[CANCER RESEARCH 61, 1214–1219, February 1, 2001]

Chromosomal Rearrangements and Oncogene Amplification PrecedeAneuploidization in the Genetic Evolution of Breast Cancer1

Karin Rennstam,2 Bo Baldetorp, Soili Kytola, Minna Tanner, and Jorma IsolaDepartment of Oncology, Jubileum Institute, University Hospital, S-221 85 Lund, Sweden [K. R., B. B., M. T., J. I.]; Institute of Medical Technology, Tampere University andUniversity Hospital, Tampere, FIN-33101 Finland [S. K., M. T., J. I.]; and Department of Molecular Medicine, Karolinska Hospital, CMML 8:01, Stockholm, S-17176 Sweden [S. K.]

ABSTRACT

Breast carcinoma is thought to arise because of multiple successivechanges in the genome of the normal epithelial cells. However, little isknown of the order of appearance of different types of genetic aberrations.We studied the ERBB2 (Her-2/neu) and CCND1 (cyclin D1) oncogeneamplification in flow cytometrically sorted diploid and nondiploid tumorcell populations by fluorescencein situ hybridization (FISH). The purity ofthe cell sorting was confirmed by static DNA image cytometry. Spectralkaryotyping was used to define differences in a genome-wide mannerbetween two distinctly different aneuploid cell clones found in each of twobreast cancer cell lines. FISH indicated the presence of gene amplificationboth in diploid and nondiploid cell clones in 17 of the 21 amplification-containing tumors analyzed. The oncogene copy numbers remained un-changed throughout aneuploidization in 11 of 17 tumors. The remainingsix tumors showed an increase in oncogene copy number as well as thenumber of chromosome 11 or 17 centromeres (the original location ofCCND1 and ERBB2, respectively). Breast carcinoma cell lines MDA-157and MDA-436 showed a significant number of chromosomal rearrange-ments in the near-diploid clones, which were present in duplicate in thecorresponding aneuploid (polyploid) clones. These results indicate thatploidy shift, i.e., aneuploidization, in breast cancer is a late genetic event,which is preceded by both oncogene amplifications as well as manychromosomal rearrangements.

INTRODUCTION

Breast carcinoma, as well as other carcinomas, arises because ofmultiple changes in the genome of the normal epithelial cells. Thesechanges include single nucleotide point mutations, amplifications, ordeletions of single genes, insertions and translocations, gains andlosses of entire, or parts of, chromosomes and chromosome arms, andeventually gross changes in chromosome number (aneuploidization;Refs. 1–3). A widely accepted model depicted from colorectal carci-noma suggests that genetic aberrations occur in stepwise manner,correlating well with morphological change from adenoma to carci-noma (4). However, studies in early breast lesions are not fullyconcordant with this model. Atypical ductal hyperplasia and DCIS3

have repeatedly been shown to contain numerous genetic aberrations,similar to invasive carcinomas (5, 6). Thus, the relationship betweengenetic pathogenesis and the morphological progression in breastcancer has remained obscure.

Relatively little is known also of the order of appearance of differ-ent types of genetic aberrations. Studies based on cytogenetic datasuggest indirectly that chromosomal imbalances may occur earlierthan gross ploidy shifts (aneuploidization), indicating the latter to bea late genetic event (7, 8). These observations, in turn, are contra-dicted by observations of aneuploidy being commonly present in earlybreast lesions, such as DCIS (9–12). Even less is known on theappearance of oncogene amplifications, which are crucial in deter-mining the clinical outcome and treatment response. The most com-monly amplified oncogenes in breast cancer involve oncogenesERBB2,MYC, andCCND1, all of which have been found amplified inDCIS (13–15).

The present study was initiated by our findings from CGH, whichdetects clonal chromosomal copy number imbalances (gains andlosses) when present in at least 60% of the cells from which the DNAis extracted (16). We have, however, repeatedly found breast tumorsshowing numerous copy number changes, despite DNA flow cytom-etry (performed from adjacent tissue sections), indicating only a smallfraction (,30%) of nondiploid cells. This suggests that chromosomalchanges may be present also in the diploid cell population. In thepresent study, we examined this aspect directly by analyzing diploidand nondiploid cell populations separately after flow cytometricallysorting them by their DNA content. Originally, we tried to performCGH on these sorted cells. This was, however, unfruitful, because thenormal cell contamination within the diploid peak was too high toallow any of the aberrations in the nonnormal diploid cells to bedetected. (All aberrations seen in the aneuploid peak were also seen inthe whole tumor CGH.) For this reason, we chose FISH, which allowsthe analysis to be made on a cell-by-cell basis. The cell populationswere analyzed separately with probes identifying gene amplificationsof CCND1andERBB2oncogenes. DNA probes for the chromosomeson which the oncogenes are located (chromosomes 11 and 17, respec-tively) were used as reference, separating the true oncogene amplifi-cations from multiplication of the entire chromosomes. SKY was usedto further clarify the order of appearance of chromosomal rearrange-ments in clonally heterogeneous breast cancer cell lines, where high-quality metaphase preparations could be prepared easily.

