| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Department of Neuroscience and 2Bobby R. Alford Department of Otorhinolaryngology—Head and Neck Surgery, Baylor College of Medicine, Houston, Texas; and 3Department of Anatomy and Cell Biology, University of Illinois College of Medicine, Chicago, Illinois
Submitted 21 June 2006; accepted in final form 19 October 2006
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
; Masetto et al. 2003; Sugihara and Furukawa 1989; Witt et al. 1994); TTX-sensitive currents that inactivate at less negative potentials (Chabbert et al. 2003; Marcotti et al. 2003); and TTX-insensitive currents that inactivate at very negative potentials (Géléoc et al. 2004; Oliver et al. 1997; Rüsch and Eatock 1997).
Two kinds of Na+ current have been described in immature rodent vestibular and cochlear hair cells but never in one report. A TTX-insensitive, very negatively inactivating current was reported by Rüsch and Eatock (1997) in hair cells of the immature mouse utricular macula [postnatal days 0–10 (P0–P10)]. In the same preparation, Géléoc et al. (2004) observed that this current peaked at embryonic day (E) 16 and decreased dramatically by P0 (E20). A Na+ current in outer hair cells of the immature rat cochlea has similar TTX sensitivity and voltage dependence (P0–P11) (Oliver et al. 1997). More recently, however, related preparations—the immature rat utricular macula and inner hair cells of the immature mouse cochlea—have yielded a TTX-sensitive and less-negatively inactivating Na+ current (Chabbert et al. 2003; Marcotti et al. 2003). Did experimental manipulations alter the currents' properties? Does expression differ between rats and mice? Or was Na+ current heterogeneity missed? For the utricle, our results favor the last explanation. We show that in the rat utricular macula, both currents are expressed between P0 and P9 but in different hair cells; which current is expressed varies with postnatal age, location in the epithelium, and hair cell type.
Voltage-gated Na+ channels comprise pore-forming (
) subunits and sometimes accessory (
) subunits that can modulate the behavior or expression of the subunits. Chabbert et al. (2003) provided RT-PCR evidence for multiple TTX-sensitive subunits in individual immature rat utricular hair cells. Here we show that the TTX-insensitive current is likely to be carried by NaV1.5 subunits, originally described in heart muscle (Rogart et al. 1989).
In cochlear hair cells, a developmental decline in Na+ current amplitude coincides with changes in Ca2+ and K+ channels with the net effect being a reduced tendency to fire action potentials (Marcotti et al. 2003). These changes occur around the onset of hearing and may mark the transition from a nonsensing epithelium, in which spikes contribute to development, to a sensing epithelium, in which spiking may interfere with outer hair cell electromotility or the graded representation of sounds. In the rat utricular macula, we see a comparable reduction in Na+ currents with a similar time frame, reflecting complete loss of the TTX-sensitive current and a decline in size of the TTX-insensitive current.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Electrophysiology
PREPARATIONS.
Most recordings were from hair cells in excised, semi-intact preparations of the rat utricular epithelium, prepared as described previously for the mouse utricular epithelium (Vollrath and Eatock 2003). The sensory part of the epithelium is called the macula. Some recordings were obtained from solitary hair cells isolated from the rat utricular macula as described previously (Wong et al. 2004).
In both types of experiment, Long-Evans rats (P0–P22, Charles River Laboratories, Wilmington, MA) were anesthetized by cooling (<P4) or intraperitoneal injection of pentobarbital sodium (Nembutal, 50 mg/kg) and decapitated. The remainder of the dissection was carried out in our standard external solution (K+-SES, see SOLUTIONS). The head was bisected, the brain was removed, and the otic capsule was opened to expose the membranous inner ear. The endolymphatic compartment of the utricle was opened and bathed for 10 min at room temperature in extracellular solution containing 100 µg/ml protease XXIV, which facilitated removal of the otolithic gel overlying the hair bundles. The utricular epithelium was excised from the otic capsule into a chamber containing K+-SES.
For the semi-intact preparation, the epithelium was trimmed, affixed with CellTak (BD Biosciences, Bedford, MA) to a coverslip in a glass-bottomed experimental chamber, and observed at x630 or x1,000 on a fixed-stage upright microscope (Axioskop FS, Carl Zeiss, Thornwood NY) with water-immersion objectives and differential interference contrast optics.
For a minority of experiments, we used isolated hair cells. To obtain these, we placed the excised utricular epithelium in K+-SES containing crude papain (500 µg/ml) and L-cysteine (300 µg/ml) for 40–60 min at 37°C. The epithelium was then transferred to external solution containing bovine serum albumin (500 µg/ml) for 10 min at room temperature (22–24°C), and finally to a recording chamber fit with a glass coverslip. The hair cells were mechanically dispersed with a fine probe and viewed at x400 or x600 on an inverted microscope with differential interference contrast optics (IMT-2, Olympus, Lake Success, NY).
SOLUTIONS.
The potassium standard external solution (K+-SES) used for dissections contained (in mM) 144 NaCl, 0.7 NaH2PO4, 5.8 KCl, 1.3 CaCl2, 0.9 MgCl2, 5.6 D-glucose, and 10 HEPES and vitamins and amino acids as in Eagle's minimal essential medium (MEM); pH 7.4 with 
7.5 mmol NaOH, 310 mmol/kg.
For voltage-clamp recording of Na+ currents, the standard external solution was altered to minimize K+ currents by adding the K+ channel blocker, 4-aminopyridine (4-AP) and replacing K+ with the less-permeant ion, Cs+. The standard external solution for recording (Cs+-SES) contained (in mM): 130 NaCl, 5 4-AP, 5 CsCl, 1 MgCl2, 5 CaCl2, 10 HEPES, and 5.6 D-glucose and vitamins and amino acids as in Eagle's MEM, pH 7.4 with 2.5 mmol NaOH, 310 mmol/kg. The experimental chamber was constantly perfused with Cs+-SES. In several voltage-clamp experiments, recordings were made in K+-SES with 1.3 mM CaCl2 and 0.9 mM MgCl2; Na+ currents were isolated kinetically from the slower K+ currents.
