|Home | About | Journals | Submit | Contact Us | Français|
Epidemiological studies suggest that human papillomavirus (HPV)-infected women who smoke face an increased risk for developing cervical cancer. We have previously reported that exposure of HPV-positive organotypic cultures to benzo[a]pyrene (BaP), a major carcinogen in cigarette smoke, resulted in enhanced viral titers. Since BaP is known to deregulate multiple pathways of cellular proliferation, enhanced virion synthesis could result from carcinogen/host cell interaction. Here, we report that BaP-mediated upregulation of virus synthesis is correlated to an altered balance between cell cycle-specific cyclin-dependent kinase (CDK) activity profile compared with controls. Specifically, BaP treatment increased accumulation of hyperphosphorylated retinoblastoma protein (pRb) which coincided with increased cdc2/CDK1 kinase activity, but which further conflicted with the simultaneous upregulation of CDK inhibitors p16INK4 and p27KIP1, which normally mediate pRb hypophosphorylation. In contrast, p21WAF1 and p53 levels remained unchanged. Under these conditions, CDK6 and CDK2 kinase activities were decreased, whereas CDK4 kinase activity remained unchanged. The addition of purvalanol A, a specific inhibitor of CDK1 kinase, to BaP-treated cultures, resulted in the production of noninfectious HPV type 31b (HPV31b) particles. In contrast, infectivity of control virus was unaffected by purvalanol A treatment. BaP targeting of CDK1 occurred independently of HPV status, since BaP treatment also increased CDK1 activity in tissues derived from primary keratinocytes. Our data indicate that HPV31b virions synthesized in the presence of BaP were dependent on BaP-mediated alteration in CDK1 kinase activity for maintaining their infectivity.
Worldwide, tobacco smoking is a major source of exposure to a number of well-characterized carcinogens which have been associated with multiple cancers (28, 32). Epidemiological studies have established a causal relationship between cigarette smoking and the development of human papillomavirus (HPV)-associated cervical cancer (10, 43, 55), with the risk of developing cancer doubling among HPV-infected women who are smokers compared with those who never smoked (79). Polycyclic aromatic hydrocarbons (PAHs), such as benzo[a]pyrene (BaP), and tobacco-specific nitrosamines, such as 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK) and their respective metabolites, which are major carcinogenic constituents in cigarette smoke, have been shown to be present in the cervical mucus of smokers (47, 63). Therefore, in combination with HPV infections, the presence of these carcinogens in the cervix may cumulatively determine cervical cancer progression (25).
Most HPV infections are cleared spontaneously (29), and therefore, viral persistence and viral load are thought to be necessary for cancer progression (76, 82). Using organotypic “raft” culture, we have recently reported on the effects of BaP exposure on the productive life cycle of three high-risk HPV types, HPV type 31 (HPV31), 16 (HPV16), and 18 (HPV18) (2). To date, organotypic culture is the best physiologically relevant in vitro model system available, which closely mimics the natural replication of the virus as it occurs in vivo (53). We showed that treatment with high concentrations of BaP resulted in a 10-fold increase in HPV31 viral titers with a concomitant decrease in the number of viral genome copies, while treatment with low BaP concentrations resulted in a 2-fold increase in the number of genome copies but no change in viral titers (2). Thus, our studies suggested that exposure to cigarette smoke carcinogens, such as BaP, could lead to manipulation of host cell and/or HPV-specific functions resulting in enhancement of the “total viral load,” with respect to increased virion synthesis as well as viral genome amplification. Both scenarios, either alone or in combination, may determine host tissue carcinogenesis as well as viral persistence. In support of our studies, we have recently reported that a high HPV16 and HPV18 DNA load in patient cervical samples was positively correlated with cigarette smoking, suggesting that exposure to these carcinogens promotes conditions which could increase viral persistence (83).
A well-studied characteristic of BaP is its ability to induce carcinogenesis, which has been attributed to its formation of DNA adducts (tumor initiator) as well as to its ability to act as a tumor promoter (61). Although the mechanisms responsible for the tumor-promoting effects and carcinogenicity of PAHs, such as BaP, have not been well established, they are known to activate pathways that control cell proliferation, differentiation, and apoptosis (8, 42, 62, 68, 75, 77, 78, 85). A few studies have attempted to address the global effects of cigarette smoke carcinogens in HPV-infected cells (48, 74), the proliferative effects of BaP on healthy human epidermal keratinocytes (60), and HPV16-positive oral keratinocytes (59). However, molecular mechanisms targeted by cigarette smoke carcinogens in HPV-infected keratinocytes have not been examined. Therefore, it is of interest to characterize the nongenotoxic or tumor-promoting effects in HPV-infected keratinocytes that may be linked to these carcinogens. BaP exposure has the potential to deregulate host cell cycle control mechanisms, which in turn could affect HPV functions or may affect both these targets individually. In the current study, we show that BaP-mediated upregulation of HPV31b titers is correlated with enhanced G2-phase-specific kinase activity associated with cyclin-dependent kinase 1 (CDK1), with simultaneous inhibition of G1- and S-phase-specific kinase activities. Purvalanol A-mediated inhibition of CDK1 kinase activity in BaP-treated tissues resulted in the synthesis of noninfectious HPV31b particles, whereas control virions were unaffected. These results suggest that virus propagated in the presence of BaP treatment was highly dependent on CDK1 kinase for determining virus infectivity. In addition, BaP treatment of HPV-negative cultures also resulted in activation of CDK1 simultaneously with CDK4/CDK6 kinase activities, suggesting that aberrant activation of cell cycle kinases by BaP in the midst of differentiation-associated cell cycle exit cues could be a measure of BaP-associated carcinogenesis. In addition to targeting cell cycle checkpoints, BaP simultaneously deregulated expression of tissue differentiation marker profiles. Our studies suggest the possibility that BaP-mediated upregulation of virion synthesis is related to its ability to manipulate differentiation and cell cycle controls and that these proteins could be utilized as markers of carcinogen exposure affecting these tissues.
Benzo[a]pyrene (BaP) was purchased from Aldrich. Dimethyl sulfoxide (DMSO) was used as solvent (Sigma). Purvalanol A was purchased from Calbiochem.
The CIN-612 9E cell line was established from a cervical intraepithelial neoplasia type I biopsy specimen containing HPV31b DNA and has been previously characterized (53). The cells were maintained in monolayer culture with E medium containing 5% fetal bovine serum in the presence of mitomycin C-treated J2 3T3 feeder cells. Primary human keratinocyte (HK) cultures were derived from newborn foreskin via trypsin digestion at 37°C as previously described (67). HKs were maintained in monolayer cultures without feeder cells, with 154 medium (Cascade Biologics, Portland, OR), supplemented with antibiotics (Cascade Biologics) and human keratinocyte growth supplement (Cascade Biologics).
Organotypic raft cultures were generated as previously described (53). These conditions have been shown to be optimal for HPV replication and production of infectious progeny. Raft cultures were lifted to the air-liquid interface and fed every other day with E medium as described previously (53) supplemented with 1 μM BaP or DMSO for a total of 12 days, also as previously described (2). On day 12, the raft tissues were harvested by removing the epithelial layer and stored at −70°C until further manipulations. For the generation of primary HK raft cultures, 80% confluent monolayer HK cultures were trypsinized, resuspended in E medium supplemented with 5 ng of epidermal growth factor (EGF) per ml, followed by seeding 106 cells/ml on top of collagen matrices. After 2 h, the medium was removed, and raft cultures were lifted to the air-liquid interface and fed every other day with E medium without EGF and supplemented with 1 μM BaP or DMSO. The raft tissues were harvested on day 12 by removing the epithelial layer and stored at −70°C until further manipulations.
