ORIGINAL ARTICLE

Tumoral epithelial and stromal expression of SMAD proteinsin pancreatic ductal adenocarcinomas

Adriana Handra-Luca • Pascal Hammel •

Alain Sauvanet • Philippe Ruszniewski •

Anne Couvelard

Published online: 12 May 2012

� Japanese Society of Hepato-Biliary-Pancreatic Surgery and Springer 2012

Abstract

Background SMAD proteins, intracellular mediators of

the transforming growth factor (TGF)-beta pathway,

function within two axes, the SMAD1/5/8 and SMAD2/3,

connected to TGF-beta and bone morphogenetic protein

(BMP) ligands. The SMAD proteins of these two axes

dimerize with SMAD4 and translocate to the nucleus.

SMAD signaling is characterized by a dichotomic func-

tioning, with tumor-suppressive functions and with loss of

normal growth inhibitory responses, depending on the

carcinogenesis stage. SMAD proteins also have pro-tumor

effects including abnormal extracellular matrix production.

Among tumors, pancreatic cancers harbor SMAD4 inacti-

vation the most frequently and the SMAD proteins are

considered to be key factors in pancreatic carcinogenesis.

Methods Our aims were to study the expression patterns

of the different types of SMAD proteins in pancreatic

ductal adenocarcinomas treated by surgical resection

(without neoadjuvant treatment) and their correlations with

morphological and clinical characteristics. We examined

the immunohistochemical expression of SMAD4, SMAD1/

5/8, and SMAD2/3 in 99 pancreatic ductal adenocarcino-

mas. Antibodies directed against the activated, phosphor-

ylated forms of proteins were used when appropriate

(SMAD1/5/8, SMAD2/3). Protein expression in the epi-

thelial tumor cells and in stromal fibroblasts was analyzed

with regard to morphological and clinical data.

Results Epithelial tumor cells showed SMAD1/5/8,

SMAD2/3, and, SMAD4 expression in 13, 93, and 45

tumors, respectively, and stromal fibroblast expression in 5,

11, and 22 tumors, respectively. Epithelial SMAD4 was

associated with a low, T1 or T2, TNM stage, and with the

presence of an abundant stroma (p = 0.05 and \0.01,

respectively). Activated stromal fibroblast SMAD2/3

expression was correlated with the presence of a fibrotic

focus (p = 0.01), whereas fibroblast SMAD4 was related

to a tendency for shorter postsurgical overall survival

(p = 0.07). The relationship of stromal, fibroblast SMAD4

to a worse outcome attained statistical significance in the

A. Handra-Luca

APHP Hopital Avicenne, Service d’Anatomie Pathologique,

Paris, France

A. Handra-Luca

EA 3406/3509, Universite Paris 13/Nord Medecine,

Bobigny, France

A. Handra-Luca (&)

Groupement Hospitalier Universitaire Paris Seine-Saint Denis,

APHP Hopital Avicenne, Universite Paris Nord/13 Medecine,

125 rue de Stalingrad, 93000 Bobigny, France

e-mail: adriana.handra-luca@avc.aphp.fr

P. Hammel � P. Ruszniewski

APHP Hopital Beaujon, Service de

Gastroenterologie-Pancreatologie, Clichy, France

P. Hammel � A. Sauvanet � P. Ruszniewski � A. Couvelard

UFR de Medecine, Universite Paris 7 Denis Diderot,

Paris, France

A. Sauvanet

APHP Hopital Beaujon, Service de Chirurgie Digestive,

Clichy, France

P. Ruszniewski � A. Couvelard

INSERM U773 Centre de Recherche Biomedicales Bichat

Beaujon CRB3, Paris, France

A. Couvelard

APHP Hopital Bichat Claude Bernard, Service d’Anatomie

Pathologique, Paris, France

123

J Hepatobiliary Pancreat Sci (2013) 20:294–302

DOI 10.1007/s00534-012-0518-6

group of patients with T1 and with N1 stage tumors

(p \ 0.01 and p = 0.04, respectively).

Conclusion In pancreatic ductal adenocarcinomas,

SMAD protein expression in epithelial tumor cells or in

stromal fibroblasts was related to stromal features and to a

shorter postsurgical overall survival. Our results point out

that the SMAD proteins play a role in the microenviron-

ment of this highly fibrotic tumor type.

Keywords Pancreas � Adenocarcinoma � Pathway �Pathology � Stroma

Introduction

The transforming growth factor beta (TGF-beta) pathway is

characterized by a dichotomic function, with tumor-sup-

pressive functions at early, precancerous tumor stages, and

with loss of the normal growth inhibitory responses and

abnormal extracellular matrix production at the established

tumor stage [1–3]. SMAD proteins are intracellular mediators

which play a central role in this pathway. They are classified

into three categories, receptor-SMAD (SMAD1, SMAD2,

SMAD3, SMAD5, and SMAD8), common partner-SMAD

(SMAD4), and regulatory, inhibitory-SMAD (SMAD6 and

SMAD7). The phosphorylated SMAD proteins belong to two

axes: the SMAD1/5/8 and the SMAD2/3 axes, which are

activated by TGF-beta and bone morphogenetic protein

(BMP) ligands, respectively [1, 4]. The SMAD proteins of

these two axes dimerize with SMAD4 and translocate to the

nucleus where they stimulate the transcription of target genes.

