MiR-21 is a Predictor for Chemoradioresistance and a Novel Therapeutic Target in Head and Neck Squamous Cell Carcinoma


 ObjectiveThis study aimed to clarify whether circulating miR-21 represents a predictive biomarker in patients with head and neck squamous cell carcinoma (HNSCC) undergoing chemoradiotherapy, and to investigate the effect of miR-21 inhibitor for chemoradiation in human SCC cells.MethodsPlasma samples were obtained from 22 patients with HNSCC and 25 non-cancer volunteers. Plasma miR-21 expression was measured using real-time quantitative reverse transcription polymerase chain reaction. The effects of miR-21 inhibitor in human SCC cells were investigated by performing 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, flow cytometry, and Western blot analysis.ResultsPlasma miR-21 expression was higher in HNSCC patients than in control patients (p< 0.001). Seven patients with recurrence showed significantly higher plasma miR-21 than the 15 patients without recurrence. Moreover, miR-21 inhibition significantly enhanced cisplatin- or radiation-induced apoptosis. Western blot analysis suggested the programmed cell death 4 (PDCD4) protein as a potential target of miR-21 in relation to apoptosis. Adding miR-21 inhibition to radiation or cisplatin treatment provided clear and potent suppression of tumor cell proliferation.ConclusionThis study provides new insights into the role of miR-21 as a predictive biomarker for HNSCC treated with chemoradiotherapy, and suggests a potential target to improve the effects of chemoradiotherapy against HNSCC.

treatment. Investigation of the molecular mechanisms underlying treatment failure and identi cation of predictive biomarkers are thus urgently required.
MicroRNAs (miRNAs) are small endogenous, non-coding RNAs of 18-25 nucleotides, acting as posttranscriptional gene expression regulators by destabilizing or inhibiting translation of mRNAs [6]. These miRNAs not only regulate gene-expression, but also are involved in carcinogenesis, such as cell proliferation, apoptosis, angiogenesis [7]. One of the most well-known miRNAs is miR-21, an oncomiR (i.e., an miRNA associated with cancer) for carcinomas such as HNSCC, esophageal squamous cell carcinoma (SCC), gastric carcinoma, lung carcinoma, colorectal carcinoma, breast cancer and pancreatic carcinoma [8, 9,10,11,12]. Increased expression of miR-21 has been deeply involved with poor prognosis of various cancers due to its oncogenic roles in processes such as cell proliferation, migration, and invasion [9]. The potential utility of miR-21 as a predictive biomarker for oncological treatment was recently elucidated, with Atrantes et al. revealing that miR-21 obtained from tissue samples was able to predict response to an organ-preservation protocol based on chemoradiation in HNSCC [13]. However, whether miR-21 circulating in blood can predict chemoradioresistance in patients with HNSCC remains unclear, despite its potential as a strong and useful biomarker in clinical situations.
In addition, emerging evidence has demonstrated that miR-21 modulates the chemosensitivity of cancer cells primarily by targeting phosphatase and tensin homolog (PTEN) or PDCD4 [14,15,16,17]. With regard to the chemoresistance induced by miR-21 in HNSCC, a few reports have shown that miR-21 induced chemoresistance in oral SCC cells [17], but not in pharyngeal or laryngeal cell carcinomas, even though chemoradiotherapy is the standard therapy for advanced oropharyngeal, hypopharyngeal and laryngeal carcinomas. Whether radioresistance is induced by miR-21 in HNSCC remains unclear.
Our aim in the study was to clarify whether circulating miR-21 offers a possible predictive biomarker for patients with locally advanced HNSCC undergoing platinum-based chemoradiotherapy, and to investigate the effects of miR-21 inhibitor together with chemoradiotherapy in human HNSCC cells. Our study revealed circulating miR-21 levels were signi cantly higher in the group with tumor recurrence after chemoradiotherapy than in the group without recurrence. Next, we showed that miR-21 inhibition signi cantly suppressed cell proliferation and also promoted cisplatin-and radiation-induced apoptosis in human hypopharyngeal SCC cells. This study provides new insights into the role of plasma miR-21 in HNSCC as a predictive biomarker for chemoradiotherapy, and miR-21 inhibition during chemoradiotherapy may improve the prognosis for advanced HNSCC.  Table 2. This study was approved by the ethics committee at Mie University Graduate School of Medicine (approval nos. 2445 and H2020-232). Written informed consent was obtained from each patient before enrolment in this study.

