Assessment of the expression and role of the α₁-nAChR subunit in efferent cholinergic function during the development of the mammalian cochlea

In addition to the well characterized α₉- and α₁₀-nicotinic acetylcholine receptors (nAChRs), the “muscle-type” α₁-nAChR subunit is expressed in developing cochlear hair cells (HCs) (Scheffer et al. 2007). The study here finds that HCs do not express nAChRs with the same pharmacology as “muscle-type” nAChRs. Remarkably, the onset of α₁-expression and efferent synaptic function occur simultaneously. Yet, α₁-expression is not necessary for synapse formation as shown in α₁-deficient mice.

Abstract

Hair cell (HC) activity in the mammalian cochlea is modulated by cholinergic efferent inputs from the brainstem. These inhibitory inputs are mediated by calcium-permeable nicotinic acetylcholine receptors (nAChRs) containing α₉- and α₁₀-subunits and by subsequent activation of calcium-dependent potassium channels. Intriguingly, mRNAs of α₁- and γ-nAChRs, subunits of the “muscle-type” nAChR have also been found in developing HCs (Cai T, Jen HI, Kang H, Klisch TJ, Zoghbi HY, Groves AK. J Neurosci 35: 5870–5883, 2015; Scheffer D, Sage C, Plazas PV, Huang M, Wedemeyer C, Zhang DS, Chen ZY, Elgoyhen AB, Corey DP, Pingault V. J Neurochem 103: 2651–2664, 2007; Sinkkonen ST, Chai R, Jan TA, Hartman BH, Laske RD, Gahlen F, Sinkkonen W, Cheng AG, Oshima K, Heller S. Sci Rep 1: 26, 2011) prompting proposals that another type of nAChR is present and may be critical during early synaptic development. Mouse genetics, histochemistry, pharmacology, and whole cell recording approaches were combined to test the role of α₁-nAChR subunit in HC efferent synapse formation and cholinergic function. The onset of α₁-mRNA expression in mouse HCs was found to coincide with the onset of the ACh response and efferent synaptic function. However, in mouse inner hair cells (IHCs) no response to the muscle-type nAChR agonists (±)-anatoxin A, (±)-epibatidine, (−)-nicotine, or 1,1-dimethyl-4-phenylpiperazinium iodide (DMPP) was detected, arguing against the presence of an independent functional α₁-containing muscle-type nAChR in IHCs. In α₁-deficient mice, no obvious change of IHC efferent innervation was detected at embryonic day 18, contrary to the hyperinnervation observed at the neuromuscular junction. Additionally, ACh response and efferent synaptic activity were detectable in α₁-deficient IHCs, suggesting that α₁ is not necessary for assembly and membrane targeting of nAChRs or for efferent synapse formation in IHCs.

NEW & NOTEWORTHY

In addition to the well characterized α₉- and α₁₀-nicotinic acetylcholine receptors (nAChRs), the “muscle-type” α₁-nAChR subunit is expressed in developing cochlear hair cells (HCs) (Scheffer et al. 2007). The study here finds that HCs do not express nAChRs with the same pharmacology as “muscle-type” nAChRs. Remarkably, the onset of α₁-expression and efferent synaptic function occur simultaneously. Yet, α₁-expression is not necessary for synapse formation as shown in α₁-deficient mice.

hair cell (hc) activity in the mammalian cochlea is modulated by medial olivocochlear efferent inputs originating in the brainstem. During development, transient efferent inputs to the inner hair cells (IHCs) influence the maturation of the IHC synaptic machinery (Johnson et al. 2013) and contribute to the maturation of central auditory pathways by modulating the IHC's input to the developing auditory nerve fibers (Clause et al. 2014; Johnson et al. 2011; Kandler et al. 2009; Sendin et al. 2014; Walsh et al. 1998). In the adult cochlea, efferent inputs provide sound-evoked feedback and influence cochlear sensitivity and frequency selectivity by modulating outer hair cell (OHC) activity (Guinan 1996; Nouvian et al. 2015; Simmons 2002). These mechanisms may contribute to selective attention and the ability to hear in a noisy environment and seem to have a protective role for the ear in case of sound injury (Liberman et al. 2014; Maison et al. 2013; Rajan 1988; Scharf et al. 1994; Winslow and Sachs 1987). It is therefore important to understand how synapses between HCs and efferent innervation are formed and maintained.

For both IHCs and OHCs, inhibitory efferent synapses are mediated by calcium-permeable nicotinic acetylcholine receptors (nAChRs) containing α₉- and α₁₀-nAChRs subunits (Elgoyhen et al. 1994, 2001) and by subsequent activation of calcium-dependent potassium channels (Evans 1996; Fuchs and Murrow 1992a,b; Glowatzki and Fuchs 2000; Oliver et al. 2000; Wersinger et al. 2010). Intriguingly, in addition to α₉ and α₁₀, the expression of α₁- and γ-nAChRs subunits, mostly known to mediate cholinergic inputs in muscle (Duclert and Changeux 1995; Hall and Sanes 1993), has been documented in HCs during the first postnatal week with levels of mRNAs of the same order of magnitude as those of α₉ and α₁₀ (Cai et al. 2015; Scheffer et al. 2015; Shen et al. 2015). These data suggest that a fetal muscle-like nAChR, with the subunits (α₁)₂β₁γδ (Mishina et al. 1986) or other combinations of α₁-containing nAChRs, might exist in HCs, at least transiently. At the mouse neuromuscular junction, α₁ is necessary for the assembly of a functional nAChR complex, and lack of α₁ leads to hyperinnervation and a complete lack of spontaneous miniature and nerve-evoked endplate potentials (An et al. 2010). The embryonic γ-nAChR subunit is required for the assembly of prepatterned nAChR clusters in the initial stages of neuromuscular synaptogenesis and its absence also leads to hyperinnervation (Liu et al. 2008). Additionally, other molecules like RIC-3, rapsyn, and MuSK, essential for the assembly, clustering, and localization of nAChRs at the skeletal neuromuscular junction have been detected in HCs (Osman et al. 2008).

These findings led to the hypothesis that some of the same molecules as found in muscle, including α₁- and γ-nAChRs, might be involved in efferent synapse formation. Here, we found that the onset of α₁-expression coincides with the onset of ACh response and efferent synaptic function in early postnatal IHCs. Pharmacological experiments excluded the possibility that functional “muscle-type” nAChRs might be expressed in HC membranes in addition to postsynaptic α₉-containing nAChRs. However, α₁ could still be part of α₉-containing nAChRs or of other yet uncharacterized nAChRs. In α₁-deficient mice, no obvious change of efferent innervation was detected in the organ of Corti at embryonic day (E)18. Secondly, ACh response and efferent synaptic function were present in the IHCs of α₁-deficient mice, suggesting that α1 is not necessary for assembly and membrane targeting of nAChRs or for efferent synapse formation in HCs.

MATERIALS AND METHODS

Animal procedures.

All experimental procedures involving animals were approved by the Johns Hopkins University Animal Care and Use Committee. C57BL/6J mice (stock number no. 000664) B6.FVB-Tg(EIIa-cre)C5379Lmgd/J (no. 003724) (Lakso et al. 1996), B6.Cg-Tg(ACTFLPe)9205Dym/J (no. 005703) (Rodriguez et al. 2000), ChAT-IRES-Cre^+/− [B6;129S6-Chat^tm1(cre)Lowl/J, no. 006410] (Rossi et al. 2011), and mTmG^+/− [Gt(ROSA)26Sor^{tm4(ACTB-tdTomato,-EGFP)Luo}/J; no. 007576] (Muzumdar et al. 2007) mice were obtained as gifts or from The Jackson Laboratory (Bar Harbor, ME) and genotyped as previously described. ChAT-IRES-Cre^+/− and mTmG^+/− mice were backcrossed for at least five generations on a C57BL/6J background before being bred for experiments. Animals of both sex were studied from E17 to 18 mo old. The first day after overnight mating was counted as E0. E19 was defined as postnatal day (P)0. Animals from at least three litters were studied per data point. All studies were performed on C57BL/6 genetic background.

