Inhibition of 11β-Hydroxysteroid Dehydrogenase Type II Suppresses Lung Carcinogenesis by Blocking Tumor COX-2 Expression as Well as the ERK and mTOR Signaling Pathways

Lung cancer is by far the leading cause of cancer death. Early diagnosis and prevention remain the best approach to reduce the overall morbidity and mortality. Experimental and clinical evidence have shown that cyclooxygenase-2 (COX-2) derived prostaglandin E2 (PGE2) contributes to lung tumorigenesis. COX-2 inhibitors suppress the development and progression of lung cancer. However, increased cardiovascular risks of COX-2 inhibitors limit their use in chemoprevention of lung cancers. Glucocorticoids are endogenous and potent COX-2 inhibitors, and their local actions are down-regulated by 11β–hydroxysteroid dehydrogenase type II (11ßHSD2)-mediated metabolism. We found that 11βHSD2 expression was increased in human lung cancers and experimental lung tumors. Inhibition of 11βHSD2 activity enhanced glucocorticoid-mediated COX-2 inhibition in human lung carcinoma cells. Furthermore, 11βHSD2 inhibition suppressed lung tumor growth and invasion in association with increased tissue active glucocorticoid levels, decreased COX-2 expression, inhibition of ERK and mTOR signaling pathways, increased tumor endoplasmic reticulum stress as well as increased lifespan. Therefore, 11βHSD2 inhibition represents a novel approach for lung cancer chemoprevention and therapy by increasing tumor glucocorticoid activity, which in turn selectively blocks local COX-2 activity and/or inhibits the ERK and mTOR signaling pathways.

Introduction 11ßHSD2 activity could suppress lung tumorigenesis due to increased tumor cell intracellular active glucocorticoids and subsequent inhibition of COX-2 expression/activity.

Ethics Statement
All animal experiments were performed according to animal care guidelines and were approved by the Vanderbilt Institutional Animal Care and Use Committe (IACUC) (M/11/130).

Animals
KrasLA2 mice were a gift of Dr. T. Jacks, MIT [24]. Since the KrasLA allele is non-functional in the germline configuration, only heterozygous mice were maintained and used in the experiments. Age-and sex-matched KrasLA2 mice were treated with water (control) or the 11βHSD2 inhibitor, glyccyrrhetinic acid (GA, 10 mg/kg/day, i.p.) from 6 to 20 weeks of age and sacrificed at 20 weeks of age. Under anesthesia with Nembutal (60 mg/kg i.p.), the lungs were weighed, filled with fixative overnight, transferred to 70% ethanol for 24 h, and examined using a dissecting microscope to count surface polyps in a blinded fashion [22]. The tumor diameter on the surface of the lung was measured with a digital caliper. Tumor volume was expressed in mm 3 and calculated as [l x w 2 ] × 0.5, where l and w denote length and width, respectively [22]. After tumors were counted, lung tissues were processed for paraffin embedding and immunohistochemical analysis. A subset of mice was sacrificed for collection of adenomas and lung tissue for immunoblotting and determination of tissue levels of corticosterone and 11-keto-corticosterone. All animals were genotyped twice (after weaning and before sacrifice).

Survival analysis
KrasLA2 mice were treated with water or 11βHSD2 inhibitor, GA at 6 weeks of age. The animals were monitored daily. The animal was sacrificed at the first sign of shortness of breath, reduced locomotion and reduced body weight (greater than 20% total body weight). Lungs from these mice were analyzed for the presence of visible tumors on the lung surface. Only mice with visible lung tumors were used for the survival analysis. The animals were euthanized by overdose carbon dioxide.

Human lung cancer tissues
Slides of lung cancer tissue array with normal lung tissue were purchased from US Biomax (Rockville, MD).

