Connexins 30 and 43 Expression Changes in Relation to Age-Related Hearing Loss

Abstract

Age-related hearing loss (ARHL), also known as presbycusis, is the number one communication disorder for aging adults. Connexin proteins are essential for intercellular communication throughout the human body, including the cochlea. Mutations in connexin genes have been linked to human syndromic and nonsyndromic deafness; thus, we hypothesize that changes in connexin gene and protein expression with age are involved in the etiology of ARHL. Here, connexin gene and protein expression changes for CBA/CaJ mice at different ages were examined, and correlations were analyzed between the changes in expression levels and functional hearing measures, such as ABRs and DPOAEs. Moreover, we investigated potential treatment options for ARHL. Results showed significant downregulation of Cx30 and Cx43 gene expression and significant correlations between the degree of hearing loss and the changes in gene expression for both genes. Moreover, dose-dependent treatments utilizing cochlear cell lines showed that aldosterone hormone therapy significantly increased Cx expression. In vivo mouse treatments with aldosterone also showed protective effects on connexin expression in aging mice. Based on these functionally relevant findings, next steps can include more investigations of the mechanisms related to connexin family gap junction protein expression changes during ARHL; and expand knowledge of clinically-relevant treatment options by knowing what specific members of the Cx family and related inter-cellular proteins should be targeted therapeutically.

Introduction

Age-related hearing loss (ARHL, presbycusis), is the most prevalent age-linked neurodegenerative and communications disorders affecting over one-third of the world’s population, including about 40 million in the US [1]. ARHL is a complex sensory disorder displaying reduced sensitivity to sound starting at high frequencies, speech perception deficits - particularly in noisy backgrounds, and slowed central processing of acoustic feature information [2, 3]. Extrinsic and intrinsic factors including genetics, aging, drugs, and loud noise contribute to the development of presbycusis [4, 5]. While there have been advances in the understanding of the etiologies of presbycusis, interventions are still limited to hearing aids and counseling, since there are no FDA-approved drugs to prevent or treat this predominant communication disorder [6].

Three types of changes in the cochlea with age include: sensory, neural, and metabolic presbycusis [7, 8]. Sensory presbycusis involves the loss of hair cells; neural presbycusis consists of the decline of the number of spiral ganglion neurons (SGNs); and metabolic presbycusis is the degeneration of the stria vascularis (SV). The latter also causes declines of the endocochlear potential (EP) [9]. Cells in the organ of Corti (OC) and SV are extensively coupled by gap junctions, which are crucial in maintaining ionic balances and hearing function. These gap junctions provide intercellular communication; thus, helping regulate and move ions, electrolytes, secondary messengers, and metabolites between cochlear cells [10, 11]. Gap junctions are also hypothesized to be involved in regulating the EP via K⁺ recycling [12].

Connexins (Cxs) are gap junction proteins that join to form gap junction ion channels and have been found in various types of cochlear tissue such as the SV, OC, supporting cells and SGNs [13–16]. Mutations in genes encoding Cx26 are responsible for up to 50% of all nonsyndromic sensorineural hearing loss, specifically the 35delG mutation [17, 18]. Cx30 and Cx43, which are encoded by gap junction genes beta-6 and alpha-1 (GJB6 and GJA1, respectively) are two of the seven connexin proteins detected thus far in the cochlea whose mutations cause hearing loss or deafness [19, 20]. We hypothesize that connexins are associated with ARHL since mutations in these proteins are known to cause hearing loss and connexin knockout mice display an absence of the EP, increases in reactive oxygen species (ROS), and cell degeneration [21, 22]. Moreover, it has previously been reported that Cx26 and Cx30 expression decreased with age in the C57BL/5J mouse cochela [23]. Additionally, aging in other areas of the body – including the retina, heart, bladder, and bones – alters connexin expression [24–28].

Aldosterone, a steroid hormone that helps regulate sodium (Na+) and potassium (K+) ion levels, has been shown to have protective effects against ARHL [6, 29–32]. These pre-clinical studies have shown that aldosterone is effective in slowing down the progression of ARHL, lowering/improving hearing thresholds, and preventing apoptosis in spiral ganglion neurons. Correlations between aldosterone and connexin expression changes have also been observed. [33] For example, Suzuki et al. (2009) noticed an upregulation of Cx43 mRNA gene expression after rat ventricular myocytes were treated with aldosterone. In light of these previous findings, we tested the hypothesis that aldosterone is beneficial for the aging cochlea in maintaining Cx30 and Cx43 expression and improving hearing thresholds. The present study provides novel insights into cochlear connexin protein expression changes with age and new treatment actions of aldosterone in delaying or preventing key aspects of ARHL.

Methods

Animal Model

CBA/CaJ mice were used because they lose their hearing progressively with age; therefore, making them an excellent model for human ARHL. The CBA/CaJ mouse breeding pairs were obtained from Jackson Laboratories (Bar Harbor, ME) and bred in-house following institutional protocols. All procedures were approved by the University of Rochester and University of South Florida Vivariums and IACUC, and were compliant with NIH policies. A total of 81 mice were used for different sets of experiments: the Affymetrix GeneChip analysis, aldosterone study, quantitative polymerase chain reaction (qPCR), and confocal microscopy. Affymetrix GeneChip

For the Affymetrix GeneChip analysis, 40 of the mice were classified into four groups according to age and degree of hearing loss [34–36]. The groups were divided as follows: young adult with good hearing (YA, N=8, 3.5±0.4 months), middle-aged (MA, N=17, 12.3±1.3 months),old with mild/moderate presbycusis (MP, N=9, 27.7±3.4 months), and severe presbycusis (SP, N=6, 30.6±1.9 months). The YA group was used as the baseline group for the gene expression analysis (i.e., for calculating fold changes). The other 20 mice used for the age comparison for the RT-qPCR portion of the present study were split into two groups: young adult (Y, N=10, 3±1.5 months) and old (O, N=10, 28±2.5 months) [34, 35].

