INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE TERM REPERFUSION injury is used to describe a phenomenon that encompasses abnormal myocardial responses with reestablishment of oxygenated blood flow to the previous ischemic myocardium. The consequences of this response include myocardial cell injury, ranging from reversibly "stunned" myocardium to accelerated necrotic cell death (10). An additional clinical consequence is the generation of potentially lethal ventricular arrhythmias that can occur within seconds of the onset of reperfusion following coronary artery occlusion (10). The precise mechanisms involved in the generation of reperfusion-induced alterations in cardiac function remain elusive. One potential mechanism involves untoward activation of populations of intrinsic cardiac neurons. Excessive activation of specific populations of atrial neurons can induce negative cardiodynamic effects that include the genesis of cardiac arrhythmias (14). It is not known if ventricular myocardial ischemia/reperfusion can be transduced by the ventricular nervous system or whether the resultant neuronal responses (if any) can be modified pharmacologically. In the context of this investigation, the term transduction is employed to represent the capacity of the intrinsic cardiac nervous system to transduce cardiac sensory information, primarily by local circuit neurons, to efferent neurons regulating regional cardiac function.

Adenosine has received attention in the setting of reperfusion injury since it is known to exert cardioprotective and anti-arrhythmia effects (8). Adenosine is a by-product of catabolism within the ischemic myocardium, increasing severalfold in the ischemic myocardium (17). In addition to cardiomyocyte modulation, adenosine has also been demonstrated to elicit neuromodulatory effects on the cardiac autonomic nervous system. Adenosine-sensitive cardiac afferent neurons are located in nodose and dorsal root ganglia (7, 12, 19). Adenosine can also modify the activity generated by atrial neurons in vitro (3) or in vivo (13). The neuromodulator effects that adenosine exerts on ventricular neurons have yet to be determined. Furthermore, if purinergic intrinsic cardiac neurons are involved in the transduction of myocardial ischemia/reperfusion from cardiac afferent neurons, it is important to determine whether purinoceptor specificity exists in that process. If such specificity does exist, the intrinsic cardiac nervous system may be capable of modification in the presence of such pathology to stabilize cardiac function.

Therefore, the present study was devised to determine 1) whether ventricular ischemia/reperfusion can be transduced by the ventricular nervous system and, if that does occur, 2) whether purinergic receptors are involved in such transduction. Furthermore, we sought to determine if the responsiveness of the intrinsic cardiac nervous system to myocardial ischemic/ reperfusion injury could be modified by pharmacological means.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

General methods. All experiments were performed in accordance with the guidelines for animal experimentation described in "Guiding Principles for Research Involving Animals and Human Beings" by the American Physiological Society. Forty-four Hamshire-Duvoc pigs of either gender, weighing 20-30 kg, were studied. The animals were sedated with a combination of ketamine (80 mg/kg iv) and Pentothal Sodium (20 mg/kg iv). After the endotracheal intubation was performed, positive pressure ventilation was initiated with 0.95 FI_O2 and 0.05 FI_CO2 using a Bird Mark 7A ventilator (Palm Springs, CA). Anesthesia was maintained with Pentothal Sodium (10-15 mg/kg iv every 5 min) during surgical preparation. Thereafter, during neuronal recording, anesthesia was maintained with -chloralose (75 mg/kg iv bolus, with repeat doses of 12.5 mg/kg iv as required). The adequacy of anesthesia was assessed at regular intervals by applying noxious stimuli to a limb to determine reflex withdrawal responses of a limb as well as to monitor jaw tone. The animals, placed in a supine position, underwent a bilateral thoracotomy through the fifth intercostal space. Once completed, the animals were then placed in a right lateral decubitus position and a ventrolateral pericardotomy was performed to expose the heart.

Operative procedures. A lead II ECG, left ventricular (LV) chamber pressure, and aortic pressure were monitored continuously throughout the experiments. LV chamber pressure was monitored via a Cordis (Miami, FL) #6 French pigtail catheter that was inserted into the outflow tract of the LV chamber via one femoral artery. Systemic arterial pressure was monitored via a Cordis #5 French catheter that was placed in the descending aorta via the other femoral artery. These catheters were attached to Bentley (Irvine, CA) Trantec model 800 transducers. Intrinsic neuronal activity, ECG, and LV pressure were recorded concomitantly on an Astromed MT9500 8 channel rectilinear chart recorder.

