Tobacco exposure results in increased E6 and E7 oncogene expression, DNA damage and mutation rates in cells maintaining episomal human papillomavirus 16 genomes

Summary

Tobacco smoking is an apparent cofactor with HPVs in cervical malignancies, but the interactions are poorly understood. We modeled tobacco smoke exposure to cervical epithelium; data suggest that tobacco smoke contributes to early stages of HPV-related cancer progression.

Abstract

High-risk human papillomavirus (HR-HPV) infections are necessary but insufficient agents of cervical and other epithelial cancers. Epidemiological studies support a causal, but ill-defined, relationship between tobacco smoking and cervical malignancies. In this study, we used mainstream tobacco smoke condensate (MSTS-C) treatments of cervical cell lines that maintain either episomal or integrated HPV16 or HPV31 genomes to model tobacco smoke exposure to the cervical epithelium of the smoker. MSTS-C exposure caused a dose-dependent increase in viral genome replication and correspondingly higher early gene transcription in cells with episomal HPV genomes. However, MSTS-C exposure in cells with integrated HR-HPV genomes had no effect on genome copy number or early gene transcription. In cells with episomal HPV genomes, the MSTS-C-induced increases in E6 oncogene transcription led to decreased p53 protein levels and activity. As expected from loss of p53 activity in tobacco-exposed cells, DNA strand breaks were significantly higher but apoptosis was minimal compared with cells containing integrated viral genomes. Furthermore, DNA mutation frequencies were higher in surviving cells with HPV episomes. These findings provide increased understanding of tobacco smoke exposure risk in HPV infection and indicate tobacco smoking acts more directly to alter HR-HPV oncogene expression in cells that maintain episomal viral genomes. This suggests a more prominent role for tobacco smoke in earlier stages of HPV-related cancer progression.

Introduction

Cervical cancer is one of the most common cancers in women worldwide, with >0.5 M new cases and nearly 275 000 deaths among females annually (1). The causal relationship between high-risk human papillomavirus (HR-HPV) infection and cervical cancer is well documented in epidemiological and functional studies, with detection of HR-HPVs in up to 99.7% of cervical malignancies (2). The HR-HPV E6 and E7 oncoproteins are expressed during and after cancer progression and contribute to cervical carcinogenesis in part by inactivating the cellular tumor suppressor proteins p53 and pRb, respectively (3). However, HPV infection alone is insufficient for cervical cancer development. An estimated 80% of women will acquire an HPV infection during their lifetime, but most infections are transient, with only a minority resulting in recognizable cervical cancer (4). Therefore, additional cofactors are required for development of cervical cancer.

Tobacco smoking exposure is associated with multiple cancers (5,6). The International Agency for Research on Cancer has classified tobacco smoking as a cause of cervical cancer (7). It is estimated that 11.8% of cervical cancer deaths are attributable to smoking. Smoking has been consistently linked with the progression of cervical neoplasia, and female smokers have up to two times higher risk of developing cervical cancer than non-smokers (8). Previous studies have focused on the impact of tobacco smoke on the prevalence (9,10), incidence (11–13) and persistence of HPV infections (14–18). Tobacco smoke contact includes mainstream tobacco smoke (MSTS) and side-stream tobacco smoke. MSTS refers to the exposure gained when a smoker inhales directly from the tobacco source, whereas side-stream tobacco smoke is that inhaled from the distal lit end of a cigarette, cigar or pipe. Both MSTS and side-stream tobacco smoke are heterogeneous mixtures of ~5000 chemical compounds, with several dozen carcinogens, cocarcinogens, mutagens and tumor promoters (5). Tobacco smoke has been shown to cause a variety of types of DNA damage (19–21), including double-strand breaks (DSBs) (22,23). Yet, the mechanisms by which tobacco smoke cooperates with HR-HPV infection to enhance cancer progression are not clear.

Few investigations have considered the effects of cigarette smoking on HPV activities directly. Xi et al. (14,24) showed current but not prior smoking is associated with higher baseline HPV16 and HPV18 DNA load; however, there was no observed dose–response relationship between cigarette smoking and HPV DNA load. Other studies showed no association of smoking status and HPV viral load for women singly infected by any HR-HPV genotype, or specifically by HPV31 or HPV16 (25). Experimentally, benzo[a]pyrene (BaP), a major carcinogen in cigarette smoke, was shown to cause DNA adducts and damage and induce a p53 response in HPV-immortalized cells (26–28). Alam et al. (29,30) demonstrated that exposure of cervical cells to a specific level of BaP could stimulate either higher levels of viral genomes or higher virion synthesis, but oddly not both, in HPV-infected cells grown as organotypic tissues. This group also showed that increased viral replication resulted from heightened signaling via the mitogen-activated protein kinase (MAPK) pathway (31). However, they failed to show dose responsiveness upon BaP exposure, and the BaP levels tested were of questionable physiologic relevance (25).

