Association of Polymorphisms in miRNA Processing Genes With Type 2 Diabetes Mellitus and Its Vascular Complications in a Southern Chinese Population

Objective: To evaluate the potential association between the genetic variants in miRNA processing genes (RAN, XPO5, DICER1, and TARBP2) and susceptibility to type 2 diabetes mellitus (T2DM) and its vascular complications, as well as to further investigate their interaction with environmental factors in type 2 diabetes. Methods: We conducted a case-control study in genotyping of five polymorphic loci, including RAN rs14035, XPO5 rs11077, DICER1 rs13078, DICER1 rs3742330, and TARBP2 rs784567, in miRNA processing genes to explore the risk factors for T2DM and diabetic vascular complications. Haplotype analyses, interactions of gene-gene and interactions of gene-environment were performed too. Results: We identified a 36% decreased risk of developing T2DM in individuals with the minor A allele in DICER1 rs13078 (OR: 0.64; 95%CI: 0.42–0.95; P: 0.026). The AA haplotype in DICER1 was also associated with a protective effect on T2DM compared with the AT haplotype (OR: 0.63; 95%CI: 0.42–0.94; P-value: 0.023). T2DM patients with the TT+TC genotype at RAN rs14035 had a 1.89-fold higher risk of developing macrovascular complications than patients with the CC genotype (OR: 1.89; 95%CI: 1.04–3.45; P-value: 0.037). We also identified two three-factor interaction models. One is a three-factor [DICER1 rs13078, body mass index (BMI), and triglyceride (TG)] interaction model for T2DM (OR: 5.93; 95%CI: 1.25–28.26; P = 0.025). Another three-factor [RAN rs14035, hypertension (HP), and duration of T2DM (DOD)] interaction model was found for macrovascular complications of T2DM (OR = 41.60, 95%CI = 11.75–147.35, P < 0.001). Conclusion: Our study provides new evidence that two single nucleotide polymorphisms (SNPs) of the miRNA processing genes, DICER1 and RAN, and their interactions with certain environmental factors might contribute to the risk of T2DM and its vascular complications in the southern Chinese population.


INTRODUCTION
Diabetes mellitus (DM) is a chronic disease that occurs when there are increased levels of blood glucose because the body is unable to produce any or enough of the hormone insulin or use insulin effectively (1). According to the International Diabetes Federation (IDF) report, there were 4.25 billion people with diabetes in 2017, and the number of diabetes patients will increase to 6.29 billion globally by 2040 (2). In 2017, more than 4 million people died because of diabetes and diabetic complications, which means that worldwide, a patient died every 8 s because of diabetes (2). The total global healthcare expenditure due to DM was estimated at 727 billion dollars in 2017 (3). As the most populous and largest developing country in the world, China has the highest number of diabetes patients worldwide (114.4 million in 2017) (2). The number of diabetes patients soared from 4.8 million in 1980 to 39.8 million in 2007 (4).
T2DM as the most common type of diabetes, accounts for approximately 90% of diabetes cases and is affected by both environmental factors and genetic factors (5)(6)(7). Type 2 diabetes mellitus is a long-term metabolic disorder characterized by damage to insulin secretion and sensitivity, resulting in hyperglycemia (8,9), which can lead to the development of diabetic vascular complications, such as diabetic nephropathy, peripheral artery disease (PAD) and coronary heart disease (CHD) (2,10,11). Risk factors for diabetic vascular complications include lifestyle (such as smoking and drinking), duration of diabetes, age of onset and genetic factors, etc. (12,13).
MicroRNAs (miRNAs) are a class of short, single-strand, non-coding, and endogenous RNA molecules of 21-23 nts in length (14). Although miRNAs constitute at most 3% of the human genome, it has been reported that approximately onethird of human genes were regulated by miRNAs (15,16). Recently, the evidences that series of miRNAs were related to the T2DM and diabetic vascular complication, have been proved by several studies (17), such as miR-200 family (18), miR-124a (19), miR-21-5P (20), and miR-125a-3P (20). RASrelated nuclear protein (RAN), exportin 5(XPO5), DICER1, and TARBP2, which are known as microRNA processing enzyme, are the key to complete the biosynthesis of mammalian miRNAs (21,22). First, RNA II polymerase transcribed miRNAs into long precursors called pri-miRNAs, which are cleaved in the nucleus to release a stem loop intermediate about 60-70 nt, known as the miRNA precursor hairpin (pre-miRNA). Secondly, XPO5 and RAN export the pre-miRNA from the nucleus to the cytoplasm. And the DICER1 and TARBP2 further cooperatively dice the pre-miRNA into a double-stranded, short miRNA duplex (23,24).
Last, the double-stranded miRNA is preferentially incorporated into RNA-induced silencing complex, targeting endogenous mRNA silencing (25,26). Therefore, if the expression and structure of these miRNA processing proteins have been altered in the biosynthesis of mammalian miRNAs, it would directly impact the biosynthesis of mature miRNAs and further change the function and structure of miRNAs (21,22) (Figure 1).
A variation in a single nucleotide is called a single nucleotide polymorphism (SNP) that occurs in polymorphisms of the genomic DNA sequence, and is the simplest form of DNA variation among individuals and might affect the expression and function of genes (27). It has been reported that SNPs are associated with the development of type 2 diabetes mellitus and its vascular complications (28)(29)(30). GWAS (genome-wide association studies) have identified over 80 SNPs were closely related to type 2 diabetes mellitus (31)(32)(33)(34). In addition, GWAS also identified over 20 SNPs related to the various chronic T2DM complications (e.g., diabetic nephropathy, diabetic retinopathy, diabetic cardiopathy and diabetic painful neuropathy) (35). RAN rs14035, DICER1 rs3742330, and DROSHA rs10719 are located in the 3 ′ -UTR, which may disturb the function of these miRNA processing proteins by impacting the binding site of their miRNAs, resulting in dysregulation in processing of related gene expression and impact the biosynthesis of related proteins (36)(37)(38)(39). We hypotheses that the SNPs in miRNA processing genes may impact the susceptibility to T2DM and its vascular complications by disrupting the structure, binding sites, or processing of miRNA (40,41). To our knowledge, few studies evaluated the potential association between the risk of T2DM and its vascular complications and specific SNPs in miRNA processing genes. Thus, the aim of our study is investigating the effects of variations in miRNA processing genes, RAN (rs14035), XPO5 (rs11077), DICER1 (rs3742330 and rs13078), and TARBP2 (rs784567), as well as their interactions with environmental factors on T2DM and its vascular complications in a Southern Han Chinese population.

