Introduction

Processing of sound in the cochlea involves both afferent and efferent innervation. The majority of afferent fibers, the myelinated type I fibers, make glutamatergic synapses with the inner hair cells (IHCs) and allow fast and precise transmission to the brain (Fuchs et al. 2003). Unmyelinated type II fibers make en passant synapses onto the outer hair cells (OHCs; Spoendlin 1973) and have a yet unknown role. Efferent feedback to the cochlea is provided by olivocochlear fibers originating in the brainstem. Activation of myelinated medial efferent terminals contacting OHCs suppresses cochlear function (Guinan 1996). Activation of the unmyelinated lateral efferent fibers that form axodendritic synapses onto type I afferent fibers causes either slow suppression or enhancement of afferent activity (Groff and Liberman 2003).

Mechanisms to maintain hyperpolarized membrane potentials and restore intracellular concentration gradients are essential for excitable cells and vary depending on their particular physiology (Blanco 2005). The Na,K-ATPase (NKA) is a membrane-bound protein that uses energy from the hydrolysis of adenosine triphosphate (ATP) to extrude three sodium ions for every two potassium ions taken into the cell. Neurons with high firing rates especially depend upon the NKA to sustain activity, since the NKA maintains concentration gradients and also hyperpolarizes the membrane. The functional NKA is comprised of two α subunits, which contain the residues necessary for ATP hydrolysis and ion transport and two β subunits. The diversity of α (1–4), β (1–3), and also FXYD (1–7) subunits (Geering 2005) and the highly regulated tissue-specific and developmental expression patterns are no doubt responsible for tailoring the NKA transport properties to specific cellular demands. In this study, we investigated the neuronal distribution of the NKA α subunit (NKAα) in the cochlea to determine its contribution to setting and modulating activity patterns in the various neuronal cell types.

Because inhibitors of the NKAα are known to reduce the endocochlear potential (Konishi and Mendelsohn 1970), previous research investigating the localization of the NKAα in the cochlea has focused on its expression in structures, like the stria vascularis and spiral ligament, responsible for maintaining the endocochlear potential (Wangemann 2006). Nonetheless, earlier work also reported expression of the NKAα in the neuronal elements of the mammalian cochlea using a variety of techniques, including enzyme cytochemistry (Kerr et al. 1982), in situ hybridization (Ryan and Watts 1991), and immunohistochemistry (Schulte and Adams 1989; Iwano et al. 1990; McGuirt and Schulte 1994; Schulte and Steel 1994; ten Cate et al. 1994; Nakazawa et al. 1995; Zuo et al. 1995; Erichsen et al. 1996; Peters et al. 2001; Weber et al. 2001).

Using immunofluorescence in whole-mount preparations of the organ of Corti and spiral ganglion, we performed a variety of double labeling experiments with antibodies against the NKAα and markers identifying particular subsets of neurons or supporting cells within the cochlea. Our results unambiguously identify the cell types expressing particular isoforms of the NKAα and, importantly, suggest novel roles for the NKAα in regulating neuronal activity in the cochlea.

Methods

Immunostaining

Immunostaining of organs of Corti was performed as described previously (Pyott et al. 2004, 2007). All animal protocols were approved by the University of North Carolina at Wilmington Animal Care and Use Committee. Whole cochleae were dissected from rats and immediately perfused through the round window with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS) at pH 7.4. Cochleae were fixed in 4% PFA/PBS for 1 to 3 h at 4°C before being rinsed with PBS. Apical turns of the organs of Corti were then dissected from the cochleae and treated with a blocking buffer (PBS with 5% normal goat or donkey serum and 0.2% Triton X-100) for 1 to 3 h at room temperature. Turns were incubated in the primary antibody diluted in blocking buffer overnight at 4°C and then rinsed three times for 20 min in PBS with 0.2% Triton X-100 (PBT). After rinsing, the turns were then incubated in the secondary antibody diluted in blocking buffer for 2 h at room temperature, rinsed three times for 20 min in PBT, and rinsed in PBS before mounting on glass slides in Vectashield mounting medium (Vector Labs). All incubations and rinses were performed on a rocking table. Rats were between 20 and 24 days old. Table 1 summarizes the primary antibodies used. Secondary antibodies (Alexa Fluor 488 and 594 generated in either goat or donkey) were purchased from Molecular Probes and diluted 1:1,000 in blocking buffer. For all antibodies, experiments performed in the absence of primary antibody showed no immunoreactivity. All data were collected from at least three independently replicated experiments.