Herein we aimed to study physiologically germane effects of all the chemicals present in MSTS-condensate (MSTS-C) on cervical cells that maintain HPV16 or HPV31 genomes either in extrachromosomal forms or in an integrated state in the host cell DNA. Results show that MSTS-C exposure leads to increased viral genome replication and early gene transcription in cells with episomal HR-HPV, but not in cells with integrated HR-HPV genomes. Consistent with increased oncogene E6 transcription, we found decreased p53 protein levels and activity. As expected from the loss of p53 activity in tobacco-exposed HPV episomal cells, DSB levels were significantly higher, but apoptosis was not activated compared with tobacco-exposed cells containing integrated HPV genomes. Furthermore, mutation frequencies were higher in surviving cells with HPV episomal genomes. These data show that tobacco smoke as a cofactor in HPV-related cancers alters viral oncoprotein activities more predominantly in cells with episomal HPV genomes, suggesting a stronger role for tobacco smoke early in cancer progression prior to HPV genome integration.

Materials and methods

Preparation of MSTS-C

A MSTS-C stock was generated from type 2R1 research cigarettes (Tobacco Health Research Institute, Lexington, KY) using a Type 1300 smoking machine (AMESA Electronics, Geneva, Switzerland). The condensate was bubbled through and collected in phosphate-buffered saline. MSTS-C exposure dosages were normalized based on nicotine concentrations, as measured by reverse-phase high-performance liquid chromatography. MSTS-C stocks were stored in aliquots at −80°C, and a fresh vial used for each experiment.

Cell culture and viability

W12 and CIN612 cell lines were established from human cervical intraepithelial neoplasia grade I (CIN1) biopsies (32,33). W12 clone 20863 (W12-E), W12 clone 201402 (W12-I), CIN612 clone 9E (CIN612-9E) and CIN612 clone 6 (CIN612-6) cell lines were derived from their respective parental cell lines. W12-E cells harbor episomal HPV16, whereas W12-I cells contain type I integration of the HPV16 genome with an intact long control region (LCR) (34). CIN612-9E cells maintain episomal HPV31 genomes; CIN612-6 cells contain wild-type p53 and an integrated HPV31 genome of unclear structure (35,36). W12-E and CIN612-9E cells express wild-type p53 genes (37). C-33A is an HPV-negative human cervical carcinoma cell line that expresses mutant p53 (38). Primary human foreskin keratinocytes (HFK) and NIKS cells, a spontaneously immortalized HFK line, both harbor wild-type p53 (39). All cell lines were authenticated. Each cell line was grown as described previously and generally grown in the presence of J2 3T3 fibroblast feeder cells (37). In most cases remaining feeder cells were removed by differential trypsinization. As W12 cells often detached with the feeders, W12-E and W12-I cells were seeded without fibroblast feeder cells prior to the 24 h MSTS-C treatments. After attachment in normal growth media for 16–20h, subconfluent cells were incubated with fresh media containing MSTS-C. Cell survival was measured by Trypan Blue exclusion staining. Toxicity was measured based on lactate dehydogenase detection in cell-free supernatants with lactate dehydogenase-based assay kit (Sigma). Apoptosis was detected by flow cytometry with Annexin V-PE kit (BD Biosciences) according to the manufacturer. By fluorescence-activated cell sorting analysis, Annexin V-PE fluorescence was recorded in FL-2 and 7-amino-actinomycin D in FL-3. Apoptotic cells were labeled with only annexin V, necrotic cells with both annexin V and 7-amino-actinomycin D, and living cells were negative for both.

Nucleic acid extraction, reverse transcription and qPCR assay

Total cellular DNA or RNA was harvested separately as described previously (37). Reverse transcription of total RNAs (0.2–0.5 µg) was performed as reported (40). Relative levels of HPV DNA or each complementary DNA (cDNA) was analyzed by quantitative PCR (qPCR) with the SYBR Green 1 kit on the iCycler (Bio-Rad) in triplicate with PCR primers as we reported (37). Primer placement and amplicons are illustrated in Supplementary Figure 1, available at Carcinogenesis Online. qPCR results from cDNA reactions were normalized by TATA-box binding protein cDNA detection, which was previously validated as a control for HPV-infected cells (37,41), and we ensured is not altered by MSTS-C exposure (data not shown). DNA amplifications were normalized to detection of the glyceraldehyde-3-phosphate dehydrogenase gene. Final results are shown as the average of three independent experiments, whose Q test <0.94. The melting curves and amplification efficiencies are shown in Supplementary Figures 2–4, available at Carcinogenesis Online.

