If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Department of Cell Maintenance, Institute of Molecular Embryology and Genetics, Kumamoto University, Kumamoto, JapanJapan Society for the Promotion of Science (JSPS), Tokyo, Japan
Correspondence: Satoshi Tateishi, Department of Cell Maintenance, Institute of Molecular Embryology and Genetics, Kumamoto University, Honjo 2-2-1 Chuoku, Kumamoto 860-0811, Japan.
Defects in DNA polymerase Eta (Polη) cause the sunlight-sensitivity and skin cancer-propensity disorder xeroderma pigmentosum variant. The extent to which Polη function depends on the upstream E3 ubiquitin ligase Rad18 is controversial and has not been investigated using mouse models. Therefore, we tested the role of Rad18 in UV-inducible skin tumorigenesis. Because Rad18 deficiency leads to compensatory DNA damage signaling by Chk2, we also investigated genetic interactions between Rad18 and Chk2 in vivo. Chk2–/–Rad18–/– mice were prone to spontaneous lymphomagenesis. Both Chk2–/– and Chk2–/–Rad18–/– mice were prone to UV-B irradiation-induced skin tumorigenesis when compared with wild-type (WT) animals, but unexpectedly Rad18–/– mice did not recapitulate the skin tumor propensity of Polη mutants. UV-irradiated Rad18–/– cells were more susceptible to G1/S arrest and apoptosis than WT cultures. Chk2 deficiency alleviated both UV-induced G1/S phase arrest and apoptosis of WT and Rad18–/– cells, but led to increased genomic instability. Taken together, our results demonstrate that the tumor-suppressive role of Polη in UV-treated skin is Rad18 independent. We also define a role for Chk2 in suppressing UV-induced skin carcinogenesis in vivo. This study identifies Chk2 dysfunction as a potential risk factor for sunlight-induced skin tumorigenesis in humans.
UV irradiation induces DNA lesions including cyclobutane pyrimidine dimers. To maintain genomic integrity, cells employ more than 150 DNA repair enzymes that recognize and repair various types of DNA lesions (
). However, DNA repair processes can be slow and incomplete, and consequently, a number of unrepaired DNA lesions may remain in the template strand during DNA replication. Conventional replicative DNA polymerases cannot replicate damaged templates and therefore become stalled at UV-induced DNA lesions. Persistent stalling of DNA polymerases can lead to replication fork “collapse” that generates double-strand DNA breaks (DSBs) (
). When replicating damaged templates, cells rely on two specialized genome maintenance mechanisms termed “template switching” and “DNA damage tolerance” to prevent the generation of dangerous genome-destabilizing DSBs. DNA damage tolerance uses specialized translesion synthesis (TLS) DNA polymerases to perform that replicative bypass of bulky DNA lesions such as UV-induced cyclobutane pyrimidine dimers.
Xeroderma pigmentosum variant (XP-V) is a rare recessive inherited disease characterized by increased incidence of the formation of various forms of cutaneous malignancies at sun-exposed skin areas (
). Cells from patients with XP-V have a defective TLS of UV-induced DNA lesions. The gene responsible for XP-V encodes the DNA polymerase eta (Polη), a TLS polymerase that has the ability to bypass cyclobutane pyrimidine dimers (
). The skin cancer propensity of individuals with XP-V and Polη-deficient mice most likely results from compensatory error-prone bypass of UV-inducible lesions by alternative TLS polymerases, and from error-prone repair of DSBs that arise from fork collapse when Polη is absent. It is well accepted that Polη activity in UV-irradiated cells is regulated by the DNA repair protein Rad18. Rad18 is an ubiquitin ligase that monoubiquitinates proliferating cell nuclear antigen (PCNA, a DNA polymerase processivity factor) in UV-irradiated cells. Polη preferentially associates with the monoubiquitinated form of PCNA. Thus, formation of the PCNA-Ub/Polη complex is thought to facilitate the engagement of Polη with replication forks and promote TLS of UV-induced lesions (
). Therefore, biochemical studies implicate Rad18 as a critical proximal component of the Polη-mediated TLS pathway.
