Transcription factor Etv5 is essential for the maintenance of alveolar type II cells

Abstract

Alveolar type II (AT2) cell dysfunction contributes to a number of significant human pathologies including respiratory distress syndrome, lung adenocarcinoma, and debilitating fibrotic diseases, but the critical transcription factors that maintain AT2 cell identity are unknown. Here we show that the E26 transformation-specific (ETS) family transcription factor Etv5 is essential to maintain AT2 cell identity. Deletion of Etv5 from AT2 cells produced gene and protein signatures characteristic of differentiated alveolar type I (AT1) cells. Consistent with a defect in the AT2 stem cell population, Etv5 deficiency markedly reduced recovery following bleomycin-induced lung injury. Lung tumorigenesis driven by mutant KrasG12D was also compromised by Etv5 deficiency. ERK activation downstream of Ras was found to stabilize Etv5 through inactivation of the cullin-RING ubiquitin ligase CRL4^COP1/DET1 that targets Etv5 for proteasomal degradation. These findings identify Etv5 as a critical output of Ras signaling in AT2 cells, contributing to both lung homeostasis and tumor initiation.

Keywords: Cop1; Etv5; Ras; lung cancer.

Fig. 1.

Expression of Etv5 in AT2 cells in adult lung. (A) Etv1, Etv4, and Etv5 gene expression in C57BL/6 mouse lung (n = 10). (B, Left) Immunostaining of wild-type mouse lung. The arrow indicates a Sftpc⁺Scgb1a1⁺Etv5⁺ cell. (Right) Percentage of Sftpc⁺Etv5⁺cells in total Sftpc⁺ cells counted in three wild-type mice. Data are shown as mean ± SEM. (C) Sftpc^CreERT2/+Rosa^tdTomato/+ lung after dosing with tamoxifen for 5 d and then s.c. dosing with bleomycin for 1 wk. Arrowheads point to the nuclei of AT2-derived AT1 cells.

Fig. 2.

Etv5 is required to maintain AT2 cell fate in vitro. (A, Left) Colonies formed by AT2 cells after culture on Matrigel with or without growth factors. (Right) Each symbol indicates cells from one mouse. Data are shown as mean ± SEM. **P < 0.01, *P < 0.05 (Student's t test). (B, Left) Heatmap shows expression of AT1 and AT2 marker genes (7) in cells sorted and then cultured with 4-OHT. RNA-seq values are scaled and centered, normalized counts to account for library size and gene length (n = 4 mice per genotype). (Right) The volcano plot indicates the effect of Etv5 deficiency on AT1 (blue) and AT2 (red) marker genes. (C) Immunostaining of colonies formed by AT2 cells after induction of Kras^d12.

Fig. S1.

Isolation and culture of primary AT2 cells. (A) Sorting of AT2 cells by flow cytometry. Lin, lineage markers CD31, F4/80, TER-119, CD16/32, CD11b, and CD45; SSC, side-scattered light. (B) Colonies formed by AT2 cells after culture on Matrigel with or without growth factors. BEGM, bronchial epithelial cell growth medium.

Fig. 3.

Etv5 regulates AT2 cell identity in vivo. (A) Scheme for deleting Etv5 from AT2 cells in the lungs. Tam, tamoxifen. (B) Volcano plot indicating the effect of Etv5 deficiency on AT1 (blue) and AT2 (red) marker gene expression in tdTomato⁺ cells isolated as in A. (C) Comparison of AT1 and AT2 marker gene expression in Etv5^+/+ versus ΔEtv5 tdTomato⁺ Epcam⁺ CD24⁻ lineage⁻ cells by single-cell qRT-PCR. The t statistic for each gene is plotted. Orange dots indicate genes significantly changed by Etv5 deficiency (Benjamini–Hochberg adjusted P < 0.05, moderated t test). Genes that are up-regulated in ΔEtv5 tdTomato⁺ cells show larger positive t statistics, whereas those down-regulated show larger negative t statistics. (D) Box-and-whisker plots showing the expression of the most differentiated AT2 and AT1 genes, Sftpc and Clic5, in tdTomato⁺ cells by single-cell RT-PCR. (E) Consensus sequence motif identified by the MEME motif-finding software in the Etv5-binding sites found in C57BL/6 AT2 cells. (F) ChIP-Seq normalized coverage of representative AT2 marker genes.

Fig. S2.

