J Cancer Res Clin Oncol (2010) 136:1489–1496 DOI 10.1007/s00432-010-0806-y

ORIGINAL PAPER

Expression and prognostic signiWcance of focal adhesion kinase in hepatocellular carcinoma Zhou Yuan · Qi Zheng · Jia Fan · Kai-xing Ai · Jie Chen · Xin-yu Huang

Received: 21 September 2009 / Accepted: 28 January 2010 / Published online: 12 February 2010 © Springer-Verlag 2010

Abstract Aim To examine the expressions of focal adhesion kinase (FAK) and its clinical signiWcance in hepatocellular carcinoma (HCC). Methods We determined the expression levels of FAK on both steady-state mRNA and protein levels in 50 HCC samples by quantitative real-time PCR and immunohistochemistry, respectively. The correlations between FAK expression and various clinicopathological parameters were analyzed. Results The expression of FAK on the mRNA level was consistent with that on the protein level. FAK mRNA levels in tumor tissues were signiWcantly higher than in the paratumor tissues (0.229 § 0.027 vs. 0.163 § 0.019; P < 0.001), but lower than in the macroscopic cancer emboli (0.506 § 0.155 vs. 0.377 § 0.176; P < 0.05). Compared within the tumor tissues, FAK expressions were signiWcantly higher in those with cancer emboli than those without (0.343 § 0.05 vs. 0.165 § 0.025; P = 0.003). Univariate and multivariate analyses revealed that FAK expression was an independent prognostic factor for disease-free survival and overall survival. Both 3-year disease-free survival rates and overall survival rates in FAK-negative group were signiWcantly higher than in positive group (52 vs. 20% and 72 vs. 29%, P < 0.001).

Z. Yuan · Q. Zheng · K. Ai · J. Chen · X. Huang (&) Department of General Surgery, The Sixth People’s Hospital AYliated to Shanghai Jiao Tong University, 600 Yishan RD, Shanghai 200233, China e-mail: [email protected] J. Fan Liver Cancer Institute and Zhongshan Hospital, Fudan University, 180 Fenglin RD, Shanghai 200032, China

Conclusion Our results suggest that FAK expression is up-regulated in HCC and its expression is an independent prognostic factor for HCC. Keywords Focal adhesion kinase · Hepatocellular carcinoma · Real-time PCR · Prognosis

Introduction Hepatocellular carcinoma (HCC) is the most common malignancy in China and the third leading cause of cancerrelated death worldwide (Parikh and Hyman 2007). Chronic infection with the hepatitis B and C virus (HBV and HCV) is the most common cause of this disease. Other important risk factors include alcohol abuse, oral contraceptives, cigarette smoking, exposure to liver toxins such as aXatoxin, etc. (Bosch et al. 2004). In clinic, although resection of the primary tumor has been routinely applied as the standard treatment for HCC, the post-surgical metastasis and recurrences are still as high as 60–70%. Local invasion and distant metastasis of the tumor cells are mainly responsible for the high mortality of HCC. On the molecular level, the interactions between the stroma and the cancer cells are the determinant factors for tumor migration and invasion. This stroma–cancer interaction is carried out by the stromal components, such as extracellular matrix (ECM), embedded signaling molecules and other stromal cell types, the cell-surface receptors and intracellular signaling molecules of the tumor cells. Focal adhesion kinase (FAK) is a cytoplasmic non-receptor tyrosine kinase that when activated, localizes to the sites of cell–matrix interactions, known as focal adhesions. Multiple signals, including integrin clustering, engagement of G-protein coupled receptors (GPCRs) and various growth