Chrna1-targeted alleles.

With the use of homologous recombination in C57BL/6NTac embryonic stem (ES) cells, a cassette including the coding sequence of β-galactosidase [β-gal; lacZ preceded by a splice acceptor site and an internal ribosome entry site (IRES)] followed by a promoter-driven neo marker, and surrounded by frt sites, was introduced into intron 3 of Chrna1 by the European Mouse Mutagenesis Consortium (EUCOMM). In addition, both the neo and the exon 4 of Chrna1 were flanked by loxP sites. Details of this allele [Chrna1 “Knockout-First-Reporter Tagged Insertion” allele; EUCOMM Project 33888; strain C57BL/6NTac-Chrna1^{tm1a(EUCOMM)Hmgu/ICS}, referred to here as Chrna1^tm1a] and targeting strategy can be found at: http://www.knockoutmouse.org/martsearch/project/33888 and in Fig. 1. Briefly, an automated high-throughput approach and genotyping by long-range PCR (for technical details see http://www.eucomm.org/) led to the identification of an ES cell clone (HEPD0507_1_B05) showing homologous targeting. Correct targeting event and unique insertion were further confirmed by Southern blot hybridization with a neo probe, after two digestions 5′ of neo and two digestions 3′ of neo. Targeted cells with a normal karyotype were injected into BALB/cN embryos at ICS (Institut Clinique de la Souris, Illkirch, France) and germline transmission was obtained. Mice carrying this recombinant allele were rederived from frozen embryos at the Johns Hopkins Transgenic Core (Johns Hopkins University, Baltimore, MD). To study Chrna1 expression, male mice carrying the Chrna1^tm1a allele were crossed to females B6.FVB-Tg(EIIa-cre)C5379Lmgd/J mice to excise the floxed promoter-driven neo selection cassette and obtain the reporter Chrna1^lacZ/+ mice (Fig. 1). These mice were backcrossed once to eliminate the transgene encoding Cre. Heterozygote mice carrying the Chrna1^lacZ allele were viable and fertile. These were crossed to C57BL/6J mice to provide both Chrna1^lacZ/+ and wild-type experimental littermates. The presence of lacZ was tested using primers lacZF 5′-CCGACGGCACGCTGATTGAAG-3′ and lacZR 5′-ATGCGGTCGCGTTCGGTTGC-3′ (provided by Michael Deans, University of Utah). Thirty-five amplification cycles were performed with an annealing temperature of 65°C with OneTaq DNA Polymerase (New England Biolabs) in the presence of 3% dimethyl sulfoxide (Sigma-Aldrich). Loading on a 1% agarose gel allowed visualization of the 1,170-bp lacZ band (Fig. 1B). Recombination between the most distant loxP sites and deletion of the neo cassette and exon 4 was shown using the internal primers neoF 5′-TGATGCCGCCGTGTTCCGGCTGT-3′ and neoR 5′-GGCCACAGTCGATGAATCCAGAA-3′ (lack of 536-bp amplicon) and resulted in the appearance of the expected 724-bp amplicon using primers T3F 5′-GTATACCCCGTACGTCTTCCCGAGC-3′ and T3R 5′-TCCCGAGAGTTGAGACCCCTG-3′ (Fig. 1B). To obtain mice deficient for α₁-nAChR, male mice carrying the Chrna1^tm1a allele were crossed to B6.Cg-Tg(ACTFLPe)9205Dym/J female mice, which express the recombinase FLPe in their germline, to excise the frt-lacZ-neo-frt cassette. The resulting Chrna1^fl/+ mice [identified by PCR amplicons of 299 (fl) and 143 (+) bp with primers Efbis 5′-GCTAGGGAGACCATAGCCATG-3′ and Er 5′-GAATTGATTTCCCCTCAGTCCTC-3′] were backcrossed once to eliminate the transgene encoding FLPe. The Chrna1^Δ4 allele was subsequently obtained by deletion of exon 4 by Cre-mediated recombination, also obtained in vivo, by crossing Chrna1^fl/+ male mice with B6.FVB-Tg(EIIa-cre)C5379Lmgd/J female mice. After one backcross to eliminate the transgene encoding Cre, the Chrna1^Δ4/+ line [identified by PCR amplicons of 310 (Δ4) and 909 (+) bp with primers Efbis and L3r 5′-AGAGTTGAGACCCCTGGGGCGG-3′; Fig. 1, A and C] was intercrossed to generate experimental Chrna1^Δ4/Δ4, Chrna1^Δ4/+, and Chrna1^+/+ littermates. Such littermates were obtained in the expected 1:2:1 ratio at E18.5. For each mouse line appropriate recombination was confirmed by sequencing. Consistent with deletion of exon 4 leading to a frame-shift and a premature STOP codon termination and previous observations of mouse mutants lacking α₁-nAChR (An et al. 2010), Chrna1^Δ4/Δ4 mice die at birth, most likely due to the loss of their diaphragm function. At E18.5, Chrna1^Δ4/Δ4 mice show characteristic kyphosis (hunchback) and carpoptosis phenotypes (Fig. 1D), as reported by An et al. (2010) and as found in choline acetyltransferase and vesicular acetylcholine transporter-deficient mice (Brandon et al. 2003; de Castro et al. 2009; Misgeld et al. 2002).

Fig. 1.

Generation of the reporter Chrna1^lacZ and knockout Chrna1^Δ4 alleles. A, top: genomic region of Chrna1. Middle: Chrna1^tm1a allele was obtained by insertion of the “Knock Out first allele” cassette flanked by frt sites, into intron 3 of Chrna1, the gene coding for the α₁-nicotinic acetylcholine receptor (α₁-nAChR) subunit. This cassette encodes β-galactosidase (β-gal; lacZ) preceded by an internal ribosome entry site (IRES) and a promoter driven neomycin resistance cassette (neo). Three loxP sites were also introduced: 5′ of the neo cassette, and 5′ and 3′ of exon 4 (EMMA Consortium). Bottom: Chrna1^lacZ allele was generated after excision of the neo cassette and exon 4 by germline Cre recombination between the 2 loxP sites further apart. SA, splice acceptor and pA, polyadenylation site. Chrna1^fl allele was generated after excision of the “Knock Out first allele” cassette by germline FLPe recombination between the 2 frt sites flanking this cassette. Chrna1^Δ4 allele was subsequently generated after excision of exon 4 by germline Cre recombination between the two loxP sites 5′ and 3′ of this exon. B: genotyping of Chrna1^lacZ/+ mice by PCR. The primers lacZF and lacZR were used to detect lacZ-containing alleles (1,170-bp amplicon). Lack of amplification with primers neoF/neoR and appearance of a 724 bp amplicon with primers T3F/T3R attest of the excision of the neo cassette and exon 4 in Chrna1^lacZ/+ mice. Note a 536-bp amplicon with primers neoF/neoR was present in Chrna1^tm1a/+, but the expected 3,417-bp amplicon with primers T3F/T3R was not detected with the PCR conditions used here. C: genotyping of embryonic day (E)18.5 wild-type (Chrna1^+/+), heterozygote (Chrna1^Δ4/+), and homozygote knockout (Chrna1^Δ4/Δ4) mouse littermates by PCR. Deletion of exon 4 leads to the appearance of a 310-bp amplicon instead of the 909-bp amplicon detected in wild-type allele with primers Efbis and L3r. D: an E18.5 Chrna1^Δ4/Δ4 embryo compared with a Chrna1^Δ4/+ control littermate. Note the hunched posture of Chrna1^Δ4/Δ4 mouse. E: confirmation of the absence of exon 4 from Chrna1 transcripts in Chrna1^Δ4/Δ4 mice. RT-PCR analysis of Chrna1 expression in inner ear tissue of E18.5 wild-type (Chrna1^+/+), heterozygote (Chrna1^Δ4/+), and homozygote knockout (Chrna1^Δ4/Δ4) mouse littermates, using primers located in exons 2 and 5. As expected, a smaller amplification product was found in Chrna1^Δ4/Δ4 mice. The 223-bp amplicon reflects the 110-bp deletion of exon 4 with respect to the 333-bp RT-PCR product detected in Chrna1^+/+ mice. Note the persistence of Chrna1 mRNA in Chrna1^Δ4/Δ4 mice, also shown with primers located in exons 5 and 7. RT-PCR amplifications of Hprt, a constitutively expressed housekeeping gene, and of Chrng, the gene that codes for γ-nAChR subunit, were used as positive controls. The presence of 2 amplicons for Chrng reflects the presence of alternatively spliced transcripts with or without exon 5, also found in muscle (Mileo et al. 1995). +/−RT indicates the presence (+) or the absence (−) of reverse transcriptase in the cDNA synthesis reaction. All studies were performed on a C57BL/6 genetic background.

