Most of our profound knowledge on oxidative stress induced by cisplatin comes from findings in different tumor cell lines. Several in vitro studies have also shown an increased production of reactive oxygen species (ROS) in rodent auditory neurons (Gabaizadeh et al. 1997), cochlear explants (Clerici et al. 1996), and auditory cell lines (Chang et al. 2013) following direct exposure to cisplatin. In general, very high concentrations of cisplatin have been used to study these redox effects. Even though in vitro studies offer well-characterized models for cisplatin-induced cell injury, their use can be limited by the fact that the vascular system and intracochlear transport systems of the drug is circumvented. For example, Laurell and coworkers showed that an equimolar dose of cisplatin and its analog oxaliplatin induced equal toxicity in OHC cultures, whereas in vivo, systemic cisplatin administration was ototoxic while an equimolar systemic dose of oxaliplatin was not (Hellberg et al. 2009). Another factor to consider when assessing the effects of cisplatin in vitro is the biotransformation of the drug, which may differ from the in vivo situation. To understand the complexity of oxidative stress pathways induced by cisplatin in the mammalian cochlea, different approaches are needed and findings that merge in vivo and in vitro studies have the potential to increase knowledge on the pathophysiology.

In vitro studies have suggested that cisplatin can interfere with the ubiquitous glutaredoxin system (Gabaizadeh et al. 1997; Sha et al. 2001; Kopke et al. 1997). Van De Water and colleagues showed in organ of Corti explants that cisplatin-induced cytotoxicity was associated with a reduction of GSH levels and an accumulation of hydrogen peroxide (H₂O₂) and that treatment with the GSH inhibitor l-buthionine sulfoximine enhanced the cytotoxic effects (Kopke et al. 1997). Cisplatin has a well-known propensity to react with thiols such as GSH (Gullo et al. 1980; Andrews et al. 1984), which may in part explain the reduced GSH levels induced by cisplatin.

So and coworkers showed that cisplatin caused cytotoxicity in HEI-OC1 auditory cells (Kim et al. 2010; So et al. 2007) and rat primary organ of Corti explants (Kim et al. 2010) through the generation of ROS, a likely source being NADPH oxidase 1 (NOX1) and/or NADPH oxidase 4 (NOX4) (Kim et al. 2010). NOX1 and NOX4 are members of the NOX family, which is composed of transmembrane isoenzymes that reduce oxygen to the free radical superoxide (O₂ ^•−) (Bedard and Krause 2007). Once produced, superoxide can generate other ROS or reactive nitrogen species (RNS), such as hydrogen peroxide (H₂O₂) by dismutation and peroxynitrite (ONOO⁻) by reaction with nitric oxide (Bedard and Krause 2007). Hydrogen peroxide can, in turn, readily generate the potent hydroxyl radical (^•OH) in the presence of transition metals, such as iron (Bedard and Krause 2007). Increased hydroxyl radical formation induced by cisplatin has been demonstrated by Clerici et al. in a study on cochlear explants (Clerici et al. 1996). So and colleagues also showed that the induction of NOX1 and NOX4 was due to upstream activation of proinflammatory cytokines, especially tumor necrosis factor-α (TNF-α), the activation of which was dependent on the extracellular-signal-regulated kinase (ERK) and nuclear factor κB (NF-κB) (Kim et al. 2010; So et al. 2007). Krause and coworkers have suggested that another member of the NOX family, NOX3, is important for cisplatin-induced ototoxicity; they found NOX3 to be highly expressed in the organ of Corti and spiral ganglia of rodents and that cisplatin enhanced the NOX3-dependent generation of superoxide in vitro (Banfi et al. 2004). In accordance with these findings, Rybak and colleagues discovered that cisplatin increased the expression of NOX3 and ROS in the organ of Corti hair cell line UB/OC-1 (Mukherjea et al. 2008). They subsequently showed that this cisplatin-induced ROS generation involved activation of the signal transducer and activator of transcription-1 (STAT1) (Kaur et al. 2011), a cytoplasmic transcription factor involved in signaling cascades initiated by cytokines and cellular stress. Besides STAT1, induced expression of the inflammatory mediators inducible nitric oxide synthase (iNOS), cyclooxygenase-2, and TNF-α as well as of the immune cell markers CD14 and CD45 was found (Kaur et al. 2011).

