Exosomal microRNAs of tuberculous pleural effusion orchestrating TGF-β signaling mediate pleural fibrosis

Abstract

Tuberculosis pleural effusion (TPE) accounts for 30 ~ 80% of all pleural effusions in developing countries and remains a major contributor to global morbidity and mortality. Although TPE is often absorbed after anti-tuberculosis treatment, approximately 60% of patients develop pleural thickening and pleural fibrosis. However, the underlying mechanisms of pleural fibrosis followed TPE remain poorly understood. In this study, firstly we found that exosomes isolated from TPE were associated with pleural fibrotic changes both in vivo and in vitro, accompanied by activation of the TGF-β signaling pathway. Disruption of exosomes or inhibition of exosome biogenesis attenuated these fibrotic responses. miRNA profiling revealed a distinct exosomal miRNA signature in TPE compared with transudative pleural effusion. Among these miRNAs, down-regulated miR-135b-5p was associated with increased SDCBP expression in pleural mesothelial cells (PMCs), while up-regulated miR-150-3p correlated with enhanced expression of the integrin ITGB6. In addition, miR-503-5p, miR-25-3p, miR-92a-3p and miR-424-3p were linked to reduced Smad7 and Smurf1 expression and decreased ubiquitination of TGF-βRI, collectively favoring TGF-β pathway. Together, these findings suggested that TPE-derived exosomal miRNAs converged on key regulatory nodes of TGF-β signaling and contributed to pleural fibrogenesis. This study provides mechanistic insight into post-tuberculous pleural fibrosis and supports exosomal miRNAs profile as a promising target for prevention and treatment of pleural fibrosis.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12931-026-03541-5.

Introduction

According to the WHO’s Global Tuberculosis Report 2022, the global incidence of tuberculosis increased by 3.6% in 2021 compared with 2020, reversing a long-standing trend of an approximately 2% annual decline over the past two decades [1]. Pleural tuberculosis is the second most common type of tuberculosis after pulmonary tuberculosis, and its main manifestations include pleuritis and pleural effusion. Tuberculosis pleural effusion (TPE) accounts for less than 1% of all exudative effusions in developed countries [2]. However, in developing countries, it is responsible for 30 ~ 80% of all pleural effusions, and thus it remains a major contributor to global morbidity and mortality [3, 4].

Pleural fibrosis, similar as fibrosis in other organs, is characterized by excessive deposition of extracellular matrix components within the pleura, leading to disruption of normal pleura architecture and impaired function. It commonly develops following pleural effusions including TPE, organised haemorrhagic effusions, and empyema. Clinically, pleural fibrosis may present as localized pleural plaques or diffuse pleural thickening [5–7]. Although TPE can be absorbed after treatment, approximately 60% of patients develop residual pleural thickening or even overt pleural fibrosis [8]. Proinflammatory and profibrotic cytokines present in TPE, particularly transforming growth factor-beta1 (TGF-β1), have been implicated in the pathogenesis of pleural fibrosis [9–12]. Our previous studies demonstrated that TPE directly induced fibrotic changes in vitro and in vivo [13, 14], and that activation of TGF-β signal pathway played a central role in this process [14, 15].

Exosomes are small extracellular vesicles with diameters of 30–150 nm. Since their discovery was more than three decades, exosomes have evolved from being considered cellular waste products to being recognized mediators of intercellular communication [16, 17]. They are released by a wide variety of cell types and can be isolated from numerous biological fluids, including blood, urine, breast milk, amniotic fluids, and bronchial alveolar lavage [18]. Increasing evidence indicates that exosomes play important roles in the pathogenesis, diagnosis and targeted therapies of chronic lung diseases, such as chronic obstructive pulmonary disease (COPD), cystic fibrosis, asthma and idiopathic pulmonary fibrosis (IPF) [19]. In IPF, for example, exosomes derived from induced-sputum have been proposed as promising disease biomarkers [20]. However, the role of exosomes in pleural fibrosis remains largely unexplored.

Exosomes carry diverse bioactive cargoes, including proteins, lipids, and nucleic acids, which can be transferred to recipient cells and modulate their biological behavior [18, 21, 22]. Among these cargoes, nucleic acids include mRNAs, microRNAs (miRNAs), and other non-coding RNAs [18]. miRNAs are small non-coding RNAs that negatively regulate gene expression at the post-transcriptional level. Exosomes have been proposed as important vehicles for miRNA-mediated intercellular communication and as sources of miRNA biomarkers in body fluids [23]. Exosomal miRNAs have also attracted attention as diagnostic biomarkers and therapeutic targets in cancer and other diseases [22]. In fibrotic disorders, exosomal miRNAs have been reported to contribute to renal, liver, pulmonary, and cardiac fibrosis [24–27]. Though delivery of specific miRNAs, exosomes can activate fibrotic effector cells, enhance expression of profibrotic mediators, and trigger key signaling pathways, particularly the TGF-β signaling pathway, thereby participating in the development of fibrosis [28]. These observations suggest that exosomal miRNAs may play a role in pleural fibrosis.

TGF-β1 is the most potent profibrotic cytokine and undergoes multiple regulatory steps from its latent to active form, with integrins playing critical roles in this process. Upon ligand binding, the type II TGF-β receptor (TGF-βRII) recruits and phosphorylates the type I TGF-β receptor (TGF-βRI), which in turn phosphorylates receptor-regulated Smads, including Smad2, Smad3, Smad1, Smad5, and Smad8. Smad4 then forms a complex with Smad2/3 to regulate transcription of profibrotic genes and promote fibroblast-to-myofibroblast differentiation [29, 30]. In parallel, TGF-β1 induces the expression of inhibitory Smads, Smad6 and Smad7, as part of a negative feedback mechanism. Smad6 and Smad7 antagonize TGF-β signaling by competing with Smad4 or TGF-βRI, while Smurf1 and Smurf2 E3 ubiquitin ligases regulate the stability of TGF-β receptors and Smads through ubiquitination [30]. Numbers miRNAs have been shown to modulate TGF-β signaling by targeting these components [28, 30]. Thus, interactions among exosomes, miRNAs, and the TGF-β/Smad pathway are likely to be involved in the development of pleural fibrosis and warrant further investigation.

In the present study, we isolated exosomes from TPE and examined their effects on pleural fibrosis in vivo and in vitro, with a particular focus on the role of exosomal miRNAs and TGF-β signaling.

Materials and methods

Reagents and antibodies

Recombinant human IL-27, TNF-α and IFN-γ protein were purchased from R&D Systems (MN, USA). ERK1/2 inhibitor (ERK-IN-3) and exosome biogenesis inhibitor (GW4869) were purchased from MedChemExpress (NJ, USA). Carbon particles were obtained from Mitsubishi Chemical Corporation (Tokyo, Japan). Bleomycin was obtained from Tianjin Taihe Pharmaceutical Co., Ltd. (Tianjin, China). Bacillus Calmette-Guerin (BCG) was obtained from Becton, Dickinson and Company (NJ, USA). Antibodies used for western blot analysis were as follows: anti-CD63 (Cat#67605-1-Ig, Proteintech), anti-TSG101 (Cat#ab30871, Abcam), anti-GAPDH (Cat#60004-1, Proteintech), anti-COL1A1 (Cat#14695-1-AP, Proteintech), anti-TGF-β1 (Cat#MAB240, R&D Systems), anti-phospho-smad2/3 (p-Smad2/3) (Cat#8828, Cell Signaling Technology), anti-smad2/3 (Cat#8685, Cell Signaling Technology), anti-CD9 (Cat#20597-1-AP, Proteintech), anti-SDCBP (Cat#68096-1-Ig, Proteintech), anti-β-actin (Cat#81115-1-RR, Proteintech), anti-phospho-ERK1/2 (p-ERK1/2) (Cat#8544, Cell Signaling Technology), anti-Fibronectin (FN) (Cat#15613-1-AP, Proteintech), anti-Vimentin (Cat#10366-1-AP, Proteintech), anti-E-cadherin (Cat#20874-1-AP, Proteintech), anti-ACTA2 (Cat#14395-1-AP, Proteintech), anti-CK8 (Cat#17514-1-AP, Proteintech), anti-ELK1 (Cat#27420-1-AP, Proteintech), anti-ITGB6 (Cat#28378-1-AP, Proteintech), anti-SMURF1 (Cat#55175-1-AP, Proteintech), anti-TGFBR1 (Cat#30117-1-AP, Proteintech), anti-SMAD7 (Cat#25840-1-AP, Proteintech), anti-Ubiquitin (Cat# 20326, Cell Signaling Technology), anti-Flag (Cat#M185-3, MBL), HA (Cat#M180-3, MBL).

