Allelic imbalance in mammary carcinomas induced by either 7,12-dimethylbenz[a]anthracene or ionizing radiation in rats carrying genes conferring differential susceptibilities to mammary carcinogenesis

To identify and compare the genetic lesions associated with tumorigenesis in rats carrying the mammary carcinoma suppressor (MCS) 1 gene, we induced mammary carcinomas in (Wistar Furth (WF) × Copenhagen (Cop))F₁ rats by using either 7,12-dimethylbenz[a]anthracene (DMBA) or radiation. The tumors were screened for allelic imbalances by using polymerase chain reaction and 65 polymorphic microsatellite markers spanning the genome. No allelic imbalance was detected at the mapped location of MCS-1 on chromosome 2; however, a scan of the genome revealed random allelic imbalances in the radiation-induced tumors. In addition, non-random loss of heterozygosity (LOH) on chromosome 1 in the DMBA-induced tumors was documented. We then screened three other subsets of DMBA- and radiation-induced mammary carcinomas from (WF × Fischer (F344))F₁, (Wistar Kyoto×F344)F₁, and (F344×Cop)F₁ rats for imbalance on chromosomes 1 and 2. No allelic imbalance was detected in the MCS-1 region of chromosome 2 in any of the tumors screened. Nonrandom imbalance on chromosome 1 was detected but only in the DMBA-induced tumors from the (F344×Cop)F₁ rats. Thus, only Cop-derived F₁ rats have mammary tumors with the chromosome 1 imbalance; however, the imbalance does not favor the Cop parental allele. We also analyzed the DMBA-induced tumors with LOH at chromosome 1 for Ha-ras codon 61 mutation and found no association. These results suggest that loss of the MCS-1 Cop allele is not required for tumor formation, that the genetic background of the F₁ rat appears to influence the type of genetic lesion identified in the mammary tumors, and that there is no association between Ha-ras codon 61 mutation and chromosome 1 imbalance in our model system. © 1996 Wiley-Liss, Inc.     

Microsatellite instability and loss of heterozygosity on chromosome 10 in rat mammary tumors induced by 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine

Microsatellite instability (MI) and loss of heterozygosity (LOH) were examined in mammary tumors induced in Sprague-Dawley × F344 F₁ female rats by 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP). Examination of 62 microsatellite loci revealed MI in nine of 15 (60%) PhIP-induced mammary tumors, and five of these MI-positive tumors had mutations in more than one microsatellite locus. In contrast, two of 12 (17%) 7,12-dimethylbenz[a]anthracene (DMBA)-induced mammary tumors were MI positive but had mutations at only one locus each. Further, by using 37 polymorphic markers specific LOH was observed in four of 15 PhIP-induced mammary tumors on distal parts of rat chromosome 10, which is homologous to human chromosome 17q, with no background level of LOH. Similarly, DMBA-induced mammary tumors showed specific LOH on the same region of chromosome 10. These data suggest that mismatch-repair deficiency and loss of chromosome 10 are involved in carcinogenesis of PhIP-induced rat mammary tumors. © 1996 Wiley-Liss, Inc.     

Effect of dietary fat on codon 12 and 13 Ha-ras gene mutations in 2-amino-1-methyl-6-phenylimidazo-[4,5-b]pyridine–induced rat mammary gland tumors

Activating mutations in and expression of the Ha-ras gene were examined in benign and malignant female Sprague-Dawley rat mammary gland tumors induced by the heterocyclic amine 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) and promoted by a diet high in polyunsaturated fat. Ha-ras mutations were detected in codons 12 and 13 by selective polymerase chain reaction amplification of mutated sequences and nucleotide sequencing. The percentage of Ha-ras mutations in carcinomas from PhIP-treated rats was significantly higher in rats on a low-fat diet than in rats on a high-fat diet (82% (nine of 11) vs 26% (seven of 27), respectively, P < 0.01). In addition, whereas 56% of the carcinomas with Ha-ras mutations from rats on a low-fat diet carried double Ha-ras mutations, none of the carcinomas from rats on a high-fat diet had double mutations. Ha-ras mutations were also detected in benign tumors (largely adenomas) induced by PhIP in rats on different diets; two of eight and three of four benign tumors examined from rats on low-fat and high-fat diets, respectively, had Ha-ras mutations, suggesting that activating Ha-ras mutations alone are not sufficient for PhIP-induced tumors to become malignant. No differences were observed in the level of Ha-ras mRNA expression in the different groups. In our animal model, a high-fat diet increased the incidence and percentage of malignant PhIP-induced mammary gland tumors yet decreased the percentage of carcinomas showing Ha-ras mutations. Thus, the complement of genetic alterations associated with PhIP-induced mammary gland carcinogenesis is probably altered by the level of dietary fat. Mol. Carcinog. 20:348–354, 1997. © 1997 Wiley-Liss, Inc.     

