Mir-17-92/P38a Dysregulation Enhances Wnt Signaling and Selects Lgr6 Cancer Stem-Like Cells During Lung Adenocarcinoma Progressi

Published OnlineFirst May 4, 2016; DOI: 10.1158/0008-5472.CAN-15-3302

Cancer Tumor and Stem Cell Biology Research

miR-17-92/p38a Dysregulation Enhances Wnt Signaling and Selects Lgr6þ Cancer Stem-like Cells during Lung Adenocarcinoma Progression Anna Guinot1, Feride Oeztuerk-Winder1, and Juan-Jose Ventura1,2

Abstract

þ Defining the molecular and cellular roots of lung cancer outgrowth of Lgr6 NSCLC cells. High levels of expression of relapse after initial treatment remains an imperative to miR-19 family members were found to target and down- improve survival. Here we report that the lung stem cell regulate levels of p38a kinase, providing a specific survival þ marker Lgr6 becomes enriched in non–smallcelllungcancer signal for Lgr6 cells as mediated by increased Wnt/ß-catenin þ (NSCLC) cells during malignant progression. Lgr6 NSCLC activity. Our results identify a specific stem-like cell popula- cells displayed self-renewal and differentiation properties tion in NSCLC with increased malignant potential, the elu- along with a higher tumorigenic potential. Mechanistic inves- cidation of which may enable earlier prognosis and possibly tigations suggested that a defective repression of the miR-17- the development of more effective targeted treatments. Cancer 92 gene cluster was responsible for evolution of a selection for Res; 76(13); 4012–22. 2016 AACR.

Introduction In this article, we have studied the cellular and molecular mechanisms behind the development and progression of human Lung cancer is the major cause of cancer-related deaths in lung adenocarcinoma. We have previously characterized a pop- western countries (1). Early detection and surgery have proven ulation of human lung cells with stem cell potential to differen- to be the most effective ways to improve survival (2). Important tiate into all bronchioalveolar lineages while being able to self- advances have been made in the last 20 years, with the detection of renew. These epithelial cells express the stem cell marker Lgr6 oncogenes and mutations linked to many lung cancer types. (Leucine-rich repeat-containing G protein–coupled receptor 6). However, therapies applied using drugs directed to those mole- Lgr6, and its homologous receptors, are amplifiers of Wnt signal- cules or related pathways have failed to significantly increase ing and are involved in maintaining stem cell self-renewal (15). survival, especially in the most prevalent cancer types, for exam- þ We have analyzed the contribution of Lgr6 cells to human lung ple, lung adenocarcinoma (3–5). Cancer relapse after radiother- adenocarcinomas. We have observed that cells-expressing Lgr6 are apy and chemotherapy is the essential reason for reduced survival enriched from early to late stages of human lung adenocarcino- (6). Previously, efforts have focused on investigating the cells and mas malignancy. These cells showed a disruption in the balance mechanisms involved in the resistance and later virulence of lung between the p38a and miR-17-92 pathways. p53 mutation pre- cancers after therapeutic treatments (7, 8). Recent theories in vents its repression of miR-17-92 expression (16). cancer research consider the existence of cell populations with Cross-talk between signaling pathways and its role in cancer has stem cell–like properties in many cancer types (9, 10). These been extensively reported elsewhere (17, 18). Reduced p38a cancer stem cells would be responsible for tumor maintenance promotes activation of Wnt signaling and, in human lung cancer and would harbor the potential to colonize other tissues as samples, we observed a correlation between a deficiency in p38a metastatic cells (11, 12). These cells would also be especially þ protein and increase of Wnt signaling, promoting Lgr6 cells resistant to drugs, and they would be selected during tumor during cancer progression. progression, generating more aggressive and metastatic disease These findings provide a new tool for early prognosis of (13, 14). human lung adenocarcinomas. We demonstrate how a novel- specific population of cells, and their regulatory mechanisms, is linked and contribute to better understanding of the malignant 1 Wellcome Trust-Medical Research Council Stem Cell Institute, Uni- progression of lung adenocarcinomas. This coadjuvant process, versity of Cambridge, Cambridge, United Kingdom. 2Translational Cell and Tissue Research, Department of Imaging and Pathology, KU in parallel to oncogenic mutations, helps select cells with a Leuven, Belgium. higher potential to perpetuate (self-renewal) and manipulate Note: Supplementary data for this article are available at Cancer Research the microenvironment, facilitating colonization of foreign tis- Online (http://cancerres.aacrjournals.org/). sues (metastasis). Corresponding Author: Juan-Jose Ventura, KU Leuven, Minderbroedersstraat 12, blok Q, Leuven 3000, Belgium. Phone: 320-1633-6615; Fax: 320-1633-6615; Materials and Methods E-mail: [email protected] Isolation and culture of human lung adenocarcinoma cells doi: 10.1158/0008-5472.CAN-15-3302 Cultures were established from freshly isolated lung adenocar- 2016 American Association for Cancer Research. cinoma cells obtained from cancer patients undergoing tumor

