Bone Marrow Mesenchymal Stem Cell-Derived Exosomes microRNA-31-5p Repress Pulmonary Fibrosis via IGFBP7

Ethics statement

The research was conducted with the approval of the ethics committee of Leshan People’s Hospital. Animal pains were minimized as much as possible.

Isolation, culture, and identification of BMSCs

BMSCs were isolated from the femur and tibia of C57BL/6 mice. The extracted cells were suspended in Dulbecco’s Modified Eagle Medium (DMEM) containing 10% fetal bovine serum (Biowest, Nuaillé, France) and penicillin–streptomycin (100 U/mL, Gibco Life Technologies, NY, USA). Medium renewal was performed after 3 days. Cells that reached 80–90% confluence were sub-cultured to a third passage, which was then prepared to a cell suspension (1 × 10⁶ cells/mL, 200 μL). ¹⁸

After subculturing to the third–fourth passage, the cultured BMSCs exhibited a typical spindle-shaped morphology and spiral arrangement, indicating uniformity and viability. Fluorescein isothiocyanate-labeled mouse anti-human antibodies (CD29, CD44, CD90, and CD45) were cultured with BMSCs at the third passage and then detected by flow cytometry. ¹⁹ Adipogenic and osteogenic differentiation experiments were performed, and then oil red O staining and alizarin red staining were conducted. The stained cells were observed under a microscope (Olympus, Tokyo, Japan). ²⁰

Isolation, extraction, and identification of BMSCs-Exos

BMSCs at third passage were cultured overnight in serum-free DMEM. The conditioned medium of BMSCs (80–90% confluence) was collected and treated with centrifugation to remove cell debris. Then, supernatant was centrifuged at 10,000 g twice and the resulting pellet was resuspended in 25 mM 4-(2-hydroxyethyl)−1-piperazineëthanesulfonic acid (pH = 7.4) and centrifuged again at 10,000 g to obtain Exos, which were stored at −80°C for backup. To further characterize the morphology and size of Exos, 10 μL of Exo suspension was applied onto carbon-coated copper grids and allowed to stand at 25°C for 10 min. After carefully removing the excess liquid with filter paper, the samples were negatively stained with 3% phosphotungstic acid solution for 5 min. The Exo samples were then observed using a 100 kV transmission electron microscope (Krios Rx Cryo-TEM, Thermo Fisher Scientific, USA), and electron micrographs were taken (magnification: 20,000× to 50,000×). Three representative TEM fields were randomly selected, and the diameters of multiple individual vesicles were measured using ImageJ software. Additionally, Western blot was employed to identify the surface antigens CD63 and Alix on BMSC-Exos.

