Recombinant Snake Venom Metalloproteinase Inhibitor BJ46A

Preclinical report 461

Recombinant snake venom metalloproteinase inhibitor BJ46A inhibits invasion and metastasis of B16F10 and MHCC97H cells through reductions of matrix metalloproteinases 2 and 9 activities Ming-Kai Jia,b, Yi Shia,c, Jian-Wen Xua, Xu Lina and Jian-Yin Lina

Studies have shown that the recombinant BJ46a (rBJ46a) activities in tumor cells were suppressed by rBJ46a in a protein can reduce matrix metalloproteinase (MMP) dose-dependent manner, and the Km values of rBJ46a activities and inhibit invasion and metastasis of melanoma against MMP2 and MMP9 activities that were expressed in cells. Here, we optimized the Pichia pastoris system to both B16F10 and MHCC97H cells were 3.6 and 1.4 lmol/l, evaluate rBJ46a protein as an anticancer agent. The respectively. Thus, rBJ46a can inhibit the invasion and Enzchek gelatinase/collagenase assay showed that metastasis of tumor cells by reducing MMP2/MMP9 rBJ46a inhibited MMP activities (IC50 = 0.119 mg/ml). activities, indicating that rBJ46a may be a novel therapeutic Kinetic analyses using a series of double reciprocal agent for antimetastasis of tumor cells. Anti-Cancer Drugs Lineweaver–Burk plots (1/V vs. 1/S) showed a competitive 24:461–472 c 2013 Wolters Kluwer Health | Lippincott mode of inhibition with rBJ46a with inhibitory efficiency Williams & Wilkins. against MMPs (K = 13.6 nmol/l). Matrigel invasion assays i Anti-Cancer Drugs 2013, 24:461–472 showed significant activity of rBJ46a on tumor cells. For lung colonization assays, C57BL/6 mice were inoculated in Keywords: antimetastasis, matrix metalloproteinase, recombinant snake venom metalloproteinase inhibitor BJ46A, tumor the lateral tail vein with B16F10 cells and were treated with three i.v. injections of rBJ46a (1, 2, and 4 mg/kg) 24 h aKey Laboratory of Ministry of Education for Gastrointestinal Cancer, Research Center of Molecular Medicine, bDepartment of Dermatology, First Affiliated before cell inoculation, and 2 and 24 h after cell inoculation. Hospital, Fujian Medical University and cDepartment of Molecular Pathology, Administration of rBJ46a suppressed lung tumor colony Fujian Provincial Tumor Hospital, The Teaching Hospital of Fujian Medical University, Fuzhou, China formation significantly. For spontaneous metastasis assays, MHCC97H cells were inoculated subcutaneously Correspondence to Jian-Yin Lin, MD, Key Laboratory of Ministry of Education for Gastrointestinal Cancer, Research Center of Molecular Medicine, Fujian into nude mice. After 24 h, rBJ46a was administered by i.p. Medical University, Fuzhou 350004, China injections: 1, 2, and 4 mg/kg once daily for 6 days. rBJ46a Tel/fax: + 86 0591 83569132; e-mail: [email protected] decreased lung tumor colony formation significantly. Received 22 July 2012 Revised 5 January 2013 Gelatin zymography showed that MMP2/MMP9 enzymatic Revised form accepted 18 January 2013

Introduction clinical trials with these MMPIs have been conducted in Tumor invasion and metastasis are usually major causes of cancer patients and have proved unsuccessful in most cancer death. Although many advances have been made cases, research into MMPIs is ongoing [2,3]. in cancer therapy, most cancer deaths still result from metastatic disease [1]. The process of cancer cell Snake venoms contain several proteolytic enzymes and metastasis can be summarized as a sequence of events proteins with different toxicological functions and special that include the formation of tumor embolus in the blood pharmacological effects [4,5]. BJ46a, a 46 kDa protein capillary, cell penetration of the extracellular matrix isolated from the serum of Bothrops jararaca,isan (ECM), trapping at a secondary site, and growth at the inhibitor of snake venom metalloproteinases (SVMPs). secondary site [1]. Degradation of the ECM mediated A member of the fetuin family and cystatin superfamily, by major proteases, including serine proteases, cysteine BJ46a interacts with metalloproteinases through the proteases, and matrix metalloproteinases (MMPs), is one formation of a noncovalent complex, specifically inhibit- of the key steps during the invasion and metastasis ing atrolysin C (class P-I) and jararhagin (class P-III) of cancer cells, and the MMP family plays an important proteolytic activities [6]. Because SVMPs share homology role in the process. MMPs are therefore amenable to with the zinc-binding environment of MMPs, it is therapeutic intervention by synthetic and natural in- possible that BJ46a may be able to modulate anticancer hibitors. Multiple therapeutic agents, called matrix invasion and metastasis through the formation of a metalloproteinase inhibitors (MMPIs), have been devel- complex with MMP. Previous studies in our laboratory oped to target MMPs in an attempt to control their have shown that the BJ46a gene can be synthesized by enzymatic activity and to inhibit the invasion and assembly PCR, and pretreatment of the recombinant metastasis of cancer cells. However, even though many BJ46a (rBJ46a) protein produced using a Bac-to-Bac

0959-4973 c 2013 Wolters Kluwer Health | Lippincott Williams & Wilkins DOI: 10.1097/CAD.0b013e32835f258d

