Biol. Chem. 2016; 397(12): 1299–1305

Short Communication

Janet C. Reid, Nigel C. Bennett, Carson R. Stephens, Melanie L. Carroll, Viktor Magdolen, Judith A. Clements and John D. Hooper* In vitro evidence that KLK14 regulates the components of the HGF/Met axis, pro-HGF and HGF-activator inhibitor 1A and 1B

DOI 10.1515/hsz-2016-0163 Received April 1, 2016; accepted August 10, 2016; previously -related peptidase (KLK) 14 (Hooper et al., ­published online August 17, 2016 2001; Yousef et al., 2001) is one of the 15 members of the human KLK family of -fold serine proteases Abstract: Kallikrein-related peptidase (KLK) 14 is a serine which also includes the prostate cancer biomarker KLK3 protease linked to several pathologies including prostate (or prostate-specific antigen; PSA) (Dong et al., 2014). cancer. We show that KLK14 has biphasic effects in vitro KLK14 spans 267 residues incorporating a signal peptide, on activating and inhibiting components of the pros- that instructs for secretion, and a short pro-sequence tate cancer associated hepatocyte growth factor (HGF)/ that is cleaved at Lys40-Ile41 to generate the catalyti- Met system. At 5–10 nm, KLK14 converts pro-HGF to cally active protease. Mature KLK14 has trypsin-like the two-chain heterodimer required for Met activation, activity with preference for cleavage after Arg (Brattsand while higher concentrations degrade the HGF α-chain. et al., 2005; Felber et al., 2005; Oikonomopoulou et al., HGF activator-inhibitor (HAI)-1A and HAI-1B, which 2006; Stefansson et al., 2006; Borgono et al., 2007b; inhibit pro-HGF activators, are degraded by KLK14 when Rajapakse and Takahashi, 2007; de Veer et al., 2012) protease:inhibitor stoichiometry is 1:1 or the protease is and there is also evidence for low-level - in excess. When inhibitors are in excess, KLK14 generates like activity (Brattsand et al., 2005; Felber et al., 2005). HAI-1A and HAI-1B fragments known to inhibit pro-HGF KLK14 is a positive regulator of the proteolytic cascade activating serine proteases. These in vitro data suggest that mediates skin desquamation, and the absence of that increased KLK14 activity could contribute at multi- its inhibition is implicated in the skin barrier disorder ple levels to HGF/Met-mediated processes in prostate and (Prassas et al., 2015). In addition, other cancers. serum KLK14 levels are significantly elevated in pros- tate cancer patients compared with healthy men, and Keywords: hepatocyte growth factor; HGF inhibitor; tissue levels correlate with prostate tumor status and ­kallikrein-related peptidase 14; Met; prostate cancer. are associated with disease progression defined by PSA relapse (Borgono et al., 2007b; Rabien et al., 2008). A functional link between increased KLK14 expression *Corresponding author: John D. Hooper, Mater Research Institute – and prostate cancer is suggested by its regulation of University of Queensland, Woolloongabba, Queensland 4102, that modulate progression of this malignancy. Australia, e-mail: [email protected] For example, KLK14 cleaves insulin-like growth factor Janet C. Reid: Mater Research Institute – University of Queensland, binding proteins, thereby increasing the availability of Woolloongabba, Queensland 4102, Australia; and Institute of insulin-like growth factor (IGF) to promote cell prolifera- Health and Biomedical Innovation, Translational Research Institute, Queensland University of Technology, Woolloongabba, Queensland tion (Borgono et al., 2007b). In addition, KLK14 modifies 4102, Australia the prostate cancer microenvironment via degradation Nigel C. Bennett, Carson R. Stephens , Melanie L. Carroll of the androgen transporter sex hormone-binding globu- and Judith A. Clements: Institute of Health and Biomedical lin (Sanchez et al., 2012), disruption of cell-cell contacts Innovation, Translational Research Institute, Queensland University via cleavage of desmoglein (Borgono et al., 2007a), and of Technology, Woolloongabba, Queensland 4102, Australia Viktor Magdolen: Klinische Forschergruppe der Frauenklinik der degradation of extracellular matrix proteins including Technischen Universität München, Klinikum rechts der Isar, Munich, collagens, laminin, fibronectin and vitronectin (Borgono Germany et al., 2007b; Rajapakse and Takahashi, 2007).

Brought to you by | EP Ipswich Authenticated Download Date | 11/9/16 5:27 AM 1300 J.C. Reid et al.: KLK14 regulates pro-HGF, HAI-1A and HAI-1B

