Abstract
Background
Congenital pulmonary airways malformations (CPAM) belong to a group of rare congenital lung anomalies whose pathological origin is still mostly unknown. The current research project aims to study the possible role of the underlying mesenchyme in CPAM pathophysiology by comparing data from fetal tissue and healthy lung with CPAM.
Methods
Tissue samples from CPAM patients and healthy adjacent parts were collected during planned surgical resections. Fetal lung tissue obtained from abortions was also used. The mesenchymal parts of pathological, healthy and fetal lung tissues were analyzed using quantitative proteomic, bulk RNA sequencing and multiplex immunohistochemistry to identify the difference in genes and proteins expressed in CPAM.
Results
When compared to healthy adjacent lung tissue, transcriptomic data of CPAM tissue highlighted downregulation of genes implicated in the transforming-growth factor-β and immune-related signaling pathways. Conversely, epithelial-mesenchymal transition genes were upregulated in CPAM, suggesting an abnormal branching associated with abnormal mesenchyme. Quantitative proteomic results showed that the expression of various proteins implicated in muscle differentiation, such as desmin, calponin1 and α-smooth muscle actin, were decreased in CPAM compared to healthy adjacent lung. Multiplex immunostaining confirmed these results with a more significant difference for CPAM type 1 compared to healthy adjacent tissue or to fetal lung. Several pathways known to be crucial in epithelial proliferation, differentiation or branching such as PI3K-AKT-mTOR pathway were dampened.
Conclusion
Our study reveals the presence of abnormal mesenchyme and muscle in and surrounding cystic tissue. Despite normal lung function during pediatric follow-up, the atypical mesenchyme with possible impaired epithelial-mesenchymal transition could potentially contribute to the development of late adult pathologies such as chronic obstructive pulmonary disease or lung tumors and should be assessed accordingly.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12931-025-03332-4.
Keywords: CPAM, Mesenchyme, Proteomic, Transcriptomic, Translational, Human
Introduction
Congenital pulmonary airway malformations (CPAM) belong to a group of rare congenital lung anomalies whose pathological origin is still barely understood [1, 2]. Postnatal clinical management of CPAM generally implies the surgical removal of the lesions to prevent two possible complications: a malignant transformation of these malformations and/or an increased risk of cystic malformation infection. Nowadays, some medical centres apply a more conservative approach consisting of a clinical follow-up without applying systematic surgery; the pro and contra arguments being adjusted between the low risks of such complications pondered by a surgical intervention [3–5]. CPAM classifications have been proposed by Stocker – CPAM types 1 to 4 [2], and by Langston – small and large cyst types [1]. These classifications rely on a macroscopic anatomical description completed by a histological description of the lesion performed after surgical resection [1, 2]. The classification of lesions is essential, as each CPAM type carries a different risk of complications. However, accurately assessing this risk remains challenging [6, 7], and there is a need for new classification schemes [8].
The pathogenesis of CPAM is not clear yet. It takes its origin during lung development, a complex process starting at the 4th gestational week and continuing after birth. As most embryological processes, the paracrine growth factor crosstalk between endothelial, epithelial and mesenchymal cells is central [9]. During pregnancy, the presence of external factors, such as toxic exposure or premature birth, can interfere with cellular interactions essential for lung branching and alveolar formation [1]. In addition, several genes regulating cellular proliferation and apoptotic processes may be altered, and potentially involved in the development of CPAM [10–12].
In the lung, mesenchymal cells provide structural support and survival cues to the adjacent bronchial and alveolar epithelium. They also play key roles in lung growth, and alveolar differentiation [13]. These cells encompass a variety of cell types, including resident fibroblasts, lipofibroblasts, myofibroblasts, smooth muscle cells (SMCs), and pericytes, all of which are located at different anatomic regions and exhibit diverse homeostatic functions. During pulmonary airways branching, mesenchymal cells act both as signaling centers for proliferation, differentiation and morphogenesis and equally as a structural support [14]. The terminal bifurcation development of pulmonary tips involves a well-established sequence of events leading to myofibroblast differentiation [15]. These myofibroblasts may develop through an epithelial-mesenchymal transition (EMT) process and also take part in alveologenesis [9, 16]. In different pathological conditions, it is well accepted that fibroblasts can differentiate into myofibroblasts exhibiting bundles of microfilaments and expressing α-smooth muscle actin (α-SMA, ACTA2), an actin isoform typical of SMCs [17]. Previous studies from our laboratory demonstrated that the presence of α-SMA-positive cells in CPAM1 was similar to that in healthy airways but decreased in CPAM2 [12]. In contrast, myosin-11 (MYH11), another marker of SMCs, exerting a central role in driving branching morphogenesis, possibly due to its ability to induce SMC contraction is decreased in CPAM2 [17]. SMCs are also described to be involved in airway diameter and branching morphogenesis [17]. Studies performed on laser microdissected fetal and postnatal CPAM epithelia showed an increased expression of different members of the transforming growth factor β (TGFβ) signaling pathway such as bone morphogenetic protein 4 (BMP4), TGFβ2 or its receptor ΤGFβR1, all implicated in EMT [10, 19]. Nevertheless, the key features and signaling pathways leading to CPAM development and the implication of muscle cells or EMT are not clear yet.
To unravel new aspects implicated in CPAM incidence, we focused our study on the possible role of the underlying mesenchyme in CPAM development. For this purpose, we compared proteomic and transcriptomic data from fetal tissue with the two most frequent CPAM subtypes, CPAM1, CPAM2 and healthy adjacent tissue.
Material and methods
Study design, subjects and description of types of lesions
Ten children with CPAM diagnosed by antenatal ultrasound and confirmed by post-natal thoracic CT-scan) were prospectively enrolled at the Children’s Hospital of Geneva at the time of surgery from November 2012 to November 2017. The institutional ethics committee approved this study and informed consent was obtained during scheduled hospital visits (Ethical Committee of the Univrsity Hospital of Geneva, CER 12–110).
