The samples (diabetic nephropathy, type-2 diabetes mellitus, and healthy controls) for this study were recruited from the subjects attending Department of General Medicine, Chettinad Hospital and Research Institute (CHRI), Tamil Nadu, India, from January to June 2016. The selection criteria for DN subjects by clinical examination including the routine biochemical assessment such as excreting urine albuminuria (>300mg/L) and serum creatinine (>1mg/dL) were considered and histopathological investigation. The T2DM subjects were diagnosed based on the criteria defined by the World Health Organization (WHO).13 The controls were selected by following the similar clinical and biochemical examination as done for DN and T2DM. All study participant were age-matched, belonging to South India, Asian ethnicity. The current study protocols were reviewed and approved by the Institutional Human Ethics Committee (03/IHEC/3-16) of the Chettinad Academy of Research and Education. Informed consent was obtained from the participants before the sampling process. Likewise, HIV positive patients, Juvenile diabetic subjects, physically challenged and subjects with cardiac problems were excluded. The demographic data and baseline clinical characteristics from each study subject including gender, age, Body Mass Index (BMI), the onset of disease, symptoms of the disease, family history, diet, and physical activity were recorded at the time sample collection.

Venous blood (3ml) was collected in EDTA coated tube (BD, USA) from the participants to isolate Genomic DNA using the phenol-chloroform method with minor modifications.14 The genetic analysis of rs1801282 was performed by Amplification Refractory Mutation System-Polymerase Chain Reaction (ARMS-PCR) using allele-specific primers (Table 1).15 For each PCR reaction, a 10μL reaction mix was setup containing 25ng DNA, 1U TAQ Brazilian Origin, 0.3mmol of each dNTP, 12pmol/μL of each primer. The ARMS-PCR program was performed using an ABI Veriti® Thermal Cycler (California, USA). The reaction condition for ARMS-PCR was set as initial denaturation (94°C, 5min), denaturation (94°C, 45s), annealing (68°C, 45s), elongation (72°C, 45s) and final elongation at 72°C for 5min. The amplicons were visualized on 1.6% agarose gel together with 100bp DNA Ladder (Dye Plus, Cat no: 3422A, Takara Bio). Finally, the identified genotypes were further confirmed in the randomly selected subset of overall samples (DN=12; T2DM =10; controls=08) using Sanger based sequencing (Applied Biosystems 3130, USA). Additionally, to determine the chromosomal interactions among the SNP variants, we used 3DSNP (http://www.cbportal.org/3dsnp/index.html) software package to generate Circos (circular) plots based on r2 values for visualizing the genomic data.

The statistical analysis for case–control study was executed by SPSS V.21 (IBM-Analytics, USA) software. The allelic and genotypic frequencies of rs1801282 polymorphism in DN, T2DM and healthy controls were calculated using the Pearson Chi-square test. To determine the distribution of SNP in controls, Hardy-Weinberg Equilibrium (HWE value>0.05) was used. Additionally, their association of polymorphism was analyzed by determining the odds ratio (OR), confidence intervals (95% CIs) in dominant (Jj+jj vs. Jj) (J-major, j-minor allele) and recessive (jj vs. JJ+Jj) genetic models.

An extensive literature to select the studies published between January 1991 to December 2017, that discuss the association between rs1801282 polymorphism with DN and T2DM from the database of MEDLINE, PubMed, Cochrane Library, EMBASE, and Google Scholar. The multiple key terms such as “Diabetic Nephropathy” “Type 2 Diabetes Mellitus”, “DN”, “T2DM”, “Peroxisome proliferator-activated receptor gamma gene”, “PPARγ gene”, “Polymorphism”, “rs1801282” were used to retrieve the relevant articles published in English Language for the meta-analysis with Diabetic Nephropathy vs. Type 2 Diabetes Mellitus. Similarly, multiple keywords such as “Type 2 Diabetes Mellitus”, “T2DM”, “Peroxisome proliferator-activated receptor gamma gene”, “PPARγ gene”, “Genetic association studies”, “SNP”, “rs1801282” were used to perform the similar meta-analysis between T2DM vs. controls.