We used the ruptured-patch method of whole cell recording. In voltage-clamp experiments of Na+ currents, the electrode usually contained the standard internal solution (Cs+-SIS) comprising (in mM) 135 CsCl, 3.5 MgCl2, 5 HEPES, 0.2 EGTA, 5 Na2ATP, 0.1 LixGTP, and 0.1 Na-cAMP, pH 7.4 with 5 mmol CsOH, 280 mmol/kg. For current-clamp recordings, the external solution was K+-SES and the electrode contained K+-SIS (in mM): 140 KCl, 3.5 MgCl2, 2.5 Na2ATP, 5 HEPES, 10 EGTA, 0.1 Na-cAMP, and 0.1 LixGTP, pH 7.4 with KOH (25 mM), 290 mmol/kg.
RECORDINGS.
All recordings were performed at room temperature (22–25°C). For the semi-intact preparation, we pulled recording electrodes from R-6 glass (Garner Glass, Claremont, CA); electrodes had resistances ranging from 2.5 to 4 M
in our standard solutions. We used the same method to record from hair cells as described in Vollrath and Eatock (2003) for hair cells in the semi-intact mouse utricular macula: we lowered the recording pipette into the epithelium 30 µm from the cell of interest, maintaining positive pressure on the pipette to clear a path for the electrode's advance within the epithelium. The electrode was advanced in between supporting and hair cells to its target membrane; if the hair cell of interest had a calyx, that calyx was dying because contact had been severed with the ganglion cell bodies and it was relatively easy to dislodge the calyx from the hair cell membrane. Acquisition of a hair cell was confirmed visually by focusing up and down from hair bundle to pipette location; also, at the end of each recording, the pipette lifted its attached hair cell—with visible bundle—out of the epithelium. Ionic currents were recorded in whole cell mode with the EPC-10 amplifier with an integrated interface (HEKA Instruments, Southboro, MA), controlled by a Pentium IV PC (Dell, Round Rock, TX) running Pulse 8.62 software (HEKA). Currents were filtered with an integrated four-pole Bessel filter at 8.3 kHz. Average residual series resistance, after electronic compensation, was 4.2 ± 0.14 (SE) M (range: 1.3–13.4 M, n = 184 cells). The mean clamp rise time (the product of residual series resistance and cell capacitance) was 27 ± 0.9 µs (range: 9.4–84.2 µs, n = 184). Potentials are corrected for liquid junction potentials, calculated with JPCalc software (Barry 1994), of –7.4 mV between Cs+-SES and Cs+-SIS and –5.4 mV between K+-SES and K+-SIS. Thus the holding potential in Cs+-based solutions was –67.4 mV. Corrections for liquid junction potentials were also applied to voltages recorded in current-clamp mode. All recorded currents were leak-subtracted with a +P,–P/4 protocol (Armstrong and Bezanilla 1974); the original, nonsubtracted data were also stored. Data were analyzed with PulseFit 8.62 (HEKA), Origin 7.0 (OriginLab, Northampton, MA) and Excel 2002 (Microsoft, Redmond, WA) software.
Hair cells in the semi-intact preparation were classified according to their macular location and cell type. The utricular macula in mammals has a quasi-central stripe, the striola, with distinctive morphology and physiology (illustrated in Fig. 6) (reviewed in Eatock and Lysakowski 2006). Hair bundles reverse orientation within the striola. Location was defined relative to the line of bundle reversal: hair cells within three cells of the reversal line were considered striolar and cells further than six cells from the line of reversal were considered extrastriolar, based on calretinin staining of calyx-only afferents (Desai et al. 2005b). Hair cells at intermediate locations and at the far edges of the epithelium were avoided. The parts of the extrastriola lateral and medial to the striola were distinguished as illustrated in Fig. 6, inset.
|
|
|
|
|
|
). Type I hair cells have an unusually large and negatively activating K+ conductance, gK,L, and are contacted by large calyceal terminals. Type II cells have smaller K+ conductances at resting potential and are contacted by bouton terminals. In most of our electrophysiological experiments, we could not use K+ currents to distinguish hair cell type as they were blocked. Instead, we classified hair cells as type I if they were contacted by calyces and as type II if they were from P9 or older animals and lacked calyces. In 184 cells, there were 50 type I (39 striolar, 11 extrastriolar) and 29 type II, leaving 105 cells unclassified. Although a few calyces are seen in the rat utricular macula as early as P0 (Gaboyard et al. 2003), in our electrophysiological sample, most (43/50) hair cells with calyces came from animals aged P3 or older. The seven cells with calyces between P0 and P2 all came from the striola, consistent with its faster development (Rüsch et al. 1998; Sans and Chat 1982).
For recording from isolated hair cells, we pulled recording electrodes from filamented thin-walled glass (TW150F-4, WPI, Sarasota, FL); pipette resistances were 1–3 M in standard recording solutions. Recordings were made with the ruptured-patch method and similar solutions to those used in the semi-intact preparation. Currents were amplified with an Axopatch 200B amplifier and digitized by a DigiData 1200 interface (Axon Instruments, Foster City, CA) controlled by Clampex 8 software (Axon). The amplifier output was low-pass filtered at 10 kHz. Data were analyzed with Clampfit 8 (Axon), Origin 7.0 (OriginLab), and Excel 2002 (Microsoft) software. Isolated hair cells were identified as type I or II by cell shape.
