Distribution of vagal afferent fibers of the guinea pig heart labeled by anterograde transport of conjugated horseradish peroxidase

Cardiac vagal afferent neurotransmission in health and disease: review and knowledge gaps

The meticulous control of cardiac sympathetic and parasympathetic tone regulates all facets of cardiac function. This precise calibration of cardiac efferent innervation is dependent on sensory information that is relayed from the heart to the central nervous system. The vagus nerve, which contains vagal cardiac afferent fibers, carries sensory information to the brainstem. Vagal afferent signaling has been predominantly shown to increase parasympathetic efferent response and vagal tone. However, cardiac vagal afferent signaling appears to change after cardiac injury, though much remains unknown. Even though subsequent cardiac autonomic imbalance is characterized by sympathoexcitation and parasympathetic dysfunction, it remains unclear if, and to what extent, vagal afferent dysfunction is involved in the development of vagal withdrawal. This review aims to summarize the current understanding of cardiac vagal afferent signaling under in health and in the setting of cardiovascular disease, especially after myocardial infarction, and to highlight the knowledge gaps that remain to be addressed.

Evaluation of the bilateral cardiac afferent distribution at the spinal and vagal ganglia by retrograde labeling

The identity of sensory neurons innervating the heart tissue and the extent of information reported to the brain via these neurons are poorly understood. In order to evaluate the multidimensional distribution and abundance of the cardiac spinal and vagal afferents, we assessed the retrograde labeling efficiency of various tracers, and mapped the cardiac afferents qualitatively and quantitatively at the bilateral nodose ganglia (NGs) and dorsal root ganglia (DRGs). From the five different retrograde tracers evaluated, Di-8-ANEPPQ yielded reproducibly the highest labeling efficiency of cardiac afferents. We demonstrated specific cardiac afferents at NGs and C4 to T11 DRG segments. Next, the 2D reconstruction of the tissue sections and 3D imaging of the whole NGs and DRGs revealed homogeneous and bilateral distribution of cardiac afferents. The quantitative analyses of the labeled cardiac afferents demonstrated approximately 5-6% of the soma in NGs that were equally distributed bilaterally. The neuronal character of Di-8-ANEPPQ labeled cells were validated by coimmunostaning with pan-neuronal marker Tuj-1. In addition, the cell diameters of labeled cardiac sensory neurons were found smaller than 20 μm, implying the nociceptor phenotype confirmed by co-labeling with TRPV1 and Di-8-ANEPPQ. Importantly, co-labeling with two distinct tracers Di-8-ANEPPQ and WGA-647 demonstrated exclusively the same cardiac afferents in DRGs and NGs, validating our findings. Collectively, our findings revealed the cardiac afferents in NGs bilaterally and DRGs with the highest labeling efficiency reported, spatial distribution and quantitation at both 2D and 3D levels, furthering our understanding of this novel neuron population.

Different brain activation under left and right ventricular stimulation: an fMRI study in anesthetized rats

BACKGROUND

Myocardial ischemia in the anterior wall of the left ventricule (LV) and in the inferior wall and/or right ventricle (RV) shows different manifestations that can be explained by the different innervations of cardiac afferent nerves. However, it remains unclear whether information from different areas of the heart, such as the LV and RV, are differently processed in the brain. In this study, we investigated the brain regions that process information from the LV or RV using cardiac electrical stimulation and functional magnetic resonance imaging (fMRI) in anesthetized rats because the combination of these two approaches cannot be used in humans.

METHODOLOGY/PRINCIPAL FINDINGS

An electrical stimulation catheter was inserted into the LV or RV (n = 12 each). Brain fMRI scans were recorded during LV or RV stimulation (9 Hz and 0.3 ms width) over 10 blocks consisting of alternating periods of 2 mA for 30 sec followed by 0.2 mA for 60 sec. The validity of fMRI signals was confirmed by first and second-level analyses and temporal profiles. Increases in fMRI signals were observed in the anterior cingulate cortex and the right somatosensory cortex under LV stimulation. In contrast, RV stimulation activated the right somatosensory cortex, which was identified more anteriorly compared with LV stimulation but did not activate the anterior cingulate cortex.

