Open Full Page

Published OnlineFirst August 25, 2009; DOI: 10.1158/0008-5472.CAN-09-0577

Cell, Tumor, and Stem Cell Biology

Matrix Metalloproteinase 12 Overexpression in Lung Epithelial Cells Plays a Key Role in Emphysema to Lung Bronchioalveolar Adenocarcinoma Transition

Peng Qu,1,2 Hong Du,3 Xi Wang,3 and Cong Yan1,2

1The Center for Immunobiology, Simon Cancer Center and 2Department of Pathology and Laboratory Medicine, Indiana University School of Medicine, Indianapolis, Indiana and 3Division of Human Genetics, Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio

Abstract extracellular matrix (ECM) and basement membrane components Chronic obstructive pulmonary disease (COPD) and lung under physiologic conditions. MMPs play very important roles in cancer are two diseases that are related to smoking in normal connective tissue turnover during morphogenesis, tissue development, wound healing, and reproduction. MMPsalsoact as humans. The molecular mechanism linking these two diseases modulatorsof inflammation and innate immunity by activating, is poorly understood. Matrix metalloproteinase 12 (MMP12) is deactivating, or modifying the activity of signaling cytokines, a member of the MMP family, which can be induced by chemokines, and receptors (2, 3). In oncology, MMPs have long smoking. Because MMP12 overexpression in epithelial cells been considered as molecules necessary to promote tumor has been reported in inflammation-triggered lung remodeling, invasion and metastasis through the degradation of the ECM a murine CCSP-rtTA/(tetO) -MMP12 bitransgenic model was 7 (4, 5). Nevertheless, their roles in directly initiating and inducing created. In this model, MMP12-Flag fusion protein over- tumor have never been reported. MMP12 isa 22-kDa metal- expression and its increased enzymatic activity were observed dependent proteinase that was first detected by Werb and Gordon in the lung in an inducible manner, which led to inflammatory in 1975 (6). It can degrade elastin and other substrates, such as type cell infiltration and increased epithelial growth. In sequential IV collagen, fibronectin, laminin, gelatin, vitronectin, entactin, events, spontaneous emphysema and bronchioalveolar ade- heparin, and chondroitin sulfates (7). In the lung, MMP12 is nocarcinoma were developed as a result of MMP12 over- identified in alveolar macrophagesof cigarette smokersasan expression. During this process, the concentration of elastolytic MMP (8). Inactivation of the MMP12 gene in MMP12 interleukin-6 was steadily increased in bronchioalveolar knockout mice suggests that MMP12 plays a critical role in lavage fluid, which activated the oncogenic signal transducer smoking-induced COPD (9). The clinical relevance of MMP12 in and activator of transcription 3 (Stat3) in alveolar type II non–small cell lung cancer (NSCLC) had been studied, in which epithelial cells. Expression of Stat3 downstream genes that are MMP12 correlateswith early cancer-related deathsin NSCLC, known to stimulate inflammation and tumor formation was especially for those associated with tobacco cigarette smoke significantly increased in the lung. When tested in humans, exposure (10). It has been reported that the MMP1-MMP3- MMP12 up-regulation was highly associated with COPD and MMP12 gene clusterplaysimportant rolesin lung cancer lung cancer in patients. Together, these studies support that development and progression (11). Studies using comparative MMP12 is a potent proinflammatory and oncogenic molecule. genomic hybridization analysis obtained a high-resolution map of MMP12 up-regulation plays a critical role in emphysema to frequent chromosomal gains and losses associated with lung lung cancer transition that is facilitated by inflammation. cancer. An amplified MMP cluster region (11q22) with overex- [Cancer Res2009;69(18):7252–61] pressed MMP1, MMP12, and MMP13 was identified (12). Although these studies showed association of MMP12 overexpression with Introduction lung cancer, the role of MMP12 up-regulation in lung cancer asa Smoking leads to chronic obstructive pulmonary disease (COPD; causer remains to be defined. In addition to macrophages, MMP12 the major phenotype isemphysema) and lung cancer, which are is overexpressed in lung epithelial cells (13, 14). During lysosomal associated with pulmonary inflammation. Human COPD patients acid lipase deficiency in the lung, blockage of cholesteryl esters and (especially with smoking history) are a high-risk population for triglyceridesto free cholesteroland free fatty acidstriggered developing lung cancer. Even after having given up smoking, lung pulmonary inflammation, emphysema, and hypercellularity À/À inflammation persists and progresses in humans with COPD (1). (14–18). MMP12 was highly overexpressed (100-fold) in the lal The molecular mechanism that links COPD and lung cancer is lung asdetermined by the Affymetrix GeneChip microarray poorly understood. analysis. Expression of the MMP12 gene isdown-regulated by lipid Matrix metalloproteinases (MMPs) are a family of >20 secreted mediatorsand anti-inflammatory peroxisomeproliferator-activat- or transmembrane proteins that are capable of degrading ed receptor g (PPARg; ref. 14). The rolesof inflammation, tumor microenvironment, and ECM remodeling during tumorigenesis are complex, as multiple cell Note: Supplementary data for thisarticle are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/). typesare involved in intricate cross-talkthat isdifficult to Requests for reprints: Cong Yan, The Center for Immunobiology and Department recapitulate in vitro. Thus, to understand the role of MMP12 in of Pathology and Laboratory Medicine, Indiana University School of Medicine, Walther Hall C418, 980 W. Walnut Street, Indianapolis, IN 46202-5188. Phone: 317-278-6005; the process of emphysema to tumor transition and tumor- Fax: 317-278-7030; E-mail: [email protected] or Hong Du, Division of Human Genetics, associated inflammation, conditional overexpression of MMP12 Cincinnati Children’sHospitalMedical Center, 3333 Burnet Avenue, Cincinnati, OH in lung epithelial cells was created by using the reverse tetracycline 45229-3039. Phone: 513-636-7136; E-mail: [email protected]. I2009 American Association for Cancer Research. transactivator (rtTA) that is under the control of the Clara cell doi:10.1158/0008-5472.CAN-09-0577 secretory protein (CCSP) promoter.