Patient recruitment and sample collection
Cell culture HNSCC FaDu and SAS cell lines (derived from human hypopharyngeal cell carcinoma and human tongue SCC, respectively) were obtained from RIKEN BRC Cell Bank (Tsukuba, Japan). These HNSCC cell lines were maintained in modi ed Eagle's medium (MEM) medium or RPMI-1640 medium with 10% fetal bovine serum and antibiotics (1% penicillin, streptomycin), respectively. All human cell lines have been authenticated using short tandem repeat (STR) pro ling this year.
RNA extraction from plasma RNA was extracted from 200 µl of plasma by using an miRNeasy Serum/Plasma Kit (QIAGEN, Hiden, Germany) in the automated QIAcube (QIAGEN) as described previously [10], according to the manufacturer's instructions. A synthetic Caenorhabditis elegans miR-39 miRNA mimic (QIAGEN) and carrier RNA (0.94 µg, MS2 bacteriophage total RNA; Roche Applied Sciences, Indianapolis, IN, USA) were spiked-in before RNA extraction. Isolated RNA was eluted in 15 µl of RNase-free water.
Reverse transcription quantitative polymerase chain reaction (RT-qPCR) for plasma miR-21 cDNA was synthesized from 5 µl of eluted RNA (containing miRNAs) in HiFlex buffer (QIAGEN), by using an miScript II RT Kit (QIAGEN) as described previously [10]. Next, quantitative RT-PCR was performed using the miScript SYBR Green PCR kit (QIAGEN) and miScript Primer Assays. MiRNA were normalized against the average of four reference miRNAs (miR-423-5p, miR-103a-3p, miR-191-5p, and miR-93), which were used as internal controls in plasma samples. In addition, syn-cel-miR-39 was used to con rm the extraction e cacy of RNA. All ampli cations were carried out in an ABI Step One Plus Real-time PCR System (Applied Biosystems, Singapore, Singapore). Ampli cation curves were analyzed using SDS software version 2.2.2 (Applied Biosystems). Expression levels of mRNA were determined by using the 2 −ΔΔC method.

Transfection experiments using miR-21 inhibitor
To transiently inhibit miR-21 expression, hsa-miR-21 mirVana™ miRNA inhibitor (Applied Biosystems) or hsa-miR-21 mirVana® mimics (Applied Biosystems) were used to transfect HNSCC cells, as described previously [18]. Veri cation of transfection e ciency was conducted using the mirVana™ miRNA Mimic Negative Control (Applied Biosystems) and mirVana™ miRNA Inhibitor Negative Control (Applied Biosystems), respectively. Forward transfections were conducted by mixing siRNA oligonucleotides (70 nM) with Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA) and Opti-MEM I (Invitrogen) and applying the mixture to cells at 24 h after plating. A series of in vitro assays was conducted after 48-72 h of incubation.
Cell viability assay A 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay was performed to assess the anti-proliferative effects of miR-21 inhibitor and proliferative effects of the miR-21 mimic on HNSCC cells, as previously described [19]. Cells were plated at a density of 4 ⋅10 3 cells/well into a 96-well plate, transfected with miR-21 mimic or inhibitor, and treated with either 48 h administration of cisplatin at various concentrations (0, 0.2, 1, or 5 µg/ml) or three exposures to radiation at various dose (0, 2, or 4 Gy), and incubated for a total of 72 h. Next, 10 µl of 5-mg/ml MTT (Sigma-Aldrich, St Louis, MO, USA) in 100 µl of medium was added to each well, followed by incubation for 4 h at 37°C. The medium was removed and replaced with 100 µl DMSO, and absorbance values were measured at 570 nm on a Bio-Rad model 680 microplate reader (Bio-Rad Laboratories, Hercules, CA, USA).

Cell apoptosis assay
Quantitative analysis of apoptotic and dead cells was performed by using An Annexin V and Dead Cell Assay Kit (Millipore, Burlington, MA, USA) with a ow cytometer (Muse™ Cell Analyzer; Millipore) as described previously [20], according to the instructions from the manufacturer. After incubating with cisplatin (0, 4, or 6 µg/ml) for 48 h or exposure to radiation (0, 4, or 12 Gy), all cells were harvested and diluted to a concentration of 5 ⋅10 5 cells/mL in MEM medium with 2% Fetal bovine serum (FBS). One hundred microliters of Annexin V and Dead Reagent and 100 µl of single cell suspension were then mixed in a microtube and analyzed by using the Muse TM Cell Analyzer (Millipore) after 30 min incubation in the dark at room temperature. All experiments were performed in quadruplicate.