ANALYSIS. Results are expressed as means ± SE. Significance was determined with the Students' t-test or one- or two-way ANOVAs with post hoc Tukey means comparisons as implemented by Origin software.
To generate activation curves [conductance-voltage, g(V), relations] for Na+ current, we converted peak inward currents to conductances by dividing by the driving force, (V – ENa), where V is membrane potential, and ENa is the Na+ equilibrium potential (+65 mV for the Cs+-based solutions and +80 mV for the K+-based solutions). Calculated conductances were plotted against voltage and the resulting curves were fit with a single Boltzmann function (Eq. 1)
 | (1) |
Na+ currents are rapidly inactivating. To study the voltage dependence of inactivation, we stepped to a near maximally activating voltage after an iterated prepulse potential and plotted the peak current as a function of prepulse voltage (inactivation curves). The curves were fit by a single Boltzmann function
 | (2) |
 | (3) |
The midpoints (V1/2 values) of single Boltzmann fits to inactivation curves had a bimodal distribution; each component was fit with a Gaussian function (Eq. 4)
 | (4) |
The inactivation time course was fit with a single-exponential function (Eq. 5)
 | (5) |
is the time constant. The recovery from inactivation was fit either with Eq. 5 or with a double-exponential function
 | (6) |
1 and 2 are time constants for each term. RT-PCR
We used the reverse-transcription polymerase chain reaction, RT-PCR, method to test for the expression of mRNA corresponding to nine known Na+ channel subunits and four subunits (Table 1). We studied utricular maculae and the sensory epithelia (cristae) of all three semicircular canals from P1 and P21 rats. In some cases, vestibular ganglia were used as a positive control (see following text).
|
). The superior division of the vestibular ganglion was also excised. The epithelia were peeled from the basement membranes by lifting with a fine eyelash and the peeled epithelia and ganglia placed in separate RNase/DNase-free tubes. Excess liquid was removed, and the tissue samples were placed on dry ice. We used the RNeasy kit (QIAGEN, Valencia, CA) to isolate RNA from utricular maculae (3 pooled at a time), cristae (6 pooled at a time) and vestibular ganglia (1 at a time). The tissue was homogenized in lysis buffer with a mortar and pestle and a QIAshredder column (QIAGEN). An equal volume of 100% ethanol was added, the combined solution was placed on an RNeasy column and washed with RW1 solution (QIAGEN), and the RNA was eluted in RNase-free H2O. The resulting RNA solution was reverse-transcribed to complementary (c) DNA using the Advantage RT-for-PCR Kit (BD Biosystems, Palo Alto, CA) with the Moloney murine leukemia virus (MMLV) reverse transcriptase and random hexamer primers.
We controlled for possible contamination of samples by genomic DNA in several ways. 1) Primer design: primer sets were designed across intron-exon boundaries as determined from rat sequences. Genomic DNA contamination would be revealed as higher molecular weight bands than the predicted PCR product. None were observed. 2) Tissue treatment: during the RNA isolation process, tissue homogenization helped break up genomic DNA. In addition, RNeasy columns contain a silica membrane that is designed to eliminate most DNA from the sample. Residual DNA was removed by treating the column with RNase-free DNase I (QIAGEN) for 30 min before eluting the RNA. 3) Negative RT controls: for each primer set, samples of inner ear tissue were prepared as described in the preceding text, substituting water for the MMLV reverse transcriptase. Thus only genomic DNA would be present as a template for amplification. No bands were observed on agarose gels (data not shown).
PCR was done with a PTC-100 thermocycler (MJ Research, Reno, NV) using the TAQ enzyme (Applied Biosystems, Foster City, CA) and the various primer sets (IDT, Coralville, IA) listed in Table 1. A "hot start" (94°C for 4 min) reduced mis-priming. The PCR protocols comprised 40 cycles of 94°C for 1 min, 58–62°C (annealing temperature) for 30 s, and 72°C for 35 s; plus a final 7 min at 72°C. PCR products were resolved on 1.2% agarose gels and visualized with ethidium bromide.
The PCR product for each primer set was sequenced at least once; gel-purified products were sequenced from the utricular macula where present, or otherwise from vestibular ganglia (SeqWright, Houston, TX). Each primer set was tested on each tissue type (utricular macula, crista) 2–14 times. For all but three primer sets, we simultaneously tested for expression in standard tissues (brain, heart or skeletal muscle), as a control for primer quality. NaV1.7, 1.8, and 1.9 are weakly expressed if at all in these standard tissues. In contrast, we found robust expression in vestibular ganglia, which we confirmed by direct sequencing. We therefore used the vestibular ganglia as our "control" tissue for the NaV1.7, 1.8, and 1.9 primer sets. As a positive control for sample tissue quality, we tested for expression of the L-type Ca2+ channel subunit, CaV1.3 (1D), which we have found to be robustly expressed in maculae, cristae and, contrary to our initial observations (Bao et al. 2003), in vestibular ganglia. The same CaV1.3 primers were used as in Bao et al. (2003) (Table 1).
SINGLE-CELL RT-PCR.
We collected individual hair cells for RT-PCR analysis from intact P1 epithelia prepared as for electrophysiological experiments. Electrode glass was cleaned and treated with RNaseZAP, rinsed with nuclease-free water and dried overnight. Low resistance pipettes (1 M) were pulled and the tips filled with 8 µl K+-SIS. Individual cells of known region and cell type were sucked into the pipette; we did not record from them. Each cell was immediately placed in 1 µl RNase inhibitor and flash-frozen at –80°C.
