Germ line mutations in the mismatch repair (MMR) genes hMSH2 and hMLH1 account for ∼98% of hereditary non-polyposis colorectal cancers. In addition, there is increasing evidence for an involvement of MMR gene expression in the response of cells to UV-induced skin cancer. The link between MMR and skin cancer suggests an involvement of MMR gene expression in the response of skin cells to UV-induced DNA damage. In this report, we have used two reporter gene assays to examine the role of hMSH2 and hMLH1 in the repair of oxidative DNA damage induced by UVA light and DNA damage caused by methylene blue plus visible light (MB+VL). UVA and MB+VL produce 8-hydroxyguanines in DNA that are repaired by base excision repair (BER). AdHCMVlacZ is a replication-deficient recombinant adenovirus that expresses the β-galactosidase (β-gal) reporter gene under the control of the human cytomegalovirus (CMV) immediate-early promoter. We show a reduced host cell reactivation for β-gal expression of UVA-treated and MB+VL-treated AdHCMVlacZ in hMSH2-deficient LoVo human colon adenocarcinoma cells compared to their hMSH2-proficient counterpart SW480 cells, but not in hMLH1-deficient HCT116 human colon adenocarcinoma cells compared to hMLH1-proficient HCT116-chr3 cells. We have also reported previously that enhanced expression of the undamaged AdHCMVlacZ reporter gene is induced by the pre-treatment of cells with lower levels of the DNA-damaging agent and to higher expression levels in transcription-coupled repair (TCR)-deficient compared to TCR-proficient cells. Here we show that pre-treatment of cells with UVA or MB+VL enhanced expression of the undamaged reporter gene to a higher level in LoVo compared to SW480 cells but there was little or no difference in HCT116 compared to HCT116-chr3 cells. These results suggest a substantial involvement of hMSH2 but little or no involvement of hMLH1 in the repair of UVA- and MB+VL-induced oxidative DNA damage by BER.
DNA mismatch repair (MMR) constitutes an important DNA repair pathway that plays a vital role in maintaining genomic stability. MMR is relatively conserved in a number of species including bacteria, yeast, and mammalian cells, and is responsible for the recognition and repair of bases incorrectly incorporated during DNA replication. In human cells, this repair process involves two principal repair proteins: (i) hMLH1 (human mutL homologue 1) and (ii) hMSH2 (human mutS homologue 2). Individuals heterozygous for a mutated MMR gene are predisposed to developing non-polyposis colorectal carcinoma (HNPCC) and tumors from patients with HNPCC have a mutator instability phenotype (1,2). Although most individuals heterozygous for a mutated MMR gene are not predisposed to skin cancer, the rare autosomal dominant Muir–Torre syndrome is characterized by the association of visceral malignancies including squamous cell carcinoma and like HNPCC shows alterations in the MMR system (3). The MSH2-MSH6 heterodimer has been shown to bind to UV adducts (4,5) and MMR proteins physically interact with nucleotide excision repair (NER) (6,7). In addition, the MMR proteins have been reported to protect mammals against UVB-induced skin cancer (8) and there is evidence that MSH2 protein levels are altered in human non-melanoma skin cancers (9–11). Both microsatellite instability and MMR gene alterations have been reported in human melanomas and may be involved in the pathogenesis of melanoma progression (12,13).
The link between MMR and skin cancer suggests an involvement of MMR gene expression in the response of skin cells to UV-induced DNA damage.
The predisposition of Msh2-deficient mice to UVB radiation-induced skin cancer may reflect an interaction between the processes of MMR and excision repair for UVB-induced DNA damage (8). In addition, since MMR proteins have been reported to promote apoptosis and contribute to cell cycle arrest in response to UVB radiation, it has also been suggested that dysfunctional MMR promotes UVB-induced tumorigenesis in mice through reduced apoptotic elimination of damaged epidermal cells (14–16).