TABLE 1 Primary antibodies used in the study

Specificity of NKAα antibodies

The NKAα antibodies used in this study were chosen because their specificity has been well characterized previously. Specificity of the antibodies for the NKAα3 has been verified by the vendor using Western blot analysis of rat brain extracts (sc-16052) or canine skeletal muscle extracts (MA3-915). Specificity of the antibodies for the NKAα1 has been verified by the vendor using Western blot analysis of rat brain microsomes (C464.6) and by Arystarkhova and Sweadner (1996) using Western blot analysis of membrane preparations from rat kidney, brainstem, and adult and neonatal heart (α6F). Specificity of the antibodies for the NKAα2 has been verified by Pressley (1992) using Western blot analysis of rat brain microsomes (anti-HERED) and by the vendor using immunohistochemistry to show membrane localization in rat retinal cells.

Microscopy

Fluorescence images were acquired using an Olympus Fluoview FV1000 confocal microscope with a 60× Olympus PlanoApo oil immersion lens (N. A. 1.42) under the control of the Olympus Fluoview FV1000 version 1.6a software. z-Stacks (30 to 90 optical sections) and were collected at 0.3 to 0.5 μm. Except for subpanels D–I in Figure 7, which are micrographs of single optical sections, all images are presented as z-projections through the entire optical stack.

Results

The NKAα3 is expressed in the spiral ganglion and the organ of Corti

To investigate the neuronal distribution of the NKAα in the spiral ganglion and organ of Corti, we immunolabeled apical cochlear turns from rats aged P20 to P25 with antibodies against NKAα3, the neuronal NKAα isoform (red; Fig. 1A). To better localize NKAα3 immunoreactivity, we double-labeled apical cochlear turns with a polyclonal anti-calretinin antibody (green; Fig. 1B). Calretinin is an endogenous calcium buffer enriched in sensory hair cells as well as type I afferent neurons (Dechesne et al. 1991; Dechesne et al. 1993; Zheng and Gao 1997). Low-magnification micrographs of apical cochlear turns show that NKAα3 immunoreactivity is localized to the spiral ganglion somata and fibers in the IHC and OHC region of the organ of Corti (Fig. 1B). Similar patterns of NKAα3 immunoreactivity were seen in middle and basal cochlear turns (data not shown).

FIG. 1.
figure 1

NKAα3 is present in the spiral ganglion somata and organ of Corti. Low-magnification micrographs of apical cochlear turns immunolabeled with a monoclonal anti-NKAα3 (red) show abundant NKAα3 immunoreactivity in the organ of Corti and spiral ganglion (A). Double labeling with a monoclonal anti-NKAα3 (red) and a polyclonal anti-calretinin (green) shows that NKAα3 immunoreactivity is localized to the spiral ganglion somata and fibers contacting the inner and outer hair cells (B). High-magnification micrographs of apical cochlear turns double-labeled with a monoclonal anti-NKAα3 (red) and a polyclonal anti-NKAα3 (green) reveal similar patterns of immunoreactivity in the membranes of the spiral ganglion somata (C–E), fibers terminating onto the inner hair cells (F–H), and fibers terminating onto the outer hair cells (I–K). Localization of NKAα3 immunoreactivity to membranes and colocalized immunoreactivity with different antibodies against the NKAα3 suggest that both antibodies are specifically labeling the NKAα3. Importantly, NKAα3 immunoreactivity appears restricted to the spiral ganglion somata and neuronal elements of the organ of Corti. For orientation, the location of the single row of inner hair cells (IHCs), three rows of outer hair cells (OHCs), and region containing the spiral ganglion cells (SGCs) have been indicated (with a triangle).