Hepatocyte growth factor (HGF) mediates cell the most poorly differentiated tumors exhibiting the growth, cell motility and morphogenesis in numerous lowest levels of expression (Bergum and List, 2010; Tsai cell and tissue types via binding to the Met tyrosine et al., 2014). kinase receptor (Cecchi et al., 2012). In prostate cancer The overlapping elevated expression of KLK14 and HGF is abundantly expressed by stromal cells present HGF in prostate cancer, and the specificity that KLK14 has within the tumor microenvironment and serum levels for cleavage after Arg residues, which potentially includes are a strong predictor of metastasis to lymph nodes and the canonical activation site of pro-HGF (Arg494-Val495), disease recurrence (Gupta et al., 2008; Yasuda et al., prompted us to explore the role of KLK14 in the regula- 2009; Varkaris et al., 2011). Met is overexpressed by tion of the HGF/Met system. For this purpose we gener- malignant cells in primary tumors and is associated ated and purified recombinant KLK14 from insect cells, with high Gleason grade and tumor vascularization, first characterizing its enzymatic properties, then evalu- and potentially tumor stage (Nakashiro et al., 2003; ating its impact in vitro on pro-HGF and the inhibitors ­Strohmeyer et al., 2004; Jacobsen et al., 2013). Thus, par- HAI-1A and HAI-1B. To generate a suitable KLK14 expres- acrine signaling via the actions of stroma-produced HGF sion construct, the previously cloned coding sequence on the Met receptor, which is elevated on the surface of (Hooper et al., 2001), including the signal and pro-region malignant cells, is thought to be important in prostate sequence, was cloned in-frame in a pIB/V5-His expres- cancer progression and metastasis (Trusolino et al., sion vector (Invitrogen, Life Technologies Australia Pty 2010). HGF is produced as a pro-factor requiring activa- Ltd, Mulgrave, Victoria, Australia) generating KLK14 with tion to generate a two-chain heterodimer that includes a carboxyl-terminal V5 tag (GKPIPNPLLGLDST) and six an α-chain, containing four kringle domains, and a β- histidine residues. Insect Spodoptera frugiperda Sf9 cells chain that has homology to trypsin-fold serine proteases. (Gibco, Life Technologies, Mulgrave, Victoria, Australia) Activation occurs at Arg494-Val495 and is mediated by were stably transfected with the KLK14-V5-His construct the circulating Arg/Lys-specific trypsin-fold serine pro- and KLK14 was purified from the conditioned media using tease HGF-activator (HGFA) (Miyazawa, 2010) and the Ni-NTA Superflow resin (Qiagen, Doncaster, Victoria, plasma membrane serine proteases (Lee ­Australia) eluting in the manufacturer’s supplied elution et al., 2000; Kirchhofer et al., 2003) and (Herter buffer (Figure 1A). Pooled fractions of purified recombi- et al., 2005; Kirchhofer et al., 2005). The actions of these nant KLK14 were concentrated and dialyzed at 4°C against serine proteases in initiation of the HGF/Met signaling 5 mm NaH2PO4, 95 mm Na2HPO4, 0.01% (v/v) Tween20, axis are commonly negatively regulated by the Kunitz- pH 8.0 (Borgono et al., 2007b). Activation was then per- type inhibitors HGFA-inhibitor (HAI)-1, which has two formed by incubating KLK14 with the metalloprotease variants HAI-1A and HAI-1B, and HAI-2 (Kawaguchi and thermolysin (Calbiochem, San Diego, CA, USA) in the Kataoka, 2014). These inhibitors block proteolysis via ratio 1:12.5 (thermolysin:KLK4) for 90 min at 37°C. Activa- tight association, as a pseudo-substrate, with the active tion was apparent from the reduction in molecular weight site of the (Shia et al., 2005). The roles of KLK14 of ~4 kDa as a result of loss of the pro-domain of HAI-1 and HAI-2 in prostate cancer are not yet well (Figure 1B). For comparison we employed thermolysin-­ defined and recent expression studies suggest poten- activated recombinant KLK4, generated and purified as tially disparate roles for these inhibitors. For example, previously described (Ramsay et al., 2008), and bovine in contrast with the usual paradigm that reduced pro- trypsin (­Worthington Biochemical, Lakewood, NJ, USA), tease inhibitor levels contribute to protease-mediated or the soluble auto-activating form of hepsin lacking processes in cancer progression, HAI-1 levels are its intracellular and transmembrane domains (Kirch- higher in serum of patients with prostate cancer com- hofer et al., 2005) (generously provided by Dr. Daniel pared with men who have benign prostatic hyperplasia, ­Kirchhofer, Genentech Inc., San Francisco, CA, USA). and are also elevated in men who have metastasis and The proportion of active KLK4, KLK14, trypsin and hepsin castrate resistant prostate cancer compared to patients present in relevant assays was determined by incubat- with organ-confined disease (Nagakawa et al., 2006). ing protease (300 nm) with the suicide pseudo-substrate Similar changes were reported in prostate tumors with 4-methylumbelliferyl 4-guanidinobenzoate (MUGB; 1 μm) HAI-1 protein levels higher in patients with prostate (Jameson et al., 1973) in 50 mm Tris, 50 mm NaCl, pH 8.8, cancer compared with BPH (Warren et al., 2009; Yasuda 0.01% (v/v) Tween20 for 10 min at 37°C. Stoichiometric 1:1 et al., 2013). In contrast with HAI-1 expression, levels of release of fluorescent 4-MU from MUGB for every active HAI-2 protein are significantly decreased in malignant protease molecule was measured using a POLARstar plate lesions compared with normal and benign lesions, with reader (BMG Labtech, Mornington, Victoria, Australia) by

Brought to you by | EP Ipswich Authenticated Download Date | 11/9/16 5:27 AM J.C. Reid et al.: KLK14 regulates pro-HGF, HAI-1A and HAI-1B 1301

