Tobacco Carcinogen Mediated Up-regulation of Ap-1 Dependent Pro-Angiogenic Cytokines in Head and Neck Carcinogenesis Wade G. swenson,"* rly RAK, Wuertz,’ and Frank G. Ondrey"# !Molecular Oncology Program, Department of Otolaryngology, University of Minnesota, Minneapolis, Minnesota *Masonic Cancer Center, Univesity of Minnesota, Minneapolis, Momesota Tobacco is notably genotoxic and associated with head and neck carcinogenesis. Cigarette carcinogens have the capacity to alter ealy response gene expression in toaacco-related malignancies via genes such as nuclear factor kappa & ((\FKB). A number of early response gene activation events are also faciitated by fox/un activator protein 1 (A2-1) associated pathways. In the resent study, we hypothesize that tobacca products may induce micoenviron- rment alterations, promoting angiogenes’s and providing a permissive environment for head and neck cancer pro- fression. In an in vitto anahss, we employed immortalized oral Keratinacyte (HOX-168) an laymngeal squamous ‘arcroma (UMSCC-1 1A) cells to investigate interleukin ()-8 and vascular endothelial growth factor (VEGA, induction by cigarette smoke condensate (CSC). IL-8 and VEGF expression is based on interactions between NFB, AP-1, ana 5, We identified at least 1.5-fold dose-dependent induction of AP-1, VEGF, and IL-8 promoterreporter gene actviy after 24 h exposure to CSC. Next, we stably transfected UM-SCC-T1A ces with A-Fos, 2 dominant negative [AP-1 protein. Treatment with CSC of the A-fos cell ines compared to empty vector contrals significantly down- regulated AP-1, VEGF, and IL-8 promoterreporter gene expression. We abs performed ELISAs and clscovered signif- cant up-regulation of iL-8 and VEGF secretion by UMSCC 11A after treatment with phorbol 12-ryristate 13-acetate, tumor necrosis factor alpha, and CSC, which was down-regulated by the A-Fos dominant negative protein, We con- clude tobacco carcinogens up-regulate AP-1 activity and AP-1 dependent I-8 and VEGF gene expression in head ard neck cancer. This up-teguation may promote an angiogenic phenotype favoring invasion in both premalignant anc squamous cancer cell ofthe head an neck. © 2011 WikrLs, he 5: head and neck cancer; activator protein 1; cigarette smoke condensate; tobacco carcinogens: VEGF: IL-8 INTRODUCTION Bach year over 40,000 people in the United States are diagnosed with head and neck cancer and tobacco andjor alcohol use is identified in lover 75% of cases (1-5). The cure rates for head and neck cancer have not improved over several decades and voids exist in the understanding of (01)-5 and IL-8 expression when NFKB is targeted and decrease vascular endothelial growth factor (VEGE) expression when AP-1 is targeted [14,17] IL-8 and VEGF are implicated in tumor pro- {gression via their contubutions to angiogenesis, a process which allows autonomous growth of tumors once they achieve a size of approximately signaling events during tobacco-induced carcino- genesis and progression [1,6]. In head and neck ‘cancer, constitutive up-regulation of eatly response genes activator protein 1 (AP-1) and nuclear factor Kappa B (NFB) controls a number of celular proc- estes associated with malignant progression from 3 pre-cancerous to metastatic phenotype [7-15]. For example, both pro-angiogenic and pro-inflamma- tory cytokines are significantly up-regulated in murine squamous carcinogenesis models of head and neck cancer during tumor growth and meta- static spread [16]. Several studies aso indicate pro- Angiogenic cytokine expression in head and neck ‘cancer cel lines is NFB dependent when AP-1 is also constitutively activated. In these studies, dominant negative and other gene knockdown approaches can significantly decrease interleukin 2 2017 WHEFSS, Ne 3 mm (18-23). 