Cancer Res 2009; 69: (18). September 15, 2009 7252 www.aacrjournals.org

MMP12 Causes Inflammation and Lung Cancer

Materials and Methods CMV-MMP12 transgenic mouse line. In this founder line, a Flag Animal care. The animal care was the same as previously described (19). sequence was added at the COOH terminus of the MMP12 cDNA to Generation of doxycycline-controlled MMP12 transgenic mice. To distinguish it from the endogenous MMP12 molecule. CCSP-rtTA/ generate the (tetO)7-CMV-MMP12 transgenic mouse line, MMP12 cDNA (tetO)7-CMV-MMP12 bitransgenic mice were identified by PCR wasamplified by PCR usinga downstream primer (5 ¶-AAGGAAAAA- using specific primers and purified mouse tail DNA (Supplemen- AGCGGCCGCTGATCACTTGTCATCGTCGTCCTTGTAGTCACAACCAAAC- tary Fig. S1). Induction of MMP12 mRNA expression in AT II CAGCTTGTAC-3¶) that contains a Flag sequence (underlined), a stop codon, epithelial cellswasachieved by doxycycline treatment as and a NotI site and an upstream primer (5¶-AGCGGATCCGCCACCAT- monitored by real-time PCR assay. WT and MMP12 single ¶ Bam GAAATTTCTCATGATG-3 ) that containsa Kozak sequenceand a HI transgenic mice were also treated as controls with no MMP12 Bam Not site. The PCR product was digested with HI/ I and subcloned mRNA induction observed (Fig. 1A). By FACS analysis, expression downstream of the (tetO) -CMV minimal promoter linked to seven Tet- 7 of the MMP12-Flag fusion protein was also significantly induced in responsive elements at the BamHI and NotI sites in the pTRE2 vector (Clontech). The expression cassette, containing the CMV minimal promoter, AT II epithelial cellsascostained with Flag antibody and SP-C the MMP12 cDNA, and the human globin polyadenylation signaling antibody (AT II cell marker) in doxycycline-treated bitransgenic sequence, was dissected out and purified for microinjection into fertilized mice (Fig. 1B). No MMP12-Flag fusion protein was detected in eggs of FVB/N mice by the Transgenic Core Facility at the University of Clara cellsand alveolar macrophagesascostainedwith Flag Cincinnati College of Medicine. Founder lineswere identified by a pair of antibody and CCSP (Clara cell marker) or CD11b antibody primerscorrespondingto a pTRE2 plasmidsequence(5 ¶-CCATC- (macrophage marker) regardless of doxycycline treatment. To CACGCTGTTTTGACC-3¶) and a MMP12 cDNA coding region sequence confirm that overexpressed MMP12-Flag fusion protein was ¶ ¶ (5 -GACTTGAGTTGTCCAGTTGCCC-3 ). The CCSP-rtTA transgenic mice secreted into alveolar lumen and had the enzymatic activity, BALF were genotyped with an upstream primer corresponding to the CCSP from doxycycline-treated (n = 3) and untreated (n = 4) bitransgenic promoter sequence (5¶-ACTGCCCATTGCCCAAACAC-3¶) and a downstream mice wascollected and incubated with a fluorescence-quenched primer corresponding to the rtTA coding region sequence (5¶-AAAAT- CTTGCCAGCTTTCCCC-3¶). CCSP-rtTA/(tetO) -CMV-MMP12 double trans- MMP12 substrate. Compared with the doxycycline-untreated BALF 7 F A genic mice were generated from cross-breeding of CCSP-rtTA transgenic samples (230 26 ng/ L), the enzymatic product of MMP12 was F mice and (tetO)7-CMV-MMP12 transgenic mice. 10 timeshigher in the doxycycline-treated BALF samples(2,270 MMP12 activity. The MMP12-specific activity from bronchioalveolar 120 ng/AL; Fig. 1C). The MMP12 enzymatic activity remained lower lavage fluid (BALF) was measured by the SensoLyte 490 MMP-12 Assay kit in other controlsaswell, including doxycycline-treated (tetO) 7- (AnaSpec). CMV-MMP12 single transgenic mice (210 F 55 ng/AL; n = 3) and Histology, tumor incidence, and MetaMorph quantification of WT mice (135 F 30 ng/AL; n = 4). Immunohistochemical staining emphysema. Histology, tumor incidence, and MetaMorph quantification revealed MMP12 overexpression on the alveolar epithelial wall of emphysema were the same as described previously (20, 21). and hyperplasia cells in the doxycycline-treated bitransgenic lung Fluorescence-activated cell sorting analysis. Fluorescence-activated (Fig. 1D). cell sorting (FACS) analysis was the same as previously described (19). Reverse transcription and real-time PCR. Reverse transcription and Emphysema and bronchioalveolar adenocarcinoma in the real-time PCR were the same as previously described (22, 23). Human lung of bitransgenic mice. To assess the pathophysiologic MMP12 primerswere asfollows:upstream,5 ¶-GATGCACGCACCTCGATGT- consequence of MMP12 overexpression in lung epithelial cells, 3¶; downstream, 5¶-GGCCCCCCTGGCATT-3¶. bitransgenic mice were treated with doxycycline for various time Annexin V binding and bromodeoxyuridine proliferation analyses. lengths. Histopathologic analyses revealed lung abnormalities in Previousprocedureswere used(18, 19). bitransgenic mice beginning at 6 weeks of doxycycline treatment. Interleukin-6 analysis. The secreted interleukin (IL)-6 protein level in At this stage, marked inflammatory cell infiltration and emphyse- BALF wasmeasuredby IL-6 ELISA kit (R&D Systems). ma were readily detectable (Fig. 2A, +Dox 6W). Quantitative + + + in vitro CD11b Gr-1 and CD4 T-cell isolation and suppression characterization revealed decreased alveolar number and increased assay. Myeloid-derived suppressor cells (MDSC) were isolated according to alveolar perimeter, alveolar radius, alveolar mean cord length, a previously described procedure (24). CD4+ T cellswere isolatedfrom the wild-type (WT) spleens with CD4 monoclonal antibody (mAb)–coated alveolar surface area, and alveolar volume in doxycycline-treated magnetic beadsand MACS-LS columnsaccording to the manufacturer’s bitransgenic lungs compared with those in doxycycline-untreated instructions. Isolated WT CD4+ cellswere labeled with carboxyfluorescein bitransgenic lungs (Supplementary Table S1). These resulted in diacetate succinimidyl diester (CFSE; Molecular Probes). Labeled cells were significant reduction of the overall alveolar surface area (69.16%) stimulated with anti-CD3 mAb plus anti-CD28 mAb for 3 d in the presence for gasexchange in doxycycline-treated bitransgenicmice. After 10 or absence of CD11b+/Gr-1+ cells isolated from the spleens of doxycycline- to 15 weeksof doxycycline treatment, the bitransgenicmice began treated CCSP-rtTA/(tetO)7-CMV-MMP12 bitransgenic mice as mentioned to develop adenomatoid hyperplasia in both parenchyma and small above. The ratiosof MDSC:CD4 + T cellswere 0:1, 1:25, 1:5, and 1:1, A, +Dox 10W + conducting airways(Fig. 2 ), which resembles to the respectively. Proliferation of CD4 T cellswasevaluated asCFSE dilution by histopathologic feature of dysplasia in clinical lesions. Bronchioal- flow cytometry. The IL-2 concentration in the cultured medium was veolar adenocarcinomas were observed in the lungs of bitransgenic measured by ELISA according to the protocol by Pharmingen. A, +Dox MMP12 immunohistochemical staining. MMP12 immunohistochem- mice after doxycycline treatment asearly as16 weeks(Fig. 2 16W ical staining was the same as previously described (13). ). To determine the tumor incidence rate, bitransgenic mice were euthanized at 9 to 11 monthsafter doxycycline treatment for histopathologic analyses (Fig. 2B). Of 40 doxycycline-untreated Results bitransgenic mice, only 1 had lung bronchioalveolar adenocarci- Generation of CCSP-rtTA/(tetO)7-CMV-MMP12 bitransgenic noma (2.5%) and 2 others developed pulmonary dysplasia without mice. A doxycycline-controlled bitransgenic mouse model was cancer. In marked contrast, 16 of 47 doxycycline-treated bitrans- generated to specifically direct MMP12 expression in lung epithelial genic mice in the same age group had lung bronchioalveolar cells. In this system, a previously established CCSP-rtTA transgenic adenocarcinoma (34.04%). Five of them developed both pulmonary mouse line (25) was crossbred with a newly generated (tetO)7- dysplasia and bronchioalveolar adenocarcinoma, and three mice www.aacrjournals.org 7253 Cancer Res 2009; 69: (18). September 15, 2009

