Skip to main content
NCBI home page
Cancer Science logoLink to Cancer Science
. 2017 Mar 16;108(3):347–353. doi: 10.1111/cas.13134

Activation of MMP‐9 by membrane type‐1 MMP/MMP‐2 axis stimulates tumor metastasis

Zichen Li 1, Takahisa Takino 1, Yoshio Endo 2, Hiroshi Sato 1,✉
PMCID: PMC5378257  PMID: 27987367

Abstract

An artificial receptor for proMMP‐9 was created by fusing tissue inhibitor of MMP‐1 (TIMP‐1) with type II transmembrane mosaic serine protease (MSP‐T1). Expression of MSP‐T1 in 293T cells induced binding of proMMP‐9, which was processed by MMP‐2 activated by membrane type 1 MMP (MT1‐MMP). HT1080 cells transfected with the MSP‐T1 gene produced activated MMP‐9 in collagen gel, and addition of proMMP‐2 to the culture augmented it, which resulted in intensive collagen digestion. These cells metastasized into chick embryonic liver more than control cells. Treatment of HT1080 cells with concanavalin A in the presence of exogenous proMMP‐2 induced activation of not only proMMP‐2 but also proMMP‐9. Knockdown of MT1‐MMP or TIMP‐2 expression with siRNA suppressed activation of both proMMP‐2 and proMMP‐9. Transfection of TIMP‐1 siRNA suppressed cell binding and activation of proMMP‐9, but not proMMP‐2 activation. Knockdown of a disintegrin and metalloproteinase 10 (ADAM10) expression reduced cell binding and processing of proMMP‐9. These results suggest that proMMP‐9, which binds to a receptor complex containing TIMP‐1 and ADAM10, is activated by the MT1‐MMP/MMP‐2 axis, and MMP‐9 thus activated stimulates cellular proteolysis and metastasis.

Keywords: Activation, MMP‐2, MMP‐9, MT1‐MMP, TIMP‐1

Matrix metalloproteinase‐2 and MMP‐9 are type IV collagenase/gelatinase, and their expression is often associated with tumor aggressiveness and a poor prognosis (reviewed in Ref. 1, 2). Membrane type 1 MMP was first identified as an activator of latent MMP‐2 (proMMP‐2).3 In addition to proMMP‐2, a variety of substrates of MT1‐MMP were identified, including ECM proteins, cell adhesion molecules, cytokines, and others.4, 5, 6, 7 ProMMP‐2 is activated in a variety of tumor tissues in which MT1‐MMP is overexpressed. In contrast, proMMP‐9 activation was detected in limited numbers of tumor samples, such as non‐small‐cell lung carcinoma and colon carcinomas,8, 9, 10 and the active form of MMP‐9 was predominantly present in colon carcinoma patients with metastases.8 ProMMP‐9 can be processed in vitro by various proteases including MMP‐2, MMP‐3, and serine proteases, however, the molecular mechanism of proMMP‐9 activation in vivo still remains to be solved.1

In proMMP‐2 activation mechanism, the MT1‐MMP/TIMP‐2 complex serves as a receptor for proMMP‐2, which is then processed by adjacent TIMP‐2‐free MT1‐MMP.11, 12, 13, 14 To mimic this model, TIMP‐2 was fused with type II transmembrane MSP to create MSP‐T2 as an artificial receptor for proMMP‐2.15 It was reported that MSP‐T2 served as a receptor for proMMP‐2, and accelerated proMMP‐2 activation by MT1‐MMP.15 ProMMP‐9 is known to bind to TIMP‐1 in a similar manner as proMMP‐2 binds to TIMP‐2; this suggested model implicates TIMP‐1 as a bridging molecule to tether proMMP‐9 on the cell surface for its processing.1, 2

In the present study, a TIMP‐1 chimera protein with MSP, MSP‐T1, was constructed, which served as an artificial receptor for proMMP‐9 and induced proMMP‐9 activation by the MT1‐MMP/MMP‐2 axis. Stable expression of MSP‐T1 in HT1080 fibrosarcoma cells stimulated proMMP‐9 activation and metastasis. ProMMP‐9 processing through TIMP‐1‐containing receptor and the MT1‐MMP/MMP‐2 axis was shown in ConA‐treated HT1080 cells.

Materials and Methods

Cell culture

Human embryonic kidney 293T and fibrosarcoma HT1080 cells were obtained from ATCC (Manassas, VA, USA) and cultured in DMEM (Sigma, St. Louis, MO, USA) supplemented with 5% FCS. HT1080 cells stably expressing MSP‐T1 were generated by transfecting MSP‐T1 plasmid, and transfected cells were selected under 5 μg/mL puromycin (Sigma). Type I collagen Cellmatrix Type I(P) was purchased from Nitta Gelatin (Osaka, Japan). Recombinant TIMP‐1 and TIMP‐2 proteins were gifts from the Daiichi Fine Chemical Co. (Takaoka, Japan). A synthetic MMP inhibitor BB94 (batimastat) was a kind gift from the Kotobuki Pharmaceutical Co. (Nagano, Japan).

