Quantitative Proteomics of Cochlear Tissues: Bilateral Comparisons in Guinea Pigs and Rats

ABSTRACT

The cochlea, an incredibly sensitive sensory system, detects sound waves and converts them into electrical signals the brain recognizes as sound. Damage to cochlear hair cells can release proteins, triggering biological responses that may impair hearing. Mass spectrometry‐based proteomics offers insights into protein expression changes in cochlear tissues, improving our understanding of inner ear diseases. In this study, we performed a comprehensive proteomics analysis of whole cochlear tissue extracted from healthy guinea pigs and rats. The study optimized protein extraction protocols and analyzed cochlear protein expression using three biological replicates for each animal model. The results included the identification of 1841 proteins in guinea pigs and 3423 proteins in rats, with a high overlap in cochlear protein expression between the left and right ears—93% in guinea pigs and 89% in rats. The findings validate the assumption that the cochlear tissues from both sides of the ears can be considered biologically equivalent. This experiment provides a comprehensive cochlear proteome for guinea pigs and rats, supporting future studies on inner ear disorders.

1.

Protein analysis, particularly proteomics, is a promising method to aid the diagnosis of inner ear disorders, with the ability to provide insights into the molecular mechanisms underlying these conditions. By investigating the protein profiles within the healthy and diseased inner ear, researchers can identify potential biomarkers related to inner ear disorders [1]. Inner ear proteomic analysis has predominantly focused on determining protein levels within the main inner ear fluid, called perilymph [2, 3]. However, isolating endolymph and perilymph is challenging due to their limited and variable availability, making it difficult to obtain consistent and sufficient samples for analysis. Whole cochlear tissue analysis helps overcome this limitation by providing a more stable and abundant source of protein [4]. Tissue proteome analysis plays an important role in clinical research, as the initiation and progression of diseases are often found at the tissue level in many organs [5]. Few investigations have been conducted on whole cochlear tissues of animal models [6]. While perilymph analysis provides valuable insights into extracellular signaling molecules and biomarkers, it lacks information on intracellular proteins, structural components, and cell‐type‐specific expression patterns. Since perilymph studies primarily detect soluble proteins, they miss membrane‐bound and cytoskeletal proteins essential for cochlear mechanics, synaptic transmission, and cellular integrity. In contrast, proteomic analysis of whole cochlear tissue offers a more comprehensive view by capturing a diverse range of proteins from various cochlear cell types, including hair cells, supporting cells, and neurons, enabling a deeper understanding of proteins involved in key structural, functional, and signaling pathways [7, 8]. Biological analysis of bilateral inner ear samples is routinely utilized in research [9, 10], particularly in studies where one side is subjected to experimental treatments or interventions while the other side serves as a control [11, 12]. Comparing the protein composition of both cochlear tissues of healthy animal models can establish a baseline to ensure the two sides are similar before any subsequent intervention. This baseline characterization ensures that any observed effects in the experimental ear are due to the intervention, not pre‐existing differences [13].

The guinea pig (Cavia porcellus) has been used as an animal model in hearing research since the 1930s–1940s and has several benefits compared to other species [14, 15]. Guinea pigs have a large hollow space surrounding the inner ear called the bulla, providing easy access to the inner ear for delicate surgical manipulations. By comparison, standard laboratory animals, like mice and rats, lack a large bulla, making experimental manipulations difficult. Moreover, the hearing range of guinea pigs (50–50,000 Hz) is more similar to humans (60–23,000 Hz) compared to rats (200–76,000 Hz) or mice (1000–91,000 Hz) [16, 17]. Generally, the anatomical similarity of guinea pigs to humans, their tolerance of anesthesia, and their suitability for cochlear manipulations make them valuable as an animal model to examine hearing problems and potential treatments [18]. Rats are also valuable models for studying changes in genetic transcripts and expression of proteins in sensory systems after trauma, leading to the expansion of scientific research on hearing problems [19]. Rats have larger inner ear structures than mice, making surgical manipulations and physiological recordings comparably easier, which may enhance the accuracy and feasibility of various auditory experiments [20].

