Breast Cancer Genes and DNA Repair

See allHide authors and affiliations

Science  05 Nov 1999:
Vol. 286, Issue 5442, pp. 1100-1102
DOI: 10.1126/science.286.5442.1100

Hopes that the cloning of two inherited breast cancer susceptibility genes–BRCA1 and BRCA2–might illuminate the common mechanisms underlying this disease remain unfulfilled. About 10% of breast cancer patients have a familial form of the disease, and of these, inherited mutations in BRCA1 or BRCA2 are found in about half. However, somatic mutations in either gene are not a feature of the 90% of breast cancers that are sporadic (that is, not inherited) [reviewed in (1)]. Therefore, the biochemical connection between the BRCA1 protein and a protein kinase called ATM (mutated in ataxia telangiectasia) reported by Cortez et al. on page 1162 of this issue (2) is cause for considerable excitement because it defines the participation of BRCA1 in a cellular pathway that may be dysfunctional in a significant fraction of all breast cancers.

BRCA1 and BRCA2 both encode large nuclear proteins (1863 and 3418 amino acids, respectively). These proteins are expressed in many tissues and are most abundant during S phase of the cell cycle [reviewed in (3)]. The proteins are quite distinct despite the misleading similarity in their acronyms. There is, however, much circumstantial evidence to suggest that they have common biological functions. Thus, inheritance of one defective BRCA1 or BRCA2 allele predisposes an individual to developing breast or ovarian cancer. Homozygosity for targeted mutations in murine Brca1 or Brca2 precipitates defective cell division, chromosomal instability, and hypersensitivity to genotoxins indicative of defects in DNA repair (4-6).

Similar abnormalities occur in human or murine cells after disruption of the ATM gene, which provokes a disease characterized by cerebellar dysfunction, chromosomal instability, and predisposition to cancer (7). ATM belongs to a family of protein kinases homologous to the catalytic subunit of phosphoinositide 3-kinase. This family includes the related ATR (AT- and Rad3-related) protein kinase in vertebrates and MEC1 and Rad3 in yeast. These kinases are essential—and quite proximal—components in the pathways that signal cell cycle checkpoint arrest after DNA damage or incomplete DNA replication.

The observations of Cortez and co-workers now place BRCA1 downstream of ATM in these pathways. They show that ATM resides in a nuclear complex that contains BRCA1, and that it phosphorylates BRCA1 after exposure of cells to g radiation. Phosphorylation occurs in a cluster of serine-glutamine amino acid residues located toward the carboxyl terminus of BRCA1, in a region (amino acids 1280 to 1524) that also contains the recently identified cyclin-CDK2 phosphorylation site (8). A number of serine residues in BRCA1 become phosphorylated in vivo, both constitutively and after g radiation. Two of these—Ser1423 and Ser1524–are particularly important. A mutant protein in which both residues are replaced with alanine fails to rescue the radiation sensitivity of a BRCA1-deficient cell line. Thus, phosphorylation of BRCA1 by ATM may be important for protecting the cell against DNA damage.

BRCA1 phosphorylation is diminished, but not entirely absent, in ATM-deficient cells. Furthermore, phosphorylation of BRCA1 after DNA damage induced by agents other than g radiation (such as ultraviolet light or hydroxyurea) appears to be independent of ATM activity (9). Collectively, these findings speak to a specific role for ATM in signaling the cellular response to DNA lesions (such as double-strand breaks) induced by g radiation. In other situations, BRCA1 phosphorylation may be carried out by alternative pathways. A precedent for this comes from studies of the phosphorylation of the p53 tumor suppressor protein, an important component of the cellular response to DNA damage. In response to different stimuli, p53 can be phosphorylated by ATM or by the ATM homolog, ATR (10). Similarly, it is possible that ATR could phosphorylate BRCA1 when ATM does not.

