[Show abstract][Hide abstract] ABSTRACT: The founder mutations in BRCA (BRCA1*185delAG, BRCA1*5382insC, and BRCA2*6174delT) account for 95% of the detectable BRCA mutations in breast and ovarian cancer families of Ashkenazi Jewish ancestry. Optimal clinical management of individuals from these high-risk families relies on the identification of BRCA founder mutations in the laboratory. We have therefore developed a rapid and reliable approach using pyrosequencing, which allows for the detection of these frequent frameshift mutations on different types of specimens. We were able to correctly identify all mutants in a blinded analysis of 177 DNA samples, including 120 DNA samples extracted from paraffin tissues, 30 samples derived from blood specimens, and 27 samples derived from saliva. The mutation detection rate of pyrosequencing was 100% for all of the DNA samples tested with neither false-positive nor false-negative results. The assay also demonstrated both high accuracy and high precision for the detection of these common mutations in paraffin tissues. Furthermore, saliva collection is a noninvasive alternative for DNA isolation in both clinical and research settings. We show that pyrosequencing is a rapid and reliable method that serves as an excellent platform for BRCA founder mutation analysis, especially when only paraffin-embedded tissues are available.
Preview · Article · Apr 2009 · The Journal of molecular diagnostics: JMD
[Show abstract][Hide abstract] ABSTRACT: For individuals genetically predisposed to breast and ovarian cancer through inheritance of a mutant BRCA allele, somatic loss of heterozygosity affecting the wild-type allele is considered obligatory for cancer initiation and/or progression. However, several lines of evidence suggest that phenotypic effects may result from BRCA haploinsufficiency.
Archival fixed and embedded tissue specimens from women with germ line deleterious mutations in BRCA1 or BRCA2 were identified. After pathologic review, focal areas of normal breast epithelium, atypical ductal hyperplasia, ductal carcinoma-in-situ, and invasive ductal carcinoma were identified from 14 BRCA1-linked and 9 BRCA2-linked breast cancers. Ten BRCA-linked prophylactic mastectomy specimens and 12 BRCA-linked invasive ovarian carcinomas were also studied. Laser catapult microdissection was used to isolate cells from the various pathologic lesions and corresponding normal tissues. After DNA isolation, real-time polymerase chain reaction assays were used to quantitate the proportion of wild-type to mutant BRCA alleles in each tissue sample.
Quantitative allelotyping of microdissected cells revealed a high level of heterogeneity in loss of heterozygosity within and between preinvasive lesions and invasive cancers from BRCA1 and BRCA2 heterozygotes with breast cancer. In contrast, all BRCA-associated ovarian cancers displayed complete loss of the wild-type BRCA allele.
These data suggest that loss of the wild-type BRCA allele is not required for BRCA-linked breast tumorigenesis, which would have important implications for the genetic mechanism of BRCA tumor suppression and for the clinical management of this patient population.
No preview · Article · Oct 2007 · Annals of Surgical Oncology
[Show abstract][Hide abstract] ABSTRACT: To characterize the role of BRCA1 in mammary gland development and tumor suppression, a transgenic mouse model of BRCA1 overexpression was developed. Using the mouse mammary tumor virus (MMTV) promoter/enhancer, transgenic mice expressing human BRCA1 or select mutant controls were generated. Transgenic animals examined during adolescence were shown to express the human transgene in their mammary glands. The mammary glands of 13-week-old virgin homozygous MMTV-BRCA1 mice presented the morphology of moderately increased lobulo-alveolar development. The mammary ductal trees of both hemizygous and homozygous MMTV-BRCA1t340 were similar to those of control non-transgenic littermates. Interestingly, both hemi- and homozygous mice expressing a splice variant of BRCA1 lacking the N-terminal RING finger domain (MMTV-BRCA1sv) exhibited marked mammary lobulo-alveolar development, particularly terminal end bud proliferation. Morphometric analyses of mammary gland whole mount preparations were used to measure epithelial staining indices of ~35% for homozygous MMTV-BRCA1 mice and ~60% for both hemizygous and homozygous MMTV-BRCA1sv mice versus ~25% for non-transgenic mice. Homozygous MMTV-BRCA1 mice showed delayed development of tumors when challenged with 7,12 dimethylbenzanthracene (DMBA), relative to non-transgenic and homozygous BRCA1t340 expressing mice. In contrast, homozygous MMTV-BRCA1sv transgenic animals were sensitized to DMBA treatment and exhibited a very rapid onset of mammary tumor development and accelerated mortality. MMTV-BRCA1 effects on mortality were restricted to DMBA-induced tumors of the mammary gland. These results demonstrate in vivo roles for BRCA1 in both mammary gland development and in tumor suppression against mutagen-induced mammary gland neoplasia.