Studies included in this meta-analysis (Diabetic Nephropathy vs. Type 2 Diabetes Mellitus; Type 2 Diabetes Mellitus vs. controls) were essential to meet the following criteria's: first, the association of PPARγ gene and rs1801282 polymorphism with DN and T2DM, T2DM and controls should be assessed. Second, study should be cases–controls (two groups). Third, the articles should provide adequate data to calculate allele and genotype frequencies for the meta-analysis. The information from the collected articles was extracted by two authors [AH and PA], any discrepancy between the authors was resolved by a group discussion (IR, SSJ, and RK). The study characteristics such as first author name, year of publication, country, ethnicity, the source of DNA, the sample size for cases-controls, genotypic frequency and genotyping method were extracted.

The methodological quality assessment of each study was evaluated by two independent variables such as the Hardy-Weinberg Equilibrium (HWE)16 and the Newcastle Ottawa Scale (NOS).17 The genotype distribution of included studies among the controls should follow HWE p-value above 0.05. The NOS grading system relies on three main factors such as selection, comparability, and exposure. The studies score up to a maximum of nine points, six or above were considered in this meta-analysis. In the meta-analysis, rs1801282 SNP with risk of DN vs. T2DM and T2DM vs. controls, were performed to calculate the pooled odds ratios (OR) and 95% confidence interval (CI) with (p-value<0.05) under allelic (j vs. J) (J-major, j-minor allele), homozygote (jj vs. JJ), heterozygote (Jj vs. JJ), dominant (Jj+jj vs. JJ) and recessive (jj vs. JJ+Jj) genetic models. The heterogeneity between the included studies in this meta-analysis was assessed by the I2 test18 and Q-statistics. Based on the results obtained from heterogeneity (I2<50), fixed effect (Mantel-Haenszel's)19 method was used, if not, the random-effect method (DerSimonian and Laird's) was applied. The publication bias for meta-analysis was assessed by Begg's funnel plot and Egger's test.20 Leave-one-out method was executed to validate the consistency of our results.21 The meta-analysis was performed using Revman V.5.0 (Cochrane Collaboration, UK) and STATA V.12.0 (Stata Corp, USA).

The baseline clinical characteristics of the study participants (DN=128; T2DM=148 and controls=148) were illustrated in Table 2. Among the samples enrolled in the study, their average±SD of age in DN, T2DM, and control groups were 56.26±8.9, 54.75±6.37 and 54.15±7.50 years respectively. Similarly, the body mass index (BMI) were (21.09±3.3), (26.23±3.68) and (23.08±2.37) for DN, T2DM and control.

The genotyping of PPARγ gene polymorphism (rs1801282) was performed using ARMS-PCR and Sanger dideoxy sequencing (Fig. 1). The genotype frequency distributions in controls (T2DM) were in concurrence with Hardy–Weinberg Equilibrium (p>0.05). The allele and genotype frequencies of rs1801282 polymorphisms were assessed to determine the odds ratio (OR), confidence intervals (95%CIs), χ2 and p-value (Table 3). The genotype distribution of rs1801282 polymorphism was 73.43% for JJ, 17.18% for Jj, 09.37% for jj in DN cases and 67.56%, 27.02%, 5.40% in T2DM, respectively. The analysis of rs1801282 polymorphism illustrated a positive association with DN compared with T2DM (p-value≤0.05) in south India. In heterozygous (Jj) and homozygous recessive genotype (jj) model of rs1801282 showed significant risk association in DN patients compared with T2DM with OR=1.19 (95%CI [0.75–1.98]) p-value=0.049, OR=2.72 (95%CI [0.96–7.67]) p-value=0.047, respectively. Further, the analysis of dominant and recessive genetic models showed an insignificant difference between DN and T2DM for rs1801282 polymorphism.

Fig. 1.

