HPV-16 E6/7 Immortalization Sensitizes Human Keratinocytes to Ultraviolet B by Altering the Pathway from Caspase-8 to Caspase-9-dependent Apoptosis*

UVB from solar radiation is both an initiating and promoting agent for skin cancer. We have found that primary human keratinocytes undergo an apoptotic response to UVB. To determine whether these responses are altered during the course of immortalization, we examined markers of apoptosis in primary human foreskin keratinocytes (HFK) transduced with either a retroviral vector expressing the E6 and E7 genes of HPV-16 or with empty vector alone (LXSN-HFK). Whereas LXSN-HFK as well as early passage keratinocytes expressing HPV-16 E6 and E7 (p7 E6/7-HFK) were both moderately responsive to UVB irradiation, late passage-immortal-ized keratinocytes (p27 E6/7-HFK) were exquisitely sensitive to UVB-induced apoptosis. After exposure to UVB, enhanced annexin V-positivity and internucleosomal DNA fragmentation were observed in p27 E6/7-HFK compared with either LXSN- or p7 E6/7-HFK. Caspase-3 fluorometric activity assays as well as immunoblot analysis with antibodies to caspase-3 and poly(ADP-ribose) polymerase revealed elevated caspase-3 activity and processing at lower UVB doses in p27 E6/7-HFK compared with LXSN-

The most common malignancy in humans is skin cancer. The incidences of basal cell carcinoma, squamous cell carcinoma, and melanoma continue to rise and approach those of all other cancer subtypes combined (1). UV 1 irradiation causes skin cancer through a series of cellular changes that are not all identified. However, since UV acts as a promoting (selective) as well as an initiating (mutating) agent (2), it is clear that in addition to genetic alterations, inappropriate or altered growth, differentiation, and/or apoptotic responses to UV play key roles in this process.
The connection between UV-induced apoptosis and skin cancer has been well studied in the context of p53. UV induces a signature pattern of p53 mutations in squamous cell carcinoma and basal cell carcinoma (3,4) as well as in normal sun-exposed skin (5). Keratinocytes harboring these p53-inactivating mutations are resistant to senescence in culture, and it has been further proposed that such keratinocytes are also resistant to growth inhibition or cytotoxicity from subsequent UV exposure (6). In support of this idea, clones of p53-mutated keratinocytes can be found within normal sun-exposed human epidermis (2,5) and can be generated in mice in which clonal expansion of p53-mutant keratinocytes is continually driven by UVB (7).
Although mucosal HPV types have been long implicated in anogenital cancer, a variety HPV types have been identified in a high percentage of basal cell carcinoma and squamous cell carcinoma in immunosuppressed patients (8) as well as in actinic keratoses and squamous cell carcinoma in those with the inherited disorder epidermodysplasia verruciformis (EV; Ref. 9). HPV has also been postulated to play a role in basal cell carcinoma and squamous cell carcinoma from immunocompetent non-EV patients as well (8,10). In most cases, carcinomas occur in sun-exposed sites, indicating cooperation between UV and HPV. Because the HPV E6 gene product inactivates p53 (for review, see Ref. 11), E6 presumably serves some of the same functions as UV-induced p53-inactivating mutations in skin carcinogenesis (12).
HPV E6 and E7 oncogenes likely play roles in the early immortalization stage of carcinogenesis, resulting ultimately in the stable activation of the telomerase catalytic subunit (hTERT) and inactivation of the p16/Rb pathway (13,14). In cell culture, spontaneous immortalization of-keratinocytes is a rare event that can be enhanced by HPV-16/18 E6 and E7. Although HPV E6 has been shown to transiently up-regulate the gene and promoter of hTERT (15,16) and E7 inactivates Rb, other genetic events are apparently required for immortalization. These other genes may be related to the stable expression of hTERT (17), since loss of a region of chromosome 6 derepresses telomerase expression in HPV-immortalized cells. Steenbergen et al. (18) also observed non-random allelic losses at 3p, 11p, and 13q during HPV-mediated immortalization, whereas Poignee et al. (19) displayed evidence for loss of a senescence locus within the chromosomal region 10p14-p15 in human foreskin keratinocyte (HFK)-expressing HPV-16 E6/7 genes.
