced suppression of SP release; and 5) C2 IBO application abolished the low level (IML cell column) sympathoexcitation associated with high cervical spinal cord stimulation concurrent with coronary artery occlusion.

The release of DYN from the thoracic spinal cord at laminae I–V in response to C2 SCS is similar to our previous finding that electrical stimulation of vagal afferents caused DYN to be released from the thoracic spinal cord, at laminae I–III and V–VII (19). Furthermore, autoradiographic data indicate there is a high concentration of κ-opioid receptors in laminae II–III (16). A number of reports demonstrate the presence of both DYN-ergic fiber endings and cell bodies within the spinal cord (7, 31, 40) and show opiates, including DYN, reduce transmitter release from nociceptive primary afferents, especially SP-containing sites (22, 29, 37–39). This correlates with the fact that high cervical SCS decreases thoracic spinothalamic tract neuronal electrical activity, one of the major pathways that conduct the ischemic afferent signal to higher centers (6, 13, 14, 25, 28, 33). Since both types of electromodulation, vagal and spinal cord stimulation, excite neurons in the upper cervical spinal cord (1, 13, 14, 25, 28), we hypothesize that the irDYN released in the thoracic dorsal horn sites is involved with the anti-nociceptive effects of electroneuromodulation.

The release data (SP and DYN) together with the changes in the pattern of stress-induced neuronal recruitment (c-Fos) provide insight into the neuronal circuitry underlying high cervical modulation of thoracic processing of the nociceptive input. First, the effects of the neurotoxin IBO, which affects soma but not fibers of passage, were studied 60–90 min following its injection. This 90-min period is well within the effective time window for IBO to disrupt the inhibitory effects of vagal afferent stimulation (34), as well as other nociceptive synaptic-dependent pathways (2). It is also well beyond the excitatory phase of IBO, which is short-lived, occurring early on (10 min) after its injection (34). While C2 SCS in the control animals resulted in a release of irDYN from laminae I–V compared with its basal resting profile, this pattern was reversed during SCS following ibotenic acid injection into the C2 spinal cord. This suggests first that a major source of the irDYN detected in the T4 spinal cord during C2 SCS would most likely be either from the terminal endings of propriospinal neurites located in the C2 spinal cord or from intervening irDYN-ergic interneurons responding to the C2 propriospinal input. There is an abundance of DYN-ergic interneurons within the thoracic spinal cord (7, 31, 40). Interestingly, applying SCS at the high cervical level does not activate supraspinal sites in the brain stem. This was demonstrated through electrophysiological studies (25, 27) as well as the immunohistochemical data in the present study, using changes in Fos expression as an index of neuronal recruitment. The present c-Fos data, which show no changes in level of neuronal recruitment in NTS sites during C2 SCS, support our previous findings (10). The c-Fos data suggest that excitation of supraspinal sites in the NTS are not the source of the irDYN release in the T4 spinal cord during C2 SCS, but these findings do not necessarily exclude the possibility that other supraspinal sites may be activated by high cervical SCS and contribute to the antinociceptive signaling at the thoracic level.

Electrical stimulation of the left vagus (19) and the C2 dorsal spinal column induce release of irDYN in the thoracic spinal cord, and reduce SP release in the T4 spinal area associated with transient myocardial ischemia. However, the c-Fos data suggest that the pathways and thus the sources for this opiate (or other antinociceptive modulators) may not be identical. Vagal stimulation caused an increase in neuronal recruitment in the NTS, as well as the C2 spinal cord (10); this suggested that some of the T4 spinal irDYN, or other inhibitory neuromediators, may have come from direct descending terminal endings from the NTS. This could function in parallel with the dynorphin projections arising as part of the C2 propriospinal pathway. Because the number of Fos-positive cells in T4 dorsal horns during C2 SCS with ibotenic acid were similar to the number of Fos-positive cells during SCS+CoAO with or without ibotenic acid (but higher than the number observed during SCS alone), the source of the irDYN during SCS was probably not thoracic DYN-ergic interneurons. If C2 SCS-activated thoracic DYN-ergic interneurons, then ibotenic acid in C2 would essentially eliminate this excitatory signal. If this were the scenario, one would expect fewer Fos-positive neurons in the T4 dorsal horn during SCS after ibotenic acid injections in C2 than in the control SCS. Taken together, these data strongly suggest a predominant role for C2 derived propriospinal projections in modulating thoracic neuronal processing of the cardiac nociceptive input by release of irDYN from their terminal endings in the thoracic dorsal horns.

As we reported previously (10), both continuous SCS and preemptive continuous SCS during cardiac ischemia increased neuronal recruitment in the T4 IML. This may be indicative of a cardioprotective effect during SCS imparted by low-level activation of cardiac adrenergic neurons (35). Specifically, the cardiac protective effects of SCS to reduce infarct size are eliminated by α1- or β-adrenergic blockade (35). Cervical IBO resulted in a lower number of Fos-positive neurons in the IML area of the T4 spinal cord during cardiac ischemia. This suggests that the ability for recruitment of sympathetic preganglionic neurons during high cervical SCS was mitigated following C2 IBO and again points to the multiple peripheral and central neural effects of neuromodulation therapy (15, 26). Further studies should evaluate the resultant effects of disrupting this intersegmental pathway with respect to end-organ function in the ischemic stressed myocardium.