UV-induced DNA lesions are repaired by NER and base excision repair (BER) in mammalian cells. The role of MMR in NER of UV-induced DNA damage is controversial and could be species specific (17). In Saccharomyces cerevisiae strains, MMR appears not to be required for removal of UVC-induced DNA lesions by the transcription-coupled repair (TCR) or global genomic repair pathway of NER (18). Several reports also suggest that there is no reduction in the repair of cyclobutane pyrimidine dimers (CPDs) by the TCR pathway of NER in MMR-deficient mouse cells (19,20). In Escherichia coli, there are conflicting reports on the role of the MMR genes in the repair of UVC-induced DNA damage by the TCR pathway of NER (21–23). Reports on the role of the human MMR genes hMLH1 and hMSH2 in the removal of UV-induced DNA lesions by the TCR pathway of NER are also conflicting (6,24–27). Most of the research into the role of the human MMR genes in the repair UV-induced DNA damage has been performed using tumor cell-derived cell lines and using UVC (germicidal, 200–280 nm) radiation as the DNA-damaging agent. It is therefore possible that the tumors used to generate some of these cell lines have acquired additional genetic mutations other than loss of MMR. In addition, since UVC is filtered out by the ozone layer in the stratosphere, little, if any, reaches the earth's surface (28) and thus is generally regarded not to have any significance on the environment or on human health. UVA (320–400 nm) and UVB (280–320 nm) make up the rest of the ultraviolet portion of solar radiation, with UVA being around 20-fold more intense than UVB in sunlight and a recent report suggests that UVA plays a role in human skin carcinogenesis (29). UVC produces predominantly CPDs and 6-4 photoproducts (6-4 PPs) that are repaired by NER, whereas UVA and UVB in addition produce oxidative DNA damage via the production of reactive oxygen species which results in damage to single DNA bases, mainly 8-oxoguanine (8-oxoG) as well as oxidized pyrimidines such as thymine glycol or uracil derivatives which are repaired by the BER pathway.
Several reports suggest a role for hMSH2 and its functional homologues in the recognition and removal of 8-oxoG, a lesion that often mispairs with adenine (30–33) and is often a target of BER. It is therefore possible that the association between MMR and skin cancer results from an effect of MMR deficiency on the removal of UV-induced DNA damage by BER.
Using a host cell reactivation (HCR) of a UVC-damaged reporter gene, we have previously presented data suggesting that the human colon adenocarcinoma cells HCT116 and LoVo bearing homozygous mutations in hMLH1 and hMSH2, respectively, are deficient in TCR of UVC-induced DNA lesions compared to their HCT116-chr3 and SW480 MMR-proficient counterparts and that the ability to detect the involvement of these MMR genes in TCR is dependent on UVC fluence to cells (27). We have also suggested that the apparent discrepancy in previously published results using these same human colon adenocarcinoma cell lines regarding the role of hMLH1 and hMSH2 in TCR results, in part at least, from differences in the UVC fluence to cells used to examine TCR.
In the present work, we use the same human colon adenocarcinoma cell lines to examine the role of the MMR genes hMSH2 and hMLH1 in the repair of UVA-induced and methylene blue plus visible light (MB+VL)-induced oxidative DNA damage using the HCR and enhanced expression assays as described previously (34–36). Surprisingly, there are currently no published studies that discuss the involvement of the hMLH1 and hMSH2 genes specifically in the repair of UVA-induced or MB+VL-induced DNA damage. Here we describe the use of UVA (320–400 nm) to induce oxidative DNA damage due to the environmental relevance of UVA exposure from the sun as well as the ability of UVA to cause 8-oxoG lesions in DNA (37–40) that are repaired by BER. We also make use of the type II photosensitizer, MB, that produces singlet oxygen (1O2) when excited by VL in the presence of oxygen (41) and which has been shown to 8-oxoG lesions in DNA (39) with minimal strand breaks (40). We report a reduced HCR of a UVA-treated and MB+VL-treated reporter gene in hMSH2-deficient LoVo human colon adenocarcinoma cells compared to their hMSH2-proficient counterpart SW480 cells, but not in hMLH1-deficient HCT116 human colon adenocarcinoma cells compared to the hMLH1-proficient HCT116-chr3 cells. In addition, we show that pre-treatment of cells with UVA and MB+VL enhances reporter gene expression to higher levels and at lower UVA fluences in LoVo compared to SW480 cells but there was little or no difference in the response of HCT116 compared to HCT116-chr3 cells. Results from both reporter gene assays suggest a substantial involvement of hMSH2 but little or no involvement of hMLH1 in the repair of UVA-induced and MB+VL-induced oxidative DNA damage. In addition, LoVo cells were significantly less sensitive to UVA-induced and MB+VL-induced cytotoxicity, consistent with an additional role for hMSH2 in radiation-induced apoptosis.
