Capsaicin Protects Against Cisplatin Ototoxicity by Changing the STAT3/STAT1 Ratio and Activating Cannabinoid (CB2) Receptors in the Cochlea

Capsaicin, the spicy component of hot chili peppers activates the TRPV1 pain receptors, and causes rapid desensitization. Capsaicin also ameliorates cisplatin-induced nephrotoxicity. Cisplatin, a commonly used anti-neoplastic agent for solid tumors causes significant hearing loss, nephrotoxicity and peripheral neuropathy. Upregulation of cochlear TRPV1 expression is related to cisplatin-mediated ototoxicity. Here we report that direct TRPV1 activation by localized trans-tympanic (TT) or oral administration of capsaicin (TRPV1 agonist) prevents cisplatin ototoxicity by sustained increased activation of pro-survival transcription factor signal transducer and activator of transcription (STAT3) in the Wistar rat. Cisplatin treatment produced prolonged activation of pro-apoptotic Ser727 p-STAT1 and suppressed Tyr705-p-STAT3 for up to 72 h in the rat cochlea. Our data indicate that capsaicin causes a transient STAT1 activation via TRPV1 activation, responsible for the previously reported temporary threshold shift. Additionally, we found that capsaicin increased cannabinoid receptor (CB2) in the cochlea, which leads to pro-survival Tyr705-p-STAT3 activation. This tilts the delicate balance of p-STAT3/p-STAT1 towards survival. Furthermore, capsaicin mediated protection is lost when CB2 antagonist AM630 is administered prior to capsaicin treatment. In conclusion, capsaicin otoprotection appears to be mediated by activation of CB2 receptors in the cochlea which are coupled to both STAT1 and STAT3 activation.


Results
Capsaicin protects against cisplatin ototoxicity. We first assessed ABRs in naïve adult male Wistar rats prior to treatment with either trans-tympanic (TT) vehicle or capsaicin (0.1 µM in 50 µl). Twenty-four hours later, we then infused cisplatin (11 mg/kg) intraperitoneally (i.p) and determined post-treatment ABRs 72 h later to assess hearing loss. Trans-tympanic administration of vehicle (sterile PBS in a volume of 50 µl) produced negligible changes in ABR threshold compared to naïve controls. ABR threshold shifts 72 h following cisplatin administration showed significant elevation. TT-Capsaicin (50 µl of a 0.1 µM solution) pretreatment 24 h prior to cisplatin significantly reduced ABR threshold shifts produced by cisplatin at all three frequencies tested (8,16 and 32 kHz) (Fig. 1A). TT-Capsaicin administered alone did not significantly alter ABR thresholds, compared to trans-tympanic vehicle-treated rats after 72 h.
Capsaicin is a nutraceutical consumed orally as a pungent spice in several Asian cultures. Therefore, we decided to test the efficacy of oral capsaicin to prevent cisplatin ototoxicity. Oral capsaicin was effective in ameliorating cisplatin-induced hearing loss when we administered it 24 h prior to cisplatin. We observed dose dependent reduction in cisplatin-induced ABR threshold shifts. Capsaicin doses were 5, 10 or 20 mg/kg. We found significant protection in rats administered 10 or 20 mg/kg (Fig. 1A). However, the lowest dose of capsaicin tested (5 mg/kg) did not protect against cisplatin-induced ABR threshold shifts. Capsaicin administered by either trans-tympanic injection or by oral gavage alone did not significantly alter ABR threshold, when assessed 72 h later.
Scanning electron microscopic (SEM) images of the basal turn of the cochlea obtained from the rats treated above showed significant morphological changes in the stereociliary bundles on the outer hair cells (OHCs), indicating damage to the OHCs in the cisplatin group (Fig. 1B). In these studies, cisplatin produced a ten-fold greater loss or damage of outer hair cells compared to vehicle controls. Rats pretreated with trans-tympanic capsaicin had no significant loss of outer hair cells compared to vehicle controls. Oral capsaicin also significantly reduced stereociliary damage induced by cisplatin. Graphical representation of the data is shown in (Fig. 1C). We stained whole mount images of cochleae from similar groups of rats with a polyclonal antibody against myosin VIIa, a hair cell-specific marker. These images supported the SEM findings, wherein cisplatin-induced hair cell loss was significantly decreased by capsaicin pre-treatment (Fig. 1D, red arrows indicative of missing OHCs). Quantitative analyses of hair cell loss indicates that cisplatin-induced outer hair cell death was 43 ± 4%; however, rats pretreated with oral capsaicin prior to cisplatin demonstrated no significant outer hair cell loss compared to control. Oral capsaicin alone produced minimal hair cell damage (Fig. 1E). Capsaicin appears to alleviate cisplatin-induced hearing loss and OHC damage whether administered locally or orally.

