Sodium-calcium exchangers in rat trigeminal ganglion neurons

Background Noxious stimulation and nerve injury induce an increase in intracellular Ca2+ concentration ([Ca2+]i) via various receptors or ionic channels. While an increase in [Ca2+]i excites neurons, [Ca2+]i overload elicits cytotoxicity, resulting in cell death. Intracellular Ca2+ is essential for many signal transduction mechanisms, and its level is precisely regulated by the Ca2+ extrusion system in the plasma membrane, which includes the Na+-Ca2+ exchanger (NCX). It has been demonstrated that Ca2+-ATPase is the primary mechanism for removing [Ca2+]i following excitatory activity in trigeminal ganglion (TG) neurons; however, the role of NCXs in this process has yet to be clarified. The goal of this study was to examine the expression/localization of NCXs in TG neurons and to evaluate their functional properties. Results NCX isoforms (NCX1, NCX2, and NCX3) were expressed in primary cultured rat TG neurons. All the NCX isoforms were also expressed in A-, peptidergic C-, and non-peptidergic C-neurons, and located not only in the somata, dendrites, axons and perinuclear region, but also in axons innervating the dental pulp. Reverse NCX activity was clearly observed in TG neurons. The inactivation kinetics of voltage-dependent Na+ channels were prolonged by NCX inhibitors when [Ca2+]i in TG neurons was elevated beyond physiological levels. Conclusions Our results suggest that NCXs in TG neurons play an important role in regulating Ca2+-homeostasis and somatosensory information processing by functionally coupling with voltage-dependent Na+ channels.

Somatosensory neurons of the peripheral nervous system, including dorsal root ganglion (DRG) and cranial neurons, are involved in conduction of sensory information from the peripheral organs, and relay to the central nervous system. Noxious and innocuous stimuli applied to the orofacial area are mainly received by trigeminal ganglion (TG) neurons. Both DRG and TG neurons consist of subpopulations of mechanosensitive and nociceptive neurons [11], which are divided into thickly-myelinated, fast-conducting Aβneurons and slower-conducting Aδ-neurons; C-neurons are unmyelinated and slow-conducting. While 90% of C-neurons and 70% of Aδ-neurons are involved in nociception, 80% of Aβ-neurons are involved in conduction of non-nociceptive information [12].
Expression of NCXs has been demonstrated in smalland large-diameter neurons of the DRG [13,14]. These NCXs usually function in Ca 2+ -efflux mode (forward mode) depending on extracellular Na + concentration, while in Ca 2+ -influx mode (reverse mode) they depend on a depolarization-induced action potential in neurons [3]. Ca 2+ -influx induces neurotransmitter release and activation of various signal transduction pathways [15]. The Ca 2+ -influx mode of NCXs has been associated with peripheral nerve injury-induced neuropathic pain [16]. Neuropathic pain is debilitating and chronic, and responds poorly to therapy. Its pathology includes a complex array of interrelated pathways leading to peripheral sensitization. A number of key factors have been hypothesized to modulate clinical status in the neuropathic pain mechanism. The NCX has been proposed as one such factor, and has been targeted in the management of peripheral nerve injury-induced neuropathic pain [16]. However, in TG neurons, PMCA has been identified as the major mechanism for removal of cytosolic Ca 2+ following electrical activity, and the role of NCXs in this process remains to be clarified [17]. The purpose of this study was to investigate the expression/localization of NCXs and their functional properties in TG neurons.