A pulmonary lobectomy was performed in all patients, and histopathological CPAM classification was established by macroscopical and histological examination of the specimens by a pathologist before further analysis [18]. All surgeries were performed preventively on asymptomatic patients between 9 and 18 months of age. For each patient, paraffin-embedded and frozen tissue was obtained from both the pathological CPAM tissue and from the healthy adjacent lung which was used as control. All the participants had a regular follow-up by the pulmonologist before and after surgery. Symptomatic patients at birth needing early postnatal surgery or patients having a sequestration associated with CPAM were excluded of this study.
Lung functions of 8-year-old patients were performed using Medisoft BodyBox 5500. Forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC), FVC/FEV1 are expressed in z-scores (normalized for the height, weight and ethnicity), total lung capacity (TLC) and diffusion (DLCO) are expressed in percent (adapted for the height, weight, ethnicity and hemoglobin fir the DLCO). Global lung initiative (GLI) reference values were used for spirometric and lung diffusing capacity results.
Human fetal samples were collected after the termination of pregnancy of fetuses without lung malformations or abnormal lung development. The age of the fetuses ranged from 14 to 19 weeks of gestation corresponding to the pseudoglandular stage of lung development. The institutional ethics committee approved this procurement, and informed consent was obtained from the parents (PB_2016–00175).
All the different experiments and analysed tissues are summarised in supplementary Table 1.
RNA extraction and transcriptomic analysis
Twenty μm slides were cut from OCT tissue blocs for RNA extractions. Four samples of CPAM1 (n = 2), CPAM2 (n = 2), and 4 control lungs were analyzed. An RNA integrity number (RIN) greater than 7 was considered satisfying for the analysis [10]. The samples were sequenced on an Illumina NovaSeq 6000 sequencer to an average depth of 27 million reads per sample. The 100 nt long reads were mapped with STAR v.2.7.4a to the Ensembl human GRCh38 genome reference. The gene features were counted with HTSeq v.0.9.1 and the differential expression analysis was performed with the R/Bioconductor EdgeR v.1.34.1 package with a GLM (generalized linear model) test. The GSEA was performed with the R/Bioconductor fgsea v.1.30.0 package.
Laser microdissection (LMD), protein extraction and quantitative mass spectrometry analysis
Ten samples (mesenchyme of 5 CPAM1 and 5 CPAM2) were analyzed. Five healthy non-cystic adjacent lungs from either CPAM1 and CPAM2 were used as control. The formalin fixed paraffin embedded (FFPE) blocks of human pulmonary tissue specimens were cut into 10 µm thick sections and mounted on Leica PET-membrane slides (76463- 320, Leica Microsystems). Samples were deparaffinized and hydrated according to standard histological procedures, stained with hematoxylin (S330130-2, Agilent), and stored at 4 °C. Laser microdissection was performed in hematoxylin-stained samples according to the manufacturer’s guidelines. Briefly, an upright Leica DM6500 microscope, equipped with a Cryslas laser and a Leica CC7000 camera was used to perform the tissue LMD, where specific tissue fragments were isolated with a UVI 5 ×/0.12 microdissection objective, and regions of interest were defined manually with the LAS-AF software (Leica). The regions of interest were defined as the mesenchyme present under the epithelium of the bronchi in healthy or fetal tissue and under the cystic epithelium in CPAM1 and CPAM2. The isolated fragments were collected in 0.5 ml tubes. Proteins were reduced and prepared as previously published [20]. The resulting peptides were sequenced by liquid chromatography coupled with tandem mass spectrometry (LC–MS/MS) at the Proteomic core facility of the Faculty of Medicine of the University of Geneva. The characteristic peptide fragmentation spectra were then blasted into the SWISSPROT protein sequence database. Database search was performed with Mascot Server (Matrix Science Ltd.) and results were analyzed and validated using Scaffold software (Proteome software Inc.). For each sample, the number of peptides assigned to each protein was normalized to the total number of peptides obtained in the same sample. Relevant differentially relative expressed proteins in samples comparisons were found with a log2 > 0.58 and a FDR < 5% and the resulting lists handpicked cleaned to discard common unwanted artifacts (blood and serum related proteins, common proteomic contaminants, etc.). Subsequently unsupervised hierarchical clustering was computed using the R language and environment (v 3.5.3) (https://www.r-project.org), and the “pheatmap” package (v1.0.12) as well as bioconductor “EnhancedVolcano” for graphical plotting.
Protein extraction and western blot analysis
Proteins were extracted from frozen samples as previously described [21]. Protein concentration was determined according using a Pierce BCA Protein Assay (Thermo Scientific, REF 23227) and were adjusted to 1ug/ul concentration in standard SB Buffer before being heat denatured at 95° for 5 min. 20 ul of samples were then loaded and separated by electrophoreses in a NuPAGE 4–12% Bis–Tris gradient minigels (Thermo-Scientific, REF NP0335) and transferred to a nitrocellulose membrane (Protran® 0.2 µm; Schleicher and Schuell), which was validated by Ponceau Red staining. Different antibodies were then incubated according to manufacturer’s guidelines: anti-β-catenin1 (8480, Cell Signaling), anti-AKT and anti-phospho-AKT (283852, Abcam) antibodies. anti-GAPDH was used as housekeeping protein (MAB 374, Millipore). Horseradish peroxidase-conjugated goat anti-mouse IgG (1706516, Biorad,) or IgM (1020–05, Southern Biotech) or goat anti-rabbit IgG (111–035-144, Jackson Immunoresearch) were used as secondary antibodies. Enhanced chemiluminescence was used for detection (K-12043-D10, WesternBright Sirius, Advansta). Signals were digitized by A Li-Cor Odyssey XF imaging system (Licorbio) and data analysed with proprietary software Empiria Studio and ImageJ.
