Molecular identity, ontogeny, and cAMP modulation of the hyperpolarization-activated current in...

Molecular identity, ontogeny, and cAMP modulation of the hyperpolarization-activated current in vestibular ganglion neurons

Angélica Almanza,1,2 Enoch Luis,1 Francisco Mercado,1,2 Rosario Vega,1 and Enrique Soto1

1Instituto de Fisiología, Universidad Autónoma de Puebla, Puebla, Mexico; and 2Instituto Nacional de Psiquiatría “Ramónde la Fuente Muñiz,” Mexico City, Mexico

Submitted 24 April 2012; accepted in final form 20 July 2012

Almanza A, Luis E, Mercado F, Vega R, Soto E. Molecularidentity, ontogeny, and cAMP modulation of the hyperpolarization-acti-vated current in vestibular ganglion neurons. J Neurophysiol 108: 000–000,2012. First published July 25, 2012; doi:10.1152/jn.00337.2012.—Proper-ties, developmental regulation, and cAMP modulation of the hyper-polarization-activated current (Ih) were investigated by the whole cellpatch-clamp technique in vestibular ganglion neurons of the rat at twopostnatal stages (P7–10 and P25–28). In addition, by RT-PCR andimmunohistochemistry the identity and distribution of hyperpolariza-tion-activated and cyclic nucleotide-gated channel (HCN) isoformsthat generate Ih were investigated. Ih current density was larger inP25–28 than P7–10 rats, increasing 410% for small cells (�30 pF)and 200% for larger cells (�30 pF). The half-maximum activationvoltage (V1/2) of Ih was �102 mV in P7–10 rats and in P25–28 ratsshifted 7 mV toward positive voltages. At both ages, intracellularcAMP increased Ih current density, decreased its activation timeconstant (�), and resulted in a rightward shift of V1/2 by 9 mV.Perfusion of 8-BrcAMP increased Ih amplitude and speed up itsactivation kinetics. Ih was blocked by Cs�, zatebradine, and ZD7288.As expected, these drugs also reduced the voltage sag caused withhyperpolarizing pulses and prevented the postpulse action potentialgeneration without changes in the resting potential. RT-PCR analysisshowed that HCN1 and HCN2 subunits were predominantly amplifiedin vestibular ganglia and end organs and HCN3 and HCN4 to a lesserextent. Immunohistochemistry showed that the four HCN subunitswere differentially expressed (HCN1 � HCN2 � HCN3 � HCN4) inganglion slices and in cultured neurons at both P7–10 and P25–28stages. Developmental changes shifted V1/2 of Ih closer to the restingmembrane potential, increasing its functional role. Modulation of Ih

by cAMP-mediated signaling pathway constitutes a potentially rele-vant control mechanism for the modulation of afferent neuron dis-charge.

inner ear; Ih; hair cell; zatebradine; ZD7288; cAMP

THE HYPERPOLARIZATION-ACTIVATED current (Ih) is a nonspecificcationic inward current that activates during membrane hyper-polarization with slow activation kinetics and is modulated byintracellular cAMP (Accili et al. 2002; Pape 1996). In mam-malian cardiac Purkinje fibers, where the current was firstreferred to as the funny current (If), it participates in cardiacpacemaking, and because of its sensitivity to cAMP If has asignificant role in the autonomic modulation of the heart rate(Accili et al. 2002; DiFrancesco 1981). Since then it has beenfound that Ih is expressed in a variety of neuronal tissues,where, besides its contribution to the resting membrane poten-tial, it also contributes to repetitive discharge. For example, Ihis essential for rhythmic oscillation of the thalamic-relay neu-

rons and in neurons that are not spontaneously active such ashippocampal CA1 and layer V cortical-pyramidal neurons.This current also contributes to the regulation of membraneinput resistance and dendritic integration of synaptic potentials(Berger et al. 2001; Magee 1998, 1999; McCormick and Pape1990; Pape 1996).

Four full-length cDNAs encoding for Ih channels were se-quenced and cloned from a variety of tissues including cardiacand neuronal tissues (Ludwig et al. 1998; Santoro et al. 1998;Seifert et al. 1999). The gene family and the correspondingproducts were named HCN1–4. The functional channels have atetrameric structure and diverge with respect to other voltage-activated potassium channels in two distinct ways: 1) the currentthrough HCN channels is carried by Na� and K� (permeabilityratio 1:4), and 2) HCN channels are activated by membranehyperpolarization and deactivated with membrane depolarization(Hofmann et al. 2005). All the HCN isoforms have a cyclicnucleotide binding domain in the carboxy-terminal portion. Thebinding of cAMP to this domain shifts the voltage dependence ofthe channel to positive membrane potentials and speeds up itsactivation kinetics (Accili et al. 2002; DiFrancesco and Tortora1991; Wainger et al. 2001). The different isoforms show quanti-tative differences as a function of their activation kinetics, voltagedependence, and cAMP sensitivity. The activation kinetics of theHCN1 current is faster than that of the HCN4, with middleactivation kinetics for HCN2–3. The half-activation voltage (V1/2)is more negative for HCN2 and HCN3 compared with HCN4 andHCN1. Unlike HCN2 and HCN4, whose open probability isstrongly influenced by cAMP, the HCN1 isoform is weaklymodulated by this nucleotide and HCN3 is inhibited rather thanactivated by cyclic nucleotides (Altomare et al. 2001, 2003;Robinson and Siegelbaum 2003).

In vestibular ganglion neurons from P5 mice, Ih had slowactivation kinetics, was blocked by extracellular Cs� andZD7288, and was carried by Na� and K� (Chabbert et al.2001b). However, because of its low V1/2 (�98 mV), theinhibition of Ih did not affect either the resting potential or theafterhyperpolarization (AHP) trajectory. Thus the participationof Ih in the discharge properties in the vestibular ganglionneurons is still undefined. Given their activation kinetics, it wassuggested that HCN2 and HCN4 could be the subunits under-lying Ih in vestibular ganglion neurons (Chabbert et al. 2001b).However, its molecular identity is still unidentified. In ourwork we studied 1) the possibility that modulation of Ih bycAMP shifts its activation range to more functional membranepotentials; 2) Ih expression in neurons obtained from rats intwo different postnatal development stages (P7–10 and P25–28); and 3) the molecular identity of Ih.

Address for reprint requests and other correspondence: E. Soto, Instituto deFisiología, BUAP, Apartado Postal 406, Puebla, Pue., CP 72000, Mexico(e-mail: [email protected]).

J Neurophysiol 108: 000–000, 2012.First published July 25, 2012; doi:10.1152/jn.00337.2012.

10022-3077/12 Copyright © 2012 the American Physiological Societywww.jn.org

AQ: 2

AQ: 3

AQ: 8 AQ: 8

MATERIALS AND METHODS

Animal care and procedures were in accordance with the AmericanPhysiological Society’s “Guiding Principles in the Care and Use ofVertebrate Animals in Research and Training” and the Reglamento dela Ley General de Salud en Materia de Investigación para la Salud ofthe Secretaría de Salud de México. Protocols involving animal re-search were reviewed and approved by the Comite Institucional deCuidado y uso de Animales de Laboratorio (CICUAL) del Consejo deInvestigación y Estudios de Posgrado de la Vicerrectoria de Investi-gación y Estudios de Posgrado de la Universidad Autónoma de Puebla(VIEP-BUAP). All efforts were made to minimize animal sufferingand to reduce the number of animals used, as outlined in the NationalInstitutes of Health Guide for the Care and Use of LaboratoryAnimals. Animals were supplied by the “Claude Bernard” animalhouse of the Universidad Autónoma de Puebla.