To confirm the specificity of the monoclonal antibody against NKAα3, double labeling experiments with a monoclonal and polyclonal anti-NKAα3 were performed (Fig. 1C–K). High magnification micrographs show colocalized immunoreactivity for both the monoclonal (red) and polyclonal (green) anti-NKAα3 for the spiral ganglion somata (Fig. 1C–E) and fibers within the inner (Fig. 1F–H) and outer (Fig. 1I–K) hair cell region. Additionally, both antibodies show immunoreactivity restricted to the cell membranes, consistent with the transmembrane localization of the NKAα. The patterns of immunoreactivity for both the monoclonal and polyclonal anti-NKAα3 suggest that both antibodies are specifically recognizing the NKAα3 in the rat organ of Corti and spiral ganglion. To unambiguously identify the structures expressing NKAα3, we undertook a series of double labeling experiments with antibodies against NKAα3 and known markers of the specific neuronal cell types in the cochlea.

NKAα3 is present in type I but not in type II spiral ganglion somata

Neurofilament 200 (NF200) is an intermediate filament enriched in the cytoplasm of both types I and II spiral ganglion somata and processes (Berglund and Ryugo 1991). Double labeling experiments with a monoclonal anti-NKAα3 (red; Fig. 2A) and a polyclonal anti-NF200 (green; Fig. 2B) revealed that the majority of NF200-positive spiral ganglion somata also show surface expression of NKAα3 (Fig. 2C). However, in all experiments, a small subset of spiral ganglion somata did not show NKAα3 immunoreactivity (marked with an asterisk in Fig. 2C). To determine if this subset of somata belong to type II spiral ganglion neurons (SGNs), we double-labeled organs of Corti with a monoclonal anti-NKAα3 (red; Fig. 2D) and a polyclonal anti-peripherin (green; Fig. 2E). Peripherin is an intermediate filament enriched in the cytoplasm of type II SGNs (Hafidi 1998). These experiments show that the peripherin-positive type II SGNs do not show surface expression of NKAα3 (Fig. 2F) and suggest that NKAα3 is restricted to the type I spiral ganglion somata.

FIG. 2.
figure 2

NKAα3 is present in the type I but not type II spiral ganglion somata. High-magnification micrographs of apical cochlear turns double-labeled with a monoclonal anti-NKAα3 (red; A) and a polyclonal anti-NF200 (green; B) show that NKAα3 immunoreactivity is detected in the cell membranes of the majority of NF200-positive spiral ganglion somata (C). To determine if the small population of NKAα3-negative spiral ganglion cells (marked with an asterisk; C) are, in fact, type II spiral ganglion somata, apical cochlear turns were double-labeled with a monoclonal anti-NKAα3 (red; D) and polyclonal anti-peripherin (green; E). Indeed, the cell membranes of the peripherin-positive, type II spiral ganglion somata do not show NKAα3 immunoreactivity (F).

NKAα3 is present in type I afferent terminals contacting the IHCs

As seen in Figure 1B, higher magnification micrographs of organs of Corti double-labeled with anti-NKAα3 (red; Fig. 3A) and anti-calretinin (green; Fig. 3B) show that the calretinin-positive afferent fibers contacting the calretinin-positive IHCs express NKAα3 (Fig. 3C). To further confirm localization of NKAα3 to the type I afferent terminals, organs of Corti were double-labeled with a monoclonal anti-NKAα3 and a polyclonal anti-NF200. These micrographs show membrane-bound NKAα3 immunoreactivity (red; Fig. 3D) surrounding cytoplasmic NF200 immunoreactivity (green; Fig. 3E) in the afferent terminal including its bouton-shaped tip (Fig. 3F). The patterns of NKAα3 immunoreactivity for calretinin-positive (Fig. 3A–C) and also NF200-positive fibers (Fig. 3D–F), in which NKAα3 immunoreactivity fully ensheathes the cytoplasmic markers, suggest that the type I afferent terminals express the NKAα3.

FIG. 3.
figure 3

NKAα3 is present in the type I afferent terminals. High-magnification micrographs of apical cochlear turns double-labeled with a monoclonal anti-NKAα3 (red; A) and a polyclonal anti-calretinin (green; B) clearly show that the NKAα3-positive fibers are the type I afferent terminals that contact the inner hair cells (C). To further confirm expression of NKAα3 in the type I afferent terminals, apical cochlear turns were double-labeled with a monoclonal anti-NKAα3 (red; D) and a polyclonal anti-NF200 (green; E). NKAα3 immunoreactivity is indeed localized to the cell membranes and bouton-shaped endings of the NF200-expressing type I afferent terminals (F). To determine if NKAα3-immunoreactivity was specifically restricted to the initial unmyelinated segments of the type I afferent terminals, apical cochlear turns were double-labeled with a polyclonal anti-NKAα3 (red) and a monoclonal pan anti-voltage-gated sodium channel (green). Micrographs show that NKAα3 immunoreactivity (red) is greatest just before and not overlapping with sites of sodium channel enrichment (green) at the heminode and is substantially less intense in the myelinated segments and subsequent unmyelinated nodes (G). For orientation, the locations of the single row of inner hair cells (IHCs) and the first heminode (HN) have been indicated (with a triangle).