Capsaicin reduces cochlear oxidative stress and inflammatory genes induced by cisplatin.
Cisplatin increased the expression of several cochlear oxidative stress and inflammatory genes. These genes regulate different cellular pathways which culminate in the death of OHCs and other cells of the cochlea, leading to hearing loss 18,[22][23][24] . We examined whether capsaicin could regulate genes linked to cochlear inflammation, such as TRPV1, NOX3 and iNOS. Real time quantitative PCR studies of the cochleae 72 h following cisplatin administration indicate 2-3 fold increases in TRPV1, NOX3 and iNOS expression. Rats pretreated with TT-capsaicin 24 h prior to cisplatin administration demonstrated significant reduction in the expression of these genes to control levels. Capsaicin administration alone did not significantly change the expression of these genes ( Fig. 2A). These data suggest an anti-inflammatory role of capsaicin in the cochlea when combined with cisplatin. We further examined gene expression in mid-modiolar sections of the rat cochlea by immunolabelling for TRPV1, NOX3 and iNOS proteins. Cisplatin increased TRPV1, NOX3 and iNOS immunoreactivity in the organ of Corti (OC), spiral ligament (SL), spiral ganglions (SG) and marginal cells of stria vascularis (SVA) compared to control. Capsaicin pre-treatment prevented this increase (Fig. 2B,C). Fluorescent intensity analyses performed using image J software revealed that cisplatin significantly increased TRPV1, NOX3 and iNOS labelling in the OHCs, SG and SVA compared to control. Capsaicin pre-treatment significantly decreased the expression of these proteins induced by cisplatin, while capsaicin treatment alone did not produce significant change compared to control (Table 1). These data provide additional support for an anti-inflammatory role of capsaicin against cisplatin in the cochlea.
Capsaicin treatment changes p-STAT3/p-STAT1 dynamics during cisplatin insult. Our previous study demonstrated that capsaicin stimulated the Ser 727 phosphorylation of STAT1 which is linked to transient inflammation in the cochlea and temporary hearing loss 18 . We propose that recovery from the transient hearing loss involves resolution of the initial inflammatory response. Capsaicin administration is linked to activation and/ or inhibition of STAT3 phosphorylation in cancer cells [8][9][10] . However, the significance of these findings to cochlear cells under physiological conditions is unclear. We have recently shown that EGCG (epigallocatechin gallate) protects against cisplatin-induced apoptotic cell death by stabilizing the STAT3/STAT1 ratio 25 . In the present study, we monitored the capsaicin-regulated STAT1 and STAT3 activation. The latter promotes anti-inflammatory responses and cell survival 24,25 . We show that capsaicin (2.5 µM) activates STAT1 Ser 727 phosphorylation in UB/OC-1 cells in a time-dependent fashion, which peaked at 45 min and returned to baseline by 120 min (Supplementary Fig. 1A). Similarly, capsaicin induced Tyr 705 p-STAT3 phosphorylation was time-dependent, which peaked at 45 min after drug administration and returned to baseline by 120 min (Fig. 3A). This was associated with increased activation of p-JAK2, the immediate upstream regulator of STAT3 phosphorylation at 45 min (Supplementary Fig. 2A). Capsaicin treatment showed significant increase in p-STAT3/p-STAT1 ratio over time with a peak value (~130%) reached in 45 min, with recovery to baseline by 60 min (Fig. 3C), thus tilting the balance of the cell fate towards pro-survival. The ratio of pSTAT3/pSTAT1 was relatively unchanged beyond that time period.
