Study design: This was an age and sex matched case control study comprising of total 100 participants with T2DM. The participants were enrolled into the study while attending medicine/nephrology outpatient clinic of University College of Medical Sciences and Guru Teg Bahadur Hospital between October 2010 and April 2012. They were divided into two groups of 50 each viz., Group I: patients with Type 2 diabetes mellitus without vascular complications (T2DM), Group II: patients with Type 2 diabetes mellitus with nephropathy (DN, Stage 2 and Stage 3). According to American Diabetes Association (ADA), Stages of Chronic Kidney Disease (CKD) with respect to Glomerular Filtration Rate (GFR) (mL/min/1.73 m² body surface area): (a) Stage 1: kidney damage with normal or increased GFR ≥90; (b) Stage 2: kidney damage with mildly decreased GFR; 60–89; (c) Stage 3: moderately decreased GFR; 30–59; (d) Stage 4: severely decreased GFR; 15–29; (e) Stage 5: kidney failure with GFR; ≤15 as per Kidney Disease: Improving Global Outcomes, KDIGO [20]. The plan of this study was approved by Institutional Ethics Committee for Human Research. Informed and written consent were taken from all the participants earlier to the inclusion into the study. Keeping 80% probability of detecting an effect using a two-tailed test with conventional levels of alpha (0.05), a quick calculation reveals that sample size will be at least 50.

Participant’s selection: The participants underwent health examination which included a detailed health questionnaire, physical examination with anthropometric measurements and necessary laboratory tests. Diabetes was diagnosed on the basis of revised ADA criteria, 2015 [21]. Nephropathy in diabetic patients was identified by qualitative analysis in early morning random midstream urine sample by dipstick method (Urine Test 11 MAU, Piramal Diagnostic, India). DN is a generic term referring to any deleterious effect on kidney structure and/or function caused by diabetes mellitus, the first being characterized by microalbuminuria and later by decreased GFR. Urinary albumin to creatinine ratio (μg/mg creatinine): (a) less than 30 is considered normal; (b) 30-300 is considered microalbuminuria; and (c) greater than 300 is considered as macroalbuminuria (overt proteinuria) [22]. It was further confirmed by measuring Urinary Albumin: Creatnine Ratio (UACR). Subjects having T2DM without any vascular complications served as a control and subjects having T2DM with nephropathy was designated as cases. All diabetic subjects with retinopathy and dipstick positive proteinuria and microalbuminuria were clubbed in Group II; however, all the patients were in pre-dialysis stage.

To avoid potential confounding factors, the patients having acute and chronic infections, fever, malignancy, other renal disorders, cirrhosis of liver and congestive heart failure, alcoholism and smoking were excluded. All the patients in Group II had retinopathy however patients with macrovascular complications like stroke and CAD were excluded. The dietary sources rich in purine like meat, seafood and sprouts were restricted from the diet of subjects for a week before the enrollment. Patients who were on inhibitors of renin– angiotensin aldosterone system, aspirin and vitamin D analogues were advised to stop these drugs for one week before inclusion in the study because these are reported to affect oxidative stress and inflammatory markers. Patients using UA lowering drugs or diuretics were also excluded.

On inclusion into the study arterial blood pressure was measured three times on the left arm in sitting position using a clinically validated automatic device (Omron, HEM-907, Matsusaka, Japan) after applying the appropriate cuff size and after a period of 10 minutes rest with the subject in the sitting position, the average of the all three values was used for analysis. Body Mass Index (BMI) was calculated as weight (in kilogram) divided by height (in meter square). Blood samples were collected in the morning after the participants had been fasting for at least 8 hours-12 hours. Blood was collected into vacutainers coated with salts of fluoride and oxalate for plasma glucose, EDTA vials for various other biochemical parameters. Blood sample collected in EDTA vials was subjected to centrifugation at 3000 rpm for 10 minute to separate the plasma at room temperature. For estimation of TNF-α and hsCRP, plasma was stored in aliquots at −80°C. Urine samples for evaluation of uMCP-1 were also stored at −80°C till further estimation. All these investigations were carried out once at the time of entry into the study.

Plasma glutathione-S-transferase; GST [23], antioxidant capacity of blood measured as Ferric Reducing Ability of Plasma; FRAP [24], malondialdehyde; MDA [25] and reduced blood glutathione; GSH [26] were estimated immediately after collection. Plasma TNF-α (Diaclone, France), hsCRP (Calbiotech, USA) and uMCP-1 (Weldon, California) were measured by sandwich enzyme linked immunosorbent assay. The sensitivity for TNF-α, hsCRP, uMCP-1 was less than 8 pg/mL, 0.005 mg/mL and 7.8125 pg/Ml respectively. Glycosylated haemoglobin (HbA1c) was estimated by ion-exchange resin chromatography using commercially available kits (Fortress Diagnostics, UK). The detection limit was < than 4.3%.

