Cell Division Cycle 42 plays a Cell type-Specific role in Lung Tumorigenesis

Cell division cycle 42 (CDC42) plays important roles in polarity establishment and maintenance as well as cell cycle progression and cell division. Although disruption of cell polarity is a prerequisite in epithelial tumor initiation, the roles of CDC42 in tumorigenesis are still poorly understood. Here we find that Cdc42 deficiency inhibits the KrasG12D-induced lung alveoli tumor formation, while conversely promotes bronchiole tumor formation in mice. Bronchial Cdc42 loss destroys contact inhibition potentially through cell polarity disruption, and results in increased tumor formation. In contrast, deletion of Cdc42 in alveoli cells prevents KrasG12D-induced cell proliferation, which leads to reduced tumor formation. Further analyses of clinical specimens uncover a significant positive correlation between CDC42 and type II alveolar epithelial cells marker SP-A, indicating the potential importance of CDC42 in this specific subset of lung cancer. Collectively, we identify the lineage-specific function of CDC42 in lung tumorigenesis potentially through the regulation of cell polarity integrity.

and development of hepatacellular carcinoma 13 . Also, induced gene targeting of Cdc42 in murine bone marrow hematopoietic stem/progenitor cells results in a loss of hematopoietic stem cell quiescence and hyperproliferation of blood progenitors 14 . Consistently, neuroblastomas with N-myc amplification display deletions of the short arm of chromosome 1 containing the Cdc42 gene in 90-95% of cases, and one copy of Cdc42 is consistently lost in this type of cancer 15 . These data suggest that the role of Cdc42 as oncogene or tumor suppressor might be lineage dependent 16 .
Lung cancer is one of the most devastating diseases worldwide with different subtypes derived from trachea, bronchiole or peripheral alveoli. Previous studies have detected high CDC42 expression in human lung cancer samples 9 and cell lines 17 and demonstrate its contribution to cancer cell migration. Moreover, down-regulation of CDC42 is found to inhibit lung cancer cell growth 18 and invasiveness 17,[19][20][21][22] . CDC42 also promotes trans-endothelial migration of lung cancer cells through β1 integrin 23 . These observation are consistent with oncogenic role of CDC42.
Here through detailed studies of Cdc42 deletion in distinct cell types using lineage specific promoter driven CRE in Kras G12D driven lung cancer mouse model, we have identified both tumor-promoting and tumor-suppressive function of CDC42 in type II alveolar epithelial cells and Club cells, respectively. Our data further show that CDC42 prevents lung bronchiole tumor formation potentially through regulation of cell polarity integrity. In accordance with its tumor promoting role in alveolar tumor formation, CDC42 expression is positively correlated with alveolar marker surfactant protein A1 (SP-A) expression in human lung adenocarcinoma patients.

Results
Cdc42 loss promotes bronchiole tumor formation but inhibits alveoli tumor formation in Kras mouse model. To investigate the potential role of CDC42 in lung tumorigenesis, we crossed the conditional Cdc42 L/L allele with Lox-Stop-Lox Kras G12D (hereafter named as Kras) allele to get Lox-Stop-Lox Kras G12D ; Cdc42 L/L allele (hereafter named as Kras/Cdc42) 24,25 . Mice were then inoculated with Cre-expressing adenovirus (Ad-Cre) by nasal inhalation as previously described 25 , and analyzed at a series of time points (Fig. 1a). In this model, it remains possible that multiple pulmonary cell lineages are infected with Ad-Cre and neoplastic lesions arise from both bronchiole and alveoli epithelial cells 25 . We initially confirmed the Cdc42 deletion in lung tumors derived from Kras/Cdc42 mouse model (Fig. 1b, Supplementary Figs S1-2). As the control, deletion of Cdc42 alone did not result in any tumor formation over 70 weeks post Ad-Cre treatment (Fig. 1c). Consistent with the essential role of CDC42 in promoting cell division and neoplastic transformation 2, 26 , Cdc42 loss significantly decreased the lesion number and percentage of alveolar tumors in Kras/Cdc42 mice ( Fig. 1d-f). Surprisingly, we observed a significant increase of the lesion number and percentage of bronchiolar tumors in this model ( Fig. 1d-f), featured with the papillae protrusion into airway lumens (Fig. 1d). These bronchiolar lesions in Kras/Cdc42 model exhibit a high cell proliferating index (presented by KI67 staining) compared with those in Kras model (Fig. 1g,h). This analysis demonstrated that Cdc42 loss increased formation of bronchial and bronchiolar epithelial tumors, but decreased Kras-induced tumor formation in alveoli.
Cdc42 loss disrupts bronchiole cell polarity. We then asked how Cdc42 loss promoted the bronchiole tumor formation. Normal bronchioles are lined by pseudostratified or single layer epithelia which potentially contribute to contact inhibition and act as the important barrier for neoplastic transformation 27,28 . Since CDC42 plays a central role in establishing and maintaining epithelial polarity which is frequently disrupted during tumor progression, we first analyzed the subcellular localization of a series of polarity proteins in Kras or Kras/Cdc42 bronchioles at three weeks post Ad-Cre treatment. Our data showed a subcellular dislocation of both ZO1 and Phalloidin, markers for tight junction and F-ACTIN assembling respectively, in Kras/Cdc42 mice bronchioles at three weeks post Ad-Cre treatment (Fig. 2a). Similar dislocation of other polarity components including PAR6 and Occludin was also detectable in Kras/Cdc42 mice (Fig. 2b). Moreover, our data from electron microscopy analyses demonstrated that the ultrastructure of tight junction in bronchioles was disrupted in Kras/Cdc42 mice but not in Kras mice (Fig. 2c).

