(Received for publication, April 30, 1996, and in revised form, April 9, 1997)
From the Departments of Medical Chemistry and
Periodontology, the || Department of Clinical Chemistry, and
the
Division of Biochemistry, University of Helsinki, FIN-00014
Helsinki, Finland, the ¶ Departments of Oral Surgery and Pathology
and the
Department of Biochemistry and
Biocenter, University of Oulu, Oulu, FIN-99020 Finland, the
** Department of Anatomy and IVth Department of Medicine and the
§§ Departments of Oral Surgery and Surgery,
University of Helsinki and Helsinki University Central Hospital,
Helsinki, Finland, and the ¶¶ Department of Biochemistry,
University of Bielefeld, Bielefeld,
Postfach 100131, 33501 Bielefeld, Germany
Increased production of proteinases, such as matrix metalloproteinases (MMPs), is a characteristic feature of malignant tumors. Some human cancers and cell lines derived from them also express trypsinogen, but the function of the extrapancreatic trypsin has remained unclear. In this study we cloned and sequenced trypsinogen-2 cDNA from human COLO 205 colon carcinoma cells and characterized the ability of the enzyme to activate latent human type IV procollagenases (proMMP-2 and proMMP-9). As shown by cloning and N-terminal amino acid sequencing, the amino acid sequence of tumor-associated trypsin-2 is identical to that of pancreatic trypsin-2. We found that both pancreatic trypsin-2 and tumor cell-derived trypsin-2 are efficient activators of proMMP-9 and are capable of activating proMMP-9 at a molar ratio of 1:1000, the lowest reported so far. Human trypsin-2 was a more efficient activator than widely used bovine trypsin and converted the 92-kDa proMMP-9 to a single 77-kDa product that was not fragmented further. The single peptide bond cleaved by trypsin-2 in proMMP-9 was Arg87-Phe88. The generation of the 77-kDa species coincided with the increase in specific activity of MMP-9. In contrast, trypsin-2 only partially activated proMMP-2. Trypsin-2 cleaved the Arg99-Lys100 peptide bond of proMMP-2 generating 62-65-kDa MMP-2 species. Trypsin-2-induced proMMP-2 and -9 conversions were inhibited by tumor-associated trypsin inhibitor added either prior to or during activation indicating that proMMPs were not activated autocatalytically. Trypsin-2 also activated proMMPs associated with tissue inhibitor of matrix metalloproteinases, the complexes of which are thought to be the major MMP forms in vivo. The ability of human tumor cell-derived trypsin-2 to activate latent MMPs suggests a role for trypsin-2 in initiating the proteinase cascade that mediates tumor invasion and metastasis formation.
Enhanced proteolytic activity is associated with malignancies and tumor invasion (1-10). Increased levels of proteolytic enzymes are found in malignant tissues and extracellular fluids of tumor-bearing animals and in conditioned medium of cultured cells transformed by a variety of oncogenic agents, including tumor viruses, radiation, chemical carcinogens, and oncogenes (1-10).
Proteinases are thought to form cascades that degrade tissue barriers and thus promote tumor invasion (1, 3, 4, 6-8). The matrix metalloproteinases (MMPs)1 comprise a gene family of proteinases that can degrade the constituents of extracellular matrix and basement membrane (1, 9-11). Increased expression and secretion of these enzymes are associated with the metastatic capacity of tumor cells (1, 3, 4, 8, 10). The MMPs are secreted as inactive zymogens in which the Cys73 residue of the propeptide is coordinated to the to the catalytic Zn2+ in the active site (12, 13). Activation is thought to change the propeptide structure by interrupting the Cys-zinc bond (5, 8, 10, 12). This can be induced by various means, including proteinases, organomercurials, chaotropic agents, and oxidants (12-16). The activation process is a key event in the extracellular regulation of the activity of MMPs (4, 8, 10).
In addition to MMPs, the proteolytic cascades associated with tumor invasion comprise serine proteinases plasminogen activators and plasmin (4, 6). We and others have found that a serine proteinase very similar or identical to pancreatic trypsinogen is expressed in human cancers such as ovarian carcinoma (7, 17-19) and cholangiocarcinoma (20-23). Production of extrapancreatic trypsinogen isoenzymes has been observed in many established tumor cell lines derived from various human tumors (17, 24-27). Trypsinogen expression by cancer cells may be associated with elevated levels of the specific inhibitor tumor-associated trypsin inhibitor (TATI) observed in serum and urine of cancer patients (7, 26). Two tumor-associated trypsinogens, trypsinogen-1 and trypsinogen-2, have been characterized, of which trypsin-2 is the major isoenzyme (17, 24, 27). Trypsinogen-1 corresponds to cationic and trypsinogen-2 to anionic trypsinogen.
