NF‐κB inhibition reverses acidic bile‐induced miR‐21, miR‐155, miR‐192, miR‐34a, miR‐375 and miR‐451a deregulations in human hypopharyngeal cells

Abstract We previously demonstrated that acidic bile activates NF‐κB, deregulating the expression of oncogenic miRNA markers, in pre‐malignant murine laryngopharyngeal mucosa. Here, we hypothesize that the in vitro exposure of human hypopharyngeal cells to acidic bile deregulates cancer‐related miRNA markers that can be reversed by BAY 11‐7082, a pharmacologic NF‐κB inhibitor. We repetitively exposed normal human hypopharyngeal primary cells and human hypopharyngeal keratinocytes to bile fluid (400 μmol/L), at pH 4.0 and 7.0, with/without BAY 11‐7082 (20 μmol/L). We centred our study on the transcriptional activation of oncogenic miR‐21, miR‐155, miR‐192, miR‐34a, miR‐375, miR‐451a and NF‐κB‐related genes, previously linked to acidic bile‐induced pre‐neoplastic events. Our novel findings in vitro are consistent with our hypothesis demonstrating that BAY 11‐7082 significantly reverses the acidic bile‐induced oncogenic miRNA phenotype, in normal hypopharyngeal cells. BAY 11‐7082 strongly inhibits the acidic bile‐induced up‐regulation of miR‐192 and down‐regulation of miR‐451a and significantly decreases the miR‐21/375 ratios, previously related to poor prognosis in hypopharyngeal cancer. This is the first in vitro report that NF‐κB inhibition reverses acidic bile‐induced miR‐21, miR‐155, miR‐192, miR‐34a, miR‐375 and miR‐451a deregulations in normal human hypopharyngeal cells, suggesting that acidic bile‐induced events are directly or indirectly dependent on NF‐κB signalling.

MicroRNA (miRNA) molecules have been considered to play an important role in both inflammation and cancer, 18 modulating the expression of genes by causing target mRNA degradation or inhibiting their translation. 19 Specifically, some miRNAs, such as "oncomirs" and "tumour suppressor" miRNAs, show altered expression levels in tumour cells compared to normal cells (up-regulated or down-regulated) and are capable of contributing to carcinogenesis, demonstrating a significant regulating role in the multistep process of cancer initiation and progression. 20 Previous studies have demonstrated that deregulation of "oncomirs" miR-21, miR-155, miR-192 and tumour suppressor miR-375, miR-451a and miR-34a is associated with laryngopharyngeal cancer. [12][13][14][15][16][17] Moreover, an independent association has been demonstrated between NF-jB activation and up-regulation of oncogenic miR-21 and/or down-regulation of tumour suppressor miR-34a and miR-451a. [21][22][23] BAY 11-7082 was selected as a reliable inhibitor of NF-jB pathway that has been widely used in many studies exploring the effect of NF-jB. 11,24,25 It has been suggested that BAY 11-7082 offers the most rapid and potent antitumour effect among other NF-jB inhibitors 24 and can possibly be used as a sensitizer of anticancer therapy, 26,27 increasing the intrinsic susceptibility of cancer cells to chemotherapeutic agents. 28 Evidence that inhibition of acidic bile-induced NF-jB activation effectively reverses the altered cancer-related miRNA phenotype will encourage the in vivo application of NF-jB inhibitors, as possible preventers of acidic bile effect in hypopharyngeal mucosa. telomerase-immortalized human hypopharyngeal keratinocytes (HHK) (4th passage) 8 to bile (pH 4.0 and pH 7.0) and corresponding controls (pH 4.0 and pH 7.0), for 10-15 minutes, 3 times per day, for 5 days, as previously described. 8 Bile fluid consisted of a mixture of conjugated bile salts (400 lmol/L) considered to be "physiologic" 29,30 (Supplementary Methods).
Experimental groups included HHPC and HHK repetitively exposed to (a) acidic bile at pH 4.0, the cut-off of reflux disease 31,32 and (b) neutral bile, containing the same bile salts mixture, at pH 7.0.
Control groups included (a) acid control (pH 4.0), and (b) neutral control (pH 7.0) with identical media used in experimental groups (Supplementary Methods).

