Cytotoxicity and Gene Expression Profiles in Cell Cultures Exposed to Whole Smoke from Three Types of Cigarettes

Toxicological Sciences, Aug 2007

The purpose of this study was to evaluate and compare the cytotoxicity and gene expression profiles in cell cultures exposed to whole smoke generated from a full flavor cigarette (Test 1), a low tar cigarette (Test 2), and an ultra-low tar cigarette (Test 3). In addition, a reference cigarette 2R4F was evaluated for cytotoxicity. Neutral red (NR) cytotoxicity assay was performed to determine relative cell death at each exposure concentration (n = 6). LC50 was generated using wet total particular matter (WTPM), cigarette number, or nicotine concentrations. The overall order of cytotoxicity was Test 1 ≫ 2R4F ≈ Test 2 > Test 3. Cell culture samples were collected for RNA extraction at WTPM concentrations of each cigarette that gave similar nicotine concentrations. Affymetrix mouse whole genome 430 2.0 array was used to characterize the gene expression profiles for each cigarette. A total of 598 genes in Test 1, 176 genes in Test 2, and 234 genes in Test 3 samples were differentially expressed compared to the concurrent sham controls. The major biological processes associated with the changed genes in Test 1 samples were down-regulated DNA replication and cell proliferation; the same biological processes were much less affected in Test 2 and Test 3 samples. The common findings in all three cigarettes types were increased glutathione biosynthesis/consumption and inflammatory response, which are known biological effects caused by smoke exposure. The most significantly up-regulated genes were CYP1A1, GSTs, Hmox1, and Procr in smoke-exposed samples, which are either related to well-studied mechanisms of smoke exposure–related diseases or potential new biomarkers for assessing and monitoring biological effects of cigarette smoke exposure in vivo and in smokers. In summary, both the NR cytotoxicity assay and gene expression profiling were able to differentiate the three types of test cigarettes, and the results demonstrated reduced biological effects for the Test 2 and Test 3 cigarettes compared to the Test 1 cigarette in BALB/c-3T3 Cells.

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

http://toxsci.oxfordjournals.org/content/98/2/469.full.pdf

Cytotoxicity and Gene Expression Profiles in Cell Cultures Exposed to Whole Smoke from Three Types of Cigarettes

Binbin Lu 1 2 Laura Kerepesi 0 Lynne Wisse 2 Keith Hitchman 2 Quanxin Ryan Meng 2 0 Battelle Biotechnology , Columbus, Ohio 43201 1 Current address: Zhengzhou Tobacco Research Institute of CNTC , Zhengzhou , China 2 Batttelle Toxicology Northwest , 902 Battelle Boulevard, PO Box 999, Richland, Washington 99352 The purpose of this study was to evaluate and compare the cytotoxicity and gene expression profiles in cell cultures exposed to whole smoke generated from a full flavor cigarette (Test 1), a low tar cigarette (Test 2), and an ultra-low tar cigarette (Test 3). In addition, a reference cigarette 2R4F was evaluated for cytotoxicity. Neutral red (NR) cytotoxicity assay was performed to determine relative cell death at each exposure concentration (n 6). LC50 was generated using wet total particular matter (WTPM), cigarette number, or nicotine concentrations. The overall order of cytotoxicity was Test 1 2R4F Test 2 > Test 3. Cell culture samples were collected for RNA extraction at WTPM concentrations of each cigarette that gave similar nicotine concentrations. Affymetrix mouse whole genome 430 2.0 array was used to characterize the gene expression profiles for each cigarette. A total of 598 genes in Test 1, 176 genes in Test 2, and 234 genes in Test 3 samples were differentially expressed compared to the concurrent sham controls. The major biological processes associated with the changed genes in Test 1 samples were down-regulated DNA replication and cell proliferation; the same biological processes were much less affected in Test 2 and Test 3 samples. The common findings in all three cigarettes types were increased glutathione biosynthesis/consumption and inflammatory response, which are known biological effects caused by smoke exposure. The most significantly up-regulated genes were CYP1A1, GSTs, Hmox1, and Procr in smoke-exposed samples, which are either related to well-studied mechanisms of smoke exposure-related diseases or potential new biomarkers for assessing and monitoring biological effects of cigarette smoke exposure in vivo and in smokers. In summary, both the NR cytotoxicity assay and gene expression profiling were able to differentiate the three types of test cigarettes, and the results demonstrated reduced biological effects for the Test 2 and Test 3 cigarettes compared to the Test 1 cigarette in BALB/c-3T3 Cells. - There is overwhelming evidence that tobacco smoke plays a major role in the pathogenesis of lung cancer, chronic obstructive pulmonary disease (COPD), cardiovascular disease, and a variety of other human diseases (Bartecchi et al., 1995). The only known way to reduce tobacco userelated disease is complete cessation, but many smokers are unable to or unwilling to quit. As an alternative, tobacco companies