Reverse transcription of RNA and amplification of specific DNA products were achieved using One-step RT-PCR (QIAGEN). To amplify PCR products corresponding to NaV1.2 or NaV1.5 subunits, we used a degenerate primer set (based on Chabbert et al. 2003) for the first round of amplification and subunit-specific primers for the second round (Table 1). To check the viability of individual cells, we included a primer set for -actin, as well as the Na+ channel degenerate primer set, in the first round PCR. The PCR protocols were: first round: 55°C for 30 min (for reverse transcription); 95°C for 15 min; 40 cycles of 94°C for 1 min, 52°C for 1 min, 72°C for 1 min; 72°C for 10 min; second round: 95°C for 15 min; 40 cycles of 94°C for 1 min, 52°C for 1 min, 72°C for 1 min; 72°C for 10 min. Each PCR product was sequenced as described for whole epithelia.
Immunohistochemistry
Long-Evans rats of various ages (P0–P21) were deeply anesthetized with Nembutal (80 mg/kg), then perfused transcardially with 10–100 ml physiological saline containing heparin (2,000 IU), followed by 2 ml/g body wt of an aldehyde fixative (4% paraformaldehyde, 1% acrolein in 0.1 M phosphate buffer (PB) with 1% picric acid and 5% sucrose, pH 7.4). Vestibular epithelia were dissected out in PB and cryo-protected in 30% sucrose-PB. Otoconia were eliminated with undiluted Cal-Ex for 1–10 min (Fisher Scientific, Pittsburgh, PA). Background fluorescence was reduced by incubating the tissues in 1% aqueous solution of sodium borohydride for 10 min. Frozen sections (35 µm) were cut with a sliding microtome.
Antibodies were from Chemicon (Temecula, CA) unless otherwise specified. Immunocytochemistry was done on free-floating sections or whole organs, permeabilized with Triton X-100 in a blocking solution of 0.5% fish gelatin and 1% BSA in phosphate-buffered saline (PBS). Samples of vestibular tissues were incubated with Triton X-100 at conditions that varied with postnatal age: P0–P1: 0.3% overnight at 4°C; P3-4: 0.5% for 1 h at room temperature (RT); P6-P8: 2% for 1 h at RT; P21: 4% for 1 h at RT. (For the younger ages, morphology was better preserved by decreasing the detergent concentration and increasing the incubation time.) Samples were then incubated with a cocktail of two primary antibodies diluted in the blocking solution: goat anti-calretinin and rabbit antibody against NaV1.5 (1:200) or NaV1.2 (1:75; Sigma, St. Louis, MO) or NaV1.6 (1:200) for 2 days at 4°C with 0.1, 0.3, and 0.5% Triton X-100 for P0–P1, P4–P21 sections, respectively. We used calretinin antibody as a marker of immature hair cells (P0–P6), type II cells (P8–P21) and calyx afferents (P4–P21). (The calretinin label is not always shown.) Specific labeling was revealed with a cocktail of two secondary antibodies: fluorescein-conjugated donkey anti-goat and rhodamine-conjugated donkey anti-rabbit (1:200 in the blocking solution). Sections were rinsed with PBS between and after incubations and mounted on slides in Mowiol (Calbiochem, Darmstadt, Germany). The sections were examined at an optical section thickness of 0.4–1 µm, depending on the magnification, on a laser scanning confocal microscope (LSM 510 META, Carl Zeiss, Oberköchen, Germany). Final image processing was done with Adobe Photoshop software (San Jose, CA).
For each antibody, we did control reactions to test for nonspecific labeling with no primary antibody and with primary antibodies preincubated with their antigenic peptide (10 µg/1 µg antibody) for 2 h at RT. Images comparing staining in control and test conditions were acquired and digitally processed identically. We also did Western blots on inner ear epithelia obtained from five adult rats (> P30; 250–300 g; NaV1.5) and six P10 rats (NaV1.2 and NaV1.6) to check that the antibody recognized a protein of the appropriate size. Membrane proteins were isolated from control tissues (heart, cerebellum, liver) using a method adapted from Moore et al. (1998). Inner ear tissues were homogenized separately in the same homogenization buffer used for the membrane preparations. Aliquots containing 50 µg of protein homogenate were mixed with 5x SDS loading buffer. The inner ear tissue was incubated at 37°C for 5 min, while the control tissues were boiled for 5 min, then both were loaded onto 4–15% Tris-HCl mini-gel wells. After electrophoresis, the proteins were transferred from the gel onto nitrocellulose membrane overnight. The membrane was then blocked with 5% milk solution for 2 h, incubated in primary antibody solution for 2 h (1:200), incubated in secondary antibody solution for 1 h (1:30,000), and washed thoroughly with TBS-Tween. Bands were visualized with chemiluminescent detection (Amersham, Little Chalfont, UK). Western blots for all three Na+ channel antibodies (data not shown) had bands at the appropriate size in the inner ear and positive control tissues (heart and cerebellum) and not in the negative control tissue (liver): NaV1.2, 228 kDa; NaV1.5, 227 kDa; NaV1.6, 226 kDa. For the NaV1.2 and NaV1.5 antibodies, we further showed that the band at the correct size was selectively blocked by preabsorption with the antigenic peptide.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
TRANSIENT CURRENT IN HAIR CELLS REQUIRED EXTERNAL NA+. Whole cell currents were recorded in situ from 184 hair cells in the utricular maculae of rats between P0 and P22. The external and internal solutions were designed to minimize K+ currents. In almost all cells (91%, 167/184), depolarizing pulses following a hyperpolarizing step evoked rapidly activating and inactivating inward currents (Figs. 1A and 2A), often accompanied by a small steady current (Fig. 1A). The peak inward current is plotted as a function of voltage in Fig. 1E. Replacement of external Na+ with the impermeant cation, N-methyl-D-glucamine+, NMDG+, eliminated the transient component (n = 4 cells, P1, Fig. 1B), showing that it was carried by Na+. As shown later, the transient current was also sensitive to the Na+ channel blocker, TTX.
