Mouse Panx1 Is Dispensable for Hearing Acquisition and Auditory Function

Panx1 forms plasma membrane channels in brain and several other organs, including the inner ear. Biophysical properties, activation mechanisms and modulators of Panx1 channels have been characterized in detail, however the impact of Panx1 on auditory function is unclear due to conflicts in published results. To address this issue, hearing performance and cochlear function of the Panx1−/− mouse strain, the first with a reported global ablation of Panx1, were scrutinized. Male and female homozygous (Panx1−/−), hemizygous (Panx1+/−) and their wild type (WT) siblings (Panx1+/+) were used for this study. Successful ablation of Panx1 was confirmed by RT-PCR and Western immunoblotting in the cochlea and brain of Panx1−/− mice. Furthermore, a previously validated Panx1-selective antibody revealed strong immunoreactivity in WT but not in Panx1−/− cochleae. Hearing sensitivity, outer hair cell-based “cochlear amplifier” and cochlear nerve function, analyzed by auditory brainstem response (ABR) and distortion product otoacoustic emission (DPOAE) recordings, were normal in Panx1+/− and Panx1−/− mice. In addition, we determined that global deletion of Panx1 impacts neither on connexin expression, nor on gap-junction coupling in the developing organ of Corti. Finally, spontaneous intercellular Ca2+ signal (ICS) activity in organotypic cochlear cultures, which is key to postnatal development of the organ of Corti and essential for hearing acquisition, was not affected by Panx1 ablation. Therefore, our results provide strong evidence that, in mice, Panx1 is dispensable for hearing acquisition and auditory function.

Panx1 forms plasma membrane channels in brain and several other organs, including the inner ear. Biophysical properties, activation mechanisms and modulators of Panx1 channels have been characterized in detail, however the impact of Panx1 on auditory function is unclear due to conflicts in published results. To address this issue, hearing performance and cochlear function of the Panx1−/− mouse strain, the first with a reported global ablation of Panx1, were scrutinized. Male and female homozygous (Panx1−/−), hemizygous (Panx1+/−) and their wild type (WT) siblings (Panx1+/+) were used for this study. Successful ablation of Panx1 was confirmed by RT-PCR and Western immunoblotting in the cochlea and brain of Panx1−/− mice. Furthermore, a previously validated Panx1-selective antibody revealed strong immunoreactivity in WT but not in Panx1−/− cochleae. Hearing sensitivity, outer hair cell-based "cochlear amplifier" and cochlear nerve function, analyzed by auditory brainstem response (ABR) and distortion product otoacoustic emission (DPOAE) recordings, were normal in Panx1+/− and Panx1−/− mice. In addition, we determined that global deletion of Panx1 impacts neither on connexin expression, nor on gap-junction coupling in the developing organ of Corti. Finally, spontaneous intercellular Ca 2+ signal (ICS) activity in organotypic cochlear cultures, which is key to postnatal development of the organ of Corti and essential for hearing acquisition, was not affected by Panx1 ablation. Therefore, our results provide strong evidence that, in mice, Panx1 is dispensable for hearing acquisition and auditory function.
Quantification of Panx1 mRNA levels by quantitative real-time polymerase chain reaction (QPCR) in mouse central and peripheral nervous system, and various organs, revealed highest values in trigeminal ganglia > bladder > spleen, followed at distance by hippocampus > cortex ∼ calvaria > heart > cerebellum, with lowest levels in kidney and spleen (Hanstein et al., 2013). Furthermore, Panx1 has been localized in eye (Ray et al., 2005;Kurtenbach et al., 2014), taste buds (Huang et al., 2007), olfactory system (Zhang et al., 2012) and inner ear (Tang et al., 2008;Wang et al., 2009;Zhao, 2016). As for the latter, the impact of Panx1 on auditory function is unclear due to recent publication of conflicting results.
