[CANCER RESEARCH55, 5400-5407. November 15, 19951

amplified ERBB-2 in tumor DNA, rare breast tumors overexpresspl85F@1t2protein or c-ERBB-2 mRNA levels in the absenceofdetectable gene amplification (7).

Although amplification of ER.8B-2 is generally considered to be asignificant prognostic indicator in patients with breast cancer, itsapplicability continues to be controversial, in part because of analytical discrepancies associated with the methods traditionally used toevaluate its amplification and/or overexpression. These techniquesinclude Southern blotting, slot blot analysis, and FISH4 for detectionof amplification, while ELISA, Western blotting, immunohistochemistry, and immunofluorescence are used to evaluate overexpression(8—15).Because FISH allows the observer to distinguish small subpopulations of amplified cells, it is more sensitive than blottingtechniques. In addition, FISH allows one to identify particular locations where aberrations exist in single tumor specimens (10, 14).Similarly, because immunohistochemically stained slides are difficultto quantify and because ELISA and Western blotting data do notprovide information concerning heterogeneity, immunofluorescencehas advantages over these other methods (13, 14).

Marked heterogeneity has been described in primary breast cancersin both the copy number of ERBB-2/cell and in the level of p185@@E1t2protein (5—12).Although cell-to-cell differences may be due in part toanalytical variation, genetic and epigenetic dispersion may also playsignificant roles. This heterogeneity provides a potential source for theselection of subclones with increased malignant and metastatic potential, especially in the context of therapeutic targeting based onERBB-2 expression.

Although amplification of ERBB-2 correlates well with overexpression of pl85@@E1t2protein in cell populations (5, 6, 9, 11, 14—16),thecorrelation has not been made on a cell-by-cell basis. The presentcommunication describes our analysis of the extent to which ERBB-2gene amplification relates to the expression of p185@1t2 on a singlecell basis. We have found that although pl85@@1t2 expression correlates with the ERBB-2 copy number/cell, p185@@1t2expression correlates poorly with the ERBB-2 copies:ch.romosome 17 copies ratio.Surprisingly, there was correlation between p185@1t2 expression andchromosome 17 copy number, suggesting that hyperploidy may berelated to the p185HER2 expression.

MATERIALS AND METhODS

Cell LInes. Humanbreastcancercell lines, BT-474, SK-BR-3, MDA-453,and MCF-7 were obtained from the American Type Culture Collection (Rockville, MD) and grown according to their specifications. The four cell lines werecharacterized previously for ERBB-2 gene amplification (10). For flow cytometric immunofluorescence measurements, cells were harvested either bytrypsin or 25 mM EDTA in PBS (pH 7.2; Ref. 17). For slide-based immuno

fluorescence measurements, cells were cultured in slide chambers (Nunc, Inc.,Naperville, IL). For BrdUrd (Sigma Chemical Co., St. Louis, MO), labelingcells were pulsed with 100 @MBrdUrd for 60 mm. Cells were washed threetimes with PBS containing 1 mMCaC12before immunofluorescence labeling.

Tumor. A biopsy froma nodepositive,T2tumor,was frozenimmediatelyafter resection. Imprint preparations were made after thawing by gently touch

4 The abbreviations used are: FISH, fluorescence in situ hybridization; BrdUrd, bro

modeoxyuridine; chr, chromosome; F!, fluorescence index.

5400

ERBB-2 (HER2Ineu) Gene Copy Number, @185HER2Overexpression, and

Intratumor Heterogeneity in Human Breast Cancer'

János Szöllosi,2 Margit Ba1IZS,2 Burt G. Feuerstein, Christopher C. Benz, and Frederic M. Waldman3

Division of Molecular Cytometry, Department of Laboratory Medicine If. S., B. M., B. G. F., F. M. W.]. Brain Tumor Research institute (B.G.F.], and Cancer Research Institute(C. C. B.], University of California, San Francisco, California 94143-0808

ABSTRACT

Amplification of the ERBB-2 (HER-2/neu) gene Is accompanied byoverexpression of Its ceH surface receptor product, pl85―@2. Heterogeneity has been observed for both the gene copy number and the level of

overexpresslon ofits proteIn product. To better understand their relationship, correlation between the level of cellular expression of pf85@@an2andERBB-2 gene amplification was studied in four human breast cancer celllines (BT-474, SK-BR-3, MDA-453, and MCF-7) and in a primary humanbreast tumor sample. The relative expression of pl85@'t2 was measured

by hnmunofiuorescence by using flow and/or image cytometry whilecorrelated DNA analysis was performed on the same cells by fluorescencein situ hybridization to determine ERBB-2 gene and chromosome 17 copynumbers. Marked heterogeneity was observed in both protein expressionand ERBB-2 copy number. Despite this heterogeneity, and in accordancewith previous studies, the average levels of p1ss@R2 expression corre.lated well with average ERBB-2 gene copy numbers in the four linesexamined (r = 0.99). When the relationship between copy number andproteIn expression was studied on a cell-by-cell basis, p185―@2 expression correlated with both the absolute number ofERBB-2 gene copies/cell(r 0.59—0.63)and chromosome 17 copy number (r 0.45—0.61).It Is ofInterest that there was weak or no correlation between p185―@2proteinexpression and the ERBB-2 copy number:chroinosome 17 copy numberratio (r = 0.0-0.25). In more than one-half of cells expressing a high levelof p185@@E@@2,the chromosome 17 copy number was@ (two or threetimes the average copy number), whereas <2% of an unselected population had a high chromosome 17 copy number. Bromodeoxyuridine incorporation indicated that the S-phase-labeling index was homogeneousacross various p185'@2-expressing subpopulations In the SK-BR-3 cellline. Analysis of the primary breast tumor sample showed results similarto the cell lines, supporting the strong possibility of a mechanistic linkamong p185@2 overexpresslon, ERBB-2 amplification, and high chromosome 17 copy number.