have developed and marketed potential reduced exposure products (PREPs) aimed at reducing the harmful constituents delivered to human body (Bombick et al., 1998a,b; Borgerding et al., 1998; Patskan and Reininghaus, 2003; Stabbert et al., 2003; Terpstra et al., 2003; Tewes et al., 2003). Harm reduction is a feasible and justifiable approach based on the Institute of Medicine report (Stratton et al., 2001). However, careful evaluation of the health effects of these claimed PREPs are essential. As an initial step, the CORESTA congress In Vitro Task Force has chosen a battery of in vitro tests to assess the acute cytotoxicity and genotoxicity of tobacco products, which include the neutral red (NR) cytotoxicity assay, the Ames test, and the micronuclei test. In addition to the traditional toxicology methods, new research tools, such as transcriptomics, proteomics, and metabonomics, have been explored by tobacco industry and other research institute to evaluate and compare the biological effects of tobacco products. The major purpose of this study was to evaluate whether or not the NR cytotoxicity assay and transcriptomics profiling can differentiate cigarettes of different configurations. A second purpose was to identify potential biomarker genes that may be used to evaluate and monitor the health effects of tobacco products in experiential animals or humans. Three types of test cigarettes with different characteristics and a reference cigarette were tested. Cigarettes were ranked based on their cytotoxicity and changed biological processes associated with differentially expressed genes in each smoke-exposed group. Individual differentially expressed genes that might be related to mechanism of pathogenesis of smoke-caused diseases were also discussed. MATERIALS AND METHODS Cigarettes. Cigarettes used in this study were three commercially available cigarettes. The smoke analysis of these cigarettes are listed Table 1. Test 1 is TPM (mg/cigarette) Tar (mg/cigarette) Nicotine (mg/cigarette) CO (mg/cigarette) Water (mg/cigarette) Note. Cigarettes were smoked under standard Federal Trade commission condition. TPM, total particular matter; CO, carbon monoxide. a full flavor cigarette, Test 2 is a low tar cigarette, and Test 3 is an ultra-low tar cigarette. In addition, a reference cigarette 2R4F purchased from the Tobacco and Heath Research Institute (University of Kentucky, Lexington, KY) was evaluated for cytotoxicity. Cell culture. Mouse BALB/3T3 fibroblast cells (CCL-163) were purchased from American Type Cell Culture (ATCC, Manassas, VA). Cells were grown in Dulbecco Modification of Eagles Medium supplemented with 10% new born calf serum, 2% L-glutamine (200mM), 2% penicillin/streptomycin (5000 U/ml, Mediatech Inc., Herndon, VA), and 1% HEPES buffer (PH 7.3 0.1, 1M, Fisher Scientific, Rockville, MD) in a CO2 incubator (6% CO2, 37 C, and 90% humidity). The medium was changed every 23 days to maintain a constant cell growth. Cells were never allowed to grow to a complete confluence. Approximately 24 h before exposure, 3T3 cells were seeded into T-75 flasks at a concentration of 0.6 3 106 cells/flask. All reagents were ordered from Invitrogen Corporation (Carlsbad, CA) unless specified otherwise. Smoke exposure. Smoke was generated using a Borgwaldt-KC Condor SM85 cigarette smoking machine from which smoke was directed to the cell culture exposure chamber. Smoke concentration was determined gravimetrically from filter weights and sample collection flow rates. The exposure chamber was kept at a target temperature of 37 C. T-75 flasks with 3T3 cells attached to the bottom of the flask were placed on a rock platform. One inlet tube delivers fresh smoke into the flask, and the aged smoke was removed through an outlet into exhaust line. During exposure, the rock platform was rocked from side to side at a speed of 4 oscillations/min so that cells were repetitively exposed to smoke or submerged in the cell culture medium. The flow rate of smoke was 300 ml/min. There were six flasks for each smoke concentration plus concurrent control flasks that were exposed to high efficiency particular air filter filtered air. The duration of exposure was 1 h. Immediately after exposure, cell culture medium was replaced with fresh medium, and flasks were incubated for ~24 h before being processed for cytotoxicity test or for 5 h before collecting cells for RNA isolation. Cytotoxicity assay. Cytotoxicity of smoke exposure was evaluated by the NR assay. The procedure described by Putnam et al. (1999) was followed with some modifications. Briefly, ~24 h after exposure, cell culture medium was removed from flasks, and flasks were rinsed with prewarmed phosphatebuffered saline. Cells were treated with NR medium (25 lg NR/ml medium) for 3 h. NR medium was removed, and 20 ml of 1% formaldehyde fixative was added to each flask. The fixative was removed, and 10 ml NR desorption solution (10 ml glacial acetic acid, 500 ml 100% ethanol, and 490 ml distilled water) was added to each flask. The flasks were rocked several times to make sure all NR is dissolved into solution, then 200 ll solution from each flask was transferred to a 96-microtitre plate. The absorbance reading