Many of the previous studies comparing the response of primary human keratinocytes to their immortalized counterparts have utilized cells that are tumor-derived or that have been cultured for long periods of time. To directly examine the effects of HPV E6/7 as well as additional early immortalizing events on the response of primary keratinocytes to UVB, we transduced HFKs with a retroviral vector expressing HPV-16 E6 and E7 or with the empty vector alone (LXSN-HFK). We then compared the UVB-induced apoptotic response of early passage cells expressing E6/7 (p7 E6/7-HFK) as well as those passaged just long enough to become immortalized (p27 E6/7-HFK). We found that p27 E6/7-HFK were much more sensitive to UVB-induced apoptosis than either LXSN-or p7 E6/7-HFK. Furthermore, all three cell types were shown to undergo caspase-3-dependent apoptosis. Examination of the upstream caspases-8 and -9 revealed that whereas caspase-8 is activated in all three cell types, caspase-9 is only activated after UV exposure of p27 E6/7-HFK. Cell cycle analysis also revealed that only p27 E6/7-HFK underwent UVB-induced accumulation in G 2 /M as well as a marked reduction in the levels of immunodetectable Bcl-2. These results suggest that the entire immortalization process rather than the expression of E6 and E7 alone is critical to the increased apoptotic response, which involves a switch from a caspase-8-to a caspase-9-dependent pathway. These results indicate that immortalization represents a "transition" stage and that UVB resistance is the result of further genetic alterations that occur subsequent to immortalization.

MATERIALS AND METHODS
Cells-Primary human keratinocytes were derived from neonatal foreskins and grown in KSF medium supplemented with human recombinant epidermal growth factor and bovine pituitary extract (Invitrogen). The primary cells were infected with an amphotropic LXSN retrovirus expressing the HPV-16 open reading frames of the E6 and E7 genes. The retrovirus was generated as described (20) with the use of existing recombinant vectors (21). Retrovirus-infected cells were selected in G418 (100 g/ml) for 10 days, and G418-resistant colonies were pooled from each transduction and passaged every 3-4 days.
Keratinocytes (from different passages after infection with HPV-16 E6/7) were grown under identical conditions to 70 -80% confluency, trypsinized before UVB exposure, and passaged at equal cell densities. Cells were allowed to recover, and after replacement of the KSF medium with phosphate-buffered saline, cells were irradiated with ultraviolet light using a UVB source with a peak wavelength of 312 nm (FS40 sunlamp (Philips) with a Kodacel filter (Eastman Kodak Co.)) at various doses. At different time points after UVB irradiation or 16 h after exposure to different doses of UVB cells were derived for further analyses.
Fluorometric Assay of Caspase-3 Activity-Cells were resuspended in lysis buffer containing 50 mM Tris HCl (pH 7.5), 150 mM NaCl, 1 mM EGTA, 0.25% sodium deoxycholate, 0.5% Nonidet P-40 (Nonidet P-40), 10 g/ml aprotinin, 20 g/ml leupeptin, 10 g/ml pepstatin, and 1 mM phenylmethylsulfonyl fluoride, incubated for 10 min on ice, and freezethawed 3 times. The cell lysate was centrifuged at 14,000 ϫ g for 5 min, and the protein concentration of cytosolic extract was determined with the Bio-Rad DC protein assay kit. For the fluorometric caspase-3 activity assay, 25 g of cytosolic extract was initially resuspended up to a volume of 50 l with Nonidet P-40 lysis buffer, to which 50 l of caspase assay buffer (10 mM HEPES (pH 7.4), 2 mM EDTA, 0.1% CHAPS, 5 mM dithiothreitol) was added. The aliquots were then mixed with an equal amount (100 l) of 40 M fluorescent tetrapeptide substrate specific for caspase-3 (Ac-DEVD-AMC; Bachem) in caspase assay buffer and transferred to 96-well plates. Free aminomethylcoumarin (AMC), generated as a result of cleavage of the aspartate-AMC bond, was monitored continuously for 30 min with a Cytofluor 4000 fluorometer (PerSeptive Biosystems, Framingham, MA) at excitation and emission wavelengths of 360 and 460 nm, respectively. The emission from each well was FIG. 1. Exposure of HFK to UVB results in nuclear fragmentation, caspase-3 activation, and PARP cleavage characteristic of apoptosis. HFK were derived from neonates as described under "Materials and Methods." Cells were irradiated with 480 J/m 2 UVB, and 16 h later, cells were fixed and stained with Hoechst (A). B, cells were irradiated with 480 J/m 2 UVB, and after the indicated times, whole cell extracts were prepared and subjected to immunoblot analysis using antibodies specific for the active form (p17) of caspase-3 or PARP.