A strain carrying the Panx1 tm1a(KOMP)Wtsi allele was generated by the Knock Out Mouse Project (KOMP 4 ) using the multipurpose tm1a knockout-first promoter-driven selection cassette, which has been adopted also by other major mouse knockout programs such as EUCOMM (EUropean Conditional Mouse Mutagenesis Program 5 ; International Mouse Knockout et al., 2007). The versatile tm1a allele contains an IRES:lacZ trapping cassette and a floxed promoter-driven neo cassette inserted into the intron of a gene, disrupting gene function 6 . However, mice homozygous for the Panx1 tm1a(KOMP)Wtsi knockout-first promoter-driven allele were found to express about 30% residual Panx1 mRNA in all organs examined, leading to the conclusion that Panx1 tm1a(KOMP)Wtsi is a hypomorphic allele (Hanstein et al., 2013). Hypomorphism has been reported also for other knockout-first alleles (Shpargel et al., 2012;Ryder et al., 2014).
In the light of these contrasting results, the goal of the present study was to clarify the role of Panx1 in hearing. To this end, we re-evaluated hearing performance and cochlear function of Panx1−/− mice using in vivo electrophysiology, plus a variety of biochemical and biophysical assays.
Panx1 mice were genotyped according to published protocols by standard PCR on extracted mouse tail tips Bargiotas et al., 2011) using the following primers: • Panx1 f: 5 -GGAAAGTCAACAGAGGTACCC-3 • Panx1 r: 5 -CTTGGCCACGGAGTATGTGTT-3 • LacZ: 5 -GTCCCTCTCACCACTTTTCTTACC-3 The Panx1+/+ allele was targeted by the above f and r primers and identified by a 330 bp band, whereas Panx1−/− was targeted by primers Panx1 f and LacZ, and was identified by a 630 bp band. Panx1+/− was identified by the simultaneous presence of a 330 bp and a 630 bp band.
Body temperature was maintained at 37 • C using a heating pad under feedback control. Corneal drying was prevented by application of ophthalmic gel to the eyes of the animals. For ABR recordings (Scimemi et al., 2014), acoustic stimuli consisted of clicks (100 µs duration) and tone bursts (1 ms rise-fall time with 3 ms plateau) of 4, 8, 16, 24 and 32 kHz, and were delivered in the free field using a MF1-M speaker. Bioelectrical potentials were collected with gauge 27, 13 mm needle electrodes (Cat. No. S83018-R9, Rochester) inserted subdermally at the vertex (active), ventrolateral to the left ear (reference) and above the tail (ground). Potentials were amplified, filtered (0.3-3 kHz) and averaged over 512 presentations of the same stimulus. Hearing threshold levels were determined offline as the SPL at which a Wave II peak, could be visually identified above the noise floor (0.1 µV).
Otoacoustic emissions (Kemp, 1978) were evoked using a pair of equal intensity primary tones, f 1 = 14,544 Hz, and f 2 = 17,440 kHz delivered at intensities ranging from 20 to 80 dB SPL in 10 dB SPL increments. Each primary tone (20.97 ms duration, 47/s) was emitted by a separate MF1-M speaker, configured for closed field stimulation, and delivered to the mouse ear via a small tube as prescribed by the manufacturer. The cubic distortion product 2f 1 -f 2 = 11,648 kHz, was detected using a small microphone (ER10B+ Low Noise Probe and Microphone, Etymotic Research, IL, USA) coupled to the ear canal.

