Inter- (cid:97) -trypsin Inhibitor Bound to Tumor Cells Is Cleaved into the Heavy Chains and the Light Chain on the Cell Surface*

Inter- (cid:97) -trypsin inhibitor (ITI), a human serum protease inhibitor of molecular mass 240 kDa which may re- lease physiological derivatives, has been shown to interact with hyaluronic acid (HA), resulting in pericellular matrix stabilization (Chen, L., Mao, S. J. T., McLean, L. R., Powers, R. W., and Larsen, W. J. (1994) J. Biol. Chem. 269, 28282–28287). The purpose of this study is to determine whether ITI binding to tumor cell surface is medi- ated by urinary trypsin inhibitor (UTI) receptor or cell-associated hyaluronic acid (HA). We demonstrated specific complex formation of the heavy (H) chains of ITI with HA. Binding of the H-chains of ITI to immobilized HA was detected and quantified using colorimetric im-munoassays. Binding was time-, temperature-, and con- centration-dependent. However, UTI and HI-8 (the car-boxyl terminus of UTI) failed to bind to immobilized HA. ITI bound to HA remained functional protease inhibiting activity. After incubation of SMT-cc1 cells with pu- rified biotinylated ITI, biotinylated ITI is bound to the cells, dissociated, and gives rise to the H-chains and UTI on the cell surface. The cell surface receptor-bound UTI derived from ITI may be the result of the limited prote- olysis on the cell surface. In the cells treated with hyaluronidase, diluted in 100 (cid:109) l of PBS, 2% (w/v) BSA was allowed to bind for 2 h at 23 °C. Then the plates were washed three times with PBS. Wells were allowed to react with anti-ITI antibody (1 (cid:109) g/ml, 1 h, 23 °C). This polyclonal antibody reacts with ITI, H-chains, UTI, and HI-8. Specific binding of ITI derivatives (total binding (HA-coated wells) (cid:50) nonspecific binding (BSA-coated wells)) approached saturation at approximately 100 n M ITI or H-chains. Bar , S.D. The experiments were performed at least three times.

Inter-␣-trypsin inhibitor (ITI), a human serum protease inhibitor of molecular mass 240 kDa which may release physiological derivatives, has been shown to interact with hyaluronic acid (HA), resulting in pericellular matrix stabilization ( . The purpose of this study is to determine whether ITI binding to tumor cell surface is mediated by urinary trypsin inhibitor (UTI) receptor or cellassociated hyaluronic acid (HA). We demonstrated specific complex formation of the heavy (H) chains of ITI with HA. Binding of the H-chains of ITI to immobilized HA was detected and quantified using colorimetric immunoassays. Binding was time-, temperature-, and concentration-dependent. However, UTI and HI-8 (the carboxyl terminus of UTI) failed to bind to immobilized HA.

ITI bound to HA remained functional protease inhibiting activity. After incubation of SMT-cc1 cells with purified biotinylated ITI, biotinylated ITI is bound to the cells, dissociated, and gives rise to the H-chains and UTI
on the cell surface. The cell surface receptor-bound UTI derived from ITI may be the result of the limited proteolysis on the cell surface. In the cells treated with hyaluronidase, bound H-chains disappeared from the surface of the cells, while most of the cell surface ITI derivatives was present in deglycosylated UTI (28 kDa). It is suggested that the binding of ITI to the cell surface is mediated by HA on the cells. This was confirmed by the fact that the hyaluronidase-treated cells can abolish the ITI binding. The cell surface UTI formation was inhibited by diisopropyl fluorophosphate, phenylmethylsulfonyl fluoride, and eglin C, suggesting that elastaselike enzyme(s) may be responsible for the UTI formation. Preincubation of the cells with UTI did not decrease in exogenously added ITI on the cell surface. A model for cell surface UTI formation is proposed in which ITI binding to cells from serum used for the culture is followed by the limited proteolysis by trace amounts of active serine proteases, to form cell-surface receptorbound UTI and the H-chains intercalated into cell surface HA. This process is subject to regulation of cell-associated UTI and of stabilization of pericellular matrix.
