Genetic Markers for Breast, Ovarian, and Prostatic Cancer

1. Field of the Invention

The field of the invention is genetic markers for inheritable breast cancer susceptibility.

2. Background

The largest proportion of inherited breast cancer described so far has been attributed to a genetic locus, the BRCA1 locus, on chromosome 17q21 (Hall et al. (1990) Science 250: 1684-1689; Narod et al. (1991) Lancet 338: 82-83; Easton et al. (1993) Am J Hum Genet 52: 678-701). Background material on the genetic markers for breast cancer screening is found in the Jan 29, 1993 issue of Science, vol 259, especially pages 622-625; see also King et al., (1993) J Amer Med Assoc 269: 1975-198. Other relevant research papers include King (1992) Nature Genet 2: 125-126; Merette et al. (1992) Amer J Human Genet 50: 515-519; NIH/CEPH Collaborative Mapping Group (1992) Science 258: 67-86.

Risks of breast cancer to women inheriting the locus are extremely high, exceeding 50% before age 50 and reaching 80% by age 65 (Newman et al. (1988) Proc Natl Acad Sci USA 85: 3044-3048; Hall et al. 1992 Amer J Human Genet 50: 1235-1242; Easton et al. (1993) supra. Epidemiological evidence for inherited susceptibility to ovarian cancer is even stronger (Cramer et at. (1983) J Natl Cancer Inst 71: 711-716; Schildkraut & Thompson (1988) Amer J Epidemiol 128: 456-466; Schildkraut et al. (1989) Amer J Hum Genet 45: 521-529). According to one study, more than 90% of families with multiple relatives with breast and ovarian cancer trace disease susceptibility to chromosome 17q21 (Easton et al. 1993 supra).

The link between increasing risk of breast and ovarian cancer and inherited susceptibility to these diseases lies in the application of genetics to diagnosis and prevention. Creating molecular tools for earlier diagnosis and developing ways to reverse the first steps of tumorigenesis may be the most effective means of breast and ovarian cancer control.

Our laboratory previously mapped the heritable breast cancer susceptibility gene locus (BRCA1 locus) to a 50 cM region of chromosome 17q (Hall et al. 1990, supra). More recently, we developed new polymorphisms at ERBB2 (Hall and King (1991) Nucl Acids Res 19: 2515), THRA1 (Bowcock et al. (1993) Amer J Human Genet 52: 718-722), EDH17B (Friedman et at. (1993) Hum Molec Genet 2: 821), and multiple anonymous loci (Anderson et al. (1993) Genomics 17: 616-623), ultimately developing a high density map of 17q12-q21 (Anderson et at. (1993 supra); see also, Simard et al. (1993) Human Molec Genet 2: 1193-1199). We also added families to the genetic study; there are now 100 families for whom transformed lymphocyte lines have been established and all informative relatives genotyped. We used our new markers and the many chromosome 17q polymorphisms developed in the past three years to test linkage in our families, refining the region first to 8 cM (Hall et al. (1992) supra, then to 4 cM (Bowcock et at. (1993 supra), then to 1 Mb based on polymorphisms from our high density map (Anderson et al. (1993supra); see also Flejter et al., (1993) Genomics 17: 624-631). We disclose here a number of mutations in BRCA1 which correlate with disease.

3. Relevant Literature

The predicted amino acid sequence for a BRCA1 cDNA and familial studies of this gene were described by Miki et al. (1994) Science 266, 66-71 and Futreal et al. (1994) Science 266, 120-122. A study of Canadian cancer families is described in Simard et al. (1994) Nature Genetics 8, 392-398. A collaborative survey of BRCA1 mutations is described in Shattuck-Eidens et al. (1995) JAMA 273, 535-541.