Isolation and culture of vestibular ganglion neurons. NeonatalLong-Evans rats from postnatal days (P)7–10 (before myelinationtakes place) of either sex were used for the experiments. A smallerexperimental sample of young rats from P25–28 was used because atthis (more mature) age the neurons of the vestibular ganglion aremyelinated, thus complicating the whole cell recording (Toesca1996). However, in some neurons this sheath had disappeared duringthe culture procedure, thus allowing the recording (Limón et al. 2005;Santos-Sacchi 1993). Animals were anesthetized with sevoflurane andkilled by decapitation. The head was cleaned rigorously with 70%ethanol. The inferior maxilla was removed and the cranium immersedin L-15 medium (GIBCO, Grand Island, NY). The upper part of theskull and the brain were removed, and under the stereoscopic micro-scope (Nikon, Tokyo, Japan) the otic capsule and the vestibularganglia were identified. The vestibular ganglia were removed andtreated with a combination of 1.25 mg/ml porcine trypsin and 1.25mg/ml collagenase IA dissolved in L-15 culture medium for 30 min at37°C. The ganglia were then rinsed with fresh culture medium,triturated with a fire-polished Pasteur pipette, and centrifuged at 4,000rpm for 5 min. The supernatant was discarded, and this procedure wasrepeated three times. The isolated ganglia neurons were plated oncover slides pretreated with 100 �g/ml poly-D-lysine (Sigma-Aldrich,St. Louis, MO) in 35-mm petri dishes (Corning, Lowell, MA) with 2ml of modified L-15 medium (supplemented with 10% fetal bovineserum, 500 IU penicillin, 25 �l/ml Fungizone, 15.7 mM NaHCO3, and15.8 mM HEPES; pH adjusted to 7.7 with NaOH). An initial pH of7.7 was used to allow the medium to reach pH 7.4 after 30 min in aCO2 incubator. The cells were maintained in a 95% air-5% CO2

humidified incubator at 37°C for 18–24 h until recording, for whichthe culture dish was mounted on the stage of an inverted phase-contrast microscope (TMS, Nikon).

Electrophysiological recording. Membrane ionic currents and cell-voltage responses were studied with whole cell voltage-clamp andcurrent-clamp techniques. Experiments were done at room tempera-ture (23–25°C). For cell recordings, the culture medium was replacedby extracellular saline solution containing (in mM) 140 NaCl, 5.4KCl, 1.8 CaCl2, 1.2 MgCl2, and 10 HEPES, at pH 7.4. Threeadditional extracellular solutions were employed: a K�-rich solutionin which the extracellular K� concentration was increased to 50 mMand the Na� concentration decreased to 90.4 mM, a Na�-free solutionin which Na� was substituted by equimolar choline chloride, and aNa�-K�-free solution where Na� and K� were substituted byequimolar N-methyl-D-glucamine (NMDG�) and choline chloride.Ionic currents from vestibular ganglion neurons were recorded with anAxopatch 200B amplifier (Molecular Devices, Union City, CA).Command pulse generation and data sampling were controlled bypCLAMP 10 software (Molecular Devices) and a 16-bit data acqui-sition system (Digidata 1440A, Molecular Devices). Data were sam-pled at 10 kHz and low-pass filtered at 2 kHz. Passive properties of thecells [membrane capacitance (Cm), membrane resistance (Rm), accessresistance (Ra), and time constant (�)] were measured online with the

pCLAMP program at �70 mV. Eighty percent of the series resistancewas electronically compensated. The liquid junction potential wascalculated with pCLAMP 10 software (Molecular Devices). Measuredliquid junction potentials were �5 mV for different solutions, forwhich these were not corrected. All patch-clamp recordings weremade with a holding potential (VH) of �60 mV. Normal extracellularsolution was employed for current-clamp experiments, and the cellmembrane potential was held at about �70 mV. The digital data werestored for off-line analysis with Clampfit 10 (Molecular Devices).Patch pipettes were pulled from borosilicate glass capillaries(TW120–3; WPI, Sarasota, FL) with a Flaming-Brown electrodepuller (80/PC; Sutter Instruments, San Rafael, CA). The pipettesolution contained (in mM) 125 KCl, 10 NaCl, 0.134 CaCl2, 5HEPES, 10 EGTA, 2 ATPMg, and 1 GTPNa, at pH 7.2. Once in thebath solution, the filled electrodes had a resistance between 1.5 and2.5 M�. In the course of an experiment, seal and series resistancewere continuously recorded to guarantee stable recording conditions.

An experimental series using perforated patch recording instead ofa whole cell configuration was also performed. For these experimentsnormal extracellular solution was used, and the pipette solution wasadded with 300 �M amphotericin B to permeabilize the cells.

Drugs used. Drug application was made with a gravity-driven flowsystem (flow rate of �0.5 ml/s) consisting of three square perfusiontubes coupled to a step motor (SF-77B; Warner Instruments, Hamden,CT) for rapid solution change. The drugs used were all fromSigma-Aldrich.

To define the inward current pharmacological profile, the effects of2 mM Ba2�, 2 mM Cs�, 30 �M zatebradine, 80 �M ZD7288, andK�-rich solution were evaluated. To study the modulation of Ih bycyclic nucleotides, the cell-permeant cAMP analog 8-BrcAMP (1mM) and cAMP (100 and 500 �M) were used. The cAMP was addedto the pipette solution, and Ih was recorded 2 min after the establish-ment of the whole cell configuration. The current amplitude wasnormalized with respect to Cm and compared to the control, withoutcAMP. The effect of ZD7288 was studied in the presence of 200 �McAMP added to the pipette solution.The effects of drugs were evalu-ated with test pulses, applied every 5 s, to �110 mV with duration of1.5 s from a VH of �60 mV. The amplitude of Ih was measured at least2 min after drug perfusion.

Data analysis. Recordings were analyzed off-line with Clampfit 10(Molecular Devices), SigmaPlot 8.0 (Systat Software, Richmond,CA), and Origin 6.0 (Microcal Software, Northampton, MA) soft-ware. The recordings of Ih were always done with 2 mM BaCl2 addedto the extracellular solution to block the IKIR. The current-voltage(I-V) relationship of Ih was generated by measuring current responseto 1.5-s voltage steps, applied every 4 s, in 10-mV increments from�120 to �60 mV from a VH of �60 mV. Recordings whereextracellular BaCl2 was tested were done with extracellular normalsolution, and inward current was generated with voltage steps between�130 and �70 mV. The amplitude of Ih was obtained as thedifference between the current at the end of the pulse (Iend) and theinstantaneous current measured just after the capacitive transient(Iinst). In some cases (specified in the text), to study Ih activationkinetics the voltage pulses were hyperpolarized to �140 mV and theIh activation � was obtained by fitting with a double-exponentialfunction after a delay (at 50 ms) (Altomare et al. 2001; Chen et al.2001). When the use of a double exponential yielded a negative �, the� was obtained by fitting current traces with a single exponentialfunction. Visual inspection and the standard deviation between the fitand the data were used to estimate the quality of the fit. In all cases thefitting correlation coefficient was �0.985.

Voltage-dependent activation was studied from the tail currentsgenerated when voltages returned to the VH (�60 mV). The tailcurrent data were normalized with the equation Rnorm � (R �Rmin)/(Rmax � Rmin), where Rmin and Rmax are the minimum andmaximum tail current measured. Data were normalized, plotted, andfitted by a Boltzmann distribution with the function Itail � 1/1 �

2 HYPERPOLARIZATION-ACTIVATED CURRENT IN VESTIBULAR GANGLION NEURONS

J Neurophysiol • doi:10.1152/jn.00337.2012 • www.jn.org

exp[(Vtest � V1/2)/s], where Vtest is the test potential, V1/2 is thehalf-maximum activation voltage, and s is the exponential slope.

Statistical analysis of the difference between various experimentalconditions was done with Student’s t-test. In some cases when thenormality test failed, a Mann-Whitney test was used, and this isindicated in the text. Effects were considered as significant with a P �0.05. Results are given as means � SE.

RT-PCR analysis. RNA was isolated from vestibular organs (utri-cle, lateral, and anterior semicircular canals were pooled), vestibularganglion, brain cortex, and heart ventricle by a standard method withTRIzol reagent. RT-PCR experiments were done three times (n � 3);in each one, tissue from inner ear of eight rats was pooled. The tissueswere homogenized with a pellet-pestle in the TRIzol reagent, and thenthe manufacturer’s instructions were followed to obtain the totalRNA. Once the RNA was isolated it was treated with DNase I(Invitrogen, Carlsbad, CA) to eliminate the genomic DNA. RNAquantity (�10 �g/�l) and purity (A260/A280 � 1.6) were verified inthe spectrophotometer and RNA integrity by electrophoresis in a 1%agarose gel. cDNA was synthesized with the SuperScript First-StrandSynthesis Supermix kit (Invitrogen) according to the manufacturer’sinstructions. PCR was done in a 25-�l reaction tube with cDNA astemplate. Primers to amplify HCN1, HCN2, HCN3, and HCN4 geneswere purchased from SABioscience-Qiagen (Frederick, MD; catalognos. PPM06969A-200, PPM06968A-200, PPM25737A-200, andPPM31706A-200), and 18S ribosomal primers were used as consti-tutive-expression controls. For PCR of negative controls the RTprocedure was omitted. Amplification was conducted in a Mini-Opticon thermal cycler (Bio-Rad, Hercules, CA). PCR products weremarked with ethidium bromide in a 1% agarose gel, with photographicrecording done in an UV transilluminator.

