Pathological remodeling of distal lung matrix in end-stage cystic fibrosis patients

Abstract

Background.

Manifestations of cystic fibrosis, although well characterized in the proximal airways, are understudied in the distal lung. Characterization of the cystic fibrosis lung ‘matrisome’ (matrix proteome) has not been previously described, and could help identify biomarkers and inform therapeutic strategies.

Methods.

We applied liquid chromatography mass spectrometry, gene ontology analysis, and multi-modal imaging (histology, immunofluorescence, electron microscopy) to provide comprehensive evaluation of distal human lung extracellular matrix structure and composition in end-stage cystic fibrosis.

Results.

Quantitative proteomic profiling identified sixty eight (68) extracellular matrix constituents that were significantly altered in end-stage cystic fibrosis. Although the total amount of matrix in cystic fibrosis lungs was not significantly different from normal lungs, over 90% of the matrix peptides detected, including structural and basement membrane proteins, were expressed at lower levels in cystic fibrosis lungs, suggesting that cystic fibrosis leads to loss of diversity among lung matrix proteins rather than an absolute loss of matrix. Visualization of distal lung matrix via immunofluorescence and electron microscopy revealed pathological remodeling of tissue architecture and loss of the alveolar basement membrane, consistent with significantly altered pathways identified by gene ontology analysis.

Conclusions.

These data indicate that obliteration of distal lung tissue structure and loss of matrix protein diversity are hallmark features of end-stage cystic fibrosis. While novel therapeutics aim to functionally restore defective CFTR, drugs that target dysregulated matrix pathways may serve as adjunct interventions to support recovery of lung function.

INTRODUCTION

Cystic fibrosis (CF) is a chronic disease caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. Defective CFTR expression results in a pathologic cascade of impaired ion transfer, dysregulation of airway surface liquid and mucus, inability to clear infection, chronic inflammation, and ultimately end-stage lung disease necessitating transplantation at a median age of 29.5 years [1,2]. While disease manifestations in the proximal airways are well characterized, alterations to the distal lung parenchyma and extracellular matrix (matrix) in CF are relatively understudied.

Lung matrix structure and composition facilitate gas exchange by providing mechanical integrity to withstand dynamic changes in pressure during breathing, elasticity to ventilate the distal alveoli, and biochemical cues to guide cell behavior and response to injury. Key structural and biological matrix constituents in the lung include collagens and laminins which comprise the basement membrane, and elastin which provides the lungs with elastic recoil enabling ventilation. In healthy patients, coordinated matrix remodeling governs lung tissue repair to restore function of an injured area. In diseased lungs, however, the matrix and its reparative wound-healing functions are compromised. Alterations in matrix composition and structure have been demonstrated in idiopathic pulmonary fibrosis (IPF), chronic obstructive pulmonary disease (COPD), emphysema, and asthma [3–5]. In CF, chronic infection and inflammation persist due to impaired immune response and poor mucociliary clearance, resulting in aberrant tissue repair that involves the matrix.

Previous studies identified elevated levels of proteolytic enzymes and matrix breakdown products in the bronchoalveolar lavage (BAL) fluid, sputum, and serum of CF patients, suggesting that matrix alterations mediated by protease/anti-protease imbalance may be a feature of CF [6–9]. However, a comprehensive quantitative analysis of the CF lung matrisome, which includes constitutive and associated matrix proteins, has not been reported. Understanding pathologic alterations to the lung matrix in CF may offer prognostic value and could uncover biomarkers of disease severity to inform therapeutic strategies. While emerging therapies aim to correct the CFTR mutation, adjunct interventions targeting matrix remodeling could be developed to slow or reverse tissue damage and lung decline.

Here, we report alterations in matrix and matrix-associated proteins in CF lung tissue through analysis of lung matrix structure and composition in end-stage CF. We hypothesized that alterations to the structure and composition of lung matrix and matrix-associated proteins accompany the well-described hallmarks of CF, including chronic infection and inflammation. Structural changes to the matrix were visualized at multiple scales by histology and electron microscopy. Profiles of collagens, glycoproteins, proteoglycans, secreted factors, matrix-regulating, and matrix-affiliated proteins were quantified by liquid chromatography-mass spectrometry (LC-MS/MS). Significant loss and structural breakdown of key matrix components, including basement membrane proteins, were observed. Our findings indicate that matrix degradation in the distal lung parenchyma accompanies stereotypical airway pathology in end-stage CF.

