Association of ARHGAP22 gene polymorphisms with the risk of type 2 diabetic retinopathy
Abstract
Background
Little is known about the contribution of ARHGAP22 polymorphism to diabetic retinopathy (DR) risk. We performed a case–control study to investigate the associations between ARHGAP22 and the risk of DR in a Chinese Han population.
Methods
A total of 341 patients with type 2 diabetes mellitus (T2DM) were selected. All patients underwent a complete eye examination. Based on this, the patients with T2DM were divided into two subgroups: 188 patients with DR and 153 patients without DR. Five single nucleotide polymorphism (SNPs) were selected and genotyped using the MassARRAY method (Sequenom, San Diego, CA, USA). The odds ratio (OR) and 95% confidence intervals (CIs) were calculated by unconditional logistic regression adjusted for age and sex.
Results
Two susceptibility SNPs in ARHGAP22 were found to be associated with an increased risk of DR both before and after the adjustment: rs10491034 under the dominant model (adjusted OR = 0.51, 95% CI = 0.27–0.95, p = 0.032) and additive model (adjusted OR = 0.47, 95% CI = 0.26–0.84, p = 0.0098) and rs3844492 under the codominant model (adjusted OR = 3.14, 95% CI = 1.10–9.01, p = 0.023) and recessive model (adjusted OR = 3.52, 95% CI = 1.26–9.85, p = 0.011).
Conclusions
Our findings reveal a significant association between SNPs in the ARHGAP22 gene and DR risk in a Han Chinese population.
1 INTRODUCTION
Diabetic retinopathy (DR) is a severe chronic microvascular complication of diabetes mellitus (DM) and it may lead to blindness as a result of continuous blood leakage from retinal pericytes and endothelial cells if left untreated.1 DR is the leading cause of blindness among working adults in the western world2 and the overall prevalence of DR was reported to be 34.6% in patients with diabetes,3 which has been a global public health and economic problem. Although hyperglycemia, hypertension and dyslipidemia have been recognized as strong risk factors for DR,4-6 genetic factors also play an important role in its development and progression.
In recent decades, several studies have linked many varieties of genetic abnormalities with the onset and development of DR. Among them, angiotensin-converting enzyme (ACE), aldose reductase (AKR1B1), vascular endothelial growth factor (VEGF) and endothelial nitric oxide synthase (NOS3) gene polymorphisms were the most common susceptibility genes associated with the risk of DR.7-10 Recently, Rho GTPase-activating protein 22 (ARHGAP22) was reported to be involved in endothelial cell angiogenesis and increased capillary permeability and has also been considered as a new candidate gene associated with DR in a Taiwanese population.11 However, an insignificant association was detected in the distribution of ARHGAP22 polymorphisms between Indian DR cases and controls.12 By contrast, little is known about the contribution of the ARHGAP22 single nucleotide polymorphism (SNP) to DR risk in the Chinese Han population. We therefore performed a case–control study aiming to investigate the associations between ARHGAP22 SNPs and the risk of DR in a Chinese Han population.
2 MATERIALS AND METHODS
2.1 Ethics statement
The present study was conducted in strict obedience with the World Medical Association Declaration of Helsinki when using human tissue and when discussing the study protocol with subjects. The protocol was approved by the Ethical Committee of Tangdu Hospital. Each participant provided their written, informed consent.
2.2 Subjects
All participants in the present study were Han Chinese individuals who lived in Shaanxi Province. A total of 341 type 2 diabetes mellitus (T2DM) patients were consecutively recruited between June 2015 and June 2016 in the Tangdu Hospital, Xi'an, China. T2DM patients were diagnosed in accordance with the 2010 American Diabetes Association diagnostic criteria for diabetes,13 based on a fasting blood glucose >7 mmol/l, a causal blood glucose >11.1 mmol/l or a postprandial 2-h blood glucose >11.1 mmol/l following a 75-g oral glucose tolerance test, or a history of therapy for diabetes. All patients underwent a complete eye examination that included dilated retinal examination and fundus photography or fundus fluorescein angiography using a TRC-50DX Mydriatic Retinal Camera (Topcon Medical Systems, Oakland, NJ, USA). Patients were diagnosed with DR in accordance with the Early Treatment Diabetic Retinopathy classification.14 Patients with T2DM were divided into two subgroups: 188 patients with retinopathy (DR) and 153 patients without retinopathy (NDR). All the participants have at least 10 years history of T2DM. Demographic and clinical data including body mass index (BMI), fasting blood glucose (FBG), fasting insulin levels (INS), total cholesterol (TC), triglyceride level (TG), high-density lipoprotein cholesterol (HDLC) and low-density lipoprotein cholesterol (LDLC), urea, creatinine (Cr), cystatin-C (Cys-C) and other factors, including smoking and drinking status, antidiabetic agents and insulin usage, were recorded for each study participant.
