Clinical & Experimental Metastasis

, Volume 30, Issue 4, pp 441–446

Podoplanin expressing cancer associated fibroblasts are associated with unfavourable prognosis in adenocarcinoma of the esophagus

Authors

    • Department of Surgery and Comprehensive Cancer Center Vienna (CCC-GET)Medical University of Vienna-General Hospital
  • Bettina Jesch
    • Department of Surgery and Comprehensive Cancer Center Vienna (CCC-GET)Medical University of Vienna-General Hospital
  • Martin F. Riegler
    • Department of Surgery and Comprehensive Cancer Center Vienna (CCC-GET)Medical University of Vienna-General Hospital
  • Florian Maroske
    • Department of Surgery and Comprehensive Cancer Center Vienna (CCC-GET)Medical University of Vienna-General Hospital
  • Katrin Schwameis
    • Department of Surgery and Comprehensive Cancer Center Vienna (CCC-GET)Medical University of Vienna-General Hospital
  • Gerd Jomrich
    • Department of Surgery and Comprehensive Cancer Center Vienna (CCC-GET)Medical University of Vienna-General Hospital
  • Peter Birner
    • Department of Surgery and Comprehensive Cancer Center Vienna (CCC-GET)Medical University of Vienna-General Hospital
    • Clinical Institute of Pathology, Medical University of Vienna
Research Paper

DOI: 10.1007/s10585-012-9549-2

Cite this article as:
Schoppmann, S.F., Jesch, B., Riegler, M.F. et al. Clin Exp Metastasis (2013) 30: 441. doi:10.1007/s10585-012-9549-2
  • 513 Views

Abstract

Overexpression of the mucin-type sialoglycoprotein podoplanin in cancer associated fibroblasts (CAFs) was recently shown to be associated with tumor progression, metastasis and poor prognosis in lung and breast cancer. Here we investigate the role of podoplanin expressing CAFs in esophagal adenocarcinoma (AC), its precursor lesions and metastases. Podoplanin expression was investigated immunohistochemically in 200 formalin-fixed, paraffin embedded specimens of invasive esophagal ACs, their corresponding metastases and 35 precursor lesions. Podoplanin expressing CAFs (CAF+) were observed in 22 % of patients with invasive AC, but not in precursor lesions. CAF+ correlated with tumor stage (p = 0.004), lymphovascular tumor invasion (p = 0.018) and lymph node metastasis (p = 0.0016). Patients with CAF+ had a significant shorter disease free and overall survival (p < 0.05, Cox regression). Podoplanin expressing CAFs were only rarely observed in lymph node and distant metastases, as well as in local recurrences of ACs. Podoplanin expression in AC tumor cells was seen in only four cases. In around 20 % of patients with esophagal AC, podoplanin expressing CAFs are evident, defining a high risk subgroup. In these patients, podoplanin expressing CAFs might represent new therapeutical targets.

Keywords

PodoplaninEsophagal cancerCancer associated fibroblastsPrognosis

Introduction

Malignant tumors consist of two cell types: a collection of mutated cells growing uncontrolled–the so called tumor cells-and various other co-opted non-malignant cell types including fibroblasts, blood and lymphatic endothelial cells and immune cells [1]. The cytokine mediated tumor-stromal interaction has been shown to play a critical role in tumor progression and metastasis [24]. Activated fibroblasts that are recruited into cancer tissue (the so called cancer-associated fibroblasts, CAFs) are supposed to promote tumor progression by secreting different cytokines and growth factors [5, 6].

The mucin-type sialoglycoprotein podoplanin, a selective marker of lymphatic endothelium, was recently shown to be overexpressed in CAFs of adenocarcinomas of the lung and breast cancer where it was associated with tumor progression, metastasis and poor prognosis [79].

Esophageal cancer is one of the most common cancers worldwide with significant growing incidences of adenocarcinomas (AC) over the past years. Despite discrete improvements in treatment in recent days, long-term survival rates in patients with advanced disease are still far below 10 %.

