Single-cell proteomic profiling of cochlear hair cells and mechanotransduction interaction mapping

Abstract

Hearing loss is a major and growing global health burden, often resulting from irreversible damage to cochlear hair cells, the sensory cells responsible for mechano-electrical transduction (MET) in the inner ear. Although genetic and transcriptomic studies have extensively characterised these cells, protein-level mechanisms underlying hair-cell function and vulnerability remain poorly understood. This thesis applies mass spectrometry-based single-cell proteomics to investigate cochlear hair cells at cell-type resolution. A workflow was developed for the proteomic analysis of individual hair cells isolated from postnatal mouse cochleae. The approach combined one-pot sample preparation optimised for ultra-low input with liquid chromatography-tandem mass spectrometry, data-dependent acquisition, and label-free quantification. In parallel, affinity-purification mass spectrometry (AP-MS) was employed to investigate the interaction network of the cytoplasmic domain of protocadherin-15 (PCDH15), a key component of the MET apparatus. The workflow reproducibly quantified approximately 400 to 500 protein groups per cell, demonstrating that single-cell proteomic profiling of these scarce and fragile sensory cells is feasible. The results indicate that inner hair cells (IHCs) and outer hair cells (OHCs) share a largely common proteome, with cell-type identity primarily reflected by quantitative differences in protein abundance rather than the presence of cell-type-specific proteins. OHCs showed a tendency towards higher abundance of metabolic, redox, and cytoskeletal proteins, consistent with their electromotile function and increased susceptibility to oxidative stress, whereas IHCs tended towards structural and synaptic proteins. Although suggestive of functional specialisation, these patterns are supported by only a small subset of statistically significant proteins and therefore require further validation. Comparison with published transcriptomic datasets revealed little correspondence between RNA and protein fold changes, highlighting the importance of direct protein-level measurements for understanding hair-cell specialisation. Canonical MET proteins were largely undetected, likely owing to their low abundance and association with membrane-bound protein complexes. AP-MS did not identify a stable PCDH15 interactome, consistent with the intrinsically disordered nature of its cytoplasmic domain and the transient, low-affinity interactions it mediates. These findings suggest that alternative approaches capable of capturing interactions in situ, such as proximity-labelling strategies, may be required to characterise such dynamic protein assemblies. Overall, this work establishes single-cell proteomics as a viable approach for rare sensory cell types and provides a protein-level perspective on cochlear hair-cell identity. By defining both the potential and the current limitations of proteomic analysis in these cells, it lays a foundation for future studies aimed at understanding, preserving, and restoring hearing.