A Elutions BSA (µg) B hydrolysis of the bond between the P1 Arg residue releas-

1234 0 5 ing fluorescent AMC. Active KLK14 (12 nm), KLK4 (12 nm) kDa 1. 0. Thermolysin– + 75 kDa 37 and trypsin (0.4 nm) were incubated at 37°C with increas- 50 37 ing concentrations (0.001–1 mm) of QAR-AMC, with fluo- 25 rescence measured each minute for 20 min using a plate 25 20 20 reader with 380 nm excitation and emission at 460 nm. α-KLK14 A standard curve of fluorescence generated from known 15 concentrations of AMC was used to determine the rate of hydrolysis. Employing non-linear regression and GraphPad

C Concentration KM Vmax kcat kcat/KM Prism 5 software (GraphPad Software, San Diego, CA, USA), –1 –1 –1 –1 Protease (nM) (µM) (µM s ) (s ) (M s ) the Michaelis-Menten enzyme kinetics model was used to calculate the substrate concentration ([S]) at which reac-

4 KLK14 12 258 ±110.145 12.1 4.7 ×10 tion velocity is half maximal, KM, and the maximal reaction

velocity, Vmax, and to derive, using the protease concentra- 4 KLK4 12 124 ±4 0.0171.4 1.1 ×10 tion and Vmax, the substrate per molecule­ of protease turno-

ver rate, kcat, and the ratio kcat/KM, as a measure of catalytic Trypsin 0.4 38 ±4 0.045114.8 3.0 ×106 efficiency. As shown in Figure 1C, KLK14 has a high KM of 258 μm, suggesting a lower affinity for the substrate than either KLK4 or trypsin. The substrate turnover rate of k D Insect Mammalian cat for KLK14 is higher than that of KLK4, which yields a higher KLK14 KLK4 KLK14 KLK4 catalytic efficiency, k /K , of 4.7 × 104 m−1s−1, for KLK14 PNGase F –+–+ –+–+ cat M kDa under the conditions tested. As expected, trypsin had the 37 lowest KM and highest kcat and, as a result, the highest cata- 25 lytic efficiency for the substrate QAR-AMC. α-V5 α-Myc α-V5 Interestingly, using the same substrate and assay para-

meters, Borgono and colleagues obtained a kcat/KM ~3-fold Figure 1: Generation, purification and characterization of recombi- higher, at 14.2 × 104 m−1s−1, for recombinant KLK14 gener- nant KLK14. ated from yeast (Borgono et al., 2007b), than the value (A) Coomassie stained SDS-PAGE gel of elution fractions 1–4 (30 μl) of pro-KLK14-V5-His from Ni-NTA agarose beads. Known amounts obtained by us. This difference may be due to differences of BSA were used to estimate KLK14 concentration. (B) Western in the approaches used to measure active levels of KLK14. blot analysis, using an anti-KLK14 antibody (Abcam ab28841), of Whereas we employed the suicide pseudo-­substrate recombinant, purified pro-KLK14-V5-His either untreated (−) or MUGB, Borgono and colleagues used α1-antitrypsin incubated with thermolysin (10:1 μg/ml) for 1 h at 37°C. (C) Kinetic and subsequent reduction in QAR-AMC hydrolysis to parameters for hydrolysis of the tri-peptide substrate QAR-AMC by thermolysin-activated KLK14 and KLK4, and trypsin. Data are aver- monitor levels of active KLK14. It is also possible that ages of triplicate experiments ± SEM. (D) Anti-V5 (Invitrogen R96025) differences in kcat/KM values may be due to differences in or anti-Myc (Cell Signaling Technology 9B11) Western blot analysis ­N-glycosylation. Borgono and co-workers showed that of purified recombinant pro-KLK14-V5-His and pro-KLK4-V5-His from recombinant KLK14 from P. pastoris is not post-transla- insect cells, and conditioned media from mammalian Cos-7 cells tionally modified (Borgono et al., 2007b), even though the transiently expressing KLK14-Myc or KLK4-V5-His. Samples were protein sequence contains a consensus N-glycosylation either untreated (−) or treated (+) with the Peptide-N-Glycosidase F (PNGase F) which removes Asn-linked glycans from glycoproteins. site at Asn184-Ile-Ser. As N-glycosylation is known to alter enzyme activity (Skropeta, 2009), we further exam- ined this issue by assessing the N-linked glycan content excitation at 355 nm with emission at 460 nm. Concentra- of KLK14 from insect cells by treating the purified pro- tion of active protease was calculated using the relative tease with the enzyme peptide-N-glycosidase F (PNGase fluorescence against a standard curve generated from F), which removes these modifications from glycopro- known concentrations of 4-MU (0–1 μm). teins. As controls, we examined the effect of PNGase F on Kinetic parameters were assessed using the synthetic insect cell generated recombinant KLK4, as well as KLK14 tri-peptide substrate N-tert-butoxycarbonyl-Gln-Ala-Arg/7- and KLK4 that were transiently expressed by monkey amino-4-methylcoumarin (QAR-AMC), which is known to kidney Cos-7 cells. As shown in Figure 1D, both insect cell- be efficiently hydrolyzed by KLK14 (Borgono et al., 2007b; expressed proteases contained ~2 kDa of N-glycosylation, ­Rajapakse and Takahashi,­ 2007). The assay relies on whereas this modification was not apparent from KLK14