Characterization of primary cultured TG neurons
Primary cultured TG cells (Figure 1, in phase contrast) were loaded with fura-2, and [Ca 2+ ] i was measured. The changes in [Ca 2+ ] i was expressed as F/F 0 units, with R F340/F380 values (F) normalized with respect to the resting value (F 0 ). To confirm that the primary cultured cells were indeed TG neurons, we assessed the cellular response to an extracellular solution containing high extracellular K + , which would induce depolarization, and hence Ca 2+ influx in neurons (according to the K + equilibrium potential changes) [17][18][19]. A series of applications of high-KCl (50 mM K + ) solutions induced an increase in [Ca 2+ ] i in 88.2% of tested primary cultured TG cells, while the remaining cells (11.8%) showed no response (Figure 2A and B).

mRNA expression of NCX isoforms in primary cultured TG neurons
Gene-specific primers for real-time RT-PCR analyses revealed the expression of NCX1, NCX2, and NCX3 mRNA in rat TG neurons (gray bars in Figure 3). We did not find any significant differences in the level of mRNA expression between the 3 NCX isoforms in TG neurons, or in cells in the cerebrum. However, myocardial cells showed significantly higher expression of NCX1 mRNA compared to NCX2 and NCX3 mRNA (black bars in Figure 3).

Immunolocalization of NCX isoforms in primary cultured TG neurons
Intense immunoreactions for NCX1, NCX2, and NCX3 were observed (red in Figure 4A to C) in the somata, dendrites, axons, and perinuclear region of primary cultured TG neurons. These cells all showed positive immunoreactivity to a blended neuronal marker cocktail containing anti-neuronal nuclei (NeuN), anti-microtubule-associated protein 2 (MAP2), βIIItubulin, and anti-neurofilament 200 (NF-H) ( Figure 4D); they also showed positive immunoreactivity to either NF-H (as an A neuron marker; Figure 4H), isolectin B4 (IB4) (as a non-peptidergic C neuron marker; Figure 4L), or calcitonin gene-related peptide (CGRP) (as a peptidergic C neuron marker; Figure 4P) alone. All NCX isoforms (green in Figure 4E to G) were co-expressed with blended neuronal marker cocktails (red in Figure 4E to G). Furthermore, all NCX isoforms (red in Figure 4I to K, M to O and Q to S) were expressed on NF-H-(green in Figure 4I to K), IB4-(green in Figure 4M to O), or CGRPpositive neurons (green in Figure 4Q to S).

Immunolocalization of NCX isoforms in the somata of TG neurons
In cryosections, intense immunoreactivity for NCX1, NCX2, and NCX3 was observed on entire cell bodies, including the perinuclear region, of the TG cells (red in Figure 5A to C). In addition, these cells showed positive immunoreactivity for NF-H (green in Figure 5D), IB4 (green in Figure 5H), or CGRP (green in Figure 5L) alone. All the NCX isoforms (red in Figure 5E to G, I to K, and M to O) were expressed in NF-H-(green in Figure 5E to G), IB4-(green in Figure 5I to K) and CGRP-positive neurons (green in Figure 5M to O).

Ca 2+ -dependence of reverse NCX activity
In order to investigate Ca 2+ influx resulting from reverse NCX activity, the [Ca 2+ ] i response was measured by equimolar substitution of Li + for Na + . TG neurons were first bathed in medium containing 150 mM Na + (Na + -ECS), and then Na + was replaced with equimolar Li + (Li + -ECS) in order to reverse the Na + gradient across the plasma membrane in the presence of extracellular Ca 2+ ( Figure 6A). An  Figure 6B).

Pharmacological characteristics of NCX-mediated Ca 2+ influx
To determine the sensitivity of the expressed NCXs to various NCX inhibitors, we examined the effects of KB-R7943 ( Figure 7A), SEA0400 ( Figure 7C), and SN-6 ( Figure 7E) on Ca 2+ influx mediated by NCXs. In the presence of 1.5 mM [Ca 2+ ] o , a series of applications of KB-R7943, SEA0400, or SN-6 significantly blocked Ca 2+ influx by reverse NCX activity in a concentration-dependent manner ( Figure 7A, C and E). A 50% blockage of inhibitory concentration (IC 50 ) was obtained at a KB-R7943 concentration of 3.08 μM ( Figure 7B), an SEA0400 concentration of 0.03 μM ( Figure 7D), or an SN-6 concentration of 0.51 μM ( Figure 7F).

