Dual airway and alveolar contributions to adult lung homeostasis and carcinogenesis

Lung adenocarcinoma (LUAD) and chronic lung diseases caused by smoking and environmental noxious agents are the deadliest diseases worldwide, sharing a partially charted pathobiology of dysfunctional alveolar repair. Here we sought to identify the respiratory epithelial dynamics and molecular signatures participating in adult lung maintenance and chemical carcinogenesis. We employed novel mouse models of respiratory epithelial marking and ablation, a battery of pulmonary toxins and carcinogens, experimental protocols of carcinogen-induced LUAD, tobacco carcinogen-induced LUAD cell lines, and human transcriptomic data and identified a prominent involvement of airway molecular programs in alveolar maintenance and carcinogen-induced LUAD. The airway-specific transcriptomic signature was redistributed to the alveoli after toxic and carcinogenic insults and resulted in marked contributions of airway-labeled cells to injury-recovered alveoli and LUAD. Airway cells maintained Kras mutations and therefore possibly contributed to lung cancer initiation, while LUAD were spatially linked to neighboring airways. Transcriptomic profiling of carcinogen-induced murine and human LUAD revealed enrichment in airway signatures, while ablation of airway cells distorted alveolar structure and function and protected mice from LUAD development. Collectively, these results indicate that airway cells and/or transcriptomic signatures are essential for alveolar maintenance and LUAD development.


INTRODUCTION
Chronic lung diseases present tremendous health burdens attributed to dysfunctional alveolar repair [1]. Lung adenocarcinoma (LUAD), the leading cancer killer worldwide, is mainly caused by chemical carcinogens of tobacco smoke that induce mutations of the Kirsten rat sarcoma viral oncogene homologue (KRAS) in distal pulmonary cells [2][3][4]. The identification of the cellular and transcriptomic events that underlie lung regeneration and carcinogenesis is extremely important, since epithelial developmental pathways are intimately related with oncogenic signaling to jointly regulating stemness and drug resistance [5,6]. To this end, lineage-specific genes encoding epithelial proteins that support the physiological functions of the lungs were recently shown to suffer non-coding insertions and deletions in LUAD, lending further support to the longstanding notion that epithelial cells that express lung-restricted proteins are the cellular sources of LUAD [7].
We combined accurate genetic marking of airway and alveolar epithelial cells with different insults to the adult lung to integrally assess the signatures that contribute to alveolar maintenance, repair, and carcinogenesis. This approach was adopted to identify the molecular signature of regenerating and malignant cells, which was recently shown to be linked with mutation processes [7]. To achieve our goals, we adapted multi-hit chemical carcinogen exposure protocols to the murine C57BL/6 strain that is resistant to chemical tumor induction [16,17], and corroborated the findings with the FVB strain that is susceptible to single-hit carcinogenesis [15,18]. We show that in the adult mouse, aging-, pneumotoxin-and carcinogen-insults result in a marked activation of airway transcriptomic 4 programs across the alveolar parenchyma, contributing to alveolar repair and carcinogenesis. We determined that, early after chemical carcinogen exposure, airway cells preferentially suffer and sustain Kras oncogenic mutations compared with alveolar cells; that LUAD are spatially associated with neighboring airways; and that ablation of airway cells hinders alveolar maintenance and carcinogenesis in mice. The data indicate remarkable plasticity of adult lung epithelial cells and suggest a central role for airway cells and transcriptional signature sin lung regeneration and cancer.

