Article

Gene expression in extratumoral microenvironment predicts clinical outcome in breast cancer patients

Department of Epidemiology, University of North Carolina at Chapel Hill, Campus Box 7435, Chapel Hill, NC 27599, USA. .
Breast cancer research: BCR (Impact Factor: 5.49). 03/2012; 14(2):R51. DOI: 10.1186/bcr3152
Source: PubMed
ABSTRACT
A gene expression signature indicative of activated wound responses is common to more than 90% of non-neoplastic tissues adjacent to breast cancer, but these tissues also exhibit substantial heterogeneity. We hypothesized that gene expression subtypes of breast cancer microenvironment can be defined and that these microenvironment subtypes have clinical relevance.
Gene expression was evaluated in 72 patient-derived breast tissue samples adjacent to invasive breast cancer or ductal carcinoma in situ. Unsupervised clustering identified two distinct gene expression subgroups that differed in expression of genes involved in activation of fibrosis, cellular movement, cell adhesion and cell-cell contact. We evaluated the prognostic relevance of extratumoral subtype (comparing the Active group, defined by high expression of fibrosis and cellular movement genes, to the Inactive group, defined by high expression of claudins and other cellular adhesion and cell-cell contact genes) using clinical data. To establish the biological characteristics of these subtypes, gene expression profiles were compared against published and novel tumor and tumor stroma-derived signatures (Twist-related protein 1 (TWIST1) overexpression, transforming growth factor beta (TGF-β)-induced fibroblast activation, breast fibrosis, claudin-low tumor subtype and estrogen response). Histological and immunohistochemical analyses of tissues representing each microenvironment subtype were performed to evaluate protein expression and compositional differences between microenvironment subtypes.
Extratumoral Active versus Inactive subtypes were not significantly associated with overall survival among all patients (hazard ratio (HR) = 1.4, 95% CI 0.6 to 2.8, P = 0.337), but there was a strong association with overall survival among estrogen receptor (ER) positive patients (HR = 2.5, 95% CI 0.9 to 6.7, P = 0.062) and hormone-treated patients (HR = 2.6, 95% CI 1.0 to 7.0, P = 0.045). The Active subtype of breast microenvironment is correlated with TWIST-overexpression signatures and shares features of claudin-low breast cancers. The Active subtype was also associated with expression of TGF-β induced fibroblast activation signatures, but there was no significant association between Active/Inactive microenvironment and desmoid type fibrosis or estrogen response gene expression signatures. Consistent with the RNA expression profiles, Active cancer-adjacent tissues exhibited higher density of TWIST nuclear staining, predominantly in epithelium, and no evidence of increased fibrosis.
These results document the presence of two distinct subtypes of microenvironment, with Active versus Inactive cancer-adjacent extratumoral microenvironment influencing the aggressiveness and outcome of ER-positive human breast cancers.

Full-text

Available from: Patricia Casbas-Hernandez, Jul 16, 2014
RESEARCH ARTICLE Open Access
Gene expression in extratumoral
microenvironment predicts clinical outcome in
breast cancer patients
Erick Román-Pérez
1
, Patricia Casbas-Hernández
2
, Jason R Pirone
3
, Jessica Rein
3
, Lisa A Carey
3
, Ronald A Lubet
4
,
Sendurai A Mani
5,6
, Keith D Amos
7,3
and Melissa A Troester
1,3*
Abstract
Introduction: A gene expression signature indicative of activated wound responses is comm on to more than 90%
of non-neoplastic tissues adjacent to breast cancer, but these tissues also exhibit substantial heterogeneity. We
hypothesized that gene expression subtypes of breast cancer microenvironment can be defined and that these
microenvironment subtypes have clinical relevance.
Methods: Gene expression was evaluated in 72 patient-derived breast tissue samples adjacent to invasive breast
cancer or ductal carcinoma in situ. Unsupervised clustering identified two distinct gene expression subgroups that
differed in expression of genes involved in activation of fibrosis, cellular movement, cell adhesion and cell-cell
contact. We evaluated the prognostic relevance of extratumoral subtype (comparing the Active group, defined by
high expression of fibrosis and cellular movement genes, to the Inactive group, defined by high expression of
claudins and other cellular adhesion and cell-cell contact genes) using clinical data. To establish the biological
characteristics of these subtypes, gene expression profiles were compared against published and novel tumor and
tumor stroma-derived signatures (Twist-related protein 1 (TWIST1) overexpression, transforming growth factor beta
(TGF-b)-induced fibroblast activation, breast fibrosis, claudin-low tumor subtype and estrogen resp onse).
Histological and immunohistochemical analyses of tissues representing each microenvironment subtype were
performed to evaluate protein expression and compositional differences between microenvironment subtyp es.
Results: Extratumoral Active versus Inact ive subtypes were not significantly associated with overall sur vival among
all patients (hazard ratio (HR) = 1.4, 95% CI 0.6 to 2.8, P = 0.337), but there was a strong association with overall
survival among estrogen receptor (ER) positive patients (HR = 2.5, 95% CI 0.9 to 6.7, P = 0.062) and hormone-
treated patients (HR = 2.6, 95% CI 1.0 to 7.0, P = 0.045). The Active subtype of breast microenvironment is
correlated with TWIST-overexpression signatures and shares features of claudin-low breast cancers. The Active
subtype was also associated with expression of TGF-b induced fibroblast activation signatures, but there was no
significant association between Active/Inactive microenvironment and desmoid type fibrosis or estrogen response
gene expr ession signatures. Consistent with the RNA expression profiles, Active cancer-adjacent tissues exhibited
higher density of TWIST nuclear staining, predominantly in epithelium, and no evidence of increased fibrosis.
Conclusions: These results document the presence of two distinct subtypes of microenvironment, with Active
versus In active cancer-adjacent extratumoral microenvironment influencing the aggressiveness and outcome of ER-
positive human breast cancers.
* Correspondence: troester@unc.edu
1
Department of Epidemiology, University of North Carolina at Chapel Hill,
Campus Box 7435, Chapel Hill, NC 27599, USA
Full list of author information is available at the end of the article
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
© 2012 Román-Pérez et al.; licensee BioMed Central Ltd. This is an open acce ss article distributed under the terms of the Creative
Commons Attribution License (http://creative commons.org/licenses/b y/2.0), which permits unrestricte d use, distribution, and
reproduction in any mediu m, provided the original work is properly cited.
Page 1
Introduction
Gene expression analysis of tissue adjacent to invasive
breast cancer and ductal carcinoma in situ has suggested
that intratumoral stromal responses contribute to dis-
ease progression. Finak et al. [1]showedthatelevated
expression of stroma-derived immune mediators in
tumor tissue predicted relapse. Chang et al. reported a
signature of fibroblast response [2] and Beck et al.
reported fibromatosis and macrophage-associated signa-
tures, each with prognostic value [3,4]. Stromal
responses are activated at early stages in carcinogenesis,
even in the absence of invasio n [5], leading to specula-
tion that for acquisition of the invasive phenotype, the
stroma is dominant over the epithelium [6]. We
recently reported an in vivo wound response signature
derived from tissue adjacent to breast cancer, which
when expressed in tumors, predicts relapse and overall
survival [7]. The vast majority of studies evaluating
stroma-derived signatures [1-5,8-11] have focused on
intratumoral stromal expression rather than extratu-
moral expression.
Growing evidence suggests that extratumoral microen-
vironment may play a role in ca ncer progression. Chen
et al. showed that some cancer patients have gene
exp ression patterns in their adjacent non-neoplastic tis-
sue that are similar to invasive breast cancer signatures,
and that these signatures may predict progression of
early premalignant lesions [12]. Graham et al. also
found that gene expression in normal epithelium of ER
positive and E R negative breast cancers ec hoes the ER
status of the adjacent tumors [13]. These observations
suggest that the stroma and/or epithelium adjacent to
tumors may harbor changes, referred to as field effects
[14], and that these changes may be of prognostic value.
However, an investigation of genomic heterogeneity in
the extratumoral microenvironment, independent of
effects caused by adjacent tumor characteristics, has not
been reported. Identification of gene expression subtypes
in the extratu moral microenvironment may provide
important insights into how stromal response alters the
progression of disease.
