The PERK/ATF4/CHOP signaling branch of the unfolded protein response mediates cisplatin-induced ototoxicity in hair cells

Abstract Cisplatin is a widely used chemotherapeutic agent. However, its clinical application remains limited due to the high incidence of severe ototoxicity. It has been reported that the unfolded protein response (UPR) is involved in cisplatin-induced ototoxicity. However, the specific mechanism underlying its effect remains unclear. Therefore, the present study aimed to explore the sequential changes in the key UPR signaling branch and its potential pro-apoptotic role in cisplatin-induced ototoxicity. The hair cell-like OC-1 cells were treated with cisplatin for different periods and then the expression levels of the UPR- and apoptosis-related proteins were determined. The results showed that the apoptotic rate of cells was gradually increased with prolonged cisplatin treatment. Furthermore, the sequential changes in three UPR signaling branches were evaluated. The expression levels of activating transcription factor 4 (ATF4) and C/EBP homologous protein (CHOP) were gradually increased with up to 12 h of cisplatin treatment. The aforementioned expression profile was consistent with that observed for the apoptosis-related proteins. Subsequently, the proportion of apoptotic cells was notably decreased in CHOP-silenced hair cell-like OC-1 cells following treatment with cisplatin. Moreover, we found significant hair cells loss and a higher level of CHOP in cisplatin-treated cochlear explants in a time-dependent manner. Overall, the present study demonstrated that the protein kinase RNA‑like endoplasmic reticulum kinase (PERK)/ATF4/CHOP signaling branch could play an important role in cisplatin-induced cell apoptosis. Furthermore, the current study suggested that CHOP may be considered as a promising therapeutic target for cisplatin-induced ototoxicity.


Introduction
Cisplatin is an antitumor drug commonly used in the clinic. However, it has been reported that 40%-80% of adults and >50% of children treated with platinum-based chemotherapy experience varying degrees of irreversible hearing loss (Breglio et al. 2017). Cisplatin-induced ototoxicity seriously affects the quality of life of these patients and provides a heavy socioeconomic burden to the families of these children (van As et al. 2018). It has been previously demonstrated that DNA damage, inflammatory signals and oxidative stress are involved in the occurrence of cisplatin-induced ototoxicity (Kros and Steyger 2019). However, The US Food and Drug Administration has not yet approved a drug that can effectively prevent or treat cisplatin-induced hearing loss.
It has been reported that cisplatin can directly interfere with the normal folding process of immature proteins in the endoplasmic reticulum (ER), thus leading to the toxic accumulation of unfolded proteins in the ER cavity, eventually triggering the unfolded protein response (UPR) (Chen et al. 1994, Ahmed-Ouameur et al. 2006, Palm et al. 2011, Sancho-Martinez et al. 2012, Wang and Kaufman 2016. Protein kinase RNA-like endoplasmic reticulum kinase (PERK), inositol-requiring enzyme 1a (IRE1a) and activating transcription factor 6 (ATF6) are three transmembrane sensors located on the ER membrane. PERK, IRE1a, and ATF6 can be activated in response to the accumulation of unfolded or misfolded proteins in the ER. Following their activation, these sensors trigger a signaling cascade, which in turn produces a series of complex biological effects in the cell, including improved cell protein homeostasis and the induction of apoptosis (Gardner et al. 2013, Iurlaro and Munoz-Pinedo 2016, Hetz and Papa 2018. It is worth noting that the relative timing of the three UPR signaling branches can determine cell fate (Walter et al. 2015). In addition, the likelihood of ER stress induction and the ability to activate the UPR may vary depending on the cell type and the pathophysiological stimuli to which the UPR pathway-related proteins respond to (Mollazadeh et al. 2018).
A previous study from our laboratory showed that the UPR and UPR-induced apoptosis could represent a novel mechanism underlying cisplatin-induced ototoxicity (Zong et al. 2017). Meanwhile, it has been suggested that cochlear hair cells are susceptible to cisplatin-induced cytotoxicity, and the loss of hair cells is closely associated with hearing loss (Kros and Steyger 2019). However, how UPR works in cochlear hair cells remains unknown. Thus, exploring the timing and dynamics of the UPR signaling branches in hair cells could improve the current understanding of the role of the UPR in cisplatin-induced ototoxicity, and facilitate the discovery of potential novel therapeutic targets for cisplatininduced ototoxicity.
The present study aimed to investigate the sequential changes in the expression of key molecules involved in the UPR signaling branches, including ATF4 and C/EBP homologous protein (CHOP) in the PERK signaling branch, apoptosis signal regulating kinase 1 (ASK1) and thioredoxin-interacting protein (TXNIP) in the IRE1a signaling branch, and the ATF6 signaling branch in cisplatin-treated hair cell-like OC-1 cells.
In addition, the current study sought to determine the potential signaling branch involved in the development of ototoxicity.