However, Polη recruitment and recovery from replication fork stalling are compromised but not completely attenuated in cultured cells from a knockin mouse harboring ubiquitination-resistant mutant PCNA (
). Thus, although cell culture studies show that Polη-mediated TLS is sometimes dissociable from Rad18-mdediated PCNA monoubiquitination, the extent to which Polη function depends on Rad18 has not been tested in a physiological setting. Rad18-deficient mice display distorted hair follicles (
), indicating that Rad18 plays key roles in the skin. However, it is unknown whether Rad18 deficiency recapitulates the hallmark pathologies of XP-V. Accordingly, a major goal of this study was to determine the role of Rad18 in preventing UV-induced skin tumorigenesis and suppressing the XP-V phenotype in vivo.
A corollary and related goal of this investigation was to define genetic interactions between Rad18 and another mediator of DNA damage signaling, namely checkpoint kinase 2 (human CHEK2; murine Chk2). CHEK2/Chk2 is a critical participant in the DSB response. We previously showed that Chk2 phosphorylation is very high in UV-irradiated Rad18–/– cells relative to wild-type (WT) cells (
). Those results suggest that UV irradiation in TLS-deficient Rad18–/– cells leads to fork collapse and DSB formation that is accompanied by elevated ATM/Chk2 signaling. Interestingly, individuals heterozygous for the nonfunctional frameshift CHEK2∗1100delC germline alteration have a twofold increased risk of developing malignant melanoma (
). However, Chk2 dysfunction has never been investigated as a risk factor for UV-induced skin tumorigenesis. Accordingly, we have used mouse genetic models to define the relationship between Rad18 and Chk2 and their roles in UV-induced skin carcinogenesis.
Results
Characterization of Chk2–/– and Rad18–/– mice
To characterize genetic interactions between Rad18 and Chk2, we generated WT, Rad18–/–, Chk2–/–, and Chk2–/–Rad18–/– mice. All genotypes were viable, fertile (for at least 2 months after birth), and grew normally for at least 12 months (data not shown). We detected spontaneous splenic lymphoma in 25% of the 12-month-old Chk2–/–Rad18–/– mice, whereas no tumors were found in other genotypes (Supplementary Figure S1a and b online). In all mice with splenic lymphoma, we detected liver infiltrates identified as B-cell lymphoma based on staining with the anti-B220 antibody (Supplementary Figure S1c). We next transplanted these lymphoma cells into immunodeficient mice (Supplementary Figure S1d and e). Four of five mice suffered from splenomegaly with massive cell infiltration. Therefore, we conclude that these cells are aggressive lymphoma cells.
Rad18 deficiency does not predispose to UV-induced skin tumorigenesis
To examine the possibility that loss of Rad18 causes XP-V pathogenesis, we chronically exposed the shaved dorsal skin of WT and Rad18–/– mice to UV-B light (Figure 1a). Control nonirradiated WT or Rad18–/– mice showed no spontaneously arising tumors on their dorsal skin. Twenty-five weeks after the beginning of the periodical UV-B irradiation, both WT and Rad18–/– mice started to develop ulcers and skin tumors around the ulcers. After 38 weeks of irradiation, approximately 50% of WT and Rad18–/– mice developed multiple dorsal skin tumors. However, there was no statistically significant difference in tumor incidence between WT and Rad18–/– mice. In similar experiments, Polη mutant mice were highly prone to skin tumors when compared with Polη-proficient mice (
). We conclude that Rad18 is dispensable for normal Polη functions in preventing skin carcinogenesis in vivo.
Figure 1Skin tumor predisposition of Chk2–/– mice chronically exposed to UV-B irradiation. (a) Tumorigenesis induced by chronic treatment with UV-B. Kaplan-Meier curves of mice free of skin tumors after chronic UV-B irradiation (2 kJ/m2/day, 3 days irradiation/week). The statistical significance of the differences between experimental groups was assessed using the log-rank test. (b) Histopathological examination of skin tissues of Rad18–/– mice with no skin tumors. (c, d) Histopathological examination of UV-B-induced well-differentiated squamous cell carcinoma predominantly observed in Chk2–/– mice (c) and Chk2–/–Rad18–/– mice (d). (e) Histopathological examination of UV-B-induced sarcoma found in a Rad18–/– mouse. Scale bar = 100 μm. WT, wild type.
