|Home | About | Journals | Submit | Contact Us | Français|
Lung cancer is more deadly than colon, breast, and prostate cancers combined, and treatment improvements have failed to improve prognosis significantly. Here, we identify a critical mediator of lung cancer progression, Rac1b, a tumor-associated protein with cell-transforming properties that are linked to the matrix metalloproteinase (MMP)–induced epithelial-mesenchymal transition (EMT) in lung epithelial cells. We show that expression of mouse Rac1b in lung epithelial cells of transgenic mice stimulated EMT and spontaneous tumor development and that activation of EMT by MMP-induced expression of Rac1b gave rise to lung adenocarcinoma in the transgenic mice through bypassing oncogene-induced senescence. Rac1b is expressed abundantly in stages 1 and 2 of human lung adenocarcinomas and, hence, is an attractive molecular target for the development of new therapies that prevent progression to later-stage lung cancers.
A devastating disease that affects millions of people each year, lung cancer is the second highest cause of morbidity and mortality worldwide. Non–small cell lung cancers (NSCLCs) represent more than 80% of all cases (1), and exposure to cigarette smoke dominates as the causative agent. Recent advances in early detection of lung cancer have made substantial progress in reducing mortality (2), but gaps in our knowledge about the molecular mechanisms of pathogenesis and the biological characteristics of lung cancers form barriers to the development of more effective targeted therapies. It is imperative that we improve our understanding of the disease mechanisms to identify new therapeutic targets and strategies for treatment.
Rac1b is an alternatively spliced isoform of Rac1, a Rho family guanosine triphosphatase (GTPase) that is known to regulate diverse cellular processes, such as actin cytoskeleton organization, membrane trafficking, cell proliferation, and gene expression (3-6). The Rac1b isoform results from inclusion of exon 3b, which contains 57 nucleotides and leads to a 19–amino acid in-frame insertion (7, 8). The resulting protein has accelerated guanosine diphosphate (GDP)/guanosine triphosphate (GTP) exchange and impaired GTP hydrolysis (9, 10), which lead to a predominance of the GTP-bound form in the cellular context and associated prolonged signaling activity (9). Rac1b shows increased expression in malignant colorectal (7) and breast cancer (8), and studies with cultured cells have shown that Rac1b may exert oncogenic activities through induction of mitochondrial reactive oxygen species (ROS) production and consequent promotion of the epithelial-mesenchymal transition (EMT) (11).
EMT is a phenotypical alteration in which epithelial cells detach from their neighbors and the underlying basement membrane and become more motile and invasive (12, 13). In cultured cells, EMT can be induced by cytokines, growth factors, and proteases (14). Matrix metalloproteinases (MMPs)—zinc-requiring enzymes that reside in and degrade components of the extracellular matrix—have been shown to induce EMT; MMP-3, MMP-7, and MMP-28 induce EMT in human A549 lung adenocarcinoma (LAC) cells (15-17), and MMP-3 also induces EMT in both mouse mammary epithelial cells (11, 18) and human breast cancer cells (19). Although MMPs are known to function in lung embryonic development and in lung cancer progression (20), the involvement of MMP-induced EMT in lung cancer has not been defined.
A major barrier to tumor progression is oncogene-induced senescence (OIS), a p19Arf−p53/p16Ink4a–retinoblastoma protein–dependent mechanism that limits the proliferative capacity of premalignant cells and that has been best defined for the transition from premalignant lung adenoma to malignant LAC (21). Recent studies indicate that aberrant expression of the EMT-associated transcription factor Twist in cultured cells allows the cells to bypass OIS (22), suggesting that induction of EMT at an early stage of tumor development might stimulate progression from adenoma to carcinoma in vivo. Here, we use cell culture and mouse models in which the role of OIS in tumor progression has been well defined in combination with new transgenic mouse models that display MMP-inducible mouse Rac1b expression and signaling to show that (i) Rac1b functions as a lung cancer oncogene in mice and humans, (ii) expression of MMPs in lung tissue leads to activation of Rac1b, and (iii) activation of EMT by MMP-induced expression of Rac1b stimulates bypass of OIS and progression to LAC.
