Effect of genetic variation in microRNA binding site in WNT1-inducible signaling pathway protein 1 gene on oral squamous cell carcinoma susceptibility

Background Oral squamous cell carcinoma (OSCC), which is the most common head and neck cancer, accounts for 1%–2% of all human malignancies and is characterized by poor prognosis and reduced survival rates. WNT1-inducible signaling pathway protein 1 (WISP1), a cysteine-rich protein belonging to the Cyr61, CTGF, Nov (CCN) family of matricellular proteins, has many developmental functions and may be involved in carcinogenesis. This study investigated WISP1 single-nucleotide polymorphisms (SNPs) to elucidate OSCC susceptibility and clinicopathologic characteristics. Methodology/Principal findings Real-time polymerase chain reaction was used to analyze 6 SNPs of WISP1 in 900 OSCC patients and 1200 cancer-free controls. The results showed that WISP1 rs2929970 polymorphism carriers with at least one G allele were susceptible to OSCC. Moreover, compared with smokers, non-smoker patients with higher frequencies of WISP1 rs2929970 (AG + GG) variants had a late stage (stages III and IV) and a large tumor size. In addition, OSCC patients who were betel quid chewers and carried WISP1 rs16893344 (CT + TT) variants had a low risk of lymph node metastasis. Conclusion Our results demonstrate that a joint effect of WISP1 rs2929970 with smoking as well as WISP1 rs16893344 with betel nut chewing causally contributes to the occurrence of OSCC. WISP1 polymorphism may serve as a marker or a therapeutic target in OSCC.


Methodology/Principal findings
Real-time polymerase chain reaction was used to analyze 6 SNPs of WISP1 in 900 OSCC patients and 1200 cancer-free controls. The results showed that WISP1 rs2929970 polymorphism carriers with at least one G allele were susceptible to OSCC. Moreover, compared with smokers, non-smoker patients with higher frequencies of WISP1 rs2929970 (AG + GG) variants had a late stage (stages III and IV) and a large tumor size. In addition, OSCC patients who were betel quid chewers and carried WISP1 rs16893344 (CT + TT) variants had a low risk of lymph node metastasis.

Conclusion
Our results demonstrate that a joint effect of WISP1 rs2929970 with smoking as well as WISP1 rs16893344 with betel nut chewing causally contributes to the occurrence of OSCC. WISP1 polymorphism may serve as a marker or a therapeutic target in OSCC. PLOS  Introduction approved this study (CSMUH No: CS13214-1), and informed written consent was obtained from each participant.

Selection of WISP1 polymorphisms
A total of six SNPs in WISP1 were selected from the International HapMap Project data for this study. The previous studies have reported the effect of WISP1 genetic polymorphisms on human cancer susceptibility [22,23,27]. Thus, we included the SNP rs62514004, rs2929970 and rs2929973 in the promoter region and 3' untranslated region, respectively. Three SNPs (rs16893344, rs2977530 and rs2977537) which locate in introns of WISP1 were selected in this study since these SNPs were suggested to be significantly associated with lung cancer susceptibility or chemotherapy response in Chinese population [22].

DNA extraction and WISP1 genotyping
Genomic DNA from OSCC group was isolated from peripheral blood using the QIAamp DNA blood mini kit and used as the template for polymerase chain reaction as described previously [28,29]. Allelic discrimination of the rs62514004, rs16893344, rs2977530, rs2977537, rs2929970 and rs2929973 of WISP1 gene was assessed with the ABI StepOne Real-Time PCR System (Applied Biosystems, Foster City, CA, USA), and analyzed with SDS vers. 3.0 software using the TaqMan assay.

Bioinformatics analysis
The stem-loop portion of the miRNA-miRNA duplex structure of pre-miRNAs was identified by miRNA target prediction using the MicroRNA.org resource. Models of the miRNA-target duplex were determined using the RNAhybrid web tool on Bielefeld Bioinformatics Server. Boxplot chart showing the differential expressions of miR-99a-5p in 420 OSCC patients and 43 normal controls, as taken from the Pan-Cancer dataset [30].

