Gene expression changes in human small airway epithelial cells exposed to Delta9-tetrahydrocannabinol

Toxicology Letters 158 (2005) 95–107

Gene expression changes in human small airway epithelialcells exposed to�9-tetrahydrocannabinol

Theodore Sarafiana,∗, Nancy Habiba, Jenny T. Maoa, I-Hsien Tsua,Mitsuko L. Yamamotob, Erin Hsub, Donald P. Tashkina, Michael D. Rotha

a Division of Pulmonary& Critical Care, Department of Medicine, 14-184 Warren Hall,David Geffen School of Medicine at UCLA, Los Angeles, CA 90095-1690, USA

b Molecular Toxicology Interdepartmental Program, David Geffen School of Medicine at UCLA, Los Angeles, CA 90095-1690, USA

Received 11 January 2005; received in revised form 1 March 2005; accepted 2 March 2005Available online 26 April 2005

Abstract

Marijuana smoking is associated with inflammation, cellular atypia, and molecular dysregulation of the tracheobronchialepithelium. While marijuana smoke shares many components in common with tobacco, it also contains a high concentrationof �9-tetrahydrocannabinol (THC). The potential contribution of THC to airway injury was assessed by exposing primarycultures of human small airway epithelial (SAE) cells to THC (0.1–10.0�g/ml) for either 1 day or 7 days. THC induceda time- and concentration-dependent decrease in cell viability, ATP level, and mitochondrial membrane potential. Using atargeted gene expression array, we observed acute changes (24 h) in the expression of mRNA for caspase-8, catalase, Bax,e 1). After7 Bax waso ggest ac xicants,2 cedp red as ac ssociatedw©

K

fedily

0

arly growth response-1, cytochrome P4501A1 (CYP1A1), metallothionein 1A, PLAB, and heat shock factor 1 (HSFdays of exposure, decrease in expression of mRNA for heat shock proteins (HSPs) and the pro-apoptotic protein

bserved, while expression of GADD45A, IL-1A, CYP1A1, and PTGS-2 increased significantly. These findings suontribution of THC to DNA damage, inflammation, and alterations in apoptosis. Treatment with selected prototypical to,3,7,8-tetrachlorodibenznzo-p-dioxin (TCDD) and carbonyl cyanide-p-(trifluoramethoxy)-phenyl hydrazone (FCCP), produartially overlapping gene expression profiles suggesting some similarity in mechanism of action with THC. THC, deliveomponent of marijuana smoke, may induce a profile of gene expression that contributes to the pulmonary pathology aith marijuana use.2005 Elsevier Ireland Ltd. All rights reserved.

eywords:Tetrahydrocannabinol; Airway epithelial cell; cDNA array; Gene expression; Marijuana

∗ Corresponding author. Tel.: +1 310 794 1979;ax: +1 310 267 2020.E-mail address:[email protected] (T. Sarafian).

1. Introduction

Marijuana is one of the most frequently smokdrugs in the United States with over 6 million da

378-4274/$ – see front matter © 2005 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.toxlet.2005.03.008

96 T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107

users and up to 35% of high school seniors reportinguse (Adams and Martin, 1996; Johnston et al., 2003).Despite this high prevalence of use, there is relativelylittle information regarding the toxicological effects ofmarijuana smoke on lung tissue. Marijuana smoke con-tains many of the same chemical components found intobacco smoke, including all of the major classes oftoxic and mutagenic species (Novotny et al., 1982).The major distinction is the presence of cannabinoids,primarily �9-tetrahydrocannabinol (THC), in mari-juana smoke as opposed to nicotine in tobacco. Inter-estingly, while habitual marijuana smoking produceshistopathologic evidence of airway inflammation andinjury that is similar to that observed in tobacco smok-ers (Fligiel et al., 1997; Roth et al., 1998), its long-termuse is not associated with the same development of ob-structive lung disease (Tashkin et al., 1997). THC hasalso been reported to induce expression of cytochromeP4501A1 (CYP1A1), an enzyme linked to the acti-vation of polycyclic aromatic hydrocarbons and lungcancer (Bartsch et al., 1992; Roth et al., 2001). Mari-juana smokers exhibit a pattern of molecular changesin their bronchial epithelial cells similar to that foundin tobacco smokers with over-expression of epidermalgrowth factor receptor, up-regulation of the Ki-67 nu-clear proliferative antigen, and changes in actin andploidy (Barsky et al., 1998), consistent with a link tobronchogenic carcinoma. However, epidemiologic ev-idence linking long-term use of marijuana to cancersof the respiratory epithelium is still limited (Zhang eta

gice seda heire 1,2 n off on-d tod n ofg ithe-l ed tot lialc atede smsu ans ert e us-i used

on stress-response pathways. SAE cells, obtained com-mercially, consist of proliferating type II epithelial cellsisolated from small airways of post mortem human lungtissue. The results demonstrate that exposure to THCimpacts on several different pathways and suggest thepotential for detrimental health effects associated withmarijuana smoking. This work presents initial gene ex-pression studies aimed at defining protein changes andtoxicological pathways induced by marijuana smoke inpulmonary epithelial cells.

2. Materials and methods

2.1. Cell culture and exposure conditions

Cryopreserved human SAE cells from disease-freelung tissue were obtained from Cambrex/BioWhittaker(Walkersville, MD) and cultured in Small Airway CellBasal Media (SABMTM) supplemented with 0.1 ng/mlretinoic acid, 0.5 mg/ml BSA, 6.5 ng/ml triiodothy-ronine, 0.5�g/ml epinephrine, 0.5�g/ml hydrocorti-sone, 5�g/ml insulin, 0.03 mg/ml bovine pituitary ex-tract, 10�g/ml transferrin, 0.5 ng/ml recombinant hu-man EGF, 50�g/ml gentamicin, and 50 ng/ml ampho-tericin B. SAE cells grown under these conditionsretain proliferative capacity for at least 15 cell divi-sions and stain positively for cytokeratin 19, identify-ing them as type II epithelial cells. Cells were platedat a density of 5× 103/cm2 in T25 flasks and subcul-t flu-e xper-i tox-i with-o dies,a dedi in-t e-d a-t );c -z I);2f os-i an-c ac-t ol-l us