Materials and methods
Cell lines and virus strains
SW480 and LoVo colon adenocarcinoma cells were purchased from the American Type Culture Collection. HCT116-Chr3 and HCT116 human colon adenocarcinoma cell lines were kindly provided by Dr Thomas Kunkel (National Institute of Environmental Health Sciences, Research Triangle Park, NC, USA) with the help of Dr Regen Drouin (Faculty of Medicine, University of Sherbrooke, Sherbrooke, Québec, Canada). All cell cultures were grown in a humidified incubator at 37°C in 5% CO2. SW480, HCT116-chr3 and HCT116 carcinomas were cultured in McCoy's modified medium; LoVo cells were cultured in a 1:1 mixture of D-minimal essential media (MEM) and F-12 media (with 4500 mg/l glucose). All cell culture media were supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin.
AdHCMVlacZ [also called AdCA17lacZ (42)] is a non-replicating, recombinant adenovirus containing the lacZ gene under the control of the HCMV-IE promoter inserted in the deleted E1 virus gene. Virus stocks were prepared as described previously (43).
UVA irradiation and MB+VL treatment of virus
UVA irradiation was performed using a 1-kW mercury-xenon arc lamp with a 335-nm cut-off filter (Sciencetech Inc., London, Ontario, Canada), which effectively reduced UVB transmittance (300–320 nm) to 0.14% (according to manufacturer's specifications). Viral suspensions were prepared in 1.8 ml phosphate-buffered saline (PBS) in 35-mm petri dishes on ice. With continuous stirring, virus suspensions were UVA irradiated at a fluence rate of 110 J/m2/sec. Aliquots of 200 μl were removed following each exposure to the virus and diluted appropriately in unsupplemented α-MEM. UVA fluence rates were assessed using a long-wavelength meter (model #J221, Ultraviolet products, San Gabriel, CA, USA).
The same procedure described above was used for MB+VL treatment of the virus except that the virus suspension in 1.8 ml of PBS also included a final concentration of 10 μg/ml (27 μM) of MB and aliquots were also removed this time following each exposure to VL. VL treatment was performed by using a 1000-W bulb (General Electric, GE R1000, Hamilton, New Zealand) at a distance of 41 cm from the source for a specified time. The viral suspension for the MB+VL experiments was always kept in the dark when not directly exposed to the VL light source to avoid activation of the photosensitizer.
HCR of the AdHCMVlacZ reporter gene
For HCR experiments, the various carcinoma cell lines were seeded at 4.0 × 104 cells per well in 96-well plates (Falcon, Franklin Lakes, NJ, USA). Between 18 and 24 h after seeding cells, growth medium was aspirated from the wells and the cells were infected with a 40-μl volume of virus treated with either UVA or MB+VL or non-treated virus at a multiplicity of infection (MOI) of 20–40 plaque-forming units (PFU) per cell. A single HCR experiment for each cell line consisted of triplicate wells for each treatment of the virus and triplicate wells containing non-infected cells were used to obtain background levels of β-galactosidase (β-gal) expression. Following viral absorption for 90 2min, cells were re-fed with the appropriate supplemented medium and allowed to incubate at 37°C for 44 h prior to harvesting. For HCR experiments using MB+VL-treated virus, following viral adsorption for 90 min, the 40 μl of virus suspension was removed by aspiration before the addition of supplemented growth medium. β-Gal activity was measured by the addition of 1 mM chlorophenolred β-D-galactopyranoside (prepared in 0.01% Triton X-100, 1 mM MgCl2, 100 mM phosphate buffer at pH 8.3; Boehringer-Mannheim, Indianapolis, IN, USA) to the infected cells and absorbance readings were taken using a 96-well plate reader (Bio-Tek Instruments EL340, Winooski, VT, Bio Kinetics Reader) at several time intervals at 570 nm.
UVA-enhanced and MB+VL-enhanced expression of the undamaged AdHCMVlacZ reporter gene
Cells were seeded at 4.0 × 104 cells per well in 96-well plates (Falcon). Between 18 and 24 h after seeding, cells were UVA irradiated, MB+VL treated, or left untreated (as a control) and subsequently immediately infected with non-treated AdHCMVlacZ at an MOI of 20–40 PFU per cell.
For UVA exposure, growth medium was aspirated from the cells and replaced with 40 μl of PBS prior to UVA treatment. Following UVA treatment at a fluence rate of 110 J/m2/sec, the PBS was removed and cells were infected with 40 μl of viral suspension.