4012 Cancer Res; 76(13) July 1, 2016

Lgr6 Labels Malignant Human Adenocarcinoma Cells

resection (Papworth hospital; Cambridge, United Kingdom). in paraffin or OCT. Paraffin blocks were cut in 5-mm sections using Approval by the Ethics Committee at Cambridge University and a microtome. OCT blocks were cut in 10-mm sections using a the Papworth Hospital was received. All patients gave their cryostat microtome at 20C. Hematoxylin and eosin (H&E) informed consent. Papworth Hospital pathologists provided staining was performed using a Leica Autostainer XL. information related to the type, stage, and grade of the tumors. Primary antibodies used: anti-human Lgr6 (Santa Cruz Bio- Human lung adenocarcinomas specimens were finely minced technology, sc-99123), anti-human TTF1 (Abcam, ab-40880), and resuspended in a mix of 1 mg/mL collagenase type 1/colla- anti-human Ki67 (Vector, VP-RM04), anti-human p53 (Novocas- genase type 3 (Worthington Biochemical, catalog no.: LS004196 tra, P53-CM5P), anti-human P-GSK3 (Ser9; Abcam, ab-75814), and LS004182) in DMEM. The suspension was incubated at 37C anti-human b-catenin (Santa Cruz Biotechnology, sc-7199), anti- with shaking (220 rpm) for 45 minutes. The suspension was human mitochondria (Thermo Fisher Scientific, MA5-12017), centrifuged for 5 minutes at 1,200 rpm, and the pellet was anti-human SP-C (Santa Cruz Bioetchnology, sc-7705), anti- resuspended in fresh DMEM before filtration through a 70-mm human CC-10 (Santa Cruz Biotechnology, sc-25554), anti- Cell Strainer (BD Falcon, catalog no.: 352350). After a second human E-Cadherin (Abcam, ab-53033), anti-human CK7 centrifugation (5 minutes at 1,200 rpm), the pellet was washed in (Abcam, ab-90083), and anti-human p63 (Santa Cruz Biotech- PBS and treated with red blood cell lysis buffer (Roche, catalog no: nology, sc-8431). 11814389001) for 5 minutes. The cells were pelleted again (5 The Ventana Discovery (Ventana Medical Systems) automated minutes at 1,500 rpm), washed in PBS, filtered through a 50-mm system was used for IHC on paraffin and frozen sections. Ventana CellTrics Sterile Filter (Partec, catalog no.: 04-004-2327) and Cell Conditioning 1 solution (CC1) was used to perform antigen placed in culture. Cells were cultured in 1:1 DMEM (Gibco, retrieval in all sections. Antibody detection was performed using catalog no.: 41965-039) and Ham's F12 (Lonza, catalog no.: the DAB Map Detection Kit (Ventana, catalog no.: 760-124), BE12-615F) supplemented with 5% FBS (Gibco, catalog no.: sections were counterstained with hematoxylin using the Leica 10270-106), hEGF (10 ng/mL, Pepro Tech, catalog no.: AF- Autostainer system and mounted manually. 100-15), FGF-2 (20 ng/mL, Pepro Tech, catalog. no.: 100-18B), For immunofluorescent staining, sections were incubated in and 0.1% penicillin/streptomycin (Sigma, catalog no.: P0781) blocking buffer (PBS, 4% donkey serum, 1% Triton) for 1 hour at antibiotics. Cells were incubated at 37 C in 93% air with 7% CO2. room temperature. Primary antibodies were incubated overnight Medium was changed every 2 to 3 days. Cells were disaggregated at 4C. Sections were rinsed three times in PBS and incubated with using accutase (PAA, catalog no.: L11-007) for 5 minutes at 37C secondary antibodies diluted at 1:1,000 for 1 hour at room to produce a single-cell suspension. Cells were then pelleted in a temperature. Slides were mounted in Vectashield mounting centrifuge (1,200 rpm for 5 minutes), washed in PBS (Sigma, media with 4',6-diamidino-2-fenylindool (DAPI). catalog. no: D8537), pelleted a second time, and resuspended in fresh medium. Immunostaining quantification All cells used were directly isolated from human samples and The number of positively stained cells on sections was quan- characterized by IHC and qPCR expression of lung and adeno- tified using Cell Profiler software. Three or more biologic repli- carcinoma markers. No commercial cell lines were used for this cates of different stages of human lung adenocarcinoma were used work. to obtain the average number of positive cells in randomly selected fields. The significance of quantitative data was tested p38a siRNA knockdown t þ using an unpaired, two-tailed test. The SD of the mean is Lgr6 human lung stem cells' isolation and culture has been represented in the graphs by error bars. previously described by our group (16). þ þ Lgr6 p38a knockdown cells (Lgr6 p38KD) were obtained by þ infecting Lgr6 human lung stem cells with lentiviral vector Western blotting pLKO.1-TRC (Sigma, catalog no: SHC001) carrying an short Cells were lysed in lysis buffer [50 mmol/L Tris-HCl pH 7.5, 150 hairpin RNA (shRNA) construct to knock down p38a (50- mmol/L NaCl, 1% (v/v) NP-40, 5 mmol/L EDTA pH 8.0, 5 m CCCGGT-GTCCATCATTCATGCGAAA-TTCAAGAGA-TTTCGCA- mmol/L EGTA pH 8.0, 20 mmol/L NaF, 0.1 mol/L PMSF, 0.1 m TGAATGATGGACTG-TTTTTT-G) and a puromycin selection mol/L NaVO3, plus complete protease inhibitor cocktail – marker. Cells were selected in puromycin (15 mg/mL, Sigma, cata- (Roche)] and cellular lysates were separated by SDS PAGE and fl log no.: P9620) for 10 days. transferred to polyvinylidene di uoride membranes. The follow- þ þ a Lgr6 p38a knockdown cells with miR-17-92 rescue (Lgr6 ing antibodies were used for protein detection: p38 MAPK (Cell þ p38KD shmiR) were obtained by infecting Lgr6 p38KD cells with Signaling Technology, #9228), Lgr6 (Santa Cruz Biotechnology, a lentiviral vector pLKO.1-TRC carrying an shRNA construct to knock sc-99123), -tubulin (Sigma, T9026), P-GSK3 (Cell Signaling down pri-miR-17-92 (50-CCGGT-AAGGAGAGCTCAATCTGCA- Technology, #9323). CA-CTCGAGTG-TGCAGATTGAGCTCTCC-TTTTTTT-G) and a gen- eticin selection marker. Cells were selected in G418 (0.5 mg/mL, Total RNA isolation and quantitative RT-PCR Invitrogen, catalog no.: 10131019) for 10 days. Total RNA was extracted from cells or tissue using miRVana In both cases, infectious virus was added to cells in the presence miRNA Isolation Kit (Ambion, catalog no.: AM1560) following of 8 mg/mL polybrene (hexadimethrine bromide, Sigma) and the manufacturer's instructions and was treated with DNAse I incubated for 6 hours. (Promega, catalog no.: M6101). miRNA expression was detected using TaqMan miRNA assays: U6 snRNA (001973), hsa-miR-17a Histology and immunostaining (002038), hsa-miR-18a (002422), hsa-miR19a (000395), hsa- Lung adenocarcinoma tissue samples and xenografts were fixed miR-19b (002425), hsa-miR-20a (000580), hsa-miR-92a overnight at 4C in 4% paraformaldehyde (PFA) and embedded (000431).

www.aacrjournals.org Cancer Res; 76(13) July 1, 2016 4013

Guinot et al.