Expression of trypsin by cancer cells may have a significant effect on their ability to attach and migrate on substratum and this could stimulate invasiveness. Indeed, trypsinogen levels in cyst fluid of ovarian tumors are higher in malignant than in benign cases (27), and the trypsinogen mRNA levels are high in particular in advanced stages of ovarian carcinoma (20-22, 26). Trypsin may be expected to promote tumor spread if it initiates a protease cascade by activating proMMPs. We have therefore studied the ability of trypsin-2 secreted by COLO 205 cells carcinoma cells to activate proMMP-9 and proMMP-2, which are type IV collagenases thought to play a key role in cell migration (1, 3, 28). We have also sequenced the cDNA for the COLO 205 cell trypsinogen-2 and found it to be identical with that of pancreatic trypsinogen-2. Our results show that human both tumor cell-derived and pancreatic trypsin-2 are highly potent activators of proMMP-9 and activate proMMPs at the lowest concentrations reported so far.
Chromatographic media, unless otherwise stated,
were purchased from Pharmacia Biotech Inc., Uppsala, Sweden.
p-Aminophenyl mercuric acetate (APMA), bovine serum albumin,
TPCK-treated bovine pancreatic trypsin, gelatin, and
phenylmethylsulfonyl fluoride (PMSF) were from Sigma. S-2222 was from
Kabi-Vitrum, Stockholm, Sweden. 125I-Labeled gelatin
(heat-denatured type I collagen) in solution was prepared as described
(29). Human neutrophil-associated lipocalin (NGAL) and rabbit
polyclonal antibodies to NGAL and to human neutrophil (PMN) MMP-9 were
kindly provided by Drs. Lars Kjeldsen and Nils Borregaard, Department
of Hematology, Granulocyte Research Laboratory, Rigshospitalet,
Copenhagen, Denmark. Human recombinant tissue inhibitors of matrix
metalloproteinases (TIMP-1 and -2) were from Oncogene Science,
Cambridge, MA. Rabbit polyclonal antibodies to human MMP-1, MMP-2,
TIMP-1, and TIMP-2 were kindly provided by Dr Henning Birkedal-Hansen
and Dr. William G. Stetler-Stevenson. Rabbit polyclonal antibody to
human neutrophil collagenase (MMP-8) was kindly provided by Dr. Jurgen
Michaelis, Department of Pathology, Christchurch Medical School,
Christchurch, New Zealand. Polyclonal and monoclonal antibodies to
human neutrophil elastase and cathepsin G as well as for the serpins
1-antitrypsin and
1-antichymotrypsin were
from Dakopatts A/S, Glostrup, Denmark.
Dinitrophenyl-Pro-Gly-Ile-Ala-Gly-Glu-D-Arg-OH was from
Bachem, Bubendorf, Switzerland. Native soluble type I collagen was
pepsin-extracted and further purified from bovine skin by selective
salt precipitations at acidic and neutral pH. The purity of type I
collagen was analyzed by CNBr cleavage and SDS-PAGE (30).
Trypsinogen-2 was purified from serum-free conditioned media of COLO 205 colon carcinoma cells or cyst fluid of ovarian tumors as described previously (17, 24). Pancreatic trypsinogen-2 was purified from urine of pancreatitis patients using ion exchange and immunoaffinity chromatography followed by reverse phase HPLC essentially as described (17). Before the final HPLC purification purified trypsinogen was autoactivated by incubating at 37 °C overnight in 50 mM Tris-HCl, 0.2 M NaCl, 20 mM CaCl2, pH 7.8 (17). Trypsin-2 activity was determined with the chromogenic substrate S-2222 as described (17). No contaminating proteases were observed in the purified autoactivated trypsin-2 preparations. TATI was purified from urine of cancer or pancreatitis patients as described (27, 31).
