GENES, CHROMOSOMES & CANCER 13:139-240 (1995)

Cytogenetic Studies of Breast Carcinomas: Different Karyotypic Profiles Detected by Direct Harvesting and Short-Term Culture Margr6t Steinad&tir, lngibjorg P h d 6 t t i r . Steinunn Snomd6ttir, J6runn E. Eyfjiird, and Helga M. *undsd6ttir

Cytogenetics Laboratory, Department of Pathology, University Hospital of Iceland (M.S., S.S.), and Molecular and Cell Biology Research Laboratory, Icelandic Cancer Society (I.P., J.E.E., H.M.O.), Reykjavik, Iceland

Chromosome analysis was performed on samples from 85 consecutive patients with breast cancer by one or more of three different methods: direct harvest, culture after mechanical disaggregation, and culture after coltagenase digestion. Metaphases suitable for karyotyping were obtained in 70% of the cases; direct harvest yielded metaphases in 29% and cultures without and with digestion in 40% and 59%, respectively. Chromosomal abnormalities were detected in 37 cases. Cells judged to be phenotypically abnormal in culture were twice as likely t o reveal chromosomal aberrations as normal-looking cells. Eight cases showed multiclonal abnormalities. Significant differences were detected in the karyotypic profile depending on the method used. With direct harvest, the yield of complex chromosomal changes was 87%, compared to 44% after culture of digested tissue (P < O.Ol), and also polyploidy was more common in direct-harvested samples. Detailed karyotypic analysis was possible in 29 primary tumors. The chromosomes most frequently involved were I, 3, 7, I I, 16, and 17. Recurrent structural abnormalities were der( 1;16)(q IOp lo), i( I)(q lo), de1(6)(q2I), and del( l)(p22). Breakpoints clustered to the centromere regions of chromosomes I, 3, I I, 15, and I6 and to the short arms of chromosomes 7, 17, and 19. Seven of twenty-nine fully analyzed cases had a family history of breast cancer, and changes of chromosomes I, 3, and 15 seemed to be more common in these cases. There was an association between karyotype and survival: The 3 year survival was 63% in patients with complex karyotypic changes and 92% in those without complex changes. Genes Chrornosorn Cancer 13239-248 (1995). 0 1995 Wiley-Liss. Inc.

INTRODUCTION

Until very recently, only few reports on chromo- somal studies of breast carcinomas using banding techniques had been published. Lately, progress in the field has picked up speed with the develop- ment of specific culture systems for mammary ep- ithelial cells (Petersen and van Deurs, 1987; Pandis e t al., 1992a). T h e methods used have varied widely, however, ranging from direct harvest (Hill et al., 1987; Mitchell and Santibanez-Koref, 1990; Ferti-Passantonopoulou et al., 1991; Lu et al., 1993) to short- or even long-term culture of the tumor tissue (Zhang e t al., 1989; Dutrillaux et al., 1990; Geleick et al., 1990; Bullerdiek et al., 1993; Pandis et al., 1993; Thompson et al., 1993) before cytogenetic analysis was attempted. With a few ex- ceptions (Gebhart et al., 1986; Hainsworth et al., 1991, 1992), only one method was applied in each study, which makes comparison between the stud- ies difficult. T h e success rate, in terms of both obtaining suitable metaphase cells and detecting clonal abnormalities, has been moderate (less than 30%), except in the recent studies by Pandis et al. (1993).

T h e chromosomes that have been shown to be most commonly involved in gross karyotypic

changes in breast carcinomas are chromosomes 1, 6, 8, 11,3, 7, 16, and 17; chromosome 17 has been studied most extensively using molecular methods (Trent, 1985; Hill et al., 1987; Dutrillaux et al., 1990; Hainsworth and Carson, 1990; Mitchell and Santibanez-Koref, 1990; Bullerdiek et al., 1993; Lu et al., 1993; Pandis et al., 1993; Thompson et al., 1993). These reports do not yet include the most recent advances in chromosome studies in malignancies: the application of fluorescence in situ hybridization (FISH; Chen et al., 1993), com- parative genome hybridization (Kallioniemi et al., 1994), or analysis of karyotypic heterogeneity within tumors (Fuhr et al., 1991; Pandis et al., 1993).

T h e aims of the present study were to establish reliable and reproducible techniques for the kary- otyping of breast carcinomas, to obtain information on the spectrum of chromosomal changes in this malignancy, to compare karyotypic results ob- tained with different methods, and to correlate the

Received November 17, 1994, accepted February 23, 1995. Address reprint requests to Margrkt Steinarsd6ttir, Cytogenetics

Laboratory, Department of Pathology, University Hospital of Ice- land, P.O. Box 1465, 121 Reykjavik, Iceland.

0 1995 Wlley-Us, Inc

240 STElNARSDdlTlR ET AL.

results with clinical data and information on family history.

MATERIALS AND METHODS

Patients and Samples

Consecutive samples from 94 breast carcinomas were obtained immediately after surgery from the Department of Pathology, University of Iceland, between September, 1990, and September, 1992. Of these 94, nine were excluded because they were too small to be processed. The remaining 85 consisted of 78 primary tumors and seven me- tastases or recurrences. Noninvolved breast tissue from the same breasts was obtained in 41 cases. The mean age of the patients was 60.8 years. Tu- mor diameter ranged from 0.4 to 11 cm, and the samples obtained weighed between 0.015 and 0.71 g (mean 0.17 g). The following tumor types were represented: 70 ductal, nine lobular, three mucoid, and one each of tubular, medullary, and papillifer- ous carcinoma.