1¢ is well established that IL is 3 ‘gene controlled by cooperative interactions in the Promoter between transcription factors and NFxB, AP-1 and nuclear factor (NF)IL6 elements, as ILS, grote, eri, Ve, vaseereneereal grec ocr Scr Gguete smoke condensate 0-1, atta poten He ‘a, hypots-rauctle fscor tes DMSO, met sulfone, PA, Bhowbol12-mynsate Isacetate, TNFa Turor neces factor "Corespondence te: Melear Oncology Program, Deparment na 270, sinreatos WN 33435. Maren 2011 , redone? Xgl 2011 Wiley Online LbatyToRACcO 8 secretion from cells in the microenvironment appears maximized when these transcriptional events are switched on [14,16,24]. VEGF has been. primarily identified as a hypoxia-inducible factor GUE) 1a and AP-1 dependent gene, and can be ‘maximally stimulated dusing periods of low oxy en tension in the tumor microenvironment when. THF-tis activated [25-29]. Both of these cytokines contribute to angiogenesis and are found in the milieu of head and neck cancer specimens as well as in oral cavity fuids of patients afficted with head and neck carcinomas (30-32) Tobacco carcinogens are principally genotoxic [83] but tobacco use is also highly associated with infammatory conditions of the oral cavity and. gingiva (34-38). The pro-inflammatory effects of tobacco carcinogens require further investigation to yield a better understanding of head and neck carcinogenesis. A. growing body of evidence indicates approximately 159% of the world’s malig- nancies are associated with chronic inflammation, [39,40], In head and neck carcinogenesis, chronic indammation Js an accompanying feature of dys- plastic lesions and at least 75% of leukoplakia lesions are associated with smoking [4144], Tn the present study, we examined influences of cigarette smoke condensate (CSC) on AP-1 depend ‘ent activation of IL-8 and VEGF in vitro. We found CSC could significantly stimulate AP-1 activation fof both genes, resulting in increased ILS and VEGF sectetion, and that these processes could be down-regulated with introduction of « dominant negative A-Fos gene, These data demonstrate a role for tobacco carcinogen stimulation of pro-angio- ‘genic cytokines, thus promoting an environment suitable for development and metastatic spread of hhead and neck cancer cell. [MATERIALS AND METHODS Materials To stimulate AP-1, cells were treated with phor- bol 12-myristate 13-acetate (PMA) (Sigm2-Aldrich, St. Louis, MO), a diester of phorbol which activates the signal transduction enzyme protein kinase C, tumor necrosis factor alpha (TNFa) (Promega, Madison, WD, or CSC (Murty Pharmaceuticals, Lexington, KY) solubilized in dimethyl sulfoxide (DMSO). The CSC is a standardized extract and is prepared by smoking University of Kentucky’s Standard Research Cigarettes on an FTC Smoke ‘Machine, The total particulate matter (IPM) on the filter is calculated by the weight gain of the filter. rom the TPM, the amount of DMSO to be used for extraction to prepare a theoretical 4% (40 mg/ml) solution is calculated. The condensate 4s extracted with DMSO by soaking and sonication, ‘then packaged 1 ml/vial in dry vials, Stock CSC, stored at 807°C, was thawed and added to ceil Moleculy Corcinogeness 1 WN HEAD AND NECK CARCINOGENESIS 669 growth media using serial dilutions to create treat- ent media of desired concentrations. Control cells were treated with medium containing an equivalent amount of DMSO. Repeated freezing and thawing of CSC solution was avoided as much as possible. The AP-1 reporter gene plasmid (Agilent, Santa Clara, CA) expresses the firefly luciferase gene ‘under the conttol of a synthetic promoter contain= ing seven direct repeats of the AP-I transcription factor binding sequence, The VEGF reporter gene plasmid (2 gift from Amit Maity, University of Pennsylvania) contains 1.5 kb of the VEGF pro- moter in the pGL-3Basic vector [45]. The IL-8 reporter gene plasmid (from Dr. Naofurm! Mukaida, Cancer Reseatch Institute, Kanazawa University, Japan) contains 133 bp of the IL-8 promoter in a firefly luciferase reporter vector (14] MerHoos coll Culture ‘The HOK-168, HPV immortalized human oral keratinocytes, were a kind gift ftom NH. Park at UCLA. These cells were grown in KGM-2 Medium, (Lonza, Walkersville, MD) supplemented with bovine pituitary extract, epidermal grovth factor, insulin and hydrocortisone at 37°C in 5% COs UMSCC+, 114, IIB, and 38 are head and neck squamous cell carcinoma cell Lines, received from ‘Thomas Carey, University of Michigan. These cells ‘were grown in minimum essential medium (MEM) supplemented with 10% fetal bovine serum (FBS) at 37°C in 5% COs UM-SCC9 was harvested from 2 tonsiifoase of tongue primary tumor, UM-SCC- IA was harvested from a primary laryngeal cancer biopsy, UM-SCC-11B was derived from the laryng- ectomy specimen of the same patient after chemo- therapy, and UM-SCC-38 was harvested from a tonsil/base of tongue primary tumor. NA and CA 9.22 are oral squamous carcinoma cell ines grown. in RPMI supplemented with 10% FBS at 37°C in 5% CO; Where indicated, cells were grown in Serum free media. The concentration of DMSO, in all cultures, including