Total protein extracts were prepared from 12-day raft cultures and quantitated as previously described (51). A total of 60 μg whole-cell extract was used in Western blots to determine expression of p21WAF1, p27KIP1, p16INK4, CDK4, CDK6, CDK2, and CDK1. To detect retinoblastoma protein (pRb) protein expression, 30 μg of whole-cell extracts were used. Protein extracts were applied to sodium dodecyl sulfate (SDS)-polyacrylamide gel (acrylamide/bisacrylamide ratio, 30:0.8). Gel compositions for resolving various proteins are as follows. A 12% gel was used to detect p21WAF1, 15% gel for p16INK4, 10% gel for p27KIP1, CDK1, CDK2, CDK6, and CDK4, 7.5% gel for retinoblastoma protein (Rb), involucrin, keratin 10 (K-10), and keratin 14 (K-14). Polyclonal antibodies against p21WAF1, p27KIP1, p16INK4, pRb, and CDK2 were each used at a dilution of 1:2,000 to detect the respective proteins and have been described previously (3). The pRb protein was detected using two different antibodies: Santa Cruz (catalog number sc-50) rabbit polyclonal antibody detects both the hyper- and hypophosphorylated form of this protein, whereas Becton Dickinson (catalog number 554164) mouse monoclonal antibody detects exclusively the hypophosphorylated form. Polyclonal antibodies against CDK1 (sc-747; Santa Cruz), CDK4 (sc-260; Santa Cruz), and CDK6 (sc-177; Santa Cruz) were each used at a dilution of 1:2,000 to detect respective proteins. The proteins were detected using the enhanced chemiluminescence (ECL) kit (Perkin Elmer) per the manufacturer's instructions. For detecting involucrin, K-10, and K-14 expression, 5-μg whole-cell extract were used. The mouse monoclonal involucrin antibody (clone SY5; Sigma) and K-10 antibody (BioGenex) were used at a dilution of 1:7,000. The mouse monoclonal K-14 antibody (Sigma) was used at a dilution of 1:3,000.
Total protein extracts from BaP-treated and mock monolayer cultures were prepared as we have previously described (3). Determination of total protein levels using Western blot analysis was performed as previously described (3).
For immunoprecipitating p53 proteins, raft cell extracts were prepared from 12-day raft tissues using a modification of methods previously described (3). Protein extracts were prepared from raft tissues as follows. Individual raft culture tissues were homogenized in 350 μl buffer A (50 mM NaPO4 [pH 7.2], 5 mM EDTA, 50 mM NaF, 0.5 μg/ml leupeptin, 0.5 μg/ml pepstatin, 20 μg/ml aprotinin, 0.2 mM Na3VO4, 1 mM dithiothreitol [DTT], 1 mM phenylmethylsulfonyl fluoride [PMSF]) with 10 strokes using a glass homogenizer. The samples were transferred into a 1.5-ml microcentrifuge tube. Four hundred micrograms of ice-cold acid-washed glass beads (0.45 mm) was added to each sample. The tubes were placed into a multitube holder and vortexed for 30 s. The samples were then chilled on ice for 30 s. This procedure was repeated a total of six times. The microcentrifuge tubes were placed on the rims of 13- by 100-mm precooled disposable glass tubes. An 18-gauge needle was used to puncture holes through the lid and bottom of the microcentrifuge tube. The piggy-back microcentrifuge tube and glass tube combination was centrifuged to attain a speed of 2,000 rpm and then centrifugation was stopped. The microcentrifuge tube was discarded. The samples were transferred into a 1.5-ml microcentrifuge tube and further clarified by centrifugation for 5 min at 4°C at 13,000 rpm. Protein concentrations were determined using Bio-Rad assays.
Immunoprecipitation buffer was prepared using the following buffer composition: 1% Triton X-100, 10 mM HEPES (pH 7.5), 2 mM EDTA, 50 mM NaF, 0.2 mM Na3VO4, 2 mM DTT, 200 mM NaCl, 20 μg/ml aprotinin, 0.5 μg/ml pepstatin, 0.5 μg/ml leupeptin, and 1 mM PMSF. For each immunoprecipitation reaction mixture, a total of 200 μg of protein extract was used. The volume was made up to 1 ml with immunoprecipitation buffer supplemented with 2% nonfat dry milk. All samples were first precleared using 1:5,000 dilution of mouse anti-cytokeratin AE1/AE3 monoclonal antibody (Chemicon International) by incubating for 1 h at 4°C with rotation. To each sample, 50 μl of preswollen protein A Sepharose beads (Amersham Biosciences) was added and incubated at 4°C for 30 min with rotation. The beads were removed by centrifugation at 13,000 rpm for 20 s. The supernatant was removed and added to a fresh tube. The precleared lysates were then subjected to immunoprecipitation using 5 μl of monoclonal antibodies against p53 (Oncogene). After the addition of antibody, all samples were incubated at 4°C for 1.5 h with rotation, followed by the addition of 50 μl of preswollen protein A Sepharose beads as the immune complex binding agent. The samples were further incubated at 4°C for 1 h with rotation. The beads were pelleted by centrifugation at 13,000 rpm for 20 s and then washed three times with 500 μl ice-cold immunoprecipitation buffer without milk supplementation. After the final wash, the beads were resuspended in 60 μl of 1× sample loading buffer and boiled for 10 min in a water bath. A total of 10 μl of each sample was used for detecting p53 in the immunoprecipitated complexes utilizing Western blot analysis.
For immunoprecipitating p53 from monolayer cultures, extracts from BaP-treated and mock-treated monolayer cultures were prepared, followed by immunoprecipitation of p53 using 5 μl of monoclonal antibodies against p53 (Oncogene), also as previously described (3).
Raft tissues were grown for 12 days, and protein extracts were prepared and quantitated as described above. CDK1- and CDK2-associated kinase activity assays were performed as previously described using histone H1 as substrate (3). CDK4- and CDK6-associated kinase activity assays were performed using the same protocols used for CDK1 and CDK2 in this study but using pRb-GST (GST stands for glutathione S-transferase) as a substrate as reported elsewhere (12).
For determining kinase activities in monolayer cultures, cell extracts from BaP-treated and mock-treated monolayer cultures were prepared as described previously (3), followed by determination of kinase activities associated with CDK4, CDK6, CDK2, and CDK1, also as previously described (3).
Raft cultures were grown for 12 days as described above, harvested, fixed in 10% buffered formalin, and embedded in paraffin, and 4-μm cross-sections were prepared as we have previously described (2). Sections were stained with hematoxylin and eosin as previously described (53).
For immunohistochemical analysis, the involucrin antibody (Sigma) was used at a dilution of 1:100, using protocols described previously (58). The K-10 antibody was purchased from BioGenex (AM201-5M). Staining procedures used were as recommended by the manufacturer. The proliferating cell nuclear antigen (PCNA) antibody rabbit polyclonal antibody (catalog number sc-7907; Santa Cruz) was used as we have described previously (33).