Cell viability assay
Cell viability was tested by Cell Counting Kit-8 (CCK-8, Dojindo, Japan) according to the manufacturer's instructions. Briefly, cells were seeded at a density of 10 3 cells/well into 96-well plates and exposed to different concentrations of cisplatin for 36 h to determine the optimal concentration for the remaining experiments. 90 lL DMEM medium and 10 lL CCK-8 reagent were added to each well at the endpoint of the culture. After incubation for 1 h at 37 C in the dark, the optical density (OD) value at 450 nm was detected by a microplate reader (Infinite F50, TECAN, Switzerland). Cells without cisplatin were used as the control group, and the corresponding wells without cells constituted the blank group. Cell viability rate¼(OD of the experimental group À OD of the blank group)/(OD of the control group À OD of the blank group). Then, the cell viability of OC-1 cells treated with 10 lM cisplatin at 12, 24, 36, and 48 h was determined separately by the same method mentioned above.

Cell apoptosis analysis
The apoptotic rate of OC-1 cells was determined using an Annexin V-FITC/PI Apoptosis Detection Kit (KGA107, KeyGen Biotech, China). Cells cultured in 6-well plates were treated with 10 lM cisplatin for 12, 24, 36 and 48 h. At the endpoint of the culture period, 200 mL of binding buffer containing 5 mL of Annexin V-FITC and 4 mL of PI was added to each sample of collected cells. Cell apoptosis was detected with a FACS Calibur flow cytometer (Becton Dickinson and Company, USA), and data was analyzed using FlowJo V10 software. The apoptotic rate was calculated as the sum of early apoptotic (Annexin V þ PI -) cells and late apoptotic (Annexin V þ PI þ ) cells.

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR)
Total RNA from OC-1 cells was extracted by the TRIzol reagent (15596018, Invitrogen, USA) following the manufacturers' instructions. Total RNA was quantified with a Nanodrop spectrophotometer (Thermo Fisher Scientific, USA). The extracted RNA was reverse transcribed into cDNA using the PrimeScript RT Master Mix kit (RR036A, TaKaRa, Japan). Subsequently, qPCR was performed using TB Green Premix Ex Taq (RR420A, TaKaRa, Japan) with an ABI StepOnePlus Real-Time PCR system (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions. The PCR condition consisted of initial denaturation step at 95 C for 30 sec, followed by 40 cycles of 95 C for 30 sec, 60 C for 45 sec, and 72 C for 30 sec. The sequence of the primers was shown in Table 1. The relative level of gene expression was calculated using the 2 À DDCq method. The expression level was normalized to that of the house-keeping gene, b-actin. The RT-qPCR experiments were not only done to test the efficacy of the RNAi but also to assess relative timing of pathway activation.

Tunel staining
Apoptotic cells were observed by using the One Step TUNEL Apoptosis Assay Kit (C1086, Beyotime, China). After cisplatin treatment, cells were rinsed by PBS and fixed with 4% paraformaldehyde (G1101, Servicebio, China) for 30 min. 0.3% Triton X-100 (P0096, Beyotime, China) was then applied for 5 min at room temperature. After incubating with the TUNEL reagent in the dark for 1 h at 37 C, cells were stained with DAPI for 5 min. Apoptotic cells were labeled in green fluorescence. Images of multiple cells from three independent experiments were obtained with a fluorescence inversion microscope (Olympus IX71, Japan).

Cochlear explant culture
The cochlear explants were dissected from SD rats at postnatal day 3 and cultured as previously described (Chen et al. 2013, He et al. 2017. Briefly, the animals were kept under specific pathogen-free (SPF) conditions and monitored for health status daily. Neonatal rats (P3, n ¼ 15) were sacrificed by cervical dislocation after euthanasia with pentobarbital solution (50 mg/kg, i.p, Sigma Aldrich, USA

Statistical analysis
Statistical analysis was conducted using GraphPad Prism 7 software (GraphPad Software Inc., USA). All data was presented as the mean ± standard deviation (SD) from at least three independent experiments. The statistical significance among multiple groups was determined using One-way ANOVA, followed by Tukey's test (for comparison with all groups) or Dunnett's test (for results comparing the 0 h group to other groups). The difference was considered significant when P < 0.05.

