Differential subcellular localisation of the tumour suppressor protein LIMD1
in breast cancer correlates with patient survival
Ian Spendlove1*, Ahmad Al-Attar1, Oliver Watherstone1, Thomas M. Webb2, Ian O. Ellis3,
Gregory D. Longmore4,5and Tyson V. Sharp2*
1Academic and Clinical Department of Oncology, University of Nottingham, Nottingham, United Kingdom
2School of Biomedical Sciences, University of Nottingham Medical School, Queen’s Medical Centre, Nottingham, United Kingdom
3Department of Histopathology, University Hospitals Nottingham, Nottingham, United Kingdom
4Department of Medicine, Washington University, St. Louis, MO
5Department of Cell Biology, Washington University, St. Louis, MO
The tumour suppressor gene (TSG) LIM domain containing pro-
tein 1 (LIMD1) has been associated with transformation of epithe-
lial cells of the lung and its expression is downregulated in all lung
tumour samples tested compared to normal lung matched con-
trols. In the first study of its kind we used an anti-LIMD1 specific
monoclonal antibody to investigate expression/localisation of the
LIMD1 protein in a well-characterised tissue microarray of breast
cancers and normal adjacent epithelia. Comparison of tumour
with adjacent normal and distant normal tissue demonstrated that
LIMD1 expression is moderate to high compared to tumour.
There was also a significant correlation with histological grade
(p 5 0.0001), tumour size (p 5 0.013) and tumour type (p 5 0.004)
indicating an association with aggressive disease. Cytoplasmic
LIMD1 expression was seen in 99.3% of cases, with 43.1% show-
ing both nuclear and cytoplasmic localisation. Absence/loss of nu-
clear staining showed a strong correlation with patient survival
and was indicative of poor prognosis (p 5 0.033). There was no
association with lymph node status and other clinicopathological
parameters. Nuclear staining was more pronounced in better
prognosis tumours and normal tissue. This study demonstrates
that LIMD1 represents a novel prognostic marker for breast can-
cer. Combined with the fact that LIMD1 expression is downregu-
lated in lung cancers this clearly indicates that LIMD1 may repre-
sent a critical TSG, the function of which is deregulated via overall
loss of expression and/or relocalisation within the cell during
tumour development. The possible functions of LIMD1 localisa-
tion within the nucleus and cytoplasm and its relationship to
tumour prognosis are discussed.
' 2008 Wiley-Liss, Inc.
Key words: LIMD1;
microarray; prognostic indicator
Ajuba; breast cancer; E-cadherin;tissue
Breast cancer is the most common cause of cancer-related death
in women in the western world. Alterations in tumour suppressor
genes (TSG) and oncogenes are common causes for the higher
proliferative rates and pathogenic characteristics associated with
tumour development. Cytogenetics and allelotyping have shown
that allelic loss from several distinct regions on chromosome 3p,
including 3p25, 3p21–22, 3p21.3, 3p12–13 and 3p14, are the ear-
liest and most frequent genomic abnormalities involved in a wide
spectrum of major epithelial cancers of lung1–6and breast
tissues7–9as well as many other epithelial derived cancers.10–13
Specifically the 3p21.3 gene cluster contains many TSG and onco-
genes and is a region of major loss of heterozygosity in lung
cancers.14–16Abnormalities at the 3p21.3 chromosomal region
include homozygous deletions, loss of heterozygosity and
expressional deficiencies that can be induced through methyla-
tion of promoters, altered mRNA transcripts and loss of protein
The chromosome 3 commonly eliminated region (C3CER1) is
located within the 3p21.3 gene cluster. The C3CER1 contains sev-
eral putative TSG and is eliminated in many forms of cancer
including breast, gastric, colorectal, ovarian and renal.18The
region spans 1.4 Mega bases and covers 19 active genes including
the LIMD1 gene, which is examined in this study.
We have shown that LIMD1 interacts with the retinoblastoma
protein (pRB) to inhibit E2F-mediated transcription.19LIMD1
blocks tumour growth and is downregulated in all human lung
cancer samples tested.19These data indicate LIMD1 is a TSG and
its loss may promote lung carcinogenesis.
Recent findings have demonstrated that LIMD1 and the Ajuba
family members are essential components of the Slug/Snail tran-
scriptional repressing complex involved in neural crest develop-
ment and possible epithelial mesenchymal transition (EMT).20–22
More specifically, LIMD1 and family member Ajuba inhibit E-
cadherin (ECAD) transcription via binding to the SNAG domain
of Snail and thus act as a co-repressing complex with Snail on the
ECAD promoter.20–22With the relationship of Snail activity and
regulation of ECAD expression in breast cancer progression and
prognosis, this suggests a possible role for LIMD1 function in the
development of this neoplasm.20–22
Further characterisation of LIMD1 expression in normal and
cancerous tissue is critical for our understanding of how early loss
or deregulation of this TSG may contribute to the development of
epithelial derived cancers. To this end we expanded our investiga-
tions of LIMD1 into breast cancer to determine if the loss of
LIMD1 expression seen in lung tumours was also recapitulated in
breast tumours, thus further supporting the role of LIMD1 as a key
TSG that undergoes loss of function via 3p21 ablation.