Gene-expression changes caused by Etv5 deficiency. (A) Heat map of the top 50 most differentially expressed genes in Etv5^+/+ Sftpc^CreERT2/+ Rosa^tdTomato/+ tdTomato⁺ cells compared with Etv5^fl/- Sftpc^CreERT2/+Rosa^tdTomato/+ tdTomato⁺ cells at 7 d after the last dose of tamoxifen. (B) Gene set enrichment analysis identifying Gene Ontology Biological Process gene sets that show significant enrichment when comparing differential expression based on RNA-seq fromEtv5^+/+ Sftpc^CreERT2/+ Rosa^tdTomato/+ and Etv5^fl/- Sftpc^CreERT2/+Rosa^tdTomato/+ tdTomato⁺ cells. (C) Cell-cycle genes identified by Gene Ontology analysis that are significantly down-regulated in Etv5^fl/− Sftpc^CreERT2/+Rosa^tdTomato/+ tdTomato⁺ cells compared with Etv5^+/+ Sftpc^CreERT2/+ Rosa^tdTomato/+ cells. (D) Volcano plots indicating the difference in gene expression between tdTomato⁺ cells and whole-lung tissue (Left) and between ΔEtv5 and Etv5^+/+ tdTomato⁺ cells (Right). There were 2,819 genes more highly expressed in tdTomato⁺ cells than in whole-lung tissue (Benjamini–Hochberg adjusted P < 0.05, fold-change >2), and 321 genes were expressed at lower levels in ΔEtv5 tdTomato⁺ cells than in wild-type cells (Benjamini–Hochberg adjusted P < 0.05, fold-change is less than −2). The Venn diagram shows the overlap of these two gene sets. (E) Single-cell qRT-PCR of AT1 and AT2 marker genes. Each point represents one gene and shows the t statistic comparing Etv5^+/+ tdTomato⁺ and ΔEtv5 tdTomato⁺ cells isolated from tamoxifen-treated Sftpc^CreERT2/+Kras^LSL/+ mice as in Fig. 3A. Genes that are significantly changed (Benjamini–Hochberg adjusted P < 0.05) are shown in orange. (F) Single-cell qRT-PCR of AT1 and AT2 marker genes. Each point represents one gene and shows the t statistic comparing Etv5^+/+ lineage⁻ CD24⁻ Epcam⁺ and ΔEtv5 lineage⁻ CD24⁻ Epcam⁺ cells isolated from tamoxifen-treated Rosa^CreERT2/+ mice. Genes that are significantly changed (Benjamini–Hochberg adjusted P < 0.05) are shown in orange. (G) Heatmap showing the row scaled signal for Etv5 and H3k4me3 on 1,000 randomly selected ETV5 sites on a 5 kb+/5 kb− window. The H3k4me3 window is centered on the Etv5 binding site. (H) The graph indicates the degree of enrichment of ChIP-seq–identified peaks adjacent to the AT1 and AT2 marker genes compared with background calculated by Macs2.

Fig. 4.

Etv5 in AT2 cells is required for the repair of bleomycin-induced lung injury and for Kras-driven tumor initiation. (A) Scheme for the study of bleomycin-induced lung injury. Tam, tamoxifen. (B) Bleomycin-treated Sftpc^CreERT2/+Rosa^tdTomato/+ lung showing tdTomato expression (red) and BrdU incorporation (green). Arrowheads indicate tdTomato⁺ BrdU⁺ cells. (C) Lungs in B with tdTomato⁺ AT1 cells pseudocolored blue after being differentiated from AT2 cells by their lower staining intensity and elongated shape. (D) Quantification of tdTomato⁺ BrdU⁺ cells and relative numbers of AT1 versus AT2 cells. Data are shown as mean ± SEM; **P < 0.01, unpaired one-tailed Student's t test with Welch's correction; *P < 0.05, unpaired one-tailed Student's t test. AT1 and AT2 cells are distinguished as in C. (E) Immunostaining of a Kras^LSL/+ lung lesion 20 wk after infection with adenovirus expressing Cre. Arrowheads indicate Etv5⁺ cells adjacent to the lesions. AAH, atypical adenomatous hyperplasia. (F) Representative lung lesions at 20 wk after infection with adenovirus expressing Cre. The graph shows mean lesion numbers. Each dot represents one mouse. Bars indicate mean ± SEM. **P < 0.01, *P < 0.05 (Mann–Whitney test). (G) The graph indicates the extent to which the floxed Etv5 allele was retained in microdissected tumors from Kras^LSL/+ Etv5^fl/fl mice infected with adenovirus expressing Cre.