Cells treated with cisplatin (2.5 µM) showed a persistent increase in Ser 727 p-STAT1, which was still elevated at 120 min ( Supplementary Fig. 1B). On the other hand, we observed that cisplatin produced a sustained inhibition of Tyr 707 p-STAT3 at 120 minutes (Fig. 3B). The ratio of p-STAT3/p-STAT1 following cisplatin treatment is shown in Fig. 3D. This indicates a significant and persistent reduction in this ratio from 30 min to 120 min. Disruption of p-STAT3/STAT1 balance by inhibiting or knockdown of STAT3 sensitizes UB/OC-1 cells to cisplatin 25 . We recorded ABR threshold shifts in naïve Wistar rats, pre-treated with either trans-tympanic capsaicin (0.1 µM, 50 µl) or TT-PBS (50 µl), oral PBS or oral capsaicin (20 mg/kg), 24 h prior to cisplatin (11 mg/kg, i.p). We measured post treatment ABR thresholds at 72 h after cisplatin and they showed significant increase with cisplatin alone, which capsaicin attenuated at all frequencies tested. Black arrows indicate significant decrease in threshold shifts when compared to cisplatin. (*Indicates significant difference from control; **indicates significant difference from cisplatin treatment, p < 0.05, n = 9). (B) SEM studies performed on rat cochleae show disruption of the stereociliary bundles indicating damage to the OHCs by cisplatin which was significantly ameliorated by pre-treatment with capsaicin. (C) Quantitative analysis of outer hair cell damage in the basal turn in the SEM images. (*Indicates significant difference from control; **indicates significant difference from cisplatin treatment, p < 0.05, n = 4). (D) Cochlear whole mount preparations from animals stained with myosin VIIA and imaged by confocal microscopy indicate that cisplatin caused hair cell loss in the basal turn (indicated by red arrows), while capsaicin pre-treatment was protective. Scale bar is 25 µm. (E) Quantitative analysis of presence or absence of outer hair cells from the basal turn of the whole mount. (*Indicates significant difference from control; **indicates significant difference from cisplatin treatment, p < 0.05, n = 4).

Figure 2.
Capsaicin inhibits cisplatin-induced stress and inflammation in the cochlea. We pretreated male Wistar rats with trans-tympanic capsaicin (0.1 µM, 50 µl) 24 h prior to cisplatin (11 mg/kg). We collected -the cochleae 72 h post cisplatin, and saved these specimens either in RNA later or fixed with freshly prepared 4% paraformaldehyde, decalcified and processed them for mid-modiolar sections. (A) Quantitative q-PCR analyses from total cochlear RNA indicate that cisplatin treatment increased the relative mRNA expression of TRPV1, NOX3 and iNOS, that was inhibited by trans-tympanic pre-treatment with capsaicin. (B,C) We used midmodiolar sections for immunofluorescent staining of TRPV1 (B), NOX3 and iNOS (C). Cisplatin increased TRPV1, NOX3 and iNOS staining in the OC, SL, SV and SG, which with trans-tympanic capsaicin pretreatment blocked. Images shown are 20X and the insets were imaged at 100x by confocal microscopy. Scale bar is 100 µm. Data presented in (A) represent the mean ± S.E.M. of cochleae from four animals from each group. Asterisk (*) indicates statistically significant difference from vehicle, while (**) indicates significant difference from the cisplatin group (p < 0.05).
www.nature.com/scientificreports www.nature.com/scientificreports/ Immunohistochemistry studies using mid-modiolar cochlear sections from rats treated with cisplatin for 72 h showed up-regulation of Ser 727 p-STAT1 in the basal turn. Similar mid-modiolar sections from oral capsaicin-treated rats did not show up-regulation in Ser 727 p-STAT1 (Fig. 3G), indicating inherent differences between these two drugs. Interestingly, oral capsaicin increased Tyr 705 p-STAT3 phosphorylation as shown in mid-modiolar cochlear sections at 72 h post drug administration, while cisplatin suppressed STAT3 phosphorylation. Mean intensity analyses of the images with image J software indicate that the p-STAT3/p-STAT1 ratio is decreased significantly by cisplatin treatment in the OHC and SG, while SVA showed decreased ratio compared to oral PBS control sections. Oral capsaicin pretreatment significantly increased this ratio compared to cisplatin treatment, while oral capsaicin alone increased the p-STAT3/p-STAT1 ratio in all the three regions of the organ of Corti significantly, compared to control as well as cisplatin (Fig. 3H). The intensity ratios of p-STAT3/p-STAT1 in the different regions of organ of Corti are presented in Table 2. These findings support our hypotheses that capsaicin protects from cisplatin-induced hearing loss not only by inhibiting the pro-inflammatory pathway but also, by stabilizing the p-STAT3/p-STAT1 ratio and tilting the delicate balance towards cell survival. We next monitored two genes, Bax and Bcl-2, whose expressions are regulated by STAT1 26,27 and STAT3 28,29 , respectively. In UB/OC1 cells, Bax/Bcl-2 ratios were 1 ± 0.2 for control, 2.9 ± 0.1 for cisplatin, 1.31 ± 0.1 capsaicin + cisplatin, and 1.1 ± 0.2 for capsaicin-treated cells, respectively (graph not shown). Rat cochlear gene expression studies showed similar expression patterns of Bax/Bcl-2 ratio. In rats treated with oral PBS we found the ratio was 1 ± 0.2. Rats treated with cisplatin demonstrated a significant increase in the Bax/Bcl-2 ratio. Rats pre-treated with oral capsaicin prior to cisplatin demonstrated a significant decrease of this ratio. Rats treated with oral capsaicin alone showed a ratio less than half of that observed following cisplatin alone (Fig. 3I).