MATERIALS AND METHODS

Tumor Samples. Tumor specimens from 21 primary invasive breast can-cers were obtained from the frozen tumor tissue bank. The specimens havebeen collected for the routine analysis of steroid hormone receptors, and DNAflow cytometry was performed at the Department of Oncology, Lund Univer-sity, Sweden. The tumors selected had all been determined previously ashaving amplification of the 11q13 region (tumors 1–10) orERBB2oncogene(tumors 11–21) by Southern blot hybridization (17, 18). For the SKY analysis,two breast cancer cell lines (MDA436 and MDA157) were obtained fromAmerican Type Culture Collection (Rochester, Michigan) and cultured inrecommended conditions.

DNA Flow Cytometry. Freshly frozen tumor samples (100–200 mg) andcell lines were prepared for flow cytometric analysis as described previously(19). Calculation of the DNA index was done after zero point adjustment of theDNA histogram using the modal values of chicken and trout RBCs. The meanchannel numbers of all G0-G1 peaks were then used for the calculation of theDNA index, with chicken and trout RBCs as reference standard (19). The

Received 5/30/00; accepted 12/11/00.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby markedadvertisementin accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1 This study was supported by The Swedish Cancer Foundation Grant 4237-099-01XAB; the Mrs. Berta Kamprads Foundation; the Gunnar, Arvid, and ElisabethNilsson Foundation; the John and Augusta Persson Foundation; the Inga Britt andArne Lundberg Foundation; the Hospital of Lund Foundation; the Franke and Mar-garetha Bergqvist Foundation; the King Gustav V Jubilee Foundation Grant 99:525;the Finnish Cancer Foundation; and the Scientific Foundation of Tampere UniversityHospital. J. I. received Fellowships 4047-B98 – 01VAA and 4047-B99 – 02VAA fromthe Swedish Cancer Society.

2 To whom requests for reprints should be addressed, at Department of Oncology,Jubileum Institute, Lund University, S-221 85 Lund, Sweden. Phone: 46-46-177567; Fax:46-46-147327; E-mail: Karin.Rennstam@onk.lu.se.

3 The abbreviations used are: DCIS, ductal carcinomain situ; CGH, comparativegenomic hybridization; FISH, fluorescencein situ hybridization; SKY, spectral karyotyp-ing; CV, coefficient of variation; DAPI, 4,6-diamino-2-phenylindole.

1214

Research. on March 7, 2021. © 2001 American Association for Cancercancerres.aacrjournals.org Downloaded from

number of cells in the aneuploid G0-G1 peak was divided by the total numberof cells in the DNA histogram to get the aneuploid population fraction. TheCVs of the diploid G02G1 peaks of 20 of the 21 tumors ranged between 2.2and 5.2. The CV of tumor number one was 7.4. All CVs were well under therecommended maximum value published previously (20).

Flow Cytometric Sorting. Simultaneous with the DNA analysis of thetumors, nuclei from the two different cell populations were separated by flowcytometric sorting. As sorting criteria, narrow electronic gates were set in theDNA histogram around the G0-G1 peaks, defined by the CVs. PBS (pH 7.0)was used as sheath fluid for sorting. The analysis rate was;150 nuclei/s.Electronic controls for the sorting were set as follows: droplet frequency, 30KHz; three droplets/sorted event; and coincidence check of five droplets, thisyielding an efficiency close to 100% and purity.95%. Up to 300,000 sortednuclei was collected onto microscope slides for each peak. Slides were imme-diately fixed in 50% Carnoy’s solution (3:1 methanol:acetic acid in water) andair dried.

DNA Image Cytometry. Before analysis, the slides with flow cytometri-cally sorted cells were fixed in 4% phosphate-buffered formaldehyde for atleast 30 min and then Feulgen stained as described earlier (21, 22). Integraloptical densitometric measurement of nuclear DNA content was performedwith a LabEye 3PC image analysis system (Innovate Vision AB, Linkoping,Sweden). For each specimen, well-preserved nuclei were selected randomly,and integrated absorbance was measured at a wavelength of 540 nm. Nucleifrom human cerebellum (fresh autopsy material) were used as diploid externalreference cells for ploidy assessment and as a control for Feulgen staining (23).

