Introduction
Cell death is a fundamental process of all organisms and inherent to life. The role of programmed cell death (PCD) in the pathology of hearing loss (HL) has been well studied and seems to play a prominent role, especially in the development of the vertebrate inner ear and in the morphogenesis of the semicircular canals (Fekete et al., 1997; Nishikori et al., 1999; Leon et al., 2004). Several genes related to PCD-induced HL have been identified and many of these are related to the mitochondria (Estivill et al., 1998; Jacobs et al., 2005; Ding et al., 2013). Mitochondria are key players during PCD and dysfunction of the mitochondria has been linked to the pathogenesis of many diseases, including HL. Mitochondria are the main producers of cellular ATP due to oxidative phosphorylation. However, this process also generates reactive oxygen species (ROS) (Raha and Robinson, 2000). Due to the mitochondrial production of ROS, the reduced DNA repair capacity and the close proximity of mtDNA to ROS generation sites, mtDNA (mitochondrial DNA) is very susceptible to mutations. Enhanced oxidative stress will have detrimental effects on cellular health and plays a major role during aging, mutagenesis, and cell death (Gredilla and Barja, 2005; Maynard et al., 2009; Gredilla et al., 2010).
Due to the relative slow speech perception detoriation, it is assumed that DFNA5-related HL is due to cochlear dysfunction. The cochlea seems to be highly susceptible to the detrimental effects of mitochondrial damage due to the post-mitotic character of its sensory epithelium (Sha et al., 2001).
It is estimated that in the Caucasian population at least 5% of the post-lingual non-syndromic HL is due to mutations in the mtDNA (Estivill et al., 1998; Jacobs et al., 2005). Most mutations in mtDNA affect the mitochondrial MTRNR1 and the MTTS1 genes encoding respectively a mitochondrial 12S rRNA and a tRNASer. Mutations in these mitochondrial genes lead to variable clinical severity of HL due to impaired mitochondrial tRNA metabolism and protein synthesis (Casano et al., 1999; Fischel-Ghodsian, 1999; Jin et al., 2007; Ding et al., 2013; Dowlati et al., 2013). Also many of the nuclear DNA mutations leading to HL are related to mitochondrial dysfunction and PCD. These include genes such as OPA1, TIMM8A, SMAC/DIABLO, MPV17, PDSS1, BCS1L, SUCLA2, C10ORF2, COX10, PLOG1, and RRM2B (Roesch et al., 2002; Antonicka et al., 2003; Payne et al., 2004; Mollet et al., 2007; Cheng et al., 2011; Meyer Zum Gottesberge et al., 2012; Luo et al., 2013). These genes contribute to various fundamental mitochondrial aspects, such as mitochondrial protein transport, mitochondrial fragmentation, and oxidative phosphorylation. Mutations in these genes are thought to induce mitochondrial stress and trigger cell death in the cochlea leading to HL. The explanation for the tissue specific effect of these genes leading to more prominent cell death in the cochlea and the hair cells remains unknown at this moment.
In addition to the genes directly associated with the mitochondria, two other genes related to apoptosis, a specific form of PCD, have been linked with hearing loss. MSBR3, a gene encoding a methionine sulfoxide reductase, is associated with caspase-3 activity. This can initiate apoptosis, eventually leading to degeneration of the inner hair cells and recessive non-syndromic HL (Ahmed et al., 2011). Overexpression of the TJP2 gene, due to a genomic duplication, altered the expression of several apoptotic genes, inducing a dominant non-syndromic form of HL (Walsh et al., 2010).
From these examples, it is clear that a prominent association exists between (mitochondria-related) PCD and non-syndromic HL. Many forms of HL, such as age-related hearing impairment (ARHI), noise-induced hearing loss (NIHL), monogenic forms of HL and ototoxicity, have been associated with dysfunctional mitochondria and PCD, underscoring the importance and the need to further investigate the role of PCD in the pathogenesis of deafness (Casano et al., 1999; Sha et al., 2009; Someya et al., 2009; Liu et al., 2010; Chen et al., 2012).