Immunostaining. To identify the expression of HCN channels in thevestibule, monoclonal antibodies (IgG) against HCN1, HCN2, HCN3,and HCN4 were used (all from UC Davis/NIH Neuromab Facility,clones N71/37, N70/28, N141/28, N114/10, respectively). Tissuesections and vestibular ganglion neuron cultures were obtained fromLong-Evans rats (P7–10 and P25–28).

For tissue slices, the rats were killed with an overdose of sevoflu-rane, the vestibule and vestibular ganglia were dissected, and thesamples were fixed for 2 h in paraformaldehyde (2%) and subse-quently placed in sucrose (20%) overnight [both diluted in phosphate-buffered saline (PBS) 0.1 M, pH 7.4]. The vestibular epithelium andvestibular ganglia were embedded in Tissue-Tek and frozen at�20°C, and sliced sections (20 �m) were obtained in a cryostat(Cryocut 2800 Reichert-Jung), with the tissue slices then placed ongelatin-coated microscope slides. For the cultured cells, glass cover-slips with the isolated vestibular ganglion neurons were fixed for30 min with paraformaldehyde at 2% and washed with PBS.

Preparations were incubated in a blocking solution (0.2% fetal goatserum, 0.03% Triton X-100, and 0.2% bovine serum albumin in PBS)for 4 h at room temperature (�20°C). Preparations were then incu-bated with HCN1, HCN2, HCN3, or HCN4 antibodies (diluted 1:40 inblocking solution) overnight at 4°C. Subsequently, slides werewashed in PBS three times and incubated with the F=ab2 fragmentanti-mouse IgG conjugated with Alexa 488 (diluted 1:500; Invitrogen,Grand Island, NY) for 2 h at room temperature. In all cases controlswere run in which the first antibody was omitted. Medium with addedpropidium iodide (VectaShield, Vector Labs) was used for mountingthe coverslip. Observations were done with a laser-scanning confocalmicroscope (LSM 5 Pascal AxiosKop2MOT, Carl Zeiss, Jena, Ger-many). For image processing and labeling, a Zeiss LSM ImageExaminer and Photoshop CS5 were used. All experiments and obser-vations were repeated at least three times.

RESULTS

Hyperpolarization-activated inward current identification.Under control conditions with normal extracellular solution,hyperpolarizing voltage steps generated in all vestibular gan-glion neurons (n � 104) an inward current with two compo-nents: a quasi-instantaneous linear component with fast acti-vation (Iinst) and a slow-activating component, which couldcorrespond to Ih (Fig. 1A). To corroborate the identity of theinstantaneous inward-rectifier component, 2 mM extracellularBaCl2 was used (Fig. 1B). Iinst amplitude (measured just afterthe capacitive current) was sensitive to BaCl2 (n � 10),decreasing from �520 � 70 pA under control conditions to�230 � 30 pA at �130 mV (P � 0.001), whereas the slowactivated component (Iend � Iinst) did not change with the useof BaCl2 (control �400 � 40 pA and �400 � 30 pA withBaCl2; n � 10; P � 0.76) (Fig. 1C). The use of 2 mM CsCl(n � 11), 30 �M zatebradine (n � 9), and 80 �M ZD7288(n � 13, data not shown) decreased the slow component 93 �2%, 94 � 5%, and 79 � 2% respectively, without significanteffect on Iinst (Fig. 2, A and B).

Perfusion of the Na�-free solution decreased the slow com-ponent amplitude 53 � 7% (n � 7), whereas perfusion of theNa�-K�-free solution decreased the amplitude 83 � 8% (n �4) (data not shown). These results indicate that the slow-activating component flows through nonspecific cationic chan-nels. Additionally, and because K� ions have been shown tohave an activation-like effect on some of the inward currents(DiFrancesco et al. 1986), an extracellular K�-rich solution

Fig. 1. Characteristics of the inward rectifier current and the effect of external BaCl2. A: typical inward currents show an initial component measured just afterthe capacitive transient (Iinst) and a slow-activating component measured as the difference between the current at the pulse end (Iend) and the Iinst. B: perfusionwith BaCl2 decreased the inward-rectifier current, with no effect on the slow-activating component. C: current-voltage relationship for the slow-activatingcomponent measured before (Control) and after BaCl2 superfusion.

3HYPERPOLARIZATION-ACTIVATED CURRENT IN VESTIBULAR GANGLION NEURONS


F2

was tested on the current caused by hyperpolarizing voltagesteps (n � 5). At �120 mV, the use of high extracellular K�

(50 mM) solution increased the amplitude of the slow-activat-ing component 721 � 140% (Fig. 2, C and D), while onlyincreasing Iinst by 60 � 29% (P � 0.1). More detailed studiesof Iinst identity were not made.

In summary, the Iinst sensitivity to extracellular BaCl2 sug-gests that this component flows at least in part through aninward-rectifier K� channel. In contrast, the lack of sensitivityto BaCl2 and the sensitivity to a relatively low concentration ofextracellular Cs�, zatebradine, ZD7288, external K� concen-tration, and a nonspecific cationic permeability are character-istics of Ih, which indicates that the slow-activating currentflows through HCN channels.

Cyclic nucleotide modulation of Ih. In all the followingrecordings, 2 mM BaCl2 was used to isolate Ih from the inwardinstantaneous rectifying current. Typically, Ih activated withhyperpolarizing steps to potentials below �60 mV (n � 44).To evaluate whether or not Ih was modulated by cAMP,initially 8-BrcAMP was used. Extracellular perfusion of 1 mM8-BrcAMP significantly increased Ih amplitude at voltagesbetween �140 mV and �110 mV (25 � 10% at �140 mV;n � 7, P � 0.05) (Fig. 3A). At �140 mV current density was

13 � 2.5 pA/pF under control conditions and 16 � 2.5 pA/pFwith 8-BrcAMP (P � 0.002). 8-BrcAMP also increased the Ihactivation rate. Current activation at �140 mV was fitted witha single exponential with � � 302 � 38 ms under controlconditions and with � � 191 � 17 ms with 8-BrcAMP (n � 7,P � 0.016; Fig. 3, B and C). Similarly, at �130 mV theactivation rate changed from � � 292 � 21 ms under controlconditions to � � 251 � 18 ms with 8-BrcAMP (n � 7, P �0.03). Activation V1/2 was not significantly modified by perfu-sion of 8-BrcAMP (V1/2 � �99 � 0.6 mV and s � 9 � 0.6 mVin control condition, V1/2 � �100 � 0.7 mV and s � 9 � 0.7mV; n � 7, P � 0.094). An increase in the amplitude of Ih withextracellular 8-BrcAMP without significant change in its V1/2 hasalso been reported in primary auditory neurons (Chen 1997).

In experiments in which 100 �M cAMP was added to theinternal solution of the recording pipette, the current densitymeasured at �140 mV increased from 12 � 2 pA/pF (n � 17) inthe control group to 18 � 2 pA/pF (n � 17, P � 0.02; Fig. 4A).It should be noted that current densities of control groups forexperiments with perfusion of 8-BrcAMP and those withintracellular cAMP were not statistically different (P � 0.69).Similarly, the increase in Ih density observed with 8-BrcAMPwas analogous to the increase in its density with intracellular

Fig. 2. Pharmacology of the slow-activating inward current. A: current under control conditions (1) and with CsCl (2) and the Cs�-sensitive current (3) obtainedby subtraction 1 � 2. B: control current (1) and with zatebradine (2) and the zatebradine-sensitive current (3). C: control current (1); use of a high extracellularK� (50 mM) solution increased the amplitude of the current (2). D: current-voltage relationship under the control condition and with high K�. Currents werenormalized as a function of control value (I/Imax).