The CIN-612 9E cell line was grown to approximately 80% confluence, trypsinized, and plated at a density of 1 × 106 cells in 100-mm plates in E medium, without the addition of mitomycin C-treated J2 feeder cells. After the cells were plated, they were incubated between 10 and 12 h, at which time about two-thirds or more of the cells are naturally synchronized at the G1 phase as previously described (3). This time point was designated time zero (t = 0). The medium was aspirated and replaced with fresh medium supplemented with DMSO or with 1 μM BaP, and cells were treated for a total of 7 days (mock treated with DMSO or treated with BaP). Both mock-treated and BaP-treated cell samples were trypsinized, inactivated by the addition of serum, pelleted, and stored in −70°C until further manipulations. Cell samples were collected at t = 0 and on days 1 through 7. On day 2 and day 5, both mock-treated and BaP-treated cells were passaged 1:2 and treated with fresh BaP.
CIN-612 9E cell monolayer cultures were synchronized as described above and also as previously described (3) and were either mock treated with DMSO or treated with 1 μM BaP for a total of 7 days. Both mock-treated and BaP-treated cell samples were trypsinized, inactivated by the addition of serum, and counted using trypan blue exclusion using standard protocols. Cell samples were collected at t = 0 and on days 1 through 7. On day 2 and day 5, both mock-treated and BaP-treated cells were passaged 1:2 and treated with fresh BaP.
For performing flow cytometry, BaP-treated and mock-treated cells were prepared for analysis as previously described (3). Briefly, CIN-612 9E cells were synchronized and treated with BaP or mock treated with DMSO from day 1 through day 7, with passaging on day 2 and day 5 as described above. On each day, cells were harvested by trypsinization, washed with phosphate-buffered saline (PBS), fixed in 70% ethanol, and stored at −20°C. The fixed cells were washed in PBS and then resuspended in PBS containing 0.1% Triton X-100 (Sigma), 200 μg/ml DNase-free RNase A (Boehringer Mannheim), and 100 μg/ml of propidium iodide (Sigma) for 30 min at 37°C. Flow cytometric analysis of 106 cells was carried out in a fluorescence-activated cell sorter (FACS), and the percentages of cells in the G1, S, and G2/M phases of the cell cycle were determined using the Cell Quest program of Becton Dickinson. Data were analyzed with the Mod Fit LT program.
HPV31b infectivity assays were performed using protocols we have previously published (2). Genome encapsidation assays were performed using protocols essentially as we have previously published (16) except that only one raft tissue per treatment was utilized for preparing crude viral preparations by Dounce homogenization. HPV31 targets were amplified using 5′ and 3′ primers as follows: 0.3 μM 5′-31E25′RTPCR (5′-CAG TAT TAA CCA CCA GGT GGT G-3′) as the forward primer (nucleotides [nt] 2848 to 2869) and 0.3 μM 3′-31E23′RTPCR (5′-GTT CAA GAC TTG TTT GCT GCA TTG TCC-3′) as the reverse primer (nt 2993 to 2967). Oligonucleotides were synthesized by Integrated DNA Technologies (Coralville, IA). A standard curve was generated by amplifying 1-μl aliquots of 104, 103, 102, and 101 serially diluted pBSHPV31 copy number controls. Acceptable R2 values for standard curves were at or above 0.99. A Bio-Rad iQ5 multicolor real-time quantitative PCR (qPCR) machine and software were utilized for PCR amplification and data analysis. The PCR thermocycling profile was as follows: a 15-min hot start at 95°C, followed by 40 cycles, with one cycle consisting of 15 s at 94°C, 30 s at 52°C, and 30 s at 72°C. The data analysis was commenced during the extension phase.
In order to determine equal loading of total immunoprecipitated samples in Western blots and kinase assays, identical blots were stained with GelCode blue stain reagent (Pierce) after proteins were transferred onto nitrocellulose membranes.
For our studies, we used raft cultures derived from the CIN-612 9E cell line which maintains episomal copies of HPV31b (53). We have recently reported that treatment of HPV31b-positive raft cultures with 1 μM BaP resulted in a 10-fold increase in viral titers (Fig. (Fig.1)1) (2), which was accompanied by increased L1 capsid protein expression as determined from Western blots (2). Treatment with 0.1 and 0.01 μM BaP resulted in dosage-independent decreases in genome amplification compared to untreated controls (2). However, treatment with 0.001 μM BaP resulted in a greater than 2-fold increase in genome amplification which could not be correlated with changes in viral titers or L1 capsid protein expression (2). Since BaP has been shown to mediate carcinogenesis by targeting cellular proliferation pathways in multiple cell types (8, 19, 34, 40, 42, 68, 69), we reasoned that BaP regulation of increased virion synthesis could be related to its ability to deregulate cell cycle checkpoints in HPV-infected keratinocytes.
In the first set of experiments, we investigated whether BaP also affected cell cycle regulation in HPV31b raft cultures. We treated HPV31b raft cultures with 1 μM BaP, with DMSO as the vehicle control, and an additional control with growth medium only. Raft cultures were treated with BaP every other day and harvested on day 12. We then prepared whole-tissue protein extracts as previously described (51) followed by Western blot analysis to determine the protein expression levels of cell cycle tumor suppressors pRb, p16INK4, p27KIP1, and p21WAF1, all of which act in the G1 phase of the cell cycle. Upon BaP treatment, pRb protein expression was significantly upregulated and stabilized in its hyperphosphorylated (inactive) form, which was not recognized by another antibody with specificity to the hypophosphorylated (active) form of pRb (Fig. (Fig.2A).2A). BaP targeting of pRb hyperphosphorylation has also been documented in rat liver epithelial cells and was correlated with the genotoxic effects of BaP (4). The observed BaP-mediated stabilization of total pRb levels could not be correlated with changes in steady-state HPV E7 oncoprotein levels (data not shown). Concurrently, p16INK4 protein expression was also upregulated (Fig. (Fig.2A).2A). In support of our studies, p16INK4 protein expression has been shown to be overexpressed in dysplastic and neoplastic epithelial cells of the cervix uteri and has been utilized as a specific biomarker for progression (39). In addition, p27KIP1 protein levels were greatly upregulated (Fig. (Fig.2A).2A). Taken together, increased expression of p16INK4 and p27KIP1 CDK inhibitors cumulatively demonstrate the ability of BaP to mediate an apparent G1-like block. However, the observed G1-like block was contradictory, given that pRb was expressed predominantly in its hyperphosphorylated form (Fig. (Fig.2A),2A), and increased expression of p16INK4 and p27KIP1 would normally be expected to mediate pRb hypophosphorylation.
p21WAF1 is a universal CDK inhibitor, and its expression has been shown to be upregulated upon DNA damage in a p53-dependent manner (13, 26, 45). It is well-known that BaP metabolites inflict DNA damage by forming DNA adducts (6). We determined that the level of expression of the p21WAF1 tumor suppressor protein was unresponsive to BaP treatment (Fig. (Fig.2A).2A). The inability of BaP treatment to affect p21WAF1 expression was not unexpected, since studies published elsewhere have reported that bypassing p21WAF1 tumor suppressor functions is a characteristic of the stealth nature of cigarette smoke carcinogens, by which the carcinogens induce DNA damage and propagate mutations without inducing cell cycle arrest (20, 35, 36). However, in the current study, the observed lack of p21WAF1 induction with BaP could also be due to the HPV E6 oncoprotein targeting of p53 functions. Since p21WAF1 expression is also a transcriptional target of p53 (26), we then determined the effect of BaP expression on the level of p53 tumor suppressor protein (Fig. (Fig.2A).2A). The p53 protein was immunoprecipitated from BaP-treated and control tissues followed by Western blot analysis. Total p53 protein levels were also found to be unchanged upon BaP treatment compared with control tissues (Fig. (Fig.2A).2A). From this first set of experiments, we established that BaP exposure of HPV-infected tissue results in the generation of a specific profile of cell cycle tumor suppressor protein expression (Fig. (Fig.2A),2A), which could be correlated with high levels of virion synthesis (Fig. (Fig.1)1) (2). These experiments have been repeated multiple times with reproducible results.
