carcinoma associated with a more aggressive phenotype
Otto Walter MDa, Manju Prasad MDa, Shaolei Lu MD, PhDb, Robert M. Quinlan MDc,
Kathryn L. Edmiston MDd, Ashraf Khan MD, FRCPatha,⁎
aDepartment of Pathology, UMass Memorial Medical Center, Worcester, MA, 01655, USA
bDepartment of Cancer Biology, UMass Medical School, Worcester, MA, 01655, USA
cDivision of Surgical Oncology, UMass Memorial Medical Center, Worcester, MA, 01655, USA
dDivision of Hematology/Oncology, UMass Memorial Medical Center, Worcester, MA, 01655, USA
Received 13 March 2009; revised 28 April 2009; accepted 6 May 2009
Invasive ductal carcinoma;
Summary IMP3, an oncofetal protein, is a member of the insulin-like growth factor-II (IGF-II) mRNA-
binding protein family. Its relevance as a novel biomarker in lung, pancreatic, renal, and cervical
adenocarcinoma was recently revealed. However, its role in breast carcinogenesis and tumor progression
is not yet established. Basal-like carcinoma was initially identified by gene expression profiling. It
accounts for 15% to 30% of all breast cancers. These tumors express basal epithelial markers including
cytokeratin 5 but lack expression of the estrogen receptor, progesterone receptor, and human epidermal
growth factor receptor 2 (HER2), therefore, are often referred to as triple negative breast cancer. They
have been found to be associated with a worse overall and disease-free survival. In this retrospective
study, we examined the IMP3 expression in invasive ductal carcinoma of the breast and correlated its
expression with morphological and biologic prognostic factors. The study group comprised 138 cases of
invasive ductal carcinoma retrieved from the surgical pathologic files for a 10-year period from 1997 to
2006. Survival data and clinical stage were available on all 138 patients. Tumor characteristics including
size, grade, lymphovascular invasion, necrosis, lymph node metastasis, estrogen receptor, progesterone
receptor, and HER2 status were obtained from pathologic reports. Immunohistochemistry was
performed on formalin-fixed paraffin-embedded tissue using mouse monoclonal antibody against IMP3
and CK5/6. Of the 138 breast cancer cases, IMP3 expression was seen in 45 (33%). Twenty-five of the
IMP3+ cases were triple negative. We found significant correlation between IMP3 expression and
higher grade (P = .001), necrosis (Pb.0001) triple negative, and CK5/6 expression (P b.0001 for each).
Cox multivariate analysis showed a hazard ratio of IMP3 expression at 3.14 (P = .05). IMP3 is a novel
biomarker for triple negative (basal-like) invasive mammary carcinoma, and its expression is associated
with a more aggressive phenotype and decreased overall survival.
© 2009 Elsevier Inc. All rights reserved.
Breast cancer is the most frequent malignancy in women
and ranks second among cancer deaths in adult females .
Despite recent progress in screening, diagnosis, and treat-
ment, 25% of breast cancer patients experience a relapse
E-mail address: email@example.com (A. Khan).
0046-8177/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
Human Pathology (2009) 40, 1528–1533
during follow-up and die of disease. Traditionally, lymph
node status, tumor size, and histologic grade were considered
to be the most useful prognostic factors in breast cancer
patients . Microarray gene expression analysis has
redefined breast cancer classification and has identified
distinct subgroups with unique patterns of gene expression
associated with disease outcome . One of these groups, the
basal-like cancers, shows characteristics of the normal basal
epithelial cells such as expression of high-molecular-weight
cytokeratin (CK5/6 or CK14). They are also defined by the
absence of estrogen, progesterone receptor (ER and PR), and
human epidermal growth factor receptor 2 (HER2) expres-
sion, therefore, are often referred to as triple negative (TN)
carcinomas. These breast carcinomas have been found in
numerous studies to be associated with a worse overall and
disease-free survival compared with other types [4,5] and is
the subtype frequently observed in familial breast cancers
related to Breast Cancer 1 (BRCA1) mutation.