Acidic bile + BAY-11-7082 treatment
In parallel with acidic bile treatment, we performed an additional procedure of repetitive exposure of HHPC (2nd passage) and HHK At the end of treatment, media were removed and cells or cell extracts were analysed.

| Luciferase assay
We performed a luciferase assay in order to monitor the transcriptional activity of the NF-jB in HHPC exposed to acidic bile and corresponding controls, with or without the pharmacologic inhibitor of NF-jB, BAY 11-7082. We used Firefly | 2923 as ratios of relative NF-jB activity (with/without NF-jB inhibitor).
(Data were obtained from three independent experiments).
We used the same pool of total RNA to determine, by qPCR, the effect of BAY 11-7082 on transcriptional levels of RELA(p65), TNFa, IL-1b, IL-6 and STAT3 in acidic bile-treated and control HHPC with or without BAY 11-7082, as previously described 33 Table S2). These genes were selected because they demonstrated an increased transcriptional activity under acidic bile exposure of HHPC 8,9 that was prevented by BAY 11-7082 in our previous study. 33

| Cell viability assay
We performed a cell viability assay, using CellTiter-Glo â Luminescent Cell Viability Assay (Promega) to monitor the effect of NF-jB inhibitor, BAY 11-7082, on viability of HHPC and HHK treated with bile at pH 4.0 and pH 7.0, and corresponding controls, as described in Supplementary Methods. We determined cell viability by comparing the mean values of cells exposed to NF-jB inhibitor against the mean value of cells that were not exposed to inhibitor, for each experimental and control group. Statistically significant difference in cell viability was determined using paired test and P value <.05 (Graph Pad Prism 6.0).

| Statistical analysis
We performed statistical analysis, using GraphPad Prism 6 software and one-way ANOVA (by Friedman and Dunn's multiple analysis test; P-values <.05) to compare expression changes of the analysed miRNA markers and the analysed genes induced by BAY 11-7082 (with/without NF-kB inhibitor) between different experimental and control groups. We also used t test analysis (multiple comparisons by Holm-Sidak) to reveal differential expression (P-values) for each analysed miRNA marker in treated cells, with and without NF-jB inhibitor.
Finally, we performed a Pearson correlation to estimate the correlation coefficient between BAY 11-7082-induced miRNA and mRNA expression levels, as well as between "oncomirs" and "tumour suppressor" miRNA levels, of different treated groups (P-values < .05). NF-jB in acidic bile-treated HHPC and HHK. We observed that cells exposed to neutral bile, acid or neutral control with NF-jB inhibitor also demonstrated a reduced transcriptional activity of NF-jB, compared to those treated without BAY 11-7082 ( Figure 1A). However, the acidic bile-treated group demonstrated the most reduced ratios of relative NF-jB transcriptional activity (NF-jB luciferase responsive element/control luciferase reporter) with/without BAY 11-7082, relative to neutral bile, acid or neutral control groups ( Figure 1B).

| NF-jB inhibitor induces a reversed
cancer-related miRNA phenotype in acidic bile-treated normal human hypopharyngeal cells, relative to controls We observed that NF-jB inhibitor induced an inverted miRNA phenotype of "oncomirs" and particularly of "tumour suppressor" miRNAs in between acidic bile and control groups (P = .0442 and P = .0139, respectively, by Friedman). The reversal of miRNA phenotypes, by NF-jB inhibitor, was particularly intense in acidic bile-treated HHPC, demonstrating a significant decrease in expression ratios of "oncomirs" ( Figure 4A-a) and increase in expression ratios of "tumour suppressor" miRNAs (with/without BAY 11-7082), compared to control (P = .0269, one-way ANOVA; by Kruskal-Wallis) ( Figure 4A-b). We also observed an inverted miRNA phenotype, by NF-jB inhibitor, in acidic bile-treated HHK, demonstrating a significant decrease in expression ratios of "oncomirs" (P = .0114, one-way ANOVA; by Kruskal-Wallis) ( Figure 5A-a) and an increase in expression ratios of "tumour suppressor" miRNAs (with/without BAY 11-7082), compared to control ( Figure 5A-b).
We further showed HHPC treated with acidic bile without BAY 11-7082 demonstrated a significantly higher miR-21/375 ratio, compared to neutral control or neutral bile, (P = .0022, and P = .0415, respectively, by Kruskal-Wallis) ( Figure 4B), and similarly HHK treated with acidic bile without NF-jB inhibitor showed a significantly higher miR-21/375 ratio compared to neutral control or acid alone (P = .0022 and P = .0415, respectively, by Kruskal-Wallis). In contrast, we observed that BAY 11-7082 induced a significant reduction in miR-21/375 ratios in acidic bile-treated HHPC, compared to neutral control and neutral bile (P = .0022 and P = .0415, respectively) ( Figure 4C-b), as well as in HHK, compared to neutral control and acid alone (P = .0022 and P = .0415, respectively) ( Figure 5C-b).
These observations suggest that NF-jB inhibition is capable of decreasing the acidic bile-induced elevated miR-21/375 ratios in both HHPC and HHK.

