The study group consisted of 100 patients (35 female and 65 male, median age of 42.7 years, age range of 18.2-83.6) who underwent functional endoscopic sinus surgery or septoplasty by a single surgeon at the Department of Otolaryngology, The First People's Hospital of Qujing (Qujing, China) between July 2018 and June 2019. Patients younger than 18 years old, with unilateral nasal polyps or with associated diseases, such as cystic fibrosis, inverted papilloma and ciliary dyskinesia were excluded from the present study. Each tissue was divided into four parts: One part was reserved for cell culture, one part was fixed for immunofluorescence evaluation using formalin for paraffin sectioning section or frozen sectioning, and the last two parts were stored at −80°C for protein and mRNA extraction. The study was approved by the medical ethics committee of The First People's Hospital of Qujing, and written informed consent was obtained from each patient before participation in the study.

A human nasal epithelial cell line (NP69; cat. no. BNCC338439) and Escherichia coli BL21 competent cells (BL21; cat. no. BNCC353591) was purchased from BeNa Culture Collection; Beijing Beina Chuanglian Biotechnology Research Institute, human embryonic kidney 293T cells were purchased from The Cell Bank of Type Culture Collection of The Chinese Academy of Sciences. The MMP-2 and MMP-9 protein expression of 100 CRSwNPs tissues was determined via western blotting, and the CRSwNPs tissues with the lowest (Fig. S1; green box) and highest (Fig. S1; red box) expression of MMP-2 and MMP-9 were used to isolate PHNECs with the lowest and highest expression of MMP-2 and MMP-9 (L-PHNECs and H-PHNECs, respectively) as previously described (32). In brief, CRSwNPs tissue samples of ~1 ml volume were rinsed with normal saline, transferred into 10 ml DMEM/F12 medium (Thermo Fisher Scientific, Inc.) containing 1% penicillin/streptomycin (Sangon Biotech Co., Ltd.), digested with 0.1% protease from Streptomyces griseus (Thermo Fisher Scientific, Inc.) and 0.1 mg/ml deoxyribonuclease I (Sangon Biotech Co., Ltd.), and incubated at 4°C overnight. Epithelial cells were removed by gentle scraping and dispersed into a single cell suspension. The medium was then transferred into a 15-ml conical tube and centrifuged at 300 × g for 5 min at room temperature. The supernatant was decanted, and the pellet was resuspended in DMEM/F12 medium.

Total RNA was extracted from CRSwNPs tissues or cells using TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. First-strand cDNA was synthesized from the total RNA according to the instructions of the PrimeScript™ RT Reagent Kit (Takara Biotechnology Co., Ltd.), and RT-qPCR was subsequently performed using the SYBR-Green qPCR kit (Thermo Fisher Scientific, Inc.), according to the manufacturer's protocols. The following primer sequences were used for RT-qPCR: miR-29b-3p forward, 5′-ACA CTC CAG CTG GGT AGC ACC ATT TGA AAT C-3′ and reverse, 5′-TGG TGT CGT GGA GTC G-3′; and U6 forward, 5′-CTC GCT TCG GCA GCA CAT A-3′ and reverse, 5′-AAC GAT TCA CGA ATT TGC GT-3′. The RT-qPCR experiments were performed on an Applied Biosystems 7900HT Fast Real-time PCR system (Applied Biosystems; Thermo Fisher Scientific, Inc.). The following thermocycling conditions were used for RT-qPCR: Initial denaturation at 95°C for 7 min; followed by 40 cycles of denaturation at 95°C for 10 sec, annealing at 60°C for 20 sec and elongation at 72°C for 20 sec; and a final extension at 72°C for 10 min. The relative expression levels were calculated using the 2^−∆∆Cq method (33) and normalized to those of internal reference gene U6.