Protein expression assays

Cells were lysed in standard radioimmunoprecipitation assay buffer. Proteins were quantified by Bradford assay and subjected to 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Following transfer to polyvinylidene difluoride membranes (Millipore), immunoblot was performed using anti-p53 antibody (Calbiochem OP43). To control for protein loading, membranes were stripped (62.5mM Tris–HCl pH 6.8; 2% sodium dodecyl sulfate; 100mM 2-mercaptoethanol) at 50°C for 20min and reprobed with anti-β-tubulin antibody (Sigma TO198). Cells were transfected with p53-luc using the Keratinocyte Nucleofector kit as reported (37). p53-luc contains 13 copies of the p53 consensus sequence driving expression of luciferase. Luciferase assays (Promega) were performed according to the manufacturer’s instructions.

Neutral single-cell gel electrophoresis (comet assay)

Neutral comet assays were performed using the Comet Assay kit (Trevigen) according to the manufacturer’s instructions. Briefly, trypsinized and resuspended cells were embedded in agarose on a slide and subjected to solution lysis followed by electrophoresis in 1× Tris/borate/EDTA buffer for 10min at 1 V/cm. The slides were then stained with SYBR Green I, photographed using a Zeiss fluorescence microscope and analyzed with Comet Assay Software Project. At least 80 cells on each slide were chosen at random for the quantification of DNA damage using Comet Assay Software Project. Olive tail moment (OTM) is defined as the product of distance (in x direction) between the center of gravity of comet head (CGH) and the center of gravity of the tail (CGT) and the percent tail DNA (DNAT) wherein OTM = (CGTx − CGHx)/%DNAT.

Mutagenesis

Cells exposed to MSTS-C at various doses were cultured with fibroblast feeder cells for 10 days to allow phenotypic expression, then seeded without feeder cells to determine the mutation frequency of the Oua^R gene: (i) 400 cells per 150mm dish, three dishes/dose, for plating efficiency (PE); and (ii) 10⁶ cells/150mm dish, five dishes/dose, in medium containing 10 μg/ml ouabain (Sigma). At 2 weeks post-seeding, colonies were fixed, stained and counted. The mutation frequency was calculated as [mutation frequency = number of colonies/(number of seeded cells × PE) × 10⁶ cells]. In parallel control experiments, cellular mutation response to 4-nitroquinoline 1-oxide (140ng/ml for 1.5h), a well-characterized mutagen, was assayed using Oua^R as described above.

Statistics

Each experiment was performed at least three times. The results are expressed as mean ± standard error of the mean. Statistical analysis was performed using Student’s t-test, and P value of ≤ 0.05 was considered as significant.

Results

Smoking-relevant concentrations of MSTS-C are non-toxic to cervical epithelial cells

Many previous studies have detected nicotine, its major metabolite, cotinine and tobacco-specific carcinogens in the cervical mucus of smokers (42–47). To study the proliferation and cytotoxicity effects of MSTS-C on HPV-infected cells, cell cultures were exposed to MSTS-C at increasing nicotine concentrations. This physiologically relevant exposure range for both lung and cervical epithelium resulted in a dose-dependent inhibition of cell proliferation for W12-E cells that maintain episomal HPV16 genomes compared with untreated cells (Figure 1A). Toxicity measured by lactate dehydogenase release was only apparent at the highest MSTS-C dose of 20 µg/ml (Figure 1B). W12-I, CIN612-9E, CIN612-6, NIKS and C33A cells responded similarly to W12-E cells in cell proliferation assays at all doses, and 2.5–10 μg/ml MSTS-C treatment was non-toxic to these cells as well (data not shown).

Fig. 1.

Effect of MSTS-C on cell proliferation and cytotoxicity. Subconfluent W12-E cells were treated with the indicated concentrations of MSTS-C at 37°C. (A) Trypan blue exclusion staining during cell quantification was used to determine the effect of MSTS-C on cell proliferation at 0, 24 and 48h posttreatment. Results are shown as the average of three separate experiments; error bars represent standard error of the mean. (B) Cell cytotoxicity was determined by detecting the release of lactate dehydogenase (LDH) in cell-free culture supernatants 24h after treatment with MSTS-C. Data represent an average from three separate experiments and error bars represent standard error of the mean. Statistical significance was achieved with P < 0.05(*).