The invention discloses methods and compositions useful in the diagnosis and treatment of breast and ovarian cancer associated with mutations and/or rare alleles of BRCA1, a breast cancer susceptibility gene. Specific genetic probes diagnostic of inheritable breast cancer susceptibility and methods of use are provided. Labelled nucleic acid probes comprising sequences complementary to specified BRCA1 alleles are hybridized to clinical nucleic acid samples. Linkage analysis and inheritance patterns of the disclosed markers are used to diagnose genetic susceptibility. In addition, BRCA1 mutations and/or rare alleles are directly identified by hybridization, polymorphism and or sequence analysis. In another embodiment, labeled binding agents, such as antibodies, specific for peptides encoded by the subject nucleic acids are used to identify expression products of diagnostic mutations or alleles in patient derived fluid or tissue samples. For therapeutic intervention, the invention provides compositions which can functionally interfere with the transcription or translation products of the breast and ovarian cancer susceptibility associated mutations and/or rare alleles within BRCA1. Such products include anti-sense nucleic acids, competitive peptides encoded by the subject nucleic acids, and high affinity binding agents such as antibodies, specific for e.g. translation products of the disclosed BRCA1 mutations and alleles.

We disclose here methods and compositions for determining the presence or absence of BRCA1 mutations and rare alleles or translation products thereof which are useful in the diagnosis of breast and ovarian cancer susceptibility. Tumorigenic BRCA1 alleles include BRCA1 allele #5803 (SEQUENCE ID NO: 1), 9601 (SEQUENCE ID NO: 2), 9815 (SEQUENCE ID NO: 3), 8403 (SEQUENCE ID NO: 4), 8203 (SEQUENCE ID NO: 5), 388 (SEQUENCE ID NO: 6), 6401 (SEQUENCE ID NO: 7), 4406 (SEQUENCE ID NO: 8), 10201 (SEQUENCE ID NO: 9), 7408 (SEQUENCE ID NO: 10), 582 (SEQUENCE ID NO: 11) or 77 (SEQUENCE ID NO: 12). These nucleic acids or fragments capable of specifically hybridizing with the corresponding allele in the presence of other BRCA1 alleles under stringent conditions find broad diagnostic and therapeutic application. Gene products of the disclosed mutant and/or rare BRCA1 alleles also find a broad range of therapeutic and diagnostic applications. For example, mutant and/or rare allelic BRCA1 peptides are used to generate specific binding compounds. Binding reagents are used diagnostically to distinguish non-tumorigenic wild-type and tumorigenic BRCA1 translation products.

The subject nucleic acids (including fragments thereof) may be single or double stranded and are isolated, partially purified, and/or recombinant. An "isolated" nucleic acid is present as other than a naturally occurring chromosome or transcript in its natural state and isolated from (not joined in sequence to) at least one nucleotide with which it is normally associated on a natural chromosome; a partially pure nucleic acid constitutes at least about 10%, preferably at least about 30%, and more preferably at least about 90% by weight of total nucleic acid present in a given fraction; and a recombinant nucleic acid is joined in sequence to at least one nucleotide with which it is not normally associated on a natural chromosome.

Fragments of the disclosed alleles are sufficiently long for use as specific hybridization probes for detecting endogenous alleles, and particularly to distinguish the disclosed critical rare or mutant alleles which correlate with cancer susceptibility from other BRCA1 alleles, including alleles encoding the BRCA1 translation product displayed in Miki et al (1994) supra, under stringent conditions. Preferred fragments are capable of hybridizing to the corresponding mutant allele under stringency conditions characterized by a hybridization buffer comprising 0% formamide in 0.9M saline/0.09M sodium citrate (SSC) buffer at a temperature of 37.degree. C. and remaining bound when subject to washing at 42.degree. C. with the SSC buffer at 37.degree. C. More preferred fragments will hybridize in a hybridization buffer comprising 20% formamide in 0.9M saline/0.09M sodium citrate (SSC) buffer at a temperature of 42.degree. C. and remaining bound when subject to washing at 42.degree. C. with 2 X SSC buffer at 42.degree. C. In any event, the fragments are necessarily of length sufficient to be unique to the corresponding allele; i.e. has a nucleotide sequence at least long enough to define a novel oligonucleotide, usually at least about 14, 16, 18, 20, 22, or 24 bp in length, though such fragment may be joined in sequence to other nucleotides which may be nucleotides which naturally flank the fragment.

