BRCA-associated cancer risk: Molecular biology and clinical practiceKenneth van Golen, Kara Milliron, Seena Davies, and Sofia D. Merajver
Ann Arbor, Michigan
Abbreviation: bp = base pair
After female sex and age, family history is the most important risk factor for breast cancer.1 Breast cancer is rare among women under 30, with the incidence increasing gradually but steadily through with age 90. The lifetime risk in women is about 1 in 200 at age 40, 1 in 50 by age 50, and 1 in 10 by age 70. Therefore, even in the presence of family history, age remains an important factor in the overall risk evaluation in a given affected or unaffected woman.
Epidemiologic and genetic studies have demonstrated that a family history of breast and ovarian cancer plays an important role in breast cancer risk. Women with affected first- and second-degree relatives have a significantly higher lifetime risk of developing breast cancer. The risk increases with the number of affected relatives, younger age of onset, bilateral disease, and the occurrence of ovarian and breast cancer in the same relative within the affected lineage. The presence of other cancers such as salivary gland cancer, gastrointestinal malignancies, and prostate cancer also appear to modulate breast cancer risk, but their quantitative influence, although not yet elucidated, is under avid study. The BRCA12 and BRCA23 genes account for the majority of familial breast cancer.
OVERVIEW OF BRCA1 AND BRCA2 FUNCTIONFunction of BRCA1 and BRCA2. According to Knudson's classic "two-hit" model that describes the inactivation of a tumor suppressor gene, both the BRCA1 and BRCA2 genes fit the traditional description of a putative tumor suppressor gene.4 In familial cancers, and individual inherits a germline mutation, thus this "first hit" is present in all cells of the body. A somatic mutation represents the "second hit" on a given cell, resulting in the loss or inactivation of the wild-type allele, thus rendering both copies of the gene inactive. In sporadic cancers, inactivation of a tumor suppressor gene is accomplished by two somatic mutations that eliminate or inactivate the alleles on both chromosomes. Knudson's model accurately accounts for the early onset of familial cancers caused by a pre-existing germline mutation, while the accumulation of two somatic mutations in a single cell may take several decades to produce a sporadic cancer.
The role of BRCA1 has been extensively investigated in sporadic breast and ovarian cancer.5-7 Several somatic mutations have been identified in the coding regions of BRCA1 in sporadic ovarian tumors; however, none have been found in sporadic breast cancer.8 All indications of involvement of the BRCA1 gene in the formation of sporadic breast carcinomas come from experiments regarding loss of heterzygosity9-11 and putative loss of protein by other mechanisms. It is apparent, however, that although the mutation studies of sporadic tumors examined the coding regions of BRCA1 and BRCA2 in detail, none of the previous studies would have detected a genomic deletion as a mechanism of BRCA inactivation. More-detailed studies aimed at detecting possible somatic genomic deletions in sporadic breast and ovarian cancers are needed.
The BRCA1 protein is a zinc-finger protein that has been found in one study to exhibit granin-like properties.12,13 Other studies provide strong evidence that BRCA1 has a role in double-stranded DNA repair14,15 and normal DNA recombination.16 Splice variants of BRCA1 mRNA have been identified that exist normally in nonmalignant breast cells.17,18 These alternatively spliced mRNAs code for truncated proteins; however, it is yet undetermined whether these truncated proteins exhibit tumor suppressor function. Truncation of the BRCA1 protein caused by inherited mutations in breast cancer-prone families is correlated with a high mitotic index of breast tumor cells from the affected patients.19 In addition, BRCA1 knockout mouse experiments indicate that at least one normal allele, coding for a full length protein, is required for normal embryonic development.20 This mouse model indicates that BRCA1 not only has tumor suppressor function but is also essential for normal embryo development.