Among these proteins, SMAD4 has been largely studied in

pancreatic carcinogenesis. Alterations of the SMAD4 gene,

occurring in approximately 50% of pancreatic adenocarci-

nomas, have been reported, and more rarely, alterations of the

SMAD3 gene have also been reported [5–7]. Nuclear

SMAD4 staining is considered to be an accurate reflection of

gene status in pancreatic adenocarcinomas [7].

Lack of tumoral SMAD4 expression is predictive of a

worse outcome [8–10] and is related to lymph node

metastases [11]. More recently, inhibitory SMAD7

expression has been reported to predict a better outcome

[12]. SMAD2 protein was found to be expressed with a

moderate or severe intensity in 67 % of pancreatic ade-

nocarcinomas [13] and to be related to poor differentiation

[14]. Although the impact of SMAD proteins in pancreatic

ductal carcinogenesis is well established, these proteins

might be relevant for the stromagenesis process at a

molecular level as well [1, 15, 16].

Our aims were to study the epithelial and stromal

expression patterns of proteins of the SMAD family in

pancreatic ductal adenocarcinomas, and the correlations

with the pathological and clinical characteristics.

Methods

Patients

Patients with pancreatic ductal adenocarcinoma treated by

surgical resection and without neoadjuvant treatment were

selected. We did not include adenocarcinomas associated

with intraductal papillary mucinous tumors.

Patients were analyzed retrospectively for clinical and

survival data (using the clinical and surgical charts), and

for pathological data (by reviewing all available slides)

according to the WHO standard criteria [17, 18]. The

analyzed clinical data were: age, gender, tumor location,

and postsurgical overall survival (time to death or last

consultation).

Tumors were analyzed for the following morphological

features by reviewing all available slides for each case:

size and T; TNM stage, histological differentiation, and

lymph node metastasis (N; TNM stage). Because the bulk

of the tumor in pancreatic ductal adenocarcinomas fre-

quently contains only a minority of neoplastic epithelial

cells [15, 19, 20], we attempted to include desmoplastic

stromal features in the analysis of tumors. Therefore, we

assessed stromal morphological features such as the pre-

dominance of stroma (tumors with abundant/predominant

stroma being considered to be present when the stroma

represented more than 50 % of the tumor surface, as

compared to the epithelial, adenocarcinomatous tumor

component), and the presence of a fibrotic focus (defined

as a compact area, frequently central, of fibrosis and

containing fibroblasts and collagen) [21]. The patients’

main clinical characteristics and the baseline morpholog-

ical features of the tumors are summarized in Table 1

[22]. Forty-seven of the patients were women (median age

61 years, range 34–76 years) and 52, men (median age

63.5 years, range 37–78 years). Tumors were treated by

pancreaticoduodenectomy (88), distal pancreatectomy (9),

or total pancreatectomy (2). Postoperative treatment con-

sisted of chemotherapy (10 patients) or chemoradiother-

apy (48 patients), and 31 patients had no adjuvant

treatment. The adjuvant treatment schemes were adapted

to the patients’ general health status (poor performance

status) and age, and the chemotherapeutic agents used

were: gemcitabine and/or 5-fluorouracil combination (26

and 35 patients, respectively). The median follow-up time

was 26 months (range 1.2–95.5).

Immunohistochemistry

Paraffin-embedded tissue blocks were selected for each

tumor after review of all available hematoxylin-and-eosin-

stained slides (reviewed by A.H.L., A.C.). Tissue micro-

arrays included several 1-mm-diameter core biopsies from

J Hepatobiliary Pancreat Sci (2013) 20:294–302 295

123

each paraffin tissue block (formalin-fixed paraffin-embed-

ded tissue specimen) using a manual tissue-arraying

instrument (Manual Tissue Arrayer-MTA1; Beecher

Instruments, Sun Prairie, WI, USA) (Fig. 1). The cores

were taken from representative zones for the presence of

both tumor epithelial cells and stromal fibroblasts.

An automated technique (streptavidin biotin procedure,

Benchmark XT IHC/ISH; Ventana, Tucson, AZ, USA) was

used for immunohistochemistry. Antigen retrieval was

conducted by pretreatment at a high temperature. Sections

were incubated for 30 min with the antibody. Ultraview/

IVIEW DAB (Ventana) was used as the detection kit.

Slides were counterstained with hematoxylin and mounted.

We used antibodies directed against the activated,

phosphorylated forms for the SMAD proteins (activated by

phosphorylation) when available and appropriate [23].