MSTS-C exposure causes increased HPV viral genome replication and early viral transcription in cells containing HPV episomal genomes

Cells maintaining HPV genomes in either episomal (W12-E, CIN612-9E) or integrated (W12-I, CIN612-6) forms were exposed to increasing amounts of MSTS-C. After 24h of treatment with MSTS-C, total DNAs and RNAs were harvested for qPCR quantification of HPV genome copies and reverse transcription–qPCR analysis of early viral transcripts. The glyceraldehyde 3-phosphate dehydrogenase gene and TATA-box binding protein cDNA were targeted as internal controls in qPCR and reverse transcription–qPCR assays, respectively. Exposure to increasing MSTS-C resulted in dose-dependent and statistically significant increases of viral genome replication and similar increases in viral E6, E7 and E1^E4 messenger RNAs (mRNAs) in W12-E or CIN612-9E cells (Figure 2A and C). MSTS-C treatment of W12-E cells resulted in a 2.4-fold increase in HPV16 viral genome levels, a 2.8-fold rise in E6 transcripts, a 2.0-fold increase in E7 mRNAs and a 3.2-fold climb in E1^E4 transcripts (Figure 2A). MSTS-C treatment of CIN612-9E cells resulted in 2.7-fold higher HPV31 viral genome levels, 3.3-fold higher E6 mRNA numbers, 2.9-fold augmentation of E7 transcripts and 4.1-fold more E1^E4 mRNAs (Figure 2C). In contrast, the MSTS-C treatment had no effect on HPV16 or HPV31 viral genome levels or viral mRNA transcription in W12-I or CIN612-6 cells harboring integrated viral genomes (Figure 2B and D, respectively). These data suggest that exposure to MSTS-C results in the increase in HPV transcription that follows increased genome load in cells containing episomal HPV DNA. The data in Figure 2 are shown relative to the levels of nucleic acids in the untreated cells for ease of comparison. It should be noted that viral oncogene levels are an average of ≈6–8 times greater in cells containing integrated cells compared with those with episomal viral DNA (Supplementary Figure 5, available at Carcinogenesis Online). Finding higher levels of RNAs containing the E7 open-reading frame (ORF) than transcripts containing the unspliced E6 ORF is consistent with other reports (48) (see Supplementary Figure 1, available at Carcinogenesis Online). Further, these data concur with the prevailing notion that HPV genome integration leads to enhanced E6 and E7 RNA levels.

Fig. 2.

Quantification of viral genome copy numbers and early viral transcripts in HR-HPV-positive cell lines exposed to MSTS-C. Subconfluent plates of cells were exposed to the concentrations of MSTS-C as indicated (A-D). After 24h of treatment, the cells were harvested for total RNA and DNA. DNase-treated total RNAs were reverse transcribed to cDNA and subjected to qPCR for E6, E7 or E1^E4 RNAs using TATA-box binding protein cDNA quantification for normalization. Viral DNA levels were quantified by targeting amplification of the viral LCR in reference to levels of the cellular glyceraldehyde 3-phosphate dehydrogenase gene. Results represent the mean normalized expressions from three separate experiments with error bars representing standard error of the mean. Statistical significance was achieved with P < 0.05(*) and P < 0.01(**) as compared with the 0mM control in each group.

Increased HPV E6 early viral transcription coincides with p53 degradation in cells maintaining episomal HPV genomes

The presence of E6 is considered a predisposing factor in the development of HPV-associated cancers, allowing the accumulation of chance errors in host cell DNA to go unchecked (49). As relative levels of p53 protein can be used as an indirect readout for E6 activity, we detected p53 protein in cells by immunoblot. After 24 h treatment with MSTS-C, both W12-E and CIN612-9E cells demonstrated a dose-dependent decrease of wild-type p53 protein levels in a manner consistent with increased E6 expression (Figure 3A). However, in W12-I and CIN612-6 cells containing integrated HPV genomes, the MSTS-C treatment lead to a mild increase in p53 protein levels consistent with their unaffected E6 mRNA levels (Figure 3A). The HPV-negative spontaneously immortalized HFK cell line, NIKS, responded to MSTS-C treatment with substantially increased p53 levels as expected in cells with functional wild-type p53 genes (Figure 3A). Lastly, MSTS-C treatment did not readily change p53 levels in the HPV-negative C-33A human cervical carcinoma cell line harboring mutant p53 (Figure 3A).

Fig. 3.

Detection of p53 expression and transcriptional activity after MSTS-C treatment in W12-E and W12-I cells. (A) Immunoblot analysis of p53 protein expression after MSTS-C treatment. Subconfluent cells were exposed to different concentrations of MSTS-C for 24h. Total protein (NIKS cells, 30 μg; C-33A cells, 0.5 μg; other cells, 10 μg) was subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis and immunoblot for p53 (53kDa); β-tubulin protein (50kDa) was detected as a loading control on the same blot after stripping the membrane. Images represent unaltered scans from films. Values directly above each lane represent the relative p53 protein levels normalized to β-tubulin with the untreated control defined as 1.0; band intensities were performed on the original autoradiographs. (B) p53 transcriptional activity assay in cells transfected with plasmid p53-luc. At 24h posttransfection, cells were treated with MSTS-C at the indicated concentration for 24h, then lysed and analyzed for luciferase activity. The luciferase activity in cells control group (0 μg/ml) was defined as 1, and the treated samples were normalized to this control. The graphs encompass averages from three independent experiments. Statistical significance was achieved with P < 0.05(*) and P < 0.01 (**) as compared with 0 μg/ml control group.