Fasting and postprandial glucose, urea, creatinine, UA, total cholesterol, TAG, HDL, LDL, VLDL as routine biochemical parameters were assayed in automated analyser (Olympus AU 400) using commercial kits.

Continuous variables were reported as mean and Standard Deviation (SD). Demographic and clinical data between two groups were analysed by Student’s t-test. Relationship between studied parameters was determined by Pearson correlation analysis. We examined the association of UA (independent variable) levels with oxidative stress parameters and inflammatory markers as dependent variables using linear regression analysis. Since PUA concentration depends on various co-variables therefore analysis was conducted without adjustment (Model 1) followed by analysis with age, sex adjustment (Model 2) and further adjustment which include age, sex along with BMI, arterial blood pressure, HbA1c, total cholesterol (Model 3) and p<0.05 was considered significant.

The demographic and clinical characteristics of the study population are depicted in [Table/Fig-1]. The participant were age (p=0.686) and sex matched (p=0.974). No significant difference was observed with respect to blood pressure i.e., Systolic Blood Pressure (SBP, p=0.077) and Diastolic Blood Pressure (DBP) (p= 0.247) and BMI (p=0.336) between both the study groups. PUA was significantly higher in Group II subjects (8.55±0.7, p<0.001) than in patients of Group I (3.40±0.64). Plasma urea and creatinine levels were higher in Group II when compared to Group I. Fasting, postprandial plasma glucose and HbA1c were significantly higher (p< 0.001) in Group II than Group I suggestive of poorer glycaemic control in later Group. Lipid profile depicted by total cholesterol, TAG, LDL-C, VLDL-C, total cholesterol to HDL-C ratio and LDL-C to HDL-C ratio were significantly higher in Group II (p<0.001) patients as compared to Group I. However, HDL-C levels were lower in Group II patients as compared to patients with Group I. Estimated Glomerular Filtration Rate (eGFR) was significantly lower in patients of Group II (p<0.001) as compared to Group I.

The comparison of mean levels of estimated oxidative stress and inflammatory markers is presented in [Table/Fig-2]. Whole blood reduced GSH and plasma FRAP as antioxidant markers are significantly reduced in Group II patients in comparison to patients of Group I. In contrast, plasma MDA levels were significantly raised in Group II as compared to Group I patients. However, plasma GST levels showed a different behaviour, it was significantly raised in Group I as compared to Group II. Mean plasma levels of TNF-α, hsCRP and uMCP-1 as markers of inflammation were significantly raised in Group II patients when compared to Group I subjects. [Table/Fig-3,4] describes the Pearson correlations of different oxidative stress markers and inflammatory cytokines with PUA in overall study population and individual study groups. With the exception of MDA which had an inverse association with uric acid, all oxidative stress markers showed a highly significant (p<0.001) negative correlation with UA. On studying the association of inflammatory markers with UA, it was observed that all markers of inflammation i.e., TNF-α (p<0.001, r = 0.552), hsCRP (p<0.001, r = 0.815), uMCP-1 (p<0.001, r = 0.811) showed a significant positive correlation with UA. These parameters when studied in individual study group, it was noticed that they follow the similar trend but lost their significance.

In the linear regression Model unadjusted for covariates, plasma UA expressed as a continuous variable was found to have a significantly negative association with levels of GST and FRAP ([Table/Fig-5], Model 1). However PUA was found to have negatively correlated with GSH and positively with MDA but were not statistically significant (p=0.935, p=0.984). After adjustment for the first set of covariates, the beta coefficients for PUA were slightly increased and they remained statistically significant except for GSH and MDA ([Table/Fig-5], Model 2). After further adjustment, the results were virtually similar ([Table/Fig-5], Model 3).

While exploring the relationship between PUA and markers of inflammation, linear unadjusted models, UA was a significant positive predictor of TNF-α, hs-CRP except uMCP-1 ([Table/Fig-6], Model 1). On adjustment for age and sex, the beta coefficients for PUA were slightly reduced and they remained statistically significant ([Table/Fig-6], Model 2). After additional adjustment for (BMI, Total cholesterol, Blood pressure) the association remains unchanged ([Table/Fig-6], Model 3).

