Hearing loss (HL) is a prevalent sensory disorder with far-reaching consequences on an individual’s life, encompassing social, functional, and psychological aspects [1]. Deprivation of auditory abilities can hinder access to spoken communication, impeding the development of language skills and executive functioning in children [2] and contributing to the risk of dementia and cognitive decline in older age groups [3].

HL affects 466 million people worldwide and it is estimated that in 2050 over 700 million people will need rehabilitation support for HL (https://www.who.int/, accessed on 27 June 2023).

Sensorineural hearing loss (SNHL) stands as the most prevalent form of hearing impairment, involving various pathologies within the inner ear and auditory nerve. The primary mechanism behind SNHL lies in the damage or loss of sensory receptor cells in the cochlea and the neurons responsible for relaying auditory information to the central auditory system. As a result, SNHL presents as a chronic condition for which rehabilitative, rather than curative, treatment option exist [4].

SNHL can be congenital or acquired. The most common causes of permanent congenital sensorineural are cytomegalovirus infection, structural abnormalities of the temporal bones, and genetic causes [5]. The principle causes of acquired SNHL are presbycusis, acoustic trauma, ototoxic drugs, sudden HL, and infectious conditions ^[6][7][6,7]. A significant challenge in managing SNHL lies in the limited understanding of its pathogenetic mechanisms. Although several models have been proposed to explain the molecular pathways leading to hair cell death and damage to spiral ganglion neurons, common pathways have been observed across different damage etiologies, including the accumulation of reactive oxygen species (ROS), activation of MAPK signaling, and apoptosis pathways [8].

Contemporary research endeavors are dedicated to identifying predictive elements with diagnostic and prognostic value, aimed at improving the management and understanding the evolution of SNHL. Indeed, the integration of precision medicine principles into translational HL research and clinical practice holds the potential to offer personalized medical care, addressing the diverse etiologies of HL with greater precision and efficacy.

The past two decades have witnessed a veritable explosion of new omics technologies, which have contrasted the reductionist approach, typical of traditional biology, with a broader holistic one [9]. This shift in perspective marked a real revolution in the biomedical field, especially in dealing with complex problems, such as HL, characterized by multifactorality and high interindividual variability [10].

Metabolomics is an emerging omics science, which studies the set of metabolites, i.e., the metabolome, contained within a biological system. As the metabolome lies downstream from the central dogma of molecular biology, it is able to reflect both the genetic characteristics and the environmental influences. Moreover, due to its dynamic nature, the metabolome undergoes rapid and well-defined changes in case of disease ^[11][12][11,12]. As a result, metabolomic analysis opens a unique window on the mechanisms underlying the different phenotypes.

There are two possible approaches in metabolomics, targeted and untargeted. The former is aimed at quantifying a set of specific metabolites, and is generally chosen to test a hypothesis. The second, on the other hand, allows the entire metabolome to be explored without starting from preexisting hypotheses, but rather allowing new ones to be formulated. Both of these approaches have already provided important results in numerous fields of medicine, ranging from the study of cancer ^[13][14][15][13,14,15] to fetal malformations ^[16][17][18][16,17,18], psychiatric, neurodevelopmental and neurodegenerative disorders ^{[19][20][21][22][23]}[19,20,21,22,23] as well as chronic and autoimmune diseases ^[24][25][24,25] and internal medicine pathologies ^[26][27][28][26,27,28], just to name a few.

2. Metabolomic Profiling in Humans with Hearing Loss

2.1. Metabolomic Analysis on Perylimph Samples

Metabolomic analyses on perilymph samples from individuals with SNHL were performed by Trinh et al. ^[29][38], who investigated the relationship between metabolomic profiles and the duration of HL in cochlear implanted patients. This study revealed significant differences in the metabolomic profiles of patients with more and less than 12 years of HL, with N-acetylneuraminate showing a strong positive correlation with the duration of HL.