Table 1. DNA sequence of the primers used for qPCR Forward Reverse Lrp6 50`-TGGGGAGAAGTGCCAAAGATAGAA-30 50TGGGGCAAGCACACTGATAAAAA-30 Lef1 50-GACGAGATGATCCCCTTCAA-30 50-AGGGCTCCTGAGAGGTTTGT-30 Dab2 50-GCTAGCTATTGCAAATGAGGG-30 50-GCCCTGTGAGAGTCTGTTGC-30 Dkk1 50-CTCGGTTCTCAATTCCAACG-30 50-GCACTCCTCGTCCTCTG-30 Axin2 50-AGTGTGAGGTCCACGGAAAC-30 50-CTGGTGCAAAGACATAGCCA-30 Tcf4 50-TATGCTCCATCAGCAAGCACTG-30 50-TGGATGCAGGCTACAGTAGCTG-30 Wif1 50-TATGGATCGATGCTCACCAG-30 50-TGCAGCTTGCCAGGTAAAAT-30 Gremlin 50-AACAGTCGCACCATCATCAA-30 50-CGATGGATATGCCAACGACAC-30 GAPDH 50- GAAGGTGAAGGTCGGAGT-30 50- GAAGATGGTGATGGGATTTC-30 HPRT 50-ACCGTGTGTTAGAAAAGTAAGAAG-30 50-AGGGAACTGCTGACAAAGATTC-30

Expression of all other genes was quantified using SYBR Green Biosciences, catalog. no.: 356230) following Lgr6 FACS sorting. technology and normalized to a housekeeping gene (GAPDH or Cells were injected under the skin of 6- to 8-week-old nude or HPRT), using specific primers (Table 1). Specificity of PCR pro- NOD/SCID mice with a 0.5 mL syringe. Mice were killed 10 to 12 ducts was tested by dissociation curves. weeks later. Subcutaneous growths were removed and prepared for histology sections. Flow cytometry Single-cell FACS (MoFlo, Beckman Coulter) was performed Tail vein injections following 1-hour incubation with anti-Lgr6 (Santa Cruz Biotech- NOD/SCID male mice were warmed for a few minutes, to nology, sc-99123) and 1 hour in the dark with anti-Rabbit PE induce vasodilatation, in a heated IVC recovery chamber set at secondary antibody (Santa Cruz Biotechnology, sc-3745). 40C. Mice were then put in a restraint tube and the tail cleaned Flow cytometry analysis was performed (BD LSRFortessa) using warm surgical scrub and water. A single-cell suspension of fi following xation in 2% PFA and incubation in anti-Ki67 (Vector, freshly isolated lung adenocarcinoma cells following FACS sort- VP-RM04) and anti-Lgr6 (Santa Cruz Biotechnology, sc-48236). ing, was resuspended in 100-mL PBS following FACS sorting and Cells were incubated in the dark for 1 hour with the following injected into the tail vein using a 30G needle and 1 mL syringe. secondary antibodies: anti-Rabbit Alexa Fluor 488 (Invitrogen, Mice were killed 10 to 12 weeks later. Lungs were removed and A21206) and anti-Goat Alexa Fluor 555 (Invitrogen, A21432). prepared for histology sections.

Soft-agar colony formation assay Results A 1% low melting agarose (Bio-Rad, catalog no.: 161-3111) We have described a population of human lung cells with the aliquot was melted in a microwave and subsequently cooled potential to differentiate both, in vitro and in vivo, into all mature down to 40C in a water bath. DMEM 10% FBS, prewarmed at bronchioalveolar cell types. These cells specifically express the 40C, was used to produce a final concentration of 0.5% agarose membrane marker Lgr6 (16). In mouse lung cancer models, by mixing equal volumes of both. A base layer of 1 to 2 mL of 0.5% bronchioalveolar stem cells are expanded during adenocarcino- agarose was placed at the bottom of each well of a 6-well plate (BD ma transformation (19, 20). We studied the expression of the Falcon, catalog no.: 353046) and allowed to cool down for 2 to 3 human bronchioalveolar stem cell marker Lgr6 in human lung minutes at 4C and then at room temperature. Different amounts adenocarcinomas. The adenocarcinoma origin of the tumors was of single cells were resuspended in DMEM 10% FBS and mixed þ confirmed as TTF1 /p63 to discard any tumors of squamous equally (1:1) with 0.7% at 40C agarose to produce a final origin (Supplementary Fig. S1). concentration of 0.33% agarose. Carefully, 1.5 mL of the cell mixtures were pipetted on top of each base layer and allowed to þ Accumulation of Lgr6 cells during human lung settle for 30 minutes at room temperature. Cells in this semisolid adenocarcinoma progression media were incubated at 37 C for 3 weeks and then stained with We have analyzed 53 human lung cancers for the expression of 0.005% crystal violet solution to allow colony counting. a number of lung specific, cancer, and stem cell markers in Cell proliferation assay samples of lung small cell, squamous, and adenocarcinoma at – In 24-well plates, 4 105 cells per well were plated in triplicates different stages. We observed that adenocarcinomas (TTF-1 pos- in general culture conditions. The number of viable cells was itive) showed increased expression of the stem cell marker Lgr6 at counted every day for 6 days using an automated Vi-CELL Cell later stages (Fig. 1A). Lgr6 expression in both earlier (stage I, stage fi Counter (Beckman-Coulter, catalog no.: 731050). II) or later (stages III, IV, and metastatic) disease was speci cally localized to the tumor area, being absent from the surrounding Mouse experiments healthy tissue (Fig. 1A). There was a direct correlation between All mouse experiments were subjected to ethical approval and adenocarcinoma stage and Lgr6 expression levels in tumor sam- performed according to United Kingdom Home Office Regula- ples, as observed by protein expression (Fig. 1B). Analysis of cell tions. CD-1 nude and NOD/SCID mice (Charles River Laborato- populations from different stage adenocarcinomas by flow cyto- þ ries) were maintained under standard pathogen-free conditions. metry confirmed the increase in the Lgr6 population during tumor progression and was shown to be particularly enriched þ Subcutaneous injections after stage II (Fig. 1C). To further understand how Lgr6 cells are Freshly isolated lung adenocarcinoma cells were resuspended selected, we screened early- and late-stage adenocarcinomas for in 100 mL of a 1:1 mix of culture medium and Matrigel (BD both proliferation and apoptosis. Immunostaining of tumor

4014 Cancer Res; 76(13) July 1, 2016 Cancer Research

Lgr6 Labels Malignant Human Adenocarcinoma Cells

A Lung Stage I Stage II Metastatic

H&E

TTF1

40 35 30 25 Lgr6 20 cells in lung

+ 15 10

adenocarcinomas 5 % Lgr6 0 Stage I Stage II Metastatic

B Lung Stage I Stage II Metastatic C Neg. Control Stage I Lgr6

0.382% 15.6% Tubulin

9 * 99.2% 79.2% 8 7 6 Lgr6 5 * SSC Stage II Metastatic 4 3 * 24.5% 2 27.3% 45.7% 1 0 FSC

