Pro-angiogenic effect of human kallikrein-related ...

DOI 10.1515/hsz-2012-0291 Biol. Chem. 2013; 394(3): 385–391

Short Communication Thomas Kryza, Gilles Lalmanach, Marion Lavergne, Fabien Lecaille, Pascale Reverdiau, Yves Courtya and Nathalie Heuzé-Vourc’ha,*

Pro-angiogenic effect of human kallikrein-related peptidase 12 (KLK12) in lung endothelial cells does not depend on kinin-mediated activation of B2 receptor Abstract: Kallikrein-12 (KLK12) may play an important role in angiogenesis modulating proangiogenic factor bioavailability and activating the kinin receptor B2 pathway. We studied whether KLK12 had an impact on angiogenesis and the activation of kinin receptor B2 results from the KLK12-dependent generation of kinins. KLK12 efficiently hydrolyzed high molecular weight kininogen, liberating a fragment containing the carboxy-terminal end of kinins. The kininogenase activity of KLK12 was poor, however, due to the cleavage resistance of the N-terminal side of the kinin sequence. A very low amount of kinins was accordingly released after in vitro incubation of high molecular weight kininogen with KLK12 and thus the proangiogenic activity of KLK12 in lung endothelial cells was not related to a kinin release. Keywords: angiogenesis; kallikrein-related peptidase 12; kininogenase; serine protease.

a

These authors contributed equally to the study design and are co-senior authors of this article. *Corresponding author: Nathalie Heuzé-Vourc’h, Université François Rabelais, EA 6305, F-37032 Tours, France, e-mail: [email protected] Thomas Kryza, Marion Lavergne, Pascale Reverdiau and Nathalie Heuzé-Vourc’h: Université François Rabelais, EA 6305, F-37032 Tours, France; and Centre d’Etude des Pathologies Respiratoires, UMR 1100/EA6305, F-37032 Tours, France Gilles Lalmanach, Fabien Lecaille and Yves Courty: Centre d’Etude des Pathologies Respiratoires, UMR 1100/EA6305, F-37032 Tours, France; INSERM, UMR 1100, F-37032 Tours, France; and Université François Rabelais, UMR 1100, F-37032 Tours, France

Human tissue kallikrein-related peptidases (KLK1– KLK15) belong to a subgroup of secreted serine proteases that are expressed in a large number of tissues

and cell populations (Shaw and Diamandis, 2007). As demonstrated in a large number of studies, this family has important roles in pathophysiological processes (Lundwall and Brattsand, 2008). For example, KLK1 is implicated in the control of vascular tone via its kininogenase (i.e., kinin-generating) activity (Meneton et al., 2001; Chao et al., 2010). Other KLKs are implicated in skin desquamation (Eissa et al., 2010), innate immunity (Sotiropoulou et al., 2009; Sotiropoulou and Pampalakis, 2010), semen liquefaction (Veveris-Lowe et al., 2007), tooth enamel formation (Lu et al., 2008), neural development and plasticity (Scarisbrick et al., 2006, 2008) and angiogenesis (Diamandis et al., 2000; Giusti et al., 2005; Guillon-Munos et al., 2011). KLK12 was originally cloned using a positional candidate gene approach (Yousef et al., 2000) and was found to be highly expressed in a variety of human tissues, such as bone and bone marrow, colon, lung and trachea, prostate, salivary glands, and stomach at the RNA and/or protein level (Yousef et al., 2000; Shaw and Diamandis, 2007). KLK12 is a serine protease with trypsin-like activity, targeting peptide bonds featuring either arginine or lysine at their P1 position. Its activity can be tightly regulated by autodegradation, by interaction with zinc ions, and by forming a covalent complex with α2-antiplasmin (Memari et al., 2007). At the present time, KLK12’s physiological function remains unclear, but circumstantial evidence indicates that it may play a critical role in angiogenesis. Our group recently showed that KLK12 cleaves several members of the CCN family, which are matricellular proteins with multiple domains interacting with various growth factors, thereby modulating the bioavailability and/or activity of growth factors, such as vascular endothelial growth factor (VEGF; Guillon-Munos et al., 2011). In addition, deregulation of KLK12 was found in systemic sclerosis, a disorder characterized by defective

Brought to you by | Universite de Tours Authenticated | 193.52.208.233 Download Date | 3/5/13 9:18 AM

