Aso 2005 04 010

Download as pdf or txt
Download as pdf or txt
You are on page 1of 9

Annals of Surgical Oncology, 12(6): 1)9

DOI: 10.1245/ASO.2005.04.010

Expression of the Transcription Factors Snail, Slug, and Twist


and Their Clinical Significance in Human Breast Cancer

Tracey A. Martin, PhD, Amit Goyal, MB ChB, Gareth Watkins, BSc, and Wen G. Jiang, MD

Metastasis & Angiogenesis Research Group, University Department of Surgery, Wales College of Medicine, Cardiff University,
Cardiff, CF14 4XN, Heath Park, United Kingdom

Background: Slug, Snail, and Twist are transcription factors that regulate the expression of
tumor suppressors such as E-cadherin. We examined the distribution and expression of these
three molecules together with the methylation of the Twist gene promoter in human breast
cancer to elucidate their clinical significance.
Methods: Frozen sections from breast cancer primary tumors (tumor, n = 114; back-
ground, n = 30) were immunostained with Slug, Snail, and Twist antibodies. RNA was
reverse-transcribed, quantified, and analyzed by quantitative polymerase chain reaction (Q-
PCR). Results were expressed as copy number of transcript per 50 ng of RNA (standardized
against b-actin).
Results: Immunohistochemistry revealed that all three molecules were stained in mammary
tissues, with an increase in Twist within tumor tissues; this was supported by Q-PCR analysis.
Q-PCR analysis showed that Slug was elevated with increasing tumor grade and prognostic
indices. Twist was elevated with increasing nodal involvement (tumor-node-metastasis status).
Conversely, Snail was reduced in expression corresponding with prognostic indices and tumor
grade. Increased levels of Slug were associated with tumors from patients with metastatic
disease or disease recurrence, and increased expression of Twist was associated with tumors
from patients who had died from breast cancer. It is interesting to note that Snail expression
was significantly reduced in patients with a poor outcome and those who had node-positive
tumors. In addition, tumors exhibited methylation of the Twist promoter.
Conclusions: These data demonstrate that all three transcription factors have inappropriate
expression in breast cancer and that this may play a part in the progression of human breast
tumors.
Key Words: Breast cancer—Twist—Snail—Slug.

Snail and Slug belong to the Snail family of genes, such as cell differentiation, cell motility, cell-cycle
which is conserved among species during evolution. regulation, and apoptosis.1,3,4 They share highly
They encode transcription factors that are expressed conserved C2H2-type zinc finger domains, which
at different stages of development in different tis- bind to the E-box and repress the transcription of
sues.1 These transcription factors encode a zinc fin- target genes.3 More recently, they have been impli-
ger–type transcription factor that is necessary for cated in the progression of human tumors via their
gastrulation and mesoderm formation2 and are in- regulatory action on the epithelial cell adhesion
volved in a broad spectrum of biological functions, molecule E-cadherin.5–7 Such transcriptional repres-
sion mechanisms have emerged as one of the crucial
Received April 8, 2004; accepted January 19, 2005; published processes for the downregulation of E-cadherin
online. expression during development and tumor progres-
Address correspondence and reprint requests to: Tracey A. sion with both Snail and Slug, acting through an
Martin, PhD; E-mail: martinta1@cf.ac.uk
interaction with proximal E-boxes of the E-cadherin
Published by Springer Science+Business Media, Inc.  2005 The Society of
Surgical Oncology, Inc. promoter.7

1
2 T. A. MARTIN ET AL.