Skin carcinogenesis propensity of Chk2 mutant mice
On the basis of cell culture studies showing compensatory Chk2 signaling on UV irradiation in Rad18–/–cells, we considered the possibility that Rad18 and Chk2 might have redundant roles in protecting against UV-induced skin carcinogenesis. Moreover, although CHEK2∗1100delC mutation confers the increased risk of melanoma, it is unknown whether CHEK2 dysfunction is a risk factor for UV exposure-induced skin tumorigenesis. Therefore, we asked whether co-deletion of Chk2 might reveal Rad18 functions in skin tumor suppression. We chronically exposed the shaved dorsal skin of Chk2–/– mice and Chk2–/–Rad18–/– mice to UV-B light. Although none of the nonirradiated Chk2–/– or Chk2–/–Rad18–/– mice showed signs of tumors on their dorsal skin, both UV-B-irradiated Chk2–/– and Chk2–/–Rad18–/– mice started to develop skin tumors around the nodules 25 and 15 weeks after the beginning of treatment, respectively (Figure 1a). Approximately 60% of Chk2–/– and Chk2–/–Rad18–/– mice developed multiple tumors on their dorsal skin 28 weeks after irradiation, whereas no tumors were found in WT mice. The skin tumor formation incidences of Chk2–/– and Chk2–/–Rad18–/– mice were higher than those of WT and Rad18–/– mice, respectively, and the differences were statistically significant. We conclude that Chk2–/– mice chronically exposed to UV irradiation were predisposed to skin tumors, indicating that a deficiency of Chk2 function is a risk factor for skin tumor formation after UV-B irradiation. Chk2–/–Rad18–/– mice seemed to develop skin tumors earlier than Chk2–/– mice, although the difference was not statistically significant. Therefore, Chk2 deficiency predisposes to overall incidence of UV-induced skin carcinogenesis regardless of Rad18 genotype. Comparing the histopathological results of the UV-B exposed mice, we found that there was a higher incidence of well-differentiated squamous cell carcinomas in Chk2–/–Rad18–/– mice than in WT and Rad18–/– mice (P < 0.05) (Table 1). There was a higher incidence of sarcomas in WT and Rad18–/– mice than in Chk2–/– and Chk2–/–Rad18–/– mice (P < 0.01).
Table 1Histopathological examination of UV-B-exposed mice
The statistical significance of well-differentiated squamous cell carcinoma (SCC) identified by histological diagnosis was measured with Fisher’s exact probability test. Bold text indicates P < 0.05 (Chk2–/–Rad18–/– vs. WT; Chk2–/–Rad18–/– vs. Rad18–/–).
The statistical significance of sarcoma identified by histological diagnosis was measured with Fisher’s exact probability test. Bold text indicates P < 0.01 (WT vs. Chk2–/–; Rad18–/– vs. Chk2–/–Rad18–/–).
Melanoma
Keratoacanthoma
Actinic karatosis
WT
20
1
6
9
0
2
0
Rad18–/–
19
1
6
6
1
0
0
Chk2–/–
19
2
7
1
0
0
1
Chk2–/–Rad18–/–
20
0
14
0
0
0
1
1 The number of mice of analyzed cohorts.
2 The statistical significance of well-differentiated squamous cell carcinoma (SCC) identified by histological diagnosis was measured with Fisher’s exact probability test. Bold text indicates P < 0.05 (Chk2–/–Rad18–/– vs. WT; Chk2–/–Rad18–/– vs. Rad18–/–).
3 The statistical significance of sarcoma identified by histological diagnosis was measured with Fisher’s exact probability test. Bold text indicates P < 0.01 (WT vs. Chk2–/–; Rad18–/– vs. Chk2–/–Rad18–/–).