123

1490

factor receptors, lead to the activation of FAK, which, in turn, transmits the signals to a vast array of downstream targets, regulating cell survival, cell cycle progression, cell motility and invasion (Schlaepfer et al. 2004). Activation of FAK by integrin-mediated cell adhesion to the extracellular matrix induces autophosphorylation of FAK at Tyr-397. In response to adhesion signals, active Src can bind and phosphorylate FAK, resulting in the phosphorylation of substrates such as p130CAS and paxillin that are involved in cytoskeleton reorganization required for cell movement. These signal transductions are key processes in cell motility and invasion. So, FAK may be intimately involved in malignancies. In consistent with the biological functions of FAK, studies have shown that the expression levels of FAK are dramatically up-regulated in multiple malignancies, such as breast cancer, prostate cancer, colon cancer and thyroid cancers (Cance et al. 2000 Owens et al. 1996; Tremblay et al. 1996). In the present study, we examined the FAK expressions on both mRNA and protein levels in HCC samples from 50 Chinese patients, and analyzed the correlations between FAK expressions and various clinicopathological parameters. We found that FAK was highly expressed in the tumor samples compared to the non-cancerous paratumor tissues, and that the high expressions was an independent indicator for more invasive phenotype and worse prognosis of HCC patients.

Materials and methods Patients and specimens This study was approved by the Ethics Committee of the Sixth People’s Hospital AYliated to Shanghai Jiao Tong University and Zhongshan Hospital, Fudan University (Shanghai, China). Informed consents were obtained from all patients recruited into this study. Fifty Chinese HCC patients undergoing liver resection at the Department of Hepatobiliary Surgery, Zhongshan Hospital, Fudan University in 2005 were included into this study. Among these 50 patients, there were 47 males and 3 females, with age ranging from 13 to 72 years (mean, 48.5 years). All patients were infected by HBV. The diagnosis of HCC was conWrmed by pathological examinations. Based on the TNM staging criteria from American Joint Committee on Cancer (AJCC) (Greene et al. 2002), these HCC tumors include 6 stage I, 17 stage II, 17 stage III and 10 stage IV. In addition, 37 patients had accompanied liver cirrhosis and 13 did not. According to the Edmondson-Steiner grading, there were 9 grade I, 32 grade II and 9 grade III tumors. By invasiveness, 32 of the HCC were low invasive as judged by localized primary tumor with no macro- or microscopic

123

J Cancer Res Clin Oncol (2010) 136:1489–1496

intra-hepatic dissemination or formation of cancer emboli in the portal vein; while the other 18 were highly invasive with multiple intra-hepatic dissemination and/or portal vein inWltration. All patients were followed up till 50 weeks after surgery. For specimen acquisition, tumor tissues were dissected from macro- or microscopic conWrmed HCC tissues. The paratumor tissues were isolated from liver tissues at least 2 cm away from tumors and with no microscopic tumor cells. The emboli tissues were from macroscopic cancer emboli in the portal vein. RNA extraction and cDNA synthesis Total RNA was extracted using Trizol reagent (Invitrogen, USA) following the manufacturer’s instructions. cDNA was synthesized from 2 g total RNA using oligo dT as the primer and M-MuLV Reverse Transcriptase (Promega, USA) as the enzyme. Real-time quantitative PCR (RT-qPCR) The following primers were used for RT-qPCR analysis: FAK, forward primer 5⬘ CCAGCCCTACCAGAGGAGT ATG 3⬘, reverse primer 5⬘ TGGCTGGTAGGAGGGGA AT 3⬘, and amplicon size 103 base pairs (bp); GAPDH (internal control), forward primer 5⬘ GTCCATGCCATCACT GGCAC 3⬘, reverse primer 5⬘ ATGACCTTGCCCACAG CCTT 3⬘, and amplicon size 131 bp. The PCR conditions were: one cycle at 95°C for 20 min followed by 40 cycles at 95°C for 30 s and 60°C for 60 s. The real-time PCR was performed on MJ Research ABI 9700 Thermal Cycler (GMI, USA) and the relative expression of FAK was calculated against GAPDH mRNA level using the 2¡Ct method (Too 2003). Immnunohistochemistry (IHC) The FAK immunohistochemistry was performed using H-1 antibody (Santa Cruz, USA) according to the manufacturer’s instructions. The working dilution is 1:100. All stainings were examined and scored blindly by a board-certiWed pathologist. The following criteria were applied to evaluate the FAK staining intensity. Strong positivity, the cytoplasm of tumor cells showed yellow or brown and the percentage of stained tumor cells out of all tumor cells ¸75%; weak positivity, the cytoplasm of tumor cells showed yellow or brown and the percentage of stained tumor cells out of all tumor cells was between 25% (including 25%) and 75%; negativity, the cytoplasm of tumor cells were not stained or the percentage of stained tumor cells out of all tumor cells was less than 25%.

