Virchows Archiv

, Volume 456, Issue 5, pp 473–482

Impact of podoplanin expression in oral squamous cell carcinoma: clinical and histopathologic correlations

Authors

    • Department for Oral and Cranio-maxillo and Facial Plastic SurgeryUniversity of Cologne
    • Centre of Integrated Oncology (CIO) Cologne-Bonn
  • Martin Scheer
    • Department for Oral and Cranio-maxillo and Facial Plastic SurgeryUniversity of Cologne
    • Centre of Integrated Oncology (CIO) Cologne-Bonn
  • Uta Drebber
    • Department of PathologyUniversity of Cologne
    • Centre of Integrated Oncology (CIO) Cologne-Bonn
  • Lutz Ritter
    • Department for Oral and Cranio-maxillo and Facial Plastic SurgeryUniversity of Cologne
    • Centre of Integrated Oncology (CIO) Cologne-Bonn
  • Joachim E. Zöller
    • Department for Oral and Cranio-maxillo and Facial Plastic SurgeryUniversity of Cologne
    • Centre of Integrated Oncology (CIO) Cologne-Bonn
Original Article

DOI: 10.1007/s00428-010-0915-7

Cite this article as:
Kreppel, M., Scheer, M., Drebber, U. et al. Virchows Arch (2010) 456: 473. doi:10.1007/s00428-010-0915-7
  • 289 Views

Abstract

Cervical lymph node metastases reduce the overall survival of patients with oral squamous cell carcinoma (OSCC) and require a neck dissection. However, elective management of a clinical N0 neck remains a controversial topic, as there are no reliable factors available predicting cervical lymph node metastases. Recent studies suggest an impact of podoplanin expression on metastatic spread to the cervical lymph nodes. Our aim was to investigate the influence of podoplanin expression on prognosis and metastatic lymphatic spread. In our retrospective study, podoplanin expression was examined in a set of 80 patients with OSCC by immunhistochemistry. We analysed associations between the level of podoplanin expression and various clinicopathologic parameters. In 67 patients (84%), podoplanin was expressed on the tumour cells. Nineteen patients (24%) showed high levels of expression. The 5-year overall survival (31%) for patients with high levels of podoplanin expression was significantly lower (p < 0.001) than for patients with low and moderate expression of podoplanin (93% and 65%, respectively). There was an association between podoplanin expression and the frequency of cervical lymph node metastases. Cervical lymph node metastases were found in 79% of the patients with high podoplanin expression, while patients with weak podoplanin expression had metastases in only 22% (p < 0.001). None of the 13 patients without podoplanin expression had cervical lymph node metastases. We concluded that podoplanin is expressed frequently in OSCC and that podoplanin expression correlates with cervical lymph node metastases and clinical outcome.

Keywords

PodoplaninImmunhistochemistryOral cancerPrognosisCervical lymph node metastases

Introduction

Carcinoma of the oral cavity is the sixth most common cancer worldwide with an annual incidence of more than 500,000 cases [1]. The incidence is influenced by lifestyle, habits, demographic and genetic factors [2]. In 2009, the estimated incidence of oral and oropharyngeal cancer in the USA was 35,000, representing 2.4% of all newly diagnosed cancers and resulting in 1.4% of cancer-related deaths [3].

Despite improved therapeutic procedures, the prognosis remains poor [4, 5] yielding 5-year overall survival rates from 45% to 58% [6, 7]. This fact can be attributed to several factors: Failure to respond to available chemotherapy, late presentation of the lesions and lack of suitable markers for early detection [8, 9]. Even in the earliest stages of disease, oral squamous cell carcinoma (OSCC) may vary dramatically in treatment response and recurrence. Consequently, a significant number of patients with early stage oral cancer die of the disease. Thus, there is a strong need to understand the development of OSCC and to identify new and additional prognostic factors that can improve the clinical staging system and establish appropriate treatment modalities [10, 11].

Cervical lymph node metastases are widely accepted as the strongest independent prognostic factor in patients with OSCC [1216]. There is a positive correlation between the size of the primary tumour and the appearance of lymph node metastases, although some patients with small tumours have lymph node metastases, whereas patients with large tumours do not develop lymph node metastases [17]. A strong positive correlation was also reported between tumour thickness and survival [18, 19]. Study results suggest that molecular features of the tumour play an essential role in determining the aggressiveness of the tumour [2024].