Aim of this study was evaluate the role and clinical impact of podoplanin expressing CAFs in primary and metastatic AC of the esophagus.

Patients and methods

Patients

All patients who underwent surgical treatment of esophagal AC between the years 1992 and 2011 at the Department of Surgery Medical University of Vienna were included into this retrospective study. Institutional review board approval was obtained. Further, tissues from corresponding lymph node and/or distant metastases as well as local recurrences when available were investigated. In addition, samples of Barrett’s mucosa with our without dysplasia were included.

Immunohistochemistry

Immunohistochemistry was performed on 3 μm thick paraffin sections applying the Benchmark Ultra immunostainer (Ventana) according to the manufacturer’s instructions. Podoplanin expression was detected with the mouse moncolonal anti-podoplanin antibody D2-40 (catalog number 760–4,395, Venatana, Tucson, AZ, ready to use). In three randomly selected cases vimentin with antibody clone SP20 (Thermo Fisher Scientific, Fremont, CA; 1:300) and alpha- smooth muscle actin with antibody clone 1A4 (DAKO, Glostrup, Denmark, 1:200) was immunostained.

Analysis was performed by two independent observers (S.F.S. and P.B.) In cases with divergent results, slides were reassessed using a multiheaded microscope.

A specimen was scored as positive for podoplanin expression in CAFs (CAF+), when ≥10 % of the fibromatous tumor stroma showed a distinct staining reaction [9]. Lymphatic vessels served as internal positive control. Lymphatic vessel invasion (LVI) of tumor cells in primary tumors was scored as positive, if tumor cells were seen in at least one podoplanin decorated vascular space.

A specimen was considered as showing podoplanin expression in tumors cells if positive cytoplasmic staining reaction was seen, regardless of percentage or intensity.

Statistics and survival-analysis

All statistics were calculated using statistical package for the social sciences (SPSS®) 19.0 software (SPSS Inc., Chicago, IL, USA).

Mann–Whitney U tests, McNemar and Chi square tests were used as appropriate. Disease free survival (DFS) was defined as the time from primary surgery to first evidence of recurrent disease. Overall survival (OS) was defined as the time between primary surgery and death from esophagal cancer. Survival until the end of the observation period or death from another cause than esophagal cancer was considered as a censored oberservation.

Univariate analysis of survival was performed using the log-rank test, multivariable survival analysis using Cox regression. Tumor stage and lymph node status according to UICC 7th edition, patients age (<60 vs. >60 years), residual tumor after surgery, histological grading and podoplanin expression in CAFs (CAF+ vs. CAF−) were included into the regression model. A two tailed p value of ≤0.05 was considered as significant.

Results

Patients

Finally 200 patients with invasive AC, 14 patients with Barrett’s mucosa without dysplasia, 4 patients with low grade dysplastic Barrett’s mucosa and 17 patients with high grade dysplasia were included into the study. In patients with premalignant lesions, resection had been performed due other than oncologic indication (e.g. esophagal perforation or end stage achalasia).

In 66 patients tissue from lymph node metastases, in 23 from distant metastases and in 7 from local recurrences was available and included into the analysis.

Sixty-eight patients with AC had received neoadjuvant chemotherapy. Of the 200 cancer patients, 166 were male (83 %), and 34 (17 %) were female. The mean age was 65 ± 10 years (standard deviation).

Histopathological and clinical data of patients are complied in Table 1.
Table 1

Shows the clinical and pathological data of the investigated patients in relation to the presence of podoplanin expressing cancer associated fibroblasts (CAFs)

Variable

CAF+

n = 44 (%)

CAF−

n = 166 (%)

p value

Tumor stage

  

0.004

 pT1a (n = 13)

0

13 (100)

 

 pT1b (n = 22)

3 (13.6)

19 (86.4)

 

 pT2 (n = 58)

10 (17.2)

48 (82.8)

 

 pT3 (n = 99)

28 (28.3)

71 (71.7)

 

 pT4 (n = 8)

3 (37.5)

5 (62.5)

 

Lymp node status

  

0.016

 pN0 (n = 67)

9 (13.4)

58 (86.6)

 

 pN1 (n = 37)

6 (16.2)

31 (83.8)

 

 pN2 (n = 37)

11 (29.7)

26 (70.3)

 

 pN3 (n = 47)

14 (29.8)

33 (70.2)

 

Histological grading

n.s.

