Tickling the Heart and the Nervous System with the Funny Current

Since their discovery in sinoatrial node cells and neurons in the late 1970s and early 1980s, Hyperpolarization-Activated Cyclic Nucleotide-Gated (HCN) channels have sparked continuing interests. The use of Alomone Labs antibodies directed towards HCN channels provides an extensive research tool for studying the importance of these channels regarding structure-function correlations, elucidating their location in various tissues, identifying interaction sites with various proteins and elucidating mechanisms underlying pathology and diseases caused by specific defects in HCN channels.

Introduction

HCN channels comprise a small subfamily of proteins within the superfamily of pore-loop cation channels. In mammals, the HCN channel family includes four members (HCN1-4) that are expressed in the heart and in the nervous system. Hyperpolarization-activated cation currents produced by HCN channels (I_h, I_f or I_q), contribute to a wide range of physiological functions including cardiac and neuronal pacemaker activity, the setting of resting potentials, dendritic integration of synaptic transmission and learning^8,42. Throughout this article I_f refers to cardiac currents and I_h refers to non-cardiac currents.

HCN channels are unique in three ways: i) channel gating by membrane hyperpolarization. In general, I_h currents activate with hyperpolarizing steps to potentials negative to -50 to -60 mV. Unlike most voltage-gated currents, I_h does not display voltage-dependent inactivation. ii) I_h is regulated by cyclic nucleotides (cAMP). iii) I_h is a mixed cation current that is carried by both K⁺ and Na⁺ under normal physiological conditions⁸.

In the heart, HCN1 and HCN2 are somewhat expressed, while HCN4 is the most abundant subunit. In the mammalian brain, all four HCN isoforms are expressed and are particularly wellcharacterized in the hippocampus. The four HCN subunits can form either homo or hetero tetrameric structures that are arranged around the centrally located pore. HCN channels can also coassemble with auxiliary subunits to form channels with modified properties⁸.

Developmental Regulation

Although I_f is an important component of spontaneous activity in the embryonic heart, the contribution of this current to the differentiation of embryonic stem (ES) cells into cardiac cells remains unknown.

Changes in the expression and function of HCN channels during the differentiation of cardiac precursor cells derived from mouse ES cells were examined. HCN1 and HCN4 were predominantly expressed using Anti-HCN1 (#APC-056) and Anti-HCN4 (#APC-052) antibodies along with I_f currents which were predominantly observed in cardiac precursor cells on day 15⁵² (Figure 1).

Another work used spontaneously beating cells derived from mouse embryonic stem cells (mESCs) in culture. Using Anti-HCN2 (#APC-030) and Anti-HCN3 (#APC-057) antibodies it was demonstrated that the proportion and density of I_f increased during development. Results from western blot analysis are consistent with the presence of HCN2 and HCN3 suggesting that I_f confers regular and faster rhythmicity in mESCs⁴⁰.

While all four HCN channels are expressed in the cerebellum, the contribution of the individual subunits to the proposed cerebellar functions of I_h is currently unknown. The presented study investigated the developmental profile, the selective distribution pattern and the cellular and subcellular localization of HCN1 in the cerebellum²² (Figure 2). Using Anti-HCN1, the authors found that HCN1 is indeed expressed in the rat cerebellum at P60 (postnatal day 60) (Figure 2A). In an attempt to study the expression of HCN1 during development, rat brain membranes were extracted (P0-P60) and subjected to western blot analysis. HCN1 expression became evident at P5 and peaked at P15 (Figure 2B). Confocal microscopy of immunohistochemical studies showed similar results and suggested that HCN1 is localized in axon terminals of cerebellar basket cells (Figure 2C). HCN1 expression is also developmentally regulated in the integrated compartments of the layer 5 pyramidal neurons as determined in immunohistochemical analysis using Anti-HCN1⁵. HCN2 expression was also found to be developmentally regulated in the hippocampus using its respective Alomone Labs antibody⁴⁸.