Immunohistochemistry was performed with an anti-

SMAD4 antibody (clone B8, dilution 1:100, Santa Cruz

Biotechnology Clinisciences, Montrouge, France) [24, 25];

and with admixed antibodies directed against the phosphor-

ylated forms of SMAD1/5/8 (p-SMAD1/5/8: polyclonal

antibody directed against SMAD1 dually phosphorylated at

serines 463 and 465, as well as SMAD5 and SMAD8 when

phosphorylated at equivalent sites: serines 463 and 465,

respectively, serines 426 and 428; dilution 1:10, Cell Sig-

nalling Technology Ozyme, St Quentin en Yvelines, France)

and of SMAD2/3 (p-SMAD2/3: polyclonal antibody detect-

ing SMAD3 phosphorylated at serine 433 and serine 435 and

corresponding phosphorylated SMAD2, dilution 1:500,

Santa Cruz Biotechnology Ozyme, St Quentin en Yvelines,

France). The expression of these antibodies in inflammatory

cells/lymphocytes was considered as an internal control.

Immunohistochemical data were not available, due to tissue

loss, in 3, 2, and 8 tumors for the epithelial tumor component

and, in 1, 2, and 5 tumors for the stromal component, for

p-SMAD1/5/8, p-SMAD2/3 and, SMAD4, respectively.

Evaluation of immunohistochemistry was performed

without knowledge of the other data. For both epithelial

tumor cells and stromal fibroblasts (defined as stromal

spindle-shaped cells with a bland, regular, oval or round

nucleus), nuclear expression of SMAD proteins was eval-

uated [7, 26]. SMAD protein expression was considered as

negative or positive.

Statistical analysis

The relationships between morphological, immunohisto-

chemical, and clinical characteristics were analyzed using

the v2 test or Fisher’s exact test. The data were analyzed and

graphics were constructed with MedCalc (v11.1.1, Mari-

akerke, Belgium) statistical software. All tests were 2-sided.

The accepted level of statistical significance was p \ 0.05.

Associations with postsurgical overall survival were

examined by univariate survival analyses for all morpho-

logical and immunohistochemical as well as clinical

Table 1 Baseline clinical characteristics of the patients

Characteristic Number of patients

Gender

Female 47

Male 52

Surgery type

Pancreaticoduodenectomy 88

Left/distal pancreatectomy 9

Total pancreatectomy 2

Tumor T stage (TNM)

pT1 26

pT2 32

pT3 41

Lymph node metastasis

Absent (N0) 19

Present (N1) 80

Post-surgical treatment

Radiochemotherapy 48

Chemotherapy 10

Death 52

Fig. 1 Slide from tissue microarray blocs with tissue cores of 1 mm (a, b 92.5 and 5, respectively)

296 J Hepatobiliary Pancreat Sci (2013) 20:294–302

123

characteristics. The four patients with perioperative death

(within the month after surgery) were not considered for

this analysis, nor the 6 patients with unavailable follow-up

data. Survival curves were constructed by the Kaplan–

Meier method, and were compared by the logrank

Cox–Mantel test [27]. Variables associated with a p value

of less than 0.1 were used to construct multivariate Cox

models.

Results

Baseline characteristics of the tumors

The tumor T TNM stage was T1 in 26 patients, T2 in 32,

and T3 in 41 patients and the node N TNM stage was 1 in

80 patients. On whole hematoxylin-and-eosin slide analy-

sis, fibrotic foci were observed in 35 patients; 34 tumors

showed an abundant stromal tissue component, and in 20

tumors these 2 lesions co-existed. Tumor T stage and the

presence of a fibrotic focus were correlated with a tumor

size of 3 cm or more (3 cm was the median tumor size)

(p \ 0.01 and p = 0.02, respectively). The presence of

fibrotic foci was more frequent in tumors with abundant

stroma than in those without abundant stroma (58 vs. 20 %,

p \ 0.01).

Immunohistochemical expression of SMAD proteins

in the epithelial tumor component

Within the epithelial tumor component, SMAD4 expres-

sion was noted in 45 tumors; p-SMAD1/5/8 expression was

noted in 13, and p-SMAD2/3 expression was noted in 93

pancreatic ductal adenocarcinomas (Table 2; Fig. 2).

Among the 99 studied pancreatic adenocarcinomas, 41

(45 %) were negative for both SMAD4 and p-SMAD1/5/8

(of 91 PDAC with available immunohistochemistry for

these antibodies). In two of these tumors p-SMAD2/3 was

also negative.

Nuclear epithelial SMAD4 was related to a low tumor T

stage (p = 0.05). Tumors with abundant stroma were more

frequently SMAD4-positive (p \ 0.01) (Table 3).

Immunohistochemical expression of SMAD proteins

in the stromal tumor component

Stromal fibroblast expression of p-SMAD1/5/8 and

p-SMAD2/3 was observed in 5 and 11 tumors, respec-

tively, whereas SMAD4 was expressed in the stromal

fibroblasts of 22 tumors (Tables 2, 4). There was no sta-

tistically significant correlation between the different

SMAD proteins when expressed by stromal fibroblasts.

There was no correlation between epithelial and fibroblast

expression for any of the SMADs.

Fibroblast SMAD2/3 expression was correlated with the

presence of a fibrotic focus (p = 0.01), while lack of

fibroblast SMAD1/5/8 was correlated with the presence of

lymph node metastases (p = 0.04) (Table 3).

