Prostacyclin reverses the cigarette smoke-induced decrease in pulmonary Frizzled 9 expression through miR-31.

Half of lung cancers are diagnosed in former smokers, leading to a significant treatment burden in this population. Chemoprevention in former smokers using the prostacyclin analogue iloprost reduces endobronchial dysplasia, a premalignant lung lesion. Iloprost requires the presence of the WNT receptor Frizzled 9 (Fzd9) for inhibition of transformed growth in vitro. To investigate the relationship between iloprost, cigarette smoke, and Fzd9 expression, we used human samples, mouse models, and in vitro studies. Fzd9 expression was low in human lung tumors and in progressive dysplasias. In mouse models and in vitro studies, tobacco smoke carcinogens reduced expression of Fzd9 while prostacyclin maintained or increased expression. Expression of miR-31 repressed Fzd9 expression, which was abrogated by prostacyclin. We propose a model where cigarette smoke exposure increases miR-31 expression, which leads to decreased Fzd9 expression and prevents response to iloprost. When smoke is removed miR-31 is reduced, prostacyclin can increase Fzd9 expression, and progression of dysplasia is inhibited. Fzd9 and miR-31 are candidate biomarkers for precision application of iloprost and monitoring of treatment progress. As we continue to investigate the mechanisms of prostacyclin chemoprevention and identify biomarkers for its use, we will facilitate clinical trials and speed implementation of this valuable prevention approach.

The lung tumor promoter, butylated hydroxytoluene (BHT), causes chronic inflammation in promotion-sensitive BALB/cByJ mice but not in promotion-resistant CXB4 mice.

An inflammatory response accompanies the reversible pneumotoxicity caused by butylated hydroxytoluene (BHT) administration to mice. Lung tumor formation is promoted by BHT administration following an initiating agent in BALB/cByJ mice, but not in CXB4 mice. To assess the contribution of inflammation to this differential susceptibility, we quantitatively characterized inflammation after one 150 mg/kg body weight, followed by three weekly 200 mg/kg ip injections of BHT into male mice of both strains. This examination included inflammatory cell infiltrate and protein contents in bronchoalveolar lavage (BAL) fluid, cyclooxygenase (COX)-1 and COX-2 expression in lung extracts, and PGE(2) and PGI(2) production by isolated bronchiolar Clara cells. BAL macrophage and lymphocyte numbers increased in BALB mice (P<0.0007 and 0.02, respectively), as did BAL protein content (P<0.05), COX-1 and COX-2 expression (P<0.05 for each), and PGI(2) production (P<0.05); conversely, these indices were not perturbed by BHT in CXB4 mice. BALB mice fed aspirin (400 mg/kg of chow) for two weeks prior to BHT treatment had reduced inflammatory cell infiltration. Our results support a hypothesis that resistance to BHT-induced inflammation in CXB4 mice accounts, at least in part, for the lack of effect of BHT on lung tumor multiplicity in this strain.

Cyclin D1 as a proliferative marker regulating retinoblastoma phosphorylation in mouse lung epithelial cells.

Elevations in cyclin D1 content increase the phosphorylation status of retinoblastoma (Rb) protein to encourage cell cycle transit. We sought to determine if cyclin D1 content could be used as an index of cell proliferation in mouse lung epithelia following growth manipulations in vitro and in vivo. Rb protein concentration was high in 82-132 and LM2, two fast-growing neoplastic mouse lung epithelial cell lines. The hyperphosphorylated form of Rb predominated in these two cell lines, while Rb in slower-growing cell lines was predominantly hypophosphorylated. Consistent with this, more cyclin D1 protein was expressed in the fast-growing cell lines than in slower-growing cells. We therefore tested whether cyclin D1 content varied with growth status. The amount of cyclin D1 decreased upon serum removal coincident with growth inhibition and then increased upon serum re-addition which stimulated resumption of proliferation. This correlation between cyclin D1 content and growth status also occurred in vivo. Cyclin D1 content increased when lungs underwent compensatory hyperplasia following damage caused by butylated hydroxytoluene administration to mice and in lung tumor extracts as compared with extracts prepared from uninvolved tissue or control lungs. We conclude that elevated cyclin D1 levels account, at least in part, for the hyperphosphorylation of Rb in neoplastic lung cells, and are associated with enhanced lung growth in vitro and in vivo.

Regulation of lung epithelial cell morphology by cAMP-dependent protein kinase type I isozyme.

