GENES, CHROMOSOMES & CANCER 4:113-121 (1992)
Genetic Alterations of the Tumour Suppressor Gene Regions 3p, I Ip, 13q, 17p, and 17q in Human Breast Carcinomas l o n e 1. Andersen, Astrid Gaustad, Lars Ottestad, George W. Farrants, JahnM. Nesland, Kjell M. Tveit, and Anne-Lise Bgrresen Department of Genetics, Institute for Cancer Research (T.I.A., A.G., A.-L.B.), Department of Clinical Oncology (L.O., K.M.T.) and Department of Pathology, Institute for Cancer Research (G.W.F., J.M.N.), The Norwegian Radium Hospital, Montebello, Oslo, Norway
Fifty-nine primary breast carcinomas and I I metastases were examined t o identify genetic alterations in the tumour suppressor gene regions 3p, I Ip, I3q, I7p, and I7q. Loss of heterozygosity (LOH) was frequently observed on chromosome arms I7p (p 144D6 lost in 75%, pYNZ22. I in 55%, and TP53 in 48% of the primary tumours), 13q (RBI lost in 40% of the primary tumours), and 17q (pRMU3 lost in 35%. pTHH59 in 29%, and NM23H I in 26% of the primary tumours). Loss of all the markers except p144D6 w a s observed even more frequently in the metastases. Pairwise comparisons for concordance of allele losses on I7p indicated that there might be two genes on I7p implicated in breast cancer development; the TP53 gene and a gene located close t o the p144D6 and pYNZ22.I markers. LOH of the RBI gene w a s associated with LOH of pYNZ22.I and p144D6, but not with LOH of TP53. LOH of RBI and TP53 was associated with occurrence of ductal carcinomas, RBI and p 144D6 losses with tumour size, and p I44D6 losses with positive node status as well. LOH of TP53 and the three 17q markers NM23H I, pTHH59, and pRMU3 was most frequently observed in tumours from postmenopausal women. p144D6 losses occurred most frequently in progesterone receptor-negative tumours, whereas pTHH59 losses occurred most frequently in oestrogen receptor-negativetumours. LOH of the investigated loci was not associated with E R B M protooncogene amplification, with positive Family history of breast cancer, o r with survival. Genes Chrom Cancer 4:113-121 (1992). INTRODUCTION
It is increasingly recognized that loss of function mutations in critical genes, tumour suppressor genes, may play an important role in the pathogenesis of many malignancies (Green, 1988). In patients with retinoblastoma, loss or inactivation of both alleles of one such gene is associated with twnorigenesis (Cavenee et al., 1983; Friend et al., 1986). Tumour suppressor genes are normal genes whose products regulate cellular growth and differentiation. Through this action, they play an important role as inhibitors of the uncontrolled cellular proliferation characteristic of cancer. Several sites of allele losses have been reported in human breast carcinomas. The following chromosome arms have been found to be involved l q in l3-26% of primary tumours, 3p in l9-46%, l l p in 10-27%, 13q in 21-40%, 16q in &YO,17p in M 1 % , 17q in 3 W % and 1% in 2149% (Theillet et al., 1986, Ali et al., 1987;Lundberg et al., 1987; Mackay et al., 1988; Chen et al., 1989 Devilee et al., 1989; Bdrresen et al., 199oa; Coles et al., 1990; Cropp et al., 1990; Sat0 et al., 1990). Known putative tumour suppressor genes are located in some of these chromosomal regions; the RIK gene on 3p (Erlandsson et al., 1990), the Wilms’ tumour suppressor gene on l l p (Haber et al., 1990),the retinoblastoma (RBI) gene on 0 1992 WILEY-LISS, INC.
13q (Cavenee et al., 1983; Friend et al., 1986), and the deleted in colon cancer (DCC) gene on 18q (Fearon et al., 1990). On the short arm of chromosome 17, two distinct regions of loss of heterozygosity (LO@ have been identified, indicating that at least two genes on 17p are involved in breast carcinogenesis; the tumour suppressor gene TP53 and a gene suggested to regulate the expression of TP53, located telomeric to the TP53 gene (Mackay et al., 1988, Coles et al., 1990). In a previous paper that included preliminary results of the present study, we were the first to describe an association between LOH of 13q and 17p sequences (Wrresen et al., 1990a).These findings were confirmed by Sat0 et al. (1990),who further suggested a sequence of genetic alterations during tumour progression in breast cancer, with loss of 17p sequences, including TP53 as the first step. Loss of 13q material involving the R B I gene and/or amplification of the protooncogeneEmB2 on 17q were claimed to be later alterations in the cascade of genetic events. Loss of chromosome arm 1%was suggested to play a significant role in the development of lymph node metastases. Another putative metastatic suppressor gene, Received May 13, 1991; accepted August 5, 1991. Address reprint requests to Anne-Lise Wrresen, Department of Pathology, Institute for Cancer Research, The Norwegian Radium Hospital, Montebello, 0310 Oslo 3, Norway.
I14
ANDERSEN ET A L
NM23H1, has also been claimed to be involved in breast cancer (Steeg et al., 1988). This gene has been shown to be highly expressed in nonmetastatic, but not in metastatic, breast tumours (Bevilacqua et al., 1989). The aims of the present study were 1)to identify and determine the frequencies of losses of constitutional alleles in the tumour suppressor gene regions on chromosome arms 3p, Up, 13q,17p, and 17q; 2) to confirm our previous finding that LOH on 13q is associated with LOH of 17p sequences; 3) to study whether LOH of the regions investigated was associated with amplification of ERBB2; and 4) to study whether the LOH pattern is associated with particular clinical and histopathological parameters, a family history of breast cancer, or survival.
Blood Samples
Peripheral venous blood (10-20 ml) was collected in EDTA and stored at - 40°C until isolation of DNA. DNA Analyses
Material for this study was obtained from 68 breast carcinoma patients admitted to the Norwegian Radium Hospital in the period 1987-1989. All, except for one 66-year-oldmale, were female Norwegian whites, with a median age at diagnosis of 58 years, ranging from 31.0 to 80.5 years. Survival data were available for 63 patients. Mean observation time for these patients was 32 months (14-154 months). Eighteen patients had premenopausal and 43 had postmenopausal disease. Menopausal status was unknown for six patients. None had received treatment prior to surgery. A family history of breast cancer was obtained from patients or relatives. All cancer diagnoses reported were confirmed by the Norwegian Cancer Registry. A family history was considered positive if at least one first-degree relative had suffered from breast cancer. Families among whom no first-degree relative had reached the age of 35 years were not scored.
DNA was isolated from cell nuclei of tumour tissue and whole blood using standard procedures (phenol/ chloroform and ethanol precipitation) (Kunkel et al., 1977) and digested with appropriate restriction endonucleases. Digested DNA threads were separated by agarose gel electrophoresis and transferred to Zeta probe membranes (Biorad,Richmond, CA) using alkaline blotting (0.4 M NaOH/O.6 M NaCl), a modification of the Southern procedure (Southern, 1975). The blots were hybridized with 32P-labelledprobes, prepared according to the random oligolabelling method (Feinberg and Vogelstein, 1983),and autoradiographed for 1-14 days at - 70°C using intensifying screens. The samples were scored for LOH and gene amplification, after adjustment of the DNA loaded, by comparing the hybridization signals from the same blots using a control probe. Quantitation of the autoradiograms was performed by computer-aided image processing (Kontron, Munich, Federal Republic of Germany), using a Grundig FA76 video camera with a Pasecon tube and was based on integrated optical density using local background correction in a halo around the band. A reduction after adjustment for loading differences of more than 50% of a band, compared with the signal in blood, was recorded as an allelic loss. Allele losses of TP53 were also scored using polymerase chain reaction (PCR) amplification of exon 4, followed by digestion with AccII, which detects a polymorphic site (Am et al., 1990). Analysis of gene amplification of E B B 2 was according to B$rresen et al. (1990a).
Pathology
Probes
Tumour tissue was obtained from each patient at surgery. The samples included 59 primary carcinomas, three axillary lymph node metastases, seven locoregional recurrences, and one distant metastasis. Two axillary metastases and one primary tumour were obtained from the same patient. One part of the tumour sample was immediately frozen and stored in liquid nitrogen for DNA analysis and immunocytochemical/biochemical determinations of steroid receptors. Formalin-fixed material from each case was processed for light microscopy and classified by the pathologist U.M.N.) according to WHO recommendations. Each case was then TNM classified according to the UICC guidelines (version 1982).