Inter-␣-trypsin inhibitor (ITI) 1 is the precursor of urinary trypsin inhibitor (UTI), which is one of the Kunitz-type protease inhibitors present in human serum and urine (1)(2)(3)(4). ITI was found to act as a stabilizer of pericellular matrix. The serum factor, identified as an ITI, is a structural component of the matrix. ITI is required to stabilize the fully expanded cumulus cell-oocyte complexes matrix, thus supporting the process of ovulation (5). Huang et al. (6) showed previously that hyaluronic acid (HA) synthesized by cultured fibroblasts firmly bound 85-kDa proteins, which were derived from serum used for the culture and appeared to be covalently linked to HA. The serum-derived HA-associated protein was confirmed to be the heavy (H)-chains of ITI, suggesting that cooperative binding to HA of the H-chains of ITI is required to stabilize the pericellular matrix (7).
Hyaluronic acid is a glycosaminoglycan of the extracellular matrix in most mammalian tissues. It consists of a linear polysaccharide chain with repeating glucuronic acid-(␤1-3)-Nacetylglucosamine-(␤1-4) structure. It is bound by cell surface receptors and extracellular matrix proteins (8). HA is also believed to play an important role in controlling tumor cell growth, migration, invasion, and differentiation. Synthesis and degradation of HA are significantly related to these cellular functions such as cell proliferation. Yoneda et al. (9,10) have shown that HA added exogenously or supplied endogenously by increased synthesis may act as a modulator of fibroblast proliferation.
We showed previously that highly purified human UTI efficiently inhibits soluble and tumor cell-associated plasmin and subsequently prevents tumor cell invasion and metastasis (11). Inhibition of cell-bound plasmin by UTI is associated with significantly reduced tumor cell invasiveness in vitro and a decreased number of metastasis in vivo (11).
Recently we found that some tumor cells have specific binding sites for UTI on the cell surface (12). Exogenously applied UTI may be bound to specific binding sites on the surface of tumor cells. This potentially leads to the build up of a substantial amount of UTI at the surface of the tumor cells (12). There is good evidence that UTI may play an important role in prevention of tumor cell invasion and metastasis (13,14).
In the present study, we investigated whether ITI binding to tumor cell surface is mediated by UTI receptors or by cellassociated HA. Available information on the interaction of ITI derivatives with cells or with immobilized HA is mainly based on the use of biotinylated compounds as ligands. We show that binding of ITI to the cell-associated HA, but not to the UTI receptor, may be essential to the formation of the aggregates comprising the H-chains of ITI and HA which constitutes pericellular matrix.

MATERIALS AND METHODS
Polyclonal Antibody against ITI-Rabbits were immunized with 2 mg of purified ITI in complete Freund's adjuvant followed by booster injections in incomplete Freund's adjuvant every 3 weeks. The serum was purified by protein A chromatography according to the instructions of the manufacturer (Pierce). Polyclonal antibody against ITI (anti-ITI antibody) reacts with native ITI and all of its derivatives including H-chains, UTI, and HI-8. This antibody was used in saturating concentrations.
Cells and Cultures-Promyeloid leukemia U937 cells were cultured in RPMI 1640 supplemented with 10% fetal calf serum (FCS) (15). The choriocarcinoma cell line SMT-cc1 was established from a pulmonary metastatic region of human gestational choriocarcinoma (16). These cells were maintained under an atmosphere of 5% CO 2 in RPMI 1640 (Nissui, Tokyo) medium supplemented with 10% FCS. SMT-cc1 cells were grown to confluency, removed from the flask by adding 4 mM EDTA and 0.01% DNase solution for 2 min at 23°C, followed by gentle tapping of the tray against the bench. Complete detachment was confirmed by direct visualization. Single-cell suspensions were made by repeated pipettings through a 0.4-mm-diameter canula. After this, the cells could be kept in suspension at 4°C by shaking. The cell viability was determined by trypan blue dye exclusion prior to use. SMT-cc1 cells were also used for binding studies.
