Targeting STAT3 prevents bile reflux‐induced oncogenic molecular events linked to hypopharyngeal carcinogenesis

Abstract The signal transducer and activator of transcription 3 (STAT3) oncogene is a transcription factor with a central role in head and neck cancer. Hypopharyngeal cells (HCs) exposed to acidic bile present aberrant activation of STAT3, possibly contributing to its oncogenic effect. We hypothesized that STAT3 contributes substantially to the bile reflux‐induced molecular oncogenic profile, which can be suppressed by STAT3 silencing or pharmacological inhibition. To explore our hypothesis, we targeted the STAT3 pathway, by knocking down STAT3 (STAT3 siRNA), and inhibiting STAT3 phosphorylation (Nifuroxazide) or dimerization (SI3‐201; STA‐21), in acidic bile (pH 4.0)‐exposed human HCs. Immunofluorescence, luciferase assay, Western blot, enzyme‐linked immunosorbent assay and qPCR analyses revealed that STAT3 knockdown or pharmacologic inhibition significantly suppressed acidic bile‐induced STAT3 activation and its transcriptional activity, Bcl‐2 overexpression, transcriptional activation of IL6, TNF‐α, BCL2, EGFR, STAT3, RELA(p65), REL and WNT5A, and cell survival. Our novel findings document the important role of STAT3 in bile reflux‐related molecular oncogenic events, which can be dramatically prevented by STAT3 silencing. STA‐21, SI3‐201 or Nifuroxazide effectively inhibited STAT3 and cancer‐related inflammatory phenotype, encouraging their single or combined application in preventive or therapeutic strategies of bile reflux‐related hypopharyngeal carcinogenesis.


| INTRODUC TI ON
The carcinogenic effect of bile reflux in hypopharyngeal mucosa has been recently documented by our established in vitro and in vivo models. 1- 7 We have shown that acidic bile (BA) can cause a progressive mutagenic effect characterized by a dramatic activation of a characteristic molecular phenotype, 2-4 assigned as 'BA-induced mRNA oncogenic phenotype', including central molecules in head and neck squamous cell carcinoma (HNSCC), such as nuclear factor kappa B (NF-κB) and signal transducer and activator of transcription 3 (STAT3). [8][9][10][11] Exploration of the BA effect using pharmacologic and dietary inhibitors of NF-κB revealed the key role of NF-kB in this process by activating early neoplastic molecular events. [12][13][14][15][16][17][18] However, the role of STAT3 in bile reflux-related hypopharyngeal carcinogenesis remains unclear. We hypothesized that STAT3 contributes substantially to the BA-related oncogenic effect, by inducing transcriptional activation of inflammatory and cancer-related genes, according to previously established BA-induced 'mRNA oncogenic phenotype', 1-8,12-18 in exposed human hypopharyngeal primary cells (HCs) and preserving cell survival.
The STAT3 is considered an oncogene 8,19 and its upregulation has been associated with development and progression of head and neck cancer, 9,20,21 and especially of HPV-negative HNSCC. 10 The mechanism by which STAT3 is activated has been described by others. 8,22,23 STAT3 proteins, as transducers of cytoplasmic signals from extracellular stimuli, including cancer-related cytokines and growth factors, can be activated through Janus-kinase family members (JAK). 24,25 JAKs after their dimerization can phosphorylate a tyrosine residue (Tyr705) of STAT3 within its Src homology 2 (SH2) domains. Then, STAT3 can homodimerize or heterodimerize with other STAT proteins to translocate to the nucleus and activate the transcription of inflammatory and cancer-related genes. 26,27 JAK/STAT3 activation plays a crucial role in inflammatory-related carcinogenesis. 8,25,26 EGFR can also activate STAT3 by its intrinsic tyrosine kinase activity. 9,27 IL-6/ STAT3 signalling 11 has also been previously suggested to play a role in bile reflux-related oncogenesis, through NF-κB activated pathway. [12][13][14][15][16][17][18] To investigate whether STAT3 actually contributes to early oncogenic molecular events, previously linked to bile-induced hypopharyngeal carcinogenesis, 2,4 we targeted the STAT3 pathway by knocking down STAT3 gene expression, as well as by inhibiting its activation using specific STAT3 pharmacologic inhibitors. STAT3 pharmacologic inhibition consisted of blocking (i) the upstream extracellular receptor JAK2, which affects STAT3 phosphorylation; or (ii) the SH2 domain of phosphorylated STAT3, which blocks its dimerization and DNA binding, and therefore preventing the subsequent target-gene transcription. 22,25,28,29 Identification of the mechanism by which STAT3 affects acidic bile-induced oncogenic molecular events may not only contribute to the characterization of its role in hypopharyngeal carcinogenesis but may also be more important to the development of new effective targeted agents in the prevention and therapy of this process.