Cell shape is mediated in part by the actin cytoskeleton and the actin-binding protein vinculin. These proteins in turn are regulated by protein phosphorylation. We assessed the contribution of cAMP-dependent protein kinase A isozyme I (PKA I) to lung epithelial morphology using the E10/E9 sibling cell lines. PKA I concentration is high in flattened, nontumorigenic E10 cells but low in their round E9 transformants. PKA I activity was lowered in E10 cells by stable transfection with a dominant negative RIalpha mutant of the PKA I regulatory subunit and was raised in E9 cells by stable transfection with a wild-type Calpha catalytic subunit construct. Reciprocal changes in morphology ensued. E10 cells became rounder and grew in colonies, their actin microfilaments were disrupted, and vinculin localization at cell-cell junctions was diminished. The converse occurred in E9 cells on elevating their PKA I content. Demonstration that PKA I is responsible for the dichotomy in these cellular behaviors suggests that manipulating PKA I concentrations in lung cancer would provide useful adjuvant therapy.

Butylated hydroxytoluene (BHT) induction of pulmonary inflammation: a role in tumor promotion.

Chronic pulmonary inflammatory diseases predispose towards lung cancer by unknown mechanisms. Butylated hydroxytoluene (BHT) administration to mice causes lung injury and a subsequent inflammatory response, and when administered chronically to certain inbred strains following carcinogen treatment, increases lung tumor multiplicity. We hypothesize that inflammation promotes lung tumor growth in this model system and have begun to examine this hypothesis by assessing inflammatory parameters in inbred strains that vary in their susceptibility to promotion. Positive correlations were found between susceptibilities to tumor promotion and BHT induction of alveolar macrophage and lymphocyte infiltration into alveolar airspaces, and increased vascular permeability (P < .03, P < .04, and P < .005, respectively). The amounts of pulmonary cyclooxygenase (COX)-1 and COX-2 did not strongly correlate with promotion. Because persistent elevation of macrophage content is the hallmark of a chronic inflammatory response, the alveolar macrophage population was depleted by adding chlorine to the drinking water prior to carcinogenesis. This treatment reduced lung tumor multiplicity following 2-stage carcinogenesis (P < .05). These correlations between inflammatory and tumorigenic responses to BHT, along with decreased tumorigenesis after macrophage depletion, are consistent with a role of inflammation in promotion. Inflammatory mediators may provide targets for early diagnosis and chemoprevention.

Defective gap junctional intercellular communication in lung cancer: loss of an important mediator of tissue homeostasis and phenotypic regulation.

Gap junctions provide direct pathways for the exchange of molecules and ions between neighboring cells, a process known as gap junctional intercellular communication (GJIC). This GJIC is important for homeostasis and regulation of mitosis, differentiation, and apoptosis. Gap junctions are present in lung airway and alveolar epithelial cells and, in addition to the above roles, might coordinate ciliary beating and surfactant secretion. GJIC is decreased in human and mouse lung carcinoma cells because of reduced expression of the gap junction protein, connexin43 (Cx43), and defects in signal transduction pathways that mediate Cx43 function. This reduced GJIC is important in the behavior of lung carcinoma cells because forced expression of Cx43 in lung carcinoma cells inhibits their growth and tumorigenicity. In this report, we summarize our studies on the role of GJIC in lung neoplasia.

Studies using structural analogs and inbred strain differences to support a role for quinone methide metabolites of butylated hydroxytoluene (BHT) in mouse lung tumor promotion.

Chronic treatment of BALB and GRS mice with BHT (2,6-di-tert-butyl-4-methylphenol) following a single urethane injection increases lung tumor multiplicity, but this does not occur in CXB4 mice. Previous data suggest that promotion requires the conversion of BHT to a tert-butyl-hydroxylated metabolite (BHTOH) in lung and the subsequent oxidation of this species to an electrophilic quinone methide. To obtain additional evidence for the importance of quinone methide formation, structural analogs that form less reactive quinone methides were tested and found to lack promoting activity in BHT-responsive mice. The possibility that promotion-unresponsive strains are unable to form BHTOH was tested by substituting this compound for BHT in the promotion protocol using CXB4 mice. No promotion occurred, and in-vitro work demonstrated that CXB4 mice are, in fact, capable of producing BHTOH and its quinone methide, albeit in smaller quantities. Incubations with BALB lung microsomes and radiolabeled substrates confirmed that more covalent binding to protein occurs with BHTOH than with BHT and, in addition, BHTOH quinone methide is considerably more toxic to mouse lung epithelial cells than BHT quinone methide. These data are consistent with the hypothesis that a two-step oxidation process, i.e. hydroxylation and quinone methide formation, is required for the promotion of mouse lung tumors by BHT.

