1. Introduction
Pulmonary fibrosis (PF) is a chronic respiratory disease with a global estimated incidence of 9-130 cases per million people and a prevalence of 33-451 cases per million people [
1]. The mortality rate associated with PF is increasing [
2]; thus, treating this disease is a significant medical challenge in modern society. PF is characterized by impaired lung function, extracellular matrix (ECM) deposition, tissue remodeling, and fibrotic scar formation accompanied by alveolar epithelial damage and inflammation [
3]. Collagen, which makes up 30%-70% of the ECM, plays a pivotal role in fibrosis [
4]. The effectiveness of antifibrotic drugs is frequently evaluated by assessing reductions in collagen synthesis and deposition [
5]. Over 16 collagen subtypes have been linked to fibrosis formation [
6]. Therefore, it is evident that the synthesis and composition of collagen subtypes play essential roles in PF progression.
Increasing evidence suggests a sharp increase in PF incidence with age [
7], [
8], [
9], [
10]. Surveys have shown that individuals over 70 face a 6.9 times greater risk of developing PF than those under 40 [
11]. With advancing age, the body’s response to external stimuli diminishes, cells undergo aging, and regenerative capabilities weaken, rendering lung tissue more susceptible to damage and prone to fibrosis progression. However, specific drugs for PF, such as pirfenidone and
N-acetylcysteine, primarily target this treatment by reducing inflammation, cytokine production, and oxidative stress. While the pathogenic role of immune cells, including T cells and macrophages, has been extensively documented in PF [
12], [
13], the impact of inflammation in this disease remains highly debated, given clinical trial results; anti-inflammatory treatments have been demonstrated to be ineffective or even associated with adverse effects for PF patients [
14]. Moreover, previous studies [
15], [
16] have shown that patients with nonspecific interstitial pneumonia and collagen-vascular-disease-associated PF exhibit severe fibrotic lung symptoms but relatively mild inflammatory pathological features. Hence, the primary factor initiating the onset of PF may not be inflammation but aging itself.
The accumulation of aging lung cells and the evolution of the microenvironment during the aging process may play crucial roles in PF development. Studies [
17], [
18] have identified a substantial increase in age-related biomarkers in the lungs of PF patients. As lung cells age, they display an altered secretion pattern that can impact the proliferation and differentiation of fibroblasts by releasing senescence-associated secretory phenotype (SASP) factors, various cytokines, chemokines, matrix remodeling enzymes, and profibrotic factors [
18], thus promoting the progression of fibrosis. During aging, the age-related decline in alveolar epithelial function may trigger progressive PF [
19], [
20], [
21]. Regional depletion of epithelial progenitor cells in fibrotic areas has been observed in PF patients [
22], [
23], with reduced alveolar type 2 (AT2) cell function/quantity, extensive secretion of SASP factors and profibrotic mediators [
24], [
25], and aberrant activation of multiple cellular senescence-related pathways [
17], [
19], [
26], [
27]. Recent data from an
in vitro organ model study on PF in mice and humans suggest that aged AT2 cells can directly influence the transformation of lung fibroblasts into myofibroblasts through the autocrine p53 and TGF-β signaling pathways, thereby contributing to PF progression [
28]. Furthermore, the literature [
29], [
30] have indicated that targeting the clearance of senescent AT2 cells in mouse lungs using senolytic drugs
in vitro can stabilize the function of epithelial cells, reduce the expression of fibrotic mediators, and decrease fibrosis. Similarly, selectively eliminating senescent AT2 cells can reverse persistent PF symptoms caused by ionizing radiation in mice [
31]. It is clear that senescent AT2 cells contribute significantly to the development of PF.
Although the specific mechanism of PF development remains unclear, a growing body of research indicates that PF involves the persistent activation of fibroblasts due to endogenous or exogenous factors, leading to excessive collagen synthesis and the over-deposition of ECM, ultimately culminating in organ failure [
5]. In the fibrotic process, multiple signals collectively regulate collagen synthesis, including transforming growth factor-β (TGF-β), wingless (Wnt)/β-catenin, epidermal growth factor receptor (EGFR), yes-associated protein (YAP)/transcriptional co-activator with PDZ-binding motif (TAZ), mechanistic target of rapamycin (mTOR), bone morphogenetic protein (BMP), and inflammation [
32], [
33], [
34]. However, the specific composition of collagen subtypes and the regulatory mechanisms governing their synthesis in PF with natural aging remain elusive. Although ample evidence suggests that dysfunctional AT2 cells are central to PF initiation and progression, the role of aging AT2 cells in collagen synthesis remains poorly understood. Direct communication between aging AT2 cells and fibroblasts and the specific regulatory factors and molecular mechanisms involved have yet to be fully elucidated.
This study aims to explore the major components of collagen in age-induced PF and elucidate the regulatory effects of aging AT2 cells on collagen synthesis. We used a natural aging mouse model and conducted a single-cell data analysis of human lung tissue. Our study provides essential evidence for understanding the influence of epithelial dysfunction on PF progression and further identifies novel targets for clinical treatment.
2. Methods
2.1. Mice and tissue sampling
Animal trials were carried out in compliance with Animal Welfare Guidelines (EU Directive 2010/63/EU) and approved by the Animal Care and Use Committee of China Agricultural University (approval number AW41203202-5-3). Due to their high genetic and physiological resemblance to humans, well-defined genetic makeup, relatively stable phenotype, and relatively short lifespan, C57BL/6J mice are widely employed in research related to human aging [
35]. Pathogen-free C57BL/6J mice were procured from Beijing Vital River Laboratory Animal Technology Co., Ltd., China and housed individually to naturally age. The study included six mouse age groups ranging from 3 to 28 months (i.e., 3m, 6m, 12m, 18m, 24m, and 28m were mouse at 3, 6, 12, 18, 24, and 28 months, respectively).
