Involvement of inflammation-related miR-155 and miR-146a in diabetic nephropathy: implications for glomerular endothelial injury

Six patients with type 2 DN were included in this study (two women and four men, age 51.7 ± 7.8 years), and the kidney tissues were obtained from renal biopsies. Renal clinical staging and pathologic classification of these six patients were referred to Mogosen stage of DN [20] and pathologic classification of DN established by the Renal Pathology Society [21]. Three patients who were diagnosed with renal carcinoma and were undergoing nephrectomy but were shown to be free from hypertension, diabetes, and other co-morbidities were included as controls (one woman and two men, age 44 ± 12.9 years, urine test and serum creatinine were normal) (Table 1). The OGTT of the three controls included in our study was normal, thus, we thought they were not diabetes. Kidney tissues were obtained from renal biopsies (patients) or nephrectomy surgery (controls, primarily kidney cortexes) after receiving signed informed consent from the individuals. Tissues were used for a miRNA microarray and real-time polymerase chain reaction (PCR) or fixed in formalin and embedded in paraffin for histological and in situ hybridization. Changes in the renal morphology were examined in formalin, paraffin-embedded tissue sections (4 μm) stained with Periodic Acid-Schiff (PAS). The study was approved by the Ethics Committee on Human Research at Sichuan University (Sichuan, China).

Dissected kidney tissue was homogenized, and RNA was extracted using TRIzol (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. The RNA was purified using a miRNeasy mini kit (QIAGEN) according to the manufacturer’s instructions. The RNA’s quality and quantity were determined using a Nanodrop 1000. To study the miRNA-expression in the diabetic patients, we performed a miRNA expression profiling of the kidney samples from six diabetic patients and three healthy controls using miRNA microarrays (Exiqon, Vedbaek, Denmark) according to the manufacturer’s instructions; the miRNA microarray service was provided by KangChen Bio-Tech, Inc., Shanghai, China. Briefly, after the RNA isolation, the miRCURY™ Hy3™/Hy5™ Power labeling kit (Exiqon, Vedbaek, Denmark) was used according to the manufacturer’s guidelines for miRNA labeling. One microgram of each sample was 3’-end-labeled with a Hy3TM fluorescent label by combining the RNA (one microgram) in 2.0 μL of water with 1.0 μL of CIP buffer (0.5 μL) and CIP (0.5 μL) (Exiqon). The mixture was incubated for 30 min at 37°C, and the reaction was terminated by incubation for 5 min at 95°C. Then, 3.0 μL of the labeling buffer, 1.5 μL of the fluorescent label (Hy3TM), 2.0 μL of the DMSO, and 2.0 μL of the labeling enzyme (this enzyme was actually a labeling enzyme from the miRCURYTM Array Power Labeling kit (Cat #208032-A, Exiqon)) were added to the mixture. The labeling reaction was incubated for 1 h at 16°C, and the reaction was stopped by incubation for 15 min at 65°C. The Hy3TM-labeled sample (25 μL) with 25 μL of hybridization buffer was denatured for 2 min at 95°C, incubated on ice for 2 min, and then hybridized to the miRCURYTM LNA Array (v.16.0) (Exiqon), according to the manufacturer’s instructions, for 16–20 h at 56°C in a 12-Bay hybridization system (Hybridization System; Nimblegen Systems, Inc., Madison, WI, USA). Following hybridization, the slides were washed several times using a wash buffer kit (Exiqon) and dried by centrifugation for 5 min at 400 rpm (the slides were dried by centrifugation for 5 min. at 1000 rpm). The slides were then scanned using the Axon GenePix 4000B microarray scanner (Axon Instruments, Foster City, CA, USA), and the images were imported using GenePix Pro 6.0 software (Axon Instruments) for the grid alignment and data extraction. The replicated miRNAs were averaged, and those with intensities ≥50 were selected to calculate the normalization factor. After normalization, the significant differentially expressed miRNAs were identified through Volcano Plot filtering. Hierarchical clustering was performed using MEV software (v4.6, TIGR) (Figure 1).

Locked nucleic acid-based in situ detection of miRNA in kidney sections was performed. Briefly, the kidney samples were fixed at 4°C in formalin, embedded in paraffin wax, and cut into 4 μm sections. Dig-labeled probes were designed and synthesized by Invitrogen Inc. The sequences of the Dig-labeled probes against miR-155 and miR-146a mRNA were as follows: miR-155: 5′-CCCCTATCACGATTAGCATTAA-3′, miR-146a: 5′-AACCCATGGAATTCAGTTCTCA-3′. A scrambled probe sequence: 5′-TTCACAATGCGTTATCGGATGT-3′ was used as negative control for non-specific hybridization signal. In situ hybridization was performed using a detection kit on the sections of 4% paraformaldehyde-fixed (containing 0.1% diethylpyrocarbonate) kidney tissues according to the manufacturer’s instructions.

