The endoderm-lineage transcription factor FOXA2 has been shown to inhibit lung tumorigenesis in in vitro and xenograft studies using lung cancer cell lines. However, FOXA2 expression in primary lung tumors does not correlate with an improved patient survival rate, and the functional role of FOXA2 in primary lung tumors remains elusive. To understand the role of FOXA2 in primary lung tumors in vivo, here, we conditionally induced the expression of FOXA2 along with either of the two major lung cancer oncogenes, EGFRL858R or KRASG12D, in the lung epithelium of transgenic mice. Notably, FOXA2 suppressed autochthonous lung tumor development driven by EGFRL858R, whereas FOXA2 promoted tumor growth driven by KRASG12D. Importantly, FOXA2 expression along with KRASG12D produced invasive mucinous adenocarcinoma (IMA) of the lung, a fatal mucus-producing lung cancer comprising approximately 5% of human lung cancer cases. In the mouse model in vivo and human lung cancer cells in vitro, FOXA2 activated a gene regulatory network involved in the key mucous transcription factor SPDEF and upregulated MUC5AC, whose expression is critical for inducing IMA. Coexpression of FOXA2 with mutant KRAS synergistically induced MUC5AC expression compared with that induced by FOXA2 alone. ChIP-seq combined with CRISPR interference indicated that FOXA2 bound directly to the enhancer region of MUC5AC and induced the H3K27ac enhancer mark. Furthermore, FOXA2 was found to be highly expressed in primary tumors of human IMA. Collectively, this study reveals that FOXA2 is not only a biomarker but also a driver for IMA in the presence of a KRAS mutation.

Significance:

FOXA2 expression combined with mutant KRAS drives invasive mucinous adenocarcinoma of the lung by synergistically promoting a mucous transcriptional program, suggesting strategies for targeting this lung cancer type that lacks effective therapies.

Lung cancer is a devastating disease causing 1.76 million deaths worldwide every year (1). Recent advances in molecular genetics using human specimens revealed driver oncogenes and tumor suppressors that are responsible for causing lung cancer, including non–small cell lung cancer (NSCLC, the most common type; refs. 2, 3) and small-cell lung cancer (SCLC; ref. 4). These human molecular studies also revealed lung lineage-specific transcription factors such as NKX2–1 (also known as TTF-1; refs. 3, 5–8) and SOX2 (2, 9) that initiate and/or promote lung tumorigenesis, which have also been validated by autochthonous mouse models (10–12). In addition to the genetic studies, comprehensive pathological studies have determined consensus biomarkers that define lung cancers, especially NKX2–1, that have been clinically used to identify primary non-mucinous lung adenocarcinoma (LUAD, the most prevalent pathological type of NSCLC) and SCLC (13, 14). Supporting the pathological observation that NKX2–1 is expressed in non-mucinous LUAD but not in mucinous LUAD [also known as invasive mucinous adenocarcinoma (IMA) of the lung comprising ∼5% of lung cancer cases], mutant KRAS (a driver oncogene most frequently seen in LUAD) along with the reduced expression of NKX2–1 resulted in the development of mucinous lung tumors (IMA) in autochthonous mouse models (10, 15), suggesting that lung lineage-specific transcription factors are critically involved in lung tumor development and pathogenesis.

FOXA2 is a transcription factor that is expressed in the endodermal lineage, including normal lung epithelium (16). FOXA2 is also expressed in lung cancer, including NSCLC and SCLC (17). Reduction of FOXA2 using shRNA in an A549 human lung carcinoma line (KRASG12S;CDKN2Adel) increased the expression of EMT (epithelial–mesenchymal transition) markers, cell migration and metastasis whereas induction of FOXA2 in an H446 human SCLC cell line (TP53G154V;RBnull;PTENnull) and an H1299 human NSCLC cell line (NRASQ61K;TP53null), both of which are cell lines derived from metastatic sites, decreased the expression of the EMT markers, cell migration and metastasis in vitro and in a nude mouse model. FOXA2 did not influence proliferation in either the loss- or gain-of-function study in these cell lines (18). However, another report indicates that induction of FOXA2 in an H358 bronchioalveolar carcinoma cell line (KRASG12C;TP53null; NSCLC cells), which are also derived from a metastatic site, suppressed growth of the cells by arresting proliferation and increasing apoptosis in vitro (19). In a study using a mouse lung cancer cell line (KrasG12D; Tp53 null; derived from a nonmetastatic primary tumor), reduction of Foxa2 using shRNA also increased the growth of the cells on the subcutaneously transplanted site as well as increased the rate of lung metastasis in a nude mouse model (20). These cell line–based in vitro studies and xenograft models suggest that FOXA2 is a tumor suppressor especially for regulating metastasis. However, the expression level of FOXA2 in primary lung tumors does not significantly correlate with the survival outcome in human NSCLC cases (17, 20). In addition, FOXA2 has been shown to promote tumorigenesis in other cancers, including colon, pancreas, prostate, and esophagus (21–24), suggesting that a further study using a different experimental model is required to elucidate the role of FOXA2 in lung tumorigenesis, especially for primary lung tumors. In the present study, we sought to determine the role of FOXA2 in autochthonous primary lung tumors by conditionally inducing FOXA2 in lung epithelium of a KRAS-mutant or an EGFR-mutant lung cancer transgenic mouse model. KRAS and EGFR mutations are the most prevalent genetic alterations seen in LUAD (1, 3). Consistent with the previous findings in the in vitro and xenograft mouse studies (18–20), FOXA2 suppressed the growth of autochthonous EGFR-mutant lung tumors; however, FOXA2 promoted the growth of autochthonous KRAS-mutant lung tumors. Importantly, FOXA2 along with mutant KRAS produced IMA-like mucinous lung tumors in this autochthonous mouse model. FOXA2 also induced the in vitro expression of mucous genes seen in IMA, including MUC5AC and SPDEF, in an H441 human lung adenocarcinoma cell line that carries a KRAS mutation (KRASG12V;TP53R158L). Here, we demonstrate a novel context-dependent role of FOXA2 in autochthonous primary lung tumors developed in transgenic mice carrying an EGFR or a KRAS mutation.

Mice

[tetO]-Foxa2 (rat Foxa2) mice were obtained from Jeffrey Whitsett and Gang Chen at Cincinnati Children's Hospital Medical Center (CCHMC) and the University of Cincinnati College of Medicine, Cincinnati, OH (25) and crossed with Scgb1a1-rtTA;[tetO]-EGFRL858R or Scgb1a1-rtTA;[tetO]-Kras4bG12D (26, 27) to develop Scgb1a1-rtTA;[tetO]-EGFRL858R;[tetO]-Foxa2 (FVB/N;B6;CBA mixed strain) or Scgb1a1-rtTA;[tetO]-Kras4bG12D;[tetO]-Foxa2 (FVB/N strain) as described previously (10). Transgenic mice were provided chow containing doxycycline (625 mg/kg chow) beginning at 4–5 weeks of age. Mouse maintenance and procedures were approved in accordance with the institutional protocol guidelines of Cincinnati Children's Hospital Medical Center Institutional Animal Care and Use Committee. See Supplementary Data and Supplementary Tables S1–S4 for further mouse information.

Human specimens

Paraffin sections for lung adenocarcinoma were obtained from Kawasaki Medical School, Okayama (approval # 1310) and Hiroshima University (approval # E-1919) in accordance with institutional guidelines for use of human tissue for research purposes. Written informed consent was obtained from all participants. Patients' information is summarized in Supplementary Table S5.

Histology and IHC

Staining [hematoxyin and eosin (H&E), Alcian blue, and IHC] was performed using 5 μm paraffin-embedded lung sections as described previously (10). The antibody information is available in Supplementary Data. The number of different types of tumors per H&E-stained section was counted in at least 3 mice of each group (see Supplementary Table S3 and S4 for details).

Cell culture, lentivirus and/or retrovirus infection, siRNAs, CRISPRi, immunoblotting, coimmunoprecipitation, RNA-seq, TaqMan gene expression analysis

H441 and A549 human lung cell lines were obtained from the ATCC on May 3, 2004 and July 26, 2004, respectively. The BEAS-2B lung cell line was obtained from Thomas Korfhagen at CCHMC on April 27, 2006. Cell line authentication was conducted by Genetica DNA Laboratories (Labcorp) on December 3, 2010 for H441 and A549 cells and May 13, 2013 for BEAS-2B cells. Mycoplasma testing was performed by Universal Mycoplasma Detection Kit on December 8, 2022 (cat# 30–1012K, ATCC). Cell passage numbers used in this study were p81-p100 for A549 cells, p51-p84 for H441 cells and p24-p41 for BEAS-2B cells. Lentiviral vector delivering FOXA1, FOXA2 or NKX2–1 was made by inserting mouse Foxa1, rat Foxa2 or rat Nkx2–1 into the PGK-IRES-EGFP vector as described previously (10). CRISPR interference (CRISPRi; CRISPR/dCas9-KRAB; ref. 28) lentiviral vector (pLV hU6-sgRNA hUbC-dCas9-KRAB-T2a-Puro; Plasmid #71236), pBabe (Plasmid #1764) and pBabe K-Ras 12V (human KRASG12V; Plasmid #12544) retroviral vectors were obtained from Addgene. The siRNA-mediated knockdown analysis is described previously in Supplementary Data.

