INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE OCCURRENCE OF SOME FATAL arrhythmias has been attributed to cardiac autonomic imbalance and to a reduced cardiac parasympathetic autonomic activity in particular (2, 23). The cause of the autonomic imbalance is unknown, but it may involve dysfunction of the command neurons of the limbic system that maintain cardiac autonomic tone. Dysfunction of the command neurons also may trigger a sequence of events that can lead to acute myocardial infarction during extreme emotional and physical states in patients without preexisting cardiovascular pathology. Emotionally stressful conditions occurring immediately before the onset of symptoms are reported by about 4-18% of survivors of an acute myocardial infarction (10).

Already in 1937, Papez (14) had recognized that the limbic lobe forms a neural circuit that provides the anatomic substrate for emotions. Later the role of the limbic system in regulation of autonomic and baroreceptor reflex activity was recognized. Anatomic information about the locations in the forebrain of command neurons for cardiac autonomic tone maintenance is limited.

For characterization of the morphology of neuronal networks, recently a retrograde transneuronal viral labeling technique was introduced. We, and later others, have employed this method on heart innervation (15, 16, 20) to localize the cardiac vagal (CVM) and sympathetic motoneurons and higher-order heart activity-controlling neurons (20). The aim of the present investigation is to selectively localize forebrain parasympathetic command neurons in spinal cord-transected rats.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animals. Male Wistar rats of ~250 g bred by Harlan (Zeist, Netherlands) were used. Experimental protocols involving the application of pseudorabies virus (PRV) in the myocardium and spinal cord transection at thoracic segment T₁ were approved by the Groningen Medical School Animal Welfare Committee (FDC-93720/95982).

Inoculations. All rats were intubated and artificially ventilated (MK2 infant ventilator, Loosco; Amsterdam, Netherlands) with a mixture of halothane-nitrous oxide-oxygen. The Sauerbruch procedure was employed to allow access to the thoracic cage. In 40 rats, 2 µl PRV solution (3 × 10⁶ plaque-forming units/ml) were applied into the left ventricular myocardium (Fig. 1E) using a 30-gauge needle connected to a 10-µl Hamilton syringe. Intravenous virus (30 µl/animal) injections in healthy (n = 7) and spinal cord-transected rats (n = 5) were employed to check whether virus spillover in the blood gave rise to infections of the central nervous system in the regions that lack a blood-brain barrier. After the injection all wounds were sutured, and the animals were then detubated and placed in the home cage for recovery.

View larger version (123K): [in this window] [in a new window]   Fig. 1.   Photomicrographs illustrating viral infections of various regions of the rat brain after left ventricular inoculations of pseudorabies virus (PRV). A: group of PRV-labeled dysgranular insular cortical cells (cc; bar, 125 µm). B: virus-infected pyramidal cell of the anterior cingulate cortex (bar, 75 µm). C: overview of the midbrain periaqueductal gray matter (PAG) showing infections of neurons in dorsal and intermediate parts of the PAG and Edinger-Westphal nucleus (bar, 175 µm). D: virus-positive cells of the periambiguus area (bar, 200 µm). Amb, nucleus ambiguus; LRt, lateral reticular nucleus; PCRt, parvocellular reticular area; Gi, gigantocellular reticular area. E: horizontal cryostat section of left ventricular myocardium illustrating Evans blue-labeled sites of PRV injection (bar, 1 mm).

Spinal cord transection. In pilot experiments it was established that only a very small number of animals could survive spinal cord transection at C₈-T₁ immediately before or after viral infection of the ventricular myocardium. Therefore, the spinal cord transection was performed after overnight recovery, after viral injections in the heart. It has been established that spinal infection in the pig (the natural PRV host) is not detectable until 36 h after the inoculation (11). The upper segments of the thoracic spinal cord were exposed, and the dorsal parts of vertebrae C₇-T₂ were removed. Ligatures were placed round the spinal cord at C₈ and T₁ where, after the cord was transected between the ligatures, Spongostan (Ferrosan; Soeborg, Denmark) was placed in the wound to keep apart the ends of the cervical and thoracic cord. The pudendal nerves were then cut to produce incontinence (8). After closure of all wounds, the animals were housed in a temperature-controlled environment. Body temperature was maintained at 35-37°C. For preservation of fluid homeostasis, the rats received a subcutaneous injection of saline (1 ml/100 g body wt) every 12-14 h. Postinoculation survival times of 6 days were used. A large percentage of the spinal cord-transected animals died on the first or second day after this surgery, most likely of shock. In the survivors, no seizures or any other untoward adverse responses were observed.

