BRCA2 is a breast cancer susceptibility gene. Germline mutations of BRCA2 account for about 10–30% of familial breast cancer cases. Consistent with its tumor-suppressor activity, BRCA2 plays an important role in DNA repair. To assess the susceptibility of carriers of mutant BRCA2 to tumorigenesis induced by DNA-damaging carcinogens, we generated a Brca2 knockout mouse strain and studied its susceptibility to chemically induced tumorigenesis. Similar to previously reported Brca2 knockout mice, our Brca2−/− embryos die at E8.5–9.5, while the Brca2+/− mice are tumor-free and fertile. Unexpectedly, Brca2+/− mice developed tumors slower than did their wild-type littermates when treated with a potent carcinogen 7,12-dimethylbenz[a]anthracene (DMBA). In vitro experiments showed that Brca2+/− mouse cells and Capan-1 cells, a human pancreatic cancer cell line deficient of BRCA2, were more sensitive to DMBA-induced apoptosis, than were Brca2+/+ mouse cells and a derivative of Capan-1 cells that expressed exogenous wild-type BRCA2, respectively. Our results suggest that enhanced sensitivity of Brca2 mutant cells to DMBA-induced apoptosis at the dose of DMBA we used contributes to the delayed tumorigenesis of Brca2+/− animals. This suggestion may also provide a rational explanation for a previous unexpected finding that cigarette smoking appears to reduce the breast cancer risk of BRCA2 mutation carriers.
Approximately 10–30% of familial breast cancer patients harbor heterozygous germline mutations on the BRCA2 gene. The finding that the wild-type BRCA2 allele is frequently lost in breast cancer from mutant BRCA2 carriers suggests that BRCA2 is a tumor-suppressor gene for breast cancer in these patients (Welcsh and King, 2001; Schwab et al., 2002; Scully and Puget, 2002). BRCA2 has been found to be involved in several fundamental cellular processes such as DNA repair and recombination, and cell cycle checkpoint control (Orelli and Bishop, 2001; Venkitaraman, 2001a, 2001b, 2002). Unexpectedly, heavy smoking appears to reduce the breast cancer risk of mutant BRCA2 carriers (Brunet et al., 1998). This finding has raised an important issue concerning the relationship between BRCA2 mutants and environmental carcinogens with respect to breast tumorigenesis (Brunet et al., 1998).
To study the functions of BRCA2, several Brca2 knockout mouse lines have been generated (Connor et al., 1997; Ludwig et al., 1997; Sharan and Bradley, 1997; Suzuki et al., 1997; Friedman et al., 1998; McAllister et al., 2002). Brca2 appears to have a critical role in embryo development. Brca2−/− animals of three Brca2 knockout lines die early in embryogenesis (Ludwig et al., 1997; Sharan and Bradley, 1997; Suzuki et al., 1997), while only some Brca2−/− animals of another three Brca2 knockout lines survived to birth (Connor et al., 1997; Friedman et al., 1998; McAllister et al., 2002). A correlation between the location of Brca2 truncations and the viability of Brca2−/− embryos has been suggested based on the different viability of Brca2−/− embryos of these Brca2 knockout mouse lines (Friedman et al., 1998; Morimatsu et al., 1998). Consistent with BRCA2 being a tumor-suppressor gene, surviving Brca2−/− mice frequently develop thymic lymphoma and other types of tumor (Connor et al., 1997; Friedman et al., 1998; McAllister et al., 2002). Moreover, mice carrying mammary gland-specific Brca2 knockout develop mammary gland tumors frequently (Ludwig et al., 2001). Different from humans carrying heterozygous BRCA2 mutations, Brca2+/− mice do not have increased risk for spontaneous tumor. However, there has been no published study on the susceptibility of Brca2−/− animals to carcinogen-induced tumorigenesis.
Our goal has been to understand the role of BRCA2 mutations in carcinogen-induced tumorigenesis. Here we report the intriguing findings from our investigation on treating heterozygous Brca2 knockout mice and their wild-type littermates with an environmental carcinogen, DMBA.