2.2. Metabolomic Analysis on Blood-Derived Samples

Wang et al. ^[30][42] investigated sudden sensorineural hearing loss (SSNHL) and its metabolic implications. The research included 20 SSNHL patients and 20 healthy controls (HCs), with serum samples collected for analysis. Metabolites were detected using LC-MS, and distinct metabolic profiles were observed in SSNHL patients compared to HCs, with a significant alteration in fatty acids metabolism. The patients were then categorized into recovery and non-recovery groups, and 12 distinctive metabolites were observed between the two groups, with significant positive associations between serum N4-Acetylcytidine, sphingosine, and nonadecanoic acid levels with hearing recovery in SSNHL patients. Metabolomic investigations were also used to assess the correlation between age ARHL and other pathological conditions. Using a publicly available dataset of the Alzheimer’s Disease Neuroimaging Initiative (ADNI), Llano et al. ^[31][41] investigated the serum lipidomic profiles in Alzheimer’s Disease (AD) subjects with and without HL. An interesting result was obtained for phosphatidylcholines, known to be one of the most important phospholipids of cell membrane, that showed a decreased abundance in AD patients with HL. A recent study ^[32][39] investigated the plasma metabolomic profile of individuals suffering from noise-induced hearing loss (NIHL). As a result, a total of seven pathways emerged as being involved, including glycerophospholipid metabolism, glycosylphosphatidylinositol (GPI) anchor biosynthesis, autophagy pathway, choline metabolism, the alpha-linolenic acid metabolism and linoleic acid metabolism, and retrograde endocannabinoid pathway. To further investigate how noise exposure affects the autophagy pathway, as part of this study the authors assessed the expression of three autophagy-related genes, namely PI3K, AKT and ATG5, which were found to be significantly down-regulated in NIHL patients compared to controls. Zhang et al. ^[33][40] studied noise-induced metabolic dysregulation by analysing a total of 90 samples collected from 60 noise exposed individuals, half of whom suffering from NIHL, and 30 HC. A total of 6 pathways resulted to be associated with both NIHL and noise exposure, including retrograde endocannabinoid signaling, sphingolipid signaling pathway, vitamin digestion and absorption, Fc gamma R-mediated phagocytosis, phospholipase D signaling pathway, and central carbon metabolism in cancer. Among the 7 metabolites emerged as crucial in differentiating NIHL patients from the other two groups, Pro-Trp dipeptide, adenine, and dimethylglycine exhibited a progressive decline pattern, starting from the control group to the non-NIHL group, and reaching the lowest levels in the NIHL group. These altered levels suggest that these metabolites may play significant roles in the development of noise-induced disorders. In addition, the researchers conducted cochlear gene expression comparisons between mice susceptible and resistant to NIHL using the GSE8342 dataset from Gene Expression Omnibus (GEO). The analysis revealed a strong association between immune response and cell death-related processes and noise exposure, suggesting their crucial involvement in noise-induced disorders.

