Increased Ultraviolet Sensitivity and Chromosomal Instability Related to P53 Function in the Xeroderma Pigmentosum Variant¹ Free

XP⁴ is a rare autosomal recessive disorder characterized by sun sensitivity, cutaneous and ocular deterioration, and premature malignant skin neoplasms on sun exposure (1). The mechanism of cancer development in XP remains a matter of conjecture, despite the appealing association of defective excision repair of UV damage to DNA in the cells of the skin with increased risk of actinic cancer (2, 3). One of the difficulties with this notion is that about one-fourth of XP patients with characteristic symptoms of sunlight-induced cancer do not have deficiencies in DNA excision repair (1, 4). These have been classified with a different abnormality in the response to UV irradiation involving semiconservative DNA replication and are known as XPVs (5, 6).

Most XP patients have fibroblasts with DNA repair deficiencies and are represented by one of the seven excision repair defective complementation groups (A-G) that encode proteins involved in recognition and excision of damaged sites from DNA (7). The eighth group, XPV, includes patients with clinically diagnosed XP whose cells are only slightly more sensitive than normal cells to UV light and have normal DNA excision repair (1, 5, 6, 8, 9). After irradiation, XPV cells replicate their DNA in abnormally short fragments, indicating a defect in the bypass replication of photoproducts in the parental strands, making a positive diagnosis possible (6, 8, 10, 11, 12, 13, 14). This defect is associated with increased rates of UV-induced mutagenesis (15, 16, 17, 18, 19), suggesting that the XPV gene, or genes, can be regarded as a form of mutator gene.

We report here that transformation of XPV fibroblasts by SV40 or HPV16 genes, which inactivate p53 (20, 21, 22), results in a dramatic increase in the sensitivity of XPV cells to both the lethal and the cytogenetic effects of UV light. The in vitro phenotype of XPV cells, therefore, exhibits secondary alterations that are informative about the nature of the XPV defect, indicating, especially, an increased genomic instability after p53 inactivation, but which may also confound attempts to clone the gene by functional complementation.

Primary fibroblasts, SV40- and HPV16-transformed cells, were grown in Eagle’s MEM supplemented with 2 mm glutamine, penicillin (100 IU/ml), streptomycin (100 μg/ml), and 10% fetal bovine serum (all from Life Technologies, Inc., Gaithersburg, MD). The majority of this study was carried out with an isogenic set consisting of one normal and two XPV cell lines, with some individual experiments involving various other cell lines. The primary normal (GM037 and AG1521), XPV (GM3617, GM3379, and GM3618), and XPC (XP6CA and XP1BE) fibroblasts were obtained from the Human Genetic Mutant Cell Repository (Camden, NJ). The normal fibroblast cell line (GM037) was permanently transformed by SV40 and maintained for many generations as GM637; we also transformed this fibroblast with HPV16(E6 and/or E7) to generate the cell lines GM037(E6), GM037(E7), and GM037(E6/7). The XPV fibroblast GM3617 (also designated XP30RO) was transformed with SV40 to produce the permanently transformed line XP30RO(sv) and, with HPV16(E6 and/or E7), to generate the cell lines XP30RO(E6), (E7), and (E6/7), respectively. The XPV fibroblast GM3379 (XP5MA) was similarly transformed to generate XP5MA(sv), XP5MA(E6), and others.

Other cell lines that were used included the XPVs XP29SF, XP203SF, XP204SF, and others described in the legend to Fig. 1, from our own stocks. The SV40-transformed XPV XPCTag was a gift from Drs. W. K. Kaufmann and M. Cordeiro-Stone (Univ. North Carolina). The normal primary fibroblast FS was obtained from a local foreskin biopsy. The SV40-transformed group A cell line XP12RO was from our own long-term cultures. The Cockayne’s syndrome group B cell line, GM10905, was also from the Human Genetic Mutant Cell Repository and transformed with SV40 ori^- to obtain an immortalized culture.