Evaluation of a panel of LAC biopsies and adjacent normal tissue samples enriched in patients with stage 1 and stage 2 disease revealed that Rac1b mRNA expression was significantly up-regulated (P < 0.0001, unpaired t test) in LAC compared to that in normal adjacent tissue (Fig. 1A and fig. S1, A and B); furthermore, Rac1b mRNA expression was increased in stage 2 relative to stage 1 disease (Fig. 1B), suggesting the involvement of Rac1b during the early stages of lung cancer progression. Using immunohistochemistry (IHC), we found consistently high expression of the Rac1b protein in adenocarcinoma tumor cells (Fig. 1C), indicating a potentially direct role for Rac1b in tumor cell progression. We also found that Rac1b mRNA was significantly increased (P = 0.0238, unpaired t test) in primary tumors of patients who showed node positivity relative to tumors in patients with node-negative disease, suggesting a potential link between expression of Rac1b and spread beyond the primary tumor site (Fig. 1D). Because cigarette smoke is a major lung cancer carcinogen, we evaluated Rac1b expression in association with smoking status in patients. Nonmalignant adjacent lung tissue from patients who were current smokers had significantly (P = 0.011, unpaired t test) higher expression of Rac1b than normal adjacent lung tissue from patients who were never smokers (Fig. 1E). To assess whether there is a relationship with Rac1b status in emphysematous lung disease, we assessed Rac1b expression in a cohort of patients diagnosed with idiopathic pulmonary fibrosis and found that those with peripheral or subpleural distribution of emphysema had significantly higher (P < 0.05, Mann-Whitney test) expression of Rac1b compared to patients who had no axial distribution of emphysema (Fig. 1, F and G).
To define the role of Rac1b in induction of EMT and lung cancer in vivo, we created bitransgenic mice with inducible expression of Rac1b in lung epithelial cells. tetO-YFP-Rac1b transgenic mice were crossed with mice bearing the CCSP-rtta activator transgene, which expresses the reverse tetracycline transactivator (rtTA) under the control of the Clara cell secretory protein (CCSP) promoter (23). Through quantitative polymerase chain reaction (qPCR) analysis of transgene expression, two stable transgenic mouse lines were identified, designated A and F. Expression of the transgene was selectively induced in lung but not liver by feeding the mice doxycycline (Fig. 2, A and B). IHC performed with an antibody specific for the insertion sequence of the Rac1b isoform (11) showed typical CCSP-driven expression patterns (Fig. 2C) (23). Transgene expression progressively increased when mice were fed larger doses of doxycycline (Fig. 2D). Exogenous expression of Rac1b in cultured epithelial cells results in the activation of cells with characteristics of myofibroblasts (12, 14, 24, 25), which are key effector cells in the development of tissue fibrosis—an accumulation of collagen fibrils that can disrupt tissue structure and significantly increase risk of tumor development (12). Thus, we examined tetO-YFP-Rac1b/CCSP-rtta mice fed doxycycline and found that dose-dependent expression of Rac1b in airway epithelial cells was accompanied by the development of peri-airway fibrosis, increased collagen expression, and deposition of collagen fibrils (Fig. 2E). These results reveal that even short-term expression of Rac1b causes alterations in tissue architecture consistent with activation of EMT.
Persistent fibrosis has been linked to an increased risk of tumor development (25); thus, we evaluated spontaneous tumor formation in tetO-YFP-Rac1b/CCSP-rtta lines A and F, which express differential amounts of the transgene (Fig. 2F), and in CCSP-rtta control mice maintained on a doxycycline diet for an extended time. A single adenoma was detected in the lungs of CCSP-rtta control mice (n = 6), consistent with previous reports of sporadic lung adenoma formation in wild-type Friend virus B-type (FVB) strain mice (26). However, numerous tumors were detected in the lungs of 7 of 16 tetO-YFP-Rac1b/CCSP-rtta mice, with 4 of 16 containing multiple tumors (Fig. 2G). Many of these tumors had progressed to adenocarcinoma (Fig. 2H), having features such as a tumor diameter greater than 5 mm (Fig. 2G), greater numbers of atypical and mitotic cells relative to sporadic adenomas that developed in nontransgenic mice (Fig. 2I), tumor cell invasion of large airways (Fig. 2J) and surrounding tissue (Fig. 2K), and variation in architectural growth patterns of neoplastic epithelium. Rac1b expression produced tissue fibrosis even in lungs that did not develop tumors (Fig. 2L), indicating that fibrosis preceded tumor formation. Lungs that developed spontaneous tumors also showed significantly increased expression of Rac1b (P < 0.01, Mann-Whitney test) as well as increased expression of mesenchymal marker transcripts, including smooth muscle actin (P < 0.01, Mann-Whitney test), collagen 1A1 (P < 0.01, Mann-Whitney test), and vimentin (P < 0.05, Mann-Whitney test; Fig. 2M). Immunohistochemical analysis of transgenic mouse lung sections identified intratumoral areas that displayed increased expression of collagen, vimentin, and α-smooth muscle actin, as well as decreased expression of epithelial markers (Fig. 2N and fig. S2), which is consistent with the possibility that an EMT-type transition had occurred within the tumors in the tetO-YFP-Rac1b/CCSP-rtta mice.
