Localisation of SCN10A Gene Product Nav1.8 and Novel Pain-Related Ion Channels in Human Heart Paul Facer, 1 Mphil, Prakash P

Localisation of SCN10A Gene Product Nav1.8 and Novel Pain-Related Ion Channels in Human Heart Paul Facer, 1 MPhil, Prakash P. Punjabi,2 FRCS, Andleeb Abrari,4 MD, Riyaz A. Kaba,5 MRCP, Nicholas J. Severs,5 PhD, John Chambers,3 PhD, Jaspal S. Kooner,3 FRCP and Praveen Anand,1 FRCP

Summary

We have shown that the gene SCN10A encoding the sodium channel Nav1.8 is a susceptibility factor for heart block and serious ventricular arrhythmia. Since Nav1.8 is known to be present in nerve fibres that mediate pain, it may be related to both cardiac pain and dysrhythmia. The localisation of Nav1.8 and other key nociceptive ion channels, including Nav1.7, Nav1.9, capsaicin receptor TRPV1, and purinergic receptor P2X3, have not been reported in human heart. The aim of this study was to determine the distribution of Nav1.8, related sodium and other sensory channels in human cardiac tissue, and correlate their density with sympathetic nerves, regenerating nerves (GAP-43), and vascularity. Human heart atrial appendage tissues (n = 13) were collected during surgery for valve disease. Tissues were investigated by im-

munohistology using specific antibodies to Nav1.8 and other markers. Nav1.8 immunoreactivity was detected in nerve fi- 2 bres and fascicles in the myocardium, often closely associated with small capillaries. Nav1.8 nerve fibres per mm correlated significantly with vascular markers. Nav1.8-immunoreactivity was present also in cardiomyocytes with a similar distribution pattern to that seen with connexins, the specialised gap junction proteins of myocardial intercalated discs.

Nav1.5-immunoreactivity was detected in cardiomyocytes but not in nerve fibres. Nav1.7, Nav1.9, TRPV1, P2X3/P2X2, and GAP43 positive nerve fibres were relatively sparse, whereas sympathetic innervation and connexin43 were abun-

dant. We conclude that sodium channel Nav1.8 is present in sensory nerves and cardiomyocytes of human heart. Nav1.8 and other pain channels provide new targets for the understanding and treatment of cardiac pain and dysrhythmia. (Int Heart J 2011; 52: 146-152)

Key words: Gene, Sodium channel, Cardiac, Pain, ECG

e have recently reported the presence of SCN10A peralgesia and allodynia.7) Increased immunoreactivity for

encoding the sodium channel Nav1.8 in human Nav1.8 was also observed in rhinitis, dental pulp and causalgic heart,1) and showed that it influences PR, P-wave finger, hence an association with painful and/or inflammatory W 8-10) and QRS time intervals in the ECG. Although Nav1.8 mRNA conditions. The majority of sodium channels in the mam- and its protein have been shown to be present in human senso- malian heart are TTX-resistant (TTX-R) and the major subtype 11) ry neurons, its localisation has not been investigated in the is Nav1.5, mutations of which are associated with Long QT heart. The aim of the present study was to determine the pres- syndrome and Brugada syndrome.11,12) Patch clamp recordings ence and distribution of sodium channel Nav1.8, in the human have demonstrated TTX-resistant ion channels eg. Nav1.5 and 13,14) heart, and compare its localisation with other markers of cardi- Nav1.8 in adult rat and mouse cardiomyocytes. Electrical ac innervation and conduction. coupling in the heart is mediated by gap junctions, comprising In mammals, voltage gated sodium channels (VGSC) are clusters of membrane channels composed of connexins.15) Three essential for the upstroke of the action potential and are crucial main connexins are expressed in the heart, connexin43(Cx43), for generation and propagation of action potentials.2) Nine Cx40, and Cx45.15) Cx43 has been established as the principal 3) α-subunits (Nav1.1-1.9) have been identified, of which the tet- connexin of ventricular myocytes in a wide range of mamma- 16,17) rodotoxin-sensitive (TTX-S) subunits Nav1.1, Nav1.6, and lian species, and is abundant in atrial myocardium. Nav1.7 and the TTX-resistant (TTX-R) subunits Nav1.8 and Apart from Nav1.8, other key molecular transducers of Nav1.9 are expressed in nociceptive dorsal root ganglion noxious stimuli in inflammatory and neuropathic disorders 4) (DRG) neurons. The neuronal subtype Nav1.8 has a key role have been discovered in recent years, including the sodium in the propagation of action potentials in nociceptive nerve fi- channels Nav1.7 and Nav1.9, but not as yet reported in human bres in animal models.5,6) We have shown previously that cardiac tissues. It has been proposed that the noxious heat and