| Cell culture and treatment conditions
Normal human hypopharyngeal cells (HCs; Celprogen Inc.) were cultured, in Human Hypopharyngeal Normal Cell Culture Media (Celprogen Inc.), at 37°C in humidified air and 5% CO 2 , as previously described. 7,13,16 An intermittent exposure of HCs was performed, to experimental and control media for 7 min, twice per day, for 4-5 days, as previously described. 9 Table S1).
All experimental groups were repetitively exposed to 400 μM of a mixture of conjugated primary bile acids (Sigma Aldrich) as previously described, 1,7,13,16 in serum-free medium (Dulbecco modified Eagle's medium/F12, 1% pen/strep, Gibco ® ). pH at 4.0 was adjusted by 1 M HCl (using a pH meter). 'BA' or acidic bile at pH 4.0, the cutoff of reflux disease 30 was used according to our prior findings. [1][2][3][4]7 Control media included serum-free medium, as used in experimental groups, at either pH 4.0 (acid control) or neutral pH 7.0 (control including vehicle) (Supplementary material; Table S1). After each treatment, experimental or control media were replaced by serum-free media until the next exposure cycle.
Experimental and control groups were cultured in parallel. All the experiments were independently repeated three times. Cells were harvested immediately after the last treatment by 0.05% trypsin-EDTA (Gibco ® ).

| STAT3 knockdown
To knockdown STAT3, STAT3 siRNA (STAT3 siRNA(h); sc-29493, Santa Cruz Biotechnology Inc.) and Control siRNA, as a negative control (Control siRNA-A; sc-37007, Santa Cruz Biotechnology Inc.) were used. STAT3 and Control siRNAs were diluted to a final concentration of 5 nM in serum-free culture medium (Opti-MEM ® serumreduced growth medium, Gibco™ by Thermo Fisher Scientific), including HiPerFect ® Transfection Reagent (3 μl/well; Qiagen), according to the manufacturers' instructions, and incubated for 10 min at room temperature. Cells were mixed with transfected complexes, seeded at 1.5 × 10 5 cells/well of six-well plates and incubated under normal growth conditions (at 37°C and 5% CO 2 ) (Supplementary Material; Table S2).
Sixteen hours after transfection, media were changed with complete growth medium and 4 h later were treated with 'BA' for 7 min. The media were removed and replaced with serumfree medium until the next day, we repeated the 7-min treatment two times with an interval time 6 h. After 12 h, cells were treated for 7-min and immediately after this last treatment media were removed, the cells were washed once with PBS and harvested either (i) for total protein isolation, using M-PER reagent (mammalian protein extraction reagent; Thermo Scientific), or (iii) for total RNA isolation using RNA mini kit (Qiagen). Assays were carried out according to the manufacturer's instructions and performed in triplicate. All experiments were independently repeated two times.