The residual inward current in NMDG+ (Fig. 1, B and E) was carried by voltage-gated Ca2+ channels that show little rapid inactivation (Bao et al. 2003). It was difficult to eliminate Ca2+ current in these cells for two reasons. First, one component of the Na+ current was sensitive to the standard Ca2+ channel blocker, cadmium (Cd2+) (Wooltorton et al. 2007). Second, the Ca2+ current in mammalian vestibular hair cells is only partly blocked by high doses of dihydropyridines (Bao et al. 2003; Dou et al. 2004). Thus most of our records include a small Ca2+ current component; as discussed later (Effect of Ca2+ current), it had little impact on our Na+ current measurements.
The NMDG+-sensitive current (Fig. 1D) included, in addition to the expected transient inward current, a small sustained component which may be a "persistent" Na+ current similar to that described in mammalian neurons (e.g., Crill 1996; Do and Bean 2004; Vreugdenhil et al. 2004) and expression systems (e.g., Mantegazza et al. 2005; Qu et al. 2001). In four isolated rat utricular hair cells, the maximum persistent current was 10 ± 2.8% of the peak transient NMDG+-sensitive (difference) current. This is a larger percentage than reported for persistent currents in other cell types (1.5–5%). In the hair cells, some of the steady-state NMDG+ difference current may have been Ca2+ current that ran down during the NMDG+ application and so appeared to be blocked by NMDG+. For example, in the hair cell in Fig. 1, the maximum steady-state current ran down 17% from the control records to the wash records. We did not study the persistent current further. A persistent NMDG+-sensitive component can also be seen in recordings from rat outer hair cells (Oliver et al. 1997) (estimated at 7% of peak current from their Fig. 1, A and B).
VOLTAGE DEPENDENCE OF TRANSIENT INWARD CURRENTS. Inactivation protocols (Fig. 2B) were used to generate inactivation curves for all 167 hair cells with Na+ current. For 97 of these cells, activation curves were also generated from activation protocols (Fig. 2A). Conductances were calculated from peak currents and driving forces and normalized (Fig. 2C). Most activation and inactivation curves were well fit by a single Boltzmann function (Eq. 1, METHODS, Fig. 2C). Figure 2 also shows the distributions of V1/2 and S values for inactivation (Fig. 2, D and E) and V1/2,act values (Fig. 2G) for hair cells between P0 and P4. We restricted the histograms to this age range because, as shown later (Fig. 8C), values change with age.
|
Although the bimodal V1/2,inact distribution is consistent with most hair cells expressing one or the other current, the distribution of S values (Fig. 2E), a measure of the width of the curves, suggests that some hair cells may have expressed both currents. For V1/2,inact values between –90 and –75 mV (double-headed arrow in Fig. 2E), inactivation curves tended to be broader than those for cells at the extremes of the distribution (P < 0.05; 1-way ANOVA). S values at –82 mV were significantly broader than those between –66 and –75 mV and between –87 and –99 mV. S values at –85 mV were significantly broader than those between –69 and –75 mV and between –90 and –96 mV (P < 0.05; Tukey post hoc means comparison). The mean S values at the extremes also differed significantly (P < 0.01): S was 6.1 ± 0.15 mV for pooled V1/2,inact values negative to –90 mV (n = 46) and 5.3 ± 0.18 mV for pooled V1/2,inact values positive to –75 mV (n = 26). A plausible explanation for the broader inactivation ranges of the intermediate cells is that they expressed appreciable amounts of both currents, as illustrated for one cell in Fig. 2F (open circles). Its inactivation curve was well fit either by a single Boltzmann with a large S value (8.7 mV; not shown) or by a sum of two Boltzmanns with the V1/2,inact and S values set at the means for INa,1 and INa,2 (thick curve).
The distributions of activation V1/2 values for the two currents showed more overlap than did the inactivation ranges but also were consistent with two populations (Fig. 2G). Mean Boltzmann parameters for activation curves were: INa,1: V1/2,act = –38 ± 0.5 mV, S = 7.9 ± 0.16 mV, gmax = 13.1 ± 1.07 nS (n = 54 cells); INa,2: V1/2,act = –31 ± 0.6 mV, S = 6.7 ± 0.21 mV, gmax = 10.3 ± 1.32 nS (n = 28 cells). Again, the V1/2 values, but not S or gmax, differ significantly (P < 0.001). Note that gmax values from activation curves were larger than those from inactivation curves, which were obtained for a sub-maximal depolarization (to –17 mV). Inactivation gmax was 83 ± 2.2% of activation gmax for INa,1 (n = 54) and 77 ± 3.7% for INa,2 (n = 28).
Some experiments were done on hair cells dissociated from the utricular macula (Figs. 1 and 5). The maculae were treated with papain, which is known to affect the voltage dependence of other channels (Armstrong and Roberts 1988). We found that the voltage dependence of Na+ current inactivation in 12 papain-dissociated hair cells (P1–P2) was consistent with two populations of Na+ channels, but the peak currents were smaller and inactivation ranges were shifted negatively relative to data recorded in situ. For the more negative group of dissociated cells (n = 8, P1–P2), V1/2,inact = –109 ± 2.3 mV, 18 mV negative to the mean for INa,1 from the intact macula at the same ages (–91 ± 0.8 mV; n = 39; P < 0.001). The maximum current density was also 37% smaller in the dissociated cells: –78 ± 10.4 versus –123 ± 14.6 pA/pF; P < 0.05. For the more positive group of dissociated cells (n = 4, P1–P2) V1/2,inact values were significantly more negative than in situ values for INa,2 over the same age range: –82 ± 2.0 mV (n = 4) versus –74 ± 1.0 mV (n = 11; P < 0.05). Maximum current density was also lower in the dissociated cells (–49 ± 9.9 pA/pF, n = 4, vs. –120 ± 39 pA/pF, n = 11; P < 0.05). In contrast, V1/2,act values for both groups of dissociated hair cells did not differ significantly from those recorded in situ. Thus for both INa,1 and INa,2, papain dissociation negatively shifted the voltage dependence of inactivation and reduced total current but did not selectively eliminate either type of Na+ current or affect the voltage dependence of activation.