Here we report the first large-scale tissue microarray (TMA)
analysis of normal and breast cancer samples for LIMD1 protein
expression. In comparison to the loss of expression of LIMD1 in
lung cancer samples compared to normal lung the expression of
LIMD1 in breast cancer was very distinct. Rather than a loss of
expression we observed a variety of differential subcellular local-
isation of the LIMD1 protein which correlated with tumour type.
Our investigation determined that the downregulation and/or elim-
ination of LIMD1 expression from the nucleus of neoplastic cells
correlated strongly with poor patient prognosis, aggressive forms
of breast carcinoma, increased tumour size and high histological
grade. Conversely, strong nuclear LIMD1 staining and localisation
correlated with low-tumour grade and better patient prognosis.
Abbreviations: EMT, epithelial-mesenchymal transition; ER, oestrogen
receptor; IHC, immunohistochemistry; LIMD1, LIM domains-containing
protein-1; NES, nuclear export signal; NPI, Nottingham Prognostic Index;
TMA, tissue microarray; TSG, tumour suppressor gene.
Grant sponsors: British Lung Foundation, BBSRC.
*Correspondence to: School of Biomedical Sciences, University of
Nottingham Medical School, Queen’s Medical Centre, Nottingham, NG7
2UH, United Kingdom. Fax: 111-508-203-0142. E-mail: tyson.sharp@
nottingham.ac.uk or Academic and Clinical Department of Oncology,
University of Nottingham, Nottingham, United Kingdom.
Received 10 April 2008; Accepted after revision 2 July 2008
Int. J. Cancer: 123, 2247–2253 (2008)
' 2008 Wiley-Liss, Inc.
Publication of the International Union Against Cancer
Material and methods
This research was approved by the Nottingham Local Research
Ethics Committee and its design and reporting seeks to adhere to
the REMARK recommendations.23
A TMA representing 819 (of which 495 were valid for IHC)
cases of primary operable invasive breast carcinoma, from patients
aged 70 years or less, diagnosed between 1987 and 1998 from the
Nottingham Tenovus Primary Breast Carcinoma Series was used.
These well-characterised breast cancer cases have been treated by
a conventional algorithm and this series has been successfully
used to study the proteomics of breast cancer; being powerful
enough to detect small variations in clinical outcome.24–28
Patient characteristics including age and menopausal status
along with information concerning recurrence and survival were
maintained on a prospective database. Patients were followed up
at 3-month intervals initially, then 6-month and annually for a me-
dian period of 76 months.
Breast cancer tissue arrays with self-matching adjacent normal
tissue and distal normal tissue were purchased from Tissue Array
Network (Rockville, MD), Catalogue number BR721 and stained
for LIMD1 via IHC as below.
The characteristics of these cancers including histological grade
(grade 1 5 low; grade 2 1 3 5 high), lymph node stage, tumour
size and Nottingham Prognostic Index (NPI) were prospectively
recorded in the database. The tumours were classified into the fol-
lowing 4 prognostic types for analysis.
1. Excellent prognosis type (>80% 10-year survival), including
tubulo-lobular, tubular mucinous and invasive cribiform car-
2. Good types (60–80% 10-year survival), including tubular
mixed, mixed ductal with special type and alveolar lobular
3. Moderate prognosis type (50–60% 10-year survival), includ-
ing classical lobular, medullary, atypical medullary and lob-
ular mixed carcinoma.