Clinical samples

All experiments involving clinical specimens were approved by the Institutional Review Board of the Tongji Medical College, Huazhong University of Science and Technology (2020-IEC-J-187). Informed consent was obtained from all patients. Pleural effusion samples were collected from 50 patients (15 women and 35 men) with tuberculous pleural effusion and 10 patients (5 women and 5 men) with transudative pleural effusion. The patient’s age ranges from 18 to 85 years old. Tuberculous pleural effusion was determined following Light’s criteria. The collected specimens were immediately immersed in ice and then centrifuged at 300 g for 10 min at 4 °C. The resulting supernatants were aliquoted and stored at -80 °C for further experiments. Clinical characteristics of patients are shown in Table S1.

Isolation and identification of exosomes from pleural effusion

Pleural effusion samples were centrifuged at 300 g for 10 min at 4 °C. The supernatant was collected and subjected to four consecutive rounds of centrifugation (2,000 ×g for 10 min, 12,000 ×g for 30 min, 100,000 ×g for 70 min and 100,000 ×g for 70 min). After ultracentrifugation (Instrument was from Beckman Coulter, USA), the pellet was suspended in 50 µl of PBS buffer. The concentration and size of exosomes derived from pleural effusion was evaluated using nanoparticle tracking analysis (NTA) (Particle Metrix, GER). The morphology of exosomes was observed using a transmission electron microscope (TEM) (HT7700, Hitachi, Japan). The exosomal markers were detected through western blotting.

Isolation and identification of Jurkat cells-derived exosomes

Exosomes were obtained from culture supernatant of Jurkat cells after five consecutive rounds of centrifugation (300 ×g for 10 min, 2,000 ×g for 10 min, 10,000 ×g for 30 min, 100,000 ×g for 90 min and 100,000 ×g for 90 min). After ultracentrifugation, the pellet was suspended in 50 µl of PBS buffer. The exosomes were identified as exosomes from pleural effusion.

Cell line cells cultureCe

The human pleural mesothelial cell (PMC) line Met-5 A cells, and human T-cell line Jurkat cells were purchased from the American Type Culture Collection (ATCC, VA, USA). The PMCs and Jurkat cells were cultured in RPMI 1640 medium (Gibco, NY, USA) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 µg/ml streptomycin. HEK293T cells were cultured in DMEM medium (Gibco, NY, USA) supplemented with 10% FBS. All cultures were maintained at 37 °C in 5% CO2 humidified atmosphere. Cells were subcultured at a ratio of 1:3 and the culture medium was changed every 3 days.

Isolation and culture of rat primary PMCs

Primary rat pleural mesothelial cells (PMCs) were isolated using protease as our previous studies (15). Briefly, the thorax isolated under sterile conditions and injected with 1 mg/ml protease from streptomyces griseus (Sigma, MI, USA) in 5 ml RPMI-1640 medium. It was then digested at 4 °C overnight. The cells were harvested and then centrifuged at 1,500 rpm for 5 min. Rat PMCs were suspended in epithelial cell medium supplemented with 2% FBS, 1% epithelial cell growth factor (ScienCell, CA, USA), and 100 U/ml penicillin and 100 µg/ml streptomycin (ScienCell, CA, USA). The cells grew to confluence at 7 days and were then replaced with RPMI-1640 containing 20% FBS. Cells equilibrated overnight in medium containing 1% FBS were used for all experiments.

Exosomes uptake assay

Met-5 A cells were grown to 70% confluence and cultured in serum-free RPMI 1640 medium for 12 h. The cellular uptake of exosomes was assessed using PKH67 Fluorescent Cell Linker Kits (Sigma Aldrich, USA) according to the manufacturer’s instructions. Exosomes were labeled with PKH67 and incubated at 37℃ for 6 h. PMCs were fixed with paraformaldehyde (4%) and permeabilized with TritonX-100 (0.5%). Following treatment, the cytoskeleton in PMCs was stained for F-actin for 30 min. The nuclei were stained for DAPI for 10 min in the dark. The samples were examined using confocal microscopy.

Western blot assay for total cell lysates

Western blotting was performed as standard, and indicated proteins were probed with primary antibodies (listed as ‘indicated protein’ (dilution factor)): anti-CD63 (1:3000), anti-TSG101 (1:3000), anti-GAPDH (1:10000), anti-COL1A1 (1:1000), anti-TGF-β1 (1:1000), anti-p-Smad2/3 (1:500), anti-smad2/3 (1:1000), anti-CD9 (1:1000), anti-SDCBP (1:2000), anti-β-actin (1:2000), anti-p-ERK1/2 (1:500), anti-FN (1:1000), anti-Vimentin (1:1000), anti-E-cadherin (1:1000), anti-ACTA2 (1:1000), anti-CK8 (1:1000), anti-ELK1 (1:1000), anti-ITGB6 (1:1000), anti- SMURF1 (1:1000), anti-TGFBR1 (1:1000 ), anti-SMAD7 (1:1000).

Ultrasound broken down exosomes

The exosomes derived from pleural effusion were suspended in 50 µl of PBS buffer and sonicated using an ultrasonic machine (Uibra Cell, USA) for 1 min (parameters: 130 W, 20HZ, Amp1 30%). The sonicated exosomes were then stored at -80 °C.

RNA extraction and quantificational real-time PCR (qRT-PCR)

After treatment, total RNA was extracted using TRIzol reagent (Vazyme, China) according to the manufacturer’s protocol. Reverse transcription of 1 µg of total RNA to complementary DNA (cDNA) was performed with the HiScript II 1st Strand cDNA Synthesis Kit (Vazyme, China). qRT-PCR was performed using Cham Q SYBR qPCR Master Mix (Vazyme, China) on the Quantagene q225 (Kubo Technology, China). Target gene expression levels were normalized to GAPDH expression. The 2-ΔΔCt method was used to calculate the expression levels. The primer sequences are listed in online supplementary Table S2.

Quantification of MiRNA

Total RNA was extracted using TRIzol reagent (Vazyme, China). RNA was reversely transcribed into miRNA complementary DNA (cDNA) using the miRNA 1st Strand cDNA Synthesis Kit (Vazyme, China). Real-time PCR was performed on the Quantagene q225 (Kubo Technology, China) using miRNA Universal SYBR qPCR Master Mix (Vazyme, China). U6 snRNA was used as an internal control. The primer sequences for miRNA can be found in Online Supplementary Table S2.

Cell proliferation assay

Cell proliferation was assessed using CCK-8 assays. PMCs were trypsinized and centrifuged. The cells were seeded at a density of 2 × 10³ cells per well in 96-well plates. The Cell Counting Kit-8 (Dojindo, Japan) was used to measure cell proliferation following the manufacturer’s instructions.

Histological analysis

Mouse specimens were preserved by immersion in 4% paraformaldehyde for 48 h. Subsequently, specimens were dehydrated and embedded in paraffin. After drying, xylene was used to remove the paraffin, followed by specimen rehydration. For morphological analysis and evaluation of collagen deposition, the 5 μm thick tissue sections were stained with Trichrome and Sirius Red for subsequent analysis. The slides were examined using a microscope connected to a digital camera. Images were analyzed using Image-Pro Plus software.

