Specimens of keloid scar tissue (N = 20) and adjacent normal skin tissue (N = 20) were collected from the chest or face of patients who visited the Department of Dermatology at the Hainan Hospital of Traditional Chinese Medicine. The patients were 22 to 35 years old and did not have any systemic diseases. All patients provided their informed consent for study participation, and the study protocol was approved by the Ethics Committee of Hainan Hospital of Traditional Chinese Medicine.

Primary human keloid fibroblasts were isolated from keloid tissues as previously described (Lei et al., 2019). In brief, we preserved the dermal layer of fibrous tissue by cutting away the epidermis and fat from excised human keloid scar tissues. After cutting the preserved tissue specimens into smaller pieces, the dermal layer of fibrous tissue in each specimen was digested for 6 h in DMEM medium containing 0.25% collagenase Type I (Sigma-Aldrich, C0130, St. Louis, MO, USA). Next, the digested tissue was centrifuged, and the precipitates were cultured in DMEM contained 15% fetal bovine serum (Gibco, 10099141C, Grand Island, NY, USA) at 37°C in a humidified atmosphere containing 5% CO₂.

Two small interfering RNAs targeting hsa_circ_0002198 via different sequences (siRNA#1 and siRNA#2), an NLRP3 overexpression plasmid, and the corresponding negative control (NC), were synthesized by Gene Pharma (Shanghai, China). Knockdown of hsa_circ_0002198 in keloid fibroblasts was achieved by transfecting the fibroblasts with siRNA#1 and siRNA#2. In rescue experiments, the keloid fibroblasts were co-transfected with a pcDNA3.1 empty vector (NC and siRNA#1 group) or siRNA#1 with/without NLRP3 (siRNA#1+NLRP3 and siRNA#1 group). All transfections were performed using by using Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA), a 50 nM concentration of each siRNA, 2.5 μg of plasmid, and a transfection duration of 48 h.

Total RNA was isolated from tissues or cells using TRIzol reagent (Invitrogen, USA) and reverse transcription was performed using a Prime Script™ RT kit (Takara, Dalian, China) according to the manufacturer’s instructions. Cytoplasmic RNA and nuclear RNA were isolated using a Cytoplasmic & Nuclear RNA Purification Kit (BioBeck Corporation, Canada). Briefly, cells were lysed in lysis buffer J solution; after which, the lysis buffer was centrifuged to produce a cell supernatant (cytoplasmic fraction) and pellet (nuclear fraction). Next, SK buffer was added to the RNA fractions, followed by addition of ethanol reagent. The mixture was then transferred onto a filtration column and the filtered liquor was discarded after centrifugation. The RNA remaining in the filtration column was washed using washing solution followed by centrifugation. Finally, RNA was collected by adding elution buffer and performing a subsequent centrifugation. The filtered liquor was retained and stored at −20°C for further analysis. A 1 µg sample of total RNA was analyzed for its levels of hsa_circ_0002198, PDE7B, and NLRP3 RNA by using SYBR green PCR Master Mix (Toyobo, Japan) on an ABI 7500 fast PCR System (Foster City, CA, USA) using the following conditions: pre-incubation at 95°C for 5 min, followed by 40 cycles of amplification at 95°C for 10 s, 60°C for 20 s, and a final elongation at 72°C for 15 s. Actinomycin D (AAT Bioquest, AAT-17505, Sunnyvale, CA, USA) was used to block transcription when detecting the half-life of RNA. The PCR primer sequences used were as follows: 5’-GTATCACAGGCTGCTGTAGCTCC-3’ and 5’-GCTCTTCGGTGCAGCTACTGG-3’ for hsa_circ_0002198; 5’-CTCGCTTCGGCAGCACA-3’ and 5’-AACGCTTCACGAATTTGCGT-3’ for GAPDH. Relative gene expression was calculated by 2^−ΔΔCt method. 5’-TCTCATGCTGCCTGTTCTCA-3’ and 5’-CAAGGAGATGTCGAAGCAGC-3’ for NLRP3; 5’-GGTGTGGCGAAATCTTGTTT-3’ and 5’-TTTTCTTGGTGCCAATCTCC-3’ for PDE7B.