4. Poor prognosis types (?50% 10-year survival), including
ductal NST, solid lobular, mixed ductal and lobular carci-
Patient management was guided by NPI and oestrogen receptor
(ER) status. The assays to determine the ER status of each tumour
section were conducted by a dextran-coated charcoal technique at
the Tenovus Institute, Cardiff. A cut-off of 10 fmol/mg protein
indicated positivity. Those women with an NPI score ?3.4
received no adjuvant therapy and those with an NPI score >3.4
received tamoxifen if ER positive (6Zoledex if pre-menopausal)
or CMF (classical chemotherapy) if ER negative and fit enough to
The excised tumours were sliced, fixed immediately in neutral-
buffered formalin and processed through embedding in paraffin
wax. Tumour samples were then arrayed as previously
described.29Tumour biopsies with a diameter of 0.6 mm were
taken from representative regions of each donor block using man-
ual tissue arrayer (Beecher Instruments, US) and then precisely
arrayed into a new recipient paraffin block. Contrary to expecta-
tion, tissue heterogeneity does not significantly influence the pre-
dictive power of the TMA results.30Single cores were analysed in
Primary antibodies. The characterised anti LIMD1 monoclo-
nal antibody 3F2/C619was titrated on breast tissue sections. A
working dilution of 1/200 antibody in 5% normal swine serum
(NSS)/TBS was subsequently used to stain the arrays. ECAD im-
described using the standard streptavidin-biotin complex.32,33
Staining protocol. Deparaffinised 4-lm thick sections of TMA
and similar thickness whole sections of breast tumours (103)
were immersed in 0.3% hydrogen peroxide for 15 min to block en-
dogenous peroxidase activity. Microwave pre-treatment in EDTA
buffer (pH 8.5) was performed at 10 min high power and 10 min
low power to retrieve antigenicity. The primary antibody was
incubated on the slides for 60 min at a 1:200 dilution. Primary
antibody was omitted from the negative control, which was left
incubating in normal swine serum (NSS). Sections of breast
tumour found to express LIMD1 acted as positive and negative
controls. The sections were then incubated with an HRP-conju-
gated rabbit anti-mouse secondary antibody (Dako, Denmark -
P0141) for 45 min. 3,30-diaminobenzidne (DAB) (Dako, Den-
mark) was applied for visualisation. The sections were lightly
counterstained with haematoxylin (Dako, Denmark), dehydrated
in alcohol, cleared with xylene (Genta Medica, York, UK) and
mounted with distyrene, plasticizer and xylene (DPX) (BDH,
Evaluation of staining
On reviewing the distribution, variety and frequency of staining
within cores we adopted a binary system of quantifying the
expression of LIMD1. Those cores with >5% of cells staining
were deemed positive and others negative. The full scoring was
performed by one author (A.A.), a high-level of concordance with
a second author (I.S.) was achieved over an initial 100 cases.
SPSS v. 14.0 was used for the statistical analyses (Chicago,
US). Univariate associations of LIMD1 staining with the clinico-
pathological parameters of breast cancer, the expression of ECAD
and 5-year survival were analysed by Pearson v2. The effect of
LIMD1 expression upon overall disease-specific survival was ana-
lysed using Kaplan-Meier curves and the log rank test. To estimate
the size and independence of effects, Cox’s regression was
employed; returning hazards ratios and 95% confidence intervals.
p values < 0.05 indicated statistically significant differences.
LIMD1 expression in normal breast tissue
In this study we wished to determine if LIMD1 protein was
expressed in normal breast tissue and whether such expression is
lost or changed in breast cancer samples; hence reflecting the pos-
sible role of regulation of this TSG in initiation of carcinogenesis.
We screened a range of normal breast tissue sections to deter-
mine LIMD1 expression pattern. LIMD1 was specifically
expressed in the epithelial cells of terminal duct lobular units
(TDLUs) (Fig. 1). Within normal epithelial cells we observed
intense stain for LIMD1 in the cytoplasm and nucleus (Fig. 1). In
some cases the staining for LIMD1 expression was so dark that
many of the nuclei could not be seen. Such staining patterns sup-
port previous studies showing LIMD1 family members localising
to the cytosol19,34and more recently LIMD1 itself localised to the
nucleus.35In contrast LIMD1 was not present in foam cells or the
surrounding stroma of the breast (Fig. 1).
LIMD1 has distinct patterns of expression and subcellular
localisation in different breast cancer patient samples
To evaluate the potential contribution of LIMD1 protein expres-
sion to the development and progression of breast cancer, we
screened a 495 breast cancer TMA (Refs. 26 and 27). The initial
SPENDLOVE ET AL.
observation was that LIMD1 protein expression in breast cancer
tissue had distinct nuclear and cytoplasmic localisation patterns,
but the majority of samples expressed LIMD1 (Figs 2a and 2b).
This is in contrast to analysis of lung cancer where loss of LIMD1
expression predominates, supporting the observation that muta-
tions or deletions in LIMD1 in breast cancers at the C3CER1
region may not occur as frequently compared to lung.18,36
This immunohistochemical analysis of the breast cancer TMAs
indicated that the majority of cellular staining was weak to moder-
ate and localised to the cytoplasm (Figs. 2a and 2b). LIMD1
staining was present in 489 of 495 cores tested (98.8%). Strong
cytoplasmic staining had a frequency of 14.7% (73/495) and was
associated with strong or moderate nuclear staining when present
(Fig. 3d). The moderate cytoplasmic staining had a prevalence of
44% (218/495) with the weak staining 40% (198/495) and no cyto-
plasmic staining accounted for 1.2% (6/495).