Baculovirus Expression System could reduce MMP BJ46a was amplified to an OD600 of 4.0 in log-phase activities and inhibit invasion and metastasis of melanoma growth in 25 ml of BMGY medium [1% (w/v) yeast cells (B16) in vitro and in vivo [7–9]. However, the Bac-to- extract, 2% (w/v) peptone, 1.34% (w/v) yeast nitrogen Bac Baculovirus Expression System has the disadvantages base, 0.1 mol/l potassium phosphate buffer pH 6.0, of low yields, a requirement for costly equipment, and it 4 Â 10 –5% (w/v) biotin, 1% (v/v) glycerol, 4 Â 10 –3% is tedious, which limits the large-scale production of (w/v) histidine] at 301C using a shaking incubator recombinant protein. To advance research on the (300 rpm). Cells from the medium were collected by biological features of BJ46a and broaden the possibilities centrifugation (1500g) for 8 min and the pellet was of its applications in the pharmaceutical industry, we resuspended to an OD600 of 1.0 in BMMY [similar to described the high-level expression, purification, char- BMGY medium but containing 0.5% (v/v) methanol acterization, and structural prediction of rBJ46a in Pichia instead of 1% glycerol] medium for induction and pastoris [10]. To evaluate further the pharmaceutical expression in a 1 l baffled flask covered with six layers potential of rBJ46a protein as a cancer therapeutic agent, of sterile gauze. The production of protein was carried in this study, we first optimized the P. pastoris system for out in a flask at a final concentration of 1% methanol, recombinant protein synthesis, and then investigated the with replacement every 24 h. The expression culture effects of rBJ46a on the invasion and metastasis of supernatants were collected at 24, 48, 72, and 96 h to B16F10 cells and MHCC97H cells in vitro by Matrigel determine the optimal time for induction. The expres- invasion assays. We also examined the actions of rBJ46a sion levels were analyzed using 11% Tris-tricine poly- in vivo by two experimental animal models: human acrylamide gels. The collected supernatants were MHCC97H tumors grafted into nude mice and the concentrated using an Amicon ultrafiltration device (GE dissemination of B16F10 melanoma cells to the lungs of Healthcare, Piscataway, New Jersey, USA). syngeneic mice. The results showed that rBJ46a pro- Purification of polyhistidine-containing recombinant produced by the P. pastoris expression system inhibited the teins was completed using immobilized metal affinity invasion and metastasis of B16F10 cells and MHCC97H chromatography (IMAC) using Ni Sepharose 6 Fast Flow cells in vitro and in vivo, possibly through the reduction of (Code Number 17-0921-07; GE Healthcare, Piscataway, MMP2/MMP9 activities. New Jersey, USA) according to the manufacturer’s Materials and methods protocol. Briefly, a concentrated expression solution after dialysis was loaded on an IMAC column equilibrated with Cell culture and animals binding buffer [20 mmol/l sodium phosphate (pH 8.0), MHCC97H cells were purchased from the Liver Cancer 500 mmol/l NaCl, and 20 mmol/l imidazole]. The speci- Institute at Fudan University (Shanghai, China) and fically bound proteins were eluted with elution buffer cultured in Dulbecco’s modified Eagle’s medium (Gibco [20 mmol/l sodium phosphate (pH 8.0), 500 mmol/l BRL, Gaithersburg, Maryland, USA) supplemented with NaCl, and 500 mmol/l imidazole]. The eluted fractions 10% fetal bovine serum (Gibco BRL). MHCC97H cells that showed inhibitory activity and high purity (98.06%) were cultivated by means of alternating cell culture were pooled, and the resulting mixture was subjected to in vitro and growth in nude mice, as described by Tian ultrafiltration and desalting. et al. [11]. MHCC97H cells are a human hepatocellular carcinoma cell line with a high metastatic potential, which were established from a subcutaneous xenograft of Deglycosylation a spontaneous lung metastasis model of human hepato- Purified recombinant proteins were enzymatically degly- cellular carcinoma. B16F10 cells were purchased from the cosylated using endoglycosidase H (Endo H) using the China Center for Type Culture Collection (Wuhan protocol recommended by the manufacturer [15]. Briefly, University, Hubei, China) and were grown in RPMI 20 mg of recombinant protein was denatured with 1 mlof 1640 medium (Gibco BRL) supplemented with 10% fetal 10 Â glycoprotein denaturing buffer, heated at 951C for bovine serum. B16F10 cells are derived from a sponta- 10 min, deglycosylated by adding 2 ml10Â G5 buffer, 1 ml neous melanoma in C57BL/6 mice and undergo pulmonary of 500 000 U/ml of Endo H, and 6 mlH2O, and incubated metastasis after injection through the tail vein [12,13]. at 371C for 1 h. The reaction was stopped by boiling and This is a classic model of experimental lung metastasis. analyzed by SDS polyacrylamide gel electrophoresis Pathogen-free 5–6-week-old C57BL/6 mice and BALB/c (PAGE) and western blotting. nude mice were obtained from the Chinese Academy of Sciences (Shanghai, China); both the C57BL/6 and the Sodium dodecyl sulfate polyacrylamide gel BALB/c nude mice included equal numbers of males and electrophoresis and immunoblot analysis females. The purified proteins were electrophoresed in 11% Tris- tricine polyacrylamide gels to determine purities and Production and purification of recombinant BJ46a molecular weights. The separated proteins were used for rBJ46a protein was produced according to the manufac- immunoblot analysis by a Western Breeze chemilumines- turer’s instructions [10,14]. Briefly, recombinant GS115- cent immunodetection kit (catalogue number: WB7106;