Processing of Tissue Samples

The samples were placed in 35 mm dishes (Nunc, Roskilde, Denmark) and minced finely in a drop of culture medium. One-half of each sample was then put aside for DNA analysis, and the re- mainder was processed for chromosome analysis.

Direct Harvesting

After removal of the bulk of the sample for DNA isolation or tissue culture, the dish was rinsed with RPMI medium (Gibco, Paisley, Scotland) and the cell suspension centrifuged at 1,000 rpm for 5 min. The cell pellet was treated with 10 ml of 75 mM KCI for 35 min at room temperature and then cen- trifuged at 1,500 rpm. T h e cells were resuspended for fixation by slow addition of 3 ml of freshly pre- pared 3: l methano1:acetic acid and were washed twice in this fixative. After final resuspension in fixative, the cell suspension was dropped onto ice- cold microscope slides. T h e slides were kept dry at room temperature until they were stained and an- alyzed (1-24 months).

Tissue Culture

T h e method for primary culture of breast epi- thelium was based on that described by Petersen and van Deurs (1987), with some modifications ac- cording to Pandis et al. (1992a) and Ogmundsdottir et al. (1993). T h e minced tissue was either plated out directly or digested overnight in 450 U/ml col- lagenase I (Sigma, St. Louis, MO) at 37°C with

gentle rotation. T h e washed cell suspension was left briefly to sediment, yielding a fraction contain- ing debris and blood cells (top layer); a fraction containing a mixture of fibroblasts, single epithelial cells, and endothelial cells (intermediate layer); and a bottom layer of clusters of epithelial cells. T h e cells were then seeded into 25 cm2 culture flasks (Nunc) that had been precoated with a 0.007% solution of Vitrogen 100 (Celtrix Laborato- ries, Palo Alto, CA) and cultured in highly supple- mented serum-free medium. T h e cultures were usually harvested after 3-1 1 days.

Harvesting of Cultures

Cells were arrested in mitosis by standard col- chicine treatment, and chromosomes were har- vested following the method of Pandis et al. (1992a). Chromosome spreads were made as after direct harvesting.

G-Banding and Karyotype Analysis

Chromosomes were G-banded with Wright’s stain following the recommendations of Pandis et al. (1992a). T h e number of cells analyzed de- pended on the material and ranged from 5 to 300 cells per sample. Karyotypes were described ac- cording to the ISCN (1991). Simple aberrations were defined as single structural changes or up to four numerical alterations. All other aberrations were called complex.

Comparison With Clinical Parameters and Family History

Information on clinical history, staging, receptor status, and family history was obtained from the Department of Pathology and the Icelandic Cancer Registry. For statistical analysis, the x2 test and Fisher’s exact test were used. All P values were two sided, with 5% confidence limits. Survival curves were calculated according to standard meth- ods.

RESULTS

Cellular Morphology of Breast Cancer Cultures Related to Karyotype and Number of Mitotic Figures

A subset of 64 breast carcinoma samples was studied specifically for cellular morphology in cul- ture in relation to chromosome analysis. Some growth of epithelial cells was obtained from ap- proximately two-thirds of the samples regardless of tissue preparation. Cell types in culture were judged by the morphological criteria we have de-

KARYOTYPIC PROFILES IN BREAST CARCINOMAS

TABLE I. Cellular Morphology In Culture Related to Karyotype and Mitotic Figures

241

Cultured after mechanical disaggregation Cultured after collagenase digestion

Cell type’ Cell type’

Cancer Mixed Normal Total Cancer Mixed Normal Total

Number of cultures I I 9 12 32b 6 14 22 42‘ Mitotic figures detected in 5 7 7 19 3 13 20 36 Abnormal clone identified in 2 4 2 9 2 9 6 17 Average number of mitotic figures 14 62 18 39 71 61

‘Cell type identified by morphological criteria (see text) as cancer epithelium only; a mixture of cancer and normal epithelium; and normal epithelium only. b%erived from 31 tissue samples and from 37 tissue samples, respectively. Some of the samples were divided into several culture flask The table includes information on 64 tissue samples. The subgroups do not add up because in several cases only one method produced results.

scribed previously (Ogmundsdbttir et al., 1993). Cells were scored as carcinoma cells if they showed irregular outlines, poor attachment to neighboring cells, marked vacuolation, and nuclear abnormali- ties. T h e relationship between karyotype, mitotic figures, and morphology is summarized in Table 1. Considering both culture methods together, karyo- typically abnormal clones were identified in 17 of 28 (61%) cultures in which cancer cells were seen, alone or in mixture, and in eight of 27 (30%) cul- tures in which no cancer cells were recognized. This implies an association between morphology and karyotype, a conclusion further strengthened by the observation that all abnormal clones from cultures morphologically identified as cancer showed complex chromosomal changes. Mixed cultures yielded both complex and simple changes, whereas simple changes were more common in cells with normal morphology.

Cumulative Success Rate of Obtaining Metaphases Using Different Methods

Table 2 shows an analysis of the overall success rate in obtaining metaphases and identification of abnormal karyotypes for the same group of 64 sam- ples as described above. T h e proportion of abnor- mal clones was much higher for direct harvested than for cultured specimens, but the mitotic effi- ciency was considerably lower. This was reflected in the number of cells available for chromosomal analysis, which usually ranged between 20 and 100 after culture but between two and 14 for direct harvest. T h e final yield in terms of the number of abnormal cases identified appeared to be similar using either direct harvest or culture after diges- tion, but the proportion of abnormal metaphases was much higher after direct harvest (see Table 4).