Immunohistochemical staining

The 5 μm thick paraffin-embedded sections of mouse tissue were deparaffinized and rehydrated. Antigen retrieval was performed by boiling Citrate Antigen Retrieval Solution (G1201, Servicebio) for 15 min. Standard immunohistochemistry staining was carried out using the UltraSensitiveTM SP IHC Kit (KIT-9710, MXB Biotechnologies, China) following the manufacturer’s instructions. The sample sections were treated with primary antibodies, specifically anti-ELK1 (1:1500) and anti-ITGB6 (1:1500), at 4 °C overnight. Biotinylated secondary antibodies were then applied for 20 min. The slides were developed using DAB working solution, counterstained with hematoxylin, and mounted with mounting medium. Finally, the slides were scanned using a microscope connected to a digital camera, and images were analyzed using Image-Pro Plus software.

KEGG pathway enrichment analysis

The annotation for differentially expressed miRNA pathways was obtained from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. Fisher’s exact test was then used for pathway enrichment analysis in KEGG.

Immunofluorescence staining

To determine intracellular localization of ELK1, ITGB6, PMCs were fixed with 4% paraformaldehyde for 20 min and then washed with PBS three times. Cells were exposed to 0.1% Triton X100 for 10 min and blocked with 5% BSA in PBS for 40 min. Then cells were stained using antibodies against ELK1 (dilution 1:200) and ITGB6 (dilution 1:200) at 4℃ overnight and then with FITC-labeled secondary antibody IgG (SA00003, Proteintech; dilution 1:2000) for one hour. The nuclei in PMCs were stained for DAPI for 5 min in dark. The fluorescence images were obtained using laser-scanning confocal microscopy (LSM 780, Zeiss, Germany). Images were acquired using the same exposure settings and further processed with ZEN software (Carl Zeiss).

miRNA sequencing and analysis

Exosomes from pleural effusion were obtained by ultracentrifugation of samples isolated from nine control individuals and nine patients with tuberculous pleurisy patients and processed by Novogene Co., Ltd (Beijing, China) for miRNA microarray analysis (lllumina SE50). Differential miRNAs were identified with follow criteria: adjusted p-value less than 0.05 and a minimum fold change of 1. The raw data were deposited in GEO under submission number GSE172150. Statistical methods ANOVA was employed to identify miRNAs that exhibited significant differences.

Nano-flow cytometry analysis

Exosomes from pleural effusion were analysed by the flow nanoanalyser instrument (nanoFCM). Exosomes were labeled with 0.20 µg of FITC-conjugated CD3 (#317315, BioLegend, USA) for 60 min at room temperature. Fluorescently‐labelled samples were washed with PBS and purified using the Beckman Coulter Ultracentrifuge at 100,000 × g at 4 °C for 90 min. The pellets were then resuspended in PBS and measured at 1 kPa sampling pressure, 10% SS decay and 10 mW laser power. The data were analysed by FlowJo software.

miRNAs overexpression and Inhibition

Unique miRNA mimics, inhibitors and non-targeting negative control (NC) oligonucleotides were synthesized by RiboBio (Guangzhou, China). miRNA mimics (50 nM/L) and negative control (50 nM/L) were transfected into PMCs using Lipofectamine RNAiMAX (Invitrogen, USA) according to the manufacturer’s protocol. miRNA inhibitors (100 nM/L) and NC (100 nM/L) were transfected into PMCs using Lipofectamine RNAiMAX transfection reagent (Invitrogen, USA) following the manufacturer’s instructions. The sequences of miRNA mimics and inhibitors are listed in Online Supplementary Table S2.

Viral constructs with lentivirus vectors encoded lenti_OE-miR-150-3p blending pri-mmu-miR-150-3p, and GFP protein was designed to induce overexpression of miR-150-3p, and lenti_OE-scramble with only GFP protein as a negative control virus controlled by the CMV promoter. Recombinant lentivirus encoding a siRNA targeted against mmu-miR-150-3p was used to knockdown miR-150-3p in animal models. Control lentiviruses expressed scrambled sequence RNAs. All recombinant lentivirus vectors were made by Genechem (Shanghai, China). The sequences for mmu-miR-150-3p overexpression and inhibition can be found in online supplementary Table S2.

Dual luciferase reporter assay

The correlation between miRNAs and their target genes was investigated using dual-luciferase reporter system. The wild-type target gene 3′ UTR sequence containing the putative seed binding sequence for miRNA was synthesized and sub-cloned into pMIR-Report luciferase. A control was generated by mutating of the miRNA binding sequence in the target gene 3′ UTR. The plasmids were transfected into HEK293T cells using lipofectamine 2000 (Invitrogen, USA) according to the manufacturer’s instructions. Furthermore, dual-Luciferase activity was determined using a dual-luciferase reporter detection system (Promega, USA) with a Synergy 2 Multi-Mode microplate reader (BioTek Instruments). The relative luciferase signals were standardized based on the firefly/renilla ratio.

Fluorescence in situ hybridization (FISH)

To investigate miR-150-3p levels in the human pleura, in situ hybridization histochemistry on pleura tissues was performed. Standard FISH experiment was carried out using the RiboTM Fluorescent in Situ Hybridization Kit (RIBOBIO, China) following the manufacturer’s instructions. The hybridization process utilized Cy3-labeled miR-150-3p probes was from BOSTER (Wuhan, China). After adding blocking buffer, slides were incubated with a biotin-conjugated IgG fraction mouse anti-digoxin antibody, and subjected to streptavidin-biotin complex. The slides were undergoing chromogenic reaction with the DAB kit. Finally, images were analyzed using laser-scanning confocal microscopy (LSM 780, Zeiss, Germany).

Co-immunoprecipitation (Co-IP)

Co-IP assays in PMCs were conducted to study interaction between Smad7 and Smurf1. First, IgG and primary antibodies were cross-linked to protein A + G magnetic beads. Then PMCs were lysed using IP lysis buffer. Next, the cell lysate was incubated with the cross-linked antibody magnetic beads at room temperature for 1 h. The immunocomplexes were analyzed by western blotting after rinsing with lysis buffer. Immunoprecipitants were examined using indicated primary antibodies: anti-Smurf1 (1:1000 for IB, 1:500 for IP), anti-TGF-βRI (1:1000 for IB, 1:500 for IP), anti-Smad7 (1:1000 for IB, 1:500 for IP), HA (1:1000 for IB, 1:500 for IP), Flag (1:1000 for IB), Ubiquitin (1:1000 for IB).

Ubiquitination assays

After treatment, PMCs were pelleted with 1% SDS and boiled at 100 °C for 10 min. Subsequently, the samples were then diluted in TNE buffer with 0.1% SDS for immunoprecipitation. Immunoblotting was used to analyze the lysates and immunoprecipitates.

Exosome-, bleomycin- and TPE- induced pleural fibrosis model and treatments

All animal experiments were conducted following the Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee (IACUC) of the Tongji Medical College, Huazhong University of Science and Technology (approval ID: 2020-S411). Mouse pleural fibrosis models were made using bleomycin, as previously described. Briefly, male C57BL/6J mice (male, 6 ~ 8 weeks of age, 18 ~ 20 g) were kept under specific-pathogen-free conditions. Exosomes derived from 50 mL of pleural effusion were suspended in 100 µl of PBS buffer. Based on the results of Bicinchoninic Acid Assay (BCA) protein assay, different volumes of PBS buffer were added to ensure that the concentration of exosomes for each patient was 1000 µg/ml. Intrapleural injection was at a volume of 100 µl (100 µg) per mouse on days 1, 5, 9. A mixture of bleomycin (0.48 µg/mouse) and carbon particles (0.1 mg/mouse) was administered in a volume of 100 µl saline solution by intrapleural injection on day 1. Or a mixture of TPE and carbon particles (0.1 mg/mouse) was administered in a volume of 100 µl saline solution by intrapleural injection on days 1, 5, and 9. Mice in control group received intrapleural injection with 100 µl PBS buffer or saline solution. All mice were euthanized after 21 days, and their tissues were taken for histological analysis.

