Search tips
Search criteria 


Logo of carcinLink to Publisher's site
Carcinogenesis. 2009 August; 30(8): 1314–1319.
Published online 2009 May 29. doi:  10.1093/carcin/bgp136
PMCID: PMC2718077

Human RIF1 encodes an anti-apoptotic factor required for DNA repair


Human Rap1-interacting protein 1 (RIF1) contributes to the ataxia telangiectasia, mutated-mediated DNA damage response against the dexterous effect of DNA lesions and plays a critical role in the S-phase checkpoint. However, the molecular mechanisms by which human RIF1 conquers DNA aberrations remain largely unknown. We here showed that inhibition of RIF1 expression by small interfering RNA led to defective homologous recombination-mediated DNA double-strand break repair and sensitized cancer cells to camptothecin or staurosporine treatment. RIF1 underwent caspase-dependent cleavage upon apoptosis. We further found that RIF1 was highly expressed in human breast tumors, and its expression status was positively correlated with differentiation degrees of invasive ductal carcinoma of the breast. Our results suggest that RIF1 encodes an anti-apoptotic factor required for DNA repair and is a potential target for cancer treatment.


Rap1-interacting protein 1(Rif1p) in the budding yeast interacts in vivo with the major duplex telomeric DNA-binding protein Rap1p and negatively regulates telomere length (1). However, human RIF1 is localized on abnormal telomeres, but not on normal telomeres, forms nuclear foci after DNA damage dependent on both ataxia telangiectasia, mutated (ATM) and 53BP1 and functions in the S-phase checkpoint (2,3). RIF1 is phosphorylated in response to DNA damage, probably by ATM/ATR, and during mitosis (46). In this report, we sought to determine if RIF1 is involved in DNA repair and apoptosis and to examine RIF1 expression status in human breast cancer samples.

Materials and methods

Cell lines, plasmids and antibodies

Human cervical cancer cell line HeLa and human breast cancer cell line MCF-7 were obtained from the American Type Culture Collection (Rockville, MD). These cell lines were cultured in Dulbecco's modified Eagle's medium (Hyclone, Logan, UT) with 10% fetal bovine serum (Hyclone). DRGFP U2OS cells with one copy of the DRGFP gene stably integrated into its genome and the I-SceI expression construct were gifts from Maria Jasin (Memorial Sloan Kettering Cancer Center). All cell lines were grown at 37°C in the presence of 5% CO2. Enhanced green fluorescent protein (EGFP)–RIF1 expression construct was kindly provided by Elizabeth H.Blackburn (University of California, San Francisco).

Mouse monoclonal antibody against β-actin (clone AC15) was purchased from Sigma (St Louis, MO). Antibodies against RIF1 (BL2000–2003) were from the Bethyl Laboratories (Montgomery, TX). Rabbit anti-PARP1 antibody (9542) that recognizes both the full-length protein and the large cleaved fragment was purchased from the Cell Signaling Technology (Danvers, MA). Peroxidase-conjugated secondary antibodies were from JacksonImmuno Research (West Grove, PA).

Western blot analysis and immunoprecipitation

Cell lysates were prepared in 0.5% NP-40 lysis buffer [50 mmol/l Tris (pH 8.0), 250 mmol/l NaCl, 5 mmol/l ethylenediaminetetraacetic acid, 0.5% NP-40] containing protease inhibitor cocktail (Roche, Indianapolis, IN). The protein concentration was determined using DC assay kit (Bio-Rad, Hercules, CA). Equal amounts of proteins were resolved on 4–18% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels, transferred to nitrocellulose membranes (Bio-Rad) overnight, and the membranes were probed with antibodies as indicated in each experiment. For immunoprecipitation assays, cell lysates were incubated with appropriate antibodies (normal rabbit immunoglobulin G (IgG) as control) at 4°C for at least 4 h. The immunocomplexes were pulled down with protein A-sepharose (Amersham, Piscataway, NJ). The immunoprecipitates were washed three times in 0.5% NP-40 lysis buffer, resuspended in sample loading buffer, resolved on sodium dodecyl sulfate–polyacrylamide gel electrophoresis gel and analyzed with antibodies as indicated in each blot.