Full-text · Article · Feb 2007 · International journal of biological sciences
[Show abstract][Hide abstract] ABSTRACT: A major limitation in counseling unaffected women from families with inherited breast and ovarian cancer is that a "true-negative" interpretation of wild type BRCA analysis of the proband cannot be inferred in the absence of demonstration of a BRCA mutation segregating in the kindred. Documentation of familial BRCA mutations from paraffin-derived DNA of deceased patients has been limited due to reports of technical complications leading to lack of reproducibility of BRCA testing of archival material.
DNA was extracted from formalin-fixed paraffin-embedded (FFPE) morphologically normal tissue of 161 blinded, coded samples from women previously genotyped for the three Ashkenazi Jewish BRCA founder mutations from lymphocyte-derived DNA. Multiplex PCR followed by denaturing polyacrylamide gel electrophoresis was performed for the three founder mutations to determine if analysis on FFPE tissue could produce results concordant with those of the lymphocyte-derived DNA.
After disclosure of the sample codes, the results were compared with the original lymphocyte-derived DNA genotypes. Excluding one sample unevaluable due to PCR failure, there was 100% concordance of 160 genotypes (120 mutation samples) derived from DNA from archival FFPE tissue compared to peripheral lymphocytes.
The method described reliably detected BRCA founder mutations in archival DNA derived from FFPE tissue. These results suggests that this technique may be useful in clinical settings to inform wild type BRCA results of unaffected probands, leading to avoidance of unnecessary intensified surveillance or risk-reducing surgery. With further validation this approach can also be applied to other populations where founder mutations are observed.
[Show abstract][Hide abstract] ABSTRACT: The goal of this study was to determine whether distinct gene expression profiles are associated with intrinsic and/or acquired chemoresistance in epithelial ovarian carcinoma.
Gene expression profiles were generated from 21 primary chemosensitive tumors and 24 primary chemo-resistant tumors using cDNA-based microarrays. Gene expression profiles of both groups of primary tumors were then compared with those of 15 ovarian carcinomas obtained following platinum-based chemotherapy ("post-chemotherapy" tumors). A theme discovery tool was used to identify functional categories of genes involved in drug resistance.
Comparison of primary chemosensitive and chemo-resistant tumors revealed differential expression of 85 genes (P < 0.001). Comparison of gene expression profiles of primary chemosensitive tumors and post-chemotherapy tumors revealed more robust differences with 760 genes differentiating the two groups (P < 0.001). In contrast, only 230 genes were differentially expressed between primary chemo-resistant and post-chemotherapy groups (P < 0.001). Common to both gene lists were 178 genes representing transcripts differentially expressed between post-chemotherapy tumors and all primary tumors irrespective of intrinsic chemosensitivity. The gene expression profile of post-chemotherapy tumors compared with that of primary tumors revealed statistically significant overrepresentation of genes encoding extracellular matrix-related proteins.
These data show that gene expression profiling can discriminate primary chemo-resistant from primary chemosensitive ovarian cancers. Gene expression profiles were also identified that correlate with states of intrinsic and acquired chemoresistance and that represent targets for future investigation and potential therapeutic interventions.
Full-text · Article · Oct 2005 · Clinical Cancer Research
[Show abstract][Hide abstract] ABSTRACT: To identify germ line CDH1 mutations in hereditary diffuse gastric cancer (HDGC) families and develop guidelines for management of at risk individuals.
We ascertained 31 HDGC previously unreported families, including 10 isolated early-onset diffuse gastric cancer (DGC) cases. Screening for CDH1 germ line mutations was done by denaturing high-performance liquid chromatography and automated DNA sequencing.