IMP3 is a member of the insulin-like growth factor
(IGF-II) mRNA binding protein (IMP) family that consists of
IMP1, IMP2, and IMP3 . IMP family members are
important in RNA trafficking and stabilization, cell growth,
and cell migration during the early stages of embryogenesis
. IMP3 has been also referred in previous publications as
L523S. IMP3 is also identical to the KOC (KH-domain
containing protein overexpressed in cancer) protein that was
originally cloned from a pancreatic tumor complementary
DNA screen, and its gene is located on 7p11.5 . IMP3 is
expressed in developing epithelium, muscle, and placenta
during early stages of human and mouse embryogenesis but
is expressed at low or undetectable concentrations in normal
adult tissues . IMP3 is also expressed in multiple
malignant tumors including pancreatic, lung, endocervical,
endometrial, renal cell, and transitional cell carcinoma
[9-15]. Moreover, IMP3 has been shown to promote
proliferation of human leukemia cells , suggesting that
it is an oncofetal protein that could have a crucial role in
regulating cell proliferation. Although substantial clinical,
epidemiologic, and experimental evidence indicate that the
different IGFs and their receptors are intimately involved in
normal mammary gland growth and differentiation as well as
in breast cancer initiation and progression [17,18], IMP3
expression in breast cancer and its relationship to tumor
behavior are unknown.
In this retrospective study, we examined IMP3 expression
in invasive ductal carcinoma of the breast and correlated it
with known morphological and biologic prognostic factors.
2. Materials and methods
2.1. Case selection
One hundred thirty-eight patients with the diagnosis of
invasive ductal carcinoma were identified by chart review at
UMass Memorial Medical Center (Worcester, MA) from
January 1997 to December 2006 and enrolled in the study.
The case selection was based on results of a pilot study that
suggested increased IMP3 expression in TN carcinomas. We
therefore initially collected retrieved grade III invasive ductal
carcinoma from our files and later added 18 grade I and 25
grade II tumors on a random basis from the same period for
comparison. The diagnoses were confirmed by at least 2
pathologists independently. The specimens included core
biopsies, lumpectomies, and partial and complete mastec-
tomies. Tumor characteristics including tumor size, histolo-
gic grade, lymphovascular invasion (LVI), necrosis, lymph
node metastasis, ER, PR, and HER2 status were obtained
from pathologic reports. The histologic grade was assessed
by using the Nottingham modification of the Scarff-Bloom-
Richardson grading scheme (NSBR) .
Clinical staging was based on pathologic findings and
clinical information from the hospital tumor database using
the system of the American Joint Committee on Cancer
. The survival data were obtained from the Hospital
2.2. Immunohistochemical study
Immunohistochemical staining for IMP3 was performed
as described in previous publications [9,13]. Briefly, 5-μm
tissue sections of the selected blocks were deparaffinized,
washed, immersed in buffer (citrate, 0.01 mol/L; pH 6.0 for
IMP3; and EDTA, 1 mmol/L; pH 8.0 for CK5/6), and
subjected to heat-induced antigen retrieval (microwave,
770 W, 14 minutes). The slides were cooled, rinsed with
PBS, and stained on Dako Autostainer (Dako Corporation,
Carpinteria, CA) at room temperature. Endogenous perox-
idase activity was blocked with 3% hydrogen peroxide, and
the slides were rinsed and treated with a blocking protein
(ChemMate; Ventana, Tucson, AZ) to prevent nonspecific
staining. The sections were then incubated with L523S
mouse monoclonal antibody specific for IMP3 (2 mg/mL;
Corixa Corporation, Seattle, WA) for 45 minutes and D5/16
B4 mouse monoclonal antibody specific for CK5/6 (Dako)
for 30 minutes. After brief buffer washes, the sections were
treated with a cocktail of biotinylated antirabbit IgG and
antimouse IgG/IgM (ChemMate) for 30 minutes, washed,
and incubated with avidin/biotin/peroxidase complex
(ChemMate) for 30 minutes. The sections were rinsed,
developed with diaminobenzidine (ChemMate) and hydro-
gen peroxide (10 minutes), rinsed with tap water, and
counterstained with hematoxylin. Representative sections of
pancreatic carcinoma were used as positive controls for
IMP3 expression, and sections of normal skin were used as
positive control for CK5/6. Negative controls were
performed by replacing the primary antibody with non-
A positive stain for IMP3 was defined as brown stain
seen in the cytoplasm and/or the cell membrane. Positive
1529 Novel biomarker for triple negative invasive mammary carcinoma
stain for CK5/6 was defined as cytoplasmic brown stain.