INTRODUCTION

A characteristic feature of cancer cells is the unregulated expressionof genes involved in cellular growth control. One of these genes is theERBB-2 (HER-2/neu) proto-oncogene, which encodes a Mr 185,000transmembrane glycoprotein @l85@.@@2)that belongs to a subfamilyof growth factor receptors having intrinsic tyrosine kinase activity,including the epidermal growth factor receptor and the receptorsHER-3 and HER-4 (1-3).

Amplification and overexpression of ERBB-2 is found in 25—30%of primary human breast cancers and is associated with a poor clinicaloutcome (4—6).This suggests that overexpression of pl85}@1t2 playsa role in the pathogenesis of some human breast cancers (5, 6).Although overexpression of p185HER2 is usually accompanied by

Received 6/8/95; accepted 9/14/95.The costs of publication of this article were defrayed in part by the payment of page

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

I This research was supported by NIH Grants CA-49056 and CA-44768 and by United

States-Hungarian Joint Fund for Science and Technology (JF292/92B).2 Present address: Department of Biophysics, Medical University School of Debrecen,

Nagyerdei krt. 98, H-4012 Debrecen, Hungary.3 To whom requests for reprints should be addressed, at Department of Laboratory

Medicine, MCB-230, Box 0808, University of California at San Francisco, San Francisco,CA 94143-0808. Phone: (415) 476—3821; Fax: (415) 476—8218; E-mail: waldman@dmc.ucsf.edu.

on April 8, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ERBB.2 EXPRESSION AND AMPLIFICATION

ing the slide surface with tumor material. Slides were then fixed in 1%formaldehyde for 60 mm at room temperature and subsequently fixed andstored in 70% ethanol. The autofluorescence of air-dried touch imprint preparations was too high for reliable immunofluorescence analysis. Fresh fixationof slides in 1% formaldehyde and subsequently in 70% ethanol reducedautofluorescence significantly.

Immunolabeling. For flow cytometry,unfixed trypsinizedcells were incubated with 5 ,.tg/ml mAbi (Triton, Alameda, CA) raised against the extracellular domain of p18SHER2,in the presence of 1% BSA on ice for 45 mm,washed three times with PBS, and incubated with fluoresceinated rabbitantimouse IgG (1:100 dilution; Sigma) at 0°Cfor 45 mm. After washing with

PBS, cells were fixed in 1% formaldehyde solution and stored for not morethan 3 weeks at 4°C before analysis.

For image analysis, cells were first fixed in 0.5% formaldehyde solution for20 mm at room temperature and in 70% ethanol at 4°C overnight. Cells on

slides could be stored in ethanol at 4°Cfor not more than 2 months. Slides werethen preblocked in 5% Carnation dry milk, 0.1% Triton X-100 in 4X SSC (1X

SSC is 0.15 M NaCl and 0.015 M sodium citrate) for 10 mm at roomtemperature. Staining was at room temperature for 45 mm. Samples were firstincubated with CB1I antibody (BioGenex, San Ramon, CA) specific to theintracellular domain of the p185HER.2protein, diluted (1:200) in the blockingbuffer, washed twice with the blocking buffer, and incubated with fluorescein

ated rabbit antimouse IgG (1:100; Sigma). After washing, samples wererefixed in 1% formaldehyde solution in PBS and kept at 4°Cfor not more than3 weeks before microscopic analysis. During this time, no significant deteri

oration of the fluorescence signal was observed.To control for nonspecific staining, cells were preincubated with irrelevant

monoclonal antibody of the same isotype before staining with fluoresceinconjugated rabbit antimouse IgG. We also compared immunofluorescencelabeling of MDA-453 and SK-BR-3 cells harvested with either trypsin or 25mM EDTA in PBS. Trypsinization caused a 10—15% loss in fluorescence

intensity as compared to cells harvested with 25 mr@iEDTA (data not shown).Because this loss was not significant, and the two other cell lines could not beharvested with 25 mM EDTA, we used trypsin to harvest cells for flowcytometric analysis. Results from monoclonal antibody (mAbi) raised againstthe extracellular domain of @l85H@.2 protein were similar to those from

monoclonal antibody (CB11) raised against the intracellular domain of the

protein. With mAbI , prefixation was unnecessary, resulting in lower non

specific binding.Flow Cytometry. Cell suspensions were filtered through a 35-g.@mnylon

mesh to remove aggregates before flow cytometric analysis. Analysis wasperformed on a FACScan flow cytometer (Becton Dickinson, San Jose, CA)equipped with a 15 mW argon laser (488 nm) and pulse-width doubletdiscrimination. A total of 10,000 events were recorded in list mode afterlogarithmic amplification of the fluorescence signal.

Digital Image Analysis. The fluorescence of cells stained on slides wasanalyzed by using a digital image analysis system based on a Zeiss Axioplanmicroscope equipped with the Microlmager 1400 Digital camera (Xillix Technologies Corp., Vancouver, British Columbia, Canada). Images were capturedthrough a fluorescein excitation filter, beam splitter, and emission filter byusing a X 20, NA: 0.5 Plan Neofluar objective. Images were processed and

quantitatively analyzed with a Sun IPX workstation using Scil-Image software

(National Research Institute, Delft, The Netherlands). Local background fluorescence was determined for each image, and the average autofluorescence ofthe isotypic control cells was subtracted from the total fluorescence intensity oflabeled cells. The Fl was defined as a ratio of the corrected total fluorescenceintensity of labeled cells to the mean autofluorescence of the isotypic controlcells.

