This study included 32 patients diagnosed with chronic rhinosinusitis with nasal polyps who underwent treatment at the Department of Otorhinolaryngology between December 2022 and September 2023. Twelve patients with nasal septum deviation, treated during the same period, were recruited as the control group. Nasal polyp tissues in the CRSwNP group were obtained during endoscopic sinus surgery, while normal nasal mucosa samples in the control group were collected from hypertrophic or bullous middle turbinate mucosa during corrective septal procedures. The institutional ethics committee of the First Affiliated Hospital of Chongqing Medical University granted approval for this study (Approval No.2022-K301).

The human bronchial epithelial cell line (BEAS-2B) was obtained from Pronas Biotechnology (China). Primary human nasal epithelial cells (hNECs) were isolated from nasal polyp or middle turbinate mucosa collected during surgery at the Department of Otorhinolaryngology. Tissues were transferred to the laboratory within 30 minutes and processed under sterile conditions to extract epithelial cells using standard enzymatic dissociation protocols. Cells were cultured in Bronchial Epithelial Cell Growth Medium (BEGM) supplemented with necessary growth factors. All cells were maintained at 37°C in a humidified incubator with 5% CO₂.

In screening tool cells, we selected the immortalized bronchial epithelial cell line BEAS-2B based on the following rationale: (1) Nasal polyp epithelium and bronchial epithelium share pathological similarities, both being pseudostratified ciliated columnar epithelia, allowing BEAS-2B to effectively mimic nasal polyp biology; (2) the hNECs exhibit significant limitations, including poor tolerance to lipofectamine transfection, low transfection efficiency, and high cellular heterogeneity. In contrast, BEAS-2B, due to its immortalized nature, overcomes these technical challenges, providing a stable and reproducible model for functional validation. This selection ensures both pathological relevance and experimental feasibility.

Cell migration was assessed using Transwell chambers following protocols reported in our prior publication (23). Briefly, cells were seeded into the upper chambers without Matrigel coating, and migrated cells were fixed, stained, and quantified after 24 h.

BEAS-2B and primary hNECs were subjected to a wound healing assay as previously described (23). In brief, a scratch was created across a confluent monolayer, and cell migration into the wound area was assessed over 24 hours.

The integrity of the epithelial barrier was evaluated using a FITC–dextran (4 kDa; Sigma-Aldrich) translocation assay. Cells were grown on 0.4 μm Transwell filters until a continuous monolayer was established. After rinsing with phosphate-buffered saline (PBS), 500 μL of FITC–dextran solution was applied to the upper (apical) chamber, while 1.5 mL of serum-free medium was added to the lower (basolateral) chamber. Following a 2-hour incubation at 37 °C, 100 μL of the basolateral medium was collected for fluorescence measurement using a microplate reader (excitation at 485 nm, emission at 535 nm). Elevated fluorescence intensity reflected increased paracellular flux, indicative of barrier dysfunction.

To visualize the F-actin cytoskeleton, cells were stained using Actin-Tracker Red-Rhodamine (Phalloidin; Beyotime, China). Cells were cultured on sterile glass coverslips until they reached approximately 75% confluence. After completing the designated treatments, cells were fixed with 4% paraformaldehyde at room temperature for 15 minutes, then permeabilized using 0.1% Triton X-100 for 10 minutes. Subsequently, cells were incubated in Actin-Tracker Red-Rhodamine solution for 30 minutes in the dark to stain F-actin. Nuclear staining was performed with DAPI for 5 minutes. Finally, coverslips were mounted with anti-fade medium, and fluorescence images were acquired using a fluorescence microscope. F-actin distribution and cytoskeletal remodeling were analyzed qualitatively.

Total RNA extraction, cDNA synthesis, and quantitative real-time PCR (qPCR) were performed as previously described in our published study (23). The specific primer sequences used are provided in Supplementary Table 2 .

Western blotting was conducted as previously reported (24), with slight modifications. Briefly, proteins from RIPA lysates were resolved via SDS-PAGE, transferred onto PVDF membranes, and incubated with specific primary antibodies. Antibody information are listed in Supplementary Table 3 .