The non gene-specific VNTR markers used were pYNH24 (D2S44, using MspI/HpaII), pEFD64.1 (D3S42, using TaqI), p144D6 (D17S34, using RsaI), pYNZ22.1 (D17S5, using TaqI), pTHH59 (D17S4, using TaqI), and pRMU3 (D17S24, using TaqI). These probes were kindly supplied to us by Drs. Y. Nakamura and R. White (Salt Lake City, Utah), and their regional chromosome localizations have been published and updated at the HGM 10 (1989). The gene-specificprobes used were pGDaI ( H M S ) detecting a VNTR polymorphism (using MspIIHpaII) provided by Dr. T.G. Krontiris (Boston, Massachusetts) (Krontiris et al., 1987) and p123M1.8 and p68RS2.0, both located within the R B I gene (detectinga BamH polymorphism and a VNTR using RsaI), provided by
MATERIALS AND METHODS
Patients
115
LOSS OF HETEROZYGOSITY IN BREAST CARCINOMAS
Figure I. Constitutional and tumour genotypes of some of the markers investigated for loss of heterozygotity (LOH), illustrating LOH of pBHP53 (a), LOH of p68RS2.0 (b), LOH of pYNZ22.I (c), duplication of pYNZ22.I (d). and LOH and duplication of pEFD64.I (e). The size of each allele in kilobases is indicated t o the left of each autoradiogram. L, lymphocyte DNA; T, tumour DNA.
Dr. R.T. Dryja (Boston, Massachusetts) (Wiggs et al., 1988).The pbeta8-2probe (CRYB1)(Ms$I/H#aII) was from the ATCC repository. The pbetaHP53 probe (BamHI) was derived from a cosmid library of genomic TP53 and was kindly provided by Dr. B. Hdyheim (Salt Lake City, Utah) (Hdyheim et al., 1989).The NM23H1 probe detecting a BgAI polymorphism was provided by Dr. P.S. Steeg (Bethesda, Maryland) (Rosengard et al., 1989).For ERBBZ amplification analysis, the ERBBZ cDNA probe (using EcoRI) provided by Dr. T. Yamamoto (Tokyo, Japan) was used (Yamamoto et al., 1986). Oestrogen and Progesterone Receptor Determinations
Oestrogen and progesterone receptors were measured by a standard dextran-coated charcoal method or using monoclonal antibodies in an enzyme immunoassay for quantitative measurement (Abbott ER and PgR-EIA monoclonal). lmmunostainingof TP53 Protein
Frozen sections from the breast carcinomas were immunostained applying a polyclonal sheep antibody against TP53 (Cambridge Research BiochemicalsOM-11-918). The avidin-biotin-peroxidase complex (ABC) method was used. The sections were treated with 0.3% hydrogen peroxide in methanol for 30 min to block endogenous peroxidase. They were incubated
overnight at 4°C with the specific anti-p53 antibody diluted 1:1,0oO, followed by sequential incubations with biotin-labelledsecondary antibody and then with the ABC. The peroxidase reaction was developed using diaminobenzidineas a chromogen. Positive and negative controls, as well as absorption controls of the primary antibody, gave satisfactory results. Statistical Analyses
All comparisons between groups and/or parameters were performed using Fisher’s exact test. Onetailed P values 2cm
12/20 20123
(60.0) (87.0)
0.047
8/15 14/26
(53.3) (53.8)
NS
411 I 10119
(36.4) (52.6)
NS
13/22 19/21
(59.1) (90.5)
0.0 I 6
7/18 15/23
(38.9) (65.2)
NS
7/13 7/17
(53.8) (4 I .2)
NS
Premenopausal Postmenopausal
911 I 23/32
(81.8) (71.9)
NS
6/13 16/28
(46.2) (57. I )
NS
1/8 13/22
( I 2.5) (59.1)
0.030
Oestrogen receptor Positive Negative
21/29 11/14
(72.4) (78.6)
NS
16/27 6/14
(59.3) (42.9)
NS
8/2 I 619
(38. I ) (66.7)
NS
Progesterone receptor Positive Negative
16/25 15/16
(64.0) (93.8)
0.032
14/24 8/16
(58.3) (50.0)
NS
7/19 711 I
(36.8) (63.6)
NS
Family history Positive Negative
619 27/35
(66.7) (77.1)
NS
317 I8/3 I
(42.9) (58. I)
NS
317 I0122
(42.9) (45.5)
NS
Survival Alive, well Relapse or dead
22/32 1011I
(68.8) (90.9)
NS
13/26 711 I
(50.0) (63.3)
NS
9/22 618
(40.9) (75.0)
NS
Node status
=O ?I
199Oa).When six additional markers on chromosome 17 were investigated, it was found that only four of 11 tumours that had lost both pYNZ22.1 and pRMU3 had lost all other informative markers, indicating loss of the entire chromosome (Fig. 2; patients 32, 34, 83, 84). In one tumour (patient l l l ) , there was evidence of duplication of the whole chromosome. ERBB2 amplification, which was seen in 21% of the primary tumours, was not associated with LOH of any of the other loci investigated, particularly not with any LOH of 17q markers. p53 protein was detected in 13% of the tumours investigated by immunostaining.The relationship between clinical and histopathological parameters and LOH on chromosomes 13q, 17p, and 17q is shown in Tables 4, 5, and 6. Loss of R B I occurred in invasive ductal carcinomas only, and LOH of TP53 was also most frequently lost in ductal carcinomas. R B I and p144D6 were lost more frequently in tumours with diameter larger than 2 cm than in smaller tumours, and p144D6 was most frequently lost in tumours from patients with node positive disease. Postmenopausal disease was associated with LOH of TP53 and with LOH of the three 17q markers
NM23H1, pTHH59, and pRMu3. LOH of p144D6 occurred most frequently in progesterone receptor-negative tumours and pTHH59 most frequently in oestrogen receptor-negative tumours. Neither positive family history of breast cancer nor frequency of recurrences was associated with LOH of any of the markers investigated. DISCUSSION
The pattern and frequencies of LOH observed in the present study in general are in good agreement with those described by others (Lundberget al., 1987; Mackay et al., 1988; Devilee et al., 1989 Coles et al., 1990; Cropp et a]., 1990, Sat0 et al., 1990). We find frequent LOH from 17p (>70%) and 13q (40%). There is also a significant number of losses of 17q markers (35%). The observed frequency of LOH of NM23H1 (26%) is considerably lower than the 64% described by Leone et al. (1991). This discrepancy may be explained by different age distributions in the two populations studied, since LOH of NM23Hl seems to be more frequent in postmenopausal disease. Another explanation may be that the existence of two
119
LOSS OF HETEROZYGOSITY IN BREAST CARCINOMAS
TABLE 6. Loss of 17q Alleles in Primary Breast Carcinomas in Relation to Different Clinical and Histopathologic Variables
NM23HI Number with Loss/ Number with Information (‘A)
pTHH59
Significance Level (P)
Number with Loss/ Number with Information (‘A)
pRMU3
Significance Level (P)
Number with Loss/ Number with Information (%)
Significance Level (P)
Histopathologic type Ductal Nonductal
4/15 Y8
(26.7) (25.0)
NS
6/23 3/1 I
(26.1) (27.3)
NS
9/25 y9
(36.0) (22.2)
NS
Tumour size < 2 cm 2 2 cm
3/e 3/15
(37.5) (20.0)
NS
3/15 7/20
(20.0) (35.0)
NS
3/13 9/21
(23.1) (42.9)
NS
4/9 2/14
(44.4) ( 14.3)
NS
5/14 5/21
(35.7) (23.8)
NS
6/14 6/20
(42.9) (30.0)
NS
Premenopausal Postmenopausal
0/8 6/15
(40.0)
0.050
10/25
(40.0)
0.018
1/12 11/22
(8.3) (50.0)
0.017
Oestrogen receptor Positive Negative
2/13 4/10
( I 5.4) (40.0)
NS
4/24 6/l I
(16.7) (54.5)
0.031
6/22 6/12
(27.3) (50.0)
NS
Progesterone receptor Positive Negative
1/1 I 4/10
(9. I) (40.0)
NS
4/20 6/13
(20.0) (46.2)
NS
6/17 6/15
(35.3) (40.0)
NS
Family history Positive Negative
0/3 6/18
(33.3)
NS
1/6 8/27
(16.7) (29.6)
NS
4/9 7/23
(44.5) (30.4)
NS
Survival Alive, well Relapse or dead
5/15 1/7
(33.3) (14.3)
NS
5/22 5/9
(22.7) (55.6)
NS
7/20 5/9
(35.0) (55.6)
NS
Node status
=O ?I
(0)
0/10
(0)