Cancer Research had dysplasia without bronchioalveolar adenocarcinoma. In whereasanother group of mice continued to be treated. most tumor incidences, one tumor per lung was observed Interestingly, both groups of mice developed hyperplasia 2 months (Fig. 2A, +Dox 40W). A few mice showed tumors on both sides of later (Fig. 2D). This observation suggests that a MMP12-initiated lung lobes (lobectomy). None of WT and single transgenic pathogenic mechanism is irreversible to induce emphysema to mice developed bronchioalveolar adenocarcinoma regardless of hyperplasia transition. doxycycline treatment (n = 10, respectively). In the soft agar Proliferation and apoptosis of alveolar epithelial cells in assay, tumor cell colonies were observed from lung cells that bitransgenic mice. Lung adenocarcinoma formation dependson were isolated from doxycycline-treated bitransgenic mice. No both uncontrolled epithelial cell growth and apoptotic inhibition. tumor cell colony was observed from those of untreated mice To evaluate if MMP12 overexpression stimulates epithelial cell (Fig. 2C). This observation suggests that MMP12 overexpression proliferation during tumor initiation and progression, bromodeox- facilitates the neoplasia process. Finally, a doxycycline on/off yuridine (BrdUrd) wasinjected into 3-, 6-, and 9-month experiment was performed. First, bitransgenic mice were treated doxycycline-treated or untreated bitransgenic mice (n = 3). After with doxycycline for 2 monthsto allow emphysemaformation. 72 hours, the BrdUrd-labeled cells were analyzed by FACS after One group of mice wasterminated of doxycycline treatment, double staining of whole lung cells with fluorochrome-conjugated

Figure 1. Expression of MMP12 in doxycycline-inducible bitransgenic mice. A, total RNA was purified from AT II epithelial cells of WT mice, (tetO)7-CMV-MMP12 single transgenic mice, and CCSP-rtTA/ (tetO)7-CMV-MMP12 bitransgenic transgenic mice that were treated [DOX(+)] or untreated [DOX(À)] with doxycycline for 3 mo. Real-time PCR was used to quantify mRNA expression levels of MMP12 as normalized by glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA expression. B, lung single cells from 3-mo doxycycline-treated (red line)or untreated bitransgenic mice (blue line)or WT mice (green line) were analyzed by flow cytometry in histograms (n = 10) by anti-Flag antibody and SP-C antibody or CCSP antibody. Alveolar macrophages (marker CD11b) were negative control. The shaded areas were isotypic antibody control. C, MMP12-specific enzymatic activity was analyzed in BALF from WT mice (n = 4), doxycycline-treated (tetO)7-CMV-MMP12 single transgenic mice (n = 3), and doxycycline-treated (n = 3) and untreated (n =4) bitransgenic mice. **, P < 0.01. D, immunohistochemical staining of MMP12 in the doxycycline-treated and untreated bitransgenic lung.