as a solvent control, was <0.1%, All nes were mycoplasma free by PCR testing. HUVEC cells were obtained {tom Invitro- gen (Carlsbad, CA) and were grown in Medium 200 (Invitrogen) supplemented with 296 FBS, 1 wg/ ml hydrocortisone, 10 ng/ml human epidermal {growth factor, 3g/ml baste fibroblast growth fac- tor, and 10 ug/ml heparin at 37°C in 5% CO>. Reporter Gene Assay ‘The cell Lines were plated at 40,000 cells/well in i2-well plates and wansiently co-transfected via ‘TransIT LT1 (MirusBio, Madison, WD) with either AP-1, VEGE, or IL-8 reporter gene plasmids 24 h later along with a pCMV Lac-Z reporter containing670 SUIENSON ETAL the CMV promoter and Lae-Z gene in peDNA3 to adjust for transfection efficiency. After ovemight transfection, cells were treated for 24h with DMSO (solvent contro), PMA, TNFa, or CSC in serum free media, Cell lysates were analyzed via Tropix Dual Light Reporter Gene Assay (Applied Biosystems, Carlsbad, CA) on a Tristar dual injec- tion flash luminometer (Berthold Technologies, ‘Oak Ridge, TN). Nine replicates were measured per data point and each experiment was repeated in triplicate UM-SCC-11A cells were stably transfected with a dominant negative AP-I plasmid (A-Fos), a gift from Charles Vinson (Laboratory of Metabolism, Center for Cancer Research, NCI/NIH), and selec” tion performed with neomycin. The A-Fos protein has been extensively studied in the past. Tt con- tains an acidic amphipathic protein sequence appended onto the N-terminus of the Fos leucine zipper. This acidic extension interacts with the Jun basic region and extends the leucine zipper which prevents a JUN/A-fos heterodimer from interacting With AP-1 consensus sequences [46,47] HOK-L6B cells were co-transfected with A-Fos and pCDNA3.L-hygro and selected with hygromy- cin, as they are neomycin resistant from HPV transfection. Polyclonal populations were used for experiments EUSA UMSCC-IIA and HOK-168 cells were plated in. 25 cm flasks and treated with DMSO, PMA, TNFa, for increasing CSC concentrations in serum free ‘media, After 24 h, media was collected and assayed for total IL-8 and VEGF using the DuoSet ELISA Development Kit for each analyte (R&D Systems, Minneapolis, MN). Assays were carried out accord ing to the manufacturers instructions and were run in triplicate. All experiments were independ- ently repeated two to Uiree times ‘Anglogenesis Microtubule Assay Cayopreserved HUVEC cells were seeded at 2x 10° viable cells in a 75 cm ussue culture flask using LSGS-supplemented Medium 200. Culture ‘medium was changed every other day until the culture was approximately 80% confluent. A 24- well tissue culture plate was coated with Geltrex (nvitrogen) and incubated at 37°C, allowing solid ification. HUVEC cells were harvested via trypsini- ation and centrifugation. The cell pellet was resuspended with non-supplemented Medium 200, Cells were mixed with conditioned media from HOK.168 cells treated with varying concentrations fof CSC (supematants were treated and collected precisely as described in the ELISA methods above) Molec Carcinogenesis to a concentration of 8 x 10" cells per 400 wl and added to the coated wells, The 24-well plate was incubated at 37°C, 5% CO, for 4h. To quantitate microtubule formation, capillary tube branch points were counted in’ two 100% fields per well (two replicate wells per Weatment for a total of our replicates). Western Blt UMSCC-LIA and HOK-16B cells were plated in. 75 cm? asks and tweated the following day with DMSO, PMA or 30 g/ml CSC in serum free media, ‘Nuclear extracts svere prepared using the NE-PER nuclear extract kit (Pierce Biotechnologies, Rock- ford, IL). Protein quantification was performed using the bicinchoninic acid assay (Pierce). Ten ‘micrograms of nuclear protein per lane was separ ated on 2 4-12% BisTris NuPage Gel (Invitrogen, Carlsbad, CA), Proteins were transferred (o Immo- bilon-PSQ 0 2. um PYDF membrane (Millipore, Bill- erica, MA). Detection was performed using the SNAP LD. (Millipore). Membranes were blocked ‘with 0.59% milk im TBS/0.1% Tween-20 then blot- ted with a primary antibodies to AP-1 family mem- bets (AP-1 Family Antibody Screening Kit, Active ‘Moti, Carlsbad, CA) individually followed by a horseradish peroxidase conjugated anti-rabbit sec- ‘ondary antibody Jackson ImmunoResearch, West Grove, PA). Proteins were visualized with a chemi- luminescence assay system (Pierce), Membranes were stripped and re-probed