The BRCA2 gene is notably similar to the BRCA1 gene. Both genes have more than 20 exons, with a very large exon 11 and a code for highly charged proteins with a putative granin domain. However, the distribution of detected mutations in BRCA2 is somewhat different than that in BRCA1. Whereas frame shift and various point mutations have been detected in BRCA1, the gene of patients with breast and ovarian cancer have been microdeletions resulting in a frame shift, while very few point mutations have been detected.21-25 Similar to BRCA1, founder mutations have been identified in the BRCA2 gene in members of defined ethnic groups.26,27 This once again well illustrated in the Ashkenazi Jewish population. A recurrent germline mutation, 6147delT, was detected in 8% of women diagnosed with early-onset breast cancer.28
Loss of heterozygosity at the BRCA2 locus has been observed in 30% to 40% of sporadic breast and ovarian cancers29; however, very few somatic mutations have been found in the remaining allele.30,31 This suggests that either BRCA2 is an infrequent target for somatic inactivation or that intron sequences or genomic deletions may be the target of somatic mutation. In a study of grade 3 breast carcinomaswhich were selected without regard to age, familial history, or mutation statusa concurrent loss of heterozygosity in the BRCA1 and BRCA2 loci was seen.31 This suggests a common pathway of tumorigenesis in familial and sporadic breast cancer that inactivates the two BRCA genes and may give rise to high-grade lesions.
Profile of mutations. It is estimated that mutations in BRCA1 alone account for 50% of all familial early-onset female breast cancers,33 while mutations in BRCA2 may be responsible for up to 35% of the remaining hereditary breast cancers.34,35 Several methods have been used to screen for mutations in these genes. These include direct sequencing of anomalous single-strand conformational polymorphism products,36 heteroduplex analysis,37,38 and a protein truncation test.39,40 During mutation screening a host of polymorphisms have been identified for both the BRCA1 and BRCA2 genes.41,42 Polymorphisms are missense alterations that cause either no change or a one amino acid substitution in the protein sequence. In general, polymorphisms do not, by definition, significantly modify the protein's function. Polymorphisms are found to varying degrees in the general population and are generally not associated with disease. It is possible, however, that certain missense alterations cause sublet modifications of protein structure that may affect function without causing overt clinical disease.
More than a hundred mutations have been identified in the BRCA1 gene since its isolation in 1994. The majority of these mutations were identified in individuals who belong to families with either breast cancer or both breast and ovarian cancer.43-45 Founder-effect mutations, common mutations presumably originating from a single ancestor within a historically isolated ethnic group, have been identified in several populations such as the Askhenazi Jews,46 the Japanese,47 the Italians,48 the Swedes,49 and the Dutch.50 A salient example is the 185delAG mutation in individuals of Ashkenazi Jewish decent.51-54 This frame shift mutation is a 2 bp deletion at base 185 in exon 2 that causes a premature truncation of the protein by producing a premature stop signal at codon 39.55 Other types of frame shift mutations, caused either by deletions of 1 to 40 bp or insertions of up to 11 bp, have been identified that also result in a premature stop codon.56,57 Additionally, single base pair substitutions resulting in missense, nonsense, or splicing mutations have been distinguished. Nearly 80% of mutations described in the BRCA1 gene would produce a truncated protein,58-63 although many mutations likely lead to little or no expression of the altered product primarily because of instability of the corresponding mRNA. A somewhat different set of mutations occur in the regulatory regions of the gene, such as methylation-sensitive CpG islands in promoters, enhancers, and repressors.64 These mutations cause an alteration in the level of gene transcription and are generally characterized by the complete absence of mRNA.
In addition to an increase in the risk of familial female breast and ovarian cancer,65-69 mutations in the BRCA2 gene are also associated with an increased risk of sporadic and familial forms of pancreatic,70-72 hepatic,73 prostate, and particularly male breast cancers.74 Mutations found in these types of cancers are similar to those found in breast and ovarian carcinomas.