Dye Transfer Assays in Cochlear Organotypic Cultures
To visualize gap junction coupling among non-sensory cells of the lesser epithelial ridge (LER), we performed dye-transfer assays using the fluorescent tracer Lucifer Yellow (LY, CH Lithium Salt, Thermofisher, #L12926) for microinjection. Cochlear cultures were transferred on the stage of a spinning disk confocal microscope (DSU, Olympus) and perfused for 5 min at 1 ml/min with EXM, an extracellular solution containing (in mM): NaCl 135, KCl 5.8, CaCl2 1.3, NaH 2 PO 4 0.7, MgCl 2 0.9, Hepes−NaOH 10, d−glucose 6, pyruvate 2, amino acids and vitamins (pH 7.48, 307 mOsm). For dye delivery, patch pipettes were fabricated from glass capillaries (G85150T-4, Harvard Apparatus, Edenbridge, UK) using a double stage vertical puller (PP-830, Narishige) and were filled with LY dissolved at 220 µM (final concentration) in a 320 mOsm intracellular solution containing (in mM): KCl 134, NaCl 4, MgCl 2 1, HEPES 20, EGTA 10 (adjusted to pH 7.3 with FIGURE 6 | Analysis of the 2f 1 − f 2 cubic distortion product. (A) Representative spectra from WT (Panx1+/+, blue) and Panx1−/− mice (red) for a pair of 80 dB sound pressure level (SPL) primary frequencies (f 1 and f 2 ); the inset shows a magnified view of the spectrum in the region of the 2f 1 − f 2 cubic distortion product. (B) Growth function of the 2f 1 − f 2 cubic distortion product; error bars represent SEM; p-values were determined by two-tail t-test. KOH) and filtered through 0.22 µm pores (Millipore). Pipette resistances were 3-4 MOhm when immersed in the bath. One cell (donor) was patch clamped and maintained in the cell-attach configuration for a few seconds to establish a baseline. The patch of membrane under the pipette sealed to the donor cell was subsequently ruptured, allowing the LY to fill the cell, while leaving the seal intact (whole cell recording conditions). The cell was held at its zero current level using the current clamp configuration of the patch clamp amplifier (Axopatch 200B, Molecular devices). LY diffusion among (first order) cells adjacent to the injected cell was monitored over time by acquiring fluorescence images at a rate of 1 frame per second with a typical exposure time of 70 ms. Fluorescence images were displayed as (F − F bck )/(F max − F bck ), where F max is the maximal value reached in the injected cell at the end of the recordin interval and F bck is autofluorescence. Experiments were performed at room temperature (22-25 • C).

Multiphoton Microscopy and Ca 2+ Imaging in Cochlear Organotypic Cultures
To record spontaneous intercellular Ca 2+ signal (ICS) activity in nonsensory cells of the mouse cochlea, organotypic cultures of sensory epithelium were incubated for 45 min at 37 • C in DMEM/F12 supplemented with the acetoxymethyl (AM) ester of the selective Ca 2+ sensor Fluo-Forte (16 µM, Enzo Life Science, #ENZ-52014). The incubation medium contains also pluronic F-127 (0.1% w/v, Sigma-Aldrich, #P2443) and sulfinpyrazone (250 µM, Sigma-Aldrich, #S9509) to prevent dye sequestration and secretion. Cultures were then transferred to an upright microscope stage (see below) and continually perfused with EXM (see above) for 15 min at 1 ml/min in a dark environment at 25 • C to allow for dye de-esterification. All subsequent imaging experiments were also performed at 25 • C.
To record Ca 2+ signals, we used a two-photon microscope (Bergamo II, Thorlabs) equipped with a resonant scanner and coupled with a mode-locked Ti:Sapphire pulsed laser (Chameleon-Ultra II, Coherent). Fluo-Forte was excited at 940 nm by focusing the Ti:Sapphire beam onto the sample through a water-immersion objective (XLPlan N, 25× 1.05 NA, Olympus). Average power at the sample was ∼20 mW. The fluorescence signal, collected by the same objective, was reflected towards the detection arm of the microscope by the 705 nm primary dichroic mirror of the microscope (Semrock, FF705-Di01), placed at 45 • above the objective. After traversing a 680 nm short pass filter (FF01-680/SP-25, Semrock) and 495 nm dichroic beam-splitter (T495lpxru, Chroma Technology), the Fluo-Forte signal was selected in the range 435-485 nm by a single band-pass filter (ET460/50m-2p, Chroma Technology) placed in front of a non-descanned GaAsP detector (H7422-50, Hamamatsu). Mechanical ultra-fast shutters were used to limit light exposure to the bear minimum required for image acquisition.