Histochemical staining.

For all staining methods, the cochleae from E17- to 18-mo-old mice were harvested, perfused through the round and oval windows (for P4 and older), and fixed for 15 min at 4°C with 4% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA) freshly diluted in phosphate buffered saline (PBS). After three 5-min washes in PBS, the neuroepithelia were microdissected. For most preparations, tissue from one ear was used for X-gal staining and from the contralateral ear for β-gal immunolabeling. Consistent with the strong expression of Chrna1 in muscle, Chrna1^lacZ/+ embryos showed an intense X-gal staining in skeletal muscle (data not shown; Klarsfeld et al. 1991).

To detect β-gal activity, the whole-mount preparations were incubated with filtered prewarmed X-gal solution {1 mg/ml X-gal/5% N,N dimethylformamide, 5 mM potassium ferricyanide crystalline [K₃Fe(CN)₆], 5 mM potassium ferrocyanide trihydrate [K₄Fe(CN)₆-3H₂O], and 2 mM MgCl₂ diluted in PBS (pH 7.3) with 0.02% NP40 and 0.1% sodium deoxycholate}, protected from light, for 42 h at 37°C, with gentle agitation. After a rapid wash in PBS, the preparations were postfixed overnight at room temperature (RT) with 4% paraformaldehyde diluted in PBS. After three washes in PBS, the cochlear tissue was mounted on slides and imaged.

For immunolabeling, the whole-mount preparations were incubated for 1 h at RT in a blocking and permeabilizing solution (BPS: PBS with 10% of normal goat serum and 0.3% Triton X-100), before being incubated for 42 h at 4°C with the primary antibodies diluted in the same solution. Chicken anti-β-gal (ab9361; Abcam, Cambridge, MA) and rabbit anti-myosin VI (Sigma, St. Louis, MO) antibodies were used as primary antibodies at 1:1,000 and 1:500 dilutions. After three 30-min washes in BPS, the samples were incubated for 1 h at RT with the secondary antibodies diluted at 1:750 in BPS. Alexa 488 goat anti-chicken IgG and Alexa Fluor 568 goat anti-rabbit (Molecular Probes, Eugene, OR) were used as secondary antibodies. Samples were rinsed twice for 30 min in BPS and once for 30 min in PBS at RT with agitation before the organs of Corti were mounted on slides in a drop of FluorSave mounting medium (Calbiochem, San Diego, CA) and observed using a LSM 700 confocal microscope (Zeiss) with a ×10 objective and a C-Apochromat 63×/1.20 W Korr M27 objective N.A. 1.2 (z-axis step size: 0.20 μm). Reconstructions and analysis were carried out using AIM 4.2 and Zen softwares (Zeiss), and Imaris 8.1.2 (Bitplane). For all experiments, Chrna1^lacZ/+ mice and wild-type control littermates were processed in parallel to assure specificity of the labeling, and three to four animals of each genotype obtained from at least three different litters were used. The presence of β-gal immunoreactivity (punctate labeling) was found in HCs and was quantified for each tissue sample in about 25 IHCs and 75 OHCs per region of the cochlea and presented as the percentage of HCs labeled (Fig. 2). Unless indicated otherwise, the images presented were taken from the apical turn of the cochlea to match the area of recording, but medial and basal turns of the cochlea were also studied (see quantification Fig. 2E). For cholinergic innervation labeling, samples were processed as described above with minor modifications. Samples were fixed for 1 h. BPS contained PBS with 10% of donkey serum and 0.3% Triton X-100. Goat anti-GFP (SICGEN, Carcavelos, Portugal) and rabbit anti-myosin VI antibodies were used as primary antibodies at 1:5,000 and 1:250 dilutions. Alexa 488 donkey anti-goat and Alexa Fluor 647 donkey anti-rabbit (Molecular Probes) were used as secondary antibodies (1:750 dilutions).

Fig. 2.

Developmental expression pattern of Chrna1 in cochlear hair cells. A-B: developmental expression pattern of Chrna1 in the cochlea as revealed by X-gal staining of Chrna1^lacZ/+ reporter mice. A: X-gal staining of Chrna1^lacZ/+ whole-mount cochlear preparation shows Chrna1 expression in the organ of Corti all along the cochlea at postnatal day (P)8. Scale bar: 100 μm. B: close-up views of the apical region of the cochlea at different ages. X-gal staining (blue dots) was detected in the organ of Corti in the inner hair cell (IHC) region at P4 and in the IHC and outer hair cell (OHC) regions at P8. From P16 onward, X-gal staining was restrained again to the IHC region and was still detectable at 6 wk, 6 mo, and 18 mo. No staining was detected in wild-type (WT) littermates (here shown at 18 mo). Scale bars = 10 μm. C–E: confocal analysis of Chrna1^lacZ/+ and WT littermates whole-mount cochlear preparations labeled with antibodies directed against β-gal (green) and myosin VI (hair cell marker, red). C: examples of maximal intensity projection (xy plane) at P8. Scale bar = 10 μm. D: single plane images in 3 perpendicular planes showing that β-gal immunoreactivity is localized in IHCs and OHCs. E: quantification of the percentage of IHCs and OHCs with β-gal labeling in the apical, medial, and basal regions of the cochlea at different ages. The number of HCs analyzed is indicated above the histogram bars.

Rat mAb35 (recognizes α₁- and α₅-nAChRs), mAb61 (recognizes α₁-nAChR), mAb210 (recognizes α₁- and possibly α₅-nAChRs; kind gifts from Dr. Jon Lindstrom, Medical School of the University of Pennsylvania, Philadelphia, PA; Lindstrom 2000), and rabbit ANC-001 (Alomone Labs), a putative α₁-nAChR antibody, were tested at different concentrations on wild-type and, when appropriate, α1-deficient cochlear tissue. None gave unambiguous specific results in either Western blot (at P4) or immunohistochemistry (at P0 and P4) experiments.

Electrophysiological recordings.

Excised apical or basal cochlear coils of E17 to 6-wk-old mice were placed into a chamber under an upright microscope (Axioskope; Zeiss, Oberkochen, Germany) and superfused at 2–3 ml/min (chamber volume ∼2 ml), with an external solution (ExS) containing the following (in mM): 5.8 KCl, 144 NaCl, 1.3 CaCl₂, 0.9 MgCl₂, 0.7 NaH₂PO₄, 5.6 d-glucose, and 10 HEPES [300 mOsm, pH 7.4 (NaOH)]. For external solutions with 80 mM KCl, KCl was substituted for NaCl. IHCs were visualized on a monitor via a ×40 water immersion objective, ×4 magnification, differential interference contrast optics using a green filter, and camera with contrast enhancement (NC70 Newvicon camera; Dage MTI). The internal solution contained the following (in mM): 135 KCl, 3.5 MgCl₂, 0.1 CaCl₂, 5 EGTA, 5 HEPES, and 2.5 Na₂ATP [290 mOsm, pH 7.2 (KOH)].