Besides the glutaredoxin system, the thioredoxin defense system plays an important role against oxidative stress. The importance of thioredoxin reductase as a potential molecular target in cisplatin ototoxicity has recently been shown in the organ of Corti cell culture (Dammeyer et al. 2014). The predominantly cytosolic isoenzyme thioredoxin reductase 1 has a reactive nucleophilic selenocysteine residue which is readily derivatized by cisplatin (Arner et al. 2001). This may be a desirable therapeutic effect in the case of cancer therapy but it is unwanted in the inner ear and may contribute to the ototoxic effect of cisplatin.

The homeostasis of the cochlea is dependent upon antioxidant system activity. Overproduction of ROS alters redox balance and signaling function, which may activate apoptosis mechanisms, eventually causing cell death and loss of function. Several experimental animal studies have found increased cochlear levels of the oxidative stress markers malondialdehyde (Whitworth et al. 2004; Teranishi et al. 2001; Ravi et al. 1995; Campbell et al. 2003a, b; Qu et al. 2012; Xiong et al. 2011; Rybak et al. 1995, 1999, 2000) and 8-iso-prostaglandin F2α (8-iso-PGF2α) (Qu et al. 2012) in rodents treated with cisplatin. But what are the underlying mechanisms?

Several in vivo studies suggest that cisplatin causes depletion of the glutaredoxin system. In studies on rodents, cisplatin reduced the cochlear levels of GSH (Ravi et al. 1995; Rybak et al. 1999, 2000; Lautermann et al. 1997) glutathione peroxidase (Ravi et al. 1995; Rybak et al. 1999, 2000), glutathione reductase (Ravi et al. 1995; Campbell et al. 2003a, b; Rybak et al. 1999, 2000), and glutathione S-transferase (GST) (Lautermann et al. 1997). Rybak and coworkers have suggested that cisplatin caused an overall decrease in GSH levels rather than an increased GSH oxidation rate, since they were unable to detect oxidized glutathione (i.e., glutathione disulfide) (Ravi et al. 1995), in agreement with a later study performed by the same research group (Rybak et al. 2000). The decrease in GSH could be due to the proclivity of cisplatin to react with GSH, as discussed in a previous section. Rybak and coworkers have also found reduced cochlear levels of superoxide dismutase and catalase in cisplatin-treated rats (Campbell et al. 2003a, b; Rybak et al. 1999, 2000), whereas González-García et al. instead found increased cochlear levels of superoxide dismutase in the same species (Gonzalez-Garcia et al. 2010). In the former studies, a much higher dose of cisplatin was administered than in the latter study, 16 versus 5 mg/kg. Possibly, when given in a very high dose, cisplatin exhausts the redox systems whereas when given in a more moderate dose, triggering antioxidant effects can instead be detected. However, in one study by Rybak and coworkers, increased levels of superoxide dismutase as well as catalase were found, for unknown reasons (Ravi et al. 1995).

As previously mentioned, the mammalian inner ear has been shown to express very high levels of NOX3 (Banfi et al. 2004). Rybak and colleagues demonstrated that large doses of cisplatin in rats caused elevated cochlear expression of NOX3 (Mukherjea et al. 2006, 2008, 2010). The increase could be localized to OHCs (Mukherjea et al. 2008, 2010), spiral ganglion cells (Mukherjea et al. 2008, 2010), stria vascularis (Mukherjea et al. 2008, 2010), and supporting cells (Mukherjea et al. 2008), thus structures that are known to be susceptible to damage by cisplatin. So and coworkers demonstrated elevated cochlear expression of NOX3 as well as of NOX1 and NOX4 in cisplatin-treated mice (Kim et al. 2010). The high levels of NOX1 and NOX4 were localized to the spiral ligament, spiral limbus, spiral ganglion neurons, OHCs, and IHCs, whereas low levels were found in the stria vascularis (Kim et al. 2010). It is noteworthy that the cellular site of ROS production is not established. Even though the OHCs are mostly damaged by cisplatin, ROS may be produced by different cochlear cell types as this is a general phenomenon of many mammalian cells.

Rybak and coworkers found that NOX3-generated ROS in cochleae from cisplatin-treated rats caused activation of STAT1 (Kaur et al. 2011). Cisplatin-induced STAT1 activation was found in OHCs, stria vascularis, and spiral ganglion cells (Kaur et al. 2011). Besides STAT1, cisplatin-induced expression of iNOS, cyclooxygenase-2, TNF-α, CD14, and CD45 was found (Kaur et al. 2011). In agreement with those results, So et al. found that cisplatin treatment caused elevation of TNF-α throughout the stria vascularis, spiral ligament, spiral limbus, modiolar spiral veins, and lacunae, and the organ of Corti in rats (So et al. 2007). The expression of the proinflammatory cytokines IL-1β, and IL-6 were also elevated but less ubiquitously (So et al. 2007). The elevated expressions likely involved activation of the transcription factor NF-κB (So et al. 2007).