Antibodies

Monoclonal antibodies against MMP‐9 (56‐2A4), MT1‐MMP (222‐2D12), TIMP‐1 (7‐6C1), and TIMP‐2 (67‐4H11) were gifts from the Daiichi Fine Chemical Co. An antibody against ADAM10 was purchased from R&D Systems (Minneapolis, MN, USA).

Plasmids

Expression plasmids for MT1‐MMP, MSP‐T1, MMP‐2, and MMP‐9 were constructed in pEAK8 vector (Edge BioSystems, Gaithersburg, MD, USA) as described previously.16, 17, 18, 19 The fusion gene for MSP‐T1, which encodes amino acid residues 1–188 of MSP and 24–207 of TIMP‐1, was constructed by replacing the fragment encoding TIMP‐2 of MSP‐T2 with that of TIMP‐1.15 The human TIMP‐1 cDNA fragment encoding amino acid residues 24–207 was amplified by PCR using 5′ and 3′ primers with extra EcoRI and XbaI sites, respectively (GAATTCTGCACC TGTGTCCCA CCCCAC and TCTAGATCAGGCTA TCTGGGACCGCAG; restriction enzyme site). The cDNA encoding mutant MMP‐9 lacking C‐terminal hinge and hemopexin‐like domains (ΔMMP‐9) was amplified by PCR using a primer TCTAGATCA ACCATAGAGGTGCCGGATGCC (restriction site). The cDNA encoding MMP‐9 mutants with amino acid substitution at Met41 and Phe88 with Val was amplified by PCR using mutation primers CTCGGGTGGCAGAGGTGCGTGGAGAGTCGA and CAGA CCTGGGCAGAGTCC AAACCTTTGAGG (mutated nucleotide), respectively.

RNA interference

RNA interference technology was used to generate specific knockdown of MT1‐MMP, TIMP‐1, TIMP‐2, ADAM10, and MMP‐9 mRNA transcription. Small interfering RNA was prepared by FASMAK (Kanagawa, Japan). The siRNA target sequences were as follows: MT1‐MMP‐I, CCAGAAGCTGAA GGTAGAA; MT1‐MMP‐II, GCGATGAAGTCTTCACTTA; TIMP‐1‐I, TCAACCAG ACCACCTTATA; TIMP‐1‐II, GATGTATAAAGGGTTCCAA; TIMP‐2‐I, GGATCCA GTATGAGATCAA; TIMP‐2‐II, GAGATCAAGCAGATAAAGA; ADAM10‐I, GATA TCCAGTCATGTTAAA; ADAM10‐II, CTGGAATTATTACTGTTCA; MMP‐9‐I, CAC AACAUCACCUAUUGGA; MMP‐9‐II, GGAGUACUCGACCUGUACC. Scrambled oligo was used as a negative control. Small interfering RNA reverse transfection was carried out using Lipofectamine RNAiMAX (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instructions.

Zymography

ProMMP‐2, proMMP‐9, and proMMP‐9 mutant supernatants were prepared from 293T cells transfected with the respective expression plasmids, as previously described.21 293T cells cultured in a 24‐well microplate were transfected with expression plasmids, cultured in DMEM supplemented with 5% FCS for 24 h, and incubated in 300 μL Opti‐MEM (Invitrogen) containing proMMP‐2 and proMMP‐9. Both MMP‐2 and MMP‐9 in the supernatant were examined by mixing with a same volume of zymography sample buffer. For the detection of cell‐bound MMP‐2 and MMP‐9, cells were washed twice with PBS, and then dissolved in 200 μL sample buffer by sonication. These samples were incubated at 37°C for 20 min, and subjected to gelatin zymography gel containing gelatin labeled with Alexa Fluor 680 (Molecular Probes, Eugene, OR, USA). Gels were processed and monitored by a LI‐COR Biosciences Odyssey IR imaging system (Lincoln, NE, USA) as described previously.15

Western blot analysis

Cell lysates or proteins precipitated from conditioned medium with 10% trichloroacetic acid were analyzed by Western blotting with the indicated antibody. Goat anti‐mouse antibody conjugated with Alexa Fluor 680 was used as a secondary antibody. The signal was monitored using a LI‐COR Odyssey IR imaging system. Precision Plus Protein Standards were used as molecular weight standards (Bio‐Rad, Hercules, CA, USA).