Cdc42 loss promotes the proliferation of KRAS-activated polarized cells potentially through disruption of cell-cell contact inhibition.
To test if Cdc42 loss indeed promotes lung bronchiolar tumorigenesis through disruption of polarity, we utilized two types of cells with distinct polarities including mouse trachea epithelial cells (MTEC) and mouse embryonic fibroblasts (MEF) to analyze the functional consequence of Cdc42 deletion. In comparison with KRAS activation alone, deletion of CDC42 together with KRAS activation almost completely disrupted the tight junction in MTEC, indicated by ZO1 staining (Fig. 3a), and dramatically increased cell growth (Fig. 3b). On the contrary, in MEF cells which did not form tight junction, Cdc42 deletion abrogated the promotive effect of KRAS upon cell proliferation (Fig. 3c,d).
We further took advantage of Madin-Darby canine kidney (MDCK) cells, which is widely used for polarity study. Our data showed that Cdc42 knockdown remarkably inhibited KRAS-activated MDCK cell growth at low cell density but the cell growth was significantly enhanced after reaching 100% confluence (Fig. 3e,f). These data clearly demonstrated that Cdc42 loss plays opposite roles in cell proliferation, depending on whether cell-cell junction exists or not.
The three-dimensional (3D) system is an excellent model of epithelial morphogenesis in which cells embedded in matrigel form acini, a spherical monolayer enclosing a central lumen. To study the potential role of CDC42 in restricting neoplastic transformation, we infected pre-formed MDCK acini (day 6) with a low dose of lenti-virus expressing Kras G12D with or without Cdc42 knockdown. Consistent with previous study, KRAS activated single cell remained quiescent in growth-arrested acini, with apical location of ZO1 and Phalloidin (Fig. 3g, upper panel). In contrast, single cell with Cdc42 knockdown in context of KRAS activation initiated inner-growing cell SCIENTIfIC RepoRts | 7: 10407 | DOI:10.1038/s41598-017-10891-0 mass from the epithelial layer with disordered ZO1 and Phalloidin localization (Fig. 3g, bottom panel), which morphologically resembled papillae protruding into the airway lumens in Kras/Cdc42 mice bronchiole (Fig. 1d). These results substantiated the key role of CDC42 and polarity integrity in inhibiting tumor formation.
Lineage-specific deletion of Cdc42 in Kras-driven lung cancer mouse model. Mouse bronchiolar and alveolar tumors are considered to be derived from Club cells and type II alveolar epithelial cells (AECII) respectively 29,30 . CCSP and the surfactant protein-C (SP-C) are commonly-used markers to distinguish Club cell and AECII 25 . We found that the bronchiolar tumors from Kras/Cdc42 model were stained positive for CCSP but negative for SP-C, whereas the alveolar tumors were opposite (Fig. 4a). Based on these data, we hypothesize that CDC42 might play a cell type-specific role in lung tumorigenesis. We then utilized lineage specific ablation of Cdc42 to dissect the potential roles of Cdc42 in bronchiolar and alveolar tumor formation. For this, we first constructed the SPC-Cre-ERT2 (mainly targeting AECII cells) and CCSP-Cre-ERT2 transgenic alleles (mainly targeting Club cells in bronchiolar epithelia 31 ) (Fig. 4b). Given that the CCSP-Cre-ERT2 strain was novelly constructed, we crossed these two alleles to Rosa26R-LacZ reporter mouse strain 32 and found that tamoxifen-induced LacZ expression from Rosa26 locus was mainly restricted to Club cells that located at bronchioles, and AECII cells located at alveoli, respectively (Fig. 4b). These data are consistant with our previous study 33 . We then crossed SPC-Cre-ERT2 or CCSP-CreERT2 allele with either Kras or Kras/Cdc42 mice to generate SPC-Cre-ERT2/Kras (SK) 34 , SPC-Cre-ERT2/Kras/Cdc42 (SKC), CCSP-Cre-ERT2/Kras (CK) and CCSP-Cre-ERT2/Kras/Cdc42 (CKC) mice and performed the detailed tumor analyses after tamoxifen treatment (Fig. 4c). As expected, the majority of tumors in the SK model were located at alveoli; SKC mice demonstrated significantly decreased tumorigenesis (alveolar tumors) compared to the SK group (Fig. 4d,e). On the other hand, we found that the majority of lung tumors in CKC mice were located at bronchioles, and stained positive for CCSP but negative for SP-C (Fig. 4d). Interestingly, a dramatic increase of tumor formation mainly in bronchiole was detected in CKC mice compared to CK mice (Fig. 4d,f,g). Less than 20% bronchioles tumor was found in CK mice, suggesting that KRAS G12D activation alone is difficult to transform the bronchiole epithelium cells with intact cell polarity.
Collectively, these data support that CDC42 functions oppositely in different lineage tumor formation: it promotes tumor growth in alveoli while prevents tumor formation in bronchioles.

