Báo cáo khoa học: Activation of hepatocyte growth factor activator zymogen (pro-HGFA) by human kallikrein 1-related peptidases docx

Activation of hepatocyte growth factor activator zymogen (pro-HGFA) by human kallikrein 1-related peptidases Shoichiro Mukai1,2, Tsuyoshi Fukushima1, Daiji Naka3, Hiroyuki Tanaka1, Yukio Osada2 and Hiroaki Kataoka1 Section of Oncopathology and Regenerative Biology, Department of Pathology, Faculty of Medicine, University of Miyazaki, Japan Department of Urology, Faculty of Medicine, University of Miyazaki, Japan Mitsubishi Chemical Medience Corporation R&D and Business Development Segment, Tokyo, Japan Keywords hepatocyte growth factor; HGF activator; KLK4; KLK5; tissue kallikrein Correspondence H Kataoka, Section of Oncopathology and Regenerative Biology, Department of Pathology, Faculty of Medicine, University of Miyazaki, 5200 Kihara, Kiyotake, Miyazaki 889-1692, Japan Fax: +81 985 85 6003 Tel: +81 985 85 2809 E-mail: mejina@med.miyazaki-u.ac.jp (Received December 2007, revised 29 December 2007, accepted January 2008) doi:10.1111/j.1742-4658.2008.06265.x Hepatocyte growth factor activator (HGFA) is a serine protease and a potent activator of prohepatocyte growth factor ⁄ scatter factor (proHGF ⁄ SF), a multifunctional growth factor that is critically involved in tissue morphogenesis, regeneration, and tumor progression HGFA circulates as a zymogen (pro-HGFA) and is activated in response to tissue injury Although thrombin is considered to be an activator of pro-HGFA, alternative pro-HGFA activation pathways in tumor microenvironments remain to be identified In this study, we examined the effects of kallikrein 1related peptidases (KLKs), a family of extracellular serine proteases, on the activation of pro-HGFA Among the KLKs examined (KLK2, KLK3, KLK4 and KLK5), we identified KLK4 and KLK5 as novel activators of pro-HGFA Using N-terminal sequencing, the cleavage site was identified as the normal processing site, Arg407–Ile408 The activation of pro-HGFA by KLK5 required a negatively charged substance such as dextran sulfate, whereas KLK4 could process pro-HGFA without dextran sulfate KLK5 showed more efficient pro-HGFA processing than KLK4, and was expressed in 50% (13 ⁄ 25) of the tumor cell lines examined HGFA processed by these KLKs efficiently activated pro-HGF ⁄ SF, and led to cellular scattering and invasion in vitro The activities of both KLK4 and KLK5 were strongly inhibited by HGFA inhibitor type 1, an integral membrane Kunitz-type serine protease inhibitor that inhibits HGFA and other proHGF ⁄ SF-activating proteases These data suggest that KLK4 and KLK5 mediate HGFA-induced activation of pro-HGF ⁄ SF within tumor tissue, which may thereafter trigger a series of events leading to tumor progression via the MET receptor In the pericellular microenvironment of tumor tissues, growth factors and proteases are critically important for tumor progression These factors enhance tumor cell proliferation, survival, motility, invasion, and angiogenesis Proteolytic activities are also essential for degrading components of the extracellular matrix or initiating coagulation and fibrinolytic systems In addition, several growth factors require proteolysis to gain full biological activity Hepatocyte growth factor ⁄ scatter factor (HGF ⁄ SF) is a multifunctional Abbreviations ACT, human a1-antichymotrypsin; AT, a1-antitrypsin; CHO, Chinese hamster ovary; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HAI, hepatocyte growth factor activator inhibitor; HAI-1KD-1, secreted form of hepatocyte growth factor activator inhibitor type consisting of the first Kunitz domain; HGF ⁄ SF, hepatocyte growth factor ⁄ scatter factor; HGFA, hepatocyte growth factor activator; KLK, human kallikrein 1-related peptidase; MDCK, Madin–Darby canine kidney; pAb, polyclonal antibody; PAI-1, plasminogen activator inhibitor-1; PC50%, processing concentration 50%; a2-AP, a2-antiplasmin FEBS Journal 275 (2008) 1003–1017 ª 2008 The Authors Journal compilation ª 2008 FEBS 1003 Activation of HGFA by tissue kallikreins S Mukai et al growth factor known