Aldosterone Animals

Aldosterone treatment was conducted as described in Frisina et al. (2016). Eighteen CBA/CaJ mice (15–18 months at baseline) were assigned randomly to control (n=10) or treated groups (n=8). The treated mice received a custom d-aldosterone pellet (0.0016 mg/day, 120-day release, Innovative Research of America, Sarasota, FL) and the control group received a saline placebo pellet. The pellets were implanted subcutaneously behind the shoulders while the mice were anesthetized with ketamine/xylazine (120 and 10mg/kg). A baseline ABR recording was done prior to receiving the pellets, and two subsequent recordings were done at 2 and 4 months. At the conclusion of the treatment period, the mice were sacrificed by decapitation and their cochleae were dissected and stored in −80°C for future experiments. A subset of these 3 treated cochleae, were collected for RT-qPCR and confocal microscopy experiments done in this study to be compared with control, young (n=3, 2±2 months) and old (n=3, 23±1.5 months) mice.

RT-qPCR

Young (n=10, 2±2 months) and old (n=12, 27±3months) SV and OC cochlear tissue from CBA/CaJ mice were obtained for RT-qPCR aging experiments.

Hearing Assessments

Auditory Brainstem Responses (ABR) and ABR Gap-In-Noise (GIN)

CBA/CaJ mice were anaesthetized with a mixture of ketamine/xylazine (120 and 10mg/kg body weight, respectively, via an intraperitoneal injection) before auditory examinations and kept on a heating pad inside a soundproof acoustic chamber (IAC lined with Sonex) to maintain their body temperature (37°C) throughout the experiments. Acoustic stimuli were synthesized digitally using the System III Tucker-Davis Technology (TDT, Alachua, FL) signal-processing platform. The stimuli were then attenuated and filtered (low-pass cutoff at 5kHz). Stimulus sounds were presented through an electrostatic speaker (TDT EC1) connected to the external ear canal by 4cm tubes, which is via a calibrated, closed system. A ¼ inch B&K microphone (Type 4938, Bruel & Kjaer, Naerum, Denmark) attached to a .1cm³ coupler was used to calibrate the TDT system daily. Three platinum needle electrodes were subcutaneously inserted at the vertex (non-inverted and in the mastoid area muscle of the ipsilateral (testing) side (inverted), with a ground electrode being inserted into the muscle posterior of the contralateral pinna to record the ABR responses. These electrodes were connected to a bioamp head stage (HS4 Fiber Optic, TDT). ABR waveforms were evoked with 5 msec tone pips (0.5- msec rise-fall times) with a cos2 onset envelope, delivered at 21/sec though electrostatic speakers (TDT EC1) connected by 4 cm tubes to the opening of the external ear canal. The response was amplified (10,000 X), filtered (100 Hz–3 kHz), and averaged using the BioSig (TDT, Gainesville, FL) data acquisition system. A total of 200 responses were averaged (with stimulus polarity alternated), using an ‘artifact reject’ algorithm, whereby response waveforms were discarded when peak-to-peak amplitude exceeded 7 μV, to prevent contamination by muscle activity. Intensity was varied in 5 dB steps starting at 80 dB and decreasing to at least 20 dB below threshold for a specific test frequency. Each intensity was replicated, and threshold was defined as the lowest intensity at which a response was replicated. [6] For ABR threshold experiments, the mice were presented with tone pips of 3, 6, 12, 16, 20, 24, 32, and 48 kHz. A wide band noise (WBN) stimulus, having a bandwidth from 0 to 48kHz, was also used. The duration for each ABR stimulus was 5ms, presented at a rate of 21 bursts/second [37].

Distortion-Product Otoacoustic Emissions (DPOAEs)

CBA/CaJ mice were placed in the same conditions as ABR experiments: anaesthetized with a mixture of ketamine/xylazine (120 and 10mg/kg body weight, respectively, via an intraperitoneal injection) before electrophysiology experiments and placed inside a soundproof booth (IAC lined with Sonex) on a heating pad to maintain body temperature. Before recording, the stimulus probe and microphone coupler were placed in the test ear near the tympanic membrane with the aid of an operating stereoscope. Ipsilateral acoustic stimulation and simultaneous measurement of DPOAEs were accomplished with a TDT BioSig System III. Stimuli were digitally synthesized at 200kHz using SigGen software with the ratio of f2/f1 constant at 1.25, and L1 = 65 dB and L2 = 50 dB SPL, as calibrated in a 0.1cm³ coupler simulating the mouse ear canal. After synthesis, f1 and f2 were passed through an RP2.1 D/A converter to PA5 programmable attenuators. Following attenuation, the signals went to ED1 speaker drivers which fed into the EC1 electrostatic loudspeakers coupled to the ear canal via short flexible tubes with rigid plastic tapering tips. For DPOAE measurements, the resulting ear canal sound pressure was recorded with an ER10B+ low noise microphone (gain 20x) and probe Etymotic, Elk Grove Village, IL) housed in the same coupler as the f1 and f2 speakers. The output of the ER10B+ amplifier was inputted to an MA3 microphone amplifier, the output of which went to an RP2.1 A/D converter for sampling at 200kHz. A fast Fourier transform (FFT) was performed on the resultant waveform. The magnitude of f1, f2, the 2f1-f2 distortion product, and the noise floor of the frequency bins surrounding the 2f1-f2 component were measured from the FFT. The procedure was repeated for geometric mean frequencies ranging from 5.6–44.8kHz (8 frequencies/octave) to assess adequately the neuroethological functional range of mouse hearing. Duration of the testing was approximately one hour per animal [38].

GeneChip Data Access

The entire microarray probe-set from each GeneChip M430A (Affymetrix Inc., Santa Clara, CA) and individual CBA mouse phenotypic (hearing measures) data have been submitted to the GEO-NCBI database and have been approved with the following Series reference #: GSE GSE49543. These GeneChip data can be accessed via, http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE49543. The probe-sets used were derived from the above GEO-NCBI accession number GSE49543. All steps were conducted according to the MIAME (Minimum Information About a Microarray Experiment) checklist [36].