with each of the screening kit members and histone H3 (Cell Signaling Technologies, Danvers, MA). Statistical Analysis Differences in reporter gene activity, released otein levels, and capillary tube branch point ‘counts were compared by Student T-test. P-values fof <0.05 were considered statistically significant, GraphPad Prism version 4,00 for Windows was used for these analyses, (GraphPad Software, San Diego, CA) RESULTS Pro-angiogenic cytokines such as IL-8 and VEGF are downstream from a variety of transcription fac- (ots Specifically, IL-8 has been shown to be under ppincipal conttol of AP-1, NF«B, and NF-IL6 [48— 51]. VEGF is under great influence by low oxygen conditions stimulating the transcription factor HF-1a, but is also under significant control by AP- 1 (52,S3). Presently, we concentrated on the AP-1 associated activities of both genes as contributed to by CSC. To establish the baseline levels of AP-1 ‘ot VEGF activity in head and neck squamous cell carcinoma, we fist investigated the degree of acti- vation of AP-I dependent reporter genes via luci- forase assay. Twenty-four hours after transfectionTopACco & RLU VEGF RLU 00 08 40 1 WN HEAD AND NECK CARCINOGENESIS or B 2 a0 2 20 1 RLU ts 20 25 30 35 AP-4 RLU serv celine ese ad nec 2 rose coro rst ‘he fet or VEG? perotettucace vere Bs er and HONE cals were nse vaneced ds A At-t°f) VEC cansttve sero” ache eal gorse pals vn Grphfad fom eve ste conelton serween conse promote acivbes "= Dee Pe O.025) ALL rave Mg uns ‘of seven cell lines with AP-1 luciferase promoter/ reporter plasmids, we assayed baseline transcrip- tional AP-1 activity (Figure 1A). We observed a 1- fold difference in activation between the highest land lowest activity of cell lines tested. Interest- ingly, the HPV-transfected premalignant oral kera- Unocytes (HOK-168 cells) had the greatest degree fof AP-1 activity compared to the other cell lines ‘The invasive carcinoma cell lines had a 13-fold difference in baseline AP-1 activation between the highest and lovrest cell line activities, Next, wwe repeated the experiment using the ‘VEGF reporter plasmid. We again identified a sig- nificant degree of baseline activation of the VEGF reporter gene, which was highest in the HOK-16B premalignant Keratinocytes compared to. the ‘malignant cells (Figure 1B). As VEGF is down. stream of AP-1, we sought to establish the degree Cf correlation between baseline AP-I reporter gene activity and VEGF reporter gene activity. We per- formed linear regression correlation analysis of the Dateline activities of both genes. We identified a high level of correlation (7 = 0.8616, P = 0.0025) Moleculy Corcinogeness between AP-1 and VEGF baseline gene activation in head and neck cancer cell lines, as judged by luciferase reporter gene activity (Figure 10) In order to further demonstrate whether pro- angiogenic cytokines are under AP-1 control, we established stably-transfected cell lines utilizing the dominant negative A-Fos construct in both the HOK-16B and UM-SCC-I1A cell lines, Once these cell populations contained either the empty vector control or the dominant negative A-fos, we com- pared the levels of AP-1, VEGE, and IL-8 promoter activity in each cell line. First, we identified con- stitutive AP-1 promoter activity was significantly decreased by at least 2.3-fold in both cell ines after insertion of dominant negative A-Fos (Figure 2A) (@ < 0.0001). Next, we assayed the level of VEGF or IL activation in empty vector control or A-Fos cell lines. After the insertion of A-Fos, we found ‘he level of constitutive VEGF activation was sup- pressed by 7.6-fold in HOK-16B cells, and by 7.1- fold in the UM-SCC-11A cell lines (Figure 2B) ( < 0.0001). Similar findings were identified for TL luciferase activation after A-Fos insertion into672 SVIENSON ETAL Z “cm RLU, Bo. aa a L. ce ‘. ° oe oki Figue 2, Dornan negative AP (Fs) decreases 82 VEGF, and tot promoter atty, HOKE sea Mee aia cis sabysepresing ery veer of Fos vere vatse nly wartece vith At a) VEGF forlueferace aciviy. Sh eines deorstates sgavcan: decease % AP, VESE, and l-B nth addon 9 ‘os shou ini vac of = 0.000 ‘each cell line (Figure 2C) (P < 0.0001). These find: ings, taken together, demonstrate that both pro- angiogenic cytokines have significant activation, ‘which is contaibuted to by constitutive AP-1 activity fn head and neck squamous carcinoma cell Lines. BLISA was utilized to further evaluate the relationship between