distinct NM23 genes, NM23Hl and NM23H2, may confuse the typings. The probe used in the present work has been shown to crosshybridize to NM23H2 (Stahl et al., 1991)under certain conditions of hybridization. An extra band frequently appears on the autoradiograms,leading to misinterpretation of homozygotes as heterozygotes. This artifact could result in a falsely high LOH frequency. We did not find LOH of 3p and l l p in frequencies higher than the backgiound LOH. This is in contrast to the findings in some previous studies (Theillet et al., 1986; Ali et al., 1987; Mackay et al., 1988). Our results do not support the claimed existence of tumour suppressor genes important for the development of breast cancer in these chromosomal regions. The observation that allele losses occur more frequently in metastases than in primary tumours fits with the concept of tumour progression as a multistep process. The trend was, however, not significant in the small number of metastases investigated. Our results show that loss of the two 17~13.3markers p144D6 and pYNZ22.1 occurs independently of TP53 losses
(0)
(17~13.1).TP53 losses are, in contrast to both pl44D6 and pYNZ22.1 losses, not associated with loss of RBl sequences. These findings indicate the existence of two different genes on 17p involved in breast carcinogenesis, as has been suggested by Coles et al. (1990). The absence of any association observed in LOH between two genes on the same chromosomal arm, and even within the same band, gives reason to urge caution when interpreting results of allelotyping. It seems insdicient to use only one DNA marker from each chromosomal arm. That p144D6 and pYNZ22.1 are the markers most frequently lost makes it reasonable to suspect that loss of a gene located close to these markers is an early step in breast cancer progression. It should be emphasized, however, that LOH of p144D6 occurred in only one of two axillary metastases investigated from the same patient. This indicates that loss of the gene close to p144D6 and pYNZ22 does not represent the very first alteration. Allele losses of pYNZ22.1 have been reported to be significantly correlated with increased TP53 mRNA expression (Thompson et al.,
I20
ANDERSEN E l A L
1990). Accordingly, the gene close to pYNZ22.1 is postulated to regulate the expression of TP53. In the present study all six tumours with TP53 expression had lost p144D6 or pYNZ22.1. Presence of immunoreactive p53 protein was not significantly associated with loss of any of the markers. p53 protein was detected in only 6 of 46 tumours (13%)compared with increased TP53 mRNA in 43 of the 76 tumours (57%) investigated by Thompson et al. (1990). The difference may partly be due to insufficient sensitivity of the antibody used in our immunostaining method. Until recently, TP53 was considered an oncogene. Several studies have concluded that wild-type TP53 does not function as an oncoprotein during in vitro transformation (Finlay et al., 1989; Prosser et al., 1990); only mutated forms have this capacity. Wildtype TP53 protein has tumour suppressor activity. The TP53 and RB1 proteins have similar tumour suppressor phenotypes. The DNA tumour viruses SV40, adenovirus type 5, and human papilloma viruses 16 and 18 have evolved mechanisms to interact with both TP53 and RB1 protein (Levine,1990;Levine and Momand, 1990). The associationsbetween LOH of RBI and p144D6 and between RBI and pYNZ22.1 might indicate that the alterations in the R B I gene and in the gene located close to p144D6 and pYNZ22.1 are biologically dependent on each other. The finding that none of the 20 tumours informative for RBI, TP53, and p144D6 or pYNZ22.1 had lost R B I sequences and retained heterozygosity for the other markers indicates that the R B I gene is lost later in the cascade of genetic events than the two genes located on 17p. Another important candidate gene that may be altered in one of the later steps of carcinogenesis is the ERBBZ protooncogene located on 17q. In a previous study on parts of the present material (Bdrresen et al., 1990a), ERBBZ gene amplification in primary turnours was significantly associated with decreased survival (P = 0.006). The mechanism was suggested to be increased resistance to tumour necrosis factor and the ability to switch off the production of local hormones and receptors involved in growth modulation. Preliminary results indicated that ERBBZ gene amplification was also associated with loss of 17q markers (Bdrresen et al., 1990a,b). This association was not confirmed when the number of tumours investigated for LOH on 17q was extended. NMZ3Hl maps to 17pllLqll by in situ hybridisation and to 17q21 by linkage analysis based on the CEPH (Centre &Etude du Polymorphism human) database families (Leone et al., 1991).Our finding that loss of NMZ3HI is associated with loss of pTHH59 and/or pRMU3 supports the mapping of NMZ3Hl to
17q21. We examined 23 NMZ3Hl informative primary tumours for LOH without finding significant associations with survival, tumour size, or node status. Firm conclusions concerning relation to metastatic potential cannot yet be drawn, since the number of informative cases is small and the observation time short. A positive family history for breast cancer is known to be a significant risk factor (Lynch et al., 1984). The risk is increased by the occurrence of breast cancer in mother or sisters, and, the younger the related cancer patients, the greater the increase in risk. In the present study, a positive family history was not significantly associated with LOH of markers in the tumour suppressor gene regions 3p, llp, 13q, 17p, and 17q. Great progress has recently been made in two areas of familial breast cancer genetics. Chromosome band 17q21 appears to contain a gene for inherited susceptibility to early-onset, site-specific breast cancer (Hall et al., 1990). Genetic analysis yielded a lod score of 5.98 for linkage of breast cancer susceptibility to D17S74 in early-onsetand a negative lod score in late-onset families. In addition, two independent studies (Malkin et al., 1990;Srivastava et al., 1990) found that inherited mutations in TP53 are associated with the high rate of breast and other cancers in families with the Li-Fraumeni syndrome. We have recently shown that germline TP53 mutations also occur in breast carcinoma patients outside classical Li-Fraumeni families (Bqirresen et al., 1991).On the other hand, one recent report on TP53 mutations studied in five early-onset breast cancer families did not identify germline mutations (Prosser et al., 1991). Mapping of the genes responsible for inherited breast cancer may allow the identification of early lesions critical also for the development of sporadic breast cancer in the general population. An effort should be made to identify the gene located on 17q that is responsible for familial early-onset breast cancer. Among the candidate genes are ERBB2 and NM23Hl. Although the sequential steps of the cascade of genetic alterations leading to breast cancer are now gradually being identified, much work remains before the complete sequence of events is known. It will be of great importance to separate alterations predicting increased metastatic potential from alterations reflecting only genetic instability. Hopefully this knowledge will turn out to be prognostically useful and help us devise new strategies for breast carcinoma therapy. ACKNOWLEDGMENTS
We thank Kirsten Lycke and Martina Skrede for excellent technical assistance. This work was sup-
LOSS OF HETEROZYGOSITY IN BREAST CARCINOMAS