Cloning and Sequencing of cDNA for Trypsinogen-2 from COLO 205 CellsTotal RNA was purified with the acid phenol guanidium
thiocyanate method of Chomzcynski et al. (32) from
approximately 108 COLO 205 cells. Poly(A) RNA was purified
by oligo(dT) affinity chromatography (Pharmacia) according to the
instructions of the manufacturer. Oligonucleotides were synthesized
with a Pharmacia Biotech Inc. gene assembler plus oligosynthesizer
using phosphoramidite chemistry (Pharmacia). A biotinylated
phosphoramidite (Clontech, Palo Alto, CA) was
introduced into the 5 end of the oligonucleotides. Reverse
transcription of poly(A) RNA was performed with avian myeloblastosis
virus reverse transcriptase (Promega) as described (33). The antisense
primers were 5
GCT GGA TCC GCT ATT GGC AGC TAT GGT GTT 3
or 5
ATA GGA TCC GCT GTT GGC AGC TAT GGT GTC 3
corresponding to the sequence 727-747 in pancreatic trypsinogen-1 and
-2 cDNAs, respectively (sequence identification number M27602 in
GenBankTM). The underlined sequences represent a BamHI
restriction site. PCR amplification was performed with one
nonbiotinylated primer and one biotinylated primer. The reverse primers
were the same as used for reverse transcription of RNA and the forward primer was 5
GAT GGA TCC TGC TGC CCC CTT TGA TGA 3
corresponding to the sequence 48-68 of pancreatic trypsinogen-1 and
-2. PCR mixtures (100 µl) for amplification of cDNA contained 50 mM KCl, 10 mM Tris-HCl, pH 8.3, 1.5 mM MgCl2, 0.001% gelatin (w/v), 200 µmol of
each dNTP, 2.5 units of Taq polymerase and 5 pmol of each primer. Thirty cycles of amplification were performed at 94 °C for
40 s, 55 °C for 1 min, and 72 °C for 3 min. PCR products
were directly sequenced as described (34, 35). Briefly, the
biotinylated PCR product was bound to streptavidin coated magnetic
particles (Dynabeads, Dynal, Norway), and the strands were dissociated
with 0.1 M NaOH. The bound biotinylated strand was
sequenced by the dideoxy chain termination method (36) with Sequenase
2.0 kit (U. S. Biochemical Corp.) according to the instructions of the manufacturer.
Neutrophil progelatinase B (92-kDa MMP-9) was purified essentially as described by Hibbs et al. (37). Briefly, neutrophils were prepared from buffy coats (37, 38) and resuspended in Krebs-Ringer phosphate buffer: 130 mM NaCl, 5 mM KCl, 1.27 mM MgSO4, 0.95 mM CaCl2, 5 mM glucose, 10 mM NaH2PO4/Na2HP04, pH 7.4. The neutrophils were stimulated with 2 µg/ml of phorbol myristic acetate (Sigma) for 15 min and pelleted by centrifugation. The resulting supernatant was collected and subjected to ion exchange chromatography on a DE-52 column (Whatman) followed by affinity chromatography on heat-denatured type I collagen (gelatin, Sigma St Louis, Mo, USA) coupled to Sepharose-4B (38). MMP-2 and MMP-9 were also purified by the modified protocol described by Chen et al. (39) from serum-free culture medium of human gingival fibroblasts and human gingival keratinocytes, respectively, using affinity chromatography on gelatin-Sepharose and gel filtration (40, 41). The purity of isolated proMMP-9 and proMMP-2 was estimated by SDS-PAGE with silver staining, zymography, Western blotting, and N-terminal sequence analysis (42, 43). Human recombinant mutant 72 kDa proMMP-2, in which Glu375 was substituted with Ala, was prepared and purified as described by Bergmann et al. (44). Purified human 92-kDa proMMP-9 did not contain neutrophil collagenase (MMP-8) as evidenced by lack of collagenase activity. Lack of contaminating proteinases and TIMP-1 and TIMP-2 was further demonstrated by zymography using gelatin and casein as substrate as well as by immunoblotting using specific antisera to MMP-8, elastase, cathepsin G, TIMPs-1 and -2, and serpins. The purity of the MMP-9 preparation was also confirmed by N-terminal analysis showing the expected sequence (42, 43, 45). The purified proMMP-9 was latent and displayed no enzymatic activity against natural (gelatin) or a synthetic substrate. The proMMP-9 remained inactive for 24 h at 37 °C. Recombinant human proMMP-9 was kindly provided by Dr. Karl Tryggvason, MBB, Matrix Biology, Karolinska Institutet, Stockholm, Sweden.