Cell Culture and Treatments

SV-k1 epithelial and HEI-OC1 cells were cultured in permissive conditions [(33°C, 10% CO₂, Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal calf serum (FCS)]. At least 24hr before treatment, the cells transferred to grow in the non-permissive condition (39°C, 5% CO₂, DMEM with 5% FCS) were split into 100mm plates (Falcon). Experiments were performed at 70–80% confluence. All treatment started after the switch from permissive to non-permissive condition. The cells were split into 5 groups for each treatment, with each group receiving different dosages of the treatment. The aldosterone treatment group was divided into: Control, 1nM, 10nM, 1μM, 5μM. The dosages chosen were determined from previous studies [39–41]. After 24hr treatment, the adherent cells were scraped off the plates using a cold plastic cell scraper, then gently transferred into an Eppendorf tube. The collected cells were stored in −80°C until future use [37, 42].

Real-time Quantitative RT-PCR

This procedure has been previously described by Ding et al. (2018). SV tissue samples extracted from the whole cochlea of mice from the aging and aldosterone studies, and the cells collected were used for the following experiment. RNA from mouse samples and cells were extracted using the RNAeasy Mini Kit (Qiagen, Valencia, CA). Samples were vortexed for 1 minute to shear genomic DNA before loading onto the RNeasy columns, and then eluted in a minimum of 30mL and a maximum volume of 2×50 mL RNAsefree water. RNA obtained with the procedure was essentially free of genomic DNA. 50ng of RNA was then used to synthesize 20uL of complimentary DNA (cDNA) using an iScript cDNA kit (Bio-Rad Laboratories; Hercules, CA). Once the sample mixtures were made, they were incubated for 2 min at 94°C (initial denaturation), 15 s at 94°C (denaturation), 30s at 55°C (annealing), 1 min at 68°C (extension), and 5 min at 68°C (final elongation). The denaturation, annealing, and extension steps were repeated for 35 cycles. Primer sequences used to detect the genes were as follows: Cx30 Forward: 5’-GAAGTGTGGGGTGATGAGCAGGAG-3’, Cx30 Reverse: 5’-CGTGGACTGCTTCATTTCGAGGCC-3’, Cx43 Forward: 5’-CTCCTCCTGGGTACAAGCTG-3’, and Cx43 Reverse: 5’-GTTCGATTTTGCTCTGCGCT-3’.

Primer specificity was checked using melting curves as previously described in Ding et al. (2018). Triplet repeated quantitative PCR (qPCR) experiments were executed by creating a master mix using the aforementioned primers, SsoFast^™ Evagreen (Bio-Rad, Hercules, CA), RNase-free water and the cDNA samples. The samples were then placed in a Analytik Jena qTOWER³ (Thuringia, Germany) to generate a quantitative analysis of the genes expressed in both the cells and in vivo tissue samples. Each experiment always included a no-sample negative control (NTC) to ensure no contamination and that the primers were working. A parallel PCR reaction was performed using glyceraldehyde 3-phosphate dehydrogenase (GAPDH)-specific primer as the housekeeping gene. Calculations of gene expression were performed with the 2ΔΔCT method [37, 42, 43].

Western Blot

SV-K1 cell lysates were prepared in radioimmunoprecipitation assay buffer (RIPA, Pierce #89901, Thermo Scientific, Waltham, MA) with protease inhibitor cocktail (#78430, Thermo Scientific, Rockford, IL). Cells were homogenized in buffer, followed by centrifugation at 13,000 rpm for 10 min at 4°C. Supernatants were subjected to western blot analysis by loading 30ug protein per lane, after the protein concentrations were determined utilizing the Bradford protein assay. Proteins were fractionated by SDS-PAGE gel electrophoresis and transferred to a PVDF blotting membrane overnight at 30mV. The membrane was then washed 3 times with phosphate-buffered saline with Tween 20 (PBS-T) for 15 min each wash and blocked using 5% non-fat milk for 1 hr. The blot was then washed PBS-T 3×15 min and then incubated with primary antibodies overnight at 4°C. The membrane was washed with PBS-T 3×15min and incubated with the secondary antibody for 2 hrs. After a final 3×15min with PBS-T, the membrane was placed in 20X LumiGLO Reagent and 20X Peroxide (#7003, Cell Signaling, Danvers, MA) for 1 min before imaging.

The primary antibodies used included: Cx30 (1:250, #71–2200, Invitrogen), Cx43 (1:250, ab11370, Abcam), β-actin (1:1000, Cell Signaling, Danvers, MA), and GAPDH (1:1000, ab9485, Abcam). The secondary antibody was horseradish peroxidase-conjugated goat anti-rabbit IgG (1:2000; Cell Signaling) [42].

Immunohistochemistry

Preparation of Samples

Six cochleae collected from our aldosterone study (Frisina et al. 2016), control old, n=3, 22–24 months; aldosterone-treated, n=3, 22–24 months) and three young adult mice cochlea (3–4 months) were stored at −80°C prior to this experiment. Prior to being stored, the cochleae were placed in 10mL of fresh 4% paraformaldehyde (Thermo Scientific, Rockford, IL) in PBS (0.1M, pH 7.6) at 4°C overnight. After fixation, the samples were washed 3×10min in PBS and then decalcified in 20mL 10% EDTA (ethylenediaminetetraacetic acid, Fisher Scientific, Pittsburgh, PA) in PBS at 4°C. The decalcification took approximately 1 week to complete; however, the cochleae were checked daily. Following this, the cochleae were washed 3×10min in PBS, and then transferred into 10% and 20% sucrose (Acros, Geel, Belgium) in PBS for 2 hours. The samples were then placed in 30% sucrose in PBS overnight at 4°C. Lastly, the cochleae were embedded into degassed OCT (Tissue-Tek, Torrance, CA) at 4°C, oriented into a cryomold (Tissue-Tek) with degassed OCT for 1 hour, and frozen at −80°C. Cryosectioning was done at 5μm/section and then mounted on glass slides.