Distribution and Function Studies

The distribution of HCN channels in the SAN was investigated using histology and immunolabeling applications. Anti-HCN1, Anti-HCN2 and Anti-HCN4 antibodies revealed that HCN4 is the only HCN isoform detectable in SAN cells. Strands of HCN4-positive nodal cells also protruded into the atrial muscle (Figure 3). The data suggest that this specialized interface between the SAN and the surrounding atrium may be necessary for the SAN to drive the more hyperpolarized atrial muscle²¹. In a different study, immunolabeling using Anti-HCN4 antibody demonstrated that the cells in the central SAN area are predominantly distributed in islets interconnected by cell prolongations (Figures 4A and 4B) and single-cell HCN4 labeling suggested sites of channel clustering (Figure 4C). It was therefore proposed that expression of HCN4 is an appropriate tool to map and identify the cardiac SAN pacemaker region⁹.

The distribution of HCN4 using Anti-HCN4 was studied in the center, periphery and atrial muscles of the mouse SAN pacemaker cell area. In addition to HCN4, two Na⁺ channel isoforms were found important for proper SAN function: neuronal voltage-gated Na⁺ channel (putatively Na_v1.1) and cardiac Na_v1.5 isoforms¹⁸.

The possibility that both HCN4 and HCN1 isoforms contribute to the native I_f in SAN cells by forming heteromeric channels was explored in HEK293 cells, using Anti-HCN4 antibody. Co-transfected (HCN4 + HCN1) and various chimeric constructs with N- and C- termini of either channel were compared with those of the native I_f current. Results indicate that HCN1 and HCN4 may contribute to native I_f channels, but a ‘context’-dependent mechanism is also likely to modulate the channel properties in native tissues³. MiRP1 was also found to contribute to modulate these currents (determined using Anti-KCNE2 (MiRP1) antibody (#APC-054)).

The co-assembly of HCN2 and HCN4 in live CHO cells was studied using bioluminescence resonance energy transfer (BRET2), a novel approach for studying tetramerization of ion channel subunits. Together with results from electrophysiological and imaging approaches, BRET2 data showed that HCN2 and HCN4 subunits self-assemble and co-assemble with equal preference. The colocalization of HCN2 and HCN4 in the embryonic mouse heart in immunohistochemistry studies was also demonstrated using Anti-HCN2 and Anti-HCN4 antibodies. Together, these data support the formation of HCN2-HCN4 heteromeric channels in native tissues⁵¹.

It was recently discovered that constitutively active Src tyrosine kinase can enhance HCN4 channel activity by binding to the channel. In order to investigate the mechanism of HCN4 modulation by Src, the effects of PP2, a selective inhibitor of Src, on HCN4 was studied by using Anti-HCN4 antibody in immunoprecipitation and western blot analyses in HEK 293 cells. Results suggest that Src enhances HCN4 currents via Tyr531¹⁹.

Cysteine scanning substitutions were introduced into the descending portion of the HCN1 P-loop. Although expressed (as demonstrated by western blot analysis using Anti-HCN1), all mutated channels were devoid of functional currents. It was concluded that specific pore residues are important determinants of the structure-function properties of HCN channels in general⁶. Cysteine scanning substitutions were also introduced in the P-S6 linker of HCN1. The novel residues identified display extracellular accessibility which is steric and state dependent. These data provide the first functional evidence consistent with a pore-to-gate model for HCN channels⁴⁵.

Alomone Labs Anti-HCN2 and Anti-HCN4 antibodies were both used in western blot and immunoprecipitation experiments and revealed that in the myocardium, the molecular mass of HCN2 is significantly lower, lacking the C-terminus cAMP binding domain. In heterologous cells, the C-terminal-truncated HCN2 protein co-assembles with HCN4 to form functional heteromeric HCN channels, which activate faster than homomeric HCN2 or HCN4 channels, and display properties similar to endogenous myocardial I_f channels. Taken together, the results suggest that functional myocardial I_f channels reflect the heteromeric assembly of HCN2 and HCN4 and furthermore, that HCN4 underlies the cAMP-mediated regulation of cardiac I_f channels⁵³. Heteromeric association between HCN2 and HCN4 was also demonstrated in Xenopus oocytes with Anti-HCN2 and Anti-HCN4. These structures also displayed faster activation kinetics than the homomeric channels⁵⁵.

