Genetic polymorphism of the Dab2 gene and its association with Type 2 Diabetes Mellitus in the Chinese Uyghur population

Human subjects and biochemical analysis

The Ethics Committee of the first affiliated hospital of Xinjiang Medical University (Urumqi, China) approved this retrospective study (20190505-01), and informed consent was obtained from each study participant. We used PASS 15.0.5 software to calculate that the minimum sample size needed for this study was 489 (Power ≥ 0.9, α < 0.05) (Xu et al., 2013a). We randomly recruited 2,500 inpatients of the heart center of the first affiliated hospital of Xinjiang Medical University from September 1, 2016 to December 31, 2021. A total of 2,157 subjects were included (528 T2DM, 1,629 control) in the final case-control study; among them, 1,326 subjects were Han Chinese (331 T2DM, 995 control) and 831 subjects were Uyghur Chinese (197 T2DM, 634 control). T2DM was identified using discharge diagnoses obtained from the electronic medical records of the patients, and any patient with a new or existing T2DM diagnosis was included in the T2DM group. The inclusion criteria of the T2DM group were: subjects with a fasting plasma glucose (FBS) ≥ 126mg/dL ( ≥7.0 mmol/L) or two-hour post prandial blood sugar (PPBS) concentrations ≥ 200 mg/ dL (≥11.1 mmol/L) in the oral glucose tolerance test (Joseph et al., 2022), or subjects who had previously used anti-diabetic drugs (American Diabetes Association Professional Practice C, 2022). Control subjects also received related laboratory testing to verify they did not have T2DM. Subjects were not included in the study if they had been diagnosed with: type 1 diabetes, gestational diabetes, or secondary diabetes; any severe liver, kidney, or hematopoietic system disease; or thyroid dysfunction. A standard venipuncture technique was used to obtain blood samples from the participants. Complete blood testing and biochemical assays were performed using the equipment for chemical analysis in the clinical laboratory department of the First Affiliated Hospital of Xinjiang Medical University. We extracted the DNA from the blood leukocytes at the periphery using the blood genome extraction kit (Adi et al., 2021). BMI was obtained by calculating the participant’s weight divided by the square of their height (kg/m²). We defined hypertension as a systolic blood pressure ≥140 mmHg and/or a diastolic blood pressure ≥90 mmHg in at least two clinic blood pressure measurements on different days (Kollias et al., 2022). We defined smoking as currently or previously smoking, and drinking was defined as drinking alcohol at least once a week for at least a year (Xie et al., 2010).

Genotyping

Using the Haploview 4.2 software and International HapMap Project website phase II data base, we obtained four single nucleotide polymorphisms (SNPs) of the Dab2 gene: rs1050903, rs2255280, rs2855512 and rs11959928, by using minor allele frequency (MAF) ≥0.05 and linkage disequilibrium patterns with r² ≥0.8 as a cutoff. The linkage disequilibrium (LD) map of these four tag SNPs is shown in Fig. 1. The genotyping of the SNPs was performed using an improved multiplex ligation detection reaction (iMLDR) technique (Genesky Biotechnologies Inc., Shanghai, China). Genotyping was performed in a blinded fashion without knowledge of the participants’ clinical data, and a total of 10% of the genotyped samples were duplicated to monitor genotyping quality.

Statistical analysis

IBM SPSS Statistics 27.0 for Windows and R version 4.2.1 (R Core Team, 2022) were used to perform the statistical analyses in this study. We chose the double-tail P-value of 0.05 as the statistically significant threshold. The mean ±standard deviation (SD) represents the continuous variable under normal distribution, and the median (25th, 75th percentiles) represents the continuous variable with non-normal distribution. The difference between two groups was assessed with the two-independent-sample T-test or the Mann–Whitney U test. The categoric variables were described using numbers and percentages (%), and the χ² test was used to analyze differences. The allele and genotype differences among groups were evaluated by the Hardy-Weinberg equilibrium (HWE) and an unconditional logistic regression analysis was performed to evaluate the value of Dab2 polymorphism on T2DM, according to computing crude or adjusted ORs and 95%Cls.

A nomogram was used to build a multi-factor regression model, score the value level of each influencing factor in the model according to the contribution degree (the size of regression coefficient) add all the scores to get the total score, and finally calculate the predicted value of the individual’s final event through the functional transformation relationship between the total score and the probability of the final event (Zhang et al., 2021). Nomograms are often used by clinicians to visualize the regression equation to evaluate a patient’s condition and prognosis (Balachandran et al., 2015; Lv et al., 2022) The scoring system in the nomogram in this study was generated by the Regression Modeling Strategies (RMS) package in R (version 4.2.1).

The nomogram model was evaluated from three aspects: discrimination ability, calibration ability, and clinical effectiveness. A receiver operation characteristic curve (ROC) was used to evaluate the discrimination ability. The value of the area under curve (AUC) is always between 0.5 and 1, with an AUC value closer to 1 indicating a good performance of the predictive model. The C-index is used to measure how well a model predicts disease risk, with an absolute value close to 1 indicating strong predictive ability of the model. Calibration plots assess the predictive accuracy and agreement between predicted and observed severity.