Invitrogen, Shanghai, China) with a rabbit anti-BJ46a polyester membranes. Each bottom chamber was coated polyclonal antibody as the primary antibody (BioAsia, with 10 ml of fibronectin (0.5 mg/ml). A cell suspension Shanghai, China). The purified proteins were sent to the (500 ml; 1 Â 105 cells) in serum-free medium after in- Biology Medical Institute of Fudan University to deter- cubation with or without rBJ46a (25, 50, 100, 200, or mine the molecular weight of rBJ46a by MALDI-TOF/ 400 mg/ml) was seeded on top of each membrane. After TOF mass spectrometry (MS) analysis. 24 h, the migrated cells were counted. For Matrigel invasion assays, the bottom chamber was precoated with Assay for the inhibitory activity of recombinant BJ46a 20 mg of Matrigel (BD, San Jose, California, USA). All The ability of the recombinant protein to inhibit the other conditions were as described for the migration activities of MMPs was determined using an Enzchek assays. Antitumor effects were evaluated using the Gelatinase/Collagenase Assay Kit (Invitrogen), which following equation: ratio of inhibition (%) = (C – T)/ provides speed, high sensitivity, and convenience re- C Â 100, where T and C are the relative mean numbers quired for measuring gelatinase or collagenase activity for of invading cells in the treatment and the control groups, screening inhibitors in a high-throughput format [16,17]. respectively. The assay works on the principle that as MMPs (gelatinase) degrade the substrate DQ gelatin, the Experimental lung colonization assay C57BL/6 mice (12 mice/group) were inoculated in the digestion products can be detected by measuring their 5 absorption maxima at B495 nm and their fluorescence lateral tail vein with B16F10 cells (2 Â 10 cells/animal). emission maxima at B515 nm. Treatment with rBJ46a (1, 2, or 4 mg/kg) was carried out 24 h before cell inoculation, and 2 and 24 h after cell First, 80 mlof1Â reaction buffer was mixed with 20 mlof inoculation. Control mice received 200 ml of 0.9% (w/v) DQ gelatin solution in the wells. Then, 100 ml of the dilute NaCl. The mice were killed on day 21. The lungs were enzyme or 100 mlof1Â reaction buffer as a negative control removed following tumor colonization, weighed, and was added to start the reactions. All the samples were processed for histology examination [18,19]. The tumor incubated at room temperature and protected from light for colonization was enumerated. Antitumor effects were 2 h. The digestion products were detected by measuring evaluated using the following equation: ratio of growth their absorption maxima at B495 nm and their fluorescence inhibition (%) = (C – T)/C Â 100, where T and C are the emission maxima at B515 nm using an RF-5310PC spectro- relative mean tumor weights in the treatment and the fluorophotometer (Ying Si Te, Hangzhou, China). control groups, respectively; ratio of colonization inhibi- rBJ46a was diluted with 1 Â reaction buffer to a final tion (%) = (C – T)/C Â 100, where T and C are the relative volume of 80 ml at a series of concentrations and preloaded mean numbers of lesions in the treatment and the control into each assay well. Next, 20 ml of DQ gelatin substrate groups, respectively. solution was added, followed by 100 ml of the dilute enzyme Spontaneous metastasis assay in nude mice or 100 mlof1Â reaction buffer as a blank. All the samples 7 were incubated at room temperature and protected from MHCC97H cells (1 Â 10 /animal) in 0.2 ml of 0.9% NaCl light for 2 h. The digestion products were detected were inoculated (s.c.) into the right flank of BALB/c nude by measuring their absorption maxima at B495 nm and mice (8 mice/group). After 24 h, rBJ46a was administered their fluorescence emission maxima at B515 nm using an by i.p. injections at 1, 2, or 4 mg/kg once daily for 6 days. RF-5310PC spectrofluorophotometer (Ying Si Te). The mice were killed on day 42, and tumor weights were measured. The tumor volume was calculated using the following equation: V = 4/3p Â L/2 Â (W/2)2, where V is Cell growth assays 3 3 the volume (mm ), L is the largest diameter (mm), and Cells (4 Â 10 cells/well) were separately seeded onto 96- W is the smallest diameter (mm) [19,20]. The tumor and well microtiter plates. After 24 h of incubation, increasing lung tissues were removed for histological examination. doses of rBJ46a (25, 50, 100, 200, or 400 mg/ml) were The number of metastases were counted. The ratios of added to the cells. Cells were cultured for an additional growth and metastasis inhibition were calculated as 72 h and then 0.5 mg/ml MTT was added. After a 4-h described above. incubation, medium was aspirated and dimethyl sulfoxide was added to dissolve the MTT-formazan crystals Gelatin zymographic assay for matrix completely. The light absorbance at 570 nm wavelength metalloproteinases (A570), which correlates to the number of cells, was MMP2 and MMP9 enzymatic activities were assayed by measured on a microplate reader (Model Elx800; Bio-Tek, gelatin zymography [21]. Tumor cells were plated on six- Shanghai, China). well dishes and grown to 90% confluence in 2 ml of growth medium. The cells were maintained in serum-free media Migration and Matrigel invasion assays and treated with various concentrations of rBJ46a (the The invasive behavior of the cells in vitro was determined final concentrations of rBJ46a were 25, 50, 100, 200, or using 24-well transwell inserts fitted with 8 mm pore size 400 mg/ml) for 24 h. Conditioned medium was collected

Fig. 1

(a) (b) kDa 123456kDa 123456789 170 130 141 95 72 66 55 48 43 38 kDa 35 34 26 27 24 17 19

10 13

OD 1 0.5 0 01020 30 40 50 60 Tube number

Production, purification, and identification of recombinant BJ46a (rBJ46a). (a) SDS polyacrylamide gel electrophoresis (PAGE) analysis of the culture supernatant of GS115-BJ46a. 1, protein standard marker; 2, GS115-pPICZaA culture supernatant (48 h); 3–6, 24, 48, 72, and 96 h culture supernatant of GS115-BJ46a. (b) Separation of the culture supernatant of GS115-BJ46a by immobilized metal affinity chromatography (IMAC) separation of the culture supernatant of GS115-BJ46a by IMAC on Ni Sepharose 6 FF (16 Â 100 mm) equilibrated with 20 mmol/l sodium phosphate, 0.5 mol/l NaCl, 40 mmol/l imidazole, pH 8.0; eluted with 20 mmol/l sodium phosphate, 0.5 mol/l NaCl, 500 mmol/l imidazole, pH 8.0. Effluent of 5 ml/tube was collected at 3 ml/min. The absorbance at 280 nm was measured. The inhibitory effect of collagenase was eluted in peak BJ46a III, IV, and V. (c) SDS-PAGE analysis of BJ46a III, IV, and V and purified recombinant protein by Endo H enzyme digestion. 1, protein standard marker; 2, GS115-pPICZaA culture supernatant (48 h); 3–5, BJ46a III, IV, and V by IMAC; 6, GS115-BJ46A culture supernatant (48 h); 7, rBJ46a after purification by ultrafiltration and desalting; 8, Endo H enzyme digestion of purified rBJ46A protein; 9, Endo H enzyme. (d) Western blot analysis of the culture supernatant of GS115-BJ46A and rBJ46A after purification. 1, GS115-pPICZaA culture supernatant (48 h); 2, culture supernatant of GS115-BJ46A (48 h); 3, rBJ46A after purification.

and concentrated at 10 000g for 30 min in a SpeedVac Statistical analyses concentrator (Savant; E-C Instruments, Niantic, Con- All the experiments were repeated at least three times. necticut, USA). The protein concentration was measured Data are represented as mean±SD. One-way analysis of using BCA protein assay reagents. Equal amounts of variance, followed by Fisher’s least significant difference conditioned media were mixed with no reducing sample post-hoc test or Tamhane’s t-test (if equal variances were buffer, incubated for 15 min at room temperature, and not assumed) was used to compare the data from then electrophoresed on 10% SDS-PAGE gels containing different groups. P less than 0.05 was considered 1 mg/ml gelatin. After electrophoresis, the gels were statistically significant. washed with 2.5% Triton X-100 twice for 30 min, rinsed three times for 30 min with a 50 mmol/l Tris-HCl buffer Results (pH 7.6) containing 5 mmol/l CaCl2, 0.02% Brij-35, and Production, purification, and identification of 0.2% sodium azide, and then incubated at 371C over- recombinant BJ46a night. The gels were stained with a 0.5% Coomassie The expression of the rBJ46a supernatant was the highest brilliant blue R-250 solution containing 10% acetic acid at 48 h after induction, and the maximum yield was and 20% methanol for 30 min, and then destained with 26 mg/l (Fig. 1a). The culture supernatant was eluted by 7.5% acetic acid solution containing 10% methanol. Areas IMAC, and seven peaks were evident (Fig. 1b). The of gelatinase activity were detected as clear bands against eluates of peaks III–V showed inhibitory activity. There- the blue-stained gelatin background and the gels were after, the eluates of peaks III–V were pooled for scanned with Image Scanner II and analyzed with ultrafiltration and desalting. Approximately 10.15 mg/l ImageQuant TLv 2003.03 Software (GE Healthcare, of bioactive BJ46a was obtained with a purity of 98.06%. Pittsburgh, Pennsylvania, USA). As BJ46a contains four putative N-glycosylation sites,

