2.1 Search strategy and databases

A search was conducted on the following databases after consulting a specialist librarian (VP): Medline (via Ovid), Embase (via Ovid), Web of Science (Core Collection), and Scopus from inception to April 2024. The search strategy was peer-reviewed by two librarian colleagues of VP using the Peer Review of Electronic Search Strategies (PRESS) checklist [26], and evaluated against the PRISMA-S guidelines [27]. Databases were searched separately rather than multiple databases being searched on the same platform. The search syntax was adapted for each database, to account for variation in thesaurus terms/controlled vocabulary between databases. In the preliminary stage of this study, Google Scholar was used to identify a broad range of target papers that otherwise may not have been captured by conventional databases. Subsequently, a formal search strategy was performed using only established databases, as listed above.

Keywords were divided into two groups according to biomarker type and hearing loss pathologies. The first group included: Biomarker, Metabolome, Proteome, RNA, miRNA, or Inflammasome, and the second group included: Hearing loss, Deafness, Ménière, or Otosclerosis. Excluding animal studies resulted in a minority of the target papers being omitted as not all human studies were tagged appropriately. Therefore, animal studies were included in the primary search and subsequently removed at screening. This was achieved by excluding studies if the population described was clearly animal-based (e.g. rodent, primate, or other non-human models). Where abstracts were unclear, full texts were retrieved to assess the study population; this was performed by two independent reviewers. This process ensured that only human studies relevant to the review question were retained in the final analysis.

2.2 Study eligibility, inclusion, and exclusion criteria

Inclusion criteria for this study included: peer-reviewed studies; biomarkers from perilymph associated with hearing loss; human studies; international publications; and correlational studies. Correlational studies included cohort studies in which participants were grouped into different SNHL aetiologies with or without a control group, and associated perilymph biomarkers were identified. Exclusion criteria included: animal studies; case studies; inner ear tissue biopsy/cellular samples; endolymphatic sac fluid; review articles; vertigo; tinnitus; serum biomarker; genetic markers; conference poster articles; non-peer-reviewed studies. Only primary studies were included in this systematic review. Review articles were excluded to avoid double-counting evidence and introducing bias into the study conclusions. Similarly, case studies were also excluded due to their limited generalisability.

2.3 Study screening process

After duplicate removal, 7471 papers were identified. The title and abstract screening were completed by two independent reviewers (KW and NC) using the online platform, Rayyan. The most common reasons for exclusion at the title and abstract screening stage were: patients with other types of hearing loss (non-SNHL), reviews, and commentaries. 191 studies were subsequently selected for full text screening. At this stage, the most common reason for exclusion was different sample type (typically serum, CSF, or inner ear tissue). Previously, studies were reliant on these indirect systemic sample types when investigating the cochlear microenvironment. However, due to recent advances in safe, intraoperative sampling techniques, researchers are now able to include direct perilymph samples in their analysis. In turn, this review focused on the biomarker profile of these localised perilymph samples specifically, to help inform future studies utilising this novel technique in different hearing pathologies going forward. 14 eligible papers were identified from this screening. Finally, the reference list of each of the remaining papers were screened, and one additional paper was identified. A total of 15 papers were therefore included in this review, as outlined in (Fig 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. PRISMA flow diagram.

https://doi.org/10.1371/journal.pone.0334351.g001

2.4 Data extraction and analysis

The extracted data included: author, year of publication, country of publication, study design, level of evidence according to ‘The Oxford Centre of Evidence-Based Medicine’, perilymph extraction procedure (e.g., cochlear implantation/ vestibular schwannoma resection/ stapedectomy), biomarker identified, and key findings. Some studies included a large volume of screened biomarkers. Those that produced statistically significant results were included in the study analysis.

2.5 Risk of bias

The ROBINS-I tool was used to assess the risk of bias in the included studies. This tool evaluates seven domains of potential bias: confounding, selection of participants, classification of interventions, deviations from intended interventions, missing data, measurement of outcomes, and selection of reported resources. Each domain was assessed as low, moderate, or series risk of bias. The risk of bias was assessed by two independent reviewers (KJ and SC), and discrepancies were resolved through discussion or consulting a third reviewer (NC).

4.1 Heat Shock Proteins

Heat shock proteins (HSP) are a superfamily of proteins that play an important role in cellular protection, particularly under stress conditions such as heat, oxidative stress, or toxins [38]. Mechanical and chemical stressors can result in various changes in genomic transcription and protein folding. In response, however, HSPs work to protect cells from injury by promoting the folding of denatured proteins and preventing aggregation. Moreover, HSPs can suppress apoptotic pathways by interacting with proteins associated with signal transduction in active cell death [39]. These proteins are classified according to their weight into 6 families: small HSPCs, HSP40, HSP60, HSP70, HSP90, and HSP110 [40]. The neuroprotective role of HSPs have also been implicated beyond the cochlea, with HSP70 and small HSPs associated with enhanced retinal ganglion cell survival in optic neuropathies such as glaucoma [39].