Activation of ProMMP-9 and ProMMP-2 by Trypsin-2Purified
proMMP-9 (11 µM) and proMMP-2 (10-14 µM)
were treated with trypsin-2 or bovine pancreatic TPCK-trypsin (10 nM to 2 µM) for indicated time periods at
37 °C. The reactions were terminated by heating at 100 °C or by
adding a 4-10-fold molar excess of TATI. In control experiments,
trypsin-2 was pretreated with a 10-fold molar excess of TATI for 60 min. Treatments with bovine pancreatic TPCK-trypsin were stopped by
adding 1 mM PMSF. Activation of proMMPs was analyzed by
SDS-PAGE, zymography, peptide substrate assay, -casein degradation
assay, and immunoblotting. Coomassie Brilliant Blue-stained gels were
quantitated by densitometry with a LKB Ultroscan laser densitometer
(model 2202) (30), and the results were expressed as percentage of
active enzyme generated.
Gelatinase activity was assayed by zymography in 1.5-mm-thick 7.5-10% SDS-polyacrylamide gels impregnated with 1 mg/ml gelatin, which had been labeled fluorescently with 2-methoxy-2,4-diphenyl-3-[2H]furanone (Fluka, Ronkonkoma, NY) (O'Grady et al. (46)). The degradation of gelatin was visualized under long wave UV light. Zymography was also conducted using unlabeled gelatin as a substrate. Samples were electrophoresed without reduction. The gels were then incubated in 50 mM Tris-HCl buffer, pH 7.8, containing 150 mM NaCl, 5 mM CaCl2, 1 µM ZnCl2, for 10 h at 37 °C and stained with Coomassie Brilliant Blue (40, 41).
Gelatin Degradation AssayThe degradation of gelatin was also assayed using 125I-labeled gelatin as a substrate (29). Samples to be tested were incubated with soluble 125I-labeled gelatin (1.5 µM) for 1 h at 37 °C. Undegraded gelatin was precipitated with 20% trichloroacetic acid. The radioactivity in the supernatants and precipitates were counted. Radioactivity in the supernatant reflected gelatinase activity (41).
Casein Degradation AssayThe degradation of -casein by
MMP-9 was determined essentially as described (30, 47). ProMMP-9 was
activated with trypsin-2 for the time periods indicated at 37 °C,
after which TATI and 21-kDa
-casein substrate (52 µM)
were added. Following caseinolysis for 60 min at 30 °C, the proteins
were resolved by SDS-PAGE and the 21-kDa
-casein, fragments of
-casein, and the 77-kDa active MMP-9 generated were scanned from the
gels as above. Caseinolysis induced by MMP-9 activated with 1 mM APMA was regarded as 100%.
Gelatinase activity was measured with the synthetic octapeptide dinitrophenyl-Pro-Gln-Gly-Ile-Ala-Gly-Gln-D-Arg-OH according to Masui et al. (48). One unit of peptidolytic activity corresponded to the degradation of 1 µmol of substrate/min at 37 °C. Control experiments showed that under the assay conditions trypsin-2 did not degrade the synthetic MMP substrate.
ImmunoblottingSamples were run on 7-10% SDS-polyacrylamide gels and transferred to nitrocellulose membranes (Schleicher & Schüll, Dassel, Germany). Nonspecific binding was blocked with 5% non-fat dry milk (Difco) for 90 min at 37 °C. The membranes were incubated with rabbit polyclonal antibodies (1:1000 dilution) against MMPs for 3 h at 37 °C and followed by peroxidase-conjugated goat anti-rabbit immunoglobulins (1:200-dilution; DAKO A/S, Glostrup, Denmark) for 1 h at 22 °C. After washing, the blot was developed with a solution of 60 mg/ml diaminobenzidine tetrahydrochloride in 50 mM Tris-HCl, pH 8.0, and 0.003% H202 (30).