Only few samples yielded abnormal clones with more than one method (see rows 4 and 5 in Table 2). Thus each method contributed separately to an increased overall success rate.

Karyotype Analysis of Normal Breast Tissue

A sample of noninvolved tissue was obtained from the same breast as the carcinoma samples in 41 cases. Two of these showed constitutional in- versions; one case had the known variant inv(l0) (pllq21), the other an inv(X)(qZZq23). In one poorly differentiated cancer, the supposedly nor- mal sample contained the same clonal aberration that was found in the tumor.

Karyotypic Profile of 37 Breast Carcinomas Showing Clonal Abnormalities

Clonal karyotypic abnormalities were detected in 37 tumors (34 primary, 2 metastases, one locally recurrent tumor) from the total group of 85 breast carcinomas examined in this study. A normal karyotype and/or nonclonal aberrations were found in 36 tumors, and no metaphases could be found in 12 cases. Table 3 shows that there were marked differences in the karyotypic profile according to the method used. In seven cases, the same abnor- mal clone was detected by different methods. Over one-half of the clones obtained after culture of di- gested tissue were of the simple type, whereas the majority of clones seen after direct harvest were complex. These differences were statistically sig- nificant, for direct harvest compared with culture after collagenase digestion tested with Fisher’s ex- act test (P = 0.014), and P < 0.01 for all three methods tested with the xz test for trend. Further- more, over one-half of the complex clones #ere polyploid, whereas 18 of 19 simple clones were

242 STEINARSD6TTlR ET AL.

TABLE 2. Cumulative Success Rate of Obtaining Metaphases and Abnormal Clones Using Different Methods

Method Metaphases Abnormal clones identified in identified in

Direct harvest I0/34 (29%) 9/34 (26%) Culture after mechanical 19/48 (40%) 8/48 ( I 7%)

Culture after collagenase 34/58 (59%) 16/58 (28%)

Culture and direct harvest 26/33 (79%) 1Eb/33 (55%) Culture without and with 34/44 (77%) 17‘/44 (39%)

Total 45/64 (70%)” 25/64 (39%)

disaggregation

digestion

digestion

Total number of samples processed was 64. Because of shortage of material. not every sample could be treated with all three methods; therefore, each of the three subgroups contains fewer than 64 samples. but there is overlap between the subgroups, which is indicated in rows 4 and 5. bOr I8 abnormal clones, seven came from direct harvest only, eight from culture only, and three from both culture and direct harvest. W I7 abnormal clones, three came from culture after mechanical dis- aggregation, nine from culture after collagenase digestion, and five from both.

diploid. There was thus a general tendency for an increased proportion of simple abnormal clones and normal karyotypes going from direct harvest to cul- ture without digestion to culture after collagenase digestion. T h e same tendency was also apparent for increased culture time. T h e mean time in cul- ture for all samples was 7.3 days vs. 6.3 days for samples showing chromosomal abnormalities.

Karyotype Analysis of 29 Abnormal Primary Carcinomas

Complete G-banding analysis was not possible in all cases because of inadequate chromosomal mor- phology or difficulties in banding. Table 4 lists the karyotypes for 29 of 34 abnormal samples derived from primary tumors. Eighteen of these karyotypes were obtained after culture, nine from direct har- vest, and two from both. Twenty-one tumors con- tained single abnormal clones; 12 of these showed multiple aberrations, and nine were of a simple type. Seven cases had two or more unrelated clones. In one case, three related subclones were identified. All the multiclonal cases were derived from cultures except for case 7, in which a complex clone was identified from direct harvest and an un- related simple clone from culture.

Recurrent structural abnormalities were der( 1; 16) (qlO;plO), i(l)(qlO), de1(6)(q21), and del(l)(p22). The most frequently lost chromosomes were 17 in five cases, 10 in four cases (once as the sole abnor-

mality), 15 in four cases, and an X chromosome in three cases (twice as a sole abnormality). Gain of a whole chromosome was less common. It occurred twice for chromosomes 10 (once as the only abnor- mality) and 20. Figure 1 shows examples of the karyotypes and abnormalities.

Chromosomes 1, 3 , 7, 11, 16, and 17 were the ones most frequently involved in abnormalities, both in total and in complex changes, whereas chromosomes 3 and 20 were most often involved in simple clones. Complex changes involved chromo- somes 1 and 12 more often in cultured than in directly harvested samples, whereas abnormalities of chromosomes 3, 10, and 15 were more promi- nent after direct harvest.

Breakpoints

A breakpoint map for the 29 primary tumors listed in Table 4 is shown in Figure 2. Breakpoints clustered to the centromere regions of chromo- somes 1, 3, 11, 1.5, and 16 and to the short arms of chromosomes 7, 17, and 19.

Karyotypic Changes Related to Clinical Features and Family History

No association could be identified between karyotypic changes in general and the presence of lymph node metastases or hormone receptors. Analysis of changes in specific chromosomes indi- cated a possible association between lymph node metastases and an altered chromosome 1. Mortality was higher in the patient group (24 cases) with complex clonal abnormalities than among the 43 patients with normal karyotypes or simple changes only. Survival at 36 months was only 63% in the former group but was 92% in the latter. Seven of the twenty-nine patients with primary tumors con- taining clonal karyotypic changes had a first-degree relative with breast cancer. T h e karyotypic profile with respect to complex vs. simple changes was not different in this subgroup compared to the whole group. Patients with a family history showed a somewhat higher frequency of structural changes of chromosomes 1, 3, and 15.