Brought to you by | EP Ipswich Authenticated Download Date | 11/9/16 5:27 AM 1302 J.C. Reid et al.: KLK14 regulates pro-HGF, HAI-1A and HAI-1B

and KLK4 transiently expressed by mammalian cells. It is 00.1 0.51 51050100 100 Protease (nM) ++++++++– Pro-HGF (1 µM) not yet known whether this N-glycan content could alter kDa A 100 protease:substrate interactions sufficiently to reduce the 75 Pro-HGF HGF α-chain kinetics of insect cell expressed KLK14 compared with the 50 37 HGF degradation non-glycosylated protease from yeast. However, N-glyco- KLK14 HGF β-chain sylation is required for matriptase autoactivation (Oberst 25 KLK14 et al., 2003; Miyake et al., 2010), and attachment of aber- 20 HGF 15 rant N-glycan structures increases matriptase stability degradation and metastasis (Ihara et al., 2002, 2004; Ito et al., 2006). B 100 Pro-HGF 75 HGF α-chain To examine the role of KLK14 in activation of pro-HGF, 50 HGF degradation increasing concentrations of the activated protease were Hepsin 37 HGF β-chain incubated with the recombinant growth factor. Hepsin and 25 Hepsin 20 KLK4 were assayed concurrently as examples of serine pro- HGF teases that, respectively, activate (Herter et al., 2005; Kirch- 15 degradation hofer et al., 2005; Owen et al., 2010) and degrade (Mukai C 100 Pro-HGF 75 et al., 2008) pro-HGF. Active protease from 0.1 to 100 nm HGF α-chain 50 was incubated with recombinant pro-HGF (1 μm) for 4 h at HGF KLK4 37 degradation 37°C. Reaction mixtures were resolved by SDS-PAGE under HGF β-chain 25 KLK4 reducing conditions then transferred onto nitrocellulose 20 membranes and stained using Coomassie Blue R-250 HGF 15 degradation (Sigma-Aldrich Pty Ltd, Castle Hill, NSW, Australia) to visu- alize protein bands. As shown in Figure 2A, KLK14 cleavage Figure 2: Proteolysis of pro-HGF by KLK14, hepsin and KLK4. generates pro-HGF fragments in a concentration-depend- Recombinant KLK14 (A), hepsin (B) and KLK4 (C) (0–100 nm) were incubated with recombinant pro-HGF (1 μm) for 4 h at 37°C. Reaction ent manner. Between KLK14 concentrations of 1 and 5 nm mixtures were resolved by SDS-PAGE under reducing conditions (protease:substrate ratio 1:1000 and 1:200), pro-HGF is then transferred onto nitrocellulose membrane and stained using converted to the two-chain heterodimer, with the protein Coomassie Blue R-250 to visualize protein bands. Each protease- fragments the same as reported when HGFA, matriptase mediated conversion of pro-HGF above background levels in a and hepsin generate active HGF (Lee et al., 2000; Kirch- concentration-dependent manner. hofer et al., 2005; Miyazawa, 2010). We are yet to confirm that KLK14 mediated cleavage of pro-HGF occurs at the canonical activation site Arg494-Val495. At higher con- examination of HAI-1A and HAI-1B as inhibitors and as sub- centrations (10–100 nm; protease:substrate ratio 1:100 to strates of KLK14. We selected HAI-1 for analysis, and not 1:10), the HGF α-chain is completely degraded and the HAI-2, because levels of the former are elevated in prostate β-chain remains largely intact. As expected, hepsin dem- cancer, whereas HAI-2 levels are decreased (Nagakawa onstrates efficient concentration-dependent conversion of et al., 2006; Warren et al., 2009; Bergum and List, 2010; pro-HGF to the two-chain form (Figure 2B). At concentra- Yasuda et al., 2013; Tsai et al., 2014). Inhibition was exam- tions of hepsin as low as 0.1 nm (molar ratio of 10 000:1), ined by assessing the impact of increasing concentrations of there is near complete conversion of pro-HGF. In contrast the inhibitors on KLK14 cleavage of the tri-peptide substrate with the effect of KLK14, even at higher hepsin concen- QAR-AMC. The assay employed soluble forms of HAI-1A and trations there was little evidence of degradation of the HAI-1B that lack transmembrane and intracellular regions HGF α-chain (Figure 2B). At concentrations up to 10 nm, (Kirchhofer et al., 2003; Fan et al., 2005) (provided by Dr. KLK4 demonstrates little activity toward pro-HGF, and at Daniel Kirchhofer). Each inhibitor was incubated over the higher concentrations it proceeds to degrade the HGF α- concentration range 0.25–50 nm with active KLK14 (5 nm) chain (Figure 2C). These data suggest the possibility that for 30 min at 37°C before addition of QAR-AMC (200 μm). KLK14 has a biphasic role in regulation of pro-HGF that Fluorescence of released AMC was measured as detailed involves activation when the substrate is in vast excess above. As shown in Figure 3A, there is no inhibition of (protease:substrate ratio 1:1000 to 1:200) and degradation KLK14 up to equimolar concentrations (5 nm) of HAI-1A or as pro-HGF concentration increases (protease:substrate HAI-1B. As the reaction shifts to containing excess inhibi- ratio 1:100 to 1:10). tor, KLK14 is mildly inhibited, with inhibition of ~35% and We next explored KLK14 regulation of mechanisms ~26% obtained when HAI-1A and HAI-1B, respectively, were that negatively impact on HGF/Met signaling. This involved present at 10-fold excess over protease.