RESULTS

We investigated alterations to the distal lung matrix in end-stage CF, using tissue from the right middle lobe that was obtained from explanted lungs of CF patients undergoing lung transplantation. Patient mutations were determined to be ΔF508/ΔF508 (n=5), ΔF508/G542K (n=1), and ΔF508/unknown (n=2) (Supplementary Table 1). CF lung tissues were obtained from a random sample of patients with end-stage CF, and none of the samples were excluded from the study. As lung matrix provides structural and biochemical cues to resident cells which together govern development, function, and injury response, we sought to investigate pathologic alterations to matrix structure and composition in end-stage CF, using microscopy and liquid chromatography mass spectrometry (LC-MS/MS) analysis (Figure 1).

Figure 1. Experimental overview.

Cystic fibrosis and normal lung tissues were procured to extract and analyze the distal lung matrix structure and composition through proteomic profiling, multimodal imaging (e.g., histology, immunostaining, electron microscopy), and gene ontology.

Ultrastructural abnormalities in CF distal lung matrix

The lung matrix provides a complex, well-organized, 3D structure that supports lung function by providing a high surface area for gas exchange and allowing for lung breathing. For a global comparison of matrix structure and tissue morphology in CF compared to normal control lung, histological staining was performed on lung specimens from all patients to visualize overall tissue histomorphology (H&E), and the distribution of collagen (Masson’s trichrome, blue), and elastin (elastic van Gieson, black) (Figure 2A, Supplementary Figures 1 – 8). Histological staining of airway sections confirmed stereotypical airway morphology, including elevated levels of glycoproteins and mucins (Alcian blue, blue) in CF (Supplementary Figure 9). Compared to normal lung, CF specimens displayed septal thickening, fragmented fibers, and proteinaceous fluid-filled airways and alveoli with varying degrees of severity.

Figure 2. Structural comparison of cystic fibrosis and normal lung tissue.

Overall tissue structure and composition of the lung parenchyma were evaluated on (A) histology with hematoxylin and eosin (H&E), Masson’s trichrome (blue, collagens), and elastic van Gieson (EVG) (black, elastin), and (B) scanning electron microscopy (SEM) and transmission electron microscopy (TEM). * = Airspace, E = elastic fibers, ATII = alveolar type II pneumocyte, arrows = lamellar bodies.

Lung parenchymal ultrastructure and alveolar architecture were grossly abnormal in CF with airway and alveolar obstruction and collapse, as evidenced by scanning electron microscopy (SEM) imaging, a finding indicative of parenchymal destruction. Visual assessment of the epithelium, endothelium, and alveolar basement membrane by transmission electron microscopy (TEM) revealed basement membrane disruption and alveolar collapse (Figure 2B).

Quantitative analysis of CF distal lung matrisome

We performed LC-MS/MS on intact distal lung tissues from CF patients and normal controls to identify changes to the lung proteome. Gene ontology (GO) analysis of detected peptides revealed that the extracellular space (p = 4.0E-4) and extracellular organelles (p = 1.68E-5) were significantly altered in CF compared to normal control (Supplementary Figures 10 – 14). Identification of significant differences in pathways involving the extracellular environment informed deeper analysis of the CF lung matrix.

Towards focused analysis of lung matrix in CF and normal lungs, we (i) extracted matrix from CF and normal distal lung tissues, (ii) quantified composition of extracted matrix using LC-MS/MS, and (iii) filtered the resultant LC-MS/MS dataset to identify constituents of the CF and normal lung matrisome [10,11]. Matrix was extracted by decellularization of lung tissues, which led to 95% reduction in DNA content and nearly complete removal of DAPI-stained nuclei (Supplementary Figure 15), while maintaining intact matrix. Matrix was then processed using LC-MS/MS by modifying previously published methodologies [3,12–14].