Capsaicin mediated protection is dependent on STAT3 activation. We next examined the role of the STAT3 pathway in UB/OC-1 cell growth and survival. Capsaicin, added prior to cisplatin, blocked cisplatin-induced cell killing of UB/OC-1 cells. Pre-treatment of UB/OC1 cells with STATTIC (a small molecule inhibitor of STAT3) abrogated the protective action of capsaicin against cisplatin-induced cell killing. Surprisingly, STATTIC also unmasked a toxic action of capsaicin (Fig. 4A). STATTIC (100 nM) significantly reduced the Tyr 705 p-STAT3 levels (Fig. 4B). This suggests an active role of STAT3 in cytoprotection when cells are challenged with a cytotoxic agent like cisplatin. The cytotoxic effect of capsaicin produced by STAT3 inhibition (see Fig. 4A) implicates this transcription factor in the cell survival/proliferative actions of capsaicin. The lack of effect of STATTIC when added alone would suggest that the STAT3 pathway is not tonically active under normal resting conditions (Fig. 4B).
www.nature.com/scientificreports www.nature.com/scientificreports/ capsaicin induced p-STAT3 Tyr 705 phosphorylation (Fig. 5B). However, inhibition of CB2 receptors with AM630 significantly decreased capsaicin-induced Tyr 705 p-STAT3 and Ser 727 p-STAT1. Western blots showed that AM630 pretreatment decreased capsaicin-induced Ser 727 p-STAT1 phosphorylation (Fig. 5C) Table 2. Capsaicin pretreatment induces pro-survival p-STAT3/p-STAT1 signal. Fluorescent intensity analyses of the pro-survival signal in rat cochlea: Mid-modiolar sections of rat cochlea were probed with p-STAT3 and p-STAT1 antibodies and imaged using confocal microscopy. Fluorescent intensity analyses of the images were performed using Image J software. Mean gray values of p-STAT3 and p-STAT1 were calculated and the ratio for each sample in each group were then calculated as %. (p < 0.05, *significant difference compared to oral PBS group, **significant difference compared to cisplatin treatment group). www.nature.com/scientificreports www.nature.com/scientificreports/ p-STAT3 phosphorylation (Fig. 5D). Moreover, the CB2 receptor agonist JWH-015 increased p-STAT3 and this effect was reversed by AM630 (Fig. 5E). Pre-treatment of UB/OC-1 cells with AM630 30 minutes prior to capsaicin for 24 h significantly decreased cell viability compared to capsaicin (Fig. 5F). Thus, capsaicin likely activated STAT3 via CB2 receptors, which may be essential for cell survival.
Cannabinoid receptors are expressed in the rat cochlea, and CB2 receptor is essential for capsaicin-induced protection of cisplatin-mediated hearing loss. Previous studies have reported that CB2 receptors are expressed in HEI/OC1 cells 29 and in the albino rat cochlea 32 . Another study found CB1 expression in SAMP8 mice cochleae 33 . We sought to determine whether cannabinoid receptors (CB1 and CB2) are expressed in UB/OC1 cells and in the rat cochlea. We investigated whether capsaicin treatment could increase the expression of CB receptors. We performed western blotting on UB/OC1 cells treated with capsaicin for 24 h. Capsaicin increased CB2 expression in UB/OC-1 cells (Supplementary Fig. 3A). We determined rat cochlear gene expression levels 72 h after oral capsaicin treatment. Capsaicin increased the expression of CB2 receptor in the rat cochlea ~2 fold relative to basal expression ( Supplementary Fig. 3B). We observed CB1 and CB2 receptor expression by immunofluorescent labeling in the organ of Corti in the OHCs, IHC's, the supporting cells (SC's), the spiral ganglion and the stria vascularis in mid-modiolar sections of untreated rats. (Supplementary Fig. 3C,D).
