Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Aug 5;13(1):4557.
doi: 10.1038/s41467-022-32052-2.

Club cells employ regeneration mechanisms during lung tumorigenesis

Yuanyuan Chen #  1 , Reka Toth #  1   2   3 , Sara Chocarro #  1   4 , Dieter Weichenhan  2 , Joschka Hey  2   4 , Pavlo Lutsik  2 , Stefan Sawall  5 , Georgios T Stathopoulos  6   7 , Christoph Plass  2   7   8   9 , Rocio Sotillo  10   11   12   13
Affiliations

Club cells employ regeneration mechanisms during lung tumorigenesis

Yuanyuan Chen et al. Nat Commun. .

Abstract

The high plasticity of lung epithelial cells, has for many years, confounded the correct identification of the cell-of-origin of lung adenocarcinoma (LUAD), one of the deadliest malignancies worldwide. Here, we employ lineage-tracing mouse models to investigate the cell of origin of Eml4-Alk LUAD, and show that Club and Alveolar type 2 (AT2) cells give rise to tumours. We focus on Club cell originated tumours and find that Club cells experience an epigenetic switch by which they lose their lineage fidelity and gain an AT2-like phenotype after oncogenic transformation. Single-cell transcriptomic analyses identified two trajectories of Club cell evolution which are similar to the ones used during lung regeneration, suggesting that lung epithelial cells leverage on their plasticity and intrinsic regeneration mechanisms to give rise to a tumour. Together, this study highlights the role of Club cells in LUAD initiation, identifies the mechanism of Club cell lineage infidelity, confirms the presence of these features in human tumours, and unveils key mechanisms conferring LUAD heterogeneity.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Mouse model of lung adenocarcinoma and identification of adenovirus-infected cell types.
a Schematic of the Eml4-Alk mouse LUAD model used. Mice were intratracheally instilled with Ad-EA and μCT was used to monitor tumour development. Representative H&E staining of an early lesion (hyperplasia) and a tumour are shown on the right. b Immunofluorescent staining of SPC (green) and CCSP (red) antibodies showing that Club cells in the bronchi start expressing SPC upon Eml4-Alk rearrangement. DAPI staining in blue. Original (left), magnified overlay (middle), and their single-channel images (right) from the dashed areas are shown. Arrowheads show cells that are double-positive CCSP+SPC+. Scale bar: 20 μm. A minimum of 6 independent animals were analysed. c Immunohistochemistry of CCSP and SPC antibodies from Eml4-Alk tumours showing CCSP+ bronchi as well as cells inside the tumours. Scale bars: 200 μm (left panel) and 100 μm (right panel). A minimum of 6 independent animals were analysed. d Experimental schematic of mT/mG mice transduced with Ad-Cre indicating the cell types that can get infected. e Immunofluorescent staining of the indicated antibodies on lung sections from mT/mG mice transduced with Ad-Cre, showing that Club, Ciliated, BASCs, AT2 and AT1 cells are infected. Scale bars 10 μm. A minimum of 6 independent animals were analysed.
Fig. 2
Fig. 2. Lineage-tracing models.
a Immunofluorescent staining of GFP and the indicated antibodies in the different lineage-tracing mouse models (Scgb1a1, Sftpc, Hopx and Foxj1) showing the specific labelling (CCSP, SPC, PDPN and acTUB). Images of single-channel and overlay are displayed sequentially. Scale bars: 10 μm. A minimum of 6 independent animals were analysed. b Immunofluorescent staining of GFP and the indicated antibodies (CCSP, SPC and acTUB) in the lung sections from lineage-tracing mouse models showing the labelling specificity. Arrows indicate the unspecific labelling of the cells. Scale bars: 10 μm. A minimum of 6 independent animals were analysed. c Schematic of the labelling and tumour induction of lineage-tracing mice. d Percentage of GFP+ tumours in the respective lineage tracing mice. Scgb1a1, 3 mice and 141 tumours analysed; Sftpc, 6 mice and 214 tumours analysed; Hopx, 3 mice and 117 tumours analysed and Foxj1 4 mice and 113 tumours analysed. One-way ANOVA, Tukey’s multiple comparison test. Data are presented as mean values +/− SD. e Immunofluorescent staining of GFP and RFP antibodies in the respective mouse models. The percentages on the upper right corners represent the number of green labelled tumours out of the total number of tumours analysed in each line. Scale bars: 500 μm. Insert in the Foxj1 example is provided to show the labelling of ciliated cells in the bronchi. A minimum of 6 independent animals were analysed.
Fig. 3
Fig. 3. Characterization of Krt5 and Hopx mice.