Cell transfection was performed using Lipofectamine 3000 (Invitrogen, USA). Before the official experiments, a total of 3 siRNAs were applied to perform knockdown, among which the one with the best knockdown efficiency was selected to perform the following functional experiments ( Figure 1D ). For miRNA modulation, miR-199-3p mimics, inhibitors, or corresponding negative controls were transfected to upregulate or suppress endogenous miRNA expression. For overexpression studies, PAK4 expression plasmids or empty vectors were transfected into cells. After 24–48 hours of transfection, cells were harvested for subsequent assays including qPCR, Western blot, or functional analyses. All oligonucleotides and plasmid constructs were obtained from GenePharma. The corresponding sequences are provided in Supplementary Table 4 .

The interaction between miR-199-3p and its predicted targets NEAT1 and PAK4 was examined using a dual-luciferase reporter assay. The wild-type (WT) and mutant (MUT) 3′-UTR sequences of NEAT1 and PAK4 were cloned into the pmirGLO dual-luciferase vector, which expresses both firefly and Renilla luciferase. HEK293T cells were seeded in 24-well plates (1.5 × 10² cells/well) and cultured overnight to reach ~70% confluence. Cells were co-transfected with 200 ng of reporter plasmid and 50 nM of miR-199-3p mimic or negative control (miR-NC) using Lipofectamine 3000 (Invitrogen) according to the manufacturer’s instructions. To ensure transfection efficiency and account for experimental variability, Renilla luciferase activity (expressed from the same pmirGLO vector) was used as an internal control. Cells were incubated for 48 hours post-transfection under standard culture conditions (37 °C, 5% CO₂). Luciferase activity was measured using the Dual-Luciferase^® Reporter Assay System (Promega) on a microplate luminometer (Promega GloMax). Firefly luciferase signals were normalized to Renilla luciferase activity, and results were presented as relative luciferase activity (firefly/Renilla).

BEAS-2B cells were transfected with si-NEAT1 or si-NC and harvested after 48 h. RNA from three replicates was extracted for sequencing, while one was used to confirm knockdown efficiency by qPCR. RNA-Seq was conducted by BGI Genomics using the DNBSEQ platform (SE50 mode). DEGs were identified using DESeq2 (|log₂FC| ≥ 0.4, adjusted p < 0.05). Functional enrichment analyses, including GO and KEGG pathway analysis, were performed using clusterProfiler. Candidate downstream targets of NEAT1 were selected based on expression trends and functional relevance, with PAK4 identified for further validation.

Twelve SPF-grade female BALB/c mice (4 weeks old, 18–22 g) were randomly divided into control and model groups (n = 6). To induce nasal polyps, mice in the model group received intraperitoneal OVA (25 μg) with aluminum hydroxide (2 mg) on days 0 and 5, followed by intranasal 3% OVA alone or with SEB (10 ng) for 90 days. Controls received PBS on the same schedule. Mice were euthanized 24 h after the final exposure, and nasal mucosa was collected, fixed in 4% paraformaldehyde, paraffin-embedded, and sectioned (4 μm) for subsequent analysis(Approval No. IACUC-CQMU-2023-0249).

Paraffin-embedded nasal sections (4 μm) were deparaffinized, rehydrated, and stained with hematoxylin and eosin to visualize cellular and tissue structures. After dehydration and mounting, slides were examined under a light microscope to assess histology and inflammation.

Paraffin-embedded nasal sections (4 μm) were deparaffinized, rehydrated, and subjected to antigen retrieval in citrate buffer (pH 6.0). After blocking with 3% hydrogen peroxide, sections were incubated overnight at 4 °C with primary antibodies, followed by HRP-conjugated secondary antibodies. DAB was used for visualization, and nuclei were counterstained with hematoxylin. Slides were dehydrated, mounted, and examined microscopically.