l., 1999; Rosenblatt et al., 2004).As a mechanism for understanding the toxicolo

ffects of marijuana smoke and THC, we have uvariety of in vitro exposure systems to assay t

ffects on lung cells (Sarafian et al., 1999, 200003). THC induces oxidative stress, suppressio

as-induced apoptosis, and impairment of mitochrial function. The goal of the present study wasetermine if THC produced a reproducible patterene expression changes in human pulmonary ep

ial cells, and whether these changes might be relathe observed effects of THC on pulmonary epitheell function. In order to better assess the integrffects of THC exposure and potential mechaninderlying these effects, we cultured primary hummall airway epithelial (SAE) cells with THC and othoxicants in vitro and assayed the biologic responsng a targeted gene expression array technique foc

ured when they reached approximately 85% connce. Media was replaced every 2–3 days and e

ments were conducted at passage 3–5. For 24-hcant exposures, agents or vehicle were addedut media replacement. For 1-week exposure stugents or the ethanol-containing vehicle were ad

mmediately following media replacement at 2-dayervals. Agents studied included: THC diluted in mia from a 50 mg/ml stock obtained from the N

ional Institute on Drug Abuse (NIDA; Bethesda, MDarbonyl cyanide-p-(trifluoramethoxy)-phenyl hydraone (FCCP) from Sigma–Aldrich (St. Louis, M,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) obtained

rom the National Cancer Institute Carcinogen Reptory (Bethesda, MD). A549 cells, a human lung cer cell line with bronchoalveolar epithelial chareristics (CCL-185, American Tissue Culture Cection, Bethesda, MD), were grown in continuo

T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107 97

culture and used for comparative purposes in someassays.

2.2. Cell viability

As an overall measure of toxic injury, cell via-bility was determined before and after either 24-hor 7-day exposures by propidium iodide (PI) assay(Sarafian et al., 2003). PI was added to cell culturesto a final concentration of 50�M and fluorescencemeasured 15 min later with a Cytoflour 2300-platereader (PerSeptive Biosystems, Farmingham, MA) atEx = 546 and Em = 590 (sensitivity = 3). Backgroundmeasurements were derived from cell-free wells con-taining media and PI. Digitonin (160�M) was thenadded and fluorescence measurements repeated after20 min to obtainFmax, a function of total cell number.Percentage viability was calculated as [(F-blank/Fmax-blank)× 100], whereF is the measured PI fluores-cence. Six replicates were analyzed for each treatmentcondition.

2.3. Mitochondrial membrane potential

Mitochondrial membrane potential was measuredusing the fluorescent probe JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolyl-carbocyanineiodide from Molecular Probes, Eugene, OR) and aCytoflour 2300-plate reader as described previously(Sarafian et al., 2003). Cells cultured in 48-well plates

d

ingre

s,lln,-s--

insin

2.5. RNA isolation

SAE cells were harvested using a trypsin–EDTA so-lution, resuspended in PBS containing 0.04% RNAse-cure inhibitor (Ambion, Austin, TX), and lysed. Totalcellular RNA was isolated using the RNeasy kit (Qia-gen, Alameda, CA) and purity determined by measur-ingA260/A280ratio. 16S and 28S ribosomal RNAs wereevaluated by agarose/formaldehyde gel electrophore-sis. Samples withA260/A280> 1.7 and 28S/16S bandsdisplaying∼2:1 relative intensity ratio were stored at−70◦C for analysis.

2.6. Gene expression array

Gene expression arrays were performed using a lu-minescence detection reagent with the “Human Stressand Toxicity PathwayFinder kit” (SuperArray Inc.,Frederick, MD). A list of the genes included in thisarray is provided asSupplementary data, Table S1.Samples containing 1–2�g RNA were used to generatebiotinylated cDNA sequences with the AmpoLabeling-LPR kit. Following hybridization to membranes, wash-ing, and labeling, signals were detected by autoradiog-raphy with XAR-5 film (Kodak) exposed for five to sixtimed periods ranging from 5 s to 2 min. Autoradio-grams were scanned and densitometry of each tetradgene spot was performed using Scion Image for Win-dows Software. Average background densities were de-termined using four to six internal blank spots and val-u eep-i nase.T par-a ex-p werep

2

pen-d testg for� weres s ini icalr kersa kea 8;

were treated for 1 day or 7 days with THC as describeabove, then exposed to 1�M JC-1 (60 min, 37◦C) andred/green fluorescence ratios were determined usbackground-subtracted values. Six replicates weanalyzed for each condition.

2.4. ATP assay

As an independent measure of cellular energeticcellular ATP levels were determined using the CeTiter-Glo luminescent assay kit (Promega, MadisoWI) according to the manufacturer’s specifications. Luminescence was quantified in opaque white plates uing a microplate luminometer (Berthold Detection Systems, Orion Instruments, Baton Rouge, LA). Protevalues were determined from triplicate parallel wellfor each treatment condition using the Bio-Rad Proteassay (Hercules, CA).

es normalized to average densities for the “housekng gene” glyceraldehyde-3-phosphate dehydrogeoxicant and control samples were always run inllel to assure consistency in the cell growth andosure conditions. Three separate determinationserformed for each condition.

