LKB1/PEA3/ΔNp63 pathway regulates PTGS-2 (COX-2) transcription in lung cancer cells upon cigarette smoke exposure.

This is the first study to show that cigarette smoking induced the LKB1/PEA3/ΔNp63-dependent transcriptional regulation of inflammatory molecules, such as COX-2/PTGS-2. Using mainstream smoke extract (MSE) and sidestream smoke extract (SSE) as modeling tools for primary and second-hand smoking, we found that both MSE and SSE down regulated protein levels for LKB1, while up regulated protein levels for PEA3 and COX-2 in a dose-dependent manner. Using the endogenous ChIP analysis, we further found that the C/EBPβ, NF-kB, NF-Y (CHOP), PEA3 (ETS), and ΔNp63 proteins bound to the specific area (-550 to -130) of the COX-2 promoter, while forming multiple protein complexes in lung cancer cells exposed to MSE and SSE. Our results define a novel link between various transcription factors occupying the COX-2 promoter and cellular response to cigarette smoke exposure bringing a new component, ΔNp63α, showing a critical role for cooperation between various chromatin components in regulation of COX-2 expression and, therefore strengthening the central role of inflammatory process in tumorigenesis of epithelial cells, especially after cigarette smoke exposure (both primary and second-hand.

1. Introduction

Tobacco smoke contains over 4,700 chemical components that have been implicated in the etiology of oxidative stress-related diseases e.g., chronic obstructive pulmonary disease, Parkinson's disease, Alzheimer disease, asthma, cancer and cardiovascular disease [1-7]. Epidemiological studies support the notion that lung cancers are directly caused by cigarette primary and secondhand smoking [1-3]. Carcinogen exposure and chronic inflammation are two important events in tumor development, and both have been implicated in the development of many human epithelial cancers [4-6]. Exposure to external factors including cigarette smoking, infectious agents, dietary carcinogens and hormonal imbalances could injure the tissue (e.g., lung) and lead to chronic inflammation [5, 6]. At the cellular and molecular levels, cigarette smoking might induce oxidative stress and DNA damage, implicated in the etiology of cancer and resulting in modulation of reactive oxygen species (ROS) production and the cell's own antioxidant defenses, therefore leading to activation of numerous signaling pathways underlying apoptosis and autophagy [8-10].

Epithelial/mesenchymal transition (EMT) and increased cell motility/migratory/invasive phenotype were also found to occur during the development and progression of lung epithelial cancers [11-13]. Thus, the understanding of mechanisms underlying these processes (apoptosis, angiogenesis, cell migration, invasiveness) in lung cancer would assist development of new therapeutic strategies [5, 6, 13].

Studies of genetic mechanisms underlying lung cancer, along with other human cancers, demonstrated that tobacco exposure is causing inactivation of tumor suppressor genes via genetic/epigenetic changes affecting many cellular processes [1, 13, 14]. LKB1 tumor suppressor gene (also known as serine/threonine kinase-11, STK11) is capable to regulate other protein's function by phosphorylation, thereby affecting cell proliferation and survival [15-21]. Smoking has been linked to human epithelial cancers, which overexpress proteins implicated in inflammatory signaling pathways [e.g., NFκB, cyclin D1, cyclooxygenase (COX)-2 (also known as PTGS-2, prostaglandin-endoperoxide synthase-2)] [7, 22–32].

We previously showed that that LKB1 physically and functionally associates with PEA3 leading to the PEA3 phosphorylation and subsequent PEA3 protein degradation via proteasomedependent pathway [16]. We also showed that the downregulated PEA3 expression and activity leads to a subsequent downregulation of COX-2/PTGS-2 expression [16]. We further showed that cells expressing mutant LKB1 deficient of kinase activity failed to downregulate PEA3 and activate COX-2 transcription, while increased cell invasiveness compared to cells with wild-type LKB1 [16]. Similarly, LKB1 knockdown by siRNA dramatically increased migration/invasiveness shown by lung cancer cells in vitro. However, lung cancer cells transfected with PEA3siRNA displayed decreased invasiveness, while the PEA3 forced expression resulted in decrease of epithelial markers and increase of mesenchymal markers suggesting that PEA3 stabilization due to LKB1 inactivation leads to a greater cancer cell invasiveness [16].