DNA Probes and Probe Labeling. Two contiguous ERBB-2 cosmidclones (cRCNeul and cRCNeu4), together spanning 55 kb of genomic DNA(10), were used in combination with a probe specific for the chromosome 17

pericentromeric sequence (p17H8; Ref. 18) for two-color FISH analysis.Probes were directly labeled with fluorescein-11-dUTP or tetramethylrhodaminc-I 1-dUTP (Boehringer Mannheim, Indianapolis, IN) by nick translation by

using commercially available kits (Bethesda Research Laboratories, Gaithersburg, MD).

fluorescence in Situ Hybridization and Staining for BrdUrd. Dualanalysis of gene copy number and protein expression was done as a two-stageprocedure. Slides were first stained for protein expression and fluorescence

images acquired; they were then hybridized for gene copy number and scoredafter relocating cells analyzed previously. After immunofluorescence analysis,slides were refixed in methanol:acetic acid (3:1; Carnoys solution) and air

dried. FISH was performed as described previously (19) with modifications.Briefly, cells on slides were denatured in 70% formamide-2X SSC at 73°Cfor3.0 mm, dehydrated in graded ethanols, treated with 0.25 @g/mlproteinase K(Sigma) in 20 mMTRIS buffer (pH 7.5) containing 2 mt@iCaCl2 for 7.5 min at37°C,and again dehydrated. The hybridization mixture was denatured at 73°Cfor 5 min, reannealed for 30 mm at 37°C,and applied to warmed slides. Tenp@lof hybridization mixture contained 6 ng of fluoresceinated chromosome 17centromeric probe, 34 ng of rhodaminated ERBB-2 probe, and 10 ng ofunlabeled, sonicated human placental DNA (Sigma) in 50% formamide, 2XSSC, and 10%dextransulfate. Hybridizationwas overnightat 37°C.Slideswere washed three times for 10 mm each in 55% formamide-2X SSC, once in2X SSC at 45°C,and once in 2X SSC at room temperature. Nuclei werecounterstained by using 4',6-diamidino-2-phenylindole hydrochloride (Molecular Probes, Eugene, OR) at 0.01 g@g/m1in antifade solution (20).

Simultaneous detection of BrdUrd incorporation and dual FISH stainingwas performed with three fluorescent dyes (fluorescein- and rhodamine-labeled probes and a Cascade Blue-conjugated antibody; Ref. 21). Cells andprobes were denatured and hybridized as described above. After washing,slides were preblocked in 5% Carnation dry milk and 0.1% Triton X-100 in 4XSSC for 10 min at room temperature.All staining reactionswere at roomtemperature for 30 min. Slides were incubated with IU4 mouse anti-BrdUrd(1:400; Caltag, La Jolla, CA), diluted in blocking buffer, washed twice withblocking buffer, and incubated with Cascade Blue-antimouse IgG (1:300;Molecular Probes), and coverslipped with antifade solution alone.

Scoring of Interphase Nuclei. Cells analyzed previously for cell surfaceexpression of p185HER2 protein were relocated on the basis of their coordinates and scored for chromosome 17 and ERBB-2 signals by using a X100NA:1.3 Plan Neofluar oil immersion objective and a computer-controlledstage. ERBB-2 doublets were counted as separate signals. Broken, torn,squashed, smeared, or overlapping nuclei were ignored. Each hybridization

was accompanied by a control hybridization using normal lymphocytes. Thescoring results were expressed both as an absolute ERBB-2 copy number/celland as the ERBB-2copy number relative to the 17 centromere copy number.

Three color images were acquired by using the digital imaging analysissystem described previously. A triple-band-pass beam splitter and emissionfilters were used (22). Excitation of each fluorochrome was accomplished byusing single-band-pass excitation filters in a computer-controlled filter wheel.This made it possible to collect sequential, properly registered images of the

three fluorochromes (4',6-diamidino-2-phenylindole hydrochloride or CascadeBlue, fluorescein, and rhodamine). The three-color images were processedwith a Sun IPX workstation using Scil-Image software for pseudocolor display.

Statistical Analysis. Significance levels for differences in gene copy number between the @185HER.2bright and total cell populations were determined

by contingency table analysis.

RESULTS

ERBB-2 Amplification and Expression in Breast Cancer CellLines Four breast cancer cell lines, MCF-7, MDA-453, SK-BR-3,and BT-474, known to have various levels of amplification of theERBB-2 gene (10) were studied for distribution of ERBB-2 gene copynumber and chromosome 17 centromere copy number (Fig. 1). Amplification of the ERBB-2 gene can be expressed as copy number/cellor as copy number relative to chromosome 17 copy number. Using arelative measure is especially important for those cell lines that areaneusomic for chromosome 17. Amplification of ERBB-2 gene wasobserved in MDA-453, SK-BR-3, and BT-474 cell lines, using eitherthe definition of amplification as total ERBB-2 copies/cell or the ratioof ERBB-2 copy number to chromosome 17 copy number. There wasmarked heterogeneity for ERBB-2 copy number, chromosome 17 copynumber, and their ratios in the three cell lines with ERBB-2 amplification. In MCF-7, the ERBB-2 gene was deleted (ERBB-2 gene copynumber was less than the chromosome 17 copy number/cell) and therewas less heterogeneity in ERBB-2 gene copy number/cell and in the

5401

on April 8, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Table 1ERBB-2 amplification and expression in breast cancer celilinesBreast

cancercell linesERBB-2'@Chr 17bERBB-2/Chr17FPMCF-72.2

±05d3.8 ±1.00.6 ±0.220±9MDA-45311.0±3.94.1±1.62.8±1.0186±75SK-BR-331.0

±9.06.9 ±1.04.5 ±1.2326 ±114BT-47452.0

±11.36.0 ±1.19.0 ±2.3549 ±165

ERBB.2 EXPRESSION AND AMPLIFICATION

ERBB-2:chromosome 17 ratio. The mean values and the SDs of thecopy number distributions are summarized in Table 1.

We next characterized the expression levels of the ERBB-2 geneproduct p185HER2 by flow cytometry. Fig. 2 shows the fluorescenceintensity histograms of the four cell lines labeled with mAbi againstp185@@E@2. Heterogeneity of expression of p185'@@2 was similar in

the four cell lines. The MCF-7 cell line was the least positive, onlytwice background, whereas the BT-474 cells were the most positive.

I-..