ported by grants from Torsteds Legacy and Blix Legacy. T.I.A. is a research fellow of The Norwegian Cancer Society. REFERENCES Ali IU, Lidereau R, Theillet C, Callahan R (1987) Reduction to homozygosity of genes on chromosome 11 in human breast neoplasia. Science 238185188. Ara S, Lee PSY, Hansen MF, Saya H (1990) Codon 72 polymorphism of the TP53 gene. Nucleic Acids Res 184961. Bevilacqua G, Sobel ME, Liotta LA, Steeg PS (1989) Association of low am23 RNA levels in human primary infiltrating ductal breast carcinomas with lymph node involvement and other histopathological indicators of high metastatic potential. Cancer Res 4951855190. m e s e n A-L, Ottestad L, Gaustad A, Andersen TI, Heikkila R, Jahnsen T, Tveit KM, Nesland JM (1990a) Amplification and protein overexpression of the neu/HER-Z/c-erbB-2protooncogene in human breast carcinomas: relationship to loss of gene sequences on chromosome 17, family history and prognosis. Br J Cancer 62:5&590. ryrreSen A-L, Ottestad L, Gaustad A, Nesland J, Heikkila R (1990b)Gene amplification of the neu oncogene in breast cancer tumors; correlation to loss of gene sequences on chromosome 17q. Abstract, ESHG meeting. Clin Genet 37292. w e s e n A-L, Hovig E, Smith-%ensen B, Lystad S, Brbgger A (1991) Mutation analysis in human cancers using PCR and constant denaturant gel electrophoresis (CDGE). Adv Mol Gen 4&71. Cavenee WK, Dryja TP, Phillips RA, Benedict WF, Godbout R, Gallie BL, Murphree AL, Strong LC, White RL (1983) Expression of recessive alleles by chromosomal mechanisms in retinoblastoma. Nature 305779-781. Chen L-C, Dollbaum C, Smith HS (1989) Loss of heterozygozity on chromosome l q in human breast cancer. Proc Natl Acad Sci USA 867204-7207. Coles C. Thompson AM, Elder PA, Cohen BB, Mackenzie IM, Cranston G, Chetty U, Mackay J, Macdonald M, Nakamura Y, Hbyheim B, Steel CM (1990) Evidence implicating at least two genes on chromosome 17p in breast carcinogenesis. Lancet 336761-763. Cropp CS,Lidereau R, Campbell G, Champene MH, Callahan R (1990) Loss of heterozygosity on chromosomes 17 and 18 in breast carcinoma: Two additional regions identified. Proc Natl Acad Sci USA 8717737-7741. Devilee P, van den Broek M, Kuipers-DijkshoornN, Kolluri R, Khan PM, Pearson PL, CornelisseCJ (1989) At least four different chromosomal regions are involved in loss of heterozygosity in human breast carcinoma. Genomics 5:554-560. Erlandsson R, Bergerheim USR, Boldog F, Marcsek Z, Kunimi K, Lin BY-T, Ingvarsson S, Castresana JS, Lee W-H, Lee E, Klein G, Sumegi J (1990) A gene near the D3F15S2 site on 3p is expressed in normal human kidney but not or at a severely reduced level in 11 of 15 primary renal cell carcinomas (RCC). Oncogene 51207-1211. Fearon ER, Cho KR, Nigro JM, Kern SE, Smons JW,Ruppert JM, Hamilton SR, Preisinger AC, Thomas G, Kinzler KW, Vogelstein B (1990) Identification of a chromosome 1% gene that is altered in colorectal cancers. Science 2474%55. Feinberg AP, Vogelstein B (1983) A technique for radiolabelling DNA restriction endonuclease fragments to high specific activity. Anal Bicchem 132G13. Finlay CA, Hinds PW, Levine AJ (1989) The p53 protooncogene can act as a suppressor of transformation. Cell 571085-1093. Friend SH, Bemards R, Rogelj S, Weinberg RA, Rapaport JM, Albert DM, Dryla TP (1986) A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature 323643-646. Green AR (1988) Recessive mechanisms of malignancy. Br J Cancer 58:115127. Haber DA, Buckler AJ, Glaser T, Call KM, Pelletier J, SohnRL, Douglas EC, Housman DE (1990) An internal deletion within an llp13 zinc finger gene contributes to the development of Wilms’ tumor. Cell 61:1257-1269. Hall JM, Lee MK, Newman B, Morrow JE, Anderson LA, Huey B, King M-C (1990)Linkage of early-onset familial breast cancer to chromosome 17q21. Science 250:1684-1689.
121
HGM 10 (1989) Tenth international workshop on human gene mapping. Cytogenet Cell Genet 51:622-947. Hbyheim B, Nakamura Y, White R (1989) A BamHl-polymorphism is detected by a genomic p53-clone (pBHP53). Nucleic Acids Res 1 7 8898. Krontiris TG, Dimartino NA, Mitcheson HD, Lonergan JA, Ekgg C, Parkinson DR (1987) Human hypervariable sequences in risk assessment: Rare Hams alleles in cancer patients. Environ Health Perspect 76147-153. Kunkel LM, Smith KD, Boyer SH, Borgaonkar DS, Wachtel SS, Miller OJ, Breg WR, Jones Jr HW,Rary JM (1977) Analysis of human Y-chromosome-specific reiterated DNA in chromosome variants. Roc Natl Acad Sci USA 7412451249. Levine AJ (1990) The p53 protein and its interactions with the oncogene products of the small DNA tumor viruses. Virology 177:41%426. Levine AJ, Momand J (1990) Tumor suppressor genes: The p53 and retinoblastoma sensitivity genes and gene products. Biochim Biophys Acta 1032119-136. Leone A, McBride OW, Weston A, Wang MG, Anglard P, Cropp CS, Goepel JR, Lidereau R, Callahan R, Linehan WM, Rees RC, Harris CC, Liotta LA. Steeg PS (1991)Somatic allelic deletions of NMz3 in human cancer. Cancer Res 51:2490-2493. Lundberg C, Skoog L, Cavenee WK, Nordenskjold M (1987) Loss of heterozygosity in human ductal breast tumors indicates a recessive mutation on chromosome 13.Proc Natl Acad Sci USA 84:2372-2376. Lynch HT, Albano WA, Danes BS,Layton MA, Kmberling WJ, Lynch JF, Cheng SC,Costello KA, Mulcahy GM, Wagner CA, Tindall SL (1984) Genetic predisposition to breast cancer. Cancer 53612422, Mackay J, Steel CM, Elder PA, Forrest APM, Evans HJ (1988) Allele loss on short arm of chromosome 17 in breast cancers. Lancet 213841385. Malkin D, Li FP, Strong LC, Fraumeni Jr JF, Nelson CE, Kim DH, Kassel J, Gryka MA, Bischoff FZ, Tainsky MA, Friend SH (1990) Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms. Science 255:123%1238. Prosser J, Thompson AM, Cranston G, Evans HJ(1990) Evidence that p53 behaves as a tumour suppressor gene in sporadic breast tum o m . Oncogene 5157F1579. Prosser J, Elder PA, Condie A, MacFadyen I, Steel CM, Evans HJ (1991) Mutations in p53 do not account for heritable breast cancer: a study in five affected families. Br J Cancer 63181-184. Rosengard AM, Krutzsch HC, Shearn A, Biggs JR,Barker E, Marguiles LMK,King CR, Liotta LA, Steeg PS (1989) Reduced Nm23/Awd protein in tumor metastasis and aberrant Dvosophila development. Nature 342177-180. Sat0 T, Tanigami A, Yamakawa K, Akiyama F, Kasumi F, Sakamoto G, Nakamura Y (1990) Allelotype of breast cancer: cumulative allele losses promote tumor progression in primary breast cancer. Cancer R e 50.7184-7189. Southern EM (1975) Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98505-517. Srivastava S, Zou Z, Pirollo K, Blattner W, Chang EH (1990) Germline transmission of a mutated p53 gene in a cancer-prone family with Li-Fraumeni syndrome. Nature 348:747-749. Stahl JA, Leone A, Rosengard AM, Porter L, King CR, Steeg F5 (1991) Identification of a second human nm23 gene, nmZ3-HZ. Cancer Res 51:44M9. Steeg PS, Bevilacqua G, Pozzatti R, Liotta LA, Sobel ME (1988) Altered expression of NMZ3, a gene associated with low tumor metastatic potential, during adenovirus 2 Ela inhibition of experimental metastasis. Cancer Res 4865504554. Theillet C, Lidereau R, Escot C, Hutzell P, Brunet M, Gest J, Schlon J, Callahan R (1986) Loss of c-H-vas-1 allele and aggressive human primary breast carcinoma. Cancer Res fi4776-4781. Thompson AM, Steel CM, Chetty U, Hawkins RA, Miller WR, Carter DC, Forrest APM, Evans HJ (1990)p53 gene mRNA expression and chromosome 17p allele loss in breast cancer. Br J Cancer 61:7478. Wiggs J, Nordenskjold M, Yandell D, Rapaport J, Grondm V, Janson M, Werelius B, Petersen R, Craft A, Riedel K, Liberfarb R, Walton D, Wilson W, Dryja T P (1988) Prediction of risk of hereditary retinoblastoma, using DNA polymorphisms within the retinoblastoma gene. N Engl J Med 318:151-157. Yamamoto T, Ikawara S, Akiyama T, Semba K. Nomura M, Miyajima N, Saito T, Toyoshima K (1986)Similarity of protein encoded by the human c-erb-B-2 gene to epidermal growth factor receptor. Nature 31923CL234.