Multiplex immunohistochemistry and image acquisition
Three μm slides were cut from FFPE tissue blocks for immunohistochemistry (IHC). CPAM1, CPAM2 and control lung samples were assayed on two separate experiments. Multiplex tissue staining and data acquisition were performed as previously published [20]. Briefly, stainings were performed according to the protein of interest revealed with 3-amino-9-carbazole (AEC), counterstained with hematoxylin and coverslipped with Aquatex mounting solution (1.08562.0050, Merck) or PBS before whole-slide image scanning at 20X magnification using a Zeiss AxioScan.Z1. After image acquisition, slides were de-coverslipped by immersion in demineralized H2O (dH2O). AEC staining was then removed by sequential washing in 50% (1 min)/75% (3 min)/95% (8 min, shaking every 2 min)/75% (2 min)/50% (1 min) ethanol solutions, followed by rehydration in dH2O. Antigen and antibody complex stripping was achieved via HIER (10 min at high pressure). Mouse (1:25, 115–007-003; Jackson ImmunoResearch) and rabbit (1:25, 111–007-003, Jackson ImmunoResearch) fragment antigen-binding solutions targeted against primary antibodies host species were used to prevent undesired binding between not fully stripped antigen–antibody complex from the previous stain and the secondary biotinylated antibodies of the subsequent staining. At this point, slides re-entered the same process of staining as described above, starting with the Serum-Free Protein Block. All antibodies and streptavidin/horseradish peroxidase were diluted in DAKO Antibody Diluent (S0809, Agilent). Data was analyzed and quantified using QuPath0.5.1.
Statistical analysis
Data are presented as median or mean ± standard error of mean (SEM) for each sample or patients depending of the data analyzed. Statistical analysis was performed using GraphPad Prism software (GraphPad Software). One-way ANOVA was used to compare groups. The results were considered significant if p < 0.05.
Results
Patient characteristics
Ten operated patients were enrolled, 5 CPAM1 and 5 CPAM2 (Table 1). Most of the patients were born at full term (median age: 38 weeks of gestation) and had a birth weight adapted to the gestational age (median: 3200 g, range: 2280–4000 g). Mechanical ventilation was required for one patient with CPAM1, and supplemental oxygen was needed for one patient with CPAM2. CPAM was associated with other malformations in one patient (pectus excavatum appeared at the age of 10), although none of the patients had a clear family history of malformations or lung diseases. Lobectomy was performed preventively in asymptomatic patients between 9 and 18 months old. A significant number of viral pulmonary infections were also observed in these patients, both before and after surgery. A follow-up of pulmonary functions was performed when patients were 8 years old and results reported in supplementary Table 2.
Table 1.
Demographic data of CPAM patients expressed in percent, or median
| CPAM1 | CPAM2 | ||
|---|---|---|---|
| n = 5 | n = 5 | ||
| Male sex (%) | 40 | 60 | |
| Prenatal diagnosis | Gestational age (wk) | 20 | 21 |
| Perinatal characteristic | Gestational age at birth (wk) | 38 | 39 |
| Birth weight (gr) | 2990 | 3290 | |
| Respiratory mechanical support (%) | 20 | 0 | |
| Days of support (n) | 7 | 0 | |
| Oxygen administration (%) | 0 | 20 | |
| Days of support (n) | 0 | 2 | |
| Familial medical history | Congenital malformations (%) | 20 | 0 |
| Pulmonary disease (%) | 0 | 0 | |
| Personal respiratory medical history | Pulmonary viral infections before surgery (%) | 40 | 20 |
| Pulmonary viral infections after surgery (%) | 40 | 60 | |
| Surgery | Age (months) | 11 | 12 |
Dysregulated epithelial-mesenchymal pathways and decreased immune activation in CPAM1
Bulk RNA-sequencing of CPAM1 and CPAM2 samples versus adjacent healthy tissues was performed to decipher the differences in gene regulation and activated pathways. RNA sequencing of CPAM1 and healthy adjacent tissue revealed significant differences in transcript profiles between the two tissues (Fig. 1). A total of 19021 genes were detected of which 984 were differentially expressed with a fold threshold > 2, false discovery rate (FDR) < 5% in CPAM1 and analyzed. One-hundred and fifty-one genes were downregulated, and 833 genes were upregulated compared to healthy adjacent tissue (Fig. 1A). In CPAM2, out of 18566 detected genes, only 1 gene was downregulated (CXCL2) and 22 upregulated (Fig. 1B). Most of the upregulated genes in CPAM2 (Fig. 1B) are related to cell migration, tissue remodelling (MMP10) and epithelial integrity (MUC13, MUC16). Due to the low number of genes modified in CPAM2 compared to healthy tissue and for more relevance, gene set enrichment analysis (GSEA, Fig. 2) and gene ontology (GO, Fig. 3) analysis were only performed on CPAM1. This trend of higher number of up- or downregulated genes in CPAM1 versus CPAM2 is consistent with previously published transcriptomic data [11, 22].
Fig. 1.
Volcano plots of (A) CPAM1 (N = 2) and (B) CPAM2 (N= 2) versus healthy adjacent tissues (N = 4) showing differentially expressed genes. Only the 50 most differentially expressed genes are reported on the figure
Fig. 2.
Gene set enrichment analysis (GSEA) pathways performed on CPAM1 healthy adjacent tissues. Classification was performed depending on the p value
Fig. 3.
Dot plot of the 50 most differentially expressed genes identified through Gene Ontology (GO) enrichment analysis of (A) biological pathways, (B) cellular component, and (C) molecular function in CPAM1, compared to healthy adjacent tissues
The upregulation of cytokine-cytokine receptor and genes involved in immune responses, such as, IL5RA, IL 13RA, CXCL13, and CXCL6, suggested that CPAM1 development could be driven by inflammatory signaling pathways, leading to chronic inflammation regardless of number of patient’s infections (Table 1, supplementary Table 3). Nevertheless, GSEA analysis in CPAM1 showed a downregulation of immune-related functions and inflammatory responses, underlying the disruption of inflammatory signaling pathways in CPAM1 while the implication of this set of genes was more prominent in pathways related to EMT which was highly upregulated (Fig. 2, supplementary Table 3 and 5). Additionally, the increase in matrix metalloproteases (MMP1, MMP7) responsible of tissue remodelling and cellular migration might further correlate with EMT upregulation [23, 24].
