Discussion
This study further investigated the biological effects induced by DFNA5 expression in Saccharomyces cerevisiae and in human HEK293T cells. A previous study in human cell lines revealed that mutDFNA5 is a cell death-inducing gene (Op de Beeck et al., 2011). As described previously (Van Rossom et al., 2012), transformation of mutDFNA5 in yeast resulted in a growth defect associated with four different mitochondria-related proteins. This study was the first to establish a possible link between the mitochondria and DFNA5. In addition, the 2012 study observed that the protein quality control system, responsible for correct protein folding and degradation, had problems coping with mutDFNA5. A failing protein quality control system could indicate the presence of ER stress, as the ER can be involved in protein folding and degradation.
This study confirmed the involvement of mitochondria-related processes upon expression of mutDFNA5 in Saccharomyces cerevisiae, especially the ATP-coupled electron transport. Several genes related to either the glycolysis and the PPP were significantly down-regulated upon mutDFNA5 transformation in yeast. Furthermore, we show that the JNK and ERK MAPK pathways are activated in vitro after transfection of mutDFNA5 in HEK293T cells and that inhibition of this pathway is able to partially attenuate the resulting cell death. Additionally, this study also revealed an association of GO annotations related to the ER and protein folding in both model organisms.

Oxidative stress and hearing loss
The up-regulation of different cytochrome c oxidase (COX) genes revealed a potential association of the mutDFNA5-related cell death mechanisms with oxidative stress in yeast. Increased oxidative stress was already demonstrated in yeast (Van Rossom et al., 2012) and was later confirmed in human cell lines (unpublished results). Enhanced COX activity has indeed been associated with increased oxidative stress (Singh et al., 2009; Srinivasan and Avadhani, 2012). COX is the rate limiting enzyme of respiration which regulates the bio-energetic status of the cell. Dependent on the cellular energetic requirements, the COX activity can be rapidly adapted. The ratio of ATP/ADP is one of the regulators of the COX activity. High ADP levels or ATP utilization will increase the enzyme activity and stimulate the respiration (Napiwotzki and Kadenbach, 1998; Arnold and Kadenbach, 1999). Activation of the COX activity could result in higher oxidative stress generated at the mitochondria.
Additionally, the correlation of mutDFNA5 with several genes of the PPP can also be linked to oxidative stress in yeast as the PPP plays a major role in the anti-oxidant response. Reduced expression of CDC19, the yeast homolog of PK, and of TPI has been correlated with the activation of the PPP and the redirection of the metabolic flux from glycolysis to PPP both in human cell lines and in yeast (Christofk et al., 2008; Gruning et al., 2011, 2014). This study revealed reduced gene expression of both TPI and CDC19, indicating a shift in redox sensing in eukaryotes mediating a fast response to oxidative stress. Activation of the PPP is correlated with the inhibition of ROS accumulation and enhancement of the anti-oxidant response upon shift from fermentation to respiration. PPP activation will enhance the anti-oxidant response and hence increase the tolerance for oxidative stress (Ralser et al., 2007; Gruning et al., 2011; Kruger et al., 2011). These data clearly demonstrate a change in redox homeostasis due to mutDFNA5 expression which was shown previously in yeast (Van Rossom et al., 2012).
Furthermore, enhanced oxidative stress is often related to a failing protein quality control system (Davies, 2001; Shang et al., 2001; Bender et al., 2010, 2011; Shang and Taylor, 2011). This possibility was already suggested by the proteolytic degradation previously seen upon wtDFNA5 transformation, but which was absent in mutDFNA5-transformed yeast cells (Van Rossom et al., 2012). MutDFNA5 seemed to escape this quality control system, in contrast to wtDFNA5 which was subject to the normal clearance system. Moreover, the authors suggested a possible link between protein degradation and the mitochondria in mutDFNA5-induced cell death as yeast seemed to have problems with proper mutDFNA5 protein turnover. The same link was confirmed in both model organisms used in the current study.