Recording pipettes were fabricated from a 1-mm borosilicate glass (WPI, Sarasota, FL) and coated with Sylgard (Dow Corning, Midland, MI). Pipettes were pulled with a multi-step horizontal puller (Sutter, San Rafael, CA) and pipette resistances were ∼3–5 MΩ. Recordings were performed at RT (22–25°C). Currents were recorded either with an Axopatch 200B amplifier or a Muliclamp 700A (Molecular Devices, Union City, CA) and digitized at 10–50 kHz with a Digidata 1322A or 1440A board and low-pass filtered at 1–10 kHz. Data were stored with PClamp10.3 or PClamp 9.2 software (Molecular Devices). Voltages were corrected offline for liquid junction potentials (approximately −4 mV) but were not corrected for the voltage drop across the uncompensated series resistance. Rs was ∼9–12 MΩ. Only recordings with holding currents <200 pA at a holding potential of −94 mV were included. An inward current >10 pA at a holding potential of −94 mV was accepted as a positive response to the agonist tested.

Nicotine was purchased from MP Biomedicals (Santa Ana, CA) as (−)-nicotine tartrate, a tartrate salt (nicotine hydrogen tartrate and nicotine bitartrate) with an anhydrous molecular weight of 462.4. However, the tartrate salt crystallizes as the dihydrate and, therefore, a molecular weight of 498 was used for calculating concentration (Matta et al. 2007). Nicotine solution was prepared daily from powder and adjusted for pH with NaOH. (±)-Anatoxin A fumarate was purchased from Abcam Biochemicals (Cambridge, UK). The conotoxins α-RgIA and αA-OIVA were synthesized as previously described (Ellison et al. 2006; Teichert et al. 2004). Conotoxins were dissolved daily from lyophilized aliquots and applied with 0.1 mg/ml bovine serum albumin. All other chemicals, including acetylcholine chloride, 1,1-dimethyl-4-phenylpiperazinium iodide (DMPP), (±)-epibatidine dihydrochloride hydrate, atropine, cadmium chloride, strychnine hydrochloride, and (+)-tubocurarine hydrochloride, were purchased from Sigma and dissolved in external solution to their final concentrations from frozen stocks daily.

Drugs were applied by using either a gravity-driven flow pipette (100-μm diameter), connected to a VC-6 channel valve controller (Warner Instrument, Hamden, CT), or a SmartSquirt Micro-Perfusion System (with a 250-μm diameter) connected to a ValveLink8.2 Controller (Automate Scientific, Berkeley, CA) (20–30 kPa), both positioned ∼50 μm from the recorded IHC. With these application systems, which allow application to the same cell of several agonists at different concentrations, drugs were washed in and out within ∼5–15 s. Even if the α₁-containing muscle-type nAChR shows fast and close to complete desensitization, when investigated with piezo-driven fast application and outside out patches (Papke et al. 2011; Papke and Grosman 2014), when investigated with a slower application system, it will still provide a response. This is due to the summed response of the small percentage of nondesensitized receptors in the cell. Therefore, the “slow” application time course used here was appropriate for testing for a muscle-type nAChR in HCs. Studies with application systems of a similar time course have shown responses of the muscle-type nAChRs to similar concentrations of the agonists used here, in native tissue (frog neuromuscular junction; Feltz and Trautmann 1982) and transfected Xenopus oocytes expressing Torpedo electric organ muscle and human muscle nAChRs (Gerzanich et al. 1995), human adult muscle nAChRs (Jonsson Fagerlund et al. 2009), and embryonic mouse muscle nAChRs (Demuro et al. 2001; Garcia-Colunga and Miledi 1999; Papke et al. 2010; Scheffer et al. 2007; Yost and Winegar 1997). Subsets of experiments were performed in the presence of 200 nM atropine to block the potential activation of muscarinic acetylcholine receptors and in 200 μM Cd²⁺ to block efferent synaptic transmission.

Data were analyzed offline using Clampfit 10.3 (Axon Instruments) and Origin 8 (OriginLab, Northampton, MA). For presentation, data were low-pass filtered offline using an 8-pole Bessel filter with a 1.4-kHz cut-off, and reduced 50 times. Values are presented as mean ± SD with n representing the number of recordings.

Reverse transcription-polymerase chain reaction.

Material was isolated in RNAlater and quickly frozen. RNA extractions were performed using Trizol reagent (Invitrogen). Reverse transcriptase (RT) reactions were performed using Superscript III (Invitrogen) with random primers p(dN)₆ (Roche) and polymerase chain reaction (PCR) reactions using GoTaq Flexi polymerase (Promega). Primers located in exon 5 (5′-CTGGGCACCTGGACCTATGA-3′) and exon 7 (5′-GACGGTGATGATGATGGACG-3′) of Chrna1 (Scheffer et al. 2007) were used to attest of the presence of Chrna1 mRNA in the neuroepithelium of the apical region of the cochlea at different ages or in whole inner ear tissue. Additionally, primers located in exon 2 (5′-GTACAAGTCACCGTGGGTCTACAG-3′) and exon 5 (5′-CTCATCGAAGGGAAAGTGAGTGAC-3′) were used to confirm the lack of exon 4 in Chrna1^Δ4/Δ4 mice inner ear at E18.5 (223-bp amplicon instead of the 333-bp amplicon found in wild-type). This deletion was confirmed by sequencing of the amplification products. RT-PCR amplification of Hprt transcript with primers 5′-GCTGGTGAAAAGGACCTCT-3′ and 5′-CACAGGACTAGAACACCTGC-3′ (248-bp amplicon) was used as a positive control. Primers located in exons 3/4 (5′-GATGCAATGGTGCGACTATCGC-3′) and exon 6 (5′-GCCTCCGGGTCAATGAAGATCC-3′) of Chrng were used to confirm the expression of the γ-nAChRs subunit (360- and 204-bp amplicons, the latter resulting from alternative splicing of exon 5) (Mileo et al. 1995; Yamane et al. 2002). To avoid any contamination, the two muscles of the tympanic cavity, the stapedius muscle and the tensor tympani muscle, were carefully removed before the tissue was harvested.

RESULTS

Developmental expression pattern of Chrna1 in the mammalian cochlea.

To study the expression pattern of the α₁-nAChR subunit in the developing and adult cochlea, a reporter mouse was used, in which β-gal expression provides a correlate of the expression of Chrna1, the gene coding for the α₁-nAChR subunit. The Chrna1^lacZ allele was obtained by germline Cre recombination using a mouse line carrying an allele with a targeted PreConditional mutation in Chrna1, here referred to as Chrna1^tm1a (Friedel et al. 2007) (materials and methods; Fig. 1, A and B). In the Chrna1^lacZ allele, the lacZ coding sequence preceded by a splice acceptor site and an IRES replaces exon 4 of Chrna1, so that lacZ/β-gal expression is under the control of Chrna1 transcriptional regulatory sequences.

X-gal staining and immunolabeling were used to detect β-gal expression in whole-mount preparations of inner ear tissue of Chrna1^lacZ/+ mice. Both methods provided consistent results. X-gal staining resulted in a blue precipitate and could be analyzed for a wide age range (E17 to 18 mo). Along the cochlear coil, X-gal staining was detected exclusively in the organ of Corti in the IHC and OHC regions (Fig. 2). Similarly, X-gal staining was found in the HC region of the vestibular sensory epithelia (data not shown). To identify the cells expressing Chrna1, coimmunolabeling of β-gal and the HC marker myosin VI was performed. β-gal immunolabeling appeared as intracellular punctate labeling (Fig. 2C), similar to what has been shown with other reporter mouse lines showing β-gal expression in HCs, like Atoh1-lacZ and Neurod1-lacZ (Matei et al. 2005). After confocal imaging and three-dimensional reconstruction, β-gal labeling was viewed in three perpendicular planes and showed that Chrna1 was exclusively expressed in IHCs and OHCs (Fig. 2D). This approach could only be performed for ages up to P12, as background labeling increased with age. However, based on the β-gal immunolabeling data, it is highly likely that the X-gal labeling at older ages also represents HC expression of Chrna1.