Nav1.8 is present in human normal DRG, and increased in capsaicin receptor TRPV1 may be involved in ischaemia-in- painful neuromas and skin from patients with chronic local hy- duced firing of cardiac spinal afferents, and its activation may

From the Departments of 1 Clinical Neuroscience, 2 Cardiothoracic Surgery and 3 Cardiology, Hammersmith Hospital, Faculty of Medicine, Imperial College London, London, UK, 4 Department of Histopathology, Max Super Speciality Hospitals, New Delhi, India, and 5 Royal Brompton Hospital, Imperial College London, London, UK. Address for correspondence: Praveen Anand, MD, Peripheral Neuropathy Unit, Imperial College London, Area A, Ground Floor, Hammersmith Hospital, Du Cane Rd, London W12 ONN, UK. Received for publication December 27, 2010. Revised and accepted February 28, 2011. 146 Vol 52 No 3 NAV1.8 AND PAIN-RELATED ION CHANNELS IN HEART 147 release neuropeptides such as substance P and CGRP, to pro- Methods tect the heart from ischaemic injury.18) TRPV1 has been demonstrated by immunohistochemistry to be present in rat heart.19) Human heart tissues: Specimens (n = 13) were obtained ATP is among the substances released as a result of ischaemia from patients [median age (range), 67 years (52-81); 6 male] or inflammation, and its receptors P2X3 and P2X2 have been undergoing surgical intervention of the mitral valve with con- shown to be sensitised in the rat nodose ganglion, which sup- comitant surgical treatment of atrial fibrillation, at Hammer- plies sensory afferents to the heart.20,21) Neuronal synaptic vesi- smith Hospital, London. Disease history for patients included cles contain conserved membrane proteins which can bind diabetes (n = 5), ischaemia (n = 6), and both diabetes and is- toxins; of these SV2A has been identified as the receptor for chaemia (n = 4). Fully informed consent was obtained for all botulinum neurotoxin A.22) We have shown that immunoreac- tissues which were collected with approval of the Local Ethics tivity for SV2A is enriched in cholinergic and sensory neurons, Committee. All patients underwent surgical excision of the for which it serves as a useful marker. atrial appendage as part of the surgical treatment of atrial fibril- Primary afferent neurons innervating the heart transmit lation; this redundant excised tissue, which would otherwise sensory signals generated by cardiac sensory receptors.23) The have been discarded, was used for this study. Where sufficient cardiac sensory afferents and their cell bodies in the dorsal root tissue was available, samples were divided into two, one part ganglion (DRG) transmit cardiac nociception to the dorsal snap frozen and stored at -70°C and the remainder fixed by im- horn of the upper thoracic spinal cord.24) Increased production mersion in Zamboni’s fluid and then washed in phosphate of certain metabolites, such as protons, bradykinin and lactic buffered saline cryoprotectant (PBS; 0.05M phosphate, 0.9% acid, during myocardial ischemia may contribute to excitation saline, 15% w/v sucrose cryoprotectant and 0.01% w/v sodium of cardiac nociceptors.23,25-27) The heart is innervated by vagal azide, anti-fungal reagent) overnight before snap freezing in afferents28) essential for the cardiogenic sympatho-excitatory OCT (RALamb Ltd, Eastbourne, UK) embedding medium. reflex during myocardial ischaemia, and include nociceptive, Immunohistochemistry: Frozen sections (15 μm thick) were afferent nerve fibres. Innervation of the human heart has been collected onto coated glass slides (Sigma, Poole, UK). Unfixed described before and after transplantation29) along with the sections were postfixed in freshly prepared 4% w/v parafor- anatomy and localisation of neuropeptides and the innervation maldehyde in 0.15M phosphate buffered saline (PBS) for 30 of the conduction system.30,31) minutes. For immunoperoxidase staining, endogenous perox- Pathologic changes of the innervation have been studied ide was blocked in all tissue sections by incubation in industri- in clinical cardiac disorders. Insulin-dependent diabetics had al methylated spirit containing hydrogen peroxide (0.3% w/v) clinical evidence of cardiac denervation, and the vagus nerve for 20 minutes. After rehydration, sections were incubated and sympathetic trunks showed severe loss of myelinated fi- overnight with primary antibodies (Table). Sites of primary an- bres.32) It has been proposed that the absence of pain in diabet- tibody attachment were revealed using nickel-enhanced, avi- ics with myocardial infarction could be due to a lesion of the din-biotin peroxidase complex (ABC-Vector Laboratories, Pe- afferent nerves that mediate pain.33) There appears to be an as- terborough, UK)37) and counter-stained for nuclei in 0.1% w/v sociation between a history of spontaneous ventricular arrhyth- aqueous neutral red, dehydrated and mounted in xylene-based mia and an increased density of sympathetic nerves in patients mountant (DPX; BDH/Merck, Poole, UK), prior to photomi- with severe heart failure, suggesting that abnormally increased crography as described previously. For specificity, Nav1.8 pep- postinjury sympathetic nerve density may be in part responsi- tide antigen (GSK) was used at 1 to 1 × 10-5 mg/mL of diluted ble for the occurrence of ventricular arrhythmia and sudden Nav1.8 antibodies, preincubated overnight and used for immu- cardiac death in these patients.34) Human myocardial nerve fi- nostaining the following day. Other negative controls included bres regenerate after necrotizing injuries, and at least some of omission/replacement of primary antibody with normal rabbit the resulting scar-associated nerve fibres have structural fea- or mouse serum. For routine histology, frozen sections of atrial tures differing from those in uninjured myocardium. These appendage were stained using a standard haematoxylin and structural differences might be associated with functional alter- eosin procedure. ations that may affect the triggering of ventricular arrhythmi- Image analysis: Nerve fibres were quantified by computerized as.35) Changes in the content of Cx43 in gap junctions have image analysis (Olympus Analysis Five DP Soft, UK) where been associated with heart disease and spontaneous pain be- analogue images were captured via video link to an Olympus haviour has been reported in rats after reduction of Cx43.36) BX50 microscope and converted into digital monochrome im- In this study we have determined the localisation of sodi- ages by the computer. The grey-shade detection threshold was um channel Nav1.8, in the human heart, alongside markers of set at a constant level to allow detection of positive immuno- cardiac conduction (Nav1.5, connexin 43) and innervation. We staining and the area of highlighted immunoreactivity obtained have examined the presence and distribution of other key noci- as a percentage (% area) of the field scanned. Five images ceptor ion channels (Nav1.7, Nav1.9, TRPV1, P2X2 and P2X3), were captured (x40 objective magnification) from each of one and compared these with markers of sympathetic innervation tissue section per sample and the mean values used for statisti- (NPY, SV2A), regenerating nerve fibres (GAP43), and vascu- cal analysis. Image analysis of the same samples was per- larity (eNOS, vWF). We used immunohistological methods formed by two independent observers where comparison of re- and well characterised primary antibodies to these targets in sults showed no significant difference between values for % samples of human heart atrial appendage. area for any of the markers. For Nav1.8 immunoreactive nerve fibres, individual nerve fibres and fascicles (consisting of more than three fibres) were counted throughout the entire section and the area of section measured using the image analyser. Re- sults were then expressed as nerve fibres per mm2 section. The Int Heart J 148 FACER, ET AL May 2011