plotted against time, and linear regression analysis of the initial velocity (slope) for each curve yielded the activity.
Cell Cycle Analysis-Nuclei were prepared for flow cytometric analysis as described (22). Cells were exposed to trypsin, resuspended in 100 l of a solution containing 250 mM sucrose, 40 mM sodium citrate (pH 7.6), and 5% Me 2 SO, and subsequently lysed in a solution containing 3.4 mM sodium citrate, 0.1% Nonidet P-40, 1.5 mM spermine tetrahydrochloride, and 0.5 mM Tris-HCl (pH 7.6). Lysates were incubated for 10 min with RNase A (0.1 mg/ml), after which nuclei were stained for 15 min with propidium iodide (0.42 mg/ml), filtered through a 37-m nylon mesh, and analyzed with a dual-laser flow cytometer (FACScan, BD PharMingen).
Analysis of DNA Fragmentation-Cells were harvested and lysed in 0.5 ml of 7 M guanidine hydrochloride. The lysate was mixed with 1 ml of Wizard Miniprep resin (Promega, Madison, WI), incubated at room temperature for 15 min with occasional mixing, and then centrifuged at 10,000 ϫ g for 5 min. The resulting pellet was resuspended in 2 ml of washing solution (90 mM NaCl, 9 mM Tris-HCl (pH 7.4), 2.25 mM EDTA, 55% (v/v) ethanol) and drawn by vacuum through a Wizard Minicolumn

FIG. 2. Immortalization of HFK with HPV-16 E6/E7 results in a UVB dose-dependent increase in annexin V-positive cells and decreased survival.
LXSN-HFK, p7 E6/7-HFK, and p27 E6/7-HFK were prepared as described under "Materials and Methods." Cells were irradiated with the indicated doses of UVB, and 16 h later, cells were prepared, and extracts were subjected to immunoblot analysis using antibodies for p53 or E7 (A) or assayed for annexin V binding plus PI staining at the indicated doses by FACS analysis (B and C). The percentage of cells exhibiting annexin V binding (B) or that were negative for annexin V binding PI staining (C) as determined by FACS analysis are shown. All the data in B and C are presented as the means Ϯ S.D. of three replicates of a representative experiment; essentially the same results were obtained in three independent experiments.  4. Immortalization of HFK with HPV-16 E6/E7 increases UVB-dependent caspase-3 activity. LXSN-HFK, p7 E6/7-HFK, and p27 E6/7-HFK were prepared as described under "Materials and Methods." Cells were irradiated with the indicated doses of UVB and assayed for caspase-3 activity 16 h after UVB exposure using a quantitative fluorometric assay. All the data are presented as the means Ϯ S.D. of three replicates of a representative experiment; essentially the same results were obtained in three independent experiments.
(Promega) mounted onto a vacuum manifold. The column was washed twice with 4 ml of washing solution and dried by centrifugation at 10,000 ϫ g for 2 min. DNA was eluted from the column by the addition of 50 l of deionized H 2 O, incubation at room temperature for 15 min, and then centrifugation at 10,000 ϫ g for 5 min. Residual RNA in the eluate was removed by incubation with 10 g of RNase A (5 Prime 3 3 Prime, Inc., Boulder, CO) at 37°C for 30 min. DNA samples were loaded onto a 1.5% agarose gel in Tris borate-EDTA buffer and subjected to electrophoresis at 4 V/cm. DNA ladders were visualized by staining with ethidium bromide (0.5 g/ml), and images were captured with the Kodak EDAS 120 gel documentation system.