Progression to adenocarcinoma requires the cells to bypass senescence, which is induced by deactivation of the retinoblastoma and p53 tumor suppressor pathways (21, 27). Although activation of EMT processes was shown to stimulate bypass of Ras-induced senescence in cultured mammary epithelial cells (22), it is not clear how senescence is bypassed in vivo or which, if any, EMT processes are involved. The spontaneous adenocarcinoma formation we observed in the tetO-YFP-Rac1b/CCSP-rtta transgenic mice (Fig. 2) might have resulted from senescence bypass induced by Rac1b-induced EMT. To test this possibility, we treated tetO-YFP-Rac1b/CCSP-rtta and control CCSP-rtta/ntg mice with ethyl carbamate (urethane), a DNA alkylating agent present in cigarette smoke that induces lung adenomas through the generation of activated Kras protein (28). Tumors that developed in tetO-YFP-Rac1b/CCSP-rtta mice showed substantial expression of Rac1b protein and mRNA, as measured by IHC (Fig. 3A) and qPCR (Fig. 3B), respectively. Although the number of tumors per mouse did not significantly differ between tetO-YFP-Rac1b/CCSP-rtta and control mice, the average tumor size (0.47 mm2 versus 0.37 mm2, respectively) was significantly increased (P < 0.05, Mann-Whitney test; Fig. 3C) in tetO-YFP-Rac1b/CCSP-rtta mice, and the presence of these large and invasive tumors (Fig. 3D) indicated progression to adenocarcinoma, a phenomenon consistent with bypass of senescence. A Rac1b-induced proliferation of mesenchymal cells was evidenced by a significant increase, in tetO-YFP-Rac1b/CCSP-rtta mice relative to control mice, in the expression of EMT marker transcripts (P < 0.05, Mann-Whitney test) in whole-lung homogenates (Fig. 3E) and EMT marker proteins within tumors (Fig. 3F). Dual-fluorescence tissue-labeling experiments supported the association of Rac1b expression with the increase in mesenchymal cell components through activation of EMT: Specifically, in tetO-YFP-Rac1b/CCSP-rtta mice, the tetO-YFP-Rac1b transgene was coexpressed with the epithelial cell marker CCSP in Clara cells (Fig. 3G) in airways and with the mesenchymal marker vimentin in the lung tumors (arrowheads in Fig. 3H and fig. S3). These observations are consistent with the interpretation that Rac1b can induce EMT in Rac1b-expressing cells. Expression of the senescence marker p16Ink4a (Fig. 3, I and J) was decreased, and expression of the proliferation marker Ki67 (Fig. 3K) was increased in tetO-YFP-Rac1b/CCSP-rtta mice relative to controls, suggesting that Rac1b-induced EMT in urethane-initiated tumors is associated with senescence bypass.
Rac1b is known to be induced by MMPs in cultured mammary epithelial cells (11, 24, 29, 30), and cigarette smoke is known to induce MMP expression in lung epithelial cells (31). Therefore, because our studies thus far implicated Rac1b in lung cancer progression, we evaluated the response of H661 human lung cancer cells, which exhibit intermediate invasiveness and malignancy (32), to a panel of recombinant MMPs. Rac1b expression was effectively induced by MMP-3 and MMP-9, but not by MMP-1 (Fig. 4A). Induction of the mesenchymal marker vimentin correlated with up-regulation of Rac1b (Fig. 4B) by these same MMPs.
MMP-3 expression has been found in respiratory epithelial cells [see the Human Protein Atlas (33)] and in lung cancer cells (34-36), and in our studies, overlapping expression of MMP-3 and Rac1b was observed in human lung cancer biopsies (Fig. 4, C and D). Because expression of MMP-3 has also been found in human lung fibrosis, an EMT-associated phenomenon (37-39), we concentrated on MMP-3 as an experimental EMT inducer and assessed a specific role for Rac1b in the increased expression of vimentin in the H661 lung cancer cells treated with MMP-3. Rac1b expression rose in a dose-dependent manner after incubation of the H661 lung cancer cells with MMP-3 (Fig. 4E); this enhanced expression was accompanied by progressively increasing expression of vimentin (Fig. 4F). Acquisition of a flattened cellular morphology (Fig. 4G) was associated with increased expression of myofibroblast-associated smooth muscle actin and, as seen previously, during the early EMT response of mammary epithelial cells to MMP-3 (24). Selective knockdown of Rac1b by small interfering RNA (siRNA) targeting of the unique 57-nucleotide insertion (Fig. 4H and fig. S4) effectively blocked MMP-3–induced expression of vimentin (Fig. 4I).