Sequences of 512 × 512 pixels frames were acquired, averaged in lots of nine and presented at a final rate of 5 per second. Illumination intensity, frame average, frame rate and the number of pixels in each frame were adjusted so as to minimize photobleaching and phototoxicity, while achieving sufficient signal to noise ratio and temporal resolution. Image sequences were acquired using ThorImage LS 3.1 software (Thorlabs). Ca 2+ signals were quantified as pixel-by-pixel relative changes of fluorescence emission intensity, i.e., ∆F(t)/F 0 where t is time, F(t) is fluorescence at time t, F 0 is the fluorescence at the onset of the recording and ∆F(t) = F(t) − F 0 . All data were processed off-line and presented using Vimmaging (F. Mammano and C. Ciubutaru, VIMM, Padova, Italy), a custom-made software routine developed under MATLAB TM environment (The MathWorks Inc., Natick, MA, USA).

RESULTS
Panx1 Is Absent in the Cochlea of Panx1−/− Mice Generation and genotyping of Panx1−/− mice were previously described Bargiotas et al., 2011). Here we report an additional data set based on Panx1 expression analyses by RT-PCT and Western immunoblotting. Panx1 mRNA transcript expression was detected in the cochlea and brain of WT mice at both P5 and adult stage, but not in Panx1−/− mice (Figure 1). Consistent with these results, Western blots failed to reveal Panx1 expression in the cochlea of Panx1−/− mice at both P5 and in the adult stage, whereas bands with the correct molecular weight (∼48 kDa) were present in Panx1+/+ (i.e., WT) extracts (Figure 2).
Using a previously validated antibody that targets an extracellular epitope of the Panx1 protein (AvesLab #6358; Hanstein et al., 2013), we detected strong immunoreactivity in epithelial cells lining the endolymphatic surface of the sensory epithelium (inner sulcus, outer sulcus), supporting and epithelial cells of the organ of Corti, Reissner's membrane, and spiral ganglion neurons of WT mice. Weak immunostaining was also detected in the spiral limbus and spiral ligament of these mice, whereas the AvesLab #6358 antibody failed to label cochlear tissue from Panx1−/− mice (Figure 3). Altogether these results confirm successful ablation of Panx1 in the (brain and) cochlea of Panx1−/− mice.

ABRs and DPOAEs in Panx1−/− Mice Are Indistinguishable from WT Controls
Next, we sought to corroborate and extend the results obtained by Anselmi et al. (2008) by analyzing in greater detail the hearing performance of Panx1−/− mice. In humans and mice alike, sound-evoked ABR potentials appear as a series of consecutive relative maxima (peaks), termed Waves and labeled with Roman numerals, which arise from the synchronous short-latency synaptic activity of successive nuclei along the peripheral afferent auditory neural pathway (Zheng et al., 1999;Legatt, 2002;Zhou et al., 2006). The first peak (Wave I) arises from the cochlea and/or compound action potential of the auditory nerve ∼1 ms after the stimulus onset (latency). Waves from II to V originate from cochlear nuclei, contralateral superior olivary complex, lateral lemniscus and contralateral lateral inferior colliculus. For reference, Table I of Scimemi et al. (2014) presents means and standard deviations (SD) of latency and amplitude values of ABR peaks I-V for C57BL/6 mice.
Hearing threshold estimates from click and pure-tone ABR analysis, as well as latency and amplitude of Wave I and Wave II in Panx1−/− mice aged between P30 and P90, were indistinguishable from those of FIGURE 9 | (A) Lucifer yellow fluorescence emission averaged over the cell body of (first order) cells (gray solid lines, n = 5) adjacent to the injected cell and normalized to the maximal fluorescence emission detected in the injected cell (black solid lines); data are mean values ± SEM (dot lines) for n = 8 cells in each condition. (B) For each experiment, the interpolating line of the curve related to first order cells (with computed slope m 2 ) and to the injected cell (with computed slope m 1 ) was computed over the first 10 s of recording. Histograms show mean values of the ratio between m 2 and m 1 for WT (Panx+/+, dashed bars) and Panx1−/− mice (filled bars). Error bars represent SEM.