AP-1 activation and down- stream IL-8 and VEGF expression in HOK.16B and. UMSCC-IIA cell lines, Levels of constitutive IL-8 secretion were significantly decreased in each cell line after insertion of dominant negative A-Fos, In addition, treatment of cells with 10 nM PMA or 10ng/ml TNFa as positive controls induced Inigh levels of ILS sectetion which were largely diminished with A‘Fos insertion (Figure 3A,C) (P < 0.05). Constitutive VEGF secretion vas sim flatly decreased in both cell lines by insertion of A- Fos. PMA and TNFa treatments alto caused increased secretion of VEGF in both cell lines, Which was significantly decreased in cells stably, expressing A-Fos (Figure 3B,D) (P < 0.05). CSC has recently been shown to up-regulate NF«B activity in. transformed oral keratinocytes and squamous cell carcinoma [15]. We hypathes: fzed CSC could also increase AP-I activity in these cell types. Treatment of the HOK-165 cell ine with, ESC demonstrated a dose-dependent increase in AD-L promoter activity, with as much as a L6-fold ‘merease in AP-1 after Weatment with 60 ng/ml Mole Carcinogenesis SC (Figure 4A) (P< 0.0001), AP-1 promoter activity in the UMSCC-11A cell line was also up- regulated in a dose-dependent fashion with CSC lueatment, demonstrating a peak 2.7-fold increase in acuvity with 60 g/ml CSC teatment (Figure 4D) (P < 0.0002), We then examined VEGF and IL-8 promoter activities in response to CSC in transformed oral keratinocytes, Similar to AP-1, we demonstrated a dose-dependent increase in these promoter activities after CSC. treatment (igure 48,0) (P< 0.05). In the UM-SCC-11A cell line, VEGF and IL-8 promoter activities were also significantly increased with CSC treatment, ‘again in 2 dose-dependent fashion (Figure 4E,F) (F = 0.0001) Next, we sought to determine whether the domi ant negative (DN) A-os plasmid could affect AP-1, VEGE, or IL-8 promoter activities in response to CSC treatment, We chose to treat HOK-16B and UMSCC-11A cells with 30 ugiml CSC, based on four previous study [15]. In the presence of A-Fos; AP-1, VEGF, and IL-8 promoter activities for HOK.16B were decreased by 1.7-f0ld, 5.9-fold, and. 2,2-fold, respectively (Figure SAC) (P < 0.05). The UM-SCC-I1A cell line expressing A-Fos showed reduction in AP-1 by 2.3old, VEGF by S.8-fold, and IL-8 by 19-fold (Figure 3D-) (P = 0.0001) ‘Others have demonstrated that dominant negative AP-1 insertion in mouse keratinocytes canTopACco & IL-8 (pgimLi24h/1x108cells) > o 8 8% 8s IL-8 (pgimL2an/txt0%cells) © [ Pu Nee 1 WN HEAD AND NECK CARCINOGENESIS B 2 2 Bo : £ 2 é g 673 3 3 ae 3 VEGF (pg/mL24hitxt0%cetls) F eure 3. Afos decreases release of proargagencctlines in response to PMA ar TNFa, HOKC168 and UMSCE-T1A cle sts exoresng Atos o ely vec were ested fo" 26 nr To Ml MiA 0 10.na ‘hfs ondwayed for alae elese vie 5A. tod moe 1 65) Vem NORA, C) Loe m UNECE Tot ce es demons agree dian of dom’ negitve APT fos, show swe Plu ot = O08 Tia ana Wer n UnSce coordinately down-regulate NF«B activity as well [54,55]. We tested whether A-Fos would act in a similar fashion in the oral cancer cells. However, ‘we found only 0.1-0.2 fold increases in NF«B luci- ferase activity when this was tested in UM-SCC- IA cells (not shown). This is far less than effects observed with TAMG7 dominant negative insertion in murine cells. We then wanted to further delineate the relationship between CSC-mediated secretion of pro-angiogenic cytokines and their dependence ‘upon AP-1 activity, HOK-16B and UM-SCC-I1A cells stably expressing A-Fos or empty vector were treated with 30 ygiml CSC for 24 h, Cultured cell supernatants were collected and assayed for cyto- Kine release via ELISA. In both HPV-transformed premalignant keratinocytes and invasive squamous carcinoms, up-tegulation of IL-8 and VEGF secretion via CSC treatment was largely decreased, with insertion of A-Fos (Figure 6-D) (P < 0.05), Next, we wanted to investigate whether super- natants from CSC treated A-Fos cells would func tionally down-regulate angiogenesis, We utilized conditioned media from either A-fos or empty vec tor transfected HOK-16B cells treated with CSC for Moleuly Corcinogene en grotne rele wth ‘rahe t= S60 24h and examined endothelial ube formation, as a standard measure of angiogenesis. We observed ‘that supernatents from CSC treated EV HOK-16 B cells could stimulate endothelial tube formation by up to 61% compared to CSC treated A-Fos cells ( = 0.0003 for 60 ug/ml