To evaluate the hypothesis that extratumoral microen-
vironment influences disease progression, we used can-
cer-adjacent non-neoplastic tissue fr om 72 invasive
breast cancer and ductal carcinoma in situ ca ses to
identify distinct gene expression subtypes in extratu-
moral tissue. Biological features of these subtypes were
defined by comparison with established and novel gene
expression signatures and by performing histological
and immunohistochemical analyses. The novel microen-
vironment subtypes identified were also evaluated for
associations with overall survival. Our data suggest two
biologically distinct subtypes of extratumoral
microenvironment with distinct biological features and
clinical outcomes.
Materials and met hods
Patient samples
Patients were women u ndergoing mastectomy a t Uni-
versity of North Carolina Hospitals in Chapel Hill, NC.
All patients enrolled voluntarily under Institutional
Review Board-approved protocols. For histologically
normal tissue adjacent to breast cancer or ductal carci-
noma in situ, a pathologist from the Lineberger Com-
prehensive Cancer Centers Tissue Procurement Facility
at the University of North Carolina in Chapel Hill con-
firmed that a mirror specimen adjacent to that used for
RNA isolation was histologically benign.
Patients with invasive cancers (n = 68) and ductal car-
cinoma in situ (DCIS) (n = 4) were considered. Patien ts
receiving neoadjuvant therapy were excluded. ER-posi-
tive patients were identified by pathologic report and
endocrine-treated patients were defined as cancer
patients whose medical records reflect that they received
anti-estrogen therapy, such as tamoxifen or aromatase
inhibitors. Median follow-up time for the 72 patients
was 39 months. All tissues were handled by snap freez-
ing immediately after surgery. RNA was isolated using
established protocols as described previously [7].
To evaluate intra-individual variation in the signatures
we identified, we used duplicate specimen samples from
the same individuals. For these a nalyses, biospecimens
included five pairs of cancer-adjacent tissue pairs from
the University of North Carolina Hospitals in Chapel
Hill, NC and five pairs from a study conducted by the
National Cancer Institute in Warsaw and Lodz, Poland
[15]. For all 10 pa irs, both peritumoral (< 2 cm) and
remote (2+ cm) tissues were sampled.
Cell lines and reagents for generating TWIST and TGFb
signatures
To identify a gene expression profile assoc iated with
TWIST overexpression, the non-tumorigenic immorta-
lized human mammary epithelial cells (HMLE) and the
HMLE cells stably expressing the transcription factor
TWIST1 (HMLE-TWIST) [16] were cultured in
HuMEC media with HuMEC Supplement and Bovine
Pituitary Extract (GIBCO, Carlsbad CA, USA). HMLE-
TWIST cells were harvested after 48 h in culture (at
80% confluence) and compared with H MLE cells har-
vested under an identical protocol. To identify a signa-
ture of fibroblast activation by transforming growth
factor beta (TGF-b), hTERT-immortalized fibroblasts
from reduction mammoplasty patients (RMF) [17]
(maintained as described previously [18]), were treated
with 50 pg TGF-b1 (PeproTech , Rockyhill, N J, USA) for
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 2 of 12
Page 2
48 h. TGF-b1 was reconstituted in 10 mM Citric Acid
and 2 mg/mL albumin in PBS, with supplemented
media replaced every 24 h. Gene expression profiles of
TGFb-treated RMF cells were compared to those in
untreated RMFs.
Microarray analysis
Microarrays were two-col or Agilent 4X44k G4112F
arrays (Agilent, Santa Clara, CA, USA) or custom 244 k
human arrays (Agilent G4502A). The probe set common
to both platforms was used fo r all analyses as previously
describ ed [7]. Cy3-labeled reference was produced from
total RNA from Stratagene Universal Human Reference
(spiked with 1:1,000 with MCF-7 RNA and 1:1,000 with
ME16C RNA to increase expression of breast cancer
genes) following amplification with Agilent Quick Amp
labeling kit following the manufacturersprotocolwith
minor modifications as described in Hu et al. [19]. The
identical protocol was applie d to Cy5-label total RNA
from cell lines and cancer-adjacent tissue. Data are pub-
licly available through the Gene Expression Omnibus
(GSE31589).
Unsupervised clustering
All data were loaded to the UNC Microarray Data base
for normalization and filtering. Only probes with signal
intensity greater than 10 dpi in both channels were
included. Data were lowess normalized, probes with at
least 80% good data were selected, and missing data
were imputed using k-nearest neighbors imputation
with k = 10. To identify subtypes of microenvironment,
we selected probes with an inter-quartile range (IQR) of
at least 0.8. This probe set was used to perform unsu-
pervised clustering of cancer-adjacent tissues (Cluster
3.0). Functional analysis of gene clusters in the unsuper-
vised clust er was performed to identify significant func-
tions with P-values less than 0.05 (using Ingenuity
Pathway Analysis (IPA), with Benjamini-Hochberg mul-
tiple testing correction).
Associations between extratumoral microenvironment
subtypes and biologically-defined gene expression
signatures
To characterize the biological phenotypes of the extratu-
moral subtypes, gene expression in each sample was
compared to existing and novel gene expression signa-
tures. TWIST1-related signatures reflecting cellular de-
differentiation and activated stroma were identified via
two-class unpaired Significance Analysis of Microarrays
(SAM) [20] with a false discov ery rate (FDR) < 0.01
comparing HMLE-TWIST with parental HMLE cell line
(vector only) (Cell lines used are described in [21]).
TGF-b1-dependent responses of fibroblasts were identi-
fied using two-class unpaired SAM with FDR < 0.01
comparing treated versus sham RMFs (as described
above). The median centered expression profile of each
individual patient was then evaluated for correlation
with these signatures by calculating Pearson correlation
coefficients, using the method of Creighton et al. [22].
Three published signatures, for desmoid-type fibrosis
(DTF) [4], Claudin-low breast cancer [23] and estrogen
response [24], were assessed the same way. Briefly, vec-
tors corresponding to all genes in each of the five signa-
tures were constructed, with 1 assigned to up-regulated
genes and -1 assigned to down-regulated genes. A Pea r-
son correlation coefficient was calculated for this stan-
dard vec tor versus the vector of median centered gene
expression for each patient. Patients were classified as
positive for a given signature if the Pearson correlation
coefficient was greater than zero, and negative if the
coefficient was less than zero.
After evaluating the 72 patient sample set as
described, additional samples were microarrayed to eval-
uate prevalence of Active subtype as a function of dis-
tance to tumor. For each of 10 patients, 2 samples (one
peritumoral and one remote ) were analyzed by microar-
rayandthe20-samplesetwas median-centered. Each
sample was classified as Active or Inactive using the
Pearson correlation coefficient for median-centered
expression of that sample v ersus the vector of up (1)
and down (-1) genes in the Active patients from the 72
patient set. Patients with a positive Pearson correlation
coefficient were classified as Active (others were
Inactive).
Survival analysis
Associations between cluster group (Active versus Inac-
tive) and overall survival were evaluated using Kaplan-
Meier and Cox-Proportional Hazards analyses in R
2.11.0 (survival package ). Univariate Cox models were
used to estimate the hazard ratio (HR) and 95% CI.
Hazard Ratios did not change substantially with exclu-
sion of the one DCIS patient in this dataset. Relapse-
free survival data were incomplete for a large number of
patients and were underpowered in this dataset, but sur-
vival trends similar to those for overall survival were
observed (data not shown).
Immunohistochemistry
Formalin fixed mirror samples of non-neoplastic tissue
adjacent to breast cancer were paraffin embedded and
assayed by IHC to compare markers across the 72
Active and Inactive patients and in 5 pairs of peritu-
moral and remote specimens collected at UNC (where
paraffin was available). Five micrometer paraffin sections
were obtained, deparaffinized, hydrated, and antigen
retrieval (HIER) was performed using Dako Target
Retrieval Solution S1699(Dako, Glostrup Denmark).