H441 cells that stably express dCas9-KRAB were developed using the CRISPR/dCas9-KRAB lentiviral vector as we described previously (29). H441 cells that stably express dCas9-KRAB were infected with the PGK-IRES-EGFP lentiviral vector carrying rat Foxa2 or empty control to develop H441 cells that stably express dCas9-KRAB with or without FOXA2 for the CRISPRi experiments. CRISPRi using these H441 cells were performed as described previously (29) by transiently transfecting synthetic sgRNA from the Invitrogen custom TrueGuide gRNA (sgRNA) ordering tool (Thermo Fisher Scientific). Nontargeted gRNA (sgRNA) was used as a negative control (cat# A35526, Thermo Fisher Scientific). H441 or BEAS-2B cells were infected with the lentiviral vector expressing FOXA1, FOXA2 and/or the retroviral vector expressing KRASG12V. A549 cells were infected with the PGK-IRES-EGFP lentiviral vector carrying rat Foxa2, rat Nkx2–1 or empty control to develop A549 cells with or without ectopic FOXA2 or NKX2–1. The empty vectors were used as controls.

Protein and RNA were extracted as described previously (10). Coimmunoprecipitation analysis is described previously in Supplementary Data. Immunoblotting assays were performed as described previously (10). The antibody information with RRID is available in Supplementary Data. RNA-seq using biological triplicates were performed as described previously (29) except that RNA was obtained from H441 cells that were infected with lentivirus (Control or rat Foxa2). TaqMan gene expression analysis was performed as described previously (10). The TaqMan probe information is available in Supplementary Data.

TCGA LUAD data and ChIP-seq analysis

Normalized gene expression data from The Cancer Genome Atlas (TCGA) LUAD RNA-seq datasets were retrieved from (3). Chromatin immunoprecipitation sequencing (ChIP-seq) datasets were retrieved from GSE43252 (15), ENCODE using accession number ENCFF686MSH (ENCSR000BRE; refs. 30, 31), GSE48930 (32) and our previous study (33). See Supplementary Data for FOXA2 ChIP-seq analysis in H441 cells in details.

Statistical analysis

Statistical differences were determined using two-tailed and unpaired Student or Welch t test or the Kolmogorov–Smirnov test. Error bars represent SEM. The difference between two groups was considered significant when the P value was <0.05.

Data availability

The RNA-seq and ChIP-seq data have been deposited in the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE210121.

FOXA2 suppresses growth of EGFR-mutant lung tumors but promotes growth of KRAS-mutant lung tumors

To assess the role of FOXA2 in autochthonous primary lung tumors, we conditionally induced the expression of FOXA2 along with either of the lung oncogenes EGFRL858R or KRASG12D in the lung epithelium by generating triple transgenic mice [Scgb1a1-rtTA; (tetO)-Kras4bG12D; (tetO)-Foxa2 or Scgb1a1-rtTA; (tetO)-EGFRL858R; (tetO)-Foxa2; hereafter KrasG12D;Foxa2 or EGFRL858R;Foxa2] that express transgenes regulated by the Tet-On system (Fig. 1A; refs. 25–27). Consistent with previous reports, lung tumors developed in EGFRL858R- (Fig. 1BD; Supplementary Table S1; ref. 27) or KRASG12D-expressing mice (Fig. 1EG; Supplementary Table S2; ref. 26) but not in mice expressing FOXA2 alone (Fig. 1BG; Supplementary Table S1 and S2; ref. 25). The number and volume of lung tumors that developed in EGFRL858R and FOXA2 co-expressing mice significantly decreased compared with those in the EGFRL858R-expressing mice (Fig. 1BD; Supplementary Table S1), consistent with previous in vitro and xenograft studies (18–20). In contrast, the volume (but not the number) of lung tumors developed in KRASG12D and FOXA2 coexpressing mice was significantly promoted compared with that in KRASG12D-expressing mice (Fig. 1EG; Supplementary Table S2). These in vivo results indicate that FOXA2 suppresses initiation and promotion of EGFR-mutant autochthonous lung tumors whereas FOXA2 promotes the growth of KRAS-mutant autochthonous lung tumors.

Figure 1.

EGFRL858R-autochthonous lung tumors are suppressed by FOXA2, whereas KrasG12D-lung autochthonous tumors are promoted by FOXA2 in mice. A, Top, schematic view of the transgenic mouse model that conditionally induces the expression of EGFRL858R or KrasG12D along with Foxa2 in lung epithelial cells, including airway club cells and alveolar type 2 cells, using the Tet-On inducible system upon doxycycline administration. Bottom, timeline for doxycycline administration and sacrifice. B, Representative microCT images detecting mouse lung tumors induced by EGFRL858R and/or FOXA2 that were conditionally induced in lung epithelium. C, Shown are per mouse tumor number counted using microCT images (n ≥ 10). FOXA2 significantly suppressed the number of lung tumors developed by EGFRL858R (Kolmogorov–Smirnov test). D, Tumor volume per mouse measured using microCT images (n ≥ 10) is shown. FOXA2 significantly suppressed the volume of lung tumors developed by EGFRL858R (Kolmogorov–Smirnov test). E, MicroCT images detecting mouse lung tumors developed by KRASG12D and/or FOXA2 that were conditionally induced in lung epithelium. F, Per mouse tumor numbers counted using microCT images (n ≥ 6) are shown. FOXA2 did not influence the number of lung tumors developed by KRASG12D (Kolmogorov–Smirnov test). G, Tumor volume per mouse measured using microCT images (n ≥ 6) is shown. FOXA2 significantly induced the volume of lung tumors developed by KRASG12D (Kolmogorov–Smirnov test). Data are presented as mean ± SEM.

Figure 1.

EGFRL858R-autochthonous lung tumors are suppressed by FOXA2, whereas KrasG12D-lung autochthonous tumors are promoted by FOXA2 in mice. A, Top, schematic view of the transgenic mouse model that conditionally induces the expression of EGFRL858R or KrasG12D along with Foxa2 in lung epithelial cells, including airway club cells and alveolar type 2 cells, using the Tet-On inducible system upon doxycycline administration. Bottom, timeline for doxycycline administration and sacrifice. B, Representative microCT images detecting mouse lung tumors induced by EGFRL858R and/or FOXA2 that were conditionally induced in lung epithelium. C, Shown are per mouse tumor number counted using microCT images (n ≥ 10). FOXA2 significantly suppressed the number of lung tumors developed by EGFRL858R (Kolmogorov–Smirnov test). D, Tumor volume per mouse measured using microCT images (n ≥ 10) is shown. FOXA2 significantly suppressed the volume of lung tumors developed by EGFRL858R (Kolmogorov–Smirnov test). E, MicroCT images detecting mouse lung tumors developed by KRASG12D and/or FOXA2 that were conditionally induced in lung epithelium. F, Per mouse tumor numbers counted using microCT images (n ≥ 6) are shown. FOXA2 did not influence the number of lung tumors developed by KRASG12D (Kolmogorov–Smirnov test). G, Tumor volume per mouse measured using microCT images (n ≥ 6) is shown. FOXA2 significantly induced the volume of lung tumors developed by KRASG12D (Kolmogorov–Smirnov test). Data are presented as mean ± SEM.

Close modal

FOXA2 along with mutant KRAS induces IMA of the lung

To determine whether FOXA2 influences tumorigenesis through proliferation and/or apoptosis in the autochthonous lung tumors developed by EGFRL858R or KRASG12D, the expression of PHH3 and caspase-3 was assessed. The expression of PHH3 (arrows), a marker for dividing cells, was frequently observed in lung tumors of mice expressing only EGFRL858R but not in the lungs of mice coexpressing EGFRL858R and FOXA2 (Supplementary Fig. S1A and S1B and Supplementary Table S3). The expression of caspase-3, a marker for apoptosis, was not altered in the mice expressing only EGFRL858R compared with the mice coexpressing EGFRL858R and FOXA2 (Supplementary Fig. S1A and S1C and Supplementary Table S3). Lung tumors were hardly seen in the mice coexpressing EGFRL858R and FOXA2 compared with the mice expressing only EGFRL858R (Supplementary Fig. S1D and Supplementary Table S3). These results indicate that FOXA2 suppresses the growth of the EGFR-mutant lung tumors by reducing cell proliferation associated with tumor initiation by mutant EGFR, which is consistent with previous studies (19, 20) but not with another study (18) though these studies used isolated human KRAS-mutant cancer cell lines (18, 19) or a mouse Kras-mutant cancer cell line (20). In contrast, the expression of PHH3 (arrows) was induced in mice coexpressing KRASG12D with FOXA2 compared with mice expressing only KRASG12D (Fig. 2A and B; Supplementary Table S4). The expression of caspase-3 was not altered in mice coexpressing KRASG12D and FOXA2 compared with mice expressing only KRASG12D (Fig. 2A and C; Supplementary Table S4). In addition, lung tumors developed in mice coexpressing KRASG12D and FOXA2 were comprised of more invasive adenocarcinoma cells than those developed in mice expressing only KRASG12D (Fig. 2D). Importantly, 48% of lung tumors in mice coexpressing KRASG12D with FOXA2 were mucinous lung tumors that stained with Alcian blue, the majority being located in the alveolar region of the lung (Fig. 2D; Supplementary Fig. S2 and Supplementary Table S4). Mice expressing only KRASG12D or EGFRL858R did not develop mucinous lung tumors (Fig. 2; Supplementary Fig. S1 and Supplementary Tables S3 and S4). Consistent with Fig. 1 and Supplementary Table S1 and S2, lung tumors were not observed in mouse lungs overexpressing FOXA2 (Fig. 2; Supplementary Fig. S1 and Supplementary Tables S3 and S4). These results indicate that FOXA2 promotes the growth of the KRAS-mutant lung tumors by inducing cell proliferation and a mucinous phenotype in the tumors.