 

n.s

 G1 (n = 7)

1 (14.3)

6 (85.7)

 

 G2 (n = 85)

15 (17.6)

70 (82.4)

 

 G3 (n = 108)

28 (25.9)

80 (74.1)

 

Complete resection (R0)

  

n.s

 Yes

35 (20.7)

134 (79.3)

 

 No

9 (29)

22 (71)

 

Neoadjuvant chemotherapy

  

n.s.

 No (n = 132)

32 (24.2)

100 (75.8)

 

 Yes (n = 68)

12 (17.6)

56 (82.4)

 

Patients age

n.s.

 

n.s.

 <=60 years (n = 64)

14 (21.9)

50 (78.1)

 

 >60 years (n = 136)

30 (22.1)

106 (77.9)

 

LVI

p = 0.018

 

0.018

 No (n = 113)

18 (15.9)

95 (84.1)

 

 Yes (n = 87)

26 (29.9)

61 (70.1)

 

For each variable, the number of podoplanin CAF+ and CAF− samples is given, the p value indicates if the various variables associate with CAF+. A significant association of podoplanin+ CAFs with higher tumor and lymph node stage and with lymphovascular invasion of tumor cells (LVI) was observed

CAF+ positive for podoplanin expressing cancer associated fibroblasts, CAF− negative for podoplanin expressing cancer associated fibroblasts, n.s. non significant

Podoplanin expressing CAFs in primary esophagal cancer

In precursor AC lesions no podoplanin expressing fibroblasts were observed.

Forty-four (22 %) patients with invasive AC were considered as CAF+ (see Fig. 1a, b). In three randomly selected ACs considered as positive for podoplanin expressing CAFs, immunostaining for vimentin and smooth muscle actin was performed. The vast majority of cells in podoplanin positive stroma areas showed distinct positive staining reaction with both antibodies, identifying them as CAFs (see Fig. 2 e, f).
https://static-content.springer.com/image/art%3A10.1007%2Fs10585-012-9549-2/MediaObjects/10585_2012_9549_Fig1_HTML.gif
Fig. 1

a, b Specimen of esophagal adenocarcinoma with a large number of podoplanin expressing cancer associated fibroblasts (CAF+), surrounding the tumor formations (T). Note the stronger stained lymphatic vessels (Ly). (Immunoperoxidase, original magnification ×200 and ×630, respectively). c Lymphovascular invasion of tumor cells (T) into podoplanin decorated vascular spaces (arrow). The surrounding fibroblasts show distinct cytoplasmic podoplanin expression (CAF+arrow). Note the podoplanin negative blood vessel with erythrocytes in the lower part of the picture. (Immunoperoxidase, original magnification ×400). d Plot showing the correlation between patients with CAF+ and lymphovascular invasion (p = 0.018, Chi square test). e, f Consecutive slides with vimentin (e) and podoplanin (f) positive stroma (see arrows) areas showing distinct positive staining reaction with both antibodies, identifying them as CAFs. (both Immunoperoxidase, original magnification ×200)

https://static-content.springer.com/image/art%3A10.1007%2Fs10585-012-9549-2/MediaObjects/10585_2012_9549_Fig2_HTML.gif
Fig. 2

Kaplan Meier curves showing the cumulative disease-free survival (upper curve) and cumulative overall survival (lower curve) of patients with podoplanin-expressing CAFs (CAF+) and without (CAF−) in esophagal adenocarcinoma (p < 0.01, respectively, log-rank test). Note the 5, 10 and 15 years survival rates

Positive statistical correlations between CAF-status and tumor stage (p = 0.004, Mann–Whitney U test) and lymph node staging (p = 0.016, Mann–Whitney U test) were observed (Table 1).