Brought to you by | EP Ipswich Authenticated Download Date | 11/9/16 5:27 AM J.C. Reid et al.: KLK14 regulates pro-HGF, HAI-1A and HAI-1B 1303

A 140 HAI-1A or HAI-1B (Figure 3B and C). Instead, each inhibi- 120 tor is extensively degraded when the stoichiometry is 1:1 100 or the protease is in excess. When HAI-1A and HAI-1B are 80 in excess, KLK14 processing generates predominant prod- HAI-1A 60 ucts at ~15 and 40 kDa with less intense protein bands at HAI-1B ~ 40 22 and 35 kDa. It has previously been reported that the

KLK14 percent activity 40 kDa fragment contains Kunitz domain 1 of HAI-1 (Lin 20 et al., 1999) which is the region that mediates inhibition 0 0.1 110 100 of the membrane anchored serine protease and pro-HGF

HAI-1 (nM) activator TMPRSS13 (Hashimoto et al., 2010) and forms a stable complex with another activator matriptase (Benaud

010.5 1 5 10 0 Inhibitor (µM) et al., 2001). Thus, when the inhibitor is in excess, KLK14

kDa +++ ++– KLK14 (1 µM) releases HAI-1A and HAI-1B fragments that likely mediate 250 B HAI-1A complexes inhibition of pro-HGF convertases. 100 75 In summary, we generated, purified and activated HAI-1A 50 KLK14 from an insect cell expression system, and used this HAI-1A 37 HAI-1A cleaved to examine the role of the protease in the regulation of com- ponents of the HGF/Met receptor system. The lower catalytic 25 KLK14 20 efficiency observed by us against a peptide substrate for HAI-1A cleaved insect cell expressed KLK14, compared with the previously 15 reported yeast expressed protease (Borgono et al., 2007b), C 250 is possibly due to N-glycosylation present on the former but 100 75 not the latter. If this is the case it is possible that the non- HAI-1B 50 glycosylated KLK14 observed by us from mammalian cells, 37 HAI-1B cleaved could have even greater activity against the macro-molec- HAI-1B ular substrates pro-HGF and HAI-1A and HAI-1B than we 25 KLK14 observed above. Our data indicate that KLK14 has biphasic 20 HAI-1B cleaved effects on pro-HGF, and HAI-1A and HAI-1B in vitro. When 15 pro-HGF is in vast excess it is activated by KLK14, and as the substrate concentration increases above a threshold level Figure 3: KLK14 interactions with the Kunitz serine protease inhibi- tors HAI-1A and HAI-1B. (protease:substrate ratio above 1:200), HGF is inactivated. (A) HAI-1A and HAI-1B (0–50 nm) inhibition of KLK14 (5 nm) prote- Similarly, HAI-1A and HAI-1B are extensively degraded olysis of the tri-peptide substrate QAR-AMC (200 μm). The rate of when the protease:inhibitor stoichiometry is 1:1, or KLK14 hydrolysis of QAR-AMC by uninhibited KLK14 is set as 100% activity is in excess. When the inhibitors are in excess, KLK14 pro- and the rate of HAI-1A- and HAI-1B-inhibited KLK14 as a ­percentage cessing generates fragments that likely mediate inhibition thereof. Data are averages of triplicate experiments ± SEM. (B, C) KLK14 (1 μm) proteolysis of HAI-1A and HAI-1B (0.5–10 μm) after of pro-HGF convertases. These data suggest that increased incubation for 30 min at 37°C. Reactions were stopped with reduc- KLK14 activity could contribute at multiple levels to HGF/ ing SDS-PAGE loading buffer and subjected to gel electrophoresis, Met-mediated processes in prostate and other cancers. then transferred to PVDF membranes and stained with Coomassie Further experiments are required to examine this in disease R-250. relevant models.

To monitor KLK14 interactions with HAI-1A and Acknowledgments: We thank Dr. Yaowu He (­University HAI-1B, the protease and inhibitor were incubated in of Queensland) for technical advice, and Dr. Daniel molar ratios of 1:0.5, 1:1, 1:5 and 1:10 for 30 min at 37°C. ­Kirchhofer (Genentech) for the generous gift of recom- Reaction mixtures were resolved by SDS-PAGE under binant hepsin, HAI-1A and HAI-1B. This work was sup- reducing conditions then transferred onto nitrocellulose ported by funding from the National Health and Medical membranes and stained using Coomassie Blue R-250. In Research Council of Australia (grant 614206), Cancer contrast with the proteases HGFA (Kataoka et al., 2002), Council Queensland (grant APP1021827), Prostate Can- prostasin (Fan et al., 2005) and matriptase (Kiyomiya cer Foundation of Australia (grant PG 3810) and Austral- et al., 2006) that form stable complexes with HAI-1, there ian Research Council (Future Fellowship FT120100917) was no evidence of complex formation between KLK14 and to J.D.H., a mobility grant (Personalized Medicine) from