SMT-cc1 cell monolayers or U937 cell suspensions (1 ϫ 10 6 cells/ml) were grown in the above medium, then in complete medium without serum for 16 h. The cells were then incubated with increasing concentrations of biotinylated ITI (0 -2 M) in RPMI 1640 medium containing ITI-depleted FCS for 16 h at 37°C. Then, the cells were washed twice with PBS containing 0.1% bovine serum albumin (BSA) (washing buffer). The cells were homogenized in washing buffer containing 1 mM phenylmethylsulfonyl fluoride (PMSF), and 1 M each aprotinin, pepstatin, and cystatin. 50 l of anti-ITI antibody-coupled CNBr-activated Sepharose 4B were incubated with the supernatants of cell lysates (16 h, 4°C) to detect the cell-associated ITI and its derivatives. The samples were washed several times with 50 mM PBS, 0.5 M NaCl, and 0.1% Triton X-100, and the final pellet (immobilized immunocomplex) was resuspended in electrophoresis sample buffer, heated, and analyzed by SDS-PAGE followed by Western blot using avidin-peroxidase.
Preparation of Inter-␣-trypsin Inhibitor (ITI) and Its Derivatives-Human ITI was isolated from serum according to Salier et al. (17,18). A highly purified preparation of human UTI with an activity of 2330 units/mg of protein and a molecular mass of 40 kDa (by SDS-PAGE) was kindly supplied by Mochida Pharmaceutical Co., Tokyo, Japan. The covalent structure of the polypeptide chain of the physiological inhibitor UTI has been already determined (2,4). A trypsin-Sepharose affinity column was used in order to obtain purified HI-8 (domain II of UTI). UTI bound to a trypsin-Sepharose column was treated with 80 mg/ml trypsin (37°C, overnight). After washing the column with PBS, trypsindigested UTI bound to the column was eluted with 10 mM HCl containing 0.2 M NaCl. The major active fractions from a trypsin-Sepharose affinity column chromatography were applied to a Sephadex G-100 gel chromatography. Each fraction was pooled and amino-terminal amino acid sequencing was carried out to obtain purified HI-8. A purified preparation of HI-8 with a molecular mass of 8 kDa was supplied by Dr. Dan Sugino (Nissin Food Products Co., Kusatsu, Shiga, Japan).
ITI was biotinylated according to the method of Guesdon et al. (19), using N-hydroxysuccinimidylbiotinamide caproate (Sigma). ITI purified from human serum was treated with hyaluronidase (hyaluronidase/ITI ϭ 1/100 molar ratio; 37°C, 6 h). Hyaluronidase-treated ITI was subjected to preparative gel electrophoresis. Eluted materials were concentrated by ultrafiltration (Amicon) for analysis on reverse phase high performance liquid chromatography. The C18 silica columns were packed for high performance and equilibrated with 5% acetonitrile, 0.1% trifluoroacetic acid before loading. Fractions containing H-chains were pooled.
Hyaluronidase Treatment-SMT-cc1 cell monolayers or U937 cell suspensions were incubated with biotinylated ITI (2 M) for 16 h at 37°C as described above. Hyaluronidase digestion in the cells saturated with biotinylated ITI was performed. The cells (1 ϫ 10 6 cells/ml) or cell monolayers were treated with 20 g/ml testicular hyaluronidase (Sigma) in RPMI with 1% BSA containing the protease inhibitors 1 g/ml pepstatin, 1 g/ml leupeptin, 1 g/ml aprotinin, 1 mM PMSF, and 1 mM EDTA, at 37°C for 60 min. Thereafter, the suspensions or the monolayers were rinsed twice with PBS, 0.1% BSA. The control cells were incubated similarly but without hyaluronidase.