Similar to findings in BRCA1 transgenic mouse experiments, BRCA2 knockout mice demonstrated that at least one normal copy of the BRCA2 gene is needed for embryogenesis.75 moreover, like BRCA1, BRCA2 is also associated with playing a role in DNA repair.76 This was further verified in experiments that introduced a truncation mutation in the mouse BRCA2 gene.77 These animal experiments have helped to interpret the in vitro work done on human normal and breast tumor cells that show cell cycle-dependent regulation of BRCA2 corresponding to an up-regulation of mRNA during S-phase and mitosis.78,79
The characterization of BRCA1 and BRCA2 mutations has enabled the study of cancer risk within extensively affected families. Population-wide studies have generally not been done on a large scale, nor do they appear yet warranted, given the cost and paucity of potential benefits. Some exceptions have been certain higher risk subpopulations such as Ashkenazi Jews,46 very young affected individuals, and specific Scandinavian cohorts.49
Truncation and certain missense mutations of BRCA1 and BRCA2 correlate strongly with the development of breast and ovarian cancer, while other mutations and polymorphisms have unknown significance. The risks conferred by BRCA1 and BRCA2 disease-associated mutations are best provided as ranges, as the data continue to evolve. BRCA1 carriers have a 60% to 85% lifetime risk of developing breast cancer and a 20% to 50% risk of developing ovarian cancer, in contrast to the 2% to 3% risk in the general population. BRCA2 carriers have similar risks, although the average age of onset for ovarian cancer appears to be similar to the age of sporadic cases. In addition, male BRCA2 carriers are at increased risk for male breast cancer, with an approximately 6% to 8% lifetime risk, as compared with the general male population risk of 0.1%.
Women affected with breast cancer who are BRCA carriers also have increased risk of another primary breast cancer in the remaining or contralateral breast of 30% to 50% by age 70. Thus carriers who are also survivors of breast or other cancers warrant as much counseling for adherence to surveillance as non-affected carriers. Although overall they appear to be less severely affected than females, male carriers in BRCA families transmit the defective allele and are possibly themselves at increased risk for prostate and gastrointestinal malignancies, in addition to male breast cancer in the case of BRCA2 carriers. Medical and psychosocial issues concerning males in BRCA families are important but largely unexplored areas of investigation.
In the Ashkenazi Jewish population, it is estimated that 2% to 3% of all women carry a defective BRCA gene (185delAG or 5382insC inBRCA1 and 6147delT in BRCA2). The founder mutations account for most of the excess breast cancer ascertained in this population. BRCA mutations account for approximately 10% of all breast cancers and have not been shown to play a significant role in sporadic breast cancers. Inactivation of the BRCA proteins by mechanisms other than coding region gene mutations may, however, turn out to play a role in the development of sporadic breast and ovarian cancer.
As more information becomes available about risks and outcomes in different subpopulations, more-specific risk information will be available to patients. For the time being, however, considerable caution is required when counseling all individuals about specific risks, especially non-caucasian persons who have, so far been extremely poorly represented in most genetic epidemiologic studies.
THE EVOLVING ROLE OF A COMPREHENSIVE CANCER RISK EVALUATION CLINICSpecific gene mutations play an important role in the determination of a cancer risk in certain individuals. However, and individual's concerns about cancer risk comprise many more variables than BRCA-associated mutations alone. Therefore, the role of the risk evaluation clinic has evolved, in some centers, into a comprehensive service of educational and counseling services, referrals, and research that focus on the client's and the family's questions and concerns.
The centerpiece of the precounseling evaluation is a thorough general medical and family history, as summarized in Table I. The medical history includes the detailed history of any cancers and surgeries, cardiovascular disease, osteoporosis, hormonal and reproductive history, medications, lifestyle (diet, exercise, exposures such as radiation), and cancer-screening practices. The family history not only details the specific cancer cases, including the confirmation of diagnoses, but also covers the history of cardiovascular disease, osteoporosis, connective tissue disease, spontaneous abortions, other genetic diseases, and chromosomal or anatomic abnormalities, all of which may bear on the cancer risk management recommendations discussed during the clinic visit. To achieve a meaningful clinic visit that addresses multiple patient concerns, much of the information detailed here must be collected before the patient ever meets the clinicians, in a rather drastic departure from the traditional model of clinical history-taking in the outpatient setting.