Previous work has shown that voltage-gated sodium channels in type I afferent fibers are localized at the heminode just before the first site of myelination and at subsequent unmyelinated nodes (Hossain et al. 2005). To determine if the region of greatest NKAα3 immunoreactivity corresponded to the initial unmyelinated segment of the type I afferent terminals, we double-labeled cochlear turns with a polyclonal anti-NKAα3 (red) and a monoclonal antibody known to recognize all neuronal voltage-gated sodium channels, including those expressed in the cochlea (green; K58/35; Rasband et al. 1999). Indeed, NKAα3 immunoreactivity appears greatest in the initial unmyelinated segment of the type I afferent fibers and does not overlap with the site of sodium channel enrichment in the first heminode (Fig. 3G). Although others have reported expression of the NKAα3 in the myelinated segments (Nakazawa et al. 1995) and nodes of Ranvier (Schulte and Adams 1989) of the afferent fibers, our experiments showed NKAα3 immunoreactivity that was most intense in the initial segments of the afferent fibers directly contacting the IHCs and not enriched in the subsequent unmyelinated nodes.

NKAα3 is not present in lateral olivocochlear efferents that contact the IHC afferent terminals

Lateral olivocochlear efferent fibers make axodendritic synapses onto the unmyelinated terminals of type I afferent fibers close to where they contact the IHCs (Bergeron et al. 2005). To test whether NKAα3 immunoreactivity detected in the IHC region may also be due to expression of NKAα3 in lateral efferent fibers, we performed double labeling experiments with a monoclonal anti-NKAα3 and a polyclonal anti-synapsin. Synapsin is known to label synaptic vesicle proteins found in the cytoplasm of traditional synapses (like the efferent synapses) but not in ribbon synapses (von Kriegstein and Schmitz 2003). These double labeling experiments show that NKAα3-positive fibers to the IHCs (red; Fig. 4A) are contacted by synapsin-positive efferent presynapses (green; Fig. 4B) at their unmyelinated terminals (Fig. 4C). Importantly, NKAα3 immunoreactivity (red; Fig. 4A) does not ensheathe the efferent presynapses identified by their cytoplasmic synapsin labeling (green; Fig. 4B). Instead, synapsin-positive efferent presynapses appear to abut upon the membrane of the type I afferent fibers identified by their NKAα3 immunoreactivity (Fig. 4C). Because projections through the entire optical z-stack (Fig. 4) cannot be used to discriminate colocalization of objects above one another, single optical sections were used to examine the colocalization of synapsin-positive efferent presynapses that appeared colocalized with NKAα3-positive terminals (yellow in the z-projections shown in Fig. 4C). Analysis of series of single optical sections (data not shown) confirmed that NKAα3 immunoreactivity does not colocalize with or ensheathe the lateral efferent terminals identified by their cytoplasmic labeling for synapsin. Thus, this pattern of immunoreactivity, in which efferent presynapses labeled cytoplasmically with synapsin do not show membrane-bound NKAα3 immunoreactivioty, is consistent with the axodendritic contacts formed between the lateral efferents and type I afferent terminals and suggests that NKAα3 expression is restricted to the type I afferent terminals.

FIG. 4.
figure 4

NKAα3 is not present in the lateral olivocochlear efferent fibers. Lateral olivocochlear efferent fibers make axodendritic synapses onto the unmyelinated terminals of type I afferent fibers close to where they contact the inner hair cells. To determine if these lateral efferent fibers express the NKAα3, apical cochlear turns were double-labeled with a monoclonal anti-NKAα3 (red; A) and a polyclonal anti-synapsin (green; B). These experiments show that the NKAα3-positive terminals (red) are, in fact, contacted by the synapsin-expressing efferent terminals (green) extracellularly (C). This pattern of immunoreactivity is consistent with the axodendritic contacts formed between the lateral efferents and type I afferent terminals and suggests that NKAα3 expression is restricted to the type I afferent terminals. For orientation, the location of the single row of inner hair cells (IHCs) has been indicated (with a triangle).