Protease Inhibiting Activity of ITI Bound to HA-The chromogenic substrates S-2222, S-2251, and S-2484 (Kabi Vitrum) were used to determine trypsin, plasmin, and leukocyte elastase activity, respectively. 96-well microtiter plates (Costar, Cambridge, MA) were coated at 4°C overnight with 0.1 mg/ml HA that subsequently blocked with 2% BSA. ITI (10 g/ml) was applied into 96-well plates coated with HA and incubated for 2 h at 23°C. To investigate the inhibiting effects of ITI bound to immobilized HA on trypsin (plasmin or elastase), wells were incubated with 10 nM trypsin (plasmin or elastase) for varying periods of time at 23°C, followed by the addition of S-2222 (S-2251 or S-2484) in PBS, 0.1% BSA, pH 7.4, according to the manufacturer's instructions. The reactions were terminated by adding 10 l of 20% acetic acid. The absorbance at 405 nm was measured with an EIA reader (Model 2550, Bio-Rad).
K i was calculated using the experimentally determined association (k 1 ) and dissociation (k Ϫ1 ) rate constants. k 1 was monitored by allowing proteases and ITI bound to immobilized HA to react for given periods of time before addition of substrate which stopped the association process and then measuring the remaining enzyme activity.  (20,21).
Binding of ITI Derivatives to HA-coated Wells-96-well microtiter plates were coated at 4°C overnight with 0.1 mg/ml HA that subsequently was blocked with 2% BSA. ITI and its derivatives (H-chains (a mixture of H-chain 1, H-chain 2, and H-chain3), UTI, and HI-8) were applied into the 96-well plates coated either with HA or BSA (as a control) and incubated for 2 h at 23°C. Wells were allowed to react with anti-ITI antibody (1 g/ml, 1 h, 23°C). Antibody specifically bound to ITI derivatives was detected with biotin-conjugated anti-rabbit IgG (1.5 g/ml, Dako, Copenhagen, Denmark; 1 h, 23°C), followed by the addition of avidin-peroxidase (1.0 g/ml, Dako, 1 h, 23°C). The reactions with tetramethylbenzidine were terminated, and the A 450 of each well was measured in an EIA reader.
We made an additional experiment to determine the equilibrium dissociation constant of ITI and HA. ITI was radioiodinated using the chloramine-T method (25). The specific activity was adjusted to 1.5 ϫ 10 4 cpm/g ITI by adding unlabeled ITI. HA was conjugated to AH-Sepharose 4B using the carbodiimide method (25). Various amounts of 125 I-labeled ITI were added to HA-Sepharose and incubated (16 h, 23°C). Bound and free 125 I-labeled ITI were separated by centrifugation and counted with a ␥-counter. Nonspecific binding was measured in the presence of an excess amount of unlabeled ITI. The data were analyzed by a Scatchard plot.
Enzyme-linked Immunosorbent Assay (Cell ELISA)-SMT-cc1 cells were grown to semiconfluence in 96-well microtiter plates, washed with PBS, 0.1% (w/v) BSA, and then allowed to react with anti-ITI antibody (1:100 dilution, 2 h, 23°C). Nonspecific binding was ascertained using reactions with rabbit immunoglobulins as a control. Specifically bound anti-ITI antibody was detected with biotin-conjugated anti-rabbit IgG (1.5 g/ml; 1 h, 23°C), followed by the addition of avidin-peroxidase (0.4 g/ml; 1 h, 23°C). Peroxidase activity was detected using a tetramethylbenzidine substrate solution. The reactions were terminated after 10 min at 23°C, and the A 450 (absorbance at 450 nm) of each well was measured in an automated EIA reader.
SDS-Polyacrylamide Gel Electrophoresis and Western Blot-SDS-PAGE (nonreducing conditions) and Western blot were carried out as described elsewhere (15).

Immunoprecipitation of ITI Bound to Tumor
Cells-After addition of purified biotinylated preparations of human ITI to cultures of SMT-cc1 cells growing in a medium containing ITI-depleted serum, biotinylated ITI could be recovered as a bound fraction from the cells by an immunoprecipitation method (Fig. 1, lane 1). Some ITI derivatives, which may correspond to the H-chains of ITI and UTI, were also detected after a 12-h incubation with biotinylated ITI at 37°C (Fig. 1,  lane 2). Membrane-associated ITI was dissociated and gave rise to the H-chains and UTI after a 16-h incubation (Fig. 1, lane 3).