The exacerbated G1/S checkpoint of Rad18–/– cells is Chk2 mediated
To assess the role of Chk2 in preventing UV-B-induced skin tumorigenesis, we generated embryonic stem (ES) cells from WT, Rad18–/–, Chk2–/–, or Chk2–/–Rad18–/– mice. To quantify the G1/S phase checkpoint, we measured the ratio of S phase to G1 plus S phase fractions in ES cells from each genotype after UV-C irradiation. The ratios of G1, S, and G2/M fractions were similar among the ES cells from each genotype cultured without UV-C irradiation (Figure 2a). The ratio of S phase to G1 plus S phase fractions decreased to approximately 92% and 84% at 18 h after UV-C irradiation in WT and Rad18–/– ES cells, respectively (Figure 2e). In contrast, those in Chk2–/– and Chk2–/–Rad18–/– ES cells kept almost constant after UV-C irradiation. Therefore, Chk2 mediates the G1/S checkpoint in response to UV-C-induced DNA damage. Because our skin carcinogenesis experiments were performed with UV-B-irradiated mice (not UV-C), we also determined the impact of Rad18 and Chk2 deficiencies on cell cycle responses to UV-B irradiation (Supplementary Figure S2a online). The ratio of S phase to G1 plus S phase fractions decreased to approximately 62% and 42% at 15 hours after UV-B irradiation in WT and Rad18–/– ES cells, respectively. In contrast, those in Chk2–/– and Chk2–/–Rad18–/– ES cells rather increased. Therefore, similar to results obtained with UV-C-irradiated cells, Chk2 mediates the G1/S checkpoint in response to UV-B treatment. Furthermore, to exclude the possibility of cell type-specific effects (of Rad18 and Chk2 on DNA damage responses), we also used murine embryonic skin fibroblasts (MEFs) with different genotypes to determine the impact of Rad18 and Chk2 deficiencies on cell cycle responses to UV-B irradiation (Supplementary Figure S3a and b online). The ratio of S phase to G1 plus S phase fractions decreased to approximately 64% and 78% at 48 hours after UV-B irradiation in WT and Rad18–/– MEFs, respectively. In contrast, those in Chk2–/– and Chk2–/–Rad18–/– MEFs decreased to only approximately 84% and 99% at 48 hours. Therefore, similar to results obtained with UV-B- or UV-C-irradiated ES cells, Chk2 mediates the G1/S checkpoint in MEFs after UV-B treatment.
Figure 2Cell cycle analysis of ES cells after UV-C irradiation. (a) The top plots show the cell cycle in ES cells of each genotype without UV-C irradiation, and the bottom plots show the cell cycle of ES cells of each genotype after 18 hours of incubation after UV-C irradiation (2 J/m2). (b–d) Time course of cell cycle of G1 (b), S (c), and G2/M (d) profiles of each genotype was analyzed by culturing ES cells after UV-C irradiation. The assays were performed in triplicate, and relative means ± standard deviation are shown. (e) Time course of S/(G1 + S) ratios was obtained by dividing the results of (c) by (b + c). ES, embryonic stem; WT, wild type.
UV sensitivity of Rad18–/– cells is rescued by loss of Chk2
To further elucidate the interplay between Rad18 and Chk2 in ES cells, we examined the UV sensitivity of ES cells. Rad18–/– ES cells were more sensitive to UV-C irradiation than WT ES cells (Figure 3a). Interestingly, Chk2–/– ES cells were UV-C resistant compared with WT ES cells. Similar to the relative UV-C resistance of Chk2–/– ES cells, Chk2–/–Rad18–/– ES cells were more UV-C resistant than Rad18–/– ES cells, indicating that the UV-C sensitivity of Rad18–/– ES cells was rescued at least partially by the loss of Chk2. These results suggest that Chk2 induces cell death in response to UV-C irradiation. Chk2 is involved in the induction of apoptosis in cells and tissues after ionizing radiation (
). We performed TUNEL assays to measure cell death-associated DNA fragmentation in control and UV-C-treated ES cells (Figure 3b). Surprisingly, Rad18–/– ES cells exhibited fourfold higher numbers of TUNEL-positive cells when compared with WT cells, both basally and after UV-C irradiation. Therefore, Rad18 protects ES cells against UV-induced genotoxicity as well as intrinsically arising replicative stress. Furthermore, WT ES cells exhibited fivefold higher numbers of TUNEL-positive cells than those of Chk2–/– ES cells both basally and after UV-C irradiation. Similarly, Rad18–/– ES cells, exhibited threefold higher numbers of TUNEL-positive cells than those of Chk2–/–Rad18–/– ES cells regardless of whether the cells were UV-C irradiated. These results suggest that Chk2 directs cell death in response to UV-induced genotoxicity as well as intrinsically