Allelic imbalance and altered expression of genes in chromosome 2q11-2q16 from rat mammary gland carcinomas induced by 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine

2-Amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), a compound found in cooked meat, is a mammary gland carcinogen in rats. Comparative genomic hybridization of PhIP-induced rat mammary gland carcinomas revealed loss in the centromeric region of 2q, a region known to carry the mammary carcinoma susceptibility 1 (Mcs1) gene and several other genes relevant to carcinogenesis. Allelic imbalance, specifically microsatellite instability and loss of heterozygosity, was examined in mammary gland carcinomas induced by PhIP in Sprague-Dawley (SD)xWistar Furth F1 hybrid rats. In a polymerase chain reaction (PCR)-based assay with 34 microsatellite markers coinciding to 2q11-2q16, nine markers revealed allelic imbalance. The frequency of imbalance in the tumors varied from 10 to 100% depending on the specific marker. However, none of the markers coinciding with the Mcs1 gene locus showed allelic imbalance, suggesting that alterations at this locus were not associated with PhIP-induced rat mammary gland cancer. The expression of several genes physically mapped to 2q11-2q16 and potentially involved in carcinogenesis including Ccnb (cyclin B1), Ccnh (cyclin H), Rasa (Ras GAP), Rasgrf2, Pi3kr1 (p85alpha), and Il6st (gp130) was also examined by quantitative real-time PCR and immunohistochemistry (IHC) across a large bank of PhIP-induced SD rat mammary gland carcinomas. By quantitative real-time PCR, the mRNA expression of Rasa, Pi3kr1, Ccnh, and Il6st in carcinomas was, respectively, 22-, 20-, three- and threefold higher in carcinomas than in control mammary gland tissues (P<0.05, Student's t-test). A statistically sixfold lower expression of Rasgrf2 was detected in carcinomas whereas no significant change in Ccnb1 expression was observed. The findings from quantitative real-time PCR were confirmed by IHC for each gene. In addition, the proliferation index in mammary gland carcinomas as assessed by PCNA was found to correlate with the overexpression of Cyclin H by IHC analysis (P<0.05, Spearman Rank Order Correlation). The findings from the current study implicate molecular alterations in the proximal region of 2q in PhIP-induced rat mammary gland carcinomas.     

A deletion map of chromosome 17q in sporadic human mammary carcinomas

The long arm of chromosome 17 is a frequent target of allelic losses at multiple sites during breast cancer formation and progression. Several genes linked to breast carcinomas have been mapped on this chromosome such as BRCA1, NME and erbB2 genes. The aim of this work was to delineate a deletion map on chromosome 17q and to examine the role of loss of heterozygosity (LOH) during breast tumor development and progression looking for correlation between LOH on 17q and various histopathological parameters. A series of 71 human mammary carcinomas and the corresponding peripheral blood lymphocytes has been studied for loss of heterozygosity at 6 different polymorphic loci on chromosome 17q. 46 out of 71 (65%) tumors showed LOH on 17q. A positive correlation was found between allelic loss for BRCA1 flanking markers and young age at diagnosis. The absence of estrogen receptors was more frequently observed in tumors with deleted BRCA1 flanking markers.     

Frequent allelic losses on chromosome 13q in human male breast carcinomas

Loss of genetic material on chromosomes 13q and 17 has been suggested to be of importance in the initiation and progression of female breast cancer, but their involvement is less well illustrated in male breast carcinomas. The present study was designed to investigate the incidence of allelic loss and microsatellite instability for chromosomes 13q, 17p and 17q in 13 sporadic male breast carcinomas using matched normal-tumour DNA samples and seven polymorphic microsatellite markers. Genetic imbalance was found in one or more informative markers in 85% of the patients, with more frequent loss of heterozygosity and microsatellite instability at loci on chromosome 13q. Thus, a high incidence of allelic losses was observed at the retinoblastoma gene (4/6) and likewise at the D13S263 locus (7/12), which also exhibited the highest frequency of microsatellite instability. The intragenic microsatellite in intron 1 of the TP53 gene on chromosome 17p revealed loss of heterozygosity in 3 of 8 informative patients. The investigated proximal region of chromosome 13q is postulated to harbour several potential tumour suppressor genes associated with female breast cancer. The high incidence of allelic losses at the D13S263 microsatellite, located distal to both the BRCA2 and the Brush-1 loci but proximal to the retinoblastoma gene, possibly indicates the presence of an additional tumour suppressor gene which may be involved in male breast carcinomas. However, this hypothesis needs verification in an extended study of male breast carcinomas.     

