DNA Damage–Induced Cytotoxicity Is Dissociated from BRCA1's DNA Repair Function but Is Dependent on Its Cytosolic Accumulation

Mutations of the breast cancer susceptibility gene BRCA1 have been implicated in the development of familial breast cancers (1). The exact role of BRCA1 as a tumor suppressor, however, has not been fully defined. BRCA1 functions in a number of cellular processes, including the repair of chromosomal breaks, DNA replication, chromatin remodeling, cell cycle checkpoints, and apoptosis (2). Disruption of any or all of these processes may contribute to the increased risk for carcinogenesis seen in carriers of germ line BRCA1 mutations.

In contrast, somatic BRCA1 mutations are rare in sporadic cancers (3). It has been hypothesized that deregulation of wild-type BRCA1 function could contribute to the pathogenesis of nonfamilial cancers, which represent up to 90% of breast cancers. Consistent with this, reduced or absent BRCA1 expression has been found in high-grade sporadic breast cancer and occurs at the transcriptional level (4). In addition to transcriptional regulation, BRCA1 function is also regulated by other mechanisms such as protein-protein interactions and posttranslational modification (5). More recently, we and others have shown that BRCA1 is a nuclear/cytoplasmic shuttling protein and its functions may be controlled via active shuttling between cellular compartments (5–7).

BRCA1 nuclear shuttling is a complex process. There are two nuclear localization signals within BRCA1, which target it to the nucleus via the importin α/β transport system (8). BRCA1 also contains two nuclear export sequences (NES) at the NH₂ terminus in its ring domain, which transports BRCA1 to the cytoplasm through the chromosome region maintenance 1 (CRM1)/exportin pathway (9, 10). BRCA1 shuttling can also be influenced by its association with other proteins. The BRCA1-associated ring domain protein (BARD1) has been shown to bind to the ring domain and mask the NES of BRCA1, subsequently preventing the nuclear export of BRCA1 through CRM1 (6, 7). The tumor suppressor p53 could also bind BRCA1 and affect BRCA1 cellular localization (5, 11). In contrast to BARD1, however, p53 mediates BRCA1 nuclear export in response to ionizing radiation (IR)–induced DNA damage (5). Consistent with the role of p53 in BRCA1 nuclear export, breast cancer cells with p53 mutations exhibit a predominantly nuclear BRCA1 staining pattern (5).¹

Given these diverse functions of BRCA1 and the poorly defined mechanisms by which its localization regulates these functions, we assessed the repair of DNA double-strand breaks (DSB) and the DNA damage–induced cytotoxic response as a function of BRCA1 location. The data suggest that BRCA1 shuttling may not only control its nuclear function in DNA repair but might also facilitate additional cellular processes involved in the execution of DNA damage–induced cell death.

MCF7 cells were maintained as previously described (12). Isogenic lines derived from HCC1937 cells carrying wild-type BRCA1 expression plasmid (HCC^wt-BRCA1), mutated BRCA1 Chk2 phosphorylation site S988A (HCC^BRCA1S988A), mutated BRCA1 RING domain C64G (HCC^BRCA1C64G), or mutated BRCA1 BRCT domain P1749R (HCC^BRCA1P1749R) were maintained as previously described (12). MCF7 and HCC1937 cells were purchased from American Type Culture Collection. The genetic background, including expression and function of BRCA1, p53, and caspase 3, 8, and 9, as well as the growth characteristics and their response to genotoxic agents, was tested most recently in May 2010 using Western blot analysis, immunohistochemistry, and colony formation assays. All transfections were performed using FuGene6 according to the manufacturer's recommendations (Roche).

Cells were cultured and mounted onto sterile glass slides and subjected to various treatments as indicated. Immunohistochemistry was performed as previously described (5). Primary antibodies were 1:500 rabbit anti-Rad51 antibody (Calbiochem) and 1:150 mouse anti-BRCA1 (Ab-1; EMD Chemicals, Inc.). Secondary antibodies were 1:1,000 anti-mouse Alexa 488–conjugated antibody (Molecular Probes) and 1:1,000 anti-rabbit Alexa 594–conjugated antibody (Molecular Probes). Staining patterns were visualized via fluorescence microscopy (Carl Zeiss). A total of at least 50 cells were counted per field, and a total of 10 fields were assessed. For BRCA1 localization, cells were assessed as having predominantly nuclear staining (N), predominantly cytoplasmic staining (C), or mixed nuclear/cytoplasmic staining (N/C).