Loss of expression of the E-cadherin cell/cell potential regulators of cell adhesion and migration,
adhesion molecule is important in carcinoma this study aimed to determine the levels of expression
development and progression.8 It has been shown of Snail, Slug, and Twist in human breast cancer
that Slug and Snail are potential repressors of E- tissues and to elucidate whether these levels are clin-
cadherin transcription in carcinomas that lack E- ically significant.
cadherin expression,6 and analysis of the expression
patterns of Slug, Snail, and E-cadherin in breast
cancer cell lines demonstrated that the expression of METHODS
Slug was strongly correlated with a loss of E-cadherin
transcripts. Slug is the more likely in vivo repressor of
Cell and Tissue Collection and Preparation
E-cadherin expression in breast cancer.
In vertebrates, the basic helix-loop-helix protein Human breast cancer cell lines (BT-549,
Twist may be involved in the negative regulation of BT-474KC, MDA-MB-231, MDA-MB-453, MDA-
cellular determination and in the differentiation of MB-231, MDA-MB-468, MDA-MB-436, MCF-7,
several lineages, including myogenesis, osteogenesis, MCF-10a, and ZR-751), putative melanoma cell line
and neurogenesis.9–11 Twist is able to inhibit onco- MDA-MB-435S, prostate cancer cell line PC-3, and
gene- and p53-dependent cell death and is thus HeLa and human fibroblast cell line MRC-5 (both
known as an antiapoptotic factor. Twist is also from ECACC, Wiltshire, UK) were maintained in
known to trigger epithelial-mesenchymal transition DulbeccoÕs modified EagleÕs medium with 10% fetal
(EMT) mechanisms, possibly regulating the E- to calf serum. Breast tissue samples23 (tumor and mat-
N-cadherin switch during EMT.12 Twist has been ched background; Table 1) were collected and
found to be inappropriately suppressed in 50% of immediately frozen in liquid nitrogen before pro-
rhabdomyosarcomas.10 Twist is also an activator of cessing—a portion of each sample for quantitative
GLI1 reporter expression, where defects in the sig- PCR (Q-PCR) analysis, a portion for immunohisto-
naling pathway lead to severe birth defects and can- chemical analysis, and a portion for routine histo-
cer formation in humans.13 Methylation-specific logical examination.
polymerase chain reaction (PCR) has been used to RNA was isolated from tissue samples and cells by
detect cancer cells from ductal lavage fluid for cyclin using standard RNAzol procedures. For reverse
D2, RAR-b, and Twist,14 which was frequently found transcription-PCR, complementary DNA (cDNA)
to be methylated in patients with carcinomas (17 of was synthesized in a 20-lL reaction mixture with 1 lg
20). Upregulation of Twist is associated with cellular of RNA, as described in the protocol (AB Gene Re-
resistance to paclitaxel but not with other drugs that verse Transcription System; ABGene, Surrey, UK).
have different mechanisms of action.15 This indicates
a novel mechanism that leads to resistance to
Quantitative PCR
microtubule-disrupting anticancer drugs through
upregulation of Twist. The Q-PCR system used the Amplofluor Unipri-
EMT results in the loss of polarity of epithelial cell mer system (Intergen Company, Oxford, UK) and
layers and cell/cell contacts, with a corresponding Thermo-Start (ABgene, Epsom, Surrey, UK), as we
remodeling of the cell cytoskeleton.16 Evidence has recently reported.16 Specific primer pairs for Snail,
accumulated showing that EMT is an important in Slug, and Twist (Table 2) were designed by the au-
vitro correlation of late-stage tumor progression.17–20 thors by using Beacon Designer software and were
In mouse mammary cells, it has been demonstrated manufactured by Invitrogen (Invitrogen Life Tech-
that EMT occurs before the acquirement of an nologies, Paisley, Scotland, UK), each amplifying a
invasive, metastatic tumor phenotype.21 Twist has region that spans at least one intron (primer details
been shown to play a role in metastasis in murine are given in supplement 1), thus generating an
models.22 Ectopic expression of Twist resulted in the approximately 100–base pair product from both the
loss of E-cadherin–mediated cell/cell adhesion, acti- control plasmid and cDNA.
vation of mesenchymal markers, and induction of cell By using the iCycler IQ system (Bio-Rad), which
motility; this suggests that Twist may contribute to incorporates a gradient thermocycler and a 96-chan-
metastasis by promoting EMT. Increased Twist nel optical unit, the plasmid standards and breast
expression has also been correlated with invasive cancer cDNA were simultaneously assayed in dupli-
lobular cancers associated with E-cadherin loss in cate reactions with a standard hot-start Q-PCR
human breast cancer.22 Because Snail and Slug are master mix. Q-PCR conditions were as follows:

Ann. Surg. Oncol. Vol. 12, No. 6, 2005


TRANSCRIPTION FACTOR EXPRESSION IN BREAST CANCER 3

TABLE 1. Clinical information of patient samples analyzed in the breast cancer samples were calculated. Q-PCR
Variable n for b-actin was also performed on the same samples
to correct for any residual differences in the initial
Tissue type
Background 33
level of RNA in the specimens (in addition to spec-
Tumor 124 trophotometry). The products of Q-PCR were veri-
Tumor grade fied on agarose gels (not shown).
1 24
2 42
3 58
NPI
Immunohistochemistry
1 68
2 38
Cryostat sections of frozen tissue were cut at 6 lm,
3 16 placed on Super Frost Plus (Optech Scientific In-
Unknown 2 struments, Oxon, UK) slides, air-dried, and fixed in a
Tumor-node-metastasis
1 70
50:50 solution of alcohol and acetone. The sections
2 40 were then air-dried again and stored at )20C until
3 7 use. Immediately before immunostaining began, the
4 4
Unknown 3
sections were washed in buffer for 5 minutes and
Histology treated with horse serum/buffer solution for 20 min-
Ductal 94 utes as a blocking agent to nonspecific binding. Sec-
Lobular 14
Other 16
tions were stained with Snail, Slug, and Twist
Outcome antibodies (Insight Biotechnology, Wembley, Cam-
Disease free 85 bridgeshire, UK). Primary antibodies were used at a
Metastatic disease 7
Local recurrence 5
1/50 dilution for 60 minutes and then washed in
Death 15 buffer. The secondary biotinylated antibody at 1/100
Poor outcomesa 27 dilution (universal secondary, Vectastain Elite ABC;
NPI, Nottingham Prognostic Index. Vector Laboratories Inc., Burlingham, CA) was ad-
a
Metastatic disease, local recurrence, or death. ded (in horse serum/buffer solution) for 30 minutes,
followed by numerous washings in buffer. Avidin/
biotin complex was added for 30 minutes, again fol-
TABLE 2. Quantitative polymerase chain reaction primer
pairs for analysis of angiogenic marker molecules lowed with washes in buffer. Diaminobenzidine was
used as a chromogen to visualize the antibody/anti-
Molecule Primer sequence gen complex. Sections were counterstained in MayerÕs
Slug hematoxylin for 1 minute, dehydrated, cleared,
SLUG1F ctccaaaaagccaaactaca mounted in dextropropoxyphene, and screened with
SLUG1ZR actgaacctgaccgtacagaggatctctggttgtggta
Snail an ·25 objective.
SNAILF1 tctttcctcgtcaggaagc
SNAILZR actgaacctgaccgtacactgctggaaggtaaactctg
Twist Evaluation of Twist Promoter Hypermethylation
TWISTF aagctgagcaagattcagac
TWISTZR actgaacctgaccgtacagaggacctggtagaggaagt Genomic DNA from multiple sections from frozen
b-Actin tissue and human breast cancer cells were extracted
b-Actin F atgatatcgccgcgctcgt
b-Actin R cgctcggtgaggatcttca by using a standard DNA-extraction protocol.
Twist methylated Methylation-specific PCR was performed as previ-
TwistMF tttcggatggggttgttatc ously described.17 One microgram of DNA was
TwistMR aaacgacctaacccgaagg
Twist unmethylated denatured in NaOH (to a final concentration of .2
TwistUF tttggatggggttgttattgt mol/L) for 10 minutes at 37C, followed by 10 mmol/
TwistUR cctaacccaaacaaccaacc L of freshly prepared hydroquinone and 3 mol/L of
F, forward; R, reverse primer. sodium bisulfite (pH 5.0). After incubation at 50C
for 16 hours, the DNA was purified with the Wizard
enzyme activation at 95C for 12 minutes for 1 cycle, DNA purification kit (Promega UK Ltd., South-
followed by 60 cycles of denaturing at 95C for 15 ampton, UK) and standard ethanol precipitation.
seconds, annealing at 55C for 40 seconds, and Hypermethylation was analyzed by PCR with prim-
extension at 72C for 25 seconds. With use of purified ers listed in Table 2. Methylation-specific PCR con-
plasmids as internal standards, the levels of each tight ditions were as follows: enzyme activation at 95C for
junction molecule cDNA (copies per 50 ng of RNA) 5 minutes for 1 cycle, followed by 36 cycles of

Ann. Surg. Oncol. Vol. 12, No. 6, 2005


4 T. A. MARTIN ET AL.

denaturing at 95C for 15 seconds, annealing at 58C


for 15 seconds, and extension at 72C for 30 seconds.
The PCR products were separated on standard 10%
polyacrylamide gel electrophoresis gels.

Statistical Analysis
Statistical analysis was performed with Minitab
version 13.32 (Minitab Inc., State College, PA) by
using a two-sample StudentÕs t-test and the non-
parametric Mann-Whitney confidence interval and
test, where appropriate.