Our usual external medium, Cs+-SES, contained elevated Ca2+ (5 mM), which might shift voltage dependence positively (Hille 2001). We obtained some measurements in physiological divalent levels (1.3 mM Ca2+) for comparison. V1/2,inact values for INa,1 were 4 mV more negative (P < 0.001): –96 ± 0.9 mV (n = 20) versus –92 ± 0.6 mV (n = 68) in 5 mM Ca2+. For INa,2, however, V1/2,inact values were unchanged: –75 ± 1.3 mV (1.3 mM Ca2+; n = 10) versus –75 ± 0.5 mV (5 mM Ca2+; n = 48).
EFFECT OF CA2+ CURRENT ON VOLTAGE DEPENDENCE. With our solutions and protocols, Ca2+ currents were elicited as well as Na+ currents. To check for effects of the Ca2+ current on our V1/2 and S values for the Na+ current, we isolated Ca2+ currents in two ways: by replacing Na+ with NMDG+ to eliminate Na+ current and by applying an activation protocol without a prepulse to inactivate INa,1 (INa,2 was not completely inactivated at the holding potential of –67 mV). Comparison of V1/2 and S values for total inward current and for currents with the Ca2+ component subtracted suggests that Ca2+ currents had little impact. For eight hair cells in the semi-intact utricular macula, there was no significant difference in V1/2,act (–38 ± 1.9 vs. –39 ± 1.9 mV), S (8.4 ± 0.40 vs. 7.9 ± 0.41 mV), or gmax (7.6 ± 1.90 vs. 7.8 ± 2.10 nS) between the records with a prepulse (Na+ plus Ca2+ current) and the difference records (prepulse - no pulse; Na+ current alone), respectively. For three isolated utricular hair cells in which currents were recorded in control and NMDG+ solutions, V1/2,act was –35 ± 1.5 mV and S was 8.1 ± 0.40 mV for the peak total inward current versus –35 ± 1.0 mV and 7.8 ± 0.45 mV for the peak NMDG+-blocked current. gmax was significantly reduced (P < 0.05; 6.0 ± 1.17 vs. 3.9 ± 0.68 nS), probably reflecting rundown of current in the NMDG+ condition.
ACTIVATION AND INACTIVATION KINETICS.
Some Na+ currents can be distinguished by variations in kinetics: for example, TTX-insensitive currents carried by NaV1.8 and NaV1.9 subunits have relatively slow kinetics (Dib-Hajj et al. 2002). Both INa,1 and INa,2 had faster kinetics than NaV1.8 or NaV1.9 channels. To compare kinetics of activation and inactivation and of recovery from inactivation of INa,1 and INa,2, we chose hair cells well separated in the distribution shown in Fig. 2D and therefore likely to express mostly one or the other current: six hair cells with INa,1 (range of V1/2,inact values: –90.2 to –93.4 mV) and six with INa,2 (range of V1/2,inact values: –69.9 to –74.3 mV). We fit the time course of current decay, recorded with the usual activation protocol, with a single-exponential function (Eq. 5; Fig. 2A, thick curve) and log(time constant) values were plotted against membrane potential (Fig. 3A). For both INa,1 and INa,2, the inactivation time constant and the time to peak, a measure of the speed of activation, became faster with more depolarized steps. Kinetic differences between the two currents were small but significant (2-way ANOVA on logarithmically transformed values of Fig. 3, A and B; P < 0.001). The difference did not reflect an average difference in uncompensated series resistance for the two groups, which produced similar average maximum voltage errors (the product of residual series resistance and maximum currents): 4.4 ± 0.89 mV for INa,1 (range: 2.9–7.1, n = 6) and 4.5 ± 1.29 mV for INa,2 (range: 2.4 –10.4, n = 6). The inact values for INa,2 are slightly slower than those for a similar Na+ current in mouse inner hair cells, reflecting at least in part the higher temperature of the inner hair cell recordings (34–37°C) (Marcotti et al. 2003). Times to peak for both INa,1 and INa,2 were shorter than the time to peak for INa in chick crista cells obtained with a similar protocol at room temperature (Masetto et al. 2003) (–120 mV prepulse): 400 versus 600 µs at approximately –10 mV.
For currents that are inactivated at resting potential, the time course of inactivation removal at hyperpolarized potentials is functionally relevant. We investigated the time required for a step to –137 mV to remove the inactivation accumulated at the holding potential (–67 mV) in 12 hair cells with INa,1 (Fig. 3C) and in five hair cells with INa,2 (Fig. 3D). In six hair cells with INa,1 and four hair cells with INa,2, the recovery of the peak response to a step to –17 mV was best fit with a double-exponential function (dashed fits in Fig. 3, C and D). Mean time constants were 1.1 ± 0.20 and 12.6 ± 2.52 ms for INa,1 and 0.5 ± 0.05 and 15.4 ± 3.83 ms for INa,2. The fast time constants for the two types of current differed significantly (P < 0.05). In both cases, the size of the fast component exceeded the size of the slower component by at least two orders of magnitude. The slow second time constant may represent recovery from a more inactivated state.
In the other hair cells (6 with INa,1 and 1 with INa,2), recovery followed a single-exponential time course with time constants of 1.8 ± 0.25 ms (INa,1) and 1.9 ms (INa,2). The dominant fast constants of recovery are consistent with data from type II hair cells of the chick crista (Masetto et al. 2003), after extrapolation for our more negative conditioning voltage.