Sound generated within the mammalian inner ear as a reflection of outer hair cell (OHC) mechanical activity (Nobili and Mammano, 1996;Nobili et al., 2003) can be detected with a sensitive microphone placed in the auditory meatus (Kemp, 1978(Kemp, , 2002. Therefore, as a further non-invasive indicator of cochlear function, we measured the cubic (2f 1 − f 2 ) DPOAE (see ''Materials and Methods'' section) and found no significant differences in the DPOAE growth function of Panx1−/− mice and age-matched WT controls (Figure 6).
Altogether, these results indicate absence of detectable defects in auditory function of Panx1−/− mice. This conclusion is based on their normal hearing sensitivity, normal function of the outer hair cell-based ''cochlear amplifier'' (Frolenkov et al., 1998;Nobili et al., 1998;Ashmore, 2008) and absence of cochlear nerve defects.

Connexin Expression and Function Is
Normal in Panx1−/− Mice Pannexins bear significant sequence homology with the invertebrate gap junction proteins, innexins, and more distant similarities in their membrane topologies and pharmacological sensitivities with the gap junction proteins, connexins (Sosinsky et al., 2011).
Non-sensory cells of the mammalian cochlea express two closely related gap junction proteins, connexin 26 (Cx26) and connexin 30 (Cx30; Lautermann et al., 1998;Ahmad et al., 2003;Forge et al., 2003;Zhao et al., 2006), the expression of which is coordinately regulated . Mouse models indicate that altered expression levels of these connexins in the early postnatal days impacts on organ of Corti development and hair cell maturation (Johnson et al., 2017), preventing normal hearing acquisition (Cohen-Salmon et al., 2002;Teubner et al., 2003;Ahmad et al., 2007;Sun et al., 2009;Crispino et al., 2011Crispino et al., , 2017Qu et al., 2012;Zhu et al., 2013). As regulatory mechanism may potentially be shared between connexins and pannexins, we examined the expression of Cx26 and Cx30 by Western blot analysis and QPCR, and found no significant differences between Panx1−/− mice and age-matched WT controls (Figure 7). The spatial distribution of Cx26 at P5 was investigated also by immunofluorescence and, again, no differences between Panx1−/− mice and age-matched WT controls were detected (Figure 8).
To assess whether the expressed connexins confer cell-tocell connectivity, we quantified dye transfer in the LER of organotypic cochlear cultures from P5 mice (see ''Materials and Methods'' section). To gauge transfer rate, we measured the slope of the LY fluorescence growth function at the onset of dye delivery in the donor cell (m 1 ) and in its nearest neighbors (m 2 ), and found no significant differences in the m 2 /m 1 ratio of Panx1−/− mice and WT controls (Figure 9).
Altogether the results presented in Figures 7-9 indicate that lack of Panx1 in Panx1−/− mice impacts neither on connexins expression, nor on gap-junction coupling in the developing organ of Corti.
Here, we used multiphoton microscopy to monitor spontaneous ICS activity (Tritsch et al., 2007;Schütz et al., 2010;Rodriguez et al., 2012;Wang and Bergles, 2015;Johnson et al., 2017;Mammano and Bortolozzi, 2017) in organotypic cochlear cultures from P5 mice loaded with the Ca 2+ indicator Fluo Forte AM (Mammano and Bortolozzi, 2017). Specifically, we examined the frequency of occurrence of spontaneous Ca 2+ transients (events) in the apical cochlear turn by counting all occurrences within the portion of the GER in the field view from P5 Panx1−/− mice and age-matched WT siblings (Figure 10). We found a similar mean frequencies of occurrence (15.5 ± 6.0 events/min in Panx1+/+ vs. 15.0 ± 3.4 events/min in Panx1−/− cultures). Likewise, amplitude and inter-peak interval distributions of spontaneous Ca 2+ transients of Panx1−/− cultures overlapped with those of Panx1+/+ cultures. Altogether, these results indicate that spontaneous ICS activity in the GER of the postnatal cochlea is not affected by Panx1 ablation.