Growth inhibition in G(1) and altered expression of cyclin D1 and p27(kip-1 )after forced connexin expression in lung and liver carcinoma cells.

Gap junctional intercellular communication (GJIC) and connexin expression are frequently decreased in neoplasia and may contribute to defective growth control and loss of differentiated functions. GJIC, in E9 mouse lung carcinoma cells and WB-aB1 neoplastic rat liver epithelial cells, was elevated by forced expression of the gap junction proteins, connexin43 (Cx43) and connexin32 (Cx32), respectively. Transfection of Cx43 into E9 cells increased fluorescent dye-coupling in the transfected clones, E9-2 and E9-3, to levels comparable to the nontransformed sibling cell line, E10, from which E9 cells originated. Transduction of Cx32 into WB-aB1 cells also increased dye-coupling in the clone, WB-a/32-10, to a level that was comparable to the nontransformed sibling cell line, WB-F344. The cell cycle distribution was also affected as a result of forced connexin expression. The percentage of cells in G(1)-phase increased and the percentage in S-phase decreased in E9-2 and WB-a/32-10 cells as compared to E9 and WB-aB1 cells. Concomitantly, these cells exhibited changes in G(1)-phase cell cycle regulators. E9-2 and WB-a/32-10 cells expressed significantly less cyclin D1 and more p27(kip-1) protein than E9 and WB-aB1 cells. Other growth-related properties (expression of platelet-derived growth factor receptor-beta, epidermal growth factor receptor, protein kinase C-alpha, protein kinase A regulatory subunit-Ialpha, and production of nitric oxide in response to a cocktail of pro-inflammatory cytokines) were minimally altered or unaffected. Thus, enhancement of connexin expression and GJIC in neoplastic mouse lung and rat liver epithelial cells restored G(1) growth control. This was associated with decreased expression of cyclin D1 and increased expression of p27(kip-1), but not with changes in other growth-related functions.

Quantitative trait locus mapping of genes regulating pulmonary PKC activity and PKC-alpha content.

Strain A/J mice, which are predisposed to experimentally induced asthma and adenocarcinoma, have the lowest pulmonary protein kinase (PK) C activity and content among 22 inbred mouse strains. PKC in neonatal A/J mice is similar to that in other strains, so this difference reflects strain-dependent postnatal regulation. PKC activity is 60% higher in C57BL/6J (B6) than in A/J lungs, and the protein and mRNA concentrations of PKC-alpha, the major pulmonary PKC isozyme, are two- to threefold higher in B6 mice. These differences result from more than a single gene as assessed in F(1), F(2), and backcross progeny of B6 and A/J parents. Quantitative trait locus (QTL) analysis of 23 AxB and BxA recombinant inbred strains derived from B6 and A/J progenitors indicates a major locus regulating lung PKC-alpha content that maps near the Pkcalpha structural gene on chromosome 11 (D11MIT333; likelihood ratio statistic = 12.5) and a major locus controlling PKC activity that maps on chromosome 3 (D3MIT19; likelihood ratio statistic = 15.4). The chromosome 11 QTL responsible for low PKC-alpha content falls within QTLs for susceptibilities to lung tumorigenesis and ozone-induced toxicity.

Differential expression and localization of the mRNA binding proteins, AU-rich element mRNA binding protein (AUF1) and Hu antigen R (HuR), in neoplastic lung tissue.

Modulation of gene expression at the level of mRNA stability has emerged as an important regulatory paradigm. In this context, differential expression of numerous mRNAs in normal versus neoplastic tissues has been described. Altered expression of these genes, at least in part, has been demonstrated to be at the level of mRNA stability. Two ubiquitously expressed mRNA binding proteins have recently been implicated in the stabilization (Hu antigen R/HuR) or destabilization (AU-rich element mRNA binding protein [AUF1]/heterogeneous nuclear ribonucleoprotein D) of target mRNAs. Further, their functional activity appears to require cytoplasmic localization. In the present study, we demonstrate a strong correlation between increased cytoplasmic expression of both AUF1 and HuR with urethane-induced neoplasia and with butylated hydroxytoluene-induced compensatory hyperplasia in mouse lung tissue. In addition, when compared with slower growing cells, rapidly growing neoplastic lung epithelial cell lines expressed a consistently higher abundance of both AUF1 and HuR proteins. Moreover, in nontumorigenic cell lines, both AUF1 and HuR protein abundance decreased with confluence and growth arrest. In contrast, in spontaneous transformants, AUF1 and HuR abundance was unaffected by changes in cell density. We suggest that growth-regulated alterations in AUF1 and HuR abundance may have pleiotropic effects on the expression of a number of highly regulated mRNAs and that this significantly impacts the onset, maintenance, and progression of the neoplastic phenotype. Mol. Carcinog. 28:76-83, 2000.