SMAD protein expression and survival

On univariate Kaplan–Meier survival analysis, only stro-

mal SMAD4 expression showed a trend relating to a

shorter survival (p = 0.07, Fig. 3). This relationship

attained statistical significance when considering the group

of patients with T1 tumors (p \ 0.01) or the group of

patients with N1 TNM stage (p = 0.04) (Fig. 4).

A multivariate Cox model could be constructed only for

the group of patients with N1 tumor stage, because, besides

fibroblast SMAD4 expression, histological differentiation

and tumor size were related to survival (logrank p = 0.07

and 0.03, respectively). This model is summarized in

Table 5. Although associated with a statistical significance

inferior to that of histological differentiation in this model

(p = 0.05, odds ratio 1.96 vs. p = 0.02 and odds ratio

2.16), fibroblast SMAD4 expression maintained its rela-

tionship to a worse outcome.

Long-term survival (C40 months) was more frequent in

patients with tumors showing negative stromal fibroblast

SMAD4 (p = 0.05). Patients with a survival of C40

months showed tumors of small size (\30 mm) and less

frequent lymph node metastases (data not shown)

(p = 0.04 and 0.08, respectively) as compared with

patients with \40 months survival. There was no signifi-

cant relationship between stromal fibroblast SMAD4 and

tumor size, tumor extrapancreatic invasion, or metastases

in any of the groups of patients (whole series, T1 or T2 or

T3 stage patients, N0 or N1 stage patients, patients with

C40 months survival, patients with \40 months survival,

patients with tumors [20 mm, or those with tumors

B20 mm).

Table 2 Cellular expression patterns of SMAD proteins in the epi-

thelial and stromal fibroblastic tumor components in pancreatic ductal

adenocarcinomas

p-SMAD1/5/8 p-SMAD2/3 SMAD4

Epithelial tumor cells

Number of positive

tumors

13 93 45

Percentage 13.5 % 96 % 49 %

Stromal fibroblasts

Number of positive

tumors

5 11 22

Percentage 5 % 11 % 23 %

J Hepatobiliary Pancreat Sci (2013) 20:294–302 297

123

Discussion

In this study, we found that SMAD proteins were expressed

in pancreatic ductal adenocarcinomas both in the epithelial

tumor cells and in stromal fibroblasts.

SMAD4 tumor epithelial expression was significantly

related to a low T TNM stage, which is consistent with the

well-established growth inhibitory function of TGF-beta

signaling. However, we did not find a significant relation-

ship between SMAD4 tumor epithelial expression and a

shorter postsurgical overall survival, although such a

relationship has been reported by several authors [8, 9]; this

lack of a relationship in our study was probably due to the

smaller number of tumors we studied [27]. In our series,

SMAD4 epithelial tumor cell expression was correlated

with the presence of an abundant stroma in the tumor,

suggesting that SMAD4 might also interfere with pro-

tumor effects, such as extracellular matrix production [1, 2,

28]. Although the relevance of this relationship might have

been limited by comparing SMAD4 expression on tissue

microarray with the characteristics of the stroma evaluated

on whole tumor section slides, we were confident of this

result, because SMAD4 protein tissue microarray analyses

have been shown to reflect the analysis of whole section

samples [29, 30]. Moreover, the relationship between

SMAD4 and stroma is not surprising, because SMAD4,

along with SMAD3, has been reported to interfere with

collagen stimulation [31], repression of matrix metallo-

protease 1 expression [1, 32–34], and the induction of

connective tissue growth factor deposition [35] in human

and mouse fibroblast cell lines.

Tumor epithelial expression of activated p-SMAD2/3

was observed in almost all tumors in the present study, in

contrast to p-SMAD1/5/8, which was observed in only

15 % of the cases. In a very limited number of tumors,

SMAD protein expression of the two axes coexisted. None

Fig. 2 Immunohistochemical expression of SMAD proteins in

pancreatic ductal adenocarcinomas. The left side column shows

tumors negative for p-SMAD1/5/8 (a), p-SMAD2/3 (d), and SMAD4

(g), the middle column shows tumors positive for these proteins (b, e,

and h for p-SMAD1/5/8, p-SMAD2/3, and SMAD4, respectively;

arrows) and the right side column shows stromal fibroblasts

expressing these antibodies (c, f, and i for p-SMAD1/5/8,

p-SMAD2/3, and SMAD4 respectively) (arrows) (940)

298 J Hepatobiliary Pancreat Sci (2013) 20:294–302

123

of these activated proteins were related to tumor morpho-

logical features when considering adenocarcinoma cell

expression.