Table. Antibodies for Immunostaining

Antibodies to Source Ref # Titre

Nerve marker PGP9.5 Ultraclone RA101 1:40000 Nerve marker neurofilaments ‘cocktail’ Dakocytomation and Novocastra M0726/NCL-PERIPH 1:500/10000 Sodium channel Nav1.8 GlaxoSmithKline K107 1:200-400 Connexin43 Millipore MAB 3067 1:5000 # Sodium channel Nav1.5 Alomone ASC-005 lot AN-08 1: 200

Sodium channel Nav1.7 GlaxoSmithKline K241 1:200-400

Sodium channel Nav1.9 GlaxoSmithKline K186 1:200-400 Capsaicin receptor TRPV1 GlaxoSmithKline C22 1:5000-10000

Purinergic receptor P2X3 Neuromics RA10109 1:1000 # Purinergic receptor P2X2 Roche 1/8/01 36886 1:1000 Neuropeptide Y (CPON/NPY) Chemicon/Millipore AB9608 1:10000-20000 Botulinum toxin receptor; Synaptic vesicle 2A (SV2A) Sigma HPA007863 1:500 Growth Associated Protein (GAP43) Sigma Clone 7B10 prod # G9264 1:50000 Von Willebrand Factor (vWF) Novocastra Labs NCL-vWFp 1:1600 Nitric oxide synthase endothelial (eNOS) Santa Cruz sc-654 1:1000

Mann-Whitney test was used for statistical analysis (P values < 0.05 were considered statistically significant).