Cell culture
Lewis lung carcinoma (LLC) cells, H2073, H1435, H25, A549, H1944, H1792, H1793, H1299, and H1395 cells were grown in RPMI1640 supplemented with 4,500 mg/l glucose, 2 mM l-glutamine, 10% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin in 5% CO2 and 95% air at 37°C [25]. All cell lines were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). LLC cells or H1435 cells were cultured in medium with 1% fetal bovine serum overnight and then treated with corticosterone (CS) with or without 11βHSD2 inhibitors, GA (10 μM) or carbenoxolone (CBX, 10 μM), for 8 hours and cells were harvested for COX-2 immunoblotting. The cells were seeded in a 96-well plate (5 x 10 3 cells in 200 μl medium) for 16 hours, then starved for 24 hours in medium containing 0.1% fetal bovine serum, and various concentrations of corticosterone (CS) or CS plus 10μM GA were added for additional 4 hours. Then BrdU solution was added to each well for 4 hours. Analysis was performed according to the manufacturer's protocol (BrdU cell proliferation assay kit, Cat#6813, Cell Signaling) and our recent report [26].

Primary tumor growth
For mouse LLC tumor experiments, male C57/B6 mice (8-12 weeks old) were given 2 dorsal s. c. injections (5 x 10 5 cells in 100 μl PBS) with administration of water (control, i.p) or GA (10 mg/kg/day, i.p.) begun 1 day prior to cell injection. Tumors were harvested and weighed after 18 days of growth. In another set of experiment, LLC cell suspension containing 5 x 10 5 cells in 100 μl PBS were injected into tail vein with or without GA administration as described above. The time of death, or when mice became moribund (by the same criteria as the endpoints described for survival studies), was recorded for survival analysis.

Measurement of corticosterone and 11-keto-corticosterone
Lung tissues from KrasLA2 mice were collected and stored at-80°C. Tissue levels of corticosterone (active form in rodent) and 11-keto-corticosterone (inactive form) were measured using high-performance liquid chromatography coupled with electrospray tandem mass spectrometry [22]. Corticosterone was from ICN Biomedicals, and 11-keto-corticosterone was from Steraloids.

Immunohistochemical staining and immunoblotting
Immunostaining was carried out as in previous reports [27]. For immunostaining of phosphorylated proteins, antigen retrieval was achieved by boiling in citric acid buffer (100 mM, pH 6.0) for 3 x 5 min. Lung tumors isolated from 20 weeks old KrasLA2 mice were used for immunoblotting as described previously [28].

Quantitative image analysis
Immunostaining was quantified by using the BIOQUANT image analysis system (R & M Biometrics, Nashville, TN) [22]. Bright-field images from a Leitz Orthoplan microscope with DVC digital RGB video camera were digitized and saved as computer files. Contrast and color level adjustment (Adobe Photoshop) were performed for the entire image; i.e., no region-or object-specific editing or enhancements were performed.

Statistical analysis
Values are presented as means ± S.E.M. ANOVA and Bonferroni 2-tailed t-test were used for statistical analysis, and differences were considered significant for P values < 0.05.