The percentage of ductal carcinoma that was positive for
either IMP3 or CK5/6 was assessed. IMP3 staining of more
than 10% of the tumor cells was considered positive. The
intensity of positive staining was recorded as weak,
moderate, or strong.
2.3. Statistical analysis
Correlation of IMP3 expression with other pathologic
tumor characteristics was performed using χ2test or Fisher
exact test. Cox proportional hazard model was used to
perform multivariate analysis to determine the independent
effects of IMP3 expression, age, stage, histologic grade,
lymph node metastasis, lymphovascular invasion, and tumor
necrosis, as well as ER, PR, and HER2 expression. Overall
survival curves were estimated and compared using Kaplan-
Meier estimates and log-rank test. All P values were 2-sided,
and P values of less than .05 were considered statistically
significant. These analyses were carried out using JMP 6.0
(SAS Institute, Cary, NC).
All the 138 enrolled patients were females, with an age at
diagnosis ranging between 29 and 94 years (median, 59.2
years). Using the NSBR grading system, we found 18 grade
I, 23 grade II, and 97 grade III breast cancers. By using the
American Joint Committee on Cancer (AJCC) classification
system, we identified 49 tumors at stage I, 40 at stage IIA
(T1, N1 and T2, N0), 28 at stage IIB (T2, N1 and T3, N0), 14
at stage III (IIIA, IIIB and IIIC), and 3 at stage IV, whereas
the staging information was not available in 4 cases.
Of the 138 breast cancer cases, IMP3 expression was seen
in 45 cases (33%); 93 tumors (67%) did not express IMP3.
Among the IMP3-positive cases, the staining was weak in
17, moderate in 16, and strong in 12 cases. Normal and
40× and D, 200×).
Negative IMP3 staining in normal ducts (A, 200× and B, 400×). Strong positive cytoplasmic IMP3 staining in a TN carcinoma (C,
1530O. Walter et al.
benign breast tissue was always negative for IMP3
expression (Fig. 1). No significant correlation was seen
between IMP3 expression and tumor size, lymphovascular
invasion, or lymph node metastasis.
IMP3 expression was found significantly more often in
high-grade tumors (41.67% in grade III versus 5.56% in
grade I and 13.04% in grade II) (Table 1). Similarly, IMP3
expression was found in significantly higher percentage
associated with tumor necrosis (49% versus 18% with
absence of necrosis). The ER-negative tumors were
significantly more often IMP3-positive (63% versus 11%
of the ER-positive). Similar data were found with PR-
negative (65% versus 16%) and HER2-negative (42%
versus 16%). The correlation was positive with CK5/6-
positive (79% versus 8% CK5/6-negative). Of 32 TN
tumors, 25 (78%) were IMP3-positive as opposed to 20
(19%) of 106 non-TN tumors being IMP3-positive.
Furthermore, of the 96 grade III tumors, 29 were TN and
67 non-TN, and of the 29 TN tumors, 24 (79%) were
IMP3-positive, whereas only 17 (25%) of 67 non-TN
tumors showed IMP3 expression; this difference was
statistically significant (P b .0010). IMP3 expression was
found in only 7 of the CK5/6-negative tumors, and 5 of
them showed only a weak staining intensity.
Survival data were available for 138 patients (mean
follow-up, 62.4 months; median, 71.5; range, 1-122
months). Fifty-one patients died during the follow-up
period, 19 of them had IMP3-positive (42% of total 45
IMP3+), and 32 of them had IMP3-negative tumor groups
(34% of total 93 IMP−).