For MB+VL treatment, the cells were treated with 10 μg/ml (27 μM) MB in a volume of 100 μl of PBS for 1 h prior to VL treatment performed using a 1000-W bulb (General Electric, GE R1000) at a distance of 41 cm from the source for a specified time. The PBS+MB was then aspirated and the cells rinsed with 100 μl of unsupplemented growth medium prior to infection with 40 μl of viral suspension.
Following viral absorption for 90 min, cells were re-fed with the appropriate supplemented medium and allowed to incubate at 37°C for 12, 24 and 44 h prior to harvesting for β-gal (as previously described for HCR experiments).
Colony survival assays
SW480 and LoVo cells were seeded in six-well plates (Falcon) at a density of 400 cells per well. For UVA exposure of cells, at 6–8 h after seeding (after which time the cells had adhered to the plate), medium was replaced with 1 ml of PBS and then cells were UVA irradiated at a fluence rate of 100 J/m2/sec (or left untreated). Cells were then re-fed with 2 ml of the appropriate medium and incubated at 37°C, 5% CO2, in a humidified incubator for 9–10 days.
For MB+VL exposure of cells, at 6–8 h after seeding, the medium was aspirated from each well and replaced with 1 ml of supplemented α-MEM containing the desired concentration (0–80 μg/ml) of MB. MB concentrations were prepared using a set volume of supplemented α-MEM and variable volumes of PBS and MB to obtain the desired 1-ml volume. Plates were incubated again, this time for an overnight period (20–24 h) and kept under aluminium foil to minimize the effects of ambient lighting. Following the 20- to 24-h incubation period, the medium was aspirated from each well and the cells were washed with 1 ml of warmed PBS. The PBS was then aspirated from the wells and replaced with another 2 ml of warmed PBS. Plates requiring exposure to VL were then placed under a 1000-W bulb (General Electric, GE R1000) at a distance of 83 cm from the source for a specified period of time. After illumination (or after adding PBS in the case of non-illuminated plates), the overlaying PBS was aspirated from each well and replaced with 2 ml of supplemented α-MEM. Plates were placed under aluminium foil in humidified air at 37°C and 5% CO2 for an incubation period of 9–10 days.
After the incubation period, colonies were stained with ∼1 ml of crystal violet solution (63% absolute ethanol, 27% H2O, 10% methanol, 5 g/l crystal violet) and counted. Colonies containing a minimum of 32 cells were considered to have survived treatment.
Stock MB solution was prepared by dissolving 0.02 g methylene blue trihydrate (colour index 52015; methylthionine chloride, C16H18ClN3S·3H2O, molecular weight 373.9) in 20.0 ml PBS [140 mM NaCl, 2.5 mM KCl, 10 mM Na2HPO4, 1.75 mM KH2PO4 (pH 7.4)] that was warmed to 37°C. The resulting mixture was filter sterilized in the dark through a 0.2-μm filter to create a 1000 μg/ml solution. Stock solution was then aliquoted and stored at −20°C in the dark.
HCR of reporter gene expression of UVA-irradiated and MB+VL-treated AdHCMVlacZ in MMR-proficient and -deficient cell lines
Typical survival curves of β-gal activity for UVA-irradiated and MB+VL-treated AdHCMVlacZ in hMSH2-deficient LoVo cells and their SW480 hMSH2-proficient counterparts (left panels) as well as in hMLH1-deficient HCT116 cells and their HCT116-chr3 hMLH1-proficient counterparts (right panels) cells are shown in Figures 1 and 2, respectively. It can be seen that LoVo cells have a significantly reduced HCR compared to SW480 cells, whereas no significant differences were observed between the HCT116 and HCT116-chr3 cell lines. This indicates a deficiency in the repair of UVA-induced and MB+VL-induced oxidative DNA damage in hMSH2-deficient but not in hMLH1-deficient cells.