DISCUSSION

This study shows that the use of a combination of different methods enhances the success rate of breast cancer chromosomal analysis and, further- more, that the types of clonal chromosomal abnor- malities differ depending on the method used. Pre- vious studies have largely employed one method only, direct harvest or culture. The terminology

KARYOTYPIC PROFILES IN BREAST CARCINOMAS

TABLE 3. Karyotypic Profile of 49 Abnormal Clones From 37 Breast Carcinomas

243

Ploidy

Method Type of clones Numbers 2n 3n 4n 25n

Direct harvest Simple Complex

Culture after mechanical disaggregation Simple Complex

Culture after collagenase digestion Simple Complex

Total (not additive because of Simple overlap between methods) Complex

2(13%) I 3 (87%) 4 (30%) 8 (70%)

18 (56%) 14 (44%) I9 (39%) 30(6l%)

I I 6" 4 I b 2a 4' 2' 3' 2' I'

I gb.' 6=' 5' I 24c

18 I I2 12 4 3

'Same clone obtained from direct harvest and culture from the same sample. bunrelated clones from the same sample. 'Same clone obtained from the same sample with both culture methods.

used is not standardized, and some authors (Geb- hart et al., 1986) refer to overnight incubation as short-term culture, whereas the same term is used by Dutrillaux et al. (1990) for incubations for 48-72 hr. T h e success rate of obtaining clonal changes is not always obvious in previously published series, but, when it is mentioned, it ranges from 4% to 29%, with the exception of the study by Pandis et al. (1993), who found clonal abnormalities in 80% of their cases. We found clonal changes in 39% of all received cases, or 56% of the cases that gave metaphases. Our culture methods are very similar to those of Pandis et al. (1992a), and our lower success rate may be related to smaller sample size. We found surprisingly little overlap between meth- ods, and the probability of detecting clonal abnor- malities was therefore increased considerably by the combined approach.

T h e abnormal karyotypes detected after culture and direct harvest were different. T h e trend for fewer complex changes observed with increasing time in culture is in agreement with several studies using either direct harvest or culture. Authors using direct harvest describe mainly very complex karyotypes (Hill et al., 1987; Mitchell and Santi- banez-Koref, 1990; Hainsworth et al., 1991; Feni- Passantonopoulou et al., 1991; Lu et al., 1993), whereas in studies based on cultured cells simple karyotypes are more common or are found in equal proportions (Zhang et al., 1989; Geleick et al., 1990; Bullerdiek et al., 1993; Pandis et al., 1993; Thompson et al., 1993). T h e method of Dutrillaux et al. (1990) is based on a very short time in cul- ture, and they found a high frequency of complex changes. T h e choice of culture medium is obvi- ously very important if analyzable mitotic cells are

to be obtained. Fetal calf serum stimulates the growth of stromal fibroblasts, whereas some spe- cially designed serum-free media may favor the growth of normal mammary gland epithelium (Zhang et al., 1989). T h e medium and approach used by ourselves and by Pandis et al. (1992a) were chosen in an attempt to circumvent both of these problems. Unfortunately, there are no clear or ac- cepted morphological markers that allow us to rec- ognize mammary carcinoma cells in primary cul- ture. T h e associations reported here between certain morphological characteristics and complex karyotypic changes, as well as the frequency of such changes in our cultures, indicate that this method can support the proliferation of primary breast carcinoma cells.

T h e chromosomes most frequently involved in abnormalities in our study were chromosomes 1, 3, 7, 11, 16, and 17, which is in line with the results obtained in other studies. Chromosome 1 has re- peatedly been reported to be the one most fre- quently aberrant in breast cancer. In our study, three of four recurrent structural changes involved chromosome 1; der(1; 16)(qlO;plO) and i(1q) were described as karyotypic subgroups in breast cancer by Pandis et al. (1992b) and are believed to be early events in breast cancer formation. These ab- errations result in gain of lq , the first one also in loss of 16q. In case 9, we found gain of distal l q through an unbalanced 1;4 translocation as the sole abnormality. Deletions of lp22 are commonly found in breast carcinomas with complex karyo- types (Mitchell and Santibanez-Koref, 1990).

About 40% of our abnormal cases had chromo- some 17 alterations, which is similar to the fre- quency noted by Mitchell and Santibanez-Koref

244 STEINARSDdlTlR ET AL.

TABLE 4. Cytogenetic Findings in 29 Primary Tumors

Sample No. Type pTNM Harvest Karyotypes

I 2

3

4

5b 6 7

8 9

I I I 2d

I 3= 14 15 16 17 184' 19 20 23 24

27

28" 29b*d

30d

31"

32'

33

34b

L ID

ID

ID

ID L ID

M M ID ID

ID ID ID ID ID MED ID ID ID L

ID

ID ID

ID

ID

ID

ID

ID

I II

111

0

II II II

I 111 111 II

I I I I II I I II I I

II

II I

111

II

I

II

II

C C

C

D

D C D/C

D D C C

D D C C C C C D C C

C

C C

D/C

D

C

D

C

44-45,XX,der( l)t(1;8)(pl3;ql I)x2,- I I,- 16,- I7,add( 17)(p13),+2mar[cp9]/46,XX[31] 45,X,-X[5]/57-64,-X,add(X)(pZZ),del( I)(pl3),add(3)(p I3).add( I I)(p I )),add( 17)(pl I), + r,inc[4]/46,XX[70] 43-46,del(X)(q24),der(X)t(X8)(q24;q I 3),del( l)(p36),add( I)(p34),add(2)(p25).add(2)(q37), - 3,-5,t(5;9)(q3 I ;q34),add(6)(q I s),add(7)(~22),hsr(7)(q32-36),-8,add( I O)(q24),- I I, add(12)(q22),- 13,- 14,- 15,hsr(16)(pl I-I3),hsr(l6)(q22-24),- 17,der(18)t(3;18) (q I3;pl l),-2l ,der(?)t(?;3)(!;p21), + 5-8mar[9]/92,idemx2[5]/46,XX,inv(3)(p12q27)[5]146,