Proteins were extracted from CRSwNPs tissues or cells using radioimmunoprecipitation assay (Beyotime Institute of Biotechnology), and the concentrations were determined according to the standard protocols of BCA protein assay kit (Beyotime Institute of Biotechnology), respectively. The total protein (30 µg/well) in the supernatant was separated via SDS-PAGE on 10% gel, and then transferred to PVDF membranes. After blocking with 5% skimmed milk for 1 h at room temperature, the membranes were incubated overnight at 4°C with rabbit anti-MMP-2 (1:1,000; no. ab92536; Abcam), rabbit anti-MMP-9 (1:1,000; no. ab76003; Abcam), rabbit anti-TIMP-1 (1:1,000; no. ab211926; Abcam), rabbit anti-integrin β1 (1:1,000; no. ab52971; Abcam), rabbit anti-α-tubulin (1:5,000; no. ab18207; Abcam), rabbit anti-acetyl-α-tubulin (1:1,000; no. ab179484; Abcam) and rabbit anti-β-actin (1:5,000; no. ab8227; Abcam) antibodies. After three washes with TBS with 0.1% Tween-20, the immunoblots were incubated for 1 h at room temperature with goat alkaline phosphatase-labeled anti-rabbit antibody (1:1,000; cat. no. 14708; Cell Signaling Technology, Inc.). The immunoreactive bands were visualized using an enhanced chemiluminescence reagent (Beyotime Institute of Biotechnology). The blots were semi-quantified using ImageJ software (version 1.47; National Institutes of Health).

miR-29b-3p mimic (50 nM; 5′-UAG CAC CAU UUG AAA UCA GUG UU-3′), NC mimics (50 nM; 5′-UUG UAC UAC ACA AAA GUA CUG-3′), miR-29b-3p inhibitor (100 nM; 5′-UUC UCC GAA CGU GUC ACG UTT-3′), NC inhibitor (100 nM; 5′-CAG UAC UUU UGU GUA GUA-3′), specific small interfering RNAs (siRNAs) targeting MMP-9 (si-MMP-9; 1.0 µg; 5′-ACC ACA ACA TCA CCT ATT GGA TC-3′), siRNA-negative control (si-NC; 1.0 µg; 5′-UUC UCC GAA CGU GUC ACG UTT-3′) were purchased from Shanghai GenePharma Co., Ltd. To overexpress MMP-9, the sequences of MMP-9 were inserted into a pcDNA3.1 plasmid to obtain the MMP-9 overexpression plasmid pcDNA3.1-MMP-9 (OE-MMP-9; 1.0 µg; Shanghai GenePharma Co., Ltd.), and an empty pcDNA3.1 plasmid was used as the negative control (OE-NC; 1.0 µg). Plasmid DNA, siRNA, miR-mimic or miR-inhibitor was transfected into L-PHNECs, H-PHNECs or NP69 cells (1×10⁵), which were subcultured at a density of 80%, with Lipofectamine^® 2000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.) at 37°C. After transfection for 48 h, the transfection efficiency was detected via RT-qPCR and western blotting, and then subsequent experiments were performed.

NP69 cells (1×10⁵ cells/well) were seeded in 12-well plates and transfected as described above. When the NP69 cells reached 80-90% confluence, the cells were washed with phosphate-buffered saline (PBS; 37°C, pH 7.4), and fresh culture medium was added, along with LPS (Sangon Biotech Co., Ltd.) at a concentration of 1 µg/ml and incubated at 37°C for 48 h.

StarBase (Version 2.0; http://starbase.sysu.edu.cn/) online software was used to predict the binding sites of miRNAs to target mRNAs. pmirGLO-MMP-9-wild-type (WT)/mutant (Mut) and pmirGLO-MMP-2-WT/Mut reporter plasmids were provided by Shanghai GenePharma Co., Ltd. 293T cells (2×10⁵/well) were co-transfected with pmirGLO-MMP-9-Wt/Mut or pmirGLO-MMP-2-Wt/Mut plasmid (1.0 µg) and NC mimic or miR-29b-3p mimic (50 nM) using Lipofectamine^® 2000 reagent at 37°C. At 48 h post-transfection, luciferase activity was determined using the dual-luciferase reporter assay system (Promega Corporation). Firefly luciferase activities were normalized to Renilla luciferase activities.