In this study, we aimed to develop a comprehensive proteome database for the cochlear tissue in healthy rats and guinea pigs. Moreover, the proteome of the left and right ears was compared in each animal to understand significant changes in the protein abundance of bilateral cochlear tissues.

All experiments performed in this study were approved by the Macquarie University Animal Ethics Committee (AEC Reference No.: 2023/001). All animals' health and welfare were monitored daily by the Macquarie Animal Research Services Technicians, who were independent of the research team. All methods were carried out in accordance with the relevant guidelines and regulations, which included the Australian Code for the Care and Use of Animals for Scientific Purposes (8th edition, 2013) and the ARRIVE guidelines. A detailed description of the material and methods is provided in Supplementary Data 1. Briefly, four rats and four guinea pigs without a history of inner ear dysfunction were euthanized by 2% carbon dioxide inhalation and 1 mL intraperitoneal injections of “Lethabarb” (pentobarbitone sodium; 150 mg/kg) while under deep isoflurane anesthesia (5%), respectively. Then, animals were decapitated, and microsurgery was performed to extract both cochlear tissues from the left and right temporal bones. Cochlear tissues were stored in PBS at –80°C until the day of processing to minimize protein degradation.

The protein extraction process was optimized using three technical replicates from the right and left cochlear tissues from one guinea pig and one rat (details of all identified proteins are shown in Supplementary Data 2 and 3). Probe sonication proved superior to bath sonication for extracting cochlear proteins (Table S1). Probe sonication was selected to repeat experiments using three biological replicates of each animal. Briefly, each tissue sample was homogenized by adding 100 µL ice‐cold RIPA lysis buffer containing 25 mM Tris HCl pH 7.6, 150 mM NaCl, 1% NP‐40, 1% sodium deoxycholate, and 0.1% SDS (Thermo Scientific, San Jose, CA). After homogenizing on ice for approximately 1 min, each cochlear tissue was homogenized using a probe sonicator (Branson 450 Digital Sonifier, Mexico) for 10 s (Duty cycle 25%—output control: 2). Samples were then centrifuged for 15 min at 12,000 × g at 4°C. The supernatant was collected to determine protein concentration using a Pierce BCA protein assay kit (Thermo Scientific, San Jose, CA) according to the manufacturer's instructions.

Three biological replicates from guinea pig and rat cochlear protein extractions, each containing approximately 90 µg of protein, were prepared for SDS‐PAGE gel electrophoresis. In‐gel digestion was performed using the protocol described previously [21]. Briefly, each gel lane was cut into pieces. Pieces were washed with 100 mM NH₄HCO₃ and twice with 50% acetonitrile (ACN)/50% 50 mM NH₄HCO₃. The gel pieces were dehydrated with 100% ACN and reduced with 10 mM dithiothreitol (DTT) at 37°C for 1 h. Alkylation was performed with 55 mM iodoacetamide (IAA) in the dark for 45 min. Then, gel pieces were washed with 100 mM NH₄HCO₃ and 50% ACN/50% 50 mM NH4HCO3. Gel pieces were dehydrated by adding 100% ACN, and the liquid was removed in a vacuum centrifuge. Once dried, gel pieces were covered with 12.5 ng/µL trypsin solution (Promega, Sequencing grade) in 50 mM NH₄HCO₃ and incubated at 4°C for 90 mins before digesting at 37°C overnight. Peptides were extracted by adding 50 µL 50% ACN/2% Formic acid for 20 mins. Peptides were dried in a vacuum centrifuge and resuspended in 1% formic acid. The peptide samples were stored in a −20°C freezer, ready for analysis by nanoLC‐MS/MS.

