Clinical samples

A total of 30 type 2 diabetes mellitus (DM) patients and 30 healthy donors were enrolled into this study. Serums were collected and immediately frozen in liquid nitrogen for further analysis. All subjects have been informed before study, and this study was approved by the Ethical Committee of Shaanxi Provincial People’s Hospital.

Cell culture

The mouse MCs (Shanghai Academy of Life Science, Shanghai, China) were grown in DMEM (Gibco, Shanghai, China) plus 1% penicillin/streptomycin (Gibco) and 20% fetal bovine serum (FBS, Gibco) at 37℃ with 5% CO2.

MCs at 80% confluency were treated with 30 mmol/l D-glucose (Solarbio, Beijing, China) (high glucose group, HG) to imitate the growth environment of MCs in DN condition. Cells incubated with 5.5 mmol/l D-glucose plus 19.5 mmol/l mannitol were used as the control to imitate normal growth environment.

Reverse transcription and quantitative real-time PCR (qRT-PCR)

Total RNA was extracted using TRIzol reagent (Takara, Dalian, China). The PrimeScript RT Reagent Kit (Takara) was applied for the synthesis of first-strand cDNA. The quantification of molecules was measured by qRT-PCR analysis with the TB Green Premix Ex Taq II (Takara). The relative fold changes were represented by a cycle threshold (Ct) value. GAPDH or U6 was as an internal control. The sequence of primers was listed in Table 1.

Table 1 Primers sequences used for qRT-PCR.RNase R digestion and actinomycin D treatment

For RNase R digestion, approximately 3 µg of total RNA was treated with RNase R (3 U/µg, Geneseed, Guangzhou, China) or Mock for 30 min at 37℃. For Actinomycin D treatment, 5 µg/ml Actinomycin D (ActD; Geneseed) was used to incubate with MCs to block the de novo RNA synthesis. Finally, levels of circ_Arf3 and Arf3 mRNA were detected by qRT-PCR.

Cell transfection

pCD5-ciR/circ_Arf3 overexpression plasmids and scrambled pCD5-ciR plasmids (Vector), miR-107-3p inhibitor (anti-miR-107-3p), mimic (miR-107-3p) and negative control oligos (anti-miR-NC or miR-NC), siRNA targeting Tmbim6 (si-Tmbim6) and the nontarget siRNA (si-NC) were constructed by Sangon (Shanghai, China). Lipofectamine 2000 (Invitrogen, Camarillo, CA, USA) was employed for cell transfection. After 48 h of transfection, MCs were stimulated with HG condition for subsequent functional experiments.

5-Ethynyl-2’-deoxyuridine (EdU) assay

MCs were seeded into a 96-well plate overnight and incubated with 50 µM EdU labeling solution (RiboBio) in growth medium for 2 h. After being settled with 4% paraformaldehyde for 30 min, cells were stained with Click-It reaction mixture for 30 min, followed by Hoechst 33,342 staining (100 µl). Finally, EdU positive cells were observed and calculated using a fluorescence microscope (Leica, Wetzlar, Germany).

Cell counting kit-8 (CCK-8) assay

Single MCs (5 × 104 cells/ml) were grown in each well of a 96-well plate overnight, and then mixed with 10 µl CCK-8 solution in 100 µl DMEM medium for 2 h incubation. Lastly, the OD value was read at 450 nm using a microplate reader at indicated times to assess cell proliferation.

Western blotting

Proteins were isolated by using RIPA lysis buffer (Beyotime, Shanghai, China) plus 1% proteinase inhibitor and quantified employing the BCA kit (KenGen Biotech, Nanjing, China). About 40 µg of proteins were loaded onto 8% SDS-PAGE for separating and then electrophoretically shift onto PVDF membranes (Millipore, Temecula, CA, USA). After being blocked with 5% skim milk powder at 37℃ for 1 h, membranes were probed with the specific primary antibodies against PCNA (1:2000, ab18197), α-smooth muscle actin (α-SMA) (1:1000, ab32575), Collagen I (Col I) (1:2000, ab138492), Fibronectin (FN) (1:1000, ab32419), Collagen IV (Col IV) (1:1000, ab6586) and GAPDH (1:5000, ab181602) (Abcam, Cambridge, MA, USA) all night at 4℃, and then incubated with HRP-conjugated secondary antibody (D110058, 1:4000, Sangon Biotech, Shanghai, China) at 37℃ for 2 h. Protein bands were quantified using an ECL reagent (Beyotime), and gray values were quantified by Image Lab software.

Dual-luciferase reporter assay

The fragments of circ_Arf3 or Tmbim6 3’UTR comprising the binding sites of miR-107-3p or the mutant version without miR-107-3p binding sites were amplified and cloned into pGL3-Basic (Promega, Madison, WI, USA) to establish wild-type (WT) or mutated (MUT) circ_Arf3-WT/MUT or Tmbim6-WT/MUT luciferase reporter vector. Then 50 ng above recombinant plasmids, 10 ng pRL-TK Renilla and 50 nM miR-107-3p mimic or the control (miR-NC) were co-transfected into MCs using Lipofectamine 2000 for 48 h, and luciferase activities were detected using the Dual Luciferase Assay System (Promega).

RNA pull-down assay

Biotin-labeled circ_Arf3 WT/MUT probe (Bio-circ_Arf3-WT/MUT) or biotin-labeled Tmbim6 WT/MUT probe (Bio-Tmbim6-WT/MUT) and nonsense control probe (Bio-NC) were synthesized by Geneseed Biotech (Shanghai, China). MCs were lysed by using lysis buffer, and cell lysates were then incubated with above biotinylated probes for 2 h. Upon incubation with M-280 Streptavidin magnetic beads (Invitrogen) for 1 h, the RNA complexes bound to the beads were purified, and the relative RNA level of miR-107-3p was detected with qRT-PCR.

Animal experiments

This animal experiment was approved by the Ethical Committee of Shaanxi Provincial People’s Hospital. Healthy C57BL/6J mice (25–30 g, n = 12, male, 8-week-old) were purchased from Hunan Lake Jingda (Hunan, China), and maintained in standardized conditions. Diabetes mice were fed for 4 weeks with high-fat diet and then induced by injecting 55 mg/kg streptozotocin (STZ, Sigma-Aldrich, St. Louis, MO, USA) after 12 h of fasting every day for 5 consecutive days. Mice in the Sham group were fed with a standard rodent food and tap water for 4 weeks and then injected with an equal volume of citrate buffer. 72 h later, the fasting blood glucose (FBG) of mice was determined, and mice with FBG higher than 11.1 mmol/L were selected as successful diabetic mice model. At the end of the third week after injection, when the blood glucose of mice was ≥ 16.7mmol/L, and urine protein positive, the diabetic nephropathy mouse model was successfully established. Then all mice were anesthetized and kidneys tissues were harvested.

Statistical analysis

The data from three repetitions were manifested as mean ± standard deviation (SD). The significance of differences between groups was analyzed using Student’s t test, Mann-Whitney or ANOVA with Tukey’s post-test. *p < 0.05 suggested statistically significant.