Recent studies identified podoplanin as a potential predictor of lymphatic dissemination in OSCC [1, 23, 25, 26]. Podoplanin is a mucin-type transmembrane glycoprotein that is specifically expressed in lymphatic endothelial cells but not in blood endothelial cells [27, 28]. Podoplanin knockout mice have lymphatic defects associated with diminished lymphatic transport, congenital lymphedema and dilation of lymphatic vessels [29], suggesting that, under physiological conditions, podoplanin plays an important role in regulating peripheral lung cell proliferation and lymphatic vascular development. Apart from head and neck squamous cell carcinoma, particularly in oral cancer [23, 30], podoplanin is expressed in carcinoma of the skin [31], lung [32], uterus [33], colon [34], oesophagus [35, 36], in mesothelioma [37] in astrocytoma [38, 39]. Investigations have shown that podoplanin, the oncofetal protein M2A recognised by the D2-40 antibody and the type I alveolar cell marker hT1ά-2 are identical proteins [25].

In the past, podoplanin has been used frequently to assess intra- and peritumoural lymphatic vessel density in oral squamous cell carcinoma, which is correlated with metastatic spread to the lymph nodes and a poor prognosis [28, 40, 41]. Recent studies have shown that podoplanin is also expressed in oral dysplastic and hyperplastic lesions with a risk of cancer development raising the question whether podoplanin may have further biological functions in premalignant oral lesions [23, 25, 30]. It has been reported that podoplanin acts as a mediator of tumour cell invasion in the absence of epithelial cell dedifferentiation [42].

In this study, we used the monoclonal antibody D2-40, which specifically stains podoplanin to assess the expression of podoplanin in OSCC cells and to determine potential associations between the expression patterns and clinicopathologic characteristics, especially overall survival, disease-specific survival and the occurrence of cervical lymph node metastases. So far, no reliable parameters have been found to assess the risk of lymph node metastases in a clinically N0-situation of the neck [4345].

Materials and methods

Patients and specimens

The retrospective study included 80 treatment-naive patients with a biopsy proven primary squamous cell carcinoma of the oral cavity of stages I–IV who were treated at the Department of Cranio-maxillo-facial Surgery at the University of Cologne between October 2002 and June 2005. All patients underwent primary surgery including a neck dissection. Histopathologic staging (pTNM) was performed according to the 6th edition of the UICC [46]. All patients with stage II, III, and IV additionally received an adjuvant concurrent chemoradiation with carboplatin AUC 5 in weeks 1 and 5 and a radiation dose of 61–65 Gy. Patients who did not receive the adjuvant scheme were excluded from this study. The patients' clinical characteristics are listed in Table 1.
Table 1

Patient and tumour characteristics

Patients' characteristics, N (%)

Patients

80

 Male

51 (63.7%)

 Female

29 (36.3%

Age (years)

 Mean ± SD

61.5±

 Median

61.8

Smoking

 Yes

63 (78.8%)

 No

17 (21.2%)

Alcohol

 Yes

53 (66.2%)

 No

27 (33.8%)

Tumour site

 Floor of mouth

35 (43.8%)

 Tongue

18 (22.5%)

 Upper and lower gingiva

14 (17.5%)

 Hard palate

5 (7.5%)

 Cheek

8 (10.0%)

pT Stage

 T1

24 (30.0%)

 T2

27 (33.8%)

 T3

6 (7.5%)

 T4

23 (28.7%)

pN Stage

 N0

49 (61.3%)

 N1

9 (11.3%)

 N2

22 (27.5%)

pUICC Stage

 I

23 (28.7%)

 II

13 (16.3%)

 III

8 (10.0%)

 IV

36 (45.0%)

Podoplanin Expression

 No Podoplanin

13 (16.3%)

 Weak

18 (22.5%)

 Moderate

30 (37.5%)

 High

19 (23.8%)

Tissue processing and immunohistochemical analysis

Formalin-fixed, paraffin-embedded specimens were obtained from the operation theatre and handled in the Department of Pathology, at the University of Cologne for further processing. The tissue samples were obtained from the peripheral invasion front of the tumour. Clinicopathologic parameters were obtained from the medical charts including the histopathologic and surgical reports. Follow-up data were gathered from a combination of chart review and the local government office for registration of residents. The average follow-up time was 42.9 months.