The determination of predominantly hyperphosphorylated (inactive) pRb upon BaP treatment of HPV-infected cultures was an intriguing observation, since its appearance conflicted with the imposition of a G1-like arrest characterized by upregulated p16INK4 and p27KIP1 expression (Fig. (Fig.2A).2A). Therefore, aberrant regulation of cell proliferation in the midst of cell cycle exit cues may be an indication of BaP-regulated carcinogenesis. Since cell cycle progression is dependent on differential phosphorylation of pRb by multiple cyclin-dependent kinases (CDKs), we wanted to determine the identity of the CDK kinase activity which correlated with pRb hyperphosphorylation in response to BaP treatment. Therefore, we examined the kinase activities associated with multiple CDKs which regulate G1-, S-, and G2-specific functions.
We first examined kinase activities associated with CDK4 and CDK6 which act in the G1 phase of the cell cycle (72, 73). Raft tissues derived from 12-day HPV31b-positive keratinocytes treated with 1 μM BaP exhibited CDK4 kinase activity which was similar to the kinase activity in untreated and DMSO controls (Fig. (Fig.2B).2B). In contrast, CDK6 kinase activity was significantly decreased following 1 μM BaP treatment compared with controls (Fig. (Fig.2B).2B). Changes in CDK6 kinase activity could not be correlated to changes in expression of the total protein level of CDK6 (Fig. (Fig.2C).2C). Under these same conditions, the levels of expression of the p16INK4 tumor suppressor were upregulated (Fig. (Fig.2A),2A), which is consistent with inhibition of CDK6 kinase activity but not that of CDK4. In healthy cells, p16INK4 binds to and inhibits kinase activities associated with both CDK4 and CDK6 (72, 73). We further used immunoprecipitation experiments to show that inhibition of CDK6 was correlated with increased binding of p16INK4 to CDK6 compared with p16INK4 binding to CDK4 (Fig. (Fig.2D2D).
Next, we determined the effect of BaP treatment on CDK2-associated kinase activity, which acts in G1/S transition and in the S phase of the cell cycle (72, 73). Under these conditions, BaP treatment inhibited CDK2-associated kinase activity compared with controls (Fig. (Fig.2B).2B). Inhibition of CDK2-associated kinase activity could not be correlated with changes in the levels of total CDK2 protein in raft tissues (Fig. (Fig.2C).2C). Inhibition of CDK2 kinase activity could be correlated with increased p27KIP1 expression but not steady-state p21WAF1 levels (compare Fig. Fig.2A2A with Fig. Fig.2B).2B). Both p21WAF1 and p27KIP1 are known to inhibit kinase activities associated with CDK2 (72).
Last, we examined kinase activity associated with CDK1, which is known to be activated in the G2 phase of the cell cycle (70). HPV31b raft cultures treated with 1 μM BaP exhibited a 2-fold increase in CDK1 kinase activity compared with untreated and DMSO controls (Fig. 2B and E) and which could not be correlated with changes in the levels of total CDK1 protein (Fig. (Fig.2C).2C). Thus, increased CDK1 kinase activity (Fig. 2B and E) was positively correlated with the appearance of hyperphosphorylated pRb (Fig. (Fig.2A)2A) and enhanced virion synthesis (Fig. (Fig.1)1) (2) in raft cultures treated with 1 μM BaP. These results suggest that BaP-targeted activation of altered G2-phase-specific functions, with a further arrest in G1- and S-phase functions, are conditions which appear to be highly conducive for enhancing HPV virion synthesis (Fig. (Fig.1)1) (2). These experiments have been repeated multiple times with reproducible results.
We have previously demonstrated that the complete HPV life cycle is strictly dependent on host tissue differentiation (5, 22, 49, 50, 52, 53, 58). Therefore, it was of interest to determine whether the observed changes in the expression profile of the multiple tumor suppressors upon BaP treatment was specific to differentiation, conditions which in general are conducive for virion synthesis. We synchronized monolayer (nondifferentiating) cultures of HPV31b-positive cells as described in Materials and Methods and as we have previously described (3), followed by treatment of the cells with 1 μM BaP over a period of 7 days. On each day, control cells and BaP-treated cells were trypsinized, pelleted, and frozen in −70°C until further manipulations. The cultures were passaged at a ratio of 1:2 on day 2 and day 5 followed by treating the cells with fresh BaP at these times. We then performed Western blot analysis to compare the levels of expression of the cell cycle tumor suppressors in BaP-treated cells and control cells. Very little pRb protein is expressed in monolayer cultures, and its levels were unchanged in response to BaP treatment (Fig. 3A and B). Further, pRb phosphorylation status, in response to BaP treatment, was indistinguishable from those of controls (compare Fig. Fig.2A2A with Fig. 3A and B). In contrast to their increased protein levels observed in raft cultures, p16INK4 and p27KIP1 protein levels were decreased in monolayer cultures treated with BaP compared with controls (compare Fig. Fig.2A2A with Fig. 3A and B). The levels of expression of p21WAF1 were also decreased in BaP-treated monolayer samples, which was not observed in raft cultures (compare Fig. Fig.2A2A with Fig. 3A and B). The p53 protein levels in monolayer cultures were unchanged as was observed in HPV31b raft cultures (compare Fig. Fig.2A2A with Fig. 3A and B). From these results, we conclude that the process of tissue differentiation and activation of differentiation-associated signaling pathways greatly influence BaP targeting of cell cycle checkpoints.
We also determined kinase activities associated with CDK4, CDK6, CDK2, and CDK1 in BaP-treated HPV31b monolayer cultures. Kinase activities associated with all the CDKs tested were generally downregulated in BaP-treated monolayer cultures (Fig. 4A and B), which was correlated with a general decrease in the total protein levels of most of the CDKs tested (Fig. (Fig.4C).4C). Decreased kinase activity in monolayer cultures observed in the current study is a reflection of the growth inhibitory properties of BaP (9, 21, 62, 64, 65). Analysis of growth curves over the 7-day period revealed that the viability of cells in the BaP-treated monolayers was dramatically decreased compared with DMSO controls (Fig. (Fig.5A).5A). In contrast, DMSO-treated control cells reached confluence and were passaged 1:2 on day 2 and day 5 (Fig. (Fig.5A).5A). Cell passaging on day 5 also resulted in the observed decrease in the total number of cells on day 6 (Fig. (Fig.5A).5A). In order to correlate cell viability with cell cycle progression, we also performed fluorescence-activated cell sorting (FACS) analysis of BaP-treated and DMSO-treated HPV31b monolayer cultures. Decreased viability of the BaP-treated cells (Fig. (Fig.5A)5A) could be correlated with progressively increased G1/S-phase progression, with a further accumulation of these cells with S- and G2/M-phase DNA content and associated growth arrest (Fig. 5B and C).