Cells grown on glass coverslips were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100. After blocking with 5% BSA, samples were incubated overnight at 4°C with primary antibodies, followed by fluorophore-conjugated secondary antibodies (Alexa Fluor 488 or 594) for 1 h in the dark. Nuclei were stained with DAPI, and coverslips were mounted with anti-fade medium. Images were captured using a fluorescence microscope.

Potential miRNAs interacting with lncRNA NEAT1 were predicted using the ENCORI database (https://rnasysu.com/encori/) (25), which integrates CLIP-Seq-supported RNA–RNA interaction data. Candidate miRNAs targeting PAK4 were identified using TargetScanHuman (https://www.targetscan.org/vert_72/) (26) and miRDB (https://mirdb.org/) (27). Overlapping miRNAs predicted by multiple databases were selected for further validation in subsequent functional assays.

Data are presented as mean ± standard deviation (SD). Statistical analyses were conducted using GraphPad Prism 9.0. Comparisons between two groups were made using unpaired Student’s t-test. Pearson correlation analysis was used to assess the linear relationship between the expression levels of NEAT1, miR-199-3p, and PAK4. A p-value < 0.05 was considered statistically significant.

To evaluate NEAT1 expression in CRSwNP, fluorescence in situ hybridization (FISH) was performed on nasal mucosal tissues from CRSwNP patients and healthy controls. NEAT1 expression was markedly increased in CRSwNP epithelial cells and predominantly localized in the cytoplasm ( Figures 1A, B ), suggesting a potential post-transcriptional regulatory role. This finding was further confirmed by qPCR analysis in an expanded sample set (n = 32 CRSwNP, n = 12 control), which showed significantly higher NEAT1 expression in the CRSwNP group ( Figure 1C ). To investigate the biological function of NEAT1, siRNA-mediated knockdown was performed in BEAS-2B cells and hNECs qPCR confirmed efficient suppression of NEAT1 expression. Before the official experiments, a total of 3 siRNAs were applied to perform knockdown, among which the one with the best knockdown efficiency was selected to perform the following functional experiments. To ease the concern of off-target issues, here we displayed the experimental results obtained by using the other two siRNA ( Figures 1F, G ). Functional assays revealed that NEAT1 knockdown significantly enhanced epithelial barrier integrity, as evidenced by reduced FITC-dextran permeability in both BEAS-2B and hNECs ( Figure 1E ). Moreover, wound healing ( Figure 1F ) and Transwell migration assays ( Figure 1G ) demonstrated a marked decrease in cell motility following NEAT1 knockdown.

Since epithelial-mesenchymal transition (EMT) is closely linked to cytoskeletal remodeling and enhanced migration, we evaluated F-actin distribution using Actin-Tracker Red-Rhodamine staining after TGF-β1 stimulation. TGF-β1 induced prominent stress fiber formation and spindle-shaped morphology, indicative of EMT, whereas NEAT1 knockdown attenuated these changes, restoring a cortical actin distribution and epithelial morphology ( Figure 1H ).

To further confirm the effect of NEAT1 on EMT at the molecular level, immunofluorescence, Western blot, and qPCR analyses were performed. Knockdown of NEAT1 reversed TGF-β1-induced EMT marker expression, as evidenced by increased E-cadherin protein levels and decreased N-cadherin, vimentin, Snail, and Slug at both mRNA and protein levels ( Figures 1I–K ).

To explore the downstream mechanisms of NEAT1 in epithelial–mesenchymal transition (EMT), transcriptome sequencing was performed in BEAS-2B cells following NEAT1 knockdown. After outlier exclusion, differential gene expression analysis revealed predominant downregulation of genes and significant enrichment in the focal adhesion pathway, which is closely associated with cytoskeletal remodeling and cell–matrix interactions ( Figures 2A–C ). Among the downregulated genes, PAK4 showed the most pronounced decrease, as validated by qPCR ( Figure 2D ), suggesting its potential role as a key mediator of NEAT1-driven EMT. In addition to PAK4, qPCR validation of selected genes from the enriched KEGG pathways confirmed significant downregulation of LAMA5,PPP1R12C,following NEAT1 knockdown ( Supplementary Materials ). These genes are involved in focal adhesion and cytoskeletal regulation, further supporting a broader role for NEAT1 in epithelial remodeling.