U BT-474U MDA-4530 SK-BR-3

@ MCF-7

,n © an © ,,@@ In © $fl 0 V@——e@ ri@ @‘i‘@‘‘@‘Siin

ERBB-2 Signals/Cell

in m in@@ N t'@Fig. 2. Frequency distribution of fluorescence intensity after immunofluorescence

staining for p1@5HER.2.Trypsinized cells were labeled with mAbi raised against theextracellular domain of @185UER.2and then with fluorescein-conjugated rabbit antimouse

B IgG.Anirrelevantprimaryantibodyofthesameisotype,followedbyfluoresceinconjugated rabbit antimouse IgG, was used for the blank control (SK-BR-3 cells). Themean values of these distribution curves are summarized in Table 1. Note that the levelof heterogeneity (width of the intensity profiles on this log intensity scale) is similar in thefour cell lines, although the absolute amount of p185'@'@2 varies greatly from line to line.c)

0

: II

The mean values and the SDs of the fluorescence intensity histogramsare summarized in Table 1. The mean fluorescence intensity wasstrongly correlated with the mean ERBB-2 copy number/cell(r 0.99; Table 1). A strong correlation was also observed betweenthe mean protein expression and mean ERBB-2:chromosome 17 ratio(r 0.99), whereas there was a weaker correlation with averagechromosome 17 copy number (r = 0.75).

ERBB-2 Gene Expression and Amplification on a Single CellBasis. Protein expression and copy number were measured in thesame individual cells to study their correlation on a single cell basis.This was especially relevant given the wide range in both copynumber and immunofluorescence observed (Figs. 1 and 2). Immunofluorescence intensity of individual SK-BR-3 and MDA-453 cells wasstudied by image microscopy, and the same cells were identified andscored for ERBB-2 gene and chromosome 17 copy number after dualFISH labeling. The fluorescence intensity was too low in MCF-7 toperform quantitative image cytometry, and BT-474 cells could not beseparated from each other during image analysis because of theirpiled-up growth pattern.

Correlated measurement of @185Han@2expression and ERBB-2copy number was performed in the same cells by consecutive analysis(Fig. 3). The fluorescence images of cells displayed in Fig. 3B areshown after double-target hybridization in Fig. 3C. The green signalscorrespond to chromosome 17 centromere, and the red signals toERBB-2 signals. The heterogeneity of pl85I@@t2 expression in 5K-BR-3 cells by image microscopy (Fig. 3A) was similar to that foundby flow cytometry (Fig. 2).

The linked analysis of p185H@2 expression and ERBB-2 geneamplification in SK-BR-3 cells is shown in Figs. 4 and 5. Note the use

of a Fl for these measurements, rather than absolute intensity (as wasused for the flow measurements), in order to control for the increasedlevels of autofluorescence in these fixed samples. ERBB-2 copynumber showed a significant correlation with protein expression on acell-by-cell basis. The correlation was stronger using absoluteERBB-2 copy number/cell (Fig. 4A) than when using a relative

40

20

•1•2

ERBB-2 Signals/Chr 17 Signals

Fig. 1. Number of ERBB-2and chromosome 17centromere copies in four breast cancercell lines. A, frequency distribution of ERBB-2 signals/cell; B, chromosome 17 signals]cell; C, ERBB-2:chromosome (Chr) 17 ratio. The values along the abscissa represent thelower limits of the range of values for each category. At least 100 cells were scored tocreate the distribution histograms. The mean values and their SDs are summarized inTable 1. Note the wide heterogeneity present in all but the MCF-7 distributions.

a ERBB-2 copy number/cell.

b Chr, chromosome 17 copy number/cell.C Mean fluorescence intensity determined from flow cytometric histograms.

d Data expressed as mean, ±SD.

5402

‘,@

.@@

E

z—

A L)

U . .—--@@‘ r— r“ ;-@-..

0 1 2 3 4 5 6 7 8 9 10 11 12

Chr 17 Signals/Cell

80 C

60

Lu1@@@@.

0 0.5 1 2 3 4 5 6 7 8 9 10 11 12 13

10 100Fluorescence Intensity

on April 8, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

a?

4/

ERBB.2 EXPRESSION AND AMPLIFICATION

A B

I

4,

7

C

I

,@ ..,

0‘4

C

..

Fig. 3. Linked detection of pl85@@ER2expression and gene amplification in individual cells. A, SK-BR-3 cells display immunofluorescence staining for pi85@R2 expression (X20objective) after staining with mAbl. B, computer magnification (X5) of the rectangle in A. C, FISH detection of ERBB-2 (red) and chromosome 17 centromeres (green) in identicalcells shown in B (X 100 objective). Cells were refixed after immunofluorescence labeling and denatured and hybridized with directly labeled ERBB-2 and chromosome 17centromere-specific probes. Not all signals are visible in this image because the plane of focus is thinner than the specimen. Anti-BrdUrd labeling (blue) is positive in the top cells.These are pseudocolor, contrast-enhanced digital images.

measure of ERBB-2 amplification (ERBB-2:chromosome 17 copy to phenotypic dispersion, the genetic composition of “bright―cells (withnumber ratio; Fig. 4B). There was also a correlation seen between more than four times more fluorescence intensity than the nonspecificp185H@@@2expression and copy number of chromosome 17 (Fig. 4C), staining of isotypic control cells) was analyzed as a separate group. Weperhaps due to an second association between aneuploidy and ERBB-2 compared the distribution of ERBB-2 copy number (Fig. 5A), ERBB-2:amplification. chromosome 17 copy number ratio (Fig. SB) and the chromosome 17

A subpopulation of cells was seen, which stained especially brightly copy number (Fig. SC) of bright cells to an unselected population andfor p185H@2•To test whether this was due to genetic heterogeneity or found that these differences were all highly significant.

5403

on April 8, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

C

p < 0.0001

40

20

ERBB-2 EXPRESSION AND AMPLIFICATION

100

10

1

0

significant difference (P = 0.29) in the average fluorescence intensity.A S-phasecellshada higheraverageERBB-2genecopynumberand

0 ERBB-2:chromosome 17 copy number ratio than did non-S-phase

cells (43.7 versus 38.7 and 6.3 versus 5.0, respectively), perhapsbecause doublets forming during DNA synthesis were scored as twoseparate gene copy numbers as described in “Materialsand Methods.―The labeling index of the whole cell population (39.7% of 224 cells)

r=0.60 . .and for cells with >10 chromosome 17 copies (38.3% of 60 cells) did‘ . 1'@0 not differ.