In the current study, we provide the first evidence that cigarette smoke treatment [mainstream smoke extract (MSE) and sidestream smoke extract (SSE)] of cultured lung normal and cancer cells decreased LKB1 expression and elevated expression of both PEA3 (polyomavirus enhancer activator 2, also known as E1A enhancer binding protein 4) factor and COX-2/ PTGS-2 inflammatory signaling molecule. We showed that cigarette smoking affects the molecular processes underlying EMT of lung cancer/epithelial cells and involving LKB1 inactivation and PEA3/ΔNp63-mediated regulation of COX-2/PTGS-2. We defined a novel LKB1/PEA3/ΔNp63α molecular pathway leading to COX-2/PTGS-2 regulation in human lung cancer cells exposed to mainstream and sidestream smokes therefore linking ΔNp63α to tumorigenesis and inflammation processes as a new biomarker for oxidative stress and DNA damage [33].

2. Results

LKB1 is downregulated in lung tumor samples from patients affected by smoking. Recent report shows that normal human lung epithelial tissues overexpress LKB1, while in primary lung tumors (squamous cell carcinomas), LKB1 levels were dramatically downregulated [15, 16]. We previously showed that LKB1 overexpression led to a binding to and subsequent phosphorylation of the PEA3 transcription factor followed by downregulation of PEA3 and in turn inhibition of COX-2/PTGS-2 expression [16]. In this study, we examined whether cigarette smoking exposure of lung normal and cancer cells affect the expression levels of LKB1, PEA3 and COX-2. We exposed normal human bronchiolar epithelial (NHBE) cells and H1299 lung cancer cells to 0.5% MSE for 48 h, and found that LKB1 protein levels were downregulated, while levels for PEA3 and COX-2/PTGS-2 were upregulated in both cell lines upon MSE exposure (Fig. 1A). We found that H1299 cells exposed to MSE in dose-dependent manner displayed altered expression of LKB1, PEA3 and COX-2/PTGS-2 (Fig. 1B). We further examined the effect of SSE on the protein levels of LKB1, PEA3 and COX-2/PTGS-2 in H1299 cells. We then found that 1% SSE added to cells for 48 h decreased the LKB1 protein levels, while upregulated the PEA3 and COX-2 protein levels (Fig. 2). We thus observed that both MSE and SSE decreased LKB1 protein levels in lung cancer cells.

PEA3 physically associates with the COX-2 specific transcription factors in lung cancer cells upon cigarette smoke exposure. We further examined molecular mechanisms underlying LKB1 downregulation and its effects on PEA3 transcriptional regulation of COX-2/PTGS-2 in CSE-exposed lung cancer cells and lung normal epithelial cells. We thus tested whether various putative transcription factors involved in regulation of COX-2 expression in lung cancer cells upon cigarette smoke exposure. First, we defined the consensus sequences for the potential transcription factors in the COX-2 promoter sequence (Sup. Material) using the TFSEARCH web-engine (http://mbs.cbrc.jp/rese-arch/db/TFSEARCH.html).

The following cis-regulatory elements were found in the 1,700 bp COX-2 promoter sequence: C/EBPβ, NFκB, NF-Y, PEA3 and p63 (Fig. 3).

We then accessed whether these transcription factors endogenously bind to the COX-2 promoter in lung cancer cells after cigarette smoke exposure. We found that both MSE and SSE induced binding of the C/EBPβ, NFκB, NF-YA, PEA3, p53 and p63 proteins to COX-2 promoter evidenced by ChIP assay followed by RT-PCR assay given rise to a 420 bp product (Fig. 4). Since lung cancer cells were previously shown to predominantly express ΔNp63α isoform of p63 (reviewed in ref. 34), the antibody that exclusively recognizes the ΔNp63 was used in these experiments [34, 35].