was collected at 540 nm using a Bio-Tek Synergy HT plate reader (Bio-Tek Services, Inc., Richmond, VA). Cell survival was expressed as the percentage of optical density values in treated samples versus concurrent sham controls. Statistical comparison of cytotoxicity of cigarettes. Dose-response curves, LC50, and 95% confidence intervals (CIs) of LC50 were generated using the Calcusyn software (Biosoft, Cambridge, UK) based on a general equation for dose-effect relationships derived by Chou (1980). LC50 values were considered significantly different if the 95% CIs of LC50 did not overlap. Additionally, a four-parameter logistic regression model was used to assess relative cell death between cigarette types. This model was used to account for the s shape response curves, which are typically seen in dose-response studies. A hypothesis test of equal slopes was performed to determine if there was evidence of a statistically significant difference between cigarette types after accounting for dose, i.e., wet total particular matter (WTPM) (lg/l), nicotine (mg/m3), and cigarettes (cigarette/m3), each with its own model. The statistical analysis was performed in R using the DRC library (Ritz and Streibig, 2005). Sample collection and RNA isolation. Cells were collected 5 h after exposure for RNA isolation. Cells were detached from flasks with 0.4% trypsin. Samples were then centrifuged to remove trypsin and culture medium. Cell pellet was loosen with a pipette in the residual medium (usually 50100 ll), and 600 ll RNA lysis tissue buffer (Qiagen, Valencia, CA) was added to the tube. Cells were homogenized by passing the lysate buffer 510 times through a 20-gauge needle. Samples were snap frozen in liquid nitrogen and store at 70 C until RNA isolation. Total RNA was extracted from ~13 million cells using RNeasy Mimi Kits purchased from Qiagen. RNA was further concentrated using YM-100 Microcon centrifuge filter devices (Millipore, Billerica, MA). The ratio of A260/280 was measured using a spectrophotometer to evaluate the purity of RNA. Sample quality was also evaluated using an Agilent Bioanalyzer 2100 (Agilent Technologies, Inc., Foster city, CA). Microarray data acquisition and preprocessing. Samples were prepared for hybridization to Affymetrix GeneChip microarrays using Affymetrix reagents and protocols. cDNA was synthesized from RNA using a one cycle cDNA synthesis kit. Biotinylated RNA was then synthesized from cDNA using an IVT labeling kit after incubating at 37 C for 16 h. The biotinylated RNA was fragmented and hybridized to an Affymetrix Mouse Genome 430 2.0 microarray for 16 h at 45 C in a hybridization oven rotating at 60 RPM. The microarray was washed and stained with streptavidin-phycoerythrin using an Affymetrix Fluidics Station 450. The array was then scanned using an Affymetrix GeneChip Scanner 3000. The chip data were analyzed using Affymetrix GeneChip Operating Software version 1.2. Microarray data analysis. The microarray data were imported into Genespring (version 7.2, Agilent Technologies, Inc., Redwood City, CA) for further analysis. The quality of each sample was first evaluated by visually inspecting the distribution of genes in graphs and by calculating sample similarity values as correlation coefficients. All measurements less than 0.01 were set to 0.01, and gene values on each chip were normalized to 50% percentile. Samples were filtered by flag (marginal or present in at least 20% of samples), followed by raw intensity value (>10 in at least 20% of samples). Genes that changed 1.5-fold or greater in the treated groups compared to the sham control group were identified by filtering with fold change. These genes were subject to one-way ANOVA parametric test using Benjamini and Hochberg False Discovery Rate for multiple testing correction. A p value of <0.05 was considered significant. Differentially expressed gene lists were imported into Metacore (GeneGo Inc., St Joseph, MI) for functional and mechanistic analysis. Biological processes associated with changed genes in each group were generated by querying against the MetaCore database. Cytotoxicity Assay The cytotoxicity of fresh smoke was evaluated by quantifying the ability of cells to incorporate NR into the lysosomes of the cells. The percent cell death was plotted against smoke concentration (WTPM, lg/L) in Figure 1. With the increase FIG. 1. Relative cell death rates in mouse BALB/3T3 fibroblast cells based on WTPM (lg/l). Cells were exposed to fresh cigarette smoke for 1 h by inhalation. Cell viability was determined by the NR cytotoxicity assay 24 h after exposure. Cell viability in the concurrent sham control samples was set at 100%. Symbol, average; bar, SD. of smoke concentration, cell death rate increased rapidly and reached 50% at ~50 lg/l for the Test 1 cigarette. Cell death increased much slower for the Test 2 and Test 3 cigarettes. The overall rank of cytotoxicity was Test 1 Test 2 > Test 3. The cytotoxicity of 2R4F reference cigarette was more like the Test 2 cigarette, especially at higher smoke concentrations (>60 lg/l), where the dose-response curves were better differentiated between cigarettes. In Figure 2, WTPM was converted to cigarettes/m3 