ERBB-2 Gene Expression and Amplification on a PrimaryTumor Sample. The results of consecutive analysis of pl85H@@2expression and ERBB-2 gene amplification in primary tumor cells

B (caseno.B372)areshowninFigs.8and9.Positivecorrelationswerefound between p185HER2 expression and ERBB-2 gene copy number(Fig. 8A), the ERBB-2:chromosome 17 ratio (Fig. 8B), and the chromosome 17 copy number (Fig. 8C). There were significant differences in the distribution of the ERBB-2 copy number (Fig. 9A), theratio of ERBB-2:chromosome copy number (Fig. 9B), and the chromosome 17 copy number (Fig. 9C) when bright cells were compared

I'2 1'4 to the unselected population. In general, the correlation patterns

observed in this touch imprint preparation were similar to thoseobserved in tumor cell lines.

C

SK-BR-3

0 20 40 60 SO 100 120ERBB-2 Signals/Cell

0 0@ 000

0 0@6'@0@

@ °0 0@ 0 0 0@ 0 0

i@'6'@o@t'@ 0F = 0.25

‘0

‘I

0

100

10

1

0 2 4 6 8 10ERBB-2 Signals/Chr 17 Signals

100

10

I

0

0

000 o 0

A 0

0900

40

20

A0

r = 0.45

U AverageD Bright(Fl>4)

p < 0.00010 10 20 30

Chr 17 Signals/Cell

Fig. 4. ERBB@2gene expression and amplification in single SK-BR-3 cells. Expression in Q in Q in ©in@ in ©in@ in@ in ©in@ in@ in ©inlevel of p185@@E@2protein@ plotted against: A, ERBB-2 copy number; B, ERBB-2: ,, @,@ ,,@@ .@@@@@@@@@

Vchromosome 17 ratio; and C, chromosome 17 copy number. Cells were labeled with ERBB-2 Signals/Cellantibody (CB11) against the intracellular domain of p18SHER2protein. Data from 184cells are shown. There is a good correlation between p185―@2 expression and copynumber of either ERBB.2 or chromosome 17 centromere (A and C). The correlationbetween @185HER.2expression and ERBB.2:chromosome 17 copy number ratio was@ I Bweaker(B). ,.@ I

ERBB-2 gene amplification and protein expression (Figs. 6 and 7). b@ ] I LI LL@m@@@

c) 20@@ p<0.002

0 IThe results in MDA-453 cells were similar to SK-BR-3 cells for@ @o-I

However, there was no correlation between protein expression and I01

ERBB-2:chromosome 17 copy number ratio (Fig. 6B), and the distri-@ if@©@lf@©@in@@ in ©@I(@@ in@@ in@@ V@Q@in@©@©@ ,- — r-@ e@ e@ r')@@ in in@ @cr-.@@@@ ©

bution of ERBB-2:chromosome 17 ratio of bright cells did not differERBB.2 Signals/Chr 17 Signals

significantly from the unselected population (Fig. 7B). In both cell

lines, the centromere 17 copy number was high (two or three times theaverage copy number) in >50% of the bright cells (expressing a highlevel of p185HER2), whereas <2% of the unselected population had a 60high chromosome 17 copy number.

Relationship between DNA Synthesis and ERBB-2 Gene Expression and AmplificatiOn. We next addressed the issue ofwhether the bright cells having high pl85'@@2 expression and highchromosome 17 copy number were proliferatively active. SKBR-3cells were pulse labeled with BrdUrd, p185H@@@@2expression was © e@r.@@ in @cr..@@ o _ ei@@ an @ar. aodetermined before fixation, and then BrdUrd incorporation and dualcolor FISH were detected simultaneously (for demonstration see Fig. @‘17 SignalS/Cell3C) Correlation between ERBB-2 gene amplification and protein Fig.5. ERBB.2copy number(A), ERBB.2:chromosome(Chr) 17 ratio(B), andexpression in these cells (data not shown) was similar to that found in chromosomei7 copy number(C) in unselectedand in highlyp185@52-expressing

(Fl > 4) SK-BR-3 cells. Distributions of bright cells were determined from cells plottedprefixed cells (Figs. 4 and 5). When BrdUrd-positive cells (cells in S in Fig.4. Cellsthatexpressedp185HER2at a highlevelhadsignificantlymorecopiesofphase) were compared with BrdUrd-negative cells, there was no ERBB-2andchromosome17,anda higherratioof thetwo,thananunselectedpopulation.

5404

SK-BR-3

00

c@0 oo0 ° o@°°8o@

o@@d?

008°@°

00

0

on April 8, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1 2 3 4 5 6 7

ERBB-2 Signals/Chr 17 Signals

C0

0

8 °0i0@0@ 000

00

0@ o° 0

e@- . -

ERBB.2 EXPRESSION AND AMPLIFICATION

linkage between gene copy number and protein expression was still100 A present within each cell line (r = 0.59—0.72) but was weaker than that

0 observed when cell lines were compared at the population level

10 0 @0 (r 0.99).Several reasons may account for the observed dissociation between

ERBB-2 copy number and protein expression. A normal dispersion inI ERBB-2 transcription and translation rates, or in the half-lives of RNA

r = 0.63 transcripts and protein products, might lead to a weakened linkage.1 - ‘ . , . , . I@ • between genotype and phenotype. ERBB-2 transcript levels in ampli

0 10 20 30 40 50 fled SK-BR-3 and BT-474 cells are 20—40times that of immortalized

ERBB-2 Signals/Cell but nontumorigenic HBL-100 breast epithelial cells, although the

average level of ERBB-2 gene copy number in the amplified cell lines100 B was only 8-fold that of the HBL-100 cells (determined by Southern

0@ blotting) (26). Another cause of the lower association on a cell-by-cell

0 %oo@ 9 0 basis is analytical variation, for either gene copy number or immu

10 0 @c@cP6oo0@ o@ nofluorescence intensity, which is much less of a factor when the0 0 0 0 8@ 0 . . . . . . . HER-2

2@@@ ,@°o entire population is measured. Asymmetric distribution of p185

I 8@ 8I'@@@ o@ :@ protein occurring during mitosis, which occurs in exponentially grow

gc@@@@@ 8 r = 0.04 ing cellpopulations (27),might also lead to lesslinkage between gene

I ° 0 copy number and expression.