Induction of PF in mice by BLM

Seventy 14-week-old C57BL/6 mice (Charles River Laboratories, MA, USA) were housed in plastic cages and acclimatized for 1 week in a controlled environment (20 ± 2°C, 50 ± 10% humidity, and a 12-h light-dark cycle). All mice were anesthetized with isoflurane and kept under specific pathogen-free conditions. A single dose of BLM (3 U/kg, 50 μL, Shanghai Maokang Biotechnology Co., Ltd., Shanghai, China) was administered intratracheally to induce PF. Mice in the blank control group received an intratracheal injection of 0.9% normal saline (50 μL) and were designated as the normal group (n = 10). ¹⁰ The remaining 60 BLM-treated mice were randomly divided into six groups (n = 10 each), including the Model group, Model-Exo group (200 μL of phosphate buffer solution [PBS] containing 20 μg of Exos was injected through the tail vein), Model-Exo negative control (NC) group (25 μg of Exos containing miR-31-5p agomir NC was injected through the tail vein), Model-Exo-miR-31-5p agomir group (25 μg of Exos containing miR-31-5p agomir was injected through the tail vein), Model-Exo-miR-31-5p agomir + overexpression (oe)-NC (25 μg of Exos containing miR-31-5p antagomir and oe-IGFBP7 NC lentivirus were injected through the tail vein), and Model-Exo-miR-31-5p agomir + oe-IGFBP7 (25 μg of Exos containing miR-31–5p antagomir and oe-IGFBP7 lentivirus were injected through the tail vein). All treatments were administered two times a week for 5 weeks. ²¹ Both the oe-IGFBP7 and control lentiviruses were synthesized and provided by Shanghai Genechem Co., Ltd. (China). The packaging virus and target vector were cotransfected into 293T cells (with a confluence of 80–90%) using Lipofectamine 2000 (11668500, Invitrogen, USA). After 48 h of cell culture, the supernatant was collected, filtered through a 0.45 μm membrane, and then the viral particles were concentrated by ultracentrifugation (25,000 g, 2 h, 4°C). The viral pellet was resuspended in PBS, and the viral titer was determined using qPCR, yielding a final concentration of 1 × 10⁹ TU/mL. In animal experiments, the lentivirus and miR-31-5p antagomir-modified Exos were coinjected via the tail vein into BLM-treated mice. Each injection consisted of 200 μL (containing approximately 2 × 10⁸ TU of viral particles), administered twice a week for 5 consecutive weeks. The mice were euthanized at the end of the experiment, and bronchoalveolar lavage fluid (BALF) and lung tissues were collected. Tissues were either fixed in 4% paraformaldehyde for histological analysis or stored at −80°C for subsequent assays. ²²

Determination of pulmonary function

Mice were anesthetized via intraperitoneal administration of ketamine and xylazine. The trachea was intubated and connected to a mandatory lung ventilation system (Buxco Research System; Buxco Electronics, NC, USA). The respiratory rate was set at 150 breaths/min. Functional Residual Capacity (FRC) was determined according to Boyle’s Law. Lung compliance (C chord), inspiratory volume (IC), vital capacity (VC), total vital capacity (TLC), and forced vital capacity (FVC) were measured by the Buxco system. ²³

Enzyme-linked immunosorbent assay

Tumor necrosis factor-α (TNF-α), interleukin (IL)−1β, and IL-6 levels were measured. PBS (1 mL) was injected into the lungs for lavage. The supernatant was collected by centrifugation at 700 g. BALF (1 mL) was collected ²⁴ and centrifuged at 300 g. TNF-α, IL-1β and IL-6 levels in the supernatant were measured with enzyme-linked immunosorbent assay (ELISA) kits (SEKH-0047, SEKH-0013, Solarbio, Beijing, China). Absorbance value was recorded at 450 nm. ²⁵

Hematoxylin–eosin staining

Lung tissues were processed into paraffin-embedded sections (4 μm) and stained with hematoxylin and 1% eosin solution. ²⁶ Then, lung tissues were fixed in malol and observed under a BX51 microscope (Olympus). Lung injury was scored as follows: 0 point: normal lung tissue with no pathological features such as inflammation, edema, or hemorrhage; 1 point: mild lung injury with slight inflammatory infiltration and alveolar edema without hemorrhage; 2 points: moderate lung injury with obvious inflammatory cell infiltration, edema, and minor alveolar hemorrhage; 3–4 points: severe lung injury with substantial inflammatory cell infiltration and alveolar edema, accompanied by obvious hemorrhage, necrosis, and other pathological changes. Lung injury was independently scored by two pathologists. ^24,27,28 Additionally, images were analyzed using ImageJ image analysis software (http://rsb.info.nih.gov/ij/). The mean linear intercept (MLI) was determined as a measure of the average distance between gas exchange surfaces, following the guidelines outlined in the American Thoracic Society/European Respiratory Society quantitative assessment criteria for lung structure. ²⁹

Masson staining

Lung tissues were made into paraffin-embedded sections (4 μm) and routinely deparaffinized. Cell nuclei were stained with Regaud hematoxylin, followed by Masson-Ponceau red acid fuchsin solution. Afterward, the sections were treated with 1% phosphomolybdic acid aqueous solution and counterstained with aniline blue or light green. The sections were observed under a light microscope (DM 4000B, Leica, Germany), and the optical density values of collagen staining were quantified using Image J v1.8.0 software. ³⁰