Transformation was carried out de novo for some experiments, using either strain 776 of SV40 whole virus (XP30ROsv) or a plasmid containing the early region of SV40 ori^- (GM3618, GM3379, and GM10905), or by using a retroviral-mediated route for HPV16 E6 or E7 genes. Depending on the experiment, 1 × 10⁶ to 4 × 10⁶ cells were plated to obtain transformed cell lines. We immortalized primary XP30RO cells with extreme difficulty using strain 776 of SV40 whole virus, as described previously, and designated the transformed cell line as XP30RO(sv) (23). The other XPV fibroblasts, GM3379 and GM3618, were transformed with the ori^- plasmid and used during the early period of rapid proliferation; although satisfactory data could be obtained with GM3379 during this period, GM3618 entered crisis rapidly and no SCE experiments were possible. The HPV16 E6 and E7 genes were introduced into the cell lines using a retroviral vector carrying a neomycin-resistance gene, as previously described (24). The HPV16-transformed cells were maintained in the antibiotic gentamicin (G418; 200 μg/ml) to select for neomycin resistance; this was because when the retroviral constructs containing either E6 or E7 alone the cells were not fully transformed, and even when both genes were used, the cells did not become immortalized during the course of these experiments.

Aliquots of primary and HPV16-transformed cells treated with and without UV irradiation were subjected to PAGE, transferred to membranes, and probed with antibodies for p53, p21^WAF1, and tubulin, as described previously (24), to demonstrate abrogation of p53 function. Similar experiments with SV40-transformed cells showed a uniform high level of p53 independent of UV irradiation (data not shown).

For cell survival, a rapid well assay was used involving measuring the growth of cells for 5–7 days in 24-well plates, as described previously (25), and a colony formation assay in which cells were grown into visible colonies over 2–3 weeks was used. In some experiments, cells were grown in caffeine (1 mm) after irradiation, which is known to sensitize XPV cells to UV light (9, 26).

For SCE analysis, 5 × 10⁵ cells were plated in MEM in 100-mm Petri dishes and analyzed, as we described previously (27). Twenty-four hours later, the medium was removed, and the exponentially growing cultures were exposed to doses of UV (254 nm) light up to 5.2 J/m². Fresh medium containing 10 μm bromodeoxyuridine was added to the plates, and the cells were grown for two cell cycles, approximately 54 h. Colcemid (2 × 10^-7 M, final concentration) was added to the cultures 4 h before fixation, and the mitotic cells were collected by shake-off. Cells were swollen by exposure to 0.075 M KCl for 7 min at 37°C. The cultures were centrifuged, and 8–10 ml of methanol was added to the pellets. Pellets were washed twice with fixative (methanol-acetic acid, 3:1 v/v) and dropped on wet slides. Air-dried slides were stained by a modification of the fluorescence-plus-Giemsa technique of Perry and Wolff (28). The slides were immersed in Hoechst 33258 (5 μg/ml, final concentration) in M/15 Sørensen’s buffer (pH 6.8) for 20 min. The slides were then washed, dried, mounted with buffer under a coverslip, and exposed to black light for 7 min on a 56°C slide warmer. This was followed by staining the slides in 4% Giemsa in Sørensen’s buffer. A minimum of 25 cells were scored for each UV dose in each cell line.

The effects of UV damage on transcription were determined in SV40-transformed cells by assaying the expression levels of the luciferase reporter gene in transient assays. Plasmid pGL2, which expresses the luciferase gene (Promega Corporation, Madison, WI), was grown in competent bacteria, isolated, and purified in CsCl gradients. Plasmid preparations were irradiated in buffer [Tris-HCl (pH 7.4)] at UV fluences up to 2400 J/m². Transfection mixtures consisting of pGL2 (10 μg), 10–20 μl of lipofectin (Life Technologies, Inc.), and culture medium without serum or antibiotics to a final volume of 200 μl were then prepared. These mixtures were allowed to stand at ambient temperature for 15 min before being added to 60-mm Petri dishes containing 1–2 × 10⁶ cells/dish. Cultures were incubated overnight at 37°C (17–24 h) before being washed with PBS and harvested. Cells were scraped into PBS, washed once, counted, and centrifuged. The pellets were lysed in 250 μl of cell lysis buffer (Promega Corp.) for 20 min. For 20 μl of each lysate, 130 μl of luciferase assay reagent (Promega Corp.) was added, and the mixture was immediately placed in a Monolight 2010 Luminometer (Analytical Luminescence Laboratory, San Diego, CA) and measured for luminescence over 10 s. Three aliquots were assayed for each lysate, and the luminescence/cell was calculated.

The excision of pyrimidine dimers from total genomic DNA was determined at 5 and 24 h after 13 J/m² by isolating the DNA, absorbing aliquots onto immunoassay plates, and incubating with an antibody to the photoproducts, followed by secondary reaction with peroxidase and quantification by absorbance (29). The total amount of repair synthesis was determined by counting the number of grains/nucleus of 25 cells/culture undergoing unscheduled DNA synthesis, measured in autoradiographs of cells labeled with [³H]thymidine (10 μC/ml; Amersham Corp.) for 2 h after 10 J/m² (30).