CONCLUSION/SIGNIFICANCE

This study provides the first evidence for differences in brain activation under LV and RV stimulation. These different brain processes may be associated with different clinical manifestations between anterior wall and inferoposterior wall and/or RV myocardial ischemia.

Characterization of thromboxane A₂ receptor and TRPV1 mRNA expression in cultured sensory neurons

Thromboxane A(2) (TxA(2)) is an arachidonic acid metabolite that stimulates platelet aggregation and vasoconstriction when released from platelets and other cell types during tissue trauma. More recent research has demonstrated that TxA(2) can also stimulate vagal and spinal sensory nerves. The purpose of this study was twofold. One, we compared the expression of the TxA(2) receptor (TxA2R) in neurons from two sensory ganglia: the nodose ganglion (NG) containing cell bodies of vagal afferent nerves and the thoracic dorsal root ganglion (DRG) containing cell bodies of spinal afferent nerves. Two, we determined if TxA2R co-localizes with mRNA for the nociceptive marker, TRPV1, which is the receptor for the noxious substance capsaicin. We found a greater percentage of neurons in the NG that are positive for TxA2R expression than in the DRG. We also found that there was no correlation of expression of TxA2R with TRPV1. These data suggest that while TxA2R is expressed in both vagal and spinal neurons, TxA(2) may elicit stronger vagal or parasympathetic reflexes in the rabbit when released during tissue trauma depending on the location of release. Our data also indicate that TxA(2) is likely to stimulate both nociceptive and non-nociceptive neurons thereby broadening the types of neurons and reflexes that it may excite.

Differential expression of vesicular glutamate transporters by vagal afferent terminals in rat nucleus of the solitary tract: projections from the heart preferentially express vesicular glutamate transporter 1

The central projections and neurochemistry of vagal afferent neurones supplying the heart in the rat were investigated by injecting cholera toxin B-subunit into the pericardium. Transganglionically transported cholera toxin B-subunit was visualized in the medulla oblongata in axons and varicosities that were predominantly aggregated in the dorsomedial, dorsolateral, ventrolateral and commissural subnuclei of the caudal nucleus of the solitary tract. Unilateral vagal section in control rats prevented cholera toxin B-subunit labeling on the ipsilateral side of the nucleus of the solitary tract. Fluorescent and electron microscopic dual labeling showed colocalization of immunoreactivity for vesicular glutamate transporter 1, but only rarely vesicular glutamate transporters 2 or 3 with cholera toxin B-subunit in terminals in nucleus of the solitary tract, suggesting that cardiac vagal axons release glutamate as a neurotransmitter. In contrast, populations of vagal afferent fibers labeled by injection of cholera toxin B-subunit, tetra-methylrhodamine dextran or biotin dextran amine into the aortic nerve, stomach or nodose ganglion colocalized vesicular glutamate transporter 2 more frequently than vesicular glutamate transporter 1. The presence of other neurochemical markers of primary afferent neurones was examined in nucleus of the solitary tract axons and nodose ganglion cells labeled by pericardial cholera toxin B-subunit injections. Immunoreactivity for a 200-kDa neurofilament protein in many large, cholera toxin B-subunit-labeled nodose ganglion cells indicated that the cardiac afferent fibers labeled are mostly myelinated, whereas binding of Griffonia simplicifolia isolectin B4 to fewer small cholera toxin B-subunit-labeled ganglion cells suggested that tracer was also taken up by some non-myelinated axons. A few labeled nucleus of the solitary tract axons and ganglion cells were positive for substance P and calcitonin gene-related peptide, which are considered as peptide markers of nociceptive afferent neurones. These data suggest that the population of cardiac vagal afferents labeled by pericardial cholera toxin B-subunit injection is neurochemically varied, which may be related to a functional heterogeneity of baroreceptive, chemoreceptive and nociceptive afferent fibers. A high proportion of cardiac neurones appear to be glutamatergic, but differ from other vagal afferents in expressing vesicular glutamate transporter 1.