Brought to you by | EP Ipswich Authenticated Download Date | 11/9/16 5:27 AM 1304 J.C. Reid et al.: KLK14 regulates pro-HGF, HAI-1A and HAI-1B the German Academic Exchange Service (DAAD) to J.A.C., 1 and activates pro-hepatocyte growth factor. FEBS J. 277, V.M., and J.D.H., a National Health and Medical Research 4888–4900. Herter, S., Piper, D.E., Aaron, W., Gabriele, T., Cutler, G., Cao, P., Council of Australia Principal Research Fellowship to Bhatt, A.S., Choe, Y., Craik, C.S., Walker, N., et al. (2005). J.A.C., and an Australian Postgraduate Award, and Smart Hepatocyte growth factor is a preferred in vitro substrate State PhD stipend and Queensland University of Technol- for human hepsin, a membrane-anchored serine protease ogy top-up and write-up scholarships to J.C.R. implicated in prostate and ovarian cancers. Biochem. J. 390, 125–136. Hooper, J.D., Bui, L.T., Rae, F.K., Harvey, T.J., Myers, S.A., ­Ashworth, L.K., and Clements, J.A. (2001). Identification and References characterization of KLK14, a novel kallikrein serine protease located on human 19q13.4 and expressed in Benaud, C., Dickson, R.B., and Lin, C.Y. (2001). Regulation of the prostate and skeletal muscle. Genomics 73, 117–122. activity of matriptase on epithelial cell surfaces by a blood- Ihara, S., Miyoshi, E., Ko, J.H., Murata, K., Nakahara, S., Honke, K., derived factor. Eur. J. Biochem. 268, 1439–1447. Dickson, R.B., Lin, C.Y., and Taniguchi, N. (2002). Prometastatic Bergum, C. and List, K. (2010). Loss of the matriptase inhibi- effect of N-acetylglucosaminyltransferase V is due to modifica- tor HAI-2 during prostate cancer progression. Prostate 70, tion and stabilization of active matriptase by adding beta 1–6 1422–1428. GlcNAc branching. J. Biol. Chem. 277, 16960–16967. Borgono, C.A., Michael, I.P., Komatsu, N., Jayakumar, A., Kapa- Ihara, S., Miyoshi, E., Nakahara, S., Sakiyama, H., Ihara, H., dia, R., Clayman, G.L., Sotiropoulou, G., and Diamandis, E.P. ­Akinaga, A., Honke, K., Dickson, R.B., Lin, C.Y., and Taniguchi, (2007a). A potential role for multiple tissue kallikrein serine N. (2004). Addition of beta1–6 GlcNAc branching to the oligo- proteases in epidermal desquamation. J. Biol. Chem. 282, saccharide attached to Asn 772 in the serine protease domain 3640–3652. of matriptase plays a pivotal role in its stability and resistance Borgono, C.A., Michael, I.P., Shaw, J.L., Luo, L.Y., Ghosh, M.C., against trypsin. Glycobiology 14, 139–146. Soosaipillai, A., Grass, L., Katsaros, D., and Diamandis, E.P. Ito, Y., Akinaga, A., Yamanaka, K., Nakagawa, T., Kondo, A., (2007b). Expression and functional characterization of the Dickson, R.B., Lin, C.Y., Miyauchi, A., Taniguchi, N., and cancer-related serine protease, human tissue kallikrein 14. J. Miyoshi, E. (2006). Co-expression of matriptase and N-acetyl- Biol. Chem. 282, 2405–2422. glucosaminyltransferase V in thyroid cancer tissues – its possi- Brattsand, M., Stefansson, K., Lundh, C., Haasum, Y., and Egelrud, T. ble role in prolonged stability in vivo by aberrant glycosylation. (2005). A proteolytic cascade of in the stratum Glycobiology 16, 368–374. corneum. J. Invest. Dermatol. 124, 198–203. Jacobsen, F., Ashtiani, S.N., Tennstedt, P., Heinzer, H., Simon, R., Cecchi, F., Rabe, D.C., and Bottaro, D.P. (2012). Targeting the HGF/ Sauter, G., Sirma, H., Tsourlakis, M.C., Minner, S., Schlomm, Met signaling pathway in cancer therapy. Expert. Opin. Ther. T., et al. (2013). High c-MET expression is frequent but not asso- Targets 16, 553–572. ciated with early PSA recurrence in prostate cancer. Exp. Ther. de Veer, S.J., Swedberg, J.E., Parker, E.A., and Harris, J.M. (2012). Med. 5, 102–106. Non-combinatorial library screening reveals subsite cooperativ- Jameson, G.W., Roberts, D.V., Adams, R.W., Kyle, W.S., and Elmore, ity and identifies new high-efficiency substrates for kallikrein- D.T. (1973). Determination of the operational molarity of solu- related peptidase 14. Biol. Chem. 393, 331–341. tions of bovine alpha-chymotrypsin, trypsin, and fac- Dong, Y., Harrington, B.S., Adams, M.N., Wortmann, A., Stephenson, tor Xa by spectrofluorimetric titration. Biochem. J. 131, 107–117. S.A., Lisle, J., Herington, A., Hooper, J.D., and Clements, J.A. Kataoka, H., Itoh, H., Nuki, Y., Hamasuna, R., Naganuma, S., Kita- (2014). Activation of membrane-bound proteins and receptor mura, N., and Shimomura, T. (2002). Mouse hepatocyte growth systems: a link between tissue kallikrein and the KLK-related factor (HGF) activator inhibitor type 2 lacking the first Kunitz peptidases. Biol. Chem. 395, 977–990. domain potently inhibits the HGF activator. Biochem. Biophys. Fan, B., Wu, T.D., Li, W., and Kirchhofer, D. (2005). Identification of Res. Commun. 290, 1096–1100. hepatocyte growth factor activator inhibitor-1B as a poten- Kawaguchi, M. and Kataoka, H. (2014). Mechanisms of hepatocyte tial physiological inhibitor of prostasin. J. Biol. Chem. 280, growth factor activation in cancer tissues. Cancers (Basel) 6, 34513–34520. 1890–1904. Felber, L.M., Borgono, C.A., Cloutier, S.M., Kundig, C., Kishi, T., Kirchhofer, D., Peek, M., Li, W., Stamos, J., Eigenbrot, C., Kadk- Ribeiro Chagas, J., Jichlinski, P., Gygi, C.M., Leisinger, H.J., hodayan, S., Elliott, J.M., Corpuz, R.T., Lazarus, R.A., and Diamandis, E.P., et al. (2005). Enzymatic profiling of human Moran, P. (2003). Tissue expression, protease specificity, and kallikrein 14 using phage-display substrate technology. Biol. Kunitz domain functions of hepatocyte growth factor activa- Chem. 386, 291–298. tor inhibitor-1B (HAI-1B), a new splice variant of HAI-1. J. Biol. Gupta, A., Karakiewicz, P.I., Roehrborn, C.G., Lotan, Y., Zlotta, Chem. 278, 36341–36349. A.R., and Shariat, S.F. (2008). Predictive value of plasma Kirchhofer, D., Peek, M., Lipari, M.T., Billeci, K., Fan, B., and hepatocyte growth factor/scatter factor levels in patients Moran, P. (2005). Hepsin activates pro-hepatocyte growth with clinically localized prostate cancer. Clin. Cancer Res. 14, factor and is inhibited by hepatocyte growth factor activator 7385–7390. inhibitor-1B (HAI-1B) and HAI-2. FEBS Lett. 579, 1945–1950. Hashimoto, T., Kato, M., Shimomura, T., and Kitamura, N. (2010). Kiyomiya, K., Lee, M.S., Tseng, I.C., Zuo, H., Barndt, R.J., Johnson, TMPRSS13, a type II transmembrane serine protease, is M.D., Dickson, R.B., and Lin, C.Y. (2006). Matriptase activation inhibited by hepatocyte growth factor activator inhibitor type and shedding with HAI-1 is induced by steroid sex hormones in

