Patients with typical keloid characteristics who had no prior treatment were recruited for the present study, as previously described (14). The diagnosis of keloid pathogenesis in the patient specimens was confirmed through routine pathological examination by an experienced dermatopathologist. The present study was conducted with the consent of each recruited patient and with the approval of the Ethics Committee of the Lanzhou General Hospital of Chinese People's Liberation Army (Lanzhou, China). A total of 10 patients with KFs, aged 17–42 years, were recruited from February 2016 to June 2017 at Burns and Plastic Surgery Center of PLA in Lanzhou General Hospital of Chinese People's Liberation Army, and the control group consisted of corresponding tissue sections obtained from the NFs adjacent to the lesion in these 10 keloid cases (Table I). As described previously (15), the excised 5-mm³ tissue biopsies (keloid explants or keloid OC) were embedded in a collagen gel matrix and were then preserved in serum-free William's medium E (WE medium; Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) (15).

The primary fibroblasts were cultured as described previously (9). Fibroblasts that had undergone 0–3 passages were used in the present study. All synthetic miRNAs and miRNA inhibitors (including the scrambled RNA as a negative control) were purchased from Shanghai GeneChem Co., Ltd. (Shanghai, China). The scrambled RNA was used as a control for mimic or inhibitor transfection experiments. All sequences are presented in Table II. Cells were transduced with miR-96 inhibitor (70 nM) or control molecules (70 nM) with Lipofectamine^® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.), 48 h prior to subsequent experiments. miR-96 mimics or control molecules (50 nM) were transfected using Lipofectamine^® 2000 (16).

The degree of fibrosis was examined using Masson's trichrome staining, according to standard protocols (1,17), and the stained tissue sections were then examined with a light microscope at magnification, ×400 (Olympus Corporation, Tokyo, Japan).

TargetScan Human 7.2 (www.targetscan.org/vert_72) was used to predict Smad7 as a target of miR-96. The wild-type (wt) or mutant (mut) Smad7 3′-untranslated region (UTR) was cloned into the pGL3 vector (Promega Corporation, Madison, WI, USA). Subsequently, the pGL3-Smad7-3′UTR-wt or pGL3-Smad7-3′UTR-mut vector, along with the miR-96 mimic or the mimic-control (as described above), were transfected with Lipofectamine^® 2000 into 293T cells (ATCC, Rockville, MD, USA). Following 48 h, the luciferase activity was measured with a Luciferase Reporter Gene Assay kit (Promega Corporation), using a GloMax-Multi Jr Single Tube Multimode Reader (Promega Corporation). Renilla luciferase activity was used for normalization of the firefly luciferase activity (18).

TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.) was used to extract the mRNA of cells and tissues, according to the manufacturer's protocol. Universal primer and the miScript reverse transcription kit (both Qiagen GmbH, Hilden, Germany) were used for reverse transcription of miRNA. RETROscript™ reverse transcription kit (cat. no. AM1710; Ambion; Thermo Fisher Scientific, Inc., Waltham, MA, USA) was used for mRNA reverse transcription. Triplicate RT-qPCR reactions and analyses were performed using a Bio-Rad C1000 Thermal Cycler (Bio-Rad Laboratories, Inc., Hercules, CA, USA). The qPCR reactions were performed with miScript SYBR Green PCR kit (cat. no. 218200; Qiagen GmbH) using the following thermocycling conditions: Initial denaturation for 15 sec at 95°C; 45 cycles of denaturation at 94°C for 15 sec, annealing at 55°C for 30 sec and extension at 70°C for 30 sec. The internal loading controls used for the mRNAs and miRNAs were GAPDH and RNU6B, respectively. The PCR primers used for mRNA and miR quantification are listed in Table III. Expression levels were determined using Applied Biosystems 7500 software version 2.0.1 (Applied Biosystems; Thermo Fisher Scientific, Inc.) and analyzed using the 2^−ΔΔCq method (19).