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 3 of 12
Page 3
Sections were washed in 0.05 M Tris buffer pH 7.6 and
treated with 3.0% hydrogen peroxide in dH
2
0for10
minutes to reduced endogenous peroxidase. Sections
were blocked in Dako protein serum free solution
(X0909) for 30 minutes and incubated overnight at 4°C
with antisera against TWIST1 (Abcam, ab50887, Cam-
bridge, UK)) diluted 1:200 in Dako antibody diluent
(S08 09). TWIST1 staining was revealed through the use
of a biotinylated secondary ( Jackson ImmunoResearch
115-065-166, Reston VA, USA), ABC complex (Standard
Elite Vector Laboratories, PK6100, Burli ngame, CA,
USA) and reacted in diaminobenzidine (Invitrogen DAB
Substrate Kit 00-2014, Carlbad, CA, USA). Subsequent
hema toxylin or Masson s trichrome counterstaining was
performed, slides were dehydrated, cleared in xylene and
DPX (BDH 360294H) and cover-slipped for light micro-
scopy at 10X. All slides were scanned using Aperio
ScanScope CS and a nalyzed with Aperio Image Scope
V11 .0.2.725 (Vista, CA, USA) using the nuclear analysis
algorithm Nuclear V9. TWIST-positive (2+ or 3+)
nuclear counts per unit area of epithelium and per unit
area of stroma were calculated.
Results
Identification of distinct extratumoral microenvironment
subtypes adjacent to breast cancer
To identify gene expression subtypes of extratumoral
microenvironment, we analy zed microarray data on 72
extratumoral tissues from women with invasive breast
cancer or ductal carcinoma in situ. Unsupervised clus-
tering o n approximately 3,500 of the most variable
probes resulted in two distinct clusters (Figure 1A, with
complete gene list presented in Additional file 1: Table
S1). Clusters were defined as Active and In active as
describ ed in Methods. Kaplan-Meier analysis for overall
survival across all patients showed that patients with an
Active signature in their extratumoral microenviron-
ment had poorer overall surviv al (HR = 1.4, 95% CI: 0.6
to 2.8, P = 0.337), but the association was weak and not
statistically significant when considering all breast can-
cer patients. However, extratumoral microenvironment
is unlikely to show dominance over highly aggressive
tumor biology and even tumor-derived signatures have
poor accur acy in pr edicting ER-negative patient survival
due to the uniformly poor survival of this group [25].
Therefore, we also assessed prognostic value of Active
versus Inactive subtype among ER-positive (n = 43) and
endocrine-treated (n = 42) patients. Among these
patients, there was a strong association between Active
subtype and overall survival (Figure 1B, C; ER-positi ve
patients Active versus Inactive HR = 2.5, 95% CI: 0.9 to
6.7, P = 0.062; endocrine-treated patients Active versus
Inactive HR = 2.6, 95% CI: 1.0 to 7.0, P = 0.045 ). These
results suggest that the phenotype of the extratumoral
microenvironment may have value as an independent
predictor o f ER-positive/hormone-treated patient
outcome.
To further evaluate clinical implications of extratu-
moral subtypes, we analyzed the association of common
clinical paramet ers with Active versus Inactive subtype.
We found no statistically significant association between
Active/Inactive subtype and standard prognostic clinico-
pathological parameters, including breast cancer sub-
type, ER status, tumor size, and tumor grade (Table 1).
Thus, the extratumoral microenvironment subtypes are
independent of tumor subtype and do not appear to be
strongly driven by the biology of the tumor. We also
evaluated several clinical factors as potential confoun-
ders of the relationship between Active/Inactive subtype
and survival. While this dataset is not large enough for
well-powered m ultiv ariable analyses, the HR comparing
Active to Inactive remained stable and above 2.0 even
after a djusting for grade (high and medium versus low
as two indicator variables), subtype (normal-like, luminal
B, HER2 and basal-like versus luminal A as four indica-
tor variables), size (< 2 cm versus >/ = 2 cm) and age
(continuous).
Biological relevance of microenvironment subtypes
To establish the biological character of the two clusters
in Figure 1, we performed ontology anal ysis using Inge-
nuity Pathway Analysis (IPA) (Table 2). Genes highly
expressed in the Active group (Figure 1A, upper right
panel), were genes related to movement of cells (P =
3.94E-17), inflammatory response (P = 3 .09E-15), con-
nective tissue disorder (P = 6.22E-15), fibrosis (P =
2.63e-07), chemotaxis of cells (P = 6.23E-07) and
recruitment of macrophages (P = 4.24E-04). Genes that
were expressed at lower levels in the Active groups (Fig-
ure 1A, lower right panel) over-represented functional
categories, such as adhesion of cells (P = 3.38E-06), dif-
ferentiation of epithelial cells (P = 1.08E-03), formation
of cell-cell contacts (P = 1.41E-03). Additional file 2:
Table S2 shows full ontology analyses for the gene clus-
ters highlighted in Figure 1 (IPA results).
The aggregation of biological behaviors observed in
the Active group, as well as some of the individual gene
expression changes, represented processes associated
with activated stroma or active dedifferentiation/epithe-
lial-to-mesenchymal transition (EMT). Because extratu-
moral microenvironments show activation of wound
response gene expression [7], and because EMT occurs
normally during wound healing [26] and in cancer [23],
we performed additional analyses and experiments to
evaluate the hypothesis that EMT-like signatures distin-
guishe d these two clusters. EMT is a process that alters
the polarity of epithelial cells and reshapes them for
movement. TWIST1, TWIST2 and ZEB2 are
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 4 of 12
Page 4
transcription factors that bind to E-boxes within the
CDH1 promoter, suppressing transcription and ulti-
mately down-regu lating E-cadherin protein [27,28]. Loss
of E-cadherin is a key step in the dedifferentiation of
epithelial cells to a mesenchymal phenotype [16]. Simul-
taneously, there is a loss of cell-cell contact proteins,
such as claudins, and gain of mesenchymal markers,
such as vimentin and S100A4 (also known as fibroblast-
specific protein 1, FSP1) [29,30]. Additional file 3: Figure
S1 visualizes the expression of the markers listed above
and other established markers associated with a dediffer-
entiated phenotype, and de monstrates higher expression
(relative to median) of TWIST1, TWI ST2 and ZEB1
genes across samples in the Active Cluster. We also
observed the same pattern of expr ession for the
intermediate filament protein vimentin, the cytoskeletal
protein S100A 4 and the recept or tyrosine kinase DDR2.
On the other hand, we found lower e xpression of tight
junction associated proteins claudins 3, 4, 7, occludin
and the calcium dependent cell-to-cell adhesion prote in
E-cadherin in the Active cluster.
Consistent with these observations, we identified and
evaluated an epithelium-derived signature of TWIST1
activatio n (derived using cell-line experiments described
in Methods) and the claudin-low tumor signature [23],
associated with tumor cell EMT [21], for association
with Active subtype. Table 3 shows Active samples were
strongly associated with both the TWIST expression sig-
nature and c laudin-low tumor signatures. In addition,
the association with claudin-low expression was
Figure 1 Ident ification of extratumoral microenvironment subtypes adjacent to breast cancer. (A) Unsupervised clustering dendrogram
for 3,518 probes yields two main clusters of non-neoplastic tissues. Cluster 1/Active subtype (left, orange) is associated with high expression of
genes involved in cellular movement, inflammatory response, fibrosis and low expression of genes involved in cellular adhesion and
differentiation (See also Table 2 for gene ontology analyses). Cluster 2/Inactive Subtype has higher cellular adhesion and differentiation-related
gene expression. Kaplan-Meier analysis and univariate Cox Proportional Hazards analysis (to estimate hazard ratios) were conducted comparing
overall survival for Active versus Inactive. Results are presented for ER positive patients only (B) and for endocrine-treated patients (C).
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 5 of 12
Page 5
independent of whether the adjacent tumor expressed
claudin-low features, and claudin-low phenotype was
much more prevalent in the normal tissue (> 40%) than
in the tumors (< 15%).