Another work directly investigated the role of cAMP-dependent regulation of HCN channels in SAN activity. Mice with heart-specific and inducible expression of a human HCN4 mutation (abolishing the cAMP-dependent regulation) were generated. The expression of HCN4 was assessed by western blot analysis, using Anti-HCN4 antibody. This mutant mouse model suggests that cAMP-mediated regulation of HCN determines basal and maximal heart rates but does not play an indispensable role in heart rate adaptation during physical activity².

Another work investigated biological pacemakers (BPM) implanted in canine, which function competitively with electronic pacemakers (EPM). The engineered BPM (BPM/EPM tandem) was implanted in rats mutated in HCN2 (thereby mimicking arrhythmia) and found that the BPM/EPM tandems confer advantage over either approach used on its own. In this study, Anti-HCN2 antibody was used to show that mutated rat cardiomyocytes express much lower levels of HCN2 than wild-type¹⁰.

In the ear, the axis of the cochlea encloses the cell bodies of the spiral ganglion cells (SGCs). Type I and type II SGCs are responsible for afferent innervation of the inner and outer hair cells of the cochlea respectively. HCN channels and I_h currents have been identified in type I SGCs and I_h currents have been observed in type II SGCs^11,14,29. Recently, all four HCN4 subunits were identified in both types of SGCs⁷ . Western blot analysis, using whole spiral ganglion lysates was done in order to verify the specificity of the antibodies. As shown in Figure 5, Anti-HCN2 and Anti-HCN3 detected specific bands (Figure 5A). Fluorescent immunohistochemistry applied on cochlear free-floating preparations as well as wax-embedded tissues showed that HCN2 and HCN3 immunoreactivities are present on the cell bodies as well as the processes (Figures 5B and 5C).

Pyramidal cells are one of the most prominent cell types in the dorsal cochlear nucleus. Immunocytochemical labeling identified HCN1, HCN2 and HCN4 from acutely isolated dorsal cochlear nuclear neurons using Alomone Labs antibodies and all three channels were found expressed by pyramidal cells³⁵. In addition, HCN1 distribution was found to overlap with that of K_v1.1 in cellular regions of the cochlear nucleus³⁴.

The medial septum/diagonal band (MS/DB) iis a region in the brain which initiates and/or modulates the hippocampal theta rhythm, important for memory and learning^46,49. Although in situ hybridization studies did show expression of HCN1, 2, and 4 in MS/DB^30,41, the extent of the overlapping expression of these channels is not clear. In accordance with the mRNA detection of all three channels in MS/DB^30,41, immunohistochemical analysis using Anti-HCN1, 2, and 4 antibodies, showed that all three channels are expressed in MS/DB nuclei³¹. HCN1 detection was strong in somatic and proximal dendritic surface membrane of neurons. Expression of HCN2 was similar to that of HCN1 but mostly cytoplasmic. HCN4 on the other hand had a more widespread staining.

I_h currents are also present in the soma of sensory neurons that do not exhibit any of the pacemaker or oscillatory potentials. The aforementioned neurons were used to study the expression of HCN1, 2, and 4, using their respective Alomone Labs antibodies in immunohistochemical studies¹². HCN1 immunoreactivity was confined to only a subpopulation of nodose neurons, whereas HCN2 and HCN4 were expressed in all neurons of the nodose ganglion. Inhibiting I_h in the sensory neurons rendered the neurons more excitable and no role for oscillatory behavior was found.