Cisplatin exerts its cytotoxic effects on OC-1 cells in a time-dependent manner
The results of the CCK-8 assay showed that the cell viability was decreased in response to increasing concentrations of cisplatin (0-80 lM). More specifically, a significant decrease in viability was observed in cells treated with 10 lM cisplatin for 36 h, with a survival rate of 54.33 ± 7.32% (P < 0.05; Figure  1(a)). To further explore the sequential changes in cell viability following treatment with 10 lM cisplatin for 0-48 h, the cytotoxic effects of cisplatin on OC-1 cells were evaluated. Prolonged treatment of OC-1 cells with 10 lM cisplatin gradually decreased cell viability in a time-dependent manner. The survival rate was 32.7 ± 3.3% at 48 h following cisplatin treatment (Figure 1(b)). In addition, the Annexin V/PI apoptosis assay results showed that the proportion of apoptotic cells was gradually increased with prolonged cisplatin treatment (Figure 1(c,d)). This finding was consistent with those obtained using the CCK-8 assay.

Sequential changes in the expression of UPR-and apoptosis-related proteins in cisplatin-treated OC-1 cells
To investigate the sequential expression changes of key UPRrelated molecules, we performed RT-qPCR analysis and found that the mRNA expression levels of ATF4, CHOP and TXNIP were notably elevated in a time-dependent manner after cisplatin treatment, while those of ASK1 and ATF6 were not significantly altered (Figure 2(a)). Consistent with the mRNA results, western blot analysis further verified the timedependent changes in the protein expression levels of ATF4 and CHOP after cisplatin treatment, while TXNIP, XBP1s and p-ASK1/ASK1 showed no significant change until the 48 h cisplatin treatment (Figure 2(b,c)). In addition, after treatment of OC-1 cells with cisplatin for 48 h, the expression levels of Bax and cleaved caspase-3, two key apoptosis-related proteins, were significantly increased in a time-dependent manner, while the expression levels of the anti-apoptotic protein Bcl-2 exhibited the opposite trend (Figure 2(d,e)). Together, the expression profile of ATF4 and CHOP was consistent with that observed for the apoptosis-related proteins.

Early changes in the activation of UPR signaling in cisplatin-treated OC-1 cells
To further investigate the sequential changes in the activation of UPR signaling within the early 12 h of cisplatin treatment, the mRNA and protein levels of UPR-related key molecules were determined. At the point of 12 h, ATF4 and CHOP started responding to cisplatin stimulation at the mRNA level, while no significant changes were observed in the ATF6 and XBP1s branches (Figure 3(a)). In the protein level, TXNIP and ASK1 did not respond to cisplatin in the early period. The protein expression levels of CHOP were also notably elevated at 12 h, whereas the expression levels of apoptosis-related proteins were not significantly altered (Figure 3(b,c)). In addition, the expression levels of the UPR signaling-related molecules remained unchanged in cells cultured in medium without cisplatin for 12 and 48 h, suggesting that the UPR was affected by cisplatin ( Supplementary Figure 1).

Atf4/CHOP signaling branch mediates cisplatininduced apoptosis in OC-1 cells
To evaluate the role of CHOP in cisplatin-induced apoptosis, OC-1 cells were transfected with si-CHOP to knockdown CHOP expression in cells. An interference efficiency of approximately 80% was obtained (Supplementary Figure 2). Following treatment with cisplatin for 36 h, the results of the CCK-8 assay demonstrated that CHOP knockdown ameliorated cisplatin-induced cytotoxicity in OC-1 cells. The relative cell viability in the cisplatin þ si-CHOP group was 56.44 ± 4.69% compared with 74.34 ± 2.27% in the cisplatin þ si-NC group (P < 0.001; Figure 4(a)). Furthermore, the apoptotic rate in the cisplatin þ si-CHOP group was significantly lower compared with that in the cisplatin þ si-NC group (11.28 ± 1.28% vs. 19.16 ± 1.40%; P < 0.01; Figure   4(b,c)). In addition, CHOP silencing inhibited cisplatin-induced OC-1 cell apoptosis, as evidenced by the reduced proportion of TUNEL-positive cells (Figure 4(d,e)). Subsequently, the effect of CHOP knockdown on the mitochondrial apoptotic pathway was investigated. The expression levels of the proapoptotic proteins, Bax, cleaved caspase-9 and cleaved caspase-3 were notably reduced, while those of the antiapoptotic protein Bcl-2 were increased (Figure 4(f,g)), indicating that CHOP may serve as an essential pro-apoptotic target of the UPR signaling branches in cisplatin-induced damage.

