Correlation between polymorphisms in microRNA-regulated genes and cervical cancer susceptibility in a Xinjiang Uygur population

We explored the correlation between single nucleotide polymorphisms (SNPs) and susceptibility to cervical cancer (CC) in a Xinjiang Uygur population. Ten SNPs in eight miRNA-regulated genes were selected for analysis. Odds ratios (ORs) and 95% confidence intervals (95% CIs) were calculated using unconditional logistic regression analysis. Multivariate logistic regression analysis was used to detect correlations between SNPs and CC. We found that minor allele “C” of rs512715 in NEAT1 was associated with an increased risk of CC in the allele, codominant, dominant, overdominant and log-additive models. Minor allele “C” of rs4777498 in CELF6 was associated with an increased risk of CC in the recessive model. Minor allele “C” of rs3094 in RNASE4 was associated with increased risk of CC in the allele, dominant and log-additive models. In clinical stage III/IV CC patients, minor allele “C” of rs3094 in RNASE4 and minor allele “C” of rs8004334 in JDP2 were associated with increased risk. In subtype squamous carcinoma CC patients, minor allele “C” of rs512715 in NEAT1 and minor allele “C” of rs3094 in RNASE4 were associated with increased risk. In subtype adenocarcinoma CC patients, minor allele “C” of rs3094 in RNASE was associated with increased risk.


INTRODUCTION
Cervical cancer (CC) is the fourth most common malignancy in women, with 528,000 occurrences and 266,000 deaths in 2012 [1]. More than 85% of CC occurs in developing regions such as Eastern Africa, Melanesia, and Southern and Central Africa, where cervical CC accounts for more than 60% of gynecological cancers [2]. Cervical cancer is mainly attributable to human papillomaviruses (HPVs or PVs), which belong to the large Papillomaviridae family [1]. HPV particles include an approximately 8000-bp, double-stranded, closed circular DNA harboring eight genes [3].
Although previous studies have reported several biomarkers of CC, including p16 INK4a and Ki-67, few have investigated the relationship between CC risk and microRNA (miRNA)-regulated genes, including CDK6, PTEN and NEAT1, among others. MiRNAs are small (~19-25 nucleotides) non-coding RNA sequences that regulate gene expression through translational  [4]. In this way, miRNAs play key roles in a variety of biological process, including cell apoptosis, proliferation, differentiation, development and tumorigenesis, which are involved in the pathogenesis of a variety of ailments, including cancer, nephropathy and renovascular disease [5][6][7]. Consequently, how miRNA-regulated genes affect CC risk would seem to be a potentially meaningful investigation. We therefore investigated the relationships between single nucleotide polymorphisms (SNPs) in miRNA-regulated genes and the risk of CC. In the present case-control study, we selected 10 SNPs in eight miRNA-regulated genes and performed a comprehensive association analysis in a Xinjiang Uygur population.

