Circulating hsa-miR-221 as a possible diagnostic and prognostic biomarker of diabetic nephropathy

Background Diabetic nephropathy (DN), which is a chronic outcome of diabetes mellitus (DM), usually progresses to end-stage renal disease (ESRD). The DN pathophysiology, nevertheless, is not well-defined. Several miRNAs were reported to be either risk or protective factors in DN. Methods, and results The present study sought to inspect the potential diagnostic and prognostic value of hsa-miR-221 in DN. The study included 200 participants divided into four groups: Group 1 (50 patients with DN), Group 2 (50 diabetic patients without nephropathy), Group 3 (50 nondiabetic patients with CKD), and Group 4 (50 healthy subjects as a control group). Patients in groups 1 and 3 were further classified based on the presence of macroalbuminuria and microalbuminuria. Hsa-miR-221 expression was measured by RT- qRT-PCR. DN patients had significantly elevated serum hsa-miR-221 levels than the other groups, while diabetic patients without nephropathy exhibited elevated levels compared to both nondiabetic patients with CKD, and the control group. The DN patients with macroalbuminuria revealed significantly higher mean values of hsa-miR-221 relative to the patients with microalbuminuria. Significant positive associations were observed in the DN group between serum hsa-miR-221 and fasting insulin, fasting glucose, HOMA IR, ACR, and BMI. The ROC curve analysis of serum hsa-miR-221 in the initial diagnosis of DN in DM revealed high specificity and sensitivity. Conclusions It is concluded that hsa-miR-221 has the potential to be a useful biomarker for prognostic and diagnostic purposes in DN.


Introduction
Between 1980 and 2021, the number of diabetic patients globally increased from 108 million to 536.6 million, climbing to 783.2 million in 2045.Type 2 diabetes represents about 90-95% of these cases [1,2].
Micro-and macrovascular complications in diabetes mellitus (DM) affect various systems and organs [3].Diabetic nephropathy (DN) is among the most serious and common diabetes complications and develops in 20% of diabetic patients [4,5].Renal replacement would be necessary for approximately half of all chronic diabetes patients having end-stage renal disease (ESRD) [6].
Diabetes is considered responsible for 80% of ESRD cases worldwide.Type 2 diabetes is responsible for 91% of all new cases of diabetes related ESRD, following the United States Renal Data System (USRDS) [7].
Proteinuria and glomerulopathy are developed gradually with years-long latent periods of DN [5].Although substantial recent studies, the specific etiology of DN remains unknown.However, the development of DN is known to involve various variables and mechanisms.
The pathophysiology of DN is marked by hypertrophy and hyperplasia of the mesangium of the glomeruli and dysfunction of podocyte [8].Long-term oxidative stress, advanced glycation end products (AGEs), hyperglycemia, and inflammation are all contributors to the development of DN [9].
MicroRNAs (miRs) have been attributed to the DN pathogenesis in two ways: Binding upregulated miRs to the protective genes in the kidney, repressing them, and downregulated miRNAs, in turn, can cause upregulation of genes that promote the mesangial expansion of or podocyte dysfunction [10].
Previous studies on hsa-miR-221 reported that its expression was upregulated and involved in the physiopathology of diabetes and type 2 diabetes (T2D) macrovascular and microvascular complications via targeting specific genes [13].
It has been found that aberrant overexpression of hsa-miR-221 accelerates retinal neovascularization in diabetic retinopathy by promoting cell migration, apoptosis, and proliferation, boosting hyperplasia and retinal artery blockage, and causing retinal ischemia and hypoxia [14].
As we know so far, no research has been accomplished to examine the potential of serum hsa-miR-221 as a biomarker for prognostic and diagnostic purposes in DN.The current study explores the hypothesis based on earlier research: whether serum hsa-miR-221 constitutes an innovative biomarker with significant accuracy for the diagnosis of DN.

Study design
This case-control study was conducted on a total of 200 participants chosen from the Internal Medicine Department at Beni-Suef University Hospital.The participants were divided into four groups: Group 1 (50 diabetic nephropathy patients), Group 2 (50 diabetic patients without nephropathy), Group 3 (50 nondiabetic patients having chronic kidney disease (CKD), and Group 4 (50 apparently healthy subjects as a control group).Then, the fifty patients with diabetic nephropathy were subdivided into 27 patients with macroalbuminuria and 23 patients with microalbuminuria.In contrast, the 50 patients with nondiabetic CKD were subdivided into 20 patients with macroalbuminuria and 20 patients with microalbuminuria.