Role of ER stress in the pathology of hearing loss
The increase in mitochondrial metabolism seen in this study and the decrease of protein folding processes, can also be explained by the presence of cellular ER stress leading to an unfolded protein response. Mitochondria and ER form an interconnected network which is important for several biological processes mediating an adaptive response under various cellular stress conditions (De Brito and Scorrano, 2010; Marchi et al., 2014). The association between mutDFNA5 and the GO terms related to lipid metabolism, protein targeting to ER and the ER membrane, suggest the presence of ER stress (Schroder, 2008). Mitochondria depend on the ER for the import of several proteins and lipids and for Ca2+ exchange involved in cell death and mitochondrial metabolism (Sauner and Levy, 1971; Zecchini et al., 2007; Stiban et al., 2008; Wiel et al., 2014). Enhanced Ca2+ supply will increase ATP production and mitochondrial respiration, processes which were indeed both up-regulated at the post-diauxic shift upon mutDFNA5 expression. Prolonged enhancement however will eventually have a detrimental effect on the mitochondria.
Despite limited knowledge correlating mutDFNA5 with ER stress, a correlation has been established between ER stress and certain causes of HL. Ototoxicity (HL due to the use of pharmaceuticals such as aminoglycoside antibiotics and platinum-based chemotherapeutics) was shown to be correlated with ER stress-dependent pathways. Certain pain relievers, contributing to tinnitus and progressive bilateral sensorineural HL, were shown to induce ROS overproduction, altered ER morphology and changes in ER stress markers, such as CHOP (Kalinec et al., 2014).
Taken together, the previous observation of the importance of the mitochondria in mutDFNA5-related cell death and the known correlation between the ER and the mitochondria points to a potential role for the ER in DFNA5-induced cell death. The failing of the protein quality control system in mutDFNA5 suggests the involvement of the ER, but this remains unclear at this moment and needs to be further investigated in the future.

Contribution of the mitochondria in MAPK-related cell death
In addition to the mitochondria, the MAPK pathways seem to play a prominent role in mutDFNA5-induced cell death in HEK293T cells. The link between the mitochondria and MAPK in DFNA5-related cell death is unknown at this moment, but several mitochondria-MAPK correlations have been described. It is known that MAPK pathways can be induced by ROS production generated by the mitochondria (Chambers and Lograsso, 2011). As increased oxidative stress has been shown in an earlier study in yeast and in human cell lines (unpublished results), this can provide a direct link between these two processes. This could suggest that ROS activates the MAPK pathway and hence lays up-stream of the MAPK pathway in DFNA5-induced cell death. However, we showed in this study that a specific MAPK inhibitor was able to attenuate the cell death and we have unpublished results showing that several anti-oxidants did not inhibit DFNA5-induced cell death. This implies that ROS is either a secondary event not directly causing cell death, or that the activation of MAPK is an early event in the cell death process, up-stream of ROS production. Furthermore, comparison of yeast and the human microarray results identified UCP2, a gene associated with the mitochondria, which showed reduced gene expression in mutDFNA5 compared to wtDFNA5. Interestingly, mitochondrial stress has been linked to the down-regulation of UCP2 by activation of the MAPK pathway and of JNK activation in particular (Emre et al., 2007; Selimovic et al., 2008). UCP2 reduction was an early event required for the amplification of the activated MAPK pathway enabling mitochondrial ROS production (Emre et al., 2007; Basu Ball et al, 2011). This down-regulation enables mitochondrial ROS production providing the amplification loop stimulating the MAPK pathway. UCP2 can therefore provide the link between MAPK and the mitochondria, regulating the ROS production, a feature deregulated in both model organisms.
However, we did not observe an up-regulation of MAPK-related pathways in yeast. This could be explained by the differences in timing between the experiments. For human cell lines, the experiment was performed 12 h post-transfection, while yeast RNA was collected at the post-diauxic shift.
In conclusion, this study confirms the role of the mitochondria in mutDFNA5-induced toxicity in yeast. Additionally, it shows that mutDFNA5-induced cell death is mediated by the MAPK pathway, especially through the ERK and the JNK branch. Inhibition of this pathway could significantly enhance the cell viability of mutDFNA5-transfected HEK293T cells which suggests the importance of this signaling cascade for DFNA5. How the MAPK pathways perform their role in cell death and the connection with DFNA5-related mechanisms remains uncertain at this moment but is a good starting for future studies. Future studies further unraveling the DFNA5-induced cell death mechanism are important as they may lead to new insights in the involvement of mitochondria in HL and have the potential to lead to new future therapies.

Conflict of interest statement
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Joris Winderickx declares that he is co-founder of the KU Leuven spin-off companies reMYND and ADxNeuroSciences, but this did not influence study design, data collection, analysis, publication or the preparation of the manuscript. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.