We identified eight inactivating and one missense CDH1 germ line mutation. The missense mutation conferred in vitro loss of protein function. Two families had the previously described 1003C>T nonsense mutation. Haplotype analysis revealed this to be a recurrent and not a founder mutation. Thirty-six percent (5 of 14) of the families with a documented DGC diagnosed before the age of 50 and other cases of gastric cancer carried CDH1 germ line mutations. Two of 10 isolated cases of DGC in individuals ages <35 years harbored CDH1 germ line mutations. One mutation positive family was ascertained through a family history of lobular breast cancer (LBC) and another through an individual with both DGC and LBC. Occult DGC was identified in five of six prophylactic gastrectomies done on asymptomatic, endoscopically negative 1003C>T mutation carriers.
In addition to families with a strong history of early-onset DGC, CDH1 mutation screening should be offered to isolated cases of DGC in individuals ages <35 years and for families with multiple cases of LBC, with any history of DGC or unspecified GI malignancies. Prophylactic gastrectomy is potentially a lifesaving procedure and clinical breast screening is recommended for asymptomatic mutation carriers.
Preview · Article · Sep 2005 · Clinical Cancer Research
[Show abstract][Hide abstract] ABSTRACT: Activation of the phosphatidylinositol 3-kinase (PI3K)-AKT pathway, resulting in increased cell proliferation, survival, and motility, is believed to play an oncogenic role in many cancer types. The PIK3CA gene encodes the p110alpha catalytic subunit of PI3K, and is amplified in some ovarian cancers, whereas the AKT2 gene is amplified in some ovarian, breast, and pancreatic cancers. Recently, in a mutational screen of eight PI3K genes and eight PI3K-like genes, PIK3CA was found to be the only gene affected by somatic mutations, which were observed frequently in gastrointestinal and brain cancers. Here, we test whether PIK3CA is subject to mutation in ovarian and breast cancers.
Exons 9 and 20, encoding the highly conserved helical and kinase domains of PIK3CA, were subjected to sequence analysis in 198 advanced stage epithelial ovarian carcinomas and 72 invasive breast carcinomas (48 of ductal histology and 24 of lobular histology).
Somatic missense mutations were observed in 24 of 198 (12%) ovarian carcinomas, and in 13 of 72 (18%) breast carcinomas.
These data indicate that mutations of PIK3CA play an oncogenic role in substantial fractions of ovarian and breast carcinomas, and in consideration of mutation of other components of the PI3K-AKT pathway in both tumor types, confirm the major oncogenic role of this pathway in ovarian and breast carcinomas.
Preview · Article · May 2005 · Clinical Cancer Research
[Show abstract][Hide abstract] ABSTRACT: Recently BRCA1 has been implicated in the regulation of gene expression from the X chromosome. In this study the influence of BRCA1 on expression of X chromosome genes was investigated. Complementary DNA microarrays were used to compare the expression levels of X chromosome genes in 18 BRCA1-associated ovarian cancers to those of the 13 "BRCA1-like" and 14 "BRCA2-like" sporadic tumors (as defined by previously reported expression profiling). Significance was determined using parametric statistics with P < 0.005 as a cutoff. Forty of 178 total X-chromosome transcripts were differentially expressed between the BRCA1-associated tumors and sporadic cancers with a BRCA2-like molecular profile. Thirty of these 40 genes showed higher mean expression in the BRCA1-associated samples including all 11 transcripts that mapped to Xp11. In contrast, four of 178 total X chromosome transcripts showed significant differential expression between BRCA1-associated and sporadic tumors with a BRCA1-like molecular profile. All four mapped to Xp11 and showed higher mean expression in BRCA1-associated tumors. Re-expression of BRCA1 in HCC1937 BRCA1-deficient breast cancer cell resulted in the repression of 21 transcripts. Eleven of the 21 (54.5%) transcripts mapped to Xp11. However, there was no significant overlap between these Xp11 genes and those found to be differentially expressed between BRCA1-associated and sporadic ovarian cancer samples. These results demonstrate that BRCA1 mediates the repression of several X chromosome genes, many of which map to the Xp11 locus.
Full-text · Article · Oct 2004 · Journal of Translational Medicine
[Show abstract][Hide abstract] ABSTRACT: The aims of this study were to determine the incidence of BRCA mutations among Ashkenazi Jewish patients with fallopian tube carcinoma (FTC) or primary peritoneal carcinoma (PPC), to study the clinicopathologic features of these cancers, and to estimate the risks of these cancers in association with a BRCA mutation.