Stroma-derived signatures were also considered in
association with Active versus Inactive subtype. TGFb
signaling in the microenvi ronment has been implicated
in tumor progression and is enriched in active wound
healing, so we evaluated association of Active subtype
with TGFb signaling. Active samples were more likely
to be positively correlated with the TGFb signature,
though the associ ation was weaker than for TWIST and
claudin-low signatures. We also evaluated correlation
with desmoid-type fibrosis signature [4], which is asso-
ciated with breast cancer progression when measured in
tumors, and found no strong association between Active
subtype and fibrosis. The relatively weaker association
between Active subtype and each of these two stroma-
derived signatures suggests that th e biological character-
istics of Active versus Inactive are more s imilar to the
cellular dedifferentiation capture d in the TWIST and
claudin-low signatures.
To identify protein level overexpression of TWIST
andtoidentifythesourceoftheTWISTproteinin
breast cancer samples, we performed immunohisto-
chemistry along with Massons trichrome in clinical spe-
cimens. Aperio scanned images analyzed by ImageScope
software demonstrated that the density of TWIST posi-
tive cells was higher in both epithelium and stro ma of
Active patients (73% increase in density of TWIST posi-
tive cells in epitheli um, 44% increase in density of
TWIST positive cells in stroma relative to Inactive
group). Two representative ductal specimens, one each
from Active and Inactive patients, depicting Masson s
Trichrome, the TWIST-stained samples, and Aperio
ImageScope scoring for nuclear markup/quantitative
scoring are shown in Figure 2. The images show that
even though the pictured Inactive specimen had more
epithelial cells, there are far fewer red and orange (3+
and 2+, respectively) cells relative to wha t is seen in the
Active sample, and there are a larger number of yellow
and red (negative or 1+) cells. These images illustrate
the trend observed across the entire IHC sample set: the
density of intense TWIST staining was higher among
Active patients.
In tumors, estrogen response gene expression strongly
distinguishes two groups of patients and often drives
progression [31]. To ensure that the subtypes were not
simply measuring estrogen responsiveness of the tumors
and, thereby, acting as a surrogate for ER status of the
tumors , we evaluate d a published estrogen response sig-
nature [24]. Consistent with previous reports that ER
positive tumors are more likely to show high estrogen
response gene expression in adjacent epithelium, the
estrogen re sponsiveness signature was more likely to be
positive in tissue adjacent to ER positive breast cancers.
However, the Active and Inactive subtypes were inde-
pendent of the estrogen response gene set, with no asso-
ciation between Active microenvironment and estrogen
response gene expression. In other words, the Active
patients are not defined by estrogen responsiveness.
Prevalence of subtypes as a function of distance from
tumor
Spatial variation of the Active signature across a
patients tissue is an important consideration if the sig-
nature represents a candidate biom arker of prognosis
among ER positives. For our unsupervised clustering in
Figure 1, samples were taken peritumorally, but tissue
was not sampled distant from the tumor and precise
distances were not recorded. Thus, in a subset of sam-
ples, we specifically evaluated both peritumoral and
remote tissue to identify whether two locations from the
same patient were concordant (Additional file 4: Table
Table 1 Clinical characteristics of 72 patients evaluated
for extratumoral gene expression subtypes
Total Active, N (%) Inactive, N (%)
Menopausal Status
Pre 32 10 (31.3) 22 (68.7)
Post 32 15 (46.9) 17 (53.1)
Missing 8 2 (25.0) 6 (75.0)
Breast Cancer Subtype
Normal-like 4 1 (25.0) 3 (75.0)
Luminal A 22 8 (36.4) 14 (63.6)
Luminal B 9 3 (33.3) 6 (66.7)
Her2+ 7 3 (42.9) 4 (57.1)
Basal-like 12 4 (33.3) 8 (66.7)
Missing 18 8 (44.4) 10 (55.6)
ER status
Positive 44 19 (43.2) 25 (56.8)
Negative 23 7 (30.4) 16 (69.6)
Missing 5 1 (20.0) 4 (80.0)
HER2 status
Positive 13 5 (38.5) 8 (61.5)
Negative 38 12 (31.6) 26 (68.4)
Missing 21 10 (47.6) 11 (52.4)
Tumor size (cm)
0 to 2 25 14 (56.0) 11 (44.0)
2.1 to 4 24 9 (37.5) 15 (62.5)
4+ 19 4 (21.1) 15 (78.9)
Missing 4 0 (0.0) 4 (100)
Tumor grade
1 6 3 (50.0) 3 (50.0)
2 18 4 (22.2) 14 (77.8)
3 28 13 (46.4) 15 (53.6)
Missing 20 7 (35.0) 13 (65.0)
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 6 of 12
Page 6
Table 2 Pathway analysis of enriched ontological categories, major gene clusters demarcated in Figure 1
Category Function annotation Adjusted P-value
a
Highly expressed in Active
Cellular Movement Movement of cells 3.94E-17
Homing of cells 3.17E-07
Chemotaxis of cells 6.23E-07
Infiltration of cells 4.22E-05
Inflammatory Response Inflammatory response 3.09E-15
Infiltration by neutrophils 1.91E-04
Recruitment of macrophages 4.24E-04
Connective Tissue Disorders Connective tissue disorder 6.22E-15
Organismal Injury and Abnormalities Fibrosis 2.63E-07
Highly expressed in Inactive
Cell-to-Cell Signaling and Interaction Adhesion of cells 3.38E-06
Formation of cell-cell contacts 1.41E-03
Formation of tight junctions 1.96E-02
Formation of intercellular junctions 2.40E-02
Cellular development Differentiation of epithelial cells 1.08E-03
Maturation of epithelial cells 8.05E-03
Differentiation of epithelial tissue 5.13E-02
Tissue Development Morphogenesis of epithelial tissue 6.13E-03
a
Benjamini-Hochbe rg adjusted P-values calculated in Ingenuity Pathway Analysis.
Table 3 Correlations of cancer-adjacent gene expression with biologically relevant signatures, by microenvironment
subtype and tumor characteristics
Microenvironment subtype Tumor subtype/tumor characteristics
Signature Number of genes
a
Active
N (%)
N =27
Inactive
N (%)
N =45
Claudin low
N (%)
N =9
Non- Claudin low
N (%)
N =63
ER positive
b
N (%)
N =44
ER negative
N (%)
N =22
TWIST 831/794/664
positive 25 (92.6) 7 (15.6)
negative 2 (7.4) 38 (84.4)
P = 5.7e-11
Claudin Low 807/754/478
Positive 26 (96.3) 7 (15.6) 6 (66.7) 27 (42.9)
Negative 1 (3.7) 38 (84.4) 3 (33.3) 36 (57.1)
P = 3.6e-12 P = 0.28
TGFb 217/207/181
Positive 18 (66.7) 15 (33.3)
Negative 9 (33.3) 30 (66.7)
P = 7.7e-3
DTF 511/451/351
Positive 16 (59.3) 16 (35.6)
Negative 11 (40.7) 29 (64.4)
P = 0.086
ER Response 754/700/407
positive 14 (51.9) 21 (46.7) 25 (56.8) 5 (22.7)
negative 13 (48.1) 24 (53.3) 19 (43.2) 17 (77.3)
P = 0.80 P = 0.010
a
Correlations are presented using data for genes that mapped and had IQR > than median IQR. The number of genes for each signature are given as total
number of genes in published signature/number of genes that mapped to the dataset/number of genes in datase t with IQR > median IQR. This final number of
genes was used to perform association analyses.
b
six tumors were missing ER status.
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 7 of 12
Page 7
S3). Interestingly, 60% of patients had concordant nor-
mal tissue signatures at both locations, while 40% of
patients had Active signatur e in one location and not in
the other, (that is, differences existed between peritu-
moral and remote signatures). This s uggests that dis-
tance to t umor may be an important source of
intraindividual variation in expression of this phenotype.