K + -dependence of Na + -Ca 2+ exchangers in TG neurons
In order to investigate K + -dependence of Ca 2+ influx mediated by reverse Na + -Ca 2+ exchange activity, the [Ca 2+ ] i Figure 3 Detection of NCX isoforms in rat TG cells by real-time RT-PCR. PCR products were monitored by the increase in fluorescence caused by binding of SYBR green dye to doublestranded DNA. Data was analyzed by the 2 −[ΔΔCt] method, with GAPDH as an internal control (which was positive in all samples). Note that the relative mRNA expression level of NCX isoforms does not indicate the absolute level of expression. Each bar denotes mean ± S.D. of mRNA expression relative to GAPDH. Statistically significant differences between columns are indicated by asterisk: *p<0.05. response was measured by equimolar substitution of Li + for Na + with or without extracellular 5.0 mM K + . In the presence of 1.5 mM [Ca 2+ ] o , application of Li + -ECS either with or without extracellular 5.0 mM K + induced a robust increase in [Ca 2+ ] i in the TG neurons ( Figure 8A). No significant differences were observed in these [Ca 2+ ] i responses ( Figure 8B).

Prolonged inactivation kinetics in voltage-dependent Na + channels induced by NCX blockage
The effect of NCX inhibitors on voltage-dependent Na + currents (I Na ) was examined to investigate the contribution of NCX activity to excitability in TG neurons. Inward currents were elicited from TG neurons in 20-ms voltage clamp steps ranging from −80 to +80 mV. Voltage steps were applied in 10-mV increments from a holding potential (V h ) of −70 mV, in the presence of standard extracellular solution, and intracellular solution (ICS) containing either 44 nM, or 26 μM free Ca 2+ (physiological or [Ca 2+ ] i -overloaded condition, respectively). Current-voltage (I-V) relationships were determined by plotting the peak current amplitudes against the applied membrane potentials (circles; Figure 9A and B). Differences in cell size were accounted for by normalizing for the measured capacitance, and current amplitudes were expressed in terms of current densities (pA/pF). At a V h of −70 mV, inward currents developed at a membrane potential of −60 mV and reached maximum amplitude at a membrane potential of −30 mV under both conditions (44 nM and 26 μM [Ca 2+ ] i , open black and red circles; Figure 9A and B). There was no significant difference between the current densities under the physiological (44 nM [Ca 2+ ] i ) and [Ca 2+ ] ioverloaded conditions (26 μM [Ca 2+ ] i ). Inward currents were, however, significantly blocked by the voltagedependent Na + channel blocker tetrodotoxin (TTX) (1.0 μM) under both conditions (filled black and red circles; Figure 9A), while a small residual TTX-resistant component was observed. This indicates that these currents (total I Na ) were carried mainly by TTX-sensitive voltage-dependent Na + channels, with small component carried by TTX-resistant channels.
As shown in Figure 9B, in the physiological condition (open and filled black circles), there was no significant change in the peak amplitude of total I Na without (open black) or with (filled black) a cocktail of NCX inhibitors (3.0 μM KB-R7943, 0.03 μM SEA0400 and 0.5 μM SN-6). This was also the case in the [Ca 2+ ] i -overloaded condition (open and filled red circles in Figure 9B). The mean peak To further investigate the effect of NCX inhibitors on total I Na , the inactivation kinetics were examined by analyzing current decay during depolarization. The timecourse of inactivation with or without the cocktail of NCX inhibitors was well described by the single exponential function. In the absence of NCX inhibitors, we did not observe any significant changes in the mean values for the time constants of inactivation (τ) of total I Na between the physiological (open black circles) and [Ca 2+ ] i -overloaded condition (open red circles), as shown in Figure 9C. In the presence of NCX inhibitors, there were no significant changes in the values of τ in the physiological condition (filled black circles), compared with those in the absence of NCX inhibitors (open black circles in Figure 9C). However, as shown in the inset of Figure 9C, under the [Ca 2+ ] i -overloaded condition and in the presence of NCX inhibitors, the values of τ (black lines as described by the single exponential function) were significantly prolonged in a voltage-dependent manner (filled red circles in Figure 9C).