The mice in each group were anesthetized and perfused with saline to remove blood; this was followed by perfusion and fixation with 4% paraformaldehyde (PFA). Lung lobes were dissected, fixed, processed, and embedded in paraffin. Tissues not treated with PFA were stored at -80 °C for subsequent analysis.
2.2. Sampling of human lung tissues
Human paratumor tissues were collected in accordance with the Code of Ethics of the World Medical Association (declaration of Helsinki), and the protocol was approved by the Human Research Ethics Committee of China Agricultural University (approval number CAUHR-20230502). The lung aging study involved five younger individuals (≤ 40 years) and five older individuals (≥ 60 years). Paratumor tissue samples were collected 2 cm from the tumor edge. Some samples were frozen at -80 °C, while others were fixed in formalin overnight at 4 °C and processed into 5 μm paraffin sections for histological analysis.
2.3. Assessment of pulmonary function in mice
In this study, we employed the DSI Buxco whole body plethysmography (WBP) system (Wilmington, USA) to assess pulmonary function in naturally aged mice. The WBP system is a noninvasive procedure for analyzing respiratory function in awake animals. Each mouse underwent a 30 min acclimation period in an enclosed chamber to attain a stable and natural respiratory state before respiratory parameter monitoring began. Once the experiment begins, sensors captured and automatically recorded pressure changes (ΔP) associated with mouse respiration every 2 s. Detailed information regarding the typical respiratory parameters detected by the experiment is provided in Table S1 in Appendix A. The average values of each parameter from a 20 min measurement per mouse were applied in the subsequent analysis using SPSS 26.0 software (IBM, USA).
2.4. H&E staining
We used a hematoxylin and eosin (H&E) staining kit (G1120, Solarbio, China) to assess the pulmonary tissue structure of aging mice. In accordance with the manufacturer’s instructions and supplemented by methods described in Refs. [
36], [
37], lung tissue sections were stained and examined under a bright-field microscope (DM6B, Leica, Germany) at 20× magnification.
2.5. Masson’s trichrome staining
We assessed the collagen fiber content in lung tissue using a Masson’s trichrome staining kit (G1340, Solarbio) per the manufacturer’s directions. The blue-stained areas were observed under a conventional optical microscope (DM6B, Leica) at 10× magnification.
2.6. Sirius Red staining
A modified Sirius Red staining kit (G1472, Solarbio) was employed to differentiate the types of collagen fibers and quantify their content in the lung tissue. The stained sections were photographed under a polarizing microscope (DM4B, Leica). ImageJ software was used to extract and analyze the red and green fluorescence signals and calculate the average fluorescence intensity.
2.7. Transmission electron microscopy
The lung samples (approximately 1 mm3) were fixed in a 2.5% glutaraldehyde solution, washed with phosphate-buffered saline (PBS) three times, and then fixed in a 1% osmium tetroxide solution for 2 h. After dehydration using graded ethanol solutions and acetone treatment, the samples were infiltrated with embedding resin and acetone, polymerized, and sectioned into 90 nm ultrathin slices using an ultramicrotome (EMUC7, Leica). The slices were stained with 2% uranyl acetate and lead citrate for 10 min each prior to examination under a transmission electron microscope (TEM; Hitachi, H-7650, Japan).
2.8. Immunofluorescent staining
The tissue sections were subjected to citrate antigen retrieval (ZL1-9064, ORIGEN, China; pH 6.0), followed by a 15-min immersion in PBS solution with 1% Triton X-100 (T8200, Solarbio). Afterward, the sections were blocked with 10% goat serum (C01-03001, Bioss, China) for 1 h. Following overnight incubation at 4 °C with primary antibodies, the tissue sections underwent a 1 h incubation at ambient temperature in the dark with secondary antibodies (antibody information is provided in Table S2 in Appendix A). After that, the sections underwent staining with 4’,6-diamidino-2-phenylindole (DAPI) working solution (C0065, Solarbio) for a duration of 10 min and were sealed with an anti-fade mounting solution (P0126, Beyotime, China) under a coverslip. Subsequently, the samples were examined and imaged using a laser confocal microscope (LSM 900/Axio Observer 7, ZEISS, Germany).
2.9. Cell culture
Primary AT2 cells were cultured in custom medium (CM-M003, Procell, China) supplemented with growth factors. The culture vessels used for these cells were all coated with 5 μg·cm−2 of mouse tail collagen (354236, Corning, USA) to improve cell adhesion. The A549 and NIH3T3 cell lines were procured from Wuhan Procell Biotechnology Co., Ltd., China and cultured in Ham’s F-12K medium (PM150910, Procell) and Dulbecco’s modified eagle medium (DMEM; PM150210, Procell), respectively. The WI-38 cell line was obtained from the Cell Bank of the Typical Culture Preservation Committee, China, and was cultured in minimum essential medium (MEM; PM150410, Procell).
2.10. Cellular aging model
We constructed a cellular senescence model for the A549 cells and AT2 cells following methods outlined our team’s recent publication [
38]. The cells were seeded at 1.5 × 10
5 cells per well in a six-well plate and cultured for 24 h to ensure attachment and proliferation. Then, 2 mL of culture medium containing 30 g·L
−1 D-galactose was added and incubated for 48 h to induce cell senescence.
