Acoustic overstimulation produces many anatomical, biochemical and physiological changes in the inner ear. However, the changes in gene expression that underlie these biological changes are poorly understood. Our approach to investigating this problem is to use gene microarrays to measure the changes in gene expression in the chinchilla inner ear following a 3 h or 6 h noise exposure (95 dB SPL, 707-1414 Hz). This noise exposure causes a temporary threshold shift (~40 dB) and a temporary reduction in distortion product otoacoustic emissions (DPOAE), but no permanent hearing loss or hair cell loss. Here, we present data showing (1) the suitability of mouse and human complementary DNA (cDNA) clones for detecting chinchilla cochlear gene transcripts, and (2) the change in cochlear gene transcripts in noise exposed chinchillas. Chinchilla cochlear transcript probes exhibited strong and discrete signals on both mouse and human cDNA filter arrays. Since the strongest hybridization occurred with mouse clones, mouse cDNA microarrays were used to study noise-induced changes in gene expression. Chinchilla cDNA probes were differentially labelled with Cy3 (control) or Cy5 (noise exposed) by random primed synthesis, hybridized to 8750 mouse cDNAs arrayed on microscope slides and analysed by laser fluorescent microscopy. Several classes of genes exhibited time-dependent up regulation of transcription, including those involved in protein synthesis, metabolism, cytoskeletal proteins, and calcium binding proteins. The results are discussed in relationship to previous studies showing noise-induced changes in structural proteins, calcium binding proteins, metabolic enzymes and membrane bound vesicles.

Acoustic overstimulation has very different outcomes in birds and mammals. When noise exposure kills hair cells in birds, these cells can regenerate and hearing will recover. In mammals, however, the hair cell loss, and resulting hearing loss, is permanent. Changes in gene expression form the basis for important biological processes, including repair, regeneration, and plasticity. We are therefore using a battery of molecular approaches to identify and compare changes in gene expression following noise trauma in birds and mammals. Both differential display and subtractive hybridisation were used to identify genes whose expression increased in the chick basilar papilla immediately following exposure to an octave band noise (118 dB, centre frequency 1.5 kHz) for 4-6 hr. Among those upregulated genes were two involved in actin signalling: the CDC42 gene encoding a Rho GTPase, and WDR1, which encodes a protein involved in actin dynamics. A third gene, UBE3B, encodes an E3 ubiquitin ligase involved in protein turnover. A fourth gene encodes a cystein-rich secreted protein that may interact with calcium channels. To examine the mammalian response, gene microarrays on nylon membranes (Clontech Atlas Gene Arrays) were used to examine global changes in gene expression 30 minutes after TTS (110 dB broadband noise 50% duty cycle) or PTS (125 dB, 100% duty cycle) noise overstimulation (each for 90 minutes) in the rat cochlea. Several genes, including classic immediate early response genes such as c-fos, EGR1/NGFI-A, and NGFI-B, were upregulated at this early time point following the PTS exposure, but were not upregulated following the TTS exposure.

Environmental inner ear insults often lead to hair cell injury and loss. Therapeutic measures for the prevention of hair cell loss are currently limited. Several reports have demonstrated the applicability of growth factors for hair cell protection. The goal of the experiments presented here was to assess the protective capability of the human GDNF transgene against noise trauma in the guinea pig cochlea. The left ears of guinea pigs were inoculated with a recombinant adenovirus with a human GDNF insert (Ad.GDNF). Four days later, animals were exposed to noise trauma. One week later, animals were sacrificed and hair cells counted in the left (inoculated) and right (non-inoculated) ears. Auditory brainstem thresholds were measured before the inoculation and just prior to sacrifice. Control groups included inoculation with a reporter gene vector (Ad.lacZ) and Ad.GDNF in normal ears with no noise exposure. The results show that intracochlear inoculation with adenovirus into normal ears does not compromise hair cell counts and ABR thresholds. Both Ad.GDNF and Ad.lacZ vectors can protect the cochlear hair cells and hearing from the noise insult. The difference between the protection afforded by Ad.GDNF and that of the Ad.lacZ vector is not statistically significant. The mechanism of Ad.lacZ protection needs to be elucidated. The data demonstrate the general feasibility of gene therapy for over-expression of neurotrophic factors against noise trauma, and emphasize the complexity of the technique and the problems of variability between subjects.