We, and others, have previously demonstrated that XPV primary fibroblasts are slightly sensitized by caffeine after UV irradiation, (4, 9), but that the degree of sensitization increases markedly after SV40 transformation (4, 26). We confirmed that this occurred in the cell lines we used in this study for both the immortalized cells (XP30ROsv and XPCTag) and those recently transformed by SV40 (Fig. 1). The increased sensitization was expressed as soon as transformation had occurred and cells had entered a rapid growth phase and did not require complete immortalization.

Although XPV cells have been intensely investigated and only subtle, if any, defects have been identified in excision repair, we still considered the possibility that the extreme sensitivity seen in XPV cells after transformation and growth in caffeine might have uncovered a cryptic defect in excision repair. We, therefore, assayed the recovery of luciferase expression after transfection of UV-damaged plasmids into normal and XPV cells, with and without growth in caffeine (Fig. 2). Despite the high sensitivity of UV-damaged XPV cells grown in caffeine, the recovery of luciferase expression was the same as in normal cells. This is consistent with additional analysis that indicates that excision of pyrimidine dimers occurs at the same rate and extent in SV40-transformed XPV cells as from untransformed and transformed wild type cells (see next section and Fig. 6). The high sensitivity of XPV cells in caffeine, therefore, does not involve defects in excision repair that act on exogenously irradiated plasmids.

Early studies of XPV fibroblasts showed that the induction of SCEs by UV was in the same range as seen in normal cells (31), and we confirmed this for normal and XPV fibroblasts (Fig. 3,A). After SV40 transformation, however, SCE frequencies in XPV cells were extremely sensitive to induction by UV damage (Fig. 3,A). Induction increased very rapidly at low doses and saturated at ∼1 J/m². SCEs could not be determined above 2.6 J/m² because SV40-transformed XPV cells did not enter mitosis above these doses. We confirmed, by flow cytometry, that at these doses XPV cells accumulated in the S phase and further progression was very slow (data not shown). For comparison, we also measured SCE yields in transformed and immortalized XP group A (XP12RO) and Cockayne’s syndrome group B cells (GM10905sv), and neither showed such high SCE yields as the immortalized XPV cells, XP30ROsv and XPCTAG (Fig. 3 B). The immortalized XPV cells, therefore, were more sensitive to SCE induction than cells defective in excision repair (XPA) or in transcription coupled repair (CSB).

The enhancement of SCE production occurred both in immortalized cells (XP30ROsv and XPCTAG) and in XPV cells assayed soon after transfection with SV40 DNA (XP5MAsv), which were rapidly proliferating but not permanently transformed (Fig. 3B). XP5MA(sv) had a near diploid chromosome number in contrast to the other immortalized cell lines (Table 1), and its yield of SCEs was not as great as in the immortalized cells. The high SCE induction might, therefore, be a marker for increased genomic instability in the XPV that can be further destabilized with increasing chromosome imbalance, but this will need to be investigated further in a larger range of XPV cells.

The previous results suggested to us that there might be a relationship between the p53 status of XPV cells that is altered by SV40 transformation and their genomic instability. We initially attempted to investigate transformation with various SV40 mutants, but then decided to use a different route of transformation using HPV16 E6 and E7 genes in a retroviral vector that would provide more experimental flexibility.

Primary normal (GM037) and XPV (XP5MA and XP30RO) fibroblasts were transfected with retroviral constructs that expressed either the E6 or the E7 genes of HPV16; transfections were carried out separately and with both. Transformation with E6 or with E6 and E7 together almost completely suppressed expression of p53 and p21^WAF1 (Fig. 4). The UV sensitivity of XPV cells was greatly enhanced by the E6, but not the E7 genes; use of both together did not result in any further increase over that seen with the E6 gene alone (Fig. 5). In these experiments, caffeine enhancement was not required to obtain this increased sensitivity, and the resulting sensitivity was as great as seen in an XPC cell line (GM2995 or XP6CA) that lacks the capacity to carry out global genomic repair (32). Addition of caffeine after irradiation of the HPV16-transformed XPV cells only gave a small increase in UV sensitivity, in contrast to SV40 transformation (data not shown).