.7. Real-time RT-PCR

Real-time RT-PCR was performed as an indeent measure of relative gene expression for fourenes: PTGS-2, IL-1A, CYP1A1, and Bax, and-actin as an internal standard. These geneselected for confirmation because of their rolenflammation, carcinogensis, and apoptosis; biologesponse pathways that occur in the lungs of smond in vitro following exposure to marijuana smond/or THC (Fligiel et al., 1997; Roth et al., 199

98 T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107

Sarafian et al., 2001). RNA samples from 7-dayTHC-treated SAE cells and control vehicle-treatedcells (4�g each) were treated with RNase-freeDNase I (Qiagen), then purified using the RNAeasyprotocol and reverse-transcribed to prepare cDNA.Real-time PCR was then performed with a MiIQCycler (Bio-Rad) using amplification cycle andmelt curve sequences consisting of one cycle at95◦C for 3 min followed by 40 cycles of 30 s at95◦C and 45 s at 55◦C. The melt curve gradientwas analyzed from 55 to 95◦C with a temperatureincrease of 0.5◦C every 10 s. Standard curves weregenerated using control cDNA samples with serialfive-fold dilutions from 25 ng/�l to 40 pg/�l of cDNAproduct. Assays were performed in triplicate witheach 50�l reaction mixture containing 12�l cDNA,25�l Bio-Rad Supermix, 200 nM each of sense andanti-sense primers, and 13�l of nuclease-free water.Primers were purchased from Invitrogen (Carlsbad,CA) and consisted of the following sequences:�-actin sense—5′-GTACCACTGGCATCGTGAT,�-ac-tin anti-sense—5′-ATCTTCATGAGGTAGTCAGTCA;PTGS-2 sense—5′-CCAGTTTGTTGAATCATTCA-CCAG, PTGS-2 anti-sense—5′-AAGCGTTTGCG-GTACTCATTAAA; IL-1A sense—5′-GCCACAAA-GCAAGACTACTGG, IL-1A anti-sense—5′-ATAT-GAACTGTCAACACTGCACAA; Bax sense—5′-A-GGATGCGTCCACCAAGAAG, Bax anti-sense—5′-GTCCACGGCGGCAATCAT; CYP1A1 sense—5′-TAATACGACTCACTATAGGCCAGGAGGAGCTA-GAT

2

aluesw de-t thet wereb oad-i theh epingg ntala sonswc d ast ant

change with THC exposure. Linear regression analysiswas used to compare fold changes in gene expressionas measured by the array with the fold changes in geneexpression as measured by RT-PCR.p-Values <0.05were considered significant.

3. Results

3.1. Effects of THC on cell viability and energetics

Concentration–response studies were performed toassess the overall impact of short-term (24-h) or long-term (7-day) exposure to THC on cell viability andenergetics (Fig. 1). While exposure to THC for 24 hhad minimal effects on A549 cell viability, there wasa significant concentration-dependent effect on SAEcells with viability reduced by approximately 50%at 5�g/ml. Measurements of ATP and mitochondrialmembrane potential were also more sensitive to the ef-fects of THC in SAE cells than in A549 cells, indicat-ing toxic effects on cell energetics at concentrations aslow as 0.75�g/ml. Pretreatment of SAE cells with cy-closporin A (Cys A; 4�M), known to stabilize the mi-tochondrial permeability transition pore, significantlyprotected against loss of ATP, confirming the link be-tween the mitochondrial membrane potential and ATPassays.

The toxic effects of THC increased with time inculture (Fig. 1). After a 7-day exposure, significant ef-f m-b ationo nep after1 ntrola fc

3 dS

Ar-r s in-d ion,a ut oft 1a oveb B1

ACCCAGTGATT-3′, CYP1A1 anti-sense—5′-ACTCAGATGGGTGCTTCATAGCTTCTGGTCA-GGTTGA-3′.

.8. Statistical analysis

Functional assays were expressed as mean vith standard error of the means for the replicate

erminations. Densitometry measurements fromhree replicate gene arrays for each experimentackground-subtracted and then normalized for l

ng by comparison to internal measurements ofousekeeping genes. Gene of interest/housekeene ratios <0.04 for both control and experimerrays were excluded from analysis and compariere performed using a pairedt-test with two tails. Noorrection for multiple comparisons was performehe majority of genes (>85%) displayed no signific

ects on viability, ATP level, and mitochondrial merane potential were observed at a THC concentrf 0.25�g/ml. The effect on mitochondrial membraotential was more prominent than that observed-day exposure, decreasing to less than 55% of cot a THC concentration of 0.25�g/ml and to 25% oontrol at 1�g/ml.

.2. Baseline gene expression detected in cultureAE cells

The Human Stress and Toxicity PathwayFinderay contains representative genes from pathwayicative of heat shock, oxidative stress, inflammatpoptosis, DNA damage, and carcinogenesis. O

he 96 test genes (seeSupplementary data, Table S),pproximately one-third were not detectable abackground, including cytochrome P450s (CYP)1

T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107 99

Fig. 1. Time- and concentration-dependent effects of THC on the viability and cellular energetics of cultured small airway epithelial (SAE)cells. SAE cells were cultured for either 1 day (top row) or 7 days (bottom row) with varying concentrations of THC (0–10�g/ml) and assayedfor viability by PI assay (left panels), ATP content by luminescence assay (middle panels), and mitochondrial membrane potential by red/greenfluorescence ratio after loading with JC-1 indicator dye (right panels). For 1-day exposures, the impact of THC on viability was also assessedfor the A549 lung epithelial tumor cell line. Cyclosporin A (Cys A) was added in some experiments to stabilize the mitochondrial permeabilitytransition pore. Data represent mean values of 3–6 replicate measurements± S.E.M. *p< 0.05 compared with vehicle-treated control usingone-way ANOVA and Fisher’s post hoc test.

and 2E, inducible nitric oxide synthase (iNOS), super-oxide dismutase 1 (SOD1), p53, tumor necrosis factor(TNF), and TNF receptor. Included among the mostprominently expressed genes were catalase, hemeoxy-genase 2 (HMOX2), heat shock factor 1 (HSF1), andSOD2 suggesting a chronic oxidative stress associatedwith cell culture. Several heat shock proteins (HSPs),epoxide hydrolase 2 (EPHX2), and X-ray repair cross-complementing group 1 (XRCC1) were also stronglyexpressed. When expression profiles from the controlSAE cells obtained on different days were compared,some degree of baseline variability was detected. Thesedifferences suggest that cell passage, culture density,nutrient levels, and other factors related to the cultureenvironment impact on the baseline stress and toxicityprofile. These effects were controlled for by runningcomparison assays in parallel from the same passage.

