Down Regulation of miR-34a and miR-143 May Indirectly Inhibit p53 in Oral Squamous Cell Carcinoma: a Pilot Study

Background: Aberrant microRNA expression has been associated with the pathogenesis of a variety of human malignancies including oral squamous cell carcinoma (SCC). In this study, we examined primary oral SCCs for the expression of 6 candidate miRNAs, of which five (miR-34a, miR-143, miR-373, miR-380-5p, and miR-504) regulate the tumor suppressor TP53 and one (miR-99a) is involved in AKT/mTOR signaling. Materials and Methods: Tumor tissues (punch biopsies) were collected from 52 oral cancer patients and as a control, 8 independent adjacent normal tissue samples were also obtained. After RNA isolation, we assessed the mature miRNA levels of the 6 selected candidates against RNU44 and RNU48 as endogenous controls, using specific TaqMan miRNA assays. Results: miR-34a, miR-99a, miR-143 and miR-380-5p were significantly down-regulated in tumors compared to controls. Moreover, high levels of miR-34a were associated with alcohol consumption while those of miR-99a and miR-143 were associated with advanced tumor size. No significant difference was observed in the levels of miR-504 between the tumors and controls whereas miR-373 was below the detection level in all but two tumor samples. Conclusions: Low levels of miR-380-5p and miR-504 that directly target the 3’UTR of TP53 suggest that p53 may not be repressed by these two miRNAs in OSCC. On the other hand, low levels of miR-34a or miR-143 may relieve MDM4 and SIRT1 or MDM2 respectively, which will sequester p53 indicating an indirect mode of p53 suppression in oral tumors.


Introduction
Oral squamous cell carcinoma (OSCC) is one of the major public health problems and is the sixth common human malignancy worldwide (Parkin et al., 2005). However, in India, it is the leading cancer in males and ranks within the top 5 malignancies in females (Dikshit et al., 2012;Ferlay et al., 2015). Characterization of the somatic mutational landscape of OSCC either alone or together with other head and neck neoplasms has revealed that TP53 is the most commonly mutated gene (Agrawal et al., 2011;Stransky et al., 2011;India Project Team of the International Cancer Genome, 2013;Pickering et al., 2013). The p53 protein, encoded by the TP53 gene, has a pivotal role in tumor suppression and is a mainstay of intrinsic anticancer defence systems. Its importance is reflected in the popular name "guardian of the genome" (Lane, 1992). Besides genetic mutation, the function and expression of p53 is controlled at several levels including transcription, post-transcriptional modification etc.
MicroRNAs (miRNAs) are 17-to 23-nucleotides, short, non-coding RNA molecules that negatively regulate gene expression at post-transcriptional level (Bartel, 2009). Mammalian miRNAs are versatile as a complementarity between the nucleotides spanning the position 2 to 7 of 5' end (known as miRNA 'seed') and the 3' untranslated region (3'UTR) of target mRNAs is enough for effective gene silencing. Thus, a single miRNA can target hundreds of mRNA and the 3'UTR of a single mRNA can be targeted by several miRNAs (Lewis et al., 2005). Based on this principle, computational approximations suggest that nearly half of all protein coding genes are regulated by miRNAs (Friedman et al., 2009), and therefore miRNAs participate in all vital cellular functions (Bushati and Cohen, 2007). As a consequence, deregulation in functioning or expression of miRNAs has been well-documented to result in a variety of human diseases, especially cancer, wherein miRNA themselves are known to function as oncogenes and tumor suppressors (Nelson and Weiss, 2008;Shiiba et al., 2010;Nohata et al., 2013). Several previous publications has substantiated the involvement of miRNAs in the pathogenesis of OSCC, and a compendium of miRNAs were shown to be differentially expressed (Chang et al., 2008;Chen et al., 2013). However, there has been limited research in OSCC on miRNAs that regulate or are regulated by TP53. Moreover, TP53 mutations were reported to be less frequent in Indian OSCC cases, when compared to the rest of the world (Kannan et al., 1999;Saranath et al., 1999;Zanaruddin et al., 2013), opening up the possibility of exploring the other dimensions of p53 inactivation.

Materials and Methods
The present study was approved by the Ethical Review Board of Government Royapettah Hospital, Chennai (Letter No.371/RMO/2010dt.29.1.2010 and was conducted within the ethical framework of Dr. ALM PG Institute of Basic Medical Sciences, Chennai. Informed consent forms were signed by all participants.