In families with evidence of autosomal dominant transmission of a BRCA mutation, genetic testing for BRCA1, BRCA2, TP53, or other tumor suppressor genes is discussed. However, whether testing actually proceeds in an interested individual depends on a large number of variables very specific to each client and each family. For example, the living individual "most likely" to evidence a BRCA mutation may not be the client in clinic. The affected individuals may all be diseased or known to the uninterested in genetic testing, also restricting the choices for the client. In some of these cases it may be appropriate to proceed with testing of an unaffected individual as the first person tested in a family, after informed consent is discussed. In other cases, especially is if the outcome of testing is unlikely to affect behavior or screening strategies, it is reasonable to discuss a plan for screening, prevention, and follow-up without specific genetic testing. A comprehensive clinic can serve patients and their families in all of these and many more scenarios as long as its role comprises tasks that extend far beyond those of a genetic testing station.
The increasing knowledge about specific mutations bring questions from carriers about hormone replacement use, chemoprevention strategies, and prophylactic surgery options that cannot all be thoroughly addressed in any other single clinical setting. The comprehensive cancer risk evaluation clinic thus operates as a clearing house for these risk-related concerns, as outlined in Table II. Specifically, it can serve to screen for eligibility for chemoprevention protocols, advisability and repercussions of prophylactic surgery strategies, and methods of prevention of the consequences of relative estrogen deficiency. These decisions are increasingly dependent on each other, and the ultimate plan a woman adopts may include several of these options in different combinations, depending on her cancer risk status, cancer history, age, and lifestyle preferences.
The advantage of counseling at-risk individuals in the setting of a comprehensive-service clinic is that the pros and cons of implementation of these varied strategies can be viewed within a single plan. A major task of the risk evaluation clinic in assisting at-risk individuals and their doctors in arriving at multi-dimensional, comprehensive plans is to help the parties factor into their decisions elements about which forthcoming data are expected to help clarify potential risks and benefits.
Therefore, postponing making a decision for an irreversible alternative (such as prophylactic surgery) is a reasonable course of action for some patients. Interim, short-term chemopreventive measures and life-long beneficial lifestyle measures can also be implemented in these situations, while deferring a more aggressive strategy. This is especially appropriate in very young premenopausal women who may be seeking short-term chempreventive strategies that may allow more accurate screening while preserving reproductive ability for the future.
Strategies for the early detection of cancer in higher-risk cohorts are emphasized in the risk evaluation clinic. Clinical breast examination performed by health care practitioners, breast self-examination performed by women, and mammography have all been shown to lead to early detection, and in the case of mammography screening, to improved outcomes from breast cancer. Because women with a genetic predisposition to breast and ovarian cancer are at risk for the development of these diseases at an earlier age than the general population, individualized screening strategies for these individuals are formulated in the risk evaluation clinic, taking into account the wealth of information gathered in the process. Mammograms in younger women are less sensitive because of increased breast density; however, novel protocols are being formulated to help improve mammographic sensitivity and specificity in younger women. One approach is to utilize a gonadotropin-releasing hormone agonist, low-dose estrogen, and testosterone in very high-risk premenopausal women to improve mammographic sensitivity. Other strategies are being planned to improve the detectability of early breast cancers in young mutation carriers.
Although yearly mammograms are typically initiated in high-risk women 5 years earlier than the youngest age of breast cancer diagnosis in the family, the efficacy of this strategy has not been proved. There are no accepted effective screening modalities for early detection of ovarian cancer. However, until better methods become available, CA-125 and transvaginal ultrasound with Doppler imaging are used to screen women over 40 years of age in high-risk families.
The legal and ethical issues surrounding genetic testing and risk information in general are important points of discussion in the risk evaluation clinic. The right to privacy and autonomy within the family and in society are discussed routinely. Specific issues affecting the potential for discrimination in health insurance coverage, the ability to acquire and maintain disability and life insurance, and the restriction of employment and adoption are placed in the context of the specific concerns and personal history of the patient. Both federal and state laws, as well as court cases that affect these issues, are brought into the discussion of different risk management scenarios if warranted.
The emerging ethical guidelines that help sort out the individual's right to privacy in light of other family members' right to know are incorporated in the decision making on genetic testing before the test is undertaken. The availability of confidential family history information are the cancer risk evaluation of interested relatives while protecting specific mutation information on the members already evaluated. In this way the clinic achieves its obligation to protect the confidentiality of a given individual while accommodating the need-to-know of another.