Reverse transcription quantitative polymerase chain reaction

Total RNA of the lung tissues was extracted by Trizol (Invitrogen), quantified by NanoDrop2000 spectrophotometer (Thermo Fisher Scientific), and reverse transcribed using the PrimeScript RT kit with gDNA Eraser Kit (Takara, Shiga, Japan). Reverse transcription quantitative polymerase chain reaction (RT-qPCR) was performed using an ABI7500 quantitative PCR instrument (Thermo Fisher Scientific) with the SYBR PreMixExTaq (TliRNaseHPlus) kit (Takara). U6 small nuclear RNA and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) were taken as the internal controls of miR-31-5p and IGFBP7, respectively. The 2^−ΔΔCt method was adopted to quantify gene expression. Table 1 lists the sequences of the primer. ¹⁹

Western blot assay

Protein samples were extracted by utilizing radioimmunoprecipitation assay lysis buffer containing protease inhibitors (Beyotime, Shanghai, China), and protein concentrations were tested according to the instructions of BCA kits (Boster, Wuhan, Hubei, China). Samples were fully mixed with the loading buffer and boiled at 100°C, followed by an ice bath and centrifugation. An equal amount of protein was loaded onto 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis gels for separation and then transferred to the nitrocellulose membranes. Membranes were blocked overnight in 5% skimmed milk. Primary antibodies included IGFBP7 (1:500, IPD-ANP11965, Apti Biotechnology, Hubei, China), GAPDH (1:3000, ab8245, Abcam), CD63 (1:1000, ab315108, Abcam), and Alix (1:1000, ab275377, Abcam). After incubation with the appropriate HRP-conjugated secondary antibody (1:1000, Boster), the membranes were developed using ECL detection reagent (Pierce, USA). Images were captured following exposure and development, and band intensities were analyzed by ImageJ software (version 1.8.0, National Institute of Mental Health, USA). GAPDH served as the internal control. ³¹

Dual-luciferase reporter gene assay

pLUC-IGFBP7 wild-type (WT)/pLUC-IGFBP7 mutant type (MUT) and miR-31-5p mimic/mimic NC were transfected into HEK293T cells via liposome 2000 (Thermo Fisher Scientific). Plasmids were purchased from GenePharma (Shanghai, China). Luciferase activity was detected at 560 nm with a dual-luciferase reporter assay kit (Promega, WI, USA). ²⁴

Statistical analysis

Statistical analysis was implemented using SPSS 21.0 software (IBM, NY, USA) and GraphPad Prism 9.5 (California, USA). Measurement data were expressed by mean ± standard deviation. For normally distributed data with homogeneity of variance, intergroup comparisons were performed using an unpaired t-test. For data analysis among multiple groups, one-way analysis of variance was adopted. A rank sum test was applied to compare data with skewed distribution or variance heterogeneity. Statistical significance was determined with p < 0.05. ³²

Successful establishment of PF mice by BLM

BLM-treated PF mouse models were successfully established and assessed to ensure the success of induction. First, pulmonary function was tested: in BLM-treated mice, lower FRC, C chord, IC, VC, TLC, and FVC values were observed (Fig. 1A, B). In Masson’s trichrome staining, collagen content in the interstitial tissue was found to be significantly higher than that in the normal group mice (Fig. 1C). Hematoxylin–eosin (HE) staining (Fig. 1D) revealed that the main tracheal wall of mice thickened 14 days after modeling, with fibroblast and immune cell accumulation in the pulmonary interstitial tissue. Meanwhile, measurements of the MLI showed a significant increase in alveolar spacing in the Model group, further reflecting lung tissue destruction and a reduction in ventilation area (right panel of Fig. 1D). Next, inflammation was evaluated by ELISA (Fig. 1): BLM-treated mice showed increased TNF-α, IL-1β, and IL-6 levels. The above results displayed that the BLM-induced PF model in mice was successfully established.