Genetic Alterations of the Tumour Suppressor Gene Regions 3p, I Ip, 13q, 17p, and 17q in Human Breast Carcinomas l o n e 1. Andersen, Astrid Gaustad, Lars Ottestad, George W. Farrants, JahnM. Nesland, Kjell M. Tveit, and Anne-Lise Bgrresen Department of Genetics, Institute for Cancer Research (T.I.A., A.G., A.-L.B.), Department of Clinical Oncology (L.O., K.M.T.) and Department of Pathology, Institute for Cancer Research (G.W.F., J.M.N.), The Norwegian Radium Hospital, Montebello, Oslo, Norway
Fifty-nine primary breast carcinomas and I I metastases were examined t o identify genetic alterations in the tumour suppressor gene regions 3p, I Ip, I3q, I7p, and I7q. Loss of heterozygosity (LOH) was frequently observed on chromosome arms I7p (p 144D6 lost in 75%, pYNZ22. I in 55%, and TP53 in 48% of the primary tumours), 13q (RBI lost in 40% of the primary tumours), and 17q (pRMU3 lost in 35%. pTHH59 in 29%, and NM23H I in 26% of the primary tumours). Loss of all the markers except p144D6 w a s observed even more frequently in the metastases. Pairwise comparisons for concordance of allele losses on I7p indicated that there might be two genes on I7p implicated in breast cancer development; the TP53 gene and a gene located close t o the p144D6 and pYNZ22.I markers. LOH of the RBI gene w a s associated with LOH of pYNZ22.I and p144D6, but not with LOH of TP53. LOH of RBI and TP53 was associated with occurrence of ductal carcinomas, RBI and p 144D6 losses with tumour size, and p I44D6 losses with positive node status as well. LOH of TP53 and the three 17q markers NM23H I, pTHH59, and pRMU3 was most frequently observed in tumours from postmenopausal women. p144D6 losses occurred most frequently in progesterone receptor-negative tumours, whereas pTHH59 losses occurred most frequently in oestrogen receptor-negativetumours. LOH of the investigated loci was not associated with E R B M protooncogene amplification, with positive Family history of breast cancer, o r with survival. Genes Chrom Cancer 4:113-121 (1992). INTRODUCTION
It is increasingly recognized that loss of function mutations in critical genes, tumour suppressor genes, may play an important role in the pathogenesis of many malignancies (Green, 1988). In patients with retinoblastoma, loss or inactivation of both alleles of one such gene is associated with twnorigenesis (Cavenee et al., 1983; Friend et al., 1986). Tumour suppressor genes are normal genes whose products regulate cellular growth and differentiation. Through this action, they play an important role as inhibitors of the uncontrolled cellular proliferation characteristic of cancer. Several sites of allele losses have been reported in human breast carcinomas. The following chromosome arms have been found to be involved l q in l3-26% of primary tumours, 3p in l9-46%, l l p in 10-27%, 13q in 21-40%, 16q in &YO,17p in M 1 % , 17q in 3 W % and 1% in 2149% (Theillet et al., 1986, Ali et al., 1987;Lundberg et al., 1987; Mackay et al., 1988; Chen et al., 1989 Devilee et al., 1989; Bdrresen et al., 199oa; Coles et al., 1990; Cropp et al., 1990; Sat0 et al., 1990). Known putative tumour suppressor genes are located in some of these chromosomal regions; the RIK gene on 3p (Erlandsson et al., 1990), the Wilms’ tumour suppressor gene on l l p (Haber et al., 1990),the retinoblastoma (RBI) gene on 0 1992 WILEY-LISS, INC.
13q (Cavenee et al., 1983; Friend et al., 1986), and the deleted in colon cancer (DCC) gene on 18q (Fearon et al., 1990). On the short arm of chromosome 17, two distinct regions of loss of heterozygosity (LO@ have been identified, indicating that at least two genes on 17p are involved in breast carcinogenesis; the tumour suppressor gene TP53 and a gene suggested to regulate the expression of TP53, located telomeric to the TP53 gene (Mackay et al., 1988, Coles et al., 1990). In a previous paper that included preliminary results of the present study, we were the first to describe an association between LOH of 13q and 17p sequences (Wrresen et al., 1990a).These findings were confirmed by Sat0 et al. (1990),who further suggested a sequence of genetic alterations during tumour progression in breast cancer, with loss of 17p sequences, including TP53 as the first step. Loss of 13q material involving the R B I gene and/or amplification of the protooncogeneEmB2 on 17q were claimed to be later alterations in the cascade of genetic events. Loss of chromosome arm 1%was suggested to play a significant role in the development of lymph node metastases. Another putative metastatic suppressor gene, Received May 13, 1991; accepted August 5, 1991. Address reprint requests to Anne-Lise Wrresen, Department of Pathology, Institute for Cancer Research, The Norwegian Radium Hospital, Montebello, 0310 Oslo 3, Norway.
I14
ANDERSEN ET A L
NM23H1, has also been claimed to be involved in breast cancer (Steeg et al., 1988). This gene has been shown to be highly expressed in nonmetastatic, but not in metastatic, breast tumours (Bevilacqua et al., 1989). The aims of the present study were 1)to identify and determine the frequencies of losses of constitutional alleles in the tumour suppressor gene regions on chromosome arms 3p, Up, 13q,17p, and 17q; 2) to confirm our previous finding that LOH on 13q is associated with LOH of 17p sequences; 3) to study whether LOH of the regions investigated was associated with amplification of ERBB2; and 4) to study whether the LOH pattern is associated with particular clinical and histopathological parameters, a family history of breast cancer, or survival.
Blood Samples
Peripheral venous blood (10-20 ml) was collected in EDTA and stored at - 40°C until isolation of DNA. DNA Analyses
Material for this study was obtained from 68 breast carcinoma patients admitted to the Norwegian Radium Hospital in the period 1987-1989. All, except for one 66-year-oldmale, were female Norwegian whites, with a median age at diagnosis of 58 years, ranging from 31.0 to 80.5 years. Survival data were available for 63 patients. Mean observation time for these patients was 32 months (14-154 months). Eighteen patients had premenopausal and 43 had postmenopausal disease. Menopausal status was unknown for six patients. None had received treatment prior to surgery. A family history of breast cancer was obtained from patients or relatives. All cancer diagnoses reported were confirmed by the Norwegian Cancer Registry. A family history was considered positive if at least one first-degree relative had suffered from breast cancer. Families among whom no first-degree relative had reached the age of 35 years were not scored.
DNA was isolated from cell nuclei of tumour tissue and whole blood using standard procedures (phenol/ chloroform and ethanol precipitation) (Kunkel et al., 1977) and digested with appropriate restriction endonucleases. Digested DNA threads were separated by agarose gel electrophoresis and transferred to Zeta probe membranes (Biorad,Richmond, CA) using alkaline blotting (0.4 M NaOH/O.6 M NaCl), a modification of the Southern procedure (Southern, 1975). The blots were hybridized with 32P-labelledprobes, prepared according to the random oligolabelling method (Feinberg and Vogelstein, 1983),and autoradiographed for 1-14 days at - 70°C using intensifying screens. The samples were scored for LOH and gene amplification, after adjustment of the DNA loaded, by comparing the hybridization signals from the same blots using a control probe. Quantitation of the autoradiograms was performed by computer-aided image processing (Kontron, Munich, Federal Republic of Germany), using a Grundig FA76 video camera with a Pasecon tube and was based on integrated optical density using local background correction in a halo around the band. A reduction after adjustment for loading differences of more than 50% of a band, compared with the signal in blood, was recorded as an allelic loss. Allele losses of TP53 were also scored using polymerase chain reaction (PCR) amplification of exon 4, followed by digestion with AccII, which detects a polymorphic site (Am et al., 1990). Analysis of gene amplification of E B B 2 was according to B$rresen et al. (1990a).
Pathology
Probes
Tumour tissue was obtained from each patient at surgery. The samples included 59 primary carcinomas, three axillary lymph node metastases, seven locoregional recurrences, and one distant metastasis. Two axillary metastases and one primary tumour were obtained from the same patient. One part of the tumour sample was immediately frozen and stored in liquid nitrogen for DNA analysis and immunocytochemical/biochemical determinations of steroid receptors. Formalin-fixed material from each case was processed for light microscopy and classified by the pathologist U.M.N.) according to WHO recommendations. Each case was then TNM classified according to the UICC guidelines (version 1982).