RESULTS

Slug, Snail, and Twist Show Increased Expression in


Tumor Tissues
Levels of transcripts of Slug, Snail, and Twist were
increased in tumor samples in comparison to back-
ground (expressed as transcript copy number per 50 lg
of messenger RNA and standardized with b-actin;
Fig. 1). Transcript copy numbers for Slug were
.54 ± .19 for tumor and .29 ± .11 for background
(P = .25); for Snail, they were .18 ± .07 for tumor and
.05 ± .02 for background (P = .09); and for Twist, FIG. 1. Levels of transcripts of Slug, Snail, and Twist in tumor
samples in comparison to background (expressed as transcript copy
they were 68.1 ± 20.9 for tumor and 52.4 ± 18.8 for number per 50 lg of messenger RNA and standardized with b-
background (P = .58), although these data did not actin).
reach significance. This pattern agreed with the results
from immunohistochemistry: in tumor sections, more
cells stained positively for all three regulatory factors tive factors (Tables 3–5). Slug was sequentially in-
(Fig. 2). From the sections themselves (Fig. 2A and creased with increasing Nottingham Prognostic Index
2B), it is observable that in normal background tissue, (NPI) status from .51 ± .55 in NPI 1 (NPI <3.5)
the cells surrounding the vessels and ducts stain very tumors to .84 ± .59 in NPI 3 tumors (NPI >5.4;
intensely for these transcription regulators. In the Table 3). Twist was increased in NPI 2 (NPI 3.5–5.4)
tumor tissues, although the intensity of staining is less, tumors (106.8 ± 54.4 vs. 51.5 ± 20.7 in NPI 1) but
most cells stain positively (Fig. 2). showed little change in NPI 3 tumors (32.5 ± 18.2).
It is interesting to note that both Slug and Twist Conversely, Snail expression was reduced with
had increased expression in node-positive tumors increasing NPI status (NPI 1, .25 ± .13; NPI 2,
compared with node-negative tumors (slug: node .13 ± .08; NPI 3, .03 ± .03); however, these changes
positive, .63 ± .36; node negative, .51 ± .20; P = did not reach significance (P = .2).
.78; Twist: node positive, 89 ± 21; node negative, When expression was compared with tumor grade
52 ± 40; P = .40). Snail was reduced in node-posi- (Table 4), a similar change was observed for Slug
tive tumors (node positive, .10 ± .06; node negative, —i.e., expression increased with increasing tumor
.25 ± .13; P = .32), although these values were not grade (grade 1, .39 ± 019; grade 3, .57 ± .24), and
significantly different (Fig. 3). Twist again showed no obvious correlation (Table 4).
Snail once again showed reduced expression with
increasing tumor grade (grade 1, .38 ± .25; grade 2,
Examination of Predictive Factors Shows That Snail
.18 ± .14; grade 3, .09 ± .05), although this did not
Has a Different Expression From Slug and Twist
reach significance (P = .3).
All three transcription molecules were assessed by Twist was sequentially increased with increasing
using Q-PCR and compared with a series of predic- tumor-node-metastasis status of tumors (tumor-node-

Ann. Surg. Oncol. Vol. 12, No. 6, 2005


TRANSCRIPTION FACTOR EXPRESSION IN BREAST CANCER 5

FIG. 2. Immunostaining of Slug (A), Snail (B), and Twist (C) in background (left panel) and tumor (right panel) sections. Representative cases
of increased positive staining in tumor sections are shown at ·4, ·10, and ·20.

TABLE 3. Analyses of messenger RNA from samples


showing levels of the three transcription factors tested and
NPI status (transcript copy number)
Transcription factor
NPI status Slug Snail Twist
1 .51 ± .55 .25 ± .13 51.5 ± 20.7
2 .56 ± .45 .13 ± .08 106.8 ± 54.4
3 .84 ± .59 .03 ± .03 42.5 ± 18.2

NPI, Nottingham Prognostic Index.

TABLE 4. Analyses of messenger RNA from samples


showing levels of the three transcription factors tested and
tumor grade (transcript copy number)
Transcription factor
Grade Slug Snail Twist
1 .39 ± .19 .38 ± .25 70.4 ± 30.3
2 .60 ± .46 .18 ± .14 45.3 ± 16.0
3 .57 ± .24 .09 ± .06 82.2 ± 40.4

TABLE 5. Analyses of messenger RNA from samples


showing levels of the three transcription factors tested and
tumor-node-metastasis (TNM) status
(transcript copy number)
Transcription factor
FIG. 3. Levels of transcripts of Slug, Snail, and Twist expression in
node-positive (Node +ive) tumors compared with node-negative TNM status Slug Snail Twist
(Node )ive) tumors. 1 .77 ± .33 .17 ± .08 56.0 ± 14.3
2 .11 ± .03 .29 ± .19 81.1 ± 54.3
3 .23 ± .19 .05 ± .03 197.0 ± 165.0

metastasis 1, 56.0 ± 14.3; tumor-node-metastasis 2,


81.1 ± 54.3; tumor-node-metastasis 3, 197.0 ±
165.0). Neither Slug nor Snail showed a correlation increased nodal involvement only, whereas Slug was
with tumor-node-metastasis status (Table 5). It is positively associated with NPI status and grade of
interesting that Twist was progressively increased with tumor and Snail was negatively associated.