J Cancer Res Clin Oncol (2010) 136:1489–1496

Western blot analysis Tissue extract was prepared in RIPA buVer and the protein concentration determined by the BCA method (Pierce, USA). 20 g of total proteins was loaded onto 10% SDS– PAGE gel and transferred to PVDF membrane. The membrane was incubated with anti-FAK H-1 antibody (1:100) followed by horseradish peroxidase-conjugated secondary antibody. The signal was detected using ECL substrate and GAPDH was used as internal control. Statistical analysis All statistical analyses were performed using SPSS 13.0 software. The correlation between FAK expressions and various clinicopathological parameters was analyzed either using Mann–Whitney test (for comparison between two groups) or using Kruskal–Wallis test (from comparison among three or more groups). The association between FAK expressions and TNM staging or tumor grading was analyzed by Kruskal–Wallis test. The post-surgical survival rates were calculated using Kaplan–Meier survival analysis and the diVerence between survival curves were analyzed using the log-rank test. The signiWcance of prognostic factors was determined by Cox Regression analysis. P < 0.05 was considered statistically signiWcant.

Results FAK steady-state mRNA levels signiWcantly increased in the HCC tumor tissues To examine the clinical signiWcance of FAK expression in HCC, we Wrst determined the steady-state mRNA levels of FAK in the tumor tissues using RT-qPCR and compared that to the corresponding paratumor control tissues. As shown in Fig. 1a, for the majority of the 50 paired samples, FAK mRNA levels in the tumor tissue was higher than in the control tissue. On average, this diVerence was statistically signiWcant (Fig. 1b, 0.229 § 0.027 vs. 0.163 § 0.019; P < 0.001). In addition, when compared within the tumor tissues, those accompanied with cancer emboli in the portal vein (N = 18) had signiWcantly higher FAK mRNA than those without (N = 32) (Fig. 1c, 0.343 § 0.05 vs. 0.165 § 0.025; P < 0.05). When comparing the cancer emboli to paired tumor tissue from the same patient (N = 9), we found that the FAK mRNA levels were signiWcantly higher in the former samples than in the latter (Fig. 1d, 0.506 § 0.155 vs. 0.377 § 0.176; P < 0.05). By comparing the paratumors tissues from samples with liver cirrhosis to those without, we noticed the FAK mRNA is slightly higher in the former group than

