rna interference-based functional dissection of the 17q12 amplicon in breast cancer reveals...

RNA Interference-Based Functional Dissectionof the 17q12 Amplicon in Breast Cancer RevealsContribution of Coamplified Genes

Jessica Kao and Jonathan R. Pollack*

Departmentof Pathology,Stanford University,Stanford,California

DNA amplification is a frequent occurrence in cancer genomes. While tumor amplicons may harbor known oncogenes ‘‘driv-

ing’’ amplification, amplicons rarely comprise only single genes. The potential functional contribution of coamplified genes

remains largely unexplored. In breast cancer, 20–30% of tumors exhibit amplification within chromosome band 17q12, con-

taining the ERBB2 oncogene. Analysis of array-based comparative genomic hybridization and expression profiling data indicate

that the minimum region of recurrent amplification (i.e., the amplicon ‘‘core’’) at 17q12 includes two other genes, GRB7 and

STARD3, which exhibit elevated expression when amplified. Western blot analysis confirms overexpression of each at the

protein level in breast cancer cell lines SKBR3 and BT474 harboring amplification. In these cell lines (but not in control MCF7

breast cancer cells lacking 17q12 amplification), targeted knockdown of ERBB2 expression using RNA interference (RNAi)

methods results in decreased cell proliferation, decreased cell-cycle progression, and increased apoptosis. Notably, targeted

knockdown of either GRB7 or STARD3 also leads to decreased cell proliferation and cell-cycle progression, albeit to a lesser

extent compared with ERBB2 knockdown. We conclude that the amplification and resultant overexpression of genes coam-

plified with ERBB2 at 17q12 can contribute to proliferation levels of breast cancer cells. Our findings validate the utility of

RNAi in the functional interrogation of tumor amplicons, and provide evidence for a contribution of coamplified genes to tu-

mor phenotypes. VVC 2006 Wiley-Liss, Inc.

INTRODUCTION

Breast cancer is the second leading cause of can-

cer death in women in the United States, affecting

�1 in 8 women over the course of their lifetime

(Ries et al., 2005). Clinicopathological parameters

such as tumor size, grade, stage, estrogen receptor

(ER) positivity, and lymph node involvement have

been used to predict tumor recurrence and guide

the selection of therapies (Subramaniam and Isaacs,

2005). More recently, tumors exhibiting amplifica-

tion and overexpression of ERBB2 (HER2/NEU), a

member of the epidermal growth factor receptor

(EGFR) family of receptor tyrosine kinases, have

been shown responsive to directed therapy targeting

ERBB2 (Slamon et al., 2001). Defining additional

pathogenetic lesions may provide new targets for

molecularly directed therapies.

DNA amplification, leading to the deregulated

expression of oncogenes, is one such class of patho-

genetic lesion occurring frequently in breast cancer

(Knuutila et al., 1998). Many localized amplicons

harbor known oncogenes ‘‘driving’’ DNA amplifica-

tion, including FGFR1 (8p11), MYC (8q24), CCND1(11q13), MDM2 (12q13), ERBB2 (17q12), and ZNF217(20q13) (Courjal et al., 1997; Nonet et al., 2001).

Other localized amplicons contain as yet uncharac-

terized driver oncogenes. Regardless, a common

feature of localized amplicons is that they seldom

contain just a single gene. This observation has

led to speculation that the retained amplification

of coamplified genes (i.e., those neighboring the

presumptive driver oncogene) might contribute to

tumor phenotypes, thereby providing selective

growth advantage to tumor cells (Schuuring et al.,

1992). Indeed, such speculation is fueled by obser-

vations that nearly half of all amplified genes ex-

hibit elevated expression (Hyman et al., 2002; Pol-

lack et al., 2002).

However, the true functional contribution of co-

amplified genes has remained largely unknown, in

part because of the laboriousness of traditional ap-

proaches of study, requiring individually cloning,

transfecting, and characterizing candidate genes in

heterologous cell culture systems (e.g., NIH/3T3

*Correspondence to: Jonathan R. Pollack, M.D., Ph.D., Departmentof Pathology, Stanford University School of Medicine, 269 CampusDrive, CCSR-3245A, Stanford, CA 94305-5176, USA.E-mail: [email protected]

Received 22 December 2005; Accepted 18 April 2006

DOI 10.1002/gcc.20339

Published online 17 May 2006 inWiley InterScience (www.interscience.wiley.com).

Supported by: NIH; Grant number: CA97139.

Abbreviations: array CGH, array-based comparative genomichybridization; RNAi, RNA interference; siRNA, small interferingRNA.

VVC 2006 Wiley-Liss, Inc.