Results

Histology Routine histology of atrial tissues showed good preserva- tion and typical cardiomyocytes in all samples (Figure 1A, B). In serial sections, nerve fibres and fascicles immunoreactive for Nav1.8 were seen to course within and between cardiac muscle bundles as large and small nerve fascicles and some single nerve fibres within atrial connective tissue (Figure 1C- F). Innervation of myocardium Structural nerve markers: Atrial tissues were highly innervated as demonstrated by the nerve marker PGP9.5 which showed nerve fibres of large and small calibre throughout the myocardium (Figure 1G). A similar distribution of nerve fibres was detected using neurofilaments (neurofilament ‘cocktail’) although, as expected with this marker, these were fewer (Figure 1H). Image analysis of nerve markers showed no significant correlation with age. Nociceptive markers Sodium channels: Serial sections immunostained for vascular markers and Nav1.8 suggested a close relationship between sensory innervation and small myocardial capillaries (Figure 2A, B) and there was a significant correlation between the structural nerve marker neurofilaments and PGP9.5 (Figure

2C). The mean (± SEM) value for counts of Nav1.8-immunoreactive nerve fibres per mm2 in atrial myocardium was 0.85 (± 0.37) and there was no significant correlation with age. Counts of Na 1.8 fibres per mm2 significantly correlated with both v Figure 1. Standard H&E staining and immunostaining for structural vWF (P = 0.02) and eNOS (Figure 2D; P = 0.02). Immunore- nerve markers and Na 1.8. Typical low magnification staining of human activity for Na 1.8 was present also in cardiomyocytes with an v v atrium using H&E (A, B) and serial sections immunostained with Nav1.8 appearance similar to that seen with connexin43, whose im- (C-F). Examples of human atrium immunostained with antibodies to munoreactivity was concentrated at the perimeter of the inter- structural nerve markers PGP9.5 (G), neurofilaments (H). Scale bars = 200 μm A, C; 100 μm B, D; 50 μm E-H. calated disc or as vertical bands (Figure 3). The Nav1.8 immunoreactivity in cardiomyocytes was abolished by preincubation of Nav1.8 antibodies with homologous antigen. Immunostain- ing with antibodies to Nav1.5 showed no immunoreactivity in sections immunostained with the nerve marker PGP9.5. How- nerves despite the presence of abundant nerve fibres in serial ever, Nav1.5 immunoreactivity was present within cardiomyo- Vol 52 No 3 NAV1.8 AND PAIN-RELATED ION CHANNELS IN HEART 149

Figure 2. Nav1.8-immunoreactive nerve fibres in the myocardium (A) follow the same trajectory (arrowed) as vascular capillary in serial section (B) immunostained with vascular marker eNOS. Image analysis (% area per mm2) of structural nerve marker neurofilaments and PGP9.5 (C), and counts of Nav1.8 and eNOS (D) are significantly correlated. Scale bars = 50 μm.

Figure 4. Nociceptive nerve markers in atrium. Human atrium immuno-

stained for sodium channels Nav1.7 (A), and Nav1.9 (B), TRPV1 (C),

P2X3 (D) and P2X2 (E). Arrows indicate fine calibre nerve fibres. Scale bars = 50 μm

tissue were fewer than seen with Nav1.8. Since Nav1.7 and Nav1.9-immunoreactive nerve fibres were very few, counts per mm2 were not attempted. TRPV1: A few, distinct, single nerve fibres immunoreactive for TRPV1 were found between myocardial muscle fibres (Figure 4C). The sparse distribution of TRPV1-immunoerac- tive nerve fibres precluded image analysis. Purinergic nerve markers: Nerve fibres immunoreactive for

P2X3 and P2X2 were detected within nerve fascicles in myocardium and often associated with vascular structures (Figure 4D, E). These fibres were consistent with nerves detected using the structural nerve marker neurofilaments in serial sections. Sympathetic nerve markers Antibodies to neuropeptide NPY (CPON) revealed an abundance of mostly fine, small/medium calibre, varicose nerve fibres throughout the myocardium (Figure 5A). Image analysis of NPY immunoreactivity (% area) gave a mean (± SEM) value of 0.44 (± 0.11) and there was no significant cor- Figure 3. Nav1.8 and Nav1.5-immunoreactivity in cardiomyocytes. Hu- man atrium immunostained with antibodies to Na 1.8 (A, B) and serial relation with age. However, there was a significant (P = 0.002) v correlation of % area between CPON and the nerve marker sections immunostained with antibodies to connexin43 (C, D) or Nav1.5 (E, F). Scale bar = 50 μm. PGP9.5, but not with other markers. Other neural markers Botulinum toxin receptor; Synaptic vesicle protein 2A (SV2A): Immunostaining with SV2A antibodies showed fine, varicose cytes (Figure 3E, F). SV2A-immunoreactive nerve fibres with a similar distribution Nerve fibres and fascicles were detected with antibodies to NPY, and were present throughout the myocardium and to sodium channels Nav1.7 (Figure 4A) and Nav1.9 (Figure 4B) around vascular structures (Figure 5B). Image analysis of which showed a similar prominent distribution to Nav1.8 in SV2A immunoreactivity (% area) gave a mean (± SEM) value nerve fascicles but single nerve fibres in the atrial connective of 1.18 (± 0.22), and there was no significant correlation with Int Heart J 150 FACER, ET AL May 2011

nerve fibres. Interestingly, there was a significant correlation of 2 the number of Nav1.8-immunoreactive fibres per mm and the vascular marker eNOS. In addition, some Nav1.8-immunoreactive nerve fibres notably took the same course as small myocardial capillaries. These observations emphasize a close relationship between sensory nerve fibres and vascularity, also seen between C fibres and cardiac mast cells.38) Further studies are needed in a larger number of patients to investigate any re-

lationship of Nav1.8-immunoreactive nerve fibres and gene variation in cardiac ischaemia or pain.

