Association of Genetic Variants of ELMO1 Gene With Diabetic Nephropathy in the North Indian Population


 Diabetic nephropathy (DN) is a major cause of renal failure globally including chronic kidney disease and end-stage renal disease (ESRD). Using comprehensive linkage disequilibrium mapping, we genotyped five polymorphisms from engulfment and cell motility 1 (ELMO1) gene (rs741301, rs7799004, rs1882080, rs11769038 and rs1345365) to evaluate its association with DN. BMI was observed to be low in DN cases as compared to the control groups, which is the result of haemodialysis and high doses of medication. Physical inactivity, lipid profile, urea and creatinine were observed to be the confounding factors correlated with DN. This study comprehensively evaluated ELMO1 in DN patients, T2D without Nephropathy and healthy controls from North Indian population and revealed significant association with DN. Haplotypes G-G-C-C and G-A-T-T provided ~2-fold risk towards DN development. In conclusion, the present study suggests the significant role of ELMO1 gene polymorphisms in the pathophysiology of DN in North-Indian population.


Introduction
Diabetic nephropathy (DN) is a multifactorial renal disorder triggered by hyperglycaemia-induced damage to kidney in genetically predisposed individuals. It is the major microvascular complication of type 2 diabetes (T2D) leading to end stage renal failure (ESRD) 1 . Furthermore, IDF also reported the 10-fold increased risk of ESRD development in T2D patients, suggesting chronic hyperglycaemia as a leading aetiological factor in DN development 2 . DN is characterized with persistent albuminuria and a progressive decline in the glomerular ltration rate, hence, renal function, reducing the overall quality of life 3 . The increase in risk is linked with hypertension, duration of diabetes and the degree of glycemic control. Environmental and genetic factors must, therefore, play entwined roles in the pathogenesis of DN 4 .
Increased prevalence of DN amongT2D individuals from South Asian populations like India has been reported 5 with the higher incidence of micro-and macro-albuminuria among urban T2D individuals 6 .
This can be attributed to rapid urbanization, demographic evolution, rural-to-urban migration, high fat nutrition and sedentary lifestyle along with genetic predisposition 7 . High body fat percentage is correlated with increased in ammation, insulin resistance and higher risk of diabetes, predisposing Indian population to metabolic derangements 8 . Hence, it has become evident that in ammatory mechanisms contribute signi cantly to the development and progression of DN. These include the lymphocyte and macrophage in ltration of renal compartments with cytokines/chemokines production in the kidney 9,10 . Genetic variants involved in in ammatory pathway have demonstrated strong positive association with the pathogenesis of T2D and DN [11][12][13] .
Genome-wide SNP genotyping analysis on a large cohort of Japanese patients with T2D identi ed engulfment and cell motility 1 gene (ELMO1) as a candidate gene conferring susceptibility to DN 14 . The ELMO1 gene locus 7p14, is a mammalian homologue of the C. elegans gene, ced-12, vital for cell migration and engulfment of apoptotic cells 15 . ELMO1 has also been reported to promote phagocytosis and changes in cell morphology. Additional studies on mouse model of chronic glomerulonephritis demonstrated the increased expression of ELMO1, emphasizing its plausible role in the development of glomerular disorders, including DN 16 . Studies on different ethnic groups like African Americans 17 , American Indians 18 , South Indians 19 , European Americans 20,21 and Chinese 22 have positively replicated ELMO1 association with DN 2 .while others failed to report the same for DN and T2D in Indian as well as other ethnicities [23][24][25] .
The ethnicity speci c distribution of the single nucleotide polymorphism (SNPs) can in uence the disease susceptibility 5 . Due to ethnic disparity in genetic studies, the present case-control association study is the rst study to elucidate the role of selected ve SNPs (rs741301, rs7799004, rs1882080, rs11769038 and rs1345365) of ELMO1 gene in DN pathogenesis among DN cases and controls from North Indian population. In the present study, these ve polymorphisms from ELMO1 gene were selected after reviewing the literature and the information available in public databases such as dbSNP 26 and Haploreg 27 . Furthermore, these variants were reported in the aetiology of T2D, its secondary complications like DN and other metabolic disorders as shown in Fig. 1. Thus, were included in this study to investigate their role in the development of DN in presently studied North Indian population.