The results of our study suggest that SMAD proteins

were also expressed in the stromal fibroblasts, although less

frequently than in the epithelial tumor component. Among

the studied SMAD proteins, SMAD4 was the most fre-

quently expressed in stromal fibroblasts. Interestingly,

when expressed in stromal fibroblasts, SMAD4 tended to

be correlated with a shorter postsurgical survival when

considering the whole series. This relationship is sustained

by our finding showing a statistically significant relation-

ship of fibroblast SMAD4 with an adverse outcome in the

groups of patients with T1 and N1 stage disease. Moreover,

in the Cox model constructed for the group of N1 stage

patients, SMAD4 fibroblast expression was related to an

adverse outcome, independently of histological differenti-

ation and tumor size, but with a lesser statistical signifi-

cance than tumor differentiation. The impact of fibroblast

SMAD4 on overall survival became significant in the group

of patients with survivals of C40 months, patients char-

acterized by less frequent lymph node metastases and by

the presence of tumors of small size. This same type of

impact was observed in the N1 TNM stage group, but not

in the T1 TNM stage group of patients. However, we did

not find any significant relationship of stromal fibroblast

SMAD4 to tumor size or tumor invasive characteristics

such as extrapancreatic invasion or lymph node metastases

which could explain this impact. Other stromal fibroblast-

expressed proteins such as alpha-smooth muscle actin [36]

and SPARC [29] have been reported to be related to or to

predict survival in pancreatic ductal adenocarcinomas, but

the underlying mechanism/process and whether fibroblast

proteins promote aggressive neoplastic cell behavior [16]

also remain difficult to establish. The divergent implica-

tions we observed for stromal fibroblast SMAD4, as

compared to tumor epithelial SMAD4, are difficult to

explain, but are probably related to the known cell type

specific functioning of TGF-beta signaling in tumor epi-

thelial malignant cells versus non-malignant stromal

fibroblast cells [37]. Therefore, both tumor and stroma

features could be considered when establishing and eval-

uating tumor prognostic markers.

Although less frequent, the expression of the activated,

phosphorylated forms of SMAD2/3 and SMAD1/5/8 pro-

teins is of interest because this confirms the presence of an

activated fibroblast component in the complex tumor

stromal microenvironment [16]. Whether the activation of

Table 3 Morphological characteristics of pancreatic adenocarcinoma with regard to SMAD protein expression in epithelial tumor cells

Tumor epithelial component

p-SMAD1/5/8 p-SMAD2/3 SMAD4

Negative

n = 83

Positive

n = 13

p Negative

n = 4

Positive

n = 93

p Negative

n = 46

Positive

n = 45

p

Tumor size 0.53 0.13 0.83

\30 mm 29 6 3 32 18 16

C30 mm 54 7 1 61 28 29

Tumor TNM 0.76 0.64 0.05

T1 and T2 50 7 3 54 23 32

T3 33 6 1 39 23 13

Differentiation 1 1 1

Well and moderate 57 9 3 63 31 31

Poor 26 4 1 31 15 14

Abundant stroma 1 0.60 <0.01

Absent 55 9 2 63 38 23

Present 28 4 2 30 8 22

Fibrotic focus 0.76 1 0.67

Absent 55 8 3 60 31 28

Present 28 5 1 33 15 17

Lymph node

metastasis

1 0.54 0.57

Absent 14 2 1 16 9 6

Present 69 11 3 77 37 39

Immunohistochemical data were not available in 3, 2, and 8 tumors for p-SMAD1/5/8, p-SMAD2/3, and SMAD4, respectively

Significant p values are shown in bold

J Hepatobiliary Pancreat Sci (2013) 20:294–302 299

123

SMAD proteins in stromal fibroblasts is the result of

epithelial-stromal interactions or whether it is part of an

epithelial-mesenchymal transition type process remains

difficult to establish [1]. Of note would be the correlation

we found between fibroblast p-SMAD2/3 expression and

the presence of a fibrotic focus, because the presence of a

fibrotic focus has been reported to correlate with aggressive

tumor behavior and to predict a shorter overall survival

[21, 38]. Processes such as collagen induction by SMAD3-

dependent TGF-beta signaling, as reported for skin fibro-

blasts [39], might be responsible for the relationship we

found. Interestingly, the lack of fibroblast SMAD1/5/8 was

correlated with a more aggressive tumor behavior related to

the presence of lymph node metastases. The relationship

we have found is consistent with reported data on SMAD1

SMAD5 knockout mice [40], which show a high degree of

metastases, and thus, with the known dual role of signaling

of the TGF-beta pathway [37].

In conclusion, the results of our study indicate that

SMAD proteins, including the SMAD1/5/8 and SMAD2/3

proteins in their activated, phosphorylated forms, are

expressed in pancreatic ductal adenocarcinomas both in

epithelial tumor cells and stromal fibroblasts. This finding

suggests that these molecules are related both to carcino-

genesis and to stroma development, the SMAD2/3 signal-

ing axis being dominant in pancreatic adenocarcinomas.