The FISH assay was performed to detect the subcellular localization of hsa_circ_0002198 with a RiboTM Fluor 488-labeled probe (RiboBio, Guangzhou, China) as described in the manufacturer’s instructions. In brief, cells were inoculated in a 24-well plate at a density of 6 × 104 cells/well, and then fixed with 4% paraformaldehyde for 20 min at room temperature. Next, the cells were incubated overnight with pre-hybridization solution and then washed 3 times with PBST; after which, the nuclei were stained with 4’, 6-diamidino-2-phenylindole (DAPI, Invitrogen) for 5 min. Finally, the cells were observed and photographed under a fluorescence microscope (Olympus, Japan).

The proliferative ability of cells was examined using a Cell Counting kit-8 (CCK-8, Dojindo Molecular Technologies, Inc., Kumamoto, Japan). Briefly, cells from different groups were seeded into 96-well plates (3 × 103 cells per well) and cultured in complete medium. At 0, 24, 48, and 72 h, respectively, 10 μL of sterile CCK-8 solution was added to each well. After another 2 h incubation at 37°C, the colorimetric absorbance of each well at 450 nm was recorded using a microplate reader (Thermo Labsystems, Helsinki, Finland).

The migration and invasion abilities of keloid fibroblasts were determined using 24-well Transwell chambers (Corning, Corning, New York, USA) that had either been pre-coated or not pre-coated with 50 μL (120 μg/mL) of Matrigel, respectively. Briefly, cells were resuspended in FBS-free medium and then seeded in the upper chamber, and 600 µL of culture medium with 20% FBS (Gibco, 10099141C) was added to the lower chamber. After a 24 h incubation, the cells that had penetrated through the membranes were fixed with 4% paraformaldehyde for 20 min and then stained with 0.5% crystal violet. The numbers of migrated or invasive cells were counted under a microscope at × 200 magnification.

Total cellular protein was extracted from tissue samples or cells using RIPA lysis buffer (Beyotime, P0013E, Shanghai, China) containing a protease inhibitor cocktail (Sigma-Aldrich, P9599). The protein concentration in each extract was determined using the bicinchoninic acid (BCA) method. An equal amount of protein from each extract was subjected to 10% SDS-PAGE, and the separated protein bands were transferred onto PVDF membranes, which were subsequently blocked with 5% fat-free milk for 2 h at room temperature. Next, the membranes were incubated with primary antibodies against NLRP3 (Boster, Pleasanton, CA, USA, BA3677), Collagen I (Abcam, Cambridge, UK, ab260043), α-SMA (Abcam, ab124964), Caspase-3 (Abcam, ab32042), and GAPDH (Servicebio, China, GB11002) overnight at 4°C. Next, the membranes were incubated with a horseradish peroxidase-labeled secondary antibody (Abcam, ab150077) for 2 h at room temperature. The immunostained protein bands were visualized with enhanced chemiluminescence reagent (ECL, Bio-Rad, Hercules, CA, USA).

For IF staining, keloid fibroblasts were plated onto glass coverslips and cultured to 80% confluence; after which, they were incubated in complete medium overnight. Next, the fibroblasts were rinsed twice with PBC and fixed with 4% formaldehyde solution for 30 min. The cells were then blocked with 5% bovine serum albumin (BSA, Sigma-Aldrich) for 30 min at room temperature and subsequently incubated with a primary antibody against NLRP3 (Boster, BA3677, Cambridge, UK) overnight at 4°C, followed by incubation with an Alexa-conjugated secondary antibody (Abcam, ab150077) for 2 h. The nucleus was stained with DAPI (Invitrogen), and images were obtained with an LSM510 confocal microscope (Molecular Devices, Sunnyvale, CA, USA).

The levels of cytokines (TGF-β, IL-1ß, and IL-33) in the supernatants of keloid fibroblasts from different groups were quantified using commercially available enzyme ELISA kits (R&D Systems, DB100B, DLB50, D3300B, Minneapolis, Minnesota, USA).

All statistical analyses were performed using Graphpad Prism version 6.0 software (GraphPad Software Inc., San Diego, CA, USA), and results are expressed as a mean value ± standard deviation. Differences among multiple groups were assessed by one-way analysis of variance (ANOVA), followed by the Dunnett’s post-test or Turkey’s post-hoc test. P-values < 0.05 were considered statistically significant.