Study Subjects
This case-control study included a total of 1,275 unrelated Han Chinese Southern residents in Guangzhou. A total of 743 T2DM subjects were inpatients of the Endocrinology Departments of the First Affiliated Hospital of Jinan University in Guangzhou from September 2011 to January 2018, and all patients were diagnosed using the 2003 American Diabetes Association criteria (42). Patients with impaired glucose tolerance and type 1 diabetes mellitus were excluded from the study. During the same period, a total of 532 healthy control individuals without a family history of diabetes and with normal fasting glucose were also randomly recruited to match T2DM patients by sex and age (±5 years). Before participating in the study, all of the included individuals agreed and signed informed consent, and the study protocol was approved by the Ethics Committee of Medical School at Jinan University. Within the 743 included T2DM patients, 46, 96, 108, and 227 cases were diagnosed with T2DM with other complications, micromacrovascular complications, macrovascular complications, and microvascular complications, respectively, while 266 cases were categorized into T2DM without any complication. According to the 8th edition IDF Diabetes Atlas (2) and Gregg et al. (43), macrovascular complications included coronary artery disease (CAD), resulting in myocardial infarction or stenocardia, and peripheral artery disease (PAD), leading to diabetic encephalopathy and cerebral infarction, whereas microvascular complications are retinopathy, neuropathy and nephropathy. Micro-macrovascular complications were defined as presenting with both macrovascular microvascular and complications.

Selection of SNPs
We determined the targeted SNPs of RAN/XPO5/DICER1/TARBP2 were through an electronic search of the HapMap database based on the genotype information of Han Chinese in Beijing, China. And, the tarSNPs were determined based on the following criteria: (1) the minor allele frequency (MAF) of SNPs >5%; (2) pairwise tagging with r 2 ≥ 0.8; and (3) P-value of Hardy-Weinberg > 0.001. Five SNPs were identified, RAN rs14035, XPO5 rs11077, DICER1 rs3742330 and rs13078, and TARBP2 rs784567. Finally, five target SNPs were selected for further analysis. The detailed information of the five tarSNPs, including alleles, MAFs, genes locations, the result of Hardy-Weinberg equilibrium and call rate, is shown in Table S1. However, rs14035 in the RAN gene did not meet the Hardy-Weinberg equilibrium (P < 0.05). Therefore, the further analyses of the polymorphism associations within T2DM group and health control group excluded rs14035 in the RAN gene.