Moreover, stromal expression of SMAD4 was related to a

shorter overall survival. These results, obtained in a

Table 4 Stroma-related tumor characteristics and SMAD protein expression in stromal fibroblasts of pancreatic ductal adenocarcinomas

Tumor stromal fibroblastic component

p-SMAD1/5/8 p-SMAD2/3 SMAD4

Negative

n = 93

Positive

n = 5

p Negative

n = 86

Positive

n = 11

p Negative

n = 72

Positive

n = 22

p

Tumor size 1 0.32 0.13

\30 mm 34 2 33 2 31 5

C30 mm 59 3 53 9 41 17

Tumor TNM 0.64 0.19 0.13

T1 and T2 54 4 53 4 47 10

T3 39 1 33 7 25 12

Abundant stroma 1 1 0.21

Absent 61 3 56 7 50 12

Present 32 2 30 4 22 10

Fibrotic focus 0.16 0.01 0.61

Absent 59 5 60 3 48 13

Present 34 0 26 8 24 9

Lymph node

metastases

0.04 0.68 0.55

Absent 15 3 17 1 15 3

Present 78 2 69 10 57 19

Differentiation 1 0.32 0.30

Well or moderate 63 4 60 6 46 17

Poor 30 1 26 5 26 5

Immunohistochemical data were not available in 1, 2, and 5 tumors for p-SMAD1/5/8, p-SMAD2/3, and SMAD4, respectively

Significant p values are shown in bold

Fig. 3 Stromal fibroblast SMAD4 expression and postsurgical

survival in patients with pancreatic ductal adenocarcinomas. Stromal

SMAD4 expression was related to a shorter postsurgical overall

survival (p = 0.07). Kaplan–Meier plots: the dashed line represents

the patients whose tumors showed fibroblast SMAD4 expression

300 J Hepatobiliary Pancreat Sci (2013) 20:294–302

123

relatively large series of pancreatic ductal adenocarcino-

mas, warrant further studies on stroma determinants and

their impact on tumor biology and patients’ outcome.

Acknowledgments This work was performed with the financial

support of the Societe Nationale Francaise de Gastroenterologie. We

are grateful to Claude Lesty for help with statistical analysis. The

authors thank Professor J.F. Flejou and Professor P. Bedossa, as well

as Cecile Badoual, Vinciane Rebours, Cristina Goia, and Jacques

Paries. The authors thank Annick Mariot, Florence Gantelmo, and

Nathalie Colnot for help in collecting data; Sandrine Tessiore, Claire

Cichon, Veronique Creusot, Catherine Lefevre, and Sylvie Dubois for

assistance in the technical procedures; and Corine Zylberberg, Joelle

Runfola and Veronique Colmant, as well as Latoufa Lahaya, Julie

Balthase, Mathieu Bechani, Christelle Adam, Nadia Kemache,

Marielle Abbondandolo, and Serge Morard for administrative help.

References

1. Rahimi RA, Leof EB. TGF-beta signaling: a tale of two

responses. J Cell Biochem. 2007;15(102):593–608.

2. Truty MJ, Urrutia R. Basics of TGF-beta and pancreatic cancer.

Pancreatology. 2007;7:423–35.

3. Zhang B, Halder SK, Kashikar ND, Cho YJ, Datta A, Gorden DL,

Datta PK. Antimetastatic role of Smad4 signaling in colorectal

cancer. Gastroenterology. 2010;138:969–80.

4. Whitman M. Smads and early developmental signaling by the

TGFbeta superfamily. Genes Dev. 1998;12:2445–62.

5. Blackford A, Serrano OK, Wolfgang CL, Parmigiani G, Jones S,

Zhang X, Parsons DW, Lin JC, Leary RJ, Eshleman JR, Goggins

M, Jaffee EM, Iacobuzio-Donahue CA, Maitra A, Cameron JL,

Olino K, Schulick R, Winter J, Herman JM, Laheru D, Klein AP,

Vogelstein B, Kinzler KW, Velculescu VE, Hruban RH. SMAD4

gene mutations are associated with poor prognosis in pancreatic

cancer. Clin Cancer Res. 2009;15:4674–9.

6. Jones S, Zhang X, Parsons DW, Lin JC, Leary RJ, Angenendt P,

Mankoo P, Carter H, Kamiyama H, Jimeno A, Hong SM, Fu B,

Lin MT, Calhoun ES, Kamiyama M, Walter K, Nikolskaya T,

Nikolsky Y, Hartigan J, Smith DR, Hidalgo M, Leach SD, Klein

AP, Jaffee EM, Goggins M, Maitra A, Iacobuzio-Donahue C,

Eshleman JR, Kern SE, Hruban RH, Karchin R, Papadopoulos N,

Parmigiani G, Vogelstein B, Velculescu VE, Kinzler KW. Core

signaling pathways in human pancreatic cancers revealed by

global genomic analyses. Science. 2008;321:1801–6.

7. Wilentz RE, Su GH, Dai JL, et al. Immunohistochemical labeling

for Dpc4 mirrors genetic status in pancreatic adenocarcinomas. A

new marker of DPC4 inactivation. Am J Pathol. 2000;156:37–43.

8. Biankin AV, Morey AL, Lee CS, Kench JG, Biankin SA, Hook

HC, Head DR, Hugh TB, Sutherland RL, Henshall SM. DPC4/

Smad4 expression and outcome in pancreatic ductal adenocarci-

noma. J Clin Oncol. 2002;20:4531–42.