Immunhistochemistry was performed using the avidin–biotin–peroxidase complex technique (ABC-technique) [30, 47]. Paraffin sections measuring 4 μm in thickness of the excised tumours were placed on positively charged glass sides. The slides were deparaffinised and rehydrated in descending concentrations of ethanol. For antigen unmasking, the slides were steamed with 10 mmol/L citrate buffer, pH 6.0 (Dako Cytomation, Carpinteria, CA) for 20 min. After cooling down for 45 min the slides were blocked against endogenous peroxidase by incubation in 3% hydrogen peroxide for 10 min, followed by incubation with 10% horse serum for 30 min at room temperature. The slides were then incubated with the monoclonal antibody D2-40 diluted 1:100 (Vector Laboratories, Burlingame, CA) at 4°C overnight followed by signal development processes according to the manufacturer's protocol (Vectastatin Elite ABC kit, Vector Laboratories, Burlingame, CA). Counterstaining of the slides was achieved with Mayer's haematoxylin (Dako Cytomation). Expression of podoplanin in lymphatic endothelial cells in the stroma served as a positive internal control. Only one slide for each case was analysed.

In consideration of a previously published scheme to evaluate immunohistochemical samples [48], the epithelial podoplanin staining was quantified on a scale from 0 (negative) to 3 (strong staining) in consensus of two investigators (U.D. and M.S.) without knowledge of the patient's clinical history. The final score took into account both the intensity of the staining and the amount of positive cells according to the following scheme: both were classified from 0 to 3 (0 = no podoplanin, 1 = weak expression, 2 = moderate expression, 3 = high expression), added up and divided by 2. If the final score was not an even number, it was assigned to the lower group, for example a score of 1.5 was regarded as weak expression.

Statistical analysis

Associations between podoplanin expression and clinicopathologic variables were assessed using the Wilcoxon rank sum test for continuously distributed variables and the χ²-test and Fisher exact test for categorical variables.

The Kaplan–Meier survival analysis method was used to estimate the events of interest for overall survival (time interval from beginning of primary therapy until death. Patients who did not die were censored at their last date of follow-up) and disease-specific survival (time interval from beginning of primary therapy until death through OSCC. Patients who died of other reasons or did not die were censored at their last date of follow-up) [49]. The log rank test was used to compare survival times among patients with different characteristics. P values of less than 0.05 were considered as significant and printed in bold. A Cox proportional hazard model was calculated to estimate the prognostic impact of patient and tumour-related factors in a multivariate analysis on survival, which were significant in univariate analysis [50].

All calculations were conducted using SPSS 17.0

Results

In OSCC, two different patterns of podoplanin expression emerged (Fig. 1): the first type showed a diffuse expression of podoplanin in the tumour cells. The membrane of the tumour cells showed a stronger staining than the cytoplasm. In the second type, focal expression of podoplanin in the peripheral layer of the tumour proliferation area was found.
https://static-content.springer.com/image/art%3A10.1007%2Fs00428-010-0915-7/MediaObjects/428_2010_915_Fig1_HTML.jpg
Fig. 1

Patterns of podoplanin expression: Top left: tumour without podoplanin expression at 200× (score 0), top right: tumour with weak podoplanin expression at 200× (score 1), bottom left: moderate expression of podoplanin 100× (score 2), bottom right: high podoplanin expression at 200× (score 3)

Thirteen tumours did not express podoplanin at all.

Podoplanin expression and associations with clinicopathologic parameters in patients with primary OSCC

In 13 tumours (16.3%), no sign of podoplanin expression was observed; weak expression was found in 18 patients (22.5%). A moderate expression was visualised in 30 patients (37.5%), while, in 19 cases (23.8%), the tumours exhibited a high level of podoplanin expression. As shown in Table 2, age (p = 0.420) and gender (p = 0.832) did not show an association with podoplanin expression.
Table 2

Associations of podoplanin and patient characteristics

Associations of podoplanin and patient characteristics

 

No podoplanin

Weak expression

Moderate expression

High expression

P

Age (years)

    

0.420

 Mean ± SD

64.7 ± 12.1

61.6 ± 13.4

61.9 ± 10.1

58.8 ± 13.8

 

 Median (min, max)

64 (44, 86)