To determine whether MMP-3 could increase Rac1b expression and activate EMT in vivo, we generated tetO-HA-MMP-3 transgenic mice and crossed them with mice bearing the CCSP-rtta activator transgene, identifying three tetO-HA-MMP-3/CCSP-rtta founder lines (Fig. 5A). The MMP-3 transgene was highly inducible by doxycycline (Fig. 5B) and was specifically expressed in the lung (Fig. 5C) in tetO-HA-MMP-3/CCSP-rtta mice. Control and tetO-HA-MMP-3/CCSP-rtta mice were treated with urethane to induce lung tumors and then with doxycycline to induce transgene expression, and multiple lesions were detected in both groups of mice. The lesions were predominantly solid and papillary adenoma, but progression to adenocarcinoma was observed in a subset of tetO-HA-MMP-3/CCSP-rtta mice, as indicated by tumor sizes greater than 5 mm in diameter, the presence of pleomorphic nuclei, and evidence of invasion at the tumor-host interface (Fig. 5D and figs. S5 and S6). Tumor number (Fig. 5E; P < 0.001, Mann-Whitney test) and size (Fig. 5F; P < 0.05, Mann-Whitney test) were greater in the tetO-HA-MMP-3/CCSP-rtta mice, indicating that MMP-3 might potentiate tumor formation as well as stimulate tumor progression. Expression of Rac1b was increased in the airways and in lung tumors of the tetO-HA-MMP-3 transgenic mice (Fig. 5, G to I). This increase was associated with a decrease in the nuclear localization of the heterogeneous nuclear ribonucleoprotein A1 (hnRNP A1) (Fig. 5, J to M), which previously was shown to bind to exon 3b in Rac1 mRNA and to inhibit exon inclusion, thus blocking expression of Rac1b (40).
Associated with the increased tumor progression in the tetO-HA-MMP-3/CCSP-rtta mice, we observed increased deposition of collagen adjacent to airways, as assessed by trichrome stain; decreased expression of E-cadherin in airway epithelial cells, as assessed by IHC (Fig. 6A); and increased expression of mesenchymal marker mRNA in lung homogenates, as assessed by qPCR (Fig. 6B), all compared to control mice. The notion that these effects were associated with MMP-3–induced EMT was supported by dual-fluorescence labeling of cells adjacent to the airways, which showed coexpression of the HA-MMP-3 transgene in vimentin-expressing cells (Fig. 6C and fig. S7). To provide a broader view of the EMT-promoting cell signaling alterations induced by MMP-3/Rac1b, we performed transcriptional profiling of RNA derived from homogenates of lung tissue isolated from tetO-HA-MMP-3/CCSP-rtta (n = 3) and control (n = 3) mice. We found that 6361 transcripts showed expression changes greater than twofold in tetO-HA-MMP-3/CCSP-rtta versus control mice, and clustering analysis efficiently segregated the controls from the MMP-3–expressing bitransgenic mice (annotated gene expression data are shown in data set S1). This segregation led to the identification of 2948 transcripts that were differentially expressed between control and bi-transgenic mice [P < 0.05 by analysis of variance (ANOVA)] (annotated expression data are shown in data set S2). Gene ontology analyses showed specific up-regulation of transcripts associated with lung development (P = 0.00707), response to oxidative stress (P = 0.0018), and alveolar liquid surface tension (P = 2.67 × 10−7; all categories with significant overlap are presented in data set S3). We subjected the list of differentially expressed genes to a NextBio meta-analysis (41) and found significant overlap with differentially expressed genes in four experimental data sets (for all analyses, P < 0.0001); these studies assessed cigarette smoke–induced effects in human bronchial epithelial cell lines (42), mice exposed to cigarette smoke (43), airway epithelial cells from smokers versus nonsmokers (44), and LAC samples from smokers versus nonsmokers (45) (fig. S8). Significance was found for overlap of our data set with each of these, indicating a possible role for MMP/Rac1b-induced signaling in cigarette smoke–induced pathologies. These findings are consistent with several previous studies showing MMP up-regulation in response to cigarette smoke (31).
