Effect of Inhaled Budesonide in Smokers with Bronchial Dysplasia

1. 2000 1800 Current Smoker 1600 (x104) Former Smoker 1400 1200 1000 Tags Per Million 800 600 400 200 0 CYP/(GST+GPX) CYP1B1 CYP1A1 GST2 GPX2 Effect of Inhaled Budesonide in Smokers with Bronchial Dysplasia BC Cancer Research Centre 675 West 10th Avenue Vancouver BC, V5Z 1L3 ph:604-675-8000 ext.7703 email: rchari@bccrc.ca Raj Chari1,2, Kim Lonergan1, Luc Girard4, John Minna4, Adi Gazdar4, Ruisheng Yao3, Ming You3, Raymond Ng2, Wan Lam1,2, Calum MacAulay1,2, Stephen Lam1,2 [1]British Columbia Cancer Agency, Vancouver, BC, Canada, [2]University of British Columbia, Vancouver, BC, Canada, [3]Washington University in St. Louis, St. Louis, MO, USA, [4]UT Southwestern Medical Centre4, Dallas, TX, USA Abstract Methods Results Comparison of Affymetrix U133A and SAGE Effect of Active Smoking Effect of Budesonide Background: Budesonide, an inhaled corticosteroid used in the treatment of asthma, has been shown to be an effective chemopreventive agent in an animal model of adenocarcinoma [Carcinogenesis 1997 Oct 18(10):2015-7]. In humans, although inhaled budesonide for 6 months was not effective in regression of bronchial dysplasia, a significantly higher rate of resolution of CT detected small lung nodules was observed, some of which may represent pre-neoplastic lesions in the peripheral lung [Clinical Cancer Research 2004 Oct 1 10(19): 6502-11]. Despite being used as a drug for treatment of asthma and COPD for over two decades, the in-vivo effects of inhaled steroids on gene expression profiles of bronchial epithelial cells in smokers are still poorly understood. Objective: The objective of this study is to characterize the effects of inhaled budesonide on the gene expression profiles of bronchial cells from current and former smokers with bronchial dysplasia. Methods: Bronchial cells were obtained before and after six months of treatment with budesonide 800 mcg BID by inhalation. After two rounds of linear amplification of the extracted RNA, the gene expression profiles were analyzed using the Affymetrix U133A microarray chip. Results: Using Principal Component Analysis, the effect of active smoking was found to be stronger than the effect of budesonide. In current smokers, more Phase 1 enzyme genes were up-regulated compared with former smokers. However, Phase 2 enzyme genes were up-regulated in former smokers but down-regulated in current smokers. Specifically, CYP1B1 was shown to have a two-fold increase in current smokers after Budesonide treatment and approximately a 1.5-fold decrease in former smokers after treatment. In a separate analysis, it was found that genes up-regulated in current smokers had the tendency to be down-regulated in former smokers after treatment. As well, potential genes have been identified which correlate with the response to Budesonide treatment. Conclusions: A differential effect of Budesonide on gene expression profile was found between current and former smokers. Furthermore, we have characterized potential genes which correlate with the level of response. Targeting genes and pathways that correlate with patient response may be beneficial in screening of new chemopreventive agents. Supported by NIH contract N01-CN85188, & NCI SPORE CA 70907 & NIH 1P01 CA096964 Genes Correlating with Clinical Response Higher in Current Smokers Higher in Former Smokers Description of Dataset Specimens Human bronchial epithelial cells were obtained in 10 smokers with bronchial dysplasia by bronchial brushings before and after daily treatment with 1600 µg of inhaled budesonide/placebo for six months (Clin Cancer Res 2004;10:6502-6511). Specimens were also obtained from 7 subjects before and after one month treatment with budesonide • Affymetrix U133A data was obtained from the paper by Spira et al. [Proc Natl Acad Sci U S A. 2004;101(27):10143-8] • SAGE libraries were generated from human bronchial epithelial cells obtained by bronchial brushings during bronchoscopy of heavy smokers • Comparison of the number of probes or tags that are differentially expressed based on t-value, permutation test score, fold-change and all 3 criteria at significance level of p ≤ 0.001 • Using the 3 criteria • In current smokers, SAGE data shows 30 tags (mapping to 27 unique genes) and Affymetrix U133A gives 7 probes (mapping 5 unique genes) that are higher compared to former smokers • In former smokers, SAGE data suggests 22 tags (mapping to 18 unique genes) and Affymetrix U133A has none that are higher relative current smokers Platform • Example of genes with increased (left) and decreased expression (right) in complete responders. Among the up-regulated genes are: caveolin 2 (CAV2), hydroxyprostaglandin dehyrogenase 15-(NAD) (HPGD), TU3A protein (TU3A) and the down-regulated genes are: toll-like receptor 3 (TLR3), TIR domain containing adaptor inducing interferon-beta (TRIF), mitogen-activated protein kinase kinase 6 (MAP2K6). • HPGD has been recently shown to mediate anti-inflammatory effects in lung adenocarcinoma cell lines [Arch Biochem Biophys. 