Patent Application
20110137362
The present invention relates in general to methodologies for the treatment quenching preconditioning and communication between the peripheral cardiac nervous system and an electrical stimulus. In particular, the present invention utilizes spinal cord stimulation to alter and/or affect the peripheral cardiac nervous system and thereby protect cardiac function.
Recently, the emergence of novel views of the anatomic pathways and neural mechanisms involved in the regional control of the heart have led to the presently claimed and disclosed intrinsic cardiac nervous system modalities and treatments. In fact, it has been determined that a level of processing occurs that permits independent intrinsic cardiac as well as intrathoracic extracardiac and central spinal integration of afferent and efferent autonomic influences, and local neural coordination without necessarily involving the higher brain centers. This knowledge has led to the development of the presently claimed and disclosed invention(s). Lathrop and Spooner [24] have postulated that a “hierarchy of control mechanisms among these different elements, and that they interact as a system of autonomous efferent feedback loops rather than simply as relay stations subservient to central command.” Indeed, disruption of neuronal circuitry leads to numerous cardiac pathologies. Neuronal interactions that occur within this circuitry or hierarchy modulate different regions of both healthy and diseased hearts. Thus, the knowledge of this circuitry and methodologies of modulating this circuitry (as disclosed and claimed herein) have allowed for the development and treatment of cardiac pathologies using novel therapeutic approaches to ameliorate specific cardiac pathologies.
Regional control of cardiac function is dependent upon the coordination of activity generated by neurons within intrathoracic autonomic ganglia and the central nervous system. The hierarchy of nested feedback loops therein provides precise beat-to-beat control of regional cardiac function. Contrary to classical teaching, studies undertaken and disclosed in the present specification utilizing electrophysiological and neuropharmacological techniques applied from the level of whole organ to that of neurons recorded in vitro indicate that intrathoracic autonomic ganglia act in a manner greater than simple relay stations for autonomic efferent neuronal control of the heart. It has been determined that within this hierarchy of intrathoracic ganglia and nerve interconnections, complex processing takes place that involves spatial and temporal summation of sensory inputs, preganglionic inputs from central neurons and intrathoracic ganglionic reflexes activated by local cardiopulmonary sensory inputs. The activity of neurons within intrathoracic autonomic ganglia is likewise modulated by circulating hormones, chief among them being circulating catecholamines and angiotensin II.
The progressive development of cardiac disease is associated with maladaptation of these neurohumoral control mechanisms. Recent data indicate that conventional therapy of cardiac diseases such as myocardial ischemia and heart failure exert their beneficial effects not only on cardiomyocytes directly, but indirectly via the intrinsic cardiac nervous system. The presently disclosed and claimed inventions of the complex processing that occurs within the intrathoracic nervous system, as well as between peripheral and central neurons, will provide a basis for understanding the role that the cardiac nervous system plays in regulating not only the normal heart, but the diseased heart. Information derived from research and experimentation of this complex neuronal hierarchy provides for novel therapeutic approaches for the effective treatment of cardiac dysfunction including protection of cardiac myocytes and stabilization of myocardial electrical activity by targeting various populations of neurons regulating regional cardiac behavior.
Varying elements within the cardiac neuronal hierarchy exert more influence over regional cardiac function than has been traditionally understood. For example, it is now well recognized that the cardiac nervous system is fundamental to the management of heart failure. As such, this nervous system represents a novel and previously unrecognized target for the treatment of heart failure. Control of regional cardiac function is dependent upon intrinsic properties of the cardiac electrical and mechanical tissues as modulated by neural inputs arising from neurons in the intrathoracic autonomic ganglia, spinal cord and brainstem. Disruptions in neural inputs to the heart or alterations in the cardiac interstitial milieu can be associated with deleterious cardiac structural remodeling and, as a consequence, cardiac dysfunction. In the most extreme case, this becomes evident in congestive heart failure. Excessive activation of the intrathoracic cardiac efferent nervous system, as with myocardial ischemia, can evoke ventricular dysrhythmias involving changes within the cardiac nervous system in addition to alterations in cardiomyocyte ion channel function. Maladaptation of neurohumoral control mechanisms can likewise adversely remodel the cardiac extracellular matrix.
Patients with coronary artery disease often experience a crushing, constrictive, suffocating pain, usually in the upper substernal area, but possibly radiating to the arms (especially left), and sometimes the neck, jaw, and teeth. Pain usually occurs because inadequate delivery of blood to cardiac muscle results in tissue ischemia. This ischemic pain generally results from an imbalance between myocardial oxygen consumption and coronary blood flow (demand vs. supply). This imbalance occurs when vessel obstruction or vasospasm reduces the local blood flow to cardiac muscle or the oxygen demand of the muscle is increased. The increased demand may result from events such as physical activity or stress.
Patients suffering chronic refractory angina pectoris commonly have a long history of coronary artery disease. Most patients are relatively young, predominantly male, and have moderately comprised left ventricular ejection fraction. As coronary artery disease worsens, they often require numerous hospital admissions to control the pain resulting in a very poor quality of life. These patients suffer from the ravages of pain even after being treated with conventional therapies such as multiple revascularization procedures and continued treatment with antianginal medication. Patients suffering from chronic refractory angina pectoris and resistant to conventional therapies are classified as survivors of their coronary artery disease. Several adjuvant therapies are presently available for treating these patients. However, several of these therapies have problems including a relatively short period of effectiveness, high costs, intolerable side effects, increased mortality and morbidity.
The conventional treatment for reducing the frequency and intensity of angina pectoris and arrhythmias resulting from myocardial ischemia is anti-ischemic, anti-arrhythmic therapy by depending primarily on pharmacological agents. These therapies are based on an improvement in the balance between myocardial oxygen supply and myocardial oxygen demand. Pharmacological agents and revascularization procedures (CABG and PTCA) are conventional treatments for such disease states. Pharmacological medications used to lower myocardial demand usually are calcium antagonists or beta blocking agents and to increase coronary blood flow delivery to the myocardium are nitrates and calcium channel blockers. For nonreconstructible patients, percutaneous myocardial revascularization (PMR) using laser-drilled holes has been used. Yet there are a significant number of patients that do not experience adequate relief of their anginal symptoms with these treatments or are poor candidates for these therapies. Thus, alternative approaches utilizing direct electrical activation of neural elements within the spinal cord have been devised, with the resultant modulation of the intrathoracic neurohumoral milieu thereby eliciting anti-ischemic, antiarryhtymic, and anti-anginal effects.
The most successful adjuvant therapy for treating chronically ill patients is modulation of the nervous system through spinal cord stimulation (SCS) resulting in improved quality of life, improved cardiac function, and reduction in the number and frequency of anginal attacks. The mechanisms producing the salutary effects of SCS remain unknown, therefore clinicians (especially in North America) do not treat chronic pain patients with SCS and the FDA has yet to approve this treatment. As a result, many patients are suffering needlessly. Though some experimental data indicate that SCS inhibits impulse transmission within the spinothalamic tract, most clinical observations support the notion that SCS alters the ventricular oxygen demand-supply ratio. In this regard, SCS improves myocardial lactate production because of reducing cardiac-myocyte metabolism and thus oxygen demand. It has also been proposed that SCS redistributes myocardial blood flow to regions of ischemia. However, preliminary experiments indicate that SCS does not alter distribution of myocardial blood flow to normal or ischemic ventricular zones in canine preparations. SCS also does not alter cardiac chronotropism or inotropism. In a clinical setting, the antianginal effects of SCS far outlast the duration of the stimulation period.
In recent studies, it has been shown that SCS modulates the activity generated by intrinsic cardiac neurons (ICN). SCS was effective in reducing intrinsic cardiac neuronal activity, whether it was applied before, during or following the onset of a 2-minute coronary artery occulation. This SCS-induced suppression of ICN activity persisted after cessation of SCS implying that the neural suppressing effects of SCS are long-lived and supports the clinical studies which indicate a similar cardio-protective benefit even after SCS is discontinued. It has also been shown that SCS continued to suppress the activity generated by the intrinsic cardiac neurons even when coronary arteries were occluded for periods of time up to 15 minutes. In either case, transection of the subclavian ansae eliminated the suppressor effects of SCS on ICN activity, indicating that the responses were due primarily to the influence of spinal cord neurons acting via the sympathetic nervous system. It appears that SCS may influence the function of the final common neuronal pathway of the heart, the intrinsic cardiac nervous system, in the presence of severe ischemic challenge.
In the canine model, the anti-anginal effects of SCS are not dependent upon redistribution of coronary blood flow or alterations in cardiac work. In another study, it was shown that regional cardiac blood flow distribution evoked by transient occlusion of the LAD in dogs was unaffected by SCS. Moreover, left ventricular pressure-volume loops evoked by transient LAD occlusion were likewise unaffected. SCS by itself was ineffective in changing ventricular blood flow patterns or the left ventricular pressure-volume loops. Therefore, the anti-anginal effects of SCS do not reflect modulation of the cardiac supply/demand balance, but rather involve other neurohumoral mechanisms which protect the heart from some of the deleterious consequences attending myocardial ischemia and the resultant angina.
To protect the heart, SCS activates efferent and afferent neuronal projections to and from the heart. These projections may activate intrinsic cardiac neural processes that release various endogenous neuromediators and neuromodulators (i.e., norepinephrine, purinergic agents, neurokinins, etc.). The net effect of the SCS induced release of neurochemicals stabilizes the heart during myocardial ischemia and protects the heart against the resultant reperfusion injury.
Further, SCS induces release of neurochemicals that provide cardiac myocytes with a state of transient cardioprotection, such that these myocytes have an increased resistance to cell damage during subsequent transient myocardial ischemia. Therefore, pretreatment with SCS reduces the infarct size within ischemic (risk) zones of the heart.
SCS requires activation of the alpha receptor to reduce risk zone for cell death that results from episodes of myocardial ischemia. This effect is analogous to the effects of ischemic preconditioning in that “preconditioning” the heart with SCS reduces the potential for cell death within the risk zone and the effect is mediated by alpha receptors. SCS modulates the activity of the cardiac nervous system and influences the release of neurotransmitters that contribute to the remodeling process on a short term and long term basis. Thus, SCS affects remodeling of the cardiac nervous system and the heart. SCS may be provided for 20 minutes prior to cardiovascular interventional procedures or as soon as possible in the case of a myocardial infarction, i.e., even on the way to the hospital. This treatment may also be used to reduce arrhythmias and heart failure on a long-term basis.
Approximately one-third of people having ischemic heart disease die immediately or die as a result of heart failure. The cause of death is generally a myocardial infarction resulting from death of heart tissue because the coronary artery delivering blood supply is blocked. The present invention discloses that spinal cord stimulation of the dorsal columns of the upper thoracic spinal segments reduces the deleterious effects of interrupted blood delivery to the heart muscle. During spinal cord stimulation, the size of myocardial infarction is reduced by one-half when compared to the size of an infarction without stimulation. The infarct reduction means that the heart has a much greater amount of healthy tissue when spinal cord stimulation is activated prior to and during the period of stopped blood flow because of the coronary artery occlusion.
A disturbance of the fine balance within the whole cardiac neuraxis can result in dramatic changes in cardiac efferent neuronal outflow. Experimental studies have been performed to demonstrate that pathological processes can change the integrative behavior of the cardiac neuraxis. These changes occur when cardiac sensory neurites are activated intensely and for long periods, as when cardiac tissue becomes damaged during regional ventricular ischemia. On the other hand, central processing of cardiac sensory output may become deranged leading to conflicting signals that interfere with the maintenance of cardiac function. This has led to the proposed scheme that the hierarchy of cardiac neurons interact effectively if there is an appropriate balance therein.
Under normal, physiological conditions stimuli applied to the heart do not elicit marked changes in cardiac efferent neuronal activity because central neurons can suppress excessive cardiac sensory information processing. Information has been obtained to support the conclusion that, in the hierarchy of cardiac control, activation of spinal neuronal circuits modulate the intrathoracic cardiac nervous system. Experimental studies have shown that activation of the dorsal columns at the T1-T2 segments significantly reduces the activity generated by the intrinsic cardiac neurons in their basal conditions as well as when activated in the presence of focal ventricular ischemia induced by occluding the left coronary artery. Not only does dorsal column activation modulate the intrinsic cardiac nervous system, but it also modifies the activity of spinal neurons within the T3-T4 segments. In addition, experimental evidence indicates that the central nervous system maintains a tonic inhibitory influence over intrathoracic cardiopulmonary-cardiac reflexes. One of the present inventors has also shown that reflexes mediated through the middle cervical ganglion are increased after decentralization. Based on this evidence, it is postulated that disease processes change the balance between the central and peripheral neuronal processing of cardiac sensory information. Thus, use of electrical currents to activate spinal neuronal circuits can reverse or halt disease processes of the heart preconditioning the heart—i.e., applying electrical activation prior to disease—also is contemplated as a means to pro-actively treat a patient with high susceptibility to cardiac pathologies including arrhythmias.
Within the hierarchy for cardiac control, neurons of the upper cervical segments modulate information processing in the spinal neurons of the upper thoracic segments. In human studies, spinal cord stimulation of the C1-C2 spinal segments relieved the pain symptoms in patients with chronic refractory angina pectoris. Experimental studies in support of the presently claimed and disclosed invention have shown that spinal cord activation of the upper cervical segments of the spinal cord suppressed the activity of spinal neurons in T3-T4 segments. Furthermore, chemical stimulation with glutamate of cells in the C1-C2 segments also reduced upper thoracic spinal neuronal activity. The upper cervical region is intriguing because it is positioned between supraspinal nuclei and spinal circuitry. Neurons in C1-C2 could serve as a filter, an integrator, or as a relay for afferent information, since these neurons receive inputs from vagal afferents from the heart.
Very little information has been published to address underlying mechanisms explaining how central and peripheral cardiac neurons process cardiac sensory information and interact in the maintenance of adequate cardiac output. The presently claimed and disclosed invention shows that disease processes change the balance between the central and peripheral neuronal processing so involved. For instance, when the activity generated by cardiac sensory neurons becomes excessive (such as during focal ventricular ischemia), cardiac function is profoundly affected, cardiac myocyte protection is reduced and arrhythmias are increased. A disturbance of the fine balance within the whole cardiac neuraxis results in dramatic changes in cardiac efferent neuronal outflow. Over the past 30 years, the anatomy and function of the peripheral cardiac nervous system has been studied, focusing during the last decade on its intrinsic cardiac component. The classical view of the autonomic nervous system presumes that its intrinsic cardiac component acts solely as a parasympathetic efferent neuronal relay station in which medullary preganglionic neurons synapse with parasympathetic efferent postganglionic neurons therein. In such a concept, the latter neurons project to end effectors on the heart with little or no integrative capabilities occurring therein. Similarly, intrathoracic extracardiac sympathetic ganglia have been thought to act solely as efferent relay stations for sympathetic efferent projections to the heart. Neural control of regional cardiac function resides in the network of nested feedback loops made up of the intrinsic cardiac nervous system, extracardiac intrathoracic autonomic ganglia, the spinal cord and brainstem. Within this hierarchy, the intrinsic cardiac nervous system functions as a distributive processor at the level of the target organ. Thus, the intrinsic cardiac nervous system plays an important role in the functioning of the heart and in its diseased pathologies. This novel information thereafter leads to numerous methodologies (some of which are claimed and disclosed herein for the treatment, preconditioning and/or quenching of disease pathologies through the use of spinal cord stimulation.
Experimental studies have also shown that pathological processes can change the integrative behavior of the cardiac neuraxis. These changes occur when populations of cardiac sensory neurites are activated intensely and for long periods of time when local cardiac tissue becomes damaged during, for instance, regional ventricular ischemia. Thus, under normal, physiological conditions stimuli applied to the heart do not elicit marked changes in cardiac efferent neuronal activity because central neurons suppress cardiac sensory information processing. On the other hand, central processing of cardiac sensory output may become deranged during excessive inputs leading to conflicting signals that interfere with the maintenance of cardiac function. This has led to the novel concept that the hierarchy of cardiac neurons interact effectively if there is an appropriate balance therein. Fundamental to this hierarchy is its component on the target organ—the intrinsic cardiac nervous system and its influence on the heart.
Consistent coherence of activity generated by differing populations of neurons is indicative of principal and direct synaptic interconnections between them or, conversely, the sharing by such neurons of common inputs. Such relationships have been identified among medullary and spinal cord sympathetic efferent preganglionic neurons, as well as among different populations of sympathetic efferent preganglionic neurons. Different populations of neurons, distributed spatially within the intrinsic cardiac nervous system, respond to cardiac perturbations in a coordinate fashion. If neurons in one part of this neuronal network respond to inputs from a single region of the heart, such as the mechanosensory neurites associated with a right ventricular ventral papillary muscle, then the potential for imbalance within the different populations of neurons regulating various cardiac regions might occur and, thus, its neurons display little coherence of activity. In other words, relatively low levels of specific inputs on a spatial scale to the intrinsic cardiac nervous system result in low coherence among its various neuronal components. On the other hand, excessive input to this spatially distributed nervous system would destabilize it, leading to cardiac arrhythmia formation, etc.
A specific receptor in the heart called the alpha-adrenergic receptor plays a significant role in providing the heart with protection so that the size of infarction is reduced by one-half during coronary artery occlusion. Unmasking the alpha-adrenergic receptor provides critical, new insights about the possible mechanisms contributing to cardioprotection during spinal cord stimulation. Reducing the size of the myocardial infarct will save the lives of many individuals who otherwise would have died from a massive heart attack with a large infarction.
The present invention relates to a method of using spinal cord stimulation to mitigate transient ischemia induced myocardial infarction via cardiac adrenergic neurons as well as using spinal cord stimulation to suppress neuronally induced atrial tachyarrhythmias.
Another aspect of the present invention is the use of spinal cord stimulation for suppressing neuronally induced atrial tachyarrhythmias. Atrial arrhythmias and fibrillation are abnormal. This disorder is found in approximately 2.2 million Americans. Normally, the two upper chambers of the heart (“atria”) contract and empty the blood into the ventricles. However, when the atria fibrillate (quiver), the blood is not pumped completely out of them causing pooling or clotting of blood. If a piece of a blood clot in the atria breaks off and leaves the heart, the clot can become lodged in the brain thereby causing the individual to have a stroke. About 15% of strokes occur in people with atrial fibrillation. Medications, electrical cardioversion, radiofrequency ablation, surgery and atrial pacemakers are often used to treat atrial arrhythmias and fibrillation. These treatments are often ineffective and have multiple side effects.
Previous studies have shown that spinal cord stimulation can protect the heart by influencing the function of the intrinsic cardiac nervous system. However, the present invention uses spinal cord stimulation to reduce or eliminate the ability of the intrinsic cardiac nervous system to generate atrial arrhythmias, specifically tachyarrhythmias that are generated by stimulation of specific nerves to the heart. Spinal cord stimulation is therefore an effective therapy for treating atrial arrhythmias and fibrillation. This finding is important because spinal cord stimulation has no known side effects unlike medications, electrical cardioversion, radiofrequency ablation, surgery and atrial pacemakers. Potentially, over two million stimulation units could be implanted in patients suffering with atrial fibrillation and arrhythmias resulting in potential income as high as $3 billion.
Thus it is an object of the present invention to use the identification of the peripheral cardiac nervous system along with the experimental data and results to provide methodologies utilizing spinal cord stimulation for the (1) treatment of cardiac disease pathologies; (2) communication between an external point and the peripheral cardiac nervous system; (3) preconditioning of the peripheral cardiac nervous system in order to promote a protective effect against cardiac disease pathologies; and (4) quenching aberrant neuronal activity occurring within the and peripheral cardiac nervous system.
This and numerous other objects of the present invention will be appreciated in light of the present specification, drawings, and claims.
The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee.
Before explaining at least one embodiment of the invention in detail, it is to be understood that the invention is not limited in its application to the details of construction and the arrangements of the components set forth in the following description of illustrated in the drawings. The invention is capable of other embodiments or of being practiced or carried out in various ways. Also, it is to be understood that the phraseology and terminology employed herein is for purpose of description and should not be regarded as limiting.
The intrinsic cardiac nervous system has been classically considered to contain only parasympathetic efferent postganglionic neurons that receive inputs from medullary parasympathetic efferent preganglionic neurons. As such, intrinsic cardiac ganglia have been viewed as simple relay stations and major autonomic neuronal control of the heart was believed to reside solely in the brainstem and spinal cord. However, the data supporting the presently claimed and disclosed invention indicate that centripetal as well as centrifugal processing occurs within the mammalian intrathoracic nervous system (i.e., the intrinsic cardiac nervous system). This involves afferent neurons, local circuit neurons (i.e., neurons that interconnect neurons within one ganglion and neurons in different intrathoracic ganglia), as well as sympathetic and parasympathetic efferent postganglionic neurons.
The intrinsic cardiac nervous system consists of multiple aggregates of neurons and associated neural interconnections, localized to discrete atrial and ventricular regions. Among these distinct ganglionated plexi, preferential control of specific cardiac functions has been identified. For example, right atrial ganglionated plexus neurons have been associated with primary, but not exclusive, control of SA nodal function and inferior vena cava-inferior atrial ganglionated plexus neurons primarily, but not exclusively, with control of AV nodal function. One population of intrinsic cardiac neurons, the parasympathetic postganglionic ones, receives direct input from medullary parasympathetic preganglionic neurons. Another population, adrenergic efferent ones [8,9], receives input from more centrally located neurons in intrathoracic ganglia and the spinal cord. The fact that ventricular sensory neurites continue to influence the activity generated by neurons on the heart following chronic decentralization of the intrinsic cardiac nervous system indicates that the somata of afferent neurons, some of which project axons to central neurons, are located within the intrinsic cardiac nervous system. This concept has received anatomical confirmation. Functional data also indicate that the intrinsic cardiac nervous system contains local circuit neurons interconnecting intrinsic cardiac afferent with efferent neurons.
Sub-populations of right atrial neurons that receive afferent inputs from sensory neurites in both ventricles are responsive to local mechanical stimuli and the nitric oxide donor nitroprusside. Neurons in at least one ganglionated plexus locus were activated by epicardial application of veratridine, bradykinin, the β1-adrenoceptor agonist prenaterol or the excitatory amino acid glutamate. Epicardial application of angiotensin II, the selective β2-adrenoceptor agonist terbutaline or selective α1- or α2-adrenoceptor agonists elicited inconsistent neuronal responses. The activity generated by both populations of atrial neurons studied over 5 minute periods during basal states displayed periodic coupled behavior (cross correlation coefficients of activities that reached, on average, 0.88±0.03; range 0.71-1) for 15-30 seconds periods of time. These periods of coupled activity occurred every 30-50 second during basal states, as well as when neuronal activity was enhanced by chemical activation of their ventricular sensory inputs. It has been observed that neurons throughout one intrinsic cardiac ganglionated plexus receive inputs from mechano- and chemo-sensory neurites located in both ventricles. That such neurons respond to multiple chemical stimuli, including those liberated from adjacent adrenergic efferent nerve terminals, indicates the complexity of the integrative processing of information that occurs within the intrinsic cardiac nervous system. Thus, the interdependent activity displayed by populations of neurons in different regions of one intrinsic cardiac ganglionated plexus, responding as they do to multiple cardiac sensory inputs, forms the basis for integrated regional cardiac control.
Recent anatomical and functional data indicate the presence of the multiple neuronal subtypes within intrathoracic extracardiac and intrinsic cardiac ganglia. Within this neuronal hierarchy, the intrinsic cardiac nervous system functions as a distributive processor at the level of the target organ. The redundancy of function and non-coupled behavior displayed by neurons in intrathoracic extracardiac and intrinsic cardiac ganglia minimizes the dependency for such control on a single population of peripheral autonomic neurons. In this regard, network interactions that occur within the intrinsic cardiac nervous system to integrate parasympathetic and sympathetic efferent outflow to the heart do so in coordination with intrathoracic extracardiac neurons that process afferent information from multiple sites in the heart during each cardiac cycle. As no consistent coherence of activity generated has been identified among neurons in intrinsic cardiac and intrathoracic extracardiac ganglia, different populations of neurons, distributed spatially within the intrathoracic cardiac nervous system, respond to cardiac perturbations in a coordinate fashion. If neurons in one part of this neuronal network respond solely to inputs from a single region of the heart, such as the mechanosensory neurites associated with a right ventricular ventral papillary muscle, then the potential for imbalance within the different populations of neurons regulating various cardiac regions might occur. A relatively low level of inputs on a spatial scale to the intrinsic cardiac nervous system would result in low coherence among its components. In contrast, excessive input to this spatially distributed nervous system would destabilize it, leading to cardiac arrhythmia formation, etc.
Regional control of cardiac function is dependent upon the coordination of activity generated by neurons within intrathoracic autonomic ganglia and the central nervous system. The hierarchy of nested feedback loops therein provides precise beat-to-beat control of regional cardiac function. Contrary to classical teaching, intrathoracic autonomic ganglia act as more than simple relay stations for autonomic efferent neuronal control of the heart. Within the hierarchy of intrathoracic ganglia and nerve interconnections, complex processing takes place that involves spatial and temporal summation of sensory inputs, preganglionic inputs from central neurons and intrathoracic ganglionic reflexes activated by local cardiopulmonary sensory inputs. The activity of neurons within intrathoracic autonomic ganglia is likewise modulated by circulating hormones, chief among them being circulating catecholamines and angiotensin II.
The progressive development of cardiac disease is associated with maladaptation of these neurohumoral control mechanisms. Differences exist in autonomic control of the heart before any overt cardiovascular disease occurs and such differences critically influence the outcome at the time of ischemic heart disease onset. Differential remodeling of the cardiac neuron hierarchy (central and peripheral) for reflex control of the heart occurs in animals susceptible verses resistant to development of ventricular fibrillation during the evolution of chronic myocardial ischemia/infarction. Understanding neuronal reorganization/remodeling that occurs within the peripheral autonomic nervous system and the interactions that occur between this neural remodeling and the remodeling of the myocardium leads to the novel approaches as presently disclosed and claimed for anti-arrhythmia therapy and also to therapies directed at ischemic heart disease and protection of the heart.
With respect to neural control of the heart, the intrathoracic ganglia and their interconnections form the final common pathway for autonomic modulation of cardiac function. Data summarized and presented herein indicates in support of the presently claimed and disclosed invention that intrathoracic autonomic ganglia contain a heterogeneous population of cell types including afferent, efferent and local circuit neurons. Yet, as a group, the intrathoracic reflexes mediated within these peripheral autonomic ganglia function in a coordinated fashion with central neurons located in the spinal cord, brainstem and supraspinal regions to regulate cardiac output on a beat to beat basis.
Cardiac afferent neurons. Sensory afferent neurons provide the autonomic nervous system with information about blood pressure, blood volume, blood gases as well as the mechanical and chemical milieu of the heart. For sensory inputs from cardiopulmonary regions, the nodose and dorsal root ganglia are classically recognized as providing sensory inputs to the brainstem and spinal cord respectively. Data indicates that intrathoracic extracardiac (i.e., stellate and middle cervical ganglia) and intrinsic cardiac ganglia also contain afferent neurons whose sensory neurites lie variously within the heart, lungs and great thoracic vessels. Additional sensory inputs for the control of cardiac autonomic neurons arise from baroreceptors and chemoreceptors located along the aortic arch, carotid sinus and carotid bodies as well as from other afferent neural elements within the CNS, especially the hypothalamus.
Nodose Ganglia Afferent Neurons. The nodose receive cardiac afferent inputs from sensory neurites located in atrial and ventricular tissues. These sensory neurites preferentially sense chemical stimuli, with a few responding to mechanical stimuli or both modalities. The response characteristics to induced stimuli are likewise divergent with mechanical stimuli exerting short-lived effects, while the augmentation in activity elicited by chemical stimuli far outlast the applied stimulus. While inputs from these receptors contribute to overall cardiovascular regulation, they are not normally perceived.
Dorsal Root Ganglia (DRG) Afferent Neurons. The cell bodies of DRG afferent neurons, receiving input from cardiac sensory neurites, are located in C6-T6 dorsal root ganglia. The sensory neurites of most of these afferent neurons transduce chemical and mechanical stimuli. The inputs from this subpopulation of cardiac afferent neurons subserve normal cardiovascular regulation, as well as nociception when excessively activated.
Intrathoracic Afferent Neurons. Functional and anatomical data indicate that intrathoracic autonomic ganglia contain afferent soma. The sensory neurites associated with these afferent neurons are variously located in atrial, ventricular, major vascular and pulmonary tissues. Most are responsive to mechanical and chemical stimuli. These afferent neurons continue to influence intrathoracic efferent postganglionic outflows to the heart even after long-term decentralization of intrathoracic ganglia. Such intrathoracic afferent neurons provide inputs to the intrathoracic short-loop feedback control circuits that involve intrinsic cardiac and intrathoracic extracardiac neurons. These intrathoracic neural circuits, acting in concert with CNS mediated reflexes, dynamically control regional cardiac function throughout each cardiac cycle to maintain electrical stability of the heart and protect the myocytes.
Aortic and Carotid Artery Baroreflexes. Stretch receptors, sensitive to changes in vessel size, are found on thoracic and cervical arteries, being concentrated on the aortic arch and the carotid sinus. They provide inputs to neurons within the medulla and spinal cord proportional to systemic arterial blood pressure. Inputs from these sensory neurites course centrally in the IX and X cranial nerves to synapse with neurons located in the nucleus of the medullary solitary tract. Via multi-synaptic connections, these afferent inputs modulate the activity of cardiac parasympathetic efferent preganglionic neurons located primarily in the nucleus ambiguus. They also influence sympathetic efferent neuronal outflow to the heart via brainstem projections to the intermediolateral (IML) region of the spinal cord. The baroreflex so involved represents a negative feedback system that modulates cardiac function and peripheral vascular tone in response to everyday stressors.
Sympathetic efferent neurons. The somata of sympathetic preganglionic efferent preganglionic neurons which regulate the heart are located within the intermediolateral (IML) cell column of the spinal cord, projecting axons via the rami T1-T5 to synapse with sympathetic postganglionic neurons contained within various intrathoracic extracardiac and intrinsic cardiac ganglia. Activation of these sympathetic efferent projections augments heart rate, changes patterns and speed of impulse conduction through the electrical system of the heart and increases contractile force in atrial and ventricular tissues. Sympathetic efferent postganglionic somata that project axons to various cardiac effector tissues are localized in intrathoracic extracardiac and intrinsic cardiac ganglia. Classically, the somata of sympathetic efferent postganglionic neurons that innervate the heart have been thought to be restricted to the stellate ganglia. However, cardiac sympathetic efferent postganglionic soma have also been identified in thoracic middle cervical, mediastinal and intrinsic cardiac ganglia. A subpopulation of intrinsic cardiac neurons express the catecholaminergic phenotype, these neurons thus contain the necessary enzymes to convert L-DOPA to dopamine and norepinephrine. The intrinsic cardiac nervous system also contains a separate population of small intensely fluorescent (SIF) cells that display tyrosine hydrolyase immunoreactivity. Some of these project to adjacent principal intrinsic cardiac neurons.
Parasympathetic efferent neurons. The somata of cardiac parasympathetic efferent preganglionic neurons within the brainstem are located primarily within the nucleus ambiguous, with lesser numbers being located in the dorsal motor nucleus and regions in between. Axons from these preganglionic soma project via the X cranial nerve to synapse with parasympathetic efferent postganglionic neurons located within various intrinsic cardiac ganglia (see hereinafter below). Activation of parasympathetic efferent neurons depresses heart rate, slows the speed of impulse conduction through the heart, induces major suppression of atrial muscle contractile force and evokes negative inotropic effects on ventricular contractile force.
A subpopulation of neurons contained within extracardiac and intrinsic cardiac intrathoracic autonomic ganglia function to interconnect neurons within individual ganglia and between neurons in separate intrathoracic ganglia; these are called local circuit neurons. Preliminary data indicate that these neurons are involved in processing of afferent information to coordinate sympathetic and parasympathetic efferent outflows to cardiac effector sites. Interactions within this neuron population form the substrate for generation of the basal activity within peripheral autonomic ganglia, especially when intrathoracic ganglia are disconnected from the influence of central neurons.
The cardiac nervous system consists of distinct ganglia clusters that function in an interdependent manner to modulate regional cardiac function. To date, eight separate ganglia clusters have been identified within the canine intrinsic nervous system, five associated with atrial tissues and three with ventricular tissue.
The five atrial ganglionated plexuses include: 1) the right atrial ganglionated plexus localized in fatty tissue on the ventral surface of the common right pulmonary vein complex; 2) the inferior vena cava-inferior atrial ganglionated plexus located on the inferior right atrium adjacent to the inferior vena cava; 3) the dorsal atrial ganglionated plexus located on the dorsal surface of the atria between the common pulmonary veins, immediately caudal to the right pulmonary artery; 4) the ventral left atrial ganglionated plexus contained within fat on the caudal-ventral aspect of the left atrium adjacent to the AV groove; and 5) the posterior atrial ganglionated plexus.
The three major ventricular ganglionated plexi are: 1) the right lateral ventricular ganglionated plexus located adjacent to the origin of the right marginal artery; 2) the left lateral ventricular ganglionated plexus located adjacent to the origin of the left marginal artery; and 3) the cranial medial ventricular ganglionated plexus located in fatty tissues surrounding the base of the aorta and main pulmonary artery. Of these eight clusters of ganglia, functions have been primarily ascribed to five of them: neurons in the right atrial and posterior atrial ganglionated plexus have been shown to exert preferential control over the sinoatrial node; those in inferior vena cava-inferior atrial ganglia exert predominant control over inferior atrial and atrioventricular conductile tissues. Neurons in dorsal atrial and cranial medial ventricular ganglia are principal modulators of contractile tissue.
The term “peripheral cardiac nervous system” as used herein includes all neural elements outside the dura matter of the brain and spinal cord. It includes all intrathoracic and cervical neural elements involved in cardiac regulation. For example, the specific elements of the peripheral cardiac nervous system include the intrinsic cardiac nervous system, intrathoracic ganglia including stellate ganglia, middle cervical ganglia, and mediastinal ganglia, and cervical ganglia including superior cervical ganglia and nodose ganglia.
The term “cardiac malfunction” as used herein means any electrical disturbances, mechanical disturbances, and/or coronory blood flow distrubances of normal cardiac behavior including, for example, congestive heart failure, atrial and ventricular arrhythmia, fibrillation, flutter, angina, atrial tachycardia, ventricular tachycardia, ischemia, malfunction of sinoatrial and atrioventricular nodes, and the like.
The term “neural structure” as used herein means a structure that is part of the nervous system including, for example, the brain and spinal cord, the cranial and spinal nerves, autonomic ganglia, and plexuses.
The term “mammal” as used herein means any category of mammailian vertebrate animals including, for example, primates, humans, dogs, cats, horses, rabbits and rodents.
It should be understood that the stimulation frequency of the electrical signals used in association with the present invention can range from 25 Hz to 150 Hz; 30 Hz to 145 Hz; 35 Hz to 140 Hz; 40 Hz to 135 Hz; 45 Hz to 130 Hz; 50 Hz to 125 Hz; 55 Hz to 120 Hz; 60 Hz to 115 Hz; 65 Hz to 110 Hz; 70 Hz to 105 Hz; 75 Hz to 100 Hz; 80 Hz to 95 Hz; and 85 Hz to 90 Hz. The time in which the electrical signal is generated can range from 20 seconds to at least 20 minutes.
With respect to intrathoracic autonomic ganglia, cholinergic and adrenergic efferent neurons in these ganglia represent the output elements that project axons to cardiac electrical and mechanical tissues. Local circuit neurons interconnect adjacent neurons within one ganglion or link neurons in separate clusters of intrathoracic ganglia. These interneurons are involved in coordination of neuronal activity within these peripheral autonomic ganglia, thereby providing the underlying inputs necessary for the maintenance of basal autonomic neuronal discharge. Intrathoracic afferent neurons provide mechanosensitive and chemosensitive inputs from cardiopulmonary regions directly to intrinsic cardiac and extracardiac neurons, forming the basis of the intrathoracic neural feedback system. Superimposed on activities generated by neurons in peripheral autonomic ganglia are efferent inputs from preganglionic neurons in the brainstem and spinal cord that together exert tonic influences on regional cardiac tone. CNS preganglionic inputs are, in turn, influenced by inputs from higher centers in the central nervous system and by afferent feedback from central and peripheral sensory afferent neurons.
Cardiac performance is modulated by both sympathetic and parasympathetic efferent neuronal inputs. The induced change in any regional cardiac function ultimately depends upon the intrinsic characteristics of the cardiac end-effector being innervated, the level of efferent activity from the CNS to the periphery and interactions occurring within peripheral autonomic ganglia and at the respective cardiac end-effectors.
