A majority of human melanoma cell lines exhibits an S phase-specific defect in excision of UV-induced DNA photoproducts

Abstract

It is well established that efficient removal of highly-promutagenic UV-induced dipyrimidine photoproducts via nucleotide excision repair (NER) is required for protection against sunlight-associated malignant melanoma. Nonetheless, the extent to which reduced NER capacity might contribute to individual melanoma susceptibility in the general population remains unclear. Here we show that among a panel of 14 human melanoma strains, 11 exhibit significant inhibition of DNA photoproduct removal during S phase relative to G0/G1 or G2/M. Evidence is presented that this cell cycle-specific NER defect correlates with enhanced apoptosis and reduced clonogenic survival following UV irradiation. In addition, melanoma strains deficient in S phase-specific DNA photoproduct removal manifest significantly lower levels of phosphorylated histone H2AX at 1 h post-UV, suggesting diminished activation of ataxia telangiectasia and Rad 3-related (ATR) kinase, i.e., a primary orchestrator of the cellular response to UV-induced DNA replication stress. Consistently, in the case of DNA photoproduct excision-proficient melanoma cells, siRNA-mediated depletion of ATR (but not of its immediate downstream effector kinase Chk1) engenders deficient NER specifically during S. On the other hand simultaneous siRNA-mediated depletion of ataxia telangiectasia mutated kinase (ATM) and DNA-dependent protein kinase catalytic subunit (DNA-PKcs) exerts no significant effect on either phosphorylation of H2AX at 1 h post-UV or the efficiency of DNA photoproduct removal. Our data suggest that defective NER exclusively during S phase, possibly associated with decreased ATR signaling, may constitute an heretofore unrecognized determinant in melanoma pathogenesis.

Figure 1. Cell cycle-specific 6–4PP excision in normal melanocytes and melanoma cell lines.

A) Graphical representation of 6–4PP removal as a function of cell cycle at 6 h post-UVC for 3 primary melanocyte strains and XPA-deficient skin fibroblasts. B) Same as A, but for 14 melanoma strains. In the case of all SPR-deficient strains, excision during S phase is significantly slower relative to other phases (p<0.01; two-tailed paired t-test). C) Representative bivariate dot plots showing WM35 and WM3248 (SPR-proficient and -deficient, respectively) either UV- or mock-irradiated as indicated, and stained with PI and anti-6-4PP antibody. Cells were gated in each phase of the cell cycle as shown for WM35 (no UV). D) Graphical representation of 6–4PP excision as a function of cell cycle at 6 h post-UVB (290–320-nm; 300 J/m²) in WM35 and WM3248. For all panels in this figure, values represent the mean ± SEM of three independent experiments.

Figure 2. Cell cycle progression in melanoma cells post-UVC.

Cells were irradiated with UVC or mock-irradiated and then pulse-labeled with BrdU to mark cells in S-phase at the time of irradiation. Cultures were harvested immediately (0 h) or further incubated for 6 h. Cells were stained with PI and anti-BrdU (Alexa-647), and analyzed by flow cytometry. A) SPR-proficient WM35. BrdU+ and BrdU- populations were gated as shown in the bivariate dot plots (top panels). Histograms represent DNA content of BrdU+ and BrdU- cells under each condition (lower panels). B) Same as A, but for SPR-deficient WM3248. Similar data for all other melanoma cell lines used in this study are presented in Figure S1.

Figure 3. Cell cycle-specific CPD excision in melanoma cell lines.