To determine whether MMP-3 could induce EMT-like changes in primary lung epithelial cells, we isolated alveolar epithelial cells from transgenic mice that expressed either yellow fluorescent protein (YFP) under the control of the collagen A1 promoter (col-YFP) or red fluorescent protein (RFP) under the control of the smooth muscle actin promoter (sma-RFP; a gift from D. Brenner), plated them on three-dimensional Matrigel to maintain tissue-specific characteristics (46), and supplemented the medium with recombinant MMP-3. These experiments showed activation of fluorescence in cells treated with recombinant MMP-3 relative to untreated cells (Fig. 6, D and E). Primary alveolar epithelial cells isolated from nontransgenic wild-type mice and treated with recombinant purified MMP-3 showed decreased expression of the lung Clara cell differentiation marker CCSP and increased expression of the myofibroblast markers smooth muscle actin and vimentin (Fig. 6F), indicating an MMP-3–induced EMT. To determine whether transgenic MMP-3 could induce EMT-like changes, we isolated primary alveolar epithelial cells from tetO-HA-MMP-3/CCSP-rtta transgenic mice and cultured the cells on Matrigel to maintain tissue-specific gene expression without addition of exogenous MMP-3; addition of increasing amounts of doxycycline to the culture stimulated a progressive invasive branching phenotype (Fig. 6G and fig. S9) characteristic of the development of contractility. To study these effects in vivo, we instilled the lungs of col-YFP/sma-RFP double-transgenic mice with one of three adenoviral vectors, ad-LacZ, ad-MMP-3, or ad-TGFβ (transforming growth factor–β), which expressed either the LacZ (control), the MMP-3, or the TGFβ, respectively (39). We observed a fibrotic reaction after 3 weeks in the MMP-3– and TGFβ-expressing mice (Fig. 6H) and found that, when analyzed by flow cytometry, isolated primary alveolar epithelial cells from ad-MMP-3– or ad-TGFβ–treated mice revealed an increased number of cells that expressed both the YFP and the RFP markers compared to control animals that received ad-LacZ (Fig. 6I). These results support the conclusion that MMP-3/Rac1b–induced changes in cell signaling drive alveolar epithelial cells to undergo EMT toward a myofibroblast phenotype.
To define the stage of tumor progression during which MMP-3 and Rac1b exert their influence, we evaluated the effects of MMP-3 induction at either the time of weaning, the time of urethane injection, or 3 weeks after urethane injection and then stopped MMP-3 expression at either the time of urethane injection, 3 weeks after urethane injection, or at harvest (Fig. 6J). Comparison of tumors in mice that expressed MMP-3 during the tumor initiation period (weeks 1 to 3) with those from mice that expressed MMP-3 during tumor progression (weeks 3 to 14) revealed that tumor size was significantly increased in tetO-HA-MMP-3/CCSP-rtta mice only when MMP-3 expression was initiated after the urethane treatment and then maintained until organ harvest (Fig. 6, K and L). Tumors in hematoxylin and eosin (H&E)–stained lung sections were scored for atypical adenomatous hyperplasia (AAH), adenoma, and adenocarcinoma, and we found adenocarcinoma formation only in mice treated with doxycycline, and thus expressing MMP-3, continuously until the time of harvest (Fig. 6M). Thus, expression of MMP-3 specifically during tumor progression promotes adenocarcinoma. When three additional MMP-3 founder lines were treated with urethane and fed doxycycline from the time of urethane injection until harvest, adenocarcinoma formation was observed in these lines as well (fig. S10). We conclude that MMP-3 expression does not alter response to urethane but acts as a promoter of tumor progression, consistent with EMT-associated senescence bypass.
To further define the MMP-3/Rac1b–associated bypass of Kras-induced senescence, we performed IHC to detect (OIS) markers—p15/INK4B, p16/INK4A, and tumor necrosis factor receptor superfamily member decoy receptor (DCR2) (21)—and one proliferation marker, Ki67, in serial lung sections from urethane-treated tetO-HA-MMP-3/CCSP-rtta or CCSP-rtta/ntg mice fed doxycycline to induce transgene and thus MMP-3 expression. The staining intensity of p15/INK4B and DCR2 was significantly lower in lung tumors from tetO-HA-MMP-3/CCSP-rtta mice (Fig. 6N and fig. S11) relative to control CCSP-rtta/ntg mice, which is consistent with bypass of senescence. Tumors that showed intense staining for all four OIS markers were not detected in tetO-HA-MMP-3/CCSP-rtta mice, whereas they were readily observed in control mice (Fig. 6N). In contrast, most tumors in tetO-HA-MMP-3/CCSP-rtta mice showed little to no staining of OIS markers accompanied by a higher percentage of nuclei positive for the cell proliferation marker Ki67 relative to control mice (Fig. 6N). These data indicate that expression of MMP-3 could act as a tumor promoter by causing urethane-induced adenomas to develop to adenocarcinoma through bypass of OIS. Under conditions of extended transgene expression, MMP-3/Rac1b may also function in cell transformation. We previously observed in mouse mammary cell models that MMP-3–induced Rac1b can spur genomic instability (11) and cannot rule out a potential role for MMP-3/Rac1b in a similar process in lung.