air based on the number of cigarettes needed to generate the measured WTPM concentration by dividing WTPM (lg/l) with total particular matter (mg/cigarette). For example, the highest WTPM dose for Test 1 cigarette was 132.5 lg/l, which correlated to 7.7 cigarettes/m3 (132.5 lg/l/17.3 mg/cigarette). The separation of dose-response curves for three test cigarettes was greater using cigarette number as the parameter for exposure compared to using WTPM; however, the overall order of cytotoxicity for cigarettes were the same, i.e., Test 1 test 2 2R4F > Test 3. Nicotine was also used as the parameter of exposure to address the concern of smokers compensating for the amount of nicotine inhaled (Fig. 3). Similar cytotoxicity curve as in Figure 1 was observed for each cigarette, which was due to that the ratio of nicotine/WTPM for the three test cigarettes and the reference cigarette was similar (Table 1). Individual sample values at each exposure concentration were input into computer software to calculate the LC50 value for each cigarette. The LC50 values and the 95% CIs for each cigarette are listed in Table 2. The LC50 values for the three test cigarettes were significantly different as demonstrated by nonoverlapping 95% CIs. To facilitate comparison between cigarettes, the LC50 value for the Test 3 cigarette was set as 100%, and LC50 values for other cigarettes were plotted proportionally (Fig. 4). While the magnitude of differences in LC50 values varied among the three parameters used, the overall order of cytotoxicity was consistent. In addition, the dose-response curves of cytotoxicity of the test cigarettes were compared using a logistic regression model. The Test 1 cigarette was statistically different from the Test 2 or Test 3 cigarettes (p values ranging from <0.0001 to 0.03). However, the Test 2 cigarette was not significantly different from the Test 3 cigarettes using this logistic regression model ( p > 0.05), mainly due to the overlapping cell death rates at the low dose range. FIG. 2. Relative cell death rates in mouse BALB/3T3 fibroblast cells based on cigarette number (cigrette/m3). Cells were exposed to fresh cigarette smoke for 1 h by inhalation. Cell viability was determined by the NR cytotoxicity assay 24 h after exposure. Cell viability in the concurrent sham control samples was set at 100%. Symbol, average; bar, SD. FIG. 3. Relative cell death rates in mouse BALB/3T3 fibroblast cells based on nicotine concentration (mg/m3). Cells were exposed to fresh cigarette smoke for 1 h by inhalation. Cell viability was determined by the NR cytotoxicity assay 24 h after exposure. Cell viability in the concurrent sham control samples was set at 100%. Symbol, average; bar, SD. WTPM (lg/l) 49.2 (46.052.7) 78.0 (70.686.2) 95.2 (90.3100.3) 83.7 (79.787.8) Cigarette/m3 2.8 (2.63.0) 9.3 (8.410.3) 38.1 (36.140.1) 7.7 (7.38.1) Nicotine (lg/m3) 3.5 (3.33.8) 7.2 (6.57.9) 9.9 (9.410.4) 5.8 (5.56.1) Note. Values in parenthesis are the upper and lower range of 95% CIs. Microarray Analysis In addition to ranking the cigarettes by their cytotoxicity, a second purpose of the NR assay was to select the appropriate dose to collect cells for RNA isolation and microarray analysis. For microarray analysis, it is preferred to collect samples without causing significant cell death. In the current study, cells were collected at WTPM concentrations of 20.5, 23.8, and 28.5 lg/l WTPM (corresponding to 1.2, 2.8, and 11.4 cigarette/ m3 for the Test 1, Test 2, and Test 3 cigarettes, respectively, and giving similar concentration of nicotine in the smoke), which caused about 1020% cell death. The expression of 598 genes was changed in the Test 1 group (341 increased and 257 decreased at least 1.5-fold compared to sham control). Genes with known function that were up- or down-regulated more than threefold are listed in Table 3. The greatest increase was 8.6-fold (cytochrome P450 1A1), and the greatest decrease was 5.3-fold (nanos homolog 1). A total of 176 genes were differentially expressed in the Test 2 group (78 gene up-regulated and 98 down-regulated). Genes with known function that were up- or down-regulated more than threefold are listed in Table 4. Similar to the Test 1 group, gene FIG. 4. Relative LC50 values for test and reference cigarettes based on WTPM (lg/l), cigarette number (cigrette/m3), and nicotine concentration (mg/ m3). Absolute LC50 values were generated by the Calcusyn software. Relative LC50 values were generated by setting the value for the Test 3 cigarette at 100% to facilitate comparison between cigarette types. Averages and SDs were shown. TABLE 3 Genes Up- or Down-regulated More than 3-Fold Compared to the Sham Control Group in the Test 1 Cigarette Group Fold changeab Cytochrome P450, family 1, subfamily a Glutathione s-transferase, alpha 2 (Yc2) AMP deaminase 3 Glutathione s-transferase, alpha 4 Protein C receptor, endothelial Prostaglandin I receptor (IP) Phenylalanine-tRNA synthetase-like, alpha subunit Heme oxygenase (decycling) 1 Early growth response 3 Developmental pluripotency-associated 5 Serine (or cysteine) proteinase inhibitor, clade B, member 1a Tolloid like Cell division cycle 6 homolog (Saccharomyces