The staining within the nucleus was less frequent than cytoplas-
mic staining with no nuclear staining accounting for 66.9% of all
the tumour cores (Figs. 2a and 2b). Nuclear staining was predomi-
nantly moderate/strong accounting for 16.8 and 15.2% of all cores
tested, respectively (Figs. 2a and 2b). Weak staining within the
nucleus was at a lower frequency than weak staining of the cyto-
plasm accounting for just 1.2% of all cores compared to 41.3% of
The level of LIMD1 expression in the cytoplasm of tumour cells
directly correlates with that in the nucleus
LIMD1 and its family members constantly shuttle between the
nucleus and cytoplasm in a CRM1/Ran-GTP dependent nuclear
export manner.37The equilibrium of which is toward nuclear
export thus resulting in most family members including LIMD1
being predominantly localised to the cytoplasm.19,37,38In light of
the observation of differential staining patterns and intensities for
LIMD1 in different breast tumour samples, we examined whether
there was a direct correlation between high cytoplasmic levels of
LIMD1 and high nuclear levels and if this would therefore indi-
cate a change in the equilibrium of distribution due to increased
cytoplasmic expression. Statistical analysis showed a significant
correlation between high cytoplasmic LIMD1 expression and high
nuclear LIMD1 expression (p < 0.0001). Figure 3 shows represen-
tative nuclear and cytoplasmic staining patterns for LIMD1 in
breast tumour cores from the TMA used in this study, demonstrat-
ing the different nuclear and cytoplasmic patterns of expression.
The significant correlation above was confirmed in representative
examples showing that low cytoplasmic expression correlated
with negative/low nuclear staining (Fig. 3a) and high cytoplasmic
LIMD1 expression with high nuclear LIMD1 expression (Fig. 3d)
and gradations of expression between these 2 extremes (Figs. 3b
Absence of nuclear LIMD1 staining correlates
with a decrease in patient survival
The observation of clear differences in LIMD1 distribution was
further examined. To evaluate the potential contributions of this
differential expression pattern in different tumours, we examined
the association of LIMD1 expression with clinicopathological
We first examined patient survival with the nuclear intensity of
LIMD1 expression. These data indicate a correlation between a
reduced LIMD1 staining intensity (strong > moderate > no/weak)
and the expected patient survival (Table I). Absent/weak nuclear
staining intensity had strong correlation with poor patient progno-
sis (p 5 0.033), tumour type (p 5 0.004), tumour size (p 5
0.013), histological grade (p 5 0.00001), ER status (p 5 0.015)
and NPI (p 5 0.007) but not lymph node status (Table II). A
Kaplan-Meier plot further illustrates the correlation between no/
weak staining intensity of the nucleus and poor patient prognosis
The observation of correlation between the level of expression
and localisation of LIMD1 with prognosis and grade, etc (above)
led us to reason that if we were to take paired normal breast and
matched tumour TMA and compare them for LIMD1 expression
FIGURE 1 – Normal breast tissue staining for the cellular localiza-
tion of the LIMD1 protein. Breast epithelial cells show the brown
pigment which denotes expression of LIMD1 protein. Staining is
negative for adipocytes and the connective tissue cells. 3200 magnifi-
cation, 3F2/C6 1:500 dilution. Inset: higher magnification to show
FIGURE 2 – Charts illustrating the distribution of staining patterns
and level for LIMD1 between the nucleus and the cytoplasm from the
analysis of the TMAs. (a) The distribution of staining patterns and lev-
els for LIMD1 in the nucleus and the cytoplasm from the analysis of
495 breast cancer tissue microarray cores. (b) The distribution of
breast cancer cores that express both nuclear and cytoplasmic LIMD-1
(solid bar) versus those that are missing either cytoplasmic or nuclear
staining (white bar). ‘‘Frequency’’ indicates number of TMA cores
with the indicated expression and distribution. Specific values for
TMA cores are shown on top of bars.
TUMOUR SUPPRESSOR PROTEIN LIMD1 IN BREAST CANCER
pattern and level we would predict that normal breast would
express both high levels of cytoplasmic and nuclear LIMD1 and
its matched adjacent tumour would express low cytoplasmic levels
and low (or undetectable) levels of nuclear LIMD1. We therefore
obtained an additional TMA with 11 paired breast cancer and
matched adjacent and distal normal control tissue and performed
anti-LIMD1 IHC. Figure 5 shows that this prediction was indeed
the case and all of normal breast tissue had higher levels of cyto-
plasmic LIMD1 expression compared to their adjacent matched
tumour. Scoring the normal and cancer tissue for LIMD1 intensity
revealed a significant reduction in expression of LIMD1 in the
tumour tissue compared to normal (p 5 0.006) (Fig. 5b). This was
maintained when comparing either adjacent or distal normal tis-
sue. Similarly there was no significant difference in staining
between adjacent and distal normal LIMD1 intensity (p 5 0.8)
FIGURE 3 – Representative nuclear and cytoplasmic LIMD1 staining patterns in breast cancer TMA cores. (a) Non-nuclear staining for
LIMD1 protein. This low cytoplasmic staining category showed weak levels LIMD1 staining and weak to no nuclear staining. (b) Weak to mod-
erate cytoplasmic staining with moderate to strong nuclear LIMD1 stain. (c) Moderate cytoplasmic stain with strong nuclear LIMD1 stain. (d)
Strong cytoplasmic and strong nuclear LIMD1 stain. All cores were fixed and stained as described in ‘‘Material and methods.’’ The LIMD1
mAb (3F2/C6) was used at 1:200 dilution. Panels are at 3200 magnification with digital zoom for inset details.