The Kaplan-Meier curves revealed a tendency to impaired
survival in patients with IMP3-positive tumors (Fig. 2);
however, it did not reach significance (P = .113; log-rank).
To evaluate and compare the individual risk of IMP3
expression, age, tumor grade, stage, lymphovascular inva-
sion, lymph node metastasis, and hormone receptor status,
we performed a Cox proportional hazard analysis (Table 2).
IMP3 was found to be a significant independent predictor of
risk for the overall survival by increasing the risk of death
more than 3-fold (hazard ratio, 3.14 ± 1.83; P = .05). Age is
another independent predictor in the analysis (hazard ratio,
1.03 ± 0.02; P = .031).
To identify patients with high risk of recurrent or
metastatic disease and those who benefit from adjuvant
chemotherapy is an important task for the oncologist. Their
decision heavily depends on the histopathologic and
immunophenotypical characteristics of the tumor. For a
long time, the traditional clinicopathologic factors such as
lymph node status, tumor size, and histologic grade were
considered to be the most useful prognostic factors
determining further therapy . With the emergence of
hormone mediator therapy (tamoxifen) and targeted HER2
therapy (trastuzumab), determination of hormone receptor
and HER2 status became essential.
Furthermore, gene expression profiling has identified 5
molecular subtypes (luminal A, luminal B, normal breast-
IMP3 expression versus other tumor characteristics
(N = 93)
(N = 45)
ER and PR
doublenegative(DN) 12 (30.77%)
ER, PR, and HER2
triple negative (TN)
18 (18.18%)81 (81.82%)
20 (18.87 %)
aOne case of spindle cell carcinoma was not graded with NSBR,
therefore not included.
diagnosis to death) of patients with IMP3 (+) carcinoma and ones of
IMP3 (−) (P = .113, log-rank test).
Kaplan-Meier plot of overall survival curves (time from
1531Novel biomarker for triple negative invasive mammary carcinoma
like, HER2 overexpressing, and basal-like) that are refining
our understanding of breast cancer biology. These mole-
cular subtypes are classified on the basis of expression
patterns of several genes [3,4]. Unfortunately, microarray
studies are not widely available, and most of the proteins
transcribed from these genes do not have diagnostic
antibodies that can be used for immunohistochemistry on
formalin-fixed paraffin-embedded archival tissue. These
limitations make routine clinical identification of the basal-
like breast carcinoma challenging. Most of the breast
tumors are luminal subtypes with ER/PR positivity,
whereas HER2 expression defines the HER2 overexpres-
sing subtype. The tumors of basal-like subtype are
incompletely defined. Although multiple publications
have attempted to address the histologic and immunophe-
notypic characteristics of the basal-like breast cancers, a
clear consensus has not been reached.
Livasy et al  examined 23 microarray-proven basal-
like carcinomas. The most consistent immunophenotype in
their series was ER and HER2-negative and vimentin,
EGFR, CK8/18, and CK5/6-positive. Rakha et al 
proposed that the basal cytokeratins are the most reliable
marker of the basal-like carcinomas regardless of the
expression of other markers. Nielsen et al  found that
ER, HER1, HER2, and CK5/6 can accurately identify basal-
like tumors with high specificity. In accordance to these prior
data, we defined “basal-like cancer” as TN and CK5/6-
positive. These carcinomas have been found in numerous
studies to be associated with a worse overall and disease-free
survival compared with other breast carcinoma types
[3-5,24]. Some evidence suggests that the effect of adjuvant
chemotherapy with anthracycline and taxane is more
effective among ER- and PR-negative breast cancer
cases—either in the TN (basal-like) or HER2 overexpressing
group , whereas there is a higher relapse rate and
progressive disease among those tumors that could not be
eradicated by the chemotherapy . It is highly desirable to
define and identify with commercially available markers this
“chemotherapy responsive” group.
Studies suggest that IMP3 up-regulation and overexpres-
sion in malignant lesions may play an integral role in the
proliferation of malignant cells. Multiple studies proved
increased metastatic potential in IMP3-positive tumors.