Cx30 and Cx43 Staining

Slides were rinsed with 1xPBS + .1% Triton X-100 3×5 minutes. Next the tissues were placed in a damp black box and incubated in blocking solution (10% goat serum + 0.1% Triton X-100 + 1% BSA in PBS) for 1hr. If mouse-on-mouse staining was going to be done, a subsequent block was done utilizing anti-mouse-IgG blocking reagent (Vector Laboratories, cat. No. MKB-2213). Afterwards, the samples were incubated with the primary antibodies at 4°C overnight. The next day, the samples were washed 3×15 min with PBS and incubated with the secondary antibodies for 1hr. The secondary antibodies were washed out with DI H₂O, 3×10 min). The samples were then mounted onto slides with mounting medium (Prolong^™ Gold antifade reagent with DAPI, 1:1000) for 2–5 minutes and observed using a confocal laser scanning microscope (Nikon Corporation, Tokyo, Japan). The primary antibodies used were Cx30 anti-rabbit (1:100, #712200, Invitrogen) and Cx43 anti-mouse (1:100, #13–8300. The secondary antibodies used were Alexa Fluor 488 goat anti-rabbit IgG and Alexa Fluor 594 goat anti-mouse IgG (1:1000, ThermoFisher Scientific, Waltham, MA).

Laser Scanning Confocal Imaging

Confocal imaging was done utilizing a Nikon A1 HD25 Multi-Photon & Laser Scanning Confocal combined with a Nikon Eclipse Ti2 inverted microscope. Images were taken at 20X via excitation at 405, 488, and 594 nm wavelength for DAPI, Cx30, and Cx43 at a scan speed of 0.5 frames/second. Z-stack images were taken at 0.5μM steps at digital zoom levels of 4.21 and 8.96. Each stack was exported into individual images used for fluorescence intensity analysis and cell counting of the apex, middle, and basal regions for each animal’s cochlea.

Statistical Analysis

GeneChip Expression Analysis

The data from the GeneChip microarray were imported into Microsoft Excel showing the raw gene expression level values. The gene expressions for each sample were converted into signal log ratio (SLR) to determine the difference in expression of the studied gene in that sample from the mean expression of that gene in all samples from the young adult mice. A signal log ratio of 1.0 indicates an increase of the transcript level by 2-fold and −1.0 indicates a decrease by 2-fold. A signal log ratio of zero indicates no change.

For the six connexin GeneChip probes, the ROUT method (Q=1%) was applied to find any outliers within the dataset. A total of 4 and 6 outliers were discovered for Cx30 and Cx43, respectively. Then the raw gene expression was converted into SLR. One-way Analysis of Variance (ANOVA) (95% confidence limit) was used to compare the signal log ratio values of the different subject groups. In addition, fold changes of all samples were calculated from signal log ratios using the following equations:

$$\begin{gathered} Fold\hspace{0.17em}Change\hspace{0.17em}=\hspace{0.17em}{2}^{signal\hspace{0.17em}\log \hspace{0.17em}ratio}if\hspace{0.17em}signal\hspace{0.17em}\log \hspace{0.17em}ratio\hspace{0.17em}\ge \hspace{0.17em}0 \\ =\hspace{0.17em}or\hspace{0.17em}\left(-1\right)x{2}^{\wedge }\hspace{0.17em}\left(\left(-signal\hspace{0.17em}\log \hspace{0.17em}ratio\right)\right)\hspace{0.17em}signal\hspace{0.17em}\log \hspace{0.17em}ratio\hspace{0.17em}<0 \end{gathered}$$

Linear regression comparing the changes in gene expression and hearing ability measures was also conducted. These statistical tests were conducted using GraphPad Prism 6.0 (GraphPad Software, La Jolla, CA).

Real Time qPCR Analysis

The threshold cycle (${\mathrm{C}}_T$) values were measured to detect the threshold of each of four significant genes of interest and GAPDH gene in all samples. Each sample was measured in triplicate and normalized to the reference control GAPDH gene expression. The ${\mathrm{C}}_T$ value of each well was determined and the average of the three wells of each sample was calculated. For samples that showed no expression of the test gene, the value of minimum expression was used for statistical analysis.

$\mathrm{Delta}{\hspace{0.17em}}_{\text{CT}}\left(\Delta {\text{C}}_{\text{T}}\right)$ for test gene of each sample was calculated using the equation:

$$\Delta C_T\hspace{0.17em}=\hspace{0.17em}{C}_{T}TestGene\hspace{0.17em}-\hspace{0.17em}{C}_{T}GAPDH$$

$\mathrm{Delta}\hspace{0.17em}\text{delta}\hspace{0.17em}{\text{C}}_{\text{T}}(\Delta \Delta {\text{C}}_{\text{T})}$ was calculated using the following equation:

$$\Delta \Delta C_T\hspace{0.17em}=\hspace{0.17em}\Delta C_{T}SampleAverage\hspace{0.17em}-\hspace{0.17em}\Delta C_{T}YoungGroup$$

Lastly, the fold change in the test gene expression was calculated from the formula:

$$Fold\hspace{0.17em}Change\hspace{0.17em}=\hspace{0.17em}{2}^{\left(-\Delta \Delta C_T\right)}if\hspace{0.17em}\Delta \Delta C_T\hspace{0.17em}\le \hspace{0.17em}0\hspace{0.17em}or\hspace{0.17em}\frac{-1}{2^{\left(-\Delta \Delta C_T\right)}}\hspace{0.17em}if\hspace{0.17em}\Delta \Delta C_T\hspace{0.17em}<0$$

A statistical evaluation of real time PCR results was performed using a one-way ANOVA followed by Bonferroni post hoc tests (correction for multiple comparisons) to compare between the gene expression values for the young adult, middle age, old age mild presbycusis, and old age severe presbycusis groups from the GeneChip (N=40 mice), along with comparing the young adult and old animals from the PCR aging study (N=20 mice). For the cell treatment experiments, two-tailed, independent t-tests were utilized to compare the difference in gene expression values between the different dosages used.