Accurate genetic marking of airway and alveolar cells.
For genetic marking, we crossed a Cre-reporter strain that features Cre-mediated permanent switch of somatic cells from membranous tdTomato (mT) to GFP (mG) fluorescence (mT/mG) to six different Cre-driver strains, including a novel airway-specific Cre-driver (Scgb1a1.Cre, Sftpc.Cre, Lyz2.Cre, Sox2.Cre, Vav.Cre, and Nes.Cre; all C57BL/6 background; [14, [19][20][21][22][23][24]). In double heterozygote offspring at six postnatal weeks (i.e., after lung development is complete; [8,14]), mG+ labeling was completely and exclusively confined to the airway epithelium of mT/mG;Scgb1a1.Cre mice, promiscuous to airway and alveolar epithelia of mT/mG;Sftpc.Cre mice, partial and exclusive to the alveolar epithelium of mT/mG;Lyz2.Cre mice, and not informative in the remaining intercrosses ( Figure 1A and Figure S1). Co-localization of mG+ labeling with lineage marker proteins revealed complete mG+ labeling of all AEC but not of alveolar cells in mT/mG;Scgb1a1.Cre mice, of most AEC and all ATII cells in mT/mG;Sftpc.Cre mice, and of some ATII cells and all alveolar macrophages in mT/mG;Lyz2.Cre mice ( Figure 1A and Figures S1B and S1C). Lung flow cytometry of six-week-old mT/mG;Scgb1a1.Cre, mT/mG;Sftpc.Cre, and mT/mG;Lyz2.Cre mice estimated the proportions of mG+ marked cells in concordance to microscopy ( Figure 1A and Figures S1A and S1D). Thus mT/mG;Scgb1a1.Cre and 5 mT/mG;Lyz2.Cre mice display 100% distinct labeling of AEC versus ATII cells and AMΦ at conclusion of development.

Airway and alveolar signatures in chemical-induced lung adenocarcinoma
We reproducibly induced LUAD in the above genetically-marked mice using repetitive exposures to the tobacco-contained carcinogens urethane (ethyl carbamate, EC; stand-alone mutagen and tumor promoter) or 3-methylcholanthrene followed by butylated hydroxytoluene (MCA/BHT; a two-hit mutagen/tumor promoter regimen; Figure S2A). In both models, preneoplastic (airway epithelial hyperplasia, atypical alveolar hyperplasia) and neoplastic (adenoma and LUAD) lesions [25] were both airway-and alveolar-located and hence inconclusive on tumor origins ( Figure 1B).
Interestingly, both airway-and alveolar-located hyperplasias and tumors of mT/mG;Scgb1a1.Cre mice were partially mG+, of mT/mG;Sftpc.Cre ubiquitously and non-informatively mG+, and of mT/mG;Lyz2.Cre mice either mG+ or mG-( Figure 1C and Figure S2B-S2E). Immunostaining revealed that genetically-marked Scgb1a1+, Sftpc+, and Lyz2+ tumor cells were CCSP-TUBA1A-SFTPC+LYZ2± ( Figure 1D). We further tested these observations using a one-hit model in which a single exposure of susceptible FVB mice to urethane causes LUAD featuring the full-blown mutation spectrum of the human disease [15,18]. For this, we backcrossed mT/mG, Scgb1a1.Cre, Sftpc.Cre and Lyz2.Cre strains >F12 to the FVB background, set up all the relevant intercrosses, and subjected double heterozygote offspring to a single urethane exposure. Examination of neoplastic lesions at six months after carcinogen again revealed mixed contributions of both airway and alveolar cells/signatures to LUAD ( Figure S3A-S3C). Baseline genetic marking at six postnatal weeks was similar to that observed in the respective genotypes in the C57BL/6 strain (mG+ marking in Figure S3D), as was the expression pattern of specific lineage markers ( Figure S3D), corroborating ubiquitous marking of all CCSP+ AEC in mT/mG;Scgb1a1.Cre mice, of both CCSP+ AEC and SFTPC+ ATII cells in mT/mG;Sftpc.Cre mice, and of some SFTPC+ ATII cells in mT/mG;Lyz2.Cre mice. Collectively, these data support that chemical-induced LUAD carry both airway and alveolar signatures and suggest that tumor cells can originate from Scgb1a1+ airway 6 cells that up-regulate SFTPC with or without LYZ2 during carcinogenesis, from Scgb1a1+Sftpc+ airway cells that feature CCSP loss with/without LYZ2 gain, or from Sftpc+ alveolar cells that transiently express CCSP with/without LYZ2.