Previous reports pointed out the possibility for LKB1 to form protein-protein complexes with p53 (reviewed in ref. 36). We therefore tested whether p53 homolog ΔNp63α is forming protein complexes with LKB1 and the other transcription factors occupying the COX-2 promoter in lung cancer cells after cigarette smoke exposure. We found that, indeed, ΔNp63α formed protein-protein complexes with LKB1 in untreated cells (Fig. 5A), while the amount of these ΔNp63/LKB1 protein complexes dramatically decreased upon cigarette smoke exposure (Fig. 5A). In addition, the ΔNp63α protein associated with C/EBPβ, NFκB, NF-YA and PEA3 in cells exposed to both MSE and SSE (Fig. 5A).

We previously showed that the PEA3 ETS domain interacted with the LKB1 kinase domain pLexA-LKB1-KD (1–300) in vitro [16]. We also showed that the LKB1/ PEA3 protein complexes formed in transfected cells and endogenously in H1299 lung cancer cells found exclusively at the nucleus [16]. To further understand the effect of molecular interactions between LKB1 and PEA3 in lung cancer cells exposed to cigarette smoke extract, we performed immunoprecipitation analysis of the LKB1/ PEA3 protein-protein complexes. We thus found that, while LKB1 is downregulated in lung cancer cells upon cigarette smoke exposure by both MSE and SSE (Figs. 1 and 2), the formation of protein complexes between LKB1 and PEA3 is dramatically diminished during cellular response to cigarette smoke (Fig. 5B).

Cigarette smoking exposure induces invasiveness, foci formation and survival of lung cancer cells. Our previous results supported the notion that the PEA3 overexpression can mediate increased cell migration/cell invasiveness of lung cancer cells potentially through EMT-dependent mechanism [16].

We further examined whether MSE or SSE exposure leads to change of cellular characteristics of lung cancer cells and whether siRNA knockdown of PEA3 or COX-2 expression would affect these potential changes. A427 lung cancer cells with altered LKB1 protein kinase16 were transiently transfected with scramble siRNA (Fig. 6A–C, samples 1, 4 and 7), PEA3 siRNA(Fig. 6A–C, samples 2, 5 and 8) and COX-2/PTGS-2 (Fig. 6A–C, samples 3, 6 and 9) for 48 h. Cells then were exposed to control medium (sample 1–3), 1% SSE (samples 4–6), 0.5% MSE (samples 7–9). We then tested the ability of A427 cells to migrate in Matrigel, form foci/clones in soft agar, and undergo apoptosis/survival after cigarette smoke exposure. We found that both SSE and MSE increased cell migration/cell invasiveness (Fig. 6A, samples 4 and 7), cellular ability to form foci/ clones in soft agar (Fig. 6B, samples 4 and 7) and cell survival (Fig. 6C, samples 4 and 7). However, siRNA against PEA3 (samples 5 and 8) and COX-2 (samples 6 and 9) dramatically modulated this increase in cell migration, foci formation and cell survival (Fig. 6A–C).

3. Discussion

Cigarette primary and secondhand smoking is a direct cause of lung cancer associated with inactivation of tumor suppressor genes (e.g., LKB1) via genetic/epigenetic changes therefore affecting many cellular processes including oxidative and DNA damage, inflammatory processes, cell migration and invasiveness. Search and “assessment of novel biomarkers” and therapeutic targets affected by oxidative stress/DNA damage implicated in the cellular response to tobacco-induced pathologies “may have critical clinical utility for the formulation of novel therapeutic options” [33].

This is the first study to be undertaken to support a notion that cigarette smoking is an initiating event for loss of LKB1 expression/function and the LKB1-mediated transcriptional regulation of inflammatory molecules, such as COX-2/PTGS- 2. We hypothesized that cigarette smoking affects the molecular processes underlying EMT of lung cancer/epithelial cells, therefore linking together inactivation of LKB1 tumor suppressor and PEA3-mediated transcriptional regulation of inflammatory signaling molecules (e.g., COX-2/PTGS-2).

Using MSE and SSE as modeling tools for primary and secondhand smoking, we found that both MSE and SSE downregulated protein levels for LKB1, while upregulated protein levels for PEA3 and COX-2/PTGS-2. We further found that MSE and SSE induced these changes in a dose-dependent manner. We next determined that the COX-2/PTGS-2 promoter sequence contains many regulatory sequences recognized by key transcriptional regulators, such as C/EBPβ, NFκB, NF-YA, PEA3 (ETS), STAT and p53.

