Research over the last decades has identified a major role for reactive oxygen species (ROS) and oxidative stress in hearing loss, including overexposure to noise, ototoxic drugs (e.g. cisplatin), and age-related hearing loss. While the sources of ROS are diverse, NADPH oxidases (NOX) have been identified as the only family of enzymes dedicated to the production of ROS. In particular NOX3 is highly and specifically expressed in the inner ear and its expression has been shown to be dramatically increased in the cochlea following insult. The present work is based on the hypothesis that inhibiting NOX3 related ROS production, associated with presbycusis, noise trauma, and cisplatin ototoxicity, would result in an otoprotective effect.
In order to test this hypothesis, in the first part of this thesis, we developed preclinical models of noise, age and cisplatin-induced hearing loss. As systemic delivery of cisplatin at ototoxic doses results in important morbidity in the mouse, we developed an alternative model through direct middle ear delivery of cisplatin, in line with the 3R principles. Interestingly, this novel model recapitulated the main functional and histological correlates of cisplatin ototoxicity without any systemic recirculation of cisplatin, thus with a major positive impact on animal welfare. Using NOX3 deficient mice, preclinical models allowed to demonstrate the involvement of NOX3 in the pathogenesis of the above-mentioned acquired forms of hearing loss. Indeed, mice devoid of NOX3 activity exhibited remarkable preservation of hearing and cochlear histology compared to wild type littermates, identifying NOX3 as a prime target to combat hearing loss.
In the second part of this thesis, we developed genetic tools to silence pathogenic NOX3 expression as a putative otoprotective treatment. As acquired hearing loss can be acute (i.e. noise trauma) or chronic (i.e. presbycusis), both small interfering RNA (siRNA) and vector based molecular tools were developed to respectively achieve a transient or long term NOX3 silencing. Genetic therapies were directly delivered in the inner ear through the posterior semi-circular canal. First, we screened 10 siRNA targeting different parts of the NOX3 mRNA sequence in NOX3 expressing cell lines. Finally, the most potent siRNA identified in vitro, was further tested in vivo, allowing robust silencing of NOX3 expression in the mouse cochlea. Interestingly, this sequence is conserved between mouse, guinea pig, and human, facilitating an eventual translation.
To date, the absence of viral vectors able to achieve efficient transduction in the adult cochlea limits the options for therapeutic transgene delivery. By using RNAscope in situ hybridization, NOX3 expression was primarily found in the Rosenthal canal area, colocalizing with auditory neurons. Based on this knowledge, we characterized the cochlear tropism of several AAV derived vectors in vivo in the adult cochlea and identified new candidates with promising tropism and expression in the inner ear.
In conclusion, the data presented along this thesis demonstrates the rationale of targeting NOX3 and provides new insight about molecular treatment and genetic therapies against acquired forms of hearing loss.