Annexin V and Propidium Iodide (PI) Staining and FACS Analysis-Cells were plated in culture plates and exposed to various doses of UVB. 24 h after induction of apoptosis, the cells were trypsinized, washed with ice-cold phosphate-buffered saline, and subsequently incubated in the dark with 100 l of annexin V incubation reagent, which includes fluorescein isothiocyanate-conjugated annexin V (Trevigen, Gaithersburg, MD) and PI for 15 min at room temperature. Flow cytometric analyses were conducted on a BD PharMingen FACStar Plus cytometer using a 100-mW air-cooled argon laser at 488 nm.

UVB Induces Caspase-3-mediated Apoptosis in HFK-We
have previously shown that the DNA alkylating agent sulfur mustard induces markers of terminal differentiation and apoptosis in normal human epidermal keratinocytes, including the early activation and late cleavage of PARP (23). To determine whether UVB induces the apoptotic response in HFK, cells were exposed to UVB, and markers of apoptosis were examined. Nuclear fragmentation (Fig. 1A) as well as the proteolytic processing of caspase-3 to its active form (p17; Fig. 1B) occurs 16 h after exposure to 480 J/m 2 UVB. PARP is also catalytically cleaved by caspases-3 from a 116-kDa full-length into an 89-kDa fragment that contains the C-terminal catalytic and automodification domains, a hallmark of apoptosis (Fig. 1B).
To determine the effects of HPV-16 E6 and E7 as well as immortalization on the response of HFK to UVB, we transduced HFK with a LXSN retroviral vector expressing the E6 and E7 genes of HPV-16 or with empty retroviral vector alone. Although E6 levels are technically difficult to detect, we examined the levels of p53 by Western blot analysis, since E6 induces the degradation of p53. Fig. 2A, top, shows that p53 is detectable in LXSN-HFK but not in p7 or p27 E6/7-HFK. In addition, comparison with HFK shows no effect of transfection with vector alone. The expression of E7 was directly detected by Western analysis in both p7 and p27 E6/7 HFK but not in control LXSN-HFK ( Fig. 2A, bottom). Phosphatidylserine is exposed on the surface of apoptotic cells (24), and the presence of these residues can be detected by their ability to bind to annexin V (25). To further examine the FIG. 6. The caspase-3 inhibitor DEVD-CHO partially blocks UVB-induced apoptosis. LXSN-HFK, p7 E6/7-HFK, and p27 E6/7-HFK were prepared as described under "Materials and Methods." Cells were irradiated with the 60 J/m 2 UVB in the presence or absence of the caspase-3 inhibitor DEVD-CHO, and 16 h later, cells were prepared and assayed for annexin V binding plus PI staining by FACS analysis (A and  B), or extracts were derived and subjected to immunoblot analysis using antibodies for PARP (C). The percentage of cells exhibiting annexin V binding (A) or that were negative for annexin V binding PI staining (B) as determined by FACS analysis are shown. All the data in A and B are presented as the means Ϯ S.D. of three replicates of a representative experiment; essentially the same results were obtained in three independent experiments. sensitivity to UVB-induced apoptosis in control and HFK E6/7 transfectants, we analyzed the cells for annexin V binding by FACS analysis 16 h after irradiation. Fig. 2B shows that p27 E6/7-HFK are much more sensitive to UVB-induced apoptosis at all doses tested than either p7 E6/7-or LXSN-HFK. A plot of the survival rates (PI negative, annexin V-negative) also confirms that p27 cells are more sensitive to UVB-mediated killing at all doses (Fig. 2C), thus indicating that the additional step(s) in immortalization plays a role in the sensitization to UVB.
Apoptosis is also characterized by the internucleosomal cleavage of DNA. We irradiated the three HFK cell types with the indicated doses of UVB, after which DNA was extracted and resolved by agarose gel electrophoresis. Fig. 3 shows that although nonspecific smearing is observed only at higher doses in p7-and LXSN-HFK (superimposed on two diffuse white bands that are the negative images of bromphenol blue and xylene cyanol dyes), a dose-dependent increase in internucleosomal DNA fragmentation is only observed in p27-HFK, with DNA ladders observed at the lowest dose of UVB used (60 J/m 2 ).