For immunofluorescence staining, 10-µm-thick tissue sections of CRSwNPs samples were fixed with 4% paraformaldehyde for 2 h at 4°C. CRSwNPs were incubated with blocking solution [5% bovine serum albumin (Thermo Fisher Scientific, Inc.) with 0.2% Triton X-100] for 1 h at room temperature, and incubated overnight at 4°C with antibodies against MMP-9 (1:500; cat. no. ab76003; Abcam) and integrin β1 (1:100; cat. no. ab52971; Abcam). After washing, the sections were incubated with Alexa Fluor 555-conjugated anti-rabbit IgG (1:500; no. ab150062; Abcam) at room temperature for 1 h, and the nuclei were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) for 30 min at room temperature. Fluorescence images were collected with a Nikon Eclipse 80i microscope (Nikon Corporation).

Myc-integrin β1, Myc-α-tubulin and/or MMP-9-WT-HA plasmids were purchased from Transomic Technologies, Inc., and were transiently transfected into 293T cells with Lipofectamine^® 2000 reagent (Invitrogen; Thermo Fisher Scientific, Inc.) at 37°C. At 24 h posttransfection, the cells were harvested and lysed with 500 µl IP lysis buffer containing protease inhibitor cocktail (Thermo Fisher Scientific, Inc.). After incubating on ice for 5 min, the cell lysates were centrifuged (4°C) at 13,000 × g for 10 min. Then, ~25% of the supernatant was subjected to input assays, and the remaining supernatant was used for the Co-IP assay with an anti-HA agarose affinity gel (Shanghai Yeasen Biotechnology Co., Ltd.) according to the manufacturer's instructions. Rabbit anti-HA (1 µg; cat. no. ab9110; Abcam) and rabbit anti-Myc (1 µg; cat. no. ab9106; Abcam) were used for IP. Beads alone were used as the negative control. Briefly, 500 µl agarose affinity gel was centrifuged at 13,000 × g for 30 sec at 4°C to remove the glycerol and washed with cold TBS. Then, 500 µl cell lysate was added to the equilibrated resin and rocked gently on a rotating platform for 2 h at 4°C. The resin was washed with cold TBS, and the protein samples were evaluated by western blotting.

GST pull-down assays were carried out as previously described (34). Briefly, the pGEX-GST-MMP-9 plasmid was transformed into BL21 cells to GST-MMP-9 proteins. The pcDNA-Myc-integrin β1 or pcDNA-Myc-α-tubulin plasmid was transfected into 293T cells to express Myc-integrin β1 and Myc-α-tubulin proteins. The Pierce™ GST Protein Interaction Pull-Down Kit (cat. no. 21516; Pierce; Thermo Fisher Scientific, Inc.) was used according to the manufacturer's instructions. BL21 cells expressing GST-MMP-9 proteins were treated with pull-down lysis buffer and immobilized on equilibrated glutathione agarose resin at 4°C for 2 h. The resin was washed with wash solution (TBS with Pull-down lysis buffer), and 293T lysates containing Myc-integrin β1 or Myc-α-tubulin protein were added, followed by incubation at 4°C for 12 h. After washing with a wash solution (TBS with Pull-down lysis buffer), the resin was eluted with glutathione elution buffer. The protein samples were evaluated via western blotting.

After 24 h of treatment with 1 µg/ml LPS, cell viability was assessed using Cell Counting Kit-8 (CCK-8; Beyotime Institute of Biotechnology). Briefly, cells were seeded in 96-well plates at a density of 1×10⁵ cells/well; 10 µl CCK-8 solution was added to each well and incubated for 4 h. The OD value at 450 nm was measured utilizing a microplate reader (BioTek Instruments, Inc.).