Chick embryo assay

The assay was done as originally described by Endo et al.17 Briefly, HT1080 cells stably transfected with pEAK control plasmid or MSP‐T1 expression plasmid (1.0 × 105 cells per egg) were injected into the chorioallantoic membrane vein of chicken eggs 11 days after fertilization, and incubated for a further 7 days. Ten eggs were used for the injection of each type of cell. The number of HT1080 cells that colonized liver was estimated by quantitative real‐time PCR as described previously.15, 17

Gelatin degradation

Alexa Fluor 680‐labeled gelatin (150 ng/200 μL Opti‐MEM) was incubated with HT1080 cells in a 24‐well microplate for 1 h, separated on 10% SDS‐PAGE, and detected by a LI‐COR Odyssey IR imaging system.15

Results

MSP‐T1 functions as a receptor for proMMP‐9

293T cells transfected with MSP‐T1 and/or MT1‐MMP plasmids were incubated with proMMP‐2 and proMMP‐9, and cell‐bound MMP‐2 and MMP‐9 were examined by gelatin zymography. ProMMP‐9 bound to the cells expressing MSP‐T1 more effectively than to control cells (Fig. 1a, Cells). Expression of MT1‐MMP alone induced activation of proMMP‐2, but not proMMP‐9. Coexpression of MSP‐T1 with MT1‐MMP induced activation of both proMMP‐2 and proMMP‐9. To confirm that proMMP‐9 activation is mediated by transfected cells, transfected cells were incubated with proMMP‐2, and then supernatants from transfected cells were incubated with proMMP‐9 (Fig. 1a, Sup). ProMMP‐9 was not processed in the supernatant from cells expressing MSP‐T1, MT1‐MMP, and MMP‐2. Next, to examine the requirement of MMP‐2 in proMMP‐9 activation through MSP‐T1/MT1‐MMP, cells coexpressing MSP‐T1 and MT1‐MMP were incubated with proMMP‐9 and 2‐fold serially diluted proMMP‐2 (Fig. 1b). ProMMP‐9 activation was shown to depend on MMP‐2 concentration added to the culture. The time course experiment indicated that proMMP‐9 activation by the cells expressing MSP‐T1/MT1‐MMP/MMP‐2 was induced at 0.2 h and reached a plateau at 1–2 h after incubation (Fig. 1c). ProMMP‐9 binds to TIMP‐1 through the carboxy‐terminal hemopexin‐like domain. ProMMP‐9 mutant lacking hemopexin‐like domain (ΔMMP‐9) was compared with wild‐type proMMP‐9 for binding and processing by cells expressing MSP‐T1 and MT1‐MMP in the presence of proMMP‐2 (Fig. 1d). Wild‐type proMMP‐9 bound to the cells expressing MSP‐T1 and was activated by the cells expressing MSP‐T1 and MT1‐MMP in the presence of proMMP‐2. However, ΔproMMP‐9 failed to bind to MSP‐T1‐expressing cells and was not activated by the cells expressing MSP‐T1 and MT1‐MMP. ProMMP‐9 was shown to be cleaved in vitro by activated MMP‐2 at the Glu40–Met41 amide bond to generate an activation intermediate form, which is converted to the fully active form through the autocleavage at the Arg87–Phe88 amide bond.22 To confirm proMMP‐9 cleavage sites by the cells expressing MSP‐T1 and MT1‐MMP in the presence of proMMP‐2, wild‐type or mutant proMMP‐9 with the amino acid substitution at Met41 (Fig. 1e, Mut1) or Phe88 (Fig. 1e, Mut2) was incubated with the cells expressing MSP‐T1 and MT1‐MMP in the presence of proMMP‐2. Wild‐type proMMP‐9 was converted to fully active form by these cells, but proMMP‐9 mutant with amino acid substitution at Met41 (Mut1) was not processed. ProMMP‐9 mutant with an amino acid substitution at Phe88 (Mut2) was cleaved to generate an activation intermediate form. These results indicate that proMMP‐9 is processed by the cells expressing MSP‐T1, MT1‐MMP in the presence of proMMP‐2 through the cleavage at the same sites as in vitro activation by active MMP‐2. The activation efficiency by these cells was dramatically improved compared with activation by MMP‐2 in vitro.22

Figure 1.