Lgr6 relative protein levels relative Lgr6 Lung Stage I Stage II Metastatic

69.2% 46.6%

D Stage I Stage II Metastatic FSC

Ki67

70 E 60 Neg. Control Tumor 50 0.268% 0.067% 11.9% 28.3% 40 30 Lgr6 20

0.08% 45.4% 14.4% % Ki67-positive cells % Ki67-positive 10 99.6% 0 Stage I Stage II Metastatic Ki67

Figure 1. Increased Lgr6þ cells during human lung adenocarcinoma progression. A, histopathology images of normal lung (left) and different stages of lung adenocarcinoma (TTF1-positive tumors). Lgr6 expression is higher in later stages. Representative images of 15 stage I, 30 stage II, and 8 metastatic lung adenocarcinomas. B, detection by Western blotting of Lgr6 protein expression correlates with imaging at different adenocarcinoma stages. Numbers are the mean of three independent experiments SD. Significant differences are marked (, P 0.05). C, flow cytometry analysis of control, stage I, stage II, and metastatic lung adenocarcinoma tumors for Lgr6-expressing cells. Representative experiment of three biologic replicates. FSC, forward scatter; SSC, side scatter. D, immunohistology images show expression of the proliferation marker Ki67. The graph shows the quantification of Ki67þ cells at different stages. N ¼ 5 stage I, 10 stage II, 5 metastatic tumors (5 slides per sample). Percentage of positive cells SEM in randomly selected fields. E, flow cytometry analysis of Lgr6 and Ki67þ cells. The plots are representative of five stage II lung adenocarcinoma tumor biologic replicates.

www.aacrjournals.org Cancer Res; 76(13) July 1, 2016 4015

Guinot et al.

A B Stage I Metastatic Figure 2. þ 12 Lgr6 cell accumulation correlates Lung with dysregulation of the miR-17-92/ p38a axis during adenocarcinoma 10 Stage I p53 development. A, graph shows qPCR Stage II 8 analysis of miR-17-92 members' Stage III relative expression at different stages of lung adenocarcinoma compared 6 Metastatic 1.2 with normal lung. Bars depict the mean, SD, of three independent 4 1 experiments with ﬁve biologic 0.8 replicates per stage. B, 2 0.6 immunohistochemical staining of Relative expression levels to U6 levels expression Relative 0.4 early and late stages of lung 0 adenocarcinoma showing nuclear

p21 mRNA relative 0.2 miR-18 miR-20 miR-92 expression of total p53. miR-17 miR-19a miR-19b

expression levels to GAPDH levels expression 0 Control Stage I Stage II Metastatic Representative images of three biologic replicates. Bottom panel

Lgr6 shows mRNA levels of p21, a direct D Lgr6 miR-19 C Stage I Stage II Metastatic target of p53, used to test p53 activity. C, Western blot analysis shows the α p38 CTL TUM CTL TUMCTL TUM expression of p38a in healthy and Tubulin tumor tissue from patients at different Tubulin stages of lung adenocarcinoma, with metastatic referring to an invasive α p38 primary tumor. The numbers 1 1 0.68 1 0.59 1 0.30 represent the relative expression in 0.8 each tumor compared with its

0.6 healthy counterpart tissue. One + – Lgr6 Lgr6 representative experiment out of 5 is 0.4 E Lung Tumor Tumor depicted. D, expression of p38a in 0.2 normal Lgr6þ lung cells, with or

relative protein levels p38 α relative Tubulin 0 without overexpression of the miR-17- Lgr6 92 member miR-19. The graph shows the relative levels of p38a SEM P-GSK3 (N ¼ 5). E, protein expression of Lgr6 þ p38α and p38a in normal lung or Lgr6 or 1 0.2 0.59 Lgr6 cells from stage II lung F adenocarcinoma. The numbers are 6 relative to the lung control and normalized with tubulin levels. One representative experiment of three 4 biologic replicates is presented. F, graph shows qPCR normalized Stage I numbers representing miR-17-92 2 tumor cells

– family member expression in FACS Stage IIa sorted Lgr6þ compared with Lgr6 cells from early (stage I) or later stages 1 Stage IIb

vs Lgr6 (stage II) of lung adenocarcinoma. + Results are the mean, SD, of three

Lgr6 –2 independent experiments with ﬁve Relative expression levels in levels expression Relative biologic replicates per stage. Signiﬁcant differences are marked –4 (, P 0.05).

samples revealed an increase of the proliferation marker Ki67 at addition, miR-17-92 has been reported to be overexpressed in later stages (Fig. 1D). Analysis by TUNEL staining could not detect lung tumors (21, 22). Analysis of miR-17-92 expression by qPCR þ any difference in apoptosis at any stage (data not shown). Ki67 in tumors from different adenocarcinoma stages confirmed the þ cells at stage II are mostly Lgr6 cells (Fig. 1E). Thus, there is a increased expression of the cluster members during cancer pro- proliferative advantage that may contribute to specifically select gression (Fig. 2A). We have previously shown that expression of Lgr6 cells during adenocarcinoma progression. the miR-17-92 cluster is transcriptionally repressed by p53 (16). Analysis of p53 by IHC showed that nuclear p53 is lost during Selection of Lgr6 cells with a disrupted p38a/miR-17-92 axis adenocarcinoma progression correlating with higher miR-17-92 þ We have previously shown that human lung Lgr6 stem cells levels (Fig. 2B). In addition, enhanced miR-17-92 expression are regulated by a fine-tuned regulated cross-talk between the correlated with decreased levels of p38a in late-stage adenocarci- p38a MAPK pathway and the miRNA cluster miR-17-92 (16). In nomas, when compared with their early-stage counterparts

