A total of 200 type 2 diabetic patients who attended the specialized DM and nephrology clinics at the Internal Medicine Department of Beni-Suef University Hospital, (Beni-Suef, Egypt) from November 2021-May 2022 provided written informed consent for participation in the present study and agreed to the use of their samples in scientific research. Eligible patients were classified according to the value of their eGFR into six groups. Furthermore, 40 healthy subjects were included in the study as controls. The study protocol followed the Declaration of Helsinki and good clinical practice guidelines and was approved by the Ethics Committee of Beni-Suef University Hospital (approval no. BSU:7-2021; Beni-Suef, Egypt).

Healthy individuals, diabetic patients without nephropathy [glycated hemoglobin (HbA1c) >6.5%; eGFR ≥60 ml/min/1.73 m²)], and diabetic patients with nephropathy [HbA1c >6.5%; eGFR <60 ml/min/1.73 m²)] were enrolled in the present study. The present study excluded patients with any history of chronic and acute infections, diabetic retinopathy, hepatic diseases, malignancy and other endocrine dysfunctions. According to the clinical data and eGFR values, enrolled participants were classified into seven groups as follows: i) Healthy controls (n=40); ii) G1, eGFR ≥90 ml/min/1.73 m² (n=35); iii) G2, eGFR 60-89 ml/min/1.73 m² (n=30); iv) G3a, eGFR 45-59 ml/min/1.73 m² (n=30); v) G3b, eGFR 30-44 ml/min/1.73 m² (n=35); vi) G4, eGFR 15-29 ml/min/1.73 m² (n=30); and vii) G5, eGFR <15 ml/min/1.73 m² (n=40). The G5 group of patients was classed as being in kidney failure. The sample size for the present study was determined based on several factors, including the desired level of confidence, expected effect sizes of the biomarkers, variability within the population and statistical power considerations. Specifically, the aim was to achieve a power of 80% to detect statistically significant differences in biomarker expression levels between diabetic patients with and without complications, as well as among different severity groups of diabetic nephropathies. With each group comprising 30-40 participants and considering an effect size of 0.5, the sample size was deemed sufficient for detecting medium to large effect sizes.

A total of two blood samples (4 ml/sample) were collected from healthy controls and diabetic participants after overnight fasting. EDTA was used in the collection of one of the blood samples and the other was collected using a plain collection tube. Samples were incubated for 30 min at room temperature and blood in the plain tubes was centrifuged at 4,000 x g at 4˚C for 20 min to isolate serum. Blood samples in EDTA were used for complete blood count, DNA extraction and HbA1c level measurements. Samples were stored at -80˚C until used.

HbA1c was measured using a Stanbio™ Glycohemoglobin (HBA1 and HBA1C) Pre-Fil^™ Test (cat. no. SB-P350-50; Stanbio), while fasting blood sugar (FBS), Na⁺, Ca²⁺ and K⁺ levels were determined in serum samples using a commercial SPINREACT diagnostic kit (cat. nos. 1001380, MD1001065 and #1001390 for Na⁺, Ca²⁺ and K⁺, respectively; Spinreact, S.A.U.). Blood uric acid was measured using a commercial SPINREACT diagnostic kit (cat. no. MD41001; Spinreact, S.A.U.). Serum activity of glutamate pyruvate transaminase (sGPT) and glutamate oxaloacetate transaminase (sGOT) enzymes were measured using a commercial SPINREACT GOT/AST diagnostic kit (cat. no. MD41264; Spinreact, S.A.U.) according to the manufacturer's instructions. Blood urea (BUN reagent; cat. no. BK-443350D) and creatinine uric acid levels (creatine reagent kit; cat. no. BK-472525D) were determined in the serum samples (Diamond Diagnostics). Fasting insulin levels were assayed using Diagnostic Products Corporation radioimmunoassay kits (Coat-A-Count; cat. no. TKIN-4; Diagnostic Products Corporation). Insulin resistance was investigated by calculating the HOMA-IR. HOMA-IR=[(fasting insulin, µU/ml) x (fasting glucose, mmol/l)]/22.5, where 22.5 is the normalizing factor.