In addition to several p53 consensus sequences (-1,446 to -1,431; 1,283 to -1,266; -986 to -969 and -941 to -928), we defined a couple of p63 responsive elements (RE) in the COX-2 promoter sequence located at the positions -460 to -441 and -200 to -182.

As a member of the p53 gene family, p63 regulates downstream target gene expression by binding to sequence-specific response elements similar to those of p53 (reviewed in ref. 37). However, p63RE was shown to be distinct from the canonical p53RE, suggesting that p53 preferentially binds to the RRR CAT GYY Y sequence, whereas p63 preferentially recognizes RRR CGT GYY Y (reviewed in ref. 37). Using the endogenous ChIP analysis of the COX-2/PTGS-2 promoter in lung cancer cells upon cigarette smoke exposure (MSE and SSE), we found that the C/EBPβ, NFκB, NF-YA, PEA3 (ETS) and ΔNp63 proteins bound to the narrow area of the COX-2/PTGS-2 promoter spanning from -550 to -130 upstream of the transcription start site (Fig. 3). We then showed that after cigarette smoke exposure (MSE and SSE) of lung cancer cells, these transcription factors C/EBPβ, NFκB, NF-YA, PEA3 (ETS) and ΔNp63 formed protein complexes. While levels of ΔNp63, p-ΔNp63 (reviewed in ref. 35), C/EBPβ and PEA3 increased, the levels of NFκB and NF-YA remained the same after cigarette smoke exposure.

Several transcription factors (C/EBPβ, NFκB, STAT3, p53 and PEA3) shown to activate COX-2 expression were previously reported playing critical roles in gene expression during lung cancer cellular response to tobacco smoking exposure [21–24, 27, 28, 36–41]. Moreover, PEA3 was found to be activated by cigarette smoke exposure and regulate MMP-1 transcription by binding to a smoke response region in the distal MMP-1 promoter suggesting that PEA3 function has implications for smoking-related cancer, heart disease and emphysema [30]. Furthermore, by association with other chromatin/transcription regulators, PEA3 was shown to activate Twist expression and promote cell migration/cell invasiveness in human cancer cells suggesting its role in cancer metastasis [11, 42, 43]. PEA3 along with CBP/p300 was also shown to be a key transcriptional regulator of many genes (MMP-1, 2, 7 and 9, COX-2/ PTGS-2) implicated in inflammation and cell invasiveness [44-47].

The current study defines a novel link between various transcription factors occupying the COX-2/PTGS-2 promoter and cellular response to cigarette smoke exposure bringing a new component, ΔNp63α, showing a critical role for cooperation between various chromatin components in regulation of COX-2/ PTGS-2 expression and, therefore strengthening the central role of inflammatory process and oxidative stress in tumorigenesis of epithelial cells, especially after cigarette smoke exposure (both primary and secondhand).

Numerous signaling pathways are implicated in cellular response to oxidative and genotoxic stress potentially induced by smoke exposure. For example, the ROS-induced interplay between the DNA-damage and oxidative stress through an activation of ATM-dependent phosphorylation of p53 family members and TSG tumor suppressor via the PARP1-LKB1-AMPK-mTOR metabolic pathway known to stimulate autophagy [48-50].

Furthermore, as a substrate for ATM-dependent phosphorylation, ΔNp63α was shown serving as a pro-survival factor by upregulating a glutathione peroxidase (GPX2) to reduce the p53-dependent oxidative stress-induced apoptotic response [51]. p63 was also shown to transcriptionally regulate REDD1 that implicates ROS in the p53-dependent DNA damage response and in p63-mediated regulation of epithelial differentiation [52]. ROS were found to activate p53 family members (p53, p63, and p73) inducing the expression of ferredoxin reductase (FDXR), which in turn sensitizes cells to ROS-mediated apoptosis [53].

As a target for p53 family proteins, COX-2 was shown producing oxidative damage and affecting the stress-induced cellular senescence [54]. Hyperoxia was also shown to induce the cellular senescence through the p53-LKB1-AMPK pathway [55].