The non gene-specific VNTR markers used were pYNH24 (D2S44, using MspI/HpaII), pEFD64.1 (D3S42, using TaqI), p144D6 (D17S34, using RsaI), pYNZ22.1 (D17S5, using TaqI), pTHH59 (D17S4, using TaqI), and pRMU3 (D17S24, using TaqI). These probes were kindly supplied to us by Drs. Y. Nakamura and R. White (Salt Lake City, Utah), and their regional chromosome localizations have been published and updated at the HGM 10 (1989). The gene-specificprobes used were pGDaI ( H M S ) detecting a VNTR polymorphism (using MspIIHpaII) provided by Dr. T.G. Krontiris (Boston, Massachusetts) (Krontiris et al., 1987) and p123M1.8 and p68RS2.0, both located within the R B I gene (detectinga BamH polymorphism and a VNTR using RsaI), provided by
MATERIALS AND METHODS
Patients
115
LOSS OF HETEROZYGOSITY IN BREAST CARCINOMAS
Figure I. Constitutional and tumour genotypes of some of the markers investigated for loss of heterozygotity (LOH), illustrating LOH of pBHP53 (a), LOH of p68RS2.0 (b), LOH of pYNZ22.I (c), duplication of pYNZ22.I (d). and LOH and duplication of pEFD64.I (e). The size of each allele in kilobases is indicated t o the left of each autoradiogram. L, lymphocyte DNA; T, tumour DNA.
Dr. R.T. Dryja (Boston, Massachusetts) (Wiggs et al., 1988).The pbeta8-2probe (CRYB1)(Ms$I/H#aII) was from the ATCC repository. The pbetaHP53 probe (BamHI) was derived from a cosmid library of genomic TP53 and was kindly provided by Dr. B. Hdyheim (Salt Lake City, Utah) (Hdyheim et al., 1989).The NM23H1 probe detecting a BgAI polymorphism was provided by Dr. P.S. Steeg (Bethesda, Maryland) (Rosengard et al., 1989).For ERBBZ amplification analysis, the ERBBZ cDNA probe (using EcoRI) provided by Dr. T. Yamamoto (Tokyo, Japan) was used (Yamamoto et al., 1986). Oestrogen and Progesterone Receptor Determinations
Oestrogen and progesterone receptors were measured by a standard dextran-coated charcoal method or using monoclonal antibodies in an enzyme immunoassay for quantitative measurement (Abbott ER and PgR-EIA monoclonal). lmmunostainingof TP53 Protein
Frozen sections from the breast carcinomas were immunostained applying a polyclonal sheep antibody against TP53 (Cambridge Research BiochemicalsOM-11-918). The avidin-biotin-peroxidase complex (ABC) method was used. The sections were treated with 0.3% hydrogen peroxide in methanol for 30 min to block endogenous peroxidase. They were incubated
overnight at 4°C with the specific anti-p53 antibody diluted 1:1,0oO, followed by sequential incubations with biotin-labelledsecondary antibody and then with the ABC. The peroxidase reaction was developed using diaminobenzidineas a chromogen. Positive and negative controls, as well as absorption controls of the primary antibody, gave satisfactory results. Statistical Analyses
All comparisons between groups and/or parameters were performed using Fisher’s exact test. Onetailed P values 2cm
12/20 20123
(60.0) (87.0)
0.047
8/15 14/26
(53.3) (53.8)
NS
411 I 10119
(36.4) (52.6)
NS
13/22 19/21
(59.1) (90.5)
0.0 I 6
7/18 15/23
(38.9) (65.2)
NS
7/13 7/17
(53.8) (4 I .2)
NS
Premenopausal Postmenopausal
911 I 23/32
(81.8) (71.9)
NS
6/13 16/28
(46.2) (57. I )
NS
1/8 13/22
( I 2.5) (59.1)
0.030
Oestrogen receptor Positive Negative
21/29 11/14
(72.4) (78.6)
NS
16/27 6/14
(59.3) (42.9)
NS
8/2 I 619
(38. I ) (66.7)
NS
Progesterone receptor Positive Negative
16/25 15/16
(64.0) (93.8)
0.032
14/24 8/16
(58.3) (50.0)
NS
7/19 711 I
(36.8) (63.6)
NS
Family history Positive Negative
619 27/35
(66.7) (77.1)
NS
317 I8/3 I
(42.9) (58. I)
NS
317 I0122
(42.9) (45.5)
NS
Survival Alive, well Relapse or dead
22/32 1011I
(68.8) (90.9)
NS
13/26 711 I
(50.0) (63.3)
NS
9/22 618
(40.9) (75.0)
NS
Node status
=O ?I
199Oa).When six additional markers on chromosome 17 were investigated, it was found that only four of 11 tumours that had lost both pYNZ22.1 and pRMU3 had lost all other informative markers, indicating loss of the entire chromosome (Fig. 2; patients 32, 34, 83, 84). In one tumour (patient l l l ) , there was evidence of duplication of the whole chromosome. ERBB2 amplification, which was seen in 21% of the primary tumours, was not associated with LOH of any of the other loci investigated, particularly not with any LOH of 17q markers. p53 protein was detected in 13% of the tumours investigated by immunostaining.The relationship between clinical and histopathological parameters and LOH on chromosomes 13q, 17p, and 17q is shown in Tables 4, 5, and 6. Loss of R B I occurred in invasive ductal carcinomas only, and LOH of TP53 was also most frequently lost in ductal carcinomas. R B I and p144D6 were lost more frequently in tumours with diameter larger than 2 cm than in smaller tumours, and p144D6 was most frequently lost in tumours from patients with node positive disease. Postmenopausal disease was associated with LOH of TP53 and with LOH of the three 17q markers
NM23H1, pTHH59, and pRMu3. LOH of p144D6 occurred most frequently in progesterone receptor-negative tumours and pTHH59 most frequently in oestrogen receptor-negative tumours. Neither positive family history of breast cancer nor frequency of recurrences was associated with LOH of any of the markers investigated. DISCUSSION
The pattern and frequencies of LOH observed in the present study in general are in good agreement with those described by others (Lundberget al., 1987; Mackay et al., 1988; Devilee et al., 1989 Coles et al., 1990; Cropp et a]., 1990, Sat0 et al., 1990). We find frequent LOH from 17p (>70%) and 13q (40%). There is also a significant number of losses of 17q markers (35%). The observed frequency of LOH of NM23H1 (26%) is considerably lower than the 64% described by Leone et al. (1991). This discrepancy may be explained by different age distributions in the two populations studied, since LOH of NM23Hl seems to be more frequent in postmenopausal disease. Another explanation may be that the existence of two
119
LOSS OF HETEROZYGOSITY IN BREAST CARCINOMAS
TABLE 6. Loss of 17q Alleles in Primary Breast Carcinomas in Relation to Different Clinical and Histopathologic Variables
NM23HI Number with Loss/ Number with Information (‘A)
pTHH59
Significance Level (P)
Number with Loss/ Number with Information (‘A)
pRMU3
Significance Level (P)
Number with Loss/ Number with Information (%)
Significance Level (P)
Histopathologic type Ductal Nonductal
4/15 Y8
(26.7) (25.0)
NS
6/23 3/1 I
(26.1) (27.3)
NS
9/25 y9
(36.0) (22.2)
NS
Tumour size < 2 cm 2 2 cm
3/e 3/15
(37.5) (20.0)
NS
3/15 7/20
(20.0) (35.0)
NS
3/13 9/21
(23.1) (42.9)
NS
4/9 2/14
(44.4) ( 14.3)
NS
5/14 5/21
(35.7) (23.8)
NS
6/14 6/20
(42.9) (30.0)
NS
Premenopausal Postmenopausal
0/8 6/15
(40.0)
0.050
10/25
(40.0)
0.018
1/12 11/22
(8.3) (50.0)
0.017
Oestrogen receptor Positive Negative
2/13 4/10
( I 5.4) (40.0)
NS
4/24 6/l I
(16.7) (54.5)
0.031
6/22 6/12
(27.3) (50.0)
NS
Progesterone receptor Positive Negative
1/1 I 4/10
(9. I) (40.0)
NS
4/20 6/13
(20.0) (46.2)
NS
6/17 6/15
(35.3) (40.0)
NS
Family history Positive Negative
0/3 6/18
(33.3)
NS
1/6 8/27
(16.7) (29.6)
NS
4/9 7/23
(44.5) (30.4)
NS
Survival Alive, well Relapse or dead
5/15 1/7
(33.3) (14.3)
NS
5/22 5/9
(22.7) (55.6)
NS
7/20 5/9
(35.0) (55.6)
NS
Node status
=O ?I
(0)
0/10
(0)