a Immunofluorescent staining of GFP and KRT5 antibodies in the trachea of Krt5 mice showing labelled Basal cells. Overlay and single-channel images are sequentially presented. Scale bars: 10 μm. A minimum of 6 independent animals were analysed. b Immunofluorescent staining of GFP and the indicated antibodies in the distal lung of Krt5 mice. AT1 cells, but not Club or AT2, are labelled. Blue (DAPI), green (GFP), red (as indicated). Scale bars: 10 μm. A minimum of 6 independent animals were analysed. c Overview of GFP immunofluorescent staining in Krt5 mice. The percentage on the lower left corner represents the percentage of GFP+ tumours, 0 out of 130 tumours analysed in a total of 4 animals. Scale bars: 1 mm. A minimum of 6 independent animals were analysed. d, e GFP, SPC, and CCSP immunofluorescent staining on Hopx mice transduced with Ad-EA. Both early (d) and late (e) stages of tumorigenesis are shown; the original, magnified and single-channel images are sequentially shown from left to right; arrowheads indicate the labelled Club cells under lineage switch into AT2 cells. Scale bars: 10 μm. A minimum of 6 independent animals were analysed.
Fig. 4
Fig. 4. DNA methylation patterns of normal and tumorigenic cells.
a Principal component analysis of DNA methylation at TSS-distal and TSS proximal poised enhancer regions (En-Pd and En-Pp, based on ENCODE postnatal 0 days mouse lung, ENCSR538YJF). The shape of the dots reflects the sample type, as normal lung (Normal) or GFP+ tumours (Tumour). The colour indicates mouse lineages used. Coloured shading is drawn around the Normal (light blue) and Tumour (purple) samples. b DNA methylation deconvolution by MeDeCom analysis shows three identified latent methylation components (LMC 1–3). Colours indicate the proportion of LMCs for each sample. Samples are clustered within each mouse lineage (Sftpc, Scgb1a1 or Hopx) and according to sample type (Normal/Tumour). c Association of LMCs with cell type-specific gene signatures based on published single-cell RNA sequencing data. Genes are ranked by their correlation of promoter methylation with the respective LMC. Enrichment scores are running sums calculated using the Gene Set Enrichment Analysis algorithm. d Heatmap of regions that were differentially methylated between Scgb1a1 and Sftpc normal samples. The methylation value is shown as beta values ranging from 0 to 1 visualized by the heatmap colours blue to red. For tumours, the label of each column shows the proposed originating cell type, while the colours below depict their originating lineage. e Average methylation across the Scgb1a1 and Sftpc genes for normal samples from Scgb1a1 and Sftpc lines and Club and AT2 originated tumours. Pink bars highlight regions with DNA methylation differences between normal lineages, while blue bars highlight regions with differences in tumours from distinct origins. ENCODE cCRE, showing candidate Cis-Regulatory elements, cHMM lung P0 showing ChromHMM regions in mouse lung, postnatal 0 days, and ATAC lung P0 showing ATAC-seq peaks in postnatal 0-day-old lung are UCSC tracks under the same name. f TF motif enrichment analysis of the differentially methylated regions (DMRs). Each column represents one comparison (e.g. Scgb1a1 normal vs. tumour) in one direction (e.g. hypermethylation in second, as in hypermethylated in tumours). The colour of the dots shows the enrichment of the DMRs for the motifs compared to random genomic regions. The size of the dots reflects on the –log10(p value) of the enrichment analysis. Empty lines show the lack of significant enrichment.
Fig. 5
Fig. 5. Single-cell RNA sequencing.
a Experimental strategy of scRNA sequencing analysis. b UMAP of the high-quality cells. Cells are coloured by the clusters identified using a shared nearest neighbour (SNN) modularity optimization-based clustering. c UMAP embedding showing the distribution of the cells collected from the different time points. Coloured dots represent the cells collected at the indicated time point; grey dots represent cells from the other time points. d Percentage of cells in each cluster, coloured by sample origin. The relative contributions were normalized to the number of cells in each sample. e UMAP coloured by the cell-type annotation of the study set. The annotation combines previously published cell-type-specific markers, and manual curation and is based on the most significant marker genes of each cluster or cell group (Supplementary Table 5). f UMAP with cells coloured by expression of Club (Scgb1a1, Scgb3a2), AT2 (Sftpc, Lyz2) and AT1 (Hopx, Ager) cell-type-specific markers. Colours represent the normalized expression levels for each marker in each cell. g Violin plot depicting the normalized expression of Alk by cluster indicating elevated expression in clusters 11, 12, 13 and 14. h Cell-type composition of each collected sample.