NKAα3 is present in medial olivocochlear efferents that contact the OHCs

To further investigate the expression of the NKAα3 in fibers contacting the OHCs, double labeling experiments with antibodies against the NKAα3 and other established markers of structures in the OHC region were performed. Double labeling of organs of Corti with a monoclonal anti-NKAα3 (red; Fig. 5A) and a polyclonal anti-prestin (green; Fig. 5B), a protein enriched in the cell membrane (especially lateral walls) of the OHCs (Belyantseva et al. 2000), shows the NKAα3-positive fibers traveling through the tunnel of Corti and contacting the three rows of OHCs with their bouton-shaped terminals (Fig. 5C). Double labeling with a monoclonal anti-NKAα3 (red; Fig. 5D) and a polyclonal anti-synapsin (green; Fig. 5E) show that NKAα3 and synapsin immunoreactivity are localized to the same set of terminals (Fig. 5F). Higher magnification micrographs clearly show membrane-bound NKAα3 immunoreactivity (Fig. 5G) completely ensheathing cytoplasmic synapsin immunoreactivity (Fig. 5H) in the bouton-shaped terminals contacting the OHCs (Fig. 5I). The known expression of synapsin in the efferent presynapses identifies these NKAα3-positive fibers as the medial efferent fibers that innervate the OHCs.

FIG. 5.
figure 5

NKAα3 is present in the medial efferent fibers. Double labeling of apical cochlear turns with a monoclonal anti-NKAα3 (red; A) and a polyclonal anti-prestin (green; B) shows that NKAα3-positive fibers contact the three rows of outer hair cells (C). To identify these fibers, double labeling experiments were performed with a monoclonal anti-NKAα3 (red; D) and a polyclonal anti-synapsin (green; E). These experiments show that NKAα3 and synapsin immunoreactivity are localized to the same set of terminals (F). Higher magnification micrographs clearly show membrane-bound NKAα3 immunoreactivity (red; G) surrounding cytoplasmic synapsin immunoreactivity (green; H) in the bouton-shaped terminals contacting the outer hair cells (I). Together, these experiments identify the NKAα3-positive fibers targeting the OHCs as the medial efferent fibers. For orientation, the location of the three rows of outer hair cells (OHCs) and the tunnel of Corti (ToC) have been indicated (with a triangle).

NKAα3 is absent in type II afferent fibers that contact the OHCs

Double labeling experiments were performed to test whether NKAα3 immunoreactivity detected in the OHC region may also be due to expression of the NKAα3 in the type II afferent fibers that contact the OHCs. Double labeling experiments with a monoclonal anti-NKAα3 (red; Fig. 6A) and a polyclonal anti-NF200 (green; Fig. 6B) show distinct, non-overlapping patterns of immunoreactivity for the NKAα3-positive fibers and the NF200-positive fibers that appear in the outer spiral bundles under the OHCs (Fig. 6C). This lack of colocalization is better viewed in the x-axis cross-section through the z-projection, which reveals that the NKAα3-positive fibers (red; Fig. 6D) are located in a separate plane compared to the NF200-positive fibers (green; Fig. 6E, F). Similar results were found when double labeling with a monoclonal anti-NKAα3 and a polyclonal anti-peripherin (data not shown). Together, these experiments suggest that the NKAα3 is not expressed in type II afferent fibers that contact the OHCs.

FIG. 6.
figure 6

NKAα3 is not present in the type II afferent fibers. To determine if NKAα3-immunoreactivity detected in the outer hair cell region may also be due to expression of the NKAα3 in the type II afferent fibers that contact the outer hair cells, double labeling of apical cochlear turns with a monoclonal anti-NKAα3 (red; A) and a polyclonal anti-NF200 (green; B) was performed. High-magnification micrographs show that the NKAα3-positive fibers are distinct from the NF200-positive fibers that form the outer spiral bundles under the outer hair cells (C). This lack of colocalization is better viewed in the x-axis cross section through the z-projection, which reveals that the NKAα3-positive fibers (red; D) relative to the NF200-positive fibers (green; E) are located in a separate plane (F). These results suggest that the NKAα3 is not expressed in type II afferent fibers that contact the outer hair cells. For orientation, the location of the three rows of the outer spiral bundle (OSB) and the tunnel of Corti (ToC) have been indicated (with a triangle).