BCG- and H37Ra-induced pleural fibrosis model and treatments

BALB/c mice (male, 6 ~ 8 weeks of age, 18 ~ 20 g) were maintained under specific-pathogen-free conditions. Bacillus Calmette- Guerin (BCG) (106 CFU/ml) or heat-inactivated H37Ra (106 CFU/ml) was administered in a volume of 100 µl saline solution by intrapleural injection on day 1. Some mice received intraperitoneal injection of 100 µl exosome biogenesis inhibitor GW4869 (2.5 µg/g) on days 14, 16, 18, and 20. The control group was given intrapleural injection of 100 µl saline solution. After 21 days, all mice were euthanized, and their tissues were taken for histological analysis.

miR-150-3p overexpression in mice

C57BL/6J mice (male, 6 ~ 8 weeks of age, 18 ~ 20 g) were divided into three groups (vector control, vector control+carbon, and miR-150-3p over-expression). Mice in the vector control group received intrapleural injections of vector control lentivirus (2 × 10⁶ TU/mouse, 100 µl/mouse) on days 5, 7 and 10. A mixture of vector control lentivirus (2 × 10⁶ TU/mouse) and carbon particles (0.1 mg/mouse) was administered in a volume of 100 µl saline solution by intrapleural injection to the vector control plus carbon group at the same times. The miR-150-3p overexpression group received intrapleural injections of lentivirus expressing miR-150-3p (2 × 10⁶ TU/mouse) and carbon particles (0.1 mg/mouse) on days 5, 7, and 10. All mice were euthanized after 21 days, and tissues were taken for histological and immunological fluorescence staining.

miR-150-3p inhibitor in mice

C57BL/6J mice (male, 6 ~ 8 weeks of age, 18 ~ 20 g) were divided into three groups (vector control+carbon, miR-150-3p inhibitor+carbon, miR-150-3p inhibitor+ carbon + TPE). In the first two groups, a mixture of vector control lentivirus or miR-150-3p inhibition lentivirus (2 × 10⁶ TU/mouse) and carbon particles (0.1 mg/mouse) was administered in a volume of 100 µl saline solution by intrapleural injection at days 11, 13, and 15. A mixture of TPE and carbon particles (0.1 mg/mouse) was administered in a volume of 100 µl saline solution by intrapleural injection on days 1, 5, and 9. Some mice received intrapleural injections of miR-150-3p inhibition lentivirus (2 × 10⁶ TU/mouse, 100 µl/mouse) on days 11, 13, and 15. All mice were euthanized after 21 days, and the tissues were taken for histological and immunological fluorescence staining.

Triple miRNAs inhibitor in mice

A triple miRNAs inhibitor which restrained expressions of miR-150-3p, miR-424-3p and miR-503-5p was designed. C57BL/6J mice (male, 6 ~ 8 weeks of age, 18 ~ 20 g) were intra-pleural injected by using PBS (100 µl/mouse), TPE exosomes (100 µl/mouse), TPE exosomes plus control inhibitor, or TPE exosomes plus triple miRNAs inhibitor with carbon particles (0.1 mg/mouse). Intra-pleural injection was administered at days 1, 5, and 9. All mice were euthanized at day 21, and tissues were taken for analysis.

Statistical analysis

The potential mRNA targets of miRNAs were analyzed using the Targetscan database (http://www.targetscan.org). Expression profiling by miRNAs-seq data was reanalyzed using the R Project (Vension 4.2.2). All other experimental data were analyzed using GraphPad Prism 8. Data were presented as mean ± SEM of n individual experiments. Statistical analyses between two groups were conducted using a two-sided Student’s t test, while comparisons between multiple groups were performed using an ANOVA test. P values < 0.05 were accepted as statistically significant, indicating a significant difference between the two groups or multiple groups.

Results

Exosomes existed in TPE

To uncover whether exosomes exist in TPE, pleural effusion was collected from patients with tuberculous pleurisy. Using gradient ultracentrifugation, exosomes were collected from the pleural effusion (Fig. 1A). As shown in Fig. 1B, transmission electron microscope (TEM) revealed exosomes really existed. The diameter of these exosomes from TPE (TPE-Exosomes) was among 50 ~ 300 nm, and the mean diameter was 127.1 nm (Fig. 1C). Western blotting showed that exosome biomarkers CD63 and TSG101 expressed in these TPE-Exosomes without GAPDH expression (Fig. 1D). To confirm biological activity of these exosomes, the exosomes were labeled PKH67 (with green fluorescence) and used to treat pleural mesothelial cells (PMCs). It was clearly found PMCs took in exosomes at 2 h. Exosomes were almost entered into cells after 6 h (Fig. 1E). Exosomes were also obtained in transudative pleural effusion from patients with heart failure (Fig. S1). However, the exosome concentration in transudative pleural effusion was no more than 1/8 of that in TPE. Furthermore, cell source of TPE-Exosomes was explored. The main cells in TPE are T cells. As shown in Fig. S2, 16.2% TPE-Exosomes expressed CD3, a T cell marker. Then a T line cell, Jurkat cell was cultured. Exosomes in medium of Jurkat cells were detected. TEM, nanoparticle tracking analysis, and western blotting revealed that Jurkat cells produced exosomes (Fig. S2B-D). These data revealed that exosomes existed in TPE, and T cells were one of main sources of exosomes.

Fig. 1.

Isolation and identification of exosomes in tuberculous pleural effusion. A Workflow diagram for isolating exosomes from tuberculous pleural effusion. Exosomes were isolated via ultracentrifugation. B Transmission Electron Microscope (TEM) images of exosomes in tuberculous pleural effusion. C Nanoparticle tracking analysis (NTA) of exosomes in tuberculous pleural effusion. D Western blot analysis for exosome markers CD63 and TSG101 (TPE-Exo: exosomes in tuberculous pleural effusion). (E) Cellular uptake of PKH67-labeled (green) exosomes in human pleural mesothelial cells. Scale bar: 50 μm

TPE-Exosomes activated TGF-β/smad2/3 signaling and promoted collagen I synthesis in vitro

Exosomes entered into PMCs, the role of these exosomes in the cells needed to be investigated. As shown in Fig. S3, exosomes from different patients were used to treat human PMCs. TGF-β1 levels, p-smad2/3 levels and collagen I levels were increased in PMCs which treated by TPE-Exosomes compared with exosomes from transudative pleural effusion. These data suggested that TPE-Exosomes activated TGF-β/smad2/3 signaling and induced PMC fibrotic changes in vitro.

TPE-Exosomes induced mouse pleural fibrosis in vivo

Next, effects of TPE-Exosomes in mouse models were studied. As shown in Fig. 2, mouse pleural fibrosis models were induced by intra-pleural injections of bleomycin plus carbon particles as positive control. At the same time, TPE-Exosomes and exosomes from transudative pleural effusion were used to administrate mice by intra-pleural injections. Pleural fibrosis was found in TPE-Exosomes-treated mice as mice which were treated by bleomycin plus carbon particles. No pleural fibrosis occurred in mice which were administrated with exosomes from transudative pleural effusion. As shown in Fig. 2A-C, collagen deposition, pleural thickness and collagen percentage in pleura obviously increased in TPE-Exosomes-treated mice compared with negative controls.

Fig. 2.