Whole cell lysates were prepared using radioimmunoprecipitation lysis buffer [150 mmol/L NaCl, 50 mmol/L Tris (pH 8.0), 5 mmol/L EDTA, 0.5% sodium deoxycholate, 0.1% SDS, and 1.0% NP40] with protease and phosphatase inhibitor cocktails (Sigma) and subjected to SDS-PAGE analysis. BRCA1 antibody (Ab-1) was used at 1:1,000 dilution. Caspase antibodies [cleaved caspase-3 (Asp175; Cell Signaling Technology), cleaved caspase-9 (Asp330; Cell Signaling Technology), and caspase-8 (H-134; Santa Cruz Biotechnology)] were used at 1:1,000 dilution. Hsp70 (1:1,000) or actin (1:1,000; both from Santa Cruz Biotechnology) levels were also analyzed by Western blot as a loading control.

The colony-forming ability following treatment was evaluated in MCF7 and HCC1937 cells as previously described (12). Briefly, cells subjected to the indicated treatment and doses were maintained in culture in plates for 10 to 15 days. Colonies in plates were fixed with 70% ethanol and stained with 1% methylene blue. Colonies consisting of >50 cells were scored using a microscope. Survival fraction was calculated as: (number of colonies / number of cells plated) / (number of colonies for corresponding control / number of control cells plated).

HCC1937 breast cancer cells were transfected with pcDNA3 vector control, wild-type BRCA1, or various mutant BRCA1 constructs using FuGene6 according to the manufacturer's recommendations (Roche). Twenty-four hours posttransfection, cells were subjected to either mock or 10 Gy IR. Forty hours post-IR, cells were harvested, and apoptotic cells were stained using the Annexin V-FITC Apoptosis Detection kit (BD PharMingen). The percentage of apoptotic cells was determined by flow cytometry.

Analysis of cleaved caspase 3, 8, or 9 was also performed in transfected cells as above, and subjected to either mock or 6 Gy IR. Forty-eight hours following IR, whole cell lysates were prepared and subjected to Western blot analysis for cleaved caspase 3, 8, or 9.

Data were analyzed via one-way ANOVA followed by a Bonferroni or Dunnett post-test using GraphPad Prism version 4.02 for Windows (GraphPad Software).

Previously, we and others have shown that in response to DNA damage, BRCA1 shuttles between the nucleus and cytoplasm through the importin and CRM1 pathways, respectively (5, 6). Additionally, it has been shown that the protein BARD1 prevents BRCA1 export to the cytosol through its binding to the NH₂-terminal region of BRCA1. This, in turn, masks the BRCA1 NES and blocks BRCA1 interaction with CRM1/exportin (6).

Given the central role of BRCA1 in promoting the repair of DSBs, we sought to determine whether BRCA1 nuclear/cytosolic shuttling regulates its function in DSB repair. We first analyzed Rad51 foci, an in vivo functional marker of homologous recombination (HR) repair activity, in relation to BRCA1 localization in MCF7 human breast cancer cells. The subcellular localization of BRCA1 was assessed by immunohistochemical staining. Similar to previous studies by us and others (5, 6), cells were categorized into three groups based on BRCA1 staining pattern: strictly nuclear (N), strictly cytoplasmic (C), and mixed nuclear and cytoplasmic staining (N/C). Rad51 nuclear foci were subsequently determined in the cells of each category (Fig. 1A). As shown in Fig. 1B, IR-induced Rad51 foci were found predominantly in MCF7 cells with BRCA1 present in the nucleus (90% of cells with nuclear BRCA1, 70% of cells with mixed nuclear/cytoplasmic BRCA1). In contrast, the percentage of cells with IR-induced Rad51 foci was drastically diminished in cells with BRCA1 not present in the nucleus (10% of cells with cytoplasmic BRCA1). These results suggest that the repair of DSBs is strongly associated with nuclear BRCA1.