To investigate regulatory cues affecting NEAT1, BEAS-2B cells were treated with inflammatory stimuli representing different CRSwNP endotypes. IL-13, a Th2 cytokine central to eosinophilic nasal polyps (ENP), significantly upregulated NEAT1 expression in a time-dependent manner ( Figures 2E, F ). Furthermore, IL-13 stimulation increased both NEAT1 and PAK4 expression, while NEAT1 knockdown suppressed IL-13-induced PAK4 upregulation at both mRNA and protein levels ( Figures 2G, H ). Pearson correlation analysis confirmed a strong positive correlation between NEAT1 and PAK4 expression under IL-13 stimulation (R = 0.94, p = 0.0052; Figure 2I ).

To validate these findings in vivo, a murine nasal polyp model was established using OVA/SEB sensitization. HE staining confirmed characteristic polypoid lesions ( Figure 2J ). qPCR and FISH analyses showed significant upregulation of NEAT1 in nasal tissues from model mice ( Figures 2K, M ), while IHC and qPCR revealed concurrent overexpression of PAK4 ( Figures 2L, N ). Additionally, immunohistochemical analysis of clinical samples revealed that PAK4 expression was higher in nasal polyp tissues compared to normal nasal mucosa ( Supplementary Figure 1 ). These results support a functional NEAT1–PAK4 axis in CRSwNP pathogenesis, particularly in Th2-driven ENP.

To assess whether PAK4 mediates the effects of NEAT1 on epithelial dynamics, we overexpressed PAK4 in BEAS-2B cells with NEAT1 knockdown. Wound healing and Transwell migration assays showed that PAK4 overexpression significantly restored the reduced migration capacity caused by NEAT1 silencing, as indicated by increased wound closure and a higher number of migrated cells ( Figures 3A–D ).

Since NEAT1 knockdown was previously shown to inhibit cytoskeletal remodeling in response to TGF-β1, we examined whether PAK4 overexpression could reverse this effect. Phalloidin staining revealed that while NEAT1 knockdown suppressed stress fiber formation and promoted cortical actin distribution, PAK4 overexpression restored cytoplasmic stress fiber assembly and a mesenchymal-like morphology ( Figure 3E ). Furthermore, PAK4 overexpression attenuated the barrier-protective effect of NEAT1 knockdown, as shown by increased FITC-dextran permeability ( Figure 3F ).

At the molecular level, qPCR and Western blot analyses demonstrated that PAK4 overexpression reversed NEAT1 knockdown-induced changes in EMT markers. Specifically, E-cadherin expression decreased, while N-cadherin, Vimentin, Snail, and Slug levels increased, consistent with EMT activation ( Figures 3G, H ). These findings support that PAK4 acts downstream of NEAT1 to promote EMT, cytoskeletal reorganization, and cell motility in nasal epithelial cells.

Given NEAT1’s cytoplasmic localization and its role in upregulating PAK4, we hypothesized that NEAT1 may act as a competing endogenous RNA (ceRNA) by binding and suppressing miR-199-3p. RNA-FISH revealed that NEAT1 was predominantly localized in the cytoplasm of BEAS-2B and hNECs, and its expression was significantly induced by TGF-β1 stimulation ( Figure 4A ). In addition, to further confirm the authenticity of the cytoplasmic-enrichment of NEAT1, qPCR was performed on the nuclear/cytoplasmic RNA extracts in BEAS-2B cells before/after TGF-β1(10 ng/ml 48h) stimulation, and the results indicated that NEAT1 was originally localized in nuclei, but the expression of NEAT1 in cytoplasm at transcriptional level was increased after TGFb1 stimulation, and the difference in nucleus/cytoplasm ratio of NEAT1 levels before/after stimulation was statistically significant (p = 0.003) ( Figure 4B ). Integrated miRNA target prediction using ENCORI, TargetScan, and miRDB identified miR-199a/b-3p as a shared miRNA candidate predicted to bind both NEAT1 and PAK4 ( Figure 4C ). Given that miR-199a-3p and miR-199b-3p share identical mature sequences, we referred to this functional miRNA as miR-199-3p in subsequent experiments. qPCR analysis confirmed that miR-199-3p was significantly downregulated in CRSwNP tissues compared to controls ( Figure 4D ), and Pearson correlation analysis showed a strong inverse correlation between NEAT1 and miR-199-3p expression (R = −0.9043, p = 0.0001) ( Figure 4E ).