Western blot analysis
After transfection with miR-21 mimic or inhibitor for 48 h, cells were harvested and lysed using RIPA buffer (Cell Signaling Technology, Dancers, MA, USA) supplemented with phenylmethylsulphonyl uoride (Nacalai Tesque, Kyoto, Japan). Equal amounts of protein were runned by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to polyvinylidene uoride membranes (0.45 µm; Millipore). Membranes were washed with Tris-buffered saline (TBST) containing 0.1% Tween-20 (Nacalai Tesque) and 5% bovine serum albumin (Sigma-Aldrich, Saint Louis, MO, USA), and incubated overnight at 4°C with primary antibodies (PDCD4 antibody from Cell Signaling Technology, Danvers, MA, USA; β-actin antibody from MP Biomedicals, LLC, Irvine, CA, USA). After washing with TBST, membranes were further incubated with horseradish peroxidase-conjugated secondary antibody (1:10,000; Santa Cruz Biotechnology, Dallas TX, USA) for 1 h at room temperature, and developed by using an electrochemiluminescence system in the end (GE Healthcare, Little Chalfont, UK). Protein bands were detected by an LAS4000mini imaging system (Fuji lm, Tokyo, Japan), and band intensities of Western blots were quantitatively measured by calculating integrated grayscale densities in consistently sized windows incorporating each band using ImageJ version 1.48 software as described previously [21] .

Statistical analysis
All statistical analyses were performed with SPSS version 23 statistical software. Two-group comparisons of ΔCt values between plasma samples from controls and HNSCC patients, as well as relapse-free and recurrent patients, were performed by using the Mann-Whitney U-test. Receiver operating characteristic (ROC) curves were chosen to evaluate the predictive power of miR-21 for tumor recurrence and to obtain the miR-21 cut-off values offering maximal sensitivity and speci city. The Kaplan-Meier method with log-rank comparison was selected to calculate OS, measured from the date chemoradiotherapy was initiated until the date of death from any cause (i.e., cancer-unrelated deaths were not censored), or last known follow-up for patients who were still alive. Statistical differences between groups were calculated by χ 2 text, unpaired t-test, Mann-Whitney test, and one-way ANOVA with post-hoc multiple comparisons. Values of p<0.05 were considered to be a statistical signi cance.

Results
Plasma miR-21 level and OS rate in HNSCC patients treated with chemoradiotherapy First, peripheral blood samples were collected from 22 patients with advanced HNSCC treated using platinum-based chemoradiotherapy and 25 healthy subjects without HNSCC, and levels of plasma miR-21 expression were detected by qPCR. MiR-21 is an abundantly expressed miRNA in mammalian cells. MiR-21 showed low expression level in steady state, and it would be upregulated in malignancy [22].
As expected, relative levels of miR-21 were signi cantly higher in HNSCC patients than in control subjects ( Fig. 1A; p<0.001). Next, we sought to determine whether expression levels of plasma miR-21 were associated with tumor recurrence after chemoradiation. Our results demonstrated a signi cant difference in plasma miR-21 levels between the tumor-free and recurrent groups ( Fig. 1B; p<0.05). Factors including age, sex, stage, subsite, Brinkman index, CRP, neutrophil-lymphocyte ratio and SCC antigen were compared between the tumor-free and recurrent groups. The difference between groups was not signi cant (p>0.05; Table 2). Though we also examined the correlation plasma miR-21 and several clinical parameters like CRP, neutrophil-lymphocyte ratio and SCC antigen, there were no correlation. This result implied that plasma miR-21 could offer a predictive biomarker for tumor recurrence after chemoradiotherapy in HNSCC. We generated the ROC curve to assess the potential utility of plasma miR-21 as a biomarker for early detection of recurrent HNSCC after chemoradiotherapy (Fig. 1C). The area under the ROC curve (AUC) for plasma miR-21 was 0.7771. Using a cut off value of 1.295, miR-21 showed 85.7% sensitivity and 66.7% speci city. These ndings highlighted the potential for plasma miR-21 as a useful biomarker for predicting recurrent HNSCC after chemoradiotherapy.
Several reports have suggested miR-21 as a candidate prognostic biomarker for HNSCC [12,13,23], and a relationship between high miR-21 expression and tumor recurrence was identi ed in the present study. However, whether tumor recurrence exerts a big impact on prognosis has remained unclear. We therefore assessed the correlation between tumor recurrence and OS. The results demonstrated a signi cant difference in OS between the tumor-free and recurrent groups (Fig. 1D), in turn suggesting the importance of miR-21 as a prognostic biomarker.