Histology. The rats were anesthetized with an intraperitoneal injection of a 6% pentobarbital sodium solution and were then perfused transcardially with a 4% paraformaldehyde solution. After the perfusion the spinal cord was examined for incomplete transection, and it was complete in all surviving animals. The brain and spinal cord were removed and then postfixed and cryoprotected overnight. Serial 40-µm coronal sections of the brain were prepared on a cryostat microtome and collected in 0.02 M potassium phosphate-buffered saline (KPBS, pH 7.4) containing azide. The presence of PRV in the brain was revealed immunocytochemically in free-floating sections. The sections were incubated subsequently with rabbit anti-PRV (1:2,500 for 24 h), goat anti-rabbit immunoglobulin G (IgG) (1:800 for 18 h; Sigma, St. Louis, MO), and rabbit peroxidase anti-peroxidase (1:800 for 4 h; Dakopatts; Dako, Glostrup, Denmark). Intermittent washing was done with KPBS. The presence of peroxidase was revealed with a solution consisting of 0.05% 3,3-diaminobenzidine (Sigma), 2.5% nickel ammonium sulfate, 0.04% ammonium chloride, and 0.005% hydrogen peroxide. The specificity of the antibodies was tested previously (20) on PRV-positive sections that were either incubated with anti-PRV without secondary antibodies or with the secondary antibody without a preceding anti-PRV incubation. All such control experiments were negative. The distribution of the virus-labeled cells was studied with bright-field microscopy and charted in camera lucida drawings.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Fourteen rats survived the experimental procedures, and six of those rats exhibited viral labeling in the forebrain. This infection rate is slightly higher than the success percentage in our previous study with the PRV-Bartha strain, in which we obtained an infection rate of ~30% in healthy animals (20). The eight survivors without forebrain infections either showed only a few labeled preganglionic cells in the brain stem, in the dorsal motor vagus nucleus and/or the periambiguus area, or were not infected at all. Overinfection, present in ~20% of non-T₁-transected animals in our previous study (20), was not found in the survivor group, which may be a consequence of the slower progression of infection in the parasympathetic system compared with the sympathetic system (8).

The included cases showed a bilateral labeling pattern at the various levels of the neuraxis, but the labeling in the right hemisphere dominated. The number of PRV-infected cells in the forebrain varied among the cases, but the infected regions were the same in all six animals. In brief, virus-infected cells were found in the following structures.

Brain stem. In accordance with previous viral labeling studies dealing with heart innervation (15, 16, 20), infected CVMs were found in the periambiguus area (Fig. 1D), the dorsal motor vagus nucleus, and the lateral reticular formation (Fig. 2G). Retrograde transneuronal infected neurons were located in the nucleus of the solitary tract (NTS), the area postrema, the spinal trigeminal nucleus caudalis, the ventrolateral reticular formation, and the raphe magnus (Fig. 2, F and G). At the pontomedullary level the locus ceruleus, the subceruleus nucleus, the Kölliker-Fuse nucleus, the parabrachial nucleus, and the A5 catecholaminergic area contained virus-infected cells bilaterally (Fig. 2F).

View larger version (48K): [in this window] [in a new window]   Fig. 2.   Series of schematic coronal sections showing, from rostral to caudal (A-G, respectively), the locations of virus-infected cells at various levels of the neuraxis after inoculation of left ventricular myocardium in T1-transected rats. A: Acb, nucleus accumbens; ACG, anterior cingulate gyrus; CC, corpus callosum; CPU, caudate putamen; GI, granular insular cortex; IL, infralimbic cortex; PIR, piriform cortex; PL, prelimbic cortex. B: AC, anterior commissure; BST, bed nucleus of the stria terminalis; DI, dysgranular insular cortex; LPO, lateral preoptic nucleus; LSV, lateral septum; LV, lateral ventricle; MPO, medial preoptic nucleus; S1BF, somatosensory cortex, barrel fields; ST, stria terminalis; TO, optic tract. C: 3, 3rd ventricle; AHN, anterior hypothalamus; CA3, Ammon's horn, area 3; CEA, central amygdala; FR, frontal cortex; LHA, lateral hypothalamus; MEA, medial amygdala; PVN, paraventricular hypothalamus; PVT, paraventricular thalamus; RSG, retrosplenial gyrus; SS, somatosensory cortex. D: CA1, Ammon's horn, area 1; CM, centromedial thalamus; DG, dentate gyrus; DMH, dorsomedial hypothalamus; ECT, ectorhinal cortex; LA, lateral amygdala; LH, lateral habenula; PH, posterior hypothalamus; VMH, ventromedial hypothalamus; VPL, ventroposterolateral thalamus; ZI, zona incerta. E: IPN, interpeduncular nucleus; Lm, medial lemniscus; MGN, medial geniculate; MRF, midbrain reticular formation; PP, peripeduncular nucleus; R, red nucleus; S, subiculum; SC, superior colliculus; SNR, substantia nigra, reticular part. F: 4, 4th ventricle; DR, dorsal raphe; KF, Kölliker-Fuse nucleus; Ldt, lateral dorsal tegmental nucleus; M5, trigeminal motor nucleus; Pb, parabrachial area; Pr5, principal spinal trigeminal nucleus; PRN, pontine reticular nucleus; Py, pyramidal tract; RMG, raphe magnus. G: 10, dorsal motor vagus nucleus; 12, hypoglossus nucleus; LRN, lateral reticular nucleus; Sp5, spinal trigeminal nucleus; Io, inferior olive. H: location of rostral-to-caudal coronal sections (A-G, respectively) of rat neuraxis. (Sections are adapted from electronic graphics files of Ref. 17.)