Characteristics of study participants

The general data and clinical characteristics of the T2DM group and the control group of the Han and Uygur population included in this study are described in Tables 1 and 2. For both populations, there were no significant differences in age and sex between the T2DM group and the control group. In the Han population, there were no significant differences in body mass index (BMI) and triglyceride (TG) levels and the ratio of reported drinkers (all P > 0.05). The Han population T2DM group had higher levels of systolic blood pressure (SBP), HbA1c, left ventricular ejection fractions (LVEF), total cholesterol (TC), LDL-C, smoking, and a higher hypertension ratio than the control group (all P < 0.05, Table 1). In the Uygur population, there were no significant differences in high-density lipoprotein cholesterol (HDL-C) levels and the ratio of reported smokers and drinkers between the two groups (all P >  0.05), but the T2DM group had a significantly higher ratio of hypertension and significantly higher BMI, SBP, HbA1c, TG, TC, and LDL-C levels (all P < 0.05, Table 2) than those in the control group.

The genotype and allele distributions of the selected SNPs in the T2DM group and controls

The genotype distributions of the four SNPs were under the Hardy-Weinberg equilibrium (all P > 0.05) in both the T2DM and the control groups. However, there may be significant differences in gene variations when considering the genetic backgrounds of different populations. To test this possibility, we compared the distributions of genotype and allele frequency between the two study populations (Table S1). The distribution of the genotypes, genetic models and allele frequency all showed significant differences between the Han and Uygur populations for rs2255280, rs2855512, and rs11959982 (all P < 0.001). Based on these results showing the impact of population stratification, we divided the subjects into two groups, Han and Uyghur Chinese population group, and all analyses were performed separately.

Table 3 shows the distribution of genotypes, genetic models (dominant model, recessive model, and overdominant model) (Sabourin, Nobel & Valdar, 2015) and allele frequencies for the four SNPs (rs1050903, rs2255280, rs2855512, and rs11959982) of the Dab2 gene in the Han population group. There were no significant differences in genotypes, all genetic models, or allele frequencies between the T2DM and control group in the Han population group (all P > 0.05, Table 3).

Table 4 shows the distribution of genotypes, genetic models, and alleles for the four SNPs (rs1050903, rs2255280, rs2855512, and rs11959982) of the Dab2 gene in the Uygur population group. For rs2255280, the distribution of the genotypes and the recessive model (CC vs. AA + CA) showed significant differences between the T2DM and the control group (P = 0.011 and P = 0.011, respectively). This suggests that patients with the CC genotypes of rs2255280 may have a lower risk of T2DM. For rs2855512, the distribution of the genotypes and the recessive model (CC vs. AA + CA) also showed significant differences between T2DM and the control group (P = 0.017 and P = 0.015), indicating that patients with the CC genotype of rs22855512 might also have a lower risk of T2DM than those with AA or CA genotypes. There were no significant differences for rs1050903 or rs11959982 in the distribution genotypes and different genetic models between the T2DM and the control group in the Uygur population group (all P >0.05).

Multivariable logistic regression analysis for T2DM and control group in the Uygur population group

Tables 5 and 6 show the multivariable logistic regression analyses of the major confounding factors for T2DM in the Uygur population group. After multivariate adjustment for the confounders, such as BMI, TG, TC, LDL-C, and prevalence of smoking and hypertension, the recessive model of rs2255280 (CC vs. CA+AA: OR = 5.303, 95% CI = 1.236–22.755, P = 0.025, Table 5) and the recessive model of rs2855512 (CC vs. CA+AA: OR = 4.892, 95% CI = 1.136–21.013, P = 0.033, Table 6) remained significantly associated with T2DM.

Genotypes and glucose levels

In order to further judge the relationship between SNPs and blood glucose level in the Uygur population group, we drew a bar chart of the correlation between rs2255280 and rs2855512 genotype and blood glucose level including 831 Uyghur Chinese patients (197 T2DM, 634 control). Blood glucose level was represented by FBG and HbA1c. Figure 2 shows that the plasma glucose and HbA1c levels were significantly lower in the CC genotype than that in AA or CA genotypes both in rs2255280 (Figs. 2A, 2B) and rs2855512 (Figs. 2A, 2B); P < 0.05). The plasma glucose and HbA1c levels were significantly lower in the recessive model (CC) than that in the AA+CA model both in rs2255280 (Figs. 2C, 2D) and rs2855512 (Figs. 2B, 2D; P < 0.05). After excluding 98 patients who had received drug treatment, the correlation between blood glucose level and gene polymorphism was analyzed in the remaining 733 untreated Uygur subjects (Fig. S1). Even with the possible effects of drugs on blood glucose levels excluded, the plasma glucose and HbA1c levels were still significantly lower in the recessive model (CC) than in the AA+CA model in rs2255280 (P < 0.05).

Predictive nomogram for T2DM and validation of the nomogram

Figure 3 shows the nomogram based on the results of the logistic regression analysis. Hypertension, BMI, TG, TC, LDL-C, uric acid and rs2255280 (all P < 0.05) were included in the final model to develop the nomogram. The nomogram is a graphic depiction of the model, predicting the risk of T2DM. Refer to the nomogram legends for how to predict risk. For example, a person with hypertension (9.6 points), BMI of 27 (7.1 points), uric acid of 250 mmol/L (30 points), TG of 1.5 mmol/L (3.1 points), TC of 5.2 mmol/L (72.5 points), LDL-C of 2.3 mmol/L (20 points) and rs2255280 AA genotype (21 points), would have a nomogram total point score of 163.3, indicating an estimated 42.1% chance of experiencing T2DM.

This nomogram was validated internally, based on discrimination and calibration, using the bootstrap method with 1,000 resamples. As shown in Fig. 4A, the AUC value of the nomogram was 0.704 (95% CI [0.662–0.746]; P < 0.001) and the C-index was 0.704 (95% CI [0.663–0.776]; P < 0.001), indicating the model had a good predictive power. The calibration curve of the model is shown in Fig. 4B.