Like tobacco smoke, marijuana smoke has been associated with numerous adverse pulmonary effects in humans including airway inflammation, chronic bronchitis, edema, mucus hypersecretion, and the impairment of large airway function and lung efficiency.Moreover, Aldington et al. showed that the impairment of large airway function and lung efficiency is 2.5–5 times greater in marijuana smokers than tobacco smokers.Like tobacco smoke, previous studies have also shown marijuana smoke to be genotoxic both in vitro and in vivo.In addition, it is suspected that marijuana smoke may be carcinogenic.Indeed, some agencies such as the California Environmental Protection Agency have placed marijuana smoke on their list of chemicals known to cause cancer.However, since there is a paucity of marijuana-only smoking populations to complete definitive studies, epidemiological studies conducted to date are limited in scope, and often confounded by concurrent tobacco smoking.Therefore, a clear and widely accepted empirical link between marijuana smoking and cancer does not exist. Information on the pharmacokinetics of marijuana smoke, and the mechanisms by which it may cause adverse effects, is also limited. Several mechanisms have been proposed including genotoxicity,alterations in endocrine function,alterations in cell signaling pathways,and immune suppression. However, many of these findings are based on the testing of individual cannabinoids found in marijuana smoke, as opposed to the whole smoke or smoke condensate. Genome-wide expression profiling may provide information to permit a better understanding of the toxicological pathways perturbed by exposure to marijuana smoke. Currently, there are no published studies that have used a whole genome toxicogenomics approach to evaluate responses to marijuana smoke.
However, Sarafian et al. employed a targeted stress response gene expression array to evaluate the effects of 9-tetrahydrocannabinol,plant benches the main psychoactive component of marijuana, on human small airway epithelial cells.They observed significant changes in genes related to xenobiotic metabolism,DNA damage response,inflammation and apoptosis.Microarray technology has been used more extensively to evaluate gene expression changes following exposure to tobacco smoke. For example, Sen et al. reviewed 28 studies examining transcriptional responses to complex mixtures including whole cigarette smoke and cigarette smoke condensate, and included in vivo and in vitro studies using human and rodent tissues.It was determined that the pathways most frequently affected by tobacco smoke were oxidative stress response, xenobiotic metabolism, inflammation/immune response, and matrix degradation. Other microarray studies have noted a DNA damage response leading to cell cycle arrest and apoptosis to be among the top pathways affected by tobacco smoke.The results of this study showed extensive overlap with the affected pathways highlighted in the review by Sen et al..Our study also showed that gene expression is remarkably similar across cigarette brands, and there is limited variation in the genotoxic potency of cigarette smoke condensates. In contrast to these findings, our earlier work revealed that tobacco and marijuana smoke condensates differ substantially in terms of their genotoxicity.More specifically, MSC were observed to be significantly more cytotoxic and mutagenic than matched tobacco smoke condensates.In addition, TSC appeared to induce chromosomal damage in a concentration-dependent manner, whereas matched marijuana condensates did not. The mechanisms underlying these differences in toxicity are unclear and warrant further investigation. As an extension of our previous work, the objective of the present study is to employ a toxicogenomics approach to compare and contrast the molecular pathways that are perturbed by MSC and TSC.