Apoptosis was detected using an Annexin V combined fluorescein isothiocyanate (FITC)/propidine iodide (PI) cell apoptosis detection kit (Beijing Solarbio Science & Technology Co., Ltd.). In brief, cells were collected using cold PBS buffer and then cultured with 5 µl Annexin V-FITC reagent and 5 µl PI in the dark for 15 min at room temperature. Subsequently, 400 µl 1X binding buffer was added, and the cells were analyzed using a BD FACSCanto II flow cytometer (BD Biosciences) with FlowJo software (version 10; FlowJo LLC). Early and late apoptosis were both analyzed.

The cell culture media was centrifuged at 2,000 × g for 10 min to remove debris, and the cell-free culture supernatants were collected after treatment. The secretory levels of IL-6 (cat. no. 900-T16), TNF-α (cat. no. 900-M25) and IL-1β (cat. no. 900-M95) were analyzed using their corresponding ELISA kits (PeproTech China.) according to the manufacturer's protocol. The concentrations of inflammatory cytokines were measured using a microplate spectrophotometer (BioTek Instruments, Inc.) at a wavelength of 450 nm.

Differential expression of miR-29b-3p between three nasal cell samples and four nasal polyps cell samples from airway epithelia samples was analyzed using the NCBI GEO DataSets portal: http://www.ncbi.nlm.nih.gov/geo/. The miRNA expression data (GEO accession no. GSE159708) from high-throughput sequencing were submitted by Pommier et al (35). The web-accessible analysis tool GEO2R (https://www.ncbi.nlm.nih.gov/geo/geo2r/) was utilized on its default settings to screen the miR-29b-3p expression of each dataset. The cut-off value for the filtration criteria was set at FDR<0.05 and log2 fold-change>1.

In the 100 CRSwNPs tissue samples, the levels of MMP-2, MMP-9 and TIMP-1 protein were detected via western blotting, the blots were semi-quantified by ImageJ software and normalized to those of the internal reference gene β-actin. The levels of miR-29b-3p were determined via RT-qPCR, and the relative expression levels were calculated using the 2^−ΔΔCq method and normalized to those of internal reference gene U6. Bivariate correlation analyses of the correlations between MMP-2 and MMP-9 expression and miR-29b-3p, or TIMP-1 expression were performed.

All the experiments were repeated three times. GraphPad Prism 8 (GraphPad Software, Inc.) was used for statistical analysis, and the data are presented as the mean ± standard deviation. Data between two groups were analyzed using an unpaired Student's t-test, and data among multiple groups were analyzed by one-way analysis of variance (ANOVA) followed by a Tukey's post hoc test. P<0.05 was considered to indicate a statistically significant difference.

Previous studies have shown that miR-29b-3p regulates the expression of MMP-2 and MMP-9 (28,29), and the balance between MMPs and TIMPs is important in tissue homeostasis within CRSwNPs (5). To examine the relationships between miR-29b-3p, MMP-2/MMP-9 and TIMP-1 in CRSwNPs, the mRNA expression of miR-29b-3p was detected using RT-qPCR, and the protein expression levels of MMP-2, MMP-9 and TIMP-1 were determined via western blotting in 100 CRSwNPs tissue samples (Fig. S1). The results showed that miR-29b-3p expression was moderately positively correlated with the expression of MMP-2 (r=0.4420; P<0.001; Fig. 1A) and MMP-9 (r=0.4799; P<0.001; Fig. 1B), and TIMP-1 expression was moderately negatively correlated with the expression of MMP-2 (r=-0.4672; P<0.001; Fig. 1C) and MMP-9 (r=−0.4484; P<0.001; Fig. 1D). Immunofluorescent images showed MMP-2, MMP-9 and TIMP-1 protein expression in CRSwNPs tissue samples (Fig. 1E). These observations suggested that miR-29b-3p plays a protective role in the regulation of the balance in MMP and TIMP expression.

