Functional Characterization of a Novel BRCA1-Null Ovarian Cancer Cell Line in Response to Ionizing Radiation

The breast and ovarian cancer susceptibility gene BRCA1 plays a major role in the DNA damage response pathway. The lack of well-characterized human BRCA1-null cell lines has limited the investigation of BRCA1 function, particularly with regard to its role in ovarian cancer. We propagated a novel BRCA1-null human ovarian cancer cell line UWB1.289 from a tumor of papillary serous histology, the most common form of ovarian carcinoma. UWB1.289 carries a germline BRCA1 mutation within exon 11 and has a deletion of the wild-type allele. UWB1.289 is estrogen and progesterone receptor negative and has an acquired somatic mutation in p53, similar to the commonly used BRCA1-null breast cancer cell line HCC1937. We used ionizing radiation to induce DNA damage in both UWB1.289 and in a stable UWB1.289 line in which wild-type BRCA1 was restored. We examined several responses to DNA damage in these cell lines, including sensitivity to radiation, cell cycle checkpoint function, and changes in gene expression using microarray analysis. We observed that UWB1.289 is sensitive to ionizing radiation and lacks cell cycle checkpoint functions that are a normal part of the DNA damage response. Restoration of wild-type BRCA1 function in these cells partially restores DNA damage responses. Expression array analysis not only supports this partial functional correction but also reveals interesting new information regarding BRCA1-positive regulation of the expression of claudin 6 and other metastasis-associated genes and negative regulation of multiple IFN-inducible genes. (Mol Cancer Res 2007;5(1):35–45)

Ovarian carcinoma is the fourth leading cause of cancer death in American women and has the highest mortality of the gynecologic cancers. Approximately 10% to 15% of ovarian carcinomas occur secondary to an inherited mutation in BRCA1 or BRCA2 (1, 2) Inherited mutations in BRCA1 confer lifetime risks of breast cancer of 60% to 80% and of ovarian cancer of >25% to 50% (3, 4). A better understanding of the molecular events of ovarian transformation would facilitate improved surveillance, prevention, and therapy. Although recent advances have been made in developing mouse models for sporadic ovarian carcinoma (5-8), in vitro models for hereditary ovarian carcinoma are limited.

BRCA1 encodes a 1,863-amino-acid protein that associates with a multitude of proteins involved in cell cycle checkpoint regulation, DNA repair, transcription, chromatin remodeling, and ubiquitination (9-13), suggesting that BRCA1 plays a critical role in the cellular response to DNA damage. However, the exact mechanism by which BRCA1 functions in the DNA damage response is not completely understood, partly due to the paucity of human BRCA1-null cell lines available for study. Cellular models of hereditary ovarian cancer are especially lacking, with one ovarian endometrioid BRCA1-null cell line available for study (14), and two ovarian, papillary serous BRCA1-null lines recently developed from xenografts but not yet functionally characterized (15).

Ionizing radiation (IR) induces breaks in double-stranded DNA, the most severe DNA lesions that compromise genomic integrity if left unrepaired. Breaks in double-stranded DNA induce a robust DNA damage response that includes delays in cell cycle progression (induction of cell cycle checkpoints) and activation of DNA repair mechanisms (16, 17). Cell cycle checkpoints and DNA repair mechanisms are coordinated by multiple, complex signal transduction pathways and are highly dependent on the protein kinase function of the ataxia telangiectasia mutated (ATM) protein (18, 19). In response to DNA damage, ATM phosphorylates multiple proteins involved in cell cycle checkpoint regulation and DNA repair, including BRCA1 (20).

Our goal was to develop and characterize a unique BRCA1-null cell line to be used specifically to study BRCA1 function in ovarian cancer. In response to DNA damage induced by IR, our cell line (UWB1.289) lacks DNA repair and cell cycle checkpoint functions. We have also developed UWB1.289+BRCA1, a stable cell line in which wild-type BRCA1 function has been restored, as shown by partial rescue of the response to DNA damage. In addition, comparative expression array analysis of the two cell lines revealed BRCA1 involvement in the positive regulation of cell adhesion, invasion, and metastasis-associated genes, including members of the claudin family of genes implicated in cancer invasion and metastasis (21, 22). Array data also revealed a strong association between BRCA1 expression and the down-regulation of multiple IFN-inducible genes. Importantly, the direct propagation of UWB1.289 from an ovarian tumor of papillary serous histology will facilitate further critical analyses of BRCA1 function in ovarian cancer.