Nav1.8-immunostaining was also present in atrial cardiac myocytes similar to that seen with connexins (40, 43 and 45), which are markers of the intercalated discs in which the specialised gap junctions between cardiac myocytes are locat- 39,40) ed. Nav1.5-immunoreactivity was abundant in cardiomyocytes but not detected in nerve fibres. While this is not likely to Figure 5. Nerve and vascular markers in human heart. Human atrium im- be relevant in the heart, the connexins, in particular Cx43, are munostained for neuropeptide Y (A), botulinum toxin receptor SV2A (B), present in glial cells in the peripheral nervous system, and may 41,42) GAP43 (C), and the vascular marker vWF (D). Scale bars = 50 μm play a role in the transmission of neuropathic pain. Nerves immunoreactive for Nav1.7 or Nav1.9 were present in larger fascicles adjacent to the epicardium, but were

fewer than Nav1.8, particularly in the myocardium, where they age. were too sparse to quantify. Studies of gene expression in

Growth associated protein (GAP43): The localisation and dis- mouse heart indicate a greater expression of Nav1.7 RNA in tribution of GAP43-immunoreactive nerve fibres was similar the sinoatrial and atrioventricular nodes in contrast to atrium,43) to that seen with SV2A (Figure 5C), but there was considera- which could provide an explanation for the relative lack of ble variation in distribution within the tissue sections, being Nav1.7-immunoreactive nerves in our samples. The distribu- more frequent in some regions and diffuse or absent in others. tion and frequency of Nav1.9-immunoreactive nerve fibres was For this reason image analysis or counting was not attempted. similar to that seen with Nav1.7, and there was no detectable Vascular markers variation between samples. The sodium channel Nav1.9 is Von Willebrand Factor (vWF) and endothelial nitric oxide syn- thought to play a significant role in the maintenance of inflam- 44,45) thase (eNOS): Immunoreactivity for vWF was intense, abun- matory pain, and increased Nav1.9-immunoreactive nerves dant (Figure 5D) and detected in all vascular structures have been reported in inflammation-related diseases such as throughout all atrial samples. Immunoreactivity with eNOS rhinitis and arthritis.8,46) Our atrial tissues showed little or no gave a similar distribution to vWF but was less intense. The evidence of inflammation. difference of intensity was reflected in the image analysis (% Nerve fibres immunoreactive for other sensory nerve area) of vascular markers where the mean (± SEM) values for markers included TRPV1 and the purinergic receptors P2X3 vWF and eNOS were 9.14 (± 0.36) and 4.15 (± 0.58) respec- and P2X2. The production of bradykinin during cardiac ischae- tively. There was no significant correlation with age although mia may sensitise cardiac nociceptive sensory nerve endings both markers correlated with counts of Nav1.8 nerve fibres, as responsible for pain. Antagonists of TRPV1 have been shown described above. to reduce cardiac spinal afferent nerve excitation in response to bradykinin suggesting that TRPV1 fibres may be responsible for the detection of ischaemia.18) TRPV1-immunoreactive Discussion nerve fibres have been demonstrated on the epicardial surface of the ventricle, sparse in deeper myocardium, and depleted af- We have recently reported1) that genetic variation in SC- ter resiniferatoxin (RTX) treatment in rat.19) Epicardial TRPV1 N10A influences cardiac conduction (PR, P-wave and QRS in- fibres were not detected in human atrium; this may reflect a tervals), and that rs6795970, a common gain-of-function vari- species difference, otherwise the sparse distribution of TRPV1 ant, is a susceptibility factor for heart block and serious in human atrial myocardium is similar. We have shown that ventricular arrhythmia. SCN10A encodes Nav1.8, an alpha sub- nerve fascicles near to the epicardium contained fibres immu- unit of voltage-gated sodium channels. The sodium channel noreactive for purinergic receptors P2X3 and P2X2. There was Nav1.8 is known to play a key role in the propagation of action no evidence of immunoreactivity in myocytes, which have potentials in nociceptive (pain) nerve fibres in animal models, been shown to express P2X4, a purine receptor involved in 47) and we have previously shown increased levels of Nav1.8 in contractility. In the rat heart, P2X3 and P2X2 receptors have tissues from patients in a number of pain and inflammatory been shown to be coexpressed in nodose ganglion neurons.21) conditions.7,8,10) ATP can be released when there is tissue damage, and may act We have now demonstrated, for the first time in any spe- as a signal for nociceptive sensory neurons. Activation of car- cies, Nav1.8-immunoreactive sensory (nociceptor) nerve fibres diac sensory afferents from the nodose ganglion may lead to in human atrial myocardium. The density of these fibres was the perception of pain. relatively low compared to the pan-neuronal structural marker Diabetic cardiomyopathy involves contractile cardiomyo- PGP9.5, suggesting its expression by a sub-set of nociceptor cytes and the sensory and autonomic innervation of the heart.48) Vol 52 No 3 NAV1.8 AND PAIN-RELATED ION CHANNELS IN HEART 151