The proteins were extracted from KF and NF tissues using a mammalian protein extraction reagent (M-PER™; Thermo Fisher Scientific, Inc.) and concentrations were determined using a bicinchoninic acid assay kit (Thermo Fisher Scientific, Inc.). The lysates were mixed with laemmli buffer (2X; Bio-Rad Laboratories, Inc.) supplemented with 5% β-mercaptoethanol (1:1 ratio, Thermo Fisher Scientific, Inc.). Samples were boiled for 5 min at 95°C. Proteins (15 µg/lane) were separated using SDS-PAGE on 10% gel and subsequently the proteins were transferred to polyvinylidene fluoride membranes by cold transfer (25V at 4°C overnight). The membranes were blocked by a solution of 5% non-fat dried milk in Tris buffered saline with Tween (TBST; 25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.05% Tween-20) for 1 h at room temperature (20). The membrane was then incubated at 4°C overnight with primary antibodies for COL1A1 (cat. no. sc-293184, 1:1,000), COL3A1 (cat. no. sc-271249, 1:500), Smad7 (cat. no. sc-365846, 1:300; all Santa Cruz Biotechnology, Inc., Dallas, TX, USA), and β-actin (cat. no. sc-70319, 1:5,000; Sigma-Aldrich; Merck KGaA). All primary antibodies were detected using anti-mouse- or anti-rabbit-horseradish peroxidase conjugated secondary antibodies (1:3,000; cat. nos. sc-390944 and sc-5162, respectively; Santa Cruz Biotechnology, Inc.) at 37°C for 2 h. Membranes were washed three times with TBST and the chemiluminescent signals were detected using Clarity ECL Western Blotting substrate (cat. no. 1705061; Bio-Rad Laboratories, Inc.) and detected using the Odyssey Infrared Imaging System (LI-COR Biosciences, Lincoln, NE, USA). The relative protein expression of COL1A1 and COL3A1 was normalized to that of β-actin with Image-Pro Plus software (version 6.0; Media Cybernetics, Inc., Rockville, MD, USA).

The level of Col1 supernatant protein was assessed using COL1A1 and COL3A1 ELISA kits (cat. nos. ab210966 and ab7778, respectively; Abcam, Cambridge, MA, USA) according to the manufacturer's protocol. Briefly, the conditioned medium from cell culture was added to an ELISA kit plate precoated with an anti-COL1A1 and COL3A1 antibody. Subsequently a biotinylated secondary antibody was added and the plate was incubated at room temperature for 2 h. The optical density was measured at a wavelength of 450 nm (21).

The chemically synthesized miR-96 inhibitor, antagomir, functions to decrease the miR-96 expression level in the keloid OC (22–24). miR-96 antagomir (100 µg; Guangzhou RiboBio Co., Ltd., Guangzhou, China) was added to 1 ml keloid OC in serum-free WE medium every 72 h for a period of five weeks at 37°C, while the control group received an equal amount of scrambled RNA in the same antagomir medium, and the medium was changed every three days. Keloid shrinkage in vitro was assessed through measuring the dry weight of the 5-µm Masson's trichrome-stained sections on an analytical balance.

si-Smad7 sequences were chemically synthesized and purified by high-performance liquid chromatography (Shanghai GenePharma Co., Ltd., Shanghai, China). All the oligonucleotides were 2′-OMe modified. Briefly, cells were transfected with siRNA-Smad7 at a final concentration of 50 nM using Lipofectamine^® 2000. siRNA-NC was transfected as a negative control. At 24 h post-transfection, the culture medium was changed according to the protocols of manufacturer. After 48 h, cells were harvested for analysis. All transfections were performed in triplicate (25).

Statistical analysis was performed using a two-tailed unpaired Student's t-test, paired Student's t-test, Pearson's linear correlation test and one-way analysis of variance with Tukey's post hoc test. All data were obtained from triplicate experiments and were presented as the mean ± standard deviation. P<0.05 was considered to indicate a statistically significant difference.

The expression levels of Smad7 and miR-96 were compared between the KFs and NFs obtained from 10 keloid patients. It was demonstrated that the mRNA expression of miR-96 was increased whereas that of Smad7 was decreased in the KFs, as compared with the NFs (Fig. 1A-C). Furthermore, their expression levels were inversely correlated (Fig. 1D). Given that type I collagen is the major component of the ECM that participates in keloid formation, the correlation between miR-96 and COL1A1 was also investigated. Endogenous miR-96 was demonstrated to be positively correlated with the COL1A1 expression levels (Fig. 1E). Masson's staining of KF tissues revealed relatively high COL1A1 and miR-96 expression levels in case #5, yet a low Smad7 expression level (Fig. 1F); whereas case #9 exhibited relatively low expression levels of miR-96 and COL1A1, yet a high Smad7 expression level (Fig. 1G). Examination of the structural characteristics of the keloid scars of cases #5 and #9 revealed that case #5, which had a higher expression level of miR-96 and a lower level of Smad7, exhibited a thicker collagen fiber deposition and a relatively non-uniform collagen density distribution (Fig. 1F). Conversely, case #9, which had a lower level of miR-96 and a higher level of Smad7, exhibited a thinner collagen fiber deposition and a relatively uniform density distribution (Fig. 1G).