Cancer Res 2009; 69: (18). September 15, 2009 7254 www.aacrjournals.org

MMP12 Causes Inflammation and Lung Cancer

Figure 2. Pathology of bitransgenic lung. A, lung tissue sections from representative doxycycline-treated or untreated bitransgenic mice at various time points were H&E stained. A histopathologically normal lung tissue from 15-wk doxycycline-untreated bitransgenic mice (ÀDox 15W) was used as control. A gross picture represents the tumor size of a 40-wk doxycycline-treated bitransgenic mouse (+Dox 40W)as indicated by a green arrow. B, the tumor incidence rates of doxycycline-treated [DOX(+)] or untreated [DOX(À)] bitransgenic mice at 9 to 11 mo of age (n > 40). C, soft agar assay for tumor cell colony formation. Lung single cells from doxycycline-treated (with tumors) or untreated bitransgenic mice were grown on soft agarose gel. D, doxycycline on/off study. Bitransgenic mice were treated with doxycycline for f2 mo. One group of mice was terminated of treatment and another group of mice continued to be treated. Two months later, H&E staining was performed on doxycycline-untreated mice (ÀDox), doxycycline-treated mice (+Dox), and doxycycline-treated and off-treated mice (+Dox on/off).

anti-BrdUrd and anti–SP-C antibodies. As shown in Fig. 3A, there bitransgenic mice after 3 months of doxycycline treatment as was a steady increase of BrdUrd pulse-labeled AT II epithelial cells analyzed by real-time PCR (n =3;Fig.3C). in the doxycycline-treated bitransgenic mice compared with Activation of oncogenic pathways in the lung of bitransgenic untreated littermates. The increase of MMP12-induced cell mice. In searching for molecular mechanism by which MMP12 proliferation wastime dependent. To evaluate if MMP12 over- overexpression stimulates emphysema to bronchioalveolar adeno- expression alters apoptotic activity in lung epithelial cells during carcinoma transition, the IL-6/signal transducer and activator tumor initiation and progression, whole lung cells were labeled of transcription 3 (Stat3) pathway was studied. This is because with Annexin V and fluorochrome-conjugated anti–SP-C antibodies IL-6 up-regulation causes emphysema (26). More importantly, (Fig. 3B). FACS analysis showed a more than 5-fold decrease of persistent activation of IL-6 downstream Stat3 in AT II epithelial Annexin V–labeled AT II epithelial cellsin 9-month doxycycline- cellsdirectly induced lung inflammation and bronchioalveolar treated bitransgenic mice compared with untreated littermates adenocarcinoma (22). To determine secreted IL-6 in BALF, (21.41% versus 4.91%). The decrease of MMP12-induced apoptosis bitransgenic mice at 1 month old were treated with doxycycline was also time dependent. Furthermore, expression of a spectrum of for 0, 1, 3, 6, and 9 months. Protein concentrations of secreted IL-6 proapoptotic genes( Apaf-1, Bax, Bid, FasL, Casp3, Casp7, Casp9, in BALF were measured by ELISA. As shown in Fig. 4A, IL-6 and CideA) wasdown-regulated in AT II epithelial cellsof concentration in BALF of doxycycline-treated mice wassteadily www.aacrjournals.org 7255 Cancer Res 2009; 69: (18). September 15, 2009

Cancer Research increased compared with those in untreated bitransgenic mice in a bitransgenic mice were isolated and double labeled with fluoro- time-dependent fashion, correlated well with the process from chrome-conjugated anti–phospho-Stat3Y705 and SP-C (AT II cell emphysema to bronchioalveolar adenocarcinoma transition. In- marker) antibodies. Labeled cells were analyzed by flow cytometry. crease of the IL-6 concentration can potentially activate down- In gated SP-C–positive cells, percentage (32.46% versus 2.07%), stream Stat3 (phosphorylation) in AT II epithelial cells to induce total positive cells, and mean fluorescence intensity (MFI) of bronchioalveolar adenocarcinoma. To test this assumption, cells phospho-Stat3Y705 were dramatically increased in doxycycline- from whole lungsof 9-month doxycycline-treated or untreated treated bitransgenic mice compared with untreated bitransgenic

Figure 3. Proliferation and apoptosis of alveolar epithelial cells in bitransgenic mice. A, whole lung cells from BrdUrd-injected 3-, 6-, and 9-mo doxycycline-treated [DOX(+)] or untreated [DOX(À)] bitransgenic mice or WT mice were double stained with SP-C and BrdUrd antibodies for FACS analysis. Example of histogram analysis of 9-mo–treated bitransgenic mice. Isotypic antibody (shaded area) served as control. The right two panels showed total BrdUrd-positive SP-C cells and MFIof 3-, 6-, and 9-mo doxycycline-treated and untreated mice or WT mice (n = 3) by FACS analysis. *, P < 0.05; **, P < 0.01. B, whole lung cells from 3-, 6-, and 9-mo doxycycline-treated or untreated bitransgenic mice were doubly stained with SP-C and Annexin V antibodies for FACS analysis. Top, example of dot plot analysis of 9-mo–treated bitransgenic mice. Isotypic antibody served as control (data not shown). Bottom, summary studies of 3-, 6-, and 9-mo doxycycline-treated or untreated mice (n = 3) by FACS analysis. *, P < 0.05; **, P < 0.01. C, total RNAs were purified from AT II epithelial cells of 3-mo doxycycline-treated or untreated bitransgenic mice. The expression levels of proapoptotic molecules were measured by real-time PCR. Data represent three independent (n = 3) studies.