Several transcription factors known to suppress malignant transformation including EGFL7 (Fig. 1) and TBX3 (supplementary Table 3) were downregulated in CPAM1 compared to healthy tissue [25, 26]. Additionally, the upregulation of KRAS signaling, a well-known oncogene (Fig. 2), might promote abnormal proliferation and inhibit terminal differentiation of lung tissue [6, 11]. This could suggests either a possible progression of CPAM toward malignancy or abnormal lung branching and alveolar differentiation as recently proposed by others [14, 27–29]. Additionally, significant decreased expression of BMP family members (BMP6, BMPR2, BMP2, BMPR1, BMP8A), SMAD regulators (SMAD6, SMAD7) and TGFβR3 was observed and corroborated a disruption of TGF-β signaling (Fig. 2, supplementary Table 3 and 4), a key molecule in normal or abnormal lung development and in extracellular matrix remodeling and muscle regeneration.
TBX2, a transcription factor also critical for lung development, was significantly decreased in CPAM1 compared to healthy adjacent tissues while stem cell-related markers such as SOX2, CDH2, CYP2J2, FBXO15, KIF19, KRT17, and PROM1 were highly expressed in CPAM1, aligning with increased EMT and recent findings by Zhang et al. [30] (supplementary Table 3 and 4). CPAM1 also presented an increased level of markers of basal and club cells (TP63 and SCGB1A +), suggesting that a dysregulation in cellular composition might also be implicated in CPAM development.
GO enrichment analysis (Fig. 3) performed on the set of differentially expressed genes between CPAM1 and adjacent healthy tissue revealed significant enrichment in several biological processes (BP) associated with the negative regulation of immune function, negative regulation of cell growth and increased tissue remodeling. In the molecular function (MF) category, genes were significantly enriched for cell motility (microtubule motor activity) and extracellular matrix remodeling. Cellular component (CC) analysis revealed enrichment in epithelia related pathway (Fig. 3).
Taken together, the transcriptomic results support the dysregulation of pathways related to cell migration, growth regulation, differentiation and EMT in CPAM1 compared to healthy tissue.
Desmin downregulation, a sign of abnormal smooth muscle in CPAM
To systematically characterize the protein expression within the mesenchyme surrounding CPAM malformations, quantitative proteomics was conducted after laser microdissection of the mesenchymal parts present under the cystic tissue and healthy adjacent airway. A side-by-side analysis of protein profiles from microdissected mesenchyme of 4 patients in each group (CPAM1, CPAM2, healthy and fetal lung tissue) was performed. A total of 1066 proteins were differentially expressed with a Log2 > 0.58 and FDR < 5%. CPAM1 exhibited a large cluster of downregulated proteins compared to fetal samples, while partially similar differentially expressed protein profiles were shared between the CPAM and healthy lung tissue. Of note, these proteins were upregulated when comparing healthy CPAM adjacent tissue with fetal tissue indicating that the downregulation of these proteins in CPAM is due to abnormal development rather than a differentiation process from fetus to newborn (supplementary Fig. 1). Various proteins involved in muscle differentiation were identified as downregulated in both types of CPAM as already suggested by our group (Fig. 4) [18]. Only a few inflammation proteins were significantly modulated, such as S100A9 upregulated in CPAM1 or cathepsin G downregulated in CPAM2 (supplementary Table 4). We found that several proteins expressed in the mesenchyme were modified in CPAM tissue, some of which are discussed below.
Fig. 4.
Volcano plots depicting the proteomic data derived from laser microdissected mesenchymal tissues of CPAM1 (N = 5), CPAM2 (N = 5), healthy adjacent (N = 5) and fetal (N = 5) lungs in two distinct comparisons: (A) between CPAM and healthy adjacent samples and (B) CPAM and fetal samples. Colored arrows indicate the selected proteins of interest for the immunohistochemistery experiments: Desmin (blue arrows), ACTA2 (red arrows), MYH11 (green arrows)
α-SMA, an important muscle component and typical marker of SMCs, was decreased in both CPAM subtypes compared to healthy adjacent tissue and with a stable expression in fetal tissue compared to healthy samples (Fig. 4). Desmin, a cytoskeletal protein essential for smooth muscle integrity, upregulated during lung development to support airway formation. It was significantly decreased in both CPAM subtypes compared to healthy adjacent samples (Fig. 4A) and in CPAM1 compared to fetal lung tissue (Figs. 4B and 5B) Conversely its level was also decreased in fetal samples relative to the pediatric healthy lung. In addition, more pronounced decrease in desmin expression in CPAM1 compared to CPAM2 indicates a difference between both subtypes of CPAM as already suggested by our group [18, 31].
Fig. 5.
Representative multiplex immunohistochemistry of (A) healthy adjacent airway, (B) CPAM1 cyst and (C) fetal airway. Hematoxylin: dark blue, α-SMA: red, desmin: green. (colocalisation signal of green and red color appears as light green to yellow) Coloration are highlighted by the white arrows. Scale bar: 200 μm (D) Quantification of desmin/a-SMA in percentage of co-expression. N = 10 for CPAM1 cyst and healthy adjacent bronchi. N = 4 for fetal bronchi in a total of 4 technical replicates. ** P < 0.01
Of note, both CPAM1 and CPAM2 had similar expression patterns in regards to α-SMA and desmin when compared to fetal tissue (Fig. 5B) with most of the protein expression differences highlighting protein production and extracellular matrix modulation.
Laminin, a crucial component of the basal membrane, was decreased in both CPAM subtypes, with a more significant reduction in CPAM1. MYH11, another cytoskeletal protein involved in muscle contraction, was also reduced in both CPAM subtypes, indicating potentially compromised muscle contractility [31]. However, laminin and MYH11 were not differentially expressed in the CPAM/fetal tissue comparison, suggesting that the muscle tissue in CPAM may be less mature than the one in healthy adjacent tissue.
Abnormal mesenchymal and muscle development in CPAM1
To validate our quantitative proteomic results, we conducted multiplex immunostaining for various smooth muscle markers in CPAM1 cystic borders, healthy adjacent and fetal airways. The samples were stained for desmin and α-SMA (ACTA2). We observed a significant reduction of desmin-positive area, calculated as a % of α-SMA-positive area, in CPAM1 compared to both fetal or healthy adjacent airways (Fig. 5). Staining for MYH11 and calponin1, two muscle proteins involved in muscle contraction, were colocalized around cystic, fetal, and adjacent airways with no significant differences among them, suggesting a partially functional muscle (Fig. 6).