A germline BRCA1 mutation 2594delC was identified by the protein truncation test and confirmed by DNA sequencing of lymphocyte DNA from a woman with ovarian cancer. The BRCA1 mutation 2594delC leads to a stop at codon 845 of BRCA1. The 2594delC BRCA1 mutation has been previously reported in women with familial breast and ovarian cancer.⁶

The proband developed breast cancer at age 42, ovarian cancer at age 54, and died at age 56. The proband's mother developed premenopausal breast and then ovarian cancer at age 32 and 40, respectively, and died at age 42. The patient had received no previous radiation. Chemotherapy after optimal cytoreduction consisted of taxol and carboplatinum for six cycles with a complete response, followed by three cycles of taxol on a consolidation trial. Three months after completing chemotherapy, she presented with a symptomatic intra-abdominal recurrence. She received three cycles of topotecan with progressive disease then received gemcitabine and Adriamycin chemotherapy with a partial response. Tumor was collected from a surgery done to bypass a persistent enterovaginal fistula. The interval between chemotherapy and surgery was 1 month. The patient's postoperative course was complicated and consistent with progressive disease. After 6 weeks, she elected hospital discharge with hospice care. She expired 1 week later.

The recurrent ovarian tumor (Fig. 1A) and the cell line derived from the tumor carry the germline BRCA1 mutation 2594delC (Fig. 1B). Protein truncation tests further showed the lack of a wild-type BRCA1 coding sequence in the cell line (Fig. 1C). Genotyping of intragenic markers in tumor and cell line DNA confirmed loss of the wild-type BRCA1 allele (data not shown). A somatic p53 frameshift mutation in exon 6 (625delAG; Fig. 1D) and loss of the wild-type p53 allele were detected in tumor and cell line DNA, but not in germline DNA. Sequencing of the entire coding region of phosphatase and tensin homologue deleted on chromosome 10 (PTEN) revealed no germline or somatic mutations (data not shown).

Two PCR primer sets were used to amplify BRCA1 message from UWB1.289 cells: primer set 1 (spanning exons 2-7) and primer set 2 (exons 9-11). One-step, real-time reverse transcription-PCR analysis of total RNA revealed expression of mutant BRCA1 transcript in UWB1.289 cells at about half the level of wild-type BRCA1 message when compared with control MCF7 cells (Fig. 1E).

UWB1.289 cells revealed a complex near-tetraploid karyotype that included two copies of chromosome 17 containing the mutant BRCA1 allele and consistent abnormalities of all chromosomes except chromosome 22 (Fig. 1F). Unstable alterations included dicentric chromosomes, triradials, quadriradials, and double minutes. UWB1.289 cells stained positive for proteins characteristic of the immunophenotype of ovarian carcinoma (cytokeratin 7, calretinin, and Wilms' tumor protein) but negative for the gastrointestinal epithelium marker cytokeratin 20 (data not shown). UWB1.289 cells were estrogen receptor and progesterone receptor negative and did not overexpress HER-2/neu as determined by immunohistochemistry (data not shown). Paraffin sections of the patient's primary ovarian tumor as well as the recurrent tumor from which the cell line was derived showed similar staining compared with the UWB1.289 cell line for all proteins investigated (data not shown).

A pcDNA3 plasmid carrying wild-type BRCA1 and a hemagglutinin tag (23) was stably transfected into UWB1.289 cells. Nuclear extracts from UWB1.289 cells revealed no wild-type BRCA1 protein expression, whereas UWB1.289+BRCA1 cells showed a high level of wild-type BRCA1 protein expression, comparable with that expressed by control HBL100 mammary epithelial cells (Fig. 2A). Mutant BRCA1 protein expression (predicted to migrate to ∼ 93 kDa) in the UWB1.289 cell line could not be discerned from that of alternatively spliced isoforms of BRCA1 (data not shown; refs. 24-27). We confirmed that the wild-type BRCA1 protein observed in UWB1.289+BRCA1 cells was expressed from the stably integrated pcDNA3 plasmid by immunoblotting with an anti-hemagglutinin antibody (Fig. 2B). Growth rates of the UWB1.289 (doubling time = 15.3 h) and UWB1.289+BRCA1 (doubling time = 20.0 h) cell lines were not significantly different.

Because BRCA1 is phosphorylated in response to DNA damage (20, 28, 29), we wished to show that the wild-type BRCA1 protein expressed in UWB1.289+BRCA1 cells was potentially functional in the DNA damage response. We irradiated the mutant and reconstituted cell lines 5 h before harvesting nuclear protein extracts. Western blot analysis revealed a slight mobility shift in the migration of wild-type BRCA1 protein in response to irradiation in UWB1.289+BRCA1 cells, indicating at least partial phosphorylation of wild-type BRCA1 in response to IR (Fig. 2C).