Figure 2.

FOXA2 promotes the proliferation of KRASG12D lung tumor cells in mice. A, Analyses of histology and IHC on mouse lungs that conditionally expressed FOXA2 and/or KRASG12D were performed. Alcian blue staining was performed to detect mucins. Staining of PHH3 (phosphohistone H3; a marker for proliferation) and caspase-3 (a marker for apoptosis) was performed to assess whether FOXA2 influenced proliferation and apoptosis of KRASG12D lung tumors. Black dots in the nuclei indicate the expression of FOXA2. The x-axis indicates mouse genotype. Scale bar, 20 μm. B, Number of PHH3-positive cells in lungs of different mouse genotypes was counted in three different views per section per mouse. Number of PHH3-positive cells was significantly increased in lungs of KrasG12D;Foxa2 mice compared with those of KrasG12D mice (n = 3 mice). C, Number of caspase-3–positive cells in lungs of different mouse genotypes was counted in three different views per section per mouse. Caspase-3 (a marker for apoptosis) was not altered in lungs of different mouse genotypes (n = 3 mice). D, Differences in histology of mouse lung tumors developed by KRASG12D and/or FOXA2 were assessed. Invasive mucinous adenocarcinoma was significantly induced by FOXA2 in the presence of KrasG12D (n ≥ 9). AAH, atypical adenomatous hyperplasia. All tests are unpaired, two-tailed Student t tests. Data are presented as mean ± SEM.

Figure 2.

FOXA2 promotes the proliferation of KRASG12D lung tumor cells in mice. A, Analyses of histology and IHC on mouse lungs that conditionally expressed FOXA2 and/or KRASG12D were performed. Alcian blue staining was performed to detect mucins. Staining of PHH3 (phosphohistone H3; a marker for proliferation) and caspase-3 (a marker for apoptosis) was performed to assess whether FOXA2 influenced proliferation and apoptosis of KRASG12D lung tumors. Black dots in the nuclei indicate the expression of FOXA2. The x-axis indicates mouse genotype. Scale bar, 20 μm. B, Number of PHH3-positive cells in lungs of different mouse genotypes was counted in three different views per section per mouse. Number of PHH3-positive cells was significantly increased in lungs of KrasG12D;Foxa2 mice compared with those of KrasG12D mice (n = 3 mice). C, Number of caspase-3–positive cells in lungs of different mouse genotypes was counted in three different views per section per mouse. Caspase-3 (a marker for apoptosis) was not altered in lungs of different mouse genotypes (n = 3 mice). D, Differences in histology of mouse lung tumors developed by KRASG12D and/or FOXA2 were assessed. Invasive mucinous adenocarcinoma was significantly induced by FOXA2 in the presence of KrasG12D (n ≥ 9). AAH, atypical adenomatous hyperplasia. All tests are unpaired, two-tailed Student t tests. Data are presented as mean ± SEM.

Close modal

FOXA2 along with mutant KRAS induces biomarkers for IMA

Previously, we determined a gene signature for human IMA of the lung by identifying the genes that are specifically expressed in both human and mouse IMAs (33). To characterize the mucinous lung tumors developed in mice coexpressing KRASG12D with FOXA2 in the lung epithelium, we assessed the expression of selected IMA-signature genes (MUC5AC, MUC5B, AGR2, SPDEF, FOXA3, and HNF4A) along with NKX2–1, which is known to be deficient in human IMA (14). As shown in Fig. 3, these IMA-signature genes were expressed in the mucinous lung tumors developed by coexpression of KRASG12D with FOXA2 whereas anti-mucous NKX2–1 was reduced compared with the other mouse groups (Control, Foxa2 or KrasG12D). House dust mite (HDM)–induced nontumorigenic asthma-like mouse lungs (HDM) also express mucous genes and lack NKX2–1 as we previously reported (10); however, HDM lungs do not express HNF4A contrary to the mucinous lung tumors (KrasG12D;Foxa2). The transgenic expression of FOXA2 alone induced goblet (mucous) cells accompanied by mucous genes though the goblet cells appeared immature compared with those seen in the mucinous lung tumors (KrasG12D;Foxa2) or HDM lungs (Figs. 2 and 3; Supplementary Fig. S3). Of note, although MUC5AC, AGR2, and SPDEF do not initiate lung tumor formation in vivo, they do promote the growth of lung tumors (33, 34–36), suggesting that FOXA2 promotes the growth of KRAS-mutant lung tumors in part through MUC5AC, AGR2, and SPDEF. These results suggest that mucinous lung tumors developed by FOXA2 along with KRASG12D mimic human IMA.

Figure 3.

FOXA2 together with mutant KRAS induces mucinous lung tumors in mice. IHC detecting mucous genes in mouse lungs conditionally induced to express FOXA2 and/or KRASG12D was performed as described previously in Materials and Methods. HDM-challenged mouse lungs (an asthma model) were used as a reference for nontumorigenic lungs that produce mucus. Coexpression of FOXA2 and KRASG12D induced mucinous lung tumors accompanied by the expression of markers for human IMA of the lung. The x-axis indicates mouse genotype. Scale bar, 20 μm.

Figure 3.

FOXA2 together with mutant KRAS induces mucinous lung tumors in mice. IHC detecting mucous genes in mouse lungs conditionally induced to express FOXA2 and/or KRASG12D was performed as described previously in Materials and Methods. HDM-challenged mouse lungs (an asthma model) were used as a reference for nontumorigenic lungs that produce mucus. Coexpression of FOXA2 and KRASG12D induced mucinous lung tumors accompanied by the expression of markers for human IMA of the lung. The x-axis indicates mouse genotype. Scale bar, 20 μm.

Close modal

FOXA2 binds to the locus of the mucus transcription factor Spdef in KrasG12D lung tumors lacking Nkx2–1 in vivo

Snyder and colleagues (15) have performed ChIP-seq to identify FOXA1/FOXA2-binding sites in KrasG12D lung tumors in the presence or absence of Nkx2–1 (an antibody recognizing both FOXA1 and FOXA2 was used for their ChIP-seq). Consistent with our previous report (10), KrasG12D in the absence of Nkx2–1 induced mucinous lung tumors whereas KrasG12D alone induced non-mucinous lung tumors (15). Unexpectedly, the ChIP-seq data indicated that FOXA1/FOXA2 bound to the locus of Muc5b in the NKX2–1-positive lung tumors (non-mucinous lung tumors) but not in the NKX2–1–deleted lung tumors (mucinous lung tumors; Fig. 4). FOXA1/FOXA2 did not bind to the locus of Muc5ac regardless of the expression of Nkx2–1 (Fig. 4A). These results indicate that FOXA1/FOXA2 do not directly induce two major mucins Muc5ac or Muc5b in mucinous lung tumors in mice. However, the ChIP-seq data indicated that FOXA1/FOXA2 bound to the locus of Spdef, which is the key transcription factor that is indispensable for the expression of mucous genes, including Muc5ac and Muc5b (33, 37), in the NKX2–1-deleted mouse lung tumors (mucinous lung tumors; Fig. 4B), suggesting that FOXA2 may directly induce the expression of Spdef, which may in turn increase the expression of Muc5ac and Muc5b in mucinous lung tumors in mice (Fig. 4C).

Figure 4.

FOXA2 binds to the locus of Spdef in mucinous lung tumors in mice. A, ChIP-seq data (15) using an antibody that recognizes FOXA1 and FOXA2 (FOXA1/2) indicated that FOXA1 and FOXA2 bound to the locus of Muc5b in non-mucinous KrasG12D mouse lung tumors (Nkx2–1 positive) but not in the mucinous KrasG12D-mouse lung tumors (Nkx2–1 deleted). B, ChIP-seq data using the antibody described above indicated that FOXA1 and FOXA2 bound to the upstream and intronic regions of Spdef gene in mucinous KrasG12D mouse lung tumors (Nkx2–1 deleted) but not in non-mucinous KrasG12D mouse lung tumors (Nkx2–1 positive). C, Schematic indicates a proposed model of the mechanism by which FOXA2 induces two major IMA mucin genes MUC5AC and MUC5B in mice through induction of the mucus transcription factor SPDEF.

Figure 4.

FOXA2 binds to the locus of Spdef in mucinous lung tumors in mice. A, ChIP-seq data (15) using an antibody that recognizes FOXA1 and FOXA2 (FOXA1/2) indicated that FOXA1 and FOXA2 bound to the locus of Muc5b in non-mucinous KrasG12D mouse lung tumors (Nkx2–1 positive) but not in the mucinous KrasG12D-mouse lung tumors (Nkx2–1 deleted). B, ChIP-seq data using the antibody described above indicated that FOXA1 and FOXA2 bound to the upstream and intronic regions of Spdef gene in mucinous KrasG12D mouse lung tumors (Nkx2–1 deleted) but not in non-mucinous KrasG12D mouse lung tumors (Nkx2–1 positive). C, Schematic indicates a proposed model of the mechanism by which FOXA2 induces two major IMA mucin genes MUC5AC and MUC5B in mice through induction of the mucus transcription factor SPDEF.