Fig. 6
Fig. 6. Activity programmes.
a UMAP embedding of combined RNA velocity and PAGA. Connections between clusters are established based on PAGA, while the direction of the arrows is inferred from the RNA velocity analysis. b Heatmap of the module scores of each cluster (Y axis) for the identified activity programmes (cNMF modules, “Methods”). The top 10 genes were used to calculate the module score. c Activity of the transcription programmes identified by the cNMF analysis for each identified cluster. The heatmap colours represent the z-score of the gene expression for the top 10 marker genes for each activity programme. d RNA velocity and gene expression of candidate genes Ltf (upper) and Trp63 (lower). The left side of the plots shows the current gene expression levels. The right side of the plots shows the RNA velocity. The green colour represents a high velocity, therefore upregulation, while the red colour denotes downregulation. e Immunohistochemistry of Ltf in Ad-Cas9 control samples and in representative Eml4-Alk lung sections at different time points after adenoviral instillation. Scale bar 20 μm. Below, a quantification of the number of Ltf positive cells mm−2 of bronchi of the animals. Each dot represents a bronchus. Mann–Whitney test, two-tailed. Data are presented as mean values +/− SEM. n = 3 mice per group were analysed except Ad-Cas9 2 weeks and Ad-EA 4 weeks (n = 4) and Ad-Cas9 6 endpoint (n = 2). f Schematic representation of the two routes followed by Club cells upon Eml4-Alk transformation.
Fig. 7
Fig. 7. Validation of the tumorigenic routes.
a Violin plot depicting cNMF module scores in human LUAD based on single-cell transcriptome of patient-specific tumour cells. Patients 3 to 7 had a LUAD. b Violin plot depicting cNMF module scores in human LUAD based on single-cell transcriptome. TS1-TS3 represent 3 tumour-specific transcription states of LUAD, and non-lung metastases labelled as “Malignant cells”. c Forest plot depicting the Hazard Ratio (HR) and its confidence interval (CI) based on a Cox proportional hazards model in LUAD patients in TCGA for each gene expression module (n = 533). d Correlation-based heatmap of the cluster composition of our original clusters and the clusters from ref. . The rows are the integrated clusters, while columns are the original clusters from both studies. To avoid confusion, the clusters from ref. , are referred by name and coloured in blue. e AT1-like, Regeneration-like and Stem-like tumour module z-scores for cells belonging to the original clusters from this study and the one from ref. . The clusters from the Yang study are referred by name and coloured in blue. Cluster 13 vs. AT1-like, AT1-like tumour module, p-value 2.11 × 10−41; Cluster 12 vs. Late Gastric, Regeneration-like tumour module p value = 2.29 × 10−31. n1 = 3502, n2 = 4710, n3 = 3091, n4 = 3027, n5 = 2717, n6 = 1706, n7 = 2684, n8 = 506, n9 = 518, n10 = 594, n11 = 64, n12 = 1526, n13 = 103, nMesenchymal-1 = 9950, nMesenchymal-2 = 9028, nMesenchymal-1 (Met)=818, nMesenchymal-2 (Met)=3829, nEarly EMT-1 = 3513, nEarly EMT-2 = 460, nAT2-like = 4680, nPre-EMT = 2108, nEarly gastric = 2175, nLung progenitor-like = 1074, nGastric-like = 4906, nEndoderm-like = 10272, nLate Gastric = 714, nHigh plasticity = 2439, nAT1-like = 1788. The p values were calculated based on a two-sided t test with the following results: t = 20.677, df = 120.21, 95% CI = 0.554–0.671 and t = 5.2966, df = 1678.6, 95% CI = 0.0497–0.108, respectively.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Yau C, et al. Cell-of-origin patterns dominate the molecular classification of 10,000 tumors from 33 types of cancer. Cell. 2018;173:291.e6–304.e6. - PMC - PubMed
    1. Capper D, et al. DNA methylation-based classification of central nervous system tumours. Nature. 2018;555:469–474. doi: 10.1038/nature26000. - DOI - PMC - PubMed
    1. Sill M, Plass C, Pfister SM, Lipka DB. Molecular tumor classification using DNA methylome analysis. Hum. Mol. Genet. 2020;29:R205–R213. doi: 10.1093/hmg/ddaa147. - DOI - PubMed
    1. Lipka DB, et al. RAS-pathway mutation patterns define epigenetic subclasses in juvenile myelomonocytic leukemia. Nat. Commun. 2017;8:2126. doi: 10.1038/s41467-017-02177-w. - DOI - PMC - PubMed
    1. Oakes CC, et al. DNA methylation dynamics during B cell maturation underlie a continuum of disease phenotypes in chronic lymphocytic leukemia. Nat. Genet. 2016;48:253–264. doi: 10.1038/ng.3488. - DOI - PMC - PubMed

Publication types