HR- or NHEJ-mediated DSB repair assay

Homologous recombination (HR)-mediated double-strand break (DSB) assay was performed essentially as described before (7,8). Chromatin immunoprecipitation (ChIP) assay was performed as described before (9). ChIP polymerase chain reaction (PCR) primers are HRChIP.S: 5′-TCTTCTTCAAGGACGACGGCAACT-3′ and HRChIP.R: 5′-TTGTAGTTGTACTCCAGCTTGTGC-3′, annealing temperature was 55°C, the size of PCR products was 145 bp. The non-homologous end-joining (NHEJ) analysis was based on PCR and Southern blotting as described previously (10) with modifications. PCR primers for the NHEJ assay were NHEJ.S: 5′-AGGGCGGGGTTCGGCTTCTGG-3′ and NHEJ.R: 5′-CCTTCGGGCATGGCGGACTTGA-3′, annealing at 68°C. The 723 bp PCR products of mock or I-SceI-transfected cells were purified and digested with I-SceI, followed by BcgI enzymes. After Southern blotting, blots were probed with a digoxigenin-labeled DNA fragment corresponding to the internal green fluorescent protein fragment according to commercial procedure of DIG DNA Labeling Kit (Roche) and developed with chemiluminescent CSPD substrate. The total NHEJ efficiency was quantified by the density of DNA bands resistant to I-SceI and BcgI cleavage using the Quantity One software (Bio-Rad).

Cell proliferation assay

Cell proliferation assays for MCF-7 cells were performed using the CellTiter 96® AQueous One Solution Cell Proliferation Assay system (Promega, Madison, WI) essentially according the manufacturer's instructions. Briefly, HeLa or MCF-7 cells were transfected twice with a control small interfering RNA (siRNA) oligo (CONTROLsi), RIF1-specific siRNA oligo (3), XLF-specific siRNA oligo (11) or PP4C-specific siRNA oligo (PP4CsiM) (12) using oligofectamine (Invitrogen, Carlsbad, CA). The siRNA oligo sequences were as follows: CONTROLsi, CGUACGCGGAAUACUUCGAdTdT and RIF1-2si, AAUGAGACUUACGUGUUAAAAdTdT; RIF1-6si, AAGAGCAUCUCAGGGUUUGCUdTdT; XLFsi, GCAUUACAGUGCCAAGTGAdTdT. PP4Csi is a mixture of four predesigned OnTarget plus siRNA oligonucleotide duplexes: GCACUGAGAUCUUUGACUA, GACAAUCGACCGAAAGCAA, GCACUUAAGGUUCGCUAUC and GGAGCCGGCUACCUAUUUG. All siRNA oligo duplexes were purchased from Dharmacon (Lafayette, CO). Transfectants were trypsinized 24 h after the second transfection and were plated, at least in quadruplicate, into 96-well plates at 8000 per well in 100 μl Dulbecco's modified Eagle's medium containing camptothecin (CPT) at the indicated concentrations with 10% fetal bovine serum. Following a 24 h incubation, 20 μl of One Solution Reagent, which contains a novel tetrazolium compound [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS), inner salt] and an electron-coupling reagent (phenazine ethosulfate), were added to each well. After 1–2 h of further incubation, absorbance was measured at 490 nm and expressed as a percentage of the absorbance reading of untreated cells (relative survival).

Tissue specimens

Human breast tumor sections were obtained from the Department of Pathology at the City of Hope National Medical Center with institutional review board approval. It was not known if these patients had received any treatment before tissue collection. All specimens had been fixed in 10% formalin-buffered solution and then embedded in paraffin. The Summary of specimen information is shown in Table I.

Table I.
Summary of RIF1 immunohistochemistry results and statistical analysis in breast tumors


Formalin-fixed, paraffin-embedded tissues were deparaffinized in xylene and rehydrated in a graded alcohol series followed by blocking in 1:20 normal horse serum. The primary rabbit anti-human polyclonal RIF1 (BL2001) antibodies (1 μg/μl, Bethyl Laboratories, Montgomery, TX) were used irretrievably in a 1:1000 dilution in a humid container overnight and washed three times (5 min each) with phosphate buffered saline tween-20. The slides were then incubated with biotinylated anti-rabbit IgG antibody (Vector Laboratories, Burlingame, CA) at a dilution of 1:200 followed by three washes (5 min each) of phosphate buffered saline tween-20. Then, after 45 min incubation with AB Complex (Vector Elite Kit, 1:200 dilution) and three washes, diaminobenzidene (0.05 g diaminobenzidene and 100 μl 30% H2O2 in 100 ml phosphate-buffered saline) and 1% copper sulfate were applied for 5 and 10 min, respectively. Each slide was counterstained with Mayer's hematoxylin. For all immunohistochemistry studies, normal rabbit IgG was used as a negative control. The evaluation of staining intensity was performed by two independent pathologists.