The putative binding sites of miR-96 in the region of the 3′UTR of Smad7 were detected through bioinformatics analysis. Therefore, the predicted region, containing the wt or mut seed sequence of miR-96 in the 3′UTR of Smad7, was cloned into the luciferase reporter plasmid (Fig. 2A and B). The 293T cells transfected with the miR-96 mimic and the Smad7-3′-UTR had a lower luciferase intensity, but the mutant reporter-transfected group did not (Fig. 2C), thus indicating that Smad7 may be a direct target of miR-96. Western blotting and RT-qPCR were used to confirm the role of miR-96 in regulating Smad7 expression in the KFs and NFs. The miR-96 expression level was higher in the KFs than in the NFs (Fig. 1A), thus, the miR-96 mimic was transfected into NFs, whereas the miR-96 inhibitor was transfected into the KFs (Fig. 2D). The results demonstrated that Smad7 was decreased in the miR-96 mimic-transfected NFs at the mRNA and protein levels; however, the Smad7 expression level was increased in the KFs transfected with the miR-96 inhibitor, compared with the corresponding control groups (Fig. 2E and F). These results suggest that miR-96 may regulate Smad7 expression in KFs and NFs at both the mRNA and protein levels.

Type I and Ш collagen are the most common types of collagen that are deposited in the abnormal ECM of KFs (1–3,10,26). Therefore, these two collagen types were chosen to evaluate the regulatory effect of miR-96 on ECM deposition. As Smad7 is a well-known inhibitory cytokine in the TGF-β pathway and is under the regulation of miR-96, it was assumed that miR-96 inhibition may reduce the COL1A1 and COL3A1 expression levels in KFs. ELISA was used to confirm that transfection of the miR-96 inhibitor into KFs led to a reduction in COL1A1 and COL3A1 expression (Fig. 3A and B). However, the introduction of siRNA-Smad7 was demonstrated to significantly reverse the miR-96-inhibitor-induced Smad7 upregulation (Fig. 3C and D), thereby leading to increased COL1A1 and COL3A1 production (Fig. 3E and F).

Local gene therapies targeted at reducing fibrosis through the application of liposome-complex small interfering RNA or an miRNA mimic, or through the direct administration of plasmid, have been demonstrated to be effective and successful (22). Given the validation of the degradation ability of antagomirs (23), miRNA antagomir was added to the keloid OC media, as described previously (22–24). The RT-qPCR data demonstrated that the miR-96 expression level was barely changed at each time point in the control group. However, the miR-96 antagomir-treated keloid OC tissues exhibited decreased miR-96 expression levels at each time point, compared with the control groups (Fig. 4A). These findings demonstrated that the direct delivery of miR-96 antagomir into the keloid OC may lead to a reduction in miR-96 expression.

Compared with the scrambled RNA-treated control, the dry weight of the keloid OC was significantly reduced by 8% at week 3, and by 14% at week 5 (Fig. 4B). The keloid OC models were treated with miR-96 antagomir to investigate its effect on collagen degradation. In the miR-96 antagomir-treated group, both the COL1A1 and COL3A1 mRNA expression levels in the keloid tissues at the end of the 4th week of OC were significantly reduced (Fig. 4C). Similarly, the protein expression levels of COL1A1 and COL3A1 (Fig. 4D and E) were also significantly inhibited in the miR-96 antagomir-treated group. These findings indicated that the addition of miR-96 antagomir to the keloid OC decreased the quantity of type I and Ш collagen. The weight of the keloid OC was measured weekly for a period of 5 weeks post-treatment to assess the antifibrotic potential of miR-96 antagomir. Visible keloid shrinkage was observed at the end of week 4 (Fig. 4F).

At the end of 4 weeks of OC, the harvested keloid tissues were subjected to Masson's staining. Thinner and more loosely arranged collagen fibers were observed in the miR-96 antagomir-treated keloid tissues, compared with the control (Fig. 4F).