4016 Cancer Res; 76(13) July 1, 2016 Cancer Research

Lgr6 Labels Malignant Human Adenocarcinoma Cells

(Fig. 2C). p38a protein, but not mRNA expression, has been mentary Fig. S1B) and epithelial markers (Supplementary Fig. þ shown to be reduced in human lung tumors (20). Hence, down- S4A–S4C), while increasing Lgr6 cell numbers in later malignant regulation of p38a must be a posttranscriptional event. Using stages (Supplementary Fig. S4E). public databases (www.targetscan.org), we noted that p38a was To test some of the tumorigenic and stem-like properties of the þ a putative target for one of the miR-17-92 members, miR-19. Lgr6 cells, we sorted cells from different stages of lung adeno- þ þ Using a lentiviral vector in cultured lung Lgr6 cells, we observed carcinoma. Both Lgr6 and Lgr6 tumor cells express lung mar- that miR-19 overexpression induces proliferation (Supplementa- kers at similar levels lung markers (Fig. 4A). Interestingly, only the þ ry Fig. S2A and S2B). miR-19–specific targeting of p38a 30-UTR marker AQP5 (aquaporin 5) is overexpressed in Lgr6 tumor cells. was shown using a luciferase reporter assay (Supplementary This protein is a marker for lung cancer malignancy and has been Fig. S2C), and results in decreased p38a protein (Fig. 2D). This reported that overexpression links lung cancer with tumor inva- is a previously unreported negative cross-talk regulation and it sion, proliferation, and metastatic features (30, 31). creates a negative-feedback loop between the miR-17-92 and In culture, cells from later stages showed an increase in prolif- p38a pathways (Supplementary Fig. S2D). eration, expanding at higher rates (Supplementary Fig. S4D). As þ þ Repression of p38a is enhanced in Lgr6 cells from stage II stage II seemed to be the switching point where Lgr6 cells start to tumors, when compared with counterpart Lgr6 cells (Fig. 2E). prevail, we used tumor cells from adenocarcinomas at this stage to þ þ Reduced p38a protein in Lgr6 correlates with a change in the compare the properties of Lgr6 and Lgr6 cells. A widely used in expression of miR-17-92 members, larger than its increase in vitro assay to test the tumorigenic capacity and self-renewal Lgr6 cells, during cancer progression (Fig. 2F). potential of cancer cells is colony-forming potential. Freshly þ isolated Lgr6 and Lgr6 cells from stage II adenocarcinomas þ p38 regulates the expression of Wnt signaling components were seeded at low confluency (105). Only Lgr6 cells produced Wnt signaling is a key pathway related to cancer development visible macroscopic colonies, confirming a higher tumorigenic (23). Inhibitory phosphorylation of GSK3 and nuclear translo- and self-renewal capacity (Fig. 4B). þ cation of b-catenin are hallmarks of canonical Wnt signaling and Lgr6 and Lgr6 cells from stage II adenocarcinoma were tested are present in many cancers (24, 25). We observed increased P- in vivo by intravenous injection into immunodeficient mice. This GSK3 levels with increasing lung adenocarcinoma stage (Fig. 3A). procedure has been extensively used to deliver cancer cells into the Activation of the canonical Wnt pathway and enhanced signal was lungs, allowing the secondary tumor potentiality of the cells to be confirmed by IHC, showing specific P-GSK3 and nuclear b-catenin investigated. Similar to the in vitro assays, we only observed tumor þ induction in late-stage lung adenocarcinomas (Fig. 3B). Further burdens in the lungs of mice injected with Lgr6 cells. We failed to confirmation was established by qPCR analysis of mRNA expres- detect any tumors from Lgr6 cells. Newly formed lung tumors þ sion of Wnt components with elevated levels of enhancers (Lrp6) from Lgr6 cells were usually located close to blood vessels and and targets (Lef1, Tcf4, Axin2) and reduced levels of Wnt repres- their human lung adenocarcinoma origin was confirmed by sors (Wif1, Gremlin, Dkk1, Dab2) in late stages (Fig. 3C; Table 1). staining with TTF1 and a human-specific mitochondrial antibody þ Lgr6 is a promoter of Wnt receptor signaling. The coactivator (Fig. 4C). Interestingly, stage II Lgr6 tumor cells retained differ- Lrp6 interacts with Lgr6 enhancing the response to Wnt ligands, entiation potential. The tumor masses formed from injected þ amplifying the receptors signal (26, 27). Activation in of Wnt Lgr6 human tumor cells were only partially positive for Lgr6, signaling was analyzed in Lgr6 tumor cells. qPCR analysis of Wnt whereas they expressed lung-specific markers such as SP-C or CC- þ components in Lgr6 or Lgr6 cells from stage II or stage III 10 (Fig. 4D). This confirms a putative cancer stem cell role for þ adenocarcinomas showed enhanced Wnt signaling in later stage Lgr6 human lung adenocarcinoma cells. We recapitulated the þ þ adenocarcinomas (Fig. 3D and Supplementary Fig. S3). tumors formed by Lgr6 cancer cells using normal Lgr6 lung cells Using inactivating phospho-GSK3 (Ser9) as a sensor of Wnt transformed by overexpression of oncogenic K-RasG12D. These þ signal, we observed that normal lung Lgr6 cells, deficient in p38a cells coexpressed GFP and could be visually traced (green) in the (p38KD), showed an increase in P-GSK3. This is independent lung tumors (Supplementary Fig. S5A). Histologic analysis þ of miR-17-92, as combined downregulation of this pathway showed features similar to those from Lgr6 adenocarcinoma (p38KD/shmiR) failed to rescue Wnt activation (Fig. 3E). Loss cells (Supplementary Fig. S5B), and they expressed GFP and the of p38a modified the expression of other components of the Wnt adenocarcinoma markers TTF1 and CK7 but not the squamous þ pathway. Interestingly, p38a-deficient Lgr6 lung cells showed an carcinoma marker p63 (Supplementary Fig. S5C). Immunofluo- increased expression of the Wnt coreceptor Lrp6 and decreased of rescent staining confirmed the human origin of the tumors þ the Wnt antagonist DKK1 (Fig. 3F). Inactivation of GSK3 corre- formed by Lgr6 transformed lung cells and the adenocarcinoma lates with augmented expression of genes regulated by canonical features, as they expressed TTF1 (Supplementary Fig. S5D). þ Wnt, confirming the enhanced activation of the pathway in Lgr6 Then in vivo tumor-colonizing potential of human cancer cells cells lacking p38a (Fig. 3G). The specific role of p38a as a negative was tested using subcutaneous injections into immunodeficient þ regulator of canonical Wnt signaling in Lgr6 cells was confirmed mice. This method assesses the capacity of the tumor cells to grow þ by overexpression of p38a in Lgr6 /p38KD cells, restoring the new burdens in an alien environment, one of the hallmarks of þ P-GSK3 to control levels (Fig. 3H). Partial rescue, by p38a over- metastasis. We isolated Lgr6 and Lgr6 tumor cells from early expression, was confirmed by qPCR analysis with significant (stage I), medium (stage II), and late (metastatic) human lung reduction of Wnt targets and increase in Wnt repressors (Fig. 3I). adenocarcinomas. Freshly isolated cells from single tumors were injected in parallel in both flanks of single mice (Fig. 5A). As has Selected Lgr6 cells have cancer stem cell–like properties been previously observed in intravenous injections, Lgr6 cells þ þ Activation of Wnt signal may account for maintaining Lgr6 failed to form subcutaneous tumors. Tumors from Lgr6 cells cells in an undifferentiated state while retaining stem cell features were histologically analyzed (Fig. 5A). The human lung adeno- (28, 29). Lung adenocarcinomas showed a loss of lung (Supple- carcinoma origin of the resected tumors was confirmed, with

www.aacrjournals.org Cancer Res; 76(13) July 1, 2016 4017

Guinot et al.