EVOKED SPIKING IN CURRENT-CLAMP RECORDINGS.
Voltage-gated Na+ channels are usually assumed to play a role in action potential firing or spiking. We looked for spiking in current-clamp recordings made with K+-based, rather than Cs+-based, solutions (K+-SES, K+-SIS) from hair cells in the semi-intact utricular macula (10 cells with INa,1, mean input resistance 615 ± 117 M, zero-current potential –54 ± 4.1 mV; 8 cells with INa,2, mean input resistance 576 ± 45.8 M, zero-current potential –55 ± 1.9 mV). We never saw spontaneous spiking, and depolarizing current steps usually failed to evoke spikes unless preceded by a hyperpolarizing current step. In this regard, our data resemble previous recordings from immature hair cells of rat and mouse utricular macula (Chabbert et al. 2003; Géléoc et al. 2004), immature rat outer hair cells (Oliver et al. 1997), and more mature chick crista hair cells (Masetto et al. 2003).
Figure 4 shows, for a cell with INa,2 (A) and a cell with INa,1 (B), how the size and time course of the spike-like event at the onset of a large depolarizing current step depended on prestep voltage. [We consider these events to be spike-like because shoulders in the depolarizing phase (see arrows) indicate a threshold for regenerative current.] In three cells with INa,2, including the cell in Fig. 4A, the drop in peak potential with depolarization of the prestep interval (Fig. 4C) strongly overlapped the INa,2 inactivation range (V1/2 values of –68 ± 2.0 vs. –76 ± 1.3 mV, respectively). Thus the fall-off in peak potential with prestep depolarization may be attributable to INa,2 inactivation. Chabbert et al. (2003) blocked similar spikes in a rat utricular hair cell with INa,2 with 100 nM TTX. Thus Na+ channel openings for INa,2 might boost the spikes elicited from potentials negative to –60 mV. In contrast, when the same current protocol was applied to three cells with INa,1, the peak potential did not change over the inactivation range of INa,1 (Fig. 4, B and D), suggesting that INa,1 did not determine the peaks. The fall-off in peak potential for prestep potentials positive to –65 mV might reflect the activation of outwardly rectifying K+ channels, which can also be seen in the current record (arrows, Figs. 4, C and D).
THE TWO NA+ CURRENTS HAD DIFFERENT TETRODOTOXIN SENSITIVITIES.
TTX separates Na+ currents into those that are TTX sensitive, with KD's in the low nanomolar range, and those that are TTX insensitive, with KD's in the high nanomolar or micromolar range. We tested the TTX sensitivity of the Na+ currents in hair cells dissociated from the rat utricular macula. For five hair cells with INa,1, 500 nM TTX blocked Na current (INa,1) by just 53 ± 4.8% (Fig. 5A). If the underlying channels were a uniform population, this block is consistent with a dissociation constant, KD, in the TTX-insensitive range (440 nM, calculated assuming a 1:1 ratio of TTX molecules to channels) (Hille 2001). Although the presence of a small contaminating Ca2+ current could reduce the apparent block, note that in the example shown in Fig. 5A, the current in TTX had no sustained component—i.e., no Ca2+ current—and clearly had a transient component, i.e., Na+ current. Thus there is no doubt that the cell had a TTX-insensitive Na+ current.
For three hair cells with INa,2, a much lower concentration of TTX (50 nM) blocked the Na current (INa,2) by 76 ± 8.1% (Fig. 5B). Under the same assumptions, such a block is consistent with a TTX-sensitive KD of 16 nM. In the example in Fig. 5B, the presence of a TTX-insensitive sustained component (see data in TTX) is consistent with a Ca2+ current. Any Ca2+ current contamination of the peak current, however, would lead us to underestimate the TTX sensitivity of the Na+ current.
In summary, most rat utricular hair cells between P0 and P4 expressed one of two Na+ currents with different voltage dependence and sensitivity to block by TTX. INa,1 had TTX sensitivity and kinetics comparable to those of cardiac Na+ currents. The cardiac Na+ current also has a more negative voltage dependence than most brain Na+ currents (DISCUSSION). These similarities suggest that INa,1 might be carried by channels similar to those responsible for the cardiac Na+ current. INa,2 had TTX sensitivity, kinetics, and voltage dependence in the range of several neuronal channels. In the next sections, we show how expression of the two Na+ currents depended on hair cell location in the utricular macula, hair cell type, and postnatal age.
Expression of the two Na+ currents varied with location in the immature epithelium
The utricular macula has two regions, the striola and extrastriola, with distinct morphology and afferent physiology. We examined how the two Na+ currents were distributed relative to these regions (Fig. 6A). Because of developmental changes (see next section), we included only cells from ages P0 to P4 (n = 110) in the histogram. In this period, all but four cells expressed Na+ current. Most (90%, 45/50) striolar hair cells expressed INa,1; the remaining five cells expressed INa,2. In contrast, just over half (33) of 60 extrastriolar cells expressed INa,2, with 23 cells expressing INa,1 and four cells lacking any Na+ current. Breaking the extrastriolar data of Fig. 6A into values from the lateral and medial zones (LES and MES, Fig. 6B) revealed that nearly all extrastriolar cells with INa,1 were from the MES. The lateral edge of the extrastriola, furthest from the point of entry of the utricular nerve branch, was almost devoid of INa,1. The difference in our LES and MES data might alternatively reflect a peripheral versus central difference because MES cells were taken mostly from the central part of this zone while LES cells are perforce near the peripheral edge of the epithelium (see Fig. 6A, inset).
