BRCA: What we now know

September 2006
Feature Story

William Check, PhD

It is well known that the identification in the early 1990s of the BRCA1 and BRCA2 genes as the genetic elements underlying inherited breast and ovarian cancer, followed by rapid implementation of clinical testing for mutations in these genes, has been one of the most medically important achievements to date of genomics techniques. But how many people know what the initials BRCA stand for? Is it breast cancer, Berkeley, Calif., or Paul Broca?

In fact, all three are correct. “Breast cancer” is obvious. And the scientist who discovered the BRCA genes, Mary-Claire King, PhD, did her doctoral work and the two decades of research leading to the identification of the genes at the University of California, Berkeley. But Paul Broca? Dr. King notes that the French pathologist who discovered the speech production center in the frontal lobes also was one of the first to recognize breast cancer pedigrees, in the 1860s. “I wanted it to stand for Broca, but I was only allowed four letters,” says Dr. King, who is now American Cancer Society professor of medicine (medical genetics) and genome sciences at the University of Washington School of Medicine, Seattle.

Several more important aspects of the epidemiology of the BRCA genes and the clinical implications of being a carrier are also not so obvious, and have emerged only after more than a decade of further research by Dr. King and others. Some of these aspects are:

• Risks from having a BRCA mutation are influenced by ethnicity.
• Women carrying a BRCA mutation have a greater than 80 percent lifetime risk of having breast cancer.
• Expression of the BRCA genes is affected by many of the same nongenetic modifiers that increase or decrease risk of sporadic breast cancer, such as weight, diet, exercise, and age at menarche.
• An inherited predisposition to breast cancer is often found in women with no family history of breast or ovarian cancer.
• It is still not understood why, if inherited BRCA mutations are present in all cells, only certain tissues are more susceptible to cancer.
• Large rearrangements of BRCA genes may play a part in sporadic breast cancer.
• Surveillance of women carrying BRCA mutations has raised an alternative theory of the histological origin of ovarian cancer.

Dr. King’s work on the BRCA genes followed her dissertation research with the preeminent biochemical geneticist Allan Wilson, whose groundbreaking work on the timespan of hominid evolution revolutionized evolutionary thinking (King MD, Wilson AC. Science. 1975;188:107–116). Her decision to take up the search for breast cancer predisposition genes arose from the fact that “it was clearly an important problem,” but also for scientific reasons. “I thought of it as a problem in quantitative genetics and integrative molecular genetics, what is now called genomics,” she says. Awareness of families affected with breast cancer to a degree greater than could be explained by chance went back to the Greeks. Modern work started with Broca and continued with a study of British mortality records in the 1920s, followed by decades of epidemiologic genetics.

To the standard technique of using linkage analysis as proof of the existence of a genetic basis for an inherited disease, Dr. King added two newer ideas. First was the two-hit model proposed by Knudsen for retinoblastoma, which postulated that familial cancer was a consequence of an inherited mutation and a somatic mutation in the same gene. Second was the modification that, for breast cancer, only a small fraction of families would be of that sort. A typical patient would get breast cancer through a series of specific mutations in a somatic cell and its daughters. Only a minority would have cancer through inheriting one mutation followed by somatic mutations. “For persons inheriting a BRCA mutation, fewer somatic steps are required, because the first step is already present,” Dr. King explains. “If the other copy of that gene is lost in breast or ovarian tissue, that cell lineage goes on to become malignant.”

Previously, linkage analysis techniques had been used only for traits that were clearly inherited in a Mendelian way, such as cystic fibrosis and Huntington disease. “What I brought to the problem,” Dr. King says, “was the notion that you can use the same approach for complex traits that are mostly not inherited, and if you are right, there will be a gene there. We didn’t know how big that subset would be,” she recalls, “so we did that first with models” (Schwartz AG, et al. J Natl Cancer Inst. 1985;75:665–668; Newman B, et al. Proc Natl Acad Sci U S A. 1988;85: 3044–3048). Linkage was established to chromosome 17q (Hall JM, et al. Am J Hum Genet. 1992;50:1235–1242); a high-density map of that segment was made (Anderson LA, et al. Genomics. 1993;17: 618–623); and the identity of a defined region as BRCA1 was established by linking germline mutations to breast and ovarian cancer in 10 families (Friedman LS, et al. Nat Genet. 1994; 8:399–404).