UVA-enhanced and MB+VL-enhanced expression of the undamaged reporter gene in MMR-proficient and -deficient cell lines
We have reported previously that UVC-enhanced expression of the undamaged reporter gene is induced at lower UVC fluences to cells and at higher levels in TCR-deficient compared to TCR-proficient human fibroblasts indicating that persistent damage in active genes triggers increased activity from the HCMV-driven reporter construct (35,36). Using the same experimental protocol, we have examined expression of the undamaged reporter gene in UVA-treated and MB+VL-treated cells. Results for expression of the undamaged reporter gene in MMR-proficient and -deficient colon carcinoma cells are shown in Figure 3 for pre-treatment of cells with UVA and in Figure 4 for pre-treatment of cells with MB+VL. Expression of β-gal for the undamaged AdHCMVlacZ virus was examined at multiple time points after infection. At the 12-, 24- and 44-h time points, expression levels were consistently observed at higher levels in LoVo compared to SW480 cells following UVA exposure and MB+VL exposure to cells, suggesting a deficiency in the repair of oxidative DNA damage from the transcribed strand of active genes in LoVo compared to SW480 cells (top three panels of Figures 3 and 4). In contrast, little or no difference in β-gal expression was observed between HCT116 and its MMR-proficient HCT116-chr3 counterpart (bottom 3 panels of Figures 3 and 4). This suggests little or no repair deficiency for UVA-induced and MB+VL-induced DNA damage in the hMLH1-deficient HCT116 cell line compared to hMLH1-proficient HCT116-chr3 cells.
Clonogenic survival of SW480 and LoVo cells in response to UVA exposure and MB+VL
Survival curves for SW480 and LoVo colon carcinoma cells in response to UVA and MB+VL are shown in Figures 5A and B, respectively. It can be seen that LoVo cells have increased resistance to UVA and MB+VL compared to SW480 cells.
In the present work using the HCR and enhanced expression assay (Figures 1–4), we present data suggesting that hMSH2-deficient LoVo cells exhibit a deficiency in the removal of UVA-induced and MB+VL-induced DNA damage as compared to their hMSH2-proficient SW480 counterparts. By contrast, little or no differences were observed in HCR or enhanced expression of AdHCMVlacZ between the hMLH1-deficient HCT116 and hMLH1-proficient HCT116-chr3 counterpart. Taken together, these results suggest a substantial involvement of hMSH2 but little or no involvement of hMLH1 in the removal of oxidative damage (predominantly 8-oxoG) induced by UVA or MB+VL.
These finding are in agreement with results from several research groups that have suggested a role for hMSH2 and its functional homologues in the recognition and removal of 8-oxoG, a lesion that often mispairs with adenine (30) and is often a target of BER. Mouse embryonic fibroblasts and colon tumor lines that were Msh2−/−, as well as a human ovarian carcinoma line that was hMLH1−, both exhibited increased 8-oxoG levels before and after H2O2 treatment compared to wild-type cells (31). Using radiolabeled DNA in gel mobility shift assays, it was found that the MSH2-MSH6 heterodimer in S. cerevisiae forms complexes around 8-oxoG:A mispairs and is required for the removal of adenine misincorporated opposite 8-oxoG (31). Furthermore, the DNA glycosylase activities of hMYH were increased through the physical interaction with hMutSα (hMSH2-hMSH6 heterodimer) (32). Finally, various mispairs of 8-oxoG were found to activate hMutSα through binding affinity studies and its associated increases in ATPase and ADP → ATP exchange activites (33). It is therefore possible that the association between MMR and skin cancer results, in part at least, from an effect of MMR deficiency on the removal of UV-induced DNA damage by BER.
BER is thought to represent the most essential and diverse pathway in correcting damage induced by oxidizing and alkylating agents, mainly due to the incorporation of a large number of DNA glycosylases needed to recognize a variety of base modifications. At least 12 different human DNA glycosylases have been identified (44) with several functional homologues present in bacteria and yeast. In spite of comprehensive research into the nature of BER, there are no known naturally occurring human mutations that specifically confer defects in BER (unlike XP and CS in NER and hMSH2 and hMLH1 mutants in MMR). Conversely, no distinct human disease has been associated with defective BER (44).
UV irradiation as well as several other DNA-damaging agents induce expression from a number of cellular and viral promoters (reviewed in references 45,46). This response can be divided into two components: the ‘immediate’ response and the ‘delayed’ response (reviewed in reference 46). The immediate response arises at the cell membrane with the activation of cell surface receptors and their corresponding signaling cascades, and the strength of the immediate response is directly proportional to the strength of the initial UV exposure. The delayed response is observed several hours after UV exposure and the duration and strength of the delayed response are inversely related to the cells’ ability to repair transcriptionally active genes, which is mediated by the TCR sub-pathway of NER (47–49). Delayed responsive genes include numerous viral promoters including the human CMV immediate-early promoter, which controls expression of the lacZ reporter gene in the recombinant adenovirus Ad5HCMVlacZ (35,36,46,50).