TABLE I – THE NO/WEAK STAINING INTENSITIES OF THE NUCLEUS FOR
LIMD1 COMPARED WITH PATIENT SURVIVAL
The no/weak group had a mortality rate of 15.7% compared with
8.4% for moderate staining and 5.3% for strong staining p 5 0.033,
indicating a correlation between poor prognosis and loss of nuclear
TABLE II – PROGNOSTIC MARKERS AND CORRELATION
WITH LIMD1 STAINING
Prognostic factors Cut off points
Histological grade Well/moderate/poor
LN (2ve) LN
<10, 11–20, 21–30,
31–40, 41–50 mm
<40, 41–50, 51–60,
p 5 0.00001
Lymph node (LN)
p 5 0.013
p 5 0.007
Tumour type in
p 5 0.004
(log rank test)
p 5 0.015
p 5 0.033
Summary of LIMD1 correlation with clinicpathology of breast can-
p < 0.05 is considered significant.
SPENDLOVE ET AL.
This current report of LIMD1 protein expression in breast can-
cer represents the first large-scale (n 5 495) IHC TMA analysis
study of its kind examining a member of the Ajuba/Zyxin LIM do-
main containing family of proteins. Previously there has only been
a small pilot study reported for the Zyxin family member, LPP, on
ductal carcinomas of the breast (n 5 53); with no report of any
correlation with clinicopathological data.39In contrast, we show
the specific subcellular staining pattern for LIMD1 (no nuclear
and weak to moderate cytoplasmic staining) correlated sig-
nificantly with poor patient prognosis (p 5 0.033), tumour type
(p 5 0.004), tumour size (p 5 0.013) and the (NPI) (p 5 0.007)
(Table II). These data are the first reported for an Ajuba/Zyxin
family member exhibiting differences in nuclear/cytoplasmic
localisation in tumour tissue samples and in addition a correlation
of this expression and distribution pattern with clinicopathological
features of breast cancer. The ability of this family of proteins to
shuttle between the nucleus and cytoplasm is well documented.37
Furthermore, many biochemical signal transduction pathways
have been associated with this inherent ability.37However, this is
the first demonstration that differentiation in intracellular localisa-
tion of one family member, LIMD1, may correlate in situ with dis-
ease pathogenesis. There is a precedent for such an observation of
TSG subcellular localisation correlating to breast cancer progno-
sis.40Specifically, the ability of HER-2/neu level to influence the
relocalization of p21WAFI/CIPIfrom the nucleus to the cytoplasm,
resulting in a loss of p21WAFI/CIPItumour suppressor function
which correlates with poor prognostic tumours.41Furthermore, it
has been suggested in 2 independent studies that loss of nuclear
BRCA1 expression in breast cancers is associated with highly pro-
liferative tumour phenotypes.42,43In light of these commonalities
for re-compartmentalisation of TSG function from the nucleus to
the cytoplasm, it will be of interest in future studies to determine
the regulatory mechanisms that control re-distribution of these
TSG or indeed if there are any common mechanisms which may
regulate one or more critical pro-proliferative related gene products.
On the basis of the correlation of the expression and subcellular
distribution of LIMD1 with diseases pathogenesis and prognosis,
we propose a model which shows how LIMD1 expression, as
determined by IHC, correlates with mortality rate (Fig. 5). This
figure also summarises the key observations of this study. The va-
lidity of this model was demonstrated by the comparison of a sep-
arate breast cancer TMA containing adjacent and distal normal tis-
sue. When stained by IHC for LIMD1 expression this showed that
normal breast had a staining pattern similar to that of good prog-
nosis breast cancer whereas tumour stained with a similar pattern
to poor prognosis cancers observed in the first array (Fig. 6).
Although this was not independently significant, this may be due
to the small population size of the groups identified. Further analy-
sis on larger TMAs may reveal statistical significance with respect
to this analysis, but this is currently beyond the scope of this
The fact that LIMD1 is expressed to varying degrees in all nor-
mal and tumour tissues examined in this study corroborates the
recent report by Huggins et al.,36where they analysed the genomic
DNA from 235 breast cancer for LIMD1 polymorphisms and
found only 0.85% to have any changes. These were 4 coding
region alterations, including 2 amino acid substitutions at posi-
tions 255 and 302 with no known attributed effects on LIMD1
function. Their data suggests that LIMD1 is not mutated or deleted
in the majority of breast cancers. More recently the same group
have demonstrated that 80% (16/20) of breast cancers have
unchanged levels of LIMD1 mRNA compared to controls35and
this supports our findings that LIMD1 was present and expressed
in all breast cancer samples tested. Our findings would suggest
that it is the regulation of LIMD1 protein levels, perhaps by epige-
netic regulation that account for differential expression. However,
the preliminary pilot study by Huggins and Andrulis,35where they
compared the LIMD1 promoter methylation status in 4 breast
cancer samples with LIMD1 mRNA, indicated that the degree of
FIGURE 4 – A Kaplan-Meier plot measuring patient survival time in
months against cumulative survival for nuclear LIMD1 staining.