P-GSK3 β−Catenin ABStage I Stage II Metastatic

Tubulin Lung

pGSK3

1 2.1 5.2 C 4 Stage I 3.5 Control 3

2.5 Early stage

2 Late stage Stage II 1.5

0.5 mRNA Relative levels to HPRT levels mRNA Relative 0 Lrp6 Lef1 Tcf4 Axin2 Dkk1 Gremlin Dab2 Wif1 Metastatic

D 4 Lgr6– cells 3.5 Lgr6+ cells 3 Lgr6+ Lgr6+ p38KD 2.5 E + Lgr6 p38KD shmiR 2 1.5 P-GSK3 1 Tubulin 0.5 mRNA Relative levels to HPRT levels mRNA Relative 0 2.5 Lrp6 Lef1 Axin2 Tcf4 Wif1 Gremlin Dkk1 Dab2 F Lgr6+ G 3 2 Lgr6+/ p38KD

2.5 Lgr6+ 1.5 + 2 Lgr6 / p38KD 1 1.5 0.5 1

0 0.5 Relative expression levels to GAPDH levels expression Relative LRP6 Fzd10 Wifi1 Dkk1 Gremlin

0 I Lef1 Dab2 Axin2 Tcf4 Relative expression levels to GAPDH levels expression Relative 2.5 Control p38α H p38KD Control p38KD rescue 2 p38α α−Tub 1.5 rescue

p38α 1

P-GSK3 0.5 0.1 1 0.4 0 Relative expression levels to GAPDH levels expression Relative Lrp6 Lef1 Axin2 Tcf4 Dab2 Dkk1 Wif1 Gremlin

Figure 3. p38a reduction mediates Wnt signaling enhancement during adenocarcinoma progression and selection of Lgr6þ tumor cells. A, Western blot analysis showing activation of Wnt canonical pathway (P-GSK3) during adenocarcinoma development. Representative image of three independent replicates. B, immunohistochemical analysis of canonical Wnt activation at early and late stages of lung adenocarcinoma. Representative images of five biologic replicates per stage. C, graphs show mRNA relative levels of Wnt components, targets, and repressors in normal lung (white bars), early (black) or late (gray) stages of adenocarcinoma. Results are the mean, SD, of three independent experiments with five biologic replicates per stage. Significant differences are marked (, P 0.05). D, relative mRNA expression in Lgr6 and Lgr6þ cells from stage II tumors. Results are the mean, SD, of 5 tumor samples in triplicates. Significant differences are marked (, P 0.05). E, Western blot anlaysis showing P-GSK3 levels in normal lung Lgr6þ cells or cells lacking p38a (p38KD), or cells combining lack of p38a and miR-17-92 downregulation (p38KDþshmiR). Representative image of three independent replicates. F, relative mRNA expression of Wnt components in Lgr6þ or Lgr6þ cells lacking p38a (p38KD). G, mRNA expression by qPCR of Wnt targets in Lgr6þ or Lgr6þ/p38KD cells. H, overexpression of p38a in Lgr6þ cells deficient in p38a restores pGSK3 levels. I, p38a rescue in Lgr6þ/p38KD cells restores Wnt components and Wnt targets mRNA expression. All results are the mean of three to five experiments, SD.

4018 Cancer Res; 76(13) July 1, 2016 Cancer Research

Lgr6 Labels Malignant Human Adenocarcinoma Cells

A 4 Lgr6– Lgr6+ 3

Figure 4. 1 Stage II Lgr6þ tumor cells are more tumorigenic than Lgr6 cells and

mRNA Relative levels to HPRT levels mRNA Relative 0 retain self-renewal and differentiation p38α Lgr6 SP-C CC-10 AQP5 Muc5ac potential. A, cells from tumor masses at different stages of lung B Lgr6– Lgr6+ adenocarcinoma were sorted into Lgr6þ and Lgr6 populations and analyzed for mRNA expression of Lgr6, p38a, and lung-specific markers. Results are the mean, SD, of three tumor samples in triplicates. B, tumor cells from stage II adenocarcinoma were sorted for Lgr6þ and Lgr6 populations and seeded in soft agar to test for tumorigenic potential. C Colonies that formed were stained with crystal violet. C, Lgr6þ and Lgr6 cells from stage II adenocarcinomas were injected (104 cells) intravenously in SCID mice (n ¼ 4). Only Lgr6þ cells produced tumors. The images show H&E, TTF1 IHC, and staining of paraffin sections with a human mitochondrial antibody (green) and TTF1 (red). Representative images of three D biologic replicates. D, intravenously injected Lgr6þ cells were detected in the lungs of SCID mice and stained with human specific mitochondrial antibody and Lgr6 or lung-specific markers (SP-C, CC-10). Representative images of three biologic replicates.

expression of TTF1 and human mitochondrial antigen evident Discussion (Fig. 5B). Further analysis by immnuoﬂuorescent staining Despite all of the advances in unraveling the molecular and revealed the loss of epithelial (E-Cad) and the gaining of stem biologic initiation and progression of cancer (32, 33), the knowl- cell (Lgr6) markers during the progression from early- to late-stage edge of these processes in lung adenocarcinoma is still poor, disease (Fig. 5C). Lung differentiation was also affected. Early- resulting in little improvement in patient survival (5). stage tumors showed expression of alveolar (SP-C) and bronchi- In this article, we revealed the potential of a population of olar (CC-10) markers, with only marginal expression of these human lung adenocarcinoma cells, characterized for expressing markers in late tumors (Fig. 5C). The lack of cellular differenti- Lgr6, that are selected through the progression of the cancer and ation within the tumor correlated with the observed increase in þ that harbor cancer stem cell–like properties. As a consequence of Wnt signaling and deﬁciency in p38a observed in Lgr6 cells. p53-mutating inactivation, there is a dysregulation of the balance Loss of p38a during lung adenocarcinoma progression favors þ between two pathways, the miR-17-92 family of miRNAs and the the selection of Lgr6 stem cell–like cells with reduced differen- þ p38a kinase, that control normal homeostasis in human Lgr6 tiation potential while maintaining their colonizing and self- lung stem cells. Lack of p53 leads to loss of its repressive effects on renewal capabilities. This changes result in tumors containing miR-17-92 cluster expression. miR-17-92 overexpression is a poorly differentiated cells with a high metastatic potential.

www.aacrjournals.org Cancer Res; 76(13) July 1, 2016 4019

Guinot et al.