ROS were further shown to stimulate cancer cell growth by regulating AMPK-COX-2 pathway [56]. Sp1/Sp3-dependent transcriptional regulation of COX-2 was shown playing an essential role in the modulation of COX-2 expression that mediates neuronal homeostasis and survival by preventing DNA damage [57]. Furthermore, oxidative stress was shown to induce cGMPprotein kinase-mediated thioredoxin peroxidase 1 transcription through PEA3, AP-1, c-Myc and c-Jun transcription factors [58].

Accumulated data strongly suggest that continuous (chronic) upregulation of pro-inflammatory mediators (e.g., TNFalpha, IL-1beta, IL-6, COX-2, NOS-2) are induced during the aging process due to an age-related redox imbalance that activates many pro-inflammatory signaling pathways, including the NFκB signaling pathway [59]. Both ROS and pro-inflammatory genes (e.g., COX-2) were found contributing to the expansion of cellular inflammatory responses and reduce the expression of genes required to maintain synaptic structure and function ultimately leading to progressive dysfunction, apoptosis and/or necrosis and brain cell death [60]. Pro-inflammatory genes shown playing a role in neurodegeneration (e.g., Alzheimer disease) are transiently activated by the heterodimeric oxygen-sensitive protein-protein complexes between NFκB and HIF-1α (reviewed in ref. 62).

Many cellular responses to tobacco smoke, such as oxidative stress/DNA damage, EMT, altered adhesion-mediating signaling pathways and altered protein degradation, chromatin modifications/epigenetic changes, angiogenesis and autophagy/apoptosis complement the inflammatory/neoplastic processes as the key underlying mechanisms in both chronic obstructive pulmonary disease, cardiovascular disease, lung cancer, aging and age-related diseases [4–13, 24, 28, 49, 54–56, 59, 60]. Thus, understanding the cellular and molecular mechanisms underlying these processes will provide novel venues for devising therapeutic strategies against smoke-related diseases.

4. Materials and Methods

Preparation of CSE

Mainstream smoke extract (MSE) and sidestream smoke extract (SSE) made from research-grade cigarettes (2R4F, from Tobacco Health Research, University of Kentucky, Louisville) contain nicotine: 0.85 mg/cigarette and tar: 9.70 mg/cigarette as previously described [61]. SSE was collected from the burning end of the cigarettes without puffing at the rate of 200 ml/min and MSE was collected with 35 ml/min puff per 2 sec using the opposite end of two smoking machines (MasterFlex Pump Systems, Cole-Parmer Instrument). Briefly, the smoke of 20 cigarettes for MSE and 40 cigarettes for SSE was bubbled into each flask containing 20 ml of pre-warmed phosphate buffer saline. The aqueous smoke extract was filtered through 0.22 μm pore syringe filter to remove large particles. The smoke bubbled into MSE flask was acidic and that into SSE flask appeared to be basic, therefore the pH of each solution was adjusted to 7.4. The solution was aliquoted and kept frozen at -80°C until use. The concentration of SSE was monitored at the absorbance of 1 at A230 was considered 100%. The concentration of MSE solution was considered 100% [61]. MSE and SSE were used to imitate cigarette primary smoking and secondhand smoking, respectively.

Cell cultures and transfections

Human lung cancer cell lines (A427 and H1299) and normal human bronchiolar epithelial (NHBE) cells were purchased from the American Type Culture Collection (ATCC) were grown in the recommended media. The 200 pmol/six-well plate of scramble siRNA, siRNA against PEA3 (sc-36205) and COX-2 (sc-29279) were purchased from Santa Cruz Biotechnology and were transiently introduced into cells for 24 h using FuGENE-6 (Roche Molecular Biochemicals) as previously described [16]. Cells were then exposed to control medium or various concentrations of both MSE and SSE for indicated periods of time [61].

