ANTICANCER RESEARCH 24: 2681-2688 (2004)

Deletions on Chromosome 4 in Sporadic and BRCA Mutated Tumors and Association with Pathological Variables

HREFNA K. JOHANNSDOTTIR1, GUDRUN JOHANNESDOTTIR1, BJARNI A. AGNARSSON1, HANNALEENA EEROLA2,3, ADALGEIR ARASON1, OSKAR TH. JOHANNSSON4,5, PÄIVI HEIKKILÄ6, VALGARDUR EGILSSON1, HAKAN OLSSON5, ÅKE BORG5, HELI NEVANLINNA2 and ROSA B. BARKARDOTTIR1

1Department of Pathology and 4Department of Oncology, Landspitali-University Hospital by Hringbraut, IS-121, Reykjavik, Iceland; 2Department of Obstetrics and Gynecology, Biomedicum Helsinki, P.O. Box 700, 00029 HUS, Helsinki; 3Department of Oncology, P.O. Box 180, 00029 HUS, Helsinki; 5Department of Oncology, Klinikgatan 7, University Hospital of Lund, SE-22185; 6Department of Pathology, P.O. Box 400, 00029 HUS, Helsinki, Finland

Abstract. Background: Chromosomal aberrations in breast 4p16.3-p16.1, 4q27-q32.1 and 4q35.1-4qter, have not been tumors from BRCA1 and BRCA2 germ-line mutation carriers reported in breast cancer previously. Conclusion: The results are considerably more frequent than what is seen in sporadic manifest the frequent alterations of chromosome 4 in BRCA1- breast tumors. According to Comparative Genomic associated breast tumors and indicate the location of several Hybridisation analysis (CGH), deletions on chromosome 4 are genes of potential importance in breast cancer development. one of the most frequent events in BRCA1-associated tumors, suggesting inactivation of specific tumor suppressor genes. Breast cancer is the most frequently diagnosed malignancy Materials and Methods: In the present study, 16 microsatellite in women in the Western countries (1). Of the many risk markers covering chromosome 4 were used to map loss of factors that have been identified for breast cancer, family heterozygosity (LOH) in tumors from BRCA1 (n=41) as well history is the best established one (2). Approximately half of as in tumors from BRCA2 (n=66) mutation carriers and in high risk familial breast cancer cases can be explained by tumors from unselected cases of breast cancer (n=68). Results: inactivation of mutation in one of the two major breast The frequency of LOH in these groups ranged from 16-73% in cancer susceptibility genes, BRCA1 or BRCA2, although the BRCA1-associated tumors, 13-42% in BRCA2-associated proportion varies in different populations (3-5). Tirkkonen tumors and 8-33% in unselected tumors. LOH was significantly et al. have used comparative genomic hybridisation (CGH) more frequent in BRCA1-associated tumors as compared to to define genetic alterations in breast tumors from BRCA1 BRCA2-associated tumors and unselected tumors, and and BRCA2 mutation carriers (6). The study revealed a high particularly high (over 70%) at 4q35.2. Pathological variables frequency of overall chromosomal copy number changes in that were found significantly associated (p≤0.05) with LOH at BRCA1 and BRCA2 mutation carriers in comparison with specific markers were: high percentage of cells in S-phase, tumors from unselected breast cancer patients. This elevated negative estrogen receptor status, young age at diagnosis and frequency of genetic aberrations could possibly reflect the large tumors. Deletion mapping indicates the existence of seven important roles of BRCA1 and BRCA2 in maintaining non-overlapping regions at chromosome 4, which were identified genomic integrity and stability (reviewed in (7)). in all three groups of tumors. Three of these seven regions, Furthermore, the pattern of genetic changes observed suggest that BRCA1 and BRCA2 tumors have a different spectrum of genetic changes, indicating distinct progression pathways depending on the genetic background. Correspondence to: Rosa B. Barkardottir, Department of Pathology, In the CGH analysis by Tirkkonen et al., deletions on House 14 by Eiriksgata, Landspitali-University Hospital by chromosome 4q were among the most frequently observed Hringbraut, IS-121, Reykjavik. Tel: 354-5438033, Fax: 354-5434828, e-mail: [email protected] events in BRCA1 tumors (6). As the resolution of CGH is limited, a more refined approach is needed to define the Key Words: Breast cancer, BRCA1, BRCA2, loss of heterozygosity. regions involved. In the present study, mapping of LOH

0250-7005/2004 $2.00+.40 2681 ANTICANCER RESEARCH 24: 2681-2688 (2004)

Table I. Frequency of loss of heterozygosity at chromosome 4 in BRCA1 and BRCA2 mutated tumors and unselected tumors and the association between loss of heterozygosity at respective markers and pathological variables.