The glomerular filtration rate (eGFR) was measured based on serum creatinine levels and calculated according to the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation. Moreover, the CKD-EPI value was calculated taking into account sex and serum creatinine (39). Females: Creatinine ≤0.7 mg/dl: eGFR=144x (creatinine/0.7)^-0.329 x (0.993) age x 1.159 (for Black or African American individuals). Creatinine >0.7 mg/dl: eGFR=144 x (creatinine/0.7)^-1.209 x (0.993) age x 1.159 (for Black or African American individuals). Males: Creatinine ≤0.9 mg/dl: eGFR=141 x (creatinine/0.9)-0.411 x (0.993) age x 1.159 (for Black or African American individuals). Creatinine >0.9 mg/dl: eGFR=141 x (creatinine/0.9)-1.209 x (0.993) age x 1.159 (for Black or African American individuals). Adjustments for Black or African American individuals were made in eGFR calculations to account for higher average serum creatinine levels in Black individuals for more accurate kidney function estimates (40). The calculation used for the rest of the individuals studied in this study who were not of Black/African American ethnicity was based on the following CKD-EPI creatinine-cystatin equation (41) eGFR=135 x min (SCr/κ,1)^α x max (SCr/κ,1)^-0.544 x min (Scys/0.8,1)^-0.323 x max (Scys/0.8,1)^-0.778 x 0.9961^Age x 0.963 [if female], where eGFR=estimated GFR in ml/min/1.73 m², SCr=standardized serum creatinine in mg/dl, Scys=standardized serum cystatin C in mg/l, κ=0.7 (females) or 0.9 (males), α=-0.219 (females) or -0.144 (males), min=indicates the minimum of SCr/κ or 1, max=indicates the maximum of SCr/κ or 1, age=years. ELISA kits were used to measure serum TNF-α (cat. no. SEA133Mu; Cloud-Clone Corp.), cardiac troponin I (Human Cardiac Troponin IELISA Kit; cat. no. ab200016; Abcam) and Nrf2 levels (cat no. #80593-1-PBS; Wuhan Fine Biological Technology Co., Ltd.).

To investigate hyperglycemia-induced cardiac injury, circulating creatine kinase (CK) activity, CK-myocardial band (MB) activity, lactate dehydrogenase (LDH) activity and troponin I levels were measured. Manufacturers' instructions were followed to determine the ELISA CK-MB (Rat CK-MB ELISA kit; cat. no. DEIA-FN285; Creative Diagnostics), troponin I (cat no. ELH-Troponin1-1; RayBiotech, Inc.) and LDH (Rat LDH kit; cat. no. LS-F5026; LifeSpan Biosciences, Inc.).

miRNA was extracted from serum samples using the Plus kit (cat. no. R2072; Zymo Research Corp.). Isolated RNA samples were reverse-transcribed, then miR-133a, miR-132 and lnc-MGC expression levels were determined via reverse-transcription quantitative PCR (RT-qPCR) using the One-Step RT-PCR kit (cat. no. 12594100; Thermo Fisher Scientific, Inc.). The cycling conditions for RT-qPCR started with initial HotStar Taq DNA Polymerase activation step at 95˚C for 15 min, then 40 cycles each of three steps (94˚C for 15 sec, 55˚C for 30 sec, and 70˚C for 30 sec), and then the dissociation curve stage was added to verify specificity of the PCR products. The primer sequences for GAPDH were obtained from OriGene Technologies, Inc. (cat. no. HP205798) and were used as internal reference controls. The primer sequences for miR-132, miR-133a and lnc-MGC were designed using the National Center for Biotechnology Information Primer-BLAST tool (https://www.ncbi.nlm.nih.gov/tools/primer-blast/). The primer sequences used were as follows: miR-123 (accession no. NC_000017.11) forward (F), 5'-CGACCATGGCTGTAGACTGT-3' and reverse (R), 5'-GTCTCCAGGGCAACCGTG-3'; miR-133a (accession no. NC_000018.10) F, 5'-TTTGGTCCCCTTCAACC-3' and R, 5'-GAACATGTCTGCGTATCTCA-3'; lnc-MGC (accession no. MW802745) F, 5'-GCTACAGCTGGTTGAAGGG-3' and R, 5'-TGCTTTGCTAGAGCTGGTAAAATG-3'; small nucleolar RNA C/D box 68 (SNORD68; accession no. NC_000016.10) F, 5'-CGTGATGACATTCTCCGGAATC-3' and R, 5'-AATCAGATGGAAAAGGGTTCAAATG-3; and GAPDH (accession no. NM_002046) F, 5'-GTCTCCTCTGACTTCAACAGCGC-3' and R, 5'-ACCACCCTGTTGCTGTAGCCAA-3'. SNORD68 was used as an endogenous housekeeping gene for miR-132 and miR-133a as its expression remains stable and does not vary under different experimental conditions or in different states of the same sample (for example, ‘disease’ vs. ‘normal’ samples) (42-44). The relative quantities of each target gene were measured and standardized against the specified internal control according to the 2^-∆∆Cq method (45).