Recording neuronal activity. A collection of porcine intrinsic cardiac neurons is located at the bifurcation of the left main as well as ventral interventricular (VIV) (analogous to the left anterior descending coronary artery in human) and circumflex coronary arteries (Fig. 1). Neurons in this ganglionated plexus, termed the VIV ganglionated plexus, are representative of those found throughout the porcine intrinsic cardiac nervous system. This ganglionated plexus was chosen for study because of the relative density of its neurons and the fact that its arterial blood supply is spared when the VIV coronary artery is occluded distal to its first diagonal branch as the multiple smaller arteries that supply blood to this collection of neurons originate cranial to that location.

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Schematic representation of the location of the ventral ventricular ganglionated plexus [ventral interventricular (VIV)]. This figure illustrates not only the location of investigated neurons in the VIV but also the location where the catheter (thick black arrow) was introduced into the VIV coronary artery (CA) before being threaded superior to the branches arising from that artery that perfused the VIV. RV, right ventricle; LV, left ventricle; PA, pulmonary artery; LA app, left atrial appendage; RSVC, right superior vena cava.

Interventions. Ventricular mechanical sensory inputs to identified neurons were tested by gently touching epicardial loci on the ventral epicardial surface of the right and left ventricles with a saline-soaked cotton applicator. Mechanosensory inputs to identified neurons were further tested by transient occlusion (5 s) of the inferior vena cava. Chemosensory inputs to these neurons were then tested by applying the Na⁺ channel modifier veratridine (7.5 µM) and, subsequently, adenosine (100 µM) to ventricular epicardial sensory fields. These chemical agents were applied individually to the epicardium for 60-120 s via 2-cm × 2-cm gauze squares soaked with one chemical (~0.5 ml). After each chemical was removed, sensory fields were washed with normal saline (~2 ml/s) for 30 s, at least 5-10 min being allowed to elapse before the next intervention. The application of these chemicals to epicardial loci was repeated to ascertain whether spurious results could occur due to tachyphylaxis. Gauze squares soaked with room temperature normal saline were also applied to identified epicardial sensory fields to determine whether neuronal responses elicited by epicardial chemical application were due to vehicle effects or any mechanical effects elicited by the gauze squares.

Administration of chemicals to identified neurons via their local arterial blood supply. To administer various chemicals to the somata and dendrites of neurons identified in the ventral ventricular ganglionated plexus, a 24-French catheter was inserted in the VIV coronary artery at the level of its first diagonal branch (Fig. 1, thick black arrow). The cannula was threaded proximally (retrograde to flow) such that its tip was positioned just cranial of the origin of the small arteries that supplied blood to the ventral ventricular ganglionated plexus. The position of the catheter tip was confirmed by gentle palpation through the artery wall. The cannula was fixed in place with 2-3 ml of adhesive applied to the arterial wall. PE-15 tubing was inserted into this cannula, with a stopcock at its other end, to permit the administration of chemicals into the local arterial blood supply of identified neurons. Monitored hemodynamic indexes were unaffected by the placement of this cannula. Postmortem examination of appropriate catheter placement was confirmed by injecting methylene blue dye via this catheter into the regional coronary arteries.