Figure 1

Open in a new tab

ProMMP‐9 activation by cells expressing artificial receptor for proMMP‐9 (MSP‐T1), membrane type 1 (MT1)‐MMP, and MMP‐2. (a) Control plasmid or expression plasmid for MT1‐MMP (200 ng) and/or MSP‐T1 (300 ng) was transfected into 293T cells in a 24‐well microplate. Twenty‐four hours after transfection, cells were incubated in Opti‐MEM containing proMMP‐2 (100 nM) and proMMP‐9 (20 nM) for 2 h. Cell‐bound MMP‐2 and MMP‐9 were analyzed by gelatin zymography (Cells). Cells were incubated with proMMP‐2 for 2 h, and then the supernatant was incubated with proMMP‐9 (20 nM) for a further 2 h. ProMMP‐9 processing was monitored by zymography (Sup). Expression of MSP‐T1 and MT1‐MMP was detected by Western blotting using anti‐tissue inhibitor of MMP‐1 and anti‐MT1‐MMP antibodies, respectively (lower panels). (b) 293T cells transfected with expression plasmids for MSP‐T1 and MT1‐MMP as described above were incubated with proMMP‐9 (20 nM) and serially diluted proMMP‐2 for 2 h, and then cell‐bound MMP‐2 and MMP‐9 were subjected to gelatin zymography. Gels were processed and monitored by an infrared imaging system with low (top) and high sensitivity (bottom). (c) 293T cells transfected as above were incubated with proMMP‐2 (100 nM) for 1 h, and then proMMP‐9 (20 nM) was added and incubated for the indicated period. Cell‐bound MMP‐2 and MMP‐9 were detected as above. (d) 293T cells transfected with control, MSP‐T1 and/or MT1‐MMP plasmid as described above were incubated with proMMP‐2 and wild‐type proMMP‐9 or proMMP‐9 mutant lacking hemopexin‐like domain (ΔMMP‐9) for 2 h, and cell‐bound (Cell) or supernatant (Sup) MMP‐2 and MMP‐9 were detected by gelatin zymography. (e) MT1‐MMP plasmid (1 μg) was cotransfected with control or MSP‐T1 plasmid (1 μg) into 293T cells cultured in a 35‐mm dish. Twenty‐four hours after incubation, 50 nM wild‐type or mutant proMMP‐9 with an amino acid substitution at Met41 (Mut1) or Phe88 (Mut2) was incubated with cells in the presence of proMMP‐2 (50 nM) for 3 h. Supernatant was concentrated with trichloroacetic acid, and subjected to Western blotting using an antibody against MMP‐9. Expression of MSP‐T1 and MT1‐MMP was detected by Western blotting as described above (lower panels).

Expression of MSP‐T1 in HT1080 cells

HT1080 cells were stably transfected with MSP‐T1 plasmid, and cultured in collagen gel in the presence or absence of exogenously added proMMP‐2 to examine proMMP‐9 activation and degradation of collagen gel (Fig. 2). HT1080 cells migrated out of the collagen gel, but collagen degradation was not obvious in the culture of cells without exogenous proMMP‐2. In the presence of exogenously added proMMP‐2, collagen gel was degraded intensively by the cells expressing MSP‐T1, and moderately by control cells (Fig. 2b). Gelatin zymography analysis showed that endogenous proMMP‐2 was activated equally by the cells transfected with control or MSP‐T1 plasmid, but proMMP‐9 processing was detected only in the cells expressing MSP‐T1 in the absence of exogenously added proMMP‐2. Addition of proMMP‐2 to the culture produced a large amount of active MMP‐2, which resulted in significant proMMP‐9 processing in control cells and a more dramatic effect in MSP‐T1‐expressing cells (Fig. 2c). The metastatic ability of HT1080 cells expressing MSP‐T1 was analyzed by the chick embryo system. HT1080 cells expressing MSP‐T1 metastasized into liver 4‐fold more efficiently than control cells (Fig. 2d).

Figure 2.

Figure 2

Open in a new tab

Stable transfection of artificial receptor for proMMP‐9 (MSP‐T1) into HT1080 cells. (a) HT1080 cells stably transfected with control or MSP‐T1 plasmid were subjected to Western blotting using an antibody against tissue inhibitor of MMP‐1. (b) HT1080 cells transfected with control or MSP‐T1 plasmid (2 × 105 cells) were suspended in 100 μL collagen mixture. A 20‐μL drop of collagen mixture containing cells was polymerized on the bottom of a 24‐well microplate, cultured in DMEM supplemented with 5% FCS for a day, and then in Opti‐MEM with or without 50 nM proMMP‐2 for 2 days. Cells in collagen culture were photographed under a phase contrast microscope. (c) Cells cultured in collagen gel as described above were dissolved in a sample buffer, and subjected to gelatin zymography. (d) HT1080 cells transfected with control or MSP‐T1 plasmid were injected into the chorioallantoic membrane vein of chicken eggs, and cells that metastasized into liver were analyzed. *P < 0.05.

Activation of proMMP‐9 by HT1080 cells

HT1080 cells cultured in collagen gel in the presence of exogenous proMMP‐2 activated proMMP‐2 and slightly activated proMMP‐9. To induce more efficient proMMP‐2 activation in HT1080 cells, cells were treated with 2‐fold serially diluted ConA in the presence of exogenous proMMP‐2 (Fig. 3a). ProMMP‐2 activation was induced even at low ConA concentration. Treatment with ConA induced cell binding and processing of proMMP‐9 in HT1080 cells, but they required higher ConA concentration than proMMP‐2 activation.

Figure 3.