DNA sequence electropherograms of rs1801282 polymorphism in the PPARγ gene. Examples of homozygous dominant (JJ genotype) and homozygous recessive (jj genotype) condition of the studied SNP.

(0.35MB).

The susceptibility analysis of rs1801282 polymorphism with T2DM and controls was performed. The genotypic frequency distribution in healthy controls was in agreement with HWE (p>0.05). The allele and genotype frequencies of rs1801282 polymorphisms were used to estimate the odds ratio (OR), confidence intervals (95% CIs), χ2 and p-value (Table 4). The genotypic distribution of rs1801282 polymorphism was 67.56% for JJ, 27.02% for Jj, 05.40% for jj in T2DM and 81.08%, 16.21%, 02.70% in controls, respectively. The analysis of rs1801282 polymorphism showed a positive association with T2DM compared with controls (p-value≤0.05). The minor ‘G’ allele showed a positive association among T2DM and controls with OR=1.92 (95%CI [1.20–3.07]) p-value=0.003. In heterozygous (Jj) genotype of rs1801282 showed significant risk association among T2DM and controls with OR=2.00 (95%CI [1.12–3.54]) p-value=0.011, respectively. Likewise, the analysis of dominant genetic model showed significant association among T2DM and controls with OR=1.48 (95%CI [0.28–3.83]) p-value=0.005 for rs1801282 polymorphism.

The susceptibility analysis of rs1801282 variant with DN and controls was performed (Table 5). The genotype frequency distributions in controls were in agreement with HWE (p>0.05). The genotype distribution of rs1801282 polymorphism was 73.43% for JJ, 17.18% for Jj, 09.37% for jj in DN cases and 81.08%, 16.21%, 02.70% in control, respectively. The minor ‘G’ allele showed a positive association among T2DM and control with OR=1.80 (95%CI [1.13–2.93]) p-value=0.011. However, the analysis of rs1801282 polymorphism showed a negative association between DN and control with p-value≥0.05 in the heterozygote, homozygous recessive, dominant and recessive genetic models, respectively. The nucleotide sequences of PPARγ genotypes were deposited (Accession numbers: KY547832, KY547833, KY547834, KY547835, and KY547836) in the NCBI Genbank [https://www.ncbi.nlm.nih.gov/genbank/] database. The Circos plot (outer circle to inner) illustrates the chromatin states, annotated genes, histone modifications, transcription factors, currently studied variant (rs1801282) with related SNPs (r2) and 3D chromatin interactions (Fig. 2).

Fig. 2.

Circos plot illustrating the chromosomal interactions among the current SNP (rs1801282) and its associated SNPs. The Circos plot (outer to inner) denotes chromatin states, annotated genes, histone modifications, transcription factors, rs1801282 variant with associated SNPs (r2).

(0.7MB).

In meta-analysis, a total of 678 papers published before December 2017 was retrieved from the initial search in literature databases. The editorials, review articles, duplicates, and case reports were removed after a screening of the abstracts. Further, the studies were assessed for quality control methods such as Hardy-Weinberg Equilibrium and Newcastle Ottawa Scale. Thus, four studies22–25 were added to the meta-analysis of rs1801282 polymorphism and the characteristics of included studies in this meta-analysis were explained in supplementary Table 1. The genotypic, allelic frequencies and HWE/Chi-square values for the included studies were presented in Table 6.

The literature search published from 1991 up to December 2017 retrieved 846 relevant studies from several electronic databases. After removing the duplicate studies, case reports, studies on cell lines, 236 records remained. According to the quality assessment criteria, followed by HWE analysis, nine studies T2DM=6365, controls =5322 participants26–33 were eligible for this meta-analysis. The characteristics (author name, year of publication, country, ethnicity, DNA source, a sample size of cases-controls and genotyping methods) of each study included in the meta-analysis were defined in supplementary Table 2. The genotype, allele frequencies, and HWE/Chi-square values for the studies included in this meta-analysis were presented in Table 7.