The full list of peptides identified using LC-MS/MS was filtered using the open-access ‘Matrisome Project’ which categorizes matrix proteins as: core matrisome (i.e.,collagens, glycoproteins, proteoglycans), matrix-regulators, matrix-affiliated, and secreted factors [10,11]. The number of matrisome proteins and their relative intensities across each of these categories were compared for CF and normal lungs (Supplementary Figure 16).

Overall, 243 significantly different proteins were detected in the CF matrix, including 14 collagens, 18 glycoproteins, 4 proteoglycans, 10 matrix-regulators, 6 matrix-affiliated, 9 secreted factors, and 182 others (i.e., significantly altered but not considered to be matrisome constituents) (Figure 3B). All except 6 of the significantly altered matrix proteins were downregulated (MMP11, IL16, TGFB1, VWA5A, LGALS1 and HRG were significantly upregulated). We observed similar changes in the relative intensities of the proteins in the collagens, glycoproteins, and secreted factors categories (Supplementary Figure 16), indicating similar coverage in both groups. However, there was a decreased abundance of proteoglycans in CF compared to normal lung (p = 0.016), suggesting the loss of proteoglycan activity in CF. To validate the matrix extraction method, LC-MS/MS was performed on whole lung tissues, and a decrease in matrix proteins was also detected (Supplementary Figure 10).

Figure 3. Global analysis of cystic fibrosis lung matrisome.

(A) Principal component analysis (PCA) of CF and normal lung matrix. Shape of marker indicates patient mutations: circle, ΔF508/ΔF508; triangle, ΔF508/G542K; hexagon, ΔF508/unknown. (B) Volcano plot of differentially expressed matrisome components. (C) Gene ontology (GO) analysis of pathways significantly altered in CF. BP, biological processes; CC, cellular components; MF, molecular factors.

Using a principal component analysis (PCA) on the extracted matrix LC-MS dataset, we found that 7 out of 8 CF patients clustered together, while healthy patient samples were randomly distributed (Figure 3A). One patient (ΔF5O8/unknown) did not cluster, suggesting that some mutations may impact the matrix in a mutation-specific manner. Nevertheless, for patients with the most common mutations (ΔF5O8 and G542K), the altered matrix was of the same phenotype.

The full list of peptides identified through LC-MS/MS was subjected to GO analysis, which revealed significant differences in matrix-related pathways in CF vs normal lung matrix. Significantly different GO terms included: extracellular matrix (p = 3.16E-26), degradation of the extracellular matrix (p = 2.73E-7), collagen trimer (p = 1.09E-11), structural matrix constituents conferring tensile strength (p = 5.78E-11), collagen chain trimerization (p = 2.93E-10), and matrix proteoglycans (p = 5.86E-9) (Figure 3C). Additional significant GO terms related to cell-matrix interactions and cell signaling were identified, including anion binding (p = 1.19E-6), focal adhesion (p = 1.39E-6), cell-substrate junction (p = 1.97E-6), extracellular vesicle (p = 1.36E-28), and extracellular exosome (p = 3.38E-28). Complete list of identified GO pathways can be found in Supplementary Figures 17 – 20.

Degradation of core proteins in distal lung matrix

To further investigate structural abnormalities visualized on electron microscopy, core matrisome proteins (i.e., collagens, proteoglycans, glycoproteins) were quantified and compared in CF and normal lung using LC-MS/MS. In total, 36 core matrisome proteins were differentially expressed in CF matrix compared to normal lung. The majority of collagens (COL4A1/2/3/6, COL5A3, COL6A2/3/5/6, COL11A1, COL21A1, COL26A1, COL12A1, and COL18A1) were significantly downregulated in CF compared to normal tissue (Figure 4A). Other collagens (COL15A1, COL23A1, COL10A1, COL5A1, COL1A2) were upregulated in CF, albeit not significantly (Supplementary Figure 21). Other key basement membrane constituents, including multiple isoforms of laminin (LAMA2, 4, and 5) and nidogen1 (NID1) were downregulated in CF (LAMA2, log2FC = − 1.6, p = 0.025; LAMA4, log2FC = −1.2, p = 0.012; LAMA5, log2FC = −1.3, p = 0.047; NID1, log2FC = −1.3, p = 0.030), indicative of dysregulation of basement membrane in CF lung disease and consistent with TEM and SEM findings. Elastin (ELN, log2FC = −1.6, p = 0.026) which confers lung elasticity was decreased in CF compared to healthy lung, a trend indicative of possible emphysema (Figure 4B).