35-4 I ,XX,- 3,-8,- 10,- I3,add( I4)(p I I),-- 16,- I7,add( I7)(p I l),der(20)t(3;20)(q I2;p I I), -2 I ,-22,der(?) t(5 I I )(?;q I 3). + 2r, + mar[30] 33-47,XX,i( I)(q lo), + I O,inc[3] 46,X,t(X6)(p I I ;p23-25)[20]/92,idemx2[2]/46,XX[33] <80>,X?,?t( I ; I9)(q I I ;p I 3),hsr(der(5)),der( I5)t( 15; I7)(p I I ;q I 2).inc[2]/47,XX, + 8[2]/46.

45-47,XX, + r[2] 92,XXXX,der(4)t( I ;4)(q25;q35)x2[5] 48,XX, + 20, + 2 I [7]/46,XX[ I 591 42,X,-X,del( I )(p I3),del( I)(p22),del(S)(p I3),add( I2)(q24),add( I7)(p I 3).inc[2]/63-74, idemx2,add( I2)(p I 3),inc[27]/128-13 I ,idemx4,add( I2)(p I3)x2,-add( IZ)(q24),inc[4]/46,

45,XX,+der(l;16)(qIOp10),-4,-7,?t(l 1;14)(p13;ql3),- 15,- l6,+2mar[5]/46,XX[5] 63-69,add(7)(q?),i( I I )(q I O),hsr( I7)x2,inc[5] 47,XX. + I8[2]/46,XX[ I71 46,XX,t( );7)(q29;q I I )[ 8]/46,XX,t(2;3)(~25;qZ I ) [ 5]/45,XX, - 20[4]/46,XX[ 961 45,XX,- I O [ 3]/46,XX[72] 39-40,?XX,der( I; 16)(q I Op IO),de1(3)(q21),de1(6)(q2I),inc[cp4]

59-65,XX,del( I)(q I2),de1(3)(p I2),add( I l)(q23),add( I 5)(p I I),inc[ I41

44-47,XX,dup( l)(q21q44),t( 1;16)(q2I;ql2),add(3)(q28),der(8)t(8;lb)(pl I;pl I),del( I I) (q23),- 16, add( 17)(pl 1)[24]/46,XX[66] 43-45,XX,der( I)del( l)(p32)ins( I ;?)(qZI-25;?),add( I2)(q24), + r,inc[cp8]/86-89,idemx 2[cp6]/47,XX, + 7[ 3]/46,XX[ 3001 46,XX, + der( I ; I6)(q I O p I O),- I6[4]/46.XX[ I961 76-87,XXXX,del( I)(p22), + i( I)(q I O),-2,-3,del(3)(q I l),der(6)t(3;6; I6)(pl I;q I l;q24),-6, -6,-7,de1(7)(q22),-8,-8,-9,- IO.add(l l)(pl2),- 14,- 15,- 15,del(l6)(q21)~2,der(l6) t(3;6; I6)(p I I ;q I I ;q24),add( 17) (p I 3)x2, + 5-7mar[cp I8]/45,X,-X[9]/46,XX[75] 106-1 I2,XX,del(X)(q I I),del( l)(p22),del( I)(q2l),add(2)(p I3),de1(3)(pl3),de1(3)(q23),add (5)(q35),add(6)(q I 5),der(6)dup(6)(p2 I p25)add(b)(q27),dup(6)(~2 I p25),de1(7)(q22), dic( IO20)(q26;p I 3),del( I O)(q24),add( I 3)(p I 3),der( I 3)t(?2; I 3)(pl I ;p I I),dic( I $?)(PI I ;?), add( I5)(p I I),add( I7)(q I I ),add( I8)(q23),add( I9)(p I 3),dic( I %?)(PI 3;?),inc[cp33]/46,XX[37] 48-5I,XX,+X,+der( 1;16)(q1Op1O)x2,+ 3,der(8)t(8;15)(~22;q22),+ lo,+ 15,- 16,- 17[ lo] 75-86,<4n>,XX,del(X)(pI l),der(X3)(pIOqlO),del( l)(p31)x2,der( I)t( 1;9)(q21;ql3)x2. t( I;5)(p IOpIO)x2,-2,-2,del(3)(q I3)~2,-4,-4,de1(4)(pI 3),del(6)(q2l).del(7)(q21),de1(7) (q22),der(7; I3)(p I Oq I O),hsr(7)(p22)~2.-8,-8,add(8)(p2 I ),- 10,- I O,inv( I O)(p I I q22)x2, - I I,ins(l I;?)(q13;?),der(l2)t(3;12)(q23;qI5)t(3;12)(q23;p13),-13,- l3,-14,- 15,der(15) t(5;15)(ql3;pl 1-1 3),- 16,- 16,- 17,der( 17)t( 16;17)(pl I;pl I),- 18,- 18,add( 19)(pl3), del(20)(qI 3),-21,-22, + 3-6mar[50] 68-70,XXX,add(3)(q2 I ),del( 3)(q22q25).add(7)(~22),inv(7)(~22q I l),?der(8)t(8;9)(p2 I ;q I 3), del(l I)(pl2),add(l5)(pl I-13),der(19)t(1;19)(ql2ql3),add(2l)(q22),inc[cp4] 46,XX,t( I I ;20)(q I 3;p I 3)[5]/64-7 I ,?XXX,i( I )(q I O),add( I6)(p I 3),inc[5]/45,XX,