Airway cells sustain Kras mutations
We next determined the lung lineage in which LUAD driver mutations are inflicted. Since urethaneinduced LUAD of mice exclusively harbor Kras Q61R mutations [15], we sought these at early timepoints after urethane. For this, FVB mT/mG;Scgb1a1.Cre and mT/mG;Lyz2.Cre mice received urethane, lungs were harvested one and two weeks later, and digital droplet PCR was performed with probes targeting mT and Kras Q61R sequences. Interestingly, mG+;Kras Q61R cells in the lungs of mT/mG;Scgb1a1.Cre mice (i.e. Kras Q61R -mutant AEC) survived and increased in number, while mG+;Kras Q61R cells in the lungs of mT/mG;Lyz2.Cre mice (i.e. Kras Q61R -mutant ATII cells) did not persist over time (Figure 2A). Supporting the importance of AEC in LUAD development, threedimensional reconstruction of chemical carcinogen-inflicted lungs of FVB mice using highresolution micro-computed tomography (μCT) revealed that most lung tumors were spatially linked with neighboring airways, either sprouting from or even contained in bronchi ( Figure 2B). These results support a significant role of airway cells in chemical-induced LUAD of mice.

Dissemination of airway signatures across the tumor-initiated lung
Since airborne carcinogens act globally on the respiratory field, we examined non-neoplastic alveolar areas of carcinogen-treated mT/mG;Scgb1a1.Cre, mT/mG;Sftpc.Cre, and mT/mG;Lyz2.Cre mice. Dissemination of the airway transcriptomic signature was evident by the markedly increased mG+ cell numbers in the alveoli of carcinogen-treated mT/mG;Scgb1a1.Cre mice compared with saline-treated controls ( Figure 3A and Figure S4A and S4B). Immunostaining revealed that these Scgb1a1+ marked cells were CCSP+TUBA1A-SFTPC-when located near airways and CCSP-TUBA1A-SFTPC+ in alveoli and tumors ( Figure 3B-3D and Figure S4C). Expansion of Scgb1a1+ marked cells after urethane treatment was also documented using bioluminescent imaging of double 7 heterozygote offspring of R26.Luc [26] intercrosses with Scgb1a1.Cre mice, yielding a strain featuring light emission from Scgb1a1+ cells ( Figure S4D and S4E). In addition, co-staining of human LUAD [27] for SFTPC, CCSP, and KRT5 showed co-localization of SFTPC with KRT5 but not with CCSP ( Figure 3E-3G). These results indicate that dynamic changes in alveolar cell composition and/or gene expression occur globally during field cancerization by tobacco carcinogens. Altogether, the findings suggest that some alveolar cells are recycled by Scgb1a1+ cells or transiently express CCSP during carcinogenesis, resulting in expansion of the Scgb1a1+ signature across the lungs. Moreover, that human and murine LUAD carry mixed epithelial signatures although their location and protein expression suggests an alveolar origin [18,[28][29][30].

Scgb1a1+ marked cells in lung injury and repair
We next examined the dynamics of epithelial signatures during aging, injury, and repair. While mG+ cell abundance in the lungs of aging mT/mG;Sftpc.Cre and mT/mG;Lyz2.Cre mice did not change, Scgb1a1+ marked cells progressively increased in the alveoli of aging mT/mG;Scgb1a1.Cre mice, and these cells expressed SFTPC ( Figure 4A). Bleomycin treatment, which depletes ATII cells [31], accelerated the accumulation of Scgb1a1+ marked cells in the alveoli and in urethanetriggered LUAD ( Figure 4B and Figure S5A and S5B). Alveolar Scgb1a1+ cells also increased in response to perinatal hyperoxia that damages forming alveoli, as well as to naphthalene treatment that kills AEC ( Figure 4C and 4D) [31]. However, no SFTPC-expressing or Lyz2+ marked AEC were found in naphthalene-treated mT/mG;Scgb1a1.Cre or mT/mG;Lyz2.Cre mice, respectively ( Figure 4D and Figure S5C and S5D). Hence de novo Scgb1a1+ signatures appear in the alveolar regions of the aging and injured adult mouse lung, a fact that can be explained by peripheral lung migration of AEC or, as suggested by a previous study [32], by transient up-regulation of CCSP expression by regenerating alveolar cells. In addition, ATII-restricted transcriptomic signatures are not observed in the airways after injury, in line with previous work [14].
Airway and alveolar contributions to alveolar maintenance and adenocarcinoma in the adult lung 8 To test the role of Scgb1a1+, Sftpc+, and Lyz2+ marked cells in alveolar homeostasis and carcinogenesis, we ablated them by crossing Scgb1a1.Cre, Sftpc.Cre, and Lyz2.Cre mice to Dta mice expressing Diphtheria toxin in somatic cells upon Cre-mediated recombination [33]. Triple transgenic mT/mG;Driver.Cre;Dta intercrosses were also generated to evaluate ablation efficiency. However, Scgb1a1.Cre;Dta mice displayed widened airway caliper, enlarged alveoli, and inflammatory interalveolar septal destruction evident by increased mean linear intercept, bronchoalveolar lavage of CD45+CD11b+ cells, and static compliance ( Figure 5B and 5C), mimicking human chronic obstructive pulmonary disease [1]. Finally, we exposed control and ablated mice to ten consecutive weekly urethane exposures. All mice survived six months into carcinogen treatment, and Scgb1a1.Cre;Dta and Lyz2.Cre;Dta mice were equally protected from LUAD development compared with controls ( Figure 5D). Taken together, these results show that 9 Scgb1a1+ marked cells maintain postnatal alveolar structure and function, and, together with Lyz2+ marked cells, are required for LUAD development.