Interactions at the organ level. Anatomical and functional studies indicate that sympathetic and parasympathetic efferent postganglionic nerve endings lie in close proximity to each other in the target tissues. Interactions among sympathetic and parasympathetic efferent projections to the heart involve pre- and postjunctional mechanisms at the end-effectors in cardiac tissue. Post-junctional interactions involve differential modulation of adenylate cyclase via G-protein coupled receptor systems (
Interactions within the ICN. Various lines of evidence indicate that peripheral sites that are separate from the end-effectors contribute to mediating sympathetic-parasympathetic interactions for the control of regional cardiac function. Stimulating parasympathetic and/or sympathetic efferent projections to the heart activates subpopulations of intrinsic cardiac neurons. These extrinsic autonomic projections converge on separate aggregates of intrinsic cardiac neurons, each of which exhibit preferential control over regional cardiac function. With respect to control of chronotropic function, surgical disruption of the right atrial ganglionated plexus eliminates direct vagal projections to the sinoatrial node. Sympathetic efferent neuronal control of chronotropic function and the vagal inhibition of the sinus tachycardia produced by cardiac sympathetic efferent neurons are maintained. These residual sympathetic-parasympathetic efferent neuronal interactions occur at the level of the heart and are prejunctional to the sinoatrial node. As shown herein, these residual interactions occur within the intrinsic cardiac nervous system. Whether such intraganglionic autonomic interactions play correspondingly roles in modulation of dromotropic and inotropic function has yet to be determined.
Intraganglionic interactions within the intrinsic cardiac nervous system depend in large part on common shared afferent inputs and/or interconnections mediated via local circuit neurons. In order to evaluate these interactions, separate populations of neurons were recorded in the ventral right atrial ganglionated plexus (RAGP) in basal states and during discrete mechanical and chemical stimuli of ventricular neurites. In basal states, the coherence of activity generated by the two populations of RAGP neurons fluctuated with a periodicity of 30-50 s and with an average peak coherence of 0.88±0.03. Coherence was increased in conjunction with the enhanced neuronal activity evoked during exposure of ventricular sensory inputs to mechanical and chemical (nitroprusside, veratridine, bradykinin, adrenergic agonists or glutamate) stimuli. The interdependent activity displayed by the population of neurons in different regions of one intrinsic cardiac ganglionated plexus, depending as they do on multiple cardiac sensory inputs, forms the basis for coordination of regional cardiac function within the intrinsic cardiac nervous system.
Interactions within the intrathoracic nervous system. Coordination of autonomic outflows from intrathoracic neurons to cardiomyocytes depends to a large extent on sharing of inputs from higher centers along with interactions among and between various peripheral ganglia. Interactions within and between intrathoracic ganglia involve local circuit neurons (see herein above). Activities generated by neurons in intrinsic cardiac ganglia demonstrate no consistent short-term relationship to neurons in extracardiac ganglia. However, the sharing of cardiopulmonary afferent information acting through both intrathoracic and brain stem/spinal cord feedback loops permits an overall coordination of effector control. Together, these nested feedback control systems allow for a redundancy in neural control of the heart while at the same time maintaining the flexibility to differentially modulate regional cardiac function.
In Vivo Studies. Cardiac neurons generate spontaneous activity in situ, frequently exhibiting activity that is temporally related to the cardiac or respiratory cycles. Of the neurons that displayed cardiac-related activity, many are affected by mechano- or chemosensory inputs from the heart. Trains of electrical stimuli delivered to axons in the T1-T5 ventral roots activate a substantial population of stellate and middle cervical neurons. These data indicate a convergence of preganglionic inputs onto the extracardiac postganglionic soma, reflective of a functional amplification of such sensory input. In contrast, trains of electrical stimuli delivered to the vagosympathetic trunks or stellate ganglia activate a much smaller population of intrinsic cardiac neurons. Moreover, few intrinsic cardiac neurons are activated after a fixed latency when extracardiac efferent neurons that innervate the heart are stimulated electrically, a finding indicative of monosynaptic interconnections to such neurons. These data indicate that, in contradistinction to extracardiac ganglia, substantial spatial and temporal summation of inputs are required to modify the activity generated by neurons on the heart. Intrinsic cardiac neurons generate low level activity in such a state consistent with a nerve network that functions as a “low pass filter”, thereby minimizing the potential for imbalances within autonomic efferent neuronal inputs to the heart, a process which by itself could be arrhythmogenic.
In Vitro Studies. Intrinsic cardiac ganglia contain a heterogeneous population of neurons. An intracellular recording from isolated whole mount aggregates of intrinsic cardiac ganglia indicates that complex neural interactions occur within the heart. Studies on aggregates of intrinsic cardiac ganglia derived from different species further indicate that the resting membrane potentials of these neurons is approximately −60 mV, with relatively low input resistances and thresholds for the generation of action potential being approximately 20 mV more positive than the resting membrane potential. These properties are consistent with neurons functioning with low excitability. No evidence for ramp-like pacemaker activities has been found within mammalian intrinsic cardiac neurons in vitro. Thus spontaneous activity generated by such neurons in vivo likely reflects underlying cell-cell interactions. For orthodromic stimulation there is substantial dispersion in time of the evoked excitatory postsynaptic potentials (EPSP's) generated by a given intrinsic cardiac neuron, indicative of polysynaptic inputs to neurons within the intrinsic cardiac nervous system. After the generation of action potentials, prolonged afterhyperpolarizations are produced by these cells, an additional factor which limits the excitability of intrinsic cardiac neurons in situ. Chronic disruptions of nerve inputs to these ganglia evokes changes in membrane properties which may result in increased excitability within the ganglionated plexus.
Intracellular recordings from isolated aggregates of intrinsic cardiac ganglia have identified both cholinergic and non-cholinergic synaptic mechanisms coexisting within intrinsic cardiac ganglia. In rats and pigs only fast excitatory postsynaptic potentials are displayed by intrinsic cardiac neurons in response to orthodromic stimulation of closely adjacent intraganglionic axons. These postsynaptic potentials are substantially attenuated, but not completely eliminated, by nicotinic cholinergic blockade. In the dog, orthodromic stimulation of presynaptic fibers in these nerves elicits fast and slow postsynaptic potentials within intrinsic cardiac neurons. Fast excitatory postsynaptic potentials are mediated by cholinergic nicotinic receptors, while the slow excitatory and slow inhibitory potentials are mediated by cholinergic muscarinic receptors. In the pig direct application of norepinephrine modifies the properties of about 25% of identified intrinsic cardiac neurons. These data indicate that intrinsic cardiac neurons possess muscarinic cholinergic, nicotinic cholinergic as well as adrenergic receptors. As detailed hereinafter, many other putative neurotransmitters likewise modify electrical events of intrinsic cardiac neurons. These neurochemicals may play important roles in the modulation of intrinsic cardiac neuronal activity.
In summary, intrathoracic autonomic ganglia do not function as obligatory synaptic stations for autonomic efferent neuronal input to the heart. Instead, they are capable of complex signal integration involving afferent, local circuit as well as parasympathetic and sympathetic efferent neurons. While the physiological properties of extracardiac autonomic ganglia tend to amplify CNS and afferent feedback inputs, those of the intrinsic cardiac nervous system act to limit cardiac excitability. As such, the final common pathway of cardiac control—the intrinsic cardiac nervous system—appears to function as a “low pass” filter to minimize transient neuronal imbalances arising from separate sympathetic and parasympathetic efferent neuronal inputs to the heart. In conjunction with this local afferent feedback mechanism, neurons in intrathoracic ganglia also mediate local cardio-cardiac reflexes at sites separate from those on the heart and the CNS. The synaptic events underlying such intraganglionic interactions involve multiple neurotransmitters that interact with various neuronal receptors to exert rapid acting neuronal membrane conductance and/or longer-term modulation of synapses within the intrinsic cardiac nervous system.
Synaptic Mechanisms Associated with Neurons in Intrathoracic Autonomic Ganglia
Cholinergic Mechanisms. Synaptic transmission in autonomic ganglia principally involves the release of acetylcholine by presynaptic terminals and subsequent binding of that neurotransmitter to nicotinic cholinergic receptors on postganglionic neurons. In mammals this synaptic junction is not obligatory, indicating that a significant convergence of inputs may be necessary to evoke postganglionic neuronal activity. Thus the potential for synaptic integration exists within intrathoracic autonomic ganglia. Nicotinic and muscarinic cholinergic receptors have been associated with intrathoracic autonomic neurons. Furthermore, blockade of nicotinic receptors attenuates, but does not eliminate, activity generated by the intrinsic cardiac neurons. Muscarinic blockade attenuates excitatory and inhibitory synaptic function within intrinsic cardiac ganglia as well. These sets of data indicate that acetylcholine exerts both mediator and modulator effects at synapses within intrathoracic autonomic ganglia.
Application of nicotine to intrathoracic autonomic neurons can alter their activity and induce concomitant changes in regional cardiac function, whether the neurons are located in extracardiac or intrinsic cardiac ganglia. Nicotinic activation of intrinsic cardiac neurons evokes a biphasic cardiac response, with initial suppression in regional cardiac function being followed by augmentation. Acute decentralization of intrathoracic ganglia from the CNS attenuates, but does not eliminate, such effects. In time, following chronic decentralization of intrathoracic ganglia including those on the heart as with cardiac transplantation, peripheral nerve networks remodel to sustain cardiac function. For cholinergic receptor systems, the remodeling primarily involves augmentation of excitatory influences mediated by muscarinic receptors.
Non-cholinergic Mechanisms. Blockade of nicotinic cholinergic receptors attenuates, but does not eliminate, the activity generated by neurons within the intrathoracic autonomic ganglia. These data indicate that non-nicotinic putative neurotransmitters act as mediators for synaptic transmission within the intrathoracic neuronal system. Anatomical and physiological studies have identified multiple putative neurotransmitters in association with the mammalian intrinsic cardiac ganglia which include purinergic agonists, catecholamines, angiotensin II, calcitonin gene-related peptide, neuropeptide Y, substance P, neurokinins, endothelin and vasoactive intestinal peptide. Many of these putative neurochemicals arise from neurons whose cell bodies are located in stellate, middle cervical or mediastinal ganglia, while others may be synthesized by neurons intrinsic to the heart. Direct application of various neurotransmitters adjacent to neurons in intrinsic cardiac ganglia modifies the activities they generated, often resulting in concomitant changes in cardiac pacemaker and/or contractile behavior.
Intrinsic cardiac ganglia contain a heterogeneous population of neurons that utilize cholinergic and non-cholinergic synapses to control intraganglionic, interganglionic and nerve effector organ cell activities. Some of these neurotransmitters subserve short duration synaptic actions (e.g., acetylcholine) while others modulate pre- and/or post-synaptic function over longer periods of time (e.g., neuropeptide Y).
Neural Remodeling in the Heart Associated with Myocardial Ischemia
Myocardial ischemia and infarction can induce substantial changes in the intrathoracic nerve networks and their reflex control of regional cardiac function. Chen and co-workers [(41; 42; 122)] have recently proposed the sprouting hypothesis of sudden cardiac death: namely, “Myocardial Ischemia results in nerve injury, followed by sympathetic nerve sprouting and regional myocardial hyperinnervation. The coupling between augmented sympathetic nerve sprouting with electrical remodeled myocardium results in VT, VF and SCD.” The results of these studies and others have indicated that the evolution of cardiac pathologies may be associated with a heterogeneous distribution of efferent projections to cardiac end-effectors. Myocardial ischemia may also alter the neurochemical profile of that innervation; e.g., expression of vasoactive intestinal polypeptide and calcitonin gene-related peptide is enhanced in sympathetic neurons after myocardial infarction. Finally, the evolution of cardiac pathology can be associated with disruptions of the intrinsic cardiac nervous system and its ability to process afferent information. Such changes compromise the abilities of the peripheral nerve networks to maintain homogeneity for reflex control of regional cardiac function. This neural remodeling, when coupled with the ischemic-induced heterogeneous electrical remodeling of cardiac myocytes, creates a synergistic substrate for arrhythmias and sudden cardiac death.
Myocardial ischemia reflects an imbalance in the supply:demand balance within the heart with resultant activation of cardiac afferent neurons and, as a consequence, the perception of symptoms (i.e., angina pectoris). In addition to such nociceptive responses, activating cardiac afferent neurons can elicit autonomic and somatic reflexes. Pharmacological, surgical and angioplasty therapies are commonly used to improve symptoms and cardiac function in patients exhibiting angina pectoris. While these treatments are usually successful, some patients still suffer from cardiac pain following these procedures. Recently, epidural stimulation of the spinal cord (SCS or Dorsal Cord Activation, DCA) has been suggested as an alternative to bypass surgery in high-risk patients. With DCA, high frequency, low intensity electrical stimuli are delivered to the dorsal aspect of the T1-T3 segments of the thoracic spinal cord. This therapy decreases the frequency and intensity of anginal episodes. DCA reduces the magnitude and duration of ST segment alteration during exercise stress in patients with cardiac disease, improves myocardial lactate metabolism and increases workload tolerance. The mechanisms whereby this mode of therapy produces such beneficial effects are, to date, poorly understood and although used extensively in Europe, are not a standard of practice within the United States.
Since intrathoracic cardiac neurons have been found to play important modulatory roles in cardiac regulation, the use of DCA and its effects on the activity generated by intrinsic cardiac neurons has been studied and is at least one component of the presently claimed and disclosed invention. Transient cardiac ventricular ischemia increases the activities generated by intrathoracic ganglia, including those on the heart. Excessive focal activation of intrathoracic neural circuits can induce cardiac dysrhythmias, even in normally perfused hearts. DCA results in an immediate suppression in intrinsic cardiac neuronal activity. A neuro-suppressor effect imposed in the intrinsic cardiac nervous system occurs whether DCA is applied immediately before, during or after coronary artery occlusion (
The synaptic mechanisms and specific pathways mediating these responses likely involve both sympathetic afferent and efferent neurons. Dorsal cord activation excites sensory afferent fibers antidromically such that endorphins or neuropeptides such as calcitonin gene-related peptide or substance P are locally released in the intrinsic cardiac ganglia and myocardium. Opiates and neuropeptides can also influence intrinsic cardiac neurons (see herein-above). Spinal cord stimulation also suppresses intrinsic cardiac adrenergic as well as local circuit neurons as the result of altered sympathetic efferent preganglionic neuronal activity. It is also known that activation of sympathetic efferent preganglionic axons suppresses many intrathoracic reflexes that are involved in cardiac regulation. Thus these neuro-suppressor effects appear to be due, in part, to activation of inhibitory synapses within intrathoracic ganglia. Recent clinical experience with DCA highlights the dynamic interactions that can occur between central and intrathoracic neurons, demonstrating the potential for effective clinical treatment of cardiac pathology via modulation of the intrathoracic nervous system or the intrinsic cardiac nervous system.
Coordination of Activities within and Between Ganglia of the Intrinsic Cardiac Nervous System
As shown in
Differential Control of Neurotransmitter Release within the Cardiac Interstitium
Exogenous administration of ANG II into the blood supply for the right atrial ganglionated plexus increased NE concentration in the cardiac interstitial fluid (ISF) to the same extent as achieved during electrical stimulation of the stellate ganglia (
Disruption of nerve projections to or within the intrinsic cardiac nervous (ICN) system are associated with alterations in the passive and active properties of the cardiac neurons. chronic interruption of the extrinsic nerve inputs to the ICN has been shown to produce changes in membrane properties that lead to increased network excitability within this ganglionated plexus. Intrinsic cardiac neurons remain responsive to cholinergic synaptic inputs. The cholinergic receptor systems are differentially affected by disruption of nerve inputs to the ICN, with muscarinic responsiveness being enhanced (
Data indicate that the progression of cardiac disease is associated with myocyte and neural remodeling. The neural remodeling likely includes degenerative and regenerative aspects. The net result is the potential for heterogeneous innervation to various regions of the heart. Chen et al. have therefore proposed the “nerve sprouting hypothesis of sudden cardiac death”. As illustrated in
Interactions within the intrinsic cardiac nervous system depend in large part on common shared afferent inputs and/or interconnections mediated via local circuit neurons. The degree of coordination between aggregates of intrinsic cardiac neurons is influenced by proximity and the activation state of afferent inputs. In basal states, the degree of coherence of activity within a single cardiac ganglia waxes and wanes with a periodicity of approximately 20-30 sec. In response to enhanced neuronal activity, evoked during activation of their associated sensory inputs, that coherence increases. For neurons contained within different intrinsic cardiac ganglia, lesser degrees of coherence of basal activity between them, but this coherence increases during stimulation of afferent inputs owing to common shared inputs between them.
There are two distinct classes of sensory input affecting ICN activity: a phasic input, whose influence is short-lived and subserves rapid feedback processes within the ICN and a dynamic input whose influence is determined by the context/history of its activation and whose influence on ICN activity is long-lived. Mechano-sensitive neurites subserve the phasic inputs and chemo-sensitive neurites subserve the neural “memory”.
Myocardial ischemia and infarction induce substantial changes in the intrathoracic nerve networks and their reflex control of regional cardiac function including protection and stabilization of electrical activity of the heart. Chronic myocardial ischemia induces a heterogeneous distribution of efferent projections to cardiac end-effectors. Heterogeneous distribution of sympathetic fibers to the left ventricle results in similar heterogeneous release of catecholamines into the interstitial space during stimulation of the efferent nerves. Myocardial ischemia alters the neurochemical profile of that innervation, with differential increases in neuropeptide content within subsets of neurons contained within the intrinsic cardiac nervous system. The evolution of cardiac pathology is associated with disruptions of the intrinsic cardiac nervous system and its ability to process afferent information and such changes are more evident in the CMVPG than the RAGP intrinsic cardiac ganglia. Animals that exhibit indices of higher vagal tone (higher baroreflex sensitivity and higher heart rate variability) demonstrate lesser degrees of ischemic-induced neural remodeling.
DCA can exert long-term modulation of the activities within the intrinsic cardiac nervous system. While initial studies indicate that catecholamines are released in response to DCA, it is anticipated that DCA also will activate sensory afferent fibers antidromically such that endorphins or neuropeptides such as calcitonin gene-related peptide or substance P are locally released in the intrinsic cardiac ganglia and myocardium. Opiates and neuropeptides can also influence the intrinsic cardiac neurons. DCA also suppresses intrinsic cardiac adrenergic as well as local circuit neurons via altered sympathetic efferent preganglionic neuronal input. Activation of sympathetic efferent preganglionic axons suppresses many intrathoracic reflexes that are involved in cardiac regulation. Thus these neuro-suppressor effects may be due, in part, to activation of inhibitory synapses within intrinsic ganglia.
Heterogeneous alterations within the intrinsic cardiac ganglia or at the end-terminus of the autonomic innervation to the ischemic myocardium are major contributors to the increased incidence of sudden cardiac death in patients with coronary artery disease. Chronic DCA ameliorates ischemia-induced remodeling within the intrinsic cardiac nervous system and thereby reduces the heterogeneous neural substrate that predisposes the susceptible animals to ventricular arrhythmias and sudden cardiac death.
Control of regional cardiac electrical and mechanical function is dependent upon varied neural inputs from intrathoracic autonomic ganglia, the spinal cord and brainstem, as well as by circulating neurohumoral agents. Neural control of the heart is dependent upon the coordination of activity generated by neurons within intrathoracic autonomic ganglia and the CNS. The hierarchy of nested feedback loops therein provides precise beat-to-beat control over regional cardiac function. Within the hierarchy of intrathoracic ganglia and nerve interconnections, complex processing takes place that involves the summation of preganglionic inputs from central neurons with those derived from cardiopulmonary sensory inputs.
Excessive activation of the intrathoracic cardiac efferent nervous system can provoke cardiac arrhythmias, as can myocardial ischemia. These maladaptations likely involve changes within the cardiac nervous system in addition to alterations in cardiomyocyte function. Differential adaptations of cardiomyocyte ion channels (e.g., IK and ICa) and intercellular connections during the progression of cardiac disease have been termed “electrical remodeling.” Recent data indicates that neurohumoral control mechanisms likewise reorganize during progression into certain cardiac diseases and are referred to as “neurohumoral remodeling.”
Changes in autonomic outflow accompany and influence the progression of cardiac disease. Sympathetic efferent neuronal activation contributes to sudden cardiac death in patients with ischemic as well as non-ischemic heart disease. The ATRAMI study demonstrates that baroreflex sensitivity and heart rate variability predict risk for cardiovascular mortality and myocardial infarction. Electrical stimulation of vagal efferent neurons suppresses the tendency to ventricular fibrillation formation in dogs with depressed vagal reflex activity as measured by baroreflex sensitivity. Yet, pharmacological agents that increase vagal efferent neuronal tone, such as a low-dose scopolamine, do not confer similar degrees of protection.
The mechanism(s) whereby activation of sympathetic efferent neurons and/or withdrawal of parasympathetic efferent neuronal tone increase the risk for sudden death are not clear. However, post-infarction heterogeneous remodeling of cardiac innervation, including extracardiac sympathetic and intrinsic cardiac efferent neural elements, likely contributes to the resultant cardiac electrical instability. The present claimed and disclosed invention, as disclosed herein, outlines the evolution of neural remodeling associated with chronic myocardial ischemia and infarction and thus provides a stepping off point for the development of treatments for cardiac pathologies utilizing SCS or DCA.
After decades of progress, improvement in the management of cardiac arrhythmias appears to have leveled off. The problem of sudden cardiac death occurring as the result of an initial arrhythmic event has not been addressed (except perhaps through palliative public health strategies which include public access defibrillators, PAD). This state of affairs is due, in part, to the fact that key pieces of information regarding cardiac arrhythmia formation are still missing, including the ability to identify the apparently normal individual at risk before an event occurs. While changes in myocardial electrical events have been well characterized in the diseased heart, information concerning the complex neuronal organization regulating cardiac rhythm remains limited. The comprehensive presently claimed and disclosed invention which includes the knowledge of the complex processing which occurs within the intrathoracic nervous system, as well as between peripheral and central cardiovascular neurons, provides a basis for understanding the role that the cardiac nervous system plays in regulating the electrical behavior of not only the normal heart, but the diseased heart as well, thus providing for novel therapeutic approaches for the effective treatment of cardiac arrhythmias, sudden cardiac death or syncope of cardiac origin by targeting discrete populations of neurons regulating regional cardiac behavior.
Control of regional cardiac function is dependent upon properties intrinsic to cardiac electrical and mechanical tissues as modulated by neuronal reflexes arising at the level of the intrinsic cardiac and intrathoracic extracardiac nervous systems, in addition to well-known spinal cord and brainstem reflexes. The proper function of this cardiac neuronal hierarchy is ultimately dependent on ongoing cardiovascular sensory and spinal cord neuronal inputs. The synergism of function within the cardiac autonomic hierarchy and cardiac myocytes results in a finely balanced, rapidly responsive control system that is continuously being upgraded to maintain adequate cardiac output. As outlined hereinabove, the intrinsic cardiac ganglia form the principal final common pathway for autonomic modulation of regional cardiac function. Maintenance of cardiac output depends not only on the Frank-Starling mechanism and circulating catecholamines, but also on inputs from this nervous system. Disruptions of the sensory inputs to the hierarchy of autonomic neurons regulating the heart due to alterations in the mechanical and/or chemical milieu of the heart can be associated with compromised control of the heart.
Divergent populations of cardiac neurons within different intrathoracic ganglia interact on an ongoing basis to maintain adequate cardiac output, requiring little ongoing input from spinal cord neurons. neurons in this hierarchy interact to regulate normal cardiac function on a beat-to-beat basis. The development of novel strategies to manage cardiac disease necessitates not only a thorough understanding of the processing of information arising from cardiac and major intrathoracic vascular sensory neurites, but also inputs from central neurons. Neurons in the spinal control exert preferential control over such intrathoracic neuronal processing of cardiac sensory information.
Human studies have shown that stimulation of the dorsal T1-T2 segments of the spinal cord suppresses angina pectoris (sensory information arising from the heart) without masking awareness of acute myocardial ischemic episodes. The mechanisms whereby activation of the dorsal aspect of the cranial thoracic spinal cord produces improved cardiac function and reduces symptoms of the ischemic myocardium are not currently understood. The experiments and resulting data from the presently claimed and disclosed invention show that the anti-anginal and cardiac stabilization effects of such spinal cord modulation are mediated via stabilization of the intrathoracic nervous system, especially its intrinsic cardiac component. Neural control of cardiac function resides in the network of nested feedback loops made up of the intrinsic cardiac nervous system, extracardiac intrathoracic autonomic ganglia, the spinal cord and brainstem. Within this hierarchy, the intrinsic cardiac nervous system functions as a distributive processor at the level of the target organ. The redundancy of function and non-coupled behavior displayed by neurons in intrathoracic extracardiac and intrinsic cardiac ganglia minimizes the dependency for such control on a single population of peripheral autonomic neurons. On the other hand, network interactions occurring within the intrinsic cardiac nervous system integrate parasympathetic and sympathetic efferent outflow with cardiovascular afferent feedback to modify cardiac rate and regional contractile force throughout each cardiac cycle. Thus, neural control of cardiac function resides in the network of nested feedback loops made up of the intrinsic cardiac nervous system as well as the extracardiac intrathoracic nervous system, spinal cord and brainstem (
The redundancy of function and non-coupled behavior displayed by neurons in intrathoracic extracardiac and intrinsic cardiac ganglia minimizes the dependency for such control on a single population of peripheral autonomic neurons. Furthermore, network interactions occurring within the intrinsic cardiac nervous system integrate parasympathetic and sympathetic efferent outflow with afferent feedback to modify cardiac rate and regional contractile force throughout each cardiac cycle.
The intrinsic cardiac nervous system has been classically considered to contain only parasympathetic efferent postganglionic neurons that receive inputs from medullary parasympathetic efferent preganglionic neurons. As such, intrinsic cardiac ganglia are viewed as simple relay stations and major autonomic neuronal control of the heart is purported to reside solely in the brainstem and spinal cord. However, current data indicates that centripetal as well as centrifugal processing occurs within the mammalian intrathoracic nervous system. This involves afferent neurons, local circuit neurons (i.e., neurons that interconnect neurons within one ganglion and neurons in different intrathoracic ganglia), as well as sympathetic and parasympathetic efferent postganglionic neurons. The divergent populations of neurons within the intrinsic cardiac nervous are influenced by spinal cord neurons on an ongoing basis in the maintenance of adequate cardiac output.
The development of novel therapeutic strategies to manage abnormal cardiac states necessitates a thorough understanding of not only of the processing of information arising from sensory neurites in various regions of the heart and great thoracic vessels, but how spinal control neurons exert preferential control over the intrathoracic cardiac nervous system with particular reference to its target organ. Similarly, intrathoracic extracardiac sympathetic ganglia have been thought to act solely as efferent relay stations for sympathetic efferent projections to the heart. However, recent anatomical and functional data indicate the presence of the multiple neuronal subtypes within the intrinsic cardiac nervous system. The intrathoracic nervous system, including its intrinsic cardiac component, is made up of different neuronal subtypes. These include afferent, local circuit as well as adrenergic and cholinergic efferent postganglionic neurons. These neurons form the intrathoracic component of the central and peripheral neuronal feedback loops that regulate regional cardiodynamics on a beat-to-beat basis.
The intrinsic cardiac nervous system consists of multiple aggregates of neurons and associated neural interconnections, localized to discrete atrial and ventricular regions. Among these distinct ganglionated plexuses, preferential control of specific cardiac functions has been identified. For example, right atrial ganglionated plexus (RAGP) neurons have been associated with primary, but not exclusive, control of SA nodal function and inferior vena cava-inferior atrial ganglionated plexus neurons primarily, but not exclusively, with control of AV nodal function. One population of intrinsic cardiac neurons, the parasympathetic postganglionic ones, receives direct inputs from medullary parasympathetic preganglionic neurons. Another population, adrenergic efferent ones, receives inputs from more centrally located neurons in intrathoracic ganglia and the spinal cord. That ventricular sensory neurites continue to influence the activity generated by neurons on the heart following chronic decentralization of the intrinsic cardiac nervous system has been interpreted as indicating that the somata of afferent neurons are located within the intrinsic cardiac nervous system, some of which project axons to central neurons. This latter concept has received anatomical confirmation. Intrinsic local circuit neurons interconnect cardiac afferent to efferent neurons that innervate each region of the heart.
Neurons in intrathoracic ganglia, including those on the heart, receive constant inputs not only from spinal cord neurons, but also from cardiac afferent neurons to modulate cardiac efferent neurons. The activity generated by most intrinsic cardiac neurons increases markedly in the presence of focal ventricular ischemia. Furthermore, excessive activation of limited populations of intrinsic cardiac neurons induces cardiac dysrhythmias that lead to ventricular fibrillation, even in normally perfused hearts. Therapies that act to stabilize such heterogeneous evoked activities within cardiac reflex control circuits have obvious clinical importance. Proper information exchange among the intrathoracic components of the cardiac nervous system act in concert to stabilize the electrical and mechanical behavior of the heart, particularly in the presence of focal ventricular ischemia. Thus, use of SCS or DCA is a means to stabilize the heart prior or post ischemia. An object of the present invention is to provide such treatment methodologies.
Consistent coherence of activity generated by differing populations of neurons is indicative of principal, direct synaptic interconnections between them or, conversely, the sharing by such neurons of common inputs. Such relationships have been identified among medullary and spinal cord sympathetic efferent preganglionic neurons, as well as among different populations of sympathetic efferent preganglionic neurons. Different populations of neurons, distributed spatially within the intrathoracic cardiac nervous system, respond to cardiac perturbations in a coordinate fashion. If neurons in one part of this neuronal network respond to inputs from a single region of the heart, such as the mechanosensory neurites associated with a right ventricular ventral papillary muscle, then the potential for imbalance within the different populations of neurons regulating various cardiac regions might occur and, thus, its neurons would display little coherence of activity. In other words, relatively low levels of specific inputs on a spatial scale to the intrathoracic cardiac nervous system would result in low coherence among its various neuronal components. On the other hand, excessive input to this spatially distributed nervous system would destabilize it, leading to cardiac arrhythmia formation, etc.
One must know how neurons in intrinsic cardiac versus intrathoracic extracardiac ganglia interact to regulate regional contractile function in order to understand not only the complexity of cardiac control, but also how the cardiac neuroaxis can be targeted therapeutically to manage specific cardiac disease entities. Over the past 30 years studies of the anatomy and function of the peripheral cardiac nervous system have taken place, focusing during the last decade on its intrinsic cardiac component. The classical view of the autonomic nervous system presumes that its intrinsic cardiac component comprises a parasympathetic efferent neuronal relay station in which medullary preganglionic neurons synapse with parasympathetic efferent postganglionic neurons therein. In such a concept, the latter neurons project to end effectors on the heart with little or no integrative capabilities occurring therein. Similarly, intrathoracic paravertebral ganglia have been thought to represent synaptic stations for sympathetic efferent postganglionic neurons controlling the heart.
The intrinsic cardiac nervous system functions, according to the presently claimed and disclosed invention, as a distributive processor at the level of the target organ. The redundancy of function and non-coupled behavior displayed by neurons within intrathoracic extracardiac and intrinsic cardiac ganglia minimizes the dependency for such control on a single population of peripheral autonomic neurons. In that regard, network interactions occurring at the level of the heart integrate parasympathetic and sympathetic efferent inputs with local afferent feedback to modify cardiac rate and regional contractile force throughout each cardiac cycle. A recent editorial by David Lathrop and Pete Spooner of the NIH highlights the potential clinical relevance of altered processing of information by these populations of neurons such that a lack of coordination of data exchange within the cardiac neuronal axis may lead to the genesis of cardiac arrhythmias. Hence the importance of determining how neurons in intrathoracic extracardiac and intrinsic cardiac ganglia interact in the maintenance of adequate cardiac output.
The different populations of neurons distributed spatially within the intrathoracic cardiac nervous system respond to cardiac perturbations in a complex fashion. Neurons in intrathoracic extracardiac ganglia do not respond to cardiac perturbations in a fashion similar to that displayed by intrinsic cardiac ones. Consistent coherence of activity generated by differing populations of neurons has been identified among medullary and spinal cord sympathetic efferent preganglionic neurons, as well as among different populations of sympathetic efferent preganglionic neurons. A relatively low level of inputs on a spatial scale to one population of intrathoracic cardiac neurons results in low coherence among its components. In contrast, excessive input to this spatially distributed nervous system destabilizes it, leading for instance to cardiac arrhythmia formation. Since neurons in one part of the intrathoracic neuronal network respond solely to inputs from a single region of the heart, such as from mechanosensory neurites in a right ventricular ventral papillary muscle, then the potential for imbalance within the different populations of neurons in various levels of the intrathoracic neuronal hierarchy arises.
Ultimately, the outflows of efferent neuronal signals to the various regions of the heart depend to a large extent on the direct or indirect inputs they receive from cardiac and major intrathoracic vascular sensory neurites in addition to pulmonary mechanosensory neurites. The redundancy of function and non-coupled behavior displayed by neurons in intrathoracic extracardiac and intrinsic cardiac ganglia minimizes the dependency for regional cardiac control on a single population of intrathoracic neurons. This may be particularly relevant with respect to supporting the output of the ischemic heart. In that regard, network interactions occurring among intrathoracic extracardiac and intrinsic cardiac neurons secondary to inputs from cardiovascular afferent neurons involve local circuit neurons feeding information foreword to cardiac parasympathetic and sympathetic efferent neurons. These network interactions are under the constant influence of spinal cord neurons
Overview of cardiac sensory neuronal transduction. It has been known for some time that cardiac sensory neurites (nerve endings) are associated with somata located in ganglia relatively distant from the heart, nodose and dorsal root ganglia. It has recently become evident that cardiac sensory neurites are also associated with somata located in intrathoracic ganglia, including those on the heart. The relative distance between these sensory neurites and their associated somata represents a major determinant of their function. The somata of many cardiac afferent neurons located near to or on the target organ display high frequency (phasic) activity that directly affects target organ efferent neurons (TABLE I). In this manner, high fidelity information content can exert rapid control over efferent neurons adjacent to or on the heart that modulate regional contractility. In contrast, cardiac afferent neuronal somata located relatively distant from their sensory neurites (i.e., in nodose or dorsal root ganglia) are, of necessity, involved in longer latency influences on second order neurons in the cardiac neuroaxis. These relatively distant cardiac afferent neurons, as such, are involved in relatively long latency cardio-cardiac reflexes, being spatially removed from the target organ they display memory. A division of cardiac sensory neuronal function into two broad, functional categories can be based on spatially derived cardiac sensory transduction (TABLE I). It should be noted that some cardiac sensory neurons within dorsal root ganglia generate high frequency phasic activity, particularly when their sensory neurites are exposed to increasing concentrations of local chemicals.
That two broad categories of cardiac afferent neurons exist (TABLE I) indicates unique transduction capabilities such that cardiac information provided to second order cardiac neuroaxis neurons depends not only on the location of their sensory neurites, but on the location of their somata. The sensory information transduced by fast responding cardiac afferent neurons, impinging as it does directly on cardiac motor neurons, is one of the primary determinants of the input function to cardiac efferent neurons that coordinate regional cardiac behavior. Fast responding cardiac sensory neurons normally generate relatively high frequency (10-100 Hz) activity patterns reflective of regional cardiodynamics. Slow responding cardiac afferent neurons generally transduce alterations primarily in the local chemical milieu and, thus, by their nature are generally not responsive to regional alterations that occur on a short time scale. During physiological states, they generate tonic activity at lower frequencies (0.1-1 Hz).
The sensory neurites associated with intrinsic afferent neuronal somata are located in atrial and ventricular tissues, as well as the adventitia of major coronary arteries. The sensory neurites associated with the somata of afferent neurons in intrathoracic extracardiac ganglia are concentrated in the same cardiac regions, in addition to being found around the origins, of vena cava and on the thoracic aorta. Cardiovascular afferent neurons within the thorax provide feed-forward information to efferent neurons in intrathoracic ganglia, some via local circuit neurons. Those in the nodose ganglion influence medullary nucleus tractus solatarius neurons, while those in dorsal root ganglia influence spinal cord neurons. As discussed hereinabove, the varied transduction properties displayed by cardiac afferent neurons in nodose, dorsal root and intrathoracic ganglia reflect to a considerable extent the anatomical location of their somata, i.e., the distance between their somata and associated sensory neurites. Cardiac afferent neurons with somata close to or on the heart influence cardiac efferent neurons to initiate short-loop reflexes with short latencies of activation while those located in nodose ganglia initiate longer latency reflexes. That is why cardiac afferent neurons in intrathoracic ganglia display different transduction capabilities than those in dorsal root and nodose ganglia.
Afferent axons arising from cardiac or intrathoracic vascular sensory neurites vary in diameter (degree of myelination), according to the location of the cardiopulmonary nerve in which they course. For instance, most aortic mechanosensory neurites are associated with Ad axons, most of which are located in the intrathoracic dorsal cardiopulmonary nerve. Many ventricular sensory neurites are associated with c class axons. On the other hand, carotid artery mechanosensory neurites associated with afferent axons are divisible into the A□ and C fiber categories, each population displaying unique transduction properties.
Function. The majority of cardiac sensory neurons, particularly those located distant from the heart, generate sporadic, low frequency activity. As the activity generated by the most cardiac afferent neurons is of low frequency (i.e., 0.01-0.1 Hz), information content cannot reside in the interspike intervals of activity. Rather, it resides primarily in their average activity over time unless their activity becomes entrained to cardiodynamics in the presence of increased sensory neurite chemical milieu. Information transduced by multimodal sensory neurites associated with each axon connected to individual cardiac afferent neuronal somata also depends on their cardiac spatial distribution. Arrays of atrial sensory neurites are concentrated in the region of the sino-atrial node (right atrium) and the dorsal aspects of both atria, others are scattered throughout the rest of the atria. Ventricular sensory neurites are concentrated in the outflow tracts of the two ventricles as well as the right and left ventricular papillary muscles. Another concentration of sensory neurites is located in the adventitia on the inner arch of the thoracic aorta.