Fig. 6.
Representative multiplex immunohistochemistry of (A) healthy adjacent airway, (B) CPAM1 cyst and (C) fetal airway. Hematoxylin: dark blue, α-SMA: red, myosin 11: magenta, and calponin1: green (colocalised signal of red, magenta and green colors appears as pink). Colorations are highlighted by the white arrow. Scale bar: 100 μm. Quantification of (D) calponin1/α-SMA and (E) myosin11/α-SMA co-expression. N = 5 for CPAM1 and healthy adjacent bronchi, N = 3 for fetal bronchi in a total of 3 technical replicates
We also analyzed vimentin expression, an intermediate filament protein involved in fetal mesenchymal differentiation of airways or vessels, which decreases during airways maturation [31]. Vimentin and α-SMA did not colocalize, indicating only partially differentiated muscle tissue (Fig. 7). Vimentin was present in fetal tissue, its expression decreasing between the pseudoglandular and saccular stages (data not shown), as expected.
Fig. 7.
Representative multiplex immunohistochemistry of (A) healthy adjacent airway, (B) CPAM1 cyst and (C) fetal airway. Hematoxylin: dark blue, α-SMA: red, vimentin: green (colocalised signal of green and red colors appear as light green to yellow). Colocalisations are highlighted by the white arrow head. Scale bar: 500 and 100 μm. N = 3 for CPAM1 and healthy adjacent bronchi, N = 3 for fetal bronchi in a total of 2 technical replicates
Altered β-catenin and PI3K-AKT-mTOR pathways expression in CPAM1
The TBX2 and TBX3 decrease could affect regulation of the WNT/β-catenin pathway implicated in lung development particularly during the pseudoglandular stage, where it influences branching morphogenesis and epithelial proliferation and differentiation [32–34]. The WNT-β-catenin pathway also regulates immune cell infiltration in the tumor microenvironment [35]. This decrease was suggested by Western blotting, which showed a tendency to reduced β-catenin expression in the CPAM1 compared to healthy adjacent tissue (Fig. 8A and B). However, to properly address this question, further analysis of nuclear β-catenin should be performed.
Fig. 8.
Western blot showing the expression and quantification of (A, B) β-catenin, (C, D, E) AKT and phosphorylated AKT (AKTPhos) in CPAM1 and healthy adjacent samples. Protein extract (20 μg) from 3 biological samples of each condition were analyzed. GAPDH was used a housekeeping protein. Quantification was performed after normalization on GAPDH for (B) β-catenin and (E) AKT/AKTphos. N = 3 for CPAM1 and for healthy adjacent tissue
The PI3K-AKT-mTOR pathway, which regulates cell survival, growth, cell cycle, and metabolism, is also critical in lung development and branching morphogenesis. Somatic mutations in this pathway have been linked to structural birth defects The downregulation of the PI3K-AKT-mTOR pathway was already suggested as taking part on the CPAM development [11, 36]. Our results showed low AKT phosphorylation aligning with findings from our transcriptomic analysis with cell cycle gene modification such as TBX2 or TBX3 (Figs. 1 and 8C, D and E) [27].
Discussion
CPAM is a congenital disorder characterized by localized abnormal lung development. Despite significant research, knowledge concerning the genesis of these malformations and their evolution into adulthood is limited. Most studies have focused on CPAM lesions in their entirety. In our study, we specifically analyzed the role of the mesenchymal tissue adjacent to CPAM using laser microdissection and advanced molecular techniques. To our knowledge, this is the first study employing diverse sample types and methodologies of the same patients to elucidate the role of mesenchyme in CPAM development. Our study highlights significant molecular and structural alterations related to immune responses, lung development, epithelial-mesenchymal interactions as well as muscle contraction particularly in the adjacent mesenchymal tissue of CPAM1 patients, underscoring the complexity of CPAM pathogenesis. Further study of adjacent tissues obtained after surgical resection of CPAM could have long-term clinical significance for these patients. Indeed, some patients may benefit from clinical and lung function monitoring into adulthood as well as from tailored preventive strategies aimed at preserving their respiratory health and potentially avoiding the development of oncologic complications.
The use of multiple techniques such as multiplex immunohistochemistry on paraffin-embedded sections, laser microdissection, and quantitative proteomic and transcriptomic analysis on the same patient samples is an innovative and an advantage for understanding the different levels of regulation that can influence the occurrence of CPAM at both individual and collective levels.
Over the past five years, several studies including small cohorts of patients have been published focusing on the transcriptome in order to find a biomarker or genetic profile explaining the abnormal lung development observed in CPAM [10, 11, 14, 22, 27, 30]. They also analyzed transcriptomic modifications by studying miRNA, IncRNA or methylation [11, 14, 22]. The altered inflammatory profile observed in CPAM compared to healthy subjects appears in all studies, as confirmed by us. Only one study used different techniques to analyze CPAM, such as transcriptomics and immunohistochemistry [36]. In this study, they highlighted the disruption of Ras and PI3K-AKT-mTOR signalling, as demonstrated in our study. The PI3K-AKT-mTOR pathway is essential for regulating cell survival, growth, cycle progression, and metabolism. It also plays a role in lung branching morphogenesis, and somatic mutations in this pathway have been linked to congenital structural lung defects. Previous studies have reported downregulation of this pathway in CPAM patients, and our findings confirmed a decrease in AKT phosphorylation, consistent with our transcriptomic data [11, 36]. Additionally, transcription factors known to prevent malignant transformation were decreased, highlighting potential targets for monitoring CPAM progression toward malignancy, as suggested by Tan et al. and Patrizi et al. [11, 21]. Moreover, the RAS pathway is upregulated in our results, as previously proposed by Swarr et al. [36]. These highlighting Ras regulations are in line with previous publications describing KRAS mutation in CPAM lung epithelial tissue with a potential role in tumor progression later in life [29].