The analysis of rs1801282 polymorphism among DN and T2DM revealed no heterogeneity in allelic (I2=22%) and dominant (I2=18%). Whereas, moderate heterogeneity was noticed in homozygote (I2=49%), heterozygote (I2=52%) and recessive (I2=44%) genetic models. Based on values of heterogeneity (I2) the fixed effects (Mantel–Haenszel's) model was implemented, which showed insignificant association with DN risk in allelic (j vs J) with (Q test p=0.27), OR=0.89, (95%CI [0.71–1.10]), p=0.26, homozygote (jj vs JJ) (Q test p=0.12), OR=0.87, (95% CI [0.66–1.12]), p=0.27, dominant (Jj+jj vs JJ) (Q test p=0.30), OR=0.86, (95% CI [0.66–1.12]), p=0.27 and recessive (jj vs JJ+Jj) (Q test p=0.15), OR=0.88, (95% CI [0.54–1.44]), p=0.61 genetic models, respectively. Whereas, the random effects model was used which showed insignificant association with DN risk in the heterozygote (Jj vs JJ) (Q test p=0.10), OR=1.03, (95% CI [0.44–2.41]), p=0.94. The results of the meta-analysis were shown as forest plots (Fig. 3). Further, funnel plot (Fig. 5) and Egger's test was executed which revealed no publication bias in the studied genetic models. Since a low number of studies were from Caucasian and Asian ethnicity, sub-group meta-analysis was not performed.

Fig. 3.

Forest plots showing individual and pooled ORs (95% CI) of the risk for rs1801282 polymorphism with Diabetic Nephropathy and Type 2 Diabetes Mellitus. The Forest Plots provides an insignificant association between the PPARγ polymorphism with DN and T2DM under five genetic models. The error bars indicate 95% CIs. Squares represent individual studies in the meta-analysis. Diamonds represent the pooled OR.

(0.98MB).

Fig. 5.

Funnel plots of publication biases on the relationship of PPARγ (rs1801282) polymorphism. Funnel plots for publication bias (five genetic models). Each point indicates the individual study included in this meta-analysis. A) Diabetic Nephropathy and Type 2 Diabetes Mellitus B) Type 2 Diabetes Mellitus and Controls.

(0.24MB).

The analysis of rs1801282 polymorphism various T2DM vs. controls revealed moderate heterogeneity in the heterozygote (I2=31%) and high heterogeneity was noticed in allelic (I2=90%), homozygote (I2=51%), dominant (I2=88%) and recessive (I2=45%) genetic models. Based on values of heterogeneity (I2), the random-effect method (DerSimonian and Laird's) was adopted which showed insignificant association with T2DM risk in allelic (j vs J) with (Q test p=0.00001), OR=0.91, (95%CI [0.64–1.30]), p=0.61, homozygote (jj vs JJ) (Q test p=0.05), OR=1.22, (95% CI [0.62–2.41]), p=0.57, dominant (Jj +jj vs JJ) (Q test p=0.00001), OR=0.91, (95% CI [0.64–1.31]), p=0.62 and recessive (jj vs JJ+ Jj) (Q test p=0.08), OR=1.23, (95% CI [0.65–2.30]), p=0.52 genetic models respectively. The fixed effects (Mantel-Haenszel's) model was used, which showed significant association with DN risk in heterozygote (Jj vs JJ) (Q test p=0.18), OR=0.56, (95% CI [0.43–0.74]), p≤0.0001. The results of meta-analysis were illustrated as forest plot (Fig. 4). Further, Begg's funnel plot (Fig. 5) and Egger's regression test were performed which suggested no publication bias in the analyzed genetic models. Since a low number of studies were from Caucasian ethnicity, sub-group analysis was not performed.

Fig. 4.

Forest plots showing individual and pooled ORs (95% CI) of the risk for rs1801282 polymorphism with Type 2 Diabetes Mellitus and controls. The Forest Plots provides an insignificant association between the PPARγ polymorphism with T2DM and controls under five genetic models. The error bars indicate 95% CIs. Solid squares represent individual studies included in the meta-analysis. Solid diamonds represent the pooled OR.

(1.2MB).