2.3 SNP selection and genotyping
Candidate ARHGAP22 SNPs were selected from previous studies that associated polymorphisms with DR.11 SNPs with minor allele frequencies (MAF) > 5% in the HapMap CHB population were selected. We validated five SNPs in ARHGAP22. A GoldMag-Mini Purification Kit (GoldMag Co. Ltd, Xian, China) was used to extract genomic DNA from whole blood samples. DNA concentrations were measured using a DU530 ultraviolet/visible spectrophotometer (Beckman Instruments, Fullerton, CA, USA). Using MassARRAY Assay Design, version 3.0 (Sequenom, San Diego, CA, USA), we designed a multiplexed SNP MassEXTENDED assay.15 SNPs were genotyped using the standard protocol recommended by the manufacturer of the MassARRAY RS1000 (Sequenom) and data were analyzed using Typer, version 4.0 (Sequenom). The primers used in the present study are listed in Table 1.
| SNP_ID | 1st-PCRP | 2nd-PCRP | UEP_SEQ |
|---|---|---|---|
| rs1051509 | ACGTTGGATGTGCTCCAACCATGCAGTTCC | ACGTTGGATGGATATTTCCTGGGTAAGGGC | tGTAAGGGCTAAGCGCT |
| rs3813864 | ACGTTGGATGGGAAACTGGGTAGAGTGCAT | ACGTTGGATGCAATCGCACATACAATGCCC | ACAATGCCCACTCCC |
| rs10491034 | ACGTTGGATGGCTATTAAGGAGGCTCCTTC | ACGTTGGATGGACATTGACACTCCTGCTAC | TGGTTAAAGCTCCTATAACTT |
| rs3844492 | ACGTTGGATGACTTACCCTAGTCCCATTCC | ACGTTGGATGAATGATACTGTCACCAGGGC | TGCAAGACTCTTGCTTA |
| rs1822861 | ACGTTGGATGCACACTCAGCATGTTTCTTG | ACGTTGGATGTCTTGTCCTGATTCCAGACC | CCGGCATGGAAACTGGA |
2.4 Statistical analysis
We used Excel (Microsoft Corp., Redmond, WA, USA) and SPSS, version 17.0 (SPSS, Chicago, IL, USA) to perform the statistical analyses. p < 0.05 (two-sided) was considered statistically significant. All continuous data are presented as the mean ± SD. Pearson's chi-squared test and a Student's t-test were used to compare the distribution of categorical variables and continuous variables, respectively. Fisher's exact test was applied to each SNP in the controls to test for departure from Hardy–Weinberg equilibrium (HWE). Odds ratios (ORs) and 95% confidence intervals (CIs) for the allele and genotype frequencies were calculated using the Pearson chi-squared test adjusted for age and sex.16 PLINK software (http://pngu.mgh.harvard.edu/ purcell/plink) was used to assess SNP associations with DR risk in different genetic models (codominant, dominant, recessive, overdominat and additive). We used unconditional logistic regression analysis to calculate ORs and 95% CI adjusted for age and sex.17
3 RESULTS
We recruited 188 DR cases (90 males and 98 females; average age at diagnosis: 60.75 years) and 153 NDR controls (63 males and 90 females; average age: 53.82 years) for the present study. The clinical and biochemical characteristics of the cases and controls are shown in Table 2. Age, sex, BMI, smoking and drinking status, and antidiabetic agent usage of cases and controls were well-matched (p > 0.05). There is no significant difference in the distribution of FBG, INS, TC, HDLC, LDLC, urea, Cr and Cys-C levels between cases and controls (p > 0.05).