To functionally assess p53 transcriptional activity, a p53-responsive luciferase reporter plasmid was transfected into the cell lines prior to MSTS-C treatment. We found wild-type p53 activity was reduced significantly in W12-E and CIN612-9E cells following MSTS-C treatment (Figure 3B) corresponding to the decreased p53 levels (Figure 3A) and increased E6 RNA levels (Figure 2A and C). In the cells with integrated HPV genomes, MSTS-C treatment induced a slight increase in p53 transcriptional activity concordant with the higher p53 protein levels (Figure 3B). Also consistent with the levels of p53 induced by MSTS-C in HPV-negative, p53 wild-type NIKS cells, MSTS-C treatment lead to a robust and dose-dependent increase in p53 transcriptional activity (Figure 3B). Finally, in agreement with the unchanged p53 protein levels in C33A cells, MSTS-C treatment did not alter p53 transcriptional activity (Figure 3B). These data show that p53 activity is directly related to p53 levels in each cell line, and the levels of HPV E6 protein present in the HPV-infected cells impacted both.

MSTS-C induces more DNA DSBs, but lower apoptosis, in cells containing HPV episomes compared with cells with integrated HPV genomes

Tobacco exposure causes DNA DSBs in mammalian cells (22,23), and DSBs are the most cytotoxic of DNA lesions. As we found p53 activity to be lower upon MSTS-C exposure in cells maintaining episomal HPV16 or HPV31 genomes, compared with cells with integrated HPV genomes, we reasoned that DSBs would be more pronounced in cells maintaining HPV episomes after MSTS-C exposure. To test this idea, DNA DSBs were evaluated in W12-E and W12-I cells using the neutral comet assay wherein cells sustaining DNA damage appear as a comet with a specific bright head and tail. In contrast, cells with undamaged DNA show an intact nucleus with no tail in this assay. As predicted, exposure to increasing doses of MSTS-C resulted in significantly higher DNA DSBs in W12-E cells compared with W12-I cells (Figure 4A and B). HPV-negative HFK cells showed little evidence of DSBs after MSTS-C exposure.

Fig. 4.

Effect of MSTS-C on DSBs and cell apoptosis during a 24h time course of treatment. (A) Neutral comet assay for DNA DSBs is shown subsequent to MSTS-C treatment with representative images of the nuclei depicted. (B) Quantification of the comet tail moments for each condition were calculated on ≥80 cells per group from A; data are shown as a box-and-whisker plot to illustrate the distribution and statistics of the datasets. The box bottom and top are 25th and 75th percentiles of each dataset, respectively; the central band is 50th percentile (median). Vertical lines extend to 10th and 90th percentiles. Outliers are indicated as dots. (C) Apoptosis analysis on cells treated for 24h using double staining with Annexin V-PE and 7-amino-actinomycin D (see Materials and methods). Graphed are the averages of three separate experiments; statistical significance was achieved with P < 0.05(*) and P < 0.01 (**) as compared with 0 μg/ml control group.

In normal cells, DNA DSBs induce expression of p53 to trigger either DNA repair or apoptotic cell death as a protective mechanism (50). To determine whether apoptosis was induced or correlated with lost wild-type p53 activity and MSTS-C-related DNA damage in HPV-positive cells, annexin V and 7-amino-actinomycin D staining was employed after 24 h of MSTS-C exposure. At 0 and 5 μg/ml MSTS-C treatment there were similar, but not statistically significant increases of cell apoptosis in W12-E and W12-I cells (Figure 4C). However, when treated with 5 μg/ml MSTS-C, the HPV-negative NIKS cells showed a statistically significant increase of apoptosis (P < 0.05), which is consistent with the doubled levels of p53 protein and activity induced in these cells under the same conditions (Figure 3). Also consistent with induced levels of p53 activity in the cells when treated with the higher MSTS-C dose (10 μg/ml), both W12-I cells and NIKS cells displayed statistically significant increases in apoptosis (Figure 4C). However, W12-E cells failed to mount a significant apoptotic response to MSTS-C at the higher dose (Figure 4C) where significant DNA DSBs were detected (Figure 4B). As expected, MSTS-C treatment did not induce apoptosis in C-33A cells that express mutant p53 (Figure 4C). Thus, although MSTS-C exposure led to more DSBs in cells with HPV16 episomal genomes, the increased E6 levels suppressed both the p53 activity and the protective function of apoptosis in these cells.