59 (37, 91)

63 (38, 82)

62 (30, 85)

 

Gender (%)

    

0.832

 Male

9 (17.6%)

10 (19.6%)

19 (37.3%)

13 (25.5%)

 

 Female

4 (13.8%)

8 (27.8%)

11 (37.9%)

6 (20.5%)

 

Smoking

    

0.161

 No

1 (5.9%)

2 (11.8%)

10 (58.8%)

4 (23.5%)

 

 Yes

12 (19.1%)

16 (25.4%)

20 (31.7%)

15 (23.8%)

 

Alcohol

    

0.434

 No

2 (7.5%)

7 (25.9%)

12 (44.4%)

6 (22.2%)

 

 Yes

11 (20.8%)

11 (20.8%)

18 (33.9%)

13 (24.5%)

 

pT stage (%)

    

0.016

 T1 & T2

12 (23.1%)

14 (26.9%)

18 (34.6%)

8 (15.4%)

 

 T3 & T4

1 (3.6%)

4 (14.3%)

12 (42.8%)

11 (39.3%)

 

pN Stage (%)

    

<0.001

 N0

13 (26.5%)

14 (28.6%)

18 (36.7%)

4 (8.2%)

 

 N1 & N2

0

4 (12.9%)

12 (38.7%)

15 (48.4%)

 

EPS

    

0.231

 EPS-negative

13 (19.4%)

16 (23.9%)

23 (34.3%)

15 (22.4%)

 

 EPS-positive

0

2 (15.4%)

7 (53.8%)

4 (30.8%)

 

pUICC Stage

    

<0.001

 UICC I & II

12 (32.4%)

11 (29.7%)

11 (29.7%)

3 (8.2%)

 

 UICC III & IV

1 (2.3%)

7 (16.3%)

19 (44.2%)

16 (37.2%)

 

The χ² test revealed a strong association between T stage and podoplanin expression (p = 0.016). In 50.0% of the patients with stage T1 and T2 the tumour cells had no or weak podoplanin expression, while only five patients (17.9%) with an advanced primary tumour of stage T3 and T4 showed no or weak podoplanin expression. Our analyses showed that 39.3% patients with locally advanced tumours of stage T3 and T4 had a strong podoplanin expression, which was only diagnosed, in 15.4% of the patients with tumours of stage T1 and T2.

The presence of lymph node metastases had the strongest association with podoplanin expression (p < 0.001). Among the 31 patients with lymph node metastases, 15 (48.4%) had high levels of podoplanin expression in the primary tumour compared with 8.2% of the patients with no detectable cervical lymph node metastases. Extra-capsular spread (ECS) of the tumour in patients with cervical lymph node metastases tended to be associated with podoplanin expression in the primary tumour. In 11 out of 13 patients with ECS, a moderate or high level of podoplanin expression was observed; however, the association was not significant (p = 0.231).

In addition, podoplanin expression was significantly associated with a higher UICC stage (p < 0.001). In 81.4% of the patients who presented with stage III and IV, a moderate or high level of podoplanin expression was found in the primary tumour, while only 37.9% patients with UICC stage I and II showed moderate or high levels of podoplanin expression.

Smoking (p = 0.161) and alcohol consumption (p = 0.434) did not have a significant association with the level of podoplanin expression in the tumour cells.

Parameters predicting overall survival and disease-specific survival in patients with OSCC

One of our main aims was determine whether podoplanin expression in primary tumours of OSCC is a feasible parameter to predict the patients' outcome with respect to survival. Using the Kaplan–Meier product limit method we calculated the 5-year survival rates for overall survival and disease-specific survival with regard to the podoplanin expression. Table 3 presents the 5-year survival rates for overall survival and disease-specific survival.
Table 3

Identification of prognostic factors by univariate analysis (log rank test)

Univariate analysis of prognostic factors

Parameter

5-year OS

P value

5-year DSS

P value

Age

 

0.067

 

0.116

 ≤61 years (lower half of median)

72.2%

 

84.8%

 

 >61 years (upper half of median)

50.9%

 

55.2%

 

Gender

 

0.750

 

0.567

 Male

64.8%

 

69.5%

 

 Female

67.4%

 

69.3%

 

Tumour site

 

0.669

 

0.865

T stage

 

0.059

 

0.072

 T1 & T2

75.6%

 