2.11. Coculture and conditioned culture model
Conditioned medium culture: A cellular senescence model was established with A549 or primary AT2 cells treated with 30 g·L−1 D-galactose. Simultaneously, NIH3T3 cells were seeded at 2 × 105 cells per well in a six-well plate and subjected to a 24 h serum starvation period in 0.4% serum medium. The conditioned media collected from the A549 or primary AT2 cells were then used to culture the NIH3T3 cells. The total cellular proteins were extracted after 48 h of incubation.
Coculture using a Transwell system: A549 cells were cultured in the upper chamber of the Transwell system to establish a cellular senescence model with 30 g·L−1 D-galactose. NIH3T3 or WI-38 cells were seeded in the lower chamber at a density of 2 × 105 cells/well and starved for 24 h in 0.4% serum medium. Then, coculture was conducted for 48 h prior to protein extraction.
2.12. Stimulation of fibroblasts by PAI-1
NIH3T3 cells were seeded at 1.5 × 105 cells/well in a six-well plate and cultured until they reached 80% confluence. After overnight starvation in medium containing 0.4% serum, the NIH3T3 cells were exposed to various concentrations of PAI-1 (HY-P71133, MCE, China) for 48 h. The total cellular proteins were subsequently extracted and analyzed.
2.13. Small interfering RNA transfection
NIH3T3 cells were plated at a density of 3 × 104 cells per well in a 24-well plate and grown to 60%-70% confluency. Twenty-five picomoles of small interfering RNA (siRNA) and 1 µL of KEL-R Transfection Reagent (KC127, GENCEFE, China) were diluted separately in 50 µL of Opti-MEM (31985070, Gibco, USA) and incubated for 5 min. The two solutions were gently mixed and incubated for 20 min to create the transfection vector mixture. After 8 h of transfection, the NIH3T3 cells were exposed to medium containing 2 μg·mL−1 PAI-1 (HY-P71133, MCE) and 0.4% serum. Transfection efficiency was assessed using fluorescein amidite (FAM) green fluorescence. The total cellular proteins were then extracted after 48 h for subsequent analysis. The specific siRNA sequences of the primers used are provided in Table S3 in Appendix A.
2.14. Inhibitor treatment
NIH3T3 cells were seeded at 1.5 × 105 cells/well in a six-well plate, to 80% confluence. The cells were then starved overnight with 0.4% serum and pretreated with 10 μmol·L−1 SB431542 for 45 min, followed by treatment with 2 μg·mL−1 PAI-1 (HY-P71133, MCE) and 10 μmol·L−1 SB431542. After 48 h, the total cellular proteins were extracted, and the TGF-β pathway-related proteins were analyzed.
2.15. Extraction of primary AT2 cells
Prior to the experiment, a digestion enzyme solution was prepared by combining 780 U of Type I collagenase (17100017, Gibco; 260 U·mg−1), 40 U of elastase (LS002294, Worthington, China; ≥ 3 U·mg−1), and 20 U of dispase II (4942078001, Roach, China; 0.9 U·mg−1) in 10 mL of Ca2+- and Mg2+-supplemented PBS solution (FB13356, FEIMOBIO, China).
The mice were anesthetized, perfused with physiological saline to clear blood from the lung tissues, and injected with 1.0 mL of enzyme solution via tracheal intubation. Next, 0.6 mL of 1% low-melting-point agarose was used to seal the trachea. The lung tissue was dissected and digested in enzyme solution at 37 °C for approximately 40 min to create a cell suspension. After filtration, centrifugation, and red blood cell lysis, the cells were resuspended in complete culture medium (Procell, CM-M003) and incubated for 45 min to remove the adhered fibroblasts. Finally, the AT2 cells in the culture medium were seeded onto a collagen-coated (354236, Corning; 5 μg·cm−2) culture flask for further culture.
2.16. RNA sequencing analysis of AT2 cells
We isolated highly pure AT2 cells from the lungs of 3- and 24-month old mice using flow cytometry techniques, following the methods recently published by our team [
38]. The total RNA was extracted from the AT2 cells and a quality assessment was performed. The RNA meeting the quality standards was used for library construction. We prepared paired-end (PE) libraries following the instructions provided in an ABclonal mRNA-seq Lib Prep Kit (ABclonal, China). The mRNA from 1 μg of total RNA was purified using oligo(dT) magnetic beads; this was followed by mRNA fragmentation in ABclonal (China) first strand synthesis reaction buffer. Subsequently, we utilized the mRNA fragments as templates to synthesize the double strands of complementary DNA (cDNA). We then ligated the synthesized double-stranded cDNA fragments with adapter sequences for polymerase chain reaction (PCR) amplification to obtain the final cDNA library. The PCR products were purified, and their library quality was assessed using an Agilent Bioanalyzer 4150 (USA). Transcriptome sequencing analysis was conducted using the Illumina NovaSeq 6000 platform (sequencing mode: PE150) at Shanghai Applied Protein Technology (China).
2.17. RNA extraction and RT-qPCR
The lung tissue samples were lysed with pre-chilled TRNzol reagent (DP424, TIANGEN, China) for 30 min for RNA extraction. The cDNA synthesis was conducted using a cDNA reverse-transcription (RT) kit (G592, Abm, China). Gene expression analysis was carried out using reverse transcription quantitative polymerase chain reaction (RT-qPCR) on a real-time PCR system (QuantStudioTM 5, Thermo Scientific Fisher, Singapore). The results were calculated as the fold change using the ΔΔCT approach, GAPDH as the reference gene for normalization. The sequences of the primers are provided in Table S4 in Appendix A.