F4

enrique

Texto escrito a máquina

enrique

Texto escrito a máquina

|

cAMP (P � 0.44). Intracellular cAMP shifted the activationV1/2 from �102 � 1.4 mV to �93 � 2 mV (P � 0.001; Fig.4B), with no significant change of the slope in the I-V curves(s � 8 � 0.8 mV and 9 � 0.9 mV). Under control conditions,in 11 of 17 cells the Ih activation kinetics was fitted with asimple exponential whereas in 6 of 17 cells the activationkinetic was better fitted with a double exponential (Table 1).Figure 4C shows family traces of Ih activation fitted with asingle- or double-exponential function in control conditionsand with 100 �M cAMP (no difference in the activation ratewas found between control groups for 8-BrcAMP and intra-cellular cAMP; P � 0.26). The activation kinetics of Ih wasvoltage dependent; the activation � for two voltages in differentconditions are shown in Table 1. The activation rate measuredat �140 mV and fitted with a simple exponential was signif-icantly different between the controls and cells treated withcAMP (P � 0.03), whereas between the groups in which theactivation kinetics were fitted with a double exponential onlythe �slow was significantly different (P � 0.017 for �slow and

P � 0.052 for �fast; Mann-Whitney test used in both cases)(Table 1). Analysis of the voltage dependence separately forcells fitted with a single or a double exponential shows thatactivation kinetics in the control condition and with intracel-lular cAMP did not differ. At higher cAMP concentrations, i.e.,500 �M, V1/2 shifted to �87 � 1.2 mV (n � 13; P � 0.001;Fig. 4B) without significant change in the slope (s � 8 � 1.2mV; P � 0.09). A good linear correlation between Cm ofvestibular ganglion neurons in culture and their neuronal size iswell established, indicating that Cm can be used as an indirectmeasure of soma size (Limón et al. 2005). Cm was not signif-icantly different in the group of neurons in which fitting of Ihactivation was single compared with those in which it wasdouble exponential (Fig. 4D; P � 0.4 and 0.7 for controlconditions and with intracellular cAMP, respectively), indicat-ing that Ih activation characteristics are not segregated betweenneurons with different soma sizes.

Current-clamp recordings. To analyze the functional role ofIh, current-clamp experiments were done. Current was injected

Fig. 3. Effect of 8-BrcAMP on Ih. A: current-voltage relationship under control conditions (n � 7) and after the application of 1 mM 8-BrcAMP (n � 7). Currentswere normalized as a function of the control value. B: the control current obtained at �140 mV is compared with those taken after the application of 1 mM8-BrcAMP. Solid lines are a single-exponential fit to Ih activation phase. C: activation time constant (�) of Ih at �140 and �130 mV in the control conditioncompared with that obtained after 8-BrcAMP use (n � 7). Solid lines connect the � values in each condition for a cell.

Fig. 4. Modulation of Ih by cAMP. A: current-voltage relationship shows that Ih amplitudein cells recorded with 100 �M intracellularcAMP was higher than in control cells. B:activation curve for Ih in control conditionand with 100 and 500 �M intracellularcAMP. Data were fitted with a Boltzmannfunction (solid lines). Half-activation voltage(V1/2) shifted from a control value of �102mV to �93 mV and �87 mV. Itail, tail current.C: Ih current family recorded between �140mV and �60 mV, in control condition andwith 100 �M intracellular-cAMP. Simple anddouble exponential fits are shown by graylines. D: correlation between membrane ca-pacitance (Cm) and single- or double-exponen-tial Ih kinetics. Plot shows fits obtained for Ih

activation kinetics in control condition andwith 100 �M intracellular cAMP at both de-velopment stages studied (P7–10 and P25–28).



to hold the cells around �70 mV, a potential near the restingmembrane potential (�62 � 4 mV, n � 45), and voltageresponses were produced with current pulses between 1.7 and3 nA and a duration of 1 ms. Action potential (AP) character-istics and the zatebradine effect were analyzed at the minimumcurrent that caused an AP. The waveform of the AP andquantitative characteristics were determined in accordancewith Bean (2007). Under control conditions (n � 10) thethreshold of the AP was �30 � 1.5 mV, the overshoot was44 � 3 mV, the spike height was 118 � 4 mV, the AP widthwas 1 � 0.1 ms (measured at half-maximum spike amplitude),and the AHP was 8.7 � 1.8 mV. Perfusion of 30 �M zatebra-dine (n � 10) did not significantly modify the cell membranepotential or the AP threshold (P � 0.06 and P � 0.7).However, it decreased the overshoot to 30 � 4 mV (P � 0.004)and the spike height to 103 � 5 mV (P � 0.005), increased theAP width to 1.2 � 0.1 ms (P � 0.006), and decreased the AHPto 4.4 � 1.4 mV (P � 0.04) (Fig. 5A). The reduction of APamplitude produced by zatebradine might be due to nonspecificaction of the drug, which has been reported to also affect the Na�

current (Soto et al. 2008). In another experimental series cellresponses were caused with 200-ms pulses between �0.5 and 0.9nA (n � 12). In this condition, typically after the onset ofhyperpolarization current, a depolarizing sag, thought to be due toIh activation, slowly develops (the sag was measured as the ratiobetween the voltage change at the end and at the beginning of thepulse). At the end of the hyperpolarizing current step, in 83% ofthe cells a postpulse rebound AP was produced upon repolariza-tion. The use of 30 �M zatebradine significantly decreased the sagfrom 0.7 � 0.04 to 0.9 � 0.04 (P � 0.001). In 90% of the cellsthe postpulse AP was prevented (Fig. 5B), and in 10% of the cellszatebradine increased the postpulse AP latency (from 20.8 ms to33.6 ms). Perfusion of 2 mM Cs� (n � 5) and ZD7288 (n � 12)had a similar effect on the voltage response produced withhyperpolarizing current injection and on the postpulse AP. Thus,although the action of zatebradine may be in part due to its effecton currents participating in AP generation other than Ih, the actionof Cs� and ZD7288 did not affect the height of the AP and stillprevented rebound discharge, indicating that activation of Ih playsa significant role in the increase in excitability and AP generationthat follows a hyperpolarization.

Perforated patch recordings. An experimental series usingperforated patch clamp in P7–10 rats was developed with theaim of defining the functional expression of Ih when the

intracellular milieu is not dialyzed. Hyperpolarizing voltagesteps generated an inward current similar to that found inwhole cell recordings, showing two components: a quasi-instantaneous linear component with fast activation (Iinst) and aslow-activating component corresponding to Ih (Fig. 6A). At�140 mV, the mean activation rate (128 � 21 ms; n � 4) wasfaster than that found under control conditions in whole cellrecordings (P � 0.01), and also faster than that found when100 �M cAMP was added to the internal solution (P � 0.05).Ih conductance measured at the tail showed V1/2 � �93 � 4

Table 1. Ih activation kinetics at two voltages in the different conditions studied

Group

P7-10 P25-28

�single �slow �fast �single �slow �fast

Controln 11 6 6 13 9 9�140 mV 356 � 28 ms* 2.3 � 0.8 s* 168 � 23 ms 224 � 17 ms* 663 � 6 ms 98 � 13 ms�110 mV 1.5 � 0.4s 4.8 � 1.2s 897 � 292 ms 840 � 171 ms 2.3 � 0.9 s 220 � 17 msP �0.001 0.1 0.029 0.01 0.024 0.003

cAMPn 12 5 5 7 2 2�140 mV 263 � 29 ms* 471 � 108 ms* 86 � 16 ms 134 � 11 ms* 757 ms, 451 ms 93 ms, 104 ms�110 mV 911 � 178 ms 1.7 � 0.7 s 257 � 30 ms 374 � 48 ms 1.4 s, 2 s 332 ms, 330 msP 0.001 0.016 �0.001 �0.001 — —

Values (expressed in ms and in s) are mean � SE activation time constants (�); n for each case is indicated. P value corresponds to statistical comparisonbetween activation kinetics at 2 voltages in the same group. Because n � 2 for �slow and �fast at P25–28 with cAMP, no P value is reported. *Statistical differencebetween control and cAMP groups.

Fig. 5. Effect of zatebradine on electric activity of vestibular ganglion neurons.A: 30 �M zatebradine significantly modified the action potential waveform.B: voltage response produced with a 200-ms 400-pA hyperpolarizing pulse inthe control condition and with 30 �M zatebradine. The voltage changefollowing the onset of hyperpolarizing pulse shows a depolarizing sag andrebound action potential; both are blocked by zatebradine. In both graphs,dotted lines indicate the zero voltage; current was injected to hold the cell atabout �70 mV.



mV and s � 11 � 1 mV (Fig. 6B), which is 9 mV rightwardshifted compared with control whole cell V1/2 but not signifi-cantly different from that found with 100 �M cAMP (P �0.05). Ih open probability obtained from perforated patchrecordings was 6.4% (at �70 mV), significantly larger thanthat measured from whole cell recordings with cAMP (1%).

Under current clamp, injection of hyperpolarizing pulsesproduced a voltage sag whose trajectory followed that of the Ihactivation under voltage clamp; upon repolarization of themembrane potential to �60 mV, a rebound AP was producedin all recordings (n � 4; Fig. 6C). Depolarizing pulses pro-duced from one to three (at most) APs in the cells studied.