In the cochlea, HSPs may play a protective role in hearing preservation following cochlear implantation. The variability in outcomes of cochlear implant patients remains largely unexplained. Clinical factors alone have proved insufficient for predicting patient performance [41], suggesting that additional factors such as the patient’s genetic profile and their individual cochlear microenvironment may influence their postoperative outcomes, particularly with regards to hearing preservation following cochlear implantation. Schmitt et al. evaluated the HSPs associated with residual hearing after cochlear implantation and found that subtypes 1 and 6 of HSP70 were found in all patients with preserved residual hearing [32]. The protective effect of HSP70 is supported by experimental mouse models in which HSP70-overexpressing mice were significantly protected against aminoglycoside-induced hearing loss and hair cell death compared to their wild-type counterparts [42].

Interestingly, not all HSPs exert a protective effect. HSP90 was found in higher levels in patients with a complete loss of residual hearing following cochlear implantation and only one patient with preserved hearing had HSP90 detected in the perilymph following surgery. In this study, 3 patients undergoing translabyrinthine vestibular surgeries were also included, however, no significant difference in HSP distribution was identified when compared to the cochlear implantation group. Furthermore, Edvardsson Rasmussen et al. identified three variations of HSP70 in only one of the 15 patients with vestibular schwannoma included in their study [31].

Of note, Schmitt et al.’s [32] findings were not replicated in Durisin et al.’s [22] analysis of perilymph proteins post-CI. These researchers evaluated hearing performance using speech intelligibility (HSM sentence test in noise at 10dB and Freiberg monosyllable word test) and grouped participants into excellent and poor speech intelligibility groups post-CI. They found higher levels of HSP70 1A/B in the poor speech intelligibility group, in other words, higher levels of this HSP in the group with poorer CI outcomes. Overall, the association of HSP70 with CI outcomes is inconsistent, with further investigation needed to clarify the role of HSPs as a useful marker for CI prognostication.

4.2 Immune/ Inflammatory mediators

Historically, the inner ear, including the cochlea, was assumed to be an immune-privileged organ similar to the eyes or brain, due to the presence of a blood-labyrinthine-barrier (BLB) with tight junctions [43]. This forms a protective barrier blocking immune cells, inner ear antigens, and antibodies from entering. However, recent studies have demonstrated the presence of a complex immune microenvironment within the cochlea itself [44]. An initial study demonstrating the responsiveness of certain types of SSHL to steroid treatment was the first to challenge the idea of the cochlea as an immune-privileged organ [45]. Later, Zhang and colleagues revealed two cell types of macrophage lineage in the lateral wall of the cochlea: cochlear macrophages and perivascular macrophage-like melanocytes, which are involved in clearing cell debris, maintaining cochlear fluid homeostasis, and regulating BLB permeability [46]. Animal studies have demonstrated changes in inflammatory gene expression in the cochlea following noise exposure, including the upregulation of genes such as CXL10, SOCS3, and TCl1b1 [47]. Other studies have identified increased pro-inflammatory cytokines such as TNF-α, IL-1β, and IL-6 following noise exposure [48]. Since these initial findings, an extensive list of inflammatory gene regulations has been reported by Karayay and colleagues [49]. Collectively, these findings suggest a role for inflammatory cells/mediators in the development of cochlear damage and SNHL.

Soluble components of the innate immune system have recently been identified in the perilymph which correlate with cochlear implant performance. Warnecke and colleagues identified higher concentrations of inflammatory cytokines, IL-13 and IL-9, in patients with complete hearing loss following CI surgery when compared to patients with some residual hearing following surgery, although the difference in IL-9 did not reach statistical significance [25]. The complement system is another crucial part of the innate immune system and has roles in pathogen opsonisation, chemotaxis and activation of leukocytes, cytolysis, as well as clearance of apoptotic cells and immune complexes [50]. Higher levels of complement component C8 alpha chain were found in patients with excellent speech intelligibility at 1-year post-CI compared to the poor speech intelligibility group [22]. Complement C8 is an important antibacterial immune effector and is part of the membrane attack complex that forms a pore in bacterial membranes and leads to cell lysis. Complement C1r subcomponent and complement H were also found to be in higher abundance in the perilymph of patients with Ménière’s disease compared to the perilymph of otosclerosis controls [33]. Furthermore, complement C4 was found in a higher abundance in adults when compared to children following cochlear implantation, potentially providing insight into differences in perilymph composition in presbycusis by age [11]. Finally, one study identified alpha-2-HS glycoprotein as an independent variable for tumour-associated hearing loss [31]. Alpha-2-HS glycoprotein is an acute phase response protein and is involved in neutrophil degranulation [51]. Recent studies have also hypothesised that alpha-2-HS may be excreted from vestibular schwannomas into the extracellular perilymph where it exerts pro-inflammatory activity [31]. These findings are in congruence with a previous study by Schmitt et al. that identified alpha-2-HS in several samples from patients with an unknown cause of sensorineural hearing loss [11].