DISCUSSION
The Panx1−/− strain we have analyzed was the first with a reported global ablation of Panx1 Bargiotas et al., 2011). The present results confirm successful ablation of Panx1 in Panx1−/− mice, while our ABR and DPOAE data indicate normal hearing sensitivity, normal function of the outer hair cell-based ''cochlear amplifier'' (Frolenkov et al., 1998;Nobili et al., 1998;Ashmore, 2008) and absence of cochlear nerve defects, in agreement with the initial observation that Panx1−/− mice do not exhibit a detectable hearing phenotype (Anselmi et al., 2008 8 ).
We also confirmed that lack of Panx1 affects neither the expression of inner ear connexins nor gap junction communication in the organ of Corti. Furthermore, our experiments with cochlear organotypic cultures indicate that the ATP-release mechanism underlying the spontaneous ICS activity of cells in the GER is intact. Preservation of this mechanism is essential for hearing acquisition (Schütz et al., 2010;Rodriguez et al., 2012;Mammano and Bortolozzi, 2017) and maturation of sensory hair cells (Johnson et al., 2017). Recent results with a monoclonal antibody that inhibits Cx26 hemichannels substantiate the notion that, in the cochlear sensory epithelium, ATP is released from such hemichannels (Xu et al., 2017).
Ensuring that an allele derived from the tm1a cassette is a full null, rather that a hypomorph such as the Panx1 tm1a(KOMP)Wtsi strain, and alleviating potential off-target gene mis-regulation, requires modification of tm1a, which can be performed in embryonic stem (ES) cells or in crosses with transgenic Flp and Cre mice. Flp deletion converts tm1a to a conditional allele (tm1c), restoring gene activity, whereas the promoterdriven selection cassette and floxed exon of the tm1a allele can be deleted by Cre to generate a lacZ-tagged allele (tm1b; Skarnes et al., 2011). This is usually accomplished by breeding the mice to a source of Cre expressed in the germline, followed by outcrossing and selection of knockout offspring that fail to carry the Cre driver (Skarnes et al., 2011). This was also the strategy followed by the International Mouse Phenotyping Consortium (IMPC 9 Brown and Moore, 2012) to convert the tm1a allele to tm1b for subsequent phenotyping (Ryder et al., 2014). The Panx1 tm1b(KOMP)Wtsi strain generated in this way was analyzed by the IMPC and reported to have no significant hearing/vestibular phenotype 10 .
Thus, altogether, our results and conclusions are consistent with hearing assessments both in the Genentech-Panx1−/− strain (Abitbol et al., 2016) and in the Panx1 tm1b(KOMP)Wtsi strain generated and analyzed by the IMPC.
The lack of any measurable auditory phenotype in three strains, which are all global knockouts of Panx1, is in stark contrast with the phenotype reported for Pax2-cPanx1−/− , and even more so for Foxg1-cPanx1−/− mice that, despite descending from the hypomorphic Panx1 tm1a(KOMP)Wtsi strain, retained strong immunoreactivity for the Panx1 #4515 chicken anti-human antibody (which is not specific: Bargiotas et al., 2011;see ''Introduction'' section) in the organ of Corti and the spiral limbus . However, it should be considered that both the Pax2-cPanx1−/− and the Foxg1-cPanx1−/− strain must necessarily express a Cre recombinase (Friedel et al., 2011), whereas the three global knockout strains mentioned above do not express Cre. It is well known that Cre expression in mammalian cells can induce chromosomal aberrations and toxicity that is dependent on the level of Cre activity (Loonstra et al., 2001). Indeed, Caspase-3 activation and cell degeneration are reported hallmarks of the organ of Corti in Pax2-cPanx1−/− mice . Furthermore, Foxg1 Cre mice that are homozygous for the targeted mutation die perinatally (Tian et al., 2012 11 ), whereas heterozygous Gfi1 Cre mice, one of the many driver lines used for conditional cell-specific gene deletion/reporter gene activation in the inner ear (Cox et al., 2012), were recently reported with an early onset progressive hearing loss, which was absent in their wild-type littermates (Matern et al., 2017).