TRPC6 immunoreactivity is colocalized with neuronal nitric oxide synthase in extrinsic fibers innervating guinea pig intrinsic cardiac ganglia

Tachykinins depolarize guinea pig intracardiac neurons by activating nonselective cationic channels. Recently, members of the transient receptor potential family of membrane channels (TRPC) have been implicated in the generation of G protein-coupled receptor-activated nonselective cationic currents. We have investigated whether guinea pig cardiac neurons exhibit immunoreactivity to TRPC. Our results showed that nerve fibers within guinea pig intrinsic cardiac ganglia exhibited immunoreactivity to TRPC6. After culture of cardiac ganglia whole-mount explants for 72 hours, the TRPC6-IR fiber networks were absent. Therefore, the TRPC6-IR fibers were derived from sources extrinsic to the heart. A small percentage ( approximately 3%) of intracardiac neurons also exhibited TRPC6 immunoreactivity in control preparations, and the percentage of cells exhibiting TRPC6 immunoreactivity was not changed following explant culture for 72 hours. The few intrinsic TRPC6-IR neurons also exhibited nitric oxide synthase (NOS) immunoreactivity, indicating that they were nitrergic as well. We compared the immunohistochemical staining patterns of TRPC6-IR fibers with the staining patterns of a number of other neurotransmitters or neurotransmitter synthetic enzymes that mark specific extrinsic inputs to the intrinsic cardiac ganglia. The TRPC6-IR fibers were not immunoreactive for choline acetyltransferase, tyrosine hydroxylase, or substance P. However, the TRPC6-IR fibers exhibited immunoreactivity to neuronal NOS. Therefore, we propose that the TRPC6-IR fibers within the guinea pig intrinsic cardiac ganglia are vagal sensory fibers that also contain NOS. We found, in support of this conclusion, that TRPC6-IR cells were also present in sections of nodose ganglia.

Thromboxane A(2) mimetic evokes a bradycardia mediated by stimulation of cardiac vagal afferent nerves

Injections of the thromboxane A(2) mimetic U-46619 (10 and 20 microg) into the left atrium of anesthetized rabbits evoked decreases in heart rate (HR) and arterial blood pressure (ABP) followed by an increase in ABP. Bilateral, cervical vagotomy abolished the U-46619-induced bradycardia and attenuated the hypotension. Injections of U-46619 into the ascending aorta did not evoke the bradycardia and hypotension but did cause arterial hypertension. To further define the origin of the vagal reflex, recordings of nerve impulses were made from 11 chemosensitive cardiac vagal afferent nerves. Impulse frequency increased in all 11 fibers in response to left atrial injections of phenylbiguanide (20-30 microg) and U-46619 (5-10 microg). Onset time of nerve activity induced by U-46619 correlated with the onset time of bradycardia. We conclude that U-46619 injections into the left heart elicit decreases in HR and ABP via a vagal reflex that originates from the heart similar to the coronary chemoreflex described for other agents.

Distribution of cocaine- and amphetamine-regulated transcript peptide in the guinea pig intrinsic cardiac nervous system and colocalization with neuropeptides or transmitter synthetic enzymes

This study was conducted to establish the presence of cocaine- and amphetamine-regulated transcript peptide (CARTp) immunoreactivity in neurons and fibers within guinea pig atrial whole-mount preparations containing the intrinsic cardiac ganglia. Many cardiac ganglia, but not all, in a given whole-mount preparation, were innervated by CARTp-immunoreactive (IR) fibers. Following explant culture of whole mounts for 72 hours, the CARTp-IR fiber networks were absent, but the number of CARTp-IR neurons was increased markedly. These observations suggested that the majority of the CARTp-IR fibers in the intracardiac ganglia were derived from sources extrinsic to the heart. In control whole-mount preparations, very few CARTp-positive neurons were present. The few intrinsic CARTp-IR neurons also exhibited choline acetyltransferase (ChAT) immunoreactivity, indicating that they make up a small subpopulation of cholinergic postganglionic neurons. Some CARTp-IR neurons also exhibited nitric oxide synthase (NOS) immunoreactivity, indicating that they were nitrergic as well. We compared the immunohistochemical staining patterns of CARTp-IR fibers with the staining patterns of a number of other neurotransmitters or neurotransmitter synthetic enzymes that mark specific extrinsic inputs. The CARTp-IR fibers were not immunoreactive for ChAT, tyrosine hydroxylase, calcitonin gene-related peptide, or substance P. However, virtually all CARTp-IR fibers exhibited immunoreactivity to neuronal NOS (a marker for nitric oxide-producing neurons). CARTp-IR cells and NOS-IR cells were present in the nodose ganglia. In addition, CARTp-IR neurons in the nodose also were stained positively for NADPH-diaphorase. Thus, we propose that most CARTp-IR fibers within the guinea pig intrinsic cardiac ganglia are vagal afferent fibers that also contain NOS.