Fig. 2

Matched peptides are shown in bold type

1 MNSLVALVLL GQIIGSTLSS QVRGDLECDE KDAKEWTDTG VRYINEHKLH

51 GYKYALNVIK NIVVVPWDGD WVAVFLKLNL LETECHVLDP TPVKNCTVRP

101 QHNHAVEMDC DVKIMFNVDT FKEDVFAKCH STPDSVENVR RNCPKCPILL

151 PSNNPQVVDS VEYVLNKHNE KLSDHVYEVL EISRGQHKYE PEAYYVEFAI

201 VEVNCTAQEL HDDHHHCHPN TAGEDHIGFC RATVFRSHAS LEKPKDEQFE

251 SDCVILHVKE GHAHSHLIQQ HVEKDSISPE HNNTALNFVH PHNDTSTSHE

301 SHEHLAEVPV AFVKKELPKD ISDRHTTPVK GCPGKVHHFE L

Alignment of the amino acid sequence of BJ46a. The alignment indicated that the amino acid sequence by mass spectrometry analysis matched that expected for BJ46a (gi|48428681); matched peptides are shown in bold type. The score is 91 and the expected value is 0.0052. Ion score is –10Â log (P), where P is the probability that the observed match is a random event. Protein scores greater than 81 are significant (P < 0.05). Protein scores are derived from ions scores on a nonprobabilistic basis for ranking protein hits.

located at amino acids 56–59, 165–168, 243–246, and Inhibitory effect of recombinant BJ46a on the invasion 254–257; the anomalous higher value obtained by SDS- of tumor cells in vitro PAGE is probably an artifact as a result of the glycosylated To determine the effects of rBJ46a on the invasion of the state of BJ46a. Fortunately, Endo H can cleave various tumor cells in vitro, B16F10 and MHCC97H cells were high mannose glycans. After treatment with Endo H, the treated with five doses of rBJ46a at 50, 100, 200, or deglycosylated recombinant protein migrated as a single 400 mg/ml for 24 h and Matrigel invasion assays were band of 38 kDa, which confirmed the occurrence of performed. As shown in Fig. 4, the numbers of invading glycosylation on the recombinant protein (Fig. 1c). B16F10 cells after treatment with rBJ46a were signifi- Western blot analyses showed that the bioactive rBJ46a cantly lower than those of the untreated control had a molecular weight of 57.9 kDa (Fig. 1d). (P < 0.01) (Fig. 4A and C), and the inhibition values were 18.6, 29.2, 33.3, 38.3, and 49.7%, respectively. The The identity of the purified protein was determined by numbers of invading MHCC97H cells treated with MS. As shown in Fig. 2, the recombinant protein rBJ46a were significantly reduced (P < 0.01) (Fig. 4B produced by P. pastoris shared 100% identity with the and C) and the inhibition values were 24.2, 28.3, 33.5, deduced amino acid sequence of BJ46a. The actual and 35.3%, respectively. These data show that rBJ46a can molecular mass was 38 745 Da by MS and consistent with inhibit tumor cell invasion in vitro. However, no effect was the literature, in which BJ46a in the native state has a observed on the proliferation, apoptosis, and migration of mass of 38 166 a.m.u. by MALDI-TOF MS following tumor cells (data not shown). chemical deglycosylation.

Recombinant BJ46a shows high inhibitory activity Recombinant BJ46a inhibits the invasion, metastasis, against matrix metalloproteinases and growth of tumor cells in vivo The inhibitory profile of rBJ46a against MMPs was To further evaluate the antitumor activity of rBJ46a, analyzed over 2 h in the presence of increasing concen- experimental lung colonization assays in C57BL/6 mice trations of the inhibitor using the substrate gelatin and and spontaneous lung metastases assays in nude assessment of the residual activity was carried out using mice were carried out. The lung colonization in C57BL/ fluorescence detection. MMP activities decreased with 6 mice after treatment at 1, 2, and 4 mg/kg rBJ46a was increasing concentrations of the recombinant protein. significantly inhibited as compared with untreated mice At 0.5 mg/ml, rBJ46a effectively inhibited the activity of (P < 0.05) (Fig. 5A and C), and the inhibition values were the MMPs with an IC50 of 0.119 mg/ml (Fig. 3a and b). 24.4, 28.1, and 29.4%, respectively (Fig. 5D). In nude Kinetic analyses by a series double reciprocal mice, rBJ46a treatment at 1, 2, and 4 mg/kg significantly Lineweaver–Burk plots (1/V vs. 1/S) method indicated a decreased the formation of lung tumor colonies competitive mode of inhibition, with rBJ46a showing (P < 0.05) (Fig. 6A and C), and the inhibition values highly effective inhibitory efficiency against MMPs were 21.8, 23.1, and 27.8%, respectively (Fig. 6D). These (Ki = 13.6 nmol/l, Fig. 3c). These results suggest that results show that rBJ46a can inhibit the invasion and rBJ46a is an active MMPI. metastasis of tumor cells in vivo.

Fig. 3

(a) (b) 90 0.05 U/ml 0.2 U/ml 0.4 U/ml 80 )

1000 % 70 60 800 50 600 40

400 30

Fluorescence 20

200 Fluorescence intensity ( 10 0 0 0 50 100 150 200 250 300 350 0 0.1 0.2 0.3 0.4 0.5 Time (min) BJ46a (mg/ml)

1 0.06 0.04 0.02 0 −0.04 −0.02 0 0.02 0.04 0.06 0.08 0.1 0.12 1/S

Activity assessment of recombinant BJ46a (rBJ46a). (a) Kinetics of the Enzchek gelatinase/collagenase reaction. When the activity of the matrix metalloproteinase (MMP) was 2 Â 10 –3U/ml, a 200-min incubation period was the optimal time to run the reaction. (b) The curve of BJ46a inhibitory activity. When increasing the concentration of recombinant protein, MMP activity decreased at the same time of 200 min. (c) The double-reciprocal Lineweaver–Burk plots (1/V vs. 1/S) of rBJ46a inhibitory activity. a, Without rBJ46a; b, with rBJ46a.