Myocardial ischemia. A 3-0 silk ligature, passed around the VIV coronary artery just distal to its first diagonal branch (Fig. 1, black arrow), was led through a snare to occlude that artery later in the experiments. Because investigated neurons derived their arterial blood supply from arteries proximal to this ligature, their arterial blood supply was not affected by these transient coronary artery occlusions. A 30-s coronary artery occlusion time was chosen to create each ischemic episode. This relatively short period of occlusion was chosen for two reasons. First, longer duration occlusions performed multiple times during the course of each experiment might have induced preconditioning. It has been demonstrated that myocardial preconditioning can be avoided if the periods of coronary artery occlusion are less than 1 min in duration (20). Second, in four preliminary experiments in which coronary artery occlusions of various durations (15 and 30 s as well as 1 and 2 min) were tested, it was found that multiple 30-s periods of occlusion elicited consistent neuronal responses. Occlusions longer than 30 s did not produce any further modification of neuronal activity, whereas occlusions less than 30-s duration produced inconsistent neuronal responses. Thus, 30-s occlusion times were employed for the remainder of the experiments. A minimum of 10 min was allowed to elapse between each coronary artery occlusion to allow neuronal and cardiovascular variables to return to baseline values. To ensure reproducibility of the ischemic/ reperfusion responses, multiple occlusions were performed during the course of each experiment. In a preliminary set of experiments involving eight animals, adenosine and then DPCPX plus DMPX were administered locally to investigated neurons during the ischemic episodes (at concentrations described above).

Myocardial reperfusion. In 15 separate pigs, following testing the epicardial sensory inputs depicted above, adenosine (100 µM, 0.1 ml) was administered into the arterial blood supply of the ventral ventricular ganglionated plexus. The local infusion of adenosine was begun at the time of onset of myocardial reperfusion (immediately on release of the vascular occluder); its administration lasted for the first 2-5 s after coronary arterial flow was reestablished.

Data analysis. Heart rate and LV chamber systolic pressure were measured for 30-s periods of time before and during the peak responses elicited by each intervention. Similarly, the activity generated by identified ventricular neurons was analyzed for 30-s periods of time before and during each intervention. Data obtained before and during each intervention are presented as means ± SE. Repeated-measures ANOVA or paired t-tests with Bonferroni correction were used for statistical analysis of the effects elicited by each intervention, where appropriate. Each chemical tested elicited neuronal responses in each animal. As has been found in the past (13), intrinsic cardiac neuronal activity either increased or decreased when exposed to the interventions described above, including exogenous adenosine administration, depending on the population of neurons investigated. Thus, change in neuronal activity from baseline values was also assessed during each intervention [expressed as absolute value of the delta change (||) in neuronal activity]. Significance values of P < 0.05, 0.02, or 0.01 were used for these determinations.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Ventricular mechanosensory and chemosensory inputs. When a gentle touch was applied to the epicardial surface of the right or left ventricles of nine pigs, the activity generated by identified neurons changed in each animal (Table 1). Neuronal activity increased or decreased in response to gently touching a ventricular epicardial locus, depending on the population of neurons studied. Epicardial application of veratridine also altered neuronal activity, increasing or decreasing activity depending on the population studied. On the other hand, when adenosine was applied to epicardial loci at the dose studied, the activity generated by investigated neurons was not modified. Transient occlusion of the inferior vena cava induced minor reductions in LV chamber systolic pressure that were accompanied by concomitant changes in intrinsic cardiac neuronal activity. As has been found in the past (18), monitored indexes were unaffected when equal volumes of saline were administered into the blood supply of investigated neurons. Epicardial application of saline did not alter monitored indexes.

Effects of ischemia on monitored indexes. In 15 pigs, when the VIV coronary artery was occluded for 30 s, the activity generated by the more centrally located ventricular neurons increased in 10 animals (Fig. 2); the activity generated by neurons identified in the five remaining animals decreased. Neuronal activity was changed from baseline values (|| change) by the brief periods of coronary artery occlusion (Table 2). Activity changed even more during the early reperfusion period (Figs. 2 and 3). Heart rate remained unchanged throughout these brief periods of focal ventricular ischemia as well as during reperfusion (control 154 ± 11 beats/min; ischemia 156 ± 12 beats/min; reperfusion 155 ± 11 beats/min). During ischemia, LV chamber systolic pressure fell slightly (113 ± 4 to 101 ± 4 mmHg; P < 0.01). This index returned to baseline values immediately with reperfusion (114 ± 4 mmHg).

View larger version (97K): [in this window] [in a new window]   Fig. 2.   Effect of ischemia/reperfusion on intrinsic cardiac neuronal activity. The activity generated by identified ventricular neurons (D) increased when the distal left ventral descending coronary artery was occluded (black bar below) for 30 s. Their excitability persisted during the reperfusion phase. The break in the diagram signifies a 3.5-min period removed from the recording. AP, aortic pressure; LVP, left ventricular pressure.