Ethical considerations
The current study's protocol followed the ethical principles of the Helsinki Declaration [15].The study was granted approval by the Ethical Committee of the Faculty of Medicine, Beni-Suef University, with approval number (FMB-SUREC /09072023/Abd Elazeem).Informed written consent was obtained from all participants.

Patients' selection
According to the American Diabetes Association, both diabetic patients [16] and patients with diabetic nephropathy [17] were diagnosed.DN criteria of diagnosis were chronic albuminuria (> 200 g/min or > 300 mg/d) demonstrated on at least two occasions at least 3-6 months apart or an albumin-to-creatinine ratio (ACR) > 300 mg/g, with a consistent reduction in the glomerular filtration rate (GFR), without symptoms or signs of other primary renal diseases.
All patients' clinical and medical histories were obtained.Next, the patients were submitted to a comprehensive clinical evaluation to select individuals with diabetic nephropathy who were over the age of 18. Exclusion criteria were hypertension, congestive heart failure, severe illness, current infection, peripheral artery disease, pregnancy, and malignancy.By measuring weight and height, body mass index (BMI) was estimated [18].

Sampling and laboratory investigations
Five milliliters of venous blood were drawn and put in either plain or fluoride-containing vacutainer tubes.Then, Urine samples were collected to evaluate albuminuria and estimate the urinary ACR and eGFR.Following overnight fasting for 10 h at least, biochemical tests (fasting insulin, fasting glucose, total cholesterol, HDL-c, LDL-c, HbA1c, triglycerides, creatinine, and urea) were conducted.The formula [19] adopted to assess the Homeostasis model assessment for insulin resistance (HOMA-IR) was as follows: HOMA-IR = fasting insulin (mlU/L) X fasting glucose (mg/dL) ÷ 405.Insulin resistance was deemed high if it was ˃ 2 [20].Additionally, the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation [21] was employed to determine eGFR.LDL-c levels were calculated by Friedewald equation [22]: LDL-c = (Total Cholesterol mg/dL) − (HDL-c mg/dL) − (TGs mg/dL /5).

Real time quantitative reverse-transcription polymerase chain reaction (qRT-PCR)
According to manufacturer's instructions, Real-time qRT-PCR was accomplished utilizing a single-plex TaqMan twostep stem-loop real-time RT.TaqmanTM miRNA assay and the TaqmanTM miRNA Reverse Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA) were utilized to produce cDNA.The RT reaction total volume was 15 µL.Primer Premier 6.00 was used to design the primers (Premier Biosoft, Palo Alto, CA, USA).The micro-RNA specific stem loop primer of miRNA-221, and miRNA-16 (a reference gene) were added.TaqMan ® Small RNA Assay, and the TaqMan ® Universal PCR Master Mix II PCR were used to obtain amplified PCR from cDNA.

Analysis of data
The cycle threshold (CT) was adopted to quantify the levels of miRNA-221 expression.The comparative CT method (ΔΔ CT) was calculated by the difference between miRNA-221 CT and the reference gene average CT in each sample.Furthermore, miRNA expression fold change was calculated according to the formula 2 −ΔΔCt [23] using the controls as calibrators.

Statistical analysis
SPSS (Statistical Package for Social Sciences) version 23 (SPSS Inc., Chicago, IL, USA) was used.The data were presented as mean ± SD.The statistical analysis of group differences involved the use of the Kruskal-Walli test, the Shapiro-Wilk test, and the t test.To compare the discrepancies among groups, Tukey post-hoc testing was employed.To evaluate the linear relationships between the study genes and clinical factors, a simple linear correlation (Pearson correlation coefficient test) (r) was conducted.If p was 0.05, the p value was considered significant.Receiver operating characteristic curves (ROC curves) were employed to appraise the diagnostic performance of the studied parameter.

Results
The Frequency of some demographic, and clinical data in the studied groups were demonstrated in Table 1.

The mean levels of demographic and clinical data in the studied groups
The mean levels of demographic and clinical data in the investigated groups were recorded in Table 2.
The current study was conducted on 121 males (60.5% of subjects) and 79 females (39.5% of subjects).The mean levels of age and LDL-c revealed no significant variations across the studied groups.
In diabetic nephropathy, the mean levels of BMI, serum fasting insulin, serum fasting glucose, HOMA IR, serum HbA1c, serum creatinine, and ACR exhibited significantly higher mean levels.In contrast, the eGFR exhibited significantly lower mean levels than diabetic patients without nephropathy, nondiabetic patients with CKD, and healthy controls.Diabetic nephropathy patients had a significantly longer disease duration than diabetic patients without nephropathy and nondiabetic patients with CKD.
Compared to healthy controls and nondiabetic patients with CKD, both diabetic nephropathy and diabetic patients without nephropathy showed a significant decline in mean serum HDL-c levels along with significant increases in mean serum levels of TG and cholesterol.
The mean levels of serum BUN demonstrated a significant increase in diabetic nephropathy compared to diabetic patients without nephropathy, healthy controls, and nondiabetic patients having CKD.