3.3. Gene expression changes associated withexposure to THC

Cell viability curves were used to identify concen-trations of THC that were relatively non-toxic duringthe 1-day or 7-day culture periods (0.75 and 0.25�g/mlTHC, respectively). Consistent with this, total cell andRNA yields from cultures were always comparable forcontrol and THC-exposed cells. Of the 96 test genes,expression of 10 were reproducibly altered followingthe 7-day THC exposure (p≤ 0.05;Fig. 2). When nor-malized to internal housekeeping genes and then com-pared to control cells, four genes demonstrated higherexpression following exposure to THC and six demon-strated lower expression. Gene expression that in-creased following the 7-day exposure to THC includedCYP1A1 (19% increase), IL-1A (51% increase),

100 T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107

Fig. 2. Gene expression profiles for SAE cells cultured for 7 days with THC. (A) SAE cells were cultured in medium containing 0.1%ethanol (control) or medium containing 0.25�g/ml THC with fresh medium containing these additives replaced every 2 days. mRNA washarvested after 7 days and the resulting cDNA hybridized to Human Stress and Toxicity PathwayFinder Arrays followed by chemiluminescencedetection. Experiments were carried out in triplicate and scanned images from one pair of representative arrays are shown. Gene spots aresequentially numbered for identification purposes and boxes added to identify genes where the mean relative density values are higher in onecondition compared to the other (p< 0.05). (B) Autoradiograms were scanned and densitometric analysis performed using Scion Image forWindows Software. Background-subtracted values were normalized for the expression of the housekeeping gene, glyceraldehyde-3-phosphatedehydrogenase (GAPD, spots 103 and 104), to determine relative expression values. Gene spots with densities less than 0.04 for both treated andcontrol arrays relative to GAPD were excluded from results. Ten genes were reproducibly and differentially expressed upon exposure to THC(p≤ 0.05), including four up-regulated genes (left panel) and six down-regulated genes (right panel). Relative THC:control expression ratios foreach of the identified genes are indicated by values above the bar graphs.

GADD45A (63% increase), and PTGS-2 (3.38-foldincrease). Decreases in expression, comparing vehi-cle control to THC-treated SAE cells, ranged from13% for Bax to 43% for HSP701A. PLAB expressionwas undetectable in THC-treated cells, but consistently

expressed at very low levels in control cells, making itdifficult to assess the relative fold difference. Five ofthe six down-regulated genes were heat shock-related.Gene expression array results for all experiments aresummarized inSupplementary data, Table S2.

T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107 101

Real-time RT-PCR was performed as an indepen-dent approach for measuring the relative differencein gene expression for four genes: PTGS-2, IL-1A,CYP1A1, and Bax. Relative expression of�-actinmRNA was used for normalization. In each case, thepattern of THC-induced expression change reported byRT-PCR was similar to that observed using the array.A high correlation coefficient was obtained when thetwo approaches were compared by linear regression

analysis (R2 = 0.9553). However, the relative differ-ences in gene expression between control and THC-treated cells were approximately two- to three-foldlarger when analyzed by RT-PCR, suggesting that thegene array method may be less sensitive at detectingdifferences between control and test samples (data notshown).

Gene expression profiles for the 1-day expo-sure differed significantly from the 7-day exposure

F(asa(

ig. 3. Gene expression profiles for SAE cells cultured for 1 day withcontrol) or medium containing 0.75�g/ml THC for 24 h. mRNA was harnd Toxicity PathwayFinder Arrays (SuperArray) followed by chemilucanned and densitometric analysis performed using Scion Image fornd differentially expressed upon exposure to THC (p≤ 0.05), including tright panel). Relative THC:control expression ratios for each of the id

THC. (A) SAE cells were cultured in medium containing 0.1% ethanolvested after 24 h and the resulting cDNA hybridized to Human Stressminescence detection as described inFig. 2. (B) Autoradiograms wereWindows Software as described forFig. 2. Nine genes were reproducibly

wo up-regulated genes (left panel) and seven down-regulated genesentified genes are indicated by values above the bar graphs.

102 T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107

Fig. 4. Gene expression profiles for SAE cells cultured for 1 day with either TCDD or FCCP. SAE cells were cultured under identical conditionsin either control medium containing 0.1% DMSO vehicle or medium containing 100 nM TCDD or 10�M FCCP. mRNA was harvested after 24 hand the resulting cDNA hybridized to Human Stress and Toxicity PathwayFinder Arrays followed by chemiluminescence detection. Experimentswere carried out in triplicate and scanned images from one set of representative arrays are shown. Gene spots are sequentially numbered foridentification purposes and boxes added to identify genes where the mean relative density values are higher in one condition compared to theother (p< 0.05).

T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107 103

Fig. 5. The overlapping effects of THC, TCDD, and FCCP on the expression of genes measured with the Toxicity PathwayFinder arrays. SAEcells were cultured for 24 h with either control medium, 100nM TCDD, 10�M FCCP, or THC; results for THC treatment represent changes inexpression due to either 24-h exposure to 0.75�g/ml THC or 7-day exposure to 0.25 mg/ml THC. (A) Genes whose expression increased as aresult of the respective treatments indicated (up-regulated genes). (B) Genes which were down-regulated by the treatments. Genes which wereeither up-regulated or down-regulated by more than one treatment are represented by the overlapping areas. The numbers indicate the numberof genes affected by that particular treatment.

(Fig. 3). With respect to the comparison betweenTHC-exposed and control cells, a decrease in Baxexpression was the only change common to boththe 1- and 7-day exposure paradigms, the effect be-ing of a greater magnitude in the acute exposure(65% decrease,p= 0.035). Of the nine changes in-duced by a 1-day exposure to THC, seven genesdecreased in expression while two increased. De-creases ranged from 17% to 65%, while increaseswere 1.32-fold for MT-1A (p= 0.001) and 2.56-foldfor PLAB (p= 0.006). Both PLAB and CYP1A1 ex-pression were altered in opposite directions relativeto that caused by prolonged THC exposure; PLABincreased while CYP1A1 decreased. Expression ofPTGS-2, IL-1A, GADD45A, XRCC1, and most of theheat shock genes were not affected by 24-h THC ex-posure.