Immunolocalization of NCX isoforms in axons innervating dental pulp
We further investigated the distribution of NCX isoforms in the axons of TG neurons innervating peripheral tissue. Intense immunoreactivity for the Pan neuronal maker was observed on axons innervating dental pulp (red in Figure 10A, D, and G), and this was co-localized with positive immunoreactivity for NCX1 (green in Figure 10B and C), NCX2 (green in Figure 10E and F), and NCX3 (green in Figure 10H and I).

Discussion
Our results demonstrate the functional expression of NCX1, NCX2, and NCX3 in TG neurons. Application of a high-KCl solution resulted in a depolarization-induced increase in [Ca 2+ ] i in 90% of the primary cultured TG cells. These depolarization-induced increases in [Ca 2+ ] i were observed in all tested cells showing reverse NCX activity (data not shown). In addition, these cells, which were obtained from rat TG primary cultures, expressed total I Na , indicating that the tested cells comprised mostly TG neurons rather than glial cells [17][18][19].
Real-time RT-PCR revealed mRNA expression of all NCX isoforms in the primary cultured TG neurons. NCX1 was ubiquitously characterized and cloned (in organs such as the heart, brain, and kidney and at lower levels in other tissues), while NCX2 and NCX3 were selectively expressed in the brain, skeletal muscle, and selected neuronal populations [1,3,4,10]. In the current study, NCX1, NCX2, and NCX3 expression pattern in the TG neurons was similar to that in the cerebrum [3], but not to that in myocardium. Expression of all NCX isoforms has been reported to be much higher in neurons than in astrocytes [20,21].
Intense NCX immunoreactivity was observed on the somata, dendrites, axons and perinuclear region of TG neurons (Figures 4 and 5). These NCX isoforms appeared to be localized on the cell membrane rather than in the  Figure 5). Moreover, expression of NCXs was observed on axons innervating the peripheral tissue (dental pulp), indicating that all NCX isoforms are distributed throughout trigeminal A-, non-peptidergic C-, and peptidergic C-neurons [12].
It has been reported that NF-H (a marker for myelinated medium-to large-diameter afferents) [23] and IB4 (which binds to unmyelinated small-diameter nociceptive afferent neurons) [24,25] co-localize with P2X 2 and P2X 3 receptors in rat TG neurons, as Aδ and C nociceptors, respectively [26,27]. In addition, CGRP is released from primary afferent terminals onto secondary afferents in response to various noxious stimulation [24,28]. Considering these findings, the current results indicate that NCX isoforms are expressed not only in mechanosensitive Aβ neurons, but also in nociceptive afferent neurons.
Pharmacological identification of NCX was achieved by using benzyloxyphenyl derivatives, which are NCX inhibitors. The potencies of 3 benzyloxyphenyl derivatives, KB-R7943, SEA0400, and SN-6, on Na + -dependent Ca 2+ uptake have been reported in fibroblasts [29]. These 3 inhibitors partially block Ca 2+ influx through NCX and have different isoform selectivities. KB-R7943 is 3-times more potent against NCX3 than against NCX1 or NCX2, while SEA0400 predominantly inhibits NCX1, only mildly inhibits NCX2, and exerts almost no influence upon NCX3 [29]. SN-6 shows 3-to 5-times greater inhibition of NCX1 than either NCX2 or NCX3 [29]. In the present study, the values of IC 50 for KB-R7943 (3.08 μM) and SEA0400 (0.056 μM) were consistent with the general characteristics of those reported for fibroblasts transfected with NCX1, NCX2, or NCX3. However, reverse NCX activity in TG neurons showed a strong sensitivity to SN-6 (IC 50 =0.51 μM). Therefore, our results suggest that NCX1 is the predominantly functional isoform in TG neurons, although in TG neurons, the mRNA expression levels of the 3 isoforms were not different. During postnatal development, however, the expression pattern of NCX isoforms in rat cultured neurons changes: NCX1 was predominant in neonatal rat cortical cultured neurons, but the level of NCX2 increased, while NCX1 and NCX3 level decreased in adult rat neurons [21]. These results are in line with our results obtained from neonatal (4-7 days old) rat TG neurons. Further study is required to precisely characterize the developmental changes in NCX isoform expressions in TG neurons.
In addition to NCX, the intracellular Ca 2+ efflux system on the plasma membrane involves another type of transporter, the K + -dependent Na + -Ca 2+ exchangers (NCKXs) which are comprises NCKX1 to NCKX5 and constitutes the SLC24 gene family [30,31]. The NCKXs are also reportedly distributed throughout the central nervous system [32][33][34]; however, Ca 2+ influx by reverse Na + -Ca 2+ exchange did not show any K + -dependence (Figure 8), indicating that NCKX activities did not contribute to the reverse NCX activities recorded in the present study.