| Luciferase assay
Luciferase assay was performed to measure the transcriptional activity of STAT3 dimers (homodimers or heterodimers) in HCs exposed to (a) BA with knockdown of STAT3 or (b) BA with pharmacologic inhibitors of STAT3, Nifuroxazide, S3I-201 or STA-21, compared to BA alone and controls, as described in Supplementary Material (Tables S1 and S2). STAT3 dual-luciferase reporter assay was used (Cignal reporter assay by Qiagen), including (i) a firefly luciferase reporter for STAT3 and a constitutively expressing Renilla luciferase construct (Creport-STAT3), and (ii) a cignal negative control with a non-inducible reporter construct and a constitutively expressing Renilla luciferase construct (Creport-NC). A reverse transfection was performed, using Lipofectamine ® 2000 (Invitrogen™), according to manufacturer's procedure.

| Immunofluorescence cell staining
Immunofluorescence (IF) assay was performed for p-STAT3 (Tyr705) and p-NF-κB (p65 S536), as previously described 9 Zeiss Confocal microscope and imaging software by Zen were used to examine stained slides and captured images, respectively (Zen imaging software, Carl Zeiss, microscopy GmbH). 9,10 Expression levels of p-STAT3 and p-NF-κB were assessed by fluorescence intensity (mean ± SD bin count) from at least two intendent images (>10 cells; Zen imaging software).

| Protein expression analysis
Western blot analysis and a direct enzyme-linked immunosorbent assay (ELISA) were performed, as previously described 7,13,16,31 and in Supplementary Material, to monitor the successful knockdown of

| Quantitative real-time polymerase chain reaction
Quantitative real-time polymerase chain reaction (qPCR) analysis (Bio-Rad real-time thermal cycler CFX96™; Bio-Rad) was performed, as previously described 7,13,16 Table S3) and data were analysed by CFX96™ software. 7,13,16 Relative mRNA expression levels were estimated for each target gene compared to the reference control gene (ΔΔCt).

| Cell viability assay
Cell Titer-Glo ® Luminescent Cell Viability Assay (Promega Corp.) was used as described in Supplementary Material. All values were normalized to mean value of untreated controls. All experimental groups and controls were performed in triplicates.

| Knockdown of STAT3 reduces BA-induced total STAT3, p-STAT3 and Bcl-2 protein levels
Immunofluorescence (IF) assay revealed that STAT3 knockdown prevented the BA-induced activation of STAT3, as indicated by weak nuclear straining of p-STAT3 (Tyr705) in treated HCs compared to BA alone ( Figure 1A-a). IF also revealed that STAT3 knockdown had a minimal effect on BA-induced activation of NF-κB, as shown by a less intense nuclear staining of p-NF-κB (p65 S536) relative to BA alone ( Figure 1A-b). Western blot analysis confirmed that silencing of STAT3 effectively suppressed STAT3 protein levels produced by BA exposure in HCs ( Figure 1B). Quantification by ELISA also showed that STAT3 knockdown significantly reduced the BA-induced total p-STAT3 (Tyr705; Figure 1C-a) and Bcl-2 protein levels ( Figure 1C-b).

| Pharmacologic inhibition of STAT3 prevents BA-induced STAT3 activation and Bcl-2 overexpression
Immunofluorescence (IF) assay presented that BA induced an intense nuclear staining of p-STAT3 (Tyr705), as it was expected. 7,14 However, application of Nifuroxazide, which blocks the upstream extracellular receptor JAK2, affecting STAT3 phosphorylation, 28  IF assay also revealed that application of Nifuroxazide, SI3-201 or STA-21, diminished the BA-induced activation of NF-κB, as shown by a less intense nuclear staining of p-NF-κB (p65 S536) relative to BA alone ( Figure 2B-a). However, scoring of IF staining did not reveal a significant difference between p-NF-κB (p65 S536) nuclear levels measured in HCs exposed to BA plus STAT3 pharmacologic inhibitors, compared to BA alone ( Figure 2B-b).
In order to confirm the above IF data, and further determine if targeting JAK/STAT3 phosphorylation can affect the acidic bileinduced STAT3 activation, Bcl-2 overexpression and NF-κB activation, we performed Western blot analysis. Our analysis revealed that Nifuroxazide inhibited the activation of STAT3 and Bcl-2 overexpression, caused by acidic bile. This was shown by significantly reduced nuclear levels of p-STAT3(Tyr705) and cytoplasmic levels of Bcl-2 in HCs exposed to BA plus Nifuroxazide compared to BA alone ( Figure 2C-a,b). However, our analysis showed that Nifuroxazide did not reduce BA-induced nuclear levels of p-NF-κB ( Figure 2C-c).