Ann. Surg. Oncol. Vol. 12, No. 6, 2005


6 T. A. MARTIN ET AL.

TABLE 6. Analysis of messenger RNA to investigate cor-


relations between survival status and transcription factor
(transcript copy number)
Transcription factor
Survival status Slug Snail Twist
Disease free .31 ± .12 .24 ± .11 57.2 ± 17.0
Metastatic disease 1.00 ± .80 .18 ± .10 37.9 ± 15.2
Local recurrence .37 ± .37 .01 ± .007 57.1 ± 48.5
Death from 1.91 ± 1.29 .26 ± .02 169.0 ± 138.0
breast cancer

Correlation of Snail, Slug, and Twist With Patient


Prognosis
Our patient follow-up of 72.2 months allowed us to
examine the Q-PCR data with regard to patient
outcome. Patients were classified as (1) disease free,
(2) with metastatic disease, (3) with local recurrence
of breast cancer, and (4) dead as a result of breast
cancer. Slug showed the lowest expression in patients
who had remained disease free (.31 ± .12) and was
highest in patients who had died of breast cancer
(1.91 ± 1.29; P = .2). Slug was also increased in
patients with metastatic disease (1.00 ± .8), but there
FIG. 4. Comparison of transcript levels of Slug, Snail, and Twist
was little change of expression in patients with local considering poor-outcome patients compared with those who re-
recurrences (Table 6). Twist showed the highest mained disease free. Snail was significantly reduced in patients with
expression in patients who had died of breast cancer poor outcome compared with those who remained alive and well
after 72.2 months (P = .05).
(169.0 ± 138.0). Snail was reduced in patients who
had metastatic disease (.18 ± .10) and was signifi-
cantly reduced in those with local recurrences
(recurrence, .01 ± .007; alive and well, .24 ± .11; P
= .04) but was increased in patients who had died of
breast cancer (.26 ± .02; P = .05). When all poor
outcomes are considered together, it can be observed
that both Slug and Twist were increased in patients
with poor outcome (metastatic disease and recurrence
of and death from breast cancer), whereas Snail was
significantly reduced in patients with a poor outcome
compared with those who remained alive and well
after 72.2 months (Fig. 4).
FIG. 5. Agarose gel indicating hypermethylation of the Twist
promoter in representative tumors (3 of 8) tested (A). Of the 10
breast cancer cell lines analyzed, only 3 showed Twist promoter
Twist Shows Hypermethylation in Human Breast methylation (B). N, normal; T, tumor.
Cancer
Hypermethylation of the Twist promoter was
observed in 50% of the tumors (four of eight) tested showed only a weak signal (Fig. 5B). It is interesting to
(Fig. 5A), whereas only one background tissue showed note that control cells (such as human endothelial cells
methylation (one of eight). Of the 10 breast cancer cell from vein), the melanoma cell line MDA-MB-435S,
lines analyzed, only 3 showed Twist promoter meth- and HeLa cells also showed promoter methylation of
ylation: MCF-7 cells showed a strong signal for Twist. The fibroblast cell line MRC-5 was not methy-
methylation, and ZR-751 and MDA-MB-231 cells lated.