Discussion
Pathways and processes that are part of normal homeos-
tasis, such as wound healing and TGF-b signaling,
become tumor promoting in the presence of initiated
cells [ 32,33]. While this has been proven experimentally,
the r ole of t hese processes in predic ting human cancer
outcomes has been less well studied in observational
contexts. We p reviously observed that wound healing
signatures can be detected in more than 90% of tumor
adjacent tissues [7], and in the current study, we
demonstrate that extratumoral wound response can be
of two types: Active and Inactive. Active cancer-adjacent
tissue shows features of activ e EMT or cellular dediffer-
entiation and was associat ed with poor survival. Inactive
extratumoral tissue maintains higher expression of cellu-
lar adhesion genes and shows lower expression levels of
EMT-related transcription factors. Our preliminary esti-
mate of the hazard ratio (HR) associated with Active
(compared to Inactive) microenvironment was approxi-
mately 2.5, which is s ubstantial considering that es tab-
lished genomic phenotypes such as p53 m utation status
[34] and Cyclin E overexpression [35] have HRs near
two. Acknowledging the limitations of this study, which
included heterogeneously treated patients of various
stages, the magnitude of this preliminar y effect estimate
suggests that the association merits further investigation
for clinical relevance and lends credence to the influ-
ence of tumor microenvironment on breast cancer
outcome.
The clinical relevance of our gene expression subtypes
appeared, in our study, to be restricted to ER-positive
patients. It is common for tumor-derived gene expres-
sion signatures to have prognostic value only among
ER-positive tumors, which may be attributable to rela-
tively uniform, poor survival among ER-negative
patients. A recent paper simultaneously reviewed multi-
ple gene expression signatures and showed that very few
tumor signatures had value in predicting ER-negative
breast cancer survival [25]. However, it may also be the
case that there are specific biological interactions
Figure 2 A ctive phenotype is associated with increased d ensity of T WIST positive cells . Representative Active and Inactive tissues are
shown. Two trichrome images are shown (different magnifications) and the TWIST1-DAB stained section illustrates that despite a greater density
of epithelial cells in the Inactive patient, there is a lower density of TWIST staining and the intensity of staining (nuclear markup image) is
reduced relative to Active patients. The nuclear markup scale ranges from blue (no staining) to red (3+ intensity).
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 8 of 12
Page 8
between ER-positive tumors and their microenviron-
ments. For example, TGF-b is a well-established para-
crine mediator of breast cancer aggressiveness [3 6,37],
and consistent with our findings, TGF-b signatures are
prognostic solely among ER-positive cancers [38]. This
example is particularly relevant given th at Active micro-
environments are correlated with a TGF-b signature.
Increased TGF-b signaling along with other biological
characteristics of the Ac tive microenvironment subtype
suggested dedifferentiation or EMT-like gene expression,
which may seem surprising given that EMT is a devel-
opmental prog ram. However, EMT is known to be acti-
vatedinnormal,adulttissuesundergoingwound
healing (referred to as type 2 EMT, in contrast to type 1
EMT that occurs during development) [26]. In tumors
and cancer cells, EMT-associated expression has been
observed [39-41] (referred to as t ype 3 EMT), and has
recently been linked with the claudin-low subtype and
with resistance to therapy [21,23,26,42]. In our study,
we observed that claudin-low gene expression features
are more common (more than twice as likely to occur)
in cancer-adjacent tissues than in breast tumors, and
occur independent of whether the adjacent tumor is
claudin-low. In fact, claudin-low gene expression in the
adjacent tissue appears to be independent of tumor sub-
type altogether. This suggests that the Active signature
may be an endogenous response, dependent upon
patient biology and not tumor biology. It also appears
that the signature can occur early in disease pathogen-
esis, being present in some DCIS patients in the current
study.
A recent study has suggested that claudin-low features
may al so occur (though much less frequently) in normal
tissues from non-diseased women [43]. In that study,
claudin-low features were associate d with high risk of
breast c ancer. Future studies should assess whether the
prevalence of this subtype differs according to disease
progression (DCIS versus Invasive), or other patient
characteristics, such as ag e, smoking history or race. In
one study of esophageal cancers, it has been hy pothe-
sized that EMT could be artificially induced in surgery
[44], consistent with the idea that EMT is induced dur-
ing wounding. However, tumors displaying this signa-
ture were exposed to ischemic conditions for at least
four hours. In collections at UNC, more than 95% of
specimens are sna p frozen within two hours of devascu-
larization, with the majority snap frozen in u nder an
hour. Thus induction of ischemic conditions i s unlikely
to explain our observation of EMT-like gene expression.
In support of an exogenous or tumor-specific induc-
tion of EMT-like gene expression, EMT marker expres-
sion has been previously reported as a highly localized
response to cancer progression. Trujillo et al. [45]
observed high expression of SNAIL1 and TGFb in the
breast epithelium adjac ent to tumors. Interestingly,
across the five patients evaluated in that study, these
markers were most highly expressed in the peritumoral
region (< 1 cm) and expression was much lower at a
distance of 5 cm from the tumo r margin. In our ana-
lyses, many pairs of peritumoral and remote tissue had
pervasive expression of an EMT-like signature, both
near (< 2 cm) and remote from the tumor (> 2 cm). Dif-
ferences between the number of markers (hundreds of
genes versus few markers) and differences in the sensi-
tivity of RNA-based and protein-based analyses may
account for differences between our study and the
report by Trujillo et al. However, future work should
continue to evaluate the role of distance from tumor in
modifying these biological processes. If there is a wide
geographic range of the tumor-effects on adjacent tissue,
this has potential implications for surgery strategies.
One might speculate that patients for whom a wide-
spread gene expression alteration is present might bene-
fit from mastectomy. If a promoting wound healing
signature or Active signature persists in the lumpectomy
bed, it raises the question of whether this could account
for the higher rate of recurrences in this region after
breast conserving therapy. If this hypothesis is to be
advanced, it will be important to evaluate whether
Active/Inactive signatures persists after breast conser-
ving therapy.
Other studies of stroma-derived signatures have sup-
ported both exogenous and endogenous orig ins for stro-
mal response [46]. Whether the signature is tumor-
dependent or a host factor varies signature by signature.
In our findings, estrogen response signatures in extratu-
moral tissue w ere correlated with tumor ER status.
However, neither the tumor ER status nor the estrogen
responsiveness of the adjacent tissue were correlated
with the extratumoral subtypes we ident ified, suggesting
that the extratumoral phenotypes are not driven by hor-
monal exposures or endogenous ho rmone response.
These results are important because breast tumor biol-
ogy is so strongly driven by e strogen response and
because previous reports have shown that estrogen posi-
tive tumor features are reflected in the adjacent normal
tissue of these patients [13]. However, while our Active
and Inactive subtypes are independent of ho rmone
receptor status of the adjacent tumor, we did recapitu-
late the finding of Graham et al. [13] that extratumoral
estrogen responsiveness mirrors the ER expression phe-
notype of the adjacent tissue, despite the fact that we
did not micro-dissect the e pithelium from the stroma.
Thus, we are able to report that even signatures such as
estrogen responsivenes s are detectable in whole (not
micro-dissected) tissue. The use of whole tissue facili-
tates microarray a nalysis of a much larger number of
samples (more than 90 in the current study) facilitated
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 9 of 12
Page 9
the assessment of clinical associations and inter-and
intra-individua l variation that would not ha ve been pos-
sible with smaller sample sizes that have been used in
micro-dissection studies [1,13].
The higher prevalence of Active subtype in our study
of women with DCIS and cancer (compared to the
study of non-diseased high risk patients) suggests that
either the signature is associated with risk or that initial
activation of the signature requires the presence of
tumor or another early benign lesion [7]. Further
research to determine temporality and distinguish risk
and response is needed. It will also be important to
determine whether there are some extratumoral signa-
tures that are dictated by tumor characteristics, to fully
explore the role of t he extratumoral tissue in disease
progression. For example, grade is known to induce
widespread stromal alterations, but the specif ic molecu-
lar signatures associ ated with high grade tumors are not
well studied. Likewise, extratumoral microenvironments
that are common to basal-like breast cancers and c ould
account for higher risk of loco-regional recurrence
could provide novel insights abo ut the biology of basal-
like breast cancer progression.