Close modal

FOXA2 induces IMA signature genes in a KRAS-mutant human lung cancer cell line in vitro

Having determined that coexpression of mutant KRAS with FOXA2 induced IMA-signature mucous genes in autochthonous lung tumors in mice in vivo, we next sought to determine whether FOXA2 directly or indirectly induced these IMA-signature genes in human lung cancer cells that carry a KRAS mutation in vitro. First, we performed a gain-of-function study by infecting the H441 human lung papillary adenocarcinoma cell line (KRASG12V;TP53R158L) with control lentivirus, lentivirus expressing Foxa1 (mouse; 94% similarity to human FOXA1) or Foxa2 (rat; 95% similarity to human FOXA2) and confirmed their protein expression using specific antibodies (Fig. 5A; Supplementary Fig. S4A). Using the H441 cells, mRNA-seq and TaqMan gene expression analyses were performed. As shown in Fig. 5B and C, Supplementary Fig. S5 and Supplementary Tables S6 and S7, FOXA2 induced 42 genes that are highly expressed in human IMA compared with normal lung (33). Among the 42 genes, 12 genes, including AGR2, BCAS1, CAPN5, CREB3L1, EHF, MMP7, MUC5AC, MUC5B, SLC44A4, SPDEF, TNS4, and TOX3 (red highlighted, Fig. 5B and C; Supplementary Fig. S5 and Supplementary Tables S6 and S7), were the IMA-signature mucous genes (143 genes; ref. 33) highly expressed in both human and mouse IMA. Among these 12 genes, AGR2, EHF, MUC5AC, MUC5B, and SPDEF are involved in mucus production (35, 38–41). Of note, FOXA2 alone did not affect the expression of anti-mucous transcription factor NKX2–1 (Fig. 5C; Supplementary Table S7). We also performed a loss-of-function study using siRNAs targeting endogenous FOXA2 in A549 lung carcinoma cells (KRASG12S;CDKN2Adel), which indicated that FOXA2 is required for the expression of MUC5AC, MUC5B, and SPDEF (Supplementary Fig. S6A and S6B and Supplementary Table S8), consistent with the gain-of-function study results above. ChIP-seq analysis with the FOXA2-specific (no cross-reactivity with FOXA1) antibody (Fig. 5D; Supplementary Fig. S7A and S7B and Supplementary Table S9) using H441 cells (KRASG12V;TP53R158L) and A549 cells (ENCSR000BRE; refs. 30, 31) indicated that FOXA2 bound to the loci of MUC5 and SPDEF in both cell lines, suggesting that FOXA2 may directly induce the expression of MUC5AC, the most highly expressed human IMA mucous gene (33), in human lung carcinoma cells (Fig. 5D and E) contrary to mouse lung carcinoma cells (Fig. 4). Notably, 2 distinct noncoding regions at the MUC5 locus (upstream of MUC5AC and intergenic region of MUC5AC and MUC5B) bound by FOXA2 were enhancer regions bound by SPDEF (a pro-mucous transcription factor), which we previously reported (Fig. 5D, top; ref. 33). Importantly, coimmunoprecipitation experiments showed that FOXA2 interacted with SPDEF (Supplementary Fig. S8), indicating that the enhancer regions function as gene regulatory hubs bound by multiple interacting transcription factors for the expression of MUC5AC and MUC5B (Fig. 5E). Of note, SPDEF did not bind to its own SPDEF locus as FOXA2 did (Fig. 5D, bottom).

Figure 5.

FOXA2 binds to the loci of mucous genes (e.g., MUC5AC and SPDEF) and induces their transcriptional expression in human lung adenocarcinoma cells. A, Immunoblotting (IB) was performed using antibodies against FOXA1 or FOXA2. H441 cells were infected with a lentiviral vector carrying mouse Foxa1 or rat Foxa2 and cell extracts were used for immunoblotting. Control, an empty lentiviral vector. Shown are representative images from three independent experiments. B, RNA-seq was performed using RNAs from H441 cells expressing ectopic FOXA2 as described in A (n = 3). Control, an empty lentiviral vector. Shown are FOXA2-induced genes that are highly expressed in human IMA of the lung. Red, genes that are also significantly expressed in mouse IMA. C, TaqMan gene expression analysis was performed using RNAs as described in B. Results are expressed as mean ± SEM of biological replicates for each group. Control, an empty lentiviral vector.A P value of <0.05 versus control was considered significant (n = 3, unpaired, two-tailed Student t test). Gene expression was normalized by comparison with the constitutive expression of GAPDH. The expression of selected mucous genes identified by the RNA-seq in B was confirmed. Of note, the expression of ectopic rat Foxa2 was induced as expected, whereas the endogenous human FOXA2 was reduced (See Supplementary Fig. S4B). The expression of anti-mucous gene NKX2–1 was not altered by ectopic rat Foxa2. D, Shown is the combined analysis of ChIP-seq datasets. Each genome browser (MUC5AC, MUC5B loci, and SPDEF locus) indicates bam files demonstrating FOXA2 binding in H441 cells at the top and bed files, indicating FOXA2 or SPDEF binding in H441 (lung papillary adenocarcinoma), A549 (lung carcinoma), and MCF7 (breast adenocarcinoma) cells at the bottom. E, Schematic indicates a proposed model of the mechanism by which FOXA2 induces two IMA major mucin genes MUC5AC and MUC5B in humans directly and indirectly through the induction of the mucus transcription factor SPDEF.

Figure 5.

FOXA2 binds to the loci of mucous genes (e.g., MUC5AC and SPDEF) and induces their transcriptional expression in human lung adenocarcinoma cells. A, Immunoblotting (IB) was performed using antibodies against FOXA1 or FOXA2. H441 cells were infected with a lentiviral vector carrying mouse Foxa1 or rat Foxa2 and cell extracts were used for immunoblotting. Control, an empty lentiviral vector. Shown are representative images from three independent experiments. B, RNA-seq was performed using RNAs from H441 cells expressing ectopic FOXA2 as described in A (n = 3). Control, an empty lentiviral vector. Shown are FOXA2-induced genes that are highly expressed in human IMA of the lung. Red, genes that are also significantly expressed in mouse IMA. C, TaqMan gene expression analysis was performed using RNAs as described in B. Results are expressed as mean ± SEM of biological replicates for each group. Control, an empty lentiviral vector.A P value of <0.05 versus control was considered significant (n = 3, unpaired, two-tailed Student t test). Gene expression was normalized by comparison with the constitutive expression of GAPDH. The expression of selected mucous genes identified by the RNA-seq in B was confirmed. Of note, the expression of ectopic rat Foxa2 was induced as expected, whereas the endogenous human FOXA2 was reduced (See Supplementary Fig. S4B). The expression of anti-mucous gene NKX2–1 was not altered by ectopic rat Foxa2. D, Shown is the combined analysis of ChIP-seq datasets. Each genome browser (MUC5AC, MUC5B loci, and SPDEF locus) indicates bam files demonstrating FOXA2 binding in H441 cells at the top and bed files, indicating FOXA2 or SPDEF binding in H441 (lung papillary adenocarcinoma), A549 (lung carcinoma), and MCF7 (breast adenocarcinoma) cells at the bottom. E, Schematic indicates a proposed model of the mechanism by which FOXA2 induces two IMA major mucin genes MUC5AC and MUC5B in humans directly and indirectly through the induction of the mucus transcription factor SPDEF.

Close modal

FOXA2 induces the expression of both MUC5AC and MUC5B through the upstream enhancer region of MUC5AC

CRISPR/Cas9 technology, including CRISPRi, has enabled researchers to assess the functional roles of noncoding regions bound by transcription factors, beyond mere binding knowledge obtained by ChIP-seq (29, 33). To validate whether the FOXA2-binding regions described above function as gene-regulatory regions to induce the expression of MUC5AC and/or MUC5B, we perturbed the enhancer regions using CRISPRi, which uses dCas9-KRAB and single-guide RNA targeting the enhancer regions [sgRNA#1 for the upstream enhancer region of MUC5AC (hereafter, MUC5AC enhancer region) or sgRNA#2 for the intergenic enhancer region between MUC5AC and MUC5B (hereafter, MUC5B enhancer region)], and assessed whether FOXA2-mediated induction of MUC5AC and/or MUC5B is functionally affected (Fig. 6A). Importantly, FOXA2-mediated induction of MUC5AC was repressed by CRISPRi targeting the MUC5AC enhancer region (sgRNA#1) but not the MUC5B enhancer region (sgRNA#2) in H441 cells (Fig. 6B; Supplementary Table S10), indicating that the MUC5AC enhancer region but not the MUC5B enhancer region is required for FOXA2-mediated induction of MUC5AC in H441 cells. Notably, FOXA2-mediated induction of MUC5B was repressed by CRISPRi targeting both the MUC5AC enhancer region (sgRNA#1) and the MUC5B enhancer region (sgRNA#2) in H441 cells, indicating that both enhancer regions are essential for the expression of MUC5B (Fig. 6B; Supplementary Table S10), which is consistent in part with the results shown by Helling and colleagues (42) on the MUC5B-enhancer region using A549 cells. Of note, CRISPRi targeting both regions did not influence the expression of pro-mucous transcription factors SPDEF or FOXA3 (Supplementary Fig. S9), indicating the specificity of CRISPRi and direct regulation of MUC5AC and MUC5B by FOXA2 (not mediated by other pro-mucous transcription factors). These results suggest that the MUC5AC enhancer region bound by FOXA2 has a dual role to regulate the expression of two genes (MUC5AC and MUC5B) whereas the MUC5B enhancer region bound by FOXA2 has a single role to regulate the expression of only one gene MUC5B (Fig. 6C).