Results and discussion
Generation of Brca2 knockout mice
We generated a Brca2 knockout mouse strain using a homologous recombination replacement technique to study the susceptibility of BRCA2 mutant carriers to environmental carcinogens (Figure 1a). Figure 1b shows the typical result of genotyping Brca2+/+ and Brca2+/− mice when probe A was used in a Southern blot analysis, which is consistent with the expected results. Similar to previously reported Brca2+/− mice (Connor et al., 1997; Ludwig et al., 1997; Sharan and Bradley, 1997; Suzuki et al., 1997; Friedman et al., 1998; McAllister et al., 2002), our Brca2+/− mice did not develop spontaneous tumors and have remained tumor-free for more than one year. Crosses between Brca2+/− mice yielded about two-thirds Brca2+/− mice and one-third of Brca2+/+ mice. Examination of embryos showed that Brca2−/− embryos died at E8.5–E9.5. Based on the observation of previously reported Brca2 mutant mouse lines (Ludwig et al., 1997; Sharan and Bradley, 1997; Suzuki et al., 1997), it was suggested that embryos carrying Brca2 mutants that did not encode any BRC repeat could not survive, whereas embryos carrying Brca2 mutants that encoded at least three BRC repeats were partially viable (Connor et al., 1997; Friedman et al., 1998; McAllister et al., 2002). However, the mutant Brca2 allele of our knockout mice encodes a truncated protein containing the N-terminal 1593 amino acids of Brca2, which includes four BRC repeats. The viability of Brca2−/− embryos could also be influenced by their genetic background and whether the mutant Brca2 allele is expressed or not (Connor et al., 1997; Friedman et al., 1998; Bennett et al., 2000; McAllister et al., 2002).
Delayed DMBA-induced tumorigenesis in Brca2+/− mice
To assess the cancer susceptibility of Brca2 mutant carriers exposed to environmental DNA-damaging agents, we treated female virgin Brca2+/− mice and their wild-type littermates with DMBA in the presence of a hormone, medroxyprogesterone acetate (MPA) (Aldaz et al., 1996a, 1996b). Based on the two-hit tumor-suppressor model (Knudson, 1971), the inactivation of one Brca2 allele in Brca2+/− mice may lead to faster tumorigenesis than in the wild-type mice. In a pilot experiment, we unexpectedly found that Brca2+/− mice (n=15) developed tumors significantly slower than did Brca2+/+ mice (n=6) (P=0.0058, log-rank test) (Figure 2a). Consistent with the reported observations (Aldaz et al., 1996a, 1996b), we also found that MPA/DMBA treatment caused predominantly mammary tumors. Brca2+/− mice developed mammary carcinoma (13/15), lymphoma (7/15), and ovarian carcinoma (2/15). The wild-type mice developed mammary carcinoma (5/6), lymphoma (2/6), and ovarian carcinoma (1/6). To confirm the unexpected finding described above, we performed a second experiment with a larger sample size and found that Brca2+/− mice (n=39) indeed developed tumors significantly slower than did wild-type mice (n=22) (P=0.0227) (Figure 2a). The combined data from these two experiments were plotted using the Kaplan and Meier method, and showed a significant delay of tumorigenesis in Brca2+/− mice, as compared with that of wild-type mice (P=0.0007) (Figure 2a, b).