A Lgr6+ Lgr6– Lgr6+ Flank

12 weeks B

C Stage I Stage II Metastatic

Figure 5. Lgr6þ adenocarcinoma cells retain tumorigenic potential and loss of differentiation capability due to increased Wnt signaling during cancer development. A, Lgr6þ and Lgr6 tumor cells from different stages of lung adenocarcinoma were injected (103–104 cells) in parallel flanks of SCID mice (n ¼ 10). Histologic analysis show a tumor formed from stage II Lgr6þ cells. B, immunofluorescence analysis of the tumors with TTF1 and human mitochondrial antibodies confirmed their human adenocarcinoma origin. C, histologic and immunofluorescence images of tumors from early, medium, and late adenocarcinoma stages. The tissues were stained with antibodies to detect human (hMito), epithelial (E-Cad), stem (Lgr6), and lung-specific (SP-C, CC-10) markers. Representative images of eight biologic replicates.

4020 Cancer Res; 76(13) July 1, 2016 Cancer Research

Lgr6 Labels Malignant Human Adenocarcinoma Cells

common feature in lung cancer. We have now shown that a miR- of cells with characteristics of cancer stem cells, higher potential to 17-92 member, miR-19, downregulates p38a. form new tumors, capacity to establish in other tissues, and rapid We have previously reported a deficiency in p38a protein in proliferation. human lung cancer (20). There is some evidence of an inhibitory In this article, we are contributing to the elucidation of mechan- role of p38a on Wnt signaling (31, 34). Here we have found that isms that promote malignancy throughout lung adenocarcinoma p38a deficiency results in a wide-ranging dysregulation of Wnt development. A coadjuvant process, combined with the side effect components, repression of Wnt inhibitors, and promotion of its of oncogenic mutations, allows the selection of a population of enhancers. cells with features of cancer stem cells, providing more malignant Wnt signaling is a regulator of stem cell self-renewal and a properties and advancing tumor progression. These findings add signature pathway in many cancers (15). One of the components new potential cellular and molecular targets for detection, prog- overexpressed in the absence of p38a is the low-density lipopro- nosis, and therapies to tackle this devastating disease. tein-related receptor 6 (LRP6), which plays a central role in canonical Wnt signal (26, 35). We have demonstrated the spec- Disclosure of Potential Conflicts of Interest fi a i city of this p38 function, as it can be rescued by recovering No potential conflicts of interest were disclosed. p38a expression. LRP6 function can be enhanced by interacting with R-spondin receptors like Lgr6 (27). Cells expressing Lgr6 can Authors' Contributions respond to Wnt activation, enhancing the canonical signal and Conception and design: F. Oeztuerk-Winder, J.-J. Ventura hence promoting growth. þ Development of methodology: J.-J. Ventura Lgr6 adenocarcinoma cells proliferate more than Lgr6 as the Acquisition of data (provided animals, acquired and managed patients, cancer progresses. From early to late stages, there is an increase in provided facilities, etc.): A. Guinot miR-17-92 levels that diminishes p38a, allowing further activa- Analysis and interpretation of data (e.g., statistical analysis, biostatistics, þ tion of Wnt signaling. Lgr6 cells are then selected, as they have a computational analysis): J.-J. Ventura Writing, review, and/or revision of the manuscript: J.-J. Ventura proliferative advantage over the cells lacking this R-spondin Administrative, technical, or material support (i.e., reporting or organizing receptor. data, constructing databases): A. Guinot Similar to other related R-spondins (e.g., Lgr5), Lgr6 is also Study supervision: F. Oeztuerk-Winder, J.-J. Ventura considered a marker for adult stem cells and a potential marker for þ cancer stem cells (27). We now show that Lgr6 cells from Acknowledgments different stages of adenocarcinoma have higher tumorigenic and The authors thank P. Humphreys, J. Soffe, H. Skelton, J. Brown, and þ self-renewal potential than Lgr6 cells. Lgr6 cells from early A. Riddell for technical help. stages still maintain their capacity to differentiate into bronchioalveolar lineages, but this functionality is lost in later stages. This Grant Support is likely due to enhanced Wnt signaling, that promotes self- This work was supported by grants from the Medical Research Council (MRC; RG51968/RG57589) and Cancer Research UK (RG52191). A. Guinot was renewal, and deficient p38a activity, that controls differentiation þ supported by a Herchel-Smith PhD student fellowship. of lung stem cells. As a consequence, Lgr6 accumulate in later The costs of publication of this article were defrayed in part by the payment of stages, retaining cancer stem cell features, with active self-renewal page charges. This article must therefore be hereby marked advertisement in and poor differentiation, both signatures of advanced stages in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. cancer. We have therefore found a coadjuvant mechanism, activated Received December 7, 2015; revised March 24, 2016; accepted April 27, 2016; through human lung adenocarcinoma, which allows the selection published OnlineFirst May 4, 2016.