In many applications, the nucleic acids are labelled with directly or indirectly detectable signals or means for amplifying a detectable signal. Examples include radiolabels, luminescent (e.g. fluorescent) tags, components of amplified tags such antigen-labelled antibody, biotin-avidin combinations etc. The nucleic acids can be subject to purification, synthesis, modification, sequencing, recombination, incorporation into a variety of vectors, expression, transfection, administration or methods of use disclosed in standard manuals such as Molecular Cloning, A Laboratory Manual (2nd Ed., Sambrook, Fritsch and Maniatis, Cold Spring Harbor), Current Protocols in Molecular Biology (Eds. Ausubel, Brent, Kingston, More, Feidman, Smith and Stuhl, Greene Publ. Assoc., Wiley-Interscience, NY, N.Y., 1992) or that are otherwise known in the art.

The subject nucleic acids are used in a wide variety of nucleic acid-based diagnostic method that are known to those in the art. Exemplary methods include their use as allele-specific oligonucleotide probes (ASOs), in ligase mediated methods for detecting mutations, as primers in PCR-based methods, direct sequencing methods wherein the clinical BRCA1 nucleic acid sequence is compared with the disclosed mutations and rare alleles, etc. The subject nucleic acids are capable of detecting the presence of a critical mutant or rare BRCA1 allele in a sample and distinguishing the mutant or rare allele from other BRCA1 alleles. For example, where the subject nucleic acids are used as PCR primers or hybridization probes the subject primer or probe comprises an oligonucleotide complementary to a strand of the mutant or rare allele of length sufficient to selectively hybridize with the mutant or rare allele. Generally, these primers and probes comprise at least 16 bp to 24 bp complementary to the mutant or rare allele and may be as large as is convenient for the hybridizations conditions.

Where the critical mutation is a deletion of wild-type sequence, useful primers/probes require wild-type sequences flanking (both sides) the deletion with at least 2, usually at least 3, more usually at least 4, most usually at least 5 bases. Where the mutation is an insertion or substitution which exceeds about 20 bp, it is generally not necessary to include wild-type sequence in the probes/primers. For insertions or substitutions of fewer than 5 bp, preferred nucleic acid portions comprise and flank the substitution/insertion with at least 2, preferably at least 3, more preferably at least 4, most preferably at least 5 bases. For substitutions or insertions from about 5 to about 20 bp, it is usually necessary to include both the entire insertion/substitution and at least 2, usually at least 3, more usually at least 4, most usually at least 5 basis of wild-type sequence of at least one flank of the substitution/insertion.

In addition to their use as diagnostic genetic probes and primers, BRCA1 nucleic acids are used to effect a variety of gene-based therapies. See, e.g. Zhu et al. (1993) Science 261, 209-211; Gutierrez et al. (1992) Lancet 339, 715-721; Nabel et al. (1993), Proc. Nat'l. Acad Sci USA 90: 11307-11311. For example, therapeutic nucleic acids are used to modulate cellular expression or intracellular concentration or availability of a tumorigenic BRCA1 translation product by introducing into cells complements of the disclosed nucleic acids. These nucleic acids are typically antisense: single-stranded sequences comprising complements of the disclosed relevant BRCA1 mutant. Antisense modulation of the expression of a given mutant may employ antisense nucleic acids operably linked to gene regulatory sequences. Cell are transfected with a vector comprising such a sequence with a promoter sequence oriented such that transcription of the gene yields an antisense transcript capable of binding to the endogenous tumorigenic BRCA1 allele or transcript. Transcription of the antisense nucleic acid may be constitutive or inducible and the vector may provide for stable extrachromosomal maintenance or integration. Alternatively, single-stranded antisense nucleic acids that bind to BRCA1 genomic DNA or mRNA may be administered to the target cell, in or temporarily isolated from a host, at a concentration that results in a substantial reduction in expression of the targeted translation product.