Our results demonstrate that cell cycle pathways activated upon keratinocyte differentiation in three-dimensional tissues (Fig. (Fig.2)2) serve to bypass a majority of the growth inhibitory effects of BaP observed in nondifferentiating monolayer cultures (Fig. (Fig.3,3, ,4,4, and and5).5). These results suggested that in HPV31b-infected cells, CDK1 kinase activity is targeted by BaP in a differentiation-dependent manner and that increased CDK1 kinase activity in HPV31b raft tissues (Fig. (Fig.2B)2B) is correlated with induction of high levels of virion synthesis (Fig. (Fig.1)1) (2). In addition, differentiation-dependent cell signaling pathways may act in tandem with BaP-regulated growth modulatory signals, which cumulatively result in the observed profile of cell cycle tumor suppressors and kinase activities in raft tissues, and which in turn correlate with high levels of virion synthesis.
For an additional control, we wanted to determine whether the BaP-induced expression profile of tumor suppressors in the HPV31b raft cultures was specific to the presence of HPV. Primary human keratinocytes (HKs) were isolated as previously described (67), followed by their growth and treatment with BaP in raft cultures for 12 days. Raft tissues were harvested, whole-tissue extracts were prepared, and pRb, p16INK4, p27KIP1, p21WAF1, and p53 were detected by Western blotting as described above. We determined that in the HPV-negative tissues, pRb expression levels were eliminated in response to BaP treatment (Fig. (Fig.6A),6A), in contrast with HPV31b raft cultures treated with BaP (compare Fig. Fig.2A2A with Fig. Fig.6A6A).
BaP treatment also increased p16INK4 protein expression levels compared with controls in HPV-negative cultures similar to that observed in HPV31b-positive raft tissues treated with BaP (compare Fig. Fig.2A2A with Fig. Fig.6A).6A). We were unable to detect p21WAF1 and p27KIP1 protein expression in HK raft cultures (Fig. (Fig.6A).6A). The failure to detect p21WAF1 is presumably due to the differentiation-dependent downregulation of the expression of this protein as has been previously described elsewhere (56). Our failure to detect p27KIP1 protein in HPV-negative raft tissue extracts may also be due to downregulation of this protein in a differentiation-dependent manner utilizing similar mechanisms. The p21WAF1- and p27KIP1-specific antibodies used in the current study have been used previously to demonstrate abundant expression of both species in HK monolayer cultures as previously reported (3). In contrast, stabilized p21WAF1 and p27KIP1 protein levels in HPV-infected monolayer cultures as well as in differentiating tissues is known to be due to binding of the HPV E7 oncoprotein to the carboxy termini of both tumor suppressors (23, 24). We tested the effect of BaP treatment on p53 expression in HK raft tissues and found that its levels of expression were also unchanged in response to BaP (Fig. (Fig.6A).6A). Collectively, our results from this set of experiments suggest that BaP targeting of pRb steady-state levels are common events in HPV-infected as well as HPV-negative tissues but with opposite consequences. Also, BaP-mediated upregulation of p16INK4 protein expression appears to be a common event in both tissue types. The BaP-regulated differences in cell cycle protein expression observed in the two tissue types could be due to the presence of HPV-specific proteins.
Since BaP treatment was also shown to mediate pRb loss in HK raft cultures (Fig. (Fig.6A),6A), we asked whether the presence or absence of HPV in BaP-treated tissue played a role in CDK1 activation which could be correlated with the observed loss of pRb in these tissues (Fig. (Fig.6A).6A). We treated HK raft cultures with 1 μM BaP for 12 days and then determined the kinase activity associated with CDK4, CDK6, CDK2, and CDK1 as described above. We observed a robust upregulation of kinase activity associated with CDK4 and CDK6 upon BaP treatment (Fig. (Fig.6B),6B), which could be positively correlated with the loss of pRb in these tissues (Fig. (Fig.6A),6A), as well as with increased total protein levels of these proteins (Fig. (Fig.6C).6C). On the other hand, CDK2-associated kinase activity was unchanged upon BaP treatment compared with controls (Fig. (Fig.6B),6B), but the levels of total CDK2 protein in BaP-treated tissues were elevated (Fig. (Fig.6C).6C). Finally, we determined the effect of BaP treatment on modulating CDK1-associated kinase activity and observed that BaP treatment of primary tissues also resulted in upregulation of CDK1-associated kinase activity (Fig. (Fig.6B6B).
The results from this set of experiments suggest that in HPV-negative tissues, BaP targeting of pRb expression is correlated with the simultaneous activation of CDK4/CDK6 as well as CDK1-associated kinase activities. In HPV-negative tissues, BaP stimulation of CDK4/CDK6 kinase activities (Fig. (Fig.6B)6B) was observed in contradiction to the simultaneous increase in p16INK4 protein levels (Fig. (Fig.6A),6A), suggesting that BaP targeted and deregulated kinase inhibitory activities of p16INK4 in these tissues. In addition, BaP-targeted modulation of CDK1 kinase activity was common to differentiating keratinocytes irrespective of the presence of HPV (compare Fig. Fig.2B2B with Fig. Fig.6B).6B). These results suggest the possibility that BaP targeting of increased CDK1-associated kinase activity is a common, yet fortuitous, cellular event which is correlated with high levels of HPV titers (Fig. (Fig.1)1) (2). These experiments have been repeated multiple times with reproducible results.
The in vitro organotypic raft culture system faithfully reproduces the three-dimensional architecture of fully stratified and differentiated epithelial tissue (5, 22, 49, 50, 52, 53, 58), and we have previously demonstrated that the complete HPV life cycle is intimately associated with host tissue differentiation (53). We have also established that differentiation of raft tissue derived from both low-grade or invasive carcinomas is characterized by expression of physiologic markers of keratinocyte differentiation (58). We wanted to determine whether BaP upregulation of virion synthesis could be correlated with changes in the patterns of differentiation markers. We treated HPV31b and HK raft cultures with BaP over a period of 12 days and then determined the total protein levels of the keratinocyte differentiation markers involucrin, K-10, and K-14. Normally, expression of involucrin and K-10 proteins are simultaneously induced in the spinous layer of the differentiating epithelium (22). In contrast, K-14 is expressed only in the basal layers (22). We found that both primary and HPV-infected raft tissues treated with BaP displayed marked upregulation of involucrin expression levels, whereas K-10 expression was severely diminished at this concentration (Fig. 7A and B). Expression of K-14 was unaltered in 1 μM BaP-treated HPV31b and HK raft cultures (Fig. 7A and B). These results are interesting, since in BaP-treated HPV31b raft tissues, the increased involucrin to decreased K-10 ratio (Fig. (Fig.7A)7A) appears to be a marker which could be correlated with enhanced virion synthesis (Fig. (Fig.1)1) (2), but it could also be a marker for cigarette carcinogen exposure in tissues not infected with HPV (Fig. (Fig.7B7B).