Demographic and clinical parameters
A total of 1584 samples were included in this study, comprising of 344 DN cases, 1240 non-diabetic nephropathic controls (NDN) [970 healthy controls which are non-nephropathic and non-T2D (NDNT) and 270 were T2D without Nephropathy disease controls]. Baseline characteristics of the DN cases and control groups are shown in Table 1. It was observed that males were more affected with DN as compared to females. The family history of T2D was observed to be the highest among the T2D without Nephropathy controls. Alcohol consumption and smoking were observed to be highest in T2D without Nephropathy (Data not shown) controls while, BMI was observed to be highest in T2D without Nephropathy controls and lowest in DN cases.   Haplotype Analyses ELMO1 gene haplotype frequencies (rs11769038, rs1882080, rs741301, rs7799004, and rs1345365) for all the compared groups are given in the Table 3. The sequence of haplotypes is in the direction of rs11769038, rs1882080, rs741301, and rs7799004. The polymorphism rs1345365 was not in LD with any of the other polymorphism thus was omitted from the haplotype analysis. The haplotypes G-G-C-C and G-A-T-T provided 2.

Rna Secondary Structure Prediction
Elucidation of RNA structure and the access to correctly annotated RNA structure is of great importance, especially in the predictions of its secondary and 3D structures 28 . Protein secondary structure and initiation codon in the mRNA are known to in uence the translation e ciency 29 . Therefore, in the present study, the RNA secondary structure of the wild allele and mutant allele of the ELMO1 gene polymorphisms with reference allele were analysed. There was a slight reduction in the energy of the wild type allele as compared to the variant type allele as shown in the Fig. 3. The changes in their structure of wild type and mutant are shown in the enclosed circle. There is a slight decrease in the free energy for each polymorphism, thus causing the change in the secondary structure of RNA after folding. This is the rst study that strongly suggests the association of ve polymorphisms of ELMO1 with DN in the North Indian population (Fig. 3). Novelty of this study also lies in the categorization of controls into three groups (NDN, NDNT and T2D without Nephropathy disease controls), which generate better association of confounding factors and genotypes/alleles with DN. ELMO1 is a proin ammatory gene that has been identi ed as a putative candidate for DN pathophysiology in a Genome-wide SNP genotyping study on a Japanese cohort 14 . In the present study, we found signi cant association of rs741301 (T > C) polymorphism with DN in comparison with NDN controls and NDNT controls; and Callele provided 1.5 and 1.6-fold increased risk towards DN, respectively. These results are in concordance with a previous study from North India that has also reported a signi cant association of rs741301 between DN cases and T2D without Nephropathy controls 23  population. This is the rst study to report the association of rs7799004 (T > C) polymorphism in Indian population and can be considered as rst baseline data for future studies in different ethnicities of India. Another SNP rs1882080 (G > A) re ected a signi cant association with DN; genotype AA conferred 2.01fold risk towards DN. The results of present study are similar to a previous study in European population which has also observed a signi cant risk associated with rs1882080 (G > A) polymorphism in diabetic nephropathy 21 . They have also demonstrated that rs1882080 in combination with polymorphism rs11769038 showed a strong risk towards development of DN. In accordance to the previous study, we also observed increased risk provided by AA genotype of rs1882080 towards development of DN.
The current study demonstrated a signi cant association of rs11769038 with DN in all three comparisons and G-allele distribution showed signi cant risk towards DN. The results of present study are in accordance with a previous study on European (Caucasian) T1D subjects in which a strong association of rs11769038(G > T) polymorphism of ELMO1 gene with DN was observed 21 , while it is contradictory to another study which has demonstrated non-signi cant association of rs11769038 polymorphism with DN in Chinese population 22 . The differences with Chinese population can be due to the wide range of difference in ethnic and cultural backgrounds of two populations with the addition of the role of environment in it. In our study, the G-allele of rs1345365 (G > A) variant provided 2 fold risk towards DN aetiology consistent with another study, which has also shown signi cant association for DN under dominant genetic model in African-American population 17 . Intriguingly, a study on Mexicans have reported protective role of this polymorphism emphasizing the notion of allelic heterogeneity 39 . However, other studies on American Indians and Chinese population did not report any association with rs1345365 (G > A) polymorphism of ELMO1 gene 18,22 .