Various techniques may be employed for introducing of the nucleic acids into viable cells. The techniques vary depending upon whether one is using the subject compositions in culture or in vivo in a host. Various techniques which have been found efficient include transfection with a retrovirus, viral coat protein-liposome mediated transfection, see Dzau et al., Trends in Biotech 11, 205-210 (1993). In some situations it is desirable to provide the nucleic acid source with an agent which targets the target cells, such as an antibody specific for a surface membrane protein on the target cell, a ligand for a receptor on the target cell, etc. Where liposomes are employed, proteins which bind to a surface membrane protein associated with endocytosis may be used for targeting and/or to facilitate uptake, e.g. capsid proteins or fragments thereof tropic for a particular cell type, antibodies for proteins which undergo internalization in cycling, proteins that target intracellular localization and enhance intracellular half-life. In liposomes, the decoy concentration in the lumen will generally be in the range of about 0.1 .mu.M to 20 .mu.M. For other techniques, the application rate is determined empirically, using conventional techniques to determine desired ranges. Usually, application of the subject therapeutics will be local, so as to be administered at the site of interest. Various techniques can be used for providing the subject compositions at the site of interest, such as injection, use of catheters, trocars, projectiles, pluronic gel, stents, sustained drug release polymers or other device which provides for internal access. Systemic administration of the nucleic acid using lipofection, liposomes with tissue targeting (e.g. antibody) may also be employed.

The invention also provides isolated translation products of the disclosed BRCA1 allele which distinguish the wild type BRCA1 gene product. For example, for alleles which encode truncated tumorigenic translation product, the C-terminus is used to differentiate wild-type BRCA1. Accordingly, the invention provides the translation product of BRCA1 allele #5803 (SEQUENCE ID NO: 13), 9601 (SEQUENCE ID NO: 14), 9815 (SEQUENCE ID NO: 15), 8203 (SEQUENCE ID NO: 17), 388 (SEQUENCE ID NO: 18), 6401 (SEQUENCE ID NO: 19), 4406 (SEQUENCE ID NO: 20), 10201 (SEQUENCE ID NO: 21), 7408 (SEQUENCE ID NO: 22), 582 (SEQUENCE ID NO: 23) or 77 (SEQUENCE ID NO: 24), or a C-terminus fragment thereof; and that of #8403 (SEQUENCE ID NO: 16), or a fragment thereof comprising Gly at position 61.

The subject mutant and/or rare allelic BRCA1 translation products comprise an amino acid sequence which provides a target for distinguishing the product from that of other BRCA1 alleles. Preferred fragments are capable of eliciting the production of a peptide-specific antibody, in vivo or in vitro, capable of distinguishing a protein comprising the immunogenic peptide from a wild-type BRCA1 translation product. The fragments are necessarily unique to the disclosed allele translation product in that it is not found in any previously known protein and has a length at least long enough to define a novel peptide, from about 5 to about 25 residues, preferably from 6 to 10 residues in length, depending on the particular amino acid sequence.

The subject translation products (including fragments) are either isolated, i.e. unaccompanied by at least some of the material with which they are associated in their natural state); partially purified, i.e. constituting at least about 1%, preferably at least about 10%, and more preferably at least about 50% by weight of the total translation product in a given sample; or pure, i.e. at least about 60%, preferably at least 80%, and more preferably at least about 90% by weight of total translation product. Included in the subject translation product weight are any atoms, molecules, groups, etc. covalently coupled to the subject translation products, such as detectable labels, glycosylations, phosphorylations, etc. The subject translation products may be isolated, purified, modified or joined to other compounds in a variety of ways known to those skilled in the art depending on what other components are present in the sample and to what, if anything, the translation product is covalently linked.