NKAα1 is present in supporting cells surrounding the IHCs

To determine if other isoforms of the NKAα are also expressed in the neuronal structures of the organ of Corti, we immunolabeled apical cochlear turns with various antibodies against NKAα1 and NKAα2. No immunoreactivity was detected for either of two antibodies used against the NKAα2 (sc-16049 and anti-HERED, data not shown). Interestingly, both monoclonal antibodies used against the NKAα1 showed patterns of immunoreactivity that were distinct from that seen for the NKAα3. Double labeling of organs of Corti with a monoclonal NKAα1 (C464.6; red; Fig. 7A) and a polyclonal anti-NKAα3 (green; Fig. 7B) show NKAα1 immunoreactivity in the IHC region that does not overlap with the NKAα3-positive type I afferent fibers that contact the IHCs (Fig. 7C). Subsequent double labeling experiments with a monoclonal anti-NKAα1 (C464.6; red, Fig. 7D) and a polyclonal anti-calretinin (green, Fig. 7E) show that NKAα1 immunoreactivity is specific to supporting cells surrounding the calretinin-positive IHCs, including the inner phalangeal cells that separate the IHCs (Fig. 7F). Inner phalangeal cells, in addition to closely neighboring supporting cells, are known to express the glutamate transporter GLAST and are thought to take up glutamate released from the IHCs (Furness and Lehre 1997; Hakuba et al. 2000; Furness and Lawton 2003; Glowatzki et al. 2006). To determine if the NKAα1-positive cells in the IHC region include the GLAST-expressing inner phalangeal cells, we performed double labeling experiments with a monoclonal anti-NKAα1 (C464.6; red, Fig. 7G) and a polyclonal anti-GLAST (green, Fig. 7H). The overlapping patterns of NKAα1 and GLAST immunoreactivity (Fig. 7I) suggest that the NKAα1 is also expressed by the GLAST-positive inner phalangeal cells and other supporting cells that surround the IHCs.

FIG. 7.
figure 7

NKAα1 is present in the GLAST-expressing inner phalangeal and supporting cells of the organ of Corti. Double labeling of apical cochlear turns with a monoclonal anti-NKAα1 (red; A) and a polyclonal anti-NKAα3 (green; B) show NKAα1 immunoreactivity in the IHC region that does not overlap with the NKAα3-positive type I afferent terminals that contact the IHCs (C). Subsequent double labeling experiments with a monoclonal anti-NKAα1 (red; D) and a polyclonal anti-calretinin (green; E) show that NKAα1 immunoreactivity is specific to cells surrounding and separating the calretinin-positive IHCs (F). The similar patterns of immunoreactivity seen in double labeling experiments with a monoclonal anti-NKAα1 (red; G) and a polyclonal anti-GLAST (green; H) identify the NKAα1-positive cells in the IHC region as the GLAST-expressing inner phalangeal and supporting cells of the organ of Corti (I). For orientation, the location of the inner phalangeal cells (IPCs) and single row of inner hair cells (IHCs) has been indicated (with a triangle).

Discussion

The aim of this study was to characterize the neuronal expression of NKAα in the organ of Corti and spiral ganglion. Compared to earlier work, our study utilized double immunofluorescent labeling and higher resolution micrographs of whole-mount preparations of the rat cochlea to unambiguously assign NKAα immunoreactivity to specific subtypes and subcellular structures of neurons. We identified NKAα3 in the somata and postsynaptic afferent terminals of type I SGNs that contact the IHCs and in the presynaptic medial efferent terminals that contact the OHCs. No NKAα3 immunoreactivity was detected in type II SGNs that contact the OHCs or in the lateral efferent terminals that contact the type I afferent terminals. Thus, NKAα3 expression appears restricted to the myelinated nerve fibers of the cochlea. Such somatic, axonal, and dendritic expression of NKAα3 has been reported in a variety of preparations (reviewed in Dobretsov and Stimers 2005). Thus, there does not seem to be a particular subcellular compartment to which NKAα3 expression is restricted in either the central or peripheral nervous system.