Dual-luciferase assays validated the direct interaction between NEAT1 and miR-199-3p. Co-transfection of miR-199-3p mimic significantly suppressed luciferase activity in cells carrying the NEAT1-WT reporter but not the mutant (NEAT1-MUT), confirming direct binding ( Figure 4F ). Transfection of miR-199-3p mimic or inhibitor successfully modulated miR-199-3p expression in BEAS-2B and hNECs ( Figures 4G, H ), and inversely regulated PAK4 expression at both the mRNA and protein levels ( Figures 4H–J ). Finally, miR-199-3p mimic significantly reduced luciferase activity of the PAK4-WT 3′UTR reporter, but not the mutant construct (PAK4-MUT), confirming PAK4 as a direct downstream target of miR-199-3p ( Figure 4L ).

These results demonstrate that NEAT1 regulates PAK4 by acting as a ceRNA to sponge miR-199-3p, thereby promoting EMT and contributing to CRSwNP pathogenesis.

To assess the functional role of miR-199-3p in counteracting PAK4-mediated EMT, we performed a series of rescue assays in BEAS-2B cells co-transfected with PAK4 overexpression vector and miR-199-3p mimic. Wound healing and Transwell assays demonstrated that PAK4 overexpression significantly enhanced the migratory capacity of BEAS-2B cells, as evidenced by increased wound closure rate and the number of migrated cells. However, co-transfection with miR-199-3p mimic effectively suppressed this enhancement, reducing both wound healing and migration to near baseline levels ( Figures 5A–D ). These findings suggest that miR-199-3p mitigates the pro-migratory effects induced by PAK4. To evaluate epithelial barrier function, we conducted a FITC-dextran permeability assay. PAK4 overexpression markedly increased paracellular dextran flux, indicating compromised barrier integrity. This effect was significantly reversed by co-expression of miR-199-3p, which restored barrier function to levels comparable to controls ( Figure 5E ). Phalloidin staining was used to assess actin cytoskeletal changes associated with EMT. Cells overexpressing PAK4 exhibited enhanced stress fiber formation and increased cytoplasmic F-actin accumulation, indicative of cytoskeletal remodeling. Notably, miR-199-3p mimic co-expression reversed these changes, leading to reduced stress fiber density and a return to cortical actin distribution, consistent with an epithelial phenotype ( Figure 5F ). these results demonstrate that miR-199-3p can effectively reverse the pro-EMT phenotypes induced by PAK4, including enhanced migration, impaired barrier function, and cytoskeletal reorganization.

Next, we examined whether miR-199-3p also mediates NEAT1-driven phenotypes. Knockdown of NEAT1 suppressed cell migration and barrier dysfunction, as shown by decreased wound healing, reduced Transwell migration, and lower FITC-dextran permeability. These effects were partially reversed by miR-199-3p inhibition, suggesting that miR-199-3p is required for NEAT1-mediated EMT activation ( Figures 6A–E ).

At the molecular level, qPCR and Western blot analyses confirmed that NEAT1 knockdown downregulated PAK4 and mesenchymal markers (N-cadherin, Vimentin, Snail, Slug), while upregulating E-cadherin, consistent with EMT suppression. These effects were reversed by miR-199-3p inhibition, which restored PAK4 expression and EMT marker levels ( Figures 6F, G ).

Together, these findings demonstrate that miR-199-3p is a critical intermediate in the NEAT1/PAK4 regulatory circuit, and highlight the NEAT1/miR-199-3p/PAK4 axis as a key modulator of EMT and epithelial barrier dysfunction in nasal epithelial cells.