To evaluate potentiation of tumor development by MMP-3 in mice with a specific activating mutation of Kras, we performed a series of parallel experiments using mouse strains that carry an oncogenic allele of KRAS (KrasLA2) that can be activated by spontaneous recombination and are highly predisposed to early-onset lung cancer (47). Triple-transgenic (tetO-HA-MMP-3/CCSP-rtta/KrasLA2) and control (CCSP-rtta/KrasLA2 or ntg/KrasLA2) mice were fed doxycycline continuously from the time of weaning until harvest 14 weeks later. The triple-transgenic mice that expressed MMP-3 had numerous lesions visible in the lung (fig. S12A), with increased tumor number (fig. S12B; P < 0.05, Mann-Whitney test) and size (fig. S12C; P < 0.001, Mann-Whitney test) as well as increased expression of Rac1b (fig. S12D; P < 0.05, Mann-Whitney test) relative to MMP-3–expressing control mice. Consistent with a bypass of Kras-induced senescence, the MMP-3–expressing triple-transgenic mice displayed decreased expression of p16Ink4a and increased proliferation (as assessed by Ki67 expression; fig. S11E) compared to control mice. We thus conclude that MMP-3 can drive tumor progression in KrasLA2 mutant mice.
Increased expression of MMPs is found in nearly every tumor type, and specific MMPs have been associated with many aspects of tumor biology, including tissue invasion, metastasis, and neoangiogenesis. Our results provide insight into a new mechanism by which MMP expression can promote tumor progression. In previous studies using cultured mammary epithelial cells, we demonstrated that treatment with MMP-3 stimulated induction of Rac1b and consequent activation of the EMT transcriptional program (11). Herein, we have identified Rac1b as an inducer of EMT in vivo and as a key intermediate in MMP-induced lung cancer progression. The EMT program can be initiated by a variety of intrinsic signals (for example, gene mutations) or extrinsic signals (such as growth factor signaling). Because EMT characteristics are now being recognized in human cancers, there is an increasing emphasis toward identifying therapeutic approaches that target EMT. It has been shown that the cellular changes associated with EMT-like transitions alter the dependence of carcinoma cells on epidermal growth factor receptor (EGFR) signaling networks and are associated with NSCLC tumor cell, xenograft, and patient insensitivity to selective EGFR tyrosine kinase inhibitors (TKIs) (48-50). Novel therapeutics that inhibit pathological inducers of EMT may sensitize tumor cells and enhance response to existing therapies, whereas new therapies that are targeted specifically against downstream EMT-directed pathways might independently reduce tumor progression and metastasis of lung cancer.
Rac1b, in particular, shows promise as a point of therapeutic intervention. In cultured cells, Rac1b displays many of the properties of GTPase-deficient Rac1 mutants, but the presence of a 19–amino acid insertion enables Rac1b to interact with a distinct set of effector proteins, potentially contributing to loss of the epithelial phenotype (10, 51-53). From the viewpoint of therapeutic targeting, Rac1b has been identified uniquely in tumors and cancer cell lines and has no known functions in physiological tissue homeostasis. In addition, it seems likely that the unique structural and dynamic properties conferred by the 19–amino acid insertion may also offer the opportunity for selective targeting of Rac1b in preference to Rac1 by novel small-molecule or protein therapeutics that block EMT-associated progression to advanced tumor stages.
EMT is a physiological process best described in embryonic development (54). Although physiological activation of the complete EMT program is a highly coordinated process involving multiple mediators, there is an increasing consensus that aberrant activation of EMT inducers in the tumor microenvironment can stimulate aspects of the EMT program that drive tumor progression (55). Early investigations into this process have focused on how acquisition of motility characteristics facilitates tumor metastasis; however, other consequences of EMT induction have also been identified, including development of fibrosis, increased resistance to apoptosis, and the markedly enhanced tumorigenic potential of the cancer stem cell phenotype (56, 57). Production of myofibroblasts by EMT can also result in a stiffened extracellular matrix, which can potentiate tumor progression by disrupting tissue structure and activating cellular invasiveness (58). Experiments in cultured cells have indicated that EMT mediators could induce bypass of senescence (22), but previous studies did not indicate whether or when this process might happen in vivo. Our results implicate EMT-induced senescence bypass as a key step in lung cancer progression and identify Rac1b as a key mediator of this process. Furthermore, our findings that Rac1b is expressed in stages 1 and 2 of lung tumor progression (Fig. 1, A and B) and in response to cigarette smoke (Fig. 1, E and G) combined with the similarities between MMP-3/Rac1b-induced molecular alterations in transgenic mice and lung tissue from human smokers (fig. S8) indicate that cigarette smoke–induced changes in lung tissue that drive lung cancer may occur through MMP-induced Rac1b. Our new experimental models allow for the development of new therapeutic strategies that target these processes.