SMT-cc1 cell monolayers were treated with ITI-depleted serum containing 2 M biotinylated ITI and incubated for 16 h at 37°C. After the cells were washed twice with washing buffer, the cells were incubated in the presence or absence of hyaluronidase (20 g/ml, 30 min, 37°C). After washing, the cells were homogenized. Anti-ITI antibody coupled to CNBr-activated Sepharose 4B were incubated with the supernatant of SMT-cc1 cell lysate (16 h, 4°C). Immobilized immunocomplexes were analyzed by SDS-PAGE, followed by Western blot. Fig. 2 shows that in the absence of hyaluronidase, sharp bands of Ͼ200 kDa, 120 kDa, 90 -100 kDa, and 80 kDa as well as a polydisperse one of 40 kDa were obtained in the cell lysate sample. In the hyaluronidase-treated samples, the staining intensity of Ͼ200-kDa, 120-kDa, 90 -100-kDa and 80-kDa bands disappeared, and a new band (deglycosylated UTI, molecular mass ϭ 28 kDa) appeared. It has been reported that the 28-kDa form is the deglycosylated UTI. UTI contains a lowsulfated chondroitin 4-sulfate chain, and its apparent molecular mass upon SDS-PAGE shifts from 40 kDa to 28 kDa upon chondroitinase ABC or hyaluronidase treatment (26). SDS-PAGE of UTI yields a molecular mass of 40 kDa, whereas the amino acid sequences yield values of 15-16 kDa. The discrepancy between these value is due to the chondroitin sulfate chain, whose removal results in a shift of the UTI band to 28 kDa. Hyaluronidase will not digest the whole glycosaminoglycan chain but will leave the innermost disaccharide repeat and the linkage region intact.
Although deglycosylated UTI could be recovered from the cell lysate, the H-chains of ITI could not be detected in the immunoprecipitable proteins (Fig. 2). It is likely that the serumderived HA-associated proteins may be the H-chains of ITI, but without the light chain.
After being treated with hyaluronidase, cell-associated ITI and the H-chains disappeared. This means that H-chains were released into the medium by hyaluronidase treatment. In a parallel experiment, soluble ITI purified from human serum was incubated with hyaluronidase at the same conditions as described above (20 g/ml, 30 min, 37°C). This treatment did not cause the complete disappearance of the 240-kDa band and the appearance of several new bands, major ones of 80 kDa and 28 kDa as well as minor bands of 70 kDa, 60 kDa, and 50 kDa (Fig. 3). In addition, soluble UTI purified from human urine was incubated with hyaluronidase at the same conditions. This treatment caused the incomplete disappearance of the 40-kDa band and the appearance of new bands (molecular mass of 60 kDa and 28 kDa). Broad staining with several bands shown in Fig. 3 (lane 2, d) may represent the difference in the extent of deglycosylation.
We speculated that the cell surface UTI might have been derived from ITI bound on the cell surface, but not from UTI formed in the medium and subsequently bound to the cells. It was confirmed by Western blot analysis that biotinylated ITI does not contain UTI fragments. In a parallel experiment, the cells were treated with the biotinylated ITI (0.2 M) in the presence of an excess of unlabeled ITI (2 M). Under this condition, H-chains and UTI could not be detected in the immunoprecipitable proteins (Fig. 4).