Recordings in neurons from young rats (P25–28). Becauseof the negative activation voltage reported for Ih in vestibularganglion neurons from neonatal mice (Chabbert et al. 2001b),the possibility that in more mature stages the activation rangeof Ih could shift to more positive voltages has been considered.

During postnatal development the amplitude and current den-sity of Ih increased significantly in all neurons studied (Fig. 7,A and B). However, the increment of Ih density was greater insmaller neurons, which probably means that there is a selectiveexpression among the different neuron sizes present in thevestibular ganglion (Kevetter and Leonard 2002). On the basisof previous observations (Limón et al. 2005; Mercado et al.2006), neurons were divided into small (Cm � 30 pF)- andlarge (Cm � 30 pF)-soma diameter neurons. In young rats(P25–30) Ih density was 37 � 6 pA/pF in neurons with Cm �30 pF (mean capacitance of 20 � 1 pF; n � 11) and 14 � 1pA/pF in neurons with Cm � 30 pF (mean capacitance of 46 �3 pF; n � 11) (Fig. 7B). Ih density in small neurons fromP25–30 was 410% higher than in small neurons from P7–10(9 � 1 pA/pF; mean capacitance of 19 � 1 pF, n � 24; P �0.001) and 200% higher than in large neurons from P7–10 (7 �1 pA/pF; mean capacitance of 36 � 2 pF, n � 10; P � 0.001).

The V1/2 of Ih in P25–28 rats was �95 � 1 mV (n � 22), avalue shifted to positive voltages compared with the V1/2 inneonatal rats (�101 � 0.8 mV; n � 24; P � 0.001; Fig. 7C).With 200 �M cAMP (in the recording pipette) the V1/2 ofP25–28 rats was �86 � 1 mV significantly shifted to morepositive potentials (n � 9; P � 0.001; Fig. 7C). No significantchanges were observed in the s factor (P � 0.6). The activationkinetics of Ih in P25–28 rats was significantly faster than inneonatal P7–10 rats. In 59% of cells from P25–28 rats theactivation kinetics was well fitted with a simple exponential,and in the remaining 41% of cells it was better fitted with adouble-exponential function (Table 1). With 200 �M intracel-lular cAMP in seven cells the Ih activation kinetics was fittedwith a simple exponential function and in two cells was betterfitted with a double exponential (Table 1). Both simple anddouble exponential fits were faster in P25–28 rats than inP7–10 rats (P � 0.05 for each comparison).

Similar to neurons obtained from P7–10 rats, Cm of P25–28neurons was independent of whether Ih activation was fitted witha single or double exponential (P � 0.2 and 0.5 for controlconditions and with intracellular cAMP respectively) (Fig. 4D).

RT-PCR identification of HCN in vestibular ganglion andepithelia. To determine the molecular identity of the HCNsunderlying the Ih, RT-PCR of vestibular ganglia and vestibularepithelia (anterior and lateral semicircular canals and utricle) wasdone. Specific primers against the four HCN genes were used, andcDNAs from the brain and the heart ventricle were used aspositive controls. HCN1, HCN2, HCN4, and, to a lesser extent,HCN3 were amplified from the vestibular ganglia (Fig. 8). In thesensory neuroepithelia (including hair cells, supporting cells, andafferent and efferent dendrites) the HCN1 and HCN2 genes wereamplified, whereas the expressions of HCN3 and HCN4 were notsignificant. Similar results in both tissues were found in rats fromP25–30 (data not shown); although the results are not quantitative,transcripts for HCN3 were more evident in vestibular ganglia. Thepositive control from the brain tissue amplified the four HCNgenes as expected (Thollon et al. 2007), and the heart ventricleonly expressed HCN2 and HCN4, as previously reported (Moos-mang et al. 2001; Shi et al. 1999).

HCN immunoreactivity in vestibular ganglia and epithelia.To determine the tissue expression of the HCNs, the immuno-reactivity for each of the HCN subunits was studied in P7–10and P25–28 rats, both in vestibular tissue (semicircular canalcrista and ganglia) and in cultured vestibular ganglion neurons

Fig. 6. Perforated patch recording of Ih in P7–10 rats. A: inward current showsa large slow-activating component (pulses from �140 to �60 mV in every 20mV). B: current-voltage relationship for the slow-activating component.C: voltage response produced with 2-s hyperpolarizing pulses from �200 to�50 pA (every 50 pA) show a depolarizing sag and rebound action potential.Depolarizing pulse to 250 pA produced 2 action potentials. Dotted lineindicates zero voltage.



F8

(n � 3 for each experimental condition). Immunostaining forthe four HCN subunits was found in vestibular ganglia and incultured vestibular ganglion neurons at both ages studied. Theexpression of the four subunits was not the same: immunore-activity was clearly higher in all the cases studied for HCN1,and qualitatively the staining intensity was graded as HCN1 �HCN2 � HCN3 � HCN4 (Figs. 9 and 10). In all cases controlswere clearly negative (Fig. 9, J–L).

In the vestibular ganglion at P7–10, staining with HCN1,HCN3, and HCN4 antibodies was homogeneous, with noparticular selectivity in relation with the cell soma size (Fig. 9,A–D), although HCN3 and HCN4 intensity were very lowcompared with HCN1. The HCN2 antibody produced a lesshomogeneous staining, with large cells scattered through tissuesections. At P25–28 HCN1 staining was homogeneous and themost intense within vestibular ganglion (Fig. 10A). Selectiveneuronal body staining for HCN2 and HCN4 was clearlynoticeable in P25–28 rats (Fig. 10, B and D), and staining forHCN3 was just above noise level (Fig. 10C).

In the cultured neurons the four HCN subunits were found atboth ages studied, although differences in the staining intensityand staining patterns were noted (Fig. 9, E–H). HCN2 was themost difficult to identify, and stained cells were scarce. At P25–28HCN4 was clearly identifiable at the neuronal membrane (Fig. 10E).

In the neuroepithelia, HCN1 staining of hair cells wasintense and homogeneous in both P7–10 and P25–28 rats (Fig.9I). This coincides with previous reports showing that HCN1 isthe main subunit sustaining Ih in mouse hair cells (Horwitzet al. 2011). HCN2 and HCN3 produced a nonhomogeneousstaining of vestibular neuroepithelia that did not allow defini-tion of the cellular site of expression. At both ages studied,HCN4 produced a defined staining suggesting its expression atafferent calyceal endings. The HCN4 antibody also produced apunctuate staining within the neuroepithelia suggesting itsexpression in bouton endings of either efferent or afferentterminals (Fig. 10, F and G). This result raises the interestingpossibility that HCN4 could be preferentially expressed at thesynaptic regions of vestibular afferent neurons.

Fig. 7. Ih in the vestibular ganglion neurons from young rats. A: typical currents recorded in neurons from neonatal (P7–10) and young (P25–28) rats. B: plotof maximum Ih density vs. Cm of cells from neonatal and young rats. C: Ih current activation as a function of voltage from neonatal and young rats in the presenceof 200 �M cAMP. V1/2 shifted from �101 mV to �95 mV and �86 mV.

Fig. 8. Expression of HCN subunits in thevestibular ganglion (P7–10). Electrophoresisgel showing the cDNA amplification of thehousekeeping gene, the ribosomal 18S sub-unit, and the HCN1–HCN4 from brain (B),heart ventricle (HV), vestibular epithelia(VO), and vestibular ganglia (VG) and in thesame sequence the negative controls (no RT)where the RT procedure was omitted.



F10

DISCUSSION

We found the expression of mRNA for all HCN subunits inrat vestibular ganglion neurons harvested from neonatal (P7–10) and young (P25–28) animals. The neurons from neonatalrats predominantly express the mRNA for HCN1, HCN2, andHCN4 and, to a lesser extent, for HCN3, whereas in young ratsthe neuron transcripts for HCN3 were more evident, suggestingan increase in HCN3 expression with inner ear maturation.Immunolabeling with specific primary antibody showed that inthe ganglia and the cultured neurons the four HCN subunitswere expressed with a rank order HCN1 � HCN2 � HCN3 �HCN4. Notably, immunoreactivity for HCN4 was identifiablein the calyceal terminals specifically in domain 1 (Lysakowskiet al. 2011). In the vestibular neuroepithelia the HCN1 andHCN2 mRNAs were the most abundant in both neonatal andyoung rats. Our results are in concordance with data obtainedin the rainbow trout saccular hair cell layer, where the expres-sion of HCN1:HCN2:HCN4 is 7:2:1 (Cho et al. 2003), and inmouse utricle, where expression of HCN1–4 mRNA wasshown and hair cells were stained by HCN1, 2, and 4 antibod-ies (Horwitz et al. 2011). Vestibular hair cells obtained fromHcn1 knockout mice lacked Ih entirely, while in Hcn2 knock-out mice Ih was similar to wild type, suggesting that HCN1 isboth necessary and sufficient to form the channels that carry Ih

(Horwitz et al. 2011). Our results also match up with results inguinea pig spiral ganglion neurons, where all HCN subunits arelocalized (Bakondi et al. 2009), although no immunolabelingfor HCN3 was found in either the spiral ganglion or theafferent dendrites in the rat cochlea (Yi et al. 2010).