Our findings are consistent with earlier research using immunocytochemistry that identified the NKAα3 in the SGNs and cochlear nerve in the gerbil (Schulte and Adams 1989; McGuirt and Schulte 1994; Nakazawa et al. 1995), rat (ten Cate et al. 1994; Peters et al. 2001), and mouse (Erichsen et al. 1996). NKAα immunoreactivity was localized to the nerve terminals underlying the IHCs and OHCs in the human cochlea (Weber et al. 2001). Ultrastructural analysis attributed NKAα immunoreactivity in the terminals underlying the IHCs to the type I afferent and not lateral efferent terminals in the gerbil (Nakazawa et al. 1995). Additional work has shown that NKAα immunoreactivity in the terminals underlying the IHCs and OHCs is due to expression of the NKAα3 in both the gerbil (Schulte and Adams 1989; McGuirt and Schulte 1994) and rat (ten Cate et al. 1994).

We also tested for the presence of the other NKAα isoforms in the organ of Corti and spiral ganglion. We found NKAα1 in supporting cells surrounding the IHCs (see discussion below). We found no NKAα1 immunoreactivity in the afferent terminals. Previous studies have reported weak to moderate NKAα1 immunoreactivity in the SGNs and cochlear nerve of the rat (ten Cate et al. 1994; Peters et al. 2001; Delprat et al. 2007) and mouse (Erichsen et al. 1996), and one study reports expression of the NKAα in the afferent terminals underlying the OHCs in the gerbil (Nakazawa et al. 1995). Although we did not find NKAα1 immunoreactivity in either type I or II afferent terminals, we did see NKAα1 labeling in the spiral ganglion (data not shown) but were unable to identify unequivocally the NKAα1-expressing cells as either the SGNs or ensheathing satellite cells. In the region of the spiral ganglion, NKAα1 immunoreactivity in the gerbil cochlea (McGuirt and Schulte 1994; Schulte and Steel 1994) and NKAα immunoreactivity in the human cochlea (Weber et al. 2001) were attributed to the satellite cells. We were unable to detect NKAα2 immunoreactivity in the organ of Corti or spiral ganglion with either of two antibodies used (data not shown). While some groups have reported exclusive expression of the NKAα2 in the supralimbal fibrocytes (McGuirt and Schulte 1994; Schulte and Steel 1994; Peters et al. 2001), others have reported additional weak or transient NKAα2 immunoreactivity in the cochlear nerve (ten Cate et al. 1994; Erichsen et al. 1996).

Our results suggest that the myelinated nerve fibers of the organ of Corti express only NKAα3. Although NKAα1 is often considered to be ubiquitously expressed, the exclusive expression of NKAα3 has been observed in a variety of neurons (McGrail et al. 1991) including those that maintain tonic and/or rapid rates of firing, such as the afferent and efferent nerve fibers innervating the muscle stretch receptor (Dobretsov et al. 2003), neurons of the rat medial geniculate nucleus in auditory thalamus (Senatorov et al. 1997), as well as the calyx of Held afferent terminals in the medial nucleus of the trapezoid body in auditory brainstem (Kim et al. 2007). The selective expression of NKAα3 provides unique insight into the physiological demands on the myelinated fibers in the cochlea. The NKAα3 is distinguished from the NKAα1 by its relatively low affinity for intracellular sodium (Jewell and Lingrel 1991; Munzer et al. 1994; Zahler et al. 1997; Crambert et al. 2000), high affinity for ATP (Jewell and Lingrel 1991), and lack of inhibition at hyperpolarized potentials (Balshaw et al. 2000; Crambert et al. 2000). Thus, NKAα3 may be especially suited for the myelinated fiber types of the cochlea that sustain high rates of activity that would be expected to raise intracellular sodium concentrations, deplete ATP, and maintain hyperpolarized membrane potentials. In fact, the NKAα1 isoform, which has a much steeper voltage dependence of its transport rate compared to the NKAα3 isoform, may respond too quickly to action potential depolarization and prevent sustained firing (Dobretsov and Stimers 2005).