DNA Extraction and Genotyping
A QIAmp Blood DNA Mini Kit (Qiagen, Hilden, Germany) was used to extract the genomic DNA from peripheral whole blood samples. A Sequenom MassARRAYiPLEX Gold analyzer (Sequenom, Life Technologies, Shanghai) was used to genotype the five selected SNPs. MassARRAY Assay Design 3.1 software were used to design PCR conditions and primers (Table S1).

Statistical Analysis
The demographic data and clinical characteristics of the included individuals are presented as the mean ± S.D. or number (percentage). The Shapiro-Wilk test was performed to identify the normality of the included demographic data and clinical characteristics in each comparison group (Tables S2-S5). For continuous variables with normal distribution, the t-test and ANOVA were performed on them, otherwise the non-parametric test (Kruskal-Wallis test and Mann-Whitney test) was conducted. For categorical variable, the χ 2 -test was performed. The logistic regression with different genetic model (codominant, dominant, and recessive models) was performed to calculate the ORs and 95% confidence intervals (CI) for estimating risk (44). Hardy-Weinberg equilibrium (HWE) of each SNP in the control group was calculated by χ 2 statistics. All the above statistical analyses were performed by SPSS software v. 22.0 (SPSS, Inc.).
The MDR (multifactor dimensionality reduction) software (45) was used to performed to identified the best model of gene-gene interaction models and gene-environment interaction models of the five selected SNPs. The environmental factors we included in MDR were age, gender, BMI (body mass index), TG (triglyceride), TC (total cholesterol), LDL (low density lipoprotein), HDL (high density lipoprotein), FBG (fasting blood glucose), post-prandial blood glucose, HbA1c, duration of T2DM, patients with family history of diabetes, former smoking, current smoking, current smoking, current drinking, and patents with hypertension. SHEsis, an online software (http://analysis.bio-x.cn/myAnalysis.php), was used to identified the effects of haplotype frequencies on T2DM and its vascular complications (46)(47)(48).

RESULT Population Characteristics
The baseline population characteristics of all included individuals are presented in Table 1. In comparison with healthy controls, the patients in T2DM group had higher levels of body mass index (BMI), triglyceride (TG), fasting blood glucose (FBG), GPT (glutamic-pyruvic, transaminase), serum creatinine, and blood uric acid but lower levels of low-density lipoprotein (LDL) and high-density lipoprotein (HDL), when compared with healthy controls. In comparison with T2DM patients without any complications, all T2DM patients with microvascular complications, macrovascular complications and micromacrovascular complications had higher average age, longer duration of T2DM, and high levels of serum creatinine, but they had lower estimated glomerular filtration rate (eGFR) and GPT levels. Additionally, in comparison with T2DM patients without any complications, hypertension was more prevalent in T2DM patients with vascular complications (macrovascular, microvascular, and micro-macrovascular complications). And T2DM patients with macrovascular complications included a larger number of former smokers ( Table 2). XPO5 rs11077, TARBP2 rs784567, DICER1 rs13078, and rs3742330 met the requirement of the HWE in healthy controls. However, RAN rs14035 deviated from HWE in healthy controls (Tables S2, S3).

Association Between T2DM and XPO5/DICER1/TARBP2 Polymorphisms
We identified the effects of XPO5, DICER1, and TARBP2 SNPs on T2DM. As presented in Table 2, there was a significant association between rs13078 in the DICER1 gene and a decreased risk of developing T2DM under the allelic mode (A vs. T: OR: 0.64; 95CI%: 0.42-0.95; P: 0.026), suggesting that individuals carrying the A allele had a 36% lower risk of developing

Association Among T2DM, T2DM Vascular Complications, Gene-Environment, and Gene-Gene Interaction
To explore the association among gene-environment interaction, T2DM and its vascular complications, MDR analysis was conducted to analysis the five SNPs in the RAN, XPO5, DICER1, and TARBP2 genes as well as environment factors ( Table S4).
The best gene-environmental interaction model on T2DM and its macrovascular complications were identified (DICER1 rs13078, BMI and TG; RAN rs14035, DOD and HP), with a significant TBA value and the highest CVC value ( Table 5). The risk analysis of the three-way interaction in the model was also performed ( Table S5). As shown in Figure 2, in comparison with the reference group (wild-type rs13078, normal BMI, and normal TG), the individuals with all three factors (mutation of rs13078, high BMI, and high TG) exhibited a 5.93-fold higher possibility of developing T2DM (OR: 5.93; 95%CI:1.25-28.26; P: 0.025). Figure 3 shows that the individuals with mutation of RAN rs14035, DOD more than 5 years and hypertension had a 40.60-fold higher possibility of suffering diabetic macrovascular However, no gene-gene interaction was found to be associated with the increase risk of T2DM and its vascular complications using MDR analysis.