To chart the spatiotemporal expression pattern of Chrna1 in cochlear HCs, whole-mount preparations were studied at multiple ages (E17, P0, P2, P4, P8, P12, P16, 6 wk, and 6 mo) in apical, medial and basal cochlear turns. For the ages E17 to P12, quantitative analysis of β-gal immunolabeling provided the percentage of HCs expressing Chrna1 (Fig. 2E). At E17, in the apical turn of the cochlea, no Chrna1 expression was detected (Fig. 2, B and E). At P0, Chrna1 expression was found in 14% of IHCs. The percentage of IHCs expressing Chrna1 highly increased with age reaching 61% by P2, 87% by P4, and 100% by P8. Chrna1 expression in IHCs persisted after hearing onset (at ∼P10; Ehret 1976) as shown at P16 and P21 and even into adulthood (6 wk, 6 mo) (Fig. 2B). At 18 mo, X-gal staining was still found in most apical IHCs and in a subset of medial IHCs (base not tested). The persistent expression of Chrna1 in the adult cochlea was further confirmed by RT-PCR experiments using tissue from 6-mo-old apical cochlear coils (wild-type C57BL/6J mice; data not shown).

For OHCs, in the apical turn of the cochlea, β-gal labeling was found in a narrow time window, at P8 and P12 (in 58 and 20% of OHCs, respectively) (Fig. 2E). By P16, X-gal staining was no longer detected (Fig. 2B). For Chrna1 expression in medial and basal IHCs and OHCs, the sequence of events was similar to what was found in the apex but shifted to younger ages (Fig. 2E).

The onset of Chrna1 expression in IHCs at ∼P0-P4 and in OHCs at ∼P4-P8 in the apex of the cochlea seems to coincide with the approximate time at which ACh receptors and cholinergic efferent synapses appear in these two cell types (Bruce et al. 1997; Marcotti et al. 2004; Roux et al. 2011; Sobkowicz and Emmerling 1989) (for review see Simmons 2002). Intriguingly, Chrna1 expression in IHCs persists into adulthood, whereas the ACh response and transient efferent synapses onto IHCs disappear after hearing onset (Katz et al. 2004; Marcotti et al. 2004; Roux et al. 2011). On the other hand, Chrna1 expression in OHCs is detected only transiently around the time of efferent synapse formation, whereas OHC efferent synapses stay active throughout adulthood (for review see Guinan 1996; Nouvian et al. 2015; Simmons 2002).

Functional development of ACh receptors and cholinergic efferent synapses in mouse apical IHCs.

To allow for a direct comparison of the developmental expression profile of Chrna1 and the functional development of nAChRs and cholinergic efferent synapses, ACh-activated currents and efferent synaptic activity in IHCs were recorded in C57BL/6J mice. Whole cell voltage-clamp recordings at a holding potential of −94 mV were performed from IHCs in acutely excised apical cochlear turns. To measure the ACh response, 1 mM ACh was applied via local perfusion for 15 s. To induce synaptic activity, efferent terminals were depolarized by a 4-min long local perfusion of extracellular solution with 80 mM K⁺ (Glowatzki and Fuchs 2000; Katz et al. 2004; Roux et al. 2011). The high concentrations of ACh and K⁺ and the long application times have been proven to uncover small ACh responses and synaptic currents in developing HCs that had been missed with shorter application times or lower concentrations (100 μM ACh; 40 mM K⁺). Figure 3A shows an example of ACh-induced inward currents recorded at P4. ACh current amplitudes ranged from 10 to 1356 pA, depending on age. At E17, IHCs did not respond to ACh. ACh-evoked currents were first detected at P0 in 23% of IHCs, and the percentage of responding IHCs increased to 84% at P2 and was 100% at P4, P8, and P12 (Fig. 3B). After hearing onset, the percentage of IHCs responding to ACh was still 100% at P16, but then decreased to 32% at P21. At 6 wk, no ACh-evoked current was detected (Fig. 3B).

Fig. 3.

Functional development of nAChRs and cholinergic efferent synapses in mouse apical IHCs. A: whole cell patch-clamp recordings from IHCs in excised apical cochlear coils; holding potential −94 mV. Application of 1 mM ACh evoked an inward current, and application of 80 mM K⁺ activated synaptic currents (*) in IHCs at P4. However, no ACh response or synaptic current was evoked at P0 or at 6 wk. 80 mM K⁺ induced a steady inward current in IHCs at all ages due to the change in potassium equilibrium potential. B: percentage of apical IHCs with ACh responses and synaptic currents at different ages. ACh responses were detected earliest at P0, and synaptic currents at P2. By P4, 100% of IHCs exhibited ACh responses and synaptic currents. At P16 and P21, the percentage of IHCs showing ACh responses and synaptic currents declined and reached 0% by 6 wk. The number of IHCs tested per age is indicated at the top of the histogram bars.

When extracellular solution with 80 mM K⁺ was applied, IHCs responded with a steady inward current due to the change in the K⁺ equilibrium potential (Fig. 3A, all ages), and when efferent synapses were functional, a flurry of synaptic events were detectable (Fig. 3A, at P4). As expected from cholinergic currents mediated by activation of α₉-containing nAChRs, these synaptic currents were completely and reversibly blocked by 1 μM tubocurarine (P4; n = 2), a general nAChR antagonist, and by 1 μM strychnine (P3-P4; n = 7), a glycine receptor antagonist that also blocks α₉-containing nAChRs (Elgoyhen et al. 1994, 2001; Sgard et al. 2002) (data not shown). No synaptic currents were evoked at E17 or P0. Efferent synaptic activity was first detected at P2 in 57% of IHCs (Fig. 3B). At P4 to P12, 100% of IHCs showed efferent synaptic currents, and at P16 and P21, the fraction of responding IHCs decreased to 26% and 15%. No more synaptic events were detected in 6-wk-old IHCs.

These results are consistent with what has been shown before in rat (Katz et al. 2004; Roux et al. 2011). Within a few days during early postnatal development, at first functional nAChRs are formed in apical IHC membranes, and the percentage of IHCs with ACh responses increases until it reaches 100%. A couple of days after the first ACh responses are seen, efferent synaptic responses appear in an increasing number of IHCs, until all IHCs show synaptic currents. In mouse, these events occur with a delay of 1–2 days compared with rat (Roux et al. 2011). After hearing onset, in mouse similarly as in rat, ACh response and synaptic currents are lost (Katz et al. 2004; Marcotti et al. 2004; Roux et al. 2011). Interestingly, Chrna1 expression, nAChRs currents, and efferent synapse function start at about the same time during development, at P0 to P2, and “ramp up” during the same time window, with an increasing percentage of IHCs expressing Chrna1, exhibiting nAChR currents or synaptic activity between P0 and P4. Additionally, as shown in rat, the onset of nAChR currents and efferent synapse function are delayed by a few days in OHCs compared with IHCs (Roux et al. 2011). Similarly, Chrna1 expression in mice occurs a few days later in OHCs compared with IHCs, suggesting that also in the OHC region the onset of Chrna1 expression, nAChR currents, or synaptic activity may coincide. This parallel development of Chrna1 expression and efferent synapse formation further supported the hypothesis that the expression of α₁-nAChR subunit might be functionally associated with efferent synapse formation or function in the cochlea.

Test for the presence of functional “muscle-type” α₁-containing nAChRs in IHCs.

It is well established that ACh activated currents and efferent synaptic currents in cochlear HCs are mediated by nAChRs containing α₉- and α₁₀-nAChRs subunits (Elgoyhen et al. 1994, 2001; Ellison et al. 2006; Gomez-Casati et al. 2005; Luo et al. 1998; Morley and Simmons 2002; Sgard et al. 2002; Simmons and Morley 1998; Vetter et al. 2007; Zuo et al. 1999). However, the presence of the mRNA of α₁-nAChR subunit in HCs (Cai et al. 2015; Scheffer et al. 2007; Sinkkonen et al. 2011; Fig. 2) led to the question whether the α₁-nAChR subunit is involved in forming functional ion channels in the HC membrane, either independently of α₉α₁₀ or as part of an α₉α₁₀-containing nAChRs. Some evidence points to the possibility that a muscle-type α₁-containing nAChR might form in HCs. The mRNAs of the α₂-, α₃-, α₄-, α₅-, α₆-, α₇-, β₁-, β₂-, β₄-, γ-, and δ-nAChRs subunits have been found in immature mouse cochlear HCs by in situ hybridization, RT-PCR, or RNA-Seq (Cai et al. 2015; Scheffer et al. 2007, 2015; Shen et al. 2015). These data suggest that a fetal muscle-like nAChR, with the subunits (α₁)₂β₁γδ or other combinations of α₁-containing nAChRs may exist in HCs.