To determine whether DSB repair is dependent on BRCA1 in the nucleus, we next examined whether depletion of nuclear BRCA1 could attenuate DSB repair. We thus exogenously expressed the small peptide “tr-BRCA1,” a truncated form (1–301 amino acids) of BRCA1 that contains the NES and BARD1 binding site (13), in MCF7 cells. Translocation of BRCA1 by tr-BRCA1 was subsequently determined by immunohistochemical staining. As shown in Fig. 1C, expression of tr-BRCA1 alone in MCF7 cells effectively decreased nuclear BRCA1 by 2-fold compared with vector control (40% versus 20%). As a reference control, vector-transfected MCF7 cells were irradiated and exhibited a similar 2-fold nuclear export of BRCA1 in response to 5 Gy IR as previously reported (5). IR, however, did not further enhance tr-BRCA1's effects on BRCA1 localization (18% versus 15% nuclear staining). This suggests that tr-BRCA1 is as efficient as IR in shifting BRCA1 to the cytosol.

Having verified that tr-BRCA1 could efficiently drive BRCA1 to the cytosol, we next confirmed its effects on DNA repair. We reasoned that compromised DNA repair by the expression of tr-BRCA1 would result in increased levels of persistent DSBs following IR. In these experiments, we examined IR-induced γ-H2AX foci, a commonly used in situ marker of DNA DSBs, with or without tr-BRCA1. As shown in Fig. 1D, the expression of tr-BRCA1 significantly decreases DSB repair efficiency as evidenced by persistent γ-H2AX foci compared with control cells (1 hour post-IR, 32% versus 25%, P < 0.001; 8 hours post-IR, 25% versus 12%, P < 0.001). These results suggest that BRCA1-mediated DSB repair depends on nuclear BRCA1. Additionally, in support for the role of BRCA1 localization/shuttling in its repair function, we have previously shown that tr-BRCA1 significantly decreased HR efficiency in MCF7 human breast cancer cells (14).

Given that integral cell survival processes such as DNA repair are dependent on nuclear BRCA1 (Fig. 1), we hypothesized that depletion of nuclear BRCA1 could potentially serve as a molecular strategy to render cells more susceptible to genotoxic agents. Support for this hypothesis has been shown by sensitization of tumor cells to IR or cisplatin when DSB repair is deficient (15). Because tr-BRCA1 could modify BRCA1 localization and DSB repair function (Fig. 1C and D; ref. 14), we next examined whether the expression of tr-BRCA1 could sensitize cells to DNA-damaging agents by decreasing the nuclear functions of BRCA1 in DSB repair. The survival fractions, as measured by colony formation assays, were used to determine cytotoxicity. As shown in Fig. 2, increasing cytosolic translocation of BRCA1 by tr-BRCA1 expression in MCF7 breast cancer cells significantly enhanced the cytotoxic response to IR (Fig. 2A; P < 0.001) or cisplatin (Fig. 2B; P < 0.001) compared with vector control cells. These effects were also observed in HCT116 human colon cancer cells (data not shown). To determine whether the cytotoxic effects of tr-BRCA1 are specific to its effects on BRCA1, radiosensitization was assessed in the BRCA1-deficient human breast cancer cell line HCC1937, which is hemizygous for BRCA15382insC mutation. This mutation renders the BRCA1 protein unstable and predominantly cytosolic (16). As shown in Fig. 2C, tr-BRCA1 did not radiosensitize HCC1937 cells at 4 Gy IR, the dose at which tr-BRCA1 already significantly enhanced radiosensitivity in MCF7 breast cancer cells (Fig. 2A). This lack of radiosensitization by tr-BRCA1 in BRCA1-deficient cells suggests a dependence of these effects on BRCA1.