| Characteristics | DR (n = 188) | NDR (n = 153) | p-value |
|---|---|---|---|
| Sex | 0.080 | ||
| Female | 90 | 63 | |
| Male | 98 | 90 | |
| Duration of T2DM, years (mean ± SD) | 21.44 ± 8.12 | 20.03 ± 7.84 | 0.076 |
| Age, years (mean ± SD) | 60.75 ± 12.33 | 53.82 ± 14.35 | 0.443 |
| Smoking status | 0.210 | ||
| Yes | 50 | 54 | |
| No | 138 | 99 | |
| Drinking status | 0.538 | ||
| Yes | 27 | 28 | |
| No | 161 | 125 | |
| Taking antidiabetic | 0.689 | ||
| Yes | 67 | 62 | |
| No | 121 | 91 | |
| Insulin therapy | 0.001* | ||
| Yes | 133 | 55 | |
| No | 55 | 98 | |
| BMI (kg/m2) | 24.63 ± 3.45 | 25.18 ± 3.27 | 0.856 |
| FBG (mmol/l) | 10.08 ± 5.14 | 9.75 ± 4.07 | 0.057 |
| HbA1c (%) | 9.24 ± 2.47 | 9.39 ± 2.47 | 0.064 |
| TC (mmol/l) | 4.40 ± 1.21 | 4.89 ± 1.38 | 0.287 |
| TG (mmol/l) | 2.04 ± 1.42 | 3.05 ± 2.87 | 0.001* |
| LDL (mmol/l) | 2.62 ± 0.88 | 2.96 ± 0.99 | 0.436 |
| HDL (mmol/l) | 2.09 ± 0.75 | 1.27 ± 0.85 | 0.157 |
| Urea (mmol/l) | 6.44 ± 2.33 | 6.32 ± 4.27 | 0.643 |
| Cr (μmol/l) | 67.08 ± 31.38 | 60.41 ± 25.11 | 0.559 |
| Cys-C (mg/l) | 0.86 ± 0.25 | 0.77 ± 0.21 | 0.137 |
| ALBP (μg/l) | 9.96 ± 8.84 | 9.71 ± 7.95 | 0.279 |
| GFR (ml/min) | 116.31 ± 33.66 | 131.14 ± 38.15 | 0.255 |
| CP (ng/ml) | 1.12 ± 0.74 | 1.67 ± 2.16 | 0.008* |
| INS (ulU/ml) | 10.95 ± 4.77 | 10.55 ± 4.44 | 0.208 |
| UCRP (nmol/mmol) | 0.40 ± 0.33 | 0.40 ± 0.24 | 0.512 |
| RBP (mg/l) | 38.80 ± 11.48 | 38.75 ± 11.25 | 0.531 |
- BMI, body mass index; FBG, fasting blood glucose; TC, total cholesterol; TG, triglyceride; LDL, low-density lipoprotein; HDL, high-density lipoprotein; Cr, creatinine; Cys-C, cystatin-C; ALBP, adipocyte lipid binding protein; GFR, glomerular filtration rate; CP, C-peptide; INS, fasting insulin levels; UCRP, urine C-peptide per 24-h urine; RBP, retinol binding protein.
A total of five ARHGAP22 SNPs were identified in the cases and controls. All SNP call rates exceeded 98.2%, which was considered sufficiently high to perform association analyses. In the controls, all SNPs were in HWE (p > 0.05) (Table 3). However, no associations were observed between the alleles and DR risk in an allele model.
| SNP id | Chromosome | Position | Gene | Alleles | MAF | HWE p-value | DR | NDR | OR | 95% ci | p-value | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| A | B | DR | NDR | A | B | A | B | ||||||||
| rs1051509 | chr10 | 49654195 | ARHGAP22 | G | A | 0.340 | 0.359 | 0.863 | 128 | 248 | 110 | 196 | 0.920 | 0.670–1.262 | 0.604 |
| rs3813864 | chr10 | 49659843 | ARHGAP22 | A | G | 0.239 | 0.281 | 0.696 | 90 | 286 | 86 | 220 | 0.805 | 0.571–1.135 | 0.216 |
| rs10491034 | chr10 | 49810367 | ARHGAP22 | T | G | 0.096 | 0.144 | 0.769 | 36 | 340 | 44 | 262 | 0.630 | 0.394–1.008 | 0.052 |
| rs3844492 | chr10 | 49822801 | ARHGAP22 | G | A | 0.293 | 0.248 | 0.291 | 110 | 266 | 76 | 230 | 1.251 | 0.889–1.761 | 0.198 |
| rs1822861 | chr10 | 49834326 | ARHGAP22 | G | T | 0.452 | 0.493 | 0.874 | 170 | 206 | 151 | 155 | 0.847 | 0.626–1.146 | 0.282 |
- A, minor allele; B, reference allele; MAF, minor allelic frequency; HWE, Hardy–Weinberg equilibrium.