Marker Cytogenetic location Loss of heterozygosity/informative samples (%) Association with (Mb)* BRCA1 BRCA2 Unselected Total pathological variables** (p value)

D4S2936 4p16.3 (0.7) 11/26 (42) 14/33 (42) 4/32 (13) 29/91 (32) S-phase (0.01) D4S43 4p16.3 (2.3) 6/14 (43) 9/26 (35) 3/31 (10) 18/71 (25) ER (0.001) D4S1614 4p16.3 (2.7) 9/15 (60) 6/22 (27) 4/27 (15) 19/64 (30) D4S394 4p16.1 (7.0) 10/18 (56) 15/42 (36) 7/54 (13) 32/114 (28) S-phase (0.03) D4S1546 4p15.31 (20.5) 8/22 (36) 14/38 (37) 6/42 (14) 28/102 (27) D4S2971 4q12 (53.6) 3/19 (16) 6/45 (13) 6/29 (21) 15/93 (16) S-phase (0.02) D4S3000 4q12 (57.2) 7/19 (37) 8/47 (17) 4/42 (10) 19/108 (18) D4S2964 4q21.21 (81.2) 14/27 (52) 12/38 (32) 6/31 (19) 32/96 (33) S-phase (0.005) D4S414 4q22.1 (92.9) 7/20 (35) 12/40 (30) 3/40 (8) 22/100 (22) D4S2917 4q25 (109.2) 8/18 (44) 14/43 (33) 12/36 (33) 34/97 (35) D4S427 4q27 (121.9) 9/22 (41) 12/38 (32) 4/29 (14) 25/89 (28) age (0.02) D4S424 4q31.21 (142.8) 11/24 (46) 13/48 (27) 4/33 (12) 28/105 (27) S-phase (0.004), size (0.02) D4S413 4q32.1 (158.9) 14/23 (61) 12/48 (25) 10/42 (24) 36/113 (32) S-phase (0.03) D4S1595 4q34.1 (175.0) 12/19 (63) 12/44 (27) 9/30 (30) 33/93 (35) S-phase (0.003) D4S1554 4q35.1 (185.4) 11/24 (46) 15/47 (32) 7/44 (16) 33/115 (29) age (0.05) D4S426 4q35.2 (189.8) 11/15 (73) 14/38 (37) 7/36 (19) 32/89 (36) ER (0.02)

* localisation of markers in megabases according to UCSC Genome Bioinformatics (http://genome.uscs.edu/) ** the pathological variables tested were estrogen receptor status (ER), progesterone receptor status (PgR), histological type of tumor, tumor size, age at diagnosis, percentage of cells in S-phase, ploidy and tumor grade