Current estimates are that about one in 1,000 women carry a BRCA mutation and that inheritance of BRCA mutations accounts for five percent to seven percent of all breast cancer. Also, says Dr. King, “The mutational spectrum of BRCA genes is enormously broad, with thousands of mutations known in each. If a family walks into my lab right now, more than 10 years after the genes were defined, it is still possible that we will see a mutation that we have never seen before.” More than 1,500 unique and independently arising deleterious mutations are in the database of Myriad Genetic Laboratories, says Brian Ward, PhD, senior vice president at Myriad, which holds the patent on the BRCA genes and is the only commercial laboratory in the U.S. to test for BRCA mutations in the general population. A well-known exception to all these statements is the Ashkenazi Jewish population, in which 10 percent of breast cancer is due to BRCA mutations, 90 percent of inherited breast cancer is accounted for by three founder mutations, and one of these mutations is found in about one percent of young adult Ashkenazis in New York. Also, Myriad licenses several laboratories to test the Ashkenazi population.

Differences have also been seen between Caucasian and African-American women, both in the spectrum of mutations (Nanda R, et al. JAMA. 2005; 294:1925–1933) and in the fraction of breast cancer cases accounted for by BRCA mutations (Malone KE, et al. Cancer Res. 2006;66:8297–8308). Among white women with breast cancer, five percent carry a BRCA mutation, compared with four percent of black patients. Although slightly lower, the incidence of BRCA-associated breast cancer in African-American women supports genetic testing for BRCA gene mutations in high-risk families in this population.

In addition to breast and ovarian cancer in females, mutations in BRCA genes are also associated with an increased risk of male breast cancer; of pancreas cancer in men and women, especially for BRCA2 (though the absolute risk is low, less than one percent); and, for BRCA2, of a rare ocular melanoma in men and women. It is not always appreciated that men who carry a BRCA mutation should be screened. Dr. King also finds that some physicians working with patients at high risk of breast cancer focus only on inheritance from the mother. “It is important to reinforce to one’s clinical colleagues that susceptibility to sex-specific cancers can be inherited maternally or paternally,” she says. “The risk is not inherited as a sex-linked trait.”

Establishing a clinical testing program and monitoring its value requires the expertise of a clinical molecular geneticist. While Myriad is the only U.S. laboratory with extensive experience in BRCA testing, in April 2000 the provincial government of Ontario authorized the provincially funded molecular genetics laboratories to test for BRCA mutations. Diane J. Allingham-Hawkins, PhD, FCCMG, is the director of the Molecular Genetics and Cytogenetics Laboratory at North York General Hospital, Toronto. “For any molecular test, what you are interested in is clinical utility and clinical validity,” Dr. Allingham-Hawkins says. “With the BRCA genes we are pretty clear there is clinical validity. When you find a mutation in one of these genes, there is a strong association with the risk of developing breast or ovarian cancer over the lifetime of that individual.”

With regard to clinical utility—can you do anything for the patient?—BRCA testing again rates a positive grade from Dr. Allingham-Hawkins. “When you get a positive result, there is some benefit,” she says. Possible interventions for carriers range from increased surveillance with breast examination or ultrasound to chemoprevention with oral contraceptives or tamoxifen to prophylactic mastectomy or oophorectomy. (For recommendations from the Cancer Genetics Studies Consortium, see Burke W, et al. JAMA. 1997;277:997–1003.) Family members can benefit by having the option of knowing what is implied for their future health.

Evidence supporting the benefit of prophylactic surgery has been published. However, the authors of a review of BRCA testing on the GeneTests Web site conclude that “None of these strategies has been assessed by randomized trials or case control studies in high-risk women, and no studies have evaluated the effects of these interventions in individuals with a BRCA1 or BRCA2 cancer-predisposing mutation. As a result, current recommendations are made on the basis of expert opinion”. A simpler summary of facts and recommendations, suitable for patients, is on the Myriad Web site. A site for geneticists, which includes a list of known mutations, is also available research.nhgri.nih.gov/bic/.