Exosome from TPE induced pleural fibrosis in vivo. C57BL/6 mouse pleural fibrosis model was induced by intra-pleural injections of pleural effusion exosomes, bleomycin and carbon particles as the description in the Methods. TPE-Exo or Trans Exo (100 µl/mouse) from 50 ml pleural effusion or a mixture of bleomycin (0.48 µg/mouse) with carbon particles (0.1 mg/mouse) or 0.9% NaCl (100 µl) was administered by intra-pleural injection at days 1, 5, 9. All mice were euthanized at day 21, and then tissues were taken for analysis. A Representative Masson’s trichrome staining images of visceral pleura from lung sections, parietal pleura from chest wall and diaphragm sections. Original magnification, ×400. B Changes in pleural thickness.C Changes in collagen percentages of visceral and parietal pleura. Data are expressed as mean ± SEM. n = 5 mice. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (One-way ANOVA followed by the Bonferroni’s test)

Blockade of exosome generation inhibited experimental tuberculous pleural fibrosis

To further confirm the role of TPE-Exosomes in pleural fibrosis, TPE-Exosomes were destroyed by ultrasonic in vitro, and then used to treat PMC again. As shown in Fig. 3A, vesicles were destroyed after treatment of ultrasonic. Solution with destroyed-TPE-Exosomes did not induce expression of TGF-β1, p-smad2/3 or collagen I in PMCs (Fig. 3B and C). In the in vivo experiments, exosome biogenesis inhibitor GW4869 was used to decrease exosomes. Bacillus Calmette-Guerin (BCG) induced pleural fibrosis in mice, and this was prevented by GW4869 (Fig. S4). Moreover, mouse pleural fibrosis model was also made by intra-pleural injections of heat-inactivated mycobacterium tuberculosis H37Ra, GW4869 inhibited the pleural fibrosis as well (Fig. S5). These data revealed that blockade of exosome generation inhibited experimental tuberculous pleural fibrosis.

Fig. 3.

Broken down of exosomes from TPE by ultrasonic attenuated COL1A1, TGF-β and p-Smad2/3 expression in human PMCs. A Transmission electron microscope (TEM) images of exosomes before and after destruction by ultrasonic. B, C Human PMCs were incubated with TPE-Exo (100 µg/ml), Ultrasonic Exo (100 µg/ml) and TPE (5%) for 24 h, after which intracellular protein levels of COL1A1, TGF-β, p-Smad2/3 (phosphorylated) and t-Smad2/3 (total) protein was measured by western blotting. Representative western blots (B), and changes in relative density of COL1A1 and TGF-β to GAPDH or p-Smad2/3 to t-Smad2/3 and GAPDH were presented (C). Data are expressed as mean ± SEM. n = 5 (COL1A1), n = 5 (TGF-β), n = 5 (p-Smad2/3). *P < 0.05 (One-way ANOVA followed by the Bonferroni’s test)

Differentially expressed miRNA in TPE-Exosomes and transudative pleural effusion exosomes

Above data indicated that TPE-Exosomes mediated pleural fibrosis in tuberculous pleurisy, the underlying mechanisms needed to be explored. Firstly, micro-RNA (miRNA) in exosomes were detected by miRNA sequencing. As shown in Fig. 4A, differentially expressed miRNA in TPE-Exosomes and exosomes from transudative pleural effusion was analyzed. Compared with exosomes from transudative pleural effusion, there were 247 differentially expressed miRNAs unique to TPE-Exosomes, among which 77 miRNAs up-expressed and 73 miRNAs down-expressed in TPE-Exosomes (Fig. 4B and C). Pathway enrichment analysis revealed that changed-miRNAs in TPE-Exosomes were highly correlated with several pathways including TGF-β signaling pathway (Fig. 4D). Among differentially expressed-miRNAs in TPE-Exosomes, some miRNAs were further detected by qRT-PCR in TPE-Exosomes and exosomes from transudate pleural effusion. As shown in Fig. 4E, decreased-miR-135-5p, was confirmed. On the contrary, miR-150-3p and miR-424-3p increasing was verified.

Fig. 4.

Differentially expressed miRNA in exosomes from tuberculous pleural effusion and transudative pleural effusion (as control) by miRNAs sequencing. A Heatmap of differential exosomes miRNAs expression in pleural effusion between control individuals and tuberculous pleurisy patients; n = 9. RPKM values are represented by gradient colors and shown for each sample. Red represents a higher RPKM; blue represents a lower RPKM. Results are based on nine RNA sequencing samples. B Venn diagram showing the overlap between differentially expressed miRNAs in TPE exosomes and transudate exosomes. C Volcanic map of differential exosomes miRNAs expression in pleural effusion between control individuals and tuberculous pleurisy patients. Adjusted P value < 0.05 and fold change > 1 was set as restrictive conditions to identify the differentially expressed genes. D Pathway enrichment analysis showed the significant target genes of differentially expressed miRNAs associated with various KEGG pathways. E The expression of three differentially expressed miRNAs in TPE exosomes and transudate exosomes was verified by RT-qPCR and normalized by the U6. Data are expressed as mean ± SEM. n = 6, ****P < 0.0001 (Paired student’s t-test)

Decreased miR-135b-5p was associated with increased syntenin (SDCBP) expression and enhanced exosome biogenesis

SDCBP is a protein which mediates exosome production and secretion. If TPE-Exosomes increase SDCBP expression, PMCs will mediate more exosomes generation after they took TPE-Exosomes. Firstly, whether SDCBP increases in tuberculous microenvironment was investigated. Tuberculosis-related cytokines IL-27, TNF-α, IFN-γ, and heat-inactivated mycobacterium tuberculosis H37Ra were used to treat PMCs, SDCBP expression was examined by western blotting. As shown in Fig. S6, H37Ra, IL-27 and TNF-α increased SDCBP synthesis. These results indicated that tuberculosis-related cytokines increased SDCBP.

Bioinformatics analysis revealed that miR-135b-5p should have effect on SDCBP expression. miRNA mimics efficiently exhibit functions of the corresponding miRNA, miR-135b-5p mimics were used. As shown in Fig. S7A, miR-135b-5p mimics increased levels of miR-135b-5p in human PMCs. Wild type miR-135b-5p mimics inhibited transcription of SDCBP gene, mutant of miR-135b-5p failed to inhibit SDCBP gene transcription (Fig. S7B and C). Not surprisedly, miR-135b-5p mimics decreased SDCBP protein synthesis (Fig. S7D). On the contrary, miR-135b-5p inhibitor increased mRNA and protein levels of SDCBP (Fig. S7E-G). Thus, decreased miR-135b-5p was associated with increased SDCBP expression which promoted exosome generation.

miR-150-3p was associated with integrin ITGB6 upregulation and profibrotic responses

miR-150-3p magnificently increased in TPE-Exosomes (Fig. 4C and E), the role of miR-150-3p was studied. Firstly, miR-150-3p expression in parietal pleura of patients with tuberculous pleurisy was examined. In situ hybridization histochemistry (ISH) staining revealed hsa-miR-150-3p mRNA increased in the tuberculous pleural tissue compared with control (Fig. 5A). Then, miR-150-3p mimics was used (Fig. 5B). As shown in Fig. 5C-G, miR-150-3p mimics increased mRNA and protein levels of collagen I, fibronectin and vimentin, decreased E-cadherin and CK8 in PMCs. miR-150-3p mimics promoted cell proliferation of human PMCs (Fig. 5F) and rat PMCs (Fig. 5G). On the contrary, miR-150-3p inhibitor expressed opposite effect to miR-150-3p mimics (Fig. 5H and I). In vivo, overexpression of miR-150-3p with carbon exactly resulted in pleural fibrosis just as bleomycin with carbon (Fig. 6, Fig. S8). TPE with carbon induced pleural fibrosis, and this was prevented by miR-150-3p inhibitor (Fig. S9).