CAF+ cases showed significantly more often LVI of tumor cells compared to CAF− cases (59 % vs. 39 %; p = 0.018, Chi square test, odds ratio 2.25, 95 % interval of confidence 1.138–4.447) (see Fig. 1c, d).

No association of CAF+ with the patients sex or age, application of neoadjuvant chemotherapy nor histological tumor grading was seen.

Podoplanin expressing CAFs in corresponding metastases

In 2 out of 66 investigated lymph node metastases (3 %) podoplanin expressing CAFs were observed. A significant disconcordance of CAF scoring between the primary tumor and lymph node metastases was found (p < 0.001, McNemar test): The 2 patients with CAF+ lymph node metastases were also CAF+ in the primary tumors, but in 17 patients with CAF+ primary tumors, lymph node metastases were CAF−. None of the distant metastases and one local recurrence was scored CAF+.

Podoplanin expression in tumor cells

No podoplanin expression was observed in AC precursor lesions. In cancer cells of primary tumors and metastases, podoplanin expression was rare: Only four patients showed a weak to moderate cytoplasmic podoplanin expression in their primary tumors (one case 1 %, two cases 5 % and one case 70 % of tumor cells). No podoplanin expression in tumor cells of any metastatic site was observed.

Survival analysis

The mean observation time was 54 ± 4 months (standard error). During this period of time, 92 patients with AC (46 %) developed recurrent disease, and 80 patients died from their cancer.

CAF+ was associated with shorter DFS and OS (p < 0.01, respectively, log-rank test) in esophagal adenocarcinoma (see Table 2 and Fig. 2).
Table 2

Disease free and overall survival of 200 patients with esophagal adenocarcinoma (AC) with regard to clinical and pathological parameters and the presence of podoplanin expressing CAFs

Factor

p value univariate

p value multivariable

Relative risk

95 % Confidence interval

Disease free survival

 Patients age (≤60 vs. >60)

0.262

0.074

 Tumor stage pT

<0.001

0.046

1.406

1.007–1.965

 Tumor stage pN

<0.001

<0.001

1.859

1.455–2.376

 R0 resection

0.001

0.516

 Histological grading

0.001

0.937

 Podoplanin+ CAFs

0.004

0.034

1.697

1.042–2.764

Overall survival

 Patients age

0.248

0.076

 Tumor stage pT

<0.001

0.15

 Tumor stage pN

<0.001

<0.001

1.87

1.431–2.445

 R0 resection

0.002

0.629

 Histological grading

0.003

0.902

 Podoplanin+ CAFs

0.001

0.021

1.843

1.097–3.096

For each variable, univariate (log-rank test), and multivariable p values (Cox regression) are shown, as well as its relative risk and 95 % interval of confidence in multivariable analysis. The presence of podoplanin+ CAFs is a strong prognostic factor in esophagal AC

At multivariable analysis including patient’s age, tumor and lymph-node staging as well as histological grading and the CAF−status, CAF+ was associated with shorter DFS and OS. Mean DFS was 89 ± 12 months in patients scored CAF−, and 42 ± 10 months in those with CAF+. Mean OS was 105 ± 11 months for CAF− and 42 ± 10 months for CAF+ (DFS: p = 0.034 and OS: p = 0.021, Cox regression) (Table 2).

No influence of podoplanin expression in tumor cells on DFS and OS was seen (p > 0.05, log-rank test).

Discussion

While in the last decades cancer research has been mainly focused on the tumor cells themselves, increasing evidence exists that cancer progression and dissemination is greatly influenced by the tumor microenvironment [10]. Especially activated cancer associated fibroblasts (CAFs) are known to influence tumor growth and cancer cell dissemination by producing and secreting various cytokines and growth factors [11]. Additionally, stimulated CAFs can distribute extracellular matrix proteins and collagens and hence influence the cell–matrix and cell–cell adhesion [12].