While HCN1-4 expression has been described in the thalamus, HCN2 and HCN4 are the major isorforms in the glutamatergic ventrobasal complex (VB) and the GABAergic reticular thalamic nucleus (RTN)³³. However, their exact compartmental location is still unclear. Immunohistochemical studies with AntiHCN2 and Anti-HCN4 showed that they are individually expressed in both RTN and VB neurons, and mostly overlap in the VB¹. HCN2 but not HCN4 was distributed in dendritic spines in both RTN and VB neurons, important in the regulation of neuronal excitability and synaptic plasticity¹. The role of HCN2 in RTN neurons was therefore investigated54. I_h was abolished in Hcn2^{-/ –} cells strongly suggesting that HCN2 is the major contributor to I_h currents in RTN. Confocal microscopy images demonstrated punctate and sparse somatic labeling of HCN2 immunoreactivity in wild-type cells and was absent in all cellular compartments in Hcn2^-/- cells. On the other hand, Hcn2 deletion did not affect HCN4 somatic immunoreactivity (Figure 6). The electrophysiological data in the study strongly suggested that I_h mediated by HCN2 acts as a leak current path in order to “turn off” excitatory inputs. This implies that HCN2 and glutamate receptor subunit GluR4 may be in close proximity, as they are both expressed in RTN neurons⁵⁴.

All four HCN channels have unique but overlapping expression patterns in the CNS. The distribution of HCN1 in the brainstem and spinal cord was examined in immunohistochemical studies²⁸. At all levels of the spinal dorsal horn, HCN1 immunoreactivity was quite dense in laminae III. In the brainstem, HCN1 expression was included in neurons in sensory pathways as well as motoneurons.

HCN channels are abunddantly expressed in DRGs. A study was initiated to see whether HCN2 is localized in the DRG soma or whether it is transported from the soma to the central axons of the DRG neurons that terminate in the spinal dorsal horn⁴ (Figure 7). HCN2 was located in primary sensory neurons in the superficial spinal dorsal horn of rats, more specifically in laminae I and outer laminae II (Figure 7A). Furthermore, dorsal rhyzotomy completely abolished the immunoreactivity of HCN2 (Figure 7B), strongly suggesting that HCN2 is confined to axon terminals of nociceptive primary afferents. To further expand this notion, HCN2 was found to colocalize with CGRP, a neuropeptide expressed by peptidergic nociceptive primary afferents (Figure 7C). Having gained insight on the localization of HCN2 in nociceptive primary afferents, a subsequent study was done to investigate the functional significance of HCN2 in nociceptive primary afferents³⁶. HCN2 colocalized with Substance P (SP), suggesting that HCN2 may be involved in pain transmission from SPcontaining nociceptive primary afferents to spinal neurons (Figure 7D).

Afferent pathways innervating the urinary bladder consist of dorsal root ganglion (DRG) neurons (medium- and small-sized cell populations). Since I_h currents have been identified in various peripheral sensory neurons, the expression of HCN channels in bladder afferent neurons from L6-S1 spinal cord DRGs was examined25. Using Alomone Labs Anti-HCN1, 2 and 4 antibodies, HCN2 showed the most intense immunoreactivity in DRGs. In contrast to HCN1, HCN2 is found in laminae I and II of the dorsal horn, whereas HCN1 is found in laminae III²⁸ as mentioned above.

Modulating I_h and I_k-K-leak (mediated by TASK K⁺ channels) plays an important role in determining the resting potential. Thalamocortical relay (TC) neurons are an ideal model for understanding the interaction between I_h and I_k-K-leak, and resting membrane potential. Immunohistochemical studies showed that 100% of TC neurons are positive for both HCN2 and TASK3 and that both currents contribute to the resting membrane potential²⁶.

HCN and Protein-Protein Interactions

Channel activity depends in many cases on the interaction with auxiliary subunits and scaffolding proteins. HCN channel subunits are believed to be modulated by additional regulatory proteins, which only recently have been identified.

MinK-related protein (MiRP1 or KCNE2) was previously found to interact with the HCN family of pacemaker channels, and to alter channel gating in heterologous expression systems. In order to evaluate the effect of MiRP1 on HCN expression and function in a physiological context, overexpressed HA-tagged MiRP1 (HA-MiRP1) and HCN2 expressed in neonatal rat ventricular myocytes was investigated. Coimmunoprecipitation experiments using AntiHCN2 antibody demonstrated that HA-MiRP1 and HCN2, as well as endogenous MiRP1 and HCN2, co-assemble in ventricular myocytes. The results indicate that MiRP1 acts as a β subunit for HCN2 and alters channel gating in cardiac cells³⁹.