IR induces double-stranded DNA breaks in cellular DNA, and deficiencies in double-stranded DNA break repair are apparent in cells that are sensitive to IR. Impaired colony formation is characteristic of cells following IR exposure. We examined colony formation capacity in the UWB1.289 and UWB1.289+BRCA1 cell lines compared with control MDA MB-231 (BRCA1 wild-type and p53 mutant) and HCC1937 (BRCA1-null and p53 mutant) cancer cell lines. The UWB1.289 cell line showed increased radiation sensitivity (decreased clonogenic survival) with increasing doses of IR (Fig. 3), consistent with the response observed in the HCC1937 cell line. Loss of p53 in tumor cell lines does not induce radiation sensitivity, illustrated by the response of MDA MB-231 cells to IR. UWB1.289+BRCA1 cells showed a partial correction of radiation sensitivity, evidenced by increased clonogenic survival at higher IR doses. This response was similar to, but relatively greater than, the gain of function observed in HCC1937+BRCA1 cells.

To repair damaged DNA, cells use DNA damage checkpoints to slow down or halt DNA replication (S phase) and progression to mitosis (M phase; ref. 30). Although p53 is important in the G₁-S phase checkpoint (31), BRCA1 has been implicated in both the S and G₂-M phase DNA damage checkpoints (32, 33). To determine the effect of BRCA1 on cell cycle checkpoints in UWB1.289 cells, we examined cell cycle distribution following DNA damage induced by IR in UWB1.289 and UWB1.289+BRCA1 cells. Fifteen and 24 h after IR, UWB1.289 cells show a decrease in the percentage of cells in G₁, consistent with other cells that have acquired mutations in p53 (Fig. 4A). UWB1.289 cells also show a lack of S-phase arrest, suggesting a loss of capacity to slow DNA replication in the presence of IR-induced DNA damage. This cell cycle distribution phenotype is consistent with that observed in HCC1937 cells.

Restoration of BRCA1 function in the UWB1 cell line results in an increase in the percentage of cells in S phase 15 h after IR, suggesting that wild-type BRCA1 functions to arrest cells in S-phase arrest and prevent them from entering G₂-M phase in the presence of DNA damage. We further evaluated the role of BRCA1 in S-phase arrest by examining radioresistant DNA synthesis in UWB1.289 and UWB1.289+BRCA1 cells. Radioresistant DNA synthesis occurs when DNA synthesis continues in the presence of DNA damage, indicating a compromised S-phase checkpoint (34). Fifteen hours after IR, UWB1.289+BRCA1 cells showed a partial correction of the S-phase checkpoint defect noted in UWB1.289 cells, as illustrated by a significant decrease in DNA synthesis (Fig. 4B). This corrective effect was more pronounced in UWB1.289+BRCA1 cells than in HCC1937+BRCA1 cells.

The G₂-M DNA damage checkpoint arrests cells in G₂, preventing cell division and transmission of damaged DNA to daughter cells. Because the amount of DNA in both G₂ and M phase cells is 4n, the cell cycle distribution assay (which shows accumulation of UWB1.289 cells in G₂-M phase by measuring cellular DNA content; see Fig. 4A) cannot distinguish between cells arrested in G₂ and those allowed entry into M phase. We therefore used a mitotic cell marker, phosphorylated histone H3, in a checkpoint assay to delineate G₂ versus M phase cells (32, 35). In comparison with ATM-positive control cells, which have a functional G₂-M checkpoint (32, 36), UWB1.289 cells seem to have lost this critical DNA damage checkpoint. This is illustrated by a high percentage of mitotic cells, versus cells arrested in G₂, that accumulate after IR-induced DNA damage in UWB1.289 cells (Fig. 5). Consistent with previous reports that ascribe a role for BRCA1 function in the G₂-M phase DNA damage checkpoint, we observed a significant reduction in the percentage of mitotic cells after exposure to IR in UWB1.289+BRCA1 cells.

We next sought to define how gene expression is altered after IR exposure when wild-type BRCA1 is restored to a BRCA1-null ovarian cancer cell line. We did gene expression array experiments to examine the effect of wild-type BRCA1 expression in UWB1.289 cells following IR-induced DNA damage. Table 1 lists genes both up- and down-regulated at least 3-fold in UWB1.289+BRCA1 cells compared with UWB1.289 cells, 5 h after IR-induced DNA damage. Consistent with our protein expression data, wild-type BRCA1 expression was almost 9-fold higher in UWB1.289+BRCA1 cells compared with UWB1.289 cells (P < 0.005). When a P cutoff level of <0.0005 was used in the analysis, 28 genes were found to be up-regulated, whereas 23 genes were observed to be down-regulated at least 3-fold in UWB1.289+BRCA1 cells compared with UWB1.289 cells after IR-induced DNA damage. If P < 0.001 was used with a threshold fold change of 2.0, then 95 genes were up-regulated, and 89 genes were down-regulated (data not shown). Ontology analysis (fold change threshold = 2.0, P < 0.005) showed that multiple cell adhesion genes were up-regulated (z score = 2.8), whereas genes involved in the immune response were down-regulated (z score = 5.9; data not shown).