The mean level of serum hsa-miR-221 in the studied groups
As demonstrated in Fig. 1a, the mean levels of serum hsa-miR-221 were significantly (p value ˂0.001) higher in Table 2 The demographic, and clinical data of the subjects in the studied groups DN diabetic nephropathy, DM diabetes mellitus without nephropathy, ND CKD Nondiabetic chronic kidney disease, N number of subjects, M ± SD mean ± standard deviation, a Significant differences compared to diabetic nephropathy group, b Significant differences compared to DM without nephropathy, c Significant differences compared to Nondiabetic CKD, d Significant differences compared to controls *Significant difference (p value < 0.05), #High significant difference (p value < 0 0.

The mean levels of hsa-miR-221 in the subgroups
According to Table 3, the mean levels of hsa-miR-221 in males and females in each group were not significantly different.Additionally, they showed no significant variations between patients who used rosiglitazone, glyburide, or metformin in any group of diabetic patients, However, they were significantly higher as obesity developed in patients with diabetic nephropathy.

The relation between hsa-miR-221 and levels of albuminuria in the DN group
As shown in Fig. 1b, in diabetic patients with nephropathy, the mean levels of hsa-miR-221 were significantly higher (P < 0.001) in patients with macroalbuminuria (3.11 ± 0.43) than those with microalbuminuria (2.43 ± 0.14).However, no considerable variation was found in the mean levels of hsa-miR-221 between patients with macroalbuminuria (0.992 ± 0.03) and those with microalbuminuria (0.996 ± 0.2) in the nondiabetic CKD group (Fig. 1c).

Logistic regression of significant predictor parameters for prediction of diabetic nephropathy
Univariate logistic regression revealed hsa-miR-221 to be a significant predictor of the development of diabetic nephropathy in diabetic patients, along with HbA1c, fasting glucose, fasting insulin, HOMA IR, and BMI (Table 4).This result remained in multivariate logistic regression models, whether conventional risk factors (BMI) or the significant variables from the univariate models (hsa-miR-221, HbA1c, fasting glucose, and HOMA IR; Table 4) were specified as covariates.According to Table 4, the combination of hsa-miR-221 and either fasting glucose or HbA1c showed the best performance and explained 70% and 79% (Nagelkerke R2) of the variation, respectively.

ROC curve analysis of serum hsa-miR-221 in the early detection of diabetic nephropathy in diabetes mellitus
The AUC of serum hsa-miR-221 was 0.953 with a sensitivity of 96% and specificity of 88% at the cut-off value ≥ 2.55, with a significantly accurate diagnosis for patients with diabetic nephropathy, as shown in Fig. 3.