Another important criterion for any test is accuracy, which is difficult to assess for BRCA. A major challenge to evaluating methods is that many changes are clearly disease-causing, such as truncating mutations, but other changes—called unclassified variants—are ambiguous. “Sometimes a lab has to report that a change may or may not cause disease,” Dr. Allingham-Hawkins says. “This complexity of reporting occurs with any test where we are doing ‘mutation-hunting.’ When an unclassified variant does occur, it can be helpful to test family members and see whether the genetic change segregates with the disease.

“There is controversy around testing for these genes,” Dr. Allingham-Hawkins says. “What is the best way to do it?” Myriad does full sequence determination of both BRCA1 and BRCA2 and detection of five specific large genomic rearrangements of the BRCA1 gene. It also offers “single-site” testing for known familial mutations. Dr. Allingham-Hawkins uses a protein truncation test, or PTT, with selective sequencing. “We are really selecting for disease-causing mutations,” she says. “If we find a frameshift, we sequence that fragment and look for a mutation.” She notes that estimating the sensitivity of any test is difficult, especially for genes like the BRCA genes where thousands of genetic changes have been identified. “The sensitivity of the PTT is estimated to be in the 85 to 90 percent range,” she says. “Everyone thinks their test is the most appropriate one.” However, as technology improves, so do testing methods, and the Ontario labs are in the final stages of a large study comparing the sensitivity of an alternative BRCA testing algorithm using denaturing high-performance liquid chromatography, or DHPLC, and the multiple ligation probe amplification assay, or MLPA, to the one currently used.

Dr. King agrees that methodology is an issue for BRCA testing. She likes the multiple ligation probe amplification assay (used in Europe), because it detects whether a protein or gene has fewer or more than two copies, indicating a deletion or duplication. For clinical testing, she recommends full sequencing of both genes, followed by MLPA if no mutation is found by sequencing. Earlier this year, Dr. King and colleagues published results of using MLPA and other methods to examine BRCA genes in 300 women from families with four or more cases of breast or ovarian cancer but with negative commercial genetic test results. They found undetected mutations in 52 women (17 percent) (Walsh T, et al. JAMA. 2006;295:1379–1388). In response, on Aug. 1 Myriad introduced the BRACAnalysis Rearrangement DNA sequencing assay to detect large duplications and deletions in high-risk women. Myriad’s Dr. Ward says the new assay, which is a modification of quantitative real-time PCR and which has been validated in-house, will be used as an adjunct for “exceptionally” high-risk women—basically any woman with approximately 30 percent or greater a priori risk—who test negative with Myriad’s primary sequencing test.

“I haven’t seen Myriad’s new test in action,” Dr. King says, “but it has not been independently verified.” She adds, “It is interesting that genetic testing does not require FDA approval. Anyone can market any kind of genetic test.” An even bigger controversy concerned the lifetime risk of carrying a BRCA mutation. Several groups claimed that the true value was lower than Dr. King’s estimates, especially after age 60. “There are two basic ways to determine the risk associated with mutations,” Dr. King explains. In one method you work with families that have multiple cases of disease and do complete molecular testing of everyone. Then you extrapolate back from severely affected families to the population as a whole. In the second method, you start with breast cancer cases without regard to mutations, then try to model risk of the disease. “We initially did the first,” Dr. King says. Groups that disputed her figures used the second method (for example: Antoniou, et al. Am J Hum Genet. 2003;72:1117–1130).

In an attempt to settle the disagreement, Dr. King led a study combining the strengths of both methods (King MC, et al. Science. 2003;302:643–646). Starting in the mid-1990s, mutation analysis was done on 1,008 consecutive women self-identified as being of Ashkenazi Jewish ancestry who presented at any of 12 sites in New York with breast or ovarian cancer independent of family history. A BRCA mutation was found in 103 (10.2 percent) women. Molecular analysis was performed on every at-risk woman in the families of these 103 probands. For deceased relatives, genotypes were constructed from children’s genotypes or from archival tissue. A very high cumulative risk of breast cancer was verified for women who had a BRCA mutation—37 percent by age 50, rising to 82 percent by age 80. The risk of ovarian cancer for BRCA1 mutation carriers was 21 percent and 54 percent at these ages. When the analysis was repeated leaving out probands and mothers to eliminate ascertainment bias, the same level of breast cancer risk was found. An exchange of letters ensued (Science. 2004;306:2187–2191). In a subsequent publication, one group that originally advocated lower estimates found agreement with Dr. King’s figures at least for BRCA1 Ashkenazi mutations (Antoniou AC, et al. J Med Genet. 2005;42:602–603).