Intrathoracic cardiac afferent neurons influence (via intrathoracic local circuit neurons) cardiac efferent postganglionic neurons with latencies as short as 40 milliseconds. Nodose ganglion cardiac afferent neurons influence cardiac parasympathetic efferent preganglionic neurons in the medulla via short latency reflexes (75 ms) as well. On the other hand, dorsal root ganglion cardiac afferent neurons influence sympathetic efferent postganglionic neurons via longer latency (100-500 ms) reflexes. Thus, the differing populations of cardiac sensory neurons located at each level of the cardiac neuroaxis not only displaying unique transduction characteristics, but subserve cardio-cardiac reflexes that of necessity differ in latency and form.
Nodose ganglion afferent neurons. Using neuroanatomical tracing techniques, about 500 somata associated with cardiac sensory neurites have been identified throughout the right and left nodose ganglia. Their axons belong to the Ad and c classes, as defined by Erlanger and Gasser. Histochemical evidence indicates that the somata of nodose ganglion afferent neurons express receptors for a variety of neurochemicals, including adenosine, bradykinin and substance P receptors. Most of these cardiac afferent neurons transduce multiple chemicals, including purinergic agents such as adenosine (
Doral root ganglion afferent neurons. Despite the widely held opinion that the majority of cardiac afferent neurons are located primarily in left-sided dorsal root ganglia, anatomic evidence indicates that cardiac afferent neurons are distributed relatively equally among right and left dorsal root ganglia from the C6 to the T6 levels of the spinal cord. Afferent neuronal somata lie scattered predominantly, but not exclusively, around the centrally located axons in these ganglia. Over 500 cardiac sensory neurons have been identified anatomically in canine dorsal root ganglia from the T1 to the T3 levels of the spinal cord, ganglia containing up to 50 cardiac afferent neuronal somata. The axons connecting cardiac sensory neurites with somata in dorsal root ganglia belong to the Ad or c classes of axons, each having little bearing on their sensory transduction capabilities.
Intrathoracic extracardiac ganglion afferent neurons. Functional evidence indicates the presence of cardiac sensory neuronal somata in stellate, middle cervical and mediastinal ganglia. Axons connecting atrial or ventricular mechanosensory neurites with somata in extracardiac ganglia belong in the Ad class of axons. Those connecting intrathoracic vascular mechanosensory neuritis with somata in intrathoracic extracardiac ganglia belong to Ad class of axons as well. On the other hand, ventricular endocardial mechanosensory neurites connected with somata in intrathoracic ganglia belong to c class axons. Cardiac and aortic chemosensory neurites connected with somata in intrathoracic ganglia also belong to Ad class axons.
Intrinsic cardiac ganglion afferent neurons. Unipolar neurons are located throughout atrial and ventricular intrinsic cardiac ganglionated plexuses. Based on anatomical and functional data, the somata of some intrinsic cardiac afferent neurons project axons centrally; the remainder interacting directly with other intrinsic cardiac neurons exclusive of central neuronal inputs. Sensory neurites associated with intrinsic cardiac afferent neurons are located in all four chambers of the heart (particularly in the cranial aspect of the ventricles) are multimodal in nature (transduce mechanical and chemical stimuli). Unfornunately, little is currently known of their transduction capabilities.
Cardiac afferent neurons with sensory neurites located primarily in the atria and the outflow tracts of the ventricles or major intrathoracic vessels initiate short, intermediate and relatively long duration cardiovascular-cardiac reflexes, depending on their multimodal transduction capabilities.
Cardiac sensory input to the multiple nested feedback loops within the intrathoracic cardiac neuronal axis displaying redundancy of function and non-coupled behavior within the different anatomical levels of this hierarchy (
Intrathoracic cardiac afferent neurons are multimodal in nature (i.e., responsive to local mechanical and chemical stimuli), transducing a host of chemicals that include ion channel modifying agents (i.e., veratridine; c.f.,
On the other hand, neurons in intrathoracic extracardiac (middle cervical or stellate) and intrinsic cardiac ganglia do not display such function, despite the fact that neurons in intrathoracic extracardiac and intrinsic cardiac ganglia receive inputs from cardiac mechanosensory and chemosensory neurites. That is due in part because neurons in intrathoracic extracardiac ganglia receive many inputs mechanosensory neurites located on the inner arch of the aorta. That some of these neurons are still influenced by cardiac sensory inputs when decentralized from central neurons indicates that intrathoracic cardiac afferent neurons can influence the intrathoracic neuronal hierarchy independent of central neuronal inputs.
Neurons in intrathoracic extracardiac and intrinsic cardiac ganglia exhibit non-coupled behavior, even when they are mutually entrained to cardiac events by cardiovascular afferent feedback (
Direct application of neurochemical agonists or selective antagonists has been used to survey receptor subtypes associated with neurons within the intrathoracic cardiac nervous system and to characterize the functional differences of neurons within its various ganglia. Chemical stimulation of specific intrathoracic neurons with low doses of chemicals such as nicotine, neuropeptides, catecholamines, amino acids and purinergic agents can induce changes in their activity. When neuronal changes so induced are of sufficient magnitude, alterations in cardiac pacemaker, conductive and regional contractile function occur. The cardiac responses so induced reflect activation of specific populations of neurons in intrathoracic extracardiac or intrinsic cardiac ganglia as similar application of such neurochemicals to intracardiac axons of passage does not effect neuronal activity or cardiodynamics. In agreement with that, transection of all extrinsic neuronal inputs to the intrathoracic nervous system (acute decentralization) attenuates cardiac responses so elicited. This data indicates the importance of the connectivity of neurons within the thorax with central ones in mediating cardio-cardiac reflexes.
Cholinergic Mechanisms: Synaptic transmission in cardiac autonomic ganglia has been thought to be principally involved in the release of acetylcholine by presynaptic terminals and subsequent binding of that neurotransmitter to nicotinic cholinergic receptors on postganglionic neurons. In mammals this synaptic junction is not obligatory, indicating that a significant convergence of inputs may be necessary to evoke postganglionic activity. Nicotinic and muscarinic cholinergic agonists and antagonists modify intrinsic cardiac neurons in vitro and in vivo, as well as neurons in intrathoracic extracardiac ganglia. Local application of nicotine to intrinsic cardiac or intrathoracic extracardiac neurons induces alterations in cardiac rate and regional contractile function. Activation of intrinsic cardiac neurons with nicotine induces either augmenter or depressor cardiac effects, depending on the population of neurons so affected. Blockade of nicotinic receptors attenuates, but does not eliminate, these cardiac reflexes. Muscarinic cholinergic blockade attenuates synaptic function within intrathoracic ganglia, as well, indicating that acetylcholine exerts both mediator and modulator effects at synaptic junctions within intrathoracic ganglia.
Noncholinergic Mechanisms: Blockade of nicotinic cholinergic receptors attenuates, but does not eliminate synaptic transmission within intrathoracic ganglia indicating that non-nicotinic synapses act as primary mediators of synaptic transmission within the intrathoracic nervous system. Anatomical and physiological studies have identified multiple putative neurotransmitters in association with neuronal somata in mammalian intrathoracic extracardiac and intrinsic cardiac ganglia. These chemicals include purinergic agents (adenosine and ATP), alpha- and beta-adrenergic agonists, angiotensin II, bradykinin, calcitonin gene-related peptide, neuropeptide Y, histamine, serotonin, substance P and vasoactive intestinal peptide as well as excitatory and inhibitory amino acids. Many of these putative neurotransmitters arise from neurons whose cell bodies are intrathoracic extracardiac (stellate and middle cervical) ganglia, while other may be synthesized by neurons intrinsic to the heart. Direct application of various putative neurotransmitters adjacent to neurons in intrinsic cardiac or intrathoracic extracardiac ganglia modifies cardiac pacemaker and contractile activities. Such responses presumably reflect varied receptor mediated activation of adjacent neurons and their associated dendrites since when identical concentrations of these neurochemicals are applied directly to intracardiac axons of passage neuronal activity and cardiac indices remain unaffected.
Thus, when taken together, synapses interconnecting intrathoracic afferent, local circuit and efferent neurons utilize a host of neurochemicals in the regulation of regional cardiodynamics, even when disconnected from the influence of central neurons.
Memory Function within the Intrathroacic Nervous System
Cardiovascular-cardiac reflexes exert long-term control over cardiodynamics, including those initiated solely within the intrathoracic neuronal hierarchy. Intrathoracic cardio-cardiac reflexes display different latencies of activation in as much as intrathoracic cardiac afferent neurons influence local circuit neurons in ganglia at different levels of the thorax (intrinsic cardiac, mediastinal, middle cervical and stellate ganglia) following different latencies. The differing populations of cardiac afferent neurons that initiate these varied intrathoracic reflexes display unique transduction characteristics that are suited to the cardio-cardiac reflexes that they sub serve. For instance, intrathoracic cardio-cardiac reflexes have latencies as short as 40 ms, whereas those involving spinal cord neurons have latencies that exceed 450 millisecond. Thus, the relative distance between cardiac sensory neurites and neuronal somata is a significant determinant of cardio-cardiac reflex latencies they initiate.
Inherent in this issue is the fact that intrathoracic neurons involved have two types of memory: (1) The first type involves non-computational memory displayed by cardiac chemosensory neurons, in as much as the previous status of their transduction behavior is a major determinant of their responsiveness to a chemical stimulus. The slowly varying, long-term transduction capabilities exhibited by cardiac chemosensory neurites is characteristic of passive memory. (2) The second type of memory displayed in the intrathoracic nervous system is represented by active processing of sensory inputs from: (i) cardiovascular afferent neurons and (ii) central efferent preganglionic neurons. This computational memory resides in the network interactions that are dependent on intrathoracic local circuit neurons. Their state dependent memory represents hysteretic computation of cardiovascular sensory information that, along with inputs from central neurons, exerts ongoing control over cardiac efferent neurons. Such computation ability is necessary if a population of neurons is to simultaneously process information arising from many sources—an important characteristic of the intrathoracic nervous system. The presence of such complex information processing within the intrathoracic autonomic nervous system has led to the discovery that this nervous system functions as a distributive processor of centripetal and centrifugal information arising from and going to the heart that, of necessity, requires state-dependent memory.
Memory displayed by slow responding cardiac afferent neurons. Most of the cardiac afferent neuronal somata located near or on the target organ transduce high frequency (phasic) information directly to target organ efferent neurons that control regional contractile behavior. In this manner, high fidelity information content can exert rapid control over cardiac efferent neurons coordinating regional contractile patterns. On the other hand, the cardiac sensory neurites located anatomically distant from their associated somata (i.e., in nodose and dorsal root ganglia) take longer to influence second order neurons (c.f., neurons in the CNS). These latter cardiac afferent neurons are involved in relatively long latency cardio-cardiac reflexes. Presumably because of that function (lack of necessary short term influences), for the most part they display relatively long term memory function since they transduce slowly varying chemical signals. It should be noted that some intermediary cardiac afferent neurons also generate tonic activity, only generating high frequency, phasic activity when exposed to increasing concentrations of chemicals reflective of their multimodal transduction properties. The passive memory function displayed by these slow responding cardiac sensory neurons resides in the state dependent properties of their cardiac chemosensory neurites. This is also indicative of the fact that chemical excitation of these afferent neurons remains long after removal of the stimulus—yet another form of memory.
Local circuit neuronal memory function. It has been postulated that active memory resides in the multi-synaptic processing of cardiac sensory information that takes place within the intrathoracic neuronal hierarchy. This is particularly relevant with respect to the processing of cardiopulmonary sensory information by intrathoracic local circuit neurons. It has been determined that memory is displayed by remodeled intrathoracic local circuit neurons following chronic removal of their central neuronal inputs. Short duration (10 msec.) cardiopulmonary sensory inputs to neurons in chronically decentralized intrathoracic ganglia results in the activation of local circuit neurons therein for up to 2 seconds. Such data indicates a memory capacity that lasts for a number of cardiac cycles subsequent to sensory inputs arising during one cardiac cycle.
As there is little direct relationship between sensory inputs and output in the intrathoracic cardiac nervous system, its local circuit neurons act to compute state-dependent information on a beat-to-beat basis. This permits inputs from multiple sources (peripheral cardiovascular afferent neurons and spinal cord efferent neurons) to influence restricted cardiac efferent neuronal outputs to the heart in an efficient manner and over time. Thus, one function of intrathoracic local circuit neurons is key to understanding hysteretic information processing (memory) since their capacity to compute relatively minor sensory input alterations without adapting out represents an important characteristic of this neuronal hierarchy. In such a scenario, local circuit neurons function to reduce ‘noise’ to ensure restricted (not excessive) output in the presence of multiple sensory inputs.
This processing of cardiovascular sensory information by intrathoracic local circuit neurons accounts for the stability of the control exerted over regional cardiac function during relatively prolonged period of time in normal cardiovascular states. On the one hand, simple state switching among excitatory versus inhibitory neurons in this population would generate oscillatory behavior such as occurs among excitatory and inhibitory neurons in the spinal cord. This would, in fact, lead to instability of function since computational analysis would become deranged and noise reduction capabilities would be lost. On the other, memory function associated with intrathoracic local circuit neurons, driven by ongoing cardiac sensory inputs, ensures stable control over cardiac efferent neuronal outputs. For that reason, hysteretic memory related to the active processing of cardiac sensory information is important for the ultimate stability of cardiac efferent neuronal control.
Current data indicates that passive memory resides in cardiac sensory transduction and active memory in the processing of that information by local circuit neurons within the intrathoracic neuronal hierarchy. Control based memory residing in intrathoracic extracardiac and intrinsic cardiac local circuit neuronal interactions, driven as they are by cardiovascular sensory inputs, it is not a passive process. Neurons in chronically decentralized intrathoracic ganglia also display hysteretic memory. In the situation where there is a loss of inputs from central neurons, 100 millisecond long bursts of sensory information (such as arise from aortic mechanosensory neurites during each cardiac cycle) affect local circuit neurons for up to 3 seconds after their discontinuance. That period of time is sufficient for the next five or six cardiac cycles to be generated.
Thus, events in one cardiac cycle influence regional cardiac behavior throughout a few subsequent cardiac cycles via fed-foreword reflexes residing solely within the intrathoracic nervous system. By utilizing such fed-foreward reflexes, the intrinsic cardiac nervous system can be pre-conditioned through the use of SCS or DCA to thereby “override” quench neuronal signals which would place the heart into a diseased state. Such pre-conditioning may take the form of constant SCS or DCA stimulation; and/or long pulses of SCS or DCA stimulation followed by short or long resting periods. In this manner the intrinsic cardiac nervous system is pre-conditioned to resist ischemic neuronal overloading.
Myocardial ischemia. The importance of the processing of sensory information arising from the ischemic myocardium by the intrathoracic cardiac nervous system in the maintenance of adequate cardiac output is only now beginning to be appreciated by those of ordinary skill in the art. One of the major challenges to neurocardiology is understanding the response characteristics of each component of the cardiac neuronal hierarchy to myocardial ischemia so that focused neurocardiological strategies can be devised to stabilize cardiac function in such a state. For that reason, one of the objects of the present invention is to remodel the heart using SCS or DCA stimulation to combat remodeling that occurs within the intrathoracic cardiac nervous system in the presence of focal ventricular ischemia. The selective nature of the responses elicited by each component of the intrathoracic neuronal hierarchy to myocardial ischemia depends on how each population is affected by the content of their altered cardiac sensory inputs. Neuronal interactions in diseased states are relevant given the fact that pharmacological agents proven for use in treating heart failure (i.e., beta-adrenoceptor or angiotensin II receptor blocking agents) target not only cardiomyocytes directly, but also indirectly by altering their inputs from cardiac efferent neurons secondary to altering the intrathoracic neuronal interactions.
Data also indicates that the cardiac neuronal hierarchy becomes obtunded by a variety of interventions, including multiple transmural laser ‘revascularization’ therapy or heart failure. Intrathoracic neuronal function also remodels in the presence of focal ventricular ischemia. Given the fact that certain populations of intrathoracic neurons, when activated, can induce ventricular fibrillation even in the normally perfused heart, therapy directed at the intrinsic cardiac nervous system, whether pharmacological or surgical in nature, or through use of SCS or DCA stimulation are of benefit in managing the ischemic heart and one of its sequellae-ventricular arrhythmias.
Activation of the dorsal columns of the cranial thoracic spinal cord results in a suppression of the activity generated by neurons not only on the target organ, but also in middle cervical and stellate ganglia. It is known that neurons in middle cervical and stellate ganglia are under the constant influence of spinal cord neurons such that following their decentralization the activity generated by many of the latter increased upon removal of such control. Furthermore, removal of spinal cord inputs to the intrathoracic extracardiac nervous system results in enhancement of many intrathoracic extracardiac cardio-cardiac reflexes. It has also been shown that excessive activation of spinal cord neurons suppresses the intrinsic cardiac nervous system, i.e., preconditioning the intrinsic cardiac nervous system.
Heart failure has been considered to be primarily a hemodynamic disorder. Only recently has the importance of neurohumoral mechanisms that act to maintain adequate cardiac output in the presence of heart failure become appreciated, particularly with respect to arrhythmia formation. This recognition and other clinically relevant findings have forced a reappraisal of neuronal mechanisms involved in regulating the ischemic myocardium.
Upper cervical neuronal modulation of upper thoracic cell activity and interactions within and between upper cervical and upper thoracic spinal neurons involved in this processing have been examined. More specifically, these experiments have determined that different populations of neurons within and between segments of the spinal cord exhibit coherence and correlation of activity and may, on occasion, act independently. It has been determined that neurons in the upper cervical (C1-C2) spinal cord are organized to process cardiac sensory information and coordinate the interactions between the C1-C2 and the T3-T4 spinal neurons, to thereby determine autonomic outflow to the intrinsic cardiac nervous system. Coupling of neuronal processing fluctuates within and between two cell populations during increased cardiac sensory stimulation. In the present application it is shown that chemical activation of upper cervical neurons modulates the stimulus locked and long lasting responses of thoracic spinal cord neurons to myocardial algogenic chemical stimuli. It has also been demonstrated that cardiac sensory information arising via thoracic sympathetic afferent activity ascends in the spinal cord via propriospinal neurons to influence neurons in the upper cervical spinal cord. In addition, vagal sensory inputs excite neurons in upper cervical spinal segments. Thus, the upper cervical spinal cord is an area that processes cardiac sensory information transduced by afferent somata in nodose and dorsal root ganglia. Based on these experiments and data, it was determined that, within the hierarchy of control that regulates cardiac function (
Well-established evidence links the autonomic nervous system to life-threatening arrhythmias and cardiovascular mortality. The autonomic imbalance of increased sympathetic activity and reduced vagal activity increases the likelihood for ventricular fibrillation during myocardial ischemia. Clinical studies using various markers of impaired vagal activity supports the experimental evidence that this type of autonomic imbalance increases cardiovascular risk. Disease processes may change the balance between the central and peripheral neurons involved in such regulation. For instance, when the activity generated by cardiac sensory neurons becomes abnormal, cardiac function can be affected profoundly. Therefore, a disturbance of the fine balance within the cardiac neuraxis will produce dramatic changes in cardiac efferent neuronal outflow. Within the hierarchy of central neurons that control the heart, complex sensory processing involves spatial and temporal summation of cardiac sensory inputs to affect central preganglionic autonomic efferent neurons that modulate autonomic efferent postganglionic activity, as well as intrathoracic ganglionic reflexes and the intrinsic cardiac nervous system. Experimental studies, as discussed herein, show that pathological processes can change the integrative behavior of the central cardiac neuraxis. For example, arrhythmias generated by occlusion of the coronary artery are significantly decreased after transection of upper thoracic dorsal roots. This observation indicates that the spinal cord receives and processes information that is generated during an ischemic episode. Furthermore, spinal neurons of the upper thoracic segments are sensitive to changes associated with arrhythmias. These changes can occur when cardiac sensory neurites are activated intensely and for long periods when cardiac tissue becomes damaged during regional ventricular ischemia.
Sympathetic afferents from the heart convey noxious and mechanical, presumably innocuous, information via the dorsal roots primarily in the upper thoracic segments. We herein show that both centrally projecting as well as non-projecting neurons respond to noxious stimuli applied to the heart. We also demonstrate that chemical stimulation of cardiac nociceptors produces either a stimulus-locked or long lasting evoked response of superficial and deep spinal neurons of the upper thoracic spinal cord.
The classical concept of acute cardiac nociception is based on a serial neuronal system that transmits information from cardiac afferents to spinal neurons. The transfer of information is mediated by classical neurotransmitters, such as excitatory amino acids, that lead to membrane potential changes within a time span of milliseconds to seconds. Thus, nociceptive stimulation of cardiac afferents evokes discharge rates of spinal neurons that increase as long as the nociceptors are stimulated. It is generally assumed that impulses of spinal neurons responding to cardiac stimuli constitute a simple renewal process with a very high number of degrees of freedom. Electrophysiological studies show that nociceptive responses of spinal neurons are the basis of mean discharge rates of single neurons. It was further shown that discharge rates are often correlated in a generally linear manner to the intensity of noxious stimulation and antinociception is consequently defined as a reduction in discharge rates of nociceptive neurons.
Under normal, physiological conditions stimuli applied to the heart do not elicit marked changes in cardiac efferent neuronal activity because central neurons suppress excessive cardiac sensory information processing. In the hierarchy of cardiac control, activation of spinal neuronal circuits modulate the intrathoracic cardiac nervous system. Activation of the dorsal columns at the T1-T2 segments significantly reduces the activity generated by the intrinsic cardiac neurons in their basal conditions as well as when activated in the presence of focal ventricular ischemia induced by occluding the left coronary artery. Not only does dorsal column activation modulate the intrinsic cardiac nervous system, but it also modifies the activity of spinal neurons within the T3-T4 segments. In addition, the central nervous system maintains a tonic inhibitory influence over intrathoracic cardiopulmonary-cardiac reflexes. Reflexes mediated through the middle cervical ganglion are increased after decentralization. Thus, disease processes change the balance between the central and peripheral neuronal processing of cardiac sensory information. For instance, when the activity generated by cardiac sensory neurons becomes excessive, e.g., during focal ventricular ischemia, cardiac function can be profoundly affected. Thus, a disturbance of the fine balance within the cardiac neuraxis results in dramatic changes in cardiac efferent neuronal activity. Nests of neural networks in the hierarchy of cardiac control, therefore, appear to interact effectively when an appropriate balance is achieved therein.
Within the hierarchy for cardiac control, neurons of the upper cervical segments modulate information processing in the spinal neurons of the upper thoracic segments. In human studies, spinal cord stimulation of the C1-C2 spinal segments relieves pain symptoms in patients with chronic refractory angina pectoris. Experimental studies disclosed and discussed herein show that spinal cord activation of the upper cervical segments of the spinal cord suppresses the activity of spinal neurons in T3-T4 segments. Furthermore, chemical stimulation with glutamate of cells in the C1-C2 segments also reduces upper thoracic spinal neuronal activity and that chemical stimulation of C1-C2 cells suppresses the activity of lumbosacral spinal neurons. It is especially important to note that this suppression of lumbosacral neuronal activity is sustained even after the spinal cord is transected at the spinomedullary junction. Glutamate was chosen as the stimulant because it only activates cell bodies but not the axons passing through the upper cervical segments.
Little information about descending pathways that originate from C1-C2 segments is available, but anatomical studies provide some evidence for a subpopulation of C1-C2 cells that are involved in propriospinal modulation of spinal sensory neurons. Horseradish peroxidase (HRP) injection into the thalamus of cats labeled cells in the lateral cervical nucleus; however, a subpopulation of cells in the medial part of this nucleus was unlabeled, and axons of these unlabeled cells appear to descend to caudal spinal segments. Others in the art have confirmed those descending projections by injecting HRP in the C8-T5 segments of one monkey and finding labeled cells in the lateral cervical nucleus and in the C1-C2 gray matter. Furthermore, another group skilled in the art has shown in cats that neurons in the medial portion of the lateral cervical nucleus respond to noxious stimuli. In addition, spinal sensory neurons in upper cervical segments receive noxious inputs from large areas of the body and thus, may project to more caudal spinal segments, as well as to the thalamus. Indeed, it has been proposed that the lateral spinal nucleus, which extends down the entire spinal cord, may participate in inhibition of efferent nerve activity in rats. The data as presently disclosed also shows that after selective spinal transections in rats supported the concept that spinal inhibitory effects in sensory neurons utilize upper cervical segments.
Spinal Relay for Vagal Inputs
A very interesting finding is the differential processing of cardiac vagal afferent information in the cervical and thoracic spinal cord. Electrical and chemical stimulation of vagal afferent fibers primarily excites neurons of the C1, C2 segments. It should be pointed out that these cervical cells also receive input that was carried to the thoracic spinal cord via the sympathetic afferents. However, the vagal information elicits larger evoked responses. In contrast to excitation of upper cervical spinal neurons, vagal input from the cardiopulmonary region generally reduces neuronal activity in sensory cells of rats, cats and monkeys in segments below C3; vagal facilitation of responses to noxious inputs are reported only at low stimulus intensities. Antinociceptive effects of vagal stimulation also are found in the tail-flick response in rats, and vagotomy attenuates opioid-mediated and stress-induced analgesia.
Disruption of the C1-C2 neurons with the excitotoxin, ibotenic acid, eliminates the suppressor effects on thoracic spinal neurons with vagal stimulation. Vagal suppression of evoked activity of thoracic spinal neurons resulting from intrapericardial injections of algogenic chemicals is attenuated or eliminated after ibotenic acid was placed on the dorsal surface of the C1-C2 spinal segments.
Neurons in the upper cervical (C1-C2) and upper thoracic (T3-T4) spinal cord process cardiac sensory information to coordinate the interactions within and between these populations of spinal cord neurons and thereby modulate efferent neurons that regulate regional cardiac function. Specifically, neurons in C1-C2 spinal cord process cardiac sensory information to coordinate the interactions within and between C1-C2 and T3-T4 spinal neurons and thereby determine autonomic outflow to the intrinsic cardiac nervous system.
Simultaneous Recordings of Two Neurons
Different populations of neurons within and between segments of the spinal cord exhibit coherence and correlation of activity and may, on occasion, act as independent units. Valuable data was gathered by recording from two cells simultaneously by using two microelectrodes.
A typical example of an upper thoracic cell responding to somatic and noxious chemical stimulation of the heart is shown in
Cell Response to Coronary Artery Occlusion
Chemical stimulation of the heart using algogenic chemical stimulation of cardiac afferents provides a global method for activating cardiac afferents. The effects of coronary artery occlusion on upper cervical and upper thoracic cell activity is included herein because it specifically provides a means of activating nociceptive afferents in regional areas of the heart and clearly demonstrates that SCS or DCA stimulation has a preconditioning or protective effect on the heart thereby dampening neuronal activity of the intrinsic cardiac nervous system. Such an effect leads or lends itself to therapeutic SCS or DCA stimulation of the intrinsic cardiac nervous system to prevent and/or lessen the effects of cardiac pathologies.
Neurochemistry
Experiments were performed to show changes in c-fos expression in the upper thoracic segments in response to activation of cardiac afferents by injecting algogenic chemicals into the pericardial sac (
Effects of Vagal Stimulation on c-fos Expression in C1-C2 Neurons
In order to determine whether input from the vagus would activate C1-C2 neurons, c-fos immunohistochemical studies following vagal electrical stimulation were performed. Three groups have been evaluated: unoperated controls, rats with the vagus nerve crushed for 2 hrs, and rats with the vagus nerve stimulated with the following parameters: (20 Hz, 30 V, 0.2 ms, 5 min on, 5 min off for 1 hr. Abundant c-fos immunoreactive neurons were found in the superficial dorsal horn (marginal zone, substantia gelatinosa), nucleus proprius, central gray region (area X), and ventral horn (
C1 Modulation of Upper Thoracic Cell Activity
Originally, it was assumed that supraspinal pathways are necessary for descending inhibitory effects of visceral afferents on sensory neurons. However, evidence shows that in rats that high cervical neurons can mediate inhibitory effects of cardiopulmonary spinal input in lumbar spinothalamic tract (STT) and dorsal horn (DH) neurons. Thus, it appears that the upper cervical segments play an important role in the hierarchy that controls the efferent outflow to the intrathoracic and intrinsic cardiac nervous system. Based on this knowledge and evidence from previous studies, we conclude that cell bodies located in the gray matter of C1-C2 spinal segments can modulate nociceptive cardiac-evoked activity of spinal neurons in the upper thoracic spinal cord. The effects of glutamate activation of cell bodies in the upper cervical spinal cord on the activity of cells in the T3-T4 spinal cord evoked by injections of bradykinin (BK) into the pericardial sac have been examined. Others have used Glutamate to activate cell bodies in the cervical spinal cord in the art. Glutamate (1M) was absorbed onto filter paper pledgets (2×2 mm) and was placed on the dorsal surface of the C1-C2 segments. Saline control pledgets were applied at the same sites before and after glutamate. Saline did not elicit any responses.
Chemical Stimulation of C1-C2 Cells Before and After Rostral C1 Transection
The evoked activity of one T3 cell to glutamate is shown before (
Neuroanatomy and Immunohistochemistry
Retrograde tracing was used to detect upper cervical neurons with propriospinal projections to the lumbosacral spinal segments in the rat. Termination sites of the upper cervical propriospinal neurons in the gray matter of the upper thoracic segments are shown in
Vagal Modulation of T3-14 Neurons Via the C1-C2 Segments
Electrical Stimulation of the Vagus
Electrical stimulation of the vagal afferents, in general, suppresses the activity of the upper thoracic spinal neurons (
Chemical Disruption of the C1-C2 Neurons
Chemical disruption of the C1-C2 spinal neurons alters the effects of stimulation of the cardiac afferent input on the regulation of information processing in the cervical and thoracic spinal cord. In order to produce chemical disruption of cells, ibotenic acid was chosen because it is an excitotoxin that has been used effectively in previous studies. Ibotenic acid is a structurally rigid glutamate analog that destroys neuronal perikarya, but spares axons and non-neuronal cells. After ibotenic acid is injected into a nucleus or applied to the surface of the spinal cord, the cells in the region are initially excited and then enter a phase of depolarization block.
Vagal Effects after Chemical Disruption of C1-C2 Cell Bodies
At least part of the vagal inhibitory effects of the upper thoracic neurons depend on the C1-C2 relay. Approximately 20 min. after ibotenic acid was placed on the spinal cord, the inhibitory effects to vagal stimulation observed in
Using high frequency, low intensity electrical stimulation of the dorsal aspect of the T1-T2 spinal cord, the modulatory effects on the final common integrator of cardiac function, the intrinsic cardiac nervous system, have been determined. Dorsal cord activation by itself decreases basal intrinsic cardiac neuronal activity by 77%. This suppression of neuronal activity persisted for 30-45 minutes after terminating the dorsal cord stimulation. When LAD occlusion was initiated during dorsal cord activation, neuronal activity remained suppressed. Thus, use of SCS or DCA cord stimulation to precondition and/or remodel the neuronal activity of the intrinsic cardiac nervous system has been shown.
Thus, dorsal cord activation suppresses intrinsic cardiac neuronal activity in both normally perfused and ischemic hearts and dorsal cord activation suppresses the activity of upper thoracic spinothalamic tract neurons evoked by chemical stimulation of cardiac afferents. Dorsal cord activation or SCS can modulate the activity of cells in central nervous system and the intrinsic cardiac nervous system. Dorsal cord activation can be used at either the thoracic or the cervical levels. The cervical segments are particularly interesting, because this is a key region for hierarchical control, and dorsal cord activation of the upper cervical segments has been used to relieve the symptoms in patients with chronic refractory angina pectoris. Dorsal cord activation of the upper cervical segments suppresses the responses of a T3 spinal neuron evoked by algogenic chemical stimulation of the cardiac afferents is shown in
Chemical stimulation of the upper cervical cell bodies suppresses upper thoracic cell responses to nociceptive (chemical) and non-nociceptive (mechanical) input. In contrast, chemical stimulation of the upper thoracic cell bodies excites the upper cervical spinal neurons. Furthermore, the responses to nociceptive and non-nociceptive stimuli are enhanced. Inactivation of the upper cervical cell bodies eliminates the suppression of spontaneous and evoked activity of the upper thoracic neurons. In fact, the nociceptive and non-nociceptive responses are facilitated because elimination of the upper cervical spinal neurons reduces the tonic inhibition that continually impinges on the upper thoracic spinal neurons. Elimination of the upper thoracic cell bodies does not have an appreciable effect on the spontaneous activity and the evoked responses of the upper cervical spinal neurons, because vagal input produces larger responses of the upper cervical neurons than do the inputs that originate from sympathetic afferents.
Vagotomy also changes the modulation of spontaneous activity and nociceptor evoked responses of C1-C2 and T3-T4 spinal neurons. Inactivation of C1-C2 cell bodies eliminates vagal effects of chemical and mechanical stimulation on the activity of the upper thoracic neurons. Vagotomy also eliminates the nociceptive and non-nociceptive responses of the C1-C2 spinal neurons after elimination of input from the T3-T4 spinal neurons. C-fos expression at the upper thoracic segments increases after C1-C2 ablation before and after activation of the vagal afferents, because the cells of the upper cervical segments tonically suppress cell activity in the upper thoracic spinal cord. Perturbations change the correlation characteristics of the pairs of neurons. In addition, the responses of the individual neurons, which are recorded simultaneously, change their response characteristics independently after the interventions are made. The cfos expression does not change in the cervical segments because of the disruption of the cells by ibotenic acids that participate in producing the suppression of thoracic activity. In anterograde tracing studies with PHAL, a clearer picture of the reciprocal innervation between the C1-C2 and T3-T4 segments is seen.
The experiments were designed in order to study the activity and responses of individual cells (192) as well as pairs (96) of cells. Results indicate that coronary artery occlusion evokes responses in the C1-C2 and T3-T4 neurons. That ischemic responses differ from the algogenic chemical responses because chemical stimulation provides a global activation of the afferents; whereas, coronary artery occlusion limits the stimulus to a specific region of the heart. Since vagal input provided the strongest input to the C1-C2 neurons and sympathetic afferents provided the excitatory inputs to T3-T4 neurons, different patterns of activity are demonstrated. Since activation of the vagus excites C1-C2 cells and suppresses the activity of T3-T4 cells, vagotomy reduced the responses of upper cervical neurons but enhances upper thoracic responses to coronary artery occlusion. Coronary artery occlusions will increase the number of cells filled with c-fos expression in both C1-C2 and T3-T4 spinal segments.
The experiments were also designed to study the activity and responses of individual cells (384) as well as pairs (192) of cells to address information processing of the effects of algogenic chemical stimulation and coronary artery occlusion before and after the cells of C1-C2 are disrupted using ibotenic acid. Dorsal cord activation of the T1-T2 or C1-C2 segments suppresses the evoked T3-T4 cell activity to algogenic chemical stimulation and coronary artery occlusion. Since disruption of cells with ibotenic acid reduces or eliminates vagal suppression of the evoked activity of the T3-T4 cells, inhibitory effects of dorsal cord activation are reduced or eliminated, because synaptic activity occurs in the same segments that are stimulated electrically with dorsal cord activation. Disruption of C1-C2 cells with ibotenic acid might reduce the effectiveness of T1-T2 dorsal cord activation on the evoked responses of T3-T4 spinal neurons due to the vasodilator effects of dorsal cord activation being eliminated when the spinal cord was transected at least four to six segments rostral to the site of stimulation. Dorsal cord activation changes the correlation of cell activity in the pairs of cells. These changes are responsible for the suppressed activity of the intrinsic cardiac nerve activity. Dorsal cord activation generates patterns of activity in the spinal neurons that act to stabilize the activity generated by the intrinsic cardiac neurons.