Conclusions
We have explored novel subtypes o f extratumoral
microenvironment that appear to be independent of
breast cancer subtype and that may h ave prognostic
value. We have shown that these subtypes have distinct
biological characteristics that may determine their clini-
cal impact. While the extratumoral subtypes we identi-
fied were not dependent on tumor characteristics, we
have confirmed the findings of others that some extratu-
moral characteristics vary according to the ER-status of
the adjacent t umor. These results demonstrate that
studying the cancer-adjacent tissue can provide novel
biological and clinically-relevan t insights ab out breast
cancer progression.
Additional material
Additional file 1: Table S1. Complete gene list used for
unsupervised clustering and ingenuity pathway analysis. Each gene
in Figure 1A is presented in this table showing average median-centered
log2(R/G) in Active and Inactive groups. Fold change is calculated from
the ratio of columns C and D. Genes highlighted in red are those in the
gene cluster marked in orange in Figure 1A and genes highlighted in
green are the gene cluster marked in grey in Figure 1A. These gene
clusters were used to perform the IPA analyses presented in Table 1 and
Table S2.
Additional file 2: Table S2. Ingenuity Pathway Analysis of Molecular
and Cellular Functions associated with gene clusters in Figure 1A.
Gene categories, functions, function annotation, Benjamini-Hochberg P-
value and lists of molecules detected per category are shown for each of
the two clusters (orange and grey) identified in Figure 1 and
enumerated in Table S1.
Additional file 3: Figure S1. Identification of EMT markers in
extratumoral microenvironment subtypes in Active and Inactive
patients. EMT-associated genes selected from the literatu re are visualized
across the two sample groups from Figure 1.
Additional file 4: Table S3. Concordance of extratumoral subtypes
in paired tissues from the same patient. At least two patient samples
were used for microarray analysis and Active versus Inactive subt ype was
evaluated in each. Samples include specimens from the University of
North Carolina at Chapel Hill Normal Breast Study and samples collected
in the NCI-funded Polish Womens Breast Cancer Study.
Abbreviations
DCIS: ductal carcinoma in situ; DTF: desmoid-type fibrosis; EMT: epithelial-to-
mesenchymal transition; ER: estrogen receptor; FDR: false discovery rate;
HMLE: human mammary epithelial cells; HR: hazard ratio; IPA: Ingenuity
Pathway Analysis; IQR: interquartile range; RMF: reduction mammoplasty
fibroblast; SAM: Significance Analysis of Microarrays; TGF-β: transforming
growth factor beta.
Acknowledgements
This project was supported by grants from the National Cancer Institute and
National Institutes of Environmental Health Sciences (U01-ES019472, R01-
CA138255), a Breast SPORE (P50CA058223) Career Development Award to M.
A.T., a grant from the Avon Foundation, and the University Cancer Research
Fund at the University of North Carolina. P.C-H is a Howard Hughes Medical
Institute (HHMI) Med into Grad Scholar supported in part by a grant to the
University of North Carolina at Chapel Hill from HHMI through the Med into
Grad Initiative. S.A.M. laboratory is supported by a grant from National
Institute of Cancer (R01 CA155243). We are grateful to Jonine Figueroa, Mark
Sherman and Gretchen Gierach at the National Cancer Institute for
contributing samples from the Polish Womens Breast Cancer study and for
helpful comments on this work and to Amy Drobish for assistance with
medical records abstraction for patient clinical samples.
Author details
1
Department of Epidemiology, University of North Carolina at Chapel Hill,
Campus Box 7435, Chapel Hill, NC 27599, USA.
2
Department of Pathology
and Laboratory Medicine, University of North Carolina at Chapel Hill, Campus
Box 7525, Chapel Hill, NC 27599, USA.
3
Lineberger Comprehensive Cancer
Center, University of North Carolina at Chapel Hill, Campus Box 7295, Chapel
Hill, NC 27599, USA.
4
Division of Cancer Prevention, National Cancer Institute,
6130 Executive Blvd, Bethesda, MD 20892 USA.
5
Department of Molecular
Pathology, University of Texas, Houston, TX 77053, USA.
6
Metastasis Research
Center, M. D. Anderson Cancer Center, University of Texas, Unit Number 951,
Houston, TX 77053, USA.
7
UNC Department of Surgery, University of North
Carolina at Chapel Hill, Campus Box 7050, Chapel Hill, NC 27599, USA.
Authors contributions
ERP performed microarray and clinical data analyses, analyzed images for
staining, and drafted the manuscript. PCH helped to coordinate the sample
processing and contributed to drafting the manuscript. JRP performed
microarray analysis and contributed to drafting the manuscript. JR
performed RNA extraction from tissues, labeled the RNA and performed
hybridizations to microarrays. RAL helped design the study and interpret the
results. LAC provided access to clinical data and contributed to data
interpretation. SAM participated in the selection and interpretation of EMT-
associated markers and helped to draft the manuscript. KDA and MAT
conceived of the study, participated in its design and coordination, and
helped to draft the manuscript. All authors read and approved the final
manuscript.
Competing interests
The authors declare that they have no competing interests.
Received: 22 June 2011 Revised: 13 January 2012
Accepted: 19 March 2012 Published: 19 March 2012
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 10 of 12
Page 10
References
1. Finak G, Bertos N, Pepin F, Sadekova S, Souleimanova M, Zhao H, Chen H,
Omeroglu G, Meterissian S, Omeroglu A, Hallett M, Park M: Stromal gene
expression predicts clinical outcome in breast cancer. Nat Med 2008,
14:518-527.
2. Chang HY, Sneddon JB, Alizadeh AA, Sood R, West RB, Montgomery K,
Chi JT, van de Rijn M, Botstein D, Brown PO: Gene expression signature of
fibroblast serum response predicts human cancer progression:
similarities between tumors and wounds. PLoS Biol 2004, 2:E7.
3. Beck AH, Espinosa I, Edris B, Li R, Montgomery K, Zhu S, Varma S,
Marinelli RJ, van de Rijn M, West RB: The macrophage colony-stimulating
factor 1 response signature in breast carcinoma. Clin Cancer Res 2009,
15:778-787.
4. Beck AH, Espinosa I, Gilks CB, van de Rijn M, West RB: The fibromatosis
signature defines a robust stromal response in breast carcinoma. Lab
Invest 2008, 88:591-601.
5. Ma XJ, Dahiya S, Richardson E, Erlander M, Sgroi DC: Gene expression
profiling of the tumor microenvironment during breast cancer
progression. Breast Cancer Res 2009, 11:R7.
6. Schedin P, Borges V: Breaking down barriers: the importance of the
stromal microenvironment in acquiring invasiveness in young womens
breast cancer. Breast Cancer Res 2009, 11:102.
7. Troester MA, Lee MH, Carter M, Fan C, Cowan DW, Perez ER, Pirone JR,
Perou CM, Jerry DJ, Schneider SS: Activation of host wound responses in
breast cancer microenvironment. Clin Cancer Res 2009, 15:7020-7028.
8. Sharma M, Beck AH, Webster JA, Espinosa I, Montgomery K, Varma S, van
de Rijn M, Jensen KC, West RB: Analysis of stromal signatures in the
tumor microenvironment of ductal carcinoma in situ. Breast Cancer Res
Treat 2009, 123:397-404.
9. Allinen M, Beroukhim R, Cai L, Brennan C, Lahti-Domenici J, Huang H,
Porter D, Hu M, Chin L, Richardson A, Schnitt S, Sellers WR, Polyak K:
Molecular characterization of the tumor microenvironment in breast
cancer. Cancer Cell 2004, 6:17-32.
10. Cichon M, Degnim A, Visscher D, Radisky D: Microenvironmental
Influences that Drive Progression from Benign Breast Disease to Invasive
Breast Cancer. Journal of Mammary Gland Biology and Neoplasia 2010,
15:389-397.
11. de Kruijf EM, van Nes JG, van de Velde CJ, Putter H, Smit VT, Liefers GJ,
Kuppen PJ, Tollenaar RA, Mesker WE: Tumor-stroma ratio in the primary
tumor is a prognostic factor in early breast cancer patients, especially in
triple-negative carcinoma patients. Breast Cancer Res Treat 2011,
125:687-696.