Figure 3

Open in a new tab

Induction of proMMP‐2 and proMMP‐9 activation in HT1080 cells. (a) HT1080 cells were cultured in DMEM supplemented with 5% FCS for 24 h, and then in Opti‐MEM containing proMMP‐2 (50 nM) and serially diluted concanavalin (ConA) for 12 h. Cell‐bound MMP‐2 and MMP‐9 were analyzed by gelatin zymography. (b) HT1080 cells cultured as above were incubated in Opti‐MEM containing ConA (50 μg/mL) and serially diluted proMMP2 for 12 h, and cell‐bound MMP‐2 and MMP‐9 were analyzed by gelatin zymography. (c) HT1080 cells cultured as above were incubated in Opti‐MEM containing ConA (50 μg/mL) and proMMP‐2 (50 nM) for the indicated period, and cell‐bound MMP‐2 and MMP‐9 were analyzed by gelatin zymography. (d) HT1080 cells cultured as above were incubated in Opti‐MEM containing ConA (50 ng/mL) and proMMP‐2 (50 nM) with mock (−), tissue inhibitor of tissue inhibitor of metalloproteinase (TIMP)‐1 (10 μg/mL), TIMP‐2 (10 μg/mL), or BB94 (1 μM) for 12 h. Cell‐bound and supernatant MMP‐2 and MMP‐9 were analyzed by gelatin zymography.

As shown in Figure 1, activation of proMMP‐9 by cells expressing MSP‐T1 and MT1‐MMP required a considerable concentration of proMMP‐2. Thus, requirement of MMP‐2 in proMMP‐9 activation by ConA‐treated HT1080 cells was examined (Fig. 3b). Addition of exogenous proMMP‐2 to ConA‐treated HT1080 cells induced activation of proMMP‐2 and proMMP‐9, depending on the proMMP‐2 concentration. The time course experiment showed that proMMP‐9 was processed to the intermediate activation form in 1 h and to fully active form in 2–3 h (Fig. 3c).

The effect of MMP inhibitors on proMMP‐9 activation by ConA‐treated HT1080 cells was examined (Fig. 3d). Addition of TIMP‐2 or BB94 interfered with activation of both proMMP‐2 and proMMP‐9 by ConA‐treated HT1080 cells. Cells treated with TIMP‐1 processed proMMP‐2 to the intermediate activation form, and failed to activate proMMP‐9. These results suggest that proMMP‐2 activation by MT1‐MMP is associated with proMMP‐9 processing in ConA‐treated HT1080 cells.

Analysis of proMMP‐9 activation mechanism by siRNA

To confirm the association of MT1‐MMP with proMMP‐9 processing, MT1‐MMP siRNA was transfected into HT1080 cells, and proMMP‐9 activation was examined (Fig. 4). Knockdown of MT1‐MMP expression suppressed activation of not only proMMP‐2 but also proMMP‐9 (left panel). MSP‐TIMP‐1 was shown to act as a receptor for pro‐MMP‐9, which facilitated proMMP‐9 activation by the MT1‐MMP/MMP‐2 axis. This suggested a possibility that TIMP‐1 acts as a part of the proMMP‐9 receptor and induces its processing in ConA‐treated HT1080 cells. To test it, TIMP‐1 expression was downregulated by the transfection of siRNA, and proMMP‐9 activation was examined. Knockdown of TIMP‐1 did not alter proMMP‐2 activation, but suppressed binding of proMMP‐9 to the cells and subsequent activation in ConA‐treated HT1080 cells (middle panel). Knockdown of TIMP‐2 did not alter proMMP‐9 binding but interfered with activation of both proMMP‐2 and proMMP‐9. These results suggest that TIMP‐1 serves as a part of the proMMP‐9 receptor, which induces subsequent proMMP‐9 processing. Expressed at a considerable level in HT1080 cells, ADAM10 is known to be inactivated by TIMP‐1.23 Transfection of ADAM10 siRNA did not affect proMMP‐2 activation, but suppressed cell binding and processing of proMMP‐9 (right panel).

Figure 4.

Figure 4

Open in a new tab

Suppression of proMMP‐9 processing by siRNA. HT1080 cells transfected with control RNA or siRNA for membrane type 1 (MT1)‐MMP (MT1‐MMP‐I, MT1‐MMP‐II), tissue inhibitor of tissue inhibitor of metalloproteinase (TIMP)‐1 (TIMP‐1‐I, TIMP‐1‐II), TIMP‐2 (TIMP‐2‐I, TIMP‐2‐II), or a disintegrin and metalloproteinase 10 (ADAM10) (ADAM10‐I, ADAM10‐II) were cultured in Opti‐MEM containing concanavalin A (50 ng/mL) and proMMP‐2 (50 nM) for 12 h. Cell‐bound MMP‐2 and MMP‐9 were analyzed by gelatin zymography. The expression of each protein was analyzed by Western blotting.