. ) It is of interest that p185'@@2 expression and the ERBB-2 copy

number:chromosome 17 copy number ratio were not closely linked

0

‘0

I100

10

40@

30

20

10

A. Average

@ Bright(Fb4)

p < 0.004

0 20 F]e@ ‘@‘000@ e@@@ c@ ‘@‘C@ © ei ‘@‘@C@

—————e@ e@ r-4e@ e-1r'@@@@@ ‘@

ERBB-2 Signals/Cell

B

p = 0.72

0 in@ in 0 in@ in 0 in— — 1'1 e@ @e e'i ‘@‘@in in

ERBB-2 Signals/Chr 17 Signals

80

60

40

20

C

p < 0.00001

ei@ ‘@in@ I-.@ 0@@@ ei@ ‘@in@ I'.@

5405

MDA-453

10

Chr 17 Signal/Cell

Fig. 6. ERBB.2 gene expression and amplification in individual MDA-453 cells. Flplotted against: A, ERBB-2 copy number; B, ERBB-2:chromosome 17 ratio; and C,chromosome i7 copy number. Cells were labeled with antibody (CB1 1) against theintracellular domain of pi8S'@'@2protein. Data from 239 cells are shown. There is a goodcorrelation between p185HER2 expression and copy number of either ERBB-2 or chromosome 17 centromere (A and C) but no correlation between p185@@2 expression andERBB.2 to chromosome (Chr) 17 ratio (B).

MDA-453

Chr 17 Signals/Cell

Fig. 7. ERBB-2copy number(A), ERBB-2:chromosome17 ratio (B), and chromosome(Chr) 17 copy number (C) in unselected and in bright (Fl > 4) MDA-453 cells.Distribution of bright cells was determined from cells plotted in Fig. 6. Cells that expressp185M5@@.2 at a high level have more copies of both ERBB-2 and chromosome 17 than does

an unselected population. There was no significant difference between unselected andbright cells in the distribution of ERBB-2:chromosome 17 ratio.

1

DISCUSSION

Overexpression of p185―@2 can occur as a result of either DNAamplification or by increased levels of RNA transcription. ConcordantERBB-2 gene amplification and p185HER2 overexpression has beenfound in both human mammary cancers and cell lines, with goodcorrelation between the level of ERBB-2 gene amplification and theaverage p185HER2 protein overexpression (5, 6, 9, 11, 14—16,23—25).However, there has been no prior analysis of this genotype-phenotypeassociation on a single-cell basis.

To investigate the cell-by-cell basis for the correlation betweenERBB-2 amplification and overexpression, several well-establishedbreast cancer cell lines having a wide range of gene amplification andprotein expression were studied, as was a primary breast tumorsample. The levels of gene amplification in the cell lines, as detectedby FISH, were equivalent to those reported previously by Southernand slot blot analysis (10). We found that the population mean valuesfor pl85H@2 protein expression were concordant with their averageERBB-2 gene copy number. However, on a cell-by-cell basis, significant heterogeneity was present both in gene copy number (detectedby FISH) and in protein expression level (by immunofluorescenceusing two different p185HER2@specific monoclonal antibodies). The

on April 8, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

20' A

0 101@ EL@@i @j.@ Bright (Fb.4)

U Average

10 12 14 J .

r = 0.49 p < 0.0002

@;lFt@ .1@@

0 ‘ . .@ r-1—

@ .@...

e't ‘@in ‘CI'-@ O@@ r@ r@ ‘@in @CI'-ERBB-2

Signals/Chr 17SignalS40Q

@C.J@ii,j@*J@@:-;

ERBB.2 EXPRESSION AND AMPLIFICATION

The positive correlation observed between pl85@°@2overexpres0 A sion and chromosome 17 copy number supports the likelihood of a

biological link between ERBB-2 gene amplification and increasedtumor cell DNA index. In fact, most primary tumors with p185HER2overexpression have DNA indices in the triploid to tetraploid range(13, 28—30).Only a few cases have been reported where @185Han@2overexpressing cells had DNA content close to diploid, and even these

r = 0.72 cases had aneuploid elements (13, 28). It is possible that the same. , factors that predispose a tumor cell to a DNA index above diploid,

120 140 together with chromosome 17 polysomy, also predispose to the gen

eration of ERBB-2 gene amplification with p185HE@@2overexpression.ERBB-2 amplification also may be associated with mitotic nondisjunctional events, destabilizing the genetic integrity of the cell and

B increasingtheopportunityforselectionof highlytumorigeniccellswith polyploid DNA content. The fact that we detected cells with12—24chromosome 17 copies, having high p185H@2 expression, isconsistent with the possibility of a mechanistic link between ERBB-2

gene amplification, polyploid DNA content, and chromosome 17r —0.39 polysomy. Although the nature of this biological link remains largely

— unknown, its importance is inferred by the established association

‘ between p185HER@2 overexpression and aggressive breast cancer be

20 havior (9, 15, 29—32).

100

10

1

.1

0

0

000o8

0

‘0

0

I

1008

000

00

0

10

1

0 10

ERBB-2 Signals/Chr 17 Signals

100

10

I

0 C8°°:8§8000

0

0@

0 2 4 6 8

Chr 17 Signals/Cell@1©@@@ ©@@@ ©in@ in©in©in@ in@ in@ in©in

,- ,- N r@@@@@ in in @O @C I'- t'-@@ 0@ 0@@@ ‘ @l

Fig. 8. ERBB-2 gene expression and amplification in single interphase cells of a touchpreparation from a primary breast tumor (case no. B372). Fl plotted against: A, ERBB.2copy number; B: ERBB-2:chromosome (Chr) 17 ratio; and C, chromosome 17 copynumber. Cells were labeled with antibody (CB11) against the intracellular domain ofpl85'@51@2protein. Data from 129 cells are shown. p185hlFl@2expression correlated wellwith copy numbers of both ERBB-2 or chromosome 17 centromere (A and C); p185HER2expression and ERBB-2:chromosome 17 ratio (B) correlated only weakly.