Evaluation of immunostaining

The positive signal was detected in the nucleus and cytoplasm. The immunoreactivity of the cells was assessed for both the nuclear staining intensity and the percentage of tumor or normal cells stained. No positive signal (−) was observed or <10% cells were stained weakly; (+) >10% cells were stained weakly; (++) >10% cells were stained moderately and (+++) >10% cells were stained strongly.

Statistical analysis

Statistical significance was examined by non-parametric Mann–Whitney test or Pearson's Chi-square test when two different groups were compared. P values <0.05 (two tailed) were considered significant in all analyses. Correlation between RIF1 expression status and tumor differentiation degrees was examined by Spearman's rank correlation coefficient.

Results and discussion

To reveal functions of human RIF1, we generated four peptide polyclonal antibodies against human RIF1 in collaboration with Bethyl Laboratories (BL2000–2003). We identified both BL2001 and BL2003 suitable for immunoblotting and immunoprecipitation (data not shown). HeLa or MCF-7 cells transfected with two independent siRNA duplexes targeting sequences specific to RIF1 (2) were assayed for RIF1 by immunoblotting 2 days later. Individually each siRNA had a significant knockdown effect on RIF1 levels, whereas RIF1-6si had a better knockdown effect than RIF1-2si (Figure 1A). Indirect immunofluorescence using the BL2001 antibody showed that RIF1 was exclusively localized in the nucleus and formed discrete nuclear foci 1 h after treatment with the topoisomerase inhibitor CPT, which colocalized with the DSB marker γ-H2AX (Figure 1B and data not shown). All these data demonstrate that the RIF1 antibodies we developed are able to specifically recognize and bind to RIF1.

Fig. 1.
RIF1 is recruited to DSBs. (A) Inhibition of RIF1 expression by siRNA. Transfection with control siRNA oligo or individual siRNA oligo against RIF1 was performed using Oligofectamine in HeLa or MCF-7 cells. RIF1 protein levels were determined 2 days later ...

To directly confirm that RIF1 is recruited to the DSB sites after DNA damage, we examined if RIF1 localized at the unique DSB generated by the I-SceI endonuclease in vivo (13). The genome of DRGFP U2OS cells has stably integrated one copy of the DRGFP gene. The DRGFP construct carries a tandem repeat of the GFP gene, in which one copy is inactivated by the I-SceI sequence and the other by truncations at the N- and C-termini. Expression of the I-SceI endonuclease leads to a single DSB in each cell. As shown in Figure 1C, 6.15 ± 1.07% of DRGFP U2OS cells transfected with the I-SceI expression construct (384–438 cells per experiment were randomly examined in three independent experiments) showed discrete RIF1 foci, which were colocalized with γ-H2AX; whereas only 1.42 ± 0.55% of DRGFP U2OS cells transfected with the control vector (340–434 cells per experiment were randomly examined in three independent experiments) showed discrete RIF1 foci. They also colocalized at unique DSBs in HeLa cells transiently transfected with the DRGFP construct and the I-SceI construct (data not shown). ChIP analysis of RIF1 in DRGFP U2OS cells showed that RIF1 bound to the region adjacent to the I-SceI cutting sequence only upon induction of the unique DSB site by the I-SceI endonuclease (Figure 1D). These results support that RIF1 is indeed recruited to the DSB sites in responses to DNA damage.

Recruitment of RIF1 to the DSB sites indicates that RIF1 may be involved in DNA damage checkpoint signaling and/or DNA repair. It has been demonstrated that RIF1 is required for the IR-induced S-phase checkpoint in an ATM/53BP1-dependent manner (2). However, it is not clear if RIF1 is involved in DNA repair. We first investigated how RIF1 knockdown affects γ-H2AX levels. MCF-7 cells transfected with control or RIF1 siRNAs were treated with CPT for 1 h and then washed free of the drug, γ-H2AX levels were analyzed thereafter by immunoblotting (Figure 2A). In the absence of RIF1 siRNA, γ-H2AX was not detected in cells untreated with CPT, began to be detected immediately after adding CPT, peaked at the time of CPT removal and returned to background levels by 5 h. However, in the presence of RIF1 siRNA, γ-H2AX levels were sustained at significantly higher levels and were only slightly decreased 7 h after CPT removal. Persistence of high γ-H2AX levels in RIF1 knockdown cells indicates that RIF1 is required for efficient DNA repair.