2.18. Western blotting
For the Western blotting of the lung tissue samples, 20 mg of fresh tissue was lysed on ice for 30 min with 250 µL of lysis buffer (R0010, Solarbio), followed by homogenization and centrifugation. For the adherent cells in six-well plates, 100 μL of lysis buffer was introduced to collect the protein supernatant. The high-molecular-weight proteins were separated using sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and subsequently transferred onto a polyvinylidene fluoride (PVDF) membrane. The membrane was then incubated with specific antibodies targeting the desired protein (see Table S5 in Appendix A for detailed antibody information), followed by the detection and visualization of the bound antibodies.
2.19. Analyses of publicly available single-cell RNA sequencing databases
To compare gene expression differences between normal donors and idiopathic PF (IPF) patients, we analyzed three public databases (accession numbers GSE135893, GSE136831, and GSE122960) from the National Center for Biotechnology Information’s (NCBI) gene expression omnibus (GEO) platform. After data conversion, quality control, and filtering, a differential gene expression analysis between groups was conducted using the DESeq2 package (v 1.40.2), with thresholds of |log2 (fold change)| > 1 and adjusted p-value (Padj) < 0.01 to select differentially expressed genes (DEGs). The identified DEGs were further analyzed for average gene expression using the average expression function in the Seurat package (v 5.0), followed by visualization using the ggplot2 package (v 3.4.4). A correlation analysis between the DEGs and inflammation or the SASP in normal donors and IPF patients was conducted using a canonical correlation analysis (CCA) and gene set enrichment analysis (GSEA).
To compare SERPINE1 gene expression in young and aged lung tissue samples, we analyzed single-cell RNA sequencing (scRNA-seq) data from normal donors in two public databases (GSE227136 and GSE122960). The samples were stratified according to age into younger (< 40 years) and older (> 50 years) groups. Seven datasets from younger donors aged 21 (GSM3489197), 22 (GSM3489193), 25 (GSM7103283), 27 (GSM7135600), 29 (GSM7135601 and GSM3489187), and 34 (GSM7135584), as well as seven datasets from older donors aged 55 (GSM7135599 and GSM3489185), 57 (GSM3489189), 58 (GSM7103281), 60 (GSM7103319), 63 (GSM3489182), and 71 (GSM7135596), met the inclusion criteria for subsequent gene expression analysis. The downloaded sample data from each study were normalized using SCTransform in Seurat (v 5.0), combined with CCA and the Integration function for batch correction. The cells were filtered, retaining those with more than 250 detected genes (nFeature_RNA) and less than 30% mitochondrial gene content (percent.mt). In addition, genes were required to be expressed in at least three cells to be retained. Doublet cells were identified and removed using DoubletFinder. Subsequent steps included dimensionality reduction, clustering analysis, cell type annotation, and an analysis of target gene expression. The data analysis scripts used were exclusively written in the R programming language (the New Zealand). The correlation between SERPINE1 and ECM-related genes was characterized using Pearson correlation coefficients.
2.20. Statistical analyses
This study initially analyzed the correlation between aging and PF through scRNA-seq databases of lung tissue obtained from both normal donors (number of samples (n) = 45) and fibrotic patients (n = 51). Subsequently, explorations were conducted at both the animal level (n ≥ 3) and the cellular level (n = 3). Finally, the findings were validated using human lung tissue scRNA-seq databases (n = 7) and samples (n ≥ 4). The statistical analyses were performed using IBM SPSS Statistics 26.0 software. For comparisons between two groups, two-tailed t tests were utilized to calculate P values (with *P < 0.05 and **P < 0.01 indicating significance, and “ns” denoting non-significance). For multiple group comparisons, post hoc multiple comparisons were conducted via Duncan’s test within one-way analysis of variance (ANOVA), with lowercase letters indicating intergroup differences (P < 0.05). We utilized Prism 9.0 software (USA) for data visualization.
3. Results
3.1. Aging is the primary factor underlying PF
To pinpoint the predominant initiating factors of PF, we analyzed three public databases (accession numbers GSE135893, GSE136831, and GSE122960) from the NCBI GEO platform, which collectively included scRNA-seq data from the lung tissues of 45 normal donors and 51 IPF patients. Our analysis of the age distribution revealed a predominant occurrence of IPF in individuals aged 50 and above. Among them, the highest number of cases (62.7%) was observed in the 60- to 70-year old age group (
Fig. 1(a)), indicating a robust correlation between PF and advanced age. There were significant differences in the gene expression profiles of healthy donors versus IPF patients. We subsequently conducted differential gene expression screening in the two groups and identified 593 upregulated and 171 downregulated DEGs in the lungs of IPF patients versus those in the lungs of normal donors (
Fig. 1(b)). Through CCA of these DEGs, we observed that these DEGs had a stronger correlation with SASP factors than inflammatory factors (
Fig. 1(c)). The GSEA results further demonstrated significant SASP enrichment in the DEGs (
P < 0.01), with 17 out of 45 genes overlapping and an odds ratio of 6.7 (
Fig. 1(d)). These findings imply that aging might serve as the primary catalyst for PF development.