Sequential expression changes of CHOP in cisplatintreated apoptosis in cochlear explants
Furthermore, we investigated the sequential expression changes of CHOP in cochlear explants. Based on the analysis of the immunofluorescence staining, we found that the number of phalloidin-labeled hair cells was decreased with prolonged cisplatin treatment (Figure 5(a)). Notably, the positive rate of CHOP in hair cells was gradually upregulated in cochlear explants as the duration of cisplatin treatment increased, which exhibited the same changes in the expression of CHOP as the OC-1 cells (Figure 5(b)). In conclusion, both the results of OC-1 cells and cochlear explants revealed that PERK/ATF4/ CHOP signaling branch played an essential role in cisplatininduced ototoxicity (Figure 6).   . Schematic diagram displaying the specific regulatory role of the UPR in cisplatin-induced ototoxicity and the major signaling branch exerting pro-apoptotic effects. ATF6, IRE1a, and PERK are the three main sensors of the UPR, among which PERK and its associated PERK/ATF4/CHOP signaling branch were suggested to serve a predominant role in cisplatin-induced ototoxicity. Notably, CHOP could inhibit Bcl-2 expression, activate the mitochondrial apoptotic pathway and further induce cisplatin-mediated cochlear cell apoptosis. Abbreviations: UPR: unfolded protein response; ATF: activating transcription factor; IRE1a: inositol-requiring enzyme 1a; PERK: pancreatic endoplasmic reticulum kinase; CHOP: C/EBP homologous protein.
aminoglycosides are considered to be the most important ototoxic drugs (He et al. 2017, Li et al. 2018a, 2018b. However, to date, there is still no effective treatment strategy for cisplatin-induced hearing loss based on known mechanisms underlying cisplatin-induced ototoxicity. A previous study from our laboratory demonstrated that the UPR was involved in cisplatin-induced cochlear cell apoptosis in vivo (Zong et al. 2017). However, how UPR mediates cisplatininduced ototoxicity requires further investigation. The current study aimed to explore the sequential changes in the key UPR signaling branch and its potential pro-apoptotic role in cisplatin-induced ototoxicity. Notably, the ATF4/CHOP signaling branch was activated in cisplatin-treated OC-1 cells in a time-dependent manner. The results indicated that the UPR may serve an important role in affecting cell fate under cisplatin treatment. Furthermore, CHOP was demonstrated to act as a key molecule in cisplatin-induced hair cell apoptosis. Cisplatin-induced ototoxicity results from injury of cochlear cells from the region of the organ of Corti, stria vascularis and spiral ganglia (Gentilin et al. 2019, Liu et al. 2019a. At the cellular level, the present study investigated the cytotoxic effect of cisplatin at different concentrations and time points. The results showed that the apoptotic rate of OC-1 cells was markedly increased within 48 h of cisplatin treatment. This finding was consistent with that reported in our previous study, which demonstrated that the number of TUNEL-positive cells and the expression of UPR markers GRP78, Caspase-12, and CHOP were increased in the cochlea of cisplatin-treated rats (Zong et al. 2017). Furthermore, the mitochondrial apoptotic pathway was activated within the consecutive changes in cell apoptosis.
ER stress, which caused by the toxic aggregation of unfolded or misfolded proteins in the ER, is considered the main reason for activating the UPR. The destruction of disulfide bonds by cisplatin prevents the protein from being folded into the correct spatial structure in the endoplasmic reticulum, resulting in the accumulation of unfolded or misfolded proteins in the ER (Sancho-Martinez et al. 2012). Therefore, the present study hypothesized that the UPR signaling branches could exert vital effects on cisplatin-induced ototoxicity. When ER stress occurs, activated PERK non-selectively inhibits the synthesis of total cellular proteins and selectively promotes ATF4 expression via phosphorylating eukaryotic translation initiation factor 2a. When the extent of the injury exceeds the compensatory capacity of the UPR, ATF4 induces apoptosis via upregulating CHOP (Nakamura et al. 2006, Gardner et al. 2013, Urra et al. 2013. Herein, the expression levels of ATF4 and CHOP were gradually increased in a time-dependent manner following treatment with cisplatin for 12 h. The above expression profile was consistent with that observed for the expression levels of apoptosisrelated proteins. The aforementioned findings indicated that the PERK/ATF4/CHOP signaling branch could play an essential role in cisplatin-induced apoptosis in OC-1 cells and