We also investigated the coupling mechanism between voltage-dependent Na + channels and NCXs in TG neurons. In the nociceptive neurons, voltage-dependent Na + channels are pharmacologically categorized as either TTX-sensitive or TTX-resistant, depending on their activation and inactivation kinetics, activation threshold, and voltage dependence of activation and steady-state inactivation [35][36][37]. In this study, although NCX expression in small-to large-diameter neurons was observed to co-localize with NF-H-, IB4-and CGRP-positive TG neurons, voltage-dependent inward currents were significantly blocked by TTX. This indicates that the inward currents were carried primarily by TTX-sensitive voltage-dependent Na + channels; however, we could not separate total I Na into TTX-sensitive and TTX-resistant components.
As an essential pain modulator, NCXs are of interest in the management of peripheral nerve injury-induced neuropathic pain [16]. NCX activation in Ca 2+ -influx mode is associated with [Ca 2+ ] i increase and various pathological conditions, including hypoxia-anoxia, white matter degeneration after spinal cord injury, brain trauma, and optical nerve injury [38]. An overload of [Ca 2+ ] i via NCXs in Ca 2+ -influx mode driven by axonal Na + accumulation via TTX-sensitive voltage-dependent Na + channels has been observed in injured myelinated axons in dorsal root [39]. This increase in [Ca 2+ ] i was not Page 9 of 15 http://www.molecularpain.com/content/9/1/22 dependent on L-type voltage-dependent Ca 2+ channels [39]. In addition, in the case of axonal injury, stretchinjured traumatic deformation of axons induces abnormal Na + influx via TTX-sensitive voltage-dependent Na + channels, which subsequently triggers an increase in intra-axonal Ca 2+ via the activation of NCX in Ca 2+ -influx mode [40]. In an anoxic situation, myelinated axons lose K + , while intra-axonal Na + concentrations increase primarily via Na + influx through TTX-sensitive voltagedependent Na + channels; elevated axoplasmic Na + and  axolemmal depolarization promote Ca 2+ influx mediated primarily by NCX reverse activity [41]. This pathway is considered the primary mechanism of Ca 2+ influx in neuropathic pain states [41].
In our study, the [Ca 2+ ] i -overloaded condition had no effect on either peak current amplitudes, or inactivation kinetics of total I Na (black and red open circles in Figure 9A and C). Additionally, in the physiological condition, the inhibition of NCXs did not influence the peak amplitudes or inactivation kinetics of total I Na (see Figure 9B and C). These findings indicate that neither [Ca 2+ ] i -overload itself nor NCX inhibition seem to be directly related to neuronal excitability; rather our results support previous findings (see above) [41] that [Ca 2+ ] i -overload caused by Ca 2+ influx via NCXs contributes to neuropathic pain. In contrast, the inactivation process of I Na was significantly and voltagedependently prolonged by the application of NCX inhibitors, only when the [Ca 2+ ] i in TG neurons was overloaded (red filled circles in Figure 9C), suggesting that NCXs play important roles in maintaining the normal functioning of voltage-dependent Na + channels in TG neurons under the pathologically [Ca 2+ ] i -overloaded condition, but not under physiological conditions. Therefore, not only [Ca 2+ ] i -overload from Ca 2+ influx via NCXs, but also the reduced activity of NCXs under pathological [Ca 2+ ] i -overload in TG neurons might be involved in modulating neural excitability.
In contrast, KB-R7943 (at 10 μM), but not SEA0400, has been shown to have an inhibitory effect on the peak amplitude of I Na in myocardium [42]. In our study, at the IC 50 for inhibition of the reverse Ca 2+ influx mode of NCX, not only KB-R7943, but also SEA0400 and SN-6 did not significantly affect the I Na peak current density in TG neurons, under both the physiological and [Ca 2+ ] i -overloaded conditions. In addition, a cocktail of these 3 NCX inhibitors (KB-R7943, SEA0400 and SN-6) prolonged the inactivation process of I Na under the [Ca 2+ ] i -overloaded condition, but not under the physiological condition (red filled circles in Figure 9C). Thus, a non-specific effect of these NCX inhibitors on total I Na is unlikely.
Voltage-dependent Na + channels comprise a transmembrane glycoprotein consisting of a pore-forming α subunit, and an auxiliary β subunit which modulates the kinetics and voltage dependence of Na + channel activation and inactivation [43,44]. The β subunit, in particular, regulates the propagation of action potentials through critical intermolecular and cell-cell communication events [44]. Additionally, the β subunit appears with the α subunit during early neural development, and its expression is increased during rapid neuronal growth and differentiation [45]. In DRG neurons, similar expression patterns of voltage-dependent Na + channel α subunits (e.g. Na v 1.6, Na v 1.7, Na v 1.8, and Na v 1.9) and NCX2 have been demonstrated along the neurites and at the neurite tips [13]. Furthermore, NCX1 colocalizes with Na v 1.6 in the spinal cord [46]. Therefore, the NCXs in TG neurons functionally couple with voltage-dependent Na + channels, most likely via their β subunit. The current results together with those of earlier reports suggest that functional coupling between NCXs and voltage-dependent Na + channels in TG neurons is one of the key factor(s) in the development and/ or modulation of neuro-pathological conditions such as neuropathic pain.