DNA excision repair was measured to ensure that a cryptic repair defect was not uncovered by HPV16 transformation of XPV cells. Overall, pyrimidine dimer excision and unscheduled DNA synthesis were, however, not significantly different in normal and XPV cells (Fig. 6 and Table 2). Although the sensitivity of the XPV cells transformed by E6/E7 was the same as that for primary fibroblasts of XP group C (XP6CA; Fig. 5, B and C) cells, unscheduled synthesis was not reduced to the low levels seen in group C fibroblasts (XP1BE and XP6CA; Table 2). Incidental to this study, we did not see changes in the repair capacity or UV sensitivity in the normal cell line after HPV16 transformation, in contrast to recent reports that a functional wild type p53 is required for global repair of the nontranscribed regions of the genome (33, 34). Detailed comparisons, however, of unscheduled DNA synthesis in autoradiographs from different cell types with different morphologies can confound the quantification, and only large differences should be regarded as informative.

We then investigated the induction of SCEs in the E6/E7-transformed cells and found, as with SV40 transformation, that the XPV cells were much more sensitive to their induction than normal cells containing E6 and E7 (Fig. 7). The SCE yield in the E6/E7-transformed normal cells was similar to that seen previously in nontransformed fibroblasts (FS; Fig. 3 A). The highest dose at which we could obtain second division cells for this assay was 0.3 J/m², suggesting that this route of transformation resulted in cells that were much more sensitive to UV-induced inhibition of cell cycle progression than with SV40. As with SV40-transformed cells, flow cytometry demonstrated an accumulation of cells in the S phase after higher doses of UV light (data not shown).

These results indicate that the phenotype of XPV cells in culture is much more plastic than we had come to expect from the development of transformed cell lines of other groups of XP and other mutagen-sensitive disorders. In most previous cases, the phenotypes of the transformed cells showed expected changes in cell cycle regulation (35), but the processes related to repair and UV or mutagen sensitivity were similar. In the case of the XPV, however, both SV40 and HPV16 transformation led to phenotypes in which the sensitivity to induction of SCEs by UV light was greatly increased. Primary fibroblasts and cells obtained after SV40 transformation showed only small increases in their sensitivity to UV-induced cell killing, unless sensitized by caffeine, which preferentially sensitized the XPV cells. Transformation with HPV16(E6), however, increased the UV sensitivity of XPV cells to a degree equivalent to that seen in repair-defective XPC cells. The increased sensitivity seen after either route of transformation was not associated with any detectable change in reactivation of an exogenous shuttle vector nor in excision repair of the total genome. In contrast, previous work with EBV-transformed lymphoblastoid cell lines, which usually exhibit a wild type p53 response, did not show increased SCE formation nor increased UV sensitivity with caffeine (36).

Taken together, these observations suggest that the UV sensitivity and chromosomal instability of XPV cells is restrained when p53 is able to function normally. This is particularly clear in the comparison between transformation by the E6 and E7 genes of HPV16. Inactivation of p53 by E6-mediated ubiquitination (20, 21, 24, 37) resulted in increased UV sensitivity; E7, which targets the Rb gene product, was without effect (38).

Several previous studies have implicated p53 in the regulation of global nucleotide excision repair (33, 34, 39), but this is not the mechanism that gives rise to the UV sensitivity of the XPV because we observed no change in global repair in either normal or XPV cells. Instead, we need to reconcile our observations with the well-established defect in semiconservative DNA chain growth that is characteristic of the XP variant.

The XP variant exhibits a defect in the ability to replicate past pyrimidine dimers with consequent arrest of the polymerase δ-dependent leading strand (40, 41). Replication fork arrest results an increase in the amount of single stranded regions and DNA termini (42). These could be sites for initiation of strand exchange and recombination. Several attempts have been made in the past to determine whether the recovery of replication in UV-damaged cells occurred by strand elongation and bypass or by strand exchange mechanisms, but evidence for strand exchange was not found (43, 44). The exchange models were envisaged to be similar to the recA-mediated exchange seen during recovery of DNA replication in UV-irradiated in E. coli (45). The early mammalian cell experiments were carried out in fibroblasts that presumably had competent p53 (10, 14, 43, 46). Our results raise the possibility that the mechanism of recovery of DNA replication from UV damage could either be by a replicative bypass or by recombination, depending on p53 function. If SCEs are regarded as S phase exchange events involving homologous recombination between chromatids (47), then our results would be consistent with enhanced recombinational repair occurring when p53 is nonfunctional. Enhanced recombination after UV irradiation has been observed in SV40-transformed XP groups A and F (48).