FISH. After air drying, slides were fixed with 50, 70, and 100% Carnoy’ssolution, 10 min each. Samples were then further fixed with 1% paraformal-dehyde in PBS (10 min at 4°C), dehydrated in graded ethanol series (70, 85,and 100%), air dried, and baked at 80°C for 30 min in a hybridization oven.Two-color FISH was carried out as described previously (8) with minormodifications. Slides were denatured in a 70% formamide-23 SSC at 72°C for3 min. The directly labeled dual-color probes forCCND1(and chromosome 11centromere) andERBB2(and chromosome 17 centromere) were obtained fromVysis, Inc. (Downers Grove, IL). The hybridization mixture for each slidecontained 3.4ml of master mix (70% formamide and 10% dextran sulfate in23 SSC), 0.4ml of placental DNA, and 0.25ml of the probe solution. Theprobe mixture was denatured at 75°C for 5 min and applied onto slides. Thehybridization was carried out overnight at 42°C. Posthybridization stringencywashes were done at 72°C (0.43SSC for 2 min) and room temperature (23SSC for 1 min). After a short rinse in distilled water, the slides were air driedand counterstained with 0.2mM DAPI in an antifade solution (Vectashield;Vector Laboratories, Burlingame, CA).

Hybridization signals were analyzed using a Zeiss Axioplan 2 epifluores-cence microscope equipped with dual band-pass fluorescence filter (Chroma-technology, Brattleboro, NV), which enables simultaneous detection of bothgreen (500–600 nm) and red (600–700 nm) fluorescence. Hybridizationsignals from at least 50 nuclei were scored to assess the chromosome centro-mere and oncogene copy numbers. The nuclei was determined to carry an

amplification if the number of gene probe signals divided by the number ofcentromere signals was$1.5. Digital images were taken with a cooled CCDcamera (Sensys; Photometrics, Tucson, NV) operated via Quips FISH imageanalysis software (Vysis, Inc.).

Fig. 1. A flow cytometric DNA histogram (A) and the CGHcopy number profiles of chromosomes showing gains and losses(B) from a primary breast tumor that contained only a smallproportion (23%) of nondiploid cells of the total cell count(nondiploid DNA index, 2.8). Inpanel A, Cand T refer tochicken and trout erythrocytes, respectively, used as internalcontrols.D andNon-D,diploid and nondiploid cell populations,respectively. Inpanel B,the curves shown are mean values ofgains/losses from at least six chromosomes (bold)6 95% con-fidence interval. The threshold values for losses and gains wereset to 0.85 and 1.15, respectively. This tumor was not furtheranalyzed by FISH.

Fig. 2. Confirmation of the purity of the flow cytometrically sorted diploid and nondiploidcell clones (tumor 1).A, the original DNA flow cytometry histogram with clearly distinguish-able diploid and nondiploid cell populations.B andC, the DNA image cytometry histogramfrom the flow cytometrically sorted diploid and nondiploid clones, respectively.Insets,histograms of diploid reference cells (nonsorted cells from human cerebellum). Note that thepropidium iodide staining made for sorting has impaired the Feulgen staining, thereby shiftingboth peaks to theleft in the image cytometry histograms. The diploid DNA index was 1.03 byflow cytometry and;1.0 by image cytometry. The nondiploid DNA index was 2.11 by flowcytometry and;2.1 by image cytometry.

1215

GENE AMPLIFICATION AND ANEUPLOIDIZATION IN BREAST CANCER

Research. on March 7, 2021. © 2001 American Association for Cancercancerres.aacrjournals.org Downloaded from

CGH. CGH was done according to a protocol published previously (16,24). Briefly, tumor DNA was extracted from freshly frozen tumor tissue(Qiagen, Hilden, Germany) and labeled with FITC-dUTP and FITC-dCTP(DuPont, Boston, MA) using standard nick translation. Labeled DNAs (400–800 ng each, labeled reference DNA; Vysis) and 10mg of unlabeled Cot-1DNA (Life Technologies, Inc., Gaithersburg, MD) were hybridized ontocommercially available normal metaphase chromosomes (Vysis, Inc.). Thehybridizations were evaluated using the QUIPS digital image analysis system(Vysis, Inc.). At least five metaphases from each tumor were analyzed.