Enhanced apoptosis induced by DMBA in Brca2+/− cells
DMBA is a DNA-damaging agent, and is known to induce both tumorigenesis and apoptosis (Pena et al., 1998). Limited DNA damage may induce tumorigenesis; however, extensive DNA damage could cause cell death. Thus, the unexpected results shown in Figure 2 prompted us to hypothesize that there was a difference in the DMBA-induced apoptosis between Brca2+/− and wild-type cells that contributed to the slower rate of tumorigenesis in Brca2+/− mice. To test this hypothesis, we established cell strains from a mammary tumor, each from two Brca2+/− mice and a wild-type mouse. The Brca2 genotypes of these cell lines were confirmed by PCR (Figure 3a). The growth rates of these cell lines were measured by direct cell count. As shown in Figure 3b, although there is no significant difference in the growth rates among these cell lines on day 2, Brca2+/− cell lines grew significantly faster than Brca2+/+ cell line after day 3 (Figure 3b and data not shown). This observation is consistent with the previous findings that increased BRCA2 expression inhibits cell growth (Wang et al., 2002) and increased cell proliferation was observed in BRCA2 hereditary tumors, as compared with that in sporadic tumors (Levine et al., 2002). Thus, this result indicates that the growth rate cannot simply account for the delayed tumorigenesis of Brca2+/− mice. To test the sensitivity of these cell lines to DMBA-induced apoptosis, we treated these cell lines with or without DMBA (1 μg/ml), followed by apoptosis assays such as TUNEL (Figure 3c) and PARP cleavage (Figure 3d). Our results clearly show both Brca2+/− cell lines are highly sensitive to apoptosis induced by DMBA. In contrast, little or no apoptosis was detected in the DMBA-treated Brca2+/+ cell line. These results suggested that Brca2+/− cells were more sensitive to DMBA-induced apoptosis, presumably caused by the reduced expression of Brca2 in Brca2+/− cells. To further test this hypothesis, we performed a similar experiment using derivatives of a human pancreatic cancer cell line, Capan-1, which expresses only a truncated BRCA2 protein (Goggins et al., 1996; Abbott et al., 1998; Su et al., 1998). These derivatives of Capan-1 either expressed a moderate level of the wild-type BRCA2 protein (236BRCA2) or carried only the empty vector (Capan-1/neo) (Wang et al., 2002). Consistent with our hypothesis, we found that 236BRCA2 were less susceptible to DMBA-induced apoptosis, than were Capan-1/neo (Figure 3e). This result indicates that expression of the wild-type BRCA2 attenuates the DMBA-induced apoptosis in the BRCA2-null Capan-1 cells, and supports the notion that the enhanced apoptosis in mammary epithelium of the DMBA-treated Brca2+/− mice resulting from reduced DNA repair activity contributes to the delayed tumorigenesis of these mice (Figure 4).
Although Brca2 plays a role in DNA double-strand break repair (Zheng et al., 2000; Venkitaraman, 2002), the degree of its importance in repairing DNA double-strand break may depend on cell type. For instance, while Brca2 mutant MEFs were more sensitive to the apoptosis induced by ionizing radiation (e.g., X-rays) than were the wild-type MEFs, the repair of double-strand DNA break during V(D)J recombination in Brca2 mutant lymphoid cells remained intact (Patel et al., 1998). It is interesting to note Brca2+/− MEFs were slightly but consistently more sensitive to apoptosis induced by UV, X-rays, and MMS, than were Brca2+/+ MEFs (Patel et al., 1998). Similar to that observation, our Brca2+/− tumor cell lines are more sensitive to DMBA-induced apoptosis than are Brca2+/+ tumor cells (Figure 3c, d). Together, these observations suggest that the role of Brca2 in DNA double-strand break repair may be cell type-specific, and supports a potential role of Brca2 in removing DNA adducts induced by alkylating carcinogens such as DMBA.
BRCA2 has an essential function in DNA repair through binding to Rad51 (Zheng et al., 2000; Venkitaraman, 2002). Thus, a decreased level of BRCA2 may render the cancer cells to be sensitive to certain chemotherapy. Consistent with that prediction, it has been shown that women with BRCA2-associated hereditary tumors have an improved recurrence-free interval following initial chemotherapy (Boyd et al., 2000). Furthermore, a lower level of BRCA2 appears to be associated with a better response to chemotherapy (Egawa et al., 2001). Here we report that BRCA2 mutant cells exhibited an increased sensitivity to apoptosis when exposed to the DNA damage agent DMBA, and propose a model to explain these observations (Figure 4). We assume that the number of cells with different degrees of DNA damage caused by environmental carcinogens such as DMBA follows a Poisson distribution. The data shown in Figure 3 suggest that Brca2+/− cells could not efficiently repair DNA damages and, thus, would cause the distribution shift to the right in relation to that of Brca2+/+ cells. As shown in Figure 4, the hypothetical degree of DNA damage beyond which tumorigenesis occurs is designated as critical point 1 (C-1). Critical point 2 (C-2) is a hypothetical point beyond which apoptosis takes place due to severe DNA damage. Thus, cells with the degree of DNA damage range between C-1 and C-2 will have a high probability to develop tumors, whereas cells with the degree of DNA damage beyond C-2 will likely undergo apoptosis. Since the distribution patterns of cells with various degrees of DNA damage differ between Brca2+/− and wild-type mice, this model provides a plausible interpretation for the delayed tumorigenesis observed in the DMBA-treated Brca2+/− mice. This model can also be used to explain the intriguing finding that heavy smoking was correlated with a reduced risk of breast cancer in BRCA1 or BRCA2 mutant carriers (Brunet et al., 1998). Like DMBA, cigarette smoking induces DNA damages, such as adduct formation, and that can lead to tumorigenesis (e.g. lung cancer) and apoptosis (Hecht, 2002). Based on this model, heavy smoking predominantly causes apoptosis in the breast epithelial cells of BRCA2 mutant carriers, thus eliminating some of the premalignant breast epithelial cells. One of the consequences would be the apparent reduced risk of breast cancer. Our working model also raises a possibility of using the noncarcinogenic agents that induce apoptosis in breast epithelial cells to prevent breast cancer for BRCA2 mutant carriers. Further systematic investigation will be required to address these interesting and critical issues.