(r 0.00—0.25). Because chromosome 17 copy number generallyreflects the total DNA content (ploidy) of the ERBB-2 amplified cells(data not shown), ERBB-2:chromosome 17 ratio is a measure ofERBB-2 “amplification―as traditionally characterized by Southernblotting analysis. Also, we have reported previously that much of theERBB-2 amplification in SKBR3 and BT474 was not located onchromosome 17 (10), thus, reducing the utility of the ERBB-2:chromosome 17 ratio. The weak correlation we observed between ERBB2:chromosome 17 ratio and protein expression is consistent withdiscrepancies reported previously between ERBB-2 Southern analysisand p185HER2 immunohistochemistry (11). Thus, ERBB-2 amplification is dependent on the specific method of analysis used and may notalways reflect the absolute copy number of the gene.

Surprisingly, cellular expression correlated with increasing chromosome 17 copy number (r = 0.45—0.61). The brightest cells byimmunofluorescence, representing <2% of the entire cell population,had two to three times the overall population's mean chromosome 17copy number. This small subpopulation of highly p185'@@2 expressing, highly aneuploid cells had the same fraction of S-phase cells asthe remainder of the population. A similar highly overexpressing andchromosome 17 polysomic subpopulation was also observed in theprimary tumor sample, suggesting that this is not an in vitro artifact.

Chr 17 Signals/Cell

Fig. 9. ERBB-2 copy number (A), ERBB-2:chromosome (Chr) 17 ratio (B), andchromosome 17 copy number (C) in unselected and in bright (Fl > 4) interphase cells ofa touch preparation from a primary breast tumor (case no. B372). Distributions of brightcells were determined from cells plotted in Fig. 8. Cells that express p185―@'2at a highlevel have significantly more copies of ERBB-2 and chromosome 17, and a higher ratioof the two, than an unselected population.

5406

Touch Imprint # B372

0@o 0

0 6o@°@o@ %°

0 @r@0:°° 0

0 20 40 60 80 100

ERBB-2 Signals/Cells

00 s/@0o 0 0

B@@ 000@ oo@

0c@%@c@oø 000

0 @0öck@0a@0°c@l0

Touch Imprint # B372

ERBB-2 Signals/Cell

B

I I@ 1;'.1

p < 0.0002

30

c) 20

0@ 10

on April 8, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ERBB.2 EXPRESSION AND AMPLIFICATION

17. Fendly, B. M., Wignet, M., Hudziak, R. M., Lipari, M. T., Napier, M. A., and Ullrich,A. Characterization of murine monoclonal antibodies reactive to either the human

epidermal growth factor receptor or HER2/neu gene product. Cancer Res., 50:1550—1558, 1990.

18. Willard, H. F., and Waye, J. S. Structure organization and sequence of alpha satelliteDNA from human chromosome 17: evidence for evolution by unequal crossing over

and an ancestral pentamer repeat shared with the human X chromosome. Mol. Cell.Biol., 6: 3156—3165, 1986.

19. Sauter, G., Moch, H., Moore, D., Carroll, P., Kerschmann, R., Chew. K., Mihatsch.M. J., Gudat, F., and Waldman, F. W. Heterogeneity of ERBB-2 gene amplificationin bladder cancer. Cancer Res., 53: 2199—2203, 1993.

20. Johnson, G,. Davidson, R., McNamee, K., Russel, G., Goodwin, D., and Holborow,E. Fading of immunofluorescence during microscopy: a study of its phenomenon andits remedy. J. Immunol. Methods, 55: 231—242,1982.

21. BalSzs, M., Matsumara, K., Moore, D., Pinkel, D., Gray, J. W., and Waldman, F. M.Karyotypic heterogeneity and its relation to labeling index in interphase breast tumorcells. Cytometry, 20: 62—73,1995.

22. Szöliösi,J., Lockett, S. J., Balázs, M., and Waldman, F. M. Autofluorescencecorrection for fluorescence in situ hybridization. Cytometry, 20: 356—361,1995.

23. Slamon, D. J., aark, G. M., Wong, S. G., Levin, W. J., UIlrich, A., and McGuire,W. L. Human breast cancer: correlation of relapse and survival with amplification ofthe HER-2/neu oncogene. Science (Washington DC), 235: 177—181,1987.

24. Kraus, M. H., Popescu, N. C., Amsbaugh, S. C., and King, C. R. Overexpression ofEGF receptor related proto-oncogene c-erb-B2 in human mammary tumor cell lines

is by different molecular mechanisms. EMBO J., 6: 605—610, 1987.25. Corbett, I. P., Henry, J. A., Angus, B., Watchorn, C. J., Wilkinson, L, Hennessy, C.,

Gullick, W. J., Tuzi, N. L, May, F. E., Westley, B. R., and Home, C. H. W.NCL-CB1 1, a new monoclonal antibody recognizing the internal domain of thec-erbB-2 oncogene protein effective for use on formalin-fixed, paraffin-embeddedtissue. J. Pathol., 161: 15—25,1990.

26. Pasleau, F., Grooteclaes, M., and Gol-Winker, R. Expression of the c-erbB2 gene inthe B474 human mammary tumor cell line: measurement of c-erbB2 mRNA half-life.Oncogene, 8: 849—854, 1993.

27. Czemiak, B., Hera, F., Wersto, R. P., and Koss, L G. Asymmetric distribution ofoncogene products at mitosis. Proc. Nail. Acad. Sci. USA, 89: 4860—4863, 1992.

28. Bacus, S. S., Bacus, J. W., Slamon, D. J., Press, M. F. HER-2/Neu oncogeneexpression and DNA ploidy analysis in breast cancer. Arch. Pathol. Lab. Med., 114:164—169,1990.

29. Visscher, D. W., Sarkar, F. H., and Crisman, J. D. Correlation of DNA ploidy withc-erbB-2 expression in preinvasive and invasive breast tumors. Anal. Quant. Cytol.Histol., 13: 418—424, 1991.