CSC and P = 0.0019 for 30 ugiml CSC) (Figure 68). These data, taken together demonstrate that CSC has the capacity to stimulate both proangiogenic cytokine secretion and angiogenesis in premalignant oral cavity (HOK-16B EV cells) when AP-1 is intact compared to when the Adfos dominant negative is transfected, ‘Next, we examined the nuclear levels of two AP- 1 family members, efor and cjun, to determine if expression of the dominant negative A-fos affected expression in an autocrine manner, Cells were Uweated with DMSO, 10 ng/ml TNFa, or 30 g/l CSC, nuclear extracts isolated and separated via PAGE, Western blot was performed with antibodies to fos, cjun, and histone 13. In the HOK-16B (igure 7A), we found cfos expression was visibly decreased with addition of the dominant negative AJfos. The cjun expression decreased when the cells were treated with DMSO or CSC. An increase674 SWENSON ET AL gree ote Content) Spat ate and ht) E F t a : = i a 8 ss Clowes ae Cnt) nt} Crete tet Cont et Flaws &, CSC upregubtesAPt, VEGE an 18 promatey acti ip HOK-16B and UMESCC-11A cls. HOK 160 and UMeSecovin els vere Cans wanstetes with Abt, WEG, or ted prometerreports” Has {nd undernet 24h tenant with vung concent sons of CSC: Celt were hares ed and acre ‘nay perfor, Promoter ste: wen aon (8) Bo 8) VEGE ana fC) hod m HOR 2.0) APs) Weel and tn NISCETIN Se ante peegbd AP VEG promoter acts o's [Sse ependens manner for Snel ines Phe of < 0.5 coma, "Pslye ot # O0S01 cant A, B . c . re a" ge : Uc [] sie) Bie ro Se See, See D E F tem 7 2 | ee BO Clem tn kets) Cyne btn Coe Ss nest Flgue 5. aos decreases AP, VEG, aed Ie promoter acl in tsponse to CSC eabnent HOK- 168 ane UMSCC118 ces Sas epresing AZos or empty vector vere arin nstcted ty AP, EGE. oF CE brareneore purse 24m vexsrent wh 30 wg CS” mata siya sale va ‘epoter cone my. araer aves ae a0 flaw, a) P= 8) WEG, and) Hm NOR) AP (Wives sna fe eg in UMSCCIA, APA, VEGR, ana Ls promorer acy was sanieary Seceased Seth celines wih aden of os shown nwt aie = Gos ws cal “Rake of © 000 Molecule CorcinogenessTOBACCO & APN HEAD AND NECK CARCINOGENESIS 675 B 2 zo 3 3 S son oe as : § sos 3 Z : z z Eo 2 = 8 Zo - go ° % s ° % Cigarette Smoke Condensate (\.gimL) Cigarette Smoke Condensate (g/mL) c D BE 6000 F 2000. 3 B 500 . e & tt00- E «000 = z & 000 3 Z . Z Z 2 = B00 : son : = 3 5 Zo go ° % S ° % Cigarette Smoke Condensate (igimL) Cigarette Smoke Condensate (g/mL) Ee :° Ee = © § 2 3 © 2 Cigarette Smoke Condensate (ugimL) EME ofr deca rae of proangiogenic ok n report 0 CSC. HOR IGE nd MLSE reese FSR Isa ora) Lo aed) Wen HORT: hed ad EGE n UNAS 1A Btn ce deren sghfeanedececse n cytokine felaze wt son of Ares, shawn nah (Paalve Snrdotrea ose fomatonaeraysncressedy 6c? 0.0003) aren comparing eet of consoned iPeas rom 60 agin Coc eaed ey vec’ cool WOK IB® cls to Nios NOK cals, An nese ot {in expression was seen with TNFe treatment. Also addition of the dominant negative, A-fos, In the in the HOK-16B, the antibody to Histone H3 non- UMSCC-1IA, the particular antibody to cjun swe specifically bound to a slightly smaller protein, employed also bound nonspecifically to a protein UMSCC-11A igure 7B) experienced a visible marginally smaller than cjun, the lower band decrease in fos and cfun expression with the shown in the cjun panel in Figure 7A, Moleuly Carcinogenesis678 SUIENSON ETAL Histone Hs Ete iS REET APES cng aoe ‘Reo ect es wah Sg WPS pga ES Bot ii I SEIT Sea in emartsess fds n eos anid un expresion vith DISCUSSION In the present study, we investigated the role of AP-L in the production of a pro-inflammatory and pro-angiogenic cellular microenvironment in pre- ‘malignant oral keratinocytes and HNSCC cell lines. We studied the importance of AP-1 at base- line conditions as well as in the presence of CSC, a known stimulator of cellular carcinogenesis, AP-1 is composed of a family of dimeric basic region leucine zipper (bZIP) proteins, which regulate cel~ lular proliferation, transformation, and death (56) A strong connection between AP-I and chronic inflammation has been established in colon, breast, and prostate cancer [57-59]. Cytokines such, as L$ and VEGF permit a pro-tumorigenie cellular environment via propagation of inflammation and angiogenesis. Accordingly, we frst evaluated the potential association between constitutive AP-1 and VEGF promoter activities in multiple head and neck cancer cell lines. Results demonstrate the level of VEGF activation at baseline is