Fig. 2.
Depletion of RIF1 leads to defects in DNA repair. (A) Depletion of RIF1 leads to sustained γ-H2AX levels after DNA damage. MCF-7 cells were transfected with RIF1-2si, RIF1-6si or CONTROLsi. Transfectants were treated with CPT (10 μM) for ...

CPT induces DNA breaks by blocking DNA replication. DSBs are repaired by HR or NHEJ. DSBs generated during DNA replication are mainly repaired by the HR pathway (14). To directly test whether RIF1-knockdown cells displayed defects in DSB repair, we carried out assays for HR- or NHEJ-mediated repair using the DRGFP/I-SceI system (8,13). DRGFP U2OS cells have one single copy of the DRGFP gene stably integrated into its genome. These cells transiently transfected with the I-SceI expression construct were used for this assay. A functional green fluorescent protein (GFP) gene including the endogenous BcgI site can be reconstituted if the DSB is repaired by HR using another partial GFP gene as a template and can be detected by fluorescence activated cell sorting. Thus, the number of GFP-positive cells is a measure of HR-mediated DSB repair. NHEJ results in the imprecise fusion of the broken ends, causing a loss of the I-SceI site but no restoration of GFP or the BcgI site. Repair by NHEJ can be estimated by PCR amplification of the GFP gene and subsequent restriction digest with BcgI and I-SceI. The fraction of PCR products that are insensitive to either I-SceI or BcgI digestion reflects the NHEJ frequency (8). It has been demonstrated that PP4C and XLF are required for HR- or NHEJ-mediated DSB repair, respectively (7,11). We recapitulated these DNA repair defects in PP4C- or XLF-depleted DRGFP cells (Figure 2B and C). Furthermore, there was significant reduction in HR efficiency in cells with RIF1 siRNA compared with the control (Figure 2B), whereas the NHEJ-mediated repair efficiency was similar between the RIF1-depleted cells and the mock-depleted cells (Figure 2C). This result suggests that RIF1 deficiency mainly impacts HR repair pathway but not the NHEJ repair pathway. How RIF1 is involved in this process warrants further exploration.

To investigate the biological significance of defective HR repair in RIF1-knockdown cells, we examined their sensitivity to CPT. MCF-7 or HeLa cells were transfected with a control siRNA oligo or a RIF1-specific siRNA oligo. When transfected cells were treated with various concentrations of CPT, RIF1-depleted cells became more sensitive to CPT treatment for 48 h (Figure 3A and data not shown). We also examined whether ecotopic expression of RIF1 would increase resistance of cancer cells to CPT treatment. Indeed, when treated with higher concentration of CPT, HeLa cells harboring the GFP–RIF1 expression construct enhanced cell proliferation that was not observed with the GFP control vector (Figure 3B). Furthermore, endogenous levels of the large fragment (89 kDa) of human PARP1 [poly (adenosine diphosphate-ribose) polymerase family, member 1] produced by caspase cleavage increased more significantly in RIF1-depleted HeLa cells than those in mock-depleted HeLa cells upon CPT treatment (Figure 3C). Cleavage of PARP1 serves as a marker of cells undergoing apoptosis (15). These results indicate that RIF1 is probably involved in the CPT-induced apoptotic process.

Fig. 3.
Depletion of RIF1 sensitizes cancer cells to CPT treatment. (A) Depletion of RIF1 sensitizes MCF-7 cells to CPT treatment. MCF-7 cells were transfected with RIF1-2si, RIF1-6si or CONTROLsi. Transfectants were treated with CPT at different concentrations ...

We then sought to determine if RIF1 is a substrate of caspases during CPT-induced apoptosis. When HeLa cells were treated with CPT (10 μM) for 12 h, the levels of full-length RIF1 polypeptide decreased and concomitantly a faster migration form of RIF1 (~160 kDa) was detected using the RIF1 antibody (BL2001) whose epitope was mapped to a region between residues 1–50 of human RIF1 (NP_060621.3) (Figure 3D). Cotreatment with the pan caspase inhibitor Z-VAD-FMK blocked the production of the faster migration form of RIF1 (Figure 3D). This indicates that RIF1 undergoes caspase-dependent cleavage during CPT-induced apoptosis.