Protein Determination and N-terminal SequencingProtein concentrations were determined by the Bradford method (49) or by measuring absorbance at 280 nm. Concentration of purified trypsinogen was estimated from its specific absorbance as described (17). In some experiments active site titration of human trypsin-2 and bovine TPCK-trypsin were performed in which S-2222 was used as substrate and TATI as a titrant. N-terminal amino acid sequences of the activated MMPs were determined from proteins electroblotted onto polyvinylidene difluoride membranes using a modified Applied Biosystems 1 477 A/120 A Sequencer (50). The N-terminal amino acid sequence of purified trypsin-2 from COLO 205 cells was obtained by automatic sequence determination using a microsequencer (model 810, Knauer, Berlin, Germany) with on-line amino acid phenylthiohydantoin separation (51).
To confirm the similarity of tumor cell-derived trypsin to pancreatic trypsin, we used PCR with trypsin-1 and trypsin-2 primers to amplify cDNA obtained from COLO 205 colon carcinoma cells. This resulted in a product of 718 base pairs, which corresponds to the size of trypsinogen. The PCR product was sequenced in its entity and the sequence showed identity with that of human pancreatic trypsinogen-2 in GenBankTM (sequence identification number M27602) with the exception for one conservative base substitution (G to A) at the base position 276. We also determined the N-terminal amino acid sequence of autoactivated trypsin-2 purified from the culture medium of the COLO 205 cells. The N-terminal sequence obtained, IVGGYIXEENSVPYQVSLNSGY, confirms the identity of the COLO 205 cell trypsin-2 to pancreatic trypsin-2 (52) and ovarian tumor-associated trypsin-2 (17).
Tumor Cell-derived and Pancreatic Trypsin-2 Activate ProMMP-9Trypsin-2 from COLO 205 medium was found to rapidly
activate proMMP-9. We used a high (1:5.5) molar ratio of trypsin-2 to proMMP-9 to visualize both trypsin-2 and proMMP-9 on gelatin zymography (Fig. 1). Trypsin-2 migrated as 27- and
28-kDa bands on zymography (Fig. 1A, lanes 2-5, 7, and
8). Under these conditions proMMP-9 was converted to a
single active product of 77 kDa within 15-30 min as shown by
zymography (Fig. 1A, lanes 2 and 3) and by
immunoblotting with anti-MMP-9 antibodies (Fig. 1B, lanes 3 and 4). The conversion of the 92 kDa proMMP-9 to the 77-kDa
MMP-9 species occurred without detectable formation of intermediates,
and no further fragmentation of the 77-kDa form was observed after
incubation for 60 min (Fig. 1A, lane 4 and Fig. 1B,
lane 5) or even 24 h with 60 ng of trypsin-2 (not shown). The
activation was completely inhibited by pretreatment of trypsin-2 with a
10-fold molar excess of TATI (Fig. 1A, lane 7). A similar
activation was obtained with trypsin-2 purified from cyst fluid of
ovarian tumors or with pancreatic trypsin-2 purified from urine from
pancreatitis patients (not shown).
The action of human trypsin-2 was compared with that of bovine trypsin and APMA, which have been widely used as activators of MMPs (5, 43-45). TPCK-treated bovine trypsin, at a molar ratio 1:5.5, also activated proMMP-9 to about 77-kDa major product (Fig. 1A, lanes 10-12). However, a notable difference to human trypsin-2 was that bovine trypsin induced further fragmentation of proMMP-9 to multiple lower molecular weight products (Fig. 1A, lanes 10-12). Similar bovine trypsin-mediated fragmentation of proMMP-9 has been reported previously (53-56). Chemical activation with APMA also induced multiple fragments of proMMP-9 (Fig. 1A, lane 13) in agreement with previous findings (43-45).
When trypsin-2 was used at a molar ratio of 1:1100 to proMMP-9, the
conversion of the 92-kDa proMMP-9 to the 77-kDa MMP-9 species was
completed in 3 h. No major intermediate activation species or
other fragments of MMP-9 were detected (Figs.
2A). Scanning of the SDS gels
stained with Coomassie Blue indicated that the proportion of the 77-kDa
product steadily increased, while the proportion of the 92-kDa proform
decreased (Fig. 2E). The trypsin-2-induced conversion was
halted by adding a 10-fold molar excess of TATI during activation
(Figs. 2, B and F). Recombinant human proMMP-9
and 92-kDa proMMP-9 purified from human gingival keratinocytes were
similarly activated by trypsin-2 (not shown). Sequencing the N terminus
of the 77-kDa MMP-9 species showed that the peptide bond cleaved by
trypsin-2 was Arg87-Phe88.