In testing for other NKAα isoforms in the organ of Corti, we found NKAα1 immunoreactivity in the glial-like supporting cells surrounding the IHCs, including the inner phalangeal cells. Previous ultrastructural work reported NKAα1 immunoreactivity of the inner phalangeal and border cells in rat (Iwano et al. 1990). These supporting cells express the glutamate transporter GLAST and are thought to remove glutamate from the extracellular space that has been released from the IHC (Furness and Lehre 1997; Hakuba et al. 2000; Furness and Lawton 2003; Glowatzki et al. 2006). The expression of NKAα1 in these supporting cells is not unexpected: glutamate transporters rely on sodium gradients to transport glutamate (Danbolt 2001). The NKAα inhibitors ouabain and vanadate have been shown to block the clearance of extracellular glutamate in cultured rat cortical astrocytes, suggesting that glutamate transporters are dependent on NKAα activity (Abe and Saito 2000). Our experiments suggest that GLAST-mediated glutamate uptake by the supporting cells adjacent to IHCs may be dependent upon the activity of NKAα1.

Our knowledge about the functional involvement of the NKAα in hearing has come from experiments involving inhibitors of the NKAα and also from experiments using mice with genetic deletions of the NKAα. Perilymphatic perfusion of ouabain, an inhibitor of the NKAα, causes a rapid reduction of the endocochlear potential (Konishi and Mendelsohn 1970; Kuijpers and Bonting 1970; Kusakari et al. 1978), and researchers have identified NKAα1 in the strial marginal cells and NKAα2 in the lateral wall fibrocytes (McGuirt and Schulte 1994; Schulte and Steel 1994; reviewed in Wangemann 2006). The contribution of the NKAα1 and NKAα2 in maintaining the endocochlear potential is supported by the characterization of mice carrying deletions of these subunits. Although homozygous deletion of either NKAα1 or NKAα2 is lethal at birth, heterozygous deletions are viable (Moseley et al. 2007). Despite normal cochlear morphology, NKAα1+/− and NKAα2+/− mice show reductions in the endocochlear potential and progressive, age-dependent hearing loss (Diaz et al. 2007). A possible neuronal contribution to hearing loss in these mice has not been excluded.

The contribution of the NKAα3 to hearing is less obvious. Homozygous deletion of NKAα3 is lethal; however, heterozygous deletion is viable (Moseley et al. 2007). The hearing of NKAα3+/− mice has not been reported. More recent studies investigating the perfusion of ouabain directly through the round window of the gerbil cochlea provide evidence for an additional neuronal role of the NKAα in the cochlea. Specifically, these experiments show threshold increases or complete abolishment of the compound action potential without a reduction of the endocochlear potential (Schmiedt et al. 2002; Wang et al. 2006). Moreover, chronic exposure to ouabain is associated with degeneration of type I but not type II SGNs in the gerbil cochlea (Lang et al. 2005). These findings are consistent with our observation of the NKAα3 in type I but not type II spiral ganglion afferent terminals. Especially in rodents, the NKAα3 isoform is much more sensitive to inhibition by ouabain than the NKAα1 isoform and slightly more sensitive than the NKAα2 isoform (O’Brien et al. 1994). Thus, specific doses of ouabain could be expected to inhibit the NKAα3 isoform and reduce or abolish the compound action potential without necessarily affecting the NKAα1-and NKAα2-dependent endocochlear potential.

The expression pattern of the NKAα in the organ of Corti and spiral ganglion suggests that both the NKAα1 and NKAα3 are poised to play an essential role in the regulation of the type I afferent synapses, the medial efferent synapses, and also glutamate transport from the afferent-IHC synapse. Thus, in addition to maintaining the endocochlear potential, the NKAα may contribute directly to the functioning of these synapses in the inner ear. Future investigation of the particular β (1–3) and FXYD (1–7) subunits coexpressed with the NKAα1 and NKAα3 will identify the additional molecular players involved in tailoring the NKAα transport properties at each of these synapses. Ultimately, clarifying the molecular components and neuronal contribution of the NKA in the cochlea will suggest ways in which its activity may be altered by exogenous inhibitors as a possible means to treat hearing loss (Aperia 2007) and also, very excitingly, by naturally occurring, endogenously expressed inhibitors (Nesher et al. 2007; Schoner and Scheiner-Bobis 2007) that may contribute to both normal and pathological hearing states.