View larger version (106K): [in this window] [in a new window]   Fig. 3.   Effects of adenosine on reperfusion-induced neuronal responses. During the reperfusion period subsequent to that depicted in Fig. 2, when adenosine was administered locally to identified ventricular neurons (between arrows below), the activity they generated was suppressed.

Adenosine and adenosine receptor antagonists. In the normally perfused hearts of 15 pigs when adenosine was administered into the local coronary artery blood supply of identified ventricular neurons, neuronal activity changed by a small amount (Table 2). No change in cardiovascular variables was identified during this intervention (heart rate 138 ± 14 to 141 ± 16 beats/min; LVP systolic 116.9 ± 4.3 to 116.7 ± 3.7 mmHg). Local vehicle administration did not alter monitored variables.

Adenosine administered during ischemia. In eight separate animals, neuronal activity responses to ischemia were not affected by locally administered adenosine (92 ± 52 vs. 127 ± 41 || impulses/min; P = not significant). Neither were ischemia-induced neuronal responses affected by the presence of DPCPX and DMPX (51 ± 18 and 63 ± 24 || impulses/min, respectively; P = not significant). Therefore, for the rest of the experiments, we focused on neuronal responses elicited during the reperfusion (immediate postischemia) phase.

Adenosine administered during reperfusion. Local infusion of adenosine obtunded neuronal responses to reperfusion (n = 15 pigs; Fig. 3). Adenosine reduced their responsiveness to reperfusion by 75% (|| impulses/min of 167.7 ± 34.2 to 41.9 ± 10.7 impulses; P < 0.01). Overall, adenosine returned intrinsic cardiac neuronal activity toward baseline (preischemic) values whether excitatory or depressor neuronal responses were induced initially during reperfusion (Fig. 4). Thus, in the 13 animals in which ventricular neuronal activity increased during reperfusion, local administration of adenosine suppressed neuronal activity during repeat reperfusion. In the two animals in which neuronal activity was diminished during the reperfusion phase, adenosine administered during reperfusion returned neuronal activity to baseline values (i.e., activity was greater than in the absence of adenosine).

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Average change in neuronal activity elicited during occlusion and reperfusion in the absence or presence of adenosine and adenosine antagonists. VIV coronary artery occlusion altered ventricular neuronal activity in all 15 pigs, something that was further enhanced on reperfusion (reperfusion baseline). Reperfusion-induced neuronal responses were obtunded when studied in the presence of locally administered adenosine (adenosine during reperfusion). Right: 2 columns present average neuronal activity changes induced during reperfusion in the presence of either 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) or 3,7-dimethyl-1-propargylxanthine (DMPX) (n = 7 for both). Note that adenosine-induced reperfusion responses persisted in the presence of DMPX but not DPCPX. Changes in neuronal activity are absolute delta change from control values (impulses/min). *P < 0.02, comparing reperfusion data with occlusion data.

Reperfusion and adenosine receptor blockade. Adenosine was tested during reperfusion in the presence of the A₁ receptor antagonist DPCPX in seven other animals. After preadministered DPCPX, exogenously administered adenosine failed to modify the neuronal effects of reperfusion (230.3 ± 63.8 vs. 216.6 ± 94.7 || impulses/min, comparing control responses elicited during reperfusion to those induced in the presence of DPCPX, P = not significant; Fig. 4). Conversely, preadministration of the A₂ receptor antagonist DMPX to seven different animals did not diminish the capacity of adenosine to modify neuronal activity during reperfusion (230.9 ± 63 || impulses/min before and 72.3 ± 41.3 || impulses/min after DMPX; P < 0.02; Fig. 4).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

As is found with canine atrial neurons (11), the porcine ventricular nervous system can transduce myocardial ischemia and the subsequent period of reperfusion injury (Fig. 2). This occurred despite the fact that the somata of investigated neurons were not directly involved in the ischemic process. The responsiveness (excitatory or inhibitory activity responses) of ventricular neurons to distal ventricular ischemia apparently was due to the summation of various inputs from an array of ventricular sensory neurites that were capable of transducing the mechanical and chemical milieu of the ventricle. Exogenously administered adenosine, acting principally via adenosine A₁ receptors, modified ventricular neuronal responses to myocardial reperfusion but not to ischemia.