to play an important role in tumor progression via its specific receptor tyrosine kinase MET, the c-met proto-oncogene product [1,2] HGF ⁄ SF is secreted primarily by stromal cells as an inactive single-chain precursor (pro-HGF ⁄ SF) that lacks biological activity and requires proteolytic cleavage to become the active, two-chain mature form [3] To date, several proteases have been reported to be HGF ⁄ SF-converting enzymes [3] Hepatocyte growth factor activator (HGFA) has been identified as a very potent serum activator of pro-HGF ⁄ SF [4,5] The membrane-anchored, cellular surface serine proteases, matriptase and hepsin, have also been reported as cellular activators of pro-HGF ⁄ SF [6–8] HGFA is primarily synthesized by the liver, and circulates in blood as an inactive zymogen (pro-HGFA) at a concentration of approximately 40 nm [3,9] Thrombin proteolytically activates pro-HGFA through cleavage of the Arg407–Ile408 bond in response to tissue injury, generating the two-chain active form consisting of a disulfide-linked 66 kDa long chain and 32 kDa light chain [10,11] The light chain exhibits the enzymatic activity, and the long chain is further cleaved by proteases [10,11] Activation can also occur in tumor tissue, and we have reported enhanced activity of HGFA and its involvement in the activation of HGF ⁄ SF in tumors [12–15] Although thrombin is considered to be an activator of pro-HGFA [3,10], alternative pathways in the activation of pro-HGFA in the tumor microenvironment remain to be clarified As the endogenous inhibitor of HGFA, HGFA inhibitor type (HAI-1), an integral membrane Kunitztype serine proteinase inhibitor, has been identified as a cell surface regulator of HGFA activity [3,16,17] Protein C inhibitor may also act as a serum inhibitor of HGFA [18] Human kallikrein 1-related peptidase (KLK) genes consist of 15 homologous serine protease genes located in tandem on chromosome 19q13.4 [19–21] The KLK proteins are translated as single-chain preproproteases, with cleavage of the signal peptide prior to secretion An additional cleavage of the propeptide is required for activation of KLKs After activation, KLK3, KLK7 and KLK9 have chymotrypsin-like activity, whereas other KLKs have trypsin-like activity in proteolysis [20] Autoactivation of KLK2 and KLK5 has been confirmed, and these enzymes can activate other coexpressed KLKs [22] Many members of the KLK family are associated with various human tumors, such as prostate, breast, ovary, colon, urothelial and renal cancers [23–28] For example, KLK3 (prostate-specific antigen) is a well-known tumor 1004 marker for prostatic cancer [23], and the expression of KLK5 is associated with unfavorable prognosis and invasiveness in breast, ovary and urothelial cancers [26–28] However, little is known regarding the underlying biological significance of KLK expression in tumors In this study, we examined the possible roles of KLKs (KLK2, KLK3, KLK4, and KLK5) in the HGF ⁄ SF signaling axis, focusing on their ability to generate active HGF ⁄ SF in the pericellular microenvironment via activation of HGFA We found that KLK5 is an efficient activator of pro-HGFA, and KLK4 also activates pro-HGFA to a lesser degree Results Expression of KLKs in cancer cell lines We characterized the expression of KLK2, KLK3, KLK4 and KLK5 mRNAs in a panel of human tumor cell lines The expression of KLKs such as KLK4 and KLK5 had already been extensively studied in ovarian and breast cancers For that reason, we examined cell lines of human tumors that were derived from other organs, in which activation of the HGF ⁄ SF signaling axis has been reported [1–3,8,12,13] Thus, we characterized colon (RCM-1, HT29, CaCo-2, HCT116, SW837, DLD-1, LoVo, Colo205), pancreas (S2-007, SUIT-2, AsPC1, Panc1, MiaPaCa), lung (HLC1, LC2, LC-1, T3M11, LU139), kidney (MRT-1, Caki-1), prostate (LNCap, PC3, DU145), and urinary bladder (KU-1 and UMK-1) In addition, expression of MET, HGF ⁄ SF, HGFA and HAI-1 was examined As shown in Fig 1, KLK2 was expressed only by the prostatic cancer cell line LNCap The expression of KLK3 (also known as PSA) was