High cyclooxygenase 1 (COX-1) and cyclooxygenase 2 (COX-2) contents in mouse lung tumors.

Mouse lung tumorigenesis is a convenient model for examining all stages of lung adenocarcinoma (AC) progression. Because enhanced cyclooxygenase 2 (COX-2) expression has been observed in advanced human AC, we investigated the intracellular concentrations of the two cyclooxygenases, cyclooxygenase 1 (COX-1) and COX-2, at different times after carcinogen administration to A/J mice. The concentrations of both proteins were much higher in urethane-induced adenomas and carcinomas compared with control A/J mouse lung tissue (P < 0.03 and P < 0.01 in adenomas and AC, respectively, for COX-1; P < 0.003 and P < 0.004 in adenomas and AC, respectively, for COX-2). Small benign tumors that arose spontaneously in 13-month-old mice also stained for COX-1 and COX-2, showing that this elevated enzyme content does not depend on chemical induction. COX-1 and COX-2 immunostaining was observed in normal bronchiolar and alveolar epithelia, alveolar macrophages and bronchiolar smooth muscle. This is the first report of the cellular distribution of COX-1 and COX-2 in murine lungs and the first in any species to demonstrate their co-localization. COX content in isolated bronchiolar Clara cells, a putative cell of tumor origin, was equal to that found in tumors, suggesting that the high enzyme content in neoplasms is due to their proportionally high concentration of these tumor precursor cells. Different patterns of COX-1 and COX-2 expression were observed in tumors of different growth patterns; only occasional small foci stained in solid adenomas, while most cells in papillary adenomas were immunoreactive. This staining pattern was also seen in adenocarcinomas, but some of the papillary portions also included focally stained and unstained regions. The continued expression during neoplastic progression of these specialized enzymes present in normal cells of tumor origin suggests their function in maintenance of the neoplastic state.

Selective induction of apoptosis in mouse and human lung epithelial cell lines by the tert-butyl hydroxylated metabolite of butylated hydroxytoluene: a proposed role in tumor promotion.

Butylated hydroxytoluene (BHT) causes lung injury in mice and promotes tumor formation. Hydroxylation of a tert-butyl group on BHT to yield the metabolite, 6-tert-butyl-2-[2'-(2'-hydroxymethyl)-propyl]-4-methylphenol (BHTOH), may be required. BHTOH is more potent than BHT on an equimolar basis in causing lung damage, enhancing lung tumor development, killing isolated bronchiolar non-ciliated Clara cells, and inhibiting lung epithelial gap junctional intercellular communication. One mechanism proposed for tumor promoting agents is selective cytotoxicity; killing normal cells allows uninhibited clonal expansion of neighboring initiated cells. We compared the abilities of BHT, BHTOH, and other BHT metabolites to kill non-tumorigenic and tumorigenic mouse and human lung cell lines, and examined the contribution of apoptosis to this cytotoxicity. These cells lack the cytochrome P450 2B isozyme necessary for converting BHT to BHTOH. BHTOH and 4-hydroperoxy-4-methyl-2,6-di-tert-butyl-2,5-cyclohex-adienone+ ++ (BHTOOH) were most toxic, BHT and 2,6-di-tert-butyl-1,4-benzoquinone (BHTQu) were less potent, and 4-methyl BHT metabolites that are not pneumotoxic were ineffective. BHTOH most strongly induced apoptosis, based on nuclear condensation and transmission electron microscopy. Non-tumorigenic cells were as susceptible to cell death as the neoplastic cell lines when apoptosis and necrosis are not distinguished, but more sensitive to BHTOH-induced apoptosis. An apoptotic mechanism may underlie the lung tumor promoting actions of BHTOH.

Inhibition of 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone-induced mouse lung tumor formation by FGN-1 (sulindac sulfone).