To test if α₁-containing nAChRs form functional ion channels in HCs independently from α₉α₁₀ nAChRs, classical agonists of the muscle nAChRs, (±)-anatoxin A, (±)-epibatidine, DMPP, and (−)-nicotine, were applied during whole cell recording (Akk and Auerbach 1999; Cooper et al. 1996; Gerzanich et al. 1995; Michelmore et al. 2002; Papke et al. 2010; Swanson et al. 1986; Yost and Winegar 1997). These agonists of α₁ do not, or only poorly, activate α₉- and α₉α₁₀-nAChRs when expressed in oocytes (Baker et al. 2004; Elgoyhen et al. 1994, 2001; Sgard et al. 2002; Verbitsky et al. 2000). Recordings were performed in apical IHCs at P4, as at this age almost all IHCs express Chrna1 (Fig. 2). A subset of experiments was repeated at P8, to account for the possibility that a delay between mRNA expression and formation of the functional receptor might occur. ACh (1 mM) activated an inward current in 100% of HCs at P4 and P8 (Fig. 3). However, neither of the classical muscle nAChRs agonists evoked any response in IHCs [10 μM (±)-anatoxin at P4: n = 7; at P8: n = 13; 30–100 μM (±)-epibatidine at P4: n = 8; at P8: n = 8; 100 μM DMPP at P4: n = 13; at P8: n = 4; 1–5 mM (−)- nicotine at P4: n = 14; at P8: n = 3; Fig. 4A]. Additionally, in OHCs that responded to 1 mM ACh, 10 μM (±)-anatoxin, the most potent of the muscle nAChR agonists used in this study, did not elicit any response (n = 3). The experiments in OHCs were performed at the time when the first ACh responses can be detected in the apex, at P5 to P6. Similarly, Scheffer et al. (2007) found a negative response to nicotine in IHCs at P4 at concentrations of 100 μM to 1 mM. These observations argue against the hypothesis that α₁ is involved in forming a functional “muscle-type” (α₁)₂β₁γδ-receptor in IHCs. However, it cannot be excluded that α₁ in association with a different combination of nAChR subunits may form a yet uncharacterized receptor with an unknown pharmacological profile.

Fig. 4.

Test for a functional correlate associated with the expression of the α₁-nAChR subunit in developing IHCs. Whole cell voltage-clamp recordings from IHCs in excised apical cochlear coils at P4. A: sample traces of recordings from 2 IHCs. ACh (1 mM) evoked inward currents in all IHCs tested at P4. However, the “muscle-type” α₁-nAChR agonists 1,1-dimethyl-4-phenylpiperazinium iodide (DMPP), (±)-anatoxin A, (±)-epibatidine, and (−)-nicotine, which do not or only poorly activate α₉-containing nAChRs, did not activate any response in IHCs. Holding potential was −94 mV. B: response to 1 mM ACh was not significantly blocked by αA-conotoxin OIVA (αA-OIVA), a peptide toxin that blocks the fetal muscle form of α₁-nAChRs by specifically binding to its α₁γ interface. However, α-conotoxin RgIA (α-RgIA), a subtype specific blocker of α₉-containing nAChRs, did mostly block the ACh response. Holding potential of −80 mV.

If α₁ and γ were part of a nAChR with a new subunit combination, and their association would recreate the same α₁γ-interface as has been found in the “muscle-type” receptor, αA-conotoxin OIVA (αA-OIVA), a peptide toxin that specifically binds at the α₁γ subunits interface, and thereby blocks the ACh response in fetal muscle (Teichert et al. 2004, 2008), should block such a new nAChR type. However, ACh-activated currents in IHCs were not significantly blocked by 1 μM αA-OIVA (at P4; n = 5), whereas they were reversibly blocked by α-RgIA, a specific blocker of α₉-containing nAChRs (Ellison et al. 2006) (Fig. 4B). At P4, 300 nM α-RgIA mostly blocked the response to 1 mM ACh (n = 3) and completely blocked the response to 100 μM ACh (n = 3). Altogether these results are consistent with the idea that an α₉-containing nAChR without the contribution of a separate muscle-type nAChR (α₁)₂β₁γδ mediates the ACh response in IHCs. However, additional pharmacological tools and antibodies need to be developed to further test if α₁ is part of the α₉-containing nAChRs in cochlear HCs.

Test for a role of α₁-nAChR subunit in setting up the cochlear efferent innervation pattern.

To investigate α₁-nAChR's possible role in setting up cochlear efferent innervation pattern, Chrna1^Δ4/Δ4 knockout mice were generated (see MATERIALS AND METHODS and Fig. 1, A and C). These mice lack exon 4 of Chrna1, leading to a frame-shift and a premature STOP codon (Fig. 1E). Consistent with the phenotype of mice lacking Chrna1 expression described by An et al. (2010), Chrna1^Δ4/Δ4 mice exhibit a hunched back and die at birth, due to the critical role of α₁-nAChR in diaphragm function (Fig. 1D). However, experiments could be performed on embryos shortly before birth, at E18 to E18.5.

In muscle, lack of α₁ results in an increased number of motoneurons and in increased nerve branching (An et al. 2010). It was investigated if Chrna1^Δ4/Δ4 mice present a similar phenotype regarding their efferent cochlear innervation pattern. A genetic labeling approach was used to visualize the developing efferent neurons in Chrna1^Δ4/Δ4 mice and control littermates. As choline acetyltransferase (ChAT), an enzyme involved in synthesizing ACh, is expressed by efferent cholinergic neurons, ChAT-IRES-Cre (Rossi et al. 2011) was used to drive the expression of membrane-targeted green fluorescent protein (mG) from the double-fluorescent Cre reporter allele mTmG (Muzumdar et al. 2007) and visualize the efferent neurons. Chrna1^Δ4/+;ChAT-IRES-Cre^+/− male mice were bred with Chrna1^Δ4/+;mTmG/mTmG females. In their progeny, all cell membranes were labeled with red fluorescent protein (mTomato), except for the cell membranes of cells expressing ChAT; these membranes were labeled with green fluorescent protein (mG, for membrane bound EGFP).

In control ChAT-IRES-Cre^+/−;mTmG^+/−;Chrna1^+/+, at E18, cholinergic EGFP labeled radial fibers with endings in close vicinity to IHCs were present all along the cochlea, with a higher density in the median and basal region compared with the apex (Fig. 5B). Magnified views of the HC area in the base of the cochlea show incoming efferent fibers to the IHC area (Fig. 5E) and efferent boutons endings at IHCs (Fig. 5F). These findings are consistent with previous observations using anterograde labeling with the diffusible carbocyanines dye DiI applied to the cut central axons of olivocochlear neurons (Bruce et al. 1997). No obvious difference was found in the gross innervation patterns between ChAT-IRES-Cre^+/−;mTmG^+/−;Chrna1^Δ4/Δ4 knockout mice (Fig. 5, A, C, and D) and their control littermates (ChAT-IRES-Cre^+/−;mTmG^+/−;Chrna1^+/+) (Fig. 5, B, E, and F) suggesting that α₁-nAChR is not necessary for setting up the branching pattern of efferent fibers in the cochlea, at least as can be observed just before birth.

Fig. 5.