Because we observed partial restoration of DNA repair and cell cycle checkpoint function in UWB1.289+BRCA1 cells after IR, and because BRCA1 associates directly with multiple proteins involved in DNA repair and cell cycle checkpoint regulation (9, 10), we hypothesized that some of these genes would be up-regulated in UWB1.289+BRCA1 cells compared with UWB1.289 cells after IR. Of the BRCA1-associated genome surveillance protein complex proteins described to date (37), MRE11 was up-regulated 1.3-fold (P = 0.001), RAD50 was up-regulated 1.6-fold (P = 0.01), and MSH6 was up-regulated 1.4-fold (P = 0.04). Expression of ATM, ATR, NBS1, MLH1, and MSH2 was not significantly altered. The double-strand break repair gene RAD51 (1.6-fold, P = 0.004) and the Bloom syndrome gene BLM (1.3-fold, P = 0.04763) were both down-regulated in UWB1.289+BRCA1 versus UWB1.289 cells in response to IR.

We also observed significant down-regulation of multiple IFN-inducible (IFI) genes in UWB1.289+BRCA1 cells after IR (see Table 1). Of the 23 down-regulated genes listed in Table 1, the four genes that showed the greatest fold-change in expression were IFI16 (17×), IFI44 (10×), IFI27 (8×), and IFITM1 (6×). In addition, MX1 (IFN-inducible protein p78) was down-regulated 4-fold in UWB1.289+BRCA1 cells following IR. This effect was not due to irradiation, as microarray analyses of nonirradiated UWB1.289 and UWB1.289+BRCA1 cells also revealed down-regulation of IFN-responsive genes in UWB1.289+BRCA1 cells (Table 2). Also listed in Table 2 are expression data for a selected set of genes identified as significant in our analysis in both irradiated and untreated cells.

We propagated a new human, ovarian BRCA1-null cell line (UWB1.289) from a recurrent tumor of a patient with ovarian cancer. The patient's clinical course was consistent with primary platinum refractory disease according to standard definitions of platinum sensitivity in patients with ovarian cancer, and her overall survival was only 24 months. Women with BRCA1-associated ovarian cancers are generally thought to have a better survival than those with sporadic ovarian cancer, and BRCA1-deficient tumors generally display increased sensitivity to platinum chemotherapy in vitro (38-43). It has been difficult to generate cell lines from BRCA1-associated ovarian cancers. It is interesting to speculate that the molecular features that contributed to this patient's unusually aggressive clinical course may also have facilitated propagation of the tumor in vitro.

We characterized the UWB1.289 cell line's molecular features and also studied its function relative to the DNA damage response. Consistent with previous reports for the BRCA1-null breast cancer cell line HCC1937, the UWB1.289 cell line is deficient in DNA repair and cell cycle checkpoint function. Stable transfection of wild-type BRCA1 into the UWB1.289 cell line (UWB1.289+BRCA1) showed partial correction of DNA repair defects and cell cycle checkpoint function in response to IR-induced DNA damage. In addition, results of expression array experiments reveal a novel function for BRCA1 in the positive regulation of selected set of cell adhesion, invasion, and metastasis-related genes and the negative regulation of multiple IFN-inducible genes.

The SNU251 cell line is the only ovarian cancer BRCA1-null cell line that has been examined relative to BRCA1 function in the DNA damage response (14). SNU251 is of endometrioid histology and thus represents a less common histologic type (∼15%) of ovarian cancer (44). SNU251 carries a heterozygous missense mutation in MLH1, and it is unknown if the BRCA1 mutation in SNU-251 originated in the patient's germline or as a somatic mutation in the tumor or cell line (14). SNU251 has a nonsense mutation (5564G>A) in exon 23 of BRCA1, resulting in mutant BRCA1 protein that is truncated at the COOH terminus by only 48 amino acids yet is expressed at a high level in the mutant cell line (14).