The sulfone derivative of the non-steroidal anti-inflammatory drug (NSAID), sulindac, has been reported to inhibit mammary and colon tumor formation in rodent models of chemically-induced carcinogenesis. Unlike its parent compound, this metabolite lacks cyclo-oxygenase inhibitory activity. A tumor induction protocol, consisting of NNK administration in the drinking water over several weeks to model chronic human exposure, was used to test whether the sulfone (called FGN-1) could inhibit the formation of primary lung tumors in mice. A total of 150 female, AIN76A-fed, A/J mice received 9 mg of NNK each. Concentrations of FGN-1 that had been previously determined not to affect body weight gain were added to the food at levels of 0, 250, 500 and 750 mg/kg of diet (30 mice/group) starting 2 weeks before NNK administration and continuing for 22 weeks. At that time pleural surface tumors were counted. Tumor incidence decreased significantly from 96 % in the control diet and 93% in the 250 FGN-1 mg/kg diet to 63 and 67% in the 500 and 750 mg FGN-1/kg diet groups, respectively (P < 0.001 by chi-square analysis). Lung tumor multiplicity decreased from 18.1+/-3 tumors/ mouse (mean+/-SEM, control diet) to 12.3+/-3 (250), 5.3+/-1 (500) and 2.1+/-1 (750) (P < 0.0005 by post hoc ANOVA). In previous studies using this carcinogenesis protocol, the maximum tolerated dose of sulindac inhibited lung tumor multiplicity by no more than 50% with no effect on incidence. This dose-dependent reduction in tumorigenesis by a non-toxic dose of FGN-1 indicates a strong chemopreventive activity against experimental induction of lung carcinogenesis. The greater potency of the sulfone over sulindac and its lack of toxic side effects because of its inability to affect cyclo-oxygenase activity suggests that clinical testing in individuals at high risk for lung cancer should be considered.

Mouse lung epithelial cell lines--tools for the study of differentiation and the neoplastic phenotype.

Several dozen lung epithelial cell lines have been established in culture over the past 20 years from normal lung explants and their spontaneous transformants, and from lung tumors that arose spontaneously or were induced with chemicals, viruses, or oncogenic transgenes. To provide information from which to choose appropriate lines for investigating problems in lung cell biology and pulmonary neoplasia, this review describes the origins of these lines and some of their characteristics. These include growth, morphology, tumorigenicity, ability to metastasize, xenobiotic metabolism, mutational status, signal transducing activities, cytogenetics, ability to form domes, and electric conductance. In addition to collecting this information in a single place for the first time, we describe previously unpublished apoptosis features of some of these lines. An increasing number of investigations are beginning to use these lines and this review contains references into 1997.

Cytoskeletal architecture in mouse lung epithelial cells is regulated by protein-kinase C-alpha and calpain II.

Brief exposure to 12-O-tetradecanoylphorbol 13-acetate (TPA) caused a uniformly flattened population of mouse lung epithelial cells to become more heterogeneous; some cells rounded up, and others detached to overlap with flatter cells. Actin stress fiber organization was disrupted, and F-actin accumulated in lemellipodia. Vinculin dissociated from the focal adhesion plaques to diffuse throughout the cytoplasm. Inhibition of protein kinase C (PKC) activity blocked these effects of TPA. After 8 h of TPA exposure, actin filaments reassembled and vinculin again localized to the cell periphery. Calpain inhibition attenuated the decrease of PKC-alpha protein and PKC activity from the membrane fraction, and prevented the redistribution of cytoskeletal elements. Talin immunostaining was widespread throughout control cells but was localized to the periphery 8 h after treatment with TPA or with inhibitors of PKC and calpain. Both vinculin and talin concentrations increased with prolonged TPA treatment. PKC-zeta and calpain II were not appreciably affected by TPA exposure. Translocation of PKC-alpha to the membrane, followed by its calpain-induced downmodulation, is apparently required for the reversible pattern of cytoskeletal changes caused by TPA.

Very early changes in pulmonary protein kinase C-alpha and calpain II contents following injection of butylated hydroxytoluene (BHT) into mice.

Butylated hydroxytoluene (BHT)-induced lung damage in mice is an excellent model system for studying mechanisms of chemically-induced, reversible alveolar injury. Changes in the pulmonary contents of protein kinase C (PKC) and the calcium-dependent protease, calpain, were previously noted during the repair phase following BHT-induced pneumotoxicity. Calpain is believed to initiate PKC down-regulation. PKC-alpha is the major PKC isozyme and calpain II the major calpain isozyme in mouse lung. We have now studied the time course of these enzymatic changes in detail. Pulmonary PKC-alpha concentrations decreased as early as 45 min after an i.p. injection of 200 mg/kg BHT. Calpain II levels rose within the first 40 min after BHT injection, and then declined below control levels. The rapidity of these changes implies a role of these enzymes in mediating the onset of injury. Lung damage and repair, as estimated by measuring the lung weight/body weight ratio, is maximal 6 days after administration of this dose of BHT. The extent of the decreased PKC-alpha and calpain II concentrations at this time was linearly related to the estimated degree of injury based on increased lung weight. This correlation suggests the value of monitoring these enzymes as putative early biomarkers of alveolar injury.