W 6 I 1

~ 3 0 1

x " 3 1

45-53.XX, + 2, + I I, + 20[~~4]/46,XX[76]

48-50,XX, + l2[~~2]/46,XX[52]

- I9[3]/46,XX[ I251 ~ ~~ ~~~

&Recurrent structural changes: 4 der(1;16)(q1Op10); b, i( I)(qlO); c, del(6)(q2I); d, del( l)(p22). D. direct harvest; C, culture; ID. infiltrating ductal; L, lobular; M, rnucoid; MED, medullary.

KARYOTYPIC PROFILES IN BREAST CARCINOMAS 245

Figure I. il: Complex -type from case 3. Clonal description is given in Table 4. but in this particular cell a few nonclonal changes are seen, i.e., -4. hsr(7)(~15-22)(q32-36). and - 14, and three clonal changes are not present, i.e., add(6)(q IS), - 15. and -21. See also num- ber I I in b. b Partial karyotypes giving examples d chromosomal ab- enatiomfromrimpleand complexclones.Clonal descriptions aregiven in Table 4. 1-6: Recurrent abnormaliies and/or karyotypic subgroups. I :

der(1;16)(q10plO) case 28. 2 i(lq) case 5. 3 del(l)(p22) case 12. 4 de1(6)(q21) case 32.5 + 18 ca5e 15. 6 + 7 case 27.7 partial gain of I q from case 9.8: Abnormaliies of 7p22 from case 33.9.10: Two of three unrelated clones from case 16. I I : Second clone from case 3. 12-14: Abnormalities of chromosome 17. add(17)(p13) case 2 9 der(l7) t(16;17)(pl I;pl I ) case 32. and add(l7)(pl I) case 4. 1 5 Simple clone from case I I ( + 20, + 2 I ).

246 -1

STEINARSD6TTIR ET AL.

2

3 9 4 4

1

7

13

2

8

14

3

9

15

4

10

16

I 9 20 21 22

5

11

Figure 2. Breakpoint map of 29 primary breast carcinomas.

17

6

12

X

(1990). This is interesting in the light of molecular data on chromosome 17 in breast cancer, 17q being the location of the breast-ovarian cancer suscepti- bility gene BRCAI. T h e alterations detected were mainly loss of a whole chromosome or partial loss of

17p. In a previous study, we examined changes in the TP53 gene on 17p in breast tumors and found a significant association between TP53 mutations and allelic imbalance on both arms of chromosome 17 (Thorlacius et al., 1993). Furthermore, we have

KARYOTYPIC PROFILES IN BREAST CARCINOMAS 247

shown that TP53 abnormalities are significantly as- sociated with complex chromosomal changes in pri- mary breast tumors (Eyfjord et al., 1995).

Five more changes that have been proposed to identify karyotypic subgroups of breast carcinomas (Pandis et al., 1995) occurred in one case each in our material: +7, + 18, +20, del(l)(qll-12), and t( l)(qlO-1 1). T h e recurrent numerical changes ob- served were those commonly found in highly ab- normal karyotypes. In a few cases, numerical changes were seen as sole abnormalities. Some of these changes ( + 7, + 10, and -X) have been re- ported in various tumors, but it has been ques- tioned whether they were derived from neoplastic cells (Lindstrom et al., 1991; Dal Cin et al., 1992).

Seven cases were shown to contain two or more unrelated clones. Most of these clones were iden- tified in cultured cells. Multiclonality has been de- tected in various tumors, such as colorectal cancer (Bardi et al., 1993a), head and neck tumors (Jin et al., 1993), and pancreatic cancer (Bardi et al., 1993b). These studies used short-term cultures af- ter collagenase digestion of tumor tissue to obtain metaphases, and, furthermore, Jin et al. (1993) found that the culture conditions influenced the frequencies of unrelated clones. We are aware of only one other group reporting multiclonality in breast carcinomas (Pandis et al., 1993), and they used the same culture conditions as we used. In- tratumoral heterogeneity is well known in breast carcinoma from DNA flow cytometry studies (Fuhr et al., 1991) and immunohistochemical anal- ysis of TP53 (Fisher et al., 1994; Thorlacius et al., in press). T h e importance and frequency of multiclonality among solid tumors is as yet un- known.

Several lines of evidence point in the direction of poorer prognosis for patients with complex chro- mosomal changes. We saw shorter survival for such patients. A very high rate of complex karyotypes was reported by Trent et al. (1993) for malignant effusions, i.e., in very advanced cases of breast carcinomas. Pandis et al. (1993) found higher his- tological grade to be more common in tumors with complex changes, and Zafrani et al. (1992) noted that homogeneously staining regions were more frequent in breast cancer patients with poor prog- nostic factors. Several studies, including our own (Thorlacius et al., 1993), have demonstrated an association between TP53 abnormalities and poor short-term survival in breast cancer, and, as men- tioned above, these abnormalities were signifi- cantly associated with complex karyotypic changes. There is very little information on clinical associa-

tions with specific chromosomal changes, but Hainsworth et al. (1992) found a significant corre- lation between rearrangements on chromosome 1 and poor prognosis. Our data indicate an associa- tion between changes in this chromosome and the presence of lymph node metastases.