We have reported previously that enhanced expression of the undamaged CMV-promoted reporter gene is induced by pre-treatment of cells with lower levels of a DNA-damaging agent and to higher expression levels in TCR-deficient compared to TCR-proficient human fibroblasts indicating that persistent damage in active genes (which is not repaired by the TCR pathway) triggers increased activity from the HCMV-driven reporter construct (35,36). Although the exact mechanism leading to the enhanced expression of the reporter gene is not known at present, results from our laboratory suggest a role for one or more of the pRb family members in mediating UV-induced expression from the CMV promoter (50). We have reported previously an increased expression of the CMV-promoted reporter gene in LoVo compared to SW480 cells, and in HCT116 compared to HCT116-chr3 cells following exposure to UVC, suggesting an involvement of both hMSH2 and hMLH1 in the TCR of UVC-induced DNA damage (27), consistent with the earlier report of Mellon et al. (6).
In the present work and using the same experimental protocol and the same human colon adenocarcinoma cell lines, we have examined expression of the undamaged reporter gene in UVA-treated and MB+VL-treated cells. SW480 and HCT116-chr3 cells are proficient in both hMSH2 and hMLH1 and following pre-treatment with MB+VL show a change in β-gal expression of the undamaged reporter gene of 0.5–1.0 for SW480 and 1.5–2.5 for HCT116-chr3 over the range of MB+VL treatment given (Figure 4). In contrast, the same range of MB+VL treatment of hMSH2-deficient LoVo cells resulted in an increased β-gal expression of the undamaged reporter gene of 4.2- to 6.8-fold. This suggests there is a greater amount of unrepaired MB+VL-induced DNA damage in the actively transcribing genes of LoVo compared to SW480 and HCT116-chr3 cells, consistent with a deficiency in the TCR of MB+VL-induced DNA damage (mainly 8-oxoG) in hMSH2-deficient LoVo cells. In contrast, following MB+VL treatment to cells, β-gal expression of the undamaged reporter gene was similar in hMLH1-deficient HCT116 cells and hMLH1-proficient HCT116-chr3 cells when β-gal was scored at 12 and 44 h after infection and only a small difference in β-gal expression was detected at the highest light exposure to cells (20 min) when infected cells were scored at 24 h after infection. This suggests little or no difference in TCR of MB+VL-induced DNA damage in hMLH1-deficient HCT116 cells compared to hMLH1-proficient HCT116-chr3 cells.
UVA treatment of cells resulted in greater β-gal expression in LoVo (1.2- to 1.4-fold at 400 kJ/m2) compared to SW480 cells (∼0.8 fold at 400 kJ/m2) at all times after infection was examined. This suggested reduced TCR of UVA-induced DNA lesions in hMSH2-deficient LoVo cells compared to hMSH2-proficient SW480 cells. In addition, following UVA treatment to cells, β-gal expression of the undamaged reporter gene was similar in hMLH1-deficient HCT116 cells and hMLH1-proficient HCT116-chr3 cells when β-gal was scored at 12 and 44 h after infection and only a small difference in β-gal expression was detected at the highest exposure to cells (400 kJ/m2) when infected cells were scored at 24 h after infection. This suggests little or no difference in TCR of UVA-induced DNA damage in hMLH1-deficient HCT116 cells compared to hMLH1-proficient HCT116-chr3 cells.
It should also be noted that following UVA treatment of cells, the increased expression in hMSH2- and hMLH1-proficient HCT116-chr3 cells (1.4- to 2.2-fold at 400 kJ/m2 to cells) was greater compared to that in hMSH2- and hMLH1-proficient SW480 cells (∼0.8-fold at 400 kJ/m2 to cells). We have also previously reported an increased expression in hMSH2- and hMLH1-proficient HCT116-chr3 cells compared to hMSH2- and hMLH1-proficient SW480 cells following pre-treatment of cells with UVC (27), suggesting a reduced TCR of UVC-induced DNA damage in HCT116-chr3 compared to SW480 cells. Since repair of UVC-induced CPDs in the transcribed strand of an active gene is similar in SW480 compared to HCT116-chr3 cells (6,25), our results showing increased expression of the reporter gene in HCT116-chr3 compared to SW480 cells following UVC treatment to cells (27) suggest that one or more of the factors responsible for enhanced expression of the reporter gene is altered in SW480 compared to HCT116-chr3 cells. Notwithstanding, the results of Figures 3 and 4 are consistent with a deficiency in TCR of UVA- and MB+VL-induced oxidative DNA damage in LoVo compared to SW480 cells. This pair of human colon adenocarcinoma cells has been used in previous studies in several laboratories to examine the role of hMSH2 in DNA repair (6,25,27). However, since hMSH2-deficient LoVo cells are not an isogenic cell line for hMSH2-proficient SW480 cells, differences in the response of these two cell lines to UVC, UVA and MB+VL could, in part at least, reflect differences other than the hMSH2 deficiency in LoVo cells.