SKY. The probe mixture containing 24 differentially labeled, chromosome-specific painting probes and Cot-1 blocking DNA (SKY kit; ASI AppliedSpectral Imaging, Migdal Ha’Emek, Israel) was denatured and hybridized todenatured tumor metaphase chromosomes according to the protocol recom-mended by ASI. After hybridization and washing, the chromosomes werecounterstained with DAPI. Image acquisitions were performed using a SD200Spectracube system (ASI) mounted on a Zeiss Axioskop microscope with acustom designed optical filter (SKY-1; Chroma Technology, Brattleboro, VT).The emission spectra were then converted to the pseudocolors matching thefluorochrome combinations of each chromosome. For each cell line and clone,at least seven metaphases were analyzed (25).

RESULTS

This study was initiated by our CGH findings of breast tumors con-taining only a small proportion of aneuploid cells when studied by flowcytometry. An example of such a tumor is given in Fig. 1. The flowcytometric DNA histogram (Fig. 1A), shows only a small proportion(23%) of nondiploid cells (DNA index, 2.8); yet, the CGH made from thesame tumor specimen reveals a large number of chromosomal gains andlosses (Fig. 1B). CGH typically detects clonal chromosomal copy numberimbalances (gains and losses) when present in at least 60% of the cellsfrom which the DNA is extracted. This suggests the presence of genet-ically deviant diploid cells, because the small fraction of nondiploid cellscould not possibly be detected by CGH.

Analysis of ERBB2 and Cyclin D1 Copy Numbers in FlowCytometrically Sorted Diploid and Nondiploid Tumor Cell Pop-ulations. To study diploid and nondiploid tumor cell clones sepa-rately, the propidium iodide-stained tumor cell suspensions weresorted by their DNA content with a flow cytometer and collected onto

Fig. 3. A (diploid) and B (nondiploid) show FISH ofERBB2 oncogene (red) and chromosome 17 centromere(green; tumor 14).C (diploid) andD (nondiploid) show FISHof CCND1 oncogene (red) and chromosome 11 centromere(green; tumor 4). The amplifications of both oncogenes areclearly distinguishable in both cell clones when dividing thenumber of gene probe signals (red) by the number of centro-mere probes (green; amplification if$1.5).Lower right cor-ners, average probe:reference ratio for each sample. DAPI(blue) was used as a nuclear counterstain.

Table 1 Amplification of CCND1 oncogene in flow cytometrically sorted diploid and nondiploid tumor cells

Tumor no.DNA indexes

by FCMaDNA indexes byimage cytometry

MeanCCND1/cen11 in diploid sort

MeanCCND1/cen11 in nondiploid sort Clones amplified

1 1.03/2.11 ;1.0/;2.1 9/3 10/3.5 Both2 1.03/1.35 ;1.0/;1.7 7/2.6 8/3 Both3 1.00/1.59 ;1.0/;1.7 4/2 9/4 Both4 1.00/2.15 ND 8/2 11/3 Both5 0.98/1.72 ND 2/2 7/2 Nondiploid6 1.00/1.96 ;1.0/;1.9 3/2 7/4 Both7 0.99/1.48 ;1.0/;1.5 4/2 6/3 Both8 0.99/2.22 ND 10/2 9/3 Both9 0.99/2.22 ND 2/2 9/2.3 Nondiploid

10 1.00/1.51 ND 11/3 7/3.6 Botha FCM, flow cytometric analysis, cen; centromere; ND, not determined.

1216

GENE AMPLIFICATION AND ANEUPLOIDIZATION IN BREAST CANCER

Research. on March 7, 2021. © 2001 American Association for Cancercancerres.aacrjournals.org Downloaded from

microscope slides. Static DNA image cytometry was performed toconfirm the purity of sorted cell clones. The cell clones with the DNAindexes matching flow cytometry were found by image cytometry,and no evidence for significant impurity or contamination was found(Fig. 2 and Table 1).

The FISH hybridizations, performed on the sorted cells, indicatedthat both diploid and nondiploid cell clones often showed amplifica-tion of ERBB2andCCND1(Fig. 3). The gene copy numbers from 21sorted tumors are presented in Tables 1 and 2. Eight of 10 tumors(80%) withCCND1amplification were amplified in both diploid andnondiploid cell (copy number relative to 11 centromeres,$1.5). Intwo tumors (18%), only the nondiploid clone was found amplified(Table 1). In the 11 tumors previously determined as havingERBB2oncogene amplification, 9 (82%) showed amplification in both clonesand two in the nondiploid clone only (18%; Table 2). Tumors withamplification in the diploid clone only were never seen.