Algogenic chemical stimulation evokes short lasting and long lasting excitatory as well as inhibitory responses of the C1-C2 and T3-T4 neurons. If two neurons recorded simultaneously receive common input from algogenic chemical stimulation of cardiac afferents, they have more synchronous action potentials than statistically expected, and their cross-correlation function correspondingly shows a sharp central peak (i.e., when the mutual delay is at zero). However, the central peak is widened to a variable extent when several neuronal connections are interposed between the locus of common input and the neurons from which the activity is recorded. There are stronger correlations in pairs of neurons when one neuron is in the superficial dorsal horn and the other one is in the deeper dorsal horn. Experiments have shown that latency to the onset of the evoked response of superficial cell to algogenic chemical stimulation of cardiac afferents is shorter than the latency to the onset of the evoked response in a deeper cell. This difference in the latency suggests that the superficial neurons serve as an interneuron between the input from the primary afferents and the activation of the deeper cells. Since vagal input provided the strongest input to the C1-C2 neurons and the sympathetic afferents provided the excitatory inputs to the T3-T4 neurons, different patterns of activity have been demonstrated. Since activation of the vagus excites C1-C2 cells and suppresses the activity of T3-T4 cells, vagotomy will reduce the responses of the upper cervical but enhance upper thoracic responses to nociceptive algogenic chemical stimulation. No effects ocurred using saline controls. With respect to mechanical studies, some of the cells discharge in response to the premature ventricular contraction. Some of the bursts occur early in the compensatory phase, but more commonly the burst is associated with the potentiated contraction. Cells were analyzed individually and as pairs. Vagotomy does not prevent responses of the neurons to chemical stimulation, but most likely modulates some of the mechanical responses. Chemical stimulation increases the number of cells filled with c-fos expression in both the upper cervical and upper thoracic spinal segments. After bilateral vagotomy, a decreased number of cells with cfos, but the number of thoracic spinal cells with c-fos increases because vagal activation of upper cervical neurons suppresses the activity in thoracic neurons. As shown in
Chemical stimulation of the upper cervical cell bodies suppresses upper thoracic cell responses to nociceptive (chemical) and non-nociceptive (mechanical) input. In contrast, chemical stimulation of the upper thoracic cell bodies excites the upper cervical spinal neurons. Furthermore the responses to nociceptive and non-nociceptive stimuli are enhanced. Inactivation of the upper cervical cell bodies eliminates the suppression of spontaneous and evoked activity of the upper thoracic neurons. In fact, the nociceptive and non-nociceptive responses are facilitated because elimination of the upper cervical spinal neurons reduces the tonic inhibition that continually impinges on the upper thoracic spinal neurons. Elimination of the upper thoracic cell bodies does not have much effect on the spontaneous activity and the evoked responses of the upper cervical spinal neurons because vagal input produces larger responses of the upper cervical neurons than do the inputs that originate from sympathetic afferents. Vagotomy changes the modulation of spontaneous activity and nociceptor evoked responses of C1-C2 and T3-T4 spinal neurons. Inactivation of C1-C2 cell bodies eliminates vagal effects of chemical and mechanical stimulation on the activity of the upper thoracic neurons. Vagotomy also eliminates the nociceptive and non-nociceptive responses of the C1-C2 spinal neurons after elimination of input from the T3-T4 spinal neurons. C-fos expression at the upper thoracic segments increases after C1-C2 ablation before and after activation of the vagal afferents, because the cells of the upper cervical segments tonically suppress cell activity in the upper thoracic spinal cord. The perturbations change the correlation characteristics of the pairs of neurons. In addition, the responses of the individual neurons, but recorded simultaneously, change their response characteristics independently after the interventions are made. The cfos expression is not changed in the cervical segments because of the disruption of the cells by ibotenic acids that participate in producing the suppression of thoracic activity. Anterograde tracing studies with PHAL, have shown that a clearer picture of the reciprocal innervation between the C1-C2 and T3-T4 segments is obtained.
Differential remodeling of the peripheral and central cardiac nervous hierarchy and its nerve-cardiac myocyte junction in the presence of a healed myocardial infarction specifically as related to the genesis of ventricular fibrillation occurs. Tests utilize a well-defined canine model of ventricular fibrillation that combines three elements relevant to the genesis of malignant arrhythmias in man: a healed myocardial infarction, acute myocardial ischemia, and physiologically elevated sympathetic efferent neuronal activity have shown that differential remodiling is at least partially responsible for cardiac pathologies. Test also reveal and demonstrate that SCS or DCA stimulation of the intrinsic cardiac nervous system has a preconditioning effect pre-remodeling and a quenching effect post re-modeling. Based on an “exercise and ischemia test”, animals in this model separate into two groups: 1) animals that develop ventricular fibrillation and are thereby classified “susceptible” to fibrillation; and 2) dogs that don't develop sustained ventricular tachycardia/fibrillation and are thus defined as “resistant”. Thus, differential remodeling of the cardiac neuron hierarchy (central and peripheral) for reflex control of the heart occurs in susceptible versus resistant animals.
Autonomic Nervous System and Sudden Death after Myocardial Infarction.
A canine model of lethal ventricular arrhythmias developed in 1978 has been used to elaborate the mechanisms of sudden death after myocardial infarction (MI). In this model, animals with a chronic anterior wall infarction undergo a sub-maximal exercise stress test, culminating in transient total occlusion of the circumflex coronary artery for 2 minutes. During that 2-minute period of transient myocardial ischemia, 40% of the dogs develop ventricular fibrillation (VF); the remaining animals do not generate sustained ventricular arrhythmias. This model produces clinically relevant information by incorporating a healed anterior MI in the setting of elevated sympathetic efferent neuronal tone (induced by exercise), coupled with acute, regional myocardial ischemia distant from the original infarction. This model was developed to duplicate the clinical situation of a patient with multi-vessel coronary artery disease who begins sub-maximal exertion in the convalescent phase of an uncomplicated MI, patients who then develop transient myocardial ischemia. In the dog model, those destined to develop VF display persistent tachycardia in response to transient, acute myocardial ischemia. In contrast, VF resistant animals have been found to possess active vagal reflexes that control heart rate during the ischemic event. Thus, this model produces two distinct groups of animals, based on the occurrence of VF, that have very different characteristics of autonomic control of heart rate.
This model gives rise to the data that non-invasive markers of cardiac vagal reflexes predict risk for sudden death after myocardial infarction. This is shown through baroreflex sensitivity (BRS) data relating a rise in systolic blood pressure to RR interval slowing was prospectively tested prior to exercise and the induction of myocardial ischemia to predict outcomes. BRS was reduced in chronic MI dogs destined to develop VF during exercise and acute, regional myocardial ischemia. Interestingly, BRS was lower before MI in dogs that either died after coronary artery ligation or developed VF within 30 days of acute MI during exercise. Clinical conformation of these results has been published and shows that BRS was lower in patients who subsequently died suddenly after their first myocardial infarction. Results establish that autonomic markers add critical predictive information to the sudden death risk profile after MI. Baroreflex sensitivity measurements provide one index of the cardiac parasympathetic nervous system. Heart rate variability (HRV) quantifies cardiac autonomic interactions by measuring the impact of vagally mediated respiratory sinus arrhythmia via beat-to-beat RR interval variability derived from resting ECG recordings. It is shown herein dogs at high risk for sudden death had low variability measurements, suggesting low tonic vagal input to the heart. Tonic autonomic activity was influenced significantly by MI and recovered only in dogs at low risk for sudden death. In contrast, dogs at high risk for sudden death displayed little recovery during the first 30 days following MI. This persistent blunting of vagal tone was associated with a high risk for VF during exercise and myocardial ischemia. These experiments provide evidence that autonomic control of heart rate remodels during the progression of coronary artery disease.
If depression of vagal efferent neuronal tone to the heart and, as a consequence, cardiovascular reflexes are important for the development of lethal ventricular arrhythmias, does augmentation of cardiac vagal efferent neuronal activity prevent sudden cardiac death in such a model? This issue was addressed using the Schwartz and Stone model of sudden death by electrically stimulating the vagus nerve by means of chronically implanted electrodes. When the vagosympathetic trunk was electrically stimulated during exercise initiated at the onset of coronary artery occlusion, the incidence of VF was prevented in over 80% of high-risk dogs tested. This effect was largely independent of the heart rate reduction associated with vagal activation. Furthermore, augmentation of tonic vagal activity by daily exercise training prevented VF in 100% of the high-risk dogs, either in the presence or absence of acute myocardial infarction. Finally, left stellate ganglionectomy was effective in reducing VF in these high-risk animals. Thus, abnormal autonomic control of the infarcted heart associated with sympathetic efferent neuronal dominance and weak vagal input, results in ventricular electrically instability that increases the risk for sudden cardiac death.
Remodeling of the Cardiac Neuronal Hierarchy after Myocardial Infarction.
What comprises the cardiac neuronal hierarchy and why is it important for the management of cardiac arrhythmias in chronically infracted hearts? Neurons in intrathoracic extracardiac and intrinsic cardiac ganglia have long been thought to act as simple efferent information relay stations involving one synapse, for instance in paravertebral sympathetic ganglia or parasympathetic ganglia on the heart. Recently, this concept has been extended in recognition of the fact that cardiovascular afferent information is also processed within the intrathoracic nervous system, including its component intrinsic to the heart. Neurons in intrathoracic ganglia, including those on the heart, receive constant inputs from spinal cord neurons to modulate their behavior. They also receive sensory inputs from cardiac afferent neurons on an ongoing basis. That is why the activity generated by most intrinsic cardiac neurons increases markedly in the presence increased sensory inputs arising from the ischemic myocardium. Indeed, excessive activation of limited populations of intrinsic cardiac neurons induced cardiac dysrhythmias that lead to ventricular fibrillation. Thus, therapies that act to stabilize heterogeneous evoked activities within cardiac reflex control circuits such, as the SCS or DCA stimulation of the intrinsic cardiac nervous system of the presently claimed and disclosed invention, has obvious clinical importance.
Proper information exchange among the intrathoracic components of the cardiac nervous system act in concert to stabilize the electrical and mechanical behavior of the heart, particularly in the presence of focal ventricular ischemia. Different populations of neurons, distributed spatially within the intrathoracic cardiac nervous system, respond to cardiac perturbations in a coordinate fashion. If neurons in one part of this neuronal axis respond to inputs from a single region of the heart, such as the mechanosensory neurites associated with a right ventricular ventral papillary muscle, then the potential for imbalance within the different populations of neurons regulating various cardiac regions occurs and, thus, its neurons display little coherence of activity. On the other hand, relatively low levels of specific inputs on a spatial scale to the intrathoracic cardiac nervous system results in low basal coherence among its various neuronal components, thereby acting to stabilize cardiac regulation. Alternatively, excessive input to the spatially distributed intrathoracic nervous system destabilizes cardiac electrical behavior, leading to cardiac arrhythmia formation. Intrathoracic extracardiac and intrinsic cardiac neurons receive tonic inputs not only from cardiac and major intrathoracic vascular sensory neurites, but also from spinal cord neurons in the integration of efferent neuronal inputs to the heart.
The infarct matrix is important in determining risk for VF in the Schwartz and Stone model discussed hereinabove. This is illustrated by data showing epicardial conduction mapping across the infarct zone in high and low risk dogs. It has been found that conduction delays are much more profound across the infarct zone in high-risk dogs (>85 millisecond) compared with low risk animals (
The importance of the peripheral cardiac nervous system in the maintenance of normal cardiac output can be appreciated from the presently claimed and disclosed invention. The selective nature of the responses elicited by each component of the intrathoracic neuronal hierarchy to myocardial ischemia depends on how each population of peripheral autonomic neurons is affected, as well as the nature and content of their sensory inputs. That ischemia sensitive cardiac afferent neurons in nodose and dorsal root ganglia influence the behavior of central autonomic neurons which, in turn, modify cardiovascular autonomic efferent preganglionic neurons represents yet another level of this regulatory hierarchy.
Myocardial ischemia. Recent anatomical and functional data indicate the presence of the multiple neuronal subtypes within intrathoracic extracardiac and intrinsic cardiac ganglia. Its intrinsic cardiac component functions as a distributive processor at the level of the target organ. The redundancy of function and non-coupled behavior displayed by neurons within intrathoracic extracardiac and intrinsic cardiac ganglia minimizes the dependency for such control on a single population of peripheral autonomic neurons. In that regard, network interactions occurring at the level of the heart integrate parasympathetic and sympathetic efferent inputs with local afferent feedback to modify cardiac rate and regional contractile force throughout each cardiac cycle. A recent editorial by David Lathrop and Pete Spooner of the NIH highlights the potential clinical relevance of altered processing of information by these populations of neurons such that a lack of coordination of data exchange within the cardiac neuronal axis may lead to the genesis of cardiac arrhythmias.
Interactions Among Neurons in the Cardiac Neuronal Hierarchy.
The different populations of neurons distributed spatially within the intrathoracic cardiac nervous system respond to cardiac perturbations in a complex fashion. For instance, neurons in intrathoracic extracardiac ganglia do not respond to cardiac perturbations in a similar fashion as intrinsic cardiac ones. Consistent coherence of activity generated by differing populations of neurons has been identified among medullary and spinal cord sympathetic efferent preganglionic neurons, as well as among different populations of sympathetic efferent preganglionic neurons. If neurons in one part of the intrathoracic neuronal network respond solely to inputs from a single region of the heart, then the potential for imbalance within the different populations of neurons in various levels of the intrathoracic neuronal hierarchy might occur. A relatively low level of inputs on a spatial scale to populations of intrathoracic cardiac neurons would result in a low basal coherence among its components and stabilize that system. In contrast, excessive input to this spatially distributed nervous system would destabilize it, leading for instance to cardiac arrhythmia formation.
Arterial reflexes can become blunted during the evolution of heart disease. Focal ventricular ischemia is known to alter cardio-cardiac reflexes. Furthermore, ischemia induced liberation of chemicals such as adenosine or hydroxyl radicals within the affected myocardium can suppress ventricular myocyte electrical and contractile behavior. On the other hand, locally released adenosine or hydroxyl radicals can influence the cardiac nervous system via excitation of its afferent neuronal components. Thus, when devising a therapy to modify the outcome of myocardial ischemia one must consider not only altered cardiac myocyte behavior, but autonomic neuronal alterations. A brief summary of some of the issues concerning autonomic neuronal control of the ischemic myocardium is presented below, including its importance in one sequellae of myocardial ischemia-ventricular arrhythmia formation.
Symptomatology. The somata of isolated afferent neurons are sensitive to adenosine. ATP and, to a lesser extent, adenosine influence sensory neurites of dorsal root ganglion neurons. The importance of adenosine in the genesis of cardiac pain became evident when Christer Sylvén and his colleagues administered adenosine into the blood stream of patients with diseased coronary arteries. Indeed, the symptoms induced by adenosine in these patients mimicked those that they experienced during effort. These data are in accord with the fact that dorsal root ganglion purine cardiac afferent neurons play an important role in the genesis of pain and that the ventricular sensory neurites of these neurons become non-responsive to ischemia in the presence of adenosine receptor blockade.
Cardiovascular reflexes secondary to myocardial ischemia. Alterations in heart rate secondary to ventricular ischemia can be due, in part, to altered neural control of cardiac pacemaker cells. Myocardial ischemia can be attended by not only by tachycardia, but also by bradycardia.
Most ventricular sensory neurites associated with nodose ganglion cardiac afferent neurons are sensitive to purinergic agents. Activation of a sufficient population of nodose ganglion afferent neurons by exposing their sensory neurites to purinergic agents can result in the induction of bradycardia via medullary reflexes. Bradycardia can also be induced when sufficient populations of intrinsic cardiac neurons projecting axons to medullary neurons are activated by purinergic agents. In contrast, activation of cardiac sensory neurites associated with dorsal root ganglion neurons with adenosine results in the reflex excitation of sympathetic efferent neurons that innervate the heart. The details of the various reflex responses induced when specific populations of cardiac afferent neurons in nodose as opposed to dorsal root ganglia are modified by local ischemia remain to be fully elucidated. Coordination of autonomic outflows to the heart depends to a large extent upon the sharing of inputs from higher centers concomitant with interactions among neurons in various intrathoracic ganglia. That sharing of cardiac afferent information occurs within the intrathoracic and brainstem/spinal cord feedback loops of
Cardiac arrhythmias. Another sequel of myocardial ischemia is the development of cardiac arrhythmias. As neurons from the level of the insular cortex to the intrinsic cardiac nervous system can be involved in the genesis of cardiac arrhythmias, it is important to recognize that such neurons can induce untoward cardiac electrical events in the presence of myocardial ischemia. For instance, activation of a relatively minor population of intrinsic cardiac neurons in anesthetized canine preparations by exogenous application of an alpha- or beta-adrenoceptor agonist, endothelin I or angiotensin II can induce ventricular dysrhythmias or even fibrillation. DCA and SCS do reduce or ameliorate these effects.
The selective nature of the responses elicited by each component of the cardiac neuronal hierarchy to focal, ventricular ischemia depends on how each population of neurons within this autonomic neuronal hierarchy is affected and that depends in large part on the nature and content of their ventricular sensory inputs. It also depends, in part, on any alteration in ventricular efferent postganglionic axon function secondary to their presence within the ischemic zone.
Cardiac Afferent Neurons.
The chemical milieu of the sensory neurites associated with intrinsic cardiac afferent neurons also change when the blood flow in a coronary artery is compromised. Locally liberated adenosine, ATP, oxygen free radicals and peptides can affect the sensory neurites associated with afferent neuronal somata in nodose, dorsal root or intrathoracic ganglia. Oxygen free radicals also affect the functional integrity of ventricular nerves. The quantities of purinergic agents liberated into the local blood stream and pericardial fluid, increases during ventricular ischemia, as peptides or hydrogen peroxide can affect the activity generated by intrathoracic and central cardiac afferent neurons in an indirect fashion as chemicals accumulated in myocardial tissues and pericardial fluid modify their sensory neurites. When coronary arterial blood flow is restored, during the reperfusion phase various metabolites that accumulate upstream can influence intrinsic cardiac neurons and their sensory neurites supplied by that blood even more.
That ischemia sensitive cardiac afferent neurons in relatively distant (nodose and dorsal root) ganglia versus the somata of cardiac afferent neurons relatively closer to the affected tissue (intrathoracic extracardiac and intrinsic cardiac afferent neurons) influence the behavior of cardiac efferent postganglionic neurons via central and intrathoracic local circuit neurons represents yet another issue of importance within this regulatory hierarchy (
Intrinsic cardiac neurons. Intrinsic cardiac neurons are modified by myocardial ischemia in two fashions: one direct and the other indirect. Transient occlusion of the coronary arterial blood supply to a population of intrinsic cardiac neurons directly affects the function of their somata and/or dendrites. Presumably a lack of energy substrates normally available to them via their local arterial blood supply accounts in part for their altered behavior, as well as the fact that they are bathed by local products of ischemia such as oxygen free radicals and purinergic agents. Each major intrinsic cardiac ganglionated plexus on human or dog hearts is perfused by two or more arterial branches arising from different major coronary arteries. Intrinsic cardiac neurons and cardiomyocytes are affected by hypoxia. Myocardial ischemia of short duration affects not only cardiac myocyte function, but also the capacity of intrinsic cardiac neurons to respond to their sensory inputs. Metabolites accumulating locally when the regional coronary arterial blood supply of intrinsic cardiac neurons is compromised also influence the somata and dendrites of such neurons in a direct manner. Thus, regional ventricular ischemia influences the cardiac neuronal hierarchy in a number of ways, depending on whether the arterial blood supply affected by the arterial lesion directly affects the somata and dendrites of somata therein or indirectly via affecting sensory neurites in the infarct zone.
Data indicate that adaptations occur within the cardiac neuronal hierarchy in the presence of acute, focal ventricular ischemia. The cardiac nervous system remodels during chronic ischemic/infarction to maintain control over regional cardiac dynamics.
Myocardial infarction is induced by ligation of the left anterior descending coronary artery in an open chest procedure during surgical anesthesia. The circumflex coronary artery is instrumented with a pneumatic occluder so that reversible myocardial ischemia can be induced at a later time. After 30 days of recovery, dogs have autonomic tests performed including baroreflex sensitivity (Sleight phenylephrine method) and heart rate variability (time and frequency domain). Then animals run on a treadmill using a protocol in which workload (belt speed and elevation) are increased every 3 minutes. Once heart rate reaches 210 beats per minute the circumflex occluder is inflated for 2 minutes, the first minute the dogs continue to run on the treadmill and the treadmill is stopped for the last minute. Forty percent of the post-MI animals develop ventricular fibrillation (VF) during the 2 minutes of coronary occlusion. The other 60% do not have sustained ventricular arrhythmias. An example of the arrhythmia that susceptible dogs develop is illustrated in
These findings were clinically validated in the multicenter trial called ATRIAMI in which baroreflex sensitivity was found to be an independent risk factor for post-MI sudden cardiac death. Subsequently, the indication that weak vagal reflexes was responsible for susceptible dogs developing VF was tested using electrical stimuli delivered to vagal efferent preganglionic axons to augment cardiac vagal control. Vagal stimulation was started at the time of coronary artery occlusion and continued until the occluder was released. Vagal stimulation prevented VF in over 80% of the susceptible dogs. Even during subsequent exercise testing in which vagal stimulation was coupled with atrial pacing to maintain heart rate at control levels, VF was prevented in about 50% of the animals. Therefore, electrical stimulation of cardiac vagal efferent neurons prevents ventricular electrical instability that develops during exercise and transient myocardial ischemia in susceptible dogs.
One susceptible and one resistant dog were implanted with a spinal cord stimulator and allowed to recover for 7 days. Control exercise and ischemia testing and heart rate variability were studied prior to and during dorsal cord activation (DCA, 50 Hz, 200 μs, 90% motor threshold). The stimulator was activated for 4 hours daily for 4 days; then testing was repeated with the stimulator on.
Both spectral analysis (
It is very difficult to predict how central and intrathoracic autonomic neurons involved in cardiac regulation remodel to sustain cardiac output in the presence of chronic, regional ventricular infarction. Data indicate, however, that the cardiac neuronal hierarchy becomes obtunded by a variety of interventions, including chronic regional ventricular injury.
Information processing within the intrinsic cardiac nervous system and its control of regional cardiac function.
Myocardial ischemia and infarction induce substantial changes in the intrathoracic nerve networks and their reflex control of regional cardiac function. Chronic myocardial infarction/ischemia induces a heterogeneous distribution of efferent projections to cardiac end-effectors. Myocardial infarction/ischemia alters the neurochemical profile of that innervation, with differential increases in neuropeptide content within subsets of neurons contained within the intrinsic cardiac nervous system. The evolution of cardiac pathology is associated with disruptions of the intrinsic cardiac nervous system and its ability to process afferent information and such changes will be more evident in the CMVPG than the RAGP intrinsic cardiac ganglia. Animals that exhibit indices of higher vagal tone (higher baroreflex sensitivity and higher heart rate variability) demonstrate lesser degrees of ischemic/infarct-induced neural remodeling.
The functional connectivity of intrinsic cardiac and intrathoracic extracardiac neurons in normal and acutely ischemic hearts.
Little direct functional interconnectivity exists among intrinsic cardiac neurons and their intrathoracic extracardiac counterparts. Independent function as such indicates that little reliance on one such population normally occurs when regulating regional cardiac function; i.e., dysfunction of one population occurs without a major loss of regional cardiac control. Significant alterations in the cardiac milieu, such as occurs during acute, focal ventricular ischemia, induces greater coherence of activity among populations of intrathoracic and intrathoracic extracardiac neurons.
Chronic myocardial ischemia induces a heterogeneous distribution of efferent projections to cardiac end-effectors. We anticipate that this heterogeneous distribution of sympathetic fibers to the left ventricle results in similar heterogeneous release of catecholamines and neuropeptides into the interstitial space during stimulation of the efferent nerves. Finally, animals that exhibit indices of higher vagal tone (higher baroreflex sensitivity and higher heart rate variability) demonstrate lesser degrees of ischemic/infarct-induced remodeling of the efferent outflow of the left ventricle.
Activation of the dorsal columns of the cranial thoracic spinal cord suppresses the activity generated by neurons not only on the target organ, but also in middle cervical and stellate ganglia. It is known that neurons in these ganglia are under the constant influence of spinal cord neurons such that following their decentralization their activity increases (i.e., spinal cord neurons exert tonic suppression of their function). Removal of spinal cord inputs to the intrathoracic nervous system enhances many intrathoracic cardio-cardiac reflexes is tied to the principle and thus excessive activation of spinal cord neurons suppress the intrinsic cardiac nervous system.
Heterogeneous alterations within the intrinsic cardiac ganglia or at the end-terminus of the autonomic innervation to the ischemic myocardium are major contributors to the increased incidence of sudden cardiac death in patients with coronary artery disease. The increased incidence of sudden death often result from lack of protection of the myocytes and instability of the cardiac electrical system. Chronic DCA ameliorate ischemia-induced remodeling within the intrinsic cardiac nervous and thereby reduces the heterogeneous neural substrate that predisposes the susceptible animals to ventricular arrhythmias and sudden cardiac death.
Heart failure has traditionally been considered to be primarily a hemodynamic disorder. The importance of neurohumoral mechanisms that act to maintain adequate cardiac output in the presence of ventricular ischemia is apparent. This recognition has forced a reappraisal of neuronal mechanisms involved in regulating the ischemic myocardium leading to the development of the presently claimed and disclosed invention.
Spinal cord-peripheral neural interactions and modulation of peripheral nerve function in the ischemic heart. Dorsal column activation stabilizes the intrinsic cardiac nervous system in acute myocardial ischemia experiments were conducted. The purpose of these experiments was to determine if dorsal column activation (DCA) induces long-term effects on the intrinsic nervous system, the final common integrator of cardiac function, particularly in the presence of myocardial ischemia. Methods: Activity generated by right atrial neurons was recorded in 10 anesthetized dogs during basal states, and during 15 min occlusions of the LAD coronary artery, with and without background DCA. For DCA, dorsal T1-T4 spinal segments were stimulated for 17 min. at 90% of motor threshold (50 Hz; 0.2 ms duration). For combined effects, the coronary occlusion commenced 1 min into DCA. Results: Ischemia-induced excitatory effects on the intrinsic cardiac nervous system were suppressed (−76%) during DCA and for approximately 20 min after DCA termination. Conclusions: DCA suppresses basal activity within the intrinsic cardiac nervous system and prevents the ischemia-induced activation of these peripheral neural networks. This stabilization of intrinsic cardiac neuronal function, induced by higher elements of the neural hierarchy for cardiac control, is maintained for prolonged periods post-stimulation and is reflective of the neural memory of these processes. These long-term effects may partially explain the prolonged effects patients with angina experience not only during DCA, but also for a time thereafter.
Coronary artery occlusion induces differential catecholamine release in the normal and ischemic myocardium. (
Dorsal column activation stabilizes peripheral adrenergic function in acute myocardial ischemia. The purpose of this study was to determine whether DCA modulates NE and EPI release into the canine ISF in both normal hearts and those exposed to transient myocardial ischemia. Methods: In anesthetized dogs, left ventricular ISF samples were collected by microdialysis during electrical stimulation (50 Hz, 0.2 ms) of the dorsal T1-T4 segments of the spinal cord at an intensity of 90% of motor threshold with and without concurrent 15 min occlusions of the circumflex coronary artery. Results: ISF EPI doubled by 10 min and tripled by 20 min of DCA (239 to 935 pg/ml, respectively). ISF EPI remained twice baseline 20 min post-DCA. DCA increased left ventricular NE by 43% (890 to 1273 pg/ml); ISF NE returned to baseline values 20 min post-DCA. Heart rate and left ventricular inotropic function were not affected by DCA. When 15 min CAO was instituted during DCA (
Spinal cord-peripheral neuronal interactions modify myocardial electrical stability. Dorsal column activation reduces ventricular fibrillation accompanying acute myocardial ischemia.
The purpose of this study was to determine if the stabilization of peripheral neural function exerted by DCA reduces the potential for ventricular fibrillation induction in acute myocardial ischemia. Methods: Under anesthesia, atrial and ventricular electrograms were recorded during basal states, and during 15 min occlusions of the proximal circumflex artery with and without pre-existing DCA. DCA involved electrical stimulation (50 Hz, 0.2 ms) of the dorsal T1-T4 segments of the spinal cord at an intensity of 90% of motor threshold for 36 min, with coronary artery occlusion commencing 15 min into DCA (
Dorsal column activation stabilizes ischemic myocardial electrical dysfunction. The purpose of this study was to determine whether DCA modulates electrical imbalance within the chronically ischemic ventricle. Methods: An ameroid constrictor was implanted around the left circumflex coronary artery to gradually occlude that vessel. Four weeks later, under general anesthesia multiple ventricular unipolar electrograms were recorded in the normal and ischemic left ventricle during basal states and when ANG II (40 μg/min; 1 minute) was administered to right atrial neurons before and after DCA. Results: ANG II increased the area and magnitude of regional ST segment changes in the ischemic ventricle. ANG II induced minimal changes in the electrical behavior of the normal myocardium. DCA (50 Hz, 0.2 ms, 0.32 mA for 15 min) modified ischemic indices, even suppressing regional ventricular ST segment abnormalities previously induced by ANG II. Conclusions: DCA suppresses ischemia induced ventricular electrical disturbances. This may occur, in part, via stabilizing intrathoracic adrenergic neurons that modulate the ischemic ventricle.
Processing of cardiac sensory information by neurons in the upper thoracic (T3-T4) spinal cord. Chemical activation of cardiac receptors differentially affect activity of superficial and deeper spinal neurons in rats. The purpose of this study was to evaluate responses of superficial (depth <300 μm) versus deeper thoracic spinal neurons to chemical stimulation of cardiac afferent neurons and to determine if descending central neuronal inputs modulate these effects. Methods: Extracellular potentials of single T3-T4 neurons were recorded in pentobarbital anesthetized, paralyzed and ventilated male rats. A catheter was placed in the pericardial sac to administer 0.2 ml of an algogenic chemical mixture that contained adenosine (10−3 M), bradykinin, histamine, serotonin, and prostaglandin E2 (10−5 M). Results: Intrapericardial chemicals elicited responses in 27% of the superficial neurons and in 47% of the deeper neurons. All superficial neurons that responded to cardiac afferents were excited. Of the deeper neurons, approximately 80% were excited, 15% were inhibited and 5% showed excitation-inhibition. Spontaneous activity of superficial neurons with short-lasting excitatory responses was significantly lower than that of deeper neurons (P<0.05). After cervical spinal transection, spontaneous activity generated by superficial and deeper neurons increased significantly, as did responses to chemical activation of cardiac afferents neurons. Conclusions: Chemical stimulation of cardiac afferent neurons excites superficial T3-T4 spinal neurons; deeper neurons exhibit multiple patterns of responses. These data further indicate that thoracic spinal neurons that process cardiac nociceptive information are tonically inhibited by higher center neurons.
Descending modulation of thoracic cardiac nociceptive transmission by upper cervical spinal neurons. The purpose of this study was to examine effects of stimulating upper cervical spinal neurons on spontaneous and evoked activity of thoracic spinal sensory neurons that responded to noxious cardiac stimuli. Methods: Extracellular potentials of single T3 neurons were recorded in pentobarbital anesthetized male rats. A catheter was placed in the pericardial sac to administer bradykinin (10−5M, 0.2 ml, 1 min) as a noxious cardiac stimulus and saline as control. A glutamate pledget (1 M, 1-3 min) was placed on the surface of C1-C2 segments to chemically activate upper cervical spinal neurons. Results: In 77% of the T3 neurons tested, glutamate at C1-C2 inhibited spontaneous activity and/or excitatory responses to intrapericardial bradykinin. After transection at the rostral C1 spinal cord, excitatory amino acid (glutamate) excitation of C1-C2 neurons still reduced the spontaneous activity of T3 neurons, as well as excitatory inputs from cardiac sensory neurons. Conclusions: Chemical activation of C1-C2 spinal neurons evokes a descending inhibition in thoracic spinal cord cardiac neurons during basal states as well as in the presence of noxious cardiac stimuli. Furthermore, modulation of cranial thoracic neurons by upper cervical spinal neurons does not require supraspinal connectivity.
Interdependence of cardiac sensory information processing by neurons in the upper thoracic (T3-T4) spinal cord. The purpose of these studies was to evaluate the coordination of activity among upper thoracic neurons that process cardiac sensory inputs. To date, we have evaluated the correlation of spontaneous and evoked activity of 15 pairs of T3 spinal neurons. Included is
Mechanical activation of spinal neurons using programmed ventricular arrhythmias. Data demonstrating the feasibility of recording the responses of spinal neurons to premature ventricular contractions and compensatory beats. To generate these events, an electrical stimulus was applied through a pair of stainless steel electrodes that were inserted in the free wall of the left ventricle.
Without the present specification, one of ordinary skill in the art would not have appreciated or known to use SCS or DCA stimulation as a means to (1) electrically influence the intrinsic cardiac nervous system to protect cardiac myocytes from initial ischemic damage or from being further damaged during subsequent ischemic episodes; and (2) preserve the electrical stability of the intrinsic cardiac nervous system and the heart itself prior, during, or post an ischemic episode. As such, the presently claimed and disclosed invention would be non-obvious in light of the prior art showing the use of SCS stimulation for the treatment of angina. In fact, those of ordinary skill in the art that SCS alleviated widely and traditionally believe angina pain by either changing blood flow within ischemic or non-ischemic myocardium or modifying left ventricular (LV) pressure-volume dynamics. As the following experiments show, however, SCS does not alter these blood parameters—rather, SCS influences and effects the modulation of neuronal activity within the intrinsic cardiac nervous system. Thus, use of SCS to treat, modify, protect, and influence neuronal activity within the intrinsic cardiac nervous system is a novel and non-obvious approach to the pre- and post-treatment of an ischemic heart.
In the first series of experiments, it is shown that (1) SCS modifies the capacity of the intrinsic cardiac nervous system to generate electrical activity; (2) SCS suppresses the excitatory effects that local myocardial ischemia exerts on the neurons of the intrinsic cardiac nervous system; and (3) SCS does not change heart indices such as blood pressure. Thus, the underlying principle that SCS can and does stimulate and provoke an effect in the intrinsic cardiac nervous system is shown and demonstrated.
Electrical stimulation of the dorsal aspect of the upper thoracic spinal cord is used increasingly to treat patients with severe angina pectoris refractory to conventional therapeutic strategies. Clinical studies show that spinal cord stimulation (SCS) is a safe adjunct therapy for cardiac patients, producing anti-anginal as well as anti-ischemic effects. The effects of SCS on the final common integrator of cardiac function, the intrinsic cardiac nervous system, was studied during basal states as well as during transient (2 min) myocardial ischemia. Activity generated by intrinsic cardiac neurons was recorded in 9 anesthetized dogs in the absence and presence of myocardial ischemia before, during and after stimulating the dorsal T1-T2 segments of the spinal cord at 66 and 90% of motor threshold using epidural bipolar electrodes (50 Hz; 0.2 ms; parameters within the therapeutic range used in humans). The SCS suppressed activity generated by intrinsic cardiac neurons. No concomitant change in monitored cardiovascular indices was detected. Neuronal activity increased during transient ventricular ischemia (46%), as well as during the early reperfusion period (68% compared to control). Despite that, activity was suppressed during both states by SCS.
Thus, SCS modifies the capacity of intrinsic cardiac neurons to generate activity. SCS also acts to suppress the excitatory effects that local myocardial ischemia exerts on such neurons. Since no significant changes in monitored cardiovascular indices were observed during SCS, it is concluded that modulation of the intrinsic cardiac nervous system might contribute to the therapeutic effects of SCS in patients with angina pectoris.
Patients who suffer from severe angina pectoris following coronary artery revascularization or whose clinical status render them inappropriate candidates for such a procedure can obtain relief from their angina by spinal cord stimulation (SCS) (Jessurun et al., 1997; Schoebel et al., 1997). High frequency, low intensity electrical stimuli delivered to the dorsal aspect of the T1-T2 thoracic spinal cord suppresses the pain associated with myocardial ischemia without affecting awareness of the symptoms from a possible myocardial infarction (Anderson et al., 1994; Eliasson et al., 1996; Hautvast et al., 1998; Sanderson et al., 1994). Application of SCS does not appear to induce any adverse effects in patients experiencing transient ischemia of the myocardium (Sanderson et al., 1992), patients retain their capacity to sense angina during increased workload (Mannheimer et al., 1993).
The effects of SCS have been attributed to improved myocardial perfusion and/or alterations in the oxygen demand and supply ratio as reflected in a reduction in stress-induced alterations in the ST segment of the ECG (Sanderson et al., 1992). Spinal cord stimulation also improves myocardial lactate metabolism (Mannheimer et al., 1993). Spinal cord stimulation has recently been suggested as an adjunct to coronary artery bypass surgery in high-risk patients (Mannheimer et al., 1998).
Spinal cord stimulation has been shown to influence information processing within the central nervous system (Chandler et al., 1193; Yakhnitsa et al., 1999). This treatment modality has also been demonstrated to influence peripheral blood flow (Augustinsson et al., 1995; Augustinsson et al., 1997; Linderoth et al., 1991; Linderoth et al., 1994; Croom et al., 1997). In order to understand the mechanisms underlying SCS in cardiac control, we studied the effects of SCS upon the intrinsic cardiac nervous system. Intrinsic cardiac neurons receive constant inputs from spinal cord neurons to regulate regional cardiac function on a beat-to-beat basis (NAMES). Transient regional ventricular ischemia markedly increases the activity generated by intrinsic cardiac neurons (Huang et al., 1993). Furthermore, excessive activation of limited populations of intrinsic cardiac neurons induces cardiac dysrhythmias, even in normally perfused hearts (Huang et al., 1994).
The experiments and data detailed hereinbelow show that SCS, applied with clinically employed electrical stimulation parameters, modifies the activity generated by intrinsic cardiac neurons in situ. SCS does not change cardiac dynamics. Effects of SCS on intrinsic cardiac neural activity were characterized during coronary arterial occlusion as well as during the subsequent reperfusion period and it was determined that SCS modifies intrinsic cardiac neuronal function in the presence of myocardial ischemia. These experiments show that SCS influences the behavior of intrinsic cardiac neurons markedly, changes that are involved in the clinically observed effects of SCS during acute myocardial ischemia.