9. Tascilar M, Skinner HG, Rosty C, Sohn T, Wilentz RE, Offer-

haus GJ, Adsay V, Abrams RA, Cameron JL, Kern SE, Yeo CJ,

Hruban RH, Goggins M. The SMAD4 protein and prognosis of

pancreatic ductal adenocarcinoma. Clin Cancer Res. 2001;7:

4115–21.

10. Takahashi H, Oda T, Hasebe T, Aoyagi Y, Kinoshita T, Konishi

M, Nakagohri T, Inoue K, Takahashi S, Kawahira H, Monden M,

Ochiai A. Biologically different subgroups of invasive ductal

carcinoma of the pancreas: Dpc4 status according to the ratio of

intraductal carcinoma components. Clin Cancer Res. 2004;10:

3772–9.

11. Toga T, Nio Y, Hashimoto K, Higami T, Maruyama R. The

dissociated expression of protein and messenger RNA of DPC4 in

human invasive ductal carcinoma of the pancreas and their

implication for patient outcome. Anticancer Res. 2004;24:

1173–8.

12. Wang P, Fan J, Chen Z, Meng ZQ, Luo JM, Lin JH, Zhou ZH,

Chen H, Wang K, Xu ZD, Liu LM. Low-level expression of

Fig. 4 Stromal fibroblast SMAD4 expression and postsurgical

survival in patients with T1 and N1 stage pancreatic ductal

adenocarcinomas. Stromal SMAD4 expression was related to a

shorter postsurgical overall survival in patients with T1 stage (a) and

those with with N1 TNM stage (b) disease (p \ 0.01 and p = 0.04,

respectively). Kaplan–Meier plots: the dashed lines represent the

patients whose tumors showed fibroblast SMAD4 expression

Table 5 Cox model for the patients with N1 TNM pancreatic ductal

adenocarcinomas

p value 95% Confidence

interval

Odds

ratio

Tumor differentiation 0.02 1.12-4.14 2.16

Stromal fibroblast SMAD4 0.05 0.98-3.91 1.96

Tumor size (C30 mm) 0.16 0.81-3.33 1.65

J Hepatobiliary Pancreat Sci (2013) 20:294–302 301

123

Smad7 correlates with lymph node metastasis and poor prognosis

in patients with pancreatic cancer. Ann Surg Oncol. 2009;

16:826–35.

13. Kleeff J, Beckhove P, Esposito I, Herzig S, Huber PE, Lohr JM,

Friess H. Pancreatic cancer microenvironment. Int J Cancer.

2007;121:699–705.

14. Ottaviano AJ, Sun L, Ananthanarayanan V, Munshi HG. Extra-

cellular matrix-mediated membrane-type 1 matrix metallopro-

teinase expression in pancreatic ductal cells is regulated by

transforming growth factor-beta1. Cancer Res. 2006;66:7032–40.

15. Mahadevan D, Von Hoff DD. Tumor-stroma interactions in

pancreatic ductal adenocarcinoma. Mol Cancer Ther. 2007;6:

1186–97.

16. Neesse A, Michl P, Frese KK, Feig C, Cook N, Jacobetz MA,

Lolkema MP, Buchholz M, Olive KP, Gress TM, Tuveson DA.

Stromal biology and therapy in pancreatic cancer. Gut. 2011;60:

861–8.

17. Hruban RH, Boffetta P, Hiraoka N, Iacobuzio-Donahue C, Kato

Y, Kern SE, Klimstra DS, Kloppel G, Maitre A, Offerhaus GJA,

Pitman MB. Ductal adenocarcinoma of the pancreas. In: Bosman

F, Carneiro F, Hruban H, Theise N, editors. WHO classification

of digestive tumors. 4th ed. Lyon: IARC Press; 2010. p. 281–90.

18. Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A,

editors. American Joint Committee on Cancer AJCC Cancer

Staging Manual. 7th ed. New York: Springer; 2010. p. 241–51.

19. Hruban RH, Bishop Portman M, Klimstra DS. Tumors of the

pancreas. 6th ed. Washington, D.C.: American Registry of

Pathology; 2007.

20. Salaria SN, Illei P, Sharma R, Walter KM, Klein AP, Eshleman

JR, Maitra A, Schulick R, Winter J, Ouellette MM, Goggins M,

Hruban RH. Palladin is overexpressed in the non-neoplastic

stroma of infiltrating ductal adenocarcinomas of the pancreas, but

is only rarely overexpressed in neoplastic cells. Cancer Biol Ther.

2007;6:324–8.

21. Couvelard A, O’Toole D, Leek R, Turley H, Sauvanet A, Degott

C, Ruszniewski P, Belghiti J, Harris AL, Gatter K, Pezzella F.

Expression of hypoxia-inducible factors is correlated with the

presence of a fibrotic focus and angiogenesis in pancreatic ductal

adenocarcinomas. Histopathology. 2005;46:668–76.