Previous studies have shown that miR-29b-3p targets the 3′-untranslated region of MMP-2 and MMP-9 (28,29). Using the bioinformatics database StarBase to search for potential targets, a putative interaction between miR-29b-3p and MMP-2/MMP-9 was found, and the target binding sequence is shown in Fig. 2A and B. A dual-luciferase reporter assay demonstrated that MMP-2-WT/MMP-9-WT and miR-29b-3p mimic co-transfection significantly decreased the luciferase activities in 293T cells (P<0.01; Fig. 2C and D), while MMP-2-MUT/MMP-9-MUT and miR-29b-3p mimic co-transfection failed to affect the luciferase activity in 293T cells (P>0.05; Fig. 2C and D). Next, transfection of miR-29b-3p inhibitor and miR-29b-3p mimic into PHNECs significantly decreased and increased miR-29b-3p expression, respectively (P<0.01; Fig. 2E and G). Notably, transfection of miR-29b-3p inhibitor into PHNECs significantly increased MMP-2 and MMP-9 protein expression in L-PHNECs (P<0.01; Fig. 2F), and transfection with miR-29b-3p mimic significantly decreased MMP-2 and MMP-9 protein expression in H-PHNECs (P<0.01; Fig. 2H). Altogether, the aforementioned results indicated that miR-29b-3p directly targeted both the MMP-2 and MMP-9 genes.

Compared with MMP-2, the r value (Fig. 1A and B) and the targeting capability (Fig. 2F and H) were higher for MMP-9. Thus, the miR-29b-3p/MMP-9 axis was investigated in subsequent assays. In a recent study, integrin β1 and α-tubulin were identified as potential MMP-9-interacting proteins (25). To examine the interaction between MMP-9 and integrin β1/α-tubulin, Myc-integrin β1/Myc-α-tubulin and HA-MMP-9 were transiently co-expressed in 293T cells and Co-IP assays were performed using a HA antibody. Indeed, MMP-9 readily immunoprecipitated Myc-tagged integrin β1 and α-tubulin (Fig. 3A and B). To further determine whether the interaction between integrin β1/α-tubulin and MMP-9 was direct, GST pull-down assays were performed using the GST-MMP-9 protein expressed in and purified from bacteria. As shown in Fig. 3C and D, GST-fused MMP-9 pulled down integrin β1, but not α-tubulin. These observations suggested that integrin β1 could directly interact with MMP-9, but that α-tubulin indirectly interacts with MMP-9. These results suggested that MMP-9 binds with integrin β1.

MMP-9 and integrin β1 activation can increase α-tubulin acetylation (23), and Smith (24) and the present study found that MMP-9 directly bound to integrin β1 (Fig. 3C). We hypothesized that MMP-9 binds to integrin β1 to promote α-tubulin acetylation by downregulating miR-29b-3p in CRSwNPs. MMP-9 expression was knocked down in H-PHNECs by transfection with si-MMP-9 (P=0.024; Fig. 4A), whereas expression of MMP-9 was upregulated in L-PHNECs by transfection with OE-MMP-9 (P<0.01; Fig. 4B). To test this hypothesis, miR-29b-3p was overexpressed in H-PHNECs. The western blotting results showed that MMP-9 expression was reduced after overexpression of miR-29b-3p (P<0.01), but this effect was reversed after overexpression of MMP-9 (P<0.01; Fig. 4C). By contrast, the expression of acetyl-α-tubulin was elevated after overexpression of miR-29b-3p (P<0.01), which was repressed by overexpression of MMP-9 (P<0.01). In addition, miR-29b-3p expression was inhibited in L-PHNECs. These results showed the expression of MMP-9 was elevated by miR-29b-3p inhibitors (P<0.01), but this effect was reversed by MMP-9 knockdown (P=0.002; Fig. 4D). By contrast, the expression of acetyl-α-tubulin was decreased by the inhibition of miR-29b-3p (P<0.01), but it was enhanced by MMP-9 knockdown. Of note, the protein expression of integrin β1 was not affected (P>0.05). Immunofluorescence was performed to examine the protein expression of MMP-9 and integrin β1 in tissues with lower and higher MMP-9 expression. As shown in Fig. 4E, there was no notable difference in the expression of integrin β1 in low and high MMP tissues. Next, the correlation between MMP-9 and acetyl-α-tubulin expression was further analyzed in 100 CRSwNPs tissue samples, and a moderate negative correlation was observed between MMP-9 and acetyl-α-tubulin (r=-0.3435; P=0.004; Fig. 4F). Taken together, we propose a novel model in which MMP-9 binds to integrin β1 to promote α-tubulin deacetylation-mediated tissue remodeling by downregulating miR-29b-3p in CRSwNPs (Fig. 4G).