We investigated capsaicin induced CB2 receptor activation. We pretreated rats with either vehicle or AM630 by the trans-tympanic route, followed by oral capsaicin (20 mg/kg) or vehicle 4 h later. We administered cisplatin (11 mg/kg, i.p.) 24 h after oral capsaicin and recorded ABR thresholds 72 h post cisplatin. Relative fluorescent intensity (rfi) analyses of mid-modiolar section using confocal microscopy by ImageJ software indicated www.nature.com/scientificreports www.nature.com/scientificreports/ that cisplatin treatment did not significantly change the level of CB2 labelling in the OHCs, SG's or in the SVA. However, capsaicin pretreatment significantly increased the expression of CB2 in all the three regions of the cochleae, while capsaicin treatment alone increased CB2 labelling in all the three regions to an even greater extent compared to control (Fig. 6A). Cisplatin produced a 20-40 dB ABR threshold shift with increasing frequencies ranging, from 8-32 kHz, which was abrogated by oral capsaicin. Trans-tympanic administration of CB2 receptor antagonist AM630 reversed the protective effect of oral capsaicin against cisplatin-induced hearing loss. Interestingly, AM630 added either alone or in combination with capsaicin resulted in significant increases in ABR thresholds (Fig. 6B). These latter data suggest that CB2 exerts a tonic protective function in the cochlea and could represent a novel target for otoprotection.
Capsaicin does not interfere with the antitumor efficacy of cisplatin. We employed a SCID mouse xenograft model to determine the effect of oral capsaicin treatment chemotherapeutic efficacy of cisplatin. We injected mice with 1 × 10 6 head and neck squamous cell carcinoma (UMSCC-10b) cells subcutaneously in the flank region. Tumor growth plotted over a four week period showed a substantial increase in tumor volumes over time which was significantly reduced by cisplatin. Administration of oral capsaicin shifted the cisplatin tumor www.nature.com/scientificreports www.nature.com/scientificreports/ growth curves to the right, suggesting improved efficacy of this drug combination over cisplatin alone. Tumor weights obtained at the end of the treatment period support these findings (Fig. 7).

Discussion
The goal of this study was to test the hypothesis that capsaicin could protect against cisplatin-induced ototoxicity and to delineate the mechanisms involved. We developed this hypothesis based on our previous finding that capsaicin produced transient hearing loss in the rat which recovered over a three day period, unlike cisplatin which produced persistent hearing loss 18 . We propose that the transient nature of capsaicin's action could precondition the cochlea to a subsequent cochlear trauma (i.e. to an ototoxic drug) (Fig. 8). This study confirms the hypothesis that capsaicin could precondition the cochlea and reduce cisplatin ototoxicity. We show that capsaicin, unlike cisplatin, produces transient activation of STAT1 in UB/OC-1 cells and in the cochlea. These findings contrast with the more persistent activation of STAT1 by cisplatin in both UB/OC-1 cells and cochlea and likely result from the more persistent activation of TRPV1 by cisplatin.
Preconditioning utilizes the innate ability of an organ to mount a defense against minor trauma and using that to combat severe trauma. Investigators have utilized this strategy to mitigate pain and nephrotoxicity 7,11,12 . Several groups have studied preconditioning with different agents to prevent noise, age-related or cisplatin-induced hearing loss. Investigators using preconditioning sound in rat and mouse models have demonstrated protection from noise injury as well as age-induced hearing loss [34][35][36] . Sound preconditioning appears to protect against cisplatin-induced hearing loss 37 . A more recent study suggests that sound preconditioning may work via the induction of heat shock proteins (HSP's) 38 . A study in zebrafish also showed that dexmedetomidine (DEX) (an alpha-2 adrenoreceptor selective agonist) is effective in preconditioning against cisplatin ototoxicity 39 .
The differences in the ABR threshold shifts in the capsaicin and cisplatin-treated rats were unexpected, given that both capsaicin and cisplatin target TRPV1 13 . Data from our laboratory indicate the involvement of TRPV1 in both capsaicin and cisplatin mediated inflammatory response leading to either a temporary or a permanent ABR threshold shift respectively. Capsaicin mediated TRPV1 activation is transient and resolved within 72 h. Our data further indicate that capsaicin causes an induction of pro-apoptotic STAT1 transcription factor as well as the pro-survival STAT3 transcription factor. This induction of p-STAT3 in the cochlea persists even at 72 h after capsaicin. However, capsaicin transiently induced p-STAT1. This tilts the balance of the p-STAT3/p-STAT1 ratio towards cell survival. Activation of TRPV1 by capsaicin promotes STAT1 activation and the transient induction of inflammatory genes in the cochlea, leading to temporary hearing loss for 24 hr 18 . Similarly, activation of TRPV1 in small diameter C-fiber neurons promotes neurogenic inflammation 4,40 . However, following cisplatin treatment, STAT1 activation persists for 72 h while STAT3 activation is decreased. This alters the balance of p-STAT3/p-STAT1 towards cell death. Cisplatin treatment results in sustained increase in cochlear ROS generation. This leads to TRPV1 activation and persistent inflammation 13 . Thus, TRPV1 appears to be a target for the induction of hearing loss by cisplatin. Knockdown of this protein provided significant protection against Figure 7. Oral capsaicin pre-treatment does not affect cisplatin's chemotherapeutic ability in the SCID mouse xenograft model. We injected twenty seven SCID mice with UMSCC-10b cells (1 × 10 6 cells) subcutaneously and divided them into four treatment groups: Oral PBS + PBS (i.p, 1 ml) (n = 8), cisplatin (2 mg/kg, i.p) (n = 7), oral capsaicin (0.5 mg/kg) + cisplatin (2 mg/kg, i.p) (n = 7) and oral capsaicin alone (n = 5). We administered drugs by oral gavage and/or i.p injections every other day (three times a week, once the tumors were palpable, until the end of the study. We monitored tumor growth closely. Oral capsaicin + Cisplatin and the cisplatin treated mice showed suppressed tumor size (C-F and I) and weight (J). Capsaicin treatment alone decreased tumor size and weight, but not significantly. We show representative images of the tumor bearing mice and excised tumors at the end of the study from each group in (A-H). Line graphs and histograms in (I,J) represent mean ± SEM from each treatment group. Asterisk (*) indicates statistically significant difference (p < 0.05) from vehicle-treated mice.