Fig. S3.

Etv5 deficiency impairs recovery from bleomycin-induced lung injury. (A) Trichrome staining of lungs. Fibrous connective tissue is blue. Cells are red. (B) Distribution of histologic scores of lung lesions in bleomycin-treated animals. Histologic lesions were scored according to the following matrix: 0, no histologic lesions identified; 1, focal increased cellularity with minimal to no fibrosis; 2, focal to locally extensive subpleural fibrosis with little to no extension into adjacent alveoli; 3, subpleural fibrosis and inflammation with moderate, focal, or mild multifocal extension and disruption of the alveolar architecture; 4, extensive, multifocal fibrosis and inflammation with marked disruption of the alveolar architecture. P = 0.07, Student t test.

Fig. 5.

ERK activation stabilizes ETV5. (A) Immunoblots of normal (N) and tumor (T) lung tissue from two Kras^LSL/+ and two Kras^LSL/+Trp53^fl/fl mice after infection with adenovirus expressing Cre. B6, C57BL/6 wild-type. pRSK1 (phospho-Ribosomal protein S6 kinase alpha-1) is a readout of ERK activity. (B) Etv5 gene expression in lung tissue from wild-type (B6) mice and in normal or tumor lung tissue from Kras^LSL/+Trp53^fl/fl (KP) mice after infection with adenovirus expressing Cre. (C) Immunoblots of HT55 cells. Treatments shown on the left were for 1 h, with the exception of MG-132, which was added for 3 h. Numbers indicate relative ETV5 mRNA expression by real-time RT-PCR. (D) Immunoblots of E1A/Hras-transformed MEFs. Inhibitor treatments shown on the left were for 1 h. Numbers indicate relative Etv5 mRNA expression by real-time RT-PCR.

Fig. 6.

Active ERK stabilizes ETV5 by inhibiting the COP1 ubiquitin ligase. (A) Organization of the CRL4^COP1/DET1 ubiquitin ligase. NED, Nedd8. (B) Immunoblots of E1A/Hras-transformed MEFs treated with MEK inhibitor (MEKI) for 1 h. (C) Immunoblots of H460, H1299, and Kras^d12ΔTrp53 (KP1 and KP2) cells transfected with Cop1 siRNAs and then treated with ERK inhibitor (ERKI) for 1 h. SiNC1, siRNA negative control 1. (D) Immunoblots of human NSCLC lines treated with MLN4924 for 1 h before the addition of ERK inhibitor.

Fig. S4.

Localization, abundance, and interactions of components of the Cop1 complex and activity of the core Cullin complex are unchanged by ERK activity. (A) Immunoblots of E1A/Hras-transformed wild-type or Det1^3×f/3×f MEFs that were treated with 1 µM MEK inhibitor (MEKI) for 1 h. 3×Flag-DET1 was immunoprecipitated (IP) with anti-Flag beads where indicated. (B) Immunoblots of 293.Kras cells transfected with Flag-ETV5 were treated with 100 ng/mL doxycycline (Dox) overnight, then with 1 µM MLN4924 for 1 h, and then with 1 µM MEK inhibitor for another 1 h. Flag-ETV5 was immunoprecipitated with anti-Flag beads where indicated. (C) Immunoblots of E1A/Hras-transformed MEFs that were treated with 1 µM MEK inhibitor for 1 h and then were fractionated into cytosol (Cyto) and nuclei (Nucl). WCL, whole-cell lysate starting material. (D) Immunoblots of E1A/Hras-transformed MEFs treated with DMSO or 1 µM ERK inhibitor (ERKI) for 1 h and fractionated through a Superose 6 10/300 GL column. (E) Immunoblots of COP1-containing complexes affinity purified from Cop1^fh/fh MEFs treated with 1 µM ERK inhibitor or DMSO for 1 h. Flag.HA-COP1 was immunoprecipitated with anti-Flag beads and eluted with 3×Flag peptide. The eluates were fractionated through a Superose 6 10/300 GL column and concentrated by acetone precipitation for Western blot analysis. (F) Immunoblots of 293.Kras cells treated with doxycycline overnight and then with MEK inhibitor or MLN4924 for 1 h before UV irradiation. CDT1, chromatin licensing and DNA replication factor 1; CDT2, chromatin licensing and DNA replication factor 2; DDB2, damage-specific DNA-binding protein 2; NED, Nedd8. Each small blue ball represents a ubiquitin in a polyubiquitin chain.

Fig. S5.