We further examined if RIF1 was involved in the general apoptotic process. Staurosporine (STS), a general protein kinase inhibitor, has been characterized as a strong inducer of apoptosis in many different cell types. It is generally believed that caspase activation is required for staurosporine-induced apoptosis (16). We found that RIF1-depleted HeLa cells were more sensitive to staurosporine-induced apoptosis than that in mock-depleted cells (Figure 4A), whereas HeLa cells transiently transfected with the GFP–RIF1 expression construct were less sensitive to staurosporine-induced apoptosis than that in HeLa cells transfected with the GFP control vector (Figure 4B). We also observed that staurosporine treatment (0.5 μM) for 4 h in HeLa cells induced production of the cleaved form of RIF1 and this cleavage was blocked by cotreatment with the pan caspase inhibitor Z-VAD-FMK (Figure 4C). This suggests that RIF1 undergoes caspase-dependent cleavage during CPT- or staurosporine-induced apoptosis. Proteolysis of RIF1 by caspases may facilitate the progression of apoptosis.

Fig. 4.
Depletion of RIF1 sensitizes cancer cells to staurosporine treatment. (A) Depletion of RIF1 enhances staurosporine-induced apoptosis. HeLa cells were transfected with RIF1-2si, RIF1-6si or CONTROLsi. Transfectants were treated with staurosporine (0.5 ...

Recently, it was found that high expression of genes involved in DNA repair pathways is a signature for metastasis in melanoma patients (17,18). We found that the protein phosphatase PP4, which is required for efficient DSB repair through the HR pathway (7), is overexpressed in human breast and lung tumors (12). We therefore sought to determine the expression status of RIF1 in human breast tumors. To this end, we tested whether our antibodies (normal rabbit IgG and blocking peptides were used as controls) were suitable for imunohistochemistry staining. We found that the BL2001 antibody specifically stained on formalin-fixed and paraffin-embedded tissues (data not shown). Sections of breast tissues (175 cancerous lesions and 25 benign lesions) were stained with this antibody. Immunoreactive signals were evenly distributed mainly in the nucleus (Figure 5). A summary of the immunohistochemistry results and statistical analysis is shown in Table I. The non-parametric Mann–Whitney test demonstrated that expression levels of RFI-1 were significantly higher in cancerous lesions than those in benign lesions (Table I). Pearson's Chi-square test for RIF immunostaining scores (Negative and + versus ++ and +++) in invasive ductal carcinoma (IDC)-I and invasive lobular carcinoma versus IDC-II and IDC-III yields an asymptotic significance (two-sided) of 0.000. We further found a significant correlation (Spearman's rank correlation coefficient r2 = 0.426 with an asymptotic standard error of 0.071 and P = 0.000) between the RIF1 staining intensity and differentiation status of the invasive ductal carcinoma of the breast. Our finding indicates a potential link between RIF1 expression levels and breast tumorigenesis.

Fig. 5.
RIF1 is highly expressed in human breast tumors. Formalin-fixed, paraffin-embedded human breast tumor sections were immunostained with rabbit anti-human polyclonal RIF1 (BL2001) antibodies (1 μg/μl) at a 1:1000 dilution. Normal rabbit ...

In this report, we have demonstrated that RIF1 is recruited to the DSB sites and is required for efficient DNA repair by HR in addition to its role in the S-phase checkpoint. Human RIF1 encodes an anti-apoptotic factor involved in both staurosporine-induced apoptosis and DNA-damaging agent-induced apoptosis. RIF1 is highly expressed in breast tumors and its expression levels correlate with differentiation status of IDC of the breast.

High expression of DNA repair factors potentially enhances DNA repair capacity. In return, enhanced DNA repair capacity may support tumor cells to replicate in a fast and error-careless mode. At the same time, high expression of anti-apoptotic factors confers resistance to therapeutic agents (radiation and chemotherapy drugs) dependent on DNA damage-induced apoptosis. Our results indicate that RIF1 is an anti-apoptotic factor involved in DNA repair. Thus, RIF1 could be a promising diagnostic marker for IDC and a potential druggable target for cancer treatment.