22. Handra-Luca A, Lesty C, Hammel P, Sauvanet A, Rebours V,

Martin A, Fagard R, Flejou JF, Faivre S, Bedossa P, Ruszniewski

P, Couvelard A. Biological and prognostic relevance of mitogen

activated protein kinases in pancreatic adenocarcinoma. Pancreas.

2012;41:416–21.

23. Zhou S, Kinzler KW, Vogelstein B. Going mad with Smads.

N Engl J Med. 1999;341:1144–6.

24. Handra-Luca A, Condroyer C, de Moncuit C, Tepper M, Flejou

JF, Thomas G, Olschwang S. Vessels’ morphology in SMAD4

and BMPR1A-related juvenile polyposis. Am J Med Genet A.

2005;138A:113–7.

25. Handra-Luca A, Olschwang S, Flejou JF. Smad4 protein

expression and cell proliferation in colorectal adenocarcinomas.

Virchows Archiv 2011;459:511–9.

26. Montgomery E, Goggins M, Zhou S, Argani P, Wilentz R,

Kaushal M, Booker S, Romans K, Bhargava P, Hruban R, Kern S.

Nuclear localization of Dpc4 (Madh4, Smad4) in colorectal

carcinomas and relation to mismatch repair/transforming growth

factor-beta receptor defects. Am J Pathol. 2001;158:537–42.

27. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed.

New York: Wiley; 2000.

28. Haber PS, Keogh GW, Apte MV, Moran CS, Stewart NL,

Crawford DH, Pirola RC, McCaughan GW, Ramm GA, Wilson

JS. Activation of pancreatic stellate cells in human and experi-

mental pancreatic fibrosis. Am J Pathol. 1999;155:1087–95.

29. Infante JR, Matsubayashi H, Sato N, Tonascia J, Klein AP, Riall

TA, Yeo C, Iacobuzio-Donahue C, Goggins M. Peritumoral

fibroblast SPARC expression and patient outcome with resectable

pancreatic adenocarcinoma. J Clin Oncol. 2007;25:319–25.

30. Maitra A, Adsay NV, Argani P, Iacobuzio-Donahue C, De Marzo

A, Cameron JL, Yeo CJ, Hruban RH. Multicomponent analysis of

the pancreatic adenocarcinoma progression model using a pan-

creatic intraepithelial neoplasia tissue microarray. Mod Pathol.

2003;16:902–12.

31. Yuan W, Varga J. Transforming growth factor-beta repression of

matrix metalloproteinase-1 in dermal fibroblasts involves Smad3.

J Biol Chem. 2001;276:38502–10.

32. Chen SJ, Yuan W, Mori Y, Levenson A, Trojanowska M, Varga

J. Stimulation of type I collagen transcription in human skin

fibroblasts by TGF-beta: involvement of Smad 3. J Invest

Dermatol. 1999;112:49–57.

33. Ghosh AK, Yuan W, Mori Y, Varga J. Smad-dependent stimu-

lation of type I collagen gene expression in human skin fibro-

blasts by TGF-beta involves functional cooperation with p300/

CBP transcriptional coactivators. Oncogene. 2000;19:3546–55.

34. Zhang W, Ou J, Inagaki Y, Greenwel P, Ramirez F. Synergistic

cooperation between Sp1 and Smad3/Smad4 mediates trans-

forming growth factor beta1 stimulation of alpha 2(I)-collagen

(COL1A2) transcription. J Biol Chem. 2000;275:39237–45.

35. Holmes A, Abraham DJ, Sa S, Shiwen X, Black CM, Leask A.

CTGF and SMADs, maintenance of scleroderma phenotype is

independent of SMAD signaling. J Biol Chem. 2001;276:10594–

601.

36. Erkan M, Michalski CW, Rieder S, Reiser-Erkan C, Abiatari I,

Kolb A, Giese NA, Esposito I, Friess H, Kleeff J. The activated

stroma index is a novel and independent prognostic marker in

pancreatic ductal adenocarcinoma. Clin Gastroenterol Hepatol.

2008;6:1155–61.

37. Pardali K, Moustakas A. Actions of TGF-beta as tumor

suppressor and pro-metastatic factor in human cancer. Biochim

Biophys Acta. 2007;1775:21–62.

38. Watanabe I, Hasebe T, Sasaki S, Konishi M, Inoue K, Nakagohri

T, Oda T, Mukai K, Kinoshita T. Advanced pancreatic ductal

cancer: fibrotic focus and beta-catenin expression correlate with

outcome. Pancreas. 2003;26:326–33.

39. Leask A, Abraham DJ. TGF-beta signaling and the fibrotic

response. FASEB J. 2004;18:816–27.

40. Pangas SA, Li X, Umans L, Zwijsen A, Huylebroeck D, Gutierrez

C, Wang D, Martin JF, Jamin SP, Behringer RR, Robertson EJ,

Matzuk MM. Conditional deletion of Smad1 and Smad5 in

somatic cells of male and female gonads leads to metastatic

tumor development in mice. Mol Cell Biol. 2008;28:248–57.

302 J Hepatobiliary Pancreat Sci (2013) 20:294–302

123