Very recently, two ovarian cancer BRCA1-null cell lines (PD-OVCA1 and PD-OVCA2) were reported that, similar to UWB1.289, represent papillary serous cancer, the most common histologic subtype of ovarian cancer (15). These two lines were developed from tumor xenografts and were characterized genetically and histologically but not functionally. Interestingly, PD-OVCA1 revealed multiple mutant BRCA1 alleles, similar to our finding that chromosome 17 is duplicated in the UWB1.289 cell line. It is unclear if the amplification of the chromosome containing the mutant BRCA1 allele contributes to expedited ovarian tumorigenesis, as others have observed this genetic phenomena in breast tumors (45). Thus, studies designed using both the PD-OVCA1 and UWB1.289 cell lines have the potential to reveal a role for BRCA1 in the most common type of ovarian cancer.

The HCC1937 cell line is the only BRCA1-null cell line in which the DNA damage response to IR has been extensively studied (46). This cell line carries a BRCA1 mutation in exon 20 (5382insC), resulting in the loss of 34 amino acids from the COOH terminus. Although the BRCA1 mutation in HCC1937 cells is in the COOH terminus (versus exon 11 in UWB1.289 cells), many molecular characteristics of the two cell lines (estrogen receptor and progesterone receptor status, p53 mutation, and ploidy) are similar, as well as the impaired response to IR-induced DNA damage, as revealed in this study. Yet, homozygous deletion of PTEN in HCC1937 cells (46) could complicate evaluation of the role of BRCA1 in the DNA damage response in this cell line. UWB1.289 has no discernible defect in PTEN, providing a complementary cell line for analyses of BRCA1 function. Initial studies concluded that HCC1937 cells were sensitive to IR and deficient in double-strand break repair (20, 23, 47) but retained normal S-phase and G₂-M cell cycle checkpoints (23). In contrast, subsequent studies of HCC1937 cells showed defective G₂-M and S-phase cell cycle checkpoints in response to IR (32, 48). Additional conflicting results from HCC1937 cells were reported recently with respect to G₂-M checkpoint function (49, 50) and nonhomologous end-joining of double-strand break repair (29, 51-53), suggesting that additional experimental systems are necessary to thoroughly delineate the function of BRCA1 in the DNA damage response.

In expression array experiments that compared gene expression following exposure to IR between UWB1.289+BRCA1 and UWB1.289 cells, we found that although DNA repair and cell cycle checkpoint genes were generally up-regulated in UWB1.289+BRCA1 cells compared with UWB1.289 cells, the effect was modest. This finding is consistent with the results of our functional assays that showed only a partial (but in most experiments, significant) correction of deficiencies in the DNA damage response in UWB1.289+BRCA1 cells.

In our expression array experiments, more pronounced gene expression alterations were found in extracellular matrix and IFN-regulated genes. We observed significant up-regulation of claudin 6 and down-regulation of claudin 11 in UWB1.289+BRCA1 cells compared with UWB1.289 cells, both in untreated cells and those exposed to irradiation. Because regulation of the claudin genes is not due to irradiation in this study, it is likely the result of expression of wild-type BRCA1 in UWB1.289 cells. To our knowledge, a link between claudins and BRCA1, or between claudin 11 and cancer, has not been shown previously. Claudin genes, which code for tight junction proteins, are believed to play a role in cancer cell invasion and metastasis (21, 22) and have been implicated in breast and ovarian cancer. Reduced expression of claudin 7 was associated with invasion and metastasis of esophageal cancers (54). Claudin 3 was recently shown to be down-regulated in invasive, compared with noninvasive, breast cancer cells (55). Interestingly, claudin 6 expression, which we found to be highly up-regulated in the presence of wild-type BRCA1, was preferentially expressed in mammary epithelial cells from Copenhagen rats, which are resistant to mammary cancer, versus more susceptible Buffalo rats (56). The same authors found reduced expression of claudin 6 in human breast cancer cell lines and tissue samples in comparison with normal breast tissue. Our observations are intriguing in that they suggest a role for BRCA1 in the maintenance of claudin 6 expression and thus the metastatic potential of ovarian cancer.

In addition to the claudin genes, our expression array experiments revealed BRCA1 regulation of multiple additional genes associated with cell adhesion processes. Strikingly, many of these genes encode extracellular matrix proteins (CLDN6, CLDN11, CDH6, FLRT2, COL8A1, and ECM2) that function as cell-cell adhesion molecules. These genes normally play a key role in cell morphogenesis and maintenance of orderly structure and, when disrupted, may contribute to the invasion and metastasis of cancer. Interestingly, a previous study found that reverse transcription-PCR detected FLRT2 expression in a number of human tissues, with highest expression in the ovary (57).