tetO-HA-MMP-3 mice were engineered by cloning the coding sequence of MMP-3 (59) downstream of the Tet operator in pTet-Splice (Promega). tetO-YFP-Rac1b mice were similarly engineered with the coding sequence of Rac1b (11). tetO-HA-MMP-3 and tetO-YFP-Rac1b founder lines were generated by pronuclear injection into FVB mouse oocytes by standard transgenic procedures (Mayo Transgenic Core). Mice were maintained as heterozygotes by breeding with nontransgenic FVB mice (The Jackson Laboratory). Transgenic mice bearing the tetO-YFP-Rac1b or tetO-HA-MMP-3 transgenes had no detectable tissue pathologies and survived normally in the vivarium. To produce mice with regulatable MMP-3 or Rac1b production, we mated tetO-HA-MMP-3 or tetO-YFP-Rac1b transgenic mice with FVB mice bearing the CCSP-rtta activator transgene (obtained from J. Whitsett) that express the rtTA under the control of the rat CCSP (rCCSP) promoter (23). For urethane treatment, 6-week-old mice were weighed and given a single intra-peritoneal injection of 25 mg of urethane (300 μl of an 83.3 mg/ml solution; Sigma) for an approximate dose of 200 mg/kg. The injection was repeated (20 mg) on a weekly basis for 2 or 5 weeks as indicated. At either 14 or 20 weeks after injection, mice were euthanized by CO2 asphyxiation. The lungs were removed, formalin-fixed, and paraffin-embedded. Because significant gender differences have been observed in response to urethane (60-62), results presented are for male mice only.
The mice were housed in pathogen-free conditions according to protocols approved by the Institutional Animal Care and Use Committee at Mayo Clinic. For induction of tetO-HA-MMP-3 and tetO-YFP-Rac1b transgene expression in double-transgenic animals, the mice were fed doxycycline (Bio-Serv) at a concentration of 200 mg/kg.
Mice were euthanized, and the lungs were removed intact and fixed overnight in 10% buffered formalin. For histological analysis, lungs were embedded in paraffin, serially sectioned (5 μm), and stained with H&E. For IHC, sections were deparaffinized by placing slides into three changes of xylene and rehydrated in a graded ethanol series. The rehydrated tissue samples were rinsed in water, and most sections were subjected to heat antigen retrieval as described by the manufacturer (Dako) at either standard pH (6.9) or high pH (9.0). Slides were incubated with the appropriate primary antibody (see Supplementary Materials and Methods) with anti-rabbit–labeled polymer–horseradish peroxidase (HRP) (Dako) or, for anti-Ki67, the Rat-on-Mouse HRP Kit (Biocare).
Mouse pulmonary lesions were classified according to the recommendations of Nikitin et al. (63). AAH was identified as focal and diffuse lesions consisting of relatively uniform atypical cuboidal to columnar cells with dense chromatin. Lesions with well-circumscribed areas smaller than 5 mm in diameter were considered adenomas. Adenomas were categorized as solid (no evidence of papillary structures), mixed (both papillary and solid structures present), or papillary (consists primarily of papillary structures lined by cuboidal to columnar cells). Adenocarcinomas showed greater cytological atypia, increased frequency of mitoses, regional variation in growth pattern, more papillary structures, size greater than 5 mm in diameter, and invasion of vessels or large airways.
The human H661 large cell lung cancer line was obtained from the American Type Culture Collection and maintained as suggested by the supplier. Methods for selective siRNA knockdown of Rac1b were described previously (11).
Murine alveolar epithelial type II cells were isolated from adult double-transgenic tetO-HA-MMP-3/CCSP-rtta mice (6 to 12 weeks old) following the method of Rice et al. (64) and were cultured on tissue culture plates coated with Matrigel (BD Biosciences). Cells were maintained in bronchial epithelial cell growth medium (BEGM; Clonetics) with 5% charcoal- and dextran-treated fetal bovine serum (FBS) in the presence or absence of doxycycline hydrochloride (Sigma) in a 37°C, 5% CO2 incubator. Branching morphogenesis was assessed as the fraction of cell clusters per field with two or more extensions longer than the diameter of the central cell cluster.