3.2. Pulmonary function declines and tissue structures are remodeled in aging mice
To further investigate the influence of aging on PF formation, we conducted a comprehensive and thorough study employing a natural aging mouse model. First, we utilized the WBP system to elucidate the associated changes in pulmonary function in aging mice and detected typical functional parameters, as illustrated in
Figs. 2(a)-(i). As mice age, their respiratory frequency (
F) decreases (
Fig. 2(b)), necessitating a greater tidal volume (TVb) to maintain an adequate minute volume (
Fig. 2(c)). The results indicate that, with increasing age, the respiratory pattern of mice transitions to deeper and slower breathing. A significant age-dependent decrease in mid-expiratory flow (EF50;
Fig. 2(d)) and an increase in relaxation time (Tr;
Fig. 2(e)) suggest obstructive ventilation impairment and increased gas exchange resistance in aging mice. In addition, we observed a lengthening of the inspiratory time (Ti) and expiratory time (Te) with age (
Figs. 2(f) and
(g)), along with a decrease in the peak inspiratory flow (PIFb) and peak expiratory flow (PEFb;
Figs. 2(h) and
(i)). These findings indicate weakened respiratory muscle strength, reduced lung compliance, and the onset of obstructive ventilation impairment, resulting in an overall decline in pulmonary function during the natural aging process.
Pulmonary function is endowed by an intact lung structure; therefore, we isolated the lungs of the mice and examined their tissue structure. We observed that the lung tissue in young mice at 3m and 6m was pinkish-white, soft, intact in the lobes, smooth on the surface, elastic, and glossy. However, middle-aged mice at 12m exhibited dull and lusterless lung tissue. At 18m, congestion occurred primarily in the lung hilum. In 24 month old mice, congestion was worsened, with bleeding spreading from proximal to distal parts of the lung and individual lobes exhibiting sclerotic changes. At 28m, the mice exhibited brownish lung tissue with severe sclerotic changes, visible cracks, and even developing tumors (
Fig. 2(j)).
Further pathological observations of the alveolar structures were conducted on the distal and proximal lung through H&E staining (
Fig. 2(k)). Here, the proximal lung refers to the region closer to the pulmonary hilum, where we found that young mice at 3m and 6m exhibited a greater number of well-formed alveoli with clear bronchial wall layers. However, at 12m, the number of alveoli began to decrease. By 18m, alveolar collapse and damage were observed, along with the local bleeding and infiltration of inflammatory cells, such as lymphocytes and macrophages. At 24m, the aged mice showed severe tissue disruption in the form of alveolar distortion and disorganization. Inflammatory cells, epithelial debris, and fibrous exudates clustered together, leading to obstruction of the alveolar cavity and the formation of sleeve-like structures in the trachea, accompanied by significant diffuse bleeding areas. At 28m, the lesions had worsened, leading to significant lung tissue remodeling. In addition, the structural changes in the distal lung were mainly characterized by age-dependent alveolar expansion and collapse, with inflammation occurring at 24m of age.
3.3. The occurrence of PF accompanies dynamic shifts in collagen fiber subtypes during lung aging
Observations of pulmonary tissue morphology have revealed that lung lobes undergo gradual sclerotic changes with aging—a clinical manifestation of PF. To assess fibrosis in the lung tissues of mice at various ages, we quantified the protein and mRNA expression of α-smooth muscle actin (α-SMA), a key marker for evaluating the extent of fibrosis. Our results revealed an age-dependent increase in α-SMA (
Fig. 3(a)) and
Acta2 (
Fig. 3(b)) expression in mouse lung tissues across various age groups, indicating that aging contributes to increased lung tissue fibrosis. In addition, we observed distinct differences in the collagen fiber ultrastructure in lung tissue from 3 and 28 month old mice (
Fig. 3(c)). Specifically, the collagen fibers at 3m exhibited loose cross-linking, an orderly arrangement, and a bundled distribution, whereas those at 28m displayed tight cross-linking and a diffuse distribution. We further performed Masson staining to evaluate the overall collagen fiber content in the lung tissues. Microscopic examination revealed blue staining, indicating collagen fibers, and red staining, indicating muscle fibers. As shown in
Fig. 3(d), the results revealed extensive blue staining in the mouse lung sections at 3m and 6m, primarily around the trachea and bronchi, in a patchy or speckled distribution. However, blue staining was minimal in the middle-aged group (12m), indicating a significant reduction in collagen fibers in comparison with the young mouse group. In the aged group at 18m, there was a slight increase in the area of blue staining. Collagen fibers were clearly proliferating in the lung tissues of the 24 and 28 month old mice, as large continuous blue-stained areas were visible, forming a mesh-like structure surrounding the alveoli. Thus, the total content of collagen fibers in the mouse lung tissue exhibited a dynamic pattern of initial decrease followed by an increase during aging. The aforementioned results suggest that the subtypes of collagen fibers in lung tissue may undergo dynamic variations with age.
Subsequently, Sirius Red staining was used to identify alterations in the collagen fiber subtypes in lung tissue during aging. As shown in
Fig. 3(e), Type I collagen fibers (Col I) and Type III collagen fibers (Col III) were the predominant types in the pulmonary interstitium, with an increase in red-stained Col I and a decrease in green-stained Col III with age. The ratio of Col I to Col III showed a significant upward trend. These results indicated that aging alters Col I and Col III quantity, spatial arrangement, and the Col III/Col I ratio. To confirm this discovery, we further quantitatively analyzed the expression of collagen type I alpha 1 (Col1a1) and Col3a1 in lung tissue. The findings revealed that the expression level of Col1a1 in the lungs of the aged mice was approximately twice that in the young mice (
Fig. 3(f)). Col I governs lung stiffness, while Col III determines lung compliance. As a result, the increased presence of Col I in lung tissue significantly contributes to the development of fibrosis.