For the two significantly different genes on GeneChip and/or qPCR, linear regression analyses were employed to find correlations between the signal log ratio values or fold changes, and the functional hearing measurements (ABRs, DPOAEs). These measurements were the distortion product otoacoustic emission (DPOAE) amplitudes at low frequencies (5.6 kHz to 14.5 kHz), mid frequencies (15.8 kHz to 29.0 kHz) and high frequencies (31.6 kHz to 44.8 kHz), in addition to auditory brainstem response thresholds (ABR) at 3, 6, 12, 24, 32 and 48 kHz. Microsoft Excel and GraphPad Prism 9.0 software (GraphPad, La Jolla, CA) were used to do the conversions and calculate the one-way ANOVA and the linear regression statistics [36].

Densitometric Analysis

Images from Analytic Jena UVP Chem Studio SA2 were imported to NIH ImageJ software for the densitometry analysis and the data are reported as mean±SD. Using ImageJ, the image was saved in a TIFF format and processed using the Process > Subtract Background command to reduce background noise. The “rectangle” tool from ImageJ was used to frame the largest band for each row. For each protein band across the lanes, each single region of interest was defined using the same frame as the largest band. With the same frame as the protein (row) a background measurement was taken. This procedure was done for the other rows or loading controls and the measurements for the bands and their backgrounds were exported to Excel. The pixel density for all data (bands/controls + their backgrounds) were inverted and put in new columns. The inverted values are expressed as 255 – X, where X is the value recorded by ImageJ. For the protein bands and loading controls, the net values were expressed by deducting the inverted background from the inverted band value. When the net bands and loading controls were calculated as the final step, a ratio of a net band value over the net loading control of that lane was calculated. The final relative quantification values are the ratio of the net band to the net loading control. Differences were analyzed with a 1-way or 2-way ANOVA as appropriate, with an ANOVA followed by Bonferroni post hoc tests (corrected for multiple comparisons), with p<0.05 considered significant, using GraphPad Prism 9.0. [39, 42].

Confocal Analysis

On NIS-Elements Advanced Research (Nikon Instruments Inc., Melville, NY), each image was imported to analyze the fluorescence intensity of each protein and count the cells. First a region of interest (ROI) was selected for each region of the cochlea - SV, OC, SGN. Each region’s ROI size was determined by ensuring that critical regions were included, i.e., for the OC, hair cells must be in the ROI. The sizes of the ROIs were consistent in size for all three turns. Cell counting was done utilizing the software’s automated measurement counter identifying DAPI, which labels DNA within the nucleus of each cell. Parameters for the counting were adjusted for the intensity ranges of the 405 (blue) channel, separation, and size to ensure accurate cell counting. The fluorescence intensity was achieved following the Ding method. [42] by subtracting out the background of the image in the DAPI channel and NIS-Elements Advanced Research software quantified the mean fluorescence intensity for each wavelength 405 [(blue), 488 (green), and 594 (red)] in the ROI. The intensity averages of each animal were added up within their group and analyzed for significant differences between each group utilizing GraphPad Prism 9.0. The statistical analysis included a one-way ANOVA followed by Bonferroni post hoc tests, with p>0.05 considered significant.

Results

GeneChip Analysis Shows Cx30 and Cx43 Gene Expressions Decline with Age

CBA/CaJ mice (N=40) had their hearing tested with ABRs and DPOAEs and classified into groups based on age and degree of hearing loss. The four resulting groups included: Young Adult (YA) with good hearing (N=8, 4 males and 4 females, age: 3.5±0.4 months), Middle-Aged (MA) with good hearing (N=17, 8 males and 9 females, age: 12.3±1.3 months), old with Mild/Moderate Presbycusis (MP, N=9, 4 males and 5 females, age: 27.7±3.4 months), and old with Severe Presbycusis (SP, N=6, 2 males and 4 females, age: 30.6±1.9 months) [34–36]. Auditory data demonstrating age-related hearing loss from these animal groups was previously published by Tadros et al. [34]

Twelve connexin (Cx) gene probes were identified in our Affymetrix M430A GeneChip array. The gene expression levels were first tested for outliers using the ROUT method (Q=1%) using GraphPad Prism. The remaining gene expression levels for each sample were converted into fold change and signal log ratio (SLR) to determine the difference in expression of the YA mice compared to the other groups. Comparisons of the fold changes and SLRs between the MA, MP, and SP groups revealed statistically significant fold changes in Cx30 (probe set ID: 1448397_at) and Cx43 (probe set ID: 1415800_at) gene expression levels (Figure 1). The average fold change values for these genes along with the one-way ANOVA results are summarized in Table 1. These gene expression changes for the two Cx genes were then subjected to further validation with RT-qPCR.

Figure 1.

Cx30 and Cx43 gene expression decreases with age and hearing loss. Fold changes of (A) Cx30 and (B) Cx43 gene expression displayed significant downregulation in the GeneChip data collected from the cochlea of mild/moderate presbycusis, and severe presbycusis groups compared to the young adult control group. *p<0.05, ****p<0.0001.

Table 1.

ANOVA Results of Fold Change Values

	Fold Change Correlation
Hearing Measures	Cx30			Cx43
Gene Name	Average Fold Change			ANOVA
	Middle Age	Mild Presbycusis	Severe Presbycusis
Cx30	0.1370	−1.066	−1.014	P=<0.0001, F(3,32) = 17.34
Cx43	−1.506	−1.743	−1.811	P=<0.0001, F(3,30) =34.65, df = 3,30

Decreases of Cx30 and Cx43 gene expressions with age in the cochlea are related to ABR, but not DPOAE changes.

Linear regression tests were used to analyze if there were any correlations between the auditory ability of the animals and the gene expression levels for Cx30 and Cx43. Significant correlations were found between all hearing measures and fold change values for Cx30 as seen in Table 2. There were slight correlations found for Cx30 in relation to the ABR thresholds at all frequencies versus the SLR values; however, there were near significant correlations (for 48kHz), see Figures 2A, C. No significant correlations were found between Cx43 SLR and ABR thresholds (Figure 2B). There were no statistically significant correlations between Cx30 and Cx43 gene expression and DPOAE amplitudes (Figure 3). It is important to note the binary nature of the Cx30 gene expression results in Figure 2, due to the data points within the groups having little to no variability. For this reason, further in vitro and in vivo experiments using multiple regression assays would be required to verify these findings.