Deletion mapping of chromosome 4q in hepatocellular carcinoma

Hepatocellular carcinoma (HCC) frequently shows a loss of heterozygosity (LOH) on chromosome 4q. In order to define the commonly affected region on chromosome 4q for further positional cloning of the putative tumor suppressor gene, we carried out allelic imbalance (AI) studies in 41 HCCs using a panel of 43 microsatellite markers. Thirty-four cases (82.9%) showed AI at one or more loci. Detailed deletion mapping identified 7 independent, frequently deleted regions on this chromosome arm. These were the (1) D4S1615 locus, (2) D4S1598 locus, (3) D4S620 locus, (4) D4S1566 and D4S2979 loci, (5) D4S1617 and D4S1545 loci, (6)D4S1537 locus; and (7) from the D4S2920 to D4S2954 locus. Among these 7 frequently deleted regions, 5 were associated with tumor differentiation. Our results suggest that several putative tumor suppressor genes may be present on chromosome 4q and that the AI of chromosome 4q may play a role in the aggressive progression of HCC. Int. J. Cancer (Pred. Oncol.) 79:356–360, 1998. © 1998 Wiley-Liss, Inc.     

Nonrandom abnormalities involving chromosome 1 and Harvey-ras-1 alleles in rat mammary tumor progression.

Little is known about the role of chromosomal abnormalities in the widely used models of rat mammary carcinogenesis. In this study, we cytogenetically analyzed nitrosomethylurea-induced rat mammary adenocarcinomas at different time points of development. As tools to study more advanced stages of malignant progression, we also analyzed the cytogenetic progression of tumors transplanted into younger syngeneic hosts, and of tumors that did not regress or that developed after host ovariectomy. Our results indicate that rat mammary adenocarcinomas appear to start development as diploid lesions with cytogenetically "normal" karyotypes. However, upon progression, tumors showed coexistence of normal diploid clones with abnormal clones bearing specific abnormalities affecting mainly chromosomes 1 and 15. Almost every ovary-independent tumor presented stem lines with specific nonrandom chromosomal abnormalities. Numerical chromosomal abnormalities such as specific trisomies started to develop mainly after subsequent in vivo transplantations. The abnormalities affecting chromosome 1 observed in many tumors were: (a) interstitial deletions and breakpoints for translocation in region 1q22; and (b) partial or complete overrepresentation of chromosome 1 in the form of direct duplication of region 1q22q43 or as trisomy 1. Interestingly, Harvey-ras-1 gene maps to rat chromosome 1, and by Southern analysis we observed that 4 of 8 primary tumors and 6 of 9 ovary-independent tumors showed considerable loss of Harvey-ras-1 signal indicating probable allele loss. However, analyses of some tumor transplants in more advanced stages of progression showed, paradoxically, an increased copy number of the Harvey-ras-1 oncogene coinciding with the presence of the direct duplication observed in chromosome 1 or with trisomy 1 as possible mechanisms for gene amplification. Interestingly, rat chromosome 1 is the homologue to human chromosome 11, and in numerous cases of human breast cancer loss of heterozygosity of several genes in chromosome 11p15.5 has been reported. Some of the rat chromosome 1 abnormalities observed may be equivalent to those affecting 11p15 in human tumors. We also observed 8 tumors with abnormalities affecting chromosome 15. At least 3 genes of interest in breast cancer have been previously mapped to that rat chromosome. The similarities observed with human breast cancer may point to common mechanisms of tumor progression in both species.     