Contribution of MMP‐9 to gelatin degradation by HT1080 cells

In order to examine the contribution of MMP‐2 and MMP‐9 to ECM degradation by HT1080 cells, MMP‐9 expression was knocked down through siRNA transfection, and gelatin degradation by these cells was monitored (Fig. 5). Gelatin was not significantly degraded by ConA‐treated HT1080 cells in the absence of exogenous proMMP‐2. Addition of proMMP‐2 to ConA‐treated cells generated active MMP‐2 and MMP‐9, which resulted in intensive gelatin degradation. Knockdown of MMP‐9 expression in these cells considerably reduced gelatin degradation by ConA‐treated cells. These results indicate that not only MMP‐2 but also MMP‐9 activated by HT1080 cells contributes to gelatin degradation.

Figure 5.

Figure 5

Open in a new tab

Gelatin degradation by MMP‐2 and MMP‐9 in HT1080 cells. HT1080 cells transfected with control RNA or MMP‐9 siRNA (MMP‐9‐I, MMP‐9‐II) were cultured in Opti‐MEM containing concanavalin A (50 ng/mL) with or without proMMP‐2 (50 nM) for 12 h. MMP‐2 and MMP‐9 bound to cells were analyzed by gelatin zymography (left panel). These cells were incubated with fluorescence‐labeled gelatin (−) for 1 h, and gelatin degradation was monitored (right panel).

Discussion

Activation of proMMP‐2 by MT1‐MMP is a feature of the malignant phenotype of various tumors.4, 5, 6, 7 Tumor‐specific proMMP‐9 activation is restricted to a limited numbers of tissues, and the molecular mechanism of proMMP‐9 activation in vivo still remains unsolved.8, 9, 10 ProMMP‐2 and proMMP‐9 bind to TIMP‐2 and TIMP‐1 through the carboxy‐terminal domain of each molecule, respectively.1, 2 Formation of a trimolecular complex consisting of proMMP‐2/TIMP‐2/MT1‐MMP is the initial step for proMMP‐2 activation by MT1‐MMP.11, 12, 13 Previously, MSP‐T2 was constructed as an artificial receptor for proMMP‐2, which accelerated proMMP‐2 activation by MT1‐MMP.15 Assuming a similar scenario in proMMP‐9 activation process, in which binding of proMMP‐9 to the receptor containing TIMP‐1 is the initial step for activation, MSP‐T1 was created as an artificial receptor for proMMP‐9. Indeed, MSP‐T1 served as a receptor for proMMP‐9, which was effectively processed by the MT1‐MMP/MMP‐2 axis (Fig. 1). Membrane‐bound serine proteases, such as MSP and matriptase, were negative as a cell‐surface proMMP‐9 activator, when coexpressed with MSP‐T1 (data not shown). Toth et al.24 reported activation of proMMP‐9 by membrane fraction from cells expressing MMP‐2 and MT1‐MMP, however, incubation of proMMP‐9 with these cells failed to activate proMMP‐9, which was consistent with the present study. Coexpression of MSP‐T1 with MT1‐MMP induced dramatic proMMP‐9 activation in the presence of proMMP‐2. ProMMP‐9 is activated by these cells through the cleavage at the site identical with those by active MMP‐2 in vitro.22 These results suggest a proMMP‐9 activation process in which proMMP‐9 is anchored to the cell surface through a TIMP‐1‐containing receptor and is subsequently processed by MMP‐2 activated and anchored on the cell surface by MT1‐MMP. HT1080 cells in collagen gel did not show detectable proMMP‐9 processing, but addition of proMMP‐2 induced significant proMMP‐9 processing. Treatment of HT1080 cells with ConA caused more efficient proMMP‐2 activation than collagen culture, and induced cell binding and processing of proMMP‐9 in the presence of exogenously added proMMP‐2 (Fig. 3). Knockdown of MT1‐MMP or TIMP‐2 expression, which is essential for proMMP‐2 activation, suppressed processing of not only proMMP‐2 but also proMMP‐9. Knockdown of TIMP‐1 expression downregulated cell binding and processing of proMMP‐9, but did not affect proMMP‐2 activation. These results indicate that TIMP‐1 is involved in cell‐binding of proMMP‐9, which is subsequently processed by MMP‐2 activated by MT1‐MMP. ADAM10 is known to be inactivated by TIMP‐1, suggesting the possibility that ADAM10 acts as a cell‐surface binding molecule of TIMP‐1.23 Knockdown of ADAM10 expression by siRNA reduced both binding and processing of proMMP‐9. These results may suggest that ADAM10 is one of the receptors for TIMP‐1 and that the ADAM10/TIMP‐1 complex acts as a receptor for proMMP‐9, which is subsequently processed by the MT1‐MMP/MMP‐2 axis. However, reconstitution of proMMP‐9 receptor by coexpressing ADAM10 and TIMP‐1 was not successful (data not shown). Actually, ADAM10 and TIMP‐1 protein expression was not significantly affected by ConA treatment of HT1080 cells (data not shown). The cell surface event induced by ConA treatment seems to be complicated, and the mechanism by which ADAM10 participates in proMMP‐9 processing still remains unsolved. Several TIMP‐1 binding molecules such as CD44, CD63, and integrins were reported, which may function as proMMP‐9 receptors by forming complexes with TIMP‐1.23 Association of CD44 and MT1‐MMP with proMMP‐9 activation in osteoclasts and formation of the proMMP‐9/CD44/TIMP‐1 complex were reported, although the molecular mechanism and physiological significance of them in relation to proMMP‐9 processing still remain to be solved.20, 25