Figure 6.

FOXA2 induces the expression of both MUC5AC and MUC5B through the upstream enhancer regions of MUC5AC. A, Top, two enhancer regions at the upstream region (#1) of MUC5AC and at the upstream region (#2) of MUC5B (between MUC5AC and MUC5B) that are targeted by CRISPRi, in which dCas9 (deactivated Cas9)-KRAB repressor is recruited to a region (#1 or #2) selected by sgRNA sequences (green highlighted) located at presumed FOXA2 binding sites (yellow highlighted). Bottom, sgRNA target sequences (green) and adjacent PAM sequence (blue) and presumed FOXA2-binding sequences (yellow). B, TaqMan gene expression analysis indicates that a synthetic sgRNA targeting a region at the upstream region (#1) of MUC5AC significantly represses the FOXA2-mediated induction of MUC5AC and MUC5B in H441 KRAS-mutant lung adenocarcinoma cells that stably express dCas9-KRAB along with FOXA2. Control, empty vector. A synthetic sgRNA targeting an upstream region (#2) of MUC5B significantly represses the FOXA2-mediated induction of MUC5B but not that of MUC5AC. Nontargeted synthetic sgRNA was used as control for sgRNAs targeting the loci of MUC5AC or MUC5B. Results are expressed as mean ± SEM of three biological replicates for each group. Unpaired, two-tailed Student t test. C, Schematic indicates a proposed model of the mechanism by which FOXA2 induces two IMA major mucin genes MUC5AC and MUC5B in human lung cancer cells through the upstream region of MUC5AC (MUC5AC enhancer) and the intergenic region between MUC5AC and MUC5B (MUC5B enhancer).

Figure 6.

FOXA2 induces the expression of both MUC5AC and MUC5B through the upstream enhancer regions of MUC5AC. A, Top, two enhancer regions at the upstream region (#1) of MUC5AC and at the upstream region (#2) of MUC5B (between MUC5AC and MUC5B) that are targeted by CRISPRi, in which dCas9 (deactivated Cas9)-KRAB repressor is recruited to a region (#1 or #2) selected by sgRNA sequences (green highlighted) located at presumed FOXA2 binding sites (yellow highlighted). Bottom, sgRNA target sequences (green) and adjacent PAM sequence (blue) and presumed FOXA2-binding sequences (yellow). B, TaqMan gene expression analysis indicates that a synthetic sgRNA targeting a region at the upstream region (#1) of MUC5AC significantly represses the FOXA2-mediated induction of MUC5AC and MUC5B in H441 KRAS-mutant lung adenocarcinoma cells that stably express dCas9-KRAB along with FOXA2. Control, empty vector. A synthetic sgRNA targeting an upstream region (#2) of MUC5B significantly represses the FOXA2-mediated induction of MUC5B but not that of MUC5AC. Nontargeted synthetic sgRNA was used as control for sgRNAs targeting the loci of MUC5AC or MUC5B. Results are expressed as mean ± SEM of three biological replicates for each group. Unpaired, two-tailed Student t test. C, Schematic indicates a proposed model of the mechanism by which FOXA2 induces two IMA major mucin genes MUC5AC and MUC5B in human lung cancer cells through the upstream region of MUC5AC (MUC5AC enhancer) and the intergenic region between MUC5AC and MUC5B (MUC5B enhancer).

Close modal

Expression of mucous genes are primed by FOXA2 and amplified by mutant KRAS

Our current study demonstrates that FOXA2 alone induces the development of immature goblet (mucous) cells in the airways whereas coexpression of FOXA2 with mutant KRAS induces fully developed mucinous lung tumors in mice (Figs. 2 and 3; Supplementary Fig. S3), suggesting that FOXA2 is a pro-mucous transcription factor, especially in the presence of mutant KRAS. To understand the role of FOXA2 or mutant KRAS in mucous gene regulation associated with the in vivo mucous phenotypes, we assessed chromatin modification at the locus of MUC5AC in the presence of FOXA2 and/or mutant KRAS (KRASG12V) using BEAS-2B human-transformed bronchial epithelial cells. Consistent with the mouse results (Fig. 3), FOXA2 alone (Foxa2) moderately induced the mRNA expression of MUC5AC and MUC5B in the absence of mutant KRAS in BEAS-2B cells (Fig. 7A and B; Supplementary Fig. S10 and Supplementary Table S11). Mutant KRAS alone (KRASG12V) did not induce the expression of MUC5AC or MUC5B, which is also consistent with the mouse data (Fig. 3). Notably, coexpression of FOXA2 with mutant KRAS (KRASG12V;Foxa2) synergistically induced the expression of MUC5AC and MUC5B compared with that induced by FOXA2 alone, indicating that mutant KRAS amplifies the expression of MUC5AC and MUC5B primed by FOXA2. Importantly, H3K27ac (an enhancer mark of histone) was detected at the upstream locus of MUC5AC in the BEAS-2B cells that express FOXA2 but not mutant KRAS (Fig. 7C; Supplementary Fig. S11A–S11C and Supplementary Table S12), suggesting a novel role of FOXA2 as a pioneer factor that modifies chromatin by introducing H3K27ac at the locus of MUC5AC, which in turn allows mutant KRAS-induced transcriptional activators access to induce expression of MUC5AC (Fig. 7D).

Figure 7.

Mutant KRAS amplifies the expression of MUC5AC that is primed by FOXA2. A, Immunoblotting (IB) was performed using antibodies against FOXA2 or KRASG12V. BEAS-2B human–transformed bronchial epithelial cells were infected with a lentiviral vector carrying rat Foxa2 and/or a retroviral vector carrying human KRASG12V and cell extracts were used for immunoblotting. Controls, empty lentiviral and retroviral vectors. Shown are representative images from four independent experiments. B, TaqMan gene expression analysis indicates that the expression of MUC5AC and MUC5B is primed by FOXA2 and enhanced by a mutant KRAS (KRASG12V). BEAS-2B cells were infected with a lentiviral vector carrying rat Foxa2 and/or a retroviral vector carrying human KRASG12V and RNA was extracted for analysis. Results are expressed as mean ± SEM of four biological replicates for each group. Controls, empty lentiviral and retroviral vectors. A P value of <0.05 versus control was considered significant (Student t test). Gene expression was normalized by comparison with the constitutive expression of GAPDH (P values are from unpaired, two-tailed Student t tests). C, ChIP-seq using an antibody against H3K27ac (an enhancer mark of histone) was performed using BEAS-2B cells that ectopically expressed FOXA2 and/or KRASG12V as described previously in A. Shown is the genome browser demonstrating H3K27ac gain compared with control at the MUC5AC locus in BEAS-2B cells expressing FOXA2 or KRASG12V; FOXA2 but not KRASG12V alone. D, Schematic indicates a proposed model of the mechanism by which FOXA2 induces two IMA major mucin genes MUC5AC and MUC5B in human lung cancer cells by initiating H3K27ac enhancer marks at the upstream region of MUC5AC. (D, Created with BioRender.com.)

Figure 7.

Mutant KRAS amplifies the expression of MUC5AC that is primed by FOXA2. A, Immunoblotting (IB) was performed using antibodies against FOXA2 or KRASG12V. BEAS-2B human–transformed bronchial epithelial cells were infected with a lentiviral vector carrying rat Foxa2 and/or a retroviral vector carrying human KRASG12V and cell extracts were used for immunoblotting. Controls, empty lentiviral and retroviral vectors. Shown are representative images from four independent experiments. B, TaqMan gene expression analysis indicates that the expression of MUC5AC and MUC5B is primed by FOXA2 and enhanced by a mutant KRAS (KRASG12V). BEAS-2B cells were infected with a lentiviral vector carrying rat Foxa2 and/or a retroviral vector carrying human KRASG12V and RNA was extracted for analysis. Results are expressed as mean ± SEM of four biological replicates for each group. Controls, empty lentiviral and retroviral vectors. A P value of <0.05 versus control was considered significant (Student t test). Gene expression was normalized by comparison with the constitutive expression of GAPDH (P values are from unpaired, two-tailed Student t tests). C, ChIP-seq using an antibody against H3K27ac (an enhancer mark of histone) was performed using BEAS-2B cells that ectopically expressed FOXA2 and/or KRASG12V as described previously in A. Shown is the genome browser demonstrating H3K27ac gain compared with control at the MUC5AC locus in BEAS-2B cells expressing FOXA2 or KRASG12V; FOXA2 but not KRASG12V alone. D, Schematic indicates a proposed model of the mechanism by which FOXA2 induces two IMA major mucin genes MUC5AC and MUC5B in human lung cancer cells by initiating H3K27ac enhancer marks at the upstream region of MUC5AC. (D, Created with BioRender.com.)