highly corre- od with the degree of AP-1 activation in the cell lines tested Figure 1). These data are consistent ‘with published literature demonstrating that VEGF relies on AP-1 as a principal contral mechantsi in cancer [25-29] ‘We hypothesized inhibition of AP-1 activity in premalignant keratinocytes and invasive carci- roma cell lines would yield markedly decreased transcriptional and translational activities of VEGE and ILS, which are implicated in the promotion ff cellular inammation and angiogenesis. We Molec Carcinogenesis Utilized a dominant negative to AP-1, which inhib- its FosJun heterodimer binding to DNA in equi ‘molar competition, This dominant negative, termed A-Fos, is more potent than other previously constructed DNS to AP-1, which carried a deletion ff the Fos or Jun transactivation domains [46]. Our present investigation is the first to utilize A-Fos to study the importance of AP-1 activity in the con- text of HNSCC. We established the level of AP-1 activation in head and neck cancer cell lines is associated with the levels of both VEGE and ILS activation in HLV-transéormed oral keratinocytes fand head and neck cancer cell lines, as introduc- tUon of A‘os resulted in a proportional decrease in VEGF and IL-8 promoter activity compared to AP-1 (Figure 2). This suggests that both genes are AP-1 dependent to a significant degree in head and neck ‘cancer cell lines. Further, data shown in Figure 3 supports the fact that IL-S and VEGF secretion in hhead and neck cancer cel lines is highly depend- fent upon the activation of APA, Thus, AP-1 appears to play a critical role in the promotion of a pro-inflammatory environment in both HPV-trans- formed premalignant oral keratinocytes and HINSCC cell lines. And since chronic indammation is often a central process in carcinogenesis, aber- rant increases in AP-1 may be a primary mechan- ism for tumorigenesis in this setting. As cigarette smoking is a significant risk factor in the development of head and neck cancer and demonstrates pro-inflammatory properties, We sought to characterize the associations between acute CSC exposure and AP-1 activity in HPV. ‘transformed oral Keratinocytes and HNSCC cell lines, We found, in both cell lines, CSC could sig- nificantly stimulate AP-1 activation above constit- tusive levels. A proportional increase in VEGF and IL-8 transcriptional activity was also demonstrated. in time- and dose-dependent manners (Figure 4). Cytokine sectetion of VEGF and IL-8 were also up- regulated with exposure to CSC (Figure 6), We again utilized A-Fos to demonstrate inhibition of AAP-I activity abrogated the vehement pro-inflam- matory response initiated by CSC (Gigures § and 6). These findings suggest AP-1 plays a signifi- ‘ant role in the heightened proinflammatory response that occurs with CSC exposure through, downstream activation of VEGF and ILS, Tn a subset of head and neck cancers, HPV infec tion has been identi, predominantly the high risk subtype HPV-16, Thus, oral HPV infection is tor for HINSCC [60]. Asa virus, LIPV does not produce its own transcription factors land relies on those provided by the host cell. OF ‘great interest, it has been shown that AP-1 is necessary for E6 and E7 oncogene expression in HPV-16 and HPV-18 infected cells (61,62). Our pre- ‘malignant oral Keratinocyte cell line, HOK-16B, has been immortalized with HPV-16, Initially, weTOBACCO & APN HEAD AND NECK CARCINOGENESIS a7 ‘expected to see markedly increased AP-1 levels and downstream pro-inflammatory cytokines in the invasive carcinoma cell line, as compared to the premalignant oral keratinocytes. In acwality, the HPV-transformed cell line had far greater AP-1 pro- moter activity than the HNSCC line, both in the presence of absence of CSC treatment (Figures 2 and 4), Although IL-$ and VEGF activity was greater in the IIA cell line compared to HOK-168, ‘the expression variance between premalignant and, ‘malignant cells was less than expected (Figures 3 and 6). Itis possible that in HOK-168 cells, AP-1 is being utilized to augment viral replication, with less AP-L available to bind the IL-8 and VEGE pro- ‘moters. In the setting of CSC exposure, we do see ‘multfold increases in AP-1 for HPV-transformed. ‘oral keratinocytes. This suggests that HPV and CSC. may work synergistically to mount a strong pro- inflammatory response and promote malignant transformation, as IIPV needs AP-1 to replicate land CSC provides an increased