Staining was categorised into 3 groups; strong, moderate and weak/no
staining (which have been culminated).
FIGURE 5 – Expression of LIMD1 in matched normal and adjacent
tumour samples. (a) Anti-LIMD1 IHC staining of 11 matched sample
breast TMA sets of tumour, adjacent normal and distal normal. Two
representative sets of stained cores are shown. Panels are 320 magni-
fication with digital zoom for inset details. (b) Scatter plot summary
of anti-LIMD1 staining for the 11 sets of matched tumour, adjacent
normal and distal normal breast TMA cores. Cores were scored nega-
tive, weak, moderate or strong (0, 1, 2 and 3, respectively).
TUMOUR SUPPRESSOR PROTEIN LIMD1 IN BREAST CANCER
hypermethylation did not affect expression levels. However, how
informative such a small sample size can be (only 2 breast cancer
samples were examined with reduced LIMD1 mRNA) remains to
be determined. If epigenetic regulation of LIMD1 does not tran-
spire to be the main regulatory mechanism for levels of expression
then protein stability and posttranslational modification may play
an important role. This may certainly be the case with respect to
differential localisation of LIMD1 protein which may ultimately
result in causing its dysfunction thus contributing to breast cancer
development and progression. The observation for LIMD1 expres-
sion in breast cancer seems to be in contrast to complete loss of
expression observed in lung cancer which is most probably due to
3p21.3 gene deletions.16
The Ajuba LIM proteins have very recently been shown to be
recruited to the endogenous ECAD promoter in a Snail-dependent
manner.20In vivo, Ajuba, LIMD1 and WTIP cooperate with Snail
and Slug to mediate Xenopus neural crest development.22Thus, in
addition to regulating epithelial cell–cell adhesion, Ajuba LIM
proteins contribute to epithelial-mesenchymal transitions as Snail
co-repressors during neural crest development.22Furthermore,
with the relationship of Snail activity and downregulation of
ECAD expression in invasion and metastasis, this suggests a pos-
sible role for LIMD1 function in the development of advanced dis-
seminated disease.21,22The TMA used in this study had previously
been analysed for expression of ECAD, demonstrating that loss of
ECAD was an independent indicator of poor prognosis/survival
(p 5 0.013).33We therefore analysed this group for co-expression
of both ECAD and LIMD1. With respect to the Snail-LIMD1
interaction and functions we reasoned that LIMD1 expression
would inversely correlate with that of overall ECAD. Although
this held true for cytoplasmic LIMD1 and ECAD expression (p 5
0.035) this did not maintain significance for LIMD1 nuclear
expression (p 5 0.059) possibly due to small numbers in this latter
group. This inverse correlation supports very recent findings by
Langer and co-workers, demonstrating the ability of LIMD1 to
co-repress with Snail the transcription and thus expression of
The expression pattern of LIMD1 in breast cancer has not previ-
ously been investigated. Our results provide strong evidence that
the specific subcellular localisation and levels of LIMD1 between
the nucleus and cytoplasm of breast cancer cells is important in
breast cancer progression. Furthermore, these effects may be in
part through the ability of LIMD1 to regulate Snail function
through differential subcellular localization and thus, ultimately,
ECAD expression as well as the pRb/E2F cell cycle control
The authors thank Ms. Anne Webber and Mr. Daniel E. Foxler
for technical assistance.
1. Lerman MI, Minna JD. The 630-kb lung cancer homozygous deletion
region on human chromosome 3p21.3: identification and evaluation
of the resident candidate tumor suppressor genes. The international
lung cancer chromosome 3p21.3 tumor suppressor gene consortium.
Cancer Res 2000;60:6116–33.
Sekido Y, Fong KM, Minna JD. Molecular genetics of lung cancer.
Annu Rev Med 2003;54:73–87.
Zabarovsky ER, Lerman MI, Minna JD. Tumor suppressor genes on
chromosome 3p involved in the pathogenesis of lung and other can-
cers. Oncogene 2002;21:6915–35.
Minna JD, Fong K, Zochbauer-Muller S, Gazdar AF. Molecular
pathogenesis of lung cancer and potential translational applications.
Cancer J 2002;Suppl 1:S41–S46.
Wistuba II, Gazdar AF, Minna JD. Molecular genetics of small cell
lung carcinoma. Semin Oncol 2001;28(2, Suppl 4):3–13.
Yan PS, Shi H, Rahmatpanah F, Hsiau TH, Hsiau AH, Leu YW, Liu
JC, Huang TH. Differential distribution of DNA methylation within the
Miller BJ, Wang D, Krahe R, Wright FA. Pooled analysis of loss of
heterozygosity in breast cancer: a genome scan provides comparative
evidence for multiple tumor suppressors and identifies novel candi-
date regions. Am J Hum Genet 2003;73:748–67.
Maitra A, Wistuba II, Washington C, Virmani AK, Ashfaq R, Milch-
grub S, Gazdar AF, Minna JD. High-resolution chromosome 3p alle-
lotyping of breast carcinomas and precursor lesions demonstrates fre-
quent loss of heterozygosity and a discontinuous pattern of allele loss.