In terms of the tumor growth in vivo, the lung weights in concentrations of rBJ46a, the enzymolysis strip volumes C57BL/6 mice after rBJ46a treatment at 1, 2, and 4 mg/kg of proMMP-2, actMMP-2, proMMP-9, and actMMP-9 were increased more slowly than those in untreated mice, and reduced in a dose-dependent manner (Fig. 7a). In the the inhibition values were 21.9, 25.2, and 27.6%, MHCC97H cells treated without rBJ46a, the enzymolysis respectively (P < 0.05) (Table 1). In nude mice, the strip volumes of proMMP-2, actMMP-2, proMMP-9, and weights of the primary tumors after rBJ46a treatment actMMP-9 were 169±9, 83±15, 710±33, and 210± (4 mg/kg) were lower than those in control mice and the 11 mm3, respectively, and in the MHCC97H cells treated inhibition value was 26.7% (P < 0.05) (Table 1). These with rBJ46a, the enzymolysis strip volumes of proMMP-2, data show that rBJ46a can inhibit the growth of B16F10 actMMP-2, proMMP-9, and actMMP-9 were reduced in a cells and MHCC97H cells in vivo. dose-dependent manner (Fig. 7b) and the Km values of rBJ46a against MMP2 and MMP9 enzymatic activities that may be expressed in both B16F10 and MHCC97H cells Recombinant BJ46a suppresses matrix were 3.6 and 1.4 mmol/l, respectively. These results indicate metalloproteinase 2 and matrix metalloproteinase 9 that rBJ46a can suppress MMP9 and MMP2 activities by enzymatic activities tumor cells. It has been shown that MMP2 and MMP9 are the prominent MMPs responsible for ECM degradation. Therefore, to test whether the enzymatic activity of Discussion these two MMPs is regulated by rBJ46a, the tumor cells To further evaluate the anticancer effects of BJ46a, in this (B16F10/MHCC97H) were treated with various concen- study, we first optimized the P. pastoris system for trations of rBJ46a at 25, 50, 100, 200, or 400 mg/ml for 24 h the production of rBJ46a on the basis of previous and the activity of MMP2 and MMP9 was assayed by reports [10,22–24]. The conditions were optimized as gelatin zymography. As shown in Fig. 7, in the B16F10 cells follows: culture volume of 30% of the total flask volume; treated without rBJ46a, the enzymolysis strip volumes of flasks covered with sterile gauze to provide adequate proMMP-2, actMMP-2, proMMP-9, and actMMP-9 were ventilation; and the addition of yeast extract and peptone 723±35, 719±25, 770±32, and 250±21 mm3, respectively, in the expression medium to improve the yield and whereas in the B16F10 cells treated with various reduce the degradation of secretory proteins [25,26].

Fig. 4

(A) (B) ab ab

cd cd

ef ef

Number of invasion cell 5

0 B16F10 MHCC97H

Effects of recombinant BJ46a (rBJ46a) on Matrigel invasion of B16F10 cells and MHCC97H cells. Representative images (crystal blue stain, Â 200) of invading B16F10 cells (A) and MHCC97H (B) cells in the absence and presence of rBJ46a. a, Control; b, 25 mg/ml; c, 50 mg/ml; d, 100 mg/ml; e, 200 mg/ml; f, 400 mg/ml. (C) The numbers of invading cells. *P < 0.01 versus control.

The maximum yield was 26 mg/l by shaking-flask (Ki = 13.6 nmol/l). This specific activity was comparable fermentation. Approximately 10.15 mg/l of bioactive with those of other MMPIs, such as recombinant TIMP-4 material was obtained with a purity of 98.06%. Enzchek (Ki = 14.5 nmol/l) from P. pastoris expression [32], re- gelatinase/collagenase assay showed that rBJ46a effec- combinant TIMP-1 (Ki = 11.8 nmol/l) from CHO-K-1 tively inhibited the activity of the MMPs with an IC50 of cells, and recombinant TIMP-2 (Ki = 14.9 nmol/l) from 0.119 mg/ml. The inhibitory activity was comparable with 293-EBNA cells [33]. These data show that the those of other MMPIs, such as batimastat (IC50 values optimized P. pastoris system for recombinant protein 10 ng/ml) and marimastat (IC50 values 3 ng/ml) of production, as reported here, allowed for high recovery, peptidomimetic MMPIs [27,28], BAY12-9566 (IC50 high purity, and high activity, which provided the values below 0.13 mg/ml) and AG3340 (IC50 values below preconditions for future evaluation of its applications in 0.13 ng/ml) of nonpeptidic MMPIs [29,30], and Col-3 the pharmaceutical industry. (IC50 values 12 mg/ml) of tetracycline derivatives. Overall, the inhibitory spectrum of rBJ46a was similar to those of Subsequently, we showed that the rBJ46a significantly synthesized MMPIs [31]. Kinetic analyses showed a blocked the invasion of B16F10 cells and MHCC7H cells competitive mode of inhibition with rBJ46a showing in vitro in a dose-dependent manner by Matrigel invasion highly effective inhibitory efficiency against MMPs assays, indicating that rBJ46a has the potential to degrade

Fig. 5

(A) a (B)

c c d

20 %

) ( 40 C 16 ∗ / ∗ ∗ T − 30 12 C

8 20

4 Inhibiting ratio of 10 colonization ( Number of colonization 0 0 abcd abcd Recombinant BJ46a Recombinant BJ46a

Effects of recombinant BJ46a (rBJ46a) on B16F10 cells lung colonization in vivo. B16F10 cells were injected into the tail veins of C57BL/6 mice. C57BL/6 mice inoculated in the lateral tail vein with B16F10 cells were treated with three i.v. injections of rBJ46a (1, 4, and 4 mg/kg) 24 h before cell inoculation, and 2 and 24 h after cell inoculation. After 21 days, the surface pulmonary metastases were counted. (A) The colonization (arrows) of treated and untreated mice. (B) Histological examination of lung tumor colonies (arrows) (hematoxylin and eosin stain, Â 200). (C) The numbers of colonies were lower after treatment as compared with untreated mice (control). *P < 0.05 versus control. (d) Inhibition values in the treated and control groups. a, Control; b, 1 mg/kg; c, 2 mg/kg; d, 4 mg/kg.