cerevisiae) Ankyrin repeat domain 1 (cardiac muscle) Nanos homolog 1 (Drosophila) aMinus values indicate down-regulated genes. bp < 0.05, one-way ANOVA parametric test. with the greatest increase was cytochrome P450 1A1 (9.4fold). There were 234 genes differentially expressed in the Test 3 group (160 increased and 74 decreased). Genes that were upor down-regulated more than threefold are listed in Table 5. The greatest increase was 7.6-fold (cytochrome P450 1A1). There were significant overlaps between differentially expressed genes among three cigarette types as demonstrated by a Venn diagram (Fig. 5). Among the 176 differentially expressed genes in Test 2 group, 87 (49%) were also seen in the Test 1 group, while 166 genes (71%) of the Test 3 group gene list also appeared in the Test 1 group gene list. The top biological processes associated with differentially expressed genes in each test cigarette group were demonstrated in Figure 6. There were overlaps among these biological processes. The first major group of biological processes was related to cell cycle regulation and cell proliferation, including DNA replication, regulation of progression through cell cycle, TABLE 4 Genes Up- or Down-regulated More than 3-Fold Compared to the Sham Control Group in the Test 2 Cigarette Group Fold changea Cytochrome P450, family 1, subfamily a Glutathione s-transferase, alpha 2 (Yc2) Heme oxygenase (decycling) 1 Protein C receptor, endothelial Fold changea Cytochrome P450, family 1, subfamily a Glutathione s-transferase, alpha 2 (Yc2) Glutathione s-transferase, alpha 2 (Yc2) Phenylalanine-tRNA synthetase-like, alpha subunit Glutathione s-transferase, alpha 4 Heme oxygenase (decycling) 1 Colony-stimulating factor 3 (granulocyte) Protein C receptor, endothelial ap < 0.05, one-way ANOVA parametric test. positive regulation of cell proliferation, DNA replication initiation, DNA repair, cell proliferation, etc. These biological processes were overall down-regulated. A second biological process was inflammatory response, which was up-regulated. Another major biological process was related to glutathione biosynthesis, which was also up-regulated. A significant number of genes were changed only in the Test 1 cigarette samples (Fig. 5). The biological processes associated with these genes were characterized by querying the MetaCore database. The top biological processes are listed in Table 6. The dominant biological processes were related to cell cycle regulation and cell proliferation. Three of five samples in each cigarette group were used as a training set in the class prediction function in the Genespring. FIG. 5. Venn diagram of differentially expressed genes in cell cultures exposed to smoke generated from the Test 1, Test 2, or the Test 3 cigarettes. The rest of samples were treated as unknown samples. One hundred genes were selected as the predictor genes from the initial 746 genes that were differentially expressed in the smoke-exposed samples. The two unknown samples in each cigarette group were correctly assigned to the respective group based on the expression profiles of these 100 predictor genes in each sample (Table 7). Cigarette smoke condensate (CSC) and liquid extract of vapor phase from smoke are often used for the evaluation of in vitro toxicity of cigarettes due to the difficulty of exposing cells to whole smoke. In the current study, we used an in vitro whole smoke exposure system that can efficiently transfer the vapor phase to the cell exposure chamber, such as aldehydes (almost 100% except for formaldehyde, which had an transfer efficiency of 50% due to its high reactivity), NOx (~100%), CO, and CO2 (~100%). Little cytotoxicity can be attributed to formaldehyde in conventional cigarette. In comparison, acrolein, which is likely to be a major contributor to the cytotoxicity of gas phase, had a transfer efficiency of 100% (Tewes et al., 2003). The NR cytotoxicity assay has been used to evaluate the cytotoxic activity of cigarette whole smoke, the CSC, and the gas/vapor phase (Bombick et al., 1997; 1998a,b; Roemer et al., 2002). While all three fractions were cytotoxic and able to differentiate cigarette types, the vapor/gas phase was found to be the major contributor to the cytotoxicity of whole cigarette smoke (Meng et al., 2005). The major contributors to the in vitro cytotoxic effects of whole smoke, for example, acrolein, are also responsible for the respiratory tract irritation in experimental animals and humans and cause histopathological changes in the upper respiratory tract. Therefore, in vitro cytotoxicity screening represents an important initial step in the toxicological evaluation of tobacco products. In this study, the exposure concentration was based on WTPM measured from Cambridge filters. To demonstrate better the quantity of cigarettes consumed and to address the issue of smokers compensating for amount of nicotine inhaled, the exposure concentrations were converted to the number of cigarettes and quantity of nicotine per cubic meter of air. The NR cytotoxicity results were able to differentiate the three types of test cigarettes while the greatest separation of dose-response curves were achieved by using the number