1491

in the latter, but the diVerence was not statistically signiWcant (Fig. 1e, 0.191 § 0.029 vs. 0.141 § 0.034; P > 0.05). FAK protein levels increased with the development of HCC Although the steady-state mRNA level is a strong indicator for gene expression level, since protein is the functional form for FAK, we next examined the expression of FAK on protein levels by IHC and Western blot analysis. By IHC, 50% (N = 25) of the HCC tumor samples showed positive staining of FAK, among which eight were strong positive and 17 were weak positive. Detailed analysis revealed that FAK were only positively stained in the cytoplasm of carcinoma cells, with no stain in the mesenchymal cells or paratumor tissues but the strongest stain in cancer emboli (Fig. 2a–d). Similar to the trend as seen for FAK mRNA levels, FAK protein was highest in cancer emboli, followed by tumor tissues and lowest in paratumor tissues (Fig. 2e). The diVerential expressions of FAK during HCC development were consistent on both mRNA and protein levels Given the increasing trend of FAK on steady-state mRNA and protein levels during HCC development, we investigated whether the diVerential expressions of FAK were consistent on these two levels. As shown in Fig. 3, by comparing the mRNA levels of FAK in tumor tissues with positive staining (N = 25) to those in tumor tissues with negative staining (N = 25) as determined by IHC, we found that the former group had signiWcantly higher mRNA expressions than the latter group (P = 0.001), suggesting the consistency of both levels to represent FAK changes during HCC development. FAK mRNA expression level signiWcantly correlated with tumor invasiveness and the presence of cancer emboli To assess the clinical signiWcance of FAK mRNA expression in HCC, we examined the correlation between FAK mRNA levels and a variety of clinicopathological parameters (Table 1). High mRNA expression of FAK signiWcantly correlated with a more invasive phenotype (P = 0.02) and the formation of cancer emboli in the portal vein (P = 0.003). No signiWcant correlation was found between FAK mRNA expression and other variables including patient gender, status of liver cirrhosis, serum alpha fetal protein (AFP) level, pathological grading, tumor diameter, TNM staging or the presence of tumor capsule (P > 0.05).

123

1492

J Cancer Res Clin Oncol (2010) 136:1489–1496

Fig. 1 FAK steady-state mRNA levels signiWcantly increased in the HCC tumor tissues. a The steady-state mRNA levels in 50 pairs of liver tissues were examined by quantitative real-time PCR (qRT-PCR) and presented as relative ratio to GAPDH (internal control). Each pair consisted of HCC tumor tissue and paratumor tissue from the same patient. b The relative mRNA levels of FAK in the tumor and paratumor tissues from all 50 patients were determined as in a and presented as mean § SD. c The relative mRNA levels of FAK in tumors accompanied with formation of cancer emboli (w/emboli) (N = 18) and in those

without (w/o emboli) (N = 32) were determined as in a and presented as mean § SD. d The relative mRNA levels of FAK in macroscopic cancer emboli and paired tumor tissues were determined by qRT-PCR and presented as mean § SD (N = 9). e The relative mRNA levels of FAK in paratumor tissues with (w/cirrhosis) and without liver cirrhosis (w/o cirrhosis) (N = 37 vs. 13) were determined by qRT-PCR and presented as mean § SD. **P < 0.001, *P < 0.05 as compared to the other group in each Wgure

FAK expression was an independent prognostic factor for patients with HCC by univariate and multivariate analyses

To investigate the potential of using FAK mRNA expression in predicting the postoperative prognosis of HCC patients, we performed univariate analysis using Cox regression analysis. The results showed that FAK mRNA expression level, serum AFP level and presence of cancer emboli were all independent and signiWcant predictors of patient prognosis (P = 0.009, 0.035 and 0.002, respectively) (Table 2), while other parameters, such as patient gender, tumor size, pathological grading, TNM staging, tumor invasiveness, status of liver cirrhosis and presence of capsule, were not (data not shown). The risk of death for HCC patients with positive FAK expression was 3.734-fold higher than those with negative FAK

To evaluate the prognostic signiWcance of FAK expression, we analyzed the disease-free survival rates and overall survival rates of FAK-positive patients and FAK-negative patients, as determined by IHC analysis. As shown in Fig. 4, both 3-year disease-free survival rates and overall survival rates in FAK-negative group were signiWcantly higher than in positive group (52 vs. 20% and 72 vs. 29%, respectively), and the diVerences between these groups were statistically signiWcant (P < 0.001).