Subjects
Fifty two OSCC tissue samples and 8 independent normal tissues were collected from patients of Government Royapettah Hospital, Chennai. The patients' background and clinico-pathological characteristics were documented in a standard questionnaire. The tumor specimens were collected under local anesthesia with a 3-mm punch biopsy, immersed in RNAlater solution (Ambion, Austin, TX, USA) and transported to laboratory in cold-storage container. In case of control tissues, biopsies were punched 2 cm away from the tumor margins. It should be noted that either the tumor tissue or the control tissue were excised from a patient but not both, according to the patient's will and consent. The tissues were minced into smaller pieces and RNAlater was allowed to percolate for a day at 4°C. On the subsequent day, RNAlater was drained off and tissues were stored at -80°C until RNA extraction.

RNA isolation
Homogenization of tissue samples were carried out in a MicroSmash MS-100 automated homogenizer (Tomy, Japan) with Zirconium beads, followed by RNA isolation with miRNeasy kit (Qiagen, Germany) according to the manufacturer's protocol and instructions. The quantity and quality of the extracted RNA was checked by spectrophotometry (Nanodrop 2000, Thermo Fisher Scientific Inc., USA) and samples of optical density higher than 1.9 at 260/280 nm were used for further reverse transcription.

Two-step reverse transcription quantitative PCR
Reverse transcription quantitative PCR (RT-qPCR) was performed using TaqMan MicroRNA Reverse transcription kit and TaqMan miRNA assays (Table  1), following the manufacturer's protocol. All reagents were obtained from Applied Biosystems (USA) and the experiments were carried out in a 7500HT Real Time PCR System. cDNA was synthesized in a 15 μL total volume containing 100 mM dNTPs, Multiscribe Reverse Transcriptase (50 U/μL), RT buffer, RNase inhibitor, RNA sample (10 ng/15 μL) and specific 5X TaqMan microRNA RT primers. After reverse transcription, the products were diluted 15 times, added to 384 well optical plates and RT-qPCR reactions (in triplicates) were performed in a 10 μL total volume containing the required proportion of TaqMan 2X Universal Master Mix (No AmpErase UNG) and specific 20X TaqMan MicroRNA assay. A negative control without cDNA was also included in parallel. RNU44 and RNU48 were used as endogenous references. Expression fold changes were computed using the 2− ΔΔCt calculation (Livak and Schmittgen, 2001), where ∆C t = C t (test miRNA) -C t (average of RNU44 and RNU48) and ∆∆C t = ∆C t (individual sample) -∆C t (control median)

Statistical analyses
Statistical analyses of the data were carried out in GraphPad Prism 6 (GraphPad software Inc., La Jolla, CA, USA). The fold change ratios of each miRNA in the samples were log2 transformed and tested for normality by the Kolmogorov-Smirnov test. In case of Gaussian distribution (miR-99a, miR-143 and miR-504), the difference between the two groups were analysed using Student's t-test. Whenever, a significant difference in variance was observed between the groups, a Welch correction was applied post t-test (miR-504, miR-99a). In case of non-Gaussian distribution (miR-34a and miR-380-5p), Mann Whitney test for independent samples was applied. The expression of each miRNA in tumors was further tested for association with the clinico-pathological characteristics by means of Student's t-test or Mann Whitney test in case of two subgroups, and by One-way ANOVA or Kruskal-Wallis test in case of more than two subgroups. All tests were two tailed and a P < 0.05 was considered as significant. The relative expression levels are provided as mean or median as and when applicable.

Patient characteristics
The clinico-pathological characteristics of the 52 oral cancer patients included in this study are summarized in Table 2. The age of the patients ranged from 32 to 86 yrs. with a mean of 52.7 yrs. There were a higher proportion of males (67.3%) than females (32.7%). Gingivo-buccal complex was the most common site of tumor occurrence (84.6%), which is related to the chewing of various tobacco products. Most of the patients presented with advanced TNM stage (90.4%) and 86.5% were positive for regional lymph node involvement. Predominantly, tumors were of moderately differentiated histology (53.85%).
Approximately 85% of the patients had practiced at least one of the behavioral risk habits associated with OSCC.

The relative expression of microRNAs in oral tumors
We examined the expression of the 6 candidate miRNAs in OSCC samples (n=52) compared to independent adjacent normal tissues (n=8) using TaqMan probes (Applied Biosystems, USA). Two smallnucleolar RNAs namely RNU44 and RNU48 were used as endogenous references for normalization. Statistical analyses of the data obtained from RT-qPCR showed that the expression of four miRNAs were significantly low in OSCC sample group compared to controls (Figure 2). These are hsa-miR-34a (0.39 in tumor Vs 1.01 in control; P=0.0007), hsa-miR-99a (0.27 in tumor Vs 1.16 in control; P=0.0219), hsa-miR-143 (0.46 in tumor Vs 1.07 in control; P=0.0146) and hsa-miR-380-5p (0.04 in tumor Vs 1.03 in control; P=0.0071). Hsa-miR-504, despite its down regulation in tumors, was not statistically significant (0.46 in tumor Vs 1.86 in control; P=0.0896). The Ct value for hsa-miR-373 was undetermined in all controls and tumors except for two OSCC specimens. This indicates that miR-373 expression is either below the detection limit or is not Figure 1. Regulation of p53 Activity by the MicroRNAs Selected for Evaluation in the present study. The microRNAs miR-380-5p and miR-504 directly target the 3' untranslated region (3'UTR) of TP53 mRNA and suppresses its translation, while miR-34a, miR-143 and miR-373 indirectly regulate p53 activity by targeting SIRT1and MDM4, MDM2, and LATS2, respectively. Additionally, miR-143 and miR-34a are under the transcriptional control of p53. The candidate miR-99a activates AKT/mTOR signaling pathway and p53 is known to repress this pathway in response to various stress signals