iNOS and other NOS isoenzymes generate the free radical nitric oxide (^•NO) by metabolizing l-arginine to l-citrulline. Nitric oxide readily reacts with superoxide, producing the reactive peroxynitrite (ONOO⁻), which, in turn, can generate other, even more potent ROS/RNS, such as nitrosoperoxycarbonate (ONOOCO₂ ⁻) and the hydroxyl radical. Elevated nitric oxide in cochlear tissue following cisplatin treatment has been found in rodents (Xiong et al. 2011; Kelly et al. 2003). Using higher doses of cisplatin, Watanabe et al. ( 2000a, b) and Kaur et al. (2011) increased the cochlear expression of iNOS in rodents (Kaur et al. 2011; Watanabe et al. 2000a, b). The expression was mainly localized to the stria vascularis, spiral ganglion cells, and the organ of Corti, but not to the IHCs and OHCs (Watanabe et al. 2000a, b). It has also been shown that cisplatin treatment of mice induced the expression of iNOS in the lateral wall, an alteration which was accompanied by increased levels of NF-κB and, in stria vascularis, ssDNA (Watanabe et al. 2002).

In conclusion, findings from experimental animals show that cisplatin treatment can cause oxidative stress in the cochlea, e.g., due to induction of NOX, depletion of the glutaredoxin system, and/or induction of iNOS. The oxidative stress may be a consequence of toxic platinum-DNA adducts formed by cisplatin and/or MHC. However, oxidative stress can likely be generated independently of platinum-DNA adducts due to the avidity of cisplatin and/or MHC to react with nucleophiles involved in the cellular antioxidant defense, e.g., GSH.

There are no known methods to study specific biochemical toxic processes in the cochlea of patients during cisplatin treatment; our tools are limited to functional monitoring. Moreover, very few studies based on examinations of the human inner ear postmortem have been published. Thus, there is a lack of clear-cut evidence that cisplatin-induced hearing loss in patients is associated with oxidative stress. Thus, the primary evidence for the key role of oxidative stress in cisplatin-induced ototoxicity comes from animal laboratory studies.

There are some clinical data implying that the ototoxic effects of cisplatin might depend on interindividual variabilities in redox systems. The GST supergene family encodes isoenzymes that catalyze the detoxifying reactions of GSH. Some GST loci are polymorphic, demonstrating alleles that are null, encode low-activity variants, or are associated with variable inducibility. In a Norwegian study on testicular cancer long-term survivors that had received cisplatin-based therapy, Oldenburg et al. discovered that the genotype 105Val/105Val-GSTP1 was associated with better hearing than 105Ile/105Ile-GSTP1 and 105Val/105Ile-GSTP1, especially when combined with GSTT1*1 (wild genotype) and GSTM1*1 (Oldenburg et al. 2007). In a study on cisplatin-treated Asian pediatric cancer patients, Choeyprasert et al. found that subjects with GSTT1*1 had a significantly increased risk of hearing impairment compared to those with GSTT1*0 (null genotype), whereas no significant association was found with GSTM1 polymorphism (Choeyprasert et al. 2012). Peters et al. found that the GSTM3*B allele was more common among subjects with normal hearing after cisplatin therapy than among those with deteriorated hearing (Peters et al. 2000), while Ross et al. did not find any associations at all between GST polymorphism and hearing loss (Ross et al. 2009). It is possible that otoprotection by a certain genetic pattern is due to physiological functions not related to the inner ear, such as a higher clearance of cisplatin and/or MHC in the kidneys, leading to less cisplatin/MHC reaching the inner ear. Clearly, more clinical research is needed within this field.

Irreversible cochlear damage is a dose-limiting side-effect of the frequently used anticancer drug cisplatin. According to experimental animal studies, the damage involves several oxidative stress pathways, which may be targeted by administration of an antioxidant in conjunction with cisplatin therapy. A main challenge is to reduce the ototoxicity without interfering with cisplatin’s antineoplastic effects. Two major routes for administration of otoprotectors are plausible, a systemic route and a local route. Systemic administration of an antioxidant seems less attractive in systemic cisplatin therapy owing to the risk for lowering of the antineoplastic effects unless the cancer disease is highly confined. Local administration of an otoprotective agent or a combination of substances offers the possibility of protecting the cochlea without compromising the antitumoral efficacy, according to preclinical data, and needs to be studied in randomized trials. Studies are ongoing investigating the clinical applicability of this approach in humans (Table 11.1).