Perhaps the most important result thus far from these studies is that there was a difference in the release of irSP from T4 dorsal horn sites during SCS+CoAO after ibotenic acid was injected into the C2 spinal cord. Unlike the effects of preemptive continuous C2 SCS during cardiac ischemia, which inhibited SP release (10), chemical disruption of this inhibitory signal from the cervical propriospinal neurons increased SP release of from T4 laminae I–II and IV–VII during activation of the cardiac nociceptive signal. In fact, the SP release profile returned to that evident during CoAO alone (21). This increased irSP release following cervical IBO injection supports the hypothesis that high cervical SCS modulates the transmission of the cardiac nociceptive signal, that the nociceptive signal is, in part, mediated by irSP, and that this central neuromodulation is mediated in part by irDYN. This suggestion that irDYN, in part, mediates the inhibitory effects of spinal cord stimulation on SP release is confirmed by the findings with the κ-opioid antagonist, nor-binaltorphimine. When injected into the thoracic spinal cord, at sites that would target laminae I–VII, nor-binaltorphimine resulted in an increased SP release from the same laminae in which an increased release of SP was observed in response to CoAO alone and in response to the cervical ibotenic acid. There was no difference between the effects of ibotenic acid (at C2) and nor-binaltorphimine (at T4) on SP release at the T4 spinal level during SCS and cardiac ischemia; i.e., interrupting the release of irDYN (by the excitotoxin) or preventing irDYN from interacting with its receptor (by the antagonist) allowed SP release to occur in response to the cardiac ischemia. This suggests that irDYN release, caused by high cervical SCS, most likely interacts with SP-containing sites within the thoracic spinal cord to alter the transmission of the cardiac ischemic nociceptive signal.

Perspectives and Significance

There is a growing body of evidence that points to a neural hierarchy controlling the heart that involves not only central nervous system but also intra-spinal neural pathways interacting with each other, as well as with intrathoracic, extracardiac, and intrinsic cardiac neuronal pathways (3–5, 14). This integrated view of cardiac control is especially pertinent to our evolving understanding of not only what is happening within the myocardium during ischemia but also to how the individual with chronic angina and coronary artery disease should be treated. Our summary diagram (Fig. 7) illustrates schematically some of the potential central interactions between the various components of this hierarchy and some of the possible neurochemical mediators for some of these interactions within the spinal cord and focuses on the analgesic effect of locally released irDYN on SP.

Fig. 7.

Simplified schematic diagram of the basic circuitry and components involved with signaling from the ischemic myocardium into the central nervous system. Localized ischemia activates cardiac ischemia-sensitive afferent neurons (CISAN) with cell bodies in the nodose ganglia (NG) and dorsal root ganglia (DRG) of spinal segments C1–C3 and T1–T6, respectively (13, 24). The CISAN also project to intrinsic cardiac ganglia (ICN) (5, 28). CISAN with cell bodies in DRG synapse in laminae I–V, VII, and X, where neurons of the spinothalamic tract (STT) are excited (red-colored pathway and solid black large•pathway). Some CISAN project directly into the brain stem (cell bodies located in NG), primarily at the nucleus tractus solitarius. As illustrated, CISAN contain different profiles of substance P (SP), neurokinin A (NKA), calcitonin gene-related peptide (CGRP), and “other” putative neurotransmitters. These are presumably released at the dorsal horn sites to activate STT neurons, especially during cardiac stress. These nociceptive reflex pathways can be modulated by site-specific electrical stimuli delivered to either the spinal cord or vagus (LVS) nerves. The centrally derived neuromodulation of nociceptive pathways, evoked from LVS or high cervical SCS electrical stimulation (represented by thunderbolt), can arise from pathways involving propriospinal neurons (PSN) {1}, interneurons {2}, or from brain stem sites either on STT or via PSN {3}, to release neurotransmitters (e.g., dynorphin, DYN) to modulate the myocardial ischemia stress-induced release of SP from CISAN.

We suggest there are three possibilities for the source of this DYN within the thoracic spinal cord. DYN can be released from the terminal endings of propriospinal neurites located in the cranial cervical spinal cord which project to the thoracic spinal cord (see Fig. 7 pathway {1}). Alternatively, DYN can be released from dynorphinergic interneurons located within laminae I–V in the thoracic spinal cord (see Fig. 7, pathway {2}); these interneurons would be activated by an excitatory projection from the cranial cervical spinal cord. Lastly, DYN could be released from descending terminal endings of supraspinal neurites located within the brain stem (or other higher centers) that project to the thoracic spinal cord (see Fig. 7, pathway {3}). Given the ability of SCS to mitigate intractable angina (26), to induce states of cardioprotection during transient myocardial ischemia (35) and to stabilize the myocardial ischemia-induced neuronal imbalances within central and peripheral components of the cardiac neuronal hierarchy (10, 15), future studies should be directed at further unraveling the multiple interactions existent within the cardiac neuronal hierarchy and the resultant changes in reflex modulation in the presence of a stressed myocardium.

GRANTS

This work was supported by a grant-in-aid from the American Heart Association-SE-region (0555269B, C. A. Williams, principle investigator) and the National Institutes of Health (HL-71830, J. L. Ardell, principle investigator). K. Sutherly was supported by an American Heart Association Health Science Summer Research Fellowship (J. L. Ardell, principle investigator). The authors wish to thank Professor J. Andrew Armour for his comments, discussion, and suggestions regarding this manuscript.