Overexpression of LMO4 induces mammary hyperplasia, promotes cell invasion, and is a predictor of poor outcome in breast cancer.

Abstract

The zinc finger protein LMO4 is overexpressed in a high proportion of breast carcinomas. Here, we report that overexpression of a mouse mammary tumor virus (MMTV)-Lmo4 transgene in the mouse mammary gland elicits hyperplasia and mammary intraepithelial neoplasia or adenosquamous carcinoma in two transgenic strains with a tumor latency of 13-18 months. To investigate cellular processes controlled by LMO4 and those that may be deregulated during oncogenesis, we used RNA interference. Down-regulation of LMO4 expression reduced proliferation of human breast cancer cells and increased differentiation of mouse mammary epithelial cells. Furthermore, small-interfering-RNA-transfected breast cancer cells (MDA-MB-231) had a reduced capacity to migrate and invade an extracellular matrix. Conversely, overexpression of LMO4 in noninvasive, immortalized human MCF10A cells promoted cell motility and invasion. Significantly, in a cohort of 159 primary breast cancers, high nuclear levels of LMO4 were an independent predictor of death from breast cancer. Together, these findings suggest that deregulation of LMO4 in breast epithelium contributes directly to breast neoplasia by altering the rate of cellular proliferation and promoting cell invasion.

Full text

Page 1

Overexpression of LMO4 induces mammary
hyperplasia, promotes cell invasion, and is a
predictor of poor outcome in breast cancer
Eleanor Y. M. Sum*, Davendra Segara†, Belinda Duscio*, Mary L. Bath*, Andrew S. Field‡, Robert L. Sutherland†,
Geoffrey J. Lindeman*, and Jane E. Visvader*§
*The Walter and Eliza Hall Institute of Medical Research and Bone Marrow Research Laboratories, 1G Royal Parade, Parkville VIC 3050, Australia;†Garvan
Institute of Medical Research, 384 Victoria Street, Darlinghurst NSW 2010, Australia; and‡Department of Anatomical Pathology, St. Vincent’s Hospital,
Darlinghurst NSW 2010, Australia
Communicated by Suzanne Cory, The Walter and Eliza Hall Institute of Medical Research, Melbourne, Victoria, Australia, April 12, 2005 (received for review
November 23, 2004)
The zinc finger protein LMO4 is overexpressed in a high proportion
of breast carcinomas. Here, we report that overexpression of a
mouse mammary tumor virus (MMTV)-Lmo4 transgene in the
mouse mammary gland elicits hyperplasia and mammary intraepi-
thelial neoplasia or adenosquamous carcinoma in two transgenic
strains with a tumor latency of 13–18 months. To investigate
cellular processes controlled by LMO4 and those that may be
deregulated during oncogenesis, we used RNA interference.
Down-regulation of LMO4 expression reduced proliferation of
human breast cancer cells and increased differentiation of mouse
mammary epithelial cells. Furthermore, small-interfering-RNA-
transfected breast cancer cells (MDA-MB-231) had a reduced ca-
pacity to migrate and invade an extracellular matrix. Conversely,
overexpression of LMO4 in noninvasive, immortalized human
MCF10A cells promoted cell motility and invasion. Significantly, in
a cohort of 159 primary breast cancers, high nuclear levels of LMO4
were an independent predictor of death from breast cancer.
Together, these findings suggest that deregulation of LMO4 in
breast epithelium contributes directly to breast neoplasia by al-
teringtherateofcellularproliferationandpromotingcellinvasion.
LIM domain ? oncogene ? proliferation
L
One of the central functions of the LIM domain is to mediate
protein–protein interactions, allowing LMO proteins to act as
adaptors for the assembly of multiprotein complexes. LMO pro-
teins have critical roles in normal development. LMO2 is essential
forembryonichematopoiesisandangiogenesis(2,3),whereasmice
lacking both LMO1 and LMO3 die shortly after birth from un-
known causes (4). Targeted deletion of Lmo4 (4, 5) results in
perinatal lethality accompanied by failure of neural tube closure
and homeotic transformations in the rib cage and cervical verte-
brae,suggestingthatLmo4functionsasamodulatorofHoxprotein
function.