Cancer Res 2009; 69: (18). September 15, 2009 7256 www.aacrjournals.org

MMP12 Causes Inflammation and Lung Cancer

Figure 4. Activation of the IL-6/Stat3 pathway in AT II epithelial cells of bitransgenic mice. A, bitransgenic mice were treated [DOX(+)] or untreated [DOX(À)] with doxycycline for 0, 1, 3, 6, and 9 mo. BALF from the lungs was analyzed by IL-6 ELISA kit. P < 0.05 (n = 3–4). B, whole lung single cells from 9-mo doxycycline-treated (red line)or untreated (blue line) bitransgenic mice or WT mice (green line) were double labeled with anti–phospho-Stat3Y705 and SP-C antibodies and analyzed by flow cytometry. Isotypic antibody (shaded area) served as control. In gated SP-C–positive cells, the percentage numbers of phospho-Stat3Y705–positive cells were presented in histograms. Total phospho-Stat3Y705–positive cells and MFIof phospho-Stat3Y705 in gated SP-C–positive cells were also determined. The study represents five independent experiments (n =5).C, total RNA was purified from AT II epithelial cells of bitransgenic mice that were treated or untreated with doxycycline for 9 mo. Real-time PCR was used to quantify mRNA expression levels of IL-6, LIF, and Stat3 and normalized by GAPDH mRNA expression. P <0.05(n = 3–4). D, total RNA was purified from AT II epithelial cells of bitransgenic mice (Bi Tg), WT mice, and MMP12 single transgenic mice that were treated or untreated with doxycycline for 1 mo. Real-time PCR was used to quantify mRNA expression levels of Stat3 downstream genes and MMP genes and normalized by GAPDH mRNA expression.

mice and WT mice (Fig. 4B). At the gene transcriptional level, both pathway and MMP12-induced tumorigenesis, expression levels of IL-6 and Stat3 mRNA levelswere up-regulated in responseto these molecules were measured in bitransgenic mice before tumor MMP12 overexpression (Fig. 4C). Other intracellular oncogenic formation. Total RNAswere purified from AT II epithelial cellsafter moleculeswere alsoanalyzed (Supplementary Fig. S2). NF nB, 1 month of doxycycline treatment. The expression levels of Stat3 extracellular signal-regulated kinase (ERK), p38, and AKT were all downstream cytokines and chemokines were quantitatively deter- significantly activated in doxycycline-treated bitransgenic mice mined by real-time PCR. In comparison with untreated bitrans- compared with untreated bitransgenic mice. genic mice, these molecules were highly induced in response to We identified multiple Stat3 downstream cytokines and chemo- MMP12 overexpression. Among them, proinflammatory cytokine kinesin AT II epithelial cellsthat play important rolesin initiating and chemokine gene Gp130 wasup-regulated to 199.4-fold, LIF to inflammation-triggered bronchioalveolar adenocarcinoma in vivo 80.57-fold, IL-6 to 84.8-fold, CSF-2 to 162.2-fold, TNFsf9 to 128.8- (22). To further establish the connection between the IL-6/Stat3 fold, CCL5 to 98-fold, CCL8 to 208-fold, CXCR2 to 1,930-fold, and www.aacrjournals.org 7257 Cancer Res 2009; 69: (18). September 15, 2009

Cancer Research

VEGF to 160.3-fold in doxycycline-treated bitransgenic mice MMPs was also observed in doxycycline-treated bitransgenic mice compared with those untreated mice (Fig. 4D). CSF-1 remained compared with untreated mice. unchanged. In addition, Stat3-induced downstream developmental Inflammatory cell infiltration in the lung of bitransgenic gene HNF4a wasup-regulated to 78.6-fold, Foxa3 to 9.3-fold, and mice. One major manifestation in Stat3-induced bronchioalveolar SHH to 3.3-fold (Fig. 4D). None of these molecules was induced in adenocarcinoma isinflammatory cell infiltration in the lung (22). WT and MMP12 single transgenic mice regardless of doxycycline Because persistent activation of the IL-6/Stat3 pathway was treatment. Interestingly, increased expression of several other observed by MMP12 overexpression, it is necessary to assess if

Figure 5. Inflammatory cell infiltration in bitransgenic mice. A, whole lung cells were isolated from the lungs of 3-, 6-, and 9-mo doxycycline-treated [DOX(+)]or untreated [DOX(À)] bitransgenic mice and age-matched WT mice. A representative of 3-mo–treated and untreated bitransgenic mice and age-matched WT mice is presented. B, the time-dependent study of total Gr-1+CD11b+ cell numbers in the lung of 3-, 6-, and 9-mo doxycycline-treated and untreated bitransgenic mice and age-matched WT mice. Columns, mean of at least four independent animals (n > 4); bars, SD. *, P < 0.05. C, CFSE-labeled normal CD4+ T cells were stimulated with anti-CD3 mAb plus anti-CD28 mAb in the presence or absence of MDSCs isolated from doxycycline-treated bitransgenic mice. The ratios of MDSC:T cell were 0:1, 1:25, 1:5, and 1:1, respectively. Proliferation of CD4+ T cells was evaluated as CFSE dilution by flow cytometry. D, inhibition of IL-2 secretion in supernatants of stimulated CD4+ T cell and MDSC cocultured cells measured by ELISA.

Cancer Res 2009; 69: (18). September 15, 2009 7258 www.aacrjournals.org

MMP12 Causes Inflammation and Lung Cancer

Figure 6. MMP12 up-regulation in human adenocarcinoma, squamous cell carcinoma, and COPD. Expression of MMP12 mRNA was measured by real-time PCR and normalized by GAPDH mRNA expression. Numbers in each column represent average dCt. Average fold changes were determined by 2(ddCt), in which dCt = dCt (cancer or COPD samples) À dCt (normal human samples). Numbers on the bottom represent average fold changes compared with the normal human samples. Normal human samples were set up as one.