Recordings of the inward current showed two components:1) an instantaneous component (that we referred to as Iinst)decreased by extracellular BaCl2, suggesting that this componentflows at least in part through an inward-rectifier K� channel, and2) a slow-activating component whose pharmacology and non-specific cationic permeability are characteristic of Ih. The V1/2 forIh in whole cell recording (around �101 mV) is coincident withthe V1/2 reported in mouse vestibular primary neurons (Chabbertet al. 2001b), in guinea pig and murine spiral ganglion neurons(Chen 1997; Mo and Davis 1997), and in afferent dendrites in therat cochlea (Yi et al. 2010), where V1/2 values were in the rangeof �96 to �110 mV. It is worth noting that V1/2 under perforatedpatch was significantly lower (around �93 mV). In heterologousexpression systems the voltage dependence of the HCN channelsis limited by the isoform that activates at the most negativemembrane voltages (Ulens and Tytgat 2001). This can explainwhy, although the immunoreactivity signal was stronger forHCN1, the V1/2 determined for Ih is closer to that of the HCN2 andHCN4 subunits. This may also account for the fact that even

Fig. 9. Immunostaining with the 4 HCN subunit antibodies in P7–10 rats. Green, signal from specific antibody for each of the HCN subunits; red, propidiumiodide staining the cell nuclei. Left to right: HCN1, HCN2, HCN3, and HCN4, respectively. A–D: vestibular ganglion slices. E–H: isolated vestibular ganglionneurons maintained in culture. I: slice of the crista ampullaris showing the highly significant staining of hair cells by HCN1 antibody. J–L: control slices andcultured neurons where the secondary antibody was omitted. The highest immunostaining was found for the HCN1 subunit, and the immunostaining for othersubunits was of lower intensity, although all of them gave a specific staining. Scale bars, 20 �m for the ganglia and crista ampullaris slices (A–D, I–K) and 10�m for the cultured neurons (E–H, L).



COLOR

though voltage dependence was analyzed separately for cells withdouble-exponential or single-exponential activation kinetics, sig-nificant differences were not found.

In our experiments, two cellular groups could be differenti-ated. In the first set, Ih activation was fitted with a simpleexponential falling within the activation range for the HCN2,HCN3, and HCN4 channels (Moosmang et al. 2001; Santoroet al. 2000; Seifert et al. 1999). In the second set, Ih activationwas better fitted with a double exponential, for which the �slowis in the activation range for HCN2 and HCN4 and the �fast iscloser to the HCN1 and HCN2 activation range (Altomareet al. 2003; Santoro et al. 2000). The presence of multiple HCNisoforms in vestibular ganglion neurons and the variability of Ihactivation kinetics suggest that current results from both ho-momeric and heteromeric channels.

We found that raising intracellular cAMP increased the ampli-tude of the Ih, shifted its V1/2 toward positive membrane voltages,and speeded up its activation kinetics. However, Ih open proba-bility at �70 mV remained around 1% with or without cAMP.Removal of intracellular factors that might be lost during record-ing in the whole cell configuration could produce a negative shiftof Ih (Bal and Oertel 2000; Biel et al. 2009; Seifert et al. 1999;Southan et al. 2000). The use of perforated patch recordingsshowed that this is the case, and maintaining a more homogeneousand long-lasting intracellular milieu produced a significant shift ofV1/2 toward positive values and the open probability at �70 mVincreased to �6.5%. The V1/2 shift produced by cAMP in thevestibular ganglion neurons is similar to that found in auditoryneurons, where displacements between 7 and 16 mV have beenreported (Banks et al. 1993; Mo and Davis 1997; Rodrigues andOertel 2006; Yi et al. 2010). The increase of cAMP significantlydecreased the � in cells whose Ih activation was fitted with asimple exponential and the slow � in cells whose Ih activation was

fitted with a double exponential. The fast � was not significantlymodified by cAMP, consistent with the expression of HCN1.

The Ih density in vestibular ganglion neurons shows adevelopmental increase. The V1/2 was more depolarized inyoung than in neonatal rats, thus increasing the channel openprobability at �70 mV to 3%. Also, the current activationkinetics was developmentally regulated, being faster in youngrats. These changes determine an increase of Ih participation inthe setting of the membrane potential and in shaping neuronalresponses. The changes observed in the V1/2 and activationkinetics of Ih can be caused by a variation in the expressionratio of HCN subunits or by changes in the expression of theauxiliary subunit TRIP8b (Santoro et al. 2009).

The increase in Ih amplitude when the extracellular K�

concentration is increased may have a special relevance in thesynapse between the type I hair cells and the primary neurons,because K� may accumulate in the intersynaptic space (Gold-berg 1996; Lim et al. 2011; Soto et al. 2002b). The accumu-lation of K� in the synaptic cleft in calyx afferents will alsoaffect the postsynaptic element (Eatock and Lysakowski 2006;Goldberg 1996). Thus the amplitude of Ih in situ may be higherthan that observed under experimental conditions that use alow K� concentration similar to the perilymphatic fluid.

The voltage responses produced by hyperpolarizing pulsesshowed a sag back toward the resting potential, enough tocause postpulse APs in 83% of the cells. Zatebradine andZD7288 decreased the sag amplitude and blocked the postpulsespike discharge. Both drugs decreased the AHP and increasedAP duration and decreased its amplitude, with no significantchanges in the resting membrane potential. In our recordingconditions the AHP reached values close to �80 mV (similarto mouse vestibular ganglion neurons; Chabbert et al. 2001b),a voltage at which Ih opening probability increased to 14%

Fig. 10. Immunostaining with HCN subunit antibodies in ganglion from P25–28 rats and in neuroepithelia from P7–10 and P25–28 rats. A–D: slices fromvestibular ganglion stained with HCN1–HCN4, respectively. A: HCN1 immunoreactivity was the highest and most homogeneous. B: HCN2 staining showed aselectivity staining some large cell body neurons (arrows). C: HCN3 staining was close to noise level. D: although staining with HCN4 antibody was also oflow intensity, a distinct pattern of cell membrane staining can be seen (arrows). E: in ganglion cells, although HCN4 produced a faint staining in some cells itcan be clearly located at the cell membrane (arrow). F: slice of crista ampullaris showing the staining with HCN4 at P7–10, which marked some hemicalyx-likestructures beneath the hair cells (*) and some bouton-like structures in other cells (arrows). G: slice of crista ampullaris stained with HCN4 antibodies at P25–28shows calyx-like structures (*). Scale bars, 20 �m in A–D and 10 �m in E–G.



COLOR

with cAMP. After a hyperpolarization, the activation of Ih willlead to an increase in neuron excitability and AP discharge.

Electrophysiological properties of central and peripheral termi-nals of vestibular afferent neurons most probably differ from thecell body. Even though there are differences in the specific ionicchannels between afferent process and somas, data suggest thatthe use of cultured cells is adequate for study of vestibular afferentneuron electrical properties. Furthermore, short-term primary cul-ture of vestibular neurons gives rise to the outgrowth of neurites,and to the expression of various membrane proteins includingionic channels and neurotransmitter receptors that are usuallylocated at the nerve terminals (Soto et al. 2002a). In addition, partof the set of voltage-activated ionic channels that underlie thefiring pattern generation at the dendrites seem to be expressed atthe cell body. In both peripheral dendrites and cell body regionsvoltage-activated Na� channels (Chabbert et al. 1997; Rennie andStreeter 2006), KCNQ channels (Hurley et al. 2006; Pérez et al.2009), calcium-activated K� channels (Limón et al. 2005; Mer-edith et al. 2011), and voltage-activated K� currents (Chabbertet al. 2001a; Dhawan et al. 2010; Iwasaki et al. 2008; Kalluri et al.2010; Risner and Holt 2006) have been described. Coincidentlywith our immunohistochemical data showing the expression ofHCN4 at the calyx, evidence for the expression of Ih was obtainedby direct recording in the calyx endings, indicating that, alongwith other conductances, Ih is functionally expressed in a regioncritical in modulating the input to afferent neurons from hair cells(Meredith et al. 2011).