Since the function of the hMSH2 protein (in the form of the hMSH2-hMSH6 heterodimer hMutSα) is to recognize and bind to base–base mispairs as well as insertion/deletion loops (51), it is quite conceivable that other mispairs such as 8-oxoG:A, 8-oxoG:G and 8-oxoG:T serve as ‘mismatch’-binding substrates for hMutSα. Indeed, it has been reported that these particular mispairs activate hMutSα through increased binding affinity of the heterodimer for these lesions as well as increases in its associated ATPase and ADP → ATP exchange activities (33).
Although in the past it has been widely believed that UVA induces predominantly oxidative DNA lesions, recent studies into the spectrum of lesions produced by UVA have shown that even UVA is capable of producing significant levels of CPDs (37) even at levels greater than 8-oxoG production (38,52). It is therefore possible that the reduced HCR capacity for the UVA-treated reporter gene in LoVo compared to SW480 cells (Figure 1) reflects a deficiency in the repair of CPDs and 6-4 PPs by NER as reported previously (27). Notwithstanding, the spectrum of lesions induced by UVA is significantly different than that produced by UVC, which produces primarily CPDs and 6-4 PPs. This difference in the spectrum of lesions is reflected in the differential involvement of hMLH1 in the HCR of UVC-induced lesions (27) but not in lesions induced by UVA (Figure 1). This is further supported by the results achieved using MB+VL as a DNA-damaging agent since it has been shown to induce predominately 8-oxoGs (39,40) (Figures 2 and 4).
Our desire to utilize UVA radiation as a source of producing oxidative DNA lesions, namely 8-oxoG, was due to the clinical implications of the omnipresence of UVA in our environment from the sun, since chronic sunlight exposure is unambiguously associated in increased cancer risk. In addition to the HCR and enhanced expression assays, we also performed colony survival assays with the SW480 and LoVo cells to test their sensitivity to UVA and MB+VL (Figure 5). The hMSH2-deficient LoVo cells were actually more resistant to UVA and MB+VL exposure compared to hMSH2-proficient SW480 cells. This is consistent with previous reports in which Msh2−/− mouse embryonic stem cells and mouse fibroblasts exhibited increased survival in response to ionizing radiation compared to their wild-type counterparts, presumably due to a failure to execute apoptosis in response to radiation exposure (53,54). Indeed, it has been suggested that the Msh2 gene protects cells from UVB-induced tumorigenesis by facilitating apoptosis and p53 activation (55). Additionally, there is evidence that hMSH2 protein levels are altered in human non-melanoma skin cancers (9). These effects may also be occurring in UVA-irradiated LoVo cells, which would potentially have clinical implications for individuals that are genetically deficient in hMSH2, resulting in their predisposition to cancer as a result of chronic UVA exposure.
In summary, HCR of β-gal activity for UVA-irradiated and MB+VL-treated AdHCMVlacZ virus was substantially reduced in LoVo compared to SW480 cells, but not in HCT116 compared to HCT116-chr3 cells. Enhanced expression of β-gal for untreated AdHCMVlacZ virus was significantly greater for cells pre-treated with UVA and MB+VL in LoVo compared to SW480 cells but generally not in HCT116 compared to HCT116-chr3 cells. These results suggest a substantial involvement of hMSH2 but little or no involvement of hMLH1 in BER of UVA- and MB+VL-induced oxidative DNA damage. In addition, hMSH2-deficient cells are more resistant compared to hMSH2-proficient cells following UVA and MB+VL exposure, presumably due to a failure to execute apoptosis in response to UVA and MB+VL. These results suggest that the hMSH2 gene may protect cells from UV-induced tumorigenesis due to a role in facilitating apoptosis as well as a role in repairing oxidative DNA damage.
This work was supported by the National Cancer Institute of Canada with funds from the Canadian Cancer Society.