Other important observations derived from Dr. King’s study of Ashkenazi women. Half of all women with BRCA mutations came from families with no history of breast or ovarian cancer. Dr. King says there is a simple explanation for this observation. Half of the time a woman inherits a BRCA mutation from her father, which virtually eliminates parents from expression of the trait. Alternatively, a woman might inherit the mutation from her mother who died early before expressing it. Since the current generation in the U.S. often has small families, a woman may have no sisters, or her sister(s) may have inherited a normal BRCA allele by chance. It all adds up to a family in which a mutation co-exists with a low incidence of cancer. An explicit analysis of such families in the study reinforced that “low incidence was not equivalent to low risk.”

Further analysis showed that the risk for mutation carriers differed over historical time: Women born before 1940 had lower age-specific rates of breast cancer than those born after 1940. “Among women born in the early part of the 20th century who carried BRCA mutations, the risk age-for-age was lower than for women born in the 1950s who carry BRCA mutations,” Dr. King says. There have to be non-genetic influences on risk, she concludes, probably including risk factors for breast cancer generally. For instance, women who were never pregnant have an earlier onset of breast cancer, which is true also of women who carry mutations. For women with mutations, being physically active as young girls shifts the curve for onset of breast cancer to the right. The same is true for healthy weight. “So activity and healthy weight protect somewhat for age of onset, just as in nonmutation carriers,” Dr. King says. In another study, age at menarche was inversely associated with the risk of breast cancer, at least among BRCA1 carriers.

Notably, these are all factors that regulate estrogen production, which may eventually help to explain what Dr. King calls “the remaining most important theoretical question about the BRCA genes.” A mutation inherited through the germline is present in all cells in the body. So why do only some tissues have a higher risk of developing cancer? “All tissues that get cancer if the second of the BRCA gene gets lost are estrogen-responsive tissues,” Dr. King points out. “So the role for estrogen in them is quite logical.” Current thinking is that BRCA proteins may be ligases that play a role in DNA repair. But it is not yet clear what about the two proteins made by BRCA1 and BRCA2 affects estrogen-mediated proliferation of breast and ovarian epithelium.

Two revolutionary ideas deriving from research on the BRCA genes are now in play. First, some BRCA1 mutations may be active in sporadic breast cancer. People have traditionally thought that only inherited BRCA mutations play a role in breast cancer. “I think they were proceeding from incorrect reasoning,” Dr. King says. “People were using as a model either retinoblastoma or Li-Fraumeni cancers due to inherited mutations in the p53 genes. In both of those situations inherited mutations are point mutations.” Since no point mutations were found in BRCA genes in women with sporadic breast cancer, people concluded that BRCA mutations are not important if they are not inherited.

Now, however, there is evidence to the contrary. Large somatic mutations have been found in women with non-inherited breast cancer that delete all or part of the BRCA1 gene or the entire arm of a chromosome or that silence the BRCA1 gene. When molecular profiling is done on these non-inherited breast cancers, they show the same profile as breast cancer due to inherited mutations in BRCA1 genes, in addition to non-expression of BRCA1 itself. “So now very recently we are seeing, because of molecular pathology, a convergence of BRCA1 profiles between women with inherited mutations and non-inherited breast cancer,” Dr. King says. (Non-inherited mutations can be distinguished because they are found only in affected tissues rather than throughout the body.)

Finally, prevailing notions of the histological origin of ovarian cancer are being brought into question. Because women with BRCA mutations are intensely screened, Dr. King says, one sometimes detects small, easily treated tumors in the fallopian tubes. “Does that mean that BRCA-associated ovarian cancer arises in the fallopian tubes?” she asks. “If that is true, could some non-inherited ovarian cancer also arise in the fallopian tubes?” Such an idea changes the target of surveillance. Work is ongoing between geneticists and gynecologic oncologists to investigate this possibility. “It would be fabulous if that proved useful for women generally,” Dr. King says. It would also be one more clinical benefit of the discovery of the BRCA genes.