The 18 months middle-aged period was found to be a crucial time point for fibrosis development. Col I increased in the lung tissue, with no significant inflammation observed, indicating that aging—rather than inflammation—was the primary factor driving the upregulation of Col I expression. Consequently, our investigation focused on elucidating the mechanisms by which aging influences the upregulation of Col I in lung tissue.
3.4. Senescent AT2 cells promote Col I synthesis
Col I is a primary component of the interstitium of fibrotic lung tissue. It is synthesized from two Col1a1 chains and one Col1a2 chain (
Fig. 4(a)), while Col III is synthesized from three Col3a1 chains. We further elucidated the influence of senescent AT2 cells on collagen secretion from fibroblasts through conditioned media culture and Transwell coculture methods and performed quantitative analyses of specific collagen chains (
Fig. 4(b)). The limited yields and viability of primary AT2 cells from aged mice presented significant challenges in acquiring a substantial quantity of cells for subsequent experiments. To address this issue, we opted to utilize primary AT2 cells obtained from four week old mice as well as A549 cell lines in our follow-up studies.
First, we extracted primary AT2 cells from the lungs of four week old mice and induced cellular aging using 30 g·L
−1 D-galactose. We observed a 1.44-fold increase in the expression of the aging-associated protein p-21 and a 0.79-fold increase in the expression of the DNA damage marker Histone family member X phosphorylated on Ser-139 (γ-H2AX), while the levels of the cell cycle-related proteins Cyclin B1 and p-Histone H3 decreased by 0.6-fold and 0.5-fold, respectively (
Fig. 4(c)), indicating successful establishment of the aging model. Subsequently, by culturing NIH3T3 fibroblasts with senescent AT2 cell-conditioned media, as shown in
Fig. 4(d), we observed an increase in the α-SMA, Col1a1, and Col1a2 protein levels in NIH3T3 cells, whereas Col3a1 expression decreased. A further analysis of the collagen-synthesis-related proteins revealed elevated phosphoserine transaminase 1 (PSAT1), phosphoglycerate dehydrogenase (PHGDH), and heat shock protein 47 (Hsp47) expression levels, indicating a notable increase in Col I synthesis relative to Col III degradation. Excessive deposition of Col I could lead to PF development. Thus, the secretions from senescent AT2 cells may primarily facilitate fibrosis formation by enhancing the synthesis of Col I in fibroblasts.
To validate this concept, we further established an aging model using A549 cells, a typical alternative cell line for AT2 cells, and cultured NIH3T3 cells with A549-cell-conditioned media. Similarly, we observed a marked upregulation in the expression of α-SMA and Col1a1 and the collagen synthesis-related proteins PSAT1, pyrrolin-5-carboxylate synthase (P5CS), and Hsp47, with a decrease in Col3a1 expression (
Fig. 4(e)). Next, to further demonstrate the profibrotic effect of senescent AT2 cell-conditioned media, we conducted Transwell cocultures of aging A549 cells with two types of fibroblasts: NIH3T3 cells (derived from mice) and WI-38 cells (derived from humans). As shown in Figs. S1(a) and (b) in Appendix A, we observed similar results in both cell types. Thus, we conclude that secretions from senescent AT2 cells can stimulate fibroblasts to produce Col I, thereby promoting fibrotic lesions in aging lung tissue.
3.5. Secretion of the PAI-1 factor from aging AT2 cells promotes Col I synthesis
To explore the impact of the SASP on fibrosis, we first conducted extensive literature screening for key profibrotic factors involved in the SASP and quantified their levels in lung tissue and primary AT2 cells (
Fig. 5(a)). Based on the RT-qPCR results, lung tissues from 28 month old mice showed marked upregulation of
Serpine1,
Tnf,
Fn1,
Tnc,
Csf3, and
Mmp3 at the mRNA level compared with those from 3-month-old mice (
Fig. 5(b)). In particular,
Serpine1 had the highest average expression level in the 28m lung tissue—1.67 times greater than that in the 3m lung tissue. Furthermore, an analysis of RNA-seq data from the primary AT2 cells of mice aged 3m and 24m revealed distinct expression patterns of the profibrotic factors between the two age groups, with notable increases in
Mmp2,
Edn1,
Vegfa,
Jag2,
Serpine1, and
Fn1 observed with advancing age (
Fig. 5(c)). Notably, the expression of Serpine1 (plasminogen activator inhibitor-1 (PAI-1)), a critical SASP factor, significantly increased in aging lung tissue and A549 cells (
Fig. 5(d)). Therefore, PAI-1 may be a pivotal profibrotic factor in the secretome of aging AT2 cells. We stimulated fibroblasts with various concentrations (0, 0.5, 1.0, and 2.0 μg·mL
−1) of PAI-1 for 48 h to validate this hypothesis. The results revealed a dose-dependent upregulation of α-SMA and Col1a1 expression as the concentration of PAI-1 increased (
Figs. 5(e) and
(f)). Concurrently, the levels of collagen-synthesis-related proteins increased, particularly PSAT1, P5CS, and Hsp47. These results suggest the substantial involvement of PAI-1 in promoting Col I synthesis in fibroblasts and triggering fibrogenesis.