We also compared the genotype frequencies between cases and controls (Table 4). For SNP rs3844492, the genotype frequency distributions differed between cases and controls. Compared with the AA genotype, the GG frequency of rs3844492 polymorphism among cases was significantly different from controls (GG versus AA: OR = 3.043, 95% CI = 1.179–7.857, p = 0.021). After further adjustment by age and sex, the difference for the ‘GG’ genotype in rs3844492 remained significant (adjusted OR = 3.072, 95% CI = 1.096–8.611, p = 0.032), which suggested that the rs3844492 polymorphism had an increased effect on the risk of DR.
| SNP id | Genotype | Genotype frequencies | Without adjustment | With adjustment | |||
|---|---|---|---|---|---|---|---|
| DR | NDR | Or (95% ci) | pa | Or (95% ci) | pb | ||
| rs1051509 | AA | 79 (42.0%) | 62 (40.5%) | 1 | 1 | ||
| GA | 90 (47.9%) | 72 (47.1%) | 0.981 (0.623–1.546) | 0.934 | 1.098 (0.683–1.767) | 0.699 | |
| GG | 19 (10.1%) | 19 (12.4%) | 0.785 (0.383–1.608) | 0.508 | 0.813 (0.386–1.710) | 0.585 | |
| rs3813864 | GG | 107 (56.9%) | 78 (51.0%) | 1 | 1 | ||
| AG | 72 (38.3%) | 64 (41.8%) | 0.820 (0.525–1.280) | 0.383 | 0.832 (0.524–1.323) | 0.438 | |
| AA | 9 (4.8%) | 11 (7.2%) | 0.596 (0.234–1.509) | 0.275 | 0.643 (0.245–1.686) | 0.369 | |
| rs10491034 | GG | 152 (80.9%) | 112 (73.2%) | 1 | 1 | ||
| TG | 36 (19.1%) | 38 (24.8%) | 0.698 (0.416–1.171) | 0.173 | 0.787 (0.459–1.349) | 0.383 | |
| TT | 0 (0.0%) | 3 (2.0%) | – | – | – | – | |
| rs3844492 | AA | 100 (53.2%) | 83 (54.2%) | 1 | 1 | ||
| GA | 66 (35.1%) | 64 (41.8%) | 0.856 (0.546–1.343) | 0.499 | 0.861 (0.541–1.371) | 0.529 | |
| GG | 22 (11.7%) | 6 (3.9%) | 3.043 (1.179–7.857) | 0.021* | 3.072 (1.096–8.611) | 0.032* | |
| rs1822861 | TT | 62 (33.0%) | 39 (25.5%) | 1 | 1 | ||
| GT | 82 (43.6%) | 77 (50.3%) | 0.670 (0.403–1.112) | 0.122 | 0.636 (0.374–1.081) | 0.095 | |
| GG | 44 (23.4%) | 37 (24.2%) | 0.748 (0.413–1.353) | 0.337 | 0.715 (0.385–1.326) | 0.287 | |
- a p-values were calculated from unconditional logistic regression analysis.
- b p-values were calculated by unconditional logistic regression analysis with adjustments for age and sex.
- * p ≤ 0.05 indicates statistical significance.
We assumed the minor allele of each SNP was a DR risk factor compared to the wild-type allele and analyzed associations between SNPs and DR in various inheritance models (Table 5). Two susceptibility SNPs were found to be associated with an increased risk of DR both before and after the adjustment: rs10491034 under the dominant model (adjusted OR = 0.51, 95% CI = 0.27–0.95, p = 0.032) and additive model (adjusted OR = 0.47, 95% CI = 0.26–0.84, p = 0.0098) and rs3844492 under the codominant model (adjusted OR = 3.14, 95% CI = 1.10–9.01, p = 0.023) and recessive model (adjusted OR = 3.52, 95% CI = 1.26–9.85, p = 0.011).