Conclusion
In the present study, we showed that TG neurons functionally expressed all NCX isoforms in axon, dendrites and somata, and that they localized on A-, non-peptidergic C-, and peptidergic C-neurons. The present results suggest that coupling of NCXs with voltage-dependent Na + channels in TG neurons may play a key role in the modulation of nociceptive and/or neuropathic pain in the oral and maxillofacial region.

Cell isolation and primary culture
All animals were treated in accordance with the Guiding Principles for the Care and Use of Animals in the Field of Physiological Sciences approved by the Council of the Physiological Society of Japan and the American Physiological Society, and the guidelines established by the National Institutes of Health (USA) regarding the care and use of animals for experimental procedures. The study was approved by the Ethics Committee of Tokyo Dental College (approval No.232507/No.242502). The Trigeminal ganglion (TG) was dissected and isolated from neonatal Wistar rats of both sexes (4-7 days old) under pentobarbital sodium anesthesia (50 mg/kg intraperitonially). The cells were dissociated by enzymatic treatment with Hank's balanced salt solution (Invitrogen, Grand Island, NY, USA) containing 20 U/ml papain (Worthington, Lakewood, NJ, USA) for 20 min at 37°C followed by dissociation by trituration [47,48]. Dissociated TG cells were plated onto poly-L-lysine-coated dishes (Corning, Corning, NY, USA). Primary culture was performed with Leibovitz's L-15 medium (Invitrogen) containing 10% fetal bovine serum, 1% penicillin-streptomycin (Invitrogen), 1% fungizone (Invitrogen), 26 mM NaHCO 3 and 30 mM glucose (pH 7.4) [47,48]. The cells were incubated and maintained at 37°C in a humidified atmosphere of 95% air and 5% CO 2 for 24-48 hr.