Consistent with the particular importance of ATR and its yeast homologs in monitoring genome integrity during S phase, such an alternative phosphorylation pathway is germane to the idea that BRCA1 and BRCA2 participate in DNA repair through homologous recombination (see the figure) (3). Recombination mechanisms are essential not only for the repair of DNA damage during S phase, but also for normal DNA replication (11). The evidence linking BRCA1 and BRCA2 to recombination mechanisms has thus far been indirect. Both molecules are reported (6, 12) to associate with mRad51, the mammalian homolog of Escherichia coli RecA, which is essential for double-strand break repair through recombination. Furthermore, chromosomal aberrations reflecting defects in mitotic recombination occur in murine Brca1-deficient (7) and Brca2-deficient cells (5). Direct evidence substantiating a role for BRCA1 in recombination is now provided by new data that demonstrate a considerable reduction in the frequency of recombination between homologous DNA substrates integrated into the genome of Brca1-deficient cells (13). A model is now emerging in which phosphorylation of BRCA1 by ATM (and perhaps ATR) is an essential prelude to the recombination repair of DNA lesions that occur or persist during DNA replication (see the figure).

Repairing damaged DNA.

BRCA1 is phosphorylated by the ATM protein kinase in response to DNA damage induced by g radiation. Phosphorylated BRCA1 activates DNA repair through homologous recombination, in cooperation with BRCA2, mRad51, and other molecules related to members of the yeast RAD52 epistasis group. Phosphorylated BRCA1 may also regulate transcription and transcription-coupled DNA repair.

However, it is important to appreciate the multiplicity of additional—or alternative—functions proposed for BRCA1, many of which were discussed at a recent workshop (Second International Workshop on the Function of BRCA1 and BRCA2, 9–10 September 1999, Cambridge, UK). Some of the proposed functions, such as BRCA1's participation in cell cycle checkpoints during G2/M phase (7), are in accord with processes regulated by ATM and its relatives. The proposed involvement of BRCA1 in transcription, supported by manifold data (14), could also be regulated by phosphorylation. BRCA1 copurifies with the RNA polymerase II holoenzyme. Certain BRCA1 domains activate transcription when fused to heterologous DNA binding proteins, and may induce genes such as GADD45 that mediate apoptosis. Disruption of murine Brca1 cripples the repair of oxidative base damage on the transcribed DNA strand, consistent with Brca1's proposed involvement in transcription-coupled repair.

How BRCA1 may affect such apparently disparate processes can only be guessed at. A single function in transcriptional regulation could underpin multiple effects (see the figure). Alternatively, distinct functions may be performed through cooperation with different protein partners. Regardless, Cortez and colleagues show that a large fraction of cellular BRCA1 undergoes phosphorylation, at least after g radiation, suggesting that this is a crucial step in several different pathways.

As yet, it is unclear how the function of BRCA1 relates to that of BRCA2. The similarities in the phenotypes induced by disruption of these molecules, as well as their reported colocalization (15) along with mRad51 to the nucleus of somatic cells and synaptonemal complexes in meiotic cells, certainly suggest that they perform common functions. Nevertheless, there are important differences. In particular, BRCA2 binds with relatively high stoichiometry to mRad51 through the conserved BRC repeat region in exon 11 as well as through a carboxyl-terminal motif (6, 12), possibly speaking to a more direct role in recombination repair. On the other hand, halting cells with damaged DNA at a cell cycle checkpoint seems less affected by disruption of Brca2 than by disruption of Brca1. The new evidence placing BRCA1 downstream of ATM (and, by implication, ATR) suggests that BRCA1 may link the DNA repair functions of BRCA2 to the pathways that signal DNA damage or incomplete DNA replication.

Heterozygous mutations in ATM may predispose to breast cancer (16), although contrary evidence also exists (17). ATM heterozygosity affects between 0.5% and 1% of individuals, creating a large population who are potentially at risk. On this basis, up to 4% of sporadic breast cancers in women younger than 40 years of age may occur in ATM heterozygotes (18). Thus, the links between ATM, BRCA1, and (by inference) BRCA2 established by Cortez et al. define a cellular pathway that could be dysfunctional in a significant fraction of breast cancer patients. This fraction may increase still further if mutations affecting other components in the pathway (see the figure) also contribute to breast cancer. Further epidemiological studies will be needed to clarify this important issue.

Many important biological questions also remain. Not least among them is the paradox that although ATM, BRCA1, and BRCA2 are ubiquitously expressed and important in processes apparently fundamental to all cells, their disruption leads to an excess of cancer predisposition in particular tissues such as breast epithelium. The resolution of this paradox will surely be central to a better understanding of the genesis of breast cancer.


Stay Connected to Science

Navigate This Article