High CDC42 level correlates with SP-A level in human lung adenocarcinoma.
To further study the potential clinical relevance of our findings from mouse model, we analyzed a set of 84 lung adenocarcinoma for expression patterns of CDC42 and SP-A, a marker for human AECII 35 . Immunohistochemistry staining were performed and blindly scored according to Allred score system (Fig. 5a). Consistent with previous study 9 , we found that 57.5% of tumors were positive for CDC42 expression. Pearson correlation analysis showed that CDC42 level positively correlated with SP-A level (Fig. 5b,c; r = 0.3436, P = 0.0014). Although SP-A is a biomarker for normal human AECII, human lung cancer frequently lost the expression of lineage biomarkers. Thus, our clinical data analyses only provide an implication that CDC42 expression might be specifically high in those lung cancer derived from AECII.

Discussion
In this study, we have identified an unexpected cell type-dependent role of CDC42 in lung tumorigenesis. In AECII, Cdc42 loss strongly prevents Kras-driven neoplastic transformation, establishing the tumor-promotive role of CDC42. These results are consistent with early studies carried out in various mammalian cell lines by overexpressing dominant-negative and/or constitutively active mutants showing the necessity of CDC42 in transformation 11,12,36 . In contrast, our data shows that Cdc42 loss not only disrupted epithelial cell polarity and branching   morphogenesis of the developing lung as previously described 37 , but also promotes KRAS-induced overgrowth and tumor formation in adult Club cells.
Our results further suggest that Cdc42 loss induces Kras-derived bronchiole tumor formation potentially through disruption of cell polarity and cell-cell contact inhibition. Compared with alveolar cells, bronchiolar epithelia cells are close-packed and express polarity-related protein specifically located on apical or basal, which function as a non-cell-autonomous tumor suppressor to restrict KRAS-induced cell proliferation. Since CDC42 plays an essential role in regulation of cell polarity in these cells, its deletion leads to disrupted cell-cell junction, the loss of contact inhibition, eventually triggers cell overgrowth. Together, these data suggest that the cell polarity maintained by CDC42 might serve as an important barrier for bronchiolar tumor formation in lungs. Our findings are consistent with Previous studies showing that the dysfunction of certain polarity genes is associated with tumor initiation or progression [38][39][40] .The role of CDC42 in cell proliferation (mainly in alveoli) and contact inhibition (mainly in bronchiole) during the tumorigenesis is summarized in Fig. 6.
In conclusion, our data show that CDC42 functions in a cell-type specific manner in lung tumorigenesis in context with different cell lineages. Loss of Cdc42 promoted KRAS-induced Club cell tumor formation, underscoring an important tumor initiation function of polarity loss which was previously considered a by-product of abnormal cell accumulation. In addition, studies from mouse model and cell lines can be further applied to human pathological conditions. Of course, more detailed studies are needed to determine the feasibility and therapeutic benefit of targeting CDC42 in human disease.