MSTS-C causes accumulation of DNA mutations in cells containing HPV episomes

DNA damage repair of DSBs is error prone and increases the likelihood of DNA mutations that can lead to cancer. As p53 is thought to be involved in the repair process and data above show decreased p53 activity in W12-E cells, but not in W12-I cells after MSTS-C treatment, we quantified the accumulation of mutations in the cells after MSTS-C exposure. The data in Table I show the sensitivity of W12-E and W12-I cells to ouabain. The spontaneous (0 μg/ml MSTS-C) Oua^R gene mutation frequencies were similar in W12-E (11.3±2.9 per 10⁶ cells) and in W12-I (9.4±2.1 per 10⁶ cells). After 24h treatment with 5 or 10 μg/ml MSTS-C, we detected 1.5- or 2.2-fold increases (P < 0.05) of Oua^R gene mutation frequencies in W12-E cells, but only 1.1- and 1.4-fold increases in W12-I cells as compared with their untreated control groups. As a positive control, mutagen 4-nitroquinoline 1-oxide induced Oua^R gene mutations at frequencies significantly higher in W12-E cells (P < 0.01) than in W12-I cells (P < 0.05) when compared with their untreated control groups (Table I). Therefore, MSTS-C exposure results in higher DNA mutation frequencies in W12-E cells where early HPV16 gene transcription was induced, p53 levels decreased and apoptosis was suppressed in response to MSTS-C. This contrasted to W12-I cells where early HPV16 gene transcription was unaltered.

Table I.

MSTS-C induced Oua^R gene mutation frequency after 24h of exposure (N = 3)

Cell	MSTS-C (μg/ml)	OuaR mutants/106 cells
W12-E	0	11.3±2.9
	5	16.1±5
	10	24.4±4*
	4-NQO, 140ng/ml	28.1±1.8**
W12-I	0	9.4±2.1
	5	11.1±3.5
	10	14.3±3
	4-NQO, 140ng/ml	21.7±6.3*

4-NQO, 4-nitroquinoline 1-oxide.

*P < 0.05, compared with 0 μg/ml MSTS-C controls; **P < 0.01, compared with 0 μg/ml MSTS-C controls.

Discussion

Nearly all cases of cervical cancer are linked to HR-HPV infections, with more than half attributable to HPV16 (3). Increasing numbers of oropharyngeal squamous cell carcinomas also are due to HPV16 infections (51). However, HR-HPV infection is an insufficient carcinogen. Epidemiological studies demonstrate a causal relationship between tobacco smoking and cervical cancer development (52–56). Chemical constituents of tobacco and their metabolites, including nicotine, nicotine-derived cotinine, carcinogens 4-(methyl nitrosamino)-1-(3-pyridyl)-1-butanone, BaP and others are present in the cervical mucus of active and passive smokers (42–47). Although these observations suggest that tobacco smoking acts as a cofactor with HR-HPV infection in cervical cancer progression, little is understood about the mechanisms by which the viral, cellular and environmental factors may cooperate in carcinogenesis. In this study, we found that exposure of HPV-infected cervical cells to physiologically relevant doses of MSTS-C elicited increased viral genome replication and oncogene expression, which is known to be essential for HPV-mediated cellular transformation. We showed that the MSTS-C-induced HPV oncoprotein expression, in turn, neutralizes the protective effects of the p53 tumor suppressor protein and leaves HPV-infected cells at increased susceptibility to tobacco smoke-induced DNA DSBs and higher accumulated mutation rates. Interestingly, the MSTS-C actions on HR-HPV activities were seen in cells that maintain episomal viral genomes, but not in isogenic cells with integrated viral genomes. Additionally, cells with episomal genomes had higher DNA strand breaks and lower apoptosis than their integrated counterparts. These observations suggest tobacco smoke exposure plays an important role as a promoter of HPV-initiated carcinogenesis, with key effects on HR-HPV activities occurring prior to, and perhaps promoting, viral genome integration.