A protein-protein interaction between HCN2 and the K⁺ channel regulator protein 1 (KCR1) was revealed. Coimmunoprecipitation experiments, using Anti-HCN2 antibody, showed that KCR1 and HCN2 associate. The results further showed that KCR1 modulates I_HCN2/I_f channel gating in cardiac cells²⁷. In addition, usage of Anti-HCN2 and Anti-HCN4 antibodies showed that both HCN2 and HCN4 isoforms co-distribute with the adapter protein SAP97, an important component in the SAN. HCN4, but not HCN2, also co-distributes with the post-synaptic marker β-catenin, thus identifying diverse organized domains within this tissue. Together, their data suggest that SAP97 contributes to isoform specific organization of HCN channels within specific domains in the SAN³⁷.

In the brain, a yeast two hybrid screen showed that HCN2 interacts with Tamalin, a scaffold protein, through the PDZ-binding motif¹⁵. These two proteins indeed interact as they could both be immunoprecipitated using Anti-HCN2 antibody. In addition, immunoprecipitating rat brain membrane showed that along with HCN2, S-SCAM and Mint2, two neuronal scaffold proteins, are both pulled down with the channel. These scaffold proteins may unequivocally direct proper localization, influence trafficking and signal transduction of HCN channels in neuronal cells. In a different yeast two hybrid screen, Filamin A, also a scaffold protein was found to interact with the C-terminal of HCN1¹³. in vivo interaction was also demonstrated using Anti-HCN1 antibody. Filamin A may be involved in clustering HCN1¹³. Pex5R was first identified via a yeast two hybrid screen as an HCN interacting protein⁴³. In a recent study, Pex5R was also identified as an HCN interacting protein when Anti-HCN2 antibody was used to immunoprecipitate the channel from membrane fractions that were prepared from rat brain⁵⁷. The role of Pex5R is actually to slow the time course of HCN channel activation by allosterically hindering cAMP binding in the brain.

HCN, Disease and Channelopathies

The overexpression of I_f has been functionally demonstrated in ventricular myocytes in failing human hearts. The molecular basis however of I_f overexpression in human cardiac disease is still unknown.

Little is known about the expression of HCN channels in HCN remodeling by diseases, such as congestive heart failure (CHF), associated with disturbances of cardiac rhythm. The expression of HCN1, HCN2 and HCN4 in normal dogs was assessed compared to dogs subjected to 2-week ventricular tachypacing-induced CHF. Western blot and immunohistochemistry analyses, using Anti-HCN2 and Anti-HCN4 antibodies showed that HCN2 and HCN4 expression are greater at both protein and mRNA levels in the SAN than in the right atrium (RA). In the RA, HCN4 expression but not that of HCN2 is significantly upregulated by CHF. The overall findings provide new information about the molecular basis of normal and disease-related impairments of cardiac impulse formation⁵⁶.

A missense mutation in the ion channel pore domain of HCN4 is associated with inherited sinus bradycardia. Expression of the wild-type and mutant HCN4 channel was assessed in HEK293 cells using Anti-HCN4. Results show that cells transfected with mutated HCN4 express much lower levels than those transfected with wildtype. Despite its critical location, this mutation carries a favorable prognosis without the need for pacemaker implantation during long-term follow-up³².

The mechanisms involved in the impairment of arterial baroreflex in type-1 diabetes (T1D), have yet to be deciphered. The nodose ganglion (NG) contains the cell bodies of the aortic baroreceptor (AB) neurons which express HCN channels important for regulating the cell excitability. Immunofluorescence and western blot analyses using Anti-HCN1 and Anti-HCN2 antibodies showed that the expression of HCN1 and HCN2 is higher in T1D rats compared to wild-type. The results indicate that HCN channels contribute to the decreased excitability of AB neurons in T1D rats²⁰.

The genetic and molecular mechanisms underlying human sinus node dysfunction (SND) are limited. Penetrant and severe SND was mapped to the human ANK2 (ankyrin-B/AnkB) locus. AnkB is essential for normal membrane organization of SAN cell channels and transporters, and is also required for physiological cardiac pacing. Anti-HCN4 antibody was used as a SAN marker in immunohistochemistry in AnkB^+/- and AnkB^+/+ SAN cells from mouse sections (Figure 8). Their findings reveal abnormal channel targeting (but not HCN channels) with SND and highlight the critical role of local membrane organization for SAN excitability¹⁷.