Cardiac nerves contain a broad spectrum of neurotransmitters and their roles in electrogenesis within dorsal root ganglion neu- including numerous neuropeptides, the most abundant is neurons. J Physiol 2007; 579: 1-14. (Review) ropeptide Y (NPY), a 36 amino acid peptide produced by 5. Laird JM, Souslova V, Wood JN, Cervero F. Deficits in visceral 49) pain and referred hyperalgesia in Nav1.8 (SNS/PN3)-null mice. J cleavage from a large precursor, preproNPY. NPY (CPON)- Neurosci 2002; 22: 8352-6. immunoreactive fibres have been demonstrated previously in 6. Zimmermann K, Leffler A, Babes A, et al. Sensory neuron sodium heart as being among the most abundant in atrium, although channel Nav1.8 is essential for pain at low temperatures. Nature they were not quantified.29) We found an abundance of NPY- 2007; 447: 855-8. immunoreactivity in fine calibre nerve fibres in most of our 7. Coward K, Plumpton C, Facer P, et al. Immunolocalization of samples, and quantification of these correlated with the struc- SNS/PN3 and NaN/SNS2 sodium channels in human pain states. tural nerve marker PGP 9.5. We found no difference in NPY- Pain 2000; 85: 41-50. 8. Keh SM, Facer P, Simpson KD, Sandhu G, Saleh HA, Anand P. immunoreactive nerves (% area) between atria from diabetics Increased nerve fiber expression of sensory sodium channels and non-diabetics in the limited numbers of patients studied, Nav1.7, Nav1.8, And Nav1.9 in rhinitis. Laryngoscope 2008; 118: and these results are in agreement with experimentally induced 573-9. diabetes where NPY was diminished in the ventricle but un- 9. Shembalkar PK, Till S, Boettger MK, et al. Increased sodium changed in atrium.50) In ischaemic conditions NPY stimulates channel SNS/PN3 immunoreactivity in a causalgic finger. Eur J neovascularisation,51) although we found no correlation be- Pain 2001; 5: 319-23. 10. Renton T, Yiangou Y, Plumpton C, Tate S, Bountra C, Anand P. tween NPY immunoreactive nerve (% area) and vascular Sodium channel Nav1.8 immunoreactivity in painful human den- markers (vWF and eNOS). tal pulp. BMC Oral Health 2005; 5: 5. GAP43-immunoreactive nerve fibres were variable in 11. Liu CJ, Dib-Hajj SD, Renganathan M, Cummins TR, Waxman most atrial samples. In normal canine atrial appendage SG. Modulation of the cardiac sodium channel Nav1.5 by fibro- GAP43-immunoreactive fibres are of low density until repeat- blast growth factor homologous factor 1B. J Biol Chem 2003; ed electrical stimulation induces nerve sprouting.52) The rela- 278: 1029-36. tive abundance of GAP43-immunoreactive nerve fibres in our 12. Wu L, Nishiyama K, Hollyfield JG, Wang Q. Localization of Nav1.5 sodium channel protein in the mouse brain. Neuroreport specimen suggests an ongoing regeneration. A time-course 2002; 13: 2547-51. study of GAP-43 nerve fibres in disease states is feasible, and 13. Nikolaev VO, Moshkov A, Lyon AR, et al. Beta 2-adrenergic re- may demonstrate a relationship to pathophysiological mecha- ceptor redistribution in heart failure changes cAMP compartmen- nisms following ischaemia or myocardial infarction. tation. Science 2010; 327: 1653-7. The highest density of innervation was obtained with the 14. Korchev YE, Negulyaev YA, Edwards CR, Vodyanoy I, Lab MJ. structural neuronal marker PGP9.5, as expected. Our median Functional localization of single active ion channels on the surface of a living cell. Nat Cell Biol 2000; 2: 616-9. values for % area (PGP9.5; CPON) of immunostaining ob- 15. Severs NJ. The role of connexins in the heart. In: Connexins:A tained for atrial samples are lower than those described earlier guide. Harris AI, Locke, D, (ed). Humana Press Vol; 2009: 429- for normal human atrial myocardium adjacent to the AV 50. node.31) In our study we have used samples from diseased 16. Vozzi C, Dupont E, Coppen SR, Yeh HI, Severs NJ. Chamber-re- hearts, which may have contributed to lower values. lated differences in connexin expression in the human heart. J Mol In summary, we present evidence that heart innervation Cell Cardiol 1999; 31: 991-1003. expresses several nociceptive ion channels including Na 1.7, 17. Dupont E, Ko Y, Rothery S, et al. The gap-junctional protein con- v nexin40 is elevated in patients susceptible to postoperative atrial Nav1.8 and Nav1.9 and TRPV1 and P2X3. Our results may pro- fibrillation. Circulation 2001; 103: 842-9. vide a basis for the further study and understanding of pain in 18. Pan HL, Chen SR. Sensing tissue ischemia: another new function cardiac disorders, and they indicate that clinical trials with an- for capsaicin receptors? Circulation 2004; 110: 1826-31. tagonists of these ion channels may be warranted in chronic 19. Zahner MR, Li DP, Chen SR, Pan HL. Cardiac vanilloid receptor intractable cardiac pain. 1-expressing afferent nerves and their role in the cardiogenic sympathetic reflex in rats. J Physiol 2003; 551: 515-23. 20. Burnstock G. Pathophysiology and therapeutic potential of purinergic signaling. Pharmacol Rev 2006; 58: 58-86. (Review) Acknowledgments 21. Wan F, Li G, Liu S, et al. P2X(2/3) receptor activity of rat nodose ganglion neurons contributing to myocardial ischemic nociceptive We thank Dr Pervez Ahmed and Max Healthcare, Delhi, India for signaling. Auton Neurosci 2010; 158: 58-64. their advice and support. 22. Dong M, Yeh F, Tepp WH, et al. SV2 is the protein receptor for botulinum neurotoxin A. Science 2006; 312: 592-6. 23. Baker DG, Coleridge HM, Coleridge JC, Nerdrum T. Search for a cardiac nociceptor: stimulation by bradykinin of sympathetic af- References ferent nerve endings in the heart of the cat. J Physiol 1980; 306: 519-36. 1. Chambers JC, Zhao J, Terracciano CM, et al. Genetic variation in 24. Kuo DC, Oravitz JJ, DeGroat WC. Tracing of afferent and efferent SCN10A influences cardiac conduction. Nat Genet 2010; 42: 149- pathways in the left inferior cardiac nerve of the cat using retro- 52. grade and transganglionic transport of horseradish peroxidase. 2. Catterall WA. From ionic currents to molecular mechanisms: the Brain Res 1984; 321: 111-8. structure and function of voltage-gated sodium channels. Neuron 25. Pan HL, Longhurst JC, Eisenach JC, Chen SR. Role of protons in 2000; 26: 13-25. (Review) activation of cardiac sympathetic C-fibre afferents during ischae- 3. Catterall WA, Goldin AL, Waxman SG. International Union of mia in cats. J Physiol 1999; 518: 857-66. Pharmacology. XLVII. Nomenclature and structure-function rela- 26. Immke DC, McCleskey EW. ASIC3: a lactic acid sensor for cardi- tionships of voltage-gated sodium channels. Pharmacol Rev 2005; ac pain. ScientificWorldJournal 2001; 1: 510-2. (Review) 57: 397-409. (Review) 27. Sutherland SP, Benson CJ, Adelman JP, McClesky EW. Acid- 4. Rush AM, Cummins TR, Waxman SG. Multiple sodium channels sensing ion channel 3 matches the acid-gated current in cardiac Int Heart J 152 FACER, ET AL May 2011