Association Among T2DM, Its Vascular Complications, and DICER1 Haplotype Frequencies
The two SNPs (rs3742330 and rs13078) constitute a haplotype block spanning 3 kb of the DICER1 gene (Figure 3). We conducted a haplotype analysis among the DICER1 gene, T2DM and T2DM vascular complications. As shown in Table S6, compared with the highest frequency haplotype AT, the haplotype AA was significantly related to a lower risk of T2DM (OR: 0.63; 95%CI: 0.42-0.94; P: 0.023).

DISCUSSION
Alterations in the miRNA machinery play important roles in the pathogenesis of a variety of disorders (49), which may account for abnormal profiles of miRNA in various diseases. Recently, more and more studies have provided evidence that alterations in miRNA machinery may result in dysregulation or upregulation of miRNAs in diabetes (40,50,51). In our study, a significant 0.64-fold decrease was found in the A allele frequency at DICER1 rs13078 in T2DM patients than healthy individuals in the allelic model, suggesting that individuals carring the A allele at DICER1 rs13078 had a decreased possibility of developing T2DM than those carrying the T allele. Additionally, we also identified a protective effect of the AA haplotype in DICER1. The bold in the table means statistically significant (P < 0.05). DICER1, which situate in chromosome 14q32.13, contains 1922 amino acids in humans, encoding an approximately 218 kDa RNase III endonuclease (21,52). The DICER1 enzyme is responsible for the processing of gene-encoded pre-miRNAs into mature miRNAs, and it plays a key role in the highly conserved cellular pathway (52). DICER1 is also well known to be an important component in the oncogenic process of several cancers, such as breast cancer (53), hepatocellular carcinoma (54), lung cancer (55), and ovarian cancer (56). In metabolic diseases, Noren Hooten et al. (57) hypothesized that alterations FIGURE 2 | Risk analysis of gene-environment interaction Model: DICER1 rs13078, BMI, TG. The reference group is the interaction of wild-type for rs13078, normal BMI and normal TG. The OR value is presented in the figure. *P < 0.05 and **P < 0.001. of the levels of DICER1 gene may play an important role in organismal aging and the upregulation of expression of DICER1 gene may provide us a new pharmacotherapeutic approaches for age-related disease, such as T2DM. Furthermore, it has been demonstrated that different levels of expression of microRNAs were identified in exosomes isolated from the serum and in the blood of healthy control individuals compared with patients with T2DM, hypercholesterolemia or metabolic syndrome (58,59). It has been reported that the site of DICER1 rs13078 located in the 3 ′ -UTR of the DICER1 gene and might be an important component in the expression process of the DICER1 gene by influencing the miRNAs binding site, which might impact the DICER1 enzyme's function by DICER1 gene regulation and sequentially impact the expression of microRNAs (52,60,61). Additionally, rs13078 genetic variants in DICER1 have been reported to be related to other diseases, such as gestational hypertension (62) and larynx cancer (63).
In the association analysis between diabetic vascular complications and SNPs, we found that T2DM patients who carried TT+TC genotype of rs14035 in RAN gene had a 1.89fold increased risk of developing macrovascular complications compared with those with the CC wild-type genotype. RAN rs14035 is also located in the 3 ′ -UTR of the RAN gene at chr12:130876696 (64). The RAN enzyme as the member of the Ras superfamily of GTPases, plays a key component in the translocation of pre-miRNAs from the nucleus to the cytoplasm through the nuclear pore complex (65). RAN-GTP is depleted as a result of RAN guanine nucleotide exchange factor inhibition, and pre-miRNA export is greatly reduced, indicating that miRNA transport is mediated by a RAN-GTPbinding export receptor (66). These data not only suggest that mutations in RAN might play an important role in pathology-related changes in the transport and expression of miRNA but also imply that polymorphisms of the RAN gene might also be related to the biosynthesis and translation of miRNAs. Vorpahl et al. (67) suggested that inhibition or modulation of RAN-GTPase-activating protein expression might be sufficient to attenuate the vascular proliferative response following vascular injury. Additionally, RAN-GTPase signaling might also be essential for postnatal pancreatic islet development and glucose homeostasis (68). It has been reported that a significant upregulation of circulating miR-126 has been detected in patients with angina and acute myocardial infarction, while downregulation of miR-126 has been detected in plasma from patients with cancer, diabetes or heart failure (58,59). Moreover, it has been reported that circulating miR-126-3p, which may be a reliable biomarker of physiological endothelial senescence in elderly subjects with normoglycemia, underlies a mechanism that may be disrupted in elderly patients with DM (69). These studies give evidence the rs14035 polymorphism in RAN might be related to the occurrence and deterioration of the macrovascular complications of T2DM by regulating the expression level of RAN.