A murine pulmonary epithelial cell line was employed for in vitro exposures to both MSC and TSC. The results show that the pathways perturbed by MSC as compared to TSC are largely similar. However, subtle differences in gene expression provide insight into mechanisms underlying the observed differences in toxicities.Hierarchal clustering using all genes that were statistically significant revealed that the controls and the marijuana high concentration clustered independently from the rest of the samples. The remaining samples clustered first by concentration ,then by condensate type,with the last branching resulting from time.When cells exposed to TSC and MSC were analyzed separately, samples clustered first by concentration and then by time point, suggesting that concentration has the largest overall effect on gene expression. For MSC, the high concentration samples were on the first main branch, followed by control, low and medium concentrations. The results indicate that the expression profiles of the high concentration MSC exposed cells are quite distinct.For TSC, the controls branched separately from all the treatment groups.The top 10 genes with the largest overall fold changes are listed in Table 2. All of the top 10 genes were significantly up-regulated with the exception of low density lipoprotein receptor,which was down-regulated in MSC exposed cells. Of the top 10 changing genes, five genes were common to both MSC and TSC. The GO terms associated with these commongenes included multicellular organismal development, vasculogenesis, regulation of transcription, and regulation of inflammatory response. Ingenuity Pathway Analysis was used to define the pathways that were significantly altered following exposure to MSC or TSC. Fig. 3 shows the overlap in all the significant pathways between the two condensate types. The top five most significantly altered pathways for cells treated with MSC or TSC are listed in Table 3. NRF2-Mediated Oxidative Stress Response was the most significant pathway for cells exposed to TSC at all concentrations and time points, with the exception of lowest concentration attime 6 + 4 h where LXR/RXR Activation was the most significant.
For cells exposed to MSC, the most significantly altered pathways were Biosynthesis of Steroids, as well as NRF2-Mediated Oxidative Stress Response, Aminoacyl-tRNA Biosynthesis and HMGB1 Signaling.Some ofthe top five pathways were common to both the MSC and TSC including those related to oxidative stress and xenobiotic metabolism. However, inflammation pathways were more predominant for the MSC, whereas cell cycling and cancer signaling pathways were more predominant for the TSC. To further elucidate differences between the two smoke condensates, the genes that were uniquely expressed following TSC exposure or uniquely expressed following MSC exposure at the highest concentrations for the two separate time points were compared in IPA.The findings confirm the importance of inflammation and steroid biosynthesis pathways in MSC exposed cells and highlight the significance of apoptotic pathways particularly at the 6 h time point. For cells exposed to TSC,Mphase cell cycle pathways appear to be of particular importance. Gene Ontology in the Database for Visualization, Annotation and Integrated Discovery was used to apply functionalannotation to all the significantly differentially expressed genes for each condensate. The full results are shown in Supplementary Tables 1 and 2. For cells exposed to MSC, significant perturbations were associated with steroid/cholesterol/lipid biosynthesis, NODlike receptor signaling,tRNA aminoacylation, transcription regulation, unfolded protein response and DNA binding. Like MSC, cells exposed to TSC had significant perturbations in transcription regulation, unfolded protein response and DNA binding. In addition, perturbations in cell cycle, p53 signaling, oxidative stress, andcancer signaling were alsonoted in TSC exposed cells. Fig. 5 shows the overlap of all the significantly affected ontologies between the two condensate types. Functional annotation clustering in DAVID was used to minimize redundancy in the GO terms. This analysis revealed 19 clusters with enrichment scores greater than 2 for MSC and 19 clusters for TSC.The top clusters for MSC relevant to toxicological processes include lipid/steroid biosynthesis,RNA processing,cellular response to unfolded protein,tRNA aminoacylation,and positive regulation of transcription.The top clusters for TSC relevant to toxicological processes include cellular response to unfolded protein,cell cycle,positive regulation of transcription,response to steroid hormone stimulus,and positive/negative regulation of apoptosis and cell death.To investigate early versus downstream effects, functional annotation was applied to significantly differentially expressed genes at the two separate time points. The results are shown in Supplementary Tables 5–8.