ECM and basement membranes. In addition, the admin- inhibit invasion and metastasis of MHCC97H cells in vivo istration of rBJ46a by multiple i.p. injections prevented (data not shown). Taken together, these data further show spontaneous pulmonary metastases of MHCC97H cells in that rBJ46a played an inhibitory role in the invasion and nude mice, and the administration of rBJ46a by three i.v. metastasis of tumor cells in vitro and in vivo. It seems that injections was sufficient to inhibit pulmonary colonization rBJ46a not only inhibits hematogenous invasion of tumor of B16F10 cells in C57BL/6 mice. The inhibition rate of cells but also prevents tumor cells from escaping from the tumor cell invasion and metastasis is similar to many primary tumor and establishing colonies at local and MMPIs, including synthetic and natural inhibitors [29–34]. distant sites. These findings provide novel evidence that However, the inhibitory effect of rBJ46a on tumor the use of rBJ46a may be an effective approach to inhibit metastasis in vivo is relatively modest, which might be the invasion and metastasis of tumor cells through associated with the rBJ46a concentration entering the multiple targets. However, the molecular mechanism tumor. In addition, the molecular complexity of cancer underlying the inhibitory action of rBJ46a on tumor progression suggests that the appropriate combination of invasion and metastasis remains unclear. rBJ46a and other MMPIs or molecular targeted agents may increase drug efficacy. In this respect, we recently It has been shown that the proteolytic activity of MMPs determined the effects of combining rBJ46a and snake is required for the degradation of ECM and basement venom cystatin (sv-cystatin) on invasion and metastasis of membranes during invasion and metastasis [2], and this MHCC97H cells. The results showed that the common study showed that rBJ46a shows high inhibitory activity expression of BJ46a and sv-cystatin could synergistically against MMPs. Therefore, it is reasonable to conclude

Fig. 6

(A) (B)

a a

b b

c c

d d

12 % 25 ∗ ) (

∗ C

10 / ∗ 20 T

8 − C 15 6 10 4 2 Inhibiting ratio of 5 Number of metastasis metastasis ( 0 0 abcd abcd Recombinant BJ46a Recombinant BJ46a Effects of recombinant BJ46a (rBJ46a) on the metastasis of MHCC97H cells in nude mice. MHCC97H cells were inoculated s.c. into the right flank of nude mice. The treatment was initiated 24 h after cell inoculation. rBJ46a was administered by i.p. injections once daily for 5 days. After 6 weeks, lung tissues were removed and stained. (A) The representative images of nude mice (left) and tumor tissues (right) at 6 weeks. (B) Representative lung sections of metastases (arrows) (hematoxylin and eosin stain, Â 200). (C) The numbers of metastases were lower after treatment as compared with untreated mice (control). *P < 0.05 versus control. (D) Metastasis inhibition values of treated and control groups. a, Control; b, 1 mg/kg; c, 2 mg/kg; d, 4 mg/kg.

Table 1 Effects of recombinant BJ46a on tumor growth in vivo

C57BL/6 mice Nude mice

Group Weight (g) Inhibiting ratio (%) Weight Inhibiting ratio (%) Tumor volume (cm3)

Control 0.63±0.06 – 3.81±0.73 – 1.43±0.12 1 mg/kg 0.49±0.06* 21.9 3.26±0.85 14.5 1.27±0.06 2 mg/kg 0.47±0.06* 25.2 3.23±1.04 15.0 1.19±0.10 4 mg/kg 0.46±0.05* 27.6 2.80±0.39* 26.7 1.07±0.05*

The lung tissues from C57BL/6 mice and tumor tissues from nude mice were weighed. The approximate primary tumor volumes from MHCC97H cells were calculated. The ratio of growth inhibition (%) = (C–T)/C Â 100, where T and C are the relative mean tumor weights in the treatment and control groups, respectively. *P < 0.05 vs. control. that rBJ46a inhibits tumor invasion and metastasis by ECM [35,36]. In particular, MMP2 and MMP9 are blocking MMP activities. However, the MMP family gelatinases that play key roles in the invasion and consists of a variety of proteolytic enzymes that mediate metastasis of cancer cells through the destruction of the the degradation of many different components of the basal membrane and ECM [37]. To further identify the

Fig. 7

(a) 1 234567

130 pro-MMP-9 95 act-MMP-9 72 pro-MMP-2

55 act-MMP-2

Control 25 μg/ml 50 μg/ml 100 μg/ml 200 μg/ml 400 μg/ml

900 800 ∗ 700 600 500 ∗ 400 ∗ ∗ 300 200

Enzymolysis strip volume 100 0 pro-MMP-2act-MMP-2 pro-MMP-9 act-MMP-9

(b) 1 234567

130 pro-MMP-9 95 act-MMP-9 72 pro-MMP-2 55 act-MMP-2

Control 25 μg/ml 50 μg/ml 100 μg/ml 200 μg/ml 400 μg/ml 800 700 600 500 400 300 ∗ 200 ∗ ∗ ∗ 100 ∗ Enzymolysis strip volume 0 pro-MMP-2 act-MMP-2 pro-MMP-9 act-MMP-9

Gelatin zymographic assay for matrix metalloproteinases (MMPs) of tumor cells treated with recombinant BJ46a (rBJ46a). (a) Effect of rBJ46a on MMP-2/MMP-9 activity of B16F10 cells. (b) Effect of rBJ46a on MMP-2/MMP-9 activity of MHCC97H cells. 1, Protein standard marker; 2, control; 3, 25 mg/ml; 4, 50 mg/ml; 5, 100 mg/ml; 6, 200 mg/ml; 7, 400 mg/ml. *P<0.05 versus control.

molecular targets for the inhibition of invasion and may function as a target inhibitor of MMPs or TIMPs to metastasis of tumor cells by rBJ46a, a gelatin zymography reduce the activities of MMP2/MMP9 in tumor cells, experiment was performed. We showed that MMP9 and which then decreases the degradation of the ECM and MMP2 enzymatic activities in tumor cells were sup- basement membrane to inhibit tumor cell metastasis. pressed by rBJ46a in a dose-dependent manner and the Km value of rBJ46a against MMP2 and MMP9 enzymatic Until now, there has been no direct evidence showing the activities expressed in both B16F10 and MHCC97H cells link of MMP with BJ46a. BJ46a, a 46 kDa protein isolated was 3.6 and 1.4 mmol/l, respectively, which was compar- from the serum of B. jararaca, is an inhibitor of SVMPs. able with those of other MMPIs, such as TIMP-1, TIMP- BJ46a is a member of the fetuin family and cystatin 2, and TIMP-4 [32,33]. These findings show that rBJ46a superfamily, which interacts with metalloproteinases