MMP12 overexpression induces inflammatory cell infiltration carcinomas, 2.23-fold higher in lung tissues with COPD from during the course of pathogenesis in the lung of bitransgenic nonsmokers, and 2.92-fold higher in lung tissues with COPD from mice. Whole lung cellswere isolatedfrom the lungsof 3-, 6-, and smokers. Therefore, up-regulation of MMP12 is associated with 9-month doxycycline-treated or untreated bitransgenic mice and emphysema and lung cancer in humans. Stat3 up-regulation has WT mice for staining with fluorochrome-conjugated anti-mouse been observed in the same human samples as we previously CD11b (for monocytes) or Gr-1 (for neutrophils) antibody in described (23). Co–up-regulation of MMP12 and Stat3 in human FACS analysis. Compared with untreated bitransgenic mice, the patientsprovidesadditional evidence supporting that both percentage numbersof CD11b + (2.59% versus 11.84%) and Gr-1+ moleculesare connected and contribute to lung cancer formation. (0.64% versus 5.87%) inflammatory cells were significantly increased in the lung of 3-month doxycycline-treated bitransgenic mice (Fig. 5A). Importantly, the percentage number of CD11b/Gr-1 Discussion double-positivecells(2.91% versus15.58%) wassignificantly To understand the pathophysiologic roles of MMP12 in the lung, increased in the lung of doxycycline-treated mice (Fig. 5A). a conditional bitransgenic mouse model was generated to CD11b/Gr-1 double-positive cells are MDSCs. Constant influx of overexpress MMP12 in lung epithelial cells. In this bitransgenic these cells suppresses adaptive immunity and supports the model, doxycycline treatment induced MMP12-Flag fusion protein angiogenesis and stroma remodeling for tumor growth (27). With expression at both mRNA and protein levels (Fig. 1A and B). In disease progression, the total number of this cell population was BALF of bitransgenic mice after doxycycline treatment, the MMP12 gradually increased in the lungs of 3-, 6-, and 9-month doxycycline- enzymatic activity was significantly increased (Fig. 1C). Together, B treated bitransgenic mice (Fig. 5 ). MDSCsisolatedfrom these results suggest that CCSP-rtTA/(tetO)7-CMV-MMP12 bitrans- doxycycline-treated bitransgenic mice were able to inhibit T-cell genic mice are able to synthesize and secret the active form of the proliferation, in which CFSE-labeled CD4+ T cellswere stimulated MMP12-Flag fusion protein in the alveolar lumen. with anti-CD3 mAb plusanti-CD28 mAb. The inhibition wasMDSC As a pathologic consequence, persistent overexpression of the concentration dependent (Fig. 5C). IL-2 secretion from stimulated MMP12-Flag fusion protein in AT II cells caused emphysema CD4+ T cells was also inhibited by MDSCs in the cocultured study 6 weeksafter doxycycline treatment in bitransgenicmice (Fig. 2 A). asmeasured by ELISA, an indication of functional impairment of Thisisprobably due to degradation of ECM that weakensthe T cells(Fig. 5 D). These results indicate that inflammatory cell interstitial alveolar structure. With longer MMP12 exposure, infiltration and immunosuppression plays a critical role in dysplasia (10–15 weeks of doxycycline treatment) and bronchioal- emphysema to bronchioalveolar adenocarcinoma transition. veolar adenocarcinoma (16–40 weeksof doxycycline treatment) MMP12 up-regulation in human adenocarcinoma, squamous were observed. This sequential emphysema to lung cancer cell carcinoma, and COPD. Because MMP12 overexpression transition resembles the clinical situation that COPD is highly caused lung tumor formation in the CCSP-rtTA/(tetO)7-CMV- likely to develop lung cancer in humans. It seems that MMP12 MMP12 bitransgenic mouse model, it is important to determine if serves as a trigger. In the doxycycline on/off study (Fig. 2D), in up-regulation of MMP12 is associated with lung cancers and COPD which doxycycline treatment wasterminated once emphysemawas in humans. The expression levels of MMP12 mRNA in human established but before hyperplasia, the mice still developed adenocarcinomas( n = 24), squamous cell carcinomas (n = 22), hyperplasia 2 months later. This may be explained by that COPD without smoking (n = 21), and COPD with smoking (n = 25) inflammation (e.g., inflammatory cell infiltration and cytokine/ versus normal samples (n = 12) were examined by quantitative chemokine up-regulation) had already been induced, which real-time PCR (Fig. 6). In comparison with normal human lungs, triggered irreversible pathologic cascades toward tumorigenesis. the average of MMP12 mRNA expression levels was 10.06-fold In a similar clinical situation, some smokers develop lung cancer higher in adenocarcinomas, 5.31-fold higher in squamous cell even after quitting smoking. Before and during tumor formation, www.aacrjournals.org 7259 Cancer Res 2009; 69: (18). September 15, 2009