Test for the influence of α₁-nAChR subunit on efferent cholinergic fiber patterning. Comparison of the cholinergic efferent innervation of ChAT-IRES-Cre^+/−;mTmG^+/−;Chrna1^Δ4/Δ4 and ChAT-IRES-Cre^+/−;mTmG^+/−;Chrna1^+/+ mouse littermates at E18. A-C and E: maximum intensity projections of whole-mount cochlear preparations labeled with antibodies directed against EGFP (green) and myosin VI (hair cell marker, white) analyzed by confocal microscopy. A and B: basal cochlear turn. Scale bar = 30 μm. C and E: higher magnification of cholinergic fibers in the HC area. D and F: single confocal image of the same area showing bouton like endings (arrowheads) (green) in close vicinity of the IHCs (white) for both genotypes. C–F: scale bar = 10 μm. Mice lacking α₁-expression did not show a marked change in the cochlear efferent innervation at the level of the radial fibers (RF) or in the organ of Corti compared with their control littermates.

Test for a role of α₁-nAChR subunit in setting up efferent HC synapses.

To test the hypothesis that the α₁-nAChR subunit is involved in setting up efferent synapses in cochlear HCs, HC recordings were performed in Chrna1^Δ4/Δ4 knockout mice and control littermates. Fortunately, at E18.5, ACh currents and efferent synaptic events can be recorded in the basal turn of the embryonic cochlea, which is the earliest developed turn of the cochlear coil.

Similar to wild-type and Chrna1^Δ4/+, Chrna1^Δ4/Δ4 IHCs showed electrophysiological properties typical for developing IHCs (Kros et al. 1998; Marcotti et al. 1999, 2003a,b, 2004). In response to a voltage step protocol, IHCs of all genotypes showed small voltage-gated sodium currents and delayed rectifier potassium currents (Fig. 6A) and were capable of firing calcium action potentials (Fig. 6B).

Fig. 6.

Test for a role of the α₁-nAChR subunit in setting up functional nAChRs and efferent synapses in hair cells. A–D: whole cell recordings from basal IHCs in Chrna1^Δ4/Δ4 and control heterozygote littermates (Chrna1^Δ4/+) at E18.5. A: both Chrna1^Δ4/Δ4 and Chrna1^Δ4/+ IHCs showed similar current-voltage relationships dominated by delayed rectifier potassium currents. Voltage step protocol (inset): from a holding potential of −84 mV, 200-ms voltage steps to values between −104 and −36 mV were applied in 10-mV increments. B: IHCs of both genotypes fired calcium action potentials in current clamp. C: in both genotypes, in a subset of IHCs, 1 mM ACh induced an inward current response at a holding potential of −94 mV and an outward current at −34 mV, suggesting that α₉α₁₀-nAChRs were coupled to SK2 channels. D: in both genotypes, in a subset of IHCs, application of 80 mM K⁺ evoked synaptic currents. Together these data indicate that α₁-nAChR expression in IHCs is not necessary for the formation of functional efferent synapses.

Regardless of their genotype, more than 80% of IHCs tested at E18.5 responded to ACh with current amplitudes between 10 and 207 pA at a holding potential of −94 mV (n = 15/18, 20/22, 17/18 for Chrna1^+/+, Chrna1^Δ4/+, and Chrna1^Δ4/Δ4 mice, respectively) (Fig. 6C). In IHCs, calcium influx through the nAChRs activates calcium-dependent SK2 potassium channels that are clustered tightly with nAChRs at efferent postsynapses (Art et al. 1984; Fuchs and Murrow 1992a; Glowatzki and Fuchs 2000; Oliver et al. 2000). The coupling between nAChRs and SK2 channels causes an outward current at depolarized IHC membrane potentials, due to the dominant SK2 component. Therefore, to test if in Chrna1^Δ4/Δ4 mice nAChRs and SK2 channels are coupled, the ACh response was measured at a holding potential of −34 mV. For all genotypes, in a subset of HCs, an outward current was measured in response to ACh application, indicating that SK2 channels were coupled to nAChRs (n = 2/10, 5/11, 10/14 for Chrna1^+/+, Chrna1^Δ4/+, and Chrna1^Δ4/Δ4 mice, respectively; Fig. 6C). Finally, 80 mM extracellular potassium was used to depolarize efferent terminals and induce neurotransmitter release. At a holding potential of −94 mV, synaptic currents were detected in a subset of HCs in all three genotypes (n = 1/15, 4/12, 7/14 for Chrna1^+/+, Chrna1^Δ4/+, and Chrna1^Δ4/Δ4 mice, respectively; Fig. 6D). Synaptic currents from Chrna1^Δ4/Δ4 IHCs were completely and reversibly blocked by 1 μM strychnine (n = 2). The percentage of IHCs with an ACh response with an SK2 component, or with synaptic activity varies substantially between Chrna1^+/+, Chrna1^Δ4/+, and Chrna1^Δ4/Δ4 mice. As recordings took place in the time window in which ACh response and subsequent SK2 coupling to the ACh response just develop, a few hours difference in maturation of the recorded tissue can result in significant differences. Therefore, only the qualitative result that SK2 component and synaptic currents could be found in all experimental groups, was considered.

Together, these results indicate that in the absence of the α₁-nAChR subunit, clusters of nAChRs coupled to SK2 potassium channels as well as functional efferent synapses did form, showing that α₁ is not necessary for these processes to take place.

DISCUSSION

To provide clues for α₁'s highly speculative role in the cochlea, the present study was designed to assess the spatiotemporal expression pattern of α₁ in the cochlea, to test for a functional α₁-containing nAChR in HCs, and to test for a possible role of α₁-nAChR in efferent synapse formation and function. The onset of α₁-expression, ACh response, and efferent synaptic function was found to coincide in early postnatal IHCs. Pharmacological experiments excluded the possibility of a functional “muscle-type” nAChRs besides the α₉-containing nAChRs in IHCs. In α₁-deficient mice, no obvious change of efferent innervation was detected in the organ of Corti at E18. Most importantly, ACh response and efferent synaptic function were present in IHCs of α1-deficient mice, suggesting that α₁ is not necessary for assembly and membrane targeting of nAChRs nor for efferent synapse formation in HCs.

Spatiotemporal expression of α₁-mRNA in cochlear HCs.

For both IHCs and OHCs, the onset of expression of α₁-mRNA coincided with the onset of ACh response and efferent synaptic function. The di-cistronic reporter system used here with an IRES sequence does not allow for direct quantitative analysis of the α₁-expression levels. However, it still shows that the time course of α₁-expression in IHCs and OHCs differs. In apical OHCs, α₁-mRNA appears only transiently, during the onset of synaptic function (P4 to P12), and is downregulated after, below detection threshold, while OHC efferents continue to function throughout life. The OHCs transcriptome studied by GeneChip microarray analysis still shows the presence of α1 transcripts in 25- to 30-day-old CBA/J mice, at a lower level than in IHCs though (Liu et al. 2014). It is therefore unclear whether α1 might be functional in adult OHCs despite the downregulation of its transcription level. Similarly, in muscle, α₁-, β₁-, δ-, and γ-nAChR subunits are strongly expressed in differentiated myotubes before innervation, but both their mRNAs and proteins are rapidly and dramatically reduced after innervation (Albuquerque and McIsaac 1969; Evans et al. 1987; Goldman et al. 1985; Klarsfeld and Changeux 1985; Merlie et al. 1984; Merlie and Kornhauser 1989; Moss et al. 1987; Shieh et al. 1987; for reviews, see Duclert and Changeux 1995; Hall and Sanes 1993). However, although at a dramatically reduced expression level, α₁ clearly is functional in the adult muscle.

As observed in OHCs, α₁-expression in apical IHCs is upregulated during the onset of synaptic function, at P0-P4. Shortly after hearing onset, at 2–3 wk postnatally, IHCs lose their efferent innervation (Katz et al. 2004; Marcotti et al. 2004; Shnerson et al. 1981). After this denervation, α₁-expression persists for at least 6 mo and is still detectable at 18 mo in some IHCs. Similarly, α₉-expression is also maintained in IHCs after loss of efferent function (Luo et al. 1998; Morley and Simmons 2002; Simmons and Morley 1998; 2011; Zuo et al. 1999). In contrast, the expression of α₁₀ as well as of SK2 channels is correlated with the presence of IHC efferent innervation and disappears after hearing onset (Dulon et al. 1998; Katz et al. 2004; Morley and Simmons 2002; Simmons and Morley 2011).