DISCUSSION
In the present study, we found that four SNPs belonging to four miRNA-regulated genes were associated with CC risk. These were rs512715 in NEAT1 regulated by hsa-mir-342-3p, rs4777498 in CELF6 regulated by hsamir-375, and rs3094 in RNASE4 and rs8004334 in JDP2, both regulated by hsa-mir-590-5p.
In humans, miRNAs are transcribed by RNA polymerase II in the nucleus as pri-miRNAs, which may contain two or more mature miRNAs. Subsequently,   pri-miRNAs are processed by RNase III to form pre-miRNAs exported to the cytosol, carried by exportin 5, after which the pre-miRNAs are processed by Dicer in the cytosol to mature miRNAs. One strand of the mature miRNA is then incorporated with RNAinduced silencing complex (RISC), directing it to target mRNA [8]. The minor allele "C" of rs512715 increased CC risk in the allele, codominant, dominant, overdominant and log-additive models. Rs512715 belongs to NEAT1, which is regulated by hsa-mir-342-3p. We know of no other study relating NEAT1 to CC risk, though a Chinese study found a relationship between NEAT1 and bladder cancer [9,10]. In addition, an American study found hsa-mir-342-3p to be related to irritable bowel syndrome [11]. In a German study, significant upregulation of hsa-miR-342-3p was detected in the brains of macaques infected with bovine spongiform encephalopathy, and in a pilot study they also showed that hsa-miR-342-3p was upregulated in brain samples from humans with type 1 or type 2 sporadic Creutzfeldt-Jakob disease [12]. We have so far detected no direct evidence of a specific relationship between hsa-miR-342-3p and CC, and we suggest that this miRNA likely plays a general role in the regulation of multiple target genes in disease. However, the detailed mechanism by which hsa-miR-342-3p exerts gene effects in CC deserves further investigation.
The minor allele "C" of rs4777498 increased CC risk in the recessive model. Rs4777498 belongs to CELF6, which is regulated by hsa-mir-375. An American study found that CELF6 is highly expressed in diencephalic nuclei and neuromodulatory cell populations of the mouse brain [13]. Previous studies also reported hsamir-375 to be related to pancreatic cancer and early stage breast cancer [14,15]. In breast cancer, higher levels of hsa-mir-375 were expressed in ER-α-positive than ERα-negative or normal cells, which led to the suggestion that hsa-miR-375 up-regulation is a key driver of cell proliferation and an early event in tumorigenesis in ERα-positive tissues [16]. However, a detailed understanding of the mechanism by which hsa-mir-375 affects CC risks will require further investigation.
The minor allele "C" of rs3094 increased CC risk in the allele, dominant and log-additive models. In clinical stage III/IV patients, the minor allele "C" of rs3094 and minor allele "C" of rs8004334 were associated with increased CC risk. Rs3094 belongs to RNASE4 while rs8004334 belong to JDP2, and both are regulated by hsa-mir-590-5p. Previous studies showed RNASE4 to be associated with high-altitude adaptation, metabolic syndrome and neuron degeneration [17][18][19], while JDP2 was associated with heart failure [20]. Hsa-mir-590-5p is reportedly related to cardiac differentiation through down-regulation of TGFB signaling [21]. TGFB1-induced activation of Smad 2, -3, -4 leads to direct inhibition of STAT5 transactivation and STAT5-mediated transcription of downstream target genes, including miR-590 [22]. TGFB1 inhibits STAT5 expression at the protein level with no effect on mRNA expression. Whether there is a relationship between the mechanism of hsa-mir-590-5pmediated effects on CC risk and TGFB signaling warrants further investigation. There are two intrinsic limitations to this study. 1) The sample size was not large enough to obtain illative combinatory associations between SNPs and CC. 2) Selection bias may be unavoidable since this was a hospital-based study. Therefore, larger well-designed studies combined with CC classification are needed to confirm the observed associations and clarify the potential biological mechanisms of these SNPs in CC.

Study participants
A total of 532 subjects, including 247 patients with cervical cancer and 285 healthy women were recruited at the People's Hospital of Xinjiang Uyghur Autonomous Region between January 2014 and Jun 2016. The included patients were recently diagnosed with primary CC based on cervical biopsy with histopathological confirmation. We excluded patients with other cancers who underwent radiotherapy or chemotherapy. Controls were healthy, unrelated individuals selected randomly from the medical examination center of the hospital. All participants were women at least 18 years old in good mental condition who had at least three generations of paternal ancestry in their ethnicity (Xinjiang Uygur population). Tumors were staged according to International Federation of Gynecology and Obstetrics (FIGO) classification. Informed consent was obtained from all participants, and the study protocols were approved by the institutional review board of the People's Hospital of Xinjiang Uyghur Autonomous Region.

SNP selection and genotyping
Candidate SNPs were selected from among previously published polymorphisms associated with CC. Validated SNPs were selected with a MAF > 5% in the HapMap Asian population [23]. Venous blood samples (5 ml) were collected from each patient during laboratory examination. Genomic DNA was extracted from whole blood samples using a Gold Mag-Mini Whole Blood Genomic DNA Purification Kit (version 3.0; TaKaRa, Japan) [24] and stored at -80°C after centrifugation. DNA concentrations were evaluated using spectrometry (DU530 UV/VIS spectrophotometer, Beckman Instruments, Fullerton, CA, USA). We used Sequenom MassARRAY Assay Design 3.0 Software to design the Multiplexed SNP MassEXTEND assays [25]. SNP genotyping was done with a Sequenom MassARRAY RS1000 using the standard protocol recommended by the manufacturer [25]. The primer sequences used for genotyping are listed in Table 4. Data management and analyses were performed using Sequenom Typer 4.0 software as previously described [25,26].

Statistical analysis
Statistical analyses were performed using Microsoft Excel (Redmond, WA, USA) and the SPSS 17.0 statistical package (SPSS, Chicago, IL, USA). All p values in this study were two-sided, and p ≤ 0.05 after Bonferroni correction was considered the statistical significance threshold [27]. An exact test was used to assess the departure of each SNP frequency from Hardy-Weinberg equilibrium (HWE) in the controls. We compared allele frequencies between cases and controls using the χ 2 test. To assess the association of single SNPs with the risk of CC, five genetic models (codominant, dominant, recessive, over-dominant and log-additive) were applied using PLINK software (http://www.cog-genomics.org/plink2/). Odds ratios (ORs), 95% confidence intervals (95% CIs), and p values were calculated using unconditional logistic regression analysis [28][29][30].