To determine the mode of apoptotic cell death, we analyzed the amount of caspase-3 activity using a quantitative fluorometric DEVDase assay. The three cell types were irradiated with increasing doses of UVB, after which the cytosolic extracts were assayed for caspase-3 activity. Fig. 4 shows that caspase-3 is activated in all three cell types, but caspase-3 is activated at lower UVB doses in p27 E6/7-HFK than in either p7 E6/7-or LXSN-HFK.
Caspase-3, responsible for the cleavage of PARP during apoptosis, is composed of two subunits of 17 and 12 kDa that are derived from a common proenzyme (26). To further analyze the proteolytic processing of caspase-3 and its substrate PARP, we performed immunoblot analysis of extracts from cells treated with different doses of UVB using an antibody specific for the p17 subunit of caspase-3 or to PARP. Fig. 5 shows that caspase-3 is proteolytically processed to its active form (p17), and PARP is specifically cleaved at lower UVB doses in p27 E6/7-HFK than either LXSN-or p7 E6/7-HFK.
To examine whether caspase-3 was in fact responsible for UVB-induced apoptosis, we preincubated the three cell types with an inhibitor of caspase-3 (DEVD-CHO) for 30 min before and during UVB exposure. The caspase-3 inhibitor decreased the proteolytic processing of PARP (Fig. 6C), reduced the number of cells undergoing apoptosis after 480 J/m 2 UVB exposure, as determined by annexin V plus PI staining (Fig. 6A), and reduced internucleosomal fragmentation in all three cell types (Fig. 3). In addition, survival was increased (Fig. 6B). Thus UVB induces an apoptotic mode of death that is partially dependent upon caspase-3.
Immortalization of HFK Switches the Mode of Apoptosis from a Caspase-8-to a Caspase-9-dependent Pathway-Various reports indicate that UVB activates either a death receptor or mitochondrial pathway of apoptosis. The former pathway results in the rapid activation of caspase-8, whereas the latter pathway activates caspase-9. We performed immunoblot analysis of extracts derived from the three cell types treated with different doses of UVB or with a single dose of UVB to deter- FIG. 7. Immortalization of HFK with HPV-16 E6/E7 induces UVB-dependent caspase-9 processing and increases caspase-8 processing at lower UVB doses. LXSN-HFK, p7 E6/7-HFK, and p27 E6/7-HFK were prepared as described under "Materials and Methods." Cells were irradiated with the indicated doses of UVB, and 16 h later, whole cell extracts were prepared and subjected to immunoblot analysis using antibodies specific for caspase-9 (A) or caspase-8 (B). The top panel in A was exposed for a longer period of time to visualize the reduced levels of caspase-9 in the treated samples.
FIG. 8. Immortalization of HFK with HPV16 E6/E7 induces rapid UVB-dependent caspase-9 processing. LXSN-HFK, p7 E6/7-HFK, and p27 E6/7-HFK were prepared as described under "Materials and Methods." Cells were irradiated with 480 J/m 2 UVB, and after the indicated times whole cell extracts were prepared and subjected to immunoblot analysis using antibodies specific for caspase-9 (A) or caspase-8 (B). mine a time course. Fig. 7B shows that caspase-8 is activated in all three cell types, although at a lower dose in p27 E6/7-HFK. In contrast, caspase-9 (Fig. 7A) is only proteolytically activated in p27 E6/7-HFK, as revealed by the loss of the inactive caspase-9 precursor. To determine which caspase is activated first, a time course was performed. Fig. 8 shows that although caspase-9 is activated almost immediately after UVB irradiation in p27 E6/7-HFK, caspase-8 is not activated until 16 h after exposure. Consistent with the results of the dose-response experiments, caspase-9 is not proteolytically processed at any time after UVB exposure in p7 E6/E7-or LXSN-HFK.
To confirm that these results are representative of responses of E6/E7-immortalized keratinocytes, we derived two additional batches of pooled cell clones from two different E6/E7 retroviral infections of different mixed foreskins. Comparison of LXSN, p7 E6/E7-HFK, and p27 E6/E7-HFK from matched sets confirmed our earlier findings that p27 cells are more sensitive to apoptosis than either LXSN or p7 E6/E7-HFK as a result of the preferential activation of caspase-9 in the later passage cells (Fig. 9).