Trypsin digest from cochlear tissues was analyzed on a Q‐Exactive Orbitrap mass spectrometer connected to an UltiMate 3000 nanoLC system (Thermo, San Jose, CA, USA). Chromatographic separation was performed on a reversed‐phase fused silica capillary column, 75 µm ID × 30 cm, packed with Halo‐C18 160 Å, 3 µm packing material, and maintained at 45°C. Before the analytical column, peptides were loaded on a trap column 300 µm × 5 mm (Thermo Acclaim PepMap C18 100). Samples were loaded in buffer A (2% v/v ACN, 0.1% v/v FA), and peptides were resolved over a 60 min gradient with a flow rate of 300 L/min at 45°C. The 60 min gradient was developed using buffer A (0.1% v/v formic acid) and buffer B (99.9% v/v ACN, 0.1% v/v FA) as follows: 2% buffer (3 mins), 37% buffer B (50 mins), 80% buffer B (7 mins). The spectral acquisition was performed in positive ion mode over a scan range of 400 m/z to 1500 m/z (70,000 resolution, 3e6 AGC target, 100 ms maximum injection time). The data‐dependent acquisition method was used to acquire MS/MS spectra of the 10 most abundant ions (17,500 resolution, 1e5 AGC target, 50 ms maximum injection time, 4.0 m/z isolation width, dynamic exclusion enabled for 20 s).

Raw data files from the Q‐Exactive mass spectrometer were searched in MaxQuant version 1.6.14 [22] against the UniProt Guinea pig and Rat sequence databases, supplemented with the MaxQuant contaminant database. Oxidation and N‐terminal acetylation were allowed as variable modifications, and carbamidomethylation of cysteine was a fixed modification. Trypsin was selected as the enzyme specificity, with one missed cleavage allowed. Match‐between runs were selected for protein quantitation, and LFQ intensities from the biological replicates were combined. The multiplicity and LFQ min ratio count were set at 1 for label‐free quantification. Peptide and protein level identification was set to 1% using a target‐decoy‐based strategy.

Using MaxQuant outputs, all protein IDs were searched on the UniProt website after removing reversed database hits and potential contaminants. Further analysis of MaxQuant outputs was performed using Perseus 2.0.11 [23], LFQ‐Analyst [24] and Panther [25]. Perseus filtered the total identified proteins in each animal model based on those quantified in all three biological replicates based on LFQ intensity, producing an output list of reproducibly identified proteins for each tissue type (Table 1). A non‐redundant total of 1841 and 3423 proteins were identified in the cochlear tissues of guinea pigs and rats, respectively (Supplementary Data 4 and 5). Among the most abundant proteins of both guinea pigs and rats, three proteins were found in all replicates: cochlin, albumin, and hemoglobin. Previous proteomic analysis showed that cochlin is the predominant protein in cochlear tissue and one of the major components of the extracellular matrix in the inner ear [26]. Albumin also accounted for approximately 14% of the total perilymph proteome [27].

TABLE 1.

Summary of protein identification data of cochlear tissue samples for three biological replicates in guinea pigs and rats.

Samples	Reproducibly identified proteins		Proteins			Peptides			Protein	Peptide
			R1	R2	R3	R1	R2	R3	RSD (%)	RSD (%)
GP‐Right	1162	944	1498	1518	1472	14,361	14,227	14,516	1.5	1.0
GP‐Left	1090		1247	1603	1419	13,443	14,549	14,017	12.5	3.9
Rat‐Right	1976	1589	2924	2852	2170	26,301	26,770	24,461	15.7	4.7
Rat‐Left	1865		2201	2654	2758	24,320	24,651	25,101	11.6	1.6

To improve our understanding of the functional roles of total cochlear proteomes in each animal, Gene Ontology (GO) and the Panther classification were applied. Regarding the protein classes, metabolite interconversion enzyme and protein modifying enzyme constitute many proteins in both animal cochlear proteomes. Moreover, the GO analysis of whole cochlear tissues in guinea pigs and rats showed that the cellular and metabolic process and biological regulation involved the highest number of proteins in the biological process category in both animals (Figures S1 and S2). This result was similar to the biological process characterization of human perilymph in a previous study [3].

To our knowledge, this is the first mass spectrometry‐based proteomic analysis of healthy rat cochlear tissue. The result indicated a five‐fold increase in the cochlear proteome of Sprague‐Dawley compared with a previous study (649 protein) [6]. Although no previously published studies have focused on the protein profile of the whole cochlea in normal guinea pigs, Lee and colleagues found 1413 proteins in the 24 perilymph samples from right and left ears using LC‐MS/MS [13]. Based on these results, our guinea pig cochlear proteome contains 765 perilymph proteins, which account for over 50% of the proteins reported in their study (Supplementary Data 6). Since they used 12 animals, which is four times greater than our guinea pig sample size (n = 3), increasing the number of animals is expected to increase shared proteins.