Local differences in vagal afferent innervation of the rat esophagus are reflected by neurochemical differences at the level of the sensory ganglia and by different brainstem projections

The objective of the present study was to characterize further the vagal afferent fibers in the rat esophagus, particularly those in its uppermost part, their cell bodies in vagal sensory ganglia, and their central projections. We applied immunohistochemistry for calretinin, calbindin, and calcitonin gene-related peptide (CGRP); retrograde tracing with FluoroGold; and transganglionic tracing with wheat germ agglutinin-horseradish peroxidase in combination with neurectomies. Vagal terminal structures in the muscularis propria of the whole esophagus consisted of calretinin-immunoreactive intraganglionic laminar endings that were linked to cervical vagal and recurrent laryngeal nerve pathways. The mucosa of the uppermost esophagus was innervated by a very dense net of longitudinally arranged, calretinin-positive fibers that were depleted by section of the superior laryngeal nerve. Distal to this area, the mucosa was virtually devoid of calretinin-immunoreactive vagal afferents. Calretinin-positive mucosal fibers in the upper cervical esophagus were classified into four types. One type, the finger-like endings, was sometimes immunoreactive also for CGRP. About one-third of cell bodies in vagal sensory ganglia retrogradely labeled from the upper cervical esophagus expressed CGRP, whereas two-thirds coexpressed calretinin and calbindin but not CGRP. In addition to the central subnucleus of the nucleus of the solitary tract, vagal afferents from the upper cervical esophagus also projected heavily to the interstitial subnucleus. This additional projection was attributed to mucosal afferents traveling through the superior laryngeal nerve. The present study provides a possible morphological basis for bronchopulmonary and aversive reflexes elicited upon stimulation of the esophagus.

Mechanisms of cardiac pain

Angina pectoris often results from ischemic episodes that excite chemosensitive and mechanoreceptive receptors in the heart. Ischemic episodes release a collage of chemicals, including adenosine and bradykinin, that excites the receptors of the sympathetic and vagal afferent pathways. Sympathetic afferent fibers from the heart enter the upper thoracic spinal cord and synapse on cells of origin of ascending pathways. This review focuses on the spinothalamic tract, but other pathways are excited as well. Excitation of spinothalamic tract cells in the upper thoracic and lower cervical segments, except C7 and C8 segments, contributes to the anginal pain experienced in the chest and arm. Cardiac vagal afferent fibers synapse in the nucleus tractus solitarius of the medulla and then descend to excite upper cervical spinothalamic tract cells. This innervation contributes to the anginal pain experienced in the neck and jaw. The spinothalamic tract projects to the medial and lateral thalamus and, based on positron emission tomography studies, activates several cortical areas, including the anterior cingulate gyrus (BA 24 and 25), the lateral basal frontal cortex, and the mesiofrontal cortex.

Retrograde and transganglionic transport of horseradish peroxidase-conjugated cholera toxin B subunit, wheatgerm agglutinin and isolectin B4 from Griffonia simplicifolia I in primary afferent neurons innervating the rat urinary bladder