123

J Cancer Res Clin Oncol (2010) 136:1489–1496

1493

Fig. 2 FAK protein levels increased with the development of HCC. The FAK protein expressions were detected by IHC analysis and representative images for negative staining (a), weak positive staining (b), strong positive staining (c) and cancer emboli (arrowhead, d) were presented. H&E stain, £200. e The FAK protein from paratumor, tumor and cancer emboli tissues were detected by western blot analysis. GAPDH was used as internal control

emboli formation had a 4.593-fold higher chance of death than those without.

Discussion

Fig. 3 FAK protein levels correlated with its mRNA levels in HCC tumor tissues. The FAK protein levels were determined by IHC and grouped into positive (FAK (+)) and negative (FAK(¡)) staining groups (N = 25 for each group). The relative mRNA levels for all tumor samples were determined by qRT-PCR and compared between the positive and negative staining groups. *P = 0.001 as compared to the negative staining group

expression. Similarly, that for HCC patients with serum AFP level >200 ng/mL was 3.068 fold higher than those with AFP level ·200 ng/mL, and HCC patients with cancer

Postoperative recurrences and metastasis are the main reasons for poor prognosis of HCC patients. Currently, factors such as tumor vascular inWltration and TNM staging are considered as the key prognostic indicators for HCC patients. However, there is still a great need for tumor biomarkers that can suYciently and accurately predict patient outcome. The adhesion of tumor cells to ECM and vascular endothelial cells, which is mainly mediated through the integrin family of proteins, is prerequisite for tumor invasion and metastasis. FAK is a downstream signaling molecule for both intra- and extra-cellular signals conveyed through integrins or growth factor receptors. Wu et al. (2006) revealed that the cell adhesion to ECM mediated in human breast carcinoma cells was inhibited by reduced the expression of FAK and its tyrosine phosphorylation level, which

123

1494

J Cancer Res Clin Oncol (2010) 136:1489–1496

Table 1 The correlation between FAK expression and clinicopathological parameters in 50 patients with HCC Clinicopathological parameters

N

Relative FAK mRNA level (mean § SD)

P value

47

0.235 § 0.19

0.426

3

0.129 § 0.10

Positive

33

0.240 § 0.18

Negative

17

0.207 § 0.19

Gender Male Female Liver cirrhosis 0.774

Serum AFP level ·200 ng/mL

25

0.192 § 0.16

>200 ng/mL

25

0.265 § 0.21

I

9

0.198 § 0.17

II

32

0.228 § 0.19

III

9

0.262 § 0.21

·5 cm

29

0.201 § 0.19

>5 cm

21

0.267 § 0.18

I

6

0.190 § 0.16

II

17

0.180 § 0.15

III

17

0.239 § 0.22

IV

10

0.318 § 0.19

Low

22

0.161 § 0.14

High

28

0.282 § 0.20

Present

23

0.199 § 0.15

Absent

27

0.254 § 0.21

Absent

32

0.165 § 0.05

Present

18

0.343 § 0.025

0.286

Pathological grading 0.794

Tumor diameter 0.157

TNM staging 0.299

Invasiveness 0.020

Tumor capsule 0.514

Table 2 SigniWcant variables for prognosis as determined by Cox’s regression analysis

Cancer emboli 0.003

was induced by the decreased expression of integrin 51 in positive expression of E-cadherin cells. Owens et al. (1995) found that the up-regulation of FAK was a common pathway utilized by many epithelial tumors and mesenchymal tumors in the process of gaining the potentials to invade locally and metastasize distantly. In the present study, by comparing paired tissues, we found that the FAK mRNA levels in cancer emboli were signiWcantly higher than in tumor tissues, where the level was signiWcantly higher than in paratumor tissues (P < 0.05). Compared within the tumor tissues, those accompanied with emboli formation contained much higher FAK mRNA levels than those without. In the cirrhotic paratumor tissues, the FAK mRNA levels were higher than in the non-cirrhotic paratumor tissues, but the diVerence was not statistically signiWcant (P > 0.05). Further analysis on the association between