Brought to you by | EP Ipswich Authenticated Download Date | 11/9/16 5:27 AM J.C. Reid et al.: KLK14 regulates pro-HGF, HAI-1A and HAI-1B 1305

human prostate cancer cells, but not in breast cancer cells. Am. Sanchez, W.Y., de Veer, S.J., Swedberg, J.E., Hong, E.J., Reid, J.C., J. Physiol. Cell Physiol. 291, C40–C49. Walsh, T.P., Hooper, J.D., Hammond, G.L., Clements, J.A., and Lee, S.L., Dickson, R.B., and Lin, C.Y. (2000). Activation of hepato- Harris, J.M. (2012). Selective cleavage of human sex hormone- cyte growth factor and /plasminogen activator by binding globulin by kallikrein-related peptidases and effects matriptase, an epithelial membrane serine protease. J. Biol. on androgen action in LNCaP prostate cancer cells. Endocrinol- Chem. 275, 36720–36725. ogy 153, 3179–3189. Lin, C.Y., Anders, J., Johnson, M., and Dickson, R.B. (1999). Purifica- Shia, S., Stamos, J., Kirchhofer, D., Fan, B., Wu, J., Corpuz, R.T., tion and characterization of a complex containing matriptase ­Santell, L., Lazarus, R.A., and Eigenbrot, C. (2005). Confor- and a Kunitz-type serine protease inhibitor from human milk. mational lability in serine protease active sites: structures of J. Biol. Chem. 274, 18237–18242. hepatocyte growth factor activator (HGFA) alone and with the Miyake, Y., Tsuzuki, S., Mochida, S., Fushiki, T., and Inouye, K. inhibitory domain from HGFA inhibitor-1B. J. Mol. Biol. 346, (2010). The role of asparagine-linked glycosylation site on 1335–1349. the catalytic domain of matriptase in its zymogen activation. Skropeta, D. (2009). The effect of individual N-glycans on enzyme Biochim. Biophys. Acta 1804, 156–165. activity. Bioorg. Med. Chem. 17, 2645–2653. Miyazawa, K. (2010). Hepatocyte growth factor activator (HGFA): a Stefansson, K., Brattsand, M., Ny, A., Glas, B., and Egelrud, T. serine protease that links tissue injury to activation of hepato- (2006). Kallikrein-related peptidase 14 may be a major con- cyte growth factor. FEBS J. 277, 2208–2214. tributor to trypsin-like proteolytic activity in human stratum Mukai, S., Fukushima, T., Naka, D., Tanaka, H., Osada, Y., and corneum. Biol. Chem. 387, 761–768. ­Kataoka, H. (2008). Activation of hepatocyte growth factor Strohmeyer, D., Strauss, F., Rossing, C., Roberts, C., Kaufmann, activator zymogen (pro-HGFA) by human kallikrein 1-related O., Bartsch, G., and Effert, P. (2004). Expression of bFGF, peptidases. FEBS J. 275, 1003–1017. VEGF and c-met and their correlation with microvessel density Nagakawa, O., Yamagishi, T., Akashi, T., Nagaike, K., and Fuse, H. and progression in prostate carcinoma. Anticancer Res. 24, (2006). Serum hepatocyte growth factor activator inhibitor 1797–1804. type I (HAI-I) and type 2 (HAI-2) in prostate cancer. Prostate 66, Tsai, C.H., Teng, C.H., Tu, Y.T., Cheng, T.S., Wu, S.R., Ko, C.J., 447–452. Shyu, H.Y., Lan, S.W., Huang, H.P., Tzeng, S.F., et al. (2014). Nakashiro, K., Hayashi, Y., and Oyasu, R. (2003). Immunohisto- HAI-2 suppresses the invasive growth and metastasis of pros- chemical expression of hepatocyte growth factor and c-Met/ tate cancer through regulation of matriptase. Oncogene 33, HGF receptor in benign and malignant human prostate tissue. 