78.8%

 

 T3 & T4

50.7%

 

54.8%

 

N stage

 

0.032

 

0.089

 N0

76.2%

 

77.6%

 

 N1 & N2

50.6%

 

56.3%

 

UICC stage

 

0.023

 

0.032

 I & II

82.1%

 

84.1%

 

 III & IV

49.8%

 

59.3%

 

Margin status

0.113

 

0.116

 

 R0

67.4%

 

71.0%

 

 R1

58.3%

 

61.9%

 

Podoplanin expression

 

<0.001

 

0.001

 No podoplanin

90.9%

 

100%

 

 Weak

93.3%

 

93.3%

 

 Moderate

64.8%

 

69.7%

 

 High

31.1%

 

37.2%

 
In univariate analysis, the level of podoplanin expression had a strong impact on the overall survival (p < 0.001) and on disease-specific survival (p = 0.001). A continuous decrease of the 5-year overall survival rate was found from 93.3% for patients with weak expression to 64.8% for patients with moderate expression and to 31.1% for patients with high expression of podoplanin (Fig. 2). Patients with podoplanin-negative primary tumours had a slightly lower overall survival after 5 years (90.7%) than patients with weak podoplanin expression. A similar trend can be observed for disease-specific survival (Fig. 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs00428-010-0915-7/MediaObjects/428_2010_915_Fig2_HTML.gif
Fig. 2

Overall survival according to the different levels of podoplanin expression (log rank test, p < 0.001)

https://static-content.springer.com/image/art%3A10.1007%2Fs00428-010-0915-7/MediaObjects/428_2010_915_Fig3_HTML.gif
Fig. 3

Disease-specific survival according to the different levels of podoplanin expression (log rank test, p = 0.001)

N stage and UICC stage reached the 95% significance level for overall survival in univariate analysis (p < 0.05), too. However, T stage, age, gender, tumour site and status of the resection margins did not have a significant impact on the overall survival of our patients. For disease-specific survival, N stage did not quite reach the 95% significance level (p = 0.089).

The results of multivariate analysis are displayed in Tables 4 and 5. Multivariate analysis was performed with parameters that were significant in univariate analysis (podoplanin expression, nodal status for both survival parameters and additionally UICC stage for overall survival) to determine whether the effect of podoplanin expression on overall survival and disease-specific survival is dependent on other risk factors. For overall survival, a multivariate analysis was conducted with a model including podoplanin expression and N stage. N stage did not have significant impact on overall survival (p = 0.846) in multivariate analysis, whereas podoplanin expression was significantly associated with overall survival (p = 0.019). For overall survival, a second model using UICC stage and instead of N stage and podoplanin expression was created. The patients' prognosis was significantly influenced by the level of podoplanin expression (p = 0.021), whereas UICC stage did not have a significant impact on overall survival (p = 0.377).
Table 4

Identification of prognostic factors by multivariate analysis for overall survival (Cox proportional hazard regression model)

Multivariate analysis of prognostic factors

Parameter

Hazard ratio (95% confidence interval)

P value

Podoplanin expression

 

0.019

Negative vs. weak

0.134 (0.015–1.182)

0.070

Negative vs. moderate

0.062 (0.008–0.509)

0.010

Negative vs. high

0.389 (0.161–0.939)

0.036

N stage (N0 vs. N1&N2)

0.913 (0.364–2.292)

0.846

Table 5

Identification of prognostic factors by multivariate analysis for disease-specific survival (Cox proportional hazard regression model)

Multivariate analysis of prognostic factors

Parameter

Hazard ratio (95% confidence interval)

P value

Podoplanin expression

 

0.049

Negative vs. weak

0.189 (0.000–infinite)

0.971

Negative vs. moderate

0.083 (0.010–0.673)

0.020

Negative vs. high

0.394 (0.162–0.956)

0.037

UICC stage (I&II vs. III&IV)

0.511 (0.144–1.817)

0.300

For disease-specific survival podoplanin expression was found to be the only significant prognostic factor in multivariate analysis (p = 0.049), too. UICC stage did not significantly influence disease-specific survival in multivariate analysis (p = 0.300).