Binding agents specific for the disclosed tumorigenic BRCA1 genes and gene products find particular use in cancer diagnosis. The selected method of diagnosis will depend on the nature of the tumorigenic BRCA1 mutants/rare allele and its transcription or translation product(s). For example, soluble secreted translation products of the disclosed alleles may be detected in a variety of physiologic fluids using a binding agent with a detectable label such as a radiolabel, fluorester etc. Detection of membrane bound or intracellular products generally requires preliminary isolation of cells (e.g. blood cells) or tissue (e.g. breast biopsy tissue). A wide variety of specific binding assays, e.g. ELISA, may be used

BRCA1 gene product-specific binding agents are produced in a variety of ways using the compositions disclosed herein. For example, structural x-ray crystallographic and/or NMR data of the mutant and/or rare allelic BRCA1 translation products are used to rationally design binding molecules of determined structure or complementarity. Also, the disclosed mutant and/or rare allelic BRCA1 translation products are used as immunogens to generate specific polyclonal or monoclonal antibodies. See, Harlow and Lane (1988) Antibodies, A Laboratory Manual, Cold Spring Harbor Laboratory, for general methods. Specific antibodies are readily modified to a monovalent form, such as Fab, Fab', or Fv.

Other mutant and/or rare allelic BRCA1 gene-product specific agents are screened from large libraries of synthetic or natural compounds. For example, numerous means are available for random and directed synthesis of saccharide, peptide, and nucleic acid based compounds. Alternatively, libraries of natural compounds in the form of bacterial, fungal, plant and animal extracts are available or readily producible. Additionally, natural and synthetically produced libraries and compounds are readily modified through conventional chemical, physical, and biochemical means. See, e.g. Houghten et al. and Lam et al (1991) Nature 354, 84 and 81, respectively and Blake and Litzi-Davis (1992), Bioconjugate Chem 3, 510.

Useful binding agents are identified with assays employing a compound comprising mutant and/or rare allelic BRCA1 peptides or encoding nucleic acids. A wide variety of in vitro, cell-free binding assays, especially assays for specific binding to immobilized compounds comprising the subject nucleic acid or translation product find convenient use. See, e.g. Fodor et al (1991) Science 251, 767, for the light directed parallel synthesis method. Such assays are amenable to scale-up, high throughput usage suitable for volume drug screening.

Useful agents are typically those that bind the targeted mutant and/or rare allelic BRCA1 gene product with high affinity and specificity and distinguish the tumorigenic BRCA1 mutants/rare alleles from the wild-type BRCA1 gene product. Candidate agents comprise functional chemical groups necessary for structural interactions with proteins and/or DNA, and typically include at least an amine, carbonyl, hydroxyl or carboxyl group, preferably at least two of the functional chemical groups, more preferably at least three. The candidate agents often comprise cyclical carbon or heterocyclic structures and/or aromatic or polyaromatic structures substituted with one or more of the forementioned functional groups. Candidate agents are also found among biomolecules including peptides, saccharides, fatty acids, sterols, isoprenoids, purines, pyrimidines, derivatives, structural analogs or combinations thereof, and the like. Where the agent is or is encoded by a transfected nucleic acid, said nucleic acid is typically DNA or RNA.

Candidate agents are obtained from a wide variety of sources including libraries of synthetic or natural compounds. For example, numerous means are available for random and directed synthesis of a wide variety of organic compounds and biomolecules, including expression of randomized oligonucleotides. Alternatively, libraries of natural compounds in the form of bacterial, fungal, plant and animal extracts are available or readily produced. Additionally, natural and synthetically produced libraries and compounds are readily modified through conventional chemical, physical, and biochemical means to enhance efficacy, stability, pharmaceutical compatibility, and the like. In addition, known pharmacological agents may be subject to directed or random chemical modifications, such as acylation, alkylation, esterification, amidification, etc., to produce structural analogs.