Materials and methods
Isolation of a mouse genomic Brca2 clone
A mouse embryonic stem (ES) cell genomic library in Lambda DASHII (Stratagene, La Jolla, CA, USA) was screened by hybridizing with human BRCA2 exon 11 probes under a low stringent condition. Out of 107 clones screened, five appeared to be positive after five runs of screening. A 14-kb mouse Brca2 genomic DNA fragment was excised from one of these clones by NotI, and then subcloned into pBlueScript vector (Stratagene). Sequencing analysis revealed that this genomic clone spanned from exon 9 to exon 14. A restriction map was constructed by restriction enzyme digestions, and the relevant enzyme sites are shown in Figure 1.
Construction of Brca2 knockout (KO) plasmid
Southern blot analysis identified a 2.0-kb SpeI/KpnI fragment, which contained the 3′ one-third of Brca2 exon 11. We chose this region to be deleted to abolish the Brca2 function. A 1.3-kb KpnI/BamHI fragment 5′ to this region and a 1.9-kb BglII/SpeI fragment 3′ to this region were selected to be homologous recombination regions. These fragments were subcloned into a KO vector, pLG1, which harbors two selection markers, a neomycin phosphotransferase (neo) gene and a thymidine kinase (tk) gene (Figure 1a). Insertion of a BglII linker abolishes an endogenous EcoRV site in the 3′ region.
Generation of Brca2 KO mice
The targeting construct was first linearized by NotI digestion and then electroporated into the AB-1 ES cell line derived from 129SvEv (129S) mice (McMahon and Bradley, 1990). Out of 576 ES clones that survived G418 and 1-[2′-deoxy-2′-fluoro-β-D-arabinofuranosyl]-5-iodouracil (FIAU) double selection, we obtained four Brca2+/− ES clones (0.67%) using Southern blot genotyping method. The Brca2+/− ES cells were expanded and microinjected into C57BL/6J blastocysts. The chimera embryos were then transferred into the uterus of pseudo-pregnant Swiss Webster females. Taking advantage of the coat color difference between 129SvEv mice (agouti) and C57BL/6J (albino), the resulting mice with higher chimerism were backcrossed with C57BL/6J. The germ line transmission of the Brca2 KO genotype was revealed by Southern blot analysis of mice with agouti coat color in the F1 offspring.
Genotyping of Brca2
Southern blot was performed using a standard protocol (Maniatis et al., 1982). DNA extracted from ES clones or mouse tails was digested with EcoRV and subsequently hybridized with either probe A or probe B. The mutant Brca2 allele would be revealed as a 9-kb band when either probe A or probe B was used, whereas the wild-type Brca2 allele would be revealed as a 5- or a 4-kb band when probe A or probe B was used, respectively (Figure 1a). Genotyping of Brca2 by PCR was performed using three primers: P1: 5′-IndexTermCTCTAAGGAGACTGAAATGC-3′, P2: 5′-IndexTermTGTTCCAAGCTTATCACCC -3′, and P3: 5′-IndexTermGAAAGCGAAGGAGCAAAGC-3′. P1 and P2 are specific for the Brca2 allele and P3 is specific for the exogenous neomycin-resistant gene (Figure 1a). The wild-type Brca2 allele was revealed as a ∼280-bp fragment (P1/P2 PCR product), whereas the mutant allele was revealed as a ∼470-bp fragment (P1/P3 PCR product).