Littermate Sprague Dawley (SD) rats (male, aged 6 weeks, 200–220 g) were kept under laboratory standard conditions at a temperature of 22 ± 2°C, with a 12-h light/12-h dark cycle and relative humidity of 40%–60%. They were divided into two groups: the non-diabetic rats (controls, n = 6) and the streptozotocin (STZ)-induced type 1 diabetes rats (DM, n = 18). The DM rat models were induced by an intraperitoneal injection of a single dose of 60 mg/kg STZ (Sigma, St. Louis, MO, USA) in citrate buffer (pH 4.5). The random blood glucose level was tested using a glucometer (Accu-Chek, Roche, Basel, Switzerland), and blood glucose levels were determined on days 3 and 7 after the STZ injection. Only the rats with glucose concentrations ≥16.7 mmol/L for 3 consecutive days were used in the study. The DM rats were euthanized in weeks 1, 4, and 8 after the induction of diabetes, and their kidneys were harvested. Their renal cortexes were collected for real-time PCR by carefully removing the renal pelvis and medullar tissues. The experimental protocol was approved by the Animal Experimentation Ethics Committee at Sichuan University.

In a separate experiment to generate type 2 diabetes SD rats, rats (male, aged 6 weeks, 200–220 g) were allocated to one of two dietary regimens for an initial period of 6 weeks: a normal pellet diet (NPD, n = 6) or a high-fat and high-sugar diet (HFD, containing a regular diet plus 27.3% lard, 54.6% sucrose, 16.4% cholesterol, and 1.6% sodium cholate (w/w), n = 18). After this time, the HFD animals with high HOMA-IR (fasting plasma glucose (mmol/L) × fasting insulin (mIU/L) ÷ 22.5) were defined as insulin resistant and were injected with multiple low-doses of STZ (four doses of 25 mg/kg; Sigma, St. Louis, MO, USA) in citrate buffer (pH 4.5) after overnight fasting. The NPD animals were injected with a citrate buffer (1 mL/kg). The rats with fasting glucose levels ≥16.7 mmol/L 72 h after the STZ injection for 3 consecutive days were used for the study. The rats were kept on their respective diets until the end of the study. The DM rats were euthanized in weeks 0, 8, and 16 after the induction of diabetes, and their kidneys were harvested and their renal cortexes collected as before. All of the experimental protocols involving the animals were approved by the Animal Experimentation Ethics Committee at Sichuan University.

Fasting (6 h) blood glucose levels were determined weekly for the first 2 weeks then every 2 weeks thereafter by using blood glucose meter (Optium™ Xceed Systems, Victoria, Australia). Before the disease induction (week 0), urine was collected for 24 h by placing rats in metabolic cages and performing a urinary albumin excretion assay. This process was repeated after the disease induction (for type 1 diabetes: weeks 1, 4, and 8; for type 2 diabetes: weeks 0, 8, and 16). The urinary albumin was measured by a competitive ELISA according to the manufacturer’s instructions (MaxiSorp Surface Immuno Plates™; Nunc, Roskilde, Denmark). The serum creatinine was detected by the QuantiChrom Creatinine Assay Kit (BioAssay Systems, Hayward, CA, USA).

To confirm the miRNA profiling findings in the human subjects, we used a quantitative real-time PCR to measure the expression of miR-155 and miR-146a in both the human and the rat kidneys. The reverse transcription reaction used MMLV reverse transcriptase (Epicenter, Madison, WI, USA), and a quantitative PCR was performed using an ABI PRISM7500 system (Applied Biosystems, Foster City, CA, USA). miR-155 and miR-146a were further quantified with a TaqMan quantitative real-time PCR. The PCR primers used are shown in Table 2.

Normal HRGECs were purchased from ScienCell Research Laboratories (Carlsbad, CA, USA) and cultured in an endothelial cell medium (ScienCell) containing 5% fetal bovine serum and 1% endothelial cell growth supplement at 37°C in a 5% CO₂ atmosphere. The cells were serially passaged by a brief exposure to 0.25% trypsin (BD-Difco, USA) and 0.04% EDTA (Sigma, St. Louis, MO, USA), and the experiments were performed in the cells at passages 2–5 at 80% confluence. They were seeded at a density of 1 × 10⁵ cells/well in 6-well dishes or 2.5 × 10⁵ cells/25 cm² flask. Prior to the experiments, the cells were fasted for 24 h in maintenance media.