A retrospective review at two institutions identified 29 Jewish patients with FTC and 22 Jewish patients with PPC. These patients were genotyped for the three Ashkenazi Jewish BRCA founder mutations (185delAG and 5382insC in BRCA1 and 6174delT in BRCA2). Surgical and pathologic information, family history, and survival data were obtained from hospital records. All statistical tests were two sided.
Germline BRCA mutations were identified in five of 29 patients with FTC (17%) and nine of 22 patients with PPC (41%). Mutation carriers had a younger mean age at diagnosis than patients without a mutation (60 v 70 years; P =.002). The overall median survival was 148 months for mutation carriers and 41 months for patients without a mutation (P =.04). For BRCA mutation carriers, the lifetime risks of FTC and PPC were 0.6% and 1.3%, respectively.
Substantial proportions of Ashkenazi Jewish patients with FTC or PPC are BRCA mutation carriers. Patients with BRCA-associated FTC or PPC are younger at diagnosis and have improved survival compared with patients without a BRCA mutation. Although the lifetime risks of FTC or PPC for patients with BRCA heterozygotes are greater than those for the general population, the absolute risks seem relatively low.
No preview · Article · Dec 2003 · Journal of Clinical Oncology
[Show abstract][Hide abstract] ABSTRACT: Germline mutations in BRCA1 and BRCA2 are responsible for 5%-10% of epithelial ovarian cancers, but the molecular pathways affected by these mutations are unknown. We used complementary DNA (cDNA) microarrays to compare gene expression patterns in ovarian cancers associated with BRCA1 or BRCA2 mutations with gene expression patterns in sporadic epithelial ovarian cancers and to identify patterns common to both hereditary and sporadic tumors.
Tumor samples from 61 patients with pathologically confirmed epithelial ovarian adenocarcinoma with matched clinicopathologic features were studied, including 18 with BRCA1 founder mutations, 16 with BRCA2 founder mutations, and 27 without either founder mutation (termed sporadic cancers). The cDNA microarrays contained 7651 sequence-verified features. Gene expression data were analyzed with a modified two-sided F test, with P<.0001 considered statistically significant. The expression level of six genes was also studied with reverse transcription-polymerase chain reaction.
The greatest contrast in gene expression was observed between tumors with BRCA1 mutations and those with BRCA2 mutations; 110 genes showed statistically significantly different expression levels (P<.0001). This group of genes could segregate sporadic tumors into two subgroups, "BRCA1-like" and "BRCA2-like," suggesting that BRCA1-related and BRCA2-related pathways are also involved in sporadic ovarian cancers. Fifty-three genes were differentially expressed between tumors with BRCA1 mutations and sporadic tumors; six of the 53 mapped to Xp11.23 and were expressed at higher levels in tumors with BRCA1 mutations than in sporadic tumors. Compared with the immortalized ovarian surface epithelial cells used as reference, several interferon-inducible genes were overexpressed in the majority of tumors with a BRCA mutation and in sporadic tumors.
Mutations in BRCA1 and BRCA2 may lead to carcinogenesis through distinct molecular pathways that also appear to be involved in sporadic cancers. Sporadic carcinogenic pathways may result from epigenetic aberrations of BRCA1 and BRCA2 or their downstream effectors.
Full-text · Article · Jul 2002 · JNCI Journal of the National Cancer Institute
[Show abstract][Hide abstract] ABSTRACT: Recombinant adenoviruses expressing human BRCA1 (AdBRCA1), murine Brca1 (AdBrca1), three clinically relevant human mutant BRCA1 proteins (t340, C61G, and 1853Stop), or a murine Brca1 C-terminal deletion mutant were constructed and evaluated in vitro. These recombinants were capable of transducing high-level transgene expression to a wide variety of cell lines in vitro. Three independent methods were utilized to monitor cell growth following transduction with these recombinants. High-level expression of either the human or mouse wild-type BRCA1 protein was incompatible with maximal levels of cell growth. AdBRCA1 transduction inhibited the outgrowth of several human breast and ovarian cell lines in colony formation assays. Flow cytometric analysis revealed an accumulation of the transduced cells in the G0/G1 phase of the cell cycle. This BRCA1-mediated accumulation of cells in G0/G1 was accompanied by an increase in the cellular level of hypophosphorylated pRB. Ad mutant BRCA1 t340, C61G, and 1853Stop viruses were impaired, to varying degrees, in their ability to transduce a growth-arrested state to the target cells. Using these same three criteria, overexpression of murine Brca1 by AdBrca1 was also capable of transducing a growth-arrested state to human cells. Deletion of the C-terminus of Brca1 diminished this activity. This panel of adenoviruses may be useful reagents as part of an approach to understand the function of BRCA1/Brca1 in normal breast and ovary and help to define the tumor suppressor defect (s) conferred by clinical BRCA1 mutations in breast and ovarian cell tumorigenesis.