12. Chen DT, Nasir A, Venkataramu C, Fulp W, Gruidl M, Yeatman T: Evaluation
of malignancy-risk gene signature in breast cancer patients. Breast
Cancer Res Treat 2009, 120:25-34.
13. Graham K, Ge X, de Las Morenas A, Tripathi A, Rosenberg CL: Gene
expression profiles of estrogen receptor-positive and estrogen receptor-
negative breast cancers are detectable in histologically normal breast
epithelium. Clin Cancer Res
17:236-246.
14.
Heaphy CM, Griffith JK, Bisoffi M: Mammary field cancerization: molecular
evidence and clinical importance. Breast Cancer Res Treat 2009,
118:229-239.
15. Garcia-Closas M, Egan KM, Newcomb PA, Brinton LA, Titus-Ernstoff L,
Chanock S, Welch R, Lissowska J, Peplonska B, Szeszenia-Dabrowska N,
Zatonski W, Bardin-Mikolajczak A, Struewing JP: Polymorphisms in DNA
double-strand break repair genes and risk of breast cancer: two
population-based studies in USA and Poland, and meta-analyses. Hum
Genet 2006, 119:376-388.
16. Mani SA, Guo W, Liao MJ, Eaton EN, Ayyanan A, Zhou AY, Brooks M,
Reinhard F, Zhang CC, Shipitsin M, Campbell LL, Polyak K, Brisken C, Yang J,
Weinberg RA: The epithelial-mesenchymal transition generates cells with
properties of stem cells. Cell 2008, 133:704-715.
17. Proia DA, Kuperwasser C: Reconstruction of human mammary tissues in a
mouse model. Nat Protoc 2006, 1:206-214.
18. Camp JT, Elloumi F, Roman-Perez E, Rein J, Stewart DA, Harrell JC,
Perou CM, Troester MA: Interactions with fibroblasts are distinct in Basal-
like and luminal breast cancers. Mol Cancer Res 9:3-13.
19. Hu Z, Fan C, Oh DS, Marron JS, He X, Qaqish BF, Livasy C, Carey LA,
Reynolds E, Dressler L, Nobel A, Parker J, Ewend MG, Sawyer LR, Wu J, Liu Y,
Nanda R, Tretiakova M, Ruiz Orrico A, Dreher D, Palazzo JP, Perreard L,
Nelson E, Mone M, Hansen H, Mullins M, Quackenbush JF, Ellis MJ,
Olopade OI, Bernard PS, et al: The molecular portraits of breast tumors
are conserved across microarray platforms. BMC Genomics 2006, 7:96.
20. Tusher V, Tibshirani R, Chu G: Significance analysis of microarrays applied
to the ionizing radiation response. Proc Natl Acad Sci USA 2001,
98:5116-5121.
21. Taube JH, Herschkowitz JI, Komurov K, Zhou AY, Gupta S, Yang J,
Hartwell K, Onder TT, Gupta PB, Evans KW, Hollier BG, Ram PT, Lander ES,
Rosen JM, Weinberg RA, Mani SA: Core epithelial-to-mesenchymal
transition interactome gene-expression signature is associated with
claudin-low and metaplastic breast cancer subtypes. Proc Natl Acad Sci
USA 107:15449-15454.
22. Creighton CJ, Casa A, Lazard Z, Huang S, Tsimelzon A, Hilsenbeck SG,
Osborne CK, Lee AV: Insulin-like growth factor-I activates gene
transcription programs strongly associated with poor breast cancer
prognosis. J Clin Oncol 2008, 26:4078-4085.
23. Prat A, Parker J, Karginova O, Fan C, Livasy C, Herschkowitz J, He X, Perou C:
Phenotypic and molecular characterization of the claudin-low intrinsic
subtype of breast cancer. Breast Cancer Research 2010, 12:R68.
24. Oh DS, Troester MA, Usary J, Hu Z, He X, Fan C, Wu J, Carey LA, Perou CM:
Estrogen-regulated genes predict survival in hormone receptor-positive
breast cancers. J Clin Oncol 2006, 24:1656-1664.
25. Fan C, Prat A, Parker JS, Liu Y, Carey LA, Troester MA, Perou CM: Building
prognostic models for breast cancer patients using clinical variables and
hundreds of gene expression signatures. BMC Med Genomics 4:3.
26. Kalluri R, Weinberg RA: The basics of epithelial-mesenchymal transition. J
Clin Invest 2009,
119:1420-1428.
27.
Comijn J, Berx G, Vermassen P, Verschueren K, van Grunsven L, Bruyneel E,
Mareel M, Huylebroeck D, van Roy F: The two-handed E box binding zinc
finger protein SIP1 downregulates E-cadherin and induces invasion. Mol
Cell 2001, 7:1267-1278.
28. Vesuna F, van Diest P, Chen JH, Raman V: Twist is a transcriptional
repressor of E-cadherin gene expression in breast cancer. Biochem
Biophys Res Commun 2008, 367:235-241.
29. Okada H, Danoff TM, Kalluri R, Neilson EG: Early role of Fsp1 in epithelial-
mesenchymal transformation. Am J Physiol 1997, 273:F563-574.
30. Strutz F, Okada H, Lo CW, Danoff T, Carone RL, Tomaszewski JE, Neilson EG:
Identification and characterization of a fibroblast marker: FSP1. J Cell Biol
1995, 130:393-405.
31. Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR,
Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A, Williams C,
Zhu SX, Lonning PE, Borresen-Dale AL, Brown PO, Botstein D: Molecular
portraits of human breast tumours. Nature 2000, 406:747-752.
32. Bissell MJ, Hines WC: Why dont we get more cancer? A proposed role of
the microenvironment in restraining cancer progression. Nat Med
17:320-329.
33. Sieweke M, Thompson N, Sporn M, Bissell M: Mediation of wound-related
Rous sarcoma virus tumorigenesis by TGF-beta. Science 1990,
248:1656-1660.
34. Rossner P Jr, Gammon MD, Zhang YJ, Terry MB, Hibshoosh H, Memeo L,
Mansukhani M, Long CM, Garbowski G, Agrawal M, Kalra TS, Gaudet MM,
Teitelbaum SL, Neugut AI, Santella RM: Mutations in p53, p53 protein
overexpression and breast cancer survival. J Cell Mol Med 2009,
13:3847-3857.
35. Keyomarsi K, Tucker SL, Buchholz TA, Callister M, Ding Y, Hortobagyi GN,
Bedrosian I, Knickerbocker C, Toyofuku W, Lowe M, Herliczek TW, Bacus SS:
Cyclin E and survival in patients with breast cancer. N Engl J Med 2002,
347:1566-1575.
36. Barcellos-Hoff MH, Akhurst RJ: Transforming growth factor-beta in breast
cancer: too much, too late. Breast Cancer Res 2009, 11:202.
37. Bhowmick NA, Chytil A, Plieth D, Gorska AE, Dumont N, Shappell S,
Washington MK, Neilson EG, Moses HL: TGF-beta signaling in fibroblasts
modulates the oncogenic potential of adjacent epithelia. Science 2004,
303:848-851.
38. Bierie B, Chung CH, Parker JS, Stover DG, Cheng N, Chytil A, Aakre M,
Shyr Y, Moses HL: Abrogation of TGF-beta signaling enhances chemokine
production and correlates with prognosis in human breast cancer. J Clin
Invest 2009, 119:1571-1582.
39. Foubert E, De Craene B, Berx G: Key signalling nodes in mammary gland
development and cancer. The Snail1-Twist1 conspiracy in malignant
breast cancer progression. Breast Cancer Res 2010, 12
:206.
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 11 of 12
Page 11
40. Sehgal PB: Interleukin-6 induces increased motility, cell-cell and cell-
substrate dyshesion and epithelial-to-mesenchymal transformation in
breast cancer cells. Oncogene 2010, 29:2599-2600, author reply 2601-2593.
41. Tuhkanen H, Soini Y, Kosma VM, Anttila M, Sironen R, Hamalainen K,
Kukkonen L, Virtanen I, Mannermaa A: Nuclear expression of Snail1 in
borderline and malignant epithelial ovarian tumours is associated with
tumour progression. BMC Cancer 2009, 9:289.