DeregulationofLMOexpressioncanleadtooncogenesis.LMO1
and LMO2 were originally discovered by their association with
recurrent translocations in T cell acute lymphocytic leukemia and
subsequently were shown to act as T cell oncogenes in transgenic
models (1, 6–9). Remarkably, the LMO2 gene was ectopically
activated by retroviral integration in severe combined immunode-
ficient patients who developed T cell leukemia after gene therapy
(10). LMO4 was initially identified in an expression screen by using
serum from a breast cancer patient (11). Moreover, deregulated
expression of LMO4 has been demonstrated in a significant pro-
portion of breast carcinomas (12) and, more recently, in cancers of
the oral cavity (13). In breast cancer, LMO4 expression appears to
be inversely correlated with estrogen receptor ? (ER?) expression
(14). LMO4 can act as a negative regulator of mammary epithelial
differentiation in vitro (12), suggesting that LMO4 may have a role
MO4 belongs to the LIM-only (LMO) subclass of LIM domain
proteinsthataredefinedbytwotandemzincfingerdomains(1).
in governing cell proliferation. We have shown that LMO4 forms
a complex with the corepressor CtIP and BRCA1 in breast epi-
thelial cells and that LMO4 can repress BRCA1-mediated tran-
scriptional activation (15).
To further explore a role for LMO4 in breast oncogenesis, we
generated transgenic mice expressing Lmo4 under the control of
the mouse mammary tumor virus (MMTV) promoter and used
RNA interference (RNAi) to investigate the cellular processes
affected by LMO4 in breast cancer cells. Mammary hyperplasia,
mammaryintraepithelialneoplasia(MIN),and?oradenosquamous
carcinoma were observed in two MMTV-Lmo4 transgenic strains,
demonstrating that overexpression of LMO4 contributes to mam-
mary tumorigenesis. RNAi revealed that down-regulating LMO4
expression substantially reduces the proliferation, migration, and
invasion of breast cancer cells. These findings indicate that LMO4
may induce tumorigenesis by perturbation of cellular proliferation
and motility. Further, the observation that high levels of nuclear
LMO4 correlate with poor patient outcome suggests that LMO4
may provide an additional marker in breast cancer.
Materials and Methods
Generation and Analysis of Transgenic Mice. Therabbitglobinintron
(RG-IVS2) was subcloned as a HindIII–EcoRI fragment into
HindIII–EcoRI sites of the MMTV LTR expression plasmid (16)
to generate the MMTV-RG-SV40 vector to provide a longer
intronicsequenceformRNAstability.AHindIII–EcoRVfragment
spanning the mouse Lmo4 coding region was cloned into MMTV-
RG-SV40 to generate MMTV-Lmo4. A 6-kb MMTV-Lmo4 frag-
ment was microinjected into C57BL?6 fertilized mouse oocytes,
and several transgenic founder mice were identified. Four trans-
genic lines on a BALB?c background (more than six generations)
were further examined. Transgenic mice were identified by PCR
analysis using the following SV40-specific primers: 5?-CTCTA-
GAGGATCTTTGTGAAGG-3?(forward)and5?-GGACAAAC-
CACAACTAGAATGC-3? (reverse). Total RNA was isolated
from mammary tissue of transgenic mice by using TRIzol
(GIBCO?BRL) according to the manufacturer’s instructions, and
Northern blot analysis was performed by using 15 ?g of total RNA
(12). For histology, tissue was fixed in 10% (wt?vol) buffered
formalin, dehydrated, and embedded in paraffin, and sections were
stained with hematoxylin and eosin.
CellLinesandCellTransduction.BT-549,MCF-7,andMDA-MB-231
cells were maintained in RPMI medium 1640 (GIBCO?BRL)
containing 10% FCS. HC11 cells were maintained as described
Abbreviations: siRNA, small interfering RNA; LMO, LIM-only; RNAi, RNA interference;
MMTV,mousemammarytumorvirus;MIN,mammaryintraepithelialneoplasia;HR,hazard
ratio; CI, confidence interval; ER, estrogen receptor; PR, progesterone receptor.