with microsatellite markers covering chromosome 4 in al. (8). DNA was extracted from blood samples according to Miller tumors from BRCA1 as well as in BRCA2 carriers and et al. (9). Mapping of chromosome 4 was performed using 16 unselected breast cancer patients was performed. The microsatellite markers (MWG-Biotech AG, Ebersberg, Germany) (Table I). PCR was performed in 15 Ìl reactions with 25 ng DNA, deletion frequency was found to be very high at specific loci 2 mM MgCl2, 0.16 mM nucleotides (Amersham Pharmacia Biotech in BRCA1-associated tumors and, although in most cases at Inc, Buckinghamshire, England), 0.32 ÌM each primer and 0.3 u a lower frequency, the same loci are also affected in the Taq DNA polymerase together with the enclosed 10xbuffer (MBI other subgroups of breast tumors. Fermentas, St. Leon-Rot, Germany). Reactions were as follows: 30 sec at 94ÆC, 45 sec at 55ÆC and 45 sec at 72ÆC for 35 cycles. The Materials and Methods reactions were preceded with denaturation at 94ÆC for 3 min and followed by 10 min at 72ÆC. Samples. DNA from 175 paired blood and breast tumor samples were subjected to analysis of LOH on chromosome 4. The sample LOH analysis. The PCR products were separated on a 6.5% set included tumors from BRCA1 (n=41) and BRCA2 (n=66) polyacrylamide gel (8M urea) in 1 x TBE buffer, and blotted to mutation carriers as well as tumors from unselected cases of breast Hybond-N+ nylon transfer membrane (Amersham Pharmacia cancer (n=68). The Swedish samples were collected at the Biotech Inc, Buckinghamshire, England). The DNA bands were Department of Oncology/Oncogenetic clinic at the Lund University visualised using ECL (Amersham Pharmacia Biotech Inc, Hospital (18 BRCA1 mutated and 22 unselected breast tumors), Buckinghamshire, England) according to the manufacturer’s the Finnish samples at the Department of Obstetrics and protocol with the modification of employing HEPES buffer (40 Gynecology and Oncology, Helsinki University Hospital (15 mM K-HEPES, pH 7.2 with 1mM CoCl2), instead of cacodylate, BRCA1 and 17 BRCA2 mutated tumors), and the Icelandic for tailing oligomers with terminal transferase for probe samples at the Department of Pathology, Landspitali-University preparation. Absence of or significant decrease in the intensity Hospital (8 BRCA1, 49 BRCA2 mutated tumors and 46 unselected of one allele relative to the other was considered LOH. tumors). The study was approved by the respective Ethical and Data Protection Committees in each country. Statistical analysis. A Chi-square test with Yates correction was used to compare LOH frequencies between the three groups of DNA extraction and PCR. DNA extraction from paraffin-embedded breast tumors as well as to evaluate the relationship between tissue and fresh tumor tissue was performed according to Smith et LOH and different pathological variables. Fisher’s exact test was

2682 Johannsdottir et al: LOH at Chromosome 4 in Breast Tumors

Figure 1. The frequency of loss of heterozygosity on chromosome 4 in tumors from patients with germ-line mutation in BRCA1 (■), BRCA2 ( ) and unselected tumors (■). a = significant difference between the frequency of LOH in BRCA1 mutated versus unselected tumors. b = significant difference between the frequency of LOH in BRCA1 versus BRCA2 mutated tumors. c = significant difference between the frequency of LOH in BRCA2 mutated versus unselected tumors. The numbers attached to a, b and c refer to the level of significance; 1 = p≤0.05, 2 = p≤0.01 and 3 = p≤0.001.

used if the number within one or more columns was 5 or less. For figures for the BRCA2 and unselected tumors were 13 - each marker, 8 Chi-squared tests were performed. Processing of 42% and 8 - 33%, respectively (Figure 1 and Table I). the data was performed using SPSS 9.0 program package. The Of the 30 BRCA1-associated tumors informative on at level of significance was set at 0.05. least 6 markers distributed equally along the chromosome, 3 tumors (10%) showed LOH at all informative markers and Results 5 (17%) showed retention of heterozygosity on all informative markers. Corresponding figures for BRCA2- LOH on chromosome 4 and association with pathological associated (n=59) and unselected tumors (n=50) were 1 variables. In the present study, 16 microsatellite markers, (2%) and 18 (31%), and 1 (2%) and 25 (50%), respectively. covering chromosome 4, were typed for LOH and used to In Figure 2, tumor samples with regional and mosaic LOH map the deleted regions in tumors from BRCA1 mutation pattern are shown. Samples showing LOH or retention of carriers as well as BRCA2 mutation carriers and heterozygosity at all informative markers were excluded unselected cases of breast cancer. LOH in BRCA1- from the figure. associated tumors was markedly more prevalent than in The deletion mapping in BRCA1 tumors was in some BRCA2-associated tumors and unselected tumors as can cases difficult due to the extensive deletions. However, the be seen in Figure 1. Significant difference was found concordance between target regions in all three groups of between LOH frequency in BRCA1-associated tumors and tumors, and the less extensive deletion pattern in BRCA2- the unselected cases at 12 markers, between BRCA2 associated tumors and unselected tumors, facilitated the mutated and unselected tumors at 5 markers and between fine-mapping process. The LOH frequency reaches over the BRCA1 and BRCA2 groups at 3 markers (Figure 1). 70% for marker D4S426 at 4q35.2 in BRCA1 mutated For individual markers the deletion frequency in BRCA1- tumors (Figure 1). By examination of the LOH pattern in associated tumors varied from 16-73%. Corresponding BRCA1 mutated tumors at this loci (Figure 2a), it can be