In the present study, we investigated and compared the ability of the cholera toxin B subunit, wheat germ agglutinin and isolectin B4 from Griffonia simplicifolia I conjugated to horseradish peroxidase, to retrogradely and transganglionically label visceral primary afferents after unilateral injections into the rat urinary bladder wall. Horseradish peroxidase histochemical or lectin-immunofluorescence histochemical labelling of bladder afferents was seen in the L6-S1 spinal cord segments and in the T13-L2 and L6-S1 dorsal root ganglia. In the lumbosacral spinal cord, the most intense and extensive labelling of bladder afferents was seen when cholera toxin B subunit-horseradish peroxidase was injected. Cholera toxin B subunit-horseradish peroxidase-labelled fibres were found in Lissauer's tract, its lateral and medial collateral projections, and laminae I and IV-VI of the spinal gray matter. Labelled fibres were numerous in the lateral collateral projection and extended into the spinal parasympathetic nucleus. Labelling from both the lateral and medial projections extended into the dorsal grey commissural region. Wheat germ agglutinin-horseradish peroxidase labelling produced a similar pattern but was not as dense and extensive as that of cholera toxin B subunit-horseradish peroxidase. The isolectin B4 from Griffonia simplicifolia I-horseradish peroxidase-labelled fibres, on the other hand, were fewer and only observed in the lateral collateral projection and occasionally in lamina I. Cell profile counts showed that a larger number of dorsal root ganglion cells were labelled with cholera toxin B subunit-horseradish peroxidase than with wheat germ agglutinin- or isolectin B4-horseradish peroxidase. In the L6-S1 dorsal root ganglia, the majority (81%) of the cholera toxin B subunit-, and almost all of the wheat germ agglutinin- and isolectin B4-immunoreactive cells were RT97-negative (an anti-neurofilament antibody that labels dorsal root ganglion neurons with myelinated fibres). Double labelling with other neuronal markers showed that 71%, 43% and 36% of the cholera toxin B subunit-immunoreactive cells were calcitonin gene-related peptide-, isolectin B4-binding- and substance P-positive, respectively. A few cholera toxin B subunit cells showed galanin-immunoreactivity, but none were somatostatin-, vasoactive intestinal polypeptide-, or neuropeptide Y-immunoreactive or contained fluoride-resistant acid phosphatase. The results show that cholera toxin B subunit-horseradish peroxidase is a more effective retrograde and transganglionic tracer for pelvic primary afferents from the urinary bladder than wheat germ agglutinin-horseradish peroxidase and isolectin B4-horseradish peroxidase, but in contrast to somatic nerves, it is transported mainly by unmyelinated fibres in the visceral afferents.

Distribution of PGP 9.5, TH, NPY, SP and CGRP immunoreactive nerves in the rat and guinea pig atrioventricular valves and chordae tendineae

The distribution of nerves immunoreactive to protein gene product 9.5 (PGP 9.5), tyrosine hydroxylase (TH), neuropeptide Y (NPY), substance P (SP) and calcitonin gene related peptide (CGRP) antisera was investigated in the atrioventricular valves of the Sprague-Dawley rat and the Dunkin-Hartley guinea pig using confocal and epifluoresence microscopy. No major differences were noted between the innervation of the mitral and tricuspid valves in either species. For all antisera the staining was more extensive in the guinea pig valves. Two distinct nerve plexuses separated by a 'nearly nerve free' zone were identified in both species with each antiserum tested. This was most apparent on the anterior cusp of the mitral valve. The major nerve plexus extends from the atrioventricular ring through the basal, intermediate and distal zones of the valves towards the free edge of the valve cusp. These nerve bundles, arranged as primary, secondary and tertiary components, ramify to the free edge of the valve and extend to the attachment of the chordae. They do not contribute to the innervation of the chordae tendineae. The second, minor chordal plexus, runs from the papillary muscles through the chordae tendineae and passes parallel to the free edge of the cusp. The nerves of this minor plexus are interchordal, branching to terminate mainly in the distal zone, free edge of the valve cusp and adjacent chordae tendineae. Some interchordal nerve fibres loop from a papillary muscle up through a chorda, along the free edge and pass down an adjacent chorda into another papillary muscle. The nerve fibres of the major and minor plexuses intermingle although no evidence was found for interconnectivity between them. In the distal zone between the major plexus which extends from the base of the valve and the minor chordal plexus there is a zone completely free of nerves staining with antisera to TH and NPY. Occasional nerves which stained positive for PGP 9.5, SP and CGRP immunoreactivities crossed this 'nearly nerve free zone' passing either from the chordal/free edge nerves to the intermediate and basal zones or vice versa. An additional small nerve plexus which displayed immunoreactivity to CGRP antiserum extended from the atrioventricular ring into the basal zone of the valve cusp. Not all chordae tendineae displayed immunoreactive nerve fibres. It is concluded that the innervation patterns of the sensory and sympathetic neurotransmitters and neuropeptides examined in the atrioventricular valves of the rat and guinea pig are ubiquitous in nature. The complexity of the terminal innervation network of the mammalian atrioventricular valves and chordae tendineae may contribute to the complex functioning of these valves in the cardiac cycle.