The mean copy number ofCCDN1andERBB2(per cell) remainedapproximately the same in diploid and nondiploid cells in 11 of 17tumors with amplification in both clones (tumors 1, 2, 8, 10, 15, 16,17, 18, 19, 20, and 21; see Tables 1 and 2). In the remaining cases(tumors 3, 4, 6, 7, 13, and 14; see Tables 1 and 2), the gene copynumber, as well as the chromosome copy number of the respectivereference centromere, increased during aneuploidization. Whencounting the ratio between gene copy number and chromosome copynumber, we found that it remained stable throughout aneuploidization.

We analyzed the copy numbers of the reference probes (pericen-tromeric probes for chromosomes 11 and 17, analyzed separately).Surprisingly, multiple signals were found in the diploid clones of 7tumors (tumors 1, 2, 10, 16, 17, 18, and 21; see Tables 1 and 2). Allof these tumors showedERBB2 or CCND1 amplification both indiploid and nondiploid clones. The corresponding flow cytometricDNA histograms showed a small CV (,5.2% in all but one tumor) forthe diploid DNA peaks, which is generally considered as a sign oflittle or no genetic instability.

Duplication of Chromosomal Changes as Evidenced by SKY.The order of appearance of chromosomal rearrangements was furtherstudied with breast cancer cell lines MDA-157 and MDA-436, whichboth contain two distinct nondiploid cell clones despite tens of pas-sages in culture. The flow cytometric DNA indexes were 1.31 and2.50 (79 and 21% of cells in G0-G1 peaks) for MDA-157 and 0.85 and1.70 for MDA-436 5 (77 and 23% of cells in G0-G1 peaks; Fig. 4).The modal chromosome numbers in these clones when analyzed bySKY where 54 and 95 for MDA-157 and 39 and 80 for MDA-436,respectively.

The different clones were analyzed separately by SKY, as shown inFig. 5. In both cell lines, the aneuploid (polyploid) clone contains amajority of the aberrant chromosomes of the near-diploid clone induplicate. For example, derivative chromosomes containing materialfrom chromosomes 3, 5, 12, and 20 in MDA-157 and chromosomes 1,

7, 8, and 21 in MDA-436 (Fig. 5) were duplicated in the polyploidclone. Several “new” aberrations (not present in the near-diploidclone) were also found in the polyploid clones. These include chro-mosomes 1, 4, and 9 in MDA-157 and chromosomes 2, 3, and 6 inMDA-436.

DISCUSSION

This study was initiated by our findings made by CGH in which wewere able to show copy number imbalances in breast tumor samples

Fig. 4. DNA flow cytometry histograms of the two cell lines MDA-157 (A) andMDA-436 (B). In both panels,C andT refer to chicken and trout erythrocytes, respec-tively, used as internal controls.Near-D and Non-D, near-diploid and nondiploid cellpopulations, respectively. InA (MDA-157), the near-diploid population has a DNA indexof 1.31, and it constitutes 79% of the counted G0-G1 cells. The nondiploid population hasa DNA index of 2.50 and constitutes of 21% of the counted G0-G1 cells. InB (MDA-436),the near-diploid population has a DNA index of 0.85 (hypodiploid), and it constitutes 77%of the counted G0-G1 cells. (This is the same DNA index as for the trout erythrocytes. Inthe histogram, the two populations are merged into one peak.) The nondiploid populationhas a DNA index of 1.70 and constitutes 23% of the counted G0-G1 cells.

Table 2 Amplification of ERBB2 oncogene in flow-cytometrically sorted diploid and nondiploid tumor cell clones

Tumor no.DNA indexby FCMa

MeanERBB2/cen17 in diploid sort

MeanERBB2/cen17 in nondiploid sort Clones amplified

11 0.98/1.77 2/2 30/3 Nondiploid12 0.99/1.64 2/2 18/2 Nondiploid13 0.99/1.48 4.3/1.4 18/4.4 Both14 0.97/1.64 15/2 24/3.7 Both15 1.00/1.58 16/2 18/2 Both16 1.00/1.90 12/5.2 10/4.6 Both17 1.01/2.80 18/3 10/3.5 Both18 1.01/2.93 11/4.5 11/5.3 Both19 1.00/1.86 19/2 17/2 Both20 0.98/1.78 10/1.7 18/2.3 Both21 0.99/2.06 14/3 18/3 Both

a FCM, flow cytometric analysis; cen, centromere.