Cells were irradiated with 10/m² UVC and then labeled with BrdU for 1 h (except for the 0 h time point, where BrdU labeling was performed for 30 min prior to UV). At the indicated times post-UV, cells were stained with anti-CPD (FITC), PI, and anti-BrdU (Alexa-647) and analyzed by flow cytometry. A) Bivariate plots of WM35 showing the distribution of BrdU-positive cells at different times post-UV. The S' population indicated with an arrow at 24 h post-UV have traversed mitosis after increasing their DNA content and are excluded from the analysis. The extent of CPD removal is compared for cells in S phase at the time of irradiation (designated S), vs. cells in G1 at the time of irradiation which includes a minor proportion of cells that were initially in G2 and have migrated into the G1 compartment during the post-UV incubation period (designated G1/G2) (see text for details). B) Graphical representation of CPD excision in WM35 (top), WM3248 (middle), and XPA control skin fibroblasts (bottom). * and **, two-tailed paired t-test comparing CPD excision during S-phase vs G1/G2; p<0.02 and 0.001 at 12 and 24 h, respectively. For all panels in this figure, values represent the mean ± SEM of three independent experiments.

Figure 4. Determination of clonogenic survival and sub-G1 DNA content in SPR+ vs SPR- strains post-UVC.

A) Clonogenic survival in response to various doses of UVC was calculated as the number of surviving colonies (containing at least 50 cells) for irradiated cells relative to mock-irradiated counterparts. Values represent the mean +/− SEM of four independent experiments. B) Percentage of sub-G1 cells at different times after exposure to 10 J/m² of UVC in melanoma cell lines. Representative cell cycle profiles are shown for the SPR-deficient strain WM983B at 0 h and 72 h (top). Values represent the mean +/− SEM of three independent experiments.

Figure 5. Correlation between defective SPR and reduced H2AX phosphorylation in melanoma cell lines.

A) Representative bivariate dot plots illustrating relative γH2AX induction as a function of cell cycle at 1 h post-UV in ATR-proficient 1BR vs. the ATR-hypomorphic counterpart F02-98. Cells were gated in each phase of the cell cycle as shown for 1BR (no UV). B) Graphical representation of γH2AX induction as a function of cell cycle for 1BR and F02-98 at 1 h post UV. * p<0.001; two-tailed unpaired t-test comparing mean γH2AX induction in S-phase for 1BR vs F02-98. C) Graphical representation of γH2AX induction during S phase in 1BR vs F02-98 (same as panel B), compared with 3 normal melanocyte lines (MC), 3 SPR+ melanoma strains and 9 SPR- melanoma strains. Values for melanocytes and melanoma strains are illustrated as a box plot. * p<0.002; two-tailed unpaired t-test comparing mean γH2AX induction for SPR+ vs SPR- strains. D) ATR and ATRIP protein levels were determined in whole cell extracts for various melanoma strains by western blot. YY1 is used as a loading control.

Figure 6. Downregulation of ATR but not Chk1 engenders defective SPR in melanoma cells.

A) Each of the three SPR-proficient strains in our collection were incubated with siRNA pools against ATR or Chk1, or with non-targeting control siRNA, and treated with UVC or mock-irradiated. Protein levels for ATR, Chk1, and p-Chk1(S345) were determined by western blotting. YY1 and actin were used as loading control for ATR and Chk1, respectively. B) SPR-proficient strains were incubated with combined siRNA pools targeting both ATM and DNA-PKcs (A+D) or with non-targeting control siRNA, and treated with 6 Gy of IR or mock-irradiated. Protein levels for ATM, DNA-PKcs, and γH2AX were determined by western blotting. YY1 and GAPDH were used as loading controls for ATM/DNA-PKcs and γH2AX, respectively. C) γH2AX induction as a function of cell cycle at 1 h post-UV for siATR knockdown and ATM/DNA-PKcs double knockdown vs. non-targeting siRNA control. * p<0.01, two-tailed unpaired t-test comparing γH2AX induction in S-phase for siATR knockdown cells vs. non-targeting siRNA control. D) Cell cycle-specific excision of 6–4PP at 6 h post-UV in ATR, Chk1 and ATM/DNA-PK siRNA knockdown cells vs. non-targeting siRNA control. * p<0.005, two-tailed paired t-test comparing the extent of 6–4PP removal in G1 vs S. For all panels in this figure, values represent the mean ± SEM of three independent experiments.