| STAT3 knockdown or pharmacologic inhibition of STAT3 reduces BA-induced STAT3 transcriptional activity
Our analysis revealed a significant reduction of STAT3-reporter luciferase activity in BA-treated HCs with STAT3 knockdown, compared to BA alone (BA plus STAT3 siRNA vs. BA plus Control siRNA; Figure 3A). Acid and neutral control-treated groups appeared with low levels of STAT3 luciferase activity; however, silencing of STAT3 gene also suppressed its levels (STAT3 siRNA vs. Control siRNA).
Our analysis, using STAT3-reporter luciferase assay and Nifuroxazide, SI3-201 or STA-21, also revealed that pharmacologic inhibition of STAT3 affected the BA-induced STAT3 transcriptional activity in HCs, as similarly shown by knocking down of STAT3.
Specifically, our analysis showed that BA-treated HCs presented significantly higher STAT3-transcriptional activity, compared to controls ( Figure 3B). However, application of all three pharmacologic inhibitors significantly reduced BA-induced STAT3-transcriptional activity. This was indicated by significantly lower levels of luciferase activity for STAT3-reporter, in HCs exposed to BA plus STAT3 pharmacologic inhibitors, compared to BA alone ( Figure 3B).

| Silencing or pharmacologic inhibition of STAT3 affects cell viability of BA-treated HCs
We performed a cell viability assay to explore the effect of STAT3 knockdown or its pharmacologic inhibition on BA-induced cell survival rates. Our analysis revealed that silencing STAT3 (siRNA STAT3;