| Correlations between BAY 11-7082-induced changes of oncogenic miRNA markers and NF-jB-related genes in treated human hypopharyngeal primary cells
To determine the correlations between BAY 11-7082-induced expression changes (with/without BAY 11-7082) of the analysed oncogenic miRNA markers and NF-jB-related genes, in the same treated groups, we performed qPCR analysis from the same pool of total RNA. F I G U R E 3 NF-jB inhibitor (20 lmol/L ΒΑΥ 11-7082) reverses the acidic bile-induced deregulation of cancer-related miRNA markers, in human hypopharyngeal keratinocytes (HHK). A, Acidic bile induces in HHK (a) an up-regulation of the analysed "oncomirs", demonstrated by significantly higher miRNA levels, compared to controls, inverted by NF-jB inhibitor (BAY 11-7082). (b) Acidic bile induces a down-regulation of the analysed "tumour suppressor" miRNAs, demonstrated by significantly lower expression levels, compared to controls, that is also inverted by NF-jB inhibitor (BAY 11-7082) in HHK (one-way ANOVA; by Friedman; *P < .05; **<P < .005; GraphPad Prism 6.0). B, Graphs depict the significantly (a) decreased expression levels of miR-21, miR-155 and miR-192 and (b) increased expression levels of miR-34a, miR-375 and miR-451a, in HHK exposed to acidic bile with NF-jB inhibitor (BAY 11-7082), compared to HHK exposed to acidic bile without NF-jB  form of IL-6 up-regulation. 34 Our data from immunofluorescence assay demonstrated that the acidic bile-induced activated STAT3 and NF-jB were inhibited by NF-jB inhibitor in treated HHPC ( Figure 6). Specifically, we observed that acidic bile-treated HHPC showed an intense nuclear staining for both phospho-NF-jB (p-p65 S556) ( Figure 6A) and phospho-STAT3 (Tyr705) ( Figure 6B