One mechanism by which the enhanced sister chromatid recombination could occur in the XPV is through rad51-dependent recombination, because p53 normally inhibits rad51 activity (49). Alternatively, resolution of recombination intermediates could be blocked if p53 is normal because of its capacity to bind Holliday junctions and prevent nuclease-mediated resolution (50). Further investigation of recombination in transformed XPV cells would, therefore, be informative, and we have begun to do this with the vector previously used for XPA and XPF cells (48). Our results imply that a recombinational mode of recovery, on a p53 negative background, is a less successful mechanism of recovery than DNA chain elongation seen in primary fibroblasts, as judged by the enhanced UV sensitivity, caffeine sensitization, and SCE frequencies that may, therefore, contribute to further genomic instability.

An important implication of these observations is that the sensitivity of the XPV cells to UV light is strongly dependent on the route of transformation and is, therefore, derivative from the primary defect in DNA replication. There is even a distinction between SV40 and HPV16(E6) transformation because although both give rise to increased sensitivity to SCE induction, a high degree of UV sensitivity is only seen in SV40-transformed cells after caffeine sensitization. Similar differences in UV sensitivity and caffeine sensitization in p53(+) and (-) rodent cells have also been observed (51). The distinction in the two transformed states is that whereas SV40 large T binds the central region of p53 that allows the NH_2- and COOH-terminal regions to retain some functions (22), HPV16(E6) causes ubiquitination and complete loss of p53 (37). Complete loss of p53, therefore, has a more severe consequence on cell survival than T antigen binding.

The complex transformed phenotypes mean that attempts to clone the XPV gene by functional complementation with DNA libraries may result in cloning genes that are downstream of the primary defect and involved in expression of the transformed phenotype (52). This has been our experience; we have found, for example, that the high SCE induction can be complemented by a gene on chromosome 1 that is a homologue of a gene on the DNA methylation pathway (S-adenosyl homocysteine hydrolase, accession number U82761), but which does not have mutations in XPV cells (52). Several other presently uncharacterized cDNAs have also shown partial complementation.⁵ This partial correction seems to be significant because other genes that we have specifically tested did not show any correction of the caffeine sensitization XPV cells. These include 5-methylcytosine transferase, PCNA, RPA, Hsc70, elongation factor E1a, yeast Rad7, XRCC9, and Fen1. Several radiosensitive complementation groups of Chinese hamster ovary cells remain uncharacterized, and these could be XPV candidates because the parent Chinese hamster ovary cell line is p53 negative. Recently, we have excluded XRCC9, which had the potential to be involved in postreplication repair (53). Those cDNAs that show partial complementation of transformed XPV cells may, therefore, be involved in the complex pathways associated with signaling DNA chain growth arrest and cell death and may eventually be informative about the detailed expression of the XPV phenotype after transformation. The increase in UV sensitivity and SCE induction after transformation with HPV16 could also be used as another assay for improved identification and diagnosis of the XPV phenotype.

Our observations have several implications for the pathology of the XPV. Mutations occur in the p53 gene with high frequency after sun exposure (54) of normal and excision-defective XP patients; at least half the tumors carry mutations in both cases. Because we observed a large increase in chromosomal instability and UV sensitivity after loss of p53 function, our results predict that a higher proportion of tumors in XPV patients might show p53 mutations than in normal and excision-defective patients. The similarity in symptoms between excision-defective and variant patients might also be related to the dependency on p53 function; excision defective patients might generate p53 mutant cells in the skin more rapidly than normal individuals, but the XPV cells might show greater genomic instability once p53-negative cells had been produced. The net result would be similar overall increases in genomic instability, leading to actinic cancers. XPV patients might also show a unique pathology associated with HPV infection, which would be worth careful clinical consideration.

We are grateful for the assistance of patients and their families in the early development of this study. We are also grateful to the XP Society (Poughkeepsie, NY) for continued support and encouragement to one of us (J. E. C.).

Note Added In Proof

We have recently used the pTPSN vector (48) to measure recombination. At 1.3J/m², SV40-transformed normal, GM637, cells yielded 6 × 10^-6 recombinants per survivor whereas the variant, XP30RO, cells yielded 1.5 × 10^-6 recombinants. Gene conversion, which is the major event detected by this vector (48), does not therefore correlate with increased SCEs in XPV cells.