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. B16F10 MHCC97H cell inhibition by rBJ46A Ji et al.471 through the formation of a noncovalent complex, with epithelial to mesenchymal transition. Although specifically inhibiting atrolysin C (class P-I) and jararha- classified as members of the cystatin superfamily, BJ46a gin (class P-III) proteolytic activities [6]. Because SVMP is unlikely to be a viable cysteine protease inhibitor shares homology with the zinc-binding environment of because the consensus sequence for the cystatin active MMPs, it is possible that BJ46a may be able to modulate site, QXVXG, has been extensively altered in this anticancer invasion and metastasis through the formation molecule. Corroborating this view, BJ46a failed to inhibit of a complex with MMP. Recently, we predicted the 3D cathepsin B activity [6]. These data indicate that BJ46a structure of rBJ46a and proposed that BJ46a does not may function only as a target inhibitor of MMP2 and interact as a dimer with the metalloproteinase, although MMP9, but not as an off-target inhibitor of cysteine it is a dimer protein in its native state. Instead, BJ46a may protease and cathespin B, to inhibit epithelial– inhibit MMPs because of its two cystatin-like domains in mesenchymal transition, and tumor invasion and metas- one monomer, which interact with the catalytic zinc, tasis [38–40]. It will be necessary to test this hypothesis leading to enzymatic inhibition [10]. Further studies in the future. using immunoprecipitation are required to prove the In conclusion, in this study, we showed for the first time binding of BJ46a and MMPs. that rBJ46a in an optimized P. pastoris expression system Previous studies in our laboratory have shown that can inhibit the invasion and metastasis of tumor cells by transfection of the BJ46a gene into B16 cells resulted in reducing MMP2/MMP9 activities. These results indicate the inhibition of invasion potential in vitro and in vivo [8]. that rBJ46a may be a promising candidate therapeutic Currently, we have also shown that transfection of the agent for antimetastasis of tumor cells, and may have BJ46a gene into MHCC97H cells results in the inhibition potential for pharmaceutical applications. For future of invasion in vitro and in vivo (data not shown). In this studies, we hope to focus on evaluating the safety and study, we have further shown that rBJ46a can inhibit the efficacy of this agent (including pharmacokinetics, invasion and metastasis of B16F10 and MHCC97H cells pharmacodynamics, toxicology, tolerability, and immuno- in vitro and in vivo by blocking MMP2 and MMP9 genicity) to facilitate studies into pharmaceutical appli- activities. The role of MMPs is regulated at multiple cations. levels, such as by gene expression, protein amounts, and proteolytic activity. Many studies have shown that Acknowledgements the expression of MMP2 and MMP9 by tumor cells is This research was supported by the grants from Sci-Tech associated with tumor invasion and metastasis [2,3]. Projects of the Educational Department in Fujian However, further studies are required to examine the Province (No. JB12109) and the Science Foundation of effects of BJ46a on the expression of MMP2 and MMP9 Fujian (No. XZ04002, 2004-686). to gain further insights into its inhibitory mechanism of action on tumor invasion and metastasis. Conflicts of interest It has been shown that the expression and activities of There are no conflicts of interest. MMP2 and MMP9 play a key role in tumor angiogenesis because they activate several proangiogenic factors, such as vascular endothelial growth factor, basic fibroblast References growth factor, and transforming growth factor-b, and 1 Deep G, Agarwal R. Antimetastatic efficacy of silibinin: molecular mechanisms and therapeutic potential against cancer. Cancer Metastasis degrade the extracellular components, such as collagen Rev 2010; 29:447–463. type IV, XVIII, and perlecan, respectively, facilitating 2 Gialeli C, Theocharis AD, Karamanos NK. Roles of matrix metalloproteinases tumor angiogenesis [2,3]. This led us to speculate that in cancer progression and their pharmacological targeting. FEBS J 2011; 278:16–27. BJ46a may inhibit tumor angiogenesis by reducing the 3 Benjamin MM, Khalil RA. Matrix metalloproteinase inhibitors as investigative activities of MMP2 and MMP9. Further studies are tools in the pathogenesis and management of vascular disease. EXS 2012; required to determine the effects of BJ46a on down- 103:209–279. 4 El-Refael MF, Sarkar NH. Snake venom inhibits the growth of stream signaling events associated with each MMP mouse mammary tumor cells in vitro and in vivo. Toxicon 2009; activity. 54:33–41. 5 Evangelista IL, Martins AM, Nascimento NR, Havt A, Evangelista JS, More recently, we showed that the transfection of de Noro˜ es TB, et al. Renal and cardiovascular effects of Bothrops marajoensis venom and phospholipase A2. Toxicon 2010; 55:1061–1070. MHCC97H cells with sv-cystatin repressed tumor 6 Valente RH, Dragulev B, Perales J, Fox JY, Domont GB. A snake venom invasion and metastasis, mainly by inhibiting cathepsin metalloproteinase inhibitor. Isolation, characterization, cloning and insights B, MMP2, and MMP9 activities, as well as epithelial– into its mechanism of action. Eur J Biochem 2001; 268:3042–3052. 7 Xu JW, Lin JY, Lin X, Qi YL, Zhang XY. Gene synthesis and construction of mesenchymal transition, suggesting that sv-cystatin may baculovirus expression vector of a snake venom metalloproteinase inhibitor. function as a target inhibitor of cysteine protease and Chin J Biomed Eng 2006; 25:755–764. cathespin B, and as an off-target inhibitor of MMPs to 8 Xu JW, Lin JY, Lin X, Qi YL, Zhang XY, Zhu JW, et al. Inhibition effect of a snake venom metalloproteinase inhibitor on experimental pulmonary inhibit tumor invasion and metastasis. In addition, metastasis in mice inoculated with B16 melanoma cells. China Biotechnol overexpression of MMP2 and MMP9 has been associated 2009; 29:28–32.