2683 ANTICANCER RESEARCH 24: 2681-2688 (2004) seen that, although the deletions in these tumors are in most cases quite extensive, some of the samples show a specific LOH target region at 4q35.1 (D4S1554) to 4qter. The same region can be identified as a LOH target in BRCA2-associated and unselected tumors (Figure 2b and c), but the deletion frequency in these groups at these loci is considerably lower, particularly in unselected tumors. Based on our mapping results (Figure 2), we could also identify LOH target regions defined by the markers D4S1614-394 at 4p16.3-16.1, D4S3000-2964 at 4q12- q21.21, D4S427-413 at 4q27-q32.1 and D4S413-1554 at 4q32.1-q35.1. These regions all show 50-60% LOH = ROH (retention of heterozygosity), = nd unselected tumors (c). The numbers in the frequency in BRCA1-associated tumors. Two additional ● target loci were identified in all three groups of tumors, one located at 4p16.3-4pter and another surrounded by the markers D4S414 and D4S427 at 4q22.1-4q27. These two regions showed similar LOH frequency in both BRCA1 and BRCA2-associated tumors (30-40%). As the LOH frequency for these regions was among the highest seen in this study for BRCA2-associated tumors, they are likely to be of specific relevance for the progression of

these tumors. = LOH (loss of heterozygosity), As the mapping results indicated that the same loci ● were affected in all three groups of breast tumors, we decided to pool the groups in our association analysis between LOH at chromosome 4 and pathological variables. The information on pathological variables available were estrogen receptor status (ER), progesterone receptor status (PgR), histological type of tumor, tumor size, age at diagnosis, percentage of cells in S-phase, ploidy and tumor grade. Significant association was found between LOH and one or more pathological variables at 11 of the 16 markers used for the study and the results are shown in Table I. The variables that were found associated with LOH are high percentage of cells in S-phase, negative estrogen receptor status, young age at diagnosis and tumor size (Table I).

Discussion

In the current study, parallel mapping of LOH with microsatellite markers on chromosome 4 in breast tumors from BRCA1 and BRCA2 germ-line mutation carriers and unselected breast tumors is reported. Such mapping has not been performed previously. The results indicate seven target regions that were seen affected in all three groups of tumors. Three of these targets (4p16.3-p16.1, 4q27-q32.1 and 4q35.1-4qter) have not been reported previously for breast tumors. A very high frequency of deletions was seen at specific Deletion patterns in breast tumors showing regional LOH at chromosome 4 BRCA1 mutated (a), BRCA2 (b) a chromosomal regions in BRCA1-associated tumors, whereas at these same loci the deletion frequency was Figure 2. found to be more moderate, and in some cases considerably front refer to the name of chromosome 4 markers used in study and are numbers following D4S. homozygosity or not informative.