ischemia-sensing neurons. Proc Natl Acad Sci U S A 2001; 98: 218: 13-28. (Review) 711-6. 41. Jasmin L, Vit JP, Bhargava A, Ohara PT. Can satellite glial cells be 28. Hisata Y, Zeredo JL, Eishi K, Toda K. Cardiac nociceptors inner- therapeutic targets for pain control? Neuron Glia Biol 2010; 6: 63- vated by vagal afferents in rats. Auton Neurosci 2006; 126-127: 71. 174-8. 42. Ohara PT, Vit JP, Bhargava A, Jasmin L. Evidence for a role of 29. Wharton J, Polak JM, Gordon L, et al. Immunohistochemical connexin 43 in trigeminal pain using RNA interference in vivo. J demonstration of human cardiac innervation before and after Neurophysiol 2008; 100: 3064-73. transplantation. Circ Res 1990; 66: 900-12. 43. Marionneau C, Couette B, Liu J, et al. Specific pattern of ionic 30. Gordon L, Polak JM, Moscoso GJ, Smith A, Kuhn DM, Wharton J. channel gene expression associated with pacemaker activity in the Development of the peptidergic innervation of human heart. J mouse heart. J Physiol 2005; 562: 223-34. Anat 1993; 183: 131-40. 44. Leo S, D’Hooge R, Meert T. Exploring the role of nociceptor-spe- 31. Crick SJ, Wharton J, Sheppard MN, et al. Innervation of the hu- cific sodium channels in pain transmission using Nav1.8 and man cardiac conduction system. A quantitative immunohisto- Nav1.9 knockout mice. Behav Brain Res 2010; 208: 149-57. chemical and histochemical study. Circulation 1994; 89: 1697- 45. Ritter AM, Martin WJ, Thorneloe KS. The voltage-gated sodium 708. channel Nav1.9 is required for inflammation-based urinary blad- 32. Duchen LW, Anjorin A, Watkins PJ, Mackay JD. Pathology of au- der dysfunction. Neurosci Lett 2009; 452: 28-32. tonomic neuropathy in diabetes mellitus. Ann Intern Med 1980; 46. Strickland IT, Martindale JC, Woodhams PL, Reeve AJ, Chessell 92: 301-3. IP, McQueen DS. Changes in the expression of NaV1.7, NaV1.8 33. Faerman I, Faccio E, Milei J, et al. Autonomic neuropathy and and NaV1.9 in a distinct population of dorsal root ganglia inner- painless myocardial infarction in diabetic patients. Histologic evi- vating the rat knee joint in a model of chronic inflammatory joint dence of their relationship. Diabetes 1977; 26: 1147-58. pain. Eur J Pain 2008; 12: 564-72. 34. Cao JM, Fishbein MC, Han JB, et al. Relationship between re- 47. Hu B, Mei QB, Yao XJ, Smith E, Barry WH, Liang BT. A novel gional cardiac hyperinnervation and ventricular arrhythmia. Circu- contractile phenotype with cardiac transgenic expression of the lation 2000; 101: 1960-9. human P2X4 receptor. FASEB J 2001; 15: 2739-41. 35. Vracko R, Thorning D, Frederickson RG. Nerve fibers in human 48. Dhalla NS, Liu X, Panagia V, Takeda N. Subcellular remodeling myocardial scars. Hum Pathol 1991; 22: 138-46. and heart dysfunction in chronic diabetes. Cardiovasc Res 1998; 36. Severs NJ, Bruce AF, Dupont E, Rothery S. Remodelling of gap 40: 239-47. (Review) junctions and connexin expression in diseased myocardium. Car- 49. Tatemoto K, Carlquist M, Mutt V. Neuropeptide Y--a novel brain diovasc Res 2008; 80: 9-19. (Review) peptide with structural similarities to peptide YY and pancreatic 37. Shu SY, Ju G, Fan LZ. The glucose oxidase-DAB-nickel method polypeptide. Nature 1982; 296: 659-60. in peroxidase histochemistry of the nervous system. Neurosci Lett 50. Chottová Dvoráková M, Wiegand S, Pesta M, et al. Expression of 1988; 85: 169-71. neuropeptide Y and its receptors Y1 and Y2 in the rat heart and its 38. Morrey C, Brazin J, Seyedi N, Corti F, Silver RB, Levi R. Interac- supplying autonomic and spinal sensory ganglia in experimentally tion between sensory C-fibers and cardiac mast cells in ischemia/ induced diabetes. Neuroscience 2008; 151: 1016-28. reperfusion: activation of a local renin-angiotensin system culmi- 51. Lee EW, Michalkiewicz M, Kitlinska J, et al. Neuropeptide Y in- nating in severe arrhythmic dysfunction. J Pharmacol Exp Ther duces ischemic angiogenesis and restores function of ischemic 2010; 335: 76-84. skeletal muscles. J Clin Invest 2003; 111: 1853-62. 39. Severs NJ, Dupont E, Thomas N, et al. Alterations in cardiac con- 52. Swissa M, Zhou S, Tan AY, Fishbein MC, Chen PS, Chen LS. nexin expression in cardiomyopathies. Adv Cardiol 2006; 42: 228- Atrial sympathetic and parasympathetic nerve sprouting and hy- 42. (Review) perinnervation induced by subthreshold electrical stimulation of 40. Desplantez T, Dupont E, Severs NJ, Weingart R. Gap junction the left stellate ganglion in normal dogs. Cardiovasc Pathol 2008; channels and cardiac impulse propagation. J Membr Biol 2007; 17: 303-8.