distinct NM23 genes, NM23Hl and NM23H2, may confuse the typings. The probe used in the present work has been shown to crosshybridize to NM23H2 (Stahl et al., 1991)under certain conditions of hybridization. An extra band frequently appears on the autoradiograms,leading to misinterpretation of homozygotes as heterozygotes. This artifact could result in a falsely high LOH frequency. We did not find LOH of 3p and l l p in frequencies higher than the backgiound LOH. This is in contrast to the findings in some previous studies (Theillet et al., 1986; Ali et al., 1987; Mackay et al., 1988). Our results do not support the claimed existence of tumour suppressor genes important for the development of breast cancer in these chromosomal regions. The observation that allele losses occur more frequently in metastases than in primary tumours fits with the concept of tumour progression as a multistep process. The trend was, however, not significant in the small number of metastases investigated. Our results show that loss of the two 17~13.3markers p144D6 and pYNZ22.1 occurs independently of TP53 losses
(0)
(17~13.1).TP53 losses are, in contrast to both pl44D6 and pYNZ22.1 losses, not associated with loss of RBl sequences. These findings indicate the existence of two different genes on 17p involved in breast carcinogenesis, as has been suggested by Coles et al. (1990). The absence of any association observed in LOH between two genes on the same chromosomal arm, and even within the same band, gives reason to urge caution when interpreting results of allelotyping. It seems insdicient to use only one DNA marker from each chromosomal arm. That p144D6 and pYNZ22.1 are the markers most frequently lost makes it reasonable to suspect that loss of a gene located close to these markers is an early step in breast cancer progression. It should be emphasized, however, that LOH of p144D6 occurred in only one of two axillary metastases investigated from the same patient. This indicates that loss of the gene close to p144D6 and pYNZ22 does not represent the very first alteration. Allele losses of pYNZ22.1 have been reported to be significantly correlated with increased TP53 mRNA expression (Thompson et al.,
I20
ANDERSEN E l A L
1990). Accordingly, the gene close to pYNZ22.1 is postulated to regulate the expression of TP53. In the present study all six tumours with TP53 expression had lost p144D6 or pYNZ22.1. Presence of immunoreactive p53 protein was not significantly associated with loss of any of the markers. p53 protein was detected in only 6 of 46 tumours (13%)compared with increased TP53 mRNA in 43 of the 76 tumours (57%) investigated by Thompson et al. (1990). The difference may partly be due to insufficient sensitivity of the antibody used in our immunostaining method. Until recently, TP53 was considered an oncogene. Several studies have concluded that wild-type TP53 does not function as an oncoprotein during in vitro transformation (Finlay et al., 1989; Prosser et al., 1990); only mutated forms have this capacity. Wildtype TP53 protein has tumour suppressor activity. The TP53 and RB1 proteins have similar tumour suppressor phenotypes. The DNA tumour viruses SV40, adenovirus type 5, and human papilloma viruses 16 and 18 have evolved mechanisms to interact with both TP53 and RB1 protein (Levine,1990;Levine and Momand, 1990). The associationsbetween LOH of RBI and p144D6 and between RBI and pYNZ22.1 might indicate that the alterations in the R B I gene and in the gene located close to p144D6 and pYNZ22.1 are biologically dependent on each other. The finding that none of the 20 tumours informative for RBI, TP53, and p144D6 or pYNZ22.1 had lost R B I sequences and retained heterozygosity for the other markers indicates that the R B I gene is lost later in the cascade of genetic events than the two genes located on 17p. Another important candidate gene that may be altered in one of the later steps of carcinogenesis is the ERBBZ protooncogene located on 17q. In a previous study on parts of the present material (Bdrresen et al., 1990a), ERBBZ gene amplification in primary turnours was significantly associated with decreased survival (P = 0.006). The mechanism was suggested to be increased resistance to tumour necrosis factor and the ability to switch off the production of local hormones and receptors involved in growth modulation. Preliminary results indicated that ERBBZ gene amplification was also associated with loss of 17q markers (Bdrresen et al., 1990a,b). This association was not confirmed when the number of tumours investigated for LOH on 17q was extended. NMZ3Hl maps to 17pllLqll by in situ hybridisation and to 17q21 by linkage analysis based on the CEPH (Centre &Etude du Polymorphism human) database families (Leone et al., 1991).Our finding that loss of NMZ3HI is associated with loss of pTHH59 and/or pRMU3 supports the mapping of NMZ3Hl to
17q21. We examined 23 NMZ3Hl informative primary tumours for LOH without finding significant associations with survival, tumour size, or node status. Firm conclusions concerning relation to metastatic potential cannot yet be drawn, since the number of informative cases is small and the observation time short. A positive family history for breast cancer is known to be a significant risk factor (Lynch et al., 1984). The risk is increased by the occurrence of breast cancer in mother or sisters, and, the younger the related cancer patients, the greater the increase in risk. In the present study, a positive family history was not significantly associated with LOH of markers in the tumour suppressor gene regions 3p, llp, 13q, 17p, and 17q. Great progress has recently been made in two areas of familial breast cancer genetics. Chromosome band 17q21 appears to contain a gene for inherited susceptibility to early-onset, site-specific breast cancer (Hall et al., 1990). Genetic analysis yielded a lod score of 5.98 for linkage of breast cancer susceptibility to D17S74 in early-onsetand a negative lod score in late-onset families. In addition, two independent studies (Malkin et al., 1990;Srivastava et al., 1990) found that inherited mutations in TP53 are associated with the high rate of breast and other cancers in families with the Li-Fraumeni syndrome. We have recently shown that germline TP53 mutations also occur in breast carcinoma patients outside classical Li-Fraumeni families (Bqirresen et al., 1991).On the other hand, one recent report on TP53 mutations studied in five early-onset breast cancer families did not identify germline mutations (Prosser et al., 1991). Mapping of the genes responsible for inherited breast cancer may allow the identification of early lesions critical also for the development of sporadic breast cancer in the general population. An effort should be made to identify the gene located on 17q that is responsible for familial early-onset breast cancer. Among the candidate genes are ERBB2 and NM23Hl. Although the sequential steps of the cascade of genetic alterations leading to breast cancer are now gradually being identified, much work remains before the complete sequence of events is known. It will be of great importance to separate alterations predicting increased metastatic potential from alterations reflecting only genetic instability. Hopefully this knowledge will turn out to be prognostically useful and help us devise new strategies for breast carcinoma therapy. ACKNOWLEDGMENTS
We thank Kirsten Lycke and Martina Skrede for excellent technical assistance. This work was sup-
LOSS OF HETEROZYGOSITY IN BREAST CARCINOMAS