GENES, CHROMOSOMES & CANCER 45:761–769 (2006)

mouse fibroblasts). RNA interference (RNAi) me-

thods, based on the transfection of synthetic siRNAs

(small interfering RNAs) effecting the sequence-

specific post-transcriptional silencing of targeted

genes (Tuschl and Borkhardt, 2002), provide a

potentially more rapid alternative for screening

the function of candidate genes within the context

of cancer cell lines harboring DNA amplification.

Here, we apply RNAi methods to explore the

contribution of coamplified genes to breast tumor

phenotypes, focusing on the 17q12 amplicon harbor-

ing ERBB2 in breast cancer. Our findings validate

an RNAi-based strategy for functionally dissecting

DNA amplicons, and provide evidence that coam-

plified genes can indeed contribute to tumor pheno-

types.

MATERIALS AND METHODS

Microarray Data Analysis

cDNA microarray-based comparative genomic

hybridization (CGH) and expression-profiling data

for breast cancer cell lines and primary breast

tumors were previously published (Pollack et al.,

2002). To define a minimum core amplicon at

17q12-21, well-measured (intensity/background �1.4)

tumor/normal array CGH fluorescence ratios for

cDNA probes were ordered by chromosome map

position, assigned using the NCBI genome assem-

bly accessed through the UCSC genome browser

(NCBI Build 35). For genes represented by multi-

ple cDNA probes, the average fluorescence ratio

was used. For expression profiling data, ‘‘mean-cen-

tered’’ ratios (i.e., reported for each gene relative to

its mean expression across samples) were used.

Cell Culture

Breast cancer cell lines SKBR3, BT474, and

MCF7 were obtained from the American Type Cul-

ture Collection (Manassas, Virginia). Cell lines were

maintained at 378C in complete media of RPMI

1640 (Invitrogen), 10% fetal bovine serum (FBS),

50 U/ml penicillin, and 50 U/ml streptomycin. ‘‘Low

serum’’ experiments were performed by plating cells

into the above media with 2% FBS.

Small Interfering RNAs

Synthetic 21-nt siRNAs were designed and syn-

thesized to target ERBB2, GRB7, and STARD3

(Qiagen, Valencia, California). A siRNA targeting

the gene luciferase (LUC), not encoded in the

human genome, was used as a control (Qiagen).

Before use, siRNAs were reconstituted in annealing

buffer per manufacturer’s instructions at 20 lM,

heated to 908C for 1 min, incubated at 378C for

1 hr, and then stored at �808C. Sequences of siR-NAs are listed in Table 1.

Cell Transfection

For cell transfection, 350,000 cells were seeded

per 6-well plate well, and transfected using TransIT-

TKO reagent (Mirus Bio, Madison, Wisconsin) ex-

actly according to the manufacturer’s protocol. Cells

were transfected with a final concentration of 30 nM

siRNA for 18 hr, subsequently replaced with com-

plete growth media (or media with 2% FBS for low-

serum experiments).