| DISCUSSION
Nuclear factor kappa B (NF-jB) is a key factor that mediates inflammatory and early tumorigenic events in epithelial cells, 35  that are related to activated NF-jB and that may contribute to an aggressive phenotype of head and neck cancer. 23,38 Here, we present the first in vitro report that bile and acid com- Whereas the ratio of miR-21/375 has been considered a potential biomarker related to poor prognosis of supraglottic cancer, 13,14 our data demonstrate a significant increase in miR-21/375 ratios in cells exposed to acidic bile (pH 4.0), compared to controls, a relationship that is effectively inverted in the presence of NF-jB inhibitor, again strongly supporting the miR-21/375 ratio, as a potential biomarker in acidic bile-induced cancer-related molecular events in hypopharyngeal cells, mediated by NF-jB pathway.
Our novel findings also demonstrate that NF-jB inhibition significantly prevents the acidic bile-induced up-regulation of miR-155. Our current findings also demonstrate that miR-451a could be an important marker of acidic bile-related laryngopharyngeal carcinogenesis, in agreement with Fukumoto et al who previously suggested miR-451a is a tumour suppressor marker in hypopharyngeal SCC. 55 This in vitro measure is in line with our 45-day in vivo model where a significant down-regulation of miR-451a occurred. 10 We also showed a significant effect of NF-jB inhibition on "tu- suppressor and key regulator miRNA in HNSCC, 17,23 and our prior in vivo findings showed that acidic bile down-regulated miR-34a levels in treated laryngopharyngeal mucosa. Although previous studies reported NF-jB binding sites on the promoter of miR-34a, 21,56 and an increase in miR-34a levels in oesophageal cells under NF-jB activation, 21 the exact mechanism of miR-34a regulation by NF-jB is not yet obvious. Alternatively, it has been shown that STAT3 can directly repress miR-34a, while an active IL-6R/STAT3/miR-34a loop was found necessary for EMT, invasion and metastasis of colorectal cancer cell line. 56 Our current data support the notion that NF-jB inhibition is capable of preventing down-regulation of miR-34a by acidic bile.
As would be expected, the individual response of HHPC and HHK was not always congruent. The ability of BAY 11-7082 to reverse acidic bile-induced phenotype of "oncomir" miR-192 was more prominent in HHPC compared to HHK. Likewise, the effect of NF-jB inhibition regarding "tumour suppressor" miRNAs miR-34a and miR-451a was more intense in HHPC relative to HHK under acidic bile exposure. We are of the understanding that effects may be related to differences in maturational status. Primary cells HHPC, considered to be less mature and more sensitive to injurious external stimuli than immortalized keratinocytes HHK, may respond differently to NF-jB inhibition. The miRNA phenotype observed in HHPC was in fact similar to that observed in vivo, an observation in support of this view. 10 Our novel data demonstrate strong inverted correlations among the BAY 11-7082-induced levels of the analysed tumour suppressors" miR-34a, miR-375 and miR451a, and NF-jB-related genes, such as RELA(p65), STAT3, TNF-a, IL-6 and IL-1b, that previous studies documented as crucial mediators of inflammatory and neoplastic events in head and neck cancer [35][36][37][38][39][40][41][42][43][44][45][46] (Figure 7). In line with our recent study, 33 our current data suggest that NF-jB inhibition may reverse acidic bile-induced molecular events in normal human hypopharyngeal cells that are known to link inflammation to cancer, thereby in a sense shielding HHPC from the effects of bile-induced oncogenic molecular events.
We showed that BAY 11-7082 is also capable of preventing acid alone-induced deregulations of miR-21 and miR-192, but not miR-155 or "tumour suppressor" miRNAs in HHPC, suggesting that NF-jB inhibition could reverse a part of acid-induced miRNA phenotype in normal human hypopharyngeal cells. Although acid alone may upregulate selected "oncomirs", it was not capable of accelerating activation of oncogenic STAT3 or other cancer-related molecular events.
Similarly, bile at neutral pH seemed to contribute to deregulations of "tumour suppressor" miRNAs, such as miR-34a and miR-451a, but was not capable of accelerating activation of oncogenic STAT3 or other cancer-related molecules. Our data support the observation that acid and bile in combination but not acid or bile alone may contribute to cancer-related molecular events, mediated by NF-jB.
Our findings revealed that NF-jB inhibition resulted in a significant reduction in viable cells, particularly in acidic bile-treated groups. The identified decreased cell viability induced by 7082 is in line with previous studies. 27 The preferential effect of BAY 11-7082 in acidic bile-treated groups is especially interesting because it suggests its effect on cell viability is not global but related to specific events, in those groups. Interestingly, other studies have shown that increased expression of miR-34a and miR-375 reduces cell viability of cancer cell lines. 57,58 In this regard, our present data also showed that BAY 11-7082 decreases the levels of "oncomirs", miR-155, miR-192 and miR-21 levels, while increasing the levels of "tumour suppressors", miR-34a and miR-375, raising the view that specific cancer-related miRNAs interactions may be at least partially responsible for decreased cells viability. and -451a in treated hypopharyngeal mucosa.

CONFLI CT OF INTEREST
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 interest or non-financial interest in the subject matter or materials discussed in this manuscript.

AUTHOR CONTRI BUTIONS
DV, SD and CTS contributed to conceptualization and data curation.