The epithelial-mesenchymal interactions were carefully studied from the gene to the protein level to better analyze their role during either normal or abnormal lung development. Indeed, these interactions between lung airways and the surrounding mesenchymal compartment play a role in controlling airway growth. Embryonic lung mesenchymal progenitors serve as precursors for pulmonary airway SMCs, which are involved in lung branching by directing epithelial bud bifurcation. Dysregulation in the TGF-β signalling, a crucial pathway for extracellular matrix production and epithelial-mesenchymal interactions during lung branching, was previously described by Lezmi et al. [10]. Our findings confirmed a decreased expression of this pathway further suggesting abnormal lung development in CPAM1. SPOCK2, a crucial protein involved in maintaining epithelial-mesenchymal interactions during alveolarization, was also decreased in CPAM1. This reduction could affect MMP1 and MMP7 overexpression as it has been described to affect MMP-9 activation and impair the alveolarization process [27].
During fetal lung development, cytoskeletal proteins play a crucial role in regulating cell shape, migration, and tissue architecture, thereby contributing to airway branching and alveolar formation. Muscle contraction, primarily mediated by SMCs surrounding the airways, is also essential for proper lung morphogenesis, as it regulates airway peristalsis, influencing lung fluid dynamics and tissue remodeling. Disruptions in these processes may contribute to improper airway differentiation, potentially leading to the development of CPAM. A significant decrease in desmin expression was observed in CPAM tissue compared to adjacent healthy tissue. In contrast, desmin expression was higher in fetal tissue corresponding to the canalicular stage of lung development, supporting the hypothesis that CPAM1 may develop before this stage, as previously proposed by our team [18]. Vimentin, which normally decreases in favor of desmin as the airway matures, presented only a slight decrease in CPAM1 lesions [31]. Moreover, vimentin did not co-localize with α-SMA staining, suggesting the presence of only partially differentiated muscle tissue. In addition, laminin and MYH11 were decreased in CPAM tissue compared to adjacent healthy tissue. Therefore, CPAM lesions show a less differentiated muscle phenotype around cystic tissue and an impaired muscle contractility, potentially contributing to the development of these lesions. A correlation between loose mesenchyme or abnormal EMT has been suggested to predispose to chronic obstructive pulmonary disease later in life in a murine model opening a possible similarity between abnormal CPAM mesenchyme and partially impaired lung evolution during adulthood [37]. β-Catenin plays a crucial role in lung development, particularly during the pseudoglandular stage, where it regulates branching morphogenesis, epithelial proliferation, and differentiation [38]. Additionally, the WNT-β-catenin pathway is implicated in immune cell infiltration within the tumor microenvironment [35]. Our experiments demonstrated a decreasing trend in β-catenin expression in CPAM1 compared to healthy adjacent airway epithelium, reinforcing the hypothesis of developmental abnormalities in CPAM1. These results are aligned with preliminary publications in transcriptomic analysis performed on CPAM [10, 36].
The relationship between patients’ clinical evolution and lung tissue analysis remains unclear as most patients, whether operated or not, have a good clinical outcome. Although spirometric results remained within normal ranges for both CPAM subtypes, FEV1 and FVC were reduced in CPAM1 compared to CPAM2, as previously reported by Mandaliya et al. [39]. The difference in FEV1 values likely reflects involvement of medium to large airways in CPAM1, whereas CPAM2 primarily affects more distal lung structures [30, 39]. Our previous study suggested a greater resemblance of CPAM1 to airway, while CPAM2 exhibits similarities to alveolar tissue [18]. TLC was approximately 100% for both subtypes, suggesting that compensatory lung growth facilitated full lung volume recovery post-surgery, ruling out restrictive lung disease because of surgical intervention. Finally, the potential presence of stem cells in CPAM and, to a lesser extent, in adjacent healthy tissue, suggests a capacity for tissue regeneration post-surgery [12]. Diffusion capacity remained within the normal range for both CPAM1 and CPAM2, further suggesting the hypothesis that early-life surgical intervention may enable functional lung recovery. Even in a normal range level the diffusion capacity of CPAM1 was significantly lower than CPAM2 (p = 0.014) possibly in link with the different size of the lesion correlating with the literature [40].
As congenital pulmonary airway malformation is a rare disease, this implies a small number of potential patients. A limitation of our study was the difficulty in recruiting eligible patients with appropriate and complete follow-up, as has already been observed in other published studies of this pathology. Although this resulted in a relatively small proportion of all patients followed at our center, it is one of the most comprehensive study designs of a CPAM cohort using all methods in the same patients’ samples and their healthy adjacent part. The low amount of OCT-conserved CPAM samples was a also limitation to our Western blot statistical analysis.
Conclusion
Highlighting the possible role of the mesenchyme in the pulmonary evolution of these patients introduces a new aspect to consider during their follow-up. This perspective is novel and has been scarcely studied until now. Our study revealed the presence of abnormal mesenchyme in particular muscle around cystic lesions. This opens an important new focus of research that requires further investigation. Currently, patients are rarely followed by a pneumologist after the pediatric period. This could change the course of action, encouraging continued follow-up of patients throughout their lives to monitor both tumor risk and the development of chronic obstructive pulmonary disease. This translational approach is innovative and crucial for gaining a better understanding of the potential genesis of CPAM, risk factors for infectious events, optimal timing for surgery, and the risk of tumor transformation possibly due to additional molecular anomalies later in life. To our knowledge, the ability to compare all these different methods, in both fetal and cystic samples, is unique in the field and offers new perspectives for patient follow-up.
Supplementary Information
Supplementary Material 1: Figure S1. Heatmap of quantitative proteomic data from CPAM1, CPAM2, healthy and fetal lung samples comparisons. The hierarchical clustering is based on differential expression levels. Data is expressed in log2 minus log2 average.
Supplementary Material 2: Table S1. Summary of the different techniques used and sample type analyzed.
Supplementary Material 3: Table S2. Lung functions of 8 years old patients after CPAM removal.
Supplementary Material 4: Table S3. Raw transcriptomic data.
Supplementary Material 5: Table S4. Raw proteomic data.