Another risk factor that is worth considering is related to the structure of the tm1a selection cassette (Testa et al., 2004;Skarnes et al., 2011) used to generate the hypomorphic Panx1 tm1a(KOMP)Wtsi mice. As the latter were not crossed with a Flp deleter line before being used to create Pax2-cPanx1−/− and Foxg1-cPanx1−/− mice, it is not clear from the data provided if the Cre-mediated deletion removed only exon 2 of Panx1 or also the neo cassette, nor from which tissue the sample tested was obtained.
It is well known that retention of a neo cassette can cause unexpected phenotypes in ''knockout'' mice due to neighborhood effects (Pham et al., 1996;Scacheri et al., 2001;Ren et al., 2002;Meier et al., 2010). Indeed, removal of the neo cassette and critical exon from the tm1a allele is regarded as an essential procedure that alleviates potential off-target gene mis-regulation caused by the neo promoter, and to ensure that the allele is a full null rather then a hypomorph (Skarnes et al., 2011). 10 www.mousephenotype.org/data/genes/MGI:1860055#section-associations 11 http://jaxmice.jax.org/strain/004337.html In this vein, it was recently argued that the hearing loss phenotype exhibited by Cx30−/− mice (Teubner et al., 2003) depends on the cumulative effect of deletion of Cx30 and 3' insertion of a lacZ and neo cassette. Indeed, in a strictly related knockout mouse model (Cx30∆/∆) in which Cx30 was removed without perturbing the surrounding sequences, hearing thresholds determined by ABR analysis are normal (Boulay et al., 2013;Crispino et al., 2017).
In conclusion, our extended characterization of Panx1−/− mice provides strong evidence that Panx1 is dispensable for hearing acquisition and auditory function.

DATA AND CODE AVAILABILITY
Data and computer code used to analyze data are available from the authors upon request.

AUTHOR CONTRIBUTIONS
FaM designed the studies, provided resources to conduct the studies and wrote the manuscript; HM provided Panx1−/− mice and genotyping protocols; VZ, MP and FC performed animal genotyping; MR and FS were in charge of animal welfare and performed quality controls; VZ and FP performed in vivo electrophysiology; AC wrote software code to filter ABR waveforms; CDC wrote image acquisition and analysis software; VZ, FP and CN analyzed ABR and DPOAE data; VZ performed immunofluorescence studies; FP performed Western blot analyses; VZ and GZ generated organotypic cochlear cultures; GZ performed patch clamp, dye transfer in cochlear organotypic cultures and analyzed data; CP and FlM performed multiphoton microscopy and Ca 2+ imaging in cochlear organotypic cultures, and analyzed data; AMS, ARF and FaM supervised the work of junior colleagues; VZ, GC, FC, FS, ARF and AMS edited the text.

ACKNOWLEDGMENTS
The AvesLab #6358 antibody used in this study was a gift of Prof. Eliana Scemes (Dominick P. Purpura Department of Neuroscience, Kennedy Center, Albert Einstein College of Medicine, Bronx, NY, USA). The authors thank E. Perlas of EMBL-Rome Histology Facility and G. Bolasco of EMBL-Rome Microscopy Facility for assistance with histology and microscopy; I. Losso of EMMA-INFRAFRONTIER Monterotondo for technical assistance.