Real-time reverse transcription-polymerase chain reaction
The cerebrum and myocardium were dissected from neonatal Wistar rats of both sexes (4-7 days old) under pentobarbital sodium anesthesia. Total mRNA was extracted from the cultured rat TG cells or isolated cerebrum and myocardium by a modified acid guanidiumphenol-chloroform method [49]. Reverse transcription, cDNA amplification and polymerase chain reaction (PCR) were performed with the One-Step SYBR Primescript RT-PCR Kit and Thermal Cycle Dice for semiquantitative real-time RT-PCR (Takara-Bio, Shiga, Japan). The forward and reverse primer sets used are listed in Table 1 [9]. Results obtained from real-time RT-PCR experiments were quantified using the comparative threshold  ) method (Ct represents cycle threshold) [50,51]. Relative mRNA expression level was normalized by that of glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The Ct for each primer set was obtained from samples and averaged. Delta Ct represented the calculated difference between the average Ct for the target gene (NCX isoforms) and the average Ct for GAPDH as the control for total starting RNA quantity. The delta-delta Ct method of calculation was then used to assess fold-change in gene expression relative to the GAPDH gene [50,51].

Immunofluorescence analyses
In order to identify the localization of NCX isoforms in TG neurons, we examined immunofluorescence in the primary cultured TG cells, as well as in cryosections from the TG and mandible. The TG and mandible was dissected bilaterally from neonatal Wistar rats of both sexes (4-7 days old) under pentobarbital sodium anesthesia. Dissected TG and mandible were embedded together in mounting medium (Tissue-Tek O.C.T.; Sakura Finetek, Tokyo, Japan), and sectioned longitudinally on a cryostat at 15-μm thickness. Primary cultured TG cells were seeded and cultured on poly-D-lysine-coated glass slides (Matsunami, Osaka, Japan). After fixation with 50% ethanol and 50% acetone and blocking in 10% donkey serum, each section and primary cultured TG cell was incubated with goat anti-NCX1, anti-NCX2, or anti-NCX3 antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA; dilution, 1:50) and a blended neuronal maker cocktail (Milli-Mark ™ Pan Neuronal Marker, Millipore, Billerica, MA, USA; dilution 1:50) including mouse anti-NeuN, anti-MAP2, anti-βIIItubulin and anti-NF-H antibodies. The TG cells were also incubated with either rabbit anti-NF-H antibody (Millipore; dilution 1:200) as an A-neuron marker, IB4 antibody-conjugated FITC (dilution 1:200) as a nonpeptidergic C-neuron marker, or rabbit anti-CGRP antibody (Santa Cruz Biotechnology; dilution 1:50) as a peptidergic C-neuron marker, at 4°C overnight. The cells were then washed and incubated with a secondary antibody (Alexa Fluor 488 or 568 donkey anti-goat IgG, Alexa Fluor 568 donkey anti-mouse IgG, or Alexa Fluor 488 donkey anti-rabbit IgG, Molecular Probes, Eugene, OR, USA; dilution 1:100) for fluorescence staining and with 4' ,6-diamino-2-phenylindole dihydrochloride (Molecular Probes) for nuclear staining at room temperature for 5 min. Negative controls were prepared by using nonimmune IgGs diluted at a concentration equivalent to that of the primary antibodies. Cells were examined under a conventional fluorescence microscope (Zeiss, Jena, Germany).