Adenosine, a by-product of catabolism within the ischemic cell, may arise from multiple cell types (15). Myocardial release of adenosine increases severalfold during myocardial ischemia (17). Adenosine has been shown to affect neurons in the central nervous system (2). It also affects neurons in autonomic ganglia associated with the gastrointestinal tract (6) and heart (3, 13) as well as sympathetic efferent postganglionic axons innervating coronary arteries (1). Adenosine affects cardiac sensory neurites associated with afferent neuronal somata in dorsal root (12) and nodose (19) ganglia. In contradistinction to these findings, adenosine did not modify ventricular sensory neurites that input to ventricular neurons identified in this study (Table 1). On the other hand, identified neurons did receive inputs from ventricular chemosensory neurites capable of transducing ion channel modifying agents such as veratridine (Table 1). These data imply that ventricular sensory inputs to investigated porcine ventricular neurons were capable of transducing chemicals other than adenosine, in addition to local deformation. The absence of neuronal responses elicited by epicardial application of adenosine, combined with an absence of hemodynamic perturbations when adenosine was administered directly to the somata and/or dendrites of ventral ventricular neurons (i.e., lack of efferent neuronal effects), suggests adenosine neuromodulation was principally via the local circuit neurons (interneurons).

Data generated from these experiments demonstrated a differential response of the intrinsic cardiac nervous system to myocardial ischemia, as opposed to the subsequent reperfusion injury phase. The capacity of ventricular neurons to transduce regional ventricular ischemia was not affected by adenosine administered exogenously to investigated somata and/or dendrites immediately before or during the ischemic episode. Conversely, intrinsic cardiac neuronal activity was modified when adenosine administration began immediately with initiation of reperfusion (Fig. 3). Presumably, the lack of effect of adenosine induced during the ischemic period was due, in part, to the fact that mechanosensory inputs to the intrinsic cardiac nervous system secondary to ischemia-induced local myocardial dyskinesia (9) would not have been influenced by locally applied adenosine. Furthermore, oxygen free radicals liberated by the ischemic myocardium (21) are capable of activating ventricular sensory neurites (18). Thus, the enhancement of multiple sensory inputs to the intrinsic cardiac nervous system in such a state presumably overwhelmed any demonstrable effects that adenosine receptor modification elicited on identified ventricular neurons.

During reperfusion, locally administered adenosine affected intrinsic cardiac neuronal activity (Fig. 3). Under these circumstances, adenosine acted to restore neuronal activity toward baseline (preischemic) values (c. f., Fig. 2). The stabilizing effects that adenosine imparted to the intrinsic cardiac nervous system during reperfusion were abrogated by the presence of an A₁ receptor but not an A₂ adenosine receptor antagonist (Fig. 4). These data indicate that activation of intrinsic cardiac neuron adenosine A₁ receptors obtunds the responsiveness of ventricular neurons to reperfusion injury.

Adenosine has been shown to exert modulator effects on a number of cellular mechanisms associated with reperfusion injury. Specifically, adenosine A₁ receptor activation produces alterations in local free radical generation (16) and K⁺_ATP channel activity (5). Both of these cellular mechanisms are known to affect intrinsic cardiac neurons (18). In the experiments described herein, adenosine apparently influenced intrinsic cardiac local circuit neurons that were involved in translating myocardial reperfusion injury to cardiac efferent neurons, principally via their adenosine A₁ receptors. As such, modulation of adenosine A₁ receptors may be important in clinical situations such as during revascularization therapy (i.e., thrombolytic therapy or following coronary artery bypass grafting procedures) to suppress reperfusion-induced alterations in the intrinsic cardiac nervous system. Further studies are required to better elucidate the role of ventricular purinergic neurons involved in the reperfusion process to establish the possible clinical utility of adenosine receptor modulation in such circumstances.

Perspectives