3.4. TCDD and FCCP as model toxicants

Two additional agents, TCDD and FCCP, were ex-amined by the same targeted gene arrays as modelsfor the effects of specific pathway toxicants on SAEcells (Figs. 4 and 5). TCDD is a prototypic activa-tor of the aryl hydrocarbon (Ah) receptor, which me-diates virtually all of its biological effects includinginduction of CYP1A1 gene expression (Hankinson,

1995; Fernandez-Salguero et al., 1996). Twenty-four-hour exposure to 100 nM TCDD resulted in de-creased expression of 28 out of 96 test genes, in-cluding apoptosis regulatory genes of the bcl-2 fam-ily, early growth response (EGR) protein, HMOX2, and HSF-1. Several of these changes were alsoobserved in response to THC including Bax, Casp-8, CDKN1, EGR, HSF1, and GRP78. The onlygene that increased in expression in response toTCDD was CYP1A1, a gene expressed prominentlyin control, untreated SAE cells, and also regu-lated by THC. TCDD exposure resulted in a 30%increase in CYP1A1 expression under the condi-tions assayed. Twenty-four-hour exposure to FCCPcaused a decrease in expression of 33 out of 96genes, including representatives of apoptosis, anti-oxidant, heat shock, inflammation, and DNA re-pair gene classes. FCCP depolarizes the mitochon-drial membrane and uncouples oxidative phospho-rylation from electron transport (Dispersyn et al.,1999). Six genes increased in expression: CYP1A1(30%), PLAB (not detected in control), PTGS-2 (5-fold), HSPE1 (45%), HMOX1 (2.1-fold), and SOD2(2.9-fold). Two of these were also up-regulated byTHC to similar extents, although CYP1A1 inductionwas observed after 7-day THC but not after 24-hexposure.

104 T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107

4. Discussion

Marijuana is smoked primarily as a mechanism fordelivering nanomolar concentrations of THC to thebrain. However, during the process, milligram quan-tities THC are inhaled into the lung and may con-tribute to the inflammation, cellular injury, and molec-ular dysregulation that has been observed in the lungsof marijuana smokers (Wu et al., 1988; Roth et al.,2001; Tashkin et al., 2002). In vitro studies sug-gest that THC enhances oxidative stress, regulatesthe metabolism of polycyclic aromatic hydrocarbons,suppresses receptor-mediated apoptosis, and can dis-rupt mitochondrial function and cellular energetics(Sarafian et al., 1999, 2001, 2003; Roth et al., 2001).This pattern of biological consequences is distinct fromthat associated with tobacco products, suggesting a spe-cific role for THC as a toxic mediator.

This study focused on the interaction of THC withprimary cultures of human SAE cells, representing acentral target for smoke-related injury and carcinogen-esis in vivo. Similar to effects on A549 cells (Sarafianet al., 2003), THC induced a time- and concentration-dependant decrease in mitochondrial membrane po-tential, ATP level, and cell viability in SAE cells.However, SAE cells were significantly more sensi-tive to these toxic effects. While functional changes inA549 required acute exposure to THC in the range of5–10�g/ml, similar changes were observed when SAE

m-of

r-

-ithreay

nddg-ofs-

ity

to withstand a variety of biological stresses known tooccur in the lung including thermal injury, hypoxemia,ischemia, radiation, and infection (Villar et al., 1993;Wong and Wispe, 1997; Yoo et al., 2000). Bax, encod-ing a pro-apoptotic factor, was also down-regulated.Bax is involved in the normal homeostatic response tocell injury and its down-regulation would be expectedto promote the accumulation of genomic injury andundesired cells (McDonnell et al., 1996). The PLABgene was expressed at low levels in control SAE cellsbut not at all in cells exposed to THC for 7 days. PLAB(also called NAG-1) has been described as a differentia-tion and pro-apoptotic factor for airway epithelial cells(Newman et al., 2003). Its down-regulation is associ-ated with lung carcinogenesis while its up-regulationinhibits tumorigenicity (Newman et al., 2003). Collec-tively, the pattern of genes down-regulated upon ex-posure to THC suggests that exposed lung epitheliumwill be more susceptible to smoke-related injury andcarcinogenesis.

Expression levels for 4 of the 96 test genes were el-evated following a 7-day exposure to THC: CYP1A1,GADD45A, PTGS-2, and IL-1A. We previously re-ported that THC can interact with the aryl hydrocarbonreceptor to induce the expression of CYP1A1 in A549cells in vitro (Roth et al., 2001). CYP1A1 biotrans-forms polycyclic aromatic hydrocarbons into proxi-mate carcinogens and the induction of CYP1A1 bytobacco smoke plays a well-established role in the de-velopment of lung cancer (Bartsch et al., 1992). Fur-t omt andI a-l dV t al.,2 toDP thep andl12 ndinsa ,1 ina in ahh toticp c

her carcinogenic potential may be implicated frhe increased expression of GADD45A, PTGS-2,L-1A genes, each of which may play a role in mignant transformation (Huang et al., 1998; Apte anoronov, 2002; Yamasawa et al., 2002; Czekay e003). GADD45A is up-regulated following damageNA in an effort to allow repair (Fornace et al., 1989).TGS-2 and IL-1A, encoding factors involved inro-inflammatory cascade, are often co-regulated

ead to increased cellular oxidative stress (Martin et al.,994; Nam et al., 1995; Soloff et al., 2004). The COX-gene, encoded by PTGS-2, generates prostagland eicosanoids from arachidonic acid (Herschman996). While normally expressed at very low levelsirway epithelial cells, PTGS-2 is over-expressedigh percentage of lung tumors (Huang et al., 1998) andas been linked to increased levels of the anti-apoprotein survivin (Krysan et al., 2004), pro-angiogeni

cells were exposed to concentrations of 1–2.5�g/ml for24 h. When the exposure was extended to 7 days, a siilar response pattern occurred at even lower levelsTHC (0.25–0.5�g/ml).

