Abstract
Excitatory amino acids (EAAs) mediate their effects through the glutamate receptors (GluRs) in the brain. GluRs play an important role in the treatment of a variety of neuropsychiatric conditions and are central to the neurotoxicity of EAAs such as domoic and kainic acid. Unstained histological preparations of human heart tissues were used for the histopathological assessment, the anatomical identification of specific cardiac structures and the presence of the GluRs. Immunohistochemical stains with the biomarkers protein gene product (PGP 9.5) and the neurofilaments (NF 160 and NF 200) were used to identify neural structures and the components of the conducting system. Several subtypes of GluRs were differentially expressed and each had a specific distribution. In contrast to nonhuman primates, GluRs are more widely expressed in humans, where the working myocardium and the wall of blood vessels stained for GluRs. The immunolabelling was observed within the specialized structures of the conducting system, intramural nerves, and ganglia cells. These receptors may be involved in important cardiac functions such as contraction, rhythm, coronary circulation, and thus may be implicated in the pathobiology of some cardiac disease. The GluRs in the heart could be targets for the effects of excitatory compounds and is therefore an important consideration for the safety evaluation of foods and therapeutic products.
Introduction
Theamino acids glutamate and aspartate, are abundantly present in the mammalian central nervous system (CNS) and are the principal excitatory neurotransmitters in the brain (Gasic and Hollman, 1992; Cunningham et al., 1994; Hollman and Heinemann, 1994; Dingledine et al., 1999; Miller et al., 1999). There are many other exogenous substances acting as glutamate analogues that possess excitatory properties and potential excitotoxic effects. Glutamate and its analogues may enter the food supply during food preparation and food processing, as contaminants and/or additives (Zautcke et al., 1986; Iverson et al., 1990; Gasic and Hollmann, 1992; Krogsgaard-Larsen and Hansen, 1992; Olney, 1994; Peng et al., 1994; Mueller et al., 1996; Dingledine et al., 1999; Gill and Pulido, 2005, 2000; Gill et al., 1999, 2000; Silvagni et al., 2005).
Domoic acid (DA) is one of the most potent excitatory amino acids (EAAs) and neurotoxins that can enter the food chain. It is produced by a phytoplankton that is ingested and accumulated in the digestive system of seashells such as mussels (Iverson et al., 1990; Perl et al., 1990; Teitelbaum et al., 1990; Tryphonas et al., 1990; Olney, 1994; Peng et al., 1994; Watters, 1995; Silvagni et al., 2005). DA has been responsible for one episode of human intoxication and many in sea lions and other types of wildlife (Iverson et al., 1990; Krogsgaard-Larsen and Hansen, 1992; Watters 1995; Scholin et al., 2000; Silvagni et al., 2005). A variety of EAAs can also be present in foods including aspartate and monosodium glutamate (MSG) (Zautcke et al., 1986; Rockhold et al., 1989; Krogsgaard-Larsen and Hansen 1992; Cunningham et al., 1994; Olney, 1994; Gulland 2000; Scholin et al., 2000; Silvagni et al., 2005). There are differences in the potency of each compound and of their affinity for specific subtypes of the GluRs (Cunningham et al., 1994; Hollmann and Heinemann 1994; Dingledine et al., 1999; Miller et al., 1999).