Increased sensitivity to radiotherapy in HNSCC cells by miR-21 inhibitor
Having identi ed plasma miR-21 expression level as increased in the group with recurrent HNSCC comparing to the tumor-free group in this study, the role of miR-21 in chemoradiotherapy for HNSCC is still unknown. We hypothesized that miR-21 would play an important role in chemoradioresistance to platinum-based chemoradiotherapy for HNSCC. First, we con rmed expression levels of miR-21 by RT-qPCR analysis in human HNSCC cells (FaDu cells derived from hypopharyngeal SCC) with transfection of either miR-21 mimic or miR-21 inhibitor. As expected, we could con rm that miR-21 inhibitor suppressed miR-21 expression ( Fig. 2A) and miR-21 mimic induced miR-21 expression (Fig. 3A). Although miR-21 has been shown to be involved in the radioresistance of several types of cancer, including breast cancer, lung cancer, cervical cancer, and glioma [24,25,26,27], whether miR-21 inhibitor enhances the radiosensitivity of human HNSCC cells is still unclear. We therefore next assessed the effects of miR-21 inhibitor on treatment with radiation by performing an MTT proliferation assay. We determined the RT dose for MTT assay because patients with HNSCC are treated with single dose 2Gy (totally 60-70 Gy in 30-35 fractions over 6-7 weeks) in conventional radiotherapy [28]. Cell proliferation was signi cantly suppressed by miR-21 inhibitor compared with radiation alone (Fig. 2B, 2C). Similar results were also observed in the SAS cell line, as human tongue SCC cells (Fig. 2F). Consistent with these results, MTT assay results showed radioresistance with miR-21 mimic transfection (Fig. 3B). Taken together, these results suggest that miR-21 is deeply involved in the radiosensitivity of HNSCC cells.

Increased sensitivity to chemotherapy in HNSCC cells by miR-21 inhibitor
With regard to the chemoresistance of HNSCC by miR-21, we also investigated whether miR-21 is involved in the chemosensitivity of HNSCC treated with miR-21 inhibitor or mimic using the MTT assay. According to Nagai et al [29], high dose CDDP administration (80mg) after continuous infusion over 4 hours and over 2 hours provided 1.39±0.79µg/ml and 2.22±0.90 µg/ml of plasma concentration of unchanged CDDP, respectively. For this reason, we selected CDDP concentration at 0.2, 1. 5µg/ml for MTT assay. Cell proliferation was signi cantly decreased by cisplatin treatment in a concentration-dependent manner, and miR-21 inhibitor enhanced the effects of cisplatin treatment in a similar manner to that with radiation (Fig. 2D, 2E). Similar results were also observed in SAS cells (Fig. 2G). Supporting these ndings, miR-21 mimic transfection induced obvious chemoresistance in HNSCC (Fig. 3C). Our results thus clearly demonstrated that miR-21 is also involved in chemoresistance for HNSCC cells.

Apoptotic effects of miR-21 inhibitor in HNSCC cells
We provided direct evidence that miR-21 plays an important role in HNSCC chemoradioresistance.
However, the mechanisms underlying the distinct roles of miR-21 inhibitor in regulating chemoradioresistance of HNSCC remain elusive. We further analyzed the effects of miR-21 inhibitor on apoptosis from treatment with either cisplatin or radiation by ow cytometry after annexin V and 7amino-actinomycin D staining. The results demonstrated that FaDu cells transfected with miR-21 inhibitor showed synergistic enhancement of apoptosis compared with either cisplatin alone or radiation alone (Fig. 4A-4D). We also used ow cytometry to evaluate the effects of miR-21 mimic, revealing signi cant suppression of the apoptosis induced by either radiation or cisplatin treatment (Fig. 4E-4H).
PDCD4 as a possible target of miR-21 PDCD4 has been reported as a target molecule related to drug sensitivity in various cancers [30.31], so we investigated protein levels of PDCD4 following transfection of miR-21 inhibitor or mimic in FaDu cells by Western blot analysis. As expected, we could con rm that miR-21 inhibitor upregulated PDCD expression, while miR-21 mimic decreased PDCD4 expression ( Fig. 5A-5D). This result demonstrated that PDCD4 was de nitely regulated by miR-21 expression.