One consequence of noise exposure is increased production of reactive oxygen species (ROS), such as superoxide, hydrogen peroxide, and hydroxyl radicals, in the cochlea. ROS can cause oxidative damage to diverse cellular components, including membranes, proteins, and DNA, if they are not "neutralised" by antioxidant defences. Two important enzymes of the cochlear antioxidant defense system are cytosolic copper/zinc superoxide dismutase (SOD1) and selenium-dependent glutathione peroxidase (GPx1). These metalloenzymes work together to regulate ROS production in virtually every cell in the body, and they may be important for limiting cochlear damage associated with aging and acoustic overexposure. In this chapter, we describe a series of experiments using mice with targeted deletions of Sod1 or Gpx1, the mouse genes that code for SOD1 and GPx1, respectively, to study the cellular mechanisms underlying noise-induced hearing loss (NIHL). The results from Sod1 and Gpx1 knockout mice provide insights into the link between endogenous levels of antioxidant enzymes and susceptibility to NIHL.

Recent findings that glial cell line-derived neurotrophic factor (GDNF), neurotrophin-3 (NT­3), and transforming growth factor α can protect the auditory hair cells from acoustic trauma or aminoglycoside ototoxicity in vivo raise the question of whether other neurotrophic factors can also protect the hair cells in vivo. Fibroblast growth factor-2 (FGF-2) can protect hair cells from neomycin ototoxicity in vitro, and in vivo study has shown upregulation of FGF receptor­3 in the cochlea following noise exposure, suggesting that some FGF family members might play a role in protection or repair of the cochlea from damage. We therefore examined if FGF­1 and FGF-2 chronically delivered to the cochlea prior to noise overstimulation can attenuate noise-induced hair cell damage in vivo under conditions in which GDNF and NT-3 were effective. Pigmented female guinea pigs underwent left scala tympani implantation of a microcannula attached to an osmotic pump filled with artificial perilymph only or containing FGFs (10 or 1 µµg/ml FGF-1 or 10 µµg/ml FGF-2). They were exposed to noise (4 kHz octave band, 115 dB SPL, 5 hr) 4 days after surgery. Threshold shifts 10 days postexposure were essentially equivalent at all frequencies tested across different treatment groups. No significant difference in threshold shifts was observed between the treated and untreated ears in any of the groups. The extent of hair cell damage was also comparable among the different treatment groups. These findings indicate that exogenous FGF-1 or FGF-2 does not influence noise­induced hair cell damage under the experimental conditions used in this study, suggesting that these FGFs are not good candidates as auditory hair cell protectors in vivo.

This study examined the therapeutic effect of magnesium (Mg) on noise trauma in anesthetized guinea pigs exposed to an impulse noise series (1/s) of Lpeak 167 dB (Leq,1s 127 dB) for 38 min. The permanent hearing threshold shift (PTS) was measured 1 week post-exposure, using auditory brain stem response audiometry (frequency range, 0.5-32 kHz). The total Mg concentrations of perilymph, cerebrospinal fluid and plasma were analyzed by atomic absorption spectrometry. In a first series, animals maintained on physiologically low Mg received subcutaneous injections of either different Mg doses (0.11-0.33 mmol MgSO4/100 g per day) for 3 days and drinking water with an additive of 39 mmol MgCl2/l for 1 week or saline as a placebo and tap water alone. The treatment began immediately after the impulse noise exposure. The dose of 0.29 mmol Mg/100 g per day was found to be most effective and reduced the hearing loss by 13-20 dB compared to placebo. The PTS and the perilymph Mg level showed a close negative correlation, suggesting that the intracochlear Mg level plays an important role in bringing about these protective effects. In a second series, we tested the therapeutic efficacy as a function of the post-exposure time of onset of the optimal Mg treatment (1 min, 2 and 4 hours), using normal Mg animals. The therapeutic effect decreased with the length of time elapsed between the end of exposure and the beginning of treatment. In a parallel scanning electron microscopic test, we also found a Mg-related difference in the susceptibility of hair cell stereocilia to impulse noise exposure.