To determine the possible reason for the switch in the apoptotic pathway in p27 E6/7-HFK, we examined the cell cycle before and after UVB irradiation in the three cell types. Fig.  10A shows that LXSN-HFK displays a strong G 1 arrest after all doses of UVB irradiation. As might be predicted from the reduced levels of p53, neither p7 nor p27 E6/7-HFK exhibit an appreciable G 1 arrest at any dose of UVB. However, only p27 E6/7-HFK showed a marked increase in the population of cells in G 2 /M after UVB treatment (Fig. 10C). Thus, although the elimination of the G 1 arrest is associated with expression of E6/7, the UVB-induced G 2 arrest appears to be attributable to the immortalization process.
Immortalization Results in a Bcl-2-dependent Pathway for UVB-induced Apoptosis-Although a causal link has not been unequivocally demonstrated, a strong correlation has been observed between G 2 arrest and apoptosis. This link is also associated with the phosphorylation and degradation of Bcl-2, associated with a mitochondrial pathway of apoptosis (27). To determine whether this was a possible mechanism that re- FIG. 9. Immortalization of other preparations of HFK with HPV-16 E6/E7 induces rapid UVB-dependent caspase-9 processing. LXSN-HFK, p7 E6/7-HFK, and p27 E6/7-HFK were prepared from two additional batches of neonatal foreskins as described under "Materials and Methods." Cells were irradiated with 480 J/m 2 UVB, and after the indicated times, whole cell extracts were prepared and subjected to immunoblot analysis using antibodies specific for caspase-9 (A) or caspase-8 (B). Representative results from one group of cells are shown; essentially the same results were obtained with both groups of cells.
FIG. 10. Immortalization of HFK with HPV-16 E6/E7 induces UVB-dependent G 2 /M arrest. LXSN-HFK, p7 E6/7-HFK, and p27 E6/7-HFK were prepared as described under "Materials and Methods." Cells were irradiated with the indicated doses of UVB, and 16 h later, nuclei were prepared as described under "Materials and Methods" and subjected to cell cycle analysis using FACS. All data are presented as the means Ϯ S.D. of three replicates of a representative experiment; essentially the same results were obtained in three independent experiments. sulted in the switch to a caspase-9-mediated pathway for apoptosis in immortalized HFK, we performed an immunoblot analysis of UVB-irradiated cells using a Bcl-2-specific antibody. Fig. 11 shows that the lowest dose of UVB results in the disappearance of immunodetectable Bcl-2 in p27 E6/7-HFK. In contrast, Bcl-2 levels persist at higher doses of UVB in p7 E6/7and LXSN-HFK. In fact, Bcl-2 persists in the latter cell types until it is specifically cleaved into a 23-kDa fragment, most likely by caspase-3 or -7 (28), at higher UVB doses. DISCUSSION Both HPV and UV have been shown to be etiologic agents for skin cancer. Immortalization is thought to be an early step in this process that involves the selection of a population of cells that progresses to the next stage of cancer. Based on these earlier observations, we separately tested the effects of E6/7 expression and immortalization on UVB apoptosis using early (p7) and immortalized late (p27) passage E6/7-transduced HFK. We found that UVB induces a caspase-3-mediated apoptotic death in all three cell types, but that p27 E6/7-HFK cells were more sensitive than either p7 E6/7-HFK cells or cells transduced with empty vector alone (LXSN-HFK).
Caspase-3 is activated in all three cell-types, but caspase-3 is activated at lower UVB doses in p27 E6/7-HFK than in either p7 E6/7-or LXSN-HFK, although a drop in caspase-3 activity is seen in p27 E6/7-HFK at higher doses. The mechanism for this phenomenon remains to be elucidated. An important point is that although the activity at 16 h is decreased, caspase-3 is processed into its active form at the higher doses (Fig. 5). Clearly, however, the sum total of caspase-3 activity expressed within the first 16 h after higher doses of UV treatment of late passage E6/7 is sufficient to completely proteolyze PARP (Fig.  5) and activate internucleosomal cleavage (Fig. 3). Likewise, all other apoptotic markers are higher in the late passage cells.