also limited, and was observed in LNCap Interestingly, DLD-1 derived from colon cancer also expressed low level of KLK3 KLK4 was abundantly expressed by LNCap and MRT-1, and to a lesser degree by DLD-1 On the other hand, low but distinct expression of KLK5 was observed in 13 out of 25 cell lines examined MET and HAI-1 were detectable in most (22 ⁄ 25 and 20 ⁄ 25, respectively) cell lines As the primer set for HAI-1 was designed to detect both HAI-1 and its splicing variant with similar inhibitory properties, HAI-1B [29], two PCR products (289 and 337 bp products for HAI-1 and HAI-1B, respectively) were observed None of the cell lines expressed notable levels of HGF ⁄ SF mRNA Although proHGFA is produced mainly by the liver and can be supplied from plasma in vivo [5,9], some tumor cell lines, such as DLD-1, LU139, MRT-1, and LNCap, also expressed endogenous HGFA FEBS Journal 275 (2008) 1003–1017 ª 2008 The Authors Journal compilation ª 2008 FEBS Activation of HGFA by tissue kallikreins Fig RT-PCR analyses for the expression of KLKs and HGF-related molecules in various human tumor cell lines Lung cancer Prostate Bladder RCC cancer cancer KLK2 KLK3 KLK4 KLK5 MET HGF/SF HGFA HAI-1 KU-1 UMK-1 Pancreas cancer Caco-2 HCT116 SW837 DLD-1 LoVo Colo205 S2-007 SUIT-2 AsPC1 Panc1 MiaPaCa HLC-1 LC-2 LC-1 T3M11 LU139 MRT-1 Caki-1 RCM-1 HT29 Colon cancer LNCap PC3 DU145 S Mukai et al HAI-1B HAI-1 GAPDH Processing of pro-HGFA by KLKs We directly tested the possibility that KLKs could activate pro-HGFA Before assaying, the enzymatic activity of each recombinant KLK was confirmed with synthetic chromogenic substrates (Fig 2A) KLK3 and KLK4 required activation by a processing protease, whereas KLK2 and KLK5 were autoactivated [22] As reported previously [10], thrombin efficiently processed pro-HGFA in the presence of a negatively charged substance (dextran sulfate, 10 lgỈmL)1) (Fig 2B) Under the same assay conditions, KLK4 and KLK5 also processed pro-HGFA, generating a 32 kDa light- chain band similar to thrombin-treated pro-HGFA under reducing conditions (Fig 2B) The N-terminal amino acid sequences of the 32 kDa processed bands generated by KLK4 and KLK5 were identified as Ile-Ile-Gly-Gly-Ser Therefore, thrombin, KLK4 and KLK5 cleaved pro-HGFA at the same sites (Arg407– Ile408) [10] The processing activity of KLK2 was weak, and generated a 34 kDa product similar to that generated by plasma kallikrein (data not shown) KLK3 did not show HGFA-processing activity (Fig 2B) Thrombin, a known activator of pro-HGFA, requires negatively charged substances such as dextran Fig Cleavage of pro-HGFA by KLKs (A) Enzymatic activities of recombinant KLKs measured with chromogenic substrates (B) Cleavage of pro-HGFA by KLKs Pro-HGFA (52 nM) was incubated for 12 h at 37 °C in the presence of 10 lgỈmL)1 dextran sulfate with nM of one of the following: thrombin, plasma kallikrein (p-kallikrein), KLK3, KLK4, or KLK5 Recombinant KLK3 and KLK4 were preincubated with 0.04 nM thermolysin to convert them to the active forms, and this was followed by addition of mM phosphoramidon Thermolysin also activated proHGFA, and the addition of phosphoramidon completely inhibited the activity of thermolysin Processing of pro-HGFA was determined by SDS ⁄ PAGE under reducing conditions followed by immunoblot analysis using a mAb to HGFA light chain (A-1) Both KLK4 and KLK5 generated 32 kDa fragments of HGFA The N-terminal amino acid sequences of the 32 kDa fragments were identical (Ile-Ile-Gly-Gly-Ser) for KLK4-mediated and KLK5-mediated processing FEBS Journal 275 (2008) 1003–1017 ª 2008 The Authors Journal compilation ª 2008 FEBS 1005 Activation of HGFA by tissue kallikreins S Mukai et al sulfate [10] Therefore, we checked the effect of dextran sulfate on KLK4 and KLK5 activation of pro-HGFA In the presence of dextran sulfate, KLK5 processed pro-HGFA more efficiently than did KLK4 (Fig 3A) The approximate concentrations of thrombin, KLK4 and KLK5 required to activate 50% of 52 nm proHGFA in the presence of 10 lgỈmL)1 dextran sulfate after 12 h at 37 °C (PC50%) were 0.035 nm, 0.45 nm, and 0.085 