123

Fig. 4 FAK expression levels signiWcantly and reversely correlated with postoperative patient survival. All 50 HCC patients were divided into positive and negative FAK groups based on IHC staining for the tumor tissues. Both 3-year disease-free survival rates (a) and overall survival rates (b) in FAK-negative group were signiWcantly higher than in positive group (52 vs. 20% and 72 vs. 29%, P < 0.001) by Kaplan– Meier survival analysis

Variable

 value

SE

Wald

P value

Relative risk

Serum AFP

1.121

0.532

4.439

0.035

3.068

Cancer emboli

1.524

0.498

9.369

0.002

4.593

FAK expression

1.317

0.506

6.779

0.009

3.734

FAK mRNA levels and various clinicopathological parameters revealed that FAK expression signiWcantly and positively correlated with tumor invasiveness and formation of cancer emboli. These data suggest that FAK is highly expressed in malignant HCC and its expression level associates with the formation of emboli in portal vein and tumor invasion. In addition to the mRNA level, we also examined the expression of FAK on the protein level. By IHC and western blot analyses, we observed similar trends on FAK protein levels as on its mRNA levels, that is, higher FAK protein levels were detected in more malignant tumor

J Cancer Res Clin Oncol (2010) 136:1489–1496

tissues. Correlation analysis indicated that FAK protein levels signiWcantly correlated with its mRNA levels, suggesting the major regulation on FAK expression during HCC development is on or before the transcription step, but not after that. Although the high expression of FAK is associated with the invasion and metastasis of many malignant tumors, its correlation with patient prognosis is still not clear. Sood et al. (2004) discovered that in ovarian cancer, high expression of FAK is associated with lymph node and distal metastasis, as well as shorter patient survival. Mol et al. (2007) founded the elevated expression of FAK in early stage prostate carcinoma might be used as a biological marker. For colon adenocarcinomas, however, studies suggest that FAK expression levels cannot be used as an prognostic indicator (Han et al. 1997; Theocharis et al. 2003). In the current study, by survival analysis, we found that in HCC with negative FAK expression, the disease-free survival rates and the overall survival rates were signiWcantly higher than in tumors with positive FAK expression, indicating a reverse correlation between FAK expression and prognosis of HCC patients. Unlike most other epitheliumderived tumors that spread through the lymph, the major metastatic pathway for HCC, is through the blood circulation via the portal vein. The formation of cancer emboli in the portal vein, therefore, is an important indicator for patient outcome. Univariate and multivariate analyses showed that FAK expression, presence of cancer emboli in the portal vein and serum AFP levels are all independent prognostic factor for HCC patients. Patients with positive FAK expression have approximately 3.7-fold higher risk of death than those without. Some studies have revealed cancer emboli in the portal vein and serum AFP levels as prognostic factor for HCC patients. This study indicates that FAK overexpression might help to identify HCC patients with a poor prognosis. Itoh et al. (2004) identiWed that FAK expression could play an important role in promoting HCC progression, especially portal venous invasion and that FAK would be a novel target for HCC therapeutics as well as a prognostic marker. Fujii et al. (2004) showed an increase of FAK mRNA in 60 Japanese patients with HCC and that the expression was an independent prognostic factor for disease-free and overall survival. Our study also supports the role of FAK as a prognostic biomarker for HCC. However, the mechanism by which FAK expression contributes to poor prognosis of patients with HCC is still not clear and should be analyzed by further study. In summary, FAK is up-regulated in primary HCC tissues, with the highest expression in cancer emboli formed in the portal vein, implying its importance in the onset, progression and hematogenous spread of HCC. The reverse correlation between high FAK expression and poor patient

1495

prognosis indicate its potential as a prognostic marker for HCC patient. Further study on the correlation between circulatory FAK level and HCC recurrences and prognosis will provide more insights on the potential of using FAK as a novel clinical tumor marker and a promising therapeutic target for HCC. Acknowledgments This study was supported by the National Natural Science Foundation of China, No.30872512.

