In the present study we have determined the mRNA levels of prostasin, PN-1, HAI-1A and HAI-1B during colorectal cancer carcinogenesis. It has previously been shown that over-expression of prostasin in mammary and prostate cancer cells reduces the invasive properties of cancer cells [23, 24] and that high prostasin expression in gastric tumours is associated with longer survival . It could thus be speculated that the transition from severe dysplasia to cancer is accompanied by a transcriptional decrease in prostasin expression and/or a down-regulation of prostasin activity via up regulation of one or more enzymatic inhibitors of prostasin. In the present study we found modest but significantly lower levels of prostasin mRNA in mild/moderate dysplasia, severe dysplasia and colorectal cancer tissues combined as compared to corresponding normal tissues from healthy individuals. In a previous study a similar minor but significant down-regulation was seen when comparing normal and affected tissue from gastric cancer patients . Our results do not show any transcriptional down-regulation of prostasin accompanying the transition between severe dysplasia and colorectal cancer. However, the mRNA levels of the inhibitor of prostasin, PN-1 increase at both the transition between normal tissue and mild/moderate dysplasia and again at the transition between severe dysplasia and colorectal cancer.
For PN-1 to be a physiologically relevant inhibitor of prostasin it is required that PN-1 and prostasin have physical access to each other. PN-1 belongs to the serpin family of secretory proteins. We have previously shown that five other members of the serpin family are secreted to both the apical and the basolateral side of the polarized epithelial cell line, MDCK, albeit at different ratios . Similar results were obtained for the Caco-2 cell line, derived from a human colonic adenocarcinoma (Vogel et al., unpublished results). There are therefore no indications that PN-1 in the extracellular fluid has restricted access to plasma membrane-bound prostasin. Consequently, it is possible that PN-1 plays an important role in down-regulating the enzymatic activity of prostasin in cancer tissue resulting in more invasive cell behaviour.
Prostasin is inhibited not only by PN-1 but also by HAI-1A and HAI-1B. We have previously shown that the level of HAI-1 mRNA is decreased by 3-fold during colorectal cancer carcinogenesis using an assay that does not distinguish between HAI-1A and HAI-1B . In the present study we have shown that HAI-1A and HAI-1B have expression patterns virtually identical and indistinguishable from the expression pattern of HAI-1 previously found  and that the ratio of HAI-1A/HAI-1B remains the same in all tissues investigated during colorectal cancer carcinogenesis. Therefore, the splicing of HAI-1 mRNA is not regulated during carcinogenesis. In the present study we found limited co-localization of prostasin and HAI-1 in normal tissue, as HAI-1 is located mainly on the basolateral plasma membrane and prostasin is located on the apical plasma membrane. This probably does not mean that their interactions is limited due to physical separation as we have previously shown that HAI-1 is transcytosed from the basolateral plasma membrane to the apical plasma membrane . It is thus possible that HAI-1 interacts with prostasin at the apical plasma membrane.
The enzymatic activity of prostasin is influenced both by the levels of PN-1, HAI-1A and HAI-1B and probably also by the expression levels of other as yet undiscovered inhibitors of prostasin. The relative physiological importance of PN-1, HAI-1A and HAI-1B as inhibitors of prostasin is at present unclear. PN-1 is an inhibitor of not only prostasin, but also of uPA, tPA, thrombin, plasmin, trypsin  and matriptase  and it forms an essentially non-reversible binding to its target protease like most other serpins . Apart from physical accessibilities, the relative abundances of other target proteases may become an important issue as an abundant target for PN-1 may "quench" a large amount of PN-1 by forming a non-reversible complex with it. However, the co-purification of prostasin in complex with PN-1  suggests that PN-1, at least to some degree plays a role as an inhibitor of prostasin in vivo. Likewise prostasin-HAI-1B complexes have been purified from a cell line , suggesting that it is a physiologically relevant inhibitor of prostasin. Further knowledge about the binding kinetics between the proteases and their inhibitors is needed to clarify the relative importance of PN-1 and HAI-1 isoforms as inhibitors of prostasin.