Quantitative RT-PCR

Post-transfection (24 hr), total RNA was prepared

by the TRIzol method according to the manufac-

turer’s instructions (Invitrogen, Carlsbad, CA) and

quantified by UV spectrophotometry. Q-RT-PCR

was performed using the one-step QuantiTect

SYBR Green RT-PCR Kit (Qiagen) and PCR prod-

ucts quantified using an ABI PRISM 7700 Sequence

Detection System (Applied Biosystems, Foster City,

CA). Briefly, PCR primers (Table 2) were designed

using Primer3 software (Rozen and Skaletsky, 2000),

with one primer of each pair spanning an exon–

exon junction to prevent amplification from contam-

inating genomic DNA. Q-RT-PCR was carried out

in a final volume of 25 ll, containing 200 ng RNA,

13 SYBR green mix, 0.25 U RT, and 0.1 lM each

primer, under the following conditions: 508C 30 min,

958C 15 min, followed by 40 cycles each of 958C

TABLE 1. siRNA Sequences

Gene targeted Target sequence

LUC (control) 50 CGTACGCGGAATACTTCGA 30

GRB7 50 CCCAACAAGCTTCGAAATGGA 30

ERBB2 50 AAGCCTCACAGAGATCTTGAA 30

ERBB2 (second siRNA) 50 AATGGCTCAGTGACCTGTTTT 30

STARD3 50 AACCCCCGTGTTTGCACCTTT 30

STARD3 (second siRNA) 50 AATCCTTTGCAGGGTCTGACA 30

TABLE 2. Quantitative RT-PCR Primers

Gene Primer sequence

GAPDH-F 50 CAATGACCCCTTCATTGACC 30

GAPDH-R 50 GATCTCGCTCCTGCAAGATG 30

GRB7-F 50 CTTGTTGGGCTTGACACAGP 30

GRB7-R 50 GCCTGGAGGAAGAAGACAAA 30

ERBB2-F 50 CTGTGCCCGAGTGTGCTA 30

ERBB2-R 50 GTCCCCATCAAAGCTCTCC 30

STARD3-F 50 AAGGTCATCCTCTCTGAGCTG 30

STARD3-R 50 GCARGATACCATCGCTCCTC 30

Genes, Chromosomes & Cancer DOI 10.1002/gcc

762 KAO AND POLLACK

15 sec, 558C 30 sec, and 728C 30 sec. GAPDH

mRNAwas quantified as an internal standard. Melt-

ing curve analysis was performed from 60–908C to

maximize PCR product specificity; data for GRB7and ERBB2 were acquired at 76.68C and for STARD3at 808C. All Q-RT-PCR reactions were performed in

triplicate, and the comparative CT method (Livak

and Schmittgen, 2001) was used to calculate relative

mRNA levels normalized to GAPDH.

Western Blot

Post-transfection (72 hr), cells were lysed in 13RIPA Lysis buffer (Upstate/Chemicon, San Francisco,

CA) to which had been added 13 complete protease

inhibitor (Roche, Indianapolis, IN), 0.1 mM sodium

orthovanadate, 1 mM sodium fluoride, and 1 mM

PMSF, and protein was quantified using the BCA

assay (Pierce, Rockford, IL). For Western blot, 20 lgprotein lysate was electrophoresed on a 4–15% Tris/

glycine polyacrylamide gradient gel (Bio-Rad, Her-

cules, CA) and transferred to PVDF membrane

(Bio-Rad). After blocking in TBS-T buffer (20 mM

Tris-HCl (pH 7.4), 0.15 M NaCl, and 0.1% Tween

20) with 5% dry milk for 30 min, blots were incu-

bated sequentially with primary antibody for 90 min

and HRP-conjugated secondary antibody for 45 min,

each at room temperature in TBS-T buffer. Primary

antibodies were used as follows: anti-HER2/NEU

mouse monoclonal antibody (1:2,500, NeoMarkers,

Foster City, CA), anti-GRB7 rabbit polyclonal anti-

body (1:1,000, Santa Cruz Biotechnologies, Santa

Cruz, CA), anti-STARD3 mouse monoclonal antibody

(1:1,000, kind gift from Dr. M.C. Rio (Moog-Lutz

et al., 1997)), and anti-a-tubulin mouse monoclo-

nal antibody (1:1,000 for loading control, Santa

Cruz Biotechnology). Secondary antibodies used

were as follows: HRP-conjugated anti-mouse IgG

(1:20,000, Pierce), and HRP-conjugated anti-rabbit

IgG (1:20,000, Pierce). Detection was carried out

using the ECL kit (Amersham Biosciences, Piscat-

away, NJ), and quantification by densitometry

using Image J software (Abramoff et al., 2004).

Cell Proliferation Assay

Post-transfection (24, 48, and 72 hr), cell prolif-

eration was quantified by colorimetry based on the

metabolic cleavage of the tetrazolium salt WST-1

in viable cells, according to the manufacturer’s

protocol (Roche). Briefly, WST-1 reagent was added

at 1/10th the culture volume and incubated at 378Cfor 30 min. Absorbance was then measured at 450 nm

with reference to 650 nm using a Spectra Max 190

plate reader (Molecular Devices, Sunnyvale, CA).

Transfections were performed in triplicate and av-

erage (61 SD) OD reported.

Cell-Cycle Analysis

Post-transfection (72 hr), cell-cycle distribution

analysis was performed by flow cytometry using

the BrdU-FITC Flow kit (BD Biosciences, San Jose,

CA) per the manufacturer’s instructions. Briefly, cells

were incubated with 10 uM BrdU at 378C for 4 hr,

then cells were fixed and permeabilized with Cyto-

fix/Cytoperm buffer (BD Biosciences). Cellular DNA

was treated with DNase at 378 for 1 hr to expose

incorporated BrdU, then cells were stained with

anti-BrdU FITC antibody (to quantify incorporated

BrdU) and 7-aminoactinomycin D (7-AAD; to quan-

tify total DNA content). Ten thousand events were

scored by FACSCalibur (BD Biosciences) and ana-

lyzed using CellQuest software (BD Biosciences).

For SKBR3 studies, transfections were performed in

duplicate and average (61 SD) cell-cycle fractions

reported.

Apoptosis Assay

Post transfection (72 hr), apoptosis was quantified

by fluorescence microscopy of nuclear morphology,

after staining with ethidium bromide (EB) and acri-

dine orange (AO) (Cohen, 1993). Briefly, floating

cells and trypsinized adherent cells were pooled and

resuspended in 25 ll complete growth media, and

1 ll of a mixture of 100 lg/ml each EB and AO

added. Cells were examined using an Olympus BX

61 microscope fitted with a triple-pass fluorescence

filter using the 403 objective. Apoptotic cells were

scored as those with highly condensed or fragmented

chromatin. Transfections were performed in tripli-

cate, 200 cells were counted per transfection, and

average (61 SD) percent apoptosis reported.