If compromised DSB repair is the predominant underlying mechanism for tr-BRCA1–mediated sensitization of cells to DNA damage, then mutations that abolish BRCA1's repair function would similarly render cells susceptible to DNA damage. We and others have shown that the DNA damage response kinase Chk2 regulates BRCA1's repair function by phosphorylating BRCA1 on serine 988. Mutation of this Chk2 phosphorylation site renders cells deficient in DSB repair (17, 18). Thus, to determine whether cytotoxicity mediated by tr-BRCA1 is conferred by decreasing BRCA1's repair function, we used the previously established isogenic BRCA1-deficient HCC1937 cells that stably express either wild-type BRCA1 (HCC^wt-BRCA1) or BRCA1 mutated at the Chk2 phosphorylation site (HCC^BRCA1-S988A; ref. 18). As previously reported and as shown in Fig. 3A (black data points), restoration of wild-type BRCA1 in HCC1937 cells resulted in a 5-fold decrease in radiosensitivity at 6 Gy IR compared with HCC1937 cells with control pcDNA3 vector (12). This IR resistance has been hypothesized to be due to the enhanced efficiency of cells to repair IR-induced DNA DSBs. Similar effects were seen in response to the DNA-damaging agent cisplatin (75 μmol/L; Fig. 3B, black data points). Surprisingly, HCC^BRCA1-S988A exhibited IR and cisplatin resistance similar to HCC^wt-BRCA1 cells despite being deficient in DSB repair (Fig. 3A and B). Additionally, this differential response to DNA damage–induced cytotoxicity among the various BRCA1 constructs is not due to differences in protein expression (Fig. 3C) or cell cycle effects (data not shown). These results did not support our hypothesis that compromised DSB repair is the predominant underlying mechanism for the enhanced sensitivity of cells to DNA-damaging agents. Instead, our observations imply a potential dissociation of BRCA1's function in DSB repair and cellular cytotoxicity.

It has been shown that BRCA1 is required for DNA damage–induced apoptosis via the caspase 3 pathway (19). Furthermore, cytoplasmic BRCA1 accumulation alone induces apoptosis (7). We have also previously shown that DNA damage–induced BRCA1 nuclear export is dependent on p53 (5). Thus, an alternative hypothesis is that in addition to inhibition of repair function, BRCA1 nuclear export/cytosolic translocation is also required to regulate cell death processes following DNA damage. Thus, we next examined whether the observed resistance to DNA damage in HCC^wt-BRCA1 and HCC^BRCA1-S988A cells was due to dysregulated BRCA1 nuclear/cytosolic shuttling secondary to the endogenous p53 deficiency of HCC cells, given that p53 controls DNA damage–induced BRCA1 nuclear export. To test this notion, we assessed nuclear shuttling of BRCA1 in HCC1937 cells. Interestingly, in contrast to IR-induced BRCA1 nuclear export seen in p53-proficient MCF7 cells (ref. 5; Fig. 1C), the p53-deficient HCC^wt-BRCA1 cells exhibited predominantly nuclear BRCA1 with a diminished magnitude of IR-induced BRCA1 nuclear export (Fig. 3D). Similarly, mutation of BRCA1 at the Chk2 phosphorylation site did not significantly alter BRCA1 localization compared with wild-type BRCA1 (Fig. 3D). These data suggest that, in addition to repair of damaged DNA, the resistance to DNA damage–induced cytotoxicity in HCC^wt-BRCA1 and HCC^BRCA1-S988A cells may be associated with diminished shuttling of BRCA1 from the nucleus to the cytosol.

To further test this alternative hypothesis that BRCA1 nuclear export following DNA damage confers enhanced cytotoxicity, we used HCC1937 cells that express BRCA1 C64G (HCC^BRCA1C64G), which carries a mutation within the NH₂-terminal RING domain, or BRCA1 P1749R (HCC^BRCA1P1749R), which carries a mutation within the COOH-terminal BRCT domain. These mutations have been previously shown to not only affect BRCA1 localization (20) but also render cells deficient in repairing IR-induced DNA DSBs (12). In contrast to the HR-deficient Chk2 kinase mutant above, HCC^BRCA1P1749R cells expressed predominantly cytoplasmic BRCA1 with or without IR, whereas HCC^BRCA1C64G cells exhibited pronounced BRCA1 nuclear export in response to IR (Fig. 4A).