Midbrain. At the midbrain levels virus-labeled cells were located in the periaqueductal gray (Fig. 1C), the precommissural nucleus, the mesencephalic reticular formation, the peripeduncular nucleus, and the perirubral area (Fig. 2E).

Forebrain. Circumscribed labeling (Fig. 2, A-C) was encountered in the cortical areas, in particular in the anterior cingulate (Fig. 1B) and the pre- and infralimbic cortexes. Other structures that contained virus-infected cells after myocardial inoculation were the dysgranular insular cortex (Fig. 1A), located dorsal to the rhinal fissure overlying the claustrum, the central and medial amygdaloid nuclei, the subfornical organ, and the ventral part of the lateral septal nucleus. The cortical infections were restricted to the pyramidal cells of layer 5.

Controls. The intravenous injections of the virus did not cause any infection of the central nervous system, either in healthy or spinal cord-transected rats. All circumventricular organs were free of virus-infected cells, as described previously in other studies using an intravenous route of neurotropic virus application (15, 16, 20).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

In the present investigation we have revealed, for the first time, the morphology of neuronal networks that control CVM activity in the rat after application of a neuroanatomic marker substance in the ventricular myocardium. Previously, this was not possible because of loss of signal intensity during the process of transneuronal retrograde labeling. Virus replication in the infected neurons prevents the loss of signal intensity after the passage of the synaptic clefts. It has been demonstrated, both at light microscopic and electron microscopic levels, that viral infection is restricted to functional neuronal networks and that the nonspecific spread of virons to parallel neuronal networks is prevented by astrocytes and microglial cells (3).

Virus-infected cells were revealed at various levels of the neuraxis. Considering the speed of the transneuronal retrograde infection, labeled higher-order cells have either direct synaptic contacts with CVMs or indirect access through an intermediate relay area. Direct projections may derive from the locus ceruleus (19), the parabrachial area (6, 9, 18), the periaqueductal gray matter (9, 18), the dorsal hypothalamus (9, 18) and the retrochiasmatic area (9, 18), the central amygdala (4, 9, 18), the bed nucleus of the stria terminalis (4), the zona incerta (9, 18), and the prelimbic cortex (12). The anterior cingulate and the insular and somatosensory cortexes most likely influence CVM activity through intermediate relay areas, including the lateral hypothalamic area, the amygdala, the periaqueductal gray matter, and the NTS (1, 24).

Parasympathetic command neurons in cardiac disease. Decreased parasympathetic autonomic activity after myocardial infarctions may have several causes, including damage of cardiac sensory and autonomic nerves and/or limbic circuitry dysfunction (2, 23). The presence of ischemic lesions on forebrain magnetic resonance imaging scans of heart disease patients illustrates the susceptibility of the limbic regions to neuronal damage (5). Alternatively, in some of the virus-infected limbic forebrain areas a selective endothelial leakage has been demonstrated after acute myocardial infarction in rats, which could be mimicked with intravenous injections of recombinant tumor necrosis factor- (21, 22). The endothelial leakage was associated with local neuronal uptake of IgG and expression of heat shock proteins, which suggests that the leakage may be a cause of transient or sustained neuronal dysfunction in the forebrain limbic circuitry. Because myocardial infarction involves inflammatory responses (7, 13), immune-mediated limbic neuronal dysfunction should be considered an underlying cause of autonomic imbalance in heart disease patients.