Animal Preparation
Experiments performed in the present study were approved by the Institutional Animal Care and Use Committee of the OUHSC and followed the guidelines outlined by the International Association for the Study of Pain and in the NIH Guide for the Care and Use of Laboratory Animals (National Academy Press, Washington, D.C., 1996). Nine adult male dogs of mixed breed weighing between 15 and 25 kg were used. Animals were kept under standard laboratory conditions in a light-cycled environment (12 h/12 h) with free access to water at all times and to food at regular intervals. For the duration of the surgery, dogs were first anesthetized with sodium thiopental (20 mg/kg, i.v.) and maintained with sodium thiopental administered in boluses (5 mg/kg i.v.) to effect every 5-10 min. Animals were intubated and then artificially ventilated using a Harvard respirator (Palm Springs, Calif.). After the surgical preparation was completed, anesthesia was changed to alpha chloralose. An initial bolus dose of alpha chloralose (75 mg/kg, i.v.) was administered, with repeat doses (20 mg/kg) given as required during the remainder of the experiment. The level of anesthesia was checked throughout each experiment by observing pupil reaction, monitoring jaw tension and squeezing a hindpaw to determine if blood pressure and heart rate changed. This anesthetic regimen has been demonstrated to produce adequate anesthesia without suppressing autonomic neural responses (Gagliardi et al., 1988). Electrodes inserted into the forelimbs and the left hind limb were connected to an Astro-Med, Inc. (West Warwick, R.I.) model MT 9500 eight channel rectilinear recorder to monitor a modified Lead II electrocardiogram.
Implantation of Spinal Cord Stimulation Electrodes
After induction of anesthesia, animals were placed in the prone position and the epidural space of the mid-thoracic spinal column was penetrated percutaneously with a Touhy needle using A-P fluoroscopy and loss-of-resistance technique, as is routinely done in the clinic. A four-pole catheter (Medtronic QUAD Plus Model 3888; Medtronic Inc., Minneapolis, Minn.) was introduced through the cannula and its tip was advanced to the T1 level of the spinal column and placed slightly to the left of the midline (Augustinsson et al., 1995). The two poles of this stimulating lead chosen for subsequent use (inter-electrode distance of 1.5 cm) were placed at the level of the T1 and T4 vertebrae. Final placement was aided by delivering electrical current to induce motor responses using the rostral or caudal poles as cathodes, respectively. Rostral stimulation just above motor threshold resulted in proximal forepaw and/or shoulder muscle contractions while caudal electrode stimulation induced contractions in the lower trunk. Once the appropriate electrode positions were obtained, the lead was fixed to the intraspinous ligaments with a suture surrounding a Silicone protective sleeve. Extension wires were tunneled subcutaneously to the ventral surface of the animal where they were connected to a stimulator. Motor responses were rechecked after the animal had been turned to the supine position to make sure the electrodes had not moved during this maneuver and to establish the appropriate stimulus intensities for the subsequent SCS.
Cardiac Instrumentation
After placing the animal on its back, a bilateral thoracotomy was made in the fifth intercostal space to expose the heart. The subclavian ansae on both sides of the thorax were exposed and silk ligatures were placed around them so that each could be easily sectioned later in the experiments to decentralize the intrinsic cardiac nervous system. The ventral pericardium was incised and retracted laterally to expose the heart and the ventral right atrial deposit of fat containing the ventral component of the right atrial ganglionated plexus. Neurons in this ganglionated plexus are representative of those found in the various intrinsic cardiac ganglionated plexuses (Gagliardi et al., 1988).
Left atrial chamber pressure was measured via a PE-50 catheter inserted directly into the left atrial chamber via its appendage. Left ventricular chamber pressure was monitored via a Cordis (Miami, Fla.) #6 French pigtail catheter, which was inserted into that chamber through a femoral artery. Systemic arterial pressure was measured using a Cordis #7 French catheter placed in the descending aorta via the other femoral artery. These catheters were attached to Bentley (Irvine, Calif.) Trantec model 800 transducers.
Neuronal Recording
Activity generated by ventral right atrial neurons was recorded in situ, as has been done in previous studies (Gagliardi et al., 1988). To minimize epicardial motion during each cardiac beat, a circular ring of stiff wire was placed gently on the fatty epicardial tissue overlying the ventral surface of the right atrium containing the right atrial ganglionated plexus. A tungsten microelectrode (30-40 μm diameter and exposed tip of 1 μm; impedance of 9-11 M at 1000 Hz), mounted on a micromanipulator, was lowered into this fat using a microdrive. Exploration was done by driving the electrode tip through this tissue beginning at the surface of this fat, penetrating to regions adjacent to cardiac musculature. Proximity to the atrial musculature was indicated by increases in the amplitude of the ECG artifact. The indifferent electrode was attached to mediastinal connective tissue adjacent to the heart. Signals recorded via the electrode were led to a CWE BMA-831 differential preamplifier with a high impedance head stage (bandpass filters set at 300 Hz and 10 kHz), and were processed by a signal conditioner (bandpass 100 Hz-2 kHz). Signals were amplified further via a Princeton Applied Research (Princeton, N.J.) battery driven amplifier (300 Hz-2 kHz) and were displayed on an Astro-Med, Inc. (West Warwick, R.I.) MT 9500 8 channel rectilinear recorder along with the cardiovascular variables described above. Data were stored via a Vetter (Rebesburg, Pa.) M3000A digital tape system for later analysis. Action potentials generated by neurons in one site of a right atrial ganglionated plexus were recorded using extracellular recording electrodes, individual units being identified by their amplitudes and configurations. As established previously (Armour et al., 1990), extracellular action potentials so generated are derived from somata and/or dendrites rather than axons of passage. Amplitudes of identified action potentials varied by less than 25 μV over several minutes. Each potential retained the same configuration over time. Action potentials recorded in a given locus with the same configuration and amplitude (±25 μV) were considered to be generated by a single unit.
Five different protocols were employed in each animal (cf.
Protocol A—Spinal Cord Stimulation
The parameters used to electrically stimulate the thoracic spinal cord were similar to those used clinically. Stimuli were delivered to the dorsal aspect of the thoracic spinal cord via a Grass model S48 stimulator connected to the quadripolar electrode via a stimulus isolation unit (Grass model CCU1) via, a constant current unit (Grass SIU1). With the animal placed in the supine position for all subsequent experimentation, the current intensity used to evoke detectable skeletal muscle motor responses was determined as the motor threshold (MT). Stimuli (50 Hz and 0.2 ms duration) were delivered at two intensities (66 and 90% of MT). An intensity of 66% of MT has been shown to recruit low threshold, rapidly conducting axons (A-beta), whereas higher, intensity stimuli (90%) activate fast A-delta fibers as well as the other axonal populations (Croom et al., 1997; Linderoth et al., 1991). The current measured at MT varied among animals likely because of the varied anatomy of the thoracic spinal space among animals. The stimulus intensity was found to vary between 30 and 50 μA when current was set to 66% of MT. When the stimulus current was 90% of MT, it varied between 80 and 210 μA among different animals. The MT was rechecked periodically and remained stable over time in individual animals. With respect to protocol A, cardiac indices and intrinsic cardiac neural activity were monitored immediately before, during and for 30-45 s after 4 ruin of SCS at 90% of MT (
Protocol B—Regional Ventricular Ischemia
A silk (3-0) ligature was placed around the left anterior descending coronary artery and another around the circumflex coronary artery, approximately 1 cm from their respective origins. Each ligature was led through a short segment of polyethylene tubing in order to occlude these arteries later in the experiments while leaving the arterial blood supply (right coronary and sino-atrial arteries) patent to the ventral right atrial neurons that were being investigated. Fur protocol B, cardiac indices and neuronal activity were monitored before, during and immediately after occluding the two coronary arteries concurrently for 2 min (
Protocols C, D and E in which SCS and Regional Ventricular Ischemia were Combined
The effects of 2 min of myocardial ischemia on intrinsic cardiac neuronal activity and regional cardiac indices were studied in the presence of SCS (at 90% of MT for 4 min) applied at different times during the myocardial ischemia. Protocol C: The spinal cord was stimulated for 4 min and the 2 min of coronary artery occlusion began 1 min after the onset of the SCS (in the middle of the SCS; Foreman
After all of the protocols described above were completed, the right and left subclavian ansae were sectioned in five of the dogs, thereby eliminating spinal cord afferent and efferent communications with neurons in intrathoracic ganglia. After this maneuver, the five SCS and transient coronary occlusion protocols described above were repeated.
Individual action potentials, which maintained their configurations over time, were analyzed. Activity generated by the somata and/or dendrites of neurons within the right atrial ganglionated plexus was averaged during successive 30-s periods before, during and after each intervention. At the same time, heart rate, left ventricular wall (intramyocardial) and chamber systolic pressures were measured, as was aortic pressure. Neuronal activity and cardiovascular indices recorded immediately before each intervention and during the steady state response to an intervention were averaged and presented as means±S.E.M. Fluctuations in the amplitude of action potentials generated by a unit varied by less than 50 μV over several minutes, action potentials retaining the same configurations over time. Thus, action potentials recorded in a given locus with the same configuration and amplitude (±50 μV) were considered to be generated by a single unit. Action potentials with signal-to-noise ratios greater than 3:1 were analyzed. The threshold for neuronal activity changes was taken as a change of more than 20% from baseline values. Neuronal activity responses elicited by each intervention were evaluated by comparing activity generated immediately before each intervention with data obtained at the point of maximum change during the intervention. Data were expressed as means±S.E.M. Oneway ANOVA and paired t-test with Bonferroni correction for multiple tests was used for statistical analysis. A significance value of P<0.05 was used for these determinations.
Identification of Active Sites
Action potentials were identified in 1-3 loci within the ventral right atrial ganglionated plexus of each animal. Based on the different amplitudes and configurations of the recorded action potentials within these loci, ongoing activity was generated by an average of 5.1±0.9 (range 3-9) neurons. Identified neurons generated, on average, 496±112 impulses per minute (ipm) during control conditions throughout the duration of these experiments. Multiple neurons at each identified active site generated action potentials that were altered in a similar fashion by each of the different interventions tested.
(Protocol A) Effects of Spinal Cord Stimulation
Only the effects of SCS employed at 90% of MT are presented herein since 66% MT elicited minimal changes in the activity generated by the intrinsic cardiac neurons. The average activity generated by identified right atrial neurons in all animals (n=9) fell from 496±112 to 1501±71 ipm (P<0.01) during SCS at 90% MT (
(Protocol B) Effects of Transient Myodcardial Ischemia
When ventricular ischemia was induced by occluding both the left anterior descending and circumflex coronary arteries for 2 min in neurally intact preparations, the activity generated by right atrial neurons increased in each animal (
SCS Modulated Responses to Transient Myocardial Ischemia
Neuronal activity was not enhanced by coronary artery occlusion induced in the presence of SCS, irrespective of whether SCS was applied during (
(Protocol C) Occlusion in the Middle of Stimulation
When the 2-min period of myocardial ischemia occurred in the middle of the SCS (1 min after SCS began), the neurosuppressor effects of SCS persisted during the ischemic period (
(Protocol D) Occlusion Overlapped by Stimulation
During this protocol (FIG. 34(D)), the activity generated by intrinsic cardiac neuronal activity was enhanced by 42% (388±155 to 555±211 ipm; P<0.01) during the initial coronary artery occlusion period. When SCS was applied 1 min after the occlusion began, neuronal activity was suppressed by 46% (activity of 211±134 ipm) even though the myocardial ischemia persisted (Fig. CC-C). In this protocol, neuronal activity remained suppressed during the reperfusion period (227±134 ipm) while the SCS persisted neuronal activity returned to control values only after SCS ceased (394±142 ipm).
(Protocol E) Occlusion Followed by Stimulation
In this protocol (FIG. 34(E)), coronary artery occlusion alone enhanced neuronal activity (403±150 to 701±315 ipm; P<0.01). When SCS was started immediately following termination of 2 min of coronary occlusion (that is during the early reperfusion period), neuronal activity fell to 173±295 ipm (P<0.01 compared to the ischemia period). Neuronal activity remained suppressed throughout this stimulation period, being 244±98 ipm (P<0.01 compared to control values) after 4 min of SCS. This is in distinct contrast to the finding that neuronal activity remained elevated (−70% of control values) during the early reperfusion period immediately after SCS was terminated (
Acute Decentralization
After all of the experimental protocols described above were completed, the spinal cord was stimulated in 5 animals at 90% of MT before and after sectioning the right and left ventral and dorsal subclavian ansae. After surgically disconnecting intrinsic cardiac neurons from the spinal cord neurons, ongoing neuronal activity decreased from 378±34 to 162±72 ipm (P<0.01). SCS did not modify the activity generated by identified intrinsic cardiac neurons thereafter (162±72 vs. 147±61 ipm); nor did SCS affect recorded cardiac indices.
Results of the present experiments demonstrate that the activity generated by intrinsic cardiac neurons is modulated when the dorsal aspect of the thoracic spinal is stimulated electrically. That suppression of the ongoing activity generated by intrinsic cardiac neurons induced by SCS persisted for at least 30 s following termination of 4 min of SCS shows that the effects of this intervention last beyond the stimulation period. Interruption of afferent and efferent nerves traveling in the subclavian ansae eliminated the suppressor effects that SCS exerted on intrinsic cardiac neurons. These data show that the influence of spinal cord neurons on the intrinsic cardiac nervous system occur primarily via axons coursing in the intrathoracic sympathetic nervous system.
Based on results obtained when SCS was applied to the lumbosacral spinal cord, both sympathetic afferent and efferent fibers contribute to the suppression of intrinsic cardiac activity so identified. Four minutes of SCS at 66% of MT was much less effective in suppressing neuronal activity than when the spinal cord was stimulated at 90% of MT. Spinal cord stimulation at 90% MT antidromically activates sensory afferent fibers that release calcitonin gene-related peptide (CGRP) from their afferent terminals, an action that may be dependent on the presence of nitric oxide; such local release of CGRP from sensory afferent nerve terminals produces vasodilation of the rat hind paw (Croom et al., 1997). It is known that endorphins are released into the coronary circulation of humans during SCS (Eliasson et al., 1991). The release of neuropeptides by antidromic activation of sensory neurites (Croom et al., 1997) acts to change the activity generated by intrinsic cardiac neurons (Armour et al., 1993).
Activation of sympathetic efferent preganglionic axons suppresses many intrathoracic reflexes that are involved in cardiac regulation (Armour et al., 1985) as well as the activity generated by populations of neurons within intrathoracic extracardiac (Armour, 1986) and intrinsic cardiac (NAMES) ganglia, thereby reducing the capacity of intrathoracic sympathetic efferent neurons to influence cardiodynamics (Butler et al., 1988). This effect may in part be due to activating inhibitory synapses within intrathoracic ganglia, including those on the heart such as occurs when intracranial pressure raises (Murphy et al., 1995). Such suppression of neuronal activity has been demonstrated in sympathetic efferent neurons controlling the peripheral vasculature as well (Linderoth et al., 1991; Linderoth, Fedorcsak et al., 1991).
As has been shown previously (Huang et al., 1993), the activity generated by right atrial neurons increased in the presence of regional ventricular ischemia (
The data obtained in this study are in accord with clinical findings indicating that improvement of cardiac function and symptoms can occur when SCS is applied to patients with angina pectoris (deJongste et al., 1994). Since modification of the intrinsic cardiac nervous system can lead to alterations in ventricular regional flow (Kingma et al., 1994), perhaps some of the responses elicited by SCS involved subtle changes in the redistribution of coronary artery blood flow given that no detectable changes in cardiodynamics were identified with the methods used in these experiments. Thus, the effects that SCS induces in a clinical setting reside, in part, in the capacity of such therapy to stabilize this final common regulator, even in the presence of ventricular ischemia. Since the intrinsic cardiac nervous system receives inputs arising from cardiac sensory nenrites as well as from central neurons (Murphy et al., 1995), SCS may exert multiple effects on this local neuronal circuitry. Heterogeneous activation of intrinsic cardiac neurons destabilizes cardiac neuronal regulation that, in turn, leads to the genesis of ventricular tachydysrhythmias (Huang et al., 1994). Data obtained in the present experiments indicated that SCS reduces the excitability of intrinsic cardiac neurons, even in the presence of ventricular ischemia and, as such, may help to stabilize cardiac function.
In summary, electrical stimulation of the thoracic spinal cord influences the function of the final common neuronal regulator of cardiac function, the intrinsic cardiac nervous system, even in the presence of myocardial ischemic challenge. Thus, SCS acts in part to protect the heart from some of the deleterious consequences resulting from myocardial ischemia via altering the function of the intrinsic cardiac nervous system.
As stated previously, electrical stimulation of the dorsal aspect of the upper thoracic spinal cord is used to treat patients with angina pectoris refractory to conventional therapeutic strategies. The purpose of the following described experiments was to determine whether spinal cord stimulation SCS in dogs affects regional myocardial blood flow and left-ventricular LV function before and during transient obstruction of the left anterior descending coronary artery LAD. In anesthetized dogs, regional myocardial blood flow distribution was determined using radiolabeled microspheres and left-ventricular function was measured by impedance-derived pressure-volume loops. SCS was accomplished by stimulating the dorsal T1-T2 segments of the spinal cord using epidural bipolar electrodes at 90% of motor threshold MT 50 Hz, 0.2-ms duration. Effects of 5-min SCS were assessed under basal conditions and during 4-min occlusion of the LAD.
In summary, SCS alone evoked no change in regional myocardial blood flow or cardiovascular indices. Transient LAD occlusion significantly diminished blood flow within ischemic, but not in non-ischemic myocardial tissue. Left ventricular pressure-volume loops were shifted rightward during LAD occlusion. Cardiac indices were altered similarly during LAD occlusion and concurrent SCS. Thus, SCS does not influence the distribution of blood flow, within the non-ischemic or ischemic myocardium. Nor does it modify LV pressure-volume dynamics in the anesthetized experimental preparation.
The majority of patients with angina pectoris secondary to coronary artery disease can be adequately controlled with medication and revascularization procedures. However, a subset of patients exists with chronic angina that is refractory to these standard treatment strategies. Neuromodulation therapy has been advocated as an adjunct therapy for patients with chronic refractory angina pectoris (DeJongste, Hautvast et al., 1994; Mannheimer et al., 1985) or even as an alternative to coronary artery bypass grafting CABG in high-risk patients (Mannheimer et al., 1998) to improve the clinical response to ischemia. With regard to safety concerns, electrical stimulation of the dorsal aspect of the spinal cord SCS does not mask anginal symptoms elicited during acute myocardial infarction (Anderson et al., 1994; Sanderson et al., 1994). Furthermore, SCS appears to have anti-ischemic properties as demonstrated during exercise stress testing (DeJongste, Hautvast et al., 1994; Sanderson et al., 1994; Hautvast et al., 1994), ambulatory ECG monitoring (DeJongste et al., 1994; Hautvast et al., 1997), and rapid right atrial pacing (Mannheimer et al, 1993; Sanderson et al., 1994).
Chauhan et al. 1994 showed that the velocity of coronary arterial blood flow of patients with either CAD measured in the left main artery with stenosis >50% in the right coronary artery or syndrome X changed when transcutaneous electrical nerve stimulation TENS; 150 Hz at 300 ms, 10-60 mA is applied for 5 min. In agreement with that, other reports indicate that SCS can increase myocardial blood flow in low-flow regions, possibly related to recruitment of coronary collateral vessels and a decrease in flow in normally perfused myocardium (Mobilia et al., 1998). Other contrary studies indicate, however, that SCS does not improve blood flow within the ischemic myocardium of patients with significant coronary artery disease or syndrome X even though it reduces ST-segment alterations (De Landesheere et al., 1992; Sanderson et al., 1996; Jessurun et al., 1998; Norrsell et al., 1998). Furthermore, at least one study has indicated that SCS does not alter total coronary blood flow in patients undergoing dipyramidole stress testing (Hautvast et al., 1996).
In light of these divergent clinical findings, examination of the influence of SCS on the distribution of regional myocardial blood flow utilizing radiolabeled microsphere technique (Baer et al., 1984) and LV chamber dynamics utilizing the conductance catheter technique (Baan et al., 1984) in canine hearts was undertaken. As has been indicated in the previously discussed experiments hereinabove, SCS does not alter cardiac indices (Foreman et al., 2000). Thus, the effects of altered cardiac workload on regional ventricular flow elicited during SCS were expected to be minimal. For that reason, the effects of SCS on the distribution of blood flow in the acutely ischemic myocardium was also examined. The duration of ischemic period was brief in order that cardiac indices return to normal values after terminating regional myocardial ischemia. The results obtained from these disclosed experiments indicate that SCS does not affect regional myocardial blood supply nor LV dynamics in the normally perfused myocardium. Furthermore, SCS does not alter blood flow within a ventricular ischemic zone; nor does it affect the ischemia-induced rightward shift of the LV pressure-volume relationship.
Animal Preparation
The experiments performed in the present study were performed in accordance with the Guide to the Care and Use of Experimental Animals set up by the Canadian Council on Animal Care and under the regulations of the Animal Care Committee at Laval University. Adult mongrel dogs of either sex, weighing between 20 and 25 kg, were used. Dogs were tranquilized with diazepam 1 mg/kg, i.v. and fentanyl 20 mg/kg, i.v. and then anesthetized with sodium pentobarbital 25 mg/kg, i.v. Noxious stimuli were applied occasionally to a paw throughout the experiments to ascertain the adequacy of the anesthesia. Repeat doses of pentobarbital 5 mg/kg, i.v. were administered throughout the experiments as required. Dogs were intubated and mechanically ventilated with a mixture of oxygen 25% and room air 75%, maintaining an end-expiratory pressure of 5-7 cm H2O to prevent atelectasis. Respiratory rate and tidal volume were adjusted to maintain arterial blood gases within physiological values. Body temperature was monitored and kept between 37.58 C and 38.58 C by a water-jacketed Micro-Temp heating unit Zimmer, Dover, Ohio, USA.
Spinal Cord Stimulation
Implantation of the Spinal Cord Stimulation Electrodes
After the animal was placed in the prone position, the epidural space was entered with a Touhy needle via a small skin incision in the lower thoracic region. A four-pole lead Medtronic QUAD Plus Model 3888; Medtronic, Minneapolis, Minn. was advanced rostrally in the epidural space to the upper thoracic level under anterior-posterior fluoroscopy and positioned slightly to the left of midline according to current clinical practice (DeJongste et al., 1994). The most cranial pole of the lead was positioned at the T1 level. Electrical current was delivered via the rostral and caudal poles to verify their functional positioning. Increasing stimulus intensity via the rostral pole as cathode to motor threshold intensity MT-induced muscle contractions in the proximal forepaw and shoulder. Stimulation with the caudal pole as cathode at MT activated thoracic paravertebral muscles, resulting in a twisting movement of the trunk. When a satisfactory electrode position was obtained, the lead, protected by a silicon sleeve, was fixed to the interspinous ligament and then connected to an external stimulator.
Threshold Determination for Spinal Cord Stimulation
The animal was shifted to the decubitus position for the remainder of the experiment and MT was then reestablished. SCS was delivered via the indwelling electrode connected to a Grass S48 Stimulator Grass Instruments, Quincy, Mass., USA via a stimulus isolation unit Grass SIU 5B and a constant current generator GrassrCCU1A. The parameters used to stimulate the spinal cord were 50 Hz and 0.2-ms duration; these values are the same as those used previously to reduce neuronal activity of intrinsic cardiac neurons in anesthetized dogs (Foreman et al., 2000). Stimulation intensity was 90% of that evoking a motor response and corresponds to the maximum used in patients (Chandler et al., 1993; Anderson et al., 1994). The current intensity used for SCS at 90% of MT, varied between 0.16 and 0.72 mA mean: 0.44 mA among animals. The most rostral and caudal poles were chosen as cathode and anode, respectively, so that the entire spinal cord area used for angina therapy in humans would be stimulated.
CardioVascular Instrumentation
Both femoral arteries and the right femoral vein were exposed and cannulated with 8F vascular introducers. Cordis, Miami, Fla., USA. A 7F 12-electrode conductance catheter with Pigtail and vascular port Cordis, Roden, The Netherlands was advanced into the LV chamber via the left femoral artery. Pressure transducers were connected to the vascular port of the conductance catheter and to a fluid-filled catheter Cordis a4 placed in the descending aorta. A femoral vein catheter was used for periodic drug injections and for fluid replacement therapy physiological saline. Surface needle electrodes were positioned to record a standard lead II electrocardiogram. Analog data were displayed on an Astro Med model MT-9500 polygraph and stored directly on a computer hard disk at a sampling rate of 333 samples/channel using the AxoScope data acquisition software Axon Instruments, Foster City, Calif., USA. Total LV pressure-volume loops were determined from changes in the electrical impedance measured by the summed volumes using a signal conditioner-processor Leycom Model Sigma-5, Oegstgeest, The Netherlands, as described previously (Baan et al., 1984). A computer analysis system Conduct-PC, Cardiodynamics, Leiden, The Netherlands was used to assess LV pressure-volume loops.
Placement of Coronary Artery Occluder
Under fluoroscopy, a modified right Judkins catheter 8F, Cordis, USA was advanced to the left coronary ostium. Thereafter, a balloon catheter was advanced into the left anterior descending LAD coronary artery. The position of the balloon catheter was verified by injection of contrast medium Hexabrix 320, Malinckrodt Medical, Pointe-Claire, CAN into the left main coronary artery, visualized in the left anterior oblique position. A baseline coronary artery angiogram was obtained to confirm positioning of the balloon in the ventral descending coronary artery about 2 cm from its origin.
Experimental Protocol
Surgical preparation and angiographic balloon catheter placement were followed by a 30-min stabilization period. Regional blood flow and LV dynamics were obtained at: 1 baseline C1; 2 during 5-min SCS; 3 return to steady-state conditions C2; 4 4 min of LAD occlusion CO; 5 return to steady-state conditions C3; and 6 5-min SCS during which time blood flow in the LAD was stopped for 4 minutes. The experimental protocol was always begun with interventions 1 and 2 since the initial goal was to assess the effects of SCS on myocardial blood supply and LV dynamics. SCS and coronary occlusion were performed twice in each animal. Four dogs underwent the following protocol sequence baseline-SCS; baseline-LAD occlusion; baseline-SCS/LAD occlusion. In another four dogs, the protocol sequence was altered baseline-SCS; baseline-SCS/LAD occlusion; baseline-LAD occlusion. Blood flow in the LAD was totally obstructed by inflating the angiocatheter balloon ns8 to a pressure of 8 atm Inde-flator Plus 20, ACS, Tomecula, Calif., USA for 4 min. Completeness of coronary obstruction was confirmed by injection of contrast medium under fluoroscopy. At least 10 min elapsed between each intervention to stabilize the experimental preparation. The time during which the coronary artery was occluded 4 min was of sufficient duration to alter regional dynamics cf., the pressure-volume relationship, yet result in a return to control values upon restoration of coronary artery blood flow.
Measurement of Regional Myocardial Blood Flow
Regional blood flow distribution was determined using the radioactive microsphere technique (Baer et al., 1984). Six different radiolabeled microspheres Sn, Sr, Nb, Sc, Ce, In, each with a diameter of 15 mm, were obtained from NEN Boston, Mass., USA. Immediately prior to injection, the microsphere suspension was agitated in a vortex mixer for 2 min. Each injection comprised 1.6−3=106 microspheres administered into the LV chamber as a bolus over 15-20 s and flushed with 15 ml of warmed saline. For each microsphere injection, a timed collection of arterial blood was performed with a Masterflex infusion/withdrawal pump Fisher, Montreal, CAN from the right femoral artery catheter at a constant rate of 7.5 ml/min beginning 10 s before microsphere injection and continuing for 2 min. Myocardial blood flow was evaluated in all dogs at six different time points: 1 during baseline state before any intervention had commenced control, 2 during the final 2 min of the 5-min SCS period, 3 baseline control a2 i.e., 10 min after return to baseline conditions, 4 at the midpoint of coronary occlusion, 5 baseline control a3, i.e., 10 min after return to baseline conditions, and 6 during the final 2 min of the 5-min SCS plus 4-min coronary occlusion period.
Anatomic Risk Zone Analysis
At the end of each study, the angiographic balloon catheter was re-inflated; contrast medium was injected to verify that the balloon was positioned in the same location used earlier to induce regional myocardial ischemia. Monastral blue dye 5 ml was injected directly into the coronary artery distal to the occlusion site to identify the ischemic zone. During deep pentobarbital sodium anesthesia, cardiac arrest was induced by intravenous injection of saturated potassium chloride. The heart and left kidney were excised rapidly from the body, rinsed in saline at room temperature and then fixed in 10% buffered formaldehyde. For blood flow analysis, the right ventricle was removed and the LV including interventricular septum was cut into 6-mm slices from apex to base parallel to the atrioventricular groove. Four transverse myocardial sections beginning with the second most apical slice were employed for blood flow analysis. The LV was divided into anterior ischemic and posterior non-ischemic segments and further subdivided into endocardial, midmyocardial and epicardial portions. The outlines of each LV slice, cavity area and the area at risk, i.e., containing blue dye were traced onto acetate sheets.
Planimetry with Sigma Scan software; SPSS, California, USA was performed on these using a digitizing tablet Summagraphics II Plus interfaced with a personal computer to determine respective surface areas. The results so obtained were expressed as the area at risk indexed to total left-ventricular mass. Regional blood flow was also assessed in 4 kidney slices excluding the most polar slice that were further subdivided into medulla and cortex regions. Radioactivity in all tissue and blood reference samples was measured in a gamma-well scintillation counter Cobra. II, Canberra Packard Instruments, Montreal, CAN with standard window settings. Tissue counts were corrected for background, decay and isotope spillover; regional blood flow ml/min/g was calculated using the PCGERDA computer software Packard Instruments and expressed in Ml/min/g.
Heart rate, arterial pressure, LV pressure and LV pressure-volume loops were evaluated on a beat-to-beat basis and averaged for 30 s prior to and during each intervention. Comparisons of cardiac hemodynamics and distribution of myocardial blood flow during different experimental conditions was performed using analysis of variance ANOVA with repeated measures. When a significant effect of treatment was obtained, pair wise comparisons were made using Scheffe's post-hoc test. All statistical procedures were performed using the SAS statistical software package SAS, Cary, N.C., USA; a pF0.05 was considered significant.
Ten dogs entered into the study; two dogs, one during AD occlusion and one during LAD occlusion with concurrent SCS went into intractable ventricular fibrillation and were excluded from the data analysis.
CardioVascular Variables
Heart rate, LV end-systolic and end-diastolic pressures and mean aortic pressure did not change during SCS. Table III LV stroke volume and ejection fraction were likewise unaffected by SCS. Monitored cardiovascular indices did not change significantly during 4 min of LAD occlusion Table III.
The decrease in LV chamber systolic and diastolic pressures was not significant presumably due to the short duration of the individual ischemic periods. Although the heart rate/LV pressure product, an index of myocardial oxygen demand, increased slightly during acute ischemia, this index did not change significantly due to large standard deviations from mean values data were not normalized. Ventricular dynamics were not significantly altered with acute LAD occlusion and concurrent SCS. Cardiac hemodynamics at the end of the experimental protocol, i.e., recovery-10 min after the final intervention was comparable to baseline values.
Regional Myocardial Blood Flow Distribution
The overall anatomic risk zone represented 21.2″5.3% mean″1 S.D. of total LV volume. Distribution of ventricular blood flow determined by radiolabeled microspheres showed that average blood flow levels decreased significantly within the ischemic zone during LAD occlusion 0.9″0.1 to 0.2″0.1 ml/min/g p-0.02. Blood flow levels were not significantly affected in the non-ischemic LV wall, the right ventricle or kidneys during LAD occlusion Table IV.
indicates data missing or illegible when filed
During application of SCS concomitant with LAD occlusion, the level of blood flow reduction in the ischemic zone was similar to that which occurred during LAD occlusion alone Table IV and
Pressure-Volume Relations
The LV pressure-volume loops did not change during SCS
Neuromodulation therapy is utilized to alleviate angina of cardiac origin. In order to investigate the underlying mechanisms for such therapy, we studied the potential influence of SCS on myocardial blood flow and LV dynamics in the normal canine heart. The results of this study demonstrate that electrical stimulation of the upper thoracic spinal cord does not alter either transmural distribution of blood flow within the myocardial wall or overall LV dynamics. As has been found in the past (Foreman et. al., 2000), transient focal myocardial ischemia did not alter left-ventricular chamber systolic or diastolic pressures significantly Table III. On the other hand, the LV pressure-volume loop was shifted rightward during transient occlusion of the LAD indicative of volume changes
Limitations of Study
The radioactive microsphere technique for determination of regional blood flow distribution has the advantage that microspheres are trapped during the first pass through an organ with no detectable recirculation; however, the number of microspheres that can be safely injected without affecting cardiac hemodynamics is finite. Baer et al. (1984) estimated that injection of 18-27 million microspheres nine different radiolabels of 2-3 million spheres each into the left atrium had little influence on distribution of blood flow during normal coronary autoregulation or vasodilatation. In the present study systolic LV pressure and mean aortic pressure remained constant during and after microsphere injections; this indicates that these dogs were hemodynamically stable during the respective experimental protocols. It is known that radioactive microspheres have the inherent limitation that regional blood flow changes less than 10% of baseline are not readily detectable. Thus, minor changes in regional blood flow distributions between myocardial regions might not be detected. Regardless, the lack of change in measured cardiac indices during SCS suggests that the primary determinant of regional myocardial blood flow and cardiac work was not altered thereby. Coronary vascular resistance or conductance was not calculated in the present study since we did not include measurements of extravascular compressive forces or critical closing pressure; in a recent study from our laboratory we document that during autoregulation the entire coronary pressure-flow relation can shift in relation to changes in LV pressure and volume Rouleau et. al., 1999. Under steady-state conditions in the present study the endocardialrepicardial blood flow ratio was similar not during ischemia; as such, distribution of blood flow was maintained across the LV wall.
Major differences exist between the present study and some clinical studies; we used SCS, while Chauhan et al., (1994), who reported an increase in blood flow in the contralateral coronary artery used TENS. In addition, SCS at 90% motor threshold in the anesthetized canine represents an intensity used clinically when patients anticipate more strenuous activities. Normally, stimulation intensities between 60% and 66% of paresthesia threshold are adequate to reduce anginal pain.
Clinical Implications of SCS
Pain-reducing properties of neuromodulation resulting from SCS are based on the gate theory of pain (Melzack and Wall, 1965). This theory proposes that stimulation of large afferent fibers conducting innocuous information reduce the nociceptive effects of the small afferent fibers on the activity of spinal neurons. Neuromodulation is known to stimulate neurons in the dorsal horn (Melzack and Wall, 1965; Chandler et al., 1993) and higher centers (Hautvast et al., 1997; Yakhnitsa et al., 1999). Recently we documented that the activity generated by intrinsic cardiac neurons is also suppressed by SCS, even during acute myocardial ischemic challenges (Foreman et al., 2000).
The present results are in agreement with the majority of previous clinical studies that indicate a lack of effect of SCS on overall coronary blood flow (De Landesheere et al., 1992; Hautvast et al., 1996; Sanderson et al., 1996; Norrsell et al., 1998). SCS was applied for 5 min in our study, while in most clinical studies it is maintained for much longer time periods. It is unlikely that the duration of SCS stimulation determines the effects that this intervention exerts on coronary artery blood flow (Chauhan et al., 1994). The present study was performed in canine hearts that underwent brief periods of regional ventricular ischemia. In clinical studies carried out among patients with stable angina, electrical or pharmacological i.e., dipyramidole induction of cardiac stress in the presence of neuromodulation has been shown to exert no influence on their coronary blood flow (Hautvast et al., 1996; Norrsell et al., 1998). In the present study, the canine coronary vasculature was considered to be normal in contrast to these clinical investigations in which coronary artery blood flow was assessed in patients with underlying coronary vessel disease. The primary determinant of blood flow in the normal myocardium is regional myocardial metabolic demand, the latter being very dependent on LV dynamics (Hoffman, 1987). Hemodynamic alterations are accompanied by changes in distribution of blood flow patterns across the LV wall (Dole and Bishop, 1982; Messina et al., 1985). For that reason, it is important to note that the periods of regional ventricular ischemia induced in these experiments were of short enough duration to induce minor, if any change in left-ventricular pressure Table III. It is also important to point out that the hemodynamic results obtained in the canine model may not directly apply to other animal models with different coronary collateral vascular function. However, intrinsic cardiac neuronal results derived from the canine model appear to be applicable to the porcine model and even to humans undergoing bypass surgery. Thus, the effects of regional ventricular ischemia on the intrinsic cardiac nervous system depend more on the location of the neurons and the site of ventricular injury than on species investigated. The ischemic area i.e., anatomic risk zone that was produced in these experiments was significant, being 21.2″5.3% mean″1 S.D. of the total ventricular volume.