One potential explanation for our findings is that E6 and E7 are expressed at low levels in p7 E6/7-HFK. However, four lines of evidence argue against this possibility. First, HFK were infected at a high multiplicity of infection with the E6/7 retroviral construct and subsequently selected for 10 days in G418 (see "Materials and Methods"). Second, p53 levels were reduced in p7 E6/7-HFK as well as p27 E6/7-HFK, indicating that E6 was present and functional in both cell types ( Fig. 2A). Third, E7 levels were similar in both p7 E6/7-HFK and p27 E6/7-HFK (Fig. 2B). Fourth, FACS analysis revealed that the G 1 arrest induced by UV was not observed in either p7 E6/7-or p27 E6/7-HFK (Fig. 10). Thus, it appears that events subsequent to expression of E6 and E7 are critical to the increased sensitivity to UVB. The idea that immortalization requires genetic events in addition to the expression of E6 and E7 is not a new one.
Loss of a region of chromosome 6 has been associated with increased telomerase expression in HPV-immortalized cells; replacement of this chromosomal region suppressed hTERT activity (17). Non-random allelic losses have also been observed at 3p, 11p, and 13q during HPV-mediated immortalization (18), and a senescence locus within the chromosomal region 10p14-p15 is lost in HFK-expressing HPV-16 E6/7 genes (19).
To avoid many of the cell culture artifacts that arise from the use of well established lines such as HaCat, we utilized pooled clones of E6/7-HFK soon after immortalization (p27). In addition, batches of p27 E6/7-HFK-pooled cell clones from different E6/E7 retroviral infections were more sensitive to UVB-induced apoptosis than either LXSN-or p7 E6/7-HFK as a result of the preferential activation of caspase-9 in the later passage cells (Fig. 9).
p27 E6/7-HFK differed from both LXSN-and p7 E6/7-HFK with respect to G 2 /M accumulation, Bcl-2 down-regulation, and caspase-9 activation. These events are likely to be associated with a mitochondrial pathway of apoptosis, whereas the activation of caspase-8 is generally, but not exclusively associated with a death receptor-mediated pathway. In fact, evidence has been presented for both pathways for UVB-induced apoptosis in both human and mouse keratinocytes. The Fas/CD95 death receptor has been shown to be up-regulated in keratinocytes in response to UV (29), and Aragane et al. (30) present evidence that this receptor is activated on the surface of keratinocytes via ligand-independent direct cross-linking of Fas by UV. Knockout studies indicate that the tumor necrosis factor receptor p55 plays a pivotal role in murine keratinocyte apoptosis induced by UVB irradiation (31). Other investigators show that UV stimulates the up-regulation or secretion of death receptor ligands, such as tumor necrosis factor ␣ (32), although tumor necrosis factor ␣ is only partially involved in UVB-induced apoptosis of normal human keratinocytes (33). Bcl-xL, associated with a mitochondrial apoptotic pathway, also confers resistance to UV in transgenic mouse skin (34). In addition, blocking Bcl-xL expression with an antisense inhibitor sensitizes normal human keratinocytes to UVB (35). Moreover, nitric oxide protects against UVA-induced apoptosis, which was correlated with Bcl-2 up-regulation (36).
The response to UVB is therefore determined by a number of pro-and anti-apoptotic factors that are present within keratinocytes both before and after irradiation. The changes that occur with immortalization that result in the increased sensitization and alteration of apoptotic pathways are not understood. However, as an instructive first step, we have examined the changes in gene expression after UVB exposure of LXSN-HFK, p7 E6/7-HFK, or p27 E6/7-HFK utilizing microarray analysis. Similar to the apoptotic response reported in the current study, we have shown that UVB-induced mRNA profile changes in LXSN-HFK and p7 E6/7-HFK are somewhat similar, whereas both differ from the UVB-induced mRNA profile alterations in p27 E6/7-HFK. 2 The results of this study indicate that the immortalization process, rather than the expression of E6 and E7 alone, is critical to the increased apoptotic response, which involves a switch from a caspase-8-to a caspase-9-mediated pathway. The death-resistant phenotype that further contributes to clonal selection during the promotion stage of carcinogenesis may therefore be the result of further genetic alterations that occur subsequent to immortalization.