Recently, it has been reported that the susceptibility of type 2 diabetes mellitus and vascular complications of T2DM could be affected by various factors, including interactions between various environmental factors and the different polymorphisms (10). An MDR analysis was conducted to further investigate the multiple-factor, higher-order interaction of T2DM and the vascular complications of T2DM. The best three-factor interaction model among DICER1 rs13078, BMI (body mass index) and TG (triglyceride), suggested that rs13078 in DICER1 more likely interacts with overweigh and high TG to increase the possibility of developing T2DM. This outcome is consistent with the analysis of the clinical data between healthy control individuals and T2DM patients which showed that T2DM patients had significantly higher BMI and TG than healthy controls. Shai et al. (70) and Astrup et al. (71) also supported that individuals with abnormal lipid metabolism (high TG, TC, HDL, and low HDL) had a higher risk of developing T2DM. It has also been reported that BMI above normal weight level has been demonstrated to be one of the risk factors of T2DM and its complications (71)(72)(73). Moreover, our interaction results were supported by the synergistic effect and individual effect that overweight, increased triglyceride, high LDL and low HDL are associated with T2DM risk (74). In addition, another best threefactor interaction model (RAN rs14035, with hypertension and more than 5 years of T2DM duration) was also identified to be related to the increased risk of macrovascular complications of T2DM. Hypertension is known as one of the independent risk factors of T2DM (75). Moreover, the coexistence of hypertension and diabetes will significantly increase the risk of developing cardiovascular complications of DM, which means that hypertension may be an important risk factor for diabetic macrovascular complications (76,77). According to the ADA guidelines, with the increase in the duration of diabetes, the risk of developing diabetic vascular complications will significantly increase (78). Many studies also supported that the duration of diabetes played an important role in developing diabetic vascular complications (79,80). However, our study is the first time we found these two interaction models in T2DM and diabetic macrovascular complications, thus further studies with larger population samples are needed.
Similarly, there might be more environmental factors, such as diet behavior, passive smoke exposure and PM 2.5 , which might also modify the effects of genetic variants on the susceptibility of T2DM and its vascular complications (72,73). For example, the study by Eze et al. demonstrated the interaction between long-term PM 10 and gene polymorphisms at IL6-572 on T2DM (81). Therefore, more environmental factors could be included in further studies on interaction.
Our study has several limitations. First, the sample in the study was not very large, which might restrict the ability to explore weaker associations among T2DM, vascular complications of T2DM and SNPs. Second, the basic characteristics of the included individuals collected were not very comprehensive. Several clinical characteristics were not collected, such as the actual values of systolic and diastolic blood pressure. In addition, we should also include more meaningful environmental factors, such as diet, PM 10 and passive smoke, in our analysis. Therefore, more large and well-designed investigation is needed to support our findings.
As far as we know, this is the first study to investigate the association of these five polymorphisms (rs14035 in RAN, rs11077 in XPO5, rs13078 in DICER1, rs3742330 in DICER1 and rs784567 in TARBP2) with the risk of T2DM and its vascular complications. Our findings offer us the first evidence that rs13078 in DICER1 is related to a reduced risk of T2DM and that RAN rs14035 is relevant to an increased risk of macrovascular complication of T2DM in southern Chinese population. Additionally, our results also demonstrate two feasible interactions, DICER1 rs13078, BMI and TG in T2DM and RAN rs14035, hypertension and duration of T2DM in diabetic macrovascular complications. Our study may provide a new clue of epidemiology about the importance of miRNA processing genes (RAN, XPO5, DICER1, and TARBP2) in type 2 diabetes mellitus and diabetic vascular complications.

ETHICS STATEMENT
Our study was conducted in conformity with the rules of the Ethics Committee of Jinan University by written informed consent to all participant subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The Ethics Committee of Jinan University approved the protocol.

AUTHOR CONTRIBUTIONS
ZW, XZ, and XX contributed equally to the writing of this paper. LN provided all the samples for this study. SZ, XC, KZ, SD, JL, XH, DL, YW, JW, YL, YY, KL, CL, and BZ carried out data collection and the extraction of DNA. GY and CJ carried out whole design.