www.nature.com/scientificreports www.nature.com/scientificreports/ cisplatin-induced hearing loss, and apoptosis of outer hair cells 13 . However, TRPV1 agonists appear to be effective in treating bladder pain and over reactive bladder, presumably by inducing desensitization of this receptor 41 . These data suggest a bimodal function of TRPV1 activation by capsaicin but not cisplatin in the cochlea.
The mechanism underlying the transient activation of STAT1 by capsaicin is unclear but may involve rapid de-phosphorylation of this protein by phospho-serine phosphatases, such as protein phosphatase 2A and calcineurin. In addition, it is possible that the increased STAT3 activation produced by capsaicin negatively influences STAT1 activation 42,43 . Another difference between capsaicin and cisplatin is their differing ability to regulate JAK2 and STAT3 phosphorylation. In our cell culture model, we showed rapid phosphorylation of both JAK2 and STAT3 by capsaicin, which persisted for at least 120 min. In contrast, cisplatin produced rapid inhibition of JAK2 and STAT3 phosphorylation. Therefore, the cellular STAT3/STAT1 ratio and cell viability is greater in the capsaicin-treated cells as compared to the cisplatin-treated cells.
We found similar results in vivo. Capsaicin increased STAT3 activation without activating STAT1 at 3 days following exposure to cisplatin. In contrast, rats administered cisplatin demonstrated reduced p-STAT3 and greater p-STAT1 in the cochlea than vehicle-treated rats. In nerve injury, STAT3 appears to plav a pro-survival role. This transcription factor appears to be crucial for Schwann cell survival and phenotype conversion for regeneration in mice 44 . STAT3 also inhibits interferon-mediated STAT1 activity and its ability to form homodimers in myeloid cells 45 . Therefore, an increase in STAT3 by capsaicin could serve to diminish the activity of STAT1 and prevent inflammation produced by cisplatin. In contrast, inhibition of STAT3 by STATTIC enhanced killing of UB/OC-1 cells by cisplatin in vitro, without affecting survival under basal conditions. Furthermore, STATTIC unmasked a cytotoxic property of capsaicin, likely resulting from unopposed activation of STAT1. STAT3 has important cellular functions linked to cell survival 46 and the relative proportions of STAT1/STAT3 can regulate the transcriptional activity of anti-apoptotic genes such as Bcl-2 and Bcl-x 27 . Thus, inhibition of STAT3 by cisplatin could play a major role in its initiation of cell death in the cochlea.