Close modal

FOXA2 is highly expressed in human IMA

Next, to determine whether FOXA2 is associated with the mucous phenotype in human lung adenocarcinoma in vivo, we sought to determine by IHC whether FOXA2 is expressed in human IMA (Supplementary Table S5). As shown in Fig. 8A, FOXA2 was expressed in the nucleus of mucinous lung tumor cells in human IMA. The expression of FOXA2 was seen in 71% of human IMA cases (Fig. 8B; considered positive when FOXA2 expression is seen in over 50% of IMA tumor cells in each case), which is similar to a previous pathological study (100% positivity reported; ref. 43). This result suggests that FOXA2 expressed in mucinous tumor cells may drive the mucinous phenotype in human IMA. According to TCGA LUAD dataset (3), KRAS mutation did not affect the expression of FOXA2 compared with KRAS wild-type (Supplementary Fig. S12A–S12C); however, among the KRAS mutations/CDKN2A loss co-occurring genomic alteration group (44), expression of FOXA2 is significantly higher in IMA than non-IMA whereas the expression of NKX2–1, an anti-mucous transcription factor (10, 15), is significantly lower in IMA than in non-IMA (Fig. 8C). Of note, TCGA LUAD IMA cases with KRAS mutations are all associated with CDKN2A loss (3). Importantly, ectopic Nkx2–1 significantly suppressed the expression of FOXA2 in A549 cells (KRASG12S;CDKN2Adel; Fig. 8D; Supplementary Table S13). These results suggest that genetic or epigenetic loss of NKX2–1 in LUAD with KRAS mutations/CDKN2A loss induces the expression of FOXA2, which drives IMA.

Figure 8.

FOXA2 is highly expressed in human IMA of the lung. A, IHC was performed using human IMA and adjacent normal lung tissues. FOXA2 expressed in the nucleus was induced in mucinous tumor cells (right) compared with normal lung epithelial cells (left). MUC5AC and MUC5B expressed in cytoplasm and secreted in lumen were also induced in mucinous tumor cells (right) compared with normal lung epithelial cells (left). Alcian blue–staining detects mucins. Scale bar, 20 μm. B, Pie charts indicate the expression ratio of FOXA2, MUC5AC, or MUC5B in human IMA. FOXA2 was expressed in the majority of human IMA. Mixed indicates that the expression of FOXA2 in IMA was less than 50% (see Supplementary Table S5). C, Normalized and log2-transformed mRNA expression of FOXA2 (left) and NKX2–1 (right) in IMA or Non-IMA in TCGA LUAD with KRAS mutations/CDKN2A loss co-occurring genomic alteration group is shown. A P value of <0.05 was considered significant (unpaired, two-tailed Welch t test). D, TaqMan gene expression analysis was performed using RNAs from A549 cells infected with a lentiviral vector carrying rat Nkx2–1. Control, an empty lentiviral vector. Results are expressed as mean ± SEM of three biological replicates for each group. Gene expression was normalized by comparison with the constitutive expression of GAPDH. A P value of <0.05 was considered significant (unpaired, two-tailed Student t test).

Figure 8.

FOXA2 is highly expressed in human IMA of the lung. A, IHC was performed using human IMA and adjacent normal lung tissues. FOXA2 expressed in the nucleus was induced in mucinous tumor cells (right) compared with normal lung epithelial cells (left). MUC5AC and MUC5B expressed in cytoplasm and secreted in lumen were also induced in mucinous tumor cells (right) compared with normal lung epithelial cells (left). Alcian blue–staining detects mucins. Scale bar, 20 μm. B, Pie charts indicate the expression ratio of FOXA2, MUC5AC, or MUC5B in human IMA. FOXA2 was expressed in the majority of human IMA. Mixed indicates that the expression of FOXA2 in IMA was less than 50% (see Supplementary Table S5). C, Normalized and log2-transformed mRNA expression of FOXA2 (left) and NKX2–1 (right) in IMA or Non-IMA in TCGA LUAD with KRAS mutations/CDKN2A loss co-occurring genomic alteration group is shown. A P value of <0.05 was considered significant (unpaired, two-tailed Welch t test). D, TaqMan gene expression analysis was performed using RNAs from A549 cells infected with a lentiviral vector carrying rat Nkx2–1. Control, an empty lentiviral vector. Results are expressed as mean ± SEM of three biological replicates for each group. Gene expression was normalized by comparison with the constitutive expression of GAPDH. A P value of <0.05 was considered significant (unpaired, two-tailed Student t test).

Close modal

Because FOXA2 expression was observed in human lung cancer specimens (e.g., https://www.proteinatlas.org/ENSG00000125798-FOXA2/pathology/lung+cancer), the role of FOXA2 in lung cancer has been studied by multiple groups using in vitro cell culture and in vivo xenograft mouse models. However, the role of FOXA2 in genetically engineered mouse models (GEMM) that induce autochthonous lung tumors has not been well studied. Here, using a Tet-On “gain-of-function” system, we conditionally induced FOXA2 along with oncogenic mutant KRAS or mutant EGFR in lung epithelium in GEMMs and determined the role of FOXA2 in autochthonous lung tumors driven by mutant KRAS or mutant EGFR. Although ectopic expression of FOXA2 in human lung cancer cell lines that were derived from metastatic sites (H446, H1299 and H358 cells) inhibited tumorigenicity of the cells (18, 19), transgenic expression of FOXA2 in a GEMM promoted KRAS mutant-driven autochthonous primary lung tumors and suppressed EGFR mutant-driven autochthonous primary lung tumors in our study. These results suggest that FOXA2 functions as a tumor promoter or a tumor suppressor depending on the presence of different driver oncogenes (e.g., KRAS mutant vs. EGFR mutant) and tumor locations (e.g., primary or metastatic), which is reminiscent of the role of NKX2–1 as a context-dependent lung tumor promoter or tumor suppressor (45). A sequential transgenic mouse model that induces FOXA2 before or after mutant KRAS or mutant EGFR expression will further delineate how FOXA2 influences the initiation and/or promotion of KRAS-mutant or EGFR-mutant lung tumorigenesis. The use of knock-in mice carrying alveolar type II cell-specific Sftpc-Cre or club cell-specific Scgb1a1-Cre other than our rat Scgb1a1-rtTA transgenic mouse model targeting broad lung epithelial cells may further elucidate the precise origin of mucinous tumor cells influenced by FOXA2 (46). Notably, Camolotto and colleagues (47) reported using a GEMM “loss-of-function” model that ubiquitous PGK-Cre–mediated codeletion of Foxa1 and Foxa2 suppressed growth of autochthonous KRAS-mutant primary lung tumors, which is in part consistent with our “gain-of-function” results that FOXA2 promotes the growth of KRAS mutant-lung tumors (Fig. 1). Of note, FOXA1 also significantly induced the expression of SPDEF, MUC5AC, and MUC5B in H441 cells. Interestingly, FOXA1 was a major inducer of MUC5B whereas FOXA2 was a major inducer of MUC5AC (Supplementary Fig. S13 and Supplementary Table S14), suggesting that coinduction of FOXA1 and FOXA2 along with mutant KRAS may further drive IMA of the lung.

FOXA2 has also been shown to be a pro-mucous transcription factor or anti-mucous transcription factor depending on the context. Conditional codeletion of Foxa1 and Foxa2 in developing (nontumorigenic) mouse gastrointestinal (GI) tract reduced goblet (mucous) cell differentiation accompanied by the decreased expression of Muc2 and the increased expression of Muc5ac (48); however, conditional deletion of Foxa2 in asthmatic (nontumorigenic) mouse lung induced goblet (mucous) differentiation accompanied by the increased expression of Muc5ac (25). These previous studies indicate that FOXA2 suppresses goblet (mucous) cell differentiation in part by repressing the expression of MUC5AC in non-tumorigenic lung. However, in vitro loss-of-function studies by us (Supplementary Fig. S6A and S6B and Supplementary Table S8) and others (42) in A549 human lung carcinoma cells (KRASG12S;CDKN2Adel) indicated that knockdown of FOXA2 reduced the expression of mucous genes, including SPDEF, MUC5AC, and MUC5B. Consistent with these in vitro loss-of-function studies, our present in vivo and in vitro gain-of-function study by ectopic expression of Foxa2 in the presence of mutant Kras in the adult GEMM, H441 human lung papillary adenocarcinoma cells (KRASG12V; TP53R158L) and BEAS-2B human transformed bronchial epithelial cells (transformed by adenovirus-12 SV40 hybrid virus) indicated that FOXA2 in the presence of mutant KRAS induced tumorigenic goblet (mucous) cells accompanied by the expression of IMA-related genes, including MUC5AC and MUC5B, in mice and in human cells (Figs. 2, 3, 58). For therapeutic interest for IMA, IL23A is a potential target gene among the FOXA2-induced IMA genes because the IL23A–IL23R pathway is therapeutically targetable by antibody (49). Further studies are required to elucidate the role of FOXA2 in a different context in normal and diseased lungs (e.g., lung cancer with different KRAS or EGFR mutations, NRG fusions or nontumorigenic chronic diseased lungs; Supplementary Fig. S12A–S12D).