Materials and Methods
Mouse treatment. Kras G12D , and Rosa26-LacZ mice were originally generously provided by Dr. T. Jacks and Dr. L. Cheng, respectively. Cdc42 L/L mice were generated as described before 24,41 . The SPC-CreERT2 allele was made as previously reported 34 . The CCSP-CreERT2 allele was made using the rat CCSP promoter 31 driving the CreERT2 gene expression with similar strategy as previously reported for SPC-CreERT2 allele 34 . Progeny were screened using southern blot and confirmed by PCR strategy. All mice were housed in a specific pathogen-free environment at the Shanghai Institute of Biochemistry and Cell Biology and treated in strict accordance with protocols approved by the Institutional Animal Care and Use Committee of the Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences. All experiments were performed in accordance with relevant guidelines and regulations. For Kras, Kras/Cdc42 lung cancer mice model, mice were treated with 2 × 10 6 plague-forming units (PFU) of Ad-Cre (purchased from University of Iowa adenoviral core) or lenti-virus through nasal inhalation at 6~8 weeks of age 25 . For SPC-Cre-ERT2/Kras (SK), SPC-Cre-ERT2/Kras/Cdc42 (SKC), CCSP-CreERT2/Kras (CK) or CCSP-CreERT2/Kras/Cdc42 (CKC) mouse models, tamoxifen in sunflower oil (40 mg/ kg) were given to mice via intraperitoneal injection daily for five continuous doses. Mice were sacrificed at a serial time points for gross inspection, histopathological examination and molecular analyses. The numbers of lung bronchiolar and alveolar tumors were analyzed according to pathology for different mouse cohorts and the percentage of regional tumors was then calculated accordingly.

Constructs, Cell culture and in vitro
For in vitro growth assay, cells were seeded in 96-well plates (2500 cells per well), and the proliferation information were obtained by MTT (3-4, 4-dimethylthiazol-2, 5 diphenyl tetrabromide) assay every day as previously described 43 .
Mouse tracheal epithelial cell culture (MTEC). MTEC harvest was performed as previously described 44 .
Briefly, 8-10 wk Kras or Kras/Cdc42 mice were sacrificed and immersed in 70% ethanol (avoiding airway submersion), and tracheas were resected from the larynx to the bronchial main branches and collected in ice-cold Ham's F-12 pen-strep. After overnight digestion of 0.25% trypsin at 4 °C, the suspended trachea cells were collected for incubation in tissue culture plates for 3-4 hrs in 5% CO 2 at 37 °C to adhere fibroblasts, and nonadherent cells were collected by centrifugation, resuspended in MTEC/Plus medium, and maintained in collagen coated 24-well plate. After 72 hrs growth, MTEC (~1 × 10 5 ) at 24-well plate were virally infected with Ad-Cre (1 × 10 6 CFU) overnight and then changed with fresh medium. Immuno or crystal violet staining was performed after another 4 days' growth.
Electron microscopy. After the mice were sacrificed, small fragments (~1 mm 3 ) of lung were fixed overnight in a solution containing 4% paraformaldehyde, 2.5% gluteraldehyde in 0.1 M PBS (pH 7.4). Tissue fragments were washed for 20 min in 0.1 M PBS for three times and subsequently treated with 2% osmium tetroxide in PBS for 1.5 hrs. After washing for 5 min × 3 with PBS, specimens were dehydrated in a graded ethanol series and embedded in Epon. For orientation purposes, 1 µm sections were stained with toluidine blue. Ultrathin sections (70-80 nm) were then cut with a Reichert-Jung ultramicrotome, collected on formvar coated nickel grids, stained with uranyl acetate for 10 min and with lead citrate for 7 min. Samples were examined with a FEI Tecnai G2 Spirit Transmission Electron Microscope.

Human Lung Cancer Specimen analyses.
A total of 84 NSCLC patient samples were collected with the approval by the institutional review committees of Shanghai Cancer Center, Fudan University. Patients gave written informed consents. The specimens were used for immunostaining of SP-A and CDC42 and analyzed for clinical relevance. All human methods were performed in accordance with the relevant guidelines and regulations. Statistical Analysis. Data were analyzed by Student's t test or Pearson correlation test; P < 0.05 was considered significant.
Data availability. All data generated or analysed during this study are included in this published article.