To detect the time-dependent changes in the miR-155 and miR-146a expression, the cells cultured with mediums containing 25 mmol/L glucose were cultured at 37°C in 5% CO₂ for 0, 0.5, 1, 2, 4, 8, and 24 h. They were then collected for a real-time PCR. Chemically modified RNA oligonucleotides comprising a sequence complementary to mature miR-155 and miR-146a (miR-155 and miR-146a inhibitors) were used to inhibit the miR-155 and miR-146a activities. The miR-155 and miR-146a mimics were double-stranded constructs consisting of a guide and passenger strands. An oligonucleotide containing a four-base mismatch (nontargeting scramble RNA) was used as a negative control. Pyrrolidine dithiocarbamate (PDTC, Sigma) was used as an NF-κB inhibitor. According to the set time points, the HRGECs (1 × 10⁵ per well) were transfected with miR-155 and miR-146a mimics, inhibitors, and scramble control after 24 h of starvation in a serum-free medium using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions.

The proteins were extracted from the HRGECs using a RIPA lysis buffer and analyzed by western blot as described previously [17]. Briefly, after blocking the nonspecific binding with 5% bovine serum albumin, the membranes were incubated with primary antibodies against TNF-α and transforming growth factor (TGF)-β1 overnight at 4°C. After washing, the membranes were incubated with IRDye 800-conjugated secondary antibodies (Rockland Immunochemicals, Inc., Gilbertsville, PA, USA), and the signals were detected using the Odyssey Infrared Imaging System (Li-COR Biosciences, Lincoln, NE, USA) and quantitated with Image J software (NIH). The protein ratio was normalized against the GAPDH and expressed as the mean ± standard errors of the mean (SEM).

The total RNA was extracted from the HRGECs using the RNeasy kit according to the manufacturer’s instructions (Qiagen Inc., Valencia, CA, USA), and the cDNA was synthesized. A real-time PCR was performed with the Opticon®2 Real-Time PCR detector (Bio-Rad, Hercules, CA, USA) using the IQ SYBR green supermix reagent (Bio-Rad) as described previously [17]. The primers used to amplify the mRNAs of TNF-α, TGF-β1, and GAPDH are listed in Table 2. The housekeeping gene GAPDH was used as an internal standard. The ratios of specific mRNA:GAPDH mRNA were calculated and expressed as the mean ± SEM.

EMSA was performed using the LightShift™ chemiluminescent EMSA kit (Pierce Biotechnology, Rockford, Illinois, USA) according to the manufacturer’s protocol. Nuclear proteins were extracted from the HRGECs using a nuclear extraction kit (Pierce Biotechnology, Rockford, Illinois, USA) and incubated with a biotin-labeled NF-κB probe (5′-AGT TGA GGG GAC TTT CCC AGG C-3′) at room temperature for 20 min. Samples were separated by 6% non-denaturing polyacrylamide gel electrophoresis and transferred to a positively charged nylon membrane (0.45 μm, Pierce Biotechnology, Rockford, Illinois, USA).

The clinical and pathological characteristics of the patients are shown in Table 1. The cases of DN were notable for their statistically significant decreased estimated glomerular filtration rate (eGFR), increased proteinuria, serum creatinine levels, and increased glycosylated hemoglobin compared with the controls. They were also defined by the presence of diabetes, proteinuria, histological changes, and the absence of hepatitis, HIV, lupus, or other potential causes of glomerulonephritis. Three control samples were obtained from the healthy or biopsied samples of the unaffected portions of the tumor nephrectomies. The control subjects were defined as having normal serum creatinine levels and eGFR and an absence of proteinuria and hematuria.

The miRNA expression of the kidney tissues was analyzed using microarray screening. Compared with the normal controls, the 71 miRNAs in the type 2 DN patient samples were differentially expressed, with 32 elevated and 39 down-regulated (Figure 1). Thirty-one miRNAs were upregulated by ≥ 2-fold in the DN kidney samples from the DN patients, in which there were two significantly upregulated miRNAs, miR-155 (6.22 ± 2.81-fold) and miR-146a (4.87 ± 1.39-fold). These findings were further confirmed by a quantitative real-time PCR for miR-155 and miR-146a (Figure 2A,B).