IFN-responsive genes have been identified previously in gene expression array experiments in relation to BRCA1. Andrews et al. observed that multiple IFN-inducible genes were up-regulated in response to up-regulated BRCA1 expression in a tetracycline-regulated breast cancer cell line (58). These studies were only done in duplicate; therefore, significance levels were unknown. In contrast, Jazaeri et al. identified a set IFN-responsive genes that were up-regulated in ovarian tumors, including BRCA1-mutated tumors versus SV40-immortalized ovarian surface epithelial cell lines (59). In both our analysis and the Jazaeri et al. experiments, IFN-inducible transmembrane protein 1 (9-27) was down-regulated in the samples with wild-type BRCA1, suggesting a role of BRCA1 in disruption of the IFN-inducible pathway in a subset of ovarian cancers. Interestingly, a recent study found that overexpression of IFN-inducible 9-27 in gastric cancer cells was associated with increased resistance to immune cells and invasive potential (60).

IFI16, a member of the IFN-inducible p200 protein family, was down-regulated in irradiated (17-fold) and untreated (24-fold) UWB1.289+BRCA1 cells compared with UWB1.289 cells under the same conditions (see Table 2). IFI16 is one of the few IFN-inducible genes directly involved in the regulation of cell proliferation and differentiation (61). There is increasing evidence that members of the IFI p200 family (including IFI16) may be regulated by stimuli, independent of the IFNs (59, 61, 62). In our array analysis, we observed down-regulation of multiple IFN-inducible genes when wild-type BRCA1 was expressed. Strikingly, down-regulation of these genes occurred in the absence of any measured changes in expression of IFN genes, such as IFN-α or IFN-γ. Thus, our study suggests that BRCA1-mediated regulation of IFN-inducible genes may be one alternate mechanism of IFN-inducible gene activation.

In summary, we describe the molecular features of a novel BRCA1-null ovarian cancer cell line UWB1.289 and examine the role of BRCA1 in the DNA damage response in this cell line in comparison with UWB1.289 cells in which wild-type BRCA1 expression has been stably reconstituted. UWB1.289 cells are sensitive to radiation and lack both S and G₂-M cell cycle checkpoints. Restoration of wild-type BRCA1 expression in UWB1.289 cells partially corrects these DNA damage response deficits. This result is supported by increased expression of DNA repair and cell cycle genes in response to IR in UWB1.289+BRCA1 cells. In addition, our expression array experiments revealed that expression of a set of extracellular matrix genes was notably altered, whereas a set of IFN-inducible genes were significantly down-regulated, in cells not as a result of exposure to IR but as a result of restoration of wild-type BRCA1 expression in UWB1.289 cells. Because these genes are known to play a role in cell invasion and metastasis and seem to be regulated by BRCA1, we suspect that they are involved in ovarian carcinogenesis. Finally, because the vast majority of reports of BRCA1 function have been based solely on studies done in breast cancer cell line HCC1937 (20, 23, 32, 47, 50), we expect that experiments in additional BRCA1-null cell lines will enable more robust conclusions regarding BRCA1 function, particularly the elucidation of the role of BRCA1 in ovarian tumorigenesis.

The HBL-100 and MDA MB 231 cell lines were purchased from the American Type Culture Collection (Manassas, VA) and maintained as directed. The HCC1937 cell line was a generous gift from Dr. Gail Tomlinson (University of Texas Southwestern Medical Center), and HCC1937+BRCA1 cells were developed in Dr. Weixin Wang's laboratory with a protocol identical to that used to reconstitute wild-type BRCA1 in the UWB1.289 cell line (see below). The AT221JETpEBS7-YZ5 (ATM positive) cell line was a generous gift from Dr. Yosef Shiloh (Tel Aviv University, Israel) and was maintained as described (63). The UWB1.289 cell line was maintained in 50% RPMI + 50% MEGM, supplemented with 3% fetal bovine serum, and UWB1.289+BRCA1 was maintained in the same media plus 200 μg/mL G418. The vector pcDNA3-BRCA1 plasmid (23) was generously provided by Dr. David Livingston (Dana-Farber Cancer Institute, Boston, MA) to Dr. Weixin Wang and was transfected into the UWB1.289 cell line using 5 μg Fugene 6 (Roche, Indianapolis, IN) per 100-mm plate of cells, according to the manufacturer's recommendations. Stable cell lines were established by selection of positive clones that grew in media containing G418. Irradiation of cells for DNA damage experiments was done at room temperature using a Gammacell 40 ¹³⁷Cesium irradiator (Atomic Energy of Canada, Ltd., Mississauga, Ontario, Canada). All flow cytometry was done in the flow cytometry core facility of Dr. Peter Rabinovitch (University of Washington Department of Pathology).