We further wanted to determine whether the process of tissue differentiation determined BaP regulation of keratin expression. Therefore, we repeated the 1 μM BaP treatment of monolayer cultures, followed by Western blot analysis to look at the expression levels of the differentiation markers studied. Monolayer (nondifferentiating) cultures of HPV31b-positive cells were synchronized as described herein, followed by treatment of the cells with 1 μM BaP over a period of 7 days. On each day, control and BaP-treated cells were trypsinized, pelleted, and frozen in −70°C until further manipulations. The cultures were passaged at a ratio of 1:2 on day 2 and day 5 followed by the addition of fresh BaP at these times. We then performed Western blot analysis to compare the levels of expression of differentiation markers in BaP-treated cells and control cells. In contrast to the pattern observed in raft tissues, monolayer HPV31b cultures treated with BaP showed a different pattern of expression. Under nondifferentiating conditions, involucrin expression remained unresponsive to BaP treatment (Fig. (Fig.7C),7C), whereas K-10 expression was decreased in BaP-treated cells, similar to results observed in BaP-treated HPV31b raft cultures (compare Fig. Fig.7A7A with Fig. Fig.7C).7C). On the other hand, expression of K-14 did not demonstrate a specific pattern of expression in BaP-treated monolayer cultures (Fig. (Fig.7C).7C). Thus, BaP modulation of involucrin appears to be controlled via differentiation-dependent signaling pathways. Interestingly, expression of the three differentiation markers tested here have been shown to be transcriptionally regulated by proteins of the growth regulatory signaling pathways which also control cell cycle regulation (44). Our data suggest that differentiation-dependent signaling pathways act in concert with BaP-targeted growth regulatory pathways.
Finally, it was of interest to correlate the spatial expression pattern of these differentiation markers in three-dimensional tissue with BaP-regulated increase in HPV31b virion synthesis. We prepared paraffin-embedded sections derived from 12-day BaP-treated and control raft tissues derived from HPV-infected as well as HPV-negative keratinocytes. We first determined the overall morphology of the tissues by staining them with hematoxylin and eosin. There appeared to be no morphological differences between the two tissue types (Fig. (Fig.7D).7D). We then performed immunohistochemical staining of the tissue sections with monoclonal antibodies against involucrin and K-10 (Fig. (Fig.7D).7D). The decreased total levels of expression of K-10 in HPV31b- and HPV-negative raft cultures treated with BaP as determined from Western blots (Fig. 7A and B) was also reflected by the lack of K-10 staining in these sections (Fig. (Fig.7D).7D). The qualitative difference between the results of in situ staining and Western blotting is due to the lower sensitivity of the former technique compared with the latter. On the other hand, the spatial expression pattern of involucrin in HFK raft tissues treated with BaP appeared to be similar to that of the controls. The decrease in K-10 expression may provide a potential “signature fingerprint” to identify cervical tissues from biopsy samples from patients exposed to cigarette smoke. These experiments have been repeated multiple times with reproducible results.
In summary, BaP modulation of cell cycle- and differentiation-dependent signaling pathways provide specific profiles which could be correlated with BaP exposure and which correlate with high levels of virion synthesis in raft cultures. These results are important, since cellular changes which abrogate the process of differentiation are generally correlated with progression (1).
Thus far, our data show that BaP-mediated upregulation of HPV31b morphogenesis is correlated with an altered balance between cell cycle-specific CDK kinase activities compared with control tissues (compare Fig. Fig.11 with Fig. Fig.2B).2B). BaP treatment increased CDK1 kinase activity, but CDK6 and CDK2 kinase activities were decreased, whereas CDK4 kinase activity remained unchanged (Fig. (Fig.2B).2B). The observed increase in CDK1 kinase activity could be a compensatory response to the simultaneous loss of both CDK6 and CDK2 kinase activities, which could be important for one or more steps in the viral life cycle. To determine the importance of increased CDK1 kinase activity in HPV31b virus synthesis in the presence of BaP, we repeated the experiments in the presence of purvalanol A, a specific CDK1 inhibitor (50% inhibitory concentration [IC50] of 4 nM for CDK1/cyclin B), which competes with ATP for binding to the catalytic site of the kinase. Preliminary screening of purvalanol A showed that treatment with this inhibitor grossly affected tissue growth when added throughout the 12-day growth period of raft tissues (data not shown). Therefore, we started purvalanol A treatment beginning on day 8 and terminating on day 12 at which time the tissues were harvested. In our hands, tissue growth and stratification are readily visible starting on day 8 after the dermal equivalents were lifted onto the air-liquid interface. This growth period also coincides with early steps of virion morphogenesis (16, 57). When added during this time span, addition of the CDK1 inhibitor was not inhibitory to tissue growth. Tissues receiving purvalanol A treatment starting on day 8 received fresh purvalanol A inhibitor on day 10. We also examined the effect of purvalanol A treatment on day 10 and on day 11. Raft cultures were allowed to differentiate up to day 12, at which point the tissues were harvested. We also determined that a final concentration of 10 nM purvalanol A was the smallest amount which could be used to significantly inhibit CDK1 kinase activity in raft tissues (data not shown).
We repeated BaP treatment of HPV31b raft cultures in the presence of 10 nM purvalanol A. CDK1 kinase activity was determined by incubating immunoprecipitated complexes containing CDK1 with histone H1 as a substrate. Again, raft tissues treated with 1 μM BaP exhibited increased CDK1 kinase activity compared with control and DMSO-treated tissues (Fig. (Fig.8A,8A, compare lanes 1, 2, and 3). Purvalanol A inhibited CDK1 kinase activity to levels below those found in both control and BaP-treated tissues, irrespective of the time span the cultures were exposed to purvalanol A (Fig. (Fig.8A,8A, compare lanes 1 to 3 with lanes 4 to 9). To ensure that inhibition of CDK1 kinase activity was not disruptive to tissue stratification, we performed histochemical analysis of BaP-treated and control tissues using hematoxylin and eosin stain (Fig. (Fig.8B).8B). Purvalanol A treatment did not interfere with general growth/proliferation and differentiation of tissues during the time points tested. In order to determine whether purvalanol A affected cell growth functions in differentiating tissues, we stained identical tissue sections using a rabbit polyclonal antibody against proliferating cell nuclear antigen (PCNA) to stain active nuclei (Fig. (Fig.8C).8C). The pattern of nuclei expressing PCNA in control tissues was similar to the patterns in tissues treated with both BaP and purvalanol A or treated with BaP or purvalanol A alone (Fig. (Fig.8C8C).