Table 2.

Correlations between Gene Expression and Auditory Physiology Measurements

ABR 3kHz	*P=0.0371, r2=0.1567, F=4.830	P=0.0756, r2=0.1257, F=3.449
ABR 6kHz	*P=0.0229, r2=0.1836, F=5.847	P=0.1168, r2=0.0993, F=2.647
ABR 12kHz	**P=0.0090, r2=0.2343, F=7.958	P=0.3634, r2=0.0345, F=0.8585
ABR 24kHz	**P=0.0083, r2=0.2390, F=8.166	P=0.3862, r2=0.0314, F=0.7790
ABR 32kHz	**P=0.0048, r2=0.2677, F=9.505	P=0.6600, r2=0.0081, F=0.1984
ABR 48kHz	**P=0.0046, r2=0.2700, F=9.616	P=0.4691, r2=0.0191, F= 0.4691
DPOAE Low Freq.	*P=0.0360, r2=0.1583, F=4.891	P=0.3529, r2=0.0360, F=0.8975
DPOAE Mid Freq.	*P=0.0285, r2=0.1715, F=5.382	P=0.4177, r2=0.0275, F=0.6800
DPOAE High Freq.	*P=0.0971, r2=0.1023, F=2.962	P=0.4920, r2=0.0198, F=0.4870

^*p<0.05;

^**p<0.01.

Figure 2.

Correlations found between ABR thresholds and SLR for Cx30 and Cx43. Representative examples of the correlation between ABR thresholds with signal log ratios changes for Cx30 and Cx43 gene expression. (A-B) are examples of the correlations between ABR thresholds and Cx30 gene expression changes). (C-D) shows correlations between Cx43 gene expression changes and ABR test results. The solid line in the graphs indicate the line of best fit and the dotted lines represent the 95% confidence intervals.

Figure 3.

DPOAE amplitude correlated with fold changes. DPOAE amplitudes at low frequencies and mid frequencies showed significant correlations for Cx30 (Left column), but no significant correlations were found at high frequencies. Although trends were observed, there were no significant correlations seen between Cx43 fold changes and DPOAE amplitudes (Right column).

Cx30 and Cx43 Gene and Protein Expression Decreases with Age

Two groups of CBA/CaJ mice were grouped by age and hearing ability. The young adults (N=5, 3±1.5 months) and old (N=5, 28.2±2.5 months) groups underwent ABRs and DPOAE recordings to measure ARHL changes. The audiological data collected are shown in Figure 4. These hearing tests demonstrated a significant increase in thresholds and decrease in DPOAE amplitudes for old mice relative to the young adults.

Figure 4.

ARHL was seen between young adult and old subject groups. Young adult and old hearing comparisons for (A) ABR thresholds, (B) DPOAE amplitudes, and (C) DPOAE thresholds for frequencies ranging from 3–53.6 kHz. All hearing tests demonstrated a significant increase in ABR and DPOAE thresholds for the old animals compared to the young adults. The DPOAE amplitudes are also significantly decreased; thus, validating that age-related hearing loss is occurring. The values for Old NF and Young NF in Panels B and C for the DPOAE analysis are nearly identical; thus, the error bars overlap. ****p<0.0001

The validation of the GeneChip results were positive for Cx30: gene expression in the OC and SV of the cochlea was significantly downregulated with age. Cx43 PCR results were also consistent with the GeneChip results. Figure 5A shows the quantitative results comparing the fold changes of the young adult and old animals with the fold changes obtained from the GeneChip. From the GeneChip, the old animal group consisted of combined data from the mild presbycusis and severe presbycusis subjects. RT-qPCR experiments were done after separating the SV and OC. The results from these separate experiments show a significant downregulation in the SV for both, Cx30 (t(10)=3.578, p=0.005, unpaired t-test) and Cx43 (t(10)=2.592, p=0.0269, unpaired t-test), and no significant changes in the OC for both gene expressions (Figure 5B). Overall, there were no differences between GeneChip and qPCR results as both showed a general trend of downregulation in the cochlea. In addition, the downregulation of Cx30 and Cx43 gene expressions with age in whole cochlea is related with their decreasing in SV.

Figure 5.

Gene expression fold changes from RT-qPCR and GeneChip displayed changes for both Cx genes. A) Fold changes obtained from real-time PCR and GeneChip for Cx30 and Cx43 in the cochlea showed downregulation with age and hearing loss. B) Results from qPCR triplicate experiments showed a significant fold change difference for Cx30 and Cx43 between young adult and old animals in the stria vascularis (left column). There were no significant changes found with age in the organ of Corti for both genes (right column). *p<0.05, **p<0.01

Cells Treated with Aldosterone Modulate Cx Expression

After treatment with aldosterone for 24 hours, SV-k1 cells were collected and analyzed in triplicate RT-qPCR experiments. Figure 6 shows the quantitative results from these cell line treatment experiments. The gene expression of Cx30 increased in the SV-k1 cell line following aldosterone treatment at these doses: 1μM, 5μM, and 10μM (Figure 6A). Cx43 showed very little change in expression for aldosterone treatments (Figure 6B). Utilizing the same experimental conditions as the RT-qPCR cell treatment conditions, SV-k1 cells were treated and collected to analyze the protein expression changes of Cx30 and Cx43. The following triplicate experiment results mirror those obtained for the RT-qPCR findings. The densitometry analysis showed that Cx30 protein levels increased after being treated with aldosterone (Figure 6C, E). The findings affirm that aldosterone exerts an enhancing influence on the expression of the Cx30 protein, not potentially linked to the regulation of its gene expression. Additionally, the application of aldosterone resulted in an increased expression of Cx43 protein (Figure 6D, F).

Figure 6.

Aldosterone treatments upregulate Cx gene and protein expression in the stria vascularis cell line. Utilizing densitometry, gene expression of Cx30 increased in the SV-k1 cell line following 24-hr treatment with (A) aldosterone at multiple doses; but remained the same for Cx43 after being treated with (B) aldosterone. (C and E) shows a significant upregulation of Cx30 protein expression following the treatment with aldosterone. (D and F) also revealed an upregulation of Cx43 after the aldosterone combination treatment. E and F show examples of western blots from individual experiments, and were part of to the group densitometry data (C and D).