Ventricular ischemic zones of this magnitude are sufficient to induce fatal ventricular arrhythmias (Vegh et al., 1991; Curtis et al., 1989). In the study reported by Vegh et al. (1991), hearts were preconditioned by repeated episodes of rapid ventricular pacing; this resulted in significant cardioprotection against ischemia-induced ventricular arrhythmias. Whether SCS triggered a preconditioning response in the present experimental model is debatable. The lack of heart rate or hemodynamic effect, reflected by the similarity of the myocardial oxygen demand and myocardial blood flow data indicates that SCS may not have induced a preconditioning response. In addition, we did not observe an increase in coronary collateral flow within the ischemic myocardium. Whether preconditioning increases coronary collateral blood flow within the ischemic zone in dogs remains unclear. In the present study, ischemic zone size was not influenced by SCS. These data are in accord with the fact that SCS did not affect the LV pressure-volume relationships in addition to arterial perfusion of ventricular tissue. We cannot completely exclude the possibility that SCS redistributes blood flow between adjacent myocardial regions via the coronary collateral circulation since the microsphere technique may not reliably detect changes in blood flow at this level. Rather, these data suggest that the anti-anginal effects of SCS are induced by mechanisms other than changes in regional myocardial blood flow or
Summary and Conclusions
Data obtained in the present study document the fact that SCS does not affect either total myocardial blood flow or blood flow distribution across the LV wall. Neither does SCS affect the distribution of blood within the ischemic myocardium, nor that between ischemic and non-ischemic zones.
Additional studies hereinafter disclosed herein, demonstrate that (1) ischemia causes neuronal activation; (2) SCS or DCA stimulation nullifies or quenches such ischemia induced activated neurons; and (3) nullification and/or such quenching or suppressor effects on such activated neurons occur (i) prior, (ii) during, and (iii) after the cessation of SCS or DCA stimulation. Thus, SCS or DCA effectively directs the intrinsic nervous system in such a manner as to have an immediate and lasting effect on the activity of myocardial neurons and the intrinsic cardiac nervous system in general.
As mentioned previously, it is known currently that electrical excitation of the dorsal aspect of the rostral thoracic spinal cord imparts long-term therapeutic benefits to patients with angina pectoris. What is not known and is being claimed and disclosed in the present application, is that spinal cord stimulation induces short-term suppressor effects on the intrinsic cardiac nervous system. The results of the following tests show that spinal cord stimulation (SCS) induces long-term effects on the intrinsic nervous system, particularly in the presence of myocardial ischemia.
The activity generated by right atrial neurons was recorded in 10 anesthetized dogs during basal states, during prolonged (15 min) occlusion of the left anterior descending coronary artery, and during the subsequent reperfusion phase. Neuronal activity and cardiovascular indices were also monitored when the dorsal T1-T4 segments of the spinal cord were stimulated electrically (50 Hz; 0.2 ms) at an intensity 90% of motor threshold (mean 0.32 mA) for 17 min. SCS was performed before, during and after 15-min periods of regional ventricular ischemia. Occlusion of a major coronary artery, one that did not perfuse investigated neurons, resulted in their excitation. Ischemia-induced neuronal excitatory effects were suppressed (>76% from baseline) by SCS. SCS suppression of intrinsic cardiac neuronal activity persisted during the subsequent reperfusion period; after terminating 17 min of SCS, at least 20 min elapsed before intrinsic cardiac neuronal activity returned to baseline values. It is concluded that populations of intrinsic cardiac neurons are activated by inputs arising from the ischemic myocardium. Ischemia-induced activation of these neurons is nullified by SCS. The neuronal suppressor effects that SCS induces persist not only during reperfusion, but also for an extended period of time thereafter.
High frequency, low intensity electrical stimulation of the dorsal aspect of the T1-T2 spinal cord alleviates angina pectoris in patients suffering from ischaemic heart disease (Eliasson et al., 1996; Mannheimer et al., 1993; Sanderson et al., 1992). The therapeutic effects of spinal cord stimulation on angina (SCS) can persist for hours after its termination (Jessurun et al., 1999). Accumulating evidence demonstrates that SCS is a safe anti-anginal treatment modality that does not result in increased frequency of arrhythmia formation (DeJongste et al., 1994; Eliasson et al., 1996; Hautvast et al., 1998; Mannheimer et al., 1998). However, the mechanisms whereby SCS produces its long-term effects remain unknown. Clinical studies have led to the hypothesis that SCS exerts its anti-anginal effects principally by altering the ventricular oxygen supply/demand ratio (Mannheimer et al., 1993; Sanderson et al., 1992). Mannheimer et al. (1993) suggested that SCS reduces cardiac metabolism, thereby reducing oxygen demand and, consequently, the myocardial lactate production within the ischemic myocardium. In this regard, Hautvast et al. (1998) proposed that SCS redistributes myocardial blood flow from normal to ischaemic regions of the heart. However, in the canine model, SCS does not alter cardiac chronotropism or inotropism (as shown in the experiments above), suggesting that oxygen demand is minimally affected by such an intervention. Furthermore, SCS does not alter blood flow distribution within either the normal or ischaemic canine myocardium (also shown above).
As we show, the effects of SCS reflect changes within the CNS and/or changes in neurohumoral control of the heart. SCS modulates impulse transmission within the spinothalamic tracts of the spinal cord without blocking afferent neuronal signals arising from the ischaemic myocardium (Chandler et al., 1993). It also alters intrinsic cardiac neuronal function (show hereinabove). The intrinsic cardiac nervous system represents the final common regulator of regional cardiac function (Armour, 1991; Ardell, 2000). Its neurons are under the constant influence of central neurons, including those in the spinal cord (Gagliardi et al., 1988). Regional myocardial ischaemia results in the heterogeneous activation of the intrinsic cardiac nervous system (Armour et al., 1998). When sub-populations of intrinsic cardiac neurons become excessively activated, the cardiac electrophysiological consequences, such as the occurrence of ventricular tachycardia or ventricular fibrillation, may be devastating (Armour, 1991). Stabilization of the intrathoracic intrinsic cardiac nervous system, especially in the presence of myocardial ischaemia ameliorate the potential for cardiac electrical instability. Such a system is shown and demonstrated in the experiments outlined in the present application.
Short duration SCS (4 min) transiently suppresses the activity generated by intrinsic cardiac neurons (shown hereinabove). In a clinical setting, the anti-anginal effects of SCS persist long after its termination (Jessurun et al., 1999). The following experiments were devised to evaluate the effects of prolonged (17 min) SCS on the intrinsic cardiac nervous system in normally perfused and ischaemic hearts. These experiments were also designed to evaluate whether the neurohormonal effects that SCS imparts on the intrinsic cardiac nervous system persist not only throughout its application, but also for a time thereafter.
Animal Preparation
The Institutional Animal Care and Use Committee of Dalhousie University approved the experiments performed in the following experiments. These experiments followed the guidelines outlined by the International Association for the Study of Pain as well as the NIH Guide for the Care and Use of Laboratory Animals (National Academy Press, Washington, D.C., 1996). Ten adult dogs of mixed breed, weighing between 12.5 and 26 kg (mean 19.6 kg), were used for this study. The animals were kept under standard laboratory conditions in a light-cycled environment (12 h/12 h) with free access to water at all times and to food at regular intervals.
Dogs were anesthetized in a standard manner by first administering a bolus dose of sodium thiopental (20 mg kg−1, i.v.). Anesthesia was maintained throughout the surgery period by means of bolus doses of thiopental (5 mg kg−1, i.v.) administered to effect every 5-10 min. Animals were intubated and then artificially ventilated using a Bird Mark VII respirator with 100% O2. After completing the surgery, anesthesia was changed to alpha chloralose by first administering a dose of alpha chloralose (75 mg kg−1, i.v.). Thereafter, repeat doses of alpha chloralose (20 mg kg−1, i.v.) were administered, as required, during the remainder of the experiments.
The level of anesthesia was checked throughout each experiment by observing pupil reaction as well as monitoring jaw tension, heart rate and blood pressure, and by periodically checking for the withdrawal reflex by squeezing a paw. Since each bolus of alpha chloralose suppressed neuronal activity for a few minutes after its administration, these doses were administered between the interventions performed in each protocol. This anesthetic regimen produces adequate anesthesia without inordinately suppressing peripheral autonomic neural activity. Electrodes were inserted in the forelimbs and the left hind limb and connected to an Astro-Med (West Warwick, R.I.) model MT 9500 eight-channel rectilinear recorder to monitor a Lead II electrocardiogram throughout the experiments. In addition, a 12-lead electrocardiogram (ECG) strip-chart recorder (Nihon Ohden Cardiofax V model BME 7707) was employed to obtain standard lead electrocardiograms during control states and at 5-min intervals during each intervention. Heart rate and the duration of the PQ, QR and QTc intervals were analyzed during control states as well as 1, 5, 10 and 15 min after occlusions began in the absence or presence of SCS. In addition, alterations in the morphology of ST-T segments and arrhythmia formation were assessed.
Implantation of Spinal Cord Stimulation Electrodes
After induction of anesthesia, animals were placed in the prone position. The epidural space of the mid-thoracic spinal column was penetrated percutaneously with a Toughy needle (15 F). A Toughy needle has a slight angle at its tip to ease penetration between vertebral processes. Using the loss-of-resistance technique as is routinely done in a clinical setting, the tip of the Toughy needle was slowly advanced until it entered the epidural space, as visualized via A-P fluoroscopy. Once the inner cannula was removed from the Toughy needle, a four-pole catheter electrode (Medtronic QUAD Plus Model 3888; Medtronic, Minneapolis, Minn.) was introduced through the needle such that its tip could be advanced to the T1 level of the spinal column, as determined by fluoroscopy. The tip of this electrode was positioned slightly to the left of the midline, as is done in a clinical setting (Linderoth and Foreman, 1999). The rostral and caudal poles of the stimulating electrode chosen for subsequent use (inter-electrode distance of 1.5 cm) were located at the levels of the T1 and T4 vertebrae. Correct placement of the stimulating electrodes was confirmed by delivering electrical current to induce motor responses using the rostral or caudal poles as cathodes, respectively.
The rostral cathode (T1 level) and caudal anode (T4 level) of the quadripolar electrode were connected to a Grass S88 stimulator via a constant current stimulus isolation unit (Grass model CCU1 and Grass SIU5). Stimuli, delivered at 50 Hz and 0.2-ms duration, were monitored on an oscilloscope to determine the amount of current delivered. Rostral stimulation above motor threshold resulted in proximal forepaw or shoulder muscle fasciculations (or both), while caudal electrode stimulation induced contractions in the thoracic trunk. When the appropriate electrode position was confirmed, the electrode lead was covered by a Teflon protective sleeve and fixed to adjacent interspinous ligaments with a suture. Extension wires attached to the electrode leads were connected to the Grass constant current stimulator (see hereinabove). Motor responses were rechecked after the animal had been placed in the supine position to ensure that the electrodes had not moved during that maneuver.
Cardiac Instrumentation
After placing the animal on its back, a bilateral thoracotomy was made in the fifth intercostal space. The ventral pericardium was incised and retracted laterally to expose the heart and the ventral right atrial deposit of fat containing the ventral component of the right atrial ganglionated plexus. We investigated the activity generated by neurons in the right atrial ganglionated plexus because not only are they representative of those found in other atrial and in ventricular ganglionated plexuses (Armour, 1991), but they do not receive their arterial blood supply from the left ventral descending coronary artery (Huang et al., 1993). The regional arterial blood supply of these neurons and other cardiac tissues is unaffected by spinal cord stimulation (Kingma et al., in press). Thus, the blood supply of identified neurons was not affected in a significant manner by the procedures described below.
Left ventricular chamber pressure was monitored via a Cordis (Miami, Fla.) #7 French pigtail catheter that was inserted into the chamber via one femoral artery. Systemic arterial pressure was measured using a Cordis #6 French catheter placed in the descending aorta via the other femoral artery. These catheters were attached to Bentley (Irvine, Calif.) Trantec model 800 transducers.
Neuronal Recording
To minimize epicardial motion during each cardiac beat, a circular ring of stiff wire was placed gently on the fatty epicardial tissue overlying the ventral surface of the right atrium containing the right atrial ganglionated plexus (Gagliardi et al., 1988). A tungsten microelectrode (10-mm shank diameter; exposed tip of 1 mm; impedance of 9-11 MV at 1000 Hz) mounted on a micromanipulator was lowered into this fat using a microdrive. The indifferent electrode was attached to mediastinal connective tissue adjacent to the heart. The electrode tip explored this tissue at depths ranging from the surface of the fat to regions adjacent to cardiac musculature. Proximity to the atrial musculature was indicated by increases in the amplitude of the ECG artifact. Signals generated by the somata and/or proximal dendrites of intrinsic atrial neurons were differentially amplified by a Princeton Applied Research model 113 amplifier with bandpass filters set at 300 Hz to 10 kHz and an amplification range of 100-500×. The output of this amplifier, further amplified (50-200×) and filtered (bandwidth 100 Hz-2 kHz) by means of optically isolated amplifiers (Applied Microelectronics Institute, Halifax, NS, Canada), was led to a Nicolet model 207 oscilloscope and to a Grass AM8 Audio Monitor. Signals were displayed on an Astro-Med MT 9500 eight-channel rectilinear recorder along with the cardiovascular variables described above. All data were stored via a Vetter (Rebesburg, Pa.) M3000A digital tape system for later analysis. Action potentials generated by neurons in a site in the right atrial ganglionated plexus were recorded.
Individual units being identified by their amplitudes and configurations. The amplitudes of the identified action potentials varied by less than 10-50 μV over several hours; individual action potential retained the same configuration over time. Somata and/or dendrites rather than axons of passage generate individual action potentials so identified. Action potentials recorded at a given locus that displayed the same configuration and amplitude were considered to be generated by a single unit. When multiple action potentials were identified at an active site, action potentials generate by individual units were discriminated by means of a window discriminator (Hartley Instrumentation Development Laboratories, Baylor College of Medicine, Houston, Tex.).
Induction of Coronary Artery Occlusion
A silk (3-0) ligature was placed around the left anterior descending (LAD) coronary artery approximately 1.5 cm from its origin, distal to its first diagonal branch. If a relatively large number of collateral arterial branches from the apex or lateral wall were evident, ligatures were also placed around these vessels. These ligatures were led through short segments of polyethylene tubing in order to occlude these arteries later in the experiments. Since the arterial blood supply of investigated right atrial neurons arises from major branches of the right and distal circumflex coronary arteries, their blood supply remained patent during these coronary artery occlusions.
Spinal Cord Stimulation (SCS)
With the animal placed in the supine position, the intensity of the current delivered via the bipolar electrode was increased until a detectable skeletal muscle motor response was evident, as described above. This current intensity corresponds to the threshold for motor activity induction (MT). An intensity of 90% of MT was used for all subsequent stimuli as it recruits A-delta fibers and other axon populations (Linderoth and Foreman, 1999). This stimulus intensity corresponds to parameters used clinically to stimulate the thoracic spinal cord (Linderoth and Foreman, 1999). The stimulus intensity at 90% MT varied between 0.09 and 0.63 mA (mean 0.32 mA) among animals studied. Presumably the variation in current intensity at 90% MT among animals reflected slight differences in electrode position with respect to the dorsal surface of the thoracic cord. The MT was checked periodically and found to remain constant over time in individual animals.
Two separate protocols were applied to each of five animals, the order of their application being randomized among the 10 animals. These were devised to evaluate the long-term effects of successive 15-min periods of coronary artery occlusion performed with or without concurrent SCS. Electrical stimuli were delivered to the dorsal aspect of the thoracic spinal cord for 17-min periods. Protocol #1 began with two 15-min periods of coronary artery occlusion, with a 1.5-h interval elapsing between occlusions (
Protocol #2 was employed in the other five animals. In protocol #2, the effects of 17 min of SCS combined with 15 min of coronary artery occlusion were studied first. The coronary occlusion was initiated 1 min after beginning SCS (
Epicardial Application of Veratridine
Veratridine is a selective modifier of Na+ channels that excites sensory neurites associated with cardiac afferent neurons without inducing tachyphylaxis (Thompson et al., 2000). This agent (obtained from Sigma, St. Louis, Mo., USA) was dissolved in physiological Tyrode solution to make a 7.5 μM solution. Gauze squares (1×1 cm) soaked with veratridine (0.5 ml) were applied for 60-100 s to discrete epicardial loci on the right ventricular conus and the ventral surface of the left ventricle at the end of each experiment (n=10 dogs). In four animals, the effects that epicardial application of veratridine exerted on the intrinsic cardiac nervous system was also tested before the protocols described above had been performed. After removing the applied gauze, the epicardial region was flushed with normal saline for at least 30 s. Gauze squares soaked with room temperature normal saline were also applied to identify epicardial sensory fields in order to determine whether neuronal responses elicited by chemical application were due to vehicle effects or the mechanical effects elicited by gauze squares.
Individual action potentials generated by the somata or dendrites of neurons within the right atrial ganglionated plexus were averaged over 30-s periods of time prior to and during each intervention. Average heart rate, left ventricular chamber systolic pressure and aortic pressure were determined concomitantly. Changes in ECG morphology induced by the protocols were assessed. When the coronary artery occlusion was performed alone, data were assessed during baseline conditions and 14 min after the occlusion began (occlusion period), as well as starting 15 s after reperfusion began (reperfusion period). When the occlusions were performed in the presence of SCS, cardiac indices and neuronal activity were assessed at five time points: (1) control period; (2) 30 s after SCS began; (3) 12 min after coronary artery occlusion began, in the presence of SCS; (4) after terminating the occlusion while the SCS persisted; and (5) within 30-60 s of terminating the SCS. Data are expressed as means±S.E.M. One-way ANOVA and paired t-test, with Bonferroni correction for multiple tests, were employed to examine grouped responses elicited during occlusion of a coronary artery alone (first occlusion) or when SCS and occlusions were performed in each protocol. Values of P<0.05 were used to determine significance.
Identification of Active Sites
Action potentials with signal-to-noise ratios greater than 3:1 were identified in 2-3 loci within the ventral right atrial ganglionated plexus of each animal. Based on the different amplitudes and configurations of action potentials recorded at one site per animal, an average of 3.2±0.5 (range 2-6) neurons generated spontaneous activity at investigated sites during control states. Neuronal activity during basal states was usually sporadic in nature. During basal states, a few spontaneous active neurons were identified in active loci of most animals (
Effects of Transient Myocardial Ischaemia
Monitored cardiac indices did not change significantly overall during coronary artery occlusion or the reperfusion period, except when cardiac arrhythmias occurred. For instance, heart rate was 134±2 beats/min (bpm) before occlusion and 130±3, 134±3, 132±2 and 134±2 bpm after 1, 5, 10 and 15 min of ischaemia, respectively. S-T segment alterations and terminal QRS slurring was evident in the ECG pattern of each animal during ischaemic episodes (
Overall, these electrophysiological data substantiate the substantial ischaemia insult that was induced by 15-min periods of left anterior descending coronary artery (LAD) occlusion. When the LAD was occluded in either protocol in the absence of SCS, the activity generated by right atrial neurons (
Table V shows heart rate (HR), left ventricular chamber systolic pressure (LVP), aortic systolic and diastolic pressures (AP) and the activity generated by right atrial neurons recorded before (control) and during coronary artery occlusion (CAO), as well as during early reperfusion (reperfusion). These indices were also recorded when occlusion occurred during spinal cord stimulation.
Effects of Spinal Cord Stimulation in the Presence of Myocardial Ischaemia
During normal coronary artery perfusion, SCS did not alter the ECG or monitored cardiac indices (Table V). The activity generated by identified right atrial neurons was reduced from 48.2±6.5 to 23±2.5 impulses/min within 30 s of applying electrical current to the dorsal aspect of the rostral thoracic spinal cord in hearts with normal coronary arterial blood supply (
Epicardial Application of Veratridine
The activity generated by right atrial neurons increased when veratridine was applied topically to their ventricular sensory inputs. Veratridine-induced excitation of intrinsic cardiac neurons was studied both before and after application of SCS in four animals. In those cases, veratridine enhanced intrinsic cardiac neuronal activity by 124% (40.5±26.7; P<0.05) before application of SCS and by only 39% (11.8±2.5 to 16.5±4.6 impulses/min; no significant difference) after its application. When veratridine was applied to their ventricular sensory inputs after completing the protocols in all 10 dogs (following SCS and regional ventricular ischaemia), intrinsic cardiac neuronal activity increased by only 58% (25.6±5.7 to 40.6±12.5 impulses/min; P<0.05).
FIG. II shows representative ECG records obtained from one animal during control states (A), as well as a few minutes after beginning coronary artery occlusion in the presence of spinal cord stimulation (B) and at the end of occlusion while SCS was maintained (C). Note that ST segment alterations occurred throughout the period of ischaemia.
The results obtained from the experiments conducted in the present study not only confirm that spinal cord neurons can modulate the intrinsic cardiac nervous system (as discussed hereinabove), but they demonstrate that such modulation persists unabated throughout 17-min periods of stimulating the dorsal thoracic spinal cord. They also indicate that spinal cord neurons continue to exert their suppressor effects on the intrinsic cardiac nervous system long after their activation terminates. Furthermore, these data indicate that spinal cord neurons reorganize information processing within the intrinsic cardiac nervous system arising from the ischaemic myocardium, including during the reperfusion post-ischaemic phase. Finally, as indicated by the neural responses evoked by veratridine application to the ventricular epicardium, the stabilizing influence that SCS exerts on the intrinsic cardiac nervous system extends to intrinsic cardiac reflex responses evoked by activating cardiac sensory neurites associated with afferent neurons within the cardiac neuroaxis.
Given that bilateral transection of the ansae subclavia abolishes the neuro-suppressor effects that SCS imparts upon the intrinsic cardiac nervous system (as discussed hereinabove), it appears that the sympathetic nervous system is involved in the effects of SCS on the intrinsic cardiac nervous system. Activation of spinal cord neurons may inhibit intrinsic cardiac local circuit neurons in a manner similar to that which occurs when they receive increasing inputs from sympathetic efferent preganglionic neurons (Murphy et al., 1995). Based on the results obtained during application of SCS to the lumbosacral spinal cord (Linderoth and Foreman, 1999), sympathetic afferent as well as efferent axons may contribute to the suppressor effects that SCS exerts on the intrinsic cardiac nervous system.
Activation of sympathetic efferent preganglionic axons attenuates the activity generated by sub-populations of neurons within intrathoracic ganglia, including those on the heart (Armour, 1991; Murphy et al., 1995). Supramaximal stimulation of sympathetic efferent preganglionic neurons also leads to a rapid reduction in the capacity of intrathoracic sympathetic efferent neurons to influence cardiodynamics (Butler et al., 1988). It has been proposed that such suppressor effects are most likely due to inhibitory synapses within intrathoracic ganglia, including those on the heart (Armour, 1991). In accord with that, spinal cord neurons, when activated, suppress the activity generated by intrinsic cardiac neurons.
It is known that the activity generated by many intrinsic cardiac neurons increases secondary to transient ventricular ischaemia (Armour et al., 1998). Right atrial neurons are supplied by arterial blood in the sinoatrial artery arising from the right coronary artery and distal branches of the circumflex coronary artery (Huang et al., 1993). Since occlusion of the left anterior descending coronary artery does not compromise the arterial blood supply of investigated right atrial neurons (Huang et al., 1993), the effects that regional ventricular ischaemia exerted on investigated neurons were primarily the result of ischaemia-induced enhancement of ventricular sensory neurite inputs to identified neurons rather than any direct effects of ischaemia on identified somata and/or dendrites (Armour, 1991). Intrinsic cardiac neuronal activity remained elevated throughout the 15-min periods of regional ventricular ischaemia when performed in the absence of SCS. That these regional coronary artery occlusions affected the ST segments of the ECG presumably is reflective of the underlying myocardial ischaemia so induced. The processing of cardiac sensory information within the intrinsic cardiac nervous system was affected by SCS, as indicated by changed neuronal responsiveness to chemical (veratridine) activation of their ventricular sensory inputs.
Given the fact that the capacity of veratridine to affect sensory neurites associated with cardiac afferent neuron exhibits no tachyphylaxis (Thompson et al., 2000), the changed transduction properties of ventricular sensory inputs to the intrinsic cardiac nervous system is due, in part, to remodeling of the intrinsic cardiac nervous system subsequent to SCS. It should be noted that in clinical studies, the sensory effects that SCS imparts persist long after the stimulation has stopped. Patients with refractory angina pectoris continue to experience decreased episodes of pain after terminating SCS (Jessurun et al., 1999). Furthermore, the allodynia associated with neuropathic pain can be reduced for as long as 1 h after terminating SCS (Stiller et al., 1996). Application of SCS immediately prior to onset of LAD occlusion did not blunt the evolution of ischaemic-induced changes in the ECG. It is unlikely that the results obtained by SCS in a clinical setting can be ascribed to alterations in hemodynamics (Hautvast et al., 1998; Linderoth and Foreman, 1999) or coronary artery blood flow (as discussed hereinabove). This is because SCS exerts its primary effects on the intrinsic cardiac nervous system that, in turn, may influence control over regional cardiac electrical or mechanical events. These data indicate that activation of spinal cord neurons induces a conformational change in the intrinsic cardiac nervous system that persists for a considerable period of time after terminating such activation. This remodeling of the intrinsic cardiac nervous system can override excitatory inputs to it arising from the ischaemic myocardium. Thus, thoracic spinal cord neurons act to stabilize the intrinsic cardiac nervous system in the presence of ventricular ischaemia and during reperfusion. The prolonged salutary effects that SCS imparts to some patients long after it is discontinued is due, in part, to remodeling of the cardiac nervous system.
Cardiac control is represented by dynamic interactions occurring between myocyte, neuronal and hormonal factors (Ardell, 2004; Armour, 2004). Each of these factors represents a potential target for cardiac therapy. Acute coronary artery syndromes (ACS) represent multifaceted processes involving both direct effects on cardiac myocytes and their modulation by neurohumoral factors (Kloner and Rezkalla, 2004; Sanada and Kitakaze, 2004). A major adverse consequence of ACS is angina pectoris (Foreman, 1999). While early revascularization by percutaneous transmural interventions has become the primary therapy to alleviate adverse consequences of ACS (Keely et al., 2003), adjunct pharmacological therapies such as those aimed to resolve or prevent clot formation (Kloner and Rezkalla, 2004), beta-adrenoceptor blockers (Colucci, 2003; Kloner and Rezkalla, 2004) or adenosine (Cohen and Downey, 2001) can contribute to further infarct reduction.
Recently, electrical neuromodulation using spinal cord stimulation (SCS) has been shown to represent a safe, long-term adjunct therapy for patients with chronic angina pectoris refractory to standard treatments (Ekre et al., 2002). Electrical stimuli delivered to the upper thoracic spinal cord segments suppress pain associated with myocardial ischemia (Mannheimer et al., 2002). Clinically, SCS also exhibits anti-ischemic properties that include increased exercise tolerance (Hautvast et al., 1998), diminished ST-segment deviation during stress (Cardinal et al., 2004; Hautvast et al., 1998) and improved myocardial lactate metabolism (Mannheimer et al., 1993). Clinical studies have indicated a maintained relationship between anginal pain and relative levels of myocardial ischemia during SCS (Mannheimer et al., 2002).
Without wishing to be held to theory, the present desired mode of therapy appears to produce its beneficial effects via several mechanisms. Spinal cord stimulation influences the processing of information within the central nervous system (Chandler et al., 1993; Foreman et al., 2004). Animal studies indicate that SCS may inhibit spinothalamic tract neurons (Chandler et al., 1993). It also influences information processes within the intrathoracic cardiac nervous system (Armour et al.; Foreman et al., 2000). With respect to the latter, SCS exerts a stabilizing influence on both the basal activity generated by intrinsic cardiac neurons as well as reflexly evoked enhancement of such activity during transient ventricular ischemia (Armour et al.; Foreman et al., 2000). These stabilizing effects are eliminated by transection of neuronal connections between the spinal cord and the intrathoracic cardiac nervous system (Foreman et al., 2000). Since sympathetic efferent and afferent projections are disrupted by such transections (Ardell, 2004) and since catecholamines can induce cardioprotection against apoptosis via both alpha and beta-adrenergic receptor mechanisms (Sanada and Kitakaze, 2004), early-phase cardioprotection induced by preemptive SCS is dependent to a considerable extent on adrenergic efferent neurons.
Subjects. One hundred and thirty-four New Zealand White rabbits of either sex, weighing between 1.95 and 3.55 kg, were used in these acute studies. All experiments were performed in accordance with the guidelines for animal experimentation described in the “Guiding principals for research involving animals and human beings” (World Medical Association, American Physiology Association, 2002). The Institutional Animal Care and Use Committee of the East Tennessee State University approved these experiments.
Surgical preparation. Rabbits were anesthetized with intravenous sodium pentobarbital (30 mg/kg iv, supplemented as needed; i.e., 2 mg/kg iv, if the animal responded to noxious stimuli or control arterial blood pressure increased). The trachea was intubated via a cervical incision and mechanical ventilation initiated and maintained with a positive pressure ventilator (MD Industries, Mobile, Ala.) using 100% O2. Core body temperature was maintained at 38° C. via a heating pad. The right carotid artery and jugular vein were each cannulated for blood pressure monitoring and administration of additional anesthesia and drugs, respectively. Heart rate was assessed from a lead II electrocardiogram. All hemodynamic data were recorded concurrently on a Gould Model TA6000 rectilinear recorder.
A laminectomy was performed at the T1 level, followed by the subdural placement of two plate electrodes (2 mm×3 mm) slightly to the left of the midline at the C8 and T2 level. Spinal cord stimulation (SCS) was delivered via these indwelling electrodes connected to a Grass S88 Stimulator (Grass Instruments, Quincy, Mass., U.S.A.) using a constant current stimulus isolation unit (Grass PSIU 6G). The parameters used to stimulate the spinal cord were: 5 or 50 Hz; 0.2 msec duration. To determine the adequate stimulus intensity, current intensity was progressively increased until minor muscle contractions were induced in the left upper forelimb (Motor Threshold: MT). Current intensity used for the experimental protocols was set at 90% of MT; this intensity averaged 0.84±0.34 mA. Stability of the stimuli intensity was confirmed by repeat motor threshold determinations at the end of the experimental protocols.
A thoracotomy was performed in the left, fourth intercostal space. Subsequently, the pericardium was opened to expose the heart. A 2-0 silk suture on a curved tapered needle was passed around the left coronary artery at a level 1/3 of the distance from the LV base to apex. The ends of the suture were pulled through a small polyethylene tube to form a snare. Subsequently, the coronary artery was transiently occluded by pulling on the snare and secured by clamping the tube with a hemostat. Cyanosis of the muscle in the territory downstream to the vessel occlusion site confirmed the induction of regional ventricular ischemia. Myocardial tissue perfused by the snared coronary artery was considered the zone at risk; the remainder of the left ventricle was considered to be the non-risk zone. A 20-min stabilization period preceded the onset of each experimental protocol.
Pre-emptive electrical neuromodulation. Animals were subdivided into 6 separate groups; each was subjected to 30 minutes of coronary artery occlusion, followed by a 3 hr reperfusion period. For these animals, a SCS stimulation frequency of 50 Hz was employed with the exception of group 3.1 in which SCS was applied at 5 Hz.
Preemptive protocol group 2 (n=9) involved 5 minutes of spinal cord stimulation (SCS) followed by a 10 minute rest period before applying the longer duration SCS (32 min). In this group, the 30-minute period of coronary artery occlusion began 1 minute after the onset of the longer duration SCS (
Animals in protocol 3 (n=35) of the pre-emptive group were subjected to 46 minutes of SCS. In this group, the 30-minute period of ischemia was initiated 15 minutes after beginning SCS and terminated 1 minute before completing SCS. In 24 of these animals, SCS was delivered at a frequency of 50 Hz (group 3) and in 11 animals SCS was delivered, at a frequency of 5 Hz (Group 3.1).
Animals in protocol 4 (n=11) of the pre-emptive group were subjected to 30 minutes of SCS followed by a 15-minute period of rest. Thereafter, SCS was maintained for a period of 46 minutes. In this group, the 30-minute ischemic period began 15 minutes after commencing the second SCS.
In order to determine whether adrenergic receptor blockade modifies the effects of pre-emptive SCS neuromodulation, 17 additional animals underwent Protocol 3 (50 Hz) 15 minutes after administering either the α-adrenoceptor blocking agent prazosin (0.15 mg/kg, iv; n=8) or the β-adrenoceptor blocking agent timolol (2 mg/kg, iv; n=9).
Nine animals were excluded from final analysis owing to myocardial ischemia induced terminal ventricular fibrillation (VF) events and thus failure to complete the entire protocol. Terminal VF was induced in 9.1% of the control group (n=3), in 6.8% of the animals receiving pre-emptive SCS (5 animals: 3 from protocol 3 sham treatment plus 1 each from protocol 3 with prazosin or timolol pretreatment), and 4.3% of the animals receiving reactive SCS (1 animal from protocol 6).
Reactive Electrical Neuromodulation.
Animals were subdivided into 3 separate groups; each was subjected to 30 minutes of coronary artery occlusion, followed by a 3 hr reperfusion period. For these animals, SCS was delivered at a stimulation frequency of 50 Hz at 90% MT. For protocol 5 (n=8), 30 min of SCS commenced one minute after onset of the 30 minute coronary artery occlusion. For animals in protocols 6 (n=6) and 7 (n=9), SCS was initiated at 28′ or 1′ of CAO respectively and maintained throughout reperfusion.
Infarct Size Measurement.
At the end of each experiment, the hearts were rapidly excised, mounted on a modified Langendorff apparatus and perfused with room temperature 0.9% saline to remove blood from the coronary circulation. The same coronary artery site previously occluded was subsequently re-occluded. Thereafter, 2-9 μm fluorescent polymer microspheres (Duke Scientific Corporation; Palo Alto, Calif., USA) were injected into the coronary artery perfusate in order to demarcate the ventricular region at risk. After removing both atria, the rest of the heart was weighed and then frozen at −20° C. The heart was then cut into 2 mm thick slices, parallel to the AV groove. Tissue slices so obtained were incubated for 20 minutes at 37° C. in 1% triphenyltetrazolium chloride (TTC) and sodium phosphate buffer (pH 7.4). These tissue slices were then placed in 10% formalin to improve the contrast between stained and unstained tissue. Areas of infarction (TTC negative), risk zone (negative fluorescence under UV light) and the non-risk zone (positive fluorescence under UV light) were determined by planimetry. The infarct and risk zones were calculated by multiplying each area by the tissue thickness and their products were summed. Infarct size is expressed as a percentage of the risk zone.
Phosphorylation of PKCs.
Four additional animals were prepared to evaluate the capacity of SCS to induce phosphorylation of left ventricular PKC. Following laminectomy and electrode placement, the treatment group (n=2) underwent 46 min of SCS (50 Hz, 200 μs, 90% MT) followed by 30 min recovery while the sham group (n=2) served as a 76 min time control. At the end of these experiments, LV tissues were flash frozen for subsequent analysis. From these LV samples, tissue lysates were prepared in lysis buffer (1% Triton X-100, NaCl 150 mmol/L, Tris 10 mmol/L, pH 7.4, EDTA 1 mmol/L, EGTA 1 mmol/L, phenylmethylsulfonyl fluoride 0.2 mmol/l, sodium orthovanadate 0.2 mmol/L and 0.5% nonidet P-40) using homogenizer. Equal amounts of proteins (50 μg) were precleared from endogenous rabbit IgG by incubating the lysates with protein A agarose beads. The lysates were resolved by 10% SDS-polyacrylamide gels and the proteins were transferred to PVDF membrane. The membranes were blocked in TBS containing 5% milk and 0.1% Tween 20 and then incubated with primary antibodies (phospho-PKC; 1:1000 or phospho-PKC-zeta; 1:1000; Cell Signaling Technology, Beverly, Mass.). Protein loading was normalized using β-actin immunostaining.
Statistical Analysis.
All data are presented as means±SD. Sigmastat 3.1 (Systat Software, Inc), one-way analysis of variance with post-hoc comparisons (Holm-Sidak method) was used to test for differences within and between groups.
Effects of SCS Neuromodulation on Hemodynamics. Table 6 summarizes average body weight, risk zone and baseline hemodynamic data for each of the experimental groups (control, pre-emptive SCS and reactive SCS neuromodulation), as well as for the associated subgroups. Table 7 summarizes the hemodynamic data obtained consequent to producing pre-emptive neuromodulation (protocol #3; 50 Hz SCS) before, during and 2 hours after coronary artery occlusion. Table 8 summarizes the hemodynamic data for reactive SCS neuromodulation in protocols 5-7.
214.1 ± 16.4#@
212.3 ± 12.9@
211.0 ± 10.9@
75.3 ± 15.5
72.9 ± 8.0+
+p < 0.01 from baseline;
#p < 0.01 paired comparison to pre-block baseline (see Table 6).
@p < 0.02 from vehicle and prazosin pre-treatment groups.
77.3 ± 9.0#
+p < .02 from coronary occlusion;
#p < .04 from baseline.