1217

GENE AMPLIFICATION AND ANEUPLOIDIZATION IN BREAST CANCER

Research. on March 7, 2021. © 2001 American Association for Cancercancerres.aacrjournals.org Downloaded from

containing only a small fraction of aneuploid cells, as evidenced byDNA flow cytometry (26). CGH generally requires the sample tocontain at least 60–70% of genetically deviating cells to detect aber-ration (16). We have found several tumors with an aneuploid cellfraction of ,30%, and yet they show multiple chromosomal aberra-tions by CGH. This suggests indirectly that a significant fraction ofthe diploid cells must be genetically deviant and contain at least partlythe same genetic aberrations as the aneuploid cells.

To explore the order of appearance of genetic aberrations directly,we sorted tumor cells for their DNA content by flow cytometry andanalyzed the different clones by FISH. In 17 of 21 tumors, theamplification of ERBB2 and CCND1 was present both in DNA-diploid and nondiploid cell clones. When the diploid clone was foundamplified, typically only 25–50% of the cells showed oncogene am-plification. The presence of cells without amplification can be best

explained by the presence of nonmalignant cells,i.e.,stromal, inflam-matory, and benign epithelial cells that are present in every breasttumor. The presence of nonepithelial cells in breast carcinomas hasbeen shown previously (27). The contamination of the diploid cell sortby nondiploid cells was excluded by analyzing the DNA content ofthe cells after sorting, using DNA image cytometry. These experi-ments showed clearly that there was no contamination of nondiploidcells on the diploid sorted slides (orvice versa). On the basis of theimage cytometry data, we feel that the relatively high proportion ofcells with oncogene amplification among the diploid cells (25–50%)is very unlikely due to nondiploid cell contamination. Thus, weconclude that aneuploid primary breast tumors often contain DNAdiploid cell clone(s) that have undergone oncogene amplification.

If the gene amplification takes place already in diploid state (beforeaneuploidization), one would expect that the number of oncogene

Fig. 5. SKY of the different cell clones of breast cancercell lines MDA-157 (A) and MDA-436 (B). Chromosomesare seen in SKY classification colors.A: theupper left panelshows the near-diploid cell clone (mean number of chromo-somes from 10 metaphases, 54), and thelower panelshowsthe nondiploid (polyploid) clone (mean number of chromo-somes from 10 metaphases, 95). Note that many chromo-somes containing rearrangements in the near-diploid clonewere duplicated in the polyploid clone (e.g.,chromosomes 3,5, 12, and 20, extrapolated).B: theupper left panelshows thenear-diploid cell clone (mean number of chromosomes from10 metaphases, 39), and thelower panelshows the nondip-loid (polyploid) clone (mean number of chromosomes from10 metaphases, 80). Note also here that many chromosomescontaining rearrangements in the near-diploid clone wereduplicated in the polyploid clone (e.g.,chromosomes 1, 7, 8,and 21, extrapolated).

1218

GENE AMPLIFICATION AND ANEUPLOIDIZATION IN BREAST CANCER

Research. on March 7, 2021. © 2001 American Association for Cancercancerres.aacrjournals.org Downloaded from

copies becomes multiplied along with chromosome multiplication.This was seen only in 6 of the 17 tumors where both diploid andnondiploid cells were found amplified. Thus, in these 6 tumors, theamplicon-containing chromosomes have multiplied during aneu-ploidization. In the remaining 11 tumors, the mean number of copiesof ERBB2or CCND1was almost the same in the DNA diploid andnondiploid cells. The same was true also for chromosome 11 and 17centromere counts, which neither showed any clear increase. In thesetumors, the diploid clones already showed unexpectedly more thantwo centromere signals/cell. It is, therefore, possible that the multi-plication of amplification-carrying chromosomes has occurred alreadybefore gross polyploidization. The preserved diploid DNA content byflow cytometry can be explained by losses of other chromosomes orchromosome arms. An alternative explanation for unaltered oncogenecopy numbers is that the amplified gene copies are present in doubleminute chromosomes. However, according to cytogenetic literature,double minute chromosomes are considered to be rare in primarybreast cancer (28).

The endoreduplication after multiple chromosomal rearrangementswas best visualized by our data from SKY. The flow cytometryanalysis confirmed that the two different clones analyzed were notartifacts caused by induction metaphase cells in culture (by Colcemidtreatments). Distinct G0-G1 as well as G2 phases could be seen in bothcell lines. The SKY data from both breast cancer cell lines demon-strated many translocation events in one copy in the near-diploidclone but in two or more copies in the corresponding aneuploid clone.This result indicates that at least in these cell lines, polyploidizationdoes not give significant growth advantage over the already aneuploidcells, which can persist in culture despite tens of passages. In fact, inMDA-157, it seems that the proportion of cells in the near-diploidpopulation is increased during cell culturing (data not shown). Thus,the growth characteristics of these cell lines are likely to be deter-mined by the genetic aberrations present already in the near-diploidphase and not by aneuploidization.