The transmembrane sialoglycoprotein podoplanin is overexpressed in cancer cells of various human malignancies, and is associated with dismal prognosis in several tumors [1319].

Recently, podoplanin expression was also found in CAFs of various tumor types suggesting a special role for these subtypes of CAFs in cancer progression [79, 2023].

In the present study we show that podoplanin expressing CAFs (CAF+) are found in about 20 % of esophagal adenocarcinomas. In this subgroup of patients they are associated with aggressive tumor biology (e.g. lymphovascular invasion and lymph node metastasis) and dismal prognosis. These results are in good correlation with observations made in lung and breast cancer where CAFs+ are strongly predictive for poor prognosis [79, 20, 21, 24], and intrahepatic cholangiocarcinoma, where CAF+ are associated with tumor progression and lymph node metastasis [24].

On the contrary, in cervical and colorectal cancer, CAFs+ are associated with similar [22] or even improved patient’s outcome, respectively [23]. Therefor the role of CAFs+ seems to be in some way tumor-type specific. In metastatic tumor disease, the situation seems to be comparable. We observed CAFs in only 3 % of lymph node metastases and none of the distant metastases, which is not in concordance to the findings in lung adenocarcinoma: In a recent study, CAFs+ were found in around 40 % of lymph node metastases of lung adenocarcinomas, correlating with the CAFs+ in the primary tumors and indicating further dismal prognosis [25].

In contrast to podoplanin-expression in tumor cells where it significantly influences cell migration and invasion via cellular membrane rearrangements in vivo and in vitro [26], the role of podoplanin expression in CAFs in the tumor situation is less well understood.

It is believed that podoplanin in fibroblasts acts as a tumor formation enhancer during the initiation of tumorigenesis [7]. The activity of RhoA, cellular key regulator of processes such as cytokinetics, cytoskeletal regulation, and cell migration activity, is increased in podoplanin expressing fibroblasts, and is supposed to promote tumor progression by biomechanical remodelling of the microenvironment [27]. Interestingly we did not observe podoplanin expressing CAFs in esophagal AC precursor lesions, indicating that the upregulation of podoplanin in CAFs is limited to invasive disease. However, the exact mechanisms how podoplanin expressing CAFs interact with cancer cells to promote tumor progression have to remain unclear.

Recently, the phenomenon of endothelial-mesenchymal transition (EndMT), a process in which proliferating endothelial cells are known to undergo a phenotypic conversion into fibroblasts, arrested growing attention in the field of cancer research [28]. Tagged endothelial cells brought into a pancreatic carcinoma model were shown to be the major source of CAFs at the tumors invasive front [29]. These CAFs facilitated tumor progression and enhanced tumor invasion. With podoplanin specifically expressed on lymphatic endothelial cells, one could hypothesize that proliferating lymphatic endothelial cells may become CAFs through an EndMT. This would indicate that in CAF+ esophagal ACs possibly a large number of CAFs is recruited from the lymphatic endothelia [30, 31].

The correlation between CAF+ and the presence of LVI observed in our patient cohort can likely be used to support this theory, since obviously tumors with podoplanin positive CAFs do more easily invade the lymphatic vessel system, whose endothelial cells also express podoplanin.

Various therapeutic targets associated with CAFs have been described over the last years. The idea of selectively inhibiting the pro-tumorous effects of CAFs has already led to the development of first, selectively CAF− targeting therapeutical approaches [26, 32, 33].Our observation of a clear clinical role of podoplanin expressing CAFs in patients with invasive esophagal AC, provides indirect evidence that therapies targeting those cells might be of benefit for patients.

In summary our data show that podoplanin expressing CAFs are evident in about 20 % of esophagal ACs defining a subgroup of patients with high risk, warranting further investigations.

Conflict of interest

The authors declare to have no conflicts of interest.

Copyright information

© Springer Science+Business Media Dordrecht 2012