42. Thiery JP: Epithelial-mesenchymal transitions in tumour progression. Nat
Rev Cancer 2002, 2:442-454.
43. Haakensen VD, Lingjaerde OC, Luders T, Riis M, Prat A, Troester MA,
Holmen MM, Frantzen JO, Romundstad L, Navjord D, Bukholm IK,
Johannesen TB, Perou CM, Ursin G, Kristensen VN, Borresen-Dale AL,
Helland A: Gene expression profiles of breast biopsies from healthy
women identify a group with claudin-low features. BMC Med Genomics
2011, 4:77.
44. Aoyagi K, Minashi K, Igaki H, Tachimori Y, Nishimura T, Hokamura N,
Ashida A, Daiko H, Ochiai A, Muto M, Ohtsu A, Yoshida T, Sasaki H:
Artificially induced epithelial-mesenchymal transition in surgical
subjects: its implications in clinical and basic cancer research. PLoS One
2011, 6:e18196.
45. Trujillo KA, Heaphy CM, Mai M, Vargas KM, Jones AC, Vo P, Butler K, Joste N,
Bisoffi M, Griffith JK: Markers of fibrosis and epithelial to mesenchymal
transition demonstrate field cancerization in histologically normal tissue
adjacent to breast tumors. Int J Cancer 129:1310-1321.
46. Wu JM, Beck AH, Pate LL, Witten D, Zhu SX, Montgomery KD, Allison KH,
van de Rijn M, West RB: Endogenous versus tumor-specific host response
to breast carcinoma: a study of stromal response in synchronous breast
primaries and biopsy site changes. Clin Cancer Res 17:437-446.
doi:10.1186/bcr3152
Cite this article as: Román-Pérez et al.: Gene expression in extratumoral
microenvironment predicts clinical outcome in breast cancer patients.
Breast Cancer Research 2012 14:R51.
Submit your next manuscript to BioMed Central
and take full advantage of:
Convenient online submission
Thorough peer review
No space constraints or color figure charges
Immediate publication on acceptance
Inclusion in PubMed, CAS, Scopus and Google Scholar
Research which is freely available for redistribution
Submit your manuscript at
www.biomedcentral.com/submit
Román-Pérez et al. Breast Cancer Research 2012, 14:R51
http://breast-cancer-research.com/content/14/2/R51
Page 12 of 12
Page 12
  • Source
    • "An important upstream regulator of PKC-α was demonstrated to be the second most potently upregulated kinase by HFD and dramatically reduced by weight loss, the serine/threonine-protein kinase D1 (KPCD1, Prkd1), also known as PKD1. PKD1 has been shown to increase cell proliferation in breast, prostate, salivary tumors and pancreatic cancers [51, 52] . PKD1 also reduced serumand anchorage-dependence for proliferation and survival in vitro and drove tumorigenesis in xenograft models of mammary tumors [53]. "
    [Show abstract] [Hide abstract] ABSTRACT: Background: Obesity is associated with an aggressive subtype of breast cancer called basal-like breast cancer (BBC). BBC has no targeted therapies, making the need for mechanistic insight urgent. Reducing adiposity in adulthood can lower incidence of BBC in humans. Thus, this study investigated whether a dietary intervention to reduce adiposity prior to tumor onset would reverse HFD-induced BBC. Methods: Adult C3(1)-Tag mice were fed a low or high fat diet (LFD, HFD), and an obese group initially exposed to HFD was then switched to LFD to induce weight loss. A subset of mice was sacrificed prior to average tumor latency to examine unaffected mammary gland. Latency, tumor burden and progression was evaluated for effect of diet exposure. Physiologic, histology and proteomic analysis was undertaken to determine mechanisms regulating obesity and weight loss in BBC risk. Statistical analysis included Kaplan-Meier and log rank analysis to investigate latency. Student's t tests or ANOVA compared variables. Results: Mice that lost weight displayed significantly delayed latency compared to mice fed HFD, with latency matching those on LFD. Plasma leptin concentrations significantly increased with adiposity, were reduced to control levels with weight loss, and negatively correlated with tumor latency. HFD increased atypical ductal hyperplasia and ductal carcinoma in situ in mammary gland isolated prior to mean latency-a phenomenon that was lost in mice induced to lose weight. Importantly, kinome analysis revealed that weight loss reversed HFD-upregulated activity of PKC-α, PKD1, PKA, and MEK3 and increased AMPKα activity in unaffected mammary glands isolated prior to tumor latency. Conclusions: Weight loss prior to tumor onset protected against the effects of HFD on latency and pre-neoplastic lesions including atypical ductal hyperplasia and DCIS. Using innovative kinomics, multiple kinases upstream of MAPK/P38α were demonstrated to be activated by HFD-induced weight gain and reversed with weight loss, providing novel targets in obesity-associated BBC. Thus, the HFD-exposed microenvironment that promoted early tumor onset was reprogrammed by weight loss and the restoration of a lean phenotype. Our work contributes to an understanding of underlying mechanisms associated with tumor and normal mammary changes that occur with weight loss.
    Full-text · Article · Dec 2016 · Cancer Cell International
  • Source
    • "For example, extensive gene expression changes have been observed in the stroma associated with ductal carcinoma in situ (DCIS) and invasive ductal carcinoma (IDC), suggesting the co-evolution of the tumor adjacent stroma with epithelium even before tumor invasion and supporting the important role of stromal changes in the transition from pre-invasive to invasive tumor growth [4]. shown to provide novel biological and clinically relevant insights into breast cancer progression [5, 6]. These studies demonstrate that the tumor microenvironment is an important player in tumorigenesis. "
    [Show abstract] [Hide abstract] ABSTRACT: The tumor microenvironment is well known to play a role in sustaining malignant transformation of tissue, tumor progression, and in drug responsiveness; however, much remains unclear about the interplay between tumor cells, the extracellular matrix, and stroma cells. The extracellular matrix has been shown to elicit both biochemical and biophysical signaling, and matrix rigidity is an important microenvironmental parameter in the regulation of cellular behavior. Thus, tissue engineering and the development of novel biomaterials that mimic mechanical and topological properties of tumor stroma and can cope with the effect of mechanical forces are promising approaches to study this interplay. New in vitro tools to investigate the effect of mechanical signals on breast cancer cell aggressiveness and drug sensitivity include genipin-crosslinked gelatin hydrogel scaffolds with adjustable degrees of stiffness.
    Full-text · Article · Jun 2015
  • Source
    • "GASCs without such properties. In accordance with our findings, Roman-Pérez et al., (2012) [35] reported the presence of two different subtypes of extratumoral microenvironment influencing the aggressiveness and outcome of human breast cancers. Furthermore, in 2013, Al Rakan et al. [36] carried out molecular and cellular characterizations of breast stromal fibroblasts (TCFs) from negative surgical margins. "
    [Show abstract] [Hide abstract] ABSTRACT: Glioblastoma (GB) is a highly infiltrative tumor recurring within a few centimeters of the resection cavity in 85 % of cases, even in cases of complete tumor resection and adjuvant chemo/radiotherapy. We recently isolated GB-associated stromal cells (GASCs) from the GB peritumoral zone, with phenotypic and functional properties similar to those of the cancer-associated fibroblasts present in the stroma of carcinomas. In particular, GASCs promote blood vessel development and have tumor-promoting effects on glioma cells in vitro and in vivo. In this study, we characterized these cells further, by analyzing the transcriptome and methylome of 14 GASC and five control stromal cell preparations derived from non-GB peripheral brain tissues. We identified two subtypes of GASCs in surgical margins in GB patients: GASC-A and GASC-B. GASC-B promoted the development of tumors and endothelium, whereas GASC-A did not. A difference in DNA methylation may underlie these two subtypes. We identified various proteins as being produced in the procarcinogenic GASC-B. Some of these proteins may serve as prognostic factors for GB and/or targets for anti-glioma treatment. In conclusion, in this era of personalized therapy, the status of GASCs in GB-free surgical margins should be taken into account, to improve treatment and the prevention of recurrence.
    Preview · Article · Dec 2014 · Journal of Neuro-Oncology
Show more