To further demonstrate that the 77-kDa species is the active form of
MMP-9, we monitored the degradation of -casein, which is a sensitive
substrate for MMP-9 (47).
-Casein was added after activation of
proMMP-9 with trypsin-2 was completed for various time intervals, and
the remaining trypsin was inactivated. The bands corresponding to
degradation products of the 21-kDa
-casein and the 77-kDa MMP-9
generated were scanned from SDS gels. This indicated that the increase
in caseinolytic activity of MMP-9 correlates with the generation of the
77-kDa species (Fig. 3). With a molar
ratio of 1:500 of trypsin-2 to proMMP-9, the activation was completed
within 30 min, after which both the conversion to the 77-kDa form and
the caseinolysis reached plateau values. Activation by 1 mM
APMA, regarded as 100% increase in specific activity, reached a
plateau within 5-10 min (not shown).
To show that the differences in the abilities of human trypsin-2 and
bovine trypsin to activate proMMP-9 are not due to different amounts of
trypsin activity applied, we studied proMMP-9 activation with
active-site titrated trypsins. Comparison of the caseinolytic activity
of MMP-9 obtained with 10 ng of each trypsin indicated that human
trypsin-2 is a more efficient activator than bovine trypsin (Fig.
4). Trypsin-2 induced in 20 min the same
maximal caseinolytic activity (lanes 3-5) that is obtained
with 1 mM APMA (lanes 9 and 10). At a
molar ratio of 1:1000 used, bovine trypsin induced only a partial
activation of proMMP-9 (lanes 6-8). ProMMP-9 without any
activation degraded -casein to some extent (Fig. 3 and Fig. 4,
lane 2) as found previously (47). Trypsin-2 that was
inactivated with TATI did not contribute to the degradation of
-casein (lane 11).
As proMMP-9 mostly occurs in complex with TIMP-1 in tissues, we
examined in a separate set of experiments the ability of trypsin-2 to
activate complexed proMMP-9. In the absence of TIMP-1, trypsin-2 induced more than 1000-fold activity enhancement as determined with the
synthetic peptide substrate for gelatinases (Fig.
5A). The presence of TIMP-1
caused an partial inhibition on the activation but did not totally
prevent it. Complexation with another proMMP-9-binding protein, NGAL
(39), did not affect the activation (not shown). As shown by SDS-gel
electrophoresis, we also found that TIMP-1 and NGAL were degraded by
trypsin-2, but the fragmented TIMP-1 did not completely lose its
ability to inhibit MMP-9 (not shown).
ProMMP-2 Is Partially Activated by Trypsin-2
Trypsin-2 preparations induced a moderate increase in enzyme activity of proMMP-2. At a 1:1100 molar ratio of trypsin-2 to fibroblast proMMP-2, the activity increase detected with the synthetic gelatinase substrate assay was 1.4-3.5-fold (Fig. 5B) and 1.6-2.8-fold using a gelatin-degradation assay (not shown). 1 mM APMA activated proMMP-2 about 4.6-fold (substrate degradation 460 µmol/min) and 6.5-fold (substrate degradation was 650 µmol/min) after 1- and 2-h treatments, respectively (not shown).
To investigate the reason for the relative poor activation of proMMP-2 by trypsin-2, we studied a Glu375-Ala active-site mutant of proMMP-2 that cannot undergo autodegradation. We found that the 72-kDa mutant proMMP-2 was only partially converted by trypsin-2 to 62-65-kDa species at a molar ratio of 1:1000 of trypsin-2 to proMMP-2 (Fig. 2, C and D). After 1-h incubation, only 5% of the proMMP-2 was activated (Fig. 2, C and G) in comparison with the 50% of the proMMP-9 processed for the same time (Fig. 2, A and E). Addition of TATI stopped the conversion of proMMP-2 (Fig. 2, D and H). N-terminal sequencing of the 65-kDa product indicated cleavage of the Arg99-Lys100 peptide bond. After a prolonged incubation for 5 h, further fragmentation of MMP-2 to 40-kDa species was observed (not shown). The N-terminal amino acid sequence of the 40-kDa product was the same as that of the 65-kDa form, indicating additional trypsin-2-mediated processing in the C-terminal region.