Interestingly, HCC^BRCA1C64G cells, which are proficient at IR-induced BRCA1 nuclear export, and HCC^BRCA1P1749R cells, which exhibit predominantly cytoplasmic BRCA1, possessed an increased sensitivity to IR and cisplatin similar to HCC^Vector cells when compared with HCC^wt-BRCA1 and HCC^BRCA1-S988A cells (Fig. 4B and C). These effects were not due to differences in BRCA1 expression levels (Fig. 4D). Thus, our data support the notion that, in addition to impaired DNA repair, increased cytosolic BRCA1 is a critical determinant of enhanced cytotoxicity from DNA damage.

We next reasoned that if cytosolic accumulation of BRCA1 is the critical determinant in DNA damage–induced cytotoxicity, this effect should not be limited to the type of DNA lesions which specifically require BRCA1-dependent HR-mediated repair. To further test the hypothesis that increased cytosolic BRCA1 is a crucial determinant of enhanced cytotoxicity from DNA damage, we analyzed the sensitivity of cells to UV-induced DNA damage, which is predominantly repaired via the nucleotide excision repair pathway. MCF7 cells with or without the ectopic expression of tr-BRCA1, which we have shown to induce BRCA1 cytosolic localization independent of DNA damage (Fig. 1C), were subjected to UV exposure. The survival fractions, as measured by colony formation assays, were used to determine cytotoxicity. As shown in Fig. 5, increasing cytosolic translocation of BRCA1 by tr-BRCA1 expression in MCF7 breast cancer cells significantly enhanced the cytotoxic response to UV (Fig. 5; P < 0.001) compared with vector control cells. These results again support the hypothesis that cytosolic BRCA1 confers enhanced cytotoxicity to DNA damage.

One potential mechanism by which BRCA1 subcellular localization controls DNA damage–induced cytotoxicity might be the previously reported role of cytoplasmic BRCA1 in the induction of apoptosis (7). To test this possibility, we analyzed the effect of the various BRCA1 mutants on IR-induced apoptosis as a function of their subcellular location. As shown in Fig. 6A, HCC1937 cells that express BRCA1 mutants located predominantly in the cytosol exhibited the greatest IR-induced apoptosis as measured by the positivity of Annexin V staining. Further support for the involvement of apoptotic pathways is shown in Fig. 6B. In particular, cells which express cytosolic localized BRCA1 mutants exhibit a 2-fold increase in cleavage of caspase 3 (Fig. 6C). These results are in agreement with the cytotoxicity shown in Fig. 4.

There are two major apoptotic processes, consisting of the intrinsic and extrinsic pathways (21). The extrinsic pathway is activated via proapoptotic ligand-mediated stimulation of cellular death receptors. This, in turn, results in the cleavage of caspase 8 and the induction of apoptotic response. In contrast, the intrinsic pathway is triggered by stress signals from within the cell, which ultimately results in cleavage of the initiator caspase 9 and subsequent programmed cell death.

To further investigate the mechanism by which cytosolic BRCA1 mediates IR-induced apoptosis, cleavage of caspase 8 and 9 following IR was examined. As shown in Fig. 7A and B, an increased IR-induced cleavage of caspase 9 is observed when there is a predominance of cytosolic BRCA1. Interestingly, no IR-induced cleavage of caspase 8 was observed (data not shown). This is consistent with the role of cytosolic BRCA1 in the intrinsic apoptotic pathway. These results support the notion that cytosolic localization/shuttling of BRCA1 may regulate apoptotic processes through the intrinsic pathway following DNA damage. Further support of this idea is evident in the vector control HCC1937 cells, which harbor the BRCA15382insC mutation that renders the BRCA1 protein predominantly cytosolic.

Taken together, these data suggest that DSB repair deficiency alone might not be sufficient for radiosensitization, and that nuclear export–induced loss of DSB repair in combination with cytoplasmic accumulation of BRCA1 seems to be required to fully launch the DNA damage–induced cytotoxic response.