Vikesaa et al  found that IMP3 expression is associated
with increased tumor invasion. Yaniv et al  suggested
that IMP3 promotes cell migration. Yantiss et al  have
reported a correlation between tumor stage and IMP3 (KOC)
expression in pancreatic carcinoma that suggests a relation-
ship between degree of expression and the development of
progressive dysplasia and carcinoma. Jiang at al [13,28]
reported significantly increased metastatic potential in a
subset of renal cell carcinomas with IMP3 expression
independent of stage. Recently, Zheng et al  noted that
highly aggressive endometrial cancers have higher IMP3
expression than more indolent carcinomas such as endome-
trioid endometrial adenocarcinoma. Our study is the first one
to detect IMP3 expression in invasive breast carcinoma,
which significantly correlates with adverse prognostic
factors such as high-grade tumor necrosis, TN phenotype,
and CK5/6 expression. Furthermore, in a multivariate
analysis, IMP3 expression was an independent predictor of
risk for overall survival. Whether IMP3 actually initiates
malignant transformation and proliferation or simply deve-
lops as a consequence of that transformation remains to be
seen in further studies.
Although IMP3 is expressed extensively during embryo-
genesis, the expression disappears from virtually all tissues
soon after birth and is generally undetectable in adult tissue
Multivariable Cox regression analyses
Hazard ratioSEzP N |z|95%Confidence interval
1532O. Walter et al.
except in placentas and gonads . We confirmed that Download full-text
IMP3 expression is seen neither in normal ducts nor in
stromal tissue. Our findings suggest that IMP3 may
contribute to the aggressive behavior of these carcinomas.
In this study, we observed that IMP3 is expressed most
often in TN breast cancers. The IMP3-positive tumors were
concomitantly CK5/6-positive, which is also a basal-like
breast carcinoma marker. Carey et al  describes a good
response to neoadjuvant chemotherapy in a subset of TN
breast carcinomas. The defining markers of this chemother-
apy responsive subset are not well understood. Although
our results show that IMP3 expression is an independent
risk factor for poor prognosis, in this current study, we did
not have the ability to investigate the possible differences of
chemotherapeutic response between IMP3-positive versus
negative groups. Future studies need to address whether
IMP3 is a reliable prognostic marker of the chemotherapy
resistant subset of triple negative or basal-like invasive
 Jemal A, Siegel R, Ward E, Murray T, Xu J, Thun MJ. Cancer
statistics, 2007. CA Cancer J Clin 2007;57:43-66.
 Fitzgibbons PL, Page DL, Weaver D, et al. Prognostic factors in breast
cancer. College of American Pathologists Consensus Statement 1999.
Arch Pathol Lab Med 2000;124:966-78.
 Perou CM, Sorlie T, Eisen MB, et al. Molecular portraits of human
breast tumours. Nature 2000;406:747-52.
 Sorlie T, Perou CM, Tibshirani R, et al. Gene expression patterns of
breast carcinomas distinguish tumor subclasses with clinical implica-
tions. Proc Natl Acad Sci USA 2001;98:10869-74.
 Carey LA, Dees EC, Sawyer L, et al. The triple negative paradox:
primary tumor chemosensitivity of breast cancer subtypes. Clin Cancer
 Nielsen J, Christiansen J, Lykke-Andersen J, Johnsen AH, Wewer UM,
Nielsen FC. A family of insulin-like growth factor II mRNA-binding
proteins represses translation in late development. Mol Cell Biol 1999;
 Mueller-Pillasch F, Pohl B, Wilda M, et al. Expression of the highly
conserved RNA binding protein KOC in embryogenesis. Mech Dev
 Mueller-Pillasch F, Lacher U, Wallrapp C, et al. Cloning of a gene
highly overexpressed in cancer coding for a novel KH-domain
containing protein. Oncogene 1997;14:2729-33.