To test which enzymes were responsible for the cell surface ITI cleavage, we added biotinylated ITI to the serum-free culture medium of SMT-cc1 cells in the presence or absence of several protease inhibitors including diisopropyl fluorophosphate, PMSF, aprotinin, leupeptin, pepstatin, cystatin, eglin C, and E64. Several ITI derivatives were found even in the presence of aprotinin, leupeptin, pepstatin, cystatin, and E64. How- FIG. 1. Immunoprecipitation of ITI in cells saturated with biotinylated ITI. SMT-cc1 cells were cultured in RPMI 1640 supplemented with 10% FCS. We confirmed that endogenous UTI is localized on the cell surface or to the pericellular matrix by immunohistochemical staining (12). This is the significance of the usage of the biotinylated ITI. After the cells were washed twice with PBS, 0.1% BSA, the cell monolayers were incubated in serum-free medium for 16 h at 37°C. Then, the cells were treated with RPMI 1640 medium containing ITIdepleted serum plus biotinylated ITI ( [11][12][13][14]). a, ITI; b, H-chains; c, UTI; d, deglycosylated UTI (28 kDa, black arrow). White arrow, UTI degradation/hyaluronidase complex. SDS-PAGE and Western blot were carried out using anti-ITI antibody, followed by biotinanti-rabbit IgG, avidin-peroxidase, and by tetramethylbenzidine substrate solution. ever, virtually no UTI was detected on the cell surface when PMSF or eglin C was added to the medium, while there was a considerable dose-dependent and time-dependent UTI formation in the absence of PMSF or eglin C (Fig. 5). These findings indicated that the cell-bound UTI was formed by cleavage of ITI on the surface of the cells. In a parallel experiment, soluble ITI was incubated with elastase. This treatment caused the complete disappearance of ITI and the appearance of H-chains and UTI. Cell-associated elastase-like enzyme(s) may be responsible for the cell-bound UTI formation (Fig. 6).
ITI and H-chains Have an Ability to Bind to HA-ITI and its derivatives (H-chains, UTI, and HI-8) were applied into the 96-well plates coated either with BSA or with HA and incubated for 2 h at 23°C. Wells were allowed to react with anti-ITI antibody. ITI and the H-chains can bind to the wells coated with HA, while UTI and HI-8 failed to bind to immobilized HA (Fig. 7). All of the ITI derivatives do not bind to the BSA-coated wells (data not shown). In another set of experiments, 96-well plates were coated with ITI, H-chains, UTI, or HI-8, respectively. HA was applied into each plate and incubated for 2 h at 23°C, followed by the addition of peroxidase-conjugated hyaluronic acid binding protein (HRP-HABP, 0.5 g/ml). We confirmed that HA can bind to the wells coated with ITI and H-chains (Fig. 8). In addition, ITI was applied into the 96-well plates coated with HA and incubated for various periods at 4°C, 23°C, or 37°C, indicating that binding was time-and temperature-dependent (Fig. 9). These results suggest that H-chains of ITI appear to bind specifically to HA.
In the next experiment, we investigated whether ITI bound to HA has functional protease inhibiting activity. The relative enzyme activity obtained when trypsin, plasmin, or elastase was titrated with ITI bound to immobilized HA was measured (Table I). ITI bound to HA also strongly inhibited trypsin, plasmin, and elastase activities. Inhibition constant (K i ), a measure of the stability of protease inhibition, varied with a difference of more than 10 between trypsin and plasmin. Trypsin exhibited the most stable inhibition as well as the greatest association rate (Table I).
Recently, the heavy chain 2 of ITI (HC2) has been shown to be easily degraded by HLE whereas HC1 is more resistant (27). The molecular mass of fragments derived by limited proteolysis of UTI with HLE was 22 kDa (13). The 22-kDa protein does not appear to be degraded further in lower molecular mass forms, despite prolongation of the incubation time or increased protease concentrations, indicating that the 22-kDa protein is resistant to further degradation (13). The 22-kDa protein inhibited HLE with essentially the same affinity as native UTI. ITIenzyme complexes can dissociate, which allows for a cleavage of ITI by formerly inhibited enzyme molecules. It has been observed in vitro that such a cleavage releases UTI-based byproducts whose enzyme inhibitory capacity remains intact. We consider that HLE is inhibited by ITI bound to HA, although the same enzyme (2 ng) may digest ITI (5 g) (see Fig. 6).