In this report, we investigated the regulatory role of BRCA1 localization on its DNA repair function as well as on the cytotoxic response to DNA damage. Specifically, we showed that DNA repair was dependent on nuclear BRCA1. Targeted translocation of BRCA1 to the cytoplasm suppressed BRCA1's function in DSB repair and conferred enhanced cytotoxicity to DNA-damaging agents, including IR and cisplatin. Evidence for decreased HR capacity and enhanced sensitivity to DNA-damaging agents has been previously reported (15). Surprisingly, our genetic study using various BRCA1 mutants deficient in either repair or nuclear/cytosolic shuttling revealed a dissociation of DNA damage–induced cytotoxicity from BRCA1's DNA repair function but a dependence on BRCA1 shuttling. These results imply that BRCA1 localization not only controls repair but also regulates additional cell death processes in response to DNA damage. Understanding this facet of the molecular pathways regulated by BRCA1, in particular BRCA1-mediated regulation of cell death, could shed light on one potential mechanism by which BRCA1 serves as a tumor suppressor.

Using various BRCA1 mutants, we found that one of the main determinants of cellular sensitivity to genotoxic therapy might be DNA damage–induced BRCA1 nuclear export and/or cytosolic accumulation. The BRCA1-deficient HCC1937 cells exhibit exquisite sensitivity to DNA-damaging agents. Interestingly, these cells do express a mutant 5382insC BRCA1 that is exclusively cytosolic, which supports the notion that cytoplasmic BRCA1 confers the enhanced cytotoxic response to DNA damage. Stable expression of wild-type BRCA1 in these cells (HCC^wt-BRCA1) confers increased therapeutic resistance that is thought to be due to enhanced repair of therapy-induced DNA damage. However, a similar resistance is seen in isogenic cells expressing a BRCA1 mutation at the Chk2-phosphorylation site (HCC^BRCA1-S988A). Contrary to HCC^wt-BRCA1 cells, HCC^BRCA1-S988A cells are HR deficient, which argues against the idea that DSB repair alone is sufficient and could fully account for the resistance to DNA damage observed in these cells.

A potential explanation may be that the cell cycle checkpoint function of wild-type BRCA1 is responsible for rescuing the sensitivity of HCC1937 cells. However, the BRCA1 S988A mutation is deficient in S phase checkpoint (17), making this explanation less likely. We have also examined another BRCA1 mutant, S1423A/S1524A, which renders BRCA1 resistant to ATM-mediated phosphorylation. Cells stably expressing this mutant have a deficient G₂ checkpoint but are proficient in HR (22). BRCA1 S1423A/S1524A localizes predominantly in the nucleus, does not respond to DNA damage–induced nuclear export, and restores the resistance of HCC1937 cells to IR or cisplatin (data not shown).²

Sensitivity to DNA-damaging agents, however, is restored upon disruption of nuclear BRCA1 localization, as evidenced by BRCA1 NH₂- (C64G) and COOH-terminal (P1749R) mutations. These cells are not only deficient in HR-mediated repair of DSBs but also exhibit a robust DNA damage–induced BRCA1 nuclear export (C64G) or a predominantly cytosolic distribution of BRCA1 (P1749R; ref. 20). Interestingly, the cytosolic translocation/localization of these BRCA1 mutants is not dependent on p53. The C64G mutation in BRCA1 resides in the ring domain of BRCA1, which interacts with BARD1 (6, 7). This interaction prevents BRCA1 nuclear export by masking the BRCA1 NES. It is possible that the C64G mutation disrupts BRCA1-BARD1 interaction following DNA damage to allow for p53-independent BRCA1 shuttling.

In contrast, the P1749R mutation in BRCA1 resides in the BRCT domain of BRCA1, which interacts with other DNA repair proteins such as BACH1 to localize BRCA1 to sites of DNA damage (23, 24). The BRCT domain also interacts with p53 (25, 26), and p53 dysfunction has been shown to induce BRCA1 nuclear accumulation and disrupt BRCA1 nuclear export following DNA damage.³

Nevertheless, these data suggest that the unrepaired DSBs alone resulting from loss of BRCA1's repair function might not be sufficient to fully affect the cytotoxicity and sensitivity of these cells to DNA-damaging agents but rather may rely on the subsequent nuclear export/cytoplasmic accumulation of BRCA1. This is not discordant with previous observations that repair of DSBs is critical for cellular sensitivity to DNA damage. Instead, our novel finding implies that cytosolic translocation of BRCA1 following DNA damage might be the process that links failed repair of DNA damage to the induction and execution of cell death processes.