Although these receptors were once thought to be predominantly located in the CNS, recent evidence shows that they are also present in peripheral neural and non-neural tissues (for review Gill and Pulido, 2005; Gill et al., 2000, 1999, 1998; Gill et al., 2000, 1999, 1998; Gill and Pulido 2000; Mueller et al., 2003). These GluRs in peripheral tissues are potential targets for the toxic effects of EAAs present in foods and the environment (Krogsgaard-Larsen and Hansen 1992; Olney, 1994; Gill et al., 2000, 1999, 1998; Mueller et al., 2003; Gill and Pulido, 2005, 2000). They may also be viewed as potential sites for drug development. The association of arrhythmias and other cardiovascular symptoms in humans intoxicated with DA and on individuals susceptible to MSG prompted us to focus our attention to the investigation of GluRs in heart (Zautcke et al., 1986; Rockhold et al., 1989; Winter and Baker 1995; Gulland 2000; Scholin et al., 2000; Mueller et al., 2003). We have hypothesized that GluRs play a role in mediating the cardiac effects of these glutamate analogues. Hence, the observation of cardiac lesions in sea lions intoxicated with DA further supports the view that DA may be cardiotoxic (Zautcke et al., 1986; Winter and Baker 1995; Gill et al., 1999, 1998; Gulland, 2000; Mueller et al., 1996, 2003; Scholin et al., 2000). In previous publications we have described the presence and characterization of GluRs in the heart of rats and nonhuman primates (Mueller et al., 1996, 2003; Gill and Pulido, 2000, 2005; Gill et al., 1998, 1999, 2000). In order to assess their relevance to human health we investigated the distribution and localization of several subtypes of GluRs and neural biomarkers in the human heart tissues.
Materials and Methods
Histopathology
The hearts were obtained from people who had died of non-cardiac causes and who had no clinically heart disease symptoms. Histological preparations of these hearts from both males and females (ranging from 13 to 79 years of age) were obtained from archived specimens collected during postmortem examinations following the standard autopsy procedures established at the Ottawa Hospital (Ottawa, Ontario). Archival animal tissues from our own laboratory were used for comparison. All tissues were dissected for specific identification of the different components of the heart, including the conducting system (Withdran and Lev, 1951; Mueller et al., 1996, 2003). Tissues blocks were fixed in formaldehyde, embedded in paraffin, sectioned at 5–6 μ, mounted on charged slides (Snowcoat X-Tra, Surgipath), and stored until needed.
Immunohistochemistry
Deparaffinized slides were immersed in 0.5% hydrogen peroxide/100% ethanol, washed in 100% and 95% ethanol and in running distilled water. Slides were placed in 10 mM sodium citrate buffer (pH 6.0) and microwaved for antigen retrieval (Kenmore 900 W or H2200/Energy Beam Science Inc.). Microwave-treated sections were washed in PBS and blocked for endogenous avidin and biotin. The slides were then incubated overnight at 4°C with polyclonal antibodies to PGP 9.5, NMDAR1, GluR1, GluR 2/3, mGluR 1, mGluR 2/3 (Chemicon International Inc. Temecula, CA, USA), mGluR 5 (Upstate Biotechnology Inc., Lake Placid, NY, USA); monoclonal antibodies to GluR 5/6/7 (BD Biosciences, Franklin Lakes, NJ, USA); neurofilaments (NF 160 and NF 200) and connexin 43 (Sigma-Aldrich Canada, Oakville, ON, Canada). Primary antibodies were diluted in 15% normal serum. After washing in PBS, slides were then incubated at room temperature in either biotinylated F(ab′) swine-anti-rabbit or rabbit anti- mouse secondary antibody (Dakopatts, Dimension Lab) diluted in 15% normal serum.
After washing in PBS, slides were incubated for 30 minutes at room temperature in streptavidin (1:200) complex (LAB/Dakopatts, Dimension Lab., Mississauga, Ontario, Canada). Slides were washed in PBS and 50 mM Tris buffer and then treated with 3′,3-diaminobenzidinetetrachloride (DAB 80 mg/72 ul of 30% hydrogen peroxide) in 400 ml 50 mM Tris buffer (pH 7.2). The slides were then rinsed in running tap water and counterstained with hematoxylin. Slides were rehydrated and coverslipped with Micromount medium (Surgipath).
For anatomical reference some sections were stained with Mayer’s hematoxylin and eosin Y (H&E, Armed Forces Institute of Pathology, Washington, DC, USA). Photographs were taken on an Axiophot Zeiss microscope (Germany) with a digital camera, linked to an image analysis system (Progres camera, ROCHE image analysis and archiving system). The above procedures have being used in previous investigations and described in detail (Gill et al., 1998, 1999; Mueller et al., 2003).