nm, respectively Thus, KLK5 appeared to be much more potent at activating pro-HGFA than KLK4, and its specific activity was about half that of thrombin In the absence of dextran sulfate, the processing of pro-HGFA by KLK5 and by thrombin was markedly attenuated (Fig 3B) In contrast, KLK4 activated pro-HGFA even in the absence of dextran sulfate (PC50%: 0.60 nm and 0.45 nm in the absence and presence of dextran sulfate, respectively) (Fig 3B) As reported previously [10], an alternatively cleaved 80 kDa product of pro-HGFA (cleavage at the Arg88– Ala89 bond) was generated by thrombin in the absence of dextran sulfate This 80 kDa product was not apparent in the case of KLK5 Further degradation of HGFA light chain was not observed even in the presence of high concentrations of KLK4 or KLK5, confirming that both KLKs are activators of pro-HGFA (Fig 3A,B) We also generated a time course for pro-HGFA processing by KLK In the presence of dextran sulfate, KLK5 was a more efficient activator than KLK4 (Fig 3C) Taken together, these findings show that in the presence of dextran sulfate, KLK5 was five to 10 times more potent than KLK4 in the processing of pro-HGFA By using antibodies (A-1, N19, and C20) that recognize different epitopes of HGFA, we further analyzed the cleavage patterns of pro-HGFA by KLKs (Fig 4) The epitope of each antibody is indicated in Fig 4B KLK5 cleaved the activation site in the first step to generate the active light chain, and then cleaved the Fig Dose-dependent and time-dependent processing of pro-HGFA by KLK4 and KLK5 and the effect of dextran sulfate (A, B) Pro-HGFA (52 nM) was incubated with various concentrations of KLK4 (0.1–10 nM), KLK5 (0.05–5 nM) or thrombin (0.05–5 nM) in the presence (A) or absence (B) of 10 lgỈmL)1 dextran sulfate The mixtures were subjected to SDS ⁄ PAGE under reducing conditions, and analyzed by immunoblot analysis The extent of processing (%) is shown (C) Time course of pro-HGFA processing in the presence of dextran sulfate Pro-HGFA (52 nM) was incubated with nM KLK4, nM KLK5 or nM thrombin in the presence of 10 lgỈmL)1 dextran sulfate for the indicated period at 37 °C The mixtures were subjected to SDS ⁄ PAGE under reducing conditions, and then subjected to immunoblot analysis 1006 FEBS Journal 275 (2008) 1003–1017 ª 2008 The Authors Journal compilation ª 2008 FEBS S Mukai et al Activation of HGFA by tissue kallikreins Fig Analysis of cleavage sites of pro-HGFA by KLKs (A) Time-dependent cleavage patterns of pro-HGFA (52 nM) were analyzed by using three kinds of antibody to HGFA (A-1, C20, and N19) under reducing or nonreducing conditions The epitope of each antibody is indicated in (B) A schematic representation of each band in immunoblot analysis is also indicated (right and lower panels) The results obtained under nonreducing conditions indicate the existence of multiple disulfide bonds in the heavy chain, as reported previously [5,10] (B) Schematic representation of the cleavage sites of pro-HGFA and epitopes of the antibodies A-1 recognizes the light chain of active HGFA (hatched bar), but the precise position of the epitope is not known The intra-heavy-chain disulfide bonds are not indicated heavy chain, generating 41–32 kDa fragments (Fig 4A) Similar, but less efficient, cleavage patterns were also observed in KLK4 (not shown) A schematic representation of each band observed in immunoblot is indicated in Fig 4A, and that of the cleavage sites is shown in Fig 4B Inhibition of KLK-mediated pro-HGFA activation by serpins and HAI-1 The pro-HGFA-processing activity of KLK4 was inhibited by plasma serine protease inhibitors such as human a1-antichymotrypsin (ACT), a1-antitrypsin (AT), and a2-antiplasmin (a2-AP), and to a lesser degree by plasminogen activator inhibitor-1 (PAI-1) (Fig 5) On the other hand, KLK5 was inhibited strongly by PAI-1 and a2-AP, but not by ACT and AT (Fig 5) Pro-HGFA processing by both KLKs was potently inhibited by HAI-1KD1, a truncated form of recombinant HAI-1 containing the first Kunitz domain, which is the major functional inhibitor domain against cognate proteases [29,30] (Fig 5) As mature HAI-1 is a membrane-anchored inhibitor expressed on the surface of various epithelial cells and FEBS Journal 275 (2008) 1003–1017 ª 2008 The Authors Journal compilation ª 2008 FEBS 1007 Activation of HGFA by tissue kallikreins S Mukai et al Fig Inhibition of KLK-mediated pro-HGFA activation by serpins and HAI-1 KLK4 or KLK5 was preincubated with each protease inhibitor (ACT, PAI-1, AT, a2-AP, or HAI-1KD1) at a : 10 molar ratio, and each mixture was used for pro-HGFA activation assay The final concentrations of pro-HGFA, KLK4 and KLK5 were 52 nM, nM, and nM, respectively The mixtures were subjected to SDS ⁄ PAGE under reducing conditions, and then subjected to immunoblot analysis tumor cells [3,12,31,32], the activities of these KLKs could be regulated by HAI-1 within the pericellular microenvironment Degradation of pro-HGF ⁄ SF by KLK4 and KLK5 We examined whether KLK4 and KLK5 were able to activate not only pro-HGFA but also pro-HGF ⁄ SF However, as shown in Fig 6, pro-HGF ⁄ SF was degraded by KLK4 and KLK5 in a dose-dependent manner Although, the physiological cleavage of proHGF ⁄ SF may occur very inefficiently when the proHGF ⁄ SF is incubated with a low concentration of KLK5, pro-HGF ⁄ SF was almost completely degraded into small fragments when the KLK5 ⁄ pro-HGF ⁄ SF ratio was set higher than 0.1 Similar findings were obtained with KLK4 (Fig 6) On the other hand, HGFA showed very efficient activation of proHGF ⁄ SF (Fig 6) without any degradation, even at high enzyme ⁄ substrate ratios (data not shown) Biological roles of KLK-dependent activation of HGFA We wanted to investigate whether KLK-mediated activation of pro-HGFA and subsequent processing of pro-HGF ⁄ SF induced cellular responses via the MET receptor tyrosine kinase Thus, we examined the phosphorylation of MET The addition of pro-HGF ⁄ SF, which was preincubated with KLK5-treated proHGFA, rapidly activated the cellular MET receptor (Fig 7A,B) We also examined the effects on cellular scattering Madin–Darby canine kidney (MDCK) cells were treated with pro-HGF ⁄ SF in the presence or absence of KLK5-treated pro-HGFA Enhanced cellular scattering was observed at 12 h after the treatment when the cells were treated concomitantly with proHGF ⁄ SF and KLK5-activated HGFA (Fig 7C) Therefore, HGFA activated by KLK5 appears to be 1008 functional On the other hand, at 24 h after treatment, cells treated with pro-HGF ⁄ SF and pro-HGFA also showed cellular scattering This may be due to processing of pro-HGFA or pro-HGF ⁄ SF by endogenous protease of MDCK cells Finally, we used a KLK5-negative tumor cell line to determine the effect of engineered expression of KLK5 on pericellular activation of the HGFA–HGF ⁄ SF axis For this purpose, we selected SUIT-2, because this cell line did not express HGFA (Fig 1) and the expression levels of the HGF ⁄ SF activators such as matriptase and hepsin were low (data not shown) As shown in Fig 8A, cellular KLK5 induced pro-HGF ⁄ SF activation via processing of pro-HGFA It should be noted that, although HGF ⁄ SF could be degraded by KLK5 in an in vitro tube assay (Fig 6B), the cell-based assay revealed that KLK5 ⁄ HGFA-mediated activation of pro-HGF ⁄ SF worked without HGF ⁄ SF being degraded In a migration assay (Fig 8B), pro-HGF ⁄ SF enhanced migration of SUIT-2 cells even in the absence of HGFA; this may be caused by a trace of activated HGF ⁄ SF contaminating the pro-HGF ⁄ SF preparation (as shown in the immunoblot in Fig 8A) or by endogenous pro-HGF ⁄ SF-activating protease of SUIT-2 However, in the co-presence of pro-HGFA and pro-HGF ⁄ SF, the expression of KLK5 significantly upregulated cellular migratory capability as compared with corresponding control cells (Fig 8B) The data suggested that pericellular pro-HGFA derived from plasma or tumor cells may be activated by tumor cell-derived KLK5, which may thereafter trigger a series of events leading to cellular invasion via HGF ⁄ MET signaling Discussion In the present study, we showed that KLK5 activates pro-HGFA, resulting in activation of pro-HGF ⁄ SF and MET-mediated cellular