Because FOXA1 and FOXA2 can bind both nucleosome-bound and nucleosome-free DNA targets, FOXA1 and FOXA2 are considered to be pioneer transcription factors that allow the subsequent binding of other transcription factors at nearby sites (50). In our study, the mRNA expression of MUC5AC and MUC5B primed by FOXA2 was further induced by mutant KRAS in BEAS-2B bronchial epithelial cells; however, mutant KRAS alone did not prime the expression (Fig. 7B). Our data demonstrate that ectopic expression of FOXA2 alone is sufficient to introduce the H3K27ac enhancer mark at the locus of MUC5AC (Fig. 7C), suggesting that FOXA2 not only binds both nucleosome-bound and nucleosome-free DNA targets but also recruits chromatin modifiers to mark gene-regulatory enhancers at the target loci. Most of the studies investigating the role of FOXA1 and FOXA2 as pioneer transcription factors have been performed using cells derived from liver, breast, and prostate. FOXA2 may associate differently with chromatin in lung-originating cells; however, further studies looking at different histone modifications at loci of genes expressed in lung are required to understand the precise role of FOXA2 as a pioneer transcription factor and/or a recruiter for chromatin modifiers in the lung.

The association of FOXA2 expression in primary lung NSCLC with patients’ survival has been assessed by IHC using paraffin-embedded NSCLC samples (17) and analysis of mRNA-seq data from TCGA lung adenocarcinoma datasets (20). Although previous reports using lung cancer cell lines suggest that FOXA2 is a tumor suppressor especially in the context of metastasis (18–20), the mRNA or protein expression of FOXA2 in primary lung tumors does not significantly correlate with poor survival of such patients with lung cancer (17, 20). Our present data suggest that FOXA2 is a context-dependent tumor influencer driven by distinct oncogenes and tumor suppressors (e.g., CDKN2A, TP53 and/or STK11/LKB1; ref. 44) just as NKX2–1 also functions in a different manner depending on the presence of such oncogenes (e.g., mutant KRAS or mutant EGFR; ref. 45). Assessing the expression of FOXA2 based on different driver oncogenes and tumor suppressors is required to understand whether FOXA2 influences survival in lung cancer patients. In addition, because FOXA2 suppresses itself in an auto-inhibitory fashion (Supplementary Fig. S4B and S4C), assessing FOXA2 at the protein level might be more accurate than assessing FOXA2 at the mRNA level to discern the role of FOXA2 in patient survival.

In summary, using transgenic model mice that develop autochthonous lung tumors, we identified that FOXA2 suppresses EGFR-mutant lung tumors whereas FOXA2 promotes KRAS-mutant lung tumors. Notably, FOXA2 in the presence of mutant KRAS induces mucinous lung tumors in vivo. The expression of FOXA2 protein is also significantly associated with human IMA. In human lung adenocarcinoma cells, FOXA2 bound to the two enhancer regions at the MUC5 (MUC5AC and MUC5B) locus and induced the expression of both MUC5AC and MUC5B in vitro. Notably, FOXA2 but not mutant KRAS introduced the H3K27ac enhancer histone mark to the upstream enhancer region of MUC5AC, which indicates that FOXA2 primes the expression of both MUC5AC and MUC5B through this enhancer (Supplementary Fig. S14). Our present study along with the previous studies by others indicate that FOXA2 promotes KRAS-mutant primary lung tumors and suppresses EGFR-mutant primary lung tumors; and may inhibit lung tumor metastasis.

No disclosures were reported.

K. Tomoshige: Resources, formal analysis, funding acquisition, validation, investigation, visualization, methodology. W.D. Stuart: Formal analysis, validation, investigation, visualization, methodology, writing–review and editing. I.M. Fink-Baldauf: Validation, investigation, methodology. M. Ito: Resources, validation, writing–review and editing. T. Tsuchiya: Resources. T. Nagayasu: Resources. T. Yamatsuji: Resources. M. Okada: Resources. T. Fukazawa: Resources. M. Guo: Formal analysis, validation, investigation, visualization, methodology, writing–original draft. Y. Maeda: Conceptualization, formal analysis, supervision, funding acquisition, validation, investigation, methodology, writing–original draft, project administration.

This work was supported by NIH grants (U01HL134745 and R01CA240317), Cincinnati Children's Hospital Medical Center (Trustee Award grant, CF-RDP Pilot and Feasibility grant, and GAP funding to Y. Maeda), and the Rotary Foundation Global grant (to K. Tomoshige). The authors thank Rafael Rosell, Eric Snyder, Mari Mino-Kenudson, Jeffrey Whitsett, Gang Chen, Thomas Korfhagen, Kristin Hudock, and Mary Durbin for materials and discussions.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Cancer Research Online (http://cancerres.aacrjournals.org/).