Aldosterone Treatment Provides Protective Effects against ARHL

Previous study indicated that aldosterone treatment provides a therapeutic effect on auditory brainstem responses for aging mice (Frisina 2016).

A quantitative analysis of the gene expression of Cx30 and Cx43 was performed utilizing cochlear tissue samples collected from these CBA/CaJ mice after the four-month-long treatment period [6]. Figure 7 shows that the samples from the aldosterone-treated mice did not display any marked Cx30 and Cx43 gene expression changes in the aldosterone-treated mice compared to the control animals. Images taken using confocal laser scanning microscopy also show the colocalization and changes of Cx30 and Cx43 protein expression between young adult and old mice. Figures 8A and 8B show images representative of these findings while Figure 8C provides the quantified results showing the differences of fluorescence intensity with the antibody labeling. The downregulation of both Cx proteins was found to be statistically significant in the SV in the apex of the cochlea (Figure 8C–i) and the OC of the middle and basal turns (Figure 8C–ii) and. Differences in the changes between gene expression and protein expression, especially in the OC, is hypothesized to be caused by post-transcriptional modifications, as protein expression changes are not companied by gene expression changes (Figure 7 vs. Figure 8).

Figure 7.

In vivo aldosterone treatment does not cause significant changes in Cx43 and Cx30 gene expression. Triplicate qPCR experiments show some trends but resulted in no significant fold change differences in Cx30 and Cx43 between control and aldosterone-treated animals.

Figure 8.

Cx protein expression increased in the mouse cochlea following long-term aldosterone treatment while ARHL is observed between young and old mice. The expression of Cx30 (green) and Cx43 (red) visibly decreased between young adult and old animals, meanwhile aldosterone treatment maintained the expression of Cx43 comparable to that of young adult animals but increased the expression of Cx30 compared to old animals. Panels A-C illustrate the difference in Cx30 and Cx43 between young adult and old animals in different locations in the cochlea: A) SV, B) OC, and C) SGN. Panel D) highlights the quantitative results from the confocal analysis. A significant increase of Cx30 was seen in aldosterone-treated animals compared to old animals of the same age in the (i) SV, (ii) OC, and (iii) SGN. Aldosterone-treated animals also showed higher Cx43 protein levels compared to old mice of the same age (iv-vi). Normal, age-related cell loss was observed between young adult and old mice. Although a trend for aldosterone protecting cells in the old cochlea is apparent in vii-ix, the number of cells in aldosterone-treated old mice were about the same as that found in the old comparison mice (G-I). *p<0.05, **p<0.01

At the same time, enhanced Cx30 protein expression effects by aldosterone were seen in the cochleae of mice that were treated with aldosterone (22–24 months) vs. non-treatment old mice. These mice exhibited significantly higher Cx30 protein expression in SV, OC and SGNs compared to both young adult and old mice in all three turns of the cochlea under aldosterone treatment. The expression of Cx43 was not changed by aldosterone treatment (Figure 8 iv, v, vi). also higher compared to young adult and old mice, except in the OC. In the OC, Cx43 levels in the aldosterone-treated mice were only higher than the levels of the old mice. Cells were also counted, and we found that there was loss of cells between young adult and old animals, as expected with age. Fewer cells were also counted in the aldosterone-treated mice compared to the young adult animals; and we observed slightly more cells than the number found in untreated old animals. For this reason, the analysis of protein expression takes into account this loss of cells.

Discussion

This study provides insights into the possible role of connexins in the progression of ARHL. We discovered that expression of Cx30 and Cx43 decreases with age in CBA/CaJ mice, but aldosterone provided protective effects against these declines. ABR thresholds were also preserved in aldosterone-treated mice compared to controls, old mice who were still exhibiting ARHL. These results suggest that changes in connexin protein expression, affecting cochlear gap junction structure and functionality, contribute to the development of ARHL, but aldosterone can be used as a therapeutic agent to prevent some of these changes.

It is well known that mutations in genes for connexin gap junction proteins are one of the main causes of congenital deafness. Based on that, we hypothesized these proteins have a role in presbycusis which has not been investigated systematically, until now. The present study demonstrated elevated thresholds and reduced amplitudes in older animals compared to the young adult mice; which are key characteristics of mammalian presbycusis. Significant downregulation of connexin genes and proteins with age was also observed. These findings are consistent with previous literature where changes in connexin expression for other age-related diseases and disorders have been reported [44–47]. For example, it was found that Cx43 expression decreased with age in rat aortic endothelium, bladder, and bone-marrow [44–46]. Moreover, Tajima et al. showed that Cx30, along with Cx26, significantly decreased in C57BL/6J mice at 32 weeks compared with mice at 4 weeks of age [23]. It has also been seen that total deletion of Cx30 in a transgenic mouse model can intensify ARHL, including elevating inflammation biomarkers and cochlear damage [47]. In our investigation, we observed a decline in the gene and protein expressions of Cx30 and Cx43 in the cochlea with age. However, in our in vitro study using cochlear cell lines SV-K1 and HEI-OC1, aldosterone treatments revealed protein expression changes that did not correlate with gene expression changes. This suggests a potential mechanism involving post-translational modifications, and hints at a potential therapeutic approach using aldosterone to counteract the age-related downregulation of Cx30 and Cx43. Further research is necessary to understand further details of these mechanisms.