Enrichment of airway and alveolar signatures in experimental and human lung adenocarcinoma
We subsequently cross-examined the transcriptomes of LUAD cell lines isolated from urethaneinduced lung tumors [34,35] and of their originating murine lungs with the gene expression profiles of murine AEC isolated from tracheal explants, of murine ATII cells [36], and of murine bone- Similar analyses of the transcriptomes of human LUAD and corresponding healthy lungs [37] and the profiles of primary human AEC, ATII cells, and AMΦ [38][39][40], also disclosed that the AEC (but not the ATII and AMΦ) signature was significantly enriched in LUAD compared with healthy lung tissue ( Figure 6C and 6D). Gene set enrichment analyses showed that while AEC, ATII, and BMDM signatures were significantly enriched in murine lungs, the AEC signature predominated over ATII and BMDM signatures in LUAD cells. In addition, all human AEC, ATII and AMΦ signatures were significantly enriched in human LUAD compared with healthy lungs ( Figure 6E and 6F and Figure S7). These results were plausible by the early nature of the human surgical specimens examined compared with our murine cell lines that represent advanced metastatic tumorinitiating cells, and collectively indicated the presence of an anticipated alveolar, but also an unexpected airway epithelial transcriptomic signature in tobacco carcinogen-induced LUAD. 10 We characterized the dynamics of respiratory epithelial signatures in the postnatal mouse lung during aging and after challenge with noxious and carcinogenic insults. The contribution of airway and alveolar signatures to chemical-induced LUAD of mice and men is described for the first time ( Figure 7A). Although the peripheral location and molecular phenotype of these LUAD suggest an alveolar origin, we show here that both airway and alveolar-programmed cells are found in chemical-induced LUAD and that, in fact, AEC may play a more prominent role during the initial steps of chemical lung carcinogenesis. signatures. This is important because human LUAD is inflicted by chronic exposure to tobacco smoke and other environmental exposures [41]. As such, the mutation profile of the human disease is more closely paralleled by chemical-induced murine lung tumors compared with lung cancers triggered by transgenic expression of Kras G12C or Kras G12D in the respiratory epithelium [15].

