Cultured lung cancer cells were treated with cigarette smoke condensates (CSC) using a variety of concentrations and exposure durations. Under exposure conditions mimicking 1 pack per day (ppd) for five days, CSC mediated no appreciable effects regarding in-vitro proliferation rates, yet dramatically enhanced tumorigenicity of lung cancer cells in nude mice. Affymetrix array, qRT-PCR, immunoblot and ChIP experiments demonstrated that this phenomenon coincided with polycomb-mediated repression of Dickkopf-1 (Dkk-1), encoding a secreted antagonist of Wnt signaling. Focused PCR-array and immunoblot experiments demonstrated that CSC up-regulated Wnt5a and increased non-canonical Wnt signaling in lung cancer cells. Similar findings were noted following shRNA-mediated silencing of Dkk-1 in these cells. Knock-down of Dkk-1 recapitulated the dramatic pro-tumorigenic effects of CSC in lung cancer cells. Additional experiments were undertaken to specifically examine the effects of Wnt5a activation in lung cancer cells. Constitutive over-expression of Wnt5a increased proliferation, migration and invasion of Calu-6 and H841 cells expressing low endogenous levels of Wnt5a. Furthermore, over-expression of Wnt5a increased growth of Calu-6 xenografts in athymic nude mice. In contrast, constitutive over-expression of Wnt5a had no significant effects in A549 cells, which express relatively high endogenous levels of Wnt5a. shRNA-mediated inhibition of Wnt5a expression significantly decreased proliferation, migration and invasion of A549 cells. Subsequent micro-array and qRT-PCR experiments revealed that c Wnt5a mediated repression of synaptoponin 2 (also referred to as myopondin), which encodes a cytoskeletal protein, that inhibits growth, motility, and metastasis of human cancer cells. Experiments are in progress to fuler elucidate the mechanisms underlying this phenomenon. In additional experiments, we have develop a model with which to examine the effects of CSC on phenotype and gene expression profiles of lung cancer cells in-vivo without the complexity and expense of exposing mice to ambient smoke. Briefly, athymic nude mice were inoculated subcutaneously with A549 or Calu-6 lung cancer cells. When xenografts were approximately 2-4mm in diameter, mice were randomized to receive daily IP injections of CSC at varying doses, or saline. Twenty-eight day CSC exposure mediated dose-dependent increases in growth of lung cancer xenografts, and significantly increased intra-tumoral levels of adrenomedulin and cyp1A1, which are known to be regulated by AhR signaling;these alterations coincided with up-regulation of Wnt5a, and decreased Dkk-1 expression. IP administration of DZNep (2.5 mg/kg BID q4 days every 7 days) significantly attenuated CSC-mediated repression of Dkk-1, and abrogated enhancement of xenograft growth by CSC. Presently, the mechanisms establishing epigenetic signatures, as well as the sequence and clinical relevance of epigenomic alterations during initiation and progression of tobacco-induced pulmonary cancers have not been fully elucidated. In order to examine these issues, normal human small airway epithelial cells (SAEC) and cdk4/H-tert immortalized human bronchial epithelial cells (HBEC), as well as lung cancer cells from smokers and never-smokers have been cultured in normal media with or without cigarette smoke condensate (CSC) for up to 24 months under potentially relevant exposure conditions. Histone modifications, gene expression, DNA methylation, and micro-RNA profiles have been analyzed by immunoblot, quantitative RT-PCR, microarray, pyrosequencing, chromatin immunoprecipitation (ChIP), RNA cross link immunoprecipitation (CLIP), and methylated DNA immunoprecipitation (MeDIP) techniques. Under exposure conditions approximating 1 pack per day, CSC mediated progressive accumulation of cancer-associated histone marks in SAEC and HBEC. Global DNA hypomethylation was a relatively early event coinciding with a decrease in DMNT1/DMNT3b expression ratios, up-regulation of imprinted genes, and de-repression of cancer-testis genes. Stem cell polycomb target genes were significantly more likely to be hypermethylated than non-polycomb target genes in HBEC following CSC exposure. Progressive epigenetic alterations correlated with significantly increased soft agar clonogenicity of HBEC. More recent experiments have been performed to examine if epigenetic alterations induced in HBEC by CSC are the result of random stochastic events, or are more consistent with an "instructive" mechanism, a phenomenon alluded to, but not firmly established on the literature. Briefly, histone modifications, individual/genome wide gene expression, and DNA methylation profiles were analyzed by real time RT-PCR, Illumina array, pyrosequencing, ChIP techniques in SAEC and HBEC cultured in normal media (NM) with or without cigarette smoke condensate (CSC) for up to 16 months. Pyrosequencing analysis of multiple genes silenced by CSC treatment showed progressive de novo DNA hypermethylation. Illumina methylation array analysis revealed that hypermethylated- but not hypomethylated genes in CSC-treated HBEC highly overlapped with stem cell polycomb target genes (PCTG). Further analysis revealed that stem cell PCTGs were seven-fold more likely to be hypermethylated than non-PCTGs in HBEC following CSC exposure. ChIP analysis revealed that genes, which were hypermethylated and completely silenced by CSC treatment, had high levels of H3K27Me3 prior to CSC exposure. In contrast, genes that were resistant to de novo DNA methylation had very low or undetectable H3K27Me3, and high levels of the H3K4Me3 (activation mark) prior to treatment. Interestingly, progressive increases in site-specific DNA methylation coincided with gradual reductions of H3K27Me3 of several epigenetically-silenced genes. In additional studies, we have identified several micro-RNAs that are either up-regulated or silenced in normal respiratory epithelia and lung cancer cells by cigarette smoke. For example, we have observed that CSC up-regulates miR-31 in SAEC and lung cancer cells via recruitment of C/EBP-beta to LOC 554202, the host gene for miR-31. Subsequent experiments demonstrated significantly elevated miR-31 expression levels in primary lung cancers (particularly those from smokers) relative to adjacent histologically normal lung parenchyma. Further analysis revealed that miR-31 directly targets transcripts encoding several Wnt antagonists including Dkk-1, Dkk-3, SFRP1 and SFRP4, resulting in up-regulation of Wnt5a, a non-canonical ligand implicated in maintenance of cancer stem cells. Over-expression of miR-31 enhanced proliferation and tumorigenicity of lung cancer cells, confirming that miR-31 acts as an oncomir during pulmonary carcinogenesis. Additional experiments have revealed that CSC mediates epigenetic repression of miR-487b, targeting Wnt5a, SUZ12, BMI1, c-Myc and K-ras in SAEC and lung cancer cells. These targets have been implicated in establishment of stem-cell phenotype in lung cancer cells. Expression levels of miR-487b were significantly decreased in a large panel of resected lung cancers (especially those from smokers) relative to adjacent normal lung tissues. Furthermore, repression of mir-487b correlated significantly with increased expression of all five targets in primary lung cancers. Constitutive expression of miR-487b significantly decreased proliferation and tumorigenicity and in-vivo invasion/metastasis of lung cancer cells. Subsequent experiments demonstrated that epigenetic repression of mir-487b by CSC is mediated via TGF-beta 1. Collectively these data demonstrate that mir-487b is a novel tumor suppressor, which is silenced by epigenetic mechanisms during human pulmonary carcinogenesis.
|Xi, Sichuan; Xu, Hong; Shan, Jigui et al. (2013) Cigarette smoke mediates epigenetic repression of miR-487b during pulmonary carcinogenesis. J Clin Invest 123:1241-61|