Enhanced abnormal automaticity of ventricular cells contributes to hypertrophic arrhythmias. The expression of HCN2 and HCN4 is reportedly increased in hypertrophic and failing hearts, contributing to arrhythmogenesis. A study on post-transcriptional regulation of HCN2 and HCN4 by microRNAs was initiated in the rat model of left ventricular hypertrophy. The use of Anti-HCN2 and Anti-HCN4 antibodies in western blot analysis revealed the robust increase in HCN2 and HCN4 levels in a rat model of ventricular hypertrophy, accompanied by pronounced reduction of microRNA levels. Downregulation of HCN2 and HCN4 targeted microRNAs contributes to the re-expression of HCN2/HCN4 and thereby the electrical remodeling process in hypertrophic hearts²⁴.

The expression of HCN was assessed in ventricular and atrial samples from normal or failing hearts explanted from patients with endstage ischemic cardiomyopathy. HCN2 and HCN4 were detected in human myocardium, using Anti-HCN2 and Anti-HCN4 antibodies in western blot studies and their levels were significantly higher in failing ventricles and atrial samples. HCN upregulation likely contributes to increased I_f and may play a role in ventricular and atrial arrhythmogenesis in heart failure⁴⁷.

Although absence epilepsy has a genetic origin, evidence from an animal model suggests that seizures are sensitive to environmental manipulations. Manipulation of the early rearing environment of epileptic rats leads to a pronounced decrease in seizure activity later in life. The alterations in seizure activity between rats reared differently might be correlated with changes in I_h and its channel subunits. Western blot analysis of the somatosensory cortex revealed a long term increase in I_h and HCN1 expression specifically (using Anti-HCN1, 2, and 4 antibodies), in neonatal handled and maternal deprived⁴⁴. An extended work showed, that HCN1 expression undergoes a rapid decline (along with a decline in I_h), preceding the onset of seizures¹⁶. Epileptic rat cortex demonstrated weak immunolabeling of HCN1 and I_h currents as opposed to control rat cortex. Abnormal activity in DRG function may be involved in higher pain sensitivity or spontaneous pain sensation among some patients. As HCN channels have been identified in DRGs, their involvement was investigated regarding pain sensation²³. Immunohistochemistry sections demonstrated that all four HCN channels are expressed in peripheral sensory terminals within plantar skin sections (Anti-HCN2 and Anti-HCN4 were both purchased from Alomone Labs). The data strongly suggested that expression and modulation of I_h in sensory organs and nerve fibers in the periphery play a role in somatosensory processing and contribute to both spontaneous pain and tactile allodynia.

The preoptic area (PO) and the anterior hypothalamus (AH) of the rostral hypothalamus are important integrative centers in the regulation of body temperature⁵⁰. Neurons are classified as temperature sensitive (warm or cold), temperature insensitive and silent. Immnunohistochemical studies showed that HCN2 and HCN4 labeling are found (using Anti-HCN2 and Anti-HCN4) in the PO/AH at all levels and are important for their activity⁵⁰.

