Our study confirmed a marked overexpression of KLK10
in PDAC by means of a virtual subarray. Immunohistochemistry in native tumour tissue could prove not only an intense expression for KLK10
in 64.4% of the malignant cells, but also for KLK6
in 91.5%. Both proteins were located in the cytoplasm, from where they are likely to be secreted (Borgono et al, 2004
Co-expression of different kallikreins, similar to the situation found in our study, was already reported in skin and different glands. In these tissues the kallikreins can act independently, but also together as part of proteolytic cascades (Petraki et al, 2002b
; Borgono and Diamandis, 2004
). The latter seems to be an important mechanism in pancreatic cancer, because expression of KLK10
itself could not be associated with poor survival in PDAC, whereas the co-expression of both kallikreins was significantly associated with poor survival and an R1-resection status, which is an indirect sign for infiltrative and aggressive growth. In multivariate analysis, the co-expression of KLK10
was also an independent risk factor for survival.
It is most interesting, in which ways kallikreins affect cellular signalling and thereby contribute to cancer progression. It was already reported that kallikreins influence communication between malignant cells and their environment by degradation of extracellular matrix and thereby facilitate tumour invasion and metastasis (Borgono and Diamandis, 2004
). In case of KLK6
, degradation of fibrinogen, laminin, fibronectin and collagen types I and IV are documented (Bernett et al, 2002
; Magklara et al, 2003
). This cleavage of fractions of the ECM might be of specific importance in pancreatic carcinoma, which is a tumour type with a very high content of stromal tissue (Pilarsky et al, 2008
). In contrast, functional data on KLK10
are very limited. Although Zhang et al (2006
) suggested, that KLK10
was not even an active protease, it was stated in the same report that neither the protein relevant for conversion of KLK10
into its active form nor the physiological substrates for KLK10
are known. So, the importance of KLK10
in tumour progression remains unclear.
It therefore seems crucial to further pinpoint some of the components, which might be responsible for the pathophysiological effect of KLK10
. To find possible interaction partners for both kallikreins we used the in silico
method of protein interaction prediction. By means of this method we could identify four potential interaction partners for KLK6
. Although α-1 antiproteinase seems to be an inhibitor for KLK6
action, the interaction with AT III shows a branching between kallikreins and blood coagulation cascade, as already reported earlier (Borgono et al, 2004
). Another interaction partner was pigment epithelium-derived factor (PEDF), which is the major circulating inhibitor of plasmin. With this interaction, PEDF is linked to the plasminogen activator/plasmin system, which is one of the main protease systems involved in tumour cell invasion and metastasis (Hayashido et al, 2007
). Only recently PEDF was also identified as a key inhibitor of stromal vasculature in the mural pancreas. In vivo
androgen ablation increased PEDF in human cancer biopsies, which might also be an indirect sign for the interaction of the androgen-responsive kallikrein family and PEDF (Doll et al, 2003
). The interaction between KLK6
and PEDF seems highly significant and studies are under way which will further evaluate this topic. Another interaction partner we found is synuclein, which integrates presynaptic signalling and membrane trafficking in neurons. The high expression of KLK6
might thereby play an important role in various pathologic processes of pancreatic cancer.
Although a specific interaction partner for KLK10
could not be found, our study implies that it might have a role in the pathophysiology of PDAC. To ascertain the contribution of KLK10
to pancreatic cancer microenvironment, we used siRNA-mediated gene-silencing (Hammond et al, 2000
). AsPC-1 cells, which inherently express high levels of KLK10
mRNA, were transfected with specific siRNA. We could not observe an effect on proliferation or apoptosis in KLK10
-silenced cells (data not shown). But KLK10
-suppressed clones had markedly reduced cell motility in the Boyden chamber assay. The number of cells migrating through the membrane along an FCS-gradient dropped more than 50%. This is highly significant, as KLK6
was also shown to reduce cell motility (Ghosh et al, 2004
Although high expression in pancreatic carcinoma indicates that KLK6 and 10 could be promising tumour markers, we could not assess the use of KLK6 and KLK10 as serum biomarkers in PDAC. The serum levels of both proteins showed no significant differences between patients with PDAC and healthy donors. In addition, the serum concentrations were not able to predict the localisation of malignant lesions in the pancreatico-biliary tract. This circumstance can be because of the mainly local action of the kallikreins or fast degradation. Probably future studies including more patients can prove a use for KLK6 or KLK10 as tumour biomarkers in PDAC.
In conclusion, this study shows that KLK10 and KLK6 co-expression has an unfavourable influence on the survival in patients with PDAC and was significantly associated with R1 resection status. This effect might be mediated by direct or indirect interaction of the two kallikreins. The pathophysiological mechanisms are most likely degradation of the extracellular matrix and interaction with angiogenic factors by KLK6, whereas KLK10 augments cell motility. However, our findings suggest a high complexity of interactions between the kallikreins, which leaves it difficult to generally make statements about properties of single kallikreins.
It seems very promising to find out more about the physiological role of KLK10. Consequently, it might be possible to use inhibitors of kallikreins to disrupt interactions between the tumour and its environment and thereby reduce disease progression in patients with pancreatic cancer.