Under approval from the Human Subjects Division of the University of Washington Institutional Review Board, tissue and blood were collected from a patient with a germline BRCA1 mutation (2594delC) undergoing surgery for recurrent ovarian cancer. Fresh tumor was mechanically minced and plated in several types of media. Pure epithelial clones were isolated by subcloning and differential trypsinization until fibroblasts were eliminated. The tumor cell line was maintained at 37°C with 5% CO₂ in 50% RPMI + 50% MEGM, supplemented with 3% fetal bovine serum. The stable cell line was named UWB1.289 (University of Washington-BRCA1-family 289). Immunohistochemistry was done on cultured cells using cytokeratin AE1/AE3 antibodies to assess epithelial purity.

The BRCA1 mutation was originally detected by protein truncation tests done on both germline and cell line DNA using PCR amplification of BRCA1 exon 11 followed by in vitro transcription/translation with the TNT T7 Quick for PCR DNA kit (Promega, Madison, WI) and visualization with SDS-PAGE. DNA from lymphocytes and the UWB1.289 cell line were sequenced around BRCA1 nucleotide 2594 using Big Dye Terminator chemistry and an automated 3100 DNA sequencer (Applied Biosystems, Foster City, CA). Direct DNA sequencing for UWB1.289 was done for all coding exons and flanking regulatory regions of the tumor suppressor genes TP53 and PTEN. Loss of heterozygosity was evaluated at the intragenic BRCA1 marker D17S855 using standard methodology. Karyotype analysis was done by the University of Washington Cytogenetics core facility on metaphase spreads for five cells of the stable cell line.

One-step real-time reverse transcription-PCR was used to evaluate BRCA1 transcript in total RNA samples extracted from BRCA1-null cell lines with Trizol (Invitrogen, San Diego, CA). A Lightcycler-RNA Amplification kit SYBR Green I (Roche) apparatus was used to evaluate transcript expression (64). A fluorescent dye that binds double-stranded DNA and a real-time fluorescence detection system allow the number of cycles to be calibrated for each amplicon, and cDNA copy numbers are estimated based on an internal standard curve. Two different pairs of BRCA1 primers were used: BRCA1 primer set 1 amplified a region across exons 2 to 7 (forward, 5′-CAAGGAACCTGTCTCCACAAAGTG-3′) and BRCA1 primer set 2 amplified a region across exons 9 to 11 (forward, 5′-CCAACTCTCTAACCTTGGAACTGTG-3′). Glyceraldehyde-3-phosphate dehydrogenase transcript was also amplified and used to normalize the quantity of BRCA1 transcript in each cell line. All reactions were run in duplicate.

BRCA1 and hemagglutinin protein expression in cell lines was visualized by SDS-PAGE/Western blot. Nuclear protein extracts were harvested as described (65) during logarithmic cell growth from UWB1.289, UWB1.289+BRCA1, and HBL-100 cells. Protein was quantitated with a colorimetric assay (Bio-Rad, Richmond, CA), and equal amounts of nuclear protein extract were loaded on a Nu-Page 3-8% Tris-Acetate gel (Invitrogen). Proteins were transferred for 4 h to Immun-Blot polyvinylidene difluoride membrane (Bio-Rad). Western blots were blocked for one hour in PBS + 6% nonfat dry milk and probed overnight at 4°C with BRCA1 Ab-1 (MS110; Oncogene Research Products, San Diego, CA) diluted 1:250 in PBS + 0.05% Tween + 3% nonfat dry milk. Blots were washed and probed for 1 h (for BRCA1 visualization) at room temperature with goat anti-mouse horseradish peroxidase–conjugated secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA) diluted 1:20,000. For hemagglutinin visualization, the blot was stripped and reprobed with a high-affinity rat monoclonal anti-hemagglutinin antibody (Roche) diluted 1:1,000, washed, and then probed with a donkey anti-rat horseradish peroxidase–conjugated secondary antibody (The Jackson Laboratory, Bar Harbor, ME) diluted 1:10,000. Blots were developed using enhanced chemiluminescence (Amersham, Arlington Heights, IL).

Radiation sensitivity of cell lines was evaluated using methods described previously (23). Cells were plated in triplicate for optimal density at counting. During the logarithmic growth phase, cells were trypsinized and irradiated with 0, 0.5, 1, 2, or 4 Gy IR. After IR, cells were re-plated and incubated in normal culture conditions for 10 to 14 days. Cells were fixed and stained with crystal violet, and colonies having >20 cells were counted visually. The number of colonies counted in each test plate was normalized to the number of colonies counted in unirradiated control plates.