T h e lower incidence of complex karyotypes among cultured cells compared to direct harvest preparations to some extent reflects the fact that the latter method fails to pick up abnormal cases with only minor changes. This bias may also occur if cells carrying complex changes are more easily lost during enzyme digestion in preparation for cul- ture, as was suggested for prostate carcinoma by Konig et al. (1993). These cells have poor prolif- erative rates and/or yield mitoses that cannot be analyzed. T h e lack of simple karyotypic changes identified from direct harvest shows that cells with these changes require some stimulation to enter mitosis, or that they were not easily detached from the tissue clumps in direct preparations. T h e cen- tral issue in determining the biological relevance of complex and simple karyotypic changes is the question of identity of the cells harboring them. This question could be addressed by combining immunohistochemical techniques with FISH anal- ysis of chromosomes, as was described recently by Weber-Matthiesen et al. (1992).

We conclude that meaningful karyotypic data can be obtained from a significant proportion of breast carcinomas, particularly by using a com- bined approach. A possible correlation between the tumor karyotype, clinical behavior, and family his- tory is beginning to emerge. Cytogenetic alter- ations on chromosome 17 have been related to mo- lecular data. Commonly observed breakpoints on chromosomes 1, 7, 15, 16, 17, and 19 suggest that these may be locations of hitherto unknown onco- genes or tumor suppressor genes that are important in breast carcinogenesis.

ACKNOWLEDGMENTS

T h e authors thank Prof. Sverre Heim and Dr. Nikos Pandis of the University of Lund and the staff of the Icelandic Cancer Registry for their as- sistance. This work was supported by grants from T h e National Hospital Science Fund, T h e Icelan- dic Science Fund, NORFA, and T h e Memorial Fund of Bergthora Magnusdottir and Jakob J. Bjar- nason.

REFERENCES

Bardi G , Johansson B, Pandis N, Mandahl N, Bak-Jensen E, Lind- srrom C, Tornqvist A, Frederiksen H, Andrkn-Sandberg A, Mitelman F, Heim S (1993a) Cyrogeneric analysis of 52 colorecral

24% STElNARSDdTTlR ET AL.

carcinomas-Nonrandom aberration pattern and correlation with pathologic parameters. Int J Cancer 55:422-428.

Bardi G, Johansson B, Pandis N, Mandahl N, Bak-Jensen E, An- dren-Sandberg A, Mitelman F, Heim S (1993b) Karyotypic ab- normalities in tumours of the pancreas. Br J Cancer 67:1106- 11 12.

Bullerdiek J, Leuschner E, Taquia E, Bonk U, Bartnitzke S (1993) Trisomy 8 as a recurrent clonal abnormality in breast cancer? Can- cer Genet Cytogenet 6564-67.

Chen Z, Morgan R, Stone JF, Sandberg AA (1993) FISH: A useful technique in the verification of clonality of random chromosome abnormalities. Cancer Genet Cytogenet 66:73-74.

Dal Cin P, Aly MS, Delabie J, Ceuppens JL, Van Goo1 S, Van Damme B, Baert L, Van Poppel H, Van Den Berghe H (1992) Trisomy 7 and trisomy 10 characterize subpopulations of tumor- infiltrating lymphocytes in kidney tumors and in the surrounding kidney tissue. Proc Natl Acad Sci USA 899744-9748.

Dutrillaux B, Gerbault-Serueau M, Zafrani B (1990) Characteriza- tion of chromosomal anomalies in human breast cancer. A com- parison of 30 paradiploid cases with few chromosome changes. Cancer Genet Cytogenet 49:203-217.

Eytjord JE, Thorlacius S, Steinarsdottir M, Valgardsd6ttir R, og- mundsd6ttir HM, Anamathawat-Jonsson K (1995) P53 abnormal- ities and genomic instability in primary human breast carcinomas. Cancer Res 55:646-651.

Ferti-Passantonopoulou A, Panani AD, Raptis S (1991) Preferential involvement of 11q23-24 and l lp15 in breast cancer. Cancer Genet Cytogenet 51:183-188.

Fisher CJ, Gillett CE, Vojtesek B, Barnes DM, Millis RR (1994) Problems with p53 immunohistochemical staining: T h e effect of fixation and variation in the methods of evaluation. Br J Cancer

Fuhr JE, Frye A, Kattine AA, Van Meter S (1991) Flow cytometric determination of breast tumor heterogeneity. Cancer 67: 1401- 1405.

Gebhart E , Bruderlein S, Augustus M, Siebert E , Feldner J, Schmidt W (1986) Cytogenetic studies on human breast carcino- mas. Breast Cancer Res Treat 8:125-138.

Geleick D, Miiller H , Matter A, Torhorst J, Regenass U (1990) Cytogenetics of breast cancer. Cancer Genet Cytogenet 4 6 2 17- 229.

Hainsworth PJ, Garson OM (1990) Breast cancer cytogenetics and beyond. Aust NZ J Surg 60:327-336

Hainsworth PJ, Raphael KL, Stillwell RG, Bennett RC, Garson OM (1991) Cytogenetic features of twenty-six primary breast cancers. Cancer Genet Cytogenet 52:205-218.

Hainsworth PJ, Raphael KL, Stillwell RG, Bennett RC, Garson OM (1992) Rearrangement of chromosome l p in breast cancer corre- lates with poor prognostic features. Br J Cancer 66:131-135.