Complexation of proMMP-2 with TIMP-2 did not prevent the trypsin-2-mediated activation. SDS-gel electrophoresis showed TIMP-2 was degraded by trypsin-2 to a 17-kDa form, but did not entirely lose the inhibitor capacity for MMP-2 (not shown).
Our results show that human trypsin-2 is a highly efficient activator of the human gelatinases/type IV collagenases, especially of proMMP-9. ProMMP-2 is also activated but the activation is less efficient, and the resulting activity is fairly low. Trypsin-2 derived from tumor cells showed similar activity as pancreatic trypsin-2, and sequencing of the cDNA of tumor cell-derived trypsin-2 confirmed that the amino acid sequence is identical to that of pancreatic trypsin-2 (52). These results suggest that earlier observed differences in isoelectric points, and activities against synthetic substrates between tumor derived and pancreatic enzymes may be caused by posttranslational modification (17, 57). Pancreatic trypsinogen is sulfated (58), and the machinery for sulfation present in gastrointestinal tissues may be lacking from tumors. The ability of tumor cell-derived trypsin-2 to activate proMMP-9 could explain our previous results showing that TATI and antibodies to human trypsin inhibit the degradation of subendothelial extracellular matrix by four different human tumor cell lines (COLO 205 colon carcinoma, K-562 erythroleukemia, CAPAN-1 pancreatic carcinoma, and HT 1080 fibrosarcoma cells) under serum-free conditions (24). Furthermore, monoclonal anti-MMP-9 antibodies have been shown to inhibit the Matrigel invasion of HT 1080 fibrosarcoma cells, which produce both MMP-9 (28, 59) and trypsin-2 (17, 24). These results suggest the existence of a proteolytic cascade(s) in human tumors involving type IV collagenases/gelatinases (MMP-9 and -2) and trypsin-2 (7).
The factors inducing expression of trypsinogen-2 in cancer cells are not known. Whether trypsinogen expression seen in some cancer cells (17, 24, 25, 27) is a remnant of their past and reflects the site of origin of the cells, or whether trypsinogen expression is in general associated with malignant transformation of cells remains to be determined. Our recent study showing that pancreactectomized patients have measurable levels of trypsinogen-2 suggests that trypsin-2 may have an important function outside the pancreas and the digestive tract (57). Trypsin has also recently been detected by immunohistochemistry in human intrahepatic bile ducts, where trypsin has been suggested to play a role in biliary cell migration by activating MMPs (21).
In connection with pancreatitis, active trypsin-2 is released into
circulation, where it forms complexes with proteinase inhibitors. The
concentrations of the trypsin-2-1-antitrypsin complex in serum reflects the severity of the pancreatitis (60, 61). It is thought
that activation of trypsinogen plays a pivotal role in development of
pancreatitis by activating other pancreatic enzymes (61, 62). It is
conceivable that activation of proMMP-9 derived from inflammatory cells
(8-10) could also contribute to the tissue destruction seen in this
disease.
The mechanisms of activating the 92-kDa proMMP-9 in vivo have so far remained unclear (42, 43, 45, 53-56), but both proteolytic and nonproteolytic mechanisms are possible (8, 9, 12, 13). Recently, Ca2+ was suggested to be a factor controlling the activation process (63). Organomercurials generate MMP-9 species truncated at the N and the C termini (42, 43, 45), but the conserved PRCGVPD sequence thought to be important for latency is not removed (42, 43, 45). As indicated by the N-terminal sequence of the 77-kDa MMP-9, trypsin-2 released the complete propeptide domain containing the PRCGVPD motif and induced more than a 1000-fold enhancement of enzyme activity. Furthermore, generation of the 77-kDa species correlated with increase in the specific activity. The activation was not mediated by autoproteolysis, as it could be halted by adding TATI. The activation of proMMP-9 by trypsin-2 resembles superactivation of MMPs that is characterized by the generation of enzyme species having Phe as the N-terminal amino residue (15, 64, 65). Crystallographic studies of MMP-8 (66) have indicated that a salt bridge formed between the ammonium group of the N-terminal Phe79 (which corresponds Phe81 and Phe88 in MMP-1 and -9, respectively), and the carboxylate group of Asp232 stabilizes the N-terminal segment enabling optimal substrate binding to the active site. Thus a highly ordered structure of the N terminus seems to characterize the superactivation of MMPs (5, 8, 10). The Phe88-MMP-9 species generated by trypsin-2 is the first demonstration of superactivation of human proMMP by a human serine proteinase.