Ann. Surg. Oncol. Vol. 12, No. 6, 2005


TRANSCRIPTION FACTOR EXPRESSION IN BREAST CANCER 7

DISCUSSION cancer tissue, it has been suggested that Snail may


have a cancer-protective role in healthy breast tis-
In this article, we have demonstrated that the reg- sue.33 Together with our findings shown here, the
ulatory proteins Slug, Snail, and Twist are expressed reduced expression of Snail in breast cancer may have
aberrantly in human breast cancer. Expression of a direct bearing on the use of aromatase suppressors
Slug and Twist shows increased levels in tumors, in breast cancer therapy.34
whereas Snail seems reduced. Slug bears a close homology to the Caenorhabditis
The members of the Snail family have been impli- elegans CES-1 protein35 and is thought to be an evo-
cated in the triggering of EMT during embryonic lutionarily conserved transcriptional repressor whose
development,24,25 and EMT induction by Snail in activation promotes the aberrant survival and even-
epithelial cells is mediated by direct transcriptional tual malignant transformation of mammalian pro-B
repression of the cell adhesion molecule E-cadherin. cells otherwise scheduled for apoptosis. It been shown
Indeed, Snail has been shown to repress E-cadherin to be a transcriptional repressor in humans.3 Both
expression and to trigger EMT associated with epi- Snail and Slug have been shown to repress E-cadherin
thelial tumor progression.23–26 Snail expression has expression during development and tumor progres-
been demonstrated in primary human tumors,27,28 sion through their interaction with proximal E-boxes
with a correlation between Snail expression and a of the E-cadherin promoter.7 Slug is also thought to
reduction or absence of E-cadherin expression in be the likely in vivo repressor of E-cadherin in breast
several tumors. Snail expression inversely correlates carcinoma.6 This is in contrast to the relationship
with the grade of differentiation of tumors and is between Slug and Snail expression in human hepato-
expressed in infiltrating ductal carcinomas with cellular carcinoma, in which Snail is overexpressed
lymph node metastases; it has been concluded that and is the likely E-cadherin repressor,5 as it is thought
Snail is involved in the progression of breast ductal to be in malignant melanomas.36
tumors and possibly serves as a marker of metastatic In our study, we have shown that Slug expression is
potential.29,30 We have previously reported that there increased in tumors compared with normal back-
was no correlation among E-cadherin Slug, Snail, or ground tissues. Slug is also increased progressively
Twist in this dataset.31 with NPI status and tumor grade, although some
Our results show that Snail is expressed by more reduction in expression was observed corresponding
cells in tumors (assessed by immunohistochemistry), with tumor-node-metastasis status. Slug expression
although the intensity of staining is reduced in stro- was increased in node-positive tumors and was high
mal and epithelial/tumor cells compared with cells in in tumors from patients who had died from breast
microvessels. This is supported by Q-PCR analysis cancer or had metastatic disease. Moreover, Slug
and by previous work showing that Snail expression expression was reduced in lobular tumors compared
is higher in normal cells compared with breast cancer with ductal tumors or tumors of other types. These
cell lines.32 However, Snail expression was reduced results differ from those observed for Snail, even
with both NPI status and grade of tumor and was though both are transcription factors from the same
significantly reduced in patients with poor outcomes, family (Snail is upstream of Slug).37 However, evi-
particularly those with local recurrences. Further, dence is accumulating to support the idea that the in
Snail expression was reduced in node-positive tumors vivo action of different factors such as Snail and Slug,
and in both ductal and lobular carcinomas. This in E-cadherin repression, for example, can be mod-
somewhat contradicts previous findings28 that did not ulated by their relative concentrations, as well as by
use Q-PCR and had a small sample size (n = 21) and specific cellular or tumoral contexts.7 These repres-
so concentrated on the distribution of Snail expres- sion factors may act alone or in concert6 and may
sion. Moreover, it has been demonstrated6 that Snail have differing functions that are dependent on tumor
expression does not correlate as well with loss of type.
E-cadherin in breast cancer as does the other Snail Twist is able to inhibit oncogene- and p53-depen-
family member, Slug. Snail is, however, a repressor dent cell death, interferes with p53 reporter activa-
that downregulates the expression of aromatase tion, and impairs the induction of genes subsequent
(oestrogen synthetase) in healthy breast tissue via to that in response to DNA damage.10 It is thus
suppression of the I.3 promoter.30 Aromatase is up- known as an antiapoptotic factor. Twist is also
regulated in breast tumors and stimulates cancer known to trigger EMT mechanisms and possibly
growth in both an autocrine and a paracrine man- regulates the E- to N- cadherin switch during EMT.12
ner.31,32 Because Snail expression is reduced in breast Forced N-cadherin expression exerts a dominant ef-