Hill SM, Rodgers CS, Hulttn MA (1987) Cytogenetic analysis in human breast carcinoma. 11. Seven cases in the triploid/tetraploid range investigated using direct preparations. Cancer Genet Cyto- genet 2445-62.

ISCN (1991) Guidelines for Cancer Cytogenetics, Supplement to an International System for Human Cytogenetic Nomenclature. Mitelman F (ed). Basel: S. Karger.

Jin Y, Mertens F, Mandahl N, Heim S, OlegArd C, Wennerberg J, Biorklund A, Mitelman F (1993) Chromosome abnormalities in eighty-three head and neck squamous cell carcinomas: Influence of culture conditions on karyotypic pattern. Cancer Res 532140- 2146.

Kallioniemi A, Kallioniemi OP, Piper J, Tanner M, Stokke T, Cien L, Smith HS, Pinkel D, Gray JW, Waldman FM (1994) Detection and mapping of amplified DNA sequences in breast cancer by comparative genomic hybridization. Proc Natl Acad Sci USA 91:

Konig JJ, van Dongen JW, Schroder FH (1993) Preferential loss of

69: 26-3 1.

2156-2 160.

abnormal prostate carcinoma cells by collagenase treatment. Cy- tometry 14805-810.

Lindstrom E, Salford LG, Heim S, Mandahl N, Stromblad S, Brun A, Mitelman F (1991) Trisomy 7 and sex chromosome loss need not be representative of tumor parenchyma cells in malignant glioma. Genes Chromosom Cancer 3:474-479.

Lu YJ, Xiao S, Yan YS, Fu SB, Liu QZ, Li P (1993) Direct chro- mosome analysis of 50 primary breast carcinomas. Cancer Genet Cytogenet 6991-99.

Mitchell ELD, Santibanez-Koref MF (1990) lp13 Is the most fre- quently involved band in structural chromosomal rearrangements in human breast cancer. Genes Chromosom Cancer 2278-289.

Ogmundsddttir HM, Pttursdottir I, Gudmundsdottir I, Amunda- dottir L, Ronnow-Jessen L, Petersen OW (1993) Effects of lym- phocytes and fibroblasts on the growth of human mammary car- cinoma cells studied in short-term primary cultures. In Vitro Cell Dev Biol 29A936-942.

Pandis N, Heim S, Bardi G, Limon J, Mandahl N, Mitelman F (1992a) Improved technique for short-term culture and cytoge- netic analysis of human breast cancer. Genes Chromosom Cancer 5: 14-20.

Pandis N, Heim S, Bardi G, Idvall I, Mandahl N, Mitelman F (1992b) Whole-arm t(1;16) and i(lq) as sole anomalies identify gain of l q as a primary chromosomal abnormality in breast cancer. Genes Chromosom Cancer 5:235-238.

Pandis N, Heim S, Bardi G, Idvall I, Mandahl N, Mitelman F (1993) Chromosome analysis of 20 breast carcinomas: Cytogenetic multiclonaliry and karyotypic-pathologic correlations. Genes Chromosom Cancer 651-57.

Pandis N, Jin Y, Gorunova L, Petersson C, Bardi G, Idvall 1, Jo- hansson B, Ingvar C, Mandahl N, Mitelman F, Heim S (1995) Chromosome analysis of 97 primary breast carcinomas: Identifi- cation of eight karyorypic subgroups. Genes Chromosom Cancer

Petersen OW, van Deurs B (1987) Preservation of defined pheno- typic traits in short-term cultured human breast carcinoma de- rived epithelial cells. Cancer Res 472356466.

Thompson F, Emerson J, Dalton W, Yang JM, McGee D, Villar H, Knox S, Massey K, Weinstein R, Bhattacharyya A, Trent J (1993) Clonal chromosome abnormalities in human breast carcinomas I. Twenty-eight cases with primary disease. Genes Chromosom

12:173-185.

Cancer 7:185-193. Thorlacius S, Borresen A-L, Eyfiord JE (1993) Somatic ~ 5 3 muta-

tions in human breast carcinomas in an Icelandic population: A prognostic factor. Cancer Res 53:1637-1641.

Thorlacius S, Thorgilsson B, Bjornsson J, Trygvad6ttir L, Borre- sen A, Ogmundsdottir HM, Eytjord JE (in press) TP53 mutations and abnormal TP53 protein staining in breast carcinomas related to prognosis. Eur J Cancer.

Trent J, Yang JM, Emerson J, Dalton W, McGee D, Massey K, Thompson F, Villar H (1993) Clonal chromosome abnormalities in human breast carcinomas 11. Thirty-four cases with metastatic disease. Genes Chromosom Cancer 7: 194-203.

Trent JM (1985) Cytogenetic and molecular biologic alterations in human breast cancer: A review. Breast Cancer Res Treat 5221- 229.

Weber-Matthiesen K, Winkemann M, Muller-Hermelink A, Schlegelberger B, Grote W (1992) Simultaneous fluorescence im- munophenotyping and interphase cytogenetics: A contribution to the characterization of tumor cells. J Histochem Cytochem 40:

Zafrani B, Gerbault-Seureau M, Mosseri V, Dutrillaux B (1992) Cytogenetic study of breast cancer: Clinicopathologic significance of homogeneously staining regions in 84 patients. Hum Pathol

Zhang R, Wiley J, Howard SP, Meisner LF, Gould MN (1989) Rare clonal karyotypic variants in primary cultures of human breast carcinoma cells. Cancer Res 49:444-449.

171-175.

23: 542-547.