It appears that the primary effect of tobacco smoke on HR-HPV activities is that of augmented viral genome replication leading to increased early transcription, rather than genome replication driven by a boost in HPV early transcription. Our reasoning is based on the findings that the fold change in viral genome levels and early gene transcription is similar in cells with episomal HR-HPV genomes combined with the fact that we do not observe increased viral early gene transcription in cells with integrated HPV genomes. The structure of HPV genomes when integrated in cancer cells includes maintenance of the LCR that regulates gene expression (34,57–61). Many studies, including our unpublished data (A.M.Griego and M.A.Ozbun), indicate the LCR and early promoter are functional to drive expression of E6 and E7 from integrated genomes (26,62–64). Therefore, it is reasonable to expect integrated genomes would be responsive to transcriptional stimulation if provided by the MSTS-C. It is interesting to note that the MSTS-C constituent BaP was reported to activate both MAPK and ERK1/2 signaling to increase viral titer in CIN612-9E organotypic epithelial tissues, which maintain episomal HPV31 genomes (31), as well as stimulate oncoprotein expression from integrated HPV16 in CaSki cells (26). It is not surprising that BaP-induced MAPK activation stimulated HPV oncoprotein expression from the integrated HPV16 LCR (26), as the MAPK/ERK pathway activates transcription factors known to bind in the LCR and promote viral gene expression (65). Unfortunately, the effects of BaP on early viral transcription were not assessed in CIN612-9E raft tissues (29–31); such data might help to clarify confounding results in these reports indicating that genome amplification and virion production were not dose responsive and were inversely related. In a previous study, we assayed the response of the cervical cells maintaining episomal HPV16 and HPV31 genomes (W12-E and CINC12-9E cells, respectively) to the effects of nitric oxide, a free radical that is increased in the cervix in response to other factors that cooperate with HR-HPV in cervical cancer progression, including tobacco smoke (37). In contrast to the present work, physiologically relevant levels of nitric oxide promoted oncogene transcription, but not increased viral genome replication. Nevertheless, the outcome of higher oncoprotein expression was similar to what we found herein: increased DNA damage and survival of cells with higher mutational frequencies (37). We emphasize that our current work, employing the complete and more tobacco smoke exposure-relevant complement of chemical constituents in MSTS, cannot be directly compared with the prior studies where cells were exposed only to BaP or to nitric oxide. Our use of the more complex, but physiologically relevant MSTS-C is the most likely explanation for the differences observed in overall effects to HR-HPV in this study compared with those previous reports. This includes our finding no enhanced viral gene expression from integrated HPV16 and HPV31 genomes, in contrast to that seen in BaP-treated CaSki cells with integrated HPV16 genomes (26) and our observed MSTS-C induced increase in viral DNA levels, which were not found in response to nitric oxide exposure (37).

The mechanism of episomal HPV genome replication in response to MSTS-C exposure is not clear. Extrachromosomal HPV genome replication requires E1 and E2 proteins in addition to cellular replication factors, which can be influenced by E6 and E7 proteins (49). The concentrations of MSTS-C we used slightly suppressed cellular proliferation, suggesting cell proliferation was not a key factor in enhancing viral genome replication. Interestingly, the interferon-inducible protein p56 blunts HPV DNA replication via binding E1 (66), but tobacco smoke is known to suppress the interferon response. This may in part explain increased HPV DNA levels following MSTS-C exposure. However, this is a complicated interplay of systems that will take a great deal of experimentation to clarify.

The lack of reagents for quantifying E1 and E2 proteins is a detriment to our understanding of how these proteins regulate genome replication (67). As the viral RNAs that contain E1 and/or E2 ORFs also include E6 (or E6*) and E7 sequences (see Supplementary Figure 1, available at Carcinogenesis Online) (67), the levels of E1 and/or E2 RNAs in undifferentiated cells are predicted to mirror those of E6 and E7. Additionally, viral RNA levels may not be indicative of protein levels as it is unclear which ORFs in the polycistronic RNAs are actually translated. Although loss of E2 expression via HPV genome integration results in increased viral oncogene expression, viral integration would need to occur rapidly and specifically in a majority of cells, which is highly improbable in our 24 h period to assay for MSTS-C effects. Further, genome integration would preclude the increased viral DNA copies relative to genomic genes we observed. Thus, we do not feel viral integration and loss of E2 could account for our results. Nevertheless the heightened DNA strand breaks and lower apoptotic indexes are likely to support HPV integration at some point, whereby a subset of those events might lead to an eventual growth advantage, but this would require many cell generations to prevail.

Although persistent HR-HPV infections are a clear prerequisite for the development of cervical cancer (68–70), there is no consistent association of HR-HPV DNA load (the levels of viral DNA detected) with persistence (14,25). Increased viral load in the cervix has also been inconsistently reported for current and/or ever smokers (15,16,24,25). Nevertheless, any observed association is proposed to favor HPV persistence in vivo (15,16), but this is yet inconclusive (25). Without a true quantitative assay for viral load in individual sampled cells, one cannot differentiate in a cell population between a more homogeneous population with high DNA levels versus a heterogeneous cell group where most have low HPV copies and a few cells contain very high levels of HPV DNA. This limits the informative nature of the current assays in this regard. The general lack of a dose–response relationship observed between cigarette smoking and viral load could indicate a low threshold for the effect of smoking on HPV DNA load (25). Our data show cells with a ≤3-fold increase in viral DNA levels and similar increase in E6 RNAs have a significant deficit in ability to repair DNA or undergo apoptosis; this could mean that small changes in the levels of HPV genomes in a minor population of cells may have substantial functional effects in restraining tumor suppressor functions, thus leading to malignant progression of these few cells in vivo.