Hemodynamic variables. Prazosin pretreatment induced a significant decrease in basal mean arterial blood pressure (86.5±8.3 versus 66.8±8.1 mmHg). Timolol pretreatment reduced heart rate (untreated controls: 234.2±21.7 beats/min; treated: 214.1±12.9 beats/min); these changes were sustained throughout the subsequent observation periods. Within group comparisons for pre-emptive neuromodulation (Table 7) indicated that SCS by itself induced no significant change in heart rate or blood pressure. During the reperfusion period, blood pressure was reduced from baseline values in the vehicle control and timolol pre-treated groups. For reactive SCS (Table 8), heart rate did not change significantly throughout the protocols, while blood pressure was significantly reduced from baseline during CAO for reactive SCS protocols 5 and 7. By 2 hr of reperfusion, blood pressure was further reduced from both baseline and CAO levels in all reactive SCS groups.
Effects of Spinal Cord Stimulation on Infarct Size. Body weight and left ventricular risk zones were similar in all experimental groups (Table 6).
Pre-emptive SCS impacted on IS induced by 30 minutes of ischemia with 3 hr reperfusion. As summarized in
In contrast to the infarct reduction induced by pre-emptive SCS, reactive SCS did not reduce infarct size (
Effects of Spinal Cord Stimulation on PKC Phosphorylation. Phosphorylation of total myocardial PKC and PKC-zeta increased in the SCS treatment group when compared to tissues derived from sham animals (
Effects of Spinal Cord Stimulation on Ventricular Fibrillation. Myocardial ischemia induced ventricular fibrillation (VF) in 3 of 33 animals in the control group, in 5 of 73 animals with pre-emptive SCS and in 1 of 22 animals with reactive SCS. For pre-emptive SCS, all 5 terminal ventricular fibrillation events occurred in protocol 3, with 3 in the vehicle control group and 1 each in the prazosin and timolol pre-treatment groups. All terminal VF events occurred during CAO. Transient periods of VF (33.0±19.0 sec duration) were noted in 2 animals in the control group, 4 animals with pre-emptive SCS, and 3 animals with reactive SCS. Animals with terminal VF events were excluded from subsequent data analysis.
This study demonstrates for the first time that pre-emptive electrical neuromodulation therapy with SCS reduces the size of infarcts induced by transient ventricular ischemia. These data also indicate that such SCS-induced cardioprotection involves cardiac adrenergic neurons, acting principally via α-adrenoceptors, and that pre-emptive SCS increases phosphorylation of cardiac PKC pathways. The capacity of pre-emptive SCS to reduce infarct size depends on the duration and frequency of the stimulus applied to the dorsal aspect of the thoracic spinal cord. In accord with data derived from experimental animals (Armour et al., 2002; Foreman et al., 2000; and Kingma et al., 2001), this form of therapy exerts minimal effects on basal cardiac function.
Pre-emptive SCS demonstrates frequency dependence, as evidenced by differences in cardioprotection induced by 5 versus 50 Hz electrical stimuli applied to the spinal cord at the same relative intensity and for the same duration. Moreover, preemptive SCS demonstrated a threshold effect in as much as 5 minutes of SCS with 10 minutes rest was insufficient to impart a detectable benefit. Interestingly, adding a 30 min cycle of pre-emptive SCS 15 min prior to the longer duration (46 minute) SCS did not provide additional cardioprotection for infarct reduction. In contrast, reactive SCS was not effective in reducing infarct size, even when initiated within 1 min of coronary occlusion onset and maintained during reperfusion.
Catecholamines and cardioprotection. Adrenergic receptors, coupled to distinct signal transduction pathways, affect cardiac myocytes and the neurons that regulate them (Sanada and Kitakaze, 2004; Yang et al., 2004). For cardiomyocytes, PKC activation represents a major mediator in preconditioning (Cohen and Downey, 2001; Sanada and Kitakaze, 2004); it can be activated via multiple pathways that include α1-adrenergic receptors (Tsuchida et al., 1994). Exogenous activation of α1-adrenoceptors induces early and late phase preconditioning for reduction of infarct size and stunning (Bankwala et al., 1994; Marktanner et al., 2003; Stein et al., 2004; Thornton et al., 1993). During early phase preconditioning, such activation also reduces ischemia/reperfusion induced cardiac arrhythmias (Vegh and Parratt, 2002). β-adrenoceptor activation can also limit IS to transient myocardial ischemia via a non-PKC pathway that involves PKA and p30MAPK (Kloner and Rezkalla, 2004). While endogenous release of norepinephrine following tyramine administration can also reduce infarct size (Bankwala et al., 1994; Thornton et al., 1993), and depletion of cardiac catecholamines using reserpine raises PC threshold (Ardell et al., 1996; Toombs et al., 1993), most data indicate that the major myocyte triggers for ischemic preconditioning do not include catecholamines (Cohen and Downey, 2001; Thornton et al., 1993). Yet, as demonstrated herein, pre-emptive SCS reduces IS induced by transient myocardial ischemia. Such protection involves neurally-dependent activation of cardiac PKC pathways coupled to α-adrenergic receptors. In addition to direct effects on cardiomyocytes, our results indicate that α-adrenergic receptors can also modulate the cardiac nervous system to effect cardiomyocyte viabililty.
Neuromodulation and cardioprotection. Integrated control of regional cardiac function represents the dynamic interplay between local factors, such as the Frank-Starling mechanism, the cardiac nervous system and circulating hormones (e.g., angiotensin II, epinephrine) to modulate cardiac tissues (Ardell, 2004; Armour, 2004). The evolution of cardiac pathology is associated with remodeling of these elements such that imbalance of neurohumoral control, especially excessive sympathetic efferent neuronal activation, can induce adverse cardiac events (Armour, 2004; Dell'Italia and Ardell, 2004; Schwartz, 2001). The reason that cardiac pathology can be effectively managed by beta-adrenoceptor blockade and/or ACE inhibitors (Dell'Italia and Ardell, 2004; Kloner and Rezkalla, 2004) may be related to the fact that these agents act not only on cardiomyocytes directly, but also indirectly via the cardiac nervous system (Armour, 2004). Specifically, sub-populations of neurons within the intrathoracic cardiac neuronal hierarchy possess adrenoceptors (Ardell, 2004; Armour, 2004) and modulation of their activity can impact on the evolution of cardiac pathology (Dell'Italia and Ardell, 2004; Killingsworth et al., 2004; Tallaj et al., 2003).
A key issue with regard to understanding the mechanisms whereby SCS exerts its beneficial cardiac effects relates to the fact that the processing of information arising from the ischemic myocardium by the intrinsic cardiac nervous system can be modified by electrical neuromodulation therapy. Electrical activation of neurons in the upper thoracic spinal cord modifies the processing of information not only within the central nervous system (Chandler et al., 1993; Foreman et al., 2004) but also the intrinsic cardiac nervous system (Armour et al., 2002; Foreman et al., 2000). SCS does not appear to interfere with primary efferent neuronal control of nodal tissues in the heart (Olgin, 2002), but does exert immediate and long-lasting stabilizing effects on the intrinsic cardiac nervous system by suppressing its excessive reflex activation during focal myocardial ischemia (Armour et al., 2002; Foreman et al., 2000). As such, SCS may blunt reflex activation of sympathetic efferent neurons elicited by acute myocardial ischemia as first described by Malliani et al in 1969 (Malliani et al., 1969). Our results indicate that there are at least two separate processes involved in SCS-mediated effects in reducing IS, as characterized in this study. The first involves a low level neural-dependent catecholamine release into the cardiac interstitium that when evoked pre-coronary occlusion will activate cardiomyocyte PKC pathways to induce protection. The second acts via the cardiac nervous system to limit reflex activation of cardiac sympathetic efferent neurons during the ischemic insult itself. In support of the second proposed mechanism, Cardinal et al. (Cardinal et al., 2004) recently demonstrated that sympathetic efferent neuronal dependent ST segment deviations in the stressed myocardium are mitigated by SCS.
The specific descending neural pathways that are modified by SCS to affect peripheral neuronal and myocyte interactions remain poorly defined. SCS modulation of the intrinsic cardiac nervous system is eliminated by bilateral ansae transection, a surgical procedure that eliminates afferent and efferent sympathetic axons coursing between spinal cord neurons to intrathoracic neurons (Foreman et al., 2000). For instance, SCS may induce neuropeptide release from cardiac sympathetic afferent nerve terminals (e.g. CGRP, substance P, opiates) to affect adjacent neuronal and/or myocyte function (Croom et al., 1997; Eliasson et al., 1998; Hoover et al., 2000). Activation of sympathetic efferent preganglionic neurons may also alter the processing of information within the intrathoracic sympathetic nervous system (Butler et al., 1988; Murphy et al., 1995). With respect to SCS-mediated effects inducing infarct size reduction, data presented herein indicates that this involves cardiac neuronal β-adrenoceptors, with secondary contributions by β-adrenoceptors. Different populations of cardiac adrenergic neurons are known to posses both these receptor subtypes (Armour, 1997). Activation (or blockade) of them can modify regional cardiac function (Armour, 1997) and alter neurotransmitter release from sympathetic efferent nerve terminals (Tallaj et al., 2003).
Neuromodulation therapy and arrhythmias. In addition to effects on cardiomyocyte viability, myocardial ischemia impacts on cardiac electrical function. It is a classical concept that excessive sympathetic neuronal activity can increase dispersion of ventricular electrical events that ultimately result in ventricular fibrillation (Han and Moe, 1964) secondary to excessive local release of catecholamines (Han et al., 1964). The clinical importance of this is emphasized by the effectiveness of β-adrenergic blockade in reducing sudden cardiac death in patients' post-myocardial infarction (Schwartz et al., 1992). In the current study, pre-emptive SCS did not reduce the incidence of sudden cardiac death associated with transient myocardial ischemia in the rabbit model. However, it should be considered that the lack of collateral blood flow in the risk zone in the rabbit model (Maxwell et al., 1987) will impact on local neural and myocyte function and as such are not necessarily reflective of what may occur in humans. Indeed, in a canine model with regional supply/demand imbalance induced by a chronic ameroid constrictor, sympathetic efferent neuronal dependent ST segment deviations in the stressed myocardium are mitigated by SCS (Cardinal et al., 2004). Thoracic SCS can also reduce the potential for ischemic induced ventricular tachycardia/ventricular fibrillation in a canine model of healed myocardial infarction and pacing-induced heart failure (Issa et al., 2005). Taken together, these data indicate that neuromodulation therapy impacts on multiple aspects of the adverse cardiac pathology associated with myocardial ischemia.
Perspectives. Although the specific contributions that SCS initiate concerning alterations in neurotransmitter release within the intrinsic cardiac nervous system versus its direct cardiomyocyte effects remain to be fully elucidated, data derived from these experiments indicate that this form of neuromodulation therapy in addition to its anti-anginal properties is effective in mediating ventricular infarct reduction during transient myocardial ischemia. The ineffectiveness of reactive SCS to reduce infarct size in the acute setting represents a limitation. However, in clinical practice, SCS has been shown to be a long-term adjunct therapy for patients with chronic angina pectoris (Ekre et al., 2002). It should be considered that as an unrecognized benefit to chronic SCS therapy, these patients may experience a relative state of cardioprotection to transient periods of myocardial ischemia.
Activation of the dorsal aspect of the upper thoracic spinal cord with high frequency electrical stimuli (50 Hz) of short duration (0.2 ms) has been proposed as an alternative therapeutic modality for patients with chronic angina pectoris refractory to standard medical or surgical therapy (Sanderson et al., 1994: Mannheimer et al., 1998, 2002). Mechanisms involved in the anti-anginal effects of spinal cord stimulation (SCS) remain unclear. It has been suggested that dorsal column activation may inhibit pain perception, something that has received support from primate experimentation (Chandler et al., 1993). Another possibility is that spinal cord stimulation induces anti-ischaemjc effects via improved myocardial perfusion and/or decreased myocardial O2 consumption. Although, global myocardial blood flow appears to be unaffected by spinal cord stimulation (Meyerson and Linderoth, 2000), it has been suggested that coronary blood flow might be redistributed from non-ischaemic to ischaemic ventricular areas (Hautvast et al., 1998). In patients with compromised coronary artery blood supply, the magnitude of electrocardiographic ST segment changes elicited during exercise can be reduced by spinal cord stimulation at comparable workloads, as can those induced by rapid cardiac pacing (Sanderson et al., 1992; Mannheimer et al, 1993; Hautvast et al., 1998).
Implantation of an ameroid constrictor around the proximal left circumflex coronary artery (LCx) causes the artery to become slowly obliterated while collaterals develop, as the material takes up tissue water and swells over 3-4 weeks (Schaper. 1971; Tomoike et al., 1983). Scar formation is thus avoided but a collateral-dependent myocardial substrate is created (Katagiri et al., 1978) Canines with chronically implanted ameroid constrictors studied in the conscious state displayed normal subendocardial contractile function at rest, that deteriorated during transient rapid ventricular pacing or exercise, as total flow increased but was redistributed away from subendocardial layers (Fedor et al., 1980: Kunlada et al., 1982). We recently reported that, in such ameroid canine preparations, unipolar electrograms recorded from the epicardium display ST segment elevation (>4 mV) in the center of the LCx territory and ST depression (<−4 mV) at more peripheral sites (Cardinal et al., 3004). Moreover, ST segment shifts were augmented in response to transient bouts of rapid ventricular pacing in such preparations, and repolarization intervals shortened in central areas of the LCx, thereby causing local gradients >20 ms between neighbouring sites (activation of ATP-dependent potassium channels appeared to be involved in such responses). This appeared to be a useful paradigm in which to investigate the effects of spinal cord stimulation on cardiac electrical events. One objective of the present study was, therefore, to determine whether spinal cord stimulation could prevent ST segment shifts and alterations of repolarization intervals in the LCx territory of such preparations.
Previous studies performed in anesthetized canines indicate that electrical activation of the dorsal aspect of the spinal cord suppresses the activity generated by intrinsic cardiac neurons (Foreman et al., 2000). Spinal cord stimulation also obtunds the excitatory effects that regional ventricular ischaemia imparts to the intrinsic cardiac nervous system (Armour et al., 2002). Intrinsic cardiac neurons can be activated in response to local injection of several pharmacological agents, among which is angiotensin II (Levett et al., 1996), an action that can lead to the induction of arrhythmias (Huang et al., 1994). Neuronally induced positive cardiac inotropic effects are also elicited in response to angiotensin II when it is administered locally to populations of intrinsic cardiac neurons via their coronary arterial blood supply (Horackova and Armour, 1997). The second objective was, therefore, to investigate the effects of intra-coronary angiotensin II injection on ST segment displacements in canines with chronically implanted ameroid constrictors with a view of testing whether local neuronally induced responses to angiotensin II would be affected (presumably inhibited) by spinal cord stimulation.
All experiments were performed in accordance with guidelines described in “Guide for the Care and Use Laboratory Animals” (NIH publication 85-23, revised 1996), as well as guidelines of the Canadian Council for Animal Care, as monitored by a certified institutional Ethics Committee that granted approval to this study. Seventeen conditioned mongrel canines of either sex, weighing 21-44 kg, were included in this study. It was performed in two steps: (i) the initial surgery for implantation of an ameroid constrictor and (ii) later electrophysiologic study at which time data were collected.
Canine Preparations
Initial surgery was conducted under aseptic conditions and anesthesia (thiopental 25 mg/kg (i.v.), followed by isoflurane 1%). Pulmonary capillary wedge pressure, and cardiac output measurements (thermodilution) were measured by means of a Swan-Ganz catheter introduced via the left jugular vein. The lateral left ventricular (LV) wall was exposed via limited left intercostal thoracotomy (4th-5th ribs). The proximal left circumflex coronary artery (LCx) was exposed approximately 1 cm from its origin. The ameroid constrictor (purchased from Research Instruments; Escondido, Calif.) consists of a stainless steel ring filled with hygroscopic material, with a central hole and sideways aperture that allowed the device to be placed around the artery. The hygroscopic material within the metal ring swells with time as it takes up water, causing the artery to become gradually obliterated such that collateral vessels develop. Several different constrictor sizes were available at the time of surgery (i.d. 2.25-3.25 mm) in order to select one that matched the calibre of the LCx of each investigated animal, attempting thereby to ensure maximal closure of this artery during the next few weeks. The thorax was closed with interrupted sutures and air withdrawn from the thoracic cavity Postoperative care (including antibiotics) and pain management (buprenorphine 0.3 mg/kg, i.m.) were instituted thereafter.
Electrophysiologic Study
The final study was performed approximately 6 weeks after implantation of the coronary artery constrictor. After inducing anesthesia (meperidine 50 mg and thiopental 8 mg/kg i.v., followed by isoflurane 1%), pulmonary capillary wedge pressure and cardiac output measurements were repeated for comparison with initial values. Coronary angiography was performed under fluoroscopy via a catheter threaded through a carotid artery in order to assess the degree of closure of the LCx by the ameroid constrictor. The heart was exposed via a midline sternotomy and pericardial incision. The animals were kept under physiological solute IV perfusion. Respiratory rate was controlled to maintain blood gases stable within the physiological range. An octagonal silicone plaque electrode carrying 191 unipolar recording contacts (spaced 2.6 μm apart, each contact 300 pm in diameter) was positioned onto the lateral free wall of the left ventricle in the territory supplied by arterial branches arising from the LCx distal to the site of arterial occlusion; the plaque was kept in place with a stitch at each corner. The surface area of the recording field was 15 cm2. The thoracotomy was then covered with gauze soaked in saline, and the chest was warmed with a heating lamp to maintain internal temperature constant at 37° C. as monitored with a thermometer probe. Arterial and LV pressures were monitored using a Millar catheter introduced across the aortic valve from a carotid artery. Epicardial unipolar electrograms and four limb leads were connected to a multi-channel recording system (EOI 12/256, Institut de genie biomedical, Universite de Montreal and Ecole Polytechnique de Montreal) that was controlled by custom-made software, using a PC-computer (Cardiomap III: ww.crhsc.umontreal.ca/cardiomap). Unipolar epicardial electrograms (measured with reference to Wilson's central terminal derived from limb leads) were amplified by programmable-gain analogue amplifiers (0.05-450 Hz), and converted to digital format at 1000 samples/s/channel. Data were stored on hard disk from which files were subsequently retrieved for detailed analysis.
The dorsal columns of the cranial, thoracic spinal cord were stimulated as described previously (Foreman et al., 2000; Armour et al., 2002). Animals were placed in the prone position and the epidural space of the mid-thoracic spinal column penetrated percutaneously with a Touhy needle, using fluoroscopy and loss-of-resistance technique. A quadrupolar electrode catheter (Medtronic QUAD Plus Model 3888; Medtronic, Minneapolis, Minn.) was introduced through this cannula and its tip advanced to the T2 level of the spinal column, slightly to the left of the midline. The rostral and caudal poles selected for subsequent use (interelectrode distance of 1.5 cm) were placed at the levels of the T2 and T4 vertebrae. These electrodes were connected to both poles of a Grass S88 stimulator via a constant current stimulus isolation unit (50 Hz, 0.2 ms duration). Then the threshold intensity for motor responses (proximal forepaw, shoulder muscle, thoracic trunk contractions) was determined. After fixing the electrodes in place, the animals were placed in the supine position and the motor threshold rechecked. Thereafter, spinal cord stimulus intensity was set at 90% of motor threshold (Foreman et al., 2000).
Atrioventricular block was induced by formaldehyde injection (37%, 0.1 ml) into AV node to control ventricular rate. Pacing at a cycle length of 400-500 ms was performed via a bipolar electrode catheter positioned at the right ventricular apex and connected to a constant current generator controlled by a programmable stimulator. The two cardiac stressors that were applied consisted of (i) rapid ventricular pacing and (ii) activation of right atrial neurons by locally administered angiotensin II. Rapid ventricular pacing was performed for 1 min at a cycle length of 250 ms (240 beats/per min). Angiotensin II dissolved in saline (100 μg/ml) was continuously administered to right atrial neurons over 1 min periods (0.4 ml/min) via a PE-50 cannula that was inserted into the right coronary arterial blood stream with its tip placed immediately proximal to the root of the artery that supplied blood to the right atrial ganglionated plexus. This permitted angiotensin II (40 μg/min) to be administered into arterial blood perfusing right atrial neurons without compromising local coronary arterial blood flow (Levett et al., 1996).
Each animal was subjected to one of the two protocols, each comprising four phases: a conditioning phase (to account for differences between the first and subsequent pro-ischaemic trials) followed by three trials (Cardinal et al., 1981). Each phase lasted for 17 min, being separated from one another by 1-h recovery periods (
Data Analysis
The main variable of interest was the displacement of ST segment potential measured during the repolarization phase, 60 ms after the end of the activation complex of unipolar electrograms. ST segment elevation (positive voltage displacement, in mV) or depression (negative displacement) at each epicardial recording site was determined using a computer program with reference to the isoelectric lines identified during T-Q segments. Electrical alterations were identified in the territory supplied by LCx arterial branches downstream from the ameroid constrictor, in the presence or absence of the stressors. According to classical interpretation, the magnitude of the epicardial ST segment depression so identified primarily represents a measure of subendocardial ischaemia, whereas, regional ST segment elevation is considered to be indicative of underlying transmural ischaemia (Holland and Brooks, 1977; Kleber et al., 1978).
The responses to each stressor (difference between ST segment values under the stressor and basal state) were determined at maximum stressor effect. To analyze the effects of spinal cord stimulation in the experimental protocol, the differences in the responses to each stressor identified between trials 2 and 1 were plotted as a function of responses elicited in the absence of spinal cord stimulation (trial 1) at all recording sites in each protocol. These data were subjected to classical linear regression analysis (y=a+b+x) to determine the slope b of the relationship. A statistically significant linear relationship displaying a negative slope indicated that the magnitude of the responses elicited in trial 2 was reduced in comparison with to the ones elicited in trial 1. In contrast, a horizontal (“flat”) relationship displaying a slope value that was not significantly different from 0 indicated that the responses were similar between trials 1 and 2, i.e., a non-significant effect of spinal cord stimulation in the experimental protocol. The significance of the linear model was analyzed by testing the null hypothesis concerning the slope of the regression line from the statistic t=r/√(l−r)/(n−2), where r is the correlation coefficient and n is the number of data points (Glantz, 2002). Slopes measured in experiments performed under the experimental protocol were compared to the ones measured under the control protocol by Student's t-test.
Activation-recovery intervals measured from maximum slope of negative deflections in activation complexes −dVIdtmax to maximum positive slope of T wave in each unipolar electrogram (Derakhchan et al., 1998) were employed to estimate the responses of repolarization intervals to transient rapid pacing (Cardinal et al., 2004). Data are expressed as mean±S.D. Differences were considered as statistically significant when P<0.05.
All 17 preparations survived to study. The time from constrictor implantation to study was 42±16 days (n=17). Signs of pain were not observed in any of the preparations. Comparison between values measured in the anesthetized state at pre-implantation and at the time of study indicated that there was no significant effect on cardiac output (2.8±0.8 and 2.7±1.0 1/min, respectively), and pulmonary capillary wedge pressure (6.3±2.6 and 6.5±3.5 mm Hg). There was no gross evidence of segmental myocardial scarring at post-mortem. Thus, creation of a collateral-dependant myocardial substrate in the LCx territory with ameroid constrictor closure did not induce hemodynamic deficit in the resting (anesthetized) state, but regional ischaemic changes were induced in response to the stressors (transient rapid pacing, intracoronary angiotensin II).
ST Segment Responses to Transient Rapid Ventricular Pacing
ST Segment Responses to Intracoronary Angiotensin
Data presented regarding responses to intracoronary angiotensin II injection were obtained from five preparations studied under the control protocol, and 8/12 preparations studied under the experimental protocol (minimal responses to angiotensin II of ±2 mV being induced in the other four). Angiotensin II (40 mg over 1 min) delivered to right atrial neurons via their arterial supply in the same experiment illustrated in
Effects of Spinal Cord Stimulation on ST Segment Responses to Intracoronary Angiotensin II Administration and Transient Rapid Pacing
ST segment changes induced in response to angiotensin II (40 μg over 1 min) delivered to right atrial neurons were obtunded in the presence of dorsal spinal cord stimulation (
To investigate the dependence of modulation by spinal cord stimulation (trial 2) on the magnitude of ST segment changes induced by the stressor alone (trial 1), the differences in ST segment values recorded in trials 1 and 2 were plotted as a function of the magnitude of the responses to the stressors induced in trial 1. For instance, data obtained in response to intracoronary angiotensin II administration in the animal illustrated in
In contrast to ST segment responses to angiotensin II administration, the responses to rapid ventricular pacing were unchanged in the presence of spinal cord stimulation (
Responses of Ventricular Repolarization Intervals to Transient Bouts of Rapid Ventricular Pacing Effects of Spinal Cord Stimulation
Another feature of preparations with chronically implanted ameroid constrictors consists of localized shortening of repolarization intervals in response to transient rapid pacing (250 ms), which is superimposed on rate-dependent shortening that occurs uniformly in healthy preparations (Cardinal et al., 2003). This phenomenon is seen as a region with activation-recovery intervals <150 ms (minimum of 141 ms) in the LCx territory of the preparation illustrated in
A novel observation reported herein was that ST segment changes, in the form of augmented ST elevations, were evident at a number of sites in the central areas of the compromised LCx territory in response to activating populations of intrinsic cardiac neurons by locally administered angiotensin II. Angiotensin II is a potent activator of intrinsic cardiac neurons and its neuronal administration has been proposed as an approach to sustain the circulation in the postoperative setting (Levett et al., 1996). Moreover, angiotensin II may be of pathophysiological significance in ventricular dysfunction since this substance has been reported to cause neuronal catecholamine release into the cardiac interstitium in canine preparations of experimental, mitral valve regurgitation (Tallaj et al., 2003). A second finding was that such neuronally induced responses could be attenuated by spinal cord stimulation.
The chronic ameroid preparations were also found to be appropriate to investigate ST segment changes (augmented ST elevation or depression), as well as repolarization changes, in response to transient bouts of rapid ventricular pacing. However, the data reported herein indicate that responses to an extraneously imposed tachycardic stress were not affected by spinal cord stimulation.
Responses to Transient Bouts of Rapid Pacing
As reported in a separate study (Cardinal et al., 2004), unipolar electrograms generated at the most affected sites of the occluded LCx territory were typical of the electrical activity generated by populations of electrically-coupled viable cells displaying depressed action potentials (Holland and Brooks, 1977) of the type previously described in the acutely ischaemic myocardium (Kleber et al., 1978). Epicardial QS waves typical of transmural or extensive subendocardial necrosis (Durrer et al., 1964) were never detected in accordance with the fact that confluent necrosis was not detected in ameroid preparations, as has been previously reported (Katagiri et al., 1978).
ST segment changes and local repolarization interval shortening that occurred in response to transient bouts of rapid ventricular pacing, may have resulted from increased myocardial metabolic demand and/or blood flow deficit to subendocardial muscle (Macho et al., 1981). ST segment displacement recorded via an epicardial electrode is generated by an injury current vector that affects an overlying electrode in direct proportion to the difference in regional voltages, and the solid angle subtended at that site with respect to the ischaemic borders (Holland and Brooks, 1977). As such, ST segment depression recorded from an epicardial unipolar electrogram is indicative of subendocardial ischemia. On the other hand, epicardial ST segment elevation is indicative of transmural ischaemia. Activation of ATP-dependent potassium channels may be involved in shortening of repolarization intervals in the collateral-dependent myocardium as preliminary data indicate that this response to transient rapid pacing is inhibited by glibenclamide. Such metabolically determined alterations were unaffected by spinal cord stimulation.
Responses to Intracoronary Angiotensin II Administration
The rationale underlying the use of this stressor was based on the concept that angiotensin II activates intrinsic cardiac neurons (Horackova and Armour, 1997; Yin et al., 1999), thereby increasing the release of neurochemicals into the ventricular interstitium from cardiac adrenergic efferent postganglionic neurons (Farrell et al., 2001). The patterns of ST segment changes induced in response to administering angiotensin II to intrinsic cardiac neurons differed from patterns induced in the classical paradigm of tachycardic stress. ST segment displacements were augmented at sites displaying ST segment elevation; however, ST segment depression did not develop under this stressor. Note that these effects occurred independently of any positive chronotropic effect on sinus node activity since AV block had been induced and ventricular pacing was performed at a fixed rate. We have previously shown that intrinsic cardiac neuronal activity is enhanced (Levett et al., 1996) and that interstitial fluid catecholamine levels increase in response to this stressor (Farrell et al., 2001). To our knowledge, this is the first report concerning the fact that intrinsic cardiac neurons can, when activated, modify electrical events in ischaemic myocardial tissues. An interesting finding of this investigation was that the magnitude of the dorsal cord stimulation effect was proportional to the magnitude of the angiotensin II response elicited, as evidenced by a negative linear relationship between the two. These data indicate that the least compromised regions were the ones least likely to respond to that stressor and, as such, to be least likely to respond to neuronal modulation.
Limitations
Any experimental protocol using repeated pro-ischaemic trials (here: transient rapid pacing and intracoronary angiotensin II) to investigate anti-ischaemic interventions by comparing trials performed with or without an experimental intervention (here: spinal cord stimulation) is subject to the limitations imposed by variability in repeated trials. In particular, it is notorious that the first trial may be different from subsequent ones (Cardinal et al., 1981). Therefore, the first application of stressors was considered as a “conditioning” trial and the next trial-designated “trial 1” was used as the reference against which to compare data obtained in the test trial (trial 2 performed in the presence of spinal cord stimulation). Moreover, a control protocol was used in which trial 2 was performed without spinal cord stimulation to control for possible differences between trials 1 and 2. The attenuating effect of spinal cord stimulation on ventricular electrical responses to angiotensin IT were thus found to be greater than a trend for attenuation existing between trials 1 and 2 in the absence of spinal cord stimulation.
Perspective
Dorsal column activation was found to attenuate the deleterious effects that stressors exerted on ischaemic myocardial electrical events, particularly those associated with the stress that chemically activated intrinsic cardiac neurons induce. ST segment changes induced in collateral-dependent myocardium were reduced, an effect consistent with “stabilization” of intrinsic cardiac neuronal activity during spinal cord stimulation (Foreman et al., 2000; Armour et al., 2002). This occurs without inducing detectable changes in regional ventricular dynamics or blood flow (Kingma et al., 2001). That spinal cord stimulation can modify electrical events in collateral-dependent myocardium of chronic coronary artery obstruction preparations independent of alterations in cardiac dynamics has implications with regard to its efficacy in jeopardized cardiac states.
Recent clinical evidence indicates that delivering high frequency electrical stimuli to the dorsal aspect of the cranial thoracic spinal cord (SACS) can alleviate symptoms associated with chronic refractory angina pectoris (Elias son et al., 1996; Foreman et al., 2004; Linderoth and Foreman, 1999; Mannheimer et al., 1993). The physiological mechanisms underlying the beneficial effects that such therapy imparts remain poorly understood. One mechanism of action of SACS may reside in its capacity to attenuate increased inputs to the intrinsic cardiac nervous system arising from the alchemic myocardium (Armour et al., 2002; Foreman et al., 2000).
That select mediastinal nerve ablation can suppress paroxysmal atrial fibrillation (Pappone et al., 2004) is in agreement with the fact that atrial arrhythmias can be initiated as a consequence of increased inputs to the intrinsic cardiac nervous system from more centrally located neurons (Armour, 2004). In anesthetised canines, increasing extrinsic neuronal inputs to the intrinsic cardiac nervous system in normally perfused hearts can initiate atrial arrhythmias without the need for concomitant programmed electrical stimulation of atrial muscle (Amour et al., 1972 and 1975; Hagemen et al., 1973); so can excessive cervical vagosympathetic nerve inputs (Scarifov et al., 2004). Focal ventricular ischemia is known to enhance sensory inputs to the intrinsic cardiac nervous system such that its neurons may become excessively activated (Arora et al., 2004). Excessive activation of the intrinsic cardiac nervous system attending its transduction of regional ventricular ischemia, for instance, can be overcome by SCS (Armour et al., 2002; Foreman et al., 2002). Presumably that is why SCS can modulate regional electrical alterations associated with ventricular stress (Cardinal et al., 2004).
Since the capacity of SCS to impart cardioprotection has been proposed to reside in part with its capacity to influence the intrinsic cardiac nervous system (Armour et al., 2002; Foreman et al., 2000), we sought to determine whether this form of therapy can modify the capacity of excessive extracardiac neuronal inputs to the final common integrator of cardiac control—the intrinsic cardiac nervous system—to initiate atrial arrhythmias. Specifically, the main focus of this study was to determine whether SCS therapy suppresses neuronally induced atrial tachyarrhythmia formation.
Animal preparation. These experiments were performed in accordance with guidelines specified by the American Physiological Society for animal experimentation (Guiding principles for research involving animals and human beings. Am J Physiol Regul Integr Comp Physiol. 2002; 283: R281-83) and approved by the institutional animal care committee. Ten adult mongrel dogs of either sex, weighing between 15-27 kg, were employed in this study. These animals were anesthetised with sodium thiopental (25 mg/kg iv supplemented every 30 minutes or less as required) and then intubated so that respiration could be maintained with positive-pressure ventilation. Following a trans-thoracic incision, the pericardium was incised to expose the heart. Left ventricular and aortic pressures (Millar electronic pressure sensors) and a lead II EGG were recorded on a rectilinear pen recorder (Nihon Cohen, Tokyo, Japan). Atrioventricular blockade was induced by injecting formaldehyde (37%; 0.1 ml) into the AV node in order to be able to isolate atrial from ventricular electrical events. Right ventricular pacing (60 beats/min) was instituted thereafter to maintain adequate cardiac output. After the surgery had been completed, the anesthetic agent was changed to α-choralose (25-50 mg/kg iv bolus, supplemented with 25 mg/kg iv doses as required).
Atrial epicardial mapping. Multiple silicone plaques containing 191 unipolar recording contacts (4.6-5.9 mm spacing) were positioned on the ventral, lateral and dorsal surfaces of the right and left atrium, as described previously (Hélie et al., 2000). These unipolar leads and lead II ECG were connected to a multi-channel recording system (EDI 12/256, Institut de genie biomedical, & École Polytechnique de Montreal) that was controlled by custom-made software (Cardiomap III: www.crhsc.umontreal.ca/cardiomap) using a PC-computer. Unipolar electrograms (measured with reference to Wilson's central terminal derived from the 4 limb leads) were amplified by programmable-gain analog amplifiers (0.05-450 Hz) and converted to digital format at 1000 samples/s/channel. Data were stored on hard disk from which files were subsequently retrieved for detailed analysis.
Electrical stimulation of mediastinal nerves. Right-sided mediastinal nerves on i) the caudal portion of the superior vena cava (intrapericardial sites) and ii) the first 1-2 cm of the superior vena cava craniad to its pericardial reflection (extra-pericardial sites) (Amour et al., 1975) were studied. Their accompanying vessels render these mediastinal nerves identifiable. Active sites were identified functionally such that, when stimulated electrically, they induced atrial arrhythmias in these otherwise normal preparations. Once so identified, each locus was marked with ink for repeat stimulation. In that manner, electrical stimuli could be applied to neural elements located in i) on the first 1-2 cm of the superior vena cava craniad to the pericardial reflection (extra-pericardial nerves) and ii) the ventral and lateral surfaces of the intra-pericardial superior vena cava, including sites just craniad to the right atrial ganglionated plexus (intra-pericardial nerves).
Electrical stimuli were delivered to these select mediastinal nerve sites via a bipolar electrode (1-5 mm inter-electrode distance) mounted on a roving probe that was connected to a battery-driven current source controlled by a programmable stimulator (Bloom Associates, Philadelphia, Pa.). Trains of 4 electrical stimuli (1-2 mA; 1 ms duration; 5 ms pulse intervals) were delivered to these mediastinal nerves during the atrial refractory period of adjacent atrial tissues (˜30 ms after excitation of a reference electrogram) for up to 20 seconds (when no arrhythmias were induced). This stimulus protocol was used in order to avoid atrial muscle capture. In control states, the stimulation periods required to initiate atrial arrhythmias usually were 3 seconds or less. As longer periods of simulation resulted in prolonged periods of atrial fibrillation in control states, stimulus periods were restricted such that this situation did not occur. Identified sites were briefly re-stimulated two or more times to confirm reproducibility of results. About 5 minutes of recovery period was allowed to elapse between site stimulations.
Implantation of spinal cord stimulating electrodes. After placing the animal in the prone position, the epidural space was entered with a Touhy needle via a small skin incision in the lower, dorsal thorax. A four-pole lead (Medtronic QUAD Plus Model 3888; Medtronic Inc., Minneapolis, Minn.) was advanced rostrally in the epidural space to the cranial thoracic level under anterior-posterior fluoroscopy. The tip of the lead was positioned slightly to the left of midline, according to current clinical practice (Manheimer et al., 1993). The most cranial pole of the lead was positioned at the T1 level (Foreman et al., 2000). Electrical current was delivered via its rostral and caudal poles to verify their functional position. Increasing stimulus intensity via the rostral pole as cathode to motor threshold intensity (MT) induced muscle contractions in the proximal forepaw and shoulder. Stimulation with the caudal pole as cathode at MT activated thoracic paravertebral muscles, resulting in a twisting movement of the trunk. After a satisfactory electrode position was obtained the lead, protected by a silicon sleeve, was fixed to the interspinous ligament and then connected via a stimulus isolation unit and a constant current generator (Grass SIU 5B) attached to a Grass model S48 stimulator (Grass Instruments, Quincy, Mass., U.S.A.).