Supplementary Material 6: Table S5. GSEA obtained from transcriptomic data.
Acknowledgements
This work was supported by the Ligue Pulmonaire Genevoise, the Fondation Privée de HUG and the Fondation Prim’Enfance. We thank the Genomic, Proteomics and Bioimaging core facilities of the CMU, University of Geneva, for their equipment, advice and help. In particular, we thank Nicolas Liaudet from the Bioimaging core facility for his help in the development of the proper software for image processing. We thank Domitille Schvartz for her help in analysing the bioinformatics data. We thank Professor Constannce Barazzone-Argiriffo for her help and advices during this project.
Authors’ contributions
PA, YA, MLBP and IRM, contributed to conception and design of the study and the writing manuscript. PA, AH and YA performed the practical work. A-LR and CD provided the human specimens and participate to the design of the study. All authors contributed to manuscript revision, read and approved the submitted version.
Funding
Open access funding provided by University of Geneva. This work was supported by the Ligue Pulmonaire Genevoise, the Fondation Privée de HUG and the Fondation Prim’Enfance.
Data availability
Data is provided within the manuscript or supplementary files.
Declarations
Ethics approval and consent to participate
The institutional ethics committee approved this study and informed consent was obtained during scheduled hospital visits (Ethics committee of the Geneva University Hospital CER 12–110 and PB_2016-00175).
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Langston C. New concepts in the pathology of congenital lung malformations. Semin Pediatr Surg. 2003;12(1):17–37. [DOI] [PubMed] [Google Scholar]
- 2.Stocker JT, Madewell JE, Drake RM. Congenital cystic adenomatoid malformation of the lung. Hum Pathol. 1977;8(2):155–71. [DOI] [PubMed] [Google Scholar]
- 3.Cook J, Chitty LS, De Coppi P, Ashworth M, Wallis C. The natural history of prenatally diagnosed congenital cystic lung lesions: long-term follow-up of 119 cases. Arch Dis Child. 2017;102(9):798–803. [DOI] [PubMed] [Google Scholar]
- 4.Wang A, D’Amico TA, Berry MF. Surgical management of congenital pulmonary malformations after the first decade of life. Ann Thorac Surg. 2014;97(6):1933–8. [DOI] [PubMed] [Google Scholar]
- 5.Wong KKY, Flake AW, Tibboel D, Rottier RJ, Tam PKH. Congenital pulmonary airway malformation: advances and controversies. Lancet Child Adolesc Health. 2018;2(4):290–7. [DOI] [PubMed] [Google Scholar]
- 6.Hermelijn SM, Wolf JL, Dorine den Toom T, Wijnen RMH, Rottier RJ, Schnater JM, et al. Early KRAS oncogenic driver mutations in nonmucinous tissue of congenital pulmonary airway malformations as an indicator of potential malignant behavior. Hum Pathol. 2020;103:95–106. [DOI] [PubMed] [Google Scholar]
- 7.Fakler F, Aykutlu U, Brcic L, Eidenhammer S, Thueringer A, Kashofer K, et al. Atypical goblet cell hyperplasia occurs in CPAM 1, 2, and 3, and is a probable precursor lesion for childhood adenocarcinoma. Virchows Arch. 2020;476(6):843–54. [DOI] [PubMed] [Google Scholar]
- 8.Pogoriler J, Vargas SO. Cystic masses of the pediatric lung: update on congenital pulmonary airway malformation and its differential diagnosis. Virchows Arch. 2025;486(1):177–88. [DOI] [PubMed] [Google Scholar]
- 9.Zepp JA, Morrisey EE. Cellular crosstalk in the development and regeneration of the respiratory system. Nat Rev Mol Cell Biol. 2019;20(9):551–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lezmi G, Vibhushan S, Bevilaqua C, Crapart N, Cagnard N, Khen-Dunlop N, et al. Congenital cystic adenomatoid malformations of the lung: an epithelial transcriptomic approach. Respir Res. 2020;21(1):43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zeng J, Liu W, Liang J, Peng J, Wang F, Tang J, et al. Analysis of miRNA profiles and the regulatory network in congenital pulmonary airway malformations. Front Pediatr. 2021;18(9):671107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Zhang S, Ye C, Xiao J, Yang J, Zhu C, Xiao Y, et al. Single-cell transcriptome profiling reveals the mechanism of abnormal proliferation of epithelial cells in congenital cystic adenomatoid malformation. Exp Cell Res. 2020;396(2):112299. [DOI] [PubMed] [Google Scholar]
- 13.Ligresti G, Raslan AA, Hong J, Caporarello N, Confalonieri M, Huang SK. Mesenchymal cells in the Lung: evolving concepts and their role in fibrosis. Gene. 2023;859:147142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Patrizi S, Pederiva F, d’Adamo AP. Whole-genome methylation study of congenital lung malformations in children. Front Oncol. 2021;28(11):689833. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Danopoulos S, Alonso I, Thornton ME, Grubbs BH, Bellusci S, Warburton D, et al. Human lung branching morphogenesis is orchestrated by the spatiotemporal distribution of ACTA2, SOX2, and SOX9. Am J Physiol-Lung Cell Mol Physiol. 2018;314(1):L144–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Zhang M, Shi J, Huang Y, Lai L. Expression of canonical WNT/β-CATENIN signaling components in the developing human lung. BMC Dev Biol. 2012;12(1):21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Jiang Y, Luo Y, Tang Y, Moats R, Warburton D, Zhou S, et al. Alteration of cystic airway mesenchyme in congenital pulmonary airway malformation. Sci Rep. 2019;9(1):5296. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Barazzone-Argiroffo C, Lascano Maillard J, Vidal I, Bochaton-Piallat ML, Blaskovic S, Donati Y, et al. New insights on congenital pulmonary airways malformations revealed by proteomic analyses. Orphanet J Rare Dis. 2019;14(1):272. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Zhu H, Liu D, Jia H. Analysis of Wnt7B and BMP4 expression patterns in congenital pulmonary airway malformation. Pediatr Pulmonol. 2020;55(3):765–70. [DOI] [PubMed] [Google Scholar]