Figure 4. Characterization of core matrix structure and composition in cystic fibrosis and normal lung.

Expression of core matrisome components, including (A) collagens, (B) glycoproteins, and (C) proteoglycans using LC-MS/MS. (D) expression of basement membrane constituents and elastin using immunofluorescence.

Hierarchical clustering showed that patients 5 and 6 exhibited similar collagen and glycoprotein expression patterns as healthy patients, while all other CF patients clustered separately. Notably, perlecan (HSPG2, proteoglycan found in the basement membrane) downregulation accompanied the loss of collagens and laminins (HSPG2, log2FC = −1.36, p = 0.021). Other collagen-interacting proteoglycans were also downregulated, such as osteoglycin (OGN, log2FC = −1.21, p = 0.014), biglycan (BGN, log2FC = −1.08, p = 0.014), and asporin (ASPN, log2FC = −2.21, p = 0.001) (Figure 4C).

Immunofluorescent staining for basement membrane constituents, including collagen IV, elastin, and laminin, was also performed. In line with the matrisome quantification and GO analysis indicating collagen trimer and basement membrane as significantly altered pathways, collagen IV – a major component of the lung basement membrane – was sparse, poorly organized, and partially degraded in the alveolar septa in CF lung compared to normal lung where well-organized collagen IV staining was ubiquitous in the basement membrane. Elastin, which enables cyclic movement during breathing, and laminin, which provides structural support and promotes cell adhesion, appeared fragmented in CF lungs compared to controls (Figure 4D).

Dysregulation of matrix modifying proteins

To investigate matrix maintenance, regulation, and remodeling, we identified matrix-affiliates, matrix-regulators, and secreted factors that were differentially expressed in CF and normal lung. Degradation of basement membrane proteins was accompanied by decreased expression of proteins that promote cell adhesion to the basement membrane (C1QTNF5/7, log2FC = −1.3/−1.4, p = 0.017/0.002; SEMA3B, log2FC = −1.6, p = 0.046). Tetranectin (CLEC3B) was found to be downregulated in CF (log2FC = −1.62, p = 0.001). However, galectin-1 (LGALS1) was overexpressed in CF, at moderate to high levels (log2FC = 1.43, p = 0.009), indicative of infection response and myeloid cell recruitment (Figure 5A). Immunostaining of tetranectin in CF samples showed sparse staining, whereas galectin staining was moderately increased in CF compared to normal lung.

Figure 5. Analysis of matrix-associated proteins in cystic fibrosis and normal lung.

(A) matrix-affiliated, (B) matrix-regulators, and (C) secreted factors CF and normal lung matrisome. Representative immunostaining of differentially expressed proteins in each category. Arrows indicate positive cytoplasmic galectin staining. Asterisks indicate positive nuclear NFkB.

Matrix regulators are proteins responsible for crosslinking, degrading, and remodeling the matrix. Consistent with the breakdown of key matrix constituents, we observed significant downregulation of proteins that inhibit matrix degradation, including trypsin inhibitor (ITIH5, log2FC = −1.82, p = 0.002), serine protease inhibitor (SERPINA5, log2FC = −1.57, p = 0.013), and tissue inhibitor of metalloproteinase 3 (TIMP3, log2FC = − 2.62, p = 0.019). Similarly, lower levels of LOXL2 (log2FC = −1.05, p = 0.019), a collagen and elastin crosslinker, correlates with observed decreases in both collagen and elastin levels in CF lung. Matrix metalloprotease (MMP) expression varied, with upregulation of MMP11 and MMP8 and downregulation of MMP28 (Figure 5B).

Because our study was focused on tissue-level changes, alterations to protease levels in BAL or sputum were not captured in our analysis. Staining for MMP11 was more pronounced in CF compared to normal lung. Histidine rich glycoprotein (HRG) which is expressed by platelets and modulates angiogenesis, fibroblast proliferation, complement activation, coagulation, and fibrinolysis [15], was significantly upregulated (HRG, log2FC = 1.08, p = 0.014) and showed increased expression and more nuclear localization in CF compared to normal lung. Other alterations in proteins expressed by or associated with platelets were also observed, including upregulation of platelet derived growth factor subunit B (PDGF-B, log2FC = −1.7, p = 0.028), fibulin-1 (NS), and type 3 collagen (NS).