The expression levels of hsa_circ_0002198 were significantly higher in samples of keloid scar tissue than in samples of adjacent normal tissue (Fig. 1A). Furthermore, the levels of NLRP3 RNA and protein expression were obviously elevated in the scar tissues when compared with the paired control tissues (Figs. 1B and 1C). These preliminary data suggested that elevated hsa_circ_0002198 and NLRP3 levels might be involved in the regulation of scar hyperplasia.

Hsa_circ_0002198 was identified by a half-life comparison and use of a second-generation sequencing method. As shown in Fig. 2A, when measured under the influence of actinomycetes, the half-life of hsa_circ_0002198 was longer than that of its host gene (linear mRNA PDE7B), reflecting the stability of hsa_circ_0002198. Results from second-generation sequencing identified the splice sites (TCAG/GAAA) for hsa_circ_0002198 (Fig. 2B). Further detection showed that hsa_circ_0002198 mainly distributed in cytoplasm (Cy) (Fig. 2C). Moreover, the subcellular localization of hsa_circ_0002198 in cells was further explored using the FISH assay. As depicted in Fig. 2D, hsa_circ_0002198 was mainly expressed in the cytoplasm of keloid fibroblasts rather than in the nucleus. While, we also found that the signal in normal skin fibroblasts was weaker than in keloid fibroblasts (Fig. 2D).

To investigate the biological function of hsa_circ_0002198 in scar hyperplasia in vitro, we first established a hsa_circ_0002198 knockdown system by using si-circRNA in keloid fibroblasts. CCK-8 assays showed that knockdown of hsa_circ_0002198 by siRNA#1 or siRNA#2 transfection significantly suppressed the proliferation of keloid fibroblasts (Fig. 3A). Subsequently, results from Transwell assays revealed that the numbers of migratory (Fig. 3B) and invasive (Fig. 3C) keloid fibroblast cells were notably decreased in the siRNA#1 and siRNA#2 transfection group when compared with the NC or blank group. At the molecular level, TGF-β expression was significantly downregulated after knockdown of hsa_circ_0002198 by siRNA#1 or siRNA#2 (Fig. 3D). Moreover, the levels of NLRP3 mRNA expression were downregulated after hsa_circ_0002198 knockdown (Fig. 3E). A western blot analysis showed that knockdown of hsa_circ_0002198 downregulated the levels of Col I, α-SMA, and NLRP3 protein expression, and upregulated caspase 3 expression in keloid fibroblasts (Fig. 3F). The changes in NLRP3 protein expression in keloid fibroblasts were further confirmed by IF assays (Fig. 3G). In addition, ELISA assays showed that the levels of pro-inflammatory cytokines (IL-1ß and IL-33) were decreased following hsa_circ_0002198 knockdown (Figs. 3H and 3I). Finally, quantitative real-time PCR confirmed the downregulation of hsa_circ_0002198 after siRNA#1 or siRNA#2 transfection (Fig. 3J).

As NLRP3 was upregulated in scar tissues and could be suppressed by knockdown of hsa_circ_0002198, it seemed reasonable that NLRP3 might be a downstream regulator in hsa_circ_0002198 knockdown cells that attenuates cell functions in keloid fibroblasts. To help validate this expectation, keloid fibroblasts were co-transfected with siRNA#1 and the NLRP3 overexpression plasmid. A subsequent series of functional experiments showed that overexpression of NLRP3 partially reversed the suppressive effects of hsa_circ_0002198 on keloid fibroblast cell proliferation (Fig. 4A), migration, (Fig. 4B), and invasion (Fig. 4C). The decreases in TGF-β levels, induced by hsa_circ_0002198 knockdown, were recovered by NLRP3 overexpression (Fig. 4D). Expression of NLRP3 mRNA was suppressed by hsa_circ_0002198 siRNA, while that loss of expression was recovered after NLRP3 was overexpressed (Fig. 4E). As expected, NLRP3 overexpression partially reversed the effects of hsa_circ_0002198 knockdown on Col I, α-SMA, caspase-3, and NLRP3 expression in keloid fibroblasts, as reflected by western blot analyses (Fig. 4F) and IF staining of NLRP3 (Fig. 4G). In addition, restoration of NLRP3 expression attenuated the effects of hsa_circ_0002198 knockdown on the levels of IL-1ß (Fig. 4H) and IL-33 (Fig. 4I). However, NLRP3 overexpression did not affect hsa_circ_0002198 expression in the siRNA#1 transfection group (Fig. 4J).