Measurement of intracellular Ca 2+ concentration
Primary cultured TG cells were incubated for 60 min at 37°C in Hank's solution containing 10 μM fura-2 acetoxymethyl ester (Dojindo, Kumamoto, Japan) and 0.1% (w/v) pluronic acid F-127 (Invitrogen), followed by rinsing with fresh Hank's solution. A dish with fura-2-loaded TG neurons was mounted on the stage of a microscope (Olympus, Tokyo, Japan) incorporated into the Aquacosmos system and software (Hamamatsu Photonics, Shizuoka, Japan) equipped with an excitation wavelength selector and an intensified charge-coupled device camera (Hamamatsu Photonics). [Ca 2+ ] i was expressed as the fluorescence ratio (R F340/F380 ) at two excitation wavelengths of 380 nm and 340 nm.

Whole-cell recording techniques
Whole-cell patch-clamp recordings were carried out under voltage-clamp conditions [52]. Patch pipettes with a resistance of 2-5 MΩ were pulled from capillary tubes by using a DMZ-Universal Puller (Zeitz-Instruments, Martinsried, Germany) and filled with an intracellular solution (ICS). Whole-cell currents were measured by using a patch-clamp amplifier (L/M-EPC-7+; Heka Elektronik, Lambrecht, Germany). Current traces were displayed and stored in a computer by using pCLAMP (Axon Instruments, Foster City, CA) after digitizing the analog signals at 10 kHz (DigiData 1440A; Axon Instru Instruments). Current records were filtered at 3 kHz. Data were analyzed offline by using pCLAMP. The currents via voltage-dependent Na + channels (total I Na ) were recorded with a holding potential of −70 mV. The current-voltage (I-V) relationship of I Na was measured by applying 20-ms depolarizing pulses increasing from −80 to +80 mV in 10-mV steps at 2-s intervals.   [53], and 20 mM HEPES (pH 7.2). Standard solution was also used as an ECS for the patch-clamp recordings. KB-R7943 and SN-6 were obtained from Tocris Bioscience (Bristol, UK). SEA0400 (2-[4-[(2,5-difluorophenyl)methoxy]phenoxy]-5ethoxyaniline) was synthesized by Taisho Pharmaceutical Co. Ltd (Saitama, Japan), and was a gift from Professors Toshio Matsuda and Akemichi Baba, Osaka University, Japan. These reagents were diluted in the ECSs for [Ca 2+ ] i measurement and patch-clamp recording, and applied at the appropriate concentration to primary cultured TG neurons via a gravity-fed perfusion system (Warner instruments, Hamden, CT, USA). For Na + inward current recording, 1.0 μM tetrodotoxin was diluted with standard solution. Except for those noted above, all other reagents were obtained from Sigma-Aldrich (St. Louis, MO, U.S.A).

Statistics and offline analysis
The Ca 2+ -dependence of NCX activity and the doseresponse relationship for the inhibitory effect of NCX inhibitors were obtained by fitting the data with following function: where K is the equilibrium binding constant for applied extracellular Ca 2+ concentration ([Ca 2+ ] o ) with a Hill coefficient (h) equal to 1; or where IC 50 is the 50% inhibitory concentration for applied NCX inhibitors. A max is maximal, and A min is minimal response. The [X] o indicates applied concentration of extracellular Ca 2+ or inhibitors. The kinetics of inactivation was determined by fitting the experimental data with a single exponential function using the pCLAMP software. Data were expressed as the mean ± standard error (S.E.) or standard deviation (S.D.) of the mean of n observations, where n represents the number of separate experiments. The Wilcoxon t-test, Friedman test, or Kruskal-Wallis test and Dunn's post hoc test were used to determine nonparametric statistical significance. A p value of less than 0.05 was considered significant. The statistical analysis was performed using Graph Pad Prism 5.0 (Graph Pad Software, La Jolla, CA, USA).