2.3. Metabolomic Analysis on Urine Samples

Carta et al. [34] focused on individuals with idiopathic sudden-onset sensorineural hearing loss (ISSNHL) who underwent a treatment regimen involving steroids, heparin, and hyperbaric oxygen therapy. Through NMR-based urine metabolomic profiling, notable distinctions emerged between subjects who positively responded to steroid therapy and those who did not. Specifically, the non-responders exhibited elevated levels of β-alanine, trimethylamine N-oxide (TMAO), and 3-hydroxybutyrate, while citrate and creatinine showed an opposite pattern. In another study [35], through GC-MS analysis on urine samples from individuals with presbycusis and normal hearing matched controls, a total of 23 metabolites were found to be differently expressed. KEGG (Kyoto Encyclopedia of Genes and Genomes) pathway analysis revealed that these metabolites were mainly involved in the metabolism of glutathione, amino acids, glucose, and related to GABA and NMDA receptors activity. The findings also showed a notable disparity in the accumulation of kidney deficiency points between the presbycusis group and elderly individuals with normal hearing. This is a scoring system for evaluating symptoms associated with kidney deficiency, as formulated in accordance with the ‘Standards of Practice for Assessing Grading Factors of Kidney Deficiency Syndrome Differentiation’ as revised by the ‘Standards of Reference for Traditional Chinese Medicine Deficiency Syndrome’, under the auspices of the Chinese Integrated Traditional and Western Medicine Deficiency Syndrome and Geriatrics Research Professional Committee. Thus, the authors posit that these identified metabolic pathways could potentially serve as the underlying basis for the connection between presbycusis and kidney deficiency, as postulated in traditional Chinese medicine. Moreover, in a retrospective study by Pudrith and Dudley [36], the urinary levels of volatile organic compounds (VOCs) were used to investigate the implication of oxidative stress in acquired sensorineural hearing loss (ASNHL), considering that reliable biomarkers are lacking ^[37][43]. The study revealed a significant increase of around 3 to 4 dB in high-frequency pure-tone thresholds among individuals in the upper quartile groups of five specific metabolites. These metabolites were identified as glutathione-dependent mercapturic acids, originating from parent compounds such as acrylonitrile, 1,3-butadiene, styrene, acrylamide, and N,N-dimethylformamide. Notably, these associations were evident only in individuals who reported no exposure to noise. Other studies have focused on the correlation between exposure to specific chemicals and auditory system function. Two studies investigated the implications of polycyclic aromatic hydrocarbons (PAHs) in HL ^[38][39][31,33]. PAHs are environmental pollutants which are taken up by humans through inhalation, ingestion or even dermal exposure and have been associated with numerous harmful effects on human health, including the induction of oxidative stress ^[40][44]. Chou et al. found a dose-response relationship between exposure to PAHs and increased hearing thresholds in adults, emphasizing the potential role of these metabolites in causing damage to the cochlea, which may be a result of the inflammatory response they induce ^[39][33]. Following the same line of research, Li et al. ^[38][31] investigated the association between PAHs and HL as well as the potential role of systemic inflammation as a key mediator of this correlation in a large cohort of adults and adolescents. In this study, a significant association between the urinary levels of PAHs and HL was confirmed in both groups, hence suggesting that monitoring these levels may be a useful strategy to prevent HL. Interestingly, C-reactive protein, which is a valuable marker of inflammatory response, was found to have no mediating effects in the relationship between PAHs and HL. Other compounds of growing concern that have already been linked to neurotoxicity phenomena are pyrethroids, commonly used as insecticides. In a study by Xu et al. ^[41][30], urinary levels of 3-phenoxybenzoic acid (3-PBA), a pyrethroid metabolite, were evaluated to assess pyrethroid exposure in a wide cohort of adolescents. The study revealed a strong association between the exposure to these compounds and an increase in hearing thresholds. Similarly, Long & Tang ^[42][37] further investigated the ototoxicity induced by pyrethroid exposure in a large cohort of adults. In this case, an association between 3-PBA levels and both low and high frequencies hearing thresholds was described in subjects aged 20–39 years. In the pursuit of understanding the potential harmfulness of specific environmental substances, a study investigated the impact of caffeine exposure on hearing thresholds. However, no association between caffeine metabolites and hearing function was observed in US adults ^[43][32].

3. Metabolomics Investigations on Animal Models of Hearing Loss

3.1. Metabolomic Analysis on Tissues

In a study on mice by Ji et al. ^[44][49], a targeted metabolomic analysis for 220 metabolites was performed to investigate the effects of noise exposure on inner ear. Inner ear samples were collected immediately after noise exposure. A total of 40 metabolites were found to be affected by noise, with metabolic alterations depending on the intensity and duration of noise exposure. The main metabolic pathways altered by noise were purine metabolism and alanine, aspartate and glutamate metabolism, which were found to be upregulated, while phenylalanine metabolism and phenylalanine, tyrosine and tryptophan biosynthesis were downregulated. In a study carried out by Miao et al. ^[45][50] involving mice, when analyzing cochlea tissues, 17 metabolites demonstrated statistical significance in differentiating between noise-exposed mice and control subjects, with a total of 9 metabolic pathways involved. Among these metabolites, only three, namely spermidine, 3-hydroxybutyric acid, and orotic acid, showed higher levels in the noise-exposed group as compared to the controls. Interestingly, when examining rat brains exposed to acoustic trauma, the metabolomic analyses showed diverse metabolic patterns among different regions, yet regions with similar functions displayed analogous metabolite compositions ^[46][48]. Notably, in certain brain areas, such as the auditory cortex, 17 crucial metabolites were identified, distinguishing between control and acoustic trauma-exposed animals. These metabolites predominantly involved amino acid metabolism, specifically affecting the alanine, aspartate, glutamate, arginine, proline, and purine metabolic pathways.