Therapeutic applications typically involve binding to and functional disruption of a tumorigenic BRCA1 gene product by an administered high affinity binding agent. For therapeutic uses, the compositions and agents disclosed herein may be administered by any convenient way. Small organics are preferably administered orally; other compositions and agents are preferably administered parenterally, conveniently in a pharmaceutically or physiologically acceptable carrier, e.g., phosphate buffered saline, or the like. Typically, the compositions are added to a retained physiological fluid such as blood or synovial fluid. Generally, the amount administered will be empirically determined, typically in the range of about 10 to 1000 .mu.g/kg of the recipient. For peptide agents, the concentration will generally be in the range of about 50 to 500 .mu.g/ml in the dose administered. Other additives may be included, such as stabilizers, bactericides, etc. These additives will be present in conventional amounts.

The following examples are offered by way of illustration and not by way of limitation.

Contig Construction

YACs.

Primers flanking polymorphic repeats in the 4 Mb region of linkage were used to amplify pools from the CEPH, Washington University, and CEPH megaYAC libraries available. 39 YACs were selected. Of these, 23 were tested for chimerism by FISH and 12 found to be chimeric. YACs were aligned to each other by attempting to amplify each YAC with primer pairs from known sequence tagged sites (STSes). More STSes were defined by sequencing the ends of YACs, and these new STSes used for further alignment and YAC identification.

Cosmids.

A gridded cosmid library of chromosome 17 was prepared. Alu-Alu PCR products of YACs were hybridized to the cosmid grids and positively hybridizing cosmids used for subsequent studies. Contigs were constructed in two ways. Cosmids with the same restriction patterns were aligned; and, the unique sequences flanking polymorphic markers and our sequenced cDNAs were used as STSes.

Physical Mapping by Pulsed Field Gel Electrophoresis.

Physical distances were estimated by pulsed field gel electrophoresis, using DNA from lymphocyte cell lines of BRCA1-linked patients and of controls. DNA samples were digested with NotI, MluI, RsrII, NruI, SacII, and EclXI. Filters were probed with single-copy sequences isolated from cosmids and later with eDNA clones. Multiple unrelated linked patients and controls were screened to detect large insertions or deletions associated with BRCA1. Results of PFGE were used to define the region first used to screen eDNA libraries as.about.1 Mb and the current linked region as.ltoreq.500 kb.

Screening cDNA Libraries.

We began library screening when the linked region defined by meiotic recombination was.about.1 Mb. The first question was what library would optimize the length of cDNA clones, representation of both 5' and 3' ends of genes, and the chances that BRCA1 would be expressed. We chose to use a random primed cDNA library cloned into 1gt10 from cultured (not transformed) fibroblasts from a human female. This library was selected because it had inserts averaging 1.8 kb, with 80% of inserts between 1 and 4 kb, was contructed from cultured fibroblasts known to be "leaky" in gene expression, and was known to include 5' ends of genes. We simultaneously screened three other libraries (from ovary, fetal brain, and mouse mammary epithelium). With one exception (described below), all transcripts from these libraries cross-hybridized to transcripts from the fibroblast library.

The fibroblast library was screened with YAC DNA isolated by PFGE. Pure YAC DNA (100 nanograms) was random primed with both aP32-dATP (6000 mCi/mmole) and .sup.32 P-dCTP (3000 mCi/mmole), and used immediately after labelling. Filters from the library were prehybridized with human placental DNA for 24-48 hours. Labelled YAC DNA was hybridized to the filters for 48 hours at 65.degree. C. Approximately 250 transcripts were selected by screening with 7 YACs and then ross-hybridized. We also used pools of cosmids from the linked region to screen the fibroblast library. We selected 122 transcripts and cross-hybridized them to clones previously detected by the YACs.