arising replicative stress. To complement the results of the TUNEL assays, we quantified cell death in the ES cells of each genotype by flow cytometric measurements of the sub-G1 (apoptotic) fractions relative to the entire cell population (Figure 3c). In all four genotypes, the size of the sub-G1 population relative to total cell number plateaued 18 hours after UV-C irradiation and decreased to basal levels after 24 hours. In Rad18–/– ES cells, the sub-G1 population was 1.5-fold larger than the sub-G1 population of WT ES cells 18 hours after UV-C treatment. These results further show that Rad18 protects ES cells from UV-induced death. In WT and Rad18–/– ES cells, the sub-G1 populations were 1.7-fold and 3.0-fold larger than those of Chk2–/– and Chk2–/–Rad18–/– ES cells, respectively (also determined 18 hours after UV-C treatment). Next, we performed TUNEL assays to evaluate the apoptotic response to UV-B irradiation in ES cells. As shown in Supplementary Figure S2b, the number of TUNEL-positive WT and Rad18–/– ES cells increased to approximately 1.8-fold and 1.2-fold, respectively, whereas the numbers of apoptotic Chk2–/– and Chk2–/–Rad18–/– ES cells decreased. Furthermore, to exclude the possibility of cell type-specific effects of Rad18 and Chk2 on apoptosis, we also used MEFs with different genotypes to evaluate the apoptotic response to UV-B irradiation. To quantify apoptotic populations, we determined the number of sub-G1 cells and expressed these as a percentage of the total cell number. As shown in Supplementary Figure S3c, the percentage of sub-G1 cells of WT and Rad18–/– MEFs increased by approximately 1.9-fold and 1.6-fold, respectively, whereas the percentage of sub-G1 cells in Chk2–/– and Chk2–/–Rad18–/– MEFs remained constant after UV-B irradiation. We conclude that Chk2 mediates cell death (in both WT and Rad18–/– ES cells) after acquisition of UV-induced DNA damage.
Figure 3Chk2-directed cell death induction in response to UV-C irradiation. (a) Survival of ES cells in response to UV-C irradiation. The assays were performed in triplicate. Data are presented as mean survival rates ± standard error. (b) ES cells were cultured for 14 hours after UV-C irradiation (2 J/m2). Tunnel-positive cells were counted with the cells with or without UV-C irradiation. **The statistical significance of formation of tunnel-positive cell efficiency was identified (P < 0.01). (c) Time course of rates of the sub-G1 fraction in the total cell cycle was measured in the ES cells after UV-C irradiation (2 J/m2) using flow cytometry. ES, embryonic stem; WT, wild type.
Chk2 and Rad18 maintain genomic DNA stability in UV-irradiated cells
Carcinogenesis is driven by genetic change. To investigate the mechanisms by which the interplay between Rad18 and Chk2 affects spontaneous and UV-induced tumorigenesis, we sought to determine how individual and combined mutations in Rad18 and Chk2 impact genomic stability both basally and in response to UV irradiation. Therefore, to investigate genomic stability we measured frequencies of micronucleus formation in ES cells of each genotype (Figure 4a). Under normal culture conditions, the micronucleus formation frequency of Chk2–/–Rad18–/– ES cells was significantly elevated when compared with WT ES cells (Figure 4b, upper panel). After UV-C irradiation, micronuclei formation frequencies in both Chk2–/– and Chk2–/–Rad18–/– ES cells were significantly increased when compared with WT cells (Figure 4b, lower panel). Next, we measured frequencies of micronucleus formation in the ES cells after UV-B irradiation (Supplementary Figure S2c). After UV-B irradiation, the frequencies of micronuclei were significantly increased in Chk2–/– and Chk2–/–Rad18–/– ES cells when compared with WT ES cells. Furthermore, to exclude the possibility of cell type-specific effects of Rad18 and Chk2 on micronucleation, we also measured frequencies of micronucleus formation in MEFs with different genotypes. As shown in Supplementary Figure S3d, the frequencies of micronuclei were significantly increased in Chk2–/– and Chk2–/–Rad18–/– MEFs when compared with WT MEFs. The dependence on Chk2 (but not Rad18) for suppressing UV-B-induced skin tumorigenesis is consistent with the requirement for Chk2 (but not Rad18) in preventing micronuclei in UV-irradiated cultures (Supplementary Table S1 online). Our results can explain why there is no statistically significant difference in rates of skin carcinogenesis between Chk2–/–and Chk2–/–Rad18–/– mice (Figure 1a).