Results
To investigate whether 11ßHSD2 might be involved in lung tumorigenesis, its expression in normal mouse lung tissue, experimental lung cancer, normal human lung tissue and different human lung cancers was investigated with immunohistochemistry. In athymic nude mouse lung tissue, 11ßHSD2 was expressed in the epithelial cells of small airways and alveoli ( Fig 1A). 11ßHSD2 expression was increased in A549 adenocarcinoma compared to adjacent normal athymic nude mouse lung tissue ( Fig 1B). 11ßHSD2 immunostaining was strong in human adenocarcinoma and squamous cell carcinoma, moderate in papillary carcinoma as well as small cell lung cancer, but very weak in normal human lung tissue ( Fig 1C). Negative control was performed by replacing the primary antibody with purified rabbit IgG and no immunoreactivity was found (data not shown). Therefore, 11ßHSD2 expression is increased in both experimental lung cancers and human lung cancers.
To investigate whether inhibition of 11ßHSD2 activity was beneficial in lung tumorigenesis through inhibition of COX-2 expression due to increased intracellular active glucocorticoids, we screened lung cancer cell lines for 11ßHSD2 and COX-2 expression. As indicated in Fig 2A, 11ßHSD2 was expressed in all investigated lung cancer cell lines; while COX-2 was only detected in LLC, H1435, A549 and H1395 cells. LLC cell line, which is highly tumorigenic in mice and expresses both COX-2 and 11ßHSD2, was chosen to investigate the role of 11ßHSD2 inhibition in COX-2 expression. As indicated in Fig 2B, corticosterone (CS, glucocorticoid in rodent) inhibited LLC cell COX-2 expression in a dose-dependent pattern. However, even 1μM CS could not completely inhibited COX-2 expression. One possible reason for this partial inhibition is that intracellular CS is quickly inactivated by 11ßHSD2. It is also possible that some COX-2 is insensitive to glucocorticoid inhibition. CS-induced LLC cell COX-2 inhibition was markedly enhanced in the presence of GA, an inhibitor of 11ßHSD2 activity (Fig 2C). In addition, CS-induced LLC COX-2 inhibition was also enhanced by carbenoxolone (CBX, 10 μM), another inhibitor of 11ßHSD2 activity (Fig 2D) [29]. Furthermore, CS-induced COX-2 inhibition was also enhanced by GA treatment in both H1435 and A549 cells, which also express both 11ßHSD2 and COX-2 (Fig 2E and 2F).
To determine the effect of 11ßHSD2 inhibition on CS-mediated inhibition of mTOR and ERK signaling pathways, both LLC cells and A549 cells were treated with 1 μM CS with or without 10 μM GA for 24 hours and p-mTOR and p-ERK levels were determined with immunoblotting. As indicated in Fig 3A, CS treatment resulted in decreases in both p-mTOR and p-ERK levels in LLC cells, which were enhanced by 11ßHSD2 inhibition with GA. Similarly, CS-mediated inhibition of p-mTOR and p-ERK was also augmented by GA treatment in A549 cells (Fig 3B).
To investigate the direct effects of CS on lung cancer cell proliferation, we treated LLC cells (mouse) and A549 cells (human) with different dose of CS and cell proliferation was measured by using cell proliferation assay kit. As shown in Fig 3C, CS inhibited LLC cell proliferation in a dose-dependent pattern. In addition, CS-mediated inhibition was significantly enhanced by addition of GA. Similarly, CS-mediated inhibition of A549 cell proliferation was also augmented by GA treatment in A549 cells (Fig 3D). To determine whether 11ßHSD2 inhibition affected lung tumorigenesis, we examined LLC tumor growth in a Xenograft model. LLC cells were injected subcutaneously into C57BL/6 mice. GA treatment (10 mg/kg/day, i.p.) was initiated one day before LLC cell injection and tumors were harvested and weighted at the end of the experiment (18 days after injection). As shown in Fig 4A, LLC tumor growth was markedly attenuated by GA treatment (tumor weight: 414 ± 38 vs. 646 ± 58 mg of vehicle, P < 0.01, n = 8 in each group). Similarly, GA treatment inhibited tumor invasion to the lung in a LLC tail vein injection model. As indicated in Fig 4B, GA treatment led to marked increase in survival probability (average survival days: 34.8 ± 3.3 vs. 24.5 ± 3.5 of control, P < 0.001, n = 12 in vehicle group and n = 13 in GA group).
KrasLA2 mice carry an oncogenic mutation of the Kras gene (G12D) and develop spontaneous primary tumors with features of NSCLC [24]. To further investigate the role of 11ßHSD2 inhibition in lung tumorigenesis, heterozygous KrasLA2 mice were treated with water (control) or GA (10 mg/kg/day, i.p.) from 6 weeks of age. At 20 weeks of age, a subset of mice was sacrificed, lung tissue kept for measurement of glucocorticoid levels and immunoblotting. Another subset of mice was sacrificed and the lungs were filled with fixative to fix the tissue for analysis of tumor number, size and immunostaining. Lung tissue levels of glucocorticoids were measured by using high-performance liquid chromatography coupled with electrospray tandem mass spectrometry [22]. GA treatment led to marked increases in lung corticosterone levels (active glucocorticoid in rodent: 37.40 ± 12.50 vs. 5.06 ± 2.19 ng/g lung tissue of control, P < 0.05, n = 6 in each group), and led to marked decreases in 11-keto-corticosterone (inactive glucocorticoid in rodent: 1.05 ± 0.08 vs. 2.63 ± 0.64 ng/g lung tissue of control, P < 0.05) ( Fig  5A). As expected, COX-2 levels were markedly reduced in lung tumors from GA treated Kra-sLA2 mice (26 ± 5% of control, P < 0.001, n = 3 in each group) (Fig 5B). Immunostaining confirmed the reduction of tumor COX-2 expression in GA treatment mice. Glucocorticoids have been reported to induce mannose receptor (MR) expression in macrophages [30]. Therefore, we investigated MR expression in the tumor. As indicated in Fig 5C, the number of tumor MR- GA treatment caused significant inhibition of tumorigenesis in KrasLA2 mice, including decrease in adenoma number on the lung surface (40.0 ± 5.8 vs. 72.7 ± 7.8 of control, P < 0.01, n = 9 in each group) and decrease in adenoma size (mm 3 /mouse: 40.7 ± 10.7 vs. 168.8 ± 41 of vehicle, P < 0.01) (Fig 6A). GA treatment markedly increased lifespan of KrasLA2 mice (weeks: 42.1 ± 2.6 vs. 24.1 ± 1.0 of control, P < 0.01, n = 10 in GA group and n = 25 in control group) (Fig 6B). Therefore, GA treatment inhibited lung tumorigenesis in KrasLA2 mice and increased KrasLA2 mouse lifespan.
Glucocorticoids have been reported to inhibit lung cancer cell growth through inhibition of ERK activity [31,32]. Therefore, we investigated tumor ERK activity with immunostaining of phosphorylated ERK. As indicated in Fig 6C, the levels of tumor phosphorylated ERK were marked reduced by GA treatment (ratio of p-ERK area/tumor area: 1.79 ± 0.79 vs. 16.11 ± 3.61 of control, P < 0.001, n = 4 in each group). GA treatment also inhibited tumor mammalian target of rapamycin (mTOR) signaling pathway (Fig 6D). Finally, tumor cell proliferation was inhibited by GA treatment as indicated by decreased tumor cells with Ki67 positivity, a marker of cell proliferation (Fig 6E).
Recently, it was recognized that induction of endoplasmic reticulum stress (ER stress) and activation of the unfolded protein response (UPR) is an important pathway for antitumor agents to induce cancer cell death [33][34][35]. ER-stress has been reported to contribute to apoptosis induced by COX-2 inhibition in lung cancer cells as well as other cancer cells [36,37]. Furthermore, ER stress-mediated apoptosis requires eIF2α and PERK. Upon activation, PERK phosphorylates eIF2α to reduce global mRNA translation [38]. Therefore, we investigated the potential role of ER stress in 11ßHSD2 inhibition-induced suppression of lung tumorigenesis. C/EBP homologous protein (CHOP) plays an important role in ER stress induced apoptosis [36,37]. 11ßHSD2 inhibition with GA led to marked increases in BIP/GPR78 (a marker of ER stress) and CHOP expression in tumors from KrasLA2 mice ( ÃÃ P < 0.01, n = 4) (Fig 7A). In addition, 11ßHSD2 inhibition with GA led to increased phosphorylation of both PERK and eIF2α (Fig 7B).
To determine whether suppression of lung tumorigenesis by 11ßHSD2 inhibition was related to decreased cell proliferation and/or increased apoptosis, the expression levels of cyclin D1 (marker of proliferation) and cleaved caspase-3 (a specific marker of apoptosis) were investigated in KrasLA2 mouse tumors. GA treatment caused significant decreases in both cyclin D1 immunostaining density and the number of cyclin D1 positive cells in KrasLA2 lung tumors (cells/hpf: 31.1 ± 3.8 vs. 57.4 ± 4.0 of control, P < 0.01, n = 4). In contrast, the number of tumor cells that were positive for cleaved caspase-3 was significantly increased in GA-treated KrasLA2 lung tumors (cells/hpf: 59.8 ± 2.9 vs. 12.5 ± 5.2 of control, P < 0.01, n = 4) (Fig 7C). Therefore, inhibition of 11ßHSD2 activity with GA led to decreased cell proliferation and increased apoptosis in lung tumors.