Ann. Surg. Oncol. Vol. 12, No. 6, 2005


8 T. A. MARTIN ET AL.

fect over E-cadherin function in breast cancer cells.38 through upregulation of Twist expression in human
N-cadherin expression in normal epithelial cells cau- cancer.
ses a reduction in E-cadherin expression, and N- In conclusion, the EMT regulatory proteins Slug
cadherin is also able to enhance tumor cell motility and Twist are upregulated in human breast cancer,
and migration.39 Twist has been found to be inap- whereas Snail is downregulated. Such disparate
propriately suppressed in 50% of rhabdomyosarco- expression levels may contribute to the progression of
mas.10 Twist is also an activator of GLI1 reporter tumors in breast cancer, and this deserves further
expression through E-box interaction via the binding investigation.
of upstream stimulatory factors (USF) proteins.
GLI1 encodes a critical transcription activator in the
sonic hedgehog signaling pathway, which regulates ACKNOWLEDGMENTS
vertebrate patterning, where defects lead to severe We thank Cancer Research UK for supporting this
birth defects and cancer formation in humans.13 work and Dr. Anthony Douglas-Jones for his
Here we demonstrated that Twist was increased in invaluable advice on histology.
tumor tissues (both in intensity of staining with
immunohistochemistry and in Q-PCR analysis) and
showed a progressive increase with tumor-node-
REFERENCES
metastasis status and in patients who had died from
breast cancer. Twist was also increased in ductal and 1. Seki K, Fujimori T, Savagner P, et al. Mouse Snail family
lobular cancer compared with other types and also in transcription repressors regulate chondrocyte, extracellular
node-positive tumors. Twist also showed some hy- matrix, type II collagen, and aggrecan. Biol Chem
2003;278:41862–70.
permethylation of its promoter. Little work has been 2. Boulay JL, Dennefeld C, Alberga A. The Drosophila devel-
performed investigating the role of Twist in human opmental gene snail encodes a protein with nucleic acid bind-
cancer. Both Twist and Snail have been investigated ing fingers. Nature 1987;330:395–8.
3. Hemavathy K, Ashraf SI, Ip YT. Snail/slug family of
in human gastric cancer: Snail expression was in- repressors: slowly going into the fast lane of development and
creased and was associated with decreased E-cadh- cancer. Gene 2000;257:1–12.
erin expression.40 Moreover, Twist overexpression 4. Arias AM. Epithelial mesenchymal interactions in cancer and
development. Cell 2001;105:425–31.
was also seen and correlated with abnormally positive 5. Sugimachi K, Tanaka S, Kameyama T, et al. Transcriptional
or increased N-cadherin expression in gastric carci- repressor snail and progression of human hepatocellular car-
nomas. The epigenetic silencing of a small number of cinoma. Clin Cancer Res 2003;9:2657–64.
6. Hajra KM, Chen DY, Fearon ER. The SLUG zinc-finger
genes by promoter hypermethylation in lobular protein represses E-cadherin in breast cancer. Cancer Res
breast cancers has been studied.41 It has been shown 2002;62:1613–8.
that Twist was hypermethylated less often in invasive 7. Bolos V, Peinado H, Perez-Moreno MA, Fraga MF, Esteller
M, Cano A. The transcription factor Slug represses E-cadherin
lobular cancers than in invasive ductal cancers. Our expression and induces epithelial to mesenchymal transitions: a
findings show that the Twist promoter was hyper- comparison with Snail and E47 repressors. J Cell Sci
methylated in only three of eight breast cancers and 2003;116:499–511.
8. Jiang WG. E-cadherin and its associated protein catenins,
that this occurred only in ductal carcinomas. This cancer invasion and metastasis. Br J Surg 1996;83:437–46.
may suggest that promoter methylation is patient 9. Hamamori Y, Wu HY, Sartorelli V, Kedes L. The basic do-
specific and is not indicative of prognosis, as antici- main of myogenic basic helix-loop-helix (bHLH) proteins is the
novel target for direct inhibition by another bHLH protein,
pated. In mouse mammary tumors, Twist expression Twist. Mol Cell Biol 1997;17:6563–73.
can be regulated by Wnt/b-catenin signaling, and 10. Maestro R, Dei Tos AP, Hamamori Y, et al. Twist is a po-
both Twist and Wnt1 can function as inhibitors of tential oncogene that inhibits apoptosis. Genes Dev
1999;13:2207–17.
lactogenic differentiation, which could contribute to 11. Castanon I, Von Stetina S, Kass J, Baylies MK. Dimerization
mammary tumorigenesis.42 Our data suggest that partners determine the activity of the Twist bHLH protein
Twist may have a considerable effect in the ultimate during Drosophila mesoderm development. Development
2001;128:3145–59.
increase in tumor/nodal involvement in breast cancer 12. Oda H, Tsukita S, Takeichi M. Dynamic behavior of the
and, as such, could have potential as a marker in cadherin-based cell-cell adhesion system during Drosophila
human breast cancer. Upregulation of Twist has been gastrulation. Dev Biol 1998;203:435–50.
13. Villavicencio EH, Yoon JW, Frank DJ, Fuchtbauer EM,
shown to be associated with cellular resistance to Walterhouse DO, Iannaccone PM. Cooperative E-box regu-
paclitaxel and vincristine in four types of human lation of human GLI1 by TWIST and USF. Genesis
cancer (nasopharyngeal, bladder, ovarian, and pros- 2002;32:247–58.
14. Evron E, Dooley WC, Umbricht CB, et al. Detection of breast
tate),15 thus suggesting that a therapeutic strategy cancer cells in ductal lavage fluid by methylation-specific PCR.
may be developed to overcome acquired resistance Lancet 2001;357:1335–6.