of cigarettes as the exposure dose parameter. The overall order of cytotoxicity was Test 1 Test 2 2R4F > Test 3 cigarettes, which was consistent with the fact that nanomaterials and/or special catalyzers were used in the filters of the Test 2 and Test 3 cigarettes to increase surface area and chemistry reaction speed, thus reduced toxic constituents, such as PAH, free radicals, in both particular matter and vapor phases. Cell cultures for microarray assay were exposed at WTPM concentrations that gave similar nicotine concentration in each test cigarette smoke, which caused only mild cytotoxicity (1020% cell death). While the time course of the level of transcription varies for individual gene after in vitro exposure, earlier studies showed that changes of gene expression profiles, especially genes related to DNA damage, oxidative stress, inflammatory response, cell cycle regulation, etc, were most Rank 1 2 3 Process Development Cell proliferation evident after 410 h exposure to CSC or aqueous extract of cigarette smoke (Bosio et al., 2002; Fields et al., 2005; Yoneda et al., 2003). Therefore, in the current study samples were collected 5 h after exposure for RNA isolation. The number of differentially expressed genes in the Test 1 cigarette samples (598 genes) was significantly greater than those in the Test 2 cigarette group (176 genes) or in the Test 3 cigarette group (234 genes), indicating a greater response in the Test 1 treatment group in terms of changes of gene expression profiles (Fig. 5). On the other hand, about half of the changed genes in the Test 2 samples and 71% of the changed genes in the Test 3 samples were also present in the Test 1 samples. These common genes are related to some basic biological functions frequently affected by tobacco smoke exposure. The top biological processes associated with the differentially expressed genes in each treatment group were characterized by querying the MetaCore database (Fig. 6). The dominant biological changes were down-regulated DNA replication and cell cycle regulation. The degree of impact varied among the three test cigarettes, with the Test 1 cigarette having the greatest down-regulation, the Test 2 cigarette having mild down-regulation, and the Test 3 cigarette having the minimal down-regulation of cell proliferation. This finding was consistent with that the biological processes associated with Treatment group prediction p value ratio Note. Three samples in each treatment group were used as the training sets in the class prediction function in the Genespring. The rest two samples were treated as unknown samples. One hundred genes were selected as the predictor genes from the initial 746 genes that were differentially expressed in the smoke-exposed samples. p value ratio is the probability of a sample belonging to the first best class compared to the probability of the sample belong to the second best class, i.e., a p value ratio of 0.0455 indicate the probability of this sample belonging to the first best class by chance is 22 (1/0.0455) times smaller than the probability of this sample belonging to the next best class. differentially expressed genes unique to the Test 1 samples (409 genes, Fig. 5) were mainly related to cell cycle regulation and cell proliferation (Table 6). Smoke-caused inhibition of fibroblast proliferation was seen in previous in vitro studies. Nakamura et al. (1995) showed that cigarette smoke inhibited lung fibroblast proliferation and chemotaxis. Acrolein and acetyaldehyde, two volatile components of smoke, were shown to induce a dose-dependent inhibition of lung fibroblasts attachment and proliferation (Nakamura et al., 1995). Lung fibroblasts from patients with emphysema were more susceptible to cigarette smoke extract induced inhibition of cell proliferation than those from nonemphysemateous lungs (Holz et al., 2004; Nobukuni et al., 2002; Noordhoek et al., 2003). Pulmonary emphysema is characterized anatomically by the destruction of lung parenchyma. It is hypothesized that cigarette smoke might contribute to the development of emphysema by inhibiting fibroblast proliferation and migration (Nakamura et al., 1995). The results of cell proliferation assays as mentioned above and gene expression profiling in the current study provided some evidence for this hypothesis. However, careful considerations are needed in assessing the mechanistic role of reduced fibroblast proliferation in emphysema development. First, it is important to determine how much of the observed inhibition of fibroblast proliferation was due to cytotoxicity of smoke exposure. In the study of Nobukuni et al. (2002), the cytotoxicity of cigarette smoke extract was determined by measuring the release of lactase dehydrogenase (LDH). LDH levels in the smoke extracttreated samples were not increased compared to sham controls, indicating the observed reduced number of fibroblasts was not due to cell death caused by cytotoxicity of cigarette smoke extract. In the current study, samples for microarray analysis were collected at smoke levels causing mild cytotoxicity (1020% cell death). The biological process of cell proliferation were clearly down-regulated while the anti-apoptotic