In an effort to characterize the mechanisms undelying the toxic effects of THC on lung epithelial cells,we employed a targeted gene array to identify THCinduced changes in gene expression patterns. As wcell energetics, effects of THC on gene expression wenot identical at 24 h and 7 days of exposure. Seven-dexposure to 0.25�g/ml THC decreased the expressionof 6 out of 96 genes assessed by the Human Stress aToxicity PathwayFinder Array. Five heat shock-relateproteins were affected, suggesting a coordinated reulation of this stress-response pathway. ExpressionHSP70 was similarly decreased in rat hippocampal tisue following prolonged in vivo administration of THCin an animal model (Kittler et al., 2000). Low levels ofHSP suggest a reduction in the exposed cell’s capac

T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107 105

chemokines (Pold et al., 2004), and tissue invasivenessfactors (Dohadwala et al., 2002), as well as to an imbal-ance of IL-10 and IL-12 that suppresses anti-tumor im-munity and promotes tumor growth (Huang et al., 1998;Stolina et al., 2000; Mao et al., 2003). Inhibitionof PTGS-2 appears to protect against tumorigenesisand provides the rationale for numerous clinical tri-als using COX-2 inhibitors as cancer chemopreven-tion agents (Williams et al., 1999; Liao et al., 2003;Mao et al., 2003; Richardson et al., 2003). THC-induced elevation of PTGS-2 and IL-1A may alsocontribute to airway inflammation observed in habit-ual marijuana smokers (Roth et al., 1998). Elevationof PTGS-2 and IL-1A was observed following 7-dayTHC but not after 1-day exposure, suggesting thatpro-inflammatory effects do not occur acutely in air-way epithelial cells but require extended, repeated ex-posures. Similar temporal discordance was observedfor expression of the CYP1A1 gene, which was de-creased after 24 h of THC exposure but increased after7 days.

Three separate gene expression array experimentsare reported in this study, each performed at differ-ent times using different SAE cell populations. Thesemethodologic differences likely contributed to the vari-ations in expression patterns displayed by the con-trol arrays in these experiments and may have con-tributed to the different response patterns elicited bythe various chemical treatments. Other factors whichmay have influenced basal gene expression were thel ish( ehi-c um-b hichc runi ri-m es ing Indep on-fi enest matet res-s ts

ringa SAEc ikeT A1

mRNA synthesis (Roth et al., 2001). Thus, if Ah re-ceptor activation and CYP1A1 expression are primaryeffector pathways for THC action, the gene expres-sion patterns produced by TCDD and THC should bevery similar. Analogously, both THC and FCCP havebeen shown to disrupt mitochondrial function and de-crease cellular ATP levels. Similarly, we hypothesizedthat the role of THC-mediated mitochondrial injuryon gene expression might be revealed by compara-ble changes associated with exposure to FCCP. Al-though dissimilarities were apparent when compar-ing gene expression changes induced by THC ver-sus TCDD or FCCP, marked congruence was alsonoted in expression changes for several genes, par-ticularly CYP1A1, apoptosis-related, and heat shock-related genes. These results suggest that regulatorypathways associated with activation of the Ah re-ceptor and with disruption of mitochondrial functionmay partially account for the toxicological effects ofTHC.

In summary, exposure of primary human lung ep-ithelial cells to THC produced time- and concentration-dependent effects on mitochondrial membrane poten-tial, cellular energetics, viability, and gene expressionassociated with stress and toxicology pathways. Thetargeted arrays identified decreases in the expressionof Bax (1- and 7-day exposure) and caspase-8 (1-day exposure) that may account for the previously ob-served suppression of fas-mediated apoptosis that oc-

ex-in-terfor

de-a-lapmi-

rylg-ndy-n-es-ud-leshemnd

ength of residence time in the same culture d2 days versus 8 days), cell density, solvent vle (ethanol versus DMSO), and cell passage ner. Nevertheless, the experimental design, in wontrol and toxicant-exposed cells were alwaysn parallel, with triplicate samples for THC expe

ents, suggested that the reproducible differencene expression levels are toxicant-dependent.endent corroboration with real time RT-PCR crmed the expression patterns observed in all gested and indicates that the arrays underestihe magnitude of the induced changes in expion by approximately two-fold (R2 = 0.9553, data nohown).

As an approach to understand how THC may bbout the observed changes in gene expression,ells were also exposed to TCDD and FCCP. LCDD, THC has been shown to induce CYP1

-

curs when A549 lung adenocarcinoma cells areposed to either THC and marijuana smoke. Theduction of genes encoding IL-1A and PTGS-2 af7 days of exposure raises concern about the riskmalignancy and is consistent with the enhancedgree of inflammation observed in the airways of hbitual marijuana smokers. We observed partial overbetween these effects and those produced by thetochondrial toxin, FCCP, and an activator of the ahydrocarbon receptor, TCDD. These similarities sugest that effects of THC on cellular energetics aits previously described interaction with the aryl hdrocarbon receptor pathway may significantly cotribute to the observed modulation of gene exprsion that occurred in response to THC. Further sties are warranted to define and confirm precise rofor these changes in gene expression and relate tto the pulmonary toxicology of marijuana smoke aTHC.

106 T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107

Acknowledgements

The authors are grateful to Annie Balian, KarenBadal, David Russell, Laura Lin, and Laura Kurek forassistance on this project and to Dr. David Elashoff (De-partment of Biostatistics, UCLA) for consultation onanalysis of gene array data. We also thank Drs. RobertStrieter and David Heber (Department of Medicine,UCLA) for use of equipment. This work was supportedby National Institute on Drug Abuse/National Instituteof Health Grant R37DA030-20.

Appendix A. Supplementary data

Supplementary data associated with this ar-ticle can be found, in the online version, atdoi:10.1016/j.toxlet.2005.03.008.

References

Adams, I.B., Martin, B.R., 1996. Cannabis: pharmacology and tox-icology in animals and humans. Addiction 91, 1585–1614.