responses KLK4 also FEBS Journal 275 (2008) 1003–1017 ª 2008 The Authors Journal compilation ª 2008 FEBS S Mukai et al Fig Degradation of pro-HGF ⁄ SF by KLK4 and KLK5 (A, B) ProHGF ⁄ SF (53 nM) was incubated at various concentrations of KLK4 (A) or KLK5 (B) at 37 °C for 12 h, and the mixtures were subjected to SDS ⁄ PAGE and analyzed by immunoblot For positive control of processing, the same amount of pro-HGF ⁄ SF was incubated with 0.05 nM HGFA at 37 °C for 12 h and simultaneously analyzed hc, heavy chain of mature active form HGF ⁄ SF showed pro-HGFA-processing activity, although the specific activity was lower than that of KLK5 As negatively charged substances, particularly dextran sulfate, stimulated the activation of HGFA by thrombin [10], we also examined the effect of dextran sulfate on KLK-mediated HGFA activation We found that KLK5 required dextran sulfate to efficiently activate pro-HGFA, whereas KLK4 activated pro-HGFA even in the absence of dextran sulfate The mechanism by which dextran sulfate enhanced KLK5-mediated proHGFA processing remains to be determined As the Activation of HGFA by tissue kallikreins pericellular microenvironment is rich in negatively charged substances such as glycosaminoglycans, and both HGFA and HGF ⁄ SF also show affinity for negatively charged substances [3], it is reasonable to postulate that, in tumors expressing KLK4 and ⁄ or KLK5, efficient HGFA-activating machinery would be generated in the pericellular microenvironment Pro-HGFA is abundant in plasma [9], and is also expressed by certain human cancers [12,14,33–35], whereas pro-HGF ⁄ SF is produced by stromal cells and is significantly increased in tumor tissues via interactions between tumor cells and stromal cells [36] Therefore, these results reveal a novel mechanism in the control of cellular invasiveness that involves an upstream tumor cell-derived activator and downstream stromal effectors in tumor tissue Kallikrein 1-related peptidases are expressed in various tissues, and are implicated in several physiological and pathological conditions KLK5 is expressed in normal skin as stratum corneum tryptic enzyme and in the prostate [22,37] In the epidermis, KLK5 activates pro-KLK7 (known as stratum corneum chymotryptic enzyme), and cleaves the components of corneodesmosomes, leading to desquamation in a coordinated manner with KLK7 [38] In the prostate, KLK5 is secreted into the prostatic fluid Self-activated KLK5 converts KLK2 and KLK3 into their active forms After ejaculation, these KLKs degrade semenogelins, the components of seminal clots, to release sperm [22] KLK4 was first characterized as enamel matrix serine protease 1, and was reported to be involved in enamelogenesis by processing enamelin [39] KLK4 is also expressed abundantly in the prostate and secreted into the prostatic fluid; however, its physiological function in the prostate is not understood, except for the proteolytic activation of KLK3 Recently, KLKs have been studied in terms of their diagnostic and prognostic values in some cancers [23] KLK5 appears to be a potential biomarker of ovarian and breast cancers [23,26,27], and may be involved in the progression of prostate cancer [23–25] Moreover, overexpression of the KLK5 gene is associated with invasiveness of urinary bladder carcinoma cells [28] In this study, KLK5 was expressed in more than 50% of the tumor cell lines examined However, the mechanisms underlying the role of KLK5 in tumor progression are poorly understood KLK5 and other KLKs activated by KLK5 may degrade extracellular matrix proteins such as fibronectin, laminin, and type IV collagen, which facilitate the cellular invasiveness at the invasion front [37] Insulin-like growth factor-binding proteins are also possible targets of KLKs [21,40] Degradation of extracellular insulin-like growth FEBS Journal 275 (2008) 1003–1017 ª 2008 The Authors Journal compilation ª 2008 FEBS 1009 Activation of HGFA by tissue kallikreins S Mukai et al B Incubation time (m) A Pro-HGF/SF Pro-HGFA KLK5 + + + + + + + + 10 10 10 p-MET 94 Pro-HGF/SF hc 62 kDa 145 Total MET 145 kDa ratio