In summary, the four HCN channel subunits were found to beexpressed in vestibular ganglion neurons. From our data weconclude that we have recorded from a heterogeneous populationranging from the smallest to the largest afferent neuron soma. Inneonatal neurons we found no correlation between Ih amplitudeand soma size, while although all older neurons expressed Ih, inthis population there was a correlation between Ih density andsmall cell size. Thus most probably the great majority of vestib-ular ganglion neurons express the Ih, including calyx, dimorphic,and bouton afferents. In fact, the immunostaining of the ganglionneurons shows no selectivity for soma size, and staining of bothbouton- and calyxlike terminals can be identified within thevestibular neuroepithelia. To date, most studies on vestibularafferent neurons are from neonatal rats because of the difficulty ofrecording in mature neurons because of myelination of the somaafter P10 (Toesca 1996). Results obtained in P25–28 neurons arephysiologically significant, indicating that functional changes takeplace after P7–10, extending the range over which maturationprocesses are thought to occur. In hair cells from mouse utricleelectrophysiological maturity is reached after P8 (Rüsch et al.1998), while in afferent neurons some features of mature dis-charge patterns (like the regularity of the spontaneous discharge)begin after P3 (Desmadryl et al. 1986) and the sensitivity of thevestibular nerve also is increased after birth (Desmadryl 1991).Thus developmental changes after P7–10 in current density,voltage sensitivity, and kinetics could enhance differences inspiking of afferent neurons between neonatal and young rats.Furthermore, modulation of Ih by activation of cAMP-mediatedsignaling pathways and by other factors in the intracellular milieucould constitute a control mechanism by which neurotransmittersor intracellular process may modulate vestibular afferent neurondischarge.

ACKNOWLEDGMENTS

The authors thank cDr. Emmanuel Seseña Mendez for performing theperforated patch experiments, cDr. Blanca Cervantes Sanchez for her partici-pation in the immunohistochemical experiments, and Dr. Ellis Glazier forediting the English-language text.

GRANTS

This work was supported by Grants PIFI-2010 and VIEP-BUAP 2011 to R.Vega and National Council of Science and Technology of México(CONACyT) Grants I0110/127/08 to E. Soto and I0013/119257 to A. Al-manza.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: A.A., E.L., F.M., R.V., and E.S. performed experi-ments; A.A., E.L., F.M., and E.S. analyzed data; A.A. and E.S. interpretedresults of experiments; A.A., E.L., F.M., R.V., and E.S. prepared figures; A.A.and E.S. drafted manuscript; A.A., R.V., and E.S. edited and revised manu-script; A.A., E.L., F.M., R.V., and E.S. approved final version of manuscript;R.V. and E.S. conception and design of research.

REFERENCES

Accili EA, Proenza C, Baruscotti M, DiFrancesco D. From funny current toHCN channels: 20 years of excitation. News Physiol Sci 17: 32–37, 2002.

Altomare C, Bucchi A, Camatini E, Baruscotti M, Viscomi C, Moroni A,DiFrancesco D. Integrated allosteric model of voltage gating of HCNchannels. J Gen Physiol 117: 519–532, 2001.

Altomare C, Terragni B, Brioschi C, Milanesi R, Pagliuca C, Viscomi C,Moroni A, Baruscotti M, DiFrancesco D. Heteromeric HCN1-HCN4channels: a comparison with native pacemaker channels from the rabbitsinoatrial node. J Physiol 549: 347–359, 2003.

Bakondi G, Pór A, Kovács I, Szucs G, Rusznák Z. Hyperpolarization-activated, cyclic nucleotide-gated, cation non-selective channel subunitexpression pattern of guinea-pig spiral ganglion cells. Neuroscience 158:1469–1477, 2009.

Bal R, Oertel D. Hyperpolarization-activated, mixed-cation current Ih inoctopus cells of the mammalian cochlear nucleus. J Neurophysiol 84:806–817, 2000.

Banks MI, Pearce RA, Smith PH. Hyperpolarization-activated cation current(Ih) in neurons of the medial nucleus of the trapezoid body: voltage-clampanalysis and enhancement by norepinephrine and cAMP suggest a modula-tory mechanism in the auditory brain stem. J Neurophysiol 70: 1420–1432,1993.

Bean BP. The action potential in mammalian central neurons. Nat RevNeurosci 8: 451–465, 2007.

Berger T, Larkum ME, Lüscher HR. High Ih channel density in the distalapical dendrite of layer V pyramidal cells increases bidirectional attenuationof EPSPs. J Neurophysiol 85: 855–868, 2001.

Biel M, Wahl-Schott C, Michalakis S, Zong X. Hyperpolarization-activatedcation channels: from genes to function. Physiol Rev 89: 847–885, 2009.

Chabbert C, Chambard JM, Sans A, Desmadryl G. Three types of depo-larization-activated potassium currents in acutely isolated mouse vestibularneurons. J Neurophysiol 85: 1017–1026, 2001a.

Chabbert C, Chambard JM, Valmier J, Sans A, Desmadryl G. Hyperpo-larization-activated Ih current in mouse vestibular primary neurons. Neu-roreport 12: 2701–2704, 2001b.

Chabbert C, Chambard JM, Valmier J, Sans A, Desmadryl G. Voltage-activated sodium currents in acutely isolated mouse vestibular ganglionneurones. Neuroreport 8: 1253–1256, 1997.

Chen C. Hyperpolarization-activated current Ih in primary auditory neurons.Hear Res 110: 179–190, 1997.

Chen S, Wang J, Siegelbaum SA. Properties of hyperpolarization-activatedpacemaker current defined by coassembly of HCN1 and HCN2 subunits andbasal modulation by cyclic nucleotide. J Gen Physiol 117: 491–504, 2001.

Cho WJ, Drescher MJ, Hatfield JS, Bessert DA, Skoff RP, Drescher DG.Hyperpolarization-activated, cyclic AMP-gated, HCN1-like cation channel:



AQ: 6

the primary, full-length HCN isoform expressed in a saccular hair-cell layer.Neuroscience 118: 525–534, 2003.

Desmadryl G. Postnatal developmental changes in the responses of mouseprimary vestibular neurons to externally applied galvanic currents. BrainRes Dev Brain Res 64: 137–143, 1991.

Desmadryl G, Raymond J, Sans A. In vitro electrophysiological study ofspontaneous activity in neonatal mouse vestibular ganglion neurons duringdevelopment. Brain Res 390: 133–136, 1986.

Dhawan R, Mann SE, Meredith FL, Rennie KJ. K� currents in isolatedvestibular afferent calyx terminals. J Assoc Res Otolaryngol 11: 463–476,2010.

DiFrancesco D. A study of the ionic nature of the pace-maker current in calfPurkinje fibres. J Physiol 314: 377–393, 1981.

DiFrancesco D, Ferroni A, Mazzanti M, Tromba C. Properties of thehyperpolarizing-activated current (If) in cells isolated from the rabbit sino-atrial node. J Physiol 377: 61–88, 1986.

DiFrancesco D, Tortora P. Direct activation of cardiac pacemaker channelsby intracellular cyclic AMP. Nature 351: 145–147, 1991.

Eatock RA, Lysakowski A. Mammalian vestibular hair cells. In: Vertebratehair cells, edited by Eatock RA, Fay RR, Popper AN. New York: Springer,2006.

Felix R, Sandoval A, Sánchez D, Gómora JC, De la Vega-Beltrán JL,Treviño CL, Darszon A. ZD7288 inhibits low-threshold Ca2� channelactivity and regulates sperm function. Biochem Biophys Res Commun 311:187–192, 2003.

Goldberg JM. Theoretical analysis of intercellular communication betweenthe vestibular type I hair cell and its calyx ending. J Neurophysiol 76:1942–1957, 1996.

Hofmann F, Biel M, Kaupp B. Nomenclature and structure-function rela-tionship of cyclic nucleotide-regulated channels. Pharmacol Rev 57: 455–462, 2005.

Horwitz GC, Risner-Janiczek JR, Jones SM, Holt JR. HCN channelsexpressed in the inner ear are necessary for normal balance function. JNeurosci 31: 16814–16825, 2011.