Specifically, disruption of DNA damage–induced cytosolic translocation of BRCA1 inhibited the apoptotic pathway by 2-fold as measured by Annexin V positivity and cleavage of caspases 3 and 9. The magnitude of regulation of apoptosis does not reach the levels of cytotoxicity as measured by colony formation assays. As multiple pathways other than apoptosis could affect the colony-forming abilities of cells, such as inhibition of cell proliferation, cell cycle arrest, mitotic catastrophe, and autophagy, it is likely that cytosolic translocation of BRCA1 might regulate multiple cytotoxic pathways. These studies are currently ongoing.

Further support of the role of cytosolic BRCA1 in conferring cytotoxicity following DNA damage was found in cells subjected to UV damage (Fig. 5). Cells expressing tr-BRCA1, which drives BRCA1 to the cytosol, possess an enhanced sensitivity to UV. As UV damage is mediated by the nucleotide excision repair pathway, this result emphasizes the importance of cytosolic BRCA1 in the cytotoxic response. However, the transcriptional role of nuclear BRCA1 in the regulation of the repair of UV-mediated damage (27) cannot be ruled out.

Based on our findings, we speculate that nuclear depletion and cytosolic accumulation of BRCA1 might lead to one of the following results:

1. complete abolishment of all (including unknown) aspects of BRCA1's nuclear functions. In addition to DNA repair and checkpoints, for example, sequestration of BRCA1 away from the nucleus could block BRCA1's interaction with other nuclear protein partners that initiate or regulate cell death processes and thus prevent the activation of other DNA damage response pathways;

2. loss of nuclear function in combination with a gain of cytosolic function in mediating cell death, such as BRCA1 localization to the mitochondria (28); or

3. change in the ratio of nuclear/cytoplasmic BRCA1, resulting in a switch from repair and survival activity in the nucleus to cell death processes in the cytosol.

Interestingly, recent reports suggest the importance of cytosolic BRCA1 in facilitating cell death pathways (19, 28). Similarly, BRCA1 nuclear export has been shown to stimulate apoptosis, whereas nuclear sequestration of BRCA1 inhibits apoptosis (7). These findings further substantiate our observation that the enhanced cytotoxic response to DNA-damaging agents is dependent on BRCA1 nuclear shuttling/cytosolic accumulation of BRCA1. Cytoplasmic BRCA1 has also been shown to interact with the centrosome. This additional function contributes to BRCA1's tumor suppressor activities and maintenance of genomic stability (reviewed in ref. 29). It is intriguing to hypothesize that BRCA1 executes its tumor suppressor function not only by its critical role in repair in the nucleus but also signals to the apoptotic machinery in the cytoplasm and thereby eliminates cells when DNA damage cannot be successfully repaired. Ongoing investigations to test these notions may provide further insight regarding the role of BRCA1 nuclear-cytoplasmic shuttling in the control of its function and determination of cell fate (survival versus death).

BRCA1 is essential in maintaining genomic stability and controlling the cellular response to genotoxic stress. Precise regulation of these BRCA1 functions is of obvious importance from an oncologic and cell survival perspective. One emerging target is BRCA1 localization and shuttling, as sequestration of BRCA1 away from the nucleus might switch BRCA1 function from repair in the nucleus to activation of cell death signals in the cytoplasm. These data point to the potential use of BRCA1 shuttling as a novel avenue by which manipulation of BRCA1 localization could control cellular function and sensitivity to therapy. Furthermore, BRCA1 shuttling itself may be a functional biomarker to predict tumor response to therapy.

No potential conflicts of interest were disclosed.

We thank Melissa Stauffer, Ph.D., of Scientific Editing Solutions, for editorial assistance.

Grant Support: R01 CA118158-02 from the NIH (F. Xia) and RR0813 from the Radiological Society of North America (E.S. Yang).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.