The present case-control study was designed to find out the role of UA in the pathogenesis of DN. The main strength of this study is that we address the association of PUA levels with oxidative stress and inflammatory markers taking into consideration of all confounding factors under one umbrella. Data of the present study suggests that the levels of antioxidant were further reduced and pro-oxidant were raised in patients with DN in comparison to patients with T2DM without nephropathy, which was reported in our previous study [11]. Our results were supported by a study [27] which reported that there was a significant decrease in reduced GSH and increase in MDA levels in diabetic patients with micro-vascular complications. Our results were further confirmed by a related recent study [28] that showed that levels of MDA were increased and FRAP levels were decreased in patients with DN. However, GST behaved in a subtle dissimilar manner. Various studies have reported ambiguity in activity of GST in response to Oxidative Stress (OS) in diabetic patients [29,30]. Levels of GST is anticipated to increase with increased OS, however it seems that heightened levels of GST in T2DM is a compensatory mechanisms of antioxidants against OS in humans, which protects the progression from DM to DN. The study conducted by Yang Y et al., has reported that higher expression of GSTs in the endothelium shield against oxidative damage from aldehydes such as 4-hydroxynonenal (4-HNE) [31]. It has been reported in the previous study that there is inhibition of RBC-GST activity by uremic plasma [32]. Hence, there is a chance that the presence of uremic toxins may contribute to reduced GST activity in patients with DN most certainly by interfering with its induction.

To our best knowledge, this is the first case control study to assess the relationship between PUA and various oxidative stress markers. In the current case control study, as depicted in [Table/Fig-3] we found a significant negative association of PUA with anti-oxidant parameters i.e., GSH, FRAP, GST and significantly positive correlation pro-oxidant i.e., MDA. Indeed, GST and FRAP remained significantly negative correlated with PUA even after adjustment for important potential confounders. These result supported the idea that UA may be involved in the progression of micro-vascular complications in diabetes mellitus through production of ROS. Oxidative stress plays a key role in steatosis induced by UA. Cellular exposure to increased UA leads to mitochondrial oxidative stress with generation of ROS by Nicotinamide Adenine Dinucleotide Phosphate (NADPH) oxidase. As a result of which the activity of aconitase, an enzyme involved in the kreb’s cycle was inhibited leading to accumulation of citrate. Since citrate is a substrate of de novo lipogenesis and therefore increased fat generation, hence more lipid peroxidation. Furthermore, ROS production increases endoplasmic reticulum stress, which is a determinant of fat accumulation [33]. Thus, elevated PUA contributes to oxidative stress within the glomerulus and tubulo-interstitium leading to endothelial dysfunction, thrombogenesis and remodeling fibrosis of kidney [34]. It has been reported in a previous study that UA activates MAPK pathway via oxidative stress, causes preglomerular arteriolar damage [35]. Correlation of PUA with FRAP is not in accordance with study done by Lyngdoh T et al., [36]. However, till date there is no study to support our findings of association of PUA with GSH, GST and MDA.

A recent study has reported that elevated UA levels upregulate Notch-1 expression and consequently leading to inflammatory reaction via Notch signalling pathway [37]. It has been established that PUA when entering the vascular smooth muscle cell stimulates the release of CRP and chemokine MCP-1. UA also stimulates human mononuclear cell to produce TNF-α, IL-1, IL-6. The finding from the present study also showed that PUA is positively associated with markers of inflammation viz., TNF-α, uMCP-1, hsCRP. This positive significant association was still maintained after adjustment for important confounding factors. Thus, these finding supports the hypothesis that UA is involved in inflammation by triggering the release of inflammatory mediators. Few studies have investigated the relationship between PUA and systemic inflammation. Our results were in accordance to the study conducted in patients of congestive heart failure by Leyva F et al., [12]. A study conducted on healthy people demonstrated that UA was associated positively with CRP [37]. Serum UA was positively associated with TNF-α, CRP in elderly people in Italy [37] whereas, UA levels was not associated with TNF-α and CRP in elderly men in Taiwan [3].

The potential limitation of this study is small sample size and results need to be validating in larger sample size. Other limitation is that genetic factors influencing the levels of UA has not been considered in our study which may have an impact on the relationship between UA and progression of DN. Further experimental or interventional studies are needed to explore the relationship between UA and DN.

In conclusion, elevated PUA levels might contribute to the vicious cycle of inflammatory responses (chemokines) and oxidative stress (production of ROS). Our findings propose that PUA levels may be an important risk factor for further progression of renal dysfunction. Thus, our study provides new evidence that lowering the UA levels may represent a new therapeutic approach in the treatment of DN.