Statistical analysis was conducted using SPSS (version 16; SPSS, Inc.). Data were presented as mean ± SEM (46) and statistical comparisons were carried out using a one-way analysis of variance (ANOVA) with Tukey's post hoc test. P<0.05 was considered to indicate a statistically significant difference. Multiple testing corrections were carried out using the false discovery rate (Table SI). Receiver operating characteristic (ROC) curves were generated by plotting sensitivity against 1-specificity at different cut-off values. The diagnostic accuracy for each cut-off point was assessed using the area under the ROC curve (AUC). A performance level of ≥50% was deemed acceptable.

Based on the clinical characteristics, the study participants were grouped into a healthy control group and six patient groups (Table I). The age range of the study participants was 41-65 years. The mean age of the healthy group (43 years) was significantly lower compared with that of the patient groups (mean range, 50-63 years; P=0.001). The sex distribution of the individuals in the healthy and patient groups was not significantly different. Patient groups demonstrated significantly higher obesity indices as indicated by BMI, compared with the body weights recorded in the healthy group (P=0.001). Furthermore, diabetes-related parameters such as systolic blood pressure (SBP), diastolic blood pressure (DBP), HbA1c, FBS and the activity levels of sGPT and sGOT enzymes were significantly higher in the patient groups compared with the healthy controls (P<0.001), particularly in patient groups G3-G5. According to the present results, both sGPT and sGOT enzyme level in healthy and all diabetic groups were in the range from 21-27 U/l, which is within the normal range of sGPT (7-56 U/l) sGOT (5-40 U/l) (https://www.medicinenet.com/liver_blood_tests/article.htm).

High blood pressure was observed in the patients of G1-G5 with values of 124-146 mmHg for SBP and 82-90 mmHg for DBP. Moreover, there were significant increases in both FBS and HbA1c levels in the diabetic groups compared with the healthy group.

Kidney function-related parameters including urine, creatinine, urea, uric acid and mineral (Na⁺, K⁺ and Ca²⁺) levels in serum, as well as GFR, were measured (Table II). There was a significant increase in the concentrations of urea, uric acid, and creatinine in the patient urine in diabetic groups (G1-G5) compared with the healthy group (P=0.001). These increases were more pronounced in the diabetic groups G3b-5, where the concentration of creatinine increased by 78, 233 and 711% (P=0.001), respectively, compared with the healthy group. In G2 diabetic group, there was also a significant increase in serum Na⁺, also known as hypernatremia, of 5.8% (P=0.001), compared with the healthy group. This indicated the presence of a common type of electrolyte abnormality due to osmotic diuresis-induced hypotonic losses increasing serum Na⁺ levels (47). Similar to the observed increase in Na⁺ levels, the level of serum K⁺ was increased, and these increases were positively correlated with the increase in diabetic indices. There was a steady increase in the concentration across different diabetic groups, recording high increased values for the G5 group (29.5%) for K⁺ levels compared with the healthy group. On the other hand, Ca²⁺ levels and GFR, a sensitive indicator of kidney function (a low GFR value indicates that the kidneys are not functioning properly), were significantly decreased (P=0.001) compared with healthy controls. The G5 group demonstrated a lower level of Ca²⁺ (7.9±0.14) compared with the normal serum Ca²⁺ level of the healthy control group (9.3±0.07) and the other diabetic groups (G1-G4) (8.9-9.1 mg/dl). In contrast to the G1 group, there was a significant decrease in GFR across the other diabetic groups and the lowest value was recorded in the G5 group (7.9 ml/min/1.73 m²).

