A step toward genotype-based therapeutic regimens for breast cancer in patients with BRCA2 mutations?

The positional cloning of a familial tumor suppressor gene is a monumental task often involving the coordinated efforts of international consortia of many laboratories. The euphoria that accompanies the successful cloning of a cancer-causing gene signals the initiation of many new avenues of study that attempt to define the activity of that gene’s product. These goals are laudatory, because by defining such a function and understanding how it is subverted by mutations, the hope is that eventually this information can be used to develop therapeutic approaches aimed at patients and their family members who participated selflessly in the genetic studies that ultimately led to the identification of the gene. The biologic functions of gene products are hugely diverse, and our knowledge base of gene function is currently so poor that the experimental pathways that lead to the elucidation of a gene’s function are not necessarily easy to discern. Usually, many years elapse between the cloning of a familial tumor suppressor gene and the identification of its function. Moreover, although valuable information will be garnered about a gene’s function, translating this knowledge into effective therapeutic regimens, based on the knowledge of an individual’s genetic lesion, is extremely difficult. However, an article appearing in this issue of the Journal (1), as well as two other articles (2,3), suggests that genotype-based therapeutics for familial breast cancer caused by lesions in the BRCA2 gene may be possible. All of these studies show that cell lines lacking functional BRCA2 are hypersensitive to agents that cause double-strand DNA breaks. Moreover, Abbott et al. (1) demonstrate, in nude mice, that xenografts containing mutant BRCA2 genes are dramatically more sensitive to both ionizing radiation and mitoxantrone, implying that tumors arising in patients with BRCA2 mutations may also be highly sensitive to these therapies. How did this important realization come about so rapidly? The BRCA2 gene sequence was deposited in GenBank just under 2.5 years ago (4,5). However, the nucleotide sequence provided no clues as to the function of this huge protein, which suggested initially that the experimental pathway to define the function of BRCA2 was likely to be extremely long. As is often the case when such an important disease-associated gene is cloned, many groups (including ours) rushed to clone the mouse homologue (Brca2) and to generate mice lacking this gene. As a prelude to these studies, the mouse gene was cloned (6,7). The sequence of the mouse gene in comparison with its human counterpart provided some surprises. The sequence was poorly conserved for a tumor suppressor gene; for instance, although the genes APC, WT1, NF1, and RB1 exhibit greater than 90% identity between the two species, the BRCA2 gene exhibits only 59% identity with the corresponding mouse gene. This low identity required that the gene be mapped in the mouse genome to provide some confidence that the reported sequences were from the authentic homologue. It is interesting that the poor conservation of the sequence was quite helpful in the analysis of this giant protein, because within the stretches of divergent sequences were several regions that were much more conserved (6,8), suggesting their likely functional significance. BRCA2 was not an exception; we now know that several of these highly conserved domains are functionally important, as shown in Fig. 1. Among those most relevant to this editorial are the BRC repeats in exon 11 and the highly conserved carboxyl terminus of the protein, both of which have been demonstrated to interact with the recombinational repair protein Rad51 (9,10). It is interesting that, in Paul Hasty’s laboratory, yeast two-hybrid studies (9) that used Rad51 as bait identified the Brca2 polypeptide as a potent interacting partner long before the Brca2 sequence was available in the database. This result and the BRC repeat– Rad51 interaction have been observed by many groups. The Brca2–Rad51 interaction, coupled with the prior observation of the radiation sensitivity of Rad51-deficient mouse embryos (11), made clear suggestions for the appropriate phenotype to examine in Brca2 knockout mice. Acute radiation sensitivity of Brca2-deficient cells was first reported by Sharan et al. (9), by irradiating embryos with a homozygous null Brca2 mutation (Brca2). Subsequently, these observations were repeated and confirmed by Connor et al. (12) who used cells with a hypomorphic Brca2 mutation that is compatible with embryonic