Expression of focal adhesion kinase (p125 FAK) and proline-rich tyrosine kinase 2 (PYK2/CAKb) in cerebral metastases, correlation with VEGF-R-, ecNOS III-labelling and morphometric data.

BACKGROUND

Cerebral metastasis occurs in about 20% of all neurosurgical patients. Cerebral metastases have a typical spherical morphology with a common central necrosis and perifocal oedema. It has been proposed that oedema extension, tumour volumes and infiltrative behaviour are partially mediated by vascular endothelial growth factor (VEGF) and nitric oxide (NO). In several systemic tumour entities NO is suggested as a factor which influences the metastatic potential. VEGF has recently been reported to influence the matrix related migratory activity by interaction with focal adhesion kinase (p125FAK) and proline-rich tyrosine kinase beta (PYK2/CAK beta). Nitric oxide, which is produced in metastases by three different NOS isozymes is capable of antagonizing the binding of FAK to matrix integrins. NO, VEGF and FAK/PYK 2 are therefore considered to be important mediators of the cerebral metastatic incidence, growth, infiltration and oedema extension. The aim of our present study was to investigate the expression of p125FAK and the coexpression with PYK2/CAK beta, VEGF-receptor FLT-1, NOS isozymes NOS I-III, capillary density and the histology in 130 specimens of resected cerebral metastatic tumours. A further analysis was performed to morphometrically evaluate tumour and oedema volumes and to correlate the immunohistochemical data in a subgroup of 40 patients.

METHODS

Cryosections (N = 130) of metastatic resections were investigated immunohistologically using a 4-step scoring evaluation for the expression of NOS I-III, VEGF-receptor FLT-1, and capillary vessel presence by endothelial Von-Willebrand-Factor (VWF) staining. Tumour and oedema extension was measured in preoperative MRI (N = 40) scans by an image-processing device (Kontron) and the ratios of oedema volumes to total tumour volumes were calculated. The data were analysed statistically (Spearman rank order correlation and Kruskal-Wallis ANOVA) and correlated with the clinical data.

RESULTS

FAK immunoexpression was observed in 50% of the specimens (31.2% gradings 2 and 3). We observed a significant coexpression (p = 0.0001) with PYK 2 labelling which occurred frequently in 74% of the specimens (42% gradings 2 and 3). The VEGF receptor FLT-1 could be detected in 70% of them, 24% at higher expression values 2 and 3. The expression of NO synthase was frequently observed. NOS I was detected in 83.6% of the specimens, values 2 and 3 in 40.5%. NOS III, the endothelial isoform, was observed in 39.4% of the specimens (gradings 2 and 3) and inducible NOS II in 29.4% (grading 2 and 3) of them. Coexpressions were statistically significant for FAK and NOS III (Spearman p = 0.008) and FAK and VEGF-R (p = 0.03). The morphometric evaluation resulted in tumour volumes between 2.0 and 83 cm3 (mean 22.5 +/- 19.1 SD) with oedema ratios between 0 and 100% (mean 62.2 +/- 22.5 SD). FAK expression correlated significantly (p = 0.06) with tumour volumes and histology.

CONCLUSIONS

The frequent histotypic occurrence of FAK and PYK2 in metastases could be an important factor in the modulation of metastatic capacity and infiltrative behaviour and might influence the disease course. Judging from its frequent expression PYK2 may generate the more relevant signals. A further aspect is the possible interaction with endothelial NOS III and VEGF receptor, which could be important for the infiltrative behaviour in a latent hypoxic scenery and environment.