Epigenetic Alterations Induced by Tobacco Smoke

Division Of Basic Sciences - Nci

Investigators

Abstract

Cultured lung cancer cells were treated with tobacco smoke condensates (TSC) using a variety of concentrations and exposure durations. Under conditions mimicking 1 pack per day (ppd) exposures for five days, TSC mediated no appreciable changes regarding in-vitro proliferation rates, but instead dramatically enhanced tumorigenicity of A549 and Calu-6 cells in nude mice. Affimetrix long-oligo array and quantitative RT-PCR experiments revealed that TSC modulated expression of numerous genes in both cell lines. Of particular interest, TSC markedly inhibited expression of DKK1, a secreted Wnt signaling antagonist and putative tumor suppressor. ChIP experiments demonstrated that inhibition of DKK1 expression by TSC coincided with a dose-dependent increase in H3K27 trimethylation, and recruitment of polycomb repressor proteins EZH2, SUZ12 and Bim1 to the DKK1 promoter, with a concommitant decrease in total acetylated histone H4, but no appreciable change in methylated H3K4 (activation marks), or RNA pol II. Interestingly, methylation specific PCR (MSP) and pyrosequencing experiments revealed no appreciable increase in DNA methylation within the DKK1 promoter despite continuous TSC exposures ranging from 5-60 days. Consistent with these observations, cessation of TSC exposure resulted in diminution of DKK1 promoter-associated polycomb proteins, and restoration of DKK1 expression in these cancer cells. Western blot experiments indicated that TSC mediated dose-dependent increases in phospho-disheveled as well as phospho-JNK levels consistent with enhanced Wnt signaling in A549 and Calu-6 cells. Similar findings were noted following siRNA mediated knock-down of DKK1 in these cells. Knock-down of DKK1 dramatically increased tumorigenicity of Calu-6 lung cancer cells. The fact that TSC induced EZH2-mediated repression of DKK1, which is known to be down-regulated by Oct 4, suggested that tobacco smoke might enhance tumorigenicity by increasing stemness of lung cancer cells. To examine this issue, Calu-6 cells were exposed to normal media or 0.001 or 0.003 puffs per ml TSC for 5 days. Flow cytometry techniques were used to identify the side population (SP) fraction (side population of Hoechst dye uptake), which has been shown to contain cancer cells exhibiting stem cell characteristics. The percentage of cells within the SP of untreated, poorly tumorigenic Calu-6 cells was quite low. Interestingly, TSC exposure increased the SP of these cells by nearly 70 percent. Consistent with these observations, TSC mediated a 2 fold increase in Oct 4, NANOG, and CD44 expression in unsorted Calu-6 cells. Additional quantitative RT-PCR experiments demonstrated that sequential DAC/DP mediated a 2-4 fold increase in Oct 4 and NANOG in these cancer cells (Table 4). Interestingly, sequential DAC/DP exposure appeared to abolish the SP in Calu-6 cells. These preliminary results raise the possibility that tobacco smoke enhances the malignant phenotype of lung cancer cells by selecting for outgrowth of cancer stem cells (CSC). In contrast, chromatin remodeling agents may decrease the stem cell component despite induction of Oct 4, NANOG, BORIS, and CT-X genes, possibly by enhancing expression of p16, which inhibits replicative potential of stem cells, or cdx2, which is known to mediate trophectoderm differentiation. Analysis of epigenetic mechanisms in cancer cells provides limited information regarding chromatin remodeling events associated with initiation of malignant transformation. Surprisingly little information is available concerning epigenetic responses in respiratory epithelia following exposure to tobacco smoke. As such, a series of experiments were initiated to ascertain if tobacco smoke could induce epigenetic alterations in normal bronchial epithelia that were similar to those observed in lung cancer cells. Under relevant exposure conditions, TSC mediated dose-dependent growth inhibitory effects in SAEC and HBEC, and to a lesser extent cultured lung cancer cells. These effects coincided with an oxidative stress response, increased p21 and cyclin D1 expression, and depletion of wt p53 in SAEC and HBECs. Long term TSC treatment (1- more than 6 months) markedly diminished H4K16Ac and H4K20Me3 levels, while increasing relative levels of H3K27Me3 in HBECs; these histone alterations coincided with dose-dependent depletion of DNMT1, a modest increase in DNMT3b, complex DNA methylation and gene expression profiles, hypomethylation of D4Z4, NLB2, and LINE 1 elements, induction of H19, MAGE-A1 and NY-ESO-1, and activation of Wnt signaling. Prolonged TSC exposure markedly increased clonogenicity and tumorigenicity of HBEC. Collectively, these data indicate that TSC activates Wnt signaling, and mediates cancer-associated epigenetic alterations in cultured respiratory epithelia. Further analysis of this novel model system may provide considerable insight regarding novel epigenetic mechanisms associated with tobacco induced pulmonary carcinogenesis. Additional experiments were undertaken to examine micro-RNA expression profiles in untreated and TSC-exposed HBEC and SAEC relative to cancer lines or primary tumor specimens from lung cancer patient. Four sets of cell lines were chosen for this analysis including short term NHBE and SAEC, immortalized HBECs, lung cancer lines from non-smokers (H1650 and H1975), as well as lung cancer lines from smokers (H1299, H358, Calu-6, and A549). Briefly, small RNAs were isolated from untreated NHBE, SAEC, HBEC, A549, and Calu-6 cells treated with TSC for 72h, and A549 and Calu-6 cells exposed to DAC/DP using our standard treatment regimen. miRNA expression profiles were evaluated using NCode Multi-Species miRNA Microarray V2 chips containing probes that target all of the miRNA species in the Sanger mirBase 9.0 for human, mouse, rat, Drosophila, C. elegans, and zebrafish. Microarray data were processed and normalized, with statistic and clustering analyses performed using Genespring software. Two-way analysis of variance (ANOVA) techniques were used to identify statistically different miRNA expression levels in untreated and TSC-exposed cells. Similar techniques were used to compare miRNA expression in primary respiratory epithelia, HBEC, smoker-, and nonsmoker- derived lung cancer lines for cancer progression modeling. This analysis yeilded several interesting results including: 1. miRNA signature can classify human lung cancer progression. Based on the global miRNA expression clustering, primary respiratory epithelia, immortalized bronchial epithelial cells, and smoker/ non-smoker derived cancer lines could be readily distinguished from one another. Each group exhibited its own set of miRNAs that were over- or under-expressed relative to the primary cells. 2. tobacco smoke modulates miRNA expression in normal respiratory epithelia and lung cancer cells. Comparison of the miRNA expression profiles between TSC-treated and untreated SAEC, NHBE, HBEC, A549 and Calu-6 cells, revealed that tobacco smoke exposure significantly altered expression of mir-21, mir-31 and mir-487b. Of particular interest, TSC induced expression of mir-21, which previously has been shown to be over-expressed in human lung cancers(97).Target prediction algorithms identified 192 putative targets for mir-31, including L [summary truncated at 7800 characters]