ported by grants from Torsteds Legacy and Blix Legacy. T.I.A. is a research fellow of The Norwegian Cancer Society. REFERENCES Ali IU, Lidereau R, Theillet C, Callahan R (1987) Reduction to homozygosity of genes on chromosome 11 in human breast neoplasia. Science 238185188. Ara S, Lee PSY, Hansen MF, Saya H (1990) Codon 72 polymorphism of the TP53 gene. Nucleic Acids Res 184961. Bevilacqua G, Sobel ME, Liotta LA, Steeg PS (1989) Association of low am23 RNA levels in human primary infiltrating ductal breast carcinomas with lymph node involvement and other histopathological indicators of high metastatic potential. Cancer Res 4951855190. m e s e n A-L, Ottestad L, Gaustad A, Andersen TI, Heikkila R, Jahnsen T, Tveit KM, Nesland JM (1990a) Amplification and protein overexpression of the neu/HER-Z/c-erbB-2protooncogene in human breast carcinomas: relationship to loss of gene sequences on chromosome 17, family history and prognosis. Br J Cancer 62:5&590. ryrreSen A-L, Ottestad L, Gaustad A, Nesland J, Heikkila R (1990b)Gene amplification of the neu oncogene in breast cancer tumors; correlation to loss of gene sequences on chromosome 17q. Abstract, ESHG meeting. Clin Genet 37292. w e s e n A-L, Hovig E, Smith-%ensen B, Lystad S, Brbgger A (1991) Mutation analysis in human cancers using PCR and constant denaturant gel electrophoresis (CDGE). Adv Mol Gen 4&71. Cavenee WK, Dryja TP, Phillips RA, Benedict WF, Godbout R, Gallie BL, Murphree AL, Strong LC, White RL (1983) Expression of recessive alleles by chromosomal mechanisms in retinoblastoma. Nature 305779-781. Chen L-C, Dollbaum C, Smith HS (1989) Loss of heterozygozity on chromosome l q in human breast cancer. Proc Natl Acad Sci USA 867204-7207. Coles C. Thompson AM, Elder PA, Cohen BB, Mackenzie IM, Cranston G, Chetty U, Mackay J, Macdonald M, Nakamura Y, Hbyheim B, Steel CM (1990) Evidence implicating at least two genes on chromosome 17p in breast carcinogenesis. Lancet 336761-763. Cropp CS,Lidereau R, Campbell G, Champene MH, Callahan R (1990) Loss of heterozygosity on chromosomes 17 and 18 in breast carcinoma: Two additional regions identified. Proc Natl Acad Sci USA 8717737-7741. Devilee P, van den Broek M, Kuipers-DijkshoornN, Kolluri R, Khan PM, Pearson PL, CornelisseCJ (1989) At least four different chromosomal regions are involved in loss of heterozygosity in human breast carcinoma. Genomics 5:554-560. Erlandsson R, Bergerheim USR, Boldog F, Marcsek Z, Kunimi K, Lin BY-T, Ingvarsson S, Castresana JS, Lee W-H, Lee E, Klein G, Sumegi J (1990) A gene near the D3F15S2 site on 3p is expressed in normal human kidney but not or at a severely reduced level in 11 of 15 primary renal cell carcinomas (RCC). Oncogene 51207-1211. Fearon ER, Cho KR, Nigro JM, Kern SE, Smons JW,Ruppert JM, Hamilton SR, Preisinger AC, Thomas G, Kinzler KW, Vogelstein B (1990) Identification of a chromosome 1% gene that is altered in colorectal cancers. Science 2474%55. Feinberg AP, Vogelstein B (1983) A technique for radiolabelling DNA restriction endonuclease fragments to high specific activity. Anal Bicchem 132G13. Finlay CA, Hinds PW, Levine AJ (1989) The p53 protooncogene can act as a suppressor of transformation. Cell 571085-1093. Friend SH, Bemards R, Rogelj S, Weinberg RA, Rapaport JM, Albert DM, Dryla TP (1986) A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature 323643-646. Green AR (1988) Recessive mechanisms of malignancy. Br J Cancer 58:115127. Haber DA, Buckler AJ, Glaser T, Call KM, Pelletier J, SohnRL, Douglas EC, Housman DE (1990) An internal deletion within an llp13 zinc finger gene contributes to the development of Wilms’ tumor. Cell 61:1257-1269. Hall JM, Lee MK, Newman B, Morrow JE, Anderson LA, Huey B, King M-C (1990)Linkage of early-onset familial breast cancer to chromosome 17q21. Science 250:1684-1689.
121
HGM 10 (1989) Tenth international workshop on human gene mapping. Cytogenet Cell Genet 51:622-947. Hbyheim B, Nakamura Y, White R (1989) A BamHl-polymorphism is detected by a genomic p53-clone (pBHP53). Nucleic Acids Res 1 7 8898. Krontiris TG, Dimartino NA, Mitcheson HD, Lonergan JA, Ekgg C, Parkinson DR (1987) Human hypervariable sequences in risk assessment: Rare Hams alleles in cancer patients. Environ Health Perspect 76147-153. Kunkel LM, Smith KD, Boyer SH, Borgaonkar DS, Wachtel SS, Miller OJ, Breg WR, Jones Jr HW,Rary JM (1977) Analysis of human Y-chromosome-specific reiterated DNA in chromosome variants. Roc Natl Acad Sci USA 7412451249. Levine AJ (1990) The p53 protein and its interactions with the oncogene products of the small DNA tumor viruses. Virology 177:41%426. Levine AJ, Momand J (1990) Tumor suppressor genes: The p53 and retinoblastoma sensitivity genes and gene products. Biochim Biophys Acta 1032119-136. Leone A, McBride OW, Weston A, Wang MG, Anglard P, Cropp CS, Goepel JR, Lidereau R, Callahan R, Linehan WM, Rees RC, Harris CC, Liotta LA. Steeg PS (1991)Somatic allelic deletions of NMz3 in human cancer. Cancer Res 51:2490-2493. Lundberg C, Skoog L, Cavenee WK, Nordenskjold M (1987) Loss of heterozygosity in human ductal breast tumors indicates a recessive mutation on chromosome 13.Proc Natl Acad Sci USA 84:2372-2376. Lynch HT, Albano WA, Danes BS,Layton MA, Kmberling WJ, Lynch JF, Cheng SC,Costello KA, Mulcahy GM, Wagner CA, Tindall SL (1984) Genetic predisposition to breast cancer. Cancer 53612422, Mackay J, Steel CM, Elder PA, Forrest APM, Evans HJ (1988) Allele loss on short arm of chromosome 17 in breast cancers. Lancet 213841385. Malkin D, Li FP, Strong LC, Fraumeni Jr JF, Nelson CE, Kim DH, Kassel J, Gryka MA, Bischoff FZ, Tainsky MA, Friend SH (1990) Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms. Science 255:123%1238. Prosser J, Thompson AM, Cranston G, Evans HJ(1990) Evidence that p53 behaves as a tumour suppressor gene in sporadic breast tum o m . Oncogene 5157F1579. Prosser J, Elder PA, Condie A, MacFadyen I, Steel CM, Evans HJ (1991) Mutations in p53 do not account for heritable breast cancer: a study in five affected families. Br J Cancer 63181-184. Rosengard AM, Krutzsch HC, Shearn A, Biggs JR,Barker E, Marguiles LMK,King CR, Liotta LA, Steeg PS (1989) Reduced Nm23/Awd protein in tumor metastasis and aberrant Dvosophila development. Nature 342177-180. Sat0 T, Tanigami A, Yamakawa K, Akiyama F, Kasumi F, Sakamoto G, Nakamura Y (1990) Allelotype of breast cancer: cumulative allele losses promote tumor progression in primary breast cancer. Cancer R e 50.7184-7189. Southern EM (1975) Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol 98505-517. Srivastava S, Zou Z, Pirollo K, Blattner W, Chang EH (1990) Germline transmission of a mutated p53 gene in a cancer-prone family with Li-Fraumeni syndrome. Nature 348:747-749. Stahl JA, Leone A, Rosengard AM, Porter L, King CR, Steeg F5 (1991) Identification of a second human nm23 gene, nmZ3-HZ. Cancer Res 51:44M9. Steeg PS, Bevilacqua G, Pozzatti R, Liotta LA, Sobel ME (1988) Altered expression of NMZ3, a gene associated with low tumor metastatic potential, during adenovirus 2 Ela inhibition of experimental metastasis. Cancer Res 4865504554. Theillet C, Lidereau R, Escot C, Hutzell P, Brunet M, Gest J, Schlon J, Callahan R (1986) Loss of c-H-vas-1 allele and aggressive human primary breast carcinoma. Cancer Res fi4776-4781. Thompson AM, Steel CM, Chetty U, Hawkins RA, Miller WR, Carter DC, Forrest APM, Evans HJ (1990)p53 gene mRNA expression and chromosome 17p allele loss in breast cancer. Br J Cancer 61:7478. Wiggs J, Nordenskjold M, Yandell D, Rapaport J, Grondm V, Janson M, Werelius B, Petersen R, Craft A, Riedel K, Liberfarb R, Walton D, Wilson W, Dryja T P (1988) Prediction of risk of hereditary retinoblastoma, using DNA polymorphisms within the retinoblastoma gene. N Engl J Med 318:151-157. Yamamoto T, Ikawara S, Akiyama T, Semba K. Nomura M, Miyajima N, Saito T, Toyoshima K (1986)Similarity of protein encoded by the human c-erb-B-2 gene to epidermal growth factor receptor. Nature 31923CL234.