DISCUSSION
Although the latter transgenic tumors have been extensively studied [9][10][11][12][13][14], chemical-induced LUAD have not been investigated. In both of the mouse models that we used, even in the FVB onehit model involving a single dose of carcinogen administration, all developed LUAD contained the Scgb1a1+ genetic marking, in contrast with the Lyz2+ genetic marking which was dispensable for LUAD formation. These observations could support a multi-stage course of events in chemical 11 carcinogenesis, involving at some point an airway-specific transcriptomic signature. In fact, the prevalence of a different Kras mutation in urethane-induced tumors (Kras Q61R compared to Kras G12 mutations in the transgenic mouse models) has led to the suggestion that chemical carcinogens introduce Kras mutations in a different population of tumor-initiating cells than the mouse models of genetic activation of Kras [15]. Our findings of Scgb1a1+ AEC being more sensitive than Lyz2+ ATII cells toKras Q61R mutations during the initiation steps of urethane-induced lung carcinogenesis further supports this notion. The findings of chemical-triggered LUAD, as well as of their precursor hyperplastic lesions, bearing Scgb1a1+, Sftpc+, and Lyz2+ markings, implies that they can originate from: i) AEC that colonize the distal lung during carcinogenesis thereby activating obligate (Sftpc+) and dispensable (Lyz2+) alveolar transcriptomic programs; ii) alveolar cells that transit through an obligate Scgb1a1+ and a dispensable Lyz2+ stage during the process; or iii) multipotent progenitors that express multiple epithelial signatures, such as those found during pulmonary embryogenesis, in human LUAD, and in other chronic lung diseases [40][41][42][43][44]. However, in our view, the propensity of airway cells to survive Kras Q61R mutations during the early initiation steps of urethane-induced lung carcinogenesis, and the close airway association of lung tumors revealed by our high resolution μCT analysis support a bronchial origin of these tumors, in line with recent evidence of tobacco smoke inducing epigenetic changes that sensitize human airway epithelial cells to a single KRAS mutation [45]. Along these lines, the split phenotype of chemical-induced lung tumors of mT/mG;Lyz2.Cre mice indicates that Lyz2-expressing ATII cells can be dispensable for carcinogentriggered LUAD development, as opposed to what has been previously shown for geneticallytriggered LUAD [14].
Our approach focuses on integral assessment of the lung epithelial transcriptomic signatures participating in adult lung changes in response to aging, injury, and carcinogenesis. The identification of transcriptomic programs and signatures that are activated during lung repair and carcinogenesis and that team up with oncogenic signaling in a non-oncogenic addictive fashion is of great importance for LUAD biology and is likely to lead to therapeutic innovations [46]. To this 12 end, it was recently shown that insertions and deletions in lineage-restricted genes occur in human LUAD [7]. Moreover, integrin β 3 and TANK-binding kinase 1 partner with oncogenic KRAS signaling to mediate cancer stemness and drug resistance [5,6]. Along these lines, our findings of both airway and alveolar transcriptomes being involved in lung maintenance, repair, and carcinogenesis is of great importance, as it implies that these programs facilitate the survival and proliferation of lung stem cells with regenerative potential and of mutated cells with malignant potential and are thus therapeutic targets. The perpetual cell marking approach adopted was preferential over available pulsed lineage tracing models because of the unprecedented accuracy of our Scgb1a1.Cre strain in exclusively and completely marking airway epithelial cells at the conclusion of development, allowing tracking of subsequent changes in adulthood.
In conclusion, airway cells and transcriptomic programs contribute to alveolar maintenance and LUAD. Since defective epithelial repair underlies the pathogenesis of chronic lung diseases and since abundantly transcribed genes are central to the mutational processes that cause cancer, this finding is of potential therapeutic importance for chronic pulmonary diseases and lung cancer.

DECLARATION OF INTERESTS
The authors declare no competing interests.  mice were donated by their founder [21]. Mice were bred >F12 to the FVB background at the University of Patras Center for Animal Models of Disease.

Mouse models of lung adenocarcinoma
Six-week-old mice on the C57BL/6 background received ten consecutive weekly intraperitoneal urethane injections (1 g/Kg in 100 μL saline) and were sacrificed 6-7 months after the first injection, or four consecutive weekly intraperitoneal MCA (15 mg/Kg in 100 μL saline) followed by eight consecutive weekly intraperitoneal BHT injections (200 mg/Kg in 100 μL corn oil) and were sacrificed 6-7 months after the first injection. Six-week-old mice on the FVB background received one intraperitoneal urethane injection (1 g/Kg in 100 μL saline) and were sacrificed 6-7 months later [16][17][18].