Cancer Research

AT II epithelial cellsunderwent both increasedcell proliferation as disruption of the IL-6/Stat3 pathway in CCSP-rtTA/(tetO)7-CMV- determined by BrdUrd pulse-labeling analysis and decreased MMP12 bitransgenic mice by crossbreeding with IL-6 and Stat3 apoptosis as determined by Annexin V labeling analysis under knockout mice should be performed in the future. As we reported MMP12 overexpression (Fig. 3). recently, Stat3 and its downstream genes serve as biomarkers for It hasbeen proposedpreviously that degraded fragmentsfrom lung adenocarcinoma and COPD diagnosis in humans (23). ECM by MMPsserveaschemoattractantsfor inflammatory cell Importantly, up-regulation of MMP12 is associated with human infiltration in the lung (28). Thisseemsto be supported by our COPD, adenocarcinomas, and squamous cell carcinomas (Fig. 6). observation that inflammatory cells are readily detectable in the Other oncogenic molecules(e.g., ERK1/2, AKT, p38, and NF nB) alveolar region before and after tumor formation in MMP12 were also activated in doxycycline-treated bitransgenic mice bitransgenic lungs after doxycycline treatment (Fig. 5). Especially, compared with untreated bitransgenic mice (Supplementary tumor-promoting MDSCswere significantly increasedin the lung of Fig. S2), suggesting that MMP12-induced tumorigenesis is a bitransgenic mice. As we reported recently, increased concentration complex process. of MDSCs alone by apoptosis inhibition is sufficient to cause In summary, sequential appearance of emphysema, epithelial bronchioalveolar adenocarcinoma in vivo (19). Infiltration of innate dysplasia, and bronchioalveolar adenocarcinoma in the CCSP- and adaptive immune cellsinto lung changed the local microenvi- rtTA/(tetO)7-CMV-MMP12 bitransgenic mouse model supports ronment and hijacked the immune surveillance system to favor that MMP12 isa key player that controlsemphysemato tumor tumor growth. These cells change the microenvironment to assist transition in the lung. This seems achieved by activation of the tumorigenesis to influence angiogenesis, remodel ECM, and increase proinflammatory IL-6/Stat3 pathway (at least partially) and mutational rate that leadsto genetic or epigenetic instabilities. inflammatory cell infiltration, therefore providing additional One obviousmicroenvironment change in BALF of bitransgenic evidence showing that chronic inflammation facilitates formation mice is the steady increase of IL-6 concentration (Fig. 4A). of emphysema and lung cancer. As we showed previously, MMP12 Interestingly and importantly, IL-6 up-regulation has been reported isunder the control of anti-inflammatory PPAR g and itslipid to cause emphysema and to play a role in lung adenocarcinoma hormonal ligands(14). Activation of PPAR g inhibitsthe prolifer- (22, 26). Phosphorylation of IL-6 downstream effector Stat3 in AT II ation of lung carcinoma cells (29). Therefore, our study supports a epithelial cells was significantly increased as a result of IL-6 surge concept that the balance and antagonism between anti-inflamma- in BALF of doxycycline-treated bitransgenic mice (Fig. 4B). tory PPARg and proinflammatory IL-6/Stat3 determine lung cancer Simultaneously, IL-6 and Stat3 gene transcription was activated formation, similar to that observed in myeloma (30). in AT II epithelial cellsafter doxycycline treatment (Fig. 4 C). We reported previously that persistent activation of Stat3 in AT II Disclosure of Potential Conflicts of Interest epithelial cellsdirectly induced lung bronchioalveolar adenocarcinoma (22). Consequently, Stat3 downstream genes in AT II No potential conflicts of interest were disclosed. epithelial cellswere activated in bitransgenicmice (Fig. 4 D). These genesare known to promote inflammation (e.g., LIF, IL-6, CSF-2, Acknowledgments TNFsf9, CCL5, CCL8, CXCR2, and VEGF) and to stimulate epithelial HNF4a, Foxa3 SHH Received 2/13/09; revised 6/15/09; accepted 7/11/09; published OnlineFirst 8/25/09. growth (e.g., , and ). It isconceivable that up- Grant support: NIH grantsCA138759 (C. Yan), HL087001 (H. Du), HL-061803 regulation of these Stat3 downstream genes along with other MMP (C. Yan and H. Du), and HL-067862 (C. Yan and H. Du) and Indiana University Cancer genesexacerbatesinflammation and inducesepithelial cell Center ITRAC Award (C. Yan). The costs of publication of this article were defrayed in part by the payment of page transformation into bronchioalveolar adenocarcinoma in the lung charges. This article must therefore be hereby marked advertisement in accordance of bitransgenic mice. Therefore, activation of the IL-6/Stat3 with 18 U.S.C. Section 1734 solely to indicate this fact. We thank Jennifer Roberts for assisting in real-time PCR, Marjorie E. Albrecht for pathway is responsible, at least partially, for MMP12-induced assisting in tissue collection and animal maintenance, and Drs. Jay W. Tichelaar and emphysema and tumorigenesis. To formally prove this concept, Jeffrey A. Whitsett for providing CCSP-rtTA transgenic mice.