Spinal cord stimulation threshold determination. After the animals were placed in the decubitus position, electrical stimuli (50 Hz and 0.2 msec duration) were delivered to the dorsal aspect of the thoracic spinal cord via the implanted electrodes such that MT could be confirmed by ascertaining that stimulation intensity was 90% of that evoking a motor response; this corresponds to the maximum intensity used in patients (Manheim et al, 1993). The current intensity used for 90% of MT varied between 0.21 and 0.86 mA (mean 0.39 mA) among animals. The most rostral and caudal poles were chosen as cathode and anode, respectively, so that the spinal cord region used for angina therapy in humans could be stimulated.
Intenention protocol. Firstly, 8-10 neural element sites were identified in each animal that, when subjected to focal electrical stimuli, consistently initiated atrial tachyamhythmias. No responses were elicited from most of the intrapericardial or extrapericardial sites tested. Active sites were identified with an ink dot so that each could be studied repeatedly. Repeat stimulation of identified loci initiated similar atrial electrical responses. At least 5 minutes was allowed to elapse between site stimulations. This function-anatomic identification of right-sided mediastinal nerves permitted the study of mediastinal neural elements in cranial and caudal locations in each animal with precision.
After identifying active sites from which repeated focal electrical stimuli induced similar atrial electrical responses, SCS was applied for 20 minutes. After waiting 15 minutes following discontinuing SCS, identified sites were re-stimulated at least two times with about 5 minutes of recovery period being allowed to elapse between site stimulations to permit preparation stabilization. In 2 animals, hexamethonium bromide (0.1 mg/kg i.v.) was then administered and cranial and caudal mediastinal nerve sites were re-stimulated thereafter.
Data analysis. Neurally induced atrial tachyarrhythmias were grouped according to whether they could be elicited or not in each protocol. Atrial electrical activation times were also identified as the moment in the activation complex of unipolar electro grams at which the slope of the negative potential displacement (−dV/dtmax) was maximal (Lia and Nattel, 1997), displaying QS morphology at the site of origin and rS morphology at sites of later activation. Thereafter, isochronal maps (10-ms interval) were computed automatically by linear interpolation.
Mediastinal nerve stimulation prior to SCS. During basal states (i.e., during sinus rhythm) when electrical stimuli were applied to select right-sided extrapericardial or intrapericardial mediastinal nerves for less than 5 seconds, a typical sequence of atrial electrical events occurred in all animals. This response consisted of a brief period of bradycardia followed by the induction of atrial tachycardia (
On average, 8 cranial and caudal mediastinal nerve sites were identified in each animal which, when stimulated focally, initiated atrial arrhythmias. Repeat stimulation of each site elicited similar atrial arrhythmias. When the trains of electrical stimuli were delivered during the atrial refractory period to other loci on the superior vena cava (non-active sites) or directly onto the right atrial epicardium, atrial capture did not occur; nor did atrial rate change.
Mediastinal nerve stimulation post-SCS. During or after the 20 minutes of SCS, no change in atrial electrical events was identified. Following SCS, no alteration in atrial electrical events was recorded when focal electrical stimuli were applied to previously identified extrapericardial sites that had initiated atrial tachyarrhythmias before SCS application. After SCS, atrial arrhythmias were no longer induced when cranial intra-pericardial nerve sites were exposed to burst stimuli, even when they were delivered for longer period of time (
Atrial arrhythmias were initiated post-SCS when the most caudal mediastinal nerve sites were re-stimulated (
The results of the present experiments demonstrate that electrical activation of the dorsal aspect of the thoracic spinal cord modifies the capacity of excessively activated intrinsic cardiac neurons to induce atrial arrhythmias. That enhancement of spinal cord neuronal inputs to the intrinsic cardiac nervous system obtunds its capacity to initiate atrial tachydysrhythmias is in accord with the fact that spinal cord therapy is known to stabilize neuronally induced electrical changes in the ischemic, stressed ventricle (Cardinal et al., 2004).
Each of the limited number of atrial fibrillatory nerves within the mediastinum contains sympathetic and parasympathetic efferent axons that innervate the heart (Brandys et al., 1986). When electrical stimuli are applied to these mediastinal nerves as they course adjacent to the heart, both pre- and postganglionic cholinergic and adrenergic efferent neuronal element are activated (Amour et al., 1975; Butler et al., 1988). Thus, depending on the cranial or caudal location of the mediastinal nerve site investigated the effects of either population dominate (Brandys et al., 1986).
When repeatedly stimulated select mediastinal nerves, the atrial fibrillatory nerves, initiate atrial tachydysrhythmias (Armour et al., 1972 and 1975; Hageman et al., 1973). It is classical knowledge that cholinergic efferent neuronal inputs induce bradycardia, while concomitantly reducing the atrial effective refractory period in a spatially heterogeneous manner (Alessi et al., 1958; Liu and Natel, 1997). The more slowly developing adrenergic influences elicited during concurrent activation of adrenergic and cholinergic efferent neurons provides a substrate for atrial arrhythmia formation (Sharifov et al., 2004). The functional anatomy of extrapericardial and cranial intrapericardial mediastinal nerves presumably accounts for the fact that different cardiac effects accrue when cranial versus caudal components of intrapericardial mediastinal nerves were activated. For instance, the more cranial sites contain primarily efferent preganglionic axons that synapse with neurons located in the more caudally located intrinsic cardiac ganglionated plexuses while their caudal counterpart contain mostly, but not exclusively, efferent postganglionic axons (Brandys et al., 1986; Butler et al., 1980). This anatomy was confirmed by the fact that hexamethonium administration eliminated some of the atrial responses elicited when caudal sites were tested.
This functional anatomy is relevant considering the fact that SCS modified all atrial arrhythmias induce by activating caudal mediastinal nerve sites. As preganglionic efferent axons predominated at such sites, synapses located within the more caudally located intrinsic cardiac ganglia could have been affected by enhanced spinal cord inputs. Presumably, such would not be the case for the more distal sites that primarily contain autonomic efferent postganglionic axons (Brandys et al., 1986; Butler et al., 1990). In other words, it is unlikely that spinal cord inputs could have directly modulated autonomic efferent postganglionic nerves projecting to cardiomyocytes, explaining why the most caudal nerve sites continued to initiate atrial arrhythmias post-SCS (
Perspectives. That the genesis of neuronally induced atrial arrhythmias can be influenced by inputs from the thoracic spinal cord has implications for the suppression of such untoward events. This may have clinical relevance given the fact that excessive activation of select populations of intrinsic cardiac neurons in such a state is known to induce cardiac arrhythmias (Huang et al., 1993). These data are in accord with the observation that SCS suppresses the capacity of the intrinsic cardiac nervous system to transduce myocardial ischemia (Foreman et al., 2000). These data are also in accord with the fact that spinal cord inputs can stabilize the intrinsic cardiac nervous system when excessively activated by inputs from an ischemic ventricle (Cardinal et al., 2003). These data delineate the intrinsic cardiac nervous system as a target for atrial anti-arrhythmia therapy.
Spinal Cord Stimulation Suppresses Bradycardias and Atrial Tachyarrhythmias Induced by Mediastinal Nerve Stimulation in Dogs
Delivering high-frequency, low-intensity electrical stimuli to the dorsal aspect of the cranial thoracic spinal cord can alleviate symptoms associated with chronic refractory angina of cardiac origin (Eliasson et al., 1996; Foreman et al., 2004; Mannheimer et al., 1996). A primary mechanism of such spinal cord stimulation (SCS) therapy is due to its capacity to modulate the intrinsic cardiac nervous system in the presence of excessive afferent inputs arising from the ischemic ventricle (Armour et al., 2002; Foreman et al., 2000).
Increasing extrinsic neuronal inputs to the intrinsic cardiac nervous system can initiate self-terminating episodes of atrial tachyarrhythmia/fibrillation in intact hearts without the need for concomitant programmed electrical stimulation of atrial muscle (Armour et al., 1975; Arpir et al. 2005; Scherlag et al., 2002; Sharifov et al., 2004). Atrial tachyarrhythmias can be induced experimentally by electrical stimulation of mediastinal nerves during the refractory period of atrial muscle to avoid local muscle capture (Armour et al., 2005; Scherlag et al., 2002). The initial beats of the tachyarrhythmias thus elicited appear to be of focal origin (Armour et al., 2005). It is possible that this may account, in part, for the fact that select mediastinal nerve ablation can suppress paroxysmal atrial fibrillation (Pappone et al., 2004).
A primary mechanism of such SCS therapy resides in its capacity to modulate the intrinsic cardiac nervous system in the presence of excessive afferent neuronal inputs arising from the ischemic ventricle (Armour, et al., 2002; Cardinal et al., 2004; Foreman et al., 2000; Issa et al., 2005). In this study, we investigated the possibility that SCS therapy might suppress neuronally induced atrial tachyarrhythmia formation via its capacity to stabilize excessive inputs to the intrinsic cardiac nervous system.
Animals
A total of 21 adult mongrel dogs (either sex), weighing 15-27 kg, were used in this study. Experiments were performed in accordance with guidelines for animal experimentation (World Medical Association, American Physiological Society, 2002) and approved by the institutional animal care committee. Animals were anesthetized with sodium thiopental (25 mg/kg iv, supplemented as required), intubated, and maintained under positive-pressure ventilation.
Implantation of Spinal Cord-Stimulating Electrodes
With animals lying in the prone position, a small incision was made in the caudal, dorsal thoracic skin. The epidural space of the midthoracic spinal column was penetrated percutaneously with a Touhy needle, using ventral-dorsal positional fluoroscopy and loss-of-resistance technique (Eliasson et al., 1996). A quadrupolar electrode catheter with 1.5 cm interelectrode distance (QUAD Plus Model 3888, Medtronic, Minneapolis, Minn.) was introduced, and its tip was advanced under fluoroscopy to the T1 level of the dorsal surface of the spinal column, slightly to the left of the midline. The rostral pole (cathode) and caudal pole (anode) selected for subsequent use were connected to a constant current generator (WPI model #A385, 50 Hz, 0.2-ms duration) controlled by a stimulator (model S88, Grass Instruments, Quincy, Mass.). Threshold intensity for motor responses (proximal forepaw and shoulder muscle contraction) was determined. After a satisfactory electrode position was obtained, the external portion of that electrode catheter was covered with a protective silicone sleeve that was fixed to the interspinous ligament. Thereafter, the animals were placed in the supine position, and the motor threshold was rechecked. Spinal cord stimulus intensity was set at 90% of motor threshold (Foreman et al., 2000), varying between 0.21 and 0.86 mA (mean 0.39 mA) among animals.
Instrumentation for Electrophysiological Study
Left ventricular chamber and aortic pressures (Millar electronic pressure sensors) along with a lead II ECG were recorded on a rectilinear pen recorder (Nihon Kohden, Tokyo, Japan). With the animal remaining in the supine position, a sternotomy was performed, the pericardium was incised, and the heart was exposed. Atrioventricular (AV) block was induced by formaldehyde injection (37%, 0.1 ml) into the AV node to separate atrial from ventricular electrical events. Right ventricular pacing (50-60 beats/min) was instituted to assure adequate cardiac output. After this surgery, the anesthetic agent was changed to α-chloralose (50 mg/kg iv bolus, supplemented with 25 mg/kg doses as required). Silicone plaques carrying 191 unipolar recording contacts (4.6-5.9 mm spacing) were positioned on the ventral, lateral, and dorsal surfaces of the right and left atria and attached with sutures (Armour et al., 2005). The unipolar leads and lead II ECG were connected to a multichannel recording system (EDI 12/256, Institut de Ge'nie Biome'dical, Ecole Polytechnique de Montreal) controlled by custom-made software (Cardiomap III, www.crhsc.umontreal.ca/cardiomap) using a PC computer. Unipolar electrograms (measured with reference to Wilson's central terminal derived from the four limb leads) were amplified by programmable-gain analog amplifiers (0.05-450 Hz) and converted to digital format at 1,000 samples/s/channel. Data were stored on hard disk, from which files were subsequently retrieved for detailed analysis.
Electrical Stimulation of Mediastinal Nerves
In all of the experiments, the right-sided mediastinal nerves studied were located on the ventral and ventrolateral surfaces of the superior vena cava, either in its caudal portion (intrapericardial sites) or in the first centimeter of the superior vena cava craniad to the pericardial reflection (extrapericardial sites). Individual nerve branches coursing over the superior vena cava arise from the intrathoracic right vagosympathetic nerve complex, some of which can be identified by their accompanying blood vessels (Armour et al., 1975; Armour et al., 2005). In five experiments (conducted according to protocol A; see below), electrical stimuli were also delivered to left-sided mediastinal nerves that arise from the left vagosympathetic nerve complex, coursing intrapericardially cranial to the pulmonary veins (Armour et al., 1975).
A train of five electrical stimuli (1-2 mA, 1-ms duration, 5-ms pulse interval) was delivered to selected mediastinal nerve sites once per cycle of spontaneous atrial activity during the refractory period of adjacent atrial tissues (˜30 ms after excitation of a reference electrogram) to avoid muscle capture. Active sites were identified functionally such that, when exposed to focal electrical stimuli, changes in atrial rhythm were elicited (Armour, et al., 2005). The most frequently identified response consisted of bradycardia followed by an episode of spontaneously terminating atrial tachyarrhythmia/fibrillation with an average cycle length of ˜130 ms (Armour et al., 2005). No response was elicited from other intrapericardial or extrapericardial sites tested, despite periods of stimulation lasting for up to 20 s. Each active locus was marked with ink for repeat stimulation. Electrical stimuli were delivered focally via bipolar electrodes (1.5-mm spacing) mounted on a probe that was connected to a battery-driven current source controlled by a programmable stimulator (Bloom Associates, Philadelphia, Pa.). Contact between the bipolar electrodes and tissue was interrupted immediately after the onset of an episode of atrial tachyarrhythmia.
Protocol A: Effects of Spinal Cord Stimulation (11 Animals)
This protocol was performed in four steps. 1) In control states, multiple (six or more) active neural sites were identified and then restimulated to confirm reproducibility of responses (5-min recovery period between successive stimulations). 2) After the identification of active sites, SCS was applied for 20 min (Armour et al., 2002). 3) Thirty minutes after discontinuing SCS (Armour et al., 2002; Cardinal et al., 2004), the previously identified active sites were restimulated twice. 4) When sites were identified at which atrial tachydysrhythmias could still be initiated post-SCS, atropine (0.1 mg/kg iv) was administered and those mediastinal nerve sites were restimulated a final time.
Protocol B: Reproducibility
In five other animals, a protocol similar to protocol A was used but without SCS stimulation in step 2. Protocol B allowed us to verify that, in the absence of SCS, the functional responses were reproducible in a similar time frame.
Protocol C: Effects of Spinal Cord Stimulation in Animals with Bilateral Stellectomy
In five other animals, the right and left stellate ganglia were extirpated before determining the responses to electrical stimulation of mediastinal nerves in control states and after SCS (as in protocol A).
Data Analysis
Analyzing unipolar electrograms with Cardiomap III software, we identified activation times as the moment in the atrial activation complex at which the slope of the negative potential displacement (−dV/dtmax) was maximal (Armour et al., 2005). During basal states (i.e., sinus rhythm), the atrial cycle length determinations were derived from 10 atrial cycles. With respect to bradycardia or sinus tachycardia elicited by mediastinal nerve stimulation, atrial cycle length was assessed by means of the maximum interval recorded during two consecutive atrial electrograms. The reference electrogram for these determinations was recorded from the right atrial free wall. Threshold for cycle length change in response to mediastinal nerve stimulation was set at >2% variation. The duration of the episodes of tachyarrhythmia/fibrillation was calculated as the interval between the atrial premature depolarization initiating the arrhythmia and the last arrhythmia beat. Atrial cycle lengths recorded during neuronally induced atrial tachyarrhythmias were averaged over 50-s periods when the arrhythmias lasted for 50 s or more. When shorter-duration arrhythmias were elicited, that index was averaged from data collected throughout the arrhythmia period. The earliest 10-ms epicardial breakthrough area was determined from isochronal maps (10-ms interval) that were computed automatically by linear interpolation from activation times at all 191 electrode sites of the biatrial silicon plaque electrodes.
For each protocol, Student's paired t-test was used for comparisons of continuous variables measured in the responses to electrical stimulation at a given nerve site in control states (step 1) vs. post-SCS or sham (step 3). The numbers of sites from which atrial arrhythmias were elicited when individual loci were stimulated electrically before and after SCS application were compared using the x2-test. The level of certainty for rejecting the null hypothesis was P<0.05. Data are presented as means±SD.
Protocol A: Effects of Spinal Cord Stimulation
Mediastinal nerve stimulation in control states. In 11 anesthetized animals, focal electrical stimuli were delivered to multiple sites on the superior vena cava and, in five of them, cranial to the ventral left atrium during basal states (i.e., sinus rhythm). A total of 86 “active” sites were identified that, when subjected to focal electrical stimulation, elicited bradycardia alone (12 sites: 11 right-sided, 1 left-sided), bradycardia followed by atrial tachyarrhythmia/fibrillation (50 sites: 47 right-sided, 3 left-sided), tachyarrhythmia/fibrillation without preceding bradycardia (21 sites: 8 right-sided, 13 left-sided). Episodes of atrial tachyarrhythmia/fibrillation displayed average cycle lengths of 118±12 ms. Moderate “sinus” tachycardia occurred in response to stimulation of only three nerve sites, as a shortening of sinus cycle length from 392±41 to 306±55 ms (22±6% shortening). Among the total of 62 initial bradycardias thus elicited (with or without a subsequent tachyarrhythmia), there was a 30±24% prolongation in atrial cycle length (i.e., sinus cycle length changed from 416 42 to 535±83 ms; Table 9, protocol A). Similar atrial responses were elicited by repeat stimulation of any active site in control states.
As illustrated in
11 (6.5)
Mediastinal nerve stimulation after preemptive SCS. Baseline sinus cycle length increased by 5% after applying SCS (Table 9, protocol A: from 416±42 to 435±47 ms, P<0.006). After SCS, repeat electrical stimuli initiated tachyarrhythmia/fibrillation from only 29 of the 71 previously identified active sites (Table 10, protocol A). Bradycardias (alone or followed by tachyarrhythmia/fibrillation) were also elicited from significantly fewer sites after SCS (Table 9; protocol A: reduced from 62 to 47, P<0.03). The maximum cycle length prolongations elicited during the bradycardias induced after SCS (23±20%) were significantly reduced vs. the bradycardias elicited from these sites in control states (30±24%, P<0.006). Electrical stimulation applied to seven nerve sites caused slight sinus tachycardia during which cycle length shortened from 432±47 to 411±47 ms (this 5±4% shortening was significantly smaller than the 22±6% shortening seen in three such responses in control states).
Of the tachyarrhythmias still induced from 29 nerve sites by repeat electrical stimulation after SCS, 24 tachyarrhythmias were preceded by bradycardia, and 5 were elicited without a preceding bradycardia. The durations of the tachyarrhythmia/fibrillations were reduced in 16 episodes (20±10 to 7±45) and increased in 10 others (9±5 to 17±45). In the three remaining episodes, the tachyarrhythmia duration exceeded 100 s in control states but were reduced to an average of 41 s after SCS (since duration >100 s exceeded the mean duration by several SDs, the data from these three episodes were not included in the statistics presented in Table 10). Thus duration was not significantly affected overall (16±10 to 11±75; Table 10, protocol A). The latency from beginning mediastinal nerve stimulation to tachyarrhythmia onset was not affected by SCS (control: 1.7±1.9 5, post-SCS: 1.6±1.0 5).
Epicardial Mapping
Classically, the sites of earliest epicardial activation and areas of earliest 10-ms activation occurred in the superior portion of the right atrial free wall during all basal beats (“sinus” rhythm) and shifted toward the inferior right atrial regions during the bradycardias (not shown) (Armour et al., 2005; Page et al., 1995; Schuessler et al., 1986). As previously shown (Armour et al., 2005), the earliest 10-ms epicardial break-through areas during the initial tachyarrhythmia beats induced by right-sided mediastinal nerve stimulation were localized in the right atrial free wall, Bachmann bundle region, or adjacent base of the medial right atrial appendage (
Atropine administration. Any residual atrial arrhythmias that were elicited by mediastinal nerve stimulation following SCS (protocol A) were eliminated following subsequent atropine administration, even when identified sites were restimulated for periods of up to 20 s.
Protocol B: Reproducibility
The induction of the bradycardias, as well as the cycle length prolongation during bradycardia, were reproducible between control and repeat stimulation in protocol B (Table 9). Likewise, the incidence of atrial tachyarrhythmias/fibrillations was not significantly different between control and repeat stimulations in protocol B (Table 10). The tachyarrhythmia/fibrillation cycle length and duration were reproducible as well.
Protocol C: Acute Bilateral Stellectomy
The baseline sinus cycle length was significantly longer among the five animals of the stellectomy group (Table 9, protocol C: 529±71 ms) than in the animals with intact stellate ganglia (protocol A: 416±42 ms). The sinus cycle length was not affected by SCS in the animals with stellectomy (after SCS: 521±60 ms).
In the five animals of the stellectomy group, 31 active nerve sites were identified which, when stimulated electrically in control states, elicited bradycardia alone (five nerve sites), bradycardia followed by atrial tachyarrhythmia/fibrillation (22 sites), tachyarrhythmia/fibrillation without preceding bradycardia (1 site), or a moderate acceleration of sinus rate (3 sites). The induction of bradycardias in response to nerve stimulation (with or without subsequent tachyarrhythmias) and the attendant cycle length prolongation were unaffected after SCS (Table 9, protocol C: control: 34±23%, post-SCS: 39±26%). Contrasting with protocol A, in animals with bilateral stellectomy, SCS did not alter the ability of mediastinal nerve stimulation (when applied to 23 nerve sites) to evoke atrial tachyarrhythmia/fibrillation. Furthermore, tachyarrhythmia durations were similar between control states and after SCS (Table 10, protocol C).
The main finding reported herein is that thoracic spinal cord neurons, when stimulated electrically, can obtund the induction of atrial tachyarrhythmias/fibrillation associated with excessive, asymmetric activation of the intrinsic cardiac nervous system. Given that “active” sites were defined as the ones from which neuronally induced atrial rate responses (bradycardia and/or tachyarrhythmia/fibrillation) were elicited in control states, the proportion of total active sites from which tachyarrhythmias/fibrillation were elicited by mediastinal nerve stimulation was reduced from 71/86 in control states to 29/86 when repeat stimulation was applied to the previously identified nerve sites after SCS, i.e., a 60% reduction. The cycle length and duration of the residual tachyarrhythmias were not, overall, significantly affected after SCS. However, there was a tendency for shortening of duration, as seen in the majority of the episodes elicited after SCS (19/29). Removal from the statistics of data concerning the tachyarrhythmias with durations longer than 100 s in control state may have contributed to the conclusion that there were no changes in duration.
Interestingly, the initial tachyarrhythmia beats elicited after SCS displayed early epicardial breakthroughs that were localized in areas similar to those identified in control states. Taken together, the data indicate that, once elicited, the foci involved in the tachyarrhythmias induced after SCS were similar to the ones that were elicited in control states. Any residual atrial tachyarrhythmia evoked after SCS was eliminated by atropine, as previously shown in this model in the absence of SCS (Armour et al., 2005; Sharifov et al., 2004). It is classical knowledge that cholinergic efferent neuronal inputs to the atria reduce atrial refractory periods in a spatially heterogeneous manner (Alessi et al., 1958; Liu and Nattel, 1997), in addition to inducing bradycardia.
The proportion of total responsive sites from which bradycardias were elicited by mediastinal nerve stimulation was reduced from 62/86 in control states to 47/86 when repeat stimulation was applied to the previously identified nerve sites after SCS, that is, a 25% reduction. Moreover, there was a significant reduction in the magnitude of the bradycardias that were still elicited after SCS. Minor prolongation of baseline sinus cycle length also occurred after SCS, as reported by others (Issa et al., 2005). The reproducibility experiments support the notion that these differences were not related to changing conditions within the time frame of the protocol.
We thus conclude that the data reported herein support SCS effects in the form of 1) significant reductions in the inductions of both the tachyarrhythmia/fibrillations and the bradycardias elicited in response to electrical stimulation of mediastinal nerves, 2) significant reductions in the magnitude of neurally induced cycle length prolongation during the bradycardias. We were thus able to demonstrate an effect of SCS on the induction of the tachyarrhythmia/fibrillation episodes but without statistically significant change in their dynamics (cycle length) once elicited.
Linkage between enhanced spinal cord neuronal inputs and inhibition of neuronally induced atrial tachyarrhythmias was further demonstrated by the fact that the antiarrhythmic action of SCS did not occur after extirpation of both stellate ganglia before SCS. This is consistent with previous data indicating that the capacity of SCS to modulate the intrinsic cardiac nervous system, even when transducing excessive afferent inputs from the ischemic ventricle, occurs primarily via axons coursing in the subclavian ansae and stellate ganglia (Foreman et al., 2000).
It is also important to note that the blunting effect of SCS on the bradycardia responses to nerve stimulation was not elicited in the presence of bilateral stellectomy (Table 9). Thus the ability of SCS to obtund both the bradycardias and the neuronally induced atrial tachyarrhythmias can be eliminated by bilateral stellectomy. Because the parasympathetic efferent preganglionic inputs to the intrinsic cardiac nervous system remained intact in the stellectomy preparations, it is unlikely that such medullary neurons play a significant role in SCS-induced modulation of neuronally induced atrial arrhythmias. Our results indicate that neural inputs from the spinal cord modulate cardiac efferent postganglionic neurons, presumably via the effects of such inputs on intrinsic cardiac local circuit neurons (Armour et al., 2002).
That activated thoracic spinal cord neurons can modulate neuronally induced atrial tachyarrhythmias secondary to overloading the intrinsic cardiac nervous system has implications with respect to suppressing such untoward atrial events in a clinical setting. These data are in accord with the observation that enhanced spinal cord neuronal inputs to the intrinsic cardiac nervous system obtund the latter's capacity to transduce the ischemic myocardium (Foreman et al., 2000) and thereby stabilize neuronally induced ventricular electrical alterations in such a state (Cardinal et al., 2004; Issa et al., 2005).
They also support previous findings indicating that SCS neuromodulation therapy exerts its cardioprotective effects without compromising cardiac function, either at rest or during induced stress (Armour et al., 2002; Cardinal et al., 2004; Foreman et al., 2004; Foreman et al., 2000).
Perspectives
SCS obtunds the induction of atrial tachyarrhythmias resulting from excessive activation of intrinsic cardiac neurons, indicating that the intrinsic cardiac nervous system may be a target for SCS therapy in the management of atrial tachyarrhythmias. It is known that neuronally induced tachyarrhythmias can be acutely obtunded by local mediastinal neuronal ablation (Armour et al., 1975; Armour et al., 2005; Scherlag et al., 2002); however, that approach may represent temporary therapy due to the capacity of intrathoracic neurons to sprout neurites to innervate other intrinsic cardiac neurons and cardiomyocytes.
The presently claimed and disclosed invention encompasses the concept of a peripheral cardiac nervous system and the ability to stimulate this peripheral cardiac nervous system through the use of SCS. The stimulation of this peripheral cardiac nervous system results in the ability to easily and with minimal invasiveness, treat cardiac pathologies either pre-, during, or post-symptom.
The presently claimed and disclosed invention provides a method for protecting cardiac function and reducing cardiac malfunction. This methodology includes the steps of: (1) providing a stimulator capable of generating an electrical signal; (2) placing the stimulator adjacent a neural structure capable of carrying the electrical signal from the neural structure to the peripheral cardiac nervous system; and (3) activating the stimulator for a predetermined period of time to generate the electrical signal to protect cardiac function and reduce cardiac malfunction. In an alternate embodiment of this method, the neural structure is a spinal cord. In another embodiment of this methodology, the cardiac malfunction which is reduced is a ventricular malfunction. In yet another embodiment of this methodology, the ventricular malfunction is a ventricular arrhythmia. In another embodiment of this methodology, the cardiac malfunction which is reduced is an atrial malfunction. In yet another embodiment of this methodology, the atrial malfunction is an atrial arrhythmia. In yet another embodiment of this methodology, the atrial malfunction is an atrial fibrillation.
The presently claimed and disclosed invention further provides a method for treating a mammal having a cardiac pathology by protecting cardiac function and reducing cardiac malfunction. This methodology includes the steps of: (1) providing a stimulator capable of generating an electrical signal; (2) placing the stimulator adjacent a neural structure capable of carrying the electrical signal from the neural structure to at least one of the peripheral cardiac nervous system and the heart; and (3) activating the stimulator for a predetermined period of time to generate the electrical signal to modulate at least one of the peripheral cardiac nervous system and the heart, and thereby protecting at least one of the peripheral cardiac nervous system and the heart to treat the heart. In an alternate embodiment of this methodology, the neural structure is a spinal cord. In another embodiment of this methodology, the cardiac malfunction which is reduced is a ventricular malfunction. In yet another embodiment of this methodology, the ventricular malfunction is a ventricular arrhythmia. In another embodiment of this method, the cardiac malfunction which is reduced is an atrial malfunction. In yet another embodiment of this method, the atrial malfunction is an atrial arrhythmia. In yet another embodiment of this method, the atrial malfunction is an atrial fibrillation.
The presently claimed and disclosed invention also provides a method for electrically communicating with at least one of an peripheral cardiac nervous system and a heart. This methodology includes the steps of: (1) providing a stimulator capable of generating an electrical signal; (2) placing the stimulator adjacent a neural structure capable of carrying the electrical signal from the neural structure to at least one of the peripheral cardiac nervous system and the heart; and (3) activating the stimulator for a predetermined period of time to generate the electrical signal to communicate with at least one of the peripheral cardiac nervous system and the heart. In an alternate embodiment of this methodology, the neural structure is a spinal cord.
Additionally, the presently claimed and disclosed invention encompasses a method of modulating electrical neuronal and humoral responses of at least one of a peripheral cardiac nervous system and a heart. This methodology includes the steps of: (1) providing a stimulator capable of generating an electrical signal; (2) placing the stimulator adjacent a neural structure capable of carrying the electrical signal from the neural structure to at least one of the peripheral cardiac nervous system and the heart; and (3) activating the stimulator for a predetermined period of time to thereby generate the electrical signal to modulate the electrical neuronal and humoral response of at least one of the peripheral cardiac nervous system and the heart. In an alternate embodiment of this methodology, the neural structure is a spinal cord.
Furthermore, the presently claimed and disclosed invention also calls for a method of activating spinal cord neurons to induce a conformational change in an peripheral cardiac nervous system. This methodology includes the steps of: (1) providing a stimulator capable of generating an electrical signal; (2) placing the stimulator adjacent a spinal cord to carry the electrical signal from the spinal cord to a peripheral cardiac nervous system; and (3) activating the stimulator for a predetermined period of time to thereby generate the electrical signal to thereby activate spinal cord neurons in proximity of the stimulator so as to induce a conformational change in the peripheral cardiac nervous system.
The presently claimed and disclosed invention also provides for a method for the prolonged activation of spinal cord neurons to induce a conformational change in an peripheral cardiac nervous system. This methodology includes the steps of: (1) providing a stimulator capable of generating an electrical signal; (2) placing the stimulator adjacent a spinal cord to carry the electrical signal from the spinal cord to a peripheral cardiac nervous system; and (3) activating the stimulator for a predetermined period of time to thereby generate the electrical signal to thereby activate spinal cord neurons in proximity of the stimulator so as to induce a conformational change in the peripheral cardiac nervous system wherein the activation effects persist for a period of time extending beyond the activation of the stimulator.
Additionally, the presently claimed and disclosed invention includes a method for transiently nullifying neuronal activation of a peripheral cardiac nervous system caused by cardiac malfunction. This methodology includes the steps of: (1) providing a stimulator capable of generating an electrical signal; (2) placing the stimulator adjacent a neural structure capable of carrying the electrical signal from the neural structure to a peripheral cardiac nervous system; and (3) activating the stimulator for a predetermined period of time to thereby generate the electrical signal to transiently nullify neuronal activation of a peripheral cardiac nervous system caused by cardiac malfunction. In an alternate embodiment of this methodology, the neural structure is a spinal cord.
The presently claimed and disclosed invention also provides for a method for prolonged nullification of neuronal activation of a peripheral cardiac nervous system caused by cardiac malfunction. This methodology includes the steps of: (1) providing a stimulator capable of generating an electrical signal; (2) placing the stimulator adjacent a neural structure capable of carrying the electrical signal from the neural structure to the peripheral cardiac nervous system; and (3) activating the stimulator for a predetermined period of time to thereby generate the electrical signal to nullify the neuronal activation of the intrinsic cardiac nervous system caused by cardiac malfunction for a prolonged period of time extending beyond stimulator activation. In an alternate embodiment the neural structure is a spinal cord.
The presently claimed and disclosed invention also includes a method for transiently modulating neuronal activation of a peripheral cardiac nervous system caused by cardiac malfunction. This methodology includes the steps of: (1) providing a stimulator capable of generating an electrical signal; (2) placing the stimulator adjacent a neural structure capable of carrying the electrical signal from the neural structure to the peripheral cardiac nervous system; and (3) activating the stimulator for a predetermined period of time to thereby generate the electrical signal to transiently modulate the neuronal activation of the peripheral cardiac nervous system caused by cardiac malfunction. In an alternate embodiment of the present methodology, the neural structure is a spinal cord.
The presently claimed and disclosed invention includes a method for prolonged modulation of neuronal activation of a peripheral cardiac nervous system caused by cardiac malfunction. This methodology includes the steps of: (1) providing a stimulator capable of generating an electrical signal; (2) placing the stimulator adjacent a neural structure capable of carrying the electrical signal from the neural structure to the peripheral cardiac nervous system; and (3) activating the stimulator for a predetermined period of time to thereby generate the electrical signal to modulate the neuronal activation of the peripheral cardiac nervous system caused by cardiac malfunction for a prolonged period of time extending beyond stimulator activation. In an alternate embodiment of this methodology, the neural structure is a spinal cord.
The presently claimed and disclosed invention includes a method of inhibiting, treating, or reducing a cardiac malfunction. This methodology includes the steps of: (1) providing a stimultor capable of generating an electrical signal; (2) operatively associating the stimulator with a neural structure capable of stimulating a portion of the peripheral cardiac nervous system upon receiving the electrical signal from the stimulator; and (3) activating the stimulator thereby generating the electrical signal which stimulates the portion of the peripheral cardiac nervous system and inhibits, treats or reduces the cardiac malfunction. In an alternate embodiment of this methodology, the neural structure is a spinal cord. In another embodiment of this methodology, the cardiac malfunction which is reduced is an atrial malfunction. In another embodiment of this methodology, the atrial malfunction is an atrial arrhythmia. In yet another embodiment of this methodology, the atrial malfunction is an atrial fibrillation. In another embodiment of this methodology, the cardiac malfunction is a myocardial infarct induced by ventricular ischemia. In yet another embodiment of this methodology, the portion of the peripheral cardiac nervous system which is stimulated comprises adrenergic neurons. In yet another embodiment of this methodology, the stimulation of the adrenergic neurons acts via alpha-receptors. In another embodiment of this methodology, the cardiac malfunction is an atrial tachyarrhythmia or bradycardia. In yet another embodiment of this methodology, the treatment of cardiac malfunction is performed prior to conducting a subsequent procedure to treat the cardiac malfunction.
Thus it should be apparent that there has been provided in accordance with the present invention a detailed description, examples and data showing the SCS stimulation directly impacts the peripheral cardiac nervous system and that such an impact can be used to modify, treat, modulate, suppress, and/or quench the neuronal activity of the peripheral cardiac nervous system and in turn protect cardiac function and preserve the electrical stability of the peripheral cardiac nervous system and the heart itself, that fully satisfies the objectives and advantages set forth above. Although the invention has been described in conjunction with specific embodiments thereof, it is evident that many alternatives, modifications, and variations will be apparent to those skilled in the art. Accordingly, it is intended to embrace all such alternatives, modifications, and variations that fall within the spirit and broad scope of the appended claims.
The following references, to the extent that they provide exemplary procedural or other details supplementary to those set forth herein, are specifically incorporated herein by reference in their entirety as though set forth herein in particular.
This application is a continuation of U.S. Ser. No. 11/603,841, filed Nov. 21, 2006, which claims priority under 35 U.S.C. §119(e) of U.S. Provisional Application Ser. No. 60/738,641, filed on Nov. 21, 2005, entitled “ACTIVATION OF CARDIAC ALPHA RECEPTORS BY SPINAL CORD STIMULATION PRODUCES CARDIOPROTECTION AGAINST ISCHEMIA, ARRHYTHMIAS, AND HEART FAILURE.” The contents of each of the above-referenced applications are each hereby expressly incorporated in their entirety by reference.
| Number | Date | Country | |
|---|---|---|---|
| 60738641 | Nov 2005 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 11603841 | Nov 2006 | US |
| Child | 12925772 | US |