Mouse models of lung injury
Six-week-old mice (C57BL/6 background) received intratracheal bleomycin A2 (0.08 units in 50 μL saline) or intraperitoneal naphthalene (250 mg/Kg in 100 μL corn oil) [31,32]. In addition, preterm mothers of the C57BL/6 background and their offspring were exposed to room air (21% oxygen; control) or 98% oxygen for two days before and four days after birth [47]. 15 For induction of perinatal hyperoxic lung injury, preterm mothers of the C57BL/6 background and their offspring were exposed to room air (21% oxygen; control) or 98% oxygen for two days before and four days after birth, when oxygen-exposed pups were returned to room air, as described previously [47]. Oxygen levels were continuously monitored with an oxygen sensor. The gas stream was humidified to 40-70% by a deionized water-jacketed Nafion membrane tubing and delivered through a 0.22 μm filter before passage into a sealed Lexan polycarbonate chamber measuring 40 x 25 x 25 cm and accommodating 25 L gas at a flow rate of 5 L/min, resulting in complete gas exchange every 5 min. Mothers were cycled between litters on 21% and 98% oxygen every 24 hours to prevent oxygen toxicity and to control for nutritional support of the pups. After perinatal hyperoxia, mice remained at room air till sacrificed at eight weeks of age.

Human lung adenocarcinomas
Ten archival formalin-fixed, paraffin-embedded tissue samples of patients with LUAD that underwent surgical resection with curative intent between 2001 and 2008 at the University Hospital of Patras were retrospectively enrolled [27].

Statistical analysis
Sample size was calculated using power analysis on G*power, assuming α = 0.05, β = 0.05, and effect size d = 1.5. No data were excluded from analyses. Animals were allocated to treatments by alternation and transgenic animals were enrolled case-control-wise. Data were collected by at least two blinded investigators from samples coded by non-blinded investigators. All data were normally 16 distributed by Kolmogorov-Smirnov test, are given as mean ± SD, and sample size (n) always refers to biological and not technical replicates. Differences in frequency were examined by Fischer's exact and χ 2 tests and in means by t-test or one-way ANOVA with Bonferroni post-tests. Changes over time and interaction between two variables were examined by two-way ANOVA with Bonferroni post-tests. All probability (P) values are two-tailed and were considered significant when P<0.05. All analyses and plots were done on Prism v5.0 (GraphPad, La Jolla, CA).

SUPPLEMENTAL INFORMATION
Supplemental information includes supplemental methods and seven figures and can be found with this article online.    Our evidence supports the existence of distinct developmental ancestries for airway and alveolar type II (ATII) cells, notwithstanding their common descent from an early (possibly Sftpc+) lung epithelial progenitor. The developmental airway lineage (Scgb1a1+Sftpc±; green) gives rise to all types of airway cells, including club or Clara, ciliated, goblet, basal, and other cells, while the developmental ATII lineage (Sftpc+Lyz2±; red) gives rise to ATII cells formed before birth. These lineages appear to be relatively segregated in the growing unaffected lung of the mouse till the age of six weeks, which roughly corresponds to a human age of six years, where cellular proliferation in the human lungs ceases. Thereafter, and likely due to the continuous bombardment of the lungs by inhaled noxious particles and substances during normal respiration, gradual expansion of Scgb1a1+Sftpc± marked cells ensues. Upon lung injury, this process is markedly accelerated.

Figure Legends
Similarly, during carcinogenesis caused by chemical carcinogens of tobacco smoke, we show how  Figure 1A and Figure S1).
Lung lineages appear to be relatively segregated in the growing lung till the age of full lung development (six weeks in mice and 6-8 years in humans) or till lung injury ensues, whichever comes first. Schematic of postnatal redistribution of marked cells in the adult lung, based on the findings of the present work and a previous report [14]. Upon injury to airway and/or alveolar cells, during multi-stage field carcinogenesis, or even during unchallenged aging, Scgb1a1+ marked cells 28 appear in the distal alveolar regions, thereby maintaining lung structure and function. Bubble size indicates relative marked cell abundance. CCSP, Clara cell secretory protein; FOXJ1, forkhead box J1; KRT5, keratin 5; LYZ2, lysozyme 2; SFTPC, surfactant protein C; TUB1A1, acetylated αtubulin.