2684 Johannsdottir et al: LOH at Chromosome 4 in Breast Tumors lower, in the BRCA2-associated and unselected tumors. At D4S424 (4q31.21) and large tumors that could reflect the other loci, an approximately equal LOH frequency was increased proliferation indicated by a high percent of cells seen for BRCA1- and BRCA2-associated tumors, whereas in S-phase. considerably lower deletion frequency was seen for A significant association was seen between LOH at unselected breast tumors. This is in line with our previous markers D4S43 (4p16.3) and D4S426 (4q35.2) and negative results from studies comparing frequency and pattern of estrogen receptor status. A negative estrogen receptor LOH at other chromosomes in BRCA2-associated and status has been associated with BRCA1-associated tumors sporadic tumors. Those studies showed that many of the while BRCA2-associated and sporadic tumors more loci affected in tumors of BRCA2 carriers are also affected frequently show a positive estrogen receptor status (18-20). in tumors from non-carriers, although at lower frequency Interestingly, LOH at 4p16.3 was associated with negative (10-13). These observations indicate specific, rather than estrogen receptor status in all three groups, while the unique tumor progression pathways in patients predisposed association to LOH at 4q35.2 was mainly contributed by to breast cancer due to a mutation in BRCA1 or BRCA2. BRCA1-associated tumors. At 4p16.3, 73% of all tumors It might be speculated that, in some cases, by coincidence showing LOH and 21% of all tumors not showing LOH or due to reduced expression level of BRCA1 or BRCA2 were ER-negative (p<0.001), and the respective figures for (reviewed in (14)), tumors in non-carriers might follow the BRCA2-associated and unselected tumors together were same progression pathways as BRCA associated tumors. 64% and 7% (p<0.001). For 4q35.2, the figures were 52% Several recently published observations support this theory. and 25% (p=0.01) and 29% and 21% (p=0.5), respectively.) Lack of nuclear Brca1 localization in sporadic tumors has This raises the question as to whether the mechanism been associated with a younger age at diagnosis, indicating behind the down-regulation of estrogen receptors is that any disturbance of molecular pathways in which the different depending on genetic background and whether the Brca1 protein participates might lead to a specific tumor above mentioned chromosomal regions harbor genes that biology, normally associated with BRCA1-associated participate in different pathways of the down-regulation or tumors (15,16). In addition, a recently published article silencing of the estrogen receptor. In a recent study, an describes an oncogene, EMSY, that provides a link association was reported between negative estrogen between the progression pathway seen in tumors with receptor status and poorer breast cancer survival in BRCA germ-line mutation in BRCA2 and the progression of mutation carriers (20), which emphasises the importance of sporadic tumors. Emsy, which is a suppressor of Brca2 the identification of factors affecting estrogen receptor function, is amplified and overexpressed in some sporadic status of breast tumors. breast and ovarian tumors and this overexpression was A significant association was found between LOH at found to be associated with a pathological profile similar markers D4S427 (4q27), and D4S1554 (4q35.1) and young to that seen for BRCA2-associated tumors (17). Last but age at diagnosis. Young age at diagnosis is a well established not least, in a recent publication by Wang et al., frequent risk factor for BRCA mutation carriers (20-23). It is LOH at chromosome 4 in sporadic tumors was described therefore of interest to note that the association between as one of the hallmarks of the so-called basal-like breast LOH at these loci and younger age at diagnosis does not tumors that show certain similarities to BRCA1 tumors. appear to be restricted to BRCA associated tumors as we The authors suggest that basal-like tumors might represent see the same tendency in all three subgroups. This could a subclass of sporadic tumors showing a defect in the indicate that losses of heterozygosity at these loci are BRCA1 pathway. implicated in a more aggressive progression pathway In the present study a significant association was found regardless of BRCA germ-line mutation status. between LOH at markers D4S2936 (4p16.3), D4S394 Using CGH, two groups of investigators have identified (4p16.1), D4S2971 (4q12), D4S2964 (4q21.21), D4S424 chromosome 4q as frequently lost in breast tumors (6, 24). (4q31.21), D4S413 (4q32.1) and D4S1595 (4q34.1) and a Tirkkonen et al. (6) suggested that 4q is a specific target in high percent of cells in S-phase (Table I). Histopathological BRCA1-associated tumors, since loss was only seen in studies have shown that BRCA associated tumors show a around 10% of unselected tumors while Schwendel et al. higher percentage of cells in S-phase than sporadic ones (24) reported over 50% loss at 4q in their LOH study of (18-20). Despite the high incidence of LOH at these loci in unselected tumors. Shivapurkar et al. performed LOH BRCA1-associated tumors, our results indicate that all mapping on chromosome 4 in unselected breast tumors three groups of tumors contributed to the above mentioned and reported four regions of loss with a deletion associations suggesting that loss of one or more loci at frequency ranging from 50% to 76% (25). Some of these chromosome 4 causes cell cycle deregulation independent regions (4p16.3, 4q25-q26 and 4q33-q34) overlap with of BRCA germ-line mutation status. Interestingly, an regions identified in the present study, but the LOH association was also found between LOH at marker frequency is in all cases considerably higher than that