Chromatin immunoprecipitation (ChIP) assay

H1299 cells (2 x 106) were transfected exposed to control medium, MSE and SSE. ChIP was done using a ChIP assay kit (Upstate Cell Signaling Solutions). H1299 cells (2 x 106) were cross-linked with 1% formaldehyde and ChIP was performed as previously described [16, 35]. After reversing the cross-links, chromatin was purified using the QIAquick PCR Purification kit (Qiagen, Inc.) and samples were eluted with 30 μl of elution buffer. Chromatin was immunoprecipitated with 1 μg of primary antibodies against PEA3 (clone 1A2G3, sc-130661, Santa Cruz Biotechnology), NFκB p65 subunit (ab7970, Abcam), ΔNp63, C/EBP/β (#3087, Cell Signaling Technology) and NF-YA (ab6558, Abcam). Immunoprecipitates were used as templates for PCR amplification of the COX-2 promoter with the following primers: COX-2, sense primer, (-550) 5'-TCC CGA CGT GAC TTC CTC GA (-530) and antisense primer, (-150) 5'-GGA GAG GAG GGA AAA ATT TG-3' (-130) yielding the PCR product of 420 bp, which was amplified as follows: 94°C for 1 minute, 55°C for 1 minute and 72°C for 30 seconds using Platinum Taq I DNAa polymerase (Invitrogen) for 37 cycles [16].

Immunoblotting and immunoprecipitation

We used primary antibodies against PEA3 (clone 1A2G3, sc-130661) and ΔNp63(clone (3F306): sc-71825) from Santa Cruz Biotechnology; against LKB1 (clone 27D10, ab#3050) and C/EBP/β (#3087) from Cell Signaling Technology; NFκB p65 subunit (ab7970) and NF-YA (ab6558) from Abcam; against b-actin (Sigma), COX-2/PTGS-2 (Cayman Chemical) or ΔNp63 (Ab-1, EMD) for immunoblotting analysis and immunoprecipitation [16, 35]. Immunoblots were scanned and quantification was carried out by Image Quant software version 3.3 (Molecular Dynamics). Values were expressed as percentage of a control sample (defined as 100%). 5 x 105 cells per well were plated into six-well plates and exposed to control medium, MSE and SSE. Cells were lysed in buffer [50 mmol/l Tris (pH 7.5), 150 mmol/l NaCl, 5 μg/ml aprotinin, pepstatin, 1% Nonidet P-40, 1 mmol/l EDTA, 0.25% deoxycholate]. For immunoprecipitation, total cell lysates (500 μl) were precipitated with 5 μg of the indicated primary antibodies overnight at 4°C followed by incubation with anti-rabbit or anti-mouse immunoglobulins (IgG)-coupled to agarose beads (Sigma) and washed with lysis buffer before being resolved by SDS-PAGE [16, 35].

Cell invasion/Matrigel assay

Cells (1 x 104) in 0.5 ml of serum-free MEM were added to each well of 24-well/8-μm pore invasion membrane chambers coated with Matrigel (BD Discovery Labware). The lower chambers contained 10% fetal bovine serum (FBS) in MEM to serve as a chemoattractant. Cells were allowed to migrate or invade over the course of 48 h. Chambers were fixed with 100% methanol for 2 min, stained with 0.5% crystal violet for 2 min, rinsed in water and examined under a bright-field microscope. Values for invasion and migration were obtained by counting five fields per membrane (20x objective) and represented the average of three independent experiments done over multiple days [16, 61].

Colony formation assay

Soft agar colony formation assays were carried out in six-well dishes. Cells (1 x 104) suspended in 2 ml of 0.36% bactoagar (Becton Dickinson) with growth medium (RPMI 1640 supplemented with 10% FBS) were added on a base layer of 0.72% bactoagar containing culture medium [61]. The plates were incubated at 37°C in a 5% CO2 incubator for 3 weeks. Colonies were counted under low magnification (x100).

MTT survival assay

The number of cells in each well after treatment was estimated using the 3-(4, 5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (ATCC) as previously described [35, 61]. MTT labeling reagent (final concentration, 0.5 mg/ml) was added to cells in 96-well culture plates (final volume, 100 μl culture medium/well) and incubated for 4 h at 37°C in a humidified atmosphere of 10% CO2. Cells were then solubilized overnight and the samples were quantified at 570 nm using a microtiter plate reader (Bio-Rad Laboratories).

Note

Supplementary materials can be found at: Supplementary materials can be found at: http://www.landesbioscience.com/supplement/RatovitskiOMCL3-5-Sup.pdf.