The Arg87-Phe88 peptide bond of proMMP-9 cleaved by human trypsin-2 is also cleaved by bovine trypsin (67), which, however, induces additional fragmentation of MMP-9 (53-56), e.g. at the Lys73-Ala74. This is located N-terminally in relation to the conserved PRCGVPD motif responsible for MMP latency (5, 10). This fragmentation yields only partially active MMP-9 (53), and we did not detect this fragmentation with human trypsin-2. This may be the reason for the lower specific activity than obtained with human trypsin-2. Some differences between the activation induced by human trypsin-2 and bovine trypsin could be due to contaminating proteases in commercial trypsin preparations. A more likely explanation is, however, that TPCK-treated bovine trypsin exhibits broader substrate specificity than human trypsin-2. Trypsin isoenzymes differ in their substrate specificities and susceptibility to trypsin inhibitors (17, 68). As a proMMP-activator human trypsin-2 seems superior to other serine proteinase activators described so far (5, 8, 10, 14, 16, 67, 69, 70). ProMMP-9 can also be activated by other members of the MMP family, including matrilysin (MMP-7), fibroblast collagenase (MMP-1), gelatinase A (MMP-2), and stromelysin-1 (MMP-3) (53-56, 67, 71). However, these need to be used at a molar ratio of 1:1 to 1:3. Therefore the significance of this activation mechanism in vivo is unclear.
Recent studies have indicated that proMMP-2 is activated on cell membranes by MT-MMPs (2, 4, 72-89), and purified proMMP-2 is not as sensitive to activation by trypsin-like proteases as proMMP-9 (1, 3-5, 10, 44, 65, 88). We found that trypsin-2 converted the 72-kDa proMMP-2 to 62-65-kDa species by cleaving the Arg99-Lys100 peptide bond, but the activation proceeded slowly, and the activity increment was rather small. Experiments with mutant proMMP-2 that does not autodegrade indicated that the major reason for the weak activation is that the cleaved peptide bond is a poor substrate for trypsin-2. Trypsin-2 further fragments MMP-2 in the C-terminal portion. Although the conserved PRCGVPD sequence (5, 8, 10) was removed, the activity of MMP-2 against the studied substrates was low. It appears that some additional changes in the MMP-2 molecule (89), induced by MT-MMPs, are required for complete activation.
Significant amounts of proMMP-9 and-2 in tissues and body fluids occur in complex with either TIMP-1 or TIMP-2, and proMMP-9 has also been found in a 120-kDa complex with NGAL and a 180-kDa complex with fibroblast collagenase as well as an apparent 220-kDa dimer (8, 10, 38, 42-44, 90). The role of the complexation may be to prevent accidental or premature activation of proMMPs. It has been reported that stromelysin-1 (MMP-3) is unable to activate proMMP-9 associated with TIMP-1 (90). We found that trypsin-2 activated proMMPs in complex with TIMPs or NGAL, but TIMP-1 retarded the activation. TIMP-1 was fragmented by trypsin-2. Possibly trypsin-2 must fragment TIMP-1 before it can activate proMMP-9. TIMP-2 is also degraded by bovine trypsin and to lesser extent by plasmin (91). A low molecular weight TIMP form is generated in human bladder carcinoma cell line cultures (91). Thus, the efficacy of trypsin-2 as a proMMP activator may at least in part be due to its ability to degrade the associated TIMPs. Interestingly, trypsin-2 added to neutrophil and fibroblast cell culture medium also converted the latent proMMP-9 and -2 to their active forms as demonstrated by immunoblotting and zymography (not shown).
In conclusion, our results show that tumor cell-derived trypsin-2 is a potential activator of proMMP-9 and is capable of activating both the free and complexed forms of the enzyme. ProMMP-2 is also activated by trypsin-2 to some extent, but the significance of this is unclear. We have shown previously that purified trypsin-2 activates prourokinase, leading to activation of plasminogen (17). Taken together, the results suggest that trypsin-2 may have a previously unrecognized function in promoting the degradation of basement membrane and extracellular matrix associated with tumor cell invasion.
We thank Professor Karl Tryggvason, MBB/ Matrix Biology, Karolinska Institutet, Stockholm, Sweden, for fruitfull discussions.