| SNP id | Model | Genotype | DR | NDR | Or (95% ci) | p-value | AIC | BIC |
|---|---|---|---|---|---|---|---|---|
| rs1051509 | Codominant | A/A | 79 (41.8%) | 65 (41.1%) | 1 | 0.46 | 507.3 | 761.4 |
| A/G | 91 (48.1%) | 74 (46.8%) | 0.88 (0.51–1.54) | |||||
| G/G | 19 (10.1%) | 19 (12%) | 0.56 (0.23–1.39) | |||||
| Dominant | A/A | 79 (41.8%) | 65 (41.1%) | 1 | 0.43 | 506.3 | 756.5 | |
| A/G-G/G | 110 (58.2%) | 93 (58.9%) | 0.81 (0.48–1.37) | |||||
| Recessive | A/A-A/G | 170 (90%) | 139 (88%) | 1 | 0.24 | 505.5 | 755.7 | |
| G/G | 19 (10.1%) | 19 (12%) | 0.60 (0.25–1.42) | |||||
| Overdominant | A/A-G/G | 98 (51.9%) | 84 (53.2%) | 1 | 0.94 | 506.9 | 757.1 | |
| A/G | 91 (48.1%) | 74 (46.8%) | 0.98 (0.58–1.66) | |||||
| Log-additive | – | – | – | 0.80 (0.54–1.18) | 0.26 | 505.6 | 755.8 | |
| rs3813864 | Codominant | G/G | 107 (56.6%) | 81 (51.3%) | 1 | 0.2 | 505.7 | 759.8 |
| A/G | 73 (38.6%) | 66 (41.8%) | 0.65 (0.38–1.12) | |||||
| A/A | 9 (4.8%) | 11 (7%) | 0.52 (0.17–1.58) | |||||
| Dominant | G/G | 107 (56.6%) | 81 (51.3%) | 1 | 0.081 | 503.8 | 754.1 | |
| A/G-A/A | 82 (43.4%) | 77 (48.7%) | 0.63 (0.38–1.06) | |||||
| Recessive | G/G-A/G | 180 (95.2%) | 147 (93%) | 1 | 0.39 | 506.1 | 756.4 | |
| A/A | 9 (4.8%) | 11 (7%) | 0.62 (0.21–1.84) | |||||
| Overdominant | G/G-A/A | 116 (61.4%) | 92 (58.2%) | 1 | 0.17 | 505 | 755.2 | |
| A/G | 73 (38.6%) | 66 (41.8%) | 0.69 (0.41–1.18) | |||||
| Log-additive | – | – | – | 0.68 (0.45–1.05) | 0.078 | 503.8 | 754 | |
| rs10491034 | Codominant | G/G | 153 (81%) | 113 (71.5%) | 1 | 0.003 | 497.3 | 751.3 |
| G/T | 36 (19.1%) | 41 (25.9%) | 0.60 (0.32–1.14) | |||||
| T/T | 0 (0%) | 4 (2.5%) | 0.00 (0.00–Na) | |||||
| Dominant | G/G | 153 (81%) | 113 (71.5%) | 1 | 0.032* | 502.3 | 752.5 | |
| G/T–T/T | 36 (19.1%) | 45 (28.5%) | 0.51 (0.27–0.95) | |||||
| Recessive | G/G-G/T | 189 (100%) | 154 (97.5%) | 1 | 0.0024 | 497.7 | 747.9 | |
| T/T | 0 (0%) | 4 (2.5%) | 0.00 (0.00–Na) | |||||
| Overdominant | G/G-T/T | 153 (81%) | 117 (74%) | 1 | 0.12 | 504.5 | 754.7 | |
| G/T | 36 (19.1%) | 41 (25.9%) | 0.61 (0.32–1.15) | |||||
| Log-additive | – | – | – | 0.47 (0.26–0.84) | 0.0098* | 500.2 | 750.4 | |
| rs3844492 | Codominant | A/A | 100 (52.9%) | 85 (53.8%) | 1 | 0.023* | 501.4 | 755.4 |
| G/A | 67 (35.5%) | 66 (41.8%) | 0.76 (0.44–1.31) | |||||
| G/G | 22 (11.6%) | 7 (4.4%) | 3.14 (1.10–9.01) | |||||
| Dominant | A/A | 100 (52.9%) | 85 (53.8%) | 1 | 0.91 | 506.9 | 757.1 | |
| G/A-G/G | 89 (47.1%) | 73 (46.2%) | 0.97 (0.58–1.62) | |||||
| Recessive | A/A-G/A | 167 (88.4%) | 151 (95.6%) | 1 | 0.011* | 500.4 | 750.6 | |
| G/G | 22 (11.6%) | 7 (4.4%) | 3.52 (1.26–9.85) | |||||
| Overdominant | A/A-G/G | 122 (64.5%) | 92 (58.2%) | 1 | 0.12 | 504.4 | 754.6 | |
| G/A | 67 (35.5%) | 66 (41.8%) | 0.66 (0.39–1.11) | |||||
| Log-additive | – | – | – | 1.22 (0.82–1.81) | 0.32 | 505.9 | 756.1 | |
| rs1822861 | Codominant | T/T | 63 (33.3%) | 42 (26.6%) | 1 | 0.15 | 505.1 | 759.1 |
| G/T | 82 (43.4%) | 78 (49.4%) | 0.54 (0.28–1.01) | |||||
| G/G | 44 (23.3%) | 38 (24.1%) | 0.65 (0.32–1.34) | |||||
| Dominant | T/T | 63 (33.3%) | 42 (26.6%) | 1 | 0.063 | 503.4 | 753.6 | |
| G/T-G/G | 126 (66.7%) | 116 (73.4%) | 0.57 (0.32–1.04) | |||||
| Recessive | T/T-G/T | 145 (76.7%) | 120 (76%) | 1 | 0.89 | 506.9 | 757.1 | |
| G/G | 44 (23.3%) | 38 (24.1%) | 0.96 (0.52–1.75) | |||||
| Overdominant | T/T-G/G | 107 (56.6%) | 80 (50.6%) | 1 | 0.12 | 504.4 | 754.6 | |
| G/T | 82 (43.4%) | 78 (49.4%) | 0.66 (0.39–1.11) | |||||
| Log-additive | – | – | – | 0.80 (0.56–1.14) | 0.22 | 505.4 | 755.6 |