Based on whole cell capacitance (Cm) values, which are proportional to cell surface area, cells at all ages expressing INa,1 had 16% more surface area than cells expressing INa,2: 7.0 ± 0.13 pF (n = 122) versus 6.0 ± 0.23 pF (n = 45; P < 0.001). Similar values were obtained in the smaller P0–P4 data set and when striola and extrastriola were considered separately. The size difference might reflect a difference in cell type. Identified type I cells (7.2 ± 0.26 pF, n = 26) were larger than type II cells (6.1 ± 0.31 pF, n = 30; P < 0.01) between P9 (the youngest age at which type II cells can be confidently identified) and P22. Some of the capacitance difference might reflect different hair bundle surface areas. Alternatively, cells with INa,2 might be smaller if they are at an earlier developmental stage. This seems unlikely, however, as we saw no change in Cm with age (P0–P22), either overall or within groups (striola, extrastriola, INa,1 and INa,2).
Type I cells expressed INa,1
Calyces form and expand in the first postnatal week in the rat utricle. We were able to classify a hair cell as type I (mature or immature) if a partial or full calyx was visible around it (see METHODS). When no calyx was visible, a hair cell could be an immature hair cell of either type. Most calyces and therefore most identified hair cells were in the striola. All 50 hair cells with calyces (39 striolar, 11 extrastriolar) expressed INa,1.
We classified 30 cells as type II based on their lack of calyces at P9 or older; even this group may include some type I hair cells that have yet to receive a calyx ending. Nearly half (14) of these "type II" cells, approximately evenly spread across zones (striola, LES, MES), lacked any Na+ current. Of the remaining 16 type II cells, 15 expressed INa,1 (again, evenly distributed across zones); only one cell expressed INa,2. Note that we classify cells as type II only after the first postnatal week when INa,2 had largely disappeared from the macula (see next section). Many type II cells at earlier stages, here considered unclassified, must express INa,2: unclassified hair cells with INa,2 (all from P0 to P8) were unevenly distributed across zones: 6/35 (17%) of unclassified striolar cells and 38/66 (58%) of unclassified extrastriolar cells had INa,2.
In the mature rat utricular macula, 50% of hair cells in both zones are type I (Desai et al. 2005b). Similar percentages are likely to hold in the immature epithelium given evidence that vestibular hair cells in mice are born (have undergone their final division) by P3 (Ruben 1967). If so, then our data suggest that from P0 to P22 all type I hair cells and a significant fraction of immature type II cells in the striola express INa,1 (see DISCUSSION). It is possible that, at least during the first postnatal week, INa,2 is expressed only by immature type II cells. The next section shows dramatic changes after the first postnatal week.
Expression of the two Na+ currents varied with postnatal age
Cell size, as measured by Cm, did not change significantly from P0 to P22 (linear regression analysis yielded a line with slope 20 fF/day, n = 167, r2 = 0.005). Peak-current densities (peak current/Cm) for INa,1 and INa,2 varied with age and region. To calculate current densities, we took peak currents from the Boltzmann fits of inactivation data because we recorded inactivation protocols for all cells. Recall that these values were obtained for a test step to a sub-maximal depolarization (–17 mV) and thus underestimate peak current density by 20%. Figure 7 shows, for each current and each zone, how current density varied with age (A and B) and how incidence (% cells expressing a current) varied with age (C and D). Both currents were largest and detected most frequently during the first postnatal week.
|
Mean INa,2 density in the extrastriola plummeted after P3 (Fig. 7B). From P0 to P3, one-third of extrastriolar cells with INa,2 had densities 
300 pA/pF, whereas the rest had densities <180 pA/pF. After P3, the larger current densities disappeared, reducing both the mean and the variance. In the striola, only six hair cells had INa,2; their current densities seemed in line with the extrastriolar data (Fig. 7B).
With respect to incidence, INa,1 always dominated in the striola (Fig. 7C). INa,2 dominated in the extrastriola in the first week (59% of cells; 19/24 LES cells, 21/46 MES cells; Fig. 7D). From P13 onward, 63% of all cells (13/18 striolar, 3/8 LES, 8/12 MES) expressed INa,1. The remaining 37% with no detectable Na+ current matches the percentage of all cells in the first postnatal week that either expressed INa,2 or had no current (48/129). This raises the possibility that they are the same cell populations; i.e., that cells dominated by INa,2 early on did not convert to INa,1 later.
For INa,1, we had enough data to separate the current density data (Fig. 7A) according to cell type (Fig. 8, A and B). This shows that the zonal difference in Fig. 7A (larger currents in the striola than the extrastriola at all ages) holds for type I hair cells (Fig. 8A) and for unclassified and type II hair cells (Fig. 8B), considered separately. That is, there are no significant differences between the current densities of INa,1 in type I cells and in other cells (unclassified and type II cells pooled) as shown by a two-way ANOVA (P = 0.76) and post hoc Tukey means comparison (P > 0.05). Cells with low INa,1 density (<60 pA/pF) were seen at all ages, but densities >200 pA/pF were only recorded in the first week. The wide range of current densities in the first postnatal week may reflect hair cell differentiation occurring at different rates even within a single zone, as observed in the perinatal mouse utricular macula (Denman-Johnson and Forge 1999; Géléoc et al. 2004). Figure 8C shows that V1/2,inact values for both currents hyperpolarized by 1 mV/day (Fig. 8C).
In summary, almost all striolar cells expressed INa,1, no matter what age, but the mean current density dropped from the first to third weeks. Many extrastriolar cells expressed INa,2 in the first week. The overall numbers are consistent with these being immature type II cells. In addition, the numbers suggest that the population lacking all Na+ current in the third week co
). The time course of recovery of peak currents was fit with a single- or double-exponential function. Shown are 
, r2 = 0.44) and 0.8 mV/day for INa,2 (
, r2 = 0.25). —, 95% confidence intervals.