Mutagenesis analysis of ETV5 and COP1. (A) Western blots of 293.Kras cells transfected with Flag-tagged wild-type or mutant ETV5, the latter with putative ERK, MSK, RSK, and PKA phosphorylation sites mutated to alanine. Where indicated, at 1 d posttransfection, cells were treated with 100 ng/mL doxycycline (Dox) overnight, then 1 µM MLN4924 for 1 h, and then 1 µM MEK inhibitor (MEKI) for another 1 h. (B) Immunoblots of 293.Kras cells expressing Flag-tagged wild-type or mutant ETV5. Anti-Flag immunoblot samples were resolved by 7% PAGE with 100 µM Phos-tag acrylamide. (C, Lower) Immunoblots of HEK293T cells cotransfected with ETV5 (E), DET1 (D), and either wild-type or mutant COP1. All AP, all putative ERK sites mutated to alanine; All EP, all putative ERK sites mutated to glutamic acid. (Upper) Sites conserved between species are indicated in red.

Fig. 7.

Phosphorylation of DET1 by ERK. (A, Lower) Immunoblots of HEK293T cells cotransfected with ETV5, COP1, and either wild-type or mutant DET1. (Upper) The schematic of DET1 indicates the position of putative ERK sites. Residues conserved through evolution are shown in red. (B, Left) Coomassie blue-stained wild-type DET1 or DET1-COP1–tethered protein that was affinity purified from 293.Kras coexpressing constitutively active (CA) ERK2. Arrowheads indicate DET1-COP1 and DET1 bands. (Right) Percentage of DET1 phosphorylated at Ser66 or Ser458 estimated by MS. aa, the number of amino acids forming the linkers between DET1 and COP1. (C, Left) Immunoblots of 293.Kras cells coexpressing wild-type or S458A mutant DET1 with constitutively active (ca) or kinase-dead (kd) ERK2. (Right) Immunoblots of 293.Kras cells after treatment with doxycycline (Dox) and then with DMSO, 1 µM MEK inhibitor (MEKI), or 1 µM ERK inhibitor (ERKI) for 1 h before harvest. (D) Model of ERK regulation of COP1 complex activity. NED, Nedd8.

Fig. S6.

Isotopically labeled peptide standards confirm DET1 phosphorylation at Ser66 and Ser458. (A) MS/MS spectrum of isotopically labeled standard peptide representing phosphorylation on the human DET1 protein at Ser66 (pSer66). The red lowercase residue denotes the position of phosphorylation. An asterisk (P*) indicates the isotopically labeled residue. Observed (Obs.) and expected (Exp.) m/z measurements, as well as charge (z) state are shown. Ions from the b' and y' series are indicated on the spectrum in blue and red; ions demonstrating neutral loss (NL) of the phosphate are shown in green. (B) Extracted ion chromatogram (± 10 ppm) for pSer66 showing isotopically labeled standard peptide (Heavy) and the digested analyte peptide (Light). To display coelution more clearly, the y axis of the light peptide was magnified 20×. Retention time (in minutes) and observed m/z for heavy and light peptides are shown above each corresponding peak. (C) MS/MS spectrum of isotopically labeled standard peptide representing phosphorylation on the human DET1 protein at Ser458 (pSer458) with the isotope label incorporated at Leu (L*). (D) Extracted ion chromatogram (± 10 ppm) for pSer458 presented as in B.

Fig. S7.

Weak interaction between COP1 and DET1. (A) Immunoblots of HEK293T lysate fractions from a Superose 6 10/300 GL column. (B) Silver-stained proteins recovered after serial affinity purification of the indicated overexpressed tagged protein pairs from HEK293F cells. (C) Model for COP1 and DET1 residing in distinct cellular pools.

Fig. S8.

Generation of Cop1^fh, Det1^3×f, and Det1^fl gene-targeted mice. (A) The Flag.HA epitope tag sequence was inserted right after the start codon of Cop1 to create the Cop1^fh knockin allele. (B) The 3×Flag epitope tag sequence was inserted before the Det1 stop codon to create the Det1^3×f knockin allele. (C) Exon 2 of Det1 was flanked with loxP sites to create a conditional Det1 allele. Lower are Southern blots confirming targeting of the Det1 locus and generation of ES cells that are heterozygous for the Det1 floxed (fl) or knockout (−) alleles. ES cell genomic DNA was digested with the restriction enzyme indicated and hybridized to a probe sequence outside of the homology arms of the targeting vector. The locations of the probe sequences are indicated on the maps above.