The cochlear proteomes of guinea pigs and rats were analyzed based on their bilateral location in three biological replicates, using LFQ‐Analyst [24] (Figures 1a and 2a). The most abundant proteins identified in bilateral cochlear tissues, based on MS/MS count numbers, included Albumin, Cochlin, Myelin protein P0, and Hemoglobin subunit beta in guinea pigs, and Albumin, Hemoglobin subunit beta‐1, Cochlin, Hemoglobin alpha, and Myosin heavy chain 4 in rats (Supplementary Data 7 and 8).

FIGURE 1.

(a) Comparative proteomic analysis between right and left cochlear tissue in guinea pigs. Overview of the experimental design for the proteomic analysis of the three biological replicates of cochlear tissues. (b) A Venn diagram representing the overlap of protein occurrence between the left and right cochlear tissues in guinea pigs shows that 93% of the total proteome is shared between both sides. (c) A volcano plot illustrating the relationship between statistical significance (adjusted p value < 0.05) and fold change magnitude between the left and right cochlear tissues in guinea pigs. (d) A heatmap displaying 32 significantly differentially expressed proteins (rows) across three replicates of left and right guinea pig cochlear tissues (columns). (e) Gene Ontology (GO) analysis of differentially expressed proteins between guinea pigs' left and right cochlear tissues. (f) GO biological process of differentially abundant proteins; protein class of differentially abundant proteins.

FIGURE 2.

(a) Comparative proteomic analysis between right and left cochlear tissue in rats. Overview of the experimental design for the proteomic analysis of the three biological replicates of cochlear tissues. (b) A Venn diagram illustrating the overlap of protein presence between the left and right cochlear tissues in rats shows that 89% of the total proteome is shared between both sides. (c) A volcano plot displays the relationship between statistical significance (adjusted p value) and fold change magnitude between rats' left and right cochlear tissues. (d) A heatmap showing the expression levels of 33 significantly differentially expressed proteins (rows) across three biological replicates of left and right rat cochlear tissues (columns). (e) Gene Ontology (GO) analysis of differentially expressed proteins between the left and right cochlear tissues. (f) GO biological process of differentially abundant proteins; protein class of differentially abundant proteins.

Our findings showed a high level of overlap with 93% and 89% between left and right ears in guinea pigs and rats, respectively (Figures 1b and 2b). In addition, only 32 and 33 proteins were present in both left and right cochlea but significantly differentially expressed between bilateral cochlear samples in guinea pigs and rats, respectively (Figures 1c and 2c). The relative protein abundance of the significant differentially abundant proteins and their expression in all guinea pig and rat replicates are also presented as a heatmap (Figures 1d and 2d). A previous study reported a 95.8% overlap with only 23 significantly differentially expressed proteins between left and right perilymph samples in guinea pigs [13]. The higher level of overlap in their study could be due to sampling pure perilymph through the round window. Notably, researchers have determined perilymph samples taken from the round window are prone to significant contamination with cerebrospinal fluid [28]. The blood contamination documented in this study was an inevitable factor in analyzing whole cochlear tissues in rats and guinea pigs [29, 30]. A recently published study investigated human perilymph's low molecular weight (LMW) protein and metabolite content to compare the molecular profiles between the right and left ears. As a result, a strong correlation between the mass spectrometry profiles of metabolites from the left and right ears has been reported, indicating good individual reproducibility when comparing the two sides [31]. Consistent with the previously published studies, our experiments revealed that the protein expression level did not differ significantly between left and right cochlear tissues in guinea pigs and rats. This high degree of overlap indicates strong bilateral molecular symmetry, which is essential for maintaining symmetrical auditory function and balanced neural input, thus supporting the use of one ear as a control in unilateral intervention studies.