Controls for Immunohistochemistry
Rat brain sections were used as positive controls for the antibodies and peptide absorption. For negative controls, the solution (LAB/Dakopatt, Dimension Lab, Mississauga, Ontario, CA) containing the 15% normal serum was substituted without the primary antibody (Ab.). Validation of the specificity of the antibody with the corresponding peptide was performed on heart sections. Peptides for mGluR5 (BioSource International, Hopkinton, MA), GluR2/3 and mGluR1 (Upstate Biotechnology Inc., Lake Placid, NY, USA) were used against the antibodies tested. Optimum concentration of the peptide was determined using the dilution curve for the peptide against the corresponding antibody. The antibody and peptide were co-incubated at room temperature for 1 hour prior to placing the mixtures on the slides (Gill et al., 1999; Gill and Pulido, 2000; Mueller et al., 2003) this was followed by immunohistochemistry as above.
Molecular Analysis of the GluR in the Different Components of the Human Heart PCR and Cloning and cDNA Sequencing
Three glutamate receptors from the rat brain—NMDAR 1, KA-2, and GluR 3 were used as probes for Northern hybridization. The RT-PCR and PCR cloning of these have been previously described in our previous publication (Gill et al., 1999; Gill and Pulido, 2000). The cDNA sequencing reactions were performed using the BigDye terminator V3.0 cycle sequencing kit, (Applied Biosystems, Mississauga ON) according to the manufacturer’s specifications. Post reaction removal of unincorporated dyes was done using the Montage Seq96 system (Millipore, Etobicoke, ON). Sequences were resolved on an Applied Biosystems 3100 capillary sequencer (Applied Biosystems, Foster City, CA) and individual sequencing runs were assembled into contigs using Sequencher software (Gene Codes, Woburn, MA).
Northern Hybridization
Commercially prepared MTE Multiple Tissues Expression Array 3, containing normalized poly A+ RNA from different human tissues were purchased from BD Bioscience (Palo Alto, CA). Membranes were first incubated for 1 hour in prehybridization solution (20% SDS, 0.25 M Na2HPO4, 1 mM EDTA and 1% blocking agent). DIG-labelled NM-DAR1, KA-2 and GluR3 cDNA probes were prepared by PCR using alkali-labile, DIG-11-dUTP. The probes were added independently to three different and freshly prepared pre-hybridization solutions at 5 pmol/ml and hybridized overnight at 65°C. The hybridization solution was removed, and the filters washed 4 times (5 minutes each) in a solution of 2X SSC and 0.4% SDS at 65°C, followed by 4 washes (5 minutes each) of higher stringency in a solution of 0.5X SSC and 0.1% SDS at 60°C.
Membranes were blocked with 1% blocking solution in wash buffer (3M NaCl, 0.1 M Maleic acid, 0.3% Tween 20, pH 7.5) for 1 hour. Anti-DIG-AP conjugate was added into the blocking solution previously made at 1:10,000 dilution and the incubation was continued another 30 minutes. This was followed by 4 washes (10 minutes each) with the wash buffer and then incubated 2 minutes in 0.1 M Tris, pH 9.5. Finally the membrane was placed between transparency pages and incubated briefly with the substrate (CDP-Star). The excess liquid was removed and the filters were exposed to X-ray film until the bands became clearly visible (Gill et al., 1999).
Results
Table 1 show a summary of the distribution of all the biomarkers tested by immunohistochemistry. The antibodies to neurofilaments NF 160 and NF 200 (Figure 1a) showed specific affinity for nerve fibres and intramural ganglia. No staining was observed in the blood vessels, conducting system or myocardium. The neural biomarker PGP 9.5 (Figure 1b) stained the AV node, bundle of His, nerve fibres and intramural ganglia. All the human hearts examined showed differential expression of several subtypes of iGluRs and mGluRs with each displaying a variable distribution and intensity of the stain. The grading system for the distribution of the GluRs is based on the intensity of the stain- from the strongest signal (4+) to the weakest (1+).