 Yantiss RK, Woda BA, Fanger GR, et al. KOC (K homology domain
containing protein overexpressed in cancer): a novel molecular marker
that distinguishes between benign and malignant lesions of the
pancreas. Am J Surg Pathol 2005;29:188-95.
 Simon R, Bourne PA, Yang Q, et al. Extrapulmonary small cell
carcinomas express K homology domain containing protein over-
expressed incancer, but carcinoid tumors donot. HUM PATHOL 2007;38:
 Wang T, Fan L, Watanabe Y, et al. L523S, an RNA-binding protein as a
potential therapeutic target for lung cancer. Br J Cancer 2003;88:
 Li C, Rock KL, Woda BA, Jiang Z, Fraire AE, Dresser K. IMP3 is a
novel biomarker for adenocarcinoma in situ of the uterine cervix: an
immunohistochemical study in comparison with p16(INK4a) expres-
sion. Mod Pathol 2007;20:242-7.
 Jiang Z, Chu PG, Woda BA, et al. Analysis of RNA-binding protein
IMP3 to predict metastasis and prognosis of renal-cell carcinoma: a
retrospective study. Lancet Oncol 2006;7:556-64.
 Sitnikova L, Mendese G, Liu Q, et al. IMP3 predicts aggressive
superficial urothelial carcinoma of the bladder. Clin Cancer Res 2008;
 Zheng W, Yi X, Fadare O, et al. The oncofetal protein IMP3: a novel
biomarker for endometrial serous carcinoma. Am J Surg Pathol 2008;
 Liao B, Hu Y, Herrick DJ, Brewer G. The RNA-binding protein IMP-3
is a translational activator of insulin-like growth factor II leader-3
mRNA during proliferation of human K562 leukemia cells. J Biol
 Surmacz E. Function of the IGF-I receptor in breast cancer.
J Mammary Gland Biol Neoplasia 2000;5:95-105.
 Yee D, Lee AV. Crosstalk between the insulin-like growth factors and
estrogens in breast cancer. J Mammary Gland Biol Neoplasia 2000;5:
 Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. I.
The value of histological grade in breast cancer: experience from a
large study with long-term follow-up. Histopathology 1991;19:
 Greene FL, American Joint Committee on Cancer. American Cancer
Society. AJCC cancer staging manual. New York: Springer-Verlag;
 Livasy CA, Karaca G, Nanda R, et al. Phenotypic evaluation of the
basal-like subtype of invasive breast carcinoma. Mod Pathol 2006;19:
 Rakha EA, El-Sayed ME, Green AR, Paish EC, Lee AH, Ellis IO.
Breast carcinoma with basal differentiation: a proposal for pathology
definition based on basal cytokeratin expression. Histopathology
 Nielsen TO, Hsu FD, Jensen K, et al. Immunohistochemical and
clinical characterization of the basal-like subtype of invasive breast
carcinoma. Clin Cancer Res 2004;10:5367-74.
 van't Veer LJ, Dai H, van de Vijver MJ, et al. Gene expression
profiling predicts clinical outcome of breast cancer. Nature 2002;415:
 Berry DA, Cirrincione C, Henderson IC, et al. Estrogen-receptor status
and outcomes of modern chemotherapy for patients with node-positive
breast cancer. JAMA 2006;295:1658-67.
 Vikesaa J, Hansen TV, Jonson L, et al. RNA-binding IMPs promote
cell adhesion and invadopodia formation. EMBO J 2006;25:1456-68.
 Yaniv K, Fainsod A, Kalcheim C, Yisraeli JK. The RNA-binding
protein Vg1 RBP is required for cell migration during early neural
development. Development 2003;130:5649-61.
 Jiang Z, Lohse CM, Chu PG, et al. Oncofetal protein IMP3: a novel
molecular marker that predicts metastasis of papillary and chromo-
phobe renal cell carcinomas. Cancer 2008;112:2676-82.
 Yaniv K, Yisraeli JK. The involvement of a conserved family of RNA
binding proteins in embryonic development and carcinogenesis. Gene
1533 Novel biomarker for triple negative invasive mammary carcinoma