Another important finding of this study is that capsaicin-dependent activation of STAT3 is mediated, at least in part, through CB2 receptors. Immunohistochemical studies show expression of CB2 in the cochlea, localized primarily in the OHCs, IHCs, stria vascularis and spiral ganglion cells. Activation of these receptors appears to protect against cisplatin ototoxicity. Inhibition of these receptors by a CB2-selective antagonist reversed this protective action of capsaicin and produced significant hearing loss when added alone. Thus, CB2 receptor activation by endogenous cannabinoids could have a tonic influence on hearing. A recent report 32 suggests that cisplatin causes a spontaneous increase in CB2 expression in the cochlea. These authors proposed a protective role for CB2 in the cochlea. Cannabinoids appear to exert anti-inflammatory activity in various in vivo and in vitro models. These agents ameliorate inflammatory conditions 47,48 , presumably via the JAK2/STAT3 pathway. The CB2 agonist HU-308 attenuated cisplatin-induced nephrotoxicity by limiting inflammation, oxidative and nitrosative stress in the kidney 49 . www.nature.com/scientificreports www.nature.com/scientificreports/ The otoprotective function of capsaicin did not appear to compromise cisplatin-induced reduction of tumor growth in a SCID mouse xenograft model. In fact, there appeared to be a trend towards increased chemotherapeutic efficacy of combined capsaicin and cisplatin drug compared to cisplatin alone. Therefore, combining oral capsaicin with cisplatin-based chemotherapy could protect against hearing loss without compromising the chemotherapeutic efficacy of cisplatin. The ease of administering capsaicin orally could enhance the feasibility of this approach. Capsaicin is available commercially as a flavored capsule that is easily tolerated. In addition, capsaicin is already in use clinically as a topical ointment for pain and skin conditions, and cannabis derived medicines are also in clinical use for the treatment of various inflammatory diseases. We propose further investigations of capsaicin protection against hearing loss resulting from cisplatin and other inflammatory conditions. Proposed mechanism for the protective effect of capsaicin against cisplatin ototoxicity.
Capsaicin mediates a transient early inflammatory response in the cochlea. It also activates a delayed anti-inflammatory, pro-survival cascade that protects against cisplatin-induced hearing loss. Capsaicin appears to be otoprotective by desensitization of TRPV1 and also by activation of CB2 receptors. CB2 receptor activation leads to JAK2/STAT3 pro-survival signaling that negates the cisplatin-induced inflammatory/apoptotic (STAT1/ Bax:Bcl2) pathways. Capsaicin also activates the TRPV1 pathway transiently which causes an increase in Ca 2+ release, an increase in ROS generation and increased p-STAT1Ser 727 phosphorylation. However, apoptotic pathways are not recruited since p53 is not activated 50 . Cisplatin activates TRPV1 receptors chronically. This leads to increases in: Ca 2+ , ROS and p-STAT1Ser 727 phosphorylation and inflammation. This is followed by recruitment of p53, BAX and caspase cascade 19 . Thus capsaicin shifts the delicate balance of STAT3/STAT1 42 towards pro-survival signaling, enabling a CB2R activation leading to cell survival. This inhibits damage and apoptosis of outer hair cells and prevents hearing loss in cisplatin treated animals (Fig. 8). We cultured Cells in RPMI 1640 media supplemented with 10% Fetalclone II serum, and penicillin-streptomycin (1 ml/100 ml) and normocin (2 µl/ml). We grew cultures at 33 °C in an incubator with 10% CO 2 . We passaged cells twice a week (1:3-4 media ratio). Dr. Krishna Rao provided to us a head and neck cancer cell line, UMSCC 10B 51 , by Southern Illinois University (SIU) School of Medicine (Springfield, IL, USA). We cultured cancer cells in DMEM (HyClone) supplemented with 10% fetal bovine serum (Atlanta Biologicals Inc., Flowery Branch, GA, USA) and penicillin-streptomycin (Invitrogen).

Animals.
We utilized-adult male Wistar rats (200-350 g) (from Envigo) for this study. We housed the rats in a temperature controlled room with a twelve hour light/dark cycle. They had free access to commercial food and water provided by the SIU Laboratory Animal Care facility. Cisplatin is administered intraperitoneally at 11 mg/kg dose in the rat.
We injected SCID mice (from Envigo, 5-6 weeks of age) subcutaneously with 1.5 × 10 6 mycoplasma-free UMSCC 10B cells in one flank. When tumors reached palpable size (~100 mm3), attained in 10-15 days after injections, pre-treated mice with oral PBS (vehicle) or oral capsaicin (0.5 mg/kg), and we then treated each group with intraperitoneal PBS or intraperitoneal cisplatin (2 mg/kg). We subsequently treated animals were with oral capsaicin, followed by intraperitoneal cisplatin, on alternate days three times per week for a total of 11 treatments. At the time of each treatment, we calculated tumor volumes based on the formula: volume = width 2 × (length/2) 52 . We killed mice 24 h after the eleventh treatment. The SIU, Laboratory Animal Care and Use Committee approved and monitored all animal procedures. We confirm that we performed all experiments in accordance with relevant guidelines and regulations of SIU School of Medicine.

Auditory brainstem evoked responses (ABRs).