Chemically induced tumorigenesis
We used a chemically induced tumorigenesis protocol that predominantly caused mammary tumors (Aldaz et al., 1996a, 1996b). Briefly, 6-week-old female virgin mice received subcutaneous interscapular implants of two MPA pellets (40 mg/pellet) (Hormone Pellet Press, Leawood, KS, USA). The mice were then treated intragastrically with 100 μl of DMBA (Sigma) in vegetable oil at 1 mg/dose at 9, 10, 12, and 13 weeks of age. The tumor incidence was monitored weekly.
Generation of Brca2+/+ and Brca2+/− mammary tumor epithelial cell lines
Mouse mammary tumors were collected under sterile condition and chopped into small pieces using surgical scissors under sterile condition. The samples were washed with PBS followed by trypsin digestion (1.25% Trypsin in PBS) at 37oC for 20–30 min. After trypsin was removed, the tissues were placed in a complete medium (DMEM/F12 with 10% FBS) and then pipetted through the pipette tip multiple times. The tissue/cell mixture was filtered through a six-layer sterile gauze to remove tissue debris. The cell density was determined by using a hemocytometer, and cells were plated on a 100 mm tissue culture dish in 20 ml complete medium at a density of 5 × 106. The fibroblast cells were removed mechanically by physical scraping and repeated washing.
The TUNEL assay was performed as described previously (Gavrieli et al., 1992). Briefly, cells were harvested into a cytospin chamber. TUNEL assay was performed to score the apoptotic cells with or without DMBA (1 μg/ml) treatment under × 40 magnification, from 10 random fields with total more than 300 cells. Each sample was counted twice. The percentage of apoptotic cells per field was calculated. The PARP cleavage assay was performed as described previously (Ding et al., 2002). For measuring sub-G1 apoptotic cell population using flow cytometry, cells treated with or without DMBA (1 μg/ml) were washed with PBS and suspended in 0.5 ml of PBS containing 0.1% (v/v) Triton X-100 for nuclei preparation. The nuclei were re-suspended in PBS containing 0.1% (w/v) RNase and propidium iodide (50 μg/ml), followed by flow cytometry using a FACScan cytometer.
Abbott DW, Freeman ML and Holt JT . (1998). J. Natl. Cancer Inst., 90, 978–985.
Aldaz CM, Liao QY, LaBate M and Johnston DA . (1996a). Carcinogenesis, 17, 2069–2072.
Aldaz CM, Liao QY, Paladugu A, Rehm S and Wang H . (1996b). Mol. Carcinogen., 17, 126–133.
Bennett LM, McAllister KA, Blackshear PE, Malphurs J, Goulding G, Collins NK, Ward T, Bunch DO, Eddy EM, Davis BJ and Wiseman RW . (2000). Mol. Carcinogen., 28, 174–183.
Boyd J, Sonoda Y, Federici MG, Bogomolniy F, Rhei E, Maresco DL, Saigo PE, Almadrones LA, Barakat RR, Brown CL, Chi DS, Curtin JP, Poynor EA and Hoskins WJ . (2000). JAMA, 283, 2260–2265.
Brunet JS, Ghadirian P, Rebbeck TR, Lerman C, Garber JE, Tonin PN, Abrahamson J, Foulkes WD, Daly M, Wagner-Costalas J, Godwin A, Olopade OI, Moslehi R, Liede A, Futreal PA, Weber BL, Lenoir GM, Lynch HT and Narod SA . (1998). J. Natl. Cancer Inst., 90, 761–766.
Connor F, Bertwistle D, Mee PJ, Ross GM, Swift S, Grigorieva E, Tybulewicz VL and Ashworth A . (1997). Nat. Genet., 17, 423–430.
Ding Y, Wen Y, Spohn B, Wang L, Xia W, Kwong KY, Shao R, Li Z, Hortobagyi GN, Hung MC and Yan DH . (2002). Clin. Cancer Res., 8, 3290–3297.
Egawa C, Miyoshi Y, Takamura Y, Taguchi T, Tamaki Y and Noguchi S . (2001). Int. J. Cancer, 95, 255–259.
Friedman LS, Thistlethwaite FC, Patel KJ, Yu VP, Lee H, Venkitaraman AR, Abel KJ, Carlton MB, Hunter SM, Colledge WH, Evans MJ and Ponder BA . (1998). Cancer Res., 58, 1338–1343.