Figure 4Evaluation of genomic instability of ES cells. (a) Chk2–/– ES cells (left panel) and Chk2–/–Rad18–/– cells (right panel) with micronuclei stained with DAPI (upper panel), double-stranded DNA break marker phospho-H2AX (middle panel), and merged images (lower panel). Arrows indicate micronuclei. Scale bar = 10 μm. (b) Genomic instability was evaluated as the percentage of ES cells with micronuclei by counting the number of nuclei with phospho-H2AX-positive micronuclei. The white bars represent the result in ES cells without UV-C irradiation (upper panel). The gray bars represent the result in ES cells incubated for 14 hours after UV-C irradiation (2 J/m2) (lower panel). *The statistical significance of the differences in micronucleation was determined by comparison with those in WT ES cells (P < 0.05). ES, embryonic stem; WT, wild type.
Polη–/– mice are prone to developing skin tumors after UV irradiation and therefore recapitulate the cancer-propensity phenotypes of individuals with XP-V (
). Remarkably, we show here that Rad18–/– mice were not skin cancer-prone after UV-B irradiation (Figure 1a). These findings may suggest that Rad18 is dispensable for normal Polη function in preventing skin tumorigenesis in vivo. The lack of reports describing causative RAD18 mutations in patients with XP-V thus far is also consistent with the concept that Polη functions in suppressing skin tumorigenesis are Rad18 independent.
Contrary to our expectation, Rad18 deficiency did not impact UV-B-induced skin carcinogenesis. The main role of Rad18 in the DNA damage response to UV irradiation is to monoubiquitinate PCNA. However, three other E3 ligases, HLTF, RNF8, and CRL4Cdt2, are also capable of directly monoubiquitinating PCNA and promoting TLS in response to UV (
). Therefore, it is likely that the other UV-inducible and PCNA-directed E3 ubiquitin ligases are redundant with Rad18 and compensate for Rad18 deficiency by conferring DNA damage tolerance and suppressing skin carcinogenesis in mammals.
There are some hints that Rad18-mediated TLS may play a role in the etiology of some sunlight-induced cancers. For example, a cancer/testes antigen (melanoma antigen-A4 or MAGE-A4) that is highly expressed in melanoma stabilizes Rad18, promoting Polη-mediated TLS (
). It is possible therefore that “pathological translesion synthesis” due to MAGE-A4-dependent Rad18 overactivity confers DNA damage tolerance and mutability and drives some forms of carcinogenesis (
; Supplementary Figure S1) or UV-B-induced skin tumorigenesis (Figure 1) suggests that inhibiting Rad18 could be a useful strategy for sensitizing melanomas to chemotherapy without side effects.
Chk2–/–Rad18–/– mice seemed to develop skin tumors earlier than Chk2–/– mice, although the difference was not statistically significant. Our cell culture experiments do reveal extensive interplay between Rad18 and Chk2 that impacts checkpoint control and genome integrity. Chk2 is a stable protein expressed throughout the cell cycle that is largely inactive in the absence of DNA damage. In response to ionizing radiation–induced DSBs, Chk2 is activated mainly by ATM (
). The kinetics of UV-inducible Chk2 phosphorylation are also correlated with a surge in phospho-H2AX, suggesting that Chk2 phosphorylation is mediated by DSBs resulting from the collapse of stalled replication forks. Our experiments with ES cells harboring individual or combined mutations in Rad18 and Chk2 have defined the nature of the relationship between the Rad18 and Chk2 pathways as pertains to cell cycle control and survival (Figure 5).
Figure 5Working model for Rad18 and Chk2 functions in maintaining genomic stability and suppressing tumorigenesis. In UV-irradiated cells, Rad18 and other proliferating cell nuclear antigen-directed ubiquitin ligases (Crl4CDT2, Rnf8, Hltf) play redundant roles in sustaining Polη-mediated error-free bypass of cyclobutane pyrimidine dimer lesions and confer DNA damage tolerance. When stalled replication forks collapse generating DSB, Chk2 is activated and mediates cell cycle checkpoints or apoptosis thereby suppressing tumorigenesis. In the absence of Chk2, DSBs are repaired via error-prone mechanisms leading to genome instability and tumorigenesis. DSB, double-strand DNA break.