References

1. Abbas, S.Y. et al. (2006) Neuroscience 141, 1811.
2. Alig, J. et al. (2009) Proc. Natl. Acad. Sci. U.S.A. 106, 12189.
3. Altomare, C. et al. (2003) J. Physiol. 549, 347.
4. Antal, M. et al. (2004) Eur. J. Neurosci. 19, 1336.
5. Atkinson, S.E. and Williams, S.R. (2009) J. Neurophysiol. 102, 735.
6. Au, K.W. et al. (2008) Am. J. Physiol. 294, C136.
7. Bakondi, G. et al. (2009) Neuroscience 158, 1469.
8. Biel, M. et al. (2009) Physiol. Rev. 89, 847.
9. Brioschi, C. et al. (2009) J. Mol. Cell. Cardiol. 47, 221.
10. Bucchi, A. et al. (2006) Circulation 114, 992.
11. Chen, C. (1997) Hear. Res. 110, 179.
12. Doan, T.N. et al. (2004) J. Neurosci. 24, 3335.
13. Gravante, B. et al. (2004) J. Biol. Chem. 279, 43847.
14. Jagger, D.J. and Housley, G.D. (2003) J. Physiol. 552, 525.
15. Kimura, K. et al. (2004) Genes Cells. 9, 631.
16. Kole, M.H.P. et al. (2007) J. Physiol. (London) 578, 507.
17. Le Scouarnec, S. et al. (2008) Proc. Natl. Acad. Sci. U. S. A. 105, 15617.
18. Lei, M. et al. (2004) J. Physiol. 559.3, 835.
19. Li, C.H. et al. (2008) Am. J. Physiol. 294, C355.
20. Li, Y.L. et al. (2008) Cardiovasc. Res. 79, 715.
21. Liu, J. et al. (2007) Cardiovasc. Res. 73, 729.
22. Lujan, R. et al. (2005) Eur. J. Neurosc. 21, 2073.
23. Luo, L. et al. (2007) Neuroscience 144, 1477.
24. Luo, X. et al. (2008) J. Biol. Chem. 283, 20045.
25. Matsuyoshi, H. et al. (2006) Brain Res. 1119, 115.
26. Meuth, S.G. et al. (2006) J. Neurophysiol. 96, 1517.
27. Michels, G. et al. (2008) PLoS One 3, e1511.
28. Milligan, C.J. et al. (2006) Brain Res. 1081, 79.
29. Mo, Z.L. and Davis, R.L. (1997) J. Neurophysiol. 78, 3019.
30. Monteggia, L.M. et al. (2000) Brain Res. Mol. Brain. Res. 81, 129.
31. Morris, N.P. et al. (2004) Brain. Res. 1006, 74.
32. Nof, E. et al. (2007) Circulation 116, 463.
33. Notomi, T. and Shigemoto, R. (2004) J. Comp. Neurol. 471, 241.
34. Oertel, D. et al. (2008) Neuroscience 154, 77.
35. Pal, B. et al. (2003) Cell. Mol. Life Sci. 60, 2189.
36. Papp, I. et al. (2006) Eur. J. Neurosci. 24, 1341.
37. Peters, C.J. et al. (2009) J. Mol. Cell. Cardiol. 46, 636.
38. Proenza, C. et al. (2002) J. Biol. Chem. 277, 5101.
39. Qu, J. et al. (2004) J. Biol. Chem. 279, 43497.
40. Qu, Y. et al. (2008) J. Physiol. 586, 701.
41. Santoro, B. et al. (2000) J. Neurosci. 20, 5264.
42. Santoro, B. and Tibbs, G.R. (1999) Ann. N. Y. Acad. Sci. 868, 741.
43. Santoro, B. et al. (2004) J. Neurosci. 24, 10750.
44. Schridde, U. et al. (2006) Eur. J. Neurosci. 23, 3346.
45. Siu, C.W. et al. (2009) J. Membr. Biol. 230, 35.
46. Stewart, M. and Fox, S.E. (1990) Trends Neurosci. 13, 163.
47. Stillitano, F. et al. (2008) J. Mol. Cell. Cardiol. 45, 289.
48. Surges, R. et al. (2006) Eur. J. Neurosci. 24, 94.
49. Vertes, R.P. and Kocsis, B. (1997) Neuroscience 81, 893.
50. Wechselberger, M. et al. (2006) Am. J. Physiol. 291, R518.
51. Whitaker, G.M. et al. (2007) J. Biol. Chem. 282, 22900.
52. Yano, S. et al. (2008) Biomed. Res. 29, 195.
53. Ye, B. and Nerbonne, J.M. (2009) J. Biol. Chem. 284, 25553.
54. Ying, S.W. et al. (2007) J. Neurosci. 27, 8719.
55. Zhang, Q. et al. (2009) Biochim. Biophys. Acta. 1788, 1138.
56. Zicha, S. et al. (2005) Cardiovasc. Res. 66, 472.
57. Zolles, G. et al. (2009) Neuron 62, 814.