Expression of MSP‐T1 in HT1080 cells stimulated proMMP‐9 activation by the MT1‐MMP/MMP‐2 axis, which in turn enhanced metastatic ability in vivo and collagen degradation in vitro. These results showed for the first time that activation of proMMP‐9 stimulates tumor metastasis. Addition of proMMP‐2 to ConA‐treated HT1080 cells generated active MMP‐2, which consequently processed endogenous proMMP‐9. Both MMP‐2 and MMP‐9 cleave various substrates, such as ECM molecules, membrane proteins, cytokines, and growth factors that regulate key signaling pathways in cell growth, migration, invasion, and angiogenesis.1, 2 Membrane‐bound MMP‐9 thus activated by the MT1‐MMP/MMP‐2 axis may have advantage over pericellular proteolysis. Participation of MMP‐9 to coordinate action of MT1‐MMP and MMP‐2 further facilitates pericellular proteolysis for tumor invasion and metastasis. ProMMP‐2 is often abundantly expressed by stromal cells located adjacent to malignant epithelial cells, and functions as a mediator of tumor–stroma interaction by being activated by tumor‐specific MT1‐MMP, which may subsequently activate proMMP‐9. In conclusion, proMMP‐9, which is anchored though a receptor complex containing TIMP‐1, is activated by the MT1‐MMP/MMP‐2 axis. Thus activated, MMP‐9 contributes to pericellular proteolysis for tumor invasion and metastasis in collaboration with MT1‐MMP and MMP‐2.

Disclosure Statement

The authors have no conflict of interest.

Abbreviations

ADAM10

a disintegrin and metalloproteinase 10

ConA

concanavalin A

IR

infrared

MSP

mosaic serine protease

MSP‐T2

TIMP‐2 fused with type II transmembrane MSP

MT

membrane type

TIMP

tissue inhibitor of metalloproteinase

Acknowledgments

This work was supported by the grant from the Japan Society for the Promotion of Science (Kakenhi 25460382).

Cancer Sci 108 (2017) 347–353

Funding Information

Japan Society for the Promotion of Science (Kakenhi 25460382).