386 T. Kryza et al.: Kininogen proteolysis by kallikrein-related peptidase 12

A

Non reducing condition

Reducing condition

Reducing condition + Inhibitor

230 kDa 150 kDa 100 kDa 80 kDa 60 kDa 50 kDa 40 kDa 30 kDa

B

KLK1 (50 – 0.5 nM ) : HMWK (500 nM)

KLK12 (50 – 0.5 nM) : HMWK (500 nM)

2 K W M) K 1 M) n KL 0 n HM 00 (5 (5

K K1 ) W M) KL nM HM 0 n 0 0 (5 (5

230 kDa 150 kDa 100 kDa 80 kDa


60 kDa

60 kDa

50 kDa

50 kDa

40 kDa

40 kDa

30 kDa

30 kDa

C

KLK1 (10 nM) : HMWK (500 nM)

KLK12 (10 nM) : HMWK (500 nM)

K W M) n HM 00 (5

K W M) n HM 00 (5



60 kDa

60 kDa

50 kDa

50 kDa

40 kDa

40 kDa

30 kDa

30 kDa

Figure 1 Proteolysis profiles of purified high molecular weight kininogen (HMWK) after KLK1 and KLK12 degradation. (A) HMWK proteolysis patterns: purified human HMWK (500 nm; Merck Millipore, Nottingham, UK) was incubated for 8 h at 37°C with 10 nm titrated human KLK1 (RayBiotech, Norcross, GA, USA) or titrated human KLK12 (R&D Systems, Minneapolis, USA) in their respective activity buffer (50 mm Tris/HCl buffer, pH 8.3, 1 mm ethylenediaminetetraacetic acid for KLK1 and 100 mm Tris, 150 mm NaCl, 10 mm CaCl2, 0.05% (w/v) Brij35, pH 7.5 for KLK12) in the presence or absence of inhibitors (aprotinin for KLK1 or a specific KLK12 inhibitor; unpublished results). Degradation products of HMWK were analyzed in non-reducing and reducing conditions on Coomassie Blue stained 4–20% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS PAGE; Invitrogen Corporation, Carlsbad, CA, USA). (B) Dose-dependent fragmentation of HMWK. HMWK (500 nm) was incubated for 4 h at 37°C with variable concentrations of KLK1 or KLK12 (50–0.5 nm) to obtain an enzyme:substrate molar ratio of 1:10 to 1:1000. Proteolysis products were loaded on SDS-PAGE under reducing conditions. (C) Time-dependent hydrolysis of HMWK. An equal amount of HMWK (500 nm) was incubated with 10 nm KLK1 or KLK12 for different periods of time (5 min to 24 h; enzyme:substrate molar ratio of 1:50). Proteolysis products were loaded on SDS-PAGE under reducing conditions.


T. Kryza et al.: Kininogen proteolysis by kallikrein-related peptidase 12 387

angiogenesis, and inhibition of KLK12 in human skin endothelial cells with antibodies resulted in a reduction in endothelial cell growth, migration and tubule-like formation (Giusti et al., 2005). Moreover, Giusti et al. (2005) suggested that the effect of KLK12 might be mainly associated with kinin receptor B2 (B2R) activation and be mediated by kinins. However, the evidence for this was indirect and we still do not know whether KLK12 can act as a kininogenase.

A

Cleavage sites analysis (Amino terminal sequencing)

S D1

D2

D3

Protein bands extraction

S D4

D5H

Proteins separation (1D SDS-PAGE)

Proteins elimination (Ethanol 100%) Peptides purification (HILIC Column) Peptides analysis (Mass spectrometry)

C

B 230 kDa

100 kDa

I245CVGC D3

60 kDa

D4

SPFR371

50 kDa

D5H

GHTR409

D6

S372SRI D5H

40 kDa

D6

R410HDW D5H

D6

30 kDa

Peptides formed by KLK1

Peptides formed by KLK12 Bradykinin

Lys-Bradykinin

%

PPTK244

80 kDa

Incubated 24 h at 37°C

Titrated protease (KLK1, KLK12 or buffer)

In order to gain insight into the molecular mechanisms associated with KLK12 proangiogenic properties, we studied the capacity of KLK12 to degrade kininogen and release kinins. Both high molecular weight kininogen (HMWK, apparent molecular mass 90–120 KDa) and low molecular weight kininogen (LMWK, 50–68 KDa) serve as