Association with clinico-pathological characteristics
The relative fold change of miRNAs in tumors is merely an overall estimate and it should be noted that the tumor samples differ with respect to their pathological stage, grade etc. Therefore, we sub-classified the tumors based on their clinical parameters and tested their association with miRNA expression levels by univariate analysis (Table 3). Interestingly, miR-34a was associated with histological grade and alcohol consumption status. High levels of miR-99a and miR-143 were associated with advanced tumor size (T3-T4). Since miR-373 was expressed in only two of the OSCC samples, we could not test its association with the tumor characteristics. Of note, these two samples were female buccal mucosal tissue samples of advanced tumor stage (IVa) with involvement of a single/multiple ipsilateral lymph node(s) i.e., N2a or N2b.

Discussion
MicroRNAs play a pivotal role in an array of physiological functions and consequently miRNA deregulation has been linked to various diseases. Recently, this class of non-coding RNAs have been extensively investigated in various cancers including OSCC. However, on account of the limited research in OSCC on miRNAs  DOI:http://dx.doi.org/10.7314/APJCP.2015.16.17.7619 Down Regulation of miR-34a and miR-143 May Indirectly Inhibit p53 in Oral Squamous Cell Carcinoma: a Pilot Study that regulate or is regulated by p53, we undertook this study. Several miRNAs have been identified to regulate p53 either directly by binding its 3' UTR or indirectly by targeting p53 modifying enzymes (Hermeking, 2012). We chose miR-380-5p and miR-504 among the direct p53 regulators, miR-34a, miR-143 and miR-373 among the indirect p53 regulators, and miR-99a that is independent of the p53 pathway.
The miRNAs miR-380-5p and miR-504 are suggested oncomiRs since each miRNA binds to two different sites in the 3'UTR of TP53 and inhibits its translation (Hu et al., 2010;Swarbrick et al., 2010). Indicative of its oncogenic nature, high expression of miR-380-5p correlated with poor outcome in neuroblastomas showing MYCN amplification and further cooperated with activated RAS to transform primary cells, block oncogene-induced senescence and form tumors in mice (Swarbrick et al., 2010). Similarly, overexpression of miR-504 caused a decrease in p53 protein levels leading to tumor progression (Hu et al., 2010). A previous study in OSCC demonstrated that over expression of miR-504 resulted in down regulation of FOXP1 and contributed to