3.6. PAI-1 primarily promotes Col I synthesis in fibroblasts via the TGF-β pathway
To further elucidate the mechanism by which PAI-1 promotes the synthesis of Col I, we investigated the transcription factors regulating Col1a1. Small mother against decapentaplegic (Smad) 2/3 has emerged as a key regulator that binds to the promoter region and regulates the expression of the Col1a1 gene [
39], [
40]. By targeting this transcription factor and conducting cellular-level detection, we observed significant upregulation of Smad2/3 and p-Smad2/3 expression in fibroblasts after 48 h of PAI-1 stimulation (
Figs. 5(g) and
(h)). Smad2/3 plays a crucial role as a regulatory protein in the TGF-β pathway [
41]. Further examination of the TGF-β pathway receptor p-TβRI revealed increased expression, indicating substantial activation of the TGF-β pathway. Hence, it is probable that PAI-1 primarily regulates the synthesis of Col I in fibroblasts via the TGF-β/Smad pathway.
We conducted studies using both inhibitor blockade and siRNA gene-silencing methods to validate our hypothesis. The results demonstrated that blocking the TGF-β pathway with SB431542 significantly reduced PAI-1-induced Col1a1 expression (
Fig. 5(i)), confirming the role of PAI-1 in promoting Col1a1 synthesis via the TGF-β pathway. Moreover, the targeting of Smad2/3 by siRNA effectively suppressed PAI-1-induced Col1a1 expression (
Fig. 5(j)), indicating that PAI-1 is a key regulator of collagen I synthesis through the transcription factor Smad2/3. In conclusion, PAI-1 from aging AT2 cells primarily promotes Col I synthesis and fibrosis occurrence via the TGF-β/Smad2/3 pathway.
3.7. Aged human lung tissue exhibits Col I deposition and high PAI-1 expression secreted by AT2 cells
To validate the expression of PAI-1 in the aging population and its association with fibrosis, we analyzed publicly available scRNA-seq datasets from normal donors obtained from the NCBI GEO database (accessions GSE227136 and GSE122960). Significant differences were observed in the gene expression profiles of lung tissues between younger and older people, suggesting that aging results in changes in pulmonary gene expression (
Fig. 6(a)).
Serpine1 was highly expressed in the lung tissue of aged mice, similar to the expression observed in lung tissue samples from aged humans, with the mRNA level of
SERPINE1 being nearly six times greater than that in those from younger individuals (
Figs. 6(b) and
(c)). We further annotated different cell types in lung tissue (
Fig. 6(d)) and subclustered AT2 cells and found a notable reduction in the quantity of AT2 cells in aged lung tissue compared with young lung tissue (
Fig. 6(e)). Importantly, the expression level of
SERPINE1 in Older-AT2 cells was over five times greater than that in Younger-AT2 cells (
Fig. 6(f)). These results revealed a substantial increase in
SERPINE1 expression in aged lung tissue and AT2 cells. A Pearson correlation analysis revealed a correlation between
SERPINE1 and ECM-related genes (
Fig. 6(g)), providing additional evidence for the intimate connection between increased PAI-1 expression during aging and PF development.
To further validate the aforementioned findings, we collected adjacent noncancerous lung tissues from individuals in different age groups, with younger samples aged 15, 17, 31, 39, and 40 a, and older samples aged 60, 63, 64, 66, and 71 a, and conducted histological analyses. H&E staining revealed that the lung alveolar structures in the younger individuals were intact, morphologically regular, and more abundant, whereas the lung tissue in older individuals was mainly characterized by disrupted and deformed alveoli with inflammatory infiltration (
Fig. 7(a)). Masson staining revealed an obvious increase in blue-stained areas in the lung tissue of older individuals, which were distributed in a patchy pattern around the bronchi (
Fig. 7(b)).
In addition, Sirius Red staining was used to assess changes in collagen fiber types within human lung tissues. The results showed that the lung tissues of younger individuals predominantly consisted of green-stained Col III, whereas those of older individuals were primarily composed of red-stained Col I (
Fig. 7(c)), in line with the observations from the aging mouse model. Subsequently, Western blotting revealed significantly greater expression levels of the α-SMA, Col1a1, and PAI-1 proteins in the lung tissue of individuals over 50 years of age in comparison with individuals under 40 years of age (
Fig. 7(d)). Finally, we performed immunofluorescence staining to assess the colocalization of AT2 cells and PAI-1 in human lung tissue. The results revealed a notable increase in the level of PAI-1 secreted by AT2 cells with aging (
Fig. 7(e)). These findings are consistent with previous findings from animal experiments.
4. Discussion
PF is an age-dependent lung disease, and the risk of its occurrence increases significantly with age [
42], as confirmed by our experiments on aging mouse models and studies on older populations. Furthermore, prior studies have demonstrated an upward trend in Col I and Col III levels in mouse models of PF [
43], [
44]. However, our research revealed dynamic alterations in the quantity and types of collagen fibers during aging-induced PF progression, diverging from previous reports. Notably, 18 months marks a critical onset point for PF in mice, accompanied by a remarkable increase in Col I and a notable decrease in Col III levels. Col I exhibits a bundled distribution and strong tensile strength, while Col III is loosely cross-linked and possesses good elasticity. Maintaining an adequate balance between Col I and Col III is essential for preserving the structural integrity and functionality of lung tissue. Our findings revealed shifts in the quantity, spatial arrangement, and ratio of Col I and Col III during aging. These alterations contribute to heightened lung tissue stiffness, significantly affecting intercellular signaling and force transmission and thereby impairing lung tissue contraction and expansion. Collagen metabolism in the lungs becomes severely disturbed during aging, leading to collagen network restructuring. The ratio of Col I to Col III is crucial for the early detection and assessment of PF lesions.