Components of the adaptive immune system, specifically immunoglobulin (Ig) chains, have also been associated with SNHL. One study found Ig Kappa chain V-IV regions upregulated in patients with excellent hearing performance following cochlear implantation. However, a different subset of immunoglobulins was upregulated in patients with poor hearing performance, specifically genes IGHV1–2 and IGHV1–46 which encode various regions on the immunoglobulin heavy chain, and IGKV6–21 for immunoglobulin kappa variable 6–21 [22]. Two further studies have also suggested a role for attractin in SSHL [11,22]. Attractin is a circulating glycoprotein that is rapidly expressed on activated T cells [52] and was found to be more abundant in patients with Ménière’s disease but also patients with higher speech intelligibility after CI surgery. Collectively, these findings suggest that a diverse set of immune or inflammatory mediators may be associated with SNHL, with a potential impact on CI performance, ranging from interleukins and complement proteins to acute phase reactants and immunoglobulins.

4.3 microRNAs

Changes in the protein environment, or ‘proteome’ of the inner ear may also contribute to SNHL, a process regulated in part by microRNAs (miRNA). MiRNAs are 19–23 base pair single-stranded RNA sequences that regulate gene expression through mRNA degradation and silencing [53,54]. Beyond the ear, miRNAs in CSF are being evaluated as potential markers of neurodegenerative diseases such as Alzheimer’s and Parkinson’s [55] and in the aqueous humour of the eye for glaucoma [56]. Equally, miRNAs have been shown to play a crucial role in inner ear development. The miR-183 family is expressed as the otic vesicle develops from the otic placode and is later restricted in expression to hair cells and the spiral ganglion [57].

Previous studies on peripheral blood samples identified 24 differentially expressed miRNAs in SNHL patients when compared to healthy controls, with subsequent functional annotation analysis revealing target genes in arachidonic acid metabolism, complement, and coagulation cascades [58]. Focusing on perilymph samples in particular, work by Shew et al. [24,34–36] has identified a diverse set of miRNAs correlated with SNHL of varying aetiologies. In their 2018 study, three miRNAs (mi1233-5p, mi455-3p, mi638) were expressed in patients undergoing cochlear implantation who were diagnosed with Ménière’s disease, but not in the other implantation patients without Ménière’s disease [24]. Mi1233-5p, mi455-3p, and mi638 form part of a regulatory network, including two proteins, aquaporin 4, and TNFSF12, which have been implicated in the pathogenesis of Ménière’s disease [59,60]. However, caution should be exercised when interpreting these findings, as the number of control participants was not stated in this study [24]. Shew and colleagues subsequent study in 2021 corroborated these findings and identified miRNA 1299 and 1270 uniquely and differentially expressed in the perilymph of patients with Ménière’s disease [35]. Both miRNA 1299 and 1270 are linked to inflammatory and autoimmune pathways, while miRNA-1299 can be linked to aquaporin expression specifically. The author’s 2019 study also utilised a machine learning approach and identified several ‘critical miRNA’ that differentiated conductive and SNHL. The downstream targets were identified as KCNJ10, HCN, and Otoferlin which are important for propagating endocochlear potentials [61], post-synaptic potentials [62], and Ca2 + evoked vesicular endocytosis in inner hair cells [63], respectively.

In addition to aquaporin and ion channel expression, miRNAs may also play a role in the neurotrophin pathway as discussed below. Shew and colleagues found that a lower expression of miR-1207 and miR4651, which target neurotrophin receptors 2 and 3 (NTR2 and NTR3), was statistically correlated with having no residual hearing (pure tone audiometry threshold > 80dB) prior to cochlear implantation [35]. These findings demonstrate the potential clinical utility of implementing miRNA profiling before CI surgery, to optimise post-operative outcomes in patients with SNHL.