- AIC, Akaike's information criterion; BIC, Bayesian information criterion; NA, not available.
- * p ≤ 0.05 indicates statistical significance.
4 DISCUSSION
An epidemiologic study of DR has shown that only 28.8% of diabetic patients develop retinopathy early,18 which suggested that genetic factors could promote the onset of retinopathy in diabetic patients. Genome-wide association studies (GWAS) have identified several susceptibility genes for DR; however, little information has been found about ARHGAP22 polymorphisms and the risk of DR, and the results are controversial.11, 12 In the present case–control study, we investigated the association between five ARHGAP22 SNPs and the risk of DR in the Chinese Han population. We found that SNP rs3844492 is associated with an increased risk of DR, whereas rs10491034 is associated with a decreased risk of DR.
ARHGAP22 is located at locus 10q11.22. The protein encoded by ARHGAP22 is insulin-responsive and is a member of the GTPase activating protein family.19 It is dependent on the kinase Akt and requires the Akt-dependent 14–3-3 binding protein, which binds sequentially to two serine residues. To date, multiple transcript variants encoding different isoforms have been found for this gene.20 In the present study, we found that genetic polymorphisms of ARHGAP22 are associated with DR risk, which may shed a new light on in-depth studies concerning this gene.
Earlier GWAS studies revealed two common variants (rs2300782 and rs10519765) that influence DR risk in populations of Mexican-Americans.21 Replication analysis for DR in Caucasian populations provided two SNPs (rs4865047 and rs1902491) that could be pursued in subsequent studies.22 However, none of the previously reported index SNPs has yet been validated in the East-Asian population. Subsequently, Chinese investigators have identified novel loci in the Chinese population, including rs9565164, rs1399634 and rs2380261.23 Japanese investigators also conducted GWAS studies for DR in a Japanese population and found rs9362054 in an intron of RP1-90 L14.1 showing borderline genome-wide significance.24 However, little information has been found about ARHGAP22 SNPs associated with DR in Chinese or Japanese populations. In the present study, we first showed that rs3844492 and rs10491034 in ARHGAP22 are associated with DR risk in a Chinese Han population. These results need be confirmed in the further studies with a larger sample size and different populations.
The present study has several potential limitations. First, the sample size is relatively small and the participants included only those from a Chinese population. Second, DR is a very heterogeneous disease with many other risk factors, including hyperglycemia, hypertension and dyslipidemia. We could not completely eliminate the potential influences of these factors on the results.
In summary, our results indicate that ARHGAP22 rs3844492 is associated with an increased risk of DR, whereas rs10491034 was associated with a decreased risk of DR in Chinese Han T2D patients. As a preliminary study with a small sample size, the conclusions drawn from our work must be viewed with caution; however, our observations are encouraging and certainly warrant more suitably powered studies of this relationship. Further studies aim to focus on validating our findings in large-scale East-Asian populations with long-term T2D follow-up data.
ACKNOWLEDGEMENTS
This work is supported by China Postdoctoral Science Foundation funded projects (2016 M592833). The authors have no conflicts of interest to report.