Statistical analysis
The Mann-Whitney U test was used to compare differences in the distribution of age and demographic characteristics between the controls and OSCC patients. The common haplotypes were estimated by PHASE version 2.1. ORs with 95% CIs were estimated using logistic regression models. AORs with 95% CIs were used to assess association between genotype frequencies with OSCC risk and clinical factors. p values less than 0.05 were considered significant. The data were analysed with SPSS 12.0 statistical software (SPSS Inc., Chicago, IL, USA).
The interactive effects between environmental risk factors and the WISP1 genetic polymorphisms are shown in Table 3. To reduce the possible interference of confounding variables, we used adjusted ORs (AORs) with 95% CIs that were estimated using multiple logistic regression models after controlling for betel nut chewing and alcohol consumption in each comparison. Among 664 non-smokers, those with WISP1 rs2977530 AG + GG genotypes exhibited a 0.598 Mann-Whitney U test was used between healthy controls and patients with oral cancer. * p value < 0.05 as statistically significant.
To clarify the role of the WISP1 genetic polymorphisms in OSCC clinicopathologic statuses, such as clinical stage, tumor size, LN metastasis, distant metastasis, and cell differentiation, the distribution frequency of clinical statuses and WISP1 genotype frequencies in OSCC patients were estimated. The rs62514004, rs2977530, rs2977537 and rs2929973 genetic polymorphisms showed no significant association with the clinicopathologic statuses. However, among 707 OSCC patients who were betel quid chewers, those carrying the polymorphic rs16893344 gene The odds ratio (OR) with their 95% confidence intervals were estimated by logistic regression models. The adjusted odds ratio (AOR) with their 95% confidence intervals were estimated by multiple logistic regression models after controlling for betel nut chewing, alcohol and tobacco consumption.  Fig 1A, the 3 0 -untranslated region (3 0 -UTR) of the WISP1 gene is 5.0 kb long and might be among the region's most sensitive to microRNA (miRNA) epigenetic regulation The odds ratio (OR) with their 95% confidence intervals were estimated by logistic regression models. The adjusted odds ratio (AOR) with their 95% confidence intervals were estimated by multiple logistic regression models after controlling for betel nut chewing and alcohol. https://doi.org/10.1371/journal.pone.0176246.t003 ( Fig 1A). The miRNA hsa-miR-99a (miRBase ID: MI0003190, Fig 1B) shares binding site complementarily with rs2929970 in the 3 0 -UTR region (Fig 1C and 1D). In addition, compared with the [A]-allele, the OSCC-associated risk [G]-allele creates a slight kink in the WISP1 mRNA structure, which results in a less negative free energy state and less stable hybridization [MFE (minimum free energy) changes: 22.16%, from −22.6 to −18.5 kcal/mol] (Fig 1E). Furthermore, starBase analysis revealed a significant difference in hsa-miR-99a expression in 420 OSCC patients and 43 subjects from the Pan-Cancer data set [30] (Fig 1F) (p = 1.78 × 10 −15 ).