In some instances, no signal was observed, as indicated by (−)ve staining. Differential expression was observed in the specialized structures of the conducting system, the myocardium, the wall of blood vessels, nerve fibres and intramural ganglia cells depending on the antibody used (Figures 1 and 2). The mGluR 2/3 showed no staining in any structures of the heart. Antibodies to the subtypes NMDAR 1, GluR 2/3, GluR 5/6/7, mGluR 5, and mGluR1 showed positive staining in nerve fibres, wall of blood vessels, atrial and ventricular cardiocytes and in different components of the conducting system including Purkinje fibres, AV node and the bundle of His (Figures 1D, 2B,C, D).
For the 2 controls for immunohistochemistry, there was no signal, hence confirming the specificity of the antibodies (Figure 2A). The gap junction protein connexion (Cx)-43 showed strong staining in the intercalated discs (ID) of the myocardium with a wide distribution through atrial and ventricular cardiocytes (Figure 1C). This feature was in contrast to the negative staining for the IDs observed with NMDAR 1 (Figure 1D). NMDAR 1 appears to have specific intra-cellular distribution within sarcomere and myofibrils of the myocardiocyte.
The molecular analysis of the 3 different subtypes of the GluRs including NMDAR 1, GluR 3 and KA-2 on the multiple tissue array, supports the differential specificity in the different components of the heart. This array contained poly A+ from the total heart and from the specific components including aorta, left atrium, right atrium, left ventricular, right ventricular, intraventricular septum and apex of the heart. The subtype GluR receptors, KA-2 showed the strongest signal in the right ventricle, whereas the NMDAR 1 had the most expression in all areas (Figure 3). The strongest signal being in the apex of the heart, and in the right and the left ventricles. The aorta showed the least amount of both receptors Ka-2 and NMDAR 1. No signal was observed for the GluR 3 in any part of the heart.
Conclusion and Discussions
In our previous studies using monkey and rat hearts, we identified the conducting system, cardiac intramural nerve fibers and ganglia cells as the main structures expressing GluRs (Mueller et al., 1996, 2003). In this study, we extended these findings having identified and evaluated the human cardiac conducting system, as the main conduction pathway for excitation and rhythmic control (Widran and Lev, 1951; Gorza and Vitadello, 1989; Dobrzynski et al., 2003). Histological sections were obtained from specific anatomical regions of the heart to ensure visualization of the conducting system, including the sinoatrial node (Figure 2B); atrioventricular node, Bundle of His (Figures 1B and 2 C) and Purkinje fibres. Sections from the wall and septum of the atria and ventricle were also used to assess the working myocardium. The anatomical structures encompassing the conducting system were identified using the neural markers protein gene product (PGP) 9.5 (Schofield et al., 1995; Mueller et al., 1996, 2003) and neurofilaments (NF) (Mueller et al., 1996, 2003). The antibody to PGP 9.5 has been shown to have affinity in neurons and nerve fibers at all levels of the central and peripheral nervous system and in many neuroendocrine cells (Schofield et al., 1995). As in the monkey and rat, PGP 9.5 showed strong affinity for all components of the conducting system as well as nerve fibers and ganglia cells (Figure 1B). NF 160 and NF 200 showed affinity only to nerve fibers and ganglia cells, but not to the components of the conducting system (Mueller et al., 2003, 1996).
In addition to using these neural biomarkers, this is the first report showing the immunohistochemical and the molecular distribution of GluRs in specific anatomical structures in the human heart. These structures included the nerve fibers, ganglia cells, conducting system, atria, and ventricular cardiocytes. The presence on GluRs in myocardium, conducting system, nerve fibers and intramural ganglia cells in all three species, strongly supports the view that these receptors may play a role in the physiology and pathophysiology of cardiac rhythm and excitation in human. The cardiac ganglia are usually associated with interconnecting nerves that form a ganglionated plexus. They have been identified in the atria and atrioventricular regions. In humans, intrinsic cardiac ganglia are connected to the cranially located extracardiac ganglia via cardiopulmonary nerves adjacent to the larger vessels (Scherlag and Po 2006; Tan et al., 2006). At these sites, they may also serve as important target sites for the toxic effect of excitatory compounds.