Am J Pathol 2001;159:119–30.
Yang Q, Yoshimura G, Mori I, Sakurai T, Kakudo K. Chromosome
3p and breast cancer. J Hum Genet 2002;47:453–9.
10. Manderson EN, Presneau N, Provencher D, Mes-Masson AM, Tonin
PN. Comparative analysis of loss of heterozygosity of specific chro-
mosome 3, 13, 17, and X loci and TP53 mutations in human epithelial
ovarian cancer. Mol Carcinog 2002;34:78–90.
11. Simsir A, Palacios D, Linehan WM, Merino MJ, Abati A. Detection
of loss of heterozygosity at chromosome 3p25–26 in primary and met-
astatic ovarian clear-cell carcinoma: utilization of microdissection
and polymerase chain reaction in archival tissues. Diagn Cytopathol
12. Goel A, Arnold CN, Niedzwiecki D, Chang DK, Ricciardiello L,
Carethers JM, Dowell JM, Wasserman L, Compton C, Mayer RJ, Ber-
tagnolli MM, Boland CR. Characterization of sporadic colon cancer
by patterns of genomic instability. Cancer Res 2003;63:1608–14.
13. Yashiro M, Carethers JM, Laghi L, Saito K, Slezak P, Jaramillo E,
Rubio C, Koizumi K, Hirakawa K, Boland CR. Genetic pathways in
the evolution of morphologically distinct colorectal neoplasms. Can-
cer Res 2001;61:2676–83.
14. Hesson LB, Cooper WN, Latif F. Evaluation of the 3p21.3 tumour-
suppressor gene cluster. Oncogene 2007;26:7283–301.
15. Angeloni D. Molecular analysis of deletions in human chromosome
3p21 and the role of resident cancer genes in disease. Brief Funct
Genomic Proteomic 2007;6:19–39.
16. Kost-Alimova M, Imreh S. Modeling non-random deletions in cancer.
Semin Cancer Biol 2007;17:19–30.
17. Ji L, Minna JD, Roth JA. 3p21.3 tumor suppressor cluster: prospects
for translational applications. Future Oncol 2005;1:79–92.
18. Petursdottir TE, Thorsteinsdottir U, Jonasson JG, Moller PH, Huiping
C, Bjornsson J, Egilsson V, Imreh S, Ingvarsson S. Interstitial dele-
tions including chromosome 3 common eliminated region 1
(C3CER1) prevail in human solid tumors from 10 different tissues.
Genes Chromosomes Cancer 2004;41:232–42.
19. Sharp TV, Munoz F, Bourboulia D, Presneau N, Darai E, Wang HW,
Cannon M, Butcher DN, Nicholson AG, Klein G, Imreh S, Boshoff C.
FIGURE 6 – Proposed model for correlation of LIMD1 expression
pattern in breast cancer with the indicated variables. Overall TMA
staining pattern (top 4 representative panels) have been summarised
into 4 distinct pictorial representations: (a) no nuclear/weak cyto-
plasmic; (b) moderate-heavy nuclear/weak to moderate cytoplasmic;
(c) strong nuclear/moderate cytoplasmic and (d) strong nuclear/strong
cytoplasmic. We propose that LIMD1 may represent a prognostic in-
dicator whereby strong/high expression of LIMD1 both in the cyto-
plasm and nucleus is indicative of low mortality rate and thus better
patient prognosis. Conversely negative to weak-nuclear expression of
LIMD1 together with weak cytoplasmic and cell–cell junction staining
of LIMD1 correlates with higher mortality rates and poor prognosis.
SPENDLOVE ET AL.
LIM domains-containing protein 1 (LIMD1), a tumor suppressor Download full-text
encoded at chromosome 3p21.3, binds pRB and represses E2F-driven
transcription. Proc Natl Acad Sci USA 2004;101:16531–6.
20. Ayyanathan K, Peng H, Hou Z, Fredericks WJ, Goyal RK, Langer
EM, Longmore GD, Rauscher FJ, III. The Ajuba LIM domain protein
is a corepressor for SNAG domain mediated repression and partici-
pates in nucleocytoplasmic Shuttling. Cancer Res 2007;67:9097–106.
21. Hou Z, Peng H, Ayyanathan K, Yan K, Langer EM, Longmore GD,
Rauscher FJ, III. Mol Cell Biol 2008;28:3198–207.
22. Langer EM, Feng Y, Zhaoyuan H, Rauscher FJ, III, Kroll KL, Long-
more GD. Ajuba LIM Proteins are snail/slug corepressors required for
neural crest development in Xenopus. Dev Cell 2008;14:424–36.
23. McShane LM, Altman DG, Sauerbrei W, Taube SE, Gion M, Clark
GM. REporting recommendations for tumor MARKer prognostic
studies (REMARK). Breast Cancer Res Treat 2006;100:229–35.