Discussion
Early diagnosis and management of diabetic nephropathy are urgent to avoid complications and fatality [24].Although albuminuria is considered the best available test for the diagnosis of diabetic nephropathy, additional measures, such as serum creatinine, urine ACR, or eGFR, can be used to assess renal damage [25].However, none of those tests offers a significant sensitivity for early diagnosis of all patients with diabetic renal dysfunction [26,27].
The current investigations demonstrated that the mean levels of BMI, serum fasting insulin, serum fasting glucose, HOMA IR, serum HbA1c, serum creatinine, and urinary ACR were significantly higher.In contrast, the mean eGFR concentrations were significantly decreased in diabetic nephropathy patients than their corresponding levels in diabetic patients with no nephropathy, nondiabetic patients with CKD, and healthy controls, indicating the increased severity of diabetes and the deterioration of the kidney in patients with diabetic nephropathy.
Although BUN levels were deemed significantly higher in DN than in diabetic patients without nephropathy, and In the present study, the hsa-miR-221 mean levels revealed a significant increase in diabetic nephropathy and diabetic patients without nephropathy compared to patients with nondiabetic CKD and healthy controls.At the same time, they were higher in DN than in diabetic patients without nephropathy and comparable between nondiabetic CKD and healthy controls.These findings suggested the significance of hsa-miR-221 for the diabetes pathogenesis and the progression of diabetic nephropathy.
This role was explained by Li et al. [28], and Poliseno et al. [11] who coincided with these results as they reported that longstanding hyperglycemia in diabetes resulted in induction of the expression of hsa-miR-221 and dysfunction of endothelial cells.Also, Fiorentino et al. [29], Coinciding with the present findings Costantino et al. [30], Lightell et al. [13], and Li et al. [31] revealed that hsa-miR-221 expression was highly increased in diabetes and might be included in Similar findings were found in diabetic retinopathy (DR), as researchers identified hsa-miR-221 as a biomarker for DR in T2DM and proliferative DR patients, and it was thought to contribute to DM pathogenesis and associated macrovascular complications [32][33][34][35].Also, Li et al., [28] explained the importance of hsa-miR-221 in the occurrence and progression of T2DM as they reported that under high glucose levels, significantly elevated hsa-miR-221 could promote apoptosis and cell proliferation facilitating endothelial dysfunction and hyperplasia of retinal vessels contributing to diabetic microvascular complications.As a result, hsa-miR-221 seems to be abnormally increased in the serum of T2D patients, rising gradually as DR severity increased, aggravating the damage of retinal vessels and dysfunction exposed to elevated glucose levels suggesting that hsa-miR-221 could be implicated in angiogenesis, proposing an innovative hypothesis for DR in T2D patients.
In contrast to earlier and current findings, Togliatto et al. [36] discovered downregulation of hsa-miR-221 expression related to reduced angiogenesis in hyperglycemia, and Marei et al. [37] found substantial decreases in hsa-miR-221 levels in both placenta and serum in the GDM group relative to the control group.The findings on the effect of glucose levels on hsa-miR-221 expression are contentious.However, this could be attributed to modifications in cell culture procedures.Evidently, more research into the expression of hsa-miR-221 in hyperglycemia is needed.
The current study found that the mean levels of hsa-miR-221 in diabetic patients with nephropathy were greater in patients with macroalbuminuria than those with microalbuminuria but comparable in similar patients in the nondiabetic CKD group.Additionally, a significant positive correlation was noticed between hsa-miR-221 and urinary ACR in diabetic patients with nephropathy.
The current outcomes agree with those of Atya et al. [38], who discovered that the serum hsa-miR-221 expression and  ACR ratio had a significant positive correlation in diabetic patients and patients with either proliferative or non-proliferative diabetic retinopathy.The coincidence between hsa-miR-221 and DN and the level of albuminuria was explained by Su et al. [39], who reported that the proximal tubule was the target of impact in DN.In high-glucose conditions, podocytes triggered the division of epithelial cells of the proximal tubule, and extracellular vehicles mediated the connection.The level of hsa-miR-221, which was extracted from these extracellular vehicles, was significantly higher, promoting the division of the proximal tubule's epithelial cells.Additionally, in diabetic mice, the suppression of hsa-miR-221 corrected these abnormal changes in the proximal tubule's epithelial cells.
The present investigation found that in the diabetic patients with the nephropathy group, the hsa-miR-221 concentrations were positively correlated with BMI, fasting insulin, fasting glucose, HOMA IR, and urinary ACR.In contrast, in the diabetic patients without the nephropathy group, these correlations were limited to fasting insulin and HOMA IR.
The current correlations of hsa-miR-221 were similarly reported by Lustig et al. [40] who discovered that hsa-miR-221 was raised in obesity animal models, Atya et al. [38], who discovered that the serum hsa-miR-221 expression was positively correlated with HBA1c and HOMA-IR in diabetic patients without retinopathy and those with different degrees of diabetic retinopathy compared to healthy subjects, Liu et al. [14] who discovered that the serum hsa-miR-221 was significantly positively associated with metabolic HOMA IR in type 2 diabetes patients with diabetic retinopathy.
The explanation of these correlations were demonstrated by previous research [12], which revealed the binding of hsa-miR-221 to important genes and components of the insulin/insulin-like signaling pathway, as well as their relationship with metabolic syndrome, diabetes, and diabetic implications, Lustig et al. [40], who revealed that hsa-miR-221 could have a role in IR regulation via reducing adiponectin expression, and Meerson et al. [41] who showed that hsa-miR-221 expression in subcutaneous adipose tissue was positively linked with BMI in nondiabetic Pima Indians.
In contrast to the present results, Chou et al. [42] found that hsa-miR-221 concentrations in adipocytes were negatively associated with BMI.Additionally, Wang et al. [43] showed that metabolic syndrome-affected females without obesity had greater hsa-miR-221 plasma concentrations than did controls, while Ortega et al. [44] found lower plasma concentrations of hsa-miR-221 in obese patients.
The variance in these outcomes could be attributed to methodologic discrepancies or unknown attributes in the cohorts.Despite contradictory human data, hsa-miR-221 can promote IR development by suppressing adiponectin signaling and has the potential as a circulating biomarker.
This study revealed considerable declines in the HDL-c serum levels along with considerable elevations in the TG and cholesterol serum levels in diabetic nephropathy and diabetic patients without nephropathy relative to nondiabetic patients with CKD and controls.
Similar findings were discovered by Vinod et al. [45], who reported that serum cholesterol and TG levels were significantly elevated while HDL levels were significantly decreased in patients with poor control of hyperglycemia than those with good control, implying that the lipid profile could be used to predict and monitor diabetes mellitus.
The ROC curve analysis for the role of serum hsa-miR-221 in the early identification of diabetic nephropathy in diabetes mellitus revealed high substantial sensitivity and specificity with highly accurate diabetic nephropathy diagnosis.
No previous study investigated the use of serum hsa-miR-221 in the early detection of diabetic nephropathy in diabetes mellitus.Still, Atya et al. [38], and Liu et al. [14] discovered related results in diabetic retinopathy, revealing that hsa-miR-221 had clear diagnostic value in diabetic retinopathy in diabetic patients using ROC curve analysis.
By using the logistic regression, hsa-miR-221 was found to be a significant predictor of the development of diabetic nephropathy in diabetic patients, along with HbA1c, fasting glucose, fasting insulin, HOMA IR, and BMI.The combination of hsa-miR-221 and either fasting glucose or HbA1c showed the best performance and explained 70% and 79% of the variation, respectively assisting in an early diagnosis of the diabetic nephropathy, before kidney damage progresses or even before albuminuria develops.