3.2. Metabolomic Analysis on Biofluids

Fujita et al. ^[47][46] performed two distinct experiments on guinea pigs. The first involved a comparison of metabolite abundance in the inner ear fluid and plasma under normal conditions, showing differences in the levels of 15 metabolites. Then, they investigated the metabolomic alterations after exposing the guinea pigs to loud noise. As a result, a total of ten metabolites showed altered concentrations after noise exposure, namely 3-hydroxy-butyrate, glycerol, fumaric acid, galactosamine, pyruvate + oxaloacetic acid, phosphate, meso-erythritol, citric acid + isocitric acid, mannose, and inositol. The characterization of “inner ear fluid” as per the authors denotes a composite fluid resulting from the amalgamation of both endolymphatic and perilymphatic fluids. This definition aligns with the methodology employed, wherein the authors meticulously extracted this fluid subsequent to the removal of temporal bones, employing a dissecting microscope. This process involved delicately accessing the lymphatic fluid within the cochlea via the round and oval windows. While it is probable that the bulk of the fluid extracted corresponded to perilymphatic fluid, it is conceivable that endolymphatic fluid might have also been encompassed within the collection. The metabolomic alterations following acoustic trauma were also investigated on perilymph samples extracted from sheep in a study by Boullard et al. ^[48][51]. In this context, the sheep served as their own reference for comparison. Indeed, for each sheep only one ear was subjected to acoustic trauma, representing the NIHL model, while the other was used as control. As a consequence of noise exposure, 5 metabolic pathways resulted affected, namely phenylalanine/tyrosine/tryptophan metabolism, β-alanine metabolism, pantothenate and CoA biosynthesis, pyrimidine metabolism, and amino and nucleotide sugar metabolism. Moreover, an untargeted metabolomic analysis was performed on perilymph samples collected from guinea pigs exposed to a heavy impulse noise ^[49][45]. As part of this study, the potential otoprotective role of Hydrogen gas (H₂) inhalation was also investigated. Animals were divided into four subgroups: controls; animals exposed to noise only; exposed to noise and treated with H₂; only treated with H₂. Guinea pigs exposed to noise and receiving H₂ treatment were found to have a reduced HL compared to those in the noise only group. Metabolomic analysis confirmed the otoprotective role of this gas by showing that the perilymphatic metabolome of H₂-treated animals exposed to noise was more similar to that of the controls and H₂ only groups than to the noise only group. In particular, guinea pigs of the latter group exhibited higher levels of various acyl carnitines in their perilymph metabolome, suggesting an increased oxidative stress. Moreover, the noise only group showed lower levels of two osmoprotectans, namely stachydrine and homostachydrine, compared to the other three groups, indicating a potential osmolytic effect which may be attenuated by the H₂ gas. A serious and unresolved clinical issue is ototoxicity due to treatment with cisplatin which is the most used anticancer drug ^[50][52]. In a study by Pierre et al. ^[51][47] on guinea pigs, changes in serum metabolomic profile induced by cisplatin treatment were investigated. Some animals were treated with sodium thiosulfate (STS), a candidate otoprotectant, 30 min before cisplatin administration; another group of animals received sodium chloride. Blood samples were collected before and 4 days after the treatment and metabolomics investigations were carried out through LC-MS. Cisplatin treatment induced a significant increase in hearing threshold and also a loss of outer hair cell in both groups. However, metabolomics analysis revealed that cisplatin had a major effect on serum metabolome, with only minor, non-significant differences observed in subjects treated with STS. Conversely, it was observed an inverse correlation between the concentration of four metabolites, namely N-acetylneuraminic acid, ceramides, cysteinylserine, L-acetylcarnitine, and the hearing threshold shift at high frequencies (20 and 30 kHz), but only in the group treated with sodium chloride. In particular, L-acetylcarnitine is a short-chain acylcarnitine which is involved in numerous functions, including proper mitochondrial function, removal of oxidative products, and regeneration of peripheral nerves ^[50][52][52,53].