Hurley KM, Gaboyard S, Zhong M, Price SD, Wooltorton JR, LysakowskiA, Eatock RA. M-like K� currents in type I hair cells and calyx afferentendings of the developing rat utricle. J Neurosci 26: 10253–10269, 2006.

Iwasaki S, Chihara Y, Komuta Y, Ito K, Sahara Y. Low-voltage-activatedpotassium channels underlie the regulation of intrinsic firing properties of ratvestibular ganglion cells. J Neurophysiol 100: 2192–2204, 2008.

Kalluri R, Xue J, Eatock RA. Ion channels set spike timing regularity ofmammalian vestibular afferent neurons. J Neurophysiol 104: 2034–2051,2010.

Kevetter GA, Leonard RB. Molecular probes of the vestibular nerve. II.Characterization of neurons in Scarpa’s ganglion to determine separatepopulations within the nerve. Brain Res 928: 18–29, 2002.

Lim R, Kindig AE, Donne SW, Callister RJ, Brichta AM. Potassiumaccumulation between type I hair cells and calyx terminals in mouse crista.Exp Brain Res 210: 607–621, 2011.

Limón A, Pérez C, Vega R, Soto E. Ca2�-activated K�-current density iscorrelated with soma size in rat vestibular-afferent neurons in culture. JNeurophysiol 94: 3751–3761, 2005.

Ludwig A, Zong X, Jeglitsch F, Biel M. A family of hyperpolarization-activated mammalian cation channels. Nature 393: 587–591, 1998.

Lysakowski A, Gaboyard-Niay S, Calin-Jageman I, Chatlani S, Price SD,Eatock RA. Molecular microdomains in a sensory terminal, the vestibularcalyx ending. J Neurosci 31: 10101–10114, 2011.

Magee JC. Dendritic hyperpolarization-activated currents modify the integra-tive properties of hippocampal CA1 pyramidal neurons. J Neurosci 18:7613–7624, 1998.

Magee JC. Dendritic Ih normalizes temporal summation in hippocampal CA1neurons. Nat Neurosci 2: 508–514, 1999.

McCormick DA, Pape HC. Properties of a hyperpolarization-activated cationcurrent and its role in rhythmic oscillation in thalamic relay neurones. JPhysiol 431: 291–318, 1990.

Mercado F, López IA, Acuna D, Vega R, Soto E. Acid-sensing ionicchannels in the rat vestibular endorgans and ganglia. J Neurophysiol 96:1615–1624, 2006.

Meredith FL, Li GQ, Rennie KJ. Postnatal expression of an apamin-sensitiveKCa current in vestibular calyx terminals. J Membr Biol 244: 81–91, 2011.

Mo ZL, Davis RL. Heterogeneous voltage dependence of inward rectifiercurrents in spiral ganglion neurons. J Neurophysiol 78: 3019–3027, 1997.

Moosmang S, Stieber J, Zong X, Biel M, Hofmann F, Ludwig A. Cellularexpression and functional characterization of four hyperpolarization-acti-

vated pacemaker channels in cardiac and neuronal tissues. Eur J Biochem268: 1646–1652, 2001.

Pape HC. Queer current and pacemaker: the hyperpolarization-activatedcation current in neurons. Annu Rev Physiol 58: 299–327, 1996.

Pérez C, Limón A, Vega R, Soto E. The muscarinic inhibition of thepotassium M-current modulates the action-potential discharge in the vestib-ular primary-afferent neurons of the rat. Neuroscience 158: 1662–1674,2009.

Rennie KJ, Streeter MA. Voltage-dependent currents in isolated vestibularafferent calyx terminals. J Neurophysiol 95: 26–32, 2006.

Risner JR, Holt JR. Heterogeneous potassium conductances contribute to thediverse firing properties of postnatal mouse vestibular ganglion neurons. JNeurophysiol 96: 2364–2376, 2006.

Robinson R, Siegelbaum S. Hyperpolarization-activated cation current: frommolecules to physiological function. Annu Rev Physiol 65: 453–480, 2003.

Rodrigues AR, Oertel D. Hyperpolarization-activated currents regulate ex-citability in stellate cells of the mammalian ventral cochlear nucleus. JNeurophysiol 95: 76–87, 2006.

Rüsch A, Lysakowski A, Eatock RA. Postnatal development of type I andtype II hair cells in the mouse utricle: acquisition of voltage-gated conduc-tances and differentiated morphology. J Neurosci 18: 7487–7501, 1998.

Santoro B, Chen S, Luthi A, Pavlidis P, Shumyatsky GP, Tibbs GR,Siegelbaum SA. Molecular and functional heterogeneity of hyperpolariza-tion-activated pacemaker channels in the mouse CNS. J Neurosci 20:5264–5274, 2000.

Santoro B, Liu DT, Yao H, Bartsch D, Kandel ER, Siegelbaum SA, TibbsGR. Identification of a gene encoding a hyperpolarization-activated pace-maker channel of brain. Cell 93: 717–729, 1998.

Santoro B, Piskorowski RA, Pian P, Hu L, Liu H, Siegelbaum SA. TRIP8bsplice variants form a family of auxiliary subunits that regulate gating andtrafficking of HCN channels in the brain. Neuron 62: 802–813, 2009.

Santos-Sacchi J. Voltage-dependent ionic conductances of type I spiralganglion cells from the guinea pig inner ear. J Neurosci 13: 3599–35611,1993.

Satoh TO, Yamada M. Multiple inhibitory effects of zatebradine (UL-FS 49)on the electrophysiological properties of retinal rod photoreceptors. PflügersArch 443: 532–540, 2002.

Seifert R, Scholten A, Gauss R, Mincheva A, Lichter P, Kaupp UB.Molecular characterization of a slowly gating human hyperpolarization-activated channel predominantly expressed in thalamus, heart, and testis.Proc Natl Acad Sci USA 96: 9391–9396, 1999.

Shi W, Wymore R, Yu H, Wu J, Wymore RT, Pan Z, Robinson RB, DixonJE, McKinnon D, Cohen IS. Distribution and prevalence of hyperpolar-ization-activated cation channel (HCN) mRNA expression in cardiac tissues.Circ Res 85: e1–e6, 1999.

Soto E, Limón A, Ortega A, Vega R. Características morfológicas y elec-trofisiológicas de las neuronas del ganglio vestibular en cultivo. GacetaMédica México 138: 1–14, 2002a.

Soto E, Vega R, Luis E. Developmental expression of inward rectifier K�

currents in vestibular afferent neurons of the rat. Abstr 31th Assoc ResOtolaryngol Meeting 31: 261–262, 2008.

Soto E, Vega R, Budelli R. The receptor potential in type I and type IIvestibular system hair cells: a model analysis. Hear Res 165: 35–47, 2002b.

Southan AP, Morris NP, Stephens GJ, Robertson B. Hyperpolarization-activated currents in presynaptic terminals of mouse cerebellar basket cells.J Physiol 526: 91–97, 2000.

Thollon C, Bedut S, Villeneuve N, Cogé F, Piffard L, Guillaumin JP,Brunel-Jacquemin C, Chomarat P, Boutin JA, Peglion JL, Vilaine JP.Use-dependent inhibition of hHCN4 by ivabradine and relationship withreduction in pacemaker activity. Br J Pharmacol 150: 37–46, 2007.

Toesca A. Central and peripheral myelin in the rat cochlear and vestibularnerves. Neurosci Lett 221: 21–24, 1996.

Ulens C, Tytgat J. Functional heteromerization of HCN1 and HCN2 pace-maker channels. J Biol Chem 276: 6069–6072, 2001.

Valenzuela C, Delpón E, Franqueza L, Gay P, Pérez O, Tamargo J,Snyders DJ. Class III antiarrhythmic effects of zatebradine. Time-, state-,use-, and voltage-dependent block of hKv1.5 channels. Circulation 94:562–570, 1996.

Wainger BJ, DeGennaro M, Santoro B, Siegelbaum SA, Tibbs GR.Molecular mechanism of cAMP modulation of HCN pacemaker channels.Nature 411: 805–810, 2001.

Yi E, Roux I, Glowatzki E. Dendritic HCN channels shape excitatorypostsynaptic potentials at the inner hair cell afferent synapse in the mam-malian cochlea. J Neurophysiol 103: 2532–2543, 2010.



Molecular identity, ontogeny, and cAMP modulation of the hyperpolarization-activated current in...

Documents

Transcript of Molecular identity, ontogeny, and cAMP modulation of the hyperpolarization-activated current in...