The content of Col I, a primary component of the ECM, plays a dominant role in fibrosis formation. However, there is limited research on the regulatory mechanisms of Col I synthesis during aging. It is believed that the sustained damage and functional decline of AT2 cells due to aging contribute to the disrupted repair of fibroblasts and pathogenic activation, serving as intrinsic mechanisms in the development of PF [
27]. During the occurrence and development of PF, damaged AT2 cells secrete a significant concentration of senescence-associated secretory SASP proteins and cytokines, such as PDGF, TGF, TNF-α, PAI-1, prostaglandins, and clotting factors [
25], [
45]. These factors form the microenvironment within lung tissue and collectively regulate the behavior of fibroblasts. However, the critical regulatory factors involved in this process and their mechanisms of action in relation to Col I synthesis remain unclear. Here, we investigated the direct communication between AT2 cells and fibroblasts. Through factor screening and cellular validation, we identified PAI-1 as a crucial profibrotic factor secreted by aging AT2 cells, significantly enhancing α-SMA and Col1a1 expression. Our discovery is bolstered by existing research showing that PAI-1 is associated with the fibrotic processes in several organs, including the lungs, liver, and kidneys [
46].
Previous research [
47], [
48] has reported increasing PAI-1 expression with age in wild-type and aging model mice and humans with age-associated diseases, including PF. In line with prior findings, we observed an age-related increase in PAI-1 levels in the lung tissues of naturally aging mice. Furthermore, PAI-1 expression in the lung tissues of older individuals notably surpasses that in younger individuals, demonstrating a synchronized trend with the expression of fibrosis-related proteins. Previous reports have indicated that PAI-1 primarily inhibits the proteolytic activation of plasminogen. PAI-1 expression upregulation disrupts the balance of the fibrinolysis system
in vivo, impeding ECM degradation and leading to fibronectin accumulation, thereby promoting PF development [
49]. Our study indicates that PAI-1 serves not only as an SASP factor but also as a critical profibrotic factor that can promote the development of PF by participating in the synthesis of Col1a1. Furthermore, we found that AT2 cells are a significant source of PAI-1 in fibrotic lung tissue—a conclusion supported by previous studies [
45], [
50]. Histological analyses of PF in previous studies revealed that 93.4% of the PAI-1-positive cells were surfactant protein C (SP-C) positive, whereas 54.7% of the SP-C-positive cells were PAI-1 positive [
51]. More specifically, targeting AT2 cell injury significantly increased PAI-1 expression in lung tissue [
45], [
52]. This finding suggests that aged AT2 cells can establish direct communication with fibroblasts through PAI-1.
The TGF-β pathway is a crucial signaling cascade that governs fibroblast collagen secretion [
33]. Earlier research [
53] has suggested that activation of the TGF-β/Smad pathway in fibrotic tissue can upregulate the expression of the downstream gene PAI-1. However, our research revealed that PAI-1 serves not only as a downstream effector of TGF-β but also as an upstream regulatory factor of TGF-β-induced fibrotic activity. To be specific, PAI-1 can activate the phosphorylation of the transcription factor Smad2/3, which then translocates into the nucleus and binds to the promoter region of the target gene Col1a1, regulating the transcription of the Col1a1 gene and thereby facilitating the deposition of Col I (
Fig. 8). The increased degree of fibrosis further upregulates the expression of PAI-1 through the TGF-β pathway, establishing a self-regulating positive feedback loop.
Our study elucidates the direct crosstalk between aging AT2 cells and fibroblasts, systematically explaining the intricate and close connection between PAI-1 secreted by AT2 cells and Col I synthesis. These discoveries provide crucial evidence regarding the contribution of epithelial dysfunction to the onset and progression of PF. Aging AT2 cells serve as a proximal driver and therapeutic target of PF. Future therapeutic strategies may involve the early, targeted delivery of senolytic drugs to attenuate aging-related pathways and eliminate aged AT2 cells or the repeated administration of siRNAs targeting PAI-1 and small-molecule inhibitors in the lungs to inhibit excessive synthesis of Col1a1, aiming to delay or halt PF progression.
5. Conclusions
Pulmonary aging is the predominant driver of PF. Our research suggests that the ratio of Col I to Col III in lung tissue tends to increase during aging. Col I is a critical component of the ECM, and elevated Col I levels significantly contribute to PF development. Moreover, aging AT2 cells can facilitate Col I synthesis through intercellular communication with fibroblasts. Our research reveals that PAI-1, a crucial factor secreted by aging AT2 cells, exerts a pivotal role in promoting the synthesis of Col1a1 in fibroblasts, subsequently leading to Col I deposition and driving the progression of PF by mediating the TGF-β/Smad2/3 pathway. A single-cell database analysis of human lung tissue and histological analysis further confirmed the close relationship between PAI-1 and PF during aging. This study provides pivotal evidence of the impact of age-related epithelial dysfunction on the progression of PF, offering novel therapeutic targets for clinical intervention.
Acknowledgments
The work was supported by the Young Elite Scientist Sponsorship Program by CAST (2022QNRC001) and the 111 project of the Education Ministry of China (B18053).
Compliance with ethics guidelines
Rui Quan, Chenhong Shi, Yanan Sun, Chengying Zhang, Ran Bi, Yiran Zhang, Xin Bi, Bin Liu, Ziheng Dong, Dekui Jin, and Yixuan Li declare that they have no conflict of interest or financial conflicts to disclose.
Appendix A. Supplementary data
Supplementary data to this article can be fo==und online at
https://doi.org/10.1016/j.eng.2024.08.014.
PDF
(6502KB)