The con ict in results of various studies can be attributed to unknown aetiology of DN, multiple genetic factors, environmental factors, ethnicity differences, sample size variation and selection of control groups. T2D represents an intermediate disease stage between healthy controls and overt DN cases. The present association suggest the role of prolonged hyperglycaemia in causation of nephropathy and highlights the importance of considering T2D cases as internal controls in assessment of genetic architecture of DN.
Previously, there were fewer studies available on the role of ELMO1gene polymorphisms in T2D but the DN data suggested that ELMO1 may have an imperative role in the development of nephropathy in T2D patients, which might be contributing to renal function decline 21 . Later, different studies from different populations con rmed that ELMO1 variants are primarily associated with the risk of proteinuria. Several functional studies have even demonstrated that ELMO1 contributes to the progression of chronic glomerular injury through its dysregulation of extracellular matrix (ECM) metabolism, resulting in renal ECM accumulation 16 . This accumulation contributes to both glomerular and tubular basement membrane thickening, which are the main hallmarks of advanced DN 40 . Other studies have also reported the association of intronic variants in ELMO1 gene but the variants in their study are different from those reported in the present study except for one polymorphism (rs1345365) 14,17 . These associations at ELMO1 across each study represent allelic heterogeneity contributed by diverse ancestral genetic backgrounds of the different ethnic groups 21 . The ndings of the present study suggest that role of ELMO1 gene may be implicated in the development of DN.
In imputation of SNPs rst studied polymorphisms and then NCBI_GIH genotypes were phased independently and then the two types of datasets were merged and were phased again. This phased data was used as the reference for the imputations. Concordance rate for each SNP was measured to check the accuracy of imputation. Numbers of SNPs involved in imputation ranged from 1-3 for ve SNPs.
Imputation has revealed that the highest concordance rates were observed for rs741301 (89.6%) and rs7799004 (86.1%), while rest were observed to be below 70%, whereas did not show any results for rs1345365. The average concordance rate of the SNPs was observed to be 76.4% (Supplementary Table  S5). Accuracy of imputation is known to increase with the increase in reference population size and also by including the familial genotype data in the reference population 41 . The sample size of the reference population is not large, but the numbers of SNPs were good enough to make a conclusion that the accuracy of the present data is of good quality.

ELMO1, haplotypes G-G-C-C and G-A-T-T conferred risk towards the development of DN, in DN vs. NDN controls, while the haplotype T-A-T-T and G-G-T-T conferred risk in DN cases vs. T2D without Nephropathy
controls, but provided protection in DN cases vs. NDNT controls. The variants rs741301 and rs7799004 (D'=0.99, r 2 = 0.97), rs741301 and rs1882080 (D'=0.97, r 2 = 0.92), and rs7799004 and rs1882080 (D'=0.97, r 2 = 0.92) were observed to be in strong LD in DN cases. A study in African-American population has also observed that rs741301 and rs7799004 were in strong LD (D'=1.0) in their population 18 . In another study, rs1345365 was observed to be in LD with other two SNPs-rs1981740 and rs10951509, not included in this study. A study in Chinese population and Iranian population, found that polymorphisms-rs741301 and rs1345365 were in weak LD (D'=0.01 and D'=0.11, respectively) with each other 22,36 . But, rs1345365 was observed in strong LD with rs11769038 (D'=0.91) in Chinese population 22 . The polymorphism rs1345365 was not observed to be in LD with any of the studied polymorphisms in the present study. The major problem in comparison with other studies is that the markers selected for analysis are not same among all the previously studied populations which lead to the divergence of the results in different studies. The con ict in results in various studies re ect that association of different haplotype blocks and LD patterns with the disease risk may vary among populations due to diverse genetic backgrounds, complex aetiology of the disease, variation in sample sizes and heterogeneity in samples; and selection of different polymorphisms/markers. Also LD is the non-random distribution of alleles in the general population, which may be the cause that rs7799004 and rs11769038 were not in LD in case of NDNT (healthy) controls while showing mild LD in DN cases and T2D without nephropathy controls. Thus, indicating that rs7799004 and rs11769038 are following normal distribution in healthy individuals and were showing mild LD in DN cases as well as T2D without nephropathy controls.