Discussion

The specific staining of the lymphatic endothelial cells and the tumour cells is consistent with findings in previous studies [23, 51]. Most of the studies tried to show an association between lymphovascular density in OSCC and cervical lymph node metastases [20, 52, 53]. Longatto-Filho et al. [53] reported a correlation between peritumoural lymphovascular density and survival, Franchi et al. [20] showed an association between peritumoural lymphovascular density and cervical lymph node metastases; Kyzas et al., however, pointed out that only intratumoural lymphovascular density and not peritumoural lymphovascular density had an impact on the clinical outcome [54]. However, assessing lymphovascular density takes up a great deal of time and is therefore not suitable as a prognostic parameter in daily routine. In this study, we focused on the level of podoplanin expression on the tumour cells themselves to investigate the impact of podoplanin as a prognostic marker for lymph node metastases and overall survival.

The reported incidence of occult regional lymph node metastases in clinical N0 necks of patients with OSCC varies from 6% to 46%. Occult regional metastases may be found even in cases with small primary T1 tumours [55]. A reliable marker, which is able to predict the chance of regional lymph node metastases, could be a great asset in decision of an elective treatment of a clinical N0 neck. The χ² test revealed a very strong association between podoplanin expression in the primary tumour and cervical lymph node metastases (p < 0.001), which suggests that the level of podoplanin expression in tumour cells is a reliable parameter to predict the appearance of cervical lymph node metastases. Recently, Yuan et al. described this association at the same significance level [23].

Univariate analysis showed a strong impact of podoplanin expression (p < 0.001) and cervical lymph node status (p = 0.032) on overall survival. Multivariate analysis confirmed our findings. In fact, podoplanin was the only parameter that exhibited a significant prognostic impact on overall survival and disease-specific survival in univariate and multivariate analysis.

Our results confirm the assumption that high levels of podoplanin expression in the primary tumour of OSCC are associated with advanced T stage, lymphatic spread to the cervical region and poor clinical outcome. Yuan et al. described similar findings, except they could not find an association between T stage and podoplanin expression. However, their observations were limited to a set of 60 patients with tongue cancer [23].

The results suggest that podoplanin might play a role in lymphatic spread and in tumour invasion and progression.

The biological function of podoplanin is not yet fully understood, but it has been proposed that podoplanin is able to enhance tumour invasion by promoting the tumour cell mobility [25]. Cancer cell migration and invasion involves active remodelling of the actin cytoskeleton [56]. However, podoplanin does not interact directly with actin but via ERM proteins such as ezrin radixin and moesin [57]. It has been shown that overexpression of podoplanin leads to an increased phosphorylation of ezrin which links podoplanin to the observed rearrangement of the actin cytoskeleton [57]. Apart from that, podoplanin increases the activities of Rho-family GTPases, mainly RhoA [42], which is also linked to tumour cell mobility.

Dumoff et al. reported a strong correlation between low expression of podoplanin and both lymphatic invasion and nodal metastasis in uterine cervical cancer [33]. These findings indicate that the biological function of podoplanin may vary between different types of cancer. A recent study, however, showed similar correlations between the level of podoplanin expression in esophageal squamous cell carcinoma (ESCC) and OSCC, suggesting the role of podoplanin may be common in ESCC and OSCC [23, 36]. The results indicate that podoplanin may be a marker for cancer stem cells of ESCC. Knockdown of podoplanin expression in ESCC cell lines resulted in reduction of in vitro colony formation and in vivo tumourigenic activity. Podoplanin knockdown cells also showed decreased invasive ability and were more vulnerable to cisplatin and 5-FU, two anticancer drugs frequently used in chemotherapy for ESCC and OSCC [36, 58, 59]. These results support the assumption that there is an important role for podoplanin in tumourigenesis and progression in ESCC and in OSCC.

Although we were able to show significant associations between the level of podoplanin expression, T stage, N stage and the significant impact of the level of podoplanin expression on overall survival and disease-specific survival in univariate and multivariate analysis, prospective studies with larger cohorts are needed to clarify the role podoplanin in precancerous lesions and invasive carcinoma. Exploring the function of biomarkers like podoplanin offers the opportunity to move from a tumour model of temporal determinism to one of biological determinism as carcinogenesis is not defined by what stage the patient is in at detection but rather by molecular and characteristics of the tumour and the host.

Conflict of interest

The authors declare that they have no potential or actual conflict of interest.

Copyright information

© Springer-Verlag 2010