Apte, R.N., Voronov, E., 2002. Interleukin-1—a major pleiotropiccytokine in tumor–host interactions. Semin. Cancer Biol. 12,277–290.

Barsky, S.H., Roth, M.D., Kleerup, E.C., Simmons, M., Tashkin,D.P., 1998. Histopathologic and molecular alterations inbronchial epithelium in habitual smokers of marijuana, cocaine,and/or tobacco. J. Natl. Cancer Inst. 90, 1198–1205.

Bartsch, H., Petruzzelli, S., De Flora, S., Hietanen, E., Camus, A.M.,tini,d thecen-t. 98,

C 003.ellu-91.

D s, H.,d mi-ells.

D ld,rinean-4 in277,

F .M.,mice-

Fligiel, S.E., Roth, M.D., Kleerup, E.C., Barsky, S.H., Simmons,M.S., Tashkin, D.P., 1997. Tracheobronchial histopathology inhabitual smokers of cocaine, marijuana, and/or tobacco. Chest112, 319–326.

Fornace Jr., A.J., Nebert, D.W., Hollander, M.C., Luethy, J.D., Pap-athanasiou, M., Fargnoli, J., Holbrook, N.J., 1989. Mammaliangenes coordinately regulated by growth arrest signals and DNA-damaging agents. Mol. Cell. Biol. 9, 4196–4203.

Hankinson, O., 1995. The aryl hydrocarbon receptor complex. Annu.Rev. Pharmacol. Toxicol. 35, 307–340.

Herschman, H.R., 1996. Prostaglandin synthase 2. Biochim. Bio-phys. Acta 1299, 125–140.

Huang, M., Stolina, M., Sharma, S., Mao, J.T., Zhu, L., Miller, P.W.,Wollman, J., Herschman, H., Dubinett, S.M., 1998. Non-smallcell lung cancer cyclooxygenase-2-dependent regulation of cy-tokine balance in lymphocytes and macrophages: up-regulationof interleukin 10 and down-regulation of interleukin 12 produc-tion. Cancer Res. 58, 1208–1216.

Johnston, L.D., O’Malley P.M., Bachman J.G., 2003. Monitoringthe Future National Survey Results on Adolescent Drug Use:Overview of Key Findings, 2002. National Institute on DrugAbuse, Bethesda, MD, p. 56.

Kittler, J.T., Grigorenko, E.V., Clayton, C., Zhuang, S.Y., Bundey,S.C., Trower, M.M., Wallace, D., Hampson, R., Deadwyler, S.,2000. Large-scale analysis of gene expression changes duringacute and chronic exposure to [Delta]9-THC in rats. Physiol.Genomics 3, 175–185.

Krysan, K., Merchant, F.H., Zhu, L., Dohadwala, M., Luo, J., Lin, Y.,Heuze-Vourc’h, N., Pold, M., Seligson, D., Chia, D., Goodglick,L., Wang, H., Strieter, R., Sharma, S., Dubinett, S., 2004. COX-2-dependent stabilization of surviving in non-small cell lung can-cer. FASEB J. 18, 206–208.

Liao, Z., Milas, L., Komaki, R., Stevens, C., Cox, J.D., 2003. Combi-nation of a COX-2 inhibitor with radiotherapy or radiochemother-apy in the treatment of thoracic cancer. Am. J. Clin. Oncol. 26,

M es,atestionRes.

M elt-2 ex-cells.

M , P.,ion.

N esisvine69–

N ling,idalheo-ol.

Castegnaro, M., Alexandrov, K., Rojas, M., Saracci, R., GiunC., 1992. Carcinogen metabolism in human lung tissues aneffect of tobacco smoking: results from a case-control multiter study on lung cancer patients. Environ. Health Perspec119–124.

zekay, R.P., Aertgeerts, K., Curriden, S.A., Loskutoff, D.J., 2Plasminogen activator inhibitor-1 detaches cells from extraclar matrices by inactivating integrins. J. Cell Biol. 160, 781–7

ispersyn, G., Nuydens, R., Connors, R., Borgers, M., Geert1999. Bcl-2 protects against FCCP-induced apoptosis antochondrial membrane potential depolarization in PC12 cBiochim. Biophys. Acta 1428, 357–371.

ohadwala, M., Batra, R.K., Luo, J., Lin, Y., Krysan, K., PoM., Sharma, S., Dubinett, S.M., 2002. Autocrine/paracprostaglandin E2 production by non-small cell lung ccer cells regulates matrix metalloproteinase-2 and CD4cyclooxygenase-2-dependent invasion. J. Biol. Chem.50828–50833.

ernandez-Salguero, P.M., Hilbert, D.M., Rudikoff, S., Ward, JGonzalez, F.J., 1996. Aryl-hydrocarbon receptor-deficientare resistant to 2,3,7,8-tetrachlorodibenzo-p-dioxin-induced toxicity. Toxicol. Appl. Pharmacol. 140, 173–179.

S85–S91.ao, J.T., Roth, M.D., Serio, K.J., Baratelli, F., Zhu, L., Holm

E.C., Strieter, R.M., Dubinett, S.M., 2003. Celecoxib modulthe capacity for prostaglandin E2 and interleukin-10 producin alveolar macrophages from active smokers. Clin. Cancer9, 5835–5841.

artin, M., Neumann, D., Hoff, T., Resch, K., DeWitt, D.L., GoppStruebe, M., 1994. Interleukin-1-induced cyclooxygenasepression is suppressed by cyclosporin A in rat mesangialKidney Int. 45, 150–158.

cDonnell, T.J., Beham, A., Sarkiss, M., Andersen, M.M., Lo1996. Importance of the Bcl-2 family in cell death regulatExperientia 52, 1008–1017.

am, M.J., Thore, C., Busija, D., 1995. Role of protein synthin interleukin 1 alpha-induced prostaglandin production in oastroglia. Prostaglandins Leukot. Essent. Fatty Acids 53,72.

ewman, D., Sakaue, M., Koo, J.S., Kim, K.S., Baek, S.J., ET., Jetten, A.M., 2003. Differential regulation of nonsteroanti-inflammatory drug-activated gene in normal human tracbronchial epithelial and lung carcinoma cells by retinoids. MPharmacol. 63, 557–564.