Discussion
The major findings in the present studies include: 1) 11ßHSD2 is expressed in epithelial cells of small airways and alveoli of mouse lung and its expression is increased in experimental lung tumors and in human SCLC and NSCLC, 2) glucocorticoid-induced COX-2 inhibition in lung cancer cells is enhanced by 11ßHSD2 inhibition, 3) inhibition of 11ßHSD2 activity with GA reduces LLC tumor growth and invasion, and 4) 11ßHSD2 inhibition with GA suppresses lung tumorigenesis and increases lifespan in KrasLA2 mice in association with increases in lung active corticosterone levels, decreases in tumor COX-2 expression, inhibition of the tumor ERK and mTOR signaling pathways and increased ER stress.
There is a clear link between activation of the COX-2/mPGES-1/PGE 2 pathway and lung tumorigenesis and progression. In NSCLC A549 cells, COX-2-derived PGE 2 promotes cell migration, cell proliferation and apoptosis resistance, and COX-2 inhibition suppresses cell proliferation associating with inhibition of survivin expression and increased caspase-3 mediated apoptosis [39][40][41]. The terminal enzyme of COX-2-mediated PGE 2 biosynthesis, mPGES-1, plays a key role in LLC cell proliferation in vitro and tumor growth and metastasis in vivo [42]. COX-2 expression increased in experimental lung cancer models and COX-2 inhibition suppressed tumor development and growth in these models [43,44]. Selective COX-2 inhibition may act synergistically with ionizing radiation to inhibit A549 cancer cells through the activation of caspase-8 and caspase-3 [45]. Knockdown of mPGES-1 markedly reduced Xenograft A549 tumor growth [46].