2005;435(1):50-5] • TRIF and TLR3 potentially implicate down-regulation of the Toll-like receptor pathway • Affymetrix U133A array was used to profile the samples obtained from the 10 subjects in the six month study • Affymetrix U133A 2.0 array was used for the 7 subjects in the one month study Serial Analysis of Gene Expression (SAGE) SAGE deduces expression profile based on the relative abundance of sequence tags. Evaluation Criteria Background Examples of Differentially Expressed Genes - Phase 1& 2 Toxin Exposure Genes • Identify genes that show similar expression differences before and after treatment in 4 of 5 subjects • Fold-change of ≥ 1.5 (log2 fold of 0.585) between baseline and six months • Remove genes if similar expression pattern seen in the placebo arm • Findings validated in the dataset of the one-month study • Chemopreventive agents for lung cancer may exhibit a differential effect in current versus former or never smokers • Previous microarray studies showed different gene expression profiles between current and former smokers • Budesonide is one of the most effective chemopreventive agents in animal models of lung cancer but was found to be ineffective in regression of bronchial dysplasia in smokers [Clinical Cancer Research 2004 Oct 1 10(19): 6502-11] Subset of Genes Differentially Expressed Between Current and Former Smokers • SAGE is not limited to genes spotted on the array • SAGE produces quantitative, discrete (digital) data values • ~150,000 tags enumerated per library Correlation of Gene Expression with Clinical Response • Phase I enzymes and cell cycle control genes was observed in current smokers compared to former smokers Objectives Filtering of data • Response characterized as complete response (CR), partial response (PR) and progressive disease (PD) • CR is defined as complete regression of all dysplastic lesions to hyperplasia or normal, PR is regression of some but not all dysplastic lesions; PD is appearance of new lesions or progression of dysplastic lesions by ≥ 2 grades after treatment • Identify genes that have a ≥ 1.5 fold change in complete responders only • Must also exhibit a gradient of expression between CR, PR and PD. Budesonide Associated Expression Changes • To determine the effect of active tobacco smoking on gene expression profiles in human bronchial epithelial cells retrieved by bronchial brushing during bronchoscopy • To compare the gene expression profiles obtained by Affymetrix U133A array with Serial Analysis of Gene Expression (SAGE) • To characterize the effects of inhaled budesonide on the gene expression profiles of bronchial cells from current versus former smokers with bronchial dysplasia • To identify genes associated with regression of bronchial dysplasia after treatment with inhaled budesonide • Affymetrix data filtered as explained by Spira et al. [Proc Natl Acad Sci U S A. 2004;101(27):10143-8] • Data for SAGE tags was retained only if raw tag counts were greater than 1 in at least one library Subject Background and Response Conclusions • Analysis of SAGE data affords more discovery of significant genes as compared to the Affymetrix U133A platform • Budesonide exhibits anti-inflammatory effects as evidenced by increased HPGD, furthermore this gene is increased even higher in complete responders. • Expression changes of some of the genes appeared as soon as one month after budesonide exposure, while others do not appear until six months • Inhaled budesonide has a differential effect in current versus smokers. Such differences such as effects on Phase I & II enzymes gene expression may account for differences in response between current and former smokers Statistical Methods • A combination of the student’s t-test, fold-change and permutation test were used in the analysis • Differentially expressed genes were selected at a significance level of p ≤ 0.001 • T-test and fold-change are commonly used in the literature • Permutation test • suitable test for a small sample size • non-parametric test • performs well in identifying differentially expressed genes Effect of Budesonide Independent of Response and Smoking Status Hypotheses • (Below) Sample of genes up-regulated and down-regulated after treatment in at least 4 of 5 subjects • Also exhibit same pattern in at least 5 of 7 subjects in validation dataset (U133A 2.0) Smoking Status and Response to Inhaled Budesonide • SAGE identifies differentially expressed genes not captured by previous microarray analysis • Current and former smokers respond differently to inhaled budesonide • Regression of bronchial dysplasia after budesonide is associated with different gene expression profiles versus non-responders Acknowledgements • Identify genes which have ≥ 1.5 fold change exclusively in all former smokers but not in all current smokers and vice-versa • Remove genes if similar expression pattern seen in placebo arm Supported by NIH contract N01-CN85188, & NCI SPORE CA 70907 & NIH 1P01 CA096964 & Genome Canada/Genome BC