Conclusion
According to the current study, diabetic nephropathy patients had elevated levels of hsa-miR-221 in their serum than diabetic patients without nephropathy and controls.In contrast, all diabetic patients had higher levels than nondiabetic individuals.Additionally, diabetic nephropathy patients with macroalbuminuria had increased levels of hsa-miR-221 than those with microalbuminuria.The levels of hsa-miR-221 positively correlated with the BMI, serum fasting insulin, serum fasting glucose, HOMA IR, and urinary ACR in the diabetic nephropathy group.The early detection of diabetic nephropathy in diabetes mellitus using serum hsa-miR-221 exhibited significant sensitivity and specificity with an extremely accurate diagnosis of diabetic nephropathy.This study points to hsa-miR-221 as a possible biomarker for diagnostic and prognostic proposes in diabetic nephropathy, contributing to diagnose the condition early before kidney damage worsens or even before albuminuria occurs.Future studies should pay greater focus to investigate novel methods to suppress the negative miRNAs.Prior to kidney structural pathological changes, therapeutic targeting of miRNAs would try to reverse these earlier changes.

Fig. 1
Fig. 1 The mean levels of serum hsa-mir-221 in studied subjects.a mean levels of serum hsa-mir-221 in studied groups.b mean levels of serum hsa-mir-221 in subjects with different degrees of albuminuria in diabetic nephropathy group.c mean levels of serum hsa-mir-221 in subjects with different degrees of albuminuria in nondiabetic CKD group

Fig. 2
Fig. 2 Significant correlations of serum hsa-mir-221 in the studied groups.Red dots diabetic nephropathy group, green dots diabetic without nephropathy group

Fig. 3
Fig. 3 ROC Curve Analysis of serum hsa-mir-221 in early detection of diabetic nephropathy in diabetic patients

Table 1
The Frequency of some demographic, and clinical data in the studied groups DN diabetic nephropathy, DM diabetes mellitus without nephropathy, ND CKD Nondiabetic chronic kidney disease, N number of cases, % percentage of frequency

Table 3
Serum hsa-mir-221 levels in the subgroups of the studied groups DN diabetic nephropathy, DM diabetes mellitus without nephropathy, ND CKD Nondiabetic chronic kidney disease, a significant differences compared to non-obese, b significant differences compared to overweight, c significant differences compared to obese

Table 4
Logistic regression of significant predictor parameters for prediction of diabetic nephropathyWald Wald Chi-square, OR Odds Ratio (per increase of 1 SD of the parameter), CI Confidence interval