150 kDa

D6

Purified human HMWK

KLK12’s kininogen degradation capacity

%

m/z

m/z

Figure 2 Identification of high molecular weight kininogen (HMWK) cleavage sites of KLK12 by amino terminal sequencing and mass spectrometry. (A) Procedure scheme. HMWK (500 nm) was incubated with an equal amount of KLK1 or KLK12 (10 nm) for 24 h at 37°C (enzyme:substrate molar ratio of 1:50) in their respective activity buffer. KLK12-mediated HMWK fragments were analyzed using a Coomassie Blue stained 4–20% sodium dodecyl sulfate polyacrylamide gel electrophoresis. The selected bands were extracted and analyzed by N- terminal Edman sequencing (Plateforme de protéomique PISSARO-IRIB, Rouen, France); the three cleavage sites identified are presented in (B). To identify the kinin-type peptides, HMWK peptides resulting from KLK1 or KLK12 cleavage were analyzed using mass spectrometry. Proteins were eliminated by ethanol precipitation and peptides were purified using hydrophilic interaction liquid chromatography column (Protea-bio, Chicago, USA) according to the manufacturer’s instructions. Then purified peptides were analyzed by mass spectrometry using a M@ LDI-TOF LR (Waters, Manchester, UK) (Laboratoire de Spectrométrie de Masse, Nouzilly, France). (C) Mass finger prints obtained show that KLK1 mainly formed Lys-bradykinin (Mwobs: 1188.86 Da; Mwtheo: 1188.65 Da) and KLK12 formed bradykinin (Mwobs: 1060.80 Da; Mwtheo: 1060.56 Da) after HMWK proteolysis.


388 T. Kryza et al.: Kininogen proteolysis by kallikrein-related peptidase 12

precursor for kinins in humans (Lalmanach et al., 2010). First, we analyzed the proteolytic activity of KLK12 on purified HMWK and compared it to KLK1, which is the only tissue KLK-related peptidase family member with significant kininogenase activity described so far (Bhoola et al., 1992). KLK12 cleaves HMWK but the resulting proteolytic profile is different from that of KLK1 (Figure 1A) and trypsin (not shown), leading to the release of several fragments under both reducing and non-reducing conditions. Although kininogen seems to be a relevant substrate of KLK12, as shown by both dose- and time-dependent fragmentation of HMWK (Figure 1B and C), the HMWK cleavage by KLK12 was incomplete and less efficient than KLK1. HMWK and LMWK display an identical N-terminus heavy chain containing three domains (D1, D2, and D3), and a short D4 domain containing kinin peptides. They differ at their C-terminus because of alternative splicing (HMWK: domains D5 and D6; LMWK: domain D5) (Kitamura et al., 1985). KLK1 cleaves both LMWK and HMMK at two sites in the D4 domain to release the decapeptide kinin Lys-bradykinin or kallidin (KRPPGFSPFR), whereas plasma KLK that does not belong to the KLK-related peptidase family, generates the nonapeptide kinin bradykinin (RPPGFSPFR), but only from HMWK. To characterize the fragments and peptides derived from kininogen proteolysis

A

S D1

358I

D2

by KLK12, we used N-terminal sequencing and mass spectrometry (Figure 2). As shown in Figure 2B, we identified three cleavage sites: one at the beginning of the D3 domain (PPTK244-I245CVGC), the second at the end of the D4 domain (SPFR371-S372SRI) liberating the carboxy-terminal part of the kinins, and finally one in the D5 domain (GHTR409R410HDW), also called kininostatin, which has cell-binding sites, antiangiogenic properties, and sequences for heparin binding as previously described (Colman, 2006; Veillard et al., 2008). This result shows that KLK12 indifferently accommodates the Lys or Arg residue in the P1 position, as previously shown by Guillon-Munos and colleagues (2011). As shown by mass spectrometry, HMWK incubated with KLK12 released bradykinin, but not Lys-bradykinin like KLK1 (Figure 2C). To determine whether bradykinin release can be physiologically relevant, we analyzed the hydrolytic activity of KLK12 using fluorescence resonance energy transfer peptides derived from the amino- and carboxy-terminal ends of kinins (Desmazes et al., 2003). Hydrolysis of the human kininogen fragment peptides (Figure 3) showed that KLK12 efficiently cleaved Abz-SPFRSSR-3-NO2-Tyr more slowly than KLK1, but its catalytic efficiency towards AbzISLMKYRPPGF-3-NO2-Tyr was extremely low. This result suggests that the kininogenase activity of KLK12 is very poor compared with that of KLK1, although KLK12 was able