- 20.Cardoso Dos Santos LM, Avila Y, Schvartz D, Rougemont AL, Bochaton-Piallat ML, Ruchonnet-Metrailler I. Laser microdissection, proteomics, and multiplex immunohistochemistry: a bumpy ride into the study of paraffin-embedded fetal and pediatric lung tissues. Front Med. 2023;10:1191205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Sakic A, Chaabane C, Ambartsumian N, Klingelhöfer J, Lemeille S, Kwak BR, et al. Neutralization of S100A4 induces stabilization of atherosclerotic plaques: role of smooth muscle cells. Cardiovasc Res. 2022;118(1):141–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Yang W, Zhao P, Liu Y, Cao P, Ji X, Gao Y, et al. Transcriptome analysis of lncRNA expression patterns in human congenital lung malformations. BMC Genomics. 2021;22(1):861. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Pardo A, Selman M. Role of matrix metaloproteases in idiopathic pulmonary fibrosis. Fibrogenesis Tissue Repair. 2012;5(S1):S9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Agraval H, Kandhari K, Yadav UCS. MMPs as potential molecular targets in epithelial-to-mesenchymal transition driven COPD progression. Life Sci. 2024;352:122874. [DOI] [PubMed] [Google Scholar]
- 25.Liu W, Zhang Y, Huang F, Ma Q, Li C, Liu S, et al. The polymorphism and expression of EGFL7 and miR-126 are associated with NSCLC susceptibility. Front Oncol. 2022;14(12):772405. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Khan SF, Damerell V, Omar R, Du Toit M, Khan M, Maranyane HM, et al. The roles and regulation of TBX3 in development and disease. Gene. 2020;726:144223. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Tan Z, Li F, Chen Q, Chen H, Xue Z, Zhang J, et al. Integrated bulk and single-cell RNA-sequencing reveals SPOCK2 as a novel biomarker gene in the development of congenital pulmonary airway malformation. Respir Res. 2023;24(1):127. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Garinet S, Rahshenas M, Galmiche-Rolland L, Abbo O, Bonnard A, Hameury F, et al. Cancer-associated mutations in congenital pulmonary malformations: a prospective cohort. Am J Respir Crit Care Med. 2023;207(5):615–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Nelson ND, Xu F, Chandrasekaran P, Litzky LA, Peranteau WH, Frank DB, et al. Defining the spatial landscape of KRAS mutated congenital pulmonary airway malformations: a distinct entity with a spectrum of histopathologic features. Mod Pathol. 2022;35(12):1870–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Zhang G, Cai C, Li X, Lou L, Zhou B, Zeng H, et al. Application of second-generation sequencing in congenital pulmonary airway malformations. Sci Rep. 2022;12(1):20459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Mitchell JJ, Reynolds SE, Leslie KO, Low RB, Woodcock-Mitchell J. Smooth muscle cell markers in developing rat lung. Am J Respir Cell Mol Biol. 1990;3(6):515–23. [DOI] [PubMed] [Google Scholar]
- 32.Lüdtke TH, Rudat C, Wojahn I, Weiss AC, Kleppa MJ, Kurz J, et al. Tbx2 and Tbx3 act downstream of Shh to maintain canonical Wnt signaling during branching morphogenesis of the murine lung. Dev Cell. 2016;39(2):239–53. [DOI] [PubMed] [Google Scholar]
- 33.Hussain M, Xu C, Lu M, Wu X, Tang L, Wu X. Wnt/β-catenin signaling links embryonic lung development and asthmatic airway remodeling. Biochimica et Biophysica Acta (BBA). 2017;1863(12):3226–42. [DOI] [PubMed] [Google Scholar]
- 34.Ustiyan V, Zhang Y, Perl AT, Whitsett JA, Kalin TV, Kalinichenko VV. Β-catenin and Kras/Foxm1 signaling pathway are critical to restrict Sox9 in basal cells during pulmonary branching morphogenesis. Dev Dyn. 2016;245(5):590–604. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Pai SG, Carneiro BA, Mota JM, Costa R, Leite CA, Barroso-Sousa R, et al. Wnt/beta-catenin pathway: modulating anticancer immune response. J Hematol Oncol. 2017;10(1):101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Swarr DT, Peranteau WH, Pogoriler J, Frank DB, Adzick NS, Hedrick HL, et al. Novel Molecular and Phenotypic Insights into Congenital Lung Malformations. Am J Respir Crit Care Med. 2018;197(10):1328–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Benjamin JT, Plosa EJ, Sucre JMS, Van Der Meer R, Dave S, Gutor S, et al. Neutrophilic inflammation during lung development disrupts elastin assembly and predisposes adult mice to COPD. J Clin Invest. 2021;131(1):e139481. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Caldeira I, Fernandes-Silva H, Machado-Costa D, Correia-Pinto J, Moura RS. Developmental pathways underlying lung development and congenital lung disorders. Cells. 2021;10(11):2987. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Mandaliya PH, Morten M, Kumar R, James A, Deshpande A, Murphy VE, et al. Ventilation inhomogeneities in children with congenital thoracic malformations. BMC Pulm Med. 2015;15(1):25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hijkoop A, Van Schoonhoven MM, Van Rosmalen J, Tibboel D, Van Der Cammen-van Zijp MHM, Pijnenburg MW, et al. Lung function, exercise tolerance, and physical growth of children with congenital lung malformations at 8 years of age. Pediatr Pulmonol. 2019;54(8):1326–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material 1: Figure S1. Heatmap of quantitative proteomic data from CPAM1, CPAM2, healthy and fetal lung samples comparisons. The hierarchical clustering is based on differential expression levels. Data is expressed in log2 minus log2 average.
Supplementary Material 2: Table S1. Summary of the different techniques used and sample type analyzed.
Supplementary Material 3: Table S2. Lung functions of 8 years old patients after CPAM removal.
Supplementary Material 4: Table S3. Raw transcriptomic data.
Supplementary Material 5: Table S4. Raw proteomic data.
Supplementary Material 6: Table S5. GSEA obtained from transcriptomic data.
Data Availability Statement
Data is provided within the manuscript or supplementary files.