The limitation of the present study is the functional relevance of ELMO1. To overcome this limitation, prediction of the secondary RNA structures of studied polymorphisms using bioinformatics approach, supported the risk causal role of ELMO1 in DN pathophysiology. Conclusion ELMO1 gene polymorphisms are signi cantly associated with DN thus intimating its pivotal role in DN pathogenesis in North Indian population along with the elevated diabetogenic risk factors and kidney function markers-urea and creatinine. Gender based differences were reported, predisposing males towards the susceptibility of DN. The haplotypes G-G-C-C and G-A-T-T are strongly associated with DN development. The study also highlights the importance of ethnicity and recruitment of appropriate control groups in genetic association studies. However, further studies are warranted to validate the functional aspect of this gene and its comprehensive mechanism in the aetiology of DN.

Research Design And Methods
Study Participants. The present study included total of 1584 subjects from North India, comprising of age-matched 344 DN cases, 1240 non-diabetic nephropathic controls (NDN) [970 healthy controls which are non-nephropathic and non-T2D (NDNT) and 270 were T2D without Nephropathy disease controls]. This study was approved by the Ethics Committee of Guru Nanak Dev University, Amritsar (Letter No.-229/HG). Informed consent was obtained from all participants and/or their legal guardians and study has been carried out in accordance with the Declaration of Helsinki. prescribed with oral hypoglycaemic agents, or insulin, or both. Blood pressure was measured after 20 minutes of rest Omron (HEM-711 model) digital machine. Various demographic and anthropometric parameters like age, gender, family history of T2D & DN, alcohol consumption, smoking and BMI were noted for DN cases and three control groups. BMI classi cation was made according to cut off values for Asian Indian adults (23 kg/m 2 ) given by Snehalatha et al, (2003) 42 . Written informed consent was taken from all individuals before participation in the study.
DNA extraction and Clinical parameters. The genomic DNA was extracted from the blood samples using inorganic extraction 43 . Quanti cation and quality estimation of genomic DNA was done by UV spectrophotometer (Eppendorf Biospectrometer-Basic) and agarose gel electrophoresis. Blood samples were analysed for random blood sugar using glucometer (Accu-Check Active). Levels of urea, creatinine, total cholesterol (TC), triglycerides (TG), and high-density lipoprotein cholesterol (HDL-C) were measured from plasma by using commercial kits (Erba Chem 7 analyzer). Very low density lipoprotein (VLDL) and low density lipoprotein (LDL) were calculated by using the formulae given by Friedewald et al. in 1972 44 .
SNP selection and Genotyping. Genotyping was performed using high throughput Sequenom iPLEX Gold for Sequenom MassARRAY platform.
Statistical analyses. Baseline parameters were compared using unpaired student's t-test. Hardy Weinberg equilibrium was checked for all the SNPs, at rst six polymorphisms were included in the study but of the six polymorphisms, one (rs7785934) was not following HWE thus was excluded from further analyses.
Chi-square test was performed to compare the distribution of genotype/allele and genotype models of ELMO1 SNPs between DN cases and controls (IBM SPSS Inc., version 20.0; Chicago, IL, USA). The extent of association between ELMO1 SNPs was determined using Odds ratio (OR) at 95% con dence interval (CI) with p-value < 0.05. Haplotype and Linkage disequilibrium (LD) analysis was performed using Haploview (version 4.2) 45 . One of the polymorphism (rs1345365) was not in LD and thus was excluded from the analyses. Power of the present study was > 80% calculated using PS software (version 3.0) 46 .
The imputation of the selected markers was performed using Plink v1.07 47 , with the reference data downloaded from NCBI as NCBI_GIH with 2737 genetic markers of ELMO1 gene in 103 individuals. Locations of ve single nucleotide polymorphisms were identi ed from dbSNP browser and then the SNP data was merged with GIH data obtained from NCBI. RNA structures and their free energies were obtained using RNA fold software available online (http://rna.tbi.univie.ac.at/cgi-bin/RNAWebSuite/RNAfold.cgi).