RESULTS

Defining the Amplicon Core at 17q12

To define the minimum amplicon core at 17q12,

we first inspected our previous cDNA microarray-

based CGH data (Pollack et al., 2002) for gene

amplification at this locus. Six of 10 breast cancer

cell lines (BT474, MDA361, SKBR3, UACC-812,

UACC-893, ZR75-30) and 8 of 44 primary breast

tumors exhibited localized high-level DNA amplifi-

cation at 17q12. The minimum region of recurrent

amplification included three arrayed genes, ERBB2,GRB7, and STARD3, and each of the three exhibited

elevated mRNA levels when amplified (Fig. 1).

ERBB2 (HER2/NEU) is a bona fide pathogenic

oncogene in breast cancer (Reese and Slamon, 1997),


763FUNCTIONAL DISSECTION OF 17q12 AMPLICON

and a target of directed therapies such as the anti-

body-based therapeutic trastuzamab (Herceptin)

(Slamon et al., 2001). Nonetheless, amplification of

GRB7, a SH2-domain adapter protein that inter-

acts with ERBB2 (Stein et al., 1994), and STARD3(MLN64; (Moog-Lutz et al., 1997)), which stimu-

lates steroidogenesis (Watari et al., 1997), might

contribute to cancer phenotypes. Before exploring

this hypothesis, we first validated that DNA am-

plification with elevated mRNA levels was associ-

ated with increased protein levels for each of the

genes. For our analysis, we selected to study SKBR3

and BT474 breast cancer cell lines harboring ampli-

fication at 17q12-21, and MCF7 as a control breast

cancer cell line without amplification at this locus.

Western blot demonstrated markedly elevated

expression of ERBB2 (185 kDa), GRB7 (54 kDa),

and STARD3 (50 kDa) in SKBR3 and BT474 lines

compared with MCF7 (Fig. 2A).

Validating siRNA-Mediated Knockdown of ERBB2,

GRB7, and STARD3

Our strategy to assess the contribution of amplified

genes to cancer phenotypes was to separately trans-

fect synthetic siRNAs to knock down the expression

of each of the amplified genes via RNAi, and then

to monitor altered cancer-relevant phenotypes in

Figure 1. Defining the amplicon core at 17q12. Shown are DNAcopy number ratios quantified by array CGH for six breast cancer celllines (above) and eight primary breast tumors (below) harboring17q12-21 amplification. Genes are ordered by their position along thechromosome segment, and height of bars indicates the level of amplifi-cation (log2 ratio scales shown). The tumor amplicon has distinctboundaries in each specimen; the minimum region of recurrent amplifi-cation (i.e., the amplicon ‘‘core’’), highlighted in gray, includes STARD3,ERBB2, and GRB7. Parallel DNA microarray-based expression-profilingidentifies the subset of amplified genes exhibiting elevated expressionwhen amplified (gene names in bold text); gray text indicates poorlymeasured mRNA levels.

Figure 2. Validation of protein overexpression and RNAi knock-down. A: Western blot analysis of ERBB2, GRB7, and STARD3 indi-cates proteins are overexpressed in cell lines harboring 17q12 amplifi-cation (SKBR3 and BT474) compared to nonamplified line MCF7. B:Western blot analysis confirms siRNA-mediated knockdown of ERBB2,GRB7, and STARD3 in each of the three cell lines assayed. a-Tubulinserves as a loading control.


764 KAO AND POLLACK

cell culture. We first sought to validate effective knock-

down of mRNA levels by Q-RT-PCR (not shown),

and of protein levels by Western blot (Fig. 1B), in

comparison to transfection of a control siRNA tar-

geting the irrelevant gene LUC. Two of three differ-

ent siRNAs targeting ERBB2 were effectual in

knocking down both mRNA and protein levels, with

the most effective resulting in residual ERBB2 pro-

tein levels of 29%, 6%, and 3% in SKBR3, BT474,

and MCF7, respectively. For GRB7, only one of six

different siRNAs assayed significantly reduced mRNA

and protein levels, with 28%, 8%, and 36% residual

GRB7 protein levels in SKBR3, BT474, and MCF7.

For STARD3, two of three siRNAs effectively de-

creased mRNA and protein levels, with the most

efficient leaving 3% and 12% residual STARD3

protein levels in SKBR3 and BT474; protein levels

in MCF7 were too low to accurately quantify re-

duction. For subsequent experiments, unless other-

wise specified we used the most effective siRNA

(Table 1) to target each amplified gene.

Characterizing Altered Cancer-Relevant

Phenotypes in Cell Culture

If coamplification of GRB7 and STARD3 contrib-

ute to tumorigenesis, then knocking down their

expression by RNAi in cells harboring amplifica-

tion might revert cancer-relevant phenotypes. We

first characterized the effect of siRNA-mediated

knockdown on cell proliferation, quantified on

days 1, 3, and 5 following siRNA transfection using

the WST-1 assay. Targeted knockdown of ERBB2

led to significant inhibition of cell proliferation in

both SKBR3 and BT474 cells (P < 0.001; Student’s

t test) (Fig. 3A), comparable to reported results

using anti-ERBB2 antibody (Lewis et al., 1993) or

more recently RNAi (Choudhury et al., 2004; Faltus

et al., 2004). In contrast, no significant decrease in

proliferation was observed in MCF7 cells that do

not carry amplification at 17q12. Although targeted

knockdown of GRB7 and STARD3 had no effect

on cell proliferation in control MCF7 cells, knock-

down of each resulted in decreased cell prolifera-

tion in both SKBR3 and BT474 cells (Fig. 3A); for

the former, the effect was more pronounced and

reproducible when cells were grown in low serum

(2% FBS), so subsequent studies of SKBR3 were

performed in low serum. Although proliferation

was not reduced to levels seen with ERBB2 knock-

down, reductions were nonetheless significant in

both SKBR3 and BT474 cell lines for both GRB7

(P < 0.05 and P < 0.01, respectively) and STARD3

(P < 0.01) knockdown.

MCF7 cells do not carry 17q12 amplification,

and therefore, presumably do not rely on increased

levels of ERBB2, GRB7, and STARD3 to affect

tumorigenesis. Thus, RNAi-mediated knockdown

of these genes in MCF7 cells did not reduce cell

proliferation and therefore supports the specificity

of the siRNA effect. Nonetheless, to further rule

out ‘‘off-target’’ effects (Jackson et al., 2003), we

knocked down ERBB2 and STARD3 expression in

BT474 cells using for each a second siRNA (Table 1)

targeting distinct sequences within the genes, and

found comparable reductions in cell proliferation

(not shown). A similar assessment could not be

performed for GRB7, since a second effective siRNA

could not be identified among six surveyed.

The reduced cell proliferation following knock-

down of ERBB2, GRB7, or STARD3 might result

from decreased cell-cycle progression, increased

cell death (apoptosis), or both. To distinguish among

these possibilities, we quantified DNA synthesis by

BrdU incorporation, and apoptosis by nuclear mor-

phology using an EB and AO staining assay, 72 hr

post-transfection. Knockdown of ERBB2 in SKBR3

and BT474 cells resulted in reduced S-phase frac-

tion (9-fold and 10-fold reduction, respectively) with

a G0/G1 accumulation (Fig. 3C). Likewise, targeting

of GRB7 and STARD3 led to reduced S-phase frac-

tion in SKBR3 cells (2.1-fold and 1.3-fold, respec-

tively) and BT474 cells (1.7-fold and 1.3-fold,

respectively), suggesting that reduced cell prolifera-

tion was at least in part attributable to decreased

cell-cycle progression. Knockdown of ERBB2 in

SKBR3 and BT474 cells also led to increased apo-

ptosis (P < 0.01; Fig. 3B). However, no such effect

on apoptosis was observed accompanying knockdown

of either GRB7 or STARD3 in SKBR3 or BT474

cells (Fig. 3B); nor for SKBR3 did we observe dis-

proportionally (i.e., compared to targeting LUC)

increased apoptosis when experiments were carried

out in the presence of 10 nM (�LD25) of the cyto-

toxic drug cisplatin (not shown).

Assessing Additive Effects of Targeting Multiple

Amplified Targets

Since targeting GRB7 and STARD3 each led to

reduced cell proliferation, albeit to a lesser extent

compared to targeting ERBB2, we wondered whether

targeting combinations of GRB7, STARD3, and

ERBB2 might provide additive effects, with impli-

cations for directed therapies. To test this idea, we

cotransfected BT474 cells with siRNA pools tar-

geting GRB7/STARD3, ERBB2/GRB7, or ERBB2/

STARD3, and then assayed cell proliferation on

days 1, 3, and 5 post-transfection. Although targeting



Figure 3. Effects of siRNA-mediated knockdown on cancer relevantphenotypes. A: Cell proliferation quantified by WST-1 assay on days 1, 3,and 5 post-transfection. Cell lines transfected and genes targeted by siR-NAs are indicated. Transfections were performed in triplicate and mean(61 SD) ODs plotted. Note, the decreased growth observed in control(LUC)-transfected SKBR3 cells is attributable to a �2-fold decreasedS-phase fraction and a �2-fold increased apoptotic fraction consequent

to reduced serum (data not shown). B: Apoptosis quantified by fluores-cence microscopy of nuclear morphology on day 3 post-transfection.Transfections were performed in triplicate and mean (61 SD) percent ap-optosis plotted. C: Cell-cycle analysis quantified on day 3 post-transfec-tion by flow cytometry following BrdU incorporation. Fraction of cells inG1, S, and G2/M is indicated. Insets show representative FACS plots forexperiments targeting LUC (left) and ERBB2 (right) in SKBR3 cells.


766 KAO AND POLLACK

GRB7 and STARD3 together provided an additive

effect (P< 0.05), targeting either together with ERBB2

provided only minimal (and not reproducible) increased

reduction in cell proliferation compared to targeting

ERBB2 alone (Fig. 4).

DISCUSSION

Using RNAi methods to knock down the ex-

pression of amplified genes, we have shown here

that in some cell contexts genes coamplified to-

gether with ERBB2 at 17q12 contribute to cancer-

relevant phenotypes. Specifically, we have found

that knockdown of GRB7 and STARD3 expression

leads to reduced cell proliferation, which appears to

result at least in part from decreased cell-cycle pro-

gression. Therefore, we infer that overexpression of

these genes due to amplification contributes to the

elevated proliferation rates that characterize SKBR3

and BT474 cells. Indeed, preliminary experiments

using other breast cancer cell lines with 17q12

amplification (e.g., ZR75-30) demonstrate the same

to hold true (data not shown). Although previously

speculated (Stein et al., 1994; Akiyama et al., 1997;

Moog-Lutz et al., 1997; ), this is to our knowledge

the first experimental support for a functional contribu-

tion of coamplified genes within the 17q12 amplicon.

GRB7 is an SH2-domain containing adapter pro-

tein that has been shown to interact with EGFR

and ERBB2 (Han et al., 2001). Amplification and

overexpression of GRB7 may enhance signaling

through EGFR-family receptor pathways in some

contexts. STARD3 functions in cholesterol traffick-

ing (Strauss et al., 2003), and has been shown to

stimulate steroidogenesis (Watari et al., 1997) where

breast tumor growth is often responsive to steroids

(Flototto et al., 2001). While plausible then, the

mechanisms through which GRB7 and STARD3

overexpression promote cell proliferation remain to

be elucidated. The differential effect of GRB7 and

STARD3 knockdown observed in BT474 compared

with that in SKBR3 cells (where the effect is more

reproducibly observed in suboptimal growth condi-

tions using low serum) also remains to be investi-

gated, but might be attributable to differences in the

residual levels of expressed protein, or to differences

in the genetic/epigenetic alterations of the cell lines.

Our findings indicate that combined targeting of

ERBB2/GRB7 or ERBB2/STARD3 provides little

additive effect on reducing proliferation levels be-

yond targeting ERBB2 alone, at least in the context

of BT474 cells. This finding underscores the promi-

nent role of the ERBB2 oncogene within this ampli-

con, and from a practical standpoint suggests that

therapies directed against GRB7 and STARD3

would likely provide little if any additional benefit

over targeted ERBB2 therapies alone for patients

with tumors carrying 17q12 amplification. Nonethe-

less, there may be utility for such therapies in ERBB2

therapy (e.g., trastuzamab)-resistant cases.

Our study has addressed the role of two coam-

plified genes (GRB7 and STARD3) that we had

defined by our array CGH data to reside within the

core amplicon at 17q12, and for which antibodies

were available to confirm protein knockdown by

RNAi. The intent was to test the possibility that

coamplified genes might contribute to tumor pheno-

types, rather than to carry out a comprehensive

functional analysis of the 17q12 amplicon. Recent

data from our own laboratory (unpublished) and

others (Kauraniemi et al., 2003) indicate that the

core amplicon at 17q12 includes, besides GRB7,ERBB2 and STARD3, four additional annotated

genes (TCAP, PNMT, PERLD1, C17orf37), of whichthree (PNMT, PERLD1, C17orf37) when amplified

exhibit elevated mRNA levels by RT-PCR (Kaura-

niemi et al., 2003) or DNA microarray (our unpub-

lished data). Additional experiments are required to

assess the functional contribution, if any, of these

other coamplified genes.

It is notable that we succeeded in knocking

down all three targeted genes despite high levels

of DNA amplification (�10- to 15-fold) and over-

Figure 4. Effect of targeting multiple amplified targets. Cell prolifer-ation quantified by WST-1 assay on days 1, 3, and 5 post-transfection.Cell lines transfected and gene combinations targeted by siRNAs are in-dicated. Transfections were performed in triplicate and mean (61 SD)ODs plotted.



expression. Our findings support the general utility

of a strategy using RNAi to rapidly screen candi-

date oncogene(s) within localized tumor amplicons.

Notably, using RNAi to knock down the expression

of oncogene candidates in cultured cells harboring

amplification likely provides a more physiologically

appropriate context for studying gene function, in

comparison to the more traditional approach of over-

expressing candidate oncogenes in heterologous

nontumorigenic cells such as murine fibroblasts.

For RNAi studies, while assayable phenotypes in

cell culture (e.g., proliferation, apoptosis, migra-

tion, invasion, anchorage-independent growth) are

somewhat limited, in principle a similar approach

could be used to assess altered tumor growth in

vivo using murine xenografts or transgenic models.

While the transient transfection of synthetic siR-

NAs, as performed here, has provided a quick

screen for certain cell culture phenotypes, stable

and even inducible knockdown (Amarzguioui et al.,

2005) would be more suitable for lengthier assays

and in vivo studies.

Breast tumors harbor frequent high-level amplifi-

cation at other loci with known oncogenes, includ-

ing 8p12 (FGFR1), 8q24 (MYC), 11q13 (CCND1),12q13 (MDM2), and 20q13 (ZNF217) (Courjal et al.,1997; Nonet et al., 2001), as well as many amplified

sites with as yet no known oncogene (Hyman et al.,

2002; Pollack et al., 2002). Such amplicons rarely

contain single genes. This observation might

reflect physical properties of genome hotspots for

amplification, for example preferred sites of break-

age-fusion-bridge (BFB) cycles (Gisselsson, 2003).

Alternatively, our findings for the 17q12 amplicon

support the competing hypothesis that many or

most such tumor amplicons comprise several genes

because the coamplification of multiple genes pro-

vides selective advantage through contributions to

tumorigenic phenotypes. Additional studies are re-

quired to assess the generalizability of our findings

to other tumor amplicons.

ACKNOWLEDGMENTS

We wish to thank Dr. Marie-Christine Rio (IGMBC,

France) for providing anti-STARD3 antibody. We

also thank the members of the Pollack lab for help-

ful discussion.

REFERENCES

Abramoff MD, Magelhaes PJ, Ram SJ. 2004. Image processing withImageJ. Biophotonics Int 11:36–42.

Akiyama N, Sasaki H, Ishizuka T, Kishi T, Sakamoto H, Onda M,Hirai H, Yazaki Y, Sugimura T, Terada M. 1997. Isolation of a

candidate gene, CAB1, for cholesterol transport to mitochondriafrom the c-ERBB-2 amplicon by a modified cDNA selectionmethod. Cancer Res 57:3548–3553.

Amarzguioui M, Rossi JJ, Kim D. 2005. Approaches for chemicallysynthesized siRNA and vector-mediated RNAi. FEBS Lett 579:5974–5981.

Choudhury A, Charo J, Parapuram SK, Hunt RC, Hunt DM,Seliger B, Kiessling R. 2004. Small interfering RNA (siRNA)inhibits the expression of the Her2/neu gene, upregulates HLAclass I and induces apoptosis of Her2/neu positive tumor celllines. Int J Cancer 108:71–77.

Cohen JJ. 1993. Apoptosis. Immunol Today 14:126–130.Courjal F, Cuny M, Simony-Lafontaine J, Louason G, Speiser P,

Zeillinger R, Rodriguez C, Theillet C. 1997. Mapping of DNAamplifications at 15 chromosomal localizations in 1875 breasttumors: Definition of phenotypic groups. Cancer Res 57:4360–4367.

Faltus T, Yuan J, Zimmer B, Kramer A, Loibl S, Kaufmann M,Strebhardt K. 2004. Silencing of the HER2/neu gene by siRNAinhibits proliferation and induces apoptosis in HER2/neu-overex-pressing breast cancer cells. Neoplasia 6:786–795.

Flototto T, Djahansouzi S, Glaser M, Hanstein B, Niederacher D,Brumm C, Beckmann MW. 2001. Hormones and hormone antag-onists: Mechanisms of action in carcinogenesis of endometrialand breast cancer. Horm Metab Res 33:451–457.

Gisselsson D. 2003. Chromosome instability in cancer: How, when,and why? Adv Cancer Res 87:1–29.

Han DC, Shen TL, Guan JL. 2001. The Grb7 family proteins:Structure, interactions with other signaling molecules and poten-tial cellular functions. Oncogene 20:6315–6321.

Hyman E, Kauraniemi P, Hautaniemi S, Wolf M, Mousses S,Rozenblum E, Ringner M, Sauter G, Monni O, Elkahloun A,Kallioniemi OP, Kallioniemi A. 2002. Impact of DNA amplifica-tion on gene expression patterns in breast cancer. Cancer Res 62:6240–6245.

Jackson AL, Bartz SR, Schelter J, Kobayashi SV, Burchard J,Mao M, Li B, Cavet G, Linsley PS. 2003. Expression profilingreveals off-target gene regulation by RNAi. Nat Biotechnol 21:635–637.

Kauraniemi P, Kuukasjarvi T, Sauter G, Kallioniemi A. 2003. Ampli-fication of a 280-kilobase core region at the ERBB2 locus leadsto activation of two hypothetical proteins in breast cancer. AmJ Pathol 163:1979–1984.

Knuutila S, Bjorkqvist AM, Autio K, Tarkkanen M, Wolf M, Monni O,Szymanska J, Larramendy ML, Tapper J, Pere H, El-Rifai W,Hemmer S, Wasenius VM, Vidgren V, Zhu Y. 1998. DNA copynumber amplifications in human neoplasms: Review of compara-tive genomic hybridization studies. Am J Pathol 152:1107–1123.

Lewis GD, Figari I, Fendly B, Wong WL, Carter P, Gorman C,Shepard HM. 1993. Differential responses of human tumor celllines to anti-p185HER2 monoclonal antibodies. Cancer ImmunolImmunother 37:255–263.

Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expressiondata using real-time quantitative PCR and the 2(-Delta Delta C(T))Method. Methods 25:402–408.

Moog-Lutz C, Tomasetto C, Regnier CH, Wendling C, Lutz Y,Muller D, Chenard MP, Basset P, Rio MC. 1997. MLN64 exhibitshomology with the steroidogenic acute regulatory protein (STAR)and is over-expressed in human breast carcinomas. Int J Cancer71:183–191.

Nonet GH, Stampfer MR, Chin K, Gray JW, Collins CC, Yaswen P.2001. The ZNF217 gene amplified in breast cancers promotesimmortalization of human mammary epithelial cells. Cancer Res61:1250–1254.

Pollack JR, Sorlie T, Perou CM, Rees CA, Jeffrey SS, Lonning PE,Tibshirani R, Botstein D, Borresen-Dale AL, Brown PO. 2002.Microarray analysis reveals a major direct role of DNA copy num-ber alteration in the transcriptional program of human breasttumors. Proc Natl Acad Sci USA 99:12963–12968.

Reese DM, Slamon DJ. 1997. HER-2/neu signal transduction inhuman breast and ovarian cancer. Stem Cells 15:1–8.

Ries LAG, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L,Mariotto A, Feuer EJ, Edwards BK. 2005. SEER Cancer StatisticsReview, 1975–2002. Bethesda, MD: National Cancer Institute.

Rozen S, Skaletsky H. 2000. Primer3 on the WWW for generalusers and for biologist programmers. Methods Mol Biol 132:365–386.

Schuuring E, Verhoeven E, Mooi WJ, Michalides RJ. 1992. Identifi-cation and cloning of two overexpressed genes, U21B31/PRAD1and EMS1, within the amplified chromosome 11q13 region inhuman carcinomas. Oncogene 7:355–361.


768 KAO AND POLLACK

Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V,Bajamonde A, Fleming T, Eiermann W, Wolter J, Pegram M,Baselga J, Norton L. 2001. Use of chemotherapy plus a monoclonalantibody against HER2 for metastatic breast cancer that overex-presses HER2. N Engl J Med 344:783–792.

Stein D, Wu J, Fuqua SA, Roonprapunt C, Yajnik V, D’Eustachio P,Moskow JJ, Buchberg AM, Osborne CK, Margolis B. 1994. TheSH2 domain protein GRB-7 is co-amplified, overexpressed andin a tight complex with HER2 in breast cancer. EMBO J 13:1331–1340.

Strauss JF, III, Kishida T, Christenson LK, Fujimoto T, Hiroi H.2003. START domain proteins and the intracellular trafficking of

cholesterol in steroidogenic cells. Mol Cell Endocrinol 202:59–65.

Subramaniam DS, Isaacs C. 2005. Utilizing prognostic and predictivefactors in breast cancer. Curr Treat Options Oncol 6:147–159.

Tuschl T, Borkhardt A. 2002. Small interfering RNAs: A revolution-ary tool for the analysis of gene function and gene therapy. MolInterv 2:158–167.

Watari H, Arakane F, Moog-Lutz C, Kallen CB, Tomasetto C,Gerton GL, Rio MC, Baker ME, Strauss JF, III. 1997. MLN64contains a domain with homology to the steroidogenic acute reg-ulatory protein (StAR) that stimulates steroidogenesis. Proc NatlAcad Sci USA 94:8462–8467.



rna interference-based functional dissection of the 17q12 amplicon in breast cancer reveals...

Documents