Genomic imbalance in rat mammary gland carcinomas induced by 2-amino-1-methyl-6-phenylimidazo(4,5-b)pyridine

2-Amino-1-methyl-6-phenylimidazo(4,5-b)pyridine (PhIP), a compound found in cooked meat, is a mammary gland carcinogen in female Sprague-Dawley rats. PhIP-induced rat mammary gland carcinomas were examined for mutations in several genes (exons) known to regulate cell growth and apoptosis, including p53 (4-8), p21(Waf1) (coding region), Apc (14, 15), B-catenin (3), E-cadherin (9,13,15), Bcl-x (coding region), Bax (3), IGFIIR (28), and TGFBIIR (3). DNA from 30 carcinomas was examined by single-strand conformation polymorphism analysis, but no mutations were detected in these genes or gene regions. DNA from carcinomas and matching normal tissue were further screened for allelic imbalance by using a polymerase chain reaction-based approach with primers to known microsatellite regions located throughout the rat genome. Of 53 markers examined, 12 revealed allelic imbalance. Microsatellite instability (MSI) was detected at two markers, one on chromosome 4 and one on chromosome 6. Sixty-five percent and 96% of all carcinomas examined (N=23) showed MSI at these loci on chromosomes 4 and 6, respectively, supporting the notion that MSI plays a role in PhIP-induced mammary carcinogenesis. Loss of heterozygosity (LOH), an indication of a possible tumor suppressor gene, was observed at 10 markers distributed on chromosomes 3, 10, 11, 14, and X. The frequency of LOH at these markers was 75-94%, supporting that the regions of allelic imbalance were largely similar for the PhIP-induced carcinomas examined in this study. When PhIP-induced carcinomas from rats placed on high-fat and low-fat diet were compared, no unique regions of allelic imbalance or statistical differences in the frequency of allelic imbalance were observed. Therefore, the high-fat diet, known to be a promoter of PhIP-induced rat mammary carcinogenesis, did not appear to influence allelic imbalance in the carcinomas. Interestingly, 7,12-dimethylbenz[a]-anthracene-induced mammary carcinomas did not show allelic imbalance at 11 of the 12 loci that showed allelic imbalance in PhIP-induced carcinomas. These findings suggest that distinct chemical carcinogens induce different patterns of allelic imbalance during rat mammary carcinogenesis. Since several of the known genes involved in carcinogenesis did not harbor mutations in PhIP-induced carcinomas, further studies are needed to clarify the critical genes involved in PhIP-induced mammary carcinogenesis and to determine whether regions of LOH harbor potentially novel tumor suppressor genes involved in this disease.     

Deletion mapping of chromosome 19 in human gliomas

There is evidence that a putative glioma tumor suppressor locus resides on the long arm of chromosome 19. We present data on 161 gliomas from IS6 patients, which were studied by microsatellite analysis for loss of heterozygosity (LOH) on chromosome 19. Eight loci on the long arm and 2 loci on the short arm of chromosome IV were examined. LOH on I9qwas observed in 3/19 astrocytomas (WHO grade II), 12/27 anaplastic astrocytomas (WHO grade III), 16/76 cases of glioblastoma multiforme WHO (grade IV), 4/9 oligodendrogliomas (WHO grade II), 3/5 anaplastic oligodendrogliomas (WHO grade III), 5/9 mixed oligo-astrocytomas (WHO grade II) and 8/10 anaplastic oligo-astrocytomas (WHO grade III). While 31 of the tumors with LOH on chromosomal arm I9q exhibited allelic loss at every informative locus, 20 tumors showed terminal or interstitial deletions. In contrast to astrocytomas and glioblastomas, tumors with an oligodendroglial component had predominantly lost the entire long arm of chromosome 19. The common region of overlap in gliomas was located on 19q 13.2-q 13.4 between the markers D 19S 178 and D 19S 180. Our data confirm the involvement of a putative tumor suppressor gene on chromosomal arm 19q in gliomas and assign this gene to 19q 13.2-q 13.4.     

Loss of heterozygosity at loci from chromosome arm 22Q in human sporadic breast carcinomas

Forty-four breast carcinomas were studied for loss of heterozygosity (LOH) at 25 microsatellite markers distributed almost evenly along chromosome arm 22q. LOH at at least one marker were observed in 66% tumors, while 6 regions of consistent LOH were identified. The size of each region ranged between 3 and 6 cM, and the distance between each region was estimated to be 8 to 12 cM. Even if not all these regions contain a bona fide tumor-suppressor gene, it is possible that several loci from chromosome arm 22q may be involved in breast carcinogenesis. Int. J. Cancer 75:181–186, 1998. © 1998 Wiley-Liss, Inc.     