One of the monogenic deafness genes that is related to PCD is DFNA5. DFNA5 was originally identified as a gene responsible for an autosomal dominant form of HL in a Dutch family (Van Laer et al., 1998). Today, eight families with mutations in DFNA5 (mutDFNA5) associated with HL have been identified. The phenotype of the HL is very similar with the exception of the age-of-onset which varies from 15 to 50 years old. The HL is symmetric and starts in the high frequencies, but spreads to the lower frequencies in a later stage. Four of the eight mutations differ at the genomic level, but they all result in skipping of exon 8, yielding an immature truncated protein (Yu et al., 2003; Bischoff et al., 2004; Cheng et al., 2007; Park et al., 2010; Nishio et al., 2014). The structure of DFNA5 is unknown at this moment, but hydrophobic cluster analysis revealed that DFNA5 consists of two regions, domain A and domain B, connected by a hinge region (Op de Beeck et al., 2011). Mutations in DFNA5 result in a truncated protein lacking the last part of domain B due to a premature stop codon. This premature stop codon is caused by skipping of exon 8 as a result of the different genomic mutations in DFNA5 present in the families. Op de Beeck et al. (2011) have also demonstrated that only the first part, domain A, present in both wild-type and mutDFNA5, is sufficient to induce cell death after transfection in human cell lines. Domain B did not have any cell death inducing capacity. This led to the hypothesis that domain B will normally shield this cell death-inducing domain A to avoid inappropriate activation of domain A. Due to the partial lack of domain B in mutDFNA5, this shielding may not be possible, leading to a constitutive activation of mutDFNA5 (Op de Beeck et al., 2011).
In addition to HL, wild-type DFNA5 (wtDFNA5) has also been correlated with several forms of cancer, such as breast, colorectal, hepatocellular, and gastric cancer. Endogenous DFNA5 is epigenetically silenced by hypermethylation in cancer cells resulting in a decreased DFNA5 expression level. Based on these findings, it is hypothesized that DFNA5 is a tumor suppressor gene (TSG) (Akino et al., 2007; Kim et al., 2008a,b; Wang et al., 2013).
The function of DFNA5 remained unknown for a long time, but previous functional studies by Op de Beeck et al. (2011) revealed that DFNA5 induces a growth defect in mutDFNA5-transfected HEK293T cells, as well as other cells, leading to PCD (Op de Beeck et al., 2011). The cell death-inducing capacity of DFNA5 was not only restricted to human cell lines, but was also observed in the yeast model Saccharomyces cerevisiae (Van Rossom et al., 2012). This inspired us to use these two different model organisms to further elucidate the mechanisms related to DFNA5.
The Saccharomyces cerevisiae yeast model has several advantages as a model organism compared to human cell lines. The rapid growth, inexpensive media and the relatively easy genetic modifications, made the budding yeast Saccharomyces cerevisiae a valuable model to unravel regulators of different human pathologies (reviewed in Winderickx et al., 2008). Thirty percent of known genes involved in human diseases have an ortholog in yeast and due to this high degree of conservation, yeast is very suitable for fundamental research to identify core regulators of diverse signaling mechanisms (Foury, 1997). This was demonstrated in particular for mechanisms related to PCD where numerous yeast homologs of human genes related to cell death have been identified to play a common role (Wissing et al., 2004; Buttner et al., 2007; Madeo et al., 2009). The conservation of PCD mechanisms has been confirmed for DFNA5-related PCD in a previous study using the budding yeast Saccharomyces cerevisiae (Van Rossom et al., 2012). This study demonstrated the value of yeast to unravel PCD mechanisms-related to human genes. Transformation of mutant DFNA5 (mutDFNA5) in yeast led to the induction of PCD. Several mitochondrial proteins, such as Fis1, Por1, Aac1 and Aac3, were shown to be involved. Moreover, mutDFNA5 was found to co-localize with a mitochondrial marker protein (Westermann and Neupert, 2000). These former observations established a role for mitochondria in DFNA5-related cell death in yeast and demonstrated the value of yeast as a model organism to unravel DFNA5-related (mitochondrial) HL.
In the current study, we performed a transcriptomic analysis and confirmed the significance of mitochondria in DFNA5-induced cell death in Saccharomyces cerevisiae. Additionally, Gene Ontology (GO) analysis suggested a role for the endoplasmic reticulum (ER). The latter observation was not only present in Saccharomyces cerevisiae, but was also confirmed in human cell lines. Furthermore, we show that the MAPK pathways, namely the induction of the extracellular signal regulated kinase (ERK) and c-Jun N-terminal kinase (JNK), are activated upon mutDFNA5 transfection in human cell lines. Our data suggest the presence of a cellular adaptive response related to mitochondria, MAPK pathways and potentially for the ER. The exact correlation between those processes and DFNA5 remains unclear but further study will lead to a better understanding of DFNA5-induced cell death mechanisms and to HL-related to mitochondrial damage in general.