Among the secreted factors that were identified, inflammatory markers – including interleukin 16 (IL-16, log2FC = 3.9, p = 0.031) and transforming growth factor beta (TGF-β, log₂FC = 2.3, p = 0.024), were all significantly increased in CF compared to normal lung, whereas interleukin 17 (IL-17, log2FC = −2.2, p = 0.009) was downregulated. Several constituents of the WNT pathway were significantly downregulated in CF, consistent with previous findings that WNT and CFTR expression are linked [16]. To gain further insight into the WNT and inflammatory pathways in CF, expression of β-catenin and NF-kB were investigated, revealing increased nuclear localization of both proteins, indicative of WNT and NF-kB signaling (Figure 5C).

DISCUSSION

The objective of this study was to investigate alterations to the distal lung matrix in end-stage CF. Analysis of CF lung matrix composition and structure revealed pathological remodeling of tissue architecture, destruction of alveolar basement membranes, and significant loss of key constitutive and associated matrix proteins, including collagens, laminins, and elastin. Obliteration of distal lung parenchymal structure and loss of matrix protein diversity were identified as features of end-stage CF.

While matrisome analysis has been performed on lung tissue samples from other chronic lung diseases, e.g., COPD or IPF, a matrix-based analysis of the CF lung has not been reported [3]. Previous studies have revealed elevated levels of proteolytic enzymes and matrix breakdown products in CF patient airway fluid. We sought to assess if destruction of lung matrix was also a unifying feature of end-stage CF. Through LC-MS/MS analysis of end-stage CF samples in comparison to normal lung tissue, we determined that the majority of detected matrix proteins (> 90%) were expressed at lower levels in CF compared to normal lung. Interestingly, the total amount of matrix proteins detected in CF and normal lung tissue was not significantly different, suggesting that CF does not change the total amount of matrix but instead leads to loss of diversity among matrix proteins.

In previous studies, matrix-degrading enzymes including neutrophil elastase, collagenases, serine proteases, and matrix metalloproteinases have been identified in airway secretions, BAL fluid, and sputum of CF patients, with elevated levels correlating with pulmonary exacerbations and respiratory decline [6,17–19]. Pathologically low levels of protease inhibitors including TIMPs, alpha1-antitrypsin, and SLPI have also been identified in CF, consistent with our findings [7,20–24]. Matrix breakdown products including collagen and elastin fragments were previously detected in BAL, serum, and urine of CF patients [9,25,26]. Taken together, these studies imply aberrant matrix remodeling with a strong inflammatory component as key contributors to CF progression.

Basement membrane constituents, including several isoforms of collagen IV and laminin, were degraded in end-stage CF, consistent with dysregulation of the basement membrane and alveolar structure observed on SEM and TEM. Degradation of basement membrane has been identified in many lung pathologies, including emphysema, lung cancer, acute respiratory distress syndrome (ARDS), and COPD [27–29], and may be mediated by an imbalance between proteases and protease inhibitors [30,31]. In particular, alveolar basement membrane degradation mediated by MMPs in COPD correlated with mortality [28]. Loss of basement membrane integrity in CF may contribute to cytokine release that triggers paracrine-mediated inflammation, as well as edema and alveolar fluid retention, contributing to impaired gas exchange [32,33].

Elastin, a protein that confers the lung with appropriate compliance and enables dynamic movement during breathing, was degraded in end-stage CF. This finding is consistent with the features of emphysema, in which protease and antiprotease imbalance and immune cell recruitment lead to irreversible destruction of the lung parenchyma [34–37]. Evidence of tissue breakdown, similar to that in emphysematous lung, was observed on histology in CF specimens. Prior work has suggested emphysema as a unifying feature in end-stage cystic fibrosis based on histopathological and radiographic analysis [38]. In addition to elastin degradation, we observed a loss of lysyl oxidase 2 (LOXL2), which functions as an elastin crosslinker to stabilize the elastic matrix. Loss of elastin fibers compromises lung elasticity, which is necessary for proper ventilation and oxygenation.