Our next step was to determine the consequences of CDK1 inhibition on HPV31b viral titers in BaP-treated tissues. HPV31b raft tissues were treated with 1 μM BaP only (over the 12-day growth period) or treated with 1 μM BaP (12-day treatment) and 10 nM purvalanol A (added on day 8, day 10, and day 11). In addition, raft tissues were also treated with 10 nM purvalanol A alone (added on day 8, day 10, and day 11). Virus stocks were prepared, and the titers of infectious virus were determined. Viral titers were determined by infecting HaCaT cells with serially diluted viral stocks, followed by performing infectivity assays to amplify the HPV31b E1^E4 spliced transcript cDNA as we have previously described (2). For controls, primers against β-actin were included in the PCR mixture as described previously (2). PCR products were visualized by electrophoresis in a 2% agarose-ethidium bromide gel. Treatment with 1 μM BaP resulted in a 10-fold increase in HPV31b viral titers compared with those of control cultures (Fig. (Fig.9A,9A, compare lanes 1 and 3) as we have previously reported (2) (Fig. (Fig.1).1). Inhibition of CDK1 kinase activity in control tissues not exposed to BaP did not interfere with the production of infectious virus, since viral titers in raft tissues treated with purvalanol A alone were the same as medium-only and DMSO-treated cultures (Fig. (Fig.9A,9A, compare lanes 4 to 6 to lanes 1 and 2). In contrast, viral stocks prepared from raft tissues treated with both 1 μM BaP and 10 nM purvalanol A were almost devoid of infectious virus (Fig. (Fig.9A,9A, compare lanes 7, 8, and 9). Infectious virus was detected in purvalanol A-treated tissues (added day 8 to day 12) when used at a 1/20 dilution (Fig. (Fig.9A,9A, lane 7), but this result was not reproducible in different experiments. To rule out the possibility that the failure of these virus stocks to infect HaCaT cells was due to a carryover inhibitory effect of residual purvalanol A on these cells, we determined CDK1-associated kinase activity of HaCaT cells infected with the lowest dilution of the virus stocks (1/20 dilution). CDK1 kinase activities found in HaCaT cells infected with viral stocks treated with purvalanol A were similar to infections using medium-only and control DMSO-treated virus stocks (Fig. (Fig.9B).9B). In conclusion, these results suggest that BaP-induced increase in CDK1 kinase activity was important for conferring infectivity on HPV31b virions propagated in the presence of BaP.
Exposure of HPV31b virus to BaP over the 12-day growth period resulted in synthesis of virus which was dramatically dependent on CDK1 kinase activity for retaining its infectious properties (Fig. (Fig.9A).9A). These alterations were specific to BaP treatment, since inhibition of CDK1 kinase activity alone in control tissues was not sufficient to interfere with the production of infectious virions (Fig. (Fig.9A,9A, lanes 4 to 6). Changes in functional aspects of the viral capsids necessary for infection could explain these observations. One obvious possibility was that the failure of virus that had been treated with BaP and purvalanol A to infect HaCaT cells (Fig. (Fig.9A,9A, lanes 7 to 9) was either due to disassembly or lack of assembly of these particles in response to inhibition of CDK1 kinase activity. We based this possibility on reported studies which predicted that hepadnaviruses contain a cdc2 kinase-like recognition motif on their core proteins which regulate assembly and disintegration of capsids (7).
To examine the possibility that the failure of virus that had been treated with BaP and purvalanol A to infect HaCaT cells was due to either disassembly or lack of assembly of these particles in response to inhibition of CDK1 kinase activity, a qPCR-based DNA encapsidation assay was used to quantify total endonuclease-resistant viral genomes using methods we have previously described (16). This assay was based on the expectation that all viruses produced in untreated tissues are infectious and are units of encapsidated viral genomes. HPV31b raft tissues were treated with 1 μM BaP and/or cotreated with 10 nM purvalanol A as described above, harvested, and used for preparing viral stocks using protocols we have previously described (16). Viral stocks were incubated with benzonase for digestion of exogenous cellular DNA and viral DNA, and viral genomes were extracted from capsids also as previously described (16). We have previously reported that benzonase treatment of crude virus stocks effectively eliminates all chromatin and exogenous viral genomes (15, 16). Quantification of isolated HPV31b genomes was performed using primers specific to HPV31 targets, followed by SYBR green-based qPCR. Though small variations in levels were observed (discussed further below), endonuclease-resistant genomes were detected in all viral stocks tested, regardless of BaP and/or purvalanol A treatment (Fig. (Fig.9C).9C). Also, the levels of endonuclease-resistant genomes were similar among all viral stocks that received BaP and purvalanol A cotreatment. Similarly, the levels of endonuclease-resistant genomes were similar among all viral stocks that received purvalanol A treatment alone compared with medium-only and DMSO-treated raft cultures (Fig. (Fig.9C).9C). These results suggest that loss of infectivity of viral stocks treated with BaP and purvalanol A (Fig. (Fig.9A)9A) was not due to capsid disassembly or lack of assembly but rather resulted from improper capsid formation. To summarize, from these experiments, a novel observation was that virus exposed to BaP was highly dependent on the BaP-mediated increase in CDK1 kinase activity for maintaining infectivity (Fig. (Fig.9A,9A, lanes 7 to 9). Inhibition of CDK1 kinase activity interfered with infectivity of the virus, regardless of whether it was added during the middle stage (day 8) or late stage (day 10 and day 11) of tissue differentiation (Fig. (Fig.9A,9A, lanes 7 to 9), suggesting that virus-related targets of the CDK1 kinase inhibitor were available during these stages of virion morphogenesis in the presence of BaP.
We also noted that treatment with the CDK1 inhibitor alone did not affect infectivity of control virus (Fig. (Fig.9A,9A, lanes 4 to 6), suggesting that targets of CDK1 kinase were unavailable in the absence of BaP treatment. However, tissues treated with purvalanol A alone (days 10 to 12) had twice the number of encapsidated genomes than tissues treated with the inhibitor on days 8 to 12 or days 11 to 12 (Fig. (Fig.9C,9C, compare bar 4 to bars 5 and 6). These results suggest the possibility that during late stages of HPV31b morphogenesis, CDK1 kinase activity interferes with events necessary for proper morphogenesis. However, an increase in the number of encapsidated genomes did not correlate with increased infectivity of these virions, which was similar to controls (compare Fig. Fig.9C,9C, bar 5, to Fig. Fig.9A,9A, bar 5). In parallel, DMSO treatment also generated about twice the amount of encapsidated genomes than medium-only controls (Fig. (Fig.9C,9C, compare bars 1 and 2). Mild oxidizing activity of DMSO could accelerate virion maturation and formation of more nuclease-resistant particles in treated tissues compared to control tissues. This observation is supported by our previously published studies which showed that virion stability and infectivity are redox dependent within three-dimensional tissue (16).
Epidemiological studies demonstrated a close correlation between cigarette smoking and the incidence of cervical cancer (81) based on the positive relationship observed between smoking habits and abnormal cervical cytology (11, 30, 31, 54), cervical intraepithelial neoplasia grade 3, and cervical cancer (30). These studies proposed that carcinogens present in cigarette smoke modify important cellular proteins which in turn act as cofactors in cervical cancer progression (27). Our previously published studies demonstrated that BaP exposure has direct dosage-dependent effects on identifiable HPV life cycle events, such as changes in genome amplification and virion synthesis (2). Both situations pertain to creating an increase in the total viral load with respect to both the quantity of virus synthesized as well as increased number of genome copies. Our published studies presented a novel hypothesis suggesting that BaP-regulated enhancement of both virus synthesis and amplification of genome copies may potentially result in increased persistence of the virus in infected women who smoke. Since persistence of the virus is thought to be correlated with progression (82), we further hypothesize that BaP exposure and the consequent increase in viral burden could be an early milestone in the initiation of cervical cancer.