Vagal afferent innervation of the atria of the rat heart reconstructed with confocal microscopy

We have used confocal microscopy to analyze the vagal afferent innervation of the rat heart. Afferents were labeled by injecting 1,1'-dioleyl-3,3,3',3'-tetramethylindocarbocyanine methanesulfonate (DiI) into the nodose ganglia of animals with prior supranodose de-efferentations, autonomic ganglia were stained with Fluoro-gold, and tissues were examined in whole mounts. Distinctively different fiber specializations were observed in the epi-, myo-, and endocardium: Afferents to the epicardium formed complexes associated with cardiac ganglia. These ganglia consisted of four major ganglionated plexuses, two on each atrium, at junctions of the major vessels with the atria. Ganglionic locations and sizes (left > right) were consistent across animals. In addition to principal neurons (PNs), significant numbers of small intensely fluorescent (SIF) cells were located in each of these plexuses, and vagal afferents provided dense pericellular varicose endings around the SIF cells in each ganglionic plexus, with few if any terminations on PNs. In the myocardium, vagal afferents formed close contacts with cardiac muscles, including conduction fibers. In the endocardium, vagal fibers formed "flower-spray" and "end-net" terminals in connective tissue. With three-dimensional reconstruction of confocal optical sections, a novel polymorphism was seen: Some fibers had one or more collaterals ending as endocardial flower sprays and other collaterals ending as myocardial intramuscular endings. Some unipolar or pseudounipolar neurons within each cardiac ganglionic plexus were retrogradely labeled from the nodose ganglia. In conclusion, vagal afferents form a heterogeneity of differentiated endings in the heart, including structured elements which may mediate chemoreceptor function, stretch reception, and local cardiac reflexes.

Tyrosine hydroxylase- and nitric oxide synthase-immunoreactive nerve fibers in mitral valve of young adult and aged Fischer 344 rats

Using confocal fluorescence microscopy we studied, in whole mounts of heart mitral valves of young adult and aged Fischer 344 rats, the distribution of nerves containing the catecholamine marker tyrosine hydroxylase (TH) or the synthetic enzyme marker for nitric oxide, nitric oxide synthase (NOS). TH-IR was localized in two separate nerve plexuses which do not intermingle. The 'major' plexus arose from the annulus region, traversed the basal zone of the valve, and ramified in the intermediate zone to form a dense network of fine fibers. The 'minor' plexus was restricted to the distal zone and originated from bundles that ascended the chordae tendineae to enter the valve cusp. A concentric zone located between the major and minor plexuses was devoid of TH-IR nerve fibers. Both plexuses demonstrated (i) nerves that contained numerous varicosities along the length of each fiber, (ii) many terminal axons and (iii) different shaped terminal axon endings. With age, the density of TH-IR innervation in the mitral valve was markedly reduced; and nerve fibers of the minor plexus were limited to the chordae tendinae, without extending into the valve cusp itself. NOS-IR fibers in the mitral valve formed a loose network that extended from the annulus to more than halfway down the cusp. The varicose beads of the terminal NOS-IR axons appeared to become progressively smaller and less intensely fluorescent until they disappeared at the terminal endings, which showed no specializations. No NOS-IR fibers were observed in the distal zone of the valve leaflet or in the chordae. In the aged mitral valve, the density of NOS-IR nerves was decreased, as compared with NOS-IR innervation in the young adult valve. The existence of TH and NOS as well as other signal molecule markers in heart valve nerves and the disparate patterns of their distribution and localization provide evidence supporting the theory that heart valve nerves form a complex reflexogenic control system in the mitral heart valve. In summary, two distinct neural architectures are described for TH-IR and NOS-IR valve nerves, respectively. The former are believed to be axons dedicated to sympathetic motor functions. The NOS-IR valve nerves may have sensory and/or postganglionic parasympathetic motor functions. An implication of these findings is that different, but perhaps related, valve functions may be mediated by separate, dedicated circuits.