References Hydrolysis of a broad spectrum of extracellular matrix 13. Lian X, Qin Y, Hossain SA, et al. Overexpression of proteins by human macrophage elastase. J Biol Chem Stat3C in pulmonary epithelium protectsagainsthyper- 1. Mannino DM, Aguayo SM, Petty TL, Redd SC. Low 1997;272:12189–94. oxic lung injury. J Immunol 2005;174:7250–6. lung function and incident lung cancer in the United 8. Shapiro SD, Kobayashi DK, Ley TJ. Cloning and 14. Lian X, Yan C, Qin Y, Knox L, Li T, Du H. Neutral States: data From the First National Health and characterization of a unique elastolytic metalloprotei- lipidsand peroxisomeproliferator-activated receptor- g Nutrition Examination Survey follow-up. Arch Intern nase produced by human alveolar macrophages. J Biol control pulmonary gene expression and inflammation- Med 2003;163:1475–80. Chem 1993;268:23824–9. triggered pathogenesis in lysosomal acid lipase knock- 2. Page-McCaw A, Ewald AJ, Werb Z. Matrix metal- 9. Hautamaki RD, Kobayashi DK, Senior RM, Shapiro SD. out mice. Am J Pathol 2005;167:813–21. loproteinases and the regulation of tissue remodelling. Requirement for macrophage elastase for cigarette 15. Lian X, Yan C, Yang L, Xu Y, Du H. Lysosomal acid Nat Rev Mol Cell Biol 2007;8:221–33. smoke-induced emphysema in mice. Science 1997;277: lipase deficiency causes respiratory inflammation and 3. ParksWC, WilsonCL, Lopez-Boado YS. Matrix 2002–4. destruction in the lung. Am J Physiol Lung Cell Mol metalloproteinases as modulators of inflammation and 10. Hofmann HS, Hansen G, Richter G, et al. Physiol 2004;286:L801–7. innate immunity. Nat Rev Immunol 2004;4:617–29. Matrix metalloproteinase-12 expression correlates 16. Yan C, Lian X, Li Y, et al. Macrophage-specific 4. Westermarck J, Kahari VM. Regulation of matrix with local recurrence and metastatic disease in non- expression of human lysosomal acid lipase corrects metalloproteinase expression in tumor invasion. FASEB small cell lung cancer patients. Clin Cancer Res 2005;11: inflammation and pathogenic phenotypesin lal À/À J 1999;13:781–92. 1086–92. mice. Am J Pathol 2006;169:916–26. 5. Moses MA. The regulation of neovascularization of 11. Sun T, Gao Y, Tan W, et al. Haplotypesin matrix 17. Yan C, Lian X, Dai Y, et al. Gene delivery by the hSP-B matrix metalloproteinases and their inhibitors. Stem metalloproteinase gene cluster on chromosome 11q22 promoter to lung alveolar type II epithelial cellsin LAL- Cells1997;15:180–9. contribute to the risk of lung cancer development and knockout mice through bone marrow mesenchymal 6. Werb Z, Gordon S. Elastase secretion by stimulated progression. Clin Cancer Res 2006;12:7009–17. stem cells. Gene Ther 2007;14:1461–70. macrophages. Characterization and regulation. J Exp 12. Dehan E, Ben-Dor A, Liao W, et al. Chromosomal 18. Qu P, Du H, WilkesDS, Yan C. Critical rolesof Med 1975;142:361–77. aberrations and gene expression profiles in non-small lysosomal acid lipase in T cell development and 7. Gronski TJ, Jr., Martin RL, Kobayashi DK, et al. cell lung cancer. Lung Cancer 2007;56:175–84. function. Am J Pathol 2009;174:944–56.

Cancer Res 2009; 69: (18). September 15, 2009 7260 www.aacrjournals.org

MMP12 Causes Inflammation and Lung Cancer

19. Qu P, Du H, Li Y, Yan C. Myeloid-specific expression inducesinflammation and adenocarcinomasin mouse ing interleukin (IL)-11 and IL-6 in the lung. Am J Respir of Api6/AIM/Spa induces systemic inflammation and lung. Cancer Res2007;67:8494–503. Cell Mol Biol 2000;22:289–95. adenocarcinoma in the lung. J Immunol 2009;182: 23. Qu P, RobertsJ, Li Y, et al. Stat3 downstream genes 27. Sica A, Bronte V. Altered macrophage differentiation 1648–59. serve as biomarkers in human lung carcinomas and and immune dysfunction in tumor development. J Clin 20. Bolender RP, Hyde DM, Dehoff RT. Lung morphom- chronic obstructive pulmonary disease. Lung Cancer Invest 2007;117:1155–66. etry: a new generation of toolsand experimentsfor 2009;63:341–7. 28. Shapiro SD, Senior RM. Matrix metalloproteinases. organ, tissue, cell and molecular biology. Am J Physiol 24. Bronte V, Apolloni E, Cabrelle A, et al. Identification Matrix degradation and more. Am J Respir Cell Mol Biol 1993;265:L521–48. of a CD11b(+)/Gr-1(+)/CD31(+) myeloid progenitor 1999;20:1100–2. 21. Massaro GD, Massaro D. Retinoic acid treat- capable of activating or suppressing CD8(+) T cells. 29. Han SW, Roman J. Activated PPARg targets ment abrogates elastase-induced pulmonary emphy- Blood 2000;96:3838–46. surface and intracellular signals that inhibit the sema in rats [see comments] [published erratum 25. Tichelaar JW, Lu W, Whitsett JA. Conditional proliferation of lung carcinoma cells. PPAR Res 2008; appearsin Nat Med 1997 Jul;3(7):805]. Nat Med 1997;3: expression of fibroblast growth factor-7 in the develop- 2008:254108. 675–7. ing and mature lung. J Biol Chem 2000;275:11858–64. 30. Wang LH, Yang XY, Zhang X, et al. Transcriptional 22. Li Y, Du H, Qin Y, RobertsJ, CummingsOW, Yan C. 26. Kuhn C III, Homer RJ, Zhu Z, et al. Airway inactivation of STAT3 by PPARg suppresses IL-6- Activation of the signal transducers and activators of hyperresponsiveness and airway obstruction in trans- responsive multiple myeloma cells. Immunity 2004;20: the transcription 3 pathway in alveolar epithelial cells genic mice. Morphologic correlates in mice overexpress- 205–18.

www.aacrjournals.org 7261 Cancer Res 2009; 69: (18). September 15, 2009

Matrix Metalloproteinase 12 Overexpression in Lung Epithelial Cells Plays a Key Role in Emphysema to Lung Bronchioalveolar Adenocarcinoma Transition

Peng Qu, Hong Du, Xi Wang, et al.

Cancer Res 2009;69:7252-7261. Published OnlineFirst August 25, 2009.

Updated version Access the most recent version of this article at: doi:10.1158/0008-5472.CAN-09-0577

Supplementary Access the most recent supplemental material at: Material http://cancerres.aacrjournals.org/content/suppl/2009/08/24/0008-5472.CAN-09-0577.DC1

Cited articles This article cites 30 articles, 11 of which you can access for free at: http://cancerres.aacrjournals.org/content/69/18/7252.full#ref-list-1

Citing articles This article has been cited by 11 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/69/18/7252.full#related-urls

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 Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/69/18/7252. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.