11ßHSD2 and Lung Tumorigenesis
GCs are known to inhibit cell proliferation and induce cell differentiation through activation of glucocorticoid receptors. 11ßHSD2 has been thought to be pro-proliferative due to its ability to inactivate glucocorticoids [47][48][49][50]. We have recently reported that 11ßHSD2 expression is increased in epithelial cells and stromal cells in human colonic and Apc +/min mouse intestinal adenomas and is correlated with increased COX-2 expression and activity, and that inhibition of 11ßHSD2 activity genetically or pharmacologically suppresses tumor COX-2 pathway and prevents adenoma formation, tumor growth, and metastasis as a result of increased tumor intracellular active glucocorticoids [22,51]. Sustained activation of the ERK signaling is important for lung cancer cell survival and proliferation [52,53]. GCs have been reported to inhibit lung cancer cell growth through inhibition of the ERK signaling pathway. Indeed, GA treatment led to inhibition of the tumor ERK signaling pathway in KrasLA2 mice (Fig 5C). Oncogenic K-Ras regulates proliferation and cell functions in lung epithelial cells through induction of cyclooxygenase-2 and activation of metalloproteinase-9 [54]. In the current study, inhibition of 11ßHSD2 activity led to inhibition of lung tumorigenesis in KrasLA2 mice associated with COX-2 inhibition and increased ER stress as well as inhibition of the ERK and mTOR signaling pathways due to increased tumor cell active glucocorticoids (Fig 8). Therefore inhibition of 11ßHSD2 activity with glycyrrhizic acid and its analogs may represent a novel approach for lung cancer chemoprevention, particularly in long-term heavy cigarette smokers, with the following advantages [22]: 1) GA, a natural compound contained in licorice, is a nontoxic, inexpensive and powerful 11ßHSD2 inhibitor; 2) Physiologic 11ßHSD2 expression is largely restricted to colon and kidney and lung. Therefore, inhibition of 11ßHSD2 activity is not expected to incur the cardiovascular risk posed by COX-2 inhibitors that suppress COX-2-derived PGI 2 production in vascular endothelial cells; 3) Intracellular active GCs are only increased in tissues with elevated 11ßHSD2 expression. 11ßHSD2 inhibition will not produce immunosuppression or other systemic side effects of conventional glucocorticoid therapy; 4) In addition to inhibiting the COX-2 pathway, increased tumor active GCs also inhibit lung tumorigenesis through inhibiting the ERK and mTOR signaling pathways as well as induction of G1 cell cycle arrest [51]. Although 11ßHSD2 inhibition may result in salt-sensitive hypertension due to activation of mineralocorticoid receptors by GCs, development of locally acting 11ßHSD2 inhibitors that are not systemically absorbed would be a potential therapeutic means to prevent lung tumorigenesis [55]. In the future, it is worthy investigating whether MEK inhibitor or rapamycin can enhance GA-mediated anti-proliferative effect.