The morphology of the kidney glomeruli from the DN patients was examined using PAS staining. The precise location and expression levels of the miR-155 and miR-146a were determined using in situ hybridization in the patient and control kidney tissue sections stained with Dig-labeled miR-155 and miR-146a riboprobes. The miR-155 and miR-146a were primarily localized in the GECs, mesangial areas and tubular sections of the DN kidneys, and the staining was significantly stronger in the DN GECs compared with the control samples (Figure 2C,D), which is consistent with the quantitative real-time PCR results above. To understand how these differentially expressed miRNAs might contribute to the inflammatory response, we searched for their potential regulatory targets using algorithms based on miRNA-mRNA complementarity and evolutionary conservation seed sequences (Target-Scan and PicTar). Surprisingly, some of the candidate targets of miR-155 and miR-146a were involved in the inflammatory responses (data not shown); therefore, we used miR-155 and miR-146a as targets for further study.

The linear correlation analysis revealed that the miR-155 expression was closely correlated with the serum creatinine levels (R = 0.95, P = 0.004) but not with the urinary protein excretion levels (R = 0.336, P = 0.461); there were also no significant correlations between the miR-146a and serum creatinine levels (R = 0.531, P = 0.220) or the urinary protein excretion levels (R = 0.360, P = 0.427) (Figure 3).

In the rats with type 1 DN, the rats were sacrificed after the induction of diabetes at 1, 4, and 8 weeks (serum creatinine and urinary protein excretion were shown in Table 3), and a quantitative real-time PCR was used to show that the expression of miR-155 and miR-146a gradually increased during the progression of the DN (Figure 4A,B). A similar trend was observed for miR-155 and miR-146a expression in the development and progression of DN in the rats with type 2 DN (Figure 5A,B). In comparison with the NPD control group, the expression levels of miR-146a and miR-155 were significantly increased in the type 2 DN rats from week 0 to week 8 and 16 (Figure 5A,B) (serum creatinine and urinary protein excretion were shown in Table 4).

To determine the effects of high glucose on the expression of miR-155 and miR-146a, the HRGECs stimulated with a medium containing 25 mmol/L glucose were cultured at 37°C in 5% CO₂ for 0, 0.5, 1, 2, 4, 8, and 24 h. The expression of both the miR-155 and miR-146a was up-regulated 2 h after the glucose stimulation, peaked after 4 h, and then decreased until 24 h (with a minor increase of the miR-155 expression) relative to the control and the mannitol groups (Figure 6A,B).To explore the biological effect of miR-155 and miR-146a on the HRGECs under high-glucose conditions, the HRGECs were categorized as the normal control (containing 5 mmol/L glucose), mannitol (20 mmol/L mannitol), high glucose (containing 25 mmol/L glucose), high glucose + miR-155 mimic (50 nmol/L), high glucose + miR-155 inhibitor (100 nmol/L), high glucose + Scrambled miRNA (50 nmol/L), high glucose + PDTC (100 μmol/L), and high glucose + miR155 (50 nmol/L) + PDTC (100 μmol/L) groups. These cells were transfected with miR mimics or inhibitors as indicated, then harvested 24 h after transfection for the real-time PCR analysis and after 48 h for the western blot analysis. Compared with the normal control and mannitol groups, high glucose levels induced the up-regulation of both proinflammatory cytokines (TNF-α) and pro-fibrotic growth factors (TGF-β1), as shown by both the real-time PCR (Figure 6C,D) and western blotting (Figure 6E,F). Under high-glucose conditions, the HRGECs overexpression of miR-155 or miR-146a following the transfection with miR-155 or miR-146a mimics showed an additive effect on the TNF-α and TGF-β1 upregulation. However, the induction by glucose was clearly blunted by the miR-155 and miR-146a inhibitors, suggesting that miR-155 and miR-146a could significantly induce renal inflammation and fibrosis and contribute to HRGEC injury.To clarify the underlying NF-κB signaling pathway, the DNA-binding activity of NF-κB was assessed by EMSA using nuclear extracts from the HRGECs. A strong nuclear activation of NF-κB was observed in the cells from the miR-155 and miR-146a group (Figure 6G,H). Furthermore, the addition of PDTC partially repressed the up-regulated expression of TNF-α and TGF-β1 caused by the activation of NF-κB in the miR-155 and miR-146a over-expressing groups.

Written informed consents were obtained from the patients for publishing their medical data and any accompanying images. The copies of the written consents are available for review by the editor of this journal.