To determine the functional classification, GO and Panther classification systems were used to search protein IDs against C. porcellus (CAVPO) and Rattus norvegicus. The differentially expressed proteins, as well as proteins present only in the left and right cochlear tissues, were classified as the Panther GO‐slim biological process and Panther protein class. In guinea pigs, the protein class categories with the highest number of proteins included metabolite interconversion enzymes, protein‐modifying enzymes, and membrane traffic proteins, while in rats, the most abundant protein classes were metabolite interconversion enzymes, immunity proteins, and protein‐modifying enzymes (Figures 1e and 2e). The biological process GO demonstrated that most of these proteins in both animals are involved in cellular and metabolic processes (Figures 1f and 2f).

Consistent with the previous human perilymph proteome [3], several proteins linked to hearing function were identified in the cochlear proteomes of rats and guinea pigs, including Cochlin, Radixin, Ezrin, Actin‐related protein 2, Aquaporin‐1, Catalase, Cofilin‐1, Gelsolin, Dual Specificity Mitogen‐Activated Protein Kinase 1, Aldehyde dehydrogenase, Plasminogen, Apolipoprotein E, and D. Detailed information about key findings is available in Figure S3 and Table S2. In addition, the cochlear proteomes of guinea pigs and rats shared several proteins with human perilymph from patients with different causes of deafness [32] (Figure S3). Common proteins identified across both animal models and three pathological conditions—enlarged vestibular aqueduct (EVA), Meniere's disease (MD), and otosclerosis—include Complement C3, Clusterin, Prothrombin, Vitronectin, Cochlin, Hemopexin, Ceruloplasmin, Peroxiredoxin‐2, Histidine‐rich glycoprotein, and Transthyretin (Table S3).

We acknowledge that one limitation of the study is the absence of objective measurements of auditory function before proteomics sampling and analysis. Despite this limitation, we performed behavioral and physical examinations to provide a qualitative indication of the general hearing ability in our animals. This included observing the animals' reflexive responses (the Preyer Reflex) to abrupt auditory stimuli to confirm functional hearing [33] as well as visual assessment of the ear canal and tympanic membrane under the operating microscope, to rule out infection, damage, or obstruction in the auditory pathways [34]. We also recognize the limited sample size of the study, with only three biological replicates per animal model (guinea pigs and rats). While the high overlap in cochlear proteome between the left and right ears (93% in guinea pigs and 89% in rats) supports the assumption that cochlear tissues from both ears are biologically equivalent, future studies should aim to increase the number of animals per group to enhance the reliability of the findings.

In conclusion, this study developed a platform for mass spectrometry‐based proteomic analysis of cochlear tissue, comparing protein extraction methods in guinea pigs and rats. It found that probe sonication was more effective than bath sonication for extracting cochlear proteins. The study also revealed substantial similarity in protein composition between the left and right cochlear tissues in both species. This research establishes a foundation for future studies, paving the way for advancements such as increased sample sizes, optimized sample preparation, expanded sample diversity, and refined methodologies like mass spectrometry to uncover pathways and triggers of inner ear diseases.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File 1: pmic13977‐sup‐0001‐SupMat‐Data‐1.docx.

Supporting File 2: pmic13977‐sup‐0002‐SupMat‐Data‐2.xlsx.

Supporting File 3: pmic13977‐sup‐0003‐SupMat‐Data‐3.xlsx.

Supporting File 4: pmic13977‐sup‐0004‐SupMat‐Data‐4.xlsx.

Supporting File 5: pmic13977‐sup‐0005‐SupMat‐Data‐5.xlsx.

Supporting File 6: pmic13977‐sup‐0006‐SupMat‐Data‐6.xlsx.

Supporting File 7: pmic13977‐sup‐0007‐SupMat‐Data‐7.xlsx.

Supporting File 8: pmic13977‐sup‐0008‐SupMat‐Data‐8.xlsx.

Supporting File 9: pmic13977‐sup‐0009‐SupMat‐Figures.docx.

Supporting File 10: pmic13977‐sup‐0010‐SupMat‐Tables.docx.

Data Availability Statement

The mass spectrometry proteomics data have been deposited in the ProteomeXchange Consortium via the PRIDE [30] partner repository with the dataset identifier PXD059025.