30. Shackney, S. E., Pollice, A. A., Smith, C. A., Singh, S. G., Janocko, L E., Brown,K. A., Groft, D. W., and Hartsock, R. J. Genetic staging of breast tumors by

multiparameter flow cytometry. Cytometry, 57: 91, 1994.31. Press, M. F., Pike, M. C., Chasm, V. R., Hung, G., Udove, J. A., Markowicz, M.,

Dnayluk, J., Godolphin, W., Sliwkowski, M., Akita, R., Paterson, M. C., and Slamon,D. J. Her-2/neu expression in node-negative breast cancer: direct tissue quantitation

by computerized image analysis and association of overexpression with increased riskof recurrent disease. Cancer Res., 53: 4960—4970, 1993.

32. Bacus, S. S., Zelnik, C. R., Plowman, G., and Yarden, Y. Expression of the erbB-2family of growth factor receptors and their ligands in breast cancers. Implication fortumor biology and clinical behavior. Am. J. Clin. Pathol., 102: 513-524, 1994.

5407

REFERENCES

1. Schechter, A. L., Hung, M. C., Vaidyanathan, L., Weinberg, R. A., Yang-Feng, T. L.,Francke, U., Ullrich, A., and Coussens, L. The neu gene: an erbB-homoiogous genedistinct from and unlinked to the gene encoding the EGF receptor. Science (Washington DC), 229: 976—978, 1985.

2. Coussens, L, Yang-Feng, T. L, Liao, Y. C., Chen, E., Gray, A., McGrath, J., Seeburg@P. H., Libermann, T. A., Schlessinger, J., Francke, U., Levinson, A, and Ullrich, A.Tyrosine kinase receptor with extensive homology to EGF receptor shares chromosomallocation with neu oncogene. Science (Washington DC), 230: 1132—1139,1985.

3. Carraway, I. I. I., K. L., and Cantley, L. C. A neu acquaintance for erbB3 and erbB4:A role for receptor heterodimerization in growth signaling. Cell, 78: 5-8, 1994.

4. Slamon, D. J., Godolphin, W., Jones, L. A., Holt, J. A., Wong. S. G., Keith, D. E.,Levine, W. J., Stuart, S. 0., Udove, J., Ullrich, A., and Press, M. M. Studies of theHer-2/neu proto-oncogene in human breast and ovarian cancer. Science (WashingtonDC),244:707—712,1989.

5. Allred, D. C., O'Connell, P., and Fuqua, A. ‘N.Biomarkers in early breast neoplasia.J. Cell. Biochem., 17G: 125—i31, 1993.

6. Barnes, D. M. c-erbB-2 amplification in mammary carcinoma. J. Cell. Biochem.,17G: 132—138,1993.

7. Borg, A., Baldetorp, B., Ferno, M., Killander, D., Olsson, H., and Sigurdsson, H.erbB-2 amplification in breast cancer with a high rate of proliferation. Oncogene, 60:137—143,1991.

8. van Dienst, P. J., Baak, J. P. A., Chin, D., Theeuwes, J. W., and Bacus, S. S.Quantitation of HER-2/neu oncoprotein overexpression in invasive breast cancer byimage analysis: a study comparing fresh and paraffin-embedded material. Anal. Cell.Pathol., 3: 195—202,1991.

9. Tiwari, R. K., Borgen, P. 1., Wong, G. Y., Cordon-Cardo, C., and Osborne, M. P.HER-2/neu amplification and overexpression in primary human breast cancer is

associated with early metastasis. Anticancer Res., 12: 419—426,1992.10. Kallioniemi, 0-P., Kailioniemi, A., Kurisu, W., Thor, A., Chen, L-C., Smith, H. S.,

Waldman, F. M,. Pinkel, D., and Gray, J. W. ERBB-2 amplification in breast canceranalyzed by fluorescence in situ hybridization. Proc. Natl. Acad. Sci. USA, 89:5321—5325,1992.

11. Dawkins, H. J. S., Robbins, P. D., Sarna, M., Carrello, S., Harvey, J. M., and Sterrett,G. F. c-erbB.2 amplification and overexpression in breast cancer: evaluation andcomparison of Southern blot, slot blot. ELISA and immunohistochemistry.Pathology,25: 124—132,1993.

12. Cuny, M., Simony-Lafontaine,J., Rouanet, P., Grenier, J., Valles, H., Lavaill, it,Louason, G., Causse, A., Lequeux, N., Thierry, C., Jeanteur, P., and Theillet, C. Quantificationof ERBB2proteinexpressioninbreastcancer:threelevelsof expressiondefmedby their clinico-phatological correlations. Oncol. Res., 6: 169—176,1994.

13. Stal, 0., Sullivan, S., Sun, X-F., Wingren, S., and Nordenskjold, B. Simultaneousanalysis of c-erbB.2 expression and DNA content in breast cancer using flowcytometry. Cytometry, 16: 160—168,1994.

14. Pauletti, G., Singh, R., Press, M. F., Godolphin, W., and Slamon, D. J. HER-2/neugene amplification detected by fluorescence in situ hybridization: a comparative studywith other techniques. Proc. Am. Assoc. Cancer Res., 35: 545, 1994.

15. Symmans, W. F., Liu, J., Knowles, D. M., and Inghirami, G. Breast cancer heterogeneity: evaluation of clonality in primary and metastatic lesions. Hum. Pathol. 26:210—216, 1995.

16. Press, M. F., Hung, G., Godolphin, W., and Slamon, D. J. Sensitivity of HER-2/neuantibodies in archival tissue samples: potential source of error in immunohistochemical studies of oncogene expression. Cancer Res., 54: 2771—2777,1994.

on April 8, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1995;55:5400-5407. Cancer Res   János Szöllösi, Margit Balázs, Burt G. Feuerstein, et al.   CancerOverexpression, and Intratumor Heterogeneity in Human Breast

HER-2) Gene Copy Number, p185HER2/neu (ERBB-2

  Updated version

  http://cancerres.aacrjournals.org/content/55/22/5400

Access the most recent version of this article 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/55/22/5400To request permission to re-use all or part of this article, use this link

on April 8, 2020. © 1995 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from