Gene ontology analysis revealed several pathways involving cell-matrix interactions and cellular function, including cell-substrate junction, focal adhesion, extracellular vesicles, cellular component biogenesis and cellular localization, indicating that matrix degradation has downstream effects on cell populations in the lung. Significantly altered matrix-associated proteins were identified, including galectin which mediates cell-matrix interactions and proliferation, and tetranectin which binds to plasminogen and may regulate secretion and exocytosis [39,40]. IL-16 and TGF-β were both significantly increased in CF lung tissue, indicative of the inflammatory process that contributes to aberrant wound healing and can signal epithelial-mesenchymal transition (EMT), a common pathway implicated in many lung diseases (e.g., asthma, COPD, IPF) [41,42]. Proteins involved in EMT, including several WNT proteins, were altered in CF compared to normal lungs. Dysregulated EMT has deleterious effects on lung cell phenotype and function by: (i) impairing epithelial barrier function and innate defense mechanisms, (ii) enabling migration of activated stromal cells throughout the lung matrix, and (iii) promoting stromal cell hyperplasia while compromising pseudostratified airway epithelium [41,43].

Several outstanding questions remain toward better understanding the complex changes in CF matrix. This study only evaluated end-stage CF lung samples obtained at the time of lung transplant. Future studies should assess temporal changes to the matrix and determine specific effects on cell phenotype and function. It is unclear if matrix breakdown is intrinsically driven by loss of CFTR expression or if chronic infection is the primary driver of matrix destruction. Comparison of matrix composition in other organs affected by CF (i.e., pancreas, liver) that are not undergoing an infectious process could offer insights into the relative contributions of intrinsic loss of CFTR versus chronic lung infection to matrix breakdown. Similarly, analysis of distal lung matrix composition in other lung diseases characterized by persistent respiratory infection, such as ciliary dyskinesia or non-CF bronchiectasis, may yield further understanding of the pathophysiologic processes that drive matrix breakdown.

Increased collagen deposition has been reported in liver and pancreas of CF patients [44,45], suggesting that matrix degradation in the lung may be primarily driven by chronic infection and corresponding immune response. Furthermore, elevated protease levels correlate with bronchiectasis and tissue damage in primary ciliary dyskinesia, indicating that persistent respiratory infection may trigger matrix breakdown pathways even in the presence of functioning CFTR [46]. Nevertheless, the extent to which CFTR loss contributes to parenchymal obliteration in CF should be further evaluated. Proteomic analysis of multiple sample types (i.e., tissue, serum, BAL fluid, urine, etc.) from individual patients may enable mechanistic understanding of disease processes involving matrix degradation. While sampling distal lung tissue via biopsy is challenging, identification of serum or BAL markers that correlate with lung matrix degradation could inform patient care, as previous studies have shown that administration of nebulized or intravenous medication can mitigate elevated proteolytic enzyme levels in BAL fluid [17,22,47].

The advent of novel combination therapeutics such as Trikafta has revolutionized the cycle-of-care for CF patients and led to dramatic improvements in functional lung parameters [48]. As the long-term benefits (e.g., increased life expectancy) of these drugs are yet to be determined, further therapeutic development is imperative, and the ability of CF medications to indirectly recover dysregulated matrix pathways or repair degraded matrix should be evaluated. For adult CF patients with moderate to significant lung destruction, nebulized proteolytic inhibitors or therapeutic matrix cocktails could serve as adjunct therapies to support recovery of lung function.

METHODS

Detailed methods can be found in the Supplementary Methods. Briefly, lung tissue was obtained at the time of transplantation from patients with end-stage cystic fibrosis (n=8) and from uninjured regions of donor lungs for normal control comparison (n=3). CF patient mutations were determined by isolating genomic DNA and sequencing regions of the CFTR gene where common mutations are found (Supplementary Table 2). Structure and composition of lung extracellular matrix was analyzed using histology, immunofluorescence, electron microscopy, liquid chromatography mass spectrometry (LC-MS/MS), and gene ontology (GO) analysis.