9 Xu JW, Lin JY, Lin X, Zhang XY, Huang QL. Recombinant protein exerts 24 Zhou J, Chu J, Wang YH, Wang H, Zhuang YP, Zhang SL. Purification and inhibition effect on invasion and metastasis of B16 cells in vitro. Acta Anat bioactivity of exendin-4, a peptide analogue of GLP-1, expressed in Pichia Sin 2009; 40:749–754. pastoris. Biotechnol Lett 2008; 30:651–656. 10 Shi Y, Ji MK, Xu JW, Lin X, Lin JY. High-level expression, purification, 25 Shumiao Z, Huang J, Zhang C, Deng L, Hu N, Liang Y. High-level expression characterization and structural prediction of a snake venom of an Aspergillus niger endo-b-1,4-glucanase in Pichia pastoris through metalloproteinase inhibitor in Pichia pastoris. Protein J 2012; 31:212–221. gene codon optimization and synthesis. J Microbiol Biotechnol 2010; 11 Tian J, Tang ZY, Ye SL, Liu YK, Lin ZY, Chen J, et al. New human 20:467–473. hepatocellular carcinoma (HCC) cell line with highly metastatic potential 26 Zhao W, Zheng J, Zhou HB. A thermotolerant and cold-active mannan (MHCC97) and its expressions of the factors associated with metastasis. endo-1,4-b-mannosidase from Aspergillus niger CBS 513.88: constitutive Br J Cancer 1999; 81:814–821. overexpression and high-density fermentation in Pichia pastoris. Bioresour 12 Chen YL, Wang JY, Chen SH, Yang BC. Granulocytes mediates the Fas-L- Technol 2011; 102:7538–7547. associated apoptosis during lung metastasis of melanoma that determines 27 Hidalgo M, Eckhardt SG. Development of matrix metalloproteinase inhibitors the metastatic behaviour. Br J Cancer 2002; 87:359–365. in cancer therapy. J Natl Cancer Inst 2001; 93:178–193. 13 Guruvayoorappan C, Kuttan G. Effect of amentoflavone on the inhibition of 28 Wojtowicz-Praga SM, Dickson RB, Hawkins MJ. Matrix metalloproteinase pulmonary metastasis induced by B16F10 melanoma cells in C57BL/ inhibitors. Invest New Drugs 1997; 15:61–75. 6 mice. Integr Cancer Ther 2007; 6:185–197. 29 Brown PD. Clinical studies with matrix metalloproteinase inhibitors. APMIS 14 Calik P, Orman MA, Celik E, Halloran SM, Calik G, Ozdamar TH. Expression 1999; 107:174–180. system for synthesis and purification of recombinant human growth hormone 30 Shalinsky DR, Brekken J, Zou H, McDermott CD, Forsyth P, Edwards D, in Pichia pastoris and structural analysis by MALDI-ToF mass spectrometry. et al. Broad antitumor and antiangiogenic activities of AG3340, a potent and Biotechnol Prog 2008; 24:221–226. selective MMP inhibitor undergoing advanced oncology clinical trials. Ann N 15 Maley F, Trimble RB, Tarentino AL, Plummer THJ. Characterization of Y Acad Sci 1999; 878:236–270. glycoproteins and their associated oligosaccharides through the use of 31 Lokeshwar BL. MMP inhibition in prostate cancer. Ann N Y Acad Sci 1999; endoglycosidases. Anal Biochem 1989; 180:195–204. 878:271–289. 16 Della Porta P, Soeltl R, Krell HW, Collins K, O’Donoghue M, Schmitt M, 32 Bernd S, Martin F, Harald T. Characterization of C-terminally truncated Kru¨ger A. Combined treatment with serine protease inhibitor aprotinin and human tissue inhibitor of metalloproteinases-4 expressed in Pichia pastoris. matrix metalloproteinase inhibitor Batimastat (BB-94) does not prevent Biol Chem 2001; 382:987–991. invasion of human esophageal and ovarian carcinoma cells in vivo. 33 Troeberg L, Tanaka M, Wait R, Shi YE, Brew K, Nagase H. E. coli expression Anticancer Res 1999; 19:3809–3816. of TIMP-4 and comparative kinetic studies with TIMP-1 and TIMP-2: insights 17 Altemtam N, Nahas ME, Johnson T. Urinary matrix metalloproteinase activity into the interactions of TIMPs and matrix metalloproteinase 2. Biochemistry in diabetic kidney disease: a potential marker of disease progression. 2002; 41:15025–15035. Nephron Extra 2012; 2:219–232. 34 Watson SA, Morris TM, Robinson G, Crimmin MJ, Brown PD, Hardcastle JD. 18 Gil-Ad I, Shtaif B, Levkovitz Y, Nordenberg J, Taler M, Korov I, et al. Inhibition of organ invasion by the matrix metalloproteinase inhibitor Phenothiazines induce apoptosis in a B16 mouse melanoma cell batimastat (BB-94) in two human colon carcinoma metastasis models. line and attenuate in vivo melanoma tumor growth. Oncol Rep 2006; Cancer Res 1995; 55:3629–3633. 15:107–112. 35 Klein T, Bischoff R. Physiology and pathophysiology of matrix 19 Xie Q, Tang NH, Wan R, Qi YL, Lin X, Lin JY. Recombinant snake venom metalloproteases. Amino Acids 2011; 41:271–290. cystatin inhibits the growth, invasion and metastasis of B16F10 cells and 36 Lemaıˆtre V, D’Armiento J. Matrix metalloproteinases in development MHCC97H cells in vitro and in vivo. Toxicon 2011; 57:704–711. and disease. Birth Defects Res C Embryo Today 2006; 78:1–10. 20 Tu SP, Jiang XH, Lin MC, Cui JT, Yang Y, Lum CT, et al. Suppression of 37 Sullu Y, Demirag GG, Yildirim A, Karagoz F, Kandemir B. Matrix survivin expression inhibits in vivo tumorigenicity and angiogenesis in gastric metalloproteinase-2 (MMP-2) and MMP-9 expression in invasive ductal cancer. Cancer Res 2003; 63:7724–7732. carcinoma of the breast. Pathol Res Pract 2011; 207:747–753. 21 Iwai S, Nakanishi-Ueda T, Armstrong D, Oguchi K. Zymographical 38 Westermarck J, Kahari VM. Regulation of matrix metalloproteinase techniques for detection of matrix metalloproteinases. Methods Mol Biol expression in tumor invasion. FASEB J 1999; 13:781–792. 2008; 477:113–121. 39 Tang NH, Xie Q, Wang XQ, Li XJ, Chen YL, Lin X, et al. Expression of snake 22 Garde S, Fraser JE, Nematpoor N, Pollex R, Morin C, Forte A, et al. Cloning, venom cystatin in human MHCC97H cells inhibits invasion and metastasis expression, purification and functional characterization of recombinant through reduction of proteinases activity and epithelial–mesenchymal human PSP94. Protein Expr Purif 2007; 54:193–203. transition. Arch Pharm Res 2011; 34:781–789. 23 Fong MY, Asha T, Azdayanti M, Yee LL, Sinnadurai S, Rohela M. Inferring the 40 Gomis-Ru¨th FX, Maskos K, Betz M, Bergner A, Huber R, Suzuki K, et al. phylogenetic position of Brugia pahangi using 18S ribosomal RNA (18S Mechanism of inhibition of the human matrix metalloproteinase stromelysin-1 rRNA) gene sequence. Trop Biomed 2008; 25:87–92. by TIMP-1. Nature 1997; 389:77–81.