As more medical offices and centers use Internet-assisted databases, careful thought must be given to safeguarding patient-specific genetic testing information from being perused randomly when accessing the patient's records. Because many patients are concerned about the privacy of this information and their right to determine who has access to it, most risk evaluation clinics keep genetic testing results in separate files outside of the daily medical record. Individual patients may then determine who may have access to their risk evaluation records. Three principles apply, in our opinion, to the confidentiality policies of the risk evaluation clinic: (1) all patients should be aware of the confidentiality policies, and their advantages and disadvantages; (2) all patient files should be afforded the same level of confidentiality within the clinic or health system, regardless of the funding source for genetic testing; (3) all patients have the right to determine who has access to their genetic testing results, as permitted by law. These patients actively participate in decisions regarding safeguards of confidentiality and options for billing.
When referrals to the risk evaluation clinic are made from general practitioners or the patients themselves, depending on the insurance policy requirements, the initial high complexity evaluation is covered in most cases. If genetic testing is recommended by the specialized risk evaluation clinic, most insurance carriers in the sate of Michigan have covered the cost of this service and of the pot-test counseling, screening, and preventative recommendations. As these evidence-based recommendations are implemented, a significant degree of cooperation and support has be received from insurance carriers.
The psychological issues surrounding risk evaluation are challenging. The risk evaluation process is followed by a detailed documentation of the results of the evaluation, as well as the risk management plan, which is addressed to the patient with copies provided to the primary practitioners, per the patient's instructions. Because the interactions with the risk evaluation clinic are of limited extent, the continuing counseling of the individual about the psychological implications of the risk evaluation fall on the primary practitioner. Therefore, detailed information about the risk evaluation process and the recommendations is of crucial value to the practitioner in helping the patient implement the plan and cope with the repercussions of the risk information. In turn, physicians and other health care practitioners are called on to elicit referrals for extended counseling for patients who require long-term interventions. There is strong consensus among risk evaluation clinic leaders that effective and timely communication with the primary practitioners and the patients is crucial in helping ensure that the risk information benefits the patient. The availability and quality of such communication are thus important criteria that the health practitioner and the patient may use in selecting the services of a particular risk evaluation clinic.
MAJOR RESEARCH CHALLENGESResearch on the risk implications of specific mutations and on interactions of the high-penetrance genes such as BRCA1 and BRCA2 with other genetic determinants of risk will likely emanate from current and future studies conducted by consortia around the world. In the United States, the Cancer Genetics Network, which at present comprises of a handful of centers, is expected to expedite this research by making tissue and DNA materials available to researchers for specific projects. Other important consortia in North and South America, Europe, and Israel are anticipated to address population-specific features of inherited cancer risk. These mechanisms of integrated multi-institutional research are also likely to result in improved early detection strategies.
The applicability of the P1-NSABP study80 results to chemoprevention in BRCA carriers will be defined in the near future, as studies are underway to investigate the BRCA carrier status of the study participants. Innovative chempreventive strategies with molecular and radiographic end points are being designed for high-risk individuals. These clinical protocols will be accruing patients in the next few years and are expected to yield results that will help tailor the risk management plan of high-risk individuals over long periods of time.
Research into factors that affect the decision-making process in men and women of different ages and risk profiles is expected to result in innovative counseling strategies; these are likely to combine computer networking and interactive technology with human counseling to achieve improvement in risk management and patient and physician communication.
The ability to incorporate large numbers of genes and lifestyle factors into the accurate calculation of risk will result from our increasing ability to understand the clinical implications of human DNA sequence variability. Research is ongoing on the mathematic and informatic framework in which these calculations can be conceived. At present, we do not yet have the theoretical basis to undertake this comprehensive goal, but for clinical advances to derive from the Human Genome Project, it is required that appropriate theories and guidelines for information storage and management be developed. All of these advances are anticipated to improve the accuracy of cancer risk evaluation and benefit the patient by the implementation of tailored chemopreventative, screening, and lifestyle strategies.
Table I. Elements of clinical history information required for risk assessment
Table II. Elements of a comprehensive plan for breast and ovarian cancer risk management