A. Screening for Mutations in Candidate Genes.

We initially identified 24 genes in the 1 Mb BRCA1 region defined by meiotic recombination, respective locations on the YAC contig, sizes of representative cDNA clones, numbers of replicates in the library, sizes of transcripts, homologies to known genes, and variants detected. Candidate gene were characterized in the following ways:

(1) Cross-Hybridizing Clones.

cDNA clones isolated from the library are hybridized against each other. Cross-hybridizing clones are considered "siblings" of the clone used as a probe and represent the same gene.

(2) Mapping Back.

At least one clone from each sibship is mapped back to total human genomic DNA, to cosmids, to YACs, and to somatic cell hybrid lines, some of which contain deletions of 17q and one of which has chromosome 17 as its only human chromosome.

(3) Subcloning and Sequencing.

One of the longest clones from each sibship is subcloned into M13 and sequenced manually by standard methods, constructing new primers at the end of each fragment to continue sequencing until the end of the clone is reached.

(4) Extending Sequences with Sibs.

In order to find clones that contain more of the gene, the last sequencing primer for the clone and primers made from gt10 are used to amplify sibs of the first clone. Sibs that amplify the longest fragments are selected, subcloned, and sequenced. This process is continued until we reach the size of the transcript defined by Northern blot and/or until the 3' sequence is a polyA tail and the 5' sequence has features of the beginning of the coding region.

(5) Southerns.

To identify insertion or deletion mutations, genomic DNA from 20 unrelated patients from families with breast cancer linked to 17q (i.e. "linked patients") and controls are digested with BamHI/TaqI and independently with HindIII/HinfI. Each cDNA clone is used to screen Southern blots. Variants have been detected in two genes. Both of these variants are RFLPs, occuring in equal frequency in linked patients and in controls.

(6) Northerns.

To identify splice mutations and/or length mutations, we prepared total RNA and polyA+ RNA from germline DNA (from lymphoblast lines) of 20 unrelated linked patients, from ovarian and breast tissues, from fibroblasts, from a HeLa cell line, and from breast cancer cell lines. Northern blots are screened with each gene.

(7) Detection of Small Mutations.

To screen for germline point mutations in patients without encountering introns, we prepared cDNA from poly-A+ mRNA from lymphoblast cell lines of 20 unrelated linked patients and from controls. cDNA has also been made from 65 malignant ovarian cancers from patients not selected for family history. Primers are constructed every.about.200 basepairs along the sequence and used to amplify these cDNAs. Genomic DNA has also been prepared from cell lines from all family members (linked and unlinked), from malignant and normal cells from paraffin blocks from their breast and ovarian surgeries, and from malignant and normal cells from 29 breast tumors not selected for family history. For sequences without introns, cDNA and gDNA lengths are equal, and the gDNA samples are amplified as well.

Two mutation detection methods are used to screen each sequence. Amplified products are screened for SSCPs using modifications that enable electrophoresis to be done with only one set of running conditions (Keen et at. (1991) Trends Genet 7: 5; Soto and Sukumar (1992) PCR Meth Appl 2: 96-98). In order to screen longer segments of DNA (100-1500 bp) and to detect variants missed by SSCP, sequences are also screened for point mutations by CCM (Cotton (1993) Mutation Res 285: 125-144) using essentially the protocol of Grompe et al. (1989) Proc Natl Acad Sci USA 86: 5888-5892. An endonuclease developed for mismatch detection reduces the toxicity of the method (Youil et al. (1993) Amer J Hum Genet 53 (supplement): abstract 1257).

(8) Polymorphism or Mutation.

Variants are screened in cases and controls to distinguish polymorphisms from a critical mutation. Linkage of breast cancer to each variant is tested in all informative families.