Fig. 5.

miR-150-3p increased in pleura tissue, promoted collagen I, fibronectin, vitmentin protein expressions, and cell proliferation of PMCs. A Pleural tissue samples were collected from parietal pleura in patients with TPE, and performed in situ hybridization histochemistry (ISH) staining to reveal hsa-miR-150-3p mRNA levels. Representative images of miR-150-3p ISH. Scale bar: 100 μm. B Human PMCs were transfected with miR-150-3p mimics (50 nmol/ml) or negative control (NC) for 24 h, after which miR-150-3p expression levels in cells were determined by qRT-PCR and normalized by the U6. C Intracellular protein levels of collagen I (COL1A1), fibronectin (FN), vitmentin, E-cadherin, ACTA2 and CK8 were measured by western blotting. D Bar graphs revealed changes in relative ratio of COL1A1, FN, vitmentin, E-cadherin, ACTA2 and CK8 to GAPDH. E mRNA expressions of COL1A1, FN, vitmentin, E-cadherin, ACTA2 and CK8 were detected by qRT-PCR. F Human PMCs were transfected with miR-150-3p mimics (50 nmol/ml) or NC for 24 h, after which cell proliferations were detected by CCK8 assay. G Rat PMCs were transfected with miR-150-3p mimics or NC for 24 h, after which cell proliferations were detected by CCK8 assay. H Human PMCs were transfected with miR-150-3p inhibitors (100 nmol/ml) or negative control (NC) for 24 h, after which miR-150-3p expression levels in cells were determined by qRT-PCR and normalized by the U6. I mRNA expressions of COL1A1, FN, vitmentin, E-cadherin, ACTA2 and CK8 were detected by qRT-PCR. Data are mean ± SEM. n = 3 (B-E), n = 12 (F-G), n = 3 (H-I). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (unpaired student’s t-test)

Fig. 6.

miR-150-3p over-expression mediated pleural fibrosis in mice. C57BL/6 mice were intra-pleural injected by using miR-150-3p overexpression plus carbon particles as the description in the Methods. The positive control of pleural fibrosis model was induced by intra-pleural injections of bleomycin plus carbon particles. A mixture of bleomycin (0.48 µg/mouse) and carbon particles (0.1 mg/mouse) or 0.9% NaCl (100 µl) was administered by intra-pleural injection at day 1. Mice were treated with mouse miR-150-3p overexpression lentivirus or vector control lentivirus (VC) (2 × 10⁶ TU/mouse) by intra-pleural injection at days 5, 7 and 10. All mice were euthanized at day 21, and then tissues were taken for analysis. A Representative Masson’s trichrome staining images of visceral pleura from lung sections, parietal pleura from chest wall and diaphragm sections. Original magnification, ×400. B Changes in pleural thickness. C Changes in collagen percentages of visceral and parietal pleura. Data are expressed as mean ± SEM. n = 5 (VC group), n = 5 (VC+bleomycin and carbon particles group), n = 6 (VC +carbon particles group), n = 4 (miR-150-3p overexpression lentivirus plus carbon particles group). **P < 0.01, ***P < 0.001 (One-way ANOVA followed by the Bonferroni’s test)

Mechanism of miR-150-3p mediating pleural fibrosis was then explored. miR-150-3p binding to ELK1 gene and inhibited ELK1 protein expression (Fig. 7A-D). By bioinformatics analysis, protein-protein interaction indicated ELK1 inhibited an integrin protein ITGB6 expression (Fig. 7E). Then, effect of miR-150-3p on ITGB6 was examined. As shown in Fig. 7F-I, miR-150-3p inhibitor decreased ITGB6, on the contrary miR-150-3p mimics increased ITGB6. In mouse models, with carbon particle miR-150-3p overexpression down-regulated ELK1 and up-regulated ITGB6 just as bleomycin did (Fig. 7J). These data suggested increased-miR-150-3p in TPE-Exosomes was associated with pleural fibrotic.

Fig. 7.

miR-150-3p promoted pleural fibrosis through activation of ELK1/ ITGB6 signaling. A Predicted hsa-miR-150-3p target sequences in the 3′ UTRs of the ELK1 mRNA. Sequences of hsa-miR-150-3p and the putative target sequence in the ELK1 mRNA (wildtype) or an engineered mutant of this sequence (mutant) and its effect on the luciferase reporter. B Human PMCs were transfected with miR-150-3p inhibitors (100 nmol/ml) or NC for 24 h, after which intracellular protein levels of ELK1 were measured by western blotting. Bar graphs revealed changes in relative ratio of ELK1 to GAPDH. C Human PMCs were transfected with miR-150-3p mimics (50 nmol/ml) or negative control (NC) for 24 h, after which intracellular protein levels of ELK1 were measured by western blotting. Bar graphs revealed changes in the relative ratio of ELK1 to GAPDH. D Human PMCs were incubated with miR-150-3p mimics. After 24 h, ELK1 protein was detected by immunofluorescence staining and nuclei with DAPI staining. Bar scale: 50 μm. E Establishment of the protein-protein interaction (PPI) network. Identification of ELK1-related genes in PPI network. The online analysis software “STRING” (https://string-db.org/) was used. F Human PMCs were transfected with miR-150-3p inhibitors (100 nmol/ml) or NC for 24 h, after which mRNA expression of ELK1 and ITGB6 were detected by qRT-PCR. G Human PMCs were transfected with miR-150-3p mimics or NC for 24 h, after which mRNA expression of ELK1 and ITGB6 were detected by qRT-PCR. H Human PMCs were transfected with miR-150-3p mimics or NC for 24 h, after which intracellular protein levels of ITGB6 were measured by western blotting. Bar graphs revealed changes in relative ratio of ITGB6 to GAPDH. I Human PMCs were incubated with miR-150-3p mimics. After 24 h, ITGB6 protein was detected by immunofluorescence staining and nuclei with DAPI staining. Bar scale: 50 μm. J Mouse pleural fibrosis was induced by intra-pleural injections of miR-150-3p overexpression lentivirus plus carbon particles or bleomycin plus carbon particles. Mice were euthanized at day 21, and then tissues were taken for immunohistochemical (IHC) staining of ELK1 and ITGB6 proteins. Scale bar: 100 μm. Original magnification, ×400. Data are mean ± SEM. n = 7 (A), n = 3 (B, D, G, J), n = 7 (F). *P < 0.05, **P < 0.01 (student’s t-test)

miR-503-5p, miR-25-3p, miR-92a-3p and miR-424-3p activated TGF-β signaling

To confirm which miRNAs in TPE-Exosomes are involved in TGF-β signaling, overlap miRNAs between up-regulated miRNAs in TPE-Exosomes and Smad7- or Smurf1-related miRNAs were predicted. As shown in Fig. 8A and B, miR-503-5p, miR-25-3p and miR-92a-3p were the overlap ones. Then mimics of these three miRNAs were used to treat PMCs. miR-503-5p, miR-25-3p and miR-92a-3p mimics inhibited mRNA and protein expression of Smurf1, increased protein levels of type I receptor of TGF-β (TGF-βRI, Fig. 8C-E). miR-503-5p mimics-treated cells were used to perform additional experiments. As shown in Fig. 8F, miR-503-5p mimics also increased mRNA of collagen I besides changed mRNA levels of Smurf1 and TGF-βRI. Immunofluorescence staining clearly presented that miR-503-5p mimics significantly decreased protein expression of Smurf1 (Fig. 8G). On the contrary, miR-503-5p inhibitor increased Smurf1 and decreased TGF-βRI and collagen I (Fig. 8H-K).