2685 ANTICANCER RESEARCH 24: 2681-2688 (2004) observed in our set of unselected tumors. One explanation on chromosome 4 harbor tumor suppressor genes that are could be that the LOH frequency reported in the present of special importance in the progression pathway of study is an underestimation of the actual deletion tumors in BRCA1 carriers, and are also of significance in frequency, since the tumor tissue used for the analysis was the progression of a portion of BRCA2-associated tumors not microdissected and some of the tumor samples could and sporadic tumors that follow the same or similar thus contain a sufficient proportion of normal cells to pathway as BRCA1-associated tumors. The results are mask the deletions in the tumor cells of the sample. In the likely to be helpful in identifying new cancer genes that current study we identified four regions of frequent LOH might have therapeutic and diagnostic importance for (4p16.3-q16.1, 4q12-q21.21, 4q27-q32.1 and 4q35.1-4qter) breast cancer patients. that Shivapurkar et al. did not report. This could possibly be due to the different set of markers used and to the Acknowledgements difference in number of samples used in these two studies. Several interesting potential cancer genes are located We thank the personnel at the Department of Pathology for within the LOH targets identified in the present study. providing pathological information, Orn Olafsson for his assistance One of the LOH targets identified overlaps with 4q12 with the statistical analysis and Minna Merikivi for her kind help. (4q12-4q21.21), which contains the potential tumor This study was supported by The Nordic Cancer Union, The suppressor gene coding for insulin-like growth factor Research Council of Iceland, the Science Fund of the Icelandic Cancer Society, a research fund of the University of Iceland, the binding protein related protein 1 (IGFBP-rP1). The Memorial Fund of Bergthora Magnusdottir and Jakob Bjarnason, expression of IGFBP-rP1 has been reported to be reduced The Finnish Cancer Society, the Helsinki University Central or abrogated in ductal carcinoma in situ and invasive Hospital Research Fund, the Academy of Finland and the Sigrid cancers as compared to normal breast epithelia and, in the Juselius Foundation. same study, 50% LOH was detected at this loci in the breast tumors (26). IGFBP-rP1can cause transcriptional References activation of endogenous ER and mediate IGF-1-induced ER activation (26-29). Down-regulation of IGFBP-rP1 in 1 Parkin DM: Epidemiology of cancer: global patterns and breast tumors has also been associated with cyclin E trends. Toxicol Lett 102-103: 227-234, 1998. 2 Easton DF: How many more breast cancer predisposition genes overexpression. Cyclin E is a regulator of G1-S transition are there? Breast Cancer Res 1: 14-17, 1999. of the cell cycle and in line with this, its overexpression was 3 Serova OM, Mazoyer S, Puget N, Dubois V, Tonin P, Shugart found to be associated with increased cell proliferation. YY, Goldgar D, Narod SA, Lynch HT and Lenoir GM: Interestingly, our results show a significant association Mutations in BRCA1 and BRCA2 in breast cancer families: are between LOH at D4S2971 localised at 4q12 (p=0.02) and there more breast cancer-susceptibility genes? Am J Hum high percentage of cells in S-phase, which indicate an Genet 60: 486-495, 1997. increased cell proliferation. 4 Ford D, Easton DF, Stratton M, Narod S, Goldgar D, Devilee Another tumor suppressor candidate gene, SCFFbw7, is P, Bishop DT, Weber B, Lenoir G, Chang-Claude J, Sobol H, Teare MD, Struewing J, Arason A, Scherneck S, Peto J, located at 4q31.3, a loci showing high frequency of LOH in Rebbeck TR, Tonin P, Neuhausen S, Barkardottir R, Eyfjord the current study. The SCFFbw7 protein targets cyclin E for J, Lynch H, Ponder BA, Gayther SA, Zelada-Hedman M et al: ubiquitination, and has been found mutated in breast cancer Genetic heterogeneity and penetrance analysis of the BRCA1 cell lines showing a high expression level of cyclin E (30-32). and BRCA2 genes in breast cancer families. The Breast Cancer We found a significant association of LOH at both markers Linkage Consortium. Am J Hum Genet 62: 676-689, 1998. surrounding the loci, D4S424 (p=0.004) and D4S413 5 Szabo CI and King MC: Population genetics of BRCA1 and (p=0.03), and high percentage of cells in S-phase. BRCA2. Am J Hum Genet 60: 1013-1020, 1997. 6 Tirkkonen M, Johannsson O, Agnarsson BA, Olsson H, In conclusion, the results of our study suggest the Ingvarsson S, Karhu R, Tanner M, Isola J, Barkardottir RB, existence of several distinct regions at chromosome 4 that Borg A and Kallioniemi OP: Distinct somatic genetic changes are affected during tumor progression. The association associated with tumor progression in carriers of BRCA1 and between loss at particular loci and pathological variables, BRCA2 germ-line mutations. Cancer Res 57: 1222-1227, 1997. contributed to by all the three breast cancer subgroups, 7 Venkitaraman AR: Cancer susceptibility and the functions of indicates certain similarities of the biology of the tumors BRCA1 and BRCA2. Cell 108: 171-182, 2002. that have lost specific regions of chromosome 4. This 8 Smith SA, Easton DF, Evans DG and Ponder BA: Allele losses in the region 17q12-21 in familial breast and ovarian could possibly implicate that loss at some loci at cancer involve the wild-type chromosome. Nat Genet 2: 128- chromosome 4 results in a more aggressive tumor 131, 1992. progression pathway in BRCA2-associated and unselected 9 Miller SA, Dykes DD, and Polesky HF: A simple salting out tumors, resulting in a more BRCA1-like phenotype. It is procedure for extracting DNA from human nucleated cells. not unlikely that the frequently deleted regions identified Nucleic Acids Res 16: 1215, 1988.