A. BRCA1 Mutations in Chromosome 17q-Linked Families.

Our series of families includes 20 large extended kindreds in which breast and ovarian cancer (and in one family prostatic cancer) are linked to 17q21, with individual lod scores >1.5. Since linked patients in these families carry mutations in BRCA1, we have identified their mutations first.

B. Germline BRCA Mutations Among Breast Cancer Patients in the General Population.

From each breast cancer patient, not selected for family history, a 30 ml sample of whole blood is drawn into acid citrate dextrose. DNA from the blood is extracted and stored at -70.degree. C. in 3 aliquots. Germline mutations in BRCA1 are identified using the approaches described above and by directly sequencing new mutations. Paraffin-embedded tumor specimens from the same patients are screened for alterations of p53, HER2, PRAD 1, and ER. Germline BRCA1 mutations are tested in the tumor blocks.

A preliminary estimate of risk associated with different BRCA1 mutations is obtained from relatives of patients with germline alterations. For each patient with a germline BRCA1 mutation, each surviving sister and mother (and for older patients, brothers as well), DNA is extracted from a blood sample and tested for the presence of the proband's BRCA1 mutation. To ascertain men at risk for prostatic cancer, brothers of breast cancer patients diagnosed after age 55 are also interviewed and sampled. Paraffin blocks from deceased relatives who had cancer are also screened. The frequency of breast, ovarian, or prostatic cancer among relatives carrying BRCA1 mutations is a first estimate of risk of these cancers associated with different mutations.

C. Somatic Alterations of BRCA1 in Breast Tumors.

Malignant cells are dissected from normal cells from paraffin blocks. By identifying BRCA1 mutations in these series, we estimate the frequency of somatic BRCA1 alterations, determine BRCA1 mutations characteristic of any particular stage of tumor development, and evaluate their association with prognosis.

D. Characterizing Mutant and Rare Alleles of BRCA1.

Mutant or rare BRCA1 allele function and pattern of expression during development are characterized using transformed cells expressing the allele and knockout or transgenic mice. For example, phenotypic changes in the animal or cell line, such as growth rate and anchorage independence are determined. In addition, several methods are used to study loss-of-function mutations, including replacing normal genes with their mutant alleles (BRCA1-/BRCA1-) by homologous recombination in embryonic stem (ES) cells and replacing mutant alleles with their normal counterparts in differentiated cultured cells (Capecchi (1989) Science 244: 1288-1292; Weissman et al. (1987) Science 236: 175-180; Wang et al. (1993) Oncogene 8: 279-288). Breast carcinoma cell lines are screened for mutation at the BRCA1 locus and a mutant BRCA1 line is selected. Normal and mutant cDNAs of BRCA1 are subcloned into an expression vector carrying genes which confer resistance to ampicillin and geneticin (Baker et al. (1990)Nature 249: 912-915). Subclones are transfected into mutant BRCA1 breast cancer cells Geneticin-resistant colonies are isolated and examined for any change in tumorigenic phenotype, such as colony formation in soft agar, increased growth rate, and/or tumor formation in athymic nude mice. In vivo functional demonstrations involve introducing the normal BCRA1 gene into a breast carcinoma cell line mutant at BRCA1 and injecting these BRCA1+ cells into nude mice. Changes observed in tumorigenic growth compared to nude mice injected with BRCA1 mutant breast carcinoma cells are readily observed. For example, correcting the mutant gene decreases the ability of the breast carcinoma cells to form tumors in nude mice (Weissman et al. (1987) Science 236: 175-8, Wang et al. (1993) Oncogene 8: 279-88).

All publications and patent applications cited in this specification are herein incorporated by reference as if each individual publication or patent application were specifically and individually indicated to be incorporated by reference. Although the foregoing invention has been described in some detail by way of illustration and example for purposes of clarity of understanding, it will be readily apparent to those of ordinary skill in the art in light of the teachings of this invention that certain changes and modifications may be made thereto without departing from the spirit or scope of the appended claims.