Fig. 8.

miR-503-5p regulated Smurf1/Smad7 signaling pathway. A, B Venn diagram showing overlap between differentially up-regulated miRNAs in TPE exosomes and miRNAs targeting Smurf1 or Smad7 predicted by the miRTarBase. The online analysis database “miRTarBase” (https://miRTarBase.cuhk.edu.cn/) was used. C Human PMCs were transfected with miR-503-5p mimics (50 nmol/ml) or negative control (NC) for 24 h, after which miR-503-5p expression levels in cells were determined by qRT-PCR and normalized by the U6. D-E Human PMCs were transfected with miR-25-3p mimics or miR-503-5p mimics or miR-92a-3p mimics (50 nmol/ml) or negative control (NC) for 24 h, after which intracellular mRNA levels of Smurf1 were measured by RT-qPCR and normalized to GAPDH (D). The protein expression of Smurf1 and TGF-β receptor (TGFBR) were detected by western blotting. Bar graphs revealed changes in relative ratio of Smurf1 and TGFBR to GAPDH. F Human PMCs were transfected with miR-503-5p mimics (50 nmol/ml) or negative control (NC) for 24 h, after which mRNA expression of Smurf1, TGFBR and COL1A1 were detected by qRT-PCR. G Human PMCs were incubated with miR-503-5p mimics. After 24 h, Smurf1 protein was detected by immunofluorescence staining and nuclei with DAPI staining. Bar scale: 50 μm. H-K Human PMCs were transfected with miR-503-5p mimics (50 nmol/ml) or negative control (NC) for 24 h. miR-503-5p expression levels in cells were determined by qRT-PCR and normalized by the U6 (H). The protein expression of S Smurf1, TGFBR and COL1A1 were detected by western blotting (I). Bar graphs revealed changes in the relative ratio to GAPDH (J). mRNA levels of S Smurf1, TGFBR and COL1A1 were measured by RT-qPCR and normalized to GAPDH (K). Data are mean ± SEM. n = 3. *P < 0.05 (student’s t-test)

According above results, another miRNA, miR-424-3p up-regulated in TPE-Exosomes (Fig. 4C and E). miR-424-3p was predicted interacting with Smad7 by bioinformatic, and then further studied. As shown in Fig. 9A, miR-424-3p could binding to Smad7 gene and inhibit Smad7 transcription. miR-424-3p mimics decreased Smad7 and increased TGF-β and TGF-βRI as well as collagen I (Fig. 9B-E). Not surprisedly, miR-424-3p inhibitor had oppositive effect as miR-424-3p mimics (Fig. 9F and G). Moreover, TPE-Exosomes promoted fibrotic changes in PMCs in vivo, but these were attenuated by miR-424-3p inhibitor (Fig. S10).

Fig. 9.

miR-424-3p regulated SMAD7/TGF-β signaling. A Predicted hsa-miR-424-3p target sequences in the 3′UTRs of the SMAD7 mRNA. Sequences of hsa-miR-424-3p and the putative target sequence in the Smad7 mRNA (wildtype) or an engineered mutant of this sequence (mutant) and its effect on the luciferase reporter. B, C Human PMCs were transfected with miR-424-3p mimics (50 nmol/ml) or negative control (NC) for 24 h, after which miR-424-3p expression levels in cells were determined by qRT-PCR and normalized by the U6 (B). mRNA levels of Smad7, collagen I (COL1A1), TGF-β (TGFB) and type I TGF-β receptor (TGFBR1) were measured by RT-qPCR and normalized to GAPDH (C). D, E Human PMCs were transfected with miR-424-3p mimics (50 nmol/ml) or negative control (NC) for 24 h. The protein levels of Smad7, COL1A1, TGF-β, p-smad2/3 and t-smad2/3 were measured by western blotting (D). Bar graphs revealed changes in relative ratio of Smad7, COL1A1 and TGF-β to GAPDH, or p-smad2/3 to t-smad2/3 and GAPDH (E). F, G Human PMCs were transfected with miR-424-3p inhibitors (100 nmol/ml) or negative control (NC) for 24 h, after which miR-424-3p expression levels in cells were determined by qRT-PCR and normalized by the U6 (F). mRNA levels of Smad7, COL1A1, TGFB and TGFBR1 were measured by RT-qPCR and normalized to GAPDH (G). Data are mean ± SEM. n = 3. *P < 0.05 (student’s t-test)

As an inhibitory Smad, Smad7 competes with receptor-regulated Smads (Smad1, 2, 3, 5 and 8) and recruits Smurf1. Smurf1is a HECT E3 ubiquitin ligase which induces proteasomal degradation of TGF-βRI. Thus, to further confirm the role of these miRNAs on Smad7 and TGF-βRI, the binding of Smad7 with Smurf1, ubiquitination of TGF-βRI were investigated. As shown in Fig. S11A-C, miR-503-5p and miR-424-3p reduced binding of Smad7 with Smurf1, as well as decreased ubiquitination of endogenous TGF-βRI in human PMCs. On the contrary, miR-424-3p inhibitors and miR-503-5p inhibitors increased ubiquitination of TGF-βRI (Fig. S11D).

These data suggested that miR-503-5p, miR-25-3p, miR-92a-3p, and miR-424-3p converge on the Smad7/Smurf1/TGF-βRI axis and were associated with enhanced TGF-β signaling activity.

Artificially removal of miRNAs in TPE-Exosomes inhibited pleural fibrosis

To investigate orchestrating role of exosomal miRNAs in pleural fibrosis, a miRNA inhibitor which could triple suppress miR-150-3p, miR-424-3p and miR-503-5p was designed. The triple miRNAs inhibitor was transferred into TPE-Exosomes and then intra-pleural injected in mice. As shown in Fig. 10, the triple miRNAs inhibitor significantly restrained pleural fibrosis induced by TPE-Exosomes.

Fig. 10.

Triple miRNAs inhibitor attenuated TPE-Exo induced pleural fibrosis. C57BL/6 mice were intra-pleural injected by using PBS (100 µl/mouse), TPE-Exo (100 µl/mouse), TPE-Exo plus control inhibitor, or TPE-Exo plus triple miRNAs inhibitor with carbon particles (0.1 mg/mouse) as descriptions in the Methods. TPE-Exo from 50 ml TPE was administered at days 1, 5, 9. In TPE-Exo plus triple miRNAs inhibitor group, TPE-Exo was co-incubated with triple miRNAs inhibitor which restrained expressions of miR-150-3p, miR-424-3p and miR-503-5p. All mice were euthanized at day 21, and tissues were taken for analysis. A Representative Masson’s trichrome staining images of visceral pleura from lung sections, parietal pleura from chest wall and diaphragm sections. Original magnification, ×400. B Changes in pleural thickness. C Changes in collagen percentages of visceral and parietal pleura. Data are expressed as mean ± SEM. n = 6 mice. ***P < 0.001 (One-way ANOVA followed by the Bonferroni’s test)

Discussion

In this study, we demonstrated that exosomes derived from TPE were associated with pleural fibrotic changes both in vivo and in vitro, accompanied by activation of the TGF-β signaling pathway. Disruption of exosomes or inhibition of exosome biogenesis attenuated experimental pleural fibrosis, supporting a functional role for exosome-mediated intercellular communication in this process. miRNA profiling further revealed a distinct exosomal miRNAs signature in TPE compared with transudative pleural effusion. Among these miRNAs, downregulated miR-135b-5p was associated with increased SDCBP expression in PMCs, while upregulated miR-150-3p correlated with elevated expression of the integrin ITGB6. In parallel, miR-503-5p, miR-25-3p, miR-92a-3p, and miR-424-3p converged on Smad7 and Smurf1, leading to reduced ubiquitination of TGF-βRI and enhanced receptor availability. Collectively, these exosomal miRNAs appeared to converge on multiple regulatory layers of the TGF-β signaling pathway. Through coordinated modulation of exosome biogenesis, integrin expression, and Smad-dependent signaling components, these miRNAs were associated with sustained activation of profibrotic signaling and pleural fibrotic responses (Fig. 11).

Fig. 11.