cochlear explants. This hypothesis was further verified via knockdown experiments, which demonstrated that cell transfection with si-CHOP alleviated cisplatin-induced apoptosis in OC-1 cells. In the three signaling branches of the UPR, the signaling transduction downstream of IRE1a is relatively complicated. It has been reported that activated IRE1a can promote the expression of several pro-inflammatory proteins, including TXNIP (Oakes and Papa 2015). Herein, the mRNA expression of TXNIP was gradually upregulated with prolonged cisplatin treatment, while the protein level of TXNIP showed no significant change until 48 h after cisplatin treatment, suggesting that the UPR-related inflammatory signal may be involved in cisplatin-induced ototoxicity, but not a key factor contributing to hair cell apoptosis. In addition, oxidative stress is considered as a significant mechanism underlying cisplatininduced cytotoxicity (Sancho-Martinez et al. 2012). As a prooxidant molecule, it is possible that the UPR-mediated upregulation of TXNIP may promote oxidative stress in cisplatin-treated hair cells (Alhawiti et al. 2017). In addition, XBP1 mRNA is also a substrate of activated IRE1a, which cleaves 26 introns of XBP1 mRNA to produce XBP1s mRNA (Walter et al. 2015), the translation product which can promote the recovery of cellular proteostasis (Nakamura et al. 2006, Sado et al. 2009, Kanekura et al. 2015, Hetz and Saxena 2017. Herein, the mRNA expression levels of XBP1s were elevated at 36-48 h following cisplatin treatment. These results indicated that XBP1s could be produced quite late or at low levels in terms of the complex and integrated UPR signaling dynamics and therefore, could not confer cytoprotective effects. Additionally, activated IRE1a may also serve as an activation platform for ASK1 to induce JNK-dependent apoptosis (Oakes and Papa 2015). In the present study, the level of p-ASK1/ASK1 showed no significant change until 48 h after cisplatin treatment. Therefore, it was hypothesized that ASK1/ JNK signaling may not exert an essential role in cisplatininduced apoptosis.
ATF6 is another UPR sensor located in the ER. Under ER stress, ATF6 is activated in the Golgi complex and transformed into a stable protein with a molecular weight of 50 kDa, which is then accumulated in the nucleus (Haze et al. 1999). Subsequently, ATF6 induces the expression of XBP1s and other UPR-related target genes to promote the recovery of ER homeostasis (Hetz 2012, Guo et al. 2014. However, in the present study ATF6 did not to respond to cisplatin. Emerging evidence has suggested that each of the UPR sensors has a unique set of pro-apoptotic outputs contributing to cell death in the setting of irremediable ER stress (Shore et al. 2011, Oakes andPapa 2015). Herein, PERK/ATF4/ CHOP was identified as the essential branch involved in cisplatin-induced UPR activation of hair cells. As a crucial transcriptional factor connecting the terminal UPR with mitochondrial apoptosis, CHOP attenuates the expression of the anti-apoptotic protein Bcl-2, thus transmitting pro-death signals to the mitochondria. In the present study, CHOP knockdown reduced the number of TUNEL-positive cells and inhibited the mitochondrial apoptotic pathway.
The UPR acts through an extremely complex signaling transduction system with dynamic characteristics that cause different biological effects in response to various factors, including the duration of injury. In addition, the results of this study cannot rule out the possibility that the mitochondrial damage in OC-1 cells under cisplatin treatment is partly caused by factors such as oxidative stress. On the other hand, there may be extensive crosstalk between the UPR and other cellular pathophysiological processes such as oxidative stress (Xiong et al. 2021). Nevertheless, as the present study elucidated, the importance of PERK/ATF4/CHOP signaling branch as the molecular targets for therapy development in cisplatin-induced hearing loss should be put in perspective.
In conclusion, the present study determined the sequential changes and the underlying regulatory mechanism of the UPR signaling branches in cisplatin-induced ototoxicity. The results suggested that the PERK/ATF4/CHOP signaling branch may serve an important role in cisplatin-induced cell apoptosis. Overall, the current study provided novel insights into the role of the UPR in cisplatin-induced ototoxicity and identified CHOP as a promising intervention target for treating cisplatin-induced ototoxicity.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
This study was supported by grants from the National Natural Science Foundation of China (No. 81771002).