Preview · Article · Apr 2001 · Cancer Gene Therapy
[Show abstract][Hide abstract] ABSTRACT: Genetically based differences in carcinogen metabolism have been related to polymorphisms of the cytochrome P450IA1 gene (CYPIA1) and the null genotypes of glutathione S-transferase classes mu and theta (GSTM1 and GSTT1). By PCR we examined the genotypes of CYPIA1, GSTM1, and GSTT1 in relation to breast cancer risk in Caucasian and African-American women. The study included 164 Caucasian and 59 African-American women with primary invasive breast cancer and age-matched female controls. Enzyme polymorphisms included in this study were the null deletions of GSTM1 and GSTT1 and the m1 (MspI), m2 (codon 462: isoleucine-->valine), m3 (MspI-AA), and m4 (codon 461: threonine-->asparagine) polymorphisms of CYPIA1. Contrary to previous reports by other investigators, none of the enzyme genotypes, individually or combined, appear to associate with an increased risk for breast cancer in Caucasian or African-American women. We also report that the recently described m4 allele occurs at a lower frequency in African-Americans than Caucasians and is not linked with breast cancer in either race. Thus, it is unlikely that polymorphisms of GSTM1, GSTT1, or CYPIA1 represent susceptibility factors for breast cancer in Caucasians or African-Americans.
[Show abstract][Hide abstract] ABSTRACT: The estrogen receptor (ER) belongs to a family of ligand-inducible nuclear receptors that exert their effects by binding to cis-acting DNA elements in the regulatory region of target genes. The detailed mechanisms by which ER interacts with the estrogen response element (ERE) and affects transcription still remain to be elucidated. To study the ER-ERE interaction and transcription initiation, we employed purified recombinant ER expressed in both the baculovirus-Sf9 and his-tagged bacterial systems. The effect of high-mobility group (HMG) protein HMG-1 and purified recombinant TATA-binding protein-associated factor TAF(II)30 on ER-ERE binding and transcription initiation were assessed by electrophoretic mobility shift assay and in vitro transcription from an ERE-containing template (pERE2LovTATA), respectively. We find that purified, recombinant ER fails to bind to ERE in spite of high ligand-binding activity and electrophoretic and immunological properties identical to ER in MCF-7 breast cancer cells. HMG-1 interacts with ER and promotes ER-ERE binding in a concentration- and time-dependent manner. The effectiveness of HMG-1 to stimulate ER-ERE binding in the electrophoretic mobility shift assay depends on the sequence flanking the ERE consensus as well as the position of the latter in the oligonucleotide. We find that TAF(II)30 has no effect on ER-ERE binding either alone or in combination with ER and HMG-1. Although HMG-1 promotes ER-ERE binding, it fails to stimulate transcription initiation either in the presence or absence of hormone. In contrast, TAF(II)30, while not affecting ER-ERE binding, stimulates transcription initiation 20-fold in the presence of HMG-1. These results indicate that HMG-1 and TAF(II)30 act in sequence, the former acting to promote ER-ERE binding followed by the latter to stimulate transcription initiation.