Graphic abstract of this study

Exosomes are released virtually by all cell types and serve as important mediators of intercellular communication under both physiological and pathological conditions. Using nano-flow cytometry, we identified CD3⁺ T cells as one significant source of exosomes in TPE. However, TPE represents a complex inflammatory microenvironment enriched with multiple immune and non-immune cell populations, including macrophages, PMCs, B cells, neutrophils, and other infiltrating immune cells, all of which are capable of releasing exosomes. Consequently, the exosomal population within TPE is inherently heterogeneous. Exosomes derived from distinct cellular origins may carry divergent miRNA cargoes, and the coexistence of these vesicles is likely to collectively shape the complex miRNA profile observed in TPE. Such cellular and molecular heterogeneity may enable coordinated regulation of multiple profibrotic signaling nodes, rather than reliance on a single dominant miRNA or cell type, thereby amplifying TGF-β-driven fibrogenic responses within the pleural space.

SDCBP is a small PDZ domain scaffolding protein, which binds and links syndecans and ALG-2-interacting protein X (ALIX), facilitates biogenesis of exosomes, and segregates intracellular cargoes into exosomes [31]. In the present study, tuberculosis-related factors increased SDCBP protein expression in PMCs, suggesting that TPE microenvironment favors enhanced exosome generation. Consistent with this notion, reduction of exosomes by ultrasonic disruption in vitro or by pharmacological inhibition of exosome biogenesis with GW4869 in vivo attenuated pleural fibrosis. These findings supported a contributory role of exosome abundance in pleural fibrotic remodeling. Exosomes possess low immunogenicity due to their biocompatible lipid bilayer structure, which also protects their cargoes from enzymatic degradation. Their small size and membrane composition facilitated efficient transfer of miRNAs and proteins from immune cells to PMCs within the pleural space [21], thereby enabling sustained biological activity in the fibrotic microenvironment.

In addition to the intrinsic properties of exosomes, the biological effects of their cargoes, particularly miRNAs, are likely to be critical determinants of fibrotic responses. Previous studies demonstrated that miRNAs participated in fibrosis by activating profibrotic signaling pathways, most notably TGF-β signaling [28]. Aberrant exosomal miRNA profiles have also been implicated in the pathogenesis of IPF [28, 32]. In this study, we focused on the potential roles of downregulated miR-135b-5p and upregulated miR-150-3p, miR-25-3p, miR-92a-3p, miR-503-5p, and miR-424-3p in pleural fibrosis.

Among the downregulated miRNAs, miR-135b-5p was found to target SDCBP. Reduced miR-135b-5p levels were associated with increased SDCBP expression and enhanced exosome biogenesis in the TPE microenvironment. As discussed above, increased exosome availability appeared to favor the development of pleural fibrosis. Regarding the upregulated miRNAs, miR-150-3p has been reported to participate in oxidative stress responses and disease progression in several pathological contexts [33–35]. In our study, miR-150-3p was upregulated in TPE-derived exosomes and associated with suppression of ELK1 expression and increased ITGB6 levels in PMCs. ELK1 has been shown to repress αvβ6 integrin transcription by binding to the ITGB6 promoter [36]. Consistent with previous reports demonstrating a positive feedback loop between αvβ6 integrin expression and TGF-β1 activation [37, 38], miR-150-3p overexpression was associated with pleural fibrotic changes both in vitro and in vivo, while miR-150-3p inhibition attenuated fibrosis induced by TPE with carbon particles. These findings supported a profibrotic role for miR-150-3p in the context of tuberculous pleural disease.

We further identified miR-503-5p, miR-25-3p, and miR-92a-3p as targeting Smurf1, while miR-424-3p targeted Smad7. Smad7 functions as a key inhibitory regulator of TGF-β signaling by recruiting Smurf E3 ubiquitin ligases to promote degradation of TGF-βRI and by competing with Smad4 or TGF-βRI for receptor-regulated Smads [39, 40]. Previous studies have reported several miRNAs, including miRNA-519d, miR-21-5p, and miRNA-424/503 cluster members, as regulators of Smad7 [41–43]. In our study, upregulated miR-424-3p suppressed Smad7 expression, while miR-503-5p–mediated suppression of Smurf1 reduced ubiquitination of TGF-βRI. Together, these coordinated effects were associated with increased stability of TGF-βRI and enhanced Smad2/3 signaling, thereby favoring activation of the TGF-β pathway.

It should be noted that, although consistent regulatory relationships between these exosomal miRNAs and components of the TGF-β signaling pathway were observed, formal rescue experiments were not performed. Therefore, the proposed mechanisms should be interpreted as convergent regulatory associations rather than definitive linear causal pathways. Future studies incorporating targeted rescue strategies, such as restoration of SDCBP, Smad7, or Smurf1 expression in the context of miRNA modulation, will be required to further validate these mechanistic links.

Finally, the orchestrating contribution of these exosomal miRNAs to pleural fibrosis was supported by the observation that a combined inhibitor targeting miR-150-3p, miR-503-5p, and miR-424-3p attenuated TPE-exosome–induced pleural fibrosis in vivo. Although the present study highlighted exosomal miRNA signaling as a potential therapeutic axis in pleural fibrosis, several translational challenges might be carefully considered. Efficient and targeted delivery of miRNA inhibitors to the pleural space remains a key hurdle, as does ensuring sufficient stability of these molecules within the inflammatory pleural microenvironment. While exosomes offer advantages such as biocompatibility and cargo protection, their immunogenicity, biodistribution, and potential accumulation in off-target tissues require thorough evaluation.

Conclusions

In summary, exosomes derived from TPE were associated with pleural fibrotic remodeling, and exosomal miRNAs played important roles in this process by converging on key nodes of the TGF-β signaling pathway. This study provided a mechanistic framework for understanding post-tuberculous pleural fibrosis, and suggested that exosomal miRNA signaling represented a promising direction for future preclinical therapeutic exploration.

Abbreviations

BCG

Bacillus Calmette- Guerin

FBS

Fetal Bovine Serum

IPF

Idiopathic Pulmonary Fibrosis

KEGG

Kyoto Encyclopedia of Genes and Genomes

miRNA

microRNA

NC

Negative control

NTA

Nanoparticle Tracking Analysis

PMC

Pleural Mesothelial Cell

TGF-β

Transforming Growth Factor-beta

TGF-Βri

Type I TGF-β Receptor

TGF-βRII

Type II TGF-β Receptor

TPE

Tuberculosis Pleural Effusion

TEM

Transmission Electron Microscope

Authors’ contributions

Li-Juan Jiang, Yuan-Yi Zheng: conceptualization; investigation. Li-Mei Liang: visualization; writing-original draft. Zi-Heng Jia: conceptualization; investigation. Rong-Hui Du, Hai Huang, Xiyong Dai: sample collections; methodology. Pei-Pei Cheng, Li-Qin Zhao, Qian Li, Ye Han Jiang, Xiao Lin Cui, Shu Yi Ye, Shi He Hu, He De Zhang, Chen Yue Lian, Xiao Feng: methodology; investigation. Lin-Jie Song, Fan Yu, Xin Liang He, Fei Xiang, Xiaorong Wang: data analysis. Liang Xiong: review and editing. An-Dong Liu: funding acquisition. Meng Wang: conceptualization; investigation; funding acquisition. Hong Ye, Wan-Li Ma: conceptualization; writing-original draft; writing-review and editing; funding acquisition. All authors reviewed the manuscript.

Funding statement

This work was supported by the National Natural Science Foundation of China (No. 82270111 and 81973991 to WLM; No. 82270075 and 82470100 to HY; No. 82200081 to MW; No. 82403464 to ADL).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

All experiments involving animals were approved by the Institutional Animal Care and Use Committee (IACUC) of the Tongji Medical College, Huazhong University of Science and Technology (approval ID: 2020-S411) and conformed to the guidelines for animal experiments of laboratory animals. All experiments involving clinical samples were approved by the Institutional Review Board of the Tongji Medical College, Huazhong University of Science and Technology (approval ID: 2020-IEC-J-187). Informed consent was obtained from all patients, and all procedures were conducted in accordance with the principles of the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.