24. Abd El-Rehim DM, Pinder SE, Paish CE, Bell JA, Rampaul RS,
Blamey RW, Robertson JF, Nicholson RI, Ellis IO. Expression and
co-expression of the members of the epidermal growth factor receptor
(EGFR) family in invasive breast carcinoma. Br J Cancer 2004;
25. Abd El-Rehim DM, Ball G, Pinder SE, Rakha E, Paish C, Robertson
JF, Macmillan D, Blamey RW, Ellis IO. High-throughput protein
expression analysis using tissue microarray technology of a large
well-characterised series identifies biologically distinct classes of
breast cancer confirming recent cDNA expression analyses. Int J
26. Madjd Z, Pinder SE, Paish C, Ellis IO, Carmichael J, Durrant LG.
Loss of CD59 expression in breast tumours correlates with poor sur-
vival. J Pathol 2003;200:633–9.
27. Madjd Z, Durrant LG, Bradley R, Spendlove I, Ellis IO, Pinder SE.
Loss of CD55 is associated with aggressive breast tumors. Clin Can-
cer Res 2004;10:2797–803.
28. Putti TC, El-Rehim DM, Rakha EA, Paish CE, Lee AH, Pinder SE,
Ellis IO. Estrogen receptor-negative breast carcinomas: a review of
morphology and immunophenotypical analysis. Mod Pathol 2005;18:
29. Kononen J, Bubendorf L, Kallioniemi A, Barlund M, Schraml P,
Leighton S, Torhorst J, Mihatsch MJ, Sauter G, Kallioniemi OP. Tis-
sue microarrays for high-throughput molecular profiling of tumor
specimens. Nat Med 1998;4:844–7
30. Camp RL, Charette LA, Rimm DL. Validation of tissue microarray
technology in breast carcinoma. Lab Invest 2000;80:1943–9.
31. Bubendorf L, Nocito A, Moch H, Sauter G. Tissue microarray (TMA)
technology: miniaturized pathology archives for high-throughput in
situ studies. J Pathol 2001;195:72–9.
32. Torhorst J, Bucher C, Kononen J, Haas P, Zuber M, Kochli OR,
Mross F, Dieterich H, Moch H, Mihatsch M, Kallioniemi OP, Sauter
G. Tissue microarrays for rapid linking of molecular changes to clini-
cal endpoints. Am J Pathol 2001;159:2249–56.
33. Rakha EA, Abd El Rehim D, Pinder SE, Lewis SA, Ellis IO. E-cad-
herin expression in invasive non-lobular carcinoma of the breast and
its prognostic significance. Histopathology 2005;46:685–93.
34. Petit MM, Crombez KR, Vervenne HB, Weyns N, Van de Ven WJ.
The tumor suppressor Scrib selectively interacts with specific mem-
bers of the zyxin family of proteins. FEBS Lett 2005;579:5061–8.
35. Huggins CJ, Andrulis IL. Cell cycle regulated phosphorylation of
LIMD1 in cell lines and expression in human breast cancers. Cancer
36. Huggins CJ, Gill M, Andrulis IL. Identification of rare variants in the
hLIMD1 gene in breast cancer. Cancer Genet Cytogenet 2007;178:
37. Kadrmas JL, Beckerle MC. The LIM domain: from the cytoskeleton
to the nucleus. Nat Rev Mol Cell Biol 2004;5:920–31.
38. Petit MM, Meulemans SM, Alen P, Ayoubi TA, Jansen E, Van de
Ven WJ. The tumor suppressor Scrib interacts with the zyxin-related
protein LPP, which shuttles between cell adhesion sites and the nu-
cleus. BMC Cell Biol 2005;6:1.
39. Guo B, Sallis RE, Greenall A, Petit MM, Jansen E, Young L, Van de
Ven WJ, Sharrocks AD. The LIM domain protein LPP is a coactivator
for the ETS domain transcription factor PEA3. Mol Cell Biol
40. Fabbro M, Henderson BR. Regulation of tumor suppressors by nu-
clear-cytoplasmic shuttling. Exp Cell Res 2003;282:59–69.
41. Winters ZE, Hunt NC, Bradburn MJ, Royds JA, Turley H, Harris AL,
Norbury CJ. Subcellular localisation of cyclin B, Cdc2 and
p21(WAF1/CIP1) in breast cancer association with prognosis. Eur J
42. Taylor J, Lymboura M, Pace PE, A’hern RP, Desai AJ, Shousha S,
Coombes RC, Ali S. An important role for BRCA1 in breast cancer
progression is indicated by its loss in a large proportion of non-fami-
lial breast cancers. Int J Cancer 1998;79:334–42.
43. Jarvis EM, Kirk JA, Clarke CL. Loss of nuclear BRCA1 expression in
breast cancers is associated with a highly proliferative tumor pheno-
type. Cancer Genet Cytogenet 1998;101:109–15.
TUMOUR SUPPRESSOR PROTEIN LIMD1 IN BREAST CANCER