OPEN IN VIEWER

Figure 1.

BLM induction reduces lung function in mice. (A, B) Lung function indicators in mice after modeling. (C) Masson’s trichrome staining and fibrosis quantification of lung tissues from each group. (D) HE staining for assessing lung injury, histological scoring, and MLI. (E) TNF-α, IL-1β, and IL-6 contents in lung tissues of PF mice were assessed by ELISA; n = 10, ***p < 0.001 versus the normal group. ELISA, enzyme-linked immunosorbent assay; HE, hematoxylin–eosin; IL, interleukin; MLI, mean linear intercept; TNF-α, tumor necrosis factor-α.

Identification results of BMSCs and BMSC-Exos

When BMSCs grew rapidly to 90% confluency, the cell clusters displayed an obvious swirling growth pattern, and the cell morphology was long and shuttle-shaped (Fig. 2A). Following lipogenic induction, Oil Red O staining revealed orange-red lipid droplets within the cells under microscopy; after osteogenic induction, Alizarin Red staining revealed obvious red calcium nodule formation (Fig. 2B). Flow cytometry showed positive expression of the surface molecules CD29 and CD9, markers of MSCs, and negative expression of CD45, a marker of hematopoietic cells (Fig. 2C). Under transmission electron microscopy, BMSC-Exos were observed as round or oval membrane-bound vesicles with diameters of 40–100 nm and significant heterogeneity, confirming successful isolation and morphological integrity (Fig. 2D). Western blot results demonstrated (Fig. 2E) that the extracted material positively expressed Alix and CD63 proteins, further confirming the successful extraction of Exos.

OPEN IN VIEWER

Figure 2.

Identification results of BMSCs and BMSC-Exos. (A) BMSCs morphology under inverted microscope (40×). (B) Identification results of BMSCs (oil red O-stained 200×, and alizarin red-stained 200×). (C) BMSCs CD90 and CD29 positive expression and CD45 negative expression were assessed by flow cytometry. (D) Morphology of BMSC-Exo under transmission electron microscope. (E) Exo surface antigen results were tested by Western blot assay. BMSCs-Exo, bone marrow mesenchymal stem cell-derived exosome.

miR-31-5p expression level changes in Exos influence PF in mice

miR-31 has been implied to reduce inflammatory responses and oxidative stress-induced neuronal injury in ischemic stroke, ³³ and to mediate the transition from inflammation to re-epithelialization during skin wound healing, suggesting miR-31 has therapeutic potential in injury repair. ³⁴ MSCs-Exos can prevent and reverse the core features of PF. ¹⁰ To verify this, the extracted Exos were injected into the model mice (Model-Exo group) and compared with the Model group to investigate whether MSCs-Exos could have therapeutic effects on PF in mice. To further verify the role and mechanism of miR-31-5p in Exos in BLM-induced PF, a miR-31-5p agomir-modified group (Model-Exo-miR-31-5p agomir) and an NC group (Model-Exo-NC) were established based on the Model-Exo group. RT-qPCR results showed that the miR-31-5p expression in the lung tissues of PF model mice was decreased, while its expression was restored in the Model-Exo-miR-31-5p agomir group (Fig. 3A).

OPEN IN VIEWER

Figure 3.

miR-31-5p expression in Exo influences PF in mice. (A) miR-31-5p expression levels were tested by RT-qPCR. (B, C) Intensity of lung function in PF mice after miR-31-5p treatment was assessed. (D) HE staining for assessing lung injury, histological scoring, and MLI. (E) Masson’s trichrome staining for assessing pulmonary collagen deposition and its quantitative analysis. (F) ELISA for detecting the expression levels of inflammatory cytokines TNF-α, IL-1β, and IL-6 in lung tissues. n = 10, *p < 0.05, **p < 0.01, ***p < 0.001. Exo, exosome; PF, pulmonary fibrosis; RT-qPCR, reverse transcription quantitative polymerase chain reaction.