BaP has been shown to promote nongenotoxic effects independent of its capacity to form DNA adducts (66) and has been shown to result in deregulation of cell proliferation pathways in human mammary epithelial cells (77). We asked whether BaP regulation of carcinogenesis in HPV-infected keratinocytes via nongenotoxic signaling pathways could be correlated with the observed induction of increased virion morphogenesis. Treatment of HPV31b-infected raft tissues with BaP resulted in aberrant activation of CDK1 kinase activity, which is consistent with activation of G2-phase-specific functions, but with a simultaneous imposition of G1-like growth arrest signals. Thus, BaP regulation of apparently conflicting growth signals in cell cycle processes appears to be an environment which is conducive for high levels of HPV31b virus synthesis. In addition, aberrant regulation of cell proliferation signals, including modulation of CDK kinase activities, presents potential for progression.
Since increased CDK1 kinase activity was positively correlated with induction of high levels of viral morphogenesis, an interesting question was whether activity of this kinase is somehow useful for one or more aspects of the HPV life cycle. One study showed that the HPV16 E1^E4 protein binds to and retains active CDK1/cyclin B1 complexes in the cytoplasm, thus implementing cellular conditions resembling a G2 arrest (18). Retention of active CDK1/cyclin B1 complexes in the cytoplasm prevents entry of these complexes into the nucleus and concomitant triggering of mitosis (18). The authors speculated that the apparent G2 arrest may induce cellular conditions optimized for specific viral functions, including late events (18), which may be a mechanism utilized by multiple viruses (17).
In the current study, it is not clear how CDK1 kinase activity is required for the normal HPV life cycle except perhaps for a possible role in virus maturation. However, a novel observation was that BaP treatment increased CDK1 kinase activity, whereas CDK6 and CDK2 kinase activities were simultaneously decreased, thus greatly altering the balance of CDK kinase activities compared with control tissues. Under these conditions, BaP-treated virions were rendered dramatically dependent on CDK1 kinase for retaining their infectious properties. These results suggest the possibility that BaP treatment abrogated novel virus-specific and/or cellular CDK1 targets which determine virion infectivity. BaP treatment could interfere with capsid-associated functions and consequently alter the infectivity of BaP-treated virions in a CDK1 kinase-dependent fashion. Others have shown that cellular metabolism of BaP causes oxidative damage due to the generation of free radicals (80), which affects biochemical properties of multiple proteins, leading to misfolding and aggregation (46). BaP-regulated changes in L1 and L2 capsid protein conformation/integrity could potentially result in increased exposure of serine/threonine residues on the capsids to a more accessible environment for phosphorylation by the CDK1 kinase. Subsequent phosphorylation of those residues could be important for events regulating HPV infectivity. We also hypothesize that in non-BaP-treated control virus, these residues could be sequestered inside the capsids and perhaps not as accessible to CDK1-dependent phosphorylation and therefore were not amenable to purvalanol A-mediated inhibition of their infectivity, as observed in the current study.
Involvement of CDK1 kinase in viral morphogenesis could be a widespread mechanism in a number of different viruses. For example, hepadnaviruses have been shown to regulate capsid assembly and disassembly depending on a cdc2/CDK1 kinase-like recognition motif on the core protein at a location that is required for the assembly of core polypeptides into capsids (7). A related study shows that the varicella-zoster virus immediate-early transactivator IE62 was phosphorylated by CDK1/cyclin B1 (41). The IE62 protein was also shown to colocalize with CDK1/cyclin B1 complexes in the viral tegument for delivery of the active cellular kinase to nondividing cells (41). A related but divergent example is the yeast retrotransposon Ty1. The Saccharomyces cerevisiae yeast Fus3 mitogenic protein kinase phosphorylates the Ty1 capsids and triggers degradation of the viruslike particles and subsequently suppresses transposition of the virus (14, 84).
The induction of an apparent G1-like block in BaP-treated HPV31b raft tissues identified by upregulated p16INK4 and p27KIP1 protein levels was notable. BaP regulation of high levels of p16INK4 protein may also be due to our observed inactivation of pRb. Since p16INK4 expression is dependent on a negative-feedback control through pRb (37, 71), reduced or lost pRb function should result in amplification of p16INK4 levels. Our control experiments also suggested that BaP regulation of pRb stability and upregulated p16INK4 expression were independent of HPV and could be a dual marker for carcinogen exposure and dysplastic progression. Our studies also showed that BaP modulation of the increased involucrin to decreased K-10 expression ratio was similar in both HPV31b and HK tissues, suggesting that signaling pathways that regulate the expression of differentiation-related proteins targeted by BaP are not dependent on the presence or absence of HPV. On a similar note, BaP-regulated loss of pRb in HPV-negative tissues could be correlated with increased CDK1, CDK4, and CDK6-associated kinase activities. Our current studies suggest the possibility that BaP control of aberrant cell cycle regulation in HPV-negative tissues may present with highly proliferative conditions which could be a priming factor for establishing HPV infections in noninfected women who smoke but who may be exposed to the virus in the future.
An important question is how differential BaP regulation of the cell cycle and differentiation pathways translates to high levels of HPV synthesis and how these pathways could potentially implement carcinogenic progression. In the current study, BaP regulation of increased CDK1 kinase activity was common in both HPV31b-positive and HK tissues. Thus, in addition to activating CDK1-containing complexes, cell cycle and growth regulatory proteins that control activation of those cyclin B-CDK1 complexes are also potential targets of BaP. We are currently in the process of characterizing these pathways. The virus is a carcinogenic entity in its own right, and also, BaP is a well-characterized carcinogen. Activation of dysregulated mitotic entry in the midst of cell cycle exit cues may play a role in BaP stimulation and/or stabilization of late gene transcripts, capsid proteins, and concomitant viral morphogenesis. Thus, abrogated signaling in multiple cell cycle phases may contribute to BaP-mediated carcinogenesis of host tissue and could subsequently aid in the persistence of the virus.
Our current study provides another molecular link to define the effects of a known tobacco carcinogen to cell cycle regulatory functions that affect the life cycle of HPV. We have also identified BaP targets that affect host cell cycle control as well as tissue differentiation. In addition to identifying p16INK4, a known biomarker for progression (39), we have potentially identified other proteins that could potentially be developed as biomarkers, including hyperphosphorylated pRb, p27KIP1, and increased involucrin to decreased K-10 ratios. A dominant player which emerged in our current studies was the CDK1 protein, the kinase activity of which appeared to be targeted by BaP and could regulate proliferation of keratinocytes. Recent studies have established that specific activity of CDK1 and CDK2 could be used as novel prognostic markers for early breast cancer detection (38), since patient biopsy specimens displaying high levels of kinase activities were strongly associated with unfavorable prognosis (38). Our data suggest the possibility that BaP modulation of increased CDK1 kinase activity in HPV-positive lesions could favor persistent HPV infections and permissiveness for cancer progression. However, BaP regulation of cell cycle deregulation in primary tissues may increase host susceptibility to future HPV infections. BaP regulation of multiple growth regulatory pathways potentially determines the dynamics of cell cycle control and differentiation in HPV-infected keratinocytes, which greatly influences multiple viral functions and cellular proliferation controls.
We thank Lynn Budgeon for excellent technical assistance.
Research described in this article was supported by Philip Morris USA Inc. and Philip Morris International and in part by a PHS grant from the National Institute of Allergy and Infectious Disease (R01AI57988) to C.M.
Published ahead of print on 24 February 2010.