Gavrieli Y, Sherman Y and Ben-Sasson SA . (1992). J. Cell Biol., 119, 493–501.
Goggins M, Schutte M, Lu J, Moskaluk CA, Weinstein CL, Petersen GM, Yeo CJ, Jackson CE, Lynch HT, Hruban RH and Kern SE . (1996). Cancer Res., 56, 5360–5364.
Hecht SS . (2002). Environ. Mol. Mutagen., 39, 119–126.
Kaplan EL and Meier P . (1958). J. Am. Stat. Assoc., 53, 457–481.
Knudson Jr AG . (1971). Proc. Natl. Acad. Sci. USA, 68, 820–823.
Levine DA, Federici MG, Reuter VE and Boyd J . (2002). Gynecol. Oncol., 85, 431–434.
Ludwig T, Chapman DL, Papaioannou VE and Efstratiadis A . (1997). Genes Dev., 11, 1226–1241.
Ludwig T, Fisher P, Murty V and Efstratiadis A . (2001). Oncogene, 20, 3937–3948.
Maniatis T, Fritsch EF and Sambrook J . (1982). Molecular Cloning – A Laboratory Manual. Harbor Laboratory: Cold Spring.
McAllister KA, Bennett LM, Houle CD, Ward T, Malphurs J, Collins NK, Cachafeiro C, Haseman J, Goulding EH, Bunch D, Eddy EM, Davis BJ and Wiseman RW . (2002). Cancer Res., 62, 990–994.
McMahon AP and Bradley A . (1990). Cell, 62, 1073–1085.
Morimatsu M, Donoho G and Hasty P . (1998). Cancer Res., 58, 3441–3447.
Orelli BJ and Bishop DK . (2001). Breast Cancer Res., 3, 294–298.
Patel KJ, Vu VP, Lee H, Corcoran A, Thistlethwaite FC, Evans MJ, Colledge WH, Friedman LS, Ponder BA and Venkitaraman AR . (1998). Mol. Cell, 1, 347–357.
Pena JC, Rudin CM and Thompson CB . (1998). Cancer Res., 58, 2111–2116.
Schwab M, Claas A and Savelyeva L . (2002). Cancer Lett., 175, 1–8.
Scully R and Puget N . (2002). Biochimie, 84, 95–102.
Sharan SK and Bradley A . (1997). Genomics, 40, 234–241.
Su LK, Wang SC, Qi Y, Luo W, Hung MC and Lin SH . (1998). Oncogene, 17, 2377–2381.
Suzuki A, de la Pompa JL, Hakem R, Elia A, Yoshida R, Mo R, Nishina H, Chuang T, Wakeham A, Itie A, Koo W, Billia P, Ho A, Fukumoto M, Hui CC and Mak TW . (1997). Genes Dev., 11, 1242–1252.
Venkitaraman AR . (2001a). Curr. Opin. Cell Biol., 13, 338–343.
Venkitaraman AR . (2001b). J. Cell Sci., 114, 3591–3598.
Venkitaraman AR . (2002). Cell, 108, 171–182.
Wang SC, Shao R, Pao AY, Zhang S, Hung MC and Su LK . (2002). Cancer Res., 62, 1311–1314.
Welcsh PL and King MC . (2001). Hum. Mol. Genet., 10, 705–713.
Zheng L, Li S, Boyer TG and Lee WH . (2000). Oncogene, 19, 6159–6175.
We thank Dr Shih-Jen Hwang for the assistance in the statistical analysis and Dr Chun-Ming Chen for analysing the development of mouse embryos. This work was supported in part by Faculty Achievement Award (MCH), the MD Anderson Breast Cancer Research Program, and grants RO1CA58880 (MCH), RO1CA 77858 (MCH), and Cancer Center Core Grant CA16672 from the NIH.
About this article
Molecules and Cells (2015)
Chromosome instability and carcinogenesis: Insights from murine models of human pancreatic cancer associated with BRCA2 inactivation
Molecular Oncology (2014)
Deletion of cyclooxygenase 2 in mouse mammary epithelial cells delays breast cancer onset through augmentation of type 1 immune responses in tumors
Experimental and Toxicologic Pathology (2005)