The cochlear features tested here regarding efferent synapse formation did not show any phenotype in absence of α₁-nAChRs and therefore these experiments did not resolve the question why IHCs and OHCs might have different α₁-expression patterns. A highly speculative idea is that α₁- and α₉-mRNA expression in IHCs stays high after IHCs lose their innervation, to return to a state that provides the ability to newly form efferent contacts in the adult. In fact, in IHCs of aged C57BL/6J mice, efferent endings and functional efferent synaptic activity have been found to reappear (Lauer et al. 2012; Zachary and Fuchs 2015) and in guinea pig IHCs, efferent endings have been found to reappear after AMPA-induced excitotoxicity (Ruel et al. 2007).

Is α₁ part of a functional nAChR in HCs?

Besides the expression of α₁, the mRNAs of α₂-, α₃-, α₄-, α₅-, α₆-, α₇-, β₁-, β₂-, β₄-, γ-, and δ-nAChRs subunits have been found in HCs (Cai et al. 2015; Scheffer et al. 2007, 2015; Shen et al. 2015) suggesting that a fetal muscle-like nAChR, (α₁)₂β₁γδ, or another combination of α₁-containing nAChRs may exist in postnatal HCs. However, agonists of the muscle nAChR including (±)-anatoxin A, (±)-epibatidine, DMPP, or (−)-nicotine did not activate any response in IHCs, arguing against the presence of a functional “muscle-type” nAChR in addition to α₉α₁₀. Alternatively, α₁ and γ could be associated with other subunits, possibly including α₉ and/or α₁₀ to form a new type of nAChR. However, such a hypothetical receptor does not include the α₁γ-interface as found in muscle, as the specific α₁γ-interface blocker αA-OIVA (Teichert et al. 2008; Teichert et al. 2004) did not affect the ACh response in HCs.

As of now, no additional pharmacological tools are available for testing if α₁ is part of a new type of nAChR in HCs. The coexpression of α₁ with α₉ or α₉α₁₀ in oocytes did not show any significant difference compared with α₉ or α₉α₁₀ expression alone, or compared with the HC response, regarding EC₅₀, kinetics or calcium permeability (Elgoyhen et al. 2001; Gomez-Casati et al. 2005; Scheffer et al. 2007; Sgard et al. 2002; Verbitsky et al. 2000). However, such coexpression experiments did not include other subunits, such as γ, δ, β₁, β₂, or β₄, possibly necessary for forming α₁-containing channels.

Alternatively, it is possible that the expression of α₁-transcript does not lead to the presence of any functional protein at the membrane. This, in principle can be tested by fluorescent α-bungarotoxin labeling or antibody labeling. However, α-bungarotoxin recognizes both α₁ and α₉α₁₀ (Anderson and Cohen 1977; Elgoyhen et al. 1994, 2001; Lee et al. 1967) and to the best of the author's knowledge, no antibody is available that specifically recognizes α₁ in cochlear tissue. Several antibodies that show specificity in other tissues than cochlea have been tested here and found to be nonspecific in cochlear tissue (see materials and methods). Due to these methodological deficits and the limited amount of biological material offered by HCs, coimmunoprecipitation and protein microsequencing approaches to identify nAChR subunit composition have yet to be conducted. Therefore, the study here could not test if α₁-nAChR protein is expressed in IHC membranes.

One might argue that shared transcription factors important for the activation of other nAChRs may have produced “leaky” expression of Chrna1 and that the nature of expression might therefore be spurious and produce functionally irrelevant α₁-nAChR mRNA in IHCs. However, the mRNA expression levels of α₁-nAChRs, during the first postnatal week are of the same order of magnitude as those of α₉ and α₁₀ (Cai et al. 2015; Scheffer et al. 2015; Shen et al. 2015).

Interestingly, the transcription factor Atoh1/Math1 (Bermingham et al. 1999) that is both necessary and sufficient for HC development and differentiation (Kawamoto et al. 2003; Woods et al. 2004; Zheng and Gao 2000) is one of the transcription factors directly regulating the expression of the genes coding for α₁ and α₁₀ in HCs (Cai et al. 2015; Scheffer et al. 2007). As Atoh1 expression in HCs is transient (downregulated after birth) and the expression of α₁- and α₁₀-mRNAs persists in mature IHCs and OHCs, respectively, it is likely that Atoh1 is only one of the transcription factors involved in the expression of these subunits (Groves et al. 2013). As found for the genes coding for α₁ and α₁₀, candidate consensus Atoh1 binding sites (extended E-box-containing sequence termed AtEAM) are also present within 5 kb of the transcriptional start sites of the genes that code for α₉-, β₂-, and β₄-nAChR subunits, all highly expressed in HCs at P1 (Cai et al. 2015; Klisch et al. 2011), suggesting that their expression may also be in part regulated by Atoh1.

Is α₁ involved in efferent synapse formation?

Apart from a possible role of α₁ in a functional membrane bound receptor, α₁ could also act in a signaling pathway, or as a chaperone for α₉α₁₀ assembly, for example, making it more efficient. Alternatively, α₁ could be part of the scaffolding machinery involved in efferent synapse formation. Indeed, RIC-3, rapsyn, and MuSK, key molecules essential for assembly, clustering, and localization of nAChRs at the neuromuscular junction, have also been detected in HCs (Osman et al. 2008), suggesting that similar molecular mechanisms for synapse formation may apply in both systems, including the involvement of α₁ and γ (Sanes and Lichtman 2001; Wu et al. 2010).

However, in α₁-knockout mice, no obvious change in the IHC efferent innervation pattern was detected at E18 and positive responses to ACh and functional efferent synapses were found, including the coupling to SK2 channels, suggesting that α₁ is not necessary for the expression of functional nAChRs, for their transport to the membrane, for their coupling to SK2, nor for efferent synapse formation to occur. In contrast, in muscle, lack of α₁ leads to a complete lack of spontaneous miniature and nerve-evoked endplate potentials as shown both in mice (An et al. 2010) and in zebrafish (ortholog nic-1) (Westerfield et al. 1990). Without α₁, the mouse diaphragm muscle shows hyperinnervation, with an increased number of motoneurons and extensive nerve branching (An et al. 2010). In HCs, however, α₉ and α₁ could possibly compensate for each other and prevent a change in innervation. In mice lacking α₉-expression (Vetter et al. 1999), no efferent synaptic activity was detected in the OHCs (Vetter et al. 2007); however, OHC efferent innervation is maintained over several weeks with a reduced number and slightly hypertrophied morphology of efferent endings (Murthy et al. 2009a,b). Similarly, possible changes in innervation should be tested in the OHCs region in α₁-conditional knockout mice. One possibility is that α₁ contributes to maintaining efferent innervation in the absence of α₉. Such hypotheses could be tested in the future on α₁/α₉-conditional double knockout mice.

GRANTS

This work was supported by National Institutes of Health (NIH) Grants R01-DC-006476 and R01-DC-012957 (to E. Glowatzki); NOHR Foundation 2010 Seed Research Award, Hearing Health Foundation 2012 Emerging Research Grant, and NIH Grant R03-DC-013374 (to I. Roux); NIH Grants GM-103801 and GM-48677 (to J. M. McIntosh); NIH Grant P30-DK-089502 to the Hopkins Basic Research Digestive Disease Development Core Center and Ross Confocal Facility; and NIH Grants P30-DC-005211 to the Center for Hearing and Balance and NS-050274 to the Department of Neuroscience Multiphoton Imaging Core.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

I.R., J.S.W., J.M.M., and E.G. conception and design of research; I.R. and J.S.W. performed experiments; I.R. and J.S.W. analyzed data; I.R., J.S.W., J.M.M., and E.G. interpreted results of experiments; I.R. and E.G. prepared figures; I.R. and E.G. drafted manuscript; I.R., J.S.W., J.M.M., and E.G. edited and revised manuscript; I.R., J.S.W., J.M.M., and E.G. approved final version of manuscript.