To investigate hyperglycemia-induced cardiac injury, the circulating LDH and troponin levels and CK and CK-MB enzyme activity were measured (Fig. 1). Compared with the healthy control group, the diabetic patients, mainly G5, demonstrated a marked increase in the majority of measured cardiac-related parameters. The mean serum LDH cholesterol (P<0.0001) and troponin (P<0.01) levels of diabetic patients in the G5 group, were higher than those compared with the healthy control group, respectively (Fig. 1A and B). Overall, a positive correlation was demonstrated between LDH and HbA1C levels (r=0.48; Table III). The mean serum CK and CK-MB levels were also significantly higher in all diabetic groups (G5) compared with the healthy control group (P<0.0001 and P<0.001, respectively) (Fig. 1C and D). For all the aforementioned parameters, there was an increase demonstrated across the diabetic groups, with the highest levels of CK and CK-MB in the serum of diabetic patients in the G4 and G5 groups. Overall, patients with increased levels of LDH, troponin, CK and CK-MB were potentially associated with the risk of heart failure in the diabetic groups.

TNF-α and Nrf2 levels were measured as they are cytokine biomarkers of inflammation (Fig. 2). These results demonstrated that the levels of TNF-α were significantly increased in G2-G5 groups compared with the healthy controls, with the G3a, G4 and G5 groups demonstrating the highest levels of TNF-α. Nrf2 levels were significantly reduced in all diabetic groups compared with the control group, with the lowest values demonstrated in the G4 and G5 groups (-66.6 and -73.33%, respectively).

To investigate the molecular biomarkers for CVD, expression levels of the ncRNAs miR-132, miR-133a and lnc-MGC were measured using quantitative RT-qPCR (Fig. 3A-C). Lower expression levels of miR-132 in blood were demonstrated in diabetic patients compared with healthy controls (Fig. 3A). Serum miR-132 expression was steadily decreased across diabetic groups and the lowest expression levels were demonstrated in the G4 and G5 groups of patients (-83.33 and -91.67%, respectively). Compared with the healthy controls, the expression levels of circulating miR-133a were significantly higher in diabetic patients of G2-G5 who showed impaired cardiovascular function (increased LDH and troponin I levels and CK and CK-MB enzyme activity) (Fig. 3B). The expression levels of serum miR-133a in diabetic patients steadily increased across diabetic groups and reached the highest expression level in patients in the G5 group (Fig. 3B). There was an increase in the expression levels of lnc-MGC in diabetic patients compared with healthy controls, with the maximum lnc-MGC expression level recorded in the G5 group (729.29%) (Fig. 3C). The expression level of miR-133a (r=0.326; P<0.01) was negatively correlated with the levels of LDH and HbA1C (Table III). However, serum miR-132 was positively correlated with HbA1C (r=0.382; P<0.01).

Given the upregulated expression levels of miR-132, miR-133a and lnc-MGC in diabetic patients, particularly those of G3-G5, their role as potential prognostic markers was investigated using ROC curves. Serum miR-132 was differentially expressed in participants with DN compared with healthy controls, with a 93.20% sensitivity and a 100.00% specificity (AUC=0.93; 95% CI=0.886-0.95; P<0.001; Fig. 4A). Similarly, serum miR-133a expression levels also differentiated the patients of G2-G5 with DN from healthy controls, with 100.00% sensitivity and 100.00% specificity (AUC=0.998; 95% CI=0.993-1.01; P<0.001; Fig. 4B), thus suggesting its potential utility as a biomarker of CVD. Lnc-MGC was differentially expressed in participants with DN compared with healthy controls, with a 97.20% sensitivity and a 100.00% specificity (AUC=0.951; 95% CI=0.891-0.954; P<0.001; Fig. 4C). In addition, serum miR-132 expression levels were shown to differentiate the of G3b, G4 and G5 stages of DN from the G2 and G3a stages of DN with 43.90% sensitivity and a 94.80% specificity (AUC=0.74; P<0.001; Fig. 5A). Similarly, serum miR-133a and lnc-MGC expression level also showed similar results (Fig. 5B-C).