1.
Thai
AA
,
Solomon
BJ
,
Sequist
LV
,
Gainor
JF
,
Heist
RS
.
Lung cancer
.
Lancet
2021
;
398
:
535
54
.
2.
Cancer Genome Atlas Research Network
.
Comprehensive genomic characterization of squamous cell lung cancers
.
Nature
2012
;
489
:
519
25
.
3.
Cancer Genome Atlas Research Network
.
Comprehensive molecular profiling of lung adenocarcinoma
.
Nature
2014
;
511
:
543
50
.
4.
George
J
,
Lim
JS
,
Jang
SJ
,
Cun
Y
,
Ozretić
L
,
Kong
G
, et al
.
Comprehensive genomic profiles of small-cell lung cancer
.
Nature
2015
;
524
:
47
53
.
5.
Tanaka
H
,
Yanagisawa
K
,
Shinjo
K
,
Taguchi
A
,
Maeno
K
,
Tomida
S
, et al
.
Lineage-specific dependency of lung adenocarcinomas on the lung development regulator TTF-1
.
Cancer Res
2007
;
67
:
6007
11
.
6.
Kendall
J
,
Liu
Q
,
Bakleh
A
,
Krasnitz
A
,
Nguyen
KC
,
Lakshmi
B
, et al
.
Oncogenic cooperation and coamplification of developmental transcription factor genes in lung cancer
.
Proc Natl Acad Sci U S A
2007
;
104
:
16663
8
.
7.
Weir
BA
,
Woo
MS
,
Getz
G
,
Perner
S
,
Ding
L
,
Beroukhim
R
, et al
.
Characterizing the cancer genome in lung adenocarcinoma
.
Nature
2007
;
450
:
893
8
.
8.
Kwei
KA
,
Kim
YH
,
Girard
L
,
Kao
J
,
Pacyna-Gengelbach
M
,
Salari
K
, et al
.
Genomic profiling identifies TITF1 as a lineage-specific oncogene amplified in lung cancer
.
Oncogene
2008
;
27
:
3635
40
.
9.
Bass
AJ
,
Watanabe
H
,
Mermel
CH
,
Yu
S
,
Perner
S
,
Verhaak
RG
, et al
.
SOX2 is an amplified lineage-survival oncogene in lung and esophageal squamous cell carcinomas
.
Nat Genet
2009
;
41
:
1238
42
.
10.
Maeda
Y
,
Tsuchiya
T
,
Hao
H
,
Tompkins
DH
,
Xu
Y
,
Mucenski
ML
, et al
.
Kras(G12D) and Nkx2–1 haploinsufficiency induce mucinous adenocarcinoma of the lung
.
J Clin Invest
2012
;
122
:
4388
400
.
11.
Mukhopadhyay
A
,
Berrett
KC
,
Kc
U
,
Clair
PM
,
Pop
SM
,
Carr
SR
, et al
.
Sox2 cooperates with Lkb1 loss in a mouse model of squamous cell lung cancer
.
Cell Rep
2014
;
8
:
40
9
.
12.
Ferone
G
,
Song
JY
,
Sutherland
KD
,
Bhaskaran
R
,
Monkhorst
K
,
Lambooij
JP
, et al
.
SOX2 is the determining oncogenic switch in promoting lung squamous cell carcinoma from different cells of origin
.
Cancer Cell
2016
;
30
:
519
32
.
13.
Holzinger
A
,
Dingle
S
,
Bejarano
PA
,
Miller
MA
,
Weaver
TE
,
DiLauro
R
, et al
.
Monoclonal antibody to thyroid transcription factor-1: production, characterization, and usefulness in tumor diagnosis
.
Hybridoma
1996
;
15
:
49
53
.
14.
Travis
WD
,
Brambilla
E
,
Noguchi
M
,
Nicholson
AG
,
Geisinger
KR
,
Yatabe
Y
, et al
.
International association for the study of lung cancer/American thoracic society/european respiratory society international multidisciplinary classification of lung adenocarcinoma
.
J Thorac Oncol
2011
;
6
:
244
85
.
15.
Snyder
EL
,
Watanabe
H
,
Magendantz
M
,
Hoersch
S
,
Chen
TA
,
Wang
DG
, et al
.
Nkx2–1 represses a latent gastric differentiation program in lung adenocarcinoma
.
Mol Cell
2013
;
50
:
185
99
.
16.
Zhou
L
,
Lim
L
,
Costa
RH
,
Whitsett
JA
.
Thyroid transcription factor-1, hepatocyte nuclear factor-3beta, surfactant protein B, C, and Clara cell secretory protein in developing mouse lung
.
J Histochem Cytochem
1996
;
44
:
1183
93
.
17.
Basseres
DS
,
D'Alò
F
,
Yeap
BY
,
Löwenberg
EC
,
Gonzalez
DA
,
Yasuda
H
, et al
.
Frequent downregulation of the transcription factor Foxa2 in lung cancer through epigenetic silencing
.
Lung Cancer
2012
;
77
:
31
7
.
18.
Tang
Y
,
Shu
G
,
Yuan
X
,
Jing
N
,
Song
J
.
FOXA2 functions as a suppressor of tumor metastasis by inhibition of epithelial-to-mesenchymal transition in human lung cancers
.
Cell Res
2011
;
21
:
316
26
.
19.
Halmos
B
,
Bassères
DS
,
Monti
S
,
D'Aló
F
,
Dayaram
T
,
Ferenczi
K
, et al
.
A transcriptional profiling study of CCAAT/enhancer–binding protein targets identifies hepatocyte nuclear factor 3 beta as a novel tumor suppressor in lung cancer
.
Cancer Res
2004
;
64
:
4137
47
.
20.
Li
CM
,
Gocheva
V
,
Oudin
MJ
,
Bhutkar
A
,
Wang
SY
,
Date
SR
, et al
.
Foxa2 and Cdx2 cooperate with Nkx2–1 to inhibit lung adenocarcinoma metastasis
.
Genes Dev
2015
;
29
:
1850
62
.
21.
Wang
B
,
Liu
G
,
Ding
L
,
Zhao
J
,
Lu
Y
.
FOXA2 promotes the proliferation, migration and invasion, and epithelial mesenchymal transition in colon cancer
.
Exp Ther Med
2018
;
16
:
133
40
.
22.
Milan
M
,
Balestrieri
C
,
Alfarano
G
,
Polletti
S
,
Prosperini
E
,
Spaggiari
P
, et al
.
FOXA2 controls the cis-regulatory networks of pancreatic cancer cells in a differentiation grade-specific manner
.
EMBO J
2019
;
38
:
e102161
.
23.
Connelly
ZM
,
Jin
R
,
Zhang
J
,
Yang
S
,
Cheng
S
,
Shi
M
, et al
.
FOXA2 promotes prostate cancer growth in the bone
.
Am J Transl Res
2020
;
12
:
5619
29
.
24.
Chen
Z
,
Xiao
Q
,
Shen
Y
,
Xue
C
.
FOXA2 promotes esophageal cancer migration and metastasis by activating CXCR4 expression
.
Biochem Biophys Res Commun
2022
;
625
:
16
22
.
25.
Chen
G
,
Wan
H
,
Luo
F
,
Zhang
L
,
Xu
Y
,
Lewkowich
I
, et al
.
Foxa2 programs Th2 cell–mediated innate immunity in the developing lung
.
J Immunol
2010
;
184
:
6133
41
.
26.
Fisher
GH
,
Wellen
SL
,
Klimstra
D
,
Lenczowski
JM
,
Tichelaar
JW
,
Lizak
MJ
, et al
.
Induction and apoptotic regression of lung adenocarcinomas by regulation of a K-Ras transgene in the presence and absence of tumor-suppressor genes
.
Genes Dev
2001
;
15
:
3249
62
.
27.
Politi
K
,
Zakowski
MF
,
Fan
PD
,
Schonfeld
EA
,
Pao
W
,
Varmus
HE
.
Lung adenocarcinomas induced in mice by mutant EGF receptors found in human lung cancers respond to a tyrosine kinase inhibitor or to downregulation of the receptors
.
Genes Dev
2006
;
20
:
1496
510
.
28.
Thakore
PI
,
D'Ippolito
AM
,
Song
L
,
Safi
A
,
Shivakumar
NK
,
Kabadi
AM
, et al
.
Highly specific epigenome editing by CRISPR–Cas9 repressors for silencing of distal regulatory elements
.
Nat Methods
2015
;
12
:
1143
9
.
29.
Stuart
WD
,
Fink-Baldauf
IM
,
Tomoshige
K
,
Guo
M
,
Maeda
Y
.
CRISPRi-mediated functional analysis of NKX2–1-binding sites in the lung
.
Commun Biol
2021
;
4
:
568
.
30.
ENCODE Project Consortium
.
An integrated encyclopedia of DNA elements in the human genome
.
Nature
2012
;
489
:
57
74
.
31.
Gertz
J
,
Savic
D
,
Varley
KE
,
Partridge
EC
,
Safi
A
,
Jain
P
, et al
.
Distinct properties of cell-type–specific and shared transcription factor binding sites
.
Mol Cell
2013
;
52
:
25
36
.
32.
Fletcher
MN
,
Castro
MA
,
Wang
X
,
de Santiago
I
,
O'Reilly
M
,
Chin
SF
, et al
.
Master regulators of FGFR2 signalling and breast cancer risk
.
Nat Commun
2013
;
4
:
2464
.
33.
Guo
M
,
Tomoshige
K
,
Meister
M
,
Muley
T
,
Fukazawa
T
,
Tsuchiya
T
, et al
.
Gene signature driving invasive mucinous adenocarcinoma of the lung
.
EMBO Mol Med
2017
;
9
:
462
81
.
34.
Bauer
AK
,
Umer
M
,
Richardson
VL
,
Cumpian
AM
,
Harder
AQ
,
Khosravi
N
, et al
.
Requirement for MUC5AC in KRAS-dependent lung carcinogenesis
.
JCI Insight
2018
;
3
:
e120941
.
35.
Li
S
,
Wang
Y
,
Zhang
Y
,
Lu
MM
,
DeMayo
FJ
,
Dekker
JD
, et al
.
Foxp1/4 control epithelial cell fate during lung development and regeneration through regulation of anterior gradient 2
.
Development
2012
;
139
:
2500
9
.
36.
Fessart
D
,
Domblides
C
,
Avril
T
,
Eriksson
LA
,
Begueret
H
,
Pineau
R
, et al
.
Secretion of protein disulphide isomerase AGR2 confers tumorigenic properties
.
Elife
2016
;
5
:
e13887
.
37.
Chen
G
,
Korfhagen
TR
,
Xu
Y
,
Kitzmiller
J
,
Wert
SE
,
Maeda
Y
, et al
.
SPDEF is required for mouse pulmonary goblet cell differentiation and regulates a network of genes associated with mucus production
.
J Clin Invest
2009
;
119
:
2914
24
.
38.
Fossum
SL
,
Mutolo
MJ
,
Yang
R
,
Dang
H
,
O'Neal
WK
,
Knowles
MR
, et al
.
Ets homologous factor regulates pathways controlling response to injury in airway epithelial cells
.
Nucleic Acids Res
2014
;
42
:
13588
98
.
39.
Evans
CM
,
Raclawska
DS
,
Ttofali
F
,
Liptzin
DR
,
Fletcher
AA
,
Harper
DN
, et al
.
The polymeric mucin Muc5ac is required for allergic airway hyperreactivity
.
Nat Commun
2015
;
6
:
6281
.
40.
Roy
MG
,
Livraghi-Butrico
A
,
Fletcher
AA
,
McElwee
MM
,
Evans
SE
,
Boerner
RM
, et al
.
Muc5b is required for airway defence
.
Nature
2014
;
505
:
412
6
.
41.
Park
KS
,
Korfhagen
TR
,
Bruno
MD
,
Kitzmiller
JA
,
Wan
H
,
Wert
SE
, et al
.
SPDEF regulates goblet cell hyperplasia in the airway epithelium
.
J Clin Invest
2007
;
117
:
978
88
.
42.
Helling
BA
,
Gerber
AN
,
Kadiyala
V
,
Sasse
SK
,
Pedersen
BS
,
Sparks
L
, et al
.
Regulation of MUC5B expression in idiopathic pulmonary fibrosis
.
Am J Respir Cell Mol Biol
2017
;
57
:
91
99
.
43.
Kishikawa
S
,
Hayashi
T
,
Saito
T
,
Takamochi
K
,
Kohsaka
S
,
Sano
K
, et al
.
Diffuse expression of MUC6 defines a distinct clinicopathological subset of pulmonary invasive mucinous adenocarcinoma
.
Mod Pathol
2021
;
34
:
786
97
.
44.
Skoulidis
F
,
Byers
LA
,
Diao
L
,
Papadimitrakopoulou
VA
,
Tong
P
,
Izzo
J
, et al
.
Co-occurring genomic alterations define major subsets of KRAS-mutant lung adenocarcinoma with distinct biology, immune profiles, and therapeutic vulnerabilities
.
Cancer Discov
2015
;
5
:
860
77
.
45.
Yamaguchi
T
,
Hosono
Y
,
Yanagisawa
K
,
Takahashi
T
.
NKX2–1/TTF-1: an enigmatic oncogene that functions as a double-edged sword for cancer cell survival and progression
.
Cancer Cell
2013
;
23
:
718
23
.
46.
Ferone
G
,
Lee
MC
,
Sage
J
,
Berns
A
.
Cells of origin of lung cancers: lessons from mouse studies
.
Genes Dev
2020
;
34
:
1017
32
.
47.
Camolotto
SA
,
Pattabiraman
S
,
Mosbruger
TL
,
Jones
A
,
Belova
VK
,
Orstad
G
, et al
.
FoxA1 and FoxA2 drive gastric differentiation and suppress squamous identity in NKX2–1-negative lung cancer
.
Elife
2018
;
7
:
e38579
.
48.
Ye
DZ
,
Kaestner
KH
.
Foxa1 and Foxa2 control the differentiation of goblet and enteroendocrine L- and D cells in mice
.
Gastroenterology
2009
;
137
:
2052
62
.
49.
Hawkes
JE
,
Yan
BY
,
Chan
TC
,
Krueger
JG
.
Discovery of the IL23/IL17 signaling pathway and the treatment of psoriasis
.
J Immunol
2018
;
201
:
1605
13
.
50.
Golson
ML
,
Kaestner
KH
.
Fox transcription factors: from development to disease
.
Development
2016
;
143
:
4558
70
.