Prior to experiments, we anesthetized rats with ketamine HCl /xylazine mixture (24.6/3 mg/kg) and we placed them in a sound-attenuation chamber (Industrial Acoustic Company, Inc.). We maintained body temperature of the rats was at 37 °C with an animal blanket system. We performed pretreatment auditory brainstem responses (ABRs) using an IHS-High-frequency System (Intelligent Hearing Systems, Miami, FL). We exposed the rats to acoustic stimuli via HIS transducers that were directly placed at the entrance of the ear canal. ABR thresholds were obtained for 5 ms duration tone bursts at 8, 16 and 32 kHz at a rate of 50/s. We determined thresholds by visually produced lowest intensity evoked potentials (EP) that progressed in 10 dB steps. We amplified evoked potentials (Eps) to 200,000X,we band pass filtered them (100-3000 Hz) and averaged them over 1024 sweeps 53  www.nature.com/scientificreports www.nature.com/scientificreports/ We performed post treatment ABR's 72 h following cisplatin administration. At the end of auditory testing, we decapitated the anesthetized animals and harvested the cochlea. The SIU Laboratory Animal Care and Use Committee approved all animal procedures.
Trans-tympanic injections. We anesthetized male Wistar rats using ketamine/xylazine cocktail. We punctured the tympanic membrane as described in Mukherjea et al. 53 .

Cell Viability Assay (MTS Assay).
We performed the in vitro UB/OC1 cell proliferation by using the CellTiter 96 ® AQueous One Solution Cell Proliferation Assay Kit (Promega, Madison, WI), according to the manufacturer's instructions. We seeded 3,500 cells per well into a 96-well plate. We pretreated cells were with varying concentrations of capsaicin for 45 mins and cisplatin (20 µM) for 48 h. After 48 h, we added 20 µl of CellTiter 96 AQueous One Solution reagent to each well in 100 µl of total volume of media. We incubated cells for 1 h, and recorded absorbance at 490 nm using an ELISA plate reader. The absorbance is directly proportional to the number of living cells and is expressed as a percent relative to vehicle-treated cells.
processing of cochlea for immunohistochemistry. We perfused cochleae with a 4% paraformaldehyde solution, we decalcified cochleae for 7-10 days in 120 mM EDTA, embedded them in paraffin and then sectioned the specimens 53 . We deparaffined slides and rehydrated and stained them with primary and secondary antibodies. We mounted glass coverslips on the specimens with ProLong Diamond mounting medium. We obtained images using a Leica confocal microscope (Buffalo Grove, IL).
Fluorescent intensity analyses of immunohistochemical images. Image J software was used to analyze the mean gray area for the various images. These values were then used to quantify the increase or decrease in % fluorescence intensity of the various regions of the samples compared to control.
Morphological studies by scanning electron microscopy (SEM) and Hair cell count. Immediately after completion of post-treatment ABRs, we euthanized rats, harvested their cochleae and processed the specimens as described previously by 54 . We viewed sputter-coated cochlea and photographed them with a Hitachi S-500 scanning electron microscope (Hitachi Ltd.). We examined the resulting scanning electron micrographs to determine qualitative morphological characteristics in various regions including the basal, middle, and apical turns of the rat cochlea.
Hair cell counts were performed as described previously 13 . Two representative areas of the basal turn, middle turn, and apex and hook portion were photographed. In each area, outer hair cells (OHCs) were counted in an area that was 10 pillar cell heads in length. The results are presented as the percentage hair cell damage per cochlear turn. At least three cochleae from different animals per treatment group were used.

Cochlear whole mount preparation and hair cell count.
To study the organization of the outer hair cells and inner hair cells, we used the cochlear whole mount technique. We fixed the cochleae in fresh 4% paraformaldehyde overnight, and then decalcified them in 0.1 mM EDTA for 2 weeks, (pH 7.3) while stirring at room temperature for 2 weeks. After decalcification, we microdissected the cochleae into basal, middle, and apical turns for whole-mount preparation.
For further processing, we washed segments with 1X PBS twice for 5 mins. For blocking, we incubated the segments in 10% blocking serum (horse serum) for 3 hrs at room temperature. We incubated the segments with primary antibody in 0.2% Triton prepared in PBS and refrigerated overnight. The following day, we washed the segments in PBS (three washes) for 5 mins each wash. We incubated the segments in secondary antibody for 2 hrs at room temperature in the dark. We next washed the antibodies with PBS (three washes). We mounted the glass coverslips on the segments with ProLong Diamond Antifade Mountant (Thermo Fisher Scientific) and incubated them at room temperature overnight and then stored them in 4 degrees. We procured the images using a Leica confocal microscope (Buffalo Grove, IL).
Hair cell count in whole mount preparations. We selected two to three random areas from the basal turn from each treatment group separately. We counted hair cells in 50 μm and 25 μm segments of the cochlea respectively. Counts were averaged across treatment group. This was further normalized as % of PBS (control) by taking the values of control as 100%.