Discussion
To improve the prognosis of HNSCC, nding useful biomarkers to estimate chemoradiosensitivity in the individual HNSCC patient is crucial. A major nding in this study was the direct evidence that expression levels of plasma miR-21, not from tissue samples in HNSCC patients with recurrence were signi cantly higher than those in patients without recurrence. As expected, the recurrent HNSCC group showed poorer OS (Fig. 1C) In addition, further interesting evidence was provided for the involvement of miR-21 in suppressing chemoradiosensitivity in HNSCC in this study. Even though miR-21 is well known as an oncomiR for various types of carcinoma [10,11,12], whether elevated miR-21 levels are associated with chemoradioresistance and the in uence of miR-21 on radioresistance in HNSCC have remained unclear.
The present study provides rst-hand evidence that miR-21 inhibitor enhanced the anti-proliferative effects of radiation in HNSCCs, including FaDu cells and SAS cells, which means hypopharyngeal squamous cell carcinoma and tongue squamous cell carcinoma cell lines, except nasopharyngeal carcinoma cell line. [39]. In addition, miR-21 inhibitor was also found to enhance the apoptotic effects of radiation, but had little effect by miR-21 inhibitor alone. These data con rmed that miR-21 caused radioresistance by increasing cell proliferation and decreasing apoptosis, identifying a potential therapeutic target for reducing radioresistance. Additional experimental signi cance of our ndings was that miR-21 inhibitor also enhanced the anti-proliferative effects of cisplatin in HNSCCs, including FaDu and SAS cell lines. A few reports have described miR-21 as modulating chemosensitivity to cisplatin for oral SCC [17,40], but no report about oropharyngeal or hypopharyngeal cell carcinoma. Because chemoradiotherapy is the standard therapy for advanced oropharyngeal or hypopharyngeal cell carcinomas, but not oral SCC, the present study offers useful information for clinical strategies. Further, miR-21 inhibitor also enhanced the apoptotic effects of radiation, but had little effect when administered alone. Collectively, these data demonstrated that miR-21 was deeply involved in the chemoradioresistance of HNSCC, suggesting miR-21 as a potential target to maximize the effects of chemoradiotherapy on HNSCC.
Finally, our data in this study implicated the involvement of PDCD4 as a target of miR-21 ( Fig. 5A-5D).
Knockdown of miR-21 clearly elevated PDCD4 expression on Western blotting. Logically, decreases in PDCD4 by miR-21 would suppress tumor cell apoptosis, contributing to chemoradioresistance [41,42,43]. There has been shown that PDCD4 exerts its activity by interacting with eIF4A and eLF4G to suppress mRNA translation and further imhibit the growth and proliferation of tumors [30]. It is one of the possible mechanism to inhibit cell proliferation by miR-21 inhibitor in this study. Unexpectedly, miR-21 did not in uence PTEN in FaDu cells in this study [44,45] (data not shown). In addition to PDCD4, a number of targets of miR-21 have been already reported including phosphatase and tensin homologue deleted on chromosome 10(PTEN), which is a tumor suppressor gene encoding a phosphatase that regulates cell cycle Akt and p53 activity, and reversion-induced cystine-rich protein (RECK) and tissue inhibitor of metalloproteinases 3 (TIMP3), which are suppressors of malignancy and inhibitors of matrix metalloproteinases. [22] Our studies are insu cient to provide a full understanding of the mechanisms by which miR-21 induces chemoradioresistance in HNSCC. This question needs to be addressed in future investigations.

Conclusion
In conclusion, the present study appears to emphasize the importance of circulating miR-21 as a predictive biomarker in chemoradiotherapy for HNSCC. Our data reveal the importance of miR-21 to chemoradioresistance in HNSCC, and provide new insights into novel roles of miR-21 for diagnosing and treating HNSCC. This study was approved by the ethics committee at Mie University Graduate School of Medicine (approval nos. 2445 and H2020-232). Written informed consent was obtained from each patient before enrolment in this study.

Consent for publication
All data described in this manuscript are original. None of the material has been published or is being considered for publication elsewhere.

Con ict of interest
The authors have no con icts of interest to declare.
Availability of data and material The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.      Supplementary Files