In the CNS, GluRs have been implicated as the mediators for excitotoxic effects of EAAs or their structural analogues such as domoic acid and MSG. The presence GluRs within the specific components of the conducting system could explain some of the clinical symptoms that have been associated with MSG and or domoic acid ingestion (Rockhold et al., 1989; Teitelbaum et al., 1990; Tryphonas et al., 1990; Olney, 1994; Watters, 1995; Gulland, 2000; Scholin et al., 2000; Silvagni et al., 2005). Symptoms such as burning, facial pressure, palpitations, and chest pains associated with MSG consumption in individuals who cannot degrade it, are known as the “Chinese Distress Syndrome” (Zautcke et al., 1986; Olney, 1994).
Winter and Baker (34) have shown that L-glutamate increases the frequency of Ca+2 oscillations, an effect that has been positively correlated with increased contraction frequency in myocardial cells, which could lead to reduced cardiac filling, hypoxia and angina-like chest pains. Recent publications (Gulland, 2000; Scholin et al., 2000; Silvagni et al., 2005) showed histologic cardiac lesions in sea lions that died of domoic acid intoxication off the coast of California in 1998.
The NMDAR 1 showed strong immunostaining for the sarcomere with distinct labelling of the striations and myofibrils but not for the intercalated discs (ID). In our earlier study, the antibody for mGluR 5 showed positive staining for ID in the rat heart (Gill et al., 1999). Therefore, we used Cx-43 a positive control for staining for IDs. ID represents junctional complexes found between adjacent myocytes, having higher concentrations in the ventricles. The entire junctional surface is specialized in various ways for maintaining cell cohesion and to provide low resistance bridges for the spread of excitation. The significance of ionic coupling is that chains of individual cells behave as a syncytium, allowing the signal to contact and pass in a wave from cell to cell (Ganong, 1999).
Gap junctions in the heart provide low resistance pathways, facilitating electrical and metabolic coupling between cardiac muscle cells for coordinated action of the heart and tissue homeostasis. The conductance of these junctions, and therefore their function, is likely to be affected by a physiological change. Hence, any changes that alters the expression of the NMDAR 1, could lead to ionic equilibrium that could quickly lead to arrhythmias through these gap junctions. Furthermore, the close aposition of NMDAR1 in the cardiocyte and Cx-43 on the intercalated disc suggest possible cross-communication. Cx-43 is the constitutive protein for the formation of cardiac gap junctions and is essential for cell-cell and normal cardiac function. Recent studies demonstrate that desphophorylation and redistribution of Cx-43 is an early sign of cardiac injury after hypoxia (Matsushita et al., 2006).
In conclusion, the general findings are that the human heart appears to have a greater affinity and wider distribution of some of the GluRs as compared to rats and non-human primates (Mueller et al., 2003). Previously we correlated the differences in the distribution of GluRs in rat and nonhuman primates to the size of the heart and animal species ranking in evolution (Mueller et al., 1996, 2003). However, findings in this study do not support our original interpretations. Hence, the reason for the species differential distribution needs further investigation.
Regardless of the variation in affinities and distribution of specific subtypes in each of the species investigated (rat, monkey, human), the presence of GluRs in myocardium, conducting system, nerve fibers and intramural ganglia cells, supports the view that these receptors may play a role in cardiac function. Furthermore, the stronger affinity and wider distribution of both ionotropic and metabotropic GluRs in the human heart fosters our view that these receptors may be able to influence the physiology and pathophysiology of cardiac rhythm and excitation in human.
At these locations they may also serve as important target sites for the toxic effect of excitatory compounds such as DA and MSG. Individuals with premorbid cardiac pathologies may have an elevated risk of cardiotoxicity when exposed to excitatory compounds in foods or pharmacotherapy. Based on the prominent and specific cellular expression of GluRs in the human heart, we believe that these receptors play an important regulatory function such as contraction, rhythm and the pathobiology of some cardiac disease. Depending on the functionality and specificity of these GluRs, they could represent novel and specific targets for drug development in cardiac diseases.
Footnotes
Acknowledgment
We would like to thank P. McGuire, P. Smyth, and I. Greer for their excellent support, and we thank the pathology staff of the Ottawa Hospital for providing the human heart tissues. This study was supported by a grant from the Office of the Chief Scientist in Health Canada.