Full-text · Article · Aug 1997 · Molecular Endocrinology
[Show abstract][Hide abstract] ABSTRACT: The estrogen receptor (ER) content of breast cancers predicts the likelihood of benefit from antiestrogen therapy. Tumors with an ER content of less than 10 fmol/mg are considered to be ER-negative and have a poor response rate to tamoxifen. The molecular basis of the ER-negative phenotype is not well understood. This study was conducted in five ER-positive (MCF-7, T47D, ZR-75-1, MDA-MB-134, MDA-MB-361) and five ER-negative (BT-20, HBL-100, MDA-MB-157, MDA-MB-231, MDA-MB-468) human breast cancer cell lines to determine whether the ER-negative phenotype results from mutations in the coding region of the ER gene or is due to a deficiency of the transcriptional or post-transcriptional regulation of ER expression. The coding region was examined for mutations by denaturing gradient gel electrophoresis, single-strand conformation polymorphism and DNA sequence analysis. The presence and integrity of ER mRNA were investigated by Northern blotting and reverse transcription (RT)-PCR amplification. The presence and concentration of ER protein were assessed by Western blotting and hormone-binding assay. Both ER-negative and -positive cell lines contained silent mutations in codons 10, 325 and 594 of the ER gene, all of which probably represent neutral polymorphic sites. No missense or nonsense mutations were identified in any of the cell lines. About 2 kb of the 5' upstream region and 4.3 kb of the 3' untranslated region of the ER gene were present and grossly intact in all cell lines. By Northern blotting, ER mRNA was shown to be present in all ER-positive cell lines and, with the exception of BT-20 cells, absent in all ER-negative cell lines. However, ER mRNA was found to be detectable in all ER-negative cell lines when using the more sensitive RT-PCR. ER protein was undetectable by Western blotting in all cell lines that were ER-negative by ligand-binding assay. This study indicates that the ER-negative phenotype of breast cancer cell lines is not the result of ER gene mutations. It is due to the actual absence of ER protein which results from deficient ER expression at the transcriptional or post-transcriptional level.
No preview · Article · Dec 1995 · Endocrine Related Cancer
[Show abstract][Hide abstract] ABSTRACT: Background: In breast cancer patients, about two thirds of the tumors are estrogen receptor (ER)-positive and one third are ER-negative. The molecular mechanisms leading to the ER-negative phenotype are poorly understood. Nearly all ER-negative and about 40% of ER-positive cancers are resistant to endocrine therapy. Purpose: In this study, we examined the entire coding region of the ER gene in ER-positive and ER-negative primary breast tumors to determine whether deletions/insertions or point mutations might account for the ER-negative phenotype. Methods : We amplified exons 1 through 8 of the ER gene in 118 ER-positive and 70 ER-negative primary breast tumors and searched for mutations by single-strand conformation polymorphism analysis, denaturing gradient gel electrophoresis, and DNA sequencing. Results : Both ER-negative and ER-positive tumors contained neutral polymorphisms in codons 10 [TCT↑TCC (Ser)], 87 [GCG↑GCC (Ala)], 243 [CGC↑CGT (Arg)], 325 [CCG↑CCG (Pro)], and 594 [ACA↑ACG (Thr)]. There was no correlation of any of the polymorphic alleles with the ER phenotype or other clinicopathologic parameters including tumor type, size, grade, or stage. However, the polymorphism in codon 325 showed a strong association with a family history of breast cancer (P =.0005). This association was observed both in premenopausal and postmeno-pausal patients. Despite extensive searching in exons 1 through 8, we found no deletions/insertions and only two missense mutations in codons 69 [AAC (Asn)↑AAG (Lys)] and 396 (ATG (Met)↑GTG (Val)) of the same ER-negative tumor. Thus, only 1% of the primary breast cancers had point mutations in the ER gene. Conclusions : In the majority of primary breast cancers, the ER-negative phenotype is not the result of mutations in the coding region of the ER gene, but is due to deficient ER expression at the transcriptional or post-transcriptional level. Implications : The correlation reported previously, as well as our current findings, suggest that further investigations are warranted to understand the possible linkage of the ER gene locus to hereditary breast cancer. (J Natl Cancer Inst 87: 446-451, 1995)
No preview · Article · Apr 1995 · JNCI Journal of the National Cancer Institute
[Show abstract][Hide abstract] ABSTRACT: Microsatellite instability (MSI) has been described in colorectal and other cancers. The purpose of this study was to determine the presence of MSI in breast cancer and to correlate its occurrence with clinicopathological parameters. For microsatellite markers we examined mono-, di-, tri-, and tetranucleotide repeats that, due to their polymorphic nature, may also be used to investigate loss of heterozygosity. In 20 paired breast cancer-peripheral blood DNA samples we identified four tumors (20%) with somatic MSI. All four tumors were stage I or II, grade 1 or 2, and estrogen receptor positive. To study MSI in relation to tumor progression we also examined paired DNA samples from two ipsilateral and three contralateral breast cancers, as well as two matched tumor-metastatic lymph node specimens. None of these seven cases showed MSI, but two of the contralateral tumors revealed allelic loss of polymorphic repeats. These data suggest that MSI is an early event in mammary tumorigenesis while loss of heterozygosity may occur at a later stage.