Discussion
In this study, we revealed the correlations between WISP1 SNPs and OSCC. Previous studies have indicated the correlations of risk factors such as alcohol consumption, betel quid chewing, and smoking with oral cancer carcinogenesis [31][32][33]. In our study, we confirmed that betel quid chewing, cigarette smoking, and alcohol consumption are associated with OSCC (Table 1). However, the associations of these risk factors with WISP1 regulation have yet to be extensively investigated. Smoking is one of the crucial risk factors for OSCC [31,33]. In a pilot study, whole genome expression profiling of Indian patients with tobacco chewingassociated oral cancers implicated that WISP1 is one of the representative apoptosis-related deregulated genes in oral cancer [34]. Chen et al. revealed that WISP1 was overexpressed in non-small cell lung carcinoma (NSCLC) samples compared with their normal lung tissue counterparts, implicating that WISP1 might act as an oncoprotein in NSCLC [10]. However, WISP1 expression was not associated with clinical parameters such as family history, metastasis, smoking history, tuberculosis, gender, tumor type, and tumor size in NSCLC individuals [10]. WISP1 mRNA expression levels were higher in both lung and head and neck tumor tissues compared with their normal tissue counterparts [35], and smoking history was inconsistent. Therefore, it can be assumed that smoking has a limited ability to induce WISP1 overexpression.
Recent studies have suggested the vital role of WISP1 in cancer [35][36][37]. However, the WISP1 SNPs contributing to cancer progression have yet to be extensively investigated. In the present study, we found that the WISP1 SNP of rs2929970 was associated with OSCC risk ( Table 2); rs2929970 is located in the 3 0 untranslated region within a region of splicing variation [38]. Our present study confirmed the association of the clinically examined rs2929970 in the WISP1 3 0 -UTR region with OSCC risk; this association is most likely attributed to a putative hsa-miRNA-99a binding site (Fig 1). This finding suggests that the SNP corresponding to the RNA bulge region may affect the binding strength of specific miRNA/target duplexes, resulting in low minimum free energy and modulating mRNA stability.
A previous study indicated that the rs2929970 SNP may be associated with spinal osteoarthritis in postmenopausal Japanese women [24]. Yamada et al. [26] indicated that WISP1 rs2929970 was associated with hypertension in men carrying the G allele, and the men carrying this polymorphism had higher blood pressure. These results demonstrate the different effects of WISP1 rs2929970 expression in different diseases. Such inconsistent results for WISP1 rs2929970 expression may be due to different ethnicities or diseases; therefore, different  genotype distributions may be observed. Although the detailed mechanism of WISP1 rs2929970 remains unclear, the WISP1 rs2929970 polymorphism certainly plays a role in cancers or diseases.
In the present study, we determined that nonsmoker controls and OSCC patients with WISP1 polymorphic rs2977530 AG + GG genotypes had a low risk of OSCC (AOR = 0.598, Table 3). Chen et al. indicated that lung cancer patients carrying the A alleles of WISP1 rs2977530 polymorphisms may have an increased risk of lung cancer [22]. However, in that study, the smoking status of controls was not adjusted [22]. Consistent with this result, our data reveal that the G alleles of WISP1 rs2977530 were associated with a low risk of OSCC. We also analyzed the WISP1 SNP of rs1689334 and showed that 707 OSCC betel quid chewers carrying WISP1 rs16893344 SNP CT + TT genotypes exhibited a low risk of LN metastasis (OR = 0.674, p = 0.038, Table 5). However, Chen et al. [22] indicated that lung cancer patients carrying the T allele of the WISP1 rs16893344 polymorphism may have an increased risk of lung cancer. Tao et al. also showed that WISP1 rs16893344 C > T polymorphisms significantly increased myocardial infarction risk [39]. These results demonstrate that the various WISP1 SNPs may be expressed in different cancers and diseases. Moreover, a previous study suggested that WISP1 expression is regulated by methylation, and WISP1 hypomethylation contributes to LN metastasis in OSCC [14]. The WISP1 expression is correlated with the DNA methylation of its promoter, and reduced methylation levels are correlated with increased WISP1 expression [14]. Therefore, although the interaction of betel quid chewing with WISP1 expression and the functions of WISP1 rs16893344 have not been extensively investigated, the WISP1 SNP of rs16893344 may contribute to changes that influence WISP1 gene transcription.
We analyzed the correlations of WISP1 SNP expression with the clinical statues of OSCC patients. We observed that among 100 nonsmoker OSCC patients, those carrying WISP1 rs2929970 AG + GG genotypes had later stage OSCC and a larger tumor size (Table 6). Because smoking is a well-known risk factor for OSCC [31,33], this result in nonsmoker OSCC patients implicated the pivotal role of the WISP1 SNP of rs2929970 in cancer progression and WISP1 regulation. A previous study showed that WISP1 binds to αvβ3 integrin and causes the activation of the ASK1, JNK/p38, and AP-1 pathways, which upregulate ICAM-1 expression and promote the migration of human OSCC cells [15], and tumor-secreted WISP1 promotes angiogenesis through VEGF-A expression and increased angiogenesis-related tumor growth [16]. Although the mechanism and regulation of WISP1 SNPs in cancer progression have not been extensively investigated, the WISP1 SNPs of rs16893344 and rs2929970 might be involved in WISP1 regulation through WISP1-induced ICAM-1 upregulation and VEGF-A expression. The WISP1 SNPs of rs16893344 and rs2929970 may interfere with or enhance the binding activity of WISP1 to αvβ3 integrin, triggering the regulation of the ASK1, JNK/p38, and AP-1 pathways and VEGF-A expression, leading to a more favorable or poorer prognosis. Moreover, Mercer et al. also reports that, in a rat model of alcohol-induced liver disease, chronic alcohol consumption can significant upregulation in WISP1 expression [40]. Approximately half of OSCC patients (53.3%) in our study consumed alcohol (Table 1), these OSCC patients may exhibit a higher level of serum WISP1, and alcohol consumption may exert a synergistic effect on WISP1 upregulation. However, the WISP1 SNPs contributing to cancer progression and WISP1 regulation require further investigation to elucidate their detailed mechanisms.
In conclusion, WISP1 SNPs are correlated with OSCC. The WISP1 SNP of rs2929970 is associated with OSCC susceptibility, and rs2929970 A/G polymorphisms may be correlated with a worse prognosis of OSCC, such as later stage OSCC or larger tumor size. WISP1 rs2929970 may serve as a marker or a therapeutic target in OSCC.
Supporting information S1 File. Table A. Odds ratio (OR) and 95% confidence interval (CI) of oral cancer associated with WISP1 genotypic frequencies in smoker. Table B. Clinical statuses and WISP1 rs16893344 genotype frequencies in oral cancer among 193 non-betel quid chewers.