4643–4652. Oncol. Rep. 10, 1149–1153. Trusolino, L., Bertotti, A., and Comoglio, P.M. (2010). MET Oberst, M.D., Williams, C.A., Dickson, R.B., Johnson, M.D., and signalling: principles and functions in development, Lin, C.Y. (2003). The activation of matriptase requires its organ regeneration and cancer. Nat. Rev. Mol. Cell Biol. noncatalytic domains, serine protease domain, and its cognate 11, 834–848. inhibitor. J. Biol. Chem. 278, 26773–26779. Varkaris, A., Corn, P.G., Gaur, S., Dayyani, F., Logothetis, C.J., and Oikonomopoulou, K., Hansen, K.K., Saifeddine, M., Tea, I., Gallick, G.E. (2011). The role of HGF/c-Met signaling in prostate Blaber, M., Blaber, S.I., Scarisbrick, I., Andrade-Gordon, P., cancer progression and c-Met inhibitors in clinical trials. Cottrell, G.S., Bunnett, N.W., et al. (2006). Proteinase-activated Expert. Opin. Investig. Drugs 20, 1677–1684. receptors, targets for kallikrein signaling. J. Biol. Chem. 281, Warren, M., Twohig, M., Pier, T., Eickhoff, J., Lin, C.Y., Jarrard, D., 32095–32112. and Huang, W. (2009). Protein expression of matriptase and Owen, K.A., Qiu, D., Alves, J., Schumacher, A.M., Kilpatrick, L.M., its cognate inhibitor HAI-1 in human prostate cancer: a tissue Li, J., Harris, J.L., and Ellis, V. (2010). Pericellular activation microarray and automated quantitative analysis. Appl. Immu- of hepatocyte growth factor by the transmembrane serine nohistochem. Mol. Morphol. 17, 23–30. proteases matriptase and hepsin, but not by the membrane- Yasuda, K., Nagakawa, O., Akashi, T., Fujiuchi, Y., Koizumi, K., associated protease uPA. Biochem. J. 426, 219–228. Komiya, A., Saiki, I., and Fuse, H. (2009). Serum active hepato- Prassas, I., Eissa, A., Poda, G., and Diamandis, E.P. (2015). Unleash- cyte growth factor (AHGF) in benign prostatic disease and ing the therapeutic potential of human kallikrein-related serine prostate cancer. Prostate 69, 346–351. proteases. Nat. Rev. Drug. Discov. 14, 183–202. Yasuda, K., Komiya, A., Watanabe, A., Morii, A., Oya, T., Nagakawa, Rabien, A., Fritzsche, F., Jung, M., Diamandis, E.P., Loening, S.A., O., Fujiuchi, Y., and Fuse, H. (2013). Expression of Hepatocyte Dietel, M., Jung, K., Stephan, C., and Kristiansen, G. (2008). growth factor activator inhibitor type-1 (HAI-1) in prostate can- High expression of KLK14 in prostatic adenocarcinoma is asso- cer. Anticancer Res. 33, 575–581. ciated with elevated risk of prostate-specific antigen relapse. Yousef, G.M., Magklara, A., Chang, A., Jung, K., Katsaros, D., and Tumour. Biol. 29, 1–8. Diamandis, E.P. (2001). Cloning of a new member of the human Rajapakse, S. and Takahashi, T. (2007). Expression and enzymatic kallikrein gene family, KLK14, which is down-regulated in dif- characterization of recombinant human kallikrein 14. Zoolog. ferent malignancies. Cancer Res. 61, 3425–3431. Sci. 24, 774–780. Ramsay, A.J., Dong, Y., Hunt, M.L., Linn, M., Samaratunga, H., ­Clements, J.A., and Hooper, J.D. (2008). Kallikrein-related peptidase 4 (KLK4) initiates intracellular signaling via Article note: This article was submitted for the Highlight Issue in protease-activated receptors (PARs). KLK4 and PAR-2 are co- connection with the 9th General Meeting 2015 of the International expressed during prostate cancer progression. J. Biol. Chem. Proteolysis Society, published in issue 9 (2016) of Biological 283, 12293–12304. Chemistry.

Brought to you by | EP Ipswich Authenticated Download Date | 11/9/16 5:27 AM Copyright of Biological Chemistry is the property of De Gruyter and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.