Taken together, our present results demonstrate that cancer cellsthat are flow cytometrically diploid in their total DNA content oftenpersist in primary breast cancers that have undergone aneuploidiza-tion. These results are parallel to cytogenetic observations (7), whichalso suggest aneuploidization to be a late genetic event. However,cytogenetic results are indirect, and artificial clonal selection duringinvitro culture cannot be ruled out. Our present data from unculturedinterphase cells showed more specifically that oncogene amplifica-tion, which is biologically and prognostically a very important geneticdefect in breast cancer, takes place before aneuploidization. Thepresence of diploid and nondiploid malignant cells is also a clearindication of heterogeneity in the genetic composition of malignanttumor cells.

ACKNOWLEDGMENTS

We thank Ghita Fallenius for skillful technical assistance with the DNAimage cytometry measurements.

REFERENCES

1. Devilee, P., and Cornelisse, C. Somatic genetic changes in human breast cancer.Biochim. Biophys. Acta,1198: 113–130, 1994.

2. Bieche, I., and Lindereau, R. Genetic alterations in breast cancer. Genes Chromo-somes Cancer,14: 227–251, 1995.

3. van de Vijver, M. J. Molecular genetic changes in human breast cancer. Adv. CancerRes.,61: 25–56, 1993.

4. Kinzler, K., and Vogelstein, B. Lessons from hereditary colorectal cancer. Cell,87:159–170, 1996.

5. Devilee, P., Schuuring, E., van de Vijver, and Cornelisse, C. Recent developments inthe molecular genetic understanding of breast cancer. Crit. Rev. Oncog.,5: 247–270,1994.

6. Kuukasjarvi, T., Tanner, M., Pennanen, S., Karhu, R., Kallioniemi O-P., and Isola,J. Genetic changes in intraductal breast cancer detected by comparative genomichybridization. Am. J. Pathol.,4: 1465–1471, 1997.

7. Dutrillaux, B., Gerbault-Seureau, M., Remvikos, Y., Zafrani, B., and Prieur, M.Breast cancer genetic evolution. I. Data from cytogenetics and DNA content. BreastCancer Res. Treat.,19: 245–255, 1991.

8. Tanner, M., Karhu, R., Nupponen, N., Borg, Å., Baldetorp, B., Pejovic, T., Ferno, M.,Killander, D., and Isola, J. Genetic aberrations in hypodiploid breast cancer. Frequentloss of chromosome 4 and amplification ofcyclin D1oncogene. Am. J. Pathol.,153:191–199, 1998.

9. Locker, P., Horrocks, C., Gilmour, A. S., Ellis, I. O., Dowle, C. S., Elston, C. W., andBlamey, R. W. Flow cytometric and histological analysis of ductal carcinomain situof the breast. Br. J. Surg.,77: 564–567, 1990.

10. Ottosen, G. L., Christensen, I. J., Larsen, J. K., Kerndrup, G. B., Hansen, B., andAndersen, J. A. DNA ploidy in early breast cancer. Br. J. Cancer,72: 832–839, 1995.

11. Leal, C. B., Schmitt, F. C., Bento, M. J., Maia, N. C., and Lopes, C. S. Ductalcarcinomain situof the breast. Histologic categorization and its relationship to ploidyand immunohistochemical expression of hormone receptors, p53, andc-erbB-2 pro-tein. Cancer (Phila.),75: 2123–2131, 1995.

12. Li, R., Sonik, A., Stindl, R., Rasnick, D., and Deusberg, P. Aneuploidyvs. genemutation hypothesis of cancer: recent study claims mutation but is found to supportaneuploidy. Proc. Natl. Acad. Sci. USA,97: 3236–3241, 2000.

13. Maquire, H. C., Jr., Hellman, M. E., Greene, M. I., and Yeh, I. Expression of c-erbB-2in in situ and in adjacent invasive ductal adenocarcinoma of the female breast.Pathobiology,60: 177–121, 1992.

14. Watson, P. H., Safneck, J. R., Le, K., Dubuk, D., and Shiv, R. P. Relationship ofc-mycamplification to progression of breast cancer fromin situ to invasive tumor andmetastasis. J. Natl. Cancer Inst.,85: 902–907, 1993.

15. Vos, C. B., Ter Haar, N. T., Peterse, J. L., Cornelisse, C. J., and van de Vijver, M. J.Cyclin D1 gene amplification and overexpression are present in ductal carcinomainsitu of the breast. J. Pathol.,187: 279–284, 1999.