A number of aspects differentiate our work from prior studies. The use of the tobacco smoke exposure-relevant chemical mixture and physiologically relevant levels of MSTS-C sets our work apart from many previous reports that have only used BaP to study the response of HPV-immortalized cells (26–31). This provides assessment that is more similar to tobacco exposure in vivo. A limitation of our work, and that of the other studies with BaP, is that the local concentrations of the MSTS-C chemicals to the cells in the cervix is not known; however, we have striven to mimic as closely as possible by normalizing our MSTS-C using nicotine levels that are found in the cervical mucus (44). Prior studies on the effects of tobacco chemicals have not compared isogenic cells and have typically analyzed only a single HR-HPV genotype. Our use of two HR-HPV positive isogenically matched sets of cervical cell lines with differing HPV structures allow us to rule out many aspects of cell line-to-cell line variation. Further, we show consistent results comparing extrachromosomal HPV16 and HPV31 genome-carrying cells with their isogenic counterparts that have integrated HPV genomes; this provides more confidence that our results can be generalized to HR-HPVs. Lastly, the MSTS-C dose-responsive results observed in cells containing both HPV16 and HPV31 episomes lends biological plausibility to our findings.

In summary, our demonstration of preferential upregulation of HPV oncoproteins in cells with episomal genomes, followed by increased DNA damage and higher mutation frequencies provides the bases for a theory on the role for tobacco smoke exposure in the progression to cancer of HR-HPV-positive infections (Figure 5). We propose that in HR-HPV-infected cells with episomal genomes, tobacco smoke not only causes DNA adducts and strand breaks (5), but independently causes increased viral load within cells. This results in heightened HPV oncoprotein expression that further squelches cell cycle control, DNA damage repair and protective apoptosis. Increased viral genome replication in the midst of DNA DSBs is a probable promoter of viral genome integration, wherein cells that maintain higher E6 and E7 expression gain a growth advantage and can continue to accumulate mutations that permit malignant progression. The immune suppression that occurs in the context of tobacco exposure likely augments the survival of these cells (71). This experimental model provides the impetus for expanded epidemiologic studies of the relationship between smoking-related cervical cancer and HPV integration status, which may provide better understanding of tobacco smoke risk to HPV infection and cancer risk. Finally, we expect this model to be applicable to other cancers, especially those oropharyngeal malignancies that are initiated by HR-HPV infections.

Fig. 5.

Model for HR-HPV and tobacco smoking interaction in cervical cancer progression. Infection with an HR-HPV is a necessary cause of cervical and other epithelial cancers. Tobacco smoking is an epidemiologically defined cofactor for progression to cervical cancer and is known to increase nitric oxide levels and induce DNA adducts and strand breaks (5), as well as suppress the immune system (71). Similar to many previous studies, we find in normal HPV-negative cells, tobacco smoke exposure induces high levels of p53 leading to DNA repair or apoptosis wherein mutant cells generally fail to survive (right side). In HR-HPV-infected cells that contain episomal viral genomes, tobacco smoke exposure causes an increase in HPV genome copies, higher HPV oncoprotein RNA expression leading to a decrease in p53 activity and improved survival of cells with mutations (left side). Increased viral DNA load in the midst of DNA DSBs may promote integration of viral genomes into the host cell chromosomes. As the cells with integrated HR-HPV genomes showed no further increase in oncoprotein activities upon tobacco exposure, we conclude that the primary effects of tobacco smoke constituents are focused on cells with episomal genomes to promote early steps in malignant conversion.

Supplementary material

Supplementary Figures 1–5 can be found at http://carcin.oxfordjournals.org/.

Funding

Primary support was by the American Cancer Society (RSG-05-149-01-MBC to M.A.O.); we also gratefully acknowledge the Oxnard Foundation (M.A.O.); the University of New Mexico Cancer Center (National Institutes of Health P30 CA118100); the National Natural Science Foundation of China (NSFC 30902706 to L.W.); the Heilongjiang Provincial Natural Science Foundation of China (ZD201020 to L.W.).

Glossary

Abbreviations:

BaP

benzo[a]pyrene

cDNA

complementary DNA

CIN

cervical intraepithelial neoplasia

DSB

double-strand break

HFK

human foreskin keratinocyte

HR-HPV

high-risk human papillomavirus

LCR

long control region

MAPK

mitogen-activated protein kinase

mRNA

messenger RNA

MSTS-C

mainstream tobacco smoke condensate

ORF

open-reading frame

qPCR

quantitative PCR.