Cells were irradiated during logarithmic growth with 0 or 10 Gy IR. After IR, cells were incubated for 15 or 24 h and then harvested and fixed/labeled with 4′,6-diamidino-2-phenylindole fix solution [10 μg/mL 4′,6-diamidino-2-phenylindole and 0.1% NP40 detergent in TBS (pH 7.4)]. DNA content quantitation and cell cycle analysis were done as previously described (66). Briefly, the cell suspension was triturated with a 26-gauge needle and analyzed using a high-speed Influx flow cytometer and Spigot software (Cytopeia, Inc., Seattle, WA) with UV excitation and 4′,6-diamidino-2-phenylindole emission collected with a 450-40 UV filter. Cell cycle data were analyzed using WinCycle software (Phoenix Flow Systems, San Diego, CA).

Radioresistant DNA synthesis assays were done using protocols modified from previous reports (32, 67). Briefly, 5 × 10⁴ cells were plated in triplicate for 24 h. The culture medium was changed to thymidine-free RPMI + 10% fetal bovine serum + 20 mmol/L HEPES; ¹⁴C-thymidine was added to each plate; and the cells were incubated for 15 to 17 h under standard conditions. The ¹⁴C-thymidine–containing media was removed; fresh RPMI and HEPES were added; and the cells were irradiated with 0 or 10 Gy IR. Immediately after IR, ³H-thymidine was added to each plate, and the cells were incubated for 15 h. The cells were then washed with PBS and incubated in fresh RPMI for 30 min. After washing, the cells were lysed in 500 μL of 0.25 mol/L NaOH, and the mixed cell suspension was added to 7.5 mL Opti-Fluor scintillation fluid (Packard Bioscience, Meriden, CT). ³H/¹⁴C dpm ratios were counted using a dual-label counting program on a Tri-Carb 2300TR liquid scintillation analyzer (Packard Bioscience). DNA synthesis values were calculated as a percentage of the values obtained from unirradiated control cells run in parallel.

Cells were prepared for G₂-M assays as described previously (32, 50). Briefly, logarithmically growing cells were irradiated with 0 or 6 Gy IR, incubated for 3 h, and treated with 1 μg/mL Nocodazole (Sigma, St. Louis, MO) for 20 h (50). After incubation, cells were trypsinized, harvested, and fixed in 70% ethanol at −20°C for 12 to 48 h. After two washes with PBS, the cells were incubated in 0.25% Triton X-100 on ice for 15 min. Cells were then centrifuged and resuspended in 0.75 μg of rabbit polyclonal, anti–phospho-histone H3 antibody (Upstate Biotechnology, Lake Placid, NY) in 100 μL PBS + 1% bovine serum albumin for 3 h at room temperature. Cells were rinsed with PBS + 1% bovine serum albumin and incubated in Alexa 488–conjugated goat ant-rabbit IgG (The Jackson Laboratory) diluted 1:30 in PBS + 1% bovine serum albumin for 30 min in the dark. Cells were washed with PBS and fixed in 4′,6-diamidino-2-phenylindole fix solution. Flow cytometry analysis was done on the Influx flow cytometer using a 488-nm argon excitation laser and a 525-30 filter to collect Alexa 488 emission. WinCycle software was used for data analysis.

GraphPad Prism software (San Diego, CA) was used to calculate SE and statistical differences with one-way ANOVA analysis and Tukey's multiple comparison post-test using a confidence interval of 95%.

Total RNA was extracted from triplicate plates of 10⁶ UWB1.289 and UWB1.289+BRCA1 cells 5 h after exposure to 10 Gy IR with Trizol (Sigma) and chloroform. Triplicate plates of 10⁶ untreated UWB1.289 and UWB1.289+BRCA1 cells were used to control for the effects of ionizing radiation. RNA was cleaned up using RNEasy midi-columns (Qiagen, Chatsworth, CA), and quality was assessed with a Bioanalyzer 2100. Three RNA samples each from IR-treated and untreated UWB1.289 and UWB1.289+BRCA1 cells (for a total of 12 samples hybridized to 12 chips) were labeled and hybridized to GeneChip Human Genome U133 Plus 2.0 Affymetrix chips at the University of Washington Center for Expression Arrays. Chips were washed and scanned at the University of Washington Center for Expression Arrays. Genechip Operating Software v1.3 (Affymetrix, South San Francisco, CA) was used to extract images and normalize chip data. All microarray data was analyzed using Genesifter software (Seattle, WA). To generate gene lists, log transformed data was compared between groups using Welch's t test with Bonferroni correction and median normalization. Quality was set at 1.0.

We thank Dr. Danbin Xu and YiJun Wang for real-time reverse transcription-PCR analysis, Dr. Tom Walsh and Melissa Wollan for PTEN sequencing, Dr. Ming Lee for help with microarray data analysis, and Dr. Peter Rabinovitch, Maria Bravo, and Jeanne Fredrickson for helpful discussions.