4.4 Neurotrophin pathway proteins

As previously suggested, neurotrophic proteins may represent an important group of perilymph biomarkers for SNHL with several studies directly investigating their presence in the perilymph. Neurotrophins are a family of molecules that play an integral role in the development and support of neurons. In the murine inner ear, BDNF and NT-3 have been shown to regulate the connection of hair cells to neurons [5,64]. One study found that BDNF-regulated proteins were correlated with residual hearing prior to cochlear implantation and improved performance after 1 year, although the samples were obtained from patients with SNHL of various aetiologies [25]. Another study found higher levels of kallikrein in patients with excellent hearing post-CI surgery [22] which works to increase the expression of BDNF and pro-survival Bcl-2 genes [65]. However, Schmitt and colleagues did not detect endogenous BDNF in their samples [11].

Endothelial factors may also play a role in inner ear repair. Warnecke and colleagues found reduced levels of VEGF-D in patients with no residual hearing (PTA threshold > 80dB) compared to patients with residual hearing following cochlear implantation [25]. Interestingly, VEGF-D has a central role in lymphangiogenesis [66], suggesting that reduced levels of VEGF-D in patients with no residual hearing may indicate ongoing damage and reduced repair. Similarly, Schmitt and colleagues found higher levels of HSPG in patients with enlarged vestibular aqueducts and Ménière’s disease, compared to those with otosclerosis [33]. HSPG has complex regulatory roles over VEGF A, VEGFR2, and α2β1 integrin signalling axes [67]. Another growth factor regulator, insulin-like growth factor binding protein 1 (IGFBP1), which regulates insulin-like growth factor 1 (IGF-1), was higher in patients with no residual hearing [25]. Although, the precise correlation of neurotrophins and vascular factors with SNHL is unclear, they are likely to provide insight into the pathogenesis of SNHL and the aetiologies of related diseases.

4.5 Metabolome

A growing body of literature suggests that the metabolic profile of perilymph may provide insight into SNHL pathogenesis. Trinh and colleagues found a correlation between the metabolic profile of perilymph and the duration of hearing loss in patients (less than 12, or more than 12 years of hearing loss), specifically elevated levels of N-acetylneuraminate, which is a metabolite located on the terminal glycoprotein and glycolipid on the surface of cell membranes that increases following cell membrane breakage (hair cell apoptosis) [37]. Other discriminant metabolites found in this study include: glutaric acid, cysteine, butanoate, and xanthine. Similarly, another study found the protein short-chain dehydrogenase/reductase family 9C member 7 (SDR9C7) uniquely in 11 out of 12 samples from Ménière’s disease patients, but not in patients with an enlarged vestibular aqueduct or otosclerosis [33]. SDR9C7 is involved in vitamin A metabolism, specifically the conversion of retinal to retinol in the presence of NADH. Mutations in SDR9C7 have already been linked to dermatological pathology (autosomal recessive ichthyosis). However, the role of this enzyme in the context of the cochlea is yet to be explored.

Reactive oxygen species (ROS) also form a component of the metabolic profile of perilymph. ROS are responsible for direct cellular damage to lipids, proteins, and DNA, and trigger apoptosis or necrosis [68]. Elevated levels of ROS have been detected in patients with profound SNHL and have been attributed to the activity of the xanthine dehydrogenase/xanthine oxidase enzyme system [29]. One study found that proteins related to apoptotic signalling and ROS were upregulated following intratympanic steroid treatment, specifically serine leukocyte inhibitor and annexin A1 which are both involved in ROS processes and have anti-inflammatory activities [69]. In this way, several metabolic markers could provide insight into the pathogenesis of SNHL.

4.6 Structural proteins

Finally, a selection of structural proteins has been detected at higher levels in SNHL patients than in controls. Flnb and ACTB, which encode filamin B and beta-actin respectively, were found to be highly abundant in Ménière’s disease patients compared to those with alternative hearing loss aetiologies [28]. Six out of the seven highly abundant proteins in Ménière’s disease found by Arambula and colleagues [28] were also identified in Schmitt and colleagues’ [33] list of 33 proteins. Another study reported low-density lipoprotein-related protein 2 (LPR2) in the perilymph of vestibular schwannoma patients but not controls undergoing CI without vestibular schwannoma [20]. LRP2 is a transmembrane receptor protein found primarily in absorptive epithelial cells and is a key player in mediating endocytosis and mutations in LPR2 are associated with Donnai-Barrow syndrome and facio-oculo-acoustico-renal syndrome, which present with SNHL [70]. These structural proteins present potential candidates for perilymph biomarkers and should be carefully investigated in future work.