S

D3

D4

D5H

D6

– S – L – M – K – R – P – P – G – F – S – P – F – R – S – S – R – I375

Abz – 358I – S – L – M – K – R – P – P – G – F – Y(NO2) Nter BK peptide Abz - S – P – F – R – S – S – R – I375 – Y(NO2) Cter BK peptide

B

Nter BK peptide

Cter BK peptide

Kcat/Km

Km

Kcat/Km

Km

(mM -1.Sec -1)

(μM)

(mM-1.Sec -1)

(μM)

KLK1

156.0±3.0

60.0±2.9

828.3±19.6

6.1±1.4

KLK12

1.3±0.2

N.D.

34.2±0.8

21.4±1.2

Figure 3 Hydrolysis of kininogen-derived peptides by KLK1 and KLK12. (A) Structure of kininogen-derived fluorogenic substrates. The fluorogenic peptides cover sequences surrounding the amino- and carboxyterminal ends of the kinin insertion (represented in bold) in human HMWK. Nter-BK peptide and Cter-BK peptide were flanked by a donoracceptor pair: a fluorescent N-terminal Abz group and a C-terminal 3-NO2-Tyr quencher. Peptides were synthesized as peptidyl-amides (Desmazes et al., 2003). (B) Hydrolysis of kininogen-derived fluorogenic substrates by KLK1 and KLK12. Second-order rate constants were measured under pseudo first-order conditions (500 nm BK peptides were incubated with 8 nm KLK1 or KLK12) using Enzfitter software (Biosoft, Cambridge, UK), reported as mean ± standard deviation (triplicate experiments). Michaelis constant values were determined from Lineweaver-Burk linear plot (1–50 μm BK peptides were incubated with 4 nm KLK1 or 16 nm KLK12) and reported as mean ± standard deviation (triplicate experiments).


T. Kryza et al.: Kininogen proteolysis by kallikrein-related peptidase 12 389

to digest human kininogen. In agreement with this, the measurement of kinin concentration following incubation of HMWK with KLK12 by enzyme-linked immunosorbent assay resulted in a very low amount of kinin. Indeed, incubation of 500 nm HMWK with 10 pM titrated human KLK1 (enzyme: substrate molar ratio of 1:50,000) for 2 h at 37°C led to the release of 250 pg/ml of kinin whereas incubation with 5 nm titrated KLK12 (enzyme:substrate molar ratio of 1:100) led to 14 pg/ml of kinin (slightly higher than the background level). The low kininogenase activity of KLK12 is likely due to the resistance of hydrolysis of the K-R bond on the N-terminal side of the bradykinin sequence in the kininogen, as previously observed for other KLK-related peptidases (Andrade et al., 2011).

KLK12 has proangiogenic properties Using endothelial cells’ tubule-like formation and proliferation assays, Giusti et al. (2005) demonstrated the

A

B

Fold increase in tube lenght/control

2.5

* *

proangiogenic properties of KLK12 in skin endothelial cells. Moreover, they reported that the addition of either HOE 140, a B2R antagonist (Leeb-Lundberg et al., 2005), or KLK12-blocking antibodies resulted in a reduction of angiogenesis in vitro. Although, the evidence was indirect, the authors suggested that the KLK12-mediated proangiogenic effect might be associated with B2R activation and may result from kinin release by the protease. Herein, we showed that KLK12 is not a kininogenase (Figure 3), but it is possible that another protease secreted by host cells or pathogens cooperates with KLK12 to generate kinins in vivo, as previously described for other proteases (Sato and Nagasawa, 1989; Kozik et al., 1998; Imamura et al., 2002; Imamura et al., 2005). To analyze this hypothesis, we evaluated the role of the B2R pathway in KLK12-mediated angiogenesis in human lung endothelial cells, which represent the most physiologically-relevant model for the study of KLK12 due to its tissue expression (Shaw and Diamandis, 2007) and also the endogenous expression of B2R (data not shown). The addition of exogenous KLK12 significantly increased tube-like formation on Matrigel™ (BD

Control

Control

KLK12

KLK12

VEGF

VEGF

* *

2.0

1.5

1.0

0.5

Buffer

Buffer

10 nM HOE140

10 nM HOE140

* p