amount of AP-1 We have identified the potential importance of AP-1 in the promotion of a pro-inflammatory cel- lular environment through IL-8 and VEGF expres- sion in HPV-transformed oral. keratinocytes. and. HINSCC cell lines. The pro-inflammatory role of AP-1 is particularly heightened in the presence of CSC. ‘This information is valuable since AP-1 is considered a strong target for chemoprevention, ‘Agents such a8 resveratrol, curcumin, epigallocate- chin gallate, quercitin, and caffeic acid phenethyl fester all tnlibit AP-1 activation to varying degrees [63], Our results further explain why these agents may be effective inhibitors of tumorigenesis of the hhead and neck, particularly in the setting of tobacco or HPV exposure. ACKNOWLEDGMENTS This work is supported by the Lions SM Transla- tional Biomarker Initiatives for Medical Students (GO), Minnesota Medical Foundation and the ‘Office of Clinical Research Clinical Translational Research Program, University of Minnesota (WGS), P30 CA775%8-07 NIH/NCI (FO), REFERENCES: 1. Greenlee, Hlksarmon MB, Murray, Thun M, Cancer statsts, 2001. CA-cance} Clin 00%;51:15-36. 2, Cho SY, Kayo H. efect of igatelesmoking and lcahocansumtion i the etilegy of cancer of te oral uty phar ane rym Int) Epceril 198;20.878— as, 3, Mashoerg A, Bollea P, Winkelman &, Garfinkel L Tobacco smoking soho! dinking, ad cancer ofthe oh ‘uty and orophanya among. UnteaStvesvevans (Cancer 1985:72 13881375. 4. Boffeta ?, Mashberg) A Winkelmare A, Garfinkel Corcinageseie effet of iobacea snk, Sra alkohol Ginting on. sratemie. ster of the cal civ ane ‘ropharyne Int) Cancer 1982,82 5305833, Molecubr Carcinogenesis 5 Bho WI, MeLaughlin JK, Winn OM, etal. Smoking ana Grn in elation fo orl and phargeal cancer. Concer es 1988 48:5282~3207. 6. Canatho AL, NsSimotd IN, Califano JA, Kowal Trends mn intidence and proppsis for head ad neck Cancer m the Unted States Astecpece anaes of the SEER database nt) Cancer 2005,114 806-816, 7. Matiangura A, Supiasnan P. Trvekapan S, et al Telomerase acy tra leukonikls Sr heae nd neck Sguamous call carcinoma, Cancer Res 1996,56:2530— 3533 8. Calfine J, Abvent $A, Meininger G, West Wi, Koch Wis, Sieransiy D, Detecton of telomerase atv in eal ses tom ead ang neck souamous cal man oneeytc cal ine, THP-T, through acting concurrent om Boe and -coppa B-bring ses of the tele d gene) Vel 199771 5692-5695. 52, TucrerE, Mehl Hartman T, eta. The human gene for vase enaotnetal rawch Taco. Nubile pote forms are encodes twough akeratve exon spleng Nol chem 19912661 941-1958. 53, Diaz BY, Lena MC, Ladowx A, Flin C, Demarchez My IMnel Regulation of vascubr endotelal gah factor feapresion in human Keratinocytes by etnalds. J Bol Chem 2000275 642-830. 54 UAnim 15, Senge R, Vousden KH, Colburn NH Expression ‘of ‘dominant negative Jun inhibits elevated POA” and Nekapgad ttansactiaton and suppresses anchorage insependent growin of HPV immortalzed human keratinocytes. Oncogene 1998;16:2711-2721 55. U1), cao ¥. Young MR, Ceburn NH Inuced expression oF domnanenegatve cyan downregustes NHopa8 Sand AST target genes and suppresses turer phenetype inhuman kerotnacjtes. Mol Carcinog 2000,29:'58 ts. 536. Shaulan F, Karin M. M7 a6 a regulator of ce Ife and death Nat Coll Biol 2002;42 131-136, 57. Ness RE, Medugno f Erdometiosis as a mavel for Inflammatinshormone interacrore in evanan ane beast ‘Cancers. Fur) Concer 2008 22 681-703 58. von Kempen LCL, de Vise" KE, Cousens LM. Inlrm Tato, proteases ans cancer Eu) Cancer 2006,42:728— a,TOBACCO & AP-1 IN HEAD AND NECK CAROINOGENES'S 679 58, De Morz0 AM, DeWeese T, Pate A, e a Pathlogal nd molecular rechansms of prostate carcnogeness: Impliations. for" eagnoss, detection, prevention, and trentreent J Cel achem 2004 81 259-477 50, Kremer AR, Ciford GW, Boyle P. Franceschi 5. Human Papllomentus types in head and eck squamous cel Eaanomas word A systematic revev, Cancer Ep ‘dernolSomarc Prev 2005 14 46)~475, G1, Oifor EX, Beard P. A member ote aclvaterproter-1 family found 9 keratnazytes But nin Troost: Molecubr Carcinogenesis 2. required for transcption ftom » humo paoilomavius ‘goes romorer,) Viol 1990)544792-4798, Chan W, Chong T, Bernard HU, Klock G. Transcription Gt the tanstorming genes. cf the orcagene human Papllomavnuste. is stmabted oy tumor promoters Extough APT binding Stes. Nelle Acide Res" 1990, 18 762-769. Marines CP, Colburn NH, Young MR. AP- a 3rget for Career prevention, Cur Canees Drag Tages 2007.77 Sau