16. Kallioniemi, O. P., Kallioniemi, A., Piper, J., Isola, J., Waldman, F. M., Gray, J. W.,and Pinkel, D. Optimizing comparative genomic hybridization for analysis of DNAsequence copy number changes in solid tumors. Genes Chromosomes Cancer,10:231–243, 1994.

17. Borg, Å., Baldetorp, B., Ferno, M., Killander, D., Olsson, H., and Sigurdsson, H.ERBB2 amplification in human breast cancer with a high rate of proliferation.Oncogene,6: 137–143, 1991.

18. Borg, Å., Sigurdsson, H., Clark, G. M., Ferno, M., Fuqua, S. A. W., Olsson, H.,Killander, D., and McGuire, W. L. Association of INT2/HST1 coamplification inprimary breast cancer with hormone-dependent phenotype and poor prognosis. Br. J.Cancer,63: 136–142, 1991.

19. Baldetorp, B., Ferno, M., Fallenius, A., Fallenius-Vecchi, G., Idvall, I., Olsson, H.,Sigurdsson, H., Åkerman, M., and Killander, D. Image cytometric DNA analysis inhuman breast cancer analysis may add prognostic information in diploid cases withlow S-phase fraction by flow cytometry. Cytometry,13: 577–585, 1992.

20. Hedley, D. W., Clark, G. M., Cornelisse, C. J., Killander, D., Kute, T., and Merkel,D. Consensus review of the clinical utility of DNA cytometry in carcinoma of thebreast. Cytometry,14: 482–485, 1993.

21. Decosse, J. J., and Aiello, N. Feulgen hydrolysis: effect of acid and temperature.J. Histochem. Cytochem.,14: 601–604, 1966.

22. Gaub, J., Auer, G., and Zetterberg, A. Qualitative cytochemical aspects of a combinedFeulgen-naphtol yellow S staining procedure for the simultaneous determination ofnuclear and cytoplasmic proteins and DNA in mammalian cells. Exp. Cell Res.,92:323–332, 1975.

23. Baldetorp, B., Ferno, M., Bendahl P-O., Fallenius, G., Idvall, I., Åkerman, M., andKillander, D. Proliferative index obtained by DNA image cytometry–Does it addprognostic information in Auer IV breast cancer? Analyt. Quant. Cytol.,20: 144–152,1998.

24. Karhu, R., Kahkonen, M., Kuukasjarvi, T., Pennanen, S., Tirkkonen, M., andKallioniemi, O. Quality control of CGH: impact of metaphase chromosomes and thedynamic range of hybridization. Cytometry,28: 198–205, 1997.

25. Kytola, S., Rummukainen, J., Nordren, A., Karhu, R., Farnebo, F., Isola, J., andLarsson, C. Chromosomal alterations in 15 breast cancer cell lines by comparativegenomic hybridization and spectral karyotyping. Genes Chromosomes Cancer,28:308–317, 2000.

26. Persson, K., Pandis, N., Mertens, F., Borg, Å., Baldetorp, B., Killander, D., and Isola,J. Chromosomal aberrations in breast cancer: a comparison between cytogenetics andcomparative genomic hybridization. Genes Chromosomes Cancer,25: 115–122,1999.

27. Wingren, S., Stål, O., Carstensen, J., Sun, X-F., and Nordenskjold, B. S-phasedetermination of immunoselected cytokeratin-containing breast cancer cells improvesthe prediction of recurrence. Breast Cancer Res. Treat.,29: 179–187, 1993.

28. Mitelman, F., Johansson, B., Mandahl, N., and Mertens, F. Clinical significance ofcytogenetic findings in solid tumors. Cancer Genet. Cytogenet.,95: 1–8, 1997.

1219

GENE AMPLIFICATION AND ANEUPLOIDIZATION IN BREAST CANCER

Research. on March 7, 2021. © 2001 American Association for Cancercancerres.aacrjournals.org Downloaded from

2001;61:1214-1219. Cancer Res   Karin Rennstam, Bo Baldetorp, Soili Kytölä, et al.   CancerPrecede Aneuploidization in the Genetic Evolution of Breast Chromosomal Rearrangements and Oncogene Amplification

  Updated version

  http://cancerres.aacrjournals.org/content/61/3/1214

Access the most recent version of this article at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/61/3/1214.full#ref-list-1

This article cites 27 articles, 1 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/61/3/1214.full#related-urls

This article has been cited by 12 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/61/3/1214To request permission to re-use all or part of this article, use this link

Research. on March 7, 2021. © 2001 American Association for Cancercancerres.aacrjournals.org Downloaded from