Ann. Surg. Oncol. Vol. 12, No. 6, 2005


TRANSCRIPTION FACTOR EXPRESSION IN BREAST CANCER 9

15. Wang X, Ling MT, Guan XY, et al. Identification of a novel 30. Blanco MJ, Moreno-Bueno G, Sarrio D, et al. Correlation of
function of TWIST, a bHLH protein, in the development of Snail expression with histological grade and lymph node status
acquired taxol resistance in human cancer cells. Oncogene in breast carcinomas. Oncogene 2002;21:3241–6.
2004;23:474–82. 31. Goyal A, Douglas-Jones A, Mansel RE, Jiang WG. E-cadh-
16. Thiery JP. Epithelial-mesenchymal transitions in tumour erin and alpha-catenin expression in relation to clinical out-
progression. Nat Rev Genet 2002;2:442–54. comes in patients with breast cancer. Presented to the 26th San
17. Boyer B, Valles AM, Edme N. Induction and regulation of Antonio Breast Cancer Symposium. Breast Cancer Res Treat
epithelial-mesenchymal transitions. Biochem Pharmacol 2003;82:517.
2000;60:1091–9. 32. Okubo T, Truong TK, Yu B, et al. Down-regulation of promoter
18. Jechlinger M, Grunert S, Tamir IH, et al. Expression profiling 1.3 activity of the human aromatase gene in breast tissue by zinc-
of epithelial plasticity in tumor progression. Oncogene finger protein, snail (SnaH). Cancer Res 2001;61:1338–46.
2003;22:7155–69. 33. Chen S, Itoh T, Wu K, Zhou D, Yang C. Transcriptional
19. Thiery JP. Epithelial-mesenchymal transitions in development regulation of aromatase expression in human breast tissue.
and pathologies. Curr Opin Cell Biol 2003;15:740–6. J Steroid Biochem Mol Biol 2002;83:93–9.
20. Huber MA, Azoitei N, Baumann B, et al. NF-jB is essential for 34. Chen S, Zhou D, Okubo T, et al. Prevention and treatment of
epithelial-mesenchymal transition and metastasis in a model of breast cancer by suppressing aromatase activity and expres-
breast cancer progression. J Clin Invest 2004;114:569–81. sion. Ann N Y Acad Sci 2002;963:229–38.
21. Kang Y, Massague J. Epithelial-mesenchymal transitions: 35. Inukai T, Inoue A, Kurosawa H, et al. SLUG, a ces-1-related
twist in development and metastasis. Cell 2004;118:277–9. zinc finger transcription factor gene with antiapoptotic activity,
22. Yang J, Mani SA, Donahar JL, et al. Twist, a master regulator is a downstream target of the E2A-HLF oncoprotein. Mol Cell
of morphogenesis, plays an essential role in tumor metastasis. 1999;4:343–52.
Cell 2004;118:927–39. 36. Poser I, Dominguez D, de Herreros AG, Varnai A, Buettner R,
23. Jiang WG, Watkins G, Lane J, et al. Prognostic value of rho Bosserhoff AK. Loss of E-cadherin expression in melanoma
GTPases and rho guanine nucleotide dissociation inhibitors in cells involves up-regulation of the transcriptional repressor
human breast cancers. Clin Cancer Res 2003;9:6432–40. Snail. J Biol Chem 2001;276:24661–6.
24. Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB. 37. Aybar MJ, Nieto MA, Mayor R. Snail precedes slug in the
Methylation-specific PCR: a novel PCR assay for methylation genetic cascade required for the specification and migration of
status of CpG islands. Proc Natl Acad Sci U S A 1996;93:9821–6. the Xenopus neural crest. Development 2003;130:483–94.
25. Nieto MA, Sargent MG, Wilkinson DG, Cooke J. Control of 38. Nieman MT, Prudoff RS, Johnson KR, Wheelock MJ.
cell behavior during vertebrate development by Slug, a zinc N-cadherin promotes motility in human breast cancer cells
finger gene. Science 1994;264:835–9. regardless of their E-cadherin expression. J Cell Biol
26. Cano A, Perez-Moreno MA, Rodrigo I, et al. The tran- 1999;147:631–44.
scription factor snail controls epithelial-mesenchymal transi- 39. Li G, Satyamoorthy K, Herlyn M. N-cadherin-mediated
tions by repressing E-cadherin expression. Nat Cell Biol 2000; intercellular interactions promote survival and migration of
2:76–83. melanoma cells. Cancer Res 2001;61:3819–25.
27. Carver EA, Jiang R, Lan Y, Oram KF, Gridley T. The mouse 40. Rosivatz E, Becker I, Specht K, et al. Differential expression of
snail gene encodes a key regulator of the epithelial-mesenchy- the epithelial-mesenchymal transition regulators snail, SIP1,
mal transition. Mol Cell Biol 2001;21:8184–8. and twist in gastric cancer. Am J Pathol 2002;161:1881–91.
28. Batlle E, Sancho E, Franci C, et al. The transcription factor 41. Fackler MJ, McVeigh M, Evron E, et al. DNA methylation of
snail is a repressor of E-cadherin gene expression in epithelial RASSF1A, HIN-1, RAR-beta, cyclin D2 and Twist in in situ
tumor cells. Nat Cell Biol 2000;2:84–9. and invasive lobular breast carcinoma. Int J Cancer
29. Cheng CW, Wu PE, Yu JC, et al. Mechanisms of inactivation 2003;107:970–5.
of E-cadherin in breast carcinoma: modification of the two-hit 42. Howe LR, Watanabe O, Leonard J, Brown AM. Twist is up-
hypothesis of tumor suppressor gene. Oncogene 2001;20:3814– regulated in response to Wnt1 and inhibits mouse mammary
23. cell differentiation. Cancer Res 2003;63:1906–13.

Ann. Surg. Oncol. Vol. 12, No. 6, 2005

You might also like