2686 Johannsdottir et al: LOH at Chromosome 4 in Breast Tumors

10 Gudmundsson J, Johannesdottir G, Bergthorsson JT, Arason 22 Johannesdottir G, Gudmundsson J, Bergthorsson JT, Arason A, Ingvarsson S, Egilsson V and Barkardottir RB: Different A, Agnarsson BA, Eiriksdottir G, Johannsson OT, Borg A, tumor types from BRCA2 carriers show wild-type chromosome Ingvarsson S, Easton DF, Egilsson V and Barkardottir RB: deletions on 13q12-q13. Cancer Res 55: 4830-4832, 1995. High prevalence of the 999del5 mutation in icelandic breast and 11 Sigbjornsdottir BI, Ragnarsson G, Agnarsson BA, Huiping C, ovarian cancer patients. Cancer Res 56: 3663-3665, 1996. Barkardottir RB, Egilsson V and Ingvarsson S: Chromosome 8p 23 Peto J, Collins N, Barfoot R, Seal S, Warren W, Rahman N, alterations in sporadic and BRCA2 999del5 linked breast Easton DF, Evans C, Deacon J and Stratton MR: Prevalence cancer. J Med Genet 37: 342-347, 2000. of BRCA1 and BRCA2 gene mutations in patients with early- 12 Ingvarsson S, Geirsdottir EK, Johannesdottir G, Sigbjornsdottir onset breast cancer. J Natl Cancer Inst 91: 943-949, 1999. BI, Eiriksdottir G, Ragnarsson G, Agnarsson BA, 24 Schwendel A, Richard F, Langreck H, Kaufmann O, Lage H, Gudmundsson J, Jonasson JG, Sigurdsson A, Egilsson V and Winzer KJ, Petersen I and Dietel M: Chromosome alterations in Barkardottir RB: High incidence of loss of heterozygosity in breast carcinomas: frequent involvement of DNA losses including breast tumors from carriers of the BRCA2 999del5 mutation. chromosomes 4q and 21q. Br J Cancer 78: 806-811, 1998. Cancer Res 58: 4421-4425, 1998. 25 Shivapurkar N, Sood S, Wistuba, II, Virmani AK, Maitra A, 13 Bergthorsson JT, Johannsdottir J, Jonasdottir A, Eiriksdottir G, Milchgrub S, Minna JD and Gazdar AF: Multiple regions of Egilsson V, Ingvarsson S, Barkardottir RB and Arason A: chromosome 4 demonstrating allelic losses in breast Chromosome imbalance at the 3p14 region in human breast carcinomas. Cancer Res 59: 3576-3580, 1999. tumours: high frequency in patients with inherited 26 Burger AM, Zhang X, Li H, Ostrowski JL, Beatty B, Venanzoni predisposition due to BRCA2. Eur J Cancer 34: 142-147, 1998. M, Papas T and Seth A: Down-regulation of T1A12/mac25, a 14 Welcsh PL and King MC: BRCA1 and BRCA2 and the genetics novel insulin-like growth factor binding protein related gene, is of breast and ovarian cancer. Hum Mol Genet 10: 705-713, 2001. associated with disease progression in breast carcinomas. 