Since connexin proteins are involved in other age-related pathologies, they are appealing therapeutic targets to prevent or even reverse key aspects of these disorders. A potential component of a treatment targeting connexin to alleviate some effects of ARHL is aldosterone therapy. Aldosterone, a steroid hormone, is naturally produced in the human body, making the administration of it for humans safe with the proper dose controls. Previously published articles have shown that aldosterone is a good candidate for treating ARHL because it lowers physiological and behavioral hearing thresholds, prevents apoptosis in spiral ganglion neurons, and can slow down the progression of ARHL and some of its biomarker changes [6, 29–31]. In the present study, we found that mice who were given aldosterone for a 4-month treatment period had significantly higher levels of Cx30 in all regions and turns of the cochlea compared to old, comparison animals. Moreover, protein expression of Cx43 was also higher in aldosterone-treated mice relative to these old, untreated mice. The abundance of evidence seen in past reports and the present study suggests that some of the protein changes seen with aldosterone therapy are post-translational, since concomitant gene expression effects were not observed.

Previous literature has reported downregulation of mineralocorticoid receptors being reversed following aldosterone treatment; moreover, downregulation of Na⁺-K⁺−2Cl⁻ cotransporter protein (NKCC1) was inhibited via PTM ubiquitination [24, 49]. Connexins have been shown to be regulated by PTMs such as phosphorylation, ubiquitination, and methylation [50, 51]. PTMs, including methylation, phosphyrlation, hydroxylation, and acetylation, of Cx26 in sites of deafness-causing mutations have been identified utilizing mass spectrometry [51]. To the best of our knowledge, there are no studies that have explored aldosterone’s effect on connexin expression via PTMs. Further studies are required to investigate this relationship. Another potential mechanism involves the protein kinase C (PKC) and/or mitogen-activated protein kinase (MAPK) pathway. Multiple studies have shown that aldosterone activates the PKC and MAPK pathways where connexins, specifically Cx43, have been observed being phosphorylated in those pathways [33, 52, 53]. It will be important to consider these findings from previous reports when designing new studies of the auditory system and conducting further experiments observing the therapeutic effects of aldosterone on connexin expression in the cochlea and central auditory system.

In other physiological systems, previous research has shown that aldosterone modulates the expression of connexins, specifically Cx43 [33]. For example, Suzuki et al. (2009) treated rat ventricular myocytes with aldosterone and noticed an upregulation in Cx43 mRNA gene expression. Administration of aldosterone to Sprague-Dawley rats also showed an increase in Cx43 protein expression [54]. However, our RT-qPCR experiments showed that gene expression for Cx30 and Cx43 in mice did not change significantly; thus, we hypothesize that aldosterone triggers post-translational modifications (PTM) in Cx proteins.

A limitation of the current study is the small sample size of n=3 per group in the protein expression analysis of the aldosterone-treated, old, and young mice. A larger sample size in future experiments will allow us to gain further understanding of the impact of aldosterone treatment for ARHL. Furthermore, another limitation is in the cell studies where ALD treatment was within 24 hours. A longer treatment period or higher doses of ALD could induce a greater accumulation effect on Cx30 and Cx43 expression; thus, providing a greater protective effect against ARHL. Further experiments are currently underway to investigate the effects of these scenarios.

Summary and Conclusions

Our findings indicate that Cx30 and Cx43 are mechanisms of ARHL, and they and should be considered when developing therapeutic cocktails or multi-faceted therapy approaches for treating ARHL. The aging mouse model employed here showed that aldosterone can prevent or treat key mechanisms of presbycusis. More research is required to pinpoint which pathways elements are directly involved in the modulation of cochlear connexins expression by aldosterone. So, understanding what cellular pathways are involved in the progression of ARHL is critical when formulating medications and multidisciplinary treatment plans for this highly prevalent neurodegenerative disorder of our elderly population.

Table 3.

One-Way ANOVA Aging Results

Gene Name	ANOVA
	SV	OC	SGN
Cx30	Apex: *P=0.0387 t=2.917 df=12	Apex: P=0.3932 t=1.621 df=12	Apex: P=>0.9999 t=0.5543 df = 11
	Mid: P=0.2035 t=2.007 df=12	Mid: *P=0.0069 t=3.852 df=12	Mid: P=>0.9999 t=0.6665 df=11
	Basal: P=0.0802 t=2.524 df=12	Basal: *P=0.0227, t=3.204, df=12	Basal: P=0.5271 t=1.447 df=11
Cx43	Apex: P=0.7648 t=1.196 df=12	Apex: P=0.9954 t=1.011 df=12	Apex: P=>0.9999 t=0.1978 df=11
	Mid P=0.8035 t=1.162 df=12	Mid: P=>0.9999 t=0.4525 df=12	Mid: P=>0.9999 t=0.2394 df=11
	Basal: P=0.3808 t=1.640 df=12	Basal: P=0.7508 t=1.208 df=12	Basal: P=0.8715 t=1.110 df=11

Table 4.

ANOVA Bonferroni Results from Aldosterone Treatment for Protein Expression Changes

Protein Name		ANOVA
	SV	OC	SGN
Cx30	Apex: **P=0.0064 t=3.582 df=18	Apex: *P=0.0314 t=2.857 df=18	Apex: *P=0.0223, t=3.028 df = 17
	Mid: *P=0.0137 t=3.238 df=18	Mid: *P=0.0310 t=2.864 df=18	Mid: *P=0.0364 t=2.806 df=17
	Basal: **P=0.0024 t=4.017 df=18	Basal: *P=0.0354 t=2.802 df=18	Basal: **P=0.0075 t=3.540 df=17
Cx43	Apex: P=>0.9999 t=0.8679 df=18	Apex: P=>0.9999 t=0.5813 df=18	Apex: P=>0.9999 t=0.3242 df=17
	Mid: P=0.8647 t=1.094 df=18	Mid: P=>0.9999 t=0.0148 df=18	Mid: P=>0.9999 t=0.8566 df=17
	Basal: P=>0.9999 t=0.8167 df=18	Basal: P=>0.9999 t=0.6591 df=18	Basal: P=0.8715 t=1.658 df=17

Highlights.

• Mutations in connexin proteins cause congenital nonsyndromic and syndromic hearing loss

• The roles of connexins in age-related hearing loss (ARHL) are largely unexplored

• We found that Cx30 and Cx43 decrease expressions with age in the CBA/CaJ mouse cochlea

• We discovered that aldosterone treatment provides protective effects for Cx30and Cx43

• These results indicate that connexin protein declines are likely mechanisms of ARHL