ITI Binding to Tumor Cells-In the previous study, we confirmed that SMT-cc1 cells have been shown to express functional UTI receptor and binding of biotinylated UTI to the cells occurred in a dose-dependent and saturable manner (12). We investigated biotinylated ITI binding to tumor cells to test whether ITI binds to the cell surface via UTI receptors. Specific ITI binding approached saturation at about 100 nM for 2 h at 37°C (Fig. 10A). We have examined the nature of the binding of biotinylated ITI to the hyaluronidase-treated cells. Fig. 10A indicated clearly that the binding sites for ITI are virtually abolished, when SMT-cc1 cells are pretreated with hyaluronidase. The specific binding of biotinylated ITI to the surface of SMT-cc1 cells was assayed in the presence of increasing concentrations of unlabeled UTI or ITI as the competitors. The cells preincubated for 3 h with various concentrations of UTI (0 -1,000 nM) resulted in no significant decrease in surfacebound biotinylated ITI (Fig. 10B), suggesting that UTI does not compete for the ITI binding sites. In contrast, ITI can compete for the biotinylated ITI binding sites. DISCUSSION We demonstrated that cell-associated HA binds to the Hchains of ITI, but not to UTI or HI-8. They are consistent with the reports in which it has been demonstrated that HA synthesized by cultured fibroblasts firmly bound 85-kDa protein and the serum-derived HA-associated protein is confirmed to be the H-chains of the ITI. The H-chains could be involved in the formation of HA-rich extracellular matrix through their calcium-dependent HA binding activity (5-7, 28, 29). Our results are also supported by a recent study showing that pre-␣-inhibitor stabilizes the expanding cumulus extracellular matrix by binding directly with HA and that pre-␣-inhibitor may serve as a structural protein to organize the function of the cumulus extracellular matrix (28,29). It was also shown that binding of pre-␣-inhibitor and HA is thought a stereospecific charge interaction. The stabilizing action of pre-␣-inhibitor may not be indicated through protease inhibiting activity (28,29).
It was reported that stabilization of the pericellular matrix may require more than a simple binding interaction between pre-␣-inhibitor (or ITI) and HA (22)(23)(24). We found that ITI and UTI interact strongly with hyaluronic acid-binding protein (HABP) (data not shown). This binding interaction was relatively fast. We propose that HABP may further strengthen or stabilize an interaction between ITI and HA.
More recently, it has been reported that the carboxyl-terminal Asp of H-chains was esterified to the C6-hydroxyl group of an internal N-acetylglucosamine of hyaluronic acid chain, demonstrating the covalent binding of proteins to hyaluronic acid (30).
ITI is a unique protease inhibitor that can be proteolyzed by the same enzymes that are inhibited to generate smaller inhibitors. Complete digestion of ITI by each protease was not accompanied by a comparable loss of inhibition of that enzyme or a different enzyme.

TABLE I Kinetic constants for ITI inhibition
The relative enzyme activity obtained when enzymes were titrated with ITI bound to immobilized hyaluronic acid (HA) was measured. 96-well plates were coated with HA. ITI (10 g/ml) was added to each well and incubated (2 h, 23°C). ITI specifically bound to immobilized HA was incubated for varying periods of time with trypsin, plasmin, or human leukocyte elastase (each 10 nM) at 23°C, and the enzyme activities were measured with S-2222, S-2251, or S-2484, respectively. All kinetic experiments were performed in PBS at 37°C. K i , the inhibition equilibrium constant; k 1 , the association rate constant; k Ϫ1 , the dissociation rate constant.
FIG. 9. Time-and temperature-dependent binding of ITI to immobilized HA. 96-well plates were coated with HA (0.1 mg/ml, 16 h, 4°C). Biotinylated ITI (0.1 M, 100 l/well) was added to HA-coated wells and incubated at 37°C (q), 23°C (E), or 4°C (f) for the various times indicated. Biotinylated ITI specifically bound to HA was detected with avidin-peroxidase. Ⅺ, 2 M ITI was added to one sample prior to incubation (37°C, 48 h).