| DISCUSS ION
Chronic exposure to BA has recently been shown to cause malignant transformation of hypopharyngeal epithelial cells which preceded by significant transcriptional activation of cancer and inflammatory-related genes, previously assigned as 'BA-induced mRNA oncogenic phenotype', including oncogenic STAT3. [2][3][4][5] Interestingly, the oncogenic molecular phenotype induced by BA in vitro and in vivo was similarly identified in clinical specimens from bile-related HSCC, demonstrating an aberrant overexpression of STAT3 compared to controls. 39 The role of STAT3 as a crucial transcription factor in HNSCC has been discussed extensively by others. 9,10,[20][21][22]27 Although other STAT factors, such as STAT1 and and REL, 36,[38][39][40] previously linked to BA-induced oncogenic effect. 4,41 Silencing STAT3 also reduced BA-induced cell survival rates. These findings strongly support the hypothesis that STAT3 plays an important role in bile reflux-related oncogenic molecular events in HCs. 8,9,11,26,27,[32][33][34][35][36][37][38][39][40][41] Our findings from the pharmacologic inhibition of STAT3, using Nifuroxazide, 28 S3I-201 22 or STA-21, 29 demonstrated a very similar effect to its gene silencing, supporting their effective application on reducing the BA-induced oncogenic profile. Using three different inhibitors that each one can block a different step of STAT3 upstream signalling, we gained information that improve our knowledge regarding the mechanism of bile reflux-related carcinogenesis in hypopharynx. In theory, Nifuroxazide inhibits the JAK/STAT3 phosphorylation, while SI3-201 and STA-21 inhibit STAT3 dimerization and its subsequent translocation to the nucleus. As we know, STA-21 can block STAT3 dimerization, DNA binding and STAT3 dependent transcription, when STAT3 is constitutively active. Herein, we show that all three inhibitors were similarly effective to suppress STAT3 activation, with STA-21 presenting the most pronounced reduction of BA-related mRNA phenotype and cell survival. This view may support a constitutive activation of STAT3 in HCs, caused by BA, which does not seem to be exclusively depended on JAK/STAT3 upstream signal, 25 but also to alternative signalling, such as EGFR. 22 EGFR cross-talk with STAT3 has previously been identified to affect activation of STAT3 via tyrosine kinase, 44 while Bhat et al., recent findings supported EGFR and STAT3 interactions in oesophageal precancerous and cancerous cells under the acidic bile effect. 45,46 Further investigation is required to elucidate possible similar EGFR-STAT3 interactions in acidic bile-related hypopharyngeal squamous cell cancer. 41 Notably, both SI3-201 and STA-21 appeared to produce a relatively more pronounce suppressive effect on BA-induced mRNA phenotype than Nifuroxazide (Table 1) Besides the above variations, our data demonstrated that similarly to STAT3 silencing, all three pharmacologic drugs suppressed the transcriptional activation of IL6, TNF, STAT3, BCL2, RELA(p65) and EGFR (Table 1). This observation strongly supports that the constitutive activation of STAT3 can maintain the continuous production of inflammatory and cancer-related molecules that are central to the carcinogenic process. 4,41,47 Therefore, we confirm that BAinduced molecular oncogenic events can be effectively prevented by the application of either Nifuroxazide, 28 S3I-201, 22 or STA- 21. 29 IL6 plays a central role in inflammatory and cancer promotion of HNSCC and is considered as a promising therapeutic target. 22,25,27,35 Our current findings showed that BA effect can induce an abundant overexpression of IL6 in exposed HCs, as previously shown in premalignant or malignant murine hypopharyngeal mucosa and in bile-related HSCC. 4,41 IL6 overexpression can promote a persistent activation of STAT3, by the IL6/STAT3 pathway. 25,35 Our data show that silencing or pharmacologic inhibition of STAT3 can effectively reduce IL6 transcriptional levels and thus limit potential positive feedback.
Although we show that targeting STAT3, either by its knockdown or its pharmacologic inhibition, had a minimal effect on total p-NF-κB (p65 S536) levels, our findings strongly support a role of STAT3 in promoting the transcriptional activation of NF-κB. NF-κB inhibition, using BAY 11-7082, had previously supported the role of NF-κB in BA-induced activation of STAT3, among other factors. [12][13][14][15][16][17][18] In view of all the above, we propose a possible molecular cross-talk between the NF-κB and STAT3 transcription factors signatures in bile reflux-related inflammation and tumorigenesis in hypopharynx, as has been proposed by others in HNSCC 8,10,11 that requires further exploration.
Given the central role of STAT3 pathway in HNSCC, despite its complexity, our data provide supportive evidence of STAT3 inhibition as a preventive and therapeutic approach in hypopharyngeal cancer. STAT3 is indeed a demanding target not only because of the complexity of its activating signal but also because of its resistance to standard treatment in HNSCC. 21,22 However, recent preclinical studies have shown the antiproliferative effect of STAT3 inhibition, 28,[48][49][50][51][52] and the latest clinical trials have presented very promising anti-oncogenic phenotype under STAT3-target therapy. 22 SI3-201 has been extensively tested as an antitumour target in HNSCC, since it is promising to abrogate resistance to anti-EGFR therapies. 46,[48][49][50] Nifuroxazide has also been proved to be a safe drug with antitumour and anti-inflammatory effects. 51,52 Finally, STA-21 has been clinically tested and found to improve chronic inflammatory disorders. 53 Our data from the use of the three pharmacological inhibitors of STAT3 strongly support their use as promising drugs for the prevention or treatment of inflammatory or neoplastic diseases associated with laryngopharyngeal reflux.
It was very recently presented an association between bile acidreceptors, such as nuclear farnesoid X receptors (FXRs), 54

CO N FLI C T O F I NTE R E S T
The authors whose names are listed in this article certify that they have no affiliations with or involvement in any organization or entity with any financial or non-financial interest in the subject matter or materials discussed in this manuscript.