Capital Normal University; Natural Science Foundation of China (30570371, 90608014, 30711120570); Program for New Century Excellent Talents in University (NCET-06-0187); Beijing Natural Science Foundation Program and Scientific Research Key Program of Beijing Municipal Commission of Education (KZ200810028014); Project for Academic Human Resources Development in Institutions of Higher Learning Under the Jurisdiction of Beijing Municipality to X.X.; National Institutes of Health (R01CA084469) to G.P.P.


We thank Eric W.McIntush from the Bethyl Laboratories for providing the RIF1 antibodies for testing. We thank other members of the Xu's laboratory for help.

Conflict of Interest Statement: None declared.



ataxia telangiectasia, mutated
chromatin immunoprecipitation
double-strand break
green fluorescent protein
homologous recombination
invasive ductal carcinoma
immunoglobulin G
non-homologous end joining
polymerase chain reaction
Rap1-interacting protein 1
small interfering RNA


1. Hardy CF, et al. A RAP1-interacting protein involved in transcriptional silencing and telomere length regulation. Genes Dev. 1992;6:801–814. [PubMed]
2. Silverman J, et al. Human Rif1, ortholog of a yeast telomeric protein, is regulated by ATM and 53BP1 and functions in the S-phase checkpoint. Genes Dev. 2004;18:2108–2119. [PubMed]
3. Xu L, et al. Human Rif1 protein binds aberrant telomeres and aligns along anaphase midzone microtubules. J. Cell Biol. 2004;167:819–830. [PMC free article] [PubMed]
4. Dephoure N, et al. A quantitative atlas of mitotic phosphorylation. Proc. Natl Acad. Sci. USA. 2008;105:10762–10767. [PubMed]
5. Matsuoka S, et al. ATM and ATR substrate analysis reveals extensive protein networks responsive to DNA damage. Science. 2007;316:1160–1166. [PubMed]
6. Olsen JV, et al. Global, in vivo, and site-specific phosphorylation dynamics in signaling networks. Cell. 2006;127:635–648. [PubMed]
7. Chowdhury D, et al. A PP4-phosphatase complex dephosphorylates gamma-H2AX generated during DNA replication. Mol. Cell. 2008;31:33–46. [PubMed]
8. Weinstock DM, et al. Assaying double-strand break repair pathway choice in mammalian cells using a targeted endonuclease or the RAG recombinase. Methods Enzymol. 2006;409:524–540. [PubMed]
9. Rauch T, et al. 53BP1 is a positive regulator of the BRCA1 promoter. Cell Cycle. 2005;4:1078–1083. [PubMed]
10. Yang YG, et al. Conditional deletion of Nbs1 in murine cells reveals its role in branching repair pathways of DNA double-strand breaks. EMBO J. 2006;25:5527–5538. [PubMed]
11. Ahnesorg P, et al. XLF interacts with the XRCC4-DNA ligase IV complex to promote DNA nonhomologous end-joining. Cell. 2006;124:301–313. [PubMed]
12. Wang B, et al. Protein phosphatase PP4 is overexpressed in human breast and lung tumors. Cell Res. 2008;18:974–977. [PubMed]
13. Pierce AJ, et al. XRCC3 promotes homology-directed repair of DNA damage in mammalian cells. Genes Dev. 1999;13:2633–2638. [PubMed]
14. Ben-Yehoyada M, et al. The DNA damage response during an unperturbed S-phase. DNA Repair (Amst.) 2007;6:914–922. [PubMed]
15. Koh DW, et al. Mediation of cell death by poly(ADP-ribose) polymerase-1. Pharmacol. Res. 2005;52:5–14. [PubMed]
16. Zhang XD, et al. Staurosporine induces apoptosis of melanoma by both caspase-dependent and -independent apoptotic pathways. Mol. Cancer Ther. 2004;3:187–197. [PubMed]
17. Kauffmann A, et al. High expression of DNA repair pathways is associated with metastasis in melanoma patients. Oncogene. 2008;27:565–573. [PubMed]
18. Sarasin A, et al. Overexpression of DNA repair genes is associated with metastasis: a new hypothesis. Mutat. Res. 2008;659:49–55. [PubMed]

Articles from Carcinogenesis are provided here courtesy of Oxford University Press