Lung function assessments revealed that the Model-Exo-miR-31-5p agomir group showed significant improvements in lung capacity (FRC, IC, VC, TLC, FVC) and lung compliance (C chord) compared with the Model, Model-Exo, and Model-Exo NC groups (Fig. 3B, C). HE staining showed improved structural integrity of lung tissue, reduced inflammatory cell infiltration, and significantly decreased lung injury scores and MLI values in the Model-Exo-miR-31-5p agomir group (Fig. 3D). Masson’s trichrome staining indicated a marked reduction in collagen deposition in the lung tissues of this group, suggesting fibrosis alleviation (Fig. 3E). ELISA results further showed that the levels of TNF-α, IL-1β, and IL-6 in the Model-Exo-miR-31-5p agomir group were significantly lower than those in the Model-Exo-NC group (Fig. 3F). These results collectively indicate that BMSC-Exos can effectively inhibit pulmonary inflammation and fibrosis progression, thereby improving lung function, by delivering miR-31-5p. ^33,34

A targeting relation exists between miR-31-5p and IGFBP7

A targeting relationship between miR-31-5p and IGFBP7 was predicted through the Starbase website (https://rnasysu.com/encori/) (Fig. 4A). This was confirmed by a dual-luciferase reporter assay, which showed that miR-31-5p bound directly to the 3′UTR of IGFBP7, and the luciferase activity of WT 3′UTR-transfected cells was inhibited, suggesting that IGFBP7 was a direct target of miR-31-5p (Fig. 4B). Furthermore, it was found that IGFBP7 expression levels in PF mice were elevated, and when miR-31-5p was up-regulated, IGFBP7 expression levels were diminished (Fig. 4C, D). The above results display that miR-31-5p targets and suppresses IGFBP7 expression.

OPEN IN VIEWER

Figure 4.

A targeting relationship presented between miR-31-5p and IGFBP7. (A) Starbase website predicted the targeting relationship and binding sites between miR-31-5p and IGFBP7. (B) The targeting relationship between miR-31-5p and IGFBP7 was verified by dual-luciferase reporter gene assay; (C–D). IGFBP7 expression levels were tested by RT-qPCR and Western blot assay; ***p < 0.001.

BMSCs-Exos affect PF in mice by delivering miR-31-5p to regulate IGFBP7

To further investigate whether BMSC-Exos affect PF via the miR-31-5p/IGFBP7 axis, two additional groups were established: Model-Exo-miR-31-5p agomir + oe-NC and Model-Exo-miR-31-5p agomir + oe-IGFBP7 groups. In the Model-Exo-miR-31-5p agomir + oe-IGFBP7 group, IGFBP7 expression was elevated (Fig. 5A); FRC, C chord, IC, VC, TLC, and FVC were decreased (Fig. 5B, C); lung injury score and MLI values were increased, and PF degree was pronounced (Fig. 5D, E); and TNF-α, IL-1β, and IL-6 levels were elevated (Fig. 5F). Taken together, BMSC-Exos influence PF in mice by delivering miR-31-5p to regulate IGFBP7.

OPEN IN VIEWER

Figure 5.

BMSCs-Exos affect PF in mice by delivering miR-31-5p to regulate IGFBP7. (A) IGFBP7 expression in PF mice after miR-31-5p and IGFBP7 treatment was assessed by RT-qPCR. (B, C) Lung function of PF mice after miR-31-5p and IGFBP7 treatment was tested. (D) Masson’s trichrome staining and fibrosis quantification of lung tissues from each group. (E) HE staining for assessing lung injury, histological scoring, and MLI. (F) TNF-α, IL-1β, and IL-6 levels of PF mice after miR-31-5p and IGFBP7 treatment were assessed by ELISA; *p < 0.05, **p < 0.01, ***p < 0.001 versus the Model-Exo-miR-31-5p agomir + oe-NC group. NC, negative control.