In contrast with the effects of spontaneously arising DNA damage, the frequencies of UV irradiation-induced micronucleation were significantly increased in Chk2–/– ES cells relative to WT and Rad18–/–. However, combined deficiencies of Chk2 and Rad18 did not further increase the frequency of UV-induced micronucleus formation in Chk2–/– cells (Figure 4b, lower panel; Supplementary Table S1). Micronuclei are thought to originate from chromosome fragments caused by faulty repair of DSBs (
). If we suppose that DSBs are causally linked with UV-B-induced skin tumorigenesis, the patterns of UV-induced micronucleation (observed in ES cells or MEFs of different genotypes) fully explain the UV-B-induced skin tumor propensity of Chk2–/– mice, and explain why Chk2–/–Rad18–/– mice are no more skin cancer prone than Chk2–/– mice. Our results suggest that DSBs are causally linked to UV irradiation-induced skin tumorigenesis (Figure 5).
A major finding of this study then is that Chk2 prevents UV-B-induced skin tumorigenesis. Interestingly, individuals heterozygous for the nonfunctional frameshift CHEK2∗1100delC germline alteration have a twofold increased risk of developing malignant melanoma when compared with noncarriers (
), consistent with a role for Chk2 in suppressing UV-inducible tumors. Clinical surveillance of a cohort of 1434 patients with sporadic breast cancer in the Netherlands showed an increased breast cancer risk for CHEK2∗1100delC heterozygotes (odds ratio: 2.7), whereas homozygotes had an even greater risk (odds ratio: 3.4) (
). If melanoma risk is affected in a similar manner, CHEK2∗1100delC homozygote individuals may have a higher risk of susceptibility to sunlight-induced skin cancers including melanoma than CHEK2∗1100delC heterozygote individuals. Taken together, our study defines an important role for Chk2 in mediating genome stability and suppressing UV-induced skin tumorigenesis.
All animal experiments were approved by the Center for Animal Resources and Development Committee, Kumamoto University, Japan. C57BL/6J mice were used as WT mice.
UV-induced skin tumorigenesis
To examine tumor susceptibilities after UV-B irradiation, backs of the mice were shaved and irradiated at a dose of 4.2 kJ/m2 (22 J/m2/s) once every 3 days with a UVM-57 lamp (UVP, LLC, Upland, CA). We irradiated WT (n = 30), Rad18–/– (n = 30), Chk2–/– (n = 30), and Chk2–/–Rad18–/– mice (n = 30) with UV-B for 40, 40, 32, and 32 weeks, respectively.
UV survival assay of cultured cells
ES cells or MEFs were irradiated with UV-C using a germicidal lamp (Toshiba, Tokyo, Japan) or with UV-B using a UVM-57 lamp. The cells were then cultured for 6 days, fixed, and stained. The numbers of survived colonies were counted.
Cell cycle analysis and measurement of apoptosis (sub-G1)
ES cells or MEFs were irradiated with UV-C or UV-B. The cells were cultured with EdU or BrdU to label replicating DNA. Populations in different cell cycle phases and/or the apoptotic cells (sub-G1) were analyzed using flow cytometry or CellInsight (Thermo Fisher Scientific).
Micronuclei assay
The nuclei and micronuclei of ES cells or MEFs were stained with the anti-phospho-H2AX antibody, counterstained with DAPI, and then visualized by immunofluorescence microscopy.
Statistical analysis
Data are presented as means ± standard deviation from three independent experiments. The statistical significance of the differences between experimental groups was analyzed using Student’s t test, unless otherwise indicated. P values (**P < 0.01, *P < 0.05) were used to assess statistical significance.
Conflict of Interest
The authors state no conflict of interest.
Acknowledgments
We thank members of LILA of IMEG and members of CARD of Kumatmoto University for technical assistance. We also thank Dr. Hiroshi Tanaka and Dr. Mitsuyoshi Nakao (IMEG, Kumamoto Univ.) for technical advice. This work was supported by the Japan Society for the Promotion of Science (JSPS) (Grants-in-Aid for Scientific Research [KAKENHI] JP25131715 and JP23510065, JSPS KAKENHI Grant Number JP 15J01706), Agency for Medical Research and Development (AMED) (16ek0109035h0003,17ek0109229h0001), and NIH Exploratory/Developmental Research Grant Program (Parent R21, 1R21ES023895-01).
32036-0&id=gr1.jpg){kind=link}
32036-0&id=gr2.jpg){kind=link}
32036-0&id=gr3.jpg){kind=link}
32036-0&id=gr4.jpg){kind=link}
32036-0&id=gr5.jpg){kind=link}