Lack of correlation between survival and allele loss on chromosome 7q31–32 in primary breast cancer

High incidence of loss of heterozygosity (LOH), affecting the 7q31–32 chromosome region in sporadic primary human breast carcinomas suggests the presence of a tumor suppressor gene in this region which seems relevant to the development of breast cancer. To further determine the possible role of this region in the pathogenesis of human primary breast cancer and association with survival, LOH analysis was performed on 52 primary breast cancer patients using a set of highly polymorphic microsatellite markers. Our panel contained twenty biopsy cases of unknown survival, nineteen cases with more than five years survival and fourteen cases with less than two years survival. Corresponding normal and tumor DNAs were analyzed by polymerase chain reaction (PCR). The data presented here demonstrate that all patients were informative at least at one locus and 20 (38%) out of 53 cases showed LOH at one or more loci on chromosome 7q31–32. Relatively high incidence of LOH (34%) was detected at the D7S522 microsatellite marker located near to the cMet proto-oncogene while lower frequencies were observed at D7S523 (19%) and D7S495 (17%) loci, supporting the existence of a putative tumor suppressor gene at the chromosome 7q31.1 region. Our results suggest that allelic imbalance on 7q may occur at an early stage of breast carcinogenesis, as no correlation was observed between allelic loss and clinico-pathological data. This work was supported by Hungarian Research Grants ETT (T-019307) and OMFB (04-1-99-94-0328) given to Edith Olah.     

An allelotype of papillary thyroid cancer

Although papillary carcinoma accounts for approximately 70% of all thyroid cancers, preliminary studies of allelic loss have thus far not identified any areas of chomosomal deletion. We evaluated 30 papillary thyroid carcinomas for chromosomal loss/allelic imbalance by testing at least 2 microsatellite markers from every autosomal arm. Fifteen of the 30 tumors tested exhibited loss of heterozygosity/allelic imbalance (LOH/AI) at one or more loci. Chromosomal arms with frequent LOH/AI included 4q, 5p, 7p and 11p. An average of 1.1 chromosomal arms displayed LOH/AI in each individual tumor. Therefore, 4q, 5p, 7p and, to a lesser extent, 11p display significant LOH/AI in papillary thyroid cancer, which indicates the presence of putative tumor-suppressor gene loci at these chromosomal arms. © 1996 Wiley-Liss, Inc.     

High frequency of allelic imbalance at chromosome region 16q22-23 in human breast cancer: Correlation with high pgr and low s phase

The loss of genetic material from a specific chromosome region in tumors suggests the presence of tumor-suppressor genes. Loss of heterozygosity (LOH) or allelic imbalance (AI) on the long arm of chromosome 16 is a known event in sporadic breast cancer. To locate the commonly deleted regions, and therefore (a) candidate tumor-suppressor gene(s), a deletion map of chromosome 16 was made, using 10 microsatellite markers on 150 sporadic breast tumors. The 3 smallest regions of overlap (SRO) were detected on the long arm of chromosome 16. Allelic imbalance was observed with at least one marker in 67% of the tumors. One marker, D16S421, at the 16q22-23 region, showed the highest allelic imbalance, 58%. Tumors with and without AI on 16q were tested for correlation with clinico-pathological features of the tumors such as estrogen- and progesterone-receptor content (ER and PgR), age at diagnosis, tumor size, node status, histological type, S-phase fraction, AI on chromosome 3p, and ploidy. A correlation was found between AI on 16q and high PgR content, also low S-phase fraction (99% confidence limits). A comparison of tumors with and without AI at the D16S421 marker locus revealed a slight correlation with high PgR content. The survival data showed no difference between patients with AI on 16q and those with a normal allele pattern on the long arm of chromosome 16. © 1995 Wiley-Liss, Inc.     

A revised map position for the Ha-ras gene on mouse chromosome 7: Implications for analysis of genetic alterations in rodent tumors

The mouse Ha-ras gene has previously been mapped to the central region of chromosome 7, 31 cM from the centromere, using an interspecific Mus musculus/Mus spretus backcross (Saunders AM, Seldin MF, Genomics 8:525–535, 1990). However, analysis of mitotic recombinations in mouse skin tumors from intraspecific F₁ hybrid mice suggested a more distal location for the Ha-ras gene on chromosome 7 (Bremner R, Balmain A, Cell 61:407–417, 1990). In the study reported here, we demonstrated, by analysis of Ha-ras gene mutations in skin tumors from interspecific M. spretus/M. musculus F₁ hybrids, the existence only in M. spretus of a pseudogene or other Ha-ras—related sequence that is probably the sequence originally mapped by Saunders and Seldin. The functional Ha-ras gene maps to the distal region of chromosome 7, and it is this sequence that acquires mutations in chemically induced tumors.     