References Bosch FX, Ribes J, Diaz M, Cleries R (2004) Primary liver cancer: worldwide incidence and trends. Gastroenterology 127:5–16 Cance WG, Harris JE, Iacocca MV, Roche E, Yang X, Chang J, Simkins S, Xu L (2000) Immunohistochemical analyses of focal adhesion kinase expression in benign and malignant human breast and colon tissues: correlation with preinvasive and invasive phenotypes. Clin Cancer Res 6:2417–2423 Fujii T, Koshikawa K, Nomoto S, Okochi O, Kaneko T, Inoue S, Yatabe Y, Takeda S, Nakao A (2004) Focal adhesion kinase is overexpressed in hepatocellular carcinoma and can be served as an independent prognostic factor. J Hepatol 41:104–111 Greene FL, Page DL, Fleming ID (2002) AJCC cancer staging manual, 6th edn. Springer, New York, pp 131–144 Han NM, Fleming RY, Curley SA, Gallick GE (1997) Overexpression of focal adhesion kinase (p125FAK) in human colorectal carcinoma liver metastases: independence from c-src or c-yes activation. Ann Surg Oncol 4:264–268 Itoh S, Maeda T, Shimada M, Aishima S, Shirabe K, Tanaka S, Maehara Y (2004) Role of expression of focal adhesion kinase in progression of hepatocellular carcinoma. Clin Cancer Res 10:2812–2817 Mol AJ, Geldof AA, Meijer GA, van der Poel HG, van Moorselaar RJ (2007) New experimental markers for early detection of high-risk prostate cancer: role of cell–cell adhesion and cell migration. J Cancer Res Clin Oncol 133:687–695 Owens LV, Xu L, Craven RJ, Dent GA, Weiner TM, Kornberg L, Liu ET, Cance WG (1995) Overexpression of the focal adhesion kinase (p125FAK) in invasive human tumors. Cancer Res 55:2752–2755 Owens LV, Xu L, Dent GA, Yang X, Sturge GC, Craven RJ, Cance WG (1996) Focal adhesion kinase as a marker of invasive potential in diVerentiated human thyroid cancer. Ann Surg Oncol 3:100–105 Parikh S, Hyman D (2007) Hepatocellular cancer: a guide for the internist. Am J Med 120:194–202 Schlaepfer DD, Mitra SK, Ilic D (2004) Control of motile and invasive cell phenotypes by focal adhesion kinase. Biochim Biophys Acta 1692:77–102 Sood AK, CoYn JE, Schneider GB, Fletcher MS, DeYoung BR, Gruman LM, Gershenson DM, Schaller MD, Hendrix MJC (2004) Biological signiWcance of focal adhesion kinase in ovarian cancer: role in migration and invasion. Am J Pathol 165:1087– 1095 Theocharis SE, Kouraklis GP, Kakisis JD, Kanelli HG, Apostolakou FE, Karatzas GM (2003) Focal adhesion kinase expression is not a prognostic predictor in colon adenocarcinoma patients. Eur J Surg Oncol 29:571–574 Too HP (2003) Real time PCR quantiWcation of GFRalpha-2 alternatively spliced isoforms in murine brain and peripheral tissues. Brain Res Mol Brain Res 114:146–153

123

1496 Tremblay L, Hauck W, Aprikian AG, Begin LR, Chapdelaine A, Chevalier S (1996) Focal adhesion kinase (pp125FAK) expression, activation and association with paxillin and p50CSK in human metastatic prostate carcinoma. Int J Cancer 68:164–171

123

J Cancer Res Clin Oncol (2010) 136:1489–1496 Wu H, Liang YL, Li Z, Jin Zhang W, Duan L, Zha X (2006) Positive expression of E-cadherin suppresses cell adhesion to Wbronectin via reduction of alpha5beta1 integrin in human breast carcinoma cells. J Cancer Res Clin Oncol 132:795–803