Figure 2. The Differential Expression of Candidate miRNAs in Oral Cancer and Control Samples. Box
and whisker plots representing the expression of (a) miR-34a, (b) miR-99a, (c) miR-143, (d) miR-380-5p and (e) miR-504in OSCC specimens (n=52) and independent adjacent normal tissues (n=8). Error bars of (a) and (d) represents median and interquartile range while that of (b), (c) and (e) represent mean and 95% confidence limits. The solid black circles and squares represent the miRNA expression values in individual samples of control and tumor, respectively. *, ** and *** represent statistical significance at P<0.05, P<0.01 and P<0.001, respectively. 'ns' represents P = not significant. In (f), the Ct value of miR-373 in OSCC and control specimens as detected by RT-qPCR is shown invasion (Yang et al., 2012). However, the expression of these two miRNAs tend to vary based on the tissue and its microenvironment where the tumor originates. As an example, down regulation of miR-504 was reported in hypopharyngeal squamous cell carcinoma (Kikkawa et al., 2014), while down regulation of miR-380-5p was observed in majority of the OSCC samples compared to the matched leukoplakia or dysplasia samples (Cervigne et al., 2009). In the present study, miR-380-5p was observed to be significantly down regulated in OSCC compared to controls, whereas there was no statistically significant difference in the expression of miR-504. This suggests that miR-380-5p and miR-504 may not have a role in suppressing p53 in the context of oral carcinoma.
The candidates miR-34a and miR-143 indirectly regulate p53 and are further controlled by the transcriptional activity of p53 thus forming a feedback loop (He et al., 2007;Yamakuchi et al., 2008;Zhang et al., 2013a). In response to DNA damage and oncogenic stress, p53 induces the expression of the pro-apoptotic miR-34 family (Chang et al., 2007;Raver-Shapira et al., 2007;Hermeking, 2010). Therefore, miR-34 family members are expressed at very low levels in several cancers (Bommer et al., 2007;Chang et al., 2007;Tazawa et al., 2007) including OSCC (Scapoli et al., 2010). Specifically, miR-34a targets not only SIRT1, a nicotinamide adenine dinucleotide (NAD+)-dependent deacetylase which represses p53 activity by deacetylation of the protein (Yamakuchi et al., 2008), but also MDM4, another negative regulator of p53 (Mandke et al., 2012). On the other hand, miR-143 targets MDM2, which negatively controls the stability and transcriptional activity of p53. Similar to miR-34a, miR-143 which resides at chromosome 5q, is often down regulated or deleted in a number of malignancies (Lui et al., 2007;Volinia et al., 2010;Zhang et al., 2013b;Ng et al., 2014). Moreover, miR-143 was found to be down regulated in OSCC animal model and in human head and neck squamous cell carcinoma (HNSCC) (Yu et al., 2009;Hui et al., 2010). Supporting the tumor suppressor activity of miR-34a and miR-143, we observed their down regulation in OSCC. Since both miRNAs target molecules which repress p53, their down regulation will invariably result in attenuation of p53 function, which in turn may result in reduced transcription of miR-34a and miR-143. Interestingly, we found that the expression levels of miR-34a was comparatively high in patients with a history of alcohol consumption. In this context, it should be noted that treatment of human hepatocytes and cholangiocytes with ethanol induced the expression of miR-34a via hypomethylation of miR-34a promoter (Meng et al., 2012). Our observation on the differential expression of miR-34a in different tumor grades (G1, G2 and G3) and the association of miR-143 with advanced tumor size clearly needs confirmation in a larger sample size.
The gene LATS2 encodes a serine/threonine protein kinase that interacts with MDM2 thus functioning as a positive regulator of p53. Targeting of LATS2 by miR-373 thus allows MDM2 to bind p53 and direct it to ubiquitin mediated degradation (Aylon et al., 2006;Lee et al., 2009). Also, miR-373 co-operates with oncogenic RAS to induce tumorigenic growth in testicular germ cell tumors  (Voorhoeve et al., 2006). Contraindicating its oncogenic role, miR-373 was observed to be down regulated in oral/head and neck cancer cell lines (Tran et al., 2007), as well as in tumor tissues (Childs et al., 2009). In the present study, miR-373 was expressed in only two of the tumor samples showing regional lymph node metastasis. Keklikoglou et al. ( 2012) has previously reported that miR-373 was below detection levels in primary breast tumors when analysed by RT-qPCR, thus strengthening our observation. Finally, the p53 independent miRNA, miR-99a, was demonstrated to be consistently down regulated in HNSCC along with its family members (Wong et al., 2008;Hui et al., 2010;Scapoli et al., 2010;Lajer et al., 2011;Chen et al., 2012;Yan et al., 2012;Yen et al., 2014). Forced expression of miR-99 family members in HNSCC cell lines reduced cell proliferation and migration, paralleled by an increase in apoptosis (Chen et al., 2012). Therefore, the relatively high levels of miR-99a in T3/T4 tumors compared to T1/T2 tumors is not consistent with its established role and necessitates further validation in a larger sample size. Functional studies suggest that the miR-99 family members regulate AKT/mTOR signaling pathway by targeting insulin-like growth factor 1 receptor (IGF1R), mechanistic target of rapamycin (mTOR) and regulatory associated protein of mTOR (Raptor) (Doghman et al., 2010;Chen et al., 2012). Therefore down regulation of miR-99a, along with miR-143 that directly targets AKT (Noguchi et al., 2013), may activate the AKT/mTOR signaling pathway in OSCC samples.
In conclusion, the current pilot study provides a bird's eye view on the miRNA regulation of p53 in OSCC. The two selected miRNAs that directly regulate p53 were expressed at low levels and hence the possibility of translation repression of TP53 by miR-380-5p and miR-504 in OSCC is ruled out. However, the indirect p53 regulators, miR-34a and miR-143 was significantly reduced in OSCC while miR-373 expression was detectable in only two of the tumor samples. This suggests that p53 is subjected to attenuation by the negative regulators MDM2, MDM4 and SIRT1 in a subset of oral tumors. The down regulation of miR-99a along with miR-143 additionally indicates the activation of AKT/mTOR pathway in OSCC. However, more extensive studies assessing the entire repertoire of miRNAs involved in p53 regulation are warranted to gain insight on the multiple mechanisms of p53 inactivation.