Loss of heterozygosity at locus F13B on chromosome 1q in human medulloblastoma

Medulloblastoma is a primitive neuroectodermal tumor of the cerebellum with poorly understood pathogenesis. Previous studies have reported loss of heterozygosity (LOH) on chromosome arms 17p, 11p and 9q and cytogenetic abnormalities of chromosome 1 in medulloblastoma. We have used the polymerase chain reaction to amplify 10 microsatellites on the short arm and B microsatellites on the long arm of chromosome 1 to assess allelic loss in 22 medulloblastomas. Loss of heterozygosity (LOH) on chromosome 1 was found in 9 cases. Eight medulloblastomas (36%) showed an interstitial LOH on chromosome 1q. The common region of overlap was mapped between DISI604 and DIS237 and included the locus F13B in the chromosomal region 1q31–q32.1. An additional tumor had LOH in a proximal region of 1p, but did not exhibit LOH on 1q. None of the medulloblastomas exhibited LOH of the telomeric portion of chromosome 1p, which has been associated with several other human malignancies. Our data suggest the presence of a putative tumor suppressor gene located near the locus F13B on chromosome arm 1q that appears to be involved in the pathogenesis of medulloblastoma. © 1996 Wiley-Liss, Inc.     

Ki-ras mutations with frequent normal allele loss versus absence of p53 mutations in rat prostate and seminal vesicle carcinomas induced with 3,2′-dimethyl-4-aminobiphenyl

We have developed a prostate carcinogenesis model in Fischer 344 rats using 3,2′-dimethyl-4-aminobiphenyl (DMAB) as a carcinogen to examine various potential modifying factors. In this study, mutational changes in the ras and p53 genes were assessed in DMAB-induced rat prostate and seminal vesicle carcinomas by single-strand conformation polymorphism analysis and subsequent direct DNA sequencing. Eight of 22 prostate adenocarcinomas (three of nine (33.3%) from the ventral lobe and five of 13 (38.5%) from the dorsolateral lobe, including three transplantable tumors) and one of 11 seminal vesicle adenocarcinomas (9.1 %) demonstrated point mutations in the Ki-ras gene. One prostate malignant fibrohistiocytoma examined was negative. Among the positive cases, five (three ventral prostate carcinomas and two transplantable tumors) also showed loss of the normal allele. In contrast, other than one mutation in the p53 gene in the malignant fibrohistiocytoma, there were no mutations in the Ha-ras or p53 genes. These results indicate that mutational activation of the Ki-ras gene, but not of the Ha-ras or p53 genes may play a mechanistic role in prostate and seminal vesicle carcinogenesis by DMAB and that a loss of the normal allele of the Ki-ras gene may also be involved in the process. © 1995 Wiley-Liss, Inc.     

Allelic loss on chromosome 17 in human ovarian cancer

In order to identify a common region of deletion on chromosome 17 potentially containing a tumor-suppressor gene, 27 ovarian carcinomas and 3 ovarian tumors of low malignant potential (LMP) were examined for loss of heterozygosity (LOH) at 6 p arm and 10 q arm loci. Ninety percent of all tumors had deletions at one or more loci. On the p arm, there was a single near-common region of deletion on 17p 13.3 (D/7S30/ pYNZ22.1; 86% LOH), an intervening locus with a low LOH rate, and a more proximal locus on 17p11.2 (D/7S58/pEW301; 82% LOH) with a high LOH rate. In less aggressive tumors, LOH at Df 7S30 was not accompanied by LOH at p53. The q arm had a common region of deletion for high-stage carcinoma at D/7S579 (Mfd 188; 74% LOH) on q21, a locus tightly linked to the familial breast-ovarian-cancer syndrome (BRCAI) locus. D/7S579 was lost in all informative high-stage carcinomas and retained in all low-stage carcinomas and tumors of LMP. There may be at least 2 tumor-suppressor genes, an early-acting gene on the p arm and a gene on the q arm involved in tumor progression and metastasis.