T. Sarafian et al. / Toxicology Letters 158 (2005) 95–107 107

Novotny, M., Merli, F., Weisler, D., Fencl, M., Saeed, T., 1982. Frac-tionation and capillary gas chromatographic mass spectrometriccharacterization of the neutral components in marijuana and to-bacco smoke concentrates. J. Chromatogr. 238, 141–150.

Pold, M., Zhu, L.X., Sharma, S., Burdick, M.D., Lin, Y., Lee, P.P.,Pold, A., Luo, J., Krysan, K., Dohadwala, M., Mao, J.T., Ba-tra, R.K., Strieter, R.M., Dubinett, S.M., 2004. Cyclooxygenase-2-dependent expression of angiogenic CXC chemokines ENA-78/CXC ligand (CXCL) 5 and interleukin-8/CXCL8 in humannon-small cell lung cancer. Cancer Res. 64, 1853–1860.

Richardson, C.M., Sharma, R.A., Cox, G., O’Byrne, K.J., 2003.Epidermal growth factor receptors and cyclooxygenase-2 in thepathogenesis of non-small cell lung cancer: potential targets forchemoprevention and systemic therapy. Lung Cancer 39, 1–13.

Rosenblatt, K.A., Daling, J.R., Chen, C., Sherman, K.J., Schwartz,S.M., 2004. Marijuana use and risk of oral squamous cell carci-noma. Cancer Res. 64, 4049–4054.

Roth, M.D., Arora, A., Barsky, S.H., Kleerup, E.C., Simmons,M., Tashkin, D.P., 1998. Airway inflammation in young mari-juana and tobacco smokers. Am. J. Respir. Crit. Care Med. 157,928–937.

Roth, M.D., Marques-Magallanes, J.A., Yuan, M., Sun, W., Tashkin,D.P., Hankinson, O., 2001. Induction and regulation of thecarcinogen-metabolizing enzyme CYP1A1 by marijuana smokeand delta (9)-tetrahydrocannabinol. Am. J. Respir. Cell Mol.Biol. 24, 339–344.

Sarafian, T.A., Kouyoumjian, S., Khoshaghideh, F., Tashkin, D.P.,Roth, M.D., 2003. Delta 9-tetrahydrocannabinol disrupts mito-chondrial function and cell energetics. Am. J. Physiol. Lung CellMol. Physiol. 284, L298–L306.

Sarafian, T.A., Magallanes, J.A., Shau, H., Tashkin, D., Roth, M.D.,1999. Oxidative stress produced by marijuana smoke. An adverseeffect enhanced by cannabinoids. Am. J. Respir. Cell Mol. Biol.20, 1286–1293.

Sarafian, T.A., Tashkin, D.P., Roth, M.D., 2001. Marijuana smokeeathcol.

S . Inand

il-8 in cultured human myometrial cells. Endocrinology 145,1248–1254.

Stolina, M., Sharma, S., Lin, Y., Dohadwala, M., Gardner, B.,Luo, J., Zhu, L., Kronenberg, M., Miller, P.W., Portanova,J., Lee, J.C., Dubinett, S.M., 2000. Specific inhibition of cy-clooxygenase 2 restores antitumor reactivity by altering thebalance of IL-10 and IL-12 synthesis. J. Immunol. 164, 361–370.

Tashkin, D.P., Baldwin, G.C., Sarafian, T., Dubinett, S., Roth, M.D.,2002. Respiratory and immunologic consequences of marijuanasmoking. J. Clin. Pharmacol. 42, 71–81.

Tashkin, D.P., Simmons, M.S., Sherrill, D.L., Coulson, A.H., 1997.Heavy habitual marijuana smoking does not cause an accelerateddecline in FEV1 with age. Am. J. Respir. Crit. Care Med. 155,141–148.

Villar, J., Edelson, J.D., Post, M., Mullen, J.B., Slutsky, A.S., 1993.Induction of heat stress proteins is associated with decreasedmortality in an animal model of acute lung injury. Am. Rev.Respir. Dis. 147, 177–181.

Williams, C.S., Mann, M., DuBois, R.N., 1999. The role of cyclooxy-genases in inflammation in cancer, and development. Oncogene18, 7908–7916.

Wong, H.R., Wispe, J.R., 1997. The stress response and the lung.Am. J. Physiol. 273, 1–9.

Wu, T.C., Tashkin, D.P., Djahed, B., Rose, J.E., 1988. Pulmonaryhazards of smoking marijuana as compared with tobacco. N.Engl. J. Med. 318, 347–351.

Yamasawa, K., Nio, Y., Dong, M., Yamaguchi, K., Itakura, M., 2002.Clinicopathological significance of abnormalities in Gadd45 ex-pression and its relationship to p53 in human pancreatic cancer.Clin. Cancer Res. 8, 2563–2569.

Yoo, C.G., Lee, S., Lee, C.T., Kim, Y.W., Han, S.K., Shim, Y.S.,2000. Anti-inflammatory effect of heat shock protein inductionis related to stabilization of I kappa B alpha through prevent-ing I kappa B kinase activation in respiratory epithelial cells. J.

Z .P.,e andneck.

and Delta(9)-tetrahydrocannabinol promote necrotic cell dbut inhibit Fas-mediated apoptosis. Toxicol. Appl. Pharma174, 264–272.

oloff, M.S., Cook Jr., D.L., Jeng, Y.J., Anderson, G.D., 2004situ analysis of interleukin-1-induced transcription of cox-2

Immunol. 164, 5416–5423.hang, Z.F., Morgenstern, H., Spitz, M.R., Tashkin, D.P., Yu, G

Marshall, J.R., Hsu, T.C., Schantz, S.P., 1999. Marijuana usincreased risk of squamous cell carcinoma of the head andCancer Epidemiol. Biomarkers Prev. 8, 1071–1078.