References 1. Jemal A, Grauman D, Devesa S. Recent geographic patterns of lung cancer 9. Bonnet D, Dick JE. Human acute myeloid leukemia is organized as a and mesothelioma mortality rates in 49 shipyard counties in the United hierarchy that originates from a primitive hematopoietic cell. Nat Med States, 1970–94. Am J Ind Med 2000;37:512–1. 1997;3:730–7. 2. Carney DN.Lung cancer–time to move on from chemotherapy. N Engl J 10. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF. Med 2002;346:126–8. Prospective identiﬁcation of tumorigenic breast cancer cells. Proc Natl 3. Chute JP, Chen T, Feigal E, Simon R, Johnson BE. Twenty years of phase III Acad Sci U S A 2003;100:3983–8. trials for patients with extensive-stage small-cell lung cancer: perceptible 11. Cho RW, Wang X, Diehn M, Shedden K, Chen GY, Sherlock G, et al. progress. J Clin Oncol 1999;17:1794–801. Isolation and molecular characterization of cancer stem cells in MMTV- 4. Mulshine JL, Tockman MS, Smart CR. Considerations in the development Wnt-1 murine breast tumors. Stem Cells 2008;26:364–71. of lung cancer screening tools. J Natl Cancer Inst 1989;81:900–6. 12. Lobo NA, Shimono Y, Qian D, Clarke MF. The biology of cancer stem cells. 5. Mulshine JL, Hirsch FR. Lung cancer chemoprevention: moving from Annu Rev Cell Dev Biol 2007;23:675–99. concept to a reality. Lung Cancer 2003;41:S163–74. 13. Singh A, Settleman J. EMT, cancer stem cells and drug resistance: an 6. Strauss GM. The Mayo Lung Cohort: a regression analysis focusing emerging axis of evil in the war on cancer. Oncogene 2010;29:4741–51. on lung cancer incidence and mortality. J Clin Oncol 2002;20: 14. Gibbons DL, Lin W, Creighton CJ, Zheng S, Berel D, Yang Y, et al. 1973–83. Expression signatures of metastatic capacity in a genetic mouse model of 7. Bayman N, Blackhall F, McCloskey P, Taylor P, Faivre-Finn C. How can lung adenocarcinoma. PLoS One 2009;4:e5401. we optimise concurrent chemoradiotherapy for inoperable stage III 15. Schuijers J, Clevers H. Adult mammalian stem cells: the role of Wnt, Lgr5 non-small cell lung cancer? Lung Cancer 2014;83:117–25. and R-spondins. EMBO J 2012;31:2685–96. 8. Freitas DP, Teixeira CA, Santos-Silva F, Vasconcelos MH, Almeida GM. 16. Oeztuerk-Winder F, Guinot A, Ochalek A, Ventura JJ. Regulation of human Therapy-induced enrichment of putative lung cancer stem-like cells. Int J lung alveolar multipotent cells by a novel p38alpha MAPK/miR-17–92 Cancer 2014;134:1270–8. axis. EMBO J 2012;31:3431–41.

www.aacrjournals.org Cancer Res; 76(13) July 1, 2016 4021

Guinot et al.

17. Cabrita MA, Christofori G. Sprouty proteins, masterminds of receptor 26. Joiner DM, Ke J, Zhong Z, Xu HE, Williams BO. LRP5 and LRP6 in tyrosine kinase signaling. Angiogenesis 2008;11:53–62. development and disease. Trends Endocrinol Metab 2013;24:31–9. 18. Wertz IE, Dixit VM. Regulation of death receptor signaling by the ubiquitin 27. Leushacke M, Barker N. Lgr5 and Lgr6 as markers to study adult stem cell system. Cell Death Differ 2010;17:14–24. roles in self-renewal and cancer. Oncogene 2012;31:3009–22. 19. Jackson EL, Willis N, Mercer K, Bronson RT, Crowley D, Montoya R, 28. van de Wetering M, de Lau W, Clevers H. WNT signaling and lymphocyte et al. Analysis of lung tumor initiation and progression using development. Cell 2002;109:S13–9. conditional expression of oncogenic K-ras. Genes Dev 2001;15: 29. Pinto D, Gregorieff A, Begthel H, Clevers H. Canonical Wnt signals are 3243–8. essential for homeostasis of the intestinal epithelium. Genes Dev 2003;17: 20. Ventura JJ, Tenbaum S, Perdiguero E, Huth M, Guerra C, Barbacid M, 1709–13. et al. p38alpha MAP kinase is essential in lung stem and pro- 30. Chae YK, Woo J, Kim MJ, Kang SK, Kim MS, Lee J, et al. Expression of genitor cell proliferation and differentiation. Nat Genet 2007;39: aquaporin 5 (AQP5) promotes tumor invasion in human non small cell 750–8. lung cancer. PLoS One 2008;3:e2162. 21. Takakura S, Mitsutake N, Nakashima M, Namba H, Saenko VA, Rogou- 31. He J, Liu Z, Zheng Y, Qian J, Li H, Lu Y, et al. p38 MAPK in myeloma cells novitch TI, et al. Oncogenic role of miR-17–92 cluster in anaplastic thyroid regulates osteoclast and osteoblast activity and induces bone destruction. cancer cells. Cancer Sci 2008;99:1147–54. Cancer Res 2012;72:6393–02. 22. Osada H, Takahashi T. let-7 and miR-17–92: small-sized major players in 32. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100:57–70. lung cancer development. Cancer Sci 2011;102:9–17. 33. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. Cell 23. Cleary AS, Leonard TL, Gestl SA, Gunther EJ. Tumour cell heterogeneity 2011;144:646–74. maintained by cooperating subclones in Wnt-driven mammary cancers. 34. Verkaar F, van der Doelen AA, Smits JF, Blankesteijn WM, Zaman GJ. Nature 2014;508:113–7. Inhibition of Wnt/beta-catenin signaling by p38 MAP kinase inhibitors is 24. Nusse R, Varmus H. Three decades of Wnts: a personal perspective on how a explained by cross-reactivity with casein kinase Idelta/varepsilon. Chem scientiﬁc ﬁeld developed. EMBO J 2012;31:2670–84. Biol 2011;18:485–94. 25. Polakis P, Hart M, Rubinfeld B. Defects in the regulation of beta-catenin in 35. Niehrs C.The complex world of WNT receptor signalling. Nat Rev Mol Cell colorectal cancer. Adv Exp Med Biol 1999;470:23–32. Biol 2012;13:767–79.

4022 Cancer Res; 76(13) July 1, 2016 Cancer Research

miR-17-92/p38α Dysregulation Enhances Wnt Signaling and Selects Lgr6 + Cancer Stem-like Cells during Lung Adenocarcinoma Progression

Anna Guinot, Feride Oeztuerk-Winder and Juan-Jose Ventura

Cancer Res 2016;76:4012-4022. Published OnlineFirst May 4, 2016.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-15-3302

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2016/05/04/0008-5472.CAN-15-3302.DC1

Cited articles This article cites 35 articles, 9 of which you can access for free at: http://cancerres.aacrjournals.org/content/76/13/4012.full#ref-list-1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/76/13/4012. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.