15 Lee WY: Frequent loss of BRCA1 nuclear expression in young Oncogene 16: 2459-2467, 1998. women with breast cancer: an immunohistochemical study from 27 Murphy M, Pykett MJ, Harnish P, Zang KD and George DL: an area of low incidence but early onset. Appl Immunohistochem Identification and characterization of genes differentially Mol Morphol 10: 310-315, 2002. expressed in meningiomas. Cell Growth Differ 4: 715-722, 1993. 16 Bogdani M, Teugels E, De Greve J, Bourgain C, Neyns B and 28 Burger AM, Zhang X and Seth A: Detection of novel genes Pipeleers-Marichal M: Loss of nuclear BRCA1 localization in that are up-regulated (Di12) or down-regulated (T1A12) with breast carcinoma is age dependent. Virchows Arch 440: 274- disease progression in breast cancer. Eur J Cancer Prev 7 Suppl 279, 2002. 1: S29-35, 1998. 17 Hughes-Davies L, Huntsman D, Ruas M, Fuks F, Bye J, Chin 29 Lee AV, Weng CN, Jackson JG and Yee D: Activation of SF, Milner J, Brown LA, Hsu F, Gilks B, Nielsen T, Schulzer estrogen receptor-mediated gene transcription by IGF-I in M, Chia S, Ragaz J, Cahn A, Linger L, Ozdag H, Cattaneo E, human breast cancer cells. J Endocrinol 152: 39-47, 1997. Jordanova ES, Schuuring E, Yu DS, Venkitaraman A, Ponder 30 Koepp DM, Schaefer LK, Ye X, Keyomarsi K, Chu C, Harper B, Doherty A, Aparicio S, Bentley D, Theillet C, Ponting CP, JW and Elledge SJ: Phosphorylation-dependent ubiquitination Caldas C and Kouzarides T: EMSY links the BRCA2 pathway of cyclin E by the SCFFbw7 ubiquitin ligase. Science 294: 173- to sporadic breast and ovarian cancer. Cell 115: 523-535, 2003. 177, 2001. 18 Johannsson OT, Idvall I, Anderson C, Borg A, Barkardottir 31 Enders GH: Cyclins in breast cancer: too much of a good thing. RB, Egilsson V and Olsson H: Tumour biological features of Breast Cancer Res 4: 145-147, 2002. BRCA1-induced breast and ovarian cancer. Eur J Cancer 33: 32 Strohmaier H, Spruck CH, Kaiser P, Won KA, Sangfelt O and 362-371, 1997. Reed SI: Human F-box protein hCdc4 targets cyclin E for 19 Agnarsson BA, Jonasson JG, Bjornsdottir IB, Barkardottir RB, proteolysis and is mutated in a breast cancer cell line. Nature Egilsson V and Sigurdsson H: Inherited BRCA2 mutation 413: 316-322, 2001. associated with high grade breast cancer. Breast Cancer Res Treat 47: 121-127, 1998. 20 Robson ME, Boyd J, Borgen PI and Cody HS, 3rd: Hereditary breast cancer. Curr Probl Surg 38: 387-480, 2001. 21 de Sanjose S, Leone M, Berez V, Izquierdo A, Font R, Brunet JM, Louat T, Vilardell L, Borras J, Viladiu P, Bosch FX, Lenoir GM and Sinilnikova OM: Prevalence of BRCA1 and BRCA2 germline mutations in young breast cancer patients: a Received April 13, 2004 population-based study. Int J Cancer 106: 588-593, 2003. Accepted July 1, 2004

2687