References

  • 1. Mook OR, Frederiks WM, Van Noorden CJ. The role of gelatinases in colorectal cancer progression and metastasis. Biochim Biophys Acta 2004; 1705: 69–89. [DOI] [PubMed] [Google Scholar]
  • 2. Björklund M, Koivunen E. Gelatinase‐mediated migration and invasion of cancer cells. Biochim Biophys Acta 2005; 1755: 37–69. [DOI] [PubMed] [Google Scholar]
  • 3. Sato H, Takino T, Okada Y et al A matrix metalloproteinase expressed on the surface of invasive tumour cells. Nature 1994; 370: 61–5. [DOI] [PubMed] [Google Scholar]
  • 4. Seiki M, Yana I. Roles of pericellular proteolysis by membrane type‐1 matrix metalloproteinase in cancer invasion and angiogenesis. Cancer Sci 2003; 94: 569–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Sato H, Takino T, Miyamori H. Roles of membrane‐type matrix metalloproteinase‐1 in tumor invasion and metastasis. Cancer Sci 2005; 96: 212–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Egeblad M, Werb Z. New functions for the matrix metalloproteinases in cancer progression. Nat Rev Cancer 2002; 2: 161–74. [DOI] [PubMed] [Google Scholar]
  • 7. Overall CM, McQuibban GA, Clark‐Lewis I. Discovery of chemokine substrates for matrix metalloproteinases by exosite scanning: a new tool for degradomics. Biol Chem 2002; 383: 1059–66. [DOI] [PubMed] [Google Scholar]
  • 8. Zeng ZS, Huang Y, Cohen AM, Guillem JG. Loss of basement membrane type IV collagen is associated with increased expression of metalloproteinases 2 and 9 (MMP‐2 and MMP‐9) during human colorectal tumorigenesis. Carcinogenesis 1999; 20: 749–55. [DOI] [PubMed] [Google Scholar]
  • 9. Roeb E, Dietrich CG, Winograd R et al Activity and cellular origin of gelatinases in patients with colon and rectal carcinoma differential activity of matrix metalloproteinase‐9. Cancer 2001; 92: 2680–91. [DOI] [PubMed] [Google Scholar]
  • 10. Pucci‐Minafra I, Minafra S, La Rocca G et al Zymographic analysis of circulating and tissue forms of colon carcinoma gelatinase A (MMP‐2) and B (MMP‐9) separated by mono‐ and two‐dimensional electrophoresis. Matrix Biol 2001; 20: 419–27. [DOI] [PubMed] [Google Scholar]
  • 11. Strongin AY, Collier I, Bannikov G et al Mechanism of cell surface activation of 72‐kDa type IV collagenase. Isolation of the activated form of the membrane metalloprotease. J Biol Chem 1995; 270: 5331–8. [DOI] [PubMed] [Google Scholar]
  • 12. Butler GS, Butler MJ, Atkinson SJ et al The TIMP2 membrane type 1 metalloproteinase ‘‘receptor’’ regulates the concentration and efficient activation of progelatinase A. A kinetic study. J Biol Chem 1998; 273: 871–80. [DOI] [PubMed] [Google Scholar]
  • 13. Kinoshita T, Sato H, Okada A et al TIMP‐2 promotes activation of progelatinase A by membrane‐type 1 matrix metalloproteinase immobilized on agarose beads. J Biol Chem 1998; 273: 16098–103. [DOI] [PubMed] [Google Scholar]
  • 14. Kudo T, Takino T, Miyamori H et al Substrate choice of membrane‐type 1 matrix metalloproteinase is dictated by tissue inhibitor of metalloproteinase‐2 levels. Cancer Sci 2007; 98: 563–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Nishida Y, Miyamori H, Thompson EW, Takino T, Endo Y, Sato H. Activation of matrix metalloproteinase‐2 (MMP‐2) by membrane type‐1 matrix metalloproteinase through an artificial receptor for proMMP‐2 generates active MMP‐2. Cancer Res 2008; 68: 9096–104. [DOI] [PubMed] [Google Scholar]
  • 16. Takino T, Koshikawa N, Miyamori H et al Cleavage of metastasis suppressor gene product KiSS‐1 protein/metastin by matrix metalloproteinases. Oncogene 2003; 22: 4617–26. [DOI] [PubMed] [Google Scholar]
  • 17. Endo K, Takino T, Miyamori H et al Cleavage of syndecan‐1 by membrane type matrix metalloproteinase‐1 stimulates cell migration. J Biol Chem 2003; 278: 40764–70. [DOI] [PubMed] [Google Scholar]
  • 18. Nakada M, Yamada A, Takino T et al Suppression of membrane‐type 1 matrix metalloproteinase (MMP)‐mediated MMP‐2 activation and tumor invasion bytestican 3 and its splicing variant gene product, N‐Tes. Cancer Res 2001; 61: 8896–902. [PubMed] [Google Scholar]
  • 19. Li Y, Aoki T, Mori Y et al Cleavage of lumican by membrane‐type matrix metalloproteinase‐1 abrogates this proteoglycan‐mediated suppression of tumor cell colony formation in soft agar. Cancer Res 2004; 64: 7058–64. [DOI] [PubMed] [Google Scholar]
  • 20. Chellaiah MA, Ma T. Membrane localization of membrane type matrix metalloproteinase by CD44 regulates the activation of pro‐matrix metalloprotease 9 in osteoclastas. Biomed Res Int 2013; 2013: 302392. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Sato H, Kinoshita T, Takino T et al Activation of a recombinant membrane type 1‐matrix metalloproteinase (MT1‐MMP) by furin and its interaction with tissue inhibitor of metalloproteinases (TIMP)‐2. FEBS Lett 1996; 393: 101–4. [DOI] [PubMed] [Google Scholar]
  • 22. Fridman R, Toth M, Pena D, Mobashery S. Activation of progelatinase B (MMP‐9) by gelatinase A (MMP‐2). Cancer Res 1995; 55: 2548–55. [PubMed] [Google Scholar]
  • 23. Murphy G. Tissue inhibitors of metalloproteinases. Genome Biol 2011; 12: 233–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Toth M, Chvyrkova I, Bernardo MM et al Pro‐MMP‐9 activation by the MT1‐MMP/MMP‐2 axis and MMP‐3: role of TIMP‐2 and plasma membranes. Biochem Biophys Res Com 2003; 308: 386–95. [DOI] [PubMed] [Google Scholar]
  • 25. Lambert E, Bridoux L, Devy J et al TIMP‐1 binding to proMMP‐9/CD44 complex localized at the cell surface promotes erythroid cell survival. Int J Biochem Cell Biol 2009; 41: 1102–15. [DOI] [PubMed] [Google Scholar]

Articles from Cancer Science are provided here courtesy of Wiley

RESOURCES