Lung tumor tissues were analyzed by a pathologist to confirm initial diagnosis, staging, and overall integrity of the tissue samples. Seventy-four cases of NSCLC [37 LAC and 37 squamous cell carcinoma (SCC)] were chosen for extraction of RNA. Total RNA was isolated from cryostat sections of tumor and matched normal lung tissue with the RNAqueous-4PCR kit (Ambion). Reverse transcription–qPCR (RT-qPCR) analysis was performed on a 7900 HT Thermocycler (Applied Biosystems) with TaqMan Master Mix and primers for 18S complementary DNA (cDNA) purchased from Applied Biosystems. The primer/probe set for the human Rac1b transcript was a custom assay (forward primer, 5′-TATGACA-GATTACGCCCCCTATC-3′; reverse primer, 5′-CTTTGCCCCGGGA-GGTTA-3′; and probe, 5′-AAACGTACGGTAAGGAT-3′).
One central 5-μm section of each set of paraffin-embedded lungs was stained by H&E. Digital images of each lung section were captured with the Aperio ScanScope XT slide scanner (Aperio Technologies), and tumors were countered, measured, and categorized according to morphology (63).
We thank J. Campisi and J. P. Kirkland for helpful discussions, A. Hockla and J. Robinson for MMP enzyme preparations, D. Khauv for assistance with figures, and B. Edenfield for immunohistochemical analysis of lung tissues.
Funding: This work was supported by grants from the National Cancer Institute (CA122086 to D.C.R., CA128660 to C.M.N. and D.C.R., and CA081436 to A.P.F.) and the National Heart, Lung, and Blood Institute (HL090669 to G.P.D.), the James and Esther King Foundation (to A.P.F., D.C.R., and E.S.R.), funds from the Harold and Mary Zirin Chair in Pulmonary Biology at National Jewish Health, and the Monica Flynn Jacoby Professorship (to A.P.F.).
Author contributions: M.L.S.-M., J.W., M.L.G., E.S.R., A.P.F., and D.C.R. designed the experiments; M.L.S.-M., J.W., Y.Z., E.M., and D.C.R. performed the experiments; M.L.S.-M., J.W., M.L.G., D.W.V., G.P.D., E.S.R., A.P.F., and D.C.R. interpreted the experiments and assembled the figures; M.L.S.-M. and D.C.R. wrote the manuscript; and all authors participated in editing and refinement of the manuscript.
Competing interests: The authors declare that they have no competing financial interests.
Materials and Methods
Fig. S1. Expression of Rac1b mRNA in human lung adenocarcinoma homogenates.
Fig. S2. Altered expression of mesenchymal and epithelial markers in CCSP-rtTA/tet-Rac1b transgenic mouse tumors.
Fig. S3. Coexpression of Rac1b transgene and vimentin in CCSP-rtTA/tet-Rac1b transgenic mouse tumors.
Fig. S4. Expression of Rac1b protein is inhibited by siRNA for Rac1b in H661 cells.
Fig. S5. Classification of neoplasms in mouse lung.
Fig. S6. Adenocarcinoma from a urethane-treated tetO-HA-MMP-3/CCSP-rtta bitransgenic mouse.
Fig. S7. Coexpression of the tetO-HA-MMP-3 transgene and the mesenchymal protein vimentin.
Fig. S8. Gene expression profiles of RNA isolated from lungs of tetO-HA-MMP-3/CCSP-rtta or CCSP-rtta mice treated with urethane and exposed to doxycycline.
Fig. S9. Induction of morphological alterations in primary lung epithelial cells by transgenic expression of MMP-3.
Fig. S10. Progression to adenocarcinoma in tetO-HA-MMP-3/CCSP-rtta mice in the urethane-induced tumorigenesis model.
Fig. S11. Decrease in expression of senescence markers upon MMP-3 expression in tetO-HA-MMP-3/CCSP-rtta mouse lung tissue in the urethane-induced tumorigenesis model.
Fig. S12. Potentiation of tumor formation in K-RasLA2 mice by expression of MMP-3.
Fig. S13. Immunohistochemical analyses of paraffin-embedded tissues with primary antibodies.
Data set S1. Genes that differ more than twofold in any condition in lung tissue homogenates from tetO-HA-MMP-3/CCSP-rtta and control mice (Excel file).
Data set S2. Genes that are differentially expressed (P < 0.05) in lung tissue homogenates in tetO-HA-MMP-3/CCSP-rtta versus control mice (Excel file).
Data set S3. Gene ontology analysis of differentially expressed genes (P < 0.05) from lung tissue homogenates of tetO-HA-MMP-3/CCSP-rtta versus control mice (Excel file).