processes were slightly up-regulated, confirming that even at slightly toxic concentrations of smoke, the reduced cell proliferation was due to reduced expression of genes associated with DNA replication and cell cycle regulation. However, the possibility that fibroblasts respond differently to very low level of smoke exposure cannot be excluded. Transcriptomics analysis of cells exposed to multiple levels of cigarette smoke, including concentrations that cause no cell death, should help to define the profiles of biological effects of smoke exposure on fibroblast. Second, we need to use caution in extrapolating the in vitro finding to complex in vivo situation. In humans, chronic smoke exposure generates an inflammatory microenvironment involving infiltrations of alveolar macrophages, neutrophils, and T lymphocytes, which produce cytokines that can regulate the proliferative capacity of lung fibroblast. Furthermore, epithelial cells in airway, which are frequently hyperplastic in smoke-exposed rodents, can also produce inflammatory cytokines and fibroblast growth factors. Very little data is available on the proliferative capacity of lung fibroblast in chronic smoke-exposed experimental animals or in smokers. The limited available data indicate that at dose levels encountered in tissues of smokers, smoke condensate cause suppressed fibroblast proliferation while enhance fibroblast survival by stimulating fibroblasts to express stress response/survival protein (Wong et al., 2004a,b). Nevertheless, current in vitro and in vivo data indicate that cigarette smoke intrinsically alters lung fibroblasts, such as causing defective regulation of cell proliferation and migration, reducing the capacity of synthesizing and depositing extracellular matrix, results in insufficient lung tissue repair, and might eventually contribute to the development of emphysema (Carnevali et al., 1998; Nakamura et al., 1995; Noordhoek et al., 2003; Wong et al., 2004a,b). Inflammatory response and glutathione synthesis were two other major biological processes associated with differentially expressed genes in smoke-exposed cells. Cigarette smoke causing inflammation is well documented in vitro, in vivo, and in smokers and is considered an important mechanism of smokecaused health effects, including COPD (Shields, 2002; Spurzem and Rennard, 2005). Depletion of glutathione is considered a critical step in the toxicological effect of smoke, and increased glutathione synthesis is frequently seen in smokeexposed cell or animals as a stress response to smoke-induced oxidative stress (Moriarty et al., 2003; Piperi et al., 2003; Ray et al., 2002; Reddy et al., 2002). Unlike the down-regulated cell proliferation process, the up-regulated inflammatory response and glutathione synthesis were affected to similar degrees in all three groups of smoke-exposed samples, indicating some common chemical constituents, which were not reduced by the special filters used in the Test 2 and Test 3 cigarettes, were responsible for effects on the inflammatory response and glutathione synthesis and consumption. CYP1a1 was the most up-regulated gene in all three groups of smoke-exposed samples. CYP1a1 can convert many chemical constituents in cigarette smoke to biologically reactive intermediates, including benzo[a]pyrene (Ding and Kaminsky, 2003). Elevated CYP1a1 activity correlated with the adverse effects of cigarette smoke in humans. Genotype of CYP1a1 affects cancer risk in humans (Hou et al., 2005; Miller and Fain, 2003; Taioli et al., 1998). Wardlaw et al. (1998) found that CYP1a1 was induced in the nasal respiratory and olfactory epithelia and lungs of smoke-exposed F344 rats. The pulmonary CYP1a1 activity was increased by 4- to 8-fold in rats exposed to one to three cigarettes (Iba et al., 2006). Expression of CYP1a1 in lung was approximately threefold higher in current smokers or ex-smokers than nonsmokers (Kim et al., 2004). The significant increase of CYP1a1 transcription in smoke-exposed cells in the current transcriptomics study provided further evidence for a role of CYP1a1 in cigarette smokerelated health effects in humans. GSTa2 was another significantly up-regulated gene in smoke-exposed groups. GSTa2 is a phase II detoxification enzyme, which protects cells or organism from reactive oxygen species and the products of peroxidation by conjugating glutathione with free radicals or xenobiotics. The increased expression of GST, along with increased glutathione synthesis as mentioned earlier, represents one of the most important detoxification mechanisms of smoke exposure. Protein receptor C (Procr) is known to be critical for the regulation of natural anticoagulant functions. The role of Procr in smokeinduced biological effects has not been studied. Recent studies show that Procr participates in anti-inflammatory and antiapoptotic activities (Esmon, 2006; Mosnier and Griffin, 2006), and the increased Procr expression observed in the current study could be an adaptive cell protection response to cigarette smoke. Further study is needed to characterize the role of Procr in the development of smoke exposurerelated diseases. Heme oxygenase-1 (Hmox1) is a cellular enzyme catalyzing the degradation of heme-containing molecules and is found ubiquitously in all organs. Hmox1 provides protection against oxidants and aromatic hydrocarbon in cigarette smoke and is an essential component for lung to keep a balance between oxidants and antioxidants (Choi and Alam, 1996; Kikuchi et al., 2005; Maines, 1997; Yamada et al., 2000). It has been shown that Hmox1 can be induced in vitro or in vivo after smoke exposure, and it has been suggested that Hmox1 plays a role in chronic airway disease, including emphysema (Fukano et al., 2006; Maestrelli et al., 2001, 2003). In an earlier in vitro gene profiling study, Hmox1 was substantially increased in Swiss 3T3 cells exposed to cigarette smoke extract as measured by a cDNA array, which is consistent with our results using a similar cell line (Bosio et al., 2002). While many common responses were observed in various cell lines or animals after smoke exposure, differences in genetic background of cell lines or animal models should be carefully considered in studying smoke exposurerelated biological effects. The study by March et al. (2005) showed that cigarette smokeinduced emphysema was more severe in A/J mice than in B6C3F1 mice. Eosinophilic inflammation and airway thickening were less marked in the BALB/c mice than in A/J mice after intranasal antigen challenge in a study by Shinagawa and Kojima (2003). Singh et al. (2005) showed that A/J strain was significantly more susceptible for perivascular inflammation followed by BALB/c and C57 mice, while the C3H/HeJ mice were the least susceptible strain. The cell line used in the current experiment is from BALB/c mice. Changed biological processes or differentially expressed genes in this study, such as the novel gene Proc, need to be further evaluated in other cell lines or animal models. In addition to characterizing altered biological processes and identifying potential biomarkers of smoke exposure, another purpose of the current study was to search for approaches to differentiate, categorize, and predict the biological effects of new cigarette products in experimental animals and humans. Using a class prediction function in the Genespring program, we were able to classify the unknown samples into correct cigarette groups. This is a prototype example with a very small size; however, the potential power of this approach is unlimited with the accumulating database for various categories of cigarettes. Thus, transcriptomics analysis is a promising tool as part of an in vitro test battery for tobacco industry to assess the potential toxicity of their new products prior to perform expensive in vivo studies. The three test cigarettes were ranked by the biological response in the NR assay and the transcriptomics analysis. However, it should be noted that the cytotoxicity results do not necessarily agree with the transcriptomics data. While the NR assay rank cigarettes solely relying on percent cell death caused by each cigarette, the differentiation power of transcriptomics analysis include not only the number of genes changed, but also the magnitude of change of each gene, the biological process, and pathways associated with these genes. In the current study, similar numbers of differentially expressed genes were observed in the Test 2 and the Test 3 cigarettes; however, we were able to differentiate the Test 2 and Test 3 cigarettes based on the 100 selected predictor genes. In summary, we have evaluated three commercially available test cigarettes along with the 2R4F reference cigarette, and the NR cytotoxicity assay was able to differentiate the three test cigarettes. Transcriptomics analysis showed that more genes were differentially expressed in the Test 1 smoke-exposed samples than the Test 2 or Test 3 smoke-exposed samples. The dominant biological process associated with changed genes in the Test 1 samples was decreased cell proliferation, which has been proposed as a potential mechanism of emphysema development in smokers; cell proliferation was much less affected in the Test 2 and Test 3 cigarette samples. Common biological processes associated with the changed genes in all three cigarette groups were increased glutathione synthesis and consumption and inflammatory response, which are known biological effects caused by smoke exposure. Inspection of individual genes in the differentially expressed gene lists revealed some well-studied mechanisms of smoke exposure related disease and potential new biomarkers for assessing and monitoring biological effects of cigarette smoke in vivo and in smokers. ACKNOWLEDGMENTS The authors thank Ms Terry Pavel for technical assistance in the study. Melissa Matzke performed regression analysis of the cytotoxicity data. Continuous support from Dr Bruce Westerberg, Mr Barry Hayden, Dr K Monica Lee and Dr Herb Bresler is gratefully acknowledged.


This is a preview of a remote PDF: http://toxsci.oxfordjournals.org/content/98/2/469.full.pdf

Binbin Lu, Laura Kerepesi, Lynne Wisse, Keith Hitchman, Quanxin Ryan Meng. Cytotoxicity and Gene Expression Profiles in Cell Cultures Exposed to Whole Smoke from Three Types of Cigarettes, Toxicological Sciences, 2007, 469-478, DOI: 10.1093/toxsci/kfm112