For cells exposed to MSC at the 6 h time point, the analyses revealed 79 significant terms including those related to transcription activity, DNA binding, and steroid/cholesterol biosynthesis. Four KEGG pathways and 1 Biocarta pathway were also deemed significant at this time point. At the 6 + 4 h time point, 76 significant terms were identified. These terms included unfolded protein response, and tRNA aminoacylation, as well as steroid/cholesterol biosynthesis which was found at the 6 h time point. Three KEGG pathways were significant at this time point including Steroid Biosynthesis, Terpenoid Backbone Biosynthesis, and Aminoacyl-tRNA Biosynthesis. Analyses of cells exposed to TSC at the 6 hr time point revealed 67 significant terms including those associated with oxidative stress, cell death, protein unfolding, transcription regulation, DNA binding and cell cycle. In addition, 2 KEGG pathways were significant.At the 6 + 4 h time point, 32 GO terms were identified as significant with oxidative stress being the only relevant toxicological endpoint. In addition, only one KEGG pathway was significant. Overall for MSC, the DAVID analyses confirmed many of the significant pathways identified by IPA including steroid biosynthesis, tRNA aminoacylation, inflammation and apoptosis. In addition, the analyses highlighted transcription regulation, DNA binding and unfolded protein response as also significant. For TSC, the DAVID analyses confirmed the significance of IPA pathways related to oxidative stress and cell cycle. As with the MSC, the DAVID analyses also further highlighted the importance of transcription regulation, DNA binding and unfolded protein response, as well as cell death. Transcription regulation and DNA binding were significant terms common to both MSC and TSC at the 6 h time point, whereas no common terms existed for the two condensates at the 6 + 4 h time point.In our previous genotoxicity study we showed that MSC and TSC were both cytotoxic and genotoxic.However, quantitatively, MSC was more cytotoxic and mutagenic than TSC, and TSC appeared to induce chromosomal damage in a concentration-dependent manner whereas MSC did not. Our earlier chemical analyses of MSC and TSC noted that aside from the nicotine in tobacco and the cannabinoids in marijuana, rolling bench the two smoke condensates contained mixtures of chemicals that were qualitatively similar though quantitatively different.The similarities in the chemical profiles and some of the toxicity findings suggested that the two smoke condensates might elicit somewhat comparable gene expression profiles. Hierarchal clustering of all the MSC and TSC exposed samples in the present study supported this notion and samples clustered first by concentration as opposed to smoke type. In addition, analysis of the top ten greatest gene expression changes relative to control revealed that half of the genes were common to both marijuana and tobacco.
A number of previous studies have examined gene expression changes in pulmonary cells following exposure to tobacco smoke.Generally, these studies have shown that tobacco smoke stimulates xenobiotic metabolism, and that metabolized smoke constituents contribute to DNA damage. Following early insult, DNA damage leads to disruptions in the cell cycle such as arrest at the G2 checkpoint to allow time for response. Cellular response can include DNA repair, mutation induction through faulty repair or lack of repair, and programmed cell death of heavily damaged cells. Exposure to tobacco smoke can also trigger an inflammatory response and induce oxidative stress through increased levels of reactive oxygen species. Persistent induction of these processes following repeated exposure contributes to loss of normal growth control mechanisms, which is a key step in cancer development. Our study supports many of these findings, with exposure to TSC inducing the expression of genes involved in xenobiotic metabolism,oxidative stress,and DNA damage response as evidenced by changes in the expression of genes involved in cell cycle arrest, protein unfolding,transcription regulation, and inflammation.These same pathways were also significantly affected following MSC exposure, indicating that, as expected, MSC impacts many of the same molecular processes and functions as TSC. Although the effects of the condensates were largely similar, dose–response analysis indicates that the MSC is substantially more potent than TSC, with BMDs that in many instances are an order of magnitude lower than those for TSC. In addition, the results also highlighted some differences in steroid biosynthesis,apoptosis and inflammation, which were more significantly affected following MSC exposure, and cell cycle,which was more affected following TSC exposure.IPA canonical pathways related to the metabolism of xenobiotics were significantly affected in both TSC and MSC exposed cells at both time points. These pathways included Xenobiotic Metabolism Signaling, Metabolism of Xenobiotics by CYP450, and AHR Signaling. For both TSC and MSC, the number of genes that were significantly affected increased with increasing concentration and the greatest number of genes changing occurred at the 6 + 4 h time point.