MMP-9 expression was downregulated in NP69 cells by transfection with si-MMP-9 (P=0.043; Fig. 5A), whereas expression was upregulated by transfection with OE-MMP-9 (P=0.006; Fig. 5B). To explore the potential role of miR-29b-3p in CRSwNPs, miR-29b-3p expression in nasal polyps was investigated using the GEO database. This analysis showed that miR-29b-3p was expressed at lower levels in nasal polyps (P=0.008; Fig. 5C). Next, miR-29b-3p expression in PHNECs and NP69 cells was measured via RT-qPCR. As shown in Fig. 5D, the expression of miR-29b-3p was downregulated in both L-PHNECs and H-PHNECs compared with NP69 cells (P<0.001), and it was expressed at lower levels in H-PHNECs than in L-PHNECs (P=0.005). These results suggested that miR-29b-3p was downregulated in CRSwNPs. Next, a cell model of LPS-induced CRSwNPs was investigated. The results showed that miR-29b-3p expression was significantly lower in LPS-induced NP69 cells than in the control cells (P<0.01; Fig. 5E) and was significantly increased by the upregulation of miR-29b-3p compared with the LPS group (P<0.01). Notably, overexpression of MMP-9 reduced miR-29b-3p expression compared with the LPS + miR-29 mimics group (P<0.01) and knockdown of MMP-9 elevated miR-29b-3p expression compared with the LPS group (P<0.01). In addition, LPS also significantly increased MMP-9 expression in NP69 cells (P<0.01; Fig. 5F), which were effectively reversed by upregulation of miR-29b-3p and knockdown of MMP-9 (P<0.01), whereas overexpression of MMP-9 reversed the inhibitory effect of miR-29b-3p upregulation (P<0.01). Conversely, LPS significantly decreased acetyl-α-tubulin level (P<0.01; Fig. 5F), which was effectively elevated by upregulation of miR-29b-3p and knockdown of MMP-9 (P<0.05), but overexpression of MMP-9 reversed the effects of miR-29b-3p overexpression (P<0.01). Of note, the protein expression of integrin β1 was not altered (P>0.05). The results suggested that miR-29b-3p regulated acetyl-α-tubulin levels by targeting MMP-9. Next, cell viability, apoptosis and inflammatory cytokines (IL-1β, IL-6 and TNF-α) of LPS-induced NP69 cells were determined. As presented in Fig. 5G, cell viability was inhibited after LPS induction (P<0.01), and upregulation of miR-29b-3p and knockdown of MMP-9 alleviated the inhibitory effect of LPS (P<0.01). Overexpression of MMP-9 reduced the alleviatory effect of miR-29b-3p upregulation on cell viability (P<0.01). Moreover, cell apoptosis and inflammatory cytokine levels were increased after LPS induction (P<0.01; Fig. 5H and I), and these effects were reduced by the upregulation of miR-29b-3p and knockdown of MMP-9 (P<0.01); however, the overexpression of MMP-9 repressed the inhibitory effect of miR-29b-3p upregulation (P<0.01). These data indicated that overexpression of miR-29b-3p alleviated LPS-induced inflammation in NP69 cells by targeting MMP-9.