§To whom correspondence should be addressed. E-mail: visvader@wehi.edu.au.
© 2005 by The National Academy of Sciences of the USA
www.pnas.org?cgi?doi?10.1073?pnas.0502990102PNAS ?
May 24, 2005 ?
vol. 102 ?
no. 21 ?
7659–7664
MEDICAL SCIENCES

Page 2

(17). MCF10A(EcoR) breast epithelial cells were kindly provided
by J. Brugge and D. Lynch (Harvard Medical School, Boston) (18).
Full-length LMO4 cDNA carrying a 5?-FLAG tag was subcloned
into the retroviral expression vector pBabe-puro (19), and stable
pools of MCF10A(EcoR) transductants were generated (18) by
using 2 ?g?ml puromycin. MCF10A(EcoR) cells infected with
empty pBabe vector provided a control.
RNAi. Cells plated at a density of 0.5 to 1 ? 105cells per well in
six-wellplatesweretransfectedwith200pmolofLMO4-specificand
control small interfering RNA (siRNA) oligoduplexes (Dharma-
con Research, Layfayette, CO) by using oligofectamine reagent
(Invitrogen) according to the manufacturer’s protocols. The se-
quences of each oligoduplex were as follows: siRNA276, 5?-
GUCGAUUCCUGCGAGUGAAdTdT-3?; siRNA376, 5?-GAU-
CGGUUUCACUACAUCAdTdT-3?; and siRNA276mut, 5?-
GUCCAUUUCUCGGCGUUAAdTdT-3?. The siRNA276mut
oligoduplexisbasedonthesequenceofsiRNA276andcontainssix
transversions (bold and underlined).
Cell Proliferation and Differentiation Assays. Proliferation of RNAi-
treated cells was analyzed by using the Cell Titer 96 AQueous One
solution cell-proliferation assay (Promega) according to the man-
ufacturer’s instructions. Cells were replated into a 96-well flat-
bottom plate in triplicate wells at a density of 2,500 cells per well.
The assay was performed on day 0 to control for cell density and
on subsequent days 1–5. For the differentiation assay, RNAi-
treated HC11 cells were grown to confluency for 1–2 days and then
starved of EGF overnight before the addition of the lactogenic
stimulus [10?6M dexamethasone?5 ?g/ml insulin?5 ?g/ml ovine
prolactin (17), kindly provided by C. Parlow (National Institute of
Diabetes and Digestive and Kidney Diseases)]. After 48 h, cells
were harvested for RNA extraction by using TRIzol; cDNA syn-
thesis and PCR were performed by using primers for ?-casein and
Hprt as described (12).
Transwell Migration and Invasion Assays. In vitro migration and
invasion assays were performed by using 24-well, 8-?m pore
transwell inserts (Becton Dickinson). Cells were first resuspended
in Matrigel (Becton Dickinson) for invasion assays. We seeded 106
(migration) or 5 ? 106(invasion) cells in 200 ?l of serum-free
growth medium in the upper chamber, and 600 ?l of medium with
orwithoutchemoattractant(2%FCSor20ng?mlEGF)wasadded
to the lower chamber. Cells were incubated at 37°C for 5 (migra-
tion) or 16 (invasion) h, then fixed in 10% (wt?vol) buffered-
formalin and stained with 0.5 ?g?ml DAPI (Sigma). Cells on the
uppersurfacewereremovedwithacottonswab,andmigratedcells
on the underside were counted (average of 10 fields at a magnifi-
cation of ?40 per transwell).
Western Blotting. Protein extracts were generated by lysing cells in
Triton X-100 lysis buffer (12). Proteins were separated by SDS?
PAGE; transferred to polyvinylidene difluoride membranes (Mil-
lipore); and probed with rat ?-LMO4 20F8 (20), ?-FLAG (Sigma),
or ?-tubulin Ab (Sigma).
Immunofluorescence. Cells were plated on coverslips in six-well
plates, fixed in 4% paraformaldehyde, permeabilized in 0.1%
TritonX100,andblockedin5%normalgoatserum.Costainingfor
LMO4 and focal adhesion complexes was performed by incubating
withanti-LMO4(5?g?ml)(20)andanti-vinculin(1 ?g?ml;Sigma)
mAbs. Staining was detected with secondary Alexa Fluor 488 and
Alexa Fluor 594 Abs (Molecular Probes). Cells were mounted in
fluorescent mounting medium (DAKO) and visualized by confocal
microscopy (TCS.NT.SP2, Leica, Deerfield, IL).
Patient Cohort. After receiving ethical approval from the St. Vin-
cent’s Hospital Campus Human Ethics Committee, we identified a
cohort of 194 patients with a diagnosis of invasive ductal breast
carcinoma between May 1984 and March 2001. Median follow up
was49.5months(range,1.3–135).Archivalformalin-fixed,paraffin-
embeddedpathologyspecimenswereusedtoconstructeightbreast
carcinoma tissue microarrays (TMAs), which contained up to 45 ?
1.0-mm cores per slide. The pathology of each core on the TMAs
was reviewed by a specialist breast pathologist (A.S.F.) before
immunohistochemistry.
Immunohistochemistry and Scoring. Sections were subjected to tar-
get retrieval solution (DAKO) at 100°C for 20 min with a DAKO
autostainer. Endogenous peroxidase activity was quenched with
3% hydrogen peroxide in methanol, followed by avidin?biotin and
serum-freeproteinblocks(DAKO).Sectionswereincubatedfor30
min with anti-LMO4 mAb (20F8) at 7 ?g?ml, followed by biotin-
ylated rabbit anti-rat IgG (DAKO). A streptavidin-biotin peroxi-
dase detection system was used with 3,3?-diaminobenzidine as
substrate (DAKO). LMO4 immunostaining was assessed indepen-
dently by two observers (D.S. and A.S.F.). Scores were given as the
percentage of carcinoma cell nuclei staining positive, with an
absolute intensity on a scale of 0–4 (0, none; 1, pale; 2, mild; 3,
strong homogenous; and 4, intense). The following criteria were
used to achieve a positive score for LMO4 overexpression: nuclear
intensity, ?2 in ?50% of nuclei.
Statistical Analysis. Kaplan–Meier and the Cox proportional-
hazards model were used for univariate and multivariate analysis
with STATVIEW 5.0software(AbacusSystems,Berkeley,CA).Death
from breast cancer was the end point. The factors that were
prognostic on univariate analysis were assessed in a multivariable
model to identify factors that were independently prognostic. This
analysis was performed sequentially on all patients who had avail-
able tissue (n ? 159).
Results
LMO4 Induces Mammary Hyperplasia and MIN in MMTV-Lmo4 Trans-
genic Mice. To assess the role of Lmo4 as a potential oncogene in
the mammary gland, we generated transgenic mice expressing this
gene under the control of the MMTV long-terminal repeat. The
transgene included rabbit ?-globin and SV40 intronic sequences to
augment mRNA stability, as well as a polyadenylation sequence
(Fig. 1A). Analysis of four independent transgenic strains revealed
the expected ?2.6-kb transgene transcript in mammary tissue.
Moderatetransgeneexpressionwasdetectedduringmidpregnancy,
with the highest levels noted during late pregnancy, when the
MMTV promoter is most active. Western blot analysis confirmed
increased Lmo4 expression during pregnancy and lactation (data
not shown). The two strains (strains 34 and 36) that expressed
highest levels of the transgene were selected for further study (Fig.
1B). The level of transgene-derived Lmo4 mRNA was estimated to
be ?3-fold higher than the endogenous level by Northern blot
analysis(datanotshown).Neitherstrainexpressedthetransgeneat
appreciable levels in nonpregnant nulliparous mammary glands
(Fig. 1B), compatible with the hormone-responsive nature of the
MMTV-LTR(21).Becauseluminalcellsinthemammaryglandsof
both virgin and pregnant mice already express abundant Lmo4 (12,
20), it was anticipated that high levels of transgene-derived Lmo4,
expressed with altered kinetics, would be necessary for the induc-
tion of tumors.
No overt phenotypic abnormalities were detected in the
mammary glands of nulliparous mice or mice during their first
pregnancy by whole-mount and histological analyses. Further-
more, transgenic mice were capable of lactation and their glands
were histologically indistinguishable from wild-type mice (data
not shown). Given that highest Lmo4 expression in transgenic
animals was achieved during pregnancy, we assessed the capacity
of the Lmo4 transgene to promote mammary tumorigenesis in
two strains (strains 34 and 36), both on a BALB?c background,
7660 ?
www.pnas.org?cgi?doi?10.1073?pnas.0502990102Sum et al.

Page 3

after three pregnancies. For strain 34, 3 of 13 transgenic mice
monitored for 14–18 months exhibited frank tumor nodules,
which proved to be MIN (Fig. 1C), as defined by the Annapolis
guidelines (22). The MIN lesions were multilayered and exhib-
ited nuclear pleomorphism and increased mitotic figures, with
two lesions also showing evidence of squamous metaplasia.
Mammary glands from two other mice displayed diffuse acinar
hyperplasia (Fig. 1D) and one of these females developed a
bronchoalveolar carcinoma of the lung at 14 months (Fig. 1E).
For strain 36, four of five multiparous female mice developed
multifocal acinar hyperplasia (Fig. 1G), with a mean latency of
16 months (range, 14–20) and one of these mice also exhibited
Fig. 2.
of protein lysates from MCF-7 and BT-549 breast cancer cells transiently transfected with LMO4-specific siRNA276 and siRNA376, compared with control siRNA
and mock-transfected cells, 2 and 6 days after transfection by using ?-LMO4 (20F8) mAb. Anti-tubulin was used to verify protein loading. (B) The proliferation
rate of siRNA-transfected MCF-7 and BT-549 cells, and mock-transfected cells, was determined from days 0–5, in three independent experiments. Error bars
indicateSEM;n?3.(C)CellcycledefectinLMO4-deficientMCF7cells.siRNAandmock-transfectedcellswerefixedinice-cold70%ethanolfor24hbeforestaining
with0.5?g?mlpropidiumiodide(PI).FACScananalysisrevealedanincreaseinG1?G0cellsandadecreaseinS-phasecellsinLMO4-deficientMCF-7cells,compared
withcontrolcells,inthreeindependentexperiments.ErrorbarsindicateSEM;n?3.*,P?0.05.(D)Down-regulationofLMO4expressioninHC11cellstransiently
transfected with siRNA376 (or a control siRNA) was confirmed by Western blot analysis of cells harvested 3 days after transfection by using ?-LMO4 (20F8) mAb.
Anti-tubulin immunoblotting provided a control. (E) RT-PCR analysis was performed by using total RNA derived from RNAi-treated HC11 cells that were
stimulated with (?) prolactin, insulin, and dexamethasone or unstimulated (?) for 48 h. ?-casein and Hprt were used as markers of differentiation and loading,
respectively. At least three independent experiments were performed.
Down-regulation of LMO4 expression in breast epithelial cells by RNAi inhibits proliferation and augments differentiation. (A) Western blot analysis
Fig. 1.
Lmo4 cDNA was cloned into the MMTV-RG-SV40 vector. (B) Northern blot analysis of total RNA (15 ?g) from the mammary glands of transgenic strains 34 and
36. Filters were hybridized with a transgene-specific SV40 probe followed by a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA probe. dP, day
pregnant.(C–E)Hematoxylinandeosinsectionsofmammaryglandsandlungfromstrain34,showinghighgradeMIN(C)andacinarhyperplasia(D),andalveolar
bronchocarcinoma (E). (F–H) Hematoxylin and eosin sections from strain 36 showing mammary glands from multiparous mice. (F) Littermate control. (G) Diffuse
acinar hyperplasia. (H) Adenosquamous carcinoma. [Original magnification: ?400 (C–E and H) and ?200 (F and G).]
Overexpression of Lmo4 in transgenic mice leads to mammary hyperplasia and tumors. (A) Schematic representation of the MMTV-Lmo4 transgene.
Sum et al. PNAS ?
May 24, 2005 ?
vol. 102 ?
no. 21 ?
7661
MEDICAL SCIENCES

Page 4

an adenosquamous carcinoma (Fig. 1H) at 18 months. Hyper-
plasia was also observed in a third transgenic strain. In contrast,
mammary glands from 13 nontransgenic littermate mice that
underwent three pregnancies did not develop hyperplasia or
tumors during a similar follow-up period of 14–19 months (Fig.
1F). In addition to these findings, lymphoproliferative disease or
lymphomas were observed in four transgenic mice and one
wild-type mouse, presumably reflecting MMTV promoter activ-
ity in lymphoid cells (21) and aging of the animals. Although
Lmo4 is not a potent oncoprotein, these data demonstrate that
it can contribute to mammary tumorigenesis.
Down-Regulation of LMO4 Expression by siRNA Inhibits Proliferation
ofBreastEpithelialCells.TodeterminewhetherLMO4regulatesthe
proliferationofbreastcancercells,weusedRNAitodown-regulate
LMO4expression.Twobreastcancercelllines,MCF-7andBT-549,
were transiently transfected with siRNA oligoduplexes. LMO4-
specific siRNAs (siRNA276 and siRNA376) effectively suppressed
LMO4 expression in both cell lines, whereas a control mutant
siRNA had no effect, nor did mock transfection (Fig. 2A). LMO4
expression was down-regulated rapidly, with low levels evident at
day 1 after transfection, consistent with the short half-life (? 2 h)
of this protein (data not shown). A reduction in LMO4 protein
levels was observed up to day 6 after transfection in both cell lines
with siRNA376, although this was not the case for siRNA276 in
BT-549 cells (Fig. 2A). The LMO4-specific siRNAs substantially
reduced the proliferation of both MCF-7 and BT-549 breast cancer
cells, whereas the control siRNAs had no effect (Fig. 2B). No
increase in apoptosis was observed upon transfection with LMO4-
specific siRNAs (data not shown). FACScan analysis of transfected
MCF-7cellsrevealedthatdown-regulationofLMO4expressionled
to cell cycle changes. An increase in the proportion of G1?G0cells
and a corresponding decrease in S-phase cells was evident in three
independent experiments with a P value of ?0.05, whereas the
control siRNA had no effect (Fig. 2C). Thus, LMO4 appears to
promoteS-phaseentryofbreastepithelialcells,providingevidence
for a role in cell proliferation.
To examine the effect of LMO4-specific RNAi on mammary
differentiation, we used mouse mammary epithelial cells as there is
no suitable human counterpart line to study differentiation. HC11
cells have the ability to differentiate into milk-producing cells upon
treatment with a lactogenic stimulus (17). Down-regulation of
LMO4 expression mediated by siRNA376 (Fig. 2D) was found to
augment ?-casein mRNA levels compared with cells transfected
with the control siRNA (Fig. 2E). These data are consistent with
the notion that LMO4 maintains the proliferative rather than
differentiative state of mammary epithelial cells.
Reduced LMO4 Expression Impedes the Migration and Invasion of
Breast Cancer Cells. To examine the role of LMO4 in cell motility
andinvasion,weusedRNAitodown-regulateLMO4expressionin
Fig. 3.
breastcancercells.(A)WesternblotanalysisofproteinlysatesfromMDA-MB-
231breastcancercellstransfectedwithsiRNA276,siRNA376,controlsiRNA,or
mock-transfectedcells,2and6daysaftertransfectionbyusing?-LMO4(20F8)
mAb. Anti-tubulin provided a control. (B) Focal adhesions in MDA-MB-231
cells transfected with siRNA376 or a control siRNA were visualized by indirect
immunofluorescencebyusing?-vinculinmAb.Costainingwith?-LMO4(20F8)
mAb confirmed the reduction in LMO4 expression in siRNA376-transfected
cell nuclei. [Scale bars, 20 (i and ii) and 8 (iii and iv) ?m.] (C) The number of
migrating MDA-MB-231 cells transfected with either LMO4-specific siRNA276
orsiRNA376,oramutantsiRNA,wasdeterminedbycounting10randomfields
in each of three independent experiments. Error bars indicate SEM; n ? 3. (D)
The number of MDA-MB-231 cells, transfected with siRNA276, siRNA376, or a
mutant siRNA, capable of invading Matrigel was determined for three inde-
pendent experiments, as in C. Error bars indicate SEM; n ? 3.
Reduced LMO4 expression impedes the migration and invasion of
Fig. 4.
and invasion. (A) The motility of MCF10A(EcoR) cells, stably transduced with
a FLAG-LMO4 retroviral construct or vector alone (pBabe-puro), was deter-
minedbycountingthenumberofmigratedcellsin10randomfieldsfromeach
of four experiments. Error bars indicate SEM; n ? 4. (B) The number of
MCF10A(EcoR) cells expressing either FLAG-LMO4 or vector alone (pBabe-
puro), capable of invading through Matrigel, was determined in four exper-
iments, as described for A. Error bars indicate SEM; n ? 4. (C) Western blot
analysisconfirmedtheexpressionofFLAG-LMO4inthetransductants.Lysates
from cells expressing FLAG-LMO4 or empty vector were subjected to SDS?
PAGEandimmunoblottedwithmouse?-FLAGmAb.Anti-tubulinwasusedto
control for protein loading.
LMO4 overexpression augments mammary epithelial cell migration
7662 ?
www.pnas.org?cgi?doi?10.1073?pnas.0502990102 Sum et al.

Page 5

MDA-MB-231cells,aninvasivebreastcancercellline.Transfection
with either of the LMO4-specific siRNAs profoundly reduced
LMO4 expression in MDA-MB-231 cells, whereas the control
siRNA did not (Fig. 3A). Transfection of MDA-MB-231 cells with
siRNA276 and siRNA376 resulted in a 2.5- and 3.8-fold decrease
in cell motility, using transwell migration assays in the presence of
2% FCS, compared with that for control cells (Fig. 3C). Next, we
examined invasion of these cells through an extracellular matrix
(Matrigel)inamodifiedBoydenchamber,whichisanassaythathas
been shown to correlate with in vivo metastatic potential (23).
siRNA276 resulted in a 1.8-fold decrease in the invasion of MDA-
MB-231 cells, whereas siRNA376 elicited a 2.6-fold decrease (Fig.
3D). These findings indicate that LMO4 regulates the motility and
invasiveness of breast cancer cells. To determine whether LMO4
influences focal adhesions, we assessed vinculin distribution in
siRNA-transfected MDA-MB-231 cells by indirect immunofluo-
rescence. Focal adhesions appeared to be less prominent in MDA-
MB-231 cells transfected with siRNA376, compared with cells
transfected with a control siRNA (Fig. 3B). Costaining with anti-
LMO4 mAb (20F8) confirmed a substantial decrease in LMO4
levels in the nuclei of siRNA376-transfected cells (Fig. 3B).
LMO4 Overexpression Stimulates Mammary Epithelial Cell Migration
and Invasion.TofurtherinvestigatetheroleofLMO4incellmotility
and invasion, we transduced human breast epithelial MCF-10A
cells harboring the murine ecotropic receptor, MCF-10A(EcoR),
withaLMO4-expressingretrovirus.MCF10A(EcoR)cellsexpress-
ing FLAG-LMO4 showed a 2.9-fold increase in migration when
EGFwasusedasthechemoattractant,comparedwithcellsinfected
with a control virus (Fig. 4A). Moreover, a 2.1-fold increase in the
number of invasive cells was observed for MCF10A(EcoR) cells
transducedwithpBabe-FLAG-LMO4,relativetocontrolcells(Fig.
4B). Expression of FLAG-LMO4 was confirmed by Western blot
analysis using ?-FLAG Ab (Fig. 4C). Thus, either down-regulation
or overexpression of LMO4 markedly affects the motility of mam-
mary epithelial cells and their ability to invade an extracellular
matrix in vitro.
LMO4 Overexpression in Breast Cancer Correlates with Poor Clinical
Outcome. We have reported in a small series of 60 cases that LMO4
protein is overexpressed in ?50% of primary breast cancers (12).
To further investigate the relationship between LMO4 expression
and clinicopathological features of the disease, we performed
immunohistochemical analysis on tissue from a cohort of 194
primary breast cancer patients of known clinical outcome.
LMO4 overexpression, as defined in Materials and Methods, was
apparent in 68 of 159 (42.7%) ductal carcinomas from which tissue
was available (Fig. 5A). In a univariate Cox proportional-hazards
regressionanalysis,highnuclearLMO4expressionwassignificantly
associated with decreased patient survival (P ? 0.023), as were
other well established markers of outcome [i.e., tumor size; tumor
grade; axillary lymph node status; and ER, progesterone receptor
(PR), and HER2 status] (Table 1). In a multivariate analysis
incorporating these parameters, only axillary lymph node status
[hazard ratio (HR), 4.02; 95% confidence interval (CI), 1.4–11.5;
P ? 0.009] and LMO4 nuclear overexpression (HR, 2.27; 95% CI,
1.01–5.11; P ? 0.048) were independent predictors of death from
breast cancer (Table 2). Kaplan–Meier analysis confirmed a sig-
nificant relationship between LMO4 overexpression and patient
survival (Fig. 5B). Interestingly, most (15?23) tumors overexpress-
ing HER2, an adverse prognostic marker, also overexpressed
LMO4 (Fisher’s exact test; P ? 0.0247). However, HER2-positive
tumors only represented 23% of LMO4-positive tumors (64
samples).
Discussion
Here, we provide evidence that LMO4 is an important regulator of
breast epithelial cell proliferation and invasion and that it can act
asamammaryoncoproteininatransgenicmousemodel.Enforced
expression of Lmo4 in the mammary glands of transgenic mice
recapitulatesthehighlevelsofLmo4thatoccurin?50%ofprimary
breast cancers. The development of mammary hyperplasia and
MIN in MMTV-Lmo4 transgenic mice supports the notion that
Fig.5.
outcome. (A) LMO4 immunostaining was performed on tissue microarrays
containing archival breast tumor specimens using ?-LMO4 (20F8) mAb. Rep-
resentative images of tumor specimens displaying low (Left) and high (Right)
levels of LMO4 expression are shown. (B) Kaplan–Meier curves for overall
survival(OS)of159breastcancerpatientsaccordingtoLMO4nuclearstaining.
High LMO4 expression was significantly associated with decreased OS in
univariate (P ? 0.023) analysis.
OverexpressionofLMO4inbreastcancerisapredictorofpoorclinical
Table 2. Multivariate analysis of clinicopathological parameters
ParameterHR (95% CI)
P value
Tumor size (?20 mm)
Axillary lymph node-positive
High grade
ER-positive
PR-positive
HER2-positive
LMO4-positive
1.51 (0.59–3.83)
4.02 (1.40–11.5)
3.03 (0.96–9.59)
0.28 (0.07–1.12)
1.16 (0.31–4.38)
1.54 (0.61–3.89)
2.27 (1.01–5.11)
0.389
0.009
0.059
0.072
0.822
0.367
0.048
Table 1. Univariate analysis of clinicopathological parameters
ParameterCohort, % HR (95% CI)
P value
Tumor size (?20 mm)
Axillary lymph node involvement
High grade
ER-positive
PR-positive
Her2-positive
LMO4-positive
42
54
52
71
66
15
43
3.83 (1.74–8.43)
5.40 (2.06–14.18)
7.45 (2.59–21.42)
0.19 (0.09–0.41) ?0.0001
0.29 (0.14–0.60)
2.40 (1.01–5.66)
2.38 (1.12–5.04)
0.0008
0.0006
0.0002
0.0009
0.047
0.023
Although the cohort comprised 194 patients, it was